Search

The role of the replication error-positive (RER+) phenotype in the development of specific subtypes of sporadic ovarian carcinomas was examined by screening for the presence of microsatellite instability (MI) in 47 tumors. The overall frequency of ovarian MI was 17% only. However, MI occurred in 50% of the ovarian endometrioid-type tumors, which was significantly more often than in all the other histological subtypes combined (8%). Five of the 8 RER+ tumors exhibited most marked type I instability, possibly representing a different mechanism than for the remaining type 2 tumors. The cDNA of the mutation suppression gene hMSH2, the gene most often associated with MI, was screened for alterations in 8 MI-positive and 5 MI-negative ovarian tumors. Only 3 changes were found. Complete loss of hMSH2 mRNA expression was detected in I tumor, while another expressed only an abnormal transcript containing a deletion of exon 3. One additional RER+ serous adenocarcinoma contained a rare polymorphism with a non-conservative amino acid change. One of 8 RER+ tumors showed loss of heterozygosity at the hMSH2 loci. Genetic instability, caused in part by alterations in the hMSH2 gene, may play an important role in the sporadic endometrioid subtype of ovarian tumors. Other mutator-phenotype genes may be responsible for the remaining cases of RER+ ovarian tumors.

The results of single-agent chemotherapy, with 11 anti-cancer agents, of 15 human gastro-intestinal and breast cancer lines xenografted into nude mice indicate inherent individuality of chemosensitivity spectrum of each tumor. The following 9 parameters have been measured as factors possibly relevant to chemosensitivity of tumor tissue or tumor-bearing mice: grade of histological differentiation, vascularity, percentage of necrosis, VDT, 3H-thymidine LI, human LDH activity in the cancer tissue, tissue/serum LDH ratio, TdR Pase activity, and serum CEA. These parameters exhibited presumably constant values for each tumor line. Chemosensitivity, i.e., inhibition of tumor growth by a given drug, was used as the dependent variable, and values of the 9 parameters in each cancer as the explanatory variables. Multiple regression analyses with stepwise deletion were performed for each of the 11 drugs. The equations for 8 drugs exhibited coefficients of determination of over 70%, and in particular those for M-83 (a derivative of mitomycin C), nimustine hydrochloride and doxorubicin exceeded 80% by equations with 3–4 parameters. Consequently, the estimated value for each line of effectiveness derived from the equations for these 8 drugs showed remarkable coincidences with the actual values for the inhibition rates of the corresponding drugs.