- Email: [email protected]
Chromosome 4 Deletions Are Frequent in Invasive Cervical Cancer and Differ between Histologic Variants
Gynecologic Oncology 79, 90 –96 (2000) doi:10.1006/gyno.2000.5922, available online at http://www.idealibrary.com on
Chromosome 4 Deletions Are Frequent in Invasive Cervical Cancer and Differ between Histologic Variants 1 Jennifer B. Sherwood, M.D.,* ,† ,2 Narayan Shivapurkar, Ph.D.,† W. Michael Lin, M.D.,* ,† Raheela Ashfaq, M.D.,‡ David S. Miller, M.D.,* ,† Adi F. Gazdar, M.D.,† ,‡ and Carolyn Y. Muller, M.D.* ,† ,3,4 *Division of Gynecologic Oncology, †Hamon Center for Therapeutic Oncology Research, and ‡Department of Pathology, University of Texas Southwestern Medical Center, Dallas, Texas 75235 Received February 25, 2000
Key Words: allele loss; loss of heterozygosity; squamous cell carcinoma; adenocarcinoma; human papillomavirus.
Objective. Patterns of discontinuous deletion of chromosome 4 have been described in histologic variants of lung carcinomas and may represent different “hotspot” targets for gene– environment interactions. Since similar environmental risks exist for cervical cancer, we investigated patterns of discontinuous deletion in two major histologic variants. Methods. Thirteen archival cases of squamous cell cancer (SCCA) and 11 cases of adenocarcinoma (AC) were precisely microdissected. Matched normal and tumor DNA were used for polymerase chain reaction (PCR) based loss of heterozygosity (LOH) analyses using 19 polymorphic markers spanning chromosome 4. Human papillomavirus (HPV) detection was determined by PCR using general and type-specific primers (HPV 16, 18). Differences in LOH between histologic tumor types and chromosomal regions were determined using Fisher’s exact test. Results. Loss at any chromosome 4 locus occurred in 92% of all tumors studied, with the majority of deletions occurring on the long arm of the chromosome. Four discrete minimal regions of discontinuous deletion (R) were identified. For these regions, LOH frequencies were 76% (R1, 4q34 – q35), 48% (R2, 4q25– q26), 36% (R3, 4p15.1–p15.3), and 26% (R4, 4p16). Loss in SCCA predominated at 4q (4q34 – q35; 83%) and in AC at 4p (4p15.3; 50%). Overall LOH on the p arm was significant in AC (82%) compared to SCCA (31%) (P ⴝ 0.02). HPV detection was similar in SCCA (85%) and AC (73%), and HPV 16/18 subtypes were similarly represented in both histologies. Conclusions. Chromosome 4 deletions are frequent in cervical carcinomas. Different patterns of deletion between SCCA and AC may represent gene regions targeted by different gene– environment interactions in these tumor subtypes. © 2000 Academic Press
INTRODUCTION In the cervical epithelium oncogenic human papillomavirus (HPV) initiates cellular transformation, but data are limited regarding the sequence of genetic events that contribute to the histologic features and biologic behavior of invasive cancer. Malignancy develops only after long periods of latency, suggesting the necessary role for somatic events to accompany viral infection. In support of this theory, our group and others have demonstrated that allelic deletions occur at several putative tumor suppressor gene (TSG) regions in invasive cervical cancer, including 3p, 5p, and 11q [1– 4]. To date, however, no definitive critical tumor suppressor gene for cervix cancer has been identified. Studies in lung cancer provide a good model for cervical carcinogenesis, as many clinical and molecular similarities exist. Clinically, both demonstrate the field effect by which multiple dysplastic lesions are initiated by an environmental cofactor. Gene– environment interactions that result from tobacco carcinogen exposure are also common to both cancer types. On a molecular level, p53 is inactivated by mutational events (lung cancer) or degraded by viral oncoproteins (cervical cancer). Deletions on chromosome 3p, particularly at 3p14.2 within the fragile histidine triad gene locus, are also observed to be frequent events in both cancer types [1, 5, 6]. More recently, allelic losses on chromosome 4 found in lung cancers provide evidence for the existence of at least three TSGs located on this chromosome [7]. Loss of heterozygosity (LOH) analyses in breast cancer have also recently confirmed four potential TSG regions on chromosome 4 [8]. In cervix cancer, a role for chromosome 4 has been suggested by the ability of whole chromosome transfer to revert immortal HeLa cells to a senescent phenotype [9]. Comparative genomic hybridization (CGH) studies also reveal consistent deletions of chromosome 4, in up to 33–34% of invasive cervical cancers [10, 11]. A recent CGH analysis of squamous cancers found
1
Presented at the Western Association of Gynecologic Oncologists, June 2– 6, 1999, Victoria, British Columbia. 2 Supported by NIH T32 Training Grant Surgical Oncology, 1998 – 00 CA66187-04, and in part by ACOG/3M Pharmaceuticals Research Award in Lower Genital Infections, 99 – 00. 3 Supported in part by the Reproductive Scientist Development Program through NIH Grant K12HD00849 and the AAOGF. Dr. Muller is an AAOGFNICHD Fellow of the Reproductive Scientist Development Program. 4 To whom reprint requests should be addressed at Hamon Center for Therapeutic Oncology Research, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75235-9032. Fax: (214) 648-8404. E-mail: [email protected]. 0090-8258/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
90
CHROMOSOME 4 DELETIONS IN INVASIVE CERVIX CANCER
91
FIG. 1. Representative examples of microdissection of invasive squamous cell carcinoma (A) before microdissection and (B) after microdissection and adenocarcinoma (C) before microdissection and (D) after microdissection. Tumor cells are removed without contaminating stroma or infiltrating lymphocytes. (All figures, H&E ⫻100).
that losses on 4q (53%) occurred with the same frequency as those on 3p (52%) [12]. We were intrigued by the finding of unique deletion patterns on chromosome 4 in small-cell compared with non-small-cell lung cancers [7]. Certain gynecologic malignancies similarly have shown molecular characteristics unique to their histologic subtype and disease site [13, 14]. Small-cell cervical cancer is extremely rare; however, 75– 80% of cervical cancers are of squamous histology and 15–20% are adenocarcinomas. The purpose of this study, therefore, is to compare chromosome 4 allelic loss patterns between these common histologies. This pilot study was conducted initially, with an equal representation of squamous cell and adenocarcinoma, to determine the prevalence of deletions on this chromosome. MATERIALS AND METHODS Tissue specimens. Paraffin-embedded archival tissue slides from 24 cases of cervical carcinoma (13 cases of squamous cell cancer (SCCA), 11 cases of adenocarcinoma (AC); age of samples ranged from 1 to 20 years for SCCA and from 1 to 7 years for AC) were obtained from the Department of Pathology at the Parkland Memorial Hospital, Dallas, Texas (AC), and the Department of Pathology at the Illinois Masonic Medical
Center, Chicago, Illinois (SCCA). We selected all cases for which adequate tumor and nontumor (stromal) tissues were available, from operative specimens obtained as recently as possible. Adequate tumor cells for polymerase chain reaction (PCR) analysis were particularly difficult to microdissect from ACs, given the overall arcitecture of the lesion. Thus, from all possible available specimens, we were limited to 11 samples; a matched number of SCCAs was therefore selected. Tissues were obtained from hysterectomy, cold-knife conization, or large colposcopically directed biopsy. Stages of cancer were IB1–IB2, and specimens were obtained at the time of original diagnosis (not from any disease recurrence). Microdissection and DNA extraction. Areas of invasive carcinoma were identified (and reviewed by a gynecologic pathologist) on H&E-stained slides and precisely dissected under microscopic visualization (Fig. 1). SCCAs, mostly consisting of confluent islands of tumor, were microdissected using the Lasercapture technique [15]. ACs were dissected with the use of a manual technique as previously described [16] to minimize contamination of glandular-type lesions from surrounding stroma, as this is problematic with even the smallest 30-mm laser diameter. Lymphocyte-enriched stromal cells from the same cases provided a source of constitutional DNA. From multiple sections of each
92
SHERWOOD ET AL.
FIG. 2. Patterns of LOH in cervical cancers exhibiting deletions at the 19 tested loci spanning 4p and 4q. The cases have been grouped according to tumor histology, adenocarcinoma (AC), and squamous carcinomas (SCCA). Case numbers are shown at the top. Microsatellite markers and relative positions on chromosome 4 are shown on the left. Markers are placed in the predicted order from 4pter– qter. The number of cases with LOH/number of informative cases (percentage values) are shown on the right. Also shown on the far right are the four deletion hotspots designated R1 through R4, with percentages of LOH by region. (F) allele loss; (E) heterozygous; ( ) uninformative; ( ) not done.
tissue sample, 500 –1000 cells were microdissected. DNA extractions were performed by the proteinase K method, as described previously [16]. Five microliters of the digested samples, containing DNA from approximately 50 cells, was used directly for each multiplex PCR reaction. Polymorphic DNA markers, PCR, and LOH analysis. For LOH analysis, 19 primer pairs were used to identify specific dinucleotide and multinucleotide repeat polymorphisms on several loci that span chromosome 4. Primer pairs were synthesized by Life Technologies, Inc. (Gaithersburg, MD). Sequences were obtained from the Genome Database (www.gdb.org) and their relative order was ascertained from the Genethon map of chromosome 4 [17]. Selection of markers was based on regions that have demonstrated frequent LOH in other human cancers and amplification of amplicons amenable to paraffin-embedded tissues (⬃100 –200 kb). PCR conditions were performed as previously reported [7]. PCR products were separated on 6% acrylamide gels; gels were dried for 1–2 h and then subjected to autoradiographic development. LOH for the tumor was scored as the complete absence of either allele, compared to the matched normal stroma. HPV DNA detection. The presence of HPV DNA sequences was tested by PCR, using initially general primers for the L1 open reading frame of human papillomavirus, followed by subtype-specific primers for the conserved E6 regions of
high-risk HPV types 16 and 18 [18]. Primers were designed for paraffin-extracted DNA [19]. PCR was performed using two rounds of amplification to detect even low copy numbers of DNA from paraffin tissues. In the first PCR round, 5 l of DNA were used in a 50-L reaction volume; the second PCR used 5 L of first-round product in another 50-L reaction volume. DNA extracted from the human cervical cancer cell lines CaSki (HPV 16) and HeLa (HPV18) were used as positive controls. Products were separated on 2% agarose gels. Statistical analysis. Statistical comparisons between the cervical squamous cell cancers and adenocarcinomas were made using Fisher’s exact test. Significance is defined at the P ⬍ 0.05 level. RESULTS Chromosome 4 LOH—All cervical cancers. Twenty-two of 24 (92%) invasive cervical cancers showed LOH at 1 or more of the 19 chromosome 4 loci analyzed. LOH within 1 or more loci on the short arm (p arm, n ⫽ 7 markers) of chromosome 4 was seen in 13 cases (54%) and LOH on the long arm (q arm, n ⫽ 12 markers) in 21 cases (88%). A summary of the overall frequency of LOH at each locus, which was calculated as the number of cases with LOH at that locus/the total number of informative cases at that
93
CHROMOSOME 4 DELETIONS IN INVASIVE CERVIX CANCER
FIG. 3. Autoradiographs of cases demonstrating LOH at the four regions of nonoverlapping deletion (R1, R2, R3, R4). Only the deletions are depicted in this figure. Arrowheads, LOH; N, normal; T, tumor.
locus, is depicted in Fig. 2. All loci tested demonstrated at least 1 tumor with LOH within that locus. Of the 19 loci tested, D4S426 and the neighboring locus D4S171 (4q34 – q35) demonstrated LOH in the majority (70 and 67%, respectively) of cervical cancers. LOH was seen at a higher frequency within the q arm compared to the p arm. One SCCA demonstrated extensive deletions likely involving loss of the whole q arm (Fig. 2, Case No. 16). LOH results by region (R). The pattern of the partial deletions was used to identify four discrete minimal regions of nonoverlapping deletions (Fig. 2). As our group previously described, R1 and R2 were designated on the q arm and R3 and R4 on the p arm of chromosome 4 [7, 8]. A region (R) was defined when at least one case demonstrated LOH at a specific locus and at least one of its two flanking markers demonstrated retention of heterozygosity. Region (R) loss was calculated as the number of cases with LOH at any marker within a defined region/number of informative cases at any marker within that region. The noncontiguous deletions seen in these 24 cases defined four minimal areas of deletion (R1, R2, R3, R4) (Table 1, Figs. 2 and 3). On 4q, one minimal region of deletion involving markers D4S171 and D4S426 (R1, 4q34 – q35) was observed in 13 of 17 informative cases (76%) of cervical cancers overall (Figs. 2 and 3). The other minimal region of deletion was observed between markers D4S175 and D4S194 (R2, 4q25– q26), with LOH observed in 10 of 21 informative cases (48%) of cervical cancer (Figs. 2 and 3). However, the data suggest that this region may itself contain two minimal regions of deletion, one centered on D4S175 and the other centered on D4S194 (Fig. 2). Of note, marker D4S194 tends to highlight multiple bands/fragments of DNA, not discrete alleles. Although this
may appear to represent a microsatellite alteration within the tumor, a similar appearance was seen in both control genomic DNA and lung tumor DNA in our previous studies [7]. On 4p, discrete regions of nonoverlapping deletions were somewhat more difficult to assign as precisely as those on 4q, due to a higher rate of uninformative loci. One minimal region of deletion was assigned between markers D4S2366 and D4S404 (R3, 4p15.1–p15.3), with regional loss observed in 8 of 22 informative cases (36%) (Figs. 2 and 3). The other region was designated between D4S43 and D4S127 (R4, 4p16), with LOH observed in 5 of 19 informative cases (26%) (Figs. 2 and 3). However, a significant marker on the short arm was D4S2366, positioned at the junction of R3 and R4 (and potentially located within either of these two regions). Allelic losses at this specific locus occurred at a much higher rate in ACs (50%) than in SCCAs (10%) (Fig. 2). LOH results by histologic type. Overall, SCCAs demonstrated allelic loss for at least 1 chromosome 4 marker in 12 of 13 cases (92%); similarly ACs displayed loss in 10 of 11 cases (91%). Focusing at the short arm specifically, SCCAs showed LOH in 4 of 13 tumors (31%), whereas ACs showed LOH in 9 of 11 (82%) (P ⫽ 0.02, Fisher’s exact test). For the long arm of chromosome 4, losses were more equally distributed, with 11 of 13 SCCAs (85%) and 10 of 11 ACs (91%) demonstrating LOH. The distinction between histologies, however, was pronounced at the telomeric regions of chromosome 4. At the q telomere (R1 region, locus D4S171) LOH for SCCAs was seen in 5 of 6 informative cases (83%), whereas ACs showed losses in only 1 of 3 informative cases (33%). At the p telomere (R3 region, locus D4S2366) LOH was observed in ACs in 5 of 10 informative cases (50%) and in SCCAs in 1 of 10 cases (10%) (Fig. 2). TABLE 1 Cervical Cancer Cases by Histology LOH/Informative Cases (% LOH) Markers
Arm
AC (n ⫽ 11)
SCCA (n ⫽ 13)
D4S43 D4S136 D4S127 D4S2366 D4S1546 D4S404 D4S174 D4S392 D4S1089 D4S175 D4S1586 D4S194 D4S1584 D4S1554 D4S1535 D4S408 D4S171 D4S426 D4S1652
p p p p p p p q q q q q q q q q q q q
1/2 (50) 1/4 (25) 2/7 (29) 5/10 (50) 1/3 (33) 2/11 (18) 2/6 (33) 4/8 (50) 2/6 (33) 3/6 (50) 1/3 (33) 1/3 (33) 1/7 (14) 1/6 (17) 2/4 (50) 2/4 (50) 1/3 (33) 4/6 (67) 0/4 (0)
1/5 (20) 1/3 (33) 2/7 (29) 1/10 (10) 1/3 (33) 1/10 (11) 0/6 (0) 4/10 (40) 1/6 (17) 3/7 (43) 0/6 (0) 2/8 (25) 2/9 (22) 4/7 (57) 1/4 (25) 1/3 (33) 5/6 (83) 3/4 (75) 6/10 (60)
94
SHERWOOD ET AL.
TABLE 2 HPV Analysis of Adenocarcinoma and Squamous Cell Carcinoma
HPV-positive cases HPV 16 (percentage of cases positive) HPV 18 (percentage of cases positive)
AC (%)
SCCA (%)
8/11 (73) 4/8 (50) 4/8 (50)
11/13 (85) a 6/11 (55) 1/11 (9)
a
The remainder of the cases positive for HPV general primer was not further classified.
Detection of HPV sequences. Overall, 19 of 24 (79%) cervix cancers were positive for oncogenic HPV sequences using general primers (Table 2). HPV subtype distribution was as follows: SCCAs contained HPV DNA in 11 of 13 (85%) tumors. Of these cases, 6 (55%) were HPV 16 positive, while 1 (9%) was HPV 18 positive. The remainder of HPV-positive SCCAs were likely associated with other oncogenic HPV types, such as types 31, 33, and 35, but were not further classified. ACs contained HPV DNA in 8 of 11 (73%) cases, all demonstrating infection with the “classic” oncogenic subtypes, with 4 (50%) containing HPV 16 and the other 4 (50%) HPV 18. Coinfection with both HPV16 and 18 was not seen in any of the cases studied. DISCUSSION Epidemiologic risk factors that are common to both lung and cervical cancer may target similar epigenetic events leading to their development. Based on published data of genomic regions known to undergo frequent alteration in both cervical and lung cancer, we conducted the present analysis. Chromosomes 3p, 5p, 11q, 4p, and 4q are known to contain sites of frequent loss of heterozygosity and/or microsatellite instability in cervical neoplasia [1, 3, 20 –22]. Progression from preinvasive to invasive disease has also been associated with accumulation of such alterations [6, 20]. One prior study has suggested the presence of at least two independent tumor suppressor gene regions mapping to distal regions of chromosome 4, commonly altered in cervical neoplasia [22]. Recently, nonoverlapping deletions on chromosome 4 were found to be highly prevalent in small-cell and non-small-cell lung cancers, suggesting the presence of at least three distinct suppressor genes [7]. Studies in cervical cancer tend to include a predominance of SCCA or preinvasive lesions, and infrequently include an adequate number of invasive adenocarcinoma or adenocarcinoma in situ lesions. Given the known molecular and epidemiologic parallels between lung and cervical carcinoma, we performed a detailed LOH analysis on chromosome 4 using nearly equivalent numbers of SCCAs and ACs. In conjunction with novel LOH findings, our HPV findings were somewhat surprising. Although we expected the majority of SCCAs to be infected with oncogenic HPV [23], the ACs showed an equivalently high infectivity rate, with an equal
distribution between the classic oncogenic HPV 16 and HPV 18. Due to this high HPV infection rate, we found no correlation between the presence and/or oncogenic type of HPV with particular LOH events. Of interest is that 85% of SCCAs contained HPV, but only 55% were HPV 16, the subtype most commonly associated with this histologic type [23]. The ACs were positive for HPV in 73%, higher than the expected rate of 50% [24]. One-half of these were HPV 16 positive, whereas we would expect the majority to contain HPV 18 [23]. The differences seen in this cohort may reflect the geographic regions from where the cases were obtained. Further, the differences in expected percentage of HPV infection may be related to the small sample size in this pilot study. Regarding chromosomal alterations in cervix cancer, the literature to date largely describes a random sampling of chromosome 4 regions of loss. In general, many studies have utilized few markers per chromosome targeting a broader genome-wide search for putative tumor suppressor gene sites. In addition, particularly within the earlier studies, little emphasis has been placed on microdissected tumor specimens. Microdissection has permitted detection of allelic loss at higher frequencies than has been previously recognized [25]. Particularly for the ACs, this technique substantially decreases contamination with intervening stromal lymphocytes. In both SCCA and AC we observed frequent losses at four nonoverlapping sites located on the long arm (two regions) and short arm (two regions) of chromosome 4. On the q arm, the distinct minimal regions of deletion appeared to lie at 4q34 – q35 (“R1”) and 4q25– q26 (“R2”). These at least partially overlap regions of deletion in this area described in other cancer types, including breast, lung, and bladder. For example, 4q33– q34, 4q25– q26, and 4p15.1–p15.3 represented significant regions of deletion in SCLC and NSCLC [7, 8, 26]. Losses in cervical SCCA occurred predominately in the R1 region on the q arm. Overall percentage LOH on the q arm, however, was equivalent in the two subtypes (85 and 91%). In R2, the low rate of LOH observed (11%) with marker D4S1586 could simply be due to the low rate of marker informativity. Alternatively, two distinct TSGs could exist in this area, one approximating D4S194 and the other approximating D4S175. On the p arm, deletion regions were more difficult to designate as we unexpectedly observed a disproportionate number of noninformative cases for each polymorphic marker. As such, our description of regions R3 and R4 was somewhat limited. However, there was a significant difference in LOH events between AC and SCCA on the p arm (P ⫽ 0.02). The pivotal marker D4S2366, for which we observed a predominance of LOH in AC (50%) and rare losses in SCCA (10%), appears to lie between R3 and R4 and may target the region for a more detailed deletion analysis in future studies. This is a first description in cervical cancer of distinct regions of loss on a chromosome other than 3p [6]. Our results suggest the possible existence of at least two tumor suppressor genes on 4q and at least two on 4p. Of additional interest is the region represented by marker
CHROMOSOME 4 DELETIONS IN INVASIVE CERVIX CANCER
D4S392 (4q12–13), the most centromeric 4q marker used in our study. SCCA and AC demonstrated a combined rate of LOH in 8 of 18 cases (44%). The deletions identified could target the centromeric region or possibly extend to proximal p (Fig. 2, Case No. 1). In bladder carcinoma, this region has also demonstrated LOH in a moderate number of tumors, providing evidence for a putative suppressor gene [26]. The various patterns of chromosome 4 loss (with particular attention to the p arm) appear to be unique to the respective tumor histologies. To our knowledge this is the first report of a significant regional LOH difference between adenocarcinomas and squamous carcinomas of the uterine cervix. Deletion hotspots predominated at distal regions of the q arm and p arm for SCCA and AC, respectively. In lung cancers, interestingly, losses observed in the three putative suppressor gene regions on chromosome 4 were common to both small-cell (SCLC) and non-small-cell lung cancers (NSCLC) [7]. Although in this study NSCLCs were not stratified further, the frequency of LOH overall in NSCLC was much lower than that in SCLCs, possibly explaining differences in their pathogenetic mechanisms. The clinical distinction between cervical cancer variants is also likely supported by specific molecular events. Certain gene– environment interactions may be specific to certain chromosomal regions. For example, HPV infection and tobacco smoking may target distal 4q (4q34 – q35); as these environmental risks are associated with both AC and SCCA (and both histotypes share a high LOH rate in distal 4q), this finding could explain a common mechanism of carcinogenesis. That 4p LOH is unique to adenocarcinoma, however, may represent an environmental/clinical risk significant to AC. We suspect that hormonal effects may target the 4p gene region, given that adenocarcinoma of the cervix is histologically similar to estrogen-influenced endometrial adenocarcinoma. Our results have identified four discrete deletion hotspots that may harbor tumor suppressor genes implicated in cervical cancer. We suspect there may exist a target gene on 4p specific to AC and one or more within the regions of 4q involved in overall cervical cancer development. A candidate gene on chromosome 4q should now be sought to provide insight into cervical carcinogenesis.
95
sion of precancerous lesions of the uterine cervix. J Natl Cancer Inst 87:742–745, 1995 3. Rader JS, Kamarasova T, Huettner PC, Li L, Li Y, Gerhard DS: Allelotyping of all chromosomal arms in invasive cervical cancer. Oncogene 13:2737–2741, 1996 4. Lin WM, Michalopulos EA, Dhurander N, Cheng PC, Robinson WR, Ashfaq R, Coleman RL, Muller CY: Allelic loss and microsatellite alteration of chromosome 3p14.2 are more frequent in recurrent cervical dysplasias. Clin Cancer Res 6:1410 –1414, 2000. 5. Fong KM, Biesterveld EJ, Virmani A, Wistuba II, Sekido Y, Bader SA, Ahmadian M, Ong ST, Rassool FV, Zimmerman PV, Giacconne G, Gazdar AF, Minna JD: FHIT and FRA3B 3p14.2 allele loss are common in lung cancer and preneoplastic bronchial lesions and are associated with cancer-related FHIT cDNA splicing aberrations. Cancer Res 57:2256 – 2267, 1997 6. Wistuba II, Montellano FD, Milchgrub S, Virmani AK, Behrens C, Chen H, Ahmadian M, Nowak JA, Muller CY, Minna JD, Gazdar AF: Deletions of chromosome 3p are frequent and early events in the pathogenesis of uterine cervical carcinoma. Cancer Res 57:3154 –3158, 1997 7. Shivapurkar N, Virmani AK, Wistuba II, Milchgrub S, Mackay B, Minna JD, Gazdar AF: Deletions of chromosome 4 at multiple sites are frequent in malignant mesothelioma and small cell lung carcinoma. Clin Cancer Res 5:17–23, 1999 8. Shivapurkar N, Sood S, Wistuba II, Virmani AK, Maitra A, Milchgrub S, Minna JD, Gazdar AF: Multiple regions of chromosome 4 demonstrating allelic loss in breast carcinomas. Cancer Res 59:3576 –3580, 1999 9. Ning Y, Weber JL, Killary AM, Ledbetter DH, Smith JR, Pereira-Smith OM: Genetic analysis of indefinite division in human cells: Evidence for a cell-senescence related gene(s) on human chromosome 4. Proc Natl Acad Sci USA 88:5635–5659, 1991 10. Heselmeyer K, Macville M, Schrock E, Blegen H, Hellstrom AC, Shah K, Auer G, Ried T: Advanced-stage cervical carcinomas are defined by a recurrent pattern of chromosomal aberrations revealing high genetic instability and a consistent gain of chromosome arm 3q. Genes Chromosomes Cancer 19:233–240, 1997 11. Kirchhoff M, Rose H, Peterson BL, Maahr J, Gerdes T, Lundsteen C, Bryndorf T, Kryger-Baggesen N, Christensen L, Engelholm SA, Philip J: Comparative genomic hybridization reveals a recurrent pattern of chromosomal aberrations in severe dysplasia/carcinoma in situ of the cervix and in advanced-stage cervical carcinoma. Genes Chromosomes Cancer 24:144 –150, 1999 12. Dellas A, Torhorst J, Jiang F, Proffitt J, Schultheiss E, Holzgreve W, Sauter G, Mihatsch MJ, Moch H: Prognostic value of genomic alterations in invasive cervical squamous cell carcinoma of clinical stage IB detected by comparative genomic hybridization. [In Process Citation] Cancer Res 59:3475–3479, 1999
ACKNOWLEDGMENT
13. Berchuck A: Molecular basis of endometrial cancer. Cancer 2034 –2040, 1995
We thank Dr. Jan A. Nowak (Department of Pathology, Illinois Masonic Medical Center, Chicago, IL) for providing the paraffin-embedded archival squamous carcinomas used in this study.
14. Berchuck A, Carney M: Human ovarian cancer of the surface epithelium. Biochem Pharmacol 54:541–544, 1997
REFERENCES 1. Muller CY, O’Boyle JD, Fong KM, Wistuba II, Biesterveld E, Ahmadian M, Miller DS, Gazdar AF, Minna JD: Abnormalities of fragile histidine triad genomic and complementary DNAs in cervical cancer: association with human papillomavirus type [see comments]. J Natl Cancer Inst 90:433– 439, 1998 2. Mitra AB, Murty VV, Singh V, Li RG, Pratap M, Sodhani P, Luthra UK, Chaganti RS: Genetic alterations at 5p15: a potential marker for progres-
15. Emmert-Buck MR, Bonner RF, Smith PD, Chuaqui RF, Zhuang Z, Goldstein SR, Weiss RA, Liotta LA: Laser capture microdissection. [see comments] Science 274:998 –1001, 1996 16. Hung J, Kishimoto Y, Sugio K, Virmani AK, McIntire DD, Minna JD, Gazdar AF: Allele-specific chromosome 3p deletions occur at an early stage in the pathogenesis of lung carcinoma. JAMA 273:1908, 1995 17. Gyapay G, Morissette J, Vignal A, Dib C, Fizames C, Millasseau P, Marc S, Bernardi G, Lathrop M, Weissenbach J: The 1993–94 genethon human genetic linkage map. [see comments] Nat Genet 7:246 – 339, 1994 18. De Roda Husman AM, Walboomers JM, van den Brule AJ, Meijer CJ,
96
SHERWOOD ET AL. Snijders PJ: The use of general primers GP5 and GP6 elongated at their 3⬘ ends with adjacent highly conserved sequences improves human papillomavirus detection by PCR. J Gen Virol 76:1057–1062, 1995
19. Baay MF, Quint WG, Koudstaal J, Hollema H, Duk JM, Burger MP, Stolze E, Herbrink P: Comprehensive study of several general and typespecific primer pairs for detection of human papillomavirus DNA by PCR in paraffin-embedded cervical carcinomas. J Clin Microbiol 34:745–747, 1996 20. Larson AA, Liao SY, Stanbridge EJ, Cavenee WK, Hampton GM: Genetic alterations accumulate during cervical tumorigenesis and indicate a common origin for mulifocal lesions. Cancer Res 57:4171– 4176, 1997 21. Mitra AB, Murty VV, Li RG, Pratap M, Luthra UK, Chaganti RS: Allelotype analysis of cervical carcinoma. Cancer Res 54:4481– 4487, 1994 22. Hampton GM, Larson AA, Baergen RN, Sommers RL, Kern S, Cavenee
23.
24.
25.
26.
WK: Simultaneous assessment of loss of heterozygosity at multiple microsatellite loci using semi-automated fluorescence-based detection: subregional mapping of chromosome 4 in cervical carcinoma. Proc Natl Acad Sci USA 93:6704 – 6709, 1996 Iwasawa A, Nieminen P, Lehtinen M, Paavonen J: Human papillomavirus DNA in tuerine cervix squamous cell carcinoma and adenocarcinoma detected by polymerase chain reaction. Cancer 77:2275–2279, 1996 Leminen A, Paavonen J, Vesterinen E, Wahlstrom T, Rantala I, Lehtinen M: Human papillomavirus types 16 and 18 in adenocarcinoma of the uterine cervix. Am J Clin Pathol 95:647– 652, 1991 Fujii H, Zhou W, Gabrielson E: Detection of frequent allelic loss of 6q23– q25.2 in microdissected human breast cancer tissues. Genes Chromosomes Cancer 16:35–39, 1996 Polascik TJ, Cairns P, Chang WY, Schoenberg MP, Sidransky D: Distinct regions of allelic loss on chromosome 4 in human primary bladder carcinoma. Cancer Res 55:5396 –5399, 1995
Chromosome 4 Deletions Are Frequent in Invasive Cervical Cancer and Differ between Histologic Variants 1 Jennifer B. Sherwood, M.D.,* ,† ,2 Narayan Shivapurkar, Ph.D.,† W. Michael Lin, M.D.,* ,† Raheela Ashfaq, M.D.,‡ David S. Miller, M.D.,* ,† Adi F. Gazdar, M.D.,† ,‡ and Carolyn Y. Muller, M.D.* ,† ,3,4 *Division of Gynecologic Oncology, †Hamon Center for Therapeutic Oncology Research, and ‡Department of Pathology, University of Texas Southwestern Medical Center, Dallas, Texas 75235 Received February 25, 2000
Key Words: allele loss; loss of heterozygosity; squamous cell carcinoma; adenocarcinoma; human papillomavirus.
Objective. Patterns of discontinuous deletion of chromosome 4 have been described in histologic variants of lung carcinomas and may represent different “hotspot” targets for gene– environment interactions. Since similar environmental risks exist for cervical cancer, we investigated patterns of discontinuous deletion in two major histologic variants. Methods. Thirteen archival cases of squamous cell cancer (SCCA) and 11 cases of adenocarcinoma (AC) were precisely microdissected. Matched normal and tumor DNA were used for polymerase chain reaction (PCR) based loss of heterozygosity (LOH) analyses using 19 polymorphic markers spanning chromosome 4. Human papillomavirus (HPV) detection was determined by PCR using general and type-specific primers (HPV 16, 18). Differences in LOH between histologic tumor types and chromosomal regions were determined using Fisher’s exact test. Results. Loss at any chromosome 4 locus occurred in 92% of all tumors studied, with the majority of deletions occurring on the long arm of the chromosome. Four discrete minimal regions of discontinuous deletion (R) were identified. For these regions, LOH frequencies were 76% (R1, 4q34 – q35), 48% (R2, 4q25– q26), 36% (R3, 4p15.1–p15.3), and 26% (R4, 4p16). Loss in SCCA predominated at 4q (4q34 – q35; 83%) and in AC at 4p (4p15.3; 50%). Overall LOH on the p arm was significant in AC (82%) compared to SCCA (31%) (P ⴝ 0.02). HPV detection was similar in SCCA (85%) and AC (73%), and HPV 16/18 subtypes were similarly represented in both histologies. Conclusions. Chromosome 4 deletions are frequent in cervical carcinomas. Different patterns of deletion between SCCA and AC may represent gene regions targeted by different gene– environment interactions in these tumor subtypes. © 2000 Academic Press
INTRODUCTION In the cervical epithelium oncogenic human papillomavirus (HPV) initiates cellular transformation, but data are limited regarding the sequence of genetic events that contribute to the histologic features and biologic behavior of invasive cancer. Malignancy develops only after long periods of latency, suggesting the necessary role for somatic events to accompany viral infection. In support of this theory, our group and others have demonstrated that allelic deletions occur at several putative tumor suppressor gene (TSG) regions in invasive cervical cancer, including 3p, 5p, and 11q [1– 4]. To date, however, no definitive critical tumor suppressor gene for cervix cancer has been identified. Studies in lung cancer provide a good model for cervical carcinogenesis, as many clinical and molecular similarities exist. Clinically, both demonstrate the field effect by which multiple dysplastic lesions are initiated by an environmental cofactor. Gene– environment interactions that result from tobacco carcinogen exposure are also common to both cancer types. On a molecular level, p53 is inactivated by mutational events (lung cancer) or degraded by viral oncoproteins (cervical cancer). Deletions on chromosome 3p, particularly at 3p14.2 within the fragile histidine triad gene locus, are also observed to be frequent events in both cancer types [1, 5, 6]. More recently, allelic losses on chromosome 4 found in lung cancers provide evidence for the existence of at least three TSGs located on this chromosome [7]. Loss of heterozygosity (LOH) analyses in breast cancer have also recently confirmed four potential TSG regions on chromosome 4 [8]. In cervix cancer, a role for chromosome 4 has been suggested by the ability of whole chromosome transfer to revert immortal HeLa cells to a senescent phenotype [9]. Comparative genomic hybridization (CGH) studies also reveal consistent deletions of chromosome 4, in up to 33–34% of invasive cervical cancers [10, 11]. A recent CGH analysis of squamous cancers found
1
Presented at the Western Association of Gynecologic Oncologists, June 2– 6, 1999, Victoria, British Columbia. 2 Supported by NIH T32 Training Grant Surgical Oncology, 1998 – 00 CA66187-04, and in part by ACOG/3M Pharmaceuticals Research Award in Lower Genital Infections, 99 – 00. 3 Supported in part by the Reproductive Scientist Development Program through NIH Grant K12HD00849 and the AAOGF. Dr. Muller is an AAOGFNICHD Fellow of the Reproductive Scientist Development Program. 4 To whom reprint requests should be addressed at Hamon Center for Therapeutic Oncology Research, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75235-9032. Fax: (214) 648-8404. E-mail: [email protected]. 0090-8258/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
90
CHROMOSOME 4 DELETIONS IN INVASIVE CERVIX CANCER
91
FIG. 1. Representative examples of microdissection of invasive squamous cell carcinoma (A) before microdissection and (B) after microdissection and adenocarcinoma (C) before microdissection and (D) after microdissection. Tumor cells are removed without contaminating stroma or infiltrating lymphocytes. (All figures, H&E ⫻100).
that losses on 4q (53%) occurred with the same frequency as those on 3p (52%) [12]. We were intrigued by the finding of unique deletion patterns on chromosome 4 in small-cell compared with non-small-cell lung cancers [7]. Certain gynecologic malignancies similarly have shown molecular characteristics unique to their histologic subtype and disease site [13, 14]. Small-cell cervical cancer is extremely rare; however, 75– 80% of cervical cancers are of squamous histology and 15–20% are adenocarcinomas. The purpose of this study, therefore, is to compare chromosome 4 allelic loss patterns between these common histologies. This pilot study was conducted initially, with an equal representation of squamous cell and adenocarcinoma, to determine the prevalence of deletions on this chromosome. MATERIALS AND METHODS Tissue specimens. Paraffin-embedded archival tissue slides from 24 cases of cervical carcinoma (13 cases of squamous cell cancer (SCCA), 11 cases of adenocarcinoma (AC); age of samples ranged from 1 to 20 years for SCCA and from 1 to 7 years for AC) were obtained from the Department of Pathology at the Parkland Memorial Hospital, Dallas, Texas (AC), and the Department of Pathology at the Illinois Masonic Medical
Center, Chicago, Illinois (SCCA). We selected all cases for which adequate tumor and nontumor (stromal) tissues were available, from operative specimens obtained as recently as possible. Adequate tumor cells for polymerase chain reaction (PCR) analysis were particularly difficult to microdissect from ACs, given the overall arcitecture of the lesion. Thus, from all possible available specimens, we were limited to 11 samples; a matched number of SCCAs was therefore selected. Tissues were obtained from hysterectomy, cold-knife conization, or large colposcopically directed biopsy. Stages of cancer were IB1–IB2, and specimens were obtained at the time of original diagnosis (not from any disease recurrence). Microdissection and DNA extraction. Areas of invasive carcinoma were identified (and reviewed by a gynecologic pathologist) on H&E-stained slides and precisely dissected under microscopic visualization (Fig. 1). SCCAs, mostly consisting of confluent islands of tumor, were microdissected using the Lasercapture technique [15]. ACs were dissected with the use of a manual technique as previously described [16] to minimize contamination of glandular-type lesions from surrounding stroma, as this is problematic with even the smallest 30-mm laser diameter. Lymphocyte-enriched stromal cells from the same cases provided a source of constitutional DNA. From multiple sections of each
92
SHERWOOD ET AL.
FIG. 2. Patterns of LOH in cervical cancers exhibiting deletions at the 19 tested loci spanning 4p and 4q. The cases have been grouped according to tumor histology, adenocarcinoma (AC), and squamous carcinomas (SCCA). Case numbers are shown at the top. Microsatellite markers and relative positions on chromosome 4 are shown on the left. Markers are placed in the predicted order from 4pter– qter. The number of cases with LOH/number of informative cases (percentage values) are shown on the right. Also shown on the far right are the four deletion hotspots designated R1 through R4, with percentages of LOH by region. (F) allele loss; (E) heterozygous; ( ) uninformative; ( ) not done.
tissue sample, 500 –1000 cells were microdissected. DNA extractions were performed by the proteinase K method, as described previously [16]. Five microliters of the digested samples, containing DNA from approximately 50 cells, was used directly for each multiplex PCR reaction. Polymorphic DNA markers, PCR, and LOH analysis. For LOH analysis, 19 primer pairs were used to identify specific dinucleotide and multinucleotide repeat polymorphisms on several loci that span chromosome 4. Primer pairs were synthesized by Life Technologies, Inc. (Gaithersburg, MD). Sequences were obtained from the Genome Database (www.gdb.org) and their relative order was ascertained from the Genethon map of chromosome 4 [17]. Selection of markers was based on regions that have demonstrated frequent LOH in other human cancers and amplification of amplicons amenable to paraffin-embedded tissues (⬃100 –200 kb). PCR conditions were performed as previously reported [7]. PCR products were separated on 6% acrylamide gels; gels were dried for 1–2 h and then subjected to autoradiographic development. LOH for the tumor was scored as the complete absence of either allele, compared to the matched normal stroma. HPV DNA detection. The presence of HPV DNA sequences was tested by PCR, using initially general primers for the L1 open reading frame of human papillomavirus, followed by subtype-specific primers for the conserved E6 regions of
high-risk HPV types 16 and 18 [18]. Primers were designed for paraffin-extracted DNA [19]. PCR was performed using two rounds of amplification to detect even low copy numbers of DNA from paraffin tissues. In the first PCR round, 5 l of DNA were used in a 50-L reaction volume; the second PCR used 5 L of first-round product in another 50-L reaction volume. DNA extracted from the human cervical cancer cell lines CaSki (HPV 16) and HeLa (HPV18) were used as positive controls. Products were separated on 2% agarose gels. Statistical analysis. Statistical comparisons between the cervical squamous cell cancers and adenocarcinomas were made using Fisher’s exact test. Significance is defined at the P ⬍ 0.05 level. RESULTS Chromosome 4 LOH—All cervical cancers. Twenty-two of 24 (92%) invasive cervical cancers showed LOH at 1 or more of the 19 chromosome 4 loci analyzed. LOH within 1 or more loci on the short arm (p arm, n ⫽ 7 markers) of chromosome 4 was seen in 13 cases (54%) and LOH on the long arm (q arm, n ⫽ 12 markers) in 21 cases (88%). A summary of the overall frequency of LOH at each locus, which was calculated as the number of cases with LOH at that locus/the total number of informative cases at that
93
CHROMOSOME 4 DELETIONS IN INVASIVE CERVIX CANCER
FIG. 3. Autoradiographs of cases demonstrating LOH at the four regions of nonoverlapping deletion (R1, R2, R3, R4). Only the deletions are depicted in this figure. Arrowheads, LOH; N, normal; T, tumor.
locus, is depicted in Fig. 2. All loci tested demonstrated at least 1 tumor with LOH within that locus. Of the 19 loci tested, D4S426 and the neighboring locus D4S171 (4q34 – q35) demonstrated LOH in the majority (70 and 67%, respectively) of cervical cancers. LOH was seen at a higher frequency within the q arm compared to the p arm. One SCCA demonstrated extensive deletions likely involving loss of the whole q arm (Fig. 2, Case No. 16). LOH results by region (R). The pattern of the partial deletions was used to identify four discrete minimal regions of nonoverlapping deletions (Fig. 2). As our group previously described, R1 and R2 were designated on the q arm and R3 and R4 on the p arm of chromosome 4 [7, 8]. A region (R) was defined when at least one case demonstrated LOH at a specific locus and at least one of its two flanking markers demonstrated retention of heterozygosity. Region (R) loss was calculated as the number of cases with LOH at any marker within a defined region/number of informative cases at any marker within that region. The noncontiguous deletions seen in these 24 cases defined four minimal areas of deletion (R1, R2, R3, R4) (Table 1, Figs. 2 and 3). On 4q, one minimal region of deletion involving markers D4S171 and D4S426 (R1, 4q34 – q35) was observed in 13 of 17 informative cases (76%) of cervical cancers overall (Figs. 2 and 3). The other minimal region of deletion was observed between markers D4S175 and D4S194 (R2, 4q25– q26), with LOH observed in 10 of 21 informative cases (48%) of cervical cancer (Figs. 2 and 3). However, the data suggest that this region may itself contain two minimal regions of deletion, one centered on D4S175 and the other centered on D4S194 (Fig. 2). Of note, marker D4S194 tends to highlight multiple bands/fragments of DNA, not discrete alleles. Although this
may appear to represent a microsatellite alteration within the tumor, a similar appearance was seen in both control genomic DNA and lung tumor DNA in our previous studies [7]. On 4p, discrete regions of nonoverlapping deletions were somewhat more difficult to assign as precisely as those on 4q, due to a higher rate of uninformative loci. One minimal region of deletion was assigned between markers D4S2366 and D4S404 (R3, 4p15.1–p15.3), with regional loss observed in 8 of 22 informative cases (36%) (Figs. 2 and 3). The other region was designated between D4S43 and D4S127 (R4, 4p16), with LOH observed in 5 of 19 informative cases (26%) (Figs. 2 and 3). However, a significant marker on the short arm was D4S2366, positioned at the junction of R3 and R4 (and potentially located within either of these two regions). Allelic losses at this specific locus occurred at a much higher rate in ACs (50%) than in SCCAs (10%) (Fig. 2). LOH results by histologic type. Overall, SCCAs demonstrated allelic loss for at least 1 chromosome 4 marker in 12 of 13 cases (92%); similarly ACs displayed loss in 10 of 11 cases (91%). Focusing at the short arm specifically, SCCAs showed LOH in 4 of 13 tumors (31%), whereas ACs showed LOH in 9 of 11 (82%) (P ⫽ 0.02, Fisher’s exact test). For the long arm of chromosome 4, losses were more equally distributed, with 11 of 13 SCCAs (85%) and 10 of 11 ACs (91%) demonstrating LOH. The distinction between histologies, however, was pronounced at the telomeric regions of chromosome 4. At the q telomere (R1 region, locus D4S171) LOH for SCCAs was seen in 5 of 6 informative cases (83%), whereas ACs showed losses in only 1 of 3 informative cases (33%). At the p telomere (R3 region, locus D4S2366) LOH was observed in ACs in 5 of 10 informative cases (50%) and in SCCAs in 1 of 10 cases (10%) (Fig. 2). TABLE 1 Cervical Cancer Cases by Histology LOH/Informative Cases (% LOH) Markers
Arm
AC (n ⫽ 11)
SCCA (n ⫽ 13)
D4S43 D4S136 D4S127 D4S2366 D4S1546 D4S404 D4S174 D4S392 D4S1089 D4S175 D4S1586 D4S194 D4S1584 D4S1554 D4S1535 D4S408 D4S171 D4S426 D4S1652
p p p p p p p q q q q q q q q q q q q
1/2 (50) 1/4 (25) 2/7 (29) 5/10 (50) 1/3 (33) 2/11 (18) 2/6 (33) 4/8 (50) 2/6 (33) 3/6 (50) 1/3 (33) 1/3 (33) 1/7 (14) 1/6 (17) 2/4 (50) 2/4 (50) 1/3 (33) 4/6 (67) 0/4 (0)
1/5 (20) 1/3 (33) 2/7 (29) 1/10 (10) 1/3 (33) 1/10 (11) 0/6 (0) 4/10 (40) 1/6 (17) 3/7 (43) 0/6 (0) 2/8 (25) 2/9 (22) 4/7 (57) 1/4 (25) 1/3 (33) 5/6 (83) 3/4 (75) 6/10 (60)
94
SHERWOOD ET AL.
TABLE 2 HPV Analysis of Adenocarcinoma and Squamous Cell Carcinoma
HPV-positive cases HPV 16 (percentage of cases positive) HPV 18 (percentage of cases positive)
AC (%)
SCCA (%)
8/11 (73) 4/8 (50) 4/8 (50)
11/13 (85) a 6/11 (55) 1/11 (9)
a
The remainder of the cases positive for HPV general primer was not further classified.
Detection of HPV sequences. Overall, 19 of 24 (79%) cervix cancers were positive for oncogenic HPV sequences using general primers (Table 2). HPV subtype distribution was as follows: SCCAs contained HPV DNA in 11 of 13 (85%) tumors. Of these cases, 6 (55%) were HPV 16 positive, while 1 (9%) was HPV 18 positive. The remainder of HPV-positive SCCAs were likely associated with other oncogenic HPV types, such as types 31, 33, and 35, but were not further classified. ACs contained HPV DNA in 8 of 11 (73%) cases, all demonstrating infection with the “classic” oncogenic subtypes, with 4 (50%) containing HPV 16 and the other 4 (50%) HPV 18. Coinfection with both HPV16 and 18 was not seen in any of the cases studied. DISCUSSION Epidemiologic risk factors that are common to both lung and cervical cancer may target similar epigenetic events leading to their development. Based on published data of genomic regions known to undergo frequent alteration in both cervical and lung cancer, we conducted the present analysis. Chromosomes 3p, 5p, 11q, 4p, and 4q are known to contain sites of frequent loss of heterozygosity and/or microsatellite instability in cervical neoplasia [1, 3, 20 –22]. Progression from preinvasive to invasive disease has also been associated with accumulation of such alterations [6, 20]. One prior study has suggested the presence of at least two independent tumor suppressor gene regions mapping to distal regions of chromosome 4, commonly altered in cervical neoplasia [22]. Recently, nonoverlapping deletions on chromosome 4 were found to be highly prevalent in small-cell and non-small-cell lung cancers, suggesting the presence of at least three distinct suppressor genes [7]. Studies in cervical cancer tend to include a predominance of SCCA or preinvasive lesions, and infrequently include an adequate number of invasive adenocarcinoma or adenocarcinoma in situ lesions. Given the known molecular and epidemiologic parallels between lung and cervical carcinoma, we performed a detailed LOH analysis on chromosome 4 using nearly equivalent numbers of SCCAs and ACs. In conjunction with novel LOH findings, our HPV findings were somewhat surprising. Although we expected the majority of SCCAs to be infected with oncogenic HPV [23], the ACs showed an equivalently high infectivity rate, with an equal
distribution between the classic oncogenic HPV 16 and HPV 18. Due to this high HPV infection rate, we found no correlation between the presence and/or oncogenic type of HPV with particular LOH events. Of interest is that 85% of SCCAs contained HPV, but only 55% were HPV 16, the subtype most commonly associated with this histologic type [23]. The ACs were positive for HPV in 73%, higher than the expected rate of 50% [24]. One-half of these were HPV 16 positive, whereas we would expect the majority to contain HPV 18 [23]. The differences seen in this cohort may reflect the geographic regions from where the cases were obtained. Further, the differences in expected percentage of HPV infection may be related to the small sample size in this pilot study. Regarding chromosomal alterations in cervix cancer, the literature to date largely describes a random sampling of chromosome 4 regions of loss. In general, many studies have utilized few markers per chromosome targeting a broader genome-wide search for putative tumor suppressor gene sites. In addition, particularly within the earlier studies, little emphasis has been placed on microdissected tumor specimens. Microdissection has permitted detection of allelic loss at higher frequencies than has been previously recognized [25]. Particularly for the ACs, this technique substantially decreases contamination with intervening stromal lymphocytes. In both SCCA and AC we observed frequent losses at four nonoverlapping sites located on the long arm (two regions) and short arm (two regions) of chromosome 4. On the q arm, the distinct minimal regions of deletion appeared to lie at 4q34 – q35 (“R1”) and 4q25– q26 (“R2”). These at least partially overlap regions of deletion in this area described in other cancer types, including breast, lung, and bladder. For example, 4q33– q34, 4q25– q26, and 4p15.1–p15.3 represented significant regions of deletion in SCLC and NSCLC [7, 8, 26]. Losses in cervical SCCA occurred predominately in the R1 region on the q arm. Overall percentage LOH on the q arm, however, was equivalent in the two subtypes (85 and 91%). In R2, the low rate of LOH observed (11%) with marker D4S1586 could simply be due to the low rate of marker informativity. Alternatively, two distinct TSGs could exist in this area, one approximating D4S194 and the other approximating D4S175. On the p arm, deletion regions were more difficult to designate as we unexpectedly observed a disproportionate number of noninformative cases for each polymorphic marker. As such, our description of regions R3 and R4 was somewhat limited. However, there was a significant difference in LOH events between AC and SCCA on the p arm (P ⫽ 0.02). The pivotal marker D4S2366, for which we observed a predominance of LOH in AC (50%) and rare losses in SCCA (10%), appears to lie between R3 and R4 and may target the region for a more detailed deletion analysis in future studies. This is a first description in cervical cancer of distinct regions of loss on a chromosome other than 3p [6]. Our results suggest the possible existence of at least two tumor suppressor genes on 4q and at least two on 4p. Of additional interest is the region represented by marker
CHROMOSOME 4 DELETIONS IN INVASIVE CERVIX CANCER
D4S392 (4q12–13), the most centromeric 4q marker used in our study. SCCA and AC demonstrated a combined rate of LOH in 8 of 18 cases (44%). The deletions identified could target the centromeric region or possibly extend to proximal p (Fig. 2, Case No. 1). In bladder carcinoma, this region has also demonstrated LOH in a moderate number of tumors, providing evidence for a putative suppressor gene [26]. The various patterns of chromosome 4 loss (with particular attention to the p arm) appear to be unique to the respective tumor histologies. To our knowledge this is the first report of a significant regional LOH difference between adenocarcinomas and squamous carcinomas of the uterine cervix. Deletion hotspots predominated at distal regions of the q arm and p arm for SCCA and AC, respectively. In lung cancers, interestingly, losses observed in the three putative suppressor gene regions on chromosome 4 were common to both small-cell (SCLC) and non-small-cell lung cancers (NSCLC) [7]. Although in this study NSCLCs were not stratified further, the frequency of LOH overall in NSCLC was much lower than that in SCLCs, possibly explaining differences in their pathogenetic mechanisms. The clinical distinction between cervical cancer variants is also likely supported by specific molecular events. Certain gene– environment interactions may be specific to certain chromosomal regions. For example, HPV infection and tobacco smoking may target distal 4q (4q34 – q35); as these environmental risks are associated with both AC and SCCA (and both histotypes share a high LOH rate in distal 4q), this finding could explain a common mechanism of carcinogenesis. That 4p LOH is unique to adenocarcinoma, however, may represent an environmental/clinical risk significant to AC. We suspect that hormonal effects may target the 4p gene region, given that adenocarcinoma of the cervix is histologically similar to estrogen-influenced endometrial adenocarcinoma. Our results have identified four discrete deletion hotspots that may harbor tumor suppressor genes implicated in cervical cancer. We suspect there may exist a target gene on 4p specific to AC and one or more within the regions of 4q involved in overall cervical cancer development. A candidate gene on chromosome 4q should now be sought to provide insight into cervical carcinogenesis.
95
sion of precancerous lesions of the uterine cervix. J Natl Cancer Inst 87:742–745, 1995 3. Rader JS, Kamarasova T, Huettner PC, Li L, Li Y, Gerhard DS: Allelotyping of all chromosomal arms in invasive cervical cancer. Oncogene 13:2737–2741, 1996 4. Lin WM, Michalopulos EA, Dhurander N, Cheng PC, Robinson WR, Ashfaq R, Coleman RL, Muller CY: Allelic loss and microsatellite alteration of chromosome 3p14.2 are more frequent in recurrent cervical dysplasias. Clin Cancer Res 6:1410 –1414, 2000. 5. Fong KM, Biesterveld EJ, Virmani A, Wistuba II, Sekido Y, Bader SA, Ahmadian M, Ong ST, Rassool FV, Zimmerman PV, Giacconne G, Gazdar AF, Minna JD: FHIT and FRA3B 3p14.2 allele loss are common in lung cancer and preneoplastic bronchial lesions and are associated with cancer-related FHIT cDNA splicing aberrations. Cancer Res 57:2256 – 2267, 1997 6. Wistuba II, Montellano FD, Milchgrub S, Virmani AK, Behrens C, Chen H, Ahmadian M, Nowak JA, Muller CY, Minna JD, Gazdar AF: Deletions of chromosome 3p are frequent and early events in the pathogenesis of uterine cervical carcinoma. Cancer Res 57:3154 –3158, 1997 7. Shivapurkar N, Virmani AK, Wistuba II, Milchgrub S, Mackay B, Minna JD, Gazdar AF: Deletions of chromosome 4 at multiple sites are frequent in malignant mesothelioma and small cell lung carcinoma. Clin Cancer Res 5:17–23, 1999 8. Shivapurkar N, Sood S, Wistuba II, Virmani AK, Maitra A, Milchgrub S, Minna JD, Gazdar AF: Multiple regions of chromosome 4 demonstrating allelic loss in breast carcinomas. Cancer Res 59:3576 –3580, 1999 9. Ning Y, Weber JL, Killary AM, Ledbetter DH, Smith JR, Pereira-Smith OM: Genetic analysis of indefinite division in human cells: Evidence for a cell-senescence related gene(s) on human chromosome 4. Proc Natl Acad Sci USA 88:5635–5659, 1991 10. Heselmeyer K, Macville M, Schrock E, Blegen H, Hellstrom AC, Shah K, Auer G, Ried T: Advanced-stage cervical carcinomas are defined by a recurrent pattern of chromosomal aberrations revealing high genetic instability and a consistent gain of chromosome arm 3q. Genes Chromosomes Cancer 19:233–240, 1997 11. Kirchhoff M, Rose H, Peterson BL, Maahr J, Gerdes T, Lundsteen C, Bryndorf T, Kryger-Baggesen N, Christensen L, Engelholm SA, Philip J: Comparative genomic hybridization reveals a recurrent pattern of chromosomal aberrations in severe dysplasia/carcinoma in situ of the cervix and in advanced-stage cervical carcinoma. Genes Chromosomes Cancer 24:144 –150, 1999 12. Dellas A, Torhorst J, Jiang F, Proffitt J, Schultheiss E, Holzgreve W, Sauter G, Mihatsch MJ, Moch H: Prognostic value of genomic alterations in invasive cervical squamous cell carcinoma of clinical stage IB detected by comparative genomic hybridization. [In Process Citation] Cancer Res 59:3475–3479, 1999
ACKNOWLEDGMENT
13. Berchuck A: Molecular basis of endometrial cancer. Cancer 2034 –2040, 1995
We thank Dr. Jan A. Nowak (Department of Pathology, Illinois Masonic Medical Center, Chicago, IL) for providing the paraffin-embedded archival squamous carcinomas used in this study.
14. Berchuck A, Carney M: Human ovarian cancer of the surface epithelium. Biochem Pharmacol 54:541–544, 1997
REFERENCES 1. Muller CY, O’Boyle JD, Fong KM, Wistuba II, Biesterveld E, Ahmadian M, Miller DS, Gazdar AF, Minna JD: Abnormalities of fragile histidine triad genomic and complementary DNAs in cervical cancer: association with human papillomavirus type [see comments]. J Natl Cancer Inst 90:433– 439, 1998 2. Mitra AB, Murty VV, Singh V, Li RG, Pratap M, Sodhani P, Luthra UK, Chaganti RS: Genetic alterations at 5p15: a potential marker for progres-
15. Emmert-Buck MR, Bonner RF, Smith PD, Chuaqui RF, Zhuang Z, Goldstein SR, Weiss RA, Liotta LA: Laser capture microdissection. [see comments] Science 274:998 –1001, 1996 16. Hung J, Kishimoto Y, Sugio K, Virmani AK, McIntire DD, Minna JD, Gazdar AF: Allele-specific chromosome 3p deletions occur at an early stage in the pathogenesis of lung carcinoma. JAMA 273:1908, 1995 17. Gyapay G, Morissette J, Vignal A, Dib C, Fizames C, Millasseau P, Marc S, Bernardi G, Lathrop M, Weissenbach J: The 1993–94 genethon human genetic linkage map. [see comments] Nat Genet 7:246 – 339, 1994 18. De Roda Husman AM, Walboomers JM, van den Brule AJ, Meijer CJ,
96
SHERWOOD ET AL. Snijders PJ: The use of general primers GP5 and GP6 elongated at their 3⬘ ends with adjacent highly conserved sequences improves human papillomavirus detection by PCR. J Gen Virol 76:1057–1062, 1995
19. Baay MF, Quint WG, Koudstaal J, Hollema H, Duk JM, Burger MP, Stolze E, Herbrink P: Comprehensive study of several general and typespecific primer pairs for detection of human papillomavirus DNA by PCR in paraffin-embedded cervical carcinomas. J Clin Microbiol 34:745–747, 1996 20. Larson AA, Liao SY, Stanbridge EJ, Cavenee WK, Hampton GM: Genetic alterations accumulate during cervical tumorigenesis and indicate a common origin for mulifocal lesions. Cancer Res 57:4171– 4176, 1997 21. Mitra AB, Murty VV, Li RG, Pratap M, Luthra UK, Chaganti RS: Allelotype analysis of cervical carcinoma. Cancer Res 54:4481– 4487, 1994 22. Hampton GM, Larson AA, Baergen RN, Sommers RL, Kern S, Cavenee
23.
24.
25.
26.
WK: Simultaneous assessment of loss of heterozygosity at multiple microsatellite loci using semi-automated fluorescence-based detection: subregional mapping of chromosome 4 in cervical carcinoma. Proc Natl Acad Sci USA 93:6704 – 6709, 1996 Iwasawa A, Nieminen P, Lehtinen M, Paavonen J: Human papillomavirus DNA in tuerine cervix squamous cell carcinoma and adenocarcinoma detected by polymerase chain reaction. Cancer 77:2275–2279, 1996 Leminen A, Paavonen J, Vesterinen E, Wahlstrom T, Rantala I, Lehtinen M: Human papillomavirus types 16 and 18 in adenocarcinoma of the uterine cervix. Am J Clin Pathol 95:647– 652, 1991 Fujii H, Zhou W, Gabrielson E: Detection of frequent allelic loss of 6q23– q25.2 in microdissected human breast cancer tissues. Genes Chromosomes Cancer 16:35–39, 1996 Polascik TJ, Cairns P, Chang WY, Schoenberg MP, Sidransky D: Distinct regions of allelic loss on chromosome 4 in human primary bladder carcinoma. Cancer Res 55:5396 –5399, 1995