Constitutive IκB Kinase Activity Correlates with Nuclear Factor-κB Activation in Human Melanoma Cells

INTRODUCTION

Malignant melanoma is the seventh leading cancer in the United States, and despite extensive research, the prognosis remains poor for patients with inoperable melanoma. Melanoma cells have been widely used to study the aberrant gene expression to explore the molecular mechanism of melanocyte transformation. It is known that human melanocytes and melanoma cell lines produce multiple cytokines and growth factors, including basic fibroblast growth factor, transforming growth factors α and β, nerve growth factor, IL 3 -1, IL-6, TNF α and β, CXCL8 (IL-8), CXCL1 (MGSAα/GROα), CCL5 (RANTES), CXCL10 (IP-10), CXCL9 (mig), and monocyte chemotactic and activating factor (1, 2, 3, 4, 5, 6, 7, 8, 9, 10) . Overexpression of CXCL1, CXCL2, or CXCL3 (MGSAα, MGSAβ, or MGSAγ) in immortalized melanocytes results in tumor formation (11 , 12) . CXCL8 has also been shown to enhance the growth, invasiveness, and metastatic potency of melanoma cells by a way of autocrine mechanism in mouse models (10 , 13) and in human melanomas in vivo (14) .

The CXC chemokine, MGSAα/GROα or CXCL1, is highly secreted into the culture of human Hs294T malignant melanoma cells. CXCL1 protein was first purified and characterized in our lab (15 , 16) and was subsequently found to be the same as the growth-regulated gene, GRO (17) . We now know that CXCL1 or GRO is a member of the CXC family of chemotactic cytokines and is chemotactic for cells expressing its receptor, CXCR2, including neutrophils, keratinocytes, melanocytes, and dendritic cells (18, 19, 20, 21, 22) . Since then, two additional genes encoding proteins (CXCL2 and CXCL3) with high homology to CXCL1 have been identified (23, 24, 25) , and these chemokines are aberrantly expressed in viral, inflammatory, and neoplastic disease (26 , 27) , especially in melanoma cells but not in normal melanocytes (12 , 28, 29, 30) .

Aberrant overexpression of CXCL1 has been implicated in transformation and melanoma tumor progression both in vivo and ex vivo (12 , 28 , 31, 32, 33, 34) . During melanoma tumor progression, both the CXCL1 mRNA and protein levels are deregulated (34) . In Hs294T melanoma cells, CXCL1 but not CXCL2 or CXCL3 is shown to exhibit high constitutive expression (35) . The constitutive basal transcription appears to occur through activation of the NF-κB element in the proximal 5′ regulatory region of the gene (35, 36, 37) , although additional regulatory components are likely responsible for the selective constitutive expression of CXCL1 but not CXCL2 and CXCL3.

NF-κB is a heterodimeric transcription factor that is predominantly composed of M_r 65,000 and M_r 50,000 subunits of the Rel family. NF-κB modulation of gene expression is required for inflammatory response, the immune response, and cell growth and differentiation (38, 39, 40) . In resting cells, NF-κB is largely retained in the cytoplasm by a family of inhibitory proteins known as IκB (mainly α, β, and ε), which contain multiple ankyrin repeats that interact with NF-κB to mask its nuclear translocation signal and sequester NF-κB proteins in the cytoplasm (41, 42, 43) . The IκB (α, β, and ε) proteins contain two serine residues at the NH₂ terminus that regulate protein stability (44) . IκBα and IκBβ also contain a COOH-terminal PEST domain that contributes to basal protein turnover (45) . Upon stimulation of cells by such cytokines as TNF-α, IL-1, and lipopolysaccharide, serine residues in the NH₂ terminus of IκB are phosphorylated. The IκB protein is then ubiquitinated and degraded by the 26S proteasome, freeing the NF-κB proteins for translocation to the nucleus for subsequent transactivation of NF-κB responsive genes (43) . We have demonstrated that the increased IκBα degradation results in the increased nuclear basal NF-κB activity in Hs294T cells, in contrast to ARPE cells, a variant of normal retinal pigment epithelial (36) .

A family of IκB kinases (IKKα, IKKβ, and IKKγ) have been found to exist in a high molecular weight cytoplasmic complex (M_r 600,000 and M_r 900,000; Refs. 46, 47, 48, 49 ). The kinase activity of IKKα and IKKβ is induced with cytokine challenge with a consequent phosphorylation of IκB proteins (50) . Within the high molecular weight IKK complex, two members of MAP3K family, MEKK1 and NIK, were identified (51 , 52) . These kinases modulate the cytokine induction of the IKKs. NIK preferentially phosphorylates IKKα (53) , and MEKK1 preferentially phosphorylates IKKβ (54) . In contrast to IKKα and IKKβ, IKKγ/NEMO/IKKAP1 and IKAP function as scaffold proteins and bind to NIK and IKKs to assemble them into an active kinase complex (55, 56, 57) . Upon cytokine stimulation, active IKK α/β phosphorylate IκBα as major mechanism for NF-κB activation. Our previous work has shown that during melanoma tumorigenesis, elevated endogenous IKK activity and increased degradation of IκBα lead to constitutive NF-κB activation and increased basal CXCL1 transcription in Hs294T malignant melanoma cells as compared with RPE cells. These events are implicated as a major molecular mechanism for melanocyte transformation (11 , 36 , 37) . However, it needs to be determined whether: (a) the molecular alterations in Hs294T cells were common in other melanoma-derived cell lines; (b) other chemokines in addition to CXCL1 are expressed in melanoma cells; (c) CXCL1 could directly regulate IKK activity in normal human epidermal melanocytes and blockage of IKK activity could reduce CXCL1 secretion in melanoma cells; and (d) antibody to neutralize CXCL1 produced in melanoma cell cultures could block IKK activity and cell proliferation.

Toward this aim, this study has used normal human epidermal melanocytes as control cells and investigated eight human melanoma cell lines (Hs294T, Sk Mel 2, Sk Mel 5, Sk Mel 28, WM 115, WM 164, WM 852, and A375) for the endogenous activation of IKK and the level of expression of the CXC chemokines, CXCL1, CXCL5, and CXCL8 (MGSAα/GROα, ENA-78, and IL-8, respectively). We found that all of the melanoma cells involved in this study exhibit high endogenous IKK activity, hyperphosphorylated IκBα, increased nuclear localization of p65/p50, and increased promoter activity of NF-κB, CXCL1, and AP-1. In addition, we determined that CXCL1 can up-regulate IKK activity in a concentration-dependent manner, and anti-CXCL1 inhibits the IKK activity and slows melanoma cell growth. There was diversity among the melanoma cell lines with regard to whether CXCL1, CXCL8, and/or CXCL5 were expressed endogenously. In contrast to chemokines, IL-1β was either undetected or detected in very low levels in the culture medium of the melanoma cell cultures. These data suggest that constitutive activation of IKK with resultant endogenous expression of chemokines is a common feature of human melanoma.

MATERIALS AND METHODS

Cells.

Melanoma cell lines, Hs294T, Sk Mel 2, Sk Mel 5, Sk Mel 28, WM 115, WM 164, WM 852, and A375, established from human melanomas were obtained from American Type Culture Collection or from Meenhard Herlyn (Wistar Institute). NHEMs were provided by the Tissue Culture Core of the Skin Disease Research Center. Melanoma cells were cultured in DMEM:Ham’s F-12 medium containing 10% FBS, 100 units/ml of penicillin, 100 μg/ml of streptomycin, 2 mm of l-glutamine, 100 μm of MEM Non-Essential Amino Acids (Life Technologies, Inc.), and 1 mm of sodium pyruvate (Sigma Chemical Co.). NHEM were cultured in 154 medium (Cascade Biologicals, Inc.) with 1× human melanocyte growth supplement (Cascade Biologicals, Inc.).

Antibodies.

Anti-IKKα (H-744), IKKβ (H-470), IκBα (C-21), actin (I-19), NF-κB p65 (C-20), and NF-κB p50 (NLS) antibodies were purchased from Santa Cruz Biotechnology and anti-phospho-IκBα (Ser-32) antibody was from New England Biolabs.

Western Blot Analysis.

Cytoplasmic extracts were prepared from cells cultured for 24 h in serum-free medium. Cells were mechanically released from tissue culture plates by scraping in cold PBS according to standard protocols. Cells were collected by centrifugation (800 × g), then resuspended in buffer A [10 mm HEPES (pH 7.9), 10 mm KCl, 0.1 mm EDTA, 0.1 mm EGTA, 0.1% NP40, 1 mm DTT, and 0.5 mm phenylmethylsulfonyl fluoride] with Complete protein inhibitors (Boehringer Mannheim). Nuclei were collected by microcentrifugation (10,000 × g) at 4°C, washed twice in buffer A, and then spun through 0.25 m sucrose. Nuclei were lysed by shaking vigorously in buffer B [20 mm HEPES (pH 7.9), 0.4 mm NaCl, 1 mm EDTA, 1 mm EGTA, 1 mm DTT, and 1 mm phenylmethylsulfonyl fluoride] at 4°C for 20 min, and the resulting nuclear extracts were cleared by microcentrifugation (10,000 × g) at 4°C for 10 min. Cleared nuclear extracts were collected in the supernatant, and protein content was determined by Bio-Rad assay. An aliquot containing 50 μg of protein from each cell lysate was resolved on 10% SDS-PAGE, transferred to nitrocellulose membrane (Bio-Rad Lab), blotted in TBS-T with 5% Carnation fat-free instant milk, probed with the appropriate antibodies, and visualized by enhanced chemiluminescence assay.

Immunoprecipitation and Kinase Assay in Vitro.

To analyze the IKK activity, 200 μg of cytoplasmic extracts prepared using the procedures described for Western analysis were incubated with 1 μg of both IKKα and IKKβ polyclonal antibodies (Santa Cruz Biotechnology) and 60 μg of Protein A/G agarose-conjugated beads for 3 h or overnight at 4°C. After washing with buffer C [50 mm HEPES (pH 7.0), 250 mm NaCl, 5 mm EDTA, and 0.1% NP40] twice and kinase buffer once, the beads were incubated with 20 μl of kinase buffer [20 mm HEPES (pH 7.4), 10 mm MgCl₂, 2 mm MnCl₂, 25 mm β-glycerophosphate (Sigma Chemical Co. G-6251), 4 mm NaF, 0.1 mm sodium orthovanadate, and 1 mm DTT] containing 100 μm ATP, 5 μCi of [γ⁻³²P]ATP, and 1 μg of bacterially expressed glutathione S-transferase protein fused to amino acids 1–54 of IκBα (GST-IκBα) as a substrate of the IκB kinase at 30°C for 30 min. An expression vector for GST-IκBα (amino acids 1–54) was generated by inserting human IκBα cDNA encoding amino acids 1–54 into the BamHI-EcoRI site of pGEX-2T (Pharmacia) as described (58) . The p65 kinase assay was performed as described (59) with slight modification. Briefly, the immunoprecipitated IKKα/β from cytosolic extracts was washed twice with buffer C and once with kinase buffer and incubated with 1 μg of GST-fused p65 (amino acids 354–551; kindly provided by Dr. Wataru Toriumi, Discovery Research Laboratory, Tanabe Seiyaku Co., Ltd., Osaka, Japan) in kinase buffer [20 mm HEPES (pH 7.4), 20 mm MgCl₂, 2 mm DTT, 20 mm ATP, 20 mm β-glycerophosphate, 0.1 mm sodium orthovanadate, and 5 μCi of [γ-³²P]ATP] at 30°C for 30 min. The reaction mixtures were resolved by SDS-PAGE, transferred to nitrocellulose membrane, and followed by autoradiography. The same membrane was used for Western blot using the polyclonal antibodies to IKKα and IKKβ to quantitate and normalize the kinase activities.

Transfection and Luciferase Reporter Activity Assay.

Melanoma cells and NHEM cells were seeded in a six-well plate, and the next day, 80% confluent cells were transfected with the respective constructs using the Effectene Transfection Reagent kit (Qiagen) according to the manufacturer’s protocol. The HIV-long terminal repeat-luciferase construct, containing two NF-κB binding sites, was the kind gift of Dr. Richard Gaynor (University of Texas Southwest Medical Center, Dallas, TX). The MGSA/GRO luciferase promoter construct included 350-bp of the CXCL1 promoter (−306 to +45) inserted into the pGL3-Luciferase reporter vector. The pRSV β-galactosidase reporter vector was purchased from Promega Corp. (Madison, WI). Extracts from cultured cells were prepared, and the luciferase and β-galactosidase were detected using a Dual-Light kit (Tropix) according to manufacturer’s instructions.

RT-PCR.

For RNA isolation and cDNA synthesis, total RNA was purified from 24-h, serum-free cultured melanoma cells and NHEM cells using Ultraspec RNA (Biotecx Laboratory, Inc.) according to manufacturer’s protocol. For removing DNA contamination, total RNA was processed using the Message Clean Kit (GenHunter Corp.). Briefly, 20 μg of RNA were digested with 10 units of DNase I at 37°C for 30 min, extracted with phenol:chloroform (3:1), precipitated with ethanol, washed with 70% ethanol, and finally dissolved in 20 μl of diethylpyrocarbonate-treated distilled water. To generate the first-strand cDNA from polyadenylated mRNA, the SuperScript Preamplification System (Life Technologies, Inc.) was used. Briefly, 0.5 μg of oligo-dT (12–18 bp) as a primer and 200 units of SuperScript-2 reverse transcriptase were reacted with 2 μg of total RNA for 50 min at 42°C in the presence of 0.5 mm deoxynucleotide triphosphates, 10 mm DTT, and 1× First Strand buffer.

PCR Amplification.

PCR was carried out in a reaction volume of 25 μl using bulk master mixes, except template cDNA was prepared for multiple reactions. Each PCR reaction consisted of 1.25 units of Taq DNA polymerase (Sigma Chemical Co.), 1× PCR buffer, 0.2 mm deoxynucleotide triphosphates (Promega Corp.), 1 μl of cDNA, 200 nm of each pair of specific primers (CXCL1 primer, sense ATGGCCCGCGCTGCTCTCTCC and antisense CTTAACTATGGGGGATGCAGG; CXCL8 primer, sense GCAGCTCTGTGTGAAGGTGCA and antisense AACTTCTCCACAACCCTCTG) and 100 nm of each of the primers for GAPDH as a reference gene. The PCR was conducted as follows: denature at 95°C for 4 min followed by 35 cycles of 95°C for 1 min, anneal primers 1 min at 54°C for CXCL1 or 50°C for CXCL8, 72°C for 3 min, and a final extension at 72°C for 10 min. After PCR amplification, 12 μl of the PCR product mixed with loading buffer were subjected to electrophoresis in a 2% agarose gel under 1× TAE buffer [40 mm Tris · acetate, 2 mm Na₂EDTA · 2H₂O (pH 8.5)]. The PCR product size for CXCL1 or CXCL8 is 282 and 224 bp, respectively. The CXC chemokine PCR product was compared with the GAPDH PCR product as an internal control for the same cDNA, using the same master mixture and polymerase dilution, at the same time. To rule out the possibility that contaminating DNA was responsible for positive RT-PCR results, reactions carried out without cDNA (using water and/or purified RNA without reverse transcriptase treatment) did not show any positive signal in PCR product.

[³H]Thymidine Incorporation Assay.

[methyl-³H]Thymidine was purchased from Amersham Pharmacia Biotech, and the procedures described previously for monitoring incorporation of radiolabeled thymidine into DNA (15) were slightly modified. Briefly, melanoma cells were placed into 24-well plates at a density of 10⁴ cells/well in 10% FBS medium for overnight to allow attachment. The following day, cells were washed with PBS and cultured in serum-free medium with addition of 800 ng/ml of either anti-CXCL1 monoclonal antibody (MAB275; R&D) or IgG1 (MAB002; R&D) as a control antibody for 36 h. [methyl-³H]Thymidine (2 μCi/well) was then added to cultured cells, and incubation was continued for an additional 8 h. Cultured cells were fixed and washed with a 3:1 mixture of methanol:ethanol and collected in 6 ml of Scintiverse fluid. Radioactivity incorporated into DNA was counted in a Beckman Model LS-3801 liquid scintillation counter. The intraassay coefficient of variation was ∼17%.

ELISA Assay.

Melanoma cells and NHEM cells were seeded on six-well plates and cultured to 80% confluence in DMEM/F12 medium containing 10% FBS. The monolayers were washed in serum-free culture medium and then incubated in serum-free medium for an additional 24 h at 37°C, at which time the supernatant was collected and cleared by centrifugation. Aliquots were then subjected to ELISA assay (R&D system) according to the manufacturer’s protocol. Cells were lysed in RIPA buffer and assayed for total protein (Bio-Rad). Chemokine levels were calculated as chemokine/cytokine concentration/mg protein in the lysate.

RESULTS

Aberrant Production of Chemokines in Melanoma Cells.

CXCL1, a 73-amino acid polypeptide of the CXC-chemokine superfamily, was initially described as an autocrine growth factor for melanoma cells (15) . To explore the possibility that deregulation of chemokine expression through endogenous activation of NF-κB is common in melanoma tumor progression, we compared the expression of CXCL1, CXCL8, CXCL5, and IL-1β in eight human melanoma cell lines and NHEM cells. The results in Table 1 ⇓ show that NHEM cells produce low levels of CXCL1 and CXCL8, but CXCL5 and IL-1β are not detectable. The protein levels of CXCL1 and CXCL8 were considerably elevated in most of the melanoma cells, except for melanoma cell line, WM 852, which expressed very low levels of CXCL1. However, in contrast to CXCL1 and CXCL8, CXCL5 and IL-1β were barely detected in most of the melanoma cell lines (Table 1) ⇓ .

Constitutive Activation of IκB Kinases in Melanoma Cells.

Generally, activation of NF-κB by cytokines involves transient activation of IKK. The recent findings of constitutively active IKK in adult leukemia cells infected with HTLV suggests that IKKs may play a potential role in lymphocyte transformation (58) . Our previous study has demonstrated that IKK activity was higher in malignant melanoma Hs294T cells than in RPE cells (37) . This led us to further explore whether the endogenous IKK activity existed in other melanoma cell lines as well. We assessed the levels of IKK activity in eight melanoma cell lines and compared that to the activity present in primary cultured normal human epidermal melanocytes. Using the protocols described in “Materials and Methods,” IKK activity was assessed after immunoprecipitation of IKKα and IKKβ, and the GST-IκBα (amino acids 1–54) as a substrate for the IκB kinase. The results showed that, compared with NHEM, the activities of IκB kinases were 3–14-fold higher in the melanoma cells as compared with the NHEM cultures (Fig. 1A) ⇓ . In contrast, the IKK protein expression was similar in the melanoma and NHEM cells (Fig. 1B) ⇓ .

Hyperphosphorylation of Cytoplasmic IκBα in Melanoma Cells.

The activated IKKs phosphorylate IκBα, the inhibitor of transcription factor NF-κB, at serine 32 and serine 36 (44) . We studied the expression of IκBα protein and its phosphorylation in eight different malignant human melanoma cell lines as compared with NHEM. After culturing 24 h in serum-free medium, 50 μg of cytosolic extracts were subjected to Western blot for IκBα with anti-IκBα polyclonal antibody (c-21; Santa Cruz Biotechnology) and with a monoclonal antibody specific for phospho-serine 32 of IκBα (New England Biolabs). The density of the blots was quantified by densitometric scanning, and the ratio of phospho-IκBα to IκBα was calculated. As expected, the constitutive phosphorylation of IκBα in melanoma cells was obviously higher than that of normal human epidermal melanocytes (Fig. 2A) ⇓ . However, the expression of IκBα protein appeared similar in melanoma cells as compared with NHEM (Fig. 2B) ⇓ , because of de novo protein synthesis after degradation of IκBα. Moreover, the phosphorylation of IκBα is almost parallel with the IKK activity in melanoma cells as compared with normal melanocytes. The hyperphosphorylation of IκBα within the NF-κB complex in melanoma cells, taken together with its rapid degradation, is consistent with our previous results obtained from Hs294T and RPE or ARPE cells (36 , 37) .

Nuclear NF-κB Accumulation in Melanoma Cells and p65 Phosphorylation.

In the superfamily of Rel, p65/p50 heterodimers play an important role in transcription after translocation to nuclei after phosphorylation and subsequent degradation of its inhibitor, IκBα. To study the nuclear transactivation of p65/p50 in melanoma cells, we prepared cytosolic and nuclear extracts from melanoma cells and NHEM cells cultured for 24 h in serum-free medium as described in “Materials and Methods.” A 50-μg aliquot of protein from each cell lysate was loaded per lane onto a 10% SDS-PAGE, transferred to nitrocellulose membrane, and immunoblotted with anti-NF-κB p65 and p50 polyclonal antibodies. Results revealed that both p65 and p50 protein were highly enriched in nuclei of melanoma cells as compared with NHEM cells, whereas p65/p50 proteins were almost equally expressed in the cytoplasm of both melanoma cells and NHEM (Fig. 3, A and B) ⇓ . Because the cytokine-inducible phosphorylation of p65 protein may enhance DNA binding of p65 containing NF-κB and transactivating capacity of NF-κB (60) , we examined the p65 phosphorylation by IKKα/β in melanoma cells and normal melanocytes. On the basis of results from the in vitro p65 kinase assay, we found that the cytoplasmic p65 can be potentially phosphorylated by activated IKK in all melanoma cell lines involved in this experiment (Fig. 3C) ⇓ .

CXC Chemokine mRNA Is Constitutively Overexpressed in Melanoma Cells.

Because of constitutive interaction of NF-κB, transcriptional up-regulation of CXCL1 is expected to occur in other melanoma cells, in addition to the Hs294T cells (35) . To clarify this issue, total RNA was extracted from melanoma cells, and NHEM cells were cultured for 24 h in serum-free medium. DNA contamination was removed from RNA using DNase I digestion. The RT-PCR was conducted as described in “Materials and Methods” for CXCL1 mRNA in melanoma cell lines and normal human epidermal melanocytes using the reference gene GAPDH as a internal control. As expected, the expression of mRNA of both CXCL1 and CXCL8 was barely detectable in NHEMs and relatively high in most melanoma cells (Fig. 4) ⇓ .

Up-Regulation of Activities of the NF-κB, CXCL1/350 bp, and AP-1 Promoter Luciferase Reporters in Melanoma Cells.

To further explore the endogenous activation of chemokine transcription response elements in melanoma cells as compared with NHEM cells, 80% confluent melanoma cells and NHEMs were transfected with 1.0 μg of the NF-κB, CXCL1/350 bp, and AP-1 promoter-luciferase constructs and 0.5 μg of the pRSV β-galactosidase reporter vector using Effectene Transfection Reagent kit (Qiagen). Extracts from cultured cells were prepared, and the activities of luciferase and β-galactosidase were detected using the Dual-Light kit (Tropix). Results in Table 2 ⇓ show that constitutive induction of promoter activities of NF-κB, CXCL1, and AP-1 by melanoma cells is higher than in that of NHEM cells.

CXCL1 Induces IKK Activity in NHEM in a Dose-dependent Manner.

Although constitutive IKK activity in association with NF-κB nuclear localization and expression of CXCL1 protein is prevalent in most melanoma cell lines, it is not known whether CXCL1 can directly induce IKK activity in normal human epidermal melanocytes. To resolve this issue, we treated serum-starved, 80% confluent cultures of NHEMs for 24 h with CXCL1. The cytoplasmic IKKα/β complex was immunoprecipitated with anti-IKKα/β polyclonal antibodies (Santa Cruz Biotechnology). IKK activity was detected using the specific substrate of recombinant GST-IκBα (amino acids 1–54) at 30°C for 30 min in the presence of [γ-³²P]ATP. Interestingly, CXCL1 can induce IKK activity in NHEM cells in a dose-dependent manner (Fig. 5) ⇓ .

Down-Regulation of Expression of CXCL1 Protein in Melanoma Cells by Inhibition of IKK Activity.

The expression of CXCL1 in normal human epidermal melanocytes in response to IL-1β can be inhibited by IKKβ inhibitor sulindac (data not shown). To determine whether inhibition of IKKβ activity in malignant melanoma cells by the specific inhibitor, sulindac, could affect the constitutive IKK activity and CXCL1 protein production in melanoma cell culture, aliquots of the serum-starved cultures of Sk Mel 5 cells were treated with different concentrations of sulindac (Sigma Chemical Co.; 0, 0.1, 1.0, and 2.0 mm) for 12 h. The cytoplasmic IKKα/β complex was immunoprecipitated with 1 μg of anti-IKKα/β polyclonal antibodies (Santa Cruz Biotechnology). IKK activity was detected in vitro using the specific substrate of recombinant GST-IκBα (amino acids 1–54) as described in “Materials and Methods.” The results in Fig. 6A ⇓ show that sulindac can reduce IKK activity in melanoma cells in a dose-dependent manner (upper panel), whereas cells treated with 2 mm sulindac exhibited partial apoptosis and decreased IKK protein expression (lower panel). To analyze CXCL1 protein secretion by malignant melanoma cells during the process of sulindac inhibition of IKK activation, Sk Mel 5 cells were plated in six-well plates and treated with the same concentration of sulindac as above for 12 h. CXCL1 concentrations in the supernatant were measured by ELISA. Results in Fig. 6B ⇓ showed that the CXCL1 levels in the melanoma cells treated with 0, 0.1, 1.0, and 2.0 mm sulindac exhibited 5670 ± 240, 1450 ± 155, 500 ± 56, and 490 ± 14 pg/ml, respectively, which corresponded to the decrease of IKK activity inhibited by the increased sulindac treatments.

Anti-CXCL1 Reduces IKK Activity and Proliferation of Melanoma Cells.

It not clear whether correlation between CXCL1 expression and IKK activity in both normal human epidermal melanocytes and malignant melanoma cells results from a direct activation of IKK in response to CXCL1 secretion. To explore this possibility, CXCL1 monoclonal antibody was used to neutralize endogenous CXCL1 secreted from the cultured melanoma cells to allow us to examine whether CXCL1 antibody could block the IKK activity and inhibit cell proliferation. Sk Mel 5 cell cultures were treated 24 h with 0, 160, 800, and 1600 ng/ml of CXCL1 antibody, respectively, corresponding to 0-, 10-, 50-, and 100-fold CXCL1 production by this cell line in 24 h. When the concentration of CXCL1 antibody reached 800 ng/ml, IKK activity was shown to be obviously reduced (Fig. 7A) ⇓ , which is identical to the effect of 40-fold excess antibody on induction of IKK activity by CXCL1 in NHEMs (data not shown). Furthermore, the IKK activity of all eight melanoma cell lines was examined using the dosage of 800 ng/ml antibody treatment. Results in Fig. 7B ⇓ show down-regulation of IKK activity in all but the WM 852 cell line. To determine whether anti-CXCL1 is correlated with cell proliferation of malignant melanoma, melanoma cells were plated in 24-well plates at a density of 10⁴ cells prior to treatment with 800 ng/ml of anti-CXCL1 monoclonal antibody or the same amount of IgG1 antibody as a control for each cell line in serum-free medium. Serum-free culture medium containing CXCL1 antibody or control antibody was renewed every other day, and cells were counted on the seventh day. Results in Fig. 7C ⇓ show that compared with IgG1 treatment, melanoma cells treated with anti-CXCL1 antibody resulted in approximately 30–40% inhibition of cell proliferation with the exception of the WM 852 cell line, which was not significantly inhibited. Cell growth in the presence or absence of CXCL1 antibody was also assessed by monitoring [³H]thymidine incorporation into DNA. Results from the [³H]thymidine incorporation assay correspond well to the results from the cell number assay (Fig. 7C) ⇓ .

DISCUSSION

We have demonstrated previously that the Hs294T malignant melanoma cells exhibit a constitutive activation of the NF-κB complex and express a high level of CXCL1 as compared with RPE cells (36 , 37) . To extend this finding, we studied eight cell lines derived from human malignant melanomas to monitor expression of CXCL1, CXCL5, and CXCL8 and the constitutive activation of NF-κB in malignant melanoma cells as compared with primary cultures of normal human epidermal melanocytes. Melanoma cells studied here were cultured under serum-free conditions to investigate the deregulation of signal transduction in melanoma cells, because serum contains growth factors that can induce activation of the IKK-NF-κB pathway. These data showed that most of the eight melanoma cell lines, as compared with NHEM, exhibit constitutive activation of the pathway leading to chemokine transcription, resulting in deregulation of CXCL1 and CXCL8 but not CXCL5 (with exception of Hs294T). The regulation of chemokine expression is complex, involving both transcription, mRNA stability, translation, and secretion, often resulting in varied expression of CXC chemokines among the melanoma cell lines. For example, CXCL5 is overexpressed in Hs294T but not the other melanoma cell lines, or RPE cells. At this time, we do not have an explanation for elevation in expression of CXCL5 in Hs294T melanoma cells. The promoter for CXCL5 remains poorly characterized.

In normal resting cells, the activity of the transcription factor NF-κB is tightly controlled in the cytoplasm by the inhibitory proteins, IκBs. In the IκB family, IκBα plays a major role in regulating p65/p50 heterodimers of NF-κB because of the presence of a PEST sequence (P-E-D-S) in the COOH terminus of IκBα, which functions as a degradation signal to target the phospho-IκBα protein for rapid degradation (45) . Recent studies have identified two cytokine-inducible IκB kinases, termed IKKα and IKKβ, that target IκBα for degradation via phosphorylation at Ser-32 and Ser-36 (44) . In the melanoma cells used in these experiments, both IκB kinase α and β exhibit higher activities than in NHEM cells, resulting in hyperphosphorylation of IκBα followed by rapid degradation of IκB in melanoma cells. This leads to translocation of p65/p50 NF-κB heterodimers to the nuclei for increased transcription of NF-κB response genes. In agreement with the HTLV-I-infected T-cell system, our finding here further supports the concept that constitutive IKK activity and phosphorylation of IκBα are common mechanisms for the persistent NF-κB activation in the transformed cells. In fact, the cellular NF-κB signal pathway is phosphorylation dependent. In addition to phosphorylation of IκBα leading to its degradation and subsequently nuclear translocation of p65/p50, inducible phosphorylation of p65, by either TNF-α in HeLa cells and B cells (61) or by acetaldehyde in HEPG2 (60) , is reported to enhance the DNA binding activity of NF-κB p65 and the transactivating capability of NF-κB. Sakurai et al. (59) demonstrated that upon stimulation of HeLa cells by TNF-α, IKKs directly phosphorylate p65 in vivo and in vitro. Here we reveal that p65 is readily phosphorylated by IKK immunoprecipitated from melanoma cells but is barely phosphorylated by the IKK complex immunoprecipitated from NHEMs. Moreover, both p65 and IκBα become strongly phosphorylated by IKKα/β with the same kinetics (data not shown). Thus, our data suggest that p65 phosphorylation by IKK might contribute to the nuclear NF-κB activity in melanoma cells. This endogenous NF-κB in the tumor cells in turn allows escape from apoptosis and thus contributes to the malignant phenotype (62) .

Interestingly, all of the melanoma cell lines involved in this study exhibit constitutive IKK activity, hyperphosphorylation of IκBα, and persistent translocation of nuclear p65/p50. These results correlated with NF-κB luciferase reporter assay results. However, differences in the degradation rate of phospho-IκBα in the Sk Mel 2 cells may contribute to the phenomenon of high IKK activity accompanied by accumulation of only a low level of phosphorylated IκBα in this cell line. Specifically, CXCL1 and CXCL8 mRNA were highly expressed in all melanoma cells but were barely detectable in normal human epidermal melanocytes. These data are in general agreement with the high level of CXCL1 and CXCL8 protein produced in most melanoma cell lines involved in this study. These results indicate that activation of NF-κB is sufficient to induce CXCL1 expression, but other variables may also affect chemokine expression, such as mRNA stability, translation, and posttranslational processing, resulting in failure to observe a simple linear correlation between fold induction of IKK activities, NF-κB activation, CXCL1 mRNA, and secreted protein levels between melanoma cell lines. For example, induction of IL-8 transcription requires coordinate activation of NF-κB and AP-1 or CAAT/enhancer binding protein. In several of the melanoma cell lines studied here, AP-1 is endogenously activated concomitantly with NF-κB. Thus, similar to CXCL1, CXCL8 is also involved in maintenance of the malignant phenotype of some of melanoma cell lines (10 , 13) .

To explore the possibility that CXCL1 might directly activate NF-κB as a mechanism in the tumorigenesis of human melanoma, normal human epidermal melanocytes were stimulated with exogenous CXCL1, and effects on NF-κB activation were monitored. We observed that CXCL1 induces IKK activity in normal human epidermal melanocytes in a dose-dependent fashion, demonstrating that CXCL1 itself may participate in the maintenance of constitutive NF-κB activity. To obtain more insight into the causal relationship between NF-κB signaling and production of the CXCL1 oncoprotein, the constitutive IKK activity in melanoma was reduced by sulindac, an IKKβ inhibitor (63) , and this decrease in IKK activity resulted in decrease in CXCL1 production. These data suggested that IKKβ plays a major but not a exclusive role in modulation of CXCL1 expression in normal human epidermal melanocyte culture followed CXCL1 or IL-β induction (data not shown). To further explore the role that CXCL1 plays as an autocrine factor on IKK-NF-κB signaling, anti-CXCL1 antibody was added to various melanoma cell cultures to block the binding of secreted CXCL1 to the CXCR2 receptor. This resulted in down-regulation of IKK activity in the SK Mel 5 melanoma cell line in a concentration-dependent manner. Treatment of ligand with 50-fold excess of CXCL1 antibody partially blocked IKK activity and inhibited cell proliferation in all of the melanoma cell lines, with the exception of the WM 852 cell line, which showed only slight inhibition in proliferation. The low level of CXCL1 in WM 852 cells may indicate that other chemokines/cytokines are involved in the IKK-NF-κB activation. Genetic differences among melanoma cell lines may also contribute to the differences between WM 852 and the other melanomas (64) . Anti-CXCL8 also partially inhibited the growth of melanoma cells (data not shown). Even more effective inhibition of cell proliferation occurred by blockage of the CXCR2 receptor with 100 mm SB225002, a highly selective CXCR2 inhibitor (data not shown). These data support our hypothesis that autocrine effects of CXCL1 and CXCL8 participate in melanocyte transformation and maintenance of the malignant state.

NF-κB is a crucial transcription factor involved in the control of gene expression of several chemokines, cytokines, and enzymes that protect cells from death (65, 66, 67) or from apoptosis (68, 69, 70, 71) and for regulating several tumorigenic and angiogenic factors (72) . Data provided here further demonstrate that the constitutive expression of NF-κB occurs not only in Hs294T melanoma cells but also in the malignant melanoma cell lines of Sk Mel 2, Sk Mel 5, Sk Mel 28, WM 115, WM 164, WM 852, and A 375. Alterations in NF-κB activity in these melanoma cells are strongly correlated with melanocyte transformation. Our melanoma model is similar to the leukemia model, where the tax oncoprotein encoded by HTLV-I stimulates the constitutive nuclear expression of transcription factor NF-κB. This activated NF-κB regulates antigen-directed T-cell proliferation (39 , 73) , leading to the loss of cell cycle control and development of an aggressive malignant adult T-cell leukemia (74) . In the melanoma model, overexpression of CXCL1 and/or CXCL8 is highly associated with the constitutive activation of NF-κB and malignant phenotype. Antisense studies of the transcriptional regulation of NF-κB further support the role of NF-κB in the pathogenesis of cancer. Similar to HTLV-I-infected T cells, mammary carcinoma, head and neck squamous cell carcinoma, and melanoma all exhibit constitutive activation of NF-κB (37 , 75, 76, 77, 78) . The experiments described here enhance our understanding of the mechanism by which CXCL1 and CXCL8 expression increases as melanocytes progress to malignant melanoma (29 , 32 , 79) . Because both CXCL1 and CXCL8 may serve as autocrine growth factors for several melanoma cell lines (10 , 13 , 16 , 80) , antibodies to both CXCL1 and CXCL8, or inhibitors which block the CXCR2 receptor, inhibit the autocrine loop and slow the proliferation of melanoma cells. These findings strongly indicate that development of better strategies for selectively inhibiting NF-κB activity in melanoma tumor cells will have significant therapeutic benefit.