Alterations in cell cycle regulating proteins are common in many cancer types. Recent data suggest a possible link between CDC25 phosphatases overexpression and malignancy. Our objective was to evaluate the expression of the three cell cycle-related phosphatases CDC25A, CDC25B and CDC25C in patients with ovarian cancer. All the patients had a minimal follow up of three years. CDC25A, CDC25B and CDC25C expression were investigated by immunohistochemistry in 106 patients with ovarian cancer. CDC25A and CDC25B were found expressed in almost all the samples analyzed, while CDC25C was undetectable in more than 80% of the patients. The low evaluable data on CDC25C expression, did not allow any association between the expression of this phosphatase and prognosis. The expression of CDC25A and CDC25B showed some evidences of association with a poor prognosis (p = 0.034 and p = 0.058 respectively). This relationship was independent of other factors such as tumor grade, histotype, stage and residual tumor after surgery. In the same patients the examined parameters (residual tumor, grade, stage and histotype) did show the expected relation with survival. The results indicate that high CDC25A and CDC25B expression is related to a worse prognosis in ovarian cancer patients. CDC25 phosphatases expression can be regarded as a possible prognostic factor for ovarian cancer and these proteins should be evaluated as potential molecular targets of novel drugs against this human neoplasm.

The Fas/CD95 gene encodes a transmembrane protein with an extracytoplasmic domain, responsible for ligand binding, and a small cytoplasmic domain comprising the death domain, responsible for the association of signal-transducing proteins, such as the Fas-associated death domain (FADD) protein, leading to the recruitment and activation of caspases, the apoptosis effectors (Ashkenazi and Dixit, 1998). During menstrual cycles, the ovarian surface epithelium undergoes a tightly regulated process of differentiation and apoptosis of its cells (Kacinski and Sapi, 1999). The vast majority of ovarian cancer cell lines arise from ovarian surface epithelial cells that might have undergone some mistake in these processes. In vitro data suggest that the programmed death of ovarian epithelial cells is Fas-dependent (Quirk et al., 1997), and the chromosome 10q24 locus, containing the Fas gene is recurrently deleted in ovarian cancer (Mitelman et al., 1997; Mertens et al., 1997). Moreover, somatic mutations of the CD95 gene, causing the loss of protein function, have been reported in human tumors (Tamiya et al., 1998; Landowski et al., 1997; Gronbaek et al., 1998; Shin et al., 1999; Lee et al., 1999a,b). With the hypothesis that CD95 somatic mutations might contribute to the pathogenesis of epithelial ovarian cancer, we performed mutational analysis of its full coding region, comprising exon–intron junctions, in a panel of 8 ovarian cancer cell lines: OVCAR-3, OVCAR-5, OVCAR-8, SKOV-3, IGROV-1, A2780, SW626 and PA-1. The Fas/CD95 coding region was amplified by PCR using the primers previously published (exon 1, Shin et al., 1999; exons 2–9 Gronbaek et al., 1998) from genomic DNA. The CEM cell line and a known mutated lymphoma sample were amplified as positive controls for exons 7 and 9.1, respectively. PCR products were analyzed by single-strand conformation polymorphism (SSCP) on a 0.75 × MDE gel (FMC, Rockland, ME), run overnight at room temperature and visualized after silver staining. We were not able to detect somatic mutation in any of the 8 ovarian cancer cell lines. Indeed, anti-Fas antibodies can induce apoptosis and increase the chemosensitivity to cytotoxic drugs in some Fas-positive ovarian cancer cell lines (Ghahremani et al., 1998; Uslu et al., 1996). However, the SKOV-3 cell line is resistant to apoptosis induced by anti-Fas antibodies, despite Fas expression (Wakahara et al., 1997). Since Skov-3 appears to be devoid of genomic mutations within the Fas gene, other molecules involved in the apoptosis cascade subsequent to Fas activation should be altered in ovarian cancer. We thank our colleagues Drs. G. Gaidano (Novara, Italy), for kindly providing the 2 positive controls, and S. Luminari (Milan, Italy), for helpful discussion. Francesco Bertoni frbertoni@tin.it, Annarita Conconi, Stefania Carobbio, Claudio Realini, Anna Maria Codegoni, Emanuele Zucca, Franco Cavalli

The loss of mismatch repair enzymes increases the mutation rate in microsatellites and coding regions of the genome and appears to be involved in drug resistance. The replication error (RER+) phenotype, associated with microsatellite instability, has been widely described for both familial and sporadic colon cancers and for gastric and endometrial tumors. For ovarian cancer, the incidence of RER+ cases among sporadic tumors is still uncertain. We analyzed epithelial ovarian tumors and ovarian carcinoma cell lines for microsatellite instability and for mutations in the coding regions of different genes, including the recently discovered human CHK-1 gene, which has an important role in controlling cell cycle progression and whose coding region contains a poly(A)9 tract. Microsatellite instability and frameshift mutations in coding regions of BAX, TGFbetaRII, IGFIIR, E2F-4, ICE, and CHK-1 genes were analyzed in ovarian cancer samples and cell lines by polymerase chain reaction (PCR). Approximately 26% of patients showed microsatellite instability in two or more loci. BAT-26 locus showed no alteration in primary tumors. We detected a BAX mutation in one tumor sample and a TGFbetaRII mutation in one cell line. Our findings confirm the presence of the RER+ phenotype in sporadic ovarian cancer. The low rate of mutation in genes previously reported to be altered in colon and gastric cancer suggests that other not yet identified genes might be altered and could play a role in tumor progression and response to treatment in RER+ ovarian tumors.

BACKGROUND:The p73 gene is structurally related to the tumor suppressor gene p53. The role of p73 in tumor development is still unclear and no data on ovarian cancer are so far available. For this reason we have analyzed, in a panel of ovarian cancers, the allelic distribution and expression of p73.PATIENTS AND METHODS:Fifty-one samples from ovarian cancers and five human ovarian cancer cell lines growing in culture were analyzed. Allelic origin was analyzed by PCR after digestion with the restriction enzyme Sty I. Heterozygous, informative cases were selected for studies aimed at evaluating allelic expression of p73.RESULTS:We found an allelic distribution similar to that previously reported. LOH was found in two patients with ovarian cancer. In one case in which normal ovarian tissue was available biallelic expression of p73 was found.CONCLUSION:In comparisons of ovarian cancers and borderline tumors, no differences in allelic distribution and/or expression were found, suggesting that p73 does not play an important role in the pathogenesis and development of ovarian cancer.

Background. The levels and activity of topoisomerase I were determined in 35 biopsies from patients with ovarian cancer.Patients and methods: The activity was defined by the ability to relax supercoiled DNA plasmid, and levels were determined by Western blotting and immunohistochemistry.Results. We detected topoisomerase I activity in all samples, although at different levels. Enzymatic activity was the same in fresh and frozen tissues.Western blotting analysis detected topoisomerase I in 29 of 35 tumor samples but in the remaining six the levels were below the detection limit. We analyzed the distribution of topoisomerase I in tumor cells and normal infiltrating cells by immunohistochemistry. The enzyme was mainly associated with tumor cells although there were four samples in which tumor cells were negative but ill which normal cells, mainly lymphocytes, yielded a positive result.Conclusions: The results indicate that topoisomerase I enzymatic activity is detectable in human ovarian tumors, varying among patients. No correlations were found between the levels of the enzyme and its activity Because of the high heterogeneity observed, enzymatic activity assays should be combined with immunohistochemical evaluation of Topo I in the tumor and normal cells present in the tissue.

In a panel of 16 human ovarian tumours transplanted in nude mice, the expression of genes involved in cell cycle regulation and in response to drug treatment were characterised. In the 16 tumours analysed we could not detect overexpression of Erb-B2 oncogene while expression of MDR1 mRNA was not detected in 11/15 samples and was low in 4/15 tumours. Only three tumours had mutations in the p53 gene exons 5-8 and one of these mutations did not result in any amino acid alteration. The levels of mRNA for cyclins A, D1 and E were heterogeneous with some tumours expressing high levels and others not expressing them at all. The same was found for the cyclin dependent kinases (CDK) CDK2 and CDK4 and for CDK inhibitors p21/WAF1, p27/KIP1 and p16/CDKN2. Two genes belonging to the nucleotide excision repair, ERCC1 and ERCC3 were detectable in all the samples examined, as were the genes MGMT and MAG, also involved in DNA repair. The data indicate a heterogeneity in the expression of genes considered to be involved in the cellular responses to cytotoxic drug treatment and indicate the possibility of using these tumour models to test specifically molecules with a defined mechanism of action.

The expression of mismatch repair proteins hMSH2 and hMLH1 was investigated in human ovarian cancer cell lines and in biopsies of ovarian carcinomas obtained from 20 patients undergoing surgical operation. By Western blotting analysis hMSH2 protein was detected in all the tumor samples analyzed and in eight out of nine human ovarian cancer cell lines, while hMLH1 was undetectable in four out of 20 ovarian tumors and in five out of nine human ovarian cancer cell lines analyzed. The possible presence of frameshift mutations in the BAX gene, which contains a sequence of eight contiguous guanines in its third exon, was tested in all the samples. All the cell lines presented the normal alleles for the BAX gene while only in one of the tumor samples a heterozygous frameshift mutation was found. The frameshift mutation was associated to a low, almost undetectable, level of BAX protein which was instead present at much higher levels in all the other samples investigated. The results indicate that frameshift mutations in the BAX gene, possibly arising as a consequence of microsatellite instability (detectable in these tumors), is detectable in human ovarian cancer although quantitatively it does not appear to be a major determinant of the low apoptotic response to chemotherapy observed in ovarian cancer cells.

The expression of different genes potentially involved in DNA repair and in cell responses to chemotherapy was evaluated in 33 previously untreated ovarian cancer patients. In biopsies of the same patients the expression of repair genesO6-methylguanine DNA methyltransferase (MGMT), 3-methyladenine DNA glycosylase (MAG), ERCC1, MDR-1, DNA topoisomerase I, DNA topoisomerase IIα, and glutathioneS-transferase-π (GST-π) was assessed by Northern blot analysis. No direct statistical correlation was found between the expression of these genes and the response to chemotherapy (mainly platinum-based with or without doxorubicin and cyclophosphamide). Univariate analysis showed a weak negative correlation (P= 0.037) between the expression of ERCC1 and mortality, whereas no statistically significant correlation was found for other parameters. The MDR-1 gene encoding for the P-glycoprotein P-170 was mostly undetectable in these patients (as assessed by Northern blotting), whereas relatively high levels of MAG and MGMT were found in the majority of patients. A statistically significant correlation was found between the expression of DNA topoisomerase I and the expression of either ERCC1 (P= 0.0026) or GST-π (P= 0.0279).

Glutathione (GSH), glutathione S-transferase (GST) activity and GSTπ expression were measured in 10 human bladder tumors and adjacent uninvolved specimens from Egyptian patients with a history of schistosomal infection. GSH was higher in the tumor than in surrounding uninvolved tissue (not significant). Total GST activity per mg tissue protein and GSTπ expression were higher in tumor tissues (P<0.05) than in uninvolved tissues. There was a positive correlation between GST activity and GSH content and between total GST activity and GSTπ expression in both tumor and uninvolved tissues.

The cyclin-dependent kinase inhibitor p16 gene (P16, MTS1, CDKN2) has been shown to be altered by deletion or point mutation in some human tumours and cancer cell lines, suggesting that it works as a tumour suppressor. We analysed p16 gene mutation and p16 protein expression in 42 primary ovarian carcinomas and in five human ovarian cancer cell lines. Polymerase chain reaction (PCR) amplifications of exons 1 and 2 of the gene showed no deletion or gross rearrangement in the p16 gene. The lack of deletion was further demonstrated by Southern blot analysis. Looking for point mutations, we used single-strand confirmation polymorphism (SSCP) analysis and, in half of the tumours, we sequenced both strands of exons 1 and 2. No mutations were detected. In 11 out of 42 patients (26%), however, we detected no protein expression by Western blot analysis, suggesting that decreased expression of p16 rather than deletion of the gene can occur in a significant percentage of human ovarian cancers. In the same experiment CDK4 protein was found homogeneously expressed in all the tumour specimens and in the five cell lines. The lack of expression of p16 was not due to hypermethylation of the gene assessed by digestion of genomic DNAs with a methylation sensitive enzyme, suggesting that other mechanisms, not yet identified, are involved in the decreased expression of the p16 gene in human ovarian tumours.

BACKGROUND. The induction of estrogen and progesterone receptors (ER and PGR) has been reported in bl cast and endometrial cancer cells exposed to human fibroblast interferon-beta (hIFN-beta). Clinical verification of this finding might provide the rationale for new therapeutic approaches. This study was designed to evaluate whether clinical treatment with high doses of hIFN-beta induced ER and PGR in patients with endometrial adenocarcinoma.METHODS, Two biopsies were obtained, 1 before and 1 after hIFN-beta treatment (3 x 10(6) i.m. every other day For 3 weeks) from 36 patients with endometrial adenocarcinoma. ER and PGR were determined with standard procedures using radiolabeled ligands,RESULTS, hIFN-beta treatment did not affect the proportion of ER-posit ive (i.e., >15 fmol/mg protein) or PGR-positive (i.e., >20 fmol/mg protein) cases. However, in patients with detectable ER and PGR at baseline, hIFN-beta raised the levels. Using a 35% difference before and after therapy as a cut-off, 72 and 79% of cases had increases in ER and PGR, respectively. The difference was highly significant for PGR.CONCLUSIONS, In patients with endometrial adenocarcinoma with undetectable ER or PGR, hIFN-beta did not induce the expression of these receptors, When the receptors were present they were upregulated by hIFN-beta. Whether this increase in receptor levels, particularly PGR, has therapeutic applications remains to be established. (C) 1996 American Cancer Society.

We selected two clones, isolated from the human colocarcinoma cell line LoVo, showing a sensitivity to doxorubicin similar to (LoVo clone 5) or three times lower than (LoVo clone 7) the parental cell line. Since vimentin was atypically expressed in a human breast carcinoma cell line made resistant to doxorubicin, we looked at vimentin expression in these two clones with spontaneously different sensitivity to the drug. For comparison we used the parental cell line LoVo WT and LoVo/DX made resistant pharmacologically. mRNA for vimentin was undetectable by Northern blot analysis in LoVo WT and in LoVo clone 5, while expression of this gene was high in LoVo clone 7 and in LoVo/DX. This increase in mRNA levels was not related to an amplification of DNA, as suggested by Southern blot analysis. Immunofluorescence and immunocytochemistry findings confirmed, at protein level, the mRNA data. In LoVo clones 5 and 7, there were respectively 8.6% and 71% vimentin-positive cells, although the two clones showed similar expression of multidrug resistance gene 1 (mdr-1) and accumulated intracellular doxorubicin at similar levels. Similarly, drug efflux was the same for both clones. Our results show for the first time that cells resistant to doxorubicin express vimentin independently of the mdr glycoprotein. However when cells from clone 5 were transfected with human vimentin cDNA, they did not become resistant, indicating that vimentin can be considered as a marker of resistance in these cells but does not give rise to a resistant phenotype by itself.

1 The effects of acute and repeated equiactive anorectic doses (ED50) of recently marketed 5-hydroxytryptamine (5-HT) uptake inhibitors on the content of brain indoles were compared in rats in relation to the brain regional concentrations of unchanged drug and its known active metabolite.2 Single intraperitoneal (i.p.) doses of the anorectic ED50 of fluoxetine (35 mumol kg-1), fluvoxamine (60 mumol kg-1), paroxetine (20 mumol kg-1) and sertraline (49 mumol kg-1) slightly reduced brain 5-hydroxyindoleacetic acid (5-HIAA), with regional differences, this being compatible with 5-HT uptake blockade. Only fluvoxamine and sertraline significantly enhanced the content of 5-HT in the cortex.3 The regional sensitivity to the acute effect of a given drug was not related to any preferential drug distribution, as these compounds distributed almost uniformly in the brain areas considered (cortex, striatum and hippocampus).4 Repeating the same doses twice daily, i.p. for 14 days, however gave a different picture, fluvoxamine having little or no effect on the content of indoles and fluoxetine, paroxetine and sertraline lowering both 5-HT and 5-HIAA in all the brain regions compared to pair-fed control animals, 1 h after the last dose.5 One week later only fluoxetine-treated animals still had reduced brain 5-HT, this probably being related to the accumulation of its main metabolite norfluoxetine in rat brain after chronic dosing.6 Further studies on the relationship between the long-term neurochemical changes and anorectic activity are required but it appears from these results that anorectic drugs with similar acute effects on 5-HT uptake may differ in their long-term effects on 5-HT mechanisms.

Some agents that increase serotoninergic transmission in the brain show anorectic activity at doses that do not interfere with the behavior of rats and other animal species. These agents reduce food intake by a mechanism that clearly differs from that involved in the anorectic activity of d-amphetamine. d-Fenfluramine, fluoxetine, and sertraline are three drugs that have already been tested and are used in man. These compounds accumulate in the brain and are metabolized through N-dealkylation. They affect the uptake and release of serotonin at different concentrations, with mechanisms that do not completely overlap. There is pharmacological evidence that d-fenfluramine and sertraline exert their anorectic activity by enhancing the stimulation of 5-HT1nonA receptors whereas fluoxetine seems to affect at anorectic doses both serotoninergic and dopaminergic systems. The role of serotonin in controlling food intake will be discussed, and the effects of agents that reduce serotoninergic transmission will also be considered.

The present study aimed to establish the role of the metabolite norfluoxetine in the anorectic activity of fluoxetine, and to relate the anorectic doses (ED50) to the brain concentrations of the parent drug and its metabolite. Fluoxetine showed anorectic activity at increasing intraperitoneal doses (ED50=39.1, 34.7 and 21.7-mu-mol kg-1 in mouse. rat and guinea-pig, respectively) and norfluoxetine was slightly more active (24.3, 22.9 and 19.1-mu-mol kg-1, respectively) in all three species. In terms of maximum concentration (C(max)) and area under the curve (AUC) within the experimental period (0-90 min), brain concentrations varied widely and were poorly related to the dose; guinea-pig appeared to be much more sensitive to fluoxetine than was mouse or rat. Administered norfluoxetine was present in the brain of the three species in approximately the same order as fluoxetine, i.e. lower in guinea-pig than in mouse or rat. The C(max) and AUC of norfluoxetine after fluoxetine administration was 50-60% of the values after an equiactive dose of norfluoxetine in mouse and guinea-pig, and more than 80% in rat.

Summary: The effects of valproate (VPA) on carnitine and lipid metabolism and on liver function were assessed in 213 age‐ and sex‐matched outpatients from five centers, with the following distribution: VPA monotherapy, 54; VPA polytherapy, 55; other monotherapies, 51; and untreated, 53. Mean total and free carnitine levels were significantly lower in patients with polytherapy; acylcar‐nitine was significantly higher for VPA monotherapy and the ratio of acyl‐ to free carnitine was significantly higher in all patients receiving VPA. Ammonia, uric acid, and bilirubin were the only tests selectively impaired with VPA. A significant correlation was found between serum ammonia and VPA dosage. Glucose, B‐lipoproteins, try‐glicerides, acetacetate, and p‐hydroxybutyrate were unchanged in the four groups. Sex and age appeared to interact with total and free carnitine values. Adverse drug reactions were apparently unrelated to carnitine metabolism impairment. Only a few patients had abnormal carnitine values. Our data support the assumption that carnitine deficiency and abnormal liver function due to VPA are mostly subclinical events.RÉSUMÉLes effets du valproate (VPA) sur le métabolisme de la carnitine, le métabolisme lipidique et le fonctionnement hépatique ont étéévalués chez 213 patients ambulatoires appariés pour l'âge et pour le sexe et provenant de 5 centres différents, répartis dans les groupes suivants: monothérapie par VPA, 54; polythérapie incluant le VPA, 55; autres monothérapies, 51; non traités, 53. Les taux moyens de carnitine totale et libre ont été significativement dimimiés chez les patients sous polythérapie. L'acyl‐carnitine était significativement plus élevée sous monothérapie par VPA, et le rapport acyl‐carnitine/carnitine libre était plus élevé chez les patients recevant du VPA. L'amnoniémie, l'uricémie et la bilirubinémie ont été les seuls paramètres modifyés sélectivement par le VPA. Une corrélation significative a été constatée entre l'amnoniémie et le taux sanguin de VPA. II n'y a pas eu de différences entre les 4 groupes en ce qui concerne le glucose, les beta‐lipoprotéines, les tryglicérides, l'acdtaté et le béta‐hydroxybutyrate. Le sexe et l'âge ont influenceé les taux de carnitine totale et libre. La survenue d'effets secondares des médicaments n'était apparemment pas liée aux modifications du métabolisme de la carnitine. Seuls quelques patients présentaient des valeurs anormales de carnitine. Ces résultats semblent con‐firmer la notion selon laquelle un déficit en carnitine et des perturbations du métabolisme hépatique liées au VPA sont le plus souvent des évènements infra‐cliniques.RESUMENEn 213 enfermos, de edades y sexo comparables, tratados en cinco centros distintos se nan medido los efectos del valproato (VPA) sobre la camitina, el metabolismo lipídico y las funciones hepáticas. Se ha seguido la siguiente distribución: Monoterapia con VPA, 54 enfermos, politerapia con VPA, 55; otras monote‐rapias, 51; no tratados, 53. Los valores medios totales de car‐nitina libre fueron significativamente más bajos en pacientes con politerapia; la acil‐carnitina estaba significativamente más ele‐vada en los casos con monoterapia VPA y más elevada que la relación entre la acil y la camitina libre en todos los casos que recibían VPA. El amoniaco, el acido urico y la bilirrubina fueron los unicos parametros que estaban selectivamente alterados con el VPA. Se encontró una correlation significativa entre el amoniaco sérico y las dosis de VPA. La glucosa, las beta‐Iipoproteinas, los triglicéridos, el acetacetato y el beta‐hidroxibutirato no se modificaron en los cuatro grupos. El sexo y la edad parecen relacionarse con los valores de camitina total y libre. Las reacciones adversas a la mediación no parecían relacionarse con las alteraciones del metabolismo de la camitina. Sólo en pocos enfermos se encontraron valores anormales de camitina. Nuestra información apoya la presunción de que el déficit de camitina y una función hepática anormal secundaria al VPA son, en la mayoría de los casos, acontecimientos subcliini‐cos.