Figure. Examples of activating mutations at codons 542 (A), 545 (B), and 1047 (C). E indicates glutamic acid; I, isoleucine; K, lysine; H, histidine; R, arginine; T, threonine. The green boxed A indicates adenine; the blue boxed C, cytosine; the black boxed G, guanine; and the red boxed T, thymine.
Anthony C. Nichols, David A. Palma, Winsion Chow, Susan Tan, Chandheeb Rajakumar, Giananthony Rizzo, Kevin Fung, Keith Kwan, Brett Wehrli, Eric Winquist, James Koropatnick, Joe S. Mymryk, John Yoo, John W. Barrett. High Frequency of Activating PIK3CA Mutations in Human Papillomavirus–Positive Oropharyngeal Cancer. JAMA Otolaryngol Head Neck Surg. 2013;139(6):617–622. doi:10.1001/jamaoto.2013.3210
Author Affiliations: Departments of Otolaryngology Head & Neck Surgery (Drs Nichols, Chow, Tan, Rajakumar, Fung, Yoo, and Barrett), Oncology (Drs Nichols, Palma, Fung, Winquist, Koropatnick, Mymryk, Yoo, and Barrett), Pathology (Drs Nichols, Kwan, and Wehrli), and Microbiology and Immunology (Dr Mymryk), The University of Western Ontario, London, Ontario, Canada; London Regional Cancer Program, London, Ontario (Drs Nichols, Palma, Fung, Winquist, Koropatnick, Mymryk, Yoo, and Barrett and Mr Rizzo); and Lawson Health Research Institute, London, Ontario (Drs Nichols, Palma, Koropatnick, Mymryk, and Barrett).
Importance Large-scale whole-exome sequencing studies of head and neck squamous cell carcinoma (HNSCC) have established that the disease is dominated by frequent mutations in tumor suppressor genes with rare activating mutations in oncogenes that would be easily targetable with molecular agents. There was evidence in these reports, however, that activating mutations in phosphoinositide 3-kinase catalytic subunit p110α (PIK3CA) were common in patients with human papillomavirus (HPV)-positive tumors. We set out to test this prediction in oropharyngeal patient samples from our institution.
Objective To confirm whether activating mutations in PIK3CA are frequent in HPV-positive HNSCC because this mutated oncogene represents a potential therapeutic target.
Design, Setting, and Participants A retrospective search of the London Health Sciences Centre pathology database was performed to identify oropharyngeal cancer samples. DNA from pretreatment primary site biopsy samples from 87 patients were tested for high-risk HPV types 16 and 18 by real-time polymerase chain reaction.
Main Outcomes and Measures Samples were tested for activating mutations at the 3 mutational hot spots (codons 542, 545, and 1047) by polymerase chain reaction followed by Sanger sequencing using forward and reverse primers.
Results Only 4 of 41 HPV-negative tumors (10%) demonstrated PIK3CA hot spot mutations, including 3 at codon 1047 and 1 at codon 542. Of 46 HPV-positive tumors, 13 (28%) demonstrated activating PIK3CA mutations, including 7 at codon 542, 5 at codon 545, and 1 at codon 1047. The difference in PIK3CA mutation frequency was significantly different between HPV-positive and HPV-negative cancers (P = .03).
Conclusions and Relevance Although there has been a suggestion that activating PIK3CA mutations are common in HPV-positive HNSCC, to our knowledge, this is the first study to clearly identify this phenomenon. Targeting PIK3CA with molecular agents in HPV-positive patients may be a mechanism to improve cure rates and decrease treatment toxic effects in this rapidly growing cohort of patients.
Because of increasing rates of oral infection with human papillomavirus (HPV), the incidence of oropharyngeal squamous cell carcinoma (OPSCC) has been rapidly increasing.1 Patients with HPV-related cancers tend to present at a younger age and experience markedly improved survival compared with patients with smoking- and alcohol-related head and neck cancers.1 Given that these younger patients with HPV-positive tumors are likely to survive their cancer and have to cope with the adverse effects of therapy for decades, optimizing the posttreatment quality of life for patients with OPSCC has become one of the most important issues in head and neck oncology. Many institutions favor cisplatin-based concurrent chemoradiation for high-risk OPSCC.2 However, this combined-modality treatment has notable acute and late toxic effects including dysphagia, mucositis, xerostomia, fibrosis, osteoradionecrosis, neutropenia, neurotoxicity, nephrotoxicity, and ototoxicity.3 Thus, alternative treatment options are needed to decrease toxic effects while maintaining high cure rates.
Two large-scale whole-exome sequencing studies of head and neck cancer have recently defined the genetic makeup of head and neck squamous cell cancer (HNSCC).4,5 These multi-institutional efforts revealed that the mutational landscape of HNSCC is dominated by alterations in tumor suppressors including TP53, CDKN2A, and NOTCH1, NOTCH2, and NOTCH3, while activating mutations in oncogenes were observed at a very low frequency including HRAS (5%) and PIK3CA (8%).4,5 This low frequency of activating mutations in oncogenes is disappointing from a therapeutic perspective because activated oncogene products can potentially be targeted with inactivating small molecule drugs. However, effective strategies to replace lost tumor suppressor function are lacking. A small cohort of the patients in the 2 studies were HPV positive: 4 of 32 patients in the study by Agrawal et al4 (2011) and 11 of 74 patients in the study by Stransky et al5 (2011). Stransky et al5 noted a total of 6 PIK3CA mutations in all subjects (both HPV positive and HPV negative), but a much higher proportion (albeit in a small total number of subjects) were found in the HPV-positive patients (3 of 11 [27%]) vs the HPV negative patients (3 of 63 [5%]). The aim of this study was to confirm whether activating mutations in PIK3CA are, in fact, frequent in HPV-positive HNSCC because this mutated oncogene represents a potential therapeutic target.
Study approval was obtained from the University of Western Ontario Research Ethics Board. A retrospective search of the London Regional Cancer Program (LRCP) pathology database was performed to identify patients diagnosed as having OPSCC from 2003 to 2009. Patient eligibility required (1) a histologically confirmed diagnosis of squamous cell carcinoma, (2) no history of head and neck cancer, and (3) the availability of a pretreatment primary site biopsy specimen for analysis. Patient data were extracted from a retrospective medical chart review, including age at diagnosis, use of tobacco and alcohol, American Joint Committee on Cancer TNM stage, treatment regimen, and posttreatment follow-up information.
Deparaffinization, DNA extraction, and detection of high-risk HPV types by multiplex quantitative polymerase chain reaction (qPCR) was performed as described previously.6 Standard curves were produced from 10-fold serial dilutions of CaSki cell genomic DNA. CaSki cell DNA was used as a positive control for HPV16, and HeLa cell genomic DNA was used as the HPV18 control.
Tumor DNA was tested with primers flanking the known PIK3CA hot spots at codons 542, 545, and 1047 (Table 1). The PCR reactions (20 μL) were prepared using 0.2 μL of patient template DNA and amplified with Phusion DNA polymerase (Fisher Scientific) following a hot start (98°C for 30 seconds, followed by 37 cycles at 98°C for 10 seconds, 62°C for 10 seconds, and 72°C for 30 seconds to amplify the 542 and 545 regions and 98°C for 30 seconds, and followed by 37 cycles at 98°C for 10 seconds, 47°C for 10 seconds, and 72°C for 30 seconds to amplify the 1047 region). Amplified products were resolved on agarose gel, 1%. Bands of interest were extracted from agarose gel slices, purified on columns (Bio Basic), and Sanger sequenced using both forward and reverse primers. Samples with identified mutations were confirmed by repeated PCR and sequencing.
Statistical analysis was performed with the R system (R Foundation for Statistical Computing). For all analyses, P ≤ .05 was considered statistically significant. Patient variables were compared using the Fisher exact test.
Of the 87 patients, 46 were positive for HPV (53%), of which 44 were positive for type 16 and 2 were positive for type 18. A total of 17 PIK3CA mutations were noted in the 87 patients for a frequency of 19.5%. In all cases, the mutation represented a transversion. At codon 542 and 545, there was a substitution from a guanine to an adenine resulting in a nonsynonymous substitution resulting in a change from a glutamic acid (E) to a lysine (K). At codon 1047, the transversion occurred when an adenine changed to a guanine. Again, this resulted in a nonsynonymous change from a histidine (H) to an arginine (R) (examples of the Sanger sequencing results are shown in the Figure). The specific mutations observed are outlined in Table 2. PIK3CA mutations were noted in 13 of 46 HPV-positive patients (28%) and in 4 of 41 HPV-negative patients (10%), which was statistically significant (P = .03). There was a difference in the codon distribution of the mutations between the HPV-positive and negative patients, with the HPV-positive patients more likely to harbor mutations at codons 542 and 545 (12 of 13 mutations), while 3 of 4 PIK3CA mutations in the HPV-negative tumors occurred at codon 1047 (Table 2; P = .009).
Patient demographics, tumor stage, and tobacco and alcohol use were correlated with PIK3CA mutation status for the entire cohort and for HPV-positive and negative tumors separately (Table 3). Activating PIK3CA mutations were associated with less tobacco use (P = .01), and there was a trend toward fewer PIK3CA mutations in heavy drinkers (P = .09).
Two high-impact companion publications reported the mutational landscape of head and neck squamous cell carcinoma (HNSCC) using whole-exome sequencing.4,5 These studies provided new insights into the genetic underpinnings of the disease and highlighted the high frequency of mutations in tumor suppressors TP53, CDKN2A, NOTCH1, NOTCH2, and NOTCH3. Somewhat disappointingly, there was a very low incidence of activating mutations in oncogenes. Of the few observed, PIK3CA and HRAS mutations were the most frequent, comprising 8% and 5% of the samples, respectively. The ideal result would have been a frequent mutation in a targetable oncogene such as the frequent mutations in BRAF observed in approximately 50% of malignant melanomas, which are targetable with BRAF inhibitors.7
However, both studies included only a small number of HPV-positive patients, who represent a rapidly increasing fraction of contemporary head and neck oncology. A close examination of the data in the study by Stransky et al5 revealed that the frequency of PIK3CA mutations were not equally distributed. Of the 63 HPV-negative tumors, 3 (5%) harbored PIK3CA mutations, while 3 of 11 HPV-positive tumors (27%) contained mutations. The study by Agrawal et al4 demonstrated PIK3CA mutations in 1 of 4 HPV-positive tumors and 2 of 28 HPV-negative tumors.4 Combining these 2 high-impact studies; there were PIK3CA mutations in 4 of 15 HPV-positive samples (25%) and 5 of 91 HPV-negative samples, which was significant (P = .02, Fisher exact test). Other groups have assessed PIK3CA status by PCR in head and neck tumors without regard to HPV status and revealed low rates of mutations.8,9 Kozaki et al8 (2006) evaluated 108 oral cancers and found a mutation frequency of 7%. Murugan et al9 (2008) studied 37 head and neck cancers from a variety of sites and found mutations in only 5%. In contrast to these studies, which largely included sites that are generally not HPV related, we specifically focused our efforts on the oropharynx in view of the findings of exome sequencing studies.4,5 Our study confirms that activating mutations are indeed more frequent in HPV-positive tumors with a comparable frequency to the exome sequencing studies (28%). When we pooled our results with the exome sequencing studies, the findings were significant (17 of 61 HPV-positive tumors and 9 of 133 HPV-negative tumors; P < .001, Fisher exact test). This finding, in addition to the fact that PIK3CA is frequently amplified in HNSCC,10 provides significant evidence that PIK3CA represents an important therapeutic target particularly in HPV-positive disease.
We chose to examine only the known hot spots at codons 542, 545, and 1047 outlined in the COSMIC database (http://www.sanger.ac.uk/perl/genetics/CGP/cosmic?action=bygene&ln=PIK3CA&start=1&end=1069 &coords=AA:AA) and that are well described to be oncogenic and accounted for 5 of 6 mutations noted by Stransky et al.5 The nonsynonymous mutations resulting in the changes E542K and E545K are located within exon 9 in the helical domain of PIK3CA, whereas H1047R is encoded within exon 20 and is part of the kinase domain. The probable mechanism for the oncogenicity of the E542K and E545K mutations is the disruption of an inhibitory charge-charge interaction between PIK3CA and the N-terminal SH2 domain of the p85 regulatory subunit.11 In contrast, the H1047R mutation increases binding affinity of PIK3CA for the negatively charged phosphatidylinositol substrate.11
Genome-wide analysis has shown that mutations are twice as frequent in HPV-negative than in HPV-positive tumors (4.83 vs 2.28 mutations per megabase pair, respectively).5 In strong contrast, our results show that the PIK3CA gene exhibited a higher number of mutations in the HPV-positive population. This observation suggests that either HPV cancers frequently require activation of the PIK3CA pathway to progress or that an HPV-specific mechanism leads to genetic instability in this gene. Although the connection between HPV and PIK3CA has not been specifically studied in head and neck cancer, Henken et al12 observed overexpression of PIK3CA in HPV-transformed human foreskin keratinocytes. Reduction of PIK3CA activity in these human foreskin keratinocytes resulted in reduced cellular viability, migration, and anchorage-independent growth.12 These results suggest that this pathway is critical for HPV-related oncogenesis and that these cells are “addicted” to activated PIK3CA. Further study is needed to better characterize the mechanistic roles by which PIK3CA contributes to HPV-positive head and neck cancers.
Our results also demonstrate a strong correlation between increased PIK3CA mutations and decreased exposure to tobacco. This observation is unexpected because increased mutation rates, particularly guanine to thymine (G→T) transversions in non-CpG sites, have been documented in patients with a heavy tobacco history.5 Thus, one might have predicted a decreased frequency of PIK3CA mutations in individuals with decreased tobacco consumption. Similar to our observations with PIK3CA, an increased frequency of activating mutations in epidermal growth factor receptor has been reported in non–small cell lung cancers from nonsmokers than from smokers.13 The mechanisms responsible for this phenomenon remain unknown.
The observation that HPV-positive patients tend to be younger, healthier, nonsmokers, and nondrinkers who experience markedly improved survival compared with their HPV-negative counterparts is now well described.1 Given that these young patients have a high likelihood of surviving their disease, their posttreatment quality of life becomes of paramount importance. Cisplatin-based chemoradiation is highly effective, particularly in the HPV-positive patient population. Most patients have very good functional outcomes after chemoradiation. However, a subset of patients experience severe acute and late toxic effects including hearing loss, neuropathy, renal failure, dysphagia requiring long-term gastrostomy tube dependence, and osteoradionecrosis that can have deleterious effects on patient quality of life.3 Thus, there is a great deal of interest in equally effective therapies with fewer adverse effects. However, therapy still fails in 10% to 20% of HPV-positive patients, with failure occurring because of distant metastases in approximately half.14,15 New treatment options are needed for these patients as well. Targeting the PIK3CA pathway is one strategy that may improve outcomes in HNSCC for patients with tumors harboring PIK3CA mutations. Phosphoinositide 3-kinase can be treated with pan-inhibitors, or PIK3CA can be targeted with agents specific to the catalytic 110α subunit.11 Alternate approaches include inhibiting downstream pathway members AKT and mammalian target of rapamycin (mTOR). There are numerous phase 2 trials testing these strategies in other cancers.11 In addition, 2 studies testing PX-866, a pan-PI3K inhibitor, in the metastatic setting combined with either cetuximab or docetaxel are under way for HNSCC and other solid tumors.16,17 Neither of these studies require the presence of an activating PIK3CA mutation as a criterion for entry into the study. This, combined with the low failure rate of HPV-positive patients (and thus low numbers of patients likely to enrolled compared with HPV-negative patients), and the confounding presence of another simultaneously administered chemotherapy agent make it unlikely that either will provide clear information about the role of this agent in HPV-positive HNSCC. Given the high incidence of activating PIK3CA mutations in HPV-positive HNSCC, a specific trial targeting only recurrent or metastatic HPV-positive patients harboring PIK3CA mutations is warranted.
In conclusion, activating PIK3CA mutations occur at a high frequency in HPV-positive oropharyngeal cancer, offering the possibility for personalized therapy with PIK3CA pathway inhibitors to improve outcomes in this rapidly expanding population of patients. Further work is necessary to understand the mechanism of interaction between PIK3CA and HPV.
Correspondence: Anthony C. Nichols MD, Victoria Hospital, London Health Science Centre, Department of Otolaryngology–Head and Neck Surgery, Room B3-431A, 800 Commissioners Rd E, London, ON N6A 5W9, Canada (Anthony.Nichols@lhsc.on).
Submitted for Publication: November 21, 2012; final revision received January 1, 2013; accepted January 25, 2013.
Author Contributions: Dr Nichols had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Yoo and Barrett contributed equally to this manuscript: Study concept and design: Nichols, Fung, and Yoo. Acquisition of data: Nichols, Chow, Tan, Rajakumar, Rizzo, Kwan, Wehrli, and Barrett. Analysis and interpretation of data: Nichols, Palma, Fung, Wehrli, Winquist, Koropatnick, Mymryk, and Barrett. Drafting of the manuscript: Nichols, Mymryk, and Barrett. Critical revision of the manuscript for important intellectual content: Nichols, Palma, Chow, Tan, Rajakumar, Rizzo, Fung, Kwan, Wehrli, Winquist, Koropatnick, Mymryk, Yoo, and Barrett. Statistical analysis: Nichols and Palma. Obtained funding: Nichols and Yoo. Administrative, technical, and material support: Tan, Fung, Wehrli, Koropatnick, Yoo, and Barrett. Study supervision: Nichols, Mymryk, Yoo, and Barrett.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported by a London Regional Cancer Program grant.