Macrophage Syk–PI3Kγ Inhibits Antitumor Immunity: SRX3207, a Novel Dual Syk–PI3K Inhibitory Chemotype Relieves Tumor Immunosuppression

This article is featured in Highlights of This Issue, p. 729

Macrophages (MΦ) play a broad role in host defense but can also serve as major drivers of tumor growth, metastasis, and immunosuppression, observed in the tumor microenvironment (TME; refs. 1, 2). In response to various environmental signals produced by tumor and stromal cells, proinflammatory MΦs shift to immunosuppressive phenotype that block antitumor immunity (3). Hence, targeting the molecular pathways/signaling nodes in the MΦs that regulate transition of protumorigenic MΦs into anti-tumorigenic MΦs will activate immune response in cancer. Although recent studies shed some light on the molecular entities involved in the activation of protumorigenic MΦs (4–6), the identification of potential druggable targets and small molecules to control MΦ immunosuppression of antitumor immunity are still needed.

Syk kinase is a well-established cytoplasmic protein tyrosine kinase implicated in inflammation and hematopoietic cell responses, including integrin and immunoreceptor tyrosine activation motif (ITAM) signaling (7–9). As a modulator of tumorigenesis, the role of Syk is highly controversial, it acts as tumor promoter in some cancers (10, 11), and in others it is tumor suppressor (12, 13). Syk kinase has been extensively studied in adaptive immune responses, but its role in MΦ-mediated innate immune responses remains unclear (14). Previous studies by our group revealed Syk kinase as a novel component of α₄β₁integrin–Rac2 signaling axis that encodes myeloid and endothelial Rac2 specificity in vivo (8, 15). Within the TME, Syk kinase functions upstream of Rac2 GTPase and PI3K to modulate integrin (α_vβ₃/α_vβ₅&α₄β₁)-mediated migration and metastasis in vivo (16) and these reports suggest a role for Syk kinase in regulating MΦ polarization and immunosuppression.

Another target for negative regulation of antitumor immunity recently discovered and reported by our group and other laboratories is the PI3K signaling pathway, in particular p110γ in MΦs (6, 17). Our laboratory has reported that p110γ in MΦs promotes expression of protumorigenic MΦs in TME and our pan PI3K/BRD4 inhibitor SF1126 or SF2523, blocked tumor growth and MΦ-mediated immunosuppression in tumors (17, 18). Furthermore, Syk kinase is required for activation of PI3K in MΦs and B cells (19, 20). These reports lead us to propose a hypothesis, that targeting two crucial signaling entities that promote MΦ-mediated immunosuppression viz. Syk kinase and PI3K will activate the antitumor immune response. With this aim, using computational chemistry methods, our laboratory has developed a novel chemotype, SRX3207 that inhibits both PI3K and Syk, with a single molecule for maximal activation of adaptive immune responses.

In this article, using a genetic approach and pharmacologic blockade, we provide evidence that Syk kinase plays a crucial role in the control of MΦ-mediated immunosuppression and the inhibition of antitumor immunity. Moreover, our novel dual inhibitory chemotype, SRX3207 blocks both PI3K and Syk in MΦs thereby activating innate and adaptive antitumor immunity in vivo.

All procedures involving animals were approved by the UCSD Animal Care Committee, which serves to ensure that all federal guidelines concerning animal experimentation are met. Floxed Syk mice and lysozyme M (LysM) Cre recombinase transgenic mice were purchased from The Jackson Laboratory. Integrin α4Y991A mice and normal littermates in C57BL/6J genetic background have been described before (16, 21). BMDMs were isolated as described previously (16). For hypoxia experiments, MΦs were placed in a modulator incubator chamber (Billups-Rothenberg) as described before (16).

Lewis lung carcinoma (LLC), B16 melanoma, and CT26 cells were obtained from the ATCC and were cultured in DMEM or RPMI media containing 10% FBS. B16-OVA cells were obtained from Dr. A. Sharabi (University of California, San Diego, CA). All cell lines were tested for Mycoplasma and mouse pathogens and checked for authenticity against the International Cell Line Authentication Committee (ICLAC; http://iclac.org/databases/cross-contaminations/) list. LLC or B16 or B16-OVA or CT26 (1 × 10⁵) cells were injected subcutaneously into syngeneic mice and were treated with 40 mg/kg R788 administered orally or 10 mg/kg IPI549 or SRX3207 orally, starting from day 10 when tumors reached 100 mm³ until tumors were harvested on day 21. In another experiment, B16-OVA cells were injected in C57BL/6 WT mice and when tumors reached 100 mm³, mice were treated with 200 μg anti-PDL1 antibodies either alone or in combination with 40 mg/kg R788 as described in Supplementary Methods. CD8 depletion and MΦ depletion experiments were performed as described earlier (18) and in Supplementary Methods.

Tumors were isolated, minced, and then enzymatically dissociated in collagenase digestion cocktail at 37°C for 30–45 minutes and cells were prepared for magnetic bead purification of CD11b, or CD90.2 cells or for flow cytometry as reported before (16, 18). For mass cytometry, cells were subsequently stained with metal-labeled antibodies cocktail and run on mass cytometer (CyTOF, Fluidigm) as described in Supplementary Methods.

Total RNA was isolated from BMDMs and TAMs using the Qiagen RNAeasy kit (Qiagen) and cDNA was amplified by RT-PCR as described before (16). For RNA sequencing, RNA libraries were prepared and sequenced on Illumina HiSeq2000 using standard Illumina protocols described in Supplementary Methods.

To profile open chromatin, ATAC-seq was performed on LPS and IL4-stimulated BMDMs. The cells were submitted to UCSD Center for Epigenomics. The details are available in Supplementary Methods.

X-ray structures of human Syk, ZAP70 and PI3K/p110α and PI3K/p110γ (PDB codes: 4XG9, 1U59, 4JPS and 4XZ4, respectively) were obtained from the Protein Data Bank. Detailed description of in silico design of SRX3188 and SRX3207 is provided in Supplementary Methods. Detailed description of synthesis of SRX3188 and SRX3207 has been described before, where compound 1 and 3 refers to SRX3188 and SRX3207, respectively (22). PK/PD and ADME properties of the compounds were studied in collaboration with Quintara Discovery.

MΦs play an important role in promoting tumor growth and in establishing an immunosuppressive microenvironment that dampens effective T-cell responses in tumors (23). Recently, our laboratory and others reported that PI3Kγ is one of the targets that is a major driver of tumor growth and immunosuppression (Fig. 1A and B; refs. 6, 17, 24). Herein, we identify that Syk in MΦs promotes tumor immunosuppression. Given our previous reports that MΦ-specific Syk kinase functions upstream of Rac2 and downstream of α₄β₁integrin receptor (16), we sought to evaluate the role of myeloid Syk in regulating tumor growth. We generated conditional Syk knockout mice (Syk^MC-KO) in which Syk expression was controlled by LysM cre recombinase, a myeloid cell-specific promoter. Conditional deletion of Syk as confirmed by immunoblotting of BMDMs clearly demonstrate Syk expression only in wild-type (Syk^MC-WT) MΦs and not in Syk^MC-KO MΦs (Fig. 1C). Furthermore, our results in Fig. 1C and D, show that Syk kinase is expressed only in BMDMs, CD11b⁺ F4/80⁺ tumor-associated MΦs (TAM), CD19⁺ B cells and minimally in CD90.2⁺ T cells, with no expression of Syk in the murine syngeneic tumor cells used in these experiments. To investigate the functional role of MΦ Syk kinase in tumor growth, we used 2 different syngeneic tumors that included: LLC and B16 melanoma. Tumor growth was reproducibly significantly reduced in Syk^MC-KO animals compared with Syk^MC-WT animals (Fig. 1E). Most notably, intravenous injection of B16 tumor cells resulted in significantly reduced lung tumor metastasis in Syk^MC-KO (Fig. 1F) and p110γ^−/− animals reported earlier by our laboratory (17).

To evaluate the immunologic changes in the LLC tumors, CD45⁺ hematopoietic cells were profiled from Syk^MC-WT and Syk^MC-KO mice using mass cytometry (cytometry by time of flight, CyTOF). Analysis of the total CD45⁺ leukocytes using Phenograph clustering (25) revealed 18 different clusters (Fig. 1G and H; Supplementary Fig. S1). Among those clusters, we found slight expansion of CD4⁺ (cluster C18), CD8⁺ (cluster C13) T cells in Syk^MC-KO tumors with no change in B-cell population (cluster C17). Among myeloid cell population, we observed no changes in the CD11b⁺F4/80⁺ MΦs (cluster C2), or CD11b⁺F4/80⁺CD80⁺CD86⁺CX3CR1⁺CD64⁺ tissue resident MΦs (Cluster C16) or CD11b⁺ CD86⁺CD11c⁺MHCII⁺ or CD11c⁺MHCII⁺ dendritic cells (cluster C9, C15; Fig. 1H). Interestingly, we found slight expansion in CD11b⁺Ly6C⁺ classical monocytes (cluster C6), CD11b⁺F4/80⁺CD80⁺CD86⁺MHCII⁺ immunostimulatory MΦ population (cluster C5) in Syk^MC-KO tumors by comparison with decrease in immunosuppressive MΦs, CD11b⁺F4/80^hiCD206⁺TGFb⁺CD64⁺ (cluster C11) and myeloid-derived suppressor cells or neutrophils, CD11b⁺LY6C⁺Ly6G⁺ (Cluster C8 and C14). Although, the changes in different myeloid populations in Syk^MC-WT and Syk^MC-KO tumors did not reach to any statistical significance but it provides the evidence that deletion of myeloid Syk showed shifting of immunosuppressive MΦ polarization toward immunostimulatory MΦ population in Syk^MC-KO tumors that might lead to inhibition of tumor growth and metastasis observed in these KO mice.

We next investigated whether Syk deletion in MΦs alters expression of genes associated with immunosuppression and tumor progression. The expression of genes that mediates immunosuppression in TME (6, 18) were significantly higher in TAMs isolated from LLC tumors grown in Syk^MC-WT animals compared with that of Syk^MC-KO animals (Fig. 2A). In contrast, expression of immunostimulatory genes (6, 18) was significantly enhanced in the TAMs from Syk^MC-KO animals compared with that of Syk^MC-WT tumors (Fig. 2A). Moreover, increased arginase activity (P < 0.01) and decreased nitrite production (NOS; P < 0.001) was observed in TAMs isolated from Syk^MC-WT LLC tumors compared with those isolated from Syk^MC-KO LLC tumors (Fig. 2B). RNA-seq data on LLC TAMs suggest that the expression of genes that mediates immunosuppression in TME, for example, Vegf, Tgfb, Ido2, Myc, Ccl8, Ccl28 are expressed high in Syk^MC-WT and on the contrary, genes involved in antigen presentation and innate immunity are expressed high in Syk^MC-KO (Fig. 2C). Interestingly, deletion of Syk did not affect the infiltration of TAMs in the LLC tumors (Cluster C2, Fig. 1H), but it increased the expression of major histocompatibility complex (MHC) class II (Fig. 2D) and immunostimulatory cytokines and decreased the expression of immunosuppressive genes in TAMs (Fig. 2A and C), suggesting that Syk regulates immunosuppression.

To determine whether Syk kinase controls transcriptional changes in MΦs, we tested the effect of Syk deletion on lipopolysaccharide (LPS) polarized and IL4 polarized bone marrow–derived macrophages (BMDM). It is well documented that LPS induce MΦ expression of T_H1 cytokines, whereas IL4 signaling stimulates T_H2 response (26). Genes associated with immune stimulation were upregulated in LPS or IL4 stimulated Syk^MC-KO MΦs, whereas genes associated with immunosuppression were downregulated in these MΦs (Supplementary Fig. S2A–S2D). Taken together, these results confirm that Syk plays a major role in promoting the immunosuppressive MΦ epigenetic/transcriptional program.

To investigate the mechanism by which Syk regulates MΦ immune responses, ATAC seq was performed. In the IL4 exposed group, most significant immune-related motifs (AP-1, AR, C/EBPB, EGR1, EGR2, HIF2A) were found to be highly enriched in Syk^MC-WT MΦs (Fig. 2E). Conversely in the LPS group, the significantly enriched motifs (AP-1, and NF-κB) were mostly found in the Syk^MC-KO MΦs. Existing literature suggests that NF-κB promotes expression of proinflammatory cytokines whereas HIF1α and HIF2α are the transcription factors involved in immunosuppressive MΦ differentiation and suppression of T-cell function (27–30). Hence, we determined whether genetic deletion of Syk affects the phosphorylation of p65 RelA or stability of hypoxic HIF1α or HIF2α. In consistent with ATAC-seq results, we found that deletion of Syk kinase stimulated and sustained p65 RelA phosphorylation, and destabilized hypoxic HIF1α and HIF2α (Fig. 2F and G). Together, these results suggest that in addition to PI3Kγ (6), Syk is another molecular switch that controls immunosuppression or immune activation.

To investigate whether the deletion of MΦ Syk promotes the adaptive immune response, we evaluated the recruitment and activation of CD8⁺ T cells that play an important role in antitumor immune responses. Flow cytometric analysis of LLC tumors demonstrate that recruitment of total CD8⁺ T cells increased in Syk^MC-KO tumors (Fig. 3A and B), without significantly altering T-cell recruitment in the spleen (Supplementary Fig. S3A and S3B). To validate that MΦ Syk inhibits adaptive immune responses, TAMs isolated from Syk^MC-WT and Syk^MC-KO were mixed in 1:1 ratio with LLC tumor cells and were adoptively transferred into different Syk^MC-WT and Syk^MC-KO mice (Supplementary Fig. S3C). WT MΦs adoptively transferred into Syk^MC-WT and Syk^MC-KO mice showed increase in tumor growth, whereas adoptive transfer of KO MΦs suppressed tumor growth in Syk^MC-WT and Syk^MC-KO mice (Fig. 3C). In addition, Syk inhibition did not reduce tumor growth in CD8-depleted mice, suggesting that Syk inhibition reduces tumor growth by recruiting and activating CD8⁺ T cells (Fig. 3D). Moreover, we found that CD90⁺ T cells isolated from Syk^MC-KO LLC tumors expressed significantly more Ifng and Gzmb as compared with that from WT tumors (Fig. 3E), suggesting that deletion of Syk kinase activates T-cell–mediated antitumor immune response in vivo. RNA-seq analysis performed on LLC tumors implanted in Syk^MC-KO tumors showed an increase in Prf, Gzm, Klre1 and several other genes that are involved in activation of cytotoxic T cells (Fig. 3F). We did not observe any difference in the proliferative capacity of T cells between Syk^MC-WT and Syk^MC-KO animals (Supplementary Fig. S4D). Taken together, our results demonstrate that Syk kinase inhibition in MΦs indirectly promotes cytotoxic-adaptive immune responses in the tumor.

Given that the conditional deletion of Syk kinase inhibits tumor growth and intratumoral immunosuppression, we speculated that pharmacologic inhibitors of Syk could similarly inhibit tumor growth and induce an antitumor response. To test this, we used commercially available specific Syk kinase inhibitor, Fostamatinib (R788; ref. 31). Tumor growth and metastasis were significantly suppressed in the mice treated with Syk inhibitor compared with vehicle-treated tumors (Fig. 4A). Importantly, R788 had no effect on viability of LLC cells (Supplementary Fig. S4). Pharmacologic blockade of Syk kinase significantly inhibited expression of immunosuppressive genes and increased expression of immunostimulatory genes in TAMs purified from LLC tumors (Fig. 4B). Moreover, R788-treated LLC tumors showed increased infiltration of CD8⁺ T cells with no significant reduction in number of regulatory T cells (Fig. 4D). Interestingly, we observed increased CD44^hiCD62L^lo effector T cells (gated on CD8⁺T cells) and cytotoxic T cells in R788-treated tumors (Fig. 4E and F), and these findings are not due to differences in T-cell proliferation as we observed no effect of R788 on T-cell proliferation ex vivo (Supplementary Fig. S5A).

To determine whether Syk suppresses antitumor adaptive immune response in vivo, we injected B16-OVA cells in C57BL/6 mice and characterized OVA-specific CD8⁺ T cells, in R788-treated tumors. Using H-2K^b tetramers containing the OVA protein-derived peptide SIINFEKL, we found that R788 treatment significantly blocked tumor growth and increased OVA-specific CD8⁺ T cells in the tumor draining lymph nodes (Supplementary Fig. S5B and S5C). These results provide evidence that Syk kinase suppress the antigen-specific adaptive immune response in the TME in vivo.

To explore the possibility that Syk inhibitors might synergize with T-cell check point inhibitors, we determined the effect of R788 in combination with anti-PDL1 in B16 melanoma cells. Both anti-PDL1 and R788 substantially inhibited B16-OVA tumor growth. Although the combination of R788 and anti-PDL1 showed no significant additive effect on Syk inhibition (Supplementary Fig. S5B) in resistant B16 model, but CD44^hi and CD62L^lo effector T cells were significantly increased in the combination treated group compared with monotherapy-treated groups (Supplementary Fig. S5C–S5E). Although, the combination of R788 and anti-PDL1 did not work well in B16 model, further in vivo synergistic studies are ongoing in our laboratory to explore the efficacy of this combination regimen in other cancer models. Taken together, these results validate Syk as an immuno-oncology drug candidate with equivalent in vivo activity compared anti-PDL1 blockade in this immune competent model.

The concept of combinatorial therapeutics recently emerged as effective strategy to synergistically activate antitumor immunity (18, 32–34). Our laboratory has developed a drug discovery paradigm that uses computational methods to engineer a single chemotype to inhibit two separate targets in one small molecule and demonstrated proof of concept for greater antitumor efficacy in vivo (18, 34). Herein, we apply this technology to two immuno-oncology targets, PI3K (6) and Syk kinase (data presented in this article).

We used multiple X-ray crystallographic datasets in the form of PDB files of Syk, ZAP70, PI3K/p110α and PI3K/p110γ (PDB 4XG9, 1U59, 4JPS and 4XZ4, respectively) to engineer a small-molecule chemotype that will contain two distinct “warheads” within the thienopyranone scaffold (35) to bind to the ATP binding pockets of PI3K isoforms and Syk kinase. Fig. 5A shows the docking of SRX3188 in Syk active site that showed that this inhibitor is engaged via hydrogen-bond interactions with the amide and carbonyl groups of Ala451, the sulfur atom of methionine Met450, and the hydroxyl group and Arg498, whereas the azetidine nitrogen is making a charge interaction with the carboxylate group of Asp512 in the same way the co-crystallized inhibitor in 4XG9 is observed, confirming the expected binding mode and necessary hydrogen-bond interactions. When modeling SRX3188 in ZAP70 it was observed that, contrary to staurosporine which in the crystal structure of ZAP70 is found to make 2 key hydrogen bond interactions with Glu415 and Ala417, SRX3188 does not manifest this interaction predicting no affinity toward ZAP70, a tyrosine kinase homologous to Syk.

Modeling of SRX3188 in PI3K/p110α indicates that the TP core of SRX3188 binds as expected establishing a hydrogen bond interaction between the morpholine group and Val815, the pyrazole ring forming a π-π T-shaped interaction with Trp780, and the azetidine nitrogen is making a charge interaction with the carboxylate group of Glu798 (Fig. 5B). Modeling of SRX3188 in PI3K/p110γ also showed an expected hydrogen bond interaction between the morpholine group and Val882, and the predicted π-π T-shaped interaction of the pyrazole ring with Trp812 also observed with PI3K/p110α (Fig. 5C). It was also predicted the hydroxyl group of the azetidine group occupies a region where no additional interactions are engaged with this kinase.

Because the proximity of the hydroxyl group of the azetidine group to the Syk's backbone and the nonfavorable electrostatic microenvironment surrounding the hydroxyl group were a concern, the hydroxyl group was removed leading to the design and synthesis of SRX3207. In silico docking studies of SRX3207 against Syk indicated this analog would bind to Syk in the same way as SRX3188 does. Similarly, docking results of SRX3207 against PI3K/p110α and PI3K/p110γ indicate similar binding modes while exhibiting no affinity toward ZAP70. The final chemotype designed was synthesized and tested in cell free systems for potency against Syk, ZAP70 and PI3K isoforms p110α, p110β, p110γ and p110δ (Fig. 5D). The first-generation chemotype, SRX3188 had excellent Syk inhibitory activity (IC₅₀ of 20 nmol/L) but poor PI3K inhibitory potency. Extensive modeling of the p110α and p110γ crystal structures (PDB 4JPS and PDB 4XZ4, respectively) enabled a structure activity relation analysis (SAR) to develop the second-generation prototype inhibitor, SRX3207 with potent Syk (30 nmol/L) and acceptable PI3K inhibitory properties in vitro. Fig. 5D shows the chemical structure of SRX3188 and SRX3207.

Results in Fig. 5E show that SRX3207 is a dual Syk kinase/PI3K inhibitor that hits both targets in the same molecule at nmol/L potency with minimal off-target effects (Fig. 5E, Supplementary Table S1). Supplementary Figure S6A shows that both SRX3188 and SRX3207 potently blocks phosphorylation of Syk at 348 site and Y525/526 site that are required for complete activation of Syk, described in detail in next section. Moreover, SRX3207, is able to block p-AKT at 10 μmol/L concentration (Supplementary Fig. S6A). Finally, we profiled SRX3207 against a library of 468 known kinases that demonstrated a selectivity index (SI) of 0.079 (Supplementary Fig. S6B) documenting a high-level selectivity of this in silico engineered dual inhibitory compound. SRX3207 has selectivity score of 8 that is defined as number of kinases hit by the molecule with scores less than 1 [S (1); ref. 36)]. Because of both the high potency and excellent kinase selectivity of SRX3207 an oral prototype formulation of SRX3207 was prepared using Pharmatek's Hot Rod formulations to be used in preliminary in vivo studies. Preliminary pharmacokinetic studies in mice via both the intravenous and oral route indicated a half-life of about 5 hours but with a low bioavailability of only around 2% (Fig. 5F and G). Despite this being a nonoptimized formulation the oral administration of this prototype formulation of SRX3207 yielded significant antitumor activity (Fig. 6). Additional ADME studies indicate SRX3207 has sufficient solubility in water (43 μmol/L) relative to its target potencies but suffers from a metabolic liability with a CLint (μL/min/mg protein) of 74. Further studies are in progress to optimize the prototype SRX3207 molecule from an ADME standpoint while preserving the desired potencies on the Syk and PI3K target enzymes. In the meantime, the proof-of-concept benefits from having a single molecule with both Syk and PI3K inhibition properties are illustrated with the prototype molecule SRX3207.

We next determined whether SRX3207 blocks immunosuppressive MΦ polarization. Interestingly, we found that SRX3207 potently blocks immunosuppressive gene expression effectively at very low dose compared with commercially available p110γ inhibitor (IPI-549) or Syk inhibitor (R488) alone (Supplementary Fig. S6C and S6D). Moreover, SRX3207 at 10 mg/kg dose blocked tumor growth and increased survival effectively as compared with IPI549 and R788-treated groups at similar doses without toxicity (Fig. 6A and B; Supplementary S7A). Furthermore, it reduced immunosuppressive MΦ polarization and increased infiltration and cytotoxicity of CD8⁺ T cells in LLC tumors and increased expression of Ifng and Gzmb in SRX3207 treated LLC tumors (Fig. 6C–E; Supplementary S7B). Most notably, SRX3207 did not affect T-cell proliferation (Supplementary Fig. S7C). Moreover, administration of SRX3207 did not block CT26 tumor growth in NSG mice but blocked tumor growth in Balb/c mice suggesting that SRX3207 blocks tumor growth due to its effect on immune compartment (Supplementary Fig. S7D and S7E). Moreover, depleting MΦs with anti-CSF1R (CD115) antibody treatment alone significantly blocked LLC tumor growth, but the combination with R788 or SRX3207 had no additive effects (Supplementary Fig. S7F). Taken together, these results validate the efficacy of this novel dual Syk/PI3K inhibitor in blocking immunosuppression and activating the adaptive immune response and opened new opportunities to explore it in combination with check point inhibitors.

In MΦs, Syk kinase is known to be activated upon binding to phosphorylated ITAMs of Fc gamma receptor, or by binding to the cytoplasmic domains of integrin adhesion receptors, most notably β₁ and β₃ integrin (9). Our results in Supplementary Fig. S8A show that only α₄β₁ integrin can maximally activate Syk at Y348 site in MΦs isolated from the TME (Supplementary Fig. S8A), and BMDMs isolated from α₄Y991A mice are defective in phosphorylating Syk at Y348 site (Supplementary Fig. S8B). These results suggest that Syk is phosphorylated downstream of α₄β₁integrin and mediates tumor growth and immunosuppression. Interestingly, Fig. 6F illustrates that administration of R788 did not show additive reduction in tumor growth in Syk^MC-KO mice, whereas IPI549 or SRX3207 significantly decreased tumor growth in the Syk^MC-KO mice. These results suggest that Syk and p110γ are two separate molecular entities that promote tumor growth and their simultaneous inhibition with single molecule is good strategy to reduce tumor growth and immunosuppression. Our published results (17) and data presented here have shown that genetic or pharmacologic deletion of p110γ and/or Syk results in hypoxic degradation of HIF1/2α and this effect is completely blocked by MG132, a proteasome inhibitor (Supplementary Fig. S8C). Figure 6G shows the schematic of Syk and p110γ regulation on tumor immunosuppression.

TAMs are reported as the most abundant immune cells in the TME of solid tumors, where they release immunosuppressive factors that inhibit T-cell–mediated antitumor immune response (37). Therapeutic approaches that are aimed at blocking tumor immunosuppression by inhibiting MΦ polarization have shown great efficacy in murine cancer models (38, 39) and open new and effective strategies to treat cancer.

In this report, we focused on two major immuno-oncology targets, namely Syk kinase (reported in this article) and PI3Kγ (6, 17). We provided several lines of evidence to prove that inhibition of Syk kinase is associated with reduction of tumor growth, MΦ-mediated immunosuppression and activation of adaptive immune responses in TME (Figs. 1,–4). Although recent study has shown the efficacy of Syk inhibitor in solid tumors (40), but the role of MΦs in activating adaptive immune responses has never been reported. Our results have shown that Syk kinase is minimally expressed in T cells and inhibition of this kinase increases infiltration and activation of CD8⁺ T cells with reduction in percentage of CD4⁺T cells, which can be explained by reduction in the number of immunosuppressive CD4⁺ regulatory T cells (Treg). Our initial preliminary results have shown decrease in number of Tregs in R788-treated tumors but the results did not reach to significance (Fig. 4D). Recent studies have shown that similar to dendritic cells, MΦs can also prime CD8⁺ T cells to generate cytotoxic effector cells in vivo (41). Our results in Figs. 2 and 3 clearly provide evidence that Syk kinase inhibition in MΦs indirectly promotes cytotoxic adaptive immune responses and represents a major immune-oncology target.

Herein, we describe the in silico design and synthesis of a novel dual Syk–PI3K inhibitor, SRX3207 that potently inhibited Syk and PI3K signaling and augmented the antitumor immune response in lung carcinoma tumor model with no toxicity (Fig. 6). Taken in context (16, 17, 42), the results presented here elucidate the mechanistic interconnection of integrin α₄β₁, p110γ, Syk, HIF axis in MΦs. In addition, data provided in Fig. 6F provide the rationale to target both Syk and p110γ with single molecule for effective antitumor immunity and to explore SRX3207 in combination with checkpoint inhibitors in cancers driven by MΦ-mediated immunosuppression.

A.B. Sharabi is an advisory board member for Astrazeneca, Jounce Therapeutics and reports receiving commercial research grant from Varian Medical Systems and Pfizer, and has ownership interest (including patents) in Toragen Inc. G.A. Morales is a director R&D and has ownership interest (including patents) in SignalRx Pharmaceuticals. J.R. Garlich is a chief scientific officer and has ownership interest (including patents) in Signalrx Pharmaceuticals Inc. D.L. Durden has ownership interest (including patents) in SignalRx Pharmaceuticals Inc. No potential conflicts of interest were disclosed by the other authors.

Conception and design: S. Joshi, G.A. Morales, J.R. Garlich, D.L. Durden

Development of methodology: S. Joshi, K.X. Liu, T.V. Pham, D. Shiang, D.L. Durden, P.D. Sanders

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K.X. Liu, M. Zulcic, A.R. Singh

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Joshi, K.X. Liu, D. Skola, C.K. Glass, A.B. Sharabi, T.V. Pham, P. Tamayo, H.Q. Dinh, C.C. Hedrick, D.L. Durden

Writing, review, and/or revision of the manuscript: S. Joshi, K.X. Liu, A.B. Sharabi, C.C. Hedrick, D.L. Durden

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Joshi, A.B. Sharabi, G.A. Morales, J.R. Garlich, D.L. Durden

Study supervision: S. Joshi, D.L. Durden

Other (study design of mass cytometry experiments): C.C. Hedrick

This work was supported by R01 CA215651 (to D.L. Durden).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.