Estrogen receptor 1 mutations in 260 cervical cancer samples from Chinese patients

Cervical cancer is one of the leading causes of cancer-associated mortality among females worldwide ( 1 , 2 ). Despite the existence of a safe vaccine against human papillomavirus (HPV) and an early screening test for cervical cancer, the frequency of females undergoing the test is not sufficient ( 3 , 4 ). Furthermore, an insufficient number of HPV vaccines are available in countries that exhibit a high prevalence rate of cervical cancer and the worldwide incidence rate of cervical cancer has not significantly decreased ( 5 , 6 ). Therefore, there is a requirement to explore the underlying molecular mechanisms of cervical cancer. Estrogen receptor 1 (ESR1) is a transcription factor that can be activated by estrogen and other growth factors in a ligand-dependent manner; activated ESR1 dimerizes and regulates the transcription of numerous target genes ( 7 , 8 ). Previous comprehensive genomic studies have revealed frequent ESR1 mutations in metastatic types of breast cancer ( 9 , 10 ). Subsequent studies also confirmed these observations ( 11 – 13 ) and demonstrated that ESR1-mutated samples often exhibit increased resistance to aromatase inhibitors ( 14 ). In addition, ESR1 mutations have been identified in endometrial ( 15 ) and colorectal cancer ( 16 ) at low frequencies. However, to the best of our knowledge, it remains unknown whether ESR1 mutations exist in other types of cancer, including cervical cancer. Cervical cancer shares certain genetic aberrations with breast cancer, including frequent mutations in the PIK3CA, TP53, PTEN and ARID1A genes ( 17 – 21 ), and the development of cervical and breast cancer can be affected by estrogen action ( 22 – 25 ). These similarities suggest that cervical cancer may also harbor ESR1 mutations. To test this hypothesis, a total of 260 samples from Chinese patients with distinct subtypes of cervical cancer were investigated for the presence of ESR1 mutations. A total of three heterozygous missense ESR1 mutations were identified in 207 samples of cervical squamous cell carcinoma (3/207, 1.4%), whereas no mutations were detected in the 27 adenosquamous carcinoma or 26 adenocarcinomas samples. The identified ESR1 mutations could have predictive values and may provide insights into the diagnosis and molecular therapy of cervical cancer.

A total of three heterozygous missense ESR1 mutations, p.K303R (c.908A>G), p.T311M (c.932C>T) and p.Y537C (c.1610A>G), were identified from 207 cervical squamous cell carcinoma samples (3/207, 1.4%), while no mutations were detected in the adenosquamous carcinoma and adenocarcinoma samples. The mutations were absent in the paired non-cancerous tissues and were therefore considered to be somatic ( Fig. 1 ). The K303R and T311M mutations are located in the ‘hingeregion’ and the Y537C mutation is located in the ‘ligand-binding domain’ ( 9 , 10 ). Of the three individuals with ESR1 mutations, two were further diagnosed with uterine fibroid and one with ovarian endometriosis. Two publicly available bioinformatics programs, MutationTaster and PolyPhen-2, were used to predict the potential functional significance of the ESR1 mutations. The predictions by MutationTaster for the three ESR1 mutations (p.K303R, p.T311M and p.Y537C) were ‘disease causing’ and ‘protein features (might be) affected’, while PolyPhen-2 predicted these mutations to be ‘probably damaging’ (p.T311M and p.Y537C) or ‘possibly damaging’ (p.K303R), with a prediction score of >0.90. Furthermore, the Y537C (c.1610A>G) and K303R (c.908A>G) mutations were not identified in the 1,000Genomes ( https://www.ncbi.nlm.nih.gov/variation/tools/1000genomes/ ) ( 32 ) or the Exome Aggregation Consortium (EXAC; http://exac.broadinstitute.org/ ) ( 33 ) databases, while the p.T311M (c.932C>T) mutation was identified in the general population with an extremely low frequency (1/121,362) in the EXAC database. The results of evolutionary conservation analysis demonstrated that the three ESR1 mutations were associated with highly conserved amino acid changes among 18 vertebrate species, ranging from Homo sapiens to Xenopus tropicalis ( Fig. 2 ). The protein structural prediction results suggested that the three ESR1 mutations may induce the structural changes in the side chain of ESR1 protein ( Fig. 3 ); results that were consistent with the prediction results by MutationTaster.

Previous studies have identified prevalent ESR1 mutations in breast cancer ( 9 , 10 ); however, it remains largely unknown whether ESR1 mutations exist in other types of cancer, including cervical cancer. In the present study, a total of 260 samples from Chinese patients with distinct subtypes of cervical cancer were tested for the presence of ESR1 mutations. In total, three missense somatic mutations in ESR1, p.K303R (c.908A>G), p.T311M (c.932C>T) and p.Y537C (c.1610A>G), were identified from 207 squamous cell carcinomas (3/207, 1.4%), which were not present in other cervical cancer subtypes. Evolutionary conservation analysis demonstrated that the three ESR1 mutations were associated with highly conserved amino acid, which may describe the potential to cause protein structural changes. In silico analysis suggested that these mutations may be pathogenic. Furthermore, the three ESR1 mutations have previously been identified in other types of cancer. The p.K303R (c.908A>G) mutation has been observed in 206/6,556 samples of breast cancer ( 24 – 37 ). The p.T311M (c.932C>T) mutation has been identified in 2/2,218 colorectal cancer samples ( 38 ) and 1/2,105 liver cancer samples ( 38 ). In addition, the p.Y537C (c.1610A>G) mutation has been detected in 10/6,556 breast cancer samples ( 11 , 39 – 41 ). Similarly, a previous comprehensive study identified ESR1 somatic mutations p.K206R (c.617A>G) and p.L372L (c.1116C>G) in 2/306 (0.7%) cervical squamous cell carcinoma samples ( http://cancer.sanger.ac.uk/cosmic ). However, a previous genomic analysis of cervical cancer failed to detect any ESR1 mutations in 79 squamous cell carcinoma samples ( 17 ). Considering the low frequency of ESR1 mutation in cervical squamous cell carcinoma, it is suggested that the small sample size analyzed in this previous study may have caused the inconsistency in the results ( 17 ). In combination, both previous studies and the present study suggest that ESR1 mutations may participate in the carcinogenesis of cervical cancer, albeit at a low frequency. A number of previous functional assays for the identified ESR1 mutations, including p.K303R ( 42 ) and p.Y537C ( 9 , 43 ), demonstrated that these mutations are associated with acquired endocrine resistance in hormonal therapy in breast cancer ( 7 , 8 , 40 , 41 ). Therefore, it is proposed that the ESR1 mutations identified in cervical cancer in the present study may further cause acquired endocrine resistance to hormonal therapy in breast cancer. Antihormonal agents have recently been used to improve effects of chemo- and radiotherapy in cervical cancer ( 44 ). However, due to the potential acquired endocrine resistance in cervical cancer samples with ESR1 mutations, antihormonal agents should be used with caution during chemo-and radiotherapy. The three ESR1 mutations were not detected in the 27 adenosquamous carcinoma or the 26 adenocarcinoma samples of the present study, which is consistent with a prior genomic analysis of cervical cancer with distinct subtypes, in which no ESR1 mutations were detected in either 24 adenocarcinoma or 7 adenosquamous carcinoma samples ( 17 ). In summary, these results suggest that the ESR1 mutations may not be positively involved in the pathogenesis of cervical adenocarcinoma and adenosquamous carcinoma. However, the absence of ESR1 mutations in patients with adenosquamous carcinoma and adenocarcinoma may be due to the small sample sizes analyzed in the present study. In conclusion, the current study identified three potentially pathogenic ESR1 mutations in cervical squamous cell carcinoma samples from Chinese patients, which were not observed in other subtypes. These results, together with numerous previous studies, suggested that ESR1 mutations may be involved in the carcinogenesis of squamous cell carcinoma, but not in other subtypes of cervical cancer.

The sample cohort has been previously described ( 26 ). Briefly, a total of 260 FFPE cancerous and paired adjacent non-cancerous tissue sections (10 µm), including squamous cell carcinoma (n=207), adenosquamous carcinoma (n=27) and adenocarcinoma (n=26) tissues, were fixed in 10% neutral buffered formalin for 36 h at room temperature, and collected from the archives of the Department of Pathology at the Jiangxi Provincial Maternal and Child Health Hospital (Nanchang, China) between July 2008 and August 2013. The median age of patients was 43 years old (range, 22–74 years). The present study was approved by the Institutional Review Board of Jiangxi Provincial Maternal and Child Health Hospital (Nanchang, China), and performed according to the Declaration of Helsinki. All patients provided written informed consent prior to the study. Genomic DNA was isolated using QIAamp DNA FFPE Tissue kit (Qiagen GmbH, Hilden, Germany). The entire coding exons and the corresponding intron/exon boundaries of the ESR1 gene were amplified with a set of primer pairs ( Table I ). The polymerase chain reaction (PCR) amplification was performed in a 50 µl reaction volume containing 0.2 µM deoxyribonucleotide triphosphate, 5 µl 10X PCR buffer, 0.5 U rTaq DNA Polymerase (Takara Biotechnology Co., Ltd., Dalian, China) and 200 ng genomic DNA. PCR amplifications were performed in a Bio-Rad iCycler thermal cycler (Bio-Rad Laboratories, Inc., Hercules, CA, USA) with the following conditions: 94°C for 3 min, 35 cycles of 94°C for 30 sec, 50–58°C for 30 sec and 72°C for 30 sec, and a final extension step at 72°C for 10 min. The amplicons were sequenced bidirectionally on an ABI 3730 Genetic Analyzer (Thermo Fisher Scientific, Inc., Waltham, MA, USA) and the sequencing data was aligned against the corresponding genomic sequence (ESR1, NM_000125 ) in the National Center for Biotechnology Information (NCBI) database ( www.ncbi.nlm.nih.gov ). The identified somatic mutations were confirmed by sequencing the paired, adjacent non-cancerous tissues. Two online prediction programs, MutationTaster ( http://www.mutationtaster.org ) ( 27 ) and Polymorphism Phenotyping v2 (PolyPhen-2; http://genetics.bwh.harvard.edu/pph2 ) ( 28 ), were used to predict the associations between the identified ESR1 mutations and disease occurrence. These programs assessed the identified ESR1 mutations as either ‘benign’ or ‘pathogenic’, according to the automatically predicted score. The evolutionary conservation of the mutated amino acids of ESR1 was analyzed in a total of 18 vertebrate species retrieved from the NCBI database, including Homo sapiens ( NP_000116 ), Pan troglodytes ( XP_009450519 ), Mus musculus ( NP_001289460 ), Rattus norvegicus ( NP_036821 ), Ovis aries ( NP_000116 ), Bos taurus ( NP_001001443 ), Gallus gallus ( NP_990514 ), Sus scrofa ( NP_999385 ), Canis lupus familiaris ( NP_001273887 ), Equus caballus ( NP_001075241 ), Tupaia chinensis ( NP_001304001 ), Mustela putorius furo ( XP_004753629 ), Oryctolagus cuniculus ( XP_008261925 ), Pongo abelii ( XP_002817538 ), Coturnix japonica ( NP_001310118 ), Alligator sinensis ( XP_014375965 ), Ceratotherium simum simum ( NP_001266182 ) and Xenopus tropicalis ( NP_988866 ). The Molecular Evolutionary Genetics Analysis 4.0 software ( 29 ) was used for multiple sequence alignment. DeepView Swiss-PdbViewer 4.0 software ( 30 ) was used to predict the potential protein structural changes for the identified ESR1 mutations. An available 3D protein structure of human ESR1 (protein data bank code, 2OCF) ( 31 ) was retrieved from the SWISS-MODEL repository in the ExPasy web interface ( http://www.expasy.org ).