Background. Proteasome inhibitors (PI) are a backbone for the treatment of multiple myeloma (MM). However, most MM patients ultimately acquire PI-resistance and die from MM. The biology of PI resistance is poorly understood and established treatment options for PI-refractory MM are lacking. We have recently generated a next-generation in vitro model for PI-refractory MM, based on single MM cell-derived subclones with specific (50-100x) resistance to either bortezomib (AMO-BTZ) or carfilzomib (AMO-CFZ). These cells consistently show downregulation of IRE1/XBP1, a hallmark of BTZ-refractory primary MM cells, and do or do not carry the PSMB5 A310G mutation. Our recent quantitative proteomic analysis suggested that the PI-resistant phenotype of AMO-BTZ and AMO-CFZ is in particular characterized by massive metabolic changes that render MM cells largely independent from proteasome activity. To verify such metabolic adaptation as a hallmark of proteasome inhibitor resistance of MM, and to identify therapeutic targets, we performed whole metabolome profiling of proteasome inhibitor-sensitive AMO-1 cells in comparison to BTZ or CFZ-resistant clones from the same cell line.

Methods. Metabolomic profiling was performed using Ultrahigh Performance Liquid Chromatography-Tandem Mass Spectroscopy (UPLC-MS/MS); data were quantified and normalized by Metabolon, Inc. Metaboanalyst v3.0 software was used for Metabolite Set Enrichment Analysis (MSEA). Reactive oxygen species (ROS) formation was determined by flow cytometry after staining the cells with H2DCFDA and exposure to 0.08% H2O2. Protein folding was determined by MERO-GFP construct, viability was determined by MTS assay.

Results. Based on MSEA, PI-adapted cells showed substantial metabolic changes in glutathione metabolism, protein biosynthesis, malate/aspartate shuttle, metabolism of purines, pyrimidines, amino acids and tricarboxylic acid (TCA). Relative to AMO-1, in both AMO-BTZ and AMO-CFZ, glutathione (l-γ-Glutamyl-l-cysteinyl-glycine; GSH and GSSG) was consistently identified as the metabolite with the most significant quantitative increase (fold change (FC) relative to AMO-1 for GSH in AMO-BTZ=52.92; in AMO-CFZ=56.79; FC for GSSG in AMO-BTZ=36.92; in AMO-CFZ=54.17), while many other metabolites in glycine/serine and cysteine/methionine pathways that are associated with glutathione synthesis were decreased. Consistently, PI-adapted cells showed accumulation of nicotinamide adenine dinucleotide (phosphate) (NAD(P)+) and reduced nicotinamide adenine dinucleotide (phosphate) (NAD(P)H) (FC for NAD+ in AMO-BTZ=55.81; in AMO-CFZ=2.16; FC for NADH in AMO-BTZ=19.10; in AMO-CFZ=1.65), presumably to maintain balance between reduced and oxidized glutathione, together with changes in TCA and malate/aspartate shuttle fueling the NAD(P)H production and transport, respectively. PI-resistant cells in addition showed altered carbohydrate metabolism and a shift in availability of membrane lipids, from lysolipids to sphingomyelins, in comparison to non-adapted cells.The adaptive changes in these cells correspond well to the proteomic changes previously described, confirming that the metabolic adaptive changes of PI-refractory MM cells are in particular in the regulation of redox homeostasis and protein folding. Functional experiments confirmed that the PI-adapted cells show larger capacity to buffer ROS production and that protein re-folding after DTT challenge is more effective in PI-refractory MM, compared to PI-sensitive cells. Consequently, co-treatment of PI-adapted cells with respective proteasome inhibitors (bortezomib and carfilzomib) in the presence of the inhibitor of protein-disulfide isomerase (PDI), 16F16, re-sensitized cells to proteasome inhibition (FC of IC50 for bortezomib or carfilzomib alone vs bortezomib or carfilzomib + 16F16 in AMO-BTZ=1.26 or 1.88, respectively; in AMO-CFZ=1.25 or 2.59 respectively; in AMO-1=1.1 or 0.95, respectively.

Conclusion. PI refractory MM cells show massive metabolic changes in particular in glutathione and NAD(P)H production. This is associated with increased protein folding capacity and likely results in decreased proteasome load. PI-refractory MM cells can be re-sensitized to PI treatment by reducing glutathione levels or by affecting protein folding.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.