Contribution of immunohistochemistry in prognostic assessment of epithelial ovarian carcinoma --review of the literature I.

Epithelial ovarian carcinoma is worldwide the sixth most common female cancer, and this malignancy carries the highest mortality among all gynecological cancers. The high mortality is due mostly to the fact that the tumor is frequently diagnosed late, in advanced stage, as the early disease is often asymptomatic and no effective screening methods are available. The most important prognostic factors in ovarian carcinoma are the stage, size of residual tumor following surgery, presence of ascites, age and the general condition of the patient, tumor histology, and, in patients with early disease, also the grade of the tumor. Large number of studies on prognostic and predictive factors in epithelial ovarian carcinoma has been published, often with contradictory results. The most intensely studied prognostic factors are those for expression of hormonal receptors, for tumor proliferation activity (mainly by antigen Ki-67 and topoisomerase IIalpha), the markers of apoptosis (p53, p21, mdm2, bcl-2 and other proteins), or other oncoproteins (particularly HER-2/neu).


Introduction
Epithelial ovarian cancer (EOC) is the sixth most common cancer in women worldwide. In Western countries, ovarian cancer is the fifth most common malignancy and ranks fourth in cancer mortality (24). EOC carries the highest mortality among all gynecological malignancies. The high mortality is due mostly to the fact that the tumor is frequently diagnosed late, in advanced stages (III, or even IV) as the early stages are often asymptomatic, and no effective screening methods are available. The average 5-year survival is around 40%, in patients with advanced disease only 10-20% (24,43).
In the absence of preventable etiologic factors or effective tools for screening, the only possible means of improving survival currently lies with the optimal management of patients after initial diagnosis (6). In the early 1990s, the Gynecologic Oncology Group (GOG) demonstrated that age, performance status, stage, tumor grade, tumor cell type (histology), presence or absence of ascites, size and number of residual lesions after primary cytoreduction surgery, and administration of cisplatin-based chemoterapy are of prognostic significance in advanced EOC (19,32).
There has been considerable effort directed at identifying factors that accurately predict outcome in patients with advanced EOC. A variety of putative prognostic factors have been reported, but their independent prognostic significance remains unclear. Immunohistochemistry has been widely used in the search of such markers. In this review four groups of potential prognostic factors are discussed: 1. steroid receptors expressed by tumor cells, 2. cell growth kinetics in ovarian carcinoma assessed by examination of proliferation activity of the tumor cell (mainly by antigen Ki-67 and topoisomerase IIα), 3. the expression of oncoprotein HER-2/neu, (ER+/PR+, ER+/PR-, ER+/PR-, ER-/PR-) have been identified as predictive factors for response to endocrine treatment (28).
In contrast to breast carcinoma, limited, inconclusive and conflicting data regarding the prognostic significance of ER and PR expression are available for EOC, and clinical value of determining steroid hormone receptors in this malignancy is controversial. Earlier work (23,39,41) was performed using DCC method (dextran-coated charcoal). DCC detection of estrogen receptor, however, has been found to give falsely positive results if surrounding benign tissue expresses this receptor, or falsely negative results have been obtained in cases of receptors masked by endogenous estrogen. Today, immunohistochemistry is considered the method of choice because it allows an exact assignment of ER and PR expression to tissue components of interest (28).
Some of recent studies found expression of PR to be an independent indicator of favorable prognosis in EOC (1,14,15,28). Münstedt et al. demonstrated that the favorable course of PR+ ovarian carcinoma relates primarily to the subgroup ER-/PR+ expressing tumors. This tumor phenotype was associated with superior prognosis compared to tumors with other steroid hormone receptors combinations, and ER-/PR+ tumors were associated with lower volume of ascites, less advanced tumor stage and lower tumor grade, reflecting a less aggressive tumor biology. The favorable effect of the ER-/PR+ phenotype was retained in multivariate analysis (28).
Although there is no single explanation for the effect of steroid hormone receptor expression on prognosis, two hypotheses have been proposed. Estrogen-responsive cells efficiently repair DNA and avoid apoptosis, leading to clonal expansion and drug resistance (30). On the other hand, progesterone promotes cell differentiation and apoptosis, and stimulation of PR inhibits DNA synthesis and cell division (29). These mechanisms may explain why patients with ER+/PR-tumors have the worst and those with ER-/ PR+ tumors the best prognosis (28).
However, other studies did not confirm these results (12,44). Our study on 96 patients with ovarian carcinoma found prognostic significance of expression of PR, and ER-/ PR+ phenotype of tumors only in univariate, but not in multivariate analyses (44).

Proliferation activity of the tumor cell
The proliferation index showed the correlation with the prognosis and other known clinicopathologic features in several primary tumors, including lung and breast carcinomas, and lymphomas. The number of proliferating cells can be determined using a variety of methods, but many of these methods have significant technical limitations. In particular, DNA flow cytometry has been widely used in EOC. However, for DNA index only a single cell suspension is required, and the tissue architecture is lost and not evaluable (9).
Immunohistochemistry, using monoclonal antibodies on formalin-fixed paraffin-embedded archival material has been widely used. Ki-67 and topoisomerase IIα are most frequently evaluated immunohistochemically detectable proliferation antigens. Ki-67 is a nuclear non-histone protein expressed in cells in G1, S, G2, and M cell cycle phases, but absent from quiescent cells in G0. Type II DNA topoisomerases are nuclear enzymes that play a crucial role in DNA replication. They catalyze the relaxation of supercoiled DNA and separate intertwined DNA duplexes in an ATP-dependent process, through the generation of a doublestranded nick on the DNA during trascription. Topoisomerase IIα is expressed during the G1, S, G2, and M phases cell cycle (13).
Studies on DNA content and cell proliferation in EOC have yielded conflicting results regarding the prognostic significance of these parameters. Garzetti et al. (9) and Huettner et al. (20) found that malignant ovarian neoplasm had higher median percentage of Ki-67 staining than borderline and benign tumors, and they observed a significant relationship between the Ki-67 index and disease-free survival that was independent of histologic grade and stage. Similarly, Goff et al. (12), Rölke et al. (36), Sengupta et al. (38), and Kaern et al. (21) found that Ki-67 is marker that is expressed differently between the short-and long-term survivors. High cellular proliferative activity was associated with poor outcome (21). On the other hand, other studies did not confirm relationship between proliferation activity and EOC prognosis (5,34,44).
Gotlieb et al. tried to evaluate topoisomerase IIα compared to Ki-67 expression as a marker for tumor behavior and for prognosis in EOC. Ki-67 expression was more frequent in short-term survivors compared to long-term survivors, but the difference was less prominent than with topoisomerase IIα. Specificity and sensitivity as prognostic factors was 88.2% and 93.8% for topoisomerase IIα, compared to 55.6% and 88.2% for Ki-67 (13). Similar results were also reported by van der Zee et al. (45).

Oncoprotein HER-2/neu (c-erb-2)
The HER-2/neu oncogene encodes a transmembrane glycoprotein that is member of the class I receptor tyrosine kinase family, which includes the epidermal growth factor, HER-2/neu, HER-3 and HER-4 (16). The HER-2/neu oncogene is located on chromosome 17 and is not activated by a point mutation, but through amplification and over expresion of the wild-type gene (16,27).
Amplification of the HER-2/neu oncogene may be observed in 20-30 % of cases in a wide spectrum of neoplastic disorders (e.g. breast, salivary glands, or lung carcinomas), and HER-2 over-expression is correlated with a poor prognosis (22,27). In clinical practice, the monoclonal anti-HER-2/neu antibody transtuzumab (Herceptin) has been used as the first molecular targeted biologic agent for patients with HER-2/neu over-expressing metastatic breast carcinoma and, more recently, in adjuvant setting (42).
Different rates of HER-2/neu expression result from different methods and techniques of detecting HER-2/neu and different criteria for evaluating HER-2/neu expression. First, methods of detecting HER-2/neu expression include immunostaining HER-2/neu expression (with various antibodies) and the direct detection of amplification of the HER-2/neu gene such as fluorescent in situ hybridization (FISH). Second, the scoring system that used in studies was different. In recent studies staining for determining HER-2/neu protein expression was scored on a scale 0, +1, +2 and +3 according to the HercepTest (HercepTest, DAKO, Glostrup, Denmark), and +2 and +3 were regarded as over-expression. This scoring system was developed in 1997, and HercepTest has been approved by the United States Food and Drug Administration as a standardized diagnostic kit to detect HER-2/neu. Therefore, recent studies have reported more reliable results than previous studies (42).
Most studies examined relationship between HER-2/ neu expression and prognosis in various histological types of ovarian carcinoma, but not in a specific histological type. The examination of the relationship between HER-2/neu expression and outcome should focus on a prognostic importance of histological type and considered to the stage of cancer. Tanabe et al. limited their study to clear cell carcinoma that is chemotherapy-resistant EOC variant of poor prognosis and found neither association between HER-2/neu expression and outcome, nor association between HER-2/neu over-expression and the stage or lymph node metastasis. This study demonstrated that HER-2/neu expression in this histological type was not a prognostic risk factor (42).

Markers of apoptosis
Proto-oncogenes and tumor suppressor genes play a critical role in normal cell growth and in tumorigenesis. Proto-oncogenes normally stimulate differentiation and proliferation, but when altered, promote malignant transformation. In contrast, tumor suppressor genes inhibit cell division and/or promote cell death, and inactivation causes loss of the normal negative control of cell growth. Genetic alterations thus drive the transformation of normal cells into highly malignant clones (31).
The p53 tumor suppressor gene is the most frequently mutated gene in human cancer that plays a critical role in the regulation of cell cycle and apoptosis. Mutations of p53 have been found in approximately 40-80% of EOC cases (35). The p53 gene resides at a critical "crossroads" that modifies diverse cellular functions. In the event of injury, human cells are dependent on a functional p53 for DNA repair, or, if the damage is irreparable, for apoptosis. The p53 protein, acting as a transcription factor, activates or alternatively represses the transcription of genes leading to the expression of specific elements necessary for the inhibition of cell growth and the induction of apoptosis. The absence of functional p53 thus leads to deregulated cellular proliferation. Mis-sense mutations of the p53 gene result in protein that have longer half-life than their wild-type counterparts and are resistant to degradation. As a result, these mutations appear to give rise to p53 over-expression by immunohistochemical techniques with a variety of different antibodies. Null mutation (insertion, deletion, splice site aberration, and nonsense) result in a truncated protein product that cannot bind DNA or induce apoptosis. Such mutation generally do not result in increased p53 stability, and the truncated protein is often undetected by conventional immunohistochemical techniques (40).
Several lines of evidence have elucidated a major role of functional p53 gene for enhancing therapeutic response to chemotherapy or radiation. Mutant p53 therefore may directly decrease tumor cell sensitivity to chemotherapeutic agents and promote the emergence of drug resistant population of cancer cells (38,40).
The role of p53 protein in EOC is contentious, and there is a number of studies with contradictory results. Several studies have identified p53 protein as an adverse prognostic factor for survival in EOC (18,31,33,40). Other studies have suggested that alterations in p53 expression in ovarian cancer affect sensitivity to chemotherapy (35,38). In contrast, there are a number of studies that suggest that p53 expression has no prognostic value in EOC (8,25).
The aim of study Dogan et al. was to investigate the prognostic significance of p53 nad mdm2 protein expression in EOC (7). Mdm2 gene is a proto-oncogene that encodes a nuclear protein that negatively regulates the trascriptional activating function of p53. In addition, mdm2 protein can sequester the p53 protein. Thus, over-expression of mdm2 protein results in an effect similar to the mutational inactivation of p53. It was found that mdm2 expression predicts response to chemotherapy, and co-expression of p53 and mdm2 proteins was also related to poor outcome (7). However, Mano et al. did not confirm this in multivariate analysis (26).
Protein p21 waf1/cip1 is a cyclin-dependent kinase inhibitor that is usually induced through a p53 related pathway. p21 waf1/cip1 has been shown to be integral to control of the cell cycle after DNA damage. Up-regulation of p21 waf1/cip1 by p53 is integral to sustaining G2 arrest after DNA damage. In cells without functional p53, p21 waf1/cip1 can be upregulated by the activation of protein kinase C. Although p21 waf1/cip1 has been studied in EOC, the role of this protein as a prognostic indicator is still controversial (11,38). Some studies confirm the importance of the combination of p21 and p53 staining in determining EOC prognosis.
Expression of p53 protein in the absence of p21 waf1/cip1 expression was a better marker of poor prognosis than either p53 or p21 waf1/cip1 expression status alone (2,11,46).
Proteins of the Bcl-2 family are critical regulators of the apoptotic pathway. Certain members of the family promote apoptosis (e.g. Bax, Bad, Bcl-X S ) while others have an anti-apoptotic function (Bcl-2, Bcl-X L ). The ratio of proand anti-apoptotic members, such as Bax and Bcl-2 is critical in the inhibition or induction of apoptosis (38). A variety of tumors, including EOC, resistant to anticancer drugs express Bcl-2, suggesting that Bcl-2 may protect cancer cells from programmed cell death induced by a variety of antitumor agents, including cisplatin (26). Mano et al. found that Bcl-2 protein may represent a possible predictor of response to chemotherapy. Multivariate analysis revealed that Bcl-2 protein is significant independent prognostic factor (26), but other studies did not confirm this observation (4,38). Geisler et al. found that Bcl-2 protein itself is not an independent prognostic indicator, but the combination of p53 and Bcl-2 can independently predict survival (10).

Conclusion
Immunohistochemicaly detectable prognostic and predictive factors in EOC have recently been widely covered in the literature, specifically the expression of steroid receptors by tumor cells, the assessment of cell growth kinetics by examination of proliferation activity of the tumor, the expression of oncoprotein HER-2/neu, and the expression of markers of apoptosis. A number of studies have been reported, often with contradictory results. However, no single immunohistochemically detectable marker has been so far identified in EOC that would provide reliable and reproducible prognostic or predictive information.