| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Short Communication |
From the Department of Pathology,*
Federal University of
Paraná, Curitiba, Brazil; the Ludwig Institute of Cancer
Research,
São Paulo, Brazil; National
Taiwan University Hospital,
Taipei, Taiwan,
and the Division of Women's and Perinatal
Pathology,§
Department of Pathology, Brigham
and Women's Hospital and Harvard Medical School,
Boston, Massachusetts
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
A variety of karyotypic abnormalities have been associated with VSCC in general, including those associated with HPVs. Reported karyotypic changes include both losses (3p, 5q, 8p, 10q, 15q, 18q, 19p, 22q, Xp, and others) and gains (3q).2 Following a report by one group linking HPV16-associated cervical cancers with gains in 3q, others have linked the 3q gains with HPV-positive VSCCs.3,4 In contrast, HPV-negative tumors have been distinguished by mutations in p53.5
This study was a survey of a panel of HPV-positive and -negative VSCCs for allelic losses at multiple chromosomal loci. The purpose was threefold: 1) to identify common sites of allelic loss that would be potentially productive targets for analyzing HPV-positive and -negative preinvasive disease, 2) to determine whether frequencies of loss of heterozygosity (LOH) at loci were a function of HPV status, and 3) to determine whether distribution patterns of LOH between VSCC and a prototypical HPV-dependent neoplasm, cervical squamous cell carcinoma (CSCC) were similar.2,6
| Materials and Methods |
|---|
|
|
|---|
Cases of VSCC from the files of the Women's and Perinatal Pathology Division in the Department of Pathology at Brigham and Women's Hospital were selected. Attention was paid to obtaining a wide spectrum of epithelial morphology, including intraepithelial-like growth patterns (warty or basaloid tumors commonly associated with HPV nucleic acids and VIN), and keratinizing VSCCs (HPV-negative tumors typically associated with lichen sclerosis, hyperplasia, and atypical hyperplasias).
Tissue DNA Isolation
DNA was extracted from archival tissue as previously described.7,8 Briefly, 6-µm unstained serial sections were oriented and areas of tumor and control (stroma) tissue identified by referring to paired stained sections. Tumor and stroma were separately removed from two to four sections by scraping the lesional or stromal cells with a clean scalpel. Tissues were transferred to an Eppendorf microcentrifuge tube containing digestion buffer and incubated overnight at 62°C. After addition of Chelex (Biorad, Burlingame, CA) the samples were incubated in a waterbath at 100°C for 10 minutes and centrifuged, and the supernatant was transferred to a sterile tube.7,8
HPV DNA Detection
Cases were analyzed for HPV DNA as previously described using
polymerase chain reaction (PCR) and restriction fragment polymorphism
analysis (RFLP).7,8
Briefly, 2 to 5 µl of sample DNA was
amplified in PCR buffer containing primers designed to amplify a wide
range of HPV types (MY09/MY11) based on shared homology in the L1
region.9
Reactions were conducted in the presence of
[
32P]dCTP, and the radioactive PCR products were
digested with restriction endonucleases Pst1,
RsaI, and HaeIII and resolved on polyacrylamide
gels.10
After autoradiography, HPV type was assigned based
on fragment sizes expected for known sequences.
To account for potential loss of the L1 region via viral disruption during genomic integration, samples scoring negative with the L1 consensus primers were analyzed using primers designed to amplify HPV 16/31 and 18 E7 sequences.8 Using the same strategy as described above, type was assigned after Alu1 digestion.
Analysis of LOH or Microsatellite Instability (MI)
PCR primers flanking tetranucleotide repeat sequences were
selected from the Cooperative Human Linkage Center Human Screening Set
Version 1/Weber Set 8 and purchased from Research Genetics (Huntsville,
AL). To increase the likelihood of identifying microsatellite loci
distinguishing HPV-positive from -negative phenotypes, we selected
examples for detailed LOH analysis at 66 loci on 22 chromosomes (Table 1)
. Primers scoring positive for LOH in
either of two VSCC prototypes were selected for more extensive testing
of remaining HPV-positive and -negative VSCCs.
|
|
Frequencies of allelic loss at individual loci were expressed as the number of cases scoring positive for LOH over the total number of cases informative (heterozygous) at that locus. Differences in frequency of allelic loss were compared between HPV-positive and HPV-negative tumors using the Fisher's exact test. Differences in frequency were assessed as significant at P values equal to or less than 0.05.
For comparison, frequencies of LOH on chromosome arms using the primers were contrasted with one previously published cytogenetic study of VSCC2 and one comprehensive study of allelic loss in cervical carcinomas.6
| Results |
|---|
|
|
|---|
The two prototype tumors were tested with a total of 65 markers
targeting loci on all chromosomes (Table 1)
. Of these, chromosomal loci
targeted by 29 primer pairs demonstrated LOH and were used to screen
the remaining 14 tumors (Table 2)
. For
all 16 cases, a total of 464 loci were tested; 358 (77.2%) were
informative and 106 were either homozygous (89 loci; 19.2%),
demonstrated MI (8 loci; 1.7%), or failed to amplify sufficiently to
be analyzed (9 loci; 1.9%).
|
Table 3
compares the frequencies of
karyotypic losses observed by Worsham et al2
with the
findings in this study. Close parallels between PCR-ascertained LOH and
cytogenetic losses at 3p, 5q, 8p, 10q, and 15q and 22q were observed.
Other chromosome arms were identified in this study that were not
previously emphasized by Worsham et al,2
including 1q, 2q,
8q, 10p, 11p, 11q, 17p, and 21q.
|
Relationship between HPV Status and Allelic Loss
Fifty percent (8/16) of tumors scored positive for HPV nucleic
acids. Of the HPV-positive tumors, six were type 16, one was type 33,
and one was a novel HPV type. This predilection for HPV type 16 is
consistent with previous reports. Table 2
depicts the distribution
pattern of LOH as a function of HPV status. Altogether, 181 and 177
informative loci were analyzed from HPV-positive and -negative tumors,
respectively, with 54 (29.8%) and 63 (35.6%) scoring positive. MI was
noted in similar frequencies in HPV-positive and -negative tumors.
Frequencies of allelic loss at individual loci were compared between
HPV-positive and -negative tumors and differences analyzed by the
Fisher's exact test. Significant differences in frequency were
identified at two loci, D5S816 (5q) and D10S1435 (10p), where a greater
proportion of HPV-negative tumors scored positive for LOH at
P = 0.016 and 0.047, respectively (Table 2)
. All other
loci, including those scoring for LOH in 50% or more tumors, did not
exhibit a statistically significant relationship between allelic loss
and HPV status.
Relationship between Patterns of Allelic Loss in Vulvar versus Cervical Squamous Tumors
Overall, the frequency of allelic losses in this study of VSSCs
was markedly higher than that observed in cervical carcinomas by Rader
et al,6
who observed a mean allelic loss on all chromosomal
arms to be 12% in contrast to the 32% seen in this study (Table 3)
. Similar frequencies in allelic loss
between the two sites were observed in 2q, 3p, 11q, and 18p. In
contrast, higher frequencies of allelic loss were identified in VSSCs
at 5q, 8p, 8q, 10p, 10q, 11p, 11q, 17p, 21q, and 22q. These differences
in frequency relative to cervical cancers generally held for both
HPV-positive and -negative VSSCs.
| Discussion |
|---|
|
|
|---|
The frequency of LOH in VSCCs (32%) was considerably higher than that reported in cervical tumors (12%) by Rader et al.6 However, it should be emphasized that the strategy for detecting LOH between the two studies varied. Rader et al6 selected certain primers sets to enrich for loci thought to be important in the pathogenesis of tumors. The current study targeted two prototype tumors with a large number of primers, and subsequent screening was pursued only in loci initially scoring positive (29 loci). Thus, it is not possible to answer with certainty whether VSCCs have a fundamentally different pathogenesis than CSCC by comparison of LOH alone. However, because of the low frequencies of LOH reported in CSCC by Rader et al6 and others,12 this hypothesis merits further study. Generally, a wide range of allelic loss has also been reported in tumors in extragenital squamous cell carcinomas, including of the skin, lung, head and neck, and esophagus.13-20
Similar to studies of the cervix, we found high rates of LOH at 3p and 11q.6 The locus at 6p, which is also a frequent site of LOH in cervical tumors, was not examined further in our study after scoring negative in the two index cases. Allelic losses on 3p were detected in 50% of our cases. The high frequency of LOH in this chromosomal region was in accordance with previous investigations, suggesting an important role for 3p in carcinogenesis and progression of squamous-epithelium-derived tumors of the head and neck, esophagus, lung, skin, anus, vulva, and cervix.6,21-28 Losses on 3p were also detected in high frequency (five of six) in the cytogenetic analysis by Worsham et al.2 Potential sites for novel 3p tumor suppresser genes were mapped in carcinomas of the head and neck and lung as well as endometrium, cervix, and ovary.16,23,29 However, this chromosomal region has previously been identified as a fragile site (hot spot) for sequence loss or rearrangement.30 This has prompted speculation that alterations in genes, such as the FHIT (fragile histidine triad) gene, within this region are incidental to rather than a cause of the initiation of squamous neoplasia.31 In any event, there is no evidence from this study that allelic loss in this region will distinguish HPV-positive from -negative tumors. Similarly, allelic loss at 11q, although present in one-third of informative VSCCs, was not associated with HPV status.
The highest prevalence of LOH in our series of VSSC involved chromosome 17p (58.3%), which is of interest in light of the very low incidence of LOH on 17p in cervical squamous cell carcinomas, and the reported low incidence of p53 mutations in HPV positive VSCCs.5,6,12,32 Consistent genetic losses on this chromosomal arm have also been described in many solid tumors and in squamous cell carcinomas of the head, neck, lung, esophagus, and skin.14,17-19 Accordingly, carcinomas of head and neck have shown consistent mutation and/or LOH of p53 (17p13.1) as well as the nearby CHRNB1 locus at 17p12-p11.1.33,34 One of these studies also indicated 17p12-p11 as a potential site for a tumor suppressor gene.33 In the present study, LOH on 17p was present in both HPV-positive (3/7) and HPV-negative (4/5) tumors and, like 3p and 11q, failed to discriminate between the two groups.
Other loci of interest included 8p, 8q, 21q, and 22q, all of which scored high for LOH in VSCC in contrast to cervix. Chromosome 21 scored high for LOH by three different markers: CHLCGATA129D11 (46.2%), CHLCGATA188F04 (42.9%), and D21S1446 (50%). LOH on 8q, the second most frequent site containing losses in our study (53.3%) has not been conspicuous in squamous cell carcinomas of any other site except for the cervix, but this may be subject to revision after its study in a wider spectrum of tumors. High prevalence of LOH in our study was found in the short (50.0%) and long (57.1%) arms of chromosome 8. Allelic losses in 8p were described also in squamous cell carcinomas of head and neck and lung and are consistent with previous cytogenetic findings in short-term culture of vulvar carcinoma cells.2,20,35,36
One objective of this study was to determine whether HPV-positive and -negative VSCCs could be distinguished according to frequencies of chromosome-specific allelic loss. However, with the exception of 5q and 10p, no loci exhibited significant differences in frequency, and only one (5q) demonstrated a P value of less than 0.01. These findings are not compelling, given the likelihood that a random comparison of 29 loci might produce differences at one or two loci that appear significant. Moreover, LOH at 5q was reported in 23% of CSCCs.12 Further analysis of 5q by three additional primer pairs targeting flanking loci revealed six additional sites of LOH in three HPV-negative and three in two HPV-positive tumors (data not shown). However, although these findings further support the high prevalence of LOH on 5q, determining their relationship to viral status will require additional study.
Another objective of this study was to identify loci with high frequencies of LOH that could serve as targets in the study of precursor lesions and help resolve the timing of these molecular events during the evolution of HPV-positive and -negative vulvar neoplasia. LOH at 3p has been reported to be an early event in cervical and head and neck carcinogenesis, occurring in premalignant lesions.25,28,37 In a previous study, we showed non-invasive monoclonal squamous lesions of the vulva could have either a VIN or hyperplasia morphology.38 In a recent study targeting a limited number of chromosomal loci, LOH was identified in some of these intraepithelial lesions, and LOH at one locus, 5q, was shared by VIN and VSCC.11 Based on the current study, additional loci merit testing in noninvasive vulvar mucosal lesions to determine their role in the pathogenesis of HPV-positive and -negative squamous cell carcinomas at this site.
It should be emphasized that assumptions regarding allelic loss should be tempered by one technical limitation, the inability to completely exclude allelic gain. Studies using comparative genomic hybridization have revealed amplification of several chromosomal loci, either in cervical carcinomas or cell cultures of HPV-immortalized cells.3,39 When pronounced, asymmetric allelic amplification could produce differences in PCR product intensity, similar to allelic loss. Thus, the nature of any "allelic imbalance" of potential etiologic importance should be verified to determine its influence (by either loss or gain) on tumor phenotype.
| Footnotes |
|---|
Supported by a grant from the Milton Fund and the Pemberton Fellowship Fund (A.P. Pinto). Dr. Pinto is also a recipient of a Ph.D. fellowship grant from Conselho Nacional de Desenvolvimento em Pesquisa, Brazil.
Accepted for publication December 30, 1998.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
R. P. Insinga, K.-L. Liaw, L. G. Johnson, and M. M. Madeleine A Systematic Review of the Prevalence and Attribution of Human Papillomavirus Types among Cervical, Vaginal, and Vulvar Precancers and Cancers in the United States Cancer Epidemiol. Biomarkers Prev., July 1, 2008; 17(7): 1611 - 1622. [Abstract] [Full Text] [PDF] |
||||
![]() |
W. M. Schopperle and W. C. DeWolf The TRA-1-60 and TRA-1-81 Human Pluripotent Stem Cell Markers Are Expressed on Podocalyxin in Embryonal Carcinoma Stem Cells, March 1, 2007; 25(3): 723 - 730. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. A. Brooks, J. A. Tidy, B. Gusterson, L. Hiller, J. ONions, M. Gasco, M. C. Marin, P. J. Farrell, W. G. Kaelin Jr., and T. Crook Preferential Retention of Codon 72 Arginine p53 in Squamous Cell Carcinomas of the Vulva Occurs in Cancers Positive and Negative for Human Papillomavirus Cancer Res., December 1, 2000; 60(24): 6875 - 6877. [Abstract] [Full Text] |
||||
![]() |
G. L. Mutter, M.-C. Lin, J. T. Fitzgerald, J. B. Kum, J. P. A. Baak, J. A. Lees, L.-P. Weng, and C. Eng Altered PTEN Expression as a Diagnostic Marker for the Earliest Endometrial Precancers J Natl Cancer Inst, June 7, 2000; 92(11): 924 - 930. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. C. Flowers, I. I. Wistuba, J. Scurry, C. Y. Muller, R. Ashfaq, D. S. Miller, J. D. Minna, and A. F. Gazdar Genetic Changes During the Multistage Pathogenesis of Human Papiloomavirus Positive and Negative Vulvar Carcinomas Reproductive Sciences, July 1, 1999; 6(4): 213 - 221. [Abstract] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |