Loss of heterozygosity mapping at chromosome arm 16q in 712 breast tumors reveals factors that influence delineation of candidate regions

Loss of heterozygosity (LOH) at the long arm of chromosome 16 occurs in at least half of all breast tumors and is considered to target one or more tumor suppressor genes. Despite extensive studies by us and by others, a clear consensus of the boundaries of the smallest region of overlap (SRO) could not be identified. To find more solid evidence for SROs, we tested a large series of 712 breast tumors for LOH at 16q using a dense map of polymorphic markers. Strict criteria for LOH and retention were applied, and results that did not meet these criteria were excluded from the analysis. We compared LOH results obtained from samples with different DNA isolation methods, ie., from microdissected tissue versus total tissue blocks. In the latter group, 16% of the cases were excluded because of noninterpretable LOH results. The selection of polymorphic markers is clearly influencing the LOH pattern because a chromosomal region seems more frequently involved in LOH when many markers from this region are used. The LOH detection method, i.e., radioactive versus fluorescence detection, has no marked effect on the results. Increasing the threshold window for retention of heterozygosity resulted in significantly more cases with complex LOH, i.e., several alternating regions of loss and retention, than seen in tumors with a small window for retention. Tumors with complex LOH do not provide evidence for clear-cut SROs that are repeatedly found in other samples. On disregarding these complex cases, we could identify three different SROs, two at band 16q24.3 and one at 16q22.1. In all three tumor series, we found cases with single LOH regions that designated the distal region at 16q24.3 and the region at 16q22.1. Comparing histological data on these tumors did not result in the identification of a particular subtype with LOH at 16q or a specific region involved in LOH. Only the rare mucinous tumors had no 16q LOH at all. Furthermore, a positive estrogen content is prevalent in tumors with 16q LOH, but not in tumors with LOH at 16q24.3 only.