BET inhibitor suppresses melanoma progression via the noncanonical NF-κB/SPP1 pathway

Abstract

Background: Bromodomain and extra-terminal domain (BET) inhibitors have shown profound efficacy against hematologic malignancies and solid tumors in preclinical studies. However, the underlying molecular mechanism in melanoma is not well understood. Here we identified secreted phosphoprotein 1 (SPP1) as a melanoma driver and a crucial target of BET inhibitors in melanoma. Methods: Bioinformatics analysis and meta-analysis were used to evaluate the SPP1 expression in normal tissues, primary melanoma, and metastatic melanoma. Real-time PCR (RT-PCR) and Western blotting were employed to quantify SPP1 expression in melanoma cells and tissues. Cell proliferation, wound healing, and Transwell assays were carried out to evaluate the effects of SPP1 and BET inhibitors in melanoma cells in vitro. A xenograft mouse model was used to investigate the effect of SPP1 and BET inhibitors on melanoma in vivo. Chromatin immunoprecipitation (ChIP) assay was performed to evaluate the regulatory mechanism of BET inhibitors on SPP1. Results: SPP1 was identified as a melanoma driver by bioinformatics analysis, and meta-analysis determined it to be a diagnostic and prognostic biomarker for melanoma. SPP1 overexpression was associated with poor melanoma prognosis, and silencing SPP1 suppressed melanoma cell proliferation, migration, and invasion. Through a pilot drug screen, we identified BET inhibitors as ideal therapeutic agents that suppressed SPP1 expression. Also, SPP1 overexpression could partially reverse the suppressive effect of BET inhibitors on melanoma. We further demonstrated that bromodomain-containing 4 (BRD4) regulated SPP1 expression. Notably, BRD4 did not bind directly to the SPP1 promoter but regulated SPP1 expression through NFKB2. Silencing of NFKB2 resembled the phenotype of BET inhibitors treatment and SPP1 silencing in melanoma. Conclusion: Our findings highlight SPP1 as an essential target of BET inhibitors and provide a novel mechanism by which BET inhibitors suppress melanoma progression via the noncanonical NF-κB/SPP1 pathway.

Keywords: BET inhibitor; SPP1; melanoma.; noncanonical NF-κB pathway.

Figure 1

Identification of secreted phosphoprotein 1 (SPP1) as a potential melanoma driver. (A) Schematic flowchart of gene selection in GSE15605 and GSE46517 from GEO database. (B) Heatmap of 70 genes that differentially expressed among normal skin (NS), primary (PM), and metastatic (MM) melanoma in both GSE15605 and GSE46517. (C) SPP1 expression in NS, PM, and MM from GSE15605 and GSE46517. The number of sorts: N (NS) = 16, N (PM) = 46, and N (MM) = 12 in GSE15605; N (NS) = 7, N (PM) = 31, and N (MM) = 73 in GSE46517. (D) SPP1 expression in NS and melanoma tissue from Oncomine database. (E) SPP1 expression in PM and MM from Xu et al., 2018(GSE8401) and TCGA. The number of sorts: N (PM) = 31 and N (MM) = 52 from Xu et al., 2018(GSE8401); N (PM) = 103 and N (MM) = 369 from TCGA. (F) SPP1 expression in three mutational signatures (BRAF, NF1, and RAS) and wild types (WT) of melanoma based on GEPIA (Gene Expression Profiling Interactive Analysis). The number of sorts: N (NS) = 558, N (BRAF) = 147, N (NF1) = 27, N (RAS) = 91 and N (WT) = 47. Error bars represent standard deviation (SD). P-values were calculated using unpaired two-tailed Student's t test. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 2

SPP1 acts as a biomarker for diagnosis and progression of melanoma through meta-analysis. (A-B) Forest plot of SPP1 expression between nevi and PM (A), between nevi and MM (B) based on immunohistochemistry staining. (C-E) Forest plot of SPP1 expression between health persons and MM (C), between PM and MM patients (D), between patients who were disease free (DF) for at least five years and MM patients (E) based on enzyme linked immunosorbent assay. (F) Sensitivity analysis.

Figure 3

Up-regulation of SPP1 predicts poor prognosis in human melanoma. (A-B) Relative expression levels of SPP1 in HEK-293T, HaCat cells, melanoma cell lines, and patients-derived melanoma short-term cultures (STCs) quantified by RT-PCR (A) and western blotting (B). (C-D) Quantification by RNA-seq of SPP1 expression between melanoma and melanoma-adjacent tissue (N (nontumor) = 11; N (tumor) = 19) (C), between melanoma and paired melanoma-adjacent tissue from Xiangya melanoma cohort (N = 11) (D). (E-F) The representative protein expression of SPP1 (E) and percent distribution (F) between normal skin and melanoma tissues in The Human Protein Atlas (https://www.proteinatlas.org). (G) SPP1 expression grouped by pathologic grades of melanoma from GEPIA. (H) Kaplan-Meier survival analysis of high and low SPP1 expression groups in Xiangya melanoma cohort. Error bars represent SD. *, P < 0.05; **, P < 0.01.

Figure 4

Knockdown of SPP1 inhibits melanoma cell proliferation, migration and invasion. (A-B) SPP1 knockdown efficiency in A375 and SK-MEL-28 cell lines stably transduced with control shRNA (shCtrl) or two SPP1 shRNAs (shSPP1 #1 and #2) quantified by RT-PCR (A) and western blotting (B). (C-D) Cell proliferation of A375 (C) and SK-MEL-28 (D) cells transduced with shCtrl, shSPP1 #1, and shSPP1 #2 were quantified by CCK-8 assay. (E-F) Scratch-wound healing assay of A375 and SK-MEL-28 cells transduced with shCtrl, shSPP1 #1, and shSPP1 #2. Six random areas were selected. Images (at 40X magnification) were taken at 0, 24, and 48 hours. (G) Invasiveness of A375 cells transduced with shCtrl, shSPP1 #1, and shSPP1 #2 were assessed by Transwell assays. Invaded cells were determined for 18 hours. Five random areas were selected. Images were taken at 200X magnification. (H) model pattern and picture of resected subcutaneous xenografted tumors. A375 cells (10⁶) were injected subcutaneously in the right flank of NSG mice. Tumors were resected and photographed at day 21 (N = 6 in each group). (I) Tumor volume were recorded every other day and calculated as ([length×width²]/2). (J) Tumor weights were recorded after tumor resection at day 21. All data were represented as mean ± SD of three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 5

BET inhibitors impede melanoma cell proliferation, migration, and invasion through SPP1. (A) Quantification by RT-PCR of SPP1 expression in A375 and SK-MEL-28 cells after treatment with BET inhibitor (NHWD-870), BRAF inhibitor (Vemurafenib) or MEK inhibitor (Trametinib) for 24 hours. (B-C) SPP1 expression in A375 and SK-MEL-28 cells after treatment with increasing doses of NHWD-870 for 24 hours quantified by RT-PCR (B) and western blotting (C). (D) Cell proliferation of A375 treated with vehicle (DMSO), NHWD-870 (5nM), JQ-1 (500nM) or BMS-986158 (10nM) were quantified by CCK-8 assay. (E-G) Scratch-wound healing assay (E-F) and Transwell assays (G) of A375 cells treated with vehicle, NHWD-870 (5nM), JQ-1 (500nM) or BMS-986158 (10nM). (H) Pictures of resected subcutaneous xenografted tumor in nude mice. A375 cells (10⁶) were injected subcutaneously. When the tumor reached 100mm³, NHWD-870 (1mg/kg) or vehicle (0.5% methyl cellulose + 0.1% Tween 80) were given orally once a day for five successive days and then two days off. Tumors were resected photographed at day 21 (N = 6 in each group). (I) Tumor volume were recorded twice per week by Vernier caliper measurement and calculated as ([length×width²]/2). (J) Identification by RT-PCR of SPP1 expression in tumor tissue with or without NHWD-870 treatment. All data were presented as mean ± SD of three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 6

Overexpression of SPP1 partially reversed the inhibition effects of BET inhibitors in melanoma. (A) The efficiency of SPP1 overexpression in A375 cells were evaluated by RT-PCR and western blotting. (B-D) The influence of SPP1 overexpression on proliferation of A375 cells treated with NHWD-870 (5nM) (B), JQ-1 (500nM) (C), and BMS-986158 (10nM) (D). (E-G) The influence of SPP1 overexpression on migration (E-F) and invasiveness (G) ability of A375 cells treated with NHWD-870 (5nM), JQ-1 (500nM), and BMS-986158 (10nM). All data were presented as mean ± SD of three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 7

BET inhibitor inhibits SPP1 expression via transcriptional inactivation of NFKB2. (A-C) BRD4 and SPP1 expression were quantified by RT-PCR (A-B) and western blotting (C) in A375 and SK-MEL-28 cells 48 hours post-transfection with siNC or siBRD4s. (D) BRD4 binding in NFKB1 and NFKB2 promoters among DMSO treated, BET inhibitor (NHWD-870) treated, siNC treated, and siBRD4 treated A375 cells (our CHIP-seq data), or DMSO treated, JQ-1 treated, and BRD4 overexpressed cells (Zhang et al., 2017). (E) RT-PCR analysis of NFKB1, NFKB2, RELA, RELB, and c-REL expression in A375 cells 48 hours post-transfection with siNC or siBRD4s. (F) Validation of BRD4 binding to the promoter of NFKB2 in A375 and NHWD-870-treated (4nM) A375 cells by ChIP-qPCR. (G-H) NFKB2 expression in A375 cells after treatment with increasing dose of NHWD-870 for 24 hours quantified by RT-PCR (G) and western blotting (H). (I) Identification by RT-PCR of NFKB2 expression in tumor tissue with or without NHWD-870 treatment. (J-K) SPP1 expression in A375 cells 48 hours post-transfection with siNC and siNFKB2s quantified by RT-PCR (J) and western blotting (K). All data were represented as mean ± SD of three independent experiments. ns, no significance; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 8

A model depicting that inhibition of BRD4 represses SPP1 expression by transcriptionally down-regulating NFKB2 in melanoma. BRD4 promotes NFKB2 expression via direct binding to the NFKB2 promoter. Inhibition of BRD4 by BET inhibitor decreases NFKB2 expression, then represses the expression of SPP1, resulting in impeding melanoma proliferation, migration, and invasion.