Purpose:

Proteasome inhibitors (PI) are the backbone of various treatment regimens in multiple myeloma. We recently described the first in-patient point mutations affecting the 20S subunit PSMB5 underlying PI resistance. Notably, in vivo, the incidence of mutations in PSMB5 and other proteasome encoding genes is too low to explain the development of resistance in most of the affected patients. Thus, additional genetic and epigenetic alterations need to be explored.

Experimental Design:

We performed DNA methylation profiling by Deep Bisulfite Sequencing in PSMB5, PSMC2, PSMC5, PSMC6, PSMD1, and PSMD5, a subset of proteasome subunits that have hitherto been associated with PI resistance, recruited from our own previous research, the literature, or a meta-analysis on the frequency of somatic mutations. Methylation was followed up on gene expression level and by dual-luciferase reporter assay. The KMS11 cell line served as a model to functionally test the impact of demethylating agents.

Results:

We identified PSMD5 promoter hypermethylation and subsequent epigenetic gene silencing in 24% of PI refractory patients. Hypermethylation correlated with decreased expression and the regulatory impact of this region was functionally confirmed. In contrast, patients with newly diagnosed multiple myeloma, along with peripheral blood mononuclear cells and CD138+ plasma cells from healthy donors, generally show unmethylated profiles.

Conclusions:

Under the selective pressure of PI treatment, multiple myeloma cells acquire methylation of the PSMD5 promoter silencing the PSMD5 gene expression. PSMD5 acts as a key orchestrator of proteasome assembly and its downregulation was described to increase the cell's proteolytic capacity. PSMD5 hypermethylation, therefore, represents a novel mechanism of PI tolerance in multiple myeloma.

Translational Relevance

Proteasome inhibitors (PI) are widely used to treat multiple myeloma with drug resistance as one major clinical challenge. Thus far, the underlying mechanisms are largely unknown, and no predictive biomarkers of PI response have been established. Genomic mutations in different proteasome subunit genes induce PI resistance but are only present in a subset of patients with multiple myeloma. Here we report DNA hypermethylation of the PSMD5 promoter region and subsequent gene silencing as an acquired mechanism of PI resistance in 1 in 4 patients with PI-refractory multiple myeloma. Unlike genomic mutations, aberrant PSMD5 promoter methylation is reversible via demethylating agents. Moreover, it could serve as a biomarker to predict and monitor PI response in patients.

The proteasome inhibitors (PI) bortezomib (BTZ), carfilzomib (CFZ), and ixazomib (IXA) are widely used to treat multiple myeloma (1, 2). Their main target is the 26S proteasome, a multicatalytic enzyme that serves as a molecular machine for protein degradation. It consists of two parts: the 19S regulatory complex and the 20S catalytic/proteolytic core. The 19S subunit is formed by the 19S lid, (encoding genes are named ‘PSMDs’) that regulates the recognition of ubiquitinated substrates (3) and the 19S hexameric AAA ATPase (encoded by the ‘PSMC’ genes), that unfolds and introduces the substrates in the catalytic 20S complex (encoded by ‘PSMA’ and ‘PSMB’ genes; refs. 4, 5). Resistance to PI treatment is a common observation in advanced disease in multiple myeloma (6), but the underlying resistance mechanisms are poorly understood. We were the first to characterize patient-derived PSMB5 mutations inducing PI resistance (7). In cell lines, such PSMB5 mutations are frequently found as a mechanism of acquired resistance (8–13). They induce steric or conformational changes to the drug-binding site, impairing the PI's binding and decreasing the chymotryptic-like catalytic function of the proteasome (7). But PSMB5 mutations are rarely found in vivo (13). Therefore, additional mechanisms of PI drug resistance affecting other proteasome subunits (14–16) or other gene networks have been proposed (17–19). For example, the knockout or downregulation of genes encoding 19S subunits has been suggested to induce an imbalance of the 26S/20S proteasome ratio in favor of the free 20S active form, which counteracts and overcomes, at least partially, the effects of PI inhibition (14, 16).

DNA methylation of regulatory regions, e.g., promoter regions, along with chromatin remodeling, is a known mechanism of cells to regulate gene activity and plays a crucial role in B-cell development and maturation. Disruption in epigenetic regulation (epimutations) highly contribute to the onset and progression of various B-cell malignancies including multiple myeloma (20–22). As the epigenome is highly adaptive to environmental factors such as medication, the epimutation rate is generally estimated to be one or two orders of magnitude higher than the somatic DNA mutation rate (23). Moreover, in multiple myeloma the downregulation of miR-29b has been reported, a microRNA that targets the de novo DNA methyltransferases (DNMT) DNMT3A and DNMT3B (24). Also, mutations in these and other methylation modifying enzymes were found, e.g., TET1/2/3, IDH1/2, and DNMT1 (25). Unlike genetic alterations, epigenetic modifications are potentially reversible and have been shown to be druggable. In preclinical and clinical settings, using epigenetic modulators like the histone deacetylase (HDAC) inhibitor panobinostat, the EZH2 inhibitor EPZ-6438, or DNA methyltransferase inhibitors (DNMTi) like the cytidine analogs 5-azacytidine (Aza) or 5-aza-2´deoxycytidine [decitabine (Dec)], restored normal levels of chromatin accessibility and/or DNA methylation (26, 27). In multiple myeloma, numerous clinical trials have been performed or are ongoing investigating DNMTi combination or monotherapy (28, 29).

Cellular plasticity is the ability of cells to phenotypically change in response to environmental stimuli (30). In tumor cells, under treatment, epigenetic cell plasticity plays an important role in the development of resistance to targeted therapy (26, 27, 31). The emergence of drug-tolerant cells could follow two main principles. First, the Lamarckian induction, meaning the drug tolerance is transiently acquired by a small subpopulation of cancer cells, or second, Darwinian selection, meaning that a population of cells with intrinsically higher tolerance to anticancer agents gets enriched upon treatment (31). We recently described the latter, a new epigenetic mechanism of immunomodulatory drug (IMiD) resistance via CRBN enhancer methylation. The methylation degree increased from 39% in newly diagnosed multiple myeloma to 67% in IMiD-refractory patients (27). Here we explore DNA methylation alterations of proteasome subunit genes in the light of PI resistance.

Meta-analysis of whole-genome sequencing and whole-exome sequencing cohorts

We analyzed whole-genome sequencing (WGS) data from patients with multiple myeloma treated in our institution (N = 130; Supplementary Table. S1), CoMMpass (whole-exome sequencing, IA17 release, https://themmrf.org), and other published datasets (32–35) for a total of 1,584 multiple myeloma cases. Of these, 1,137 patients were newly diagnosed multiple myeloma (NDMM) baseline cases and 447 pretreated. We compared the incidence of single-nucleotide variants and small deletions (≤50 bp) in 46 genes encoding for proteasome subunits.

Functional validation of point mutations

Next, we functionally validated the relevance of two mutations, PSMC6 R256Q and PSMD1 E824K, that were derived from BTZ-resistant patients (32, 36). To establish proteasome subunit (PSM) mutant cell lines, WT-PSMs were subcloned into a lentiviral vector system, which was used as a template to create other mutants via site-directed mutagenesis. The lentivirus expressing WT-PSM and mutant-PSM were then used to infect RPMI8226 cells to overexpress both WT and mutant protein at equivalent levels (Supplementary Fig. S1). For Western Blot, anti-PSMC6 (A303–825A) and anti-PSMD1 (A303–852A) was used, both polyclonal anti-rabbit from Bethyl Laboratories. The infected cells were selected with puromycin for 24 hours. The cell viability was assessed in biological and technical triplicates by 3-(4,5-dimethylthiazol-2-yl)-2,5-dimethyltetrazolium bromide (MTT) compound using CellTiter 96 AQueous One Solution (Promega, Madison, WI) after a 48-hour incubation time with different doses of PIs (Fig. 1; Supplementary Fig. S1). The IC50 value was determined by GraphPad Prism software v.8 (La Jolla, CA) using the nonlinear regression model.

Figure 1.

Somatic mutations in proteasome subunit genes arise in low frequencies. A, Somatic gene mutations in NDMM (blue) and PMM patients (red). This meta-analysis was conducted for a total of 1,584 multiple myeloma cases, with 1,137 NDMM and 447 PMM samples. Altogether, in 44 of 46 proteasome subunit genes, mutations were identified with an increase after therapy in 19 genes and a decrease in four genes. Five genes were found to be mutated only in patients with PMM and in 16 genes mutations were only found at baseline. Still, PSMD1, the best candidate, was mutated in only 1.47% of PMM patients and 0.44% of NDMM patients. The remaining genes even to a lower extent, indicating that such mutations rarely occur in vivo. B, Patient-derived mutations impair PI response in vitro. Cytotoxicity assay and IC50 values of RPMI8226 multiple myeloma cell lines harboring stably expressed patient-derived mutations in proteasome subunits. In comparison with controls (*RPMI8226 transduced with the lentivirally-encoded WT PSMC6 or PSMD1), the mutants were less sensitive to BTZ. Shown are the means and standard deviations of three independent experiments (biological replicates) and three technical replicates for each experiment.

Figure 1.

Somatic mutations in proteasome subunit genes arise in low frequencies. A, Somatic gene mutations in NDMM (blue) and PMM patients (red). This meta-analysis was conducted for a total of 1,584 multiple myeloma cases, with 1,137 NDMM and 447 PMM samples. Altogether, in 44 of 46 proteasome subunit genes, mutations were identified with an increase after therapy in 19 genes and a decrease in four genes. Five genes were found to be mutated only in patients with PMM and in 16 genes mutations were only found at baseline. Still, PSMD1, the best candidate, was mutated in only 1.47% of PMM patients and 0.44% of NDMM patients. The remaining genes even to a lower extent, indicating that such mutations rarely occur in vivo. B, Patient-derived mutations impair PI response in vitro. Cytotoxicity assay and IC50 values of RPMI8226 multiple myeloma cell lines harboring stably expressed patient-derived mutations in proteasome subunits. In comparison with controls (*RPMI8226 transduced with the lentivirally-encoded WT PSMC6 or PSMD1), the mutants were less sensitive to BTZ. Shown are the means and standard deviations of three independent experiments (biological replicates) and three technical replicates for each experiment.

Close modal

DNA methylation screening

We collected CD138+ primary tumor samples from 145 patients with multiple myeloma from the University Hospital of Würzburg (Würzburg, Germany) and the Hospital Universitario 12 de Octubre (Madrid, Spain). Of these patients, 83 were NDMM and 62 had relapsed multiple myeloma (rMM; Table 1). The rMM cohort consists of 6 patients that were never exposed to any PI, 28 patients that were PI pretreated but not PI refractory, 25 patients defined as PI relapsed/refractory (rrMM) according to the International Myeloma Working Group criteria, and 3 patients who had an unknown PI response. Of the patients with rrMM, 18 were resistant to BTZ, 13 to CFZ, and 2 to IXA. Seven patients were multirefractory to more than two PIs and 18 patients refractory to only one PI (Table 1). As controls, we sequenced 45 peripheral blood mononuclear cells (PBMC) and 52 CD138+ plasma cells from femoral heads obtained from healthy donors that underwent hip surgery. The cutoff to define PSMD5 hypermethylation was calculated on the basis of the empirical three-sigma rule applied to the PI naïve multiple myeloma cohort (NDMM). Informed written consent of all patients was obtained according to the Declaration of Helsinki and the studies were approved by an institutional review board.

Table 1.

Patient characteristics for PSMD5 methylation analysis with progressed multiple myeloma at sampling.

Parameter
Patients, n  62 
Gender, n (%) Male 42 (68) 
 Female 20 (32) 
Age at diagnosis of multiple myeloma, years (range)  57 (35–76) 
Time between initial diagnosis and sampling, months (range)  62 (3–215) 
Subtype, n (%) IgG 36 (58) 
 Non-IgG 13 (21) 
 LC 13 (21) 
ISS stage, n (%) 23 (37) 
 II 10 (16) 
 III 7 (11) 
 N/A 22 (36) 
Cytogenetics, n (%) High riska 20 (32) 
 Standard risk 36 (58) 
 N/A 6 (10) 
Prior lines of therapies at the time point of sampling, n (%) 1–3 36 (58) 
 4–6 22 (35) 
 >6 4 (7) 
Pretreatment at the time point of sampling, n (%)   
 PI, n (%) BTZ 36 (58) 
 CFZ 17 (27) 
 IXA 4 (6) 
 None 4 (6) 
 IMiD, n (%) Lenalidomide 50 (81) 
 Pomalidomide 19 (31) 
 Thalidomide 10 (16) 
 None 12 (19) 
 Monoclonal antibody, n (%) Elotuzumab 5 (8) 
 Daratumumab 10 (16) 
 None 49 (79) 
 Autologous stem cell transplant, n (%) Yes 8 (16) 
 No 54 (87) 
PI-refractory patients, n (%)  25 (40) 
 BTZ-refractory 18 (29) 
 CFZ-refractory 13 (21) 
 IXA-refractory 2 (3) 
 Multi-refractory to ≥2 PIs 7 (11) 
Parameter
Patients, n  62 
Gender, n (%) Male 42 (68) 
 Female 20 (32) 
Age at diagnosis of multiple myeloma, years (range)  57 (35–76) 
Time between initial diagnosis and sampling, months (range)  62 (3–215) 
Subtype, n (%) IgG 36 (58) 
 Non-IgG 13 (21) 
 LC 13 (21) 
ISS stage, n (%) 23 (37) 
 II 10 (16) 
 III 7 (11) 
 N/A 22 (36) 
Cytogenetics, n (%) High riska 20 (32) 
 Standard risk 36 (58) 
 N/A 6 (10) 
Prior lines of therapies at the time point of sampling, n (%) 1–3 36 (58) 
 4–6 22 (35) 
 >6 4 (7) 
Pretreatment at the time point of sampling, n (%)   
 PI, n (%) BTZ 36 (58) 
 CFZ 17 (27) 
 IXA 4 (6) 
 None 4 (6) 
 IMiD, n (%) Lenalidomide 50 (81) 
 Pomalidomide 19 (31) 
 Thalidomide 10 (16) 
 None 12 (19) 
 Monoclonal antibody, n (%) Elotuzumab 5 (8) 
 Daratumumab 10 (16) 
 None 49 (79) 
 Autologous stem cell transplant, n (%) Yes 8 (16) 
 No 54 (87) 
PI-refractory patients, n (%)  25 (40) 
 BTZ-refractory 18 (29) 
 CFZ-refractory 13 (21) 
 IXA-refractory 2 (3) 
 Multi-refractory to ≥2 PIs 7 (11) 

Abbreviation: N/A, not applicable.

aDefined as the presence of at least one of the following: t(4;14), t(14;16), t(14;20), del17p.

Deep bisulfite sequencing

The AllPrep DNA/RNA Micro Kit (Qiagen, Hilden, Germany) was used for DNA isolation. Per sample, 100 to 200 ng of genomic DNA was bisulfite converted with the EpiTect Fast 96 Bisulfite Conversion Kit (Qiagen). Deep Bisulfite Sequencing (DBS) primers were designed using the PyroMark Assay Design 2.0 software (Qiagen; Supplementary Table S2). DBS libraries were paired-end sequenced with the MiSeq platform (Illumina, San Diego, CA) and v2 cartridges (2 × 250 cycles) according to the manufacturer's instructions for low diversity libraries. Enrichment PCRs were prepared as described elsewhere (37). FASTQ files were aligned with the Amplikyzer2 pipeline (https://pypi.org/project/amplikyzer2/). Commercially available DNA standards (Qiagen) were used to establish the assays (Supplementary Fig. S2). A locus was considered affected when the average methylation was higher than 15%.

qPCR expression analysis

AllPrep DNA/RNA kit (Qiagen) was used to isolate RNA from cell pellets. Real-time qPCR of PSMD5 was performed on 44 multiple myeloma samples. Reverse transcription was performed with the SuperScript IV VILO Mastermix (Thermo Fisher Scientific). Analysis was done with the TaqMan reagent kits and molecular probes (Thermo Fisher Scientific): Hs01092588_m1 for PSMD5 and Hs00939627_m1 for the control GUSB on the StepOnePlus. All samples were analyzed in triplicates.

Luciferase reporter assay

A dual-luciferase reporter assay was performed to functionally validate the impact of PSMD5 promoter methylation on gene regulation. A BamHI recognition site was implemented in the forward primer and a HindIII recognition site in the reverse primer, respectively (Supplementary Table. S3). The pCpGL vector has a backbone free of CpG dinucleotides. The CpG containing PSMD5 insert was ligated into the multiple cloning site and the gene expression was measured by comparing the methylated with the unmethylated status (38). The insert was located upstream of the luciferase gene. For transformation one-shot PIR1 competent cells (Thermo Fisher Scientific) were used. The selection was conducted via Zeocin resistance and confirmed by Sanger sequencing. Colony PCR was done with the forward primer CCTGTAAAGTCTTTATCACACTACC and the reverse primer CCTCACAGACATCTCAAAGTATTC. After DNA purification, in vitro methylation was performed using SssI, HhaI, and HpaII methyltransferases (New England Biolabs, Ipswich, MA). HpaII (methylation-sensitive digestion) and MspI (methylation insensitive; New England Biolabs) were used to test the methylation status. By electroporation, either the methylated or the unmethylated pCgGL-PSMD5 vector was transfected together with a Renilla control vector (20:1 ratio) into L363 cells. To correct for the transfection efficiency bias, a CD4 cotransfection was performed, followed by magnetic bead retention and an OptiPrep purification. Measurement of the luciferase activity (20.000 cells) was done with a Spark microplate reader (Tecan Life Sciences, Männedorf, Switzerland). The pCpGL reporter (firefly fluorescence) was normalized against the Renilla control vector (Supplementary Table. S4). Empty pCpGL vector and non-transfected cells served as negative controls, and pCpGL with a cytomegalovirus (CMV) promoter was inserted as a positive control.

DNMTi treatment

The human KMS-11 cell line was kindly provided by Dr. Leif Bergsagel's laboratory and tested negative for Mycoplasma. It was cultured sterile in RPMI1640 medium supplemented with 10% FCS, 1 mmol/L sodium pyruvate, 2 mmol/L glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37°C with 5% CO2. KMS-11 cells were pretreated for 72 hours in 100 nmol/L Dec (Selleck Chemicals, Houston, TX) or for 48 hours in 500 nmol/L Aza (Sigma Aldrich, St. Louis, MO). Drug solutions were always freshly prepared from concentrated stocks. The solvent for the stock solution was DMSO.

PSMD5 promoter pyrosequencing

The methylation status of the PSMD5 promoter region after DNMTi pretreatment was monitored via pyrosequencing using the following primers: forward primer GAGATGGTTATGGGGTTGTTAAGTT, reverse primer Biotin-CCCAAAATCAAAATTCCCCTCCAA, pyrosequencing primer 1 AGGGTAAAGTTTTGTTAAAT, and pyrosequencing primer 2 GAGGTTTTAAAATTAGGTTTTAG. Altogether, 10 CpGs within the DBS target region were covered by pyrosequencing; sequencing primer 1 covered CpGs 15–19, and sequencing primer 2 CpGs 23–27 of the DBS assay. Pyrosequencing was performed with the PyroMark Gold Q96 CDT Reagent Kit (Qiagen) and 10 pmol of the sequencing primer. It was analyzed on the PyroMark Q96 MD system (Qiagen) with the Pyro Q-CpG software (Qiagen). All experiments were performed in triplicates and repeated at least once. Statistical analyses were performed with IBM SPSS Statistics 26 using t test for normally distributed measurements and Mann–Whitney U for nonparametric data.

alamarBlue measurement

Cell survival was measured with alamarBlue viability measurements. A total of 20,000 cells per well were seeded into 96-well plates and incubated for 72 hours with BTZ (S1013; Selleck Chemicals, Houston, TX). Reazurin (R7017, Sigma Aldrich) dissolved in PBS was used as reagent in the alamarBlue assay. Drug solutions were always freshly prepared from concentrated stocks. DMSO was the solvent for the stock solution. All experiments were done in triplicates and repeated at least once. Data was statistically evaluated by nonlinear regression analysis (sigmoidal shape) using GraphPad Sofware v.8 (La Jolla).

Data availability statement

All requests for raw and analyzed data and materials will be promptly reviewed by the University Hospital Würzburg and the MLL Munich Leukemia Laboratory to determine whether the request is subject to any confidentiality or data protection obligations. Any data and materials that can be shared will be released via a Material Transfer Agreement. Adherence to the Datenschutz-Grundverordnung (https://dsgvo-gesetz.de/ and https://data.consilium.europa.eu/doc/document/ST-5419-2016-REV-1/en/pdf) is mandatory for sharing WGS data. Thus, a complete whole genome raw data set for a single patient cannot be shared according to European law. Processed data from which the identification of a patient is not possible can be made available. For requests, contact Martin Kortüm via e-mail at [email protected].

Frequency of proteasome mutations in patients with multiple myeloma is low

In 1,584 patients, we identified mutations in 44 of 46 investigated proteasome subunit genes (Fig. 1A) resulting in a mutation incidence of 8.1% (92/1,137) in newly diagnosed patients and 11.0% (49/447) in pretreated patients (PMM) in at least one proteasome subunit gene. The increase in the mutation frequency from NDMM to PMM did not reach statistical significance. PSMD1, a ubiquitin receptor gene (3), showed the highest mutation incidence in PMM and the highest increase in the mutation rate from 0.44% at diagnosis to 1.47% after therapy. One PSMD1 missense mutation, [g.chr2:232028430:G>A p.E824K rs17352860 (GRCh37; ref. 32)], located in the double alpha-helix domain that interacts with PSMD2, was followed up in vitro. PSMB5 missense mutations were found in three NDMM (0.3%) and 2 patients with rMM (0.5%).

Patient-derived 19S point mutations induce PI resistance in vitro

We generated RPMI8226 multiple myeloma cell lines expressing mut-PSMs (PSMC6 R256Q g.chr14:53185704G>A p.R256Q and PSMD1 E824K g.chr2:232028430:G>A p.E824K rs17352860). Both mutations were patient-derived and showed increased resistance to all tested PIs (BTZ, CRF, IXA) compared with the WT-PSM cell lines (Fig. 1B; Supplementary Fig. S1).

Screening for epigenetic dysregulation in proteasome subunit genes

Promoter methylation of PSMB5, PSMC2, PSMC5, PSMC6, PSMD1, and PSMD5, genes partly recruited from literature research (7, 16, 39) and the meta-analysis (PSMD1 with the highest mutation rate in PMM or PSMC5 at NDMM; Fig. 1A), was determined by NGS targeted DBS. The median read coverage per sample and gene was 10,031X ± 6,624 (SD). For PSMB5, PSMC5, PSMC6, and PSMD1 the investigated promoter regions (Supplementary Table. S2) were unmethylated in 30 PBMCs with 0.17% ± 0.14%, 0.20% ± 0.14%, 0.14% ± 0.21%, and 0.23% ± 0.24%, as well as in 35 multiple myeloma samples with 0.17% ± 0.17%, 0.20% ± 0.17%, 0.11% ± 0.16%, and 0.08% ± 0.10%, respectively, independent of disease stage and PI response (Supplementary Fig. S3). For PSMC2, hypermethylation occurred in both multiple myeloma cells (12.00% ± 12.37%) and PBMCs (17.82% ± 7.66%). Even though with 15.65% ± 16.15% the mean methylation was higher in PI-resistant rrMM (N = 13) compared with PI-sensitive multiple myeloma (N = 18) with 9.15% ± 8.68%, no statistical significance was reached (P = 0.158, t test). PSMC2 and PSMD5 promoter methylation did not correlate (Spearman's rho = −0.099; P = 0.596).

PSMD5 promoter methylation is commonly acquired in PI-refractory disease

Within our investigated cohort, PSMD5 promoter methylation was lowest in PBMCs (N = 45) and CD138+ plasma cells from healthy donors (N = 52) with mean methylation of 2.40% ± 2.44% and 1.90% ± 2.04% (Fig. 2A). Similarly, PSMD5 promoter methylation profiles were low in NDMM (2.87% ± 3.81%, N = 83), but were found increased after treatment (7.30% ± 11.00%, N = 62). This increase was statistically significant (P = 0.001, t test) for the mean methylation as well as for all individual CpGs except CpG 16 (P = 0.199). Patients with PI rrMM faced the highest grade of methylation (9.20% ± 11.22%, N = 25). Interestingly, a subset of N = 28 PI-pretreated, but non–PI-refractory patients exhibited mean methylation of not more than 2.03 ± 2.60% (Fig. 2B). Only 0.5% (4/83) of the NDMM cohort showed a hypermethylated state, whereas 24% (6/25) of patients with PI rrMM were affected by PSMD5 hypermethylation. Within the PBMCs (N = 45) and healthy plasma cells (N = 52), no individual was hypermethylated.

Figure 2.

PSMD5 hypermethylation in patients with PI-refractory multiple myeloma. A, Methylation landscape of the analyzed PSMD5 promoter region in PI rrMM, PI-pretreated and -sensitive multiple myeloma, NDMM, PBMCs, and CD138+ cells from healthy donors. PSMD5 promoter methylation is absent in healthy CD138+ cells and PBMCs. Within the PI-refractory cohort, highly methylated individuals were detected. B, The single CpG methylation analysis shows a significant methylation increase between NDMM and rrMM (P = 0.010, t test, error bars: 95% CI) of the mean methylation. CpGs 1–12 are not shown here as they were rather unmethylated in all groups. CI, confidence interval.

Figure 2.

PSMD5 hypermethylation in patients with PI-refractory multiple myeloma. A, Methylation landscape of the analyzed PSMD5 promoter region in PI rrMM, PI-pretreated and -sensitive multiple myeloma, NDMM, PBMCs, and CD138+ cells from healthy donors. PSMD5 promoter methylation is absent in healthy CD138+ cells and PBMCs. Within the PI-refractory cohort, highly methylated individuals were detected. B, The single CpG methylation analysis shows a significant methylation increase between NDMM and rrMM (P = 0.010, t test, error bars: 95% CI) of the mean methylation. CpGs 1–12 are not shown here as they were rather unmethylated in all groups. CI, confidence interval.

Close modal

PSMD5 promoter hypermethylation silences gene expression

Next, we investigated whether promotor methylation correlated with PSMD5 expression in our cohort (Fig. 3A). Patients with high methylation rates had significantly lower PSMD5 expression and vice versa (Spearman's rho = −0.353, P = 0.02) confirming previous work by others (16). The dual-luciferase reporter assay showed an 8-fold increased luciferase activity when the unmethylated insert was compared with the methylated insert (Fig. 3B) also confirming methylation mediated PSMD5 gene silencing (P ≤ 0.001). Therefore, our results strongly suggest a potentially clinically meaningful role of PSMD5 promoter methylation for acquired PI resistance.

Figure 3.

Methylation transmitted gene silencing. A,PSMD5 promoter methylation and gene expression showed a significant negative correlation (N = 44 patients with multiple myeloma, Spearman's rho = −0.353; P = 0.020). For patients with high methylation, gene expression was suppressed. B, Luciferase reporter assay for functional confirmation of the regulatory impact of methylation in the investigated PSMD5 promoter region. Firefly vector with the PSMD5 insert was co-transfected with Renilla control vector into L363 cells. Vector without the insert and non-transfected cells served as negative controls. Luciferase activity of the unmethylated PSMD5 construct was 8 times increased compared with the methylated counterpart, confirming gene silencing through promoter methylation. C, Effect of DNMTi treatment on PSMD5 promoter methylation and gene expression. KMS-11 cells were exposed to 500 nmol/L Aza+ for 48 hours or 100 nmol/L Dec+ for 72 hours, respectively. PSMD5 promoter methylation, measured in at least triplicates by two independent pyrosequencing primers. D, PSMD5 expression, 5 days post incubation with DNMTi, was analyzed by qPCR TaqMan assays via the ΔΔCT method, and with GUSB as housekeeping control.

Figure 3.

Methylation transmitted gene silencing. A,PSMD5 promoter methylation and gene expression showed a significant negative correlation (N = 44 patients with multiple myeloma, Spearman's rho = −0.353; P = 0.020). For patients with high methylation, gene expression was suppressed. B, Luciferase reporter assay for functional confirmation of the regulatory impact of methylation in the investigated PSMD5 promoter region. Firefly vector with the PSMD5 insert was co-transfected with Renilla control vector into L363 cells. Vector without the insert and non-transfected cells served as negative controls. Luciferase activity of the unmethylated PSMD5 construct was 8 times increased compared with the methylated counterpart, confirming gene silencing through promoter methylation. C, Effect of DNMTi treatment on PSMD5 promoter methylation and gene expression. KMS-11 cells were exposed to 500 nmol/L Aza+ for 48 hours or 100 nmol/L Dec+ for 72 hours, respectively. PSMD5 promoter methylation, measured in at least triplicates by two independent pyrosequencing primers. D, PSMD5 expression, 5 days post incubation with DNMTi, was analyzed by qPCR TaqMan assays via the ΔΔCT method, and with GUSB as housekeeping control.

Close modal

Effect of DNMTi treatment on PSMD5 promoter methylation, expression and PI response

Furthermore, we explored the PI-desensitizing effect of the DNMTis Aza and Dec. The KMS11 cell line exhibits an intrinstic PSMD5 methylation degree of 20.38% ± 1.06%. After exposure to DNMTi, 48 hours of 500 nmol/L Aza or 72 hours of 100 nmol/L Dec, the methylation was reduced to 16.86% ± 1.46% for Aza (P = 0.017) and 16.01% ± 1.68% for Dec (P = 0.012; Fig. 3C). PSMD5 expression assessed by qPCR five days post-incubation with DNMTi, was increased from 1.01 ± 0.14 to 1.50 ± 0.21 for Aza (P = 0.001) and 1.62 ± 0.36 for Dec (P ≤ 0.001; Fig. 3D). In an alamarBlue assay, after 72 hours of BTZ incubation, no difference in the PI response was observed for the DNMTi-pretreated cells compared with the DNMTi-naive counterpart (Supplementary Fig. S4). To repeat the same experiment in a second model, we have screened the following other multiple myeloma/plasma cell leukemia cell lines: AMO-1, ARH-77, INA-6, JJN-2, KMS-11, KMS12, L-363, MM.1S, MOLP-8, OPM-2, and RPMI8226. Regrettably, besides KMS11, none showed intrinsic PSMD5 promoter hypermethylation. Furthermore, we have knocked out PSMD5 with CRISPR Cas9 or knocked it down via shRNA inhibition in the multiple myeloma cell lines L363 and AMO1. However, these modified cells were not viable and no functional characterization could be conducted. The group of Tsvetkov and colleagues made a similar observation and suggested that the general long-term downregulation of 19S subunits negatively impacts the cell viability, but is beneficial under PI treatment (14).

Enhanced cancer-associated proteasome assembly is a characteristic of tumors with high protein turnover. To maintain protein homeostasis, cancer cells significantly increase their protein degradation potential (40, 41). PIs block the ubiquitin-proteasome system resulting in inhibited degradation and an accumulation of proteins within the cancer cells. The development of PI drug resistance is a major problem in the treatment of multiple myeloma and thus far, no predictive biomarker of PI response has yet been established. Here we demonstrate acquired PSMD5 promoter methylation as a potential mechanism of PI resistance present in nearly one-quarter of PI-refractory patients with multiple myeloma. The sequencing depth of our DBS PSMD5 assay on bisulfite converted DNA was 10−3 (Supplementary Fig. S2), which is in the range of circulating tumor DNA in cell-free DNA of liquid biopsies. Its applicability for minimally invasive disease monitoring and relapse anticipation needs to be further investigated. We confirm by luciferase reporter assay in cell line, as well as in patients’ primary samples, that the methylation degree of the PSMD5 promoter regulates the PSMD5 gene expression level. Furthermore, in the KMS11 cell line, we provide evidence that PSMD5 promoter methylation can be reduced by the use of demethylating agents like Aza or Dec increasing PSMD5 gene expression.

PSMD5 (alias S5b) was recently characterized as an important mediator of proteasome activity in colorectal tumors in mice and human cell lines (40, 41). It acts as a chaperone involved in the 26S proteasome assembly (42–44). The regulatory impact of PSMD5 on proteasome activity was confirmed by incubating cell extracts with excess purified PSMD5 protein. The underlying mechanism is not yet fully unraveled, but the time or dosage-dependent binding of PSMD5 with PSMC2 seems to be essential (44). A recent study determined PSMD5 as the most frequently transcriptionally suppressed 19S proteasome subunit in a dataset of multiple human tumors and diverse human cancer cell lines (16). Moreover, in vitro the suppression of 19S subunits (inter alia PSMD5) was found to be associated with BTZ resistance (14). As PSMD5 regulates the level of proteasome assembly, its silencing stimulates the 26S proteasome assembly increasing the proteolytic capacity of the tumor (45). Thus, in PI-refractory disease, PSMD5 hypermethylation might represent a way to increase PI tolerance by counterbalancing elevated proteotoxic stress by the inhibitory effect of the PI treatment (Fig. 4).

Figure 4.

PSMD5 promoter hypermethylation is a mechanism of multiple myeloma cells to increase PI tolerance. Under the selective pressure of PI treatment, multiple myeloma cells gain PSMD5 promoter methylation that silences the PSMD5 gene expression. PSMD5 (alias S5b) acts as a negative regulator of proteasome activity and as an important inhibitor of 26S proteasome assembly. Its downregulation represents an adaptive way for a tumor cell to enhance PI tolerance by increasing the proteolytic capacity and maintaining protein homeostasis even during PI exposure.

Figure 4.

PSMD5 promoter hypermethylation is a mechanism of multiple myeloma cells to increase PI tolerance. Under the selective pressure of PI treatment, multiple myeloma cells gain PSMD5 promoter methylation that silences the PSMD5 gene expression. PSMD5 (alias S5b) acts as a negative regulator of proteasome activity and as an important inhibitor of 26S proteasome assembly. Its downregulation represents an adaptive way for a tumor cell to enhance PI tolerance by increasing the proteolytic capacity and maintaining protein homeostasis even during PI exposure.

Close modal

The APEX trial compared BTZ vs. high-dose dexamethasone and patients with reduced expression levels of at least one 19S subunit, showed inferior disease (progression-free survival) under BTZ treatment (16) and in these patients, BTZ was not superior to dexamethasone treatment (16). However, other clinical data from PI monotherapy that could confirm our suggested relationship between PSMD5 promoter hypermethylation and PI resistance is not available, due to the common use of PIs in combination with other drugs.

In our experimental setup, we faced difficulties to demonstrate the mechanistic link between PSMD5 promoter hypermethylation and PI resistance in vitro. In KMS11, the use of DNMTis decreased methylation and increased gene expression, but we observed no increase in BTZ sensitivity in the treated versus the untreated cells (Supplementary Fig. S4). It is possible that the moderate methylation reduction of only 3% to 4% has not been sufficient for a resensitization effect, nor can the presence of a different, independent PI resistance mechanism be excluded. KMS11 was the only multiple myeloma cell line with intrinsic PSMD5 hypermethylation in our screening, thus, we were unable to repeat this experiment in a second cell line. We were also not able to knockout or knockdown PSMD5 using CRISPR Cas9 and shRNA in AMO1 and L363, as these modified cells were not viable. This is in line with reports from another group that also failed to create suitable cell line models using a panel of shRNA-expressing lentiviruses targeting seventeen 19S subunits (PSMD5 included; ref. 14). They reported that a long-term reduction of 19S subunits decreased the growth and cell viability, making the generation of mutants for the majority of the 19S subunits not possible. The only models that propagated in this study were PSMC5 and PSMD2 with modest reductions. However, they showed increased resistance to PIs (14).

Synergistic effects of decitabine with BTZ were recently described in a xenograft mouse model transplanted with H929 cells, which supports our hypothesis, but the effects observed may also be explained by off-target effects (46). Nevertheless, in this model, the combination of the DNMTi with BTZ efficiently inhibited tumor growth and significantly increased the percentage of TUNEL-positive cells compared with the control groups (single drug exposure with only Dec or BTZ; ref. 46).

The available quantity of material, received from primary patient samples, did not allow us to simultaneously evaluate promoter methylation, gene expression, and protein levels in the same sample and compare PI-sensitive with -resistant patients. Thus, we performed targeted DBS, that needs not more than 100 to 200 ng of DNA per patient, and qPCR for a subset of patients with enough material available. We restricted our gene panel on promoter regions of proteasome subunits that have previously been associated with PI resistance. Those genes were selected from our own research [PSMB5 (7), PSMC2 (34, 39, 47), PSMC5 (39), PSMC6 (36, 39)], or from the literature (PSMD5; ref. 16). Two genes were derived from our meta-analysis as they showed the highest mutation rates: PSMD1 in PMM and PSMC5 in NDMM. Notably, in a recent WGBS study additional DNA methylation changes were identified in regions undergoing reprogramming in 24 paired baseline/relapse samples. These loci, proximal to the genes PRKCE, MGMT, FHIT, WWOX, and HDAC9 were prognostic for outcome and are potentially also indicative of therapeutic resistance (48). However, the relapsed/refractory cohort in this study was not further clincially characterized, in particular, their PI status was unknown. Currently, publicly available genome-wide DNA methylomic datasets on patients with multiple myeloma rather aim to find new markers of pathogenesis or high-risk disease (20), but are not specifically focused on finding new drug resistance mechanisms, which is mirrored in the cohort selection and missing clinical annotation. Thus, to unravel new resistance mechanisms, that may include others than proteasome subunits genes, more epigenome-wide studies on clinically well-annotated PI-refractory patients are needed.

In summary, our study found genetic and epigenetic dysregulation of proteasome subunit genes in a significant proportion of the investigated patients with multiple myeloma with a potentially clinically meaningful role in PI resistance. Our main finding, PSMD5 hypermethylation in patients with acquired PI resistance, needs to be confirmed in prospective clinical trials and followed up longitudinally in paired diagnosis-relapse samples before and after PI-containing regimen to assess its potential to be used as a biomarker of response and PI resistance.

L. Haertle reports personal fees from DFG Research Fellowship HA 9529/1-1, as well as nonfinancial support from Michael Rehli: pCpGL and pCpGL-CMV/EF1 vectors during the conduct of the study. R.A. Fernandez reports honoraria for consulting, lectures, and speaker bureau from BMS, Janssen, GSK, Sanofi, Amgen, Pfizer and Takeda. N. Bolli reports personal fees from Takeda, Amgen, Janssen, GSK, Jazz, and BMS/Celgene outside the submitted work. R. Hajek reports grants, personal fees, and nonfinancial support from Celgene, Amgen, and Takeda; grants and nonfinancial support from Novartis; personal fees and nonfinancial support from GSK and Sanofi; and personal fees from BMS outside the submitted work. M.S. Raab reports grants from Dietmar-Hopp Foundation during the conduct of the study; M.S. Raab also reports grants and personal fees from Amgen and Sanofi, as well as personal fees and nonfinancial support from BMS and Janssen outside the submitted work. C. Haferlach reports other support from MLL Munich Leukemia Laboratory outside the submitted work. J. Martinez-Lopez reports personal fees and nonfinancial support from Janssen, as well as personal fees from Sanofi during the conduct of the study. J. Martinez-Lopez also reports grants and personal fees from BMS, as well as personal fees from Roche and Gilead outside the submitted work. H. Einsele reports grants and other support from Janssen, BMS/Celgene, Amgen, GSK, and Sanofi, as well as other support from Takeda and Novartis during the conduct of the study. L. Rasche reports personal fees from BMS, Janssen, Pfizer, Amgen, GSK, and Sanofi outside the submitted work. K.M. Kortüm reports other support from Stifterverband für die Deutsche Wissenschaft during the conduct of the study. K.M. Kortüm also reports personal fees from Celgene, BMS, AbbVie, GSK, and Takeda; grants and personal fees from Janssen; and grants from SkylineDx and German Cancer Aid – MSNZ outside the submitted work. No disclosures were reported by the other authors.

L. Haertle: Conceptualization, data curation, formal analysis, validation, investigation, visualization, writing–original draft, project administration, writing–review and editing. S. Barrio: Conceptualization, data curation, investigation, visualization, writing–original draft, writing–review and editing. U. Munawar: Data curation, investigation, visualization, methodology, writing–review and editing. S. Han: Data curation, investigation, visualization, methodology, writing–review and editing. X. Zhou: Resources, investigation, visualization, writing–review and editing. M. Simicek: Data curation, supervision, validation, visualization, writing–review and editing. C. Vogt: Data curation, validation, methodology, writing–review and editing. M. Truger: Data curation, investigation, methodology, writing–review and editing. R.A. Fernandez: Resources, formal analysis, investigation, writing–review and editing. M. Steinhardt: Resources, formal analysis, investigation, writing–review and editing. J. Weingart: Data curation, validation, writing–review and editing. R. Snaurova: Data curation, validation, writing–review and editing. S. Nerreter: Data curation, validation, writing–review and editing. E. Teufel: Data curation, validation, writing–review and editing. A. Garitano-Trojaola: Investigation, methodology, writing–review and editing. M. Da Viá: Resources, formal analysis, investigation, visualization, writing–review and editing. Y. Ruiz-Heredia: Resources, writing–review and editing. A. Rosenwald: Formal analysis, writing–review and editing. N. Bolli: Resources, formal analysis, validation, writing–review and editing. R. Hajek: Formal analysis, writing–review and editing. P. Raab: Resources, writing–review and editing. M.S. Raab: Resources, formal analysis, validation, writing–review and editing. N. Weinhold: Data curation, software, formal analysis, validation, writing–review and editing. C. Haferlach: Formal analysis, methodology, writing–review and editing. T. Haaf: Formal analysis, supervision, investigation, methodology, writing–review and editing. J. Martinez-Lopez: Resources, formal analysis, supervision, investigation, writing–review and editing. H. Einsele: Resources, formal analysis, supervision, funding acquisition, writing–review and editing. L. Rasche: Conceptualization, resources, formal analysis, supervision, funding acquisition, investigation, writing–original draft, writing–review and editing. K.M. Kortüm: Conceptualization, resources, supervision, funding acquisition, investigation, visualization, writing–original draft, project administration, writing–review and editing.

The authors thank Michael Rehli, MD (University Hospital Regensburg, Regensburg, Germany) for providing the pCpGL and the pCpGL-CMV/EF1 vectors.

K.M. Kortüm was supported by “Stiftung zur Förderung der Krebsforschung an der Universität Würzburg”, the “Stifterverband” and the “CDW Stiftung”. K.M. Kortüm and R. Hajek shared a funding of “Bayerisch-Tschechische Hochschulagentur” (BTHA). K.M. Kortüm and L. Rasche were funded by German Cancer Aid (Deutsche Krebshilfe) via the “Mildred Scheel Early Career Center” (MSNZ) program. L. Haertle is funded by the DFG (Deutsche Forschungsgemeinschaft) research fellowship HA 9529/1–1. S. Barrio was supported by the Instituto de Salud Carlos III (FIS No. PI21/00314). M. Simicek was supported by the “New Directions of Biomedical Research in the Ostrava Region” (No. CZ.02.1.01/0.0/0.0/18_069/0010060) finances provided from ERDF.

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.

Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/).

1.
Guerrero-Garcia
TA
,
Gandolfi
S
,
Laubach
JP
,
Hideshima
T
,
Chauhan
D
,
Mitsiades
C
, et al
.
The power of proteasome inhibition in multiple myeloma
.
Expert Rev Proteomics
2018
;
15
:
1033
52
.
2.
Rajan
AM
,
Rajkumar
SV
.
Treatment of newly diagnosed myeloma: bortezomib-based triplet
.
Semin Oncol
2016
;
43
:
700
2
.
3.
VanderLinden
RT
,
Hemmis
CW
,
Yao
T
,
Robinson
H
,
Hill
CP
.
Structure and energetics of pairwise interactions between proteasome subunits RPN2, RPN13, and ubiquitin clarify a substrate recruitment mechanism
.
J Biol Chem
2017
;
292
:
9493
504
.
4.
Bar-Nun
S
,
Glickman
MH
.
Proteasomal AAA-ATPases: structure and function
.
Biochim Biophys Acta
2012
;
1823
:
67
82
.
5.
Besche
HC
,
Peth
A
,
Goldberg
AL
.
Getting to first base in proteasome assembly
.
Cell
2009
;
138
:
25
8
.
6.
Laubach
J
,
Garderet
L
,
Mahindra
A
,
Gahrton
G
,
Caers
J
,
Sezer
O
, et al
.
Management of relapsed multiple myeloma: recommendations of the International Myeloma Working Group
.
Leukemia
2016
;
30
:
1005
17
.
7.
Barrio
S
,
Stühmer
T
,
Da Viá
M
,
Barrio-Garcia
C
,
Lehners
N
,
Besse
A
, et al
.
Spectrum and functional validation of PSMB5 mutations in multiple myeloma
.
Leukemia
2019
;
33
:
447
56
.
8.
Suzuki
E
,
Demo
S
,
Deu
E
,
Keats
J
,
Arastu-Kapur
S
,
Bergsagel
PL
, et al
.
Molecular mechanisms of bortezomib resistant adenocarcinoma cells
.
PLoS One
2011
;
6
:
e27996
.
9.
Verbrugge
SE
,
Assaraf
YG
,
Dijkmans
BA
,
Scheffer
GL
,
Al
M
, et al
.
Inactivating PSMB5 mutations and P-glycoprotein (multidrug resistance-associated protein/ATP-binding cassette B1) mediate resistance to proteasome inhibitors: ex vivo efficacy of (immuno)proteasome inhibitors in mononuclear blood cells from patients with rheumatoid arthritis
.
J Pharmacol Exp Ther
2012
;
341
:
174
82
.
10.
Oerlemans
R
,
Franke
NE
,
Assaraf
YG
,
Cloos
J
,
van Zantwijk
I
,
Berkers
CR
, et al
.
Molecular basis of bortezomib resistance: proteasome subunit beta5 (PSMB5) gene mutation and overexpression of PSMB5 protein
.
Blood
2008
;
112
:
2489
99
.
11.
Besse
A
,
Besse
L
,
Kraus
M
,
Mendez-Lopez
M
,
Bader
J
,
Xin
B-T
, et al
.
Proteasome inhibition in multiple myeloma: head-to-head comparison of currently available proteasome inhibitors
.
Cell Chem Biol
2019
;
26
:
340
51
.
12.
Brünnert
D
,
Kraus
M
,
Stühmer
T
,
Kirner
S
,
Heiden
R
,
Goyal
P
. et al
.
Novel cell line models to study mechanisms and overcoming strategies of proteasome inhibitor resistance in multiple myeloma
.
Biochim Biophys Acta Mol Basis Dis
2019
;
1865
:
1666
76
.
13.
Shi
C-X
,
Zhu
YX
,
Bruins
LA
,
Bonolo de Campos
C
,
Stewart
W
,
Braggio
E
, et al
.
Proteasome subunits differentially control myeloma cell viability and proteasome inhibitor sensitivity
.
Mol Cancer Res
2020
;
18
:
1453
64
.
14.
Tsvetkov
P
,
Mendillo
ML
,
Zhao
J
,
Carette
JE
,
Merrill
PH
,
Cikes
D,
et al
.
Compromising the 19S proteasome complex protects cells from reduced flux through the proteasome
.
Elife
2015
;
4
:
e08467
.
15.
Acosta-Alvear
D
,
Cho
MY
,
Wild
T
,
Buchholz
TJ
,
Lerner
AG
,
Simakova
O,
et al
.
Paradoxical resistance of multiple myeloma to proteasome inhibitors by decreased levels of 19S proteasomal subunits
.
Elife
2015
;
4
:
e08153
.
16.
Tsvetkov
P
,
Sokol
E
,
Jin
D
,
Brune
Z
,
Thiru
P
,
Ghandi
M
, et al
.
Suppression of 19S proteasome subunits marks emergence of an altered cell state in diverse cancers
.
Proc Natl Acad Sci USA
2017
;
114
:
382
7
.
17.
Soriano
GP
,
Besse
L
,
Li
N
,
Kraus
M
,
Besse
A
,
Meeuwenoord
N
, et al
.
Proteasome inhibitor-adapted myeloma cells are largely independent from proteasome activity and show complex proteomic changes, in particular in redox and energy metabolism
.
Leukemia
2016
;
30
:
2198
207
.
18.
Besse
A
,
Stolze
SC
,
Rasche
L
,
Weinhold
N
,
Morgan
GJ
,
Kraus
M
, et al
.
Carfilzomib resistance due to ABCB1/MDR1 overexpression is overcome by nelfinavir and lopinavir in multiple myeloma
.
Leukemia
2018
;
32
:
391
401
.
19.
Xu
H
,
Han
H
,
Song
S
,
Yi
N
,
Qian
C'
,
Qiu
Y,
et al
.
Exosome-transmitted PSMA3 and PSMA3-AS1 promote proteasome inhibitor resistance in multiple myeloma
.
Clin Cancer Res
2019
;
25
:
1923
35
.
20.
Duran-Ferrer
M
,
Clot
G
,
Nadeu
F
,
Beekman
R
,
Baumann
T
,
Nordlund
J
, et al
.
The proliferative history shapes the DNA methylome of B-cell tumors and predicts clinical outcome
.
Nat Cancer
2020
;
1
:
1066
81
.
21.
Agirre
X
,
Castellano
G
,
Pascual
M
,
Heath
S
,
Kulis
M
,
Segura
V
, et al
.
Whole-epigenome analysis in multiple myeloma reveals DNA hypermethylation of B cell–specific enhancers
.
Genome Res
2015
;
25
:
478
87
.
22.
Walker
BA
,
Wardell
CP
,
Chiecchio
L
,
Smith
EM
,
Boyd
KD
,
Neri
A
, et al
.
Aberrant global methylation patterns affect the molecular pathogenesis and prognosis of multiple myeloma
.
Blood
2011
;
117
:
553
62
.
23.
Horsthemke
B
.
Epimutations in human disease
.
Curr Top Microbiol Immunol
2006
;
310
:
45
59
.
24.
Amodio
N
,
Leotta
M
,
Bellizzi
D
,
Di Martino
MT
,
D'Aquila
P
,
Lionetti
M
, et al
.
DNA-demethylating and anti-tumor activity of synthetic miR-29b mimics in multiple myeloma
.
Oncotarget
2012
;
3
:
1246
58
.
25.
Alzrigat
M
,
Parraga
AA
,
Jernberg-Wiklund
H
.
Epigenetics in multiple myeloma: From mechanisms to therapy
.
Semin Cancer Biol
2018
;
51
:
101
15
.
26.
Dimopoulos
K
,
Søgaard Helbo
A
,
Fibiger Munch-Petersen
H
,
Sjö
L
,
Christensen
J
,
Sommer Kristensen
L
, et al
.
Dual inhibition of DNMTs and EZH2 can overcome both intrinsic and acquired resistance of myeloma cells to IMiDs in a cereblon-independent manner
.
Mol Oncol
2018
;
12
:
180
95
.
27.
Haertle
L
,
Barrio
S
,
Munawar
U
,
Han
S
,
Zhou
X
,
Vogt
C
, et al
.
Cereblon enhancer methylation and IMiD resistance in multiple myeloma
.
Blood
2021
;
138
:
1721
6
.
28.
Kalff
A
,
Khong
T
,
Mithraprabhu
S
,
Bergin
K
,
Reynolds
J
,
Bowen
KM
, et al
.
Oral azacitidine (CC-486) in combination with lenalidomide and dexamethasone in advanced, lenalidomide-refractory multiple myeloma (ROAR study)
.
Leuk Lymphoma
2019
;
60
:
2143
51
.
29.
Laubach
JP
,
Schjesvold
F
,
Mariz
M
,
Dimopoulos
MA
,
Lech-Maranda
E
,
Spicka
I
, et al
.
Efficacy and safety of oral panobinostat plus subcutaneous bortezomib and oral dexamethasone in patients with relapsed or relapsed and refractory multiple myeloma (PANORAMA 3): an open-label, randomized, phase II study
.
Lancet Oncol
2021
;
22
:
142
54
.
30.
Tata
PR
,
Rajagopal
J
.
Cellular plasticity: 1712 to the present day
.
Curr Opin Cell Biol
2016
;
43
:
46
54
.
31.
Boumahdi
S
,
de Sauvage
FJ
.
The great escape: tumor cell plasticity in resistance to targeted therapy
.
Nat Rev Drug Discov
2020
;
19
:
39
56
.
32.
Lohr
JG
,
Stojanov
P
,
Carter
SL
,
Cruz-Gordillo
P
,
Lawrence
MS
,
Auclair
D
, et al
.
Widespread genetic heterogeneity in multiple myeloma: implications for targeted therapy
.
Cancer Cell
2014
;
25
:
91
101
.
33.
Bolli
N
,
Avet-Loiseau
H
,
Wedge
DC
,
Van Loo
P
,
Alexandrov
LB
,
Martincorena
I
, et al
.
Heterogeneity of genomic evolution and mutational profiles in multiple myeloma
.
Nat Commun
2014
;
5
:
2997
.
34.
Giesen
N
,
Paramasivam
N
,
Toprak
UH
,
Huebschmann
D
,
Xu
J
,
Uhrig
S
, et al
.
Comprehensive genomic analysis of refractory multiple myeloma reveals a complex mutational landscape associated with drug resistance and novel therapeutic vulnerabilities
.
Haematologica
2022
;
107
:
1891
901
.
35.
Ziccheddu
B
,
Biancon
G
,
Bagnoli
F
,
De Philippis
C
,
Maura
F
,
Rustad
EH
, et al
.
Integrative analysis of the genomic and transcriptomic landscape of double-refractory multiple myeloma
.
Blood Adv
2020
;
4
:
830
44
.
36.
Zatopkova
M
,
Ševčíková
T
,
Fanfani
V
,
Chyra
Z
,
Říhová
L
,
Bezděková
R
, et al
.
Mutation landscape of multiple myeloma measurable residual disease: identification of targets for precision medicine
.
Blood Adv
2022
;
6
:
368
72
.
37.
Haertle
L
,
Maierhofer
A
,
Böck
J
,
Lehnen
H
,
Böttcher
Y
,
Blüher
M
, et al
.
Hypermethylation of the non-imprinted maternal MEG3 and paternal MEST alleles is highly variable among normal individuals
.
PLoS One
2017
;
12
:
e0184030
.
38.
Klug
M
,
Rehli
M
.
Functional analysis of promoter CpG methylation using a CpG-free luciferase reporter vector
.
Epigenetics
2006
;
1
:
127
30
.
39.
Shi
C-X
,
Kortüm
KM
,
Zhu
YX
,
Bruins
LA
,
Jedlowski
P
,
Votruba
PG
, et al
.
CRISPR genome-wide screening identifies dependence on the proteasome subunit PSMC6 for bortezomib sensitivity in multiple myeloma
.
Mol Cancer Ther
2017
;
16
:
2862
70
.
40.
Levin
A
,
Minis
A
,
Lalazar
G
,
Rodriguez
J
,
Steller
H
.
PSMD5 inactivation promotes 26S proteasome assembly during colorectal tumor progression
.
Cancer Res
2018
;
78
:
3458
68
.
41.
Walter
P
,
Ron
D
.
The unfolded protein response: from stress pathway to homeostatic regulation
.
Science
2011
;
334
:
1081
6
.
42.
Godderz
D
,
Dohmen
RJ
.
Hsm3/S5b joins the ranks of 26S proteasome assembly chaperones
.
Mol Cell
2009
;
33
:
415
6
.
43.
Kaneko
T
,
Hamazaki
J
,
Iemura
S-I
,
Sasaki
K
,
Furuyama
K
,
Natsume
T
, et al
.
Assembly pathway of the Mammalian proteasome base subcomplex is mediated by multiple specific chaperones
.
Cell
2009
;
137
:
914
25
.
44.
Shim
SM
,
Lee
WJ
,
Kim
Y
,
Chang
JW
,
Song
S
,
Jung
Y-K
.
Role of S5b/PSMD5 in proteasome inhibition caused by TNF-alpha/NFkappaB in higher eukaryotes
.
Cell Rep
2012
;
2
:
603
15
.
45.
Carter
SL
,
Cibulskis
K
,
Helman
E
,
McKenna
A
,
Shen
H
,
Zack
T
, et al
.
Absolute quantification of somatic DNA alterations in human cancer
.
Nat Biotechnol
2012
;
30
:
413
21
.
46.
Jin
Y
,
Xu
L
,
Wu
X
,
Feng
J
,
Shu
M
,
Gu
H
, et al
.
Synergistic efficacy of the demethylation agent decitabine in combination with the protease inhibitor bortezomib for treating multiple myeloma through the Wnt/beta-catenin pathway
.
Oncol Res
2019
;
27
:
729
37
.
47.
Lopez
JM
,
Beato
MS
,
Nerea
M
,
Varela
I
,
Derdak
S
,
Rapado
I
, et al
.
Characterization of subclonal changes along progression in multiple myeloma
.
Blood
2012
;
120
:
2924
.
48.
Barwick
BG
,
Skerget
S
,
Keats
JJ
,
Auclair
D
,
Lonial
S
,
Boise
LH
, et al
.
Multiple myeloma epigenetic programming prognostic of outcome converges with Loci reprogrammed in relapsed/refractory disease
.
Blood
2019
;
134
:
858
.

Supplementary data