Протекционный субъект PSMD1 является ключевой терапевтической целью в множестве миеломы.

Introduction

Multiple myeloma (MM) arises from post-germinal center B cells and is the second most common hematologic malignancy. It is characterized by the excessive production of immunoglobulins and a heavy reliance on protein degradation pathways 1 such as the ubiquitin–proteasome system, which is responsible for degrading damaged, misfolded, or unnecessary proteins, thereby maintaining cellular homeostasis and function. 2 The proteasome is a multi-subunit structure consisting of the 20S core particle (CP) and the 19S regulatory particle (RP). 3 , 4 The barrel-shaped 20S CP is responsible for the degradation of proteins, whereas the 19S RP facilitates the recognition, binding, and unfolding of ubiquitinated protein substrates, ensuring their proper entry into the CP for degradation. 5 , 6 MM and other cancers exhibit high levels of proteasome activity to allow them to efficiently manage the high protein turnover, making them sensitive to proteasome inhibitors (PIs). 7 , 8

PIs that primarily target the catalytic activity of the 20S CP, such as bortezomib, carfilzomib, ixazomib, and marizomib, have been successfully used in hematologic malignancies 9 , 10 , 11 , 12 and are being explored in various solid tumors, including breast, lung, colon, thyroid, prostate, ovarian, pancreatic, and head and neck cancers. 13 However, resistance to PIs often develops over time, limiting their long-term efficacy in treating cancer. Resistance mechanisms include mutations in the 20S CP, upregulation of proteasome subunit expression, or elevation of compensatory pathways, 14 driving the need for alternative therapeutic strategies. 15 , 16 Previous studies from our group have identified critical subunits of the 19S RP, such as the ubiquitin receptors PSMD4 and ADRM1, as essential for the survival of MM cells. Targeting these subunits triggers cytotoxicity and overcomes bortezomib resistance, which has emerged as a promising approach to bypass PI resistance and enhance cancer therapy, suggesting that other components of the 19S RP may also serve as valuable therapeutic targets. 17 , 18 , 19 Here, we turn our attention to PSMD1, the largest non-ATPase subunit of the 19S RP. It is a scaffolding protein that facilitates the docking and assembly of other proteasome subunits. 20 PSMD1 is overexpressed in various cancers, including hepatocellular carcinoma, lung adenocarcinoma, and chronic myeloid leukemia. 21 , 22 , 23 , 24 Notably, the elevated protein levels of PSMD1 have been associated with poor prognosis, particularly in MM. 25 Despite these observations, the precise biological role of PSMD1 in cancer remains poorly understood.

To target PSMD1, we investigated the use of small interfering RNA (siRNA) 26 delivered via lipid nanoparticles (LNPs), an approach that has demonstrated promising preclinical efficacy in cancers, including colorectal cancer, hepatocellular carcinoma, and breast cancer. 27 Despite these advances, there remains a significant gap in applying LNP-siRNA therapies in the context of MM. In this study, we aimed to bridge this gap by developing PSMD1 siRNA encapsulated in LNPs and, for the first time, evaluating their therapeutic potential in MM both in vitro and in vivo. Our findings suggest that PSMD1 is a novel and critical therapeutic target in MM, with LNP-mediated siRNA delivery representing a promising approach for overcoming PI resistance and improving therapeutic outcomes.

Material and methods

Cell culture and reagents

Normal peripheral blood mononuclear cells (PBMCs) and MM cell lines were cultured in RPMI 1640 complete medium. Other cell lines, including CCRF-CEM, A375, HCT116, and HEK293T, were cultured in Dulbecco’s modified Eagle medium. Informed consent was obtained from all patients in accordance with the Declaration of Helsinki. CD138-positive cells and bone marrow stromal cells (BMSCs) from patients with MM were isolated and cultured as previously described. 28

Proteomic analysis

Cell pellets were submitted for isobaric tandem mass tag multiplexed quantitative analysis for the full proteome via mass spectrometry 29 at the Thermo Fisher Scientific Center for Multiplex Proteomics. Proteins with significant differential expression modulated by PSMD1 depletion were filtered using the edgeR R package. 30 The identified proteins were analyzed using the Broad Institute’s gene set enrichment analysis (GSEA) software, 31 , 32 focusing on the Human Molecular Signature Database hallmark gene sets. 33

Preparation of LNP

The ionizable cationic lipid DLIn-MC3-DMA (MC3), cholesterol, DSPC (1,2-distearoyl-sn-glycero-3-phosphocholine), and polyethylene glycol-DMG (1,2-dimyristoyl-rac-glycerol) were purchased from Avanti Polar Lipids Inc and BroadPharm. Briefly, 1 volume of the ethanol solution containing the lipids at a molar ratio of 50:38.5:10:1.5 was combined with 3 volumes of an aqueous solution containing siRNA using the vortex mixing method. After LNP formulation, the particles were washed with phosphate-buffered saline at 4000 g for 40 minutes to remove any residual ethanol.

Human MM xenograft model

All animal experiments were approved by the Institutional Animal Care and Use Committee at the Dana-Farber Cancer Institute and conducted in accordance with relevant regulatory standards.

For the inducible knockout (iKO) experiment, 6-week-old female NSG mice (Jackson Laboratory) were inoculated with 2 × 10 6 AMO1 iKO or nontargeting control (CT) cells, followed by a doxycycline (Dox) diet to induce PSMD1 knockout. Tumor growth was measured using calipers, and volume was calculated. In the humanized model, mice were IV injected with 5 × 10 6 AMO1 iKO or CT/Luc cells. Tumor burden was monitored by bioluminescence imaging using the IVIS Imaging System and followed by a Dox diet. They then received 10 × 10 6 human PBMCs via tail vein. Mice were euthanized before the experimental end point. BM samples were collected for flow cytometric analyses.

Statistical analysis

Statistical significance was determined using the Student t test (for pairwise comparisons) and 2-way analysis of variance for cell line growth data, with a threshold for significance at P < .05. Kaplan-Meier survival curves were compared by the log-rank test. All statistical analyses were performed using GraphPad software ( http://www.graphpad.com ), R language environment, or Python package SciPy. 34 Data visualization was done using the R package ggplot2 and Python package Matplotlib. 35

Results

Clinical impact and functional significance of PSMD1 in MM

We began our investigation by analyzing PSMD1 gene expression using public data sets, finding that PSMD1 gene expression was significantly higher in MM cells compared with normal plasma cells ( Figure 1 A; supplemental Figure 1 A; available in the Blood website). Furthermore, elevated PSMD1 expression was correlated with poor overall survival in patients with MM (hazard ratio, 1.82; P < .0001; Figure 1 B). Consistent with these findings, immunohistochemistry studies of BM biopsies from both patients with MM and normal donors demonstrated markedly higher PSMD1 protein levels in MM patient samples ( Figure 1 C). Similar results were observed in both MM cell lines and patient tumor cells when compared with normal plasma cells and PBMCs, as indicated by immunoblot analysis ( Figure 1 D; supplemental Figure 1 B). Figure 1. Clinical impact and functional significance of PSMD1 in MM. (A) Gene expression data were obtained using the GPL570 platform. PSMD1 expression was analyzed in normal (n = 22) and MM samples (n = 414) (data accession numbers GSE5900 and GSE4581 ). (B) Kaplan-Meier plots of PSMD1 expression vs overall survival of patients with MM. These data were generated as part of the Multiple Myeloma Research Foundation Personalized Medicine Initiatives. (C) Immunohistochemical analysis of BM biopsies from normal donors and patients with MM, demonstrating PSMD1 expression (scale bar, 20 μm). Arrows point to plasma cells. (D) Protein lysates from a panel of MM cell lines, normal plasma cells, and primary MM patient tumor cells were analyzed by immunoblotting for PSMD1 and β-actin expression. The right membrane was blotted with PSMD1 antibody after being previously blotted with PSMD4 and β-actin antibodies (as published in Du et al 19 ). The β-actin blot result was reused here. (E) MM cell lines were transfected with scrambled siRNA (CT) or PSMD1-siRNA (KD), and cell viability was assessed 48 hours/72 hours post-transfection using CellTiter-Glo (CTG) assay (mean ± standard deviation [SD]; n = 3; P < .01). Bar graph: percentage of cell viability normalized to CT. Immunoblot analysis confirms the expression of PSMD1 in cells at 72 hours post-transfection except for H929 cells at 48 hours. (F) Dox-inducible-Cas9 AMO1 cells were generated by infecting AMO1 cells with lentivirus-packaged–inducible Cas9 plasmid. These cells were then infected with either control sgRNA or PSMD1-targeted sgRNA, followed by puromycin selection to establish a stable iKO cell line. Cells were treated with 0.5 μg/mL Dox every 2 days, and cell growth was assessed using the CTG assay. Immunoblot was performed to confirm PSMD1 knockout efficiency. Data are presented as the mean ± SD of triplicates. The blots share the same PSMD1 and β-actin bands with supplemental Figure 3 A left panel. (G) Human AMO1-inducible PSMD1-shRNA knockdown (PSMD1-shKD) cells or control-shRNA cells (shRNA-CT) were subcutaneously inoculated into NSG mice (n = 5 mice per group). Mice were then placed on a diet containing irradiated 0.0625% Dox (delivering 1-6 mg of Dox per mouse per day) starting when tumors became visible (average size of control was 37.27 mm 3 and KD was 28.10 mm 3 ). Upper panel: tumor volume (mm 3 ) over time is illustrated as mean ± SD. Internal blot: immunoblot revealing PSMD1 expression in tumor lysates. Lower panel: Kaplan-Meier survival curves for the mice are displayed. (H) Human AMO1 PSMD1-iKO cells and control sgRNA-CT cells were subcutaneously injected into each flank of NSG mice. Five days after inoculation, mice were treated with a diet containing irradiated 0.0625% Dox (n = 8). Upper panel: tumor volume (mm 3 ) over time is plotted. Lower panel: immunoblot analysis of PSMD1 expression in tumor lysates from the corresponding mice illustrated in the middle panel.

Clinical impact and functional significance of PSMD1 in MM. (A) Gene expression data were obtained using the GPL570 platform. PSMD1 expression was analyzed in normal (n = 22) and MM samples (n = 414) (data accession numbers GSE5900 and GSE4581 ). (B) Kaplan-Meier plots of PSMD1 expression vs overall survival of patients with MM. These data were generated as part of the Multiple Myeloma Research Foundation Personalized Medicine Initiatives. (C) Immunohistochemical analysis of BM biopsies from normal donors and patients with MM, demonstrating PSMD1 expression (scale bar, 20 μm). Arrows point to plasma cells. (D) Protein lysates from a panel of MM cell lines, normal plasma cells, and primary MM patient tumor cells were analyzed by immunoblotting for PSMD1 and β-actin expression. The right membrane was blotted with PSMD1 antibody after being previously blotted with PSMD4 and β-actin antibodies (as published in Du et al 19 ). The β-actin blot result was reused here. (E) MM cell lines were transfected with scrambled siRNA (CT) or PSMD1-siRNA (KD), and cell viability was assessed 48 hours/72 hours post-transfection using CellTiter-Glo (CTG) assay (mean ± standard deviation [SD]; n = 3; P < .01). Bar graph: percentage of cell viability normalized to CT. Immunoblot analysis confirms the expression of PSMD1 in cells at 72 hours post-transfection except for H929 cells at 48 hours. (F) Dox-inducible-Cas9 AMO1 cells were generated by infecting AMO1 cells with lentivirus-packaged–inducible Cas9 plasmid. These cells were then infected with either control sgRNA or PSMD1-targeted sgRNA, followed by puromycin selection to establish a stable iKO cell line. Cells were treated with 0.5 μg/mL Dox every 2 days, and cell growth was assessed using the CTG assay. Immunoblot was performed to confirm PSMD1 knockout efficiency. Data are presented as the mean ± SD of triplicates. The blots share the same PSMD1 and β-actin bands with supplemental Figure 3 A left panel. (G) Human AMO1-inducible PSMD1-shRNA knockdown (PSMD1-shKD) cells or control-shRNA cells (shRNA-CT) were subcutaneously inoculated into NSG mice (n = 5 mice per group). Mice were then placed on a diet containing irradiated 0.0625% Dox (delivering 1-6 mg of Dox per mouse per day) starting when tumors became visible (average size of control was 37.27 mm 3 and KD was 28.10 mm 3 ). Upper panel: tumor volume (mm 3 ) over time is illustrated as mean ± SD. Internal blot: immunoblot revealing PSMD1 expression in tumor lysates. Lower panel: Kaplan-Meier survival curves for the mice are displayed. (H) Human AMO1 PSMD1-iKO cells and control sgRNA-CT cells were subcutaneously injected into each flank of NSG mice. Five days after inoculation, mice were treated with a diet containing irradiated 0.0625% Dox (n = 8). Upper panel: tumor volume (mm 3 ) over time is plotted. Lower panel: immunoblot analysis of PSMD1 expression in tumor lysates from the corresponding mice illustrated in the middle panel.

According to DepMap, PSMD1 is essential for the survival of MM cell lines ( supplemental Figure 1 C). To validate this, we knocked down PSMD1 using siRNA in diverse MM cell lines, including AMO1, MM.1S, H929, JJN3, ARP1, bortezomib-resistant ANBL6-BR, and carfilzomib-resistant AMO1-CR, representing distinct cytogenetic backgrounds and p53 status. Notably, PSMD1 knockdown significantly reduced cell growth across all lines ( Figure 1 E). Inducible PSMD1 knockout (AMO1-iKO and KMS11-iKO) confirmed these findings, as Dox-induced depletion halted proliferation and induced cell death ( Figure 1 F; supplemental Figure 1 D). Consistent results were observed with short hairpin RNA (shRNA), siRNA, and single-guide RNA (sgRNA) approaches ( supplemental Figure 1 F-G). Transcriptome analysis from the tumor–node–metastasis plot revealed elevated PSMD1 expression in multiple cancers compared with nontransformed counterparts ( supplemental Figure 1 H). Notably, PSMD1 knockdown also led to cell death in non-MM tumor cell lines, including HCT116, A375, and CCRF-CEM, suggesting that PSMD1 is a potential broad-spectrum therapeutic target ( supplemental Figure 1 I).

To investigate the in vivo effects of PSMD1, mice were administered Dox to induce PSMD1 knockdown when AMO1 shRNA-transfected tumors became visible. Both groups had comparable variability at baseline. PSMD1 knockdown significantly inhibited tumor growth and prolonged survival ( Figure 1 G). In a separate AMO1-iKO xenograft murine model, all 8 mice developed tumors at the site injected with scrambled sgRNA, whereas only 5 of 8 mice developed smaller tumors at the site injected with PSMD1 sgRNA ( Figure 1 H). These findings collectively underscore the potential of PSMD1 as a therapeutic target in MM.

Mechanism of cell death induced by PSMD1 depletion

Given that PSMD1 is a scaffold subunit of the 19S proteasome, we first evaluated whether PSMD1 depletion affects the proteasome-mediated protein degradation pathway. Notably, we observed a robust accumulation of K48 polyubiquitinated proteins during the inducible depletion of PSMD1 in AMO1-iKO and KMS11-iKO cells and after transient PSMD1 knockdown in multiple MM cell lines ( Figure 2 A; supplemental Figure 2 A). Depletion of PSMD1 via siRNA also resulted in a time-dependent accumulation of K48 polyubiquitinated proteins ( supplemental Figure 2 B), indicating that the cells were not degrading these proteins efficiently. Figure 2. Mechanism of cell death induced by PSMD1 depletion. AMO1-iKO cells were treated with Dox to induce PSMD1 knockout. (A) Protein lysates were analyzed by immunoblotting using antibodies against K48 polyubiquitinated proteins. AMO1-iKO blots share the same β-actin band as the lower one in supplemental Figure 3 A left panel in AMO1-iKO samples. (B) Immunoblotting was performed to detect p-eIF2α, PERK, BiP, p53, p21, and β-actin. The blots share the same middle β-actin band with supplemental Figure 5 A. (C) Cells were cultured with Dox for the specified time points, fixed in 70% ethanol, and stained with propidium iodide. The DNA content was analyzed using flow cytometry. Bar graph reveals the percentage of cells in the G0 to G1, G2 to M, or S phases of the cell cycle. (D) Apoptosis was analyzed using annexin V/propidium iodide double staining assay (mean ± SD; n = 3). (E) Protein lysates were analyzed by immunoblotting using antibodies against caspase 3, caspase 9, PARP, and β-actin. (F) AMO1-CR cells were transfected with PSMD1 siRNA. Protein lysates were analyzed by immunoblotting using antibodies as indicated. Cells were collected and subjected to cell cycle and apoptosis analysis as in panels C-D (mean ± SD; n = 3). The blots for p-eIF2α, BiP, p53, and p21 shared the same β-actin band with PARP, caspase 9, and caspase 3. (G) AMO1 PSMD1 rescue cells were generated by transfecting AMO1-iKO cells with lentivirus-packaged HA-tagged PSMD1 plasmid, followed by blasticidin selection (referred to as RES). Control cells (CT/sgCT + pEV), PSMD1-iKO cells (KO/iKO + pEV), and PSMD1 rescue cells were cultured with Dox. Cell proliferation inhibition was measured by CTG assay (mean ± SD; n = 3); cells were induced with Dox for 72 hours and then analyzed for apoptosis with annexin V/propidium iodide double staining assay (mean ± SD; n = 3); protein lysates were analyzed by immunoblotting using indicated antibodies. The blots for PSMD1 and BiP share the same β-actin band. (H) Protein lysates from Figure 1E were analyzed by immunoblotting using antibodies against PSMD1, PSMD4, ADRM1, and β-actin. MM.1S, AMO1, ANBL6-BR, JJN3, and ARP1 cell lines shared the same β-actin with supplemental Figure 2A. (I) AMO1 PSMD1-iKO, PSMD4-iKO, ADRM1-iKO, and sgCT cells were cultured with Dox for the indicated time points, followed by growth inhibition analysis using CTG assay (mean ± SD; n = 3). Data were normalized to sgCT. Immunoblot reveals PSMD1, PSMD4, ADRM1, and K48 polyubiquitylated protein expression in cells. BiP, binding immunoglobulin protein; PARP, poly(adenosine diphosphate-ribose) polymerase; p-eIF2α, phosphorylated eukaryotic initiation factor 2α; PERK, PKR-like endoplasmic reticulum kinase.

Mechanism of cell death induced by PSMD1 depletion. AMO1-iKO cells were treated with Dox to induce PSMD1 knockout. (A) Protein lysates were analyzed by immunoblotting using antibodies against K48 polyubiquitinated proteins. AMO1-iKO blots share the same β-actin band as the lower one in supplemental Figure 3 A left panel in AMO1-iKO samples. (B) Immunoblotting was performed to detect p-eIF2α, PERK, BiP, p53, p21, and β-actin. The blots share the same middle β-actin band with supplemental Figure 5 A. (C) Cells were cultured with Dox for the specified time points, fixed in 70% ethanol, and stained with propidium iodide. The DNA content was analyzed using flow cytometry. Bar graph reveals the percentage of cells in the G0 to G1, G2 to M, or S phases of the cell cycle. (D) Apoptosis was analyzed using annexin V/propidium iodide double staining assay (mean ± SD; n = 3). (E) Protein lysates were analyzed by immunoblotting using antibodies against caspase 3, caspase 9, PARP, and β-actin. (F) AMO1-CR cells were transfected with PSMD1 siRNA. Protein lysates were analyzed by immunoblotting using antibodies as indicated. Cells were collected and subjected to cell cycle and apoptosis analysis as in panels C-D (mean ± SD; n = 3). The blots for p-eIF2α, BiP, p53, and p21 shared the same β-actin band with PARP, caspase 9, and caspase 3. (G) AMO1 PSMD1 rescue cells were generated by transfecting AMO1-iKO cells with lentivirus-packaged HA-tagged PSMD1 plasmid, followed by blasticidin selection (referred to as RES). Control cells (CT/sgCT + pEV), PSMD1-iKO cells (KO/iKO + pEV), and PSMD1 rescue cells were cultured with Dox. Cell proliferation inhibition was measured by CTG assay (mean ± SD; n = 3); cells were induced with Dox for 72 hours and then analyzed for apoptosis with annexin V/propidium iodide double staining assay (mean ± SD; n = 3); protein lysates were analyzed by immunoblotting using indicated antibodies. The blots for PSMD1 and BiP share the same β-actin band. (H) Protein lysates from Figure 1E were analyzed by immunoblotting using antibodies against PSMD1, PSMD4, ADRM1, and β-actin. MM.1S, AMO1, ANBL6-BR, JJN3, and ARP1 cell lines shared the same β-actin with supplemental Figure 2A. (I) AMO1 PSMD1-iKO, PSMD4-iKO, ADRM1-iKO, and sgCT cells were cultured with Dox for the indicated time points, followed by growth inhibition analysis using CTG assay (mean ± SD; n = 3). Data were normalized to sgCT. Immunoblot reveals PSMD1, PSMD4, ADRM1, and K48 polyubiquitylated protein expression in cells. BiP, binding immunoglobulin protein; PARP, poly(adenosine diphosphate-ribose) polymerase; p-eIF2α, phosphorylated eukaryotic initiation factor 2α; PERK, PKR-like endoplasmic reticulum kinase.

Interestingly, we found that the catalytic activity of the 20S proteasome was not inhibited, and caspase-like activity was even upregulated by PSMD1 knockout ( supplemental Figure 2 C). An in-gel proteasome activity assay revealed a reduced recruitment of the 19S RP to the 20S core, resulting in decreased formation of the 26S proteasome, as indicated by the reduced 26S/20S ratio ( supplemental Figure 2 D). This suggests that the accumulation of polyubiquitinated proteins was caused by the lack of fully assembled proteasomes. Generally, the accumulation of ubiquitinated proteins triggers the unfolded protein response and apoptosis through the endoplasmic reticulum (ER) stress pathway. To explore this further, we used western blotting to investigate ER stress markers after PSMD1 depletion. As expected, we observed activation of the downstream ER stress pathway, indicated by increased levels of PKR-like ER kinase (PERK), binding immunoglobulin protein (BiP), and phosphorylated eukaryotic initiation factor 2α (p-eIF2α) ( Figure 2 B).

ER stress often halts cell proliferation and can induce apoptosis. Concordantly, we found that Dox-treated AMO1-iKO cells demonstrated significant G0/G1 cell cycle arrest ( Figure 2 C) and increased early and late apoptosis, associated with poly(adenosine diphosphate-ribose) polymerase cleavage (PARP) and activation of caspases 3 and 9 ( Figure 2 D-E). Moreover, PSMD1 depletion induced K48 polyubiquitin accumulation, ER stress activation, cell cycle arrest, and apoptosis in AMO1-CR and ANBL6-BR cells (resistant to carfilzomib and bortezomib, respectively), indicating that targeting PSMD1 could overcome PI resistance in MM ( Figure 2 F; supplemental Figure 2 E). These effects were also observed in KMS11-iKO and JJN3-KD cells, indicating that PSMD1 depletion induces cell death regardless of p53 status ( supplemental Figure 2 F).

Importantly, when PSMD1 was reintroduced into the AMO1-iKO cells, the previously observed effects—cell growth inhibition, apoptosis induction, and activation of downstream pathways—were restored, confirming that these effects were not due to off-target actions ( Figure 2 G).

Intriguingly, siRNA-mediated knockdown of PSMD1 significantly decreased the protein expression of PSMD4 and ADRM1, 2 proteasomal ubiquitin receptors previously identified as therapeutic targets in MM 19 , 36 ( Figure 2 H). siRNAs and sgRNAs targeting PSMD1 produced similar effects in a time-dependent manner ( supplemental Figure 3 A). Notably, PSMD1 depletion also resulted in reduced levels of PSMD4 and ADRM1 in different types of cancer cell lines ( supplemental Figure 3 B). Consistently, proteomic data revealed that both siRNA and shRNA depletion of PSMD1 led to significant downregulation of PSMD4 and ADRM1, whereas the third ubiquitin receptor, PSMD2, remained unaffected ( supplemental Figure 3 C). To determine whether the decreases in PSMD4 and ADRM1 levels occurred at the transcriptional level, we analyzed RNA sequencing data after PSMD1 knockdown. Remarkably, neither PSMD4 nor ADRM1 demonstrated downregulation at the mRNA level after PSMD1 depletion ( supplemental Figure 3 D), a finding further confirmed by reverse transcription quantitative polymerase chain reaction ( supplemental Figure 3 E). These results indicate that PSMD1 regulates the levels of PSMD4 and ADRM1 at the posttranscriptional level. Co-IP studies revealed, in addition to decreased PSMD1, that levels of PSMD4 and ADRM1 within the proteasome are also reduced on PSMD1 depletion ( supplemental Figure 3 F). Further investigations demonstrated that PSMD1 depletion led to greater cell death and accumulation of K48 ubiquitinated proteins compared with the depletion of PSMD4 or ADRM1 ( Figure 2 I; supplemental Figure 3 G). In addition, PSMD1 knockout downregulated PSMD4 and ADRM1; conversely, PSMD4 or ADRM1 knockout did not affect PSMD1 ( Figure 2 I; supplemental Figure 3 G). These findings suggest that PSMD1 is a more effective therapeutic target than PSMD4 and ADRM1 in MM.

Proteomic analysis of PSMD1 depletion

To investigate the overall effects of PSMD1 depletion in MM, we performed a proteomic analysis on AMO1-iKO and control cells after Dox induction. Our analysis revealed a total of 1535 differentially expressed proteins (724 upregulated and 811 downregulated) in the PSMD1 knockout cell line (false discovery rate q-value < .05, |log 2 (fold change) | > .5) ( Figure 3 A). We, then, conducted pathway enrichment analysis on both upregulated and downregulated proteins using Hallmark gene sets and Gene Ontology Biological Process gene sets. Notably, the downregulated proteins were primarily enriched in cell cycle–related pathways, including the G2-M checkpoint and mitotic spindle pathways ( Figure 3 B; supplemental Figure 4 A). This finding aligns with our previous results revealing that PSMD1 knockout inhibits cell growth in MM. Figure 3. Proteomic analysis of AMO1 PSMD1-iKO cells. AMO1 PSMD1-iKO cells were cultured with Dox for 96 hours, and protein lysates were analyzed using multiplexed proteomics with tandem mass spectrometry. (A) The volcano plot displays changes in protein expression in PSMD1-iKO cells compared with control cells. Blue and red dots represent proteins that were significantly (FDR q < .05, |log 2 FC| > .5) downregulated or upregulated, respectively. (B) GSEA of Hallmark pathways reveals significant enrichment after PSMD1 knockout. All pathways illustrated have an FDR < 20%. (C) Gene set enrichment plots reveal that IFN-α and IFN-γ response pathways are significantly enhanced. (D) AMO1 PSMD1-iKO cells were cultured with Dox for 72 hours and 96 hours. Protein lysates were analyzed by immunoblotting using antibodies against p-IRF3, IRF3, p-TBK1, TBK1, p-STING, STING, cGAS, γ-H2AX, and β-actin. (E) AMO1 PSMD1-iKO cells and JJN3 PSMD1-siRNA–transfected cells were subjected to flow cytometry after staining with anti-CALR antibody and Zombie NIR live/dead dye. Zombie NIR-negative cells were gated out, and the surface expression of CALR was quantified. (F) AMO1 PSMD1-iKO cells, after Dox washout, and JJN3 PSMD1-siRNA–transfected cells were cocultured with total mononuclear cells isolated from patients with MM for 2 days. Degranulation marker CD107a was measured on CD3 + /CD4 + and CD3 + /CD8 + T cells. The data represent paired samples, and each line connects measurements from the same biological replicate before and after treatment. (G) Schematic representation of humanized mouse model in the left panel. Figure created with biorender.com . Song Y. (2025) https://biorender.com/6gkhnpu . BM samples were collected at end point for flow cytometry analysis and assessed for the percentage of CD138 + cells and expression levels of the degranulation marker CD107a on CD3 + /CD4 + T cells. CALR, calreticulin; FC, fold change; FDR, false discovery rate.

Proteomic analysis of AMO1 PSMD1-iKO cells. AMO1 PSMD1-iKO cells were cultured with Dox for 96 hours, and protein lysates were analyzed using multiplexed proteomics with tandem mass spectrometry. (A) The volcano plot displays changes in protein expression in PSMD1-iKO cells compared with control cells. Blue and red dots represent proteins that were significantly (FDR q < .05, |log 2 FC| > .5) downregulated or upregulated, respectively. (B) GSEA of Hallmark pathways reveals significant enrichment after PSMD1 knockout. All pathways illustrated have an FDR < 20%. (C) Gene set enrichment plots reveal that IFN-α and IFN-γ response pathways are significantly enhanced. (D) AMO1 PSMD1-iKO cells were cultured with Dox for 72 hours and 96 hours. Protein lysates were analyzed by immunoblotting using antibodies against p-IRF3, IRF3, p-TBK1, TBK1, p-STING, STING, cGAS, γ-H2AX, and β-actin. (E) AMO1 PSMD1-iKO cells and JJN3 PSMD1-siRNA–transfected cells were subjected to flow cytometry after staining with anti-CALR antibody and Zombie NIR live/dead dye. Zombie NIR-negative cells were gated out, and the surface expression of CALR was quantified. (F) AMO1 PSMD1-iKO cells, after Dox washout, and JJN3 PSMD1-siRNA–transfected cells were cocultured with total mononuclear cells isolated from patients with MM for 2 days. Degranulation marker CD107a was measured on CD3 + /CD4 + and CD3 + /CD8 + T cells. The data represent paired samples, and each line connects measurements from the same biological replicate before and after treatment. (G) Schematic representation of humanized mouse model in the left panel. Figure created with biorender.com . Song Y. (2025) https://biorender.com/6gkhnpu . BM samples were collected at end point for flow cytometry analysis and assessed for the percentage of CD138 + cells and expression levels of the degranulation marker CD107a on CD3 + /CD4 + T cells. CALR, calreticulin; FC, fold change; FDR, false discovery rate.

Intriguingly, PSMD1 knockout significantly upregulated immune-related pathways, particularly IFN-α and IFN-γ pathways ( Figure 3 B-C). RNA sequencing analysis data also revealed activation of interferon signaling in PSMD1-depleted cells ( supplemental Figure 4 B). In PSMD1-depleted cells, DNA damage was induced, leading to activation of the downstream innate immune sensing cyclic GMP–AMP synthase (cGAS)–stimulator of interferon genes (STING) pathway ( Figure 3 D; supplemental Figure 4 C). Moreover, cell surface calreticulin expression was upregulated in PSMD1-depleted cells, a hallmark of immunogenic cell death ( Figure 3 E; supplemental Figure 4 C), which would enhance the recognition and elimination of MM cells by immune cells. Consistent with cytokine-driven immune activation, surface expression of major histocompatibility complex class I (HLA-ABC) and class II (HLA-DR and HLA-DQ) molecules was significantly elevated in PSMD1-depleted cells ( supplemental Figure 4 D). Major histocompatibility complex class molecules play a critical role in antigen presentation and the subsequent activation of CD4 + and CD8 + T cells. Consistently, PSMD1-depleted MM cells (AMO1-iKO or siRNA-treated JJN3) activated CD3 + /CD4 + and CD3 + /CD8 + T cells in coculture with patient mononuclear cells ( Figure 3 F). To evaluate the immune effects of PSMD1 depletion in vivo, we used a humanized mouse model. We found that PSMD1 knockdown significantly reduced tumor burden and, importantly, led to enhanced CD4 + T-cell activation in the BM microenvironment ( Figure 3 G). Collectively, these findings suggest that PSMD1 depletion enhances the immune response, further supporting the potential of PSMD1 as a promising therapeutic target in MM.

The ubiquitin-proteasome system and autophagy are the 2 main protein degradation pathways that can compensate for each other’s deficiencies. GSEA results indicated that PSMD1 knockout inhibited the mTORC1 pathway, a critical negative regulator of autophagy ( supplemental Figure 5 A). Consistently, we observed elevated expression of autophagy markers, such as p62, LC3-II, and LAMP2, after PSMD1 depletion in AMO1-iKO cells and JJN3 cells, suggesting activation of the autophagy pathway after PSMD1 depletion. Moreover, we observed that bafilomycin A1 treatment led to a further increase in LC3-II and p62 in PSMD1 KD cells that was higher than in control cells, indicating that lysosomal degradation remains functional and autophagic flux was activated ( supplemental Figure 5 A).

GSEA revealed that PSMD1 knockout activated the tumor necrosis factor-α signaling pathway via NF-κB ( supplemental Figure 5 B). Consistently, immunoblotting results demonstrated an upregulation of phosphorylated p65 and a downregulation of the NF-κB inhibitor IκBα after PSMD1 depletion ( supplemental Figure 5 B). Notably, our previous studies have revealed that proteasome inhibition can activate the NF-κB pathway in MM cells, suggesting that PSMD1 knockout exerts a comparable effect. 37

Accumulating evidence suggests that persistent ER stress and protein misfolding can trigger reactive oxygen species (ROS) cascades, and our GSEA also revealed a significant upregulation of the ROS pathway. Western blot analysis confirmed elevated levels of key ROS pathway markers, including GCLC, GCLM, ATF4, and NRF2, after PSMD1 siRNA knockdown in AMO1 and JJN3 cell lines ( supplemental Figure 5 C).

LNP-siRNA targeting PSMD1 exerts anti-MM effects in vitro and in vivo

siRNA-based therapies have demonstrated therapeutic promise for cancer treatment by silencing genes essential for tumor cell survival. Among the various siRNA delivery vehicles, LNPs have emerged as the leading choice. Given the lack of direct drugs targeting PSMD1, we used an LNP formulation approved by the Food and Drug Administration to deliver siRNA targeting PSMD1 ( Figure 4 A). Transmission electron microscopy images revealed that siRNA-loaded LNPs exhibited a spherical morphology ( Figure 4 B; supplemental Figure 6 A). The hydrodynamic diameter of these LNPs ranged from 50 to 100 nm, with a narrow polydispersity index of 0.108 ( supplemental Figure 6 B). Figure 4. Characterization of LNP-PSMD1 siRNA and cellular uptake. (A) Schematic representation of the LNP-siRNA formulation. Figure created with biorender.com . Song Y. (2025) https://biorender.com/63r7h86 . (B) LNPs were prepared by mixing an ethanolic lipid mixture with an acidic aqueous buffer containing siRNA and then imaged by cryoEM. Scale bars, 50 nm. Image taken on a Talos Arctica transmission electron microscope at 200 kV with ×39 000 original magnification using K3 camera at 1.1 Å per pixel. (C) H929 cells were transfected with 100 nM Cy5-labeled LNPs containing scrambled control siRNA (SCT) or PSMD1 siRNA and analyzed by confocal microscopy 2 hours posttransfection. Cells were stained with CD44 (green) and counterstained with NucBlue (blue). Scale bars, 10 μm. (D) Flow cytometry analysis of Cy5-siRNA distribution in cells treated with Cy5-labeled LNPs for 2 or 24 hours. (E) MM cell lines were cultured with LNP-SCT siRNA or LNP-LNP-PSMD1 siRNA (siRNA) at the indicated concentrations. Cell viability was assessed at 48 hours using CTG assay (mean ± SD; n = 3; P < .01). Immunoblot analysis reveals PSMD1 knockdown in cells after 48 hours culture with LNP. The LNP concentration for H929 cells was 25 nM, and for AMO1 and ANBL6 cells, it was 100 nM. For H929 cell line, the blots share the lower β-actin band with Figure 6 E. For AMO1 cell line, the blots share the β-actin band with Figure 6 B and lower β-actin in Figure 6 E. For ANBL6-BR cell line, the blots share the same β-actin band with Figure 6 B.

Characterization of LNP-PSMD1 siRNA and cellular uptake. (A) Schematic representation of the LNP-siRNA formulation. Figure created with biorender.com . Song Y. (2025) https://biorender.com/63r7h86 . (B) LNPs were prepared by mixing an ethanolic lipid mixture with an acidic aqueous buffer containing siRNA and then imaged by cryoEM. Scale bars, 50 nm. Image taken on a Talos Arctica transmission electron microscope at 200 kV with ×39 000 original magnification using K3 camera at 1.1 Å per pixel. (C) H929 cells were transfected with 100 nM Cy5-labeled LNPs containing scrambled control siRNA (SCT) or PSMD1 siRNA and analyzed by confocal microscopy 2 hours posttransfection. Cells were stained with CD44 (green) and counterstained with NucBlue (blue). Scale bars, 10 μm. (D) Flow cytometry analysis of Cy5-siRNA distribution in cells treated with Cy5-labeled LNPs for 2 or 24 hours. (E) MM cell lines were cultured with LNP-SCT siRNA or LNP-LNP-PSMD1 siRNA (siRNA) at the indicated concentrations. Cell viability was assessed at 48 hours using CTG assay (mean ± SD; n = 3; P < .01). Immunoblot analysis reveals PSMD1 knockdown in cells after 48 hours culture with LNP. The LNP concentration for H929 cells was 25 nM, and for AMO1 and ANBL6 cells, it was 100 nM. For H929 cell line, the blots share the lower β-actin band with Figure 6 E. For AMO1 cell line, the blots share the β-actin band with Figure 6 B and lower β-actin in Figure 6 E. For ANBL6-BR cell line, the blots share the same β-actin band with Figure 6 B.

To evaluate the cellular uptake of LNP-siRNA in MM cells, we used Cy5-labeled siRNA as cargo. Both LNP-Cy5 SCT (scrambled control) siRNA and LNP-Cy5 PSMD1 siRNA were efficiently taken up by MM cells, as demonstrated by confocal microscopy and flow cytometry analyses ( Figure 4 C-D). Treatment with LNP-PSMD1 siRNA led to a dose-dependent decrease in cell viability across various MM cell lines, including a dexamethasone-resistant cell line MM.1R, the doxorubicin-resistant cell line Dox-40, and the bortezomib-resistant cell line ANBL6-BR. Correspondingly, PSMD1 protein levels were downregulated after treatment with LNP-PSMD1 siRNA ( Figure 4 E; supplemental Figure 6 C).

The interaction between MM cells and BMSCs triggers cytokine secretion that facilitates paracrine growth of MM cells and protects them from drug-induced cell death. In coculture experiments of BMSCs and H929 cells, BMSCs accelerated H929 cell growth, but treatment with LNP-PSMD1 siRNA led to a marked reduction in the viability of H929 cells, whereas BMSCs remained unaffected ( Figure 5 A). Further analysis of total BM mononuclear cells (BM-MNCs) from patients with relapsed MM revealed that treatment with LNP-PSMD1 siRNA decreased the viability of CD138 + MM cells without affecting CD138 − cells ( Figure 5 B). In contrast, LNP-PSMD1 siRNA treatment led to a less pronounced downregulation of PSMD1 protein levels in healthy donor PBMCs, resulting in minimal effects on PBMC viability ( Figure 5 C). Collectively, these findings indicate that LNP-PSMD1 siRNA selectively targets myeloma cells, exerting significant cytotoxic effects while sparing normal cells. Figure 5. LNP-siRNA targeting PSMD1 exerts anti-MM effect. (A) Green fluorescence-labeled H929 cells were cultured with or without BMSCs in the presence of 50nM LNP-SCT siRNA or LNP-PSMD1 siRNA for 72 hours. Cell growth was monitored using an IncuCyte system. Cell confluence, highlighted in yellow, indicates BMSC growth, whereas green fluorescence marks H929 cell growth. For coculture, the ratio of H929:BMSCs was 5:1. (B) Total BM-MNCs from patients with MM (n = 6) were treated with either LNP-SCT or LNP-PSMD1 siRNA (100nM) for 72 hours. Multicolor flow cytometry was used to assess cell lysis, with CD138 + MM cells quantified by staining with anti-CD138–FITC antibody. (C) Normal PBMCs from healthy donors were treated with LNP-SCT or LNP-PSMD1 siRNA at indicated concentrations and time points. Cell viability was then assessed using the CTG assay (mean ± SD; n = 7). Immunoblot analysis reveals PSMD1 expression in cells after 48 hours treatment. (D) Human H929 cells were subcutaneously inoculated into NSG mice. Mice bearing H929 MM tumors were treated with either LNP-SCT or LNP-PSMD1 siRNA (2 mg/kg, IV) twice weekly for 2 weeks. Tumor volume (mm 3 ) over time is illustrated as the mean ± SD (n = 6 mice per group). (E) Kaplan-Meier plot reveals survival in mice. (F) NSG mice were IV injected with H929-luc cells. On detection of MM signal (∼10 days after tumor cell injection), mice were randomly assigned to receive either LNP-SCT or LNP-PSMD1 siRNA (2 mg/kg, IV, twice a week). Whole-body bioluminescence imaging (BLI) of mice is illustrated. (G) Quantification of bioluminescence measurements (mean ± SD).

LNP-siRNA targeting PSMD1 exerts anti-MM effect. (A) Green fluorescence-labeled H929 cells were cultured with or without BMSCs in the presence of 50nM LNP-SCT siRNA or LNP-PSMD1 siRNA for 72 hours. Cell growth was monitored using an IncuCyte system. Cell confluence, highlighted in yellow, indicates BMSC growth, whereas green fluorescence marks H929 cell growth. For coculture, the ratio of H929:BMSCs was 5:1. (B) Total BM-MNCs from patients with MM (n = 6) were treated with either LNP-SCT or LNP-PSMD1 siRNA (100nM) for 72 hours. Multicolor flow cytometry was used to assess cell lysis, with CD138 + MM cells quantified by staining with anti-CD138–FITC antibody. (C) Normal PBMCs from healthy donors were treated with LNP-SCT or LNP-PSMD1 siRNA at indicated concentrations and time points. Cell viability was then assessed using the CTG assay (mean ± SD; n = 7). Immunoblot analysis reveals PSMD1 expression in cells after 48 hours treatment. (D) Human H929 cells were subcutaneously inoculated into NSG mice. Mice bearing H929 MM tumors were treated with either LNP-SCT or LNP-PSMD1 siRNA (2 mg/kg, IV) twice weekly for 2 weeks. Tumor volume (mm 3 ) over time is illustrated as the mean ± SD (n = 6 mice per group). (E) Kaplan-Meier plot reveals survival in mice. (F) NSG mice were IV injected with H929-luc cells. On detection of MM signal (∼10 days after tumor cell injection), mice were randomly assigned to receive either LNP-SCT or LNP-PSMD1 siRNA (2 mg/kg, IV, twice a week). Whole-body bioluminescence imaging (BLI) of mice is illustrated. (G) Quantification of bioluminescence measurements (mean ± SD).

Building on our previous results that PSMD1 inhibition activates immune cells, we further investigated whether the BM microenvironment could also be activated by LNP-PSMD1 siRNA. Using BM-MNCs from patients with MM, we found that treatment with LNP-PSMD1 siRNA activated MM BM-MNCs and promoted the lysis of allogeneic U266 cells ( supplemental Figure 7 A).

To evaluate the in vivo efficacy of LNP-PSMD1 siRNA, we used several myeloma xenograft murine models. First, in a subcutaneous xenograft model with human plasmacytoma H929 cells, IV injection of LNP-PSMD1 siRNA (2 mg/kg) significantly inhibited MM tumor growth and prolonged survival in tumor-bearing mice ( Figure 5 D-E). Importantly, the treatment was well tolerated, with no significant weight loss observed (data not revealed). In addition, in an H929-Luc disseminated (IV) myeloma model, LNP-PSMD1 siRNA (2 mg/kg) treatment significantly reduced tumor burden, as assessed by serial whole-body bioluminescence imaging ( Figure 5 F-G). Again, no significant weight loss was observed in the treatment group (data not revealed). To monitor the real-time accumulation of LNPs at the tumor site and in other organs, we IV injected LNP-Cy5 SCT siRNA or LNP-Cy5 PSMD1 siRNA into mice bearing H929 tumors. Imaging the organs at 5, 24, and 48 hours postinjection revealed a gradual accumulation of LNP-Cy5 siRNA at the tumor site ( supplemental Figure 7 B). Overall, these results support that LNP-PSMD1 siRNA is a promising therapeutic method for treating MM.

Mechanisms of LNP-siRNA–induced MM cell death

Treating MM cell lines with LNP-PSMD1 siRNA resulted in the downregulation of PSMD4 and ADRM1, mirroring the effects observed with direct PSMD1 depletion ( Figure 6 A). In addition, LNP-PSMD1 siRNA induced the accumulation of K48 polyubiquitinated proteins in H929, AMO1, ANBL6-WT, and ANBL6-BR cells ( Figure 6 B). To determine whether LNP-PSMD1 siRNA blocks the activity of the 20S core proteasome, H929 cells were treated with LNP-PSMD1 siRNA and assessed for catalytic activity of the 20S proteasome, with bortezomib used as a positive control. We found that LNP-PSMD1 siRNA did not significantly affect 20S proteasome activity, whereas bortezomib effectively inhibited chymotrypsin-like and caspase-like activities ( Figure 6 C). Like genetic PSMD1 depletion, LNP-PSMD1 siRNA treatment led to the activation of ER stress, cell cycle arrest, induction of apoptosis, and activation of caspases in MM cells ( Figure 6 D-E). Notably, LNP-PSMD1 siRNA exhibited a strong synergistic effect when combined with pomalidomide across multiple MM cell lines, underscoring its potential for combination therapy in MM ( Figure 6 F). Furthermore, LNP-PSMD1 siRNA exhibited comparable effects in other cancer cell lines, namely HCT116, A375, and CCRF-CEM, suggesting that targeting PSMD1 using LNP-siRNA may offer a broad-spectrum therapeutic strategy across various cancer types ( Figure 6 G). Figure 6. Mechanisms of LNP-siRNA–induced MM cell death. (A-B) MM cells were treated with LNP-SCT siRNA or LNP-PSMD1 siRNA for 48 hours, and protein lysates were subjected to immunoblotting with the indicated antibodies. (C) H929 cells were treated with either 50nM LNP-SCT siRNA or LNP-PSMD1 siRNA for 24 hours, or with dimethyl sulfoxide (DMSO) (CT), or bortezomib (BTZ) at 1μM for 3 hours. Protein lysates were analyzed for 20S proteasome activities, including chymotrypsin-like (CTL), trypsin-like (TL), and caspase-like (CL) proteasome activities. (D-E) H929 and AMO1 cells were treated with LNP-SCT siRNA or LNP-PSMD1 siRNA for 48 hours. Cells were then analyzed for apoptosis using the annexin V/propidium iodide double staining assay (mean ± SD; n = 3). Protein lysates were subjected to immunoblotting using indicated antibodies. (F) H929, AMO1, ANBL6-BR, and ANBL6-WT cells were treated with LNP-PSMD1 siRNA and pomalidomide at the indicated concentration for 48 hours. Data presented as combination index (CI) values calculated by Compusyn software. (G) Non-myeloma cell lines (HCT116, A375, and CCRF-CEM) were treated with either LNP-SCT siRNA or LNP-PSMD1 siRNA at indicated concentration for 48 hours. Cell viability was assessed using CTG assay (mean ± SD; n = 3). Protein lysates from 100nM treatment were subjected to immunoblotting to confirm the expression of PSMD1.

Mechanisms of LNP-siRNA–induced MM cell death. (A-B) MM cells were treated with LNP-SCT siRNA or LNP-PSMD1 siRNA for 48 hours, and protein lysates were subjected to immunoblotting with the indicated antibodies. (C) H929 cells were treated with either 50nM LNP-SCT siRNA or LNP-PSMD1 siRNA for 24 hours, or with dimethyl sulfoxide (DMSO) (CT), or bortezomib (BTZ) at 1μM for 3 hours. Protein lysates were analyzed for 20S proteasome activities, including chymotrypsin-like (CTL), trypsin-like (TL), and caspase-like (CL) proteasome activities. (D-E) H929 and AMO1 cells were treated with LNP-SCT siRNA or LNP-PSMD1 siRNA for 48 hours. Cells were then analyzed for apoptosis using the annexin V/propidium iodide double staining assay (mean ± SD; n = 3). Protein lysates were subjected to immunoblotting using indicated antibodies. (F) H929, AMO1, ANBL6-BR, and ANBL6-WT cells were treated with LNP-PSMD1 siRNA and pomalidomide at the indicated concentration for 48 hours. Data presented as combination index (CI) values calculated by Compusyn software. (G) Non-myeloma cell lines (HCT116, A375, and CCRF-CEM) were treated with either LNP-SCT siRNA or LNP-PSMD1 siRNA at indicated concentration for 48 hours. Cell viability was assessed using CTG assay (mean ± SD; n = 3). Protein lysates from 100nM treatment were subjected to immunoblotting to confirm the expression of PSMD1.

Discussion

In this study, we identified and validated PSMD1 as a promising therapeutic target for MM. Our analysis of publicly available databases revealed that PSMD1 is significantly overexpressed in MM cells relative to normal plasma cells and that its overexpression is associated with poorer overall survival in patients with MM. This is consistent with a previous study revealing that PSMD1 is highly expressed in several other cancer types, with elevated expression correlating with worse overall survival outcomes. 38 Importantly, our study demonstrates that PSMD1 inhibition in MM cells leads to proteotoxic and oxidative stress, triggering both p53-dependent and p53-independent mechanisms of cell cycle arrest and apoptosis, possibly mediated in part through a maladaptive unfolded protein response and activation of the intrinsic apoptotic pathway. This is particularly important because mutations in the TP53 gene are common with the progression of MM and are associated with poor prognosis and resistance to therapy. 39 Although we revealed that depletion of PSMD1 induces death in PI-resistant MM cell lines, more precise mechanistic studies are needed to rigorously define the role of PSMD1 in PI resistance. Altogether, this establishes PSMD1 as an important therapeutic target and prognostic factor in a variety of contexts.

The 19S proteasome consists of a lid and a base, with the base comprising 6 ATPases along with non-ATPase subunits, including PSMD1, PSMD2, and ADRM1. 40 PSMD1, the largest subunit of the 19S RP, is centrally located within the base subcomplex, whereas ADRM1 docks at the proteasome through its interaction with PSMD1. 41 In contrast, PSMD4 is an intrinsic ubiquitin receptor for the proteasome. 42 , 43 Both PSMD1 and PSMD4 interact with the base ATPases and play crucial roles in the assembly of the 26S proteasome. 40 , 44 Our results revealed that PSMD1 depletion in myeloma reduces PSMD4 and ADRM1 protein levels without significant changes in their mRNA levels, suggesting posttranscriptional regulation. Conversely, depleting either PSMD4 or ADRM1 did not affect PSMD1 expression at the protein level, suggesting the directionality of the regulatory relationship. Given that PSMD1 operates upstream of both PSMD4 and ADRM1, its depletion has a more pronounced effect on proteasome assembly and function compared with targeting either PSMD4 or ADRM1. These findings suggest that PSMD1 plays a central scaffolding role in maintaining the stability and integrity of proteasome, providing a mechanistic explanation for its unique essentiality in MM and nominates it as a better target than PSMD4 and ADRM1.

In addition to its direct role in proteasome regulation, PSMD1 has been associated with immune-related gene signatures that predict prognosis and responses to immunotherapy across various cancers. 43 , 44 , 45 , 46 Our proteomic analysis revealed that PSMD1 inhibition upregulated immune-related pathways, suggesting a dual role in sustaining MM cell survival while modulating the immune landscape. By altering immune pathways, PSMD1 inhibition could potentially enhance the efficacy of immunotherapies, making it an especially attractive target for therapeutic strategies aimed at both direct tumor suppression and immune modulation.

Although PSMD1 has been recognized as a prognostic marker and potential target in various cancers, 23 , 38 , 47 , 48 effective therapeutic strategies for targeted therapy are still lacking. Here, we used this delivery system to trigger selective cytotoxicity against MM cells while sparing normal cells. Importantly, LNP-PSMD1 siRNA also proved effective in overcoming resistance to PIs, a common challenge in MM treatment, and may also be useful in combination therapy in MM, as found here with pomalidomide. Notably, our studies demonstrate that LNP-PSMD1 siRNA treatment also inhibited cell growth in other types of cancers, indicating its broad therapeutic potential.

Our in vivo experiments using LNP-PSMD1 siRNA revealed a significant reduction in tumor burden and prolonged survival in xenograft models, validating PSMD1 as a novel therapeutic target. Moreover, LNP-PSMD1 siRNA was well tolerated in mouse models, suggesting a favorable therapeutic index. However, siRNA-mediated silencing is temporary and has significant challenges, often requiring specialized carriers such as LNP to ensure siRNA stability, cellular uptake, and tissue-specific targeting. Currently, delivery remains largely restricted to the liver, limiting broader therapeutic applications. In contrast, small-molecule inhibitors and protein degraders provide precise pharmacological control, including dose adjustment, temporal regulation, and reversibility, which can enhance therapeutic efficacy while minimizing toxicity risks. Our use of LNP-siRNA in this study serves as a proof of concept to establish the therapeutic relevance of PSMD1 in myeloma. In ongoing studies, we are actively pursuing small-molecule strategies targeting PSMD1 for clinical application.

In summary, our study establishes a central role for PSMD1 in MM pathophysiology, with important implications for both prognosis and therapy. Future studies will optimize PSMD1-targeting strategies and explore combination therapies to overcome drug resistance in MM. Most importantly, our study further validates the therapeutic potential of targeting the ubiquitin–proteasome cascade to improve patient outcomes in myeloma and other cancers.

Conflict-of-interest disclosure: K.C.A. is a consultant to Pfizer, AstraZeneca, Janssen, and Genentech; is a board member of C4 Therapeutics, Dynamic Cell Therapies, Window Therapeutics, and Starton; and an equity owner in C4 Therapeutics, Dynamic Cell Therapies, Window Therapeutics, Starton, and Predicta. The remaining authors declare no competing financial interests.