- Email: [email protected]
Release of Vascular Endothelial Growth Factor from a Human Melanoma Cell Line, WM35, Is Induced by Hypoxia but Not Ultraviolet Radiation and Is Potentiated by Activated Ras Mutation
ORIGINAL ARTICLE
Release of Vascular Endothelial Growth Factor from a Human Melanoma Cell Line, WM35, Is Induced by Hypoxia but Not Ultraviolet Radiation and Is Potentiated by Activated Ras Mutation Yiqun G. Shellman, Young-Lip Park, David G. Marr, Katie Casper,w Yisheng Xu, Mayumi Fujita, Robert Swerlick,w and David A. Norris
Department of Dermatology, University of Colorado Health Science Center, Denver, Colorado, USA; wDepartment of Dermatology, Emory University School of Medicine, Atlanta, Georgia, USA
Angiogenesis, the formation of blood vessels, is a major factor in£uencing tumor growth and metastatic capacity, and VEGF is the prototype angiogenic factor. VEGF expression is also found in the dermis and tumor stroma during the course of melanoma progression. Various oncogenes such as c-Src, v-Raf, and Ras, and multiple environmental stimuli, including hypoxia and ultraviolet radiation (UVR), can regulate VEGF expression under certain conditions. We have constructed several cell lines from a radial growth phase, primary human melanoma cell line, WM35. We have stably transfected WM35 cells with mutant activated H-ras, N-ras, dominant negative p53, or empty vector. In this report, we determined how VEGF expression and release from these melanoma cell lines
were a¡ected by the following important factors associated with melanoma initiation and progression: hypoxia, UVR, activated Ras, dominant negative p53, and culture conditions mimicking radial growth phase melanoma (monolayer culture) and vertical growth phase melanoma (spheroid culture). We found that hypoxia, but not UVR, up-regulates VEGF mRNA expression and protein release in these melanoma cells. In addition, activated Ras and dominant negative p53 enhances the hypoxia-induced VEGF protein release. We propose that hypoxia-induced VEGF release promotes tumor progression, especially in melanomas with Ras or p53 mutations. Key word: angiogenesis/VEGF. J Invest Dermatol 121:910 ^917, 2003
A
There is a growing body of literature that supports the role of angiogenesis in the development and spread of a variety of solid tumors. Several ¢ndings suggest that VEGF is an important angiogenic factor in melanoma angiogenesis (Cla¡ey et al, 1996; Erhard et al, 1997; Marcoval et al, 1997; Salven et al, 1997; Graeven et al, 2001). Cla¡ey et al (1996) found that by stimulating angiogenesis, VEGF promotes melanoma growth. Erhard et al (1997) found that a gradual rise in the mean vascular density correlates with melanoma progression and that the transition from horizontal- to vertical-growth-phase melanoma is accompanied by induction of VEGF protein expression and accumulation of VEGF in the stroma. In addition, it appears that VEGF is upregulated during the course of melanoma progression and dissemination and that tumor-in¢ltrating cells expressing VEGF may contribute to the progression of melanoma (Salven et al, 1997). Recently, Graeven et al (2001) concluded that despite their ¢nding that VEGF does not play an essential role in the tumorigenic phenotype of human melanoma cells, VEGF does promote melanoma progression since overexpression of VEGF accelerates melanoma tumor expansion in vivo. Certain oncogenes have also been shown to upregulate VEGF expression, including c-Src (Mukhopadhyay et al, 1995), v-Raf (Grugel et al, 1995), and Ras (Grugel et al, 1995; Rak et al, 1995a, b). Ras mutations have been shown to be an important progression factor in melanoma (Ball et al, 1994; Fujita et al, 1999), but there are currently no data on the e¡ects of Ras on VEGF release in melanoma. Mutations in the p53 tumor suppressor gene can
ngiogenesis, the formation of blood vessels, is a major factor in determining tumor growth and metastatic capacity (Fidler and Ellis, 1994; Folkman, 1995; Hanahan and Folkman, 1996). Tumor growth requires fresh nutrients (growth factors) and oxygen supplied via neovascularization. The most widely studied mechanism known to induce angiogenesis is the release of endothelial cell growth factors. Vascular endothelial growth factor (VEGF), also known as vascular permeability factor, is a potent angiogenic and endothelial cell-speci¢c mitogen (Leung et al, 1989; Senger et al, 1993). VEGF is expressed and secreted at low levels by most normal cells but is constitutively expressed at much higher levels by many human tumors and tumor cell lines, including breast, colon, ovarian, bladder, and brain tumors and melanomas (Leung et al, 1989; Shweiki et al, 1992; Senger et al, 1993; Dvorak et al, 1995; Ferrara, 1995; Kolch et al, 1995; White et al, 1995).
Manuscript received May 29, 2002; revised February 12, 2003; accepted for publication May 14, 2003 Reprint requests to: Yiqun G Shellman, Department of Dermatology, University of Colorado Health Science Center, Denver, CO 80262. Email: [email protected] Abbreviations: DNp53, dominant-negative p53; HIF-1, hypoxia-inducible factor-1; UVR, ultraviolet radiation; VEGF, vascular endothelial growth factor.
0022-202X/03/$15.00 . Copyright r 2003 by The Society for Investigative Dermatology, Inc. 910
VOL. 121, NO. 4 OCTOBER 2003
cause a slight upregulation of VEGF, which can be further enhanced by agents that activate protein kinase C (Kieser et al, 1994). Wild-type p53 also inhibits hypoxia-inducible factor-1 (HIF-1)-stimulated transcription (Blagosklonny et al, 1998). HIF1, in turn, stimulates transcription of a number of genes including VEGF, although higher levels of p53 are required to inhibit HIF1 than are necessary to transcriptionally activate p53 target genes. Numerous studies devoted to understanding the regulation of VEGF expression have found that many environmental stimuli can upregulate VEGF. In particular, hypoxia is a major inducer of VEGF expression in virtually all cell types examined (Shweiki et al, 1992; Mukhopadhyay et al, 1995). Some cytokines and growth factors are also capable of inducing VEGF, including interleukin-1b and - 6, insulin-like growth factor-1, epidermal growth factor, and transforming growth factor-a (Brogi et al, 1994; Li et al, 1995; Cohen et al, 1996; Goad et al, 1996). In addition, ultraviolet radiation (UVR) has been shown to upregulate VEGF expression in keratinocyte-derived cell lines and cultured human ¢broblasts (Brauchle et al, 1996; Mildner et al, 1999; Trompezinski et al, 2000, 2001). Melanoma is the most aggressive skin cancer and the prognosis for patients diagnosed with melanoma is often poor. The only known de¢ned environmental risk factor for this tumor is ultraviolet exposure. No previous studies have addressed the e¡ects of UVR on VEGF expression in melanomas. It has also been shown in other cell types that oncogenes and hypoxia can synergistically induce VEGF expression (Mazure et al, 1996; White et al, 1997). Previously, no studies have been carried out to address how these oncogenes, along with environmental stimuli such as hypoxia and UVR, a¡ect VEGF expression and release in melanoma. The use of both monolayer and single-cell suspension culture systems for studying cancer in vitro has been well documented. Cancer cells grown in spheroid culture mimic the growth characteristics of solid tumors by providing a three-dimensional environment with multiple and varied cell-cell and/or cell-extracellular matrix interactions. Previous studies have suggested that the signals that regulate growth in monolayers di¡er from those in spheroids (LaRue et al, 1998). In this report, we determined the e¡ects of several important factors associated with melanoma initiation and progression on VEGF expression and release from human melanoma cell lines, including hypoxia, UVR, activated Ras, dominant-negative p53 (DNp53), and culture conditions mimicking radialgrowth-phase melanoma (monolayer) and vertical-growth-phase melanoma (spheroid).
MATERIALS AND METHODS Cell lines and culture conditions WM35 is a low-malignancy, primary human melanoma cell line. Cell lines designated as H-ras, N-ras, and Neo were previously constructed by stably transfecting H-ras, N-ras, or empty vector, respectively, into WM35 cells (Fujita et al, 1999). We performed Ras activity assays and found that the H-ras and N-ras cell lines had 30% more Ras activity than the control cell line, Neo (WM35 cells stably transfected with an empty vector) (data not shown). The DNp53 cell line was similarly constructed by transfecting WM35 cells with the pc53 -SCX3 plasmid (Khlgatian et al, 2002). This plasmid contains the neo gene and the DNp53 mutant gene with a valine to alanine substitution at codon 143 (Baker et al, 1990). Approximately twice as much p53 protein can be detected by western blot in the DNp53 construct than in the control cells (Khlgatian et al, 2002). The cell lines were maintained as monolayer cultures in RPMI 1640 (Gibco BRL, Gaithersburg, MD) with 400 mg/mL G418 (Geneticin, Gemini Bio-Products, Inc., Woodland CA) and 5% fetal bovine serum (Gemini Bio-Products, Inc.). The epidermoid carcinoma cell line A431 was cultured in RPMI 1640 (Gibco BRL) with 5% fetal bovine serum (Gemini Bio-Products, Inc.). The keratinocyte-derived cell line HaCaT was cultured in DMEM (Gibco BRL) with 5% fetal bovine serum (Gemini Bio-Products, Inc.).
RAS-POTENTIATED INDUCTION OF VEGF IN WM35 CELLS
911
In monolayer culture, 3 104 to 5 104 cells in 3 mL of culture medium were seeded in each well of six-well plates for 48 h before the media were exchanged for fresh media. Cells were incubated for an additional 24 h before the media were saved and stored for measurement of VEGF release. Also at this time, the cells were harvested, counted with a hemocytometer, or stored for DNA measurement. Multicellular spheroids were generated by the liquid overlay technique as described (Rofstad et al, 1986). Brie£y, 6 - or 24 -well tissue culture plates were coated with 0.5 or 1.5 mL, respectively, of prewarmed 1% agarose (Sigma) solution in RPMI 1640 (Gibco BRL) with 5% FBS. After the agarose solidi¢ed to form a thin layer with a gentle meniscus (B20 min), a suspension of cells in the same complete medium without agarose was overlayed. The cells were incubated undisturbed at 371C, 5% CO2. Fortyeight hours later, cells with medium were harvested and centrifuged; the supernatants were saved for quanti¢cation of VEGF release and the cells were counted with a hemocytometer or stored for DNA measurement. Quanti¢cation of VEGF release The media used for growing cells were harvested as described above, and VEGF release was measured by a ELISA kit following manufacturer’s instruction (R & D Systems, Minneapolis, MN). The amount of VEGF release was normalized by total DNA content, cell number, or optical density resulting from MTS assay (cell proliferation assay) (Promega, Madison, WI) from each sample. DNA quantitation Cells were collected into Eppendorf tubes and pelleted at 325 g for 3 min. The cell pellets were dissolved in 0.5 M NaOH (0.25 mL) and were frozen before quanti¢cation. Lysates were neutralized and the DNA concentrations were quanti¢ed using a Hoechst 33258 £uorescent assay as previously described (Labarca and Paigen, 1980; Hedlund and Miller, 1994). CellTiter 96 Aqueous One solution cell proliferation assay for quantitation of cell viability (MTS assay) The reagents were obtained from Promega, and procedures were followed according to the manufacturer’s instruction. Brie£y, the assay is a colorimetric method for determining the number of viable cells. The reagent contains a tetrazolium compound (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4 -sulfopheny)-2H-tetrazolium, inner salt) and an electroncoupling reagent (phenazine ethosulfate). Assays were performed by adding 20 mL of the reagent to each 96 -well plate containing 100 mL of cells with medium. The viability of control condition was set as 100%, and other conditions were compared to this to obtain a percentage of viability. UVR treatment Melanoma cells were plated at 0.5 105 per well as monolayer cultures in triplicate wells of tissue culture treated six-well plates in 3 mL of culture media per well. Forty-eight hours later, cells were then treated with the indicated amount of UVR. Medium was removed from tissue culture plates or dishes in which cells were grown and replaced with phosphate-bu¡ered saline (PBS; without phenol red). A solar simulator (Dr Honle blue-light 2001, 220 V, 50 Hz, 400 W) was equipped with the manufacturer’s ¢lter h2, allowing light from 295 to 400 nm to pass. The solar simulator was warmed up for 10 to 15 min before exposing cells. Radiometric measurement was performed using an IL1350 radiometer/photometer (International Light, Inc., Newburyport, MA) and the manufacturer’s UVB sensor calibrated at 290 nm or the manufacturer’s UVA sensor calibrated at 360 nm. Samples were irradiated with varying doses of UVB (mJ/cm2) or sham-irradiated, with the tops of dishes removed at a distance of 8 to 12 in. from light source. PBS was aspirated and fresh growth medium added to the dish for remainder of experiment. Cells were harvested for mRNA measurement at 4 h after UVR exposure. Twenty-four hours after UVR exposure, cell culture supernatants were removed and stored for quantitation of VEGF release and cells were harvested and counted with a hemacytometer or stored for DNA measurement.We also set up the parallel experiments for MTS assays to assess the viability of cells following UVR. HaCaT cells were used as a positive control because it has been reported that low-dose UVR induces VEGF production in these cells (Brauchle et al, 1996; Mildner et al, 1999; Trompezinski et al, 2001). We used the exact same protocol described by Mildner et al for HaCaT cells (Brauchle et al, 1996; Mildner et al, 1999; Trompezinski et al, 2001), which excluded serum from the culture medium. We have performed the control experiment and found that there was no di¡erence between PBS and medium without phenol red during the UVR treatments (data not shown). We also performed UVR experiments and found that the presence (or absence) of serum had no e¡ect on our ¢nding that UVR does not stimulate VEGF release from these melanoma cells (data not shown).
912
SHELLMAN ET AL
THE JOURNAL OF INVESTIGATIVE DERMATOLOGY
Hypoxia treatment Cells were grown in monolayer culture for 2 d, and media were changed immediately before the cells were exposed to either normoxic (control) or hypoxic (o0.5% O2) conditions for indicated times. The hypoxic environment was achieved by displacing the air in the incubator with condensed N2 gas, and the N2 gas was controlled by a compact gas oxygen controller (Model PROOX110, Reming Bioinstruments Company, Red¢eld, NY). Cells were harvested for preparation of total RNA at 6 h after hypoxia treatment. In parallel experiments, cell culture supernatants were saved and stored for quanti¢cation of VEGF release, and cells were harvested and counted with a hemocytometer or stored for DNA measurement at indicated times of posthypoxia treatment. mRNA quanti¢cation Total RNA was isolated using Trizol reagent (Gibco BRL) according to the manufacturer’s instruction. cDNA was generated from 1 mg of total RNA using the random primer protocol of the Superscript II preampli¢cation system for ¢rst stand cDNA synthesis (Gibco BRL). Real-time quantitative PCR was performed using SYBR green technology with the Perkin-Elmer GeneAmp 5700 sequence detection system according to the manufacturer’s instructions. Primers for VEGF (50 -TAC CTC CAC CAT GCC AGG TG-30 and 50 -GAT GAT TCT GCC CTC CTC CTT-30) were synthesized by the Microchemical Facility of the Emory University School of Medicine and optimized according to the SYBR green protocol. Standard curves were generated from serial dilutions of a cDNA and the data normalized for di¡erences in GAPDH. Ras activation assays Ras activation assays were performed using a manufacturer’s reagents and instructions (Upstate Biotechnology, Lake Placid, NY). Brie£y, cells were harvested and lyzed using manufacturer’s lysis bu¡er. Raf-1 RBD agarose-conjugated beads were used to precipitate activated Ras proteins. Proteins were detached from beads and analyzed by western blot analysis. Statistical analyses The e¡ects of hypoxia, UVR, activated Ras, DNp53, and culture conditions on the melanoma cells tested were compared to the respective controls by using the Microsoft Excel Student’s t test (two-tailed and assuming unequal variances). Signi¢cance was established at the level of po0.05.
RESULTS VEGF release from melanoma cells is much less than that from A431 cells Our lab has constructed stable transfectants of the low-malignancy WM35 human primary melanoma cell line overexpressing activated H-ras, N-ras, and DNp53 proteins (Fujita et al, 1999, #132; Khlgatian et al, 2002, #178). To determine relative amounts of Ras activity in WM35 transfectants, we performed Ras activity assays and found that the H-ras and N-ras cell lines had 30% more Ras activity than the control cell line, Neo (WM35 cells stably transfected with an empty vector) (data not shown). Approximately twice as much p53 protein can be detected by western blot in cells transfected with the DNp53 construct than in the control cells (Khlgatian et al, 2002). Here, we used these stable transfectants to investigate the e¡ects of activated H-ras, N-ras, and DNp53 on VEGF release in WM35 cells. All results were normalized to the control cell line, Neo, for comparison (Fig 1). All the melanoma cell lines tested secreted VEGF into the culture medium. Nevertheless, the epidermoid carcinoma cell line, A431, secreted 20-fold more VEGF than these melanoma cells (Fig 1). In a more extensive study of VEGF secretion from melanoma cells grown under di¡erent culture conditions, we found that there was a small increase of VEGF release in H-ras cells (45% more, t test of H-ras vs Neo, p ¼ 0.00004) and a decrease of VEGF release in DNp53 cells (25% less, t test of DNp53 vs Neo, p ¼ 0.0006) compared to the control in monolayer culture (Fig 2). A similar trend was found in spheroid culture H-ras cells secreted more and DNp53 secreted less VEGF than Neo cells (Fig 2). A Student’s t test showed statistically signi¢cant di¡erences between H-ras and Neo (p ¼ 0.004) and between DNp53 and Neo (p ¼ 0.0007). In addition, grown in spheroid culture, only Neo cells produced statistically signi¢cantly more VEGF than grown in monolayer culture (Fig 2, 0% more and
Figure 1. Melanoma cells produce less VEGF compared to A431. Cell lines designated as H-ras, N-ras, dominant-negative p53 (DNp53), and Neo were previously constructed by stably transfecting H-ras, N-ras, DNp53, or empty vector, respectively, into a primary human melanoma cell line, WM35. These melanoma cells or an epidermoid cell line A431 were grown in monolayer cultures for 2 d before the media were replaced with fresh media. Cells were harvested for preparation of total RNA 6 h later, and samples were used for quanti¢cation of VEGF mRNA expression by real-time quantitative RT-PCR (A). The VEGF mRNA expression was normalized by GAPDH mRNA expression in each sample. In parallel experiments, cell culture supernatants were saved and stored 24 h later for quanti¢cation of VEGF release by ELISA (B). The VEGF release was normalized by total amount of DNA from each sample. The VEGF release from the control cell line Neo was set equal to 1, and the relative VEGF release from other cells was calculated as the fold of VEGF release compared to Neo cells. Data represent means7SEM.
p ¼ 0.017). Nevertheless, the di¡erences in VEGF secretion among these cell lines and between the two growth conditions were very small compared to the VEGF release from A431. Mildner et al (1999) showed that A431 produces a similar amount of VEGF to that of cultured primary keratinocytes. From both our results and their study, we can infer that the melanoma cells we tested produce very little VEGF compared to primary keratinocytes. Early-stage melanomas are surrounded by keratinocytes in vivo, suggesting that activated mutant Ras or DNp53 has an insigni¢cant e¡ect on total VEGF present in early-stage melanoma tumors. Spheroid culture condition also has no signi¢cant e¡ects on VEGF release in these cells. Hypoxia induces mRNA expression and VEGF release To examine whether hypoxia a¡ected VEGF mRNA expression and protein secretion into the culture media in these melanoma cells, we grew the cells in monolayer culture under either normoxic (control) or hypoxic (o0.5% O2) conditions. RNA samples were harvested for measuring VEGF mRNA levels after 6 h of hypoxia treatment (Fig 3A) and in the parallel
VOL. 121, NO. 4 OCTOBER 2003
RAS-POTENTIATED INDUCTION OF VEGF IN WM35 CELLS
913
Figure 2. Spheroid culture does not a¡ect VEGF release from melanoma cell lines. The following melanoma cells were used: Neo (the control cell line), H-ras, N-ras, and DNp53. Cells were grown in monolayer culture or in spheroid culture for 2 d before the media were harvested for measurement of VEGF release by ELISA. The amount of VEGF release was normalized by total amount of DNA from each sample. The VEGF release from the control cell line Neo in monolayer culture was set equal to 1, and the relative VEGF release from other conditions or other cells was calculated as relative VEGF release compared to that from Neo cells in monolayer culture. This is a summary of two independent experiments of total 12 samples per data point. Data represent means7SEM.
experiments media were harvested for measuring VEGF release following 24, 36, and 48 h of hypoxia treatment (Fig 3B and data not shown). Similar results for VEGF release were obtained from 36 and 48 h of hypoxia treatment (data not shown). These data demonstrate that hypoxia is a very potent inducer of VEGF mRNA expression and protein secretion for these melanoma cells. We also performed a Student’s t test to analyze whether the induction of VEGF expression by hypoxia in these cells was signi¢cantly higher compared to the induction of VEGF expression by hypoxia in the control cell line, Neo. Only the statistically signi¢cant p values were labeled in Fig 3. The induction of VEGF mRNA in these cell lines was comparable except for N-ras clone S1. Nevertheless, induction of VEGF protein release by hypoxia was signi¢cantly higher than the control in two di¡erent clones of H-ras, clone S1 of N-ras cells, and DNp53 cells (Fig 3B). These results suggest that both activated Ras mutations and DNp53 may enhance hypoxia induction of VEGF release from these cells, possibly at a posttranslational level. UVR does not induce VEGF expression or VEGF release from these melanoma cells We examined whether UVR a¡ected VEGF mRNA expression and protein secretion in these melanoma cells. DNp53 cells were used as a control for p53 dependence of UV experiment. VEGF mRNA expression was quantitated by real-time RT-PCR (Fig 4A). In both monolayer and spheroid culture, 100 mJ per cm2 UVB did not increase VEGF mRNA levels, but reduced VEGF mRNA levels in all melanoma cells we tested (Fig 4A). Similar results on VEGF release were observed at these relatively high doses of UVR treatment (Fig 4B,C). We found that both 50 and 100 mJ per cm2 UVB signi¢cantly reduced VEGF release in all melanoma cell lines tested. The e¡ects of UVR on VEGF expression or release were more prominent in monolayer than in spheroid culture (Fig 4B,C). These calculations were normalized with GAPDH for mRNA expression and total amount of DNA for the VEGF release. HaCaT cells were used as a positive control, because it has been reported that low-dose UVR induces VEGF production in these cells (Brauchle et al, 1996; Mildner et al, 1999; Trompezinski et al,
Figure 3. Hypoxia upregulates VEGF mRNA expression and VEGF release in melanomas. The following melanoma cells were used: Neo, two clones of H-ras (S2, C7), two clones of N-ras (S1, C8), and DNp53. Cells were grown in monolayer culture for 2 d, and media were changed immediately before the cells were exposed to either normoxic (control) or hypoxic (o 0.5% O2) conditions for indicated times. The data were plotted as the relative expression in VEGF mRNA (A) or VEGF release (B) upon hypoxia compared to respective control. (A) Cells were treated with hypoxia for 6 h before being harvested for preparation of total RNA, and samples were used for quanti¢cation of VEGF mRNA expression by realtime quantitative RT-PCR. The VEGF mRNA expression was normalized by GAPDH mRNA expression in each sample. The data were plotted as the relative expression in VEGF mRNA upon hypoxia compared to respective control in each cell line. (B) At 24 h after hypoxia treatment, cell culture supernatants were saved and stored for quanti¢cation of VEGF release. The VEGF release was normalized by total DNA content from each sample. A Student’s t test analysis was performed, and the p values were calculated as the comparisons of the induction of VEGF expression between these cells and Neo cells. Only the p values, which are statistically signi¢cant, are labeled in the ¢gure (po0.05). This is a summary of two independent experiments. Data represent means7SEM.
2001). We found that both low and high doses of UVR increased VEGF release in HaCaT cells (Fig 5 and data not shown). Furthermore, using the same low doses of UVB, 10 and 25 mJ per cm2, VEGF release from melanoma cells was una¡ected (Fig 5). We also performed MTS assays for viability of these cells, and found that there was very little cell death (0%^5%) following low doses of UVR (10 and 25 mJ/cm2) and moderate cell death (10%^25%) in the highest dosage (50 and 100 mJ/cm2) (data not shown). Under all conditions, UVR did not increase VEGF production in these melanoma cells. The presence (or absence) of serum had no e¡ect on our ¢nding that UVR did not stimulate VEGF release from these melanoma cells (data not shown).
914
SHELLMAN ET AL
THE JOURNAL OF INVESTIGATIVE DERMATOLOGY
Figure 5. Lower doses of UVR do not induce VEGF release in melanomas. Cells were grown in monolayer culture for 2 d before being subjected to UVR treatment. Cells were treated with indicated amount of UVB for 24 h before quanti¢cation of VEGF release. The data were plotted as the relative amount of VEGF release upon UVR compared to respective control in each cell line. The amount of VEGF release in the control of each cell line was set equal to 1. The VEGF release was normalized by MTS assay from each sample. Data represent mean7SEM.
DISCUSSION
Figure 4. UVR does not induce VEGF mRNA expression or VEGF release in melanomas. The following melanoma cells were grown in monolayer culture or spheroid culture for 2 d before being subjected to UVR treatment. (A) Cells were treated with UVB 100 mJ per cm2 for 4 h before being harvested for preparation of total RNA, and samples were used for quanti¢cation of VEGF mRNA expression by real-time quantitative RT-PCR. The VEGF mRNA expression was normalized by GAPDH mRNA expression in each sample. The data were plotted as the relative amount of VEGF mRNA upon UVR compared to respective control in each cell line. The relative amount of VEGF mRNA in the control of each cell line was set equal to 1. (B,C) Cells were grown in monolayer (B) or spheroid culture (C) upon UVR treatment at indicated doses. Cell culture supernatants were saved and stored for quanti¢cation of VEGF release at 24 h after UVR treatment. The data were plotted as the relative amount of VEGF release upon UVR compared to respective control in each cell line. The amount of VEGF release in the control of each cell line was set equal to 1. The VEGF release was normalized by total DNA content from each sample. Data represent means7SEM. This is one representative experiment of three experiments with triplicates each.
In this study, we measured VEGF release from several melanoma cell lines originally derived from a low-malignancy, radialgrowth-phase melanoma cell line and modi¢ed by transduction of mutated Ras or p53.We found that the baseline levels of VEGF released from these cells are quite low. These data are consistent with previous reports (Erhard et al, 1997; Salven et al, 1997). Those studies found lower levels of VEGF production in early-stage melanomas compared to melanomas representing the transition from radial-growth-phase to vertical-growth-phase melanoma or metastatic melanomas (Erhard et al, 1997; Salven et al, 1997). We found that the levels of VEGF release from these melanoma cells are much lower compared to an epidermoid cell line, A431. Mildner et al (1999) showed that A431 produces a similar amount of VEGF to that of cultured primary keratinocytes. It follows, therefore, that the melanoma cells we tested should produce very little VEGF compared to primary keratinocytes. Several studies have shown that keratinocytes are a major source of VEGF production in the skin (Detmar et al, 1994; Ballaun et al, 1995; Weninger et al, 1996). Radial-growth-phase melanoma cells may not be required to produce high levels of VEGF if the primary keratinocytes surrounding the tumor are secreting su⁄cient quantities of this growth factor. Only when a melanoma mass reaches a critical volume, and tumor cells from inside the tumor mass are removed from neighboring keratinocytes, would VEGF production be more critically important for these larger advanced melanomas. Several ¢ndings supporting this concept showing a correlation between VEGF production and melanoma progression (Erhard et al, 1997; Salven et al, 1997). We investigated the regulation of VEGF production a¡ected by environmental factors, hypoxia and UVR in these WM35 melanoma cells. Our studies show that hypoxia is a very potent inducer for VEGF expression and secretion (up to 30-fold of baseline) in these melanoma cells. In addition, under these hypoxic conditions, we found that melanoma cells secrete a similar amount of VEGF to that of A431 (see Figs 1, 3). This is consistent with a previous ¢nding by Cla¡ey et al (1998). In that study, it was found that hypoxia induces VEGF expression in human M21 melanoma cells (Cla¡ey et al, 1998). In addition, our data also suggest that mutant Ras or DNp53 seems to enhance the induction of VEGF production by hypoxia at a post-transcriptional level.
VOL. 121, NO. 4 OCTOBER 2003
The uncontrolled proliferation of cells in an organ will always lead to low oxygen tension because the di¡usion range of oxygen will eventually be exceeded. The response to hypoxia is di¡erentially regulated at the level of speci¢c cell types or layers in various organs, and it is possible that there is preferential tumor development in areas of increased hypoxia responsiveness. Transfection of cells with a mutant form of p53 has been shown to result in the potentiation of the PMA-dependent activation of VEGF mRNA expression (Kieser et al, 1994). Oncogenes have also been shown to potentiate hypoxia-induced VEGF expression in other cell types (Mazure et al, 1996; White et al, 1997). Our results suggest that oncogenes such as mutant Ras and DNp53 might provide these melanoma cells some survival advantage by increasing hypoxia responsiveness, thus resulting in increased angiogenesis as a consequence of increased production of angiogenic agents such as VEGF. Hypoxia-induced VEGF expression is regulated by multiple mechanisms including transcriptional activation, mRNA stabilization, and alternative initiation (see reviews Berra et al, 2000; Paulding and Czyzyk-Krzeska 2000; Semenza 2000a, b). Cla¡ey et al (1998) found that hypoxia-induced VEGF expression is mediated by increased transcription and mRNA stability in human M21 melanoma cells. Transcriptional activation of the VEGF gene in response to hypoxia is mediated by HIF-1 (see review, Semenza 2000a, b). HIF-1 is a heterodimeric transcription factor composed of a constitutively expressed HIF-1b subunit and an O2 - and growth factor-inducible HIF-1a subunit (Semenza 2000a, b). HIF-1a subunit expression is controlled by O2 -regulated ubiquitination and proteasomal degradation (Salceda and Caro, 1997; Huang et al, 1998; Kallio et al, 1999; Sutter et al, 2000). Increased VEGF mRNA stability induced by hypoxia is mediated, at least in part, by speci¢c interactions between a de¢ned mRNA stability sequence in the 30 -untranslated region and distinct hypoxia-induced mRNA-binding proteins (Levy et al, 1996a, b, 1997; Cla¡ey et al, 1998). Loss of function of p53 has been found to increase HIF-1a expression by decreasing its ubiquitination (Ravi et al, 2000). Activated H-ras has been shown to increase VEGF expression via the PI3 kinase or MAP kinase pathway, depending upon the cell type (Rak et al, 1995; Arbiser et al, 1997; Mazure et al, 1997; Rak et al, 2000). The MAP kinase cascade can control VEGF expression by several mechanisms (see review Berra et al, 2000): (1) p42/p44 MAPKs activate the VEGF promoter at the proximal (^88/^66) region where Sp1/AP-2 transcriptional factor complexes are recruited; (2) p42/p44 MAPKs can directly phosphorylate HIF-1a and enhance HIF-1 transcriptional activation activity; and (3) MAPKs activated under various cellular stresses (p38MAPK and JNK) can stabilize the VEGF mRNA. Our data suggest that the enhancement of hypoxia-induced VEGF expression by activated H-ras or DNp53 is not mediated by increasing VEGF mRNA levels. It is possible that activated N-ras or DNp53 might regulate other controls of VEGF expression such as alternative translational initiation. Our studies do not de¢ne why activated H-ras and N-ras appear to regulate VEGF expression di¡erently. It is possible that activated H-ras and Nras have di¡erent substrate speci¢cities, and they can activate the downstream targets at di¡erent e⁄ciencies. Nevertheless, further studies are needed to investigate these possibilities. Studies have shown that with UVR exposure human keratinocyte-derived cell lines and human dermal ¢broblasts upregulate VEGF expression and production (Brauchle et al, 1996; Mildner et al, 1999; Trompezinski et al, 2001). Nevertheless, there are con£icting results regarding the e¡ect of UVR on VEGF production in primary keratinocytes. Longuet et al reported that UVB upregulates VEGF in primary keratinocytes. In contrast, Mildner et al (1999) found that UVB does not induce VEGF production in primary keratinocytes, and high doses of UVA actually reduces VEGF production signi¢cantly. There are methodologic di¡erences between these two studies. Mildner’s group used a solar
RAS-POTENTIATED INDUCTION OF VEGF IN WM35 CELLS
915
simulator as we did in our studies; however, Longuet et al used light sources with narrow bandwidth without speci¢c information on the equipment. Therefore, Mildner’s model may be a closer model to the in vivo situation and be more relevant to the photobiology of human skin. Our studies showed that a high dose of UVB reduces VEGF expression or production in the melanoma cells tested and that lower doses of UVB have little e¡ect. These data indicate that UVR does not promote melanoma progression through VEGF release. Nevertheless, UVR has been found to increase other factors, such as IL-8 expression in primary melanomas, resulting in increased tumor growth and metastasis (Singh et al, 1995; Wood and Richmond, 1995; Luca et al, 1997; Begum et al, 1999). In addition, UVR stimulates expression of many factors in keratinocytes including b-FGF, NGF, endothelin-1, IL-8, etc., and these factors might promote melanoma progression indirectly (Schwarz and Luger, 1989; Imokawa et al, 1992; Gilchrest et al, 1996; Bielenberg et al, 1998). In our study, spheroid culture conditions did not a¡ect VEGF mRNA expression or protein release signi¢cantly. In addition, the e¡ects of UVR on VEGF expression or release were less obvious in spheroid than in monolayer culture. Nevertheless, we cannot rule out the possibility that the reason for the diminished e¡ect in spheroid culture is that UVR cannot penetrate spheroid culture as well as cells grown as monolayers. There is evidence that in di¡erent cell types VEGF expression can be regulated di¡erently in response to the same stimuli (Claffey et al, 1996). Understanding the diverse mechanisms by which cells overexpress VEGF, and which mechanisms are operating in speci¢c tumor types, may help in the design of e¡ective anticancer therapies. A recent publication also suggests that loss of TSP-1 might be a key event in the switch from an antiangiogenic to an angiogenic phenotype and might be even more critical for melanoma formation than VEGF secretion (Graeven et al, 2001). It will be important to investigate whether these oncogenes and/or environmental stimuli control TSP-1 expression in melanoma. In summary, we found that hypoxia upregulated VEGF mRNA expression and protein release in the melanoma cells derived from WM35. UVR did not increase VEGF release in these melanoma cells grown in either spheroid or monolayer culture. In addition, activated Ras enhances the hypoxia-induced VEGF protein release in WM35 cells.We proposed that hypoxia-induced VEGF release promotes tumor progression, especially in melanomas with activated Ras mutations. We thank Gary Miller (Department of Pathology, University of Colorado Health Science Center) for use of the Flurolite 1000 £urometer. This work was supported in part by NIAMS Grant R01AR26427-16 (D.A.N.) and by a Dermatology Foundation career development award (Y.G.S.).
REFERENCES Arbiser JL, Moses MA, Fernandez CA, et al: Oncogenic H-ras stimulates tumor angiogenesis by two distinct pathways. Proc Natl Acad Sci USA 94:861^866, 1997 Baker SJ, Markowitz S, Fearon ER, Willson JK, Vogelstein B: Suppression of human colorectal carcinoma cell growth by wild-type p53. Science 249:912^915, 1990 Ball NJ, Yohn JJ, Morelli JG, Norris DA, Golitz LE, Hoe¥er JP: Ras mutations in human melanoma a marker of malignant progression. J Invest Dermatol 102:285^290, 1994 Ballaun C, Weninger W, Uthman A, Weich H, Tschachler E: Human keratinocytes express the three major splice forms of vascular endothelial growth factor. J Invest Dermatol 104:7^10, 1995 Begum NA, Shibuta K, Mori M, Barnard GF: Reduced expression of the CXCR4 receptor mRNA in hepatocellular carcinoma and lack of inducibility of its ligand alpha-chemokine hIRH/SDF1alpha/PBSF in vitro. Int J Oncol 14:927^934, 1999 Berra E, Pages G, Pouyssegur J: MAP kinases and hypoxia in the control of VEGF expression. Cancer Metastasis Rev 19:139^145, 2000
916
SHELLMAN ET AL
Bielenberg DR, Bucana CD, Sanchez R, Donawho CK, Kripke ML, Fidler IJ: Molecular regulation of UVB-induced cutaneous angiogenesis. J Invest Dermatol 111:864^872, 1998 Blagosklonny MV, An WG, Romanova LY, Trepel J, Fojo T, Neckers L: p53 inhibits hypoxia-inducible factor-stimulated transcription. J Biol Chem 273:11995^11998, 1998 Brauchle M, Funk JO, Kind P,Werner S: Ultraviolet B and H2O2 are potent inducers of vascular endothelial growth factor expression in cultured keratinocytes. J Biol Chem 271:21793^21797, 1996 Brogi E, Wu T, Namiki A, Isner JM: Indirect angiogenic cytokines upregulate VEGF and bFGF gene expression in vascular smooth muscle cells, whereas hypoxia upregulates VEGF expression only. Circulation 90:649^652, 1994 Cla¡ey KP, Brown LF, del Aguila LF, Tognazzi K, Yeo KT, Manseau EJ, Dvorak HF: Expression of vascular permeability factor/vascular endothelial growth factor by melanoma cells increases tumor growth, angiogenesis, and experimental metastasis. Cancer Res 56:172^181, 1996 Cla¡ey KP, Shih SC, Mullen A, : Identi¢cation of a human VPF/VEGF 30 untranslated region mediating hypoxia-induced mRNA stability. Mol Biol Cell 9: 469^481, 1998 Cohen T, Nahari D, Cerem LW, Neufeld G, Levi BZ: Interleukin 6 induces the expression of vascular endothelial growth factor. J Biol Chem 271:736^741, 1996 Detmar M, Brown LF, Cla¡ey KP, et al: Overexpression of vascular permeability factor/vascular endothelial growth factor and its receptors in psoriasis. J Exp Med 180:1141^1146, 1994 Dvorak HF, Brown LF, Detmar M, Dvorak AM: Vascular permeability factor/vascular endothelial growth factor, microvascular hyperpermeability, and angiogenesis. Am J Pathol 146:1029^1039, 1995 Erhard H, Rietveld FJ, van Altena MC, Brocker EB, Ruiter DJ, de Waal RM: Transition of horizontal to vertical growth phase melanoma is accompanied by induction of vascular endothelial growth factor expression and angiogenesis. Melanoma Res 7 (Suppl 2):S19^S26, 1997 Ferrara N: The role of vascular endothelial growth factor in pathological angiogenesis. Breast Cancer ResTreat 36:127^137, 1995 Fidler IJ, Ellis LM: The implications of angiogenesis for the biology and therapy of cancer metastasis [comment]. Cell 79:185^188, 1994 Folkman J: Angiogenesis in cancer, vascular, rheumatoid and other disease. Nat Med 1:27^31, 1995 Fujita M, Norris DA, et al: Overexpression of mutant ras in human melanoma increases invasiveness, proliferation and anchorage-independent growth in vitro and induces tumour formation and cachexia in vivo. Melanoma Res 9:279^291, 1999 Gilchrest BA, Park HY, Eller MS, Yaar M: Mechanisms of ultraviolet light-induced pigmentation. Photochem Photobiol 63:1^10, 1996 Goad DL, Rubin J, Wang H, Tashjian AH Jr, Patterson C: Enhanced expression of vascular endothelial growth factor in human SaOS-2 osteoblast-like cells and murine osteoblasts induced by insulin-like growth factor I. Endocrinology 137:2262^2268, 1996 Graeven U, Rodeck U, Karpinski S, Jost M, Philippou S, Schmiegel W: Modulation of angiogenesis and tumorigenicity of human melanocytic cells by vascular endothelial growth factor and basic ¢broblast growth factor. Cancer Res 61:7282^7290, 2001 Grugel S, Finkenzeller G, Weindel K, Barleon B, Marme D: Both v-Ha-Ras and vRaf stimulate expression of the vascular endothelial growth factor in NIH 3T3 cells. J Biol Chem 270:25915^25919, 1995 Hanahan D, Folkman J: Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 86:353^364, 1996 Hedlund TE, Miller GJ: A serum-free de¢ned medium capable of supporting growth of four established human prostatic carcinoma cell lines. Prostate 24:221^228, 1994 Huang LE, Gu J, Schau M, Bunn HF: Regulation of hypoxia-inducible factor 1alpha is mediated by an O2 -dependent degradation domain via the ubiquitin-proteasome pathway. Proc Natl Acad Sci USA 95:7987^7992, 1998 Imokawa G, Yada Y, Miyagishi M: Endothelins secreted from human keratinocytes are intrinsic mitogens for human melanocytes. J Biol Chem 267:24675^24680, 1992 Kallio PJ,Wilson WJ, O’Brien S, MakinoY, Poellinger L: Regulation of the hypoxiainducible transcription factor 1alpha by the ubiquitin-proteasome pathway. J Biol Chem 274:6519^6525, 1999 Khlgatian MK, Hadshiew IM, Asawanonda P, et al: Tyrosinase gene expression is regulated by p53. J Invest Dermatol 118:126^132, 2002 Kieser A, Weich HA, Brandner G, Marme D, Kolch W: Mutant p53 potentiates protein kinase C induction of vascular endothelial growth factor expression. Oncogene 9:963^969, 1994 Kolch W, Martiny-Baron G, Kieser A, Marme D: Regulation of the expression of the VEGF/VPS and its receptors: Role in tumor angiogenesis. Breast Cancer Res Treat 36:139^155, 1995 Labarca C, Paigen K: A simple, rapid, and sensitive DNA assay procedure. Anal Biochem 102:344^352, 1980 LaRue KE, Bradbury EM, Freyer JP: Di¡erential regulation of cyclin-dependent kinase inhibitors in monolayer and spheroid cultures of tumorigenic and nontumorigenic ¢broblasts. Cancer Res 58:1305^1314, 1998
THE JOURNAL OF INVESTIGATIVE DERMATOLOGY
Leung DW, Cachianes G, Kuang WJ, Goeddel DV, Ferrara N: Vascular endothelial growth factor is a secreted angiogenic mitogen. Science 246:1306^1309, 1989 Levy NS, Goldberg MA, Levy AP: Sequencing of the human vascular endothelial growth factor (VEGF) 30 untranslated region (UTR): Conservation of ¢ve hypoxia-inducible RNA-protein binding sites. Biochim Biophys Acta 1352:167^173, 1997 Levy AP, Levy NS, Goldberg MA: Hypoxia-inducible protein binding to vascular endothelial growth factor mRNA and its modulation by the von Hippel-Lindau protein. J Biol Chem 271:25492^25497, 1996a Levy AP, Levy NS, Goldberg MA: Post-transcriptional regulation of vascular endothelial growth factor by hypoxia. J Biol Chem 271:2746^2753, 1996b Li J, Perrella MA, Tsai JC, et al: Induction of vascular endothelial growth factor gene expression by interleukin-1 beta in rat aortic smooth muscle cells. J Biol Chem 270:308^312, 1995 Longuet-Perret I, Schmitt D, et al: Tumor necrosis factor-alpha is involved in the contrasting e¡ects of ultrtaviolet B and ultraviolet A 1 radiation on the release by normal human keratinocytes of vascular permeability factor. Br J Dermatol 138(2):221^224, 1998. Luca M, Huang S, Gershenwald JE, Singh RK, Reich R, Bar-Eli M: Expression of interleukin-8 by human melanoma cells up-regulates MMP-2 activity and increases tumor growth and metastasis. Am J Pathol 151:1105^1113, 1997 Marcoval J, Moreno A, Graells J, Vidal A, Escriba JM, Garcia-Ramirez M, Fabra A: Angiogenesis and malignant melanoma. Angiogenesis is related to the development of vertical (tumorigenic) growth phase. J Cutan Pathol 24:212^218, 1997 Mazure NM, Chen EY, Laderoute KR, Giaccia AJ: Induction of vascular endothelial growth factor by hypoxia is modulated by a phosphatidylinositol 3 -kinase/Akt signaling pathway in Ha-ras-transformed cells through a hypoxia inducible factor-1 transcriptional element. Blood 90:3322^3331, 1997 Mazure NM, Chen EY, Yeh P, Laderoute KR, Giaccia AJ: Oncogenic transformation and hypoxia synergistically act to modulate vascular endothelial growth factor expression. Cancer Res 56:3436^3440, 1996 Mildner M, Weninger W, Trautinger F, Ban J, Tschachler E: UVA and UVB radiation di¡erentially regulate vascular endothelial growth factor expression in keratinocyte-derived cell lines and in human keratinocytes. Photochem Photobiol 70:674^679, 1999 Mukhopadhyay D, Tsiokas L, Zhou XM, Foster D, Brugge JS, Sukhatme VP: Hypoxic induction of human vascular endothelial growth factor expression through c-Src activation. Nature 375:577^581, 1995 Paulding WR, Czyzyk-Krzeska MF: Hypoxia-induced regulation of mRNA stability. Adv Exp Med Biol 475:111^121, 2000 Rak J, Mitsuhashi Y, Bayko L, Filmus J, Shirasawa S, Sasazuki T, Kerbel RS: Mutant ras oncogenes upregulate VEGF/VPF expression: Implications for induction and inhibition of tumor angiogenesis. Cancer Res 55:4575^4580, 1995a Rak J, Mitsuhashi Y, Erdos V, Huang SN, Filmus J, Kerbel RS: Massive programmed cell death in intestinal epithelial cells induced by three-dimensional growth conditions: Suppression by mutant c-H-ras oncogene expression. J Cell Biol 131:1587^1598, 1995b Rak J, Mitsuhashi Y, Sheehan C, et al: Oncogenes and tumor angiogenesis: Differential modes of vascular endothelial growth factor up-regulation in rastransformed epithelial cells and ¢broblasts. Cancer Res 60:490^498, 2000 Ravi R, Mookerjee B, Bhujwalla ZM, et al: Regulation of tumor angiogenesis by p53 -induced degradation of hypoxia-inducible factor 1alpha. Genes Dev 14:34^44, 2000 Rofstad EK, Wahl A, Davies Cde L, Brustad T: Growth characteristics of human melanoma multicellular spheroids in liquid-overlay culture: Comparisons with the parent tumour xenografts. Cell Tissue Kinet 19:205^216, 1986 Salceda S, Caro J: Hypoxia-inducible factor 1alpha (HIF-1alpha) protein is rapidly degraded by the ubiquitin-proteasome system under normoxic conditions: Its stabilization by hypoxia depends on redox-induced changes. J Biol Chem 272:22642^22647, 1997 Salven P, Heikkila P, Joensuu H: Enhanced expression of vascular endothelial growth factor in metastatic melanoma. Br J Cancer 76:930^934, 1997 Schwarz T, Luger TA: E¡ect of UV irradiation on epidermal cell cytokine production. J Photochem Photobiol B 4:1^13, 1989 Semenza GL: HIF-1: Mediator of physiological and pathophysiological responses to hypoxia. J Appl Physiol 88:1474^1480, 2000a Semenza GL: HIF-1: Using two hands to £ip the angiogenic switch. Cancer Metastasis Rev 19:59^65, 2000b Senger DR, Van de Water L, Brown LF, et al: Vascular permeability factor (VPF, VEGF) in tumor biology. Cancer Metastasis Rev 12:303^324, 1993 Shweiki D, Itin A, So¡er D, Keshet E: Vascular endothelial growth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis. Nature 359:843^845, 1992 Singh RK, Gutman M, Reich R, Bar-Eli M: Ultraviolet B irradiation promotes tumorigenic and metastatic properties in primary cutaneous melanoma via induction of interleukin 8. Cancer Res 55:3669^3674, 1995 Sutter CH, Laughner E, Semenza GL: Hypoxia-inducible factor 1alpha protein expression is controlled by oxygen-regulated ubiquitination that is disrupted by deletions and missense mutations. Proc Natl Acad Sci USA 97:4748^4753, 2000 Trompezinski S, Pernet I, Mayoux C, Schmitt D,Viac J: Transforming growth factorbeta1 and ultraviolet A1 radiation increase production of vascular endothelial
VOL. 121, NO. 4 OCTOBER 2003
growth factor but not endothelin-1 in human dermal ¢broblasts. Br J Dermatol 143:539^545, 2000 Trompezinski S, Pernet I, Schmitt D, Viac J: UV radiation and prostaglandin E2 upregulate vascular endothelial growth factor (VEGF) in cultured human ¢broblasts. In£amm Res 50:422^427, 2001 Weninger W, Uthman A, Pammer J, et al: Vascular endothelial growth factor production in normal epidermis and in benign and malignant epithelial skin tumors. Lab Invest 75:647^657, 1996 White FC, Benehacene A, Scheele JS, Kamps M: VEGF mRNA is stabilized by ras and tyrosine kinase oncogenes, as well as by UV radiationEvidence for divergent stabilization pathways. Growth Factors 14:199^212, 1997
RAS-POTENTIATED INDUCTION OF VEGF IN WM35 CELLS
917
White FC, Carroll SM, Kamps MP: VEGF mRNA is reversibly stabilized by hypoxia and persistently stabilized in VEGF-overexpressing human tumor cell lines. Growth Factors 12:289^301, 1995 Wood LD, Richmond A: Constitutive and cytokine-induced expression of the melanoma growth stimulatory activity/GRO alpha gene requires both NFkappa B and novel constitutive factors. J Biol Chem 270:30619^30626, 1995
Release of Vascular Endothelial Growth Factor from a Human Melanoma Cell Line, WM35, Is Induced by Hypoxia but Not Ultraviolet Radiation and Is Potentiated by Activated Ras Mutation Yiqun G. Shellman, Young-Lip Park, David G. Marr, Katie Casper,w Yisheng Xu, Mayumi Fujita, Robert Swerlick,w and David A. Norris
Department of Dermatology, University of Colorado Health Science Center, Denver, Colorado, USA; wDepartment of Dermatology, Emory University School of Medicine, Atlanta, Georgia, USA
Angiogenesis, the formation of blood vessels, is a major factor in£uencing tumor growth and metastatic capacity, and VEGF is the prototype angiogenic factor. VEGF expression is also found in the dermis and tumor stroma during the course of melanoma progression. Various oncogenes such as c-Src, v-Raf, and Ras, and multiple environmental stimuli, including hypoxia and ultraviolet radiation (UVR), can regulate VEGF expression under certain conditions. We have constructed several cell lines from a radial growth phase, primary human melanoma cell line, WM35. We have stably transfected WM35 cells with mutant activated H-ras, N-ras, dominant negative p53, or empty vector. In this report, we determined how VEGF expression and release from these melanoma cell lines
were a¡ected by the following important factors associated with melanoma initiation and progression: hypoxia, UVR, activated Ras, dominant negative p53, and culture conditions mimicking radial growth phase melanoma (monolayer culture) and vertical growth phase melanoma (spheroid culture). We found that hypoxia, but not UVR, up-regulates VEGF mRNA expression and protein release in these melanoma cells. In addition, activated Ras and dominant negative p53 enhances the hypoxia-induced VEGF protein release. We propose that hypoxia-induced VEGF release promotes tumor progression, especially in melanomas with Ras or p53 mutations. Key word: angiogenesis/VEGF. J Invest Dermatol 121:910 ^917, 2003
A
There is a growing body of literature that supports the role of angiogenesis in the development and spread of a variety of solid tumors. Several ¢ndings suggest that VEGF is an important angiogenic factor in melanoma angiogenesis (Cla¡ey et al, 1996; Erhard et al, 1997; Marcoval et al, 1997; Salven et al, 1997; Graeven et al, 2001). Cla¡ey et al (1996) found that by stimulating angiogenesis, VEGF promotes melanoma growth. Erhard et al (1997) found that a gradual rise in the mean vascular density correlates with melanoma progression and that the transition from horizontal- to vertical-growth-phase melanoma is accompanied by induction of VEGF protein expression and accumulation of VEGF in the stroma. In addition, it appears that VEGF is upregulated during the course of melanoma progression and dissemination and that tumor-in¢ltrating cells expressing VEGF may contribute to the progression of melanoma (Salven et al, 1997). Recently, Graeven et al (2001) concluded that despite their ¢nding that VEGF does not play an essential role in the tumorigenic phenotype of human melanoma cells, VEGF does promote melanoma progression since overexpression of VEGF accelerates melanoma tumor expansion in vivo. Certain oncogenes have also been shown to upregulate VEGF expression, including c-Src (Mukhopadhyay et al, 1995), v-Raf (Grugel et al, 1995), and Ras (Grugel et al, 1995; Rak et al, 1995a, b). Ras mutations have been shown to be an important progression factor in melanoma (Ball et al, 1994; Fujita et al, 1999), but there are currently no data on the e¡ects of Ras on VEGF release in melanoma. Mutations in the p53 tumor suppressor gene can
ngiogenesis, the formation of blood vessels, is a major factor in determining tumor growth and metastatic capacity (Fidler and Ellis, 1994; Folkman, 1995; Hanahan and Folkman, 1996). Tumor growth requires fresh nutrients (growth factors) and oxygen supplied via neovascularization. The most widely studied mechanism known to induce angiogenesis is the release of endothelial cell growth factors. Vascular endothelial growth factor (VEGF), also known as vascular permeability factor, is a potent angiogenic and endothelial cell-speci¢c mitogen (Leung et al, 1989; Senger et al, 1993). VEGF is expressed and secreted at low levels by most normal cells but is constitutively expressed at much higher levels by many human tumors and tumor cell lines, including breast, colon, ovarian, bladder, and brain tumors and melanomas (Leung et al, 1989; Shweiki et al, 1992; Senger et al, 1993; Dvorak et al, 1995; Ferrara, 1995; Kolch et al, 1995; White et al, 1995).
Manuscript received May 29, 2002; revised February 12, 2003; accepted for publication May 14, 2003 Reprint requests to: Yiqun G Shellman, Department of Dermatology, University of Colorado Health Science Center, Denver, CO 80262. Email: [email protected] Abbreviations: DNp53, dominant-negative p53; HIF-1, hypoxia-inducible factor-1; UVR, ultraviolet radiation; VEGF, vascular endothelial growth factor.
0022-202X/03/$15.00 . Copyright r 2003 by The Society for Investigative Dermatology, Inc. 910
VOL. 121, NO. 4 OCTOBER 2003
cause a slight upregulation of VEGF, which can be further enhanced by agents that activate protein kinase C (Kieser et al, 1994). Wild-type p53 also inhibits hypoxia-inducible factor-1 (HIF-1)-stimulated transcription (Blagosklonny et al, 1998). HIF1, in turn, stimulates transcription of a number of genes including VEGF, although higher levels of p53 are required to inhibit HIF1 than are necessary to transcriptionally activate p53 target genes. Numerous studies devoted to understanding the regulation of VEGF expression have found that many environmental stimuli can upregulate VEGF. In particular, hypoxia is a major inducer of VEGF expression in virtually all cell types examined (Shweiki et al, 1992; Mukhopadhyay et al, 1995). Some cytokines and growth factors are also capable of inducing VEGF, including interleukin-1b and - 6, insulin-like growth factor-1, epidermal growth factor, and transforming growth factor-a (Brogi et al, 1994; Li et al, 1995; Cohen et al, 1996; Goad et al, 1996). In addition, ultraviolet radiation (UVR) has been shown to upregulate VEGF expression in keratinocyte-derived cell lines and cultured human ¢broblasts (Brauchle et al, 1996; Mildner et al, 1999; Trompezinski et al, 2000, 2001). Melanoma is the most aggressive skin cancer and the prognosis for patients diagnosed with melanoma is often poor. The only known de¢ned environmental risk factor for this tumor is ultraviolet exposure. No previous studies have addressed the e¡ects of UVR on VEGF expression in melanomas. It has also been shown in other cell types that oncogenes and hypoxia can synergistically induce VEGF expression (Mazure et al, 1996; White et al, 1997). Previously, no studies have been carried out to address how these oncogenes, along with environmental stimuli such as hypoxia and UVR, a¡ect VEGF expression and release in melanoma. The use of both monolayer and single-cell suspension culture systems for studying cancer in vitro has been well documented. Cancer cells grown in spheroid culture mimic the growth characteristics of solid tumors by providing a three-dimensional environment with multiple and varied cell-cell and/or cell-extracellular matrix interactions. Previous studies have suggested that the signals that regulate growth in monolayers di¡er from those in spheroids (LaRue et al, 1998). In this report, we determined the e¡ects of several important factors associated with melanoma initiation and progression on VEGF expression and release from human melanoma cell lines, including hypoxia, UVR, activated Ras, dominant-negative p53 (DNp53), and culture conditions mimicking radialgrowth-phase melanoma (monolayer) and vertical-growth-phase melanoma (spheroid).
MATERIALS AND METHODS Cell lines and culture conditions WM35 is a low-malignancy, primary human melanoma cell line. Cell lines designated as H-ras, N-ras, and Neo were previously constructed by stably transfecting H-ras, N-ras, or empty vector, respectively, into WM35 cells (Fujita et al, 1999). We performed Ras activity assays and found that the H-ras and N-ras cell lines had 30% more Ras activity than the control cell line, Neo (WM35 cells stably transfected with an empty vector) (data not shown). The DNp53 cell line was similarly constructed by transfecting WM35 cells with the pc53 -SCX3 plasmid (Khlgatian et al, 2002). This plasmid contains the neo gene and the DNp53 mutant gene with a valine to alanine substitution at codon 143 (Baker et al, 1990). Approximately twice as much p53 protein can be detected by western blot in the DNp53 construct than in the control cells (Khlgatian et al, 2002). The cell lines were maintained as monolayer cultures in RPMI 1640 (Gibco BRL, Gaithersburg, MD) with 400 mg/mL G418 (Geneticin, Gemini Bio-Products, Inc., Woodland CA) and 5% fetal bovine serum (Gemini Bio-Products, Inc.). The epidermoid carcinoma cell line A431 was cultured in RPMI 1640 (Gibco BRL) with 5% fetal bovine serum (Gemini Bio-Products, Inc.). The keratinocyte-derived cell line HaCaT was cultured in DMEM (Gibco BRL) with 5% fetal bovine serum (Gemini Bio-Products, Inc.).
RAS-POTENTIATED INDUCTION OF VEGF IN WM35 CELLS
911
In monolayer culture, 3 104 to 5 104 cells in 3 mL of culture medium were seeded in each well of six-well plates for 48 h before the media were exchanged for fresh media. Cells were incubated for an additional 24 h before the media were saved and stored for measurement of VEGF release. Also at this time, the cells were harvested, counted with a hemocytometer, or stored for DNA measurement. Multicellular spheroids were generated by the liquid overlay technique as described (Rofstad et al, 1986). Brie£y, 6 - or 24 -well tissue culture plates were coated with 0.5 or 1.5 mL, respectively, of prewarmed 1% agarose (Sigma) solution in RPMI 1640 (Gibco BRL) with 5% FBS. After the agarose solidi¢ed to form a thin layer with a gentle meniscus (B20 min), a suspension of cells in the same complete medium without agarose was overlayed. The cells were incubated undisturbed at 371C, 5% CO2. Fortyeight hours later, cells with medium were harvested and centrifuged; the supernatants were saved for quanti¢cation of VEGF release and the cells were counted with a hemocytometer or stored for DNA measurement. Quanti¢cation of VEGF release The media used for growing cells were harvested as described above, and VEGF release was measured by a ELISA kit following manufacturer’s instruction (R & D Systems, Minneapolis, MN). The amount of VEGF release was normalized by total DNA content, cell number, or optical density resulting from MTS assay (cell proliferation assay) (Promega, Madison, WI) from each sample. DNA quantitation Cells were collected into Eppendorf tubes and pelleted at 325 g for 3 min. The cell pellets were dissolved in 0.5 M NaOH (0.25 mL) and were frozen before quanti¢cation. Lysates were neutralized and the DNA concentrations were quanti¢ed using a Hoechst 33258 £uorescent assay as previously described (Labarca and Paigen, 1980; Hedlund and Miller, 1994). CellTiter 96 Aqueous One solution cell proliferation assay for quantitation of cell viability (MTS assay) The reagents were obtained from Promega, and procedures were followed according to the manufacturer’s instruction. Brie£y, the assay is a colorimetric method for determining the number of viable cells. The reagent contains a tetrazolium compound (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4 -sulfopheny)-2H-tetrazolium, inner salt) and an electroncoupling reagent (phenazine ethosulfate). Assays were performed by adding 20 mL of the reagent to each 96 -well plate containing 100 mL of cells with medium. The viability of control condition was set as 100%, and other conditions were compared to this to obtain a percentage of viability. UVR treatment Melanoma cells were plated at 0.5 105 per well as monolayer cultures in triplicate wells of tissue culture treated six-well plates in 3 mL of culture media per well. Forty-eight hours later, cells were then treated with the indicated amount of UVR. Medium was removed from tissue culture plates or dishes in which cells were grown and replaced with phosphate-bu¡ered saline (PBS; without phenol red). A solar simulator (Dr Honle blue-light 2001, 220 V, 50 Hz, 400 W) was equipped with the manufacturer’s ¢lter h2, allowing light from 295 to 400 nm to pass. The solar simulator was warmed up for 10 to 15 min before exposing cells. Radiometric measurement was performed using an IL1350 radiometer/photometer (International Light, Inc., Newburyport, MA) and the manufacturer’s UVB sensor calibrated at 290 nm or the manufacturer’s UVA sensor calibrated at 360 nm. Samples were irradiated with varying doses of UVB (mJ/cm2) or sham-irradiated, with the tops of dishes removed at a distance of 8 to 12 in. from light source. PBS was aspirated and fresh growth medium added to the dish for remainder of experiment. Cells were harvested for mRNA measurement at 4 h after UVR exposure. Twenty-four hours after UVR exposure, cell culture supernatants were removed and stored for quantitation of VEGF release and cells were harvested and counted with a hemacytometer or stored for DNA measurement.We also set up the parallel experiments for MTS assays to assess the viability of cells following UVR. HaCaT cells were used as a positive control because it has been reported that low-dose UVR induces VEGF production in these cells (Brauchle et al, 1996; Mildner et al, 1999; Trompezinski et al, 2001). We used the exact same protocol described by Mildner et al for HaCaT cells (Brauchle et al, 1996; Mildner et al, 1999; Trompezinski et al, 2001), which excluded serum from the culture medium. We have performed the control experiment and found that there was no di¡erence between PBS and medium without phenol red during the UVR treatments (data not shown). We also performed UVR experiments and found that the presence (or absence) of serum had no e¡ect on our ¢nding that UVR does not stimulate VEGF release from these melanoma cells (data not shown).
912
SHELLMAN ET AL
THE JOURNAL OF INVESTIGATIVE DERMATOLOGY
Hypoxia treatment Cells were grown in monolayer culture for 2 d, and media were changed immediately before the cells were exposed to either normoxic (control) or hypoxic (o0.5% O2) conditions for indicated times. The hypoxic environment was achieved by displacing the air in the incubator with condensed N2 gas, and the N2 gas was controlled by a compact gas oxygen controller (Model PROOX110, Reming Bioinstruments Company, Red¢eld, NY). Cells were harvested for preparation of total RNA at 6 h after hypoxia treatment. In parallel experiments, cell culture supernatants were saved and stored for quanti¢cation of VEGF release, and cells were harvested and counted with a hemocytometer or stored for DNA measurement at indicated times of posthypoxia treatment. mRNA quanti¢cation Total RNA was isolated using Trizol reagent (Gibco BRL) according to the manufacturer’s instruction. cDNA was generated from 1 mg of total RNA using the random primer protocol of the Superscript II preampli¢cation system for ¢rst stand cDNA synthesis (Gibco BRL). Real-time quantitative PCR was performed using SYBR green technology with the Perkin-Elmer GeneAmp 5700 sequence detection system according to the manufacturer’s instructions. Primers for VEGF (50 -TAC CTC CAC CAT GCC AGG TG-30 and 50 -GAT GAT TCT GCC CTC CTC CTT-30) were synthesized by the Microchemical Facility of the Emory University School of Medicine and optimized according to the SYBR green protocol. Standard curves were generated from serial dilutions of a cDNA and the data normalized for di¡erences in GAPDH. Ras activation assays Ras activation assays were performed using a manufacturer’s reagents and instructions (Upstate Biotechnology, Lake Placid, NY). Brie£y, cells were harvested and lyzed using manufacturer’s lysis bu¡er. Raf-1 RBD agarose-conjugated beads were used to precipitate activated Ras proteins. Proteins were detached from beads and analyzed by western blot analysis. Statistical analyses The e¡ects of hypoxia, UVR, activated Ras, DNp53, and culture conditions on the melanoma cells tested were compared to the respective controls by using the Microsoft Excel Student’s t test (two-tailed and assuming unequal variances). Signi¢cance was established at the level of po0.05.
RESULTS VEGF release from melanoma cells is much less than that from A431 cells Our lab has constructed stable transfectants of the low-malignancy WM35 human primary melanoma cell line overexpressing activated H-ras, N-ras, and DNp53 proteins (Fujita et al, 1999, #132; Khlgatian et al, 2002, #178). To determine relative amounts of Ras activity in WM35 transfectants, we performed Ras activity assays and found that the H-ras and N-ras cell lines had 30% more Ras activity than the control cell line, Neo (WM35 cells stably transfected with an empty vector) (data not shown). Approximately twice as much p53 protein can be detected by western blot in cells transfected with the DNp53 construct than in the control cells (Khlgatian et al, 2002). Here, we used these stable transfectants to investigate the e¡ects of activated H-ras, N-ras, and DNp53 on VEGF release in WM35 cells. All results were normalized to the control cell line, Neo, for comparison (Fig 1). All the melanoma cell lines tested secreted VEGF into the culture medium. Nevertheless, the epidermoid carcinoma cell line, A431, secreted 20-fold more VEGF than these melanoma cells (Fig 1). In a more extensive study of VEGF secretion from melanoma cells grown under di¡erent culture conditions, we found that there was a small increase of VEGF release in H-ras cells (45% more, t test of H-ras vs Neo, p ¼ 0.00004) and a decrease of VEGF release in DNp53 cells (25% less, t test of DNp53 vs Neo, p ¼ 0.0006) compared to the control in monolayer culture (Fig 2). A similar trend was found in spheroid culture H-ras cells secreted more and DNp53 secreted less VEGF than Neo cells (Fig 2). A Student’s t test showed statistically signi¢cant di¡erences between H-ras and Neo (p ¼ 0.004) and between DNp53 and Neo (p ¼ 0.0007). In addition, grown in spheroid culture, only Neo cells produced statistically signi¢cantly more VEGF than grown in monolayer culture (Fig 2, 0% more and
Figure 1. Melanoma cells produce less VEGF compared to A431. Cell lines designated as H-ras, N-ras, dominant-negative p53 (DNp53), and Neo were previously constructed by stably transfecting H-ras, N-ras, DNp53, or empty vector, respectively, into a primary human melanoma cell line, WM35. These melanoma cells or an epidermoid cell line A431 were grown in monolayer cultures for 2 d before the media were replaced with fresh media. Cells were harvested for preparation of total RNA 6 h later, and samples were used for quanti¢cation of VEGF mRNA expression by real-time quantitative RT-PCR (A). The VEGF mRNA expression was normalized by GAPDH mRNA expression in each sample. In parallel experiments, cell culture supernatants were saved and stored 24 h later for quanti¢cation of VEGF release by ELISA (B). The VEGF release was normalized by total amount of DNA from each sample. The VEGF release from the control cell line Neo was set equal to 1, and the relative VEGF release from other cells was calculated as the fold of VEGF release compared to Neo cells. Data represent means7SEM.
p ¼ 0.017). Nevertheless, the di¡erences in VEGF secretion among these cell lines and between the two growth conditions were very small compared to the VEGF release from A431. Mildner et al (1999) showed that A431 produces a similar amount of VEGF to that of cultured primary keratinocytes. From both our results and their study, we can infer that the melanoma cells we tested produce very little VEGF compared to primary keratinocytes. Early-stage melanomas are surrounded by keratinocytes in vivo, suggesting that activated mutant Ras or DNp53 has an insigni¢cant e¡ect on total VEGF present in early-stage melanoma tumors. Spheroid culture condition also has no signi¢cant e¡ects on VEGF release in these cells. Hypoxia induces mRNA expression and VEGF release To examine whether hypoxia a¡ected VEGF mRNA expression and protein secretion into the culture media in these melanoma cells, we grew the cells in monolayer culture under either normoxic (control) or hypoxic (o0.5% O2) conditions. RNA samples were harvested for measuring VEGF mRNA levels after 6 h of hypoxia treatment (Fig 3A) and in the parallel
VOL. 121, NO. 4 OCTOBER 2003
RAS-POTENTIATED INDUCTION OF VEGF IN WM35 CELLS
913
Figure 2. Spheroid culture does not a¡ect VEGF release from melanoma cell lines. The following melanoma cells were used: Neo (the control cell line), H-ras, N-ras, and DNp53. Cells were grown in monolayer culture or in spheroid culture for 2 d before the media were harvested for measurement of VEGF release by ELISA. The amount of VEGF release was normalized by total amount of DNA from each sample. The VEGF release from the control cell line Neo in monolayer culture was set equal to 1, and the relative VEGF release from other conditions or other cells was calculated as relative VEGF release compared to that from Neo cells in monolayer culture. This is a summary of two independent experiments of total 12 samples per data point. Data represent means7SEM.
experiments media were harvested for measuring VEGF release following 24, 36, and 48 h of hypoxia treatment (Fig 3B and data not shown). Similar results for VEGF release were obtained from 36 and 48 h of hypoxia treatment (data not shown). These data demonstrate that hypoxia is a very potent inducer of VEGF mRNA expression and protein secretion for these melanoma cells. We also performed a Student’s t test to analyze whether the induction of VEGF expression by hypoxia in these cells was signi¢cantly higher compared to the induction of VEGF expression by hypoxia in the control cell line, Neo. Only the statistically signi¢cant p values were labeled in Fig 3. The induction of VEGF mRNA in these cell lines was comparable except for N-ras clone S1. Nevertheless, induction of VEGF protein release by hypoxia was signi¢cantly higher than the control in two di¡erent clones of H-ras, clone S1 of N-ras cells, and DNp53 cells (Fig 3B). These results suggest that both activated Ras mutations and DNp53 may enhance hypoxia induction of VEGF release from these cells, possibly at a posttranslational level. UVR does not induce VEGF expression or VEGF release from these melanoma cells We examined whether UVR a¡ected VEGF mRNA expression and protein secretion in these melanoma cells. DNp53 cells were used as a control for p53 dependence of UV experiment. VEGF mRNA expression was quantitated by real-time RT-PCR (Fig 4A). In both monolayer and spheroid culture, 100 mJ per cm2 UVB did not increase VEGF mRNA levels, but reduced VEGF mRNA levels in all melanoma cells we tested (Fig 4A). Similar results on VEGF release were observed at these relatively high doses of UVR treatment (Fig 4B,C). We found that both 50 and 100 mJ per cm2 UVB signi¢cantly reduced VEGF release in all melanoma cell lines tested. The e¡ects of UVR on VEGF expression or release were more prominent in monolayer than in spheroid culture (Fig 4B,C). These calculations were normalized with GAPDH for mRNA expression and total amount of DNA for the VEGF release. HaCaT cells were used as a positive control, because it has been reported that low-dose UVR induces VEGF production in these cells (Brauchle et al, 1996; Mildner et al, 1999; Trompezinski et al,
Figure 3. Hypoxia upregulates VEGF mRNA expression and VEGF release in melanomas. The following melanoma cells were used: Neo, two clones of H-ras (S2, C7), two clones of N-ras (S1, C8), and DNp53. Cells were grown in monolayer culture for 2 d, and media were changed immediately before the cells were exposed to either normoxic (control) or hypoxic (o 0.5% O2) conditions for indicated times. The data were plotted as the relative expression in VEGF mRNA (A) or VEGF release (B) upon hypoxia compared to respective control. (A) Cells were treated with hypoxia for 6 h before being harvested for preparation of total RNA, and samples were used for quanti¢cation of VEGF mRNA expression by realtime quantitative RT-PCR. The VEGF mRNA expression was normalized by GAPDH mRNA expression in each sample. The data were plotted as the relative expression in VEGF mRNA upon hypoxia compared to respective control in each cell line. (B) At 24 h after hypoxia treatment, cell culture supernatants were saved and stored for quanti¢cation of VEGF release. The VEGF release was normalized by total DNA content from each sample. A Student’s t test analysis was performed, and the p values were calculated as the comparisons of the induction of VEGF expression between these cells and Neo cells. Only the p values, which are statistically signi¢cant, are labeled in the ¢gure (po0.05). This is a summary of two independent experiments. Data represent means7SEM.
2001). We found that both low and high doses of UVR increased VEGF release in HaCaT cells (Fig 5 and data not shown). Furthermore, using the same low doses of UVB, 10 and 25 mJ per cm2, VEGF release from melanoma cells was una¡ected (Fig 5). We also performed MTS assays for viability of these cells, and found that there was very little cell death (0%^5%) following low doses of UVR (10 and 25 mJ/cm2) and moderate cell death (10%^25%) in the highest dosage (50 and 100 mJ/cm2) (data not shown). Under all conditions, UVR did not increase VEGF production in these melanoma cells. The presence (or absence) of serum had no e¡ect on our ¢nding that UVR did not stimulate VEGF release from these melanoma cells (data not shown).
914
SHELLMAN ET AL
THE JOURNAL OF INVESTIGATIVE DERMATOLOGY
Figure 5. Lower doses of UVR do not induce VEGF release in melanomas. Cells were grown in monolayer culture for 2 d before being subjected to UVR treatment. Cells were treated with indicated amount of UVB for 24 h before quanti¢cation of VEGF release. The data were plotted as the relative amount of VEGF release upon UVR compared to respective control in each cell line. The amount of VEGF release in the control of each cell line was set equal to 1. The VEGF release was normalized by MTS assay from each sample. Data represent mean7SEM.
DISCUSSION
Figure 4. UVR does not induce VEGF mRNA expression or VEGF release in melanomas. The following melanoma cells were grown in monolayer culture or spheroid culture for 2 d before being subjected to UVR treatment. (A) Cells were treated with UVB 100 mJ per cm2 for 4 h before being harvested for preparation of total RNA, and samples were used for quanti¢cation of VEGF mRNA expression by real-time quantitative RT-PCR. The VEGF mRNA expression was normalized by GAPDH mRNA expression in each sample. The data were plotted as the relative amount of VEGF mRNA upon UVR compared to respective control in each cell line. The relative amount of VEGF mRNA in the control of each cell line was set equal to 1. (B,C) Cells were grown in monolayer (B) or spheroid culture (C) upon UVR treatment at indicated doses. Cell culture supernatants were saved and stored for quanti¢cation of VEGF release at 24 h after UVR treatment. The data were plotted as the relative amount of VEGF release upon UVR compared to respective control in each cell line. The amount of VEGF release in the control of each cell line was set equal to 1. The VEGF release was normalized by total DNA content from each sample. Data represent means7SEM. This is one representative experiment of three experiments with triplicates each.
In this study, we measured VEGF release from several melanoma cell lines originally derived from a low-malignancy, radialgrowth-phase melanoma cell line and modi¢ed by transduction of mutated Ras or p53.We found that the baseline levels of VEGF released from these cells are quite low. These data are consistent with previous reports (Erhard et al, 1997; Salven et al, 1997). Those studies found lower levels of VEGF production in early-stage melanomas compared to melanomas representing the transition from radial-growth-phase to vertical-growth-phase melanoma or metastatic melanomas (Erhard et al, 1997; Salven et al, 1997). We found that the levels of VEGF release from these melanoma cells are much lower compared to an epidermoid cell line, A431. Mildner et al (1999) showed that A431 produces a similar amount of VEGF to that of cultured primary keratinocytes. It follows, therefore, that the melanoma cells we tested should produce very little VEGF compared to primary keratinocytes. Several studies have shown that keratinocytes are a major source of VEGF production in the skin (Detmar et al, 1994; Ballaun et al, 1995; Weninger et al, 1996). Radial-growth-phase melanoma cells may not be required to produce high levels of VEGF if the primary keratinocytes surrounding the tumor are secreting su⁄cient quantities of this growth factor. Only when a melanoma mass reaches a critical volume, and tumor cells from inside the tumor mass are removed from neighboring keratinocytes, would VEGF production be more critically important for these larger advanced melanomas. Several ¢ndings supporting this concept showing a correlation between VEGF production and melanoma progression (Erhard et al, 1997; Salven et al, 1997). We investigated the regulation of VEGF production a¡ected by environmental factors, hypoxia and UVR in these WM35 melanoma cells. Our studies show that hypoxia is a very potent inducer for VEGF expression and secretion (up to 30-fold of baseline) in these melanoma cells. In addition, under these hypoxic conditions, we found that melanoma cells secrete a similar amount of VEGF to that of A431 (see Figs 1, 3). This is consistent with a previous ¢nding by Cla¡ey et al (1998). In that study, it was found that hypoxia induces VEGF expression in human M21 melanoma cells (Cla¡ey et al, 1998). In addition, our data also suggest that mutant Ras or DNp53 seems to enhance the induction of VEGF production by hypoxia at a post-transcriptional level.
VOL. 121, NO. 4 OCTOBER 2003
The uncontrolled proliferation of cells in an organ will always lead to low oxygen tension because the di¡usion range of oxygen will eventually be exceeded. The response to hypoxia is di¡erentially regulated at the level of speci¢c cell types or layers in various organs, and it is possible that there is preferential tumor development in areas of increased hypoxia responsiveness. Transfection of cells with a mutant form of p53 has been shown to result in the potentiation of the PMA-dependent activation of VEGF mRNA expression (Kieser et al, 1994). Oncogenes have also been shown to potentiate hypoxia-induced VEGF expression in other cell types (Mazure et al, 1996; White et al, 1997). Our results suggest that oncogenes such as mutant Ras and DNp53 might provide these melanoma cells some survival advantage by increasing hypoxia responsiveness, thus resulting in increased angiogenesis as a consequence of increased production of angiogenic agents such as VEGF. Hypoxia-induced VEGF expression is regulated by multiple mechanisms including transcriptional activation, mRNA stabilization, and alternative initiation (see reviews Berra et al, 2000; Paulding and Czyzyk-Krzeska 2000; Semenza 2000a, b). Cla¡ey et al (1998) found that hypoxia-induced VEGF expression is mediated by increased transcription and mRNA stability in human M21 melanoma cells. Transcriptional activation of the VEGF gene in response to hypoxia is mediated by HIF-1 (see review, Semenza 2000a, b). HIF-1 is a heterodimeric transcription factor composed of a constitutively expressed HIF-1b subunit and an O2 - and growth factor-inducible HIF-1a subunit (Semenza 2000a, b). HIF-1a subunit expression is controlled by O2 -regulated ubiquitination and proteasomal degradation (Salceda and Caro, 1997; Huang et al, 1998; Kallio et al, 1999; Sutter et al, 2000). Increased VEGF mRNA stability induced by hypoxia is mediated, at least in part, by speci¢c interactions between a de¢ned mRNA stability sequence in the 30 -untranslated region and distinct hypoxia-induced mRNA-binding proteins (Levy et al, 1996a, b, 1997; Cla¡ey et al, 1998). Loss of function of p53 has been found to increase HIF-1a expression by decreasing its ubiquitination (Ravi et al, 2000). Activated H-ras has been shown to increase VEGF expression via the PI3 kinase or MAP kinase pathway, depending upon the cell type (Rak et al, 1995; Arbiser et al, 1997; Mazure et al, 1997; Rak et al, 2000). The MAP kinase cascade can control VEGF expression by several mechanisms (see review Berra et al, 2000): (1) p42/p44 MAPKs activate the VEGF promoter at the proximal (^88/^66) region where Sp1/AP-2 transcriptional factor complexes are recruited; (2) p42/p44 MAPKs can directly phosphorylate HIF-1a and enhance HIF-1 transcriptional activation activity; and (3) MAPKs activated under various cellular stresses (p38MAPK and JNK) can stabilize the VEGF mRNA. Our data suggest that the enhancement of hypoxia-induced VEGF expression by activated H-ras or DNp53 is not mediated by increasing VEGF mRNA levels. It is possible that activated N-ras or DNp53 might regulate other controls of VEGF expression such as alternative translational initiation. Our studies do not de¢ne why activated H-ras and N-ras appear to regulate VEGF expression di¡erently. It is possible that activated H-ras and Nras have di¡erent substrate speci¢cities, and they can activate the downstream targets at di¡erent e⁄ciencies. Nevertheless, further studies are needed to investigate these possibilities. Studies have shown that with UVR exposure human keratinocyte-derived cell lines and human dermal ¢broblasts upregulate VEGF expression and production (Brauchle et al, 1996; Mildner et al, 1999; Trompezinski et al, 2001). Nevertheless, there are con£icting results regarding the e¡ect of UVR on VEGF production in primary keratinocytes. Longuet et al reported that UVB upregulates VEGF in primary keratinocytes. In contrast, Mildner et al (1999) found that UVB does not induce VEGF production in primary keratinocytes, and high doses of UVA actually reduces VEGF production signi¢cantly. There are methodologic di¡erences between these two studies. Mildner’s group used a solar
RAS-POTENTIATED INDUCTION OF VEGF IN WM35 CELLS
915
simulator as we did in our studies; however, Longuet et al used light sources with narrow bandwidth without speci¢c information on the equipment. Therefore, Mildner’s model may be a closer model to the in vivo situation and be more relevant to the photobiology of human skin. Our studies showed that a high dose of UVB reduces VEGF expression or production in the melanoma cells tested and that lower doses of UVB have little e¡ect. These data indicate that UVR does not promote melanoma progression through VEGF release. Nevertheless, UVR has been found to increase other factors, such as IL-8 expression in primary melanomas, resulting in increased tumor growth and metastasis (Singh et al, 1995; Wood and Richmond, 1995; Luca et al, 1997; Begum et al, 1999). In addition, UVR stimulates expression of many factors in keratinocytes including b-FGF, NGF, endothelin-1, IL-8, etc., and these factors might promote melanoma progression indirectly (Schwarz and Luger, 1989; Imokawa et al, 1992; Gilchrest et al, 1996; Bielenberg et al, 1998). In our study, spheroid culture conditions did not a¡ect VEGF mRNA expression or protein release signi¢cantly. In addition, the e¡ects of UVR on VEGF expression or release were less obvious in spheroid than in monolayer culture. Nevertheless, we cannot rule out the possibility that the reason for the diminished e¡ect in spheroid culture is that UVR cannot penetrate spheroid culture as well as cells grown as monolayers. There is evidence that in di¡erent cell types VEGF expression can be regulated di¡erently in response to the same stimuli (Claffey et al, 1996). Understanding the diverse mechanisms by which cells overexpress VEGF, and which mechanisms are operating in speci¢c tumor types, may help in the design of e¡ective anticancer therapies. A recent publication also suggests that loss of TSP-1 might be a key event in the switch from an antiangiogenic to an angiogenic phenotype and might be even more critical for melanoma formation than VEGF secretion (Graeven et al, 2001). It will be important to investigate whether these oncogenes and/or environmental stimuli control TSP-1 expression in melanoma. In summary, we found that hypoxia upregulated VEGF mRNA expression and protein release in the melanoma cells derived from WM35. UVR did not increase VEGF release in these melanoma cells grown in either spheroid or monolayer culture. In addition, activated Ras enhances the hypoxia-induced VEGF protein release in WM35 cells.We proposed that hypoxia-induced VEGF release promotes tumor progression, especially in melanomas with activated Ras mutations. We thank Gary Miller (Department of Pathology, University of Colorado Health Science Center) for use of the Flurolite 1000 £urometer. This work was supported in part by NIAMS Grant R01AR26427-16 (D.A.N.) and by a Dermatology Foundation career development award (Y.G.S.).
REFERENCES Arbiser JL, Moses MA, Fernandez CA, et al: Oncogenic H-ras stimulates tumor angiogenesis by two distinct pathways. Proc Natl Acad Sci USA 94:861^866, 1997 Baker SJ, Markowitz S, Fearon ER, Willson JK, Vogelstein B: Suppression of human colorectal carcinoma cell growth by wild-type p53. Science 249:912^915, 1990 Ball NJ, Yohn JJ, Morelli JG, Norris DA, Golitz LE, Hoe¥er JP: Ras mutations in human melanoma a marker of malignant progression. J Invest Dermatol 102:285^290, 1994 Ballaun C, Weninger W, Uthman A, Weich H, Tschachler E: Human keratinocytes express the three major splice forms of vascular endothelial growth factor. J Invest Dermatol 104:7^10, 1995 Begum NA, Shibuta K, Mori M, Barnard GF: Reduced expression of the CXCR4 receptor mRNA in hepatocellular carcinoma and lack of inducibility of its ligand alpha-chemokine hIRH/SDF1alpha/PBSF in vitro. Int J Oncol 14:927^934, 1999 Berra E, Pages G, Pouyssegur J: MAP kinases and hypoxia in the control of VEGF expression. Cancer Metastasis Rev 19:139^145, 2000
916
SHELLMAN ET AL
Bielenberg DR, Bucana CD, Sanchez R, Donawho CK, Kripke ML, Fidler IJ: Molecular regulation of UVB-induced cutaneous angiogenesis. J Invest Dermatol 111:864^872, 1998 Blagosklonny MV, An WG, Romanova LY, Trepel J, Fojo T, Neckers L: p53 inhibits hypoxia-inducible factor-stimulated transcription. J Biol Chem 273:11995^11998, 1998 Brauchle M, Funk JO, Kind P,Werner S: Ultraviolet B and H2O2 are potent inducers of vascular endothelial growth factor expression in cultured keratinocytes. J Biol Chem 271:21793^21797, 1996 Brogi E, Wu T, Namiki A, Isner JM: Indirect angiogenic cytokines upregulate VEGF and bFGF gene expression in vascular smooth muscle cells, whereas hypoxia upregulates VEGF expression only. Circulation 90:649^652, 1994 Cla¡ey KP, Brown LF, del Aguila LF, Tognazzi K, Yeo KT, Manseau EJ, Dvorak HF: Expression of vascular permeability factor/vascular endothelial growth factor by melanoma cells increases tumor growth, angiogenesis, and experimental metastasis. Cancer Res 56:172^181, 1996 Cla¡ey KP, Shih SC, Mullen A, : Identi¢cation of a human VPF/VEGF 30 untranslated region mediating hypoxia-induced mRNA stability. Mol Biol Cell 9: 469^481, 1998 Cohen T, Nahari D, Cerem LW, Neufeld G, Levi BZ: Interleukin 6 induces the expression of vascular endothelial growth factor. J Biol Chem 271:736^741, 1996 Detmar M, Brown LF, Cla¡ey KP, et al: Overexpression of vascular permeability factor/vascular endothelial growth factor and its receptors in psoriasis. J Exp Med 180:1141^1146, 1994 Dvorak HF, Brown LF, Detmar M, Dvorak AM: Vascular permeability factor/vascular endothelial growth factor, microvascular hyperpermeability, and angiogenesis. Am J Pathol 146:1029^1039, 1995 Erhard H, Rietveld FJ, van Altena MC, Brocker EB, Ruiter DJ, de Waal RM: Transition of horizontal to vertical growth phase melanoma is accompanied by induction of vascular endothelial growth factor expression and angiogenesis. Melanoma Res 7 (Suppl 2):S19^S26, 1997 Ferrara N: The role of vascular endothelial growth factor in pathological angiogenesis. Breast Cancer ResTreat 36:127^137, 1995 Fidler IJ, Ellis LM: The implications of angiogenesis for the biology and therapy of cancer metastasis [comment]. Cell 79:185^188, 1994 Folkman J: Angiogenesis in cancer, vascular, rheumatoid and other disease. Nat Med 1:27^31, 1995 Fujita M, Norris DA, et al: Overexpression of mutant ras in human melanoma increases invasiveness, proliferation and anchorage-independent growth in vitro and induces tumour formation and cachexia in vivo. Melanoma Res 9:279^291, 1999 Gilchrest BA, Park HY, Eller MS, Yaar M: Mechanisms of ultraviolet light-induced pigmentation. Photochem Photobiol 63:1^10, 1996 Goad DL, Rubin J, Wang H, Tashjian AH Jr, Patterson C: Enhanced expression of vascular endothelial growth factor in human SaOS-2 osteoblast-like cells and murine osteoblasts induced by insulin-like growth factor I. Endocrinology 137:2262^2268, 1996 Graeven U, Rodeck U, Karpinski S, Jost M, Philippou S, Schmiegel W: Modulation of angiogenesis and tumorigenicity of human melanocytic cells by vascular endothelial growth factor and basic ¢broblast growth factor. Cancer Res 61:7282^7290, 2001 Grugel S, Finkenzeller G, Weindel K, Barleon B, Marme D: Both v-Ha-Ras and vRaf stimulate expression of the vascular endothelial growth factor in NIH 3T3 cells. J Biol Chem 270:25915^25919, 1995 Hanahan D, Folkman J: Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 86:353^364, 1996 Hedlund TE, Miller GJ: A serum-free de¢ned medium capable of supporting growth of four established human prostatic carcinoma cell lines. Prostate 24:221^228, 1994 Huang LE, Gu J, Schau M, Bunn HF: Regulation of hypoxia-inducible factor 1alpha is mediated by an O2 -dependent degradation domain via the ubiquitin-proteasome pathway. Proc Natl Acad Sci USA 95:7987^7992, 1998 Imokawa G, Yada Y, Miyagishi M: Endothelins secreted from human keratinocytes are intrinsic mitogens for human melanocytes. J Biol Chem 267:24675^24680, 1992 Kallio PJ,Wilson WJ, O’Brien S, MakinoY, Poellinger L: Regulation of the hypoxiainducible transcription factor 1alpha by the ubiquitin-proteasome pathway. J Biol Chem 274:6519^6525, 1999 Khlgatian MK, Hadshiew IM, Asawanonda P, et al: Tyrosinase gene expression is regulated by p53. J Invest Dermatol 118:126^132, 2002 Kieser A, Weich HA, Brandner G, Marme D, Kolch W: Mutant p53 potentiates protein kinase C induction of vascular endothelial growth factor expression. Oncogene 9:963^969, 1994 Kolch W, Martiny-Baron G, Kieser A, Marme D: Regulation of the expression of the VEGF/VPS and its receptors: Role in tumor angiogenesis. Breast Cancer Res Treat 36:139^155, 1995 Labarca C, Paigen K: A simple, rapid, and sensitive DNA assay procedure. Anal Biochem 102:344^352, 1980 LaRue KE, Bradbury EM, Freyer JP: Di¡erential regulation of cyclin-dependent kinase inhibitors in monolayer and spheroid cultures of tumorigenic and nontumorigenic ¢broblasts. Cancer Res 58:1305^1314, 1998
THE JOURNAL OF INVESTIGATIVE DERMATOLOGY
Leung DW, Cachianes G, Kuang WJ, Goeddel DV, Ferrara N: Vascular endothelial growth factor is a secreted angiogenic mitogen. Science 246:1306^1309, 1989 Levy NS, Goldberg MA, Levy AP: Sequencing of the human vascular endothelial growth factor (VEGF) 30 untranslated region (UTR): Conservation of ¢ve hypoxia-inducible RNA-protein binding sites. Biochim Biophys Acta 1352:167^173, 1997 Levy AP, Levy NS, Goldberg MA: Hypoxia-inducible protein binding to vascular endothelial growth factor mRNA and its modulation by the von Hippel-Lindau protein. J Biol Chem 271:25492^25497, 1996a Levy AP, Levy NS, Goldberg MA: Post-transcriptional regulation of vascular endothelial growth factor by hypoxia. J Biol Chem 271:2746^2753, 1996b Li J, Perrella MA, Tsai JC, et al: Induction of vascular endothelial growth factor gene expression by interleukin-1 beta in rat aortic smooth muscle cells. J Biol Chem 270:308^312, 1995 Longuet-Perret I, Schmitt D, et al: Tumor necrosis factor-alpha is involved in the contrasting e¡ects of ultrtaviolet B and ultraviolet A 1 radiation on the release by normal human keratinocytes of vascular permeability factor. Br J Dermatol 138(2):221^224, 1998. Luca M, Huang S, Gershenwald JE, Singh RK, Reich R, Bar-Eli M: Expression of interleukin-8 by human melanoma cells up-regulates MMP-2 activity and increases tumor growth and metastasis. Am J Pathol 151:1105^1113, 1997 Marcoval J, Moreno A, Graells J, Vidal A, Escriba JM, Garcia-Ramirez M, Fabra A: Angiogenesis and malignant melanoma. Angiogenesis is related to the development of vertical (tumorigenic) growth phase. J Cutan Pathol 24:212^218, 1997 Mazure NM, Chen EY, Laderoute KR, Giaccia AJ: Induction of vascular endothelial growth factor by hypoxia is modulated by a phosphatidylinositol 3 -kinase/Akt signaling pathway in Ha-ras-transformed cells through a hypoxia inducible factor-1 transcriptional element. Blood 90:3322^3331, 1997 Mazure NM, Chen EY, Yeh P, Laderoute KR, Giaccia AJ: Oncogenic transformation and hypoxia synergistically act to modulate vascular endothelial growth factor expression. Cancer Res 56:3436^3440, 1996 Mildner M, Weninger W, Trautinger F, Ban J, Tschachler E: UVA and UVB radiation di¡erentially regulate vascular endothelial growth factor expression in keratinocyte-derived cell lines and in human keratinocytes. Photochem Photobiol 70:674^679, 1999 Mukhopadhyay D, Tsiokas L, Zhou XM, Foster D, Brugge JS, Sukhatme VP: Hypoxic induction of human vascular endothelial growth factor expression through c-Src activation. Nature 375:577^581, 1995 Paulding WR, Czyzyk-Krzeska MF: Hypoxia-induced regulation of mRNA stability. Adv Exp Med Biol 475:111^121, 2000 Rak J, Mitsuhashi Y, Bayko L, Filmus J, Shirasawa S, Sasazuki T, Kerbel RS: Mutant ras oncogenes upregulate VEGF/VPF expression: Implications for induction and inhibition of tumor angiogenesis. Cancer Res 55:4575^4580, 1995a Rak J, Mitsuhashi Y, Erdos V, Huang SN, Filmus J, Kerbel RS: Massive programmed cell death in intestinal epithelial cells induced by three-dimensional growth conditions: Suppression by mutant c-H-ras oncogene expression. J Cell Biol 131:1587^1598, 1995b Rak J, Mitsuhashi Y, Sheehan C, et al: Oncogenes and tumor angiogenesis: Differential modes of vascular endothelial growth factor up-regulation in rastransformed epithelial cells and ¢broblasts. Cancer Res 60:490^498, 2000 Ravi R, Mookerjee B, Bhujwalla ZM, et al: Regulation of tumor angiogenesis by p53 -induced degradation of hypoxia-inducible factor 1alpha. Genes Dev 14:34^44, 2000 Rofstad EK, Wahl A, Davies Cde L, Brustad T: Growth characteristics of human melanoma multicellular spheroids in liquid-overlay culture: Comparisons with the parent tumour xenografts. Cell Tissue Kinet 19:205^216, 1986 Salceda S, Caro J: Hypoxia-inducible factor 1alpha (HIF-1alpha) protein is rapidly degraded by the ubiquitin-proteasome system under normoxic conditions: Its stabilization by hypoxia depends on redox-induced changes. J Biol Chem 272:22642^22647, 1997 Salven P, Heikkila P, Joensuu H: Enhanced expression of vascular endothelial growth factor in metastatic melanoma. Br J Cancer 76:930^934, 1997 Schwarz T, Luger TA: E¡ect of UV irradiation on epidermal cell cytokine production. J Photochem Photobiol B 4:1^13, 1989 Semenza GL: HIF-1: Mediator of physiological and pathophysiological responses to hypoxia. J Appl Physiol 88:1474^1480, 2000a Semenza GL: HIF-1: Using two hands to £ip the angiogenic switch. Cancer Metastasis Rev 19:59^65, 2000b Senger DR, Van de Water L, Brown LF, et al: Vascular permeability factor (VPF, VEGF) in tumor biology. Cancer Metastasis Rev 12:303^324, 1993 Shweiki D, Itin A, So¡er D, Keshet E: Vascular endothelial growth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis. Nature 359:843^845, 1992 Singh RK, Gutman M, Reich R, Bar-Eli M: Ultraviolet B irradiation promotes tumorigenic and metastatic properties in primary cutaneous melanoma via induction of interleukin 8. Cancer Res 55:3669^3674, 1995 Sutter CH, Laughner E, Semenza GL: Hypoxia-inducible factor 1alpha protein expression is controlled by oxygen-regulated ubiquitination that is disrupted by deletions and missense mutations. Proc Natl Acad Sci USA 97:4748^4753, 2000 Trompezinski S, Pernet I, Mayoux C, Schmitt D,Viac J: Transforming growth factorbeta1 and ultraviolet A1 radiation increase production of vascular endothelial
VOL. 121, NO. 4 OCTOBER 2003
growth factor but not endothelin-1 in human dermal ¢broblasts. Br J Dermatol 143:539^545, 2000 Trompezinski S, Pernet I, Schmitt D, Viac J: UV radiation and prostaglandin E2 upregulate vascular endothelial growth factor (VEGF) in cultured human ¢broblasts. In£amm Res 50:422^427, 2001 Weninger W, Uthman A, Pammer J, et al: Vascular endothelial growth factor production in normal epidermis and in benign and malignant epithelial skin tumors. Lab Invest 75:647^657, 1996 White FC, Benehacene A, Scheele JS, Kamps M: VEGF mRNA is stabilized by ras and tyrosine kinase oncogenes, as well as by UV radiationEvidence for divergent stabilization pathways. Growth Factors 14:199^212, 1997
RAS-POTENTIATED INDUCTION OF VEGF IN WM35 CELLS
917
White FC, Carroll SM, Kamps MP: VEGF mRNA is reversibly stabilized by hypoxia and persistently stabilized in VEGF-overexpressing human tumor cell lines. Growth Factors 12:289^301, 1995 Wood LD, Richmond A: Constitutive and cytokine-induced expression of the melanoma growth stimulatory activity/GRO alpha gene requires both NFkappa B and novel constitutive factors. J Biol Chem 270:30619^30626, 1995