Immunohistochemical analysis of drug resistance-associated proteins in ovarian carcinomas

Loss of function of the tumor suppressor gene p53, increased expression of glutathione-S-transferase pi (GST7pi) and the major vault protein are involved in drug resistance of ovarian carcinomas. However, a study comparing these factors has not yet been performed. Therefore, paraffin-embedded material of 213 ovarian tumors with well-documented follow-up was used for immunohistochemical analysis of p53 protein, GSTpi, and major vault protein (antibodies LRP-56, LMR-5). Forty-six percent of the cases showed nuclear p53 accumulation. Strong immunoreactivity for GSTpi, LRP-56, and LMR-5 was seen in 50%, 36%, and 47%, respectively. p53 positivity was most often found in serous carcinomas (p < 0.05). Strong GSTpi expression was the only factor that correlated with clinical resistance to chemotherapy (p = 0.04). In the whole group, as well as in FIGO III cases stratified for residual disease < or = and >2 cm, p53 and GSTpi correlated with an adverse outcome (p = 0.01 for p53 and p = 0.04 for GSTpi). Strong LRP-56 or LMR-5 staining was associated with a tendency towards poorer prognosis, without reaching statistical significance. In multivariate analysis for FIGO III, only residual disease and p53 proved to be independent prognostic factors. Our observations confirm the prognostic significance of p53 accumulation in ovarian carcinomas. Only GSTpi immunoreactivity was significantly correlated with drug resistance.

20q13 and cyclin D1 in ovarian carcinomas. Analysis by fluorescence in situ hybridization

In ovarian carcinomas, alterations of the chromosomal region 20q13 and the cyclin D1 gene have been described. This study has sought to determine their prognostic significance. Fluorescence in situ hybridization (FISH) on dissociated nuclei and paraffin sections with DNA probes for 20q13.2 and cyclin D1, as well as immunohistochemistry (cyclin D1), were applied to formalin-fixed tissue of 69 invasive ovarian carcinomas, mainly of serous type. On dissociated nuclei 33/47 cases (70%) and on tissue sections 13/66 cases (20%) demonstrated an increase of 20q13.2 copies. The presence of > or =4 copies per nucleus (isolated nuclei) and > or =3 copies per nucleus (sections) was associated with an adverse prognosis (Kaplan-Meier for FIGO stage III after stratification for residual tumour: p=0.0049 and p=0.03, respectively). Thirty-four out of 47 cases (72%) showed an increase of cyclin D1 copies. Kaplan-Meier analysis for FIGO stage III after stratification for residual tumour>2 cm or < or =2 cm revealed an unfavourable outcome for cases with more than two cyclin D1 copies (p=0.04). No correlation was seen between FISH and immunohistochemistry. Multivariate analysis identified residual tumour (p=0.0002), 20q13.2 gain (p=0.0004) and cyclin D1 gain (p=0.0343) as independent prognostic factors. It is concluded that gains of chromosomal region 20q13.2 and the cyclin D1 gene are frequent and biologically important events, with prognostic relevance, in advanced ovarian carcinomas.