BET Bromodomain Inhibition Potentiates Ocular Melanoma Therapy by Inducing Cell Cycle Arrest

Abstract

Purpose: Ocular melanoma is a common primary malignant ocular tumor in adults with limited effective treatments. Epigenetic regulation plays an important role in tumor development. The switching/sucrose nonfermentation (SWI/SNF) chromatin remodeling complex and bromodomain and extraterminal domain family proteins are epigenetic regulators involved in several cancers. We aimed to screen a candidate small molecule inhibitor targeting these regulators and investigate its effect and mechanism in ocular melanoma.

Methods: We observed phenotypes caused by knockdown of the corresponding gene and synergistic effects with BRD inhibitor treatment and SWI/SNF complex knockdown. The effect of JQ-1 on ocular melanoma cell cycle and apoptosis was analyzed with flow cytometry. Via RNA sequencing, we also explored the mechanism of BRD4.

Results: The best tumor inhibitory effect was observed for the BRD4 inhibitor (JQ-1), although there were no statistically obvious changes in the shBRD4 and shBRD9 groups. Interestingly, the inhibitory effect of JQ-1 was decrease in the shBRD4 group. JQ-1 inhibits the growth of melanoma in various cell lines and in tumor-bearing mice. We found 17 of these 28 common differentially expressed genes were downregulated after MEL270 and MEL290 cells treated with JQ-1. Four of these 17 genes, TP53I11, SH2D5, SEMA5A, and MDGA1, were positively correlated with BRD4. In TCGA database, low expression of TP53I11, SH2D5, SEMA5A, and MDGA1 improved the overall survival rate of patients. Furthermore, the disease-free survival rate was increased in the groups with low expression of TP53I11, SH2D5, and SEMA5A.

Conclusions: JQ-1 may act downstream of BRD4 and suppress ocular melanoma growth by inducing G1 cell cycle arrest.

Ocular melanoma is a common primary ocular malignant neoplasia in adults.¹ It originates from melanocytes of the skin, uveal tract and conjunctiva.² Uveal melanoma (UM) accounts for 82.5% of ocular melanomas, whereas conjunctival melanoma³ is rare.⁴ UM is one of the most common subtypes of melanoma, only behind cutaneous malignant melanoma (CM).⁵ However, the incidence of UM varies among different districts owing to different criteria for diagnosis.⁶ Primary UM arises in the choroid in approximately 90% of cases, and 70% of patients with early-stage disease present with ocular symptoms, including defects of the visual field, blurred vision, photopsia, and heterochromia iridis.⁷ With the greatest propensity to metastasize to liver, UM causes high mortality, for which there are limited effective treatment options.⁸ CM almost occurs in white people, and its incidence increases with age.⁹ Metastases occur in 21% to 26% of CM patients.¹⁰ CM has a propensity for local node invasion and systemic metastases.¹¹ Once systemic metastases develop, the median survival time is only 8 months.¹¹ Additionally, the recurrence rate after treatment is 36% to 56%.¹⁰ 

Overall, ocular melanoma is still responsible for the loss of eyesight and death of a significant proportion of patients. There are several therapies for ocular melanoma, including charged-particle radiation therapy, enucleation surgery, and systemic chemotherapy.¹² However, 50% of patients still develop metastatic disease. Furthermore, there is no accepted standard of care for the treatment of UM, and most current therapies for UM are adapted from those for CM, although UM shows different clinical and molecular features from CM.¹³ Therefore, novel strategies for the treatment of ocular melanoma are warranted to improve the prognosis of patients. 

The switching/sucrose nonfermentation (SWI/SNF) complex was first identified to be critical for cellular responses to mating type SWI and SNF in yeasts.¹⁴ The SWI/SNF complex consists of one ATPase subunit, either BRAHMA (BRM) or Brahma-related gene-1 (BRG1), and multiple core subunits, which are referred to as Brg1/Brm–associated factors.¹⁵ It has become apparent that the SWI/SNF complex is a critical epigenetic regulator of tumorigenesis that exerts its effects via its pleiotropic roles in the regulation of the cell cycle, oncogenic pathways, and metabolism.¹⁶ The SWI/SNF complex plays essential roles in a variety of cellular processes, including differentiation, proliferation, and DNA damage repair.¹⁷ Other important proteins that play pivotal roles in epigenetic processes, the bromodomain and extraterminal domain¹⁸ family proteins, include BRD2, BRD3, BRD4, and BRDT,¹⁹ and they are characterized by the presence of two tandem bromodomains (BD1 and BD2) and a C-terminal domain.²⁰ With a high affinity for proteins with multiple acetylated residues,¹⁹ the bromodomains of the bromodomain and extraterminal domain (BET) family proteins can interact with hyperacetylated histone regions along chromatin, where they recruit other regulatory complexes to influence gene expression.²¹ BRD4, a well-studied member of the BET family, plays an important role in various biological processes by mechanisms including coactivating NF-κB proinflammatory functions and inducing cell cycle and tumorigenic genes.²² Recent evidence has revealed the complexity of the roles of BRD4 in malignances; for example, Wen et al.²³ demonstrated that the inhibition of BRD4 can induce cell cycle arrest and apoptosis of glioma stem cells through the VEGF/PI3K/AKT signaling pathway, and Latif et al.²⁴ proved that BRD4-mediated repression of p53 is a target for combination therapy in acute myeloid leukemia. BRD9 has been identified as a component of the SWI/SNF complex and also harbors a single bromodomain and plays an important role in tumor development.²⁵ For example, Hohmann et al.²⁶ demonstrated that BRD9 is required to sustain MYC transcription and cell proliferation and block differentiation in acute myeloid leukemia cells. Given the important epigenetic roles of the SWI/SNF complex and BET family, inhibitors of these proteins are regarded as promising strategies for suppressing malignancies. For example, JQ-1, an inhibitor of BRD4, can competitively inhibit the binding of acetylated lysine sites with BRD4. The applications of JQ-1 have been represented in various tumors, such as retinoblastoma and salivary adenoid cystic carcinoma.^27,28 However, the mechanism of JQ-1 in ocular melanoma remains unclear. Therefore, our study aimed to reveal the effect of JQ-1 on ocular melanoma, including UM and CM, and clarify the mechanism. 

Materials and Methods

Cell Lines and Culture

The human ocular melanoma cell lines 92.1,²⁹ MEL202,³⁰ MEL270,³¹ MEL285,³⁰ MEI290,³⁰ OMM1,³² and OMM2.3³¹ were supplied by Professor John F. Marshall (Tumor Biology Laboratory, Cancer Research UK Clinical Center, John Vane Science Centre, London, UK). The human CM cell lines CRMM1, CRMM2,³³ and CM2005.1³⁴ were obtained from Leiden University Medical Center by Prof. Martine J. Jager. The HEK293T human embryonic kidney cell line was purchased from the American Type Culture Collection (Manassas, VA, USA). Human 92.1, MEL202, MEL270, MEL285, MEI290, OMM1, and OMM2.3 cells were cultured in Roswell Park Memorial Institute (RPMI) 1640 medium (Gibco, Evansville, IN, USA). The human CM cell lines CRMM1 and CRMM2 were cultured in Ham's F-12K (Kaighn's) medium (Gibco), and the CM2005.1 cell line was cultured in RPMI 1640 HEPES (Gibco). HEK293T human embryonic kidney cells were cultured in Dulbecco's modified Eagle's medium (Gibco). All media were supplemented with 10% fetal bovine serum (Thermo Fisher Scientific, Waltham, MA, USA) and 1% penicillin-streptomycin (Thermo Fisher Scientific). All cell lines were tested for mycoplasma contamination. 

Ethical Approval and consent to participate. All experimental protocols were approved by the Animal Care and Use Committee at Shanghai Jiao Tong University School of Medicine. All methods are reported in accordance with ARRIVE guidelines (https://arriveguidelines.org) for the reporting of animal experiments. All experiments were performed in accordance with relevant guidelines and regulations. 

Small Molecule Inhibitors

Anti-BRD4, anti-BRD9, and anti-BRG1 antibodies were purchased from CST (Shanghai, China). Anti–β-actin antibody was purchased from Proteintech (Wuhan, China). BET bromodomain inhibitors and SWI/SNF complex inhibitors were obtained from Selleck (Shanghai, China). 

Plasmid Construction

shRNA sequences targeting BRG1 (shBRG1-1: CCGGCCATATTTATACAGCAGAGAACTCGAGTTCTCTGCTGTATAAATATGGTTTTT), for BRD4 (shBRD4-1: CCGGCCTGGAGATGACATAGTCTTACTCGAGTAAGACTATGTCATCTCCAGGTTTTT), and BRD9 (shBRD9-1: CCGGGAGAGCACACCTATTCAGCAACTCGAGTTGCTGAATAGGTGTGCTCTCTTTTTTG) were obtained by polymerase chain reaction and then cloned into the Plko.1 vector. 

Lentivirus Packaging and Transfection

HEK293T cells were transfected with 10 µg packaging plasmid (PsPax), 5 µg envelope plasmid (pMD.2G), 5 µg shRNA and 60 µL PolyJet (Sigma-Aldrich, St. Louis, MO, USA). Twelve hours after transfection, the medium was replaced with 10 mL fresh medium. Then, the supernatant was collected and filtered with a 0.45-µm cellulose acetate filter. Twenty-four hours before transfection, the cells were seeded at 1.0 × 10⁵ cells per well in six-well dishes, and the medium was replaced with fresh medium and virus-containing supernatant at a ratio of 1:1. Then, 10 ng/mL polybrene (Sigma-Aldrich) was added. After 48 hours, the medium was replaced with fresh medium. The cells were selected with 2 mg/mL puromycin (InvivoGen, San Diego, CA, USA) for 3 weeks in fresh medium. 

Immunoblotting

Cells were harvested and washed with phosphate-buffered saline (PBS) three times. Cellular proteins were extracted with lysis buffer for 30 minutes and then centrifuged at 15,000×g for 20 minutes at 4°C. Proteins were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to 0.2-µm polyvinylidene fluoride membranes. The membranes were then blocked in 5% milk for 1 hour at room temperature and incubated with primary antibody in 5% bovine serum albumin overnight at 4°C. The blots were washed three times in TBS-T and incubated with secondary antibodies for 1 hour at room temperature. The band signals were then visualized using an Odyssey Infrared Imaging System (LI-COR, Lincoln, NE, USA). 

Cell Viability Analysis

The cells were seeded in 96-well plates at a density of 2000 cells per well for 24 hours. Cell viability was evaluated by the Cell Counting Kit-8 (CCK-8; Dojindo, CK04) assay. Then, the optical density value of each well at 485 nm was obtained with a microplate reader (Thermo Fisher Scientific). 

Colony Formation Assay

The cells were seeded in a six-well plate at a density of 1000 cells per well with different concentrations of JQ-1 and incubated for 14 to 21 days. Then, 0.5% crystal violet was used to stain the colonies. 

Flow Cytometry

The cells were seeded into 6-well plates at a density of 100,000 cells per well. These cells were treated with JQ-1 at a concentration of 0.1 µM. The cells were harvested and washed with cold PBS and then fixed with cold ethanol. After being resuspended and rehydrated with PBS, the cells were stained with propidium iodide and annexin V and incubated for 30 minutes away from light at room temperature. Finally, the cells were analyzed with a FACSCalibur flow cytometer (BD Science, Franklin Lake, NJ, USA). 

RNA Sequencing (RNA-seq)

We performed mRNA-seq of cells (MEL270 and MEL290) treated with JQ-1 (0.04 µM) or DMSO. Cells were harvested and extracted using TRIzol reagent (Invitrogen, Waltham, MA, USA) according to the manufacturer's instructions and quantified with a NanoDrop 2000 (Thermo Fisher Scientific). Then, 0.5 mg purified RNA was transcribed to cDNA. The Agilent RNA 6000 Nano Kit (Agilent Technologies, Santa Clara, CA, USA) was used to assess RNA integrity with an Agilent 2100 Bioanalyzer. The GeneChip 3′ IVT labeling kit (Affymetrix, Santa Clara, CA, USA) was used to synthesize biotin-labeled RNA, which was then hybridized onto the microarrays. After sample labeling, the microarray hybridization and washing steps were conducted following the manufacturer's instructions, and the arrays were directly scanned by an Affymetrix Scanner 3000 (Affymetrix). Differentially expressed genes (DEGs) were identified by a threshold of absolute fold change ≥ 2 (P < 0.05). The sequencing data have been uploaded into the Gene Expression Omnibus Database (https://www.ncbi.nlm.nih.gov/geo/) under the accession number GSE233589. 

Tumor Model and In Vivo Treatments

NOG/scid female mice were housed in specific pathogen-free conditions. A total of 8 × 10⁶ cells (MEL270) with Matrigel (Corning, Corning, NY, USA) were injected subcutaneously into the right flank of mice. Tumor size was measured every 4 days by electronic caliper, and the mean tumor volume was calculated using the formula: tumor length (mm) × tumor width (mm) × tumor width (mm) × 1/2. Tumor-bearing mice were treated with JQ-1 or PBS at a daily dose of 30 mg/kg in a single intraperitoneal injection. A cervical dislocation was followed by isoflurane inhalation for euthanasia. The procedure of euthanasia followed the recommendations for euthanasia of experimental animals. 

Results

The BRD4 Inhibitor (JQ-1) Demonstrated the Best Tumor-suppressive Effect

To investigate the effect of drugs on ocular melanoma cells, we treated MEL270 cells with DMSO, BRG1 inhibitor, BRM inhibitor, JQ-1 (BRD4 inhibitor), and BRD9 inhibitor (BRD9i), and colony formation assays showed different cell viabilities after drug treatment (Fig. 1A). After analyzing the colony number, we found that the viability of MEL270 cells was decreased significantly after treatment with JQ-1. This result suggests that BRD4 may play a vital role in the development of ocular melanoma cells. Then, to explore whether the expression levels of the SWI/SNF complex and BET proteins are related to the viability of ocular melanoma cells, we knocked down the expression of BRG1, BRD4, or BRD9 in MEL270. After BRG1, BRD4, and BRD9 were successfully knocked down (Fig. 1B), we performed CCK-8 assays, which showed that the viability of cells in the shBRG1 group was decreased significantly, although there were no statistically obvious changes in the shBRD4 and shBRD9 groups (Fig. 1C). These results showed that the inhibitory effect of JQ-1, an inhibitor of BRD4, was more obvious than that of BRD4 knockdown, which may indicate that JQ-1 induces the inhibition of ocular melanoma by acting on not only BRD4, but also other targets. 

Figure 1.

 

The effect of inhibitor treatment or knockdown of candidate genes on cell line proliferation. (A) Colony information assay of MEL270 cells treated with DMSO, BRG1 inhibitor (BRG1i), BRM inhibitor (BRMi), BRD4 inhibitor (BRD4i; JQ-1), or BRD9i. The colony numbers/areas were determined from soft agar plates. (B) Western blotting showing BRG1, BRD4 and BRD9 expression in the cell line (MEL270) after the stable cell lines were generated. The signals were quantified and normalized to that of β-actin. (C) Cell viability of negative control (NC) cells or cells with knockdown of BRG1, BRD4 or BRD9, **P < 0.01. (D) Colony information assay showed the changes in the proliferation of MEL270 cells after application of BRG1i, BRMi, BRD4i (JQ-1) or BRD9i in BRG1 knockdown cell line and BRD4 knockdown cell line. The colony areas were determined from soft agar plates *P < 0.05, ***P < 0.001.

The Combination of BRD Inhibitor Treatment and SWI/SNF Complex Knockdown Inhibits the Growth of Ocular Melanoma

Based on these data, we wondered whether there is some association between inhibitor treatment and knockdown of the SWI/SNF complex and BET proteins. To assess our hypothesis, we administered inhibitors of the SWI/SNF complex and BET proteins to shBRG1 group and shBRD4 group cells. As shown in Figure 1D, the viability of the shBRD4 group was decreased significantly after treatment with JQ-1, and the shBRG1 group and shBRD4 group also had a decreased colony numbers after administration of BRD9i, which suggested that BRD9i may have synergistic effects with the knockdown of BRG1 or BRD4. Similarly, there may also be synergistic effects of BRM inhibitor and shBRG1 treatment. Furthermore, after BRD4 was knocked down, the inhibitory effect of JQ-1 was decreased, which may suggest that the mechanism of JQ-1 is related to BRD4. All of these data demonstrated that the combination of inhibitors of BRD proteins and knockdown of the SWI/SNF complex may lead to a synergistic effect on ocular melanoma. 

JQ-1 Inhibits the Growth of Melanoma In Vitro and In Vivo

To further investigate the inhibitory effect of JQ-1 on melanoma cells, we applied different concentrations of JQ-1 to CM and UM cells, and the colony formation assays (Fig. 2A) showed that the number of colonies decreased sharply at a concentration of 0.2 µM. We also obtained the half-maximal inhibitory concentration (IC50) of JQ-1 in different ocular melanoma cells through data fitting (Fig. 2B). In addition, to verify the inhibitory effect of JQ-1 in vivo, we established a xenograft model in NOG/scid mice. We injected MEL270 cells into mice subcutaneously. Then, JQ-1/PBS was given to these tumor-bearing mice. We evaluated the size of tumors every 4 days and assessed tumor weight after the experiment ended. These data (Figs. 2C–E) showed that the tumor volume and weight of the MEL270 group were decreased significantly compared with those of the control group. Therefore, the efficacy of JQ-1 was revealed in vivo. These data were consistent with the in vitro results. 

Figure 2.

 

The efficacy of JQ-1 in inhibiting tumor growth in vitro and vivo. (A) Colony assays of different cells treated with JQ-1 at different concentrations (0 µM, 0.0016 µM, 0.008 µM, 0.04 µM, 0.2 µM, and 1 µM). (B) The viability of different cells was evaluated, and these data were fitted to determine the IC50 of JQ-1 in different cells. (C) Tumors derived from cells with (top) or without (bottom) JQ-1 (30 mg/kg) treatment were removed from the mice. (D) Tumor volumes were evaluated every 4 days after the injection of MEL270 cells for 24 consecutive days (n = 6 mice per group). (E) Tumor weights were obtained after the tumors were removed from mice. Data are presented as the mean ± SEM.

JQ-1 Arrests the Melanoma Cell Cycle

To investigate the mechanism of JQ-1 in ocular melanoma, we administered JQ-1 to CM and UM cells at a concentration of 0.1 µM for 3 consecutive days. Flow cytometry assays (Fig. 3) showed that JQ-1 induced G1 cell cycle arrest in CRMM1, CRMM2, MEL202, MEL270, MEL290, and 92.1 cells. Furthermore, compared with the control group, the JQ-1–treated group had no significant change in apoptosis. These results demonstrated that JQ-1 regulates melanoma progression by modulating the cell cycle, whereas it may have weak effects on apoptosis. 

Figure 3.

 

The effects on the cell cycle and cell apoptosis after JQ-1 treatment. (A) Different cell lines (OMM1, CRMM1, CRMM2, MEL202, MEL270 MEL285, MEL290, and 92.1) were treated with JQ-1 at a concentration of 0.1 µM for 3 days and harvested for the flow cytometry assay on the fifth day. The percentage of apoptotic cells was determined by flow cytometry. All histograms show the percentage (%) of apoptotic cells from one independent experimental group. (B and C) Different cell lines (CRMM1, CRMM2, MEL202, MEL270, and MEL290) were treated with JQ-1 at a concentration of 0.1 µM for 3 days and harvested for the flow cytometry assay on the fifth day. Cell cycle analysis by flow cytometry was performed to determine the percentage of cells in different cell cycle phases. All histograms show the percentage (%) of cells in each phase of the cell cycle from one independent experimental group.

Identifying Downstream Targets of JQ-1

To further study the mechanism of JQ-1, we performed RNA-seq analysis of the cells (MEL270 and MEL290) that were treated with JQ-1 (0.04 µM) and the control cells treated with DMSO. We then performed Gene Set Enrichment Analysis and found that, among the top 30 pathways, there were 6 commonly enriched biological processes in both MEL270 and MEL290 cells after treatment with JQ-1 (Fig. 4A). These six pathways were DNA methylation, ERCC6 and EHMT2, SIRT1 negatively regulated rRNA expression, B wich complex, meiotic recombination, and epigenetic regulation of rRNA expression (Fig. 4B). Interestingly, these six pathways are mostly related to the regulation of the cell cycle. In addition, we found 28 common DEGs in MEL270 and MEL290 cells (Fig. 5A), and 17 of these 28 DEGs were downregulated after cells were treated with JQ-1 (Fig. 5B). By comparing our data with those in The Cancer Genome Atlas³⁵ database, we found that four of these 17 genes, TP53I11, SH2D5, SEMA5A, and MDGA1, were positively correlated with BRD4 (P < 0.05) (Fig. 5C). After analyzing the relationship between these genes and the survival rate from the TCGA database, we found that the expression level of BRD4 had no significant effect on prognosis. However, low expression of TP53I11, SH2D5, SEMA5A, and MDGA1 improved the overall survival rate of patients (P < 0.05). Furthermore, the disease-free survival rate was increased in the groups with low (vs. high) expression of TP53I11, SH2D5, and SEMA5A (P < 0.05), and the difference in the disease-free survival rate between the low and high expression groups for MDGA1 was almost statistically significant (Fig. 6). Taken together, these data suggest that the downstream targets of BRD4 might have clinical significance in UM. 

Figure 4.

 

Gene set enrichment analysis (GSEA) of DEGs. (A) GSEA of DEGs induced by JQ-1 in MEL270 and MEL290 cells. Each column represents a signature. The red columns represent the identical signatures among the top 30 pathways for MEL270 and MEL290 cells. NES, normalized enrichment score. (B) Representative enriched pathways among DEGs through GSEA.

Figure 5.

 

Gene set enrichment analysis (GSEA) of DEGs and the correlations of downstream gene with patient outcome. (A) Heatmap of all DEGs in DMSO- versus JQ-1-treated MEL270 and MEL290 cells. Each column represents a DEG where the values represent normalized counts that were standardized to z-scores. The colors indicate upregulation (orange) and downregulation (blue) relative to the mean (white). (B) DEGs induced by JQ-1 in MEL270 and MEL290 cells. The blue circle indicates the DEGs of MEL270, while the red circle represents that of MEL290. The cross-section indicates the DEGs in both MEL270 and MEL290 cells. The cross section includes 28 genes, all of which are listed below the circles. The red color indicates that the genes are downregulated. (C) The correlations between BRD4 and some downregulated genes induced by JQ-1 in both MEL270 and MEL290 cells, including TP53I11, SH2D5, SEMA5A, and MDGA1.

Figure 6.

 

The correlations of DEGs with BRD4. Overall and disease-free survival plots of UM patients from the Gene Expression Profiling Interactive Analysis (GEPIA) database. According to the expression of different genes, patients were divided into two groups: a high-expression group (red line) and a low-expression group (blue line).

Discussion

As the molecular pathogenesis of ocular melanoma has become clarified, some available treatments that target underlying molecules have become a research focus. For example, UM is characterized by mutations in the G protein subunits GNAQ and GNA11.^36,37 These proteins modulate the signaling pathways that lead to the activation of downstream components, including RAF, MEK, and ERK. Carvajal et al.³⁸ demonstrated that, compared with chemotherapy, MEK inhibition with selumetinib had promising clinical efficacy in a randomized phase II trial. Moreover, with our improved understanding of oncogenic drivers in ocular melanoma, more promising therapeutic targets will be identified. Furthermore, it was reported that GNAQ/11 mutant cells were extremely dependent on BRD4 activity compared with those cells without the mutations,³⁹ and the inhibitor of BRD4 had proapoptotic effects in cells with GNAQ/11 mutations.⁴⁰ 

As a member of the BET family, BRD4 is involved in the control of transcriptional elongation by RNA polymerase Ⅱ through the recruitment of the positive transcription elongation factor P-TEFb.⁴¹ BRD4 was first identified as a cell cycle controlling protein that stimulates G1 gene expression by binding to multiple G1 gene promoters in a cell cycle-dependent manner.⁴² Recent studies add further complexity to the role of BRD4 in cancer, showing that BRD4 may be related to processes such as DNA damage repair and checkpoint activation.⁴³ As a transcriptional regulator, BRD4 recruits a transcriptional regulatory complex to acetylated chromatin to govern the expression of a series of proteins, such as c-MYC.⁴⁴ Much of what is known of BRD4 involvement as an oncogene or nononcogene in cancer has been discovered through the use of BET inhibitors, especially the JQ-1 BET inhibitor created by Jay Bradner and his chemist Jun Qi, for whom the molecule was named.²² Since the introduction of JQ-1, hundreds of papers have been published implicating roles of BRD4 in tumors, including lymphoma, acute leukemia, multiple myeloma, prostate cancer, neuroblastoma, breast cancer, pancreatic cancer, and non–small-cell carcinoma of the lung.^45–48 Wen et al.²³ found that JQ-1 had notable antitumor effects against glioblastoma multiforme, which may be mediated by the VEGF/PI3K/AKT pathway. Owing to the inhibitory effect of JQ-1 on the classic ERα signaling pathway, in vivo anti-breast cancer activity was observed in a tamoxifen-resistant breast cancer xenograft mouse model. This strong long-lasting effect of JQ-1 could enhance the activity of the estrogen receptor degrader fulvestrant by significantly downregulating the expression levels of estrogen receptor α and its target genes GREB1, pS2, and cyclin D1. Nevertheless, the expression of other breast cancer genes, such as Her2, FoxA1, and SRC-3, was not affected.⁴⁹ Zhanget al.²⁷ explored the role of JQ-1 in retinoblastoma and found that JQ-1 could effectively inhibit the proliferation and colony formation ability of retinoblastoma cells by inducing cell cycle arrest and promoting tumor cell apoptosis. Moreover, the myc-p21-CDK2 and myc-cyclin D3/CDK6 signaling pathways of retinoblastoma cells can be activated under the effect of JQ-1, and animal experiments further confirmed the role of JQ-1 in inhibiting tumor formation in vivo.²⁷ Although studies related to the mechanism of JQ-1 have been popular in recent years, the molecular mechanism of JQ-1 in ocular melanoma remains unclear. 

Our study found that JQ-1 can inhibit the progression of ocular melanoma, including UM and CM. Validation of JQ-1 was performed in multiple ocular melanoma cell lines with different tissue origin and biological features, which may significantly affect JQ-1 sensitivity. JQ-1 demonstrates lower selectivity for conjunctive melanoma (CM2005.1, IC50 = 0.08825 µM and CRMM2, IC50 = 0.1708 µM), as compared with cells of UM. The cell lines with different driver mutations exhibited differential drug responses. CRMM2 harbouring NRAS mutation have weak sensitivity to JQ-1.⁵⁰ UM with different GNAQ/11-mutant type may also affect sensitivity to JQ-1.⁵¹ UM cells from different genetic backgrounds responded significantly differently to same treatment, as previous studies have reported.⁵¹ Furthermore, JQ-1 may induce G1 arrest in ocular melanoma cells and has no obvious impact on apoptosis. Bioinformatic analysis showed that after JQ-1 was administered, DEGs were enriched in epigenetic pathways, such as DNA methylation, ERCC6, and EHMT2. SIRT1 negatively regulated rRNA expression, and most of these pathways were related to regulation of the cell cycle. Furthermore, high expression of the downstream genes of BRD4, including TP53I11, SH2D5, SEMA5A, and MDGA1O, were related to a poor prognosis in patients. TP53I11 was reported to be involved in the negative regulation of cell population proliferation,⁵² and SH2D5 was confirmed to regulate levels of Rac1-GTP.⁵³ Furthermore, SEMA5A was also involved in glioma cell motility and morphology,⁵⁴ while the MDGA1 contributed to the inhibitory synapse development.⁵⁵ Among these four genes, TP53I11 was frequently reported to be associated with tumor growth. For example, TP53I11 was involved in the apoptosis of MGC-803 tumor cells,⁵² and some work confirmed that the miR-1234/TP53I11 axis was associated with the tumorigenesis of gastric cancer.^52,56 TP53I11 may be an interesting gene to study further. All of these results suggest that JQ-1 not only inhibits the binding of BRD4 but also suppresses the expression of the downstream targets of BRD4. In general, BRD4, not only as the direct target of JQ1, but also as the upstream molecular, affects a cascade of multiple downstream genes, achieving better therapeutic effects. 

In conclusion, our study highlights the potential for the therapeutic application of JQ-1 in ocular melanoma. Furthermore, we demonstrated that the combination of JQ-1 and inhibitors of downstream targets of BRD4 may achieve better therapeutic outcomes than a single approach. Although the experimental results have confirmed the inhibitory effect of JQ-1 on ocular melanoma, further studies on the pathogenesis of ocular melanoma and the mechanism of JQ-1 are still needed to promote the further applications of JQ-1 in tumors. 

Acknowledgments

The authors thank Shengfang Ge and Hanhan Shi for the critical reading of this manuscript. 

Supported by the National Natural Science Foundation of China (grants 81570884 and 81770961), the Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning (No. TP2018046 to J.S.), the Shanghai Municipal Education Commission-Two Hundred Talents (No. 20191817 to J.S.), and the General Program of National Natural Science Foundation of China (No. 81972667 to J.S.). 

Disclosure: X. Chen, None; R. Huang, None; Z. Zhang, None; X. Song, None; J. Shen, None; Q. Wu, None