• Like

Loading…

Flash Player 9 (or above) is needed to view presentations.
We have detected that you do not have it on your computer. To install it, go here.

doi:10.1016/S1470-2045(04)01382-8

  • 455 views
Uploaded on

 

  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Be the first to comment
    Be the first to like this
No Downloads

Views

Total Views
455
On Slideshare
0
From Embeds
0
Number of Embeds
0

Actions

Shares
Downloads
5
Comments
0
Likes
0

Embeds 0

No embeds

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
    No notes for slide

Transcript

  • 1. Genomics and brain-tumour drug resistance Review Genomics-based hypothesis generation: a novel approach to unravelling drug resistance in brain tumours? Markus Bredel, Claudia Bredel, Branimir I Sikic No currently available chemotherapy seems likely to substantially improve outcome in most patients with brain tumours. Several resistance-associated cellular factors, which were discovered in other cancer models, have also been identified in brain tumours. Although these mechanisms play some part in resistance in brain tumours, they are not sufficient to explain the poor clinical response to chemotherapy. There could be other brain-tumour-specific genetic profiles that are associated with tumour sensitivity to chemotherapy. There is increasing awareness that drug resistance in brain tumours is not a result of changes in single molecular pathways but is likely to involve a complex network of regulatory dynamics. Further insights into chemoresistance in brain tumours could come with comprehensive characterisation of their gene expression, as well as the genetic changes occurring in response to Figure 1. Images of a cDNA microarray slide profiling gene expression. chemotherapy. Recent progress in high-throughput bioanalytical methods for genome-wide studies has made usefulness of recent progress in high-throughput possible a novel research model of initial hypothesis bioanalytical methods for genome studies as tools to identify generation followed by functional testing of the generated determinants of response and to characterise the genetic hypothesis. changes associated with chemotherapeutic perturbation in brain neoplasms (figure 1). Lancet Oncol 2004; 5: 89–100 Drug resistance in human cancer Despite the well-established use of chemotherapeutic Most cancers have heterogeneous cell populations. Goldie approaches in the treatment of malignant brain tumours, and Coldman3 postulated a mathematical model for prognosis has not improved much over the past 20 years relating drug sensitivity of human tumours to gene and continues to be dismal for the majority of patients. mutation rate and hypothesised that tumours undergo Although chemotherapy prolongs survival for some types spontaneous genetic changes that enable development of of brain tumours, such as medulloblastoma, primitive resistance to cytotoxic agents to which the tumours have neuroectodermal tumour, oligodendroglioma, germ-cell never been exposed. Selection of mammalian cells in vitro tumour, and primary central-nervous-system lymphoma, for resistance to cytotoxic agents via exposure to for most histological types chemotherapy is applied as a last incrementally increased sublethal drug concentrations resort rather than as an established beneficial component commonly results in cross-resistance to many other drugs, of a multimodality treatment regimen. The most common which share little structural similarity with the primary type of brain tumour, the large group of high-grade selective agent and act at different intracellular targets. This gliomas, tends to be resistant to chemotherapy, and long- pleiotropic process, which is a major impediment to term tumour control is rarely achieved. Chemotherapy for brain tumours poses a special MB is Assistant Professor of Experimental Neurooncology, Department of General Neurosurgery at the Neurocenter, University challenge owing to the existence of a blood–tumour barrier, of Freiburg, Germany; and visiting Assistant Professor, Division of which is intact in those tumour regions that are biologically Oncology, Stanford University School of Medicine, CA, USA. CB is and clinically most important—namely in areas of the brain a molecular biologist in the Department of General Neurosurgery at infiltrated with tumour cells. However, there is growing the Neurocenter, University of Freiburg, Germany. BIS is Professor consensus that, in addition to difficulties related to the of Medicine (Oncology and Clinical Pharmacology) at Stanford University School of Medicine, CA, USA. blood–tumour barrier and pharmacokinetic issues, the Correspondence: Dr Markus Bredel, Division of Oncology, modest effect of chemotherapy in brain tumours is mainly Stanford University School of Medicine, 269 Campus Drive, linked to tumour-cell resistance as shown by certain genetic CCSR-1110, Stanford, CA 94305-5151, USA. Tel: +1 650 498 6949. and epigenetic factors.1,2 This review discusses the potential Fax: +1 510 438 8830. Email: mbredel@stanford.edu THE LANCET Oncology Vol 5 February 2004 http://oncology.thelancet.com 89 For personal use. Only reproduce with permission from The Lancet.
  • 2. Review Genomics and brain-tumour drug resistance treating patients with cancer, has become known as agents have mutagenic potential and could cause mutations multiple-drug (or multidrug) resistance (MDR). in key cellular target genes (genetic level); and chemo- Resistance mechanisms are expressed constitutively therapy can cause surviving cells to induce the coordinated either as genes normally expressed by the tissue of origin of expression of protective stress response genes (epigenetic the tumour (eg, MDR1 in colon cancer) or as genetic level). The resistant clones formed could be selected by alterations during tumorigenesis (eg, p53 mutations). Such subsequent therapy and expand, eventually leading to intrinsic (or de novo or constitutive) resistance results in disease relapse (figure 2). little treatment response or failure of initial treatment (figure 2). In many tumours, early drug treatment can Drug-resistance in brain tumours achieve substantial cell killing, though there might be Several mechanisms of drug resistance discovered in other selection of a clonal variant of cells that confers acquired tumour models have also been implicated in brain tumours resistance to subsequent treatment, leading via clonal (figure 3). They include ATP-dependent efflux of cytotoxic expansion, to repopulation and tumour recurrence (drug agents by transmembrane transporter proteins encoded by the selection; figure 2). Two principles help to explain the genes multiple-drug resistance 1 (MDR1, ABCB1) and acquired resistance phenotype (figure 2). First, the inherent multidrug-resistance-associated protein (MRP1, ABCC1); genetic instability of tumour cells can lead pre-existing DNA damage caused by quantitative changes in expression resistant cell clones that are present before initial therapy to of DNA topoisomerase II ; increased detoxification of expand under long-term chemotherapy. Second, drug alkylating agents by glutathione and the glutathione-linked resistance can be acquired through induction of resistance enzyme system, particularly glutathione-S-transferases; and pathways in cancer cells during chemotherapy. Anticancer increased activity and expression of members of the protein kinase C family, causing changes in Anticancer drug treatment cell-cycle transition and in the expression status of various resistance markers that are relevant to drug action.1,2 In addition, a deficiency in Initial response No response pathways of DNA mismatch repair (which render tumour cells tolerant to methylation) and increased nucleotide Remission Relapse excision repair of DNA adducts owing to altered activity of poly(ADP)-ribose polymerase and the product of the Clonal selection and expansion by prolonged therapy excision-repair cross-complementing rodent repair deficiency gene 2 (ERCC2) could be involved in the Resistant clones MDR of brain-tumour cells.1,2 Since most chemotherapeutic agents kill tumour cells via apoptosis, dysfunction Mutagenesis Epigenetic stress of genes involved in apoptotic response pathways and resultant impaired ability to commit to apoptosis can also contribute to chemotherapy resistance First drug exposure of brain-tumour cells (figure 3). In addition to MDR phenotypes, resistance to chemotherapeutic drugs Pre-existing can affect single agents or a class of resistant clones related drugs that share structural similarity. This type of resistance, which is broadly referred to as Inherent genetic instability Intrinsic individual drug resistance and (epi-)genetic has also been associated with resistance profile to some chemotherapies in brain tumours (figure 3), might be caused by raised concentrations of enzymes Acquired De novo involved in intracellular drug metabolism, for example O6-methyl- guanine-DNA methyltransferase Drug sensitivity Anticancer drug resistance (MGMT), thymidylate synthase, and dihydrofolate reductase.1,2 Further- Figure 2. Relation between drug effect and drug resistance/sensitivity of human cancers. more, metallothioneins can function to 90 THE LANCET Oncology Vol 5 February 2004 http://oncology.thelancet.com For personal use. Only reproduce with permission from The Lancet.
  • 3. Genomics and brain-tumour drug resistance Review Drug Brain tumour blood vessel lumen Brain tumour cell Pgp MRP Pgp GSH MT PKC GST GSH Nucleus Basal membrane M DNA G1 G2 repair S TII dUMP Apoptosis MGMT Lethal mitosis TS Alkyl dTMP DHFR DNA damage G Apoptopic T Components DNA Anti-apoptopic Pro-apoptopic Figure 3. Subcellular location and possible mechanism of action of previously identified resistance markers in brain tumours. The MDR1-encoded P-glycoprotein (Pgp) and members of the multidrug-resistance-associated protein (MRP) family act at the cell membrane as ATP-driven drug-efflux pumps. Glutathione-S-transferase (GST) mediates conjugation of anticancer agents with reduced glutathione (GSH). MRP1 mediates excretion of glutathione–drug conjugates. Quantitative changes in topoisomerase II (TII ) can withdraw the cellular target of agents targeting this enzyme. Nuclear O6-methylguanine-DNA methyltransferase (MGMT) may act to remove drug-induced O6-alkylguanine-DNA. Increased expression of nuclear thymidylate synthase (TS) and dihydrofolate reductase (DHFR), which are involved in the conversion of deoxyuridine monophosphate (dUMP) to deoxythymidine monophosphate (dTMP), can result in maintenance of free enzyme when challenged with drugs that deplete these enzymes. Metallothioneins (MT) can inactivate drugs either by chelation or by intracellular sequestration. Protein kinase C (PKC) induces a cell-cycle arrest to allow repair of drug-induced DNA damage before the cell enters lethal mitosis, can modulate the expression of a wide variety of resistance markers, and influences apoptotic pathways. Changes in the balance of proapoptotic and antiapoptotic factors prevent drug-induced apoptosis. The yellow arrows indicate increased or decreased marker expression associated with drug resistance. inactivate anticancer drugs in brain-tumour cells. In The traditional research model has dictated that when a oligodendroglioma, loss of heterozygosity (LOH) on gene is newly found to be related to drug resistance in one chromosomes 1p and 19q, a unique cytogenetic profile, is type of tumour, the hypothesis that the same gene is related to responsiveness to chemotherapy. important in drug resistance in brain tumours is formulated and experiments are done to test this Importance of the hypothesis-driven research hypothesis. Although some advances have been made with model this approach, others are needed to elucidate the The presence of resistance factors in a subset of brain polygenetic basis of resistance to chemotherapy in brain tumours has generally been associated with the prevalence tumours (figure 4). of clinical resistance to chemotherapy. However, Many known mechanisms of drug resistance are substantial uncertainty remains about whether the expressed constitutively in brain tumours and could be the expression of these factors in brain tumours is the cause of cause of intrinsic resistance to therapy.1,5 However, other the low response to chemotherapy in patients with brain mechanisms are likely to be either induced by drug tumours. Apart from the DNA-repair gene MGMT—for exposure or selected as mutations that occur during the which expression and hypermethylation-induced gene evolution of a tumour-cell population. The observation of inactivation have been related to response of brain tumours intrinsic expression in brain tumours could also reflect the to nitrosoureas1,2,4—no causal link between these markers in known physiological role of some of these genes in non- brain tumours and failure of chemotherapy has been found. neoplastic brain, for example the contribution of the MDR1 THE LANCET Oncology Vol 5 February 2004 http://oncology.thelancet.com 91 For personal use. Only reproduce with permission from The Lancet.
  • 4. Review Genomics and brain-tumour drug resistance (ABCB1) gene encoding P-glycoprotein to the formation of expression is regulated and by which cells adjust to varying a blood–brain barrier. The dual roles of other genes, such as conditions is through variation of the abundance of mRNA those for thymidylate synthase, dihydrofolate reductase, present in the cell. Accordingly, a primary event in the and topoisomerase II , in drug resistance and cell development of tumour-cell resistance could be a change in proliferation raise uncertainty over whether changes in degree of gene transcription. Analysis of the pattern of expression of some of these genes in brain tumours are variation in expression of genes could therefore be useful in related to drug resistance, cell-cycle turnover, or both. assessment of variation in the resistance characteristics of Changes in the expression state of some of these genes tumour cells and tissues (figure 4). Pharmacogenomics is might be completely independent of any role in drug predicated on the concept that cellular responses to resistance. For example, the locus for topoisomerase II is anticancer drugs are analogous to a higher gated logic close to that of the oncogene HER2/neu on chromosome circuit, in which, in some cases, many contradictory inputs 17q. Overexpression of HER2/neu has been shown in a modulated by features such as feedback, feed-forward, subset of brain tumours,6 and the accompanying increased error checking, and redundancy, are summed to produce a expression of topoisomerase II might be a consequence of response (figure 5). malignant transformation in these neoplasms. Genomics-based generation of drug-resistance The search for brain-tumour type-specific hypotheses resistance fingerprints There is increasing recognition of the value of Evidence for the existence of a tumour type-specific comprehensive approaches to the molecular characteri- resistance profile to a certain anticancer agent in the form sation of biological phenotypes such as drug resistance. of a resistance “fingerprint” comes from recent This appreciation has highlighted the need for investigations that have used large-scale gene expression biotechnology that allows parallel, large-scale assessment of profiling and have shown substantially differing genetic many genes. The emergence of high-throughput genomics response patterns to a defined chemotherapeutic agent for platforms has enabled genome-wide studies of gene distinct histological types of cancer cells.7–10 There is expression. The key feature of successful genomic growing awareness that drug resistance in human cancer is characterisation of a drug-resistance state is the ability to probably dictated in a combined way by the abnormal measure changes in mRNA abundance accompanying the expression of groups of coregulated genes; this association formation of this state, or differences between sensitive and suggests that many more genes are involved in the resistant phenotypes. development of resistance phenotypes in brain tumours Various strategies have emerged for gene expression than the gene expression changes described thus far for a profiling, including serial analysis of gene expression limited number of known resistance genes (figure 5). (SAGE), differential display, subtractive hybridisation The phenotypes of benign and malignant cells and approaches, and massively parallel signature sequencing. tissues ultimately depend on the types and amounts of Microarray technology that uses densely spotted cDNAs or proteins present at any given time. Translational and post- oligonucleotides allows the large-scale analysis of gene translational biochemical modifications exert decisive expression in a high-throughput way (figure 1). There are effects in regulating the amounts of active forms of proteins two main types of microarray systems. In the single-colour involved in pathological and pharmacological conditions. Affymetrix system, short oligonucleotide probes are However, a primary mechanism by which protein synthesised in situ on a chip. Each gene is represented by a Traditional Research paradigm Genomics-based Observation in Hypothesis Genomics other tumours generation ? Hypothesis Brain tumour resistance phenotype formulation ? Hypothesis testing Empirical reductionism Resistance mechanism Functional genomics Figure 4. Traditional versus genomics-based research approaches. 92 THE LANCET Oncology Vol 5 February 2004 http://oncology.thelancet.com For personal use. Only reproduce with permission from The Lancet.
  • 5. Genomics and brain-tumour drug resistance Review Response to chemotherapy Tumour type-specific Drug-specific co-regulation Polygenetic network co-alteration Gene expression Gene copy number (including contradictory inputs) Clinical Microarray tumour response Array CGH genome-wide mRNA abundance Genomic reponse fingerprint genome-wide DNA copy numbers Figure 5. The molecular determination of tumour response to chemotherapy, which includes in some cases many contradictory inputs, is a feature of the genome that might be determined by the expression and copy number of many known and unknown genes. Two complementary strategies could be useful in assessing molecular response fingerprints. First, gene expression profiling with microarrays allows characterisation of genome-wide mRNA abundance. Second, microarray-based comparative genomic hybridisation (array CGH) enables assessment of DNA copy numbers throughout the genome. set of probe pairs that correspond to different gene regions. examined the sensitivity state of various established and Each probe pair consists of an exact-match base-pair low-passage parental cancer cell lines.26–28 Scherf and sequence probe and a mismatch sequence probe. Gene colleagues29 first integrated large databases on gene expression is measured absolutely by the brightness of expression and chemosensitivity. They linked the hybridisation to the multiple probes representing a gene or expression profiles of about 8000 unique genes assessed in a relatively to the hybridisation signal of a reference RNA on set of 60 human tumour cell-lines, termed the NCI60, with a separate chip. By contrast, cDNA microarrays use a dual- the activity patterns of more than 70 000 anticancer drugs colour hybridisation scheme for the comparison of gene obtained from a database of the Developmental expression in two or more samples. Typically, mixtures of Therapeutics Program (DTP) of the US National Cancer (m)RNA reverse-transcription-generated Cy5-labelled Institute. Many gene–drug relations were identified. Similar “red” cDNA from one condition (eg, resistant or sensitive analysis in the NCI60 by Staunton and co-workers30 with cell line) and mixtures of Cy3-labelled “green” cDNA from 6817-gene microarrays revealed gene-expression-based another condition (eg, reference sample) are combined and classifiers of sensitivity or resistance for 232 compounds. allowed to hybridise to the glass microarray slide (figure 1). Butte and colleagues31 identified gene regulatory networks The common reference control—for example, a pooled determining chemotherapeutic susceptibility of the NCI60 sample of (m)RNAs from a set of cancer cell-lines— to 5084 agents by use of 7245-gene oligonucleotide arrays. generally has no biological significance and simply Dan and associates24 have used 9200-gene cDNA contributes a consistently measurable signal in the microarrays to compare the sensitivity state of 39 human denominator of the assayed ratios. The amount of red and cancer cell-lines to 55 anticancer agents. Pearson green fluorescence (R/G ratio) at each spot allows the correlation and clustering analysis identified genes with amount of (m)RNA for each gene in the test sample to be expression patterns that showed significant association with compared with that in the common reference sample, patterns of drug responsiveness. Some genes were avoiding the need for absolute measurements of (m)RNA commonly associated with response to various drug classes, (figure 1). whereas others were associated only with response to drugs An increasing number of studies have used microarray with a similar mechanism of action. Similarly, Zembutsu methods to examine gene expression patterns characteristic and colleagues20 analysed the expression profile of more of sensitive or drug-resistance phenotypes in various than 23 000 genes in a panel of 85 cancer xenografts, human cancers.7–25 Many of these studies have used the in derived from nine human organs and implanted into nude vitro resistance selection/induction model, which means mice, along with chemosensitivity to nine differing that a more or less sensitive parental cell line is exposed to anticancer drugs. This approach identified various genes (sub)lethal drug concentrations until a sufficient degree of that were significantly associated with sensitivity to one or resistance is achieved (figure 6). These studies have more drugs examined. provided preliminary insights into how changes in the In addition, gene expression profiles have been linked sensitivity state of a cancer cell are reflected in changes in to drug response in clinical tumour samples (figure 6). overall gene expression. Other genomic studies have Kihara and co-workers15 used a cDNA microarray THE LANCET Oncology Vol 5 February 2004 http://oncology.thelancet.com 93 For personal use. Only reproduce with permission from The Lancet.
  • 6. Review Genomics and brain-tumour drug resistance Histological tumour type Patient In vitro system In vivo system Established cell line Surgical tumour sample drug sample pair Resistance induction/ selection Post-chemotherapy Pre-chemotherapy sample sample Stable Transient Mock step-wise single-step treatment treatment treatment Low passage cell line Resistant Stressed Sensitive Laser microbeam Laser pressure daughter daughter parental microdissection catapulting cell line cell line cell line Purified resistant tumour cells Magnetic cell sorting cDNA microarray Purified sensitive tumour Distinct gene cells expression profiles Real-time Outlier genes quantitative PCR Biostatistic algorithms Pharmacogenomics Gene(s) of clinical interest Candidate marker validation Clinical tumour response data Candidate markers Figure 6. Genomics-based resistance research can pursue a dual strategy of assessing tumour (cell) responsiveness in vitro and in vivo. In vitro resistance research commonly involves the selection/induction model in which a parental cell line is selected/induced for resistance via chronic exposure to sublethal drug concentrations until a sufficient degree of resistance is achieved. Comparison of the gene expression of the parental line and the resistant subline can identify subsets of genes that are differentially expressed. Inclusion of stress response data by transient single-step drug exposure can help in dissecting gene expression changes primarily linked to drug resistance from secondary genetic changes that may be only downstream and marginally relevant to resistance formation. Ideally, genomics-based in vivo approaches to resistance research are combined with purification techniques that allow for enrichment of tumour cells. Laser-assisted tissue microdissection has emerged as the method of choice to limit tissue heterogeneity. Gene expression profiling in surgical tumour samples can use a dual strategy. On the one hand, the expression pattern of prechemotherapy samples is directly compared with that of postchemotherapy samples; on the other hand, with a pharmacogenomics strategy, expression data are directly correlated with clinical tumour response data to identify response-associated patterns of gene expression. containing 9216 genes to predict the sensitivity of drug response score based on differential expression of oesophageal tumours to adjuvant chemotherapy. They these genes was predictive of outlook for individual identified 52 genes that were linked to prognosis and patients. Sotiriou and associates17 investigated the response possibly to chemosensitivity and/or chemoresistance, and a to systemic chemotherapy of fine-needle aspirates of breast 94 THE LANCET Oncology Vol 5 February 2004 http://oncology.thelancet.com For personal use. Only reproduce with permission from The Lancet.
  • 7. Genomics and brain-tumour drug resistance Review cancers with expression profiles from 7600-feature cDNA Gene expression profiling identified 89 genes with microarrays. Candidate gene expression profiles were upregulated or downregulated expression. These included identified that distinguished tumours with complete the changed expression of genes related to various biological response to chemotherapy from those with no response. functions, including increased expression of several metallothionein genes and reduced expression of Application of genomics methods several proapoptotic genes. Although MGMT activity of one Gene expression profiling has been used for molecular resistant clone was twice that of the sensitive parental genetic characterisation, classification, and subclassification cells, a second clone selected with carmustine and of intracranial neoplasms, including meningiomas32 and O6-benzylguanine (an irreversible inhibitor of MGMT) did various histological types of glioma.33–46 Such strategies have not show changes in MGMT, which implies that other provided initial evidence to offer the promise of refined mechanisms must have had a role in the resistance prognosis prediction in brain tumours such as diffuse phenotype of these cells. astrocytoma,42 malignant glioma,37,47 and medulloblastoma.48,49 Baseline information about differences in chemo- Microarray analysis has also identified a potential serum sensitivity between genetic subsets of oligodendroglioma marker for glioblastoma multiforme.50 The potential value of has been provided by gene expression profiling. Mukasa microarrays as a tool for assessing variation in the response and colleagues53 used oligonucleotide microarrays with characteristics and resistance state of brain tumours is just more than 12 600 human genes and expressed sequence beginning to be appreciated. tags to examine 11 tumour specimens of two cytogenetic Rhee and colleagues51 characterised cellular pathways subsets of oligodendroglioma, (characterised by the involved in the response of glioblastoma multiforme cells to presence or absence of loss of heterozygosity [LOH] of the nitrosourea carmustine. Their array contained only 588 chromosome 1p), which show profoundly different genes, of which 17 had differential expression between a response rates to chemotherapy.1,2 The researchers carmustine-sensitive glioblastoma multiforme subline and identified 209 genes expressed differentially between the a resistant parental subline. Most (13) of these genes tumour subsets, with 86 genes showing higher expression showed lower expression in the resistant variant. and 123 lower expression in tumours with 1p LOH. Expression of MGMT, which was previously thought to be Notably, only 60% of the 123 genes with reduced the primary mechanism of resistance of glioblastoma expression in tumours with 1p LOH were located on multiforme cells to carmustine, was similar in both sublines chromosomes 1 and 19 (oligodendrogliomas with 1p LOH and not altered by carmustine treatment, which suggests commonly also show 19q deletion). that the MGMT pathway did not contribute to the carmustine resistance phenotype. By contrast, six other Potential pitfalls of comprehensive gene DNA repair genes, including ERCC2, were downregulated expression profiling strategies in the sensitive subclone. The strength of microarray technology in research on drug Bacolod and co-workers52 used microarrays with 12 000 resistance is the ability to assess the expression of many elements to study changes in gene expression accompanying genes at the same time and subsequently, by use of carmustine resistance selection in medulloblastoma cells. sophisticated computer-driven pattern recognition and Functional testing via conventional Multiple upregulated and Raw microarray data biochemistry for logical coherence downregulated genes Prioritisation of probable hypotheses hiearchical gene Many hypotheses clustering shaving self- relevance organising networks maps Bioinformatics Data mining for meaningful information Identification of co-regulated genes or gene families quality terrain threshold maps principal component clustering analysis Figure 7. Hypotheses generated from a microarray experiment require sophisticated data mining for meaningful information to be obtained. In general, coregulated genes and families of genes are identified via biostatistical approaches with the assumption that genes that are similarly altered in expression are likely to be functionally related. Functional testing by conventional biochemical approaches is essential for logical coherence and to ascertain a causal link between a resistance phenotype and altered gene expression. THE LANCET Oncology Vol 5 February 2004 http://oncology.thelancet.com 95 For personal use. Only reproduce with permission from The Lancet.
  • 8. Review Genomics and brain-tumour drug resistance map building, to cluster those that are simultaneously that genes may belong to more than one cluster59 (figure 7). upregulated or downregulated when a tumour becomes In addition, computational algorithms allow organisation chemoresistant (figure 7). The studies on drug resistance of upregulated and downregulated genes on the basis of that have used microarray methods provide evidence that their ontology. For example, sophisticated software can be genomic responses of a tumour cell to a chemotherapeutic used to view and analyse gene expression data according to agent can include altered expression of a substantial biological pathways and gene relations can be directly number of genes that might well be in the range of 5–10% explored and annotated.60 Ontology-based arrangement of of the human genome; a large proportion of these changes differential gene expression can provide immediate clues have been confirmed by traditional methods of measuring about the potential involvement of certain signalling and mRNA.7–9 In addition, the intrinsic sensitivity of a tumour biochemical pathways in drug resistance states of tumour might be reflected by the expression of several genes or a cells. group rather than by single genes. For example, in the A further problem of microarray technology is its low study by Zembutsu and colleagues,20 comparison of the sensitivity to transcriptional changes in the baseline range, gene expression profiles of 85 human cancer xenografts which complicates the identification of modest changes in with sensitivities to various drugs identified 1578 genes for gene expression that might be relevant to drug resistance. which degree of expression was significantly related to In addition, normal variability in gene expression could chemosensitivity; 333 of these genes showed significant lead to differential expression patterns of a gene in a associations with two or more drugs, and 32 genes were sensitive and resistant sample without reflecting the true linked with sensitivity to six or seven drugs. effect of the resistance formation process. Microarray technology provides a rapid way of In general, measurements of gene expression provide collecting large amounts of data and can generate new direct information on the abundance of the mRNA hypotheses on mechanisms leading to drug resistance. template only, rather than on the expression of the final However, potential problems with the application of this gene product, the protein. This point is particularly approach to drug resistance research have to be recognised. important, because concentrations of proteins can vary In particular, biostatistical methods have to be available to significantly among genes with similar abundance of analyse the substantial number of hypotheses that might be mRNA. Conversely, there can be substantial variation in generated from a microarray experiment (figure 7). This the amounts of mRNAs encoding proteins expressed with approach does not differentiate between genes that might similar abundance.61 At present, the efficiency of obtaining cause a certain resistance phenotype and secondary genetic an overall view of gene expression by measurement of changes that might be downstream, marginally relevant, or mRNA is better than that for similar proteomic methods, in even irrelevant to the resistance phenotype. In fact, a terms of throughput, reproducibility, and compatibility substantial proportion of transcriptional changes observed with clinical material. The degree of protein expression of during resistance formation probably do not immediately small numbers of genes potentially associated with response relate to the drug-resistant phenotype. The early changes in to therapy can be examined in a high-throughput way by resistance selection are likely to be broader than mere tissue-microarray-based immunohistochemistry, which resistance formation, representing a stress response of a allows rapid screening of large numbers of tumour tumour cell when exposed to a chemotherapeutic agent specimens for candidate marker validation on paraffin- rather than a targeted tumour-cell response to reduce its embedded tissue. The usefulness of tissue microarrays for vulnerability to this drug. neuropathology research has been readily demonstrated.62 Identification of sets of genes that concomitantly show Recent progress in the application of microarrays altered expression during development of drug resistance to cytogenetics—particularly comparative genomic could directly relate to regulation and coregulation for hybridisation—has led to chip-based genome-wide families of genes. Therefore, the most common approach to screening for changes in gene copy numbers.63 Since gene organisation of microarray data is hierarchal clustering54 dose effects are also relevant to chemosensitivity and (figure 7). It relies on the basic premise that genes with changes in gene copy number are frequently observed in similarly altered expression profiles are likely to be tumour cells exposed to chemotherapeutic agents, the coregulated and functionally related and uses a “guilt by application of array comparative genomic hybridisation to association” strategy to identify functional clusters. research on brain-tumour drug resistance might provide Hypotheses generated by such associations require valuable information on codeleted and coamplified genes confirmation by conventional biochemical approaches, to that are potentially important to resistance formation establish logical coherence between the resistance state of a (figure 5). brain-tumour cell and an altered state of expression of single genes or a group of genes (figures 4 and 7). In vitro drug resistance versus in vivo Alternative statistical approaches for data mining include chemoresistance the application of self-organising maps,55 the construction Experiments in tumour cell-lines are an important first step of relevance networks,27,31 principal component analysis,56 in characterising genomic changes linked to resistance terrain maps,57 quality threshold clustering,58 and a development in vitro. However, culture-derived patterns of mathematical strategy called “gene shaving” that differs gene expression changes accompanying resistance from hierarchical clustering in that it takes into account formation may not correlate with resistance-associated 96 THE LANCET Oncology Vol 5 February 2004 http://oncology.thelancet.com For personal use. Only reproduce with permission from The Lancet.
  • 9. Genomics and brain-tumour drug resistance Review microdissection have been developed. For example, laser capture micro- dissection66 uses a cap coated with a thermolabile film of ethylene vinyl acetate, which is placed in contact with a tissue section stained to enable selection of the relevant cell type. Visualisation is achieved with an inverted microscope. An applied focused laser beam of variable diameter causes localised melting of the film over selected cells and these cells fuse to the cap. The selected tissue is selectively removed when the cap is lifted. Several thousand targeted laser “shots” of material can be made on a single cap, thereby enabling concentration of the cell type of interest. A second technique67 avoids mechanical contact to capture samples of between 1 µm and several 10 m hundred micrometers in diameter. In this method, the laser is used not only Figure 8. Histological section of a high-grade glioma from a child, imaged by differential contrast for microbeam microdissection but and analysed by immunocytochemistry for expression of DNA topoisomerase II protein. The white also for transfer of the selected arrow points to a nucleus positive for expression of the protein. The black arrow points to a microdissected area where a single positive nucleus has been captured by laser pressure clusters of cells or single cells from the catapulting. object plane directly into the cap of a routine microfuge tube via high gene expression in solid brain tumours. Long-term brain- photonic pressure force—laser pressure catapulting (figure tumour cell lines differ to some extent from brain-tumour 8). The feasibility and reproducibility of this approach for tissue samples in their basic gene expression profile. Hess subsequent genetic analysis with no adverse effect on and colleagues64 showed that in hierarchical cluster analysis mRNA have been clearly demonstrated.68 of microarray data, three glioblastoma multiforme cell lines Laser microdissection has been used for reduction of formed one main cluster and ten glioblastoma multiforme tissue heterogeneity in gene expression profiling in tissue samples formed a separate cluster. Multidimensional mammalian brain69 and brain disease, including temporal- scaling and principal component analysis provided further lobe epilepsy70 and brain tumours.71 Many of the studies evidence that the cell lines clearly differed from the tissue that have combined microdissection and overall gene samples and that the cell lines differed genetically more expression analysis have used mRNA amplification to between themselves than the tissues did. Sasaki and achieve sufficient amounts of mRNA.72,73 Since amplifi- associates65 showed that gene expression of tissue-cultured cation protocols are prone to representative biases, other meningiomas and corresponding in situ meningiomas studies have shown that expression profiles for thousands differed significantly for a portion of genes. Accordingly, an of genes can be successfully generated with non-amplified integrative look at resistance-related changes in gene mRNA derived from clinical cancer specimens procured by expression in response to drug application necessarily has laser microdissection.74–76 to include actual brain tumour specimens (figure 6). One of the confounding factors in studying primary Perspective clinical specimens by microarrays is the mixture of With the perception that gene networks rather than neoplastic and various normal cell-types in the tumours. individual genes determine chemoresistance, genomics will Tissue heterogeneity in brain tumours—such as the have an important role in efforts to unravel how the presence of non-neoplastic fibroblasts, endothelial cells, transcriptome and genome of a brain-tumour cell immune cells, and necrosis—could substantially interfere influence its sensitivity to chemotherapy. Ideally, the with drug-resistance analyses in brain tumour specimens genome-wide analysis of gene expression and copy obtained by surgery. Much attention has lately focused on numbers in samples obtained by surgery will be combined strategies designed to enrich selected cell types from tissue with laser microdissection techniques to ascertain that the samples to facilitate the use of material from patients in the assessed profiles truly reflect the signature characteristics of study of markers or target discovery. In this regard, laser- the tumour-cell component. Such approaches will become assisted tissue microdissection has emerged as a method of increasingly feasible with the development of microarray choice to reduce tissue heterogeneity (figures 6 and 8). techniques that depend on lower amounts of study Several distinct laser-based techniques for tissue mRNA and DNA. In addition, further insights into THE LANCET Oncology Vol 5 February 2004 http://oncology.thelancet.com 97 For personal use. Only reproduce with permission from The Lancet.
  • 10. Review Genomics and brain-tumour drug resistance array technology: analysis of interleukin (IL) 6, IL-8, and Search strategy and selection criteria monocyte chemotactic protein 1 in the paclitaxel-resistant Data for this review were identified by searches of PubMed. All phenotype. Clin Cancer Res 1999; 5: 3445–53. papers related to gene expression profiling, drug resistance, 8 Kudoh K, Ramanna M, Ravatn R, et al. Monitoring the expression profiles of doxorubicin-induced and doxorubicin-resistant cancer and laser microdissection in primary human brain tumours cells by cDNA microarray. Cancer Res 2000; 60: 4161–66. discussed in the article were reviewed. A subset of articles 9 Watts GS, Futscher BW, Isett R, et al. cDNA microarray analysis of chosen for their importance and the further reading multidrug resistance: doxorubicin selection produces multiple opportunities they provide was included in the review. Search defects in apoptosis signaling pathways. J Pharmacol Exp Ther 2001; 299: 434–41. terms included “array CGH”, “astrocytoma”, “brain tumo(u)r”, 10 Turton NJ, Judah DJ, Riley J, et al. Gene expression and “cDNA array”, “chemoresistance”, “comparative genomic amplification in breast carcinoma cells with intrinsic and acquired hybridization”, “drug resistance”, “ependymoma”, “expression doxorubicin resistance. Oncogene 2001; 20: 1300–06. array”, “gene expression profiling”, “genomics”, “glioma”, 11 Robert J. Resistance to cytotoxic agents. Curr Opin Pharmacol “glioblastoma”, “laser microdissection”, “medulloblastoma”, 2001; 1: 353–57. “microarray”, “microdissection”, “multidrug resistance”, 12 Crowley-Weber CL, Payne CM, Gleason-Guzman M, et al. Development and molecular characterization of HCT-116 cell “oligodendroglioma”, “oligonucleotide array”, “PNET”, and lines resistant to the tumor promoter and multiple stress-inducer, “tissue microarray”. deoxycholate. Carcinogenesis 2002; 23: 2063–80. 13 Dvorakova K, Payne CM, Tome ME, et al. Molecular and cellular characterization of imexon-resistant RPMI8226/I myeloma cells. chemoresistance in brain tumours will come with a Mol Cancer Ther 2002; 1: 185–95. comprehensive analysis of the epigenetic regulation of gene 14 Hoshida Y, Moriyama M, Otsuka M, et al. Identification of genes associated with sensitivity to 5-fluorouracil and cisplatin in expression in these neoplasms. Aberrant methylation of hepatoma cells. J Gastroenterol 2002; 37 (suppl 14): 92–95. cytosine guanine dinucleotide (CpG) islands is a 15 Kihara C, Tsunoda T, Tanaka T, et al. Prediction of sensitivity of predominant epigenetic mechanism of gene inactivation. esophageal tumors to adjuvant chemotherapy by cDNA microarray analysis of gene-expression profiles. Cancer Res 2001; Microarrays that allow analysis of CpG island methylation 61: 6474–79. throughout the genome (epigenomics) have recently been 16 Sakamoto M, Kondo A, Kawasaki K, et al. Analysis of gene developed. expression profiles associated with cisplatin resistance in human ovarian cancer cell lines and tissues using cDNA microarray. Genomics investigations cannot be considered the Hum Cell 2001; 14: 305–15. endpoint in research on brain-tumour drug resistance. 17 Sotiriou C, Powles TJ, Dowsett M, et al. Gene expression profiles Such strategies allow the formulation of hypotheses about derived from fine needle aspiration correlate with response to systemic chemotherapy in breast cancer. Breast Cancer Res 2002; the relation between drug sensitivity and transcriptomic 4: R3. and genomic variation at the level of correlation rather 18 Tracey L, Villuendas R, Ortiz P, et al. Identification of genes than cause and effect. Ascertainment of biological function involved in resistance to interferon-alpha in cutaneous T-cell lymphoma. Am J Pathol 2002; 161: 1825–37. requires candidate gene validation via conventional 19 Weldon CB, Scandurro AB, Rolfe KW, et al. Identification of molecular biological approaches. The challenge of mitogen-activated protein kinase kinase as a chemoresistant functional genomics will be to elucidate how validated pathway in MCF-7 cells by using gene expression microarray. Surgery 2002; 132: 293–301. candidates act and interact as components of complex gene 20 Zembutsu H, Ohnishi Y, Tsunoda T, et al. Genome-wide cDNA regulatory networks that determine drug sensitivity in microarray screening to correlate gene expression profiles with brain tumours. Better understanding will ultimately sensitivity of 85 human cancer xenografts to anticancer drugs. Cancer Res 2002; 62: 518–27. support the search for refined therapeutic strategies that 21 Vikhanskaya F, Marchini S, Marabese M, et al. P73a improve the response to cytotoxic agents and thus the overexpression is associated with resistance to treatment with mostly poor outlook for patients with brain tumours. DNA-damaging agents in a human ovarian cancer cell line. Cancer Res 2001; 61: 935–38. 22 Komatani H, Kotani H, Hara Y, et al. Identification of breast Conflict of interest cancer resistant protein/mitoxantrone resistance/placenta-specific, None declared. ATP-binding cassette transporter as a transporter of NB-506 and J-107088, topoisomerase I inhibitors with an indolocarbazole structure. Cancer Res 2001; 61: 2827–32. References 23 Levenson VV, Davidovich IA, Roninson IB. Pleiotropic resistance 1 Bredel M. Anticancer drug resistance in primary human brain to DNA-interactive drugs is associated with increased expression tumors. Brain Res Brain Res Rev 2001; 35: 161–204. of genes involved in DNA replication, repair, and stress response. 2 Bredel M, Zentner J. Brain-tumour drug resistance: the bare Cancer Res 2000; 60: 5027–30. essentials. Lancet Oncol 2002; 3: 397–406. 24 Dan S, Tsunoda T, Kitahara O, et al. An integrated database of 3 Goldie JH, Coldman AJ. A mathematic model for relating the chemosensitivity to 55 anticancer drugs and gene expression drug sensitivity of tumors to their spontaneous mutation rate. profiles of 39 human cancer cell lines. Cancer Res 2002; 62: Cancer Treat Rep 1979; 63: 1727–33. 1139–47. 4 Esteller M, Garcia-Foncillas J, Andion E, et al. Inactivation of the 25 Wittig R, Nessling M, Will RD, et al. Candidate genes for cross- DNA-repair gene MGMT and the clinical response of gliomas to resistance against DNA-damaging drugs. Cancer Res 2002; 62: alkylating agents. N Engl J Med 2000; 343: 1350–54. 6698–705. 5 Rieger L, Rieger J, Winter S, et al. Evidence for a constitutive, 26 Bao L, Guo T, Sun Z. Mining functional relationships in feature verapamil-sensitive, non-P-glycoprotein multidrug resistance subspaces from gene expression profiles and drug activity profiles. phenotype in malignant glioma that is unaltered by FEBS Lett 2002; 516: 113–18. radiochemotherapy in vivo. Acta Neuropathol (Berl) 2000; 99: 27 Moriyama M, Hoshida Y, Otsuka M, et al. Relevance network 555–62. between chemosensitivity and transcriptome in human hepatoma 6 Schwechheimer K, Laufle RM, Schmahl W, et al. Expression of cells. Mol Cancer Ther 2003; 2: 199–205. neu/c-erbB-2 in human brain tumors. Hum Pathol 1994; 25: 28 Wallqvist A, Rabow AA, Shoemaker et al. Establishing connections 772–80. between microarray expression data and chemotherapeutic cancer 7 Duan Z, Feller AJ, Penson RT, et al. Discovery of differentially pharmacology. Mol Cancer Ther 2002; 1: 311–20. expressed genes associated with paclitaxel resistance using cDNA 29 Scherf U, Ross DT, Waltham M, et al. A gene expression 98 THE LANCET Oncology Vol 5 February 2004 http://oncology.thelancet.com For personal use. Only reproduce with permission from The Lancet.
  • 11. Genomics and brain-tumour drug resistance Review database for the molecular pharmacology of cancer. Nat Genet pathways involved in glioblastoma response to the 2000; 24: 236–44. chemotherapeutic agent 1, 3-bis(2-chloroethyl)-1-nitrosourea 30 Staunton JE, Slonim DK, Coller HA, et al. Chemosensitivity (BCNU) by gene expression profiling. Oncol Rep 1999; 6: prediction by transcriptional profiling. Proc Natl Acad Sci USA 393–401. 2001; 98: 10787–92. 52 Bacolod MD, Johnson SP, Ali-Osman F, et al. Mechanisms of 31 Butte AJ, Tamayo P, Slonim D, et al. Discovering functional resistance to 1,3-bis(2-chloroethyl)-1-nitrosourea in human relationships between RNA expression and chemotherapeutic medulloblastoma and rhabdomyosarcoma. Mol Cancer Ther 2002; susceptibility using relevance networks. Proc Natl Acad Sci USA 1: 727–36. 2000; 97: 12182–86. 53 Mukasa A, Ueki K, Matsumoto S, et al. Distinction in gene 32 Watson MA, Gutmann DH, Peterson K, et al. Molecular expression profiles of oligodendrogliomas with and without allelic characterization of human meningiomas by gene expression loss of 1p. Oncogene 2002; 21: 3961–68. profiling using high-density oligonucleotide microarrays. 54 Eisen MB, Spellman PT, Brown PO, Botstein D. Cluster analysis Am J Pathol 2002; 161: 665–72. and display of genome-wide expression patterns. Proc Natl Acad 33 Watson MA, Perry A, Budhjara V, et al. Gene expression profiling Sci USA 1998; 95: 14863–68. with oligonucleotide microarrays distinguishes World Health 55 Tamayo P, Slonim D, Mesirov J, et al. Interpreting patterns of Organization grade of oligodendrogliomas. Cancer Res 2001; 61: gene expression with self-organizing maps: methods and 1825–29. application to hematopoietic differentiation. Proc Natl Acad Sci 34 Hunter S, Young A, Olson J, et al. Differential expression between USA 1999; 96: 2907–12. pilocytic and anaplastic astrocytomas: identification of 56 Hilsenbeck SG, Friedrichs WE, Schiff R, et al. Statistical analysis of apolipoprotein D as a marker for low-grade, non-infiltrating array expression data as applied to the problem of tamoxifen primary CNS neoplasms. J Neuropathol Exp Neurol 2002; 61: resistance. J Natl Cancer Inst 1999; 91: 453–59. 275–81. 57 Kim SK, Lund J, Kiraly M, et al. A gene expression map for 35 Kim S, Dougherty ER, Shmulevich L, et al. Identification of Caenorhabditis elegans. Science 2001; 293: 2087–92. combination gene sets for glioma classification. Mol Cancer Ther 58 Heyer LJ, Kruglyak S, Yooseph S. Exploring expression data: 2002; 1: 1229–36. identification and analysis of coexpressed genes. Genome Res 1999; 36 Fathallah-Shaykh HM, Rigen M, Zhao LJ, et al. Mathematical 9: 1106–15. modeling of noise and discovery of genetic expression classes in 59 Hastie T, Tibshirani R, Eisen MB, et al. ‘Gene shaving’ as a method gliomas. Oncogene 2002; 21: 7164–74. for identifying distinct sets of genes with similar expression 37 Fuller GN, Hess KR, Rhee CH, et al. Molecular classification of patterns. Genome Biol 2000; 1: RESEARCH0003. human diffuse gliomas by multidimensional scaling analysis of 60 Dahlquist KD, Salomonis N, Vranizan K, et al. GenMAPP, a new gene expression profiles parallels morphology-based classification, tool for viewing and analyzing microarray data on biological correlates with survival, and reveals clinically-relevant novel pathways. Nat Genet 2002; 31: 19–20. glioma subsets. Brain Pathol 2002; 12: 108–16. 61 Gygi SP, Rochon Y, Franza BR, Aebersold R. Correlation between 38 Gutmann DH, Hedrick NM, Li J, et al. Comparative gene expression profile analysis of neurofibromatosis 1-associated and protein and mRNA abundance in yeast. Mol Cell Biol 1999; 19: sporadic pilocytic astrocytomas. Cancer Res 2002; 62: 2085–91. 1720–30. 39 Ljubimova JY, Khazenzon NM, Chen Z, et al. Gene expression 62 Wang H, Zhang W, Fuller GN. Tissue microarrays: applications in abnormalities in human glial tumors identified by gene array. neuropathology research, diagnosis, and education. Brain Pathol Int J Oncol 2001; 18: 287–95. 2002; 12: 95–107. 40 Qi ZY, Li Y, Ying K, et al. Isolation of novel differentially 63 Pollack JR, Sorlie T, Perou CM, et al. Microarray analysis reveals a expressed genes related to human glioma using cDNA microarray major direct role of DNA copy number alteration in the and characterizations of two novel full-length genes. J Neurooncol transcriptional program of human breast tumors. Proc Natl Acad 2002; 56: 197–208. Sci USA 2002; 99: 12963–68. 41 Rickman DS, Bobek MP, Misek DE, et al. Distinctive molecular 64 Hess KR, Fuller GN, Rhee CH, Zhang W. Statistical pattern profiles of high-grade and low-grade gliomas based on analysis of gene expression profiles for glioblastoma tissues and oligonucleotide microarray analysis. Cancer Res 2001; 61: 6885–91. cell lines. Int J Mol Med 2001; 8: 183–88. 42 Sallinen SL, Sallinen PK, Haapasalo HK, et al. Identification of 65 Sasaki T, Hankins GR, Helm GA. Comparison of gene expression differentially expressed genes in human gliomas by DNA profiles between frozen original meningiomas and primary microarray and tissue chip techniques. Cancer Res 2000; 60: cultures of the meningiomas by GeneChip. Neurosurgery 2003; 52: 6617–22. 892–99. 43 Sehgal A, Boynton AL, Young RF, et al. Application of the 66 Emmert-Buck MR, Bonner RF, Smith PD, et al. Laser capture differential hybridization of Atlas Human expression arrays microdissection. Science 1996; 274: 998–1001. technique in the identification of differentially expressed genes in 67 Schutze K, Lahr G. Identification of expressed genes by laser- human glioblastoma multiforme tumor tissue. J Surg Oncol 1998; mediated manipulation of single cells. Nat Biotechnol 1998; 16: 67: 234–41. 737–42. 44 Huang H, Colella S, Kurrer M, et al. Gene expression profiling of 68 Fink L, Seeger W, Ermert L, et al. Real-time quantitative RT-PCR low-grade diffuse astrocytomas by cDNA arrays. Cancer Res 2000; after laser-assisted cell picking. Nat Med 1998; 4: 1329–33. 60: 6868–74. 69 Bonaventure P, Guo H, Tian B, et al. Nuclei and subnuclei gene 45 Markert JM, Fuller CM, Gillespie GY, et al. Differential gene expression profiling in mammalian brain. Brain Res 2002; 943: expression profiling in human brain tumors. Physiol Genomics 38–47. 2001; 5: 21–33. 70 Becker AJ, Wiestler OD, Blumcke I. Functional genomics in 46 Mischel PS, Shai R, Shi T, et al. Identification of molecular experimental and human temporal lobe epilepsy: powerful new subtypes of glioblastoma by gene expression profiling. Oncogene tools to identify molecular disease mechanisms of hippocampal 2003; 22: 2361–73. damage. Prog Brain Res 2002; 135: 161–73. 47 Nutt CL, Mani DR, Betensky RA, et al. Gene expression-based 71 Mariani L, Beaudry C, McDonough WS, et al. Death-associated classification of malignant gliomas correlates better with survival protein 3 (Dap-3) is overexpressed in invasive glioblastoma cells than histological classification. Cancer Res 2003; 63: 1602–07. in vivo and in glioma cell lines with induced motility phenotype in 48 Korenberg MJ. Gene expression monitoring accurately predicts vitro. Clin Cancer Res 2001; 7: 2480–89. medulloblastoma positive and negative clinical outcomes. FEBS 72 Luzzi V, Holtschlag V, Watson MA. Expression profiling of Lett 2003; 533: 110–14. ductal carcinoma in situ by laser capture microdissection and 49 Pomeroy SL, Tamayo P, Gaasenbeek M, et al. Prediction of central high-density oligonucleotide arrays. Am J Pathol 2001; 158: nervous system embryonal tumour outcome based on gene 2005–10. expression. Nature 2002; 415: 436–42. 73 Kitahara O, Furukawa Y, Tanaka T, et al. Alterations of gene 50 Tanwar MK, Gilbert MR, Holland EC. Gene expression expression during colorectal carcinogenesis revealed by cDNA microarray analysis reveals YKL-40 to be a potential serum marker microarrays after laser-capture microdissection of tumor tissues for malignant character in human glioma. Cancer Res 2002; 62: and normal epithelia. Cancer Res 2001; 61: 3544–49. 4364–68. 74 Alevizos I, Mahadevappa M, Zhang X, et al. Oral cancer in 51 Rhee CH, Ruan S, Chen S, et al. Characterization of cellular vivo gene expression profiling assisted by laser capture THE LANCET Oncology Vol 5 February 2004 http://oncology.thelancet.com 99 For personal use. Only reproduce with permission from The Lancet.
  • 12. Review Genomics and brain-tumour drug resistance microdissection and microarray analysis. Oncogene 2001; 20: use of laser capture microdissection and cDNA arrays. Oncogene 6196–204. 2000; 19: 3220–24. 75 Leethanakul C, Patel V, Gillespie J, et al. Distinct pattern of 76 Sgroi DC, Teng S, Robinson G, LeVangie R, Hudson JR, Jr., expression of differentiation and growth-related genes in Elkahloun AG. In vivo gene expression profile analysis of human squamous cell carcinomas of the head and neck revealed by the breast cancer progression. Cancer Res 1999; 59: 5656–61. Advances in research Small-animal imaging of tumour proliferation with PET Henryk Barthel, Pat Price, and Eric O Aboagye We consecutively imaged tumour A B proliferation and glucose utilisation in a L5178Y lymphoma-bearing mouse. This was done by the use of a high Brain resolution, small-animal PET scanner, and the administration of [ F]FLT and 18 [18F]FDG. Currently, [18F]FDG is the most commonly used tracer for cancer detection with PET, and reflects the activity of glucose transporters and hexokinase (panel a). By contrast, tumour detection by [18F]FLT is thought to depend on thymidine Kidney kinase 1 (TK1) activity—the key Kidney enzyme of the DNA salvage pathway (panel b). In mice with bilateral TK1 / TK1 / variants of L5178Y tumours (TK1 -/- TK1 / TK1 / variant implanted in lower left dorsal region, TK1 +/- variant implanted in lower right dorsal region), the +/- tumours produced 48% more TK1 enzyme on average, and also grew faster (27% shorter volume doubling time), compared with -/- tumours. Uptake of [18F]FLT by PET was higher for the +/- tumour, encourage oncological research because they could compared with the -/- tumour. A converse pattern, however, potentially support drug development and also shorten the was found in the [18F]FDG image. These findings support the time between translation of preclinical research into a clinical conclusion that the imaging of DNA synthesis is more accu- setting. With respect to the new tumour proliferation marker, rate for the assessment of the proliferative potential of [18F]FLT, our PET imaging experiments provide clarity on tumours, compared with the imaging of glucose the uptake mechanism of this radiotracer by tumours. consumption. Clinical testing of this promising new tumour proliferation This example highlights the unique possibility of modern marker for PET is already underway and may provide a new PET techniques for non-invasive investigation of different more sensitive, non-invasive approach for monitoring tumour properties in small laboratory animals with ultra- treatment response and improving the accuracy of clinical high spatial resolution (~1 mm). Such techniques should staging. Correspondence: Dr Henryk Barthel, Molecular Therapy and PET Oncology Research Group, Faculty of Medicine, Imperial College London, Hammersmith Hospital, Du Cane Road, London W12 0NN, UK. Tel: +49 (0)341 9718082. Fax: +49 (0)341 9718009. Email: barh@medizin.uni-leipzig.de Call for papers The Lancet Oncology would like to invite all potential authors to consider submitting articles to the Advances in research section of the journal. Further details on this section can be found in the Instructions to authors, and all enquiries and submissions should be sent to the Editor at: david.collingridge@lancet.com 100 THE LANCET Oncology Vol 5 February 2004 http://oncology.thelancet.com For personal use. Only reproduce with permission from The Lancet.