Handbook of the biology of aging a003Document Transcript
Handbook ofThe Biology of Aging
The Handbooks of Aging Consisting of Three Volumes Critical comprehensive reviews ofresearch knowledge, theories, concepts, and issues Editor-in-Chief James E. Birren Handbook of the Biology of Aging Edited by Edward J. Masoro and Steven N. Austad Handbook of the Psychology of Aging Edited by James E. Birren and K. Warner Schaie Handbook of Aging and the Social Sciences Edited by Robert H. Binstock and Linda K. George
Handbook ofThe Biology of Aging Sixth Edition Editors Edward J. Masoro and Steven N. Austad AMSTERDAM • BOSTON • HEIDELBERG • LONDON NEW YORK • OXFORD • PARIS • SAN DIEGO SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO Academic Press is an imprint of Elsevier
ContentsContributors xiForeword xvPreface xviiAbout the Editors xix Section I: Conceptual and Technical Issues 1. Reliability Theory of Aging and Longevity 3 Leonid A. Gavrilov and Natalia S. Gavrilova I. Introduction 3 II. General Overview of the Reliability Theory Approach 5 III. Mortality, Failure, and Aging in Biological and Technical Systems 15 IV. Explanations of Aging Phenomena Using Reliability Theory 21 V. The Idea of High Initial Damage Load: The HIDL Hypothesis 24 VI. Reliability Models of Aging for Biological Systems 28 VII. Evolution of Species Reliability 31 VIII. Conclusions 34 References 35 2. Are Age-Associated Diseases an Integral Part of Aging? 43 Edward J. Masoro I. Introduction 43 II. Concepts of Biological Gerontology 44 III. Age-Associated Diseases 45 IV. Primary Aging, Secondary Aging, and “Normal Aging” 46 V. Evolutionary Theory and Age-Associated Diseases 49 VI. Analysis of Two Major Age-Associated Disease Processes 50 VII. Summary and Conclusions 55 References 56 3. Dietary Restriction, Hormesis, and Small Molecule Mimetics 63 David A. Sinclair and Konrad T. Howitz I. Introduction 63 II. Key Discoveries 65 III. Physiological Effects of DR on Mammals 68 v
vi Contents IV. Mechanisms of DR 70 V. Small-Molecule CR Mimetics 82 VI. Conclusions 89 References 90 4. Hematopoietic Stem Cells, Aging, and Cancer 105 Deborah R. Bell and Gary Van Zant I. Stem Cells 105 II. Stem Cell Aging 108 III. Stem Cells and Cancer 115 IV. Conclusions 119 References 119 5. Mitochondria: A Critical Role in Aging 124 Tamara R. Golden, Karl Morten, Felicity Johnson, Enrique Samper, and Simon Melov I. The Mitochondrion 124 II. Evidence for Increased Oxidative Damage to Mitochondrial Components with Age 125 III. Mitochondrial Dysfunction and Aging 129 IV. Mitochondrial Dysfunction and Age-Associated Disease 134 V. Conclusions 137 References 137 6. P53 and Mouse Aging Models 149 Catherine Gatza, George Hinkal, Lynette Moore, Melissa Dumble, and Lawrence A. Donehower I. Introduction to p53 149 II. p53 and Cellular Senescence 151 III. Linkage of IGF-1, Sir2, and p53 Signaling 154 IV. Mouse Models of Aging 155 V. Mouse Models of Accelerated Aging 158 VI. Mouse Models of Delayed Aging 161 VII. Links to p53 in Mouse Aging Models 162 VIII. Mutant Mouse p53 Models, Aging, and Cancer 164 IX. Influence of p53 on Longevity in Humans 168 X. How Might p53 Influence Organismal Aging? 169 References 171 7. Complex Genetic Architecture of Drosophila Longevity 181 Trudy F. C. Mackay, Natalia V. Roshina, Jeff W. Leips, and Elena G. Pasyukova I. Introduction 181 II. Genome Scan for Quantitative Trait Loci (QTLs) 182 III. Deficiency Complementation Mapping 187 IV. Complementation Tests to Mutations at Positional Candidate Genes 193 V. Linkage Disequilibrium (LD) Mapping 207 VI. Conclusions and Future Prospects 209 References 212
Contents vii 8. Evolutionary Biology of Aging: Future Directions 217 Daniel E. L. Promislow, Kenneth M. Fedorka, and Joep M. S. Burger I. Introduction 217 II. Genetics of Senescence 220 III. From Physiology to Demography 224 IV. Parasites and Immune Function 227 V. Sex, Sexual Selection, and Sexual Conflict 230 VI. Genetic Variation in Natural Populations 232 VII. Conclusions 234 References 235 9. Senescence in Wild Populations of Mammals and Birds 243 Anja K. Brunet-Rossinni and Steven N. Austad I. Introduction 243 II. Evidence of Senescence in Wild Populations 244 III. Patterns of Senescence 255 IV. Methodological Difficulties in Evaluating Senescence in Wild Populations 257 V. Conclusions 260 References 26110. Biodemography of Aging and Age-Specific Mortality in Drosophila melanogaster 267 James W. Curtsinger, Natalia S. Gavrilova, and Leonid A. Gavrilov I. Introduction 267 II. Experimental Evidence for Age-Specific Effects 276 III. Leveling-Off of Mortality Rates 280 IV. Conclusions 289 References 28911. Microarray Analysis of Gene Expression Changes in Aging 295 F. Noel Hudson, Matt Kaeberlein, Nancy Linford, David Pritchard, Richard Beyer, and Peter S. Rabinovitch I. Introduction 295 II. Technical Issues 295 III. Biological Studies 310 IV. Conclusions, Future Directions, and Challenges 326 References 32712. Computer Modeling in the Study of Aging 334 Thomas B. L. Kirkwood, Richard J. Boys, Colin S. Gillespie, Carole J. Procter, Daryl P. Shanley, and Darren J. Wilkenson I. Introduction 334 II. Why Aging Particularly Needs Models 337 III. Different Approaches to Modeling Biological Systems 339 IV. Currently Available Models of Aging 343 V. Models, Data Collection, and Experimental Design 347 VI. Parameter Inference 348 VII. Conclusions 351 References 352
viii Contents Section II: Non-Mammalian Models13. Dissecting the Processes of Aging Using the Nematode Caenorhabditis elegans 360 Samuel T. Henderson, Shane L. Rea, and Thomas E. Johnson I. Introduction 360 II. Biology of C. elegans 362 III. The age-1 Pathway 362 IV. Mutations in Mitochondrial Components 372 V. Caloric Restriction 379 VI. Other Non-Genetic Ways to Extend Life 382 VII. Other Discoveries 384 VIII. Summary 389 References 39014. Genetic Manipulation of Life Span in Drosophila Melanogaster 400 Daniel Ford and John Tower I.Introduction 400 II.Genetic Methods for Manipulating Drosophila Life Span 400 III.Screening for Drosophila Genes Affecting Life Span 405 IV. Specific Genes Used to Extend the Life Span of Drosophila melanogaster 406 V. Conclusions 412 References 41215. Juvenile and Steroid Hormones in Drosophila melanogaster Longevity 415 Meng-Ping Tu, Thomas Flatt, and Marc Tatar I.Introduction 415 II.JH and 20E: Two Major Insect Hormones 416 III.Effects of JH and 20E on Drosophila Aging 418 IV. Candidate Genes Affecting Life Span Through JH and 20E Signaling 422 V. Hormones, Nutrition, and Life Span 433 VI. Hormonal Effects on Stress Resistance and Immunity 436 VII. Conclusions 437 References 44016. A Critical Evaluation of Nonmammalian Models for Aging Research 449 Steven N. Austad and Andrej Podlutsky I.Introduction 449 II.Key Evolutionary Relationships 451 III.Genomic Properties 452 IV. Physiological and Pathophysiological Properties 456 V. Empirically Investigating the Similarities and Differences Among Model Organisms 460 VI. Conclusions 462 References 463
Contents ix Section III: Mammalian Models17. Differential Aging Among Skeletal Muscles 470 Roger J. M. McCarter I. Introduction 470 II. Changes in Muscle Mass and Composition 472 III. Loss of Motor Units with Age 476 IV. Altered Neuromuscular Junctions with Age 479 V. Excitation-Contraction Coupling 480 VI. Mechanical Properties 481 VII. Biochemical Environment 490 VIII. Conclusions 491 References 49318. Aging, Body Fat, and Carbohydrate Metabolism 498 Marielisa Rincon, Radhika Muzumdar, and Nir Barzilai I. Introduction 498 II. Carbohydrate Metabolism and Body Composition in Aging 498 III. Conclusions 505 References 50519. Growth and Aging: Why Do Big Dogs Die Young? 512 Richard A. Miller and Steven N. Austad I. Introduction 512 II. Body Size and Aging in Dogs 512 III. Weight and Longevity in Mice 515 IV. Anecdotal Size-Longevity Reports on Horses 520 V. Height and Longevity in Humans 520 VI. Nutritional Manipulations that Modulate Longevity and Body Size 523 VII. Relation of Size to Longevity Among Different Species 524 VIII. General Discussion: Why Do Big Dogs Die Young, and Is It Worth Figuring This Out? 526 IX. Conclusions 529 References 52920. Growth Hormone, Insulin-Like Growth Factor-1, and the Biology of Aging 534 Christy S. Carter and William E. Sonntag I. Introduction 534 II. Biological Actions of Growth Hormone 535 III. Aging and the Growth Hormone Axis 538 IV. Studies of Growth Hormone/IGF-1 Replacement 540 V. Growth Hormone, IGF-1, and Life Span 549 VI. Pleiotropic Effects of Growth Hormone and IGF-1 556 VII. Conclusion 557 References 558
x Contents21. Aging of the Female Reproductive System 570 Phyllis M. Wise I. Introduction 570 II. Menopause 570 III. Definitions 572 IV. Role of the Ovary in Reproductive Aging 574 V. Role of the Central Nervous System in Female Reproductive Aging 579 VI. Conclusion 586 References 586 Author Index 591 Subject Index 645
Contributors Numbers in parentheses indicate the pages on which the authors’ contributions begin.Steven N. Austad (242, 449, 512), and Gerontology, Wake Forrest University of Texas Health Science University Health Sciences, Winston- Center, Department of Cellular and Salem, NC 27147 Structural Biology, Barshop Institute James W. Curtsinger (267), Department for Longevity and Aging Studies, San of Ecology, Evolution, and Behavior, Antonio, TX 78245-3207 University of Minnesota, MN 55108Nir Barzilai (498), Institute of Aging Lawrence A. Donehower (149), Research, Division of Endocrinology, Department of Molecular Virology Department of Medicine, Albert Einstein and Microbiology, Baylor College of College of Medicine, Bronx, NY 10461 Medicine, Houston, TX 77030Deborah Bell (105), Department of Internal Melissa Dumble (149), Department of Medicine, University of Kentucky, Molecular Virology and Microbiology, Lexington, KY 40536-0093 Baylor College of Medicine, Houston,Richard Beyer (295), Department of TX 77030 Environmental Health, University of Kenneth M. Fedorka (217), Department of Washington, Seattle, WA 98105 Genetics, University of Georgia, Athens,Richard J. Boys (334), School of GA 30602-7223 Mathematics and Statistics, University Thomas Flatt (415), Department of Ecology of Newcastle, Newcastle Upon Tyne, and Evolutionary Biology, Brown NEJ7U, UK University, Providence, RI 02912Anja K. Brunet-Rossini (243), Department Daniel Ford (400), Molecular and of Biology, Cowley Hall, University Computational Biology Program, of Wisconsin-La Crosse, La Crosse, Department of Biological Sciences, WI 54601 University of Southern California,Joep M. S. Burger (217), Department of Los Angeles, CA 90089-1340 Genetics, University of Georgia, Athens, Catherine Gatza (149), Department of GA 30602-7223 Molecular Virology and Microbiology,Christy S. Carter (534), Department of Baylor College of Medicine, Houston, Internal Medicine, Section on Geriatrics TX 77030 xi
xii ContributorsNatalia S. Gavrilova (3, 267), Center on Trudy F. C. Mackay (181), Department of Aging, NORC/University of Chicago, Genetics, North Carolina State Chicago, IL 60637-2745 University, Raleigh, NC 27695Leonid A. Gavrilov (3, 267), Center on Edward J. Masoro (43), Department of Aging, NORC/University of Chicago, Physiology at the University of Texas Chicago, IL 60637-2745 Health Science Center at San Antonio, San Antonio, Texas 78229-3900Colin S. Gillespie (334), Institute for Aging and Health, University of Newcastle, Roger J. M. McCarter (470), Department Newcastle Upon Tyne, NE46BE UK of BioBehavioral Health, Pennsylvania State University, PA 16803Tamara R. Golden (124), Buck Institute for Age Research, Novato, CA 94945 Simon Melov (124), Buck Institute forSamuel T. Henderson (360), Institute age Research, Novato, CA 94945 for Behavioral Genetics, University Richard A. Miller (512), University of of Colorado at Boulder, Boulder, Michigan, Ann Arbor, MI 48109-0940 CO 80309 Lynette Moore (149), Department ofGeorge Hinkal (149), Department of Molecular Virology and Microbiology, Molecular Virology and Microbiology, Baylor College of Medicine, Houston, Baylor College of Medicine, Houston, TX 77030 TX 77030 Karl Morten (124), Witney, Oxon, OX29Konrad T. Howitz (63), BIOMOL 6TD, UK Research Laboratories, Inc., Plymouth Meeting, PA 19462 Radhika Muzumdar (498), Institute of Aging Research, Division ofF. Noel Hudson (295), Houston, TX 77040 Endocrinology, Department of Medicine,Felicity Johnson (124), Royston Park SA Albert Einstein College of Medicine, 5070, Australia Bronx, NY 10461Thomas E. Johnson (360), Institute for Elena G. Pasyukova (181), Institute of Behavioral Genetics, University of Molecular Genetics of the Russian, Colorado at Boulder, Boulder, CO 80309 Academy of Sciences, Moscow 123182, RussiaMatt Kaeberlein (295), Genome Sciences, University of Washington, Seattle, Andrej Podlutsky (449), Geriatric WA 98195 Research, Education and Clinical Center of the South Texas, Veterans HealthThomas B. L. Kirkwood (334), Institute Care System, San Antonio, TX 78284 for Aging and Health, University of Newcastle, Newcastle General David Pritchard (295), Department of Hospital, Newcastle Upon Tyne, Pathology, University of Washington, NE46BE, UK Seattle, WA 98195Jeff W. Leips (181), Department of Carole J. Procter (334), Institute for Biological Sciences, University of Aging and Health, University of Maryland Baltimore County, Baltimore, Newcastle, Newcastle Upon Tyne, MD 21250 NE46BE, UKNancy Linford (295), Department of Daniel E. L. Promislow (217), Department Pathology, Box 357705, University of of Genetics, University of Georgia, Washington, Seattle, WA 98195-7705 Athens, GA 30602-7223
Contributors xiiiPeter S. Rabinovitch (295), Department Marc Tatar (415), Division of Biology and of Pathology, University of Seattle, WA Medicine, Department of Ecology and 98195-7705 Evolutionary Biology, Brown University, Providence, RI 02916Shane L. Rea (360), Institute of Behavioral Genetics, University of Colorado at John Tower (400), Molecular and Boulder, Boulder, CO 80309 Computational Biology Program, Department of Biology, University ofMarielisa Rincon (498), Institute for southern California, Los Angeles, CA Aging Research and Diabetes Center, 90089-1340 Bronx, NY 10461 Meng-Ping Tu (415), DepartmentNatalia V. Roshina (181), Institute of of Genetics and Development, Molecular Genetics of the Russian, College of Physicians and Surgeons, Academy of Sciences, Moscow 123182, Columbia University, New York, Russia NY 10032Enrique Samper (124), Buck Institute for Gary Van Zant (105), Departments of Age Research, Novato, CA 94945 Internal Medicine and Physiology,Daryl P. Shanley (334), Institute for Aging University of Kentucky, Markey Cancer and Health, University of Newcastle, Center, Lexington, KY 40536-0093 Newcastle General Hospital, Newcastle Darren J. Wilkenson (334), School of Upon Tyne, NE46BE, UK Mathematics and Statistics, UniversityDavid A. Sinclair (63), Department of of Newcastle, Newcastle Upon Tyne, Pathology, Boston, MA 02115 NEJ7RU, UKWilliam E. Sonntag (534), Department of Phyllis M. Wise (570), Department Physiology and Pharmacology, Wake of Physiology and Biophysics, Forrest University School of Medicine, University of Washington, Seattle, Winston-Salem, NC 27157-1083 WA 98195
This Page Intentionally Left Blank
ForewordThis volume is one of a series of three of a biological, behavioral, social, andhandbooks of aging: Handbook of the environmental nature.Biology of Aging, Handbook of the Understanding aging is one of the majorPsychology of Aging, and Handbook of challenges facing science in the 21st cen-Aging and the Social Sciences. The tury. Interest in research on aging hasHandbooks of Aging series, now in its become a major focus in science and insixth edition, reflects the exponential the many professions that serve aging pop-growth of research and publications in ulations. Growth of interest in researchaging research, as well as the grow- findings about aging and their interpreta-ing interest in the subject of aging. tion has been accelerated with the growthStimulation of research on aging by gov- of populations of older persons in devel-ernment and private foundation sponsor- oped and developing countries. As moreship has been a major contributor to the understanding has been gained aboutgrowth of publications. There has also genetic factors that contribute to individ-been an increase in the number of uni- ual prospects for length of life and life lim-versity and college courses related to iting and disabling diseases, researchersaging. The Handbooks of Aging have have simultaneously become more awarehelped to organize courses and seminars of the environmental factors that modu-on aging by providing knowledge bases late the expression of genetic predisposi-for instruction and for new steps in tions. These Handbooks both reflect andresearch. encourage an ecological view of aging, in The Handbooks are used by academic which aging is seen as a result of diverseresearchers, graduate students, and pro- forces interacting. These Handbooks canfessionals, for access and interpretation of help to provide information to guide plan-contemporary research literature about ning as nations face “age quakes” due toaging. They serve both as a reference and shifts in the size of their populations ofas organizational tool for integrating a young and older persons.wide body of research that is often cross In addition to the rise in research pub-disciplinary. The Handbooks not only lications about aging, there has been aprovide updates about what is known dramatic change in the availability of sci-about the many processes of aging, but entific literature since the first editionsalso interpretations of findings by well of The Handbooks of Aging were pub-informed and experienced scholars in lished. There are now millions of refer-many disciplines. Aging is a complex ences available on line. This increasesprocess of change involving influences the need for integration of information. xv
xvi ForewordThe Handbooks help to encourage inte- I thank Edward J. Masoro and Stevengration of information from across disci- N. Austad, editors of the Handbook ofplines and methods of gathering data the Biology of Aging, the editors of theabout aging. Handbook of Aging and the Social With so much new information avail- Sciences, Robert H. Binstock and Lindaable, one of the editorial policies has been K. George, and their associate editors,the selection of new chapter authors and Stephen J. Cutler, Jon Hendricks, andsubject matter in each successive edition. James H. Schulz; and my co-editor of theThis allows The Handbooks to present Handbook of the Psychology of Aging,new points of view, to keep current, and to K. Warner Schaie, and the associateexplore new topics in which new research editors, Ronald P. Abeles, Margaret Gatz,has emerged. The sixth edition is thus and Timothy A. Salthouse.virtually wholly new, and is not simply an I also want to express my appreciationupdate of previous editions. to Nikki Levy, Publisher at Elsevier, I want to thank the editors of the indi- whose experience, long term interest,vidual volumes for their cooperation, and cooperation have facilitated the pub-efforts, and wisdom in planning and lication of The Handbooks through theirreviewing the chapters. Without their many editions.intense efforts and experience TheHandbooks would not be possible. James E. Birren
PrefaceThe past five years has been a lifetime in recounting how genetics and demographyaging research. That amount of time has interact in Drosophila studies. Other con-passed since the previous (5th) edition of ceptual chapters cover the complex rela-the Handbook of the Biology of Aging. tion between aging and disease, theDuring the year 2000, when the chapters complexity of the genetic architecture ofof the previous edition were being written, aging, and the use of computer modelingthe research community had at its dis- in the biology of aging. The evolutionaryposal complete gene sequences of only biology of aging crops up in a host oftwo multicellular animals (C. elegans and chapters, but specifically in the chapterDrosophila melanogaster). Since then, we on senescence in nature and the thought-have added to that list mice, rats, humans, ful discussion of where future evolution-and a dozen more species, with another 30 ary studies of aging are likely to go.species “in process.” In the year 2000, we Readers will also be updated on thewere still coming to terms with claims theory of hormesis and life extensionthat a mutation in one gene could extend and how it continues to gain currencylife and preserve health in a mammal. in the field. The roles in the aging processNow the existence of nine such genes has of hematopoietic stem cells, mitochon-been documented in mice, and by the time dria, and the tumor suppressor gene p53you read this, the total will no doubt have are also covered, as are issues in thereached double figures. All this is another application of the still emerging technol-way of saying that we have a lot of ground ogy of DNA microarray analysis to agingto cover in this, the 6th edition of the studies.Handbook of the Biology of Aging. Progress in dissecting the genetics and This edition, as previous ones, provides neuroendocrinology of aging in inverte-in-depth coverage of the latest and best brate models has been so dramatic thatresearch as summarized and interpreted five years of progress is difficult to sum-by leading investigators in the field. This marize. Yet our authors in Section II, andvolume has a particular emphasis on the- several in other chapters, perform thisoretical and technical issues. The first job admirably. At the end of that section,chapter introduces to a wide audience a there is a chapter that asks what theconceptual approach to aging from the limits might be to what we can learnfield of engineering, which has consider- about mammalian aging from the studyable relevance for molecular biologists of invertebrates.who often think in terms of simple bio- The final section is devoted to mam-chemical pathways. That chapter also malian aging—either of particular systemsplaces a premium on refined demographic such as the muscles or the impact ofanalyses of aging, as does a later chapter certain processes such as carbohydrate xvii
xviii Prefacemetabolism or the growth-hormone/IGF-1 Kevin Flurkey, Jeff Halter, Eun-Soo Han,system. The mammalian genetics of Russel Hepple, Peter Hornsby, Pamela L.growth, aging, and their likely relationship Larsen, Marc Mangel, James F. Nelson,also receives a full chapter treatment. The Linda Partridge, T. T. Samaras, and Heidilast chapter covers a topic with timely Scrable. We are also thankful to ourrelevance to the modern human condi- authors, not only for their contributedtion: female reproductive aging and the chapters, but for the alacrity with whichcomplex interplay of brain and ovary that they made helpful comments on oneit involves. another’s chapters. We are grateful to the outside reviewersof our chapters: D. J. Anderson, James R. Edward J. MasoroCarey, James R. Cypser, Caleb E. Finch, Steven N. Austad
About the Editors Edward J. Masoro Dr. Masoro has held faculty positions at Queen’s University (Canada), TuftsDr. Masoro is Professor Emeritus in the University School of Medicine,Department of Physiology at the University of Washington, and MedicalUniversity of Texas Health Science Center College of Pennsyvania. Since 1975,at San Antonio (UTHSCSA) where from Dr. Masoro’s research has focused on theSeptember of 1973 though May of 1991 he influence of food restriction on aging. Heserved as Chairman. He was the founding has served or is serving in an editorialDirector of the Aging Research and role for 10 journals and from JanuaryEducation Center of UTHSCSA, which as 1992 through December 1995, he was theof 2004 became the Barshop Institute of Editor of the Journal of Gerontology:Longevity and Aging Studies. He now Biological Sciences.serves as a member of that institute. Dr. Masoro was the recipient of the1989 Allied-Signal Achievement Award in Steven N. AustadAging Research. In 1990, he received aGeriatric Leadership Academic Award Dr. Austad is currently Professor in thefrom the National Institute on Aging and Department of Cellular and Structuralthe Robert W. Kleemeier Award from the Biology and the Barshop Institute forGerontological Society of America. In Longevity and Aging Studies at the1991, he received a medal of honor from University of Texas Health Sciencethe University of Pisa for Achievements Center at San Antonio. His research cen-in Gerontology, and in 1993, Dr. Masoro ters on the comparative biology of agingreceived the Distinguished Service Award and the development of new animalfrom the Association of Chairmen of models for aging research.Departments of Physiology. In addition, Dr. Austad was the recipient ofhe received the 1995 Irving Wright Award the 2003 Robert W. Kleemeier Awardof Distinction of the American Federation from the Gerontological Society offor Aging Research and the 1995 Glenn America. He is also a Fellow of theFoundation Award. He served as President Gerontological Society of America andof the Gerontological Society of a past Chair of the Biological SciencesAmerica from 1994–1995, as Chairman of Section of that organization. Hethe Aging Review Committee of the received the Phi Kappa Phi/UniversityNational Institute on Aging (NIA), and as of Idaho Alumni Assocation’sChairman of the Board of Scientific Distinguished Faculty Award, the FifthCounselors of the NIA. Nathan A. Shock Award, and shared the xix
xx About the EditorsGeron Corporation-Samuel Goldstein of Aging Cell and Neurobiology ofDistinguished Publication Award with Aging. His trade book, Why We Ageformer graduate student John. P. Phelan. (1997), has been translated into sevenPreviously, he served on the Science languages. He frequently writes and lec-Advisory Board of National Public tures to the general public on topicsRadio. He is currently an Associate related to the biology of aging and ethi-Editor of the Journals of Gerontology: cal issues associated with medicallyBiological Sciences and a Section Editor extending life.
4 L. A. Gavrilov and N. S. Gavrilovaaging and to wonder whether aging may 2. Antagonistic pleiotropy theory:be a property of the system as a whole. In Late-acting deleterious genes may evenother words, perhaps we need to broaden be favored by selection and be activelyour vision and be more concerned with accumulated in populations if they havethe bigger picture of the aging phenom- beneficial effects early in life.enon rather than its details. To illustrate the need for a broad vision, Note that these two theories of agingconsider the following questions: are not mutually exclusive, and both evolutionary mechanisms may operate at• Would it be possible to understand a the same time. The main difference newspaper article by looking at it between the two theories is that in the through an electronic microscope? mutation accumulation theory, genes• Would the perception of a picture in with negative effects at old age accumu- an art gallery be deeper and more late passively from one generation to comprehensive at the shortest possible the next, whereas in the antagonis- distance from it? tic pleiotropy theory, these genes are Evolutionary perspective on aging and actively kept in the gene pool by selec-longevity is one way to stay focused tion (Le Bourg, 2001). The actual relativeon the bigger picture (see recent reviews contribution of each evolutionary mech-by Charlesworth, 2000; Gavrilova & anism to species aging has not yet beenGavrilov, 2002; Martin, 2002; Partridge determined, and this scientific problem& Gems, 2002). Evolutionary explana- is the main focus of current researchtions of aging and limited longevity of in evolutionary biology.biological species are based on two Evolutionary theories demonstratemajor evolutionary theories: the muta- that taking a step back from too-closetion accumulation theory (Charlesworth, consideration of the details over the2001; Medawar, 1946) and the antagonis- “nuts and bolts” of the aging processtic pleiotropy theory (Williams, 1957). helps us to gain a broader vision of theThese two theories can be briefly sum- aging problem. The remaining questionmarized as follows: is whether the evolutionary perspective represents the ultimate general theo- 1. Mutation accumulation theory: retical framework for explanations ofFrom the evolutionary perspective, aging. Or perhaps there may be evenaging is an inevitable result of the more general theories of aging, one stepdeclining force of natural selection further removed from the particularwith age. For example, a mutant gene details?that kills young children will be The main limitation of evolutionarystrongly selected against (will not be theories of aging is that they are applica-passed to the next generation), whereas ble only to systems that reproduce them-a lethal mutation that affects only selves, because these theories are basedpeople over the age of 80 will experience on the idea of natural selection and theno selection because people with this notion of declining force of natural selec-mutation will have already passed it on tion with age.to their offspring by that age. Over However, aging is a very generalsuccessive generations, late-acting phenomenon—it is also observed in tech-deleterious mutations will accumulate, nical devices (such as cars), which doleading to an increase in mortality rates not reproduce themselves in a sexual orlate in life. any other way and which are, therefore,
CHAPTER 1 / Reliability Theory of Aging and Longevity 5not subject to evolution through natu- II. General Overview of theral selection. For this simple reason, Reliability Theory Approachthe evolutionary explanation of agingbased on the idea of declining force of Reliability theory is a body of ideas, math-natural selection with age is not appli- ematical models, and methods aimed atcable to aging technical devices. Thus, predicting, estimating, understanding, andthere may be a more general explanation optimizing the life span and failure distri-of aging, beyond mutation accumulation butions of systems and their componentsand antagonistic pleiotropy theories. (adapted from Barlow & Proschan, 1975). The quest for a general explanation Reliability theory allows researchers toof aging (age-related increase in failure predict the age-related failure kinetics forrates), applicable both to technical devices a system of given architecture (reliabilityand biological systems, invites us to con- structure) and given reliability of its com-sider the general theory of systems fail- ponents.ure known as reliability theory (Barlow& Proschan, 1975; Barlow et al., 1965;Gavrilov, 1978; Gavrilov & Gavrilova, A. Definition of Aging and Non-Aging1991, 2001b, 2003b, 2004b,c; Gavrilov Systemset al., 1978). Reliability theory was historically A reliability-engineering approach todeveloped to describe the failure and biological aging is appealing because itaging of complex electronic (military) provides a common scientific languageequipment, but the theory itself is a (general framework) for scientists work-very general theory based on mathemat- ing in different areas of aging research,ics (probability theory) and a systems helping to overcome disruptive special-approach (Barlow & Proschan, 1975; ization and allowing researchers toBarlow et al., 1965). The theory may understand each other.therefore also be useful in describing and Specifically, reliability theory helpsunderstanding the aging and failure of researchers define more clearly what isbiological systems. It may be useful in aging. In reliability theory, aging isseveral ways: first, by providing a kind of defined as a phenomenon of increasingscientific language (definitions and cross- risk of failure with the passage of timecutting principles), helping researchers (age). If the risk of failure is not increas-create a logical framework for organizing ing with age (the “old is as good as new”numerous and diverse observations on principle), then there is no aging inaging into a coherent picture. Second, it terms of reliability theory, even if thehelps researchers develop an intuition calendar age of a system is increasing.and understanding of the main principles For example, clocks that count timeof the aging process through consider- perfectly are not aging according to reli-ation of simple mathematical models, ability theory (although they have a per-having some features of a real world. fect “biomarker” for their continuousThird, reliability theory is useful for gen- age changes—a displayed time and date).erating and testing specific predictions, Thus, the regular and progressiveas well as deeper analyses of already col- changes over time per se do not consti-lected data. The purpose of this chapter tute aging unless they produce someis to review some applications of reliabil- deleterious outcome (failures). In termsity theory to the problem of biological of reliability theory, the dating problemaging. of determining the system age (time
6 L. A. Gavrilov and N. S. Gavrilovaelapsed since system creation) is differ- concepts that are often confused with eachent from the performance assessment other.1problem of a system’s aging (old becom- In terms of reliability theory, it is con-ing not as good as new). Perfect clocks ceivable to imagine at least theoreticallyhaving an ideal marker of their increas- that some biological species may noting age (time readings) are not aging, but demonstrate aging in certain conditions,progressively failing clocks are aging although their age is always increasing.(although their “biomarkers” of age at “Anti-aging” intervention, according tothe clock face may stop at a “forever reliability theory, is not an oxymoronyoung” date). incompatible with the laws of Nature Moving to a biological example, we (reversing time), but rather refers to anycan say that the formation of regular sea- feasible intervention that delays or pre-sonal tree rings tells us everything about vents “the old becoming not as good astree age but little about tree aging. new.” Later we will show that non-agingMoreover, a progressive disruption of the systems are common both in reliabilityhealthy formation of tree rings would theory and in the real physical world, soindicate tree aging (although this disrup- becoming old is not synonymous withtion obscures the determination of tree aging.age). In terms of reliability theory, the“biomarkers” of age used in forensics to B. Notion of System’s Failureestimate human ages may have nothingto do with human aging, no matter how The concept of failure is important to theaccurate these “biomarkers” are in calen- analysis of a system’s reliability. In reliabil-dar age prediction. For example, an aspar- ity theory, failure is defined as the eventtate racemization in the teeth may be when a required function is terminatedideal for age estimation but not necessar- (Rausand & Høyland, 2003). In other words,ily informative for predicting an increas- failure occurs when the system deviatesing risk of death or other types of failure. from the optimistically anticipated andOn the other hand, loss of motor neuronswith age would be highly relevant to 1The term aging is commonly used bythe problem of human aging, no matter biogerontologists and the public as a synonymhow poorly this loss is correlated with a to the word senescence (progressiveperson’s age. These examples illustrate deterioration with age). This interpretation ofa fundamental difference between bio- aging fits well with the reliability-theorymarkers of age (focused on the dating approach, although the term senescence itselfproblem of accurate age determination) is not common in reliability theory. Theand biomarkers of aging (focused on the problem with the term senescence is that itperformance problem of system deterio- focuses too narrowly on old ages, when theration over time). senescent phenotypes become apparent Thus, reliability theory helps to resolve (e.g., frailty). The term aging is more inclusivea confusion that existed in biological aging because it covers any age-related decline inresearch when some really important performance, even if its starts early in life (e.g., an increase in human death rates after agechanges related to system deterioration 15). See also the second chapter of this book forover time were not properly discriminated a critique of other too-broad definitions offrom other neutral or benign changes aging (Masoro, 2005). It remains to be seenclosely correlated with calendar age. whether the reliability-theory definition ofReliability theory helps to clarify the dif- aging will be universally accepted in the futureference between age (the passage of or will be limited to its use in a specialized waytime) and aging (deterioration with age)— as presented in this chapter.
CHAPTER 1 / Reliability Theory of Aging and Longevity 7desired behavior (it “fails”). Failures are failure outcomes as disease, disability, andoften classified in two groups: death but describe a failure in performance tests for speed, strength, endurance, and so 1. Degradation failures, where the on. For example, it is possible to study thesystem or component no longer functions age dynamics of failure in sports competi-properly, and tions (marathon records, etc.), thereby 2. Catastrophic or fatal failures—the making use of rich sports records for theend of a system’s or a component’s life. purpose of scientific research on aging.Examples of degradation failures in Thus, reliability theory may be useful inhumans would be an onset of different studying “physiological” aging too.types of health impairments, diseases, or Note that a system may have an agingdisabilities, whereas catastrophic or fatal behavior for one particular type of fail-failures obviously correspond to death. ure, but it may remain as good as new forThe notions of aging and failure are some other type of failure. Thus, therelated to each other in the following notion of aging is outcome-specific—itway: when the risk of failure outcomes requires specifying a particular type ofincreases with age (“old is not as good as failure (or group of failures) via which thenew”), this is aging by definition. Note system deteriorates.that according to reliability theory, aging Consequently, legitimate anti-agingis not just growing old; instead, aging is a interventions may be outcome-specificdegradation leading to failure (adverse too, and limited to postponing some spe-health outcomes)—becoming sick, dis- cific adverse health outcomes. Aging isabled, frail, and dead. Therefore, from a likely to be a summary term for manyreliability-theory perspective, the notion different processes leading to variousof healthy aging is an oxymoron, like a types of degradation failures, and each ofhealthy dying or a healthy disease. More these processes deserves to be studiedappropriate terms instead of healthy and prevented.2aging, successful aging, or aging wellwould be delayed aging, postponedaging, slow aging, arrested aging, negligi- 2One may wonder whether hip replacementble aging (senescence), or, hopefully, surgery would qualify as an “anti-agingaging reversal. intervention” according to its description Because the reliability definition of bio- here. The answer to this question is not aslogical aging is linked to health fail- simple as the question itself. It is conceivableures (adverse health outcomes, including that hip replacement therapy may preventdeath), aging without diseases is just as some patients from physical inactivity, stress,inconceivable as dying without death. depression, loss of appetite, malnutrition, andDiseases and disabilities are an integral drug overuse. The result may be that furtherpart (outcomes) of the aging process. Not progression of some diseases and disabilitiesevery disease is related to aging, but every could indeed slow down compared to patientsprogression of disease with age has some who did not receive this treatment. In thisrelevance to aging: aging is a “maturation” case we can say that hip replacement therapy helps to oppose aging for some specific typesof diseases with age. A more detailed dis- of degradation failures in a particular group ofcussion of the relationship between aging patients (very limited anti-aging effect). It isand diseases is provided in the second true, however, that the term anti-agingchapter of this book (Masoro, 2005). intervention is usually associated with hopes Reliability theory also allows us to for something far more radical, such as agingintroduce more “physiological” defini- reversal in the future, applicable to all oldertions of failure that are not limited to such people.
8 L. A. Gavrilov and N. S. GavrilovaC. Basic Ideas and Formulas of This failure law describes “life span” dis- Reliability Theory tribution of atoms of radioactive elements and, therefore, is often called an exponen-Reliability of the system (or its compo- tial decay law. Interestingly, this failurenent) refers to its ability to operate prop- law is observed in many wild populationserly according to a specified standard with high extrinsic mortality (Finch, 1990;(Crowder et al., 1991). Reliability is Gavrilov & Gavrilova, 1991). This kind ofdescribed by the reliability function S(x), distribution is observed if failure (death)which is the probability that a system (or occurs entirely by chance, and it is alsocomponent) will carry out its mission called a “one-hit model” or a “first orderthrough time x (Rigdon & Basu, 2000). kinetics.” The non-aging behavior of a sys-The reliability function (also called the tem can be detected graphically whensurvival function) evaluated at time x is the logarithm of the survival functionjust the probability, P, that the failure decreases with age in a linear fashion.time X is beyond time x, P(X Ͼ x). Thus, Recent studies found that at least somethe reliability function is defined as cells in the aging organism might demon-follows: strate a non-aging behavior.3 Specifically, the rate of neuronal death does not S(x) ϭ P (X Ͼ x) ϭ 1 Ϫ P (X Յ x) increase with age in a broad spectrum ϭ 1 Ϫ F (x) of aging-related neurodegenerative condi- tions (Heintz, 2000). These include 12where F(x) is a standard cumulative dis- different models of photoreceptor degener-tribution function in the probability the- ation, “excitotoxic” cell death in vitro,ory (Feller, 1968). The best illustration loss of cerebellar granule cells in a mousefor the reliability function S(x) is a sur- model, and Parkinson’s and Huntington’svival curve describing the proportion of diseases (Clarke et al., 2000). In this rangethose still alive by time x (the lx column of diseases, five different neuronal typesin life tables). are affected. In each of these cases, the Failure rate, (x), or instantaneous risk rate of cell death is best fit by an expo-of failure, also called the hazard rate, h(x), nential decay law with constant risk ofor mortality force, is defined as the rela- death independent of age (death by chancetive rate for reliability function decline: only), arguing against models of progres- sive cell deterioration and aging (Clarke dSx d lnSx et al., 2000, 2001a). An apparent lack of (x) ϭ Ϫ ϭϪ Sxdx dx cell aging is also observed in the case of amyotrophic lateral sclerosis (ALS)In those cases when the failure rate is (Clarke et al., 2001a), retinitis pigmentosaconstant (does not increase with age), we (Burns et al., 2002; Clarke et al., 2000,have a non-aging system (component) 2001a; Massoff et al., 1990), and idio-that does not deteriorate (does not fail pathic Parkinsonism (Calne, 1994; Clarkemore often) with age: et al., 2001b; Schulzer et al., 1994). (x) ϭ k ϭ const 3Non-aging behavior of cells should not beThe reliability function of non-aging sys- confused with cells’ immortality or theirtems (components) is described by the ability to self-replicate indefinitely. Insteadexponential distribution: non-aging behavior means that the risk of cell death (or loss of function) does not depend on S(x) ϭ S0eϪkx cell age.
CHAPTER 1 / Reliability Theory of Aging and Longevity 9 These observations correspond well plot) to check whether the logarithm ofwith another observation that “an impres- the failure rate is indeed increasing withsive range of cell functions in most organs age in a linear fashion.remain unimpaired throughout the life For technical systems, one of the mostspan” (Finch, 1990, p. 425). These unim- popular models for the failure rate ofpaired functions might reflect the “no- aging systems is the Weibull model, theaging” property known as “old as good power-function increase in failure ratesas new” in survival analysis (Klein & with age x (Weibull, 1939):Moerschberger, 1997, p. 38). Thus, wecome again to the following fundamental (x) ϭ axbquestion about the origin of aging: howcan we explain the aging of a system built for x Ն 0, where a, b Ͼ 0of non-aging elements? This question This law was suggested by Swedishinvites us to think about the possible sys- engineer and mathematician Walodditemic nature of aging and to wonder Weibull in 1939 to describe the strengthwhether aging may be a property of the of materials (Weibull, 1939). It is widelysystem as a whole. We would again like to used to describe the aging and failure ofemphasize the importance of looking at technical devices (Barlow & Proschan,the bigger picture of the aging phenome- 1975; Rigdon & Basu, 2000; Weibull,non in addition to its details, and we will 1951). According to the Weibull law, thesuggest a possible answer to the posed logarithm of failure rate increases lin-question later in this chapter. early with the logarithm of age, with a If failure rate increases with age, we slope coefficient equal to parameter b.have an aging system (component) that This is often used in order to illustratedeteriorates (fails more often) with age. graphically the validity of the WeibullThere are many failure laws for aging sys- law: the data are plotted in the log-logtems, and the most famous one in biology scale (known as the Weibull plot) tois the Gompertz law with exponential check whether the logarithm of the fail-increase of the failure rates with age, ure rate is indeed increasing with the log-which is observed for many biological arithm of age in a linear fashion.species including humans (Finch, 1990; Both the Gompertz and the WeibullGavrilov & Gavrilova, 1991; Gompertz, failure laws have their fundamental1825; Makeham, 1860; Strehler, 1978): explanation rooted in reliability theory (Barlow & Proschan, 1975) and are the (x) ϭ Re␣x only two theoretically possible limiting extreme value distributions for systemswhere x is age, while R and ␣ are positive whose life spans are determined by theparameters. first failed component (Galambos, 1978; We will show later that there are some Gumbel, 1958). In other words, as theexceptions to the Gompertz law and that system becomes more and more complexit is usually applicable within some age (contains more vital components, eachwindows rather than the entire range of being critical for survival), its life spanall possible ages. distribution may asymptotically approach According to the Gompertz law, the one of the only two theoretically possiblelogarithm of failure rates increases lin- limiting distributions—either Gompertzearly with age. This is often used in order or Weibull (depending on the early kinet-to illustrate graphically the validity of the ics of failure of system components). TheGompertz law—the data are plotted in the two limit theorems in the statisticssemi-log scale (known as the Gompertz of extremes (Galambos, 1978; Gumbel,
10 L. A. Gavrilov and N. S. Gavrilova1958) make the Gompertz and the Figure 1.1B presents the dependence ofWeibull failure laws as fundamental as the logarithm of the failure rate on theare some other famous limiting distribu- logarithm of age (Weibull plot) for thetions known in regular statistics, such as Gompertz and the Weibull functions.the normal distribution and the Poisson Note that this dependence is strictly lin-distribution. It is puzzling, however, why ear for the Weibull function (as antici-organisms prefer to die according to the pated) and is concave-up for the GompertzGompertz law, whereas technical devices function. So the Gompertz function lookstypically fail according to the Weibull as if it is accelerating with the logarithmlaw. One possible explanation of this of age when compared to the Weibullmystery is suggested later in this chapter. function. Because of their fundamental impor- This simple graphical method of datatance for describing mortality kinetics, it analysis is useful in practice because itmay be interesting and useful to compare allows researchers to determine easilythese two failure laws and their behavior. whether particular data follow theFigure 1.1A presents the dependence of Gompertz law or the Weibull law (orthe logarithm of the failure rate on age neither).(Gompertz plot) for the Gompertz and Two fundamental differences existthe Weibull functions. Note that this between the Weibull and the Gompertzdependence is strictly linear for the functions. First, the Weibull functionGompertz function (as expected) and is states that the system is immortal atconcave-down for the Weibull function. starting age: when age x is equal to zero,So the Weibull function looks as if it is the failure rate is equal to zero too,decelerating with age when compared to according to the Weibull formula. Thisthe Gompertz function. means that the system should be initiallyA B 1 1 Gompertz function Gompertz function Weibull function Weibull function Failure rate, log scaleFailure rate, log scale 0.1 0.1 0.01 0.01 0.001 0.001 0 20 40 60 80 100 10 100 Age Age, log scaleFigure 1.1 Comparison of the Gompertz and the Weibull functions in different coordinates. (A) Semi-log(Gompertz) coordinates. In this case, the Gompertz function produces a straight line, whereas the Weibullfunction generates a concave-down curve. (B) Log-log (Weibull) coordinates. In this case, the Weibull func-tion produces a straight line, whereas the Gompertz function generates a concave-up curve. By plotting thedeath rate data in these coordinates, it is possible to determine graphically which particular formula pro-vides the best fit (a better straight line) for the empirical data.
CHAPTER 1 / Reliability Theory of Aging and Longevity 11ideal (immortal) in order for the Weibull Gompertz function with different agelaw to be applicable to it. shifts will all be linear and parallel to On the contrary, the Gompertz func- each other in the Gompertz plot.tion states that the system is already vul- The situation is very different for thenerable to failure at starting age: when Weibull function: it is linear in theage x is equal to zero, the failure rate is Weibull plot for only one particular start-already above zero, equal to parameter R ing age, and any shifts in starting age pro-in the Gompertz formula. This means duce a different function. Specifically, if athat partially damaged systems having “true” starting age is larger than assumed,some initial damage load are more likely the resulting function will be a nonlinearto follow the Gompertz failure law, concave-up curve in the Weibull plot, indi-whereas initially perfect systems are cating model misspecification and leadingmore likely to follow the Weibull law. to a bias in estimated parameters. Thus,This profound difference between the researchers choosing the Weibull functiontwo models is often obscured in real life for data analysis first have to resolve anby the period of initially high and then uneasy biological problem: at what agedecreasing juvenile mortality that could does aging start?not be explained by either model. An alternative graceful mathematical Second, there is a fundamental differ- solution to this problem would be toence between the Gompertz and the move from a standard two-parameterWeibull functions regarding their response Weibull function to a more general three-to misspecification of the starting age parameter Weibull function, which has an(“age zero”). This is an important issue additional “location parameter” ␥ (Clark,because in biology there is an ambiguity 1975):regarding the choice of a “true” age, whenaging starts. Legally, it is the moment of (x) ϭ a(x Ϫ ␥)bbirth, which serves as a starting momentfor age calculation. However, from a bio- for x Ͼ ␥, and (x) is equal to zero other-logical perspective, there are reasons to wise.consider a starting age as a date either well Parameters of this formula, includingbefore the birth date (the moment of con- the location parameter ␥, could be esti-ception in genetics, or a critical month of mated from the data through standardpregnancy in embryology), or long after fitting procedures, thus providing athe birth date (the moment of maturity, computational answer to the questionwhen the formation of a body is finally “when does aging start?” However, thiscompleted). computational answer might be shock- From a demographic perspective, the ing to researchers unless they are famil-starting age at which aging begins is iar with the concept of initial damagewhen death rates are the lowest and start load (Gavrilov & Gavrilova, 1991;to grow—this is about 10 years of age for 2001b; 2004a), which will be discussedhumans. The uncertainty in starting age later.has very different implications for data In addition to the Gompertz and theanalysis with the Gompertz and the standard two-parameter Weibull laws, aWeibull functions. For the Gompertz more general failure law was suggestedfunction, misspecification of starting age and theoretically justified using the sys-is not as important because the shift in tem reliability theory. This law is knownthe age scale will still produce the same as the binomial failure law (Gavrilov &Gompertz function with the same slope Gavrilova, 1991; 2001b), and it representsparameter, ␣. The data generated by the a special case of the three-parameter
12 L. A. Gavrilov and N. S. GavrilovaWeibull function with a negative loca- 12tion parameter: 10 (x) ϭ a(x0 ϩ x)b log10(failure rate . 1012) 4 8The parameter x0 in this formula iscalled the initial virtual age of the sys- 6 3tem (IVAS) (Gavrilov & Gavrilova, 1991,2001b). This parameter has the dimen- 4 2sion of time and corresponds to the ageby which an initially ideal system wouldhave accumulated as many defects as a 2real system already has at the starting 1age (at x ϭ 0). In particular, when the 0 20 40 60 80system is initially undamaged, the initial Agevirtual age of the system is zero, and thefailure rate grows as a power function of Figure 1.2 Failure kinetics of systems with different levels of initial damage. Dependence 1 is for an ini-age (the Weibull law). However, as the tially ideal system (with no damage load). Dependenceinitial damage load increases, the failure 2 is for a system with an initial damage load equivalentkinetics starts to deviate from the to damage accumulated by a 20-year-old system.Weibull law, and eventually it evolves to Dependencies 3 and 4 are for systems with an initialthe Gompertz failure law at high levels damage load equivalent to damage accumulated respectively by a 50-year-old system and a 100-year-oldof initial damage load. This is illustrated system. Note that high initial damage load transformsin Figure 1.2, which represents the the Weibull curve into the Gompertz-like straight line.Gompertz plot for the data generated by 1. The Weibull curve for initially ideal systems,the binomial failure law with different (x) ϭ ax10, a ϭ 10Ϫ24 yearϪ1 Graphs for initiallylevels of initial damage load (expressed in damaged systems:the units of initial virtual age). 2. (x) ϭ a(20 ϩ x)10 Note that as the initial damage load 3. (x) ϭ a(50 ϩ x)10increases, the failure kinetics evolves 4. (x) ϭ a(100 ϩ x)10from the concave-down curves typical of Adapted from Gavrilov & Gavrilova, 2004c.the Weibull function to an almost lineardependence between the logarithm of fail- from experimental data through fittingure rate and age (the Gompertz function). procedures.Thus, the binomial failure law unifies twodifferent classes of distribution. The bio- D. System Reliability and the Conceptlogical species dying according to the of Reliability StructureGompertz law may have a high initialdamage load, presumably because of A branch of reliability theory that studiesdevelopmental noise, and a clonal expan- reliability of an entire system givension of mutations that occurred in the reliability of its components and its com-early development (Gavrilov & Gavrilova, ponents’ arrangement (reliability struc-1991, 2001b, 2003a, 2004a). ture) is called system reliability theory The concept of initial virtual age could (Rausand & Høyland, 2003). System relia-be practically useful in analysis and inter- bility involves the study of the overallpretation of survival data because it performance of systems of interconnectedallows us to take into account the initial components. The main objective of sys-damage load of the system when observa- tem reliability is the construction of ations start. Moreover, this concept allows model that represents the times-to-failureus to estimate the initial damage load of the entire system based on the life
CHAPTER 1 / Reliability Theory of Aging and Longevity 13distributions of the components from Thus, the failure of any one componentwhich it is composed. Consideration of results in the failure of the whole system,some basic ideas and models of the such as in Christmas tree lighting chains.system reliability theory is important Figure 1.3A shows a schema of the logicalbecause living organisms may be repre- connectivity of the system in series.sented as structured systems comprised of This type of system is also called aorgans, tissues, and cells. weakest-link system (Ayyub & McCuen, System reliability theory tells us that 2003). In living organisms, many organshow components are arranged strongly and tissues (heart, lung, liver, brain)affects the reliability of the whole system. are vital for the organism’s survival,The arrangement of components that are making them a good example of a series-important for system reliability is also connected component. Thus, the seriescalled reliability structure and is graphi- connection indicates a logical connectivitycally represented by a schema of logicalconnectivity. It is important to understandthat the model of logical connectivityfocuses only on those components that are Arelevant for the functioning ability of thesystem. If the components do not play a ...direct role in a system’s reliability, theyusually are not included in the analyzedreliability structure (Rausand & Høyland, B2003). For example, organs of vision arenot included in the reliability structure ofa living organism if death is the only typeof failure to be analyzed (complete failureof vision does not cause an immediatedeath of the organism). On the other hand,if disability is the type of failure underconsideration, then organs of vision Cshould be included in the schema of relia-bility structure. Therefore, reliabilitystructure does not necessarily reflect aphysical structure of the object. There are two major types of componentarrangement (connection) in the system:components connected in series and com- Dponents connected in parallel (Rausand &Høyland, 2003). Here we consider a simplesystem of n statistically independent com-ponents, where failure of one componentdoes not affect the failure rate of othercomponents of the system.1. Components Connected in Series Figure 1.3 Logical schemas of systems with differ-For a system of n independent compo- ent types of elements connectivity. (A) A system connected in series. (B) A system connected in par-nents connected in series, the system fails allel. (C) A series-parallel system with equal redun-if any one of the components fails, much dancy of system components. (D) A series-parallellike electrical circuits connected in series. system with distributed redundancy.
14 L. A. Gavrilov and N. S. Gavrilovabut not necessarily a physical or an If failure rates of all components are equal,anatomical one. For example, a domi- the failure rate of the system with n com-nant deleterious mutation leading to a ponents is n. It follows from this formulafailure of a diploid organism corresponds that if a system’s components do not ageto a schema of two components (alleles) ( n ϭ const), the entire system connectedconnected in series (in terms of logical in series does not age either.connectivity), although in fact thesealleles are physically located at two dif- 2. Components Connected in Parallelferent homologous chromosomes. The reliability of a system in series A parallel system of n independent compo-(with independent failure events of the nents fails only when all the componentscomponents), Ss, is a product of the relia- fail (such as in electrical circuits connectedbilities of its components: in parallel). The logical structure of a paral- lel system is presented in Figure 1.3B. Ss ϭ p1p2 . . . pn An example of a parallel system is a system with components performing anwhere p1 . . . pn are the reliabilities of the identical function. This function will besystem’s components. destroyed only when all the components This formula explains why complex sys- fail. The number of additional compo-tems with many critical components are nents with the same function in a parallelso sensitive to early failures of their com- structure is called a redundancy or aponents. For example, for a system built of reserve of the system. In living organ-458 critical components, the initial period isms, vital organs and tissues (such as theof a component’s life when its cumulative liver, kidney, or pancreas) consist ofrisk of failure is only 1 percent corre- many cells performing one and the samesponds to the end of a system’s life, when specialized function. A recessive deleteri-99 percent of systems have already failed. ous mutation leading to a failure of aIn other words, by the age when 99 per- diploid organism represents a classiccent of components are still functional example of two components (alleles) con-(p ϭ 0.99), a system built of 458 such criti- nected in parallel.cal components has only a 1 percent For a parallel system with n independ-chance of remaining functional: Ps ϭ ent components, the probability of a(0.99)458 Ϸ 0.01. This discrepancy between system’s failure, Q, is a product of prob-the lifetimes of systems and the lifetimes abilities of failure for its components, q:of their components is increasing furtherwith growing system complexity (num- Qs ϭ q 1q 2 . . . qnbers of critical components). Therefore, ϭ (1 Ϫ p 1)(1 Ϫ p 2) . . . (1 Ϫ pn)the early failure kinetics of components isvery important in determining the failure Hence, the reliability of a parallel sys-kinetics of a complex system for almost tem, Ss, is related to the reliability of itsits entire life. This helps simplify the components in the following way:analysis of complex system failure byfocusing on the early failure kinetics of Ss ϭ 1 Ϫ Qs ϭ 1 Ϫ (1 Ϫ p 1)(1 Ϫ p 2) . . .system components. (1 Ϫ pn) The failure rate of a system connectedin series is a sum of failure rates of its The reliability of a parallel system withcomponents (Barlow et al., 1965): components of equal reliability, p, is: s ϭ 1 ϩ 2 ϩ . . . ϩ n Ss ϭ 1 Ϫ (1 Ϫ p)n
CHAPTER 1 / Reliability Theory of Aging and Longevity 15What is important here is the emergence redundancy for two special cases: (1) theof aging in parallel systems: a parallel redundancy distributed within an organ-system is aging even if it is built of non- ism according to the Poisson law oraging components with a constant failure (2) according to the binomial law. Theyrate (see more details in Section IV). found that the failure rate of such systems In the real world, most systems are initially grows according to the Gompertzmore complex than simply series and law (in the case of the Poisson distributedparallel structures, but in many cases redundancy) or binomial failure law (inthey can be represented as combinations the case of the binomially distributedof these structures. redundancy). At advanced ages, the failure rate for both systems asymptotically approaches an upper limit (mortality3. More Complex Types of Reliability plateau). Reliability models for these sys- Structures tems are described in Section VI.The simplest combination of the two Now when the basic concepts of relia-reliability structures is a series-parallel bility theory are discussed, we maysystem with equal redundancy, shown in proceed to link them to empirical obser-Figure 1.3C. vations on aging and mortality. A general series-parallel system is a sys-tem of m subsystems (blocks) connectedin series, where each block is a set of n III. Mortality, Failure, and Agingcomponents connected in parallel. It turns in Biological and Technicalout that even if the components them- Systemsselves are not aging, the system as a whole A. Failure Kinetics in Biological andhas an aging behavior—its failure rate Technical Systemsgrows with age according to the Weibulllaw and then levels off at advanced There is a striking similarity betweenages (Gavrilov & Gavrilova, 1991, 2001b, living organisms and technical devices in2003b). This type of system is important the general age pattern of their failures—to consider because a living organism can in both cases, the failure rate usually fol-be presented as a system of critical vital lows the so-called “bathtub curve” (seeorgans and tissues connected in series, Figure 1.4).while each organ consists of specialized The bathtub curve of failure rate is acells connected in parallel. The reliability classic concept presented in many text-model for this type of system is described books on reliability theory (Ayyub &in more detail in Section IV. McCuen, 2003; Barlow & Proschan, 1975; Another type of reliability structure, a Rausand & Høyland, 2003). The curveseries-parallel system with distributed consists of three periods. Initially, the fail-redundancy, was introduced by Gavrilov ure rates are high and decrease with age.and Gavrilova (1991). The series-connected This period is called the “working-in”blocks of this system have non-equal period, and the period of “burning-out” ofredundancy (different numbers of elements defective components. For example, theconnected in parallel), and the elements risk for a new computer to fail is oftenare distributed between the system’s higher at the very start, but then thoseblocks according to some particular distri- computers that did not fail initially workbution law (see Figure 1.3D). normally afterwards. The same period Gavrilov and Gavrilova (1991, 2001b) exists early in life for most living organ-studied the reliability and failure rate of isms, including humans, and it is calledseries-parallel systems with distributed the infant mortality period.
16 L. A. Gavrilov and N. S. Gavrilova in the failure rate with age. In most living organisms, including humans, this rise in 10–1 failure rates follows an explosive expo-Hazard rate per day, log scale nential trajectory (the Gompertz curve). 10–2 Drosophila For humans, the aging period lies approxi- 10–3 mately within the interval of 20 to 100 years. 10–4 Thus, there is a remarkable similarity Human in the failure patterns of technical and 10–5 biological systems. This similarity is reinforced further by the fact that at 10–6 extreme old ages there is a fourth period common to both technical devices and 10–7 living organisms (Economos, 1979, 1980, 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 Age in median lifespan scale 1983, 1985). This period is known in biology as a period of late-life mortalityFigure 1.4 “Bathtub” mortality curves for humansand fruit flies. Mortality rates (vertical axis) are cal- leveling-off (Carey & Liedo, 1995;culated in identical units (deaths per day per indi- Clark & Guadalupe, 1995; Economos,vidual) for both species, whereas the age scale 1979; Fukui et al., 1993, 1996; Vaupel et(horizontal axis) is normalized by dividing by the al., 1998), and also as the late-life mortal-median life span of the species to allow data com- ity deceleration law (Fukui et al., 1993,parison (a similar approach to age scaling was usedby Pearl & Miner, 1935, and Carnes et al., 1998). 1996; Khazaeli et al., 1996; Partridge &Mortality for Drosophila melanogaster was calcu- Mangel, 1999).lated using data published by Hall (1969). Mortality Remarkably similar failure patterns offor humans was calculated using the official biological and technical systems indicateSwedish female life table for 1985. that there may be some very general principles of system aging and failure Then follows the second period, called (which will be discussed later), despitethe normal working period, corresponding the obvious differences in specific under-to an age of low and approximately con- lying mechanisms of aging.stant failure rates. This period also existsin humans, but unfortunately it is rather B. Mortality Laws in the Biologyshort (10 to 15 years) and ends too soon.4 of Life Span Then the third period, the aging period,starts, which involves an inexorable rise Attempts to develop a fundamental quan- titative theory of aging, mortality, and life span have deep historical roots. In 1825,4In countries with low child mortality, this the British actuary Benjamin Gompertzage window with minimal death rates has discovered a law of mortality (Gompertz,recently broaden to about 5 to 15 years of age. 1825) known today as the Gompertz lawWhen the death rates in this age interval are (Finch, 1990; Gavrilov & Gavrilova, 1991;presented in logarithmic scale (sensitive to Olshansky & Carnes, 1997; Strehler,outliers that are close to zero levels of 1978). Specifically, he found that the forcemortality), this may create an impression of of mortality increases in geometrical pro-large relative differences in death rates.However the death rates are so low in this age gression with the age of adult humans.group that the absolute differences in death According to the Gompertz law, humanrates are negligible, and it is therefore safe to mortality rates double about every 8 yearsassume that death rates are “approximately of adult age (Finch, 1990; Gavrilov &constant.” Gavrilova, 1991; Gompertz, 1825;
CHAPTER 1 / Reliability Theory of Aging and Longevity 17Makeham, 1860; Strehler, 1978). An expo- human populations from 35 to 70 years ofnential (Gompertzian) increase in death age (Gavrilov & Gavrilova, 1991).rates with age is observed for many biolog- Note that the slope coefficient ␣ charac-ical species including fruit flies terizes an “apparent aging rate” (the rapid-(Drosophila melanogaster) (Gavrilov & ity of age-deterioration in mortality); if ␣Gavrilova, 1991), nematodes (Brooks et al., is equal to zero, there is no apparent aging1994; Johnson, 1987, 1990), mosquitoes (death rates do not increase with age).(Gavrilov, 1980), human lice (Pediculus At advanced ages (after age 80), thehumanus) (Gavrilov & Gavrilova, 1991), “old-age mortality deceleration” takesflour beetles (Tribolium confusum) place: death rates increase with age at(Gavrilov & Gavrilova, 1991), mice a slower pace than expected from(Kunstyr & Leuenberger, 1975; Sacher, the Gompertz-Makeham law. This mor-1977), rats (Gavrilov & Gavrilova, 1991), tality deceleration eventually producesdogs (Sacher, 1977), horses (Strehler, 1978), the “late-life mortality leveling-off” andmountain sheep (Gavrilov, 1980), and “late-life mortality plateaus” at extremebaboons (Bronikowski et al., 2002). old ages (Curtsinger et al., 1992; Gompertz also proposed the first math- Economos, 1979, 1983; Gavrilov &ematical model to explain the exponen- Gavrilova, 1991; Greenwood and Irwin,tial increase in mortality rate with age 1939; Vaupel et al., 1998). Actuaries—(Gompertz, 1825). In reality, failure rates including Gompertz himself—first notedof organisms may contain both non-aging this phenomenon and proposed a logisticand aging terms, as, for example, in formula for mortality growth with age inthe case of the Gompertz-Makeham law order to account for mortality falloff atof mortality (Finch, 1990; Gavrilov & advanced ages (Beard, 1959, 1971; Perks,Gavrilova, 1991; Makeham, 1860; 1932). Greenwood and Irwin (1939) pro-Strehler, 1978): vided a detailed description of this phe- nomenon in humans and even made the (x) ϭ A ϩ Re␣x first estimates for the asymptotic value of the upper limit to human mortalityIn this formula, the first, age-independent (see also the chapter by Curtsinger et al.term (Makeham parameter, A) designates in this volume and review by Olshansky,the constant, “non-aging” component of 1998). According to their estimates, thethe failure rate (presumably due to exter- mortality kinetics of long-lived indi-nal causes of death, such as accidents and viduals is close to the law of radioactiveacute infections), whereas the second, decay with half-time approximatelyage-dependent term (the Gompertz func- equal to 1 year.tion, Re␣x) designates the “aging” compo- The same phenomenon of “almostnent, presumably due to deaths from non-aging” survival dynamics at extremeage-related degenerative diseases such as old ages is detected in many other biolog-cancer and heart disease. ical species. In some species, the mortal- The validity of the Gompertz- ity plateau can occupy a sizable part ofMakeham law of mortality can be illus- their life (see Figure 1.5).trated graphically when the logarithms of Biologists have been well aware of mor-death rates without the Makeham param- tality leveling-off since the 1960s. Foreter ( x Ϫ A) are increasing with age in a example, Lindop (1961) and Sacher (1966)linear fashion (see Figure 1.6). The log- discussed mortality deceleration in mice.linear increase in death rates (adjusted Strehler and Mildvan (1960) consideredfor the Makeham term) with age is indeed mortality deceleration at advanced ages asa very common phenomenon for many a prerequisite for all mathematical models
18 L. A. Gavrilov and N. S. Gavrilova manufactured products (steel samples, industrial relays, and motor heat insula- tors) also demonstrates the same “non-Hazard rate, log scale 0.1 aging” pattern at the end of their “life span” (Economos, 1979). The phenomenon of late-life mortality leveling-off presents a theoretical chal- lenge to many models and theories of 0.01 aging. One interesting corollary from these intriguing observations is that there seems to be no fixed upper limit for indi- vidual life span (Gavrilov, 1984; Gavrilov 0.001 & Gavrilova, 1991; Wilmoth, 1997).5 0 10 20 30 40 This observation calls for a very Age, days general explanation of this apparentlyFigure 1.5 Mortality leveling-off in a population of paradoxical “no aging at extreme ages”4,650 male house flies. Hazard rates were computed phenomenon, which will be discussedusing the life table of the house fly Musca domes-tica, published by Rockstein & Lieberman (1959). later in this chapter. Another empirical observation, the compensation law of mortality, in itsof aging. Later, Economos published a strong form refers to mortality conver-series of articles claiming a priority in the gence, when higher values for the slopediscovery of a “non-Gompertzian para- parameter ␣ (in the Gompertz function)digm of mortality” (Economos, 1979, are compensated by lower values of the1980, 1983, 1985). He found that mortality intercept parameter R in different popu-leveling-off is observed in rodents (guinea lations of a given species:pigs, rats, and mice) and invertebrates(nematodes, shrimps, bdelloid rotifers, ln(R) ϭ ln(M) Ϫ B␣fruit flies, and degenerate medusaeCampanularia Flexuosa). In the 1990s, the where B and M are universal species-phenomenon of mortality deceleration specific invariants.and leveling-off became widely known Sometimes this relationship is alsoafter publications demonstrated mortality called the Strehler-Mildvan correlationleveling-off in large samples of Drosophila (Strehler, 1978; Strehler & Mildvan, 1960),melanogaster (Curtsinger et al., 1992) and although that particular correlation wasmedflies (Ceratitis capitata) (Carey et al., largely an artifact of the opposite biases in1992), including isogenic strains of parameters estimation caused by not tak-Drosophila (Curtsinger et al., 1992; Fukui ing into account the age-independent mor-et al., 1993, 1996). Mortality plateaus at tality component, the Makeham term Aadvanced ages have been observed for (see Gavrilov & Gavrilova, 1991; Golubev,some other insects, including the house 2004). Parameter B is called the species-fly (Musca vicina), blowfly (Calliphoraerythrocephala) (Gavrilov, 1980), fruit flies(Anastrepha ludens, Anastrepha obliqua, 5Note that there is no mathematical limit toAnastrepha serpentine), parasitoid wasp life span, even with exponential growth of(Diachasmimorpha longiacaudtis) (Vaupel mortality force (hazard rate). However, thiset al., 1998), and bruchid beetle mathematical limit exists if the Gompertz law(Callosobruchus maculates) (Tatar et al., of exponential growth is applied to probability1993). Interestingly, the failure kinetics of of death (Gavrilov & Gavrilova, 1991).
CHAPTER 1 / Reliability Theory of Aging and Longevity 19specific life span (95 years for humans), and In those cases when the compensationparameter M is called the species-specific law of mortality is not observed in itsmortality rate (0.5 yearϪ1 for humans). strong form, it may still be valid in itsThese parameters are the coordinates for weak form—i.e., the relative differences inconvergence of all the mortality trajecto- mortality rates of compared populationsries into one single point (within a given tend to decrease with age in many species.biological species), when extrapolated by Explanation of the compensation law ofthe Gompertz function (Gavrilov & mortality is a great challenge for manyGavrilova, 1979, 1991). This means that theories of aging and longevity (Gavrilov &high mortality rates in disadvantaged pop- Gavrilova, 1991; Strehler, 1978).ulations (within a given species) are com- There are some exceptions both frompensated for by a low apparent “aging rate” the Gompertz law of mortality and the(longer mortality doubling period). As a compensation law of mortality thatresult of this compensation, the relative have to be understood and explained.differences in mortality rates tend to There were reports that in some cases,decrease with age within a given biological the organisms die according to thespecies (see Figure 1.6). Weibull (power) law (Eakin et al., 1995; Hirsch & Peretz, 1984; Hirsch et al., 1994; Janse et al., 1988; Ricklefs & 1 Scheuerlein, 2002; Vanfleteren et al., 1998). The Weibull law is more com- monly applicable to technical devices (Barlow & Proschan, 1975; Rigdon & 0.1 Basu, 2000; Weibull, 1951), whereas the Gompertz law is more common in biological systems (Finch, 1990; Gavrilovlog (μ x – A) 0.01 1 & Gavrilova, 1991; Strehler, 1978). 2 Comparative meta-analysis of 129 life tables for fruit flies as well as 285 life 3 tables for humans demonstrates that the 0.001 4 Gompertz law of mortality provides a 5 much better data fit for each of these two biological species compared to the 0.0001 Weibull law (see Gavrilov & Gavrilova, 30 40 50 60 70 80 90 100 Age, years 1991, pp. 55–56, 68–72). Possible explana- tions for why organisms prefer to dieFigure 1.6 Compensation law of mortality.Convergence of mortality rates in different popula- according to the Gompertz law and tech-tions at advanced ages. Death rates (with removed nical devices typically fail according toage-independent external mortality component, the Weibull law are provided elsewhereMakeham parameter A) are plotted in a log scale as (Gavrilov & Gavrilova, 1991, 2001b) anda function of age in the following countries: will be discussed later in this chapter (see1. India, 1941–1950, males; A ϭ 0.00676 yearϪ1 Sections V–VI).2. Turkey, 1950–1951, males; A ϭ 0.00472 yearϪ1 Thus, a comprehensive theory of3. Kenya, 1969, males; A ϭ 0.00590 yearϪ1 species aging and longevity should pro-4. England and Wales, 1930–1932, females; A ϭ 0.00246 yearϪ1 vide answers to the following questions:5. Norway, 1956–1960, females; A ϭ 0.00048 yearϪ1Computed using data from the UN Demographic 1. Why do most biological speciesYearbook (1967; 1975). Adapted from Gavrilov & deteriorate with age (i.e., die more often asGavrilova, 2003b. they grow older), whereas some primitive
20 L. A. Gavrilov and N. S. Gavrilovaorganisms do not demonstrate such a clear to age-related exhaustion of progenitormortality growth with age (Austad, 2001; cells responsible for arterial repairFinch, 1990; Haranghy & Balázs, 1980; (Goldschmidt-Clermont, 2003; Libby,Martinez, 1998)? 2003; Rauscher et al., 2003). Taking 2. Specifically, why do mortality rates these progenitor cells from youngincrease exponentially with age in many mice and adding them to experimentaladult species (Gompertz law)? How should animals prevents atherosclerosis progres-we handle cases when the Gompertzian sion and atherosclerotic inflammationmortality law is not applicable? (Goldschmidt-Clermont, 2003; Rauscher 3. Why does the age-related increase et al., 2003).in mortality rates vanish at older ages? Age-dependent decline in cardiac func-Why do mortality rates eventually tion has recently been linked to the fail-decelerate compared to predictions of the ure of cardiac stem cells to replace dyingGompertz law, demonstrating mortality myocytes with new functioning cellsleveling-off and a late-life mortality (Capogrossi, 2004). Also, it was foundplateau? that aging-impaired cardiac angiogenic 4. How do we explain the so-called function could be restored by addingcompensation law of mortality (Gavrilov endothelial precursor cells derived from& Gavrilova, 1991)? young bone marrow (Edelberg et al., 2002). Any comprehensive theory of human Chronic renal failure is found to beaging has to explain these last three associated with a decreased number ofrules, known collectively as mortality, or endothelial progenitor cells (Choi, 2004).failure, laws. And reliability theory, by People with diminished numbers ofway of a clutch of equations, covers all of nephrons in their kidneys are more likelythem (Gavrilov & Gavrilova, 1991, to suffer from hypertension (Keller et al.,2001b), as will be discussed later. 2003), and the number of glomeruli decreases with human age (Nyengaard & Bendtsen, 1992).C. Loss of Redundancy (e.g., Cell Humans generally lose 30 to 40 per- Numbers) with Age cent of their skeletal muscle fibers by ageMany age changes in living organisms 80 (Leeuwenburgh, 2003), which con-can be explained by cumulative effects of tributes to such adverse health outcomescell loss (either physical or functional) as sarcopenia and frailty. Loss of striatedover time. For example, such very com- muscle cells in such places as the rhab-mon phenomenon as hair graying with dosphincter, from 87.6 percent in aage is caused by depletion of hair follicle 5-week-old child to only 34.2 percent inmelanocytes (Commo et al., 2004). a 91-year-old person, has obvious impli-Melanocyte density in human epidermis cations for urological failure: inconti-declines gradually with age, at a rate of nence (Strasser et al., 2000).approximately 0.8 percent per year A progressive loss of dopaminergic(Gilchrest et al., 1979). Hair graying is a neurons in substantia nigra resultsrelatively benign phenomenon, but in Parkinson’s disease, loss of GABAergiccell loss can also lead to more serious neurons in striatum produces Huntington’sconsequences. disease, loss of motor neurons is respon- Recent studies suggest that such con- sible for amyotrophic lateral sclerosis,ditions as atherosclerosis, atherosclerotic and loss of neurons in the cortex causesinflammation, and consequent throm- Alzheimer’s disease over time (Baizabalboembolic complications could be linked et al., 2003). A study of cerebella from
CHAPTER 1 / Reliability Theory of Aging and Longevity 21normal males age 19 to 84 revealed that IV. Explanations of Agingthe global white matter was reduced by Phenomena Using Reliability26 percent with age, and a selective Theory40 percent loss of both Purkinje andgranule cells was observed in the anterior A. Problem of the Origin of Aginglobe (Andersen et al., 2003). The aging period for most species occu- Furthermore, a 30 percent loss of vol- pies the greater part of their life span,ume, mostly due to a cortical volume therefore any model of mortality mustloss, was found in the anterior lobe, explain the existence of this period. Itwhich is predominantly involved in turns out that the phenomena of mortal-motor control (Andersen et al., 2003). ity increase with age and the subsequentEven if the loss of the volume in various mortality leveling-off is theoretically pre-brain regions is caused by cell atrophy dicted to be an inevitable feature of allrather than cell death, it is still indica- reliability models that consider aging as ative for the loss of redundancy (reserve progressive accumulation of random dam-capacity) with age. age (Gavrilov & Gavrilova, 1991). The Loss of cells with age is not limited to detailed mathematical proof of this pre-the human species; it is observed in other diction for some particular models is pro-animals as well. For example, a nematode vided elsewhere (Gavrilov & Gavrilova,C. elegans demonstrates a gradual, pro- 1991, 2001b) and is briefly described ingressive deterioration of muscle, resem- the next sections of this chapter.bling human sarcopenia (Herndon et al., The simplest schema, which demon-2002). The authors of this study also found strates an emergence of aging in a redun-that the behavioral ability of nematode dant system, is presented in Figure 1.7.was a better predictor of life expectancy If the destruction of an organism occursthan chronological age. not in one but in two or more sequential Interestingly, recent studies have random stages, this is sufficient for thefound that caloric restriction can prevent phenomenon of aging (mortality increase)cell loss (Cohen et al., 2004; McKiernan to appear and then to vanish at older ages.et al., 2004), which may explain why Each stage of destruction corresponds tocaloric restriction delays the onset of one of the organism’s vitally importantnumerous age-associated diseases and structures being damaged. In the simplestcan significantly increase life span inmammals (Masoro, 2003). It should beacknowledged, however, that the hypoth-esis that aging occurs largely because Damageof cell loss remains a subject of debate Defect(Van Zant & Liang, 2003). No redundancy Death In terms of reliability theory, the lossof cells with age is a loss of systemredundancy, and therefore this chapterwill focus further on the effects of redun- Damagedancy loss on system aging and failure.Note that the loss of redundancy doesnot necessarily imply losing cell num- Defectbers, because the loss of cell functional- Redundancy Damage accumulationity (decrease in proportion of functional (aging)cells) could produce the same adverse Figure 1.7 Redundancy creates both damage toler-health outcomes with age. ance and damage accumulation (aging).
22 L. A. Gavrilov and N. S. Gavrilovaorganisms with unique critical structures, The next section provides a mathemat-this damage usually leads to death. ical illustration of these ideas.Therefore, defects in such organisms donot accumulate, and the organisms them-selves do not age—they just die when B. A Simple Model with Paralleldamaged. For example, the inactivation of Structuremicrobial cells and spores exposed to a In this section we show that a systemhostile environment (such as heat) follows built of non-aging components demon-approximately a non-aging mortality strates an aging behavior (mortalitykinetics; their semi-logarithmic survival growth with age) and subsequent mortal-curves are almost linear (Peleg et al., ity leveling-off.2003). This observation of non-aging sur- Consider a parallel system built of nvival dynamics is extensively used in non-aging elements with a constant fail-the calculation of the efficacy of steriliza- ure rate k and reliability (survival) func-tion processes in medicine and food tion eϪkx (see also Figure 1.3B). In thispreservation (Brock et al., 1994; Davis case, the reliability function of the entireet al., 1990; Jay, 1996). A similar non- parallel system is as follows (see alsoaging pattern of inactivation kinetics is Section II.D):often observed for viruses (Andreadis &Palsson, 1997; Kundi, 1999) and enzymes S(x) ϭ 1 Ϫ (1 Ϫ p)n ϭ 1 Ϫ (1 Ϫ eϪkx )n(Gouda et al., 2003; Kurganov, 2002). In more complex systems with many This formula corresponds to the simplestvital structures and significant redun- case when the failure of elements is statis-dancy, every occurrence of damage does tically independent. More complex modelsnot lead to death (unless the environment would require specific assumptions oris particularly hostile). Defects accumu- prior knowledge of the exact type of thelate, therefore, giving rise to the phenom- interdependence in the elements’ failure.enon of aging (mortality increase). Thus, One of such models known as “the modelaging is a direct consequence (tradeoff) of of the avalanche-like destruction” isa system’s redundancies, which ensure described elsewhere (see pp. 246–251 inincreased reliability and an increased life Gavrilov & Gavrilova, 1991).span of more complex organisms. As Consequently, the failure rate of thedefects accumulate, the redundancy in entire system, (x), can be written asthe number of elements finally disap- follows:pears. As a result of this redundancyexhaustion, the organism degenerates dS(x) nk eϪkx(1ϪeϪkx)nϪ1into a system with no redundancy (that (x) ϭ Ϫ ϭ S(x)dx 1 Ϫ (1 Ϫ eϪkx)nis, a system with elements connected inseries, in which any new defect leads to Ϸ nknxnϪ1death). In such a state, no further accu-mulation of damage can be achieved, and when x Ͻ 1/k (early-life period approxi- Ͻthe mortality rate levels off. mation, when 1 Ϫ eϪkx Ϸ kx); The positive effect of a system’s redun-dancy is damage tolerance, which Ϸkdecreases the risk of failure (mortality)and increases life span. However, damage when x Ͼ 1/k (late-life period approxi- Ͼtolerance makes it possible for damage to mation, when 1ϪeϪkx Ϸ 1).be tolerated and accumulated over time, Thus, the failure rate of a system ini-thus producing the aging phenomenon. tially grows as a power function of age
CHAPTER 1 / Reliability Theory of Aging and Longevity 23(the Weibull law). Then, the tempo at Even the simplest parallel system has awhich the failure rate grows declines, specific life span distribution determinedand the failure rate approaches asymptot- entirely by a stochastic nature of theically an upper limit equal to k. Here we aging process. In order to account forshould pay attention to three significant this stochasticity, it was proposed thatpoints. First, a system constructed of researchers use a stochastic variance com-non-aging elements is now behaving like ponent of life span in addition to genetican aging object; that is, aging is a direct and environmental components of pheno-consequence of the redundancy of the typic life span variance (Gavrilov &system (redundancy in the number of Gavrilova, 1991). The stochastic nature ofelements). Second, at very high ages, the a system’s destruction also produces het-phenomenon of aging apparently disap- erogeneity in an initially homogeneouspears (failure rate levels off) as redun- population. This kind of induced hetero-dancy in the number of elements geneity was observed in isogenic strains ofvanishes. The failure rate approaches an nematodes in which aging resulted in sub-upper limit, which is totally independent stantial heterogeneity in behavioral capac-of the initial number of elements but ity among initially homogeneous wormscoincides with the rate of their loss kept in controlled environmental condi-(parameter k). Third, the systems with tions (Herndon et al., 2002).different initial levels of redundancy The graph shown in Figure 1.8 depicts(parameter n) will have very different mortality trajectories for five systemsfailure rates in early life, but these differ- with different degrees of redundancy.ences will eventually vanish as failure System 1 has only one unique elementrates approach the upper limit deter- (no redundancy), and it has the highestmined by the rate of elements’ loss failure rate, which does not depend on(parameter k). Thus, the compensation age (no aging). System 2 has two ele-law of mortality (in its weak form) is an ments connected in parallel (one extraexpected outcome of this illustrative element is redundant), and the failuremodel. rate initially increases with age (aging Note also that the identical parallel appears). The apparent rate of aging cansystems in this example do not die simul- be characterized by a slope coefficienttaneously when their elements fail by that is equal to 1. Finally, the failure ratechance. A common view in biology is the levels off at advanced ages. Systems 3, 4,idea that all members of a homogeneous and 5 have, respectively, three, four, andpopulation in a hypothetical constant five elements connected in parallel (two,environment should have identical life three, and four extra elements are redun-spans (die simultaneously) so that the sur- dant), and the failure rate initiallyvival curve of such a population would increases with age at an apparent aginglook like a rectangle. This idea stems rate (slope coefficient) of 2, 3, and 4,from the basic principles of quantitative respectively. Finally, the mortality trajec-genetics, which assume implicitly that tories of each system level off atevery animal of a given genotype has the advanced ages at exactly the same uppersame genetically determined life span so limit to the mortality rate.that all variation of survival time around This computational example illustratesa genotype mean results from the envi- the following general ideas: (1) Aging is aronmental variance. George Sacher (1977) direct consequence of a system’s redun-pointed out that this concept is not appli- dancy, and the expression of aging iscable to longevity and used an analogy directly related to the degree of a system’swith radioactive decay in his arguments. redundancy. Specifically, an apparent
24 L. A. Gavrilov and N. S. Gavrilova mortality, an additional idea should be Risk of failure, in dimensionless units, log scale 100 taken into account (see the next section). 10–2 V. The Idea of High Initial Damage Load: The HIDL Hypothesis 10–4 In 1991, Gavrilov and Gavrilova sug- gested an idea that early development of 10–6 no redundancy, no aging living organisms produces an exception- redundancy = 1, Aging rate (slope) = 1 redundancy = 2, Aging rate (slope) = 2 ally high load of initial damage, which is redundancy = 3, Aging rate (slope) = 3 redundancy = 4, Aging rate (slope) = 4 comparable with the amount of subse- 10–8 quent aging-related deterioration accu- 0.01 0.1 1 10 Age, in dimensionless units, log scale mulating during the rest of the entireFigure 1.8 Failure kinetics of systems with differ- adult life.ent levels of redundancy. The dependence of the log- This idea of High Initial Damage Loadarithm of mortality force (failure rate) on the (the HIDL hypothesis) predicts that evenlogarithm of age in five systems with different levels small progress in optimizing the earlyof redundancy (computer simulation experiment). developmental processes can potentiallyDependence 1 is for the system containing only oneunique element (no redundancy). Dependence 2 is result in a remarkable prevention offor the system containing two elements connected many diseases in later life, postponementin parallel (degree of redundancy ϭ 1). Dependencies of aging-related morbidity and mortality,3, 4, and 5 are for systems containing, respectively, and significant extension of healthy lifethree, four, and five elements connected in parallel span (Gavrilov & Gavrilova, 1991, 2001b,(with increasing levels of redundancy). The scales formortality rates (vertical axis) and for age (horizontal 2003b, 2004a). Thus, the idea of early-lifeaxis) are presented in dimensionless units (/k) for programming of aging and longevity maymortality rates and kx for age to ensure the general- have important practical implications forizability of the results (invariance of graphs on fail- developing early-life interventions inure rate of the elements in the system, parameter k). promoting health and longevity.Also, the log scale is used to explore the systembehavior in a wide range of ages (0.01 to 10 units) Although this idea may look like aand failure rates ( 0.00000001 to 1.0 units). Adapted counterintuitive assumption, it fits wellfrom Gavrilov & Gavrilova, 2003b, 2004c. with many empirical observations on massive cell losses in early development. For example, the female human fetus at age 4 to 5 months possesses 6 to 7 mil-relative aging rate is equal to the degree lion eggs (oocytes). By birth, this numberof redundancy in parallel systems. (2) All drops to 1 to 2 million and declines evenmortality trajectories tend to converge further. At the start of puberty in normalwith age so that the compensation law of girls, there are only 0.3 to 0.5 millionmortality is observed. (3) All mortality eggs—only 4 to 8 percent of initial num-trajectories level off at advanced ages, and bers (Finch & Kirkwood, 2000; Gosden,a mortality plateau is observed. Thus, the 1985; Wallace & Kelsey, 2004). It is nowmajor aging phenomena (aging itself, well established that the exhaustion ofthe compensation law of mortality, late- the ovarian follicle numbers over time islife mortality deceleration, and late-life responsible for menopause (reproductivemortality plateaus) are already observed aging and failure), and women havingin the simplest redundant systems. higher ovarian reserve have longer repro-However, to explain the Gompertz law of ductive life span (Wallace & Kelsey,
CHAPTER 1 / Reliability Theory of Aging and Longevity 252004). When young ovaries were trans- ontogenesis through a process of self-planted to old post-reproductive mice, assembly out of de novo forming andtheir reproductive function was restored externally untested elements (cells).for a while (Cargill et al., 2003). This Moreover, because organisms are formedexample illustrates a general idea that from a single cell, any defects in early lifeaging occurs largely because of cell loss, such as deleterious mutations or deleteri-which starts early in life. ous epigenetic modifications (i.e., genomic Massive cell losses in early develop- imprinting) can proliferate by mechanismment create differences between organ- of clonal expansion, forming large clustersisms in the numbers of remaining cells, of damaged cells. This proliferation ofwhich can be described by the binomial defects during development of biologicaldistribution or, at particularly high systems can make them highly damagedlevels of cell losses, by the Poisson distri- by the time they are formed.bution. This, in turn, can produce a The second property of organisms isquasi-exponential (Gompertzian) pattern the extraordinary degree of miniaturiza-of age-specific mortality kinetics with tion of their components (the micro-a subsequent mortality deceleration scopic dimensions of cells as well as the(Gavrilov & Gavrilova, 1991). In some molecular dimensions of informationspecies, including C. elegans, the devel- carriers like DNA and RNA), permittingopmental loss of cells seems to be very the creation of a huge redundancy in theprecise. If adult individuals are identical number of elements. Thus, we canin the initial numbers of functional cells, expect that for living organisms, in dis-one can expect that mortality kinetics in tinction to many technical (manufac-such cases would be closer to the tured) devices, the reliability of theWeibull law rather than the Gompertz system is achieved not by the high initiallaw. However, the Gompertz law also quality of all the elements but by theircan be expected for initially identical huge numbers (redundancy).organisms if the critical vital organs The fundamental difference in thewithin a given organism differ by their manner in which the system is formedcell numbers (Gavrilov & Gavrilova, (external assembly in the case of techni-1991, pp. 252–264; 2001b). cal devices and self-assembly in the case Mathematical proof for this statement of biological systems) has two importantwas published elsewhere (see Gavrilov & consequences. First, it leads to the macro-Gavrilova, 1991, pp. 264–272) and will be scopicity of components in technicalbriefly summarized in Section VI. Here devices compared to biosystems, sincewe concentrate on the substantive dis- technical devices are assembled “top-cussion of the idea of high initial damage down” with the participation of a macro-load in biological systems. scopic system (man) and must be suitable for this macroscopic system to use (i.e., commensurate with man). Organisms, onA. Differences Between Biological and the other hand, are assembled “bottom- Technical Systems up” from molecules and cells, resultingBiological systems are different from tech- in an exceptionally high degree of minia-nical devices in at least two aspects. The turization of the component parts.first fundamental feature of biological sys- Second, since technical devices aretems is that, in contrast to technical (arti- assembled under the control of man, theficial) devices that are constructed out of opportunities to pretest componentspreviously manufactured and tested com- (external quality control) are incompara-ponents, organisms form themselves in bly greater than in the self-assembly of
26 L. A. Gavrilov and N. S. Gavrilovabiological systems. This inevitably leads Jonason et al., 1996; Khrapko et al., 2004;to organisms being “littered” with a great Nekhaeva et al., 2002).number of defective elements. As a Loss of telomeres, eventually leading toresult, the reliability of technical devices such outcomes as genomic instability, cellis assured by the high quality of elements death (apoptosis), cell senescence, and per-(fault avoidance), with a strict limit on haps to organism’s aging (Kim et al., 2002),their numbers because of size and cost also begins before birth, and it is directlylimitations, whereas the reliability of bio- linked to DNA replication during cell divi-logical systems is assured by an excep- sions, which are particularly intensive attionally high degree of redundancy to early stages of growth and developmentovercome the poor quality of some ele- (Collins & Mitchell, 2002; DePinho &ments (fault tolerance). Wong, 2003; Forsyth et al., 2002; Kim et al., 2002). In humans, the length of telomeres declines precipitously before theB. Some Examples Illustrating the HIDL age of 4 (by 25 percent) and then declines Hypothesis further very slowly (Hopkin, 2001).The idea that living organisms start their Another potential source of extensivelives with a large number of defects is initial damage is the birth process itself.not a new one. Biological justification for During birth, the future child is firstthis idea was discussed by Dobzhansky, deprived of oxygen by compression ofwho noted that, from the biological per- the umbilical cord (Moffett et al., 1993)spective, Hamlet’s “thousand natural and suffers severe hypoxia (often withshocks that flesh is heir to” was an ischemia and asphyxia). Then, just afterunderestimate and that in reality “the birth, a newborn child is exposed to oxida-shocks are innumerable” (1962, p. 126). tive stress because of acute reoxygenation Recent studies have found that troubles while starting breathing. It is known thatin human life start from the very begin- acute reoxygenation after hypoxia mayning: the cell-cycle checkpoints (which produce an extensive oxidative damageensure that cells will not divide until through the same mechanisms that alsoDNA damage is repaired and chromoso- produce ischemia-reperfusion injury (IRI)mal segregation is complete) do not oper- and asphyxia-reventilation injury (Martinate properly at the early, cleavage stage in et al., 2000). Asphyxia is a common occur-human embryos (Handyside & Delhanty, rence in the perinatal period, and asphyx-1997). This produces mosaicism of the ial brain injury is the most commonpreimplantation embryo, where some neurologic abnormality in the neonatalembryonic cells are genetically abnormal period (Dworkin, 1992) that may manifest(McLaren, 1998), with potentially devas- in neurologic disorders in later life. Thetating consequences in later life. brain damage that occurs after asphyxia Most of the DNA damage caused by may cause long-term neurological conse-copy errors during DNA replication also quences in full-term infants (Volpe, 2000)occurs in early life because most cell and lead to cerebral palsy, epilepsy, anddivisions happen in early development. mental retardation (Hack & Fanaroff,As a result of extensive DNA damage in 2000; Hjalmarsson et al., 1988, pp. 28–36).early development, many apparently nor- Perhaps the rare geniuses are simply thosemal tissues of young organisms have a lucky persons whose early-life brain dam-strikingly high load of mutations, includ- age was less extensive than the “normal”ing abundant oncogenic mutations and level. Thus, using Hamlet’s metaphor,frequent clones of mutated somatic cells we may conclude that humans “suffer(Cha et al., 1994; Deng et al., 1996; the slings and arrows of outrageous
CHAPTER 1 / Reliability Theory of Aging and Longevity 27fortune” and have “a sea of troubles” from Sacher & Duffy (1979). It was foundthe very beginning of their lives. that the six traits (body weight and resting It follows from this concept of HIDL and average metabolic rates both at youngthat even small progress in optimizing and old ages) of parental genotypesthe processes of ontogenesis and increas- explained 95 percent of variation in meaning the numbers of initially functional life span between 16 F1-hybrid mice geno-elements can potentially result in a types, whereas the same six traits of hybridremarkable fall in mortality and a signifi- mice themselves explained only 25 percentcant improvement in life span. This opti- of variation in their mean life spanmistic prediction is supported by (Gavrilov & Gavrilova, 1991, pp. 175–182).experimental evidence (in laboratory The highest mean life span was observedmice) of increased offspring life span if in the progeny of those parents who hadfuture parents are fed antioxidants, the lowest resting metabolic rate at youngwhich presumably result in protection of age. This observation is consistent with aparental germ cells against oxidative hypothesis that the differences in progenydamage (Harman & Eddy, 1979). life span could be linked to the rates of From this point of view, parental charac- oxidative DNA damage in parental germteristics determining the quality of the cells. Interestingly, the resting metabolicgametes, and especially maternal charac- rate measured in young progeny itself wasteristics determining the accuracy of the not predictive for progeny life span (seeearly stages of development, would be Table 1.1).expected to have significant influence on Thus, certain parameters (such as rest-the life span of the offspring, which may ing metabolic rate at young age) meas-be in some cases even stronger than the ured in parents could be better predictorseffect of these same properties of the off- of progeny life span compared to thespring themselves. In other words, the reli- same parameters measured among theability concept leads us to a paradoxical progeny itself.conjecture: sometimes a better predictor The concept of high initial damage loadof life span may be found not among the also predicts that early-life events maycharacteristics of the organism itself but affect survival in later adult life throughamong the characteristics of its parents. modulating the level of initial damage. Gavrilov & Gavrilova (1991) tested this This prediction proved to be correct forcounterintuitive prediction using data on such early-life indicators as parental agelife span and metabolic characteristics of at a person’s conception (Gavrilov &21 inbred and F1-hybrid mouse genotypes Gavrilova, 1997, 2000, 2003a; Gavrilova(several hundred mice) published by et al., 2003) and the month of a person’s Table 1.1 Parental Resting Metabolic Rates at Young Age Are Better Predictors of Life Span of Mice Progeny Than the Resting Metabolic Rates (RMR) Measured in Progeny Itself*Variable Regression Standard Error t-value p-value CoefficientMaternal RMR Ϫ1054 252 Ϫ4.18 0.001Paternal RMR Ϫ795 254 Ϫ3.13 0.009Progeny RMR 42 205 0.20 0.843*Parameter values for linear regression of progeny life span on parental and progeny resting metabolic rate measured atyoung age (RMR) for 16 genotypes of F1-hybrid mice. Computed using data published by Sacher & Duffy (1979).
28 L. A. Gavrilov and N. S. Gavrilovabirth (Doblhammer & Vaupel, 2001; The failure rate of a simple parallel sys-Gavrilov & Gavrilova, 1999, 2003a; tem built of non-aging elementsGavrilova et al., 2003). The month of increases with age, although the initialbirth may influence a person’s life span failure kinetics follows the Weibull lawthrough early-life exposure to seasonal rather then the Gompertz law. This limi-vitamin deficiencies and seasonal infec- tation of the model is rooted in thetions during critical periods of child devel- assumption that the system is built ofopment (Gavrilov & Gavrilova, 2001a). It initially ideal structures where all ele-is known that deficiencies of vitamins ments are functional from the outset.B-12, folic acid, B-6, niacin, and vitamins This standard assumption may be justi-C and E appear to mimic radiation in fied for technical devices manufactureddamaging DNA by causing single- and from pretested components, but it is notdouble-stand breaks, oxidative lesions, or justified for living organisms, presum-both (Ames, 2004). Vitamin deficiencies ably replete with defects, for the reasonshad profound seasonality in the past when described earlier. Gavrilov and Gavrilovacontemporary adults were born, and these (1991) proposed a family of reliabilitydeficiencies may be particularly harmful models based on the idea of initial dam-at the early stages of human development age load, which allows us to explain all(Gavrilov & Gavrilova, 2001a). three major laws of biological aging and There is mounting evidence now in sup- mortality: the Gompertz law, the late-lifeport of the idea of fetal origins of adult deceleration law, and the compensationdegenerative diseases (Barker, 1998; Kuh & law of mortality (mortality convergenceBen-Shlomo, 1997; Leon et al., 1998; at advanced ages). A brief description ofLucas et al., 1999) and early-life program- these models is provided below.ming of aging and longevity (Gavrilov &Gavrilova, 1991, 2001a, 2003a,b). Women A. Highly Redundant System Repletemay be particularly sensitive to early-life with Defectsexposures because they are mosaics of twodifferent cell types (one with active pater- The simplest model in this family ofnal X chromosome and one with active reliability models is the model of amaternal X chromosome), and the pattern series-parallel structure with distributedof this mosaic is determined early in life. redundancy within the organism (seeIndeed, this conjecture of stronger female Gavrilov & Gavrilova, 1991, pp. 252–264;response to early-life exposures is con- 2001b). If distribution of subsystemsfirmed for such early-life predictors of within the organism according to ini-adult life span as paternal age at a person’s tially functional elements can beconception (Gavrilov & Gavrilova, 1997, described by the Poisson law because of2000, 2003a, 2004a; Gavrilova et al., 2003) high initial damage load, then the failureand the month of a person’s birth rate of such series-parallel systems can(Gavrilov & Gavrilova, 2003a; Gavrilova be approximated initially by the expo-et al., 2003). nential (Gompertz) law with subsequent mortality leveling-off. According to this model, the compen- VI. Reliability Models of Aging sation law of mortality is inevitable if for Biological Systems the “true aging rate” (relative rate of ele- ments’ loss) is similar in different popu-It was demonstrated in Section IV that lations of a given species (presumablythe aging phenomenon emerges when a because of homeostasis—stable bodysystem gains some redundancy (reserves). temperature, glucose concentration, etc.).
CHAPTER 1 / Reliability Theory of Aging and Longevity 29This suggested explanation leads to an made about possible initial differencesinteresting testable prediction that for between the organisms themselves. In alower organisms with poor homeostasis, more general case, the population hetero-there may be deviations from the com- geneity needs to be taken into accountpensation law of mortality. because there is a large variation in the numbers of cells for the organisms of the same species (Finch & Kirkwood, 2000).B. Partially Damaged Redundant System The model of heterogeneous redundantThe simplest model, which was described systems (Gavrilov & Gavrilova, 1991, pp.earlier, assumed an extremely high level 264–272) demonstrates that taking intoof initial damage load. In a more general account the heterogeneity of the popula-model, the distribution of subsystems in tion also provides an explanation for allthe organism according to the number of the basic laws of mortality. This modelinitially functional elements is described assumes that there is a distribution ofby the binomial rather than Poisson dis- organisms with regard to their initialtribution. In this case, the failure rate of a redundancy levels (e.g., number of func-system initially follows the binomial fail- tional cells) within a population underure law (Gavrilov & Gavrilova, 1991, study. If this distribution is close to either2001b). the binomial or the Poisson distribution, Thus, if the system is not initially then a quasi-exponential (Gompertzian)ideal, the failure rate in the initial period pattern of mortality increase with age isof time grows exponentially with age, expected initially, with subsequent mor-according to the Gompertz law. A numer- tality leveling-off (Gavrilov & Gavrilova,ical example provided in Figure 1.2 shows 1991, pp. 264–272).that increase in the initial system’s dam- Figure 1.9 shows computed data for aage load (initial virtual age) converts the model in which organisms have a differ-observed mortality trajectory from the ent number of elements (connected inWeibull to the Gompertz one. The model parallel) and are distributed by theiralso explains the compensation law of redundancy levels according to themortality and mortality leveling-off later Poisson distribution law, with the meanin life (see Gavrilov & Gavrilova, 1991, number of elements equal to .2001b). Note that the dependence of the loga- Thus, both reliability models described rithm of failure rate on age is almost ahere provide an explanation for a general linear one, indicating that the initial fail-pattern of aging and mortality in biologi- ure kinetics is indeed close to thecal species: the exponential growth of Gompertz law. This initial Gompertzianfailure rate in the initial period, with the period of failure rate growth can be easilysubsequent mortality deceleration and extended for the organism’s entire lifeleveling-off, as well as the compensation span in the case of more complex sys-law of mortality. tems with many vital components (built of parallel elements), each being critical for survival (serial connection of a largeC. Heterogeneous Population number of components; see Section II.D). of Redundant Organisms Figure 1.9 also demonstrates that theThe models discussed so far examined a populations of organisms with highersituation in which series-connected vital mean levels of redundancy (parameter )subsystems (blocks) have varying degrees have lower death rates, but these deathof redundancy within each organism, rates are growing steeper with age (thewhile no additional assumptions were compensation law of mortality).
30 L. A. Gavrilov and N. S. Gavrilova 101 the death of the organism. In this case,Hazard rate in dimensionless units, log scale 100 we arrive at a schema for the accumula- tion of damage in which the rate of dam- 10–1 age flow (equal to the product of the 10–2 number of elements and their failure rate) turns out to be practically constant 10–3 in view of the incommensurability of the 10–4 number of elements and the permitted 10–5 number of defects (see Gavrilov & 1 - Poisson λ = 1 Gavrilova, 1991, pp. 272–276). 10–6 2 - Poisson λ = 5 3 - Poisson λ = 10 This model also allows us to take into 10–7 4 - Poisson λ = 15 account the influence of living condi- 5 - Poisson λ = 20 10–8 tions on the value of the critical number 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 of defects incompatible with the survival Age, in dimensionless units of the organism. The key to the solutionFigure 1.9 Failure kinetics in mixtures of systems of this problem is the replacement of thewith different redundancy levels for the initial age parallel connection hypothesis (assumedperiod. The dependence of failure rate as a functionof age in mixtures of parallel redundant systems in previous models) with the more realis-having Poisson distribution by initial numbers of tic assumption that there exists a criticalfunctional elements (mean number of elements, number of defects incompatible with the ϭ 1, 5, 10, 15, 20). The scales for mortality rates survival of the organism. In this case, it(vertical axis), and for age (horizontal axis) are pre- is natural to expect that under harshersented in dimensionless units (/k) for mortalityrates, and kx for age, to ensure the generalizability conditions, the critical number of defectsof the results (invariance of graphs on failure rate of leading to death might be less than underthe elements in the system, parameter k). more comfortable living conditions. In particular, in the wild, when an animal is deprived of care and forced to acquire its own food as well as to defend itselfD. Accumulation of Defects with against predators, the first serious dam- Constant Rate of Damage Flow age to the organism can lead to death. ItAnother reliability model of aging is is therefore not surprising that the mor-obtained after a critical reinterpretation tality of many animals (in particular,of the assumptions underlying the previ- birds) is practically independent of age inously described models. In fact, these the wild. This follows directly from themodels contain an assumption that the single-stage destruction of the organismdeath of the organism occurs only when model. On the other hand, the greaterall the elements in a block fail. It is pos- the number of defects the organism cansible that this hypothesis may be justi- accumulate while remaining alive, thefied in a number of cases for some of the greater its life span will be.organism’s subsystems. However, in the The standard model of defect accumula-majority of cases, the hypothesis seems tion with constant rate of damage flowcontentious. For example, it is hard to predicts that at the initial moment inimagine that a single surviving liver cell time, mortality grows according to a(hepatocyte) can assume the functions of power (Weibull) law of mortality. If wean entire destroyed liver. Significantly assume that distribution of living organ-more realistic is the hypothesis that the isms according to the number of defectssystem initially contains an enormous they have is described by the Poisson law,number of elements that greatly exceeds then at the initial moment in time, thisthe critical number of defects, leading to model leads to the binomial law of
CHAPTER 1 / Reliability Theory of Aging and Longevity 31mortality. In this model, the compensa- et al., 1978; Strehler & Mildvan, 1960).tion law of mortality can be obtained both The existence of a multitude of compet-as a result of variation in the degree to ing models is therefore compatible withwhich the organisms are initially dam- the reliable and meaningful interpreta-aged, and of variation in the critical num- tion of a number of mortality phenom-ber of defects, dependent on the harshness ena because variability of models doesof living conditions (see Gavrilov & not preclude their agreement on a num-Gavrilova, 1991, pp. 272–276). ber of issues. Second, if different models Summarizing this brief review of lead to the same formulas—for examplereliability models, note the striking simi- the binomial law of mortality—thislarity between the conclusions of the merely makes the problem of interpret-considered models. All these models pre- ing results more complicated for the the-dict a mortality deceleration, no matter oretician, but significantly facilitates thewhat assumptions are made regarding work for the experimenter. Indeed, forinitial population heterogeneity or its the analysis of data, it is preferable to usecomplete initial homogeneity. Moreover, a formula that is supported not by a sin-these reliability models of aging produce gle model but by a whole family of mod-mortality plateaus as inevitable out- els that encompass a wide spectrum ofcomes for any values of considered possible situations.parameters. The only constraint is thatthe elementary steps of the multistagedestruction process of a system should VII. Evolution of Speciesoccur by chance only, independent of Reliabilityage. The models also predict that an ini-tially homogeneous population will Reliability theory of aging is perfectlybecome highly heterogeneous for risk of compatible with the idea of biologicaldeath over time (acquired heterogeneity). evolution, and it helps to identify keyThe similarity of conclusions obtained components that may be importantfrom several different models means that for evolution of species reliability andit is impossible on the basis of the estab- durability (longevity): initial redundancylished mortality phenomena to uncover levels, initial damage load, rate ofthe correct mechanism behind the age- redundancy loss, and repair potential.related destruction of organisms, and fur- Moreover, reliability theory helps evolu-ther studies are necessary to discriminate tionary theories explain how the age ofbetween the competing models. onset of diseases caused by deleterious One can of course derive no pleasure mutations could be postponed to laterfrom this circumstance, but there are ages (as suggested by the mutation accu-two reasons that give ground for opti- mulation theory of aging)—this could bemism. First, the different models seem to easily achieved by a simple increase inlead to very similar interpretations of the initial redundancy levels (e.g., initialcertain mortality phenomena. For exam- cell numbers).ple, the compensation law of mortality is From the reliability perspective, theonly possible when the relative rate of increase in initial redundancy levels isredundancy loss is the same in all popu- the simplest way to improve survival atlations of a given species. This interpre- particularly early reproductive ages (withtation of the compensation law of gains fading at older ages). This exactlymortality is not only a feature of the matches with the higher fitness prioritymodels described in this chapter but also of early reproductive ages emphasizedof other models (Gavrilov, 1978; Gavrilov by evolutionary theories. Evolutionary
32 L. A. Gavrilov and N. S. Gavrilovaand reliability ideas also help to under- to increase their “intrinsic” reliabilitystand why organisms seem to “choose” a was less intensive compared to humans.simple but short-term solution to the This traditional evolutionary paradigmsurvival problem through enhancing the also says that birds live longer and havesystems redundancy, rather than a more lower “intrinsic” death rates because ofpermanent but complicated solution adaptation to flight, which improvedbased on rigorous repair (with a potential their survival in the wild and increased afor negligible senescence). selection pressure to further decrease It may be interesting and useful to “intrinsic” death rates (Austad, 2001).compare failure rates of different biologi- Thus, if a bird (say, a finch) is comparedcal species expressed in exactly the same to a similar-sized shorter-lived mammalunits of risk (risk of death per individual (say, a rat), the expected picture shouldper day). Returning back to the earlier be similar to Figure 1.4: a bird shouldFigure 1.4, we can notice with some sur- have lower death rates than a rat both inprise that the death rates of young vigor- the beginning and in the end of theirous fruit flies kept in protected lives. Interestingly, this prediction oflaboratory conditions is as high as among traditional evolutionary paradigm couldvery old people! This indicates that fruit be confronted with an alternative predic-flies from the very beginning of their tion expected from a reliability paradigm.lives have very unreliable design com- Reliability paradigm predicts that birdspared to humans. This observation also should be very prudent in redundancy oftells us that young organisms of one their body structures (because it comesbiological species may have the same with a heavy cost of additional weight,failure risk as old organisms of another making flight difficult). Therefore, aspecies—that is, being old for humans is flight adaptation should force the birds toas good as being young for fruit flies. evolve in a direction of high reliability ofNote that at extreme old ages, the death their components (cells) with low levelsrates of fruit flies are well beyond human of redundancy (cell numbers). Thus, relia-death rates (see Figure 1.4). In terms bility paradigm predicts that “intrinsic”of reliability models, this observation death rates of birds in protected environ-suggests that fruit flies are made of less ments should be rather high at youngreliable components (presumably cells), ages (because of low redundancy levels),which have higher failure rates compared whereas at old ages their death ratesto human cells. might be much lower than in other We can ask ourselves a question: is it a species (because of higher reliability ofgeneral rule that shorter-lived biological their cells). This suggestion of higher reli-species should always have higher death ability of avian cells agrees with therates within comparable age groups (say, recent findings of increased resistancewithin “young” or “old” age groups)? of these cells to oxidative stress andTraditional evolutionary theories suggest DNA damage (Holmes & Ottinger, 2003;that indeed shorter-lived species should Ogburn et al., 1998, 2001).have higher “intrinsic” death rates in Figure 1.10 presents data on “intrin-protected environments because these sic” mortality in Bengalese finches asrates are shaped in evolution through compared to rats for both species livingselection pressure by death rates in the in protected environments.wild (predation, starvation, etc.). In other Note that the death rates in both specieswords, defenseless fruit flies in the wild are very close to each other at young ages,experience much higher death rates than but later a mortality divergence occurs sodo humans; therefore a selection pressure that old birds have much lower death rates
CHAPTER 1 / Reliability Theory of Aging and Longevity 33 10–2 Rat Hazard rate, day –1, log scale 10–2 10–3 Hazard rate, day–1 , log scale 10–4 Horses 10–3 Bengalese Finch 10–5 10–6 10–4 Humans 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 10–7 Age in a median lifespan scale 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 Age in median lifespan scaleFigure 1.10 Comparative mortality of rats andBengalese finches expressed in the same units of Figure 1.11 Comparative mortality of humans andmortality (per day). Data sources: Bengalese finch, horses expressed in the same units of mortality (persurvival data for 39 birds of both sexes in captivity day). Data sources: Humans, official Swedish(Eisner, 1967); Rats, survival data for 2,050 female female life table for 1985; Horses, survival data forrats kept in a laboratory (Schlettwein-Gsell, 1970). 2,742 thoroughbred mares (Comfort, 1958).than old rats. These observations match similar in horses and humans, whereasthe predictions of a reliability paradigm the rates of the early stages of the agingbut not a traditional evolutionary explana- process are vastly different in these twotion discussed earlier (the initial death species.rates for birds are much higher than These intriguing findings demonstrateexpected from the traditional evolutionary that there are promising opportunities forperspective). Thus, a comparison of further comparative studies on the evolu-species death rates may be useful for test- tion of species reliability and the merginging different ideas on evolution of species of the reliability and evolutionary theo-aging and reliability. ries of aging. This reliability-evolutionary Another interesting observation comes approach could be considered as furtherfrom a comparison of humans with development of the earlier compara-horses (see Figure 1.11). It could be tive studies of species aging and life his-expected that shorter-lived horses should tories (Austad, 1997, 2001; Gavrilov &have higher death rates than humans. Gavrilova, 1991; Holmes et al., 2001;However, this prediction is only valid for Promislow, 1993, 1994).young ages. The data demonstrate that Another promising direction for the reli-an old horse is not much different from ability-evolutionary approach is to studyan old man in terms of mortality risk the selection effects for high performance(see Figure 1.11). This example is oppo- (e.g., the ability to avoid predators). Classicsite to observations on finch–rat compar- evolutionary theories predict that anisons and demonstrates a mortality exposure to high extrinsic mortality dueconvergence between two different bio- to predation should produce shorter-livedlogical species (man and horse) at older species (Charlesworth, 2001; Medawar,ages. In terms of reliability models, this 1946; Williams, 1957). This predictionobservation may indicate that the rates could be confronted with the oppositeof the late stages of body destruction are prediction of reliability theory, which says
34 L. A. Gavrilov and N. S. Gavrilovathat elimination of weak individuals by 2. An apparent aging rate orpredators should increase species life span expression of aging (measured as agebecause of selection for better perform- differences in failure rates, includingance and lower initial damage load. death rates) is higher for systems withInterestingly, recent studies found an higher redundancy levels (all otherincreased life span of guppies evolving in a things being equal). This is an importanthigh predation environment (Reznick et issue because it helps put a correctal., 2004) as predicted by the reliability perspective over fascinating observationstheory of aging. of negligible senescence (no apparent aging) observed in the wild and at extreme old ages. Reliability theory explains that some cases of negligible VIII. Conclusions senescence may have a trivial mechanism (lack of redundancies in theExtensive studies of aging have produced system being exposed to a challengingmany important and diverse findings, environment) and, therefore, will notwhich require a general theoretical frame- help to uncover “the secrets of negligiblework for them to be organized into a senescence.” The studies of negligiblecomprehensive body of knowledge. senescence make sense, however, when As demonstrated by the success of evo- death rates are also demonstrated to belutionary theories of aging, based on a negligible.general idea of the declining force of nat- Reliability theory also persuades aural selection with age, quite general the- re-evaluation of the old belief that agingoretical considerations can in fact be is somehow related to limited economicvery useful and practical when applied or evolutionary investments in systemsto aging research (Charlesworth, 2000; longevity. The theory provides a com-Le Bourg, 2001; Martin, 2002; Partridge & pletely opposite perspective on thisGems, 2002). issue—aging is a direct consequence of In this chapter, we attempted to go one investments into systems reliability andstep further in the search for a broader durability through enhanced redundancy.explanation of aging (not limited to bio- This is a significant statement because itlogical species only) by applying a gen- helps us to understand why the expressioneral theory of systems failure known as of aging (differences in failure ratesreliability theory. Considerations of this between younger and older age groups)theory lead to the following conclusions: may be actually more profound in more 1. Redundancy is a key notion for complex redundant systems (organisms)understanding aging, and the systemic designed for higher reliability.nature of aging in particular. Systems 3. During the life course, organisms arethat are redundant in numbers of running out of cells (Gosden, 1985;irreplaceable elements do deteriorate Herndon et al., 2002), losing reserve(i.e., age) over time, even if they are capacity (Bortz, 2002; Sehl & Yates, 2001),built of non-aging elements. The and this redundancy depletion explainspositive effect of system redundancy is the observed “compensation law ofdamage tolerance, which decreases mortality” (mortality convergence atmortality and increases life span. older ages) as well as the observed late-lifeHowever, damage tolerance makes it mortality deceleration, leveling-off, andpossible for damage to be tolerated and mortality plateaus.accumulated over time, thus producing 4. Living organisms seem to be formedthe aging phenomenon. with a high load of initial damage, and
CHAPTER 1 / Reliability Theory of Aging and Longevity 35therefore their life span and aging particular manifestations of aging (types ofpatterns may be sensitive to early-life failure). Therefore, we should not be dis-conditions that determine this initial couraged by only partial success of eachdamage load during early development. particular intervention, but instead we canThe idea of early-life programming of appreciate an idea that we do have soaging and longevity may have important many opportunities to oppose aging inpractical implications for developing numerous different ways.early-life interventions promoting health Thus, the efforts to understand theand longevity. routes and the early stages of age-related degenerative diseases should not be dis-The theory also suggests that aging carded as irrelevant to understandingresearch should not be limited to studies “true” biological aging. On the contrary,of qualitative changes (like age changes the attempts to build an intellectual fire-in gene expression) because changes in wall between biogerontological researchquantity (numbers of cells and other and clinical medicine are counterproduc-functional elements) could be an impor- tive. After all, the main reason people aretant driving force in the aging process. In really concerned about aging is because itother words, aging may be largely driven is related to health deterioration andby a process of redundancy loss. increased morbidity. The most important The reliability theory predicts that a pathways of age changes are those thatsystem may deteriorate with age even if make older people sick and frail (Bortz,it is built from non-aging elements with 2002).constant failure rate. The key issue here Reliability theory suggests generalis the system’s redundancy for irreplace- answers to both the “why” and theable elements, which is responsible for “how” questions about aging. It explainsthe aging phenomenon. In other words, “why” aging occurs by identifying theeach particular step of system destruc- key determinant of aging behavior: sys-tion/deterioration may seem to be ran- tem redundancy in numbers of irreplace-dom (no aging, just occasional failure by able elements. Reliability theory alsochance), but if a system failure requires a explains “how” aging occurs, by focusingsequence of several such steps (not just a on the process of redundancy loss oversingle step of destruction), then the sys- time as the major mechanism of aging.tem as a whole may have an aging Aging is a complex phenomenon (Sehl &behavior. Yates, 2001), and a holistic approach Why is this important? Because the using reliability theory may help ana-significance of beneficial health-promot- lyze, understand, and, perhaps, control it.ing interventions is often undermined by We suggest, therefore, adding reliabilityclaims that these interventions are not theory to the arsenal of methodologicalproven to delay the process of aging approaches applied in aging research.itself, but instead that they simply delayor “cover-up” some particular manifesta- Acknowledgmentstions of aging. This work was supported in part by grants In contrast to these pessimistic views, from the National Institute on Aging.the reliability theory says that there maybe no specific underlying elementary agingprocess itself; instead, aging may be largely Referencesa property of a redundant system as a Ames, B. N. (2004). Supplements and tuningwhole because it has a network of destruc- up metabolism. Journal of Nutrition, 134,tion pathways, each being associated with 3164S–3168S.
36 L. A. Gavrilov and N. S. GavrilovaAndersen, B. B., Gundersen, H. J., & Bronikowski, A. M., Alberts, S. C., Altmann, Pakkenberg, B. (2003). Aging of the human J., Packer, C., Carey, K. D., & Tatar, M. cerebellum: a stereological study. Journal (2002). The aging baboon: comparative of Comparative Neurology, 466, 356–365. demography in a non-human primate.Andreadis, S., & Palsson, B. O. (1997). Proceedings of the National Academy of Coupled effects of polybrene and calf serum Sciences of the USA, 99, 9591–9595. on the efficiency of retroviral transduction Brooks, A., Lithgow, G. J., & Johnson, T. E. and the stability of retroviral vectors. (1994). Mortality rates in a genetically Human Gene Therapy, 8, 285–291. heterogeneous population ofAustad, S. N. (1997). Comparative aging and Caenorhabditis elegans. Science, 263, life histories in mammals. Experimental 668–671. Gerontology, 32, 23–38. Burns, J., Clarke, G., & Lumsden, C. J. (2002).Austad, S. N. (2001). Concepts and theories of Photoreceptor death: spatiotemporal aging. In E. J. Masoro & S. N. Austad (Eds.), patterns arising from one-hit death kinetics Handbook of the biology of aging (5th ed., and a diffusible cell death factor. Bulletin of pp. 3–22). San Diego, CA: Academic Press. Mathematical Biology, 64, 1117–1145.Ayyub, B. M., & McCuen, R. H. (2003). Calne, D. B. (1994). Is idiopathic Probability, statistics, reliability for Parkinsonism the consequence of an event engineers and scientists. Boca Raton, or a process? Neurology, 44, 5–10. FL: Chapman & Hall/CRC. Capogrossi, M. C. (2004). Cardiac stem cellsBaizabal, J. M., Furlan-Magaril, M., Santa- fail with aging: a new mechanism for the Olalla, J., & Covarrubias, L. (2003). Neural age-dependent decline in cardiac function. stem cells in development and regenerative Circulation Research, 94, 411–413. medicine. Archives of Medical Research, Carey, J. R., & Liedo, P. (1995). Sex-specific 34, 572–588. life table aging rates in large medflyBarker, D. J. P. (1998). Mothers, babies, and cohorts. Experimental Gerontology, 30, disease in later life (2nd ed.). London: 315–325. Churchill Livingstone. Carey, J. R., Liedo, P., Orozco, D., &Barlow, R. E., & Proschan, F. (1975). Vaupel, J. W. (1992) Slowing of mortality Statistical theory of reliability and life rates at older ages in large medfly cohorts. testing. Probability models. New York: Science, 258, 457–461. Holt, Rinehart and Winston. Cargill, Sh. L., Carey, J. R., Muller, H.-G., &Barlow, R. E., Proschan, F., & Hunter, L. C. Anderson, G. (2003). Age of ovary (1965). Mathematical theory of reliability. determines remaining life expectancy in old New York: Wiley. ovariectomized mice. Aging Cell, 2, 185–190.Beard, R. E. (1959). Note on some Carnes, B. A., Olshansky, S. J., & Grahn, D. mathematical mortality models. (1998). An interspecies prediction of the In G. E. W. Wolstenholme & M. O’Connor risk of radiation-induced mortality. (Eds.), The lifespan of animals (pp. Radiation Research, 149, 487–492. 302–311). Boston: Little, Brown. Cha, R. S., Thilly, W. G., & Zarbl, H. (1994).Beard, R. E. (1971). Some aspects of theories N-nitroso-N-methylurea-induced rat of mortality, cause of death analysis, mammary tumors arise from cells with forecasting and stochastic processes. In preexisting oncogenic Hras1 gene mutations. W. Brass (Ed.), Biological aspects of Proceedings of the National Academy of demography (pp. 57–68). London: Taylor & Sciences of the USA, 91, 3749–3753. Francis. Charlesworth, B. (2000). Fisher, Medawar,Bortz, W. M. (2002). A conceptual framework Hamilton and the evolution of aging. of frailty: a review. Journal of Gerontology: Genetics, 156, 927–931. Medical Sciences, 57A, M283–M288. Charlesworth, B. (2001). Patterns of age-Brock, T. D., Madigan, M. T., Martinko, J. M., specific means and genetic variances of & Parker, J. (1994). Biology of mortality rates predicted by mutation- microorganisms (7th ed.), Englewood Cliffs, accumulation theory of aging. Journal of NJ: Prentice-Hall. Theoretical Biology, 210, 47–65.
CHAPTER 1 / Reliability Theory of Aging and Longevity 37Choi, J. H., Kim, K. L., Huh, W., Kim, B., Curtsinger, J. W., Fukui, H., Townsend, D., & Byun, J., Suh, W., Sung, J., Jeon, E. S., Vaupel, J. W. (1992). Demography of Oh, H. Y., & Kim, D. K. (2004). Decreased genotypes: failure of the limited life-span number and impaired angiogenic function paradigm in Drosophila melanogaster. of endothelial progenitor cells in patients Science, 258, 461–463. with chronic renal failure. Arteriosclerosis, Davis, B. D., Dulbeco, R., Eisen, H. N., & Thrombosis, and Vascular Biology, 24, Ginsberg, H. S. (1990). Microbiology (4th ed.), 1246–1252. Philadelphia: Lippincott.Clark, A. G., & Guadalupe, R. N. (1995). Deng, G., Lu, Y., Zlotnikov, G., Thor, A. D., & Probing the evolution of senescence in Smith, H. S. (1996). Loss of heterozygosity in Drosophila melanogaster with P-element normal tissue adjacent to breast carcinomas. tagging. Genetica, 96, 225–234. Science, 274, 2057–2059.Clark, V. A. (1975). Survival distribution. DePinho, R. A., & Wong, K. K. (2003). The Annual Review of Biophysics and age of cancer: telomeres, checkpoints, and Bioengineering, 4, 431–448. longevity. Journal of Clinical Investigation,Clarke, G., Collins, R. A., Leavitt, B. R., 111, S9–S14. Andrews, D. F., Hayden, M. R., Lumsden, Doblhammer, G., & Vaupel, J. W. (2001). C. J., & McInnes, R. R. (2000). A one-hit Lifespan depends on month of birth. model of cell death in inherited neuronal Proceedings of the National Academy of degenerations. Nature, 406, 195–199. Sciences of the USA, 98, 2934–2939.Clarke, G., Collins, R. A., Leavitt, B. R., Dobzhansky, T. (1962). Mankind Evolving. Andrews, D. F., Hayden, M. R., The Evolution of Human Species. Lumsden, C. J., & McInnes, R. R. (2001a). New Haven and London: Yale University Addendum: a one-hit model of cell death in Press. inherited neuronal degenerations. Nature, Dworkin, P. H. (1992). Pediatrics (2nd ed.), 409, 542. Malvern, PA: Harwal.Clarke, G., Lumsden, C. J., & McInnes, R. R. Eakin, T., Shouman, R., Qi, Y. L., Liu, G. X., & (2001b). Inherited neurodegenerative Witten, M. (1995). Estimating parametric diseases: the one-hit model of survival model parameters in gerontological neurodegeneration. Human Molecular aging studies. Methodological problems and Genetics, 10, 2269–2275. insights. Journal of Gerontology: BiologicalCohen, H. Y., Miller, C., Bitterman, K. J., Sciences, 50A, B166–B176. Wall, N. R., Hekking, B., Kessler, B., Economos, A. C. (1979). A non-Gompertzian Howitz, K. T., Gorospe, M., de Cabo, R., & paradigm for mortality kinetics of Sinclair, D. A. (2004). Calorie restriction metazoan animals and failure kinetics of promotes mammalian cell survival by manufactured products. Age, 2, 74–76. inducing the SIRT1 deacetylase. Science, Economos, A. C. (1980). Kinetics of metazoan 305, 390–392. mortality. Journal of Social and BiologicalCollins, K., & Mitchell, J. R. (2002). Structures, 3, 317–329. Telomerase in the human organism. Economos, A. C. (1983). Rate of aging, rate of Oncogene, 21, 564–579. dying and the mechanism of mortality.Comfort, A. (1958). The longevity and Archives of Gerontology and Geriatrics, 1, mortality of thoroughbred mares. Journal of 3–27. Gerontology, 13, 342–350. Economos, A. C. (1985). Rate of aging, rateCommo, S., Gaillard, O., & Bernard, B. A. of dying and non-Gompertzian (2004). Human hair greying is linked to a mortality—encore . . . Gerontology, 31, specific depletion of hair follicle 106–111. melanocytes affecting both the bulb and the Edelberg, J. M., Tang, L., Hattori, K., outer root sheath. British Journal of Lyden, D., & Rafii, S. (2002). Young adult Dermatology, 150, 435–443. bone marrow-derived endothelial precursorCrowder, M .J., Kimber, A. C., Smith, R. L., & cells restore aging-impaired cardiac Sweeting, T. J. (1991). Statistical analysis of angiogenic function. Circulation Research, reliability data. London: Chapman & Hall. 90, E89–E93.
38 L. A. Gavrilov and N. S. GavrilovaEisner, E. (1967). Actuarial data for the longevity. Reviews in Clinical Gerontology, Bengalese finch (Lonchura striata: 7, 5–12. Fam.Estrildidae) in captivity. Experimental Gavrilov, L. A, & Gavrilova, N. S. (1999). Gerontology, 2, 187–189. Season of birth and human longevity.Ekerdt, D. J. (Ed.) (2002). The Macmillan Journal of Anti-Aging Medicine, 2, Encyclopedia of Aging. New York: 365–366. Macmillan Reference USA. Gavrilov, L. A., & Gavrilova, N. S. (2000).Feller, W. (1968). An introduction to Human longevity and parental age at probability theory and its applications. conception. In J.-M. Robine, Vol.1, New York: Wiley. T. B. L. Kirkwood & M. Allard (Eds.).Finch, C. E. (1990). Longevity, senescence Sex and longevity: sexuality, gender, and the genome. Chicago: University of reproduction, parenthood (pp. 7–31). Berlin, Chicago Press. Heidelberg: Springer-Verlag.Finch, C. E., & Kirkwood, T. B. L. (2000). Gavrilov, L. A., & Gavrilova, N. S. (2001a). Chance, development, and aging. New Epidemiology of human longevity: the York, Oxford: Oxford University Press. search for appropriate methodology. JournalForsyth, N. R., Wright, W. E., & Shay, J. W. of Anti-Aging Medicine, 4, 13–30. (2002). Telomerase and differentiation in Gavrilov, L. A., & Gavrilova, N. S. (2001b). multicellular organisms: turn it off, turn it The reliability theory of aging and on, and turn it off again. Differentiation, longevity. Journal of Theoretical Biology, 69, 188–197. 213, 527–545.Fukui, H. H., Ackert, L., & Curtsinger, J. W. Gavrilov, L. A., & Gavrilova, N. S. (2003a). (1996). Deceleration of age-specific Early-life factors modulating lifespan. In mortality rates in chromosomal S. I. S. Rattan (Ed.), Modulating aging and homozygotes and heterozygotes of longevity (pp. 27–50). Dordrecht, The Drosophila melanogaster. Experimental Netherlands: Kluwer Academic Publishers. Gerontology, 31, 517–531. Gavrilov, L. A., & Gavrilova, N. S. (2003b).Fukui, H. H., Xiu, L., & Curtsinger, J. W. The quest for a general theory of aging and (1993). Slowing of age-specific mortality longevity. Science of Aging Knowledge rates in Drosophila melanogaster. Environment [electronic resource]: SAGE Experimental Gerontology, 28, 585–599. KE, Vol. 2003 Jul 16, 28, RE5. Available atGalambos, J. (1978). The asymptotic theory of http://sageke.sciencemag.org. extreme order statistics. New York: Wiley. Gavrilov, L. A., & Gavrilova, N. S. (2004a).Gavrilov, L. A. (1978). Mathematical model of Early-life programming of aging and aging in animals. Doklady Akademii Nauk longevity: the idea of high initial damage SSSR: Biological Sciences, 238, 53–55 load (the HIDL hypothesis). Annals of the (English edition). New York Academy of Sciences, 1019,Gavrilov, L. A. (1980). Study of life span 496–501. genetics using the kinetic analysis. Ph.D. Gavrilov, L. A., & Gavrilova, N. S. (2004b). Thesis, Moscow, Russia: Moscow State The reliability-engineering approach to the University. problem of biological aging. Annals of theGavrilov, L. A. (1984). Does a limit of the life New York Academy of Sciences, 1019, span really exist? Biofizika, 29, 908–911. 509–512.Gavrilov, L. A., & Gavrilova, N. S. (1979). Gavrilov, L. A., & Gavrilova, N. S. (2004c). Determination of species length of life. Why we fall apart. Engineering’s reliability Doklady Akademii Nauk SSSR: Biological theory explains human aging. IEEE Sciences, 246, 905–908 (English edition). Spectrum, 9, 2–7.Gavrilov, L. A., & Gavrilova, N. S. (1991). Gavrilov, L. A., Gavrilova, N. S., & The biology of life span: a quantitative Iaguzhinskii, L. S. (1978). Basic patterns of approach, New York: Harwood Academic aging and death in animals from the Publisher. standpoint of reliability theory. JournalGavrilov, L. A., & Gavrilova, N. S. (1997). of General Biology [Zhurnal Obshchej Parental age at conception and offspring Biologii], 39, 734–742.
CHAPTER 1 / Reliability Theory of Aging and Longevity 39Gavrilova, N. S., & Gavrilov, L. A. (2002). Haranghy, L., & Balázs, A. (1980). Regeneration Evolution of aging. In D. J. Ekerdt (Ed.), and rejuvenation of invertebrates. In N. W. Encyclopedia of aging (Vol. 2, pp. 458–467). Shock (Ed.), Perspectives in experimental New York: Macmillan Reference USA. gerontology (pp. 224–233). New York:Gavrilova, N. S., Gavrilov, L. A., Arno Press. Evdokushkina, G. N., Semyonova, V. G. Harman, D., & Eddy, D. E. (1979). Free radical (2003). Early-life predictors of human theory of aging: beneficial effects of adding longevity: analysis of the 19th century antioxidants to the maternal mouse diet on birth cohorts. Annales de Démographie life span of offspring: possible explanation Historique, 2, 177–198. of the sex difference in longevity. Age, 2,Gilchrest, B. A., Blog, F. B., & Szabo, G. 109–122. (1979). Effects of aging and chronic sun Heintz, N. (2000). One-hit neuronal death. exposure on melanocytes in human skin. Nature, 406, 137–138. Journal of Investigative Dermatology, 73, Herndon, L. A., Schmeissner, P. J., Dudaronek, 141–143. J. M., Brown, P. A., Listner, K. M., Sakano, Y.,Goldschmidt-Clermont, P. J. (2003). Loss of Paupard, M. C., Hall, D. H., & Driscoll, M. bone marrow-derived vascular progenitor (2002). Stochastic and genetic factors cells leads to inflammation and influence tissue-specific decline in ageing atherosclerosis. American Heart Journal, C. elegans. Nature, 419, 808–814. 146(4 Suppl), S5–S12. Hirsch, A. G., Williams, R. J., & Mehl, P.Golubev, A. (2004). Does Makeham make (1994). Kinetics of medfly mortality. sense? Biogerontology, 5, 159–167. Experimental Gerontology, 29, 197–204.Gompertz, B. (1825). On the nature of the Hirsch, H. R., & Peretz, B. (1984). Survival function expressive of the law of human and aging of a small laboratory population mortality and on a new mode of of a marine mollusc, Aplysia californica. determining life contingencies. Mechanisms of Ageing and Development, Philosophical Transactions of the Royal 27, 43–62. Society of London, 115A, 513–585. Hjalmarsson, O., Hagberg, B., Hagberg, G.,Gosden, R. G. (1985). The biology of Kubli, F., Patel, N., Schmidt, W., & menopause: the cause and consequence of Linderkamp, O. (Eds.). (1988). Perinatal ovarian aging. San Diego: Academic Press. events and brain damage in survivingGouda, M. D., Singh, S. A., Rao, A. G., children. Berlin: Springer-Verlag, pp. Thakur, M. S., & Karanth, N. G. (2003). 28–36. Thermal inactivation of glucose oxidase: Holmes, D. J., & Ottinger, M. A. (2003). Birds mechanism and stabilization using as long-lived animal models for the study of additives. Journal of Biological Chemistry, aging. Experimental Gerontology, 38, 278, 24324–24333. 1365–1375.Greenwood, M., & Irwin, J. O. (1939). The Holmes, D. J., Fluckiger, R., & Austad, S. N. biostatistics of senility. Human Biology, 11, (2001). Comparative biology of aging in 1–23. birds: an update. ExperimentalGumbel, E. J. (1958). Statistics of extremes. Gerontology, 36, 869–883. New York: Columbia University Press. Hopkin, K. (2001). More than a sum our cells.Hack, M., & Fanaroff, A. A. (2000). Outcomes Science of Aging Knowledge Environment of children of extremely low birthweight [electronic resource]: SAGE KE, Vol. 2001 and gestational age in the 1990s. Seminars Oct 3, 1, oa4. in Neonatology, 5, 89–106. Janse, C., Slob, W., Popelier, C. M., &Hall, J. C. (1969). Age-dependent enzyme Vogelaar, J.W. (1988). Survival changes in Drosophila melanogaster. characteristics of the mollusc Lymnaea Experimental Gerontology, 4, 207–222. stagnalis under constant culture conditions:Handyside, A. H., & Delhanty, J. D. A. (1997). effects of aging and disease. Mechanisms of Preimplantation genetic diagnosis: Ageing and Development, 42, 263–174. strategies and surprises. Trends in Genetics, Jay, J. M. (1996). Modern food microbiology. 13, 270–275. New York: Chapman and Hall.
40 L. A. Gavrilov and N. S. GavrilovaJohnson, T. E. (1987). Aging can be genetically Le Bourg, É. (2001). A mini-review of the dissected into component processes using evolutionary theories of aging. Is it long-lived lines of Caenorhabditis elegans. the time to accept them? Demographic Proceedings of the National Academy of Research [electronic resource], 4, 1–28. Sciences of the USA, 84, 3777–3781. Available at http://www.demographic-Johnson, T. E. (1990). Increased life span of research.org/volumes/vol4/1/4–1.pdf. age-1 mutants in Caenorhabditis elegans Leeuwenburgh, C. (2003). Role of apoptosis in and lower Gompertz rate of aging. Science, sarcopenia. Journal of Gerontology: 249, 908–912. Biological Sciences, 58A, B999–B1001.Jonason, A. S., Kunala, S., Price, G. T., Leon, D. A., Lithell, H. O., Vågerö, D., Restifo, R. J., Spinelli, H. M., Persing, J. A., Koupilová, I., Mohsen, R., Berglund, L., Leffell, D. J., Tarone, R. E., & Brash, D. E. Lithell, U.-B., & McKeigue, P. M. (1998). (1996). Frequent clones of p53-mutated Reduced fetal growth rate and increased keratinocytes in normal human skin. risk of death from ischaemic heart disease: Proceedings of the National Academy of cohort study of 15,000 Swedish men and Sciences of the USA, 93, 14025–14029. women born 1915–29. British MedicalKeller, G., Zimmer, G., Mall, G., Ritz, E., & Journal, 317, 241–245. Amann, K. (2003). Nephron number in Libby, P. (2003). Bone marrow: a fountain of patients with primary hypertension. New vascular youth? Circulation, 108, 378–379. England Journal of Medicine, 348, 101–108. Lindop, P. J. (1961). Growth rate, lifespan andKhazaeli, A. A., Xiu, L., & Curtsinger, J. W. causes of death in SAS/4 mice. (1996). Effect of density on age-specific Gerontologia, 5, 193–208. mortality in Drosophila: a density Lucas, A., Fewtrell, M. S., & Cole, T. J. (1999). supplementation experiment. Genetica, Fetal origins of adult disease: the 98, 21–31. hypothesis revisited. British MedicalKhrapko, K., Ebralidse, K., & Kraytsberg, Y. Journal, 319, 245–249. (2004). Where and when do somatic mtDNA Makeham, W. M. (1860). On the law of mutations occur? Annals of the New York mortality and the construction of annuity Academy of Sciences, 1019, 240–244. tables. Journal of the Institute of Actuaries,Kim, Sh. S. H, Kaminker, P., & Campisi, J. 8, 301–310. (2002). Telomeres, aging and cancer: in Martin, G. M. (2002). Gene action in the search of a happy ending. Oncogene, 21, aging brain: an evolutionary biological 503–511. perspective. Neurobiology of Aging, 23,Klein, J. P., & Moeschberger, M. L. (1997). 647–654. Survival analysis: techniques for censored Martin, L. J., Brambrink, A. M., Price, A. C., and truncated data. New York: Springer- Kaiser, A., Agnew, D. M., Ichord, R. N., & Verlag. Traystman, R. J. (2000). Neuronal death inKuh, D., & Ben-Shlomo, B. (1997). A life newborn striatum after hypoxia-ischemia is course approach to chronic disease necrosis and evolves with oxidative stress. epidemiology. Oxford: Oxford University Neurobiology of Disease, 7, 169–191. Press. Martinez, D. E. (1998). Mortality patternsKundi, M. (1999). One-hit models for virus suggest lack of senescence in hydra. inactivation studies. Antiviral Research, Experimental Gerontology, 33, 217–225. 41, 145–152. Masoro, E. J. (2003). Subfield history: caloricKunstyr, I., & Leuenberger, H.-G. W. (1975). restriction, slowing aging, and extending Gerontological data of C57BL/6J mice. I. life. Science of aging knowledge Sex differences in survival curves. Journal environment [electronic resource]: SAGE of Gerontology, 30, 157–162. KE, Vol. 2003 Feb 26, 8, RE2.Kurganov, B. I. (2002). Kinetics of protein Masoro, E. J. (2005). Are age-associated aggregation. Quantitative estimation of the diseases an integral part of aging? In chaperone-like activity in test-systems E. J. Masoro & S. N. Austad (Eds.), based on suppression of protein aggregation. Handbook of the biology of aging (6th ed.). Biochemistry (Moscow), 67, 409–422. San Diego, CA: Academic Press.
CHAPTER 1 / Reliability Theory of Aging and Longevity 41Massoff, R. W., Dagnelie, G., Benzschawel, T., Olshansky, S. J., & Carnes, B. A. (1997). Ever Palmer, R. W., & Finkelstein, D. (1990). since Gompertz. Demography, 34, 1–15. First order dynamics of visual field loss in Partridge, L., & Gems, D. (2002). The retinitis pigmentosa. Clinical Vision evolution of longevity. Current Biology, 12, Sciences, 5, 1–26. R544–R546.McKiernan, S. H., Bua, E., McGorray, J., & Partridge, L., & Mangel, M. (1999). Messages Aiken, J. (2004). Early-onset calorie from mortality: the evolution of death rates restriction conserves fiber number in aging in the old. Trends in Ecology and rat skeletal muscle. The FASEB Journal, 18, Evolution, 14, 438–442. 580–581. Pearl, R., & Miner, J. R. (1935). ExperimentalMcLaren, A. (1998). Genetics and human studies on the duration of life. XIY. The reproduction. Trends in Genetics, 14, comparative mortality of certain lower 427–431. organisms. Quarterly Review of Biology,Medawar, P. B. (1946). Old age and natural 10, 60–79. death. Modern Quarterly, 2, 30–49. Peleg, M., Normand, M. D., & [Reprinted in Medawar, P.B. (1958). The Campanella, O. H. (2003). Estimating uniqueness of the individual (pp. 17–43), microbial inactivation parameters from New York: Basic Books]. survival curves obtained under varyingMoffett, D. F., Moffett, S. B., & Schauf, C. L. conditions—the linear case. Bulletin of (1993). Human physiology: foundations Mathematical Biology, 65, 219–234. and frontiers (2nd ed.). St. Louis: Mosby. Perks, W. (1932). On some experiments in theNekhaeva, E., Bodyak, N. D., Kraytsberg, Y., graduation of mortality statistics. Journal McGrath, S. B., Van Orsouw, N. J., of the Institute of Actuaries, 63, 12–57. Pluzhnikov, A., Wei, J. Y., Vijg, J., & Promislow, D. E. (1993). On size and survival: Khrapko, K. (2002). Clonally expanded progress and pitfalls in the allometry of life mtDNA point mutations are abundant in span. Journal of Gerontology: Biological individual cells of human tissues. Sciences, 48, B115–B123. Proceedings of the National Academy of Promislow, D. E. (1994). DNA repair and the Sciences of the USA, 99, 5521–5526. evolution of longevity: a critical analysis.Nyengaard, J. R., & Bendtsen, T. F. (1992). Journal of Theoretical Biology, 170, Glomerular number and size in relation to 291–300. age, kidney weight, and body surface in Rausand, M., & Høyland, A. (2003). System normal man. Anatomical Record, 232, reliability theory: models, statistical 194–201. methods, and applications (2nd ed.).Ogburn, C. E., Austad, S. N., Holmes, D. J., Hoboken, NJ: Wiley-Interscience. Kiklevich, J. V., Gollahon, K., Rabinovitch, Rauscher, F. M., Goldschmidt-Clermont, P. J., P. S., & Martin, G. M. (1998). Cultured Davis, B. H., Wang, T., Gregg, D., renal epithelial cells from birds and mice: Ramaswami, P., Pippen, A. M., Annex, B. H., enhanced resistance of avian cells to Dong, C., & Taylor, D. A. (2003). Aging, oxidative stress and DNA damage. Journal progenitor cell exhaustion, and of Gerontology: Biological Sciences, 53A, atherosclerosis. Circulation, 108, 457–463. B287–B292. Reznick, D. N., Bryant, M. J., Roff, D.,Ogburn, C. E., Carlberg, K., Ottinger, M. A., Ghalambor, C. K., & Ghalambor, D. E. Holmes, D. J., Martin, G. M., & Austad, S. N. (2004). Effect of extrinsic mortality on the (2001). Exceptional cellular resistance to evolution of senescence in guppies. Nature, oxidative damage in long-lived birds requires 431, 1095–1099. active gene expression. Journal of Ricklefs, R. E., & Scheuerlein, A. (2002). Gerontology: Biological Sciences, 56A, Biological implications of the Weibull and B468–B474. Gompertz models of aging. Journal ofOlshansky, S. J. (1998). On the Gerontology, 57A, B69–B76. biodemography of aging: a review essay. Rigdon, S. E., & Basu, A. P. (2000). Statistical Population and Development Review, 24, methods for the reliability of repairable 381–393. systems. New York: Wiley.
42 L. A. Gavrilov and N. S. GavrilovaRockstein, M., & Lieberman, H. M. (1959). age-specific mortality. Evolution, 47, A life table for the common house fly, 1302–1312. Musca domestica. Gerontologia, 3: 23–36. United Nations (1967). Demographic Yearbook,Sacher, G. A. (1966). The Gompertz 1966, 18th issue. New York: United Nations. transformation in the study of the injury- United Nations (1975). Demographic mortality relationship: application to late Yearbook, 1974, 26th issue. New York: radiation effects and ageing. In P. J. Lindop United Nations. & G. A. Sacher (Eds.), Radiation and Vanfleteren, J. R., De Vreese, A., & ageing (pp. 411–441). London: Taylor and Braeckman, B. P. (1998). Two-parameter Francis. logistic and Weibull equations provideSacher, G. A. (1977). Life table modification better fits to survival data from isogenic and life prolongation. In C. E. Finch & populations of Caenorhabditis elegans in L. Hayflick (Eds), Handbook of the biology axenic culture than does the Gompertz of aging (pp. 582–638). New York: Van model. Journal of Gerontology: Biological Nostrand Reinhold. Sciences, 53A, B393–B403.Sacher, G. A., & Duffy, P. H. (1979). Genetic Van Zant, G., & Liang, Y. (2003). The role of relation of life span to metabolic rate for stem cells in aging. Experimental inbred mouse strains and their hybrids. Hematology, 31, 659–672. Federation Proceedings, 38, 184–188. Vaupel, J. W., Carey, J. R., Christensen, K.,Schlettwein-Gsell, D. (1970). Survival curves Johnson, T., Yashin, A. I., Holm, N. V., of an old age rat colony. Gerontologia, 16, Iachine, I. A., Kannisto, V., Khazaeli, A. A., 111–115. Liedo, P., Longo, V. D., Zeng, Y., Manton,Schulzer, M., Lee, C. S., Mak, E. K., K., & Curtsinger, J. W. (1998). Vingerhoets, F. J. G., & Calne, D. B. (1994). Biodemographic trajectories of longevity. A mathematical model of pathogenesis in Science, 280, 855–860. idiopathic Parkinsonism. Brain, 117, Volpe, J. (2000). Neurology of the newborn 509–516. (4th ed.). Philadelphia: Saunders.Sehl, M. E., & Yates, F. E. (2001). Kinetics of Wallace, W. H., & Kelsey, T. W. (2004). human aging: I. Rates of senescence Ovarian reserve and reproductive age may between ages 30 and 70 years in healthy be determined from measurement of people. Journal of Gerontology: Biological ovarian volume by transvaginal sonography. Sciences, 56A, B198–B208. Human Reproduction, 19, 1612–1617.Strasser, H., Tiefenthaler, M., Steinlechner, M., Weibull, W. A. (1939). A statistical theory of Eder, I., Bartsch, G., & Konwalinka, G. the strength of materials. (2000). Age dependent apoptosis and loss of Ingeniorsvetenskapsakademiens rhabdosphincter cells. Journal of Urology, Handlingar, 151, 5–45. 164, 1781–1785. Weibull, W. A. (1951). A statisticalStrehler, B. L. (1978). Time, cells, and aging distribution function of wide applicability. (2nd ed.). New York and London: Academic Journal of Applied Mechanics, 18, 293–297. Press. Williams, G. C. (1957). Pleiotropy, naturalStrehler, B. L., & Mildvan, A. S. (1960). selection and the evolution of senescence. General theory of mortality and aging. Evolution, 11, 398–411. Science, 132, 14–21. Wilmoth, J. R. (1997). In search of limits. InTatar, M., Carey, J. R., & Vaupel, J. W. (1993). K. W. Wachter & C. E. Finch (Eds.), Long-term cost of reproduction with and Between Zeus and the salmon. The without accelerated senescence in biodemography of longevity (pp. 38–64). Callosobruchus maculatus: analysis of Washington, DC: National Academy Press.
44 E. J. Masoroadvances in our knowledge of the biology important to note that organismic senes-of aging and age-associated diseases. This cence (i.e., aging) is a process that starts inissue is of interest not only from an aca- early adult life and progresses from thatdemic conceptual perspective, but it also time on. Indeed, athletes in some sportsaffects the planning, design, and execu- exhibit performance deterioration as earlytion of biogerontologic research. Also, as as the last half of the third decade of life;pointed out by Wick and colleagues (2003), in most sports, deterioration in perform-the study of age-associated diseases may ance occurs by the middle of the fourthwell be of great value for understanding decade. It is also critical not to confusebasic aging processes. However, it is first organismic senescence with cellularnecessary to review some of the long-held senescence, a term for the loss of prolifer-concepts of biological gerontology that ative ability of cells in culture. In fact, ithave influenced perceptions of the rela- is the definition of aging as a synonym fortionship between aging and age-associated organismic senescence that underlies thediseases. concept of biological age in contrast to chronological age (Borkan & Norris, 1980) and the search for biomarkers of aging II. Concepts of Biological (McClearn, 1997). Thus, aging and senes- Gerontology cence will be used as synonyms in this chapter; although not explicitly stated,There is no better starting point than the such is generally the case in other chap-definition of aging. A broad, general defini- ters of this book.tion is that aging is what happens to an Based on this narrow definition oforganism over time (Costa & McCrae, aging, a set of putative fundamental char-1995). This definition includes (1) no acteristics of aging was developed inchange with time, (2) beneficial changes the field of biological gerontology duringwith time (such as the maturation of phy- the first half of the 20th century. These,siological processes during development), as Strehler (1977) succinctly summa-and (3) deteriorative changes with time rized, include intrinsicality, universality,(such as those that occur with advancing progressivity, irreversibility, and geneti-adult age). Although this broad definition cally programmed. Intrinsicality definesof the aging of organisms is certainly valid, aging as an inherent characteristic of theit is not what is usually meant when bio- organism rather than a response to envi-logical gerontologists use the term. Nor, ronmental factors. Universality confinesfor that matter, is it what laypersons mean aging to processes that occur in all mem-when discussing the aging of friends and bers of a species (or all members of a gen-family. Rather, both groups are usually der of a species; e.g., menopause in thereferring solely to those deteriorative human female). Progressivity describeschanges noted with advancing age, such as aging as a change that gradually increasesdiminished physiological capacity or the in magnitude over time. Irreversibilitywrinkling of the skin. professes that once a change due to aging In other words, when biological geron- occurs, it cannot be reversed. Geneticallytologists and laypersons use the term programmed originally meant that agingaging, they are most often referring to was programmed in the same sense assenescence. Senescence is defined as the developmental processes; more recently,progressive deterioration during the adult it is often modified to mean the influ-period of life that underlies an increasing ence of genotype on the aging phenotype.vulnerability to challenges and a decreas- It is these putative fundamental charac-ing ability of the organism to survive. It is teristics of aging that have, to a great
CHAPTER 2 / Are Age-Associated Diseases an Integral Part of Aging? 45extent, served as the basis for separating integral part of aging. Such diseasebiological aging from age-associated dis- processes cause morbidity and/or mor-ease. Alas, as will become evident in this tality primarily at advanced ages and arechapter, it is doubtful that these charac- chronic, or when acute are the result ofteristics have been helpful in clarifying long-term processes, such as gradualthe relationship between aging and age- loss of bone and atherogenesis. Brodyassociated disease. and Schneider (1986) divided these dis- eases into two classes: age-dependent diseases and age-related diseases. They III. Age-Associated Diseases defined age-dependent diseases as those in which the pathogenesis appears toThe American Heritage Dictionary (1996) involve basic aging processes. Mortalitydefines disease as “a pathological condi- and morbidity caused by these diseasestion of a part, an organ, or a system of an increase exponentially with advancingorganism resulting from various causes, age. As examples of age-dependent dis-such as infection, genetic defect, or envi- eases, they listed coronary heart disease,ronmental stress, and characterized by an cerebrovascular disease, type II diabetes,identifiable group of signs or symptoms.” osteoporosis, Alzheimer’s disease, andAlthough this definition is adequate for Parkinson’s disease. They defined age-the purposes of this chapter, Scully (2004) related diseases as those with a tempo-points out that any definition of disease ral relationship to the age of the hostis problematic because what is called a but not necessarily related to the agingdisease is influenced by both medical process. These diseases occur at a spe-advances and by societal culture. Thus, cific age, but with a further increase inthe definition changes with time and dif- age they either decline in frequency orfers from place to place. increase at a less than exponential rate. Although acute infectious diseases Gout, multiple sclerosis, amyotrophicoccur at all ages, their consequences may lateral sclerosis (ALS), and many, butbe more severe in the elderly because of certainly not all, cancers are examplesage-associated deterioration in immune of such diseases. The relationship offunction (Papciak et al., 1996) and in the cancer to aging will be discussed later infunctioning of most of the other physio- this chapter.logical systems (Masoro, 1995). Although Interestingly, in a paper on diseases thatgenetic diseases are not restricted to the cause the death of people 85 years or older,young, they often do occur as congenital Kohn (1982) divides the diseases into threediseases or at early postnatal ages classes, two of which are similar to(Blumenthal, 1999). Environmental stress the two classes of Brody and Schneider. Incan also cause disease at all ages, but Kohn’s first class are diseases that arebecause aging decreases the ability to themselves normal aging processes, beingcope with stressors (Shock, 1967), mor- progressive and irreversible under usualbidity and mortality are more likely out- conditions; examples are atheroscleroticcomes of stress in the old than in the diseases, degenerative joint diseases, andyoung. It is reasonable to conclude that osteoporosis. His second class covers dis-the above categories of diseases are not eases that increase with age but may notan integral part of aging, even though be part of the aging process; examples areaging can clearly affect the consequences neoplasia and hypertension. The thirdof such diseases. class includes diseases with consequences However, age-associated diseases more serious in people of advanced ageshould not be so readily dismissed as an than in young people; infectious disease
46 E. J. Masorois an example. In a paper published some It is of interest to note that in humans,15 years later, Klima and colleagues (1997) the heritability of age-associated diseasesfound the causes of death in geriatric is in the same range (Ͻ40 percent) as thepatients in Houston, Texas, and Prague, heritability of life span (Longo & Finch,Czech Republic, to be similar to what 2002). However, this claim needs toKohn reported. be tempered because heritability varies Recently, Horiuchi and colleagues among specific age-associated diseases.(2003) reported on the causes of death in Nevertheless, this fact and much of theFrance from 1979 to 1994 during two information just presented are compatibleperiods of life in subjects age 15 to 100- with the view that age-associated diseaseplus years. In one group of subjects age is an integral part of aging. In addi-30 to 54 and a second group of subjects tion, Urban and colleagues (2002) posedage 65 to 89, the age-specific death rate that age-associated autoimmune diseasesincreased exponentially with increasing result from genetic alterations causedage (a measure of population aging). by aging processes. And Ames and col-Deaths due to malignant neoplasms, leagues (1993) concluded that the oxidantacute myocardial infarctions, hyperten- byproducts of normal metabolism causesive disease, and liver cirrhosis rose rap- extensive damage to DNA, proteins, andidly during the 30- to 54-year age range. lipids, thereby contributing to aging andDuring the range of 65 to 89 years, death age-associated diseases such as brain dys-due to certain infectious diseases, acci- function, cancer, cardiovascular disease,dents, dementia, heart failure, and cere- and cataracts. Rattan (1991) proposedbrovascular disease rose rapidly. Deaths that aging and age-associated diseasedue to malignant neoplasms and acute are linked by the failure to maintainmyocardial infarction did not rise rapidly the appropriate level and structure ofin the 65 to 89 age range; thus, the frac- proteins. However, although the viewstion of deaths due to these diseases was of the Urban and Ames groups andless than during the 30 to 54 age range. Rattan are provocative, much more workHowever, of the people who died remains to be done to establish theirbetween 15 and 100 years old, only 10 validity.percent died between 30 and 54 years and67 percent died between 65 and 89 years.Thus, the absolute number of deaths due IV. Primary Aging, Secondaryto malignant neoplasms and myocardial Aging, and “Normal Aging”infarction was markedly greater in thosein the older age range compared to those The concept of primary and secondaryin the 30- to 54-year age range. aging was proposed by Busse (1969) to Other investigators have also found resolve the paradox of viewing intrinsi-that the fraction of deaths due to cancer cality as one of the fundamental charac-decreased at advanced ages. Smith (1996) teristics of aging, although it was obviousreported that cancer was the cause of that environmental factors influence40 percent of deaths that occurred between both aging and diseases associated with50 and 69 years of age but only 4 percent aging. Primary aging is defined as theof those that occurred at ages older than universal changes occurring with age that100 years. Miyaishi and colleagues (2000) are not caused by disease or environmen-found that the peak incidence of single tal influences. Secondary aging is definedcancers occurred between 60 and 64 as changes involving interactions of pri-years of age, and of multiple cancers mary aging processes with environmentalbetween 80 and 84 years of age. influences and disease processes.
CHAPTER 2 / Are Age-Associated Diseases an Integral Part of Aging? 47 Thus, the concept of primary and sec- secondary aging is useful. In fact, it hasondary aging relegates disease, including probably impeded our quest for under-age-associated disease, to that of a factor standing the biological basis of aging and,that can influence aging but is not an inte- even worse, will continue to do so.gral part of the aging process. Because the An outgrowth of the concept of primarymajor thrust of this chapter is to assess and secondary aging is the concept of nor-whether age-associated diseases are an mal aging, which is defined as senescenceintegral part of aging, further comment on in the absence of disease (Shock, 1984).this aspect of the primary and secondary This view was bolstered by the concept ofaging concept is more appropriately con- natural death, which Fries (1980) definedsidered after all other issues have been as death in the absence of disease. Givenfully presented. However, placing environ- that with increasing age, most people expe-mental influences in the category of sec- rience one or more age-associated diseases,ondary aging does require comment at this “normal aging” and “natural death” must,juncture. Although the proximate mecha- indeed, be rare occurrences. This is under-nisms underlying aging are not fully scored by the report of Hebert (2004) thatunderstood, most gerontologists would life expectancy in 2001 in the Unitedagree that the long-term accumulation of Kingdom was 75.7 years for men andmolecular damage underlies aging, and it 80.4 years for women, and that on average,is likely that reactive oxygen molecules the projection is that men and womenplay an important role in this damage would suffer from poor health during the(Barja, 2002). Indeed, many believe that last 8.7 and 10.1 years of life, respectively.reactive oxygen molecules play a major As pointed out by Gessert and colleaguescausative role in aging, but the validity (2002), declaring that someone has “died ofof this view remains to be established. old age” does not mean the individual hasMuch of the generation of these reactive died without significant pathology.oxygen molecules occurs in the mitochon- From 1988 to 1994, the prevalence ofdria during the metabolism of fuel, and diabetes in 60- to 74-year-old Americanif these molecules do, indeed, cause aging, Caucasians was 22.3 percent; it wasit would be classified as primary aging. 29.5 percent in African-Americans of theHowever, environmental factors (e.g., same age group (Sinclair & Croxon, 2003).industrial pollutants), lifestyle (e.g., ciga- The prevalence of osteoarthritis in indi-rette smoking), and pharmacological viduals over 70 years of age is more thanagents also cause the formation of reactive 30 percent (Scott, 2003). The prevalence ofoxygen molecules, and if aging results moderate to severe dementia is estimatedfrom reactive oxygen molecules from at 1 to 2 percent at ages 65 to 70 years,these sources, it would be classified as sec- 2 to 5 percent at ages 70 to 75 years, 11 toondary aging. Because in both instances 20 percent at ages 80 to 85 years, and 39aging is proposed to result from the inter- to 60 percent at ages 90 to 95 years (Elbyaction of reactive oxygen molecules et al., 1994; Skoog et al., 1996). Whenwith the macromolecules of the organism, dementia is coupled with the many otherit seems illogical to refer to one as primary diseases that become increasingly com-aging and the other as secondary aging. mon with advancing age (coronary heartIndeed, Vieira and colleagues (2000) disease, type II diabetes, congestive heartshowed that the effect of genes on the failure, stroke, osteoporosis, cataracts,life span of Drosophila melanogaster is Parkinson’s disease, many kinds of cancer,dependent on environmental factors. For and benign prostatic hyperplasia, to namethese reasons, it seems highly unlikely a few), it becomes obvious that aging inthat the concept of primary aging and the absence of disease is rare, indeed.
48 E. J. Masoro In a community-based study of 502 peo- Despite the rarity of absence of diseaseple 90 years of age or older in Stockholm, at advanced ages, investigators have usedonly 19 percent were found to be free both human subjects and animal modelsof disease; the remainder had one or to study “normal aging” by exclud-more diseases, most of which were age- ing those with discernible disease. In theassociated diseases (von Strauss et al., case of human subjects, the procedure of2000). In a study of 207 Danes who reached excluding subjects with discernible dis-their 100th birthday between April 1, eases, including age-associated diseases, is1995, and May 31, 1996, Anderson- referred to as “cleaning up” the physiologi-Ramberg and colleagues (2001) found that cal data (Rowe et al., 1990). Obviously,only one was free of age-associated disease, this procedure limits the study to a small,with a mean number of 4.3 such diseases atypical fraction of the aging population. Itper person. In contrast, based on a study of also suffers from the fact that advances in424 Americans and Canadians in the age medical technology have enhanced detec-range of 100 to 119 years, Perls’ group tion of diseases previously not discernible.(Evert et al., 2003) reported that 19 percent For example, in Lakatta’s (1985) studies onwere free of disease as compared to the less the influence of aging on cardiovascularthan 0.5 percent in the Danish study. This physiology, the thallium-stress test wasapparent discrepancy probably stems from used to exclude subjects with occult coro-two factors: Perls’ group excluded demen- nary heart disease, a technology not avail-tia and osteoarthritis in their assessment of able or not used in earlier studies. Thus,age-associated diseases, and they used it can be anticipated that advances in med-interviews and questionnaires rather than ical technology will lead to an ever-physical examinations to determine dis- decreasing fraction of the elderly in theease status. Subsequently, in reviewing “normal aging” category.findings from various worldwide centenar- Finally, the classification of age-associ-ian studies, Perls (2004) concluded that the ated functional change as either a physi-prevalence of dementia is 75 to 85 percent ological or pathophysiological process isin people age 100 years and older. arbitrary. For example, bone loss occursMoreover, in another paper (Hitt et al., with advancing age in almost all, if not1999), which emphasizes that centenarians all, humans (Kalu, 1995); thus, it is con-have been healthier than others for most sidered an age-associated physiologicalof their lives, Perls’ group reported that deterioration. However, when the loss is37 centenarians with a mean age of 102 of sufficient magnitude to have a clinicalsuffered from 4.0 diseases or chronic condi- impact, it is then viewed as a patho-tions at that age, 3.2 such conditions physiological process underlying the age-5 years earlier, and 2.6 such conditions associated disease osteoporosis. Another10 years earlier. Clearly, even in the example is the age-associated blunting ofhealthiest, disease is prevalent at advanced the baroreflex in healthy humans, whichages. is considered a physiological deteriora- Investigators have also found a progres- tion (Jones et al., 2003). However, ifsive increase with age in the prevalence hypertension results from this physio-and severity of age-associated diseases logical deterioration, the individual isin both rats (Maeda et al., 1985) and mice said to have a disease (Lakatta & Levy,(Lipman et al., 1999). Although not 2003).as exhaustively studied as laboratory Does the concept of “normal aging”rodents, such also appears to be the case have any value? Potentially, it may befor other mammalian species used as useful as a reductionist tool in aginganimal models in aging research. research. Indeed, reductionism has been
CHAPTER 2 / Are Age-Associated Diseases an Integral Part of Aging? 49a powerful tool in biological science. most organisms in the wild do not liveUnfortunately, this use of “normal long enough for these genes to apprecia-aging” is, to a great extent, undermined bly affect evolutionary fitness. However,by the fact that advances in biomedical in a protected environment where reach-technology, such as the thallium-stress ing advanced ages becomes the norm,test discussed above, will result in an the late-life detrimental effects of theseever-changing standard of what is consid- genes result in senescence. There isered “normal aging.” Moreover, based on some empirical evidence in support ofnew evidence, what was considered nor- this mechanism (Hughes et al., 2002).mal may be reclassified as disease. For Another genetic mechanism, proposedexample, in 1990, a systolic blood pres- by Williams (1957), is referred to as antag-sure of 140 to 160 mmHg was not viewed onistic pleiotropy. It proposes that thoseas hypertension, whereas currently a sys- genes that increase evolutionary fitness intolic blood pressure of 140 or above early life will be selected for, even if theyis considered hypertension (Lakatta & have catastrophic deleterious effects inLevy, 2003). late life. Again, the deleterious effects of these genes will be evident only in sub- jects in protected environments that V. Evolutionary Theory and enable a long life. Although Huntington’s Age-Associated Diseases disease has been used as a classic example of the mutation accumulation mecha-The currently held concept of the evolu- nism, there is evidence to indicate that ittion of aging provides strong support for may instead be an example of the antago-the view that age-associated diseases are nistic pleiotropy mechanism (Frontalian integral part of aging. Specifically, it is et al., 1996).believed that aging occurs because the Kirkwood (1977) proposed a thirdforce of natural selection decreases with genetic mechanism, which he terms theincreasing post-sexual maturational age disposable soma theory. Although this(Rose, 1991); that is, there is a progres- theory does not propose a specificsive age-associated post-maturational genetic mechanism, it is based on thedecrease in the ability of natural selec- concept that evolutionary forces tend totion to eliminate detrimental traits. Lee form a genome that yields the maxi-(2003) points out that this concept mum number of progeny. The premiseshould include not only the generation of of this theory is that the fundamentalprogeny but, in addition, the intergenera- life role of organisms is to utilize freetional transfer of food and care, which energy in the environment to gener-promotes survival of offspring. ate progeny. To do so requires the use Three genetic mechanisms that affect of energy for both reproduction andreproduction and survival have been pro- the maintenance of the body (whichposed, and each is compatible with this Kirkwood calls “somatic maintenance”).evolutionary concept. One is termed He proposes that the force of naturalthe mutation accumulation mechanism selection leads to apportioning the use(Medawar, 1952). It proposes that natural of energy between reproduction andselection does not eliminate mutated somatic maintenance so as to maximizegenes that are expressed throughout life evolutionary fitness (i.e., generation of abut do not have detrimental effects until maximum lifetime yield of viable prog-late in life; thus, they accumulate in the eny). As a consequence, less energy isgenome. Specifically, because of preda- used for somatic maintenance thantion and other environmental hazards, would be needed for indefinite survival,
50 E. J. Masoroand this deficit in energy for mainte- major age-associated disease processes,nance is manifested in the deterioration have been chosen for examination.of the organism referred to as aging (thelower the use of energy for somatic A. Atherosclerosismaintenance, the greater the rate ofaging). Atherosclerosis is a disorder of the large- Although Martin (2002) tends to feel and medium-sized arteries that, inthat the above three mechanisms have humans, underlies most coronary arterylikely played a role in the evolution of disease (myocardial infarction, heartaging, he suggests that other genetic failure) and peripheral vascular diseasemechanisms may also be involved. It is (aneurysms, gangrene of the extremities)clear that further work is needed to and plays a major role in cerebrovasculardetermine the relative importance of disease (Bierman, 1985; Faxon et al., 2004;each of the proposed mechanisms and Scott, 2004). Brody and Schneider (1986)of other possible genetic mechanisms classify both coronary artery diseaseas well. and cerebrovascular disease in the age- However, Cortopassi (2002) has pointed dependent subclass of age-associated dis-out that irrespective of specific genetic eases. Atherosclerosis is a progressivemechanisms, aging involves a high rate of process that proceeds over several decadesfixation of alleles that have deleterious but does not usually have clinical conse-effects in the post-reproductive phenotype quences until advanced middle age oras compared to the low rate of fixation of older (Ross, 1986), when rupture of thesuch alleles in the pre-reproductive pheno- advanced atheromatous plaque or throm-type. Age-associated disease is almost cer- bosis related to the plaque often causestainly one of the inevitable outcomes of an acute clinical event (Lusis, 2000).the high rate of fixation of these deleteri- Atherosclerosis occurs in almost allous alleles. Indeed, Wick and colleagues humans (Bierman, 1985); for example,(2000) subscribe to the view that inherent moderate to severe atherosclerosis of theadvantages of biological factors that arterial tree was found in all 23 (7 men andenhance reproduction in the young adult 16 women) centenarians. Atherosclerosisare paid for by senescent deterioration, is responsible for about 50 percent of allincluding age-associated diseases, in later mortality in the United States, Europe,life. Indeed, Wick and colleagues (2003) and Japan (Ross, 1993). Although it ishave presented evidence that suggests that almost ubiquitous in humans, atheroscle-benign prostate hyperplasia, prostate can- rosis is not commonly seen in rodents orcer, atherosclerosis, and Alzheimer’s dis- nonhuman primates (Lakatta, 2003).ease are results of antagonistic pleiotropy. However, the absence of atherosclerosis inThey further state that it remains to be these species may be due to diet; for exam-shown whether this is also true of other ple, atherosclerosis occurs in baboons fedage-associated diseases. an atherogenic diet (Babiak et al., 1984), a diet similar in composition to that of many humans. VI. Analysis of Two Major Many of the same factors that under-Age-Associated Disease Processes lie the age-associated structural and functional changes in the human cardio-A consideration of the processes that vascular system are implicated in theunderlie age-associated diseases provides pathogenesis of atherosclerosis (Lakatta,further insight into their relationship to 2003). It appears to start at the intimal sur-aging. Atherosclerosis and neoplasia, two face of the artery, with fatty streaks
CHAPTER 2 / Are Age-Associated Diseases an Integral Part of Aging? 51consisting of lipid-engorged foam cells cells accompanies the accumulation ofoverlain with intact endothelium; the lipids and fibrous material in atheroma-foam cells are formed by recruitment of tous arteries (Robbie & Libby, 2001).monocytes into the subendothelial space, Specifically, pro-inflammatory, oxidizedalong with the accumulation of modified phospholipids generated by the oxidationforms of low-density lipoproteins (Greaves of low-density lipoprotein phospholipids,& Gordon, 2001). The lesion slowly pro- particularly those containing arachidonicgresses in severity, with a continuing acid, provoke an immune responserecruitment of monocytes and T-lympho- (Navab et al., 2004). Bjorkbacka andcytes plus a migration of smooth muscle colleagues (2004) presented evidence thatcells into the lesion as well as their pro- links the pro-inflammatory signaling cas-liferation. After many decades, the lesion cade induced in the arterial wall by ele-progresses into the fibrous plaque, which vated plasma lipids to that also used byalso contains smooth muscle cells, acti- microbial pathogens. In fact, long-termvated T-lymphocytes, and monocyte- infections have been viewed as a possiblederived macrophages (Stout, 2003). The cause of the chronic inflammation thatchemokine class of cytokines has been underlies atherogenesis (Becker et al.,postulated to play a role in the migration 2001). Infection and inflammation induceof monocytes into the developing athero- the acute phase response that is charac-sclerotic lesion (Reape & Groot, 1999); terized by increased levels of certainthere is evidence that leukotrienes play serum or plasma proteins (e.g., C-reactivean important role in the recruitment of protein, amyloid A, and fibrinogen)T-lymphocytes into the lesion (Jala & and the decreased level of others (e.g.,Haribabu, 2004). albumin, transferrin, and ␣–fetoprotein). Witzum (1994) hypothesized that the Cytokines from macrophages, mono-oxidative modification of low-density cytes, T-lymphocytes, and endotheliallipoproteins is a key event in the patho- cells mediate the acute phase response; itgenesis of atherosclerosis, and Steinberg has been proposed that when inflamma-(1997) has provided experimental find- tion is sustained, the pro-inflammatoryings in support of this hypothesis. cytokines play a pathogenic role inHypercholesterolemia, hypertension, dia- atherogenesis by causing changes inbetes, and smoking, the common risk lipoprotein metabolism, which leads tofactors for atherosclerosis, are associated the formation of pro-atherogenic lipopro-with an increased vascular production of teins (Khovidhunkit et al., 2004).reactive oxygen species (Abe & Berk, The involvement of both oxidative1998; Landmesser et al., 2004). Ross damage and inflammation in the genesis(1999) pointed out that atherosclerosis of atherosclerosis indicates that thisis an inflammatory disease and that pathologic process is strongly linked tooxidized low-density lipoproteins have aging processes. Many biological geron-a role in the arterial inflammatory tologists believe that oxidative damage isprocesses involved in the pathogenesis of the major proximate mechanism respon-these lesions. sible for senescence (Barja, 2002; Finkel & Indeed, inflammation appears to play a Holbrook, 2000). Strikingly, de Nigris andfundamental role in mediating all stages colleagues (2003) proposed that oxidationof atherogenesis, from initiation through combined with apoptosis, which has alsoprogression and, ultimately, thrombotic been intimately linked to aging (Zhang &complications (Libby et al., 2003). An Herman, 2002), promotes the progressionimmune and inflammatory response of diseased arteries towards atheroscle-involving endothelial and smooth muscle rotic lesions vulnerable to rupture. These
52 E. J. Masoroare the lesions that give rise to myocar- related macular degeneration (Seddondial infarction and ischemic stroke, major et al., 2004).age-associated diseases (de Nigris et al., It is particularly striking that inflam-2003). matory processes have also been linked Significantly, Krabbe and colleagues to both Alzheimer’s disease and vascu-(2004) have shown that increased lar dementia (Hofman et al., 1997).inflammatory activity is also charac- Interestingly, in an in vitro study,teristic of aging. Compared to young cholesterol ozonolysis products andadults, elderly humans were found to a related lipid-derived aldehyde wereexhibit a two- to four-fold increase in found to modify beta-amyloid so as toplasma levels of inflammatory media- dramatically accelerate amyloidgenesistors such as cytokines (e.g., tumor (Zhang et al., 2004); this finding pro-necrosis factor-alpha and interleukin-6) vides a potential chemical link betweenand in positive acute phase proteins. hypercholesterolemia, oxidative stress,Indeed, in a cohort of 81-year-old inflammation, atherosclerosis, and spo-humans, high levels of tumor necrosis radic Alzheimer’s disease. In fact, manyfactor-alpha in the blood were found to factors contributing to atherogenesisbe associated with a high prevalence have emerged as potential contributorsof atherosclerosis (Bruunsgaard et al., to Alzheimer’s disease (Casserly &2000). Krabbe and colleagues (2004) Torpol, 2004).suggest that aging is associated with a Recently, Rauscher and colleaguesdysregulated cytokine response to stim- (2003) reported still other support for theulation and that, in addition to playing view that aging underlies atherosclerosis;this role in atherogenesis, there is also they found that aging of ApoEϪ/Ϫ miceevidence that this dysregulation is leads to a failure of bone marrow to pro-involved in other age-associated disor- duce progenitor cells capable of repairingders. In fact, Chung and colleagues and rejuvenating arteries. They propose(2001) believe that inflammation plays a that this age-associated loss of functionalkey role in aging, and they have pro- endothelial progenitor cells contributesposed the “Inflammation Hypothesis of to the development of atherosclerosis.Aging.” Moreover, Hill and colleagues (2003) Finch and Crimmins (2004) have observed a strong correlation betweenrecently proposed that long-term inflam- the number of circulating endothelialmation underlies many age-associated progenitor cells and the Framinghamdisorders. Indeed, inflammation has Risk Factor score for coronary artery dis-been implicated in many aspects of age- ease in persons free of clinical disease.associated deterioration, including age- Geiger and Van Zant (2002) suggest thatassociated diseases other than those a loss in the functional quality of stemrelated to atherosclerosis. Chronic low- cells may play an important role in aginggrade inflammation is associated with generally, and that research aimed atage-associated decrease in muscle mass assessing age-changes in the quality of(referred to as sarcopenia), which is stem cells in all tissues is needed.an important component of frailty in the However, a cautionary note is in order: itelderly (Pedersen et al., 2003). Serum has yet to be clearly established thatmarkers of inflammation, especially stem cells undergo senescent deteriora-interleukin-6 and C-reactive protein, tion during organismic aging (Park et al.,have been prospectively associated with 2004; Van Zant & Liang, 2004).cognitive decline in the well-functioning Although a considerable body of evi-elderly (Yaffe et al., 2003) and in age- dence links aging processes and atherogenic
CHAPTER 2 / Are Age-Associated Diseases an Integral Part of Aging? 53processes, a caveat is in order because thus, prostate cancer is included in themuch of this evidence relates to the role age-dependent subclass of age-associatedof oxidative damage and inflammation in diseases.atherogenesis. As discussed above, many Miller (1991) points out that speciation,gerontologists do subscribe to the view caloric restriction, and selective breeding,that oxidative damage and possibly all of which affect the rate of aging, haveinflammation also play a causal role in parallel effects on cancer incidence. Onaging; however, not all findings support the other hand, DePinho (2000) notes thatthis view. Thus, the recent study by the increase in tumors in humansFontana and colleagues (2004) is of between 40 and 80 years of age involvesparticular significance because it offers a primarily epithelial carcinomas and notdifferent connection between aging mesenchymal and hematopoietic malig-processes and atherosclerosis. They nancies, and that the opposite is the casefound that severe, self-imposed caloric for mice between 2 to 4 years (an agerestriction in humans for a period of 3 to range comparable to humans in the 40 to15 years markedly lowered the risk of 80 age range). However, this apparentdeveloping atherosclerosis. This finding species difference may not be realprovides yet another link between aging because it is based primarily on findingsand atherogenesis because caloric restric- with isogenic mouse strains developed fortion markedly retards aging in a spec- studies on the genetics of lymphoma. Intrum of animal species (Masoro, 2002); fact, in a recent study with a geneticallyfor example, in rats, a 40 to 50 percent heterogeneous mouse stock, it was foundreduction in food intake increases the that at advanced age, fibrosarcoma, mam-mortality rate doubling time from about mary adenocarcinoma, and hepatocellular100 days to 200 days (Holehan & Merry, carcinoma were also common and that1986). However, this evidence, too, must many other less common cancers alsobe viewed with some caution because it occurred (Lipman et al., 2004).has not been established that caloric Indeed, as will be discussed below,restriction retards aging in humans. there have been many studies opposing and supporting the view that most can- cers are an integral part of aging.B. Neoplasia Donehower and his coworkers (TynerAlthough cancer can occur at all ages, et al., 2002) presented findings that tendthe incidence of most types increases to disassociate cancer and aging. In awith increasing age in humans and ani- p53 mutant mouse with an enhancedmals (Dix, 1989). Although it is true that p53 activity, they found the expectedcertain tumors, such as neuroblastoma increase in resistance to cancer and also,and Wilms’ tumor, occur in childhood, surprisingly, shortened longevity and thethese may well be due to genetic predis- expression of an age-associated pheno-positions unrelated to aging. The vast type at a chronological age younger thanmajority of cancers occur at advanced in the wild type mouse. Donehowerages, and Brody and Schneider (1986) (2002) suggests that the pro-aging effectsclassify most of them in the age-related of this mutation may result from a grad-subclass of age-associated disease ual increase in the depletion of stem cellbecause they do not continue to increase functional capacity due to increased lev-exponentially from late middle age on els of p53 in the mutant mice; he furtherthrough old-old age. An exception is postulates that this is the “price” of aprostate cancer, which does continue to cancer-free existence during the repro-increase exponentially into old-old age; ductive phase of adult life. Sharpless and
54 E. J. MasoroDePinho (2004) subscribe to this view. Although Liu and colleagues (2003) alsoMoreover, Maier and colleagues (2004) pose that both aging and cancer are due tohave confirmed and extended the find- altered genomic function, they believe thatings of the Donehower group. They an epigenetic mechanism is involved.reported that the overexpression of p44 They point out that both aging and tumori-(the short isoform of p53) in mice genesis are characterized by a globaldecreased longevity and caused prema- hypomethylation of the genomic DNA.ture aging and hyperactivation of the McCullogh and colleagues (1997) foundIGF-1 signaling axis. that there was less regression of neoplas- Although at first glance the work of the tically transformed rat liver epithelialDonehower group strongly suggests that cells when the cells were transplantedcancer is not an integral part of aging, into the liver of old rats rather thanthere is another possible explanation for young rats. These investigators suggesttheir findings. Zhang and Herman (2002) that age-associated alterations in tissueproposed that a defective control of apop- microenvironment may permit the devel-tosis is a cause of aging, and Amundson opment of tumors in late life. Indeed, Belland colleagues (1998) showed that p53 and Van Zant (2004) proposed that theis a positive regulator of apoptosis. aging of the stem cell population of theThus, the enhanced p53 activity in the hematopoietic system creates conditionsmutant mice of Tyner and colleagues may that favor leukemic development, andcause aging by inappropriately enhancing they suggest that aging of stem cells mayapoptosis, an action that simultane- similarly promote the occurrence of otherously protects against the occurrence of types of cancer. Moreover, cancer hascancer. been linked to inflammation (Ho et al., Several other studies intimately link 2004; Marx, 2004; Nelson et al., 2004),cancer and aging at both cellular and and Chung and colleagues (2001) havemolecular levels. Cutler and Semsei suggested that inflammatory processes(1989) were the first to propose that a play an important role in aging. Indeed,common mechanism underlies cancer and Schwartsburd (2004) recently hypothe-aging; they suggested that both are initi- sized that aging causes chronic inflamma-ated and propagated by impaired gene reg- tion, which plays an important role in theulation driven by destabilizing processes pathogenesis of cancer.affecting regulatory elements. In a review Krtolica and colleagues (2001) proposedarticle published some 10 years later, that the age-associated increase in cancerDePinho (2000) also linked increased is due to a synergism between genetic fac-somatic mutations with age as the likely tors (oncogenic mutations) and epigeneticcause of the age-associated increase in factors (substances released from thecancer. It is striking that somatic muta- accumulated senescent cells). It has beentions have long been considered major fac- suggested that the accumulated senescenttors underlying aging (see review by Vijg cells create a microenvironment that& Dolle, 2001). Vijg and Dolle (2002) sug- favors carcinogenesis (Kim et al., 2002).gested that an age-associated increase in Although this is an intriguing concept,genome rearrangements leads to cellular it must be viewed with caution becausesenescence or neoplastic transformation neither the prevalence nor the functionalor the death of cells. Indeed, Hasty and significance of senescent cells has beencolleagues (2003) proposed that the age- established in the intact organismassociated increase in genome instability, (Hornsby, 2002).driven by oxidative damage, is a primary Very recently, Del Monte and Statutocause of aging. (2004) proposed that at least some types
CHAPTER 2 / Are Age-Associated Diseases an Integral Part of Aging? 55of cancer (e.g., epidermal cancer) are an integral part of aging. However, the pre-integral part of aging. Specifically, they ponderance of evidence indicates it ishypothesize that cancers may result from likely that at least some age-associated dis-the dysfunction of gap junction intercel- eases, in particular those classified as agelular communication because of an age- dependent by Brody and Schneider (1986),associated decrease in connexins. are, indeed, an integral part of aging. In summary, although there is a sub- In this regard, it is of interest to con-stantial body of evidence that shows a sider age-associated diseases within theclose association between aging and can- context of the time-honored, fundamen-cer, currently available findings do not tal characteristics of aging: intrinsicality,clearly discriminate among the following universality, progressivity, irreversibility,three possibilities: (1) many cancers are an and genetically programmed. Aging fitsintegral part of aging; (2) aging provides a the concept of intrinsicality only in thefavorable environment for the occurrence limited sense that damage to biologicallyof cancer; or (3) the occurrence of most important molecules, such as DNA, pro-cancers requires the passage of time, but teins, and lipids, is probably the basisin no other way does cancer relate to of the aging phenotype. Strikingly, it isaging. Ershler and Longo (1997) point out believed that molecular damage of a simi-that all three of these possibilities may be lar type underlies age-associated diseases.involved. Moreover, it is interesting to However, it is important to note thatnote that in a review of the issue of aging both in what is classically called agingand cancer, Irminger-Finger (2003) con- and in age-associated diseases, extrinsiccluded that cancer is very much a part factors often play the major role in caus-of normal aging. Clearly, this investiga- ing this intrinsic molecular damage.tor’s definition of normal aging differs As for universality, physiological deterio-markedly from that of Shock (1984). rations, classically viewed as characterizingRecently, Hasty (2005) and Campisi (2005) the aging phenotype, are not universal, norhave independently presented novel views are age-associated diseases. For example,of the relationship between aging and can- deterioration of kidney function has longcer. Hasty poses that aging acts both to been considered a hallmark of the humanincrease life span by preventing cancer in aging phenotype, a view that is based onearly life and to decrease life span because cross-sectional studies showing a markedof functional deterioration in late life. decrease in glomerular filtration rate withCampisi also proposes that aging protects increasing adult age (Rowe et al., 1976).against the occurrence of cancer in However, a longitudinal study of 446 par-early life, but that the accumulation of ticipants in the Baltimore Longitudinalsenescent cells by late life establishes a Study of Aging revealed marked individualcancer-promoting environment. The sec- differences in the change with age inond component of Campisi’s view is not glomerular filtration rate, with one-third ofsupported by the fact that there is a participants showing no change (Lindemandecreased occurrence of cancer in the et al., 1985). On the other hand, as dis-10th and 11th decades of human life. cussed above, atherosclerosis is almost uni- versal in humans. Regarding progressivity, most age- VII. Summary and Conclusions associated diseases exhibit an age- associated progression similar to theThe currently available database does not age-associated physiological deteriora-provide a definitive answer to the question tions classically considered componentsof whether age-associated diseases are an of the aging phenotype.
56 E. J. Masoro Strictly speaking, neither the classically The most compelling reason torecognized components of the aging phe- consider age-associated diseases as annotype nor age-associated diseases exhibit integral part of aging comes from ourirreversibility. For example, strength exer- understanding of the evolutionary basiscise programs can result in some increase of aging. Specifically, aging is believedin muscle mass in the elderly who have to occur because of a progressive age-experienced a massive age-associated loss associated decrease in the force of natu-in muscle mass (Evans, 1995). Also, treat- ral selection; thus, deleterious traitsment with recombinant apoA-I Milano expressed only at advanced ages do notcan reverse, to some degree, coronary ath- tend to be eliminated by natural selec-erosclerosis in aged people who suffer tion. Because age-associated diseases dofrom acute coronary syndromes (Nissen not exhibit deleterious consequenceset al., 2003). until advanced ages, even if they begin Currently, most biological geronto- at an early age, current views of the evo-logists do not believe that aging is pro- lutionary basis of aging lead to the con-grammed in the evolutionary adaptive clusion that such diseases are exactlymanner that characterizes development what one would predict to be an integral(Rose, 1991). However, the rate of aging part of aging.is the result of gene-environment inter- In conclusion, the question posed in theactions (Rowe & Kahn, 1998); this is also title of this chapter (Are Age-Associatedtrue of the occurrence and progression of Diseases an Integral Part of Aging?) hasmost age-associated disease processes. In yet to be definitively answered. However,summary, age-associated diseases fit the after carefully considering the materialstime-honored, fundamental characteris- presented in this chapter, I am in totaltics of aging at least as well as the classi- agreement with the following statementcally viewed components of the aging made by Robin Holliday (2004) in hisphenotype. debate with Leonard Hayflick: “The dis- There are many reasons that the concept tinction between age-related changes thatof “normal aging” (defined as aging in the are not pathological and those that areabsence of disease) should be discarded. pathological is not at all fundamental.” ItAging without the occurrence of age- is my view that rigidly holding to theassociated disease is such a rare event that view that age-associated diseases are notthis concept focuses attention on the an integral part of aging may have had,unique rather than the typical. Of course, and is likely to continue to have, ansome may defend the use of the concept as adverse effect on aging research. It tendsa reductionist approach to the study of to limit the scope of the design of geronto-aging, and reductionism has certainly been logic studies, and to eliminate the use ofa powerful tool in biology. However, even experimental models that might well pro-in this regard, the concept falls short vide unique insights into aging processes.because, as discussed above, distinguishingbetween physiological deterioration and Referencespathophysiology is often arbitrary. While Abe, J., & Berk, B. C. (1998). Reactive oxygenthe tools for such a determination are species as mediators of signal transductionimproving, the standards used for claiming and cardiovascular disease. Trends inabsence of disease are also changing. Thus, Cardiovascular Medicine, 8, 59–64.the base to support this kind of reduction- American Heritage Dictionary, 3rd ed. (1996).ist approach to the study of aging is Boston: Houghton Mifflin.constantly shifting, and that is likely to Ames, B. N., Shigenaga, M. K., & Hagen, T. M.continue in the foreseeable future. (1993). Oxidants, antioxidants, and
CHAPTER 2 / Are Age-Associated Diseases an Integral Part of Aging? 57 the degenerative diseases of aging. Borkan, G. A., & Norris, A. H. (1980). Proceedings of the National Academy of Assessment of biological age using a profile Sciences of the USA, 90, 7915–7922. of physical parameters. Journal ofAmundson, S. A., Myers, T. G., & Gerontology, 35, 177–184. Fornace, A. J. (1998). Roles of p53 in Brody, J. A., & Schneider, E. L. (1986). Diseases growth arrest and apoptosis: putting on and disorders of aging: a hypothesis. Journal the brakes after genotoxic stress. of Chronic Disease, 39, 871–876. Oncogene, 17, 3287–3299. Bruunsgaard, H., Skinhoj, P., Pedersen, A. N.,Anderson-Ranberg, K., Schroll, M., & Jeune, B Schroll, M., & Pedersen, B. K. (2000). (2001). Healthy centenarians do not exist, Ageing, tumour necrosis factor-alpha but autonomous centenarians do: a (TNF-␣) and atherosclerosis. Clinical population-based study of morbidity Experimental Immunology, 121, 255–260. among Danish centenarians. Journal of Busse, E. W. (1969). Theories of aging. In E. W. the American Geriatrics Society, 49, Busse & E. Pfeiffer (Eds.), Behavior and 900–908. adaptation in adult life (pp. 11–32). Boston:Babiak, J., Gong, E. L., Nichols, A. V., Little Brown. Forte, T. M., Kuehl, T. J., McGill, Jr., H. C. Campisi, J. (2005). Aging, tumor suppression (1984). Characterization of HDL and and cancer: high-wire act! Mechanisms of lipoprotein intermediates to LDL and HDL Ageing and Development, 126, 51–58. in the serum of pedigreed baboons fed an Casserly, I., & Torpol, E. (2004). Convergence atherogenic diet. Atherosclerosis, 52, of atherosclerosis and Alzheimer’s disease: 27–45. inflammation, cholesterol, misfoldedBarja, G. (2002) Endogenous oxidative stress: proteins. Lancet, 363, 1139–1146. relationship to aging, longevity, and caloric Chung, H. Y., Kim, H. J., & Yu, B. P. (2001). restriction. Ageing Research Reviews, 1, The inflammation hypothesis of aging: 397–411. molecular modification by caloricBecker, A. E., de Boer, O. J., & van der restriction. Annals of the New York Wal, A. C. (2001) The role of inflammation Academy of Sciences, 928, 327–335. and infection in coronary artery disease. Cortopassi, G. A. (2002). Fixation of Annual Review of Medicine, 52, 289–297. deleterious alleles, evolution, and humanBell, D. R., & Van Zant, G. (2004). Stem aging. Mechanisms of Ageing and cells, aging, and cancer. Oncogene, 23, Development, 123, 851–855. 7290–7296. Costa, P. T. Jr., & McCrae, R. R. (1995).Bierman, E. L. (1985). Arteriosclerosis and Design and analysis of aging studies. In E. J. aging. In C. E. Finch & E. L. Schneider Masoro (Ed.), Handbook of physiology (pp. (Eds.), Handbook of the biology of aging 25–36). New York: Oxford University Press. (5th ed., pp. 842–858). New York: Van Cutler, R. G., & Semsei, I. (1989). Nostrand Reinhold. Development, cancer and aging: possibleBjorkbacka, H., Kunjathoor, V. V., Moore, K. J., common mechanisms of action and Koern, S., Ordija, C. M., Lee, M. A., Means, T., regulation. Journal of Gerontology: Halmen, K., Luster, A. D., Golenbock, D. T., Biological Sciences, 44, 25–34. & Freeman, M. (2004). Reduced Del Monte, U., & Statuto, M. (2004). Drop of atherosclerosis in MyD88-null mice links connexins: a possible link between aging elevated serum cholesterol to activation of and cancer. Experimental Gerontology, 39, innate immunity signaling pathways. 273–275. Nature Medicine, 10, 416–421. de Nigris, F., Lerman, A., Ignarro, L. J.,Blumenthal, H. T. (1999). A view of aging- Williams-Ignarro, S., Sica, V., Baker, A. H., disease relationship from age 85. Journal of Lerman, A. O., Geng, Y. J., & Napoli, C. Gerontology: Biological Sciences, 54A, (2003). Oxidative-sensitive mechanisms, B255–B259. vascular apoptosis, and atherosclerosis.Blumenthal, H. T. (2003). The aging-disease Trends in Molecular Medicine, 9, 352–359. dichotomy: true or false? Journal of DePinho, R. A. (2000). The age of cancer. Gerontology, 58A, 138–145. Nature, 408, 248–254.
58 E. J. MasoroDix, D. (1989). The role of aging in cancer Frontali, M., Sabbadini, G., Novelleto, A., incidence: an epidemiological study. Journal Jodice, C., Naso, F., Spadaro, M., Gigcenti, P., of Gerontology, 44, (Special Issue) 10–18. Jacopini, A. G., Veneziano, L., Mantuano, E.,Donehower, L. A. (2002). Does p53 affect Malaspina, P., Ulizzi, L., Brice, A., Durr, A., organismal aging? Journal of Cellular & Terremato, L. (1996). Genetic fitness in Physiology, 192, 23–33. Huntington’s disease and spinocerebellarElby, E. M., Parhad, I. M., Hogan, D. B., & ataxia: a population genetics model of CAG Fang, T. S. (1994) Prevalence and types of repeat expansions. Annals of Human dementia in the very old: results from the Genetics, 60, 423–435. Canadian Study of Health and Aging. Geiger, H., & Van Zant, G. (2002). The aging Neurology, 44, 1593–1600. of lympho-hematopoietic stem cells.Ershler, W. B., & Longo, D. L. (1997). Aging Nature Immunology, 3, 329–333. and cancer: issues of basic and clinical Gessert, C. E., Elliott, B. A., & Haller, I. V. science. Journal of the National Cancer (2002). Dying of old age: an examination of Institute, 89, 1489–1497. death certificates of MinnesotaEvans, J. G. (1988). Aging and disease. In D. centenarians. Journal of the American Everid & J. Whalen (Eds.), Research and the Geriatrics Society, 50, 1561–1565. aging population (pp. 38–57). Ciba Greaves, D. R., & Gordon, S. (2001). Immunity, Foundation Symposium No. 134, atherosclerosis, and cardiovascular disease. Chichester, UK: Wiley. Trends in Immunology, 22, 180–181.Evans, W. J. (1995). Effects of exercise on body Hasty, P. (2005). The impact of DNA damage, composition and functional capacity in the genetic mutation and cellular responses on elderly. Journal of Gerontology, 50A, cancer prevention, longevity and aging: 147–150. observations in humans and mice.Evert, J., Lawler, E., Bogan, H., & Perls, T. Mechanisms of Ageing and Development, (2003). Morbidity profiles of centenarians: 126, 71–77. survivors, delayers, escapers. Journal of Hasty, P., Campisi, J., Hoeijmakers, J., van Gerontology: Medical Sciences, 58A, Steeg, H., & Vijg, J. (2003). Aging and 232–237. genome maintenance: lessons from theFaxon, D. P., Fuster, V., Libby, P., Beckman, J. A., mouse? Science, 299, 1355–1359. Hiatt, W. R., Thompson, R. W., Hayflick, L. (2002). Anarchy in gerontological Topper, J. N., Annex, B. H., Rundback, J. A., terminology. Gerontologist, 42, 416–421. Fabunmi, R. P., Robertson, R. M., & Hebert, K. (2004). Life expectancy in Great Loscalzo, J. (2004). Atherosclerotic vascular Britain rises—but later years are still spent disease conference group III: in poor health. British Medical Journal, pathophysiology. Circulation, 109, 329, 250. 2617–2625. Hill, J. M., Zalos, G., Halcox, J. P. T.,Finch, C. E., & Crimmins, E. M. (2004). Schenke, W. H., Waclawi, W., Inflammatory exposure and historical Quyyumi, A. A., & Finkel, T. C. (2003). changes in human life-spans. Science, 305, Circulating endothelial progenitor cells, 1736–1739. vascular function, and cardiovascular risk.Finkel, T., & Holbrook, N. J. (2000). Oxidants, New England Journal of Medicine, 348, oxidative stress, and the biology of aging. 593–600. Nature, 408, 239–247. Hitt, R., Young-Xu, Y., Silver, M., & Perls, T.Fries, J. F. (1980). Aging, natural death, and (1999). Centenarians: the older you get, the the compression of morbidity. New healthier you have been. Lancet, 354, 652. England Journal of Medicine, 303, 130–135. Ho, E., Boileau, T., & Bray, T. M. (2004).Fontana, L., Meyer, T. E., Klein, S., & Dietary influence in endocrine- Holloszy, J. O. (2004). Long-term calorie inflammatory interactions in prostate restriction is highly effective in reducing cancer. Archives of Biochemistry and the risk of atherosclerosis in humans. Biophysics, 428, 109–117. Proceedings of the National Academy of Hofman, A., Ott, A., Breteler, M. M., Sciences of the USA, 101, 6659–6663. Bots, M. L., Slooter, A. J., van Harskamp, F.,
CHAPTER 2 / Are Age-Associated Diseases an Integral Part of Aging? 59 van Duijin, C. N., Van Broeckhoven, C., & Kim, S-h., Kominker, P., & Campisi, J. (2002). Grobbee, D. E. (1997) Atherosclerosis, Telomeres, aging, and cancer: in search of a apolipoprotein E, and prevalence of happy ending. Oncogene, 21, 503–511. dementia and Alzheimer’s disease Kirkwood, T. B. L. (1977). Evolution of ageing. in the Rotterdam study. Lancet, 349, Nature, 270, 301–304. 151–154. Klima, M. P., Povysil, C., & Teasdale, T. A.Holehan, A. M., & Merry, B. J. (1986). (1997). Causes of death in geriatric patients: The experimental manipulation of a cross-cultural study. Journal Gerontology: ageing by diet. Biological Reviews, 61, Medical Sciences, 52A, M247–M253. 329–369. Kohn, R. R. (1982). Cause of death in very oldHolliday, R. (2004). The close relationship people. Journal of the American Medical between biological aging and age-associated Association, 247, 2793–2797. pathologies in humans. Journal of Krabbe, K. S., Pedersen, M., & Bruunsgaard, H. Gerontology: Biological Sciences, 59A, (2004) Inflammatory mediators in 543–546. the elderly. Experimental Gerontology, 39,Horiuchi, S., Finch, C. E., Mesle, F., & Vallin, J. 687–699. (2003). Differential patterns of age-related Krtolica, A., Parrinello, S., Lockett, S., mortality increase in middle age and old age. Despres, P-Y, & Campisi, J. (2001). Journal of Gerontology: Biological Sciences, Senescent fibroblasts promote epithelial 58A, 495–507. cell growth and tumorigenesis: a linkHornsby, P. J. (2002). Cellular senescence and between cancer and aging. Proceedings of tissue aging in vivo. Journal of the National Academy of Sciences of the Gerontology: Biological Sciences, 57A, USA, 98, 12072–12077. B251–B256. Lakatta, E. G. (1985). Health, disease, andHughes, K. A., Alipaz, J. A., Drnevitch, J. A., cardiovascular aging. In Committee on an & Reynolds, R. M. (2002). A test of the Aging Society, Institute of Medicine and evolutionary theories of aging. National Research Council (Ed.), America’s Proceedings of the National Academy of aging-health in an older society (pp. Sciences of the USA, 99, 14286–14291. 73–104). Washington, DC: NationalIrminger-Finger, I. (2003). Geneva Workshop Academy Press. 2002. Cancer, apoptosis, and aging. Lakatta, E. G. (2003). Arterial and cardiac Biochimia et Biophysica Acta, 1653, aging: major shareholders in cardiovascular 41–45. disease enterprise: part III: cellular andJala, V. R., & Haribabu, B. (2004). molecular clues to heart and arterial aging. Leukotrienes and atherosclerosis: new roles Circulation, 107, 490–497. for old mediators. Trends in Immunology, Lakatta, E. G., & Levy, D. (2003). Arterial and 25, 315–322. cardiac aging: major shareholders inJones, P. P., Christou, C., Jordan, J., & Seals, R. cardiovascular disease enterprise: part I: (2003). Baroreflex buffering is reduced with aging arteries: a “set up” for vascular age in healthy men. Circulation, 107, disease. Circulation, 107, 139–146. 1770–1774. Landmesser, U., Hornig, B., & Drexler, H.Kalu, D. N. (1995). Bone. In E. J. Masoro (2004). Endothelial function: a critical (Ed.), Handbook of physiology (pp. determinant in atherosclerosis. 395–412). New York: Oxford University Circulation, 109 [Suppl. II], II-27–II-33. Press. Lee, R. D. (2003). Rethinking theKhovidhunkit, W., Kim. M-S., Memon, R. A., evolutionary theory of aging. Transfers, Shivenaga, J. K., Moser, A. H., Feingold, K., not births, shape senescence in social & Greenfeld, C. (2004). The pathogenesis species. Proceedings of the National of atherosclerosis: effect of infection Academy of Sciences of the USA, 100, and inflammation on lipid and lipoprotein 9367–9642. metabolism and consequences to the Libby, P., Ridker, P. M., & Maseri, A. (2003). host. Journal of Lipid Research, 45, Inflammation and atherosclerosis. 1169–1196. Circulation, 105, 1135–1143.
60 E. J. MasoroLindeman, R. D., Tobin, J. D., & Shock, N. W. Age-dependent induction of hepatic tumor (1985). Longitudinal studies on the rate of regression by the tissue microenvironment decline in renal function with age. Journal after transplantation of neoplastically of the American Geriatrics Society, 33, transformed rat liver epithelial cells into 278–285. the liver. Cancer Research, 57, 1807–1813.Lipman, R., Dallal, G. E., & Bronson, R. T. Medawar, P. B. (1952). An unsolved problem (1999). Lesion biomarkers of aging in of biology. Oxford: Oxford University Press. B6C3F1 hybrid mice. Journal of Miller, R. A. (1991) Gerontology as oncology. Gerontology: Biological Sciences, 54A, Cancer, 68, 2496–2501. B466–B477. Miyaishi, O., Ando, F., Matsuzawa, R., &Lipman, R., Galecki, A., Burke, D. T., & Isobe, K-I. (2000). Cancer incidence in old Miller, R. A. (2004). Genetic loci that age. Mechanisms of Ageing and influence cause of death in a heterogeneous Development, 117, 47–55. mouse stock. Journal of Gerontology: Navab, M., Ananthramaiah, G. M., Reddy, S. T., Biological Sciences, 59A, 977–983. Van Lenten, B. J., Ansell, B. J., Fonarow, G. C.,Liu, L., Wylie, R. C., Andrews, L. G., & Vahabzadeh, K., Hama, S., Hough, G., Tollefsbol, T. O. (2003). Aging, cancer, and Kamranpour, N., Berliner, J. A., Lusis, A., & nutrition: the DNA methylation Fogelman, A. M (2004). The oxidation connection. Mechanisms of Ageing and hypothesis of atherogenesis: the role of Development, 124, 989–998. oxidized phospholipids and HDL. Journal ofLongo, V. D., & Finch, C. E. (2002). Genetics Lipid Research, 45, 993–1007. of aging and diseases: from rare mutations Nelson, W. G., De Marzo, A. M., and model systems to disease prevention. DeWeese, T. L., & Isaacs, W. B. (2004). Archives of Neurology, 59, 1706–1708. The role of inflammation in theLusis, A. J. (2000). Atherosclerosis. Nature, pathogenesis of prostate cancer. Journal 407, 233–241. of Urology, 172, S6–S12.Maeda, H., Gleiser, C. A., Masoro, E. J., Nissen, S. E., Tsunoda, T., Turzae, R. M., Murata, I., McMahan, C., & Yu, B. P. (1985). Schoenhagen, P., Cooper, C. J., Yasin, M., Nutritional influences of aging of Fischer Eaton, G. M., Lauer, M. A., Sheldon, W. S., 344 rats: II. pathology. Journal of Grines, C. L., Halpern, S., Crowe, T., Gerontology, 40, 671–688. Blankenship, J. C., & Kerensky, R. (2003).Maier, B., Gluba, W., Bernier, B., Turner, I., Effect of recombinant apoA-I Milano on Mohammad, K., Guise, T., Sutherland, A., coronary atherosclerosis in patients with Thomer, M., & Scrable, H. (2004). acute coronary syndromes. Journal of the Modulation of mammalian life span by the American Medical Association, 290, short isoform of p53. Genes and 2292–2300. Development, 18, 306–319. Papciak, S. M., Song, L., Nagel, J. E., &Martin, G. M. (2002) The evolutionary Adler, W. H. (1996). Immune system. In substrate of aging. Archives of Neurology, J. E. Birren (Ed.), Encyclopedia of 59, 1702–1705. gerontology (Vol. 1, pp. 753–759). SanMarx, J. (2004). Inflammation and cancer: the Diego: Academic Press. link grows stronger. Science, 306, 966–968. Park, I-K., Morrison, S. J., & Clarke, M. I.Masoro, E. J. (Ed.) (1995). Handbook of (2004). Bmi 1, stem cells, and senescence. physiology, Section 11, Aging. New York: Journal of Clinical Investigation, 113, Oxford University Press. 175–179.Masoro, E. J. (2002). Caloric restriction: A key Pedersen, M., Bruunsgaard, H., Weis, N., to understanding and modulating aging. Hendel, H. W., Andreassen, B. U., Eldrup, E., Amsterdam: Elsevier Science. Dela, F., & Pedersen, B. K. (2003).McClearn, G. E. (1997). Biomarkers of age and Circulating levels of TNF-alpha and Il-6- aging. Experimental Gerontology, 32, relation to truncal fat mass and muscle 87–94. mass in healthy elderly individuals and inMcCullough, K. D., Coleman, W. B., patients with type-2 diabetes. Mechanisms Smith, G. J., & Grisham, J. W. (1997). of Ageing and Development, 124, 495–502.
CHAPTER 2 / Are Age-Associated Diseases an Integral Part of Aging? 61Perlman, R. M. (1954). The aging syndrome. Scott, J. (2004). Pathophysiology and Journal of the American Geriatrics Society, biochemistry of cardiovascular disease. 2, 123–129. Current Opinion in Genetics andPerls, T. (2004). Dementia-free centenarians. Development, 14, 271–279. Experimental. Gerontology, 39, 1587–1593. Scully, J. L. (2004). What is a disease? EMBORattan, S. I. S. (1991). Ageing and disease: Report, 7, 650–653. proteins as the molecular link. Perspectives Seddon, J. M., Gensler, G., Klein, M. L., & in Biology and Medicine, 34, 526–533. Rifai, N. (2004). Association betweenRauscher, F., Goldschmidt-Clermont, P., C-reactive protein and age-related macular Davis, B. H., Wang, T., Gregg, D., degeneration. Journal of the American Ramaswami, P., Pippen, A., Annex, B. H., Medical Association, 291, 704–710. Dong, E., & Taylor, D. A. (2003). Aging, Semsei, I. (2000). On the nature of aging. progenitor cell exhaustion, and Mechanisms of Ageing and Development, atherosclerosis. Circulation, 108, 457–463. 117, 93–108.Reape, T. J., & Groot, P. H. E. (1999). Sharpless, N. E., & DePinho, R. A. (2004). Chemokines and atherosclerosis. Telomeres, stem cells, senescence and Atherosclerosis, 147, 213–225. cancer. Journal of Clinical Investigation,Robbie, L., & Libby, P. (2001). Inflammation 113, 160–168. and atherosclerosis. Annals of the New Shock, N. W. (1961). Physiological aspects of York Academy of Sciences, 947, 167–180. aging. Annual Review of Physiology, 23,Rose, M. R. (1991). Evolutionary biology of 97–122. aging. New York: Oxford University Press. Shock, N. W. (1967). Physical activity and theRoss, R. (1986). The pathogenesis of “rate of aging.” Journal of the Canadian atherosclerosis. New England Journal of Medical Association, 96, 836–840. Medicine, 314, 488–500. Shock, N. W. (Ed.) (1984). Normal humanRoss, R. (1993). The pathogenesis of aging: the Baltimore longitudinal study of atherosclerosis. Nature, 801–809. aging (NIH Publ. 84–2450). Washington,Ross, R. (1999). Atherosclerosis: an DC: U.S. Government Printing Office. inflammatory disease. New England Sinclair, A. J., & Croxon, S. C. M. (2003). Journal of Medicine, 340, 115–126. Diabetes mellitus. In R. Tallis & H. FillitRowe, J. W., & Kahn, R. W. (1998). Successful (Eds.), Brocklehurst’s textbook of geriatric aging. New York: Pantheon Books. medicine and gerontology (6th ed., pp.Rowe, J. W., Andres, R., Tobin, J. D., 1193–1218). London: Churchill Livingstone. Norris, A. H., & Shock, N. W. (1976). Skoog, I., Blennow, K., & Marcusson, J. (1996). The effect of age on creatinine clearance in Dementia. In J. E. Birren (Ed.) Encyclopedia men: a cross-sectional and longitudinal of gerontology (Vol. 1, pp. 383–413). San study. Journal of Gerontology, 31, 155–163. Diego: Academic Press.Rowe, J. W., Wang, S. Y., & Elahi, D. (1990). Smith, D. W. E. (1996) Cancer mortality at Design, conduct, and analysis of human very old ages. Cancer, 77, 1367–1372. aging research. In E. L Schneider & Steinberg, D. (1997). Low density lipoprotein J. W. Rowe (Eds.), Handbook of the biology oxidation and its pathobiological of aging (3rd ed., pp. 63–71). San Diego: significance. Journal of Biological Academic Press. Chemistry, 272, 20963–20966.Schwartsburd, P. M. (2004). Age-promoted Stout, R. W. (2003). Atherosclerosis and lipid creation of a pro-cancer microenvironment: metabolism. In R. Tallis & H. Fillit (Eds.), pathogenesis of dyscoordinated feedback Brocklehurst’s textbook of geriatric control. Mechanisms of Ageing and medicine and gerontology (6th ed., pp. Development, 125, 581–590. 355–370). London: Churchill Livingstone.Scott, D. L. (2003). Arthritis in the elderly. Strehler, B. L. (1977). Time, cells and aging. In R. Tallis & H. Fillit (Eds.), Brocklehurst’s 2nd ed. New York: Academic Press. textbook of geriatric medicine and Tyner, S. D., Venkatachalam, S., Choi, J., gerontology (6th ed., pp. 887–904). London: Jones, S., Ghebranious, N., Ingelmann, S., Churchill Livingstone. Lu, X., Soron, G, Cooper, B., Brayton, C.,
62 E. J. Masoro Park, S. H., Thompson, T., Karsenty, G., the American Geriatrics Society, 48, Bradley, A., & Donehower, L. (2002). p53 1462–1469. mutant mice that display early aging- Wick, G., Berger, P., Jansen-Durr, P., & associated phenotypes. Nature, 415, Grubeck-Loebenstein, B. (2003). A 45–53. Darwinian-evoltuionary concept of age-Urban, L., Bessenyei, B., Marka, M., & related diseases. Experimental Gerontology, Semsei, M. (2002). On the role of aging in 38, 13–25. the etiology of autoimmunity. Gerontology, Wick, G., Jansen-Dürr, Berger, P. Blasko, I., & 179–184. Grubeck-Loebenstein, B. (2000). Diseases ofVan Zant, G., & Liang, Y. (2004). The role of aging. Vaccine, 18, 1567–1583. stem cells in aging. Experimental Williams, G. C. (1957). Pleiotropy, natural Hematology, 31, 659–672. selection, and the evolution of senescence.Viera, C., Pasyukova, E. G., Zeng, Z. B., Evolution, 11, 398–411. Hackett, J. B., Lyman, R. F., & Mackay, T. Witzum, J. L. (1994). The oxidation hypothesis (2000). Genotype-environment interaction of atherosclerosis. Lancet, 344, 793–795. for quantitative trait loci affecting life span Yaffe, K., Lindquist, K., Pennix, B. W., in Drosophila melanogaster. Genetics, 154, Simonsick, E. M., Pahor, M., Kritchevsky, S., 213–227. Launer, L., Kuller, L., Rubin, S., & Harris, T.Vijg, J., & Dolle, M. E. T. (2001). Instability of (2003). Inflammatory markers and cognition the nuclear genome and the role of DNA in well-functioning African American and repair. In E. J. Masoro & S. N. Austad (Eds.), white elders. Neurology, 61, 76–80. Handbook of the biology of aging (5th ed., Zhang, Q., Powers, E. T., Nieva, J., Huff, M. E., pp. 84–113). San Diego: Academic Press. Dendle, M. A., Bieschke, J., Glabe, C. G.,Vijg, J., & Dolle, M. E. (2002). Large genome Eschenmoser, A., Wentworth, Jr., P., rearrangements as a primary cause of aging. Lerner, R. A., & Kelly, J. W. (2004) Mechanisms of Ageing and Development, Metabolite-initiated misfolding may trigger 123, 907–915. Alzheimer’s disease. Proceedings of theVon Straus, E., Fratiglioni, L., Viitanen, M., National Academy of Sciences of the USA, Forsell, Y., & Winblad, B. (2000). Morbidity 101, 4752–4757. and comorbidity in relation to functional Zhang, Y., & Herman, B. (2002). Ageing and status: a community-based study of the apoptosis. Mechanisms of Ageing and oldest old (90ϩ years). Journal of Development, 123, 245–260.
64 D. A. Sinclair and K. T. Howitzmight tease them out by studying simpler proximal causes of aging and the path-organisms. ways that regulate the pace of aging—the Most scientists studying highly com- two seemingly separate arenas of agingplex organisms such as rodents and research—now appear to be manifesta-humans have tended to focus on proximal tions of the same underlying ancientcauses of aging. In contrast, researchers mechanisms that evolved to help organ-studying less complex organisms such as isms survive periods of adversity.yeast, worms, and flies—which are There have been many theories as toamenable to unbiased genetic screens for how DR works, and many of them havelong-lived organisms—have isolated genes fallen out of favor or have been outrightthat regulate the pace of aging. In the past disproved. In this chapter, both the oldfew years, these so-called “longevity and new theories of DR will be pre-genes” have been shown to respond to sented, partly to give a sense of how theenvironmental conditions, such as a lack field has evolved, but also because itof food (see Figure 3.1). As we will see in provides a context in which to show howthis chapter, the proximal causes of aging many of these theories can be unitedand the longevity regulators are both inti- under one: the Hormesis Hypothesis ofmately connected to DR and both are DR (Anderson et al., 2003; Iwasaki et al.,essential to understanding what underlies 1988; Masoro, 1998; Mattson et al., 2002a;this phenomenon. The effect of DR on the Turturro et al., 2000). Identification of DNA Damage Disposable Soma Characterization of daf-2 genes that link diet hypothesis of aging hypothesis of aging mutation in C.elegans and stress to longevity 1950 1960 1970 1980 1990 2000 2010 Diet and Environment Longevity Genes Longevity Genes Aging genes Cell Defenses Cell Defenses Cell Defenses AGING AGING AGING AGINGFigure 3.1 Changing views about aging. Before the 1970s, the predominant view was that aging was causedby “death” genes that directed the process, as if it were simply an extension of development. Evolutionarybiologists argued that aging is not adaptive for most species, and this idea was laid to rest. During the late1980s and 1990s, genetic screens in simple organisms such as yeast and worms uncovered single mutationsthat could dramatically extend life span, seemingly contradicting the complexity of aging. Friedman andJohnson’s discovery (1988) of age-1 mutations that extend worm life span was seminal. Around the turn ofthe 21st century, it became apparent that longevity genes have evolved to protect organisms during times ofadversity and that they are activated by low nutrition and other mild biological stress.
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 65 This chapter will begin with a summary of dietary restriction was conductedof the key discoveries in DR research and between 1934 and 1935, at a time whenthen will briefly summarize the dramatic aging was considered part of an organ-physiological effects and health benefits of ism’s developmental process. To CliveDR. The central section will summarize McCay and his colleagues at Cornellthe theories about how CR works, dispel University, it seemed reasonable that ifsome reoccurring myths, and discuss how, the pace of development of an animalwith a shift in thinking, many of the cur- were slowed, say by restricting foodrent (and valid) theories on DR can be intake, its life span would correspond-united by the Hormesis Hypothesis. The ingly increase. They tested this ideachapter will conclude by presenting excit- by substituting 10 to 20 percent of theing new avenues of research into the rats’ diet with indigestible cellulose, thusdevelopment of small molecules that can cutting back on their caloric intakemimic the beneficial effects of DR, and (Krystal & Yu, 1994; McCay, 1934). Con-will attempt to predict where this rapidly sistent with predictions, the underfedmoving field might take us in the coming rats did develop slower and lived sub-decades. stantially longer, although the hypothe- sis was ultimately proven incorrect. These first experiments of McCay II. Key Discoveries were not hailed as a breakthrough, partly because they were not performed underA. The DR Paradigm strictly controlled conditions on homo-Throughout human history, numerous geneous groups of animals, and partlysocieties have touted the health benefits because the data conflicted with theof frugal eating habits, including the current dogma that that poorly fed miceAncient Greeks and Ancient Romans die sooner. But McCay, certain he was(Dehmelt, 2004). In more recent times, right, had already begun another exper-there are also accounts of specific individ- iment, this time on large cohorts withuals who have fared better on a very lean carefully controlled diets. Over 100 ani-diet. A fifteenth-century Venetian noble- mals were used. He and his colleaguesman named Luigi Cornaro, for example, Mary Crowell and Leonard Maynardis famed for his supposed 1,400-calorie published a landmark paper titled “Thediet of meat, bread, and wine, which he Effect of Retarded Growth Upon themaintained from the age of 27 until his Length of Life Span and Upon the Ulti-death at 103. More recently, Professor mate Body Size” (McCay et al., 1935).Maurice Guéniot, a president of the Paris In it they reported:Medical Academy at the turn of the 20thcentury, is famed for having lived on a the experiment is in its fourth year, butrestricted diet and for dying at the age of the results are [already] conclusive in102. But of course these are not scientific showing that the animals that maturestudies and cannot be treated as more slowly have a much greater life span thanthan traditions or anecdotes. the rapidly growing ones. . . . This exten- sion of the life span by means of retarded In 1917, Osborne & Mendel (1917) growth indicates that the potential lifepublished the first scientific study show- span for a given species is much longering that restricting food extended life, than has been anticipated. Furthermore,but it had little impact because a publi- these data suggest that the longer lifecation by Robertson and Gray reported span of the female may be related to thejust the opposite (Robertson & Ray, slower growth rate of the female sex as1920). The first widely recognized study the animal approaches maturity.
66 D. A. Sinclair and K. T. Howitz Although their interpretations of the found that they had higher cal/kg/hresults are now considered incorrect, metabolism than the ad lib (AL)-fedthe results themselves are highly repro- counterparts. The animals also failed toducible. They have been validated in undergo the usual age-related decline indozens of laboratories and with differ- metabolic rate between 850 to 1,150 daysent versions of the diet and using differ- of age (rats typically live ϳ1,000 days).ent strains of rats and mice (Krystal & This DR-dependent increase in meta-Yu, 1994; Weindruch & Walford, 1988). bolic rate has been validated subse-One recent study went so far as to test quently by multiple rodent studiesthree rat and four mouse genotypes described below, and for other specieseach with different life spans and differ- such as Caenorhabditis elegans andent spectra of disease susceptibility Drosophila melanogaster (Braeckman(Turturro et al., 1999). When subjected et al., 2002; Hulbert et al., 2004). To thisto DR, all of them lived significantly day, the misconception that DR workslonger on average and were less prone to by slowing metabolic rate persists.age-associated diseases. It is worth not-ing, however, that the optimal amount B. Timing of the Dietof DR is still debated and is likely geno-type-specific. Weindruch (1996), for Shortly after the work of McCay, a setexample, severely restricted caloric of researchers began testing the effectsintake of a strain of mouse by 65 per- of diet composition and implementingcent (i.e., feeding them 35 percent ad DR at different ages and at morelibitum amount) and this was their moderate levels of restriction. Table 3.1longest-lived group. King & Visscher (1950) reported that To characterize the long-lived animals, 33 percent food restriction was benefi-Will & McCay (1941) determined their cial, whereas 50 percent was too severe.metabolic rate and, to their surprise, Ross found that a 30 percent reduction in Table 3.1 The Effect of Different DR Regimes on Rat Mortality Rates Study Mortality Rate Doubling Time, MRDT (Years)a AL CR Ratio CR/AL Ross, 1959 0.17 0.38 2.23 Berg and Simms, 1960 0.17 0.27 1.59 Ross and Bras, 1973 10% protein 0.30 0.42 1.40 22% protein 0.28 0.65 2.32 51% protein 0.33 0.57 1.72 Goodrick et al. (1983) For 10 months 0.25 0.44 1.76 For 18 months 0.31 0.59 1.90 Yu et al. (1982) 0.30 0.53 1.77 Yu et al. (1985) After 6 weeks 0.19 0.27 1.42 6 weeks to 9 months 0.19 0.19 1.00 After 6 months 0.19 0.32 1.63 aPost-maturityMortality Rate Doubling Time (MRDT) is an indicator of the rate of aging of a population. The higher the number, the slower the rate at which mortality increases with age. Adapted from Krystal and Yu, 1994.
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 67caloric intake was close to optimal in days of being switched to a full dietrats, with a doubling of life expectancy (Good & Tatar, 2001).and an increase in maximum life spanof more than 30 percent (Ross, 1972). C. Diet CompositionCarlson and Hoetzel discovered that theycould extend the mean and maximum Several laboratories have been instrumen-life span of rats about 10 percent by fast- tal in addressing the question of whethering one day out of every four (Carlson & the DR effect is due to the total reductionHoetzel, 1946), but the benefits of in food intake or the lack of a specificrestricting food only a few days in a week component. Early on, reduced proteinare highly dependent on the animal’s consumption was suspected as the majorgenotype and the age at which the feed- cause of the effect. Indeed, low-proteining regimen is implemented (Goodrick diets effectively reduce the incidence ofet al., 1990). Nelson and Halberg (1986) kidney disease and can slightly extendshowed that the life span extension is life span (Bras & Ross, 1964; Iwasakinot due to the animals eating only once et al., 1988), but total caloric restrictionversus the AL animals who nibble always has a greater effect on life spanthroughout the day. (Masoro, 1985). Interestingly, Yu and col- In 1947, Ball and colleagues (1947) leagues found that DR animals consumereported that lifelong DR dramatically about the same total number of caloriesextends the life span of mice (max life during their life span as AL-fed animalsspan ϭ 850 days vs. 550 days), whereas (36,000 Kcal), leading to speculation thatDR for the first 240 days had little effect. life span might be related to the totalSwitching animals from DR back to a number of calories consumed per ratnormal AL diet after 8 months gave an per lifetime (Yu et al., 1985). Neither theintermediate result (max ϭ 600 days). restriction of minerals nor fat affected lifeAlthough it was not appreciated at the span (Iwasaki et al., 1988)time, this paper effectively disproved A surprise finding has been that severeMcCay’s hypothesis that DR works by restriction of a single amino acid, methio-slowing development. There have been nine, is sufficient to extend rat lifenumerous studies since, in mice, rats, span (Orentreich et al., 1993; Richie et al.,golden hamsters, worms, flies, and yeast 1994, 2004). The same observations have(Krystal & Yu, 1994; Weindruch & been made in lower organisms such asWalford, 1988). yeast, worms, and flies (Bitterman et al., Today, it is generally accepted that 2003; Braeckman et al., 2001b; GemsDR implemented during only the first et al., 2002; Tatar et al., 2003). For exam-few months of life of a rodent can have ple, yeast life span is extended not only bysmall longevity benefits, but the most restricting calories (i.e., glucose) but alsoefficacious and reliable treatment for by restricting amino acids, nitrogen, or byextending life span is long-term DR heat or osmotic stress (Bitterman et al.,(Weindruch & Walford, 1988). With 2003). Similarly, the life span of fliesregards the effect of DR on young ani- and worms can be extended not only bymals, it is likely that the diet has lasting restricting their traditional food sourceseffects into adulthood because it perma- (i.e. yeast and bacteria) but also bynently alters the endocrinological state mild stresses such as heat shock andof the animal, but this remains to be overcrowding (Braeckman et al., 2001b;proven. For flies, there is no lasting Michalski et al., 2001; Walker et al.,effect of DR. In fact, their mortality rate 2003). Thus, the important determinant ofreturns to that of well-fed flies within longevity is not necessarily the restriction
68 D. A. Sinclair and K. T. Howitzof calories, but rather any nutrient defi- A. Rodentsciency that can invoke a survival response For both mice and rats, DR is highly effec-(Braeckman et al., 2001b; Lamming et al., tive at slowing age-dependent physiologi-2004; Turturro et al., 2000). These obser- cal decline in various tissues and systems,vations support the Hormesis Hypothesis, such as muscle (i.e., sarcopenia) and thewhich states that DR is a mild stress that immune system (Dempsey et al., 1993;provokes a survival response in the organ- McKiernan et al., 2004; Pahlavani, 2004;ism. This stress response then boosts Spaulding et al., 1997), although there areresistance to biological and chemical exceptions (Sun, 2001). DR also delays theinsults and counteracts the causes of occurrence of almost every disease associ-aging. A more detailed description of this ated with aging, including heart disease,theory is presented in a later section. cataracts, diabetes, and neurodegeneration In summary, over the past 70 years, the (Weindruch & Walford, 1988). Moreover,DR paradigm has proven extremely DR is the most potent, broadly actingrobust. It can be observed for varying anti-cancer regimen we know of in ani-extents of restriction, diet composition, mal models (Hursting et al., 2003; Raffoultime and duration of the treatment, for a et al., 1999). DR lowers body temperaturevariety of different genetic backgrounds and increases the physical activity of miceand species. It is the robustness and con- and rats, although neither of these effectsservation of the DR paradigm that makes is considered a likely cause of increasedit such a powerful tool for the investiga- longevity, with perhaps the exception oftion of the mechanisms that cause aging its anti-cancer effects (Koizumi et al.,and the pathways that regulate aging in 1996; McCarter et al., 1997). Numerousresponse to environmental conditions. studies have shown that DR animals are more resistant to a variety of toxins and drugs, such as isoproternol, an asthma III. Physiological Effects of DR drug, and ganacyclovir, an antiviral drug on Mammals (reviewed by Hart et al., 1995). The beneficial action of DR against manyA large body of literature has been accu- of these toxins seems to be two-fold:mulated over the past 70 years on the altered drug metabolism in vivo, leadingeffects of DR on mammalian physiology. to increased excretion or less conversionMany of the key findings about the of the molecule to its toxic form, andphysiological effects of DR are better the resistance of individual cells to thediscussed within the context of specific toxin or drug.hypotheses about how DR works (see At the cellular level, DR modulatesbelow). As with previous editions of this several fundamental processes that maybook, it is neither possible nor appro- be intimately involved in aging. For exam-priate to present more than a summary ple, DR retards the age-related decline inof the effects of DR within a chapter certain DNA repair capacities (Guo et al.,of this length. Excellent reviews on the 1998; Lipman et al., 1989; Weraarchakulphysiological effects of DR have been et al., 1989), although not all types ofpublished elsewhere (Krystal & Yu, 1994; DNA repair or tissues are affected (Haley-Masoro, 2000; Weindruch & Walford, Zitlin & Richardson, 1993; Prapurna &1988). Here, our discussion will focus on Rao, 1996). The ability of DR to reducethe effects of DR that are likely to be an accumulation of damage to proteins,relevant to understanding the basis of lipids, and DNA during aging is alsothe phenomenon, and on the latest find- well documented in dozens of studiesings at the cellular and molecular levels. (Dempsey et al., 1993; Moore et al., 1995;
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 69Ward, 1988; Xia et al., 1995). As one of Wisconsin (Kemnitz et al., 1993). Toexample, the rate at which mutations date, the results strongly suggest thataccumulate during aging (or following the same beneficial effects observed inexposure to DNA-damaging agents such DR rodents also occur in primatesas bleomycin) is markedly lower in DR (Edwards et al., 1998; Zainal et al., 2000).animals (Aidoo et al., 1999). Hsp70 is a The monkeys matured more slowly andprotein-folding chaperone that also blocks achieved shorter stature than controlsapoptosis by binding to Apaf-1 (Ravagnan (Roth et al., 2001). Other notable physio-et al., 2001). The ability of cells to induce logical changes included lower plasmahsp70 in response to heat shock remains insulin levels; reduced total cholesterol,relatively high in DR animals, seemingly triglycerides, blood pressure, and arterialdue to the preservation of a transcription stiffness; higher HDL (good cholesterol);factor that binds to the gene’s promoter and slower decline in circulating levels(Heydari et al., 1993; Pahlavani et al., of a hormonal marker of aging, DHEAS.1995). DR also alters the susceptibility of These biomarkers suggest that DR pri-cells to apoptosis, but in which direction mates are aging more slowly and will beseems to depend on the cell type and less likely to incur diseases of aging suchthe degree of damage inflicted (reviewed as diabetes and cardiovascular diseasein Zhang & Herman, 2002). (Roth et al., 2001). Bone mass is slightly Various age-associated changes in reduced, but in approximate proportionhormones, cytokines, and neurotransmit- to the smaller body size.ters are attenuated by DR. For example, Is there evidence that DR would extendthe decrease in cholinergic and dopamin- the maximum life span of humans? Short-ergic stimulation of inositol phosphate term markers of DR in rodents appear tosignaling components is reduced (Undie occur in people on strict low-calorie diets,& Friedman, 1993), and the age-dependent and six month pilot-scale trials of DR indecrease in the secretion of growth humans have been initiated (Roberts ethormone and IGF-1 is attenuated by DR al., 2001). But in the absence of large(D’Costa et al., 1993). Recently, micro- cohorts of clinically controlled subjects,array transcriptional profiles of DR some researchers have turned to membersanimals have identified some of the genes of the population who have taken itthat could underlie the physiological upon themselves to restrict calories, someeffects of DR. In summary, the genes that for as long as a decade. The first of suchare affected by DR indicate increases in studies was recently published and, in agluconeogenesis, the pentose phosphate sample of 18 individuals who had been onpathway, and protein turnover (Cao et al., DR for an average of six years, there2001; Prolla, 2002; Sreekumar et al., 2002; was good evidence of improved cardiovas-Weindruch et al., 2001), with conflicting cular health (Fontana et al., 2004). Thedata on whether genes that regulate apop- DR group also had lower serum levels oftosis are decreased (Weindruch et al., various indicators of health such as2001) or increased (Cao et al., 2001). LDL, triglycerides, fasting glucose, fasting insulin, and C-reactive protein. Blood pressure was lower, and the intima-mediaB. Primates of the carotid artery was 40 percent thin-In 1987, the first controlled study of ner. Members of the Biosphere 2 crew,CR in rhesus and squirrel monkeys was who were subjected to a low-calorie dietinitiated at the National Institute on (1,750 to 2,100 kcal/d) for two years, alsoAging (Ingram et al., 1990). A similar experienced hematologic, physiologic,study is also underway at the University hormonal, and biochemical alterations
70 D. A. Sinclair and K. T. Howitzthat resembled those of DR rodents and into a single unifying theory. It is also themonkeys (Walford et al., 2002). Although first theory that can explain most of thethese results are encouraging, the observations in the field, many of whichlongevity benefits and potential health were seemingly disconnected. It is as ifissues arising from long-term DR in scientists have been holding the samehumans remain to be established. animal from different ends without real- izing it until now. First, we will review how the field of DR restriction evolved; IV. Mechanisms of DR we will then discuss the paradigm shift that led us to where we are today.Although there have been many theorieson how DR works, we appear to be on theverge of understanding, at least in a A. Early Theoriesgeneral sense, what underlies the phe- 1. Developmental Delaynomenon. Some might argue that there isno reason for such optimism, and that Most researchers in the first half oftoday is no different from yesterday. the 20th century believed that agingClearly, other researchers thought the was part of the developmental processsame thing about their theories, only to Table 3.2. So, just as there are genes forbe disappointed, so what is so different development, they reasoned there mustthis time? It is the first time almost all be genes for aging. This led to the ideacurrent theories of DR can be integrated that DR worked by slowing development. Table 3.2 Hypotheses to Explain Life Span Extension by Dietary RestrictionHypothesis Evidence For Evidence AgainstDevelopmental delay DR slows development; animals with DR works post-puberty slower development tend to live longerDecreased metabolic DR lowers body temperature, some DR does not lower metabolic rate evidence for lowering of mutation rate or frequency of certain frequency; good evidence for mutations in mitochondria uncoupling of oxidative phosphorylation, thus reducing ROSEndocrinological DR alters numerous endocrine factors; Only partial overlap between dwarf changes mice mutant for IGF-1 or growth and DR mice; IGF-1 mutant hormone live longer mice still respond to DREnhanced cell defenses Cells from DR and long-lived animals DR increases rates of cell death and increased cell tend to be stress resistant in some tissues survivalDecreased Inflammation underlies some diseases Reports in the literature of the inflammation of aging and DR reduces the effect of DR on inflammatory age-associated increase in the responses are inconsistent inflammatory responseHormesis (i.e., CR) CR proven to work via hormesis for Hormesis is unproven in provokes a mild stress budding yeast; long-lived animals mammals response, causing have increased stress resistance; fits enhanced cell defenses with evolutionary theories of aging and metabolic changes, coordinated by the endocrine system
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 71But the hypothesis did not survive long The Rate of Living Hypothesis quicklyas researchers showed that placing adults gave rise to the misconception that DRon a DR diet also extended life span. extends life span by lowering metabolicToday, the idea that aging is part of a rate. Between 1950 and 1981, four groupsprogram is no longer considered valid by showed that reducing the food intakeevolutionary biologists and geneticists of rats slows metabolic rate per unitalike. Rather, aging is believed to arise body mass (reviewed in Krystal & Yu,due to a lack of selection for health late 1994). One recent study in primatesin life and the inability of organisms reported that total daily energy expendi-to maintain a pristine and immortal ture was lower in the calorie-restrictedsoma due to the competing demands of monkeys than in the AL monkeys, evenactivities such as growth and reproduc- when corrected for differences in bodytion (Charlesworth, 2000; Kirkwood & size using body weight (DeLany et al.,Holliday, 1979; Medawar, 1946). 1999). However, numerous other studies, most notably those from Ed Masoro’s group, show that DR animals have equal2. Reduced Metabolic Rate or higher metabolic rates than AL ani-Existing data on the relationship mals, although they do experience anbetween metabolic rate (i.e., energy initial drop in metabolic rate in the firstexpenditure) and life span are “contradic- six weeks (Duffy et al., 1997; Masoro,tory and extremely confused” (Speakman 1998; Masoro et al., 1982; McCarteret al., 2003). In 1908, German physiolo- et al., 1985; McCarter & Palmer, 1992;gist Max Rubner reported that animals of Yu et al., 1985).various sizes utilize a similar number of In simple eukaryotes and mammals,calories per weight per lifetime (Rubner, most studies have failed to find evi-1908), which was the basis for Pearl’s dence for the Rate of Living Hypothesis.Rate of Living Hypothesis (Pearl, 1928). For example, studies by Vanfletteren’sFor about 50 years, the theory that life group in C. elegans have found no evi-span is proportional to metabolic rate dence for decreased metabolic rate dur-enjoyed center stage. Support for this ing CR (Braeckman et al., 2002), and onetheory came from indirect evidence and a study reported an increase (Houthoofdseries of inverse correlations between et al., 2002). Linda Partridge’s groupmetabolic rate and life span, and today it also found no correlation between dietis not considered valid. Although there is and metabolic rate in D. melanogastera rough correlation between metabolic (Braeckman et al., 2002; Hulbert et al.,rate and life span, there are many excep- 2004).tions. For example, a rat and a bat A recent study in rodents also found ahave similar metabolic rates, yet a bat positive association between metaboliclives around five times longer (Wilkinson activity and life span (Speakman et al.,& South, 2002). Many researchers 2003, 2004). With regards metabolicalso have tended to equate the Rate of theories of DR, this finding is consistentLiving Hypothesis with the Free-Radical with a variation on the theme known asHypothesis (see Chapter 10), while the Uncoupling to Survive Hypothesis,assuming that reduced metabolic rate which predicts that increased energymeans fewer free radicals are generated, metabolism (therefore greater uncou-when in fact there is currently no scien- pling/proton leakage) should be associ-tific merit to the linking of these two ated with increased longevity (Speakmantheories (Yu, 1993). et al., 2004).
72 D. A. Sinclair and K. T. Howitz As noted by Masoro (2001), the miscon- energy intake would result in extinctionception that DR works by inducing of the species (Austad, 2001). Austad alsohypometabolism has been perpetuated. calculates that the amount of food con-This is well illustrated by the following sumed by a typical laboratory mouse isexample. Based on a paper by Imai and similar to that in the wild (about 3colleagues in the prestigious journal kJ/g/day); clearly ad libitum-fed mice inNature, members of the scientific and lay the laboratory are not “grossly overfed”press published articles suggesting that (Austad, 2001).the yeast Sir2 longevity regulator is acti-vated by reduced metabolic rate, which 4. Glucocorticoid Cascademight free up NAD, a co-substrate of theenzyme. Clearly, claims that increased One theory of aging is that steroids thatlongevity is caused by reduced metabolic circulate in the blood stream and playrate should be avoided, no matter how a role in the body’s stress response,attractive the hypothesis (Masoro, 2001). the glucocorticoids, are a cause of aging. Supporting the Glucocorticoid Cascade Hypothesis were several studies that3. Laboratory Gluttons linked chronic stress with acceleratedNumerous researchers have speculated aging. Unfortunately for this model, DRthat the effect of DR on extending life animals have higher, not lower, levels of aspan might be a laboratory artifact. The key glucocorticoid, corticosterone, andargument is the following: DR is a more there is no increase in levels of this steroidnatural diet for the animal, similar to in older animals as the hypothesis pre-what it would be in the wild, and that dicted (reviewed in Krystal & Yu, 1994).the AL control animals might simply Although more work is needed in thisbe overfed and more prone to disease area, these results present a serious chal-(Cutler, 1982; Hayflick, 1994). For this lenge to the glucocorticoid hypothesis.reason, many labs have switched to acontrolled diet for the non-restricted 5. Decreased Fatanimals to ensure that they are notoverfeeding. These laboratories report Given the negative effects of obesity, it isthat DR still increases life span to the natural to assume that DR might work bysame extent as an AL diet in these reducing fat stores (Berg & Simms, 1960).experiments (Pugh et al., 1999). Despite early indications that the theory A fact that supports DR being a bona might be correct (Bertrand et al., 1980),fide natural physiological response is over the past 20 years it has been con-that it improves the health and life span tradicted by numerous studies showingof almost every species it has been tested no correlation between body fat andon. It seems unlikely, but possible, that life span (Harrison et al., 1984; Masoro,for all these species—yeast, worms, flies, 1995). A recent study of mice of variousfish, spiders—it is a case of overfeeding genetic backgrounds found no relationshipthe control organisms. In the wild, yeast between life span and body mass, fat massand flies often have an abundance of (Speakman et al., 2003). This study did,food, far more than is supplied in the lab. however, find evidence for higher levelsAnother argument against DR being a of proton leakage in the mitochondria oflaboratory artifact is that animals in longer-lived outbred mouse strains, whichtheir native environment must eat more supports the Uncoupling to Survivethan DR animals for at least part of their Hypothesis, as described below. In thelife to be fertile; a lifetime of reduced case of DR rodents, those that maintain a
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 73higher level of fat are often longer lived results are not always consistent (Chung(Masoro, 1995). Another study showed et al., 2002). Nevertheless, it is wellthat AL rats kept as lean as CR rats by established that DR suppresses inflam-running on an exercise wheel do not show mation both in vitro and in vivo. Forthe same degree of life extension, and example, DR suppresses COX-derivedthere was no increase in maximum ROS generation during aging and bluntslongevity in the lean runners compared the increase in the production of TXA2,with controls (Holloszy, 1997). PGI2, and PGE2, among others (Cao The Decreased Fat Hypothesis has et al., 2001; Kim et al., 2004). DR is alsobecome popular recently after the demon- effective at suppressing inflammationstration that fat cells secrete factors that in vivo, as recently demonstrated in acan promote aging if present in excess rat model of arthritis (Seres et al., 2002).(Barzilai & Gabriely, 2001; Barzilai & Whether DR extends mammalian lifeGupta, 1999; Gupta et al., 2000) and span primarily by suppressing inflamma-the 15 percent increase in life span of tion is debatable, but many researchersmice with a fat-specific knockout of the would agree that it does play some role.insulin receptor (FIRKO) (Bluher et al.,2003). While it is possible that fat cellsmodulate the pace of aging by secreting B. Current Theories on DR based onhumoral factors, and more detailed Proximal Causes of Agingstudies of specific fat depots are required, 1. Reduced Reactive Oxygenthere is currently little evidence that Species (ROS)total fat levels play a major role in lifespan extension by DR. The act of living generates reactive oxygen molecules that can damage cellular con- stituents such as proteins, DNA, and6. Decreased Inflammation lipids. These molecules are also referred toThe Inflammation Hypothesis of Aging, as free-radicals or reactive oxygen speciesas its name implies, states that the (ROS). Each human cell receives 10,000inflammatory process is a major under- ROS hits per day, which equals 7 trillionlying cause of the aging process (Chung insults per second per person. Because ouret al., 2002). Although the theory is con- cellular repair systems are not efficienttroversial, there is no doubt that inflam- enough to cope with the onslaught, ROS-mation is an important component of mediated damage accumulates with timemany age-associated diseases, including and is considered to be a major cause ofneurodegeneration, heart disease, and aging (Harmon, 1956).vascular dysfunctions of diabetes. Aging Consistent with the free-radical theoryis positively correlated with increases in of aging (Harmon, 1956), aged mammalsthe activity of NF-kB, a ubiquitous pro- contain high quantities of oxidized lipidsinflammatory transcription factor (Chung and proteins as well as damaged/mutatedet al., 2002). Cytokine production is DNA, particularly the mitochondrialaltered in such a way to promote inflam- genome (reviewed in Droge, 2003; Dufourmation, such as prostagalandins (PGE2, & Larsson, 2004). Moreover, mice withTXA2, PGH2, and PGG2), MPO, COX-2, an accelerated rate of mutation in mito-iNOS, TNFᮀ, IL-1, IL-6, IFN␥, and TGF␤ chondrial DNA exhibit signs of prema-(Chung et al., 2002; Gen Son et al., ture aging, such as weight loss, reduced2005). Unfortunately, at least for the subcutaneous fat, hair loss, osteoporosis,cytokines, most age-related changes have anemia, and reduced fertility (Trifunoviconly been investigated in vitro, and the et al., 2004). With regards DR, the diet
74 D. A. Sinclair and K. T. Howitzslows the increase in the rate of lipid trols (Koizumi et al., 1987; Mote et al.,peroxidation and the associated loss 1991), although lower levels of mRNA ofof fluidity of biological membranes, the other defense genes such as CuZn-accumulation of oxidatively damaged superoxide dismutase (SOD), glutathioneproteins, specifically “carbonylated” pro- peroxidase, and epoxide hydrolase, andteins, and the increase in oxidative dam- these do not apparently change with age orage to DNA (reviewed in Barja, 2004; diet (Mote et al., 1991). Additional supportMerry, 2002). for the theory comes from a study that Among the many scientists who reported levels of mRNA and enzymatichypothesize that ROS are a major cause of activity for SOD, catalase, and glutathioneaging, there is some debate as to whether peroxidase remain higher in DR rats ver-DR works primarily by decreasing ROS sus AL, and that ROS damage is lowerproduction or increasing ROS defenses (Rao et al., 1990). These findings haveand repair. There are data for both these been corroborated by subsequent animalhypotheses, and they are by no means studies (Armeni et al., 2003; Xia et al.,mutually exclusive. Currently, the evi- 1995). Microarray analyses of tissues fromdence better supports the lowering of ROS DR mice show that mRNA for specificproduction by DR; the antioxidant DNA defense enzymes, such as SOD1 anddefense findings are variable, depending SOD2, are increased relative to ad lib-fedon enzyme(s) and tissues examined. In mice (Sreekumar et al., 2002; Weindruchfavor of the first hypothesis, DR has been et al., 2001). Numerous other studiesshown to decrease the production of two show that DNA-repair protein levelskey ROS, superoxide radicals and hydro- and activities are higher in DR animalsgen peroxide (Sohal et al., 1994b), and to compared to controls (Cabelof et al., 2003;slow the accumulation of ROS-induced Um et al., 2003).damage (Lindsay, 1999; Sohal et al., Although the free-radical theory of DR1994a; Zainal et al., 2000). The mitochon- remains popular, and hundreds of studiesdrial complexes I and III of the electron have been published on the subject,transport chain are considered to be the the experimental evidence so far is notmajor sources of ROS (Barja, 2004; convincing (Lindsay, 1999). Proof thatGredilla et al., 2004; Lopez-Torres et al., antioxidants are beneficial is mainly2002). DR also slows the age-dependent limited to the demonstration that theyincrease in iron content of the kidney, slightly increase average life span inthereby reducing ROS damage to that rodents and flies, but there is littleorgan (Cook & Yu, 1998). A recent study evidence to support an increase in maxi-showed that COX-2-derived ROS produc- mum life span.tion during prostaglandin biosynthesis In Drosophila, there has been con-increases with age in rats, and that this is siderable work done in examining thissuppressed by DR (Chung et al., 1999). hypothesis, but the data are contradictory. There is considerable evidence that ROS Flies transgenically altered to overexpressdefenses are also upregulated by DR, but human SOD are stress resistant and liveas with most aging theories, the scientific up to 40 percent longer (Parkes et al.,literature on this topic is replete with 1998; Reveillaud et al., 1992; Spencerinconsistencies. The results seem to et al., 2003), but the effect is genotype-depend on which enzyme and tissue one and sex-specific (Parkes et al., 1998;looks at (Xia et al., 1995). In support of the Reveillaud et al., 1992; Spencer et al.,ROS repair/defense theory, DR rodents 2003). Molecules that might soak up ROS,seem to have higher activity of catalase such as melatonin, carnosine, epithala-and lower lipid peroxidation than AL con- min (a pineal peptide) and epitalon (a
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 75short peptide of Ala-Glu-Asp-Gly), have been implicated in various diseases,increase average life span of flies up to 16 including neurodegeneration, retinalpercent (Izmaylov & Obukhova, 1999; degeneration, cardiovascular disease, andKhavinson et al., 2000; Yuneva et al., frailty (reviewed in Zhang & Herman,2002). Whether these effects are due 2002).to antioxidant properties has not been The Cell Survival Hypothesis of CRconvincingly demonstrated. Contradict- states that the increased life span ofing these studies is the recent demonstra- mammals is due to an attenuation of celltion that there is no apparent correlation loss over time, particularly cells thatbetween ROS production and fly life span, are easily replaced, such as neurons andand DR does not reduce ROS production stem cells (Cohen et al., 2004b). Consis-in these animals (Barja, 2004; Merry, tent with this idea, cells cultured from2002; Miwa et al., 2004). long-lived genetic mutants, such as the In mammals the situation is no better. p66sch knockout mouse and long-livedA recent study of rat liver and brain dwarfs, are typically less prone to stress-reported that DR does not affect the induced apoptosis (Migliaccio et al.,accumulation of a common age-related 1999; Murakami et al., 2003). Numerousdeletion in mitochondrial DNA (Cassano studies have examined rates of apoptosiset al., 2004). There have also been in cells and tissues from DR animals,numerous reports that DR has little and the results have been varied. Manyeffect on ROS defense mechanisms studies have reported that DR increasesand, if anything, DR attenuates the age- rates of apoptosis (or genes that promotedependent increase (Gong et al., 1997; apoptosis), especially rapidly dividing tis-Guo et al., 2001; Luhtala et al., 1994; sues such as skin, pre-neoplastic cells,Rojas et al., 1993; Stuart et al., 2004). and the immune system (Cao et al., 2001;This discordance could be due to the Mukherjee et al., 2004; Tsuchiya et al.,fact that many of the studies examined 2004; Wachsman, 1996). This apparentdifferent tissues, or perhaps it is because increase in apoptosis is thought to besome measured mRNA levels and others a major mechanism by which DR ratsmeasured protein activity, making direct maintain healthy cells and are relativelycomparisons between studies difficult. resistant to cancer (James et al., 1998;That aside, perhaps the most convincing Zhang & Herman, 2002).evidence against the free-radical theory On the other hand, a number of recentof aging is that mice with an SOD2 defi- studies indicate that DR protects a vari-ciency, which have increased oxidative ety of cell types from apoptosis, includingdamage, do not show signs of premature neurons, liver cells, and immune cellsaging by a variety of measures (Van (Calingasan & Gibson, 2000; Hiona &Remmen et al., 2003). Leeuwenburgh, 2004; Monti & Contesta- bile, 2003; Selman et al., 2003). Recent findings indicate that DR may have a2. Alterations in Apoptosis profound effect on brain function andIn response to damage or stress, cells will vulnerability to injury and diseases byattempt to repair and defend themselves, enhancing neuroprotection and reducingbut if unsuccessful, they often undergo susceptibility to apoptosis. Two studiesprogrammed cell death, or “apoptosis.” have reported that neurons of DR animalsAging is generally associated with express high levels of two key apoptosisincreased rates of stress-induced apop- inhibitors, XIAP and ARC, and are moretosis (Higami & Shimokawa, 2000), resistant to stress-induced apoptosisand the cumulative effects of cell loss (Hiona & Leeuwenburgh, 2004; Shelke
76 D. A. Sinclair and K. T. Howitz& Leeuwenburgh, 2003). DR even pro- generally accepted that the rates oftects hippocampal neurons from apopto- apoptosis increase with age and thatsis due to a presenillin-1 mutation in a DR modulates this process, either upmouse model of Alzheimer’s disease or down, depending on the cell type.(Mattson et al., 2002b). Whether this contributes to longevity With regards the liver, primary hepa- is still hotly debated.tocytes from DR animals are less sus-ceptible to cytotoxins and genotoxins 3. Protein Turnover(Shaddock et al., 1995). At the molecularlevel, expression of p53 and Fas receptor Several protein modifications accumu-in hepatocytes goes up with aging, but late with age, the most widely studied isDR suppressed this age-enhanced increase the carbonyl addition, which results pri-(Ando et al., 2002). The pro-apoptotic marily from oxidative damage (Stadtman,gene gadd153/chop is also repressed 1995). Other modifications include gly-by DR in liver hepatocytes, causing cation, racemization, isomerization, anddecreased apoptosis and increased resist- deamination. Protein turnover is anance to hydrogen peroxide (Ikeyama efficient way for a cell to maintainet al., 2003). DR has also been reported functional proteins, and most modifiedto suppress apoptosis in aging rat livers, proteins are marked for degradation bypossibly by attenuating the activity of cytosolic proteases or the proteosome.DNase gamma endonuclease (Tanaka There is abundant evidence from bio-et al., 2004). In one notable study, DR chemical studies and gene expressionfully suppressed TNFalpha-mediated profiling in C. elegans and mammalshepatic apoptosis (Hatano et al., 2004). that protein turnover rates and overall Lymphocytes in DR mice are much autophagic processes decline with ageless susceptible to oxidative stress- and that this decline is attenuated byinduced apoptosis due to attenuation DR (Del Roso et al., 2003; Lewis et al.,of TNF-alpha and Bcl-2 levels (Avula & 1985; Tavernarakis & Driscoll, 2002).Fernandes, 2002), and transcripts Whether this is a cause or a symptom ofinvolved in suppressing apoptosis and aging is not yet clear.promoting cell survival are increasedby DR (Weindruch et al., 2001). In a 4. Decreased Glucose and Insulin levelsrecent study, levels of the SIRT1 (a pro-tein deacetylase homologous to the Sir2 Diabetes mellitus, or type II diabetes, islongevity gene in yeast and worms) were characterized by a high level of serumshown to be elevated in DR rats, which glucose, insulin, and the types of cellularattenuates cells’ susceptibility to apop- alterations seen in the elderly, includingtosis by sequestering the pro-apoptotic glycation and glycooxidation, and theprotein Bax away from mitochondria accumulation of advanced glycation(Cohen et al., 2004b). SIRT1 also deacety- end-products (AGE) (see Chapter 19).lates the FOXO3 transcription factor, Furthermore, many organs and tissuestipping the scales even further towards of type II diabetic individuals tend tocell protection and survival (Antebi, age faster than normal, particularly the2004; Brunet et al., 2004). Sinclair and cardiovascular system. Thus, one waycolleagues propose that SIRT1 protects DR might improve health is by keepingirreplaceable cells such as neurons and blood glucose and/or insulin levels downstem cells from death during times of (reviewed in Kalant et al., 1988).stress, thus maintaining physiological Many groups have studied glucosefunction with age. In summary, it is metabolism in DR animals. An important
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 77discovery was that DR rats use plasma life span are associated with profoundglucose as fuel at the same rate per unit alterations in hormonal levels, particu-of metabolic mass as the AL rats while larly reductions in IGF-1, namely themaintaining physiologically significantly Ames dwarf, which are growth hor-lower plasma glucose levels (a 15 mg/dl mone–deficient and mutant for prop-1,difference) and markedly lower plasma the Snell dwarf, which are mutant forinsulin levels (Masoro et al., 1992). This the pit-1 gene, and little mice, which areis due, in large part, to increased glucose mutant in the GH receptor Ghrhr geneuptake in skeletal muscle and fat pads (Bartke, 2000). For a detailed review, seethat results from the glucose transporter Chapter 19 by R. Miller and S. Austad inGLUT-4 localizing to the plasma mem- this book.brane (Cartee et al., 1994; Dean et al., Whether these dwarf animals are good1998a; Dean et al., 1998b). The decrease models for DR is debated (Bluher et al.,in insulin is attributable to its lower rate 2003). Although it is true that IGF-1 andof secretion by pancreatic ␤-cells (Dean insulin levels are lower in DR animals,et al., 1998b). Although these effects are Bartke and colleagues reported that theclearly associated with DR, it remains life span of the Ames dwarf, which hasto be seen whether low glucose and already low levels of these factors, can beinsulin levels are an actual cause of the further increased by CR (Bartke et al.,life span increase. In recent years, many 2001), arguing that they work via differentresearchers have turned their attention but possibly overlapping mechanisms.to other endocrine factors, including a This idea is supported by comparing therelated signaling molecule called IGF-1, gene expression profiles of DR versus theas described below. GH knockout mouse. Although there was little overlap in gene expression, the effect of CR in GH knockout mice was much5. IGF-1, Growth Hormone, and Other lower than wildtype mice (Tsuchiya et al., Endocrinological Changes 2004). Similar conclusions have beenIt has been known for decades that DR reached in the C. elegans field, primarilyresults in changes to hormonal levels, due to the work of Vanfletteren andand that some of these changes are asso- colleagues. They have provided goodciated with increased longevity, but evidence that mild DR does not involvethere is now some evidence that these altered IGF-1 signaling but that intensechanges actually contribute to the DR or starvation does, and together theselongevity of DR animals. In simple two “processes” have an additive effect oneukaryotes such as worms and flies, dis- life span (Houthoofd et al., 2003). In sum-ruption of the insulin/IGF-1 signaling mary, it is clear that the endocrine systempathway increases longevity, and we are is important in determining longevity, andnow seeing that it is also true for mam- that certain hormones such as IGF-1 seemmals (Tatar et al., 2003). In mice, three to play an important role in coordinating atransgenic “knocking outs” of endocrine systemic response to CR.genes have been shown to extend lifespan, namely the growth hormone C. The Hormesis Hypothesis and(GH) receptor, the insulin receptor Stress-Responsive Survival Pathways(specifically in fat cells), or a heterozy-gous knockout of the IGF-1 receptor Over the past five years, a novel hypo-(Bluher et al., 2003; Holzenberger, 2004). thesis to explain the effect of DR hasMoreover, spontaneous genetic alter- gained popularity. It is such a major shiftations in mice that lead to extensions in in thinking, and embraces so many of
78 D. A. Sinclair and K. T. Howitzthe current theories on DR under one defense pathways in anticipation ofumbrella, that it deserves its own adverse conditions to come (Howitzsection. A small but rapidly growing et al., 2003; Lamming et al., 2004). Thisnumber of researchers in the DR field idea, termed the Xenohormesis Hypo-are now major proponents of this thesis, is discussed in more detail innew theory known as the Hormesis Section V.Hypothesis of DR (Anderson et al., 2003;Iwasaki et al., 1988; Johnson et al., 1996; 1. Hormesis Is an Active DefenseMasoro, 1998; Mattson et al., 2002a; ResponseTurturro et al., 2000). The theory statesthat the underlying mechanism of DR is In the early 1940s, Southam and Ehrlichthe activation of a defense response that (1943) reported that a bark extract knownevolved to help organisms cope with to inhibit fungal growth actually stimu-adverse conditions. These defenses lated growth when given at very lowextend life span because they counteract concentrations. They coined the termthe proximal causes of aging (Masoro “hormesis” to describe such beneficial& Austad, 1996) (see Figure 3.2). The actions resulting from the response of antheory has been recently expanded by organism to a low-intensity stressor. TheSinclair and Howitz to include the idea word “hormesis” is derived from thethat organisms can pick up on chemical Greek word hormaein, which meanscues from other species under stress or “to excite.” The Hormesis HypothesisDR, either in their food or environment, of DR proposes that the diet imposes aand use these to activate their own low-intensity biological stress on the CR or other biological stress longevity/survival signaling proteins cell-cell communication, IGF-1, insulin, attenuation of altered other humoral factors increased cell stress-induced metabolism, defenses cell death cell cycle reduced ROS, increased secretion of factors, cellular damage, longevity of fat mobilization, mutations critical cells insulin sensitizationFigure 3.2 The Hormesis Hypothesis of DR. The theory states that DR is a mild stress that provokes asurvival response in the organism, which boosts resistance to stress and counteracts the causes of aging.The theory unites previously disparate observations about ROS defenses, apoptosis, metabolic changes,stress resistance, and hormonal changes and is rapidly becoming accepted as the best explanation for theeffects of DR.
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 79organism, which elicits a defense was affected by hundreds, if not thou-response that helps protects it against the sands, of genes. Then, genetic studiescauses of aging (Lithgow, 2001; Masoro, in model organisms in the 1990s began2000; Turturro et al., 1998). It is a major to uncover numerous single gene muta-shift in thinking from earlier hypotheses. tions that extended life span (FriedmanIt suggests that DR is due to an active & Johnson, 1988; Jazwinski et al., 1993;defensive response of the organism as Kenyon et al., 1993). Today there areopposed to passive mechanisms such dozens of mutations known to extendas altered mitochondrial metabolism or life span in model organisms (Lee et al.,lower circulating glucose. The genes that 2003b). It is worth noting that someseem to control this process are now researchers still do not accept that thesecoming to light, thanks to genetic studies genes are particularly relevant to agingin simple organisms such as worms and research (Hayflick, 1999). This begs theflies. The Hormesis Hypothesis makes question: what have these researchersfour key predictions: overlooked? The major oversight appears to be a failure to realize that organisms 1. That DR induces intracellular possess genetic pathways to promotecell-autonomous signaling pathways survival during times of adversity. Long-that respond to biological stress and term activation of these pathwayslow nutrition counteracts the causes of aging and 2. That the pathways help defend hence extends life span (Kenyon, 2001;cells (and hence the organism) against Sinclair, 2002).the causes of aging Around the same time as geneticists 3. That the pathways alter glucose, were formulating these hypotheses tofat, and protein metabolism to enhance explain the existence of longevity genes,survival during times of adversity similar ideas about DR were emerging 4. That the pathways are under the independently. Holliday first proposedcontrol of endocrine signaling pathways that the effect of DR is an evolutionarythat ensure that cells in the organism act adaptation that allows organisms toin a coordinated fashion survive periods of low food availability Clearly, many of the observations (Holliday, 1989). Masoro and Austad thenlisted under different headings in the pre- expanded this idea and provided perhapsvious section are also consistent with the the most comprehensive theory on thisHormesis Hypothesis. Obvious examples subject by merging it with the ideasinclude the DR-associated boost in cell of Kirkwood and the Hormesis conceptsurvival and the observed metabolic and (Masoro & Austad, 1996). The disposableendocrine changes. Rather than revisit soma theory of Kirkwood proposes thatthese observations here, additional evi- each organism has limited resourcesdence for and against the Hormesis and that these resources can only beHypothesis will be presented. allocated to a finite number of cellular activities, the two primary ones being reproduction and somatic maintenance2. Genes that Control Survival and (Kirkwood & Holliday, 1979). During Life Span times of perceived adversity such asPrior to 1990, very few researchers during DR, organisms divert more ofsuspected the existence of single genes their resources to maintaining theirthat control aging. This was based in soma until conditions improve. Todaypart on the valid assumption that aging there is general consensus among leaderswas an incredibly complex process that in the DR field that the health benefits
80 D. A. Sinclair and K. T. Howitzof DR derive from an organism’s defense resistance correlate (Fabrizio et al., 2001;response to a perceived threat to its Kennedy et al., 1995). Moreover, a varietysurvival. of low-intensity stresses extend yeast life span, including mild heat, increased salt, low amino acids, or low glucose, the yeast3. Proof of Hormesis in Budding Yeast equivalent of CR (Anderson et al., 2003;The use of yeast to study longevity Bitterman et al., 2003). Remarkably, thesemechanisms is now well respected, but life span extensions are facilitated by awhen this organism was first proposed single gene, PNC1, which is induced byas a model for aging in the 1950s and every treatment known to extend yeast1960s (Barton, 1950; Johnson, 1966), it life span (Anderson et al., 2003; Gallowas met with considerable skepticism. et al., 2004) (see Figure 3.3). By addingOne reason, no doubt, is that it was dif- more copies of PNC1, researchers wereficult to see how a relatively simple able to mimic the effects of DR andunicellular organism could provide infor- extend life span 60 percent. PNC1mation about human aging, which isone of the most complex of biologicalphenomena, involving trillions of cells Glucosein numerous systems and organs. It is restriction (CR)becoming clear, however, that all eukary- Amino acidotes possess surprisingly simple and restriction Osmotic stressconserved longevity pathways that gov- Nitrogen Heat shockern life span, principally by attenuating restrictionthe proximal causes of aging in responseto adverse conditions (Kenyon, 2001;Sinclair, 2002). PNC1 Yeast “replicative” life span is defined NAM depletionas the number of divisions an indivi- Sir2dual yeast cell undergoes before dying(Bitterman et al., 2003). One attractivefeature of S. cerevisiae, as opposed tomany other simple eukaryotes, is that Longevitythe progenitor cell is easily distinguishedfrom its descendants because cell divi- Figure 3.3 S. cerevisiae life span extension by DRsion is asymmetric: a newly formed is due to hormesis. Replicative life span in yeast is extended by caloric restriction (CR) and a variety of“daughter” cell is almost always smaller mild stresses. The Sir2 enzyme is a nicotinamidethan the “mother” cell that gave rise (NAM)-sensitive enzyme that extends yeast lifeto it. Yeast mother cells divide about span by deacetylating histones and stabilizing20 times before dying and undergo repetitive DNA. PNC1 encodes an enzyme thatcharacteristic structural and metabolic depletes nicotinamide and activates Sir2. PNC1 can be viewed as a “master regulator of aging” thatchanges as they age. An alternative serves as a sensor that translates CR and environ-measure of yeast aging, “chronological mental stress signals into Sir2 activation andlife span,” is the length of time a popu- longevity (Anderson et al., 2003). By having central-lation of yeast cells remains viable in ized control of longevity, the system permitsa nondividing state following nutrient new life spans to evolve rapidly in response to a changing environment. The role of NAM and adeprivation (Longo & Fabrizio, 2002). mammalian equivalent to PNC1, known as Consistent with the Hormesis Hypo- PBEF/visfatin/NAMPT, in regulating life span andthesis and findings in other species, health in mammals is under investigation bylongevity of yeast cells and stress numerous laboratories.
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 81encodes a nicotinamidase that extends mild stress, flies and worms subjectedyeast life span by depleting the cell to DR are also resistant to a variety ofof nicotinamide, an inhibitor of the stresses, including heat shock (TatarSir2 longevity enzyme. This system of et al., 2003). These observations stronglylongevity regulation in yeast explains how argue that DR does not simply changemultiple, disparate stimuli can lead to the metabolism or ROS output, as was pre-same longevity response and how species viously thought, but rather it induces amay rapidly evolve strategies to suit a defense program that provides resistancechanging environment. Whether or not to a wide array of stresses.nicotinamide catabolic pathways control Many of the life span extensions pro-Sir2 enzymatic activity in higher organ- vided by mild stress and crowding haveisms is not known. been shown to act through the insulin/ Interestingly, the SIR2 gene is con- IGF-1 pathway, which boosts the activityserved in higher eukaryotes, and its of a conserved forkhead transcription fac-ability to extend life span is conserved tor known as daf-16/FOXO. One of theat least up to worms and flies (Helfand, functions of daf-16/FOXO is to increases2004; Tissenbaum & Guarente, 2001). the transcription of cell defense genesIt is worth noting that there must be such as SOD2 and HSP70 (Lee et al.,alternative pathways for the mediation of 2003a; Murphy et al., 2003). Overexpres-life span extension because DR can still sion of FOXO in the fat body of the flyextend the life span of certain strains of extends life span via dilp-2, an insulinyeast and worms that lack the Sir2 gene homolog produced in neurons, indicating(Kaeberlein et al., 2004; Sinclair & Wood, that FOXO controls endocrine signaling2004). A detailed discussion of Sir2 may and that fat is an important tissue forbe found in the section on CR mimetics the control of life span (Giannakoubelow. et al., 2004; Hwangbo et al., 2004) (see Chapter 10). That said, clearly the insulin/IGF-1 pathway is not the whole4. Evidence for Hormesis in Worms story because worm mutants lacking and Flies Daf-16 still respond to DR (BraeckmanThere is a strong correlation between et al., 2001a).longevity of various strains of worms andflies and their resistance to various types 5. Evidence for Hormesis in Mammalsof stress, including desiccation, heatstress, acetone, ethanol, and paraquat It has been known for a decade that DR(Arking et al., 1991; da Cunha et al., boosts serum levels of glucocorticoids,1995; Harshman et al., 1999; Houthoofd particularly corticosterone, which is aet al., 2003; Mockett et al., 2001; Wang good indicator that the animals areet al., 2004). Moreover, exposure of invoking a stress response (Masoro, 2000;organisms to agents or conditions that Sabatino et al., 1991). As is the casecause mild biological stress increases for simple metazoans, DR also increaseslife span as well as resistance to other, the resistance of animals to a varietyseemingly unrelated stresses, such as of stresses, including sudden increaseslow doses of paraquat, aldehydes, irradia- in temperature (Heydari et al., 1993)tion, heat shock, crowding, and hyper- and toxins (Duffy et al., 2001; Masoro,gravity (Braeckman et al., 2001b; Hercus 1998). These observations also extend toet al., 2003; Lints et al., 1993; Minois the cellular level.et al., 1999; Sorensen & Loeschcke, Experiments at the cellular and molecu-2001). Arguing that DR is simply another lar level strongly support the Hormesis
82 D. A. Sinclair and K. T. HowitzHypothesis. The skin cells from long- glucose metabolism or ROS produc-lived mutant Snell dwarfs are relatively tion, but by triggering an evolutionarilyresistance to stress and toxins including ancient, active defense response thatUV, light, heat, cadmium, and paraquat allows organisms to survive adversity.(Murakami et al., 2003), and cells cultured The contrast between earlier hypothesesin the presence of serum from DR rats are and this one is stark, but only time willrelatively resistant to stresses and pro- tell if the theory will hold up over theapoptotic signals (Cohen et al., 2004b; de coming decades.Cabo et al., 2003). Subjecting fibroblaststo repeated heat shock also produces simi-lar effects to those seen in vivo, including V. Small-Molecule CR Mimeticsmaintenance of the stress protein profile, A. Drug Development Strategies forreduction in the accumulation of dam- Mimicking CRaged proteins, stimulation of proteolysis,and increased resistance to stressors such Work on the genetics of aging in yeast,as ethanol, hydrogen peroxide, and UV C. elegans, Drosophila, and rodents(Rattan, 2004). suggests that CR might act through It has recently been demonstrated that conserved signaling pathways to controlthe mammalian homolog of the Sir2 gene eukaryotic longevity in response to envi-that promotes life span in lower organ- ronmental conditions (Kenyon, 2001).isms, known as SIRT1, is highly induced Important elements of a major longevityin the tissues of DR rats (Cohen et al., pathway include insulin/IGF-1 receptor2004b). Interestingly, at the cellular level, signaling and the FOXO family of tran-SIRT1 promotes the resistance of cells scription factors, whose activity is regu-to stress-induced death by attenuating lated in part by lysine-acetylation and thep53 and stimulating the Ku70-Bax anti- action of “sirtuin” NADϩ-dependentapoptotic system (Cohen et al., 2004a; deacetylases (see Figure 3.4). The exis-Cohen et al., 2004b). SIRT1 also sti- tence of this pathway raises the possi-mulates metabolic changes in cells bility that small molecule modulators ofconsistent with DR, including decreases its constituent proteins could potentiallyin fatty acid synthesis in adipocytes mimic the effects of CR, thereby pro-by deacetylating the nuclear hormone viding some of its benefits, without thereceptor PPAR␥ (Picard et al., 2004) and need for actual CR. As an example of thisincreases in glucose production from type of CR-mimesis, we will discuss inhepatocytes via PGC-1␣ and PPAR␣ detail compounds that activate sirtuins.(Puigserver & Speigelman, 2004). Given Other categories of small-moleculethat SIRT1 is up-regulated by DR and CR mimetics include (1) molecules thatthat the in vitro data are consistent may mimic CR directly by effects onwith changes seen in vivo, it will be energy metabolism and (2) moleculesinteresting to examine whether the over- identified by their ability to induce gene-expression of SIRT1 produces similar expression patterns similar to thoseeffects to DR in transgenic mice. induced by CR. Potential CR mimetics In summary, although the Hormesis that exemplify these approaches are,Hypothesis of DR is young and not respectively, 2-deoxyglucose and met-embraced by the majority of researchers formin. These approaches to CRtoday, it is the best theory we have to mimetics have been reviewed recentlyexplain the multitude of data about DR (Ingram et al., 2004) and, because wein mammals and lower organisms. If the will not be discussing them further,theory is right, DR extends life span not we refer the reader to that review andsimply by passive means such as altering some of the recent primary literature on
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 83 CR enzymes which occurs in eukaryotes, the archaea and eubacteria (Laurenson & Insulin / IGF-1 Rine, 1992; Smith et al., 2000). Originally described as a factor required Insulin/IGF-1 for maintenance of silencing at telomeres Receptors and mating-type loci, Sir2 was subse- ? quently shown to be an enhancer of PI3-Kinase mother-cell replicative life span in bud- ding yeast (Kaeberlein et al., 1999). The PDK-1 Sir2 enhancement of mother-cell life span has been linked to its stabilization AKT-1/2 ? of repetitive DNA, in particular its SIRT1 suppression of the accumulation of extra- Ku70 chromosomal rDNA circles (Kaeberlein FOXOs 1, 3, 4 et al., 1999; Kobayashi et al., 2004; p53 Bax Sinclair & Guarente, 1997).Antioxidant The sirtuins represent a distinct class Defenses DNA Repair of trichostatin A (TSA)-insensitive Apoptosis protein-lysyl-deacetylases (Class IIIFigure 3.4 Regulatory interactions between SIRT1, HDACs) and have been shown to cat-the mammalian insulin/IGF-1 signaling pathway, alyze a reaction that couples lysineand factors controlling apoptosis. SIRT1 deacety-lates the FOXO transcription factors, stimulating deacetylation to the formation oftranscription of genes involved in antioxidant nicotinamide and O-acetyl-ADP-ribosedefense and DNA repair, while repressing transcrip- from NADϩ and the abstracted acetyltion of pro-apoptotic genes. SIRT1 exerts additional group (Imai et al., 2000; Tanner et al.,anti-apoptotic effects by inactivation/destabiliza- 2000; Tanny & Moazed, 2001). HDACstion of p53 and by promoting the sequestration ofBax by Ku70. Solid lines indicate that at least one are named for their role in deacetyla-mechanism for the stimulatory or inhibitory effect tion of histone N-terminal tails, anhas been experimentally established (e.g., FOXO3 action which typically leads to the for-repression of the expression of BIM, a proapoptotic mation of condensed chromatin andprotein) (Brunet et al., 2004). Dashed lines indicate transcriptional silencing (Strahl & Allis,observed effects for which the mechanism is stillunknown. For example, insulin and IGF-1 have been 2000). However, it is becoming increas-shown in cell culture to reverse the elevation of ingly clear that the effects of HDACs,SIRT1 expression elicited by serum from CR rats including those of the sirtuins, are(Cohen et al., 2004b). also implemented through deacetyla- tion of transcription factors and other proteins.2-deoxyglucose (Wan et al., 2003; Wan There are at least seven human sirtuins,et al., 2004), CR expression profiling SIRT1–SIRT7 (Frye, 2000). SIRT1, which(Dhahbi et al., 2004; Weindruch et al., is located in the nucleus, is the human2001), and metformin (Fulgencio et al., sirtuin with the greatest homology to Sir22001; Spindler et al., 2003). and has been shown to exert a regulatory effect on multiple transcription factors,B. Sirtuin Activating Compounds including p53 (Langley, 2002; Luo et al., (STACs) 2001; Vaziri et al., 2001), MyoD (Fulco et al., 2003), FOXO1-3 (Motta et al., 2004;1. Sirtuins: Conserved Longevity Factors Daitoku et al., 2004), FOXO3 (BrunetYeast Sir2 (Silent information regulator et al., 2004; Daitoku et al., 2004; Motta2) is the founding exemplar of the sirtu- et al., 2004; van der Horst et al., 2004),ins, an apparently ancient group of PPAR␥ (Picard et al., 2004), and Ku70
84 D. A. Sinclair and K. T. Howitz(Cohen et al., 2004b). The FOXO effects encodes a NAM phosphoribosyl trans-are perhaps of particular note because ferase that regulates nicotinamide levelsthey parallel the way in which C. elegans (Fukuhara et al., 2004; Rongvaux et al.,DAF16, a FOXO homolog, is required 2002). Initial reports indicate that PBEFfor life span extension by increased Sir2 retains the ability to regulate SIRT1expression (Tissenbaum & Guarente, (Revollo et al., 2004).2001). Although there is some disagree- Another leading hypothesis on thement among the FOXO studies—compare regulation of Sir2 connects deacetylaseFOXO3 effects on p27kip1 expression in activity to energy metabolism by assert-the study of Motta et al. (2004) and that ing that Sir2 is regulated by the supplyof Brunet et al. (2004)—a pattern emerges of its substrate NADϩ or by the ratioin which SIRT1 upregulates FOXO- of NADϩ to its reduced form, NADHinduced transcripts promoting DNA repair (Lin et al., 2004). In yeast and C. elegans,(GADD45, Brunet et al., 2004), oxidative added copies of sirtuin genes extend lifestress response (MnSOD, Daitoku et al., span, and Sir2 activity is boosted by2004; van der Horst et al., 2004), and caloric restriction (Anderson et al., 2003;cell-cycle arrest (p27kip1, Brunet et al., Lin et al., 2000).2004; Daitoku et al., 2004; van der Recent work on CR rats has shown thatHorst et al., 2004), while repressing SIRT1 protein levels are elevated in multi-pro-apoptotic FOXO-induced transcripts ple tissues, relative to ad libitum-fed con-(BIM, Motta et al., 2004; Brunet et al., trols (Cohen et al., 2004b). Moreover,2004) (Fas ligand, Brunet et al., 2004). serum from CR rats elevated SIRT1 It has been suggested that SIRT1 expression in cultured human cells, anpromotes cell survival and organismal effect that could be eliminated by thelongevity by delaying apoptosis enough addition IGF-1 and insulin. Of course, thisto give antioxidant defenses and repair does not negate the possibility that NADϩprocesses time to succeed (Brunet et al., and/or nicotinamide levels may also con-2004; Howitz et al., 2003). A variant of tribute to SIRT1 regulation in mammals.this idea, suggested by SIRT1 effects on Taken together, the results summarizedNF-B, is that SIRT1 may delay stress- above suggest that SIRT1 activity plays ainduced p53-dependent apoptosis but central role in the CR-responsive pathwaymay actually make cells more sensitive that affects mammalian longevity. SIRT1to apoptosis induced by death receptor expression responds to the knownligands such as TNF-␣ and TRAIL (Yeung longevity-enhancing stimuli CR andet al., 2004). diminished insulin/IGF-1 signaling Sirtuins are inhibited by nicotinamide (Cohen et al., 2004b). On the downstream(NAM), a product of the deacetylation side, SIRT1 effects on transcription factorsreaction (Bitterman et al., 2002). In yeast, and other proteins controlling apoptosisthis forms a basis for the regulation of and cell survival, DNA repair, oxidative-Sir2 activity. Expression of the yeast stress responses, and lipid metabolismnicotinamidase, PNC1, is upregulated by suggest that increased SIRT1 activity mayseveral longevity-enhancing mild stresses, be responsible for eliciting the geneincluding calorie restriction suggesting expression changes brought on by CR.that, at least in yeast, Sir2 stimulation SIRT1 effects on mammalian longevityvia nicotinamide removal may be the have yet to be directly established.common feature that links several types Nevertheless, the demonstration of con-of hormesis (Anderson et al., 2003). The nections between SIRT1 and insulin/IGF-1human functional equivalent of yeast signaling on the one hand and the FOXOPNC1 is PBEF/visfatin/NAMPT, which transcription factors on the other means
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 85that the key elements of the metazoan nol, a stilbene, and quercetin, a flavone,longevity pathway have been confirmed in as stimulators of SIRT1 deacetylasemammals. activity (Howitz et al., 2003). Stilbenes If SIRT1 is indeed a conduit of and flavones are polyphenols, as are aCR-derived signals, then interventions variety of related plant secondarythat either elevate SIRT1 expression or metabolites such as chalcones, antho-stimulate SIRT1 activity might mimic cyanidins, and catechins. Screening ofCR. Strategies for stimulation of SIRT1 additional plant polyphenols identifiedcatalytic activity include elevation of the 15 more SIRT1 activators, the mostconcentration of its substrate, NADϩ, potent of which was the stilbene resvera-and removal of its reaction product and trol. Structural features common toinhibitor, nicotinamide. Here we will these sirtuin-activating compoundsconcentrate on the evidence that a num- (STACs) include two aromatic rings, atber of plant polyphenols can mimic CR least potentially coplanar and trans tovia a third mechanism, namely allosteric one another, and hydroxyl functions inactivation of sirtuins. Plant polyphenols one or both of the meta positions of oneare abundant in various foodstuffs ring and the para position of the otherand are themselves produced by plants (see Figure 3.5).responding to a variety of environmental Kinetic analysis of the resveratrolstresses. We will end by discussing the effect on SIRT1 revealed that it loweredidea that polyphenol sirtuin stimulation the Kms for both NADϩ and the acety-in plant consumers may represent the lated peptide substrate, while havingspecies-to-species transfer of a beneficial little or no effect on Vmax (Howitz et al.,stress signal, a hypothesis for which we 2003). This result suggests that resver-have coined the term xenohormesis. atrol is a K-type allosteric effector of SIRT1 (Monod et al., 1965). Qualitatively similar effects on the Kms of the yeast2. Polyphenolic Sirtuin Activators and sirtuin Hst2 can be achieved by deletion Their CR-Mimetic Effects of sequences lying N-terminal (NADϩA screen of various small-molecule and peptide Kms) and C-terminal (NADϩlibraries for modulators of SIRT1 activ- Km only) to the core catalytic domainity identified two compounds, piceatan- (Zhao et al., 2004). On the basis of OH OH OH OH Resveratrol Butein OH Quercetin HO HO O HO OH OH OH O OH OFigure 3.5 Structures of three structurally related polyphenolic sirtuin-activating compounds (STACs):resveratrol (a stilbene in grapes and some Asian herbs), butein (a chalcone in flowers), and quercetin (aflavone in apples and onions). Other STACs include analogs of nicotinamide (NAM) such as isonicoti-namide, which activate sirtuins by preventing NAM inhibition. Metabolites of resveratrol, such assulfated-resveratrol, which can circulate in humans for more than a day, have recently been found to acti-vate sirtuins with the same potency as the native compound. Whether a typical diet can provide sufficientSTACs to activate sirtuins in vivo is not yet known, but exogenous addition of resveratrol can extend lifespan in yeast, worms, and flies—unless they lack the Sir2 gene (Howitz et al., 2003; Wood et al, 2004).
86 D. A. Sinclair and K. T. Howitzthese Hst2 structural and kinetic studies, SIRT1 can reverse the stress-inducedZhao and colleagues proposed that a acetylation of the DNA repair factorpolyphenol-binding-induced reconfigura- Ku70 at lysines 539 and 542, therebytion of the conserved ␤1-␣2 loop and/ enhancing Ku70 binding and nuclearor zinc-binding domain could act to sequestration of the pro-apoptotic proteinenhance sirtuin substrate binding. Bax (Cohen et al., 2004b). Resveratrol Thus far, STACs have been shown to diminished Bax-induced apoptosis inextend life span in Saccharomyces cere- HEK 293 cells cotransfected with Bax andvisiae, Drosophila melanogaster, and Ku70, an effect parallel to that obtainedCaenorhabditis elegans (Howitz et al., by transfection with SIRT1 (Cohen et al.,2003; Wood et al., 2004). In each of 2004b). Several STACs, including resver-these model systems, the STACs- atrol, were shown to stimulate sirtuinsinduced life span extension required the in HeLa cells (Howitz et al., 2003) and,presence of a functional Sir2/SIRT1 with the same assay system, resveratrolortholog, and in each case recombinant increased the overall deacetylation rate inpreparations of these enzymes were acti- a non-small cell lung cancer (NSCLC)vated by STACs in vitro (Howitz et al., line with a high SIRT1 expression level2003; Wood et al., 2004). In the two (Yeung et al., 2004). Also in NSCLC cells,organisms for which this was tested, SIRT1 downregulates NF-B-mediatedSaccharomyces and Drosophila, STACs transcription by deacetylation at lysinedid not increase the life span extension 310 in the RelA/p65 subunit transac-provided by CR (Howitz et al., 2003; tivation domain (Yeung et al., 2004).Wood et al., 2004). Clearly this is con- Resveratrol was shown to decrease thesistent with the idea that STACs act to acetylation level of p65 in HEK 293 cells,extend life span by the same pathway as and the ability of both SIRT1 and resvera-CR, with direct sirtuin stimulation pro- trol to repress p65 transactivation inviding the most straightforward mecha- NSCLC cells depended on the presencenistic explanation. of lysine 310 (Yeung et al., 2004). Resveratrol effects consistent with Recently, resveratrol has been shownstimulation of SIRT1 have been observed to enhance SIRT1 action in cellularin several mammalian cell culture models of processes of particular rele-systems in which one or more of the vance to CR effects in mammalian aging:relevant SIRT1-targeted acetyl-lysine neurodegeneration, fat mobilization, andresidues have been identified. Acetylation adipogenesis. Using in vitro models ofof lysine 382 increases the activity and axonal degeneration, Araki and colleaguesstability of p53, and acetyl-lysine-382 is demonstrated that the mouse wlds muta-a known target of SIRT1 (Langley, 2002; tion delays degeneration by virtue of itsLuo et al., 2001; Vaziri et al., 2001). Low overexpression of a fusion protein incor-concentrations of resveratrol (0.5 M) porating the NADϩ-biosynthetic enzymewere shown to diminish the level of Nmnat1 (Araki et al., 2004). This delayUV-induced p53 lysine-382 acetylation in in degeneration could be mimicked by aU2OS osteosarcoma cells (Howitz et al., 24-hour pretreatment of wild-type neu-2003). Parallel experiments with HEK rons with NADϩ, an effect that required293 cells showed that this resveratrol- SIRT1 expression (Araki et al., 2004).stimulated deacetylation occurred in the Resveratrol pretreatment afforded protec-presence of transfected wild-type SIRT1, tion similar to that obtained with NADϩ,but not in the presence of a dominant- suggesting that both effects were due tonegative, catalytically inactive construct stimulation of SIRT1 activity (Araki et al.,(SIRT1 H363Y) (Howitz et al., 2003). 2004). Since axonal degeneration is a
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 87feature of aging-related neurodegenerative nols that activate sirtuins in yeastdiseases (Raff et al., 2002), these results and animals are produced by plants inecho the preventive action of CR on these response to various types of environmen-pathologies. tal stress, including drought, nutrient Sirt1ϩ/Ϫ mice display significantly deprivation, and UV-irradiation (Dixon &decreased fasting induced mobilization of Paiva, 1995). It therefore seems reason-free fatty acids, an effect that the experi- able to suggest that some plant polyphe-ments of Picard and colleagues (2004) nols might function as endogenoushave connected to a deficit in SIRT1 regulators of plant sirtuins and otherrepression of the activity of the transcrip- stress responses (Howitz et al., 2003).tion factor PPAR␥ in white adipocytes. Although plants produce an enormousResveratrol promoted the mobilization of variety of polyphenols, many of whichfree fatty acids from primary cultures serve in non-signaling roles (e.g., UV-of rat adipocytes, and the resveratrol- filters, antioxidants, pigments, antibi-induced fat depletion from 3T3-L1 otics), it has been argued on evolutionaryadipocytes depended on the expression of grounds that the first polyphenols func-SIRT1 (Picard et al., 2004). Noting that tioned as signaling molecules (Stafford,fat is depleted from white adipose tissue 1991). Lunularic acid (LA), a stilbenoid,as a consequence of CR, Picard and has been proposed to function in “lower”colleagues suggested that this may be plants (e.g., liverworts) as a stress-a SIRT1-mediated effect (Picard et al., response hormone, a role fulfilled by2004). Mice that are deficient in PPAR␥ abscisic acid (ABA) in vascular plants(PPAR␥ϩ/Ϫ heterozygotes) have increased (Yoshikawa et al., 2002) (see Figure 3.6).insulin sensitivity and experience less A recent study provided evidence ofaging-related decline in insulin sensi- structural overlap between the stabletivity than PPAR␥ϩ/ϩ mice (Miles et al., conformers of ABA and LA and of ABA-2003). Thus, SIRT1 suppression of PPAR␥ like activity by LA in higher plantsactivity could represent a mechanism (Yoshikawa et al., 2002). Noting thatthat retards the onset of aging-related synthesis of ABA, unlike that of LA,diseases such as type II diabetes. requires molecular oxygen, Yoshikawa and colleagues propose that LA is the more ancient of the two molecules. InC. Is Sirtuin Stimulation by Polyphenols higher plants, there are a few known a Case of Xenohormesis? examples of polyphenols functioning in regulatory roles, for example the flavone-1. The Roles of Polyphenols in Plant induced pollen germination in petunia Stress Responses and Stress Signaling (Napoli et al., 1990; Vogt & Taylor,All eukaryotes, including plants, encode 1995), but in general this is a poorlysirtuins in their genomes (Frye, 2000; understood area. The possible existencePandey et al., 2002). The plant polyphe- of endogenous polyphenol regulators of Lunularic Acid Abscisic Acid OH OH C HO O OH O O OHFigure 3.6 Structures of the plant stress hormones, lunularic acid and abscisic acid.
88 D. A. Sinclair and K. T. Howitzplant sirtuins may provide a fruitful CR, but not sudden starvation, increasesline of inquiry. intestinal nutrient transport capacity (Casirola et al., 1996; Ferraris et al., 2001) illustrates how the anticipatory induction2. The Xenohormesis Hypothesis of adaptive pathways, by CR-mimetics,The question of plant sirtuin regulation could be critical to survival.aside, why would plant stress molecules, Although we have discussed the ideathe polyphenol STACs, activate sirtuins of xenohormesis in the context of sirtuinfrom mammals, yeast, nematodes, and stimulation, there is no necessity thatinsects? One possibility is that the individual STACs or other plant stress-STACs might mimic yet-to-be-discov- induced compounds should act only onered endogenous small-molecule regula- this target. Rather, the notion is that thetors in these species. These hypothetical overall effect of multiple stress-inducedmolecules could not be polyphenols, compounds in the diet, perhaps acting onhowever, because animals lack the multiple targets and signaling pathways,enzymes to synthesize them. There are, should strengthen the plant-consuminghowever, well-known examples of mim- organism’s own stress resistance (Liu,icry of chemically unrelated mammalian 2003). Indeed, resveratrol effects onregulatory molecules by plant secondary numerous targets have been reportedmetabolites, including the opioids, the (Pervaiz, 2003), including, to name butcannabinoids, and the weak estrogenic a few, inhibition of cyclooxygenasesactivity of some of the STACs them- (Jang et al., 1997), inhibition of ribonu-selves. The notion of mimicry carries cleotide reductase (Fontecave et al.,with it the implication of survival 1998), and direct chemical scavenging ofadvantages for the plant. These could reactive oxygen species (Frankel et al.,range from deterrence of destructive 1993). Resveratrol’s multiple biologicalplant consumers to attraction of plant effects—cancer chemoprevention, cardio-consumers essential for pollination or protection, and neuroprotection—haveseed dispersal. However, the fact that been connected, with varying degrees ofplants synthesize and accumulate success, to this plethora of molecularpolyphenol STACs in response to both targets (Gescher & Steward, 2003). Duebiotic and abiotic stress suggests to us to its capacity to regulate multiple tran-the possible advantage to plant consum- scription factors, we do feel that polyphe-ing species of developing or maintaining nol stimulation of SIRT1 shows greata capacity to respond to them within the promise as a unifying mechanism behindcontext of their own stress-response many of these effects. Xenohormesis,pathways. however, as an evolutionary concept, We propose the term xenohormesis for doesn’t exclude other possibilities.this concept in order to emphasize that The ultimate test of the Xenohormesisthe initial stress and the beneficial Hypothesis is whether it can be shown toresponse to that stress occur in separate actually operate in the realm of nutritionalorganisms. By serving as early indicators health effects. Epidemiological studiesof a deterioration of the food supply linking consumption of polyphenol-richand/or the environment in general (e.g., foods to the prevention of aging-relateddrought), STACs, acting as CR-mimetics, diseases (i.e., CR-like health benefits) aremay stimulate plant-consuming organ- consistent with the predictions of STACs-isms to direct resources to cellular and based xenohormesis (Block et al., 1992;organismal defense and maintenance. Hertog et al., 1993; Keli et al., 1996; KnektThe fact, for example, that chronic et al., 2002; Sato et al., 2002). However,
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 89some studies have raised doubts about there is a growing number of scientistswhether resveratrol or flavone STACs who believe we are close to having a grandsuch as quercetin are sufficiently bioavail- unifying theory of DR whose basis lies inable for the effects observed in cell culture the idea that DR works by provoking anto be relevant in vivo (Goldberg et al., evolutionarily ancient stress response that2003; Soleas et al., 1997). Plasma concen- keeps organisms alive during adversity.trations of individual dietary polyphenols There can be little room for argument thatrarely rise above the level of 1 to 2 M, DR meets the narrow, phenomenologicalwhereas most cell culture effects are definition of hormesis (Masoro, 1998).observed in the range of 10 to 100 M After all, hormesis, defined simply as an(Gescher & Steward, 2003; Goldberg et al., inverted U-shaped dose-response curve,2003). There are, however, exceptions to clearly applies to a graph of mortality ratethis rule (Chang et al., 2000; Basly et al., versus dietary restriction, the least being2000), and included among these are AL and the other extreme being starvationthe sub-micromolar resveratrol effects (Turturro et al., 2000).on SIRT1-dependent p53 deacetylation The broader and more controversial(Howitz et al., 2003) and Bax-induced proposition is the one that attributesapoptosis (Cohen et al., 2004a; Howitz these hormetic effects to the inductionet al., 2003). It also should be noted of a single, evolutionarily conservedthat because SIRT1 is stimulated by program, one that activates and coordi-multiple stilbenes, flavones, and chalcones nates the downstream repair and defense(Howitz et al., 2003), the sum of the mechanisms that counteract the proxi-plasma concentrations of these effectors mate causes of aging. While the cross-may be more relevant than those of indi- resistances provided by various hormeticvidual compounds. One highly valid criti- stresses are consistent with the existencecism of mammalian cell culture studies of of such a centralized regulatory mecha-polyphenol effects is that the vast majority nism, they don’t prove it. However,of them have employed the free aglycone the Pnc1/Sir2 system, in the case offorms, rather than the sulfate or glu- budding yeast, and the sirtuins togethercuronate conjugated metabolites that pre- with the insulin/IGF-1/FOXO pathway, indominate in the serum in vivo (Corder the case of metazoans, appear to haveet al., 2003). This concern has recently the characteristics one might expectbeen somewhat alleviated with respect to for essential pieces of this regulatoryresveratrol activation of SIRT1 by the find- machinery. If this is more than justing that the 3-sulfate and 4Ј-sulfate forms appearance, biochemical and geneticstimulate SIRT1 in vitro to a similar mapping of the connections betweenextent as free resveratrol (Calamini et al., these pathways, upstream hormetic stim-2004). Use of the conjugated forms of uli and downstream gene expression, andresveratrol and other STACs should be anti-aging effects may be key to validatinggiven the highest priority in future work the broader Hormesis Hypothesis of DR.on SIRT1 activation and other polyphenol As noted at the beginning of theeffects. chapter, the Hormesis Hypothesis of DR attempts to unite various theories on the proximate causes of aging. In a similar VI. Conclusions way, the Xenohormesis Hypothesis suggests that many of the health benefitsThere have been over a dozen theories to of dietary phytochemicals, particularlyexplain the life span–extending effects of those of secondary metabolites producedDR in mammals and other organisms, but by plants under stress, may work because
90 D. A. Sinclair and K. T. Howitzthey activate an evolutionarily ancient Impact of aging and life-long caloriemechanism that allows animals and restriction on expression of apoptosis-relatedfungi to pick up on chemical stress genes in male F344 rat liver. Microscopysignals from plants. As exemplified by Research and Technique, 59(4), 293–300. Antebi, A. (2004). Tipping the balance towardthe polyphenol STACs, this mechanism longevity. Developmental Cell, 6(3), 315–316.is suggested to entail direct stimulation Araki, T., Sasaki, Y., & Milbrandt, J. (2004).of the hormesis signal transduction Increased nuclear NAD biosynthesismachinery, which includes SIRT1 and and SIRT1 activation prevent axonalother proteins that promote health and degeneration. Science, 305(5686), 1010–1013.extend life span. It is important to distin- Arking, R., Buck, S., Berrios, A., Dwyer, S., &guish this mechanism from actions that Baker, G. T. 3rd (1991). Elevated paraquatdirectly counteract proximate causes of resistance can be used as a bioassay forcellular damage (e.g., chemical antioxi- longevity in a genetically based long-liveddant action) or actions that induce a strain of Drosophila. Developmentalstress response by actually causing minor Genetics, 12(5), 362–370. Armeni, T., Principato, G., Quiles, J. L.,damage (conventional hormesis). Pieri, C., Bompadre, S., & Battino, M. While the Hormesis Hypothesis of (2003). Mitochondrial dysfunctions duringDR may be young, the Xenohormesis aging: vitamin E deficiency or caloricHypothesis is positively embryonic. The restriction—two different ways ofability to confirm or reject it will depend modulating stress. Journal of Bioenergeticson identification of the components of & Biomembranes, 35(2), 181–191.hormetic signaling pathways and subse- Austad, S. N. (2001). Does caloric restriction inquent assessment of the effects of putative the laboratory simply prevent overfeedingxenohormetic agents on those compo- and return house mice to their natural levelnents. It may also be worthwhile, however, of food intake? Science of Aging Knowledgeto take the reverse approach to this process Environment, 2001(6), pe3. Avula, C. P., & Fernandes, G. (2002).and investigate proteins already known to Inhibition of H2O2-induced apoptosis ofinteract with phytochemicals for a possible lymphocytes by calorie restriction duringrole in DR physiology. Investigation of the aging. Microscopy Research and Technique,effects of DR on polyphenol-interacting 59(4), 282–292.proteins in mammals might provide a Ball, Z. B., Barnes, R. H., & Visscher, M. B.line of inquiry into DR signaling that is (1947). The effects of dietary caloricdistinct from, yet complementary to, those restriction on maturity and senescence,currently investigated. with particular reference to fertility and longevity. American. Journal of Physiology, 150, 511–520.References Barja, G. (2004). Aging in vertebrates, and theAidoo, A., Desai, V. G., Lyn-Cook, L. E., effect of caloric restriction: a mitochondrial Chen, J. J., Feuers, R. J., & Casciano, D. A. free radical production-DNA damage (1999). Attenuation of bleomycin-induced mechanism? Biological Reviews of Hprt mutant frequency in female and male the Cambridge Philosophical Society, 79(2), rats by calorie restriction. Mutation 235–251. Research, 430(1), 155–163. Bartke, A. (2000). Delayed aging in Ames dwarfAnderson, R. M., Bitterman, K. J., Wood, J. G., mice. Relationships to endocrine function Medvedik, O., & Sinclair, D. A. (2003). and body size. Results and Problems in Cell Nicotinamide and Pnc1 govern Differentiation, 29, 181–202. lifespan extension by calorie restriction in Bartke, A., Wright, J. C., Mattison, J. A., S. cerevisiae. Nature, 423(6936), 181–185. Ingram, D. K., Miller, R. A., & Roth, G. S.Ando, K., Higami, Y., Tsuchiya, T., (2001). Extending the lifespan of long-lived Kanematsu, T., & Shimokawa, I. (2002). mice. Nature, 414(6862), 412.
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 91Barton, A. (1950). Some aspects of cell Braeckman, B. P., Houthoofd, K., & division in Saccharomyces cerevisiae. Vanfleteren, J. R. (2001b). Insulin-like Journal of General Microbiology, 4, 84–86. signaling, metabolism, stress resistanceBarzilai, N., & Gabriely, I. (2001). The role of and aging in Caenorhabditis elegans. fat depletion in the biological benefits of Mechanisms of Ageing & Development, caloric restriction. Journal of Nutrition, 122(7), 673–693. 131(3), 903S–906S. Braeckman, B. P., Houthoofd, K., &Barzilai, N., & Gupta, G. (1999). Revisiting Vanfleteren, J. R. (2002). Assessing the role of fat mass in the life extension metabolic activity in aging Caenorhabditis induced by caloric restriction. Journal of elegans: concepts and controversies. Aging Gerontology Series A Biological Sciences & Cell, 1(2), 82–88; discussion 102–3. Medical Sciences, 54(3), B89–B98. Bras, G., & Ross, M. H. (1964). Kidney diseaseBasly, J. P., Marre-Fournier, F., Le Bail, J. C., and nutrition in the rat. Toxicology & Habrioux, G., Chulia, A. J. (2000). Applied Pharmacology, 61, 247–262. Estrogenic/antiestrogenic and scavenging Brunet, A., Sweeney, L. B., Sturgill, J. F., properties of (E)– and (Z)–resveratrol. Chua, K. F., Greer, P. L., Lin, Y., Tran, H., Life Science, 66(9), 769–77. Ross, S. E., Mostoslavsky, R., Cohen, H. Y.,Berg, B. N., & Simms, H. S. (1960). Nutrition Hu, L. S., Cheng, H. L., Jedrychowski, M. P., and longevity in the rat. II. Longevity and Gygi, S. P., Sinclair, D. A., Alt, F. W., & onset of disease with different levels of food Greenberg, M. E. (2004). Stress-dependent intake. Journal of Nutrition, 71, 255–263. regulation of FOXO transcription factors byBertrand, H. A., Lynd, F. T., Masoro, E. J., & the SIRT1 deacetylase. Science, 303(5666), Yu, B. P. (1980). Changes in adipose mass 2011–2015. and cellularity through the adult life of rats Cabelof, D. C., Yanamadala, S., Raffoul, J. J., fed ad libitum or a life-prolonging restricted Guo, Z., Soofi, A., & Heydari, A. R. (2003). diet. Journal of Gerontology, 35(6), 827–835. Caloric restriction promotes genomicBitterman, K. J., Anderson, R. M., stability by induction of base excision Cohen, H. Y., Latorre-Esteves, M., & repair and reversal of its age-related decline. Sinclair, D. A. (2002). Inhibition of silencing DNA Repair (Amsterdam), 2(3), 295–307. and accelerated aging by nicotinamide, a Calamini, B., Santarsiero, B., Ratia, K., putative negative regulator of yeast sir2 and Eggler, A., Malkowski, M., Pezzuto, J. M., & human SIRT1. Journal of Biological Mesecar, A. (2004). The role of resveratrol Chemistry, 277(47), 45099–45107. and its metabolites in longevity and cancerBitterman, K. J., Medvedik, O., & chemoprevention. Abstract, 42nd Annual Sinclair, D. A. (2003). Longevity regulation Medicinal Chemistry Meeting in Miniature. in Saccharomyces cerevisiae: linking Calingasan, N. Y., & Gibson, G. E. (2000). metabolism, genome stability, and Dietary restriction attenuates the neuronal heterochromatin. Microbiology and loss, induction of heme oxygenase-1 and Molecular Biology Reviews, 67(3), 376–399. blood-brain barrier breakdown induced byBlock, G., Patterson, B., & Subar, A. (1992). impaired oxidative metabolism. Brain Fruit, vegetables, and cancer prevention: a Research, 885(1), 62–69. review of the epidemiological evidence. Cao, S. X., Dhahbi, J. M., Mote, P. L., & Nutrition and Cancer, 18(1), 1–29. Spindler, S. R. (2001). Genomic profiling ofBluher, M., Kahn, B. B., & Kahn, C. R. (2003). short- and long-term caloric restriction Extended longevity in mice lacking the effects in the liver of aging mice. insulin receptor in adipose tissue. Science, Proceedings of the National Academy of 299(5606), 572–574. Sciences of the USA, 98(19), 10630–10635.Braeckman, B. P., Houthoofd, K., & Carey, J. R., Liedo, P., Muller, H. G., Vanfleteren, J. R. (2001a). Insulin-like Wang, J. L., Zhang, Y., & Harshman, L. signaling, metabolism, stress resistance (2005). Stochastic dietary restriction using and aging in Caenorhabditis elegans. a Markov-chain feeding protocol elicits Mechanisms of Ageing & Development, complex, life history response in medflies. 122(7), 673–693. Aging Cell, 4(1), 31–39.
92 D. A. Sinclair and K. T. HowitzCarlson, A. J., & Hoetzel, F. (1946). Apparent dietary restriction. Mechanisms of Ageing prolongation of the life-span of rats by & Development, 102(1), 1–13. intermittent fasting. Journal of Nutrition, Corder, R., Crozier, A., & Kroon, P. A. (2003). 31, 363–342. Drinking your health? It’s too early to say.Cartee, G. D., Kietzke, E. W., & Nature, 426(6963), 119. Briggs-Tung, C. (1994). Adaptation of Cutler, R. G. (1982). Longevity is determined muscle glucose transport with caloric by specific genes: testing the hypothesis. restriction in adult, middle-aged, and In R. C. Adelman & G. S. Roth (Eds.), old rats. American Journal of Physiology, Testing the theories of aging (pp. 24–114). 266(5 Pt 2), R1443–R1447. Boca Raton, FL: CRC Press.Casirola, D. M., Rifkin, B., Tsai, W., & D’Costa, A. P., Lenham, J. E., Ingram, R. L., & Ferraris, R. P. (1996). Adaptations of Sonntag, W. E. (1993). Moderate caloric intestinal nutrient transport to chronic restriction increases type 1 IGF receptors and caloric restriction in mice. American protein synthesis in aging rats. Mechanisms Journal of Physiology, 271(1 Pt 1), of Ageing & Development, 71(1–2), 59–71. G192–G200. da Cunha, G. L., da Cruz, I. B., Fiorino, P., &Cassano, P., Lezza, A. M., Leeuwenburgh, C., de Oliveira, A. K. (1995). Paraquat Cantatore, P., & Gadaleta, M. N. (2004). resistance and starvation conditions in the Measurement of the 4,834-bp mitochondrial selection of longevity extremes in DNA deletion level in aging rat liver and Drosophila melanogaster populations brain subjected or not to caloric restriction previously selected for long and short diet. Annals of the New York Academy of developmental period. Developmental Sciences, 1019, 269–273. Genetics, 17(4), 352–361.Charlesworth, B. (2000). Fisher, Medawar, Daitoku, H., Hatta, M., Matsuzaki, H., Hamilton and the evolution of aging. Aratani, S., Ohshima, T., Miyagishi, M., Genetics, 156(3), 927–931. Nakajima, T., & Fukamizu, A. (2004).Chung, H. Y., Kim, H. J., Kim, K. W., Silent information regulator 2 potentiates Choi, J. S., & Yu, B. P. (2002). Molecular Foxo1-mediated transcription through its inflammation hypothesis of aging based deacetylase activity. Proceedings of the on the anti-aging mechanism of calorie National Academy of Sciences of the USA, restriction. Microscopy Ressearch & 101(27), 10042–10047. Technique, 59(4), 264–272. de Cabo, R., Furer-Galban, S., Anson, R. M.,Chung, H. Y., Kim, H. J., Shim, K. H., & Gilman, C., Gorospe, M., & Lane, M. A. Kim, K. W. (1999). Dietary modulation of (2003). An in vitro model of caloric prostanoid synthesis in the aging restriction. Experimental Gerontology, process: role of cyclooxygenase-2. 38(6), 631–639. Mechanisms of Ageing & Development, Dean, D. J., Brozinick, J. T., Jr., 111(2–3), 97–106. Cushman, S. W., & Cartee, G. D. (1998a).Cohen, H. Y., Lavu, S., Bitterman, K. J., Calorie restriction increases cell surface Hekking, B., Imahiyerobo, T. A., Miller, C., GLUT-4 in insulin-stimulated skeletal Frye, R., Ploegh, H., Kessler, B. M., & muscle. American Journal of Physiology, Sinclair, D. A. (2004a). Acetylation of the 275(6 Pt 1), E957–E964. C terminus of Ku70 by CBP and PCAF Dean, D. J., Gazdag, A. C., Wetter, T. J., & controls Bax-mediated apoptosis. Molecular Cartee, G. D. (1998b). Comparison of the Cell, 13(5), 627–638. effects of 20 days and 15 months of calorieCohen, H. Y., Miller, C., Bitterman, K. J., restriction on male Fischer 344 rats. Aging Wall, N. R., Hekking, B., Kessler, B., (Milano), 10(4), 303–307. Howitz, K. T., Gorospe, M., de Cabo, R., & Dehmelt, H. (2004). Re-adaptation Sinclair, D. A. (2004b). Calorie restriction hypothesis: explaining health benefits of promotes mammalian cell survival by caloric restriction. Medical Hypotheses, inducing the SIRT1 deacetylase. Science, 62(4), 620–624. 305(5682), 390–392. Del Roso, A., Vittorini, S., Cavallini, G.,Cook, C. I., & Yu, B. P. (1998). Iron Donati, A., Gori, Z., Masini, M., Pollera, M., accumulation in aging: modulation by & Bergamini, E. (2003). Ageing-related
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 93 changes in the in vivo function of rat Fabrizio, P., Pozza, F., Pletcher, S. D., liver macroautophagy and proteolysis. Gendron, C. M., & Longo, V. D. (2001). Experimental Gerontology, 38(5), 519–527. Regulation of longevity and stressDeLany, J. P., Hansen, B. C., Bodkin, N. L., resistance by Sch9 in yeast. Science, Hannah, J., & Bray, G. A. (1999). Long-term 292(5515), 288–290. calorie restriction reduces energy Ferraris, R. P., Cao, Q. X., & Prabhakaram, S. expenditure in aging monkeys. Journals of (2001). Chronic but not acute energy Gerontology Series A Biological Sciences & restriction increases intestinal nutrient Medical Sciences, 54(1), B5–B11; discussion transport in mice. Journal of Nutrition, B12–B13. 131(3), 779–786.Dempsey, J. L., Pfeiffer, M., & Morley, A. A. Fontana, L., Meyer, T. E., Klein, S., & (1993). Effect of dietary restriction on in Holloszy, J. O. (2004). Long-term calorie vivo somatic mutation in mice. Mutation restriction is highly effective in reducing Research, 291(2), 141–145. the risk for atherosclerosis in humans.Dhahbi, J. M., Kim, H. J., Mote, P. L., Proceedings of the National Academy of Beaver, R. J., & Spindler, S. R. (2004). Sciences of the USA, 101(17), 6659–6663. Temporal linkage between the phenotypic Fontecave, M., Lepoivre, M., Elleingand, E., and genomic responses to caloric Gerez, C., & Guittet, O. (1998). Resveratrol, restriction. Proceedings of the National a remarkable inhibitor of ribonucleotide Academy of Sciences of the USA, 101(15), reductase. FEBS Letters, 421(3), 277–279. 5524–5529. Frankel, E. N., Waterhouse, A. L., &Dixon, R. A., & Paiva, N. L. (1995). Kinsella, J. E. (1993). Inhibition of human Stress-induced phenylpropanoid LDL oxidation by resveratrol. Lancet, metabolism. Plant Cell, 7(7), 1085–1097. 341(8852), 1103–1104.Droge, W. (2003). Oxidative stress and aging. Friedman, D. B., & Johnson, T. E. (1988). Advances in Experimental Medicine & A mutation in the age-1 gene in Biology, 543, 191–200. Caenorhabditis elegans lengthens lifeDuffy, P. H., Leakey, J. E., Pipkin, J. L., and reduces hermaphrodite fertility. Turturro, A., & Hart, R. W. (1997). The Genetics, 118(1), 75–86. physiologic, neurologic, and behavioral Frye, R. A. (2000). Phylogenetic classification effects of caloric restriction related to of prokaryotic and eukaryotic Sir2-like aging, disease, and environmental factors. proteins. Biochemical & Biophysical Environmental Research, 73(1–2), Research Communications, 273(2), 242–248. 793–798.Duffy, P. H., Seng, J. E., Lewis, S. M., Fukuhara, A., Matsuda, M., Nishizawa, M., Mayhugh, M. A., Aidoo, A., Hattan, D. G., Segawa, K., Tanaka, M., Kishimoto, K., Casciano, D. A., & Feuers, R. J. (2001). Matsuki, Y., Murakami, M., Ichisaka, T., The effects of different levels of dietary Murakami, H., Watanabe, E., Takagi, T., restriction on aging and survival in the Akiyoshi, M., Ohtsubo, T., Kihara, S., Sprague-Dawley rat: implications for Yamashita, S., Makishima, M., chronic studies. Aging (Milano), 13(4), Funahashi, T., Yamanaka, S., Hiramatsu, R., 263–272. Matsuzawa, Y., & Shimomura, I. (2004).Dufour, E., & Larsson, N. G. (2004). Visfatin: a protein secreted by visceral fat Understanding aging: revealing order out that mimics the effects of insulin. Science, of chaos. Biochimica et Biophysica Acta, 307(5708), 426–430. 1658(1–2), 122–132. Fulco, M., Schiltz, R. L., Iezzi, S., King, M. T.,Edwards, I. J., Rudel, L. L., Terry, J. G., Zhao, P., Kashiwaya, Y., Hoffman, E., Kemnitz, J. W., Weindruch, R., & Veech, R. L., & Sartorelli, V. (2003). Sir2 Cefalu, W. T. (1998). Caloric restriction regulates skeletal muscle differentiation as in rhesus monkeys reduces low density a potential sensor of the redox state. lipoprotein interaction with arterial Molecular Cell, 12(1), 51–62. proteoglycans. Journals of Gerontology Fulgencio, J. P., Kohl, C., Girard, J., & Series A Biological Sciences & Medical Pegorier, J. P. (2001). Effect of metformin Sciences, 53(6), B443–B448. on fatty acid and glucose metabolism in
94 D. A. Sinclair and K. T. Howitz freshly isolated hepatocytes and on specific and skeletal muscle mitochondria. Annals gene expression in cultured hepatocytes. of the New York Academy of Sciences, Biochemical Pharmacology, 62(4), 439–446. 1019, 333–342.Gallo, C. M., Smith, D. L., Jr., & Smith, J. S. Guo, Z., Heydari, A., & Richardson, A. (1998). (2004). Nicotinamide clearance by Pnc1 Nucleotide excision repair of actively directly regulates Sir2-mediated silencing transcribed versus nontranscribed DNA in and longevity. Molecular & Cellular rat hepatocytes: effect of age and dietary Biology, 24(3), 1301–1312. restriction. Experimental Cell Research,Gems, D., Pletcher, S., & Partridge, L. (2002). 245(1), 228–238. Interpreting interactions between treatments Guo, Z. M., Yang, H., Hamilton, M. L., that slow aging. Aging Cell, 1(1), 1–9. VanRemmen, H., & Richardson, A. (2001).Gen Son, T., Zou, Y., Pal Yu, B., Lee, J., & Effects of age and food restriction on Young Chung, H. (2005). Aging effect on oxidative DNA damage and antioxidant myeloperoxidase in rat kidney and its enzyme activities in the mouse aorta. modulation by calorie restriction. Free Mechanisms of Ageing & Development, Radicals Research, 39(3), 283–289. 122(15), 1771–1786.Gescher, A. J., & Steward, W. P. (2003). Gupta, G., Cases, J. A., She, L., Ma, X. H., Relationship between mechanisms, Yang, X. M., Hu, M., Wu, J., Rossetti, L., & bioavailibility, and preclinical Barzilai, N. (2000). Ability of insulin to chemopreventive efficacy of resveratrol: a modulate hepatic glucose production in conundrum. Cancer Epidemiology, aging rats is impaired by fat accumulation. Biomarkers & Prevention, 12(10), 953–957. American Journal of Physiology, 278(6),Giannakou, M. E., Goss, M., Junger, M. A., E985–E991. Hafen, E., Leevers, S. J., & Partridge, L. Haley-Zitlin, V., & Richardson, A. (1993). (2004). Long-lived Drosophila with Effect of dietary restriction on DNA repair overexpressed dFOXO in adult fat body. and DNA damage. Mutation Research, Science, 305(5682), 361. 295(4–6), 237–245.Goldberg, D. M., Yan, J., & Soleas, G. J. Harmon, D. (1956). Aging: a theory based on (2003). Absorption of three wine-related free radical and radiation chemistry. polyphenols in three different matrices by Journal of Gerontology, 11, 298. healthy subjects. Clinical Biochemistry, Harrison, D. E., Archer, J. R., & Astle, C. M. 36(1), 79–87. (1984). Effects of food restriction on aging:Gong, X., Shang, F., Obin, M., Palmer, H., separation of food intake and adiposity. Scrofano, M. M., Jahngen-Hodge, J., Proceedings of the National Academy of Smith, D. E., & Taylor, A. (1997). Sciences of the USA, 81(6), 1835–1838. Antioxidant enzyme activities in lens, Harshman, L. G., Moore, K. M., Sty, M. A., & liver and kidney of calorie restricted Emory Magwire, M. M. (1999). Stress resistance mice. Mechanisms of Ageing & and longevity in selected lines of Development, 99(3), 181–192. Drosophila melanogaster. Neurobiology ofGood, T. P., & Tatar, M. (2001). Age-specific Aging, 20(5), 521–529. mortality and reproduction respond to Hart, R. W., Keenan, K., Turturro, A., adult dietary restriction in Drosophila Abdo, K. M., Leakey, J., & Lyn-Cook, B. melanogaster. Journal of Insect Physiology, (1995). Caloric restriction and toxicity. 47(12), 1467–1473. Fundamentals of Applied Toxicology, 25(2),Goodrick, C. L., Ingram, D. K., 184–195. Reynolds, M. A., Freeman, J. R., & Cider, N. Hatano, E., Tanaka, A., Kanazawa, A., (1990). Effects of intermittent feeding upon Tsuyuki, S., Tsunekawa, S., Iwata, S., body weight and lifespan in inbred mice: Takahashi, R., Chance, B., & Yamaoka, Y. interaction of genotype and age. Mechanisms (2004). Inhibition of tumor necrosis of Ageing & Development, 55(1), 69–87. factor-induced apoptosis in transgenicGredilla, R., Phaneuf, S., Selman, C., mouse liver expressing creatine kinase. Kendaiah, S., Leeuwenburgh, C., & Barja, G. Liver International, 24(4), 384–393. (2004). Short-term caloric restriction and Hayflick, L. (1994). How and why we age. sites of oxygen radical generation in kidney New York: Ballantine.
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 95Hayflick, L. (1999). Aging and the genome. Howitz, K. T., Bitterman, K. J., Cohen, H. Y., Science, 283(5410), 2019. Lamming, D. W., Lavu, S., Wood, J. G.,Helfand, S. L. (2004). personal Zipkin, R. E., Chung, P., Kisielewski, A., communication, manuscript in press. Zhang, L. L., Scherer, B., & Sinclair, D. A.Hercus, M. J., Loeschcke, V., & Rattan, S. I. (2003). Small molecule activators of (2003). Lifespan extension of Drosophila sirtuins extend Saccharomyces melanogaster through hormesis by repeated cerevisiae lifespan. Nature, mild heat stress. Biogerontology, 4(3), 425(6954), 191–196. 149–156. Hulbert, A. J., Clancy, D. J., Mair, W.,Hertog, M. G., Feskens, E. J., Hollman, P. C., Braeckman, B. P., Gems, D., & Partridge, L. Katan, M. B., & Kromhout, D. (1993). (2004). Metabolic rate is not reduced by Dietary antioxidant flavonoids and risk of dietary-restriction or by lowered coronary heart disease: the Zutphen Elderly insulin/IGF-1 signalling and is not Study. Lancet, 342(8878), 1007–1011. correlated with individual lifespan inHeydari, A. R., Wu, B., Takahashi, R., Drosophila melanogaster. Experimental Strong, R., & Richardson, A. (1993). Gerontology, 39(8), 1137–1143. Expression of heat shock protein 70 is Hursting, S. D., Lavigne, J. A., Berrigan, D., altered by age and diet at the level of Perkins, S. N., & Barrett, J. C. (2003). transcription. Molecular & Cellular Calorie restriction, aging, and cancer Biology, 13(5), 2909–2018. prevention: mechanisms of action andHigami, Y., & Shimokawa, I. (2000). applicability to humans. Annual Review of Apoptosis in the aging process. Cell Tissue Medicine, 54, 131–152. Research, 301(1), 125–132. Hwangbo, D. S., Gersham, B., Tu, M. P.,Hiona, A., & Leeuwenburgh, C. (2004). Palmer, M., & Tatar, M. (2004). Drosophila Effects of age and caloric restriction on dFOXO controls lifespan and regulates brain neuronal cell death/survival. Annals insulin signalling in brain and fat body. of the New York Academy of Sciences, Nature, 429(6991), 562–566. 1019, 96–105. Ikeyama, S., Wang, X. T., Li, J., Podlutsky, A.,Holliday, R. (1989). Food, reproduction and Martindale, J. L., Kokkonen, G., longevity: is the extended lifespan of van Huizen, R., Gorospe, M., & calorie-restricted animals an evolutionary Holbrook, N. J. (2003). Expression of adaptation? Bioessays, 10(4), 125–127. the pro-apoptotic gene gadd153/chop isHolloszy, J. O. (1997). Mortality rate and elevated in liver with aging and sensitizes longevity of food-restricted exercising male cells to oxidant injury. Journal of Biological rats: a reevaluation. Journal of Applied Chemistry, 278(19), 16726–16731. Physiology, 82(2), 399–403. Imai, S., Armstrong, C. M., Kaeberlein, M., &Holzenberger, M. (2004). The GH/IGF-I axis Guarente, L. (2000). Transcriptional and longevity. European Journal of silencing and longevity protein Sir2 is an Endocrinology, 151 Supplement 2, NAD- dependent histone deacetylase. S023–S027. Nature, 403(6771), 795–800.Houthoofd, K., Braeckman, B. P., Ingram, D. K., Cutler, R. G., Weindruch, R., Johnson, T. E., & Vanfleteren, J. R. (2003). Renquist, D. M., Knapka, J. J., April, M., Life extension via dietary restriction is Belcher, C. T., Clark, M. A., independent of the Ins/IGF-1 signalling Hatcherson, C. D., Marriott, B. M., & pathway in Caenorhabditis elegans. Roth, G. S. (1990). Dietary restriction and Experimental Gerontology, 38(9), aging: the initiation of a primate study. 947–954. Journal of Gerontology, 45(5), B148–B163.Houthoofd, K., Braeckman, B. P., Lenaerts, I., Ingram, D. K., Anson, R. M., de Cabo, R., Brys, K., De Vreese, A., Van Eygen, S., & Mamczarz, J., Zhu, M., Mattison, J., Vanfleteren, J. R. (2002). Axenic growth Lane, M. A., Roth, G. S. (2004). up-regulates mass-specific metabolic rate, Development of calorie restriction stress resistance, and extends life span in mimetics as a prolongevity strategy. Caenorhabditis elegans. Experimental Annals of the New York Academy of Gerontology, 37(12), 1371–1378. Sciences, 1019, 412–23.
96 D. A. Sinclair and K. T. HowitzIwasaki, K., Gleiser, C. A., Masoro, E. J., Keli, S. O., Hertog, M. G., Feskens, E. J., & McMahan, C. A., Seo, E. J., & Yu, B. P. Kromhout, D. (1996). Dietary flavonoids, (1988). The influence of dietary protein antioxidant vitamins, and incidence of source on longevity and age-related disease stroke: the Zutphen study. Archives of processes of Fischer rats. Journal of Internal Medicine, 156(6), 637–642. Gerontology, 43(1), B5–B12. Kemnitz, J. W., Weindruch, R., Roecker, E. B.,Izmaylov, D. M., & Obukhova, L. K. (1999). Crawford, K., Kaufman, P. L., & Geroprotector effectiveness of melatonin: Ershler, W. B. (1993). Dietary restriction investigation of lifespan of Drosophila of adult male rhesus monkeys: design, melanogaster. Mechanisms of Ageing & methodology, and preliminary findings Development, 106(3), 233–240. from the first year of study. Journal ofJames, S. J., Muskhelishvili, L., Gaylor, D. W., Gerontology, 48(1), B17–B26. Turturro, A., & Hart, R. (1998). Kennedy, B. K., Austriaco, N. R., Jr., Zhang, J., Upregulation of apoptosis with dietary & Guarente, L. (1995). Mutation in the restriction: implications for carcinogenesis silencing gene SIR4 can delay aging in and aging. Environmental Health S. cerevisiae. Cell, 80(3), 485–496. Perspectives, 106 Supplement 1, 307–312. Kenyon, C. (2001). A conserved regulatoryJang, M., Cai, L., Udeani, G. O., Slowing, K. V., mechanism for aging. Cell, 105, 165–168. Thomas, C. F., Beecher, C. W., Fong, H. H., Kenyon, C., Chang, J., Gensch, E., Rudner, A., Farnsworth, N. R., Kinghorn, A. D., & Tabtiang, R. (1993). A C. elegans mutant Mehta, R. G., Moon, R. C., & Pezzuto, J. M. that lives twice as long as wild type. (1997). Cancer chemopreventive activity of Nature, 366(6454), 461–464. resveratrol, a natural product derived from Khavinson, V. K., Izmaylov, D. M., grapes. Science, 275(5297), 218–220. Obukhova, L. K., & Malinin, V. V. (2000).Jazwinski, S. M., Chen, J. B., & Sun, J. (1993). Effect of epitalon on the lifespan increase A single gene change can extend yeast life in Drosophila melanogaster. Mechanisms of span: the role of Ras in cellular senescence. Ageing & Development, 120(1–3), 141–149. Advances in Experimental Medicine & Kim, J. W., Zou, Y., Yoon, S., Lee, J. H., Biology, 330, 45–53. Kim, Y. K., Yu, B. P., & Chung, H. Y. (2004).Johnson, J. R. (1966). Reproductive capacity Vascular aging: molecular modulation and mode of death of yeast cells. Antonie of the prostanoid cascade by calorie van Leeuwenhoek, 32, 94–98. restriction. Journals of Gerontology SeriesJohnson, T. E., Lithgow, G. J., & Murakami, S. A Biological Sciences & Medical Sciences, (1996). Hypothesis: interventions that 59(9), B876–B885. increase the response to stress offer the King, J. T., & Visscher, M. B. (1950). potential for effective life prolongation and Longevity as a function of diet in the C3H increased health. Journals of Gerontology mouse. Faseb Journal, 9, 70–74. Series A Biological Sciences & Medical Kirkwood, T. B., & Holliday, R. (1979). The Sciences, 51(6), B392–B395. evolution of ageing and longevity.Kaeberlein, M., Kirkland, K. T., Fields, S., & Proceedings of the Royal Society of London Kennedy, B. K. (2004). Sir2-Independent Life Series B Biological Sciences, 205(1161), Span Extension by Calorie Restriction in 531–546. Yeast. PLoS Biology, 2(9), E296. Knekt, P., Kumpulainen, J., Jarvinen, R.,Kaeberlein, M., McVey, M., & Guarente, L. Rissanen, H., Heliovaara, M., Reunanen, A., (1999). The SIR2/3/4 complex and SIR2 Hakulinen, T., & Aromaa, A. (2002). alone promote longevity in Flavonoid intake and risk of chronic Saccharomyces cerevisiae by two different diseases. American Journal of Clinical mechanisms. Genes & Development, Nutrition, 76(3), 560–568. 13(19), 2570–2580. Kobayashi, T., Horiuchi, T., Tongaonkar, P.,Kalant, N., Stewart, J., & Kaplan, R. (1988). Vu, L., & Nomura, M. (2004). SIR2 Effect of diet restriction on glucose regulates recombination between different metabolism and insulin responsiveness in rDNA repeats, but not recombination aging rats. Mechanisms of Ageing & within individual rRNA genes in yeast. Development, 46(1–3), 89–104. Cell, 117(4), 441–453.
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 97Koizumi, A., Wada, Y., Tuskada, M., Kayo, T., extends yeast life span by lowering Naruse, M., Horiuchi, K., Mogi, T., the level of NADH. Genes & Development, Yoshioka, M., Sasaki, M., Miyamaura, Y., 18(1), 12–16. Abe, T., Ohtomo, K., & Walford, R.L. Lindsay, D. G. (1999). Diet and ageing: the (1996). A tumor preventive effect of dietary possible relation to reactive oxygen species. restriction is antagonized by a high housing Journal of Nutrition & Health: Aging, 3(2), temperature through deprivation of torpor. 84–91. Mechanisms of Ageing & Development, Lints, F. A., Bullens, P., & Le Bourg, E. (1993). 92(1), 67–82. Hypergravity and aging in DrosophilaKoizumi, A., Weindruch, R., & Walford, R. L. melanogaster: 7. New longevity data. (1987). Influences of dietary restriction and Experimental Gerontology, 28(6), 611–615. age on liver enzyme activities and lipid Lipman, J. M., Turturro, A., & Hart, R. W. peroxidation in mice. Journal of Nutrition, (1989). The influence of dietary restriction 117(2), 361–367. on DNA repair in rodents: a preliminaryKrystal, B. S., & Yu, B. P. (1994). Aging and study. Mechanisms of Ageing & its modulation by dietary restriction. Development, 48(2), 135–143. In B. P. Yu (Ed.), Modulation of aging Lithgow, G. J. (2001). Hormesis: a new hope processes by dietary restriction (pp. 1–36). for ageing studies or a poor second to London: CRC Press. genetics? Human Experimental Toxicology,Lamming, D. W., Wood, J. G., & Sinclair, D. 20(6), 301–303; discussion 319–320. A. (2004). Small molecules that regulate Liu, R. H. (2003). Health benefits of fruit and lifespan: evidence for xenohormesis. vegetables are from additive and synergistic Molecular Microbiology, 53(4), 1003–1009. combinations of phytochemicals. AmericanLangley, E. P. M., Faretta M., Bauer, U.M., Journal of Clinical Nutrition, 78(3 Suppl), Frye, R.A., Minucci, S., Pelicci, P.G., & 517S–520S. Kouzarides, T. (2002). Human SIR2 Longo, V. D., & Fabrizio, P. (2002). Regulation deacetylates p53 and antagonizes of longevity and stress resistance: a PML/p53-induced cellular senescence. molecular strategy conserved from yeast to EMBO Journal, 21(10), 2383–2396. humans? Cellular & Molecular LifeLaurenson, P., & Rine, J. (1992). Silencers, Sciences, 59(6), 903–908. silencing, and heritable transcriptional states. Lopez-Torres, M., Gredilla, R., Sanz, A., & Microbiological Reviews, 56(4), 543–560. Barja, G. (2002). Influence of aging and long-Lee, S. S., Kennedy, S., Tolonen, A. C., & term caloric restriction on oxygen radical Ruvkun, G. (2003a). DAF-16 target genes generation and oxidative DNA damage in that control C. elegans life-span and rat liver mitochondria. Free Radicals in metabolism. Science, 300(5619), 644–647. Biology & Medicine, 32(9), 882–889.Lee, S. S., Lee, R. Y., Fraser, A. G., Luhtala, T. A., Roecker, E. B., Pugh, T., Kamath, R. S., Ahringer, J., & Ruvkun, G. Feuers, R. J., & Weindruch, R. (1994). (2003b). A systematic RNAi screen identifies Dietary restriction attenuates age-related a critical role for mitochondria in C. elegans increases in rat skeletal muscle antioxidant longevity. Nature Genetics, 33(1), 40–48. enzyme activities. Journal of Gerontology,Lewis, S. E., Goldspink, D. F., Phillips, J. G., 49(5), B231–B238. Merry, B. J., & Holehan, A. M. (1985). The Luo, J., Nikolaev, A. Y., Imai, S., Chen, D., effects of aging and chronic dietary Su, F., Shiloh, A., Guarente, L., & Gu, W. restriction on whole body growth and (2001). Negative control of p53 by Sir2alpha protein turnover in the rat. Experimental promotes cell survival under stress. Cell, Gerontology, 20(5), 253–263. 107(2), 137–148.Lin, S. J., Defossez, P. A., & Guarente, L. Masoro, E. J. (1985). Nutrition and aging: a (2000). Requirement of NAD and SIR2 for current assessment. Journal of Nutrition, life-span extension by calorie restriction 115(7), 842–848. in Saccharomyces cerevisiae. Science, Masoro, E. J. (1995). Antiaging action of 289(5487), 2126–2128. caloric restriction: endocrine and metabolicLin, S. J., Ford, E., Haigis, M., Liszt, G., & aspects. Obesity Research, 3 Suppl 2, Guarente, L. (2004). Calorie restriction 241s–247s.
98 D. A. Sinclair and K. T. HowitzMasoro, E. J. (1998). Influence of caloric a factor in the action of dietary restriction intake on aging and on the response to on aging: effects in Fischer 344 rats. Aging stressors. Journal of Toxicology & (Milano), 9(1–2), 73–9. Environmental Health Part B Critical McCay, C. M. (1934). Cellulose in the diet Reviews, 1(3), 243–257. of mice and rats. Journal of Nutrition, 8,Masoro, E. J. (2000). Caloric restriction and 435–447. aging: an update. Experimental McCay, C. M., Crowell, M. F., & Gerontology, 35(3), 299–305. Maynard, L. A. (1935). The effect ofMasoro, E. J. (2001). Dietary restriction: retarded growth upon the life-span and an experimental approach to the study upon the ultimate body size. Journal of of the biology of aging. In E. J. Masoro & Nutrition, 10, 63. S. N. Austad (Eds.), Handbook of the McKiernan, S. H., Bua, E., McGorray, J., & biology of aging (5th ed., pp. 396–420). Aiken, J. (2004). Early-onset calorie New York: Academic Press. restriction conserves fiber number in agingMasoro, E. J., & Austad, S. N. (1996). The rat skeletal muscle. Faseb Journal, 18(3), evolution of the antiaging action of dietary 580–581. restriction: a hypothesis. Journal of Medawar, P. B. (1946). Old age and natural Gerontology Series A Biological Sciences & death. Modern Quarterly, 1, 30–56. Medical Sciences, 51(6), B387–B391. Merry, B. J. (2002). Molecular mechanismsMasoro, E.J., McCarter, R. J., Katz, M. S., & linking calorie restriction and longevity. McMahan, C. A. (1992). Dietary restriction International Journal of Biochemistry & alters characteristics of glucose fuel use. Cell Biology, 34(11), 1340–1354. Journal of Gerontology, 47(6), B202–B208. Michalski, A. I., Johnson, T. E., Cypser, J. R.,Masoro, E. J., Yu, B. P., & Bertrand, H. A. & Yashin, A. I. (2001). Heating stress (1982). Action of food restriction in patterns in Caenorhabditis elegans delaying the aging process. Proceedings of longevity and survivorship. Biogerontology, the National Academy of Sciences of the 2(1), 35–44. USA, 79(13), 4239–4241. Migliaccio, E., Giorgio, M., Mele, S.,Mattson, M. P., Chan, S. L., & Duan, W. Pelicci, G., Reboldi, P., Pandolfi, P. P., (2002a). Modification of brain aging and Lanfrancone, L., & Pelicci, P. G. (1999). neurodegenerative disorders by genes, diet, The p66shc adaptor protein controls and behavior. Physiological Reviews, 82(3), oxidative stress response and life span in 637–672. mammals. Nature, 402(6759), 309–313.Mattson, M. P., Duan, W., Chan, S. L., Miles, P. D., Barak, Y., Evans, R. M., & Cheng, A., Haughey, N., Gary, D. S., Olefsky, J. M. (2003). Effect of heterozygous Guo, Z., Lee, J., & Furukawa, K. (2002b). PPARgamma deficiency and TZD Neuroprotective and neurorestorative treatment on insulin resistance associated signal transduction mechanisms in brain with age and high-fat feeding. American aging: modification by genes, diet and Journal of Physiology, 284(3), E618–E626. behavior. Neurobiology of Aging, 23(5), Minois, N., Guinaudy, M. J., Payre, F., & 695–705. Le Bourg, E. (1999). HSP70 induction mayMcCarter, R. J., & Palmer, J. (1992). Energy explain the long-lasting resistance to heat metabolism and aging: a lifelong study of of Drosophila melanogaster having lived Fischer 344 rats. American Journal of in hypergravity. Mechanisms of Ageing & Physiology, 263(3 Pt 1), E448–E452. Development, 109(1), 65–77.McCarter, R., Masoro, E. J., & Yu, B. P. (1985). Miwa, S., Riyahi, K., Partridge, L., & Does food restriction retard aging by Brand, M. D. (2004). Lack of correlation reducing the metabolic rate? American between mitochondrial reactive oxygen Journal of Physiology, 248(4 Pt 1), species production and life span in E488–E490. Drosophila. Annals of the New YorkMcCarter, R. J., Shimokawa, I., Ikeno, Y., Academy of Sciences, 1019, 388–391. Higami, Y., Hubbard, G. B., Yu, B. P., & Mockett, R. J., Orr, W. C., Rahmandar, J. J., McMahan, C. A. (1997). Physical activity as Sohal, B. H., & Sohal, R. S. (2001).
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 99 Antioxidant status and stress resistance in Orentreich, N., Matias, J. R., DeFelice, A., & long- and short-lived lines of Drosophila Zimmerman, J. A. (1993). Low methionine melanogaster. Experimental Gerontology, ingestion by rats extends life span. Journal 36(3), 441–463. of Nutrition, 123(2), 269–274.Monod, J., Wyman, J., & Changeux, J.-P. Osborne, T. B., & Mendel, L. B. (1917). The (1965). On the nature of allosteric effect of retardation of growth upon the transitions. Journal of Molecular Biology, breeding period and duration of life of rats. 12, 88–118. Science, 45, 294–295.Monti, B., & Contestabile, A. (2003). Selective Pahlavani, M. A. (2004). Influence of caloric alteration of DNA fragmentation and restriction on aging immune system. caspase activity in the spinal cord of aged Journal of Nutrition, Health & Aging, 8(1), rats and effect of dietary restriction. Brain 38–47. Research, 992(1), 137–141. Pahlavani, M. A., Harris, M. D., Moore, S. A.,Moore, W. A., Davey, V. A., Weindruch, R., Weindruch, R., & Richardson, A. (1995). Walford, R., & Ivy, G. O. (1995). The effect The expression of heat shock protein 70 of caloric restriction on lipofuscin decreases with age in lymphocytes from accumulation in mouse brain with age. rats and rhesus monkeys. Experimental Gerontology, 41 Suppl 2, 173–185. Cell Research, 218(1), 310–318.Mote, P. L., Grizzle, J. M., Walford, R. L., & Pandey, R., Muller, A., Napoli, C. A., Spindler, S. R. (1991). Influence of age and Selinger, D. A., Pikaard, C. S., Richards, E. J., caloric restriction on expression of hepatic Bender, J., Mount, D. W., & Jorgensen, R. A. genes for xenobiotic and oxygen (2002). Analysis of histone acetyltransferase metabolizing enzymes in the mouse. and histone deacetylase families of Journal of Gerontology, 46(3), B95–B100. Arabidopsis thaliana suggests functionalMotta, M. C., Divecha, N., Lemieux, M., diversification of chromatin modification Kamel, C., Chen, D., Gu, W., Bultsma, Y., among multicellular eukaryotes. Nucleic McBurney, M., & Guarente, L. (2004). Acids Research, 30(23), 5036–5055. Mammalian SIRT1 represses forkhead Parkes, T. L., Elia, A. J., Dickinson, D., transcription factors. Cell, 116(4), 551–563. Hilliker, A. J., Phillips, J. P., &Mukherjee, P., Abate, L. E., & Seyfried, T. N. Boulianne, G. L. (1998). Extension of (2004). Antiangiogenic and proapoptotic Drosophila lifespan by overexpression of effects of dietary restriction on experimental human SOD1 in motorneurons. Nature mouse and human brain tumors. Clinical Genetics, 19(2), 171–174. Cancer Research, 10(16), 5622–5629. Pearl, R. (1928). The rate of living. New York:Murakami, S., Salmon, A., & Miller, R. A. Alfred Knopf. (2003). Multiplex stress resistance in cells Pervaiz, S. (2003). Resveratrol: from from long-lived dwarf mice. Faseb Journal, grapevines to mammalian biology. Faseb 17(11), 1565–1566. Journal, 17(14), 1975–1985.Murphy, C. T., McCarroll, S. A., Picard, F., Kurtev, M., Chung, N., Bargmann, C. I., Fraser, A., Kamath, R. S., Topark-Ngarm, A., Senawong, T., Machado Ahringer, J., Li, H., & Kenyon, C. (2003). De Oliveira, R., Leid, M., McBurney, M. W., Genes that act downstream of DAF-16 to & Guarente, L. (2004). Sirt1 promotes fat influence the lifespan of Caenorhabditis mobilization in white adipocytes by elegans. Nature, 424(6946), 277–283. repressing PPAR-gamma. Nature,Napoli, C., Lemieux, C., & Jorgensen, R. 429(6993), 771–776. (1990). Introduction of a chimeric chalcone Prapurna, D. R., & Rao, K. S. (1996). synthase gene into petunia results in Long-term effects of caloric restriction reversible co-suppression of homologous initiated at different ages on DNA genes in trans. Plant Cell, 2(4), 279–289. polymerases in rat brain. Mechanisms ofNelson, W., & Halberg, F. (1986). Meal-timing, Ageing & Development, 92(2–3), 133–142. circadian rhythms and life span of Prolla, T. A. (2002). DNA microarray analysis mice. Journal of Nutrition, 116(11), of the aging brain. Chemical Senses, 27(3), 2244–2253. 299–306.
100 D. A. Sinclair and K. T. HowitzPugh, T. D., Klopp, R. G., & Weindruch, R. Richie, J. P. Jr., Leutzinger, Y., (1999). Controlling caloric consumption: Parthasarathy, S., Malloy, V., protocols for rodents and rhesus monkeys. Orentreich, N., & Zimmerman, J. A. (1994). Neurobiology of Aging, 20(2), 157–165. Methionine restriction increases bloodPuigserver, P., & Speigelman, B. (2004). glutathione and longevity in F344 rats. Personal communication. Faseb Journal, 8(15), 1302–1307.Raff, M. C., Whitmore, A. V., & Finn, J. T. Roberts, S. B., Pi-Sunyer, X., Kuller, L., (2002). Axonal self-destruction and Lane, M. A., Ellison, P., Prior, J. C., & neurodegeneration. Science, 296(5569), Shapses, S. (2001). Physiologic effects of 868–871. lowering caloric intake in nonhumanRaffoul, J. J., Guo, Z., Soofi, A., & primates and nonobese humans. Journals Heydari, A. R. (1999). Caloric restriction of Gerontology Series A Biological and genomic stability. Journal of Nutrition, Sciences & Medical Sciences, 56 Spec No Health, & Aging, 3(2), 102–110. 1, 66–75.Rao, G., Xia, E., Nadakavukaren, M. J., & Robertson, T. B., & Ray, L. A. (1920). On the Richardson, A. (1990). Effect of dietary growth of relatively long-lived compared restriction on the age-dependent changes in with that of relatively short lived animals. the expression of antioxidant enzymes in Journal of Biological Chemistry, 42, 71–107. rat liver. Journal of Nutrition, 120(6), Rojas, C., Cadenas, S., Perez-Campo, R., 602–609. Lopez-Torres, M., Pamplona, R., Prat, J., &Rattan, S. I. (2004). Hormetic mechanisms of Barja, G. (1993). Relationship between lipid anti-aging and rejuvenating effects of peroxidation, fatty acid composition, and repeated mild heat stress on human ascorbic acid in the liver during fibroblasts in vitro. Rejuvenation Research, carbohydrate and caloric restriction in 7(1), 40–48. mice. Archives of Biochemistry &Ravagnan, L., Gurbuxani, S., Susin, S. A., Biophysics, 306(1), 59–64. Maisse, C., Daugas, E., Zamzami, N., Rongvaux, A., Shea, R. J., Mulks, M. H., Mak, T., Jaattela, M., Penninger, J. M., Gigot, D., Urbain, J., Leo, O., & Andris, F. Garrido, C., & Kroemer, G. (2001). (2002). Pre-B-cell colony-enhancing factor, Heat-shock protein 70 antagonizes whose expression is up-regulated in apoptosis-inducing factor. Nature Cell activated lymphocytes, is a nicotinamide Biology, 3(9), 839–843. phosphoribosyltransferase, a cytosolicReveillaud, I., Kongpachith, A., Park, R., & enzyme involved in NAD biosynthesis. Fleming, J. E. (1992). Stress resistance of European Journal of Immunology, 32(11), Drosophila transgenic for bovine CuZn 3225–3234. superoxide dismutase. Free Radical Ross, M. H. (1972). Length of life and caloric Research Communications, 17(1), 73–85. intake. American Journal of ClinicalRevollo, J. R., Grimm, A. A., & Imai, S. (2004). Nutrition, 25(8), 834–838. The NAD biosynthesis pathway mediated Roth, G. S., Ingram, D. K., & Lane, M. A. by nicotinamide phosphoribosyltransferase (2001). Caloric restriction in primates regulates Sir2 activity in mammalian cells. and relevance to humans. Annals of Journal of Biological Chemistry, 279(49), the New York Academy of Sciences, 928, 50754–50763. 305–315.Reznick, D. N., Bryant, M. J., Roff, D., Rubner, M. (1908). Das problem der Ghalambor, C. K., & Ghalambor, D. E. Lebensdauer und seine Beziehungen zum (2004). Effect of extrinsic mortality on the Wachstum under Ehrnarung. Munich: evolution of senescence in guppies. Nature, Oldenburg, 150–204. 431(7012), 1095–1099. Sabatino, F., Masoro, E. J., McMahan, C. A., &Richie, J. P. Jr., Komninou, D., Leutzinger, Y., Kuhn, R. W. (1991). Assessment of the role Kleinman, W., Orentreich, N., Malloy, V., & of the glucocorticoid system in aging Zimmerman, J. A. (2004). Tissue glutathione processes and in the action of food and cysteine levels in methionine-restricted restriction. Journal of Gerontology, 46(5), rats. Nutrition, 20(9), 800–805. B171–B179.
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 101Sato, M., Maulik, N., & Das, D. K. (2002). C57BL/6 mice. Mechanisms of Ageing & Cardioprotection with alcohol: role of both Development, 76(2–3), 215–224. alcohol and polyphenolic antioxidants. Sohal, R. S., Ku, H. H., Agarwal, S., Annals of the New York Academy of Forster, M. J., & Lal, H. (1994b). Oxidative Sciences, 957, 122–135. damage, mitochondrial oxidant generationSelman, C., Gredilla, R., Phaneuf, S., and antioxidant defenses during aging and Kendaiah, S., Barja, G., & Leeuwenburgh, C. in response to food restriction in the (2003). Short-term caloric restriction and mouse. Mechanisms of Ageing & regulatory proteins of apoptosis in heart, Development, 74(1–2), 121–133. skeletal muscle and kidney of Fischer 344 Soleas, G. J., Diamandis, E. P., & rats. Biogerontology, 4(3), 141–147. Goldberg, D. M. (1997). Resveratrol: aSeres, J., Stancikova, M., Svik, K., Krsova, D., molecule whose time has come? And gone? & Jurcovicova, J. (2002). Effects of chronic Clinical Biochemistry, 30(2), 91–113. food restriction stress and chronic Sorensen, J. G., & Loeschcke, V. (2001). Larval psychological stress on the development of crowding in Drosophila melanogaster adjuvant arthritis in male long evans rats. induces Hsp70 expression, and leads to Annals of the New York Academy of increased adult longevity and adult thermal Sciences, 966, 315–319. stress resistance. Journal of InsectShaddock, J. G., Feuers, R. J., Chou, M. W., Physiology, 47(11), 1301–1307. Swenson, D. H., & Casciano, D. A. (1995). Southam, C. M., & Ehrlich, J. (1943). Effects Genotoxicity of tacrine in primary of extract of western red-cedar heartwood hepatocytes isolated from B6C3F1 mice on certain wood decaying fungi in culture. and aged ad libitum and calorie restricted Phytopathology, 33, 517–524. Fischer 344 rats. Mutation Research, Spaulding, C. C., Walford, R. L., & Effros, R. B. 344(1–2), 79–88. (1997). Calorie restriction inhibits theShelke, R. R., & Leeuwenburgh, C. (2003). age-related dysregulation of the cytokines Lifelong caloric restriction increases TNF-alpha and IL-6 in C3B10RF1 mice. expression of apoptosis repressor with a Mechanisms of Ageing & Development, caspase recruitment domain (ARC) in the 93(1–3), 87–94. brain. Faseb Journal, 17(3), 494–496. Speakman, J. R., Talbot, D. A., Selman, C.,Sinclair, D. A. (2002). Longevity Genes. Snart, S., McLaren, J. S., Redman, P., In R. Abruzzi (Ed.), The encyclopedia of Krol, E., Jackson, D. M., Johnson, M. S., & aging, Macmillan reference, USA. Brand, M. D. (2004). Uncoupled andSinclair, D. A., & Guarente, L. (1997). surviving: individual mice with high Extrachromosomal rDNA circles—a metabolism have greater mitochondrial cause of aging in yeast. Cell, 91(7), uncoupling and live longer. Aging Cell, 1033–1042. 3(3), 87–95.Sinclair, D. A., & Wood, J. G. (2004). Speakman, J. R., van Acker, A., & Harper, E. J. Unpublished result. (2003). Age-related changes in theSmith, J. S., Brachmann, C. B., Celic, I., metabolism and body composition of three Kenna, M. A., Muhammad, S., Starai, V. J., dog breeds and their relationship to life Avalos, J. L., Escalante-Semerena, J. C., expectancy. Aging Cell, 2(5), 265–275. Grubmeyer, C., Wolberger, C., & Spencer, C. C., Howell, C. E., Wright, A. R., Boeke, J. D. (2000). A phylogenetically & Promislow, D. E. (2003). Testing an conserved NADϩ-dependent protein ‘aging gene’ in long-lived Drosophila deacetylase activity in the Sir2 protein strains: increased longevity depends on sex family. Proceedings of the National and genetic background. Aging Cell, 2(2), Academy of Sciences of the USA, 97(12), 123–130. 6658–6663. Spindler, S. R., Dhahbi, J. M., Mote, P. L.,Sohal, R. S., Agarwal, S., Candas, M., Kim, H. J., & Tshuchiya, T. (2003). Rapid Forster, M. J., & Lal, H. (1994a). Effect of identification of candidate CR mimetics age and caloric restriction on DNA using microarrays. Biogerontology, oxidative damage in different tissues of 4(Supplement 1), 89.
102 D. A. Sinclair and K. T. HowitzSreekumar, R., Unnikrishnan, J., Fu, A., Tavernarakis, N., & Driscoll, M. (2002). Nygren, J., Short, K. R., Schimke, J., Caloric restriction and lifespan: a role for Barazzoni, R., & Nair, K. S. (2002). Effects of protein turnover? Mechanisms of Ageing & caloric restriction on mitochondrial function Development, 123(2–3), 215–229. and gene transcripts in rat muscle. American Tissenbaum, H. A., & Guarente, L. (2001). Journal of Physiology, 283(1), E38–E43. Increased dosage of a sir-2 gene extendsStadtman, E. R. (1995). Role of oxidized amino lifespan in Caenorhabditis elegans. Nature, acids in protein breakdown and stability. 410(6825), 227–230. Methods in Enzymology, 258, 379–393. Trifunovic, A., Wredenberg, A.,Stafford, H. A. (1991). Flavonoid evolution: Falkenberg, M., Spelbrink, J. N., an enzymic approach. Plant Physiology, Rovio, A. T., Bruder, C. E., Bohlooly, Y. M., 96, 680–685. Gidlof, S., Oldfors, A., Wibom, R.,Strahl, B. D., & Allis, C. D. (2000). The Tornell, J., Jacobs, H. T., & Larsson, N. G. language of covalent histone modifications. (2004). Premature ageing in mice expressing Nature, 403(6765), 41–45. defective mitochondrial DNA polymerase.Stuart, J. A., Karahalil, B., Hogue, B. A., Nature, 429(6990), 417–423. Souza-Pinto, N. C., & Bohr, V. A. (2004). Tsuchiya, T., Dhahbi, J. M., Cui, X., Mote, P. L., Mitochondrial and nuclear DNA base Bartke, A., & Spindler, S. R. (2004). Additive excision repair are affected differently by regulation of hepatic gene expression by caloric restriction. Faseb Journal, 18(3), dwarfism and caloric restriction. 595–597. Physiological Genomics, 17(3), 307–315.Sun, D., Muthukumar, A. R., Lawrence, R. A., Turturro, A., Hass, B., & Hart, R. W. (1998). & Fernandes, G. (2001). Effects of calorie Hormesis—implications for risk restriction on polymicrobial peritonitis assessment caloric intake (body weight) as induced by cecum ligation and puncture an exemplar. Human Experimental in young C57BL/6 mice. Clinical and Toxicology, 17(8), 454–459. Diagnostic Laboratory Immunology, 8(5), Turturro, A., Hass, B. S., & Hart, R. W. (2000). 1003–1011. Does caloric restriction induce hormesis?Tanaka, K., Higami, Y., Tsuchiya, T., Human Experimental Toxicology, 19(6), Shiokawa, D., Tanuma, S., Ayabe, H., & 320–329. Shimokawa, I. (2004). Aging increases Turturro, A., Witt, W. W., Lewis, S., Hass, B. S., DNase gamma, an apoptosis-related Lipman, R. D., & Hart, R. W. (1999). endonuclease, in rat liver nuclei: effect of Growth curves and survival characteristics dietary restriction. Experimental of the animals used in the Biomarkers of Gerontology, 39(2), 195–202. Aging Program. Journals of GerontologyTanner, K. G., Landry, J., Sternglanz, R., & Series A Biological Sciences & Medical Denu, J. M. (2000). Silent information Sciences, 54(11), B492–B501. regulator 2 family of NAD- dependent Um, J. H., Kim, S. J., Kim, D. W., Ha, M. Y., histone/protein deacetylases generates a Jang, J. H., Chung, B. S., Kang, C. D., & unique product, 1-O-acetyl-ADP-ribose. Kim, S. H. (2003). Tissue-specific changes Proceedings of the National Academy of DNA repair protein Ku and mtHSP70 in of Sciences of the USA, 97(26), aging rats and their retardation by caloric 14178–14182. restriction. Mechanisms of Ageing &Tanny, J. C., & Moazed, D. (2001). Coupling Development, 124(8–9), 967–975. of histone deacetylation to NAD Undie, A. S., & Friedman, E. (1993). Diet breakdown by the yeast silencing protein restriction prevents aging-induced deficits Sir2: Evidence for acetyl transfer from in brain phosphoinositide metabolism. substrate to an NAD breakdown product. Journal of Gerontology, 48(2), B62–B67. Proceedings of the National Academy of van der Horst, A., Tertoolen, L. G., Sciences of the USA, 98(2), 415–420. de Vries-Smits, L. M., Frye, R. A.,Tatar, M., Bartke, A., & Antebi, A. (2003). The Medema, R. H., & Burgering, B. M. endocrine regulation of aging by insulin-like (2004). FOXO4 is acetylated upon peroxide signals. Science, 299(5611), 1346–1351. stress and deacetylated by the longevity
CHAPTER 3 / Dietary Restriction, Hormesis, and Small Molecule Mimetics 103 protein hSir2(SIRT1). Journal of Biological Wang, H. D., Kazemi-Esfarjani, P., & Benzer, S. Chemistry, 279(28), 28873–28879. (2004). Multiple-stress analysis for isolationVan Remmen, H., Ikeno, Y., Hamilton, M., of Drosophila longevity genes. Proceedings Pahlavani, M., Wolf, N., Thorpe, S. R., of the National Academy of Sciences of the Alderson, N. L., Baynes, J. W., Epstein, C. J., USA, 101(34), 12610–12615. Huang, T. T., Nelson, J., Strong, R., & Ward, W. F. (1988). Food restriction enhances Richardson, A. (2003). Life-long reduction the proteolytic capacity of the aging rat in MnSOD activity results in increased liver. Journal of Gerontology, 43(5), DNA damage and higher incidence of B121–B124. cancer but does not accelerate aging. Weindruch, R. (1996). The retardation of Physiological Genomics, 16(1), 29–37. aging by caloric restriction: studies inVaziri, H., Dessain, S. K., Eaton, E. N., rodents and primates. Toxicologic Imai, S. I., Frye, R. A., Pandita, T. K., Pathology, 24(6), 742–745. Guarente, L., & Weinberg, R. A. Weindruch, R., & Walford, R. L. (1988). (2001). hSIR2(SIRT1) Functions as an The retardation of aging and disease by NAD-Dependent p53 Deacetylase. Cell, dietary restriction. Springfield, IL: Charles 107(2), 149–159. C. Thomas.Vogt, T., & Taylor, L. P. (1995). Flavonol Weindruch, R., Kayo, T., Lee, C. K., & 3-O-glycosyltransferases associated Prolla, T. A. (2001). Microarray profiling with petunia pollen produce of gene expression in aging and its alteration gametophyte-specific flavonol diglycosides. by caloric restriction in mice. Journal of Plant Physiology, 108(3), 903–911. Nutrition, 131(3), 918S–923S.Wachsman, J. T. (1996). The beneficial Weraarchakul, N., Strong, R., Wood, W. G., & effects of dietary restriction: reduced Richardson, A. (1989). The effect of aging oxidative damage and enhanced and dietary restriction on DNA repair. apoptosis. Mutation Research, 350(1), Experimental Cell Research, 181(1), 25–34. 197–204.Walford, R. L., Mock, D., Verdery, R., & Wilkinson, G. S., & South, J. M. (2002). Life MacCallum, T. (2002). Calorie history, ecology and longevity in bats. restriction in biosphere 2: alterations in Aging Cell, 1(2), 124–131. physiologic, hematologic, hormonal, and Will, L. C., & McCay, C. M. (1941). Ageing, biochemical parameters in humans basal metabolism, and retarded growth. restricted for a 2-year period. Journals of Archives of Biochemistry & Biophysics, Gerontology Series A Biological Sciences & 2(481). Medical Sciences, 57(6), B211–B224. Wood, J. G., Rogina, B., Lavu, S., Howitz, K.,Walker, G. A., Thompson, F. J., Brawley, A., Helfand, S. L., Tatar, M., & Sinclair, D. Scanlon, T., & Devaney, E. (2003). Heat (2004). Sirtuin activators mimic caloric shock factor functions at the convergence restriction and delay ageing in metazoans. of the stress response and developmental Nature, 430(7000), 686–689. pathways in Caenorhabditis elegans. Faseb Xia, E., Rao, G., Van Remmen, H., Journal, 17(13), 1960–1962. Heydari, A. R., & Richardson, A. (1995).Wan, R., Camandola, S., & Mattson, M. P. Activities of antioxidant enzymes in (2003). Intermittent fasting and dietary various tissues of male Fischer 344 rats supplementation with 2-deoxy-D-glucose are altered by food restriction. Journal of improve functional and metabolic Nutrition, 125(2), 195–201. cardiovascular risk factors in rats. Faseb Yeung, F., Hoberg, J. E., Ramsey, C. S., Journal, 17(9), 1133–1134. Keller, M. D., Jones, D. R., Frye, R. A., &Wan, R., Camandola, S., & Mattson, M. P. Mayo, M. W. (2004). Modulation of (2004). Dietary supplementation with NF-kappaB-dependent transcription and 2-deoxy-D-glucose improves cardiovascular cell survival by the SIRT1 deacetylase. and neuroendocrine stress adaptation in Embo Journal, 23(12), 2369–2380. rats. American Journal of Physiology, Yoshikawa, H., Ichiki, Y., Sakakibara, K. D., 287(3), H1186–H1193. Tamura, H., & Suiko, M. (2002). The
104 D. A. Sinclair and K. T. Howitz biological and structural similarity between Bulletin of Experimental Biology & lunularic acid and abscisic acid. Bioscience, Medicine, 133(6), 559–561. Biotechnology, & Biochemistry, 66(4), Zainal, T. A., Oberley, T. D., Allison, D. B., 840–846. Szweda, L. I., & Weindruch, R. (2000).Yu, B. P. (1993). Need the free radical theory of Caloric restriction of rhesus monkeys aging be linked to the metabolic rate lowers oxidative damage in skeletal theory? Aging (Milano), 5(3), 243–244. muscle. Faseb Journal, 14(12),Yu, B. P., Masoro, E. J., & McMahan, C. A. 1825–1836. (1985). Nutritional influences on aging of Zhang, Y., & Herman, B. (2002). Ageing and Fischer 344 rats: I. Physical, metabolic, and apoptosis. Mechanisms of Ageing & longevity characteristics. Journal of Development, 123(4), 245–260. Gerontology, 40(6), 657–670. Zhao, K., Chai, X., & Marmorstein, R. (2004).Yuneva, A. O., Kramarenko, G. G., Structure and substrate binding properties Vetreshchak, T. V., Gallant, S., & of cobB, a Sir2 homolog protein deacetylase Boldyrev, A. A. (2002). Effect of carnosine from Escherichia coli. Journal of Molecular on Drosophila melanogaster lifespan. Biology, 337(3), 731–741.
106 D. R. Bell and G. V. Zant from each other during their entire life- time (Owen, 1945). Presumably, the hematopoietic stem cells (HSCs) shared in utero are responsible for this unique proliferation chimerism. Experiments in mice in the early 1960s further demonstrated the exis- differentiationStem cell tence of stem cells in the hematopoietic system. The pioneering work of Till, McCulloch, Becker, and Siminovitch ele- gantly showed that single cells in the bone self-renewal marrow of mice could give rise to myelo- erythroid colonies in the spleens of irradi- ated recipients (Till & McCulloch, 1961; Becker et al., 1963). Moreover, when cells numbers of expressed genes from these splenic colonies were trans-Figure 4.1 Hallmark functions of stem cells. ferred to secondary lethally irradiated recipient mice, all blood cell lineages were reconstituted in these animalsreplaced by a steady supply of stem cells (Siminovitch et al., 1963). These innova-in the animal (Muller, 1996). As animals tive studies determined that in the boneevolved and became more complex, marrow cells of mice resided cells thatwhich coincided with the development of (1) could self-renew to be passaged fromextremely specialized organ systems, the primary to secondary hosts and (2) couldanimal’s longevity far exceeded that of its differentiate into multiple lineages, in thisindividual cells. This process absolutely case into the various blood cell compo-requires a source of cells to maintain nents, to reconstitute a lethally irradiatedhomeostatic balance as the animal faces a animal that was completely dependent onlifetime of injury and normal replenish- the transplanted cells for survival. Thesement of worn cells in the various organ experiments marked the beginning of thesystems. This crucial resource is the stem stem cell biology field.cell pool, and the life spans of these stem Before any discussion about aging incells are either equal to or greater than the stem cell compartment, other signifi-that of the organism as a whole (Harrison, cant properties of the stem cell must be1973). introduced. Plasticity, one of the most Despite current media coverage and important and perhaps most contentiouscontroversies surrounding stem cells and stem cell properties, refers to flexibilitytheir potential use (i.e., embryonic stem in lineage commitment, thereby allow-cells in the treatment of injury and dis- ing a stem cell to cross tissue bordersease) (Holden, 2004; Malakoff, 2004; and seed unrelated tissues and organs,Weissman, 2000b), their existence and even those from different embryonicclues to their importance were recognized layers (Blau et al., 2001). For example,as early as 1945. At this time, a remark- two publications have reported the abil-able phenomenon was noted between non- ity of HSCs to convert to hepatocytesidentical bovine twins. During the rare (Lagasse et al., 2000) or neurons in theoccurrence when these fraternal twins human brain (Mezey et al., 2003). If theshare a common placenta (and therefore concept of plasticity is accurate, thencommon blood circulation), these animals stem cells from a normal, healthy organremain hematopoietic chimeras indefi- could theoretically be used to help regen-nitely and continue to produce blood cells erate or repair another unrelated tissue
CHAPTER 4 / Hematopoietic Stem Cells, Aging, and Cancer 107or organ damaged by disease or age. to become activated as necessary inUnderstandably, this amazing restorative response to some cellular signal, purifiedpotential has been the driving force HSCs express a vast array of genes asbehind the flurry of research in stem cell compared to committed progenitor cellsplasticity. More recent evidence has (Terskikh et al., 2003) (see Figure 4.1). Ofestablished criteria that would defini- the differentially expressed genes exam-tively demonstrate plasticity in stem ined, nearly half were seen in the stemcells; to date, no published reports have cell population, with relatively few genesfulfilled all such criteria. expressed by more committed progeni- To summarize, for a stem cell to be tors. In contrast, a limited but distinctconsidered “plastic” it would have to set of lineage-specific genes was upregu-satisfy the following conditions: (1) as the lated in the progenitor cells. These datacell shifts from one lineage to another, suggest that stem cells are quite activea new lineage-specific function should transcriptionally, allowing them to con-accompany that shift, (2) the genetic pro- stantly assess their environment andfile of the cell should change to fit its respond quickly to the needs of ornew identity, (3) minimal culture or han- changes experienced by the organism.dling of the cell should occur to avoid any Active transcriptional status usuallyextrinsic interference with its develop- means an open chromatin configuration,mental program, (4) no cell-cell fusion which could account for DNA damage orbetween the “plastic” cell and a mature liabilities over time, especially over thecell of the new lineage can occur, and stem cell’s (and the organism’s) long life-(5) these results must be reproducible time. In this regard, active and robust(Wagers & Weissman, 2004). Based on gene expression in a stem cell could bethese strict criteria, no bona fide exam- important during aging.ples of stem cell plasticity exist; indeed, Finally, an additional characteristic ofmost reported cases can be explained by stem cells that potentially could becell-cell fusion (Wagers & Weissman, affected by advancing age is their prolifera-2004), introduction of heterogenous popu- tive capacity. In much the same way thatlations of stem cells into a donor tissue stem cells traditionally were considered(Orkin & Zon, 2002), or the requirement “inactive” in terms of gene expressionof induced or natural organ damage or (which we now know to be untrue), theyimpaired function in the recipient to were further thought to be sedentary as farachieve enhanced performance of the as cell cycle progression is concerned.transplanted stem cells in the desired tis- Again, the stem cell (in particular thesue (Grove et al., 2004). However, infu- HSC) has proven to be much moresion of donor HSCs may contribute to dynamic than once believed. These cellsoverall healing and improved outcome in are far from quiescent; on the contrary, thecertain cases of organ transplant or dis- most primitive stem cells actively cycle.ease, albeit via indirect, unidentified Each day between 8 and 10 percent ofmechanisms. Taken together, plasticity these primitive stem cells progress throughof a stem cell, despite hope and initial the cell cycle, and all cells in the HSC poolpromise, likely will have little or no bear- undergo cell division every one to threeing on the aging of stem cell populations. months, depending on the mouse strain Another fascinating characteristic of studied (Bradford et al., 1997; Cheshierstem cells (HSCs in particular) is the et al., 1999). Not only does this implicatewide range of genes they express. the stem cell as an active entity, it intro-Counterintuitive to the prevailing belief duces additional mechanisms for DNAof stem cells as “resting” and waiting damage (mismatches, open chromatin
108 D. R. Bell and G. V. Zantstructure) that could accumulate during rescue a lethally irradiated recipientthe lifetime of the cell. Cell proliferation, mouse. Classic limiting dilution assaystherefore, is another factor to be considered by Smith and colleagues show that singleduring the aging of a stem cell. HSCs do indeed generate multilineage clones capable of long-term self-renewal (Smith et al., 1991). HSCs differentiate inB. Hematopoietic Stem Cells a hierarchical manner, with the mostMany of the preceding descriptions about primitive stem cells, or long-term HSCs,stem cell biology and characteristics cite retaining the ability to self-renew indefi-the hematopoietic system as an example. nitely (see Figure 4.2). These, in turn, giveHSCs are likely the most highly studied rise to short-term HSCs that self renewstem cells to date because of their early in approximately eight weeks and multi-discovery (Till, 1961), their isolation from potent progenitors that can self-renew inboth mice and humans (Baum et al., 1992; less than two weeks. At this point, line-Spangrude et al., 1988), and their clinical age-restricted progenitors emerge, whichapplications, such as the critical compo- ultimately give rise to all of the differen-nents in bone marrow transplantation tiated cells of the hematopoietic system(Reya et al., 2001). Because of the exten- (Reya et al., 2001). As seen in Figure 4.2,sive literature dedicated to the HSC, this as differentiation increases, the ability tostem cell population will be the focus of self renew decreases.this chapter. The HSC displays all the hallmarks of astem cell, including the capacity to self- II. Stem Cell Agingrenew and differentiate into all necessaryblood cell lineages. This is clearly demon- Stem cells are uniquely sensitive to dam-strated in the ability of a single trans- age accumulation from both intrinsicplanted stem cell to reconstitute and and extrinsic sources because they are common lymphoid progenitor Hematopoietic Stem Cells B cells lymphoid cells T cells multipotent Natural killer cells progenitors long-term short-term HSC HSC Life-long ~8 weeks <2 weeks Macrophages Neutrophils Self-renewal capacity myeloid cells Granulocytes Erythrocytes common Platelets myeloid Degree of differentiation progenitor *adapted from Reya, Morrison, Clarke, and Weissman, 2001Figure 4.2 Self-renewal and differentiation in the stem cell hierarchy.
CHAPTER 4 / Hematopoietic Stem Cells, Aging, and Cancer 109long-lived. One of the main perpetrators serious consequences for that individualof damage is reactive oxygen species cell, mutation in a stem cell could have(ROS) that are generated by normal cellu- potentially disastrous effects becauselar metabolism. Long-term exposure to this genome will be perpetuated as theROS is detrimental to many macromole- progeny of that stem cell differentiates.cules, including proteins, lipids, and Although stem cells from other tissuesnucleic acids. Whereas damage to any of have been identified (i.e., in the centralthese molecules may affect cellular func- nervous system, liver, and skin), muchtion, harm to the nucleic acids and pro- less is known about them, and virtuallyteins that constitutes chromatin in stem no data concerning their properties dur-cells is particularly deleterious because it ing aging have been published. The studycan be passed on to progeny (Harman, of stem cell aging in systems other than1981; Johnson et al., 1999). Another the hematopoietic is in its infancy.source of damage to stem cells is associ- Likewise, data will be presented describ-ated with their lifetime of cell division. ing the functional consequences of agingAs mentioned in the previous section, on HSCs. Factors such as replicativestem cells are not quiescent, and enter stress and editing errors that occur in allthe cell cycle on a regular basis. Repeated cells, including the long-lived stem cell,rounds of DNA replication can result in are important in limiting organismalthe incorporation of numerous copy longevity and hence contribute to theerrors into the genome. Under normal decline of the individual with age.circumstances, the cell has sophisticatedproofreading and editing enzymes to cor- A. Stem Cell Aging Theoriesrect these potentially hazardous mis-takes. However, over a lifetime of use, Various hypotheses have been put forth tothe function of these enzymes may explain stem cell usage as an organismdecline due to ROS exposure, for exam- ages. The first was proposed nearlyple (Johnson et al., 1999), and lead to 40 years ago and is referred to as themutation. Whereas mutation in a single clonal succession theory (see Figure 4.3a).long-lived differentiated cell may have This theory suggests that the stem cell A. Clonal succession theory: total # of stem cells B. ‘‘Equal contribution’’ theory: total # of decreases over time stem cells remains constant over time single “inactive” differentiated “active” differentiated “activated” stem cell pool progeny stem cell pool progeny stem cellFigure 4.3 Theories of stem cell usage.
110 D. R. Bell and G. V. Zantpool as a whole is maintained in a quies- renewal properties of the stem cells.cent state, held in check until needed by Therefore, depletion of stem cell num-the organism. When required, one or at bers does not occur and hence does notmost only a few stem cells become active drive aging. This model agrees with dataand leave the pool to proliferate and sup- from mouse chimera experiments thatply the necessary differentiated cells to show equal contribution by most (per-the organism (Kay, 1965). According to haps all) HSCs to blood formation simul-the tenets of this theory, the activated taneously (Harrison et al., 1987).clone would never rejoin the primitive In view of this model, advancing agepool of resting stem cells; therefore, over poses a special problem with regard totime, the supply of stem cells would stem cells: does the ability to self-renewdiminish, and aging would proceed. Injury prevent aging of this critical cell popula-or disease could accelerate this stem cell tion? Certain limited evidence exists thatdiminution and thereby contribute to seems to argue against stem cell aging.aging. To counteract this loss of critical For example, experiments analyzing stemstem cells with time, one could postulate cell properties in a large animal modelthat a sufficiently large pool of stem cells (the dog) show little age-related differ-would be present at birth to meet the ences in the HSC populations in youngneeds of the organism during periods of versus old bone marrow (Zaucha et al.,crisis or a lifetime of wear and tear, or 2001). Additionally, although not directlyboth. In this regard, a mouse with an aver- related to the self-renewal properties ofage life span of approximately 2 years stem cells, is the observation that manywould require far fewer stem cells than common diseases responsible for deathhumans with their life expectancies in mammals, including humans, are notof nearly 80 years. However, one study diseases of stem cell origin, such as heartunexpectedly showed that HSC numbers disease or renal failure (Van Zant &are relatively conserved in mammals Liang, 2003). However, greater evidence(Abkowitz et al., 2002). That is, the total is mounting in favor of age-relatednumber of HSCs in the bone marrow of changes that affect stem cell properties,mice, cats, and humans was similar and especially in the hematopoietic system.not commensurate with either the size orthe potential life span of the organism.These data argue against a model of B. Mouse Aging Datasequential activation and subsequent 1. Intrinsic Factorsdepletion of a non-renewable stem cellpopulation. Because of the similarities in the blood- An alternative theory that takes into forming processes between mice andaccount the compelling data that demon- humans and the availability of numerousstrate that HSCs actively cycle and are inbred laboratory mouse strains, much ofnot quiescent (Bradford et al., 1997; the data on stem cells and aging generatedCheshier et al., 1999) can be termed the to date has been in the mouse. Several“equal contribution” model. In this case, general characteristics have been deter-all members of the primitive stem cell mined with regard to stem cell aging. Topool are active and in a homeostatic bal- begin, in the murine hematopoietic sys-ance of self-renewal, proliferation, or dif- tem, HSCs are not immortal and can onlyferentiation (see Figure 4.3b). This theory be serially passaged to recipient mice forsuggests that the total number of stem five generations at most (Ogden &cells does not decrease during the life- Mickliem, 1976; Siminovitch et al., 1964).time of the animal due to the self- Subsequent transplants and manipulations
CHAPTER 4 / Hematopoietic Stem Cells, Aging, and Cancer 111do not fully reconstitute and rescue cells are not immune to the inevitable pas-lethally irradiated recipient mice and sug- sage of time.gest that repeated rounds of self-renewal In addition to these studies thatand lineage replenishment exert a lasting directly examine the effects of aging oneffect on the stem cells. Although these stem cells, many investigators haveserial transplantations may not precisely exploited the intrinsic differences amongrecapitulate the natural aging process and inbred mouse strains to look at age-may further be magnified by differentia- related alterations in HSCs. One of thetion of the transplanted stem cells as well, best examples of these experimentssimilar mechanisms are involved and involves analyses of allophenic mice,hence can provide clues as to what is hap- which provide hints about naturallypening as the mouse ages. Another stem occurring differences in hematopoieticcell characteristic affected by advancing cells and their life spans between twoage is the ability to “home.” Stem cells inbred strains of mice. These innovativemust be able to find their way from one experiments involve creating chimericsite in the bone marrow, enter and travel animals by aggregating cells from twothrough the bloodstream, and finally rec- different strains of mice (in this case, theognize other positions in the bone marrow commonly used C57BL/6 [B6] and DBA/2where they are needed. The same is true of [D2] mice) at the embryonic stage. Thetransplanted stem cells: they must be able resulting animals are composed of cellsto locate appropriate locations in the bone from both B6 and D2 mice. Amazingly,marrow to lodge from the bloodstream fol- stem cells derived from the D2 mice,lowing transplantation (Cao et al., 2004). which are short-lived (by 34 percent)This process of traveling to and recogniz- compared to B6 mice, stopped contribut-ing the correct spot to reside by a stem cell ing to hematopoiesis at a time correspond-is called homing, and diminished homing ing to their normal life span. After thisefficiency of HSCs in mice is observed time, all hematopoietic cells in thewith regard to age. Indeed, old stem cells allophenic mice came from the longer-have a significantly reduced ability to lived B6 mice (Van Zant et al., 1990).home to the proper location as compared This convincingly demonstrates thatto young stem cells following transplanta- HSCs do indeed age and that there aretion (Morrison et al., 1996). Finally, other functional consequences of this aging.observations between young and old mice Adding to this data are observations fromhave shown that changes occur in the other investigators who use inbreddevelopmental potential in aged stem cells strains of laboratory mice to show wideof hematopoietic origin. Specifically, as variations in HSC characteristics andmurine bone marrow is serially trans- natural life span (Van Zant et al., 1983).planted (as mentioned before, this is a Specifically, many researchers have usedprocess likened, but not identical, to recombinant mouse strains to identifyaging), or when bone marrow from aged quantitative trait loci (QTLs) responsibledonors is used in transplantation experi- for the complex traits (including aging)ments, stem cells lose their ability to exhibited by different mouse strainsdifferentiate into the lymphoid lineage. (Abiola et al., 2003). For example, studiesHence, hematopoietic maturation is from this laboratory have shown that oldskewed toward myeloid precursors and mice (24 months old) have greater num-away from the B and T cell lineages bers of primitive HSCs compared to(Spangrude et al., 1995; Sudo et al., 2000). young mice (6 weeks old), but the totalAll of these data indicate that despite their number of HSCs per animal is strain-ability to continually self-renew, stem dependent. Five commonly used mouse
112 D. R. Bell and G. V. Zantstrains were assayed in these experi- lines (called BXD strains) identified otherments (C3H/He, CBA/J, DBA/2, BALB/c, QTLs important in HSC biology andand C57BL/6). A decrease in stem cell aging. Analysis of all 35 BXD strains ofcycling activity in older mice of each mice available to determine QTLsstrain tested was demonstrated, and this responsible for the changes in stemdecrease had a statistically significant cell numbers as mice age from 2 tonegative correlation with the maximal 20 months identified several potentialexpected life span of that particular contributing loci, including those foundstrain. However, the degree of prolifera- on chromosomes 2, 14, and X (de Haan &tive decrease was also strain-dependent Van Zant, 1999b). When thinking about(de Haan et al., 1997). These results were how vital and complex stem cell regula-further investigated in light of extrinsic tion and maintenance is, it is not surpris-(such as growth factor) controls on HSC ing that several chromosomal locationsproliferation and pool size. In this study, could be responsible for this regulation.the effects of stimulation by the growth Using newer and more reliable databasefactor flk-2/flt-3 ligand (FL), which has resources with higher resolution, thebeen proposed to be important in main- potential locus on mouse chromosome 2taining HSC numbers and regulating can be considered a bona fide “agingtheir proliferation, were assessed in the locus” because in B6 mice this sitestem cell compartment of the same five appears to be responsible for themouse strains analyzed previously. increased numbers of HSCs noted in oldInterestingly, following incubation with B6 mice (Geiger et al., 2001). Senescence,FL, a correlation was noted among the or the limitation of proliferative capacityfollowing characteristics: life span of the due to exhaustive rounds of cellularvarious mouse strains, HSC proliferation, cycling, is also genetically regulatedand HSC pool size. Specifically, FL only and age-related in HSCs (Chen et al.,elicited a stimulatory effect on stem cell 2000), as evidenced by both competitivecycling in strains of mice with a natu- repopulation and serial transplantationrally larger pool of stem cells, such as the studies in mice in vivo. To be precise, B6D2 mice, which are also relatively short- mice (longer-lived) have a delayed onsetlived. Using the power of recombinant of senescence compared to D2 miceinbred mice and QTL mapping, a puta- (shorter-lived). Decreased repopulatingtive region on mouse chromosome 18 ability with increasing HSC age in D2was identified as responsible for this mice suggests that these cells do senesce;trait. In addition, this region of mouse however, HSCs from old B6 mice repopu-chromosome 18 shows synteny with late recipients better than do those fromhuman chromosome 5q, deletions of young B6 mice, an odd contradiction.which are important in hematologic can- This could be explained if B6 stem cellscers (de Haan & Van Zant, 1997). Further simply have a higher proliferative limitanalysis of HSC cell turnover and cycling than D2 stem cells. If this is true, thenin recombinant inbred mice had a dra- only repeated serial transplantationsmatic association with mean life span in requiring extensive proliferation wouldmice, and two additional QTLs on mouse uncover functional deterioration (i.e.,chromosomes 7 and 11 were identified as loss of repopulating ability) over timepossible loci for the genes responsible for in B6 stem cells (Chen et al., 2000).these traits (de Haan & Van Zant, 1999a). As stated earlier in this section, thisDetailed experiments entirely focused on is indeed the case: serial transplanta-the informative recombinant inbred mice tion recovery by HSCs is affectedgenerated by crossing the B6 and D2 with increasing age. Initial experiments
CHAPTER 4 / Hematopoietic Stem Cells, Aging, and Cancer 113indicate that a potential locus on mouse as a stem cell niche? Or, is it more likelychromosome 12 may be responsible for that other cells in the bone marrow, suchthe senescent phenotype. In all of these as the mesenchymal cells (i.e., endothelialreports, no specific candidate genes have cells), also are part of the HSC niche andbeen identified in the putative QTLs their association with osteoblasts trulyshown to be responsible for the suggested defines the niche? Data supporting thistraits. Analysis of candidate genes in model were shown by analyzing thethese loci is an area of active research by expression of certain HSC-specific recog-many investigators. nition molecules (i.e., c-Kit ligand) on the surface of osteoblasts; no such expression was detected, thereby suggesting that2. Extrinsic Factors other accessory cells that do express suchWhile most data published on the aging of necessary factors must be associated withHSCs focus on intrinsic factors, such as the osteoblasts to allow HSCs to home togenetic components specific to each cell, their proper niche (Taichman et al., 1996).an important related topic concerns cer- These data lead to an important pointtain extrinsic factors, such as the cellular regarding aging: both the endothelial cellsmicroenvironment, and how this crucial of the bone marrow and the primitivemilieu of cells, cytokines, and soluble HSCs develop from a common hematopoi-growth factors change and affect HSCs etic precursor, the hemangioblast (Zhu &with advancing age. The supporting stro- Emerson, 2004). It is therefore highlymal cells of the microenvironment help likely that these accessory cells of theto maintain HSCs by supporting their stroma also will show similar age-relatedlocalization, survival, or self-renewal, or changes that will influence the biologyany combination thereof. This concept of and chemistry of the microenvironmentan HSC niche was first proposed many (Wineman et al., 1996). Finally, to furtheryears ago (Schofield, 1978) and has since address the role of the aged environmentbeen expanded and intensely studied to and its potential impact on HSC functionprovide a more complete understanding of in mice, bone marrow transplantationstem cells to enhance their manipulation studies were performed in young and oldand applications in treating disease. recipients. Regardless of the age of the The bone-forming cells of the bone bone marrow donor, an increased autoim-marrow, or osteoblasts, now appear to be mune response in the old recipients wascritically important factors in the stem detected compared to the young (Doriacell niche (Zhu & Emerson, 2004). et al., 1997). This supports the notion ofExperiments examining hematopoietic age-related impairment of the stroma that,recovery in mice following treatment with in turn, causes improper functioning of5-fluorouracil have shown in vivo that the immune system. Thus, the microenvi-stimulated HSCs are found in direct prox- ronment of the aging HSC is anotherimity to osteoblasts (Heissig et al., 2002). important factor to consider during theHowever, an unsettling contradiction was process of cancer development, which willnoted with regard to the numbers of be discussed in Section III.osteoblasts present in the bone marrowand the numbers of HSCs: osteoblasts C. Human Aging Datavastly outnumber the HSCs (Taichman etal., 1996). This raises the question of how While data from mouse models of stemdo HSCs “choose” with which osteoblast cell aging are considerably more thor-to associate. Are certain osteoblasts some- ough, reports demonstrating similar age-how different and specialized to function related functional decline in human
114 D. R. Bell and G. V. ZantHSCs are compelling. First, studies ana- samples from the elderly, indicating alyzing the lengths of telomeres (which are potential link between hematopoieticspecialized structures on the ends of deficiencies or aging, or both, with cancerchromosomes) in both human somatic (Marley et al., 1999). Also, as seen incells and HSCs have shown progressive murine HSCs, differentiation potential istelomere shortening after cellular prolif- affected in stem cells from the elderly.eration, which leads to termination of Primitive human CD34 ϩ cells have acell division or replicative senescence. diminished capacity to generate T cellsAccumulation of senescent cells with when examined in vitro (Offner et al.,repeated cellular cycling, as occurs over 1999). This last example of reduced T andtime with progressing age, is theorized to B cell number and activity with advanc-contribute to tissue deterioration (Baird ing age is related to the well-documented& Kipling, 2004). Unlike human somatic decrease in immune system function incells, which express no (or very low levels older individuals (Globerson & Effros,of) telomerase (the enzyme responsible 2000; Miller, 2000; Miller et al., 1997),for maintaining telomere length), human and this could clearly affect an organism’sHSCs do express detectable levels of ability to achieve long life. Impairedtelomerase. Despite this fact, HSCs also T cell monitoring and destruction ofdisplay shortening of telomeres with cancerous cells is highly affected in theage (Allsopp & Weissman, 2002). A recent elderly (Aspinall, 2000). Therefore, thisinteresting observation in cultured could represent a direct link between dys-human cells also lends further evidence functional stem cell differentiation, lossto the importance of age-related telomere of immune prowess, and increased cancershortening in somatic and presumably development due to advancing age.HSCs. As mentioned earlier in this dis-cussion, oxidative stress contributes to D. Health Impact of Agingcellular aging. It is known that oxidative Hematopoietic Stem Cellsstress can also increase the rate of telom-ere shortening (Kawanishi & Oikawa, As humans age, small, seemingly innocu-2004). Thus, oxidative stress can directly ous changes in hematopoiesis occur,affect chromosomes and lead to cellular which include anemia, diminishedsenescence and aging in many human immune responses (discussed above), andcells, including HSCs. Next, studies an overall decrease in bone marrow cellu-using primitive hematopoietic cells larity (Lipschitz et al., 1981). Althoughobtained from human umbilical cord these may be tolerated sufficiently in oth-blood and bone marrow from both adults erwise healthy individuals, times ofand the elderly show a steady decline in hematopoietic stress (such as that follow-cell function that begins after birth and ing myeloablative chemotherapeuticprogresses throughout life. Specifically, treatment or infection) exacerbate theseincreases in progenitor cell numbers with subtle changes. Not only is response toadvancing age were demonstrated in the such stress slowed in the elderly, butsamples tested, which correlates with complete recovery is unusual even afterdata obtained in mice; however, there significant periods of time, resulting inwas a concomitant decrease in prolifera- anemias or other hematopoietic deficien-tive capacity in those same progenitor cies (Botnick et al., 1982; Kim et al., 2003;cells. Samples from children with bone Marley et al., 1999). With regard tomarrow failure and a propensity to hematopoietic changes in the elderly, ane-develop leukemia also showed compara- mia is common, and nearly one-third ofble proliferative deficiencies, as did the the cases seen are unexplained (Guralnik
CHAPTER 4 / Hematopoietic Stem Cells, Aging, and Cancer 115et al., 2004). As the mean age of the popu- atic, lung, prostate, and colon cancer, butlation increases, so does the percentage of also to cancers of the hematopoietic sys-accompanying anemias (Cesari et al., tem, namely leukemias. For example,2004; Penninx et al., 2003). In fact, by age acute myeloid leukemia (AML), for85, the prevalence of anemia is greater which the majority of data is available, isthan 20 percent (Guralnik et al., 2004). more than three times as likely to occurAs mentioned earlier, this usually does in someone age 65 versus an individualnot present complications in a healthy who is 35 years old (NCI, 2000).patient, and most elderly anemic patients Furthermore, the median age at diagnosisgo untreated; however, even mildly ane- of patients with AML is 65 yearsmic patients can have unfavorable out- (Lowenberg et al., 1999). Clearly, this is acomes following heart complications such disease that predominates in older per-as myocardial infarction. The decline in sons. In addition to the prevalence ofhemoglobin in older individuals is not AML in older individuals is their poorsufficient to explain the extent of anemia response to treatment. Following stan-observed. Dysregulated red blood forma- dard chemotherapy treatment, an initialtion in the bone marrow is a plausible cul- remission is seen in 65 to 75 percent ofprit of this anemia. Ultimately, the HSC younger AML patients (Ͻ60 years old).is responsible for erythropoiesis, and Unfortunately, the same standard treat-because other developmental abnormali- ment protocols in older AML patientsties are seen in aging stem cells (such as (Ͼ60 years) result in a much lower remis-reduced T and B cell production), it is pos- sion rate (30 to 50 percent) and a deathsible the aging HSC also fails to produce rate that is twice as high (20 percent) asthe number of red blood cells necessary to that observed in young persons withprevent anemia. This anemia could there- AML (Baudard et al., 1994; Bishop et al.,fore reflect an underlying weak myelodys- 1996; Estey et al., 1995; Leith et al., 1997;plasia; indeed, evidence suggests that Mayer et al., 1994; Rowe et al.,myelodysplastic syndromes are stem cell 1995; Schiller & Lee, 1997; Stone et al.,disorders (Liesveld et al., 2004). Age-asso- 1995; Taylor et al., 1995). A sustain-ciated changes in function and accumula- able remission is achieved in only 10tion of damage, particularly in chromatin, to 20 percent of elderly AML patientsin HSCs may account for the dramatic overall (Goldstone et al., 2001). As anincrease in cancer seen after the age of 65. increasing proportion of our population gets older due to modern medical advances, health challenges of the elderly III. Stem Cells and Cancer will become even more critical to sci- ence and medicine. The statistics citedProgressing age is the most significant above demonstrate that AML will be anrisk factor in cancer development; important concern in an ever-growingindeed, between the ages of 40 to 80 segment of the population.years, there is an exponential increase incancer incidence rates, after which the A. Stem Cell Origin of Leukemiaincidence of cancer levels off (DePinho,2000). Ultimately, the overall risk of Cancers of the hematopoietic system, ordeveloping an invasive cancer in one’s leukemias, originate from and are sus-lifetime is 1 in 3 for women and 1 in 2 tained by cancer stem cells. Leukemiasfor men (ACS, 2000). These statistics not provide the most convincing data thatonly apply to the solid tumors that com- normal stem cells can become cancermonly are cited today, such as pancre- stem cells through the acquisition of
116 D. R. Bell and G. V. Zantmutations, and it is this leukemic stem immunophenotypes, even though each ofcell (LSC) that is responsible for the dis- these subtypes of leukemia displays a dif-ease (Reya et al., 2001). Studies of human ferent clinical disease (Guzman &AML have begun to define the genetic Jordan, 2004). While the phenotypicallyalterations and chromosome transloca- identical LSCs (CD34 ϩ /CD38-) fromtions associated with the various sub- each of the AML subtypes when trans-types of the disease (Dash & Gilliland, planted into NOD/SCID mice contained2001). AML is a heterogeneous disease the leukemia-initiating cells, the result-clinically and includes many subtypes ing disease was similar to that observed(M0–M7) as defined by the French- in the donor patient (Wang et al., 1998).American-British (FAB) classification These results indicate that the initialsystem (Mirro, 1992). Despite this het- transformed cell was a primitive stemerogeneity, the leukemias are quite simi- cell that took the wrong developmentallar at the level of the stem cell and share pathway, presumably depending on themany of the same immunophenotypic mutation that occurred. All of these datamarkers as normal HSCs (Bonnet & point to a transformed stem cell sourceDick, 1997; Lapidot et al., 1994). With of leukemia.the advent of improved methods ofdetection and isolation of stem cells B. The Two-Hit Model(such as fluorescence activated cell sort-ing), it is now possible to analyze the An important corollary to this discussionfunctional and molecular characteristics of a stem cell origin of AML is the theoryof both normal HSCs and LSCs. Recent that cancer, in general, results from areports formally have described the exis- series of genetic changes that, over time,tence of leukemic stem cells in AML and confer unique properties to a cell thatstrongly suggest that these LSCs give rise lead to a progression from normalcyto the disease (Jordan, 2002). Specifically, to malignancy (Hanahan & Weinberg,when CD34 ϩ /CD38-cells, previously 2000). These genetic alterations include,defined as containing the stem cell popu- but are not limited to, independence inlation of normal bone marrow, were iso- growth signaling, escape from apoptosis,lated from the bone marrow of patients and endless ability to replicate. Likewise,with AML and transplanted into the non- AML progression is theorized to resultobese diabetic/severe combined immun- from a stepwise accumulation of geneticodeficient (NOD/SCID) mouse, leukemic mutations. This buildup of mutationsdisease was evident and transferable takes place over the life span of an indi-(Bonnet & Dick, 1997). This indicates vidual, which is why AML is more thanthat, just like the normal HSC, the three times as likely to occur in someoneleukemic stem cell also displays the cell aged 65 years versus an individual whosurface marker phenotype of CD34 ϩ is 35 years old (NCI, 2000). The muta-/CD38-. Other evidence supporting the tion buildup generally results fromstem cell origin of AML comes from two sources: (1) replicative stress andstudies on the various subtypes of the (2) exposure to damaging extrinsic fac-disease. As mentioned previously, AML tors. Repeated rounds of cell division, assubtype classification is based on the occurs in a long-lived stem cell, providemorphologies and genetic abnormalities ample opportunity for mistakes such asof the leukemic cells (FAB subtypes) DNA mismatches and other editing(Brendel & Neubauer, 2000). Interestingly, errors to be incorporated into thethe AML subtypes M0, M1, M2, M4, and genome. These types of errors can beM5 all contain LSCs with similar directly responsible for alterations in
CHAPTER 4 / Hematopoietic Stem Cells, Aging, and Cancer 117gene regulation or function, or both. porting evidence for this hypothesis.Additionally, exposure to repeated Cases of CML patients with BCR/ABL-extrinsic insults, such as ROS, can lead positive disease progressing to acuteto mutation accumulation in crucial leukemia with ensuing acquisition ofcells like HSCs. As mentioned earlier, either the Nup98/HoxA9 or AML1/ETOthis will obviously have serious conse- translocations are documented (Golubquences in a long-lived, terminally differ- et al., 1994; Yamamoto et al., 2000). Theseentiated cell, but it will likely have a dire data point to the need for multiple cooper-outcome in a stem cell responsible for ating mutations during leukemogenesis.life-long tissue repair and replenishment. Given the differences in life spanIn summary, given that the origin of between mice (3 years) and humans (ϳ80AML is an HSC that has undergone years), it is logical that development of aoncogenic mutations, it is reasonable to “second hit” would occur more quicklypropose that an aging stem cell is a likely in the mouse. In fact, most processestarget for this leukemic transformation. occur faster in mice compared to humans Acute myeloid leukemias are composed (i.e., metabolism, reproduction, etc.).of leukemic cells that not only have pro- Additionally, the most common cancersliferation or survival advantages, or both, that affect elderly mice and humans areover other cells of the hematopoietic not the same and therefore are likely tosystem, but also have diminished, poor develop and progress in different ways.differentiation compared to normal cells Specifically, old mice tend to develop sar-(Deguchi & Gilliland, 2002). The current comas and lymphomas, whereas elderly“two-hit” model of AML cites mutation humans more commonly develop epithe-in a cellular kinase that results in a con- lial cancers such as breast and colonstant growth-promoting signal, and muta- cancers (DePinho, 2000). Consideringtion in a hematopoietic transcription all the variance between mouse andfactor that leads to disrupted develop- human physiology, it is not surprisingmental potential as the important ele- that development of mutations necessaryments driving leukemogenesis (Dash & to transform cells occur at different rates.Gilliland, 2001; Deguchi & Gilliland,2002). Clinically, support for the acquisi- C. Age as the “Second Hit” in Cancertion of cooperating mutations in thedevelopment of AML is well documented. The work summarized in this section cor-In nearly all cases of chronic myeloid responds strongly with a model of multi-leukemia (CML), a translocation exists step leukemogenesis, in which the mainthat constitutively activates a tyrosine culprits of disease consist of an alteredkinase. Common translocations asso- transcription factor that causes a differenti-ciated with CML include BCR/ABL, ation block, and of an activated kinase thatTEL/ABL, TEL/PDGF␤R, and TEL/JAK2 can provide limitless growth and survival(Deguchi & Gilliland, 2002). Expression of signals. The combination of these events isthese constitutively active kinases results critical to disease development. But, asin increased proliferation or survival of evidenced by the increase in leukemiaaffected cells without a block in cellular incidence with advancing age, these detri-differentiation. It is the subsequent gain mental mutations obviously take placeof mutations involving transcription over a long period of time. Aging, there-factors that provides the differentiation fore, must also be a key player in this sce-block and causes the onset of acute nario. Because stem cells are known to bedisease. Examples of CML progression to the source of AML and the effects of agingAML (or CML blast crisis) provide sup- on the stem cell population, we favor a
118 D. R. Bell and G. V. Zantmodel of AML in which aging may be con- display decreased developmental potential,sidered a secondary age-related event in which could be similar to the effects of aleukemogenesis (see Figure 4.4). However, mutated transcription factor and skew thethe models in Figure 4.4 (A and B) are not differentiation of progeny. Furthermore,necessarily mutually exclusive. For the decreased monitoring by the immune sys-sake of simplicity, models A and B were tem, another issue in older individuals,shown as separate events. But secondary may allow disease to progress more effi-mutations (A) that help drive leukemogen- ciently and aggressively. Finally, theesis may be direct results of the age-related increased cellular senescence seen withchanges brought about by increasing age (B). advancing age in mice could directly con-Effects on DNA repair mechanisms and tribute to leukemic progression. Krtolicarepeated exposure to environmental insults, et al. (2001) showed that senescent fibrob-both internal and external, may cause lasts can permit epithelial cells to becomeenough genomic instability to permit the malignant. Given that HSCs and theformation and escape from detection of the surrounding stromal cells do age, senes-translocations and point mutations/dupli- cence and cancer development may becations commonly seen in human AMLs. inevitable outcomes of this process.In this regard, aging would provide and pro- Therefore, age may be just as important amote conditions permissive to leukemic factor in AML development as the alteredtransformation. Aging stem cells also genes themselves. A. Current ‘‘Two–Hit’’ theory of leukemogenesis initial mutation secondary mutation leukemia normal HSCs transcription factor tyrosine kinase B. Age as a secondary event in the development of leukemia secondary initial mutation age-related event Long life * exposure to insults: ROS, DNA damage decreases in: normal HSCs ? homing of stem cells developmental potential leukemia Lon g lif immune system function e microenvironment changes * age-related eventsFigure 4.4 Role of stem cell aging in development of leukemia.
CHAPTER 4 / Hematopoietic Stem Cells, Aging, and Cancer 119 IV. Conclusions Peirce, J. L., Pomp, D., Rosemann, M., Rosen, G. D., Schalkwyk, L. C., Seltzer, Z.,The study of stem cells, with their prom- Settle, S., Shimomura, K., Shou, S., Sikela, J. M., Siracusa, L. D., Spearow, J. L.,ise of curing injury and disease, has led to Teuscher, C., Threadgill, D. W., Toth, L. A.,discoveries that necessitate monitoring Toye, A. A., Vadasz, C., Van Zant, G.,and caution as well as those that inspire Wakeland, E., Williams, R. W., Zhang, H. G.,excitement and hope. The evidence pre- & Zou, F. (2003). The nature andsented here not only should entice fur- identification of quantitative trait loci:ther research into the incredible potential a community’s view. Nature reviews.of stem cells but should also cause reflec- Genetics, 4, 911–916.tion about the inevitabilities of being Abkowitz, J. L., Catlin, S. N., McCallie, M. T.,such a long-lived cell. The natural course & Guttorp, P. (2002). Evidence that theof aging and its undeniable association number of hematopoietic stem cells perwith cancer is not lost in a stem cell animal is conserved in mammals. Blood, 100, 2665–2667.despite its amazing ability to self-renew ACS. (2000). Cancer facts and figures. Annualand restore (some) tissues in times of Report of the American Cancer Society, 1–7.need. The identification of malignant Allsopp, R. C., & Weissman, I. L. (2002).stem cells in cancers other than the Replicative senescence of hematopoietichematopoietic system, such as the breast stem cells during serial transplantation:and brain, provides increasing evidence does telomere shortening play a role?consistent with a general model of stem Oncogene, 21, 3270–3273.cell origin of disease. To the extent that Aspinall, R. (2000). Longevity and the immuneat least some, and perhaps most, malig- response. Biogerontology, 1, 273–278.nancies are stem cell-derived, cancer may Baird, D. M., & Kipling, D. (2004). The extentprovide some of the strongest evidence and significance of telomere loss with age. Annals of the New York Academy ofyet supporting the concept of stem cell Sciences, 1019, 265–268.aging. Baudard, M., Marie, J. P., Cadiou, M.,Viguie, F., &, Zittoun, R. (1994). Acute myelogenous leukemia in the elderly: retrospective studyReferences of 235 consecutive patients. British Journal ofAbiola, O., Angel, J. M., Avner, P., Haematology, 86, 82–91. Bachmanov, A. A., Belknap, J. K., Bennett, B., Baum, C. M., Weissman, I. L., Tsukamoto, A. S., Blankenhorn, E. P., Blizard, D. A., Bolivar, V., Buckle, A. M., & Peault, B. (1992). Isolation Brockmann, G. A., Buck, K. J., Bureau, J. F., of a candidate human hematopoietic stem- Casley, W. L., Chesler, E. J., Cheverud, J. M., cell population. Proceedings of the Churchill, G. A., Cook, M., Crabbe, J. C., National Academy of Sciences of the USA, Crusio, W. E., Darvasi, A., de Haan, G., 89, 2804–2808. Dermant, P., Doerge, R. W., Elliot, R. W., Becker, A. J., McCulloch, E. A., & Till, J. E. Farber, C. R., Flaherty, L., Flint, J., (1963). Cytological demonstration of the Gershenfeld, H., Gibson, J. P., Gu, J., clonal nature of spleen colonies derived Gu. W., Himmelbauer, H., Hitzemann, R., from transplanted mouse marrow cells. Hsu, H. C., Hunter, K., Iraqi, F. F., Nature, 197, 452–454. Jansen, R. C., Johnson, T. E., Jones, B. C., Bishop, J. F., Matthews, J. P., Young, G. A., Kempermann, G., Lammert, F., Lu, L., Szer, J., Gillett, A., Joshua, D., Bradstock, K., Manly, K. F., Matthews, D. B., Medrano, J. F., Enno, A., Wolf, M. M., Fox, R., Cobcroft, R., Mehrabian, M., Mittlemann, G., Herrmann, R., Van Der Weyden, M., Mock, B. A., Mogil, J. S., Montagutelli, X., Lowenthal, R. M., Page, F., Garson, O. M., Morahan, G., Mountz, J. D., Nagase, H., & Juneja, S. (1996). A randomized study of Nowakowski, R. S., O’Hara, B. F., high-dose cytarabine in induction in acute Osadchuk, A. V., Paigen, B., Palmer, A. A., myeloid leukemia. Blood, 87, 1710–1717.
120 D. R. Bell and G. V. ZantBlau, H. M., Brazelton, T. R., & de Haan, G., & Van Zant, G. (1997). Intrinsic Weimann, J. M. (2001). The evolving and extrinsic control of hemopoietic stem concept of a stem cell: entity or function? cell numbers: mapping of a stem cell gene. Cell, 105, 829–841. Journal of Experimental Medicine, 186,Bonnet, D., & Dick, J. E. (1997). Human acute 529–536. myeloid leukemia is organized as a de Haan, G., & Van Zant, G. (1999a). Dynamic hierarchy that originates from a primitive changes in mouse hematopoietic stem cell hematopoietic cell. Nature Medicine, 3, numbers during aging. Blood, 93, 3294–3301. 730–737. de Haan, G., & Van Zant, G. (1999b). GeneticBotnick, L. E., Hannon, E. C., Obbagy, J., & analysis of hemopoietic cell cycling in mice Hellman, S. (1982). The variation of suggests its involvement in organismal life hematopoietic stem cell self-renewal span. FASEB Journal, 13, 707–713. capacity as a function of age: further de Haan, G., Nijhof, W., & Van Zant, G. evidence for heterogenicity of the stem cell (1997). Mouse strain-dependent changes in compartment. Blood, 60, 268–271. frequency and proliferation ofBradford, G. B., Williams, B., Rossi, R., & hematopoietic stem cells during aging: Bertoncello, I. (1997). Quiescence, cycling, correlation between lifespan and cycling and turnover in the primitive activity. Blood, 89, 1543–1550. hematopoietic stem cell compartment. Deguchi, K., & Gilliland, D. G. (2002). Experimental Hematology, 25, 445–453. Cooperativity between mutations in tyrosineBrendel, C., & Neubauer, A. (2000). kinases and in hematopoietic transcription Characteristics and analysis of normal and factors in AML. Leukemia, 16, 740–744. leukemic stem cells: current concepts and DePinho, R. (2000). The age of cancer. future directions. Leukemia, 14, 1711–1717. Nature, 408, 248–254.Cao, Y. A., Wagers, A. J., Beilhack, A., Doria, G., Mancini, C., Utsuyama, M., Dusich, J., Bachmann, M. H., Negrin, R. S., Frasca, D., & Hirokawa, K. (1997). Aging of Weissman, I. L., & Contag, C. H. (2004). the recipients but not of the bone marrow Shifting foci of hematopoiesis during donors enhances autoimmunity in reconstitution from single stem cells. syngeneic radiation chimeras. Mechanisms Proceedings of the National Academy of of Ageing and Development, 95, 131–142. Sciences of the USA, 101, 221–226. Estey, E. H., Kantarjian, H., Keating, M. J.Cesari, M., Penninx, B. W., Lauretani, F., (1995). Therapy for acute myeloid leukemia. Russo, C. R., Carter, C., Bandinelli, S., In R. Hoffman, Z. E. Ben, S. Shattel, B. Furce, Atkinson, H., Onder, G., Pahor, M., & & M. Cohen (Eds.), Hematology, basic Ferrucci, L. (2004). Hemoglobin levels principles and practice (2nd ed., Chapter 65). and skeletal muscle: results from the New York: Churchill Livingstone. InCHIANTI study. Journals of Gerontology. Fuchs, E., & Segre, J. A. (2000). Stem cells: Series A. Biological Sciences and Medical a new lease on life. Cell, 100, 143–155. Sciences, 59, M249–M254. Geiger, H., True, J. M., de Haan, G., & VanChen, J., Astle, C. M., & Harrison, D. E. Zant, G. (2001). Age- and stage-specific (2000). Genetic regulation of primitive regulation patterns in the hematopoietic hematopoietic stem cell senescence. stem cell hierarchy. Blood, 98(10), Experimental Hematology, 28, 442–450. 2966–2972.Cheshier, S. H., Morrison, S. J., Liao, X., Globerson, A., & Effros, R. B. (2000). Ageing & Weissman, I. L. (1999). In vivo of lymphocytes and lymphocytes in the proliferation and cell cycle kinetics of long- aged. Immunology Today, 21, 515–521. term self-renewing hematopoietic stem Goldstone, A. H., Burnett, A. K., Wheatley, K., cells. Proceedings of the National Academy Smith A. G., Hutchinson, R. M., & of Sciences of the USA, 96, 3120–3125. Clark, R. E. (2001). Attempts to improveDash, A., & Gilliland, D. G. (2001). Molecular treatment outcomes in acute myeloid genetics of acute myeloid leukaemia. Best leukemia (AML) in older patients: the results Practice & Research. Clinical of the United Kingdom Medical Research Haematology, 14, 49–64. Council AML 11 trial. Blood, 98, 1302–1311.
CHAPTER 4 / Hematopoietic Stem Cells, Aging, and Cancer 121Golub, T. R., Barker, G. F., Lovett, M., & oxidative stress. Annals of the New York Gilliland, D. G. (1994). Fusion of PDGF Academy of the Sciences, 1019, 278–284. receptor beta to a novel ets-like gene, tel, in Kay, H. E. (1965). How many cell-generations? chronic myelomonocytic leukemia with Lancet, 15, 418–419. t(5;12) chromosomal translocation. Cell, 77, Kim, M., Moon, H. B., & Spangrude, G. J. 307–316. (2003). Major age-related changes of mouseGrove, J. E., Bruscia, E., & Krause, D. S. hematopoietic stem/progenitor cells. (2004). Plasticity of bone marrow-derived Annals of the New York Academy of stem cells. Stem Cells, 22, 487–500. Sciences, 996, 195–208.Guralnik, J. M., Eisenstaedt, R. S., Ferrucci, L., Krtolica, A., Parrinello, S., Lockett, S., Klein, H. G., & Woodman, R. C. (2004). Desprez, P-Y., & Campisi, J. (2001). Prevalence of anemia in persons 65 years Senescent fibroblasts promote epithelial and older in the United States: evidence for cell growth and tumorigenesis: A ling a high rate of unexplained anemia. Blood, between cancer and aging. Proceedings of 104, 2263–2268. the National Academy of Sciences of theGuzman, M. L., & Jordan, C. T. (2004). USA, 98, 12072–12077. Considerations for targeting malignant Lagasse, E., Connors, H., Al-Dhalmy, M., stem cells in leukemia. Cancer Control, 11, Reitsma, M., Dohse, M., Osborne, L., 97–104. Wang, X., Finegold, M., Weissman, I. L., &Hanahan, D., & Weinberg, R. A. (2000). The Grompe, M. (2000). Purified hematopoietic hallmarks of cancer. Cell, 100, 57–70. stem cells can differentiate intoHarman, D. (1981). The aging process. hepatocytes in vivo. Nature Medicine, 6, Proceedings of the National Academy of 1229–1234. Sciences of the USA, 78, 7124–7128. Lapidot, T., Sirard, C., Vormoor, J., Murdoch, B.,Harrison, D. E. (1973). Normal production of Hoang, T., Caceres-Cortes, J., Minden, M., erythrocytes by mouse marrow continuous Paterson, B., Caligiuri, M. A., & Dick, J. E. for 73 months. Proceedings of the National (1994). A cell initiating human acute Academy of Sciences of the USA 70, myeloid leukaemia after transplantation 3184–3188. into SCID mice. Nature, 367, 645–648.Harrison, D. E., Lerner, C., Hoppe, P. C., Leith, C. P., Kopecky, K. J., Godwin, J., Carlson, G. A., & Alling, D. (1987). Large McConnell, T., Slovak, M. L., Chen, I. M., numbers of primitive stem cells are active Head, D. R., Appelbaum, F. R., & Willman, simultaneously in aggregated embryo C. L. (1997). Acute myeloid leukemia in the chimeric mice. Blood, 69, 773–777. elderly: assessment of multi-drug resistanceHeissig, B., Hattori, K., Dias, S., Friedrich, M., (MDR1) and cytogenetics distinguishes Ferris, B., Hackett, N. R., Crystal, R. G., biologic subgroups with remarkably Besmer, P., Lyden, D., Moore, M. A., distinct responses to standard Werb, Z., & Rafii, S. (2002). Recruitment of chemotherapy. A Southwest Oncology stem and progenitor cells from the bone Group Study. Blood, 89, 3323–3329. marrow niche requires MMP-9 mediated Liesveld, J. L., Jordan, C. T., & Phillips, G. L. release of kit-ligand. Cell, 109, 625–637. (2004). The hematopoietic stem cell inHolden, C. (2004). Stem cell research. myelodysplasia. Stem Cells, 22, 590–599. Advocates keep pot boiling as Bush plans Lipschitz, D. A., Mitchell, C. O., & new centers. Science, 305, 461. Thompson, C. (1981). The anemia ofJohnson, F. B., Sinclair, D. A., & Guarente, L. senescence. American Journal of (1999). Molecular biology of aging. Cell, 96, Hematology, 11, 47–54. 291–302. Lowenberg, B., Downing, J. R., & Burnett, A.Jordan, C. T. (2002). Unique molecular and (1999). Acute myeloid leukemia. New cellular features of acute myelogenous England Journal of Medicine 341, leukemia stem cells. Leukemia, 16, 1051–1062. 559–562. Malakoff, D. (2004). Election 2004. TheKawanishi, S., & Oikawa, S. (2004). calculus of making stem cells a campaign Mechanism of telomere shortening by issue. Science, 305, 760.
122 D. R. Bell and G. V. ZantMarley, S. B., Lewis, J. L., Davidson, R. J., Owen, R. (1945). Immunogenetic Roberts, I. A., Dokal, I., Goldman, J. M., & consequences of vascular anastomoses Gordon, M. Y. (1999). Evidence for a between bovine twins. Science, 102, 400. continuous decline in haemopoietic cell Penninx, B. W., Guralnik, J. M., Onder, G., function from birth: application to Ferrucci, L., Wallace, R. B., & Pahor, M. evaluating bone marrow failure in children. (2003). Anemia and decline in physical British Journal of Haematology, 106, performance among older persons. 162–166. American Journal of Medicine, 115,Mayer, R. J., Davis, R. B., Schiffer, C. A., 104–110. Berg, D. T., Powell, B. L., Schulman, P., Reya, T., Morrison, S. J., Clarke, M. F., & Omura, G. A., Moore, J. O., Mclntyre, O. R., Weissman, I. L. (2001). Stem cells, cancer, & Frei, E., 3rd (1994). Intensive and cancer stem cells. Nature, 414, postremission chemotherapy in adults with 105–111. acute myeloid leukemia. New England Rowe, J. M., Anderson, J. W., Mazza, J. J., Journal of Medicine, 331, 896–903. Bennett, J. M., Paietta, E., Hayes, F. A.,Mezey, E., Key, S., Vogelsang, G., Szalayova, I., Oette, D., Cassileth, P. A., Stadtmauer, E. A., Lange, G. D., & Crain, B. (2003). & Wiernik, P. H. (1995). A randomized Transplanted bone marrow generates new placebo-controlled phase III study of neurons in human brains. Proceedings of granulocyte-macrophage colony-stimulating the National Academy of Sciences of the factor in adult patients (Ͼ55–70 years of age) USA, 100, 1364–1369. with acute myelogenous leukemia: A studyMiller, R. A. (2000). Effect of aging on T of the Eastern Cooperative Oncology Group lymphocyte activation. Vaccine, 18, (E1490). Blood, 86, 457–462. 1654–1660. Schiller, G., & Lee, M. (1997). Long-termMiller, R. A., Chrisp, C., & Galecki, A. (1997). outcome of high-dose cytarabine-based CD4 memory T cell levels predict life span consolidation chemotherapy for older in genetically heterogenous mice. FASEB patients with acute myelogenous leukemia. Journal, 11, 775–783. Leukemia & Lymphoma, 25, 111–119.Mirro, J. Jr. (1992). Pathology and Schofield, R. (1978). The relationship between immunology of acute leukemia. Leukemia, the spleen colony-forming cell and the 6 Suppl 4, 13–15. haemopoietic stem cell. Blood Cells, 4, 7–25.Morrison, S. J., Wandycz, A. M., Akashi, K., Siminovitch, L., McCulloch, E. A., & Till, J. E. Globerson, A., & Weissman, I. L. (1996). (1963). The distribution of colony-forming The aging of hematopoietic stem cells. cells among spleen colonies. Journal of Nature Medicine, 2, 1011–1016. Cellular Physiology, 62, 327–336.Muller, W. A. (1996). Pattern formation in Siminovitch, L., Till, J. E., & McCulloch, E. A. the immortal Hydra. Trends in Genetics, (1964). Decline in colony-forming ability of 12, 91–96. marrow cells subjected to serialNCI. (2000). SEER Cancer Statistics Review transplantation into irradiated mice. Journal 1975–2000. of Cellular Physiology, 64, 23–31.Offner, F., Kerre, T., DeSmedt, M., & Plum, J. Smith, L. G., Weissman, I. L., & Heimfeld, S. (1999). Bone marrow CD34ϩ cells generate (1991). Clonal analysis of hematopoietic fewer T cells in vitro with increasing age stem-cell differentiation in vivo. and following chemotherapy. British Proceedings of the National Academy of Journal of Haematology, 104, 801–808. Sciences of the USA, 88, 2788–2792.Ogden, D. A., & Mickliem, H. S. (1976). The Spangrude, G. J., Brooks, D. M., & Tumas, D. B. fate of serially transplanted bone marrow (1995). Long-term repopulation of irradiated cell populations from young and old donors. mice with limiting numbers of purified Transplantation, 22, 287–293. hematopoietic stem cells: in vivo expansionOrkin, S. H., & Zon, L. I. (2002). of stem cell phenotype but not function. Hematopoiesis and stem cells: plasticity Blood, 85, 1006–1016. versus developmental heterogeneity. Spangrude, G. J., Heimfeld, S., & Weissman, Nature Immunology, 3, 323–328. I. L. (1988). Purification and
CHAPTER 4 / Hematopoietic Stem Cells, Aging, and Cancer 123 characterization of mouse hematopoietic growth and aging patterns in hematopoietic stem cells. Science, 241, 58–62. stem cell populations of allophenic mice.Stone, R. M., Berg, D. T., George, S. L., Dodge, Journal of Experimental Medicine, 171, R. K., Paciucci, P. A., Schulman, P., Lee, E. J., 1547–1565. Moore, J. O., Powell, B. L., & Schiffer, C. A. Wagers, A. J., & Weissman, I. L. (2004). (1995). Granulocyte-macrophage Plasticity of adult stem cells. Cell, 116, colony-stimulating factor after initial 639–648. chemotherapy for elderly patients with Wang, J. C. Y., Lapidot, T., Cashman, J. D., primary acute myelogenous leukemia. Doedens, M., Addy, L., Sutherland, D. R., New England Journal of Medicine, 332, Nayar, R., Laraya, P., Minden, M., 1671–1677. Keating, A., Eaves, A. C., Eaves, C. J., &Sudo, K., Ema, H., Morita, Y., & Nakauchi, H. Dick, J. E. (1998). High level engraftment of (2000). Age-associated characteristics of NOD/SCID mice by primitive normal and murine hematopoietic stem cells. Journal leukemic hematopoietic cells from patients of Experimental Medicine, 192, 1273–1280. with chronic myeloid leukemia in chronicTaichman, R. S., Reilly, M. J., & Emerson, S. G. phase. Blood, 91, 2406–2414. (1996). Human osteoblasts support Weissman, I. L. (2000a). Stem cells: units of human hematopoietic progenitor cells in development, units of regeneration, and vitro bone marrow cultures. Blood, 87, units in evolution. Cell, 100, 157–168. 518–524. Weissman, I. L. (2000b). Translating stemTaylor, P. R., Reid, M. M., Stark, A. N., and progenitor cell biology to the clinic: Brown, N., Hamilton, P. J., & Proctor, S. J. barriers and opportunities. Science, 287, (1995). De novo acute myeloid leukemia in 1442–1446. patients over 55-years-old: a population Wineman, J., Moore, K., Lemischka, I., based study of incidence, treatment and & Muller-Sieburg, C. (1996). Functional outcome. Leukemia, 9, 231–237. heterogeneity of the hematopoieticTerskikh, A. V., Miyamoto, T., Chang, C., microenvironment: rare stromal elements Diatchenko, L., & Weissman, I. L. (2003). maintain long-term repopulating stem Gene expression analysis of purified cells. Blood, 87, 4082–4090. hematopoietic stem cells and committed Yamamoto, K., Nakamura, Y., Saito, K., progenitors. Blood, 102, 94–101. & Furusawa, S. (2000). Expression of theTill, J. E., & McCulloch, E. A. (1961). A direct NUP98/HOXA9 fusion transcript in the measurement of the radiation sensitivity of blast crisis of Philadelphia chromosome- normal mouse bone marrow cells. positive chronic myelogenous leukaemia Radiation Research, 14, 1419–1430. with t(7;11)(p15;p15). British Journal ofVan Zant, G., & Liang, Y. (2003). The role of Haematology, 109, 423–426. stem cells in aging. Experimental Zaucha, J. M., Yu, C., Mathioudakis, G., Hematolology, 31, 659–672. Seidel, K., Georges, G., Sale, G., Little, M. T.,Van Zant, G., Eldridge, P. W., Behringer, R. R., Torok-Storb, B., & Storb, R. (2001). & Dewey, M. J. (1983). Genetic control of Hematopoietic responses to stress hematopoietic kinetics revealed by analyses conditions in young dogs compared with of allophenic mice and stem cell suicide. elderly dogs. Blood, 98, 322–327. Cell, 35, 639–645. Zhu, J., & Emerson, S. G. (2004). A new boneVan Zant, G., Holland, B. P., Eldridge, P. W., to pick: osteoblasts and the haematopoietic & Chen, J. J. (1990). Genotype-restricted stem-cell niche. Bioessays, 26, 595–599.
CHAPTER 5 / Mitochondria: A Critical Role in Aging 125Boveris, 1980). In isolated brain mitochon- population, but usually a single mtDNAdria, the main site of mitochondrial radi- variant populates an egg, resulting in acal production has been reported to be at homoplasmic individual. Alternatively,complex I, on the matrix side of the inner more than one variant may be passed tomitochondrial membrane (Kudin et al., the offspring, resulting in heteroplasmy, in2004). However, debate surrounds the which a single cell contains multiple dif-relative importance of each site in vivo ferent mtDNA sequences. Heteroplasmyand whether radicals are generated at high can also develop as a result of the accu-enough levels to cause damage to cellular mulation of somatic mutations to thecomponents, including the mtDNA (Chen mtDNA during the lifetime of the indi-et al., 2003; Liu et al., 2002a; St-Pierre vidual. In this case, the mutations are sto-et al., 2002). chastic, with different unique mutations Mammalian mtDNA is a circular accumulating to different frequencies ingenome of approximately 16,500 bp, each cell. The fraction of mitochondrialencoding 13 polypeptides of the respira- genomes that need to be mutated (thetory chain, two ribosomal RNAs and “mutational load” of a cell) in vivo before22 tRNAs. In addition to the transcribed mitochondrial or cellular functions areregion, mtDNA also contains a non-cod- affected remains unknown, although stud-ing D-loop region believed to be involved ies of mitochondrial disease have indi-in mtDNA maintenance, replication, and cated that for dramatic phenotypes suchtranscription (Bogenhagen & Clayton, as clinically significant myopathies, the2003; Gillum & Clayton, 1978; Holt & majority of mtDNA needs to be mutant.Jacobs, 2003). The remainder of theproteins required for mitochondrial func-tion are encoded by nuclear genes. II. Evidence for IncreasedMammalian cell culture models and stud-ies in S. cerevisiae indicate that mtDNA Oxidative Damage tomolecules exist in nucleoids, discrete Mitochondrial Componentsstructures containing 2 to 10 mtDNA with Agemolecules combined with protein factors A. Mitochondrial DNA Mutationsinvolved in replication and transcription and Aging(Iborra et al., 2004; Legros et al., 2004;Miyakawa et al., 1987). The mitochondrial The age-related increases in the levels ofnucleoids appear to be directly tethered to both oxidative damage and mutationalthe inner mitochondrial membrane and load of mtDNA predicted by the mito-closely associated with the mitochondrial chondrial theory of aging have beenprotein import machinery (Iborra et al., described in multiple species and organ2004; Legros et al., 2004). Attachment of systems. However, whether this damagemtDNA to the inner mitochondria mem- affects mitochondrial function or signifi-brane is proposed to facilitate respiratory cantly modulates the physiology of agingchain assembly but also places mtDNA in has remained controversial (Jacobs,close proximity to the major cellular site 2003a,b; Pak et al., 2003a,b).of ROS production, the ETC. Age-related duplications and concatena- With hundreds of mitochondria per cell, tions of mtDNA were first described usingand tens of mtDNA molecules per mito- electron microscopy (Piko & Matsumoto,chondrion, an average cell may contain 1977; Piko et al., 1978). With the advent ofover a thousand copies of mtDNA. The PCR, age-related increases in large scalemtDNA is inherited maternally, and deletions of the mtDNA have beennumerous variants exist in the human described in C. elegans (Melov et al.,
126 T. R. Golden et al.1994; Melov et al., 1995a), Drosophila 1991; King et al., 1992; Trounce et al.,melanogaster (Yui et al., 2003), mouse 1994). Therefore, it has remained unclear(Tanhauser & Laipis, 1995), rat (Filburn whether the age-related increases inet al., 1996), monkey (Lee et al., 1993), and mitochondrial mutations that have beenhuman (Cortopassi & Arnheim, 1990; detected can play a role in phenotypesMelov et al., 1995b; Melov et al., 1999a). associated with aging.New technologies have also shown that Two hypotheses have been formulatedpoint mutations in the control region as to how age-related mtDNA mutationsof human mtDNA increase with age may be deleterious. The first is based on(Michikawa et al., 1999). However, deter- the observation that a single mutationmining the total mutational load of the can be found at very high levels within anmtDNA in any one tissue or cell remains a individual cardiomyocyte (Khrapko et al.,difficult proposition because age-related 1999) or muscle fiber (Schwarze et al.,somatic mutations are stochastic, resulting 1995; Wanagat et al., 2001) and that thesein heteroplasmy. Generally, studies have mutations co-localize with regions offocused on quantifying a specific deletion impaired mitochondrial function (Fayet(e.g., the so-called “common deletion” et al., 2002; Wanagat et al., 2002). Thisin human mtDNA) or detecting point has lead to the hypothesis that discretemutations through sequencing of a small mutations may, through clonal expansion,segment of the mitochondrial genome. reach a prevalence within a single cell atNeither of these approaches provides a which they are capable of affecting thecomprehensive picture of the potential function of that cell. In this model, eachtotal mutational load in a cell or tissue. cell would have its own unique mutation, Studies that have quantified specific making any one mutation difficult tomutations have found that the observed detect when a homogenized tissue or pop-prevalence of any one mutation in aged ulation of cells is studied. Alternatively, ittissue has been low, generally below 1 has been hypothesized that multiple, ran-percent in homogenized tissue samples, dom mtDNA mutations may be presentthough one report found the common in an individual cell, and though each isdeletion reached more than 10 percent in in low amount, enough mutations accu-the putamen (Corral-Debrinski et al., mulate such that a substantial fraction of1992). This is far below the level of muta- the mtDNA in the cell is damaged, withtional load known to be associated with significant phenotypes arising as a result.pathology in mitochondrial diseases, in Recent work in support of the firstwhich a mutant form of mtDNA may be hypothesis includes studies of adult steminherited from the mother (Holt et al., cells. Stem cells at the base of colonic1988; Wallace, 1994). For example, char- crypts were discovered to accumulateacteristic mtDNA deletions are present at deleterious mtDNA mutations at a rela-levels between 45 and 75 percent in tively high frequency of 5 ϫ 10Ϫ5 muta-affected tissues of patients with the mito- tions per genome (Taylor et al., 2003). Inchondrial disease Kearne Sayre Syndrome several cases, the mtDNA mutations(KSS) (Zeviani et al., 1988). In addition, reached levels of prevalence in the cryptcell culture studies in which mtDNA cells at which they could be expected tofrom one cell is introduced into a cell line affect tissue function. The authors attri-depleted of its own mtDNA (0 cells), buted an age-related decline in cytochromegenerating “cybrid” cells, indicate that a oxidase activity (complex IV of the ETC)significant majority of the mitochondrial in colonic crypts to this expansion ofgenomes (60 to 90 percent) must be stem cell mtDNA mutations.mutant for a disease-related mutation The second hypothesis, that an increaseto cause a phenotype (Hayashi et al., in random mtDNA mutations can
CHAPTER 5 / Mitochondria: A Critical Role in Aging 127influence aging, is supported by work resulting in a very long polycytosine tractdemonstrating that a high load of point in the D-loop region of the mtDNA. It ismutations accumulates in agedhuman believed that this results in mitochon-brain, such that as many as three muta- drial DNA polymerase slippage duringtions per mitochondrial genome are pres- replication, resulting in a heteroplasmicent in aged individuals (Simon et al., length variation of the tract (Marchington2004). In addition, this hypothesis was et al., 1997). It has been proposed thatdirectly tested by a recent study that subsequent instability in the D-loopdemonstrated the importance of mtDNA region, which plays a role in mtDNAmaintenance (Trifunovic et al., 2004). replication and transcription, could resultIn this study, the endogenous mouse in reduced mtDNA copy number overmitochondrial DNA polymerase (PolgA) time (Poulton et al., 1998). The 16,189was replaced with an enzyme with variant has been shown to have a positivereduced exonuclease activity, resulting association with type II diabetes, dilatedin impaired proofreading during DNA cardiomyopathy, low body fat at birth,replication. The load of point mutations iron loading associations in haemachro-per mtDNA molecule increased approxi- matosis, and stroke (Casteels et al., 1999;mately threefold, although no single Kim et al., 2002; Liou et al., 2004;mutation or “hot spot” for mutations was Livesey et al., 2004; Poulton et al., 2002),observed. In addition, the total amount of although this has not been confirmedmtDNA decreased by 30 percent, in other studies (Gibson et al., 2004;although mitochondrial transcript levels Gill-Randall et al., 2001).remained unchanged. The increase in A recent study identified a novel mito-mutational load resulted in reduced life chondrial mutation that can be a riskspan, and the development of some age- factor in a set of the most common age-related pathologies, including loss of bone related diseases. Wilson and colleaguesdensity, weight loss, cardiomyopathy, ane- identified an mtDNA mutation that pre-mia, and reduced fertility. This mouse disposed a family to hypomagnesemia,model demonstrates that it is possible for hypertension, and hypercholesterolemia,an increase in random mtDNA mutations all risk factors for cardiovascular diseaseto affect tissue function and supports the (Wilson et al., 2004). Interestingly, thehypothesis that the age-related increase in mutation did not increase incidence ofmutational load described in many sys- type II diabetes or insulin resistance, phe-tems could potentially contribute to some notypes frequently observed with mito-of the phenotypes associated with aging. chondrial DNA mutations (discussed Because maternally transmitted mtDNA below). How the mutation, which impairsmutations are more likely to be homo- the ability of tRNAIle to bind to the ribo-plasmic, they are much easier to asso- some, causes the phenotypes measured isciate with a disease or aging than the unknown, but the study supports therandom somatic mtDNA mutations gen- hypothesis that mtDNA mutations canerated as a result of oxidative damage. In contribute to age-related phenotypes.combination with certain nuclear back-grounds or environmental factors, spe- B. Oxidative Damage to Mitochondrialcific mtDNA variants may affect respira- Proteins and Membranes and Agingtory chain function and aging (Chinneryet al., 1999; Niemi et al., 2003). The Numerous studies have been aimed atmtDNA D-loop variant has been best detecting age-related increases in oxidativecharacterized in relation to age-related damage to protein or lipid components ofconditions. In this variant, cytosine the mitochondrion (Berlett & Stadtman,replaces thymidine at position 16,189, 1997; Linton et al., 2001). Aging has been
128 T. R. Golden et al.associated with both an increase in oxida- hypothesized that elevated levels oftive damage to proteins and a decrease in superoxide disrupt a structurally impor-protein turnover and repair (Stadtman, tant Fe-S cluster of SDH, resulting in2004). The age-related increase in pro- reduced levels of assembled complex intein modifications includes cross-linkages, the mitochondria (Hinerfeld et al., 2004).fragmentation, carbonylation, glycation, Aconitase has been specifically exam-and advanced glycation endproduct (AGE) ined as a target of oxidative damageformation (Berlett & Stadtman, 1997). with respect to aging. In the house fly (YanRecent evidence indicates that in blood et al., 1997), oxidation to mitochondriallymphocytes, some proteins are selec- aconitase was found to increase with age,tively modified with age (Poggioli et al., and a corresponding age-related decrease in2004), indicating that there may be a non- aconitase activity was described. The ade-random pattern of protein damage asso- nine nucleotide translocase has also beenciated with aging. Oxidative damage to identified as a mitochondrial protein thatproteins may lead to dysfunction and/or increasingly is oxidatively damaged withaggregation of the damaged proteins, age in flies (Yan & Sohal, 1998). In otherwhich can result in their being targeted for studies via proteomic profiling, heartremoval and degradation (Cuervo & Dice, mitochondrial protein was compared in1998; Levine, 2002). young and old rats (Kanski et al., 2004), Numerous lines of evidence support resulting in the identification of oxidativemitochondria as the primary source of modifications to 48 proteins with age.ROS (Chance et al., 1979; Li et al., 1995). Among the group were many metabolicHence, mitochondrial proteins, including enzymes, including succinate dehydroge-the complexes of the ETC and members nase and aconitase (Kanski et al., 2004). Inof other metabolic pathways, are likely addition, numerous oxidatively modifiedto be particularly susceptible to modifi- proteins have been identified in bovinecations as a result of oxidative damage. heart sub-mitochondrial particles, theCertain cellular components are known majority localized in the mitochondrialto be sensitive to oxidative stress. m