American Journal of Orthodontics and Dentofacial Orthopedics Moss 9Volume 112, No. 1biologic cellular network theory, permit the presen- cephalic growth, at the gross anatomic level, and ittation of this latest revision. 7-1° had two explanatory constraints: methodologic and hierarchical.THE CONCEPTUAL AND A N A T O M I C BASES OF 1. Methodologic constraint. Macroscopic mea-THE REVISED FMH surements, which use the techniques of point A comprehensible revision of the FMH should mechanics and arbitrary reference frames, e.g.,indicate (a) those portions that are retained, ex- roentgenographic cephalometry, permitted onlytended or discarded, and (b) which prior deficien- method-specific descriptions that cannot be struc-cies are now resolved. turally detailed. This constraint was removed by Although the principal FMH concepts are either the continuum mechanics techniques of the finitegenerally known or easily available, 111-18 three are element method (FEM) 61921 and of the relatedof particular resonance for this revision. macro and boundary element methods. 9,22 The developmental origin of all cranial skeletal This penultimate FEM revision added objective,elements (e.g., skeletal units) and all their subsequent reference-frame-invariant, fine-grained, and con-changes in size and shape (e.g., form) and location, as ceptually integrated descriptions of the quantitativewell as their maintenance in being, are always, without aspects of localized cephalic growth kinematics toexception, secondary, compensatory, and mechanically the earlier qualitative (phenomenologic) descrip-obligatory responses to the temporally and operation- tions of growth dynamics. 4,6,9ally prior demands of their related cephalic nonskel- 2. Hierarchical constraint. However, even thatetal cells, tissues, organs, and operational volumes versions descriptions did not extend "downward" to(e.g., the functional matrices). processes at the cellular, subcellular, or molecular More precisely, the FMH claims that epigenetic, structural domains, or extend "upwards" to theextraskeletal factors and processes are the prior, multicellular processes by which bone tissues re-proximate, extrinsic, and primary cause of all adap- spond to lower level signals. All prior FMH versionstive, secondary responses of skeletal tissues and were "suspended" or "sandwiched" as it were, be-organs? It follows that the responses of the skeletal tween these two hierarchical levels.unit (bone and cartilage) cells and tissues are not Explicitly, the FMH could not describe eitherdirectly regulated by informational content of the how extrinsic, epigenetic FM stimuli are transducedintrinsic skeletal cell genome per se. Rather, this into regulatory signals by individual bone cells, oradditional, extrinsic, epigenetic information is cre- how individual cells communicate to produce coor-ated by functional matrix operations. dinated multicellular responses. The F M H postulates two types of functional At the lower cellular or molecular levels, anothermatrices: periosteal and capsularJ 617 The former, problem exists. Almost uniformly, experimental andtypified by skeletal muscles, regulates the histologi- theoretical studies of bone adaptation consider onlycally observable active growth processes of skeletal the unicellular, unimolecular, or unigenomic levels.tissue adaptation. Accordingly, their results and derivative hypotheses This new version deals only with the responses to generally are not extensible to higher multicellular,periosteal matrices. It now includes the molecular and tissue, levels.cellular processes underlying the triad of active skele- Consequently, in prior FMH versions, significanttal growth processes: deposition, resorption, and main- disjunctions exist between the descriptions at eachtenance. Histologic studies of actively adapting osse- of the several levels of bone organization. Such aous tissues demonstrate that (1) adjacent adaptational hiatus is implicit in hierarchical theory in which thetissue surfaces simultaneously show deposition, re- attributes of successively higher levels are not simplysorption, and maintenance; (2) adaptation is a tissue the sum of lower level attributes. Rather, at eachprocess. Deposition and maintenance are functions of higher level, new and more complex structural andrelatively large groups (cohorts, compartments) of operational attributes arise that cannot be pre-homologous osteoblasts, never single cells; and (3) a dicted, even from a complete knowledge of those ofsharp demarcation exists between adjacent cohorts of the lower levels23; e.g., the sum of all lower at-active, depository, and quiescent (resting) osteoblasts. tributes (biophysical, biochemical, genomic) of a bone cell cannot predict the higher attributes of aConstraints of the FMH bone tissue. Initially, the FMH ~,2 provided only qualitative At present, no unitary hypothesis provides anarrative descriptions of the biologic dynamics of comprehensive, coherent and integrated description
10 Moss American Journal of Orthodontics and Dentofacial Orthopedics July 1997of all the processes and mechanisms involved in generally evoke one; (3) osseous signal transmissionbone growth, remodeling, adaptation, and mainte- is aneural, whereas all other mechanosensationalnance at all structural levels. This newest FMH signals use some afferent neural pathways28.41; and,version, presented herein, transcends some hierar- (4) the evoked bone adaptational responses arechical constraints and permits seamless descriptions confined within each "bone organ" independently,at, and between, the several levels of bone structure e.g., within a femur, so there is no necessary "inter-and operation-from the genomic to the organ level. bone" or organismal involvement.It does so by the inclusion of two complementary This process translates the information contentconcepts: (1) that mechanotransduction occurs in of a periosteal functional matrix stimulus into asingle bone cells, and (2) that bone cells are com- skeletal unit cell signal, for example, it moves infor-putational elements that function multicellularly as mation hierarchically downward to the osteocytes.a connected cellular network. There are two, possibly complementary, skeletal It is useful to present the database and derivative cellular mechanotransductive processes: ionic andtheories, supportive of the inclusion of these two mechanical.concepts individually in a series of two coordinated Ionic or electrical processes. This involves somearticles: the first on mechanotransduction and the process(es) of ionic transport through the bone cellsecond on connected cellular networks. (osteocytic) plasma membrane. There is a subse- quent intercellular transmission of the created ionicMechanotransduction or electrical signals that, in turn, are computed by All vital ceils are "irritable" or perturbed by and the operation of an osseous connected cellularrespond to alterations in their external environment. network (CCN), as described in the second article inMechanosensing processes enable a cell to sense this series. That networks output regulates theand to respond to extrinsic loadings, a widespread multicellular bone cell responses. 1°,42biologic attribute, 24-32 by using the processes of Although no consensual agreement exists, osteo-mechanoreception and of mechanotransduction. cytic, ionic-mechanotransduction may involve sev-The former transmits an extracellular physical stim- eral, possibly parallel, cellular processes.ulus into a receptor cell; the latter transduces or Stretch-activated channels. Several types of defor-transforms the stimuluss energetic and/or informa- mation may occur in strained bone tissue. One oftional content into an intracellular signal. Mechano- these involves the plasma membrane stretch-acti-transduction33 is one type of cellular signal transduc- vated (S-A) ion channels, a structure found in bonetion. 34-36 There are several mechanotransductive cells, 43-46 in many other cell types,25 and significantlyprocesses, for example, mechanoelectrical and in fibroblasts. 4v When activated in strained osteo-mechanochemical. Whichever are used, bone adap- cytes, they permit passage of a certain sized ion ortation requires the subsequent intercellular trans- set of ions, including K +, Ca 2+, Na +, and CS+. 4648-50mission of the transduced signals. Such ionic flow may, in turn, initiate intracellular electrical events, for example, bone cell S-A chan-Osseous Mechanotransduetion nels may modulate membrane potential as well as Static37 and dynamic3s loadings are continuously Ca 2+ ion fluxY ,5~ Other bone cell mechanicallyapplied to bone tissues, tending to deform both stimulatory processes have been suggested.52extracellular matrix and bone cells. When an appro- Rough estimates of osteocytic mechanoreceptorpriate stimulus parameter exceeds threshold values, strain sensitivity have been made, 1°,53 and the calcu-the loaded tissue responds by the triad of bone cell lated values cover the morphogenetically significantadaptation processes. Both osteocytes and osteo- strain range of 1000 to 3000 txe in the literature. 54-56blasts are competent for intracellular stimulus re- Electrical processes. These include several, non-ception and transduction and for subsequent inter- exclusive mechanotransductive processes (e.g., elec-cellular signal transmission. Osteoblasts directly tromechanical and electrokinetic), involving theregulate bone deposition and maintenance and in- plasma membrane and extracellular fluids. Electricdirectly regulate osteoclastic resorption. 39,4° field strength may also be a significant parameterF Osseous mechanotransduction is unique in fourways: (1) Most other mechanosensory cells are 1. Electromechanical. As in most cells, the osteo-cytologically specialized, but bone cells are not; (2) cytic plasma membrane contains voltage-acti-one bone-loading stimulus can evoke three adapta- vated ion channels, and transmembrane iontional responses, whereas nonosseous processes flow may be a significant osseous mechano-
American Journal of Orthodontics and Dentofacial Orthopedics Moss 11Volume 112, No. 1 transductive p r o c e s s . 58596°-62 It is also possi- alternative means by which periosteal functional ble that such ionic flows generate osteocytic matrix activity may regulate hierarchically lower action potentials capable of transmission level bone cell genomic functions. through gap junctions. 63 The mechanical properties of the extracellular 2. Electrokinetic. Bound and unbound electric matrix influence cell behavior. 71 Loaded mineral- charges exist in bone tissue, many associated ized bone matrix tissue is deformed or strained. with the bone fluid(s) in the several osseous Recent data indicate that a series of extracellular spaces or compartments. 42,64 It is generally macromolecular mechanical levers exist, capable of agreed that electrical effects in fluid-filled transmitting information from the strained matrix to bone are not piezoelectric, but rather of elec- the bone cell nuclear membrane. trokinetic, that is, streaming potential (SP) The basis of this mechanism is the physical origin. 426566 The SP is a measure of the continuity of the transmembrane molecule integrin. strain-generated potential (SGP) of con- This molecule is connected extracellularly with the vected electric charges in the fluid flow of macromolecular collagen of the organic matrix and deformed bone. The usually observed SPG of intracellularly with the cytoskekeletal actin. The +2 mV can initiate both osteogenesis and molecules of the latter, in turn, are connected to the osteocytic action potentials. 6667 nuclear membrane, at which site the action of the 3. Electric field strength. Bone responds to exog- mechanical lever chain previously noted initiates a enous electrical fields. 68 Although the extrin- subsequent series of intranuclear processes regula- sic electrical parameter is unclear, field tory of genomic activity. 72-75 (See Shapiro et al., 76 for strength may play an important role. 69 A vimentin, and Green 77 for a general discussion of significant parallel exists between the param- biophysical transductions.) eters of these exogenous electrical fields 68,69 It is suggested that such a cytoskeletal lever and the endogenous fields produced by mus- chain, connecting to the nuclear membrane, can cle activity. Bone responds to exogenous elec- provide a physical stimulus able to activate the trical fields in an effective range of 1 to 10 osteocytic genome, 78 possibly by first stimulating the ixV/cm, strengths that are "...on the order of activity of such components as the cfos those endogenously produced in bone tissue genes.36,73, 78-86 during normal (muscle) activity "7° (italics It is by such an interconnected physical chain of mine). molecular levers that periosteal functional matrix Mechanical processes. Although it is probable activity may regulate the genomic activity of itsthat the intracellular, transductive process discussed strained skeletal unit bone cells, including theirlater does not initiate action potentials, it is an phenotypic expression.
222 Moss American Journal of Orthodontics and Dentofacial Orthopedics August 1997 massively parallel or parallel-distributed signal pro- sentation of CCN is redundant, assuring that the cessing occurs. 1°3-m5It computationally processes, in a network is fault or error tolerant, i.e, one or several multiprocessor network mode, the intercellular signals inoperative cells causes little or no noticeable loss in created by an electrical type of mechanotransduction network operations, 112 a matter of useful clinical of periosteal functional matrix stimuli. Subsequently significance. the computed network output informational signals The CCNs show oscillation, i.e., iterative recip- move hierarchically "upward" to regulate the skeletal rocal signaling (feedback) between layers. This at- unit adaptational responses of the osteoblasts. tribute enables them to adjustively self-organize. Fortunately, the bases of connectionist theory This behavior is related to the fact that biologic are Sufficiently secure to permit modeling of a CCNs are not preprogrammed; rather they learn bybiologically realistic osseous C C N . 1°6-11° It consists unsupervised or epigenetic "training, 114 a process of a number of relatively simple, densely intercon- probably involving structural or conformationalnected processing elements (bone cells), with many changes in the cytoskeleton. 83 The phenomena ofmore interconnections than cells. It is useful that both network "training" and "learning" are relatedbone cells form a network because individual recep- to the suggested effects of the oscillatory nature oftors cannot code unambiguously-only a population their strain history. 115 Accordingly, the structurallyof cells can do SO. 103 more complex network attributes and behavior of a In network theory, these cells are organized into CCN gradually or epigenetically self-organize and"layers": an initial input, a final output, and one or emerge during operation. These network attributesmore intermediate or "hidden" layers. Importantly, are not reducible, i.e., they are neither apparent norsuch networks need not be numerically complex to predictable from a prior knowledge of the attributesbe operationally complex. H~ The operational pro- of individual cells.cesses are identical, in principle, for all bone cells in Gap junctions, permitting bidirectional flow ofall layers. Regardless of the actual physiological information, are the cytological basis for the oscil-stipulatory process, each cell in any layer may simul- latory behavior of a CCN. All the osteoblasts of ataneously receive several "weighted" inputs (stimu- cohort engaged in an identical adaptation processli). A weight is some quantitative attribute. In the are interconnected by open gap junctions. The pres-initial layer, these represent the loadings. Within ence of sharp histological discontinuities betweeneach cell independently, " . . . all the weighted inputs cohorts of phenotypically different osteoblasts isare then summed. 112 This sum is then compared, related to their ability to close gap junctions at thewithin the cell, against some liminal or threshold boundaries between such cohorts, and so preventvalue. If this value is exceeded, an intracellular the flow of information. 116,1~7 Informational net-signal is generated, i.e., successful mechanotrans- works also can transmit inhibitory signals, a signifi-duction occurs. This signal is then transmitted iden- cant matter beyond present concerns. 118tically to all the "hidden" layer cells (adjacent osteo- A skeletal CCN displays the following attributes:cytes) to which each initial layer cell is connected by (1) Developmentally, it is an untrained self-orga-gap junctions (and there are many styles of connec- nized, self-adapting and epigenetically regulated sys-tivity). Next, similar processes of weighted signal tem. (2) Operationally, it is a stable, dynamic systemsummation, comparison, and transmission occur in that exhibits oscillatory behavior permitting feed-these intermediate layers until the final layer cells back. It operates in a noisy, nonstationary environ-(osteoblasts) are reached. The outputs of these ment, and probably uses useful and necessary inhib-anatomically superficial cells determines the site, itory inputs. (3) Structurally, an osseous CCN israte, direction, magnitude, and duration of the nonmodular, i.e., the variations in its organizationspecific adaptive response, i.e., deposition, resorp- permit discrete processing of differential signals. It istion, and/or maintenance, of each cohort of osteo- this attribute that permits the triad of histologicblasts. ~13 responses to a unitary loading event. Information is not stored discretely in a CCN, as Certain simplifications exist in this article, as init is in a conventional, single CPU computer. Rather most of the bone literature. It is assumed that boneit is distributed across all or part of the network, and cells are organized in only two dimensions, boneseveral types of information may be stored simulta- loadings occur only at discrete loci, and gradients ofneously. The instantaneous state of a CCN is a strain are not considered. However, biologic realityproperty of the state of all its cells and of all their is otherwise. In a loaded three-dimensional boneconnections. Accordingly, the informational repre- volume, gradients of deformation must exist, and
American Journal of Orthodontics and Dentofacial Orthopedics I~[OSS ~Volume t12, No. 2each osteocyte probably senses uniquely different Skeletal muscle contraction is a typical perios-strain properties. Further, it is probable that each teal functional matrix loading event, 13,14A6,12°,134135osteocyte is potentially able to transmit three differ- and frequency is one of its critical parameters.ent adaptational signals, in three different direc- Although the fundamental frequency of contractingtions-some stimulatory and some inhibitory. How- muscle is about 2 Hz, other strain-related harmonicsever, these processes have not yet been adequately of 15 to 40 Hz exist.modeled. The role of pe1~osteal functional matrices: These higher-order frequencies, significantlynew insight. related to bone adaptational responses, are The morphogenetic primacy of periosteal func- " . . . present within the [muscle contraction] straintional matrices on their skeletal units is consensually energy spectra regardless of animal or activity andaccepted. As a muscular demand alters, e.g., myec- implicate the dynamics of muscle contraction as thetomy, myotomy, neurectomy, exercise, hypertrophy, source of this energy band" (italics mine). 68,132~36 Ofhyperplasia, atrophy, augmentation, or reposition- particular significance to the FMH is the closeing, the triad of active bone growth processes cor- similarity of muscle stimulus frequencies to bonerespondingly adapts the form of its specifically re- tissue response frequencies.lated skeletal unit. Presently excluding the stimulation of neural MECHANOTRANSDUCTION: A TENTATIVEafferents in muscle, tendon, and periosteum, extrin- SYNTHESISsic physical loadings tend to deform bone tissue and The previously mentioned data suggest that theto invoke skeletal unit (bone) adaptation responsive ability of periosteal functional matrices to regulateprocesses. A classic example is the regulation of the adaptive responses of their skeletal units by ioniccoronoid process form by the temporalis muscle.~9 mechanotransductive processes is related to severalThe tension in the tendon of this contracted muscle, factors. These are that (a) normal muscle functiontransmitted through intertwined periosteal fibers strains attached bone tissue intermittently; (b) theinserted into subjacent bone, deforms the loaded dynamics of skeletal muscle contraction fit ratherskeletal unit. 12° nicely with the energetic requirements for bone cell Although some periosteal osteoNasts may be responsiveness; (c) the range of specific strain-directly stimulated, ~2~ extant data suggest osteocytic frequency harmonics of muscle dynamics are alsoprimacy in mechanosensory processes. ~22 Anatomi- those found to be morphogenetically competentcally, bone cells are competent mechanoreceptors. (i.e., osteoregulatory); (d) normal skeletal muscleTheir three-dimensional array of extensive canalic- activity produces intraosseous electric fields on theular cell processes is architecturally well-suited to order of extrinsic fields found to be similarly mor-sense deformation of the mineralized matrixJ 23 phogenetic; and, (e) bone cells may be stimulated by Although no one mechanical parameter reliably two mechanisms-directly by strain-activated plasmapredicts all bone adaptational or remodeling re- membrane channels and indirectly by electrokinen-sponses, 124strain probably plays the primary role 125-128 tic phenomena.and is a competent stimulus. 51 The significant strain These factors strongly suggest a rather preciseattribute may vary with specific conditions. 129 These matching of significant operational characteristicsinclude: (a) loading category-bone responds best to between a contracting skeletal muscle stimulus anddynamic rather static loading54; (b) frequency-osteo- the ability of loaded bone cells to transduce this intocytes may be physiologically "tuned" to the frequencies signals capable of regulating their adaptive re-of muscle function, 13°132 tunings being analogous to sponses. In a phrase, bone appears to be closelythose of specialized nonosseous sensory cells,34,35 e.g., "tuned" to skeletal muscle, i.e., skeletal units areauditory hair cells; and (c) magnitude-relatively small tuned to their periosteal functional matrices.microstrains (txe) (about 10-6 mm/mm), and strain When both the ionic membrane and the me-magnitudes of 2000 + 1000 ge, are morphogenetically chanical (molecular lever) transductive processescompetent.55,56,129.~33 are conceptually and operationally combined with Although it is reasonably presumed that mech- the data of both electric field effects and of contrac-anosensory processes, of both the ionic and mechan- tion frequency energetics, they provide a logicallyical type, involve the plasma membrane of the sufficient biophysical basis of support for the hy-osteocytic soma or canalicular processes, the recep- pothesis of epigenetic regulation of skeletal tissuetive, and subsequent transductive, processes are adaptation 1,13,16-1s,38,129,137neither well understood nor consensually agreed on. In reality, it is probable that the ionic (electrical)
224 Moss American Journal of Orthodontics and Dentofacial Orthopedics August 1997and mechanical (molecular lever) transductive pro- 16. Moss ML, Salentijn L. The primary role of the functional matrices in facial growth. Arn J Orthod 1969;55:566-77.cesses in osteocytes are neither exhaustive nor mu- 17. Moss ML, Salentijn L. The capsular matrix. Am J Orthod 1969;56:474-90.tually exclusive. While using differing intermediate 18. Moss ML, Young R. A functional approach to craniology. Am J Phys Anthrop 1960;18:281-92.membrane mechanisms or processes, they share a 19. Skalak R, Dasgupta G, Moss ML, Otten E, Dullemeijer P, Vilmann H. Acommon final common pathway, i.e., they eventually conceptual framework for the analytical description of growth. J Theor Biol 1982;94:555-77.produce signals regulatory of osteoblastic activity. 20. Skalak R, Dasgupta G, Moss ML, Patel H, Sen K, Moss-Salentijn L. TheCertainly in the ionic processes, and possibly in the application of the finite element method to the analysis of craniofacial growth and form. Am J Orthod 1985;87:453-72.molecular lever system mechanism, the transductive 21. Moss ML, Moss-Salentijn L, Skalak R. Finite element modeling of craniofacialprocess(es) also cause a transplasma membrane growth and development. In: Graber L, editor. Orthodontics: stepping stones to the future. St Louis: CV Mosby 1986:143-68.ionic flow(s), creating a signal(s) capable of inter- 22. McAlarney M, Dasgupta G, Moss ML, Moss-Salentijn L. Anatomical macroele-cellular transmission to neighboring bone cells ments in the study of craniofacial rat growth. J Craniofac Genet Dev Biol 1992;12:3-12.through gap junctions, 1~1 and then subsequent bio- 23. Pattee HH. Hiera~chy theory: the challenge of complex systems. New York:logic computation in an osseous CCN. G.Baziller, 1973. 24. Goldsmith P. Plant stems: a possible model system for the transduction of mechanical information in bone modeling. Bone 1994;15:249-50. 25. French AS. Mechanotransduction. Ann Rev Physiol 1992;54:135-52.CONCLUSION 26. Kernan M, Cowan D, Zuker C. Genetic dissection of mechanoreception-defective Where the original FMH version offered only verbal mutations in Drosophila. Neuron 1994;12:1195-206. 27. Hamill OP, McBride DW Jr. Mechanoreceptive membrane channels. Am Scien-descriptions of periosteal matrix function and skeletal unit tist 1995;83:30-7.response, the addition to the FMH of the concepts of 28. Hackney CM, Furness DN. Mechanotransduetion in vertebrate hair cells: struc- ture and function of the stereociliary bundle. Am J Physiol 1995;268:C1-13.mechanotransduction and of computational bone biology 29. Fraser D J, Macdonald AG. Crab hydrostatic pressure sensors. Nature 1994;371:offers an explanatory chain extending from the epigenetic 383-4.event of skeletal muscle contraction, hierarchically down- 30. Olsson S, Hanson BS. Action potentiablike activity found in fungal rnycelia is sensitive to stimulation. Naturwissch 1995;82:30-1.ward, through the cellular and molecular levels to the 31. Cut C, Smith DO, Adler J. Characterization of mechanosensitive channels inbone cell genome, and then upward again, through histo- Eschericia colt cytoplasmic cell membrane by whole-cell patch clamp recording. Jlogic levels to the event of gross bone form adaptational Membr Biol 1995;144:31-42. 32. Wildron De, Thain JF, Minchin P, Gubb I, Reilly A, Skipper Y, et al. Electricalchanges. Analyzing size and shape changes by reference- signaling and systematic proteinase inhibitor induction in the wounded planLframe-invariant, finite element methods produces a more Nature 1992;360:62-5. 33. Mayer EA. Signal transduetion and intercellular communication. In: Walsh JH,comprehensive and integrated description of the totality Dockray G J, editors. Gut peptides: biochemistry and physiology. New York:of the processes of epigenetic regulation of bone form Raven Press, 1994:33-73.than previously possible. 34. Martin J. Coding and processing of sensory information. In: Kandel ER, Schwartz JH, Jessel TM, editors. Principles of neural science. 3rd. ed. New York: Elsevier, 1991:329-40. 35. Martin J, Jessel TM. Modality coding in the somatic sensory system. In: KandelREFERENCES ER, Schwartz JH, Jessel TM, editors. 3rd. ed. New- York: Elsevier, 1991:341-52. 36. Wang N, Butler JP, Ingber DE. Mechanotransduction across the cell surface and 1. Moss ML. The functional matrix. In: Kraus B, Reidel R, editors. Vistas in through the cTtoskeleton. Science 1993;260:1124-7. orthodontics. Philadelphia: Lea and Febiger, 1962:85-98. 37. Claassen DE, Spooner BS. Impact of altered gravity on aspects of cell biolo~, lot 2. Moss ML. Twenty years of functional cranial analysis. Am J Orthod 1972;61:479- Rev Ctyol 1994;156:301-72. 85. 38. van der Meulen MCH, Carter DR. Developmental mechanics determine long 3. Moss ML. Genetics, epigenetics and causation. Am J Orthod 1981;80:366-75. bone allometry. J Theor Biol 1995;172:323-7. 4. Herring S. Epigenetic and functional influences on skull growth. In: Hanken J, 39. Martin TJ, Ng KW. Mechanisms by which cells of the osteoblastic lineage control Hall BK, editors. The skull 1. Chicago: University of Chicago Press, 1993:153-206. osteoclast formation and activity. J Cell Biochem 1994;56:357-66. 5. Moss ML. Integration of the functional matrix hypothesis and the finite element 40. Hill PA, Reynolds JJ, Meickle Me. Osteoblasts mediate insulin-like growth method: a new paradigm for the analysis of craniofacial growth. Le Journal de factor-I and -II stimulation of osteoclast formation and function. Endocrinol lEdgewise 1987;15:7-54. 1995;136:124-31. 6. Moss ML. Finite element comparison of murine mandibular form differences. J 41. Moss-Salentijn L. The human tactile system. In: Nicholls HR, editor. Advanced tactile Craniofac Genet Devel Biol 1988;8:3-20. sensing for robotics. Chapter 4. Singapore: World Scientific Publishing, 1994. 7. Moss ML. Bone as a connected cellular network: modeling and testing. Ann 42. Cowin SC, Moss-Salentijn L, Moss ML. Candidates for the mechanosensory Biomed Eng 1991:117-9. system in bone. J Biomed Engineer 1991;113:191-7. g. Moss ML. Alternate mechanisms of bone remodeling: their representation in a 43. van der Laarse A, Ravelsloot JH, Neiweide PJ. Voltage, calcium and stretch connected cellular network model. Ann Biomed Engineer 1991;19:636. activated ionic channels and intracellular calcium in bone cells. J Bone Miner Res 9. Moss ML. Advances in finite element modeling of cephalic growth: the integra- 1992;7:$377-87. tion of macroelement and boundary element methods with the functional matrix 44. Guggino SE, LaJeuoesse D, Wagner JA, Snyder SH. Bone remodeling signaled by hypothesis. J Jpn Orthod Soc 1994;53:357-66. a dihydropyridine- and phenylalkylaminesensitive calcium channel. Proc Nat10. Moss ML, Cowin SC. Mechanotransduction in bone. In: Lanza R, Lallger R, Acad Sci 1989;86:2957-60. Chick W, editors. Textbook of tissue engineering. New York: Springer Verlag, 45. Duncan R, Misler S. Voltage-activated and stretch activated Ba 2+ conducting 1995 (in press). channels in an osteoblast-like cell line (URM 106). Fed Eur Biochem Soc11. Moss ML. Growth of the calvaria in the rat: the determination of osseous 1989;251:17-21. morphology. Am J Anat 1954;94:333-62. 46. Keynes RD. The kinetics of voltage-gated ion channels. Q Rev Biophys 1994;27:12. Moss ML. A functional analysis of human mandibular growth. Am J Prosthet 339-44. Dent 1960;10:1149-60. 47. Stockbridge LL, French AS. Stretch-activated cation channels in human fibro-13. Moss ML. The primacy of functional matrices in orofucial growth. Trans Br Soc blasts. J Biophys 1988;54:187-90. Stud Orthod Dent Pract 1968;19:65-73. 48. Sachs F. Biophysics of mechanoreception. Membrane Biochem 1986;6:173-95.14. Moss ML. Differential roles of the periosteal and capsular functional matrices in 49. Sachs F. Mechanical transduction in biological systems. CRC Rev Biomed orofacial growth. Trans Eur Orthod Soe 1969;45:193-206. Engineer 1988;16:141-69.15. Moss ML, Rankow R. The role of the functional matrix in mandibular growth. 50. Sackin H. Mechanosensitive channels. Ann Rev Physiol 1995;57:333-53. Angle Orthod 1968;38:95-103. 51. Hatter LV, Hruska KA, Duncan RL. Human osteoblast-like cells respond to
American Journal of Orthodontics and Dent@cial Orthopedics MOSS ~Volume 112, No. 2 mechanical strain with increased bone matrix protein production independent of gone expression and hypertrophy of cardiac myocytes. Proc Nat Acad Sci USA hormonal regulation. Endocrinol 1995;136:528-35. 1992;89:9905-9.52. Harrigan TP, Hamilton JJ. An analytical and numerical study of the stability of 87. Bennett MVL, Goodenough DA. Gap junctions: electronic coupling and inter- bone remodeling theories: dependence on microstructural stimulus. J Biomech cellular communication. Neurosci Res Prog Bull 1978;16:373-485. 1992;25:477-88. 88. Schirrmacher K, Schmitz I, Winterbager E, Traub O, Brummmer F, Jones D, et53. Lanyon LE. Functional strain as a determinant for bone remodeling. Calcif Tiss at. Characterization of gap junctions between osteoblastqike cells in culture. lntl 1984;36:$56-$61. Calcif Tiss lnt 1992;51:285-90.54. Lanyon LE, Rubin CT. Static vs dynamic loads as an influence on bone 89. Jones SJ, Gray C, Sakamaki H, Arora M. Boyde A, Gourdie R, et ak The remodeling. J Biomech 1984;17:897-905. incidence and size of gap junctions between bone cells in rat calvaria. Anat 55. Rubin LE, Lanyon LE. Limb mechanisms as a function of speed and gait: a study Embryol 1993;187:343-52. of functional strains in the radius and tibia of horse and dog. J Exp Zool 90. Gourdie R, Green C. The incidence and size of gap junctions between bone cells 1982;101:187-211. in rat calvaria. Anat Embryol 193;187:343-52. 56. Rubin LE, Lanyon LE. Regulation of bone formation by applied loads. J Bone Jt 91. Civitelli R. Cell-ceil communication in bone. Calcif Tiss Int I995;56:$29-31. Surg 1984;66A:397-402. 92. Minkoff R, Rundus VR, Parker SB, Hertzberg EL, Laing JG, Beyer E~ Gap 57. McLeod KJ. Microelectric measurement of low frequency electric fields effects in junction proteins exhibit early and specific expression during intramembranous cells and tissues. Bioelectromagnetics 1992;l(suppl):161-78. bone formation in the developing chick mandible..&nat Embryo[ 1994;190:231-41. 58. Chesnoy-Marchais D, Fritsch J. Voltage-gated sodium and calcium currents in rat 93. Rodan G. Introduction to bone biology. Bone 1992;13:$3-6. osteoblasts. J Physiol 1988;398:291-31l. 94. Curtis TA, Ashrafi SH, Weber DF. Canalicular communication in the cortices of 59. Massass R, Bingmann D, Korenstein R, Tetsch P. Membrane potential of rat human long bones. Anat Rec 1985;212:336-44. calvaria bone cells: dependence on temperature. J Cell Physiol 1990;144:1-11. 95. Doty S. Cell-to-ceil communication in bone tissue. In:Davidovitch Z, editor. "File 60. Jan LY, Jan YN. Tracing the roots of ion channels. Cell 19~2;69:715-8. biological mechanism of tooth eruption and root resorption. Birmingham: 61. Ravesloot JH, van Houten RJ, Ypey DL, Nijweide PJ. High-conductance anion EBSCO Media, 1989:61-9. channels in embryonic chick osteogenic cells. J Bone Miner Res 1991;6:355-63. 96. Nowak R. Cells that fire together, wire together. J NIH Res 1992;4:60-4. 62. Ferrier J, Crygorczyk C, Grygorczyk R, Kesthely A, Langan E, Xia SL. Ba2- 97. Bingmann D, Tetsch D and Fritsch J. Membraneigenschaf~.en yon Zellen ans induced action potentials in osteoblastic ceils. J Membrane Biol 1991;123:255-9. Knochenexplantaten. Z Zahnartzl lmplantol 1989;4:277-81. 63. Schirrmacher K, Brummer F, Dusing R, Bingmann D. Dye and electric coupling 98. Bingmann D, Tetsch D, Fritsch J. Membrane properties of bone cells derived between osteoblasts-like cells in culture. Calcif Tissue lnt 1993;53:53-60. from calvaria of newborn rats (tissue culture). Pfiugers Arch 1989;$412:R14. 64. Weinbaum S, Cowin S, Zeng Y. A model for the excitation of osteocytes by 99. Edelman A, Fritsch J, Balsan S. Short-term effects of PTH on cultured rat mechanical loading-induced bone fluid shear stresses. J Biomech 1994;27:339-60. osteoblasts: changes in membrane potential. Am J Physiol 1986;251:C483-90. 65. Gross D, Williams WS. Streaming potential and the electromechanical response 100. Ferrier J, Ward-Kesthely AW, Homble F, Ross S. Further analysis of spontaneous of physiologically moist bone. J Biomech 1982;15:277-95. membrane potential activity and hyperpolariztion response to parathyroid hor- 66. Pollack SR, Saizstein R, Pienkowski D. Streaming potentials in fluid filled bone. mone in osteoblast-like ceils. J Cell Physioi 1987;130:433-51. Ferroelectrics 1984;60:297-309. 101. Hohman EL, Elde RP, Rysavy JA, Einzig S, Gebhard RL. Innervation of 67. Turner CH, Forwood MR, Otter MW. Mechanotransduction in bone: do bone periosteum and bone by sympathetic vasoeonstrictive intestine peptidn-containing ceils act as sensors of fluid flow? FASEB J 1994;8:375-8. fibers. Science 1986;232:868-71.68. McLeod KJ, Rubin CT. Frequency specific modulation of bone adaptation by 102. Carvalho RS, Scott JE, Suga DM, Yen EH. Stimuladnn of signal transduction induced electric fields. J Theor Biol 1990;145:385-96. pathways in osteoblasts by mechanical strain potentiated by parathyroid hormone.69. Brighton CT, Okerehe E, Pollack S, Clark CC. In vitro bone~cell response to a J Bone Miner Res 1994;9:999-1011. capacitatively coupled electrical field: role of field strength, pulse pattern and duty 103. Edin BB, Trulsson M. Neural network analysis of the information content in cycle. Clin Orthop Rel Res 1992;285:255-62. population responses from human periodontal receptors. SPIE 1992;1710:257-6.70. McLeod KJ, Donahue HJ, Levin PE, Fentaine M-A, Rubin CT. Electric fields 104. Denning PJ. Neural networks. Am Sci 1992;80:426-9. modulate bone cell function in a density-dependent manner. J Bone Miner Res 1993;8:977-84. 105. Martino RL, Johnson CA, Sub EB, Trus BL, Yap TK. Parallel computing in71. Halliday NL, Tomasek JJ. Mechanical properties of the extracellular matrix biomedical research. Science 1994;265;905-8. influence fibronectin fibril assembly in vitro. Exp Cell Res i995;217:109-17. 106. Dayhoff J. Neural network architecture. New York: van Nostrand Reinhold, I990.72. Clark EA, Bgugge JS. Integrins and signal transduction pathways: the road taken. 107. Zorntzer SF, Davis J, Lau C. An introduction to neural and electronic networks, Science 1995;268:233-9. San Diego: Academic Press, 1990.73. Watanabe H, Miake K, Sasaki J. Immunohistochemical study of the cytoskeleton 108. Grossberg S. Neural networks and artificial intelligence. Cambridge: MIT Press, of osteoblasts in the rat calvaria. Acta Anat 1993;147:14-23. 1988.74. Hughes DE, Salter DM, Dedhar S, Simpson R. Integrin expression in human 109. Hinton GE, Anderson JA. Parallel models of associative memory. Hillsdale: bone. J Bone Miner Res 1993;8:527-33. Lawrence Erlbaum, 1989.75. Richardson A, Parsons JT. Signal transduction through integrins: a central role 110. McClelland JL, Rumelhart DE. Parallel distributed processing. In: Psychological for focal adhesion. Bioessays 1995;17:229-36. and biological models. 2 vet. Cambridge: MIT Press, 1987.76. Shapiro F, Cahill C, Malatantis G, Nayak RC. Transmission electron microscopic lll. Kupfermann I. Neural networks: they do not have to be complex to be complex. demonstration of vimentin in rat osteoblast and osteowtic cell bodies and Behav Brain Sci 1992;15:767-8. processes using the immunoblot technique. Anat Rec 1995;241:39-48. 112. Wasserman PD. Neural computation. In: Theory and practice. New York:77. Green PB. Connecting gone and hormone action to form, pattern and organo- Nostrand Reinhold, 1989. genesis: biophysical transducflons. J Exp Bota W 1994;45:1775-88(Special Issue). 113. Parfitt AM. Osteonal and hemi-nsteonal remodeling: the spatial and temporal78. Jones DB, Bingmann D. How do osteoblasts respond to mechanical stimulation? framework for signal traffic in adult human bone. J Cell Biochem 1994;55:273-86. Cells Mater t991 ;1:329-40. 114. Fritzke B. Growing cell structures: a self-organizing network for unsupervised and79. Machwate M, Juffienne A, Moukhtar M, Marie PJ. Temporal variation of c-fos supervised learning. Neural Networks 1994;7:1441-60. proto-oncogene expression during osteoblast differentiation and osteogenesis in 115. Carter DR. The regulation of skeletal biology by mechanical stress histories. In: developing rat bone. J Cell Biocbem 1995;57:62-70. SkaIak R, Fox CF, editors. Tissue engineering. New York: Alan Liss, 1988:173-9.80. Petrov AG, Usherwood PN. Mechanosensitivity of cell membranes: ion channels, 116. Kam E, Hodgins MB. Communication compartments in hair follicles and their lipid matrix and cytoskeleton. Eur Biophys J 1994;23:1-19. implication in differentiative control. Development 1992;114:389-93.81. Yanagishita M. Function of proteoglycans in the extracellular matrix. Acta Path 1!7. Kodoma R, Eguchi G. The loss of gap junctional cell-to-cefl communication is Jpn 1993;43:283~93. coupled with dedifferendation of retinal pigmented epithelial cells in the course of82. Uitto V-J. Extracellular matrix molecules and their receptors: an overview with transdifferentiation into the lens. Int J Dev Biol 1994;38:357-64. special emphasis on periodontal tissues. Crit Rev Oral Biol Med 1991;2:323-54. 118. Marrotti G, Ferretti M, Muglia MA, Palumbo C, Palazzini S. A quantitative83. Dayhoff JE, Hameroff SR, Lahoz-Beltra R, Swenberg CE. Intracellular mecha- evaluation of osteoblast-osteocyte relationships on growing endosteal surface of nisms in neuronal learning: adaptive models. Int Jt Cnnf Neural Networks rabbit tibiae. Bone 1992;13:363-8. 1992:173-8. 119. Horowitz SL, Shapiro HH. Modifications of mandibular architecture following84. Ingber DE. The riddle of morphogenesis: a question of solution chemistry or removal of temporalis muscle in the rat. J Dent Res 1951;30:276-80. molecular cell engineering. Cell 1993;75:1249-52. 120. Moss ML, Mos>Salentijn L. The muscle-bone interface: an analysis of a85. Haskin C, Cameron I. Physiological levels of hydrostatic pressure alter morphol- morphological boundary. Monograph 8, Craniot~acial Series. Ann Arbor: Center ogy and organization of cytoskeletal and adhesion protein!; in MG-63 osteosar- for Human Growth and Development, University" of Michigan:39-72. coma cells. Biochem Cell Bio/1993;71:27-35. 121. Harrigan TP, Hamilton JJ. Bone strain sensation via transmembrane potential86. Sadoshima J, Takahashi T, Jahn L, Izumo S. Roles of mechano-sensitive ion changes in surface osteobiasts: lnading rate and microstructnral implications. channels, cytoskeleton and contractile activity in stretch-induced immediate-early J Biomech 1993;26:183-200.
226 Moss American Journal of Orthodontics and Dentofacial Orthopedics August 1997122. Aarden EM, Burger EH, Nijweide PJ. Function of osteocytes in bone. J Cell 131. Turner CH. Functional determinants of bone structure: beyond WollFs law of Biochem 1994;55:287-99. bone transformation. Editorial. Bone 1992;13:403-9.123. Lanyon LE. Osteocytes, strain detection, bone modeling and remodeling. Calcif 132. Rubin CT, Donahue HJ, Rubin JE, McLeod KI. Optimization of electric field Tiss Int 1993;53:$102-6. parameters for the control of bone remodeling: exploitation of an indigenous124. Brown TD, Pederseu DR, Gray ML, Brand RA, Rubin CT. Periosteal remodel- mechanism for the prevention of osteopenia. J Bone Miner Res 1993;8:$573-81. ing: a combined experimental and analytic approach. J Biomech 1990;23:893-905. 133. Turner CH, Forwood MR, Rho J-Y, Yoshikawa T. Mechanical loading thresholds125. Cowin SC. Strain assessment by bone cells. In: Skalak R, Fox CF, editors. Tissue for lamellar and woven bone formation. J Bone Miner Res 1993;9:87-97. Engineering. New York: Alan R. Liss, 1988:181-7. 134. Moss ML. A theoretical analysis of the functional matrix. Acta Biotheor 1969;18:126. Cowin SC. Bone biomechanics. Boca Raton: CRC Press, 1989a. 195-202.127. Cowin SC. A resolution restriction for Wolffs law of trabecular architecture. Bull 135. Moss ML. Functional cranial analysis of the mandibular angular cartilage in the Hosp Jt Dis 1989;49:205-12. rat. Angle Orthod 1969;39:209-14.128. Rubin LE, McLeod KJ, Bain SD. Functional strains and cortical bone adaptation: 136. Rodrequez AA, Agre JC, Knudston ER, NG A. Acoustic myography compared to epigenetic assurance of skeletal integrity. J Biomech 1990;25:43-54. electromyography during isometric fatigue and recovery. Muscle Nerve 1993;16:129. Martin RB, Burr DB. Structure, function and adaptation of compact bone. New 188-92. York: Raven Press, 1989. 137. Moss ML. The functional matrix hypothesis and epigenetics. In: Graber TM,130. McLeod KL, Rubin CT. The effect of low-frequency electrical fields on osteo- editor. Physiologic principles of functional appliances. St Louis: CV Mosby, genesis. J Bone Jt Surg I992;74A:920-9. 1985:3-4. AVAILABILITY OF JOURNAL BACK ISSUES As a service to our subscribers, copies of back issues of the American Journal of Orthodontics and Dentofacial Orthopedics for the preceding 5 years are maintained and are available for purchase from the publisher, Mosby-Year Book, Inc., at a cost of $11.00 per issue. The following quantity discounts are available: 25% off on quantities of 12 to 23, and one third off on quantities of 24 or more. Please write to Mosby-Year Book, Inc., Subscription Services, 11830 Westline Industrial Dr., St. Louis, MO 63146-3318, or call (800)453-4351 or (314)453-4351 for information on availability of particular issues. If unavailable from the publisher, photocopies of complete issues are available from University Microfilms International, 300 N. Zeeb Rd., Ann Arbor, MI 48106 (313)761-4700.
American Journal of Orthodontics and Dentofacial Orthopedics Moss 339Volume 112, No. 3ple, Chiclid fish are polyphyodont (have continu- with the empirical data of animal breeders, it earlierously replacing dental sets) and can exhibit pro- provided a theoretical basis for certain human eu-nounced dental phenotypic plasticity.44 When the genic theories proposing reproductive inhibition forfish are fed on hard-shelled mollusks, the replacing individuals with "undesirable and genetically (chro-teeth are large and molariform, but when soft mosomally) regulated" medical and social condi-food is fed, those teeth are gracile, conical, and tions: a policy that eventually reached historicalnonmolariform. Experimentally in aquaria, the two genocidal depths. 56,57phenotypic states may be repeatedly and arbitrarily Later, the blending of the classical chromosomalalternated in succeeding dental generations by alter- and vertebrate paleontological disciplines creatednately changing the diets consistency. Because each the neo-Darwinian synthesis, a currently accepteddental replacement cycle involves identical odonto- paradigm of phylogenetic regulation. 58genic stages, it is postulated that (1) mechanical Recently, molecular (gene) genetics extendedforces, related to differential diet "hardness," gen- the claims of the thesis to the regulation of all aspectserate epigenetic signals, mechanotransductively pro- of ontogeny (i.e., of "growth and development").cessed by dental papilla cellsg.l°; and (2) these The mega-human genome project, 59,6°~61called "thesignals control at least the temporal and spatial ultimate triumph of genetics, 4s explicitly intends to:expression of genomic products related to the de- (1) describe the complete human genome; (2) dem-velopment of differential tooth form, such as size onstrate genomic controls of all developmental pro-and shape. 45-47 cesses, at all structural levels, from the subcellular to If the epigenetic/genomic dichotomy of odonto- the organismal; and, (3) in a societal context, possi-genetic regulation is unresolved, how much more so bly lead to some type of neoeugenics.the complex topic of cephalic morphogenesis where, Many human activities now are claimed to beparenthetically, mechanical loadings also play a genomically regulated: e.g., psychological behav-significant regulatory role. 15 ior6Z; personality63; alcohol and drug abuse64; chro- nobiological cyclic behaviors65; smoking, obesity,The Genomic Thesis alcoholism, drug abuse, food-binging--indeed any The genomic thesis holds that the genome, from attention-deficiency disorder, 66 among many others.the moment of fertilization, contains all the infor- The further suggestion of genomic control of intel-mation necessary to regulate (cause, control, direct) ligence generates prodigious, biomedical contro-(1) the intranuclear formation and transcription of versy in the social sciences and politics. 67 And notemRNA and (2) importantly, without the later addi- the frequent popular press reports of the "discov-tion of any other information, to regulate also all of ery" of yet another "gene" that "controls" yet an-the intracellular and intercellular processes of sub- other developmental, physiological, psychological,sequent, and structurally more complex, cell, tissue, or sociological event, process, or state.organ, and organismal morphogenesisa2,48: suc- The Biologic Bases for the Genomic Thesiscinctly, "all (phenotype) features are ultimately de-termined by the DNA sequence of the genome. 49 While comprehensively considered else- In this thesis, morphogenesis is but the prede- brief review is useful. The somatic w h e r e , 48,49,53 atermined reading-out of an intrinsic and inherited cells of an individual metazoan inherit two classes ofgenomic organismal blueprint 48495°5152 where, in molecular information: (1) an identical diploidaddition to molecular synthesis, the genome also DNA and (2) the maternal cytoplasmic constituentsregulates the geometric attributes of cell, tissue, of the egg: e.g., mitochondria, cytoskeleton, mem-organ, and organismal size, shape, and location. For branes. Only approximately 10% of the genomeexample, "specific patterns of gene regulation seems related to phenotypic ontogenesis, whereas(cause, control, regulate, determine) the mecha- the human genome has approximately 100,000nisms by which a fertilized egg divides and genes, "well over 90% ... does not encode precur-progresses through the various decision points to sors to mRNAs or any other RNA. 53 With regardyield groups of cells that are first determined to to individual phenotypic structural attributes, whilebecome and then actually differentiate to become all somatic cells commonly share approximatelyspecialized tissues of the right dimension and in the 5000 different polypeptide chains, each specific cellproper location. s3 type is characterized only by approximately 100 The genomic thesis originated with classical specific proteins. And it is claimed that "these(chromosomal) Mendelian genetics, s4,55 Combined quantitative (protein) differences are related to dif-
340 Moss American Journal of Orthodontics and Dentofacial Orthopedics September 1997ferences in cell size, shape and internal architec- craniofacial development is controlled by two inter-ture. s3 related, temporally sequential, processes: (1) initial The encoding 10% of the DNA exists in two regulatory (homeobox) gene activity and (2) subse-families; the vastly preponderant "housekeeping" quent activity of two regulatory molecular groups:genes and the nonabundant "structural" genes. The growth factor families and steroid/thyroid/retinoicformer regulate the normal molecular synthesis of acid super-family. For example, "homeobox genesagents involved in (1) the common energetic (met- coordinate the development of complex craniofacialabolic, respiratory) activities of all cells and, (2) the structures" and in "both normal and abnormal de-specific activities of special cell types (e.g., neurons, velopment, much of the regulation of the develop-osteoblasts, ameloblasts etc.). 52,68 ment of virtually all of the skeletal and connective These genes also regulate the synthesis of the tissue of the face is dependent on a cascade ofspecific molecular gene products, whose presence, overlapping activity of homeobox genes. 12absence, or abnormal molecular configuration are It is claimed that regulatory molecules can (1)associated with the (human) pathologic conditions "alter the manner in which homeobox genes coor-said to have a unitary genetic cause--the so-called dinate cell migration and subsequent cell interac-Mendelian disorders and the "single-gene disorders tions that regulate growth" and (2) be involved inwith nonclassic inheritance, 52 such as Marfan syn- the "genetic variations causing, or contributing to,drome, achondroplasia, osteogenesis imperfecta, the abnormal development of relatively commonand Duchenne muscular dystrophy, among many craniofacial malformations . . . perhaps modifyingothers. 52 For some, such "disorders provide the Hox gene activity. 52model on which the program of medical genetics is Specific orthodontic implications of the genomicbuilt. 59 In such conditions the absence of a normal thesis include claims that "poorly coordination-type, or the presence of a structurally abnormal ordinated control of form and size of structures, ortype, of a specific biochemical or molecular struc- groups of structures (e.g., teeth and jaws) by regu-tural entity is sufficient to initiate the cascade of lator genes should do much to explain the verysubsequent abnormal developmental pathways, frequent mismatches found in malocclusions andeventuating in a specific pathological state. other dentofacial deformities." And "single regula- A physical analogy is the construction of a tory (homeobox) genes can control the developmentbuilding wall where either the proportions of the of complex structures.., indicating that single genesconcrete are incorrect or an insufficient number of can determine the morphology of at least somemetal reinforcing rods are used. In both cases, complex structures," including "how characteristiceventual structural collapse is possible. Substitution noses or jaws are inherited from generation toof intercellular proteoglycans, and of collagen generation. s2fibrils, provides a corresponding skeletal tissue anal-ogy. Here, alterations in the genomically regulated Critical Definitionsprocesses of molecular synthesis can produce an Clarification of this dichotomy is assisted byeventual "structural collapse" at the hierarchically defining the present use of four terms: epigenetics,higher level of a macroscopic bone. Anticipating an hierarchy, emergence, and causation.antithesis, note here that the claim of genomic Epigenetics. Several millennia ago epigenesis de-control of the molecular syntheses underlying the scribed the process(es) by which increasing struc-formation of such elemental (molecular) skeletal tural complexity gradually arose from an originallytissue "building blocks" does not substantiate the unstructured mass, for example the stages of in vivofurther claim that the genome regulates the growth chick development or the gradual appearance of aand development (the size, shape, location and histo- pattern during weaving on a loom. 7s-81 Over time,logical composition) of the gross anatomical bone. many alternate, often differing, definitions ap- peared. 22,82 Earlier, they were macroscopic in scaleThe Genomic Thesis in Orofacial Biology and considered only the extrinsic, extraorganismal There is extensive support for the genomic thesis environment, such as food, light, temperature, andin the orofacial biology literature, with most genetic radiations. 83 Nineteenth century physiology addedstudies of cephalic or cranial morphogenesis explic- the intrinsic, intraorganismal milieu interieur, s4 suchitly or implicitly assuming genomic regulation of as hormones, blood gases, nutrients, and ions.each anatomical structure. 69-77 Epigenetics, as defined here, includes (1) all of A characteristic article 12 claims that prenatal the extrinsic (extraorganismal) factors impinging on
American Journal of Orthodontics and Dentofacial Orthopedics Moss 341Volume 112, No. 3vital structures, including importantly mechanical mandibular angular process of a given 14-year-oldloadings and electroelectric states and (2) all of the male? The genomic thesis holds that this processintrinsic (intraorganismal) biophysical, biomechani- was predetermined; i.e, that individuals osteoblasticcal, biochemical, and bioelectric microenvironmen- genome contained, at the moment of fertilization,tal events occurring on, in, and between individual all the information necessary to regulate where,cells, extracellular materials, and cells and extracel- when, for how long, in what direction, in whatlular substances. amount, and at what rates, bone formation and Hierarchy. Biological structures are hierarchically remodeling will occur in that individual, given theorganized, with structural and functional complexity absence of disease and the presence of the usual andincreasing "upward" from the ever-expanding family necessary extrinsic (environmental) factors, such asof subatomic particles to protons, electrons, atoms, adequate nutrition, and the customary normal phys-molecules, subcellular organelles, and on to cells, iological states, such as are presumed to exist intissues, organs, and organisms. 4s While a genomic physiologys hypothetical normal human.thesis claims that each higher level is achieved by the The antithesis (and the FMH) suggests thatpredetermined activity of the genomic information, epigenetic stimuli, created by operations of relatedan epigenetic antithesis suggests that hierarchical functional matrices and their skeletal unit adaptivecomplexity results from the functioning of epi- responses, create the "new" information sequen-genetic processes and mechanisms, 3° as described tially, as mandibular ontogenesis proceeds. 9,1° Allin the disciplines of developmental mechanicsy ,86 ontogenesis exhibits developmental "cascades," withself-organization, 87 complexity, and chaos, 88,89,9°,91 multiple branching points where decisions are madeamong others,--topics considered further in the between alternate developmental pathways. Suchfollowing epigenetic antithesis. decisions are not predetermined by encoded genetic Emergence. This phenomenon occurs in all nat- information, but instead are responses to someural hierarchies. It consists of the appearance, at epigenetic stimulus(i). Hierarchy, emergence, andeach successively higher and structurally and/or causation are topics of the greatest significance inoperationally more complex level, of new attributes any critique of the genomic hypothesis, because theor properties, not present in the lower levels, whose scope and content of molecular genetics is preciselyexistence or functions could not in any way be that; it deals with only the molecular level of struc-predicted, even from a complete knowledge of all of tural organization. The genomic hypothesis pro-the attributes and properties of any or all of the poses no pathways from molecules to morphogene-preceding lower organizational levels. 92-94 sis? ° Customarily, in craniofacial literature, the For example, full knowledge of all the attributes existence of two "facts" is stated: (1) that at theand properties of an osteocyte does not permit molecular level, a particular gene (or group ofprediction of the attributes and properties of any genes) exists and (2) that at some higher, macro-type of bone tissue. And full knowledge of all scopic level, some clinical state of normal growthattributes and properties of all constituent bone and development or of malformation and/or mal-tissue types does not permit prediction of the form function is observed. Without positing any specific(size and shape), growth, or functions of a macro- mechanisms or processes at each intervening hier-scopic "bone." archical level of the developmental cascade, it is Emergence is not genomically controlled. In- simply stated that fact 1 is the cause of fact 2. Forstead, the integrated activities of all the attributes in example, "it is demonstrated that synpolydactyly, ana given hierarchical level self-organize to produce inherited human abnormality of the hands and feet,the next higher level of complexity. In every real is caused [italics mine] by expansions of a polyala-sense, biologic structures "build" themselves; that is, nine stretch in the amino-terminal region ofbones do not grow, they are grown. Epigenetic HOXD13. 97processes and mechanisms are regulatory (causal) of In the genomic thesis morphogenesis is reducedhierarchical organization and of emergence and to molecular synthesis.self-organization. 95 T h e Classification of C a u s a t i o n 1t Causation. From this vast topic, 96 we consideronly how the attributes of a given biologic structural There are four principal causes of ontogenesis:level "cause" (control, regulate, determine) the at- material (with what?), formal (by what rules?),tributes of the next higher level. For example, what efficient (how?), and final (why?). These may becauses osteogenesis on the ectofacial surface the left categorized as either intrinsic (material and formal)
342 Moss American Journal of Orthodontics and Dentofacial Orthopedics September 1997and extrinsic (efficient); final cause (teleology) is not disks, and papers. The formal cause is the software:considered further. Of importance, both material a specific word processing program, both its appar-and formal causes are classified as prior causes, i.e., ent, user-friend form and, in reality, its ultimateexisting before the creation of some specific state or expression in machine language code. No combina-structure. Efficient cause is proximate; i.e., its oper- tion of hardware and software could ever write anation immediately causes the creation of a new state article. Extrinsic, epigenetic input is required, i.e.,or attribute. Material and formal causes are intrinsic the composition and input of the text itself. Bothbecause they reside within vital structure (either intrinsic causes must be present before (prior to) theintracellularly or intercellularly); efficient causes are textual input, whereas the extrinsic, epigenetic typ-extrinsic--they represent the entire spectrum of ing is immediately (i.e., proximately) followed byepigenetic processes, mechanisms, and events capa- creation, on the hard disk, of the text itself.ble of being imposed on vital structures. Both prior (intrinsic) and proximate (extrinsic) In biology, material cause is represented by all causes are necessary causes; neither alone is athe levels of cellular and intercellular materials, sufficient cause for the creation of this manuscript.without reference to any specific structural (anatom- Only the two integrated together furnish the neces-ical) arrangement. Formal cause is the genomic sary and sufficient cause.code, i.e., a series of "rules" or "laws." These act at In ontogenesis, genomic (intrinsic, prior) andthe at the molecular level to regulate the initial epigenetic (extrinsic, proximate) factors are each acreation of the constituents of material cause. Effi- necessary cause, but neither alone is a sufficientcient cause(s) are the epigenetic factors, as defined cause. Only the interaction of both provides bothabove, whose actions immediately regulate the next the necessary and sufficient cause of morphogene-developmental branching point. sisJ 1 This conclusion foreshadows the resolving A metaphor is helpful. Consider the use of a synthesis of this dichotomy, presented in the com-computer to prepare this manuscript. The material panion article, which also contains the comprehen-cause is the hardware: the computers, printers, sive bibliography. AVAILABILITY OF JOURNAL BACK ISSUES As a service to our subscribers, copies of back issues of the American Journal of Orthodontics and Dentofacial Orthopedics for the preceding 5 years are maintained and are available for purchase from the publisher, Mosby-Year Book, Inc., at a cost of $11.00 per issue. The following quantity discounts are available: 25% off on quantities of 12 to 23, and one third off on quantities of 24 or more. Please write to Mosby-Year Book, Inc., Subscription Services, 11830 Westline Industrial Dr., St. Louis, MO 63146-3318, or call (800)453-4351 or (314)453-4351 for information on availability of particular issues. If unavailable from the publisher, photocopies of complete issues are available from University Microfilms International, 300 N. Zeeb Rd., Ann Arbor, MI 48106 (313)761-4700.