Mk. Dodson et al., COMPARISON OF LOSS OF HETEROZYGOSITY PATTERNS IN INVASIVE LOW-GRADE AND HIGH-GRADE EPITHELIAL OVARIAN CARCINOMAS, Cancer research, 53(19), 1993, pp. 4456-4460
Loss of heterozygosity (LOH) studies were performed to investigate the
genetic differences which separate low-grade (LG), high-grade (HG), a
nd borderline epithelial ovarian carcinomas. Fresh tumor samples and b
lood were obtained from 58 patients (20 LG, 34 HG, and 4 borderline tu
mor specimens) undergoing surgery for ovarian carcinoma at Mayo Clinic
. Tumors were graded using a modified Broder's classification with inv
asive grades 1 and 2 considered LG, invasive grades 3 and 4 considered
HG, and tumors with no evidence of stromal invasion classified as bor
derline. Polymorphism analysis was performed using 76 restriction frag
ment length polymorphisms and variable number of tandem repeats and 59
microsatellite markers representing all chromosome arms. Chromosome a
rms 6p, 17p, 17q, and 22q were found to be frequently lost in LG as we
ll as HG tumors. Chromosome arms 13q and 15q were lost to a significan
tly greater extent in HG tumors compared to LG neoplasms (P = 0.003 an
d P = 0.08, respectively). Conversely, 3p loss was seen more frequentl
y with LG tumors (P = 0.02). The majority of LG tumors (65%) did not s
how frequent LOH in the allelotype analysis. In fact, a subset of 7 (7
of 20) LG tumors accounted for 76% of the total allelic loss in the L
G category. These tumors showed LOH almost identical to that of the HG
neoplasms. Borderline tumors showed a low rate of allelic loss. There
were no common events found between borderline and invasive tumors. O
ur data suggest that most HG tumors and a subset of LG tumors share ge
netic alterations at putative tumor suppressor genes detected by LOH s
tudies. Chromosome 6 and 17 losses appear to be early events while 13q
and 15q losses appear to be critical late events. However, a majority
of LG tumors appear to develop as a consequence of an alternative mec
hanism(s) which is not detected by LOH studies. Possibilities include:
(a) inactivation of tumor suppressor genes without LOH; (b) dominant
negative gene(s) in which only one allele requires mutation; and (c) c
hanges in dominant acting oncogenes. This unidentified phenomenon may
be operative in borderline tumors as well.