Preclinical data suggest that PIK3CA mutations predict response to PI3K/AKT/mTOR inhibitors. Concomitant KRAS or BRAF mutations may mediate resistance. Therefore, tumors from patients referred to the phase I program for targeted therapy starting in October 2008 were analyzed for PIK3CA mutations using PCR-based DNA sequencing of exons 9 and 20. Consecutive patients with diverse tumor types and PIK3CA mutation were treated whenever possible with agents targeting the PI3K/AKT/mTOR pathway. Overall, PIK3CA mutations were detected in 25 of 217 patients (11.5%; exon 9, n = 11; exon 20, n = 14). In tumor types with more than 10 patients tested, PIK3CA mutations were most frequent in endometrial (3 of 14, 21%), ovarian (5 of 30, 17%), colorectal (9 of 54, 17%), breast (2 of 14, 14%), cervical (2 of 15, 13%), and squamous cell cancer of the head and neck (1 of 11, 9%). Of the 25 patients with PIK3CA mutations, 17 (68%) were treated on a protocol that included a PI3K/AKT/mTOR pathway inhibitor, and 6 (35%) achieved a partial response. In contrast, only 15 of 241 patients (6%) without documented PIK3CA mutations treated on the same protocols responded (P = 0.001). Of the 17 patients with PIK3CA mutations, 6 (35%) had simultaneous KRAS or BRAF mutations (colorectal, n = 4; ovarian, n = 2). Colorectal cancer patients with PIK3CA and KRAS mutations did not respond to therapy, whereas both ovarian cancer patients with PIK3CA and KRAS or BRAF mutations did. In conclusion, PIK3CA mutations were detected in 11.5% of patients with diverse solid tumors. The response rate was significantly higher for patients with PIK3CA mutations treated with PI3K/AKT/mTOR pathway inhibitors than for those without documented mutations. Mol Cancer Ther; 10(3); 558–65. ©2011 AACR.
This article is featured in Highlights of This Issue, p. 373
Recently, major therapeutic advances have been made in tumors with druggable targets (1–4). These include the highly successful use of KIT kinase inhibitors in KIT mutation-positive gastrointestinal stromal tumors (GIST), ABL kinase inhibitors in BCR-ABL–positive chronic myelogenous leukemia (CML), and BRAF inhibitors in BRAF mutation-positive melanoma (1, 2, 4). Common solid tumors, such as breast, lung, and colorectal cancer remain difficult to treat, perhaps in part because they are heterogeneous, with each subset of patients having different molecular abnormalities (3). Identifying relevant molecular subtypes within heterogeneous diseases and matching patients with appropriate targeted agents or their combinations is crucial to future therapeutic progress (5).
The phosphatidylinositol 3-kinase (PI3K)/AKT/mTOR signaling pathway is activated in many different cancers (Supplementary Fig. S1; ref. 6). Activation is frequently mediated by mutations in the p110α subunit of PI3K called PIK3CA, with most mutations (>80%) occurring either in exon 9, which codes for the helical domain, or exon 20, which codes for the kinase domain (Supplementary Fig. S2; ref. 7). Preclinical studies suggested that PIK3CA mutations may predict for response to PI3K inhibitors (8).
We investigated the PIK3CA mutation status of patients referred to the phase I clinical trials program clinic (known as the Clinical Center for Targeted Therapy). Whenever possible, patients with PIK3CA mutations were offered treatment targeting the PI3K/AKT/mTOR pathway, and their clinical outcomes were analyzed.
Patients and Methods
We investigated the PIK3CA mutation status of patients with advanced tumors and available tissue referred to the Department of Investigational Cancer Therapeutics (phase I clinical trials program) at The University of Texas M.D. Anderson Cancer Center starting in October 2008. The registration of patients in the database, pathology assessment, and mutation analysis were done at M.D. Anderson. Eligible patients were those referred for clinical trials of targeted therapeutic agents. The study and all treatments were conducted in accordance with the guidelines of the M.D. Anderson Institutional Review Board.
Tissue samples and mutation analyses
PIK3CA mutations were investigated in archival formalin-fixed, paraffin-embedded tissue blocks or material from fine needle aspiration biopsy obtained from diagnostic and/or therapeutic procedures. All histologies were centrally reviewed at M.D. Anderson. PIK3CA mutation testing was done in the Clinical Laboratory Improvement Amendment–certified Molecular Diagnostic Laboratory within the Division of Pathology and Laboratory Medicine at M.D. Anderson. DNA was extracted from microdissected paraffin-embedded tumor sections and analyzed using a PCR-based DNA sequencing method for PIK3CA mutations in codons [c]532–554 of exon 9 (helical domain) and c1011–1062 of exon 20 (kinase domain), which included the mutation hotspot region of the PIK3CA proto-oncogene by Sanger sequencing following amplification of 276 bp and 198 bp amplicons, respectively, utilizing primers designed by the M.D. Anderson Molecular Diagnostic Laboratory. Whenever possible, in addition to PIK3CA, mutation analysis was done for KRAS and NRAS c12, c13, and c61 mutations of exon 2; and BRAF c595–600 mutations of exon 15 by pyrosequencing as previously described (9).
Treatment and evaluation
Consecutive patients with underlying PIK3CA mutations were enrolled whenever possible in clinical trials containing inhibitors of the PI3K/AKT/mTOR pathway, particularly protocols with anti–mTORC1 (rapalog)-based regimens or regimens containing PI3K inhibitors. Treatment continued until disease progression or unacceptable toxicity occurred.
Treatment was carried out according to the specific requisites in the treatment protocols selected.
Assessments, including history, physical examination, and laboratory evaluations, were done as specified in each protocol, typically before the initiation of therapy, weekly during the first cycle, and then, at a minimum, at the beginning of each new treatment cycle. Efficacy was assessed from computed tomography scans and/or magnetic resonance imaging at baseline before treatment initiation and then every 2 cycles (6–8 weeks). All radiographs were read in the Department of Radiology at M.D. Anderson and reviewed in the Department of Investigational Cancer Therapeutics tumor measurement clinic. Responses were categorized per RECIST 1.0 criteria and were reported as best response (10). In brief, complete response was defined as the disappearance of all measurable and nonmeasurable disease; partial response (PR) was defined as at least a 30% decrease in the sum of the longest diameter of measurable target lesions; progressive disease (PD) was defined as at least a 20% increase in the sum of the longest diameter of measurable target lesions, or unequivocal progression of a nontarget lesion, or the appearance of a new lesion; and stable disease (SD) was defined as neither sufficient shrinkage to qualify for PR nor sufficient increase to qualify for PD.
Statistical analysis was verified by our statistician (X. Wang). Fisher's exact test was used to assess the association among categorical variables and PIK3CA mutation status. The Wilcoxon rank-sum test assessed the association between age and PIK3CA mutation status. Time to progression was defined as the time interval from the start of therapy to the first observation of disease progression or death, whichever occurred first. All tests were 2-sided, and P < 0.05 was considered statistically significant. All statistical analyses were carried out using SAS 9.1 software (SAS Institute).
A total of 217 patients with advanced tumors were analyzed for the presence of PIK3CA mutations. One hundred thirty-one (60%) patients were women and 86 (40%) were men (Table 1). The median age was 56 years (range, 13–91 years). One hundred seventy-six (81%) were Caucasians, 19 (9%) African Americans, 12 (6%) Hispanic, and 10 (4%) Asians. Fifty-four (25%) had colorectal cancer, 30 (14%) ovarian cancer, 18 (8%) melanoma, 14 (6%) breast cancer, 14 (6%) endometrial cancer, 8 (4%) squamous cell cervical cancer, 7 (3%) cervical adenocarcinoma, 8 (4%) soft tissue sarcoma (excluding GIST), and 11 (5%) squamous cell cancer of head and neck. A variety of other tumors made up the rest of the patients (Table 1).
PIK3CA proto-oncogene mutations were detected in 25 of the 217 patients (11.5%). In 11 patients, a mutation in exon 9 was detected: 8 in c545, 1 in c542, 1 in c546, and 1 in c545/c549. Exon 20 mutations were found in the 14 remaining individuals: 10 in c1047, 3 in c1049, and 1 in c1043 (Table 2). In tumor types with more than 3 patients tested, PIK3CA mutations were most frequent in endometrial cancer (3 of 14 patients, 21%). Mutations were also present in 5 of 30 patients (17%) with ovarian cancer, 9 of 54 patients (17%) with colorectal cancer, 2 of 14 patients (14%) with breast cancer, 2 of 15 patients (13%) with cervical cancer (both of whom had squamous histology), 1 of 7 patients (14%) with non–small cell lungcarcinoma (NSCLC), and 1 of 11 patients (9%) with squamous cell cancer of head and neck (Table 1). PIK3CA mutation status was not significantly associated with age, gender, or race (Fisher's exact test).
Simultaneous RAS and PIK3CA mutations
Preclinical data suggest that activation of RAS mediates resistance to PI3K inhibitors (8). Therefore, RAS mutation status was investigated whenever possible.
KRAS mutations in exon 2 were assessed in 130 patients and identified in 33 individuals (25%; Table 3). The presence of KRAS mutations was significantly associated with PIK3CA mutations (P = 0.03; Fisher's exact test). Indeed, 45% of patients (9 of 20) with a PIK3CA mutation (who were also tested for a KRAS mutation) also had a KRAS mutation, whereas only 22% of patients (24 of 110) without a PIK3CA mutation (who were also tested for KRAS) harbored a KRAS mutation. Of the 33 patients with KRAS mutations, 9 (27%) had simultaneous PIK3CA mutations. In contrast, of the 97 patients without a KRAS mutation, only 11 (11%) had a PIK3CA mutation (P = 0.03). Of the 9 patients with simultaneous PIK3CA and KRAS mutations, disease distribution was as follows: colorectal cancer, 7; pancreatic cancer, 1; ovarian cancer, 1 (Table 2).
NRAS c61 mutations were detected in 2 patients (3%) of the 62 tested. Both patients had melanoma with wild-type PIK3CA.
Simultaneous BRAF and PIK3CA mutations
BRAF exon 15 mutations were assessed in 122 patients, of whom 11 (9%) had a c600 mutation (Table 3). BRAF mutations were found in 7 patients with melanomas, 3 with colorectal cancer, and 1 with ovarian cancer. Only 1 patient had a simultaneous PIK3CA and BRAF mutation detected (ovarian cancer; Table 2).
Response in patients with PIK3CA mutations treated with PI3K/AKT/mTOR inhibitors
Of the 25 patients with an underlying PIK3CA mutation, 17 were enrolled in clinical trials that included a PI3K/AKT/mTOR inhibitor. These patients were refractory to a median of 4 prior therapies (range, 1–12). Of these patients, 6 had colorectal cancer, 4 had ovarian cancer, 3 had endometrial cancer, 2 had squamous cell cervical cancer, 1 had small intestine cancer, and 1 had breast cancer (Table 2). Sixteen patients received anti–mTORC1-based regimens and 1 patient was treated with an anti–PI3K-based regimen (Table 4; ref. 11). A PR was observed in 6 (35%) patients. Duration of response was 8.9, 17.9+, 25+, 30.6+, 35.3, and 59+ weeks (with the plus sign indicating ongoing response at the time period noted; Figs. 1, 2, 3, and 4). Seven (41%) patients had SD, including 4 (23.5%) patients with prolonged SD lasting for more than 16 weeks. In total, 10 patients (59%) achieved either SD for more than 16 weeks or a PR. Four (23.5%) patients had PD (2 with radiological and 2 with clinical progression). In comparison, patients without documented PIK3CA mutations (meaning they had no mutation or tumor tissue was unavailable for testing) treated on the same protocols showed a significantly lower response rate of 6% (15 of 241; P = 0.001). Fisher's exact test was used to assess the associations among response (PR, SD, or PD) and other patient characteristics, such as age, gender, race, number of prior therapies (>3 prior therapies vs ≤ 3 therapies), type of PIK3CA mutation (exon 9 vs exon 20), and KRAS mutation. None of these variables was significantly associated with response.
We determined that mutations in exon 9 or exon 20 of the PIK3CA proto-oncogene were present in 25 of 217 patients (11.5%) with diverse tumor types, with the incidence being highest (9%–21%) in patients with endometrial, ovarian, colorectal, breast, cervical cancer, NSCLC, and squamous cell cancer of head and neck. Although the number of patients in each tumor type is limited in our study, previous reports have also documented PIK3CA mutations in these tumor types with an incidence as follows: 23% to 36% of endometrial cancers (12, 13), 14% to 32% of colon cancers (12, 14, 15), 4% to 12% of ovarian cancers (16–18), 18% to 40% of breast cancers usually associated with expression of hormone receptors or HER2/neu (12, 16, 17, 19), 8% to 14% of cervical squamous cell cancers (12, 20), and in 11% to 33% of squamous cell cancers of the head and neck (12, 21).
Previous preclinical observations have shown that activation of the RAS/RAF/MEK pathway mediates resistance to PI3K inhibitors in PIK3CA-mutant tumors (8). Therefore, we examined our patients for coexistence of PIK3CA mutations with RAS (K- or N-) or BRAF mutations. Forty-five percent of patients (9 of 20) with a PIK3CA mutation (who were also tested for a KRAS mutation) also had a KRAS mutation, whereas only 22% of patients (24 of 110) without a PIK3CA mutation (who were also tested for KRAS) harbored a KRAS mutation. Of the 33 patients with KRAS mutations, 9 (27%) had simultaneous PIK3CA mutations. In contrast, of the 97 patients without KRAS mutation, only 11 (11%) had a PIK3CA mutation (P = 0.03). These results suggest that these mutations (PIK3CA and RAS) commonly coexist. We also observed that 7 of 9 patients (78%) with colorectal cancer who harbored a PIK3CA mutation also had a KRAS mutation. This rate of dual mutations is similar to that in a previously reported study that showed mutant KRAS in 56% of patients with colorectal cancer and a PIK3CA mutations (14). In 1 of 5 patients with PIK3CA-mutant ovarian cancer, we detected a simultaneous KRAS mutation and, in another, a simultaneous BRAF mutation. In contrast, a study from the Middle East showed no coexistence of mutated KRAS or BRAF mutations with PIK3CA mutations in ovarian cancer, though the incidence of PIK3CA mutations in the population studied was quite low (4%; ref. 18).
Whenever possible, our patients with PIK3CA mutations were entered on trials utilizing targeted inhibitors of the PI3K/AKT/mTOR pathway. Their overall response rate on these trials was 35%. Responses were seen in patients with cervical, endometrial, ovarian cancer, and breast cancer (Fig. 1). In contrast, of the 241 patients without documented PIK3CA mutations treated on the same protocols, only 15 (6%) responded (P = 0.001). The latter response rate is similar to the 4% to 11% response rate reported by our group and others when patients are treated on phase I trials without molecular selection (22–24). It should be noted that it is conceivable that some of the small group of patients without PIK3CA mutations who responded to PI3K/AKT/mTOR axis inhibitors had other aberrations in PIK3CA not detected by our assay or had other abnormalities such as PTEN loss, that are known to activate PIK3CA (6). Indeed, we have previously shown that PTEN loss can be detected in approximately 20% of patients in the phase I setting (25).
Consistent with our data, clinical trials with therapies directed against well-defined targets have shown improved results when patients are selected for the presence of those targets, even in the phase I setting (where patients tend to be heavily pretreated and refractory/resistant to multiple conventional drugs), though mostly these trials have reported results in a disease-specific setting. Examples include imatinib mesylate (a KIT and BCR-ABL kinase inhibitor), which showed response rates of more than 50% in patients with GIST (a disorder characterized by KIT kinase mutations) or BCR-ABL–positive CML (1, 26). More recently, patients with NSCLC and an underlying EML4-ALK fusion also showed a response rate more than 50% after treatment with the ALK inhibitor crizotinib, as did patients with metastatic malignant melanoma who had an underlying BRAF mutation and responded to the BRAF inhibitor PLX4032 (4, 27). In contrast, epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors were initially tested in an unselected patient population and had only modest activity (28). Subsequent bench to the bedside forays showed that anti-EGFR tyrosine kinase inhibitors are far more effective in patients with lung cancer and an underlying EGFR mutations (29).
One question that arises is whether or not the detection of additional mutations that might confer resistance would provide more predictive information. In this regard, our patients with colorectal cancer and a simultaneous KRAS mutation did not respond to PI3K/AKT/mTOR axis therapy, which is in agreement with preclinical data suggesting that KRAS activation mediates resistance to PI3K inhibitors (8). In contrast, 2 patients with ovarian cancer and simultaneously occurring KRAS or BRAF mutations achieved a PR with PI3K/AKT/mTOR axis inhibitors, thus suggesting that such resistance is not absolute or that RAS- or RAF-mutant colorectal cancers behave differently than RAS- or RAF-mutant ovarian cancers.
In conclusion, mutations in PIK3CA occur in a subset of patients with several common cancers. In this study, the response rate for patients with heavily pretreated, diverse, advanced cancers and PIK3CA mutations who were given PI3K/AKT/mTOR axis inhibitors was significantly more than that for patients without documented PIK3CA mutations treated on the same trials. The latter observation is consistent with data that show low response rates on traditional phase I trials, in which molecular testing is not used. One hypothesis that could be generated from these data is that selecting PIK3CA-mutant patients for treatment with PI3K/AKT/mTOR axis inhibitors may predict response independent of underlying histology. Patients with colorectal cancer and concomitant presence of KRAS and PIK3CA mutations did not respond, consistent with previous experiments indicating that the RAS/RAF/MEK pathway serves as a driver of resistance to PI3K inhibitors. Because the number of patients in our series was small and no randomization occurred, these data must be interpreted cautiously. However, it seems that screening for PIK3CA (and RAS or RAF) mutations warrants further investigation in the application of targeted PI3K/AKT/mTOR inhibitors to the clinic.
Disclosure of Potential Conflict of Interest
No potential conflicts of interest were disclosed.
This study was supported in part by Grant RR024148 from the National Center for Research Resources, a component of the NIH Roadmap for Medical Research (http://nihroadmap.nih.gov/clinicalresearch/overview-translational.asp).
We thank Ms. Joann Aaron for scientific review and editing of this article.
Note: Supplementary data for this article are available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/).
- Received November 3, 2010.
- Revision received December 30, 2010.
- Accepted January 1, 2011.
- ©2011 American Association for Cancer Research.