Loss of heterozygosity of the oestrogen receptor gene in breast cancer.

DNA from 67 primary breast carcinoma biopsies has been examined for loss of heterozygosity (LOH) using the microsatellite (TA)n repeat marker positioned 1 kb upstream of the oestrogen receptor (ER) gene. Forty-seven (70.1%) of the cases were informative; nine of these (19.1%) were positive for LOH. In three of the nine cases, there was total loss, and in the other six cases there was a marked reduction in the intensity of signal from one allele. LOH correlated weakly with histological grade and age, but not with ER status. This result suggests that LOH of the ER gene does not have an important role in the lack of ER function in breast cancer tissues. ImagesFigure 1

Human breast cancer is one of the typical hormone-dependent tumours, and approximately one-third of breast cancer patients respond to endocrine therapy (Rubens et al., 1980). The oestrogen receptor (ER) is a 66 kDa intranuclear protein consisting of six functional domains and is a ligand-activated transcription factor (Ponglilitmongkol et al., 1988). Cur-rently, ER content in tumours is not used to predict those patients who might benefit from endocrine therapy. ER status also provides prognostic information (Stewart et al., 1982). Tumours lacking ER and progesterone receptor (PgR) generally grow faster than tumours containing both ER and PgR (McGuire and Clark, 1989).
In this report, we examined LOH on the ER gene in 67 breast cancer patients using a highly informative microsatellite TA repeat marker positioned 1 kb upstream of the ER gene (Senno et al., 1992). We also analysed the relationship between LOH of the ER gene and ER content, as well as other clinicopathological parameters.

Patients and methods
Patients and DNA preparations Tissue from 67 breast carcinomas was obtained from the tissue bank of the breast unit at Guy's Hospital, London. Of the 67 tumours, 58 were infiltrating ductal carcinoma and nine were special types, including six infiltrating lobular carcinomas. The histopathological classifications were carried out according to the World Health Organization typing scheme for breast tumours (WHO, 1981). Patients were graded histopathologically according to the modified Bloom and Richardson method of Elston and Ellis, (1991). Blood samples also were taken from each patient. Genomic DNA from the breast cancer specimens and the blood samples was extracted by standard techniques (Sambrook et al., 1989).
Oestrogen and progesterone receptor determinations Cytosolic ER and PgR levels were measured using enzyme immunoassay (ERand PgR-EIA, Abbott Laboratories, Chicago, IL, USA). Positive ER and PgR status was defined as more than 20 fmol mg'-protein.
Statistical analvsis All comparisons between LOH and clinicopathological parameters were performed using the Kendall test. P-values <0.05 were considered statistically significant.

Results
Loss of heterozygositv of the ER gene The DNAs from a total of 67 tumours were studied for allele loss using the dinucleotide TA repeat microsatellite marker, positioned just upstream of the ER gene. Constitutional heterozygosity was observed in 47 cases (70.1%), and, of these, clear LOH was seen in nine (19.1%) (Figure 1). In three cases (DNA nos. 185, 120 and 256) there was complete loss, but in the other six cases (nos. 34, 39, 67, 207, 813 and 894) there was a marked reduction in the intensity of signal from one allele. This residual signal could be due to the presence of normal cells within the tumour sample.  Figure 2 shows a quantitative assessment of ER value vs LOH of the ER gene and demonstrates that there was no relationship between them.

4DE
In the early stages of human breast cancer the proliferation of tumour cells depends on oestrogen. After that, the cancer cells may acquire new proliferative pathways sequentially as a result of multiple genetic alterations. This then enables the tumour cells to bypass the oestrogen-dependent proliferation (Liu et at., 1988). Although there are many reports concerning variant ER genes (Sluyser, 1992;Fuqua et al., 1993;Pfeffer et at., 1993), the reason for the existence of ERnegative breast cancer is still not understood. We initially hypothesised that breast cancer with negative ER might be induced by the mutation of one allele and loss or replacement of a chromosomal segment containing the other allele. We used a highly informative polymorphic marker located very close to the ER gene to test this hypothesis. Our results suggest that, although LOH on the ER gene was seen in about 19% of the informative cases, there was no relationship between LOH and ER status. This indicated that allele loss may not play an important role in the lack of ER function in breast cancer tissues. Devilee et al. (1991) reported that the frequency of LOH on 6q was as high as 50%, and also found no relationship between LOH on 6q and ER status using the markers D6S37 and MYB (locus 6q27 and 6q22-q23 respectively).
On the other hand, it is possible for LOH to be randomly acquired and irrelevant to tumour development (Chen et al., 1992). The background incidence of random allelic losses may be higher in later stage lesions because they have had a longer time for randomly acquired lesions to be co-selected with other mutations which confer a selective advantage associated with malignant progression. The frequency of background LOH has been reported as 4-15% (Sato et al., 1990;Chen et al., 1992). In our case, the incidence of LOH on the ER gene might be higher than the background LOH.
In summary, we found no relationship between LOH on the ER gene and the lack of ER function in breast cancer tissue. Further molecular investigation is therefore required to understand the molecular basis of ER-negative breast cancer.