Runx2 deficiency in osteoblasts promotes myeloma resistance to bortezomib by increasing TSP-1-dependent TGF- β1 activation and suppressing immunity in bone marrow

Abstract

Multiple myeloma (MM) is a plasma cell malignancy that thrives in the bone marrow (BM). The proteasome inhibitor bortezomib (BTZ) is one of the most effective front-line chemotherapeutic drugs for MM; however, 15–20% of high-risk patients do not respond to or become resistant to this drug and the mechanisms of chemoresistance remain unclear. We previously demonstrated that MM cells inhibit Runt-related transcription factor 2 (Runx2) in pre- and immature osteoblasts (OBs), and that this OB-Runx2 deficiency induces a cytokine-rich and immunosuppressive microenvironment in the BM. In the current study, we assessed the impact of OB-Runx2 deficiency on the outcome of BTZ treatment using OB-Runx2^+/+ and OB-Runx2^−/− mouse models of MM. In vitro and in vivo experiments revealed that OB-Runx2 deficiency induces MM cell resistance to BTZ via the upregulation of immunosuppressive myeloid-derived suppressor cells (MDSCs), downregulation of cytotoxic T cells, and activation of TGF-β1 in the BM. In MM tumor-bearing OB-Runx2^−/− mice, treatment with SRI31277, an antagonist of thrombospondin-1 (TSP-1)–mediated TGF-β1 activation, reversed the BM immunosuppression and significantly reduced tumor burden. Furthermore, treatment with SRI31277 combined with BTZ alleviated MM cell resistance to BTZ-induced apoptosis caused by OB-Runx2 deficiency in co-cultured cells and produced a synergistic effect on tumor burden in OB-Runx2^−/− mice. Depletion of MDSCs by 5-fluorouracil or gemcitabine similarly reversed the immunosuppressive effects and BTZ resistance induced by OB-Runx2 deficiency in tumor-bearing mice, indicating the importance of the immune environment for drug resistance and suggesting new strategies to overcome BTZ resistance in the treatment of MM.

Introduction

Multiple myeloma (MM) is a plasma cell malignancy that thrives and progresses in the bone marrow (BM) (1,2). Despite advances in treatment, MM remains incurable, with a 90% relapse rate. This high relapse rate is largely due to the development of chemoresistance (3), which occurs even with frontline MM drugs such as the proteasome inhibitor bortezomib (BTZ) (4–6). Although the BM microenvironment is known to have an important role in MM progression (7,8), and the tumor microenvironment likely serves as a major contributor to MM chemoresistance (9), the mechanisms governing the development of chemoresistance in MM are not well understood.

We and others previously demonstrated that MM cells secrete soluble factors that suppress Runt-related transcription factor 2 (Runx2) expression in osteoblasts (OBs), thereby inhibiting osteoblastogenesis and bone formation (10–12). Using a syngeneic mouse model of MM in which Runx2 is specifically deleted in the immature OBs of C57BL6/KaLwRij mice (OB-Runx2^−/− mice), we showed that OB-Runx2 deficiency feeds back to promote MM dissemination to and progression in these areas by enhancing OB production and secretion of cytokines and chemokines that produce an immunosuppressive microenvironment in the BM (13). With this in mind, we hypothesized that MM-induced Runx2 suppression in OBs might be a key contributor to microenvironmentally mediated chemoresistance in MM. Therefore, in the present study, we investigated the effect of OB-Runx2 deficiency on BTZ resistance in MM using the OB-Runx2^−/− model and in vitro tools.

Materials and Methods

Cell lines and cell culture

Mouse 5TGM1 MM cells expressing firefly luciferase (5TGM1-Luc) were a gift from Dr. Fenghuang Zhan (University of Arkansas for Medical Sciences, Little Rock, AR). MPC-11 cells were purchased from American Type Culture Collection (ATCC). Both cell lines were grown in RPMI 1640 medium (Corning) supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin (Sigma Aldrich) in an incubation chamber under standard conditions (5% CO₂ and 37°C) for 7 to 10 days before being used in experiments (medium changed every 3 days). Cell authentication was conducted by assessing the following features, as previously described (13): (i) the expression of IgG2bk (a marker of both 5TGM1 and MPC-11 cells) by flow cytometry (FACS); (ii) in vitro growth curves by MTT assay; (iii) in vivo growth by injecting 5TGM1 cells into C57BL/KaLwRij mice via tail vein and measuring levels of IgG2bk (a soluble marker of 5TGM1 cells) in murine serum by ELISA. We confirmed that 5TGM1-Luc maintain the same characteristics as wild-type 5TGM1 cells and the 5TGM1 cells used in the publications of other researchers. We also confirmed that these 5TGM1 cells and MPC-11 cells are mycoplasma negative, using Immu-Mark Myco-Test Kit (MP Biomedicals). These tests were performed each time before the cells were used in vitro and in vivo experiments.

OB-Runx2^+/+ and OB-Runx2^−/− mice

We previously generated a syngeneic immunocompetent model of murine MM with specific deletion of Runx2 in immature OBs (OB-Runx2^−/− mice) (13,14). Briefly, we first generated a mouse model with conditional deletion of Runx2 in immature OBs on a C57BL6 background using Cre-recombinase driven by the Col1a promoter, then backcrossed these mice with C57BL/KaLwRij mice, an immunocompetent syngeneic model of murine MM, and intercrossed the resulting Runx2^+/− C57BL/KaLwRij mice to generate litters. Wild-type Runx2^OB+/+ C57BL//KaLwRij mice (control littermates) and homozygous Runx2^OB−/−/KaLwRij mice are referred to herein as OB-Runx2^+/+ and OB-Runx2^−/− mice, respectively (13).

For experiments conducted with MM tumor-bearing mice, 2×10⁶ 5TGM1-Luc MM cells were injected via lateral tail vein (i.v. injection) into 5-week-old syngeneic OB-Runx2^+/+ mice and OB-Runx2^−/− mice (13). On day 8 after tumor cell injection, mice were randomly assigned to treatment groups and began treatment. Tumor burden after treatment was assessed via in vivo bioluminescence imaging performed with an IVIS Lumina III bioluminescence system and by measuring serum levels of IgG2bκ (a soluble marker of 5TGM1 MM cells) by enzyme-linked immunosorbent assay (ELISA). Upon sacrifice, BM was flushed from the femurs and tibias of each mouse in 1mL PBS and centrifuged at 300 × g at 4°C for 5 min to collect BM cells and BM supernatant (BMS) for subsequent analyses (15). All animal studies were performed in accordance with University of Alabama at Birmingham (UAB) and National Institutes of Health (NIH) guidelines after institutional review and approval. Both of male and female mice were used in each experiment.

BTZ treatment in vivo

BTZ (Calbiochem, Supplementary Table S1) was dissolved in sterile phosphate buffered saline (PBS) and administered by intraperitoneal (i.p.) injection (0.5 mg/kg body weight) twice per week for 4 weeks, starting on day 8 after tumor cell injection. The control group was i.p. injected with PBS twice per week for 4 weeks.

Depletion of myeloid-derived suppressor cells in vivo

Myeloid-derived suppressor cells (MDSCs) were depleted in the BM of 5TGM1-Luc MM tumor-bearing OB-Runx2^−/− and OB-Runx2^+/+ mice by i.p. injection of 5-fluorouracil (5-Fu, Sigma Aldrich, Supplementary Table S1) or by i.p. injection of gemcitabine (GEM, Selleckchem, Supplementary Table S1) in tumor-bearing OB-Runx2^−/− mice. 5-Fu and GEM are selective inhibitors of MDSCs which do not directly affect lymphocytes (16–18). Starting on day 8 after tumor cell injection, mice were treated with 5-Fu or GEM twice per week for 4 weeks; each drug was administered at 30 mg/kg body weight/injection, equal to 180 mg/m² body surface area/week (19).

Inhibition of thrombospondin-1–dependent TGF-β1 activation in vivo

SRI31277 is a tripeptide (ser-lys-leu) that blocks thrombospondin-1 (TSP-1) binding to and activation of latent TGF-β. SRI31277 was synthesized as described by our group previously (20), and the structure of SRI31277 has been reported (20,21).

SRI31277 was administered to 5TGM1-Luc MM tumor-bearing OB-Runx2^+/+ mice and OB-Runx2^−/− mice via an osmotic pump (ALZET) at a dose of 30 mg/kg body weight per day beginning on day 8 after tumor cell injection. Two weeks after initiation of the infusion, the osmotic pumps were replaced with new pumps containing fresh compounds and the mice were continually treated for another 2 weeks (20).

Detailed methods for bioluminescence imaging, ELISAs, flow cytometry, cell co-culture, 3- (4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) cell viability assays, and Western blotting are provided in the supplementary methods. The company’s name and the product numbers of all antibodies used for this study were listed in the Supplementary Table S1.

Data Availability Statement

The raw data of animal bioluminescence imaging for this study were generated by the IVIS Lumina III Bioluminescence System at Volker Hall Animal Facility, UAB. Derived data supporting the findings of this study are available from the corresponding author upon request. The other data generated in this study are available within the article and its supplementary data files.

Statistical analysis

Data are presented as mean ± standard error of the mean (SEM). Differences in means were analyzed with Student’s t-test to evaluate continuous variables of two groups and with one-way ANOVA to evaluate continuous variables of more than two groups. A P < 0.05 was considered statistically different.

Results

OB-Runx2 deficiency induces MM cell resistance to BTZ treatment in vivo

MM cells secrete soluble factors that promote OB-Runx2 deficiency (10). To evaluate whether OB-Runx2 deficiency in turn affects the anti-MM function of BTZ, 5-week-old OB-Runx2^+/+ mice and OB-Runx2^−/− mice were i.v. injected with 5TGM1-Luc MM cells (2×10⁶). On day 7 after tumor cell injection, serum levels of IgG2bκ (a soluble marker of 5TGM1 MM cells), measured by ELISA, showed that the tumor burden was slightly higher in OB-Runx2^−/− mice than in OB-Runx2^+/+ mice but the difference was not significant (supplementary Figure S1A). The next day, mice began 4-week treatments with either BTZ (i.p. injection, 0.5 mg/kg body weight, twice/week) or PBS (Figure 1A). Bioluminescence imaging and serum IgG2bκ ELISA performed after 4-week treatment showed that BTZ reduced MM growth in OB-Runx2^+/+ mice but not in OB-Runx2^−/− mice (Figure 1B–C). Flow cytometry analysis of 5TGM1-Luc MM cells in the BM further showed that BTZ treatment failed to reduce the percentage of MM cells (CD138⁺) and the percentage of MM cells expressing Ki-67 (proliferation marker) and BCL-2 (anti-apoptotic marker) in the BM in OB-Runx2^−/− mice, in contrast to the effectiveness observed in OB-Runx2^+/+ mice (Figure 1D–F), suggesting OB-Runx2 deficiency induces MM cell resistance to BTZ treatment in vivo.

Figure 1. OB-Runx2 deficiency induces MM cell resistance to BTZ treatment in vivo.

A, Schematic diagram of the tumor injection and treatment schedule for mice used in B-F. Five-week-old OB-Runx2^+/+ and OB-Runx2^−/− mice were i.v. injected with 5TGM1-Luc MM cells (down arrow) 1 week before treatment (left up arrow) with PBS or BTZ for 4 weeks. Blood and BM were collected at study end (right up arrow) for all analyses. B, Representative bioluminescence imaging of MM tumor-bearing OB-Runx2^+/+ and OB-Runx2^−/− mice after treatment (left); graphical representation of the luminescence intensity in each group (right) (n=5 mice/group). C, Quantification of serum IgG2bκ concentration in mice after treatment, measured by ELISA (in duplicate) (n=8–10 mice/group). D-F, BM cells harvested after treatment were analyzed by flow cytometry (n=7 mice/group). D, Representative plots (left) and the percentage of CD138⁺ (a membrane marker of 5TGM1 MM cells) MM cells detected among all B220⁺ B cells (right). E, Representative plots (left) and the percentage of Ki-67⁺ MM cells detected among CD138⁺ cells (right). F, Representative plots (left) and the percentage of BCL-2⁺ MM cells detected among CD138⁺ cells (right). G-H, 5TGM1-Luc MM cells were co-cultured with Pre-OBs harvested from the calvaria of newborn OB-Runx2^+/+ or OB-Runx2^−/− mice for 12 h before treatment with BTZ (0, 2.5, 5, or 10 nM) for 36 h. G, Representative bioluminescence imaging of luciferase activity in the 5TGM1-Luc MM cells after treatment (n=6 wells/group). H, Percent viability of 5TGM1-Luc MM cells, assessed by MTT assay, after treatment (in triplicate). I, Percent viability of OB-Runx2^+/+ and OB-Runx2^−/− Pre-OBs, assessed by MTT assay, after co-culture with 5TGM1-Luc MM cells for 48 h (n=24 wells/group). J, Percent viability of 5TGM1-Luc MM cells that were cultured for 24 h in medium mixed with BMS harvested from OB-Runx2^+/+ or OB-Runx2^−/− mice (no tumor injection) and PBS or BTZ (0, 2.5, 5, or 10 nM) (n=4/group). Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.005, ****P < 0.001. ns, not significant.

The influence of OB-Runx2 deficiency on MM cell growth and the efficacy of BTZ treatment was then examined in vitro. 5TGM1-Luc MM cells were co-cultured with precursor OBs (Pre-OBs) isolated from the calvaria of newborn OB-Runx2^+/+ or OB-Runx2^−/− mice, without or with BTZ (2.5–10.0 nM). Compared with MM cells co-cultured with Runx2^+/+ Pre-OBs, those co-cultured with Runx2^−/− Pre-OBs had higher rates of proliferation and viability (Supplementary Figure S1B–D) and lower expression of the apoptotic marker cleaved caspase-3 (Supplementary Figure S1E–F) and were less sensitive to BTZ treatment (Figure 1G–H; Supplementary Figure S1E–F). However, the viability of the Runx2^+/+ OBs and Runx2^−/− OBs did not differ significantly in these co-cultures (Figure 1I)

Next, the effect of BM soluble factors alone on MM sensitivity to BTZ was determined. 5TGM1-Luc MM cells were cultured in medium mixed in a 1:1 ratio with PBS (control) or BM supernatant (BMS) from OB-Runx2^+/+ or OB-Runx2^−/− mice and treated with PBS or BTZ (2.5–10.0 nM) for 24 h. BTZ treatment significantly decreased the viability of MM cells cultured in OB Runx2^+/+ BMS at all concentrations tested, but higher concentrations of BTZ were required to significantly decrease the viability of MM cells cultured in OB Runx2^−/− BMS. (Figure 1J). These data suggest that alterations in OB-lineage cells themselves are responsible for BTZ resistance, and that soluble factors continuously released by Runx2^−/− OBs (and possibly other cells) in this BM milieu mediate the induction of MM cell resistance to BTZ observed in OB-Runx2-deficient mice.

BTZ treatment enhances OB-Runx2 deficiency-induced immunosuppression and TGF-β1 activity in BM

Our previous study demonstrated that OB-Runx2 deficiency induces BM immunosuppression (13). To test whether this BM-specific immunosuppression contributes to OB-Runx2 deficiency-promoted BTZ resistance in MM, BM (containing BM cells and 5TGM1-Luc MM cells) was harvested from tumor-bearing OB-Runx2^+/+ and OB-Runx2^−/− mice after 4 weeks of treatment with PBS or BTZ and analyzed by flow cytometry. Analyses of BM cells from the PBS-treated mice replicated our previous findings (13): compared with the BM of tumor-bearing OB-Runx2^+/+ mice, the BM of tumor-bearing OB-Runx2^−/− mice contained more activated MDSCs (22,23) (Figure 2A–B) and fewer cytotoxic CD8⁺ T cells, which had signs of exhaustion (24,25) (Figure 2C–H).

Figure 2. OB-Runx2 deficiency and BTZ treatment induce changes in the BM microenvironment in vivo.

BM cells and BMS were harvested from tumor-bearing OB-Runx2^+/+ and OB-Runx2^−/− mice treated for 4 weeks with PBS or BTZ as depicted in Fig. 1. A-H, Immunosuppressive cells and immune effector cells in the BM of tumor-bearing OB-Runx2^+/+ and OB-Runx2^−/− mice were assessed by flow cytometry (n=7 mice/group). A, Representative plots (left) and the percentage of MDSCs (Gr1⁺ CD11b^hi) detected in the BM cell population (right). B, Relative expression of iNOS, arginase 1, and IL-10 in BM MDSCs (Gr1⁺ CD11b^hi). C, Gating strategy for identifying CD8⁺ T cells and CD8⁺ T cells expressing exhaustion markers in the BM cell population. D, Percentage of CD8⁺ T cells detected among all CD3⁺ T cells (CD3⁺CD8⁺) in the BM. E-H, Percentage of BM CD8⁺ T cells expressing PD-1 (E), TIM-3 (F), granzyme B (G), and IFN-γ (H). I, Representative Western blots showing TSP-1 expression in the BMS of tumor-bearing OB-Runx2^+/+ and OB-Runx2^−/− mice after PBS or BTZ treatment (left) and densitometric quantification of TSP-1 normalized to GAPDH (in triplicate) (right). J, Quantification of active TGF-β1 concentration in the BMS of tumor-bearing OB-Runx2^+/+ and OB-Runx2^−/− mice after treatment, measured by ELISA (in duplicate). Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.005, ****P < 0.001. ns, not significant.

Importantly, treatment of tumor-bearing OB-Runx2^−/− mice with BTZ further increased the BM abundance of MDSCs (Figure 2A), especially monocytic MDSCs (M-MDSCs) (Supplementary Figure S1G–H), as well as the expression of proteins indicative of activation, including inducible nitric oxide synthase (iNOS), arginase 1, and IL-10 (22,23) in these immunosuppressive cells (Figure 2B). BTZ treatment also further reduced the abundance of total CD8⁺ T cells, and of CD8⁺ T cells expressing the cytotoxic molecules granzyme B and IFN-γ, and increased the abundance of CD8⁺ T cells expressing an exhaustion marker (programmed death 1 [PD-1] or T-cell immunoglobulin and mucin-domain containing-3 [TIM-3] (24,25)), in the BM (Figure 2C–H). BTZ did not induce these immunosuppressive effects in the BM of tumor-bearing OB-Runx2^+/+ mice (24,25) (Figure 2A–H; Supplementary Figure S1G–H), indicating that BTZ treatment further increases OB-Runx2 deficiency-induced immune suppression in BM.

Our previous study showed that total TGF-β1 in BM is upregulated by OB-Runx2 deficiency (13). Active TGF-β1 is a negative regulator of immune cell function in the tumor microenvironment (26–28) and plays an important role in cell proliferation, apoptosis, and drug resistance (29,30). Moreover, TGF-β1 has been shown to be important for tumor progression and osteolytic bone disease in MM (31). The endogenous activation of TGF-β1 is mediated in MM models by the matricellular protein TSP-1. In the current study, PBS-treated tumor-bearing OB-Runx2^−/− mice had significantly higher BM levels of TSP-1 and active TGF-β1 than their OB-Runx2^+/+ counterparts had. Interestingly, BTZ treatment further increased the level of both TSP-1 and active TGF-β1 in the OB-Runx2^−/− mice (Figure 2I–J), although as we previously reported (20), it did not affect the level of either in the BM of tumor-bearing OB-Runx2^+/+ mice. These results suggest that OB-Runx2 deficiency increases TGF-β1 activation via TSP-1 in BM, and BTZ treatment further enhances this effect.

SRI31277 inhibits MDSC proliferation and directly alleviates MM cell resistance to BTZ treatment by inhibiting TGF-β1 pathways in vitro

SRI31277, which blocks TSP-1–mediated activation of TGF-β1 (20), was then used to determine the effect of active TGF-β1 on MDSC proliferation and MM cell resistance to BTZ in vitro. BM cells from OB-Runx2^+/+ mice were cultured with GM-CSF and IL-6 to induce the differentiation of MDSCs from mononuclear cells. These MDSCs were then sorted by flow cytometry and cultured in medium mixed 1:1 with BMS from non-tumor-bearing OB-Runx2^+/+ or OB-Runx2^−/− mice (5 weeks of age), with or without SRI31277 (40 nM), for 48 h. SRI31277 prevented the increase in MDSC proliferation induced by OB-Runx2^−/− BMS (Figure 3A).

Figure 3. SRI31277 inhibits MDSC proliferation and directly alleviates MM cell resistance to BTZ by inhibiting the TGF-β1 pathway in vitro.

A, MDSCs derived from the BM of OB-Runx2^+/+ mice were cultured in medium mixed with BMS from OB-Runx2^+/+ or OB-Runx2^−/− mice (no tumor injection) and PBS or SRI31277 (40 nM) for 48 h (n=4, BMS from 8 mice per group). Cells were then stained with trypan blue and counted. B-G, 5TGM1-Luc or MPC-11 MM cells were cultured in medium mixed with BMS from OB-Runx2^+/+ or OB-Runx2^−/− mice (no tumor cell injection) and PBS, BTZ (2.5 nM), SRI31277 (25 nM), or BTZ+SRI31277 for 24 h. B, Percent viability of 5TGM1-Luc MM cells after treatment, assessed by MTT assay (n=4, BMS from 3 mice per group). C, Percent viability of MPC-11 MM cells after treatment, assessed by MTT assay (n=3, BMS from 3 mice per group). D, Representative Western blots showing the detection of cleaved caspase-3 in 5TGM1-Luc MM cells after treatment (left) and densitometric quantification of cleaved caspase-3 normalized to GAPDH (right). E-G, Representative Western blots showing the detection of total and phosphorylated SMAD2/3 and ERK1/2 in 5TGM1-Luc MM cells after each treatment (E) and densitometric quantification of the p-SMAD2/3 to t-SMAD2/3 ratio (F) and of the p-ERK1/2 to t-ERK1/2 ratio (G) normalized to GAPDH. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.005, ****P < 0.001. ns, not significant; p-SMAD2/3, phosphorylated SMAD2/3; t-SMAD2/3, total SMAD2/3; p-ERK1/2, phosphorylated ERK1/2; t-ERK1/2, total ERK1/2.

Next, 5TGM1-Luc MM cells were cultured for 24 h in medium mixed 1:1 with the BMS from non-tumor-bearing OB-Runx2^+/+ or OB-Runx2^−/− mice and PBS, BTZ (2.5 nM), SRI31277 (40 nM), or BTZ+SRI31277. BTZ only decreased the viability of the MM cells cultured with OB-Runx2^+/+ BMS. Conversely, SRI31277 only decreased the viability of the MM cells cultured with OB-Runx2^−/− BMS. Interestingly, however, treatment with BTZ+SRI31277 augmented the effect of SRI31277 alone (Figure 3B). Similar results were obtained in MPC-11 MM cells cultured for 24 h in the same manner (Figure 3C). In addition, Western blot analysis showed that BTZ treatment significantly increased the expression of cleaved caspase-3 only in the MM cells cultured in OB-Runx2^+/+ BMS. In contrast, treatment with SRI31277 significantly increased the expression of cleaved caspase-3 only in the MM cells cultured in OB-Runx2^−/− BMS, and treatment with BTZ+SRI31277 further enhanced this effect (Figure 3D).

To further explore the molecular mechanisms by which SRI31277 alleviates MM cell resistance to BTZ, the activity of the canonical (SMAD2/3) and non-canonical (ERK1/2) signaling pathways of TGF-β1 (32,33) was assessed in 5TGM1-Luc MM cells cultured in OB-Runx2^+/+ or OB-Runx2^−/− BMS. Compared with cells cultured in OB-Runx2^+/+ BMS, MM cells cultured in OB-Runx2^−/− BMS expressed significantly higher levels of phosphorylated SMAD2/3 (p-SMAD2/3) and phosphorylated ERK1/2 (p-ERK1/2) (Figure 3E–G). In MM cells cultured with OB-Runx2^+/+ BMS, treatment with BTZ significantly reduced the level of p-ERK1/2, but not of p-SMAD2/3, whereas SRI31277 had no effect. In MM cells cultured with OB-Runx2^−/− BMS, however, treatment with SRI31277 significantly reduced the level of p-ERK1/2, but not of p-SMAD2/3, and BTZ did not affect the level of either p-SMAD2/3 or p-ERK1/2. Interestingly, treatment with BTZ+SRI31277 reduced the levels of both p-SMAD2/3 and p-ERK1/2 in these cells, suggesting a synergistic effect (Figure 3E–G). These data indicate that OB-Runx2 deficiency promotes MM cell resistance to BTZ through the upregulation of TGF-β1 signaling in the BMS and that treatment with SRI31277 can reverse this resistance.

Combining SRI31277 and BTZ treatment restores the anti-MM function of BTZ in OB-Runx2^−/− mice

To determine whether blocking TSP-1–mediated TGF-β1 activation can alleviate the resistance to BTZ induced by OB-Runx2 deficiency in vivo, 5TGM1-Luc MM tumor-bearing OB-Runx2^+/+ and OB-Runx2^−/− mice were treated with PBS, BTZ, SRI31277 (osmotic pump, 30 mg/kg body weight per day), or BTZ+SRI31277 for 4 weeks (Figure 4A). Measurement of serum IgG2bκ levels showed that tumor burden did not differ between groups one day before treatment (Supplementary Figure S2A). After the 4-week treatment, tumor burden was significantly greater in PBS-treated OB-Runx2^−/− mice than in PBS-treated OB-Runx2^+/+ mice (Figure 4B). Interestingly, SRI31277 treatment reduced tumor burden in OB-Runx2^−/− mice, whereas BTZ treatment did not. Furthermore, treatment with BTZ+SRI31277 augmented the effect of SRI31277 treatment alone (Figure 4B). Flow cytometry analyses of CD138⁺ MM cells in the BM yielded consistent findings—in OB-Runx2^−/− mice, SRI31277 treatment reduced the total abundance of MM cells and the abundance of Ki-67-expressing MM cells, and BTZ+SRI31277 treatment augmented these effects, whereas BTZ treatment alone did not elicit a response (Figure 4C,D). Although neither BTZ nor SRI31277 alone enhanced the abundance of MM cells expressing cleaved caspase-3 in the BM of these OB-Runx2^−/− mice, the combination of BTZ+SRI31277 did (Figure 4E). In contrast, SRI31277 treatment did not affect the abundance of MM cells in OB-Runx2^+/+ mice, and the combinatorial drug treatment did not have a significant synergistic effect in these mice. Importantly, these results demonstrate that SRI31277 can effectively reverse MM resistance to BTZ induced by OB-Runx2 deficiency in vivo.

Figure 4. Treatment with SRI31277 restores MM sensitivity to BTZ in OB-Runx2^−/− mice.

A,Schematic diagram of the tumor injection and treatment schedule for mice used in experiments in B-O. 5-week-old OB-Runx2^−/− and OB-Runx2^+/+ mice were i.v. injected with 5TGM1-Luc MM cells (down arrow) 1 week before treatment with PBS, BTZ, SRI31277, or BTZ+SRI31277 (left up arrow) for 4 weeks. Blood and BM were collected at study end (right up arrow) for all analyses (n=5–9 mice/group). B, Quantification of serum IgG2bκ concentration in mice after treatment, measured by ELISA in duplicate (n=5–9 mice/group). C-E, BM cells harvested after treatment were analyzed by flow cytometry for the percentage of CD138⁺ MM cells among all B220⁺ B cells (C) and the percentage of Ki-67⁺ MM cells (D) and cleaved caspase-3-positive MM cells (E) among the CD138⁺ cells (n=5 mice/group). F, Percentage of MDSCs (Gr1⁺ CD11b^hi) in the BM cell population. G-I, Relative expression of iNOS (G), arginase 1 (H), and IL-10 (I) in total MDSCs (Gr1⁺ CD11b^hi). J, Gating strategy for identifying CD8⁺ T cells and CD8⁺ T cells expressing exhaustion and activation markers in the BM cell population. K, Percentage of CD8⁺ T cells detected among all CD3⁺ T cells (CD3⁺CD8⁺) in the BM. L-O, Percentage of CD8⁺ T cells in the BM expressing PD-1 (L), TIM-3 (M), granzyme B (N), and IFN-γ (O). Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.005, ****P < 0.001. ns, not significant.

Next, the cellular mechanisms by which SRI31277 mediates these anti-MM effects in tumor-bearing mice were assessed by flow cytometry. In agreement with the results in Figure 2, PBS-treated tumor-bearing OB-Runx2^−/− mice had a higher abundance of BM MDSCs than tumor-bearing OB-Runx2^+/+ mice had. MDSCs of the tumor-bearing OB-Runx2^−/− mice also expressed more iNOS, arginase 1, and IL-10 than MDSCs of their OB-Runx2^+/+ counterparts expressed. Surprisingly, in tumor-bearing OB-Runx2^−/− mice, treatment with BTZ further increased the BM abundance of MDSCs and MDSC expression of iNOS, arginase 1, and 1L-10. Significantly, treatment with SRI31277 reduced the BM abundance of MDSCs (M-MDSCs and polymorphonuclear MDSCs [PMN-MDSCs]) and the expression of activation markers in these immunosuppressive cells. Treatment with BTZ+SRI31277 overcame the enhancement in these measures induced by BTZ alone, but no synergistic effect was detected, suggesting that the modulation of MDSCs and these MDSC activation markers was primarily due to SRI31277 (Figure 4F–I; Supplementary Figure S2B–C). Neither BTZ, SRI31277, nor BTZ+SRI31277 treatment affected the expansion of MDSCs or the expression of activation markers by these cells in the BM of tumor-bearing OB-Runx2^+/+ mice.

The CD8⁺ T cell subpopulation in the BM of tumor-bearing OB-Runx2^−/− mice was similarly regulated by SRI31277. SRI31277 significantly increased the abundance of CD8⁺ T cells in the BM and, among these CD8⁺ T cells, decreased the percentage expressing PD-1 or TIM-3 and increased the percentage expressing granzyme B or IFN-γ. BTZ treatment had the opposite effect on these measures (Figure 4J–O). As with MDSCs, BTZ+SRI31277 did not appear to produce synergistic effects on CD8⁺ T cells in the BM of OB-Runx2^−/− mice (Figure 4J–O), and none of the treatments had a significant effect on CD8⁺ T cells in the BM of tumor-bearing OB-Runx2^+/+ mice.

These results demonstrate that blocking TSP-1-mediated TGF-β1 activation with SRI31277 alleviates MM resistance to BTZ induced by OB-Runx2 deficiency by inhibiting the activation of MDSCs and restoring the activity of cytotoxic T cells, which results in decreased MM cell proliferation and increased MM cell death.

Depletion of MDSCs by 5-Fu or GEM overcomes BTZ resistance in MM induced by OB-Runx2 deficiency

5-Fu and GEM are FDA-approved anti-cancer agents used in the treatment of a variety of solid tumors, and the effects of these drugs as MDSC inhibitors in the treatment of solid cancers have been reported (17,19,34,35). However, neither drug is commonly used as an anti-MM therapy. To confirm that MDSCs indeed promote OB-Runx2 deficiency-induced MM resistance to BTZ and to determine if pharmacologic depletion of MDSCs can overcome this resistance, 5TGM1-Luc MM tumor-bearing OB-Runx2^+/+ and OB-Runx2^−/− mice were treated with PBS, BTZ, 5-Fu (i.p. injection, 30 mg/kg body weight, twice per week), or BTZ+5-Fu for 4 weeks as described in Figure 5A. Serum IgG2bκ ELISA showed that tumor burden did not differ among the groups one day before treatment (Supplementary Figure S3). After 4-week treatment, bioluminescence imaging showed that treatment with BTZ reduced the tumor burden only in OB-Runx2^+/+ mice, yet treatment with 5-Fu significantly reduced the tumor burden in both OB-Runx2^+/+ and OB-Runx2^−/− mice. Importantly, in OB-Runx2^−/− mice, treatment with BTZ+5-Fu resulted in the greatest reduction in tumor burden among all treatments. In OB-Runx2^+/+ mice, however, treatment with BTZ+5-Fu did not reduce tumor burden beyond that achieved with BTZ or 5-Fu alone (Figure 5B). Results were similar when total tumor burden was assessed by serum IgG2bκ ELISA (Figure 5C). Flow cytometry analysis further showed that treatment with 5-Fu or BTZ+5-Fu caused the greatest reduction in the total abundance of MM cells and the abundance of Ki-67–expressing MM cells and the greatest increase in cleaved caspase-3–expressing MM cells in the BM of OB-Runx2^−/− mice (Figure 5D–G). However, in OB-Runx2^+/+ mice, BTZ+5-Fu did not produce effects on MM cells beyond those induced with BTZ alone (Figure 5D–G). These results demonstrate that 5-Fu can overcome MM resistance to BTZ induced by OB-Runx2 deficiency.

Figure 5. Treatment with 5-Fu overcomes MM resistance to BTZ induced by OB-Runx2 deficiency.

A, Schematic diagram of the tumor injection and treatment schedule for mice used in experiments in B-E. 5-week-old OB-Runx2^−/− and OB-Runx2^+/+ mice were i.v. injected with 5TGM1-Luc MM cells (down arrow) 1 week before treatment with PBS, BTZ, 5-Fu, or BTZ+5-Fu (left up arrow) for 4 weeks. Blood and BM were collected at study end (right up arrow) for all analyses (n=5–7 mice/group). B, Representative bioluminescence imaging of OB-Runx2^−/− and OB-Runx2^+/+ mice after treatment (left); graphical representation of the luminescence intensity in each group (right) (n=5–7 mice/group). C, Quantification of serum IgG2bκ concentration in mice after treatment, measured by ELISA in duplicate (n=5–7 mice/group). D, Flow cytometry gating strategy for identifying CD138⁺, Ki-67⁺, and cleaved caspase-3-positive cells among BM cells harvested after treatment (n=5 mice/group). E-G, Percentage of CD138⁺ MM cells detected among all B220⁺ B cells (E) and the percentage of Ki-67⁺ MM cells (F) and cleaved caspase-3⁺ MM cells (G) detected among the CD138⁺ cells. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.005, ****P < 0.001. ns, not significant.

To determine if MDSC depletion with GEM similarly overcomes MM resistance to BTZ in OB-Runx2^−/− mice, 5TGM1-Luc MM tumor-bearing OB-Runx2^−/− mice were treated with PBS, BTZ, GEM (i.p. injection, 30 mg/kg body weight, twice per week), or BTZ+GEM for 4 weeks. Similar to the results obtained with 5-Fu, treatment with GEM or BTZ+GEM caused the greatest reduction in tumor burden, as assessed by serum IgG2bκ ELISA, as well as the greatest reduction in the total abundance of MM cells and in the abundance of Ki-67– and BCL-2–expressing MM cells in the BM of tumor-bearing OB-Runx2^−/− mice (Supplementary Figure S4A–D). Notably, treatment with BTZ+GEM had a synergistic effect on tumor burden, the total abundance of MM cells, and the abundance of Ki-67–expressing MM cells in the BM. The doses of 5-Fu and GEM used in these studies are significantly lower than doses typically administered to cancer patients (seven times lower, 180 vs. 1250 mg/m² body surface area per week) (36).

Depletion of MDSCs by 5-Fu and GEM overcomes BTZ resistance by restoring anti-MM immunity in the BM of OB-Runx2^−/− mice

To determine if 5-Fu therapy overcomes Runx2 deficiency-induced BTZ resistance by suppressing the activity of MDSCs and restoring the anti-tumor activity of cytotoxic T cells in the BM, flow cytometry was used to analyze immune cells in the BM of 5TGM1-Luc MM tumor-bearing OB-Runx2^+/+ and OB-Runx2^−/− mice treated with PBS, BTZ, 5-Fu, or BTZ+5-Fu for 4 weeks. Compared with PBS or BTZ, treatment with 5-Fu or BTZ+5-Fu significantly reduced MDSC abundance (Figure 6A), including M-MDSCs and PMN-MDSCs (Supplementary Figure S5A–B), and the expression of iNOS, arginase 1, and IL-10 by these MDSCs (Figure 6B–D) in the BM of tumor-bearing OB-Runx2^−/− mice. Treatment with 5-Fu or BTZ+5-Fu also increased the total CD8⁺ T cell abundance in the BM of these mice, and among these CD8⁺ T cells, decreased the percentage expressing PD-1 or TIM-3 and increased the percentage expressing granzyme B or IFN-γ (Figure 6E–J). Notably, compared with PBS or BTZ, treatment with 5-Fu or BTZ+5-Fu also reduced the abundance of MDSCs in the BM of tumor-bearing OB-Runx2^+/+ mice, and treatment with 5-Fu alone significantly reduced the expression of iNOS and IL-10 in the BM MDSCs of these mice. However, neither 5-Fu nor BTZ+5-Fu significantly altered the MDSC expression of arginase 1, the abundance of CD8⁺ T cells, or the abundance of CD8⁺ T cells expressing exhaustion/activation markers in the BM of tumor-bearing OB-Runx2^+/+ mice (Figure 6A–J).

Figure 6. Treatment with 5-Fu depletes MDSCs and restores cytotoxic T cells in the BM of OB-Runx2^−/− mice.

Immunosuppressive cells and immune effector cells within the BM of OB-Runx2^−/− and OB-Runx2^+/+ mice were assessed by flow cytometry after 4-week treatment with PBS, BTZ, 5-Fu, or BTZ+5-Fu as depicted in Fig. 5 (n=5 mice/group). A, Percentage of MDSCs (Gr1⁺ CD11b^hi) in the BM cell population. B-D, Relative expression of iNOS (B), arginase 1 (C), and IL-10 (D) in total MDSCs (Gr1⁺ CD11b^hi). E, Gating strategy for identifying CD8⁺ T cells and CD8⁺ T cells expressing exhaustion and activation markers in the BM cell population. F, Percentage of CD8⁺ T cells among all CD3⁺ T cells (CD3⁺CD8⁺) in the BM. G-J, Percentage of CD8⁺ T cells in the BM expressing PD-1 (G), TIM-3 (H), granzyme B (I), or IFN-γ (J). Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.005, ****P < 0.001. ns, not significant.

To validate the effects of MDSC depletion on immune cells, BM immune cells of tumor-bearing OB-Runx2^−/− mice treated with PBS, BTZ, GEM, or BTZ+GEM for 4 weeks were also assessed by flow cytometry. Similar to treatment with 5-Fu, treatment with either GEM or BTZ+GEM reduced the expansion and activation of MDSCs in OB-Runx2^−/− mice to a greater extent than treatment with PBS or BTZ did (Supplementary Figure S6A–D). Treatment with GEM or BTZ+GEM also significantly increased the BM abundance of total CD8⁺ T cells and altered the abundance of CD8⁺ T cells expressing exhaustion/activation markers, with synergistic effects of the combinatorial therapy observed for some measures (Supplementary Figure S6E–I).

These results suggest that depletion of MDSCs by 5-Fu and GEM restored the anti-MM effects of cytotoxic T cells in the BM and alleviated MM resistance to BTZ induced by OB-Runx2 deficiency.

Discussion

We previously reported that MM cell-induced Runx2 deficiency in OBs alters the BM microenvironment and promotes MM dissemination (10,13). In the current study, using OB-Runx2^+/+ and OB-Runx2^−/− syngeneic MM mouse models and in vitro approaches, we further demonstrated that OB-Runx2 deficiency promotes MM resistance to BTZ and enhances immunosuppression in the BM. Mechanistically, TSP-1–mediated activation of TGF-β1 in the BM appears to contribute to both MDSC activation and MM cell survival, thereby promoting MM cell resistance to BTZ. Furthermore, our studies demonstrate that targeting immunosuppressive MDSCs with 5-Fu or GEM or/and inhibiting TSP-1–mediated TGF-β1 activation with SRI31277 can overcome OB-Runx2 deficiency-induced BTZ resistance in MM cells.

The host anti-tumor immune response is primarily mediated by cytotoxic CD8⁺ T cells (37,38), but CD8⁺ T cell function can be inhibited by immunosuppressive cells such as MDSCs (22,23). In addition, the immunosuppressive network in the MM microenvironment negatively affects the outcome of MM chemotherapy (3). Here, we found that the BM immune cell profile in OB-Runx2^−/− mice is characterized by a high abundance of immunosuppressive MDSCs and a low abundance of CD8⁺ T cells. Furthermore, CD8⁺ T cells from the BM of these mice appear to be exhausted. These analyses confirmed our previous findings that OB-Runx2 deficiency creates an immunosuppressive microenvironment (13). Interestingly, our results also suggest that, rather than eliminating MM tumor cells as it does in the context of OB-Runx2 sufficiency, administration of BTZ in the context of OB-Runx2 deficiency exacerbates immunosuppression in the BM by further enhancing the abundance and activation of MDSCs and reducing the abundance and cytotoxic activity of CD8⁺ T cells therein.

In addition, our findings indicate that TGF-β1 signaling is critically involved in the BM immunosuppression and MM resistance to BTZ induced by OB-Runx2 deficiency. TGF-β1 is a multifunctional immunosuppressive cytokine that upregulates MDSC proliferation and activation (39) and directly promotes tumor cell proliferation and drug resistance (29,30). TGF-β1 is synthesized as a latent precursor dimer comprising mature TGF-β1 and the latency-associated peptide (LAP). Multiple mechanisms, including binding to integrins or TSP-1, can convert latent TGF-β1 to its biologically active form by disrupting the interaction between the LAP and mature TGF-β1 (40–42). Previously, we showed that TSP-1 is a major regulator of latent TGF-β1 activation, MM progression, and osteolytic bone disease in both immune competent syngeneic and xenograft mouse models of MM (20). There is also evidence that TSP-1 control of TGF-β activation has a role in proliferation/anti-apoptosis in multiple immune cells, including Tregs, Th17 T cells, dendritic cells, and NK cells (31,43). In our current study, the BM of OB-Runx2^−/− mice had a significantly higher level of TSP-1 and active TGF-β1 than the BM of OB-Runx2^+/+ mice had, and treatment of OB-Runx2^−/− mice with BTZ further augmented these levels. Furthermore, studies with SRI31277, a tripeptide antagonist of TSP-1–mediated latent TGF-β1 activation, confirmed the involvement of TSP-1/TGF-β1 activation in the development of MM resistance to BTZ. SRI31277 treatment of MM tumor-bearing OB-Runx2^−/− mice effectively inhibited the expansion and activation of BM MDSCs, reversed the suppression of cytotoxic T cells, and restored the anti-MM efficacy of BTZ. By culturing MDSCs with BMS from OB-Runx2^+/+ or OB-Runx2^−/− mice and SRI31277, we further confirmed that TSP-1 activation of TGF-β1 has an important role in MDSC proliferation/ activation induced by OB-Runx2 deficiency. Our current work is the first evidence of a role for TSP-1/TGF-β1 activation in control of immunosuppression through MDSCs in MM. TSP-1 and TGF-β1 are mainly secreted by BM stromal cells, and this secretion is upregulated by cytokines such as IGF-1 and IL-6 (20,44,45). Our previous study showed high concentrations of IGF-1 and IL-6 in the BM of OB-Runx2^−/− mice (13), which may explain the increase in TSP-1 and TGF-β1 in the BM that occurs in OB-Runx2 deficiency.

Runx2 deficiency in MM tumor-bearing mice also altered the expression and activation of proteins that regulate the induction of MM cell apoptosis, which could confound BTZ-induced apoptosis; treatment with SRI31277 mitigated this effect and restored BTZ sensitivity. Therefore, we speculate that active TGF-β1 also acts directly on MM cells to promote BTZ resistance in response to Runx2 deficiency. This is consistent with observations that cancer-associated fibroblasts isolated from BTZ-resistant patients express high levels of TGF-β1, and these BTZ-resistant fibroblasts prevent MM cell apoptosis (46). BTZ treatment increases TGF-β1 activation and pro-survival autophagy in these resistant fibroblasts (46), consistent with our observation of increased TGF-β1 activation in the BM of the OB-Runx2^−/− model. TGF-β1 is also associated with drug resistance in other cancers, and SRI31277 has been shown to overcome drug resistance in thyroid carcinoma cells (47,48).

Finally, our study shows that 5-Fu- or GEM-induced depletion of MDSCs can restore BM immunity and the anti-MM efficacy of BTZ in the context of OB-Runx2 deficiency. These data demonstrate the key role of BM immunosuppression in OB-Runx2 deficiency-driven MM resistance to BTZ. 5-Fu and GEM are FDA-approved anti-tumor agents often used to treat solid tumors, but these agents are not commonly used as anti-MM drugs (49,50). The results described herein support further investigation into the potential use of 5-Fu and GEM in BTZ-resistant MM, particularly in the setting of elevated BM MDSCs induced by OB-Runx2 deficiency.

In conclusion, our findings demonstrate novel mechanisms for BTZ resistance of MM. OB-Runx2 deficiency, induced by MM cells (10), promotes BTZ resistance in MM cells through the upregulation of immunosuppressive MDSCs and downregulation of cytotoxic T cells within the BM, and treatment with BTZ exacerbates these effects. Moreover, increased activation of TGF-β1 induced by TSP-1 in OB-Runx2 deficiency and further increased by BTZ stimulation contributes to this pathologic process and directly promotes BTZ resistance in MM cells. Importantly, this study identified potential pharmacologic strategies to overcome BTZ resistance in MM patients through repurposing of 5-Fu or GEM to target MDSCs in the BM microenvironment or use of TSP-1/TGF-β antagonists.