[Mutation analysis of hMSH2 and hMLH1 genes in Chinese hereditary nonpolyposis colorectal cancer families].

To determine the germ-line mutations of hMSH2 and hMLH1 genes in Chinese hereditary nonpolyposis colorectal cancer (HNPCC) families' probands or in patients fulfilling different clinical criteria or guidelines; to clarify the nature and distribution of the mutations; to evaluate the sensitivity of different clinical criteria in mutation prediction.The entire coding regions (35 exons including exon-intron boundaries) of hMSH2 and hMLH1 genes were directly sequenced in 24 Amsterdam criteria (AC) probands, 15 Japanese criteria (JC) probands (except AC kindreds) and 19 Bethesda guidelines (BG) patients (except two former groups). All available affected and unaffected members from families of those with mutations were screened for mutation.In 16 unrelated families selected by the different clinical criteria, 17 germ-line mutations were found with 11 (64.7%) of hMLH1 and 6 (35.3%) of hMSH2. Two mutations were identified in one of the families. Among the 17 germ-line mutations, 12 had not been reported previously. A diversified mutation spectrum was found, but 6 hMLH1 mutations were found to be concentrated in the region encompassing exon 14, 15 and 16. There was a wide spectrum of mutation type including frame shift, nonsense, splice site mutation, in frame insertion or deletion and missense mutations. The mutation detection rate of hMSH2 and hMLH1 in the AC group was significantly higher than that in the JC group (12/24 vs. 3/15). On the other hand, a low mutation rate (1/19) was detected in 19 BG patients. The mutation cosegregated with disease. Besides, three different genotypes in tumors from probands of mutation-positive families were found.hMSH2 and hMLH1 mutations in Chinese HNPCC families show a wide spectrum. It seems that hMLH1 gene is involved more frequently than hMSH2 gene in Chinese HNPCC families. Different clinical criteria predict mutations with different sensitivities. The Amsterdam Criteria are most sensitive, while Japanese Criteria are highly practical and the Bethesda Guidelines are also practical to some extent. Gene mutations cosegregate with the disease phenotype. Carriers with no symptom in HNPCC families are most vulnerable groups, follow-ups are required for this group to get early diagnosis and to prevent the development of CRCs.