- Email: [email protected]
Identification of cancer genes by mutational profiling of tumor genomes
FEBS Letters 579 (2005) 1884–1890
FEBS 29348
Minireview
Identification of cancer genes by mutational profiling of tumor genomes Silvia Benvenuti, Sabrina Arena, Alberto Bardelli* The Oncogenomics Center, Institute for Cancer Research and Treatment (IRCC), University of Torino Medical School, Candiolo (To), Italy Accepted 3 February 2005 Available online 16 February 2005 Edited by Robert Russell and Giulio Superti-Furga
Abstract It is now widely accepted that cancer is a genetic disease and that alterations in the DNA sequence underlie the development of every neoplasm. The identification of mutated genes that are causally implicated in oncogenesis (Ôcancer genesÕ) has been a major goal in medical sciences for the last two decades. The availability of the human genome sequence coupled to the introduction of high throughput sequencing technologies has created an unprecedented opportunity in this field. It is now possible to perform mutational studies of entire cancer genomes thus providing a complete description of mutations underlying human oncogenesis. The recent identification of high frequency mutations in the BRAF and PI3K genes suggests that many more cancer genes remain to be discovered. In this review, we consider how the systematic mutational analysis of gene families in individual neoplasms has led to the identification of a number of cancer genes and how this information is influencing the treatment of cancer. Ó 2005 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. Keywords: Cancer genome; Oncogenomics; Gene family; Tailored therapies; Mutational analysis
1. Cancer: a genetic disease In the living cell, DNA undergoes frequent chemical changes. The majority of these changes are quickly repaired; those that are not result in a mutation. Cancer progression results by the serial accumulation of genetic accidents in somatic cells [1]. The genes that are causally involved in oncogenesis are known as cancer genes. DNA alterations that convey tumorassociated properties are favored by natural selection that ultimately drives the acquisition of the neoplastic properties [2]. Initially, these involve the deregulation of proliferation and resistance to apoptosis. Subsequently, some cancer cells acquire the ability to invade and colonize other tissues giving rise to metastasis that are primarily responsible for cancer-associated mortality. A primary way to discriminate whether a mutated gene plays a causal role in cancer progression is that the presence of mutations in the gene, rather than being attributable to chance, are thought to result in a growth advantage on the cell population from which the cancer developed [3].
* Corresponding author. E-mail address: [email protected] (A. Bardelli).
Since the first speculations on their existence, more than 100 cancer genes have been identified [4]. However, even commonly mutated cancer genes have remained undetected until very recently. For example, only in the last 3 years the BRAF [5] and the PI3K [6,7] genes have been shown to be frequently mutated in a large fraction of tumors and to play a causative role in tumorigenesis. It is therefore reasonable to assume that many more cancer genes remain to be identified. 2. Cancer genes discovery and the human genome project A simplistic interpretation divides cancer genes into three broad categories: oncogenes, tumor suppressors and stability genes. Oncogenes promote cell growth and when mutated result in constitutively or abnormally active proteins; such mutations are mainly of the dominant type. Tumor suppressors genes are negative controllers of cell cycle progression and the mutations that affect them are typically recessive. A third category of cancer genes has been proposed more recently. This comprises the so-called stability or caretaker genes. These genes are not directly involved in tumorigenesis but when altered they contribute to cancer by exposing cells to an abnormally high mutation rate. This feature ultimately leads to oncogene activation or tumor suppressor inactivation. Regardless of the specific genes involved, the end result of the molecular alteration affecting cancer genes is the deregulation of cellular homeostasis [8,9]. As human cancer is a genetic disease, the completion of the human genome project has offered unprecedented opportunities for cancer diagnosis, prognosis and treatment. The field of cancer genes discovery has been one of the first to be affected by this historic revolution that will ultimately allow the identification of all genetic alterations underlying the origin and evolution of tumors. Below we consider how systematic mutational analysis of gene families in individual neoplasms has led to the identification of a number of cancer genes and how this information has already affected the treatment of cancer.
3. Discovery of cancer genes before the human genome project 3.1. The oncogenesÕs hunt Until the completion of the human genome project a number of approaches have led to the identification of genes that when mutated are causally implicated in oncogenesis. Key breakthrough in the discovery of oncogenes resulted from studies on retroviruses. Retroviruses are RNA tumor
0014-5793/$30.00 Ó 2005 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2005.02.015
S. Benvenuti et al. / FEBS Letters 579 (2005) 1884–1890
viruses capable of infecting vertebrate cells and whose genome can be retrotranscribed into DNA. They can be divided into two groups: acute transforming and weakly oncogenic retroviruses [10]. The former induce tumor formation very rapidly (days) after infection and were found to contain modified copies of cellular genes that have been incorporated into the viral genome during retroviral transduction process. Rous and colleagues obtained key clues on acute transforming oncogenes in 1911. For the first time they reported that sarcomas could be induced in chickens by cell free filtrates of a sarcoma taken from another chicken. However, the region responsible for the oncogenicity was isolated only 60 years later. It was v-src, a mutated copy of the human c-src gene [11]. The weakly oncogenic retroviruses cause tumors only after a long incubation period (months). Contrary to acute transforming retroviruses, weakly oncogenic retroviruses do not contain viral oncogenes. They were also central in the identification of cancer genes. In fact, weakly oncogenic retroviruses induce tumors via a process called insertional mutagenesis. Specifically, they integrate in the host genome close to an oncogene thus ensuing its abnormal expression which is typically driven by the transactivating elements contained in the virus genome [12,13]. In addition to the studies on retroviruses, pioneer experiments designed to identify dominant mutations were introduced at the beginning of the 1980s together with the advancement of cell transfection techniques [14,15]. In particular a specific oncogene hunting technique (known as the ‘‘NIH 3T3 transformation assay’’) was established. This technique is based on the extraction of genomic DNA from tumor samples that is fractionated and then introduced into NIH 3T3 mouse fibroblasts. If an oncogene is present colonies of hyper proliferating cells appear in the cell culture. Each colony is generated from the clonal expansion of a single cell, which gained a proliferative advantage after ectopic gene expression and as a result has lost the cell–cell contact inhibitory control mechanisms. The process is then serially repeated by extracting the genomic DNA from the transformed foci ultimately leading to the identification of the human gene responsible for the tumorigenic properties. Several oncogenes have been discovered using the NIH 3T3 transformation assay. The methodology was initially applied to a human bladder carcinoma cell line [16,17] and has lead to the identification of one of the most common genetic lesions found in a variety of cancers. In the specific case, the genetic alteration turned out to be a single point mutation which resulted in the incorporation of Valine instead of Glycine as the 12th amino acid residue in the Ras protein [18]. hRas mutations are present in more than 25% of all human cancers. hRas encodes for small monomeric GTPases responsible for delivering external signals captured by the cells via their membrane receptors to the cell nucleus. The point mutations affecting hRas generate a protein, which fails to efficiently hydrolyze GTP and therefore remains constitutively active resulting in deregulated cellular proliferation. Significant improvements in the identification of oncogenes also stemmed from the karyotypic analysis of human cancers. In particular, tumors of the haematopoietic system such as leukemias were shown to carry specific chromosomal transloca-
1885
tions. The introduction of sophisticated chromosomal banding techniques at the beginning of the 1970s allowed the precise description of those changes. Breakpoints (both translocations and inversions), amplifications and deletions were then characterized in detail [19]. A classical example of the power of this strategy is the identification of the genetic alteration responsible for chronic myeloid leukemia (CML). In this case, karyotype analysis unequivocally determined that cells of patients affected by CML are characterized by a translocation between chromosomes 9 and 22 (t(9;22)(q34;q11)) [20]. The translocation generates the Philadelphia chromosome. In this chimeric chromosome, the 5 0 segment of the BCR gene is juxtaposed to the 3 0 region of the ABL gene and results in a constitutively active 210 kDa fusion protein: the tyrosine kinase BCR-ABL [21].
3.2. The tumor suppressor’s hunt The behavior of cancer cells suggested that not all tumor properties could be explained by the deregulated over activation of oncogenes, but the inactivation of other genes was also crucial. Mutations affecting tumor suppressor genes typically reduce or abolish the activity of the corresponding protein product. Examples are missense mutations at residues that are essential for the catalytic activity, mutations that result in a truncated protein, deletions or insertions of various sizes and epigenetic silencing. Usually, it is necessary to eliminate both copies of a tumor suppressor gene to inactivate it. Identification of genes inactivated during tumorigenesis is a challenging task. An important contribution to the identification of tumor suppressor genes was made by Harris and colleagues [22]. Using cell fusion experiments they observed that growth of murine tumors in synergic animals could be suppressed when malignant cells were fused to non-malignant cells. Moreover, it was found that tumorigenicity was associated with specific chromosomal losses and that the introduction of single chromosomes could reverse malignancy. A breakthrough in the discovery of the first tumor suppressor gene came from retinoblastoma, a rare tumor that occurs in childhood. This tumor develops from neural precursor cells in the immature retina, usually sporadically, but in some families, it displays an autosomal dominant mode of inheritance [23]. In the hereditary type, multiple tumors occur independently, affecting both eyes. In the sporadic form only one eye is affected, and by only one tumor. A key clue on the putative gene responsible for retinoblastoma was obtained when the karyotype of patients with the inherited form was shown to be abnormal. The karyotype was characterized by a deletion of chromosome 13 band q14 [24,25]. Deletions of the same locus were also observed in tumor cells from patients affected by the nonhereditary form, suggesting that the cancer was possibly caused by loss of a critical gene located in that particular chromosomal region. Subsequently, the locus was characterized in detail and the Rb gene was identified. Loss of Rb is one of the crucial steps in cancer progression. The Rb gene encodes the Rb protein, a key regulator of the cell cycle. pRb undergoes cell-cycle dependent phosphorylation, which is dependent on the activity of several cyclin-dependent kinases and occurs in a stepwise manner [26]. When hypo-
1886
S. Benvenuti et al. / FEBS Letters 579 (2005) 1884–1890
phosphorylated, pRb associates with and sequesters the E2F family of transcription factors. Upon progression through the cell cycle pRB becomes hyperphosphorylated and therefore releases E2F factors which, in turn, transactivate genes containing E2F recognition sites in their regulatory sequences [26]. Rb mutations are thought to influence the stability of the genome of cancer cells.
4. Discovery of cancer genes after the human genome project The completion of the human genome and the introduction of high throughput sequencing technologies allow the discovery of cancer genes through systematic approaches. In principle, it is now possible to sequence every gene in a given cancer genome and identify every mutation. Already, a number of laboratories have started the systematic mutational profiling of cancer genomes. In order to achieve this goal a new generation of bioinformatic tools has been developed to extract exonic and intronic sequences for each gene of interest from the available databases. The sequences are then exploited for high throughput design of exon specific amplification and sequencing primers. Automated procedures for amplification and direct sequencing of exons from tumour derived genomic DNA have also been implemented [5,27]. These require robotic handling of PCRs, purification and sequencing of PCR products. Specific mutational analysis algorithms have made it possible to identify both heterozygous and homozygous mutations in high throughput fashion (Fig. 1). Using softwareÕs programs based on these algorithms high throughput mutational analysis can be achieved by comparing
chromatograms from different tumors to a reference sequence [27]. 4.1. Discovery of cancer genes by dissecting oncogenic signal pathways Among the strategies devised to identify cancer genes in the post genomic era a genetic analysis was implemented by Davies and colleagues [5] to identify cancer-associated mutations by dissecting signaling pathways. Specifically, they decided to search for mutations in members of signaling pathways known to involve at least one gene previously found mutated in human cancers. They started by analyzing all the members of the RAS–RAF–MEK–ERK–MAP kinase pathway [5,28]. In this approach the initial high throughput screening for mutations was achieved using a capillary-based heteroduplex technology and not by the direct comparison of DNA sequence traces. The second approach was implemented later and, given its better performance, is currently the standard for large scale sequencing projects [27]. Using this strategy, the authors reported genetic alterations in the BRAF gene in a variety of cancers, in particular in 66% of malignant melanomas analyzed [5]. Further, using the canonic transformation assay they confirmed that the mutations ultimately result in hyperactive proteins capable of transforming NIH 3T3 cells. Mutations in BRAF have later been found in a variety of cancers. Interestingly, mutations in genes within this signal transduction cascade are mutually exclusive as BRAF mutations occur only in tumors that are not affected by a mutation in hRas and vice versa [28]. The work by Davies and colleagues clearly showed that despite decades of research in the field even common genetic alterations had remained
Fig. 1. Strategies for the systematic identification of cancer genes by mutational profiling of tumor genomes.
S. Benvenuti et al. / FEBS Letters 579 (2005) 1884–1890
undetected but could be unveiled by large scale genetic analysis of cancer genomes [5]. 4.2. Discovery of cancer genes by mutational profiling of gene families So far, the most efficient approach for the identification of cancer genes in the post-genomic era has proven to be the systematic mutational profiling of gene families [27]. Genes can be categorized into families using different criteria [3]. Sequence regions similar to one another are typically translated into protein domains with similar biochemical properties, which in turn reflect overlapping function. Using this approach, it is therefore possible to predict all the genes with similar function present in the genome. Members of the same gene family can substitute one another in executing cellular functions as it has been shown in model organisms such as Drosophila melanogaster and Caenorhabditis elegans. Therefore, from a genetic perspective, the analysis of an entire gene family allows to systematically evaluate the relevance of a specific biochemical function, for example the phosphorylation of tyrosine residues in a given organism or cell type [29,30]. The parallel analysis of all genes controlling the execution of a particular biochemical reaction is especially appealing in cancer genetics [3]. If there is redundancy, only the simultaneous analysis of all the genes that can perform a specific reaction can identify which member is altered in a given tumor. If the mutated genes have overlapping functions or act within the same signaling pathways, the mutations will be mutually exclusive thus providing additional information on the model [28].
5. Analysis of the kinome and the phosphatome in cancers As previously stated cancer is a genetic disease caused by mutations in oncogenes, tumor suppressor or stability genes. Regardless of the specific genes involved, the final consequence of the molecular alteration affecting cancer genes is the deregulation of cell proliferation and cell death rate [8,9]. Many of the molecules in charge of controlling cellular homeostasis are switched on and off by the addition and removal of phosphate groups. This ultimately determines the relocalization and/or the activation of the downstream effectors thus resulting in the biological output. Ultimately, the phosphorylation state of a cellular molecule results from the balanced activity of kinases, responsible for adding the phosphate residues, and phosphatases, in charge of removing the phosphate groups [31,32]. Both these gene families are central in signaling pathways involved in tumorigenesis. Interestingly, in spite of their importance, many kinases and most phosphatases have not been studied in detail until very recently. Kinases and phosphatases are characterized by the presence of specific regions in their sequences known as kinase and phosphatase domains. Exploiting bioinformatic tools such as Hidden Markov models it is possible to identify all the genes belonging or closely related to each of these two families. Recently, the first genome wide mutational analyses of entire gene families were completed in colorectal tumours (Tables 1 and 2). These included the tyrosine kinases (tyrosine kinome) [27], the lipid kinases (lipid kinome) [7] and the tyrosine phosphatases (tyrosine phosphatome) [33]. A more limited analysis
1887 Table 1 Selected list of kinase genes genetically altered in human cancers Gene
Tumor
Mutation
ABL AKT2 BRAF EGFR EPHA3 ERBB2 FES FGFR2 FGFR3 FLT3 GUCY2F JAK2 KIT MET MLK4 NTRK1 NTRK2 NTRK3 PDGFRA PIK3CA RET
CML, ALL Ovarian, pancreatic Melanoma, colon Glioma, NSCL Colon Breast, ovarian, NSCL Colorectal Gastric Bladder, MM AML, ALL Colorectal AML, ALL GIST, AML Renal, HCC, colon Colon Thyroid Colon Fibrosarcoma, breast (Secretory), colon GIST Colon, brain, breast, colorectal, brain Thyroid, MEN
T A M M, D M M, A M M M, T M M T M M, A M, N T M T, M, N M M, A M,T
Alterations are in roman or italic if they were identified before or after the genome revolution respectively. AML, acute myelogenous leukaemia; MM, multiple myeloma; ALL, acute lymphocytic leukaemia; GIST, gastrointestinal stromal tumour; MEN, multiple endocrine neoplasia; HCC, hepato-cellular carcinoma; NSCL, non-small-cell lung; M, missense mutation; T, translocation; N, non-sense mutation; D, deletion and A, amplification. Table 2 Selected list of phosphatase genes genetically altered in human cancers Gene
Tumor
Mutation
PTEN PTPN3 PTPN13 PTPN14 PTPRF PTPRG PTPRT Shp2/PTPN11
Glioma, prostate, endometrial Colon Colon Colon Colon, breast, lung Colon Colon, gastric, lung JMML, AML, MDS
D, M, N M, N M, N, D M M M M, N, D M
Alterations are in roman or italic if they were identified before or after the genome revolution respectively. JMML, juvenile myelomonocytic leukaemia; AML, acute myelogenous leukaemia; MDS, myelodysplastic syndrome; M, missense mutation; T, translocation; N, nonsense mutation and D, deletion.
involving mainly tyrosine kinases receptors was also completed in non small cell lung cancers (NSCLC) [34]. Overall such genomic analyses have already demonstrated their utility in basic and clinical cancer research. For example, the systematic sequencing of gene families in colorectal cancers allowed the identification of mutations in 8 kinase and 6 phosphatase genes (Tables 1 and 2). Thus showing that more than 50% of colorectal cancers carry a mutation in a kinase or phosphatase gene [7,27,33]. Among the genes identified by this approach was PI3K, which is mutated not only in colorectal but also in breast, brain, gastric, lung and other tumor types. The frequency of mutations in the PI3KCA gene makes it one of the most commonly mutated oncogenes in human cancer. On the other end, the systematic mutational profiling of tyrosine kinases in NSCL cancers led to the discovery that clinical response to the EGFR small molecule inhibitor gefitinib was associated to activating mutations in the EGFR gene [34–37, see also further details below].
1888
Several conceptual implications can be drawn from the systematic mutational analysis of the kinome and phosphatome in colorectal cancers. The first is that the type of somatic mutations observed appears to be at least in part gene-family specific. For example, kinases tend to be altered by heterozygous missense mutations that mainly affect residues involved in the control of the catalytic activity. This suggests that the mutations are activating and operate by increasing the enzymatic activity of the corresponding proteins. This also implies that altered kinase genes mainly act as dominant oncogenes. On the other hand, tyrosine phosphatases are frequently altered by nonsense mutations that often affect both alleles. This suggests that the mutated phosphatase genes could act as tumor suppressors. The mutational profiling of the kinome and phosphatome also show that a relatively small number of genes are affected by somatic mutations, even in gene families that are thought to play a central role in tumorigenesis. This is different from what is obtained by the systematic analysis of gene expression
S. Benvenuti et al. / FEBS Letters 579 (2005) 1884–1890
whereby large sets of genes are typically found to be differentially expressed between normal and neoplastic tissues. Importantly, somatic mutations that have been selected during tumorigenesis are by definition causally related to tumor formation while in many cases the genes identified during differential gene expression studies are not. As discussed above the technological advances and the availability of the genome sequence make it possible to identify all the genes altered by somatic mutations in human cancers. Given that mutations appear to represent legitimate targets for anti-cancer drugs (see below) the mutational analysis of cancers should be expanded to include other gene families that could be suitable for therapeutic targeting. Among those are the serine/threonine kinases, the genes involved in G protein signaling and transcription factors genes. In addition, the systematic mutational analysis of the kinome and phosphatome have so far been completed only in a fraction of cancers and therefore these families remain to be analyzed in many tumor types.
Fig. 2. Examples of targeted cancer therapies based on protein kinases inhibitors tailored to specific genetic lesions underlying drug sensitivity.
S. Benvenuti et al. / FEBS Letters 579 (2005) 1884–1890
6. Mutational analysis of cancer genomes and individualized cancer therapies Cancer treatments have significantly improved in the last decade and the discovery that cancer is a genetic disease has played a central role in this revolution. Somatic mutations that have been selected during tumorigenesis are, by definition, causally related to tumor formation and therefore represent legitimate targets for anti-cancer drugs. For quite some time this seemingly obvious paradigm has been somewhat neglected. Only recently the genetic profiling of cancers has achieved a central role in the development of new drugs and in the selection of the patients who are likely to benefit from them [38]. Recent exciting results showing that therapeutic targeting is more effective in the context of somatic genetic alterations have been instrumental in this process. Compounds targeting protein tyrosine kinases are among those shown to achieve the most dramatic effect when tailored to cancers displaying specific genetic alterations (Fig. 2). For example, the treatment of chronic myeloid leukemia (CML) [39,40] and of advanced gastrointestinal stromal tumors (GISTs) [41] has changed since the introduction of imatinib. Imatinib mesylate, initially known as STI571 and commercialized as Gleevec, was developed to inhibit the Bcr-Abl protein. It also inhibits several other tyrosine kinases (TK) including the transmembrane receptor KIT [42,43] and the platelet-derived growth factor receptor (PDGFR). Remarkably, the often dramatic response of GIST patients to imatinib was shown to be associated with the mutational status of the KIT and PDGFR receptors in the corresponding tumors [44] (Fig. 2). Another notable example of cancer therapy tailored on the basis of a specific genetic alteration is treatment of recurrent breast cancers. In this case, trastuzumab (Herceptin) a monoclonal antibody directed against the extracellular domain of the HER-2 receptor is mainly effective in patients with amplification of the HER-2 gene [45]. The last and more recent example is the treatment of NSCL cancer, here the presence of activating mutations affecting the EGFR gene was found to underlie the response to gefitinib (IressaÒ) [34–37] (Fig. 2). The tailored treatment of cancer patients based on the mutated genes present in their tumors is revolutionizing the clinical oncology field. The systematic mutational profiling of cancer genomes is a key aspect of this process and will undoubtedly open up new diagnostic and therapeutic possibilities. Acknowledgments: We apologise to colleagues whose work was not cited owing to space constraints or our oversight. We thank Dr. Federica Di Nicolantonio and Dr. Asha Balakrishnan for critical reading the manuscript. Supported by The Italian Association for Cancer Research (AIRC) and The Italian University Technology and Research Ministry (MURST). S.B. is supported by an AIRC fellowship.
References [1] Vogelstein, B. and Kinzler, K.W. (2004) Cancer genes and the pathways they control. Nat. Med. 10, 789–799. [2] Vogelstein, B. and Kinzler, K.W. (1993) The multistep nature of cancer. Trends Genet. 9, 138–141. [3] Bardelli, A. and Velculescu, V.E. (2005) Mutational analysis of gene families in human cancer. Curr. Opin. Genet. Dev. 15, 5–12. [4] Futreal, P.A., Coin, L., Marshall, M., Down, T., Hubbard, T., Wooster, R., Rahman, N. and Stratton, M.R. (2004) A census of human cancer genes. Nat. Rev. Cancer 4, 177–183.
1889 [5] Davies, H., Bignell, G.R., Cox, C., Stephens, P., Edkins, S., Clegg, S., Teague, J., Woffendin, H., Garnett, M.J., Bottomley, W., Davis, N., Dicks, E., Ewing, R., Floyd, Y., Gray, K., Hall, S., Hawes, R., Hughes, J., Kosmidou, V., Menzies, A., Mould, C., Parker, A., Stevens, C., Watt, S., Hooper, S., Wilson, R., Jayatilake, H., Gusterson, B.A., Cooper, C., Shipley, J., Hargrave, D., Pritchard-Jones, K., Maitland, N., Chenevix-Trench, G., Riggins, G.J., Bigner, D.D., Palmieri, G., Cossu, A., Flanagan, A., Nicholson, A., Ho, J.W., Leung, S.Y., Yuen, S.T., Weber, B.L., Seigler, H.F., Darrow, T.L., Paterson, H., Marais, R., Marshall, C.J., Wooster, R., Stratton, M.R. and Futreal, P.A. (2002) Mutations of the BRAF gene in human cancer. Nature. [6] Bachman, K.E., Argani, P., Samuels, Y., Silliman, N., Ptak, J., Szabo, S., Konishi, H., Karakas, B., Blair, B.G., Lin, C., Peters, B.A., Velculescu, V.E. and Park, B.H. (2004) The PIK3CA gene is mutated with high frequency in human breast cancers. Cancer Biol. Ther. 3, 772–775. [7] Samuels, Y., Wang, Z., Bardelli, A., Silliman, N., Ptak, J., Szabo, S., Yan, H., Gazdar, A., Powell, S.M., Riggins, G.J., Willson, J.K., Markowitz, S., Kinzler, K.W., Vogelstein, B. and Velculescu, V.E. (2004) High frequency of mutations of the PIK3CA gene in human cancers. Science 304, 554. [8] Hanahan, D. and Weinberg, R.A. (2000) The hallmarks of cancer. Cell 100, 57–70. [9] Bernards, R. and Weinberg, R.A. (2002) A progression puzzle. Nature 418, 823. [10] Varmus, H.E. (1982) Form and function of retroviral proviruses. Science 216, 812–820. [11] Rous, P. (1979) A transmissible avian neoplasm. (Sarcoma of the common fowl) by Peyton Rous, M.D., Experimental Medicine for Sept. 1, 1910, vol. 12, pp. 696–705. J Exp Med 150, 738–753. [12] Neel, B.G., Hayward, W.S., Robinson, H.L., Fang, J. and Astrin, S.M. (1981) Avian leukosis virus-induced tumors have common proviral integration sites and synthesize discrete new RNAs: oncogenesis by promoter insertion. Cell 23, 323–334. [13] Hayward, W.S., Neel, B.G. and Astrin, S.M. (1981) Activation of a cellular oncogene by promoter insertion in ALV-induced lymphoid leukosis. Nature 290, 475–480. [14] Shih, C., Shilo, B.Z., Goldfarb, M.P., Dannenberg, A. and Weinberg, R.A. (1979) Passage of phenotypes of chemically transformed cells via transfection of DNA and chromatin. Proc. Natl. Acad. Sci. USA 76, 5714–5718. [15] Krontiris, T.G. and Cooper, G.M. (1981) Transforming activity of human tumor DNAs. Proc. Natl. Acad. Sci. USA 78, 1181– 1184. [16] Parada, L.F., Tabin, C.J., Shih, C. and Weinberg, R.A. (1982) Human EJ bladder carcinoma oncogene is homologue of Harvey sarcoma virus ras gene. Nature 297, 474–478. [17] Tabin, C.J., Bradley, S.M., Bargmann, C.I., Weinberg, R.A., Papageorge, A.G., Scolnick, E.M., Dhar, R., Lowy, D.R. and Chang, E.H. (1982) Mechanism of activation of a human oncogene. Nature 300, 143–149. [18] Reddy, E.P., Reynolds, R.K., Santos, E. and Barbacid, M. (1982) A point mutation is responsible for the acquisition of transforming properties by the T24 human bladder carcinoma oncogene. Nature 300, 149–152. [19] Falini, B. and Mason, D.Y. (2002) Proteins encoded by genes involved in chromosomal alterations in lymphoma and leukemia: clinical value of their detection by immunocytochemistry. Blood 99, 409–426. [20] Groffen, J., Stephenson, J.R., Heisterkamp, N., de Klein, A., Bartram, C.R. and Grosveld, G. (1984) Philadelphia chromosomal breakpoints are clustered within a limited region, bcr, on chromosome 22. Cell 36, 93–99. [21] Shtivelman, E., Lifshitz, B., Gale, R.P. and Canaani, E. (1985) Fused transcript of abl and bcr genes in chronic myelogenous leukaemia. Nature 315, 550–554. [22] Harris, H. (1988) The analysis of malignancy by cell fusion: the position in 1988. Cancer Res. 48, 3302–3306. [23] Knudson Jr., A.G. (1971) Mutation and cancer: statistical study of retinoblastoma. Proc. Natl. Acad. Sci. USA 68, 820–823. [24] Orye, E., Delbeke, M.J. and Vandenabeele, B. (1974) Retinoblastoma and long arm delection of chromosome 3. Attempts to define the deleted segment. Clin. Genet. 5, 457–464.
1890 [25] Francke, U. (1976) Retinoblastoma and chromosome 13. Birth Defects Orig. Artic. Ser. 12, 131–134. [26] Kato, J., Matsushime, H., Hiebert, S.W., Ewen, M.E. and Sherr, C.J. (1993) Direct binding of cyclin D to the retinoblastoma gene product (pRb) and pRb phosphorylation by the cyclin Ddependent kinase CDK4. Genes Dev. 7, 331–342. [27] Bardelli, A., Parsons, D.W., Silliman, N., Ptak, J., Szabo, S., Saha, S., Markowitz, S., Willson, J.K., Parmigiani, G., Kinzler, K.W., Vogelstein, B. and Velculescu, V.E. (2003) Mutational analysis of the tyrosine kinome in colorectal cancers. Science 300, 949. [28] Rajagopalan, H., Bardelli, A., Lengauer, C., Kinzler, K.W., Vogelstein, B. and Velculescu, V.E. (2002) Tumorigenesis: RAF/ RAS oncogenes and mismatch-repair status. Nature 418, 934. [29] Manning, G., Whyte, D.B., Martinez, R., Hunter, T. and Sudarsanam, S. (2002) The protein kinase complement of the human genome. Science 298, 1912–1934. [30] Blume-Jensen, P. and Hunter, T. (2001) Oncogenic kinase signalling. Nature 411, 355–365. [31] Alonso, A., Sasin, J., Bottini, N., Friedberg, I., Osterman, A., Godzik, A., Hunter, T., Dixon, J. and Mustelin, T. (2004) Protein tyrosine phosphatases in the human genome. Cell 117, 699–711. [32] Forrest, A.R., Ravasi, T., Taylor, D., Huber, T., Hume, D.A. and Grimmond, S. (2003) Phosphoregulators: protein kinases and protein phosphatases of mouse. Genome Res. 13, 1443–1454. [33] Wang, Z., Shen, D., Parsons, D.W., Bardelli, A., Sager, J., Szabo, S., Ptak, J., Silliman, N., Peters, B.A., van der Heijden, M.S., Parmigiani, G., Yan, H., Wang, T.L., Riggins, G., Powell, S.M., Willson, J.K., Markowitz, S., Kinzler, K.W., Vogelstein, B. and Velculescu, V.E. (2004) Mutational analysis of the tyrosine phosphatome in colorectal cancers. Science 304, 1164–1166. [34] Paez, J.G., Janne, P.A., Lee, J.C., Tracy, S., Greulich, H., Gabriel, S., Herman, P., Kaye, F.J., Lindeman, N., Boggon, T.J., Naoki, K., Sasaki, H., Fujii, Y., Eck, M.J., Sellers, W.R., Johnson, B.E. and Meyerson, M. (2004) EGFR mutations in lung cancer: correlation with clinical response to gefitinib therapy. Science 304, 1497–1500. [35] Lynch, T.J., Bell, D.W., Sordella, R., Gurubhagavatula, S., Okimoto, R.A., Brannigan, B.W., Harris, P.L., Haserlat, S.M., Supko, J.G., Haluska, F.G., Louis, D.N., Christiani, D.C., Settleman, J. and Haber, D.A. (2004) Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N. Engl. J. Med. 350, 2129–2139.
S. Benvenuti et al. / FEBS Letters 579 (2005) 1884–1890 [36] Huang, S.F., Liu, H.P., Li, L.H., Ku, Y.C., Fu, Y.N., Tsai, H.Y., Chen, Y.T., Lin, Y.F., Chang, W.C., Kuo, H.P., Wu, Y.C., Chen, Y.R. and Tsai, S.F. (2004) High frequency of epidermal growth factor receptor mutations with complex patterns in non-small cell lung cancers related to gefitinib responsiveness in Taiwan. Clin. Cancer Res. 10, 8195–8203. [37] Marx, J. (2004) Medicine. Why a new cancer drug works well, in some patients. Science 304, 658–659. [38] Strausberg, R.L., Simpson, A.J., Old, L.J. and Riggins, G.J. (2004) Oncogenomics and the development of new cancer therapies. Nature 429, 469–474. [39] Druker, B.J., Talpaz, M., Resta, D.J., Peng, B., Buchdunger, E., Ford, J.M., Lydon, N.B., Kantarjian, H., Capdeville, R., OhnoJones, S. and Sawyers, C.L. (2001) Efficacy and safety of a specific inhibitor of the BCR-ABL tyrosine kinase in chronic myeloid leukemia. N. Engl. J. Med. 344, 1031–1037. [40] Druker, B.J. (2004) Imatinib as a paradigm of targeted therapies. Adv. Cancer Res. 91, 1–30. [41] van Oosterom, A.T., Judson, I., Verweij, J., Stroobants, S., Donato di Paola, E., Dimitrijevic, S., Martens, M., Webb, A., Sciot, R., Van Glabbeke, M., Silberman, S. and Nielsen, O.S. (2001) Safety and efficacy of imatinib (STI571) in metastatic gastrointestinal stromal tumours: a phase I study. Lancet 358, 1421–1423. [42] Hirota, S., Nishida, T., Isozaki, K., Taniguchi, M., Nakamura, J., Okazaki, T. and Kitamura, Y. (2001) Gain-of-function mutation at the extracellular domain of KIT in gastrointestinal stromal tumours. J. Pathol. 193, 505–510. [43] Rubin, B.P., Singer, S., Tsao, C., Duensing, A., Lux, M.L., Ruiz, R., Hibbard, M.K., Chen, C.J., Xiao, S., Tuveson, D.A., Demetri, G.D., Fletcher, C.D. and Fletcher, J.A. (2001) KIT activation is a ubiquitous feature of gastrointestinal stromal tumors. Cancer Res. 61, 8118–8121. [44] Heinrich, M.C., Corless, C.L., Demetri, G.D., Blanke, C.D., von Mehren, M., Joensuu, H., McGreevey, L.S., Chen, C.J., Van den Abbeele, A.D., Druker, B.J., Kiese, B., Eisenberg, B., Roberts, P.J., Singer, S., Fletcher, C.D., Silberman, S., Dimitrijevic, S. and Fletcher, J.A. (2003) Kinase mutations and imatinib response in patients with metastatic gastrointestinal stromal tumor. J. Clin. Oncol. 21, 4342–4349. [45] Leyland-Jones, B. (2002) Trastuzumab: hopes and realities. Lancet Oncol. 3, 137–144.
FEBS 29348
Minireview
Identification of cancer genes by mutational profiling of tumor genomes Silvia Benvenuti, Sabrina Arena, Alberto Bardelli* The Oncogenomics Center, Institute for Cancer Research and Treatment (IRCC), University of Torino Medical School, Candiolo (To), Italy Accepted 3 February 2005 Available online 16 February 2005 Edited by Robert Russell and Giulio Superti-Furga
Abstract It is now widely accepted that cancer is a genetic disease and that alterations in the DNA sequence underlie the development of every neoplasm. The identification of mutated genes that are causally implicated in oncogenesis (Ôcancer genesÕ) has been a major goal in medical sciences for the last two decades. The availability of the human genome sequence coupled to the introduction of high throughput sequencing technologies has created an unprecedented opportunity in this field. It is now possible to perform mutational studies of entire cancer genomes thus providing a complete description of mutations underlying human oncogenesis. The recent identification of high frequency mutations in the BRAF and PI3K genes suggests that many more cancer genes remain to be discovered. In this review, we consider how the systematic mutational analysis of gene families in individual neoplasms has led to the identification of a number of cancer genes and how this information is influencing the treatment of cancer. Ó 2005 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. Keywords: Cancer genome; Oncogenomics; Gene family; Tailored therapies; Mutational analysis
1. Cancer: a genetic disease In the living cell, DNA undergoes frequent chemical changes. The majority of these changes are quickly repaired; those that are not result in a mutation. Cancer progression results by the serial accumulation of genetic accidents in somatic cells [1]. The genes that are causally involved in oncogenesis are known as cancer genes. DNA alterations that convey tumorassociated properties are favored by natural selection that ultimately drives the acquisition of the neoplastic properties [2]. Initially, these involve the deregulation of proliferation and resistance to apoptosis. Subsequently, some cancer cells acquire the ability to invade and colonize other tissues giving rise to metastasis that are primarily responsible for cancer-associated mortality. A primary way to discriminate whether a mutated gene plays a causal role in cancer progression is that the presence of mutations in the gene, rather than being attributable to chance, are thought to result in a growth advantage on the cell population from which the cancer developed [3].
* Corresponding author. E-mail address: [email protected] (A. Bardelli).
Since the first speculations on their existence, more than 100 cancer genes have been identified [4]. However, even commonly mutated cancer genes have remained undetected until very recently. For example, only in the last 3 years the BRAF [5] and the PI3K [6,7] genes have been shown to be frequently mutated in a large fraction of tumors and to play a causative role in tumorigenesis. It is therefore reasonable to assume that many more cancer genes remain to be identified. 2. Cancer genes discovery and the human genome project A simplistic interpretation divides cancer genes into three broad categories: oncogenes, tumor suppressors and stability genes. Oncogenes promote cell growth and when mutated result in constitutively or abnormally active proteins; such mutations are mainly of the dominant type. Tumor suppressors genes are negative controllers of cell cycle progression and the mutations that affect them are typically recessive. A third category of cancer genes has been proposed more recently. This comprises the so-called stability or caretaker genes. These genes are not directly involved in tumorigenesis but when altered they contribute to cancer by exposing cells to an abnormally high mutation rate. This feature ultimately leads to oncogene activation or tumor suppressor inactivation. Regardless of the specific genes involved, the end result of the molecular alteration affecting cancer genes is the deregulation of cellular homeostasis [8,9]. As human cancer is a genetic disease, the completion of the human genome project has offered unprecedented opportunities for cancer diagnosis, prognosis and treatment. The field of cancer genes discovery has been one of the first to be affected by this historic revolution that will ultimately allow the identification of all genetic alterations underlying the origin and evolution of tumors. Below we consider how systematic mutational analysis of gene families in individual neoplasms has led to the identification of a number of cancer genes and how this information has already affected the treatment of cancer.
3. Discovery of cancer genes before the human genome project 3.1. The oncogenesÕs hunt Until the completion of the human genome project a number of approaches have led to the identification of genes that when mutated are causally implicated in oncogenesis. Key breakthrough in the discovery of oncogenes resulted from studies on retroviruses. Retroviruses are RNA tumor
0014-5793/$30.00 Ó 2005 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2005.02.015
S. Benvenuti et al. / FEBS Letters 579 (2005) 1884–1890
viruses capable of infecting vertebrate cells and whose genome can be retrotranscribed into DNA. They can be divided into two groups: acute transforming and weakly oncogenic retroviruses [10]. The former induce tumor formation very rapidly (days) after infection and were found to contain modified copies of cellular genes that have been incorporated into the viral genome during retroviral transduction process. Rous and colleagues obtained key clues on acute transforming oncogenes in 1911. For the first time they reported that sarcomas could be induced in chickens by cell free filtrates of a sarcoma taken from another chicken. However, the region responsible for the oncogenicity was isolated only 60 years later. It was v-src, a mutated copy of the human c-src gene [11]. The weakly oncogenic retroviruses cause tumors only after a long incubation period (months). Contrary to acute transforming retroviruses, weakly oncogenic retroviruses do not contain viral oncogenes. They were also central in the identification of cancer genes. In fact, weakly oncogenic retroviruses induce tumors via a process called insertional mutagenesis. Specifically, they integrate in the host genome close to an oncogene thus ensuing its abnormal expression which is typically driven by the transactivating elements contained in the virus genome [12,13]. In addition to the studies on retroviruses, pioneer experiments designed to identify dominant mutations were introduced at the beginning of the 1980s together with the advancement of cell transfection techniques [14,15]. In particular a specific oncogene hunting technique (known as the ‘‘NIH 3T3 transformation assay’’) was established. This technique is based on the extraction of genomic DNA from tumor samples that is fractionated and then introduced into NIH 3T3 mouse fibroblasts. If an oncogene is present colonies of hyper proliferating cells appear in the cell culture. Each colony is generated from the clonal expansion of a single cell, which gained a proliferative advantage after ectopic gene expression and as a result has lost the cell–cell contact inhibitory control mechanisms. The process is then serially repeated by extracting the genomic DNA from the transformed foci ultimately leading to the identification of the human gene responsible for the tumorigenic properties. Several oncogenes have been discovered using the NIH 3T3 transformation assay. The methodology was initially applied to a human bladder carcinoma cell line [16,17] and has lead to the identification of one of the most common genetic lesions found in a variety of cancers. In the specific case, the genetic alteration turned out to be a single point mutation which resulted in the incorporation of Valine instead of Glycine as the 12th amino acid residue in the Ras protein [18]. hRas mutations are present in more than 25% of all human cancers. hRas encodes for small monomeric GTPases responsible for delivering external signals captured by the cells via their membrane receptors to the cell nucleus. The point mutations affecting hRas generate a protein, which fails to efficiently hydrolyze GTP and therefore remains constitutively active resulting in deregulated cellular proliferation. Significant improvements in the identification of oncogenes also stemmed from the karyotypic analysis of human cancers. In particular, tumors of the haematopoietic system such as leukemias were shown to carry specific chromosomal transloca-
1885
tions. The introduction of sophisticated chromosomal banding techniques at the beginning of the 1970s allowed the precise description of those changes. Breakpoints (both translocations and inversions), amplifications and deletions were then characterized in detail [19]. A classical example of the power of this strategy is the identification of the genetic alteration responsible for chronic myeloid leukemia (CML). In this case, karyotype analysis unequivocally determined that cells of patients affected by CML are characterized by a translocation between chromosomes 9 and 22 (t(9;22)(q34;q11)) [20]. The translocation generates the Philadelphia chromosome. In this chimeric chromosome, the 5 0 segment of the BCR gene is juxtaposed to the 3 0 region of the ABL gene and results in a constitutively active 210 kDa fusion protein: the tyrosine kinase BCR-ABL [21].
3.2. The tumor suppressor’s hunt The behavior of cancer cells suggested that not all tumor properties could be explained by the deregulated over activation of oncogenes, but the inactivation of other genes was also crucial. Mutations affecting tumor suppressor genes typically reduce or abolish the activity of the corresponding protein product. Examples are missense mutations at residues that are essential for the catalytic activity, mutations that result in a truncated protein, deletions or insertions of various sizes and epigenetic silencing. Usually, it is necessary to eliminate both copies of a tumor suppressor gene to inactivate it. Identification of genes inactivated during tumorigenesis is a challenging task. An important contribution to the identification of tumor suppressor genes was made by Harris and colleagues [22]. Using cell fusion experiments they observed that growth of murine tumors in synergic animals could be suppressed when malignant cells were fused to non-malignant cells. Moreover, it was found that tumorigenicity was associated with specific chromosomal losses and that the introduction of single chromosomes could reverse malignancy. A breakthrough in the discovery of the first tumor suppressor gene came from retinoblastoma, a rare tumor that occurs in childhood. This tumor develops from neural precursor cells in the immature retina, usually sporadically, but in some families, it displays an autosomal dominant mode of inheritance [23]. In the hereditary type, multiple tumors occur independently, affecting both eyes. In the sporadic form only one eye is affected, and by only one tumor. A key clue on the putative gene responsible for retinoblastoma was obtained when the karyotype of patients with the inherited form was shown to be abnormal. The karyotype was characterized by a deletion of chromosome 13 band q14 [24,25]. Deletions of the same locus were also observed in tumor cells from patients affected by the nonhereditary form, suggesting that the cancer was possibly caused by loss of a critical gene located in that particular chromosomal region. Subsequently, the locus was characterized in detail and the Rb gene was identified. Loss of Rb is one of the crucial steps in cancer progression. The Rb gene encodes the Rb protein, a key regulator of the cell cycle. pRb undergoes cell-cycle dependent phosphorylation, which is dependent on the activity of several cyclin-dependent kinases and occurs in a stepwise manner [26]. When hypo-
1886
S. Benvenuti et al. / FEBS Letters 579 (2005) 1884–1890
phosphorylated, pRb associates with and sequesters the E2F family of transcription factors. Upon progression through the cell cycle pRB becomes hyperphosphorylated and therefore releases E2F factors which, in turn, transactivate genes containing E2F recognition sites in their regulatory sequences [26]. Rb mutations are thought to influence the stability of the genome of cancer cells.
4. Discovery of cancer genes after the human genome project The completion of the human genome and the introduction of high throughput sequencing technologies allow the discovery of cancer genes through systematic approaches. In principle, it is now possible to sequence every gene in a given cancer genome and identify every mutation. Already, a number of laboratories have started the systematic mutational profiling of cancer genomes. In order to achieve this goal a new generation of bioinformatic tools has been developed to extract exonic and intronic sequences for each gene of interest from the available databases. The sequences are then exploited for high throughput design of exon specific amplification and sequencing primers. Automated procedures for amplification and direct sequencing of exons from tumour derived genomic DNA have also been implemented [5,27]. These require robotic handling of PCRs, purification and sequencing of PCR products. Specific mutational analysis algorithms have made it possible to identify both heterozygous and homozygous mutations in high throughput fashion (Fig. 1). Using softwareÕs programs based on these algorithms high throughput mutational analysis can be achieved by comparing
chromatograms from different tumors to a reference sequence [27]. 4.1. Discovery of cancer genes by dissecting oncogenic signal pathways Among the strategies devised to identify cancer genes in the post genomic era a genetic analysis was implemented by Davies and colleagues [5] to identify cancer-associated mutations by dissecting signaling pathways. Specifically, they decided to search for mutations in members of signaling pathways known to involve at least one gene previously found mutated in human cancers. They started by analyzing all the members of the RAS–RAF–MEK–ERK–MAP kinase pathway [5,28]. In this approach the initial high throughput screening for mutations was achieved using a capillary-based heteroduplex technology and not by the direct comparison of DNA sequence traces. The second approach was implemented later and, given its better performance, is currently the standard for large scale sequencing projects [27]. Using this strategy, the authors reported genetic alterations in the BRAF gene in a variety of cancers, in particular in 66% of malignant melanomas analyzed [5]. Further, using the canonic transformation assay they confirmed that the mutations ultimately result in hyperactive proteins capable of transforming NIH 3T3 cells. Mutations in BRAF have later been found in a variety of cancers. Interestingly, mutations in genes within this signal transduction cascade are mutually exclusive as BRAF mutations occur only in tumors that are not affected by a mutation in hRas and vice versa [28]. The work by Davies and colleagues clearly showed that despite decades of research in the field even common genetic alterations had remained
Fig. 1. Strategies for the systematic identification of cancer genes by mutational profiling of tumor genomes.
S. Benvenuti et al. / FEBS Letters 579 (2005) 1884–1890
undetected but could be unveiled by large scale genetic analysis of cancer genomes [5]. 4.2. Discovery of cancer genes by mutational profiling of gene families So far, the most efficient approach for the identification of cancer genes in the post-genomic era has proven to be the systematic mutational profiling of gene families [27]. Genes can be categorized into families using different criteria [3]. Sequence regions similar to one another are typically translated into protein domains with similar biochemical properties, which in turn reflect overlapping function. Using this approach, it is therefore possible to predict all the genes with similar function present in the genome. Members of the same gene family can substitute one another in executing cellular functions as it has been shown in model organisms such as Drosophila melanogaster and Caenorhabditis elegans. Therefore, from a genetic perspective, the analysis of an entire gene family allows to systematically evaluate the relevance of a specific biochemical function, for example the phosphorylation of tyrosine residues in a given organism or cell type [29,30]. The parallel analysis of all genes controlling the execution of a particular biochemical reaction is especially appealing in cancer genetics [3]. If there is redundancy, only the simultaneous analysis of all the genes that can perform a specific reaction can identify which member is altered in a given tumor. If the mutated genes have overlapping functions or act within the same signaling pathways, the mutations will be mutually exclusive thus providing additional information on the model [28].
5. Analysis of the kinome and the phosphatome in cancers As previously stated cancer is a genetic disease caused by mutations in oncogenes, tumor suppressor or stability genes. Regardless of the specific genes involved, the final consequence of the molecular alteration affecting cancer genes is the deregulation of cell proliferation and cell death rate [8,9]. Many of the molecules in charge of controlling cellular homeostasis are switched on and off by the addition and removal of phosphate groups. This ultimately determines the relocalization and/or the activation of the downstream effectors thus resulting in the biological output. Ultimately, the phosphorylation state of a cellular molecule results from the balanced activity of kinases, responsible for adding the phosphate residues, and phosphatases, in charge of removing the phosphate groups [31,32]. Both these gene families are central in signaling pathways involved in tumorigenesis. Interestingly, in spite of their importance, many kinases and most phosphatases have not been studied in detail until very recently. Kinases and phosphatases are characterized by the presence of specific regions in their sequences known as kinase and phosphatase domains. Exploiting bioinformatic tools such as Hidden Markov models it is possible to identify all the genes belonging or closely related to each of these two families. Recently, the first genome wide mutational analyses of entire gene families were completed in colorectal tumours (Tables 1 and 2). These included the tyrosine kinases (tyrosine kinome) [27], the lipid kinases (lipid kinome) [7] and the tyrosine phosphatases (tyrosine phosphatome) [33]. A more limited analysis
1887 Table 1 Selected list of kinase genes genetically altered in human cancers Gene
Tumor
Mutation
ABL AKT2 BRAF EGFR EPHA3 ERBB2 FES FGFR2 FGFR3 FLT3 GUCY2F JAK2 KIT MET MLK4 NTRK1 NTRK2 NTRK3 PDGFRA PIK3CA RET
CML, ALL Ovarian, pancreatic Melanoma, colon Glioma, NSCL Colon Breast, ovarian, NSCL Colorectal Gastric Bladder, MM AML, ALL Colorectal AML, ALL GIST, AML Renal, HCC, colon Colon Thyroid Colon Fibrosarcoma, breast (Secretory), colon GIST Colon, brain, breast, colorectal, brain Thyroid, MEN
T A M M, D M M, A M M M, T M M T M M, A M, N T M T, M, N M M, A M,T
Alterations are in roman or italic if they were identified before or after the genome revolution respectively. AML, acute myelogenous leukaemia; MM, multiple myeloma; ALL, acute lymphocytic leukaemia; GIST, gastrointestinal stromal tumour; MEN, multiple endocrine neoplasia; HCC, hepato-cellular carcinoma; NSCL, non-small-cell lung; M, missense mutation; T, translocation; N, non-sense mutation; D, deletion and A, amplification. Table 2 Selected list of phosphatase genes genetically altered in human cancers Gene
Tumor
Mutation
PTEN PTPN3 PTPN13 PTPN14 PTPRF PTPRG PTPRT Shp2/PTPN11
Glioma, prostate, endometrial Colon Colon Colon Colon, breast, lung Colon Colon, gastric, lung JMML, AML, MDS
D, M, N M, N M, N, D M M M M, N, D M
Alterations are in roman or italic if they were identified before or after the genome revolution respectively. JMML, juvenile myelomonocytic leukaemia; AML, acute myelogenous leukaemia; MDS, myelodysplastic syndrome; M, missense mutation; T, translocation; N, nonsense mutation and D, deletion.
involving mainly tyrosine kinases receptors was also completed in non small cell lung cancers (NSCLC) [34]. Overall such genomic analyses have already demonstrated their utility in basic and clinical cancer research. For example, the systematic sequencing of gene families in colorectal cancers allowed the identification of mutations in 8 kinase and 6 phosphatase genes (Tables 1 and 2). Thus showing that more than 50% of colorectal cancers carry a mutation in a kinase or phosphatase gene [7,27,33]. Among the genes identified by this approach was PI3K, which is mutated not only in colorectal but also in breast, brain, gastric, lung and other tumor types. The frequency of mutations in the PI3KCA gene makes it one of the most commonly mutated oncogenes in human cancer. On the other end, the systematic mutational profiling of tyrosine kinases in NSCL cancers led to the discovery that clinical response to the EGFR small molecule inhibitor gefitinib was associated to activating mutations in the EGFR gene [34–37, see also further details below].
1888
Several conceptual implications can be drawn from the systematic mutational analysis of the kinome and phosphatome in colorectal cancers. The first is that the type of somatic mutations observed appears to be at least in part gene-family specific. For example, kinases tend to be altered by heterozygous missense mutations that mainly affect residues involved in the control of the catalytic activity. This suggests that the mutations are activating and operate by increasing the enzymatic activity of the corresponding proteins. This also implies that altered kinase genes mainly act as dominant oncogenes. On the other hand, tyrosine phosphatases are frequently altered by nonsense mutations that often affect both alleles. This suggests that the mutated phosphatase genes could act as tumor suppressors. The mutational profiling of the kinome and phosphatome also show that a relatively small number of genes are affected by somatic mutations, even in gene families that are thought to play a central role in tumorigenesis. This is different from what is obtained by the systematic analysis of gene expression
S. Benvenuti et al. / FEBS Letters 579 (2005) 1884–1890
whereby large sets of genes are typically found to be differentially expressed between normal and neoplastic tissues. Importantly, somatic mutations that have been selected during tumorigenesis are by definition causally related to tumor formation while in many cases the genes identified during differential gene expression studies are not. As discussed above the technological advances and the availability of the genome sequence make it possible to identify all the genes altered by somatic mutations in human cancers. Given that mutations appear to represent legitimate targets for anti-cancer drugs (see below) the mutational analysis of cancers should be expanded to include other gene families that could be suitable for therapeutic targeting. Among those are the serine/threonine kinases, the genes involved in G protein signaling and transcription factors genes. In addition, the systematic mutational analysis of the kinome and phosphatome have so far been completed only in a fraction of cancers and therefore these families remain to be analyzed in many tumor types.
Fig. 2. Examples of targeted cancer therapies based on protein kinases inhibitors tailored to specific genetic lesions underlying drug sensitivity.
S. Benvenuti et al. / FEBS Letters 579 (2005) 1884–1890
6. Mutational analysis of cancer genomes and individualized cancer therapies Cancer treatments have significantly improved in the last decade and the discovery that cancer is a genetic disease has played a central role in this revolution. Somatic mutations that have been selected during tumorigenesis are, by definition, causally related to tumor formation and therefore represent legitimate targets for anti-cancer drugs. For quite some time this seemingly obvious paradigm has been somewhat neglected. Only recently the genetic profiling of cancers has achieved a central role in the development of new drugs and in the selection of the patients who are likely to benefit from them [38]. Recent exciting results showing that therapeutic targeting is more effective in the context of somatic genetic alterations have been instrumental in this process. Compounds targeting protein tyrosine kinases are among those shown to achieve the most dramatic effect when tailored to cancers displaying specific genetic alterations (Fig. 2). For example, the treatment of chronic myeloid leukemia (CML) [39,40] and of advanced gastrointestinal stromal tumors (GISTs) [41] has changed since the introduction of imatinib. Imatinib mesylate, initially known as STI571 and commercialized as Gleevec, was developed to inhibit the Bcr-Abl protein. It also inhibits several other tyrosine kinases (TK) including the transmembrane receptor KIT [42,43] and the platelet-derived growth factor receptor (PDGFR). Remarkably, the often dramatic response of GIST patients to imatinib was shown to be associated with the mutational status of the KIT and PDGFR receptors in the corresponding tumors [44] (Fig. 2). Another notable example of cancer therapy tailored on the basis of a specific genetic alteration is treatment of recurrent breast cancers. In this case, trastuzumab (Herceptin) a monoclonal antibody directed against the extracellular domain of the HER-2 receptor is mainly effective in patients with amplification of the HER-2 gene [45]. The last and more recent example is the treatment of NSCL cancer, here the presence of activating mutations affecting the EGFR gene was found to underlie the response to gefitinib (IressaÒ) [34–37] (Fig. 2). The tailored treatment of cancer patients based on the mutated genes present in their tumors is revolutionizing the clinical oncology field. The systematic mutational profiling of cancer genomes is a key aspect of this process and will undoubtedly open up new diagnostic and therapeutic possibilities. Acknowledgments: We apologise to colleagues whose work was not cited owing to space constraints or our oversight. We thank Dr. Federica Di Nicolantonio and Dr. Asha Balakrishnan for critical reading the manuscript. Supported by The Italian Association for Cancer Research (AIRC) and The Italian University Technology and Research Ministry (MURST). S.B. is supported by an AIRC fellowship.
References [1] Vogelstein, B. and Kinzler, K.W. (2004) Cancer genes and the pathways they control. Nat. Med. 10, 789–799. [2] Vogelstein, B. and Kinzler, K.W. (1993) The multistep nature of cancer. Trends Genet. 9, 138–141. [3] Bardelli, A. and Velculescu, V.E. (2005) Mutational analysis of gene families in human cancer. Curr. Opin. Genet. Dev. 15, 5–12. [4] Futreal, P.A., Coin, L., Marshall, M., Down, T., Hubbard, T., Wooster, R., Rahman, N. and Stratton, M.R. (2004) A census of human cancer genes. Nat. Rev. Cancer 4, 177–183.
1889 [5] Davies, H., Bignell, G.R., Cox, C., Stephens, P., Edkins, S., Clegg, S., Teague, J., Woffendin, H., Garnett, M.J., Bottomley, W., Davis, N., Dicks, E., Ewing, R., Floyd, Y., Gray, K., Hall, S., Hawes, R., Hughes, J., Kosmidou, V., Menzies, A., Mould, C., Parker, A., Stevens, C., Watt, S., Hooper, S., Wilson, R., Jayatilake, H., Gusterson, B.A., Cooper, C., Shipley, J., Hargrave, D., Pritchard-Jones, K., Maitland, N., Chenevix-Trench, G., Riggins, G.J., Bigner, D.D., Palmieri, G., Cossu, A., Flanagan, A., Nicholson, A., Ho, J.W., Leung, S.Y., Yuen, S.T., Weber, B.L., Seigler, H.F., Darrow, T.L., Paterson, H., Marais, R., Marshall, C.J., Wooster, R., Stratton, M.R. and Futreal, P.A. (2002) Mutations of the BRAF gene in human cancer. Nature. [6] Bachman, K.E., Argani, P., Samuels, Y., Silliman, N., Ptak, J., Szabo, S., Konishi, H., Karakas, B., Blair, B.G., Lin, C., Peters, B.A., Velculescu, V.E. and Park, B.H. (2004) The PIK3CA gene is mutated with high frequency in human breast cancers. Cancer Biol. Ther. 3, 772–775. [7] Samuels, Y., Wang, Z., Bardelli, A., Silliman, N., Ptak, J., Szabo, S., Yan, H., Gazdar, A., Powell, S.M., Riggins, G.J., Willson, J.K., Markowitz, S., Kinzler, K.W., Vogelstein, B. and Velculescu, V.E. (2004) High frequency of mutations of the PIK3CA gene in human cancers. Science 304, 554. [8] Hanahan, D. and Weinberg, R.A. (2000) The hallmarks of cancer. Cell 100, 57–70. [9] Bernards, R. and Weinberg, R.A. (2002) A progression puzzle. Nature 418, 823. [10] Varmus, H.E. (1982) Form and function of retroviral proviruses. Science 216, 812–820. [11] Rous, P. (1979) A transmissible avian neoplasm. (Sarcoma of the common fowl) by Peyton Rous, M.D., Experimental Medicine for Sept. 1, 1910, vol. 12, pp. 696–705. J Exp Med 150, 738–753. [12] Neel, B.G., Hayward, W.S., Robinson, H.L., Fang, J. and Astrin, S.M. (1981) Avian leukosis virus-induced tumors have common proviral integration sites and synthesize discrete new RNAs: oncogenesis by promoter insertion. Cell 23, 323–334. [13] Hayward, W.S., Neel, B.G. and Astrin, S.M. (1981) Activation of a cellular oncogene by promoter insertion in ALV-induced lymphoid leukosis. Nature 290, 475–480. [14] Shih, C., Shilo, B.Z., Goldfarb, M.P., Dannenberg, A. and Weinberg, R.A. (1979) Passage of phenotypes of chemically transformed cells via transfection of DNA and chromatin. Proc. Natl. Acad. Sci. USA 76, 5714–5718. [15] Krontiris, T.G. and Cooper, G.M. (1981) Transforming activity of human tumor DNAs. Proc. Natl. Acad. Sci. USA 78, 1181– 1184. [16] Parada, L.F., Tabin, C.J., Shih, C. and Weinberg, R.A. (1982) Human EJ bladder carcinoma oncogene is homologue of Harvey sarcoma virus ras gene. Nature 297, 474–478. [17] Tabin, C.J., Bradley, S.M., Bargmann, C.I., Weinberg, R.A., Papageorge, A.G., Scolnick, E.M., Dhar, R., Lowy, D.R. and Chang, E.H. (1982) Mechanism of activation of a human oncogene. Nature 300, 143–149. [18] Reddy, E.P., Reynolds, R.K., Santos, E. and Barbacid, M. (1982) A point mutation is responsible for the acquisition of transforming properties by the T24 human bladder carcinoma oncogene. Nature 300, 149–152. [19] Falini, B. and Mason, D.Y. (2002) Proteins encoded by genes involved in chromosomal alterations in lymphoma and leukemia: clinical value of their detection by immunocytochemistry. Blood 99, 409–426. [20] Groffen, J., Stephenson, J.R., Heisterkamp, N., de Klein, A., Bartram, C.R. and Grosveld, G. (1984) Philadelphia chromosomal breakpoints are clustered within a limited region, bcr, on chromosome 22. Cell 36, 93–99. [21] Shtivelman, E., Lifshitz, B., Gale, R.P. and Canaani, E. (1985) Fused transcript of abl and bcr genes in chronic myelogenous leukaemia. Nature 315, 550–554. [22] Harris, H. (1988) The analysis of malignancy by cell fusion: the position in 1988. Cancer Res. 48, 3302–3306. [23] Knudson Jr., A.G. (1971) Mutation and cancer: statistical study of retinoblastoma. Proc. Natl. Acad. Sci. USA 68, 820–823. [24] Orye, E., Delbeke, M.J. and Vandenabeele, B. (1974) Retinoblastoma and long arm delection of chromosome 3. Attempts to define the deleted segment. Clin. Genet. 5, 457–464.
1890 [25] Francke, U. (1976) Retinoblastoma and chromosome 13. Birth Defects Orig. Artic. Ser. 12, 131–134. [26] Kato, J., Matsushime, H., Hiebert, S.W., Ewen, M.E. and Sherr, C.J. (1993) Direct binding of cyclin D to the retinoblastoma gene product (pRb) and pRb phosphorylation by the cyclin Ddependent kinase CDK4. Genes Dev. 7, 331–342. [27] Bardelli, A., Parsons, D.W., Silliman, N., Ptak, J., Szabo, S., Saha, S., Markowitz, S., Willson, J.K., Parmigiani, G., Kinzler, K.W., Vogelstein, B. and Velculescu, V.E. (2003) Mutational analysis of the tyrosine kinome in colorectal cancers. Science 300, 949. [28] Rajagopalan, H., Bardelli, A., Lengauer, C., Kinzler, K.W., Vogelstein, B. and Velculescu, V.E. (2002) Tumorigenesis: RAF/ RAS oncogenes and mismatch-repair status. Nature 418, 934. [29] Manning, G., Whyte, D.B., Martinez, R., Hunter, T. and Sudarsanam, S. (2002) The protein kinase complement of the human genome. Science 298, 1912–1934. [30] Blume-Jensen, P. and Hunter, T. (2001) Oncogenic kinase signalling. Nature 411, 355–365. [31] Alonso, A., Sasin, J., Bottini, N., Friedberg, I., Osterman, A., Godzik, A., Hunter, T., Dixon, J. and Mustelin, T. (2004) Protein tyrosine phosphatases in the human genome. Cell 117, 699–711. [32] Forrest, A.R., Ravasi, T., Taylor, D., Huber, T., Hume, D.A. and Grimmond, S. (2003) Phosphoregulators: protein kinases and protein phosphatases of mouse. Genome Res. 13, 1443–1454. [33] Wang, Z., Shen, D., Parsons, D.W., Bardelli, A., Sager, J., Szabo, S., Ptak, J., Silliman, N., Peters, B.A., van der Heijden, M.S., Parmigiani, G., Yan, H., Wang, T.L., Riggins, G., Powell, S.M., Willson, J.K., Markowitz, S., Kinzler, K.W., Vogelstein, B. and Velculescu, V.E. (2004) Mutational analysis of the tyrosine phosphatome in colorectal cancers. Science 304, 1164–1166. [34] Paez, J.G., Janne, P.A., Lee, J.C., Tracy, S., Greulich, H., Gabriel, S., Herman, P., Kaye, F.J., Lindeman, N., Boggon, T.J., Naoki, K., Sasaki, H., Fujii, Y., Eck, M.J., Sellers, W.R., Johnson, B.E. and Meyerson, M. (2004) EGFR mutations in lung cancer: correlation with clinical response to gefitinib therapy. Science 304, 1497–1500. [35] Lynch, T.J., Bell, D.W., Sordella, R., Gurubhagavatula, S., Okimoto, R.A., Brannigan, B.W., Harris, P.L., Haserlat, S.M., Supko, J.G., Haluska, F.G., Louis, D.N., Christiani, D.C., Settleman, J. and Haber, D.A. (2004) Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N. Engl. J. Med. 350, 2129–2139.
S. Benvenuti et al. / FEBS Letters 579 (2005) 1884–1890 [36] Huang, S.F., Liu, H.P., Li, L.H., Ku, Y.C., Fu, Y.N., Tsai, H.Y., Chen, Y.T., Lin, Y.F., Chang, W.C., Kuo, H.P., Wu, Y.C., Chen, Y.R. and Tsai, S.F. (2004) High frequency of epidermal growth factor receptor mutations with complex patterns in non-small cell lung cancers related to gefitinib responsiveness in Taiwan. Clin. Cancer Res. 10, 8195–8203. [37] Marx, J. (2004) Medicine. Why a new cancer drug works well, in some patients. Science 304, 658–659. [38] Strausberg, R.L., Simpson, A.J., Old, L.J. and Riggins, G.J. (2004) Oncogenomics and the development of new cancer therapies. Nature 429, 469–474. [39] Druker, B.J., Talpaz, M., Resta, D.J., Peng, B., Buchdunger, E., Ford, J.M., Lydon, N.B., Kantarjian, H., Capdeville, R., OhnoJones, S. and Sawyers, C.L. (2001) Efficacy and safety of a specific inhibitor of the BCR-ABL tyrosine kinase in chronic myeloid leukemia. N. Engl. J. Med. 344, 1031–1037. [40] Druker, B.J. (2004) Imatinib as a paradigm of targeted therapies. Adv. Cancer Res. 91, 1–30. [41] van Oosterom, A.T., Judson, I., Verweij, J., Stroobants, S., Donato di Paola, E., Dimitrijevic, S., Martens, M., Webb, A., Sciot, R., Van Glabbeke, M., Silberman, S. and Nielsen, O.S. (2001) Safety and efficacy of imatinib (STI571) in metastatic gastrointestinal stromal tumours: a phase I study. Lancet 358, 1421–1423. [42] Hirota, S., Nishida, T., Isozaki, K., Taniguchi, M., Nakamura, J., Okazaki, T. and Kitamura, Y. (2001) Gain-of-function mutation at the extracellular domain of KIT in gastrointestinal stromal tumours. J. Pathol. 193, 505–510. [43] Rubin, B.P., Singer, S., Tsao, C., Duensing, A., Lux, M.L., Ruiz, R., Hibbard, M.K., Chen, C.J., Xiao, S., Tuveson, D.A., Demetri, G.D., Fletcher, C.D. and Fletcher, J.A. (2001) KIT activation is a ubiquitous feature of gastrointestinal stromal tumors. Cancer Res. 61, 8118–8121. [44] Heinrich, M.C., Corless, C.L., Demetri, G.D., Blanke, C.D., von Mehren, M., Joensuu, H., McGreevey, L.S., Chen, C.J., Van den Abbeele, A.D., Druker, B.J., Kiese, B., Eisenberg, B., Roberts, P.J., Singer, S., Fletcher, C.D., Silberman, S., Dimitrijevic, S. and Fletcher, J.A. (2003) Kinase mutations and imatinib response in patients with metastatic gastrointestinal stromal tumor. J. Clin. Oncol. 21, 4342–4349. [45] Leyland-Jones, B. (2002) Trastuzumab: hopes and realities. Lancet Oncol. 3, 137–144.