Identification of a 6-cM Minimal Deletion at 11q23.1–23.2 and Exclusion of PPP2R1B Gene as a Deletion Target in Cervical Cancer¹

CC³is a substantial public health issue among women worldwide, causing high mortality, and the incidence is rising in certain countries(1). Most invasive CCs are believed to be preceded by distinct preinvasive changes called CINs (CINI–CINIII), which represents a pathological continuum from mild to severe epithelial dysplasias. Biological behavior of cervical precursor lesions vary in which only a fraction of higher-grade dysplastic and preinvasive lesions progress to invasive cancer (2, 3, 4, 5). Most precancerous and preinvasive lesions are readily curable, whereas the prognosis of invasive CCs are generally poor. The genetic events that initiate the multistep pathway in cervical tumorigenesis and cause invasion are of considerable importance in understanding the molecular basis of CCs. A large body of evidence has implicated infection of high-risk HPV types as the critical etiological factor in CCs in which HPV E6 and HPV E7 proteins interact with critical cell cycle check point genes p53 and pRB, respectively, resulting in inactivation of these genes (6). However,epidemiological data suggest that only a certain fraction of HPV-infected CIN lesions progress to invasive CCs with variable latency periods (2, 3, 4). These data, therefore, suggest that additional genetic alterations may be necessary for the progression of CCs. Delineation of such genetic changes may be of relevance in understanding cervical carcinogenesis and will have implications in early detection and identification of high-risk lesions.

Molecular genetic studies of CCs have identified frequent LOH affecting multiple chromosomal regions such as 3p, 5p, 6p, and 11q(7, 8, 9, 10), suggesting the presence of TSGs in these regions. The functional evidence for the presence of a TSG in CC was first identified on chromosome 11 by somatic cell hybrid studies using HeLa cells (11) and subsequent analysis in other CC cell lines(12, 13, 14). Pursuing these studies, chromosome 11 has been shown consistently to exhibit LOH at different regions, 11q13 and 11q22–23 (15, 16, 17). Although these data provide strong evidence for the presence of a TSG on chromosome 11 relevant to CC, the critical region or the gene involved in the development of this tumor is not yet identified. In the present study, we defined a common minimal deletion at 11q23 that spans a 6-cM genetic distance by LOH analysis and excluded the PPP2R1B gene at 11q23 as a target of deletion, which was shown earlier to be mutated in other tumor types(18).

A total of 58 tumor biopsies derived from previously untreated primary invasive CCs and the corresponding peripheral blood samples comprised the material for the study. The tissues were ascertained from patients treated at the Instituto Nacional de Cancerologia (Santa Fe de Bogota,Colombia) after appropriate informed consent and the approval of the protocol by the institutional review board. Clinically, the tumors were classified by Fédération International des Gynaecologistes et Obstetristes class 1B (n = 3), IIB(n = 15), IIIB (n = 36), and IV (n = 4). Histologically, 56 tumors were classified as squamous cell carcinomas and 2 as adenocarcinomas. The ages of the patients ranged from 28 to 85 years,with a median of 49 years.

High molecular weight DNA from tumor and peripheral blood samples were isolated using standard procedures of proteinase K digestion,phenol-chloroform extraction, and ethanol precipitation. A panel of 25 dinucelotide polymorphic markers were chosen on the basis of their map position and heterozygosity (Table 1; Gene Map 99⁴) and were obtained from Research Genetics (Huntsville, AL). A standard PCR reaction was carried out in a 10-μl reaction volume containing 1.5–2.5 mm MgCl₂, 10–15% glycerol,4 pmol of each primer (one-fifth of one of which was end-labeled with[γ-³²P]dATP), 0.2 mmdeoxynucleotide triphosphates, 25 ng of DNA, and 0.3 unit of AmpliTaq DNA polymerase (Perkin-Elmer Corp., Branchburg, NJ). The amplification was carried out for 30 cycles at annealing temperatures ranging from 50 to 58°C. The PCR products were denatured in sequence stop buffer containing formamide and electrophoresed on a 6%urea-containing polyacrylamide gel, and the dried gels were autoradiographed for 4–16 h. Criteria applied for scoring LOH was described earlier (19). All autoradiograms were independently scored visually by three investigators (H. A. P.,F. M., and V. V. V. S. M.). The definition of minimal region of deletion was based on LOH of the loci that span common deletion in several tumors and retention of heterozygosity of adjoining markers at both the boundaries in at least two tumors.

SSCP analysis was performed on all 15 exons of the PPP2R1Bgene using primers flanking intronic sequences (18). PCR was performed as described above except that 0.5 μCi of α[³²P]dCTP was included in the reaction instead of labeled primers. The PCR products were diluted in 0.1% SDS/10 mm EDTA, denatured in sequencing stop buffer, and run overnight in 6% nondenaturing polyacrylamide gels containing 10%glycerol at room temperature. Dried gels were autoradiographed and examined for conformational changes. Automated sequencing was performed on exons that showed suggestive mutations using purified PCR products.

In the present study, a panel of 58 paired normal–tumor DNAs from CCs were assayed for LOH using 25 polymorphic sequenced-tagged sites (STSs) mapped to chromosome 11 (5 on 11p and 20 on 11q; Table 1). The analysis revealed deletions in at least one locus in 46 (79%) tumors. Of these, 4 tumors (T-46, T-48, T-52,and T-114) showed loss of one allele at all informative loci,suggesting genetic monosomy of chromosome 11. The remaining 42 tumors showed LOH at one or more loci while retaining heterozygosity at the remaining loci, suggesting regional losses. These tumors with regional losses were used to identify a pattern of minimal deletions. The frequency of LOH varied among the markers from 14.6%(D11S911 at 11q13) to 60.6% (D11S4094 at 11q23.2; Table 1). The pattern of LOH identified multiple regions of discrete deletions both on the 11p and 11q. The frequency of LOH did not show any correlation with age, tumor stage, size, or histology.

Twenty-three (39.7%) tumors showed deletions on the 11p, and 8 (4 with complete monosomy of chromosome 11) of these showed LOH at all informative loci, suggesting monosomy of 11p. The remaining 15 tumors exhibited partial deletions. The patterns of LOH identified two regions of minimal deletion, one at 11p15.11 and the other at 11p12–13. The 11p15 deletion spanned by the locus D11S4189 showed LOH in 33.3% of informative cases where tumors T-29 and T-79 exhibited deletions while retaining heterozygosity at a proximal marker D11S4099. The second site of deletion at 11p12–13 spanned by the markers D11S4200 and D11S4083 exhibited LOH in 26 and 22%, respectively. This deletion was defined by tumors T-117 and T-75, with distal marker D11S4096 retaining heterozygosity, whereas 3 other tumors (T-92, T-28 and T-29) exhibited LOH at D11S4083 and retained heterozygosity at D11S4200. Thus, the patterns of LOH on 11p identified two discrete regions of deletions (Fig. 1).

The 11q LOH was observed in 40 of 58 (69%) tumors studied. In addition to 4 tumors with complete loss of chromosome 11, two others (T-105 and T-116) showed LOH at all informative loci on 11q, suggesting monosomy. The pattern of LOH in the remaining 34 tumors with interstitial allelic deletions on 11q identified one common region of loss at 11q23.1–q23.2(Fig. 2). Of the 34 tumors with partial 11q LOH, 24 tumors exhibited deletions at 11q23.1–23.2, and the remaining 10 showed scattered deletions outside the 11q23 region (Fig. 2). The common deleted region at 11q23.1–q23.2 spanned the loci D11S4167 (45.7% LOH), D11S1353 (58.5% LOH), D11S4094 (60.6% LOH), and D11S4144 (46.8% LOH). Boundaries of this deletion flanked by the marker D11S925 proximally with retention of heterozygosity in 4 tumors (T-16, T-28, T-40, and T-112) and D11S933 distally with retention of heterozygosity in two tumors (T-110 and T-76; Fig. 3).

Mutations in the PPP2R1B gene at 11q23 were earlier reported in lung and colon carcinomas as a target of deletions in this region(18). To determine whether the PPP2R1B gene is the target of 11q23 deletions in CCs, we performed mutation analysis of the entire coding region on a panel of 30 tumor DNAs that exhibited LOH on 11q23. All suspected SSCP variants were sequenced in both orientations to identify the nature of the mutations. This analysis identified three different types of sequence alterations in three tumors: the G → C transversion that changes Ala₆₀₁ → Pro (T-98); G → A transition that changes Arg₄₁₁ → Gly (T-28, T-92); and G → A transition that changes Gly₉₀ → Asp (T-98). The Arg₄₁₁ → Gly and Gly₉₀→ Asp alterations in both tumors were found to be homozygous, whereas the Ala₆₀₁ → Pro was heterozygous. The changes Ala₆₀₁ → Pro and Arg₄₁₁→ Gly were found to be population variants because the similar alterations were also seen in the corresponding control and unrelated normal individuals (data not shown). The Gly₉₀→ Asp change was also found to be present in the corresponding normal DNA from the patient T-98. The sequence analysis suggests that the wild-type (normal) allele have been deleted in the T-98 tumor DNA,consistent with the observation of LOH at 11q23 (Fig. 2). The residual GGC₉₀ signals seen in T-98 DNA may represent contaminating normal cells in the tumor tissue compared with normal DNA from blood from the same patient (Fig. 4). Thus, the normal allele of the PPP2R1B is deleted by LOH in T-98, and the variant allele is retained in this tumor.

Functional studies have identified that chromosome 11 carries a gene(s) responsible for tumor suppression in CC (11, 12, 13, 14). Allelotype studies have further identified high frequency of LOH on 11q, consistent with the presence of one or more TSGs on this chromosomal arm (7, 8, 9, 10, 15, 16, 20). These studies,however, suffer from lack of extensive and systematic deletion mapping analysis in delineating the exact location of the TSG. In CC, only a few previous studies have performed detailed deletion mapping in an attempt to identify the critical regions of deletion. These studies identified at least two putative TSG sites at 11q13 and 11q22–24(15, 17, 21). Hampton et al. (17)using 16 polymorphic markers on chromosome 11 in 32 patients have identified a minimal deleted region to a 35-cM genetic distance at 11q23. More recently, Mugica-Van Herckenrode et al.(21) have further restricted the region of deletion to a 19-cM genetic distance at the 11q23 region using only four markers at this region. These studies, thus, identified candidate TSG sites at 11q23 that may be critical for the tumor formation in cervix uteri.

To further define the critical regions of deletion on chromosome 11, we performed a systematic and high-density LOH mapping. The pattern of allelic losses in the present study identified three discrete regions of minimal deletions, two on 11p and one on 11q. The 11p deletions were localized at p15.11 (13 of 39 informative cases; 33.3% LOH) and p12–13 (17 of 52 informative cases; 32.7% LOH). The 11q deletion was localized at 11q23.1–23.2 (30 of 57 informative cases; 52.6% LOH). The 11q23 region has been implicated previously in CC, whereas the 11p deletions have not been reported earlier. Although the frequencies of allelic deletions at 11p15 and 11p12–13 regions were low, the 11q23.1–23.2 region exhibited high frequency of LOH.

The short arm of chromosome 11 between 11p11 and 11p15 has been shown to contain genes controlling the regulation of HPV-16 viral early promoter (22). Whether the LOH seen at 11p in the present study plays any role in HPV-induced transformation remains to be seen. The 11p15 and 11p13 regions also exhibit genetic alterations in a variety of tumor systems (23). At present, the significance of the low frequency of LOH in these regions is not clear,and additional studies are required to understand the importance of these genetic alterations.

In the present study, the 11q23 region exhibited the highest frequency of LOH (60.6% LOH at D11S4094), suggesting that this region contains a TSG critical for CC development. A number of previous studies have also identified frequent LOH at 11q22–24 regions in CC(9, 10, 17, 20, 21).

The 11q22-q24 region has also been implicated in a variety of other solid tumor types including carcinomas of breast (24, 25, 26, 27, 28, 29),lung (30, 31), colon (32), ovary (33, 34), and oral (35, 36) carcinomas, malignant melanoma (37, 38), and neuroblastoma (39). The chromosome 11q deletion maps generated by these studies have been reviewed, and the results are summarized in Fig. 5. That 11q22–23 is the most frequent target of LOH in solid tumors suggests that this region contains one or more TSGs. Furthermore, review of the previous studies revealed at least three distinct regions of deletions at 11q22–23, a proximal region targeting between 106.5 and 116 cM, a second region between 121 and 127 cM, and the distal region between 137 and 145 cM. The proximal deletion is frequently involved in breast,lung, and malignant melanomas, and the middle region is a target in cervical and breast carcinomas (Fig. 5). The third and the distal regions between 137 and 145 cM may be a target in colon, ovarian, and oral carcinomas. These data, thus, suggest that there are at least three TSGs that may be the target of 11q23 deletions in various solid tumor types.

The PPP2R1B gene, which maps to the proximal 11q22–23 region between 106.5 and 116 cM, encodes the β-isoform subunit of serine/threonine protein phosphatase 2A and was found to be mutated in 15% of lung and colon carcinomas (18). This region has also been implicated previously in CC (17, 21). To find out whether PPP2R1B is a candidate TSG for CC, we performed mutation analysis of the entire gene and found four cases with a single nucleotide alteration. Three of the four sequence alterations were confirmed as polymorphisms because they were also seen in matched normal DNAs and in unrelated controls. The fourth case with codon GGA₉₀ → GAC alteration was also present in matched-normal DNA but not in 100 additional normal DNAs (data not shown), suggesting that this may represent a germ-line mutation. LOH analysis on this patient tumor DNA showed loss of a normal wild-type allele, suggesting that this may represent a pathogenic alteration. To see whether this change is inherited, we studied four members of the family (son, daughter, sister, and nephew) and found that the son and daughter inherited the same variant but not by the sister and nephew. The PPP2R1B Gly₉₀ → Asp alteration was reported earlier in three tumors of non-small cell lung carcinomas as pathogenic alteration (18). In another recent study of ovarian carcinomas by Campbell and Manolitsas (40), this change was reported as a nonpathological polymorphism. Our data,however, support the view that the PPP2R1BGly₉₀ → Asp alteration may represent a germ-line change, and individuals carrying this change may be susceptible to developing cervical or other types of solid tumors. This, however, remains to be tested in families with inherited cancer. The present data, however, conclusively exclude PPP2R1B as a target of 11q23 somatic deletions in CC.

In the present study, we defined the minimal region of deletion at 11q23 in CC to span a 6-cM genetic distance between 123.4 and 129.5 cM interval (Fig. 5). This region is currently known to contain five genes. These include U90916 mRNA of unknown function, zinc finger protein 202 (ZNF202), LOH 11 chromosomal region 2(LOH11CR2A), similar to heat shock cognate M_r 70,000 protein 10(HSC71), and a protein kinase C substrate RC3(neurogranin). Of these, the genes LOH11CR2A and ZNF202 have been examined as 11q23 deletion targets in breast and lung carcinomas. Both the LOH11CR2A and ZNF202 genes were identified in an effort to find a TSG in the region between loci D11S1345 and D11S1328deleted in breast carcinoma (24, 41). However, no genetic alterations in the support of the tumorigenic potential of these genes could be established in breast carcinoma (24, 41). In addition to these known genes, at least 28 UniGene clusters were mapped to this interval of minimal deletion (Gene Map’99⁵).

In the present study, we have defined a minimal region of deletion at 11q23 in CC that may contain the critical TSG and excluded a previously known tumor suppressor PPP2R1B as a target of 11q23 deletions in CC. The current definition of minimal deletion with the available genomic resources in this region should aid in the identification of the candidate TSG.

We thank the gynecologists at Instituto Nacional de Cancerología, Santa Fe de Bogota, Colombia, for the collection of tissue specimens.