Terry Hurst

β-Catenin mutation and expression analysis in ovarian cancer: Exon 3 mutations and nuclear translocation in 16% of endometrioid tumours

The molecular mechanisms involved in the generation of epithelial ovarian cancers are poorly understood, but evidence suggests that the different histological subtypes may arise from independent tumorigenic events. β‐Catenin is emerging as an important oncogene in the transformation of a number of epithelial cancers, and mutations have been reported in a small study of endometrioid ovarian adenocarcinomas. Mutations in the NH2‐regulatory domain of β‐catenin stabilise the cytoplasmic levels of this protein, which promotes up‐regulation of the β‐catenin–T‐cell factor–lymphoid enhancer factor transcriptional complex. We report here β‐catenin (CTNNB1) exon 3 mutation analysis in 149 epithelial ovarian carcinomas. This revealed 10/63 (16%) endometrioid ovarian tumours with activating mutations of the β‐catenin gene. All mutations were missense changes within the GSK3β consensus site, affecting serine residues at codons 33 and 37 and glycine at codon 34. Immuno‐histochemical analysis identified cytoplasmic stabilisation and nuclear translocation in those endometrioid tumours with mutations. This phenotypic change was also identified in 3 other endometrioid tumours that did not have somatic mutations within exon 3 of CTNNB1. Stabilisation of the free, monomeric pool of β‐catenin and the probable resulting constitutive activation of its Tcf‐associated transcriptional complex appears to be a specific oncogenic event in endometrioid ovarian adenocarcinoma. Int. J. Cancer 82:625–629, 1999.

Increased expression of cyclin-dependent kinase inhibitor 2 (CDKN2A) gene product p16INK4A in ovarian cancer is associated with progression and unfavourable prognosis

Paraffin sections from 190 epithelial ovarian tumours, including 159 malignant and 31 benign epithelial tumours, were analysed immunohistochemically for expression of cyclin‐dependent kinase inhibitor 2 (CDKN2A) gene product p16INK4A (p16). Most benign tumours showed no p16 expression in the tumour cells, whereas only 11% of malignant cancers were p16 negative. A high proportion of p16‐positive tumour cells was associated with advanced stage and grade, and with poor prognosis in cancer patients. For FIGO stage 1 tumours, a high proportion of p16‐positive tumour cells was associated with poorer survival, suggesting that accumulation of p16 is an early event of ovarian tumorigenesis. In contrast to tumour cells, high expression of p16 in the surrounding stromal cells was not associated with the stage and grade, but was associated with longer survival. When all parameters were combined in multivariate analysis, high p16 expression in stromal cells was not an independent predictor for survival, indicating that low p16 expression in stromal cells is associated with other markers of tumour progression. High expression of p16 survival in the stromal cells of tumours from long‐term survivors suggests that tumour growth is limited to some extent by factors associated with p16 expression in the matrix. Int. J. Cancer 74:57–63.

Frequent loss of heterozygosity and three critical regions on the short arm of chromosome 8 in ovarian adenocarcinomas

Many chromosomal regions undergo loss of heterozygosity (LOH) in ovarian adenocarcinomas but few of the target regions have been finely mapped. One of the chromosome arms likely to harbour one or more tumour suppressor genes inactivated in ovarian cancer is the short arm of chromosome 8 which is frequently deleted in many other solid tumours. We have examined a large panel of microsatellite markers on 8p for LOH in 53 ovarian adenocarcinomas. LOH was observed in 27 tumours (51%), with a significant trend towards a higher frequency of LOH in more advanced tumours. Detailed examination of nine tumours with partial deletions defined three regions of overlap, two in 8p23 and one in 8p22, which suggests that there might be as many as three tumour or metastasis suppressor genes on 8p which are inactivated during ovarian tumorigenesis. LOH on 8p was significantly associated with 9p LOH which suggests that inactivation of target genes on these chromosomes may be cooperative events.

Analysis of loss of heterozygosity and KRAS2 mutations in ovarian neoplasms: Clinicopathological correlations

The molecular events that give rise to ovarian epithelial neoplasms are not well understood. In particular, it is not known whether adenocarcinomas arise from benign or low malignant potential (LMP) precursors. We have examined a large series of benign (25) and LMP (31) ovarian tumors for loss of heterozygosity (LOH) at multiple loci on 17 chromosomes. LOH was observed in benign tumors on chromosomes 6 (14%) and 9 (5%) and on the X chromosome (33%) only. LOH on these chromosomes was also detected in a small number of LMP neoplasms, suggesting that these may derive sometimes from benign precursors. In addition, we examined LOH in 93 adenocarcinomas. Analysis of associations between LOH events showed that LOH on chromosomes 5 and 17 (P = 0.0002) and on chromosomes 17 and 18 (P = 0.00007) were associated significantly with each other, which suggests that these may represent cooperative, progressive events. No novel significant associations were identified between LOH events and stage, grade, or histology, which would indicate the existence of genetic heterogeneity in ovarian neoplasms. KRAS2 mutations were detected more often in LMP neoplasms than in malignant tumors (P = 0.004) and were detected more often in Stage I/II malignant tumors than in Stage III/IV malignant tumors (P = 0.033), suggesting that LMP tumors with KRAS2 mutations are unlikely to progress to frank malignancy. Univariate (but not multivariate) survival analysis showed that LOH of chromosomes 11 (P = 0.039) and 17 (P = 0.04) was associated with a significantly worse prognosis. Replication of these novel findings is necessary, and the identification, isolation, and characterization of the critical genes affected by LOH will determine their importance in the pathogenesis of ovarian malignancies. Genes Chromosom. Cancer 18:75–83, 1997.

Rare mutations and no hypermethylation at the CDKN2A locus in epithelial ovarian tumours

The tumour‐suppressor gene CDKN2A (p16, MTS1, CDK4I) encodes a cell cycle‐regulatory protein and is located on chromosome 9p21, a region deleted in a wide variety of human cancers. To determine the role of the CDKN2A gene in the development of ovarian adenocarcinomas, we examined a large series of benign, low malignant potential (LMP) and invasive ovarian neoplasms for evidence of loss of heterozygosity (LOH), homozygous deletions, point mutations and hypermethylation of the CDKN2A locus. We have previously reported LOH on 9p in 45% of malignant ovarian neoplasms and a smaller percentage of benign and LMP tumours. In the current study, 6 malignant tumours were identified with partial deletions of 9p21. In 5 of these, the CDKN2A gene lays within the minimal deleted region. Homozygous deletions of CDKN2A were observed in only 2/88 invasive ovarian tumours and in 5/11 ovarian cancer cell lines. Of 15 primary ovarian tumours analyzed, one nonsense mutation was identified in a mucinous LMP tumour. No evidence of hypermethylation of the CDKN2A gene was found in 50 primary ovarian adenocarcinomas nor in 3 ovarian cancer cell lines. In conclusion, homozygous deletions, mutations and the de novo methylation of 5′ CpG island are not frequent modes of inactivation of the CDKN2A gene in ovarian cancer. The target of 9p LOH in ovarian adenocarcinomas is therefore unknown. Int. J. Cancer 70:508–511.

Serine protease inhibition and mitochondrial dysfunction associated with cisplatin resistance in human tumor cell lines: Targets for therapy

Indicators of mitochondrial function were studied in two different cell culture models of cis-diamminedichloroplatinum-II (CDDP) resistance: the intrinsically resistant human ovarian cancer cell line CI-80-13S, and resistant clones (HeLa-S1a and HeLa-S1b) generated by stable expression of the serine protease inhibitor-plasminogen activator inhibitor type-2 (PAI-2), in the human cervical cancer cell line HeLa. In both models, CDDP resistance was associated with sensitivity to killing by adriamycin, etoposide, auranofin, bis[1,2-bis(diphenylphosphino)ethane]gold(I) chloride ([Au(DPPE)2]Cl), CdCl2 and the mitochondrial inhibitors rhodamine-123 (Rh123), dequalinium chloride (DeCH), tetraphenylphosphonium (TPP), and ethidium bromide (EtBr) and with lower constitutive levels of ATP. Unlike the HeLa clones, CI-80-13S cells were additionally sensitive to chloramphenicol, 1-methyl-4-phenylpyridinium ion (MPP+), rotenone, thenoyltrifluoroacetone (TTFA), and antimycin A, and showed poor reduction of 1-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT), suggesting a deficiency in NADH dehydrogenase and/or succinate dehydrogenase activities. Total platinum uptake and DNA-bound platinum were slightly lower in CI-80-13S than in sensitive cells. The HeLa-S1a and HeLa-S1b clones, on the other hand, showed poor reduction of triphenyltetrazolium chloride (TTC), indicative of low cytochrome c oxidase activity. Total platinum uptake by HeLa-Sla was similar to HeLa, but DNA-bound platinum was much lower than for the parent cell line. The mitochondria of CI-80-13S and HeLa-S1a showed altered morphology and were fewer in number than those of JAM and HeLa. In both models, CDDP resistance was associated with less platinum accumulation and with mitochondrial and membrane defects, brought about one case with expression of a protease inhibitor which is implicated in tumor progression. Such markers may identify tumors suitable for treatment with gold phosphine complexes or other mitochondrial inhibitors.

Evaluation of two new assays for tumor-associated antigens, CASA and OSA, found in the serum of patients with epithelial ovarian carcinoma--comparison with CA125.

Two new assays have been developed to measure tumor-associated antigens designated ovarian serum antigen (OSA) and cancer-associated serum antigen (CASA). Both assays are dual epitope ELISAs using the same capture monoclonal antibody (BC2); the second antibodies in the OSA and CASA assays are OM-1 and BC3, respectively. Using arbitrary cutoffs of 2.5 and 3.0 units/ml, 82 and 76% of 80 serum samples from ovarian cancer patients were positive for OSA and CASA, respectively, compared with 5 and 2.5% of samples from a control population of 40 women. A strong correlation was found between the two assays (r = 0.80, P 35 U/ml), 82% for OSA and 73% for CASA. Of the 9 samples negative for CA125, 3 were positive for OSA and 3 were positive for CASA. Serum OSA, CASA, and CA125 levels were determined in serial samples from 20 ovarian carcinoma patients throughout the course of their treatment. Clinical course was accurately reflected by CA125 levels in 85% of patients, by CASA in 65%, and by OSA in 75%. In 4 patients, a rise in CASA levels and, in 2 patients, a rise in OSA levels significantly predated rising CA125 levels to predict recurrence. Six of 7 serum samples obtained prior to positive second-look laparotomy were negative for CA125, while 4 were positive for OSA and 6 were positive for CASA. These results indicate that the OSA and CASA assays could be superior to CA125 for detection of small volume occult ovarian carcinoma.

Effect of Interferon-γ and TNF-α on MUC1 Mucin Expression in Ovarian Carcinoma Cell Lines

In view of the potential uses of cell surface tumour associated antigens in novel anticancer treatment, a study was designed to investigate whether the biological response modifiers interferon-gamma (IFN-gamma) and tumour necrosis factor-alpha (TNF-alpha) could effect the expression of an epitope on the tumour associated MUC1 epithelial mucin. Four ovarian carcinoma cell lines showing high (OAW42 and GG) and low (JAM and PE01) basal expression of MUC1 were treated with 10-1000 U/mL of IFN-gamma or TNF-alpha for one or five days. Changes in MUC1 expression in cells exposed to IFN-gamma or TNF-alpha were monitored using an ELISA technique with the monoclonal antibody BC2 which reacts with a core protein epitope on the MUC1 mucin, and then corrected for the number of viable cells present. TNF-alpha had little effect on MUC1 expression, but one or five days exposure to IFN-gamma significantly increased MUC1 expression (p < 0.01) in all cell lines including the two cell lines that initially showed little or no expression.

Relationships between resistance to cross-linking agents and glutathione metabolism, aldehyde dehydrogenase isozymes and adenovirus replication in human tumour cell lines

In a panel of 10 human tumour cell lines with no prior exposure to drugs in vitro, resistance to cisplatin correlated with resistance to the nitrogen mustard derivatives Asta Z-7557 (mafosfamide, an activated form of cyclophosphamide), melphalan and chlorambucil. Simultaneous treatment with DL-buthionine-S,R-sulfoximine did not enhance the toxicity of cisplatin or Asta Z-7557, and no correlation was found between drug resistance and cellular levels of metallothioneins (as judged by sensitivity to cadmium chloride), glutathione (GSH), GSH reductase, GSH transferase, or gamma-glutamyltranspeptidase. The two cell lines most resistant to Asta Z-7557 expressed aldehyde dehydrogenase cytosolic isozyme 1, found also in normal ovary, but not isozyme 3. Treatment of resistant cells with cisplatin or Asta Z-7557 inhibited cellular DNA synthesis and replication of adenovirus 5 to a lesser extent than in sensitive cells. The virus could be directly inactivated by both drugs prior to infection, subsequent replication being inhibited to the same extent in sensitive and resistant cells. In contrast to Asta Z-7557 and other DNA damaging agents, cisplatin was much more toxic to adenovirus (D37 0.022-0.048 microM) than to cells (D37 0.25-2.5 microM). The adenovirus 5 mutant Ad5ts125 having a G----A substitution was even more sensitive to cisplatin (D37 7-8 nM) than wild type virus and another mutant. Cisplatin was detoxified less by sonicated resistant resistant cells than sensitive cells, as judged by inactivation of Ad5ts125 added to the reaction mixture. It can be inferred that (i) the major differences in cellular resistance to cisplatin and Asta Z-7557 in the present material did not involve enhanced DNA repair or protection by metallothioneins or GSH, but were associated with the ability to continue cellular and viral DNA synthesis during treatment, (ii) resistance was not associated with less template damage, and (iii) the adenovirus genome may be a suitable probe for predicting tumour resistance to cisplatin and for elucidating the DNA sequence dependence of cisplatin toxicity.

Predictive value of serial CA 125 antigen levels in ovarian cancer evaluated by second-look laparotomy

Serial serum CA 125 levels were measured before definitive surgery and during chemotherapy for 12 months or more in 64 patients with ovarian cancer. In the 42 patients who had a complete clinical remission and thus were subjected to a second-look laparotomy, an absence of disease was not predicted by patterns of CA 125 levels. Whilst rising or persistently high levels indicated the presence of tumour in 92% of patients, declining levels to negative predicted the absence of tumour in only 50%. Although the majority of these patients showed microscopic foci or a tumour mass less than 1 cm, 3 patients had a larger amount of disease. In the follow-up of 49 patients, the accuracy of prediction of a good outcome was better than that of a poor outcome on the basis of CA 125 patterns, with rates of 92% and 79%, respectively. Our findings indicate that CA 125 lacks sensitivity in detecting small tumour masses (less than 1 cm dia.) but rising or persistently high levels suggest a strong likelihood of a residual tumour to be found at a second-look laparotomy.