Targeting Lineage-specific MITF Pathway in Human Melanoma Cell Lines by A-485, the Selective Small-molecule Inhibitor of p300/CBP

Melanoma is one of the most frequent cancers with increased incidence in the Western societies. Melanoma is responsible for 80% of deaths from skin cancer. The median overall survival of patients with advanced-stage melanoma has increased from approximately 9 months before 2011 to at least 2 years (1). Melanoma has been studied extensively and is found to be related to the uncontrolled proliferation of melanocytes (2). Although there have been recent successes in developing effective treatments for melanoma such as targeted therapies against BRAF and MEK kinases, and immunotherapy, advanced melanoma still has a high mortality rate (3, 4).

The identification of a constitutively active MAPK pathway due to BRAF V600E mutation in about 40% melanoma has led to the development of selective BRAF inhibitors such as vemurafenib (5). Although great initial clinical benefits were observed in patients with BRAF V600E mutation, the long-term benefit of vemurafenib has been compromised because of the development of resistance to the therapy. Subsequent study indicated that acquired resistance is caused by additional mutations, which can reactivate the MAPK pathway directly or indirectly (6, 7).

Interestingly, approximately 15%–20% patients with BRAF mutation do not respond to vemurafenib and the mechanism underlying the intrinsic resistance remains an intense area of investigation (7). Recently, it was shown that microphthalmia-associated transcription factor (MITF)-low melanoma is associated with intrinsic resistance to multiple targeted agents including BRAF inhibitor (8, 9). MITF is a melanocyte-specific basic helix-loop-helix leucine zipper transcription factor that plays an important role in the differentiation, proliferation, and survival of melanocytes (10). MITF has been reported as a lineage-specific oncogene in melanoma as primary cultures of human melanocytes can be transformed by enhanced expression of MITF (11). The oncogenic role of MITF is also reflected by the occurrence of its gene amplification in approximately 15% of metastatic melanomas, as well as the association with resistance to conventional chemotherapy (11). However, it has been challenging to target MITF because it is a transcription factor. Because it was shown that the transcriptional coactivators p300/CBP interact with MITF and regulate the downstream target genes (12, 13), we hypothesize that targeting p300/CBP may indirectly target the MITF pathway in melanoma.

p300 and CREB-binding protein (CBP; paralogs called p300/CBP thereafter) are members of the KAT3 family of histone acetyltransferase (HAT) that have a number of biological substrates (14). p300/CBP regulates a wide range of cellular events such as senescence, apoptosis, cell cycle, DNA damage response, and cellular differentiation (15). p300/CBP act as transcriptional coactivators for multiple proteins, including oncogenes, suggesting the potential involvement of p300/CBP in cancers (16). Several p300/CBP mutations in patients with melanoma have been identified, but these mutations have not been studied extensively (17). In addition, EP300 expression is upregulated and frequently amplified in melanoma cell lines (18). Together, the evidence indicates the potential therapeutic value of targeting p300 in melanoma cells. We recently identified a first-in-class selective catalytic inhibitor of p300/CBP A-485 (ref. 19; Fig. 1A). A-485 competes with acetyl-coenzyme A, and robustly inhibited the HAT activity of p300 bromodomain-HAT-CH3 (BHC) as well as the BHC domain of CBP. A-485 also showed minimal activity against other HATs family members and negligible binding to BET bromodomain proteins and other protein targets such as G protein–coupled receptors, ion channels, transporters, and kinases (19). In this study, we evaluated the activity of this specific p300/CBP inhibitor A-485 in a panel of melanoma cell lines. We demonstrated the modulation of MITF expression and pathway in response to A-485. Our results support a rationale for testing p300/CBP inhibitors in patients with melanoma with MITF copy number gain and upregulated expression.

p300/CBP inhibitor A-485 and inactive compound A-486 were synthesized at AbbVie, Inc.

Melanoma cell lines were cultured as recommended by the supplier (ATCC). The cells were tested for Mycoplasma using MycoAlert Detection Kit (Lonza), authenticated using GenePrint 10 STR Authentication Kit (Promega).

IC₅₀ of acetylation marker H3K27 and cell viability were determined as described previously (19).

Silencing of EP300 and MITF was performed by transfecting SKMEL5 cells with EP300 or MITF siRNAs or nontargeting control siRNAs synthesized by Dharmacon (On-Target-Plus SMARTpool siRNA for each gene). Reverse transfections were performed with Lipofectamine 2000 according to the manufacturer's protocol (Invitrogen).

Senescence was determined by using a Senescence β-Galactosidase Staining Kit (Cell Signaling Technology). Cells were cultured in 6-well plate and treated with DMSO or A-485 for 5 days before staining.

The percentages of apoptotic and early apoptotic cells were detected using Annexin V FITC and propidium iodide (PI) staining (Thermo Fisher Scientific). After staining, cells were analyzed by BD LSR II Flow Cytometer (BD Biosciences). The percentage of early apoptotic (high Annexin V-FITC/low PI) and late apoptotic (high Annexin V-FITC/high PI) cells was determined.

RNA was purified using the RNeasy Mini Kit (Qiagen) and quantitative PCR (qPCR) reaction was performed using the SuperScript III Platinum One-Step qPCR Kit (Thermo Fisher Scientific) following the manufacturer's instructions. MITF and EP300 primers were designed and synthesized by Integrated DNA Technologies. qPCR was performed with CFX96 Touch real-time PCR Detection System (Bio-Rad), using GAPDH as an internal control. The siRNA-induced reduction of gene expression was calculated using the comparative C_t (ΔΔC_t) method following the manufacturer's protocol (Thermo Fisher Scientific). The results were normalized to GAPDH and to the nontargeting siRNA control.

Cells were treated with DMSO or 3 μmol/L of A-485 for 6 or 24 hours. Purified RNA was subjected to microarray analysis on the GeneChip Human Genome U133 Plus 2.0 Array following manufacturer's protocol (Affymetrix). The data normalization and processing were performed according to the previously described approach (20). The upstream regulator analysis was performed with Ingenuity Pathway Analysis (IPA; Qiagen). Z-scores of >2 or <−2 are considered significant. Microarray data is deposited in GEO (accession no. GSE116459).

Cell lysates were prepared in RIPA buffer (Sigma) plus protease inhibitor cocktail (Roche Life Science). Thirty micrograms of total protein was resolved on a 12% SDS polyacrylamide gel. Antibodies against MITF (Abcam) and control GAPDH (Cell Signaling Technology) were used to detect protein level. After incubation with secondary antibodies (LI-COR), blots were developed using Odyssey infrared imaging system (LI-COR).

Native chromatin immunoprecipitation (ChIP) was performed as described previously (21). Briefly, nuclei were initially released to generate the soluble chromatin. The immunoprecipitation was then carried out overnight at 4°C with ChIP-grade anti-acetylated H3 antibody or anti-p300 antibody coupled with protein A/G magnetic beads (all from Millipore-Sigma). The bound chromatin was eluted followed with Proteinase K (Millipore-Sigma) digestion at 55°C for 2 hours. The eluted DNA was purified by QIAquick PCR Purification columns (Qiagen), and analyzed by qPCR using SYBR Green PCR Master Mix (Thermo Fisher Scientific) following the manufacturers' instructions with CFX96 Touch Real-Time PCR Detection System (Bio-Rad). qPCR was performed using primer pairs near the transcription start sites of MITF and its putative target genes. qPCR primers are: MITF primer, forward: 5′- CATTGTTATGCTGG AAATGCTAGAA-3′, reverse: 5′- GGCTTGCTGTATGTGGTACTTGG-3′; MLANA primer, forward: 5′- GGATAGAG CACTGGGACTGG-3′, reverse: 5′- CTGACGGG GTCGTCTGTAAT-3′; TRPM1 primer, forward: 5′- AAAGCTCATGGAAAGCTG GAA-3′, reverse: 5′- GCAT CCACAGTCACCTGAAA-3′; SEMA6A, forward: 5′-GCCTAAA CCTGTGGCTGGACACAA-3′, reverse: 5′-CCCTGGAGGGTGGGATTCTCTAAA-3′; and TDRD7, forward: 5′- AGAGGGAGT GCTTCCGTTTTCA-3′, reverse: 5′- GCCATTAAAGGC TGCTCACAAC-3′). Relative binding values were calculated using the ΔΔC_t method following the manufacturer's protocol (Thermo Fisher Scientific) by comparing with DMSO enrichment relative to negative control (IgG) for ChIP is indicated; GAPDH promoter sequence is used as endogenous control for qPCR.

MITF is a melanocyte-specific transcription factor that plays an important role in the differentiation, proliferation, and survival of melanocytes (10). Data from Project DRIVE (a large-scale RNAi screen in cancer cell lines that reveals vulnerabilities to specific genes in cancer subtypes; ref. 22) shows that 18 of 34 melanoma cell lines have a dependency on MITF (Supplementary Fig. S1). Previous publications showed that MITF recruits p300/CBP in activating transcription (12, 13). We reasoned that a small-molecule inhibitor of p300/CBP could target the MITF pathway in melanoma. Because p300/CBP are HATs, we first assessed the effect of A-485 in modulating histone acetylation of Histone H3 at Lys27 (H3K27Ac) using a high-content imaging assay (23). We selected two melanoma cell lines that are vulnerable to MITF silencing (WM2664 and SKMEL5) and one cell line resistant to MITF silencing (A375) based on the Project DRIVE data (Supplementary Fig. S1). The results indicated that A-485 inhibited the H3K27Ac in all three melanoma cells lines with IC₅₀ = 0.59 μmol/L, 0.93 μmol/L, and 0.96 μmol/L for WM2664, SKMEL5, and A375 cells, respectively (Fig. 1A). In contrast, there was no effect on H3K27 acetylation mark in all three cell lines using the inactive control compound A-486 (ref. 19; IC₅₀ > 10 μmol/L). Interestingly, even though A-485 could inhibit the H3K27 acetylation mark in all three cells lines, the cellular response to A-485 was different. In particular, proliferation of WM2664 and SKMEL5 cells was inhibited by A-485 (IC₅₀ ≤ 1 μmol/L) while A375 cell line was resistant to the inhibitor (IC₅₀ > 10 μmol/L; Fig. 1A). In addition, we treated primary melanocytes (HEM cells) with A-485 and showed that HEM cells were also sensitive to A-485 with IC₅₀ < 1 μmol/L (Supplementary Fig. S2).

It was previously suggested that downregulation of MITF or p300/CBP histone acetyltransferases activates a senescence checkpoint in human melanocytes (24, 25). We first demonstrated that silencing EP300 induced senescence in SKMEL5, similar to silencing MITF (Fig. 1B). We further showed that A-485 induced pronounced cellular senescence in sensitive melanoma cell lines WM2664 and SKMEL5 but not in the resistant line A375 (Fig. 1C). Our data suggested that p300/CBP inhibitor phenocopied the effect of silencing EP300 or MITF. In contrast, we found no significant change in the percentages of early apoptotic and apoptotic cells in all three melanoma cell lines with A-485 treatment as shown by annexin V/PI staining (Fig. 1D) and terminal deoxynucleotidyl transferase–mediateddUTP nick end labeling assay (Supplementary Fig. S3). Taken together, the major activity of p300/CBP inhibitor in the sensitive melanoma cell lines is induction of cellular senescence but not apoptosis, similar to silencing the expression of EP300 or MITF.

Melanomas are known to undergo phenotype switching. In particular, tumors with high MITF are highly proliferative and tumors with low MITF are highly invasive (26, 27). We determined whether A-485 treatment increases invasiveness in cells with high MITF. We treated WM2664 cells with A-485 for 6 days and showed that A-485 treatment greatly reduced the cell number but did not increase invasiveness (Supplementary Fig. S4).

To better understand the potential mechanisms of A-485 in melanoma cell lines, we performed global gene expression analysis using microarrays on the two sensitive melanoma cell lines, WM2664 and SKMEL5, as well as the resistant cell line A375. To distinguish early and late transcriptional changes, we treated these cells with A-485 for 6 and 24 hours. Overall, global gene expression analysis showed strong inhibition of gene expression in sensitive lines WM2664 and SKMEL5. Stronger modulation in gene expression was observed after 24-hour exposure to A-485 as compared with 6-hour exposure (Fig. 2A). Gene expression changes in the resistant cell line A375 were different from the sensitive cell lines and to a lesser extent. Minimal transcriptional effect was observed with the inactive compound, A-486 (data now shown). Upon performing Ingenuity upstream regulator analysis (IPA) across the different cell lines, we showed that there was an enrichment of transcriptional network regulated by MITF with p300/CBP inhibitor treatment in the two sensitive lines (Fig. 2B). These changes were much weaker in the resistant cell line A375. In contrast, there was no enrichment of transcriptional network regulated by Sry-related HMG-box-10 (SOX10), a neural crest stem cell transcription factor that contributes to melanomagenesis (28). Figure 2C shows the gene expression changes within the MITF signature in the two sensitive melanoma cell lines WM2664 and SKMEL5, as well as the resistant cell line A375. Quantification showed a 7- to 8-fold downregulation of MITF at 6 hours and 4- to 5-fold downregulation after 24 hours post A-485 treatment in the two sensitive cell lines. An independent qPCR analysis on gene expression indicated that the transcriptional levels of MITF was indeed downregulated in a dose-dependent manner following 6 and 24 hours of A-485 treatment in the sensitive cells lines WM2664 and SKMEL5, but not in the resistant cell line A375 (Fig. 2D). It should be noted that the resistant line A375 had low expression of MITF. Similarly, MITF protein expression was reduced in sensitive cell lines WM2664 and SKMEL5 following 24-hour treatment of A-485 (Fig. 2E) and low protein expression of MITF was observed in the resistant cell line A375. Taken together, our data suggested that MITF gene expression and downstream pathway is selectively regulated by p300/CBP inhibitor in melanoma.

Because we observed that p300/CBP inhibitor–sensitive cell lines express higher MITF mRNA and protein than resistant cell line, we hypothesize that MITF level may play a role in predicting sensitivity to p300/CBP inhibitor in melanoma cell lines. We first evaluated the effect of A-485 in a panel of 17 melanoma cell lines using a 5-day senescence assay. On the basis of the staining results, we divided melanoma cell lines into the senescence (>+) and no senescence (−) groups. Representative images of the staining results were shown in Supplementary Fig. S5A. We found that eight cell lines underwent senescence and six cell lines did not with A-485 treatment while three cell lines showed low levels of senescence (+) (Supplementary Fig. S5B). To further dissect the differences between the melanoma cell lines in response to p300/CBP inhibitor, we utilized data from the Cancer Cell Line Encyclopedia (CCLE; http://cbioportal.org) to determine whether there is differential gene expression of MITF between these two groups (29). We found that senescence cell lines were associated with higher MITF expression than the no senescence cell lines (P = 0.02; Fig. 3A). In contrast, SOX10 was not differentially expressed between senescence and no senescence cell lines (P = 0.55). EP300 and CREBBP (encoding CBP) were also not differentially expressed between the two groups of melanoma cell lines (Fig. 3A). Furthermore, we identified 430 genes that are differentially expressed with at least 2-fold change and P < 0.05 between the senescence and no senescence groups. Twenty-six out of these 430 genes, including MITF, were identified by IPA to belong to the MITF gene signature (Supplementary Fig. S5C). Next, we compared the copy number of MITF in the senescence and no senescence groups. We observed that cell lines that underwent senescence have higher MITF copy number than nonsenescence cell lines (P = 0.002; Fig. 3B). Finally, we determined the protein expression of MITF in senescence and no senescence cell lines upon A-485 treatment. We showed that cell lines that undergo senescence upon A-485 treatment have higher basal expression of MITF than cell lines that do not (Fig. 3C). In addition, we showed that there was a good correlation between MITF protein expression and viability/proliferation (IC₅₀) to A-485 in melanoma cell lines (Fig. 3D). Taken together, these findings indicate that the expression and amplification of MITF may play a role in determining sensitivity to p300/CBP inhibitor in melanoma cell lines. This also suggests that MITF gain/amplification could be used to stratify patients for p300/CBP inhibitor treatment because previous study showed that up to 15% of patients with metastatic melanoma exhibited amplification of MITF (11).

Histone acetyltransferases have been suggested to function as transcriptional coactivators and associate with activated gene expression. We sought to understand whether p300 inhibition regulates transcription by affecting the p300-mediated acetylation and/or p300 binding near the gene promoter regions. We used ChIP followed by qPCR (ChIP-qPCR) to detect histone H3 acetylation and p300 levels at the promoter region of MITF and a few strongly inhibited MITF target genes (MLANA, SEMA6A, TRPM1, and TDRD7) identified from our microarray results in WM2664 and SKMEL5 melanoma cells treated with A-485 (Fig. 4A). Notably, histone H3 acetylation level at the promoter regions of MITF and its signature genes were reduced in A-485–sensitive cell lines after a 24-hour exposure to A-485 (Fig 4B), consistent with our microarray data. However, we did not observe p300 displacement at the promoter regions of these genes (Fig. 4C), suggesting that the p300/CBP inhibition affects its histone acetylation activity at the MITF gene promoter regions without interfering with p300 chromatin binding (Fig. 4D).

The role of MITF in melanoma has been described but targeting MITF has been challenging because it is a transcription factor. In this study, we show that a selective p300/CBP inhibitor A-485 induced senescence in more than half the melanoma cell lines tested. The majority of these cell lines have high MITF expression. In addition, MITF pathway was the most significantly inhibited pathway in p300/CBP inhibitor-sensitive cell lines as compared with an insensitive cell line suggesting that MITF pathway could be targeted with a p300/CBP inhibitor. MITF expression and DNA copy number also differs between the senescence versus no senescence cell lines, suggesting patients could be selected for this treatment. Interestingly, a melanoma cell state with high MITF expression dictates sensitivity to RAF inhibitor suggesting that a combination of p300/CBP inhibitor and RAF inhibitor in these patients could be more efficacious (8, 9).

Finally, p300/CBP inhibitor decreased Histone-H3 acetylation without displacing p300 at promoters of MITF and its putative downstream genes. In addition, our current finding is consistent with our recent report that A-485 inhibited proliferation in lineage-specific tumor types, such as androgen receptor–positive prostate cancer (19). On the basis of these observations, p300/CBP inhibitor could be further exploited as a potential therapy for treating MITF-amplified melanoma.

A. Lai has ownership interest (including stock, patents, etc.) at Abbvie Stock. No potential conflicts of interest were disclosed by the other authors.

Conception and design: R. Wang, Y. He, V. Robinson, Z. Yang, X. Lu, A. Lai, T. Uziel, L.T. Lam

Development of methodology: R. Wang, Y. He, Z. Yang, A. Lai, L.T. Lam

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R. Wang, Y. He, V. Robinson, Z. Yang, P. Hessler, L.M. Lasko, L.T. Lam

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R. Wang, Y. He, V. Robinson, Z. Yang, P. Hessler, L.M. Lasko, X. Lu, A. Lai, T. Uziel, L.T. Lam

Writing, review, and/or revision of the manuscript: R. Wang, Y. He, Z. Yang, P. Hessler, X. Lu, A. Bhathena, A. Lai, T. Uziel, L.T. Lam

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R. Wang, Z. Yang, L.T. Lam

Study supervision: A. Bhathena, A. Lai, T. Uziel, L.T. Lam

We thank the AbbVie Oncology Biomarkers group and epigenetic group for discussion and critical review of the manuscript. We thank Sujatha Jagadeeswaran for technical support; Saul Rosenberg, Ken Bromberg, and Josh Plotnik for critical review of the manuscript. The design, study conduct, and financial support for this research were provided by AbbVie.