Effect of autophagy on multiple myeloma cell viability

Recent work suggests it may be possible to exploit therapeutically the fact that multiple myeloma (MM) cells synthesize huge amounts of immunoglobulin (Ig). This invariably results in the presence of a significant amount of endoplasmic reticulum (ER)–localized unfolded or malfolded protein that is potentially toxic to the plasma cell. Indeed, the MM cell has exploited several molecular pathways to deal with this toxic protein and these are potential therapeutic targets. They include the proteasome (reviewed in ref. 1), up-regulated expression of heat shock protein chaperones (2), activation of the unfolded protein response (UPR) pathway (3–5), and aggresome formation (6, 7). The remarkable efficacy of the proteasome inhibitor bortezomib in MM cells and patients (8, 9) and the correlation of its in vitro effectiveness with amount of Ig expressed (3, 10) support the notion of targeting pathways that degrade or refold toxic malfolded proteins. Additional studies on the antimyeloma effects of heat shock protein inhibitors (2) and drugs that prevent aggresome formation (6, 7) are also consistent with this hypothesis.

The induction of autophagy is an additional potential pathway that might protect MM cells against toxic malfolded proteins and could be targeted. Autophagy is a degradation process of proteins and organelles in which double-membrane vesicles, termed autophagosomes, sequester cytosolic proteins and/or organelles (reviewed in ref. 11). Following vesicle fusion with lysosomes, the contents are degraded by hydrolases (12–14). Autophagy is essential for normal cell function and occurs at a basal level in all cells. Increased autophagy can be induced, most prominently by nutrient starvation or mammalian target of rapamycin (mTOR) inhibition, which shuts down protein translation. In those cases, following hydrolysis of autophagosomal cargo, the resulting macromolecules are presumably released from the vacuole for reuse to ameliorate the starvation effect. However, recent studies suggest that autophagy can also function to remove toxic aggregated proteins such as mutant huntingtin and α-synuclein (15–17). Furthermore, ER stress inducer stimulation of the UPR can up-regulate expression of autophagy-regulatory genes (18) and can activate autophagy via the downstream activation of IRE-1/c-Jun NH₂-terminal kinase (JNK) pathway (19) or the PERK/eIF2α pathway (20). These studies suggest that autophagy may be an important additional process by which the malignant plasma cell could protect itself from toxic malfolded Ig. We, thus, tested this notion in MM cell lines and primary specimens. A significant degree of autophagy was present in the basal state of these MM cells, and the proteasome inhibitor bortezomib, the ER stress inducer thapsigargin, and the mTOR inhibitor rapamycin were all capable of further increasing autophagy. Inhibiting autophagy with inhibitors and by silencing beclin-1, an important protein required for autophagy, resulted in a toxic effect on MM cells. Although we anticipated that blocking autophagy with inhibitors would synergize with the antimyeloma effects of bortezomib, we found an antagonistic effect. In contrast, inhibiting autophagy synergized with thapsigargin for enhanced cytotoxicity and apoptosis. The data support the notion that inhibiting autophagy is a potential therapeutic maneuver in myeloma but that the complex interaction between autophagy and proteasome inhibition needs further study before considering combination therapy with autophagy inhibitors plus bortezomib.

Approval for these studies was obtained from the Greater Los Angeles Veterans Administration Healthcare System institutional review board. Informed consent was provided according to the Declaration of Helsinki.

OPM-2, 8226, and U266 cell lines were purchased from the American Type Culture Collection and maintained in culture as previously described (21, 22). Primary malignant cells were isolated from bone marrow by positive selection for CD38 (>98% pure) as previously described (23).

Survival assays were done as previously described (24). Surviving cells were enumerated by trypan blue exclusion. In individual experiments, groups were run in quadruplicate and the means of the replicates were used to calculate the percentage inhibition of viable recovery compared with untreated groups. The data are presented as percentage of control, means of three or more individual experiments. Apoptosis was identified by flow cytometric staining for expression of activated caspase-3 (BD Biosciences) as previously described (24).

Protein expression was assayed as previously described (24).

The green fluorescent protein (GFP)-LC3 coding sequence was isolated from pEGFP-LC3 plasmid (kind gift of Dr. Hong-Gang Wang, Moffitt Cancer Center, Tampa, FL) by using primers SpeIGFPLC3 (5′-TTGCACTAGTATGGTGAGCAAGGGCGAGGAG-3′) and XhoIGFPLC3 (5′-GCTACTCGAGTTACACAGCCAGTGCTGTCCC-3′). SpeIGFPLC3XhoI PCR product was digested with SpeI and XhoI. Insert from pLenti6/V5IgGL vector was removed by SpeI/XhoI digestion and replaced with SpeI/XhoI-digested GFPLC3 to yield pLenti6GFPLC3. Lentivirus was produced according to Invitrogen's protocols. Briefly, pLenti6GFPLC3 was cotransfected with ViraPower DNA (Invitrogen) into 293 ft cells, and after 48 h, viral supernatant was collected. U266 and 8226 cells (2.5 × 10⁵) were infected with 1 mL of viral supernatant overnight and then selected with 2.5 μg/mL blasticidin (Invitrogen) to generate stable cell lines. GFP-LC3 expression was confirmed by Western analysis.

To knock down beclin-1, U266 cells (2 × 10⁶) were transfected with 1 μmol/L of Beclin SMARTpool small interfering RNA (siRNA; three separate siRNAs) or scramble duplex (purchased from Dharmacon) using the Amaxa Nucleofector II electroporation device and X005 program. Cells were then cultured in RPMI 1640 complete growth medium. For beclin protein knockdown analysis, cell lysates were collected after transfection, electrophoresed, and immunoblotted with anti-beclin antibody (Novus Biologicals).

Autophagic area was assessed as previously described (19). Briefly, glass coverslips were coated with 0.01% poly-l-lysine (Sigma) in 24-well tissue culture plates overnight at 4°C and rinsed with PBS just before use. U266GFPLC3 or 8226GFPLC3 cells (20,000) were plated onto poly-l-lysine–coated coverslips in 500 μL of growth medium. After 2 h in culture, adherent cells were fixed with 3.7% formaldehyde for 10 min and rinsed with PBST (Tween 20). Cell membranes were permeabilized by treating with PBS/0.2% Triton X-100 for 5 min and rinsing with PBST. Cells were then blocked with PBST/10% normal goat serum, and anti-GFP antibody (JL-8 from Clontech) in blocking buffer was added at 4 μg/mL. After three washes with PBST, anti-mouse Alexa Fluor 488 secondary antibody (Molecular Probes) was added. Cells were additionally washed with PBST and coverslips were mounted on microscope slides in Prolong Gold antifade reagent (Invitrogen). Cells were visualized with a Leica Leitz DMRBE microscope using a 40× objective lens (Leica PL Fluotor) and fluorescent and corresponding bright-field images were captured (20 random fields/group) with a Hamamatsu C4742-95 camera and OpenLab program 3.1.5. Assays were done in blinded fashion. The images were transferred to NIH Image software and each cell was analyzed for total cell area (denominator) by outlining the perimeter of a bright-field image of the cell and quantifying the area inside. In addition, the corresponding fluorescent image of each cell was quantified for GFP-positive punctate area by density slice (numerator). For each cell, autophagic area is calculated as total punctate area divided by cell area (20 fields/group for each experiment). Data are presented as mean ± SD autophagic area of three separate experiments.

After treatments, MM cells were fixed in phosphate buffer (pH 7.4) containing 2.5% glutaraldehyde and 2% paraformaldehyde at room temperature for 60 min. Cells were postfixed in 1% OsO₄ at room temperature for 60 min, dehydrated through graded ethanol solutions, and embedded in Quetol 812 (Nissin EM Co.). Sections were stained with uranyl acetate and lead citrate and examined with a transmission electron microscope. Autophagosomes were identified and quantitatively assessed as previously described (19). Data are presented as number of autophagosomes per cell, mean ± SD, n = 20.

The effect of drug combinations on cytotoxicity was assessed by the median-effect method using CalcuSyn software, version 1.1.1 (Biosoft) as previously described (25). Combination index (CI) values were calculated using the most conservative assumption of mutually nonexclusive drug interactions. CI values were calculated from median results of cytotoxicity assays, which were done in quadruplicate. CI values significantly >1 indicate drug antagonism. Linear regression correlation coefficients of the median-effect plots were required to be >0.90 to show that the effects of the drugs follow the law of mass action, which is required for a median-effect analysis.

The t test was used to determine significance of differences between groups.

We were interested to see if induction of autophagy in MM cells occurs following treatment with thapsigargin, bortezomib, or rapamycin. Thapsigargin induces ER stress secondary to its inhibition of the Ca²⁺-ATPase in the ER (26), which blocks sequestration of calcium by the ER, causing accumulation of unfolded proteins. As stated in Introduction, previous literature supports the notion that ER stress can induce autophagy. mTOR inhibitors such as rapamycin are well known to induce autophagy (27) and mTOR inhibitors have shown some potential antimyeloma efficacy in preclinical studies (28). Bortezomib is a clinically useful proteasome inhibitor that can also stimulate several components of the UPR (3, 4) and, thus, may also stimulate autophagy. We decided initially to study the U266, 8226, and OPM-2 MM cell lines because they are known to synthesize Ig molecules, which could sensitize them to autophagy inhibition. To assay autophagy in these MM cell lines, we attempted to stably express the LC3-GFP fusion protein in U266, 8226, and OPM-2 MM cells by lentiviral infection. We were successful in the U266 and 8226 cell lines. Microtubule-associated protein-1 LC3 is a component of mammalian autophagosomes, and thus, the GFP-LC3 fusion protein has been used as a reliable marker for their presence (29, 30). A punctate pattern of LC3 localization in the cell is characteristic of autophagosome formation and represents the accumulation of a membrane-bound form of LC3 onto autophagic vacuoles (30).

The U266 and 8226 cell lines are variably sensitive to cell death or cytostasis induced by thapsigargin, bortezomib, and rapamycin. Stable expression of LC3-GFP had no effect on the sensitivity of these cell lines to any of these three drugs (data not shown). Fluorescence analysis for LC3-GFP localization showed a significant amount of punctate staining in the nontreated MM cell lines, indicative of basal autophagosome formation. Constitutive expression of punctate LC3-GFP was seen in a very high percentage of 8226 and U266 cells (85% and 65%, respectively). It was, thus, difficult to show a statistically significant increase in punctate staining from drug treatment by simply enumerating positive versus negative cells. However, analysis of the autophagic area/cell (as described in Materials and Methods) clearly showed drug-stimulated enhancement of autophagy. As shown in Fig. 1A (black columns), approximately 2.5- to 3-fold increases were shown in both cell lines exposed to the drugs. The 8226 cell line treated with thapsigargin could not be analyzed as drug treatment prevented adherence of the cell line to coverslips, which was required for analysis.

Because measurement of “autophagic area” cannot distinguish between enlargement of autophagosomes versus an increase in numbers, we also analyzed the data for number of GFP-LC3 punctate signals per cell. These data also support a drug-induced increase in autophagosomes (Fig. 1A, white columns). The number of fluorescent autophagosomes in bortezomib- or rapamycin-treated 8226 cells (5.86 ± 0.47 and 4.39 ± 0.59, mean ± SE, n = 45) was significantly greater (P < 0.05) than DMSO-treated cells (2.58 ± 0.45). Similar data for thapsigargin, bortezomib, and rapamycin in treated U266 cells (4.57 ± 0.5, 4.14 ± 0.52, and 5.78 ± 0.5, respectively, n = 50) showed a significant increase (P < 0.05) in autophagosomes versus DMSO-treated cells (2.79 ± 0.42). Figure 1B shows representative assays showing the increase in punctate LC3-GFP detection in drug-treated U266 cells. A more detailed dose-dependent analysis of bortezomib-induced autophagy in U266 cells is shown in Fig. 1C, showing significant increases (P < 0.05) in autophagic area (black columns) and number of autophagosomes/cell (white columns) at all concentrations of bortezomib.

The above data showed increases in the number and area of autophagosomes. However, these increases could result from either an induced block in later autophagy events that cause accumulation of autophagosomes or from a true increase in flux through the autophagy pathway. As described (31), when GFP-LC3 is delivered to the lysosome, the LC3 part of the fusion protein is sensitive to proteolytic degradation but the GFP portion is resistant. Thus, the appearance of free GFP on Western blots can be used to monitor lysis of the autophagosome membrane and degradation of its cargo. As shown in Fig. 1D, treatment of 8226 or U266 cell lines with bortezomib resulted in generation of free GFP, best seen at 10 nmol/L drug concentration after 14 and 24 hours of incubation with 8226 cells and after 14 hours of incubation with U266 cells. An additional marker of increased autophagic flux is diminished p62 expression, as p62 becomes incorporated into the completed autophagosome and is degraded in autolysosomes (31). As shown in Fig. 1E, p62 became down-regulated in a time- and concentration-dependent fashion during exposure of MM cells to bortezomib. These data confirm that the increase in LC3-GFP signals is due to a true stimulation of autophagic flux induced by bortezomib rather than a block in later autophagy events.

Further confirmation of autophagy induction was obtained by electron microscopy. Autophagy was assayed by blindly counting the number of identified autophagosomal vacuoles. These vacuoles could be easily identified as shown in Fig. 2A and B (shown by arrows, high-power view in Fig. 2B). All three of the drugs were capable of increasing autophagy in 8226 cells assessed by electron microscopy as shown in Fig. 2C. Thapsigargin markedly and significantly (P < 0.05) increased the number of identified autophagosomal vacuoles from a mean of 5 ± 1 per cell (mean ± SD of 20 cells analyzed) to a mean of 17 ± 3 per cell (examples in Fig. 2A). Bortezomib also significantly increased the number of vacuoles/cell (12 ± 2 per cell versus 5 ± 1, n = 20, P < 0.05) and rapamycin modestly increased the number (11 ± 2 versus 5 ± 1, n = 20, P < 0.05).

During autophagy, LC3 is processed from the cytosolic form, LC3-I, to the membrane-bound form, LC3-II, and this can be detected by immunoblot (30). Our immunoblot assay for expression of endogenous LC3 confirmed the ability of thapsigargin, rapamycin, and bortezomib to induce autophagy (data not shown for rapamycin). As shown in Fig. 2D for OPM-2 and U266 cell lines, there is a dose-dependent increase in expression of LC3-II induced by bortezomib in both cell lines. Figure 2E also shows the inability of dexamethasone at 10⁻⁶ mol/L and melphalan at 5 and 20 μmol/L to induce autophagy in 8226 cells (with positive induction by thapsigargin), although they decreased viable cell recovery by 50% (comparable with thapsigargin in this experiment) and induced significant apoptosis (data not shown) in the same cells. This rules out the possibility that autophagy induction in MM cells is a nonspecific response to any agent capable of inducing MM cell injury and/or death. Figure 2F shows the ability of thapsigargin and bortezomib at 2, 5, 10, 20, or 50 nmol/L to enhance autophagy (LC3-II expression) in primary myeloma cells. Rapamycin at 10 nmol/L had only a minimal effect in one patient. Thus, specific induction of autophagy in MM cell lines and primary samples was detected following exposure to a proteasome inhibitor (bortezomib) and an ER stress inducer (thapsigargin). A mTOR inhibitor (rapamycin) was also capable of modest induction in MM cell lines, but more studies are needed to determine whether this drug can stimulate autophagy in primary samples.

To test whether autophagy helps maintain viability in MM cells, we first used 3-methyladenine (3-MA) and chloroquine, known inhibitors of autophagy. 3-MA is a class III phosphatidylinositol 3-kinase inhibitor that prevents autophagy at the earliest stage of autophagosome formation (32) and was studied at concentrations (1, 5, or 10 mmol/L) known to inhibit autophagy in other cell models. As shown in Fig. 3A (left), 3-MA used at 1, 5, or 10 mmol/L induced a dose-dependent decrease in cell recovery (white columns) and viability (black columns) in all three cell lines (U266, 8226, and OPM-2 as noted in Fig. 3A). This cytoreductive effect roughly correlated with the ability of 3-MA to inhibit autophagy. Autophagy was monitored in U266 and 8226 cell lines by assessment of autophagosome area/cell as previously described in Fig. 1 and presented in Fig. 3A as percentage of control autophagy. In the OPM-2 line, which could not be transfected with LC3-GFP, effects on autophagy were monitored by immunoblot assay. As shown, a decreased LC3-II expression resulting from 5 and 10 mmol/L exposure to 3-MA identified inhibition of autophagy, which roughly correlated with inhibitory effects on cell recovery and viability.

Chloroquine, a weak base, concentrates in acidic vesicles such as lysosomes and disrupts vesicular acidification, a process that prevents fusion of autophagosomes with lysosomes, resulting in inhibition of autophagy (33, 34). By preventing autophagosome-lysosome fusion, chloroquine leads to accumulation of LC3-II in treated cells. Thus, immunoblot assays were used to assess chloroquine effects on autophagy (Fig. 3A, right), with an increase in LC3-II levels reflecting inhibition. As shown, chloroquine, used at 5, 10, 20, or 40 μmol/L, reduced MM cell recovery and viability in all three cell lines. This was most evident at the chloroquine concentration of 40 μmol/L. This correlated well with effects on LC3-II accumulation in OPM-2 cells. However, with U266 and 8226 cell lines, there was significant LC3-II accumulation at lower chloroquine concentrations, which had insignificant or only modest effects on cell survival.

A limited number of flow cytometric analyses for apoptosis, using an antibody to activated caspase-3, showed that induction of apoptosis occurred in these experiments with autophagy inhibitors (in Supplementary Fig. S1).¹

A similar loss of viability and recovery was seen in primary MM samples (Fig. 3B). Three primary samples were incubated with 0, 5, or 10 mmol/L 3-MA for 24 hours (white columns) or 48 hours (black columns). A significant inhibition of viable MM cell recovery was identified at the 10 mmol/L concentration (P < 0.05). In at least one of these primary samples, we could show induction of apoptosis (Supplementary Fig. S1).¹

To further identify autophagy as a possible survival-promoting mechanism in MM cells, we knocked down the beclin-1 gene with siRNA in the U266 cell line. Beclin is the mammalian homologue of ATG 6, which is found in a complex with class III phosphatidylinositol 3-kinase and which is required for autophagy (35, 36). Beclin-1 siRNA was introduced into these cells by electroporation (Amanxa technique), which, in our hands, results in >80% transduction efficiency. As shown in Fig. 4A, siRNA was successful in knocking down expression of beclin-1. This knockdown muted the increase in autophagic cell area induced by bortezomib or thapsigargin (Fig. 4B). It also modestly muted the down-regulated expression of p62 induced by bortezomib (Fig. 4C). Finally, the viable recovery of beclin knocked down U266 cells was significantly decreased relative to control cells electroporated with a scrambled sequence (Fig. 4D). These gene knockdown data support the above results that use the 3-MA and chloroquine chemical inhibitors.

The induction of autophagy during treatment with bortezomib may have therapeutic implications. Because proteasome inhibition might injure MM cells via an increase in malfolded Ig, if autophagy was an additional mechanism available to the plasma cell for dealing with these toxic molecules, concurrent autophagy inhibition might result in synergistic antimyeloma effects. Thus, U266, 8226, and OPM-2 cell lines were concurrently treated with bortezomib ± chloroquine. As shown in Fig. 5, bortezomib alone showed a dose-dependent effect, with 8226 being the most sensitive line. The sensitivity of the lines to chloroquine was similar to previously shown data in Fig. 3A with 40 μmol/L having the greatest effect. Interestingly, the combination of bortezomib with chloroquine was consistently antagonistic with the greatest antagonism seen with low concentrations of both drugs. The CIs for drug combinations are shown to the right of the survival data with CI values >1 reflecting antagonism. A similar antagonistic interaction was seen when low concentrations of both drugs were used sequentially (2 hours before treatment with chloroquine followed by bortezomib; Supplementary Fig. S2).¹ CI values were also indicative of antagonism when the 3-MA inhibitor was used concurrently with bortezomib. With the U266 cell line, 3-MA at 5 mmol/L combined with bortezomib at 5 nmol/L resulted in a CI of 3, whereas 3-MA at 10 mmol/L combined with bortezomib at 10 nmol/L resulted in a CI of 1.9.

A similar antagonism was seen when both drugs were used concurrently against three primary samples (Fig. 6A). Bortezomib, used alone, showed significant antimyeloma activity against the primary specimens when used at 5 or 10 nmol/L. Likewise, 3-MA, when used alone, showed modest cytotoxic effects in all three patients (approximately 20–40% loss of survival at 10 mmol/L 3-MA compared with nontreated controls). However, combining bortezomib with 3-MA resulted in viable recovery of MM cells in all three specimens, which was higher than calculated from simple addition of the cytotoxic effects of each drug used separately. For example, in patient 1, 5 nmol/L bortezomib induces remarkable cytotoxicity by itself, resulting in ∼90% loss of survival. Although 3-MA used alone at 5 and 10 mmol/L shows modest cytotoxic effects, combining them with bortezomib produces a significantly less toxic effect than achieved by bortezomib alone.

Further support for an antagonistic effect between autophagy inhibition and proteasome inhibitors was shown by combining beclin-1 siRNA with bortezomib (Fig. 6B). U266 cells electroporated with siRNA for beclin-1 show ∼25% loss of viability (versus control scrambled sequence), whereas 10 nmol/L bortezomib induces 50% cytotoxicity. However, combining beclin-1 knockdown with bortezomib results in only 32% cytotoxicity (68% of control viability).

In contrast to this antagonism, when the autophagy inhibitor 3-MA was combined with thapsigargin, synergistic cytotoxicity was shown. As shown in Fig. 6C, U266 cells treated with 5 and 10 mmol/L 3-MA by itself show a decrease of 20% and 30% viable recovery, respectively, whereas thapsigargin at 1 μmol/L induces a 15% loss of viable recovery. However, combining thapsigargin with 10 mmol/L 3-MA results in a 60% decrease in viable recovery, which is greater than the sum of each treatment used separately (15% decrease with thapsigargin + 30% decrease with 3-MA). The synergy is even more obvious when these treated cells were analyzed for apoptosis by flow cytometry for activated caspase-3 (Fig. 6C, bottom). As shown, thapsigargin (1 μmol/L) induced ∼10% apoptosis, whereas 3-MA used alone at 5 or 10 mmol/L induced 8% and 14% apoptosis, respectively. Combining thapsigargin with 3-MA at 5 and 10 mmol/L induced 35% and 55% apoptosis, confirming a synergistic interaction.

One potential explanation for an antagonistic interaction between bortezomib and autophagy inhibitors is that a component of bortezomib-induced MM cell death is actually mediated by an autophagic pathway. To provide some support for that possibility, we exposed MM cell lines to toxic concentrations of bortezomib in the presence of ZVAD, which would prevent caspase-mediated apoptosis. As shown in Fig. 6D for U266 cells, ZVAD used at 20 or 40 μmol/L prevented bortezomib-induced caspase-3 cleavage (48-hour incubations). Nevertheless, a significant amount of ZVAD-resistant MM cell death was still induced (Fig. 6D). ZVAD-resistant cell death could be type II autophagic death, which would explain the previously shown antagonism between bortezomib and autophagy inhibitors.

This study was prompted by previous literature, which suggested that targeting pathways that defend myeloma cells against the cytopathic effect of malfolded protein could be effective therapy. A huge amount of monoclonal Ig is synthesized by MM cells every day, invariably resulting in huge amounts of malfolded protein that must be cleared. Myeloma cells have developed efficient proteasome- and aggresome-dependent mechanisms for this process as well as a very efficient UPR cascade. The latter functions to temporarily decrease Ig synthesis and increase protein chaperone expression. Malfolded protein that overwhelms chaperone and proteasome capacity becomes aggregated and is gathered in aggresomes by a highly coordinated process. These aggresome structures may function as proteolysis centers specialized in degrading aggregated proteins or as collection centers for degradation by an autophagic pathway. Supporting the latter role for autophagy in degrading toxic aggregated proteins comes from studies (15–17) on degradation pathways for proteins implicated in neurologic disease such as α-synuclein (in Parkinson's disease; ref. 15) and huntingtin (in Huntington's disease; ref. 17). However, studies on a role for autophagy in protection of myeloma cell viability have not been previously done. Our studies show that basal autophagy protects MM cell viability and that autophagy is further enhanced by exposure of MM cells to ER stress inducers, proteasome inhibitors, and mTOR inhibitors. Unexpectedly, we also found that combining autophagy inhibitors with bortezomib resulted in antagonistic cytotoxic effects.

Autophagy was induced in MM cell lines and primary samples by exposure to rapamycin, a mTOR inhibitor, bortezomib, a clinically relevant proteasome inhibitor, and thapsigargin, a drug that induces ER stress by altering calcium metabolism. mTOR is a known inhibitor of autophagy, possibly working through its phosphorylation of ribosomal protein S6 (27, 37). Thus, we were not surprised to identify activation of autophagy in MM cells exposed to rapamycin. Recent work (19, 20, 38) indicates that stimulation of the UPR can activate autophagy. It is likely that this explains the induction of autophagy by thapsigargin. It is also possible that bortezomib induces autophagy by a similar pathway. The UPR consists of three signal cascades activated when the ER reaches a threshold amount of malfolded protein. Although Lee et al. (4) described the ability of proteasome inhibitors to suppress the activity of IRE1-α, an initiator of one of the three cascades, with subsequent impairment of XBP-1 splicing in MM cells, Obeng et al. (3) documented that bortezomib activated other components of the terminal UPR cascade, including PERK, the eIF2α kinase. As eIF2α phosphorylation can mediate an essential step for autophagy formation (20), bortezomib-induced PERK activation could explain its ability to enhance autophagy. An ER stress–induced activation of JNK is also possible as a mechanism by which autophagy is stimulated (19) and bortezomib has been previously shown to activate JNK (39). In data not shown, bortezomib treatment of our MM cell lines was effective in inducing eIF2α phosphorylation and JNK activation, supporting these hypotheses. A recent publication (40) shows that proteasome inhibitors activate autophagy in colon cancer cells by a JNK-dependent pathway, further supporting this notion.

Our data support the notion that autophagy protects MM cell viability. Two independent inhibitors of autophagy, 3-MA and chloroquine, with different structures and different mechanisms of action, induced myeloma cell death of cell lines and primary specimens in a dose-dependent fashion. In addition, inhibition of autophagy, mediated by beclin-1 knockdown, also was cytotoxic to the U266 MM cell line. These data are consistent with prior work that underscored autophagy as a viability protector in cells stressed by nutrient deficiency (33, 41) or by ER stress (19, 20, 42). The induction of MM cell death was apoptotic in nature (type I programmed cell death) as shown by staining for activated caspase-3. This apoptotic response is also consistent with the above prior work on the role of autophagy during nutrient deficiency or ER stress. It is possible that the basal ongoing ER stress in these MM cell lines that synthesize Ig is sufficient to sensitize them to enhanced apoptosis if autophagy is inhibited.

Because of these viability-promoting effects of autophagy, we anticipated that cotreatment of MM cells with bortezomib and autophagy inhibitors would result in synergistic cytotoxic effects. In fact, two recent studies presented in abstract form (43, 44) suggest that synergy could be achieved in MM cells. In contrast, however, we were surprised to find an antagonistic effect of combined treatment. This suggests either that some of the cytotoxic effect of bortezomib against MM cells is mediated via autophagy or that the proteasome inhibitors and autophagy inhibitors use similar or overlapping pathways for cytotoxicity such that pathway stimulation by one agent can cancel out stimulation by the other when both are used in combination. Some support for the former contention comes from experiments shown in Fig. 6D where significant bortezomib-induced death was independent of caspase cleavage and possibly mediated by type II autophagic cell death. At 20 and 40 μmol/L of ZVAD, caspase-3 cleavage induced by bortezomib was abrogated but a significant amount of cell death was still identified. Similar antagonistic results with combinations of proteasome inhibitors (MG132) and autophagy inhibitors (3-MA) were obtained by Yang et al. (45) where combined treatment of prostate cancer cells resulted in an inhibited cytotoxic response. Furthermore, the concept that excessive induction of autophagy could be a mechanism of MM cell death (instead of a protector against death) is supported by a recent article (46) in which an inhibitor of SCF^Skp2, which stabilized p21/p27, induced caspase-independent MM death, which was mediated by autophagy. However, when we combined an autophagy inhibitor with thapsigargin, we detected synergistic MM cell death. As noted in Figs. 1A and 2C, thapsigargin induced a greater degree of autophagy than bortezomib. Thus, it is more likely that it is the context of the enhanced autophagy (i.e., in the presence of proteasome inhibition or pure ER stress) rather than the degree of autophagy induction that determines its contribution to ultimate MM cell viability.

In summary, our data show that autophagy is significantly enhanced in MM cells exposed to ER stress inducers, mTOR inhibitors, and proteasome inhibitors. Constitutive autophagy seems to be a prosurvival mechanism in MM cells and could be a potential target of novel therapies. However, caution should be used before attempting combination therapy with newer autophagy inhibitors and proteasome inhibitors.

No potential conflicts of interest were disclosed.

We thank Myrna Fisher for assistance with flow cytometry and the following University of California at Los Angeles core labs of the Jonsson Cancer Center: Janis Giorgi Flow Cytometry Laboratory and Microscopic Technique and Electron Microscopy Core Facility.