TAK-676: A Novel Stimulator of Interferon Genes (STING) Agonist Promoting Durable IFN-dependent Antitumor Immunity in Preclinical Studies Open Access

Immuno-oncology checkpoint inhibitor therapies are revolutionizing cancer treatment and have improved patient outcomes across a broad range of tumor types; however, only a small fraction of patients respond to therapy, and primary and secondary resistance remains a significant problem (1–3). Nonresponsiveness to immunotherapy may involve an inadequate T-cell response in the tumor microenvironment (TME; ref. 1). Thus, there is an urgent need to identify new therapies that activate both the innate and adaptive immune response to drive more potent and durable antitumor immunity, with a manageable safety and tolerability profile.

Type I IFNs are known for promoting robust antiviral immunity. Accumulating evidence suggest that type I IFNs produced by innate immune cells, such as dendritic cells (DCs), play a critical role in cancer immunosurveillance. Type I IFN production has been associated with the host detection of danger signals that are released from stressed, injured, or necrotic cells through a diverse family of innate pattern recognition sensors (1). One such sensor is the stimulator of interferon genes (STING). STING is an endoplasmic reticulum signaling protein, that is broadly expressed in both immune and nonimmune cell types. STING binds 2′,3′-cGAMP produced by cyclic GMP–adenosine monophosphate (AMP) synthase in response to cytosolic DNA (4) and rapidly induces type I IFNs and proinflammatory cytokines (5–7). It has been shown that productive type I IFN-mediated antitumor immunity requires an intact STING pathway within the DC population (8). Collective evidence highlights that agonism of the host STING pathway plays a critical role in eliciting robust and durable antitumor immunity (9, 10).

Recognizing the therapeutic potential of targeting the STING pathway for its positive immune-modulatory properties has prompted many groups to develop STING-activating drugs utilizing intratumoral injection as an administration route (9). Poor pharmacokinetic and physicochemical properties of 2′,3′-cGAMP and its derivatives are hurdles to the development of systemically administered STING agonists (9) . Therefore, overcoming these obstacles would facilitate the clinical development of STING agonists as novel anticancer therapeutics and potentially increase the number of patients that can benefit. TAK-676 is a novel synthetic STING agonist that is administered as an intravenous infusion and is a highly potent modulator of the innate immune system (11, 12). TAK-676 is currently under investigation in cancer patients with locally advanced or metastatic solid tumors (ref. 13; NCT04420884, NCT04879849, NCT04541108). These trials are investigating the antitumor effects of TAK-676 as a monotherapy or with combination therapies such as checkpoint inhibitors, radiation, and chemotherapy.

Here we present the first report of the structure of TAK-676, and the activity of TAK-676 in preclinical studies. We demonstrate that TAK-676 dose-dependently triggers activation of the STING signaling pathway and activation of type I IFNs, as well as robust activation of innate and adaptive immune activity both in vitro and in vivo. In syngeneic murine tumor models in vivo, TAK-676 induced dose-dependent cytokine responses and increased the activation and proliferation of immune cells within the TME and tumor-associated lymphoid tissue. Ultimately this immune activation resulted in durable antitumor responses within multiple syngeneic tumor models.

A summary methodology for all components of the analyses reported is provided below; full methodologic details for all analyses are provided in the Supplementary Appendix. For details on cell lines and use, see “Cell lines and mouse models” section at the end of this Materials and Methods section.

Detailed methods for the synthesis and purification of TAK-676 can be found within the Supplementary Appendix.

To measure the binding of TAK-676 to mouse, rat, monkey, and human STING orthologs, in vitro time-resolved fluorescence resonance energy transfer (TR-FRET) assays were used. To assess pathway activation, the stimulatory effect of TAK-676 on the STING signaling pathway in vitro in the THP1-Dual human acute myeloid leukemia (AML) and CT26.wild type (WT) and STING-deficient CT26.KO mouse colon carcinoma cell lines was evaluated. Cells were treated with increasing concentrations of TAK-676 for 3 hours, subsequently lysed, and the proteins analyzed by SDS-PAGE and immuno-blotting. For experiments in STING-deficient CT26.KO cell lines, 5,6-dimethylxanthenone-4-acetic acid (DMXAA; 50 μg/mL) was used as a control for a treatment time of 1 hour prior to cell lysis and blotting. Three STING-induced phosphorylation events, part of a signaling cascade critical for the induction of type I IFNs, were monitored by Western blotting: TANK-binding kinase 1 (TBK1) on S172 in the kinase activation loop, STING on S365/366 (mouse/human numbering), and interferon regulatory factor (IRF) 3 on S396.

STING agonists trigger the production of type I IFNs and the induction of interferon-stimulated genes (ISG) through IRFs. The potency of the STING agonist was evaluated in reporter assays using the ISRE_NanoLuc human embryonic kidney 293 (HEK293T) cells, the more physiologically relevant human monocyte derived THP1-Dual cells, and the mouse macrophage RAW-Lucia ISG cells. These cell lines, that are of mouse or human origin, allow measurement of STING agonist activity using the NanoLuc luciferase (HEK293T) or IRF-inducible Lucia luciferase as readouts (THP1-Dual and RAW-Lucia). In HEK293T cells, which lack endogenous STING, human and mouse STING isoforms were transiently transfected, and a reporter assay was developed by generating an ISRE_NanoLuc HEK293T stable cell line using the pNL (NLucP/ISRE/Hygro) vector to monitor the resulting increase in type I IFNs. The concentration at inflection point of the curve producing a half-maximal response (EC₅₀) was calculated in cultured ISRE_NanoLuc HEK293T cells that were transiently transfected with human WT STING (R232), mouse STING (WT), and the four other isoforms of STING—R232H, R293Q, G230A-R293Q (AQ), and R71H-G230A-R293Q (HAQ)—that exist in human populations (14). The EC₅₀ was also calculated in RAW-Lucia ISG cells and THP1-Dual cells. Digitonin was used to permeabilize the HEK293T and RAW-Lucia ISG cells, but not the THP1-Dual cell line as digitonin was toxic to these cells, thus preventing accurate measurement of type I IFNs.

The “hit-and-run” clustered regularly interspaced short palindromic repeats (CRISPR) method was employed for creation of STING knockout (KO) cell clones and their WT counterparts in CT26.WT and B16F10 cells by following the Lipofectamine CRISPRMAX Transfection Reagent protocol (Thermo Fisher Scientific). Confluent clones were harvested and evaluated by Western blotting for mSTING protein and PCR analysis of mSTING genomic DNA to reconfirm efficient knockdown. Several confirmed clonal mSTING KOs, as well as matched WT controls shown to have intact mSTING, were expanded, banked, tested for murine pathogens, and cryopreserved for use in subsequent in vitro and in vivo studies.

As STING proteins are highly expressed in DCs, the activation of DCs derived from human peripheral blood mononuclear cells from 5 healthy donors and mouse bone marrow (BMDC) was evaluated using flow cytometry to detect the expression of the DC activation marker cluster of differentiation (CD) 86. The human monocyte-derived DCs (MoDC) were differentiated from CD14⁺ cells (monocytes), and the mouse BMDCs were differentiated from freshly isolated bone marrow (BM) cells of femurs and tibiae of 2 donor BALB/c mice. Mean fluorescence intensity (MFI) of CD86⁺ DCs from each sample was evaluated by flow cytometry and MoDC viability (defined as LIVE/DEAD Fixable Near-IR Dead Cell Staining Dye negative population from the parent gate) was recorded. Natural killer (NK) and T-cell activation was assessed in whole blood from 5 healthy human donors using flow cytometry to detect expression of the activation marker CD69 on CD4⁺ and CD8⁺ T cells and CD56⁺/CD16⁺ NK cells. CD69 expression (as evaluated by MFI) was measured for each cell population, and mean plasma-based EC₅₀, Hill slope, and concentration producing 10% of maximal response (EC₁₀) were calculated.

Female BALB/c mice were inoculated subcutaneously with mouse A20 (B-cell lymphoma) tumor cells in the right flank and TAK-676 dosing was initiated when estimated tumor volumes reached 300–800 mm³. Mice (n = 3/group) were dosed via intravenous injection with single doses of TAK-676 at 0.025, 0.125, 0.25, 0.5, and 2 mg/kg formulated in PBS as a single agent. Exposure to TAK-676 was determined by measuring plasma and tumor concentrations of TAK-676 at various timepoints using a LC-MS/MS method.

TAK-676 activity was evaluated in syngeneic CT26.WT and A20 mouse models. In addition, a B16F10_WT (melanoma) tumor model and STING-deficient B16F10_STING KO tumor model were used to demonstrate the STING dependence of antitumor activity using both WT and STING-deficient C57BL/6J-Tmem173gt/J (Goldenticket; Jackson Laboratory) models. Mice were inoculated subcutaneously into the right flank with CT26.WT tumor cells, A20 tumor cells, or B16F10 tumor cells and were treated with either vehicle (PBS) or TAK-676. Mice in the vehicle and TAK-676 (1 or 2 mg/kg) groups received intravenous doses once every 3 days for three doses (Q3D×3, days 0, 3, and 6). Effects on tumor growth were evaluated by measuring growth rate inhibition (GRI) and tumor regression. Tolerability was assessed by measuring percent body weight loss (BWL; body weights were measured twice weekly), mortality, or any clinical signs of adverse treatment-related side effects (see Supplementary Appendix).

Pharmacodynamic cytokine response to TAK-676 was evaluated in female BALB/c mice inoculated subcutaneously with 1.0 × 10⁶ A20 syngeneic tumor cells, and in C57BL/6 WT or Goldenticket mice inoculated subcutaneously with 0.8 × 10⁶ B16F10 WT or STING KO tumor cells. When tumor volumes grew to 300–800 mm³, mice (n = 5/group) were dosed once with vehicle (PBS), or TAK-676 intravenously at 0.05, 0.125, 0.5, 1.0, and 2.0 mg/kg. The pharmacodynamic effect of TAK-676 in plasma (for animals inoculated with A20 cells) or serum (for animals inoculated with B16F10 cells) and tumor samples was profiled to quantitate the following cytokines: IFNα, IFNγ, IFNγ-induced protein 10 (IP-10) for all mice; TNFα, IL6, and monocyte chemoattractant protein-1 (MCP-1) for animals inoculated with A20 cells.

Female C57BL/6 mice bearing B16F10 tumors were treated with TAK-676 intravenously at 0.3, 1.0, 2.0 mg/kg or vehicle on day 0 (n = 4/group). In addition, female BALB/c mice bearing CT26 tumors were treated with vehicle or 0.25 mg/kg TAK-676 on day 0 (n = 4/group). Pooled tumor-draining lymph nodes (axillary, brachial, and inguinal) and tumors were collected on day 3, and immune cell populations were analyzed by flow cytometry. Samples were initially gated off total live cells (within the live/dead negative gate) and subsequently for CD45⁺ cells. In the tumor, total T cells, defined as CD3⁺, were then examined further to specifically identify tumor-infiltrating CD8⁺ T cells. Because of the strong correlation between the presence of cytotoxic CD8⁺ T cells within the tumor and clinical prognosis (15, 16), CD8⁺ T cells in the tumor samples were further characterized for markers of activation and proliferation. CD8⁺ T-cell populations were evaluated for either CD69 (A20 tumors) or Ki67 and CD25 expression (CT26 tumors). In addition, to assess the functionality of the CD8⁺ T-cell population within the A20 tumors, expression of the proinflammatory cytokine IFNγ was evaluated ex vivo following an overnight stimulation with plate bound α-CD3 and soluble α-CD28 antibodies. In the lymph node, CD8⁺ T cells were also assessed for IFNγ expression, as well as Ki-67 and CD69 expression, following overnight stimulation with α-CD3/αCD28. Lymph nodes were also evaluated for the CD11c⁺ MHC1⁺ DC population as a fraction of all viable CD45⁺ cells, as well as for the expression of DC activation markers CD80 and CD86. Lymph nodes from mice inoculated with CT26 tumors were also evaluated for the expression of CD69 on NK cells.

All animal experiments were performed in compliance with protocols as approved by an Institutional Animal Care and Use Committee. All cell lines used in the experiments were obtained from the indicated sources summarized below and included in Supplementary Table S1, cultured in appropriate media and expanded, and then banked into the Takeda Oncology Cell bank as high-density frozen vials stored in liquid nitrogen (LN₂) between passage three and six. For cell line reporter assays, cells were thawed from high-density frozen vials and then used immediately for experiments following a short recovery period. Cells for in vitro Western blot assays and for in vivo implantation were thawed from frozen vials and cultured a short time before use in the experiment. During culture time, cells were fed and split as needed to avoid over confluence. Cells were used for experiments as needed during the period of culture but never exceeding passage 10. Cells were Mycoplasma and murine pathogen tested via the IMPACT I assay by Idexx Bioresearch (the most recent testing dates are indicated in Supplementary Table S1 and summarized below). CT26.WT, A20, and B16F10 cell lines were additionally verified using IDEXX Cell Check Cell Line Authentication Service with 27-Marker STR Strain Analysis.

Individual cell line details: THP-1 Dual Huamn Acute Myeloid Leukemia cells were obtained from Invivogen (catalog no. thpd-nfis) and were last Mycoplasma tested on June 19, 2017. Human Embryonic Kidnsey 293 (HEK293T) cells were obtained from ATCC (catalog no. CRL-11268) and were last Mycoplasma tested on February 15, 2018. ISRE-Nano Luc HEK293T cells were generated by Promega with Vector CS190901 and were last Mycoplasma tested on February 15, 2018. CT26.WT cells were obtained from ATCC (catalog no. CRL-2638) and were last Mycoplasma tested on June 23, 2016. A20 cells were obtained from ATCC (catalog no. TIB-208) and were last Mycoplasma tested on June 25, 2015. B16F10 cells were obtained from ATCC (catalog no. CRL-6475) and were last Mycoplasma tested on July 22, 2016.

For in vivo studies, the differences in the tumor growth trends over time between pairs of treatment groups were assessed by fitting each animal's data to a simple exponential growth model and comparing the mean growth rates of the two groups. The difference in the growth rates was summarized by the GRI, which is the reduction in growth rate experienced by the treatment group relative to that of the reference group, expressed as a fraction of the vehicle growth rate. A positive GRI indicates that the tumors in the treatment group grew at a reduced rate relative to the reference group. A statistically significant P value (<0.05) suggests that the trends over time for the two treatment groups were different. For all other relevant studies, the mean, SD, and P value (where significant; relative to control) were calculated using GraphPad Prism 7 software (GraphPad Software).

Data are available upon reasonable request made to the corresponding authors.

The chemical structure of TAK-676 is shown in Fig. 1A (for details, see Supplementary Appendix: Methods of chemical synthesis and characterization of TAK-676). Data on TAK-676 binding from the TR-FRET assay showed highly selective binding and activation of STING proteins from various species, with dissociation constant (K_d) values estimated to be less than or equal to 0.010 ± 0.0008 μmol/L, 0.008 ± 0.001 μmol/L, 0.011 ± 0.001 μmol/L, and 0.027 ± 0.008 μmol/L for mouse (n = 3), rat (n = 3), cynomolgus monkey (n = 3), and human (n = 20) STING orthologs, respectively (Supplementary Table S2).

IFN activation was assessed in vitro by using reporter assays. ISRE_NanoLuc HEK293T cells were transiently transfected with human WT STING (R232), with and without digitonin. The EC₅₀ was 0.3 ± 0.11 μmol/L in the absence of digitonin and 0.09 ± 0.07 μmol/L in permeabilized cells. IFN activation EC₅₀ values were also assessed in commercially available THP1 Dual and RAW-Lucia ISG reporter cell lines and EC_50s were 1.53 ± 0.45 μmol/L, and 1.78 ± 0.48 μmol/L in the THP1-Dual, and RAW-Lucia ISG cell lines, respectively (Fig. 1B).

To assess the ability of TAK-676 to activate the STING pathway, the phosphorylation of STING and STING pathway proteins was evaluated in human THP1-Dual and murine CT26.WT cell lines following treatment with TAK-676, as shown in Fig. 2A and B, respectively. These data demonstrate a dose-dependent induction in the expression of pSTING (S366), pTBK1 (S172), and pIRF3 (S396) in both murine and human cell lines following treatment with TAK-676. To confirm that the induction of STING pathway activation following treatment with TAK-676, STING-deficient cells were generated from CT26.WT cell lines. As shown in Fig. 2C, in the absence of STING expression neither TAK-676 nor control STING agonist DMXAA induced the phosphorylation of pTBK1 (S172) or pIRF3 (S396). Collectively, these data demonstrate the ability of TAK-676 to dose-dependently induce STING-TBK1-IRF3 activation is critically dependent on STING expression.

STING protein is highly expressed in DCs, and activation of DCs following STING pathway activation is key to promote innate and adaptive immune response (17). TAK-676–mediated DC activation studies were completed by monitoring the induction of the DC maturation marker CD86 by flow cytometry in both mouse BMDCs and human MoDCs. Fig. 3A–C demonstrate a dose-dependent TAK-676 induction of CD86 on the surface of BMDCs (0.1–10 μmol/L, 2 donors, 1, 3, 6, 24, 48, 72 hours) and MoDCs (0.03–30 μmol/L, 5 donors, 24 hours). The average EC₅₀ value for TAK-676-induced MoDC activation at the 24-hour timepoint was 1.217 μmol/L (±0.352 μmol/L), and 0.32 μmol/L for the BMDC activation. No significant decrease of cell viability was observed in either MoDCs or BMDCs (Supplementary Fig. S1).

To further characterize the ability of TAK-676 to mediate both innate and adaptive immune cell activation, TAK-676–mediated NK- and T-cell activation was measured by evaluating the expression levels of CD69 on human peripheral blood NK and T cells (n = 5 healthy human donors), as shown in Fig. 3D–F. TAK-676 treatment resulted in a dose-dependent increase in CD69 expression in both NK and T cells. The plasma-based EC₅₀ and Hill slope for each donor was analyzed, and the mean EC₅₀ was estimated to be 0.271, 0.216, and 0.249 μmol/L for NK, CD8⁺, and CD4⁺ T cells, respectively. The mean plasma-based EC₁₀ was calculated from the mean plasma-based EC₅₀ to be 0.114, 0.0986, and 0.106 μmol/L for NK cells, CD8⁺, and CD4⁺ T cells, respectively. Collectively, these data suggest that TAK-676 is a potent inducer of both innate and adaptive immune cell activation.

To support systemic dosing of TAK-676, mean plasma and tumor concentration–time curves of TAK-676 in female BALB/c mice bearing A20 syngeneic tumors were calculated and are presented in Supplementary Fig. S2. As shown, intravenous administration of TAK-676 resulted in dose-proportional pharmacokinetics in the plasma and pharmacologically meaningful exposure in the tumor tissue versus plasma. The mean maximum observed concentration (C_max) and mean area under the concentration–time curve (AUC) from time 0 to 72 hours (AUC₇₂) after intravenous administration of TAK-676 0.025, 0.125, 0.25, 0.5, and 2 mg/kg are shown in Supplementary Table S3.

Dose-dependent in vivo efficacy of TAK-676 was tested in multiple mouse syngeneic tumor models, including A20 (Fig. 4A) and CT26.WT (Fig. 4B) syngeneic tumor models. In both models, the dosing of TAK-676 at both 1.0 and 2.0 mg/kg was well tolerated (Supplementary Fig. S3A and S3B). In mice, body weight is a meaningful safety readout to determine the MTD (18, 19), and all experiments were conducted at or below the MTD. BALB/c mice bearing A20 syngeneic tumors showed significant antitumor activity compared with vehicle treatment when TAK-676 was dosed at 1 mg/kg intravenously Q3D×3 (GRI 72%, P < 0.001), with 1 of 10 mice achieving complete response (CR). Similar to the A20 tumor model, CT26 tumor-bearing animals treated at 1 mg/kg also displayed tumor control over vehicle treated animals (Fig. 4B). Notably, when the dose of TAK-676 was increased from 1.0 to 2.0 mg/kg, there was an observable increase in the antitumor activity in both A20 (GRI = 91%, P < 0.001) and CT26 (GRI = 132%, P < 0.001) models (Fig. 4A and B; left), with more animals achieving CRs (two for A20 and four CRs for CT26). To assess the durability of the antitumor response elicited by TAK-676, mice with CRs in both groups were continuously monitored for ≥48 days with no tumor regrowth.

The ability of TAK-676 to promote durable CD8-dependent antitumor efficacy was assessed in a subsequent study in CT26.WT tumor-bearing BALB/c mice. In this study, two mice with CRs following treatment with 1 mg/kg of TAK-676 were rechallenged on the opposite flank with CT26.WT tumors on day 34. Following rechallenge, no tumor regrowth was observed in the complete responder mice, as compared with naïve control animals. To determine whether CD8⁺ T cells were responsible for the observed durable antitumor immunity in these two animals, on day 59, CD8⁺ T cells were depleted by antibody treatment, and regrowth was observed in one animal. At day 88, CD8⁺ T cells were depleted a second time, at which point regrowth was observed in the second animal (Fig. 4C). These data demonstrate that TAK-676 drives durable anti-tumor efficacy that is resistant to rechallenge and is dependent upon CD8⁺ T cells.

To assess whether TAK-676–driven antitumor response was dependent on either host or tumor cell STING expression, Goldenticket mice were utilized and compared with WT mice following implantation of either WT or STING KO B16F10 tumor cells. When TAK-676 was dosed at 2 mg/kg intravenously Q3D×3 in WT C57BL/6 mice bearing B16F10 WT tumors, significant antitumor activity was observed (GRI 136%, P < 0.001; Fig. 5A). In contrast, in STING-deficient mice bearing B16F10 WT tumors, TAK-676 treatment failed to induce significant antitumor activity, with a GRI of −2% (P = 0.862; Fig. 5B). In addition, although TAK-676 was well tolerated in both strains, STING-deficient animals treated with TAK-676 showed reduced BWL when compared with their WT counterparts (see Supplementary Results and Supplementary Fig. S4 for details of tolerability). These results underscore the importance of the STING signaling pathway in the host for TAK-676–driven efficacy. To determine the role of STING expression in tumor cells, STING KO B16F10 cells were implanted in WT C57BL/6 animals. Treatment with TAK-676 2 mg/kg intravenously Q3D×3 in this model resulted in reduced, but significant, antitumor activity (GRI 53%, P = 0.001). These data demonstrate that STING expression in the host is the major determinant in promoting robust antitumor immunity (Fig. 5C). In further support of this, treatment of STING KO B16F10 tumor-bearing STING-deficient mice with TAK-676 also resulted in a complete lack of antitumor activity similar to what was observed with WT B16F10 tumor-bearing Goldenticket mice (GRI −16%, P = 0.349; Fig. 5D).

To evaluate the role of host versus tumor cell STING expression on STING pathway activation, tumors from these experiments were assessed for expression of STING, pIRF3 (S396), and pTBK1 (S172) at 3 hours after treatment with TAK-676 2 mg/kg intravenously As shown in Fig. 5E (left), TAK-676 treatment induced expression of pIRF3 (S396) and pTBK1 (S172) in both WT and STING-deficient mice implanted with B16F10 WT tumors. In animals implanted with STING KO B16F10 tumors (Fig. 5E; right), activation of the STING pathway can be observed, albeit at a much reduced level in WT C57BL/6 mice, but is completely abrogated in STING-deficient mice.

Type I IFNs modulate the activity of innate immune effector cells and promote DC maturation and antigen presentation to T cells, propagating an adaptive immune response (20, 21) downstream of STING pathway activation. As such, we sought to investigate the potential of TAK-676 to induce type I IFNs, as well as other cytokine expression in the plasma and the tumor, following treatment in syngeneic tumor-bearing mice. As shown in Fig. 6A–C (IFNα, IFNγ, and IP-10) and Supplementary Fig. S5 (TNFα, MCP-1, and IL6), TAK-676 induced dose-dependent cytokine responses in female BALB/c mice bearing A20 syngeneic tumors. To determine whether the induction of these cytokines depended on either tumor or immune cell STING expression, cytokine responses were evaluated in the serum and tumors from both WT and STING-deficient mice bearing WT or STING-deficient B16F10 tumors following treatment with TAK-676. As shown in Fig. 6 (right), TAK-676–driven induction of IFNα, IFNγ, and IP-10 in both serum and tumor is dependent upon host STING expression and less so in the tumor cells. Collectively, the in vivo and in vitro data suggest that host STING expression is a critical determinant of robust immune activation leading to effective antitumor response.

To evaluate the ability of TAK-676 to induce T-cell infiltration, proliferation, and activation in vivo, immune modulation in the tumor and lymph nodes was assessed following treatment in tumor-bearing mice. The in vivo T-cell responses to TAK-676 in C57BL/6 mice bearing B16F10 tumors and BALB/c mice bearing CT26 tumors are shown in Fig. 7 (left and right, respectively). At day 3, the frequency of live CD45⁺ cells increased in tumors isolated from all TAK-676–treated groups (Fig. 7A and B). Intratumoral CD8⁺ T cells within the viable CD45⁺CD3⁺ cells were slightly increased at day 3 in B16F10 tumors, and significantly increased in CT26 tumors (Fig. 7C and D). While in vitro activation assays demonstrated an increase in CD69⁺ CD8⁺ T cells following treatment with TAK-676, this did not translate to in vivo assessments at day 3 in B16F10 tumors (Fig. 7G). However, an increase in IFNγ⁺CD8⁺ T cells was observed in B16F10 tumors, indicating activation of the CD8⁺ T cells (Fig. 7E). Corresponding increases in CD25 and Ki67 expression were also observed in the CD8⁺ T cells of CT26 tumors, further demonstrating CD8⁺ T-cell activation as well as increased proliferation in these tumors as early as day 3 following TAK-676 dosing (Fig. 7F and H).

Immune modulation following treatment with TAK-676 was also observed within the tumor-draining lymph nodes of B16F10 and CT26 tumor-bearing mice (Fig. 8). Within the tumor-draining lymph nodes of the B16F10 tumor-bearing mice, a measurable increase in total viable cells was initially observed at day 3 (Fig. 8A). Notably, also at day 3, the frequency of total DCs (identified as CD45⁺CD11c⁺MHCII⁺) within the lymph node was increased in both B16F10 and CT26 tumor-bearing mice (Fig. 8B and H). As observed in the in vitro DC activation assays, TAK-676 treatment also resulted in increased activation of DCs in the lymph nodes from both models at day 3, as demonstrated by increased frequency or expression of CD80 and CD86 (Fig. 8C, D, I, and J).

Similarly to what was observed within the tumors, an increase in CD69⁺ CD8⁺ T cells was not seen in the lymph nodes from B16F10 tumor-bearing animals (Fig. 8G). However, the frequency of IFNγ⁺CD8⁺ T cells was again increased within the tumor-draining lymph nodes of B16F10 tumor-bearing mice (Fig. 8E). In addition, TAK-676 induced the proliferation (as measured by Ki67⁺ staining) of CD8⁺ T cells within the lymph nodes of B16F10 tumor-bearing mice at day 3 following treatment (Fig. 8F). Finally, the frequency of CD69⁺ NK cells was increased in the lymph nodes from CT26 tumor-bearing animals, corresponding to the NK-cell activation observed in vitro (Fig. 8K).

Collectively, these data demonstrate increased NK-cell activation and DC frequency and activation in the lymph nodes accompanied by increased frequency, proliferation, and activation of CD8⁺ T cells in both the lymph nodes and the tumor following treatment with TAK-676. These data are consistent with the in vitro immune cell activation shown in Fig. 3, and confirm the ability of TAK-676 to induce both innate and adaptive immune responses in vivo.

This is the first report of the chemical structure and biological activity of the synthetic STING agonist TAK-676, designed for intravenous administration by improving the drug-like properties of natural STING agonists through chemical modifications. Here we demonstrate that TAK-676 is a potent and selective binder of STING proteins and activator of the STING signaling pathway with subsequent immune activation effects. Characterization of TAK-676 demonstrated potent activation of the STING signaling pathway in both human (THP1-Dual) and murine (CT26.WT) STING-expressing cell lines as evidenced by agonist-induced phosphorylation of TBK1, STING and the subsequent activation of IRF3, a key transcriptional regulator of type I IFN-dependent immune responses. TAK-676–mediated activation showed improved potency when cells were permeabilized with digitonin.

TAK-676 mediated STING pathway agonism and activated DCs, NK, and T cells both in vitro and in vivo in tumor-bearing syngeneic mouse models. The activation and priming of T cells observed here and with other STING agonists (22) to induce antitumor activity is of known importance in immunotherapy. The additional activation of NK cells seen here promotes an adaptive immune response by increasing cytokine and chemokine production, which enhances DC recruitment and maturation, and allows NK cells to directly kill tumors that T cells have not targeted (23, 24). The substantial activated immune cell response demonstrated with TAK-676 is important for driving the rejection of established tumors and distant metastases in addition to providing durable immunologic memory.

Significant STING-dependent antitumor activity was observed with TAK-676 compared with vehicle (PBS) in syngeneic mouse models in vivo. Mice bearing A20 tumor cells that experienced CRs in response to TAK-676 treatment did not show tumor regrowth over a 48-day period of monitoring, suggesting durable tumor suppression. In the Goldenticket mouse model TAK-676 treatment failed to induce significant antitumor activity against B16F10 syngeneic tumors, demonstrating that TAK-676 activity is mediated by STING agonism in the host, which is consistent with the presumed mechanism of action (11).

TAK-676 administered intravenously exhibited dose-proportional pharmacokinetics in plasma and pharmacologically meaningful exposure in the tumor tissue versus plasma. Direct intratumoral injection, which has been employed for clinical testing of some STING agonists (25), is designed to allow for drug penetration into the tumor lesion with lower concentrations of agent reaching healthy tissue (26). Furthermore, intratumoral administration requires needle accessible tumor lesions for treatment, therefore only some tumor types (melanoma, head and neck squamous cell carcinoma, hepatocellular carcinoma) have been clinically tested in humans (27), and there are legitimate concerns surrounding the risks of bleeding and tissue damage after repeated dosing (28). On the contrary, the intravenous administration with TAK-676 is potentially applicable to a broad population of patients with malignant cancers, including solid tumors and hematologic tumors, which may not be amenable to intratumoral administration. This expansion of tumor indications will afford opportunities to better understand how STING agonism modulates cancer immunity in humans.

Following administration in syngeneic tumor-bearing mice, TAK-676 dose-dependently induced cytokine production in both plasma and tumor, and increased the activation and proliferation of immune cells within the TME. Importantly, TAK-676 also exerted its effects within the tumor-draining lymph nodes, demonstrated by the increase in total viable cells, increased frequency of DCs, and increased activation and proliferation of CD8⁺ T cells over that of vehicle treated mice. DCs play a critical role in bridging an innate immune response into a long-term adaptive response with potential for durable memory (29, 30), and the ability of TAK-676 to increase the frequency of DCs within the tumor-draining lymph nodes is an important part of that process. The breadth of the immune response activated following TAK-676 administration suggests a broad-based effect that may be beneficial in terms of antitumor activity. Moreover, this response could be enhanced in combination with agents that disrupt DNA replication and increase antigen presentation, such as those used in radiation, chemotherapy, and anti-PD-1 therapy (31, 32).

In summary, TAK-676 is a selective and potent novel synthetic STING agonist that demonstrates immune activation effects in preclinical models bearing solid tumors as demonstrated by increased production of proinflammatory cytokines and activation of both tissue and circulating immune cells. Importantly, peripheral detection of TAK-676–mediated immune activation can be monitored in clinical samples, thereby enabling analysis of pharmacodynamic activity of TAK-676 in clinical studies (13). Significant STING-dependent antitumor activity was observed with no tumor regrowth and acceptable tolerability, suggesting durable tumor suppression. TAK-676 exhibited greater exposure in the tumor tissue than the plasma with intravenous administration, highlighting a potential benefit regarding the convenience of treating future patients versus intratumoral administration. Our findings identify TAK-676 as a promising new therapeutic candidate and support the clinical studies that are currently ongoing. TAK-676 is currently being investigated in phase I studies of patients with locally advanced or metastatic solid tumors alone or in combination with pembrolizumab, an anti-PD-1 antibody (ref. 13; NCT04879849, NCT04420884), and with various other anticancer agents (NCT04541108).

V.A. Appleman reports personal fees from Takeda Development Centers of America, Inc during the conduct of the study; personal fees from Takeda and other from Takeda outside the submitted work. V. Kolev reports other from Takeda during the conduct of the study. M. Mochizuki reports patent application WO2018/100558. S. Haridas reports personal fees from Takeda during the conduct of the study; personal fees from Takeda outside the submitted work. T.D. Thelen reports other from Takeda outside the submitted work. A.O. Abu-Yousif reports other from Takeda during the conduct of the study. M. Okaniwa reports patent application WO2018/100558. No other disclosures were reported.

E. Carideo Cunniff: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, writing-original draft, project administration, writing-review and editing. Y. Sato: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, project administration, writing-review and editing. D. Mai: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-original draft, writing-review and editing. V.A. Appleman: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-original draft, writing-review and editing. S. Iwasaki: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-original draft, writing-review and editing. V. Kolev: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-original draft, writing-review and editing. A. Matsuda: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-original draft, writing-review and editing. J. Shi: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-review and editing. M. Mochizuki: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-original draft, writing-review and editing; M. Yoshikawa: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-original draft, writing-review and editing. J. Huang: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-review and editing. L. Shen: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-review and editing. S. Haridas: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-original draft, writing-review and editing. V. Shinde: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-original draft, writing-review and editing. C. Gemski: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-original draft, writing-review and editing. E.R. Roberts: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, project administration, writing-review and editing. O. Ghasemi: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-review and editing. H. Bazzazi: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-review and editing. S. Menon: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-original draft, writing-review and editing. T. Traore: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-original draft, writing-review and editing. P. Shi: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-original draft, writing-review and editing. T.D. Thelen: Data curation, formal analysis, validation, investigation, visualization, methodology, writing-review and editing. J. Conlon: Formal analysis, validation, investigation, visualization, methodology, writing-original draft, writing-review and editing. A.O. Abu-Yousif: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, visualization, writing-original draft, project administration, writing-review and editing. C. Arendt: Writing-review and editing. M.H. Shaw: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, visualization, project administration, writing-review and editing. M. Okaniwa: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, visualization, writing-original draft, project administration, writing-review and editing.

This study was funded by Millennium Pharmaceuticals, Inc., Cambridge, MA, USA, a wholly owned subsidiary of Takeda Pharmaceutical Company Limited.

Medical writing support for the development of this article, under the direction of the authors, was provided by Philippa Lloyd, BSc, Georgia Pangratiou, PhD, and Steve Hill, PhD, of Ashfield MedComms, an Ashfield Health company, and funded by Takeda Pharmaceuticals U.S.A., Inc., in compliance with Good Publication Practice 3 ethical guidelines (Battisti et al., Ann Intern Med 2015;163:461–4).

The authors would like to acknowledge the following medicinal chemistry team members: Morihisa Saitoh, Taisuke Kato, Yayoi Yoshitomi, Tomohiro Seki, Yasuo Nakagawa, Yusuke Tominari, Masaki Setoh, Yusuke Sasaki, Tsuneo Oda, Akito Shibuya, Kosuke Hidaka, Zenyu Shiokawa, Shumpei Murata, Atsutoshi Okabe, Yoshihisa Nakada, Brian Scott Freeze, Taisuke Tawaraishi, Yasufumi Wada, Paul D. Greenspan, Hitoaki Nishikawa, Akito Hata, Ryutaro Adachi, and Hisashi Imaichi for their contributions to the lead optimization campaign leading to the discovery of TAK-676 and its derivatives. The authors would like to acknowledge Allison Berger, Sabrina C. Collins, Rick C. Gregory, Kristin H. Horn, Eric S. Lightcap, and Chirag Patel for preclinical development discussion.