At present typing of ovarian tumors is of limited therapeutic significance because of management more dependent on stage and grade of tumor, however, new era of targeted therapy according to histological type of ovarian cancer to reduce chemoresistance is in the beginning. Reproducibility of typing of epithelial ovarian cancer as borderline or malignant has significant interobserver variability for especially poorly differentiated tumors. There is a need to develop panels of immunohistochemical staining markers suitable to rule out histopathological types of ovarian tumors. Data of studies investigating the reliability of several immonohistochemical markers to differentiate ovarian tumors can lead to development of diagnostic panels successful to confirm histological type.
Tumor suppressor genes, protooncogenes, and apoptosis-related genes have been implicated in the regulation of ovarian carcinoma. Serous tumors and some mucinous tumors may show BRAF and K-RAS mutations [6–8, 14, 25–32]. The tumor suppressor gene p53 is involved in the control of cell proliferation, particularly in stressed cells. p 53 gene mutations are the most frequent genetic event found in various types of cancers [33, 34], and is the most common studied tumor suppressor gene in ovarian carcinomas. High-grade serous tumors tended to be p53 positive and p53 positivity is related to the survival rate. Although p53 is important pathway for serous cystadenocarcinoma, prognostic value of p53 is still controversial. Despite some studies have clear evidence that p53 positivity is related with grade, stage, and survival rate, some of the studies have not found any relationship [17, 18, 35, 36]. This apparent discrepancy for p53 in ovarian carcinoma can be explained by the heterogeneity of the tumors, differences in immunohistochemistry methods and examination of limited tumor cases in the studies. Fauvet et al.  have found significant difference in p53 expression between benign and borderline ovarian tumors, however, significant difference was not found between borderline and malignant tumors. Furthermore, they have not found any differences in p53 expression of serous and mucinous tumors. In our study, we found higher p53 staining in borderline and malignant tumors compared to the benign ones. However, we found comparable level of p53 staining between borderline and malignant tumors. In addition, serous tumors showed higher p53 staining than mucinous ones. Thus, our results support that especially in serous tumors, p53 expression has a role a place in the typing of ovarian tumors, and it might be helpful in differentiating benign and borderline ovarian tumors. For differentiation of borderline and malignant tumors, p53 may be used as a predictor for malignant tumors (p < 0.05).
P21, a p53 inducible gene, encodes an inhibitor of cyclin dependent kinases involved in G1 arrest. Studies about p21 expression in ovarian carcinoma are controversial [10, 29, 36]. Quellet et al.  have studied p21 expression in benign, borderline, and malignant ovarian serous tumors. In this study in serous tumors, they have found lower p21 expression than benign and borderline tumors. According to this result, they thought that lower expression of p21 might be the sign of aggressive potential of the tumor or might be poor prognostic parameter . However, in another study, p21 expression was found lower in benign tumors than borderline and malignant tumors . Fauvet et al.  have reported that borderline tumors have higher p21 expression than benign tumors. In that study, they suggested that overexpression of p21 is specific for serous tumors. In our study, we found higher p21 staining in borderline and malignant serous tumors than benign serous ones. According to the result of the study of Fauvet et al. , in our study, p21 staining in serous tumors was higher than mucinous tumors. Our data involving both p21 and p53 expression in ovarian tumors were similar with that study. Thus, we thought that both p21 and p53 might play different role in the development of serous ovarian tumors.
A proapoptotic gene bax induction by p53 is necessary to inhibit tumor growth, and that the contribution of bax to p53-mediated apoptosis is cell type dependent . Fauvet et al.  have reported that there were not statistically significant differences in bax expression in benign, borderline, and malignant ovarian serous and mucinous tumors. Schuyer et al.  have found that low bax expression might be poor prognostic parameter. Skirnisdöttir et al.  have found the lowest rate of positive bax staining in mucinous tumors. In a study, a correlation between high bax levels and improved clinical outcome was also found . In our study, there was higher bax positive staining in borderline and malignant ovarian tumors compared to the benign ones.
Regulators of apoptosis may be cell specific and the first antiapoptotic gene identified was bcl-2. In a reported study, bcl-2 has showed higher expression in benign and borderline tumors than malignant tumors . Fauvet et al.  have reported that borderline tumors have higher bcl-2 expression than benign and malignant tumors. In the same study, they have showed that although there was not difference in mucinous tumors, serous tumors have significant difference in bcl-2 positive staining. Thus, they have suggested that bcl-2 could be specific for serous tumors and might be related with ovarian serous tumor pathogenesis. In our study, bcl-2 positive staining was not found in none of the mucinous tumors. Although benign, borderline, and malignant serous tumors showed positive bcl–2 staining, we found no difference reaching statistical significance, moreover, serous and mucinous tumors did not differ in bcl–2 staining.
The c-kit proto-oncogene encodes a tyrosine kinase receptor and is expressed in various normal and tumor tissues. In a reported study, c-kit expression has been found 4.5% in low-grade serous carcinoma and 29.7% in high-grade carcinoma . Parrott et al.  have showed that in advanced ovarian tumors, c-kit expression was high and they have suggested the importance of c-kit expression in the progression of ovarian carcinomas. In a reported study, it has been found that mucinous tumors have higher c-kit expression than serous tumors . In addition, in another study, the loss of c-kit expression was found to be a poor prognostic factor and it was reported that c-kit might play role in early stage of ovarian carcinogenesis [8, 13, 14]. In this study, in accordance with those findings, the immunohistochemical staining with c-kit was remarkable in the serous ovarian tumors compared to the mucinous ones.
Telomerase activity is involved in the maintenance of telomere length and is thought to be required for oncogenesis. Including ovarian carcinoma, in most of the tumors high levels of telomerase activity was reported [12, 14]. It was demonstrated that there was strong telomerase activity in borderline and malignant tumors, however, in few or none of the benign tumors and normal ovaries [11, 12, 14, 40, 41]. In accordance with the reported cases, we found somewhat higher telomerase staining in borderline and malignant serous and mucinous tumors although it was not reached statistical significance.
Metallothioneins are involved in cell proliferation, growth, and differentiation. The significance of metallothionein expression in ovarian cancers is inadequately documented. Mccluggage et al.  have found metallothionein expression in 56% of malignant ovarian carcinoma and in 2% of benign tumors. Because of this finding, they have suggested that metallothionein may play role in ovarian tumorogenesis . In reported studies, high expression of metallothionein was found in malignant ovarian serous and mucinous tumors compared with the benign tumors. Thus, it was suggested that metallothionein has importance in ovarian tumorogenesis and it could be used as a prognostic parameter [15, 16, 27, 42]. The result of previous studies and this investigation support that during the diagnosis of borderline and malignant ovarian tumors in difficult cases, metallothionein expression could be used.
In conclusion, the immunohistochemical staining with p53, p21, bax, c-kit, and metallothionein may be helpful for the typing of ovarian tumors as benign, borderline and malignant or serous and mucinous. p53, p21, bax, c-kit, and metallothionein may have different roles in the pathogenesis of ovarian tumor types. p53 and metallothionein may be helpful in the typing of borderline and malignant ovarian tumors. The immunohistochemical staining with bcl-2 and telomerase may not provide meaningful contribution for the typing of ovarian tumors. Because of limitations of low number of cases in histopathological subtypes of ovarian tumors, the results of this study needs to be supported by further studies with addition of new tumor markers to develop a panel suitable for use in routine pathology laboratories.