Introduction: Clinical efficacy of proteasome inhibitors (PIs) has established multiple myeloma (MM) as the paradigm cancer type in which to therapeutically target protein homeostasis. However, PI resistance is one of the key clinical challenges in this disease. While mechanisms of PI resistance appear complex, proteomic analysis showed multiple heat shock proteins to be highly upregulated in PI-resistant cell line models (Soriano et al, Leukemia (2016) 30:2198). Furthermore, HSPA1B, which encodes a cytosolic Hsp70, is one of the genes which best-stratifies PI-sensitive versus resistant MM cell lines (Mitra et al, Blood Cancer J (2017) 7:e581). These findings support the hypothesis that Hsp70 family chaperones may be critical in maintaining plasma cell protein homeostasis by reducing unfolded protein burden in the PI-refractory setting, indicating potential as a therapeutic target. Targeting Hsp70 may also therefore be effective in combination with PIs or other agents targeting MM protein homeostasis. Here, we profiled a series of novel, allosteric Hsp70 inhibitors of the "JG" class (Shao et al, J Med Chem (2018) 61:6163). We aimed to evaluate their role as anti-myeloma therapeutics as well as their utility as tool compounds to better define targetable protein homeostasis networks in MM.

Methods: "JG" compounds were screened in quadruplicate against a panel of 13 cell lines: 9 MM, one AML, 3 immortalized normal bone marrow stromal, in 384-well plates with viability measured by CellTiter-Glo. The parental compound JG98 was tested in combination with the p97 inhibitor CB-5083, the PI bortezomib, and Hsp90 inhibitor 17-DMAG in RPMI-8226, MM1.S, KMS34, HS5, and CMK cell lines for 72 hours, with synergy and antagonism quantified using the ZIP score method. Murine studies were performed using luciferase-labeled RPMI-8226 implanted in NOD scid gamma (NSG) mice (n = 5 mice per arm). For cell surface mass spectrometry experiments, RPMI-8226 and AMO-1 cells were treated with JG342 at the LD20 for 48 hours prior to biotinylation of cell surface glycoproteins (n=3), followed by enrichment for biotinylated proteins with Neutravidin beads and LC-MS sample processing and analysis. For pulsed-SILAC experiments (n = 2), MM1.S cells were grown in light medium for six doublings, followed by exchange to heavy medium (Arg10 and Lys8) with JG342 treatment at the LD10 or DMSO, with timepoint collections performed at 16, 21, and 26 hours, followed by sample processing and LC-MS analysis. All mass spectrometry data was analyzed in MaxQuant and statistical analysis was performed in Perseus and R.

Results: Supporting our hypothesis, drug screening of 16 JG compounds revealed consistently greater potency versus a PI-resistant AMO-1 cell line versus the parental line (Fig. A). We focused on the lead compound JG342, which exhibits sub-µM LC50s (140-500 nM) against a panel of myeloma cell lines and a favorable therapeutic index against bone marrow stromal cells. However, all tested JG molecules, including JG342, showed unfavorable pharmacokinetics in NSG mouse models and could not demonstrate in vivo anti-MM efficacy at achievable doses. Therefore, we focused on JG molecules as tool compounds to probe protein homeostasis in MM and further validate Hsp70 as a therapeutic target in this disease. Combination assays using JG98, bortezomib, CB-5083, and 17-DMAG revealed surprisingly diverse, cell line-specific protein homeostasis wiring networks leading to differential synergism and antagonism in dual therapies targeting the proteostasis machinery. To evaluate potential combination effects of Hsp70 inhibition with immunotherapies, we employed cell surface proteomics and found that JG342 leads to broad changes in the MM surface proteome including decrease in expression of BCMA (Fig. B). Lastly, we employed pSILAC mass spectrometry to profile JG342 mediated changes in protein synthesis and degradation. Calculated protein half-lives revealed a distinct subset of differentially regulated proteins after Hsp70 inhibition including aminotransferases and DNA chaperones.

Conclusions: Inhibition of Hsp70 is a promising therapeutic strategy to overcome PI resistance in myeloma, but new molecules with more favorable in vivo properties are needed to benefit patients. The "JG" molecules also provide a novel set of tool compounds for exploring protein homeostasis biology not only in MM but across biological systems.


Wiita:Sutro Biopharma: Research Funding; TeneoBio: Research Funding.

Author notes


Asterisk with author names denotes non-ASH members.