Oral TGF-βR1 inhibitor Vactosertib promotes osteosarcoma regression by targeting tumor proliferation and enhancing anti-tumor immunity

Osteosarcoma (OS) is an aggressive malignant bone cancer, with refractory and metastatic disease remaining a significant challenge. Transforming growth factor-β1 (TGF-β) is a potent immune suppressive cytokine in OS and the TGF-β is increased in the sera of OS patients and this increase is associated with high-grade OS and lung metastases. Therefore, blocking TGF-β1 signaling may be a novel therapy for OS treatment. Here we show that blocking TGF-β1 signaling using TGF-βR1 inhibitor, Vactosertib, significantly inhibited OS proliferation in vitro and in vivo. Notably, Vactosertib inhibits c-Myc expression in the OS cells. Vactosertib increased immune effectors (IFNγ+CD8+ cells and NK cells) and inhibited immune suppressors (M2-like TAM, MDSC) in the OS tumor microenvironment. Our results suggest that inhibition of TGF-β1 signaling is an effective therapeutic strategy against OS through a multi-pronged approach that targets tumor intrinsic and extrinsic factors to achieve optimal immune-effector functions and maximal clinical response.


64
While pOS is a well-documented major cause of OS-related deaths, the exact molecular dynamically suppress the function of tumor-reactive cytotoxic T cells and natural killer (NK) cells 78 (10, 11). Previous publications have reported that one of the most important contributions of 79 myeloid cells in the TME is increased TGF-β1 production. The subsequent TGF-β1 signaling 80 pathway in myeloid cells has been found to be required for tumor metastasis (12). It has been 81 reported that TGF-β1 expression is increased in the sera of OS patients compared to those of 82 healthy donors (13) and the overexpression of TGF-β1 is observed in OS tissues (14). This 83 increase in TGF-β1 production is correlated with high-grade OS (14, 15) and the presence of 84 lung metastases (16, 17). Also, TGF-β1 stimulates the growth of several OS cell lines (18)(19)(20)(21).

85
The existing literature strongly suggests TGF-β1 is a critical cytokine that promotes OS 86 development, arguing for a strong need to critically elucidate how TGF-β1 acts on OS cells and 87 evaluate whether blocking TGF-β1 can be a novel therapeutic approach for treating OS. Several

95
Vactosertib is a highly selective and a potent small molecule inhibitor against Type 1 TGF-β 96 animal cancer models, including lung metastasis and melanoma mouse models (26,27). In this 104 current study, we first defined the ability of Vactosertib to inhibit OS in vitro and in vivo and 105 investigated the mechanism(s) through which it suppresses OS tumor growth and enhances 106 immune cell activation within the TME. We show that mouse and human OS cell growth are

119
It has been reported that TGF-β1 expressions is highly expressed in clinical OS tissue (10). Using 120 publicly available expression cohorts, we analyzed the expression levels of TGF-β1 and its 121 correlation with overall survival in OS patients and found that a high expression of TGF-β1 was 122 associated with worse survival (Suppl. Figure 1), supporting the notion that TGF-β1 is a critical 123 cytokine in OS disease progression and further supporting the hypothesis that effectively blocking 124 TGF-β1 signaling can be a novel therapeutic approach for this disease. To further test this

150
Since we identified that Vactosertib inhibited Myc target genes in the RNA-seq analysis above, 151 we investigated whether TGF-β1 could further upregulate Myc pathways in these cells, and if that 152 up-regulation could be inhibited by Vactosertib co-treatment in OS cells. To this end, SAOS2 cells 153 were treated with TGF-β1 (5 ng/ml) alone or co-treated with Vactosertib (100 nM) for 24 hours.

154
GSEA analysis of the RNA-seq. dataset found the most significantly increased gene sets after

170
To further elucidate the role of TGF-β1 gene regulation in OS, RNA-seq. analysis was 171 performed comparing untreated (UT), TGF-β1 single treatment, and TGF-β1 + Vactosertib co-172 treatment in SAOS2 cells. GO enrichment analysis of the statistically significant differentially 173 expressed genes showed that TGF-β1 down-regulated genes involved in biological processes 174 related to bone formation related biological processes such as ossification and extracellular matrix 175 production (Suppl. Figure 2A). Conversely, Vactosertib + TGF-β1 co-treatment yielded the 176 opposite results (Suppl. Figure 2C). These results imply that TGF-β1 has a significant effect on 177 bone formation and osteoblasts (33). We also observed that TGF-β1 up-regulated genes are 178 associated with ribonucleoprotein complex biogenesis and ribosome biogenesis (Suppl. Figure   179 2B), both of which were significantly inhibited by co-treatment with Vactosertib (Suppl. Figure   180 Figure 3C-D). But the 213 percentage and total number of Arg + PDL1 + cells, which were gated on CD11b + F480 + cells, were 214 significantly suppressed in Vactosertib treat tumors (Suppl. Figure 3E).
(13, 14), we investigated the effects of TGF-β1 inhibition on pOS using an established pOS model 219 in immunocompetent mice. BALB/c mice were inoculated with 1x10 6 K7M2-Luc cells (i.v.). 7 days 220 after tumor injection, mice were treated with Vactosertib via oral gavage (25 mg/kg once daily for 221 5 days with 2 days off) and monitored weekly for tumor growth by bioluminescence imaging (BLI).

241
To further elucidate the immune landscape within the OS TME after inhibition of TGF-β1 in vivo,

242
we performed multiparametric flow cytometry with t-distributed stochastic neighbor embedding after tumor injection) and treated with Vactosertib via oral gavage (50 mg/kg once daily for 5 days 245 with 2 days off) for 6 weeks. To verify our manual gating approach of the deep profiling T cell 246 panel and to visualize these data in two dimensions, lung samples of mice were stained with 247 antibodies to CD3, CD4, CD8, CD49b, Foxp3, PD-1 and IFNg and analyzed by tSNE analysis 248 (Suppl. Figure 4A and Figure 6A). The generated tSNE was classified into 31 clusters ( Many of these clusters express CD206 which is used as a marker for M2 macrophages.

299
Vactosertib (50 mg/kg p.o. 5 days/week) was administrated 3 weeks later with or without weekly 300 ICB administration (α-PD-1 or α-PD-L1 mAb, 100ug/mouse/week, i.p.), and OS tumor burden was 301 monitored by BLI ( Figure 9A). The BLI intensity 11 weeks after tumor injection is shown in Figure   302 9B. Vactosertib alone inhibited OS tumor growth significantly as expected (Figure 9A and B). with isotype control antibody, the NK population was reduced to 6.5%. When these mice were 313 treated with Vactosertib alone or Vactosertib + a-PD-L1 co-treatment the NK population was 14.3% 314 and 22%, respectively, which was significantly higher than that of isotype treatment ( Figure 9D).        anti-PD1 mAb in vivo (Fig. 9). The exact reasons for this lack of clinical efficacy by targeting PD1 390 in OS awaits additional studies. In our studies, we observed in vivo efficacy of targeting PD-L1 in 391 our mouse OS model; however, the response to anti-PD-L1 monotherapy is similar to oral 392 Vactosertib alone (Fig. 9). Surprisingly, co-treatment of Vactosertib with ICB (anti-PD-1/anti-PD-

444
The cleaned sequencing reads were aligned to reference genome using HISAT2 package.

505
In addition to conventional FACS analysis, high-dimensional clustering using t-distributed

812
Graph of BLI over time. n=5, **p<0.01 using a two-way ANOVA between groups followed by post-