PubMed1998
Strong epidemiological evidence links human papilloma viruses (HPV) with the development of cervical intraepithelial neoplasia (CIN) and invasive cancers of the uterine cervix. The localization of HPV DNA sequences high up in the female genital tract (in benign and malignant lesions) is not that uncommon, but its precise significance is uncertain. In particular, the detection of HPV DNA sequences by polymerase chain reaction (PCR) needs careful interpretation, because the source of the amplicon may emanate from tumor cells, direct contamination from the cervix, or possibly from extratumoral sites in the endometrium. We have previously reported the identification of koilocyte-like changes in the squamous epithelium of some endometrial adenoacanthomas. Adenoacanthomas (adenocarcinoma with squamous metaplasia) are mixed epithelial tumors arising in the endometrium composed of malignant glandular areas admixed with benign metaplastic squamous epithelium. The rarer adenosquamous carcinoma containing both malignant glandular and squamous areas is also described. The origin of benign/malignant squamous epithelial islands in endometrial tumors has been the subject of speculation, with some investigators considering an origin from metaplastic glandular endometrial cells. In this study, we examined 10 normal endometrial samples, 20 adenocarcinomas, 41 adenocarcinomas with squamous metaplasia, and two adenosquamous carcinomas, (including control cervical material where possible) for the presence of HPV DNA sequences using nonisotopic in situ hybridization (NISH), type-specific HPV PCR, general primer PCR (to detect sequenced and unsequenced HPVs), and PCR in situ hybridization (PCR-ISH). We did not identify HPV DNA sequences in normal endometrial tissue. In adenocarcinomas (endometrioid type), HPV was only identified in 2 of 20 cases by PCR, both of which were HPV 11 positive. We were unsuccessful in identifying HPV in endometrial carcinomas by NISH or by PCR-ISH, raising the possibility of contamination from the cervix in the two positive cases. In adenoacanthomas, a low-risk HPV type (HPV 6) was found in 19 of 41 cases. NISH signals were intranuclear in location in squamous regions of adenoacanthomas. Additional positive nuclei were uncovered using PCR-ISH, which increases the sensitivity of standard NISH detection. HPV DNA sequences were located in some malignant endometrial glandular epithelial cells, but this accounted for a minority of samples. HPV DNA sequences were not detected in extraepithelial sites. Mixed infection by two different HPV types was identified in two cases. Most cases showed similar HPV types in cervical and endometrial lesions, although discordant cases were uncovered. In adenosquamous carcinomas, one case showed mixed infection with HPV 6 and 33 by PCR. The apparent segregation of low-risk HPV type (HPV 6) with benign squamous metaplastic epithelium in adenocarcinoma with squamous metaplasia, and high-risk type (HPV 33) with malignant squamous epithelium in adenosquamous carcinoma, raises important questions in relation to the role of HPVs in mixed epithelial tumors of the endometrium and their interplay in the pathogenesis of squamous metaplasia at extracervical sites.