|Home | About | Journals | Submit | Contact Us | Français|
The proteasome inhibitor bortezomib is clinically approved for the treatment of multiple myeloma. However, long-term remissions are difficult to achieve, and myeloma cells often develop secondary resistance to proteasome inhibitors. We recently demonstrated that the extraordinary sensitivity of myeloma cells toward bortezomib is dependent on their extensive immunoglobulin synthesis, thereby triggering the terminal unfolded protein response (UPR). Here, we investigated whether verapamil, an inhibitor of the multidrug resistance (MDR) gene product, can enhance the cytotoxicity of bortezomib. The combination of bortezomib and verapamil synergistically decreased the viability of myeloma cells by inducing cell death. Importantly, bortezomib-mediated activation of major UPR components was enhanced by verapamil. The combination of bortezomib and verapamil resulted in caspase activation followed by poly(ADP-ribose) polymerase cleavage, whereas nuclear factor κB (NF-κB) activity declined in myeloma cells. Also, we found reduced immunoglobulin G secretion along with increased amounts of ubiquitinylated proteins within insoluble fractions of myeloma cells when using the combination treatment. Verapamil markedly induced reactive oxygen species production and autophagic-like processes. Furthermore, verapamil decreased MDR1 expression. We conclude that verapamil increased the antimyeloma effect of bortezomib by enhancing ER stress signals along with NF-κB inhibition, leading to cell death. Thus, the combination of bortezomib with verapamil may improve the efficacy of proteasome inhibitory therapy.
Multiple myeloma, a virtually incurable plasma cell neoplasia, is characterized by the production of large amounts of monoclonal immunoglobulins and accounts for approximately 10% of all hematologic cancers . Existing therapeutic strategies such as high-dose chemotherapy followed by hematopoietic stem cell transplantation prolong survival of multiple myeloma patients but rarely induce long-lasting complete remissions. These treatments are also associated with severe adverse effects .
The proteasome inhibitor bortezomib (Velcade) markedly improved the treatment options for patients with relapsed multiple myeloma by inducing apoptosis in myeloma cells . The dipeptidyl boronic acid derivative bortezomib is a highly selective and reversible inhibitor of the 26S proteasome, a multienzyme complex present in all eukaryotic cells. The 26S proteasome degrades supernumerous, defective, or misfolded proteins, which are targeted for proteasomal degradation by polyubiquitinylation. In addition, it plays a fundamental role in cellular homeostasis as a critical regulator of cell proliferation and apoptosis [4,5].
The antitumor effect of bortezomib has been demonstrated in vitro and in vivo for various types of cancers. Myeloma cells seem to be exceptionally sensitive. Even the growth of chemotherapy-resistant myeloma cell lines was inhibited by bortezomib treatment . Bortezomib exerts its effect through multiple pathways that target both the tumor cell and its environment. The cytotoxic effect of bortezomib seems to be partially due to the inhibition of the antiapoptotic transcription factor nuclear factor κB (NF-κB). Bortezomib stabilizes endogenous inhibitor of kappa B alpha (IκBα) that sequesters NF-κB in the cytoplasm and prevents transcriptional activation of NF-κB target genes .
Importantly, we and others demonstrated that bortezomib-induced apoptosis is caused by excessive endoplasmic reticulum (ER) stress, activating the terminal unfolded protein response (UPR), especially in cells with extensive synthesis of secretory proteins [8–11].
The UPR is a signaling pathway from the ER to the nucleus triggered by the accumulation of misfolded proteins in the ER lumen and is essential for plasma cell differentiation and survival. The UPR includes three mechanisms to handle the vast increase of unfolded proteins: transcriptional induction of target genes enhancing protein folding, general translational repression, and ER-associated degradation to eliminate misfolded proteins. However, overwhelming ER stress activates the terminal UPR, leading to apoptosis [12,13].
Some myeloma patients are resistant or become refractory to ongoing bortezomib treatment . To improve the efficacy of proteasome inhibitor-based treatments and to overcome primary and secondary resistance, drugs augmenting the antitumor properties of bortezomib in myeloma cells are required. We identified the L-type calcium channel antagonist verapamil (Isoptin; Abbott, Wiesbaden, Germany), clinically used for the treatment of cardiac arrythmias, hypertension, and, most recently, for cluster headaches, as a promising combination partner with bortezomib. The phenylalkylamine derivative verapamil potently inhibits the influx of calcium ions into cells . Further, in drug-resistant leukemic cell lines, verapamil interfered with the multidrug resistance (MDR)-based drug elimination by decreasing P-glycoprotein (P-gp) expression . In this study, we observed that verapamil enhanced the proapoptotic effect of bortezomib. Increased cell death was associated with induction of terminal UPR and autophagy; however, a causal link and the molecular mechanisms require further investigation.
For immunoblot analysis, the following primary antibodies were used: mouse monoclonal anti-GRP78 (BiP), rabbit polyclonal anti-GRP94, and mouse monoclonal anti-poly(ADP-ribose) polymerase (PARP; BD Pharmingen, Heidelberg, Germany); mouse monoclonal anti-Bcl-2, rabbit polyclonal anti-Bax, rabbit polyclonal anti-Bim, mouse monoclonal anti-caspase 9, rabbit polyclonal anti-CHOP, rabbit polyclonal anti-p-eIF2α, mouse monoclonal anti-Hsp70, rabbit polyclonal anti-inositol-requiring transmembrane kinase/endonuclease 1α (IRE1α), rabbit polyclonal anti-p-PKR-like ER kinase (PERK), and rabbit polyclonal anti-X-box binding protein (XBP-1; all from Santa Cruz Biotechnology, Santa Cruz, CA); rabbit polyclonal anti-(active) Jun N-terminal kinase (JNK) and rabbit polyclonal anti-phospho-p38 (Promega, Madison, WI); rabbit polyclonal anti-actin (Sigma, Taufkirchen, Germany); mouse monoclonal anti-ATF4 (Abnova GmbH, Heidelberg, Germany); mouse monoclonal anti-ATF6a (Acris, Herford, Germany); rabbit polyclonal anti-LC3 (MBL, Nagoya, Japan); rabbit polyclonal anti-LMP7 (Novus Biologicals, Littleton, CO); rabbit polyclonal anti-PSMB5 (Abcam, Cambridge, UK); and mouse monoclonal anti-ubiquitin (Zymed Laboratories, Invitrogen Corporation, Carlsbad, CA). As secondary antibodies, we used HRP-conjugated goat anti-mouse immunoglobulinG(IgG), goat antirabbit IgG (Jackson Immunoresearch Laboratories, Inc, West Grove, PA), and donkey antigoat IgG (Santa Cruz Biotechnology). For flow cytometry, we used mouse anti-human CD243 (MDR1) AlexaFluor 647-conjugated antibody (Serotec, Düsseldorf, Germany). ELISA was performed with goat anti-human IgG (Jackson Immunoresearch Laboratories, Inc) and HRPconjugated goat anti-human IgG (Southern Biotech, Birmingham, AL).
Bortezomib (PS-341, Velcade; Janssen-Cilag, Neuss, Germany) and verapamil (Isoptin, Abbott) were obtained from the University Hospital Pharmacy, Erlangen. PS-1145 dihydrochloride (Sigma-Aldrich).
Cell lines used in this study were grown in RPMI 1640 medium supplemented with 50 U/ml penicillin, 50 µg/ml streptomycin, 1 mM sodium pyruvate, 2 mM l-glutamine, 50 µM β-mercaptoethanol, 10% fetal bovine serum for RPMI 8226 cells (myeloma), ARH 77 cells (plasma cell leukemia), and 20% bovine serum for JK-6L cells (myeloma) . Cells were maintained at 37°C in a humidified incubator containing 5% CO2. The medium and all supplements were obtained from Gibco (Invitrogen, Karlsruhe, Germany).
Cells (1 x 105) were treated with 10 nM bortezomib and/or 70 µM verapamil for 16 hours and incubated for another 4 hours with Alamar-Blue (BioSource International, Inc, Camarillo, CA). Activity of the mitochondrial dehydrogenase results in conversion of the coloring, which was followed by measurement of the absorption using a spectrophotometer (SpectraMax 190; Molecular Devices, Sunnyvale, CA).
Surface staining was performed as described . The cells were analyzed by flow cytometry using a FACS Calibur (BD Biosciences, San Jose, CA). Data analyses were performed using the Cell Quest software (BD Biosciences).
Cells (1 x 105) were treated with 10 nM bortezomib and/or 70 µM verapamil for 16 hours. Fifty micromolars of the broad-range caspase inhibitor zVAD-FMK (Alexis Biochemicals, San Diego, CA) was used to block caspase activation. Staining with annexin V-fluorescein isothiocyanate (FITC; Genethor, Berlin, Germany) and propidium iodide (Sigma-Aldrich) was performed as described .
Cells (5 x 104) were treated with 10 nM bortezomib and/or 70 µM verapamil for 16 hours. Caspase 3/7 activity was measured with the caspase-Glo-3/7 assay (Promega) according to the manufacturer's instructions. Luminescence was measured using the 96-well plate reader SpectraMax 190 (Molecular Devices, Sunnyvale, CA).
Cells (2 x 104) were treated with 10 nM bortezomib and/or 70 µM verapamil for 8 hours. The chymotrypsin-like activity was detected using the luminogenic proteasome substrate-based Proteasome-Glo Chymotrypsin-Like Cell-Based Assay Kit (Promega) according to the manufacturer's instructions. Luminescence was measured using the 96-well plate reader SpectraMax 190 (Molecular Devices, Sunnyvale, CA).
The fluorescent dye 3,3-dihexyloxacarbocyanine iodide (DiOC6; Sigma-Aldrich) was used to reveal disruption of the mitochondrial transmembrane potential (ΔΨm). For measurement, cells (1 x 105) were treated with 40 nM bortezomib and/or 70 µMverapamil for 10 hours. Afterward, cells were incubated with 120 nM DiOC6 dye for 20 minutes at 37°C in the dark. Stained cells were resuspended in 20 ng/ml isotonic propidium iodide solution to exclude dead cells and were analyzed by flow cytometry (Epics XL; Coulter Co, Miami, FL).
Cells (1 x 105) were treated with 40 nM bortezomib and/or 70 µM verapamil and incubated for 2 hours with 10 µM 2′,7′-dichlorofluorescein diacetate (DCFH; Sigma-Aldrich) at 37°C. DCFH penetrates the cells and is in turn oxidized to DCF in the presence of reactive oxygen species (ROS). To eliminate ROS production, cells were preincubated with 10 mM N-acetylcysteine (NAC; Sigma-Aldrich) 1 hour before adding the inhibitors. H2O2 (400 µM) was selected as a donor to get maximal DCF intensity. DCF fluorescence intensity was determined by flow cytometric analysis (Epics XL; Coulter Co).
Cells (2 x 106) were treated with 10 nM bortezomib and/or 70 µM verapamil for up to 16 hours. To generate total cell lysates, equal amounts of cells were washed with PBS and directly lysed in SDS sample buffer containing β-mercaptoethanol. For detergent-insoluble fractions, cells were lysed in TNES buffer (50 mM Tris-HCl, pH 7.5, 1% NP-40, 2 mM EDTA, 100 mM NaCl), with freshly added protease inhibitor cocktail (Serva, Heidelberg, Germany). The supernatants contain the soluble fractions. NP-40 detergent-insoluble pellets were resuspended in TNES buffer and sonificated for 10 seconds at 50 Won ice using a homogenizer (SONOPLUS; Bandlin Electronics, Berlin, Germany). Protein concentrations were colorimetrically determined using bicinchoninic acid (BCA Protein Assay; Pierce, Rockford, IL). Proteins were separated on 10% or 12% SDS-polyacrylamide gels and transferred onto nitrocellulose membranes (Schleicher & Schuell, Dassel, Germany). Membranes were blocked with 5% nonfat dry milk (Roth, Karlsruhe, Germany), probed with antibodies and developed using the enhanced chemiluminescence method.
Electromobility shift analysis (EMSA) was performed using oligonucleotides for NF-κB end labeled with IRDye according to the manufacturer's instructions (Metabion, Planegg-Martinsried, Germany). Briefly, cells (5 x 106) were treated with 10 nM bortezomib and/or 70 µM verapamil for 8 hours, and nuclear extracts were prepared as previously described . Ten microliters of probe mix (1 µg poly-dIdC, 0.25% Tween 20, 0.1% NP-40, 20 µg BSA, 1 x lipage buffer [0.1 M Tris-HCl, pH 7.5, 0.5 M NaCl, 10 mM EDTA, pH 7.5, 50% glycerol, before usage add 0.5 mM DTT], 0.1 pM oligo-IRDye-700) and Orange G loading buffer (20 mg Orange G, 15% Ficoll) were added to 10 µg of nuclear extracts. The reaction mixtures were incubated by gently shaking at room temperature for 30 minutes in the dark. After a prerun for 30 to 60 minutes at 150 V, samples were loaded on a 4% nondenaturing polyacrylamide gel and run for 2.5 hours at 150 V. Gels were analyzed using an Odyssey infrared imaging system (LI-COR Biosciences, Bad Homburg, Germany).
Cells (2 x 106) were treated with 10 nM bortezomib and/or 70 µM verapamil for up to 12 hours. Total RNA was isolated using the Mini RNAeasy Kit (Qiagen, Hilden, Germany), including digestion with RNase-free DNAse. Complementary DNA (cDNA) synthesis from 1 µg total RNA was performed using the SuperScript II Reverse Transcriptase (Invitrogen, Karlsruhe, Germany) and amplified with Taq polymerase (NEB, Frankfurt/Main, Germany) using specific primers for XBP-1. Polymerase chain reaction (PCR) products were separated by electrophoresis on 2% agarose gel and visualized by ethidium bromide staining. Levels of glyceraldehyde 3-phosphate dehydrogenase messenger RNA (mRNA) were used for adjustment of the mRNA concentrations. For quantitative real-time PCR analysis, the cDNA and appropriate primers were mixed with 2x ABsolute qRT-PCR SYBR green ROX reagent (Abgene, Hamburg, Germany). The following primers were used: CHOP forward, 5′ GAAACGGAAACAGAGTGGTCATTCCCC 3′, reverse, 5′ GTGGGATTGAGGGTCACATCATTGGCA 3′; mdr1 forward, 5′ GCTCAGACAGGATGTGAGTTGG 3′, reverse, 5′ TAGCCCCTTTAACTTGAGCAGC 3′. Primer sequences for IκBa, BiP, Bcl-2, Bax, β5-subunit, and β-actin were retrieved froman online database (Primer Bank, Harvard.edu). Real-time PCR was performed in triplicates in an Applied Biosystems 7300 real-time PCR system (Applied Biosystems, Darmstadt, Germany).
Sandwich ELISA was performed to quantify total IgG in the supernatant of JK-6L cells (2 x 106) treated with 10 nM bortezomib and/or 70 µM verapamil for 16 hours. Diluted supernatants and standard sera ( Jackson Immunoresearch Laboratories, Inc) were incubated on IgG-coated Maxisorp microplates (Thermo Fisher Scientific, Roskilde, Denmark), and bound IgG was detected with horseradish peroxidase-conjugated IgG. O-phenylenediamine dihydrochloride (Sigma-Aldrich) was used as substrate. Optical density was measured at 495 nm in a SpectraMax 190 ELISA reader (Molecular Devices, Ismaning, Germany).
Cells (2 x 106) treated with bortezomib and/or verapamil were fixed in 2.5% glutaraldehyde in 0.2 M phosphate buffer (pH 7.4), postfixed in 2% buffered osmiumtetroxide, dehydrated in graded alcohol concentrations, and embedded in epoxy resin according to standard protocols. For orientation, 1-µm semithin sections were stained with toluidine blue. Ultrathin sections were stained with uranyl acetate and lead citrate and examined using a transmission electron microscope (EM 906E; Carl Zeiss NTS GmbH, Oberkochen, Germany).
We first assessed the effect of bortezomib and verapamil alone and in combination on the viability of myeloma cell lines. The concentrations necessary for efficient depletion of myeloma cells were determined by performing titration curves. On the basis of this titration, we used 70 µM verapamil and a minimum of 10 nM bortezomib in the following studies (Figure 1A). Importantly, the combination of bortezomib and verapamil markedly declined the viability of the JK-6L, RPMI 8226, and ARH-77 cell lines after 16 hours of culture (Figure 1B). Bortezomib reduced the viability of all used myeloma cell lines, whereas verapamil alone exhibited just a slight cytotoxic effect on JK-6L cells but none on RPMI 8226 and ARH-77 cells. Compared with the other cell lines, JK-6L cells were more sensitive toward bortezomib and verapamil treatment, most likely because of their higher immunoglobulin chain expression as analyzed by intracellular flow cytometry and secretion assays (data not shown), consistent with previous results [9,10]. Most of the following analyses were performed with the JK-6L cell line because it is especially sensitive against bortezomib and verapamil.
Next we analyzed the induction of cell death on bortezomib and verapamil treatment. As displayed in Figure 1C, the combination of bortezomib and verapamil markedly diminished the percentage of annexin V/propidiumiodide double-negative viable myeloma cells while increasing apoptosis and necrosis, especially in JK-6L cells. At earlier time points, we could detect almost exclusively apoptotic cells and not necrotic ones (data not shown). Hence, combination of bortezomib and verapamil may induce predominantly apoptotic cell death, and the annexin V/propidium iodide-positive cells represent late apoptotic rather than primary necrotic cells. In contrast to verapamil alone, bortezomib, and even more when combined with verapamil, strongly induced activation of caspase 3/7 (Figure 1D). To evaluate the functional importance of bortezomib- and verapamil-induced caspase activation for apoptosis, we examined the influence of the pan-caspase inhibitor zVAD-FMK. Figure 1E shows that zVAD-FMK significantly interfered with cell death triggered either by the combination of both inhibitors or by bortezomib alone. Thus, bortezomib-induced apoptotic cell death was significantly enhanced in the presence of verapamil.
To elucidate the mechanisms by which verapamil enhances the antimyeloma effect of bortezomib, we first analyzed the proteasomal activity of the JK-6L cells during the treatment. The chymotrypsin-like activity of the 20S proteasome was completely blocked by bortezomib, whereas it was not affected by verapamil alone. Also, the combination of bortezomib and verapamil did not further decrease the proteasomal activity (Figure 2A). The concentration of the subunit β5 accounting for the chymotrypsin-like activity of the constitutive 26S proteasome moderately decreased after incubating with bortezomib (Figure 2B). Beside that, the immunoproteasomal subunit LMP7 that corresponds to β5 was not altered after treatment with the inhibitors (Figure 2B). Verapamil itself exerts no major effect on proteasomal activity.
The transcription factor NF-κB was shown to be constitutively active in myeloma cells and to be essential for their survival . To quantify the NF-κB DNA-binding activity, nuclear extracts from JK-6L cells treated with bortezomib and/or verapamil were analyzed by EMSA (Figure 2C). Bortezomib together with verapamil markedly decreased the NF-κB DNA-binding activity in JK-6L myeloma cells, whereas bortezomib or verapamil alone rather somewhat increased the constitutive NF-κB activity after 8 hours (Figure 2C). Incubation of JK-6L cells with an IKK inhibitor combined with verapamil resulted in a stronger reduction of the cell viability than in cells treated with verapamil only, suggesting that this specific NF-κB blockade may account for verapamil-induced synergistic cytotoxicity (Figure 2D).
We also measured the IκBα mRNA expression, which is strongly induced by NF-κB activation and therefore reflects transcriptional activity of NF-κB . Consistent with NF-κBDNA-binding activity, the relative IκBa mRNA concentrations dropped on the combination treatment (Figure 2D), indicating that NF-κB transcriptional activity was decreased. Thus, the blockade of the survival factor NF-κB on combination of bortezomib with verapamil might be crucial for enhancement of cell death.
To investigate whether verapamil together with bortezomib induced terminal UPR signals, we analyzed prominent UPR components in lysates of JK-6L cells. Enhanced chaperone production is a hallmark of UPR activation. Induction of the ER resident chaperone µ-heavy chain binding protein (BiP) is observed in bortezomib-treated cells. However, there was no additional increase of BiP expression in the presence of verapamil (Figure 3A). GRP94 was slightly upregulated on combination therapy compared with monotherapy. Hsp70, however, was strongly induced by bortezomib and its expression was further augmented in combination with verapamil (Figure 3A).
The transmembrane factor ATF6 induces transcription of several genes including chaperones, XBP-1, and CEB/P homologous protein (CHOP) in response to ER stress . After bortezomib and/or verapamil treatment, ATF6 was activated as shown by reduced expression of its inactive 90-kDa precursor form (Figure 3A). We also observed a slight induction of XBP-1 mainly triggered by bortezomib. The spliced XBP-1 protein was present in bortezomib-treated cells and synergistically increased on bortezomib and verapamil application (Figure 3A). Splicing of XBP-1 mRNA is induced by IRE1a to generate a transcript encoding an XBP-1 protein. The splicing results in the translation of a larger XBP-1 protein . In contrast to protein expression, analysis of XBP-1 mRNA revealed a strong activation of XBP-1 evidenced by the generation of the two splicing products on treatment with bortezomib plus verapamil (Figure 3B). The proapoptotic factor CHOP is markedly induced in cells incubated with bortezomib plus verapamil (Figure 3A), suggesting that verapamil strengthened the effect of the proteasome inhibitor.
As shown in Figure 3A, IRE1 expression itself was upregulated after bortezomib/verapamil treatment compared with monotherapy, indicating that verapamil enhanced the effect of bortezomib. On ER stress, IRE1 stimulates p38, causing activation of JNK. The p38 mitogen-activated protein kinase was highly activated evidenced by a strong expression of phosphorylated p38 (Figure 3A). Bortezomib/verapamil treatment eventually induced active JNK2 proposed to promote cell death (Figure 3A). These results imply that the ATF6/ IRE1 signaling plays an important role in the bortezomib/verapamil-mediated ER stress response.
In cells with irrecoverable levels of ER stress, the IRE1 pathway can promote apoptosis by interaction with Bcl-2 family members . The proapoptotic factor Bax is strongly expressed in double-treated JK-6L cells, whereas Bcl-2 was not markedly altered, if at all there was a slight increase on treatment with either bortezomib or verapamil alone (Figure 3A). The BH3-only protein Bim, a crucial effector of the IRE1-JNK pathway , was also induced by bortezomib together with verapamil. Induction of the key UPR molecules BiP, CHOP, Bim, and Bax could be confirmed in the myeloma cell lines ARH-77 and RPMI 8226 (Figure 3A).
PERK phosphorylates and inactivates the eukaryotic elongation factor eIF2α reducing the protein load through translational attenuation. Active p-eIF2α increases expression of the transcription factor ATF4, which induces UPR target genes . Western blot analyses revealed increasing PERK phosphorylation in bortezomib-, verapamil-, and bortezomib/verapamil-treated cells (Figure 3A). We also detected highly activated eIF2α in samples incubated with bortezomib plus verapamil (Figure 3A). At early time points, ATF4 was only induced in cells treated with verapamil after 16 hours; however, ATF4 expression was augmented on all different treatment regimens (Figure 3A). These data suggest a possible role for the PERK pathway in bortezomib/verapamil-mediated cellular effects, most notably for the calcium channel blocker verapamil.
To confirm induction of the UPR target genes on the transcriptional level, we performed quantitative reverse transcription-PCR. Here, we focused on UPR genes that are mainly involved in mediating survival and apoptosis. In contrast to BiP protein synthesis, we observed a stronger increase of BiP mRNA in cells treated with the drug combination (Figure 3C). Consistent with protein expression, analysis of CHOP mRNA revealed a more than 40-fold higher expression in response to bortezomib plus verapamil compared with untreated cells (Figure 3C). Bcl-2 mRNA was slightly downregulated, whereas Bax was increased by adding bortezomib plus verapamil (Figure 3C). Hence, bortezomib-triggered transcriptional activation of the terminal UPR was enhanced by combination with verapamil.
JNK activation can promote apoptosis in cells with overwhelming stress through activation of caspases . Hence, we investigated whether the inhibitors bortezomib and verapamil lead to caspase activation. Figure 3A demonstrates that after treatment with the drug combination, procaspase 9 virtually disappeared, indicating caspase 9 activation; verapamil alone had no effect on caspase 9 activation. Activation of the procaspase 9 was also observed in the ARH-77 and RPMI 8226 myeloma cell lines (Figure 3A). Likewise, a luminescence-based assay revealed that caspase 3/7 activation is more pronounced on the combination of bortezomib and verapamil than after monotherapy (Figure 1D). Further, bortezomib induced PARP activation as evidenced by detection of the cleavage product. The cleavage band was induced by bortezomib and even stronger in combination with verapamil (Figure 3A). Together, the bortezomib plus verapamil treatment resulted in caspase activation followed by PARP cleavage, finally leading to cell death.
To be degraded by the 26S proteasome, misfolded proteins have to get ubiquitinylated. On proteasome inhibition, these proteins can accumulate as insoluble intracellular aggregates up to cytotoxic levels . As depicted in Figure 4A, both bortezomib and verapamil induced accumulation of ubiquitinylated proteins after 16 hours in the soluble and insoluble fractions; also, after 8 hours, we observed ubiquitinylation of soluble but not of insoluble proteins (Figure W1). There was no marked further increase of ubiquitinylated proteins in the soluble fractions after combination treatment, also reflected by the pool of free ubiquitin. However, the amount of ubiquitinylated proteins in detergent-insoluble fractions was much higher in the presence of bortezomib plus verapamil, indicating an increased formation of intracellular protein aggregates.
Further, verapamil as well as bortezomib leads to the reduction of IgG in the supernatant of JK-6L cells as observed by ELISA andWestern blot analysis. Intracellular IgG was upregulated by adding verapamil (Figure 4B). Hence, bortezomib reduced IgG secretion presumably by UPR-mediated blockade of protein synthesis; whereas verapamil may interfere with IgG secretion (Figure 4C). We exclude the induction of apoptosis because IgG expression was also reduced at earlier time points (4 hours; Figure 4B).
To assess subcellular alterations induced by bortezomib and verapamil, we performed transmission electron microscopy of JK-6L cells. Bortezomib treatment resulted in ER dilation after 8 hours, most likely induced through accumulation of misfolded proteins, which even further increased after 16 hours (Figure 5A). Importantly, we observed vesicle-like intracellular structures after verapamil administration, suggesting that calcium channel blockade induced autophagy in the myeloma cells (Figure 5A). Visible signs of ER stress evidenced by ER expansion together with the appearance of protein deposits were detected 8 hours after combined treatment with bortezomib plus verapamil (Figure 5A). Aside increased autophagosomes, a vast swelling of the mitochondria was also observed after incubation with verapamil (Figure 5A). Consistent with the viability and UPR data, we observed a massive ER expansion along with vacuolization accompanied by morphologic features of apoptosis after 16 hours of incubation with this drug combination (Figure 5A).
To corroborate the morphologic evidence of verapamil-induced autophagy in myeloma cells, we analyzed the autophagosome formation using the microtubule-associated protein 1 light chain 3 (LC3; Figure 5B). LC3 is processed posttranslationally into LC3-I, then converted to LC3-II that is specifically associated with autophagoso memembranes . LC3-II concentrations, correlating with autophagosome numbers, were increased by verapamil and even further in combination with bortezomib, suggesting enhanced autophagy (Figure 5B).
Changes in mitochondrial structure may correlate with dysfunction of the mitochondria. We therefore determined themitochondrial membrane potential: The percentage of DiOC6-positive cells decreased in the presence of the drugs (Figure 5C), indicating that both bortezomib and verapamil induced disruption of the mitochondrial membrane potential. However, the loss of the mitochondrial potential was more pronounced in bortezomib-treated than in verapamil-treated cells.
Importantly, in response to verapamil treatment, we observed a vast and significant accumulation of ROS, which was reduced by the antioxidant NAC (Figure 5D), possibly linking increased ROS production to depolarization of the inner mitochondrial membrane.
As verapamil diminished expression of P-gp in resistant leukemic cells , we determined the effect of verapamil and bortezomib on P-gp expression on JK-6L cells. Flow cytometric analyses revealed decreased expression of P-gp on the JK-6L cells when treated with verapamil (Figure 6A). There was a slight effect of bortezomib alone but no synergistic effect together with verapamil on P-gp expression (Figure 6A). MDR1 mRNA levels significantly declined on verapamil monotherapy (Figure 6B). Hence, we conclude that P-gp expression on JK-6L was predominantly inhibited by verapamil.
Bortezomib exhibits antitumor activities by inducing cell death, but long-term administration often leads to secondary resistance of myeloma cells . To overcome cellular resistance and to enhance bortezomib's selectivity for myeloma cells, we searched for agents that synergistically increase the antitumor effect of bortezomib. The calcium channel blocker verapamil represented a promising candidate because it was shown to reduce MDR1-mediated drug resistance in leukemia cells . Our present study revealed that verapamil enhanced the antimyeloma effect of bortezomib leading to increased cell death in myeloma cells. It remains to be investigated if such high plasma levels of verapamil required to exert synergistic effects can be achieved in vivo without serious toxicity. ROS production and autophagy-like processes induced by verapamil might contribute to elevation of proapoptotic stress signals, strongly activating the terminal UPR in bortezomib-treated cells.
One important mechanism of bortezomib is the UPR induction that can trigger tumor cell death due to prolonged ER stress. This stress response can be induced through accumulation of misfolded immunoglobulin chains/DRiPs in the ER caused by proteasome inhibition [9,28]. Our results demonstrated that verapamil is able to enhance proapoptotic stress signals triggered by bortezomib. Depending on the UPR branch, the calcium channel inhibitor verapamil alone had only slight effects on UPR induction compared with bortezomib or the combination therapy. Also, verapamil displayed no direct effect on the 26S proteasome activity, suggesting that the ER-associated degradation is not blocked by verapamil. In contrast to our findings, Fekete et al.  observed an inhibition of proteasomal activity by verapamil in human epitheloid carcinoma cells; possibly, this discrepancy is due to the different reactions of individual cell species.
The expression of the chaperone BiP did not markedly change between treated cells, likely because of the relatively high basal levels of this protein often found in myeloma cells ; the BiP mRNA, however, was strongly induced. Likewise, an increase in cellular BiP mRNA does not necessarily lead to increased synthesis of BiP, and protein levels can remain constant . A need for more folding capacity was also evidenced by an induction of the ER resident chaperone GRP94 after bortezomib/verapamil treatment. Hsp70, the cytosolic homolog of BiP, was markedly increased in the presence of bortezomib and verapamil, indicating activation of a mitochondrial-like stress . One function of chaperones is to protect cells against ER stress-induced apoptosis, but in case of prolonged or overwhelming stress signals, the UPR ends in apoptosis. In our approach, the IRE1-mediated pathway was profoundly activated. By recruiting TRAF2, IRE1 leads to increased levels of phosphorylated p38 mitogen-activated protein kinase causing JNK activation  that can foster apoptosis through activation of caspases under severe stress induction . On treatment with the combination of bortezomib and verapamil, p38 was strongly activated followed by JNK activation, suggesting that the IRE1-p38-JNK signaling plays an important role in bortezomib/verapamil-mediated cell death.
Also, CHOP expression is enhanced by verapamil when combined with bortezomib. CHOP, which is associated with the terminal UPR and apoptosis, was shown to downregulate Bcl-2 . A recent publication demonstrated that CHOP itself is not essential for physiologic cell death in mouse B and plasma cells . Expression of the proapoptotic Bax increased in response to combination treatment, whereas there was no significant difference in Bcl-2 levels. Importantly, Bax and Bcl-2 can also activate IRE1 , thereby providing a connection between UPR and apoptosis pathways. The role of CHOP in human myeloma cells is not yet resolved.
The role of the PERK pathway in cell death regulation remains to be elucidated. Its activation initiates inhibition of translation via phosphorylation of eIF2α to protect the cell from protein overload. The PERK pathway was highly activated on treatment with verapamil and the combination of bortezomib and verapamil as shown by up-regulation of p-eIF2α, suggesting attenuation of mRNA translation. In addition, PERK-mediated eIF2α phosphorylation contributes to transcriptional activation of ATF4, which has been shown to stimulate expression of autophagy genes  and to induceCHOP. Also, long-lasting suppression of protein synthesis is not compatible with cell survival and can induce autophagy known to require PERK and eIF2α phosphorylation. Using transmission electron microscopy, we detected a vast expansion of mitochondria and formation of cytosolic vesicles on verapamil exposure. These observations suggest that verapamil induces autophagy in the myeloma cells. Autophagy, an intracellular degradation system of cytoplasmic contents and organelles, is required for normal turnover of cellular components during starvation in eukaryotic cells. The autophagosomes, with two or more membranes enclosed vesicles, engulf various cellular constituents that fuse with lysosomes for degradation and recycling . Autophagy is also characterized by dilation of mitochondria that we observed in treated JK-6L cells. Besides promoting cell survival, autophagy can trigger caspase-independent cell death . Likewise, we observed no caspase activation by verapamil. A very recent study by Williams et al.  revealed that verapamil stimulates autophagy by reducing the calcium influx. Notably, verapamil diminished the intracellular calcium level also in the myeloma cell lines (Figure W2). Changes in calcium homeostasis may foster ER stress triggering autophagy by the ER-activated autophagy pathway. The latter could be mediated by limited UPR signals involving PERK and/or IRE1 as well as UPR-independent mechanism such as calcium leakage (JNK-AKT/mTOR signaling) . Here, we demonstrate that verapamil enhanced IRE1- and PERK-mediated pathways, indicating that verapamil further activates the UPR and might trigger an autophagic, caspase-independent cell death. In the presence of the autophagy inhibitor 3-methyladenine, we observed a higher viability in verapamil-treated cells compared with cells cultured with verapamil alone (Figure W3), suggesting that verapamil-induced autophagy might contribute to cytotoxicity. Verapamil treatment also increased ROS production, leading to oxidative stress that was also shown to activate autophagy and mitochondrial dysfunction . Mitochondrial ROS can further increase calcium release from the ER, thereby causing protein misfolding. ROS production and protein misfolding together activate calcium-dependent kinases such as JNK eventually leading to cell death . Thus, we showed that verapamil supports bortezomib's ability to promote proapoptotic UPR signals such as CHOP induction and JNK activation. Bortezomib itself leads usually to a slight induction of NF-κB, presumably due to ER stress and consecutive UPR activation, which is not completely blocked by bortezomib-mediated inhibition of IκB degradation, as recently described by Hideshima et al. . Remarkably, combination treatment with bortezomib and verapamil markedly inhibited the transcription factor NF-κB, a fact which might critically contribute to cell death induction in JK-6L cells. Moreover, autophagy-induced protein degradation could further decrease NF-κB activity .
Interference with MDR represents a further potential mechanism to increase cytotoxicity by the combination of the two drugs. Verapamil functions as an inhibitor of drug efflux pump proteins such as the P-gp . Many tumor cell lines overexpress drug efflux pumps, limiting the effectiveness of cytotoxic drugs. A recent study showed that efflux pump inhibitors enhanced the effect of bortezomib on drug-resistant Ewing tumors . Using P-gp-positive and P-gp-negative cells, Rumpold et al.  provided evidence that bortezomib acts as aMDR1 substrate. We observed lower P-gp expression on JK-6L cells accompanied by impaired efflux of daunorubicin in cells treated with bortezomib/verapamil (data not shown), speculating that inhibition of P-gp might result in a reduced bortezomib efflux, leading to increasing amounts of intracellular bortezomib concentrations augmenting cell stress.
Approximately 60% of nonresistant JK-6L myeloma cells express P-gp on their surface. As expected, long-term culturing in the presence of bortezomib rendered these cells resistant to bortezomib, resulting in an increased P-gp expression (data not shown). These resistant JK-6L cells could be partially sensitized against bortezomib by verapamil (Figure W4). However, we assume that the increased sensitivity mainly results from UPR/stress-induced apoptosis and simultaneous inhibition of the antiapoptotic factor NF-κB, whereas the subtle MDR-mediated effects may play only a minor role. Further investigations are required to elucidate the exact mechanisms how verapamil triggers these signals.
In summary, our data provide a biochemical base for a new potential combination therapy of bortezomib together with verapamil, which should markedly enhance the killing of myeloma cells. Moreover, the efficiency of bortezomib treatment can be enhanced by the combination with verapamil. In addition, this drug combination may allow decreasing the bortezomib dose thereby reducing adverse effects. Animal experiments exploring the use of this combination treatment in myeloma models are underway.
The authors thank Daniela Graef for technical assistance and Till T. Wissniowski for discussion (Department of Internal Medicine 1, University Hospital of Erlangen, Germany). The JK-6L cell line was kindly provided byMartin Gramatzki (Division of Stem Cell Transplantation, 2nd Department, University of Kiel, Germany). The RPMI 8226 and ARH-77 cell lines were a kind gift from the laboratory of Hans-Martin Jäck (Division of Molecular Immunology, Internal Medicine 3, University Erlangen, Germany).
1This work was supported by the Interdisciplinary Center for Clinical Research project no. N2; German Research Society (DFG) Collaborative Research Centers SFB 643 project B3 (R.V.) and FOR832 (VO 673/3-1), the ELAN fond (08.03.12.1) of the University Erlangen-Nuremberg, Germany, and the Doktor Robert Pfleger Foundation, Bamberg, Germany, and the Bavarian ImmunotheraphyNetwork (Bay ImmuNet) to R.V.