Portal de Periódicos da CAPES

Cross-Species Oncogenomics in Cancer Gene Identification

2006; Cell Press; Volume: 125; Issue: 7 Linguagem: Inglês

10.1016/j.cell.2006.06.018

1097-4172

Daniel S. Peeper, Anton Berns,

Cancer-related gene regulation

The complexity of genomic aberrations in most human tumors hampers delineation of the genes that drive the tumorigenic process. In this issue of Cell, Kim et al., 2006Kim M. Gans J.D. Nogueira C. Wang A. Paik J.-H. Feng B. Brennan C. Hahn W.C. Cordon-Cardo C. Wagner S.N. et al.Cell. 2006; (this issue)Google Scholar and Zender et al., 2006Zender L. Spector M.S. Xue W. Flemming P. Cordon-Cardo C. Silke J. Fan S.-T. Luk J.M. Wigler M. Hannon G.J. et al.Cell. 2006; (this issue)PubMed Google Scholar demonstrate that cognate mouse tumor models recapitulate these genetic alterations with unexpected fidelity. These results indicate that cross-species genomic analysis is a powerful strategy to identify the responsible genes and assess their oncogenic capacity in the appropriate genetic context. The complexity of genomic aberrations in most human tumors hampers delineation of the genes that drive the tumorigenic process. In this issue of Cell, Kim et al., 2006Kim M. Gans J.D. Nogueira C. Wang A. Paik J.-H. Feng B. Brennan C. Hahn W.C. Cordon-Cardo C. Wagner S.N. et al.Cell. 2006; (this issue)Google Scholar and Zender et al., 2006Zender L. Spector M.S. Xue W. Flemming P. Cordon-Cardo C. Silke J. Fan S.-T. Luk J.M. Wigler M. Hannon G.J. et al.Cell. 2006; (this issue)PubMed Google Scholar demonstrate that cognate mouse tumor models recapitulate these genetic alterations with unexpected fidelity. These results indicate that cross-species genomic analysis is a powerful strategy to identify the responsible genes and assess their oncogenic capacity in the appropriate genetic context. Chromosomal aberrations recurrently contribute to malignant transformation. Typically, deleted or amplified genomic regions cover large (rather than focal) areas, thus slowing down the identification of the specific genes driving tumorigenesis. In two papers published in this issue of Cell (Kim et al., 2006Kim M. Gans J.D. Nogueira C. Wang A. Paik J.-H. Feng B. Brennan C. Hahn W.C. Cordon-Cardo C. Wagner S.N. et al.Cell. 2006; (this issue)Google Scholar, Zender et al., 2006Zender L. Spector M.S. Xue W. Flemming P. Cordon-Cardo C. Silke J. Fan S.-T. Luk J.M. Wigler M. Hannon G.J. et al.Cell. 2006; (this issue)PubMed Google Scholar), integrative cross-species analysis was used to narrow down the number of candidate oncogenes within amplified DNA segments (amplicons). This approach illustrates the power of genome-wide comparison of cognate mouse and human tumors, as it identified genes found in regions commonly amplified in both species (Figure 1). To study the genes involved in the metastasis of melanoma, Kim et al., 2006Kim M. Gans J.D. Nogueira C. Wang A. Paik J.-H. Feng B. Brennan C. Hahn W.C. Cordon-Cardo C. Wagner S.N. et al.Cell. 2006; (this issue)Google Scholar used an inducible H-Ras nonmetastatic mouse model of melanoma, from which they derived two metastatic cell lines. Array-Comparative Genome Hybridization (CGH, which measures DNA copy-number differences between genomes) showed that these cell lines, relative to their parental counterparts, shared an amplified region of 850 kb on chromosome 13 encompassing eight genes. A region of much larger size, syntenic (preserved as “blocks” of genes across species) with the amplified region in the mouse, is frequently observed in human melanoma. This amplicon is more predominantly present in metastatic variants, suggesting that it might harbor a gene contributing to the metastatic potential of melanoma. Expression analysis in murine melanomas showed that one gene, NEDD9, was the most likely candidate to enhance metastasis. Subsequent analyses of NEDD9 levels in human melanoma indicated significant upregulation, with levels increasing as a function of tumor progression. Depletion of NEDD9 by RNA interference (RNAi) reduced the invasive capacity of melanoma cells and impaired experimental metastasis in vivo, as seen for both murine and human cells. Interestingly, the metastatic potential of NEDD9 could be abrogated by knocking down focal adhesion kinase (FAK), a gene previously implicated in invasive growth (Hess et al., 2005Hess A.R. Postovit L.M. Margaryan N.V. Seftor E.A. Schneider G.B. Seftor R.E. Nickoloff B.J. Hendrix M.J. Cancer Res. 2005; 65: 9851-9860Crossref PubMed Scopus (126) Google Scholar). NEDD9 and FAK appear to colocalize in focal contacts, which result from NEDD9 overexpression. Therefore, this study not only identified a gene enhancing metastasis in melanoma, it also points to a potential interesting target for intervention, FAK. Another noteworthy point is that NEDD9 expression is observed not only in metastases but also in primary tumors, indicating that NEDD9 provides a selective advantage to primary tumors as well. In keeping with this, the authors observed that RNAi-mediated depletion of NEDD9 had a significant impact on cell proliferation. This multifaceted feature of certain oncogenes likely represents a frequently occurring phenomenon contributing to both tumor initiation and progression, as previously proposed by Bernards and Weinberg, 2002Bernards R. Weinberg R.A. Nature. 2002; 418: 823Crossref PubMed Scopus (699) Google Scholar. Indeed, overexpression of NEDD9 stimulated both proliferation and invasive capacity of Ink4a;Arf−/− melanocytes as well as their metastatic potential in vivo. Interestingly, this was seen only in combination with B-RAFV600E or H-RasV12, illustrating that NEDD9 acts in a context-dependent fashion. Notably, it would have been impossible to reach these conclusions so rapidly without the aid of this mouse model. A second paper, by Zender et al., 2006Zender L. Spector M.S. Xue W. Flemming P. Cordon-Cardo C. Silke J. Fan S.-T. Luk J.M. Wigler M. Hannon G.J. et al.Cell. 2006; (this issue)PubMed Google Scholar, published in this issue of Cell describes a new mouse model for liver cancer that permits the identification of genes contributing to hepatocellular carcinomas (HCCs). The authors established hepatoblast cultures allowing in vitro genetic manipulation. As HCCs almost invariably harbor inactivating mutations in p53, the authors infected cultured hepatoblasts from p53-deficient embryos with the oncogenes c-myc, Akt, or H-RasV12. Engraftment of these cells into mice resulted in liver tumors, albeit with different pathologies. Using representational oligonucleotide microarray analysis (ROMA; Lucito et al., 2003Lucito R. Healy J. Alexander J. Reiner A. Esposito D. Chi M. Rodgers L. Brady A. Sebat J. Troge J. et al.Genome Res. 2003; 13: 2291-2305Crossref PubMed Scopus (335) Google Scholar) to scan the genome for copy-number changes at high resolution, genes that might contribute to the tumor phenotype were identified. In Akt-induced tumors, no focal genomic alterations smaller than 5 Mb were found. H-Ras-transduced hepatoblasts gave rise to tumors with, in one case, a focal amplification of c-myc and, in another, of Rnf19. Although Rfn19 has not been linked to tumorigenesis, c-myc alterations are common in human HCC. ROMA of HCCs induced by overexpression of c-Myc revealed a small amplicon on mouse chromosome 9. The amplified segment is syntenic with a region on human chromosome 11q22 that is amplified in a subset of HCC and esophageal cancers. The cross-species comparison limited the number of candidate genes in the region. Most genes encoded by the amplicon, including a number of matrix metalloproteinases, could be excluded as candidates because they were not consistently overexpressed. The remainder of the genes could thus be responsible for the phenotypic effect, e.g., by acting as the “drivers” that stimulate expansion of the cells carrying this amplicon. Two genes encoding cIAP1 (Imoto et al., 2001Imoto I. Yang Z.Q. Pimkhaokham A. Tsuda H. Shimada Y. Imamura M. Ohki M. Inazawa J. Cancer Res. 2001; 61: 6629-6634PubMed Google Scholar), an inhibitor of apoptosis, and Yap (Yagi et al., 1999Yagi R. Chen L.F. Shigesada K. Murakami Y. Ito Y. EMBO J. 1999; 18: 2551-2562Crossref PubMed Scopus (451) Google Scholar), a Src-interacting protein, appeared overexpressed in all murine and human amplicon-containing tumors analyzed. Their contribution to HCC was subsequently evaluated in the versatile hepatoblast graft model using combinations of retroviral vectors encoding c-Myc, cIAP1, and Yap. cIAP1 overexpression significantly enhanced tumor growth, but only when c-Myc was overexpressed. cIAP1 conferred no growth advantage in combination with either H-Ras or Akt, illustrating that the oncogenicity of cIAP1 was context dependent. Yap acts to enhance the efficiency of Runx and TEAD/TEF transcription factors. It can also potentiate apoptosis, an activity that seems hard to reconcile with an oncogenic role. However, overexpression of both Yap and c-Myc (but not H-Ras) in p53-deficient hepatoblasts resulted in accelerated tumor growth, demonstrating that Yap can act as an oncogene, once again as a function of its genetic context. Downregulation of Yap by short-hairpin RNAs (shRNAs) resulted in reduced cyclin E levels and impaired progression of the murine tumor cells. It will be interesting to learn whether cIAP1/Yap depletion has a similar effect on the tumorigenicity of human liver cancer cells. One often assumes that single oncogenes within amplicons bear most if not all of the tumorigenic activity, with neighboring genes merely representing passengers. Strikingly, the authors found that coexpression of c-Myc with both cIAP1 and Yap resulted in synergistic stimulation of tumor growth. These studies highlight the power of integrative, cross-species oncogenomics in cancer gene discovery. First, they elegantly illustrate the value of cross-species comparisons of cancer genomes combining genomic and expression analyses. Although the usefulness of cross-species sequence comparisons is undisputed, there has been skepticism about the added value of comparing genomic aberrations in tumors of different species. High-resolution CGH analyses, however, have revealed a striking concordance of chromosomal gains and losses in syntenic regions of tumors in mice and humans, indicating that the development of these tumors is driven by the same genes. Evidently, cross-species comparisons greatly facilitate identification of the relevant genes or genetic elements conferring oncogenicity because more amplicons can be scored and because alignment of the amplicons found in mice and humans allows narrowing down the minimal region of overlap. The candidacy of genes can further be scrutinized by assessing their consistent expression in both mouse and human tumors, thereby creating a biological sieve that allows candidate genes to be prioritized. This integrative approach effectively complements cancer genome sequencing efforts, which, surprisingly, have uncovered relatively few cancer-associated mutations so far (Davies et al., 2002Davies H. Bignell G.R. Cox C. Stephens P. Edkins S. Clegg S. Teague J. Woffendin H. Garnett M.J. Bottomley W. et al.Nature. 2002; 417: 949-954Crossref PubMed Scopus (8262) Google Scholar). Second, both studies discussed here exemplify the influence of the genetic context on oncogene function. Specifically, NEDD9 exerts its pro-oncogenic effect in Ink4a/Arf-deficient cells only in conjunction with H-RasV12 or B-RAFV600E, whereas cIAP1 and Yap exert their synergistic effect with c-Myc but not with Akt or H-Ras. Given that YAP, by activating the p53-related protein p73, acts even proapoptotically in certain settings, it seems appropriate to add it to an expanding list of “dual-function genes” that either stimulate or suppress tumorigenesis depending on their genetic context (Rowland and Peeper, 2006Rowland B.D. Peeper D.S. Nat. Rev. Cancer. 2006; 6: 11-23Crossref PubMed Scopus (457) Google Scholar). This issue should not be underestimated, as it may impact the decision of which genes we select for targeted inhibition. One might imagine that, for specific gene products, systemic treatment with targeted drugs, in addition to inhibiting tumor growth, may give rise to adverse effects in other cellular (i.e., genetic) contexts. A third interesting finding is that in a single, relatively small amplicon—two genes, cIAP1 and Yap—cooperate in driving tumorigenesis. This illustrates that meticulous analysis of candidate oncogenes is a necessity and also that this can be performed expeditiously only with the aid of mouse models as described in the two papers discussed here. If one assumes that a few thousand genes can contribute to tumorigenesis, oncogenes (or tumor-suppressor genes, for that matter) will be located quite often in close vicinity to each other, and, hence, their comutation by amplification (or by deletion for tumor suppressors) likely occurs frequently. Thus, a single genetic aberration might affect multiple genes, each of which contributes to tumorigenesis. The Ink4 locus, harboring three tumor-suppressor genes (p16INK4a, p15INK4b, and p14/p19ARF) in very close proximity, is a case in point, but this is likely true for many other chromosomal regions. It will be important to assess to what extent the individual genes described here actually drive the tumorigenic process. As RNAi depletion of cIAP1, Yap, or NEDD9 failed to cause full tumor regression or complete suppression of metastasis, it is unclear whether the corresponding tumors are similarly “addicted” to these genes as has been shown for tumor-driving mutations such as Bcr-Abl in CML (chronic myeloid leukemia) in humans and c-myc and H-Ras in several mouse tumor models. (Jonkers and Berns, 2004Jonkers J. Berns A. Cancer Cell. 2004; 6: 535-538Abstract Full Text Full Text PDF PubMed Scopus (114) Google Scholar, Weinstein, 2002Weinstein I.B. Science. 2002; 297: 63-64Crossref PubMed Scopus (1477) Google Scholar). In addition, it will be necessary to critically evaluate whether concurrent ablation of the coamplified genes can more effectively contribute to tumor eradication. The mouse models described here are exquisitely suited to address such questions, as combination of different shRNAs can be used to genetically ablate gene expression and assess the added therapeutic potential of concomitant inhibition of targets. The studies described testify to the notion that genetically tractable mouse models represent an invaluable tool not only to identify new cancer-causing genes but also to assess the context-dependent vulnerability of tumors to multitarget intervention strategies. Comparative Oncogenomics Identifies NEDD9 as a Melanoma Metastasis GeneKim et al.CellJune 30, 2006In BriefGenomes of human cancer cells are characterized by numerous chromosomal aberrations of uncertain pathogenetic significance. Here, in an inducible mouse model of melanoma, we characterized metastatic variants with an acquired focal chromosomal amplification that corresponds to a much larger amplification in human metastatic melanomas. Further analyses identified Nedd9, an adaptor protein related to p130CAS, as the only gene within the minimal common region that exhibited amplification-associated overexpression. Full-Text PDF Open ArchiveIdentification and Validation of Oncogenes in Liver Cancer Using an Integrative Oncogenomic ApproachZender et al.CellJune 30, 2006In BriefThe heterogeneity and instability of human tumors hamper straightforward identification of cancer-causing mutations through genomic approaches alone. Herein we describe a mouse model of liver cancer initiated from progenitor cells harboring defined cancer-predisposing lesions. Genome-wide analyses of tumors in this mouse model and in human hepatocellular carcinomas revealed a recurrent amplification at mouse chromosome 9qA1, the syntenic region of human chromosome 11q22. Gene-expression analyses delineated cIAP1, a known inhibitor of apoptosis, and Yap, a transcription factor, as candidate oncogenes in the amplicon. Full-Text PDF Open Archive