Expression of epidermal growth factor receptor and HER-2/neu in normal and neoplastic cervix, vulva, and vagina.

Monoclonal antibodies were used to localize immunohistochemically epidermal growth factor receptor and HER-2/neu in normal and neoplastic frozen tissue samples from the lower genital tract of women. In squamous epithelia of the cervix, vulva, and vagina, epidermal growth factor receptor and HER-2/neu both were expressed most strongly by basal keratinocytes. Expression of both of these cell surface molecules decreased as cells underwent differentiation toward the mucosal surface. In contrast, both epidermal growth factor receptor and HER-2/neu were expressed throughout the entire thickness of the epithelium by undifferentiated squamous cells in squamous metaplasia, raised condyloma, and carcinoma in situ. In 34 squamous cancers of the cervix, vulva, and vagina, all malignant cells were found to have moderate to heavy staining for epidermal growth factor receptor. Staining of 33 of these cancers for HER-2/neu was light, although one patient who presented with distant metastases had heavy staining for HER-2/neu. These data suggest that although overexpression of HER-2/neu in squamous cancers of the lower genital tract is a rare event, it may be associated with aggressive biologic behavior.

Human papillomavirus infection of the uterine cervix analyzed by nonisotopic in situ hybridization.

Some types of human papillomaviruses (HPVs) have been suggested to be strongly related to uterine cervical carcinoma. An attempt to detect these in formalin-fixed, paraffin-embedded sections was made by either immunohistochemical or by in situ hybridization. Anticapsid protein of bovine papillomavirus antibody labeled with peroxidase was used for immunohistochemistry, and biotin was used instead of radioisotopes to label probes for in situ hybridization, which resulted in low background and a rapid procedure. Condylomatous changes were stained immunochemically with this antibody even in invasive carcinoma, whereas the carcinoma itself was not stained. Direct correlation was demonstrated by in situ hybridization between the HPV genome and histopathological structure, which was impossible by Southern or dot hybridization. HPV DNAs were detected in the nuclei of koilocytes and dyskeratinocytes of condylomata and dysplasias. Furthermore, hybridization signals of HPV DNAs in basal and parabasal cells suggested that HPV infection had already begun in the basal cells. In the case of malignant neoplasia accompanied by dysplasia, the same type of HPV was detected both in the malignant neoplasia and accompanying dysplasia. In one case of intraepithelial carcinoma, the very small focus of carcinoma just arisen in the cells of dysplasia was identified, and both were positive for HPV 18. This fact supports the suggestion that the carcinoma arises in dysplasia. Invasive carcinomas were classified further into keratinized, large-cell nonkeratinized, and small-cell nonkeratinized types, and the positive frequency for HPV 16 decreased as the differentiation of the carcinoma decreased. In the case of keratinized type of invasive carcinoma, strong hybridization signals were prominent around the malignant pearl formation.

Male partners of women with genital human papillomavirus infection. An assessment of colposcopic abnormalities by histological examination and human papillomavirus hybridization.

Men whose female sexual partners showed histological evidence of human papillomavirus infection were examined. Human papillomavirus DNA was identified in 29 of 35 biopsy samples of colposcopically-identified penile lesions. Human papillomavirus strains that were related to human papillomavirus genotypes 6/11 were observed most commonly (seven of eight patients) in the partners of patients with warty atypia or condylomata, while human papillomavirus strains that were related to human papillomavirus genotypes 16/18 were most-commonly (eight of 15 patients) observed in tissue from the partners of patients with cervical intraepithelial neoplasia. Measurement of human papillomavirus DNA in lesions by the filter in-situ hybridization technique more-frequently indicated human papillomavirus infection (29 of 35 lesions) than did conventional histopathological assessment (21 of 35 lesions) in this "high-risk" group. We conclude that colposcopically-identifiable lesions in male sexual partners are likely to contain human papillomavirus DNA, even if is no definite histological evidence of human papillomavirus infection is present, and that such lesions frequently contain strains of human papillomavirus that have been associated with the development of anogenital carcinoma.

Buffered formalin is the superior fixative for the detection of HPV DNA by in situ hybridization analysis.

In situ hybridization is used commonly for detection of human papillomavirus (HPV) DNA. There is little information, however, on whether the detection of HPV DNA by in situ hybridization can be affected by the way in which the tissue is fixed. To address this question, the authors compared the hybridization signal using this technique under low stringency conditions for several genital condylomata containing HPV 6 or 11 that were randomly subdivided and fixed in various fixatives for 16 hours. In all cases, the largest proportion of cells with koilocytotic atypia that had detectable HPV DNA was in buffered formalin-fixed tissue (80%), followed by tissue fixed in unbuffered formalin (70%), Hartman's solution (40%), and Bouin's solution (10%). After a high stringency wash, the greatest decrease in the overall hybridization signal was with tissue fixed in Bouin's solution; a minimal decrease was noted with tissue fixed in buffered formalin. Fixation in Bouin's solution for 2 hours gave in situ hybridization results comparable with buffered formalin fixation but with poorer cytologic detail. It is concluded that, of the fixatives studied, buffered formalin is superior for the detection of HPV DNA by in situ hybridization analysis.

[Detection of papillomaviruses by in situ hybridization with sulfonated probes]. / Détection par hybridation in situ des papillomavirus avec des sondes marquées par sulfonation.

A technique of detection by in situ hybridization of human papillomavirus in sections of condylomatous lesions is described. The probes are labeled and modified by sulfonation and the hybrids are revealed by immunohistochemistry, using alkaline phosphatase.

Comparison of formalin, buffered formalin, and Bouin's fixation on the detection of human papillomavirus deoxyribonucleic acid from genital lesions.

Detection of nucleic acid sequences homologous to human papillomavirus (HPV) relies primarily on their extraction from unfixed tissue. We detected HPV sequences in DNA extracted from paraffin-embedded tissue fixed in formalin (buffered and unbuffered) and Bouin's solution by dot blot hybridization. A detectable hybridization signal was noted in 32% of these fixed tissues which were chosen from cases where HPV DNA was detected in the unfixed tissue. When using a homologous 32P-labeled probe and a high stringency wash, the hybridization signal was lost if DNA was extracted after Bouin's fixation and diminished after formalin fixation, more so with unbuffered formalin. Similar differences in the hybridization signals among the different fixatives after high stringency wash were noted with in situ hybridization. Southern blot analysis showed that DNA extracted from tissues fixed in Bouin's was degraded and ranged in size from 100 to 500 base pairs as compared with 100 to 900 base pairs for DNA extracted from tissue fixed with unbuffered formalin. In contrast, no degradation was noted after fixation with buffered formalin. These results demonstrate that HPV sequences can be identified in DNA extracted from paraffin-embedded, fixed tissue. However, use of some fixatives may preclude identification of HPV type, by either dot blot or in situ hybridization.

Correlation of histology with human papillomavirus DNA detection in the female genital tract.

Genital condyloma and intraepithelial neoplasia secondary to Human Papillomavirus (HPV) infection are characterized by perinuclear halos and marked nuclear atypia (koilocytotic atypia) on cytologic and histologic examination. However, at times the histologic findings, including the degree of nuclear atypia, may be suggestive but not absolutely diagnostic of an HPV related neoplasm. HPV DNA sequences were detected in 63 and 56% of colposcopically visible vaginal and cervical lesions, respectively, that were diagnosed as condyloma or intraepithelial neoplasia. HPV DNA sequences were detected in 14 and 47% of vaginal and cervical lesions, respectively, that did not fulfill the histologic criteria of condyloma or intraepithelial neoplasia (i.e., "nondiagnostic"). When examining cervices from patients with no visible lesion and no recent history of an abnormal pap smear, 5.5% had detectable HPV DNA sequences. The histologic findings in this group were equivalent to the virus-negative cases and similar to the "nondiagnostic" cervical lesions. These findings suggest that the detection rate of HPV DNA in "nondiagnostic" tissues is dependent on the site and presence or absence of a visible lesion. The rate is similar in cervical lesions regardless of the histologic findings whereas it is less in vaginal lesions when the histologic criteria of condyloma or intraepithelial neoplasia are not detected.

Condylomatous carcinoma of the vulva with special reference to human papillomavirus DNA.

Nine cases of condylomatous carcinoma (squamous cell carcinoma arising in condyloma acuminatum) of the vulva were studied for their clinical history, histopathology, and presence of human papillomavirus (HPV) DNA. Condylomatous carcinoma occurred primarily in an elderly population with a mean age of 70 years. There was an antecedent history of vulvar condyloma in 77%, with a median of nine months before the documentation of an invasive lesion. The disease had a good prognosis, with few recurrences and no metastasis or deaths from the disease. Human papillomavirus DNA was demonstrated to be present in 55% of these tumors by either filter or in situ hybridization techniques. Both HPV 6 and HPV 16 DNA were identified in an equal number of cases.

Transcriptional differences of the human papillomavirus type 16 genome between precancerous lesions and invasive carcinomas.

Human papillomavirus type 16 (HPV16) genome DNA and its transcripts in biopsied cervical neoplasias were analyzed by simultaneous extraction of DNA and RNA from one biopsied sample. Southern blot analysis revealed that 5 of 20 cervical intraepithelial neoplasias (CINs) contained HPV16 DNAs existing primarily as episomes and two of seven invasive carcinomas harbored HPV16 genome sequences integrated into the host DNA. Northern (RNA) blot analysis showed that the HPV16 genome sequences were transcriptionally active in the five CINs, as well as in the two invasive carcinomas. The pattern of HPV16-specific transcripts in the CINs was uniform, and the major transcripts were 4.2, 2.2, 1.6, and 1.4 kilobases in size. However, the pattern of HPV16-specific transcripts in the invasive carcinomas was variable and different from that in CINs, suggesting that the alteration of transcriptional pattern might play a key role in the development of malignancy.

Simultaneous presence of herpes simplex and human papilloma virus sequences in human genital tumors.

Herpes simplex virus (HSV) and human papillomavirus (HPV) sequences were analyzed in tumors of the female lower genital tract, by probing DNA from 13 intraepithelial and 30 invasive neoplastic lesions with radiolabelled HPV-16 and HPV-18 DNA as well as cloned fragments of HSV-2 DNA. Careful removal of stromal tissue from the pathological specimens allowed authentic tumor DNA to be processed. Normal genital tissue obtained from the patients and genital condylomata were included as internal controls. The presence of HPV-16 or 18 DNA was detected in 12/13 (92.3%) intraepithelial neoplasms and in 16/30 (53.3%) invasive carcinomas. No significant difference was detected in titer or frequency of antibodies to HPV group-specific antigen in sera from patients and controls. Hybridization to BgIII N fragment of HSV-2 DNA was detected in 4/13 (30.8%) intraepithelial neoplasms and 4/30 (13.3%) invasive carcinomas but in none of the control tissues. All the 8 samples harboring HSV-2 homologous sequences were also positive for HPV, supporting the hypothesis of a synergistic association between the 2 viruses. The hybridization analyses performed to study c-myc involvement in genital oncogenesis did not reveal c-myc amplification in either invasive or pre-invasive lesions.

[A comparative study of sensitivities of immunocytochemistry, koilocytosis and transmission electron microscopy in detecting HPV in the cervical condyloma and intraepithelial neoplasia as compared to DNA hybridization].

The sensitivity in detection of human papillomavirus (HPV) by immunocytochemistry, histological observation of koilocytosis and electron microscopy with reference to the results of Southern blot DNA hybridization were reviewed in 41 lesions (37 patients) of cervical and vaginal condylomata acuminata and intraepithelial neoplasia. HPV DNA was demonstrated in all but one lesion of moderate dysplasia (98%). HPV capsid antigens were demonstrated by immunocytochemistry in approximately 60% of the lesions of condyloma and mild dysplasia. Koilocytosis was present in approximately 90% of the lesions of condyloma and mild dysplasia as well. But the rate of HPV detection by immunocytochemistry and by observation of koilocytosis declined markedly in severe lesions; immunocytochemistry was positive in 22% and koilocytosis was present in 60% in the lesions of moderate dysplasia: positive in 17% and 33% respectively in the lesions of severe dysplasia: 0% and 20% respectively in the lesions of CIS. Intranuclear virus-like particles were observed in all of 8 lesions subjected to electron microscopy. The negative findings of immunocytochemistry and koilocytosis in advanced cervical intraepithelial neoplasia (CIN) have very little significance in relation to the actual presence of HPV in these lesions, although they may be useful in detecting HPV in condyloma acuminatum and mild dysplasia (over 50%).

Sensitivity of the cytologic diagnosis of cervical condyloma in comparison with HPV-DNA hybridization studies.

The cytologic diagnosis of cervical condyloma is based on criteria developed over the last 10 years. It has now become possible to document the presence of human papilloma virus (HPV) DNA directly in cervical swabs by the highly sensitive technique of DNA filter hybridization in situ. The purpose of this article is to evaluate critically the empirically established cytologic criteria of condyloma by comparing them with HPV-DNA hybridization studies in the same material. The results of this study indicate that "classic" koilocytosis and dyskeratocytosis are not highly sensitive criteria for the presence of HPV infection, identifying only 15% of the HPV-DNA-positive cases correctly. In an attempt to improve the sensitivity of the cytologic diagnosis of HPV infections, a panel of nine "nonclassic" criteria was evaluated. The five most valuable signs were "mild koilocytosis," mild dyskeratocytosis," hyperchromatic nuclei, bi- and multinucleation, and cleared cytoplasm. Using these criteria in combination, statistically discriminant analysis could correctly identify 84% of the HPV-positive group.

Detection and localization of human papillomavirus DNA in human genital condylomas by in situ hybridization with biotinylated probes.

We have examined the distribution of human papillomavirus (HPV) DNA in paraffin sections of humans warts by in situ hybridization with biotin-labeled DNA probes. Recombinant plasmid DNAs (HPV-1, -6, -11, -16) were labeled by nick translation with biotinylated deoxyuridine triphosphate. Paraffin sections were hybridized with the probes for 18 h in stringent or non-stringent conditions, and DNA-DNA hybrids were detected by immunocytochemistry. Paraffin sections of warts were also examined for the presence of HPV capsid antigen with the avidin-biotin peroxidase complex method for immunocytochemistry. HPV DNA was detected and localized in paraffin sections from a plantar wart, a laryngeal papilloma, and seven anogenital condylomas. The specific HPV type present in each lesion was determined by hybridization under stringent conditions with the homologous DNA probe. The papillomas were found to contain many more cells with replicating virus DNA, as demonstrated by in situ hybridization, than was apparent from the number of cells containing detectable virus antigen. In situ hybridization with biotin-labeled probes is an effective technique for the identification of HPV infection in routinely collected and processed tissue specimens.

Epidermal growth factor receptors in different skin tumors.

Specific binding of 125I-labeled epidermal growth factor (EGF) was measured in 62 skin tumors of different severity. Within a group of 28 benign tumors, 11 of 15 condylomata acuminata were receptor positive, whereas the investigated mesenchymal tumors and normal skin as a control were receptor negative. 6 of 18 basal cell epitheliomas bound EGF specifically. In the group of precancerous and malignant skin tumors, 7 of 8 squamous cell carcinomas had the highest number of EGF binding sites and a high affinity state, whereas 5 malignant melanomas were receptor negative. The clinical relevance of these findings is not yet clear due to the short follow-up of the patients.

Differences of expression of cytokeratin polypeptides in various epithelial skin tumors.

In normal skin, cytokeratin polypeptides are expressed in different cell-type-specific patterns, in the keratinocytes of the different epidermal cell strata as well as in different lateral epithelial domains. Using light microscopically controlled microdissection of defined regions from frozen sections of biopsies, we have prepared cytoskeletons of various benign and malignant keratinocyte-derived tumors of human skin and analyzed their cytokeratin polypeptide patterns by two-dimensional gel electrophoresis. Premalignant fibroepitheliomas and basal cell epitheliomas display a relatively simple cytokeratin pattern (cytokeratins nos. 5, 14, 15, and 17). Pseudocarcinomatous hyperplasia, some squamous cell carcinomas, and a certain subtype of condylomata acuminata present a hair-follicle-like pattern (nos. 5, 6, 14, 16, 17). In addition to these components, variable, mostly low amounts of cytokeratins nos. 1 (Mr 68,000), and 11 are detected in most squamous cell carcinomas, in keratoacanthomas, verruca vulgaris, and another type of condylomata acuminata. In molluscum contagiosum, verruca plana, solar keratosis, and seborrheic keratosis, the cytokeratin expression is shifted more towards the normal epidermal pattern (polypeptides nos. 1, 2, 5, 10, 11, 14, 15 and traces of nos. 6 and 16 in the latter two tumors). No tumor-specific cytokeratins have been found. We conclude that keratinocyte-derived skin tumors contain various combinations of cytokeratins of the subset typical for normal keratinocytes of skin, but no cytokeratins typical for internal, simple epithelia. Different groups of tumors can be distinguished by their specific cytokeratin patterns. Possible applications of cytokeratin typing in clinical diagnosis are discussed.

Human papillomavirus infection and cervical intraepithelial neoplasia: histopathology and DNA content.

To evaluate the reliability of diagnostic criteria for separating intraepithelial squamous lesions into low- and high-risk categories, 25 lesions of the cervix were diagnosed as flat condyloma, atypical immature metaplasia, or cervical intra-epithelial neoplasia with koilocytosis based on well-defined histologic criteria. The presumption was that flat condyloma and atypical immature metaplasia would be diploid/polyploid as compared to low-grade cervical intraepithelial neoplasia, which should be aneuploid. Using the major histologic parameter of the presence or absence of abnormal mitoses to distinguish the low- and high-risk lesions, it was found that all of five typical flat condylomas were diploid/polyploid and seven of eight atypical immature metaplastic lesions were diploid/polyploid; 11 of 13 cervical intraepithelial neoplasms with koilocytosis, however, were aneuploid. An additional histologic parameter of anisocytosis (variation in nuclear size) appeared much less reliable for segregating these lesions than the nature of the mitoses. Lesions for which ploidy values were particularly difficult to predict were extremely well-differentiated koilocytotic lesions with occasional abnormal mitoses. Whether these are true polyploid lesions in which the abnormal mitoses are a response to the virus, or whether they are very early aneuploid lesions that cannot be confirmed by microspectrophotometry, remains to be determined.

Immunocytochemical localization of keratin in normal, dysplastic and neoplastic cervical epithelium.

The PAP immunocytochemical technique utilizing specific keratin antibody was applied to paraffin sections from 36 cervical biopsies. Normal squamous epithelium and condylomas had similar patterns of keratin production with intense staining of intermediate and upper layers, while basal cells remained negative. Dysplasia, carcinoma in situ and infiltrating squamous carcinoma showed uneven distribution of keratin with the least amount seen in the areas with high mitotic rate and anaplasia. All large cell squamous carcinomas demonstrated presence of significant amounts of keratin. Squamous carcinomas of the small cell type were essentially keratin-free.

Histochemical studies of lectin binding patterns in keratinized lesions, including malignancy.

Histochemical detection of lectin binding was carried out using the HRP-conjugated lectin method in hyperkeratinized lesions including leukoplakia, carcinoma in situ, Paget's disease, keratoacanthoma, and condyloma acuminatum. The lectins used for demonstrating sugar residues were: Con A (hexose), PNA and RCA-1 (Gal), DBA and SBA (GalNAc), UEA -1 (Fuc), and WGA (GlcNAc). Lectin binding in normal squamous epithelium showed regional distribution patterns of keratinized, spinous and basal layer types. Histochemical localization of lectin binding was generally at the cellular surface and in the intercellular substance and sometimes in the cytoplasm of normal epithelial cells. Dysplastic cells or carcinoma cell, in contrast, displayed a loss of cellular surface and intercellular staining. Paget's cells were devoid of lectin staining. In keratoacanthoma and condyloma specimens, spinous cells, which were PAS-positive, showed an intense PA/Con A-HRP staining and moderate binding by other lectins, which was somewhat decreased when compared with that in the surrounding intact epithelium. The cytochemical distribution of epithelial lectin binding might be indicative of the expression of normal stratification and keratinocytic differentiation , and the disappearance of this typical epithelial pattern may suggest severe dysplasia and malignancy.

Human papillomavirus in cervical condylomata. An immunohistochemical study.

Twenty-seven cervical condylomata were studied by morphological and immunohistochemical methods (peroxidase-antiperoxidase technique according to Sternberger, with some modifications). The antiserum was obtained from rabbits immunized by human papillomavirus virions; 37% of condylomata stained positively and the koilocytotic cells showed a dark brown nuclear stain. This technique, not particularly useful for diagnostic purposes, could be employed to obtain better understanding of the natural history of these cervical neoplasias.

Analysis of human genital warts (condylomata acuminata) and other genital tumors for human papillomavirus type 6 DNA.

32P-labelled cloned HPV 6 DNA was used as probe to analyze human genital tumors for DNA sequences homologous to HPV 6 DNA. Ninety three percent of all condylomata acuminata (41 out of 44) were found to harbor HPV 6 DNA. Of the remaining three, one contained HPV 1 DNA. No papillomavirus DNA was identified in the two other tumors. All three invasively growing giant condylomata acuminata (Buschke-Löwenstein tumors) investigated also contained HPV 6 DNA. Two out of six atypical condylomata of the cervix hybridized with HPV 6 DNA under stringent conditions, one only under conditions of low stringency. All DNA preparations from malignant tumors studies (54 cervical carcinomas, 10 penile carcinomas, two vulvar carcinomas) failed to anneal with HPV 6 DNA, even under conditions of low stringency. Although all HPV 6-positive condylomata acuminata analyzed in this study revealed HPV 6 DNA of regular molecular weight (5.1 x 10(6)), two of the Buschke-Löwenstein tumors, as well as one of the two positive atypical condylomata of the cervix, contained HPV 6 DNA with a remarkable size classes occurred in a supercoiled form without evidence for integration into host cell DNA.