Analysis of HLA-A2 frequency as prognostic and/or risk factors in ovarian cancer patients

Abstract

Major histocompatibility complex antigens are mandatory for the immune response, and genetic imbalance may be linked to tumour escape. The HLA-A2 allele in Scandinavia has a high prevalence; it decreases with latitude and also with ovarian cancer mortality in Europe. In solid tumours, HLA-A2 has been suggested to be a risk factor and a negative prognostic factor.

Study I. 32 unrelated Swedish women with relapsing or progressive ovarian cancer were analyzed for the genotypes at the HLA-A, -B, -Cw, and -DRB1 loci by PCR/sequence-specific oligonucleotide hybridization. The frequencies of HLA alleles of healthy Swedish bone marrow donors, provided by the coordinating centre of the Bone Marrow Donors Worldwide Registries, Leiden, The Netherlands were used as controls. The HLA-A1 and HLA-A2 genes and phenotype frequencies were increased, while HLA-A3 was decreased in ovarian cancer patients compared to healthy Swedish donors. HLA-A2 homozygotes were 2-fold higher in patients. HLA-B15 and HLA-Cw3 were only present in HLA-A2 positive patients, hence the HLA-A2-Cw3 and -B15 haplotypes were segregated. Furthermore, in these patients, HLA- A2-B5, HLA-A2-B8, HLA-A2-DRB1*03 and HLA-A2-DRB1*04 haplotypes were also increased, but not segregated.

Study II. In order to confirm the assumption, that HLA-A2 may have a role in the prognosis of ovarian cancer, we have examined the presence of HLA-A2 in all patients with ovarian cancer, admitted to the department of Oncology-Pathology, Karolinska Institute, over a 1-year period (1995) and related its presence to survival. A population-based set of 97 patients with confirmed epithelial ovarian cancer were recorded in a database by age, histology, stage, surgery and treatment. At the time the study was initiated, the majority of the patients were not alive and HLA-A2 expression was therefore determined by PCR/sequence-specific oligonucleotide hybridization, using DNA extracted from paraffin-embedded tissue specimens. 88 patients with a median age of 65 years (36 87) could be evaluated. 44% were serous adenocarcinomas, 28%-endometrioid, 6%-mucinous, 13%-clear cell carcinomas, 7%-undifferentiated and 2%-other epithelial tumours. Stages I II comprised 33% and stages III IV-67%. In stages III IV and serous histology, 73% were HLA-A2 positive. Cox analysis, in this group, showed high univariate (HR7.16; CI 2.04 25.03; p=0.002) and multivariate (HR 6.8; CI 2.10 22.4; p=0.001) Hazard Ratios. None of the HLA-A2 positive patients survived 5 years, compared to more than 50% of the HLA-A2 negative patients.

Conclusion. In cohort of 32 patients with ovarian cancer, there are indications of an unusual overrepresentation of HLA-class I-II genes/haplotypes, as well as segregation for the HLA-A2-Cw3 and -B15 haplotypes. Investigation on a larger series of ovarian cancer patients confirmed that HLA-A2 is a negative factor for survival in women with serous adenocarcinomas of the ovary in stages III-IV. This finding has implications for clinical patient management. Association with known oncogenes needs further analysis.