¹Department of Urology, Herbert Irving Comprehensive Cancer Center, Columbia University, College of Physicians and Surgeons, New York, NY 10032, USA

²Department of Pathology, Herbert Irving Comprehensive Cancer Center, Columbia University, College of Physicians and Surgeons, New York, NY 10032, USA

³Department of Neurosurgery, Herbert Irving Comprehensive Cancer Center, Columbia University, College of Physicians and Surgeons, New York, NY 10032, USA

⁴Introgen Therapeutics Incorporated, Houston, Texas, TX 77030, USA

⁵Center for Mechanistic Biology and Biotechnology, Argonne National Laboratories, Argonne, Illinois, IL 60439, USA

⁶Department of Molecular Genetics and Microbiology, University of Medicine and Dentistry of New Jersey, UMDNJ-Robert Wood Johnson Medical School, Piscataway, New Jersey, NJ 08854, USA

Correspondence to: P B Fisher, Departments of Pathology and Urology, Columbia University, College of Physicians and Surgeons, BB 15-1501, 630 West 168th Street, New York, NY 10032, USA; E-mail: pbf1@columbia.edu

^aCurrent address: Bristol Myers-Squibb, Age-related Diseases, MCDD 311 Pennington-Rocky Hill Road, Pennington, NJ 08534, USA

^bCurrent address: DGI Biotechnologies Incorporated, Molecular Biology, 40 Talmadge Road, Edison, NJ 08818, USA

^cEY Huang, MT Madireddi, RV Gopalkrishnan, M Leszczyniecka, Z-z Su and IV Lebedeva contributed equally to this manuscript

Abnormalities in cellular differentiation are frequent occurrences in human cancers. Treatment of human melanoma cells with recombinant fibroblast interferon (IFN-) and the protein kinase C activator mezerein (MEZ) results in an irreversible loss in growth potential, suppression of tumorigenic properties and induction of terminal cell differentiation. Subtraction hybridization identified melanoma differentiation associated gene-7 (mda-7), as a gene induced during these physiological changes in human melanoma cells. Ectopic expression of mda-7 by means of a replication defective adenovirus results in growth suppression and induction of apoptosis in a broad spectrum of additional cancers, including melanoma, glioblastoma multiforme, osteosarcoma and carcinomas of the breast, cervix, colon, lung, nasopharynx and prostate. In contrast, no apparent harmful effects occur when mda-7 is expressed in normal epithelial or fibroblast cells. Human clones of mda-7 were isolated and its organization resolved in terms of intron/exon structure and chromosomal localization. Hu-mda-7 encompasses seven exons and six introns and encodes a protein with a predicted size of 23.8 kDa, consisting of 206 amino acids. Hu-mda-7 mRNA is stably expressed in the thymus, spleen and peripheral blood leukocytes. De novo mda-7 mRNA expression is also detected in human melanocytes and expression is inducible in cells of melanocyte/melanoma lineage and in certain normal and cancer cell types following treatment with a combination of IFN- plus MEZ. Mda-7 expression is also induced during megakaryocyte differentiation induced in human hematopoietic cells by treatment with TPA (12-O-tetradecanoyl phorbol-13-acetate). In contrast, de novo expression of mda-7 is not detected nor is it inducible by IFN-+MEZ in a spectrum of additional normal and cancer cells. No correlation was observed between induction of mda-7 mRNA expression and growth suppression following treatment with IFN-+MEZ and induction of endogenous mda-7 mRNA by combination treatment did not result in significant intracellular MDA-7 protein. Radiation hybrid mapping assigned the mda-7 gene to human chromosome 1q, at 1q 32.2 to 1q41, an area containing a cluster of genes associated with the IL-10 family of cytokines. Mda-7 represents a differentiation, growth and apoptosis associated gene with potential utility for the gene-based therapy of diverse human cancers. Oncogene (2001) 20, 7051-7063.

melanoma differentiation associated gene-7; terminal cell differentiation; gene expression; genomic structure; IL-10 cytokine family

Introduction

Cancer is a progressive disease in which evolving tumor cells exhibit profound changes in gene expression that occur in a temporal manner (DeRisi et al., 1996; Schena et al., 1996; Ross et al., 2000). These include, activation of genes serving as positive regulators of the cancer phenotype, oncogenes; inactivation of genes functioning as inhibitors of the cancerous state, tumor suppressor genes; and altered expression of genes that act at a later stage in cancer development resulting in increased cancer aggressiveness which can culminate in metastasis, progression modulating genes (Fisher, 1984; Bishop, 1991; Vogelstein and Kinzler, 1991; Su et al., 1997, 1999; Kang et al., 1998). Extensive effort is being directed toward identifying and characterizing the genetic elements that contribute to cancer development and evolution. Achieving this goal offers promise for identifying genomic targets for improved cancer diagnosis and therapy.

A failure to display normal patterns of differentiation is common in many cancer subtypes (Jiang et al., 1994; Waxman, 1996). Recent experimental approaches have targeted this defect in cancer cells by using agent(s) that modify tumor growth by inducing terminal differentiation, a process termed 'differentiation therapy' (Waxman, 1996). Although several models exist to explain how metastatic melanoma cells develop, a widely accepted scheme involves a linear set of changes from melanocyte to nevus to primary melanoma [radial growth phase (RGP) to vertical growth phase (VGP)] to metastatic melanoma (Clark, 1991; Armstrong and Kricker, 1994; Herlyn et al., 2000). This process is associated with profound changes in cellular physiology and gene expression suggesting temporal modifications in the differentiation of cells as this tumor evolves (Jiang et al., 1995a; Huang et al., 1999a,b; Herlyn et al., 2000). In human melanoma cells, the combination of recombinant human IFN- plus the protein kinase C activator MEZ results in an irreversible loss in proliferative ability concluding in terminal cell differentiation (Fisher et al., 1985; Jiang et al., 1993). Combining this model system with subtraction hybridization resulted in the identification and cloning of several genes regulated during the process of growth arrest and terminal differentiation, i.e., melanoma differentiation associated (mda) and differentiation induction subtraction hybridization (DISH) genes (Jiang and Fisher, 1993; Jiang et al., 1995a,b, 2000; Huang et al., 1999a,b; Kang et al., 2001).

A specific mda gene, mda-7, displays elevated mRNA expression in melanocytes and nevi, reduced mRNA expression in RGP and early VGP melanoma lesions and little or no mRNA expression in late VGP and metastatic melanoma (Jiang et al., 1995b). Moreover, ectopic expression of mda-7 in human melanoma cells results in growth suppression, without induction of terminal differentiation (Jiang et al., 1995b). When expressed at high levels, by means of an adenovirus expression system, mda-7 induces growth suppression and programmed cell death (apoptosis) in a broad spectrum of human cancers, including carcinomas of most tissue origins (Su et al., 1998; Madireddi et al., 2000c; Saeki et al., 2000; Mhashilkar et al., 2001). In contrast, mda-7 has a negligible effect on growth and does not induce apoptosis in normal epithelial and fibroblast cells (Jiang et al., 1996; Su et al., 1998; Madireddi et al., 2000c; Saeki et al., 2000; Mhashilkar et al., 2001). These findings indicate that this novel cancer growth suppressing and apoptosis-inducing gene may have wide applications for the gene-based therapy of multiple human cancers (Jiang et al., 1996; Su et al., 1998; Madireddi et al., 2000c; Saeki et al., 2000; Mhashilkar et al., 2001).

Studies designed to define the mechanism of mda-7 gene expression regulation during human melanoma terminal differentiation document that the mda-7 promoter is constitutively active in human melanoma cells and treatment with IFN-+MEZ does not significantly alter promoter function (Madireddi et al., 2000a,b). In contrast, mda-7 mRNA levels during terminal differentiation are partly regulated by differential post-transcriptional-message stabilization dictated by the AU-rich (ARE) sequences present in the 3'-untranslated region of the mda-7 cDNA (Madireddi et al., 2000a). To further characterize this cancer suppressor gene we have isolated mda-7 genomic DNA from human cells and defined its gene structure and analysed endogenous and inducible expression in melanocyte/melanoma and other lineage cells. Expression of mda-7 is restricted to specific normal tissues, including thymus, spleen and blood (peripheral blood leukocytes), and certain cell types, including melanocytes, nevi and some early stage melanomas. Enhanced mda-7 expression is also induced in melanoma cells following treatment with IFN-+MEZ, in K562 erythroleukemia cells treated with TPA and in specific human normal and cancer cells following treatment with IFN-+MEZ. These results demonstrate for the first time that appropriate treatment(s) can induce mda-7 mRNA in a spectrum of normal and cancer cells, in addition to melanoma cells, which do not express detectable levels of mda-7 mRNA de novo. The level of mda-7 mRNA induced, as well as the level of MDA-7 protein, is significantly less following IFN-+MEZ treatment than infection with a replication defective adenovirus expressing mda-7, Ad.mda-7. Moreover, IFN-+MEZ inhibits growth in a panel of normal and cancer cells and this effect does not correlate with induction of mda-7 mRNA, but rather corresponds with IFN- treatment which does not induce mda-7 mRNA. The chromosomal location of human mda-7 was determined and found to reside within a region of human chromosome 1q, 1q 34.2 to 1q41, which corresponds with a genomic locus containing several members of the interleukin-10 family of cytokines.

Results

Genomic structure and chromosomal localization of mda-7 to a genomic region containing an interleukin-10 related gene cluster

Restriction fragment analyses of human genomic DNA demonstrate that the human mda-7 gene is a single copy gene. Southern blot analyses reveal a single band upon hybridization with a mda-7 cDNA probe (Jiang et al., 1995b). Based on its unique presence in the human genome, a PCR-based method using mda-7 gene specific primers was used to identify and isolate its genomic DNA. Human genomic DNA from diploid fibroblasts (Clontech) was used as a template in a polymerase chain reaction amplification protocol with mda-7 gene specific primers corresponding to the 5' and 3' untranslated regions. Agarose gel electrophoresis analyses identified a 5.5 Kbp amplification product, which was cloned into pBluescript, and the nucleotide sequence was obtained. In order to obtain genomic sequence information flanking the PCR generated fragment a human placental genomic library (Stratagene) was screened. Five clones were isolated from 2´10⁶ plaques of human placental genomic library using an mda-7 cDNA probe. Restriction enzyme digestion and probing with 5'- and 3'-UTR specific sequences revealed that one of the clones encompasses the entire reading frame of the mda-7 gene. This phage clone was purified and sequenced using mda-7 gene specific primers. Using the two methods described above a contiguous mda-7 genomic sequence was obtained. The mda-7 transcription unit is 6.33 Kbp, subsequent DNA walking resulted in the cloning of an additional 2.2 Kbp of the 5'-flanking region which contains the mda-7 promoter (Madireddi et al., 2000a). The nucleotide sequence of human mda-7 and exon/intron boundaries are shown in Figure 1. The mda-7 cDNA deduced from the human exon sequence information corresponded exactly with the previously reported human mda-7 cDNA (Jiang et al., 1995b).

The identified intron-exon boundaries conform to the consensus splicing signals (GT…AG). These dinucleotide sequences are putative splice sites implicated in primary transcript splicing (Breathnach et al., 1978; Breathnach and Chambon, 1981). The exons range in size from 64 to 889 bp while the introns range from 115 to 1443 bp. The transcriptional start site was mapped by 5' RACE using total RNA from terminally differentiated HO-1 human melanoma cells (Jiang et al., 1995b). Primer extension analysis produced similar results (data not shown), and the transcription start site was designated, and its position is shown in Figure 1. Analyses of the 5'-upstream nucleotide sequence reveals the presence of a TATA element at position -30 to -25 (Figure 1). Elimination of this TATA element resulted in a complete loss in promoter activity (Madireddi et al., 2000a).

A panel of rodent-human hybrid DNAs containing most human chromosome regions were tested for the presence of the mda-7 locus by a PCR amplification based method. Hybrid DNAs were scored positive if they contained a 129 bp human mda-7 specific product. Hybrids retaining chromosome region 1q showed a mda-7 specific product (data not shown). To further refine the localization of mda-7, the Stanford and Genebridge radiation hybrid DNAs (Research Genetics) were tested. No linkage was found using the Stanford radiation hybrid DNAs. Using the Genebridge radiation hybrids and the WICGR mapping server (www.genome.wi.mit.edu/cgibin/contig/rhmapper.pl) mda-7 displayed very close linkage to markers WI-9641 (D1S306) and D1S491, which map to the region 1q32.2-q41 (data not shown). The genomic locus encoding mda-7 appears to be within a IL-10 related gene cluster containing four genes including IL-10, IL-19, IL-20 and mda-7 in linear order spanning 195 kb of genomic DNA (Blumberg et al., 2001). The functional implications of this arrangement are presently unclear.

Mda-7 displays restricted expression in normal tissue and inducible expression in human erythroleukemia cells

Multiple tissue Northern blot analyses of human tissue poly(A)⁺ RNA (Clontech^R) reveals that mda-7 expression is restricted to those tissues associated with the immune system such as, spleen, thymus and peripheral blood leukocytes (Figure 2a). These findings are in agreement with the observations made by Soo et al. (1999) who reported significantly elevated (9-12-fold) levels of c49a (rat homologue of mda-7) mRNA during wound healing, specifically in areas surrounding the edge of the wound. To further analyse this relationship the incidence of mda-7 expression in hematopoietic (erythroid, myeloid and lymphoid) cell differentiation was analysed. Northern blotting analyses of total RNA from leukemic cell lines induced to differentiate with TPA were performed. The results of this study reveal an increase in mda-7 mRNA levels in the leukemic cell line, K562 (erythroleukemia), when it is induced to differentiate into megakaryocytes upon treatment with TPA (Figure 2b). Mda-7 mRNA expression in differentiated K562 cells (Figure 2b, lane 5) is similar to that found in terminally differentiated HO-1 human melanoma cells after correction for variations in total RNA based on the levels of rRNA (Figure 2b, lane 9). In contrast, HL-60 (human promyelocytic leukemia) induced to differentiate with TPA (monocyte/macrophage) or DMSO (granulocyte), CEM-C7 (human T-cell leukemia) treated with TPA and HL534 (TPA-resistant HL-60 cell variant) (Tonetti et al., 1992) did not express mda-7 in the absence or presence of inducer (Figure 2b).

De novo expression of mda-7 occurs in normal and immortalized human melanocytes and is readily inducible in most human melanoma cell lines

Experiments have been performed to examine mda-7 expression profiles in normal and cancer-derived human cell types. In previous studies, expression of mda-7 mRNA was detected using Northern blotting in an SV40-immortalized human melanocyte cell line (FM516-SV) and in one of six metastatic melanoma cell lines (Jiang et al., 1995b). In addition, using patient-derived samples, including melanocytes (five samples), primary melanomas (seven RGP and early VGP samples) and metastatic melanomas (seven samples), an inverse correlation between melanoma progression and mda-7 RNA expression was apparent using RT-PCR based approaches, with highest levels in melanocytes, deceasing levels in primary melanomas and lowest levels in metastatic melanomas (Jiang et al., 1995b). Previous studies demonstrated that treatment of metastatic melanoma cells with IFN-+MEZ for 24 h resulted in induction or enhanced mda-7 mRNA expression in the six metastatic melanoma cell lines analysed (Jiang et al., 1995b). These studies have now been extended to include normal early passage melanocytes and a series of additional melanoma cell lines, including WM35 (an early RGP melanoma), WM278 (an early VGP melanoma) and additional metastatic melanoma cell lines (SK-MEL p53 wt, SK-MEL p53 mt, MeWo, 3S5, 70W, WM239 and C8161) (Figure 3). In these cells, elevated mda-7 mRNA was only apparent de novo in normal melanocytes, with lower levels of de novo expression in WM35 and F0-1 cells. However, 24 h treatment with IFN-+MEZ resulted in a differential enhancement or induction of mda-7 expression in all of the melanocyte/melanoma cell lines with the exception of SK-MEL p53 mt (containing a mutant p53 gene) (Figure 3). Additionally, the induction of mda-7 in C8161 cells was significantly blunted in comparison with the other metastatic melanoma cell lines. It does not appear that having a mutant p53 gene by itself can prevent mda-7 expression, since MeWo and its two subclones, 3S5 displaying a reduction in metastatic competence and 70W exhibiting enhanced metastatic potential (Kerbel and Man, 1984; Graham et al., 1991), which contain one mutant and one wild-type p53 allele are readily inducible for mda-7 expression following IFN-+MEZ treatment. These studies document de novo expression of mda-7 in cultured normal human melanocytes, as observed with patient-derived melanocytes, and they provide additional support for the idea that mda-7 may function as a negative regulator of melanoma progression (Jiang et al., 1995b).

Induction of mda-7 mRNA is not restricted to human melanoma cells, but can occur in various normal and cancer-derived human cell lines following treatment with IFN-+MEZ

Previous studies and those described above document a clear association between mda-7 mRNA expression and melanocyte/melanoma lineage cells. To determine if mda-7 expression is restricted to this cell subtype or if de novo or inducible expression occurs in additional cell types, a panel of normal and cancer-derived cells were analysed for mda-7 mRNA expression with and without a 24 h treatment with IFN-+MEZ (Figure 4). In the case of normal human prostate epithelial and human prostate carcinoma cells no expression was apparent, with or without treatment with IFN-+MEZ, in normal early passage prostate epithelial cells (HuPEC) or in two of three prostate carcinoma cell lines, i.e., LNCaP and PC-3 (Figure 4a). In contrast, although not expressed de novo, 24 h exposure to IFN-+MEZ resulted in induction of mda-7 mRNA expression in DU-145 human prostate carcinoma cells. Since DU-145 cells contain a mutant form of the tumor suppressor protein RB and they also contain a mutation in p53, induction of mda-7 can occur in specific prostate cancer cells defective in these tumor suppressor proteins. In the case of breast-derived epithelial cells, mda-7 mRNA expression was not detected de novo in six cell lines, but it was inducible by IFN-+MEZ treatment in normal HBL-100 cells and in p53 mutant MDA-MB-231 and p53-null MDA-MB-157 cells (Figure 4b). In contrast, expression of mda-7 was not apparent in untreated or combination treated MCF-7 (wild-type p53), T47D (mutant p53) or MDA-MB-453 (mutant p53) breast carcinoma cell lines. Further analysis indicates that mda-7 is also inducible by IFN-+MEZ in normal human cerebellum cells (NC cell line), one of two human glioblastoma multiforme cell lines (GBM-18, but not T98G), a human cervical carcinoma cell line (HeLa), a human nasopharyngeal carcinoma cell line (HONE-1) and a human osteosarcoma cell line (Saos2, which is null for both RB and p53) (Figure 4c). In contrast, no expression, with or without combination treatment, was apparent in SW613 human colon carcinoma or BxPC-3, PANC-1, MIA PaCa-2 or AsPC-1 human pancreatic carcinoma cells (Figure 4c and data not shown). These results demonstrate that mda-7 is not expressed de novo in most normal and cancer cell types, but expression, at least at an mRNA level, can be induced by IFN-+MEZ in a spectrum of normal and tumor cell types independent of alterations in Rb and/or p53 genotypes.

Based on the ability of IFN-+MEZ to induce mda-7 mRNA expression in cells other than melanoma, experiments were performed to evaluate the functional significance of this treatment protocol in combination treated cells. Studies were performed to determine if induction of mda-7 mRNA correlated with expression of MDA-7 protein. To address this question FM516-SV (constitutively express mda-7 mRNA) and HO-1 and DU-145 (which are inducible for mda-7 mRNA) were either infected with 1, 10, 50 or 100 pfu/cell of Ad.mda-7, a replication incompetent adenovirus expressing the coding region of the mda-7 gene (Su et al., 1998), or treated with 2000 units/ml IFN-+10 ng/ml MEZ, and the amount of mda-7 mRNA and MDA-7 protein was determined by Northern and Western blotting, respectively (Figures 5 and 6). These studies indicate quantitative differences in the levels of mda-7 mRNA and/or MDA-7 protein in the three different cell types with and without infection with Ad.mda-7 or treatment with IFN-+MEZ. In the case of DU-145, infection with Ad.mda-7 results in a dose-dependent expression of mda-7 mRNA and MDA-7 protein, with intracellular protein readily detected 24 h following infection with 50 or 100 pfu/cell but not with 1 or 10 pfu/cell (Figures 5 and 6). In contrast, the level of mda-7 mRNA following infection with 1 pfu/cell of Ad.mda-7 is ~twofold higher than observed after treatment of DU-145 with IFN-+MEZ and no intracellular protein is detected in these cells after 24 h treatment (Figure 6). In the case of FM516-SV, mda-7 mRNA is detectable de novo at a lower level than observed following infection with 1 pfu/cell of Ad.mda-7 and treatment with IFN-+MEZ elevates mda-7 mRNA in these cells (Figure 5). With respect to protein, no MDA-7 protein is detected in IFN-+MEZ treated cells and significantly less MDA-7 protein than detected in Ad.mda-7-infected DU-145 cells is present in FM516-SV cells infected with 50 or 100 pfu/cell of Ad.mda-7 (Figure 6). In HO-1, no mda-7 mRNA is present de novo and treatment with IFN-+MEZ results in induction of mda-7 mRNA at levels that exceed those observed following infection with 1 pfu/cell of Ad.mda-7 (Figure 5). In comparison with DU-145 and FM516-SV, infection of HO-1 cells with Ad.mda-7 results in less mda-7 mRNA and MDA-7 protein, with intracellular protein only detected after infection with 100 pfu/cell of Ad.mda-7 (Figures 5 and 6). These results indicate differences in the levels of mda-7 mRNA and intracellular MDA-7 protein following Ad.mda-7 infection in the three different cell types, which may reflect differences in viral infectivity or kinetics of transgene expression. Moreover, the levels of intracellular MDA-7 protein resulting from treatment with IFN-+MEZ are minimal in all three cell types. Further studies are required to determine if the levels of secreted MDA-7 protein differ in these cells upon treatment with IFN-+MEZ and/or following infection with Ad.mda-7.

Previous studies in HO-1 melanoma cells demonstrate a small induction of mda-7 mRNA following treatment with MEZ and a major induction of mda-7 mRNA following treatment with IFN-+MEZ (Jiang et al., 1995b; Madireddi et al., 2000a,b). In contrast, IFN- does not induce mda-7 expression in HO-1 cells. To determine if a similar mda-7 induction profile is apparent in non-melanoma cell types induced to express mda-7 mRNA following treatment with IFN-+MEZ the effect of treatment with IFN- and MEZ on mda-7 expression in DU-145 and HeLa cells was determined. The combination of IFN-+MEZ induced mda-7 mRNA in both cell types, whereas no induction of mda-7 mRNA occurred after MEZ or IFN- treatment (data not shown). These results provide additional evidence that the combination treatment with IFN-+MEZ is a more effective inducer of mda-7 mRNA than either agent used independently.

An important question is the physiological relevance of induction of mda-7 by IFN-+MEZ in specific cell types. To begin addressing this issue, we have compared the effect of IFN-, MEZ and IFN-+MEZ on 48 and 96 h growth and viability in cells displaying variable induction or no induction of mda-7 after combination treatment (Figures 7 and 8). When administered at 10 ng/ml, MEZ did not inhibit the growth and in some cases even stimulated growth which was cell type specific. However, treatment with 2000 units/ml of IFN-, alone or in combination with MEZ significantly inhibited growth, an effect that was apparent to different extents by 48 h in all the cell types tested (Figures 7 and 8). With many of the cell types, the combination of IFN-+MEZ induced greater growth suppression, and in some cases a decrease in cell viability, than either agent used alone. Growth suppression and decreased viability was apparent in cells displaying inducible mda-7 mRNA expression as well as in cells not showing inducible mda-7 mRNA expression, such as MCF-7 and SK-MEL p53 mt. These results, and studies demonstrating that IFN- decreases growth without inducing mda-7, indicate that induction of mda-7 by IFN-+MEZ is not mandatory for growth suppression and decreasing viability in combination treated cells.

Discussion

Treatment of human melanoma cells with IFN-+MEZ results in profound physiological changes, including an irreversible loss in proliferative potential, suppression of oncogenic potential in athymic nude mice, altered cell surface antigenicity, temporal alterations in gene expression and induction of terminal differentiation (Fisher et al., 1985; Graham et al., 1991; Jiang et al., 1993, 1994, 1995a,b, 2000; Huang et al., 1999a,b; Kang et al., 2001; Leszczyniecka et al., 2001). To obtain insights into this process we have begun to define the spectrum of gene expression changes occurring as a consequence of this combination treatment in human melanoma cells using several molecular approaches. These include, construction of temporally spaced subtracted cDNA libraries from IFN-+MEZ treated HO-1 melanoma cells combined with random clonal isolation, high density microarray analysis of subtracted cDNA clones and reverse Northern hybridization of randomly isolated subtracted cDNA clones (Jiang and Fisher, 1993; Jiang et al., 1994; Huang et al., 1999a,b). In addition, we have applied a new and highly efficient rapid subtraction hybridization, RaSH, protocol to address the question of temporal gene expression changes occurring during the induction of terminal differentiation in human melanoma cells, resulting in the cloning of additional previously identified and novel genes implicated in this process (Jiang et al., 2000; Kang et al., 2001). These studies have proven very informative and are providing a molecular snapshot of genes involved in cancer growth control, survival, apoptosis and differentiation (Huang et al., 1999a,b; Jiang et al., 2000; Leszczyniecka et al., 2001).

A potentially relevant gene in melanoma progression, isolated by subtraction hybridization, is mda-7 (Jiang et al., 1995b). When originally cloned, it was proposed that mda-7 might function as a tumor suppressor gene in the context of melanocyte/melanoma cells, displaying elevated expression in normal melanocytes but decreased expression in primary melanoma cells and a further diminution in expression as melanomas progress to a metastatic state (Jiang et al., 1995b). If was further suggested that expression of mda-7 in the context of normal melanocytes might contribute to the slower growth rate of these cells versus melanoma cells (Jiang et al., 1995b). Two lines of evidence confirm an inverse relationship between growth rate and mda-7 expression in specific cell types. Expression of an inducible mda-7 construct in human melanoma cells decreases growth rate and antisense inhibition of mda-7 expression in human cervical cancer (HeLa) cells engineered to express mda-7 enhances their growth rate (Jiang et al., 1995b, 1996). A hallmark of the terminal differentiation process is growth suppression. When mda-7 is expressed at physiological levels in melanoma cells by DNA transfection growth is suppressed, whereas expression of mda-7 at supraphysiological levels in melanoma cells following viral (Ad.mda-7) infection induces apoptosis (Madireddi et al., 2000c). In contrast, neither of these treatment protocols with mda-7 results in terminal differentiation in human melanoma cells. These observations suggest that additional genes working in combination with mda-7 or operating independently of mda-7 are involved in initiating and maintaining terminal differentiation in human melanoma cells following IFN-+MEZ treatment.

The present study provides additional support for an inverse correlation between mda-7 expression and human melanocyte to melanoma progression. Normal early passage human melanocytes and SV40-immortalized normal human melanocytes express mda-7 de novo as does WM35 cells (an early RGP primary melanoma cell line) (Figure 3). In contrast, only one of nine metastatic melanoma cells displayed endogenous mda-7 expression, and in the one positive melanoma cell line, FO-1, expression was lower than observed in the normal melanocytes or WM35 cells (Figure 3). When treated with IFN-+MEZ for 24 h, mda-7 mRNA expression was elevated in normal early passage human melanocytes, SV40-transformed human melanocytes (FM516-SV) and WM35 cells and expression was induced in eight of the nine metastatic melanoma cells. In general, no correlation was found between the level of induction of mda-7 by IFN-+MEZ and the degree of growth-suppression in the melanocyte/melanoma cell lines, with the possible exceptions of C8161 and SK-MEL p53 mt cells which showed low-level or no induction of mda-7, respectively, and were more resistant to growth suppression and decreased viability than the other metastatic melanoma cell lines displaying elevated levels of mda-7 expression following combination treatment (Figure 7). These findings support the hypothesis that mda-7 may only partially contribute to the growth and differentiation changes observed in human melanoma cells after treatment with IFN-+MEZ. Alternatively, mda-7 may exert effects on cellular physiology that are different in the context of a melanocyte versus a melanoma. This possibility is strengthened by the observation that Ad.mda-7 does not significantly alter growth in normal melanocytes, whereas it induces growth suppression and apoptosis in metastatic human melanoma cells (unpublished data).

Analysis of de novo mda-7 mRNA expression in a panel of normal and cancer cell types demonstrates limited expression in a normal cellular context, such as melanocyte, with little or no expression in cancer cells. The lack of expression in a cancer cell background supports the possibility that mda-7 may function as a cancer growth suppressor and inactivation or decreased expression of this gene may contribute to the cancer phenotype. The lack of de novo expression of mda-7 in most cancer cells could result because of permanent defects in the gene or pathways leading to stable mRNA expression, or alternatively could reflect a lack of expression based on cellular milieu or the physiological state of the cell. It does not appear that the lack of mda-7 expression in cancer cells results from mutations in the mda-7 gene since no changes have been observed in a spectrum of cancer and normal cells (Soo et al., 1999; Mhashilkar et al., 2001). At least a partial answer comes from previous studies in melanoma cells which suggest that the mda-7 promoter is constitutively active in melanoma cells, whereas mda-7 mRNA expression is restricted to melanoma cells treated with IFN-+MEZ and induced to terminally differentiate (Madireddi et al., 2000a,b). These findings suggest that post-transcriptional modifications may contribute to mda-7 mRNA levels in human melanoma cells (Madireddi et al., 2000a). To address the question of cell context specific expression of mda-7 we have evaluated the effect of IFN-+MEZ on mda-7 mRNA expression in a panel of normal and cancer cell types (Figure 4). In addition to induction of mda-7 expression in most melanomas and constitutive expression of mda-7 in normal melanocytes and WM35 RGP cells, treatment with IFN-+MEZ for 24 h induced mda-7 expression in normal human cerebellum cells (NC), one of two human glioblastoma multiforme cell lines (GBM-18), a human nasopharyngeal carcinoma cell line (HONE-1), a human cervical carcinoma cell line (HeLa), one of three human prostate carcinoma cell lines (DU-145), a normal human breast epithelial cell line (HBL-100), two of five human breast carcinoma cell lines (MDA-MB-157 and MDA-MB-231) and a human osteosarcoma cell line (Saos-2) (Figure 4). To determine the functional significance of this mRNA induction, we determined levels of MDA-7 protein following treatment of FM516-SV, HO-1 and DU-145 cells with IFN-+MEZ. For comparison, the same cell types were infected with 1, 10, 50 or 100 pfu/cell of Ad.mda-7, which contains the coding region of mda-7 without the 3' or 5' regions of this gene (Su et al., 1998). Infection with Ad.mda-7 resulted in a dose-dependent induction of both mda-7 mRNA and MDA-7 protein, with the highest expression in DU-145 cells and reduced expression in FM516-SV and HO-1 cells (Figures 5 and 6). In contrast, treatment with IFN-+MEZ efficiently induced mda-7 mRNA, with the level of induction being comparable to low levels of Ad.mda-7 infection (1 pfu/cell), which did not result in readily detectable intracellular MDA-7 protein (Figure 6). These results suggest that the level of endogenous MDA-7 protein resulting following treatment of specific cells with IFN-+MEZ for 24 h is small and may not correlate with the increase or induction in mda-7 mRNA. These results suggest a potential role for post-transcriptional processes in regulating endogenous MDA-7 protein levels in cells. Further studies are also required to determine if there are differences in the stability and/or secretion of endogenously produced MDA-7 protein versus virally generated MDA-7 protein.

Evaluation of mda-7 expression using multiple tissue Northern blots, containing poly(A)⁺ RNA from various tissues, demonstrate that mda-7 expression is restricted to the thymus, spleen and peripheral blood leukocytes (Figure 2). Previous studies document a 9-12-fold elevation in mRNA expression of a gene c49a (a rat homologue of mda-7) during the process of wound healing (Soo et al., 1999). Based on these observations, we evaluated a series of hematopoietic cells, representing erythroid, myeloid and lymphoid lineages, with and without induction of specific differentiation programs for expression of mda-7. No de novo mda-7 expression was detected in HL-60 (human promyelocytic leukemia), K562 (human erythroleukemia), CEM-C7 (human T-cell leukemia) or a TPA-resistant variant of HL-60 (HL534) (Figure 2). Similarly, induction of HL-60 differentiation by treatment with TPA (monocyte/macrophage) or DMSO (granulocyte) or treatment of CEM-C7 cells with TPA did not induce mda-7 expression. However, treatment of K562 cells with TPA, which induces these cells to differentiate into megakaryocytes, resulted in induction of mda-7 mRNA at levels similar to that found in HO-1 cells treated with IFN-+MEZ (Figure 2). If this increase in mda-7 correlates with an elevation of functional MDA-7 protein these results suggest that mda-7 might have a biological role in megakaryocytic differentiation. Differentiation of K562 leukemia cells by TPA treatment to megakaryocytic cells is characterized by an increase in platelet peroxidase positivity, enhancement of thromboxane A2 receptors and increased cell volume and DNA ploidy (reviewed in Alitalo, 1990). There is also an increase in synthesis of platelet derived growth factor (PDGF) and transforming growth factor beta 1 (TGF-1) two of the cytokines that mark the wound-healing process (Chu et al., 1999; Li et al., 1999; Miller, 1999). The molecular function of mda-7 during this program of differentiation is unclear at this time, however, it is possible that mda-7 may have multiple biological roles including an involvement in immune responses and wound healing. Of particular relevance, mda-7 may have an indirect/direct role in platelet formation during erythroid cell differentiation.

At the time of initial isolation and characterization, computational analysis of MDA-7 indicated that it was a unique cDNA, showing no homology to known sequences in the databases or containing easily identifiable structural attributes or motifs at the peptide level (Jiang et al., 1994, 1995b, 1996). Among the reported protein domains present in the conceptually translated peptide sequence was an Interleukin-10 (IL-10) signature. Recent information (Chaiken and Williams, 1996; Kotenko et al., 2001; and unpublished data) strongly suggest that mda-7 belongs to the four-helix bundle family cytokine molecules (Chaiken and Williams, 1996; Kotenko et al., 2001) most related to the IL-10 sub-family (Zhang et al., 2000; Xie et al., 2000; Kotenko et al., 2001). In addition to IL-10, this protein family in humans presently includes IL-19, IL-TIF, AK-155 and IL-20. Although the extent of amino acid homology between members is not extensive (IL-10 and mda-7 share <20% sequence identity), several similarly related molecules have been discovered and from the collectively available information there is now convincing evidence to classify mda-7 as a member of the family (Gallagher et al., 2000; Xie et al., 2000; Zhang et al., 2000; Kotenko et al., 2001). Two features of mda-7 in addition to the presence of an IL-10 family signature and predicted four-helix bundle protein conformation reinforce this idea. These features include a 49 amino acid N-terminal signal peptide classically present in secreted molecules and its location in the human genome on chromosome locus 1q32 where it shows tight linkage to other members of the family, including IL-10, IL-19 and IL-20 comprising, what appears to be a cytokine cluster. It seems from presently documented literature, even for those members of the IL-10 family whose discovery was reported within the last year (IL-19 and IL-20), that each member has a distinct set of functional attributes and tissue distribution. For example, IL-10 has a pleiotropic immunomodulatory effect and is produced in T-cell subsets, monocytes, keratinocytes and activated B-cells (Gallagher et al., 2000; Saito, 2000). IL-19 is monocyte specific (Gallagher et al., 2000) and IL-20 is expressed at low levels in skin and certain other tissues (Blumberg et al., 2001) while mda-7 appears to be primarily restricted to peripheral blood leukocytes, thymus and spleens of normal adult humans at the RNA level (Figure 2). It appears from the presently available information that each protein has distinct non-overlapping biological effects, not entirely unexpected given the limited homology of primary amino acid sequence. Of these effector functions, the transformed cell specific inhibitory activity appears to be a property unique to mda-7. Given its significant structural relatedness to IL-10 family cytokines and clustered location in the IL-10 genomic locus, studies directed toward understanding the cytokine related nature and properties of mda-7 is likely to yield important information pertaining to biological activity.

Materials and methods

Cell cultures and growth assays

A normal SV40 immortalized human foreskin melanocyte cell line, FM516-SV (FM516), was provided by Dr L Diamond (Wistar Institute, PA, USA). WM35, WM278 and WM239 were obtained from Dr M Herlyn (Wistar Institute, PA, USA) (Jiang et al., 1995a). Metastatic FO-1 and HO-1 melanoma cells were described previously (Fisher et al., 1985). C8161 metastatic melanoma cells were obtained from Dr D Welch (University of Pennsylvania) (Jiang et al., 1995a). Dr RS Kerbel provided the MeWo cell line and its reduced metastatic variant 3S5 and highly metastatic variant 70W (Kerbel and Man, 1984; Graham et al., 1991). SK-MEL p53 mt and SK-MEL p53 wt were provided by Dr A Albino (American Health Foundation, NY, USA). HBL-100, MCF-7, T47D, MDA-MB-157, MDA-MB-231, MDA-MB-453, LNCaP, PC-3, DU-145, HeLa, Saos2, MIA PaCa-2, PANC-1, BxPC-3, AsPC-1, T98G and HONE-1 cells were obtained from the American Type Culture Collection. NC, a normal human cerebellum astrocyte cell line and the GBM-18 human glioblastoma multiforme cell line were established in culture from patient-derived samples (Vita et al., 1988; Guarini et al., 1990). Early passage normal human prostate epithelial cells (HuPEC) were obtained from Clonetics Inc. (CA, USA) and cultured using reagents and medium provided by the company. Most cell lines were grown in Dulbecco's modified Eagle's minimum essential medium supplemented with 5 or 10% fetal bovine serum and antibiotics. The pancreatic carcinoma cell lines were grown in RPMI 1640 medium supplemented with 10% fetal bovine serum. All cells were cultured at 37°C in a humidified 5% CO₂ 95% air incubator. Growth assays were performed by seeding cells in complete growth medium at 1´10⁵ cells/35-mm plate and viable cell counts were determined by hemocytometer 48 and 96 h after a medium change without additions (control), 2000 units/ml of IFN-, 10 ng/ml of MEZ or 2000 units/ml of IFN-+10 ng/ml MEZ. Data is presented as average of triplicate samples which varied by 10%. An additional study was performed which varied by 15%.

Virus construction and plaque assays

The recombinant replication-defective Ad.mda-7 virus was created in two steps as described previously (Su et al., 1998). Briefly, the coding region of the mda-7 gene was cloned into a modified Ad expression vector pAd.CMV (Falck-Pedersen et al., 1994). This vector contains, in order, the first 355 bp from the left end of the Ad genome, the cytomegalovirus immediate early promoter, DNA encoding splice donor and acceptor sites, the coding region of the mda-7 cDNA, DNA encoding a poly(A) signal sequence from the globin gene, and ~3 kbp of adenovirus sequence extending from within the E1B coding region. This arrangement allows high-level expression of the cloned sequence by the cytomegalovirus immediate early gene promoter, and appropriate RNA processing (Falck-Pedersen et al., 1994). The recombinant virus was created in vivo in 293 cells (Graham et al., 1977) by homologous recombination between mda-7-containing vector and plasmid JM17, which contains the whole of the Ad genome cloned into a modified version of pBR322 (McGrory et al., 1988). JM17 gives rise to Ad genomes in vivo but they are too large to package. This constraint is relieved by recombination with the vector to create a packageable genome (McGrory et al., 1988), containing the mda-7 gene. The recombinant virus is replication defective in human cells except 293 cells, which express adenovirus E1A and E1B. Following transfection of the two plasmids, infectious virus was recovered, the genomes were analysed to confirm the recombinant structure, and then virus was plaque purified, all by standard procedures (Volkert and Young, 1983).

Northern and Western blotting assays

Multiple-tissue Northern blots (Clontech) of poly(A)⁺ mRNA extracted from different human tissues were hybridized in ExpressHyb solution (Clontech) with the coding region of the mda-7 cDNA. The normal tissue analysed included heart, brain, placenta, lung, liver, skeletal muscle, kidney, pancreas, spleen, thymus, prostate, testis, ovary, small intestine, colon and peripheral blood leukocyte. The mda-7 probe was labeled with -³²P-dCTP using a random primer labeling kit and membranes were washed according to manufacture's protocol (Amersham, IL, USA). Levels of mda-7 and GAPDH mRNA in untreated, IFN-+MEZ (2000 units/ml+10 ng/ml) treated or Ad.mda-7 infected cells were determined by Northern blotting analysis of total cytoplasmic RNA as previously described (Su et al., 1998). In brief, 10 g of total RNA from the different cell types were electrophoresed in 1% agarose gel, transferred to a nylon membrane, and hybridized with the different ³²P-labeled cDNA fragments. The membrane was stripped and hybridized with the indicated probes sequentially. Northern blots were washed in a 0.1% SDS, 1´SSC buffer at room temperature for 30 min followed by washing at 42°C for an additional 30 min in the same buffer. After hybridization, the nylon membranes were washed and exposed for autoradiography.

MDA-7 and elongation factor 1-alpha (EF-1) protein levels were determined by Western blotting as described previously (Su et al., 1995). Cells were grown in 100-cm plates and after appropriate treatment were washed twice with cold PBS and lysed on ice for 30 min in 100 l of cold RIPA buffer [50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.1% SDS, 1% NP40, and 0.5% sodium deoxycholate] with freshly added 0.1 mg/ml phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate, and 1 mg/ml aprotinin. Cell debris were removed by centrifugation at 14 000 g for 10 min at 4°C. Protein concentrations were determined using the Bio-Rad protein assay system (Bio-Rad Laboratories, Richmond, CA, USA). Aliquots of cell extracts containing 20-50 mg of total protein were resolved in 12% SDS-PAGE and transferred to Immobilon-P PVDF membranes (Millipore Corp., Bedford, MA, USA). Filters were blocked for 1 h at room temperature in Blotto A [5% nonfat milk powder in TBS-T: 10 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.05% Tween 20], and then incubated for 1 h at room temperature in Blotto A containing a 1 : 1000 dilution of rabbit anti-MDA-7, polyclonal antibody. After washing in TBS-T buffer (3´5 min, room temperature), filters were incubated for 45 min at room temperature in Blotto A containing a 1 : 10 000 dilution of corresponding peroxidase conjugated anti-rabbit secondary antibody (Amersham, Arlington Heights, IL, USA). After washing in TBS-T, ECL was performed according to the recommendation of the manufacturer.

Isolation of mda-7 genomic clones

Polymerase chain reaction amplification: Diploid human fibroblast DNA (Clontech) was used as a template with human mda-7 gene specific primers, (5'-primer) 5'-ACAAGACATGACTGTGAGGAG-3' and (3'-primer) 5'-AGACTGTTTGAAATGACACAG-3'. The proof reading and high efficiency Advantage Tth DNA polymerase (Clontech) was used in all reactions. The PCR cycling conditions were 95°C/1 min, 60°C/1 min and 72°C/6 min. The reaction was processed for 30 cycles with an additional 72°C/10 min extension performed at the end. The PCR reaction product was analysed by agarose gel electrophoresis and sequencing by the ABI method.

Library screening: A human placental genomic lambda Fix II library (Stratagene, La Jolla, CA, USA) was screened using the human mda-7 cDNA. The probe was labeled with -³²P-dCTP using a random primer labeling kit from Amersham according to manufacture's protocol. Plaque lift filters were hybridized overnight in hybridization buffer (ExpressHyb, Clontech) at 68°C. The filters were washed at 55°C for 20 min, twice in 2´SSC, 0.1% SDS buffer, and once in 0.5´SSC, 0.1% SDS buffer and exposed to X-ray (Kodak) film.

DNA sequencing of mda-7 human genomic clone: The isolated phage DNA clones were mapped by restriction enzyme analysis using standard procedures (Sambrook et al., 1989). PCR generated products were cloned into pBluescript and sequenced using universal primers with an ABI automatic sequencer model 372 (Applied Biosystems). DNA and protein sequence alignment were determined using the GCG software package (Wisconsin Package Version 10.0, Genetics Computer Group (GCG), Madison, WI, USA). The percentage of DNA and Protein sequence homology was determined using the GenBank database and GCG homology algorithms including BLAST, FATSA, BESTFIT and GAP.

Chromosomal mapping of the human mda-7 gene

For chromosomal mapping studies, oligonucleotides for generating PCR products were designed using the computer program Oligo 4.0 (National Biosciences) based on the mda-7 sequence (Jiang et al., 1995b). The two primers used to amplify a 129 bp human mda-7 specific gene product were: MDA7F, 5'GGTTTGTTCCCTGTGTCATT3' and MDA7R, 5'GCGCTGCTTAAAGAATGACT3'. These primer sets were used with PCR to determine the presence or absence of the mda-7 locus in a panel of 19 rodent-human hybrids. PCR reactions were conducted in a final volume of 12.5 l with 100 ng of template, 20 ng primers, 10 mM tris-HCL pH 8.3, 50 mM KCl, 0.1 mg/ml gelatin, 15 mM MgCl₂, 200 M dNTPs and 0.5 U Taq polymerase. Amplifications were performed in a Perkin-Elmer Cetus 9600 thermal cycler for 30 cycles at 94°C for 30 s, 58°C for 30 s and 72°C for 30 s. The PCR products were visualized in ethidium bromide stained 1.5% agarose gels. The amplification product was purified with Qiagen PCR purification kit, and 1 ng of DNA and 20 ng specific primer used with the Taq Dye Deoxy Terminator Cycle Sequencing Kit (ABI). The reaction products were electrophoresed and recorded on the 377 DNA sequencer (ABI).

Acknowledgements

This study was supported in part by National Institutes of Health grant CA35675, CA80826, Columbia Skin Diseases Research Center (P30-AR44535), the Samuel Waxman Cancer Research Foundation, the Chernow Endowment and Introgen Therapeutics, Inc. MT Madireddi was supported by a fellowship award from the Army Department of Defense Initiative on Breast Cancer (DAMD17-98-1-8053). PB Fisher is the Michael and Stella Chernow Urological Cancer Research Scientist in the Departments of Neurosurgery, Pathology and Urology. We thank Drs Theresa Druck and Kay Hubner for assistance with the chromosomal mapping of mda-7.

Alitalo R. (1990). Leuk. Res. 14, 501-514. MEDLINE

Armstrong BK, Kricker A. (1994). Cancer Surv. 19/20, 219-240.

Bishop JM. (1991). Cell 64, 235-248. MEDLINE

Blumberg H, Conklin D, Xu W, Grossmann A, Brender T, Carollo S, Eagan M, Foster D, Haldeman B, Hammond A, Haughen H, Jelinek L, Kelly J, Madden K, Maurer M, Parrish-Novak J, Prunkard D, Sexson S, Sprecher C, Waggie K, West J, Whitmore T, Yao L, Kuechle M, Dale B, Chandrasekher Y. (2001). Cell 104, 9-19. MEDLINE

Breathnach R, Benoist C, O'Hare K, Gannon F, Chambon P. (1978). Proc. Natl. Acad. Sci. USA 75, 4853-4857. MEDLINE

Breathnach R, Chambon P. (1981). Annu. Rev. Biochem. 50, 349-383. MEDLINE

Chaiken IM, Williams WV. (1996). Trends Biotechnol. 14, 369-375. MEDLINE

Chu X, Newman J, Park B, Nares S, Ordonez G, Iacopino AM. (1999). Cell Tissue Res. 296, 331-337. MEDLINE

Clark W. (1991). Br. J. Cancer 64, 631-644. MEDLINE

DeRisi J, Penland L, Brown PO, Bittner ML, Meltzer PS, Ray M, Chen Y, Su YA, Trent JM. (1996). Nat. Genet. 14, 457-460. MEDLINE

Falck-Pedersen E, Heinflink M, Alvira M, Nussenzveig DR, Gershengorn MC. (1994). Mol. Pharmacol. 45, 684-689. MEDLINE

Fisher PB. (1984). Tumor Promotion and Cocarcinogenesis In Vitro, Mechanisms of Tumor Promotion. Slaga TJ (ed). CRC Press Inc.: Boca Raton, Florida, pp. 57-83.

Fisher PB, Prignoli DR, Hermo Jr H, Weinstein IB, Pestka S. (1985). J. Interferon Res. 5, 11-22. MEDLINE

Gallagher G, Dickensheets H, Eskdale J, Izotova LS, Mirochnitchenko OV, Peat JD, Vasquez N, Pestka S, Donnelly RP, Kotenko SV. (2000). Genes Immun. 1, 442-450. MEDLINE

Graham FL, Smiley J, Russell WC, Nairn R. (1977). J. Gen. Virol. 36, 59-72. MEDLINE

Graham GM, Guarini L, Moulton TA, Datta S, Ferrone S, Giacomini P, Kerbel RS, Fisher PB. (1991). Cancer Immunol. Immunotherap. 32, 382-390.

Guarini L, Temponi M, Bruce JN, Bollon AP, Duigou GJ, Moulton TA, Ferrone S, Fisher PB. (1990). Intl. J. Cancer 46, 1041-1047.

Herlyn M, Berking C, Li G, Satyamoorthy K. (2000). Melanoma Res. 10, 303-312. MEDLINE

Huang F, Adelman J, Jiang H, Goldstein NI, Fisher PB. (1999a). Oncogene 18, 3546-3552.

Huang F, Adelman J, Jiang H, Goldstein NI, Fisher PB. (1999b). Gene 236, 125-131. Article MEDLINE

Jiang H, Fisher PB. (1993). Mol. Cell. Differ. 1, 285-299.

Jiang H, Kang D-C, Alexandre D, Fisher PB. (2000). Proc. Natl. Acad. Sci. USA 93, 9160-9165.

Jiang H, Lin J, Fisher PB. (1994). Mol. Cell. Differ. 2, 221-239.

Jiang H, Lin JJ, Su Z-Z, Goldstein NI, Fisher PB. (1995b). Oncogene 11, 2477-2486. MEDLINE

Jiang H, Lin J, Su Z-Z, Kerbel RS, Herlyn M, Weissman RB, Welch D, Fisher PB. (1995a). Oncogene 10, 1855-1864. MEDLINE

Jiang H, Su Z-Z, Boyd J, Fisher PB. (1993). Mol. Cell. Differ. 1, 41-66.

Jiang H, Su Z-Z, Lin JJ, Goldstein NI, Young CSH, Fisher PB. (1996). Proc. Natl. Acad. Sci. USA 93, 9160-9165. MEDLINE

Kang D-C, Jiang H, Wu Q, Pestka S, Fisher PB. (2001). Gene 267, 233-242. Article MEDLINE

Kang D-C, LaFrance R, Su Z-Z, Fisher PB. (1998). Proc. Natl. Acad. Sci. USA 95, 13788-13793. MEDLINE

Kerbel RS, Man MS. (1984). Invasion Metastasis 4, Suppl 1 31-43. MEDLINE

Kotenko SV, Izotova LS, Mirochnitchenko OV, Esterova E, Dickensheets H, Donnelly RP, Pestka S. (2001). J. Biol. Chem. 276, 2725-2732. MEDLINE

Leszczyniecka M, Roberts T, Dent P, Grant S, Fisher PB. (2001). Pharmacol. Therapeut in press.

Li DQ, Lee SB, Tseng SC. (1999). Curr. Eye Res. 19, 154-161. MEDLINE

Madireddi MT, Dent P, Fisher PB. (2000a). Oncogene 19, 1362-1368.

Madireddi MT, Dent P, Fisher PB. (2000b). J. Cell. Physiol. 185, 36-46.

Madireddi MT, Su Z-Z, Young CSH, Goldstein NI, Fisher PB. (2000c). Adv. Exptl. Med. Biol. 465, 239-261.

McGrory WJ, Bautista DS, Graham FL. (1988). Virology 163, 614-617. MEDLINE

Mhashilkar AB, Schrock RD, Hindi M, Liao J, Sieger K, Kourouma F, Zou-Yang XH, Onishi E, Takh O, Vedvick TS, Fanger G, Stewart L, Watson GJ, Snary D, Fisher PB, Saeki T, Roth JA, Ramesh R, Chada S. (2001). Mol. Med. 7, 271-282. MEDLINE

Miller MS. (1999). J. Foot Ankle Surg. 38, 227-231. MEDLINE

Ross DT, Scherf U, Eisen MB, Perou CM, Rees C, Spellman P, Iyer V, Jeffrey SS, Van de Rijn M, Waltham M, Pergamenschikov A, Lee JC, Lashkari D, Shalon D, Myers TG, Weinstein JN, Botstein D, Brown PO. (2000). Nat. Genet. 24, 227-235. Article MEDLINE

Saeki T, Mhashilkar A, Chada S, Branch C, Roth JA, Ramesh R. (2000). Gene Therapy 7, 2051-2057. MEDLINE

Saito I. (2000). Crit. Rev. Immunol. 20, 153-165. MEDLINE

Sambrook J, Fritsch EF, Maniatis T. (1989). Cold Spring Harbor Laboratory Press, 2nd ed. Cold Spring Harbor, NY,

Schena M, Shalon D, Heller R, Chai A, Brown PO, Davis RW. (1996). Proc. Natl. Acad. Sci. USA 93, 10614-10619. Article MEDLINE

Soo C, Shaw WW, Freymiller E, Longaker MT, Bertolami CN, Chiu R, Tieu A, Ting K. (1999). J. Cell. Biochem. 74, 1-10. MEDLINE

Su Z-Z, Goldstein NI, Jiang H, Wang M-N, Duigou GJ, Young CSH, Fisher PB. (1999). Proc. Natl. Acad. Sci. USA 96, 15115-15120. MEDLINE

Su Z-Z, Madireddi MT, Lin JJ, Young CSH, Kitada S, Reed JC, Goldstein NI, Fisher PB. (1998). Proc. Natl. Acad. Sci. USA 95, 14400-14405. MEDLINE

Su Z-Z, Shi Y, Fisher PB. (1997). Proc. Natl. Acad. Sci. USA 94, 9125-9130. MEDLINE

Su Z-Z, Yemul S, Estabrook A, Zimmer SG, Friedman RM, Fisher PB. (1995). Intl. J. Oncology 7, 1279-1284.

Tonetti DA, Horio M, Collart FR, Huberman E. (1992). Cell Growth Differ. 3, 739-745. MEDLINE

Vita JR, Edwalds GM, Gorey T, Housepian EM, Fetell MR, Guarini L, Langer JA, Fisher PB. (1988). Anticancer Res. 8, 297-302. MEDLINE

Vogelstein B, Kinzler KW. (1991). Trends Genet. 9, 138-141.

Volkert FC, Young CSH. (1983). Virology 125, 175-193. MEDLINE

Waxman S. (1996). Differentiation Therapy. Waxman S (ed). Serono Symposium Publications, Rome, Italy, pp. 1-528.

Xie MH, Aggarwal S, Ho WH, Foster J, Zhang Z, Stinson J, Wood WI, Goddard AD, Gurney AL. (2000). J. Biol. Chem. 275, 31335-31339. MEDLINE

Zhang R, Tan Z, Liang P. (2000). J. Biol. Chem. 275, 24436-24443. MEDLINE

Figure 1 Human mda-7 genomic structure. The nucleotide sequence of all exons including non-coding 5' and 3' UTR sequences and coding exons are shown in bold type. Sequences flanking the exon/intron junction including the splice consensus GT-AG sequences are indicated in normal type and shadow font respectively. Length of each intron and other landmark sequences including transcription start site (+1), putative TATA regulatory sequence, translation initiator codon (ATG) and polyadenylation consensus sequence are highlighted at the respective positions

Figure 2 Expression of mda-7 message in the human immune system and human leukemic cells. (a) Human multiple tissue Northern blot consisting of poly(A)⁺ mRNA from different tissues shows tissue specific expression of mda-7. The mRNAs immobilized on the blot are from spleen (1), thymus (2), prostate (3), testis (4), ovary (5), small intestine (6), colon (7) and peripheral blood leukocytes (8). (b) Mda-7 expression in leukemic cells induced to undergo differentiation. HL-60 (lanes 1-3, human promyelocytic leukemia) uninduced (lane 1) or induced to differentiate by TPA (lane 2) or DMSO (lane 3), K562 (erythroleukemia) uninduced (lane 4), or induced to differentiate by TPA (lane 5), CEM-C7 (human T-cell leukemia) uninduced (lane 6) or induced to differentiate by TPA (lane 7), HL 534 (lane 8) and HO-1 (human melanoma) treated with IFN-+MEZ

Figure 3 Expression of mda-7 message in human melanocyte and melanoma cell lines with and without treatment with inducer. The indicated cell types were grown for 24 h in the presence of 2000 units/ml of IFN-+10 ng/ml of MEZ, total RNA was isolated and analysed by Northern blotting and probing with a ³²P-labeled mda-7 cDNA probe. The blot was stripped and then reprobed with a ³²P-labeled GAPDH probe. FM516 is an SV40-immortalized normal melanocyte cell line (FM516-SV); WM35 is an early RGP primary melanoma cell line; WM278 is an early VGP primary melanoma cell line; WM239, FO-1, C8161, MeWo, 3S5 and 70W are metastatic melanoma cell lines; SK-MEL p53 mt is a metastatic melanoma cell line with a confirmed mutant p53 genotype; and SK-MEL p53 wt is a metastatic melanoma cell line with a confirmed wild-type p53 genotype

Figure 4 De novo and inducible expression of mda-7 mRNA in normal and human cancer cell lines. (a) Expression of mda-7 in normal early passage human foreskin melanocytes, early passage human prostate epithelial cells HuPEC and prostate carcinoma cell lines, PC-3, LNCaP and DU-145, grown for 24 h in the absence (-) or presence (+) of IFN- (2000 units/ml) plus MEZ (10 ng/ml). (b) Expression of mda-7 in a normal human breast epithelial cell line (HBL-100) and various breast carcinoma cell lines (MCF7, T47D, MDA-MB-157, MDA-MB-231 and MDA-MB-453) grown for 24 h in the absence (-) or presence (+) of IFN- (2000 units/ml) plus MEZ (10 ng/ml). (c) Expression of mda-7 in diverse human cell lines, including normal cerebellum (NC), glioblastoma multiforme (T98G or GBM-18), colon carcinoma (SW613), osteosarcoma (Saos2), cervical carcinoma (HeLa) and nasopharyngeal carcinoma (HONE-1), grown for 24 h in the absence (-) or presence (+) of IFN- (2000 units/ml) plus MEZ (10 ng/ml)

Figure 5 Expression of mda-7 mRNA in DU-145, HO-1 and FM516 cells with and without treatment with IFN-+MEZ or infection with Ad.mda-7. (a) Effect of treatment with IFN-+MEZ or infection with Ad.vec or Ad.mda-7 on mda-7 and GAPDH mRNA. (b) Effect of treatment with IFN-+MEZ or infection with 1 pfu/cell of Ad.mda-7 on mda-7 and GAPDH mRNA. The indicated cell type was untreated (Control), treated with 2000 units/ml of IFN- plus 10 ng/ml of MEZ or infected with an Ad.vec (100 pfu/cell) or 1, 10, 50 or 100 pfu/cell of Ad.mda-7 for 24 h. Total RNA was isolated and analysed by Northern blotting for mda-7 and GAPDH mRNA expression

Figure 6 Expression of MDA-7 protein in DU-145, HO-1 and FM516 cells with and without treatment with IFN-+MEZ or infection with Ad.mda-7. The indicated cell type was treated as in Figure 5 for 24 h and levels of MDA-7 and EF-1 proteins in total cell lysates were determined by Western blotting using the appropriate polyclonal or monoclonal antibody, respectively

Figure 7 Effect of treatment with IFN-, MEZ or IFN-+MEZ on the growth of melanocyte/melanoma cell lines. The indicated cell type was treated for 48 h (a) or 96 h (b) with IFN- (2000 units/ml), MEZ (10 ng/ml) or a combination of IFN-+MEZ (2000 units/ml+10 ng/ml) and cell numbers were determined. Triplicate samples varied by 10% and a replicate experiment varied by 15%. :IFN-; :MEZ; :IFN-+MEZ

Figure 8 Effect of treatment with IFN-, MEZ or IFN-+MEZ on the growth of various normal and human cancer cell lines. The indicated cell type was treated for 48 h (a) or 96 h (b) as in Figure 7 and cell numbers were determined. Triplicate samples varied by 10% and a replicate experiment varied by 15%. :IFN-; :MEZ; :IFN-+MEZ