We defined a cytokine mRNA profile of 12 ovarian cancer biopsies, 10 normal/benign biopsies, six ovarian cancer cell lines and three ovarian cancer xenografts, using RT-PCR. The profile, based on screening for 25 cytokines and 12 receptor mRNAs, was rich in growth factors, pro-inflammatory cytokines and chemokines, but weak in lymphocyte-associated cytokines. The pattern was unique to ovarian tissue, but similar in normal, benign and malignant biopsies, with > 80% samples expressing 16 cytokines in common. Fourteen of these were also expressed by > 65% cell lines, but fewer were detected in xenografts. Potential autocrine loops existed for IL-1, IGF-1, M-CSF, GM-CSF and TNF-alpha. IL-4 and IFN-gamma receptors were expressed in absence of ligand. Chemokines RANTES, MIP-1 alpha and MIP-1 beta were expressed in biopsies, but were rarely detected in cell lines and absent from xenografts. IGF-1 and its receptor was expressed in every sample, as was IFN-gamma receptor. Another 10 cytokine mRNAs and six receptors were expressed in > 80% samples. These may contribute to key survival/growth loops. Similarities between normal and malignant biopsies suggest that analogous processes of remodelling and repair occur. RT-PCR proved a rapid, reproducible screen, but further assays are required to detect quantitative differences between normal and malignant tissues and tumour models.