Tris (Dibenzylideneacetone) Dipalladium, a N-Myristoyltransferase-1 Inhibitor, Is Effective against Melanoma Growth In vitro and In vivo

Melanoma is one of the most common solid tumors and is notoriously difficult to treat. Recently, constitutive activation of several signaling pathways has been shown in melanoma. Many melanomas carry mutations in B-raf, which cause constitutive activation of mitogen-activated protein kinase (MAPK; refs. 1, 2). Even melanomas that do not carry activated B-raf show activation of MAPK, and constitutive expression of activated MAPK kinase is sufficient to transform melanocytes to melanoma (3–5). Other pathways that are known to be activated in advanced melanoma include phosphoinositol 3-kinase (PI3K)/Akt and nuclear factor-κB (6–10). All of these pathways confer survival and proliferative advantages to melanoma, such as induction of angiogenic factors, including vascular endothelial growth factor (VEGF), interleukin-8, survivin, IAP, and mcl-1 (11–13).

Platinum compounds have been the mainstay of many solid tumor regimens, especially testicular cancer. However, platinum compounds, including cisplatin and carboplatin, have also shown activity in melanoma and have been incorporated into melanoma treatment regimens (14). Other inhibitors, such as sorafenib, a B-raf inhibitor, have had modest effects on melanoma with B-raf mutation despite robust inhibition of B-raf (15). This may be due to the ability of aggressive tumors to switch signaling pathways (16). We have observed this phenomenon in Burkitt's lymphoma, in which MAPK is activated when nuclear factor-κB is down-regulated (17). Similarly, inhibition of nuclear factor-κB with Velcade has had modest effects in melanoma (18, 19).

In our screens for angiogenesis inhibitors, we have identified a small-molecule palladium complex, which has structural similarities to curcumin and chalcones, compounds with known chemopreventive activity (20, 21). Although chemopreventive agents are effective against preneoplastic lesions in mice and man, they are less effective against established tumors (22, 23). Analysis of Tris (dibenzylideneacetone) dipalladium (Tris DBA)–treated melanoma cells by gene array revealed reduction of N-myristoyltransferase-1 (NMT-1), which was confirmed by quantitative reverse-transcription PCR (RT-PCR). Myristoylation done by NMT-1 is required for most membrane-based signaling molecules. C-src is a candidate molecule that requires myristoylation for optimal activity. Tris DNA reduced expression of c-src, which is a substrate of NMT-1. Consistent with inhibition of c-src/NMT-1, Tris DBA inhibited downstream signaling pathways, including MAPK kinase, PI3K, and Stat-3. Tris DBA has activity in vivo against A375 and B16 melanoma in vivo. Further preclinical evaluation of Tris DBA is warranted.

Cells. B16 melanoma cells were cultured in DMEM (1,000 mg glucose/L; Sigma-Aldrich) supplemented with 10% fetal bovine serum, l-glutamine (14 mL/L), and antibiotic/antimycotic (14 mL/L; Sigma-Aldrich). A375 cells were cultured in DMEM (4,500 mg glucose/L; Sigma-Aldrich) supplemented with 10% fetal bovine serum, l-glutamine (14 mL/L) and antibiotic/antimycotic (14 mL/L; Sigma-Aldrich).

Cell proliferation assays. To evaluate the potential of Tris DBA as an antitumor agent, a proliferation assay was done using B16 cells (Fig. 1A) and A375 cells (Fig. 1B). The assay was done according to the method described previously by the Arbiser laboratory (24, 25). Ten thousand cells were plated per well in a 24-well plate. After incubation for 24 h at 37°C and 5% CO₂, the cells were treated at 2.5, 5, 10, 15, and 20 μg/mL from a stock solution of 10 mg/mL Tris DBA dissolved in DMSO, also used as the control. Experiments were done in triplicate. The cells were allowed to incubate for an additional 24 h and then counted using a Coulter counter.

Western blot analysis. For signal transduction analysis, B16 cells, untreated and treated with 10 μg/mL Tris DBA at timed intervals, were lysed in NP-40 lysis buffer (1% NP-40, 150 mmol/L NaCl, 10% glycerol, 20 mmol/L HEPES, 1 mmol/L phenylmethylsulfonyl fluoride, 2.5 mmol/L EDTA, 100 μmol/L Na₃VO₄, and 1% aprotinin). The lysate was spun in microfuge, and the pellet was discarded. Protein concentration of the supernatant was determined by the Eppendorf BioPhotometer. Samples were treated with Laemmli sample buffer and heated to 90°C for 5 min before SDS-PAGE (National Diagnostics) and was transferred to nitrocellulose membranes. The membranes were then blocked with 5% nonfat dry milk in 10 mmol/L Tris/0.1% Tween 20/100 mmol/L NaCl and were subsequently incubated with p42/44 MAPK antibody, phospho-p44/42 MAPK (Thr²⁰²/Tyr²⁰⁴) antibody, phospho-Akt (Ser⁴⁷³), and phospho-p70 S6 kinase (Thr⁴²¹/Ser⁴²⁴) antibody (Cell Signaling Laboratories). Monoclonal anti-β-tubulin antibody (Sigma) was used as a loading control and detected using horseradish peroxidase–conjugated secondary antibody. The immunoreactive bands were visualized by enhanced chemiluminescence (Amersham Biosciences).

Quantitative RT-PCR for and VEGF in B16 and A375 cells and NMT-1 in A375 cells treated with vehicle control and 10 μg/mL Tris DBA. B16 and A375 cells were seeded equally into two T-25 flasks each and 24 h later were treated with 0 and 10 μg/mL Tris DBA (Aldrich) in DMSO for 24 h. RNA was extracted and purified using Qiagen RNeasy mini kit and measured using spectrophotometer (Perkin-Elmer UV/visible). RNA (1 μg) was used for DNase amplification (Invitrogen) followed by first-strand synthesis for RT-PCR (SuperScript). Optical Reaction Plate (96 wells; 7500 Fast Real-Time PCR) was used for the RT-PCR. The template (2.5 μL), which had been diluted 1:10 in cross-linked water, was used in each well and the experiment was done in triplicate. VEGF-a (Applied Biosystems Taqman Gene Expression Assay, Mm00437304_ml), NMT-1 (Applied Biosystems Taqman Gene Expression Assay, Hs00221506_m1) and 18S (Applied Biosystems Taqman Gene Expression Assay, Hs99999901_s1) primers were used along with cross-linked molecular-grade water (Cellgro) and master mix [Applied Biosystems Taqman Fast Universal PCR Master Mix (2×)]. Reaction was set up at the 7500 Applied Biosystems Reader for Absolute Quantification for 96-well plate. Ct values were analyzed by ΔΔCt method, and the SE was calculated (see Fig. 4).

NMT-1 assay. [³H]Myristic acid (39.3 Ci/mmol) was obtained from NEN Life Science Products. Pseudomonas acyl-CoA synthetase and coenzyme A were obtained from Sigma-Aldrich Canada. The peptide based on the NH₂-terminal sequence of the type II catalytic subunit of cyclic AMP–dependent protein kinase (GNAAAAKKRR) was obtained from Alberta Peptide Institute, University of Alberta. The expression and purification of recombinant human NMT-1 were undertaken as described previously (26). The NMT activity was measured as described previously (27, 28). For the standard enzyme assays, the reaction mixture contained 0.4 μmol/L [³H]myristoyl-CoA, 50 mmol/L Tris-HCl (pH 7.8), 0.5 mmol/L EGTA, 0.1% Triton X-100, 500 μmol/L synthetic peptide, and purified human NMT-1 in a total volume of 25 μL. The reaction was initiated by the addition of radiolabeled [³H]myristoyl-CoA and incubated at 30°C for 10 to 30 min. The reaction was terminated by spotting aliquots of incubation mixture onto P81 phosphocellulose paper discs and drying them under a stream of warm air. The P81 phosphocellulose paper discs were washed in three changes of 40 mmol/L Tris-HCl (pH 7.3) for 90 min. The radioactivity was quantified in 7.5 mL Beckman Ready Safe Liquid Scintillation mixture using a Beckman Liquid Scintillation Counter. One unit of NMT activity was expressed as 1 pmol myristoyl-peptide formed/min/mg protein. The human NMT-1 inhibitory assay was carried out using Tris DBA according to the method described earlier (Fig. 5A; ref. 28). A control experiment was done in the absence of Tris DBA and the human NMT-1 activity was considered as 100%.

In vivo tumor growth. To determine if a compound that inhibits melanoma growth in vitro would also inhibit tumor formation in vivo, we injected 1 million B16 melanoma cells and 1 million A375 cells s.c. into six nude mice, respectively. Beginning 2 days later, the mice received i.p. injections three times per week of either Tris DBA or control. Tris DBA (40 mg/kg/d) was suspended in 0.3 mL peanut oil, and control was 0.3 mL peanut oil alone. Neither local nor systemic toxicity was observed in any of the nude mice as a result of treatment. A total of six rounds of injections were given over a period of 2 weeks, after which the mice were sacrificed due to overwhelming tumor burden in the control group. Animals were euthanized on day 15, secondary to tumor burden in the control animals. Graph represents average tumor volume (mm³) in each of the two groups with controls (Fig. 5) with bars as SE.

Tris DBA inhibits melanoma proliferation. Tris DBA was used to treat at five different concentrations on B16 and A375 cells.¹ Treatment with Tris DBA for 24 h significantly reduced the number of viable cells. A 10 μg/mL concentration of Tris DBA resulted in a 99% decrease in B16 cell count (Fig. 1A) and a 96% decrease in A375 cell count (Fig. 1B).

Tris DBA inhibits activation of MAPK, Akt, Stat-3, and phospho-S6 kinase in melanoma cells. To examine this effect, we tested 10 μg/mL Tris DBA on MAPK and Akt signaling pathway using B16 melanoma cells. Using a time-course pattern, we found that it inhibited the phosphorylated forms of both downstream (Fig. 2A). We further tested Tris DBA on human melanoma cell line and found that it inhibits phosphorylated forms of S6 kinase downstream (Fig. 2B) and down-regulates phospho-Stat-3 at shorter time intervals (1 and 4 h) than after 24 h of treatment (Fig. 2C).

Inhibition of human NMT-1 activity and NMT-1 expression by Tris DBA. Gene array was done on A375 melanoma cells treated with Tris-DBA palladium compared with control cells. Among the genes that were strongly down-regulated as a result of treatment was NMT-1. We confirmed this down-regulation by RT-PCR (Fig. 3A).

When purified human NMT was incubated in the presence of various concentrations of Tris DBA, the human NMT was inhibited in a concentration-dependent manner, with maximal inhibition at a concentration of 2.5 ± 0.97 μmol/L and half-maximal inhibition at 1.0 ± 0.26 μmol/L (Fig. 3A). Further testing of 10 μg/mL Tris DBA on NMT-1 expression on A375 cells, we found an 80% decrease compared with control (Fig. 3B and C). Thus, Tris DBA is a novel and potent inhibitor of NMT-1 activity.

Tris DBA inhibits VEGF expression in murine and human melanoma cells in vitro. We tested 10 μg/mL Tris DBA on VEGF expression on B16 cells and found a 60% decrease in VEGF expression compared with control. We further tested 10 μg/mL of the same compound on A375 cells and found an 80% decrease compared with control (Fig. 4).

Tris DBA inhibits B16 and A375 melanoma growth in vivo. To determine if compounds that inhibit VEGF, phosphorylated forms of MAPK, Akt, Stat 3, and p70 S6 kinase in vitro would affect melanoma formation in vivo, we injected 1 million B16 cells s.c. into six nude mice and 1 million A375 cells s.c. into six nude mice. I.p. treatment with Tris DBA resulted in a 97% decreased tumor volume compared with control when using the B16 murine melanoma model. In the A375 human melanoma model, there was a 65% reduction in tumor volumes compared with control (Fig. 5). Neither local nor systemic toxicity was observed in any of the nude mice as a result of treatment.

Melanoma is a common solid tumor notorious for its high rate of metastasis and resistance to chemotherapy and radiation. Several factors may account for the resistance of melanoma to current therapies. First, melanomas are derived from melanocytes, specialized neural crest cells that are specialized to produce melanin. The production of melanin results in the generation of toxic reactive oxygen species and cytotoxic phenol derivatives; thus, melanocytes are equipped with mechanisms to resist these insults. Recently, the microopthalmia gene (MITF), which is a master transcriptional switch of melanocytes, has been shown to possess antiapoptotic activity and is found in metastatic lesions at a high frequency (29, 30). Second, multiple signaling pathways are activated in melanoma. B-raf is mutated in many melanomas, resulting in constitutive activation of MAPK signaling (1, 29). N-ras is also mutated frequently in melanoma, resulting in activation of MAPK, PI3K/Akt signaling, and S6 kinase activation (30, 31). Although B-raf and constitutive MAPK activation is sufficient to cause transformation of melanocytes into melanoma (3, 4), other signal transduction events are frequently observed in B-raf mutant melanomas, such as loss of the tumor suppressor PTEN (31, 32). The consequences of PTEN loss is activation of PI3K/Akt activation.

Multiple regimens have been tried for the treatment of locally advanced and metastatic melanoma. Initial trials several decades ago used agents such as hydroxyurea, and more recent agents used against melanoma include dacarbazine and platinum-based therapies including cisplatin and carboplatin. Other therapies, including biochemotherapy, have included interleukin-2 infusion and infusion of lymphocytes, which are present in melanoma lesions and have been expanded ex vivo (33). All of these therapies have had modest success in a minority of patients, but with significant toxicity, including pulmonary leak syndrome (34–38). Currently, IFN-α is employed in high-risk patients, and prolonged therapy results in a 10% long-term survival benefit.

Targeted therapies have been attempted in melanoma. Sorafenib was developed as a B-raf inhibitor based on the observation that B-raf mutation is common in melanoma. However, results from initial trials of sorafenib in melanoma have been disappointing (14). Everolimus has also been tried against human melanoma and has not been successful as a single agent (39). Current knowledge of signaling may provide an explanation of why previous therapies have failed. PI3K activation has been shown to mediate against extrinsic pathways of apoptosis, which include apoptosis due to TRAIL, tumor necrosis factor-α, and IFNs (10). Monotherapies of these cytokines may be frustrated in the face of PI3K activation. Similarly, apoptosis induced by tumor-infiltrating lymphocytes may be frustrated by PI3K activation. PI3K also activates VEGF expression; in addition to stimulating angiogenesis, VEGF inhibits dendritic cell function, impairing immune responses to melanoma (40–45).

Targeting MAPK as monotherapy in melanoma is clearly insufficient to eliminate melanoma in most patients. MAPK is activated in a majority of human melanomas, including those that lack B-raf mutation (3). In a previous study of human melanomas, we showed that a subset of advanced melanomas had decreased MAPK activation, implying that additional signaling pathways are operative in vivo (3). Further support of this hypothesis is our previous finding that treatment of EBV-induced Burkitt's lymphomas with antioxidants resulted in compensatory MAPK activation (17). It is likely that treatment of melanoma patients with sorafenib results in compensatory activation of non-MAPK pathways. Similarly, mammalian target of rapamycin inhibition due to rapamycin and derivatives has been shown to result in compensatory Akt activation (46). Tris DBA has the advantage that it inhibits several pathways required for melanoma tumorigenesis, including MAPK activation, PI3K/Akt activation, Stat-3 activation, and S6 kinase activation, and down-regulates NMT-1 at the level of enzyme activity and the level of mRNA. Down-regulation of these pathways may lead to diminished transcription of NMT-1. Although no drug is likely to be completely effective as monotherapy in melanoma, Tris DBA is well tolerated systemically in mice and has a novel profile of action compared with other clinically used chemotherapeutic agents. Its ability to inhibit PI3K activation may enhance the activity of cytokines, which require Akt inactivation for optimal activity, and may enhance the activity of other chemotherapeutic agents. Our studies provide a rationale for the further investigation of Tris DBA in the treatment of malignant melanoma.

J.L. Arbiser has a patent on Tris (dibenzylideneacetone) dipalladium.