This article is mentioned in:

Introduction

Cervical cancer was previously one of the leading causes of cancer-related death in Japanese women and the second most common malignancy in women worldwide (1). This disease comprises several histologic types, of which squamous cell carcinoma is predominant and accounts for approximately 85–90% of cases. In contrast, adenocarcinomas/adenosquamous carcinomas are less common and represent 10–25% of cases (2–3). The incidence of invasive cervical adenocarcinoma and its variants has risen dramatically among younger women over the past few decades (4). The cause of this is not clear but is of concern as several studies have indicated that adenocarcinoma is associated with a worse prognosis compared to that of squamous cell carcinoma. Compared to those with squamous cell carcinoma, a higher proportion of lymph node involvement and distant metastases, as well as a decline in survival across stages, are typically found with adenocarcinoma (5).

Recently, the phosphatidylinositide 3-kinase (PI3K)/AKT signaling pathway has been found to be a major survival signal for cancer cells. Cell proliferation, growth, apoptosis, autophagy, invasion, and migration are regulated by the phosphatidylinositide 3-kinase (PI3K)/AKT serine/threonine kinase (AKT)/mechanistic target of rapamycin kinase (mTOR) pathways, which are putatively activated by key signals in different tumor types (6,7). Activation is commonly conferred by mutations in the p110α subunit of PI3K, PIK3CA, with most mutations (>80%) occurring either in the helical domain exon 9 or the kinase domain exon 20. Until recently, molecular genetic studies on cervical tumors have been limited. Cervical carcinomas in U.S. populations are frequently associated with mutations in PIK3CA, with a mutation frequency of 31% (8). Lou et al (9), analyzed 675 Latin American patients with cervical tumors and found that 31% of squamous carcinomas and 24% of adeno and adenosquamous carcinomas harbored mutations in the helical domain, and specifically at the E542 and E545 residues, which were thought to result in activation of this subunit and the PI3K/AKT pathway. In Chinese patients with cervical cancer, it was revealed that PIK3CA mutations are more common in squamous cell carcinomas (15.3%) than in non-squamous cell carcinomas (7.3%) (10). However, the prevalence of mutations in PI3KCA and how they are associated with clinicopathological characteristics and prognosis in Japanese patients with cervical cancer have not been studied until now. Therefore, the objective of this work was to analyze the relationship between amplification and somatic mutations in PIK3CA and various clinicopathologic variables including prognosis, in cervical cancer cell carcinomas from Japanese patients. Furthermore, we analyzed whether mutations in PIK3CA can predict response to PI3K/AKT/mTOR inhibition in cervical cancer.

Materials and methods

Tissue samples

A total of 71 paraffin-embedded tumor tissue samples were obtained from the Department of Obstetrics and Gynecology at Shimane University Hospital; all samples were cervical squamous cell carcinomas. In addition, 53 adenocarcinomas/adenosquamous carcinomas were obtained from the Department of Obstetrics and Gynecology at Seirei Hamamatsu General Hospital. Patients had received appropriate therapy at either Shimane University Hospital or Seirei Hamamatsu General Hospital between January 1994 and December 2013. All specimens from cervical cancer patients were obtained after operation and prior to any treatment. Tumor staging was performed according to the International Federation of Gynecology and Obstetrics (FIGO) classification (Shepherd, 2014). The invasive squamous cell carcinomas consisted of 28 cases of stage I disease, 11 of stage II disease, 24 of stage III disease, and eight of stage IV disease. All tumors were classified histologically according to the World Health Organization criteria. The median patient age was 55 years (range 32–84 years). The invasive adenocarcinomas/adenosquamous cell carcinomas consisted of 39 cases of stage I disease, eight of stage II disease, five of stage III disease, and one of stage IV disease. World Health Organization criteria were used to classify all tumors histologically. The median patient age was 46 years (range 27–82 years). Stage I and II patients were treated with class II or class III radical hysterectomies with pelvic lymph node dissection. Stage I patients were treated with positive lymph node metastasis or positive lymphovascular space invasion and all stage II patients received concurrent chemoradiotherapy or radiotherapy as adjuvant therapy. Stage III and IV patients were treated with concurrent chemoradiotherapy or radiotherapy alone.

Patients with an incomplete response to radiotherapy and patients with recurrent tumors were treated with a variety of salvage chemotherapy agents including cisplatin, carboplatin, and paclitaxel. The follow-up period ranged from 5 to 142 months, with a median of 65 months. Acquisition of tissue specimens and clinical information was approved by an institutional review board (Shimane University and Seirei Hamamatsu General Hospital), and written informed consent was obtained from all patients. Only patients with follow-up data were included. The paraffin tissue blocks were organized into tissue microarrays, each produced by removing 3-mm diameter cores of tumors from the block. Selection of the area for the core was made by a gynecologic oncologist (KN) and pathology technician (KI), and was based on review of the H&E slides.

DNA isolation

Unstained paraffin-embedded tissues were serially sectioned at a thickness of 10 µm and one adjacent hematoxylin and eosin-stained section was taken for identification and selection of tumor tissue. The carcinoma, comprising at least 85% of the total area was marked, and using a sterile needle, gross macroscopic dissection was performed. The dissected tissues were placed in microcentrifuge tubes and DNA isolation was performed as described previously (11).

Cell culture and cell lines

Hela and Hela P35 (adenocarcinoma), as well as ME180, TCS, and CaSki (squamous cell carcinoma) human cervical cell lines were obtained from Tohoku University (Sendai, Japan), whereas SKGIIIa, SKGIIIb, HCS2, and BOKU (also squamous cell carcinoma) cells were obtained from the Health Science Research Resources Bank (Tokyo, Japan). All human cervical cancer cell lines were maintained in Dulbecco's modified Eagle's medium (DMEM) (Life Technologies, Gaithersburg, MD, USA) supplemented with 5% fetal bovine serum, 100 U/ml penicillin, and 100 µg/ml streptomycin at 37°C in an atmosphere of 5% CO2.

Fluorescence in situ hybridization (FISH)

FISH was performed on 5-µm paraffin sections contained on a tissue microarray consisting of both carcinoma and normal tissue cores. Zytolight® SPEC PIK3CA/CEN 3 dual color probes (Zytovision, Bremerhaven, Germany) were used according to the manufacturer's protocol. Slides were denatured for 10 min at 75°C and hybridized at 37°C overnight with the probe mix. Cell nuclei were stained with 4′,6-diamidino-2-phenylindole. Signals were evaluated by two independent researchers using an Olympus Bx41 fluorescence microscope (Olympus Corporation, Tokyo, Japan). Separate narrow band-pass filters were used to detect the ZygGreen™, ZyRed™, and DAPI signals. Approximately 100 tumor cells were examined for each specimen at magnification ×60; a signal ratio of experimental probe/reference probe greater than three was considered amplification.

Mutational analysis of PIK3CA, KRAS, and BRAF

The mutational status of KRAS and BRAF was determined for all paraffin-embedded cervical cancer tissue samples, whereas PIK3CA mutation status was determined for all paraffin-embedded tissue samples as well as cell lines. Polymerase chain reaction (PCR) was performed, followed by nucleotide sequencing using the iCycler (Bio-Rad, Hercules, CA, USA). Exons 9 and 20 of PIK3CA, exon 2 of KRAS, and exon 15 of BRAF were sequenced because together, mutational hot spots in these regions harbor nearly all published mutations (12–15). The primers for PCR and sequencing were manufactured by GeneLink, Inc., (Hawthorne, NY, USA), and their sequences were described in a previous report (16). The sequences were analyzed using the Lasergene program, DNASTAR, Inc., (Madison, WI, USA).

Cell growth assays

Cells were plated in 96-well plates at a density of 3,000 cells per well and treated with or without a potent PI3K or mTOR inhibitor. An MTT growth assay was performed to determine the number of cells (17). Potent inhibitors such as temsirolimus (Selleck Chemicals, Houston, TX, USA) and NVP-BEZ235 (Selleck Chemicals) were used to treat each of the cell lines at doses of 50, 500, 1,000, and 3,000 nM to inhibit PI3K/mTOR function, and cell viability was measured 98 h later. DMSO was used at an equal amount as a control. The data were expressed as percentage relative to the DMSO control. The mean and SD were obtained from three experiments.

Statistical analysis

Progression-free survival was calculated as the time between diagnosis and recurrence of disease, whereas overall survival was calculated as the time between diagnosis of disease and death. Kaplan-Meier curves were used to plot the survival data and log-rank tests were performed to determine the statistical significance of survival differences. Data were censored when patients were lost to follow-up. The χ2 test was used for comparisons of categorical data. Student's t-test (for comparison of two groups) or one-way analysis of variance followed by Tukey's post hoc test (for comparison of more than two groups) was used to evaluate numerical data.

Results

Identification of PIK3CA, KRAS, and BRAF mutations

Somatic mutations in PIK3CA were found in four (5.6%) of 71 cervical squamous cell carcinoma samples and three (5.7%) of 53 cervical adeno/adenosquamous cell carcinoma samples (Table I). Interestingly, mutations in adeno/adenosquamous carcinomas only occurred within exon 9 (E545A) (Table II; Fig. 1A) and no mutations were identified in the catalytic domain encoded by exon 20. In contrast, mutations in squamous cell carcinoma only occurred within exon 20 (Y1021N, A1035V, and T1052K) and no mutations were identified in the helical domain encoded by exon 9 (Table II; Fig. 1B-D). Somatic mutations in KRAS were identified in three (5.7%) of 53 cases and BRAF mutations were identified in one (2%) of 53 cases of cervical adeno/adenosquamous cell carcinoma (Table I). KRAS and BRAF mutations were not present in the same samples harboring PIK3CA mutations. No squamous cell carcinomas had detectable oncogenic mutations in KRAS (0%, of 71) or BRAF (0%, of 71).

Figure 1. PIK3CA mutations in cervical cancer. (A) Upper chromatogram sequence represent wild type and lower mutant sequence E545A (1634 A >C) in exon 9. (B) Upper-wild type PIK3CA gene and lower mutant sequence Y1021N (3061 T>A) in exon 20. (C) Upper-wild type PIK3CA gene in exon 20 and lower mutant sequence T1052K (3155 C>A). (D) Upper-wild type PIK3CA gene in exon 20 and lower mutant sequence A1035V (3104 C>T). The bottom arrow indicates the position of missense mutations in each case.

Table I. Frequency of PIK3CA, KRAS, BRAF and PIK3CA amplification in cervical carcinoma.

Table I.

Frequency of PIK3CA, KRAS, BRAF and PIK3CA amplification in cervical carcinoma.

Histological subtype	PIK3CA (%)	KRAS (%)	BRAF (%)	PIK3CA amplification (%)
SCC	4/17 (5.6)	0/71 (0.0)	0/71 (0.0)	1/71 (1.4)
AD/ASC	3/53 (5.7)	3/53 (5.7)	1/51 (2.0)	11/53 (20.7)

[i] PIK3CA, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit α; SCC, cervical squamous cell carcinoma; AD/ASC, cervical adeno/adenosquamous cell carcinoma.

Table II. Association between PIK3CA mutation, amplification and cervical carcinoma histological subtype.

Table II.

Association between PIK3CA mutation, amplification and cervical carcinoma histological subtype.

		PIK3CA mutation (exon 9)			PIK3CA mutation (exon 20)			PIK3CA amplification
Histological subtype	Patients	−	+	P-value	−	+	P-value	−	+	P-value
SCC	71	71	0	0.0424	67	4	0.848	70	1	0.0003
AD/ASC	53	50	3		53	0		42	11

[i] PIK3CA, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit α; SCC, cervical squamous cell carcinoma; AD/ASC, cervical adeno/adenosquamous cell carcinoma; -, negative; +, positive.

Frequency of PIK3CA gene amplification is higher in adeno/adenosquamous cell carcinomas than in squamous cell carcinomas

PIK3CA amplification was identified in one (1.4%) of 71 cervical squamous cell carcinomas. In contrast, PIK3CA amplification was identified in 11 (20.7%) of 53 adeno/adenosquamous cell carcinomas. Amplification of PIK3CA was more frequently found of adeno/adenosquamous carcinomas than in squamous cell carcinomas (P<0.0003, χ2-test; Table II; Fig. 2). Interestingly, amplification of PIK3CA was not found in the same patients harboring PIK3CA mutations.

Figure 2. Dual-color FISH was used to detect amplification of the PIK3CA gene amplification cervical carcinomas. (A) FISH analysis revealed a homogeneously stained region in the cervical carcinoma case with PIK3CA gene amplification (↔) as indicated by multiple green signals in each nucleus. (B) In contrast, another case contained signals for both PIK3CA (green) and reference probes (red) at a ratio of approximately 1:1 (→). FISH, fluorescence in situ hybridization; PIK3CA, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit α.

Prognostic effect of PIK3CA amplification

Next, we examined the prognostic ability of PIK3CA amplification. Kaplan-Meier curves were used to plot progression free survival and overall survival (Fig. 3), and we observed no significant relationship between PIK3CA amplification and progression free/overall survival (P=0.7576 and P=0.8859, respectively).

Figure 3. Kaplan-Meier survival analysis in 53 patients with cervical carcinoma according to PIK3CA amplification. PIK3CA amplification did not correlate with (A) progression-free survival and (B) overall survival. PIK3CA, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit α.

Association between PIK3CA mutational status and growth inhibition

Squamous cell carcinoma and adeno/adenosquamous cell lines were first analyzed for PIK3CA mutational status. Of nine cervical cancer cell lines, PIK3CA mutations were observed in four (ME180, TCS, HCS-2, and CASKI), but only in Exon 9. Next, we used the potent inhibitor temsirolimus (formerly known as CCI-799) and NVP-BEZ235 to examine the relationship between mutational status and growth inhibition. Growth of the PIK3CA mutation-containing cell lines was not inhibited by any of the inhibitors (Figs. 4 and 5). Treatment with PI3K/AKT/mTOR inhibitors failed to inhibit proliferation in all four cell lines harboring PIK3CA mutations.

Figure 4. Effect of temsirolimus on cell proliferation in cervical carcinoma cell lines. Cells were counted after 72 h of temsirolimus and DMSO (control) treatment. The presence of a mutation in PIK3CA was not related to sensitivity to growth inhibition by temsirolimus. PIK3CA, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit α; M: PIK3CA mutation containing cell line; WT: wild-type cell line.

Figure 5. Effect of NVP-BEZ235 on cell proliferation in cervical carcinoma cell lines. Cells were counted after 72 h of temsirolimus and DMSO (control) treatment. The presence of a mutation in PIK3CA was not related to sensitivity to growth inhibition by NVP-BEZ235. PIK3CA, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit α; M: PIK3CA mutation containing cell line; WT: wild-type cell line.

Discussion

PIK3CA mutations were found in only four of 71 (5.6%) cervical squamous cell carcinoma samples, and were identified in exon 20 (P=0.848). Similar mutations were present in three of 53 (5.7%) cervical adeno/adenosquamous cell carcinoma samples, and these were identified in exon 9 (P=0.042). No tumors harbored mutations in both the helical and kinase domain. These results indicated that mutations are more common in exon 9 for adeno/adenosquamous cell carcinoma and in exon 20 for cervical squamous carcinomas in Japanese cervical cancer patients. Our findings suggest that adeno/adenosquamous carcinomas could be distinguished from squamous cell carcinomas based on genetic alterations. Lou et al (9), found that PIK3CA mutations in the helical domain are significantly more prevalent in squamous cell carcinomas than in adenocarcinomas (P=0.017) in Latin American patients with cervical cancer, which is not consistent with the results of our study. Xiang et al (10), identified 105 cases (13.6%) with PIK3CA mutations from a cohort of Chinese patients with cervical carcinoma. The most common mutations were found in exon 9, at residues 545 and 542, in 59 and 32 samples, respectively. The H1047R mutation has been reported in four cases, and several rare non synonymous base substitutions were also identified in Chinese patients, some of which have been reported in the Catalog of Somatic Mutation in Cancer (COSMIC) database. Two mutations were identified in Chinese patients that were not previously reported. In addition, mutations in PIK3CA were found in 93 (15.3%) of the 606 patients with squamous cell carcinomas, nine (8.9%) of the 101 patients with adenocarcinomas, and one (2.3%) of the 44 patients with adenosquamous carcinoma. Therefore, PIK3CA mutations also occurred more frequently in squamous cell carcinomas than in non-squamous cell carcinomas in Chinese patients; moreover, PIK3CA mutations were mostly activating helical domain mutations, specifically E542K and E545K (10). The cancer Genome Atlas Research Network described the genomic and molecular characterization of cervical cancer and observed that most PIK3CA mutations occurred at the helical domain of exon 9 (E542K and E545K) and that PIK3CA (P=0.01) were differentially expressed between keratin-low and keratin-high squamous cluster gene family members (18). Comparing the results of the current study with those of previous studies, we hypothesized that the difference in prevalence is due to a difference in genetic background between the Japanese population and other ethnic groups.

PIK3CA gene amplification was also detected in this study and our results suggested that 11 of 53 (20.7%) samples were associated with PIK3CA gene amplification in adeno/adenosquamous cell carcinoma. In contrast, one of 71 (1.4%) cervical squamous cell carcinomas exhibited PIK3CA gene amplification. This event occurred at a significantly higher prevalence (P=0.0003) in adeno/adenosquamous cell carcinoma than in cervical squamous cell carcinomas. Our data suggest that cervical squamous cell carcinoma and adenocarcinoma have distinct gene expression profiles that might arise from different pathways involved in carcinogenesis. We previously found that Notch3 is significantly overexpressed in cervical squamous cell carcinomas compared to expression in cervical cancer adenocarcinomas (19). Wright et al (5), identified KRAS mutations only in adenocarcinomas (17.5% (AC) vs. 0% (SCC); P=0.01), which is similar to our results; in addition, a novel EGFR mutation was previously detected only in squamous cell carcinomas (0% (AC) vs. 7.5% (SCC); P=0.24). Taken together, our results and those of previous studies suggest that cervical squamous cell carcinoma and adenocarcinoma have distinct molecular profiles.

Activating mutations and amplification of PIK3CA, the gene that encodes the catalytic subunit of phosphatidylinositol 3-kinase (PI3K), have also been reported in 23–36% of cervical cancer specimens (20–22). Based on observational studies, activation of the PI3K pathway has been associated with higher rates of local recurrence after radiotherapy and decreased survival (22,23). Next, we examined the relationship between PIK3CA amplification and prognosis and found that PIK3CA amplification did not correlate with progression free survival and overall survival. Possible reasons for the different results in our study compared to those of others, including PIK3CA mutation rate in cervical cancer and the prognostic effect of PIK3CA amplification, might be different patient cohorts or sample sizes.

Mutations in PIK3CA are becoming a promising target for newly discovered anticancer drugs. Janku et al (24), observed that in patients with advanced breast, ovarian, endometrial, and cervical cancers, an association between PIK3CA mutations and positive response to PI3K/AKT/mTOR inhibitors was apparent. McIntyre et al (22), recently identified that among cervical cancer patients with PIK3CA mutations treated with radical chemoradiotherapy, there was a strong association with overall survival in FIGO stage IB/II but not stage III/IVA. In the present study, PI3K/AKT/mTOR inhibitors such as temsirolimus and NVP-BEZ235 were tested in cervical cancer cell lines; it was found that mutational status does not correlate with growth inhibition with any of the tested inhibitors. These observations suggest that PIK3CA inhibition might not represent a potential therapeutic option for the treatment of cervical cancers with associated mutations.

In conclusion, our data demonstrate that a proportion of Japanese cervical cancer patients harbor mutations in PIK3CA. Helical domain-encoding PIK3CA mutations are more frequent in adeno/adenosquamous cell carcinomas, whereas kinase domain-encoding PIK3CA mutations are more frequent in cervical squamous cell carcinomas. Our study also suggests that PIK3CA amplification occurs at a significantly higher rate (P=0.0003) in adeno/adenosquamous cell carcinoma than in cervical squamous cell carcinomas, and that PIK3CA amplification does not correlate with progression free survival or overall survival. The mutation status of PIK3CA also did not predict sensitivity to PI3K/AKT/mTOR inhibitors in cervical carcinoma cells in vitro. This result suggests that PIK3CA mutations might not be an important parameter for predicting treatment response in Japanese cervical cancer patients.

Acknowledgements

Not applicable.

Funding

The present study was supported by grants from the Ministry of Education, Culture, Sports, Science and Technology in Japan (grant no. 15K10717).

Availability of data and materials

The datasets used and/or analyzed during the present study are available from the corresponding author on responsible request.

Authors' contributions

SR and KeN drafted the manuscript. KoN, TI, MI, and TM performed data collection, analysis and interpretation of data. SR and KI conducted the experimental trials. YO, SN and NI performed pathological diagnosis. KeN participated in the design of the study. SK conceived the study, participated in its design and coordination, and drafted the manuscript. All authors read and approved the final manuscript.

Ethics approval and consent to participate

Ethical approval was obtained from the Ethics Committee of Shimane Medical University. All patients provided written informed consent for the procedure and study participation.

Patient consent for publication

All patients approved their data for publication.

Competing interests

The authors declare that they have no competing interests.

References