Mechanisms of Proteasome Inhibitor Resistance Selected By Clonal Evolution in Multiple Myeloma

Abstract

Background: Various treatment regimen in multiple myeloma (MM) are based on proteasome inhibition (PI). Although effective at therapy start, most patients relapse and develop drug resistance over time. To better understand the molecular underpinnings associated with PI resistance, we studied genetic and epigenetic alterations of the 26S proteasome genes in PI exposed patients. Methods: We performed a meta-analysis comprised of M3P targeted sequencing datasets and other publicly available WGS/WES sets. A selection of most frequently found mutations was tested in vitro regarding their impact on PI response. DNA promoter methylation of a subset of proteasome genes was determined by targeted Deep Bisulfite Sequencing (DBS), followed up on expression (Taqman qPCR) and validated at functional level (dual-luciferase reporter assay system). Results: The meta-analysis was conducted for a total of 1,752 MM cases, with 1,241 newly diagnosed (NDMM) and 511 progressed MM (PMM) samples. We identified mutations in 32 proteasome genes, with increased incidence from NDMM (6.1% of the patients had mutations in one or more genes) to 10.2% at PMM. Besides PSMB5 encoding for the inhibitor binding site, mainly 19S subunit genes were mutated in areas that impact the recognition of ubiquitinated proteins (PSMD1 and PSMD2), are involved in protein unfolding or gate opening (PSMC1-6). We stably expressed proteasome subunit components bearing frequently observed patient-derived mutations in RPMI-8226 MM cells. All mutants (PSMC6 R242Q, PSMD1 E824K, PSMD1 A887T, PSMD2 M646I, PSMC2 Y429S) displayed an impaired PI response towards Bortezomib (Figure A), Carfilzomib and Ixazomib. Of note, in a fluorescent based, in house clonal competition assay, Bortezomib resistant PSMB5 A20T mutants were outcompeted by WT cells when the drug was removed from co-culture, demonstrating a survival disadvantage through the mutation itself, when no selective pressure of proteasome inhibition was applied. In general, somatic mutations on single gene level were relatively rare, PSMD1, our best candidate gene, was mutated in only 2% of the analyzed MM patients, the remaining genes even to an lower extent, but the proteasome as a whole structure, was frequently affected by mutations. Therefore, we hypothesized that clonal evolution might also act by selecting epigenetic alterations. To address this, we analyzed promoter methylation of PSMC2, PSMC5, PSMC6, PSMD1 and PSMD5 in 42 MM patients by DBS. For PSMD5, NDMM patients and PBMCs were nearly unmethylated (mean methylation: 2.0%±0.020 and 2.0%±0.026), but PMM displayed noticeably increased hypermethylation (6%±0.099). We demonstrated epigenetic silencing by promoter hypermethylation (methylation degree ≥15%) in 20% of PI resistant patients. Moreover, at RNA level we confirmed that patients with high methylation had low expression of PSMD5 and vice versa. To explore the regulatory impact of PSMD5 promoter methylation on gene regulation, we now cloned our amplicon into the backbone of a CpG-free vector (pCpGL). The reporter vector with either the methylated or the unmethylated insert was co-transfected with a Renilla control vector into L363 cells. Luciferase activity of the unmethylated PSMD5 construct was 8 times increased compared to the methylated vector and the controls (vector without insert and non-transfected cells) (Figure B), confirming gene regulation through methylation of this gene. PSMD5, encodes a chaperon recently characterized as the main regulator of 19S proteasome assembly. Gene silencing may represent an adaptive way of cancer cells to bypass proteasome inhibition and escape PI induced proteolytic toxicity by increasing their protein turnover capacity. Conclusion: Altogether our data give evidence that after PI exposition, MM patients harbor acquired regulatory DNA mutations as well as epimutations that affect different proteasomal subunits and, by different modes of action, compromise proteasome activity to escape PI therapy. Figure Disclosures Bittrich: German Research Foundation (DFG): Other: N/A; University of Würzburg: Other: N/A; Bristol Myers Squibb: Research Funding; Otsuka Pharmaceuticals Europe: Other: N/A; Pfizer: Other: Travel Funding; Wilhelm Sander Foundation: Research Funding; Else Kröner Fresenius Foundation: Research Funding; Celgene: Other: Travel Funding, Research Funding; JAZZ Pharmaceuticals: Other: Travel Funding; AMGEN: Other: Travel Funding; University Hospital Wuerzburg: Employment; SANOFI Aventis: Membership on an entity's Board of Directors or advisory committees, N/A, Research Funding.