In bladder cancer the observed microsatellite instability indicates that mi
smatch repair deficiency could be a frequently involved factor in bladder c
ancer progression. To investigate this hypothesis we analysed extracts of s
even bladder cancer cell lines and, as a novel approach, five clinical canc
er samples for mismatch repair activity. We found that one cell line (T24)
and three of the clinical samples had a reduced repair capacity, measured t
o similar to 20% or less. The T24 cell extract was unable to repair a G-G m
ismatch and showed reduced repair of a 2-base loop, consistent with diminis
hed function of the MSH2-MSH6 heterodimer. The functional assay was combine
d with measurement for mutation frequency, microsatellite analysis, sequenc
ing, MTT assay, immunohistochemical analysis and RT-PCR analysis of the mis
match repair genes MSH2, MSH3, MSH6, PMS1, PMS2 and MLH1. A > 7-fold relati
ve increase in mutation frequency was observed for T24 compared to a bladde
r cancer cell line with a fully functional mismatch repair system. Neither
microsatellite instability, loss of repair nor mismatch repair gene mutatio
ns were detected. However, RT-PCR analysis of mRNA levels did detect change
s in the ratio of expression of the Mut S and Mut L homologues. The T24 cel
l line had the lowest MSH6 expression level of the cell lines tested. Ident
ical RT-PCR analysis of seventeen clinical samples (normal urothelium, 7; p
Ta low stage, 5; and pT1-4 high stage, 5) indicated a significant change in
the expression ratio between MSH3/MSH6 (P < 0.004), MSH2/MSH3 (P < 0.012)
and PMS2/MLH1 P < 0.005, in high stage bladder tumours compared to normal u
rothelium and low stage tumours. Collectively, the data suggest that imbala
nced expression of mismatch repair genes could lead to partial loss of mism
atch repair activity that is associated with invasive bladder cancer. (C) 2
001 Cancer Research Campaign http://www.bjcancer.com.