Search

Clinical and molecular findings suggest that the four major histological subtypes of ovarian carcinoma (serous, clear cell, mucinous, and endometrioid) likely represent distinct disease entities. Prior studies have shown that ovarian endometrioid adenocarcinomas (OEAs) often carry mutations in the CTNNB1 gene, which encodes beta-catenin, a critical component of the Wnt signaling pathway. However, the nature of other defects in the Wnt signaling pathway in ovarian carcinomas remains largely unknown. Thus, in 45 primary OEAs and two OEA-derived cell lines, we sought to comprehensively address the prevalence of and mechanisms underlying beta-catenin and Wnt pathway deregulation. CTNNB1 missense mutations were detected in 14 primary tumors. All mutations affected the NH(2)-terminal regulatory domain of beta-catenin, presumably rendering the mutant proteins resistant to degradation. Immunohistochemical studies revealed nuclear accumulation of beta-catenin in all but two tumors with CTNNB1 mutations. Two primary tumors lacking CTNNBI mutations showed strong nuclear immunoreactivity for beta-catenin. In one of the two tumors, biallelic inactivation of the APC gene was found. In the remaining 29 primary OEAs, unequivocal nuclear beta-catenin immunoreactivity was not observed, though a nonsense mutation in AXIN1 was observed in one tumor and a truncating frameshift mutation in AXIN2 was seen in another case. Both OEA-derived cell lines studied (TOV-112D and MDAH-2774) had elevated constitutive T-cell factor/lymphoid enhancer factor transcriptional activity. TOV-112D cells were shown to harbor mutant beta-catenin, whereas a missense AXIN1 sequence alteration was identified in MDAH-2774 cells. Collectively, our findings demonstrate frequent defects of the Wnt signaling pathway in a particular subtype of ovarian carcinomas, i.e., OEAs. Although mutations in the CTNNB1 gene are the most common mechanism of beta-catenin deregulation in OEAs, beta-catenin deregulation may also result from mutations in the APC, AXIN1, and AXIN2 genes.

A better understanding of the molecular circuitry in normal ovarian tissues and in ovarian cancer will likely provide new targets for diagnosis and therapy. Recently, much has been learned about the genes expressed in ovarian cancer through studies with cDNA arrays and serial analysis of gene expression. However, these methods do not allow highly quantitative analysis of gene expression on a large number of specimens. Here, we have used quantitative real-time RT-PCR in a panel of 39 microdissected ovarian carcinomas of various subtypes to systematically analyze the expression of 13 genes, many of which were previously identified as up-regulated in a subset of ovarian cancers by serial analyses of gene expression. The genes analyzed are glutathione peroxidase 3 (GPX3), apolipoprotein J/clusterin, insulin-like growth factor-binding protein 2, epithelial cell adhesion molecule/GA733-2, Kop protease inhibitor, matrix gla protein, tissue inhibitor of metalloproteinase 3, folate receptor 1, S100A2, signal transducer and activator of transcription 1, secretory leukocyte protease inhibitor, apolipoprotein E, and ceruloplasmin. All of the genes were found overexpressed, some at extremely high levels, in the vast majority of ovarian carcinomas irrespective of the subtype. Interestingly, GPX3 was found at much higher levels in tumors with clear cell histology and may represent a biomarker for this subtype. Some of the genes studied here may thus represent targets for early detection ovarian cancer. The gene expression patterns were not associated with age at diagnosis, stage, or K-ras mutation status in ovarian cancer. We find that several genes are coordinately regulated in ovarian cancer, likely representing the fact that many genes are activated as part of common signaling pathways or that extensive cross-talk exists between several pathways in ovarian cancer. A statistical analysis shows that genes commonly up-regulated in ovarian cancer may result from the aberrant activation of a limited number of pathways, providing promising targets for novel therapeutic strategies.

Difficulties in the detection, diagnosis, and treatment of ovarian cancer result in an overall low survival rate of women with this disease. A better understanding of the pathways involved in ovarian tumorigenesis will likely provide new targets for early and effective intervention. Here, we have used serial analysis of gene expression (SAGE) to generate global gene expression profiles from various ovarian cell lines and tissues, including primary cancers, ovarian surface epithelia cells, and cystadenoma cells. The profiles were used to compare overall patterns of gene expression and to identify differentially expressed genes. We have sequenced a total of 385,000 tags, yielding >56,000 genes expressed in 10 different libraries derived from ovarian tissues. In general, ovarian cancer cell lines showed relatively high levels of similarity to libraries from other cancer cell lines, regardless of the tissue of origin (ovarian or colon), indicating that these lines had lost many of their tissue-specific expression patterns. In contrast, immortalized ovarian surface epithelia and ovarian cystadenoma cells showed much higher similarity to primary ovarian carcinomas than to primary colon carcinomas. Primary tissue specimens therefore appeared to be a better model for gene expression analyses. Using the expression profiles described above and stringent selection criteria, we have identified a number of genes highly differentially expressed between nontransformed ovarian epithelia and ovarian carcinomas. Some of the genes identified are already known to be overexpressed in ovarian cancer, but several represent novel candidates. Many of the genes up-regulated in ovarian cancer represent surface or secreted proteins such as claudin-3 and -4, HE4, mucin-1, epithelial cellular adhesion molecule, and mesothelin. Interestingly, both apolipoprotein E (ApoE) and ApoJ, two proteins involved in lipid homeostasis, are among the genes highly up-regulated in ovarian cancer. Selected serial analysis of gene expression results were further validated through immunohistochemical analysis of ApoJ, claudin-3, claudin-4, and epithelial cellular adhesion molecule in archival material. These experiments provided additional evidence of the relevance of our findings in vivo. The publicly available expression data reported here should stimulate and aid further research in the field of ovarian cancer.

Allelic losses involving chromosome 3p are frequently observed in cervical cancers. Deletion mapping studies of primary cervical carcinomas have localized common regions of deletion to 3p14.2 and 3p21. The candidate tumor suppressor gene FHIT has been mapped to 3p14.2, and previous studies have demonstrated reduced or aberrant FHIT transcripts and reduced or absent Fhit protein expression in a large percentage of cervical cancer-derived cell lines and primary cervical carcinomas. To expand these observations to preinvasive cervical epithelial lesions and to determine whether loss of Fhit protein expression might be associated with tumor progression, immunohistochemical methods were used to examine Fhit expression in 95 invasive cervical carcinomas, 33 high-grade squamous intraepithelial lesions (HSILs) associated with concurrent invasive cancer, 38 HSILs unassociated with invasive cancer, 24 low-grade squamous intraepithelial lesions, and 22 normal cervix samples. All normal cervical epithelia and low-grade squamous intraepithelial lesions exhibited diffuse cytoplasmic immunostaining of moderate to strong intensity. Fhit protein expression was markedly reduced or absent in 67 of 95 (71%) invasive cancers, 17 of 33 (52%) HSILs associated with invasive cancer, and 8 of 38 (21%) HSILs without associated invasive cancer. The results confirm that Fhit protein expression is reduced or absent in the majority of cervical carcinomas and suggest that loss of Fhit expression often accompanies cervical tumor progression. Moreover, absent or reduced Fhit protein is observed at a significantly higher frequency in HSILs associated with progression to invasive cancer than in HSILs with unknown risk for progression (P = 0.012). These findings suggest that loss of Fhit expression in HSILs could serve as a useful marker of high-grade preinvasive lesions that have an increased likelihood of progression to invasive carcinoma.

Allelic deletions involving the short arm of chromosome 3 (3p13-21.1) have been observed frequently in cervical carcinomas. Recently, a candidate tumor suppressor gene, FHIT (Fragile Histidine Triad), was cloned and mapped to this chromosomal region (3p14.2). Abnormal FHIT transcripts have been identified previously in a variety of tumor cell lines and primary carcinomas, although their significance and the molecular mechanisms underlying their origin remain incompletely defined. In addition, integration of human papillomavirus DNA has been identified at a fragile site (FRA3B) within the FHIT locus in cervical cancer. These observations motivated us to evaluate FHIT mRNA and protein expression in cervical cancer cell lines, primary cervical carcinomas, and normal tissues. Transcripts of the expected size and sequence were the predominant species identified by reverse transcription (RT)-PCR in cultured keratinocytes and all normal tissues evaluated. In contrast, aberrant FHIT transcripts were readily demonstrated in 6 of 7 cervical carcinoma cell lines and 17 of 25 (68%) primary cervical carcinomas. Northern blot analyses demonstrated reduced or absent FHIT expression in the cervical carcinoma cell lines, particularly those with aberrant RT-PCR products. Immunohistochemical analysis of Fhit expression in cervical tissues revealed strong immunoreactivity in nonneoplastic squamous and glandular cervical epithelium and marked reduction or loss of Fhit protein in 25 of 33 (76%) primary cervical carcinomas. In those cervical cancer cell lines and primary tumors with exclusively aberrant or absent FHIT transcripts by RT-PCR, Fhit protein expression was always markedly reduced or absent. The frequent alterations in FHIT expression in many cervical carcinomas, but not in normal tissues, suggest that FHIT gene alterations may play an important role in cervical tumorigenesis.

Loss of heterozygosity of chromosome 10q has been reported in approximately 40% of endometrial carcinomas. PTEN, a candidate tumor suppressor gene located at chromosome 10q23.3, was recently identified and found to be homozygously deleted or mutated in several different types of human tumors. To determine if PTEN is a target of 10q loss of heterozygosity in carcinomas of the endometrium, we examined 32 primary endometrial carcinomas for mutations in PTEN. The tumors included the two major histopathological types of endometrial carcinoma: endometrioid (n = 26; 14 microsatellite instability (MI)-positive and 12 MI-negative) and serous (n = 6). Overall, mutations were detected in 50% of the endometrial carcinomas we analyzed. Mutations were present in 12 of 14 (86%) MI-positive and 4 of 12 (33%) MI-negative endometrioid tumors. Furthermore, mutations were found in all three histological grades of MI-positive endometrioid carcinoma. All six serous endometrial carcinomas lacked detectable mutations. To evaluate the role of PTEN in other common malignancies of the female genital tract, 12 serous ovarian carcinomas and 10 squamous cervical carcinomas were analyzed and were negative for mutations. Our results support PTEN as a tumor suppressor gene and suggest that mutations in PTEN play a significant role in the pathogenesis of the endometrioid type of endometrial carcinoma.

We recently identified a novel tumor-suppressor gene, DPC4, at chromosome 18q21.1 and found that both alleles of DPC4 were inactivated in nearly one-half of the pancreatic carcinomas. Here, we analyzed 338 tumors, originating from 12 distinct anatomic sites, for alterations in the DPC4 gene. Sixty-four specimens were selected for the presence of the allelic loss of 18q and were further analyzed for DPC4 sequence alterations. An alteration of the DPC4 gene sequence was identified in one of eight breast carcinomas and one of eight ovarian carcinomas. These results indicate that whereas DPC4 inactivation is prevalent in pancreatic carcinoma (48%), it is distinctly uncommon (< 10%) in the other tumor types examined. The tissue restriction of alterations in DPC4, as in many other tumor-suppressor genes, emphasizes the complexity of rate-limiting checkpoints in human tumorigenesis.

Recent molecular studies suggest that the expression of high-risk but not low-risk human papillomavirus (HPV) oncoproteins E6 and E7 can significantly alter normal cell cycle regulation. The alterations in cell cycle regulation may be reflected by changes in the balance between cell growth and cell loss through apoptosis in cell populations expressing E6 and/or E7. We evaluated the kinetic indices of cell proliferation and apoptosis in a histopathological spectrum of cervical neoplasia and compared low-versus high-risk HPV-associated lesions. The cell proliferation index, as determined by detection of the nuclear antigen Ki67, increased with increasing lesion grade. Apoptotic cells were identified with terminal deoxynucleotidyl transferase-labeling of the 3'-hydroxyl ends of DNA nucleosomes. No apoptosis was observed in normal epithelium, and only occasional apoptotic cells were seen in low-grade lesions. However, there was a low but measurable apoptotic index in the higher grade lesions, which increased with lesion grade. There was no significant difference in the proliferative and apoptotic indices in similar grade lesions when stratified into low-versus high-risk HPV types. These findings suggest that apoptosis in HPV-infected lesions correlates with proliferative activity rather than HPV type.

We have recently demonstrated that mutation of the transforming growth factor-beta (TGF-beta) receptor type II (RII) gene is characteristic of colon cancers exhibiting microsatellite instability or replication errors (RER+). Moreover, we have shown that RII mutations in these RER+ colon cancers are characteristically frameshift mutations within a 10-bp polyadenine repeat present in the RII-coding region. We now show that RII gene mutations in this polyadenine repeat are also commonly present in RER+ gastric cancers (71%). In contrast, we find these same RII gene mutations are distinctly uncommon in RER+ endometrial cancers (17%, P < 0.02). These results suggest that RII gene mutations confer a growth advantage and are selected for in RER+ cancers of both the upper and lower gastrointestinal tract. The genesis of RER+ endometrial tumors must, however, be by a different route.

Endometrial carcinoma is the second most common tumor type in women with hereditary nonpolyposis colorectal carcinoma. Microsatellite instability (MI) has been observed in the inherited (hereditary nonpolyposis colorectal carcinoma-associated) form of endometrial carcinoma as well as in approximately 20% of presumably sporadic cases. Recent studies suggest that MI in many cell lines or xenografts derived from sporadic colorectal carcinomas is not attributable to mutations in four known human DNA mismatch repair (MMR) genes (hMSH2, hMLH1, hPMS1, and hPMS2). Mutational analyses of these four MMR genes in endometrial carcinomas have not been previously reported. We analyzed nine sporadic MI-positive primary endometrial carcinomas for mutations in the above four MMR genes. Mutations were detected in two tumors (in hMSH2), and both of the mutations were acquired somatically. Immunohistochemical staining revealed a lack of expression of hMSH2 protein in the two tumors containing hMSH2 mutations. Our data suggest that mutations in these four known DNA MMR genes are not responsible for MI in the majority of sporadic endometrial carcinomas displaying this phenotype.

To study the pathways associated with genomic instability in cancer, we examined UV-induced and spontaneous mutagenesis in clonal cell lines expressing human papillomavirus (HPV) proteins, either high-risk (HPV16) E6 or E7 or low-risk (HPV11) E6, in comparison to the parental RKO cells, a colon carcinoma cell line expressing only normal p53. High-risk E6 and E7 bind and functionally inactivate tumor suppressor proteins p53 and Rb, respectively, and both disrupt the G1 arrest in response to DNA damage. Low-risk HPV E6 proteins bind p53 with much lower affinity than high-risk E6 and fail to mediate p53 degradation or to disrupt the G1 checkpoint. We found that cells expressing HPV16 E6 had reduced survival and increased mutagenesis at the hprt locus when treated with low doses of UV. However, this analysis was complicated by the unexpected observation of a very high background of spontaneous mutagenesis in the unirradiated cells expressing the HPV16 E6 gene. Fluctuation analysis revealed a 5-fold elevated mutation rate in the cells expressing HPV16 E6. HPV11 E6 conferred a 2-fold elevation in the mutation rate, but HPV16 E7 had no effect. The increased spontaneous mutagenesis, therefore, appeared to be mediated by p53 inactivation and to be independent of Rb (which acts downstream of p53 in the G1 arrest pathway following DNA damage). Taken together, these findings suggest that the effect of p53 inactivation on spontaneous mutagenesis is manifested at the level of DNA repair, recombination, or coupling of transcription with one of these processes instead of by an alteration in G1 arrest.

Hereditary non-polyposis colorectal carcinoma (HNPCC) syndrome is characterized by early onset and multiple cancers of predominantly the proximal colon and occasionally other organs. The mode of transmission is compatible with autosomal dominant inheritance but the location and characteristics of the putative susceptibility gene are unknown. We performed linkage analyses with the aim of proving or excluding the existence of a susceptibility locus on 18q. This hypothesis was based on the frequent involvement of the DCC gene in colorectal carcinoma and on the previously reported linkage between HNPCC and the Kidd blood group locus (JK) also on 18q. Seven HNPCC families were tested with eight polymorphisms, including three from within DCC. The DCC locus could be excluded as the HNPCC susceptibility locus in five families in which the two point logarithm-of-odds scores were -3.66, -3.63, -4.12, -7.90, and -3.74 at the recombination fraction of 0.00. In the remaining two families linkage could be neither excluded nor confirmed. The added pairwise logarithm-of-odds score for all seven families was -22.65 at the recombination fraction of 0.00. Multipoint analyses of linkage in the seven families suggested exclusion of some 60 cM in the region DCC-D18S18-D18S22-D18S7 as the site for HNPCC susceptibility locus. In addition to DCC, the excluded portion comprises JK.

The relationship between DNA topoisomerase II expression and mammalian cell proliferation has been evaluated by determining enzyme levels in normal and neoplastic rat prostate tissues. By activity assay and by immunoblot analysis using anti-topoisomerase II antiserum, topoisomerase II levels were found to be elevated in both the Dunning R3327-H and the Dunning R3327-G rat prostatic adenocarcinomas over levels assayed in the normal rat dorsal prostate. Immunohistochemical studies using the antitopoisomerase II antiserum revealed that a greater fraction of nuclei contained detectable levels of topoisomerase II in tissue sections prepared from each of the Dunning tumors than in rat dorsal prostate tissue sections. The Dunning R3327-H and R3327-G tumors grow at different rates in vivo (J. T. Isaacs and D. S. Coffey, Clin. Oncol., 2: 479-498, 1983). When measured topoisomerase II levels were compared to known growth parameters for each of the tissues studied, topoisomerase II expression was found to be correlated with tissue growth rate.