Interfering with Arginine Metabolism As a New Treatment Strategy for Multiple Myeloma

Introduction

Starvation of tumor cells from the amino acid arginine has recently gained particular interest because of the downregulation of the rate-limiting enzyme argininosuccinate synthethase 1 (ASS1) in various cancer entities. ASS1-deficient cells cannot resynthesize arginine from citrulline and are therefore considered arginine auxotrophic. The arginine depleting enzyme arginine deiminase (ADI-PEG20, Polaris Pharmaceuticals) is currently tested in phase I-III clinical trials for different arginine auxotrophic cancers. The natural arginine analogue canavanine can compete with arginine for arginyl-tRNA-binding sites and consequently be incorporated into nascent proteins instead of arginine. Canavanine could therefore potentially further disturb intracellular protein homeostasis, especially under arginine deprivation. The sensitivity of myeloma cells towards arginine depletion strategies has not been analyzed so far.

Methods

Human myeloma cell lines and CD138-sorted primary human myeloma cells from patient bone marrow were screened for ASS1 expression by western blotting (WB). The cells were cultured in arginine free medium and assessed for proliferation and metabolic activity (CFSE/MTT assays), apoptosis (caspase-3 cleavage) and cell death (annexinV/propidium iodide). Canavanine was supplied in both arginine-sufficient and -deficient conditions. The level of intracellular protein stress was determined by WB and/or flow cytometry analysis for ubiquitinated proteins, phosphorylated eukaryotic initiation factor 2α (peIF2α) and the spliced isoform of the X-Box binding protein 1 (Xbp1s). Repetitive ADI-PEG20 ± canavanine application i.p. were tested in vivo in an U266 myeloma xenograft model in NOD/SCID/IL2Rcg^-/^- (NSG) mice. Arginine and canavanine levels in plasma were determined by HPLC. Tumor growth was measured, mice were assessed for survival, weight and side effects. Tumor tissues were analyzed for caspase-3 cleavage and Ki67 expression by immunohistochemistry.

Results

5 of 6 myeloma cell lines were negative for ASS1. Also, ASS1 was either not or only weakly expressed in the majority of primary CD138⁺ myeloma patient samples. Arginine starvation induced an arrest of cell proliferation and/or metabolic activity of primary myeloma cells and myeloma cell lines after 18-24 h. Addition of citrulline could only rescue ASS1 positive myeloma cells due to the intracellular resynthesis of arginine. Arginine starvation alone led to delayed induction of apoptosis (e.g. 35% cell death of NCI-H929 cells after 72 h of treatment). Addition of 100 mM canavanine strongly increased cell death specifically in the context of arginine deficiency (e.g. cell death in NCI-H929 cells: 87% after 24 h, 100 % after 48h) while it was non-toxic and had no effect on cell viability under physiological arginine conditions. Co-application of canavanine induced ubiquitination of cellular proteins and led to the prolongation of a fatal unfolded protein response (UPR) as measured by markedly elevated Xbp1s levels. Prolonged UPR ultimately led to the induction of apoptosis as reflected by annexin V binding and caspase-3 cleavage. In an U266 myeloma NSG xenograft model, systemic arginine depletion by ADI-PEG20 suppressed tumor growth in vivo and significantly prolonged median survival of mice when compared with the control group (22±3 vs. 15±3 days). Canavanine treatment alone had no influence on viability (13±0 days). However, the combination of ADI-PEG20 and canavanine demonstrated the longest median survival (27±7 days). Histological examination of explanted tumors showed the highest rates of caspase-3 cleavage in the ADI-PEG20/canavanine group.

Conclusion

Myeloma cells are mostly arginine auxotrophic and can be selectively targeted by arginine starvation. Combination of arginine depletion with the arginine analogue canavanine leads to a highly efficient and specific tumor cell eradication and should be further optimized in multiple myeloma preclinical models.