The melanoma differentiation association gene-7 (mda-7) is a novel tumor suppressor gene of which the protein expression decreases to nearly undetectable levels in metastatic melanoma. In contrast, expression of inducible nitric oxide synthase (iNOS) is increased in advanced stages of melanoma, and iNOS expression has been proposed as a potential prognostic marker in this disease. Thus, expression of these molecules in the same tumor appears to exhibit reciprocal characteristics. We hypothesize that the relative ratios of these melanoma progression molecules may define either tumor progression or tumor suppression in human melanoma. The first goal of this study was to determine whether MDA-7 expression in melanoma negatively correlates with iNOS expression. The second goal was to determine whether iNOS expression could be regulated by MDA-7 expression in melanoma cells. Expression of MDA-7 and iNOS proteins were evaluated by immunohistochemistry in a total of 81 tumor samples: 38 primary melanomas and 43 metastatic melanomas. Evaluation of these melanoma patient samples showed that expression of MDA-7 and iNOS exhibits a significant negative correlation (P = 0.03). In vitro studies revealed that iNOS expression in melanoma cell lines is lost in a dose-dependent fashion after treatment with an adenoviral vector encoding the mda-7 gene (Ad-mda7) or with rhMDA-7 protein, demonstrating that MDA-7 down-regulates iNOS expression. Furthermore, we demonstrate that the STAT-3-modulated expression of IFN regulatory factors 1 and 2 is regulated by MDA-7, which may alter iNOS gene expression. These studies demonstrate that expression of the MDA-7 tumor suppressor can negatively regulate iNOS expression in malignant melanoma cell lines.

Despite the decrease in overall cancer incidence and mortality rates, cancer remains a major public health problem in the United States. In the year 2002, it is estimated that ∼1,284,900 new cases of invasive cancer will be diagnosed in the United States, and 555,500 Americans will die from cancer (1). Melanoma is one of the most prevalent cancers, having the fifth highest cancer incidence in males and sixth highest incidence in females. It has been predicted that in 2002, a total of 34,300 new cases of melanoma will be diagnosed, and an estimated 7,400 melanoma-related deaths will occur.

NO3 is an important bioactive agent and signaling molecule that regulates a variety of biological mechanisms including vasodilatation, neurotransmission, and host defense. More recently, NO has been proposed to play a critical role in cancer progression. NO is synthesized from l-arginine by a group of enzymes termed NOS, of which three isoforms have been described. Calcium-dependent isoforms were first found in neuronal cells (nNOS or NOS1) then in endothelial cells (eNOS or NOS3). They produce very small amounts (pm) of NO for short periods. In contrast, the macrophage-type calcium-independent isoform, iNOS (NOS2), expressed on stimulation by inflammatory cytokines or bacterial lipopolysaccharide, produces high (nm) amounts of NO for extended periods of time in macrophages (2). In human cancers, the actions of the lower and apparently noninduced constitutive expression of iNOS are implicated in the growth and metastasis of solid tumors. However, NO has been reported to have both tumor promoting and reducing effects, which are probably dependent on the local concentration of the molecule as well as other interacting molecules in the tumor environment. The presence of NO and NOS has been studied in various human tumors including breast (35), pancreas (68), lung (9, 10), prostate (11, 12), colorectal (1315), renal cell cancer (16), stomach (1719), bladder (20), head and neck tumors (21, 22), and melanomas (2325). Many of these studies have shown that expression of NOS correlates with a high grade of tumor. Current data suggest that the process of melanoma tumorigenesis is enhanced by constitutive expression of the iNOS, the enzyme responsible for NO production in response to cytokines.

mda-7 was first identified in human melanoma cell lines induced to differentiate with IFN-β and mezerein (26). The mda-7 cDNA encodes a novel, evolutionarily conserved protein of 206 amino acids with a predicted size of Mr 23,800 (26). Additional studies showed that mda-7 mRNA was expressed in normal melanocytes and early stages of melanoma, but was lost during melanoma progression (27). These observations suggest that mda-7 is a novel tumor suppressor gene of which the expression must be inhibited for tumor progression. The same group has shown that elevated expression of mda-7 suppressed cancer cell growth and selectively induced apoptosis in human breast cancer cells. Jiang et al. (28) reported that mda-7 is a potent growth-suppressing gene in cancer cells of diverse origins. Growth inhibition and apoptosis induction by elevated expression of mda-7 is more effective in cancer cells than in normal cells. Furthermore, Su et al. (29) investigated the mechanism by which mda-7 suppressed cell growth in a breast cancer model. They reported that ectopic expression of mda-7 in breast cancer cell lines MCF-7 and T47D induced apoptosis without an effect on normal cells. Western blotting of lysates from cells infected with Ad-mda7 showed an up-regulation of the proapoptotic protein BAX in MCF-7 and T47D cells, but not in normal cells. Similar findings of the proapoptotic effects of overexpressed mda-7 have been reported recently in melanoma cell lines (30). As observed in other cell types, apoptosis was not induced in normal immortalized melanocytes.

Examination of the molecular pathways utilized by MDA-7 for the induction of apoptosis is a particularly active area of research. Infection of melanoma cells with Ad-mda7 leads to activation of the p38 mitogen-activated protein kinase pathway, resulting in expression of members of the growth-arrest and DNA damage-inducible gene family, some of which are involved in apoptotic processes (31). It has also been shown that Ad-mda7 infected lung cancer cells are induced to express double-stranded RNA-dependent protein kinase in a dose-dependent fashion (32). These findings are of considerable interest, as protein kinase is involved in a number of pathways resulting in cell growth inhibition and apoptosis.

Recent data from our group have provided the first correlative evidence in humans that the MDA-7 protein has the characteristics of a tumor suppressor. Our studies in melanoma have shown that, in keeping with its potential role as a tumor suppressor, MDA-7 expression is decreased in more invasive aspects of melanomas, with nearly undetectable levels in metastatic disease (33). Moreover, in a follow-up study we observed significant decreases in MDA-7-positive cell number and intensity scores when comparing the epidermal to the deep invasive portions of primary tumors, and comparing primary tumors to paired metastases (34).

Here, we hypothesize that iNOS levels in a given melanoma tumor should correlate negatively with MDA-7 levels. We additionally hypothesize that iNOS can be regulated by local MDA-7 production in melanoma, supporting either tumor progression or tumor suppression. Altering the balance between the tumor-progressive iNOS and tumor-suppressive MDA-7 may open new avenues to controlling tumor progression in melanoma and potentially other tumor types.

Patient Samples.

The tumor samples used in this study consisted of primary cutaneous melanomas and melanoma metastases from various sites. Formalin-fixed, paraffin-embedded sections of melanoma tumors were obtained from the Melanoma and Skin Cancer Core Laboratory of the M. D. Anderson Cancer Center.

Cell Culture.

Metastatic melanoma cell lines, A375 and A375.S2, were obtained from American Type Culture Collection (Rockville, MD). Radial growth phase and vertical growth phase melanoma cell lines, WM35, WM793, and their more invasive subclones were provided by Dr. Robert Kerbel (Sunnybrook Health Science Center, Toronto, Ontario, Canada). The highly metastatic melanoma cell line MeWo was provided by Dr. David Menter (M. D. Anderson Cancer Center). Melanoma cell lines used in this study were maintained in RPMI 1640 (Life Technologies, Inc., Grand Island, NY) supplemented with 10% fetal bovine serum (Life Technologies, Inc.), 100 units/ml penicillin, 100 μg/ml streptomycin, 2 mml-glutamine, and HEPES buffer (Life Technologies, Inc.). Cells were either treated with purified MDA-7 at 1–20 ng/ml, or infected with Ad-mda7 or control Ad-luc for in vitro studies.

Purification of Human MDA-7.

The full-length cDNA of mda-7 was cloned into the pCEP4 FLAG vector (Invitrogen, San Diego, CA) containing the CMV promoter. The plasmid was transfected into HEK 293 cells, and stable subclones were isolated using hygromycin (0.4 μg/ml). Supernatant containing the secreted MDA-7 was supplemented with protease inhibitors (1 μg/ml leupeptin, 1 μg/ml pepstatin, and 0.5 mm phenylmethylsulfonyl fluoride) and 0.05% sodium azide, and was concentrated 10-fold with an Amicon stirred cell (Amicon, Beverly, MA) on an YM10 membrane. Ten-ml aliquots of concentrated supernatant were separated over an S200 Superdex prep grade column (Amersham Pharmacia, Piscataway, NJ) in 1× PBS (pH 7.4), and fractions identified to contain MDA-7 by Western blot and ELISA were pooled. After buffer exchange on an Amicon stirred cell to 50 mm 4-morpholinepropanesulfonic acid (pH 6), a second purification step was performed using a Bio-Rad S column. Column conditions consisted of a 0–90-mm NaCl gradient, a 5-min hold at 90 mm NaCl, a 30-min 90–250-mm gradient at 1 ml/min, and a 5-min hold at 250 mm NaCl. The entire purification was conducted at 4°C, and MDA-7 was identified using ELISA and Western blotting procedures. The final samples contained at least 300 ng/ml MDA-7 as determined by ELISA. Individual lots of partially purified MDA-7 were tested for endotoxin using the QCL 1000 quantitative chromogenic LAL kit (BioWhittaker, Walkersville, MD).

Gene Transfer.

Replication-deficient human type 5 adenovirus (Ad5) carrying the mda-7 gene was obtained from Introgen Therapeutics (Houston, TX). The mda-7 gene was linked to an internal CMV-IE promoter and followed by SV40 polyadenylation. Ad-Luc and Ad-CMV polyadenylation (luciferase and empty vector, respectively), were used as control viruses.

Cells were plated 1 day before infection. Melanoma cells were infected with adenoviral vectors (Ad-mda7 or Ad-luc) using 1000–5000 viral particles per cell. Experimental conditions were optimized to achieve MDA-7 protein expression by >70% of cells, based on results of immunohistochemical staining.

Reagents.

Anti-iNOS mouse monoclonal antibody (Transduction Laboratories, Lexington, KY) was used for iNOS immunohistochemistry and confirmed as being cross-reactive between species. Affinity-purified polyclonal rabbit antibodies to MDA-7 were provided by Introgen Therapeutics. IRF-1 and IRF-2 polyclonal antibodies were purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, CA). Phospho-Stat1 (Tyr701) and Phospho-Stat3 (Tyr705) antibodies were obtained from Cell Signaling Tech. (Beverly, MA). Preimmune normal mouse IgG (Vector Laboratories, Burlingame, CA) was used as a negative control. Antivimentin antibody (BioGenex Laboratories, San Ramon, CA) was used as a positive control for all of the melanoma staining.

Immunohistochemistry.

Immunohistochemical labeling was performed on 10% formalin-fixed, paraffin-embedded melanoma tissue, cut 4–6-μm thick. Sections were placed on silanized slides (Histology Control Systems, Glen Head, NY), deparaffinized in xylene, and rehydrated in descending grades (from 100 to 85%) of ethanol. To enhance the immunostaining and restore the maximal antigenicity of cytokines, sections then were placed in antigen unmasking solution (Vector Laboratories) and microwaved intermittently for up to 10 min to maintain a boiling temperature. After the slides were cooled at room temperature for 30 min, they were washed in distilled water and PBS. After this initial preparation, the slides were removed from PBS and covered with 3% H2O2 (Sigma Chemical Co., St. Louis, MO) in methanol to block endogenous peroxidase activity. All of the incubations were carried out at room temperature in a humidified covered slide chamber. The slides were washed in PBS before incubation in PBS containing 0.05% Triton X-100 (Sigma Chemical Co.) for 15 min to permeabilize the cells. An avidin-biotin-peroxidase complex kit (Vectastain; Vector Laboratories) was then used to detect staining. After the slides were incubated for 30 min with the blocking serum, the primary antibodies at various dilutions (1:100 to 1:200) were added, and the slides were incubated for 60 min at room temperature. The slides were then washed, incubated for 30 min with secondary biotinylated antibody, washed again, and then incubated for 30 min with the avidin-biotin-peroxidase complex reagent. After the slides were washed in PBS, the immunostaining was developed with the use of 3-amino-9-ethylcarbazole as a chromagen for 15 min. Slides were counterstained with hematoxylin (Vector Laboratories) and mounted with Aqua-Mount (Lerner Laboratories, Pittsburgh, PA). For each sample, vimentin and isotype-matched control IgG served as positive and negative primary antibody controls, respectively. The specificity and sensitivity of these antibodies have been published previously (23, 30) All of the tissue samples from a given patient were immunolabeled in the same experiment.

Immunohistochemistry Scoring.

Immunolabeling was scored separately for two variables: first for number of positive cells, second for the overall intensity of immunoreactivity of the positive cells. Scoring for number of positive cells was defined as follows: (0) is for <5% positive cells; (1) is for 5–50% of positive cells; (2) is for 50–90% of positive cells; and finally, (3) is for >90% of positive cells. Intensity scoring was defined as follows: (0) is for no staining; (1) is for light staining; (2) is for moderate staining; and (3) is for intense staining. The slides were interpreted by two independent readers.

Immunoblotting Assays.

Two × 106 cultured melanoma cell lines were rinsed twice in ice-cold PBS and lysed in 60 μl of lysis buffer [25 mm Tris, 140 mm NaCl, and 1% NP40 (pH 7.5)] containing 5 mm EDTA, 0.2 mm orthovanadate, 10 mm NaF, leupeptin, aprotinin, and phenylmethylsulfonyl fluoride for 10 min on ice. Equal amounts of total protein (measured with DC Protein Assay Reagent; Bio-Rad Labs, Hercules, CA) were loaded on a standard 10% SDS polyacrylamide gel, and fractionated proteins were electroblotted onto a nitrocellulose membrane. Nitrocellulose membranes were blocked for 1 h at room temperature using 5% dry milk in 1× PBS and washed three times for 5 min each in PBS containing 0.05% Tween 20 at room temperature. The membranes were incubated overnight at 4°C in a sealed bag with a 1:2000 dilution of IRF-1 and IRF-2 polyclonal antibodies in 10 ml of 5% dry milk/0.1% Tween 20 in 1× PBS. The membranes were washed three times for 5 min each in PBS containing 0.05% Tween 20, and then incubated with peroxidase-conjugated antirabbit IgG secondary antibody (Transduction Laboratories) at 1:2000 dilution in PBS with 5% dry milk and 0.1% Tween 20 for 45 min at room temperature. The blots were visualized using an enhanced chemiluminescence detection kit (Amersham, Arlington Heights, IL).

Statistical Analysis.

Means and SDs for the iNOS and MDA-7 variables were computed. To investigate the association between iNOS and MDA-7 count and intensity measurements, the test for absence of correlation was performed using the Kendall τ-b test (35).

Melanoma Tumor MDA-7 Expression Negatively Correlates with Tumor iNOS Expression.

To determine whether MDA-7 expression correlated inversely with iNOS expression in the same tumor, we performed immunohistochemical analyses on sequential paraffin-embedded malignant melanoma tumor sections. Thirty-eight primary melanomas and 43 metastases (total of 81 tumor samples) were analyzed in these experiments. After immunostaining with anti-MDA-7 polyclonal antibody and anti-iNOS monoclonal antibody, sample immunoreactivity based on numbers of positive cells and staining intensity was analyzed. A direct comparison between the numbers of MDA-7-staining cells and iNOS-staining cells revealed a negative association. Fig. 1A demonstrates a significant inverse correlation between numbers of cells staining positively for iNOS and for MDA-7 (correlation coefficient = −0.209, P < 0.05, Kendall τ-b test). Similarly, the data were analyzed for an association between iNOS and MDA-7 by comparing the intensity of staining. Fig. 1B shows an inverse correlation between iNOS intensity and MDA-7 intensity (correlation coefficient = −0.201, P < 0.05, Kendall τ-b test), reflecting a significant decrease in the average iNOS intensity as MDA-7 intensity increases.

Fig. 2 demonstrates the negative association between iNOS and MDA-7 expression in sequential sections of a primary melanoma and sequential sections of a metastasis derived from it. Primary melanoma tumor MDA-7 immunoreactivity in a paired sample displays intense cytoplasmic immunolabeling (Fig. 2A), whereas both the lymph node metastasis (Fig. 2B) and brain metastasis (Fig. 2C) are negative. In contrast, primary melanoma tumor iNOS is absent (Fig. 2D), whereas the lymph node metastasis (Fig. 2E) and brain metastasis (Fig. 2F) are strongly immunolabeled.

Ad-mda 7 and rhMDA-7 Down-Regulate iNOS Expression in Human Melanoma Cell Lines.

The inverse expression of MDA-7 and iNOS demonstrated by immunohistochemistry suggested a potential cause/effect relationship. Hence, we performed a series of in vitro experiments to examine the possible modulation of iNOS expression by MDA-7. First, we infected melanoma cell lines, A375, MeWo, WM35, and WM793 with Ad-mda-7 (500, 1000, and 2000 viral particles per cell) or with Ad-luc (1000 viral particles per cell). At baseline, these melanoma cell lines express high levels of iNOS and are negative for MDA-7. Forty-eight h after vector treatment, the cells were collected, and cytospins were prepared to analyze iNOS expression (Fig. 3). Ad-mda7 at 1000 and 2000 viral particles per cell completely down-regulated expression of iNOS by 48 h (Fig. 3, C and D), whereas Ad-luc infection had no effect (Fig. 3A). The dose of Ad-mda7 vector that inhibited iNOS expression did not appear to result in significant cell death during this short incubation. These experimental results indicated that mda7 gene transfer expression specifically blocked iNOS expression in melanoma cells.

It has been shown previously that MDA-7 is secreted by Ad-mda7-infected melanoma cells (30, 36). To address whether secreted MDA-7 might also contribute to iNOS regulation, we incubated the melanoma cell lines with 0, 5, or 20 ng/ml of rhMDA-7 and stained for iNOS expression (Fig. 4). rhMDA-7 at a concentration of 20 ng/ml resulted in clear down-regulation of iNOS expression by 48 h in A375 melanoma cells (Fig. 4C).

Ad-mda7 and rhMDA-7 Induce Up-Regulation and Phosphorylation of STAT3.

rhMDA-7 protein treatment of melanoma cells resulted in potent down-regulation of iNOS expression, suggesting that MDA-7 may be functioning via a receptor-mediated pathway. It has been shown recently that MDA-7 can bind and signal through the IL-20 and IL-22 receptors. Thus, we predicted that the IL-20 and/or IL-22 receptor signal transduction pathways, both of which are class II cytokine receptors that involve STAT activation, would be active in melanoma cells exposed to MDA-7. We initially studied both STAT1 and STAT3 phosphorylation in melanoma samples incubated with rhMDA-7. Although we did not observe an alteration in STAT1 phosphorylation, we consistently detected up-regulation of STAT3 phosphorylation in MDA-7 treated MeWo cells compared with untreated cells (Fig. 5). Similar findings were observed in the melanoma cell lines A375, WM35, and A375.S2, and in peripheral blood mononuclear cells from a healthy donor (data not shown). We first observed increased STAT3 expression in the cytoplasm of MeWo cells treated with 20 ng/ml rhMDA7 (Fig. 5C). Moreover, we observed staining for phospho-STAT3 in the nuclei of these rhMDA7-treated melanoma cells (Fig. 5D) 

MDA-7 Modulates IRF-1 and IRF-2 Expression in Melanoma Cells.

On the basis of our findings of induction of STAT3 phosphorylation after exposure to MDA-7, we next focused on downstream targets of STAT proteins. Two of these targets, IRF-1 and IRF-2, oppose each other in their activities in tumor cells. Of note, IRF-1 induces iNOS gene expression (37, 38). To investigate the potential molecular pathways linking MDA-7 signal transduction to iNOS expression, we studied IRF1 and IRF2 expression in melanoma cells after treatment with rhMDA-7. Immunoblotting for IRF-1 and IRF-2 molecules in rhMDA-7-treated cell lysates demonstrated an up-regulation of IRF-2 expression within 4 h. On the other hand, IRF-1 expression was dramatically decreased by rhMDA-7 treatment of MeWo cells within 4 h (Fig. 6). Although differences did not reach significance because of the small sample size, IRF-1 expression fell by almost 4-fold, whereas IRF-2 expression increased by 4.7-fold.

Numerous studies have been conducted on the actions of NO signaling in the vascular system. However, the role of NO in cancer is less clear. Much work is being focused on the understanding of its actions and interactions with other molecules in tumorigenesis. Previous clinical and experimental data from our laboratory have indicated that iNOS expression by melanoma metastases predicts a poor clinical outcome (23). Conversely, MDA-7 expression has been found to decline with melanoma progression, in keeping with its role as a tumor suppressor (33, 34). The objectives of this study were to examine human melanoma tissue for the relationship between iNOS and MDA-7 expression, and to elucidate potential mechanisms whereby these molecules might interact. On immunohistochemical examination of human melanoma tumor tissue, we have found that the mean iNOS number and intensity scores decline significantly as the MDA-7 scores increase. These findings suggested that one of these molecules might function to control the expression of the other, although the direction of regulation could not be determined by the immunohistochemistry experiments. However, we have also shown that exposure of iNOS-expressing melanoma cells to MDA-7, either by infection with Ad-mda7 or incubation with rhMDA-7, results in decline of iNOS expression by these cells. These results indicated that MDA-7 was involved in down-regulation of iNOS and led us to pursue the mechanistic pathway by which this regulation occurred.

A milestone in MDA-7 research was the recognition of its cytokine characteristics. A number of groups predicted that MDA-7 protein was a novel cytokine, and MDA-7 has been recently classified as IL-24 (3944). The IL-24 gene encodes a secreted protein that exhibits homology to the IL-10 family of cytokines, which includes IL-19, IL-20, IL-22 (IL-TIF), and IL-26 (AK155). Dumoutier et al. (43) showed that MDA-7/IL-24 binds to the IL-20 receptor complex (IL20R1/IL20R2), and ligand binding results in STAT-3 phosphorylation. Many cytokines, hormones, and growth factors use STAT signaling pathways to modulate a remarkable variety of biological responses, including differentiation, cell proliferation, survival, and oncogenesis (4547). Thus, we examined this signaling pathway as a potential route for the regulation of iNOS expression by MDA-7. We have found that treatment of human melanoma with rhMDA-7 results in STAT-3 phosphorylation and trans-location to the nucleus, demonstrating that the JAK-STAT signaling pathway has indeed been activated. Recent studies on intracellular cytokine signaling have provided specific evidence for the transcriptional modulation of target genes via this pathway. The activated JAKs phosphorylate the STAT proteins. After undergoing tyrosine phosphorylation, dimerization, and translocation from the cytoplasm to the nucleus, STATs bind to specific cis-acting nucleotide sequences, STAT-binding elements. STAT-1 phosphorylation is mainly induced by IFN-γ, whereas IL-6 induces the phosphorylation of STAT-3 (48, 49). Both STAT-3 and STAT-1 recognize similar STAT-binding element motifs, providing a possible molecular rationale for their often antagonistic effects.

A common concept for the role of STATs in cancer biology originated from early findings that STAT-1 appeared to function as an antiproliferative molecule, and STAT-3 and STAT-5 as oncoproteins (45, 47, 50). However, more recent studies have shown that STAT-3 can actually induce apoptosis by various mechanisms (5153). One of these proapoptotic mechanisms uses IRFs. IRFs are nuclear transcription factors that respond to α and γ IFN (IFN-γ) via the JAK-STAT signaling pathway (54). The importance of IRF-1 and STAT-1 for the induction of iNOS gene expression in response to IFN-γ has been well documented (37, 38). It is known that IRF-1 acts as the effector arm of the IFN-γ response in tumor cells, whereas IRF-2 binds to the same DNA consensus sequence and opposes IRF-1 activity (55, 56). Whereas IRF-1 induces iNOS gene expression in response to IFN-γ, IRF-2 blocks this pathway. Our data suggest that MDA-7 signaling may ultimately modulate the IRF transcriptional system, to the extent that MDA-7-treated melanoma cells exhibit a decline in IRF-1 and an increase in IRF-2. These alterations in the IRF balance would be predicted to result in inhibition of iNOS expression. Thus, we propose that MDA-7 regulates iNOS expression via STAT3 activation of IRF-2 with concomitant repression of IRF-1 activity.

In summary, our data suggest that binding of MDA-7 to its receptor results in inhibition of iNOS expression. In melanoma cells, MDA-7 activates phosphorylation of STAT3 and down-regulates IRF-1 whereas up-regulating IRF-2. Thus, we hypothesize that receptor engagement by MDA-7 activates STAT3 to bind to the IRF-2 promoter and block iNOS gene expression. However, it remains possible that MDA-7 might inhibit iNOS expression through more indirect pathways, such as secondary cytokine stimulation or interaction with other molecules that coregulate iNOS expression. Future experiments will additionally examine the mechanism that we propose, expanding the investigation to additional melanoma cell lines and to other tumor types. Although the availability of purified rhMDA-7 is presently a limiting factor, with continued improvement in the purification techniques, confirmatory data should be readily attainable. Furthermore, we would like to explore other pathways whereby MDA-7 and iNOS might interact in a regulatory fashion, including the possibility that iNOS, in turn, may down-regulate MDA-7.

MDA-7/IL-24 is unique in possessing the properties of both a tumor suppressor and a cytokine, and also appears to play a key role in the control of signaling pathways. It is evident that the potential of MDA-7-based therapeutics for the treatment of melanoma is highly promising.

Fig. 1.

Melanoma tumor MDA-7 expression negatively correlates with tumor iNOS expression. A, a negative association between mean iNOS count and MDA7 count. The Kendall τ-b correlation coefficient is −0.209, and is significantly different from 0 with P < 0.05. B, a negative association between mean iNOS intensity and MDA7 intensity. The Kendall τ-b correlation coefficient is −0.201, and is significantly different from 0 with P < 0.05; bars, ±SD.

Fig. 1.

Melanoma tumor MDA-7 expression negatively correlates with tumor iNOS expression. A, a negative association between mean iNOS count and MDA7 count. The Kendall τ-b correlation coefficient is −0.209, and is significantly different from 0 with P < 0.05. B, a negative association between mean iNOS intensity and MDA7 intensity. The Kendall τ-b correlation coefficient is −0.201, and is significantly different from 0 with P < 0.05; bars, ±SD.

Close modal
Fig. 2.

Tumor MDA-7 and iNOS expression show inverse correlation in human melanoma. Immunohistochemical labeling of MDA-7 in a paired primary tumor (A), lymph node (B), and brain metastasis (C). iNOS immunoreactivity in the same primary tumor (D), lymph node (E), and brain metastasis (F). (anti-MDA-7 and anti-iNOS, AEC, hematoxylin. Original magnification, ×20)

Fig. 2.

Tumor MDA-7 and iNOS expression show inverse correlation in human melanoma. Immunohistochemical labeling of MDA-7 in a paired primary tumor (A), lymph node (B), and brain metastasis (C). iNOS immunoreactivity in the same primary tumor (D), lymph node (E), and brain metastasis (F). (anti-MDA-7 and anti-iNOS, AEC, hematoxylin. Original magnification, ×20)

Close modal
Fig. 3.

Ad-mda7 down-regulates iNOS expression in human melanoma cell line, A375. iNOS expression is decreased in cells infected with Ad-mda7 at 500 viral particles (vp)/cell (B), and cannot be detected when 1000 (C) or 2000 vp/cell (D) are used. Ad-luc (A) serves as the negative control vector. (Anti-iNOS, AEC, hematoxylin. Original magnification, ×40)

Fig. 3.

Ad-mda7 down-regulates iNOS expression in human melanoma cell line, A375. iNOS expression is decreased in cells infected with Ad-mda7 at 500 viral particles (vp)/cell (B), and cannot be detected when 1000 (C) or 2000 vp/cell (D) are used. Ad-luc (A) serves as the negative control vector. (Anti-iNOS, AEC, hematoxylin. Original magnification, ×40)

Close modal
Fig. 4.

rhMDA-7 down-regulates iNOS expression in human melanoma cell line, A375. A, immunohistochemical labeling of iNOS in untreated A375 cells. ×20 magnification shows the cytoplasmic localization with intense labeling. B, 5 ng/ml rhMDA-7 treated A375 cells shows slightly reduced number of cells immunoreactive for iNOS. C, cells treated with 20 ng/ml rhMDA7 demonstrate significantly reduced immunoreactivity for iNOS in both count and intensity of staining.

Fig. 4.

rhMDA-7 down-regulates iNOS expression in human melanoma cell line, A375. A, immunohistochemical labeling of iNOS in untreated A375 cells. ×20 magnification shows the cytoplasmic localization with intense labeling. B, 5 ng/ml rhMDA-7 treated A375 cells shows slightly reduced number of cells immunoreactive for iNOS. C, cells treated with 20 ng/ml rhMDA7 demonstrate significantly reduced immunoreactivity for iNOS in both count and intensity of staining.

Close modal
Fig. 5.

MDA-7 induces up-regulation and phosphorylation of STAT-3 in human melanoma cell line, MeWo. A, immunohistochemical labeling of STAT3 in MeWo cells. B, immunohistochemical labeling of phosphoSTAT3 in MeWo cells. ×40 magnification shows no reactivity in both cytoplasmic and nuclear compartments of the cells at baseline. C, 20 ng/ml rhMDA-7-treated melanoma cells show strong labeling with anti-STAT3 in their cytoplasm. D, 20 ng/ml rMDA-7-treated melanoma cells shows strong labeling with anti-phospho STAT3 in their nuclei. (anti-STAT3 and anti-pSTAT3, AEC, hematoxylin. Original magnification, ×40)

Fig. 5.

MDA-7 induces up-regulation and phosphorylation of STAT-3 in human melanoma cell line, MeWo. A, immunohistochemical labeling of STAT3 in MeWo cells. B, immunohistochemical labeling of phosphoSTAT3 in MeWo cells. ×40 magnification shows no reactivity in both cytoplasmic and nuclear compartments of the cells at baseline. C, 20 ng/ml rhMDA-7-treated melanoma cells show strong labeling with anti-STAT3 in their cytoplasm. D, 20 ng/ml rMDA-7-treated melanoma cells shows strong labeling with anti-phospho STAT3 in their nuclei. (anti-STAT3 and anti-pSTAT3, AEC, hematoxylin. Original magnification, ×40)

Close modal
Fig. 6.

Immunoblotting analysis of IRF-1 and IRF-2 after 4 h of treatment of the human melanoma cell line MeWo with rhMDA-7. Treatments include medium only (Lane 1, negative control); supernatant from nontransfected HEK 293 cells (Lane 2, negative control); 5 ng/ml rhMDA-7 (Lane 3); and 20 ng/ml rhMDA7 (Lane 4). The membrane was immunoblotted with anti-IRF1 and IRF-2 antibodies at 1:2000 dilutions. Shown is one representative experiment. Graphs indicate IRF-1 and IRF-2 expression after normalization to actin protein in the cell lysates, and represent the mean of two experiments; bars, ±SD.

Fig. 6.

Immunoblotting analysis of IRF-1 and IRF-2 after 4 h of treatment of the human melanoma cell line MeWo with rhMDA-7. Treatments include medium only (Lane 1, negative control); supernatant from nontransfected HEK 293 cells (Lane 2, negative control); 5 ng/ml rhMDA-7 (Lane 3); and 20 ng/ml rhMDA7 (Lane 4). The membrane was immunoblotted with anti-IRF1 and IRF-2 antibodies at 1:2000 dilutions. Shown is one representative experiment. Graphs indicate IRF-1 and IRF-2 expression after normalization to actin protein in the cell lysates, and represent the mean of two experiments; bars, ±SD.

Close modal

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

Supported by NIH Grant R41 CA 89778 and RO1 CA 90282, as well as a grant from the Keck Foundation (to E.A.G.) and T32 CA72371 training grant (to J.A.E.).

3

The abbreviations used are: NO, nitric oxide; NOS, nitric oxide synthase; iNOS, inducible form of NOS; mda-7, melanoma differentiation-associated gene-7; Ad-mda7, adenoviral-melanoma differentiation-associated gene construct; CMV, cytomegalovirus; rhMDA-7, recombinant human MDA-7 protein; IL, interleukin; JAK, Janus-activated kinase; IRF, IFN regulatory factor; AEC, 3-amino-9-ethylcarbazole; STAT, signal transducers and activators of transcription.

We thank Sandra A. Kinney for technical assistance, and Sandra T. Yekell and the University of Texas M. D. Anderson Cancer Center Melanoma Tissue Bank Core Laboratory for providing samples and the slides.

1
Jemal, A., Thomas, A., Murray, T., and Thun, M. Cancer Statistics 2002.
CA Cancer J. Clin.
,
52
:
23
–47, 
2002
.
2
Vodovotz, Y., Kwon, N. S., Pospischil, M., Manning, J., Paik, J., and Nathan, C. Inactivation of nitric oxide synthase after prolonged incubation of mouse macrophages with IFN-γ and bacterial lipopolysaccharide.
J. Immunol.
,
152
:
4110
–4118, 
1994
.
3
Vakkala, M., Kahlos, K., Lakari, E., Pääkkö, P., Kinnula, V., and Soini, Y. Inducible nitric oxide synthase expression, apoptosis, and angiogenesis in in situ and invasive breast carcinomas.
Clin. Cancer Res.
,
6
:
2408
–2416, 
2000
.
4
Mortensen, K., Skouv, J., Hougaard, D. M., and Larsson, L. I. Endogenous endothelial cell nitric-oxide synthase modulates apoptosis in cultured breast cancer cells and is transcriptionally regulated by p53.
J. Biol. Chem.
,
274
:
37679
–37684, 
1999
.
5
Chazotte-Aubert, L., Hainaut, P., and Ohshima, H. Nitric oxide nitrates tyrosine residues of tumor-suppressor p53 protein in MCF-7 cells.
Biochem. Biophys. Res. Commun.
,
267
:
609
–613, 
2000
.
6
Vickers, S. M., MacMillan-Crow, L. A., Green, M., Ellis, C., and Thompson, J. A. Association of increased immunostaining for inducible nitric oxide synthase and nitrotyrosine with fibroblast growth factor transformation in pancreatic cancer.
Arch. Surg.
,
134
:
245
–251, 
1999
.
7
Hajri, A., Metzger, E., Vallat, F., Coffy, S., Flatter, E., Evrard, S., Marescaux, J., and Aprahamian, M. Role of nitric oxide in pancreatic tumour growth: in vivo and in vitro studies.
Br. J. Cancer
,
78
:
841
–849, 
1998
.
8
Gansauge, S., Nussler, A. K., Beger, H. G., and Gansauge, F. Nitric oxide-induced apoptosis in human pancreatic carcinoma cell lines is associated with a G1-arrest and an increase of the cyclin-dependent kinase inhibitor p21WAF1/CIP1.
Cell Growth Differ.
,
9
:
611
–617, 
1998
.
9
Marrogi, A. J., Travis, W. D., Welsh, J. A., Khan, M. A., Rahim, H., Tazelaar, H., Pairolero, P., Trastek, V., Jett, J., Caporaso, N. E., Liotta, L. A., and Harris, C. C. Nitric oxide synthase, cyclooxygenase 2, and vascular endothelial growth factor in the angiogenesis of non-small cell lung carcinoma.
Clin. Cancer Res.
,
6
:
4739
–4744, 
2000
.
10
Marrogi, A., Pass, H. I., Khan, M., Metheny-Barlow, L. J., Harris, C. C., and Gerwin, B. I. Human mesothelioma samples overexpress both cyclooxygenase-2 (COX-2) and inducible nitric oxide synthase (NOS2): in vitro antiproliferative effects of a COX-2 inhibitor.
Cancer Res.
,
60
:
3696
–3700, 
2000
.
11
Aaltomaa, S. H., Lipponen, P. K., Viitanen, J., Kankkunen, J-P., Ala-Opas, M. Y., and Kosma, V. M. The prognostic value of inducible nitric oxide synthase in local prostate cancer.
B. J. U. Int.
,
86
:
234
–239, 
2000
.
12
Klotz, T., Bloch, W., Volberg, C., Engelmann, U., and Addicks, K. Selective expression of inducible nitric oxide synthase in human prostate carcinoma.
Cancer (Phila.)
,
82
:
1897
–1903, 
1998
.
13
Kojima, M., Morisaki, T., Tsukahara, Y., Uchiyama, A., Matsunari, Y., Mibu, R., and Tanaka, M. Nitric oxide synthase expression and nitric oxide production in human colon carcinoma tissue.
J. Surg. Oncol.
,
70
:
222
–229, 
1999
.
14
Ambs, S., Merriam, W. G., Bennett, W. P., Felley-Bosco, E., Ogunfusika, M. O., Oser, S. M., Klein, S., Shields, P. G., Billiar, T. R., and Harris, C. C. Frequent nitric oxide synthase-2 expression in human colon adenomas: implication for tumor angiogenesis and colon cancer progression.
Cancer Res.
,
58
:
334
–341, 
1998
.
15
Moochhala, S., Chhatwal, V. J., Chan, S. T., Ngoi, S. S., Chia, Y. W., and Rauff, A. Nitric oxide synthase activity and expression in human colorectal cancer.
Carcinogenesis (Lond.)
,
17
:
1171
–1174, 
1996
.
16
Jansson, O. T., Morcos, E., Brundin, L., Bergerheim, U. S. R., Adolfsson, J., and Wiklund, N. Nitric oxide synthase activity in human renal cell carcinoma.
J. Urol.
,
160
:
556
–560, 
1998
.
17
Koh, E., Noh, S. H., Lee, Y. D., Lee, H. Y., Han, J-W., Lee, H. W., and Hong, S. Differential expression of nitric oxide synthase in human stomach cancer.
Cancer Lett.
,
146
:
173
–180, 
1999
.
18
Doi, C., Noguchi, Y., Marat, D., Saito, A., Fukuzawa, K., Yoshikawa, T., Akira, T., and Ito, T. Expression of nitric oxide synthase in gastric cancer.
Cancer Lett.
,
144
:
161
–167, 
1999
.
19
Goto, T., Haruma, K., Kitadai, Y., Ito, M., Yoshihara, M., Sumii, K., Hayakawa, N., and Kajiyama, G. Enhanced expression of inducible nitric oxide synthase and nitrotyrosine in gastric mucosa of gastric cancer patients.
Clin. Cancer Res.
,
5
:
1411
–1415, 
1999
.
20
Klotz, T., Bloch, W., Jacobs, G., Niggemann, S., Engelmann, U., and Addicks, K. Immunolocalization of inducible and constitutive nitric oxide synthases in human bladder cancer.
Urology
,
54
:
416
–419, 
1999
.
21
Bentz, B. G., Haines, G. K., III, and Radosevich, J. A. Increased protein nitrosylation in head and neck squamous cell carcinogenesis.
Head Neck
,
22
:
64
–70, 
2000
.
22
Bentz, B. G., Simmons, R. L., Haines, G. K., III, and Radosevich, J. A. The yin and yang of nitric oxide: reflections on the physiology and pathophysiology of NO.
Head Neck
,
22
:
71
–83, 
2000
.
23
Ekmekcioglu, S., Ellerhorst, J., Smid, C. M., Prieto, V. G., Munsell, M., Buzaid, A. C., and Grimm, E. A. Inducible nitric oxide synthase and nitrotyrosine in human metastatic melanoma tumors correlate with poor survival.
Clin. Cancer Res.
,
6
:
4768
–4775, 
2000
.
24
Tschugguel, W., Pustelnik, T., Lass, H., Mildner, M., Weninger, W., Schneeberger, C., Jansen, B., Tschachler, E., Waldhör, T., Huber, J. C., and Pehamberger, H. Inducible nitric oxide synthase (iNOS) expression may predict distant metastasis in human melanoma.
Br. J. Cancer
,
79
:
1609
–1612, 
1999
.
25
Salvucci, O., Carsana, M., Bersani, I., Tragni, G., and Anichini, A. Antiapoptotic role of endogenous nitric oxide in human melanoma cells.
Cancer Res.
,
61
:
318
–326, 
2001
.
26
Jiang, H., Lin, J. J., Su, Z. Z., Goldstein, N. I., and Fisher, P. B. Subtraction hybridization identifies a novel melanoma differentiation associated gene, mda-7, modulated during human melanoma differentiation, growth and progression.
Oncogene
,
11
:
2477
–2486, 
1995
.
27
Jiang, H., and Fisher, P. Use of a sensitive and efficient subtraction hybridization protocol for the identification of genes differentially regulated during the induction of differentiation in human melanoma cells.
Mol. Cell Differ.
,
1
:
285
–299, 
1993
.
28
Jiang, H., Su, Z., Lin, J. J., Goldstein, N. I., Young, C. S. H., and Fisher, P. B. The melanoma differentiation associated gene mda-7 suppresses cancer cell growth.
Proc. Natl. Acad. Sci. USA
,
93
:
9160
–9165, 
1996
.
29
Su, Z. Z., Madireddi, M. T., Lin, J. J., Young, C. S. H., Kitada, S., Reed, J. C., Goldstein, N. I., and Fisher, P. B. The cancer growth suppressor gene-mda7 selectively induces apoptosis in human breast cancer cells and inhibits tumor growth in nude mice.
Proc. Natl. Acad. Sci. USA
,
95
:
14400
–14405, 
1998
.
30
Lebedeva, I. V., Su, Z., Chang, Y., Kitada, S., Reed, J. C., and Fisher, P. B. The cancer growth suppressing gene mda-7 induces apoptosis selectively in human melanoma cells.
Oncogene
,
21
:
708
–718, 
2002
.
31
Sarkar, D., Su, Z. Z., Lebedeva, I. V., Sauane, M., Gopalkrishnan, R. V., Valerie, K., Dent, P., and Fisher, P. B. mda-7 (IL-24) mediates selective apoptosis in human melanoma cells by inducing the coordinated overexpression of the GADD family of genes by means of p38 MAPK.
Proc. Natl. Acad. Sci. USA
,
99
:
10054
–10059, 
2002
.
32
Pataer, A., Vorburger, S. A., Barber, G. N., Chada, S., Mhashilkar, A. M., Zou-Yang, H., Stewart, A. L., Balachandran, S., Roth, J. A., Hunt, K. K., and Swisher, S. G. Adenoviral transfer of the melanoma differentiation-associated gene 7 (mda7) induces apoptosis of lung cancer cells via up-regulation of the double-stranded RNA-dependent protein kinase (PKR).
Cancer Res.
,
62
:
2239
–2243, 
2002
.
33
Ekmekcioglu, S., Ellerhorst, J., Mhashilkar, A. M., Sahin, A. A., Read, C. M., Prieto, V. G., Chada, S., and Grimm, E. A. Down-regulated melanoma differentiation associated gene (mda-7) expression in human melanomas.
Int. J. Cancer
,
94
:
54
–59, 
2001
.
34
Ellerhorst, J. A., Prieto, V. G., Ekmekcioglu, S., Broemeling, L., Yekell, S., Chada, S., and Grimm, E. A. Loss of MDA-7 expression with progression of melanoma.
J. Clin. Oncol.
,
20
:
1069
–1072, 
2002
.
35
Woolson, R. (ed.). Statistical Methods for the Analysis of Biomedical Data, pp.
260
. New York, NY: John Wiley & Sons Inc., 
1987
.
36
Mhashilkar, A. M., Schrock, R. D., Hindi, M., Liao, J., Sieger, K., Kourouma, F., Zou-Yang, X. H., Onishi, E., Takh, O., Vedvick, T. S., Fanger, G., Stewart, L., Watson, G. J., Snary, D., Fisher, P. B., Saeki, T., Roth, J. A., Ramesh, R., and Chada, S. Melanoma differentiation associated gene-7 (mda-7): a novel anti-tumor gene for cancer gene therapy.
Mol. Med.
,
7
:
271
–282, 
2001
.
37
Saura, M., Zaragoza, C., Bao, C., McMillan, A., and Lowenstein, C. J. Interaction of interferon regulatory factor-1 and nuclear factor κ B during activation of inducible nitric oxide synthase transcription.
J. Mol. Biol.
,
289
:
459
–471, 
1999
.
38
Dell’Albani, P., Santangelo, R., Torrisi, L., Nicoletti, V. G., de Vellis, J., and Giuffrida Stella, A. M. JAK/STAT signaling pathway mediates cytokine-induced iNOS expression in primary astroglial cell cultures.
J. Neurosci. Res.
,
65
:
417
–424, 
2001
.
39
Schaefer, G., Venkataraman, C., and Schindler, U. Cutting edge: FISP (IL-4-induced secreted protein), a novel cytokine-like molecule secreted by Th2 cells.
J. Immunol.
,
166
:
5859
–5863, 
2001
.
40
Huang, E. Y., Madireddi, M. T., Gopalkrishnan, R. V., Leszczynieka, M., Su, Z., Lebedeva, I. V., Kang, D., Jiang, H., Lin, J. J., Alexandre, D., Chen, Y., Vozhilla, N., Mei, M. X., Chriatiansen, K. A., Sivo, F., Goldstein, N. I., Mhashilkar, A. B., Chada, S., Huberman, E., Pestka, S., and Fisher, P. B. Genomic structure, chromosomal localization and expression profile of a novel melanoma differentiation associated (mda-7) gene with cancer specific growth suppressing and apoptosis inducing properties.
Oncogene
,
20
:
7051
–7063, 
2001
.
41
Wang, M., Tan, Z., Zhang, R., Kotenko, S. V., and Liang, P. Interleukin 24 (MDA-7/MOB-5) signals through two heterodimeric receptors. IL-22R1/IL-20R2 and IL-20R1/IL-20R2.
J. Biol. Chem.
,
277
:
7341
–7347, 
2002
.
42
Dumoutier, L., and Renauld, J. C. Viral and cellular interleukin-10 (IL-10)-related cytokines: from structures to functions. Eur.
Cytokine Netw.
,
13
:
5
–15, 
2002
.
43
Dumoutier, L., Leemans, C., Lejeune, D., Kotenko, S. V., and Renauld, J. C. Cutting edge: STAT activation by IL-19. IL-20 and mda-7 through IL-20 receptor complexes of two types.
J. Immunol.
,
167
:
3545
–3549, 
2001
.
44
Caudell, E. G., Mumm, J. B., Poindexter, N., Ekmekcioglu, S., Mhashilkar, A. M., Yang, X. H., Retter, M. W., Hill, P., Chada, S., and Grimm, E. A. The protein product of the tumor suppressor gene, melanoma differentiation-associated gene 7, exhibits immunostimulatory activity and is designated IL-24.
J. Immunol.
,
168
:
6041
–6046, 
2002
.
45
Bromberg, J., and Darnell, J. E., Jr. The role of STATs in transcriptional control and their impact on cellular function.
Oncogene
,
19
:
2468
–2473, 
2000
.
46
Jove, R. Preface: STAT Signaling.
Oncogene
,
19
:
2466
–2467, 
2000
.
47
Bowman, T., Garcia, R., Turkson, J., and Jove, R. STATs in oncogenesis.
Oncogene
,
19
:
2474
–2488, 
2000
.
48
Darnell, J. J. STATs and gene regulation.
Science (Wash. DC)
,
277
:
1630
–1635, 
1997
.
49
Stahl, N., Boulton, T. G., Farruggella, T., Ip, N. Y., Davis, S., Witthuhn, B. A., Quelle, F. W., Silvennoinen, O., Barbieri, G., and Pellegrini, S. Association and activation of Jak-Tyk kinases by CNTF-LIF-OSM-IL-6 β receptor components.
Science (Wash. DC)
,
263
:
92
–95, 
1994
.
50
Hirano, T., Ishihara, K., and Hibi, M. Roles of STAT3 in mediating the cell growth, differentiation and survival signals relayed through the IL-6 family of cytokine receptors.
Oncogene
,
19
:
2548
–2556, 
2000
.
51
Chapman, R. S., Lourenco, P. C., Tonner, E., Flint, D. J., Selbert, S., Takeda, K., Akira, S., Clarke, A. R., and Watson, C. J. Suppression of epithelial apoptosis and delayed mammary gland involution in mice with a conditional knockout of Stat3.
Genes Dev.
,
13
:
2604
–2616, 
1999
.
52
Skov, S., Nielsen, M., Bregenholt, S., Odum, N., and Claesson, M. H. Activation of Stat3 is involved in the induction of apoptosis after ligation of major histocompatibility complex class I molecules on human Jurkat T cells.
Blood
,
91
:
3566
–3573, 
1998
.
53
Minami, M., Inoue, M., Wei, S., Takeda, K., Matsumoto, M., Kishimoto, T., and Akira, S. STAT3 activation is a critical step in gp130-mediated terminal differentiation and growth arrest of a myeloid cell line.
Proc. Natl. Acad. Sci. USA
,
93
:
3963
–3966, 
1996
.
54
Mamane, Y., Heylbroeck, C., Genin, P., Algarte, M., Servant, M. J., LePage, C., DeLuca, C., Kwon, H., Lin, R., and Hiscott, J. Interferon regulatory factors: the next generation.
Gene
,
237
:
1
–14, 
1999
.
55
Coccia, E. M., Stellacci, E., Marziali, G., Weiss, G., and Battistini, A. IFN- and IL-4 differently regulate inducible NO synthase gene expression through IRF-1 modulation.
Int. Immunol.
,
12
:
977
–985, 
2000
.
56
Lowney, J. K., Boucher, L. D., Swanson, P. E., and Doherty, G. M. Interferon regulatory factor-1 and -2 expression in human melanoma specimens.
Ann. Surg. Oncol.
,
6
:
604
–608, 
1999
.