|Home | About | Journals | Submit | Contact Us | Français|
The Raf/MEK/ERK pathway is an important mediator of tumor cell proliferation and angiogenesis. Here, we investigated the growth-inhibitory and antiangiogenic properties of PD0325901, a novel MEK inhibitor, in human melanoma cells. PD0325901 effects were determined in a panel of melanoma cell lines with different genetic aberrations. PD0325901 markedly inhibited ERK phosphorylation and growth of both BRAF mutant and wild-type melanoma cell lines, with IC50 in the nanomolar range even in the least responsive models. Growth inhibition was observed both in vitro and in vivo in xenograft models, regardless of BRAF mutation status, and was due to G1-phase cell cycle arrest and subsequent induction of apoptosis. Cell cycle (cyclin D1, c-Myc, and p27KIP1) and apoptosis (Bcl-2 and survivin) regulators were modulated by PD0325901 at the protein level. Gene expression profiling revealed profound modulation of several genes involved in the negative control of MAPK signaling and melanoma cell differentiation, suggesting alternative, potentially relevant mechanisms of action. Finally, PD0325901 inhibited the production of the proangiogenic factors vascular endothelial growth factor and interleukin 8 at a transcriptional level. In conclusion, PD0325901 exerts potent growth-inhibitory, proapoptotic, and antiangiogenic activity in melanoma lines, regardless of their BRAF mutation status. Deeper understanding of the molecular mechanisms of action of MEK inhibitors will likely translate into more effective treatment strategies for patients experiencing malignant melanoma.
The mitogen-activated protein kinase (MAPK) signal transduction pathway controls key cellular processes such as proliferation, differentiation, and survival. Among four major MAPK modules, the one converging on the activation of extracellular signal-regulated kinase (ERK) and its upstream activator MAPK and ERK kinase (MEK) is the most extensively studied and perhaps the most relevant to cancer pathogenesis and therapy [1,2]. Although oncogenic mutations of either MEK or ERK have not been identified in human tumors, their constitutive activation is sufficient to transform mammalian cells; moreover, the MEK/ERK kinase module serves as a focal point in the signal transduction pathway of known oncogenes, such as RAS or RAF  and disruption of its activity by pharmacological inhibitors severely impairs the transforming ability of many upstream-acting cellular oncogenes [4,5]. As a result, aberrant activation of the MEK/ERK pathway is observed in a large proportion of human cancers, including a wide variety of solid tumors and hematological malignancies, and has recently emerged as a promising target for anticancer therapies [2,6,7]. In addition to its role in fostering cancer cells' proliferation and survival, the MAPK module converging on ERK activation is also an important regulator of angiogenesis: indeed, MAPK activity controls vascular endothelial growth factor (VEGF) expression, through both hypoxia-inducible factor 1 (HIF-1)-dependent and Sp1/AP-2-dependent mechanisms .
Constitutive ERK activation is observed in virtually all melanomas [9,10], where MAPK is activated by the production of autocrine growth factors or, more rarely, by mutational activation of growth factor receptors, such as c-kit. Most commonly, however, ERK is constitutively activated as a result of gain-of-function mutations in pathway elements that are immediately upstream of MEK, either NRAS or BRAF [11–13]. The latter is arguably the most common mutational event in human melanoma, where it is observed in up to 70% of cases; BRAF mutations result in the aberrant activation of ERK, which, in turn, provides an essential tumor growth and maintenance signal by fostering proliferation, survival, chemoresistance, and the autocrine production of proangiogenic factors, such as VEGF [10,14]. Most interestingly from a therapeutic perspective, BRAF mutations may constitute the Achilles' heel of malignant melanoma because BRAF-mutated tumors seem to be exquisitely sensitive to clinically available MEK inhibitors . Froma molecular standpoint, data from Garnett et al.  indicate that, although a small fraction of BRAF mutations generates an enzyme that is impaired in its ability to activate the downstream MEK/ERK cascade, kinase-impaired mutants also work through the mitogenic cascade culminating in ERK activation. The mechanism is a rescue of kinase-impaired mutant BRAF by wild-type CRAF through a process that involves 14-3-3-mediated hetero-oligomerization and transactivation [16,17].
Here, we investigated the therapeutic potential of the novel, potent, and selective MEK inhibitor, PD0325901, against melanoma cells. PD0325901 is a noncompetitive MEK inhibitor, with improved oral bioavailability and aqueous solubility, compared with its parent compound CI-1040, and is currently in phase 1/2 clinical development in different solid tumors, including malignant melanoma [1,2,18]. In preclinical models of human melanoma, we found that PD0325901 potently inhibits cell growth, promotes apoptosis, and decreases the production of proangiogenic factors, such as VEGF and interleukin 8 (CXCL8).
ME1007, ME4405, ME4686,ME8959, ME10538, and ME13923 human melanoma cell lines were established at the Istituto Nazionale Tumori (Milan, Italy), as previously described ; the JR8 melanoma cell line was established at the Regina Elena Cancer Institute ; all other cell lines were purchased from American Type Culture Collection (ATCC, Manassas, VA). Cell lines were maintained in RPMI 1640 medium (Invitrogen, Milan, Italy) containing 10% of FBS, 2 mM l-glutamine, and antibiotics at 37°C under 5% CO2–95% air. PD0325901 [N-((R)-2,3-dihydroxy-propoxy)-3,4-diXuoro-2-(2-Xuoro-4-iodophenylamino)-benzamide] was obtained from Pfizer Global Research and Development (Ann Arbor, MI). The drug was dissolved in DMSO as a 10-mM stock solution, stored at -20°C, and adjusted to the final concentration with culture medium.
For IC50 assays, exponentially growing cells were exposed to increasing concentrations of PD0325901 (0.1–1000 nM) for 24, 48, or 72 hours. Cells were then assayed for cell viability (by trypan blue exclusion test) and counted using a Coulter Counter (Kontron Instruments, Milan, Italy). The IC50 value was calculated according to the Chou-Talalay method using the Calcusyn software.
To evaluate colony-forming ability, melanoma cells were seeded in 60-mm Petri dishes at a density of 500 cells per dish and cultured in medium with or without PD0325901. After 7 days of incubation, colonies were fixed with methanol, stained with 2% methylene blue in 95% ethanol/5% water (v/v), and counted under a light microscope (1 colony ≥ 50 cells). All experiments were performed in triplicate.
Female CD-1 nude (nu/nu) mice, 6 to 8 weeks old, were used (Charles River Laboratories, Calco, Italy). Mice were housed under pathogen-free conditions, and all procedures involving animals and their care were in agreement with national and international laws and policies. Solid tumors were obtained by subcutaneous injection of 1.5 or 2 x 106 viable cells for M14 (BRAFV600E) and ME8959 (wtBRAF), respectively. Each experimental group included 8 to 10 animals. PD0325901 was formulated in 0.5% hydroxypropyl methyl-cellulose plus 0.2% Tween 80 and administered by oral gavage at the dosage of 50 mg/kg per day; treatment was started when the tumor mass reached 100 mg. Untreated mice and mice treated with an equal amount of vehicle were used as control groups. The drug was administered daily for 21 days, and tumor size wasmeasured every 2 to 3 days. Mice were killed when tumor volume reached more than 2000 mg, and tumors were excised and placed in 10% buffered formaldehyde. Tumor weight was calculated from caliper measurements according to the following formula: tumor weight (mg) = length (mm) x width (mm)2 / 2.
ForWestern blot analysis, 35 µg of total protein, prepared as described previously , was fractionated by SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membrane (Amersham, Chicago, IL). Membranes were probed with primary antibody (Ab), and the signal was detected using peroxidase-conjugated antimouse or antirabbit secondary Abs (Cell Signaling Technology, Inc, Beverly, MA). The following primary Abs were used: anti-HIF-1α or anti-HIF-1β/ARNT1 (BD Biosciences, San Jose, CA); Abs specific for phosphorylated (Thr202/Tyr204) and total ERK-1/2 (Cell Signaling Technology, Inc); anti-myeloid cell leukemia-1 (BD Biosciences); anti-B-cell lymphoma-2 (Dako, Carpinteria, CA); antisurvivin (R&D System, Minneapolis, MN); and anti-p27Kip1 and anti-c-Myc (Santa Cruz Biotechnology, Santa Cruz, CA). To check the amount of proteins transferred to the nitrocellulose membrane, β-actin was used and detected by anti-β-actin (clone AC-15; Sigma, St. Louis, MO).
Cells were fixed in ice-cold ethanol (70% vol/vol) and stained with propidium iodide (PI; 25 mg/ml PI, 180 U/ml RNase, 0.1% Triton X-100, and 30 mg/ml polyethylene glycol in 4 mM citrate buffer, pH 7.8; Sigma). The DNA content was determined using a FACScan flow cytometer (Becton Dickinson, San Jose, CA). Cell cycle distribution was analyzed using the ModFit LT software (Verity Software House, Topsham, ME). For annexin V binding studies, the cells were stained with fluorescein isothiocyanate-conjugated annexin V using the Vybrant Apoptosis Kit (Molecular Probes, Eugene, OR) and analyzed by flow cytometry while simultaneously assessing membrane integrity by PI exclusion.
Total RNA was extracted using the Trizol reagent (Invitrogen, Carlsbad, CA), followed by additional purification with the RNeasy kit from Qiagen (Valencia, CA). Five micrograms of each RNA sample was retrotranscribed to double-stranded complementary DNA (cDNA) and labeled through in vitro transcription by the One-cycle cDNA synthesis/GeneChip IVT labeling kit (Affymetrix, Santa Clara, CA). Twenty micrograms of biotin-labeled complementary RNA was fragmented and hybridized to the Human Genome U133 Plus 2.0 Gene-Chip array (Affymetrix) overnight at 45°C using a Hybridization Oven 640 (Affymetrix). The hybridized probe arrays were washed and stained with the streptavidin-phycoerythrin conjugate (Molecular Probes, Invitrogen) using the Fluidic Station 450 (Affymetrix) and were scanned by the GeneChip Scanner 3000 (Affymetrix).
For statistical analysis, Affymetrix gene expression data were processed with the dChip software (www.dchip.org), which uses an invariant set normalization method. The array with the median overall intensity was chosen as the baseline for normalization. Model-based expressions were computed for each array and probe set, using only perfect match probes. For unsupervised analysis, the following nonspecific filtering criteria were used: 1) gene expression level was required to be higher than 100 in at least 30% of the samples and 2) the ratio of the SD to the mean expression across all samples was required to be between 0.5 and 10. Supervised analyses were performed to compare DMSO-treated cells with PD0325901-treated cells at 6 and 24 hours. For these comparisons, a t test was used: only the genes with expression higher than 100, P ≤ .05, and a fold change of 2 or higher were retained.
The cells were cultured in serum-free medium in a humidified atmosphere with 95% air and 5% CO2 (normoxia) or were incubated in specially designed aluminum chambers flushed with a gas mixture containing 5% CO2 and 95% N2 . VEGF analyses-ELISA Kit and CXCL8 analysis-ELISA Kit (R&D Systems, Minneapolis, MN) were used to determine the amount of either VEGF or CXCL8 protein levels in the conditioned medium. The sensitivity of the VEGF and CXCL8 assays was 31.2 pg/ml.
Electrophoretic mobility shift assay (EMSA) was performed after exposure to either normoxic or hypoxic conditions for 24 hours in the presence or absence of PD0325901, as previously described . The following double-strand oligomers were used as labeled probes or cold competitors: HIF-1 (human VEGF 5′ gene promoter), 5′TCGACCACAGT-GCATACGTGGGCTCCAACAGGTCCTCTTC-3′ [21,23]; activator protein-1 (AP-1; human CXCL8 5′ gene promoter), GTG TGATGA CTC AGG TTT G . Oligonucleotides were purchased from Invitrogen. In competition assays, a 100x unlabeled competitor was added at the same time of probe addition. In supershift analyses, 2 µl (2 mg/ml) of anti-c-jun, anti-B-jun, or anti-c-fos Abs (Santa Cruz Biotechnology) were added to the reaction.
The Human Angiogenesis Antibody Array I (RayBiotech, Inc, Norcross, GA) was used according to the manufacturer's protocol in evaluating the secretion of 20 angiogenic factors into the conditioned medium of the different lines. Membranes spotted in duplicate with Abs against angiogenic factors were incubated overnight with the conditioned medium. The signals on the membranes were detected by chemiluminescence. The intensity of protein signal (two spots for each protein) was compared with the relative positive signals by densitometric analysis.
First, we evaluated the effect of PD0325901 on ERK phosphorylation in M14 and other human melanoma cell lines. PD0325901 dose-dependently inhibited the phosphorylation of ERK and its downstream target ribosomal S6 kinase (p90RSK1), without affecting total levels of ERK protein expression, in the M14 model (Figure 1A). Similar results were obtained in the other melanoma cell lines tested (Figure W1A; data not shown), regardless of their BRAF mutation status. Inhibition of ERK phosphorylation after exposure to PD0325901 was rapid (complete inhibition observed within 15 minutes) and persisted for at least 72 hours (Figure 1B).
These results indicate that PD0325901 is a potent inhibitor of MEK-to-ERK signaling in human melanoma cells.
The growth-inhibitory properties of PD0325901 were assessed in vitro on a panel of 11 human melanoma cell lines previously characterized for the presence or absence of BRAF, NRAS, and TP53 mutations and for PTEN expression .
As shown in Table 1, PD0325901 potently (IC50, 20–50 nM) inhibited the growth of human melanoma cell lines with (M14, A375P, M, and SM, ME10538, ME4686, JR8) or without (ME4405 and ME13923) BRAF mutations. ME1007 and ME8959, both of which had wild-type BRAF, were slightly more resistant to PD0325901-mediated growth inhibition (IC50, ≥ 100 nM).
As exemplified in Figure 2A for the BRAFV600E cell line M14, PD0325901-induced growth inhibition was dose- and time-dependent. Although the potency of PD0325901 in inhibiting ERK phosphorylation was essentially unchanged, its growth-inhibitory activity was strikingly potentiated under low serum (2% fetal calf serum) conditions, resulting in an IC50 of less than 1 nM in the M14 model. We further characterized the effect of PD0325901 using a clonogenic growth assay. M14 cells were seeded at cloning densities (500 cells/cm2) in serum medium and cultured for 7 days in the presence or absence of increasing PD0325901 concentrations (0.1–1000 nM); as shown in Figure 2, B and C, PD0325901 strikingly reduced M14 clonogenic growth, with 50% inhibition observed at approximately 0.1 nM.
To further evaluate PD0325901-induced melanoma growth inhibition, we tested the drug in vivo in xenograft models obtained by subcutaneous injection of either M14 (BRAFV600E) or ME8959 (wtBRAF). Preliminary experiments conducted with M14-derived tumors indicated that in vivo growth inhibition was dosage-dependent, with 50 mg/kg per day being significantly more effective than 25 mg/kg per day, without gross signs of toxicity (Figure W2A); we therefore used the 50 mg/kg per day for further experiments. As shown in Figure 3A, daily oral treatment of established tumors with 50 mg/kg per day of PD0325901 significantly impaired in vivo tumor growth (60%–65% inhibition compared with controls at the end of a 21-day treatment cycle) in both M14 and ME8959 xenografts. The effects of PD0325901 were reversible, and tumors grew back after treatment interruption (data not shown). Upon microscopic examination of M14-derived tumors, PD0325901-treated tumors lost the characteristic nodular architecture, showing almost no stroma or blood vessel formation either at the periphery or within the tumor mass; displayed larger areas of necrosis, with only cells immediately adjacent to blood vessels surviving; and seemed more differentiated with a reduction in the number of aberrant mitosis and striking normalization of the characteristic nuclear and chromatin pleomorphism (Figure 3B). Similar results were obtained in the ME8959 tumor model (data not shown). In particular, by the end of the treatment period, microvessel density was significantly decreased by PD0325901 in both xenograft models (Figure W2B).
Overall, these results indicate that PD0325901 exerts potent growth-inhibitory effects in human melanoma cell lines, regardless of BRAF mutations.
Using the BRAFV600E cell line M14 as a model, we analyzed the mechanisms of PD0325901-induced growth inhibition in further detail. Exposure of M14 cells to PD0325901 caused a dose- and time-dependent cell cycle accumulation at the G1/S boundary and depletion of cells in the S-phase (Figure 4, A and B). Moreover, exposure of M14 cells to PD0325901 caused a dose- and time-dependent increase in the percentage of cellswith sub-G1 DNA content, thus indicating induction of apoptosis (Figure 4C). Compared with the kinetics and dose-response curve of cell cycle inhibition, DNA decrease to sub-G1 levels required longer times of exposure (72 hours) and higher concentrations of the drug (≥100 nM). The apoptotic nature of PD0325901-induced cell death was further confirmed by annexin V binding, which displayed a dose-dependency similar to that observed for the appearance of a hypodiploid peak, but was detectable within 48 hours (Figure 4, C and D). Low serum conditions (2%) significantly potentiated and accelerated PD0325901-induced apoptosis.
We next analyzed the effect of PD0325901 on the expression of key regulators of cell cycle progression and apoptosis by immunoblot analysis (Figure 5). Consistent with the observed G1 accumulation, the protein expression of cyclinD1 was strikingly decreased in PD0325901-treated M14 cells; conversely, the cyclin-dependent kinase inhibitor p27KIP1 accumulated in PD0325901-treated cells in a dose- and time-dependent fashion. PD0325901 also substantially, although not completely, inhibited c-myc expression. Relatively high concentrations of PD0325901 (100 nM) strikingly downregulated Bcl-2 protein expression, leaving the expression of Bcl-xL and Mcl-1 unaffected. Survivin was also strikingly downregulated by PD0325901 treatment (Figure 5).
To gain further insights into the molecular mechanisms of action of PD0325901, changes in the gene expression profiles were analyzed in M14 cells exposed to the drug for 6 and 24 hours. Supervised comparison between treated and untreated samples after 6 hours highlighted a large set of modulated genes (n = 557), 57 of which remained concordantly modulated at the 24-hour time point (Figure 6). Among 163 gene ontology-annotated genes upregulated by PD0325901 treatment, transporters (particularly members of the solute carrier family of proteins) and transcription modulators were clearly overrepresented (19 and 17 probe sets, respectively; Figure 6B). Interestingly, several genes related to melanocyte differentiation and melanin biosynthetic pathway were upregulated by PD0325901, as were semaphorin 6A (6- to 13-fold up-regulation with four different probe sets) and cyclin G2 (4.7-fold; Table W1). Transcription modulators and transporters were also the most represented among the 225 genes downregulated by PD0325901 (34 and 20 genes, respectively; Figure 6B), immediately followed by cell cycle/cell division (20 genes), translation (11 genes), and apoptosis (8 genes) regulators. Interestingly, genes involved in the control of signal transduction and MAPK activity, such as DUSP-4 and -6 and SPRY-2 and -4, were among those most profoundly downmodulated by PD0325901 (29- to 95-fold for DUSP-6; Table W1). Cell cycle and apoptosis regulators whose protein expression was decreased by PD0325901 treatment, such as cyclin D1 and c-myc, were also found to be significantly downregulated at the messenger RNA (mRNA) level (seven- to nine-fold and six-fold, respectively). Finally, several angiogenesis-/tissue remodeling-related genes, including VEGF-A and CXCL8, were modulated on PD0325901 treatment.
We next examined the effects of PD0325901 on the production of VEGF. Exposure of M14 cells to increasing concentrations of PD0325901 (1–100 nM) for 24 hours resulted in a significant (P ≤ .01 for PD0325901 concentrations ≥10 nM) and dose-dependent inhibition of VEGF release in culture-conditioned medium, as measured by ELISA, under both normoxic and hypoxic conditions (Figure 7A). Similar results were obtained in other melanoma cell lines that did not harbor BRAF mutations (ME13923 and ME8959 [Figure W1B; data not shown]; ME1007 had barely detectable VEGF levels even after hypoxic stimulation and were therefore not evaluable for the effect of PD0325901 on VEGF production [data not shown]). PD0325901-mediated down-regulation of VEGF production took place, at least in part, at the transcriptional level, as indicated by a 2.7-fold decrease in VEGF mRNA detected by gene expression profiling. We next analyzed whether PD0325901 affected the expression and function of the transcriptional complex HIF-1. As expected , expression of the HIF-1α subunit was undetectable under normoxic conditions and was strongly induced by exposure to hypoxia in M14 cells (Figure 7B); under hypoxic conditions, exposure to PD0325901 for 24 hours strikingly reduced HIF-1α protein levels. In contrast, the levels of the HIF-1β subunit were unaffected by MEK inhibition. Moreover, PD0325901 dose-dependently decreased HIF-1 binding to the putative hypoxia-responsive element in the VEGF promoter under hypoxic conditions, as evaluated by EMSA (Figure 7C).
Overall, these results indicate that MEK inhibition by PD0325901 inhibits VEGF production by melanoma cells through inhibition of HIF-1α expression and binding.
In addition to VEGF, we screened the expression of other angiogenic factors in culture-conditioned medium from M14 cells exposed to PD0325901 using an angiogenesis-oriented Ab array. As shown in Figure 8A, treatment with PD0325901 under normoxic conditions strikingly reduced the expression levels of the proangiogenic cytokine CXCL8. Using ELISA, we confirmed that PD0325901 significantly (P ≤ .03) and dose-dependently inhibited CXCL8 production (Figure 8B). Similar results were obtained in other melanoma cell lines that did not harbor BRAF mutations (ME13923, ME8959, and ME1007; Figure W1C; data not shown). PD0325901-mediated down-regulation of CXCL8 production took place, at least in part, at the transcriptional level, as indicated by an approximately three-fold decrease in CXCL8 mRNA detected by gene expression profiling at 24 hours. We next investigated whether PD0325901-mediated downregulation of CXCL8 production might involve the AP-1 transcription factor. As demonstrated by EMSA and supershift assay, PD0325901 dose-dependently decreased the binding to the CXCL8 promoter of an AP-1 complex containing c-Jun, b-Jun, and c-Fos (Figure 8C).
Overall, these results indicate that MEK inhibition by PD0325901 potently inhibits CXCL8 production by melanoma cells through inhibition of AP-1 binding and transcriptional activity.
The MEK/ERK signaling module has recently emerged as a promising therapeutic target in malignant melanoma [1,2,9,10,26]. Here, we demonstrate that the novel MEK inhibitor PD0325901 inhibits in vitro and in vivo growth of human melanoma cell lines, by inhibiting cell cycle progression and inducing apoptosis, and decreases the production of proangiogenic cytokines, such as VEGF and CXCL8.
Sensitivity to MEK blockade-induced growth inhibition has recently been linked to the presence of BRAF mutations [15,27]. Within the panel of melanoma cell lines we examined, no clear relationship emerges between BRAF mutational status and sensitivity to PD0325901; although the wtBRAF cell lines ME1007 and ME8959 display a slightly decreased sensitivity to PD0325901-mediated growth inhibition (Table 1), the sensitivity of other wtBRAF melanoma cell lines (ME4405 and ME13923) is similar to that of cell lines harboring the classic BRAFV600E mutation (P = .28, for the comparison between wt and mutated BRAF cell lines). In addition, we have evidence that acute myeloid leukemia cell lines may be extremely sensitive to PD0325901-mediated growth inhibition, even in the absence of BRAF mutations (M.R.R., unpublished observations). Most importantly, in vivo data indicate that PD0325901 treatment induces a similar degree of growth inhibition in both M14 (harboring the BRAFV600E mutation) and ME8959 (wtBRAF) xenograft models (Figure 3A). Overall, these data leave open the possibility that sensitivity to growth inhibition by MEK-targeted agents may be sustained by molecular mechanisms other than BRAF mutations.
Consistent with the prominent role played by theMAPK pathway in the regulation of G1/S transition , the MEK inhibitor PD0325901 exerts predominantly cytostatic effects, inducing G1 cell cycle arrest. This observation is in line with recently published reports from our group and others, showing a marked cytostatic effect by first-generation (CI-1040) or second-generation (PD0325901 and AZD6244) MEK inhibitors, both in vitro and in vivo [15,27,29–32]. The molecular mechanisms by which PD0325901 induces G1 arrest in sensitive melanoma cells are consistent with current knowledge of ERK actions in cell cycle progression: the crucial point is inhibition of cyclin D/cyclin-dependent kinase 4/6 complex activity by FOS/FRA- and MYC-dependent transcriptional down-regulation of cyclin D1  and accumulation of the cyclin-dependent kinase inhibitor p27Kip1 . However, high concentrations and prolonged exposure to the drug also induced apoptosis in a sizable proportion of sensitive melanoma cells, in agreement with recently published data produced using the U0126 MEK inhibitor . Consistent with the reported role of ERK in counteracting apoptosis at both the mitochondrial and the cytosolic caspase activation levels [6,7], apoptosis induced by PD0325901 was found to correlate with down-regulation of Bcl-2 and survivin with very close time- and dose-dependency. These results are consistent with ERK's ability to phosphorylate Bcl-2 on Ser52, thereby inhibiting protein degradation , and to increase survivin expression [30,37] and may be exploited therapeutically to build pharmacological combinations endowed with highly synergistic proapoptotic activity [6,7].
Gene expression profiling experiments indicate that PD0325901 treatment counterregulates many of the genes that have been described to be differentially expressed in melanoma cells with constitutively active ERK [38,39], including CXCL1/GROα, CD73, PLAT, SPRED1, SPRY2, TFAP2C, TNC, and CXCL8. In addition to genes regulating cell cycle progression, PD0325901 modulates an array of other genes involved in molecular circuitries that participate in the regulation of MAPK signaling itself and are potentially relevant for the anti-melanoma activity of MEK inhibitors . The striking downregulation of DUSP-4 and -6 and SPRY-2 and -4 (Table W1) on exposure to PD0325901 clearly indicates the interruption of a negative feedback loop, by which ERK activation signals the inhibition of signaling through upstream components of the Ras/Raf/MEK/ERK cascade . Although the functional relevance of disruption of such a negative feedback in the context of BRAF mutation-driven constitutive ERK activation in melanoma cells remains to be determined [41,42], it is interesting to note that similar mechanisms seem to also take place in PD0325901-sensitive acute myeloid leukemia cells, in which RAF and MEK hyperphosphorylations are observed in response to MEK inhibition (M.R.R., unpublished observations). These findings are consistent with the observation of prolonged growth factor-mediated RAF activation in response to MEK inhibition  and support the investigation of vertical combination strategies aimed at inhibiting multiple signaling elements along the MAPK cascade. Another interesting finding is the marked up-regulation of genes involved in melanoma differentiation and melanin biosynthesis (e.g., TYR, TYRP1, ENDRB) in response to MEK inhibition by PD0325901 (Figure 6 and Table W1); in agreement with a recently published report , we also observed cell differentiation and increased melanin production in vivo in M14- and ME8959-derived xenograft models treated with daily oral PD0325901. These findings may be of therapeutic relevance in view of recent reports indicating that MEK inhibition may result in increased melanoma immune recognition and killing by immune cells through both upregulation of differentiation antigens  and down-regulation immunosuppressive factors [38,45].
Recent data indicate that inhibition of ERK-MAPK signaling in the tumor vasculature suppresses angiogenesis and tumor growth directly, by impairing endothelial cell survival and sprouting . Here, we demonstrate that MEK inhibition by PD0325901 may also interfere with angiogenesis indirectly by down-regulating the production of proangiogenic factors by tumor cells, in both mutant and wtBRAF cell line models of melanoma, again arguing against an exclusive role of BRAF mutational status in determining the outcome upon therapeutic MEK inhibition. In particular, we focused on the effects of PD0325901 on the production of two major angiogenesis regulators, VEGF-A and CXCL8. Consistent with previous findings from our group, demonstrating a pivotal role for ERK activation in Bcl-2 overexpression-driven VEGF production and angiogenesis in melanoma models [47,48], PD0325901 significantly decreased VEGF-A production at the mRNA and protein levels, under both normoxic and hypoxic conditions (Figure 7A). From a mechanistic standpoint, PD0325901-induced VEGF down-regulation under hypoxic conditions seems to be related to the inhibition of HIF-1α protein expression and binding activity at the VEGF promoter; although the latter is consistent with the proposed role of ERK in the regulation of HIF-1 transcriptional activity [8,49], down-regulation of hypoxia-induced HIF-1α protein expression upon MEK inhibition has not been reported in other tumor models . Whether this observation is related to the cellular model examined or to the specific MEK inhibitor used (PD0325901) and whether the regulation of HIF-1α protein expression takes place at the transcription/translation or protein stability/degradation level remains to be determined. In addition, PD0325901-induced VEGF down-regulation under normoxic conditions is not clearly related to HIF-1 expression/transcriptional activity, leaving open the possibility that other ERK-regulated transcription factors, such as the AP-2/Sp1 complex, may play a relevant role [51,52]; this hypothesis is supported by gene expression profiling experiments indicating profound (more than four-fold) downregulation of elements of the AP-2 transcriptional complex upon PD0325901 exposure (Table W1). CXCL8 is an important proinflammatory and proangiogenic chemokine involved in melanoma progression , which has been linked to constitutive ERK activation . Our results confirm a prominent role for ERK activation in the regulation of CXCL8 production by melanoma cells at both mRNA and protein levels and indicate decreased binding of the AP-1 transcriptional complex to the CXCL8 promoter as a possible molecular mechanism for PD0325901-induced CXCL8 down-regulation. These findings are consistent with results demonstrating the modulation of CXCL8 production by AP-1-mediated transcriptional regulation . In addition to decreased production of proangiogenic cytokines, direct inhibitory effects of PD0325901 on the microenvironment surrounding the tumor may also play an important role in the inhibition of angiogenesis observed in vivo (Figure W2B), as suggested by PD0325901-induced dose-dependent inhibition of endothelial cell and fibroblast proliferation in vitro (Figure W2C).
Overall, the findings reported herein support the continued development of MEK inhibitors, such as PD0325901, as promising therapeutic agents with multiple potentially relevant mechanisms of action (inhibition of proliferation, induction of apoptosis, inhibition of angiogenesis) in malignant melanoma; deeper insights into the molecular mechanisms of action of MEK-targeted agents will likely increase our chances to successfully translate such exciting preclinical findings into effective therapies for patients experiencing malignant melanoma.
The authors thank Antonio Marchetti, Department of Pathology, University of Chieti “G. D'Annunzio,” for his help with BRAF and MEK1 mutation sequencing.
1This work was supported in part by grants from the Italian Association for Cancer Research (to M.M. and D.D.B.), the Italian Ministry of Health (to M.M. and G.Z.), and the Cariplo Foundation (to M.M. and A.A.). L.C. is a fellow of the Italian Foundation for Cancer Research.