Human papillomavirus infection in women with multicentric squamous cell neoplasia Anna Marie Beckmann, PhD," b Randy Acker, MS,. Audrey E. Christiansen, BS,. and Karen J. Sherman, PhD" c Seattle, Washington Tissues from 32 women with multicentric squamous cell neoplasia of the anogenital region (72 anatomically distinct lesions at the cervix, vagina, vulva, perineum, or anus) were tested for the presence of human papillomavirus with the polymerase chain reaction or in situ hybridization. All the women had invasive carcinomas or grade 3 intraepithelial neoplasia lesions at a minimum of one site and one or two squamous cell lesions at another site(s). Human papillomavirus was detected in all of the multicentric lesions in 87.5% (28/32) of the women and in at least one lesion in 12.5% (4/32). In the 28 women with detectable human papillomavirus at all sites, 61 % (17/28) had the same virus type(s) at all sites (types 6, 16, 6 and 16, 33) and 25% (7/28) had 6 or 16 at one site and both viruses at the other site(s). Four women (15%) had different virus patterns in the separate lesions. (AM J OBSTET GVNECOL 1991 ;165: 1431-7.)
Key words: Human papillomavirus, multicentric squamous cell neoplasia, polymerase chain reaction Numerous studies have reported that infection with human papillomavirus (HPV) is associated with squamous cell neoplasia of the lower female genital tract. The strongest evidence implicating HPV infection in the pathogenesis of genital tract malignancies comes from investigations of human tissues with molecular hybridization techniques. Many such studies have demonstrated that most squamous cell intraepithelial neoplasias and invasive carcinomas of the cervix and vulva contain HPV deoxyribonucleic acid (DNA) or ribonucleic acid (RNA).1.3 Of the 60 HPV genotypes that have been identified by molecular analyses of human tissues, 22 have been found in anogenitallesions in both women and men" The spectrum of clinical conditions associated with HPV infection ranges from occult infections, where the viral DNA is present without cytologic or histologic evidence of infection, through condylomas to precancers and invasive carcinomas! HPV type 16 is the most common virus found in cervical squamous cell carcinomas and in high-grade intraepithelial lesions. 1·3 In contrast, HPV type 6 is found in the majority
From the Division ofPublic Health Sciences, Fred Hutchinson Cam",. Research Center," the Department of Pathology, School of Medicine,' and the Depa,'tment of Epidemiology, School of Public Health and Communit.~ Medicine,' University of Washington. Supported by the National Cancel' Institute (grants CA 47619, CA 50491, and CA 42792). Received for publication OctobeT 29,1990; revised March 4,1991; accepted April 9, 1991. Reprint requests: Anna Marie Beckmann, PhD, Fred Hutchinson Cancer Research Center, E602. 1124 Columbia St., Seattle, WA 98104. 6/1 /30077
of vulvar, penile, and anal condylomas and is infrequently detected in high-grade intraepithelial lesions or invasive carcinomas.!' 2 However, some studies have shown HPV type 6 to be present in vulvar squamous cell cancers.' Other virus types (HPV types 18,31,33, 35, and 39) also are found in genital lesions but are less common than HPV types 16 and 6.' The presence of HPV nucleic acids in squamous cell carcinomas of the cervix, vulva, vagina, and anus is consistent with a role for these viruses as etiologic agents in this morphologic type of anagenital cancer. The epidemiologic basis for considering cervical cancer as a sexually transmitted disease is well established. s Recent epidemiologic studies have demonstrated that vulvar,· vaginal,' and anal"' 9 cancers also have some similar risk factors with sexually transmitted diseases (number of sex partners, age at first intercourse, sexual practices). Although the proportion of squamous cell lesions that contain HPV is similar among these anatomic sites, the corresponding incidence of cancer varies markedly. In the United States the age-adjusted incidence of cervical cancer (8.0/ 100,000 in 1981 through 1985) is five to 10 times higher than that for carcinoma of the vulva (1.5), vagina (0.7), or anus (0.9) during the same time period. '0 Also, the behavioral risk factors (number of sex partners, age at first intercourse, smoking history) for development of squamous cell carcinoma at each of these sites are somewhat different. 5. 9 In this study we have investigated this apparent differential pathogenicity of HPV infection in the female genital tract. We chose a study group of women with multicentric squamous cell neoplasia to determine if 1431
November 1991 Am J Obstet Gynecol
1432 Beckmann et al.
Table I. Summary of 72 lesions from 32 women with multicentric neoplasia by site, histopathologic type, and HPV result Second lesion
Primary lesion* Subject No.
Age (yr)
Site
I 2 3 4 5 6 7 8 9 10 II 12 13
66 60 32 45 36 26 23 61 62 62 52 40 34
Cervix Cervix Cervix Cervix Cervix Cervix Cervix Vagina Vagina Vagina Vulva Vulva Vulva
14 15 16 17 18 19 20 21 22 23 24 25 26 27
41 69 28 36 59 66 60 25 36 40 35 55 42 41
28 29 30 31 32
Pathologic condition
I
I
HPVt
Site
16 35 16 6, 16 6, 16 16 (16) 16 16 16 16 16 6
Vagina Vagina Vagina Vagina Vulva Vulva Vulva Cervix Vulva Vulva Cervix Cervix Cervix
Vulva Vulva Vulva Vulva Vulva Vulva Vulva Vulva Vulva Vulva Vulva Vulva Vulva Vulva
M-Inv Inv Inv M-Inv Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade IN Grade 3 IN M-Inv Grade 3 IN Grade 3 IN/condyloma Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Gt"ade 3 IN Gt"ade 3 IN
16 Neg (16) 16 33 6, 16 33 6, 16 (16) 6, 16 (16) 16 16
Cervix Cervix Cervix Cervix Vagina Vagina Vagina Vagina Vagina Vagina Vagina Vagina Vagina Vagina
42 49 36 40
Vulva Vulva Vulva Vulva
Grade Grade Grade Grade
6, 16 16 6, 16 16
Vagina Perineum Anus Anus
32
Vulva
Grade 3 IN
3 3 3 3
IN IN IN IN
UK
16
Anus
I
Pathologic condition
Third lesion
I
I
Pathologic condition
I
HPV
Site
lnv Inv lnv Grade 3 IN Grade 3 IN Inv Condyloma Grade 3 IN Grade 3 IN Condyloma M-Inv Grade 3 IN Grade 3 IN /condyloma Grade 3 IN Grade 3 IN Grade 2 IN Grade I IN Grade 3 IN Grade 3 IN Grade 3 IN Condyloma Grade 2 IN Grade 1 IN Grade I IN Grade I IN Grade 3 IN Grade 3 IN
16 Neg 16 16 6, 16 16 6, 16 16 16 6 Neg
Vagina Anus Anus
Grade I IN Condyloma Condyloma
6, 16 33 6, 16
Vagina
Grade 3 IN
35
Anus Anus
Condyloma Grade 3 IN Grade 3 IN Grade 3 IN / condyloma Grade 3 IN
6 6,16 6, 16 6, 16
Anus
Grade 3 IN Grade 3 IN /condyloma Condyloma
6, 16
Cervix
Grade I IN
HPV
UK 6
16 Neg (16) 16 6 6, 16
UK
33 6, 16 Neg 6, 16 (16) 16 6,16
16 6,16 6
6
M-Inv, Microinvasive squamous cell carcinoma; Inv, invasive squamous cell carcinoma; IN, intraepithelial neoplasia. *Primary lesion defined as that tumor for which patient was identified for case-control studies. tHPV types in parentheses were determined from in situ hybridization analyses; Neg, negative; UK, unidentified type (not 6, 11,16,18,31,33,35,39, or 42).
infection with viruses of different oncogenic potentials (i.e., HPV type 6 or 16) was associated with individual squamous cell lesions at different epithelial surfaces in the female genital tract. Our objectives were (I) to determine if the same virus type was present in carcinoma in situ or invasive squamous cell carcinomas at different anogenital sites in the same individual and (2) to determine if different virus types were present in lesions of different malignant potentials at various epithelial surfaces in the same individual. Material and methods
Subject tissues. Women were chosen for this study by surveys of pathology reports of individuals enrolled in case-control studies of cervical, vaginal, and vulvar cancer being conducted in western Washington State (principal investigator, Janet R. Daling, PhD). Women were eligible for inclusion if they had been diagnosed
with either invasive squamous cell cancer or grade 3 intraepithelial neoplasia at one site (cervix, vulva, or vagina) and had concurrent invasive cancer, grade 3 intraepithelial neoplasia, grade 1 or 2 intraepithelial neoplasia, or HPV-related morphologic changes in at least one other anogenital site (cervix, vulva, vagina, perineum, or anus). The study group included 22 women with invasive cancer or grade 3 intraepithelial neoplasia at more than one site and 10 women with grade 3 intraepithelial neoplasia at one site and with lesions at other sites graded I or 2 intraepithelial neoplasia or HPV-related morphologic changes. These 32 women had lesions at 72 distinct anatomic sites (Table I) and were chosen for this study because paraffin-embedded fixed tissues were available from all of the lesions. Histologic criteria. Anogenital squamous cell lesions were classified in three groups for this study: (1) in-
Volume 165 Number 5, Part I
vasive or microinvasive squamous cell carcinoma, (2) cervical, vulvar, vaginal, or anal grade 3 intraepithelial neoplasia, and (3) other intraepithelial neoplasia lesions or morphologic changes indicative of HPV infection (condyloma or koilocytotic atypia at the cervix). Histopathologic diagnoses were determined from pathology reports. Detection of HPV nucleic acids. The polymerase chain reaction was used to detect HPV DNA in 66 of the 72 lesions. Many of the lesions had already been analyzed by in situ hybridization before the development of the polymerase chain reaction method, and all of the available paraffin-embedded material had been used for in situ hybridization tests. Consequently, for this study adequate amounts of remaining tissue for polymerase chain reaction tests were not available for six of the 72 lesions, and in situ hybridization with asymmetric RNA probes was the only means for detection of HPV nucleic acids in these six lesions. Polymerase chain reaction. Paraffin-embedded tissues were prepared for polymerase chain reaction analyses as described by Wright and Manos,II and the polymerase chain reaction was performed on cellular DNA extracted from single 6 IJ.m sections. Initially, tissue DNAs were amplified with oligonucleotide primers specific for the E6/E7 region of the HPV type 6 or 16 genomes. I2 All positive results obtained with primers from the E6/E7 region were confirmed by amplification of a second region of the HPV type 6 or 16 genome; these were the El open reading frame (ORF) for HPV type 6 and the Ll ORF for HPV type 16. For the HPV type 6 El ORF, oligomer primers were GTTGCTGTGGATGTGACAGCAACGT (nucleotides 699 to 723) and CCTGACCATCTCCCCCCATTTTCCGG (nucleotides 1321 to 1297). Both to confirm HPV 16 positive results and also to reanalyze tissue DNAs that initially tested negative by polymerase chain reaction with HPV type 6 or 16 E6/E7 primers, a set of degenerate consensus primers were used for the amplification reactions." These primers amplify a region of sequences common to at least 25 distinct genital HPV types, 13 and, after RsaI digestion of amplification products, HPV types can be distinguished by the pattern of restriction enzyme fragments. 13.14 To assure maximum sensitivity and specificity of the polymerase chain reaction, all reaction products were electrophoresed through 4.0% agarose gels, transferred to nitrocellulose by Southern's method,I5 and hybridized with type-specific or consensus probes derived from viral DNA sequences internal to the primer pairs. This approach for analysis of HPV genomes has been previously described by our group.I2. 14 For the HPV type 6 El ORF the probe sequence was ATTATGCGACTGTGCAGGACCTAAAACGAAAGTATTTAGG (nucleotides 1055 through 1093); the probe sequences for the HPV types 6 and 16 E6/E7 ORF and
Human papillomavirus in multicentric neoplasia
1433
the 1.1 consensus sequence have been described elsewhere. I2 . 14 The amplification reactions allow consistent and reproducible detection of 0.1 pg of plasmid viral DNA after hybridization, representing about 10 3 copies of the HPV genome. In situ hybridization. In situ hybridization used asymmetric RNA probes from the E6/E7 region of the HPV types 6 or 16 genomes essentially as described by Stoler and Broker. I6 Briefly, sections were dewaxed, digested with proteinase K (15 IJ.g/ml), and acetylated (0.25% acetic anhydride in 0.1 mol/I. triethanolamine, pH 8.0). Sections were hybridized overnight at 49° C, in stringent conditions, with 1.0 to 5.0 X 107 disintegrations/min of tritium-labeled HPV type 6 or 16 RNA. Antisense RNA probes (complementary to the messenger RNA made in vivo) were generated from HPV type 6 or 16 E6/E7 ORFs cloned into the transcription vector pSP6-T7-19 (Bethesda Research Laboratories, Inc., Gaithersburg, Md.). Hybridized sections were washed, dehydrated, coated with autoradiographic emulsion (Kodak NTB-2), and exposed for 4 weeks. Tissues were stained with hematoxylin for microscopic evaluation. Results
Anatomic distribution of lesions. All 32 women in this study had two or three squamous cell lesions at various sites throughout the anogenital region; the site and histopathologic type of these neoplasias are detailed in Table 1. Lesions at the vulva and vagina (n = 10; age range 26 to 66; median, 57) were the most common in this study. Four of these women had grade 3 vulvar and vaginal intraepithelial neoplasia, five had grade 3 vulvar intraepithelial neoplasia and less severe vaginal lesions, and one had grade 3 vaginal intraepithelial neoplasia and a vulvar condyloma. Six women had lesions at the cervix and vulva (age range, 28 to 52; median, 38). Four of these women had invasive cancer or grade 3 intraepithelial neoplasia at both sites, and two had grade 3 vulvar intraepithelial neoplasia and less severe cervical lesions. Five women had lesions at the cervix and vagina (age range, 32 to 66; median, 60); all had either invasive cancer or grade 3 intraepithelial neoplasia at both sites. Two women had grade 3 intraepithelial neoplasia lesions at the vulva and anus (ages 36 and 40), and one women had grade 3 vulvar intraepithelial neoplasia and a distinct grade 3 intraepithelial neoplasia lesion on the perineum (age 49). Eight women had lesions at three sites. Three individuals had involvement at the cervix, vulva, and anus (ages 23, 26, 32); three had lesions at the vulva, vagina, and anus (ages 28, 41, 42); and two had lesions at the cervix, vagina, and vulva (ages 36, 69). Although the numbers of women in each group outlined above are small, it appears that women with vag-
1434 Beckmann et al.
November 1991 Am J Obstet Gynecol
Table II. Detection of HPV DNA in 72 lesions from 32 women with multicentric squamous neoplasia Anatomic site and histopathologic type of lesion
Cervix Grade 1 or 2 CIN* Grade 3 CIN Invasive cancer Vulva Condyloma Grade 3 VIN Invasive cancer Vagina Condyloma Grade 1 or 2 VAINt Grade 3 VAIN Invasive cancer Anus Condyloma Grade 3 AIN Perineum Grade 3 IN
HPV type in positive specimens
6
I
16
I
6 and 16
I 33
I 35
I
Unknown
HPV DNA negative
0
2
0
0
0
0
1 0
3 2
2 1
0 0
0 1
1 0
1 1 0
0 12 2
1 6 0
0 2 0
0 0 0
0 1 0
0
0
3
Total No.
3 8 5
0 1 0
I
% 100 87.5 80
2 23 2
100 96 100
7
86
1 0
5 2
2 0
0 0
1 0
1 0
0 1
10 3
100 67
1 0
0 1
4
1
1 0
0 0
0 0
0 0
3 5
100 100
0
0
0
0
0
0
100
CIN, Cervical intraepithelial neoplasia; VIN, vulvar intraepithelial neoplasia; VAIN, vaginal intraepithelial neoplasia; AIN, anal intraepithelial neoplasia. *Includes one grade 2 cervical intraepithelial neoplasia with HPV type 16. tInciudes one grade 2 vaginal intraepithelial neoplasia with HPV types 6 and 16.
inal and either cervical or vulvar lesions were older than women with lesions involving the vulva and either the cervIx or anus. Detection of HPV nucleic acids. A total of 276 tissue blocks comprising 72 anatomically distinct lesions from 32 women were analyzed by polymerase chain reaction or in situ hybridization. Sixty-seven of the 72 lesions (93%) contained HPV DNA or RNA of types 6,16,33, 35, or unidentified HPV types (Tables I and II). HPV nucleic acids were more frequently detected in anal or perineal (100%; n = 9) and vulvar (96%; 26/27) lesions than in vaginal (90%; 18/20) or cervical (88%; I41l6) lesions (Table II). When the histopathologic type of all lesions is considered, eight of nine (89%) condylomas contained detectable HPV DNA, six of six lesions (100%) graded 1 or 2 intraepithelial neoplasia were positive, 45 of 47 (96%) grade 3 intraepithelial neoplasia lesions were positive, and eight of 10 (80%) invasive cancers were positive (Table II). The negative condyloma was a flat lesion in the vagina of a woman with a grade 3 vaginal intraepithelial neoplasia lesion that contained HPV type 16 DNA. The two negative grade 3 intraepithelial neoplasia lesions were at the vulva and cervix of a woman who also had grade 3 vaginal intraepithelial neoplasia that contained HPV type 35 DNA. Of the two negative invasive cancers, one was a second primary cancer of the vagina in a woman with invasive cancer of the cervix that contained HPV type 35 DNA, and the second was an invasive cervical cancer in a woman with invasive vulvar cancer that contained HPV type 16 DNA. All of the HPV-negative tissues had been
tested with the polymerase chain reaction; the less sensitive in situ hybridization method was not responsible for these negative results. Also, these tissues were tested with primer pairs for the human l3-g10bin gene. IS. 17 All the HPV-negative lesions gave positive results with the l3-g10bin primers (data not shown), demonstrating the adequacy of the DNA preparations for polymerase chain reaction analyses. Distribution of HPV types by anatomic location of lesions. In the 67 lesions that contained detectable HPV nucleic acids, HPV type 6 or 16 was found in tissues from all anatomic sites, whereas HPV type 33 was seen only at the vulva, vagina, and anus, and HPV type 35 only at the cervix and vagina (Table II). HPV type 16 infection was more common at the cervix, vulva, and vagina than at the anus, whereas mixed infection with HPV types 6 and 16 was most common at the anus. At all sites, as the degree of neoplasia increased, so did the proportion of HPV type 16-positive lesions (Table III). Double infections with HPV types 6 and 16 were approximately twice as common in noninvasive lesions; only one invasive cancer, at the cervix, contained both types 6 and 16 (Table II). HPV type 16 was the most common virus type found in invasive cancers or grade 3 intraepithelial neoplasia lesions of the cervix (51l1), vulva (14/24), or vagina (7 1l2). However, women with anal intraepithelial neoplasia were more likely to have a dual infection with HPV types 6 and 16 (4/5) than an infection with HPV type 16 alone (115) (Table II). Virus patterns in multicentric lesions. HPV nucleic acids were detected in all of the anatomically distinct
Human papillomavirus in multicentric neoplasia
Volume 165 Number 5. Part 1
1435
Table III. HPV type-specific disease association in virus-positive multicentric lesions HPV type in positive specimens Histopatlwlogic type of lesion
Condyloma. grade I or 2IN Grade 3 IN Invasive cancer
16
6 No.
4
I%
No.
I%
No.
29
3
21
5
7
21 6
47
15 I
3
o
o
75
33
6 and 16
I
%
No.
I%
No.
o
36
2
14
33 12.5
o
2
4
o
Unknown
35
I
%
No.
o
o
2
12.5
3
o
I
%
Total (No.)
o
14
7
45
o
8
IN, Intraepithelial neoplasia.
lesions in 87.5% (28/32) of the women in this study. Four women (subjects 2, 11, 15, and 23) had detectable HPV at only one site; three of these women had lesions at two sites and one had lesions at three sites (Table I). The proportion of women with HPV DNA or RNA at all sites was not dependent on whether lesions were present at two or three sites (87.5% positive in both groups; 21/24 and 7/8, respectively). In women with lesions at two sites, the proportion of individuals with HPV at both sites did not depend on the degree of neoplasia. For women with grade 3 intraepithelial neoplasia at two sites, 87.5% (14/16) were positive at both sites. For women with grade 3 intraepithelial neoplasia at one site and less severe neoplasia at the other involved site, 87.5% (7/8) were positive at all three sites. In women with lesions at three sites, HPV was found in five of six individuals (84%) with grade 3 intraepithelial neoplasia at all sites and in both women with one grade 3 intraepithelial neoplasia lesion and less severe neoplasia at the other two sites. In the 28 women with detectable HPV at all sites, 61 % (17/28) had the same virus type(s) at all sites and 25% (7/28) had HPV type 6 or 16 at one site and both viruses at the other site(s) (Table I). Four women (14%) had different virus patterns in the separate lesions (subjects 6, 10, 12, and 18; Table I). Occurrence of double HPV infections. Thirteen of the 32 women (41 %) with multicentric neoplasia and 22 of the 72 lesions (31 %) showed evidence of infection with HPV types 6 and 16. Six of the 32 women (19%) were infected with HPV type 6 and 16 at more than one site, and seven of 32 (22%) were infected with both types at only one site. The total proportion of women with both HPV type 6 and 16 infections was identical in those individuals with grade 3 intraepithelial neoplasia at multiple sites (41 %,9/22) and those with grade 3 intraepithelial neoplasia at one site and lesions less than grade 3 intraepithelial neoplasia at other sites (40%, 4/10). In the first group of women, four of 22 (18%) individuals had double infections at more than one site, and five of 22 (23%) had double infections at only one site. In the
second group, two of 10 women (20%) had double infections at more than one site and the same proportion had double infections at only one site.
Comment In this study of women with multicentric squamous cell neoplasia of the lower genital tract, all of the 32 women had molecular evidence of HPV infection at at least one site. Five lesions from four individuals lacked detectable HPV nucleic acids; in each instance either an invasive cancer or grade 3 intraepithelial neoplasia at another site was HPV positive. Failure to detect HPV DNA in genital squamous cell carcinomas by conventional hybridization techniques (i.e., Southern, dot blot, in situ) may reflect a sampling error (missing the lesion), the presence of small amounts of virus that would not be detected by the methods used, the presence of a virus type unrelated to the probes, or the absence of HPV in the tumor. In this study the lack of HPV in these lesions is unlikely to represent a sampling error as sections from the entire tumor were analyzed in all cases. The sensitivity of the polymerase chain reaction is such that viral DNA should be detectable if present; however, the late consensus primers may not detect all HPV types that can infect the genital tract. A more convincing explanation for these negative findings is that random nicks introduced into the viral DNA after storage of paraffin tissues removed the oligonucleotide primer binding sites. Other investigators have demonstrated random loss of amplification ability of the human l3-globin gene, which increases with the age of the paraffin-embedded tissue. 17 Alternately, the HPVnegative lesions may not contain any viral nucleic acids. Lower rates of HPV positivity than what was found in this study have been reported in previous studies of women with multicentric squamous cell genital lesions. Pilotti et al. 18 used in situ hybridization to study tissues from women with multicentric disease and found that 73% (19/26) were positive for HPV DNA. These women were selected by clinical observation of vulvar wart-like lesions and simultaneous lesions at other sites, and approximately one half of the women had grade 3 vulvar intraepithelial neoplasia and the other
1436 Beckmann et al.
half had condylomatous vulvar lesions. Bergeron et al. 19 found HPV DNA by Southern hybridization in 87.5% (21/24) of women with multiple condylomatous lesions of the anogenital region. Kulski et al.20 found evidence of HPV infection in 85% (109/128) of women with cervical abnormalities who also had vulvar tissues collected for analysis by filter in situ hybridization. The 100% patient positivity rate in this study is, most likely, a result of the use of polymerase chain reaction, a test that is many times more sensitive than other detection methods. However, when considering each lesion separately, we found lower rates of HPV detection in the cervices of affected individuals than were seen at other sites in the genital tract. These results are similar to those of Reid et al.,21 who found detection rates, in cervical and vulvar biopsy specimens that showed koilocytotic atypia or neoplasia, of 86% and 93%, respectively. Conceivably, a wider variety of HPV types infect the cervix than the vulva, and these types infecting the cervix are not detectable under our experimental conditions. The relative frequencies of vulvar, cervical, vaginal, or anal HPV infections occurring simultaneously, in any combination, are not known. Some studies have shown that in women with vulvar condyloma pr grade 3 vulvar intraepithelial neoplasia up to 60% also have cervical HPV infections or preneoplastic lesions. 22 -24 Conversely, in women diagnosed with HPV-related cervical lesions, more than one half also have vulvar lesions!O In 1984, Okagaki 25 proposed the term genital neoplasm papilloma syndrome to describe this clinical observation of synchronously or metachronously occurring multiple HPV infections or HPV-associated intraepithelial neoplasia lesions in the female genital tract. Also, epidemiologic studies have shown a significant association between the simultaneous detection or occurrence of cervical and vulvar cancers and squamous cell cancers at other anogenital sites!6.27 implicating a common etiologic agent. The patterns of viral distribution in the multicentric lesions examined in this study suggest that the majority of these neoplasias were the result of diffuse HPV infection. Four women had different HPV types found in different areas of the genital tract; of these, two had two lesions of the same histologic grade (grade 3 vulvar intraepithelial neoplasia and grade 3 vaginal intraepithelial neoplasia or grade 3 cervical intraepithelial neoplasia), and the other two had grade 3 vaginal intraepithelial neoplasia and a vulvar condyloma, or grade 3 cervical intraepithelial neoplasia, vulvar invasive cancer, and an anal condyloma. Our results are consistent with the idea that squamous cell malignancies of the lower female genital tract share a common etiology, of which one factor is HPV infection. Women with multicentric HPV infection, whether in-
November 1991 Am J Obstet Gynecol
volving benign condylomas, precancers, or invasive lesions, appear to have much higher rates of multiple infections than do groups of women with only singlesite involvement. In this study 13 of the 32 (41 %) women had mixed infections throughout the lower genital tract, and 22 of the 72 (31 %) individual lesions contained both HPV types 6 and 16 DNA. These results agree with other studies in which from 38% to 100% of women with multicentric lesions were infected with two or more HPV types.18-20.28 In contrast, studies of women with only cervical or vulvar lesions have shown low rates «5%) of multiple HPV infections. 2This study did not analyze the exact cellular location of each virus type present in a single lesion. Reid et al. 21 have shown that in individuals with multiple HPV infections the different virus types were localized in demarcated areas of differing histopathologic appearances. However, in our study most all of these lesions containing HPV types 6 and 16 DNA were of similar or identical histopathologic grades. The observation of multiple infections in women with synchronously occurring multicentric lesions suggests that this study group represents a selection of individuals in whom the effectiveness of the host response to HPV infection is inadequate. In this study different viruses infecting differing areas of the female genital tract were ruled out as a potential explanation for the observed variation in the occurrence of squamous cell cancers at the cervix, vulvaa, vagina, and anus. HPV type 16 infection was the common thread among intraepithelial neoplasia III or invasive lesions at all sites. Interestingly, simultaneous infections with HPV types 6 and 16 were more common at the vulva and perineal and perianal skin than at the cervix, perhaps indicating a site restriction of HPV type 6 infection to cutaneous epithelium. In any event, factors other than the infecting virus type appear to be responsible for the 15-fo1d increase of cervical cancer over cancers at the vulva or vagina or anus. The unique properties of the cervical transformation zone"9 may make it more susceptible to HPV-induced oncogenesis than the native squamous epithelium is elsewhere in the lower genital tract. We are currently investigating the hypothesis that the cervix is more likely to undergo malignant changes after HPV infection than is the vulva, vagina, or anus because of site-specific variations in the expression of the HPV transforming genes £6 and £7. We thank Drs. James McDougall and Denise Galloway for helpful and critical consultation. REFERENCES 1. Lorincz AT, Reid R. Association of human papillomavirus with gynecologic cancer. Curr Opinion Oncol 1989; 1: 12332. 2. PaavonenJ. Koutsky LA. Kiviat N. Cervical neoplasia and
Volume 165 Number 5, Part 1
3. 4. 5. 6. 7. 8. 9.
10. 11.
12. 13.
14.
15. 16.
other STD-related genital and anal neoplasia. In: Holmes KK, Mardh P-A, Sparling PF, Wiesner PJ, eds. Sexually transmitted diseases. ew York: McGraw-Hill, 1990:56192. zur Hausen H. Papillomaviruses in anogenital cancer as a model to understand the role of viruses in human cancers. Cancer Res 1989;49:4677-81. de-Villiers EM. Heterogeneity of the human papillomavirus group. J Virol 1989;63:4898-903. Brinton LA, FraumeriJF Jr. Epidemiology of uterine cervical cancer. J Chronic Dis 1986;39:1051-65. Brinton LA, Nasca PC, Mallin K, Baptiste MS, Wilbanks GD, Richart RM. Case-control study of cancer of the vulva. Obstet Gynecol 1990;75:859-66. Brinton LA, Nasca PC, Mallin K. et aI. Case-control study of in situ and invasive carcinoma of the vagina. Gynecol Oncol 1990;38:49-54. Daling JR. Weiss NS. Hislop TG, et aI. Sexual practices, sexually transmitted diseases, and the incidence of anal cancer. N EnglJ Med 1987;317:973-7. Holly EA, Whit!emore AS, Aston DA, Ahn DK, Nickoloff BJ, Kristiansen JJ. Anal cancer incidence: genital warts, anal fissure or fistula, hemorrhoids, and smoking. J Nat! Cancer Inst 1989;81:1726-31. National Cancer Institute. 1987 Annual cancer statistics review. Hyattsville, Maryland: US Department of Health and Human Services, 1988: NIH publication no. 88-2789. Wright DK, Manos MM. Sample preparation from paraffin-embedded tissues. In: Innis MA, Gelfand DH, Sninsky JJ, White TJ, eds. PCR protocols: a guide to methods and applications. San Diego: Academic Press, 1990: 153-8. Jenison SA, Yu XP, Valentine JM, et al. Evidence of prevalent genital-type human papillomavirus infections in adults and children. J Infect Dis 1990;162:60-9. Ting Y, Manos MM. Detection and typing of genital human papillomaviruses. In: Innis MA, Gelfand DH, Sninsky JJ, White TJ. eds. PCR protocols: a guide to methods and applications. San Diego: Academic Press, 1990:35667. Beckmann AM, Sherman KJ, Myerson DM. Daling JR, McDougallJK, Galloway DA. Comparative virologic studies of condylomata acuminata reveal a lack of dual infections with human papillomaviruses. J Infect Dis 1991; 163:393-6. Southern EM. Detection of specific sequences among DNA fragments separated by gel electrophoresis. ] Mol Bioi 1975;98:503-17. Stoler MH, Broker TR. In situ hybridization detection of human papillomavirus DNAs and messenger RNAs in genital condylomas and a cervical carcinoma. Hum Pathol 1986; 17: 1250-8.
Human papillomavirus in multicentric neoplasia
1437
17. Greer CE, Peterson SL, Kiviat N, et al. PCR amplification from paraffin-embedded tissues. Effects of fixative and fixation time. Am] Clin Pathol 1991;95:117-24. 18. Pilotti S, GuptaJ, Stefanon B, De Palo G, Shah DV, Rilke F. Study of multiple human papillomavirus-related lesions of the lower female genital tract by in situ hybridization. Hum Pathol 1989;20: 118-23. 19. Bergeron C, Ferenczy A, Shah KV, Naghashfar Z. Multicentric human papiUomavirus infections of the female genital tract: correlation of viral types with abnormal mitotic figures, colposcopic presentation, and location. Obstet Gynecol 1987;69:736-42. 20. KulskiJK, Demeter T, Rakoczy P. Sterrett GF, Pixley EC. Human papillomavirus coinfections of the vulva and uterine cervix. J Med Virol 1989;27:244-51. 21. Reid R, Greenberg M, Jenson AB. et aI. Sexually transmitted papillomaviral infections. 1. The anatomic distribution and pathologic grade of neoplastic lesions associated with different viral types. AM J OBSTET GYNECOL 1987:156:212-22. 22. Walker PG, Colley NY, Grubb C, Tejerina A, Oriel JD. Abnormalities of the uterine cervix in women with vulval warts. Br J Vener Dis 1983;59: 120-3. 23. Walker PG, Singer A, DysonJL, OrielJD. Natural history of cervical epithelial abnormalities with vulval warts. Br J Vener Dis 1983;59:327-9. 24. Pilotti S, Rilke F. Shah KV, Delle Torre G, De Palo G. Immunohistochemical and ultrastructural evidence of papillomavirus infection associated with in situ and microinvasive squamous cell carcinoma of the vulva. Am J Surg Pathoi 1984;10:751-61. 25. Okagaki T. Female genital tumors associated with human papillomavirus infection and the concept of genital neoplasm-papilloma syndrome (GENPS). Pathol Annu 1984;19:31-62. 26. Sherman KJ, Daling JR, Chu J, McKnight B, Weiss NS. Multiple primary tumours in women with vulvar neoplasms: a case-control study. Br J Cancer 1988;57:423-7. 27. Henson D, Tarone R. An epidemiologic study of the cervix. vagina and vulva based on the third National Cancer Survey in the United States. Obstet Gynecol 1977; 120:525-32. 28. McCance DJ, Clarkson PK, Dyson JL, Walker PC, Singer A. Human papillomavirus types 6 and 16 in multifocal intraepithelial neoplasias of the female lower genital tract. Br J Obstet Gynecol 1985;92:1093-100. 29. Burghardt E, Oster AG. Site and origin of squamous cervical carcinoma. A histomorphological study. Obstet GynecoI1983;62:117-27.
Key words: Human papillomavirus, multicentric squamous cell neoplasia, polymerase chain reaction Numerous studies have reported that infection with human papillomavirus (HPV) is associated with squamous cell neoplasia of the lower female genital tract. The strongest evidence implicating HPV infection in the pathogenesis of genital tract malignancies comes from investigations of human tissues with molecular hybridization techniques. Many such studies have demonstrated that most squamous cell intraepithelial neoplasias and invasive carcinomas of the cervix and vulva contain HPV deoxyribonucleic acid (DNA) or ribonucleic acid (RNA).1.3 Of the 60 HPV genotypes that have been identified by molecular analyses of human tissues, 22 have been found in anogenitallesions in both women and men" The spectrum of clinical conditions associated with HPV infection ranges from occult infections, where the viral DNA is present without cytologic or histologic evidence of infection, through condylomas to precancers and invasive carcinomas! HPV type 16 is the most common virus found in cervical squamous cell carcinomas and in high-grade intraepithelial lesions. 1·3 In contrast, HPV type 6 is found in the majority
From the Division ofPublic Health Sciences, Fred Hutchinson Cam",. Research Center," the Department of Pathology, School of Medicine,' and the Depa,'tment of Epidemiology, School of Public Health and Communit.~ Medicine,' University of Washington. Supported by the National Cancel' Institute (grants CA 47619, CA 50491, and CA 42792). Received for publication OctobeT 29,1990; revised March 4,1991; accepted April 9, 1991. Reprint requests: Anna Marie Beckmann, PhD, Fred Hutchinson Cancer Research Center, E602. 1124 Columbia St., Seattle, WA 98104. 6/1 /30077
of vulvar, penile, and anal condylomas and is infrequently detected in high-grade intraepithelial lesions or invasive carcinomas.!' 2 However, some studies have shown HPV type 6 to be present in vulvar squamous cell cancers.' Other virus types (HPV types 18,31,33, 35, and 39) also are found in genital lesions but are less common than HPV types 16 and 6.' The presence of HPV nucleic acids in squamous cell carcinomas of the cervix, vulva, vagina, and anus is consistent with a role for these viruses as etiologic agents in this morphologic type of anagenital cancer. The epidemiologic basis for considering cervical cancer as a sexually transmitted disease is well established. s Recent epidemiologic studies have demonstrated that vulvar,· vaginal,' and anal"' 9 cancers also have some similar risk factors with sexually transmitted diseases (number of sex partners, age at first intercourse, sexual practices). Although the proportion of squamous cell lesions that contain HPV is similar among these anatomic sites, the corresponding incidence of cancer varies markedly. In the United States the age-adjusted incidence of cervical cancer (8.0/ 100,000 in 1981 through 1985) is five to 10 times higher than that for carcinoma of the vulva (1.5), vagina (0.7), or anus (0.9) during the same time period. '0 Also, the behavioral risk factors (number of sex partners, age at first intercourse, smoking history) for development of squamous cell carcinoma at each of these sites are somewhat different. 5. 9 In this study we have investigated this apparent differential pathogenicity of HPV infection in the female genital tract. We chose a study group of women with multicentric squamous cell neoplasia to determine if 1431
November 1991 Am J Obstet Gynecol
1432 Beckmann et al.
Table I. Summary of 72 lesions from 32 women with multicentric neoplasia by site, histopathologic type, and HPV result Second lesion
Primary lesion* Subject No.
Age (yr)
Site
I 2 3 4 5 6 7 8 9 10 II 12 13
66 60 32 45 36 26 23 61 62 62 52 40 34
Cervix Cervix Cervix Cervix Cervix Cervix Cervix Vagina Vagina Vagina Vulva Vulva Vulva
14 15 16 17 18 19 20 21 22 23 24 25 26 27
41 69 28 36 59 66 60 25 36 40 35 55 42 41
28 29 30 31 32
Pathologic condition
I
I
HPVt
Site
16 35 16 6, 16 6, 16 16 (16) 16 16 16 16 16 6
Vagina Vagina Vagina Vagina Vulva Vulva Vulva Cervix Vulva Vulva Cervix Cervix Cervix
Vulva Vulva Vulva Vulva Vulva Vulva Vulva Vulva Vulva Vulva Vulva Vulva Vulva Vulva
M-Inv Inv Inv M-Inv Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade IN Grade 3 IN M-Inv Grade 3 IN Grade 3 IN/condyloma Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Grade 3 IN Gt"ade 3 IN Gt"ade 3 IN
16 Neg (16) 16 33 6, 16 33 6, 16 (16) 6, 16 (16) 16 16
Cervix Cervix Cervix Cervix Vagina Vagina Vagina Vagina Vagina Vagina Vagina Vagina Vagina Vagina
42 49 36 40
Vulva Vulva Vulva Vulva
Grade Grade Grade Grade
6, 16 16 6, 16 16
Vagina Perineum Anus Anus
32
Vulva
Grade 3 IN
3 3 3 3
IN IN IN IN
UK
16
Anus
I
Pathologic condition
Third lesion
I
I
Pathologic condition
I
HPV
Site
lnv Inv lnv Grade 3 IN Grade 3 IN Inv Condyloma Grade 3 IN Grade 3 IN Condyloma M-Inv Grade 3 IN Grade 3 IN /condyloma Grade 3 IN Grade 3 IN Grade 2 IN Grade I IN Grade 3 IN Grade 3 IN Grade 3 IN Condyloma Grade 2 IN Grade 1 IN Grade I IN Grade I IN Grade 3 IN Grade 3 IN
16 Neg 16 16 6, 16 16 6, 16 16 16 6 Neg
Vagina Anus Anus
Grade I IN Condyloma Condyloma
6, 16 33 6, 16
Vagina
Grade 3 IN
35
Anus Anus
Condyloma Grade 3 IN Grade 3 IN Grade 3 IN / condyloma Grade 3 IN
6 6,16 6, 16 6, 16
Anus
Grade 3 IN Grade 3 IN /condyloma Condyloma
6, 16
Cervix
Grade I IN
HPV
UK 6
16 Neg (16) 16 6 6, 16
UK
33 6, 16 Neg 6, 16 (16) 16 6,16
16 6,16 6
6
M-Inv, Microinvasive squamous cell carcinoma; Inv, invasive squamous cell carcinoma; IN, intraepithelial neoplasia. *Primary lesion defined as that tumor for which patient was identified for case-control studies. tHPV types in parentheses were determined from in situ hybridization analyses; Neg, negative; UK, unidentified type (not 6, 11,16,18,31,33,35,39, or 42).
infection with viruses of different oncogenic potentials (i.e., HPV type 6 or 16) was associated with individual squamous cell lesions at different epithelial surfaces in the female genital tract. Our objectives were (I) to determine if the same virus type was present in carcinoma in situ or invasive squamous cell carcinomas at different anogenital sites in the same individual and (2) to determine if different virus types were present in lesions of different malignant potentials at various epithelial surfaces in the same individual. Material and methods
Subject tissues. Women were chosen for this study by surveys of pathology reports of individuals enrolled in case-control studies of cervical, vaginal, and vulvar cancer being conducted in western Washington State (principal investigator, Janet R. Daling, PhD). Women were eligible for inclusion if they had been diagnosed
with either invasive squamous cell cancer or grade 3 intraepithelial neoplasia at one site (cervix, vulva, or vagina) and had concurrent invasive cancer, grade 3 intraepithelial neoplasia, grade 1 or 2 intraepithelial neoplasia, or HPV-related morphologic changes in at least one other anogenital site (cervix, vulva, vagina, perineum, or anus). The study group included 22 women with invasive cancer or grade 3 intraepithelial neoplasia at more than one site and 10 women with grade 3 intraepithelial neoplasia at one site and with lesions at other sites graded I or 2 intraepithelial neoplasia or HPV-related morphologic changes. These 32 women had lesions at 72 distinct anatomic sites (Table I) and were chosen for this study because paraffin-embedded fixed tissues were available from all of the lesions. Histologic criteria. Anogenital squamous cell lesions were classified in three groups for this study: (1) in-
Volume 165 Number 5, Part I
vasive or microinvasive squamous cell carcinoma, (2) cervical, vulvar, vaginal, or anal grade 3 intraepithelial neoplasia, and (3) other intraepithelial neoplasia lesions or morphologic changes indicative of HPV infection (condyloma or koilocytotic atypia at the cervix). Histopathologic diagnoses were determined from pathology reports. Detection of HPV nucleic acids. The polymerase chain reaction was used to detect HPV DNA in 66 of the 72 lesions. Many of the lesions had already been analyzed by in situ hybridization before the development of the polymerase chain reaction method, and all of the available paraffin-embedded material had been used for in situ hybridization tests. Consequently, for this study adequate amounts of remaining tissue for polymerase chain reaction tests were not available for six of the 72 lesions, and in situ hybridization with asymmetric RNA probes was the only means for detection of HPV nucleic acids in these six lesions. Polymerase chain reaction. Paraffin-embedded tissues were prepared for polymerase chain reaction analyses as described by Wright and Manos,II and the polymerase chain reaction was performed on cellular DNA extracted from single 6 IJ.m sections. Initially, tissue DNAs were amplified with oligonucleotide primers specific for the E6/E7 region of the HPV type 6 or 16 genomes. I2 All positive results obtained with primers from the E6/E7 region were confirmed by amplification of a second region of the HPV type 6 or 16 genome; these were the El open reading frame (ORF) for HPV type 6 and the Ll ORF for HPV type 16. For the HPV type 6 El ORF, oligomer primers were GTTGCTGTGGATGTGACAGCAACGT (nucleotides 699 to 723) and CCTGACCATCTCCCCCCATTTTCCGG (nucleotides 1321 to 1297). Both to confirm HPV 16 positive results and also to reanalyze tissue DNAs that initially tested negative by polymerase chain reaction with HPV type 6 or 16 E6/E7 primers, a set of degenerate consensus primers were used for the amplification reactions." These primers amplify a region of sequences common to at least 25 distinct genital HPV types, 13 and, after RsaI digestion of amplification products, HPV types can be distinguished by the pattern of restriction enzyme fragments. 13.14 To assure maximum sensitivity and specificity of the polymerase chain reaction, all reaction products were electrophoresed through 4.0% agarose gels, transferred to nitrocellulose by Southern's method,I5 and hybridized with type-specific or consensus probes derived from viral DNA sequences internal to the primer pairs. This approach for analysis of HPV genomes has been previously described by our group.I2. 14 For the HPV type 6 El ORF the probe sequence was ATTATGCGACTGTGCAGGACCTAAAACGAAAGTATTTAGG (nucleotides 1055 through 1093); the probe sequences for the HPV types 6 and 16 E6/E7 ORF and
Human papillomavirus in multicentric neoplasia
1433
the 1.1 consensus sequence have been described elsewhere. I2 . 14 The amplification reactions allow consistent and reproducible detection of 0.1 pg of plasmid viral DNA after hybridization, representing about 10 3 copies of the HPV genome. In situ hybridization. In situ hybridization used asymmetric RNA probes from the E6/E7 region of the HPV types 6 or 16 genomes essentially as described by Stoler and Broker. I6 Briefly, sections were dewaxed, digested with proteinase K (15 IJ.g/ml), and acetylated (0.25% acetic anhydride in 0.1 mol/I. triethanolamine, pH 8.0). Sections were hybridized overnight at 49° C, in stringent conditions, with 1.0 to 5.0 X 107 disintegrations/min of tritium-labeled HPV type 6 or 16 RNA. Antisense RNA probes (complementary to the messenger RNA made in vivo) were generated from HPV type 6 or 16 E6/E7 ORFs cloned into the transcription vector pSP6-T7-19 (Bethesda Research Laboratories, Inc., Gaithersburg, Md.). Hybridized sections were washed, dehydrated, coated with autoradiographic emulsion (Kodak NTB-2), and exposed for 4 weeks. Tissues were stained with hematoxylin for microscopic evaluation. Results
Anatomic distribution of lesions. All 32 women in this study had two or three squamous cell lesions at various sites throughout the anogenital region; the site and histopathologic type of these neoplasias are detailed in Table 1. Lesions at the vulva and vagina (n = 10; age range 26 to 66; median, 57) were the most common in this study. Four of these women had grade 3 vulvar and vaginal intraepithelial neoplasia, five had grade 3 vulvar intraepithelial neoplasia and less severe vaginal lesions, and one had grade 3 vaginal intraepithelial neoplasia and a vulvar condyloma. Six women had lesions at the cervix and vulva (age range, 28 to 52; median, 38). Four of these women had invasive cancer or grade 3 intraepithelial neoplasia at both sites, and two had grade 3 vulvar intraepithelial neoplasia and less severe cervical lesions. Five women had lesions at the cervix and vagina (age range, 32 to 66; median, 60); all had either invasive cancer or grade 3 intraepithelial neoplasia at both sites. Two women had grade 3 intraepithelial neoplasia lesions at the vulva and anus (ages 36 and 40), and one women had grade 3 vulvar intraepithelial neoplasia and a distinct grade 3 intraepithelial neoplasia lesion on the perineum (age 49). Eight women had lesions at three sites. Three individuals had involvement at the cervix, vulva, and anus (ages 23, 26, 32); three had lesions at the vulva, vagina, and anus (ages 28, 41, 42); and two had lesions at the cervix, vagina, and vulva (ages 36, 69). Although the numbers of women in each group outlined above are small, it appears that women with vag-
1434 Beckmann et al.
November 1991 Am J Obstet Gynecol
Table II. Detection of HPV DNA in 72 lesions from 32 women with multicentric squamous neoplasia Anatomic site and histopathologic type of lesion
Cervix Grade 1 or 2 CIN* Grade 3 CIN Invasive cancer Vulva Condyloma Grade 3 VIN Invasive cancer Vagina Condyloma Grade 1 or 2 VAINt Grade 3 VAIN Invasive cancer Anus Condyloma Grade 3 AIN Perineum Grade 3 IN
HPV type in positive specimens
6
I
16
I
6 and 16
I 33
I 35
I
Unknown
HPV DNA negative
0
2
0
0
0
0
1 0
3 2
2 1
0 0
0 1
1 0
1 1 0
0 12 2
1 6 0
0 2 0
0 0 0
0 1 0
0
0
3
Total No.
3 8 5
0 1 0
I
% 100 87.5 80
2 23 2
100 96 100
7
86
1 0
5 2
2 0
0 0
1 0
1 0
0 1
10 3
100 67
1 0
0 1
4
1
1 0
0 0
0 0
0 0
3 5
100 100
0
0
0
0
0
0
100
CIN, Cervical intraepithelial neoplasia; VIN, vulvar intraepithelial neoplasia; VAIN, vaginal intraepithelial neoplasia; AIN, anal intraepithelial neoplasia. *Includes one grade 2 cervical intraepithelial neoplasia with HPV type 16. tInciudes one grade 2 vaginal intraepithelial neoplasia with HPV types 6 and 16.
inal and either cervical or vulvar lesions were older than women with lesions involving the vulva and either the cervIx or anus. Detection of HPV nucleic acids. A total of 276 tissue blocks comprising 72 anatomically distinct lesions from 32 women were analyzed by polymerase chain reaction or in situ hybridization. Sixty-seven of the 72 lesions (93%) contained HPV DNA or RNA of types 6,16,33, 35, or unidentified HPV types (Tables I and II). HPV nucleic acids were more frequently detected in anal or perineal (100%; n = 9) and vulvar (96%; 26/27) lesions than in vaginal (90%; 18/20) or cervical (88%; I41l6) lesions (Table II). When the histopathologic type of all lesions is considered, eight of nine (89%) condylomas contained detectable HPV DNA, six of six lesions (100%) graded 1 or 2 intraepithelial neoplasia were positive, 45 of 47 (96%) grade 3 intraepithelial neoplasia lesions were positive, and eight of 10 (80%) invasive cancers were positive (Table II). The negative condyloma was a flat lesion in the vagina of a woman with a grade 3 vaginal intraepithelial neoplasia lesion that contained HPV type 16 DNA. The two negative grade 3 intraepithelial neoplasia lesions were at the vulva and cervix of a woman who also had grade 3 vaginal intraepithelial neoplasia that contained HPV type 35 DNA. Of the two negative invasive cancers, one was a second primary cancer of the vagina in a woman with invasive cancer of the cervix that contained HPV type 35 DNA, and the second was an invasive cervical cancer in a woman with invasive vulvar cancer that contained HPV type 16 DNA. All of the HPV-negative tissues had been
tested with the polymerase chain reaction; the less sensitive in situ hybridization method was not responsible for these negative results. Also, these tissues were tested with primer pairs for the human l3-g10bin gene. IS. 17 All the HPV-negative lesions gave positive results with the l3-g10bin primers (data not shown), demonstrating the adequacy of the DNA preparations for polymerase chain reaction analyses. Distribution of HPV types by anatomic location of lesions. In the 67 lesions that contained detectable HPV nucleic acids, HPV type 6 or 16 was found in tissues from all anatomic sites, whereas HPV type 33 was seen only at the vulva, vagina, and anus, and HPV type 35 only at the cervix and vagina (Table II). HPV type 16 infection was more common at the cervix, vulva, and vagina than at the anus, whereas mixed infection with HPV types 6 and 16 was most common at the anus. At all sites, as the degree of neoplasia increased, so did the proportion of HPV type 16-positive lesions (Table III). Double infections with HPV types 6 and 16 were approximately twice as common in noninvasive lesions; only one invasive cancer, at the cervix, contained both types 6 and 16 (Table II). HPV type 16 was the most common virus type found in invasive cancers or grade 3 intraepithelial neoplasia lesions of the cervix (51l1), vulva (14/24), or vagina (7 1l2). However, women with anal intraepithelial neoplasia were more likely to have a dual infection with HPV types 6 and 16 (4/5) than an infection with HPV type 16 alone (115) (Table II). Virus patterns in multicentric lesions. HPV nucleic acids were detected in all of the anatomically distinct
Human papillomavirus in multicentric neoplasia
Volume 165 Number 5. Part 1
1435
Table III. HPV type-specific disease association in virus-positive multicentric lesions HPV type in positive specimens Histopatlwlogic type of lesion
Condyloma. grade I or 2IN Grade 3 IN Invasive cancer
16
6 No.
4
I%
No.
I%
No.
29
3
21
5
7
21 6
47
15 I
3
o
o
75
33
6 and 16
I
%
No.
I%
No.
o
36
2
14
33 12.5
o
2
4
o
Unknown
35
I
%
No.
o
o
2
12.5
3
o
I
%
Total (No.)
o
14
7
45
o
8
IN, Intraepithelial neoplasia.
lesions in 87.5% (28/32) of the women in this study. Four women (subjects 2, 11, 15, and 23) had detectable HPV at only one site; three of these women had lesions at two sites and one had lesions at three sites (Table I). The proportion of women with HPV DNA or RNA at all sites was not dependent on whether lesions were present at two or three sites (87.5% positive in both groups; 21/24 and 7/8, respectively). In women with lesions at two sites, the proportion of individuals with HPV at both sites did not depend on the degree of neoplasia. For women with grade 3 intraepithelial neoplasia at two sites, 87.5% (14/16) were positive at both sites. For women with grade 3 intraepithelial neoplasia at one site and less severe neoplasia at the other involved site, 87.5% (7/8) were positive at all three sites. In women with lesions at three sites, HPV was found in five of six individuals (84%) with grade 3 intraepithelial neoplasia at all sites and in both women with one grade 3 intraepithelial neoplasia lesion and less severe neoplasia at the other two sites. In the 28 women with detectable HPV at all sites, 61 % (17/28) had the same virus type(s) at all sites and 25% (7/28) had HPV type 6 or 16 at one site and both viruses at the other site(s) (Table I). Four women (14%) had different virus patterns in the separate lesions (subjects 6, 10, 12, and 18; Table I). Occurrence of double HPV infections. Thirteen of the 32 women (41 %) with multicentric neoplasia and 22 of the 72 lesions (31 %) showed evidence of infection with HPV types 6 and 16. Six of the 32 women (19%) were infected with HPV type 6 and 16 at more than one site, and seven of 32 (22%) were infected with both types at only one site. The total proportion of women with both HPV type 6 and 16 infections was identical in those individuals with grade 3 intraepithelial neoplasia at multiple sites (41 %,9/22) and those with grade 3 intraepithelial neoplasia at one site and lesions less than grade 3 intraepithelial neoplasia at other sites (40%, 4/10). In the first group of women, four of 22 (18%) individuals had double infections at more than one site, and five of 22 (23%) had double infections at only one site. In the
second group, two of 10 women (20%) had double infections at more than one site and the same proportion had double infections at only one site.
Comment In this study of women with multicentric squamous cell neoplasia of the lower genital tract, all of the 32 women had molecular evidence of HPV infection at at least one site. Five lesions from four individuals lacked detectable HPV nucleic acids; in each instance either an invasive cancer or grade 3 intraepithelial neoplasia at another site was HPV positive. Failure to detect HPV DNA in genital squamous cell carcinomas by conventional hybridization techniques (i.e., Southern, dot blot, in situ) may reflect a sampling error (missing the lesion), the presence of small amounts of virus that would not be detected by the methods used, the presence of a virus type unrelated to the probes, or the absence of HPV in the tumor. In this study the lack of HPV in these lesions is unlikely to represent a sampling error as sections from the entire tumor were analyzed in all cases. The sensitivity of the polymerase chain reaction is such that viral DNA should be detectable if present; however, the late consensus primers may not detect all HPV types that can infect the genital tract. A more convincing explanation for these negative findings is that random nicks introduced into the viral DNA after storage of paraffin tissues removed the oligonucleotide primer binding sites. Other investigators have demonstrated random loss of amplification ability of the human l3-globin gene, which increases with the age of the paraffin-embedded tissue. 17 Alternately, the HPVnegative lesions may not contain any viral nucleic acids. Lower rates of HPV positivity than what was found in this study have been reported in previous studies of women with multicentric squamous cell genital lesions. Pilotti et al. 18 used in situ hybridization to study tissues from women with multicentric disease and found that 73% (19/26) were positive for HPV DNA. These women were selected by clinical observation of vulvar wart-like lesions and simultaneous lesions at other sites, and approximately one half of the women had grade 3 vulvar intraepithelial neoplasia and the other
1436 Beckmann et al.
half had condylomatous vulvar lesions. Bergeron et al. 19 found HPV DNA by Southern hybridization in 87.5% (21/24) of women with multiple condylomatous lesions of the anogenital region. Kulski et al.20 found evidence of HPV infection in 85% (109/128) of women with cervical abnormalities who also had vulvar tissues collected for analysis by filter in situ hybridization. The 100% patient positivity rate in this study is, most likely, a result of the use of polymerase chain reaction, a test that is many times more sensitive than other detection methods. However, when considering each lesion separately, we found lower rates of HPV detection in the cervices of affected individuals than were seen at other sites in the genital tract. These results are similar to those of Reid et al.,21 who found detection rates, in cervical and vulvar biopsy specimens that showed koilocytotic atypia or neoplasia, of 86% and 93%, respectively. Conceivably, a wider variety of HPV types infect the cervix than the vulva, and these types infecting the cervix are not detectable under our experimental conditions. The relative frequencies of vulvar, cervical, vaginal, or anal HPV infections occurring simultaneously, in any combination, are not known. Some studies have shown that in women with vulvar condyloma pr grade 3 vulvar intraepithelial neoplasia up to 60% also have cervical HPV infections or preneoplastic lesions. 22 -24 Conversely, in women diagnosed with HPV-related cervical lesions, more than one half also have vulvar lesions!O In 1984, Okagaki 25 proposed the term genital neoplasm papilloma syndrome to describe this clinical observation of synchronously or metachronously occurring multiple HPV infections or HPV-associated intraepithelial neoplasia lesions in the female genital tract. Also, epidemiologic studies have shown a significant association between the simultaneous detection or occurrence of cervical and vulvar cancers and squamous cell cancers at other anogenital sites!6.27 implicating a common etiologic agent. The patterns of viral distribution in the multicentric lesions examined in this study suggest that the majority of these neoplasias were the result of diffuse HPV infection. Four women had different HPV types found in different areas of the genital tract; of these, two had two lesions of the same histologic grade (grade 3 vulvar intraepithelial neoplasia and grade 3 vaginal intraepithelial neoplasia or grade 3 cervical intraepithelial neoplasia), and the other two had grade 3 vaginal intraepithelial neoplasia and a vulvar condyloma, or grade 3 cervical intraepithelial neoplasia, vulvar invasive cancer, and an anal condyloma. Our results are consistent with the idea that squamous cell malignancies of the lower female genital tract share a common etiology, of which one factor is HPV infection. Women with multicentric HPV infection, whether in-
November 1991 Am J Obstet Gynecol
volving benign condylomas, precancers, or invasive lesions, appear to have much higher rates of multiple infections than do groups of women with only singlesite involvement. In this study 13 of the 32 (41 %) women had mixed infections throughout the lower genital tract, and 22 of the 72 (31 %) individual lesions contained both HPV types 6 and 16 DNA. These results agree with other studies in which from 38% to 100% of women with multicentric lesions were infected with two or more HPV types.18-20.28 In contrast, studies of women with only cervical or vulvar lesions have shown low rates «5%) of multiple HPV infections. 2This study did not analyze the exact cellular location of each virus type present in a single lesion. Reid et al. 21 have shown that in individuals with multiple HPV infections the different virus types were localized in demarcated areas of differing histopathologic appearances. However, in our study most all of these lesions containing HPV types 6 and 16 DNA were of similar or identical histopathologic grades. The observation of multiple infections in women with synchronously occurring multicentric lesions suggests that this study group represents a selection of individuals in whom the effectiveness of the host response to HPV infection is inadequate. In this study different viruses infecting differing areas of the female genital tract were ruled out as a potential explanation for the observed variation in the occurrence of squamous cell cancers at the cervix, vulvaa, vagina, and anus. HPV type 16 infection was the common thread among intraepithelial neoplasia III or invasive lesions at all sites. Interestingly, simultaneous infections with HPV types 6 and 16 were more common at the vulva and perineal and perianal skin than at the cervix, perhaps indicating a site restriction of HPV type 6 infection to cutaneous epithelium. In any event, factors other than the infecting virus type appear to be responsible for the 15-fo1d increase of cervical cancer over cancers at the vulva or vagina or anus. The unique properties of the cervical transformation zone"9 may make it more susceptible to HPV-induced oncogenesis than the native squamous epithelium is elsewhere in the lower genital tract. We are currently investigating the hypothesis that the cervix is more likely to undergo malignant changes after HPV infection than is the vulva, vagina, or anus because of site-specific variations in the expression of the HPV transforming genes £6 and £7. We thank Drs. James McDougall and Denise Galloway for helpful and critical consultation. REFERENCES 1. Lorincz AT, Reid R. Association of human papillomavirus with gynecologic cancer. Curr Opinion Oncol 1989; 1: 12332. 2. PaavonenJ. Koutsky LA. Kiviat N. Cervical neoplasia and
Volume 165 Number 5, Part 1
3. 4. 5. 6. 7. 8. 9.
10. 11.
12. 13.
14.
15. 16.
other STD-related genital and anal neoplasia. In: Holmes KK, Mardh P-A, Sparling PF, Wiesner PJ, eds. Sexually transmitted diseases. ew York: McGraw-Hill, 1990:56192. zur Hausen H. Papillomaviruses in anogenital cancer as a model to understand the role of viruses in human cancers. Cancer Res 1989;49:4677-81. de-Villiers EM. Heterogeneity of the human papillomavirus group. J Virol 1989;63:4898-903. Brinton LA, FraumeriJF Jr. Epidemiology of uterine cervical cancer. J Chronic Dis 1986;39:1051-65. Brinton LA, Nasca PC, Mallin K, Baptiste MS, Wilbanks GD, Richart RM. Case-control study of cancer of the vulva. Obstet Gynecol 1990;75:859-66. Brinton LA, Nasca PC, Mallin K. et aI. Case-control study of in situ and invasive carcinoma of the vagina. Gynecol Oncol 1990;38:49-54. Daling JR. Weiss NS. Hislop TG, et aI. Sexual practices, sexually transmitted diseases, and the incidence of anal cancer. N EnglJ Med 1987;317:973-7. Holly EA, Whit!emore AS, Aston DA, Ahn DK, Nickoloff BJ, Kristiansen JJ. Anal cancer incidence: genital warts, anal fissure or fistula, hemorrhoids, and smoking. J Nat! Cancer Inst 1989;81:1726-31. National Cancer Institute. 1987 Annual cancer statistics review. Hyattsville, Maryland: US Department of Health and Human Services, 1988: NIH publication no. 88-2789. Wright DK, Manos MM. Sample preparation from paraffin-embedded tissues. In: Innis MA, Gelfand DH, Sninsky JJ, White TJ, eds. PCR protocols: a guide to methods and applications. San Diego: Academic Press, 1990: 153-8. Jenison SA, Yu XP, Valentine JM, et al. Evidence of prevalent genital-type human papillomavirus infections in adults and children. J Infect Dis 1990;162:60-9. Ting Y, Manos MM. Detection and typing of genital human papillomaviruses. In: Innis MA, Gelfand DH, Sninsky JJ, White TJ. eds. PCR protocols: a guide to methods and applications. San Diego: Academic Press, 1990:35667. Beckmann AM, Sherman KJ, Myerson DM. Daling JR, McDougallJK, Galloway DA. Comparative virologic studies of condylomata acuminata reveal a lack of dual infections with human papillomaviruses. J Infect Dis 1991; 163:393-6. Southern EM. Detection of specific sequences among DNA fragments separated by gel electrophoresis. ] Mol Bioi 1975;98:503-17. Stoler MH, Broker TR. In situ hybridization detection of human papillomavirus DNAs and messenger RNAs in genital condylomas and a cervical carcinoma. Hum Pathol 1986; 17: 1250-8.
Human papillomavirus in multicentric neoplasia
1437
17. Greer CE, Peterson SL, Kiviat N, et al. PCR amplification from paraffin-embedded tissues. Effects of fixative and fixation time. Am] Clin Pathol 1991;95:117-24. 18. Pilotti S, GuptaJ, Stefanon B, De Palo G, Shah DV, Rilke F. Study of multiple human papillomavirus-related lesions of the lower female genital tract by in situ hybridization. Hum Pathol 1989;20: 118-23. 19. Bergeron C, Ferenczy A, Shah KV, Naghashfar Z. Multicentric human papiUomavirus infections of the female genital tract: correlation of viral types with abnormal mitotic figures, colposcopic presentation, and location. Obstet Gynecol 1987;69:736-42. 20. KulskiJK, Demeter T, Rakoczy P. Sterrett GF, Pixley EC. Human papillomavirus coinfections of the vulva and uterine cervix. J Med Virol 1989;27:244-51. 21. Reid R, Greenberg M, Jenson AB. et aI. Sexually transmitted papillomaviral infections. 1. The anatomic distribution and pathologic grade of neoplastic lesions associated with different viral types. AM J OBSTET GYNECOL 1987:156:212-22. 22. Walker PG, Colley NY, Grubb C, Tejerina A, Oriel JD. Abnormalities of the uterine cervix in women with vulval warts. Br J Vener Dis 1983;59: 120-3. 23. Walker PG, Singer A, DysonJL, OrielJD. Natural history of cervical epithelial abnormalities with vulval warts. Br J Vener Dis 1983;59:327-9. 24. Pilotti S, Rilke F. Shah KV, Delle Torre G, De Palo G. Immunohistochemical and ultrastructural evidence of papillomavirus infection associated with in situ and microinvasive squamous cell carcinoma of the vulva. Am J Surg Pathoi 1984;10:751-61. 25. Okagaki T. Female genital tumors associated with human papillomavirus infection and the concept of genital neoplasm-papilloma syndrome (GENPS). Pathol Annu 1984;19:31-62. 26. Sherman KJ, Daling JR, Chu J, McKnight B, Weiss NS. Multiple primary tumours in women with vulvar neoplasms: a case-control study. Br J Cancer 1988;57:423-7. 27. Henson D, Tarone R. An epidemiologic study of the cervix. vagina and vulva based on the third National Cancer Survey in the United States. Obstet Gynecol 1977; 120:525-32. 28. McCance DJ, Clarkson PK, Dyson JL, Walker PC, Singer A. Human papillomavirus types 6 and 16 in multifocal intraepithelial neoplasias of the female lower genital tract. Br J Obstet Gynecol 1985;92:1093-100. 29. Burghardt E, Oster AG. Site and origin of squamous cervical carcinoma. A histomorphological study. Obstet GynecoI1983;62:117-27.
