The molecular mechanisms involved in progression of squamous cell carcinoma of the anus (SCCA) are poorly elucidated, as well as the potential role of HIV infection. Loss of heterozygosity (LOH) is one of the mechanisms responsible for inactivation of tumor suppressor genes. We hypothesized that HIV-induced immunosuppression may contribute to an alternate molecular pathway in SCCA progression, through persistence of human papillomavirus infection within the anal canal. This study was undertaken to compare the molecular biology of SCCA in HIV-positive (HIV+) and HIV-negative (HIV-) patients. We retrieved tumor specimens from 18 HIV- and 10 HIV+ patients diagnosed with SCCA in two institutions. DNA from tumor and normal tissues was extracted and then amplified by polymerase chain reaction. LOH was investigated at 14 loci: three at 18q (DCC), two at 13q (Rb), three at 17p (p53), three at 11q, one at 2p, and two at 5q (APC). LOH was defined by a tumor DNA-to-normal tissue DNA ratio of >2. HIV+ patients were younger (36 +/- 7 years versus 53 +/- 13 years, P=0.001) and showed a trend toward tumors of larger size (3.7 +/- 1.6 cm versus 2.6 +/- 1.5 cm, P=0.09). The median CD4+ count in HIV+ patients at the time of diagnosis was 74 x 10(6)/L (range, 5-900). The overall frequency of LOH was 17.3% (41 LOH of 236 informative loci). Tumors in HIV- patients were more likely to present LOH than were tumors in HIV+ patients (24.1% versus 6.6%, P=0.0004). Differences between the two groups with regard to allelic losses were also observed at specific loci, such as 18q (41% [HIV-] versus 0% [HIV+], P=0.05), 17p (43% versus 10%, P=0.09), and 5q (33% versus 0%, P=0.12). Consistent LOH on chromosomes 17p, 18q, 5q, and 11q were observed in HIV- patients with SCCA. By contrast, allelic losses at 17p, 5q, and 18q seem to be rare in tumors of HIV+ individuals. These data suggest that immunosuppression may promote SCCA progression through an alternate pathway and that persistence of HPV infection within the anal canal may play a central role in this process.