Ubiquitin receptor PSMD4/Rpn10 is a novel therapeutic target in multiple myeloma

Abstract

PSMD4/Rpn10 is a subunit of the 19S proteasome unit that is involved with feeding target proteins into the catalytic machinery of the 26S proteasome. Because proteasome inhibition is a common therapeutic strategy in multiple myeloma (MM), we investigated Rpn10 and found that it is highly expressed in MM cells compared with normal plasma cells. Rpn10 levels inversely correlated with overall survival in patients with MM. Inducible knockout or knockdown of Rpn10 decreased MM cell viability both in vitro and in vivo by triggering the accumulation of polyubiquitinated proteins, cell cycle arrest, and apoptosis associated with the activation of caspases and unfolded protein response-related pathways. Proteomic analysis revealed that inhibiting Rpn10 increased autophagy, antigen presentation, and the activation of CD4+ T and natural killer cells. We developed an in vitro AlphaScreen binding assay for high-throughput screening and identified a novel Rpn10 inhibitor, SB699551 (SB). Treating MM cell lines, leukemic cell lines, and primary cells from patients with MM with SB decreased cell viability without affecting the viability of normal peripheral blood mononuclear cells. SB inhibited the proliferation of MM cells even in the presence of the tumor-promoting bone marrow milieu and overcame proteasome inhibitor (PI) resistance without blocking the 20S proteasome catalytic function or the 19S deubiquitinating activity. Rpn10 blockade by SB triggered MM cell death via similar pathways as the genetic strategy. In MM xenograft models, SB was well tolerated, inhibited tumor growth, and prolonged survival. Our data suggest that inhibiting Rpn10 will enhance cytotoxicity and overcome PI resistance in MM, providing the basis for further optimization studies of Rpn10 inhibitors for clinical application.

Graphical abstract

Figure 1.

Characterization of ubiquitin receptor Rpn10 in MM cells. (A) Dox-inducible Cas9 AMO1 cells were made by infecting AMO1 cells with lentivirus-packaged inducible Cas9 plasmid followed by G418 selection. These cells then were infected with either control single-guide RNA (sgRNA) or Rpn13- or Rpn10-targeted sgRNA and selected with puromycin to generate stable iKO cells. Cell growth was evaluated by a CellTiter-Glo (CTG) assay. Immunoblot showing Rpn13 and Rpn10 expression, respectively, in corresponding KO cells to test the KO efficiency. Data are shown as the mean ± standard deviation (SD) of triplicates. (B) Gene expression data were collected using Affymetrix Human Genome U133A array platform. Rpn10 expression in the different stages of MM development, including normal CD138⁺ cells (n = 15), monoclonal gammopathy of undetermined significance (MGUS) (n = 22), smoldering (n = 24), newly diagnosed MM (n = 69), and relapsed MM (n = 28) (data accession number GSE6477). (C) Kaplan-Meier plots of Rpn10 expression vs overall survival of patients with MM. Analysis was performed using samples from 170 newly diagnosed patients with MM (data accession number GSE39754; P = .0064). (D) Total cellular RNA from purified CD138⁺ cells from BM of patients with MM or normal healthy donor PBMCs was subjected to reverse transcription polymerase chain reaction analysis, followed by real-time polymerase chain reaction using primers designed to recognize sequences internal to Rpn10 (mean ± SD, n = 4). (E) Immunohistochemistry analysis of BM biopsies from normal donors and MM patients showing Rpn10 expression (scale bar, 5 μm). (F) Protein lysates from a panel of MM cell lines, normal PCs, or primary cells from patients with MM were subjected to immunoblotting (IB) using Rpn10 and β-actin antibodies. Ab, antibody; CT, control; ns, not significant.

Figure 2.

Functional significance of Rpn10. (A) Dependency scores of Rpn10 across cancers based on CRISPR data sets in the DepMap web portal. A lower chronos score indicates that the gene of interest is essential in a given cell line. Score 0 means the gene is not essential, whereas score –1 is comparable with the median of all panessential genes (red line). (B) The proliferation of the MM.1S and ANBL6 inducible Rpn10-shRNA KD cell lines were analyzed using a CTG assay (mean ± SD, n = 3) when cultured with or without Dox. Inducible KD of Rpn10 was achieved using pTRIPz-mCherry vector containing Rpn10-shRNA or scramble control. (C) The stable adding back cell line was achieved when AMO1 Rpn10-iKO cells were transfected with lentivirus-packaged V5-tagged Rpn10 plasmid or empty plasmid (pEV), followed by blasticidin selection. Cell proliferation was measured by a CTG assay (mean ± SD, n = 3). Immunoblot shows the expression levels of Rpn10. Human AMO1 Rpn10-shRNA KD cells (D) or MM.1S Rpn10-shRNA KD cells (E-F) were subcutaneously inoculated into CB17 severe combined immunodeficiency mice. In the early prevention model (D-E), a cohort of mice was treated with an irradiated 0.0625% Dox diet (1-6 mg of Dox per mouse per day) continuously starting 5 days after injection. In the late prevention model (n = 10) (F), mice were treated with an irradiated 0.0625% Dox diet after the tumor became visible and the volume was ∼100 mm³. The average and standard deviation of tumor volume (mm³) are shown vs the time when the tumor was measured (mean tumor volume ± SD). Kaplan-Meier plots show survival in mice (right). Internal blot shows Rpn10 expression of the tumor lysates.

Figure 3.

Proteomic analysis of Rpn10 iKO. After being cultured with Dox for 4 days, AMO1 Rpn10-iKO cells were subjected to proteomic analysis by multiplexed proteomics with tandem mass spectrometry. (A) Volcano plot of differentially expressed genes in AMO1 Rpn10-iKO cells compared with those in control cells. Blue and red dots represent the genes (P < .05) that were downregulated or upregulated, respectively. (B) Heatmap showing significantly (false discovery rate [FDR] < .05) altered genes in Rpn10 iKO cells. (C) Gene set enrichment analysis (GSEA) and gene ontology biological process (GOBP) significantly enriched after Rpn10 iKO. For all pathways shown, FDR < 10%. (D) GSEA-derived enrichment plots for the lysosomal transport pathway (left). AMO1 Rpn10-iKO cells were cultured with Dox as described for the indicated time points (middle) and when cultured with Dox, 25 nM BafA1 was added to the culture of Rpn10 iKO cells for the last 16 hours of day 4 (right). Cells were then lysed and subjected to immunoblot using antibodies against p62, LC3, LAMP2, or β-actin. (E) GSEA-derived enrichment plots for antigen processing and presentation of peptide antigen (left). AMO1 Rpn10-iKO cells were cultured with Dox for 4 days. Cells were subjected to flow cytometry after staining with anti–HLA-DR or anti–HLA-DQ Abs and 7AAD (middle). 7AAD^– cells were gated out, and the surface expression of HLA-DR or DQ expression was quantified. Rpn10-iKO cells, from which Dox was washed out, were added to total BM-derived mononuclear cells of patients with MM (n = 4) (right). Degranulation marker CD107a was measured on CD3⁺CD4⁺ T cells or CD3^–/CD56⁺ natural killer (NK) cells. FC, fold change; MFI, mean fluorescence intensity.

Figure 3.

Proteomic analysis of Rpn10 iKO. After being cultured with Dox for 4 days, AMO1 Rpn10-iKO cells were subjected to proteomic analysis by multiplexed proteomics with tandem mass spectrometry. (A) Volcano plot of differentially expressed genes in AMO1 Rpn10-iKO cells compared with those in control cells. Blue and red dots represent the genes (P < .05) that were downregulated or upregulated, respectively. (B) Heatmap showing significantly (false discovery rate [FDR] < .05) altered genes in Rpn10 iKO cells. (C) Gene set enrichment analysis (GSEA) and gene ontology biological process (GOBP) significantly enriched after Rpn10 iKO. For all pathways shown, FDR < 10%. (D) GSEA-derived enrichment plots for the lysosomal transport pathway (left). AMO1 Rpn10-iKO cells were cultured with Dox as described for the indicated time points (middle) and when cultured with Dox, 25 nM BafA1 was added to the culture of Rpn10 iKO cells for the last 16 hours of day 4 (right). Cells were then lysed and subjected to immunoblot using antibodies against p62, LC3, LAMP2, or β-actin. (E) GSEA-derived enrichment plots for antigen processing and presentation of peptide antigen (left). AMO1 Rpn10-iKO cells were cultured with Dox for 4 days. Cells were subjected to flow cytometry after staining with anti–HLA-DR or anti–HLA-DQ Abs and 7AAD (middle). 7AAD^– cells were gated out, and the surface expression of HLA-DR or DQ expression was quantified. Rpn10-iKO cells, from which Dox was washed out, were added to total BM-derived mononuclear cells of patients with MM (n = 4) (right). Degranulation marker CD107a was measured on CD3⁺CD4⁺ T cells or CD3^–/CD56⁺ natural killer (NK) cells. FC, fold change; MFI, mean fluorescence intensity.

Figure 4.

Rpn10 blockade induces cell death. (A) Polyubiquitylated proteins were detected in AMO1 Rpn10-iKO cells, MM.1S Rpn10-shKD cells, and ANBL6 Rpn10-shKD cells by western blot analysis. (B) Protein lysates from indicated cell lines were subjected to immunoblotting using antibodies against Rpn10, p-eIF2α, PERK, BiP, p53, p21, PLK1, or β-actin. (C) AMO1 Rpn10-iKO cells were cultured with Dox for indicated time points and fixed in 70% ethanol. After washing with phosphate-buffered saline, cells were stained with propidium iodide, and the DNA content of cells was then analyzed using fluorescence-activated cell sorter. Bar graph shows percentage of cell populations in the G1, G2/M, or S phases of the cell cycle. (D) GSEA-derived enrichment plots for mitotic cell cycle pathway and cell cycle G2-M phase transition pathway (left) and heatmap showing proteins related to G2-M phase transition enriched in AMO1-CT cells (right) have been shown. MM.1S shRNA KD cells, ANBL6-BR shRNA KD cells, and AMO1 Rpn10-iKO cells were cultured with Dox for indicated time points and then analyzed for apoptosis with annexin V/4′,6-diamidino-2-phenylindole double staining assay (mean ± SD, n = 3) (E) or protein lysates were then subjected to immunoblotting using antibodies against caspase-3, caspase-9, PARP, or β-actin (F). (G) AMO1 sgRNA-CT cells transfected with lentivirus-packaged pEV, referred to as CT; AMO1 Rpn10-iKO cells transfected with lentivirus-packaged pEV vector, referred to as KO; and AMO1 Rpn10-iKO transfected with lentivirus-packaged V5-tagged Rpn10 plasmid, referred to as RES. Protein lysates from CT, KO, or RES cells were then subjected to immunoblotting using antibodies against Rpn10 or K48 polyubiquitin, p62, LC3, BiP, PERK, caspase-3, caspase-9, p21, or β-actin.

Figure 5.

Biochemical characterization of a novel inhibitor of Rpn10, SB. (A) AlphaAssay screening for Rpn10 inhibitor (binder). Human recombinant Rpn10-GST and biotinylated Ub_2-7 were used. (B) Percentage of inhibition of Rpn10-Ub_2-7 binding by the 10 000 compounds screened. Internal figure: chemical structure of Rpn10 inhibitor (binder) candidate SB. (C) Dose response AlphaScreen assay of SB binding to Rpn10 or GST protein. (D) Recombinant human Rpn10 protein was incubated with control, SB (10μM), or NSC697923 (10μM) for 30 minutes at real time, then Ub_2-7 was added and incubated for 1 hour at real time. The mixture was then immunoprecipitated with Rpn10 antibody and subjected to immunoblotting using antibodies against Rpn10 and poly-Ub. (E) Measurement of K_d of hRpn10 with SB by FEB assay. Ten micrometers of compound or DiUb was applied on a recombinant His-hRpn10–immobilized graphene chip separately. DiUb and compound UNC0638 were used as positive and negative controls, respectively. The K_d is calculated as a median average over the test points. Three independent experiments were performed. Real-time changes of the I-response are shown in circles and correspond to the 1:1 binding model shown as solid lines. (F) Measurement of K_d of hRpn10 with SB by MST assay. K_d was derived from the binding response as a function of the GFP-hRpn10 concentration. Error in K_d represents fitting errors. (G) After being cultured with Dox for 3 days, AMO1 sgRNA-CT cells, AMO1 Rpn10-iKO cells, and iKO with Rpn10 adding back cells were treated with SB at different concentrations for 24 hours, followed by the cell viability being measured by a CTG assay (mean ± SD, n = 3).

Figure 6.

Anti-MM activity of SB. (A) MM, melanoma, leukemia, and lymphoma cell lines were treated with dimethyl sulfoxide (DMSO) control or SB at different concentrations for 48 hours, followed by assessment for cell viability using a WST assay (P < .05 for all cell lines, n = 3). IC₅₀ was calculated and presented in the table. (B) Purified CD138⁺ cells from a patient with MM were treated with DMSO or SB for 48 hours, followed by assessment for cell viability using a CTG assay (mean ± SD of triplicate cultures). (C) Normal PBMCs from healthy donors were treated with DMSO or the indicated concentrations of SB for 48 hours, and then analyzed for cell viability using a CTG assay (mean ± SD, n = 3, P value is ns). (D-E) MM.1S cells were cultured with or without pDCs or BMSCs in the presence or absence of indicated concentrations of SB for 48 hours, and then cell viability was measured by a WST assay (mean ± SD, n = 3). (F) Total BM-derived mononuclear cells from patients with MM (n = 3) were treated with Rpn10 inhibitor SB (nontoxic concentration of 0.5 μM) or DMSO control for 2 days, and multicolor flow analysis was used to assess MM cell lysis. CD138⁺ MM cells were quantified by staining with CD138-FITC Ab. Representative fluorescence-active cell sorter scatter plot showing a decrease in the number of viable fluorescein isothiocyanate–positive MM cells after treatment with SB (left) and bar graph shows quantification of CD138⁺ MM cells in the left panel (right) are shown. The fold change was obtained after normalization with control data and presented as percentage of viable cells in the presence vs absence of SB (mean ± SD, P < .05). (G) Mice bearing human MM.1S MM tumors were treated with either vehicle control or SB (20 mg/kg, intraperitoneally) 3 times weekly for 14 days. Average and SD of tumor volume (mm³) is shown vs time when tumor was measured (mean tumor volume ±SD, 10 mice per group) (left top) and Kaplan-Meier plots show survival in mice (left bottom) are shown. SB-treated mice showed increased survival vs control vehicle-treated mice. Tumor lysates from control vehicle- and SB-treated mice were subjected to immunoblot analysis using anti–caspase-3, caspase-9, K48 polyubiquitin, and β-actin (right). (H) Representative hematoxylin and eosin (HE) and immunohistochemistry stains of caspase-3, K48 polyubiquitin and Rpn10 in tumor tissue from control- and SB-treated mice. Scale bars, 100 μM.

Figure 6.

Anti-MM activity of SB. (A) MM, melanoma, leukemia, and lymphoma cell lines were treated with dimethyl sulfoxide (DMSO) control or SB at different concentrations for 48 hours, followed by assessment for cell viability using a WST assay (P < .05 for all cell lines, n = 3). IC₅₀ was calculated and presented in the table. (B) Purified CD138⁺ cells from a patient with MM were treated with DMSO or SB for 48 hours, followed by assessment for cell viability using a CTG assay (mean ± SD of triplicate cultures). (C) Normal PBMCs from healthy donors were treated with DMSO or the indicated concentrations of SB for 48 hours, and then analyzed for cell viability using a CTG assay (mean ± SD, n = 3, P value is ns). (D-E) MM.1S cells were cultured with or without pDCs or BMSCs in the presence or absence of indicated concentrations of SB for 48 hours, and then cell viability was measured by a WST assay (mean ± SD, n = 3). (F) Total BM-derived mononuclear cells from patients with MM (n = 3) were treated with Rpn10 inhibitor SB (nontoxic concentration of 0.5 μM) or DMSO control for 2 days, and multicolor flow analysis was used to assess MM cell lysis. CD138⁺ MM cells were quantified by staining with CD138-FITC Ab. Representative fluorescence-active cell sorter scatter plot showing a decrease in the number of viable fluorescein isothiocyanate–positive MM cells after treatment with SB (left) and bar graph shows quantification of CD138⁺ MM cells in the left panel (right) are shown. The fold change was obtained after normalization with control data and presented as percentage of viable cells in the presence vs absence of SB (mean ± SD, P < .05). (G) Mice bearing human MM.1S MM tumors were treated with either vehicle control or SB (20 mg/kg, intraperitoneally) 3 times weekly for 14 days. Average and SD of tumor volume (mm³) is shown vs time when tumor was measured (mean tumor volume ±SD, 10 mice per group) (left top) and Kaplan-Meier plots show survival in mice (left bottom) are shown. SB-treated mice showed increased survival vs control vehicle-treated mice. Tumor lysates from control vehicle- and SB-treated mice were subjected to immunoblot analysis using anti–caspase-3, caspase-9, K48 polyubiquitin, and β-actin (right). (H) Representative hematoxylin and eosin (HE) and immunohistochemistry stains of caspase-3, K48 polyubiquitin and Rpn10 in tumor tissue from control- and SB-treated mice. Scale bars, 100 μM.

Figure 7.

Mechanisms of SB-induced MM cell death. (A) AMO1 and MM.1S cells were treated with DMSO control or SB at the IC₅₀ concentration for 24 hours; protein lysates were subjected to immunoblot analysis using anti-K48 polyubiquitin or anti–β-actin Abs. (B) MM.1S cells were treated with DMSO, SB, or bortezomib (BTZ) for 3 hours; protein lysates were analyzed for proteasome activities. The percentage of proteasome activity was normalized to a DMSO control (mean ± SD, n = 3). (C) Recombinant human proteins USP1, USP2, USP7, USP21, USP28, UchL1, UchL3, and UchL5 were incubated with SB for 30 minutes at 37°C and then analyzed for DUB activity (mean ± SD, n = 3). (D-H) Indicated cells were treated with DMSO or SB; protein lysates were then subjected to immunoblotting using ER stress-related antibodies against p-eIF2α, PERK, BiP, and calnexin (D); caspase-related antibodies against caspase-3, caspase-8, caspase-9, and PARP (G) (caspase-3/-9 (AMO1, MM.1S, ANBL6-BR) have same actin reprobe; PARP/caspase-9 (AMO1-CFZ.R) have same actin reprobe); and autophagy-related antibodies against LAMP2, p62, LC3, or β-actin (H); treated cells were subjected to cell cycle analysis (E) or apoptosis analysis (F). CL, caspase-like proteasome activity; CTL, chymotrypsin-like proteasome activity; TL, trypsin-like proteasome activity.