Molecular chaperone GRP78 enhances aggresome delivery to autophagosomes to promote drug resistance in multiple myeloma

Abstract

Despite the clinical benefit of the proteasome inhibitor bortezomib, multiple myeloma (MM) patients invariably relapse through poorly defined mechanisms. Myeloma cells inevitably develop chemoresistance that leads to disease relapse and patient-related deaths. Studies in tumor cell lines and biopsies obtained from patients refractory to therapy have revealed that myeloma cells adapt to stress by inducing expression of glucose-regulated protein 78 (GRP78), an endoplasmic reticulum (ER) chaperone with anti-apoptotic properties. Treatment of myeloma cells with bortezomib increased GRP78 levels and activated GRP78-dependent autophagy. Expression profiling indicated that GRP78-encoding HSPA5 was significantly upregulated in bortezomib-resistant cells. Co-treatment with the anti-diabetic agent metformin suppressed GRP78 and enhanced the anti-proliferative effect of bortezomib. Bortezomib treatment led to GRP78 co-localization with proteotoxic protein aggregates, known as aggresomes. Pharmacologic suppression, genetic ablation or mutational inactivation of GRP78 followed by bortezomib treatment led to the accumulation of aggresomes but impaired autophagy and enhanced anti-myeloma effect of bortezomib. GRP78 was co-immunoprecipitated with the KDEL receptor, an ER quality control regulator that binds proteins bearing the KDEL motif to mediate their retrieval from the Golgi complex back to the ER. Taken together, we demonstrate that inhibition of GRP78 functional activity disrupts autophagy and enhances the anti-myeloma effect of bortezomib.

Figure 1. Effect of bortezomib and metformin treatment on GRP78 in myeloma cells

A. Shown is the Fold-increase in the expression of individual HSP pathway genes in MM patient samples compared to MGUS samples (top panel). Also shown is the fold-increase in the expression of individual HSP pathway genes in bortezomib resistant RPMI8226 cells relative to drug-naïve cells (bottom panel). Shown is the fold-increase in relative expression determined by microarray-based profiling using Affymetrix 3.0 chips. B. Western blot comparing GRP78 levels in parental and bortezomib resistant cells. Parental and bortezomib resistant were grown in bortezomib for 18 hours prior to preparation of cell lysates. C. GRP78 staining of myeloma cells by IHC and confocal microscopy in RPMI8226 cells that had been treated with bortezomib (10nM), metformin (1mM) or both agents. Cells were treated for 18 hours under standard growth conditions. D. Quantitation of GRP78 levels based upon the relative level of fluorescent intensity detected by IHC staining. E. GRP78 staining RPMI8226 cells by IHC and confocal microscopy treated with various concentrations of bortezomib and metformin for 18 hours. Shown are representative images obtained from the same experiment performed multiple times.

Figure 2. HSPA5 knockdown effect on aggresome formation

A. Western blot of GRP78 levels in lysates from RPMI8226 cells transfected with control or HSPA5-specific shRNA. Ponceau staining of the membrane used for the GRP78 blot is shown. B. RPMI8226 cells transfected with either scrambled control or HSPA5-specific shRNA were treated with drugs as indicated and the level of GRP78 determined by IHC and confocal microscopy. C. RPMI8226 cells transfected with either scrambled control or HSPA5-specific shRNA were treated with drugs as indicated for 18 hours and aggresomes visualized by IHC and confocal microscopy. D. Quantitation of aggresome levels based upon the relative level of fluorescent intensity detected by IHC and confocal microscopy. E. RPMI8226 cells transfected with scrambled control or shRNA to inactivate the stress transducers ATF6, IRE1α or PERK were treated with drugs as indicated and GRP78 levels determined by IHC and confocal microscopy. shRNA-mediated knockdown of the three stress transducers was validated by qRT-PCR. F. Western blot of myeloma cell lysate after treatment with bortezomib or metformin as indicted for 18 hours.

Figure 3. Effect of bortezomib and metformin on autophagosome formation

A. RPMI8226 cells were treated with either bortezomib (10nM), metformin (1mM) or both agents for 18 hours under standard growth conditions. GRP78 was detected by IHC and confocal microscopy. Aggresomes and autophagosomes were detected using the dye-based methods. Shown are representative images seen on in at least three different experiments. B. Relative level of fluorescent intensity of autophagosomes after treatment of RPMI8226 cells with the indicated drugs. C. RPMI8226 cells were transfected with scrambled (control) or HSPA5-specific shRNA, treated with drugs as indicated and GRP78, aggresomes and autophagosomes detected as in Figure 3A. Shown are representative images seen on in at least three different experiments. D. Relative level of fluorescent intensity of autophagosomes after treatment of RPMI8226 cells transfected with either scrambled control or HSPA5 shRNA and then treated with drugs as indicated. E. Co-localization of GRP78 with aggresomes as determined by IHC and confocal microscopy. RPMI8226 cells were treated with bortezomib (10nM), metformin (1mM) or both and stained using a GRP78-specirfic antibody, for aggresomes using dye-based reagent or both the GRP78 antibody and the dye-based reagent. Shown are representative images from multiple experiments. F. Co-localization of GRP78 with autophagosome as determined by IHC and confocal microscopy. RPMI8226 cells were treated with bortezomib (10nM), metformin (1mM) or both and stained using a GRP78-specirfic antibody, for autophagsomes using dye-based reagent or both the GRP78 antibody and the dye-based reagent. Shown are representative images from multiple experiments. G. Effect of bortezomib and metformin on aggresomes and autophagosomes in MM patient tumor cells. Patient bone marrow was obtained, CD138⁺ cells purified, treated with drugs as indicated and aggresomes and autophagosomes detected using the dye-based methods and confocal microscopy.

Figure 4. Effect of GRP78 genetic silencing on bortezomib-induced autophagosome formation

A. U266 cells were transfected with plasmids that expressed either control (pcDNA3.1) or a GRP78 mutant (P495L). Cells were treated with bortezomib (10nM), metformin (1mM) or both for 18h. Aggresomes were detected by the by dye-based method. Shown are representative images from multiple experiments. B. U266 cells were transfected with plasmids that expressed either control (pcDNA3.1) or the GRP78 mutant. Cells were treated with bortezomib (10nM), metformin (1mM) or both for 18h. Autophagosomes were detected by dye-based methods. Shown are representative images. C. U266 cells were transfected with plasmids that expressed GRP78-WT or the GRP78 mutant. Cells were treated with bortezomib, lysates immunoprecipitated and probed by western blot to detect the association of aggresome (p62 and HDAC6) or autophagosome pathway (KDEL receptor and LC3B) effectors with GRP78. D. U266 cells were transfected with plasmids that expressed either shRNA to inactivate control (scrambled) or HSPA5. Cells were treated with bortezomib at indicated concentrations and the effect on viability determined using the XTT assay. Values represent the mean of triplicate measurements and error bars represent the standard deviation (SD). E. U266 cells were transfected with plasmids that expressed either GRP78-WT or the GRP78 mutant. Cells were treated with bortezomib as indicated and the effect on viability determined using the XTT assay. Values represent the mean of triplicate measurements and error bars represent the SD.

Figure 5. Metformin effect on myeloma viability

A. Dose dependent effect of the biguanides alone or combined with bortezomib on myeloma proliferation. RPMI8226 and U266 cells were incubated with the biguanides alone or biguanides and bortezomib (2nM) for 72 hours. Bortezomib alone at 2nM yielded ~10% reduction in cell viability. Proliferation was determined using the XTT assay and error bars represent SD values determined from triplicate measurements. B. Effect of biguanide in AMPK-WT and AMPK-DKO MEF proliferation measured using the XTT assay and error bars represent SD values determined from triplicate measurements. Annexin-positive cells were quantitated by flow cytometry.

Figure 6. Effect of HSPA5 knockdown on bortezomib-induced cleavage of caspases and PARP

A. U266 cells were transfected with plasmids that expressed either shRNA to control (scrambled) or HSPA5. Cells were treated with bortezomib, lysates prepared and probed by western blot using antibodies to Caspase 3, 8 and 10. B. U266 cells were transfected with plasmids that expressed either shRNA to control (scrambled) or HSPA5. Cells were treated with bortezomib, lysates prepared and probed by western blot using antibodies to PARP.