MAFb protein confers intrinsic resistance to proteasome inhibitors in multiple myeloma

Human multiple myeloma cell lines (HMCLs) MM1S was purchased from ATCC (Manassas, VA, Catalogue number: CRL-2974). Other HMCLs were kind provided by Dr. Stuart Rudikoff, NCI, NIH and Dr. Michael Kuehl, NCI, NIH and described in detailed previously [16, 29, 30]. Cells were cultured in growth medium as described previously [29]. CD138-expressing patient MM cells were isolated as previously described [16].

MAFb gene expression was determined by GEP analysis of CD138-expressing MM cells from patients as described previously [1]. MAFb gene expression data set from primary CD138 expressing MM cell presented in this paper have been deposited in the NIH Gene Expression Omnibus (GEO; National Center for Biotechnology Information [NCBI], http://​www.​ncbi.​nlm.​nih.​gov/​geo/​) under accession number GSE2658 as described previously [1].

Knockdown of MAFb gene was performed in HMCLs expressing high levels of MAFb using a lentiviral expressing system [16]. Briefly, two sequences (5′ -CCGGGCCTTGTCTTATGGTCAAATTCTCGAGAATTTGACCATAAGACAAGGCTTTTT-3′ and 5′- CCGGGCCCAGTCTTGCAGGTATAAACTCGAGTTTATACCTGCAAGACTGGGCTTTTT-3′) specific to MAFb gene were engineered in lentiviral expressing system. A control oligonucleotide sequence not matching any sequence in the human genome (5′- CCGGTACAACAGCCACAACGTCTATCTCGAGATAGACGTTGTGGCTGTTGTATTTTT-3′) was used as a control shRNA sequence (designated as shCon). Total RNA, isolated after 24, 48, or 72 h, was subjected to reverse transcription (RT)-PCR and qPCR to determine of the degree of target gene silencing. Whole cell lysates were subjected to immunoblotting for MAFb protein measurement using a specific antibody.

The proliferation of MM cells was determined by colorimetric 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) assay [31]. Briefly, MM cells (4X10⁴/well) in 0.1 ml of IMDM medium with 10% FCS were cultured in flat-bottom, 96- well microliter wells. At the indicated time points, 10 ul of 5 mg/mL MTT was added to each well and incubated for 4 h at 37 °C, followed by adding 0.1 ml of 10% sodium dodecyl sulfate and incubation at 37 °C. Optical density was read on a Spectra Max340 Microplate Spectrophotometer (Molecular Devices) at 570 nm. Proteasome inhibitors including Bzb and Carfilzomib (CFZ) were used as described previously [32, 33].

Cells were cultured in growth media alone for 24 h and treated with or without serial concentrations of proteasome inhibitors for indicated time points. The apoptotic cells were stained using Annexin V-FITC Apoptosis Detection Kit (BD Biosciences) as described previously [16]. Apoptotic cell numbers were detected by flow cytometry according to the manufacturer’s instructions and analyzed using BD FACSVerse™ System, (Becton Dickinson, Mountain View, CA) and BD FACSuite software. Assay for DNA content and cell cycle analysis was performed by FITC BrdU Flow Kit (BD Pharmingen™) staining following the manufacturer’s instructions and were analysized by FCS Express 5 Flow Cytometry Professional software (De Novo Software).

Co-culture of MM cells with human bone marrow stromal cells (HBMSCs) was performed to determine the effect of MAFb on MM apoptosis in the presence of the BM microenvironment as previously described [34]. MM cells were cultured on a bone marrow stromal layer in 6-well plates in the presence of serial concentrations of proteasome inhibitors (Bzb and CFZ) for 16 h. The MM cells were harvested and subjected to Annexin staining and apoptotic cells detected by flow cytometry analysis.

Proteins were isolated from cells treated with or without Bzb, CFZ, MG132, cycloheximide (CHX), SB216763 [34], and LiCl for indicated time points prior to harvesting. The concentration of protein in each sample was determined by BCA protein Assay. Equal concentrations of total protein (50 μl per lane) were separated by SDS-PAGE followed by electrophoretic transfer to Immobilon polyvinylidene difluoride membranes. The proteins in the membranes were detected by specific antibodies including Anti-MAFb, caspase-3, − 6, − 7, − 8, and − 9, PARP, lamin A/C antibodies [31]. The membranes were stripped and reblotting with anti-actin antibody for indicating protein loading.

For the study of MAFb half-life, cells were cultured in growth media in the presence of cycloheximide (CHX, 5 μg/ml, Sigma-Aldrich) for 0, 60, 120, 240 min and lysed at various time points. Cell lysate was subject to immunoblotting analysis. Protein loading was normalized by using anti-β-actin antibody. The autoradiographs were scanned and densitometry performed for quantification [33].

Immunofluorescence staining was used to determine MAFb protein expression in cell nuclei using a MAFb antibody and DAPI as described previously [32]. Cells were incubated with an anti-MAFb antibody after being fixed with 3.7% formaldehyde in PBS. The slides were stained with Alexa Fluor 488-labeled secondary anti-mouse IgG antibody or Alexa 594-labeled secondary anti-rabbit IgG antibody for visualization. All images were captured under an AxioImager fluorescent microscope (Zeiss, Jena, Germany) and digitalized with Zeiss AxioCam MRc5 and Zeiss Axiovision software.

For quantitation of mRNA expression, total RNA was extracted from MM cell lines and cDNA was generated from 1μg RNA as described previously [33]. qPCR was performed using a QuantStudio™ 6 and 7 Flex Real-Time PCR Systems. The primers /probe sets including MAFb (Hs00271378_s1), ITGB7 (Hs01565750-m1) and CCR1 (Hs00174288-m1) were purchased from Life Technologies. The indicated genes were amplified using methods described in Additional file 1 and gene expression calculated using 2ΔCT analysis. [35, 36].

We previously reported that patients in the MF subgroup including those with a t(14;20) have a poor survival and have not benefited from the addition of Bzb to their therapy [14, 15, 37]. We hypothesized that overexpression of MAFb protein confers intrinsic resistance of MM to proteasome inhibitors. To validate our hypothesis, we first assessed MAFb transcription in primary MM cells by examining MAFb mRNA from a panel of 803 patients belonging to several molecular subgroups using Affymetrix oligonucleotide microarrays. MAFb mRNA was highly expressed in 57 samples from patients in the MF subgroup, Fig. 1a. We next detected MAFb expression by q-RT-PCR in a panel of 32 HMCLs, Fig. 1b. The highest level of MAFb mRNA was seen in SACHI and EJM, both harboring t(14;20); intermediate levels of MAFb were seen in three cell lines; OPM-2 with t(4;14), XG2 with Eμ enhancer of immunoglobulin light chain (IGL) inserted near MAFb in 20q11 [10] and L363 with t(6;20) [7]. Low MAFb mRNA levels were seen in four lines; LP1, ANBL6, H929 and Delta47, while no MAFb mRNA was detected in the remaining cell lines.

Consistent with the mRNA levels in t(14;20), immunoblotting analysis demonstrated abundance of MAFb protein in SACHI and EJM, OPM-2 and XG-2 cells, with a low level in L363 cells, Fig. 1c. In contrast, MAFb protein was not detected in MAFb mRNA negative HMCLs. Comparable high levels of MAFb mRNA and protein were seen in SACHI, and EJM cells. Surprisingly, despite lower levels of MAFb mRNA, a relative high level of MAFb protein was seen in OPM-2 cells which harbors a t(4;14).

As mRNA were detected in some cell lines with no detectable protein by immunoblotting, immunofluorescent analysis was used to validate these findings. Weak expression of MAFb proteins was detected in LP1, ANBL6, and H929 cells, with high MAFb expression in SACHI, EJM, OPM-2, XG-2 and L363, Fig. 1d. These results suggest that in MM, immunofluorescent detection of MAFb protein is more sensitive than immunoblotting analysis. We therefore extended our analysis to CD138 selected MM cells using this sensitive method. We measured MAFb protein in CD138+ cells from 8 patients; two with t(14;20), two with t(4;14) and 4 belonging to other subgroups based on FISH analysis. The MAFb protein was highest in t(14;20), with intermediate levels in the two with t(4:14), and negative in the rest, Fig. 1e (data not shown). Taken together, these results indicate that MAFb protein abundance is likely correlated with the mRNA in t(14;20).

To investigate MAFb effects on the sensitivity of MM cells to proteasome inhibitors, we established the half-maximum inhibitory concentration (IC₅₀) of Bzb and CFZ for five MAFb protein positive HMCLs and a number of t(14;20) negative HMCLs including H929 by MTT assay for 48 h. The IC₅₀ of CFZ and Bzb for SACHI was 50- and 120-nM and, respectively, Fig. 2a. In contrast, treatment of H929 cells with 10-nM of Bzb or 6-nM of CFZ reached 50% of inhibition of cell proliferation, Fig. 1b. The IC₅₀ of Bzb and CFZ for SACHI cells were 12- and 8.2-fold higher than for H929 cells, respectively. These results indicate that SACHI cells are more resistant to PIs than H929. Similarly, a relatively high IC₅₀ for Bzb and CFZ was seen in EJM, Fig. 2c and XG2, Fig. 2d; respectively. The IC₅₀ values of Bzb and CFZ for L363 cells were at a midlevel, Fig. 2e. The CI₅₀ for Bzb and CFZ was 12- and 10-nM respectively, Fig. 2f, in OPM-2 cells despite similar high level of MAFb protein to SACHI and EJM. The mean IC₅₀ of Bzb (Fig. 2g) and CFZ (Fig. 2h) were significantly lower than those in cells, which lacked MAFb expression. The IC50s of Bzb and CFZ in each tested human MM cell lines are summarized at Additional file 1: Table S1. Taken together, these results suggest that myeloma cells with high MAFb protein due to translocation to 20q11 or Igλ, or insertion on 20q11 are associated with resistance to proteasome inhibitors.

To determine if protein degradation regulates the level of MAFb protein, we used cycloheximide (CHX) to inhibit new protein synthesis from mRNA and determined the protein half-life [38]. As shown in Fig. 3a, MAFb protein decreases in the presence of CHX in 5/5 HMCLs by 60 min, while the actin protein level remains stable for each cell line. The decay curves show that the time for half the amount of MAFb protein to be degraded (T1/2) for 4/5 cell lines was 75 min, and 65 min for L363 cells, Fig. 3b. The half-life of MAFb protein was shorter than beta-actin protein. These results suggest that stability of MAFb protein in MM cells is regulated by post-translational modification. The MAFb protein appeared to become phosphorylated as indicated by a gel mobility shift (a slow migrating band shifted from the main band) at OPM2, XG2 and L363, Fig. 1c and at 180-min for SACHI Fig. 3a.

Having demonstrated that protein degradation is responsible for MAFb protein levels in MM cells, we next sought to identify factors associated with regulating the stability of MAFb protein. Phosphorylation of MAF-A protein by GSK3 governs the degradation of the protein in other tissue as well as c-MAF in MM cells [16, 39]. To assess how the inhibition of GSK3 activity affects abundance of MAFb protein, we first used SB216763 to abrogate GSK3 activity. Treatment of HMCL with SB216763 led to blockage of degradation of MAFb protein, Fig. 3g. Significant stabilization of MAFb protein by SB216763 was observed by 120 min in SACHI, XG-2, L363, and OPM-2 cells, and by 180 min in EJM, Fig. 3 (h to l), indicating that activity of GSK3β is required for degradation of MAFB protein. To confirm that GSK3 activity affects MAFb stability, we treated the cell lines with LiCl, another GSK3beta inhibitor. Exposure to LiCl led to a stabilization of MAFb protein in all 5 MM cell lines within 120 min, (Additional file 1: Figure S1). These results confirm that GSK3 activity is required for degradation of MAFb protein in MM. Since MAPK regulates stability of MAF protein in other cell type [40], we determined if MAPK is involved in regulating stability of MAFb protein. To do so, we utilized U0126, which inhibits subsequent phosphorylation and activation of ERK by MEK [31]. Treatment of SACHI, EJM and XG-2 cells with U0126 did not affect MAFb protein levels (data not shown). We next looked at the effect of p38 and Jun N-terminal protein kinases on stabilization of MAFb protein by treatment of these cells with SB20350, a well-known, p38 inhibitor and SP6000125, a Jun N-terminal protein kinases, alone or in combination. Neither SB20350 alone, nor SP6000125 alone, nor combination of these two inhibitors together affect the level of MAFb protein, respectively (data not shown). Taken together, these results suggest that GSK3 activity regulates degradation of MAFb protein in MM cells independently of the MAPK pathway and p38 kinases.

Having shown MAFb protein is regulated by GSK3 activity, we next sought to see if the proteasome inhibitors had an effect on the stability of MAFb protein. Treatment of SACHI cells with Bzb resulted in an increase in MAFb, Fig. 4a. An increase in MAFb was also observed following treatment with CFZ. A similar effect of Bzb and CFZ on MAFb protein was observed in XG-2, Fig. 4b, while the proteasome inhibitors had no effect on MAFb mRNA level by qPCR analysis (data not shown) suggesting that proteasome inhibitors prevent degradation of MAFb in a dose-dependent manner.

Since MAFb is a transcription factor [41], we determined the effect of proteasome inhibitors on localization of MAFb protein using immunofluorescent staining. Treatment of SACHI cells with 20-nM of Bzb and CFZ led to an increase of MAFb in the nucleus, Fig. 4c. Similar increases in MAFb in the nucleus and cytoplasm were seen in XG-2 and EJM in response to Bzb and CFZ, Fig 4d, while no increase MAFb protein was seen in OPM-2 cells (data not shown) although inhibition of GSK3 activity stabilized MAFb protein in this cell line. Taken together, these results suggest that blockade of proteasome activity by proteasome inhibitors prevents degradation of MAFb protein in the cells with t(14;20) or Igλ insertion, independent of mRNA regulation.

To validate the ability of MAFb-mediated resistance to proteasome inhibitor, loss-of-function studies were undertaken using SACHI and EJM cells with robust MAFb expression. To reduce MAFb level, we used shRNA to reduce MAFb expression by 90% compared to cells infected with a scrambled shRNA (shCon), Fig. 5a. A decrease in MAFb protein level in shMAFb cell was observed with an associated decrease in nuclear expression, Fig. 5b and c. Upon MAFb knockdown, a reduction in expression of MAFb target genes, including CCR1, ITGB7 and CCND2 was evident, indicating a functional MAFb knockdown, Fig. 5d. We further explored the effect of MAFb knockdown on proliferation of SACHI cells, and found that proliferation in MAFb knockdown cells markedly reduced, Fig 5e. Furthermore, we evaluated the effect of knockdown of MAFb protein on sensitivity to proteasome inhibitors. In the presence of 200-, 300- and 400-nM of Bzb, the proliferation of shMAFb cells was significantly lower than those in shCon indicating knockdown of MAFb restores sensitivity to proteasome inhibitors. The effect of silencing MAFb on sensitivity to proteasome inhibitors was more obvious with CFZ, Fig. 5f. A similar increase in sensitivity to Bzb was seen in EJM upon knockdown of MAFb (data not shown). It should be noted that MM cell infected with lentivirus particle to expressing shCon or shMAFb lead to approximate 5% of cell undergoing apoptosis shCon and shMAF, compared with the cell without infected with lentivirus particle at the absence of treatment of proteasome inhibitors (data not shown). These results suggest that high expression of MAFb protein confers the intrinsic resistance of MM to proteasome inhibitors.

The effect of knockdown of MAFb on proteasome inhibitors-induced apoptosis in shMAFb/SACHI and shCon/SACHI cells was determined using Annexin V staining and flow cytometry analysis. Increases in the number of apoptotic cells were seen in shMAFb cells treated with Bzb, Fig. 6a and by CZF, Fig. 6b. These results indicate that silencing MAFb gene and protein enhances proteasome inhibitor-triggered apoptosis.

To see if the BM microenvironment affects the role of MAFb in proteasome inhibitor-mediated apoptosis, shMAFb and shCon cells were co-cultured with primary bone morrow stromal cells from myeloma patients. Treatment of shMAFb cells with CFZ led to increases in apoptotic cells at the presence of co-culture with bone marrow stromal cells, Fig. 6c, suggesting that interaction with bone marrow microenvironment does not protect MM cells from proteasome inhibitor-induced apoptosis.

Since the caspase family plays an important role in apoptosis [42], we next determined whether specific caspases regulated proteasome inhibitor-induced apoptosis in MAFb knockdown cells. We used Western blot to analyze the inactive as well as the cleaved, active forms of caspases in the loss-of-function (shMAFb and shCon) cells exposed to Bzb and CFZ. The protein levels of the cleavage fragment of caspases-8, an initiator caspases that functions as an active executioner of the caspase pathway [42], in shMAFb/SACHI were higher than those in shCon/SACHI cells, Fig. 6d. An increase in of caspases-9, another initiator of caspases, was observed in shMAFb/SACHI cells, without Bzb treatment indicating that silencing MAFb itself induces MM cell apoptosis. Importantly, an obvious increase in the cleavage fragment of caspase 9 was seen at 20 min and maintained for 160 min after to Bzb treatment in shMAFb cells. These results indicate that Bzb-induced apoptosis is mediated via activation of caspases-8 and -9, and knockdown MAFb enhances Bzb-induced activation of caspases-8 and -9 in MM cells.

Because caspases-3, − 6 and − 7 proteins are substrates of active caspase-8 and -9 [42], we examined the cleavage products in both shMAFb and shCon cells. Increases in cleavage fragment of caspase 3 in shMAFb cells were seen in the cells following treatment with Bzb, Fig. 6e. In contrast, silencing MAFb did not affect activation of caspases-7 and -6 (middle panel and data not shown). These results indicate MAFb protein diminishes Bzb-induced activation of caspases-3 but not caspase-6 and -7.

We next determine whether MAFb affects Bzb-induced activation of PARP, which is one of the major substrates of activated caspase-3 [43]. The cleavage of PARP facilitates cellular disassembly and serves as a hallmark of cells undergoing apoptosis [44]. Increases in the cleavage fragment of PARP were observed in shMAFb/SACHI cells. Moreover, the cleavage fragment of PARP was increased following treatment with Bzb. Thus, knockdown of MAFb promotes Bzb-induced activation of PARP.

Lamins are major factors in the structural organization and function of the nucleus and chromatin and cleavage of lamins results in nuclear deregulation and cell death [45]. We investigated the effect of MAFb protein on activation of lamins in response to Bzb. An increase in cleavage fragments of lamin A/C in shMAFb cells was clearly seen with Bzb, Fig 6d. Similar increases in cleavage fragments of caspases-3, − 7, − 8, and − 9, and PARP and lamin A/C were observed in shMAFb cells response to CFZ, Fig. 6e. Therefore, knockdown of MAFb was accompanied by increased activation of caspases − 8 and − 9, which in turn induced activation of caspase-3, and at the end resulting in the degradation of PARP and lamin A/C in response to the proteasome inhibitors.