Human papillomavirus infection, centrosome aberration, and genetic stability in cervical lesions.

DNA replication and centrosome duplication have to be strictly synchronized to guarantee genomic stability. p53, pRb, cyclin E, and cyclin A are reported to be involved in the synchronizing process. We investigated the relationship between papillomavirus infection, centrosome aberration and aneuploidy during genesis of cervical carcinoma. The number of centrosomes found in cells from normal cervical epithelium (n = 5), condyloma acuminata (n = 5), cervical intraepithelial neoplasia (CIN) I, II, and III (n = 14) and invasive cervical carcinoma (n = 5) was analyzed by gamma tubulin immunofluorescence staining. The nuclear DNA content was investigated by image cytometry and human papillomavirus (HPV) infection was determined by polymerase chain reaction. Normal epithelia and condyloma acuminata showed cells with one or two centrosomes, whereas CIN lesions showed cells with an increasing number of centrosomes. This abnormality was found to be lowest in CIN I lesions, increased with advancing grade of CIN and was highest in lesions of invasive carcinomas. In parallel, an increasing number of cells with aberrant DNA content was seen. All carcinomas and all except one of the CIN III lesions showed aneuploidy. Three CIN II cases were aneuploid and two cases with CIN I were tetraploid. Normal epithelia and condyloma acuminata showed diploidy. All invasive carcinomas and lesions with CIN were positive for high-risk HPV types 16, 18, or 31, except one invasive carcinoma and one CIN II lesion positive for universal primers only. Three condyloma acuminata were HPV 16-positive and one HPV 6-positive. The results suggest that high-risk HPV infection is correlated to a progressive numerical disturbance of centrosome replication followed by progressive chromosomal aberrations in CIN lesions and invasive carcinomas.

Laminin-5 as a marker of invasiveness in cervical lesions.

BACKGROUND: Treatment decisions for cervical cancer, a common disease worldwide, depend on demonstrating whether or not tumor invasion of the surrounding tissue has occurred. Invasion can be difficult to assess by standard histopathologic methods, especially when limited amounts of tissue are available. Several studies of a variety of cancers have reported increased expression of laminin-5-an important attachment protein for epithelial cells-in invasive carcinomas. This study was designed to investigate whether the presence of laminin-5 is related to the invasive capacity of cervical lesions. METHODS: We used immunohistochemical methods to stain archival, paraffin-embedded sections of cervical lesions with a polyclonal antibody specifically targeting the gamma2 chain of human laminin-5 protein. The study sample included 23 lesions of mild and moderate dysplasia (cervical intraepithelial neoplasia [CIN] 1 and 2, respectively), 32 lesions of severe dysplasia or carcinoma in situ (CIN 3), 15 lesions of microinvasive cancer, and 20 lesions of frankly invasive cancer. Cellular proliferative activity was also investigated by the use of monoclonal MIB-1 (directed against the antigen Ki-67) and anticyclin A antibodies. RESULTS: Invasiveness of cervical lesions was positively associated with immunohistochemical staining of the gamma2 chain of laminin-5 (two-sided P =.001). All CIN 1 and CIN 2 lesions-except one CIN 2 lesion later shown to be invasive cancer-and 21 CIN 3 lesions tested negative for the gamma2 chain of laminin-5. Eleven CIN 3 lesions and all invasive cancers tested positive for this protein. One lymph node metastasis and a pleural metastasis from one of the patients with invasive cancer showed strong immunohistochemical positivity. Proliferative activity increased with advancement of the lesion but was not confined to cells positive for the gamma2 chain of laminin-5. CONCLUSIONS: These data suggest that antibodies directed against the gamma2 chain of laminin-5 can identify cervical lesions with invasive capacity and thus may be useful as a sensitive marker of early invasion.

Adenocarcinoma of the uterine cervix in Ireland and Sweden: human papillomavirus infection and biologic alterations.

Paraffin-embedded samples from cervical adenocarcinomas, 19 cases from Irish patients and 19 cases from Swedish patients, were analyzed by polymerase chain reaction for the presence of infection with human papillomavirus (HPV). The results were compared with DNA ploidy, proliferation activity, and p53 and p21/WAF1 expression. The studies were performed to discover whether high-risk HPV infection in adenocarcinomas of the uterine cervix is associated with an increased proliferative activity and genomic instability. The results show that the majority (84.6%) of patients 59 years of age or younger showed HPV infection. The overall prevalence of HPV DNA was 60.5%, with the high-risk types, 16 and 18, the most frequent. HPV-16 had a prevalence of 23.7% (9 of 38), and HPV-18 had a prevalence of 26.3% (10 of 38). The HPV-positive tumors predominantly showed a tetraploid DNA distribution pattern, whereas HPV-negative tumors more frequently showed highly scattered aneuploid DNA profiles. Both HPV-positive and HPV-negative cases displayed high proliferative activity, as indicated by high Ki-67 and cyclin A immunoreactivity. Tumor suppressor gene analysis detected low p53 expression and high p21/WAF1 expression in HPV-positive patients and high p53 expression without simultaneously increased p21/WAF1 (indicative of mutated p53) in HPV-negative cases in the groups of women older than 59 years of age.

Implications of human papillomavirus type for survival in cervical squamous cell carcinoma.

In a Swedish series of 107 invasive squamous carcinomas of the cervix, DNA extraction from paraffin-embedded material was successful in 97 cases. The prevalence of human papillomavirus (HPV) in this material was 86.6%, as determined by polymerase chain reaction (PCR) using both consensus and type-specific primers. HPV type 16 was most common (42.3%; other types were 31 (12.3%), 18 (9.3%) and 33 (10.3%). Seventeen cases (17.3%) were positive for the consensus primers only and were regarded as HPV of unknown type. There was no significant difference in corrected survival between patients with HPV-positive or -negative tumors. In the HPV-positive group, patients with tumors containing HPV 33 or HPV 18 had a significantly poorer prognosis than patients with tumors containing other types of HPV DNA (relative hazard 3.18, 95% confidence interval 1.37-7.39, P = 0.007), implying a prognostic significance of HPV type.

Human papillomavirus in cell samples from Stockholm Gynecologic Health Screening.

We compared the results of cytologic screening of 500 women in the Stockholm Gynecologic Health Control with human papillomavirus (HPV) detection by polymerase chain reaction (PCR) and in situ hybridization (ISH). There were two main age groups, one 30 years and younger and the other 40 years and older. There were relatively more women with HPV infection in the younger group than in the older one (15.7% as compared to 11.1%), but the difference was not significant in our material. Most cases (8/12) with cytologic atypia were HPV positive by PCR. HPV type 16 was most common, followed by types 31 and 18. HPV of unknown types was detected in 43.7% of HPV-positive cases. There was excellent agreement between PCR and ISH in detecting and typing HPV.

HPV detection in cytological cases with condylomatous or dysplastic changes: a study with PCR and in situ hybridization on cytological material.

Cytobrush samples of 80 patients, who previously had a cytological or histopathological diagnosis of condyloma and/or dysplasia were investigated for human papillomavirus infection (HPV) by polymerase chain reaction (PCR) and in situ DNA hybridization technique (ISH). The results were compared with concomitantly obtained cytological Pap-stained smears or, in some cases, histological sections. The time between the diagnosis of the original and the concomitant cytology/histopathology was less than 1 yr. Six additional patients had similar morphological diagnoses 2-4 yr before. Five more cases were included on clinical diagnosis of HPV. Compared with the original morphological diagnoses, 70% of the cases were positive by PCR and/or ISH. The concomitant morphology was not diagnostic of HPV in 44 out of 80 cases (55%), showing a relatively high percentage of cases morphologically normalized in the interval since the first specimen was taken. After detection with PCR, 30 cases (37.5%) were negative for HPV. Only one of the patients with a previous disease 2-4 yr before was HPV positive by PCR and two out of five patients with a clinical diagnosis of HPV. ISH could be performed on 67/80 cases, 43 of which were positive for HPV. There was a good agreement between the results of ISH and PCR, but there were six cases positive by ISH and negative by PCR. In these cases, few infected cells may have escaped detection by PCR. Both methods seem to be able to detect silent HPV infections and comparison with concomitant cytology/histopathology shows that morphology alone is insufficient for HPV detection in these cases.

Human papillomavirus type 16 capsids expose multiple type-restricted and type-common antigenic epitopes.

The study of viral infectivity and detection of viral capsid antigens of the major cervical cancer-associated human papillomavirus (HPV) type, HPV-16, requires knowledge of which epitopes are exposed in clinical specimens of infected tissue or on intact capsids. To define the antigenic epitopes of HPV-16, antisera to 66 overlapping synthetic peptides corresponding to the HPV-16 capsid proteins L1 and L2 and to seven peptide analogues were tested in immunoperoxidase stainings of consecutive sections from formalin-fixed, paraffin-embedded HPV infected tissue. Antisera against eleven different peptides from L1 and against seven different peptides from L2 recognized the HPV capsid antigen. Most epitopes were only found on the capsid antigen of certain genital HPV types, but four antigenic epitopes in L1 were detectable also in cutaneous wart specimens. All antigenic epitopes in L2 were restricted to genital HPV types and four L2 epitopes were only detectable in HPV-16 or HPV-33 positive specimens. The surface exposure of the antigenic epitopes was investigated by comparing the reactivity of the antipeptide antisera with intact or disrupted virions or capsids of HPV-11, HPV-16 and bovine papillomavirus (BPV). Twenty antipeptide sera from L1 and seven antipeptide sera from L2 were reactive with intact HPV-16 capsids at titres up to 1:146,000. Sixteen of these antisera were also reactive with disrupted HPV-16 capsids. Cross-reactivity with disrupted HPV-11 and BPV was detected for eleven and six antisera, respectively, whereas intact HPV-11 or BPV virions showed only weak cross-reactivity. In conclusion, the HPV-16 L1 and L2 capsid proteins contained multiple antigenic epitopes, most of which were shared with one or several additional HPV types.

The incidence of HPV in a Swedish series of invasive cervical carcinoma.

The incidence of HPV was studied in 71 invasive squamous carcinomas of the cervix using PCR technique. We used primers, which presumably recognize all types of HPV (consensus primers), and also type-specific primers. In situ hybridization was carried out in 24 of the cases. The overall incidence of HPV was 53/71 (75%) of which 5 cases were positive with the consensus primers only. However, 21/71 cases (30%) were negative for the consensus primers but positive for one of the type-specific primer pairs. This finding indicates that subgenomic deletions may have occurred in the viral genome upon integration in the human DNA. In situ hybridization was positive in 14/24 cases (58%), showing excellent correlation with PCR results. The HPV types detected were, in descending order of frequency: type 16 (52%), 31 (23%), 18 (13%), 33 (12%). No cases of HPV type 6 or 11 were found in this series of invasive carcinomas.

Detection of human papillomavirus infection in tissue blocks by in situ hybridization as compared with a polymerase chain reaction procedure.

Tissue blocks from 25 cases of condyloma and/or dysplasia were used for human papillomavirus typing with DNA in situ hybridization, compared with the very sensitive polymerase chain reaction. Only one of these cases was negative with both methods: a case of vaginal "koilocytosis." Polymerase chain reaction, as expected, was the more sensitive method, positive in 24 cases, with seven double infections. In situ hybridization was positive in 18 cases, with only two detected double infections. There was excellent agreement between the two methods in typing results. In all cases in situ hybridization showed a positive reaction in areas of koilocytosis and/or dysplasia.