Correlation of HLA-A02* genotype and HLA class I antigen down-regulation with the prognosis of epithelial ovarian cancer

In recent years, evidence is accumulating that cancer cells develop strategies to escape immune recognition. HLA class I HC down-regulation is one of the most investigated. In addition, different HLA haplotypes are known to correlate to both risk of acquiring diseases and also prognosis in survival of disease or cancer. We have previously shown that patients with serous adenocarcinoma of the ovary in advanced surgical stage disease have a particularly poor prognosis if they carry the HLA-A02* genotype. We aimed to study the relationship between HLA-A02* genotype in these patients and the subsequent HLA class I HC protein product defects in the tumour tissue. One hundred and sixty-two paraffin-embedded tumour lesions obtained from Swedish women with epithelial ovarian cancer were stained with HLA class I heavy chain (HC) and beta(2)-microglobulin (beta(2)-m)-specific monoclonal antibodies (mAb). Healthy ovary and tonsil tissue served as a control. The HLA genotype of these patients was determined by PCR/sequence-specific primer method. The probability of survival was calculated using the Kaplan-Meier method, and the hazard ratio (HR) was estimated using proportional hazard regression. Immunohistochemical staining of ovarian cancer lesions with mAb showed a significantly higher frequency of HLA class I HC and beta(2)-m down-regulation in patients with worse prognosis (WP) than in those with better prognosis. In univariate analysis, both HLA class I HC down-regulation in ovarian cancer lesions and WP were associated with poor survival. In multivariate Cox-analysis, the WP group (all with an HLA-A02* genotype) had a significant higher HR to HLA class I HC down-regulation. HLA-A02* is a valuable prognostic biomarker in epithelial ovarian cancer. HLA class I HC loss and/or down-regulation was significantly more frequent in tumour tissues from HLA-A02* positive patients with serous adenocarcinoma surgical stage III-IV. In multivariate analysis, we show that the prognostic impact is reasonably correlated to the HLA genetic rather than to the expression of its protein products.