Isoforms of endothelin-converting enzyme-1 (ECE-1) have opposing effects on prostate cancer cell invasion.

Cross-talk between tumour and stromal cells can profoundly influence cancer cell invasion by increasing the availability of mitogenic peptides such as endothelin-1 (ET-1). Endothelin-1 is elevated in men with metastatic prostate cancer (PC), and can exert both an autocrine (epithelial) and a paracrine (stromal) influence on growth. Endothelin-1 is generated from its inactive precursor big-ET-1 by endothelin-converting enzyme 1 (ECE-1). We and others have demonstrated that ECE-1 expression is significantly elevated in tumours and surrounding stromal tissue. Our current data show siRNA-mediated knockdown of stromal ECE-1 reduces epithelial (PC-3) cell invasion in coculture. Interestingly, readdition of ET-1 only partially recovers this effect suggesting a novel role for ECE-1 independent of ET-1 activation. Parallel knockdown of ECE-1 in both stromal and epithelial compartments results in an additive decrease in cell invasion. We extrapolated this observation to the four recognised isoforms ECE-1a, ECE-1b, ECE-1c and ECE-1d. Only ECE-1a and ECE-1c were significant but with reciprocal effects on cell invasion. Transient ECE-1c overexpression increased PC-3 invasiveness through matrigel, whereas transient ECE-1a expression suppressed invasion. Furthermore, transient ECE-1a expression in stromal cells strongly counteracts the effect of transient ECE-1c expression in PC-3 cells. The ECE-1 isoforms may, therefore, be relevant targets for antiinvasive therapy in prostate and other cancers.

Stromal-epithelial interactions influence prostate cancer cell invasion by altering the balance of metallopeptidase expression.

Perturbations of stromal-epithelial interactions in the developing tumour can contribute to cancer invasion and metastasis. The structurally related metallopeptidases endothelin-converting enzyme (ECE) and neutral endopeptidase (NEP) contribute sequentially to the synthesis and inactivation of ET-1, a mitogenic peptide that has been shown to affect tumour behaviour. This study has investigated the interaction between metastatic tumour epithelial cells, which lack NEP, and stromal cells, which we have shown to express ECE-1 (stromal-epithelial interactions), using Matrigel invasion chambers. The epithelial cell lines utilised in this study include androgen-sensitive LNCaP, androgen-independent PC-3, Du145 and recently established PNT-1a, PNT2-C2 and P4E6 prostate cell lines. Specific inhibition of endogenous ECE-1 activity in stromal cells reduced PC-3 and Du145 invasion by 70 and 50%, respectively. Addition of recombinant NEP to inactivate endogenous mitogenic peptides resulted in 50 and 20% reductions in invasion in PC-3 and Du145 cells, respectively. Neutral endopeptidase effects were reversed in the presence of thiorphan, a specific NEP inhibitor. Supplementation of defined media with bradykinin and ET-1 significantly increased PC-3 invasion by 40 and 50%, respectively. Du145 cell invasion increased by approximately 100% on adding ET-1. These studies implicate the metallopeptidases NEP and ECE-1 as mediators of prostate cancer invasion via a stromal/epithelial interaction.

Methylation of the neutral endopeptidase gene promoter in human prostate cancers.

Neutral endopeptidase 24.11 (NEP) is a cell surface peptidase expressed by prostatic epithelial cells that cleaves and inactivates neuropeptide growth factors implicated in the growth of androgen-independent prostate cancer (PC). Decreased NEP expression in hormone-refractory metastatic PCs can result from hormonal therapies because NEP transcription is induced by androgens and down-regulated by androgen withdrawal. NEP is encoded by a gene that contains a 5' CpG island spanning a transcriptional regulatory region. In this study, we investigate whether DNA hypermethylation of the NEP promoter accompanies decreased NEP expression in PC cell lines and whether it occurs in human PC tissues in vivo. DNA isolated from PC cell lines and from normal and neoplastic human prostate tissues was restriction-digested with a methylation-sensitive restriction endonuclease and analyzed by Southern blot using a 5' sequence-specific NEP probe. Methylation-specific PCR was performed using PCR primers designed to discriminate between methylated and unmethylated alleles, and reverse transcription-PCR using NEP-specific primers was performed on cDNA extracted from PC cells treated with 5-aza-2'-deoxycytidine. Methylation of the NEP promoter was present in androgen-independent PC cell lines but not in androgen-dependent or small-cell derived PC cell lines and in 3 of 21 (14%) primary PCs from patients with androgen-dependent disease. Exposure of PC cells to the demethylating agent 5-aza-2'-deoxycytidine led to an increase in NEP transcripts in DU-145 and PC-3 cells. These data show that hypermethylation of the 5' CpG NEP island is associated with a loss of NEP expression in PC. Loss of NEP expression via hypermethylation of the NEP promoter may contribute to the development of neuropeptide-stimulated PCs.

Analysis of the insertion/deletion polymorphism of the human angiotensin converting enzyme (ACE) gene in patients with renal cancer.

The angiotensin I-converting enzyme (ACE) contains an insertion/deletion (I/D) polymorphism, with the DD genotype associated with benign renal diseases. The distribution frequencies of the D and I alleles, and the DD, DI and II genotypes were determined in DNA extracted from kidney tissues of 58 renal cancer patients. The observed frequencies in patients who develop renal cancer was not significantly different than the normal population.

Homologous recombination in variants of the B16 murine melanoma with reference to their metastatic potential.

Genomic instability has been accepted as providing a phenotypic variety of malignant cells within a developing tumour. Defects in genetic recombination can often lead to phenotypic differences; therefore, it is possible that metastatic variant cell lines exhibit their particular phenotype as a result of an altered ability to catalyse homologous recombination. We have investigated recombination efficiency in B16 melanoma metastatic variants, using a plasmid, pDR, as a recombination substrate. The plasmid contains two truncated, nontandem but overlapping segments of the neomycin resistance gene (neo 1 and neo 2), separated by the functional gpt gene unit. Only a successful recombination of the two neo segments will generate a functionally intact neomycin gene. Extrachromosomal recombination here was a transient measure of the cells to recombine the neo fragments in an intra- or intermolecular manner. Extrachromosomal recombination frequencies were higher in the high metastasis variants (BL6, ML8) compared with the low metastatic F1 cells. On the other hand, the frequency of chromosomal recombination (after plasmid integration) was higher for the low metastasis (F1) cell line compared with the highly metastatic variants, BL6 and ML8. Since the recombination assay measures only successful recombination events, we have interpreted the observed higher incidence of chromosomal recombination in the low metastatic variant line as indicative of a more stable genome. Similarly, a higher inherent instability in the genome of the high metastasis variants would render these less efficient at producing and maintaining successful recombination events, and this was found to be true by Southern analysis. The results presented show that frequency of recombination may be adduced as evidence for implicating genomic instability in the generation of variant cell populations during metastatic spread. Such an interpretation is also compatible with the Nowell hypothesis for tumour progression.

Induction of 18A2/mts1 gene expression and its effects on metastasis and cell cycle control.

The metastasis associated 18A2/mtsI gene was inserted into the mammalian expression vector pMAMneo placing it under the control of the dexamethasone-inducible MMTV promoter. The construct was transfected into dexamethasone receptor negative F1 and receptor positive F10 cells of the B16 murine melanoma. The transferred gene was switched on in two transfectant clones of F10, by exposure to 10(-6) M dexamethasone, but not in clones of the receptor negative F1 line. One of the F10 transfectant clones (F10-192/10) was characterized further. A 13.5-fold increase in 18A2/mts1 transcripts was found in this clone upon exposure to dexamethasone. There was also a seven-fold increase in lung colonization in an experimental metastasis assay, together with increased expression of depolymerized tubulin and enhanced detection of p53 protein. The number of cells in the S phase increased by 2.5-fold following dexamethasone treatment of the clone. These data suggest a direct involvement of the 18A2/mts1 gene in lung colonization by the tumor cells. The 18A2/mts1 protein promotes tubulin depolymerization, sequesters the p53 phosphoprotein, and induces the cells to enter the S phase, but the relevance of these in the metastatic process remains to be elucidated.

DNA repair and repair fidelity in metastatic variants of the B16 murine melanoma.

We have examined the concept of genomic instability in relation to the metastatic progression of low (F1) and high metastasis (BL6, ML8) clones of the B16 mouse melanoma, by using a mutation assay, and DNA strand break repair and repair fidelity assays. The frequency of induced ouabain resistant colonies between the variant cell lines was consistent with the difference between their metastatic properties. Survival data for X-irradiation and bleomycin were similar among the 3 cell lines. When X-rays or bleomycin were used to induce strand breakage, no difference was detectable in either the rate or extent of DNA repair using the techniques of alkaline unwinding and alkaline elution for total strand breaks, and neutral elution for double strand breaks. DNA repair fidelity was measured using the PMH16 plasmid. A Kpn I restriction site was used to introduce a break within the gpt gene of the plasmid, prior to transfection. We found that approximately 100% and approximately 65% of the highly metastatic ML8 and Bl6 clones, respectively, religated the gene with the required fidelity, compared with only approximately 25% of the low metastasis F1 clones. In summary, the metastatic variants show similar sensitivities to X-irradiation and bleomycin, but a differential response to EMS. This difference is not reflected in any subsequent DNA strand break religation, but the variants do differ in their fidelity of repair. However, although the fidelity of DNA religation is related to metastatic potential, it is not consistent with the mutation frequency data.

Genomic instability and metastatic progression.

Recently, there has been a growing body of opinion that tumour cells progress from a less malignant to a more malignant (metastatic) phenotype, due to an inherent instability within the genome. It has also been suggested that genomic instability and the rate of generation of metastatic variants both increase as the tumour cells achieve a higher state of malignancy. In this review, several different aspects of genomic instability have been discussed with particular reference to the low (F1) and high (BL6, ML8) metastatic variants of a B16 murine melanoma. The induced mutation frequency, gene amplification, and sister chromatid recombination correlated with the order of metastatic potential. Survival data for x-irradiation and bleomycin were similar between the three cell lines. When these agents were used to induce strand breakage, no difference was detectable in either the rate or extent of DNA single or double strand break repair. The fidelity of double strand break religation was measured using the PMH16 plasmid, and the frequency of homologous recombination was measured using the DR plasmid. Surprisingly, both of these parameters showed a correlation with metastasis which was opposite to the expected result, with regard to the concept of genomic instability. A basis for this might be the functional heterogeneity within the cell lines in respect of the properties measured. Therefore, some aspects of instability may have to be measured at a more critical level if they are to be of value in estimating the inherent instability of the genome.