PAX8 is a potential marker for the diagnosis of primary epithelial ovarian cancer.

Ovarian cancer (OC) is the most lethal of all types of gynecological malignancy and the third leading cause of cancer mortality among females worldwide ( 1 ). Due to the lack of specific biomarkers for the diagnosis, more than half of patients are found to be in the advanced stage of OC at the time of initial diagnosis ( 2 , 3 ). Among the multiple types of ovarian cancer that exist, primary epithelial ovarian cancer (PEOC) accounts for 85–90% in total ( 4 ). Another type, metastatic ovarian carcinoma (MOC), can form from gastrointestinal tumors and breast cancers with peritoneal dissemination and/or lymphatic metastasis ( 5 – 7 ). The clinical use of cancer antigen (CA)-125, estrogen receptor (ER) and progesterone receptor (PR) for the diagnosis of OC is currently not effective due to low sensitivity and specificity ( 8 , 9 ). Therefore, the screening and identification of a specific biomarker for OC diagnosis is important. In a previous study, paired-box 8 (PAX8) has been suggested to be a sensitive marker for tumors of the thyroid, kidney and thymus, and tumors derived from the Müllerian pipe organ ( 10 – 12 ). It was also reported that PAX8 is widely expressed in various types of ovarian tumors, particularly in serous tumors such as ovarian endometrioid carcinoma and clear cell carcinoma ( 11 ). Through the collating, summarizing and long-term follow-up inspection of preliminary clinical data on ovarian cancer, the present study mainly discusses the value of PAX8 in the differential diagnosis of primary and metastatic ovarian carcinoma, and analyzes the association between the expression of PAX8, clinicopathological features of primary ovarian carcinoma and prognosis for patients with ovarian cancer.

There were 44 patients with ovarian serous carcinoma (73.3%), 9 patients with mucinous carcinoma (15%), 2 patients with endometrioid carcinoma (3.3%) and 5 patients with clear cell carcinoma (8.3%) within the total 60 patients with PEOC ( Fig. 1 ). The IHC staining of PAX8 in PEOC had an overall positive rate of 92% (57/60). By contrast, all MOC samples showed negative staining. Statistical analysis showed a significant difference between the PEOC and MOC groups (P<0.001). However, no significant difference was observed between the PEOC groups with different pathological types (P=0.871). The positive rate of PAX8 in 20 patients with ovarian benign tumor was 85% (17/20), showing no significant statistical difference in comparison with the PEOC group (P=0.761). In addition, no significant difference with respect to the positive expression of PAX8 was found among ovarian benign tumors with different pathologies (P=0.533). All data shown in Table I . Based on the aforementioned categorization and statistical analysis, the present study revealed that PAX8 expression was associated with the degree of cancer cell differentiation and International Federation of Gynaecological Oncologists (FIGO) stage ( 3 ). The higher the PAX8 expression, the lower the cancer cell was differentiated (P=0.033) and the more advanced the FIGO stage was (P=0.003; Table II ). There was no significant difference demonstrated between the expression levels of PAX8 in PEOC (P=0.574) of different pathological types ( Table II ). Among the group of 60 patients with primary ovarian cancer, the 1-year survival rate was 98%, the 2-year survival rate was 39% and the 3-year survival rate was 13%. According to the aforementioned categorization based on the PAX8 staining intensity ( Table II ), with respect to the low PAX8 expression groups (negative, 1+ and 2+; 17 patients in total), the survival rates subsequent to 1, 2 and 3 years were 100, 60.1 and 26.3%, respectively. With respect to the high PAX8 groups (3+ and 4+; 43 patients in total), the survival rates subsequent to 1, 2 and 3 years were 97.7, 26.7 and 6.7%, respectively. The postoperative survival rate of patients with ovarian cancer in the PAX8 low-expression groups was significantly longer than that of the PAX8 high-expression groups. The difference was considered to be statistically significant (χ 2 =6.820, P=0.009; Table III ; Fig. 2 ).

PAX8 belongs to the paired-box (PAX) gene family. The PAX protein family (PAX l-9) has a highly conserved region formed of 128 amino acids. The paired box gene family is able to recognize the specific DNA binding site of the target genes to regulate gene expression ( 13 ). PAX8 proteins are composed of 450 amino acids located in the 2q12-14 section of encoding genes with a molecular weight of 48 kb ( 14 ). PAX8 was first reported to be involved in thyroid gland development in mice ( 15 , 16 ). During embryonic development, PAX8 performs a key role in the development of organs such as the thyroid, kidney, part of the central nervous system, inner ear, eye and Wolffian and Müllerian ducts ( 17 ). Furthermore, PAX8 performs a role in maintaining the normal function of cells subsequent to the birth of the fetus. According to several studies, PAX8 can be a sensitive marker for thyroid, kidney, Müllerian duct and thymus tumors ( 10 – 12 ). The female reproductive tract originates from the fetal Müllerian system and PAX8 is involved in the regulation of the formation of the tract ( 17 ). Female mice lacking PAX8 expression cannot form the genital tract, leading to sterility ( 18 ). For humans, PAX8 is expressed in the tissues and organs, including in secretory cells, derived from the Müllerian system. Additionally, PAX8 expression is found in tumors derived from the epithelial cells of the Müllerian system ( 12 , 19 ). Previous studies have revealed that PAX8 is widely expressed in various types of ovarian tumor, particularly in serous tumors such as ovarian endometrioid carcinoma and clear cell carcinoma ( 11 , 12 ). The majority of serous ovarian cancer cells are considered to originate from the normal secretory cells of the fallopian tube in which PAX8 is highly expressed. Mucinous carcinoma of the Müllerian system is distinguished from metastatic cancer of the digestive tract, and cytokeratin (CK)-7 and CK20 are found to perform an important role in the differentiation process ( 20 ). However, in some patients it is difficult to judge the cancer source based on CK7/20 positivity or negativity ( 20 ). In certain types of mucinous carcinoma in the female reproductive tract, PAX8 may be partially expressed ( 21 ). However, PAX8 is not expressed in carcinoma of the stomach, colon, bile duct, duodenal ampulla, hepatocellular or pancreatic acinar cell carcinoma, and is rarely expressed in adenocarcinoma, pancreas-pseudopapillary tumors of the pancreas or esophageal adenocarcinoma ( 21 ). Where PAX8 is expressed, metastatic tumors of the digestive tract may be excluded, so the source of PAX8 may be considered to derive from the Müllerian duct. Similarly, PAX8 can be used for the clinical differential diagnosis of breast and ovarian cancer, ovarian sex cord-stromal tumor and malignant mesothelioma ( 22 , 23 ). In the present study, PAX8 was found to be highly expressed in PEOC, but not in MOC, suggesting that PAX8 exhibits a high enough sensitivity and specificity to be used in clinical practice. In addition, the present study found that PAX8 expression levels are associated with cancer cell differentiation degree and FIGO stage. Survival rate analysis demonstrated that patients with high levels of PAX8 expression had a poor prognosis. The results of the present suggest that there may be a link between PAX8 and the progression of ovarian cancer. In conclusion, PAX8 IHC detection exhibits high sensitivity and specificity in terms of differentiating between PEOC and MOC. The level of PAX8 expression is associated with the differentiation degree of ovarian tumor cells and FIGO staging, which can help to determine prognosis. However, determining PAX8 expression cannot be used to discriminate between benign and malignant ovarian tumors.

The present study was approved by the hospital Review Board and the Ethics Committee of the Medical Faculty at the Shanghai Jiao Tong University (Shanghai, China). All patients provided written informed consent of using their tumor tissues prior to the beginning of the study. In total, 60 PEOC patients enrolled in the Gynecology and Obstetrics Department of the Shanghai Ninth People's Hospital (Shanghai, China), whom underwent radical resection between January 2008 and December 2012, were included in the present study. The age of patients ranged between 39 and 68 years, with a median age of 52. The group included 44 patients with ovarian serous carcinoma, 9 patients with mucinous carcinoma, 2 patients with endometrial carcinoma and 5 patients with clear cell carcinoma. The control group included 20 patients with benign ovary tumors and 10 patients with MOC. In the MOC subgroup, the MOC had metastasized in 5 patients from gastric cancer, from colon cancer in 3 patients and from breast cancer in 2 patients. All clinicopathological profiles were evaluated in accordance with the criteria of the National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology (Fort Washington, PA, USA) ( 3 ). None of the patients received preoperative radiotherapy or chemotherapy prior to admission to hospital, confirmed by pathology following the surgery, and all were treated with platinum-based chemotherapy in combination with taxol subsequent to the surgery. The last follow-up was on December 31, 2014, with an average follow-up time of 20.58±5.71 months. During the follow-up period, 45 patients succumbed to recurrence or metastasis. The remaining 15 patients had censored data (among which 3 succumbed to other diseases, 2 were lost to follow-up and 10 are still alive). Surviving patients had a minimum follow-up period of 16 months, with the longest being 36 months. Surgically-obtained tumor tissue specimens were prepared in the form of 4 µm thick sections, prepared with hematoxylin and eosin staining and IHC staining, respectively. The avidin-biotin complex (ABC) method of IHC staining was adopted. The sections were treated by normal dewaxing, graded ethanol hydration, 3% H 2 O 2 closed endogenous peroxidase, microwave heating antigen retrieval and PBS rinsing. The primary antibody used was an anti-PAX8 rabbit anti-human monoclonal antibody (dilution 1:200; #ab189249; Abcam, Cambridge, UK) and the biotinylated secondary antibody was a goat anti-rabbit antibody (dilution, 1:200; #ab7010; Abcam). ABC reagent was used to visualize the staining (ABC kit; Santa Cruz Biotechnology, Inc., Dallas, TX, USA) and samples were subsequently stained with 3,3′-diaminobenzidine. Next, the samples were re-stained with hematoxylin and were resin-sealed. A PAX8 positive result was taken to be the nucleus appearing yellowish-brown or brown. For the negative control, PBS was used to replace the primary antibody. According to the intensity and range of PAX8 expression, the present study divided PEOC with PAX8 expression into 5 categories: No staining was assigned as negative ( Fig. 1A and B ); 1–24% staining as 1+ ( Fig. 1C and D ); 25–50% staining as 2+ ( Fig. 1E and F ); 51–75% staining as 3+ ( Fig. 1G and H ); >75% staining as 4+ ( Fig. 1I and J ). The categorization criteria were set as follows: Each slice was observed in 5 independent images, then 100 cells were counted per image and the percentage of stained cells in each slice was calculated. The aforementioned results were obtained via a double-blind method. Each section was examined by 2 pathologists and the patients with consistent results were enrolled in the present study. SPSS 13.0 (SPSS, Inc., Chicago, IL, USA) was used for statistical analysis. The association between PAX8 and clinicopathological characteristics was tested using a χ 2 test for comparisons between groups, and a Mann-Whitney U test and Kruskal-Wallis H test for comparisons within the same population. For survival rate analysis, a Kaplan-Meier test was used to compare the differences between groups. P<0.05 was considered to indicate a statistically significant difference.