Activation of the phosphatidylinositol 3′-kinase (PI3K)/AKT pathway results in an increase in cell proliferation and survival. Somatic mutations within the PI3K catalytic subunit, PIK3CA are common cause of increasing PI3K activity and are believed to be oncogenic in many cancer types. Few reports addressed the association between PIK3CA mutations and tumor progression specifically in microsatellite instable (MSI) colorectal cancer (CRC). In the present study, we have evaluated PIK3CA mutational status in a series of 410 Middle Eastern CRC and 13 colon cell lines to study the prevalence of PIK3CA mutations in MSI cases, PTEN expression in CRC and possibility of therapeutic targeting of this set of patients. PIK3CA mutations were found in four of the cell lines tested and 51 colorectal carcinomas (12.2%). Three of these four mutated cell lines were MSI. PTEN was inactivated in 66.1% of the CRC. Furthermore, we observed a strong association between PIK3CA mutations and MSI status (P=0.0046) while PTEN loss was more frequent in microsatellite stable (MSS) CRC (P=0.043). A high prevalence of genetic alterations in PI3K/AKT pathway in Saudi cohort of CRC, predominance of PIK3CA mutations in the MSI subgroup and their possible involvement in development/progression of this subset of CRC are some of the significant findings of our study.
The phosphatidylinositol 3-kinase (PI3K)/AKT signaling pathway is activated in multiple cancers, leading to oncogenic transformation (Vivanco and Sawyers, 2002; Bader et al., 2005; Wymann and Marone, 2005). Activation may result from activating mutations in PI3-kinase genes or inactivating mutations in the tumor suppressor gene PTEN (phosphatase and tensin homolog). PI3-kinases constitute a large family of lipid kinases which phosphorylate the inositol ring 3′OH group in inositol phospholipids PIP3 to generate the tumor promoting second messenger phosphatidylinositol-3,4,5-trisphosphate (PIP3). PIP3 mainly causes AKT to translocate to the plasma membrane, where it is activated by phosphorylation allowing it to mediate many of the biological consequences of PI3K activation. For proliferation and tumorigenesis, the most important PI3K proteins are those belonging to class IA, which are heterodimeric proteins consisting of a catalytic domain (p110a), encoded by the PIK3CA gene on 3q26.3 and an associated regulatory subunit, p85 (Garcia et al., 2006).
Recently, Samuels et al. (Samuels et al., 2004) have identified somatic mutations in the PIK3CA gene in various human cancers. These cancers include cancer of the colon, stomach, breast, brain, lung, ovary and thyroid gland (Broderick et al., 2004; Campbell et al., 2004; Lee et al., 2005; Levine et al., 2005; Li et al., 2005; Velho et al., 2005). PIK3CA mutations have been reported to occur in sporadic colorectal cancer (CRC) at a frequency of 31.6% (Samuels et al., 2004), 18.8% (Campbell et al., 2004) or 13.6% (Velho et al., 2005), generally at a later stage of tumorigenesis. PIK3CA mutations have been shown to stimulate AKT signaling, and promote the cell growth and invasion of human colon cancer cell lines (Samuels et al., 2005). However, PIK3CA mutations frequencies have not been reported in the Middle Eastern CRCs. Furthermore, approximately 15% of sporadic colorectal carcinomas show mismatch repair (MMR) deficiency leading to thousands of mutations in microsatellite sequences throughout the genome microsatellite instability (MSI) phenotype. This sporadic MMR deficiency is mainly caused by silencing of one of the MMR genes hMLH1 through promoter methylation (Philp et al., 2000; Toyota et al., 2000). PIK3CA mutations were detected within the MSI and microsatellite stable (MSS) CRC phenotypes suggesting that PIK3CA mutations play a role in the development/progression of both subsets of CRC (Samuels et al., 2004; Velho et al., 2005). However, there is a controversy about the prevalence of PIK3CA mutations within the MSI CRC (Samuels et al., 2004; Velho et al., 2005).
CRC constitutes the second most common cause of cancer deaths in many Western countries. However colon cancer death rate is ten fold lower in Mexico, South America and Africa as compared to the United States and Europe (Bray et al., 2002). In Saudi Arabia, the colon cancer incidence is about half as high as that in the United States. The peak incidence for CRC in Saudi Arabia is at lower age group and the tumors tend to be larger and more aggressive than similar cancer in the West (El-Hazmi et al., 1995; Isbister et al., 2000). There is growing evidence that potential genetic differences between cancers from different groups might contribute to the epidemiologic differences.
This prompted us to screen 418 CRC samples for activating mutations in PIK3CA, inactivating alteration in PTEN and their correlation to MSI status in Middle Eastern CRC.
Mutational analysis of PIK3CA
Mutational analysis of PIK3CA was done in 418 primary CRC and 13 colorectal cell lines (SW-948, SW480, LOVO, HCT-116-HCT15, DLD-1, CX-1, COLO-320, COLO-206F, CL11, CL34, CACO-2 and HT-29). A total of 51 missense mutations were identified (Figure 1). Because all the mutations were not detected in corresponding normal tissues, these mutations were confirmed as somatic mutations. Of these 51 mutations, 38 mutations clustered in exon-9 and the remaining 13 exon 20 (Table 1). Of the 13 colorectal cell lines tested, 4 cell lines harbored PIK3CA mutations (DLD-1, HCT-116, HCT15, and SW948).
PIK3CA mutations, MSI status and clinicopathologic characteristics
The incidence of MSI cases was 20.2% (82 of 406 cases). Clinical characteristics were compared according to MSI status because of the known correlation of MSI phenotype with certain clinicopathological features (Table 2). MSI cases were significantly associated with right colon (P=0.0020), mucinous histology (P<0.0001) and with poor differentiation (P=0.0007). The frequency of PIK3CA mutations was significantly higher (P<0.0001) in MSI cases (35.9%; 28/78 cases) as compared to MSS cases (7.1%; 22/310). No correlation between PIK3CA mutations was observed with tumor grade, histology or overall survival.
p53 Mutations and clinicopathologic characteristics
p53 mutations were detected in 33.7% (130 of the 386) colorectal carcinomas (Table 3). p53 mutations showed a trend towards older age (P=0.0728), histology subtype of adenocarcinomas (P=0.0809), larger tumor size (P=0.0590) and were significantly associated with lymph node metastasis (P=0.0464). An inverse correlation was seen between MSI status and TP53 mutations which was statistically significant (P=0.0043). Only 11 of 374 (2.94%) samples analysed for both PI3KCA and p53 mutation showed double mutant. Thus a trend was seen towards mutual exclusivity (P=0.0687).
Relationship of PIK3CA mutation with PTEN alterations
PTEN was screened for decreased protein expression (inactivation) by immunohistochemistry (IHC) analysis. Unlike PIK3CA, PTEN protein expression was more often decreased or lost (Figure 2) in MSS tumors (69%; 214/310) than in MSI tumors (56.3%; 40/71) and this difference was significant (P=0.0439). Although mutual exclusivity between PIK3CA mutations and PTEN alteration has been previously reported (Broderick et al., 2004; Saal et al., 2005; Ollikainen et al., 2007), in our study 6.6% CRC (26/390) showed coexistence of PIK3CA mutation and PTEN alteration. PIK3CA mutations alone or as a CRC subset showing combination of both PIK3CA mutations and altered PTEN did not correlate with tumor grade, stage or patient outcome.
Relationship of PIK3CA mutation with p-AKT
Because it is suggested that PIK3CA mutations activate AKT function through its phosphorylation, we investigated the relationship between PIK3CA and expression of p-AKT. The frequency of PIK3CA mutation was not significantly (P=0.2247) different in p-AKT positive tumors (73%; Figure 2) than in p-AKT negative tumors (27%).
In the present study, we have identified PIK3CA mutations in 12% of Saudi CRC. The frequency of PIK3CA mutation in Saudi CRC patients is quite similar to the Caucasian North Europeans population (Velho et al., 2005) and much lower than the North American population (Samuels et al., 2004). There are several factors that can lead to such conflicting results of PIK3CA mutations between studies, such as geographical variation/influence, sample source preservation, sample size and methods used for DNA isolation. The mutations in exon 9 clustered in the following hot spots in the helical domain: E545K (33.3%) and E542K (23.5%), while the mutations in the kinase domain (exon 20) were mainly in the hot spot H1047R (11.8%; Table 1). These results indicate that the contribution of PIK3CA mutations to pathogenesis and progression of CRC might be similar between two ethnicities.
Though previous studies have investigated the role of PI3K mutation and MSI in CRC, the data are in part conflicting. While Samuels et al. (2004) found a significantly higher frequency of PIK3CA mutation in MMR deficient tumors compared to MMR proficient tumors (16/33 vs 58/201, P=0.014), others failed to highlight this correlation (Velho et al., 2005; Ollikainen et al., 2007). Our study showed significantly higher PIK3CA mutations within the MSI subset of CRC (35.9%). In support of this finding, PIK3CA mutation analysis was done in 13 colorectal cell lines and 4 of the tested cell lines (DLD-1, HCT-116, HCT-15 cell lines) carried PIK3CA missense mutations. Mismatch repair deficiency was noted in three of the four cell lines harboring PIK3CA mutation. Interestingly, 7% of MSS/MSI-L CRC bore PIK3CA mutations, suggesting that PIK3CA mutations play a role in both subsets of CRC but might represent a more important oncogenic event for the development/progression of MSI CRC than MSS CRC. Also the observed higher incidence of PIK3CA missense mutation in MSI CRC might suggest yet another mechanism for the activation of PI3K/AKT signaling pathway through mismatch repair deficiency.
PTEN was inactivated (PTEN protein expression was decreased or lost) in 271 of 410 (66.1%) CRCs. In agreement with previous report (Nassif et al., 2004), incidence of reduced or absent PTEN expression was 70% and was seen more frequently in MSS CRC. This suggests that PTEN alterations are more commonly involved in a distinct pathway of CRC tumorigenesis that is separate from the pathway of mismatch repair deficiency. Because both PIK3CA mutations and loss of PTEN function are thought to activate PI3K/AKT pathway, it is speculated that PIK3CA mutations and loss of PTEN expression are mutually exclusive (Broderick et al., 2004; Frattini et al., 2005; Saal et al., 2005). However, we have found no association between PTEN alteration and PIK3CA mutations. Coexistence of PIK3CA mutations and PTEN alterations was seen in 6.6% of CRC, suggesting that PI3K/AKT pathway can be activated by alterations in multiple genes. Furthermore, it suggests that PTEN and PIK3CA mutations may have additional nonoverlapping consequences during CRC tumorigenesis.
In our study, we analysed PIK3CA mutations only in exons-9 and -20 and found majority of the PIK3CA mutations were localized in the main reported hot spots of PIK3CA gene, E545K, E542K and H1047R. The functional consequences of these hot spots were tested by several groups through the overexpression of these hot spots or gene deletion experiments using somatic cell knockouts and have demonstrated that these mutations are in fact oncogenic. So as a read out of PIK3CA functional activation, we screened our CRC for AKT phosphorylation status by IHC. Surprisingly, in our study, there was also a lack of statistical significance when correlating p-AKT by IHC staining in tumors with PI3K alterations (PTEN and PIK3CA). Several explanations could be offered. First, it is possible that genes in the PI3K pathway, other than PTEN and PIK3CA, could also be deregulated in CRC. Second, molecular alterations in other signaling pathways (such as MAP kinase) may cross talk and interfere with PI3K signaling (Cully et al., 2006).
Then we were interested in studying the relationship between PI3K/AKT pathway and p53 (tumor suppressor gene) since interactions between the p53 and PI3K/AKT pathways play a significant role in the determination of cell death/survival. Interrelation between these pathways occurs through the transcriptional regulation of PTEN by p53, which is required for p53-mediated apoptosis (Singh et al., 2002; Maruyama et al., 2007). However, there are conflicting data about the correlation of PIK3CA mutation and p53 mutations in cancer. Singh et al. (2002) reported that PIK3CA and p53 mutations were mutually exclusive, while Maruyama et al. (2007) could not find that associations. In our study, a trend was seen towards mutual exclusivity between PIK3CA mutations and p53 mutations (P=0.0687). Only 11 of 374 cases analysed for both PIK3CA mutations and p53 mutation showed double mutant. This mutual exclusivity suggests that both pathways may be functionally redundant and complement each other in colorectal carcinomas. It also indicates that mutations in either pathway can render the cell refractive to apoptosis and prone to transformation.
In summary, we found a high prevalence of genetic alterations in PI3K/AKT pathway in a Saudi cohort of CRC. The data provides further genetic evidence supporting the notion that dysregulated PI3K/AKT pathway plays an important role in the pathogenesis of CRC regardless of ethnic background. We also showed that PIK3CA mutations are more common in the MSI CRC subgroup and their possible involvement in development/progression of this subgroup of CRC.
Materials and methods
Patient selection and tissue microarray construction
A total of 448 patients with CRC diagnosed between 1990 and 2006 were selected from King Faisal Specialist Hospital and Research Centre. All samples were analysed in a tissue microarray (TMA) format. TMA construction was performed as described earlier (Kononen et al., 1998; Bavi et al., 2006). Briefly, tissue cylinders with a diameter of 0.6 mm were punched from representative tumor regions of each donor tissue block and brought into recipient paraffin block using a modified semi-automatic robotic precision instrument (Beecher Instruments, Woodland, CA, USA). Clinical and histopathological data were available for all these patients. Colorectal Unit, Department of Surgery at King Faisal Specialist Hospital and Research Center provided long term follow up data. The Institutional Review Board of the King Faisal Specialist Hospital & Research Centre approved the study.
DNA extraction and purification
Genomic DNAs were extracted from paraffin-embedded matched normal and neoplastic primary tissues using Gentra Kit (Minneapolis, MN, USA) following a slight modification to the manufacturer's recommendation.
Microsatellite markers and analyses
Allelic imbalances were measured by performing microsatellite analysis on all matched normal and tumor tissue by PCR amplification. A reference panel of five pairs of microsatellite primers, comprising two mononucleotide microsatellites (BAT25, BAT26) and three dinucleotide microsatellites (DS123, D5S346 and D17S250) were used to determine tumor MSI status (Boland et al., 1998). Multiplex PCR was performed in a total volume of 25 μl using 50 ng of genomic DNA, 2.5 μl 10 × Taq buffer, 1.5 μl MgCl2 (25 mM), 10 pmol of fluorescent-labeled primers, 0.05 μl dNTP (10 mM) and 0.2 μl Taq polymerase (1 U μl−1) (all reagents were from Qiagen Inc., Valencia, CA, USA). PCR was performed using an MJ Research PTC-200 thermocycler. The PCR conditions were as follows: after an initial 10 min denaturation step at 95 °C, 40 amplification cycles were performed consisting of 40 s at 95 °C, 40 s at 54 °C and a 1 min elongation step at 72 °C. Amplification was completed with a final extension step at 72 °C for 7 min. The fluorescent-labeled products were finally analysed on an ABI PRISM 3100 × l Genetic Analyzer (Applied Biosystems, Foster City, CA, USA). Tumors were classified as MSI if at least two or more markers out of the five were unstable and as MSS if only one or none of the markers was unstable.
Mutation analysis of PIK3CA and p53 genes
Since vast majority of PIK3CA gene mutations in human cancers were reported in exons 9 and 20, we focused our mutation analysis on these exons (Samuels et al., 2004). Sequencing of PIK3CA exons 9, 20 was done by PCR amplification and direct sequencing of both strands for all CRC cases and their matched normal samples as previously described (Saal et al., 2005; Abubaker et al., 2007). In brief, step-down PCR was performed as follows: after a 10-min denaturing at 95 °C, the PCR was run with each temperature for 1 min at five step-down steps, for two cycles each. The denaturing temperature was 95 °C and extension temperature was 72 °C for each step, with the annealing temperatures of 66, 64, 62, 60 and 58 °C from the first to the last step. The PCR was finally run at 95, 58 and 72 °C each for 1 min for 35 cycles, followed by an elongation at 72 °C for 5 min. PCR was performed in a total volume of 25 μl using 50 ng of genomic DNA, 2.5 μl 10 × Taq buffer, 1.5 μl MgCl2 (25 mM), 0.05 μl dNTP (10 mM), 0.2 μl Taq polymerase (1 U μl−1) (all reagents were from Qiagen Inc.), 1 μl of each primer (2.5 μM) and water. The efficiency and quality of the amplification PCR were confirmed by running the PCR products on a 2% agarose gel. The PCR products were subsequently subjected to direct sequencing PCR with BigDye terminator V 3.0 cycle sequencing reagents (Applied Biosystems). The samples were finally analysed on an ABI PRISM 3100 × l Genetic Analyzer (Applied Biosystems). Primer pairs flanking PIK3CA exons-9 and -20 were selected to avoid the frequent cross-amplification of chromosome 22q (known PIK3CA pseudogene) observed with those previously reported (Campbell et al., 2004; Saal et al., 2005).
For p53 mutational analysis, exons 5–8 of the P53 gene were amplified separately using the following primer sequences: exon-5-forward: 5′GACTTTCAACTC-TGTCTC3′, reverse: 5′CTGGGGACCCCTGGGCAAC3′; exon-6-forward: 5′GAGACGACAGGGCTGGTT3′, reverse:5′CCACTGACAACCACCCTT3′; exon-7-forward: 5′CCAAGGCGCACTGGCCTC3′, reverse: 5′GCGGCAAGCAGAGGCTGG3′ and exon-8-forward: ′CCTTACTG-CCTCTTGCTT3′, reverse: 5′TGAATCTGAGGCATAA-CTGC3′. The samples were finally analysed on an ABI PRISM 3100 × l Genetic Analyzer (Applied Biosystems). Mutational analysis was done using DNA SEQMAN software (DNASTAR Inc., Madison, WI, USA).
TMA slides were processed and stained manually. The streptavidin-biotin peroxidase technique with diaminobenzidine as chromogen was applied. For antigen retrieval, Dako Target Retrieval Solution pH 9.0 (catalogue number S2368) was used, and the slides were microwaved at 750 W for 5 min and then at 250 W for 30 min. The sections were incubated overnight with monoclonal antibodies specific for PTEN (6H2-1, Cascade Bioscience, Winchester, MA, USA), p-AKT (monoclonal, clone Ser 473, Survival Marker: Signal Stain Phospho-AKT (Ser 473) IHC detection kit (Cell Signaling Technology, Beverly, MA, USA; product number 8100) and the Dako Envision Plus System kit was used as the secondary detection system with DAB as chromogen. All slides were counterstained with hematoxylin, dehydrated, cleared, and cover slipped with premount.
For PTEN scoring, the cases were grouped into normal (2+ and 3+), reduced (1+) and loss of expression (0) as described previously (Bose et al., 2002). Comparisons were made with controls that included tissue microarray sections of normal lymph nodes, normal tissues from various sites (Abbott et al., 2003). Endothelial cells in the neovascular capillaries and vessels consistently showed intense staining for PTEN (3+) and acted as an internal positive control when present. p-AKT scoring was done as described earlier (Bose et al., 2006; Uddin et al., 2006). Briefly, p-AKT was scored as levels on an intensity scale ranging from 0 to 3. Scoring was performed as follows: 0, no appreciable staining in tumor cells; 1, barely detectable staining in tumor cells; 2, appreciable staining of moderate intensity, distinctly marking tumor cells and 3, readily appreciable staining of strong intensity. For purposes of statistical analysis, all cases staining at level 0 or 1 were grouped as p-AKT negative and all cases staining at level 2 or 3 were grouped as p-AKT positive. Only fresh cut slides were stained simultaneously to minimize the influence of slide ageing and maximize repeatability and reproducibility of the experiment. Two types of negative controls were used for p-AKT. One was the negative control in the kit in which the primary antibody was omitted. A pre-absorption experiment using p-AKT Ser 473 blocking peptide (Cell Signaling Technology; Product No 1140) was used as the second negative control.
The software used for statistical analysis was Statview 5.0. (SAS Institute Inc., NC, USA). The correlation of coefficients between pairs of variables was done using Pearson's correlation. Survival curves were constructed by the Kaplan–Meier method and multivariate analysis by Cox regression; P-values less than 0.05 were considered significant. Two-sided tests were used throughout the analyses.
About this article
The Role of the Insulin-like Growth Factor-1 Receptor (IGF-1R), Phosphatase and Tensin Homolog (PTEN), c-Met, and the PI3-Kinase Pathway in Colorectal Cancer
Current Colorectal Cancer Reports (2012)