Int. J. Cancer: 122, 2255–2259 (2008) ' 2008 Wiley-Liss, Inc.

Mutations in the RAS-MAPK, PI(3)K (phosphatidylinositol-3-OH kinase) signaling network correlate with poor survival in a population-based series of colon cancers
Ludovic Barault1–4, Nicolas Veyrie5–7, Valerie Jooste1,2,4, Delphine Lecorre5–7, Caroline Chapusot1–4, Jean-Marc Ferraz5–7, ` Astrid Lievre5–7, Marion Cortet1,2,4, Anne-Marie Bouvier1,2,4,8, Patrick Rat1,9, Patrick Roignot10, Jean Faivre1,2,4,8, Pierre Laurent-Puig5–7 and Francoise Piard1–4* 1 INSERM, U866, Dijon, F-21079, France 2 Universit de Bourgogne, Dijon, F-21079, France e 3 Service d’Anatomie Pathologique, CHU Dijon, F-21079, France 4 Registre bourguignon des cancers digestifs, Dijon, F-21079, France 5 INSERM, U775, Paris, France 6 Universit Paris Descartes, Paris, France e 7 ˆpital Europe George Pompidou, assistance publique, Ho ˆpitaux de Paris, Paris, France ´en Ho 8 ´rologie, CHU Dijon, F-21079, France Service de Gastroente 9 Service de chirurgie digestive, CHU Dijon, F-21079, France 10 Centre de Pathologie, Dijon, F-21000, France
The RAS-MAPK, PI (3)K signaling pathways form a network that play a central role in tumorigenesis. The BRAF, KRAS and PI3KCA genes code 3 partners of this network and have been found to be activated by mutation in colorectal cancer; these mutations lead to unrestricted cell growth. We evaluated the clinicopathological features and the prognosis of patients with activated-network colon cancers in a population-based study. A total of 586 colon adenocarcinomas were evaluated using sequencing for mutations of KRAS and PI3KCA, and allelic discrimination for mutation of BRAF. Clinicopathological characteristics were correlated to the risk of bearing a mutation of the network using logistic regression. Three-year survival rates were compared with the Log rank test. A multivariate survival analysis using the Cox model was performed. After adjustment for age and microsatellite instability, activation of the network by mutation of at least 1 of the 3 genes was significantly associated with female sex (p 5 0.02) and proximal location (p < 0.001). Lower levels of 3-year survival were associated with activation of the network by mutation of at least 1 of the 3 genes (59.4 and 69.4%, respectively; p 5 0.009). These results remained significant in a multivariate analysis adjusted for sex, age, location, stage and microsatellite instability (HR 5 1.48; CI CI95% 5 [1.07–2.04]). Our study is the first report to underline the potential role of RAS-MAPK, PI (3)K network mutations on survival in colon cancers. Because of the role of this signaling network on anticancer agents, the evaluation of its mutations could have clinical implications. ' 2008 Wiley-Liss, Inc. Key words: colonic neoplasm; K-Ras; B-Raf; PI-3K; RAS-MAPK/ PI(3)K network; survival

clinically approved or under development focus on partners of this network. In colorectal cancer, KRAS, PIK3CA and BRAF genes have been found to be significantly mutated (by up to 50%), which leads to the activation of RAS-MAPK and PI(3)K pathways, and underlies the carcinogenic importance of this network.7 The goal of this article was to evaluate, in a population based study, the associated clinicopathological characteristics on the activation of the RAS-MAPK, PI(3)K network by mutations, and to assess the prognostic influence of these activating mutations. Our work is the first to study the impact of the activation of genes in their network environment rather than as individual members of a pathway. Patients and methods Population ˆ Tissue samples from patients living in the Cote-d’Or area (Burgundy, France) resected for colorectal cancer between January 1998 and December 2002 were collected in a frozen-tissue bank. They were provided by the 3 pathology laboratories covering the area: Dijon University Hospital, a comprehensive cancer center and a private center. Five fragments considered as surgical waste in accordance with French ethical laws (L.1211-3 to L.1211-9), were obtained from the tumor and from the adjacent normal mucosa and were immediately frozen in liquid nitrogen. One fragment was paraffin-embedded. Cases with rectal cancers (located within 15 cm of the anal verge) or with known familial adenomatous polyposis or HNPCC syndrome, ulcerative colitis, or Crohn’s disease were not eligible. All cases were adenocarcinomas. A total
This article contains supplementary material available via the Internet at http://www.interscience.wiley.com/jpages/0020-7136/suppmat. Abbreviations: CIMP, CpG island methylator phenotype; HNPCC, hereditary non polyposis colorectal cancer; MAPK, mitogen activated protein kinase; MSI, microsatellite instability; MSI-H, microsatellite instability high; MSS, microsatellite stable; PI, phosphoinositide. ˆ Grant sponsor: Ligue nationale de lutte contre le cancer (Comit de Cote e d’Or); Grant number: 2006-19; Grant sponsor: Association pour la recherche sur le Cancer; Grant number: 3329; Grant sponsor: PHRC 2003 (Epigenetic and Genetic Alterations in Colorectal Cancer). The last two authors contributed equally to this paper. *Correspondence to: Service d’anatomie pathologique, CHU Dijon, Facult de mdecine, 7 Boulevard Jeanne d’Arc, 21079 Dijon Cedex, e e France. Fax: 133-03-80-39-33-50. E-mail: francoise.piard@u-bourgogne.fr Received 21 August 2007; Accepted after revision 3 December 2007 DOI 10.1002/ijc.23388 Published online 25 January 2008 in Wiley InterScience (www.interscience. wiley.com).

The RAS-MAPK (Mitogen Activated Protein Kinase), PI(3)K (phosphatidylinositol-3-OH kinase) signaling pathways form an intersecting biochemical network that, when mutated, leads to unrestricted cell growth (Fig. 1). The KRAS gene is a member of the oncogenic RAS gene family and binds to at least 3 types of effector proteins: kinases of the RAF family (including BRAF), phosphoinositide (PI) 3-kinase and members of a family of exchange factors for the small GTPase Ral.1 The PIK3CA gene encodes the p110a subunit of PI(3)K, which can be activated by an interaction with RAS proteins.2 The RAS-MAPK and PI(3)K pathways are strongly interconnected and play a central role in tumorigenesis through the coordinated phosphorylation of proteins or transcription factors that directly regulate cell growth, differentiation and apoptosis.3–6 Mutation in 1 of the 3 genes will have the same consequences, which is activation of the network and an increase in the transcription of different oncogenes such as CMYC, CREB, NF-jB and others. Disruption of one will, in many cases, push tumor cells to increase flux through the others in a virtual chess game between proliferation, survival signals and anticancer agents. Furthermore, most of the targeted anticancer drugs
Publication of the International Union Against Cancer

2256

BARAULT ET AL.

Mutation analyses Exon 2 of the KRAS gene and exons 1, 2, 9 and 20 of the PIK3CA gene were selected for mutation analysis because they were frequently found mutated in colorectal cancer.3 These exons were sequenced after PCR amplification. Primers and PCR conditions are available upon request. Direct sequencing was done using a Big Dye Terminator cycle sequencing kit and analyzed on an ABI Prism 3100 DNA Analyzer automated sequencer (Applied Biosystems). All somatic mutations found were further validated by conducting a new independent amplification and sequencing procedure. BRAF V600E mutation detection was assessed by allelic discrimination using Taqman probes. The amplification was performed on an ABI PRISM 7900HT (Applied Biosystems) in an 8 lL mixture composed of 1 lL of extracted DNA (25 ng), 4 lL of Master Mix (Applied Biosystems), 0.2 lL of Assay Mix containing primers and probes (available on request). The reaction was performed with a preliminary step for 2 min at 50°C, followed by denaturation for 10 min at 95°C, then 50 cycles of 2 steps at 92°C for 15 s and 64°C for 1 min. Allelic discrimination was performed using SDS software (Applied Biosystems). This method was validated by comparison to sequencing.10
FIGURE 1 – RAS-MAPK, PI(3)K signaling network: RAS-MAPK and PI(3)K pathways are strongly interconnected and form a signaling network. Activation by mutation of the different partners of the network will deregulate survival, mobility and proliferation of the cells. RTK, receptor tyrosine kinase; ER, estrogen receptor; PTEN, phosphatase with tensin homology; PI3K, phophatidylinositol 3-OH kinase; RAF, Ras activating factor; ERK, extracellular-signal regulated kinase; MEK, mitogen-activated ERK protein kinase; PAK, p21 activated kinase.

of 586 cases (336 men and 250 women) were included in the analysis. Data was merged with that of the cancer Registry, which routinely collects data on patients’ characteristics, cancer site, stage at diagnosis, treatment, recurrence and survival. An active search for vital status was carried out using a standardized administrative procedure. The information was collected ‘‘at first line’’ via the public services of the birthplace or via an electronic request to the ‘‘Rpertoire National d’Identification des Personnes Physiques’’ e (RNIPP). Both procedures required the knowledge of the birthplace. As this could not be obtained for all cases, other sources of information for vital status were used (medical records or the public services of the places of residence). Finally, the life status was known for 564 patients (97%) in December 2005. Tumors occurring between the caecum and the splenic flexure were defined as proximal (n 5 250). Other sites were classified as distal (n 5 335). Cancer extension at the time of diagnosis was classified according to the 5th UICC version of the TNM classification. Among the cases included, 51 were classified as stage I, 255 as stage II, 157 as stage III and 123 as stage IV. Microsatellite instability status Microsatellite instability was assessed according to the results of a previous paper8 for patients operated on between 1998 and 2000 and according to the panel of the consensus conference for patients operated on between 2001 and 2002.9 PCR was carried out in a 12.5 mL reaction volume with a Multiplex PCR kit (Qiagen, Courtaboeuf, France). One microliter (50 ng/lL) of genomic DNA was used, as well as 8 pmol of each primer with one primer of each pair that was fluorescently labeled for allele detection (Applied Biosystems, Courtaboeuf, France). Multiplex PCR products were mixed with 0.03 lL Genscan 500HD LIZ size standard (Applied Biosystems) and 9 lL formamide. The PCR fragments were separated using an ABI Prism 3100 instrument and further analyzed using GeneMapper v3.7 (Applied Biosystems). Tumors were classified as MSI-H when they showed instability in at least 40% of the microsatellites analyzed.

Statistical analysis In this study, the chi-square test was used to test homogeneity between different groups of patients. Multivariate logistic regression models were fitted to estimate the independent effect of sex, location and age range, MSI status, and on the risk of bearing a mutation of the network. Interactions were tested when they were deemed to be of interest. Survival functions were estimated using the Kaplan-Meier method. The 37 patients who died in the postoperative period (1 month after surgical resection) were not included in these analyses, because their deaths were not considered to be associated with the disease. Survival curves were compared with the Log-Rank Test. Then a multivariate Cox model was adjusted to estimate the effect of a mutation of the network on survival after adjustment for UICC stage, location, age, gender and MSI status. An association was considered significant for p < 0.05 and 95% confidence intervals were calculated. Results Frequency of KRAS, PIK3CA and BRAF mutations and the relationship with clinicopathological parameters Three hundred and sixteen cases (56.4%) showed activation of the network by mutation in at least 1 of the 3 genes. In the total population, 198 (34.4%), 98 (17.8%) and 78 (13.3%) were mutated for KRAS, PIK3CA and BRAF, respectively (supplementary table). KRAS was mutated at codon 12 in 163 cases (82.3%), and at codon 13 in 35 cases (17.7%). PIK3CA alterations were point mutations in 95 cases (located at exon 1 (20 cases), exon 2 (6 cases), exon 9 (46 cases), and exon 20 (29 cases)) or in-frame deletions in exon 20 (3 cases). A double point mutation of PIK3CA was observed in 6 cases (6.4%). Both KRAS and PIK3CA mutations were observed in 45 cases (8.2%), which is higher than expected by chance (p 5 0.010). Twelve (2.2%) were mutated for both BRAF and PIK3CA, and 1 tumor (0.2%) was mutated for both KRAS and BRAF (p < 0.001). Relationships between clinicopathogical variables and mutations of KRAS, PIK3CA and BRAF genes are shown in Table I. In univariate analyses, a significant association was observed between the activation of the network and the presence of an MSIH phenotype: 79.8% of the MSI-H tumors were mutated for 1 of the 3 genes vs. 52.6% of the MSS tumors (p 5 6.5 3 1026). Significant associations were also observed between activation of the network and female sex (p 5 3.6 3 1025), advanced age (p 5 0.004), proximal location of the cancers (p 5 7.5 3 10211). Results of multivariate analyses are given in Table II.

RAS-MAPK, PI(3)K NETWORK AND SURVIVAL

2257

TABLE I – UNIVARIATE ANALYSIS OF THE ASSOCIATION BETWEEN CLINICOPATHOLOGICAL VARIABLES AND MUTATIONAL STATUS OF THE GENES Activation of the network WT (%) M (%) p1 WT (%) KRAS M (%) p1 WT (%) PIK3CA M (%) p1 WT (%) BRAF M (%) p1

Sex Men Women Age 65 65–74 75 Location2 Proximal Distal Stage I II III IV MSI MSS MSI-H

163 (51.1) 156 (48.9) <0.001 212 (64.4) 117 (35.6) 0.489 275 (87.6) 39 (12.4) <0.001 313 (93.2) 23 (6.9) <0.001 81 (33.6) 160 (66.4) 166 (67.2) 81 (32.8) 179 (75.2) 59 (24.8) 194 (78.0) 55 (22.1) 77 (53.1) 68 (46.9) 0.004 95 (63.8) 72 (46.5) 83 (53.6) 105 (65.6) 95 (36.5) 165 (63.5) 178 (66.7) 54 (36.2) 0.836 120 (84.5) 22 (15.5) 55 (34.4) 120 (79.0) 32 (21.1) 89 (33.3) 214 (83.0) 44 (17.1) 0.424 145 (96.0) 6 (4.0) <0.001 148 (90.8) 15 (9.2) 214 (79.0) 57 (21.0) 0.004 182 (73.1) 67 (26.9) <0.001 324 (96.7) 11 (3.3) 0.223 45 (90.0) 5 (10.0) 223 (87.5) 32 (12.6) 131 (83.4) 26 (16.6) 108 (87.8) 15 (12.2) 0.538

64 (26.8) 175 (73.2) <0.001 155 (63.5) 89 (36.5) 0.377 183 (76.9) 55 (23.1) 179 (55.9) 141 (44.0) 222 (67.1) 109 (32.9) 270 (86.3) 43 (13.7) 24 (51.1) 23 (48.9) 0.705 32 (64.0) 105 (42.9) 140 (57.1) 168 (66.7) 62 (41.6) 87 (58.4) 98 (64.1) 53 (44.5) 66 (55.5) 80 (66.1) 18 (36.0) 0.948 40 (85.1) 7 (14.9) 84 (33.3) 189 (78.8) 51 (21.3) 55 (36.0) 131 (86.8) 20 (13.3) 41 (33.9) 94 (82.5) 20 (17.5)

228 (47.4) 253 (52.6) <0.001 306 (61.7) 190 (38.3) <0.001 394 (83.1) 80 (16.9) 16 (20.3) 63 (79.8) 72 (90.0) 8 (10.0) 60 (76.9) 18 (23.1)

0.184 477 (94.6) 27 (5.4) <0.001 30 (37.0) 51 (63.0)

WT, wild type; M, mutated. 1 Chi-square test.–2Location was unknown for 1 patient.

TABLE II – INDEPENDENT FACTORS AND RISK OF MUTATION OF THE SIGNALLING NETWORK. MULTIVARIATE LOGISTIC REGRESSIONS Activation of the network OR P 95% CI OR KRAS p 95% CI OR p PIK3CA 95% CI OR p BRAF 95% CI

Sex Sex Men 1 Men Women 1.55 0.020 [1.07–2.24] Women Age Age 65 1 65 65–74 1.23 0.393 [0.77–1.97] 65–74 75 1.44 0.099 [0.93–2.23] 75 Location1 Location1,2 Prox. 1 MSS Prox. Dist. Dist. 0.37 <0.001 [0.25–0.55] MSI-H Prox. Dist. MSI MSI MSS 1 MSS MSI-H 1.68 0.109 [0.89–3.18] MSI-H

1 1.00 0.994 1 0.96 0.881 1.04 0.851

[0.69–1.45] [0.60–1.56] [0.67–1.62]

1 1 2.19 0.001 [1.37–3.51] 1.85 1 1 1.40 0.284 [0.76–2.57] 1.72 0.86 0.607 [0.47–1.55] 2.71 1 1

0.060 [0.98–3.49] 0.328 [0.58–5.14] 0.045 [1.02–7.20]

1 0.52 0.001 [0.36–0.77] 1 4.55 0.109 [0.71–29.15] 1 0.09 <0.001 [0.04–0.23]

0.58 0.031 [0.36–0.95] 0.31

0.002 [0.14–0.66]

1 1 0.95 0.871 [0.49–1.82] 13.95 <0.001 [7.23–26.91]

1 Prox., proximal colon; Dist., Distal colon.–2Because of an interaction between MSI status and location on the occurrence of the KRAS mutation, the results are shown separately in MSS group and MSI-H group.

As for the occurrence of a mutation in 1 of the 3 genes, mutants were independently associated with female sex and proximal location, suggesting the existence of a subgroup of colon cancer patients irrespective of the microsatellite status. Among women older than 75 years with proximal colon cancer whatever the MSI status, we found that more than 77.5% showed activating mutations of the network, whereas only 34.3% of males younger than 65 years with distal colon cancer showed such mutations. Survival in the presence of the RAS-MAPK, PI(3)K signaling network In univariate analyses (Table III), no significant association was detected between survival and location, sex or MSI status. A significant overall lower level of survival at 3 years was observed for patients with a network activated by mutations in at least 1 of the 3 genes (59.4 and 69.4%, respectively; p 5 0.009) (Fig. 2). Age and TNM stage were also significantly associated with survival. Multivariate analysis adjusted for gender, age, location, stage and MSI status was performed (Table III). Patients with activation of the signaling network showed a higher risk of death (HR 5

1.48; CI CI95% 5 [1.07–2.04]) independently of age and stage at diagnosis, which were also associated to prognosis.

Discussion Numerous mitogens interact with receptor tyrosine kinases to activate a number of downstream effectors through a cascade of phosphorylation. Therefore, RAS-MAPK, PI(3)K signaling pathways form a network that governs proliferation, differentiation and cell survival5 (Fig. 1). In this population-based study of colon cancers, the most frequent alteration concerned KRAS (34.4%) followed by PIK3CA (17.8%) and BRAF (13.3%). Taken individually, the results concerning the rate and distribution of KRAS, PIK3CA, and BRAF were in close accordance with the results of large population-based series.4,5,11–21 The rate of KRAS mutations was evaluated by sequencing exon 2, which contains the most frequently mutated codons (codon 12 and 13) in colon cancer. This rate was in close accordance with the results of large population-based series.11–13 Furthermore, the

2258

BARAULT ET AL.
TABLE III – SURVIVAL AND ACTIVATION OF THE NETWORK. UNIVARIATE KAPLAN-MEIR AND MULTIVARIATE COX SURVIVAL ANALYSES Variables N 1 year (%) 3 years (%) p1 HR (Cox) 95% CI p

Sex Men Women Age 65 65–74 75 Location Proximal colon Distal colon Stage I II III IV MSI MSS MSI-H Activation of the network2 Wild type Mutated

320 234 144 157 253 240 313 049 244 149 112 477 077 231 299

88.1 87.1 89.3 90.9 84.7 83.6 90.6 95.8 94.9 86.7 70.3 88.3 83.8 91.9 84.0

64.3 64.7 70.7 67.9 58.6 62.8 65.8 85.4 80.6 64.3 21.6 63.3 71.6 69.4 59.4

0.842 0.028

1 0.93 1 1.36 2.02 1 0.89 1 1.81 3.55 14.72 1 0.97 1 1.48

[0.67–1.28] [0.89–2.09] [1.37–2.98] [0.65–1.22] [0.72–4.56] [1.41–8.98] [5.94–36.49] [0.59–1.61] [1.07–2.04]

0.660 0.155 <0.001 0.463 0.207 0.007 <0.001 0.920 0.017

0.278 <0.001

0.283 0.009

Thirty-seven postoperative deaths in the month following the surgical resection were excluded. 1 Logrank test.–2Unknown in 24 cases.

FIGURE 2 – Kaplan-Meier survival estimates of individuals with colon cancers by activation of the RAS-MAPK, PI(3)K network by mutation (- - - 5 Wild type; — 5 Mutated). Cases mutated for at least 1 of the 3 genes showed significantly poorer survival at 3 years vs. wild type cases (p 5 0.009).

distribution of the different KRAS mutations (codon 12 vs. codon 13, or transition vs. transversion) was similar to that described by Brink et al.12 with a predominance of transitions G:A. PIK3CA was screened for by sequencing exons 1, 2, 9 and 20, which count nearly all of the mutations identified up to now.3 The frequency of PIK3CA observed in our study (17.8%) is lower than that observed in the first series,3 but closer to those published more recently.5,14–16 Furthermore, in 6 (5.6%) PIK3CA mutated cases, we found two mutations (probably a biallelic alteration) suggesting selective advantage to a high activation of the PI(3)K pathway.22,23 The BRAF mutation was checked by allelic discrimination. With regard to sequencing, allelic discrimination is more rapid, needs only 1 step of PCR, without purification, and a sequencing step. However, this method is only possible for clearly localized point mutations. Thus it perfectly suits the single polymorphism of the BRAF gene since there is a hotspot for the BRAF mutation which accounts for 90% of the BRAF mutations in colorectal cancer and leads to conversion from valine to glutamic acid at codon 600. Our data confirmed that colon carcinomas accumulate

BRAF mutations at a rate similar to that reported in previous studies.4,17–21 Activation of the RAS-MAPK, PI(3)K network by point mutations of KRAS, PIK3CA or BRAF was identified in 56.4% of the cases. Interestingly, we demonstrated a significant concomitant occurrence of KRAS and PIK3CA mutations in 45 cases (8.2%).23 In the same way, mutations in PIK3CA and BRAF were observed in 12 cases (2.2%). These rates confirmed the data reported by Velho et al.16 who showed that 9.7% of PIK3CA mutated tumors displayed KRAS or BRAF mutations. In our study, we observed a significant association between KRAS mutations and proximal tumor location in multivariate analysis, which did not appear in univariate analysis. This uncommon result was due to the impact of an interaction between MSI status and location on the occurrence of the KRAS mutation. When adjusted for MSI status, we clearly demonstrated that proximal location was significantly associated with the KRAS mutation in MSS patients, but not in MSI-H, where KRAS mutations tended to be more frequent in distal tumors as shown by other studies.11,12,24–26 We reported a significant increased risk of PIK3CA or BRAF mutations in patients with proximal colon cancer. Thus, activated networks showed a strong association with proximally located tumors. Furthermore in multivariate analysis, activation of the network by mutation of at least 1 of the 3 genes was significantly associated with female sex, old age and proximal location, independently of the MSI status. This work demonstrated that the activation of the RAS-MAPK, PI(3)K network is not dependent on the MSI phenotype even though it occurs more frequently in a subset of colon tumors that occur in elderly women with proximal colon cancer. This suggests the implication of a wider process of epigenetic alterations, such as that described between the CpG Island Methylator Phenotype (CIMP) and the BRAF mutation. Moreover, univariate analysis showed that activation of the RAS-MAPK, PI(3)K signaling network was associated with a significantly lower 3-year survival (56.5% vs. 69.4%). This result remains significant after adjustment for age, sex, location, stage and MSI status. Taking into account the importance of studying genes in their network environment it could be of great interest to analyze other network partner genes (i.e., PAK4, ERK, MEK, PTEN). Moreover, the link between the RAS-MAPK, PI(3)K network and the ER pathway, should be investigated since this interaction has already

RAS-MAPK, PI(3)K NETWORK AND SURVIVAL

2259 Acknowledgements

been seen in other cancers and might explain the association between activation of the network and female sex. Our study is the first to underline the potential role of RASMAPK, PI(3)K network activation on survival in colon cancers. Furthermore, the presence of alterations that are specific to some tumors, such as those involving RAS-MAPK, PI(3)K signaling network mutations may confer different sensitivity to anticancer agents.6 Mutated colon carcinomas could be good models to test targeted anticancer drugs that act on the different partners of this network.

27,28

The authors thank the Biological Resource Center of Dijon (Ferdinand Cabanne) and the participating pathologists of Comprehensive Cancer Center (Dr. L. Arnould and Dr. F. Collin), of Dijon Pathological Private Center (Dr. F. Drouot, Dr. L. Dusserre, Dr. P. Dusserre, Dr. D. Michiels-Marzais, Dr. F. Morlevat, Dr. T. Petrella and Dr. T. Ponnelle). We also thank Mrs. L. Barbier, Mr. P. Bastable, Mrs. M. Goiset, Mrs. E. Lanier, Mrs. S. Normand and Mr. D. Rageot for their contribution to this work.

References
1. 2. 3. Mor A, Philips MR. Compartmentalized Ras/MAPK signaling. Annu Rev Immunol 2006;24:771–800. Rodriguez-Viciana P, Warne PH, Dhand R, Vanhaesebroeck B, Gout I, Fry MJ, Waterfield MD, Downward J. Phosphatidylinositol-3-OH kinase as a direct target of Ras. Nature 1994;370:527–32. Samuels Y, Diaz LA, Jr, Schmidt-Kittler O, Cummins JM, Delong L, Cheong I, Rago C, Huso DL, Lengauer C, Kinzler KW, Vogelstein B, Velculescu VE. Mutant PIK3CA promotes cell growth and invasion of human cancer cells. Cancer Cell 2005;7:561–73. Davies H, Bignell GR, Cox C, Stephens P, Edkins S, Clegg S, Teague J, Woffendin H, Garnett MJ, Bottomley W, Davis N, Dicks E, et al. Mutations of the BRAF gene in human cancer. Nature 2002;417:949–54. Shaw RJ, Cantley LC. Ras, PI(3)K and mTOR signalling controls tumour cell growth. Nature 2006;441:424–30. McCubrey JA, Steelman LS, Abrams SL, Lee JT, Chang F, Bertrand FE, Navolanic PM, Terrian DM, Franklin RA, D’Assoro AB, Salisbury JL, Mazzarino MC, et al. Roles of the RAF/MEK/ERK and PI3K/PTEN/AKT pathways in malignant transformation and drug resistance. Adv Enzyme Regul 2006;46:249–79. Oikonomou E, Pintzas A. Cancer genetics of sporadic colorectal cancer: BRAF and PI3KCA mutations, their impact on signaling and novel targeted therapies. Anticancer Res 2006;26:1077–84. Chapusot C, Martin L, Puig PL, Ponnelle T, Cheynel N, Bouvier AM, Rageot D, Roignot P, Rat P, Faivre J, Piard F. What is the best way to assess microsatellite instability status in colorectal cancer? Study on a population base of 462 colorectal cancers. Am J Surg Pathol 2004;28: 1553–9. Boland CR, Thibodeau SN, Hamilton SR, Sidransky D, Eshleman JR, Burt RW, Meltzer SJ, Rodriguez-Bigas MA, Fodde R, Ranzani GN, Srivastava S. A National Cancer Institute Workshop on Microsatellite Instability for cancer detection and familial predisposition: development of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer Res 1998;58:5248–57. Benlloch S, Paya A, Alenda C, Bessa X, Andreu M, Jover R, Castells A, Llor X, Aranda FI, Massuti B. Detection of BRAF V600E mutation in colorectal cancer: comparison of automatic sequencing and realtime chemistry methodology. J Mol Diagn 2006;8:540–3. Samowitz WS, Curtin K, Schaffer D, Robertson M, Leppert M, Slattery ML. Relationship of Ki-ras mutations in colon cancers to tumor location, stage, and survival: a population-based study. Cancer Epidemiol Biomarkers Prev 2000;9:1193–7. Brink M, de Goeij AF, Weijenberg MP, Roemen GM, Lentjes MH, Pachen MM, Smits KM, de Bruine AP, Goldbohm RA, van den Brandt PA. K-ras oncogene mutations in sporadic colorectal cancer in The Netherlands Cohort Study. Carcinogenesis 2003;24:703–10. Chang SC, Lin JK, Yang SH, Wang HS, Li AF, Chi CW. Relationship between genetic alterations and prognosis in sporadic colorectal cancer. Int J Cancer 2006;118:1721–7. Frattini M, Signoroni S, Pilotti S, Bertario L, Benvenuti S, Zanon C, Bardelli A, Pierotti MA. Phosphatase protein homologue to tensin expression and phosphatidylinositol-3 phosphate kinase mutations in colorectal cancer. Cancer Res 2005;65:11227. Moroni M, Veronese S, Benvenuti S, Marrapese G, Sartore-Bianchi A, Di Nicolantonio F, Gambacorta M, Siena S, Bardelli A. Gene copy number for epidermal growth factor receptor (EGFR) and clinical response to antiEGFR treatment in colorectal cancer: a cohort study. Lancet Oncol 2005;6:279–86. Velho S, Oliveira C, Ferreira A, Ferreira AC, Suriano G, Schwartz S, Jr, Duval A, Carneiro F, Machado JC, Hamelin R, Seruca R. The prevalence of PIK3CA mutations in gastric and colon cancer. Eur J Cancer 2005;41:1649–54. Domingo E, Laiho P, Ollikainen M, Pinto M, Wang L, French AJ, Westra J, Frebourg T, Espin E, Armengol M, Hamelin R, Yamamoto H, et al. BRAF screening as a low-cost effective strategy for simplifying HNPCC genetic testing. J Med Genet 2004;41:664–8. Fransen K, Klintenas M, Osterstrom A, Dimberg J, Monstein HJ, Soderkvist P. Mutation analysis of the BRAF, ARAF and RAF-1 genes in human colorectal adenocarcinomas. Carcinogenesis 2004;25: 527–33. Lubomierski N, Plotz G, Wormek M, Engels K, Kriener S, Trojan J, Jungling B, Zeuzem S, Raedle J. BRAF mutations in colorectal carcinoma suggest two entities of microsatellite-unstable tumors. Cancer 2005;104:952–61. Rajagopalan H, Bardelli A, Lengauer C, Kinzler KW, Vogelstein B, Velculescu VE. Tumorigenesis: RAF/RAS oncogenes and mismatchrepair status. Nature 2002;418:934. Yuen ST, Davies H, Chan TL, Ho JW, Bignell GR, Cox C, Stephens P, Edkins S, Tsui WW, Chan AS, Futreal PA, Stratton MR, et al. Similarity of the phenotypic patterns associated with BRAF and KRAS mutations in colorectal neoplasia. Cancer Res 2002;62: 6451–5. Parsons DW, Wang TL, Samuels Y, Bardelli A, Cummins JM, DeLong L, Silliman N, Ptak J, Szabo S, Willson JK, Markowitz S, Kinzler KW, et al. Colorectal cancer: mutations in a signalling pathway. Nature 2005;436:792. Samuels Y, Wang Z, Bardelli A, Silliman N, Ptak J, Szabo S, Yan H, Gazdar A, Powell SM, Riggins GJ, Willson JK, Markowitz S, et al. High frequency of mutations of the PIK3CA gene in human cancers. Science 2004;304:554. Samowitz WS, Holden JA, Curtin K, Edwards SL, Walker AR, Lin HA, Robertson MA, Nichols MF, Gruenthal KM, Lynch BJ, Leppert MF, Slattery ML. Inverse relationship between microsatellite instability and K-ras and p53 gene alterations in colon cancer. Am J Pathol 2001;158:1517–24. Bleeker WA, Hayes VM, Karrenbeld A, Hofstra RM, Hermans J, Buys CC, Plukker JT. Impact of KRAS and TP53 mutations on survival in patients with left- and right-sided Dukes’ C colon cancer. Am J Gastroenterol 2000;95:2953–7. Elnatan J, Goh HS, Smith DR. C-KI-RAS activation and the biological behaviour of proximal and distal colonic adenocarcinomas. Eur J Cancer A 1996;32:491–7. Schiff R, Massarweh SA, Shou J, Bharwani L, Mohsin SK, Osborne CK. Cross-talk between estrogen receptor and growth factor pathways as a molecular target for overcoming endocrine resistance. Clin Cancer Res 2004;10:331S–336S. Yager JD, Davidson NE. Estrogen carcinogenesis in breast cancer. N Engl J Med 2006;354:270–82.

16.

17.

4. 5. 6.

18.

19.

7. 8.

20. 21.

9.

22.

23.

10.

24.

11.

12.

25.

13. 14.

26. 27.

15.

28.