Dual inhibition of survivin and MAOA synergistically impairs growth of PTEN-negative prostate cancer.

BACKGROUND: Survivin and monoamine oxidase A (MAOA) levels are elevated in prostate cancer (PCa) compared to normal prostate glands. However, the relationship between survivin and MAOA in PCa is unclear. METHODS: We examined MAOA expression in the prostate lobes of a conditional PTEN-deficient mouse model mirroring human PCa, with or without survivin knockout. We also silenced one gene at a time and examined the expression of the other. We further evaluated the combination of MAOA inhibitors and survivin suppressants on the growth, viability, migration and invasion of PCa cells. RESULTS: Survivin and MAOA levels are both increased in clinical PCa tissues and significantly associated with patients' survival. Survivin depletion delayed MAOA increase during PCa progression, and silencing MAOA decreased survivin expression. The combination of MAOA inhibitors and the survivin suppressants (YM155 and SC144) showed significant synergy on the inhibition of PCa cell growth, migration and invasion with concomitant decrease in survivin and MMP-9 levels. CONCLUSIONS: There is a positive feedback loop between survivin and MAOA expression in PCa. Considering that survivin suppressants and MAOA inhibitors are currently available in clinical trials and clinical use, their synergistic effects in PCa support a rapid translation of this combination to clinical practice.

Enhanced expression of c-myc and decreased expression of c-fos protooncogenes in chemically and radiation-transformed C3H/10T1/2 Cl 8 mouse embryo cell lines.

c-abl, c-fos, c-Ha-ras, c-myc, and c-mos were expressed whereas c-sis, c-fms, c-rel, c-src, and c-myb expression was not detectable in C3H/10T1/2 Cl 8 (10T1/2) cells and in eight chemically and radiation-transformed 10T1/2 cell lines. The expression of c-abl, c-fos, c-Ha-ras, and c-myc was growth-related in nontransformed 10T1/2 cells. c-abl and c-fos expression increased at confluence by 5- and 9-fold, respectively, compared to that in log phase cells. c-Ha-ras and c-myc transcripts were most abundant in log phase cells and decreased by 70 and 50%, respectively, in confluent cells. There were no significant growth-related changes in the expression of c-Ha-ras, c-myc, or c-abl in methylcholanthrene-transformed Cl 15 cells. The c-fos transcript was not detected in Cl 15 cell cultures. c-abl, c-fos, c-ras, and c-myc were expressed in whole C3H mouse embryo tissue, mouse liver, and 10T1/2 cells. Sizes of these protooncogene transcripts in 10T1/2 cells were the same as those in whole embryo tissue, except that 10T1/2 cells did not express the 8.2-kilobase abl transcript. At subconfluence, equivalent low levels of c-mos expression were observed in nontransformed and in the eight transformed 10T1/2 cell lines. The level of c-abl expression was similar in the nontransformed and in the eight transformed cell lines, but there was a new 8.2-kilobase transcript in the transformed MCA Cl 15 cell line. c-fos was expressed in 10T1/2 cells but was not detectable or greatly reduced in eight transformed cell lines. c-Ha-ras was expressed to a similar extent in eight transformed cell lines and in nontransformed 10T1/2 cells. In the UVC-4 transformed cell line, extra 3.3-kilobase Ha-ras and 7.5-kilobase Ki-ras transcripts were observed. c-myc was expressed at 4- to 7-fold higher levels in six transformed cell lines compared to 10T1/2 cells. There were no major rearrangements in or amplification of the c-myc gene in three transformed cells overexpressing this gene 5-fold. These studies show that enhanced expression of c-myc and decreased expression of c-fos correlate with the chemically and radiation transformed states of 10T1/2 cells. Changes in c-fos and c-myc oncogene expression may be casually linked to late stages of neoplastic transformation in these chemically and radiation transformed 10T1/2 cell lines.

Enhanced expression and state of the c-myc oncogene in chemically and X-ray-transformed C3H/10T1/2 Cl 8 mouse embryo fibroblasts.

We examined expression of the c-myc oncogene in nontransformed, in three chemically transformed, and in two X-ray-transformed C3H/10T1/2 Cl 8 cell lines. In nontransformed C3H/10T1/2 cells, c-myc was expressed when cells were logarithmically growing, and expression decreased as cells reached confluence. In a methylcholanthrene-transformed cell line, MCA Si 24, c-myc expression was similar to that observed in nontransformed cells, while in two chemically transformed cell lines, Bleo Cl 2 and DMBA Cl 2, and in two radiation-transformed cell lines, F17 and F29, steady-state levels of the c-myc transcript were 5-7-fold greater than observed in nontransformed C3H/10T1/2 cells. All cell lines, both transformed and nontransformed, produced a 2.3-kilobase c-myc transcript. There was no detectable amplification or rearrangement of c-myc DNA sequences in any of the cell lines examined as determined by DNA dot blot and restriction endonuclease-Southern blotting analyses. In addition, the c-myc gene in nontransformed and transformed cell lines showed similar methylation patterns as determined by HpaII/MpsI digestion analysis, except that F19 and F29 cells lost a 0.95-kilobase HpaII band, suggesting extra region-specific methylation in these two cell lines compared to C3H/10T1/2 cells. Therefore, increased c-myc expression in the four transformed lines did not generally correlate with changes in DNA methylation in the vicinity of the c-myc gene. Our results suggest that expression of the c-myc gene is growth related and that elevated steady-state levels of c-myc RNA in certain chemically and X-ray transformed C3H/10T1/2 cell lines, such as Bleo Cl 2, DMBA Cl 2, F17, and F29, are correlated with and may participate in conversion to or maintenance of cells in the transformed state.

The effect of fibroblast growth factor 8, isoform b, on the biology of prostate carcinoma cells and their interaction with stromal cells.

Fibroblast growth factor 8, isoform b (FGF8b), has been implicated in the oncogenesis of the prostate and mammary epithelia. We examined whether overexpression of FGF8b in a weakly tumorigenic prostate carcinoma cell line, LNCaP, could alter the growth and tumorigenic properties of these cells. LNCaP cells were infected with a lentivirus vector carrying FGF8b cDNA and the green fluorescent protein (GFP) cDNA in the same construct, and the infected cell population was sorted on the basis of GFP protein expression. It was demonstrated that, in comparison with the cells transduced with GFP-vector alone, LNCaP cells with FGF8b-GFP expression manifested an increased growth rate, higher soft agar clonogenic efficiency, enhanced in vitro invasion, and increased in vivo tumorigenesis. Most strikingly, whereas parental or vector-control LNCaP cells failed to grow at all in an in vivo tumorigenesis/diaphragm invasion assay in nude mice, the cells overexpressing FGF8b proliferated as deposits of tumor cells on the diaphragm, frequently with indications of tumor cell invasion into the diaphragm. Coculturing of primary prostatic or non-prostatic stromal cells with the infected LNCaP cells led us to observe that: (a) stromal cells, irrespective of tissue origin, strongly suppressed LNCaP cell growth; (b) FGF8b producing LNCaP cells could partially evade the stromal inhibition, perhaps from the autocrine stimulatory effect of FGF8b; and (c) production of FGF8b in the coculture had a stimulatory effect on the proliferation of the stromal cells, prostatic or non-prostatic. This stimulation was not attributable to the direct action of FGF8b on stromal cells. Instead, it appears that epithelial-stromal cell-cell contact and some unknown soluble factors secreted by LNCaP cells upon stimulation of FGF8b are required for the maximal effect. Together, these results suggest that the growth rate and biological behavior of prostatic cancer cells can be altered to a more aggressive phenotype by up-regulation of FGF8b expression. These changes in phenotype also influence the interaction of the affected cells with stromal cells. The data obtained may have direct relevance to the progression of prostate cancer, recognizing that FGF8b is naturally overexpressed in advanced disease.

Downregulation of human FGF8 activity by antisense constructs in murine fibroblastic and human prostatic carcinoma cell systems.

Previously, we described cloning of three alternatively spliced mRNA forms of human FGF8, a, b, and e, of which the b form is the major expressed species in both normal and tumor prostatic epithelial cells. In this report, we describe construction and overexpression of sense and antisense sequences of either the full length FGF8b coding region (215-amino acids or 215aa), 103aa N-terminal part or a smaller N-terminal region (34aa), each including the 23aa putative signal peptide domain, via a retrovirus system. While the morphologic transforming activities of the sense 215aa and 103aa constructs were similar in NIH3T3 cells, 103aa displayed reduced soft agar clonogenic activity. The 34aa construct was practically inert in these assays, although its expression could mimic the ability of 215aa or 103aa in conferring cell growth under reduced serum condition. Overexpression of any of the three constructs in antisense orientation, however, was similarly effective in reversing the morphology and anchorage-independent growth property of FGF8b-transfected NIH3T3 cells. The expression of the antisense 215aa construct significantly reduced the growth rate of the human prostatic carcinoma DU145 cells and inhibited their soft agar clonogenic activity and in vivo tumorigenicity in nude mice. Taken together, these results identify N-terminal portions of FGF8 protein isoform for having the domains necessary for one or more of the biologic effects examined, and suggest that low levels of FGF8 expressed in prostatic epithelial cells may contribute significantly to their growth and tumorigenic properties.

Retrovirus receptor PiT-1 of the Felis catus.

We isolated a cDNA encoding a feline homolog of human PiT-1, a sodium-dependent phosphate symporter which is utilized by gibbon ape leukemia virus (GALV) as a receptor for entry into host cells. The overall homology between the human and feline receptors is 92 and 93% at the nucleotide and deduced amino acid levels, respectively. Hydropathy analyses implied ten potential membrane spanning regions and, in analogy to human and murine homologs, five extracellular and four intracellular loops. Strikingly, the amino acid sequence of the fourth extracellular loop, which is critical for GALV surface glycoprotein binding, has complete identity between the human and feline PiT-1s, while the mouse PiT-1, non-functional for GALV entry, is quite divergent. Ectopic expression of the feline PiT-1 in guinea pig cells, which are non-permissive to feline leukemia virus (FeLV), subgroup B virus, conferred susceptibility to FeLV-B infection confirming the functional ability of the cloned product to serve as a receptor for a natural retrovirus of the homologous species.

North-American twins with IDDM. Genetic, etiological, and clinical significance of disease concordance according to age, zygosity, and the interval after diagnosis in first twin.

In 224 twin pairs (132 monozygotic, 86 dizygotic, and 6 of uncertain zygosity) in whom the index twin had developed IDDM before 30 yr of age, 51 of the co-twins (38 monozygotic, 10 dizygotic, and 3 of uncertain zygosity) subsequently became diabetic. On the basis of concordance ratios, which were significantly discrepant (P < 0.01) between monozygotic and dizygotic twins, the substantial genetic role in IDDM etiology is confirmed. For the monozygotic co-twin of an IDDM case, the relative risk is significantly related to an early age at proband diagnosis (P < 0.01 for 0-4 vs. 5-9 yr of age). However, among monozygotic co-twins at any age, IDDM risk decreases as time passes after the proband diagnosis (P < 0.01 for 0-23 vs. > or = 24 mo after a proband diagnosis at 5-9 yr of age). Moreover, a structural-equation analysis suggests a profound contribution to liability (as much as 79%) from the twins' shared environment. Risk to like-sex male dizygotic co-twins is as high as that to monozygotic co-twins, significantly higher than that to like-sex female dizygotic co-twins (P < 0.005), and even higher than that to male co-twins in unlike-sex dizygotic pairs (P < 0.05). Overall, the risk to the dizygotic co-twin of a case is significantly higher (P < 0.001) than that to a non-twin sibling, as reported in the literature. The observed male excess is consistent with reported patterns of IDDM in experimental animals, and in certain circumstances in humans. Taken together, these observations suggest an important early acquired determinant of IDDM, independent of genetic determinants. On the basis of Kaplan-Meier IDDM-free survival curves, if the proband is diagnosed before 15 yr of age, the long-term risk to the co-twin is estimated at 44% (monozygotic) and 19% (dizygotic); it reaches 65% for the co-twin of a monozygotic proband diagnosed before 5 yr of age. An IDDM discordant period of no more than 3 yr was observed in 60% of the pairs destined to become concordant, offering a very brief window for intervention following the recognition of high risk.

Generation of a prostate epithelial cell-specific Cre transgenic mouse model for tissue-specific gene ablation.

To facilitate the elucidation of the genetic events that may play an important role in the development or tumorigenesis of the prostate gland, we have generated a transgenic mouse line with prostate-specific expression of Cre recombinase. This line, named PB-Cre4, carries the Cre gene under the control of a composite promoter, ARR2PB which is a derivative of the rat prostate-specific probasin (PB) promoter. Based on RT-PCR detection of Cre mRNA in PB-Cre4 mice or Cre-mediated activation of LacZ activity in PB-Cre4/R26R double transgenic mice, it is conclusively demonstrated that Cre expression is post-natal and prostatic epithelium-specific. Although the Cre recombination is detected in all lobes of the mouse prostate, there is a significant difference in expression levels between the lobes, being highest in the lateral lobe, followed by the ventral, and then the dorsal and anterior lobes. Besides the prostate gland, no other tissues of the adult PB-Cre4 mice demonstrate significant Cre expression, except for a few scattered areas in the gonads and the stroma of the seminal vesicle. By crossing the PB-Cre4 animals with floxed RXRalpha allelic mice, we demonstrate that mice, whose conventional knockout of this gene is lethal in embryogenesis, could be propagated with selective inactivation of RXRalpha in the prostate. Taken together, the results show that the PB-Cre4 mice have high levels of Cre expression and a high penetrance in the prostatic epithelium. The PB-Cre4 mice will be a useful resource for genetic-based studies on prostate development and prostatic disease.

Genetically defined mouse models that mimic natural aspects of human prostate cancer development.

This review is focused on mouse models for prostate cancer that have been designed on the basis of genetic alterations that are frequently found in human prostate cancer. It begins with an analysis of the similarities and differences in the gross and microscopic anatomy of the mouse and human prostate glands, and extends to the pathologies induced in the genetically manipulated mouse prostate in comparison with the sporadic development of the disease in humans. Major achievements have been made in modeling human prostate cancer in mice in recent years. There are models which display slow, temporal development of increasingly severe preneoplastic lesions, which are remarkably restricted to the prostate gland, a property similar to the aging-related progression of these lesions in humans. Other models rapidly progress to local invasive adenocarcinoma, and, in some of them metastasis is manifested subsequently with defined kinetics. Global assessment of molecular changes in the prostate of the genetically manipulated mice is increasingly underscoring the validity of the models through identification of 'signature' genes which are associated with the organ-confined primary or distant metastases of human prostate cancer. Taken together, various 'natural' models depicting stages of the disease, ranging from the early preneoplastic lesions to metastatic prostate cancer, now provide new tools both for exploring the molecular mechanism underlying prostate cancer and for development or testing of new targeted therapies.

Structure of the polymorphic feline c-myc oncogene locus.

Two feline c-myc DNA clones (CM-2 and CM-3), isolated from a cat DNA library, are structurally very similar. However, they differ at a SmaI site in exon III which is present only in CM-2. In the outbred feline population, cats heterozygous for this site or homozygous for the CM-3-type gene have been observed. The results provide a physical map of the feline c-myc locus, and define hitherto unidentified alleles of this gene.

Nucleotide sequence of the 1.2-kb 3'-region and genotype distribution of two common c-myc alleles of the domestic cat.

Nucleotide (nt) sequence analyses of the 1.2-kb BamHI-EcoRI cloned 3'-fragment encompassing the polymorphic SmaI restriction site of the feline c-myc gene reveal that the SmaI site, present in CM2 allele but absent from CM3 allele, is located in intron 2, 134 nt 5' of the exon 3. A G-to-C transversion in CM2 results in the creation of the SmaI site. Additionally, the alleles differ at four other nt positions in intron 2, three of these changes being in a region of the intron which exhibits 80% homology between the feline and human c-myc. The alleles also differ in two nt positions in exon 3 in the third position of the codon resulting, however, in no amino acid alteration. Genotype distribution analysis based on the SmaI polymorphism shows that CM2 homozygosity is rare and its frequency deviates significantly from the expected distribution patterns for independently segregating alleles.

Fusion activity dissociated from replication ability in feline immunodeficiency virus (FIV) in human cells.

Multinucleated-giant-cell formation followed by cell killing was observed after cocultivation of the feline immunodeficiency virus (FIV)-producing feline T-cell line 3201/FIV with various human cells, including T-cell lines carrying human T-cell lymphotropic virus type I (HTLV-I). The susceptibility to giant cell formation varied with the cell lines tested. Cocultivation of irradiated 3201/FIV cells with MT-2 cells resulted in giant cell formation as early as 2 h in culture, with striking resemblance to that induced by human immunodeficiency virus (HIV). MT-4 cells (HTLV-I positive) and H9 cells (HTLV-I negative) were less susceptible than MT-2 to the induction of syncytia. MOLT-4 cells (HTLV-I negative) had intermediate sensitivity to syncytia formation. No syncytia were observed in the monocytic cell line U-937 (HTLV-I negative). Syncytia formation between 3201/FIV and MT-2 cells was inhibited by polyclonal cat anti-FIV antisera but not polyclonal cat anti-feline leukemia virus (FeLV) antisera, goat anti-FeLV, uninfected specific-pathogen-free cat serum, human anti-HTLV-I antisera, or normal human and goat serum. Concentrated cell-free FIV supernatant from 3201/FIV also induced giant cells of MT-2 cells that were indistinguishable from those induced by cocultivation. Giant cells and extensive cell killing associated with giant cell formation declined and disappeared within 10 days. Surviving cells appeared to be of normal size and grew continuously without expressing FIV antigen or releasing infective virus. Although Southern blot analysis using probes specific for FIV could not detect proviral DNA in any of the five human cell lines cocultured with irradiated 3201/FIV cells, the polymerase chain reaction (PCR) technique detected FIV-specific DNA in MOLT-4 cells. DNA from the FIV-PCR positive MOLT-4 cells was PCR negative for endogenous FeLV-specific sequences, indicating that the MOLT-4 cell DNA was not contaminated with DNA from feline cells (i.e., 3201 cells). The FIV-MOLT-4 cells remained PCR positive for FIV after 40 passages, suggesting stable integration in the human cell line. These findings indicate that FIV is capable of forming proviral DNA in human T-lymphoid cells by cocultivation, although this FIV-carrying human cell line failed to produce replication-competent viruses.

Mouse strains for prostate tumorigenesis based on genes altered in human prostate cancer.

Animal models of prostate cancer have been limited in number and in relevance to the human disease. With the advancement of transgenic and knockout technologies, combined with tissue specific promoters and tissue-specific gene ablation, a new generation of mouse models has emerged. This review will discuss various animal models and their inherent strengths and weaknesses. A primary emphasis is placed on mouse models that have been designed on the basis of genetic alterations that are frequently found in human prostate cancer. These models display slow, temporal development of increasingly severe histopathologic lesions, which are remarkably restricted to the prostate gland, a property similar to the ageing related progression of this disease in humans. The preneoplastic lesions, akin to what is considered as prostatic intraepithelial neoplasia, are consistent major phenotypes in the models, and, therefore. are discussed for histopathologic criteria that may distinguish their progressions or grades. Finally, considering that prostate cancer is a complex multifocal disease, which is likely to require multiple genetic/epigenetic alterations, many of these models have already been intercrossed to derive mice with compound genetic alterations. It is predicted that these and subsequent compound mutant mice should represent "natural" animal models for investigating the mechanism of development of human prostate diseases, as well as, for preclinical models for testing therapeutics.

The long terminal repeat of feline endogenous RD-114 retroviral DNAs: analysis of transcription regulatory activity and nucleotide sequence.

Six cloned 5' long terminal repeat (LTR) and adjoining cellular DNA regions of partially deleted feline endogenous RD-114 proviral loci were linked to the chloramphenicol acetyltransferase (CAT) gene and assayed for their ability to promote transient CAT expression. One endogenous LTR (clone CRL-3) and the LTR from an infectious RD-114 provirus, EX-LTR, were capable of actively expressing the CAT gene. DNA sequence comparison of these LTRs with an inactive endogenous LTR (CR-1) revealed extensive homology in all regions except in the 5' half of U3. The homologous portion contained transcriptional regulatory sequences including CAT, TATA, polyadenylation signal boxes and an octamer enhancer, which is rarely seen in retroviruses. Variations in the 5' half of U3 were primarily due to insertions and deletions. A major difference was in number of copies and integrity of tandemly repeated sequences. EX-LTR contained two pairs of tandem direct repeats, while the two endogenous LTRs contained different deletions of repeated sequences. DNA sequence data also revealed that the primer binding site for RD-114 loci was complementary to a glycine tRNA isotype, the use of which is distinct from any other known retrovirus. An analysis of the steady state RNA levels in T-lymphoid cell lines showed that at least three different incomplete proviral transcripts and their spliced products made up the majority of expressed RD-114 mRNA, and further demonstrated that partially deleted proviral loci have the potential to be transcriptionally vigorous in certain feline cell types.

Genetic analysis of the feline RD-114 retrovirus-related endogenous elements.

The structure of the multiple copies of the endogenous RD-114 retroviral sequences in the domestic cat were compared by restriction mapping of cloned and genomic DNAs. Six different clones, each representing a different loci, were randomly chosen from 20 clones isolated from a cat DNA library. Whereas all clones contained viral structural sequences which were extensively homologous to the gag and pol regions of exogenous RD-114, they varied from each other in the extent of deletions (0.7-1.9 kilobase pairs) in the env region. Each clone contained a pair of LTRs which bracketed the structural genes, implicating an origin from exogenous infection of a distant ancestor of the domestic cat. Analysis of genomic DNA from six separate domestic cats also indicated that the majority of the RD-114 endogenous loci contained variably deleted env and conserved gag-pol regions. The occurrence of at least 13 unique RD-114-related loci could be detected in each cat DNA examined. It appeared that among these multiple copies inherited in the cat genome, there is only a single copy which would correspond to the undeleted RD-114 provirus, and which may represent the locus for the inducible RD-114 virus.

In vivo evolution and selection of recombinant feline leukemia virus species.

Ecotropic feline leukemia viruses subgroup A (FeLV-A) is known to recombine with endogenous FeLV (enFeLV) env elements yielding polytropic FeLV-B viruses. However, scattered nucleotide differences exist between enFeLV env elements and corresponding sequences of exogenous FeLV-B isolates. To address this disparity, we examined recombinant FeLV (rFeLV) viruses obtained from three experimentally-induced feline thymic tumors, along with rFeLVs derived from one naturally-occurring thymic tumor. Two of the three experimental cats were challenged with a FeLV-A/Rickard preparation, while one cat received this FeLV-A along with a mixture of in vitro-generated rFeLVs. The FeLV-A/Rickard preparation employed in this study was shown to be free of detectable rFeLVs since no recombinant products were observed in this preparation following nested PCR analyses. For each of the four tumor DNAs, nucleotide sequence analysis was performed on multiple clones of rFeLV-specific PCR products derived from the surface glycoprotein (SU) portion of the recombinant proviral env gene. Relative to the parental enFeLV sequence used to generate the rFeLVs, a total of 19 nucleotide differences were found scattered within the SU region of the env gene in these in vivo-derived rFeLV clones. Most interestingly, this set of 19 differences led to complete sequence identity with natural FeLV-B isolates. Our results indicate these differences are present early in the in vivo evolution of recombinant viruses, suggesting that rFeLVs harboring these differences are strongly selected. We also present evidence indicating an in vivo selection pattern exists for specific recombinant species containing relatively greater amounts of enFeLV-derived SU sequence. This in vivo selection process appears to be gradual, occurring over the infection timecourse, yielding rFeLV species which have recombination structural motifs similar to those seen in natural FeLV-B isolates.