TTK inhibitor OSU13 promotes immunotherapy responses by activating tumor STING

TTK (MPS1) spindle assembly checkpoint kinase is an emerging cancer target. This preclinical study explored the anti-tumor mechanism of TTK inhibitor OSU13 to define a strategy for clinical development. We observed prominent anti-tumor activity of OSU13 in melanoma, colon, and breast cancer cells, melanoma patient-derived organoids, and mice bearing colon tumors associated with G2 cell cycle arrest, senescence, and apoptosis. OSU13-treated cells displayed DNA damage and micronuclei that triggered the cytosolic DNA-sensing cGAS-STING pathway. STING was required for the induction of several proteins involved in T cell recruitment and activity. Tumors from OSU13-treated mice showed an increased proportion of T and NK cells and evidence of PD-1/PD-L1 immune checkpoint activation. Combining a low-toxicity dose of OSU13 with anti-PD1 checkpoint blockade resulted in prominent STING-and CD8 T cell-dependent tumor inhibition and improved survival. These findings provide a rationale for utilizing TTK inhibitors in combination with immunotherapy in STING-proficient tumors.


Introduction
Monopolar spindle kinase 1 (MPS1, or TTK) is an essential component of the spindle assembly checkpoint (SAC).The SAC ensures proper chromosome segregation during mitosis and meiosis by halting division until sister chromatids make it to opposite ends of the dividing cell (1).Dysfunctional SAC may lead to the accumulation of chromosome segregation errors and DNA damage.Cancer cells divide frequently and, therefore, are sensitive to SAC inhibition.Drugs targeting essential components of the SAC, such as TTK, are promising emerging therapeutic tools in cancer treatment.
Several TTK inhibitors (TTKi) have been developed and tested in human trials (2)(3)(4)(5), but none has been approved for clinical use to date.While TTKi show a clear effect on tumor cell growth and viability in vitro, the extent of their in vivo activity may not be sufficient to justify the development of single-agent TTKi therapeutics.For instance, in a mouse xenograft study, TTKi CFI-402257 did not outperform single-agent carboplatin therapy (6).Similarly, S81694, which was tested as a single agent in adults with solid tumors, showed relatively mild toxicities but primarily induced disease stabilization (7).The study was terminated early due to the prioritization of using S81694 in combination with additional cytotoxic agents.This decision was based on preclinical data that TTK inhibition augments responses to microtubule poisons paclitaxel and docetaxel that emerged during the trial (8)(9)(10).Without functional SACs, microtubule poisons induced chromosomal segregation defects and death of tumor cells in preclinical models.A combination of TTK inhibitor BAY1217389 and paclitaxel was recently tested in a Phase 1 clinical trial in solid tumors (11).However, severe hematological and other toxicities were observed in combination-treated patients.Other studies tested similar regimens of the dual TTK/PLK1 inhibitor BAL0891 as monotherapy and in combination with carboplatin or paclitaxel (NCT05768932).Another TTK inhibitor, CFI-402257, was tested in combination with paclitaxel in patients with advanced HER2-negative breast cancer (NCT03568422).Interim reports from these studies indicate some evidence of efficacy; however, overall reported response rates are not high (about 11-14%).Furthermore, there are toxicities reported, including fatigue, neutropenia, anemia, alopecia, diarrhea, and nausea (12)(13)(14)(15).Overall, the clinical utility of paclitaxel and TTKi combinations remains to be demonstrated.
Aside from paclitaxel and other chemotherapies, there is an ongoing investigation of hormonal therapy in combination with TTK inhibitors.For instance, the NCT02792465 study evaluated CFI-402257 as monotherapy and in combination with fulvestrant in a phase 1 trial in patients with advanced solid tumors, including HER2-negative breast cancer.This study reported early signs of anti-tumor activity with a manageable safety profile, both as a monotherapy and in combination with fulvestrant in patients with ER-positive/HER2-negative breast cancer who have failed CDK4/6 inhibitors (16).The follow-up study is currently ongoing (NCT05251714).
OSU13 is a small molecule inhibitor of TTK designed by computer-assisted docking analyses.Initial studies showed an antiproliferative effect of OSU13 in breast cancer cell lines (17) and in human multiple myeloma cells and xenografts (18).It has been shown to have high selectivity for TTK over almost 400 other human kinases in the kinome scan assay (17).Here, we focused on the mechanism of action of OSU13 in solid tumors, particularly its ability to enhance tumor immune recognition.We hypothesized that OSU13 would promote tumor infiltration by the immune cells based on our experience with other antiproliferative drugs.For example, inhibitors of cyclin-dependent kinases 4 and 6 (CDK4/6i) that are used to treat metastatic ER+ breast cancer can induce the production of T cell-attracting chemokines in tumors and promote T cell responses (19,20).Similarly, a drug targeting the activity of a mitotic kinase, Aurora A, can facilitate immune cell infiltration into tumors by inducing the secretion of chemokine CCL5 from tumor cells.The release of immune cell-attracting chemokines results from senescence and the senescence-associated secretory phenotype (SASP) induced in tumor cells under prolonged drug-mediated cell cycle arrest and damage (21)(22)(23).Senescence is a type of stress response characterized by a loss of proliferative potential and the increased secretion of pro-inflammatory mediators, growth factors, stromal components, and other molecules that affect the local microenvironment (24,25).This damage-associated pro-inflammatory response in tumors may promote their recognition by the immune system, which could be beneficial in the context of immunotherapy, a type of cancer treatment that enhances the activity of anti-tumor immune cells.This includes immune checkpoint blockade (ICB) therapy that targets inhibitory receptors on T cells or their ligands expressed in tissues (26).Unlike many other cancer treatments that only provide short-lived benefits, the effects of ICB can be long-lasting.However, many tumors are resistant to ICB because they are ignored by the immune system.Combining ICB with drugs promoting tumor immune recognition can overcome ICB resistance, which is an urgent and unmet clinical need (27).Our results provide strong rationale for clinical development of OSU13 and ICB combination to enhance anti-tumor immunity.

Results
The objective of this study was to investigate the anti-tumor efficacy of a TTK inhibitor OSU13.To determine the direct anti-tumor activity of OSU13, we treated colon, melanoma, and breast cancer cells with various concentrations of OSU13 or vehicle for 1, 3, 5, and 7 days and counted the cells.There were significantly fewer cells after treatment with OSU13 compared to vehicle control, and this effect was dose and time-dependent (Fig. 1A, B).To define the mechanism of tumor cell inhibition by OSU13, we employed a flow cytometry-based assay measuring apoptosis, replication, and DNA damage.Approximately half of the cells treated with OSU13 displayed cleavage of PARP, indicative of apoptosis (Fig. 1C, D).The remaining live cells failed to incorporate BRDU, indicating an arrested cell cycle (Fig. 1C, D).A more detailed analysis of cell cycle distribution revealed decreased proportion of cells in G1 and S phases and an accumulation of cells in G2 (Fig. 1E, F).There was an increased proportion of cells with <2n amount of DNA (sub-G1), which is a marker of apoptosis.Consistent with inhibition of the cell cycle, we observed a decreased expression of the cell cycle protein Cyclin A2 and induction of cell cycle arrest mediator p21 after OSU13 treatment (Fig. 1G).Furthermore, many OSU13treated cells had high levels of phosphorylated histone H2AX (γH2AX), a DNA damage marker (Fig. 1 C, D).Persistent cell cycle arrest in conjunction with DNA damage is often observed in cells undergoing senescence.Accordingly, we detected a prominent induction of senescenceassociated β-galactosidase activity after OSU13 treatment in all three tested cancer lines (Fig. 1H, I).Altogether these results show that OSU13 induced DNA damage, G2 cell cycle arrest, senescence, and apoptosis in cancer cells.
Our next objective was to test if TTK is the main cell target of OSU13.Previously, our OSU colleagues have shown that OSU13 has high selectivity for TTK over almost 400 other human kinases in the cell-free kinome scan assay (18).Specifically, OSU-13 half maximal inhibitory concentration (IC50) was shown to be the lowest for TTK (4.3 nM).The only other kinase that was targeted by OSU13 with a single-digit nM concentration was LRRK2 (7.5 nM).Therefore, we tested OSU13 in a LRRK2 NanoBRET™ assay to determine whether OSU13 can prevent the binding of an ATP analog to the kinase of interest (LRRK2) in mammalian cells.The receptor protein tyrosine kinase inhibitor CEP-701 was used as the positive control.Though our colleagues previously showed that OSU13 achieves an EC50 of 10 nM in a TTK NanoBRET™ in-cell target engagement assay (17), its EC50 in the analogous LRRK2 NanoBRET™ in-cell target engagement assay was 216 nM, indicating at least 20-fold selectivity for TTK over LRRK2 in the cellular environment (Fig. S1A).These data suggest that TTK is the predominant target of OSU13 in mammalian cells.
TTK is a key regulator of the spindle assembly checkpoint (SAC).Cells without functional TTK can proceed with division even if the spindle assembly is faulty.We have performed a spindle assembly checkpoint (SAC) inhibition assay to verify that OSU13 targets the SAC.Nocodazolearrested CAL-51 human breast adenocarcinoma cells treated with OSU13 or the reference TTK inhibitor BOS-172722 showed a dose-dependent decrease in phosphorylated Histone H3 (Ser10), indicating the ability of OSU13 to effectively bypass the activated spindle assembly checkpoint despite the presence of aberrant mitotic spindles (Fig. S1B).
To test the anti-tumor effect of OSU13 in a clinically relevant model, we used melanoma patientderived organoids previously developed by our group (23,28).Organoids originating from four melanoma patients were treated for 5 days with OSU13, and live/dead fluorescent viability assay was performed.We detected a statistically significant reduction in cell viability after OSU13 treatment in all tested models (Fig. 2A, B).We next tested potential immunogenic properties of OSU13.We showed above that OSU13treated cells display features of senescence (Fig. 1).In addition to the loss of proliferative potential, senescent cells are known to display senescence-associated secretory phenotype (SASP), which involves increased secretion of various molecules, including immunoreactive molecules, such as cytokines and chemokines (29).We have previously shown that chemokines CCL5 and CXCL10 play an important role in recruiting immune cells into senescent tumors (19,21,22).Thus, we tested if OSU13 promotes CCL5 and CXCL10 secretion by tumor cells using ELISA.We observed a dose-and time-dependent induction of chemokines in all three tested cell lines (Fig. 3A-D).
Chemokines are commonly regulated on a transcription level.Accordingly, we detected an increase of CCL5 and CXCL10 mRNA in OSU13-treated cells (Fig. 3E).We also tested the ability of OSU13 to activate NF-κB and IRF transcription factors that are major regulators of chemokines and inflammatory transcriptional programs.Indeed, the activity of IRF and NFκB reporters increased in the B16 dual reporter cells after OSU13 treatment (Fig. 3F).These innate immune pathways are triggered by pattern recognition receptors in response to pathogens and danger signals.Based on the knowledge that TTK is important for chromosome segregation, we postulated that OSU13 might trigger inflammatory transcription by inducing micronuclei, which are extranuclear DNA-containing structures that form by unsegregated chromosome fragments or lagging chromosomes.As predicted, we observed an increased proportion of cells with micronuclei after OSU13 treatment (Fig. 3G, H).Micronuclei and DNA damage are reported to activate the STING-cGAS pathway, which is a pattern-recognition receptor pathway responsible for detecting and responding to cytosolic DNA (30,31).Notably, OSU13 induced micronuclei and DNA damage in cancer cells (Fig. 3G, H, Fig. 1C, D).Accordingly, we detected increased phosphorylation of downstream STING pathway nodes, such as STAT1 and TBK1, in OSU13treated cells (Fig. 3I).In addition, the induction of IRF reporter was diminished in the presence of STING inhibitor (Fig. 3J).Furthermore, the induction of IRF reporter and CCL5 secretion by OSU13 was abrogated in mouse melanoma and human colon cancer cells with Sting gene knockout (Fig. 3J, K).To confirm that STING is responsible for OSU13-induced inflammatory phenotype in human cancer cells, we generated two A375 human melanoma cell lines with STING knockout using CRISPR-Cas9 editing (Fig. S2).Multiplex ELISA analysis revealed that STING knockout cells displayed a distinct secretome compared to parental STING wild type cells (Fig. 4A).This was further validated by a principal component analysis, where OSU13treated STING-KO cells were stratified from OSU13-treated WT cells on the PCA plot (Fig. 4B).To gain further insight into the role of STING in regulation of OSU-13-triggered inflammatory phenotype, we selected proteins differentially secreted by OSU13-treated STING WT and STING KO cells for further analysis (Fig. 4C).Only proteins that were up-or downmodulated in both STING KO clones with p value <0.1 were chosen (Table S1).The induction of T cell-recruiting chemokines CXCL10, and CXCL11 was abrogated by STING knockout.Similarly, cytokines known to promote cytotoxic T cell responses, such as Type III Interferons IL28A and IL29 were not induced in the absence of STING.In addition, loss of STING also limited the OSU13-mediated induction of MIF that suppresses anti-inflammatory effects of glucocorticoids, which play pivotal role in tumor immune evasion (32).Another intriguing effect of STING knockout was a complete loss of the treatment-induced IL9, which has been implicated in anti-tumor immunity and immunotherapy responses (33)(34)(35)(36).
Interestingly, there were a number of proteins that displayed increased secretion in OSU13treated cells with STING KO compared to STING WT cells, suggesting that STING signaling down modulates their induction in response to stress.One such protein was CXCL8 (IL8), a chemotactic cytokine for granulocytic cells implicated in tumor promotion (Fig. 4C).We also observed a decrease of the secretion of GP130 (IL6ST) in STING KO cells.Soluble GP130 inhibits signaling from IL6 and other tumor-promoting cytokines (Fig. 4C).Of note, secretion of IL6 and IL8 is a key characteristic of senescent cells and an integral component of senescenceassociated secretory phenotype (SASP) linked to chronic inflammation and tumor promotion (37).Thus, the observed modulation of IL8 and IL6 signaling by STING loss suggests that STING activation may limit tumor-promoting effects of senescence and SASP.
Using a knowledgebase of reported and predicted protein-protein interactions we constructed a network connecting proteins differentially secreted by OSU13-treated WT and STING KO cells (Fig. 4D).List of proteins used as input is provided in Table S2.IL8 (CXCL8) was identified as the most interconnected protein within the network, suggesting its potential importance in STING-mediated regulation of the phenotype of OSU13 response.Altogether, the network analysis suggested that STING regulates secretory phenotype of tumor cells in response to OSU13 to enhance anti-tumor immunity and limit tumor-promoting inflammation.
To determine if OSU13 has anti-tumor and pro-immunogenic activity in vivo, we established subcutaneous tumors in immunocompetent C57BL/6 mice using MC38 murine colon carcinoma cells.Treatment was administered daily without a break for 9 days by oral gavage.We noted diarrhea developing after a week of dosing in the OSU13 group, and several mice in this group died before the experiment's completion.We adjusted the schedule to 5 days on and 2 days off in all of the following experiments to minimize toxicity.There was a prominent anti-tumor effect with the continuous dosing of OSU13, with statistically significant tumor growth inhibition (Fig. 5A, C) and reduction in the final tumor weight (Fig. 5B).Spectral cytometry analysis revealed an enrichment of NK and CD8 T cells and a reduction of regulatory T-cells within the total CD45+ tumor immune infiltrate in the OSU13-treated mice (Fig. 5D, E).This conclusion was validated using unbiased clustering analysis of the spectral cytometry data that showed enrichment of clusters of cells expressing NK and CD8 T cell markers in tumors from OSU13-treated mice (Fig. S3A, B).Traditional gating based on NK1.1+ and CD8+ surface markers also confirmed an increase in the proportion of NK and CD8 T cells infiltrating tumors in OSU13-treated mice (Fig. 5F).On the contrary, the proportion of Foxp3+CD4+ (Tregs) cells were significantly reduced in mice treated with OSU13 resulting in overall reduction in CD4+ T cells (Fig. 5G).On further analysis we found that OSU13 treatment significantly enhanced CD8-to-Treg ratio (Fig. 5G).
To determine whether NK cells or CD8 T cells recruited by OSU13 treatment contributed to anti-tumor activity of this therapeutic, we used the NK cell and CD8 T cell depletion approach (Fig. S4, A-D).Tumor inhibition by OSU13 was not dramatically affected by NK cell depletion.In contrast, OSU13-treated tumors grew faster in mice depleted of CD8 T cells compare to nondepleted controls (Fig. 5H).Based on these findings, we further investigated the phenotype of CD8+ T cells in OSU13-treated tumors.A greater proportion of tumor-infiltrating CD8+ T cells expressed the activation marker CD69 after OSU13 treatment compared to no-drug control.However, CD8+ T cells also displayed higher expression of exhaustion marker PD-1 and lower expression of proliferation marker Ki67+ after OSU13 treatment (Fig. 5I).The expression of CD69 and PD-1 on CD4 T cells was not significantly affected by OSU13, while the proportion of Ki67+ CD4 T cells was reduced (Fig. S4E, F).These data suggest that CD8 T cells were engaged in anti-tumor activity during OSU13 treatment; however, their activity was restrained by PD-1 immune checkpoint.Furthermore, the increase of PD-L1-expressing CD45+ cells in tumors of OSU13-treated mice also suggests the engagement of PD-1/PD-L1 immune checkpoint (Fig. 5J).Quantification and statistical analysis of the phenotype markers discussed above is shown on Fig. 5K.
To overcome the PD-1/PD-L1 immune checkpoint, we treated mice bearing MC38 tumors with OSU13 and checkpoint-blocking anti-PD-1 antibody.MC38 tumors are commonly used for studies of immune checkpoint blockade and have been shown to exhibit low-moderate response to anti-PD1 that is CD8 T cell-dependent (38).There was partial inhibition of tumor growth with anti-PD-1 alone and OSU13 alone.Notably, when both treatments were combined, there was complete inhibition of growth (Fig. 6A).Furthermore, mice treated with OSU13 and PD-1 combination displayed improved survival compared to other treatment groups (Fig. 6B).Mice that achieved complete response on OSU13 and anti-PD-1 therapy and remained tumor-free two months post-treatment were re-challenged with the same tumor.Only 1 out of four mice (25%) developed a tumor after a rechallenge.This contrasts with 83% of mice that developed tumors in the tumor-naïve group (Fig. 6C).This suggests a presence of immunologic memory against tumor antigens in mice that responded to OSU13 and PD-1 combination therapy.To confirm that the anti-tumor immune response in OSU13 and anti-PD-1-treated mice was CD8 T celldependent, as we previously observed with OSU13 alone (Fig. 5H), we compared MC38 tumor inhibition in CD8 T cell-depleted and non-depleted mice.The inhibitory effect of OSU13 and anti-PD-1 therapy was abrogated in CD8 T cell-depleted mice (Fig. 6D).We also tested OSU13 and anti-PD-1 combination in an alternative model where murine CT26 colon carcinoma tumors were grown in BALB/c mice.In this model, a combination of a 50 mg/kg OSU13 administered once a week with anti-PD-1 immunotherapy demonstrated statistically significant inhibition of tumor growth compared to vehicle and single-agent treatments and improved survival in mice (Fig. S5A, B).
Our in vitro data showed that the OSU13-mediated induction of interferon responses, chemokine CCL5, and several other pro-immunogenic factors in tumor cells was STING-dependent.Therefore, we next asked if the stimulation of anti-tumor immunity by OSU13 and anti-PD-1 therapy is dependent on STING activation in tumor cells.We injected mice with wild-type or STING knockout tumor cells and treated them with OSU13 and anti-PD-1.As expected, treatment abrogated the growth of STING wild type tumors (Fig. 6E).Out of 11 mice in the treatment group, 6 showed tumor regression to the point where they became unnoticeable/unmeasurable (1 mm or less in diameter).We also observed statistically significant growth inhibition in mice inoculated with STING knockout tumors.However, the anti-tumor effect was less potent.Furthermore, none of the treated mice showed prominent tumor regression (Fig. 6E).The partial growth inhibition of STING KO tumors could be explained by the direct inhibitory effect of OSU13 on tumor cells.These data suggest that activation of tumor cellintrinsic STING is a key mechanism of anti-tumor immunity stimulation by OSU13.
Finally, we investigated potential side effects that may occur with OSU13 treatment.Tumor-free C57Bl/6 (Fig. 7) and BALB/c (Fig. S6) mice were treated with 10 mg/kg OSU13 or drug vehicle for 3 weeks on a 5 days on/2 days off schedule.C57Bl/6 mice were also treated with anti-PD1 or anti-PD1 and OSU13 combination to test for potential unexpected toxicities from combining OSU13 with immunotherapy.There were no significant changes in body weight, indicating no severe GI toxicities with the tested dosing schedule (Fig. 7A, Fig. S6A).Serum levels of liver enzyme ALT were not significantly affected in any of the treatment groups and were within the normal range in C57Bl/6 and BALB/c mice.The levels of AST were slightly elevated in the C57Bl/6 mice; however, this increase was not statistically significant (Fig. 7B).The AST serum levels did not exceed the normal range in either of the groups in BALB/c mice (Fig. S6B).We also performed a complete blood count in C57Bl/6 mice testing for potential cytopenia.No changes to total white blood cells, lymphocytes, neutrophils, monocytes, eosinophils, basophils, and immature blood cells were detected in any of the treatment groups and all values were within the normal range (Fig. 7C-I).Finally, we tested for potential bone marrow toxicities (myelotoxicities).Spectral cytometry analysis of the bone marrow from C57Bl/6 and BALB/c mice revealed no major myelotoxicities, as the relative distribution of different myeloid cell populations were comparable in vehicle and any of the experimental treatment groups (Fig. 7J and Fig. S6C, D).Traditional gating also did not reveal statistically significant changes in the neutrophil, monocyte, or B cell populations within the bone marrow (Fig. 7K).
In summary, we show that OSU13 exhibited anti-tumor activity against various tumor types, stimulated CD8+ T cell-mediated anti-tumor immune responses, and improved responses to immune checkpoint blockade therapy.

Discussion
Here we investigated the mechanism of anti-tumor activity of a TTK inhibitor OSU13 to identify promising combination strategies for clinical development of this drug and its derivatives.Our results indicated a strong rationale for combining OSU13 with immune checkpoint blockade.Firstly, we found that OSU13-treated tumor cells activated a pro-immunogenic transcriptional program that translated to increased secretion of chemokines CCL5 and CXCL10, which are known to recruit activated T cells and other immune effectors to the sites of inflammation (39).Accordingly, mice treated with OSU13 displayed enrichment of immune effectors of the adaptive and innate immune system, such as CD8 T cells and NK cells, within the immune infiltrate of their tumors.Lack of immune infiltrate into the tumor microenvironment, specifically T cells, has been linked with poor patient prognosis and low likelihood of responding to ICB (40).Furthermore, it has been suggested that recruitment of T cells into the tumor during ICB is an important determinant of therapeutic outcome (41).
Secondly, OSU13-treated cells secreted proteins known to facilitate anti-tumor immunity, such as type III interferons IL28A and IL29.These cytokines display Type I interferon-like anti-viral and anti-tumor activities (42,43).Furthermore, OSU13 induced secretion of soluble gp130 from tumor cells.Gp130 is most known for its function as a transmembrane signal-transducing receptor that forms part of the receptor complex for several cytokines, including IL-6 (44).However, in its soluble secreted form, gp130 acts as an inhibitor of IL6 signaling (45,46).IL6 signaling drives the proliferation, survival, invasiveness, and metastasis of tumor cells, while strongly suppressing the anti-tumor immune response (47).Thus, secretion of an inhibitor of IL6 signaling by OSU13-treated cells is likely to limit tumor progression and promote anti-tumor immunity.In addition, OSU13-treated tumor cells produced IL9.IL-9 garnered increased attention based on a number of recent studies that demonstrated its powerful antitumor role in solid tumors.It is reported to have a direct inhibitory effect on tumor cell proliferation and metastasis and a tumor-extrinsic activity via promoting innate and adaptive anti-tumor immunity and immunotherapy response (33)(34)(35)(36).Conventionally, production of IL9 is attributed to Thelper 9 (Th9) cells.Thus, the fact that tumor cells themselves can secrete IL9 in response to drugs, such as OSU13, is intriguing and warrants further investigation beyond this study.
Of note, increased secretion is a key characteristic of cellular senescence, which is a type of stress response associated with damage and loss of proliferative potential that often occurs in tumors undergoing therapy.Senescence-associated secretory phenotype (SASP) is extremely diverse and can have pro-and anti-tumor effects depending on the damaging stimuli and a type of cells undergoing senescence.Cytosolic DNA sensor STING has been implicated as an inducer of SASP (48,49).Contrary to this notion, we observed increased secretion of immune-related proteins in both STING WT and knockout cells after OSU13 treatment.However, loss of STING completely abrogated secretion of the proteins with anti-tumor activity mentioned above, such as CXCL10, CXCL11, IL28A, IL29, IL9, and GP130.Thus, it is plausible that, in the context of OSU13 treatment, STING acted not as the inducer of SASP, but rather a SASP modifier that established a secretome favorable to tumor growth inhibition, tumor surveillance by T cells, and anti-tumor immune response.Biologically, the link between STING activation and cytotoxic immune responses makes sense.STING is a key factor of immune defense against DNA viruses (50).When viral DNA is present inside the infected cells, activation of STING would arrest those cells and promote their T-cell mediated killing.In contrast, after the virus is cleared and STING pathway is no longer triggered, it would be more beneficial to secrete factors that inhibit cytotoxic T cell responses and facilitate proliferation to promote tissue repair.
The third observation in support of combining OSU13 with immunotherapy came from the analysis of tumor immune infiltrate that revealed increased expression of PD-1 on T cells from OSU13-treated mice.We also observed that CD45+ cells from OSU13-treated mice expressed higher levels of PD-L1 on their surface.These phenotypic changes are indicative of activation of the PD-1-PD-L1 immune checkpoint that likely limited the amplitude of anti-tumor immune response induced by OSU13 treatment.Based on these three factors, we tested OSU13 in combination with anti-PD1 antibody and observed prominent T cell-dependent tumor inhibition and improved survival in mice.Notably, a study by another group reported improved response to PD-1 blockade in mice treated with a different TTK inhibitor CFI-402257, suggesting that the induction of pro-immunogenic phenotype by OSU13 is likely an on-target effect of TTK inhibition and, therefore, other TTK inhibitors are likely to compliment immunotherapy as well (6).
Our studies into the mechanism of pro-immunogenic activity of OSU13 revealed the essential role of the anti-viral cGAS-STING pathway.Recently it has become appreciated that the STING pathway can be induced by non-viral stimuli.For example, the presence of micronuclei, small DNA compartments encapsulated by a nuclear envelope and spatially separated from the primary nucleus, has been linked with STING activation.Rupture of micronuclei causes the release of the DNA content into the cytosol, where it is recognized by the STING activator cGAS (30).Furthermore, severe DNA damage has been reported to activate STING presumably due to the escape of broken DNA fragments from the nucleus (31).We showed that OSU13 induced micronuclei and DNA damage in tumor cells.This is likely to be an on-target effect, based on prior studies of other TTK inhibitors reporting micronuclei and STING activation (5,51).Interestingly, one study reported that TTK loss can cause chromatin bridges that also activate cGAS-STING pathway (52).Notably, we showed that genetic ablation of STING in tumor cells desensitized mice to OSU13 and anti-PD1 therapy.These data point to an essential role of the tumor-intrinsic STING in establishing effective anti-tumor immunity.
According to pan-cancer analysis, a subset of tumors (up to 25%) downregulates the STING pathway through epigenetic silencing of genes encoding STING or cGAS (53).While the exact biological outcome of STING loss in tumors is not fully understood, it has been suggested as a potential mechanism of tumor immune escape (53)(54)(55).Notably, a recent study showed that KRAS-LKB1 mutant lung cancer that often silences the STING pathway is resistant to TTKi and immunotherapy combinations, however epigenetic drugs can improve response by inducing STING re-expression (51).Considering the dichotomy of STING pathway silencing in tumors, the status of STING and cGAS could be used as biomarkers for patients' recruitment onto an OSU13 and anti-PD1 clinical trial.An effective biomarker would improve the response rates, prevent unnecessary treatment for trial participants, and increase the overall likelihood of establishing TTK inhibitors as an effective tool for boosting immunotherapy efficacy in clinic.This is an unmet clinical need as no reliable TTKi response biomarkers have been developed to date.In fact, a study of TTKi CFI-402257 revealed no correlation between treatment response and key cancer-associated mutations (e.g., in APC, BRAF, CDKN2A, PIK3CA, RAS and TP53) or cell doubling time across a panel of cancer cell lines (6).Of note, a recent clinical study revealed that STING expression is a biomarker for overall survival in PDL1-negative, TMB-low non-small cell lung cancer treated with ICB (56).This further supports the idea to use STING as a marker for recruiting patients on TTKi and ICB trials.
One caveat of our study is that we focused on one TTKi, OSU13, which means we cannot fully exclude the possibility that some of the observed effects could be off target.However, the kinase activity and SAC inhibition assay data presented here and previously (18) indicate that TTK is a predominant target of OSU13 in cells.Based on our findings we foresee the use of TTKi for pharmacological engineering of anti-tumor immunity in tumors with functional STING pathway.For example, TTKi can be used to jump start the immune response against tumors that are poorly immunogenic, such as tumors with low mutational burden and minimal T cell infiltrate.Since tumors with the evidence of ongoing immune response are more likely to respond to ICB immunotherapy compared to "immune-cold" tumors, it is promising to further explore OSU13 and other TTKi clinically in combination with ICB.

Sex as a biological variable
Our study examined male and female animals, and similar findings are reported for both sexes.

Cell viability, SA-β-Gal staining, and micronuclei assay
Cells were seeded at 0.2×10 6 -0.3×10 6 cells per well in flat-bottom 6-well plates and allowed to attach overnight prior to treatment and treated as described in figure legends.Cells were photographed in phase-contrast using EVOS microscope and counted by a pathologist.For crystal violet staining, cells were fixed with 100% methanol (Sigma) for 10 minutes and stained with 1% (w/v) crystal violet (Sigma) for 15 minutes, followed by at least three washes with water.
Fluorescent viability assay in organoids was described previously (23).Briefly, PDOs were seeded in an ultra-low attachment 96 well plate.OSU13 or vehicle (DMSO) were added to the wells in triplicate, and PDOs were incubated for 72 hours.To visualize dead cells, DNA, and live cells, cultures were incubated with 50 μg/ml propidium iodide, 10 μg/ml Hoechst 33342 and 5µM Calcein AM (Thermo Fisher Scientific), respectively.Images were taken on an inverted fluorescent microscope (EVOS™ M7000, Thermo Fisher).Live (green) and dead (red) cells per organoid region were quantified in Image J. Organoid regions were identified in a trans channel image.The ratio of live to green cell counts were plotted.
Senescence-associated (SA) β-galactosidase was performed using a kit (Sigma) according to the manufacturer's protocol.Micronuclei were detected using PicoGreen (Quanti-iT PicoGreen dsDNA reagent).Cells were incubated for 1.5h with 3 μl per mL of PicoGreen diluted in complete medium and subsequently fixed in 4% paraformaldehyde in PBS.Coverslips were mounted with ProLong Diamond Antifade Mountant (Invitrogen).Images were acquired from four representative fields of each condition using an inverted fluorescent microscope (EVOS™ M7000, Thermo Fisher).The number of micronuclei per cell was quantified in ImageJ using the Cell Counter plugin.Micronuclei structures were defined as minor DNA aggregates separate from the primary nucleus.In addition to PicoGreen staining, DAPI and transmitted light images were used to confirm micronuclei structure and localization.

Real-time PCR, ELISA, multiplex ELISA and bioinformatics analysis
CCL5 mRNA was measured using real-time PCR as described previously (19).Human CCL5 and CXCL10 secretion was assessed using ELISA kits purchased from R&D. Mouse CCL5 was detected using ELISA kit from R&D.For the analysis of cell secretome, cell conditioned media was submitted to RayBiotech for quantitative multiplex ELISA assay that provided absolute concentrations of tested proteins in the media.These data were used to calculate PCA, plot the heat map, and identify proteins differentially expressed between treatment group.To construct a PCA plot and heatmap, expression data was uploaded to ClustVis web portal.Individual proteins were organized in rows, and treatment conditions in columns.For PCA, unit variance scaling was applied to rows; SVD with imputation was used to calculate principal components.X and Y axis show principal component 1 and principal component 2. For heatmap, rows were centered; unit variance scaling was applied to rows.Both rows and columns were clustered using correlation distance and average linkage.
To identify proteins differentially expressed between OSU13-treated STING WT and KO cells multiple unpaired t-test was used in GraphPad Prism with individual variances computed for each protein.Proteins were considered to be differentially expressed if t-test p-value was smaller than 0.1.A list of proteins that passed the FDR threshold was uploaded to OmicsNet and network analysis of one list of molecules was performed using the "Proteins" option.STRING database of known and predicted protein-protein interactions was selected for network creation and Streiner Forest (PCSF) network tool was applied to simplify the network by identifying subnetwork enriched with input values.The networks were visualized in 2D.

Mouse experiments
Experiment with CT26 tumors in BALB/c mice was performed by Charles River Laboratories (Morrisville, NC).All other animal experiments were approved by the OSU IACUC.Male and female C57BL/6 purchased from Jackson laboratories were used for in vivo tumor experiments.C57BL/6 and BALB/c mice were used for toxicity studies.To inoculate CT26 tumors, female mice were injected subcutaneously with 3 x 10 5 cells.To inoculate MC38 tumors, 10 5 cells were injected subcutaneously into male and female mice.Tumors were measured using calipers and tumor volume was calculated as 0.5 x length x width x width.Mice were euthanized if their tumors exceeded 16 mm in diameter or became perforated, or if a mouse lost >20% body weight.
OSU13 was provided by the OSUCCC Drug Development Institute.The vehicle used to prepare dosing solution of OSU13 was 12.5% Ethanol : 12.5% Cremophor EL in 5% dextrose in water.OSU13 was prepared by dissolving an appropriate amount of powder stepwise in ethanol, Cremophor EL, and D5W to yield dosing solutions at 1.16 and 5.78 mg/mL in vehicle.These provided active doses of 10 and 50 mg/kg, respectively.
For the studies of toxicities, terminal blood collection was performed by severing brachial plexus with scissors.For white blood cell count, blood was collected into EDTA-containing BD vacutainer tubes (Catalog # 367861), for AST and ALT analysis blood was collected in 1.5 mL microcentrifuge tubes, allowed to clot, and centrifuged to obtain the serum.Samples were submitted to the Comparative Pathology and Digital Imaging Shared Resource or CPDISR (part of the Comprehensive Cancer Center at The Ohio State University) for following services: Hematology: Complete Blood Count/Differential, Single Chemistry: Aspartate aminotransferase (AST), and Single Chemistry: Alanine aminotransferase (ALT).

Flow cytometry
Spectral flow cytometry was used to study the expression of surface or intracellular (nuclear) immune markers (Table S3) in bone marrow and tumor-derived immune cells.Mouse bone marrow cells were obtained by flushing mouse tibias and femurs with PBS.To obtain tumor cell suspensions, tumors were minced and processed on gentleMACS dissociator in the presence of tissue digesting enzymes (Miltenyi Biotech).Cell suspensions were filtered through a cell strainer (BD), incubated with FC blocking antibody (BD Pharmingen), and a fixable viability dye (Live/Dead Aqua, ThermoFisher).After 30-min incubation with fluorescent antibodies recognizing surface markers, FOXP3 staining was performed using eBiosciences/Invitrogen FOXP3/Transcription factor staining buffer set (Ref 00-5523-00) according to the manufacturer's instructions.Cells were fixed in 0.5% buffered paraformaldehyde and analyzed on 5-laser Cytek Aurora.We used the UMAP algorithm implemented in OMIQ software for dimensionality reduction of data.FlowSOM clustering algorithm with consensus clustering metaclustering was used to identify distinct cell populations.Immune cell clusters were annotated by visually investigating the expression of immune markers in cells distributed on UMAP space.

NanoBRET™ In-Cell Target Engagement Assay
HEK293 cells were transiently transfected with 1 μg of LRRK2-NanoLuc fusion vector and 9 μg of transfection carrier DNA.The transfected cells were treated for 1 h with several concentrations of OSU13 (3-fold dilutions starting at 5 μM) or the positive control receptor protein tyrosine kinase inhibitor CEP-701 (3-fold dilutions starting at 1 μM).Kinase-ligand affinity was measured by competitive displacement with 0.016 μM K-9 tracer.The 600 nm/460 nm ratio was calculated and the normalized BRET response (%) was established prior to curve fitting.

SAC inhibition assay
CAL-51 cells were incubated with 100 ng/mL nocodazole for 20 hours and subsequently treated with several concentrations of OSU13 or the reference TTK inhibitor BOS-172722 (3-fold dilutions starting at 1 μM) without removal of nocodazole.After two hours, the cells were washed once with cold PBS and lysates prepared and phospho-Histone H3 (Ser10) measured using the Meso Scale Discovery Phospho-Histone H3 (Ser10) Assay Whole Cell Lysate Kit (Catalog # K150EWD-2).

Statistical analysis
Two-tailed Student's t-test and ANOVA were used to compare the means of two samples and multiple samples, respectively.Tukey method was applied to adjust the multiple comparisons.Linear mixed model was used for repeated measurement analysis of tumor growth.The survival difference between the different treatments was assessed by Logrank test.GraphPad's Prism 7.03 software was used for the statistical analysis.Two-sided p<0.05 was considered statistically significant.

Study approval
All mouse experiments were approved by the Ohio State University IACUC (Columbus, OH).Human melanoma PDOs were established and deidentified previously.No new human tissue was collected as a part of this study.Percentages of indicated cell types within the total live bone marrow cells from the analysis described in J. Viable cells were gated based on positivity for GR1, Ly6C, and B220 markers and analyzed using one-way ANOVA.

Figure 1 .
Figure 1.OSU13 inhibits tumor cells by inducing apoptosis, cell cycle arrest, and senescence. A. Representative images of colon (HCT116), melanoma (A375), and breast (MCF7) tumor cells treated with vehicle or 1μM OSU13 for 5 days and stained with crystal violet.Scale bar -150µm.B. Cell numbers in indicated cultured cells treated with 0.3 or 1 μM OSU13 or vehicle.Cells were imaged at 1-, 3-, 5-, and 7-days post-treatment.N=9 microscopic fields from 3 wells in a 6-well plate per condition.Statistics performed using 2-way ANOVA with Dunnett's post-test.C. Representative flow cytometry plots of MCF7 cells treated with vehicle control or 0.3 μM OSU13 for 5 days and stained with fluorescent antibodies recognizing cleaved PARP, BRDU, and γH2AX.BRDU was added to the cell culture media 4 hours before cell collection at the concentration of 25µg/mL.D. Quantified data from flow cytometry analysis shown in C in indicated cancer cell lines.N=3 replicates per condition.Statistical analyses using multiple unpaired t-tests.E. Representative results of cell cycle phase gating in A375 cells analyzed as described in C. F. Quantified cell cycle distribution data from analysis shown in E across three indicated cancer cell lines.N=3 biological replicates per condition.G.Western blot analysis of cell cycle-related proteins in A375, HCT-116, and MCF7 cells treated with 0.3 μM OSU13 or vehicle for 5 days.H. Representative microphotographs of MCF7 cells treated for 3 days with 0.3 or 1 μM OSU13 or vehicle and stained for the activity of SA-β-Gal.Scale bar 150 μm.I. Percentages of SA-β-Gal-positive cells quantified from the experiment shown in H. Three indicated cancer cell lines were used.N=12-15 microscopic fields per condition (across 3 wells in a 6-well plate).Statistics using one-way ANOVA with Dunnett's multiple comparisons tests.

Figure 2 .
Figure 2. OSU13 reduces melanoma patient-derived organoids viability.A. Representative images of individual organoids from 4 distinct patients treated with OSU13 at 0.5, 1, or 2 μM for 3 days.Live (green) and dead (red) cells were visualized using Calcein AM and propidium iodide, respectively.DNA (blue) was visualized with Hoechst 33342.Scale bar = 50 μm.B. Quantified data from the PDO experiment shown in A. N=8-15 individual organoids per patient and treatment condition.Statistical analysis using ANOVA with Tukey's multiple comparisons test.

Figure 3 .
Figure 3. OSU13 induces inflammatory chemokines and transcription factors in a STINGdependent manner.A. Results of ELISA testing CCL5 in the conditioned media from cancer cells treated with indicated concentrations of OSU13 for 5 days.Statistical analysis using oneway ANOVA with Dunnett's multiple comparisons test.B. CCL5 ELISA results performed on tumor cell conditioned media collected at 1, 3, 5, and 7 days of treatment with 0.3 or 1 μM

Figure 4 .
Figure 4. STING promotes a pro-immunogenic and tumor-inhibitory secretome in OSU13treated tumor cells.A. Heat map depicting relative content of indicated immune proteins in conditioned media from STING wild type (p) or knockout (2.1 and 3.1) A375 cells treated for 3 days with vehicle (V) or 1 μM OSU13.Proteins significantly upregulated by OSU13 are shown.Rows are centered; unit variance scaling is applied to rows.Rows are clustered using correlation distance and average linkage.B. Principal component analysis of secreted proteins based on data shown in A. C. Secretion of select individual proteins from the data shown in A. D. Protein networks constructed based on data in A. Circles represent individual proteins differentially expressed in OSU13-treated STING WT and KO cells.Connecting lines represent known/validated protein-protein interactions.The size of individual circles reflects how many connections this protein has with other network proteins.

Figure 5 .
Figure 5. OSU13 inhibits tumor growth and increases tumor immune infiltrate.A. Growth of MC38 tumors in mice treated with 10mg/kg OSU13 or vehicle 5 days a week (5 days on/2

Figure 6 .
Figure 6.OSU13 augments immunotherapy responses in mice.Growth of MC38 tumors treated with vehicle or OSU13 (10 mg/kg, 5 days a week) in the absence or presence of 100 µg of anti-PD-1 or isotype antibody treatment (dose).N=6-10 mice per group.Statistics using a mixed model with Tukey's post-test.B. Survival of mice shown in A. Statistical comparison between vehicle and OSU13 and anti-PD-1 combination group is indicated.C. Results of the

Figure 7 .
Figure7.OSU13-treated C57Bl/6 mice display no signs of severe toxicities when treated 5 days a week.A. Changes in body weight over time in tumor-free C57BL/6 mice.Mice were treated with OSU13 5 days a week or anti-PD-1 twice a week for 3 weeks.N=5 mice per group.B. Analysis of indicated liver proteins in the serum of mice described in A. The dotted lines indicate the normal range for C57BL/6 mouse.C-I.Analysis of white blood cells in mice whole blood.The dotted lines indicate the normal range for each type of cell.J. Spectral cytometry analysis of bone marrow cells from mice described in A. Second and third row indicate the dimension reduction analysis and distribution of select immune marker expression.K. Percentages of indicated cell types within the total live bone marrow cells from the analysis described in J. Viable cells were gated based on positivity for GR1, Ly6C, and B220 markers and analyzed using one-way ANOVA.