Introduction
The current understanding of human malignancy is that it mainly arises due to the accumulation of multiple genetic alterations, transforming normal cells into malignant cells (1,2). Of these genetic alterations, a myriad of genomic mutation data derived from a high-throughput DNA sequencing technique provided a unique opportunity to profile the mutation spectra underlying human cancers and a large number of significant functional mutations in multiple genes were identified in diverse types of cancer (1,3,4). These genes can be defined as oncogenes or tumor suppressor genes and are being used as molecular markers for diagnosis, staging and prognosis of human cancers (5,6).
Ovarian carcinoma constitutes a heterogeneous group of malignancies with significantly different clinical expression, pathological characteristics and genetic etiology (7,8). However, the majority of ovarian carcinomas shared certain common genetic alterations, such as frequent tumor protein p53 (TP53) and PIK3CA, catalytic subunit α mutations (9,10), and patients also exhibited subtype-specific mutations (11–13), which are possibly essential for the differential clinical expression and molecular-targeted therapy in ovarian carcinomas (14,15). These observations emphasized the requirement to identify novel subtype-specific molecular genetic aberrations in ovarian carcinomas.
Recently, large-scale sequencing has identified frequent mutations of the ribonuclease type III (DICER1) gene in Sertoli-Leydig cell tumors of the ovary (3), CCCTC-binding factor (CTCF) gene in transient abnormal myelopoiesis (16) and endometrial cancer (17), ribosomal protein L22 (RPL22) gene in endometrial cancer (18), DNA (cytosine-5-)-methyltransferase 3α (DNMT3A) gene in hematological malignancies (4), the transformation/transcription domain-associated protein (TRRAP) gene in melanoma (19) and isocitrate dehydrogenase 1 and 2 (IDH1 and IDH2) genes in gliomas (1,20) and acute myeloid leukemia (AML) (21), respectively. Some of these mutations were closely associated with cancer progression (22) and prognosis (23,24).
Thus far, the mutation statuses of DICER1, CTCF, RPL22, DNMT3A, TRRAP, IDH1 and IDH2 mutational hotspots in ovarian carcinomas remain largely unknown. One critical concern in cancer genetics is whether those cancer-associated mutations identified in one type of cancer are also common in other types of cancer. Therefore, a cohort of 251 Chinese patients with distinct subtypes of ovarian carcinomas was recruited in the present study to examine whether the hotspot mutations in these genes also existed in these samples.
Results and Discussion
The available clinical data of these patients are as described previously (25,26). In the present study, a total of 251 Chinese samples with distinct subtypes of ovarian carcinoma were screened for the presence of potential hotspot mutations in the DICER1, CTCF, RPL22, DNMT3A, TRRAP, IDH1 and IDH2 genes. However, no mutations in these genes were detected in the 251 samples (Table I and Fig. 1).
Previous studies have found frequent DICER1 p.E1705-D1709 and p.D1810-E1813 mutations in Sertoli-Leydig cell tumors (3,28). However, no DICER1 mutations were detected in the two patients with Sertoli-Leydig cell tumors. Therefore, it can be speculated that this discrepancy may be caused mainly by the small sample size of the Sertoli-Leydig cell tumors analyzed in the present study. In addition, DICER1 mutations were not identified in other subtypes of ovarian carcinomas in the samples, which is consistent with previous large-scale sequencing results in which the DICER1 hotspot mutations were absent in 12 mucinous (29) or 316 serous ovarian carcinomas (9). Collectively, these results indicated that the DICER1 hotspot mutations may not be actively involved in the pathogenesis of ovarian carcinoma, except for Sertoli-Leydig cell tumors.
CTCF p.T204fs* and RPL22 c.43delA mutations have been observed frequently in endometrial carcinoma in previously studies (17,18). Considering the fact that ovarian carcinoma have certain overlapped genetic aberrations with endometrial cancer, such as frequent TP53 (9,30) and polymerase (DNA directed) ε, catalytic subunit (POLE1) mutations (26,30), we hypothesized that ovarian carcinomas may also harbor these mutations. However, neither CTCF p.T204fs* nor RPL22 c.43delA mutations were identified in the samples in the present study. The absence of the CTCF and RPL22 mutations in ovarian cancer in a previous study (29) and the present study suggested that the CTCF and RPL22 hotspot mutations may play an extremely limited role in the pathogenesis of ovarian cancer.
Prevalent TRRAP p.S722 mutation was initially identified in melanomas in a whole-exome sequencing study (19). Subsequent extended studies failed to identify these mutations in thyroid cancer (31) or splenic marginal zone lymphoma (32). In the present study, no TRRAP p.S722 mutations were detected in our ovarian cancer patients with distinct subtypes. Also, TRRAP p.S722 mutations were not found in 12 mucinous (29) or 316 serous ovarian carcinomas (9). These negative results led us to speculate that TRRAP p.S722 mutations may not play a crucial role in the malignant transformation of ovarian carcinoma.
DNMT3A p.R882 mutations were identified almost exclusively in hematological malignancies, including AML (33), acute lymphoblastic leukemia (34) and myelodysplastic syndromes (35), and are generally infrequent or absent in some solid tumors (9,29,36). DNMT3A p.R882 mutations were not detected in the 251 samples with distinct subtypes of ovarian carcinoma. Similarly, whole-exome sequencing studies suggested that DNMT3A p.R882 mutations were absent in 12 mucinous (29) or 316 serous ovarian carcinomas (9). Taken together, the absence of DNMT3A p.R882 mutations in ovarian carcinoma analyzed in the present study and in previous studies (9,29) indicated that DNMT3A p.R882 mutations may be infrequent in ovarian carcinoma.
Frequent IDH1 p.R132, and IDH2 p.R140 and p.R172 mutations were identified in the central nervous system tumors and AML (1,20,27). However, no IDH1 or IDH2 mutations were detected in the present samples. Similar results were observed in previous studies in which IDH1 p.R132 mutations were not detected in 168 ovarian carcinomas or 8 ovarian cancer cell lines (20,37–39). In addition, IDH1 and IDH2 hotspot mutations were also not identified in 12 mucinous (29) or 316 serous ovarian carcinomas (9). These combined results suggested that IDH1 and IDH2 potential hotspot mutations may not be common in patients with ovarian carcinoma.
Among these patients, the POLE1 mutation has been previously found to be frequent in 37 ovarian endometrioid carcinomas (26), whereas ring finger protein 43 (RNF43) mutations were recurrent in 15 mucinous ovarian carcinomas (25) (Table I). In the present study, neither endometrioid nor mucinous ovarian carcinomas were detected to harbor DICER1, CTCF, RPL22, DNMT3A, TRRAP, IDH1 and IDH2 hotspot mutations. These results suggested that these potential hotspot mutations observed in other (sub)types of cancer may not play synergistic roles with POLE1 or RNF43 mutations in the carcinogenesis of endometrioid or mucinous ovarian carcinomas, respectively.
The main limitation of the present study was that only short DNA fragments spanning the potential hotspot mutations were screened in the seven genes, and therefore, there is a possibility that mutations in other residues of these genes may exist in these samples. However, due to the shortage of DNA amounts, this hypothesis was not tested.
In conclusion, DICER1, CTCF, RPL22, DNMT3A, TRRAP, IDH1 and IDH2 hotspot mutations were not identified in 251 Chinese patients with diverse subtypes of ovarian carcinoma. These results were generally consistent with previous studies and these combined results indicated that the hotspot mutations in these genes may not be actively involved in the carcinogenesis of Chinese patients with ovarian carcinoma, except for DICER1 mutations in Sertoli-Leydig cell tumors.