Runx2 deficiency in osteoblasts promotes myeloma resistance to bortezomib by 1 increasing thrombospondin 1-mediated TGF- β1 activation and suppressing immunity in 2 the bone marrow

Background: Multiple myeloma (MM) is a plasma cell malignancy that thrives in the bone 33 marrow (BM). Although the proteasome inhibitor bortezomib (BTZ) is one of the most effective 34 front-line chemotherapeutic drugs for MM, 15–20% of high-risk patients do not respond to this 35 drug or become resistant to treatment. The mechanisms driving this chemoresistance remain 36 unclear. Previous studies showed that the tumor microenvironment contributes to cancer 37 chemoresistance. Our recent studies demonstrated that Runt-related transcription factor 2 38 (Runx2) deficiency in osteoblasts (OBs) creates a cytokine-rich and immunosuppressive 39 microenvironment in the BM and promotes MM progression. However, the impact of Runx2 40 deficiency in OBs on the efficacy of BTZ in treatment of MM is still unknown. 41 Methods: We assessed the effects of OB-Runx2 deficiency on the outcome of BTZ treatment in 42 OB-Runx2 +/+ and OB-Runx2 -/- mouse models of MM using bioluminescence imaging, ELISA and 43 flow cytometry. In addition, we used a co-culturing in vitro system to explore the mechanism of 44 BTZ resistance and assessed this system by MTT assays and Western blot analysis. 45 Results: We discovered that OB-Runx2 deficiency induces MM cells resistance to BTZ via the 46 suppression of immunity and increased active TGF-β1 in the BM. We further demonstrated that 47 depletion of myeloid-derived suppressor cells (MDSCs) by gemcitabine or inhibition of TGF-β1 48 activity by SRI31277, a compound that blocks thrombospondin 1 (TSP1)-mediated TGF-β1 49 activation, restores anti-tumor immunity in the BM and overcomes BTZ resistance induced by 50 OB-Runx2 deficiency. In addition, SRI31277 also directly inhibits the activity of canonical 51 (Smad2/3) and non-canonical (Erk1/2) signaling pathways of TGF-β1 in 5TGM1-Luc MM cells and sensitizes MM cells to BTZ, resulting in increased apoptosis of MM cells. OB-Runx2 deficiency promotes BTZ resistance in MM cells through the regulations of MDSCs and TSP1-mediated TGF- β1 activation in the BM. These data identify novel mechanisms of BTZ resistance in MM and suggest new strategies to overcome BTZ resistance in treatment of MM. These our previous that OB-Runx2 immunosuppression in these mice as detected by higher numbers and increased activation of MDSCs and fewer cytotoxic + T cells with less activity in Furthermore, our study shows that GEM-induced depletion of MDSCs in OB-Runx2-deficient mice restored BM immunity and the anti-MM efficacy of BTZ. These data demonstrate the key role that BM immunosuppression plays in BTZ resistance driven by OB-Runx2 deficiency and suggest that targeting immunosuppressive MDSCs can overcome this BTZ resistance. GEM is an FDA-approved anti-tumor agent, mostly used in solid tumors, and it is not commonly used as an anti-MM drug The results described herein support further investigation into the potential use of GEM in BTZ-resistant MM, particularly in the setting of elevated BM MDSCs induced by OB-Runx2 deficiency. restored BM immunity MDSCs GEM significantly overcomes OB-Runx2 deficiency- 335 induced BTZ resistance of MM by restoring BM immunity and by inhibiting Smad/Erk signaling 336 pathways of TGF-β1 in MM cells. This study identifies new potential strategies to overcome BTZ 337 resistance in MM patients and illustrates the continued need for combination studies that 338 address multiple aspects of the tumor microenvironment in MM.

7 markers that disable the CD8 + T cell response required to eliminate cancer cells (23, 24) (Fig. 2f,  158 g). 159 BTZ treatment did not appear to alter immune cells in the BM of OB-Runx2 +/+ mice, 160 whereas BTZ significantly worsened anti-MM immunity in OB-Runx2 -/mice. Compared to PBS-161 treated controls, BTZ treatment in OB-Runx2 -/mice resulted in increased Tregs (Fig. 2a) and 162 MDSCs ( Fig. 2b) with greater MDSC activation (Fig. 2c), but fewer CD8 + T cells and less 163 activation ( Fig. 2d-g). 164 Our previous study also showed that the level of total TGF-β1 in BM is upregulated by 165 OB-Runx2 deficiency (13). To investigate the effect of BTZ on TGF-β1 with BM of OB-Runx2 166 deficiency, we measured the concentrations of active TGF-β1 and its activator TSP1 in BM 167 supernatants from OB-Runx2 +/+ and OB-Runx2 -/tumor-bearing mice treated with either PBS or 168 BTZ for 4 weeks. BTZ treatment did not change the levels of active TGF-β1 and TSP1 in the BM 169 of OB-Runx2 +/+ mice, which is similar to data obtained using human MM cells in in vitro studies 170 (18). Interestingly, BTZ treatment resulted in significantly higher levels of BM active TGF-β1 (Fig.  171 2h) and TSP1 (Fig. 2i) in OB-Runx2 -/mice than in mice treated with PBS. 172 These results indicate that the BTZ treatment further increases OB-Runx2 deficiency-173 induced expansion and activation of MDSCs, immune suppression, active TGF-β1, and its 174 potential activator TSP1 in the BM. 175

Depletion of MDSCs overcomes BTZ resistance in MM induced by OB-Runx2 deficiency 176
Our previous findings showed that MDSCs have a central role in OB-Runx2 deficiency-177 induced immune suppression in the BM (13). Because BTZ treatment worsens OB-Runx2 178 deficiency-induced immunosuppression, we depleted BM MDSCs with i.p. GEM injection in OB-179 Runx2 +/+ and OB-Runx2 -/mice to determine whether MDSC depletion can restore BM immunity 180 and overcome BTZ resistance. Compared to treatment with PBS or single agent BTZ or GEM, 181 treatment with the combination of GEM + BTZ resulted in the greatest reduction in tumor burden 182 8 (Fig. 3a-c) and the fewest Ki-67 + and BCL-2 + MM cells (Fig. 3d, e) in both OB-Runx2 +/+ and OB-183 Runx2 -/mice. Importantly, GEM treatment in OB-Runx2 -/mice successfully overcame MM BTZ 184 resistance induced by OB-Runx2 deficiency. 185 Combining GEM and BTZ restores anti-tumor immunity in the BM of OB-Runx2 -/mice 186 To determine if the GEM + BTZ combination therapy overcomes Runx2-deficiency-187 induced chemoresistance by restoring BM immunity, we used flow cytometry to analyze the 188 number and activity of the immune cells in the BM harvested from OB-Runx2 +/+ and OB-Runx2 -/-189 mice treated with PBS, GEM, BTZ, or GEM + BTZ. GEM treatment indeed resulted in fewer 190 MDSCs and less activity (measured by iNOS, Arginase 1 and IL-10 expression in MDSCs) in 191 OB-Runx2 -/mice, and these reductions were more significant in GEM + BTZ-treated OB-Runx2 -192 /mice than in mice treated with PBS ( Fig. 4a-d). GEM + BTZ-treated OB-Runx2 +/+ mice also 193 had reduced MDSC expansion and activation, but the reduction in OB-Runx2 -/mice was higher 194 than that of OB-Runx2 +/+ mice (2.06-fold vs. 1.28-fold in mean MDSCs number compared to its 195 respective BTZ treatment) (Fig. 4a). In addition, GEM + BTZ treatment in OB-Runx2 -/mice had 196 the greatest effect on BM T cell subpopulations. This drug combination resulted in significantly 197 fewer Tregs (Fig. 4e), more CD8 + T cells (Fig. 4f), as well as increased production of cytotoxic 198 markers granzyme B, IFN-γ, and perforin in CD8 + T cells ( Fig. 4g-i) with less expression of the 199 exhaustion markers PD-1 and TIM-3 in CD8 + T cells (Fig. 4j, k). These results suggest that GEM 200 in combination with BTZ significantly restored BM immunity and alleviated MM BTZ resistance 201 induced by OB-Runx2 deficiency. 202

BTZ in OB-Runx2 -/mice 204
Since BTZ treatment in OB-Runx2 -/mice results in significantly higher levels of TSP1 205 and active TGF-β1 in BM supernatants than in the PBS-treated control mice, but does not affect 206 the BM concentration of TSP1 and active TGF-β1 in OB-Runx2 +/+ mice, we asked whether 207 9 TSP1 might be involved in controlling latent TGF-β1 activation and whether TSP1-regulated 208 TGF-β1 is important in BTZ resistance in the altered BM microenvironment of OB-Runx2 -/mice. 209 To address this question, we treated OB-Runx2 -/mice with SRI31277, a compound that blocks 210 TSP1-mediated TGF-β1 activation (18), in combination with BTZ in OB-Runx2 -/mice. In 211 previous studies, we showed that SRI31277 blocked myeloma progression and bone 212 destruction and when combined with BTZ, both drugs had an enhanced inhibitory effect on 213 myeloma (18). Five-week-old OB-Runx2 -/mice were i.v. injected with 5TGM1-Luc MM cells one 214 week before being treated with PBS, BTZ, SRI31277, or SRI31277 + BTZ for 4 weeks. The 215 combination treatment of SRI31277 and BTZ resulted in significantly lower tumor burden as 216 shown by serum IgG2bκ ELISA than treatment using either BTZ or SRI31277 alone (Fig. 5a). 217 Furthermore, there were fewer CD138 + total MM cells and Ki-67 + proliferating MM cells, but 218 more cleaved-caspase-3 + apoptotic MM cells present in the BM of the SRI31277 + BTZ treated 219 group than in BM from any single reagent-treated group (Fig. 5b-d). These results demonstrate 220 that SRI31277 effectively reversed BTZ resistance in MM induced by OB-Runx2 deficiency. 221 To explore the mechanisms of the action of SRI31277, we first assessed the effect of 222 SRI31277 on BM immune cells. Compared to either PBS or BTZ treatment, treatment with 223 either SRI31277 or SRI31277 + BTZ resulted in significantly less expansion and activation of 224 MDSCs in OB-Runx2 -/mice ( Fig. 5e, f). However, there was no significant difference in MDSC 225 expansion and activation detected between the SRI31277 and SRI31277 + BTZ treated groups, 226 suggesting that the primary effect on MDSCs is attributed to SRI31277 action independent of 227 BTZ sensitivity (Fig. 5e, f). T cell subpopulations had a similar pattern of response: treatment 228 with SRI31277 + BTZ resulted in significantly more total CD8 + T cells and greater numbers of 229 granzyme B, IFN-γ, and perforin expressing cytotoxic CD8 + T cells, but fewer PD-1 and TIM-3 230 expressing CD8 + T cells than treatment with BTZ alone (Fig. 5g-

SRI31277 inhibits the TGF-β1 pathway in MM cells and directly alleviates MM cell 235 resistance to BTZ treatment in vitro 236
Next, we determined whether increased TGF-β1 in the BM supernatant of OB-Runx2 -/-237 mice can directly promote MM cell resistance to BTZ and whether blocking TGF-β1 activation 238 and downstream signaling by SRI31277 can sensitize MM cells to BTZ treatment. We collected 239 BM supernatants from the femur and tibia of OB-Runx2 +/+ and OB-Runx2 -/mice (without tumor 240 cell injection) (n= 3-4 mice/group). 5TGM1-Luc MM cells were cultured in medium mixed in a 1:1 241 ratio with PBS (control) or BM supernatants from OB-Runx2 +/+ or OB-Runx2 -/mice for 24 h. Cell 242 viability was measured using MTT assays. 5TGM1-Luc MM cells was sensitive to BTZ when co-243 cultured with the BM supernatants from OB-Runx2 +/+ mice, but MM cell viability was not affected 244 by BTZ treatment (2.5 nM) when co-cultured with the BM supernatants from OB-Runx2 -/mice 245 ( Fig. 6a and Supplementary Fig. S1a). The addition of SRI31277 to 5TGM1-Luc MM cells 246 cultured with BM supernatants from OB-Runx2 -/mice resulted in fewer viable cells than 247 treatment with PBS; and this inhibition of MM cell growth was further enhanced when SRI31277 248 was combined with BTZ in the culture ( Fig. 6a and Supplementary Fig. S1b, S1c). Western blot 249 analysis showed that caspase-3 expression was higher in 5TGM1-Luc MM cells that were co-250 cultured with the BM supernatants of OB-Runx2 -/mice and treated with SRI31277 + BTZ than 251 in cells treated with either BTZ or SRI31277 alone (Fig. 6b, c). 252 Finally, to explore molecular mechanisms of SRI31277 in alleviating MM cell resistance 253 to BTZ treatment, we used Western blot analysis to assess the activity of canonical (Smad2/3) 254 and non-canonical (Erk1/2) signaling pathways of TGF-β1 (25, 26) in 5TGM1-Luc MM cells. 255 Compared to corresponding PBS treatment, BTZ did not affect TGF-β1 signaling in 5TGM1-Luc 256 cells that were cultured in BM supernatants of either OB-Runx2 +/+ mice or OB-Runx2 -/mice.

11
However, BTZ + SRI31277 treatment on 5TGM1-Luc cells that were cultured in the BM 258 supernatants of OB-Runx2 -/mice resulted in significantly less phosphorylation of Smad2/3 and 259 Erk1/2 than treatment with either PBS or BTZ alone. In contrast, BTZ + SRI31277 did not result 260 in significant changes in Smad2/3 and Erk1/2 signaling in 5TGM1-Luc cells that were cultured 261 with the BM supernatants of OB-Runx2 +/+ mice (Fig. 6d-f). 262 These data indicate that the increased active TGF-β1 in the BM supernatants (induced 263 by OB-Runx2 deficiency) promotes MM cell resistance to BTZ by upregulating TGF-β1 signaling 264 pathways (Smad2/3 and Erk1/2) in MM cells, and that inhibiting TSP1-mediated TGF-β1 265 activation using SRI31277 can reverse this BTZ resistance. 266

Discussion 267
We previously reported that MM cell-induced Runx2 deficiency in OBs can alter the BM 268 microenvironment and promote MM progression (12,13). In these current studies, using OB-269 Runx2 +/+ and OB-Runx2 -/syngeneic MM mouse models, we showed that OB-Runx2 deficiency  Furthermore, CD8 + T cells from the BM of these mice are less cytotoxic and more exhausted. 12 These analyses confirmed our previous findings that OB-Runx2 deficiency creates an immune 282 suppressive microenvironment that favors MM progression (13). 283 Interestingly and surprisingly, the administration of BTZ to OB-Runx2 -/mice further 284 enhanced immunosuppression in these mice as detected by higher numbers and increased 285 activation of MDSCs and Tregs, and fewer cytotoxic CD8 + T cells with less activity in the BM.

TGF-β1 is a multifunctional immunosuppressive cytokine that upregulates MDSC and 295
Treg activation (29-32). TGF-β1 also directly promotes tumor cell proliferation and drug 296 resistance (33, 34). TGF-β1 is synthesized as a latent precursor dimer comprising mature TGF-297 β1 and the latency-associated peptide (LAP). Multiple mechanisms, including binding to 298 integrins and TSP1, can convert latent TGF-β1 to its biologically active form by disrupting LAP 299 interactions with the mature domain (35-37). Active TGF-β1 binds with high affinity to receptors 300 that activate the canonical Smad pathway and also non-canonical Smad-independent pathways 301 such as those mediated by p38 mitogen-activated protein kinases (MAPK), phosphoinositide 3-302 kinase (PI3K)-AKT and JUN N-terminal kinase (JNK) (38). Through these signaling pathways, 303 TGF-β1 plays an important role in cell proliferation, apoptosis, and drug resistance (33, 34). 304 TGF-β has been shown to be important for MM progression and the osteolytic bone disease 305 characteristic of MM (39). Previously, we showed that TSP1 is a major regulator of latent TGF-β 306 activation, MM progression, and osteolytic bone disease in both immune competent syngeneic 307 13 and xenograft mouse models of MM (18). SRI31277 is a tripeptide derived from the LSKL 308 sequence of the LAP of latent TGF-β, and this peptide acts as a competitive antagonist to block 309 TSP1-mediated TGF-β1 activation in MM (18, 40). We found that the BM of OB-Runx2 -/mice 310 had a significantly higher level of active TGF-β1 than the BM of OB-Runx2 +/+ mice and that BTZ 311 treatment augmented the level of BM-active TGF-β1 in OB-Runx2 -/mice. Moreover, SRI31277 312 treatment in OB-Runx2 -/mice bearing 5TGM1-Luc tumors effectively inhibited the expansion 313 and activation of BM MDSCs, and successfully restored BM immunity and anti-MM efficacy of 314 BTZ. These results demonstrate that TSP1-mediated TGF-β1 activation is responsible for OB-315 Runx2 deficiency-induced MM BTZ resistance via suppressing BM immunity. Importantly, 316 SRI31277 is effective at overcoming BTZ resistance through inhibiting TGF-β1 activation and 317 restoring BM immunity. Treg: regulatory T cell; TSP1: thrombospondin-1. 346

Ethics approval and consent to participate 347
All animal studies were performed in accordance with University of Alabama at Birmingham 348 (UAB) and National Institutes of Health (NIH) guidelines after institutional review and approval. 349 Every effort was made to minimize the numbers and suffering of the included animals. 350

Consent for publication 351
Not applicable 352