IFNγ Production by Functionally Reprogrammed Tregs Promotes Antitumor Efficacy of OX40/CD137 Bispecific Agonist Therapy

Abstract Regulatory T cells (Treg) are highly enriched within many tumors and suppress immune responses to cancer. There is intense interest in reprogramming Tregs to contribute to antitumor immunity. OX40 and CD137 are expressed highly on Tregs, activated and memory T cells, and NK cells. In this study, using a novel bispecific antibody targeting mouse OX40 and CD137 (FS120m), we show that OX40/CD137 bispecific agonism induces potent antitumor immunity partially dependent upon IFNγ production by functionally reprogrammed Tregs. Treatment of tumor-bearing animals with OX40/CD137 bispecific agonists reprograms Tregs into both fragile Foxp3+ IFNγ+ Tregs with decreased suppressive function and lineage-instable Foxp3− IFNγ+ ex-Tregs. Treg fragility is partially driven by IFNγ signaling, whereas Treg instability is associated with reduced IL2 responsiveness upon treatment with OX40/CD137 bispecific agonists. Importantly, conditional deletion of Ifng in Foxp3+ Tregs and their progeny partially reverses the antitumor efficacy of OX40/CD137 bispecific agonist therapy, revealing that reprogramming of Tregs into IFNγ-producing cells contributes to the anti-tumor efficacy of OX40/CD137 bispecific agonists. These findings provide insights into mechanisms by which bispecific agonist therapies targeting costimulatory receptors highly expressed by Tregs potentiate antitumor immunity in mouse models. Significance: The bispecific antibody FS120, an immunotherapy currently being tested in the clinic, partially functions by inducing anti-tumor activity of Tregs, which results in tumor rejection.


Introduction
Immunomodulatory therapies targeting the PD1/PDL1 and CTLA4 effector T (Teff)-cell inhibitory signaling pathways induce striking objective clinical responses in certain cancer types but are ineffective at inducing durable responses in a majority of patients (1).These findings provide a rationale for developing new mechanistically distinct immunomodulatory therapies for cancer.CD4 + regulatory T cells (Treg), in which development is dependent on the transcription factor Foxp3, powerfully suppress Teff cells and prevent immune-mediated rejection of tumors (2,3).Low Treg-to-Teff cell ratios are associated with favorable prognosis and survival in the absence of immunotherapeutic treatment in ovarian cancer (4,5), breast cancer (6), nonsmall cell lung carcinoma (7), hepatocellular carcinoma (8), renal cell carcinoma (9), pancreatic cancer (10), gastric cancer (11), cervical cancer (12), and colorectal carcinoma (13).In murine tumor models, ablation of Tregs results in activation of CD4 + or CD8 + Teff cells and rejection of solid tumors (14)(15)(16)(17).Tregs therefore represent an attractive target for immunotherapy.Despite recent advances in developing an optimized anti-CD25 antibody with the potential to selectively deplete Tregs in vivo (18), there are currently no Treg-targeting therapies approved for clinical use.
The immunosuppressive tumor microenvironment, to which Tregs are a major contributing factor, presents a significant barrier to effective antitumor immunity (19).The stability of Tregs defined by their ability to maintain expression of Foxp3, is a requirement to sustain their suppressive function (20).Tregs exhibit high levels of lineage stability under steady state conditions and in experimental models of Th1 and autoimmune inflammation (21).However, Treg lineage instability has been observed in some conditions.Zhou and colleagues (22) observed accumulation of "Ex-Foxp3" IL17-expressing cells in inflamed joints in response to synovial IL6.Conversion of purified populations of Foxp3 + Tregs into Foxp3 � cells has also been observed in adoptive transfer models, both during conditions of extreme inflammation induced through allogeneic bone marrow transplantation (23) and extensive lymphopenia-induced proliferation (24).
Tregs can also lose suppressive capacity or gain the ability to produce effector cytokines while retaining Foxp3 expression (20).This has been referred to as Treg fragility.Agonists of the costimulatory receptor Glucocorticoid-Induced TNFR-Related (GITR) have been proposed to drive fragility of Tregs through the reduction of Helios expression and induction of IFNγ and TNFα expressions by Foxp3 + Tregs (25).Neuropilin 1, a receptor expressed by Tregs, suppresses Akt signaling at inflammatory sites to promote Treg functional stability (26).This is particularly the case in the context of tumor-associated Tregs, such that the loss of neuropilin 1 drives Treg fragility and promotes antitumor responses (27).CD137 (4-1BB), a member of the TNF receptor (TNFR) superfamily, is expressed on several cell types, including activated T cells and Tregs.CD137 costimulates activated T cells, resulting in proliferation, memory cell formation, increased survival, and the production of proinflammatory cytokines (28).CD137 agonistic antibodies can provoke rejection of tumors in multiple mouse models but nonspecific agonism results in generalized T cell activation, cytokine release, and systemic inflammation (29).Despite the initial signs of efficacy, clinical development of the CD137 agonist urelumab has been hampered by inflammatory liver toxicity at moderate systemic doses, whereas utomilumab functions as a less potent agonist despite a favorable safety profile (30).Thus, there is a need to develop more specific approaches to activate CD137 signaling.
OX40 is also a costimulatory molecule belonging to the TNFR superfamily (31).It is constitutively expressed by Tregs but not by resting conventional T (Tconv) cells (32).Upon immune activation, however, activated Teff cells gain OX40 expression and costimulatory signaling via OX40/OX40 ligand engagement, which supports survival, differentiation, and memory phenotype transition (33).Stimulation of OX40 on Tregs has been shown to interfere with their regulatory functions (34)(35)(36), and OX40 signaling may contribute to competitive fitness in cellular reconstitution models, indicating an important role in Treg proliferation and survival (37).Tregs expanded by OX40 stimulation are poorly suppressive, due to a relative deficiency of IL2 signaling and an "exhausted" phenotype, which requires exogenous IL2 to overcome (38).Clinical trials using OX40 agonist antibodies exhibited peripheral CD4 + and CD8 + T-cell activation and proliferation without toxicity but showed limited clinical efficacy (39,40).Mechanisms to improve the clinical efficacy of agonist antibodies are therefore of significant interest.
TNFR superfamily agonist antibodies generally require secondary crosslinking of antibody-receptor complexes to induce sufficient receptor clustering and activation by imitating the TNF superfamily ligand superclusters (41).Secondary cross-linking in vivo usually requires interaction of agonistic antibodies with Fcγ receptors (FcγR; ref. 42).The low-affinity interaction between FcγRs and the fragment crystallizable (Fc) region of IgG antibodies and the relatively low abundance of FcγR-expressing cells in the tumor microenvironment are both limitations to agonist activity and resulting antitumor immunity.Consideration should also be given to the antibodymediated effector functions that are induced by interaction with FcγRs which could result in the depletion of tumor-reactive T cells (43).Methods to improve the clinical activity of agonist antibodies targeting receptors like OX40 and CD137 without inducing toxicity are therefore required.
Targeting both OX40 and CD137 together is of interest due to their overlapping but distinct expression profiles (44).FS120 is a novel tetravalent bispecific antibody targeting OX40 and CD137 (45).Previous work has shown that it can activate both CD4 + and CD8 + T cells.The introduction of FcγR-disabling LALA mutations (46) prevents antibody cross-linking through FcγR binding and limits receptor cross-linking until coengagement of the bispecific antibody with both OX40 and CD137, most frequently when they are coexpressed on the same cell.This provides celltype specificity and avoids the depletion of OX40-or CD137-expressing cells.
Importantly, a mouse-specific surrogate version of FS120 (hereinafter referred to as FS120m) induced antitumor activity in a CT26 tumor model, which was associated with reduced liver T-cell infiltration when compared with other agonist antibodies targeting CD137, as an indication of reduced off-target toxicity (45).
Here, we show that the efficacy of OX40/CD137 bispecific agonist FS120m is

Mice and reagents
Foxp3 EGFP-Cre-ERT2 , Rosa26 flSTOPfl-tdRFP , Foxp3 IRES-EGFP , Ptprc a (CD45.1),and Rag2 �/� mice were obtained from The Jackson Laboratory.Ifng flox/flox mice were generated as previously described (47).C57BL/6 mice were purchased from Charles River Laboratories and housed in the University Biological Services (UBS) Gurdon facility (University of Cambridge) for at least 1 week of acclimatization period prior to the start of the experiment.Genetically modified animals were crossed in the Biological Support Unit at the Babraham Institute and at the UBS Gurdon facility to obtain Ifng WT Foxp3 EGFP-Cre-ERT2 Rosa26 flSTOPfl-tdRFP and Ifng flox/flox Foxp3

Antibodies
The bispecific antibody targeting mouse OX40 and CD137 (FS120m) and its isotype control were designed, produced, subjected to quality control, and provided by F-Star Therapeutics (45).All antibodies were diluted from their stock concentrations in PBS and injected in volumes of 100 μL at the timepoints indicated above.

Lymphocyte isolation
Single-cell suspensions from lymphoid tissues were prepared by mechanical dissociation through 40-μm cell strainers (Thermo Fisher Scientific).
Erythrocytes in splenocyte and blood samples were lysed using ice-cold ACK lysing buffer (Gibco) for 3 to 5 minutes before the samples were filtered for a second time.Tumors were dissected and minced in media containing 20 μg/ mL DNase I (Roche) and 1 mg/mL collagenase (Sigma-Aldrich), followed by incubation with agitation for 30 minutes at 37 °C.Digested tissue was then passed through a 40-μm cell strainer and pelleted by centrifugation.Lymphocytes were further isolated by using a density gradient with Percoll solution (Cytiva, 17-0891-01).Nine parts of Percoll were mixed with one part 10� PBS to create a stock solution.The pelleted cells were resuspended with 5 mL of 40% Percoll (made by mixing 40% stock solution with 60% complete RPMI 1640; Invitrogen, Thermo Fisher Scientific), an then a 1 mL underlayer of 80% Percoll was gently introduced.Centrifugation was carried out for 23 minutes at 2,300 rpm at room temperature.Lymphocytes were isolated from the interface and washed with PBS.

Flow cytometry
Cells requiring intracellular staining of cytokines were stimulated prior to flow cytometry analysis using phorbol 12-myristate 13-acetate, ionomycin, and brefeldin A (all from Sigma-Aldrich) for 4 hours in RPMI supplemented with 10% FBS and antibiotics.Viable cells were discriminated by staining with Fixable Viability Dye eFluor 780 (eBioscience) alongside surface-only antibodies in PBS containing 2% FBS and 0.2 mmol/L EDTA (Invitrogen).
Some viable cells were discriminated by staining with Live/Dead Blue or Live/Dead Violet fixable stain in PBS (Thermo Fisher Scientific), or with DAPI.Samples for which the preservation of fluorescent proteins was not necessary were then fixed using the eBioscience Foxp3/Transcription Factor Staining Buffer Set (Invitrogen, Thermo Fisher Scientific) before the rest of the markers were stained intracellularly overnight (48).For samples that required preservation of fluorescent proteins during intracellular staining, the cells were fixed using the BD Fixation/Permeabilization Solution Kit (BD Biosciences).For this, after the first fixation step, the cells were frozen overnight before thawing to permeabilize, refixing, and then staining intra-

FACS
Pre-enrichment of CD4 + T cells from single-cell suspensions was done using the MagniSort Mouse CD4 + T Cell Enrichment Kit (Invitrogen, Thermo Fisher Scientific) according to the manufacturer's protocol.Any markers required for cell sorting were stained using flow cytometry cell surface antibodies after enrichment and incubated alongside eFluor 780 fixable viability dye for discrimination of dead cells.Cell sorting was performed using a BD Aria instrument (BD Biosciences).The cells were sorted into solutions of complete RPMI 1640 supplemented with an additional 25% FBS (Sigma-Aldrich) before being prepared for experiments as described below.

Treg suppression assay
The suppressive capacity of Tregs treated with FS120m was tested as previously described (49).Foxp3 IRES-EGFP mice were dosed with FS120m or isotype control antibody (1 mg/kg) intraperitoneally on days 1, 3, and 5. On Cell division was evaluated by flow cytometry after 4 days of culture.

IL2 in vitro culture assay
Foxp3 IRES-GFP mice were dosed with FS120m or isotype control antibody (1 mg/kg) intraperitoneally on days 1, 3, and 5, and on day 9, GFP + Tregs were sorted by FACS.The sorted Tregs were then cultured in complete RPMI supplemented with IL2 in indicated concentrations for 72 hours before analysis by flow cytometry.

Statistical testing
Data were analyzed using unpaired two-tailed Student t test, one-way or twoway ANOVA as indicated.Bonferroni correction or Tukey correction for multiple comparisons were applied where stated.For tumor experiments, female mice were randomized, and the operator was blinded to genotype while conducting tumor measurements.Mice were randomised to treatment groups following measurement on days 7 to 9, and animals without tumors were removed from the experiment before treatment started on day 10.
Samples used for flow cytometry were also blinded to genotype and treatment group during dissection and remained so until the final stage of analysis.Experimental sample sizes were chosen using power calculations or preliminary experiments, or were based on previous experience of variability in similar experiments.Samples which had undergone technical failure during processing were excluded from subsequent analysis.

OX40 and CD137 are highly coexpressed on tumor-associated Tregs
We first examined the expression profile of OX40 and CD137 on different T-cell subsets within tumors and lymphatics, given that FS120 activity is enriched on cells coexpressing OX40 and CD137.We injected MC38 colorectal carcinoma cells subcutaneously into C57BL/6 mice and allowed tumors to develop for 18 days.Resting (CD44 � CD62L + ) and activated (CD44 + CD62L � ) CD8 + T cells, CD4 + Tconv cells, and Tregs were analyzed by flow cytometry (Fig. 1A).We found that Tregs had the highest percentage of OX40 + CD137 + double-positive cells in both spleens and tumors, whereas the percentage of OX40 + CD137 + cells increased among activated compared with resting CD8 + and CD4 + Tconv cells (Fig. 1B and C).Strikingly, more than 80% of activated Tregs from tumors were double-positive for OX40 and CD137 expression.
Tregs also tended to have the largest percentage of OX40 + and CD137 + singlepositive cells among the cell types examined (Supplementary Fig. S1).This led us to hypothesize that therapies targeting coexpression of these costimulatory molecules may have a greater effect on Tregs than they do on CD4 + Tconv and CD8 + T cells.

OX40/CD137 dual agonism with FS120m drives functional fragility and lineage instability of Tregs
Tregs are a remarkably stable lineage, capable of maintaining expression of Foxp3 even in the presence of inflammatory signaling (21).However, in certain settings, a proportion of Tregs, known as exTregs, lose Foxp3 expression and acquire a proinflammatory phenotype (20,26).To assess lineage stability of Tregs when treated with OX40/CD137 bispecific agonist FS120m, we utilized Foxp3 EGFP-Cre-ERT2 Rosa26 fl-STOP-fl-tdRFP lineage-tracking reporter mice (50).In these mice, Cre recombinase activity induced within for GFP and RFP expression (Fig. 2B).We found that the percentage of exTregs (defined as the percentage of RFP + GFP � cells out of all RFP + cells) was significantly increased in the spleens of mice treated with FS120m, whereas in tumors, there was a trend toward an increase (Fig. 2C).The percentage of CD8 + T cells infiltrating tumors of FS120mtreated animals was also increased, but the percentage of Foxp3 + RFP + Tregs was not significantly altered (Supplementary Fig. S2A and-S2B).Quantitation of the absolute numbers of Foxp3 � RFP + exTreg and Foxp3 + RFP + Treg populations revealed a significant increase in both cell types per gram of tumor in FS120m-treated mice, reflecting a general increase in lymphocyte infiltration induced by FS120m treatment.We also found that there was a significant increase in exTreg numbers in spleens of treated animals, but this was not accompanied by changes in the Foxp3 + RFP + Treg population (Supplementary Fig. S2C).These findings indicate that treatment with FS120m induces instability of Tregs, associated with an increase in the abundance of exTregs, but this is not associated with a concomitant decrease in Foxp3 + Tregs.
The increase in Foxp3 � exTregs indicated that FS120m treatment can induce Treg lineage instability, which has been associated with the upregulation of proinflammatory cytokine expression.We therefore assessed the expression of cytokines IFNγ and TNF in exTregs, Foxp3 + RFP + Tregs, CD4 + Tconv cells and CD8 + T cells.FS120m treatment resulted in a general increase in proinflammatory cytokine expression in all cell types analyzed, particularly those within tumors (Fig. 2D and E).In particular, we noted an increase in the percentage of Foxp3 + RFP + Tregs, alongside Foxp3 � exTregs, expressing these cytokines.This indicates that bispecific agonism of OX40 and CD137 with FS120m can induce Treg fragility, whereby Tregs expressing Foxp3 begin to express proinflammatory cytokines, in addition to inducing inflammatory cytokine expression by lineage-instable exTregs.
Importantly, we also noted that although there were general increases in the absolute numbers of cytokine-producing exTregs and Foxp3 + RFP + Tregs in both tumor and the spleen, it was the "fragile" cytokine-producing Foxp3 + Tregs which were more numerous (Supplementary Fig. S2D and S2E).
To test the suppressive function of Tregs from mice receiving OX40/ CD137 bispecific agonist treatment, we compared Tregs sorted from mice that received FS120m with those from control-treated animals for their ability to suppress the proliferation of naïve CD45.1 Tconv cells (49,51).
Consistent with evidence that Tregs from mice treated with FS120m are functionally fragile, we found that Foxp3 GFP+ cells sorted from the spleens of Foxp3 IRES-EGFP mice that were given FS120m showed reduced suppressive capacity (Fig. 2F).Therefore, we hypothesized that FS120m acts via Tregs to promote antitumor immunity through a direct mechanism involving their production of proinflammatory cytokines and through an indirect mechanism which results from their impaired ability to suppress tumor-targeting Tconv cells.

OX40/CD137 dual agonism results in decreased CD25 expression and diminished IL2 responsiveness of Tregs
We wished to investigate the mechanisms by which treatment with OX40/ CD137 bispecific agonist FS120m induces Treg fragility and instability.It has been previously reported that blockade of IL2 signaling can result in Treg lineage instability (21).However, when we analyzed IL2 production by CD8 + T cells and CD4 + Tconv cells, we found that FS120m induced marginally increased expression among cells from the spleen and tumor-draining lymph nodes (Supplementary Fig. S3).CD25 is the high-affinity IL2 receptor expressed by most Tregs (52)(53)(54).We observed a significant reduction in CD25 expression on Tregs from FS120m-treated mice (Fig. 3A).We hypothesized that this could be responsible for the instable phenotype of Tregs induced by FS120m treatment and that treatment with excess exogenous IL2 would be able to overcome the deficiency in signaling and restabilize the cells.To test if excess IL2 administered in vivo is capable of reversing the instability phenotype of Tregs from FS120m-treated mice, we treated tumor-bearing Foxp3 EGFP-Cre-ERT2 Rosa23 flSTOPfl-tdRFP mice with FS120m or isotype control, then treated half of each group with a stabilised IL2:anti-IL-2 complex (rIL2 combined with an anti-IL2 mAb to increase half-life in vivo; ref. 55) on the last 3 days before analysis (Fig. 3C).We found that treatment of mice with FS120m and rIL2/anti-IL2 complex reversed the lineage instability of Tregs seen in FS120m-treated animals (Fig. 3D).This supports the hypothesis that FS120m acts to destabilize Tregs by decreasing IL2 signaling through reduction of surface expression of CD25, and this can be overcome by providing exogenous IL2, which increases signaling through the remaining surface receptors.

OX40/CD137 dual-agonist-driven antitumor immunity is dependent upon IFNγ signaling
We observed that FS120m treatment induces increased expression of IFNγ by a variety of T cell lineages, including Tregs.To evaluate the requirement for IFNγ in antitumor responses driven by FS120m treatment, we treated MC38 tumor-bearing C57BL/6 mice with FS120m or isotype control and anti-IFNγ-blocking antibody (Fig. 4A).Consistent with our previous findings (Fig. 2A), treatment with FS120m resulted in significantly reduced tumor growth.However, anti-IFNγ treatment was able to completely reverse this phenotype, showing that IFNγ signaling is essential for the antitumor efficacy of FS120m (Fig. 4B).

Treg fragility driven by OX40/CD137 dual agonism is partially dependent upon IFNγ signaling
Because blockade of IFNγ signaling could reverse the antitumor efficacy of OX40/CD137 bispecific agonist treatment, we wished to assess whether IFNγ drives changes to the T cell compartment upon treatment with FS120m.Treatment resulted in increased expression of IFNγ by CD8 + T cells, CD4 + Tconv cells, and Tregs from lymphoid tissues and tumors (Fig. 4C and D).However, combined treatment with FS120m and anti-IFNγ resulted in reduced expression of IFNγ by Tregs compared with FS120m treatment alone, a phenomenon which was only observed consistently in the Treg lineage.This indicates that IFNγ signaling is required to promote Treg fragility upon FS120m treatment.
We next asked if induction of Treg fragility is a general property of immunotherapy responses.We compared the ability of FS120m and anti-PD1 therapy to induce IFNγ expression by analyzing the expression of IFNγ in CD4 + Tconv and Tregs from the blood of animals treated with FS120m or anti-PD1.We found that FS120m was capable of inducing IFNγ expression in both cell types, whereas anti-PD1 treatment did not share this ability (Fig. 5A and B).This identifies a mechanism of action for OX40/CD137 bispecific agonist treatment that is distinct from that of anti-PD1 and identifies a potential biomarker of FS120m activity.

IFNγ production by Tregs and/or exTregs is required for full efficacy of OX40/CD137 dual-agonist therapy
Our results showed that Tregs are functionally reprogrammed to produce IFNγ upon OX40/CD137 bispecific agonist treatment, whose efficacy was dependent upon IFNγ production.We therefore wished to test whether the efficacy of FS120m is dependent on IFNγ produced specifically by Tregs and their lineage-instable progeny.To do this, we crossed Foxp3 EGFP-Cre-ERT2 mice with Ifng fl/fl mice (47).PCR analysis was used to confirm Treg-specific Cre recombinase activity in this model, given that other constitutive Foxp3 Cre alleles can induce substantial nonspecific/leaky excision of floxed alleles in non-Treg lineages (56).We treated Ifng fl/fl Foxp3 EGFP-Cre-ERT2 and Ifng WT Fox-p3 EGFP-Cre-ERT2 animals with tamoxifen and sorted CD8 + T cells, CD4 + Foxp3-GFP � Tconv cells, and CD4 + Foxp3-GFP + Tregs by FACS.Excision of the Ifng gene was observed only in Tregs from Ifng fl/fl Foxp3 EGFP-Cre-ERT2 mice (Fig. 6A), validating use of this system.
To determine whether the efficacy of FS120m is dependent on IFNγ produced specifically by Tregs and their lineage-instable progeny, we treated Ifng fl/fl Foxp3 EGFP-Cre-ERT2 and Ifng WT Foxp3 EGFP-Cre-ERT2 mice with tamoxifen before tumor implantation.The mice were then treated with FS120m, and tumor growth was monitored (Fig. 6B).Strikingly, FS120m treatment induced a more effective antitumor response in Ifng WT Foxp3 EGFP-Cre-ERT2 mice than in Ifng fl/fl Foxp3 EGFP-Cre-ERT2 mice (Fig. 6C).
The analysis of IFNγ expression by Tregs showed a complete loss of IFNγ + Foxp3 + Tregs from Ifng fl/fl Foxp3 EGFP-Cre-ERT2 mice after treatment with tamoxifen (Fig. 6D).We also found that although treatment with FS120m was associated with an increase in the numbers of infiltrating T cells, deletion of IFNγ within Foxp3 + Tregs was associated with a reduction in infiltration by CD8 + T cells (Fig. 6E).Because the loss of IFNγ expression was confined to the Foxp3 + Treg population, we hypothesize that IFNγ production by fragile and instable Tregs may contribute to increased infiltration of the tumor by CD8 + T cells in FS120m-treated mice, resulting in enhanced antitumor immunity.Combined with results indicating that antitumor efficacy is reduced when Tregs are unable to produce IFNγ, we conclude that FS120m exerts its mechanism of action in part by destabilizing Tregs and causing their functional fragility.This reduces their suppressive capability and results in the production of IFNγ which subsequently aids the antitumor response.
Together with its direct effects on CD4 + Tconv cells and CD8 + T cells, this results in enhancement of antitumor immunity.

Discussion
Overall, our data shows that treatment with the bispecific OX40/ CD137 dual-agonist FS120m induces functional reprogramming of murine Tregs, resulting in their production of effector cytokines and loss of their suppressive function, thus promoting tumor regression through both direct and indirect mechanisms.However, we found that the absolute number of IFNγ-producing Tregs was greater than that of lineage-instable exTregs upon FS120m treatment, suggesting that fragile Tregs make a more significant contribution to the therapeutic efficacy of OX40/CD137 bispecific agonism.
A key area for future study would be to ask why IFNγ produced by Tregs is so important in driving the therapeutic efficacy of FS120m, given that other Teff-cell subsets (e.g., CD8 + T cells) produced more IFNγ upon treatment.It In this study, we find that the efficacy of OX40/CD137 dual agonism is in large part dependent upon IFNγ production by Tregs induced upon FS120m treatment.Anti-PD1 therapy did not induce such Treg fragility, revealing a distinct mode of action and supporting combination therapy approaches that exploit this to improve patient outcomes.Indeed, FS120 is currently being evaluated alone, or in combination with anti-PD1 antibodies (pembrolizumab) in a phase I firstin-human study in patients with advanced malignancies (NCT04648202).
partially dependent upon Treg functional reprogramming into fragile and lineage-instable Tregs producing IFNγ.Using Foxp3 fate-tracking reporter mice, we find that Tregs from mice treated with FS120m undergo reprogramming into fragile Foxp3 + IFNγ + cells with decreased suppressive function and instable IFNγ + Foxp3 � exTregs.Treg fragility is partially dependent upon IFNγ signaling, whereas Treg lineage instability is associated with reduced CD25 expression by Tregs upon FS120m treatment in vivo.The efficacy of FS120m therapy is abolished upon antibody blockade of IFNγ, and strikingly, the conditional deletion of Ifng in Foxp3 + Tregs and their progeny in large part reverses the antitumor efficacy of OX40/CD137 bispecific agonist FS120m.
Foxp3 + Treg cells upon tamoxifen treatment removes a stop codon flanked by LoxP sites in the Rosa26 locus to induce red fluorescent protein (RFP) expression, which is maintained even if Foxp3 is downregulated.Therefore, after tamoxifen-induced labeling, Tregs become GFP + and RFP + but Tregs which subsequently lose Foxp3 expression (known as exTregs) become RFP + and GFP � , indicating lineage instability.We treated Foxp3 EGFP-Cre-ERT2 Rosa26 flSTOPfl-tdRFP mice with tamoxifen 10 days prior to injection with MC38 cells.The tumor-bearing mice were then treated with FS120m or isotype control antibodies from day 10 after tumor injection.Strikingly, treatment with FS120m resulted in significantly reduced tumor growth using the MC38 tumor model, similar to previous findings using a CT26 tumor model (Fig. 2A; ref. 45).At 21 days after tumor injection, Tregs from both the tumor and spleen were assessed

First, we
cultured GFP + Tregs sorted from FS120m and control-treated Fox-p3 IRES-EGFP animals in vitro in the presence of titrated quantities of IL2.The lowest concentrations of IL2 resulted in impaired survival of Treg cells, consistent with a critical role for IL2 in limiting Treg apoptosis(32).Tregs cultured with the highest concentrations of IL2 had improved survival and minimal evidence of lineage instability as indicated by high levels of Foxp3-GFP expression.Intriguingly, at intermediate concentrations of IL2, Tregs that had been sorted from mice treated with FS120m exhibited lower Foxp3-GFP expression and reduced cell counts than Tregs from control-treated animals (Fig.3B).This

FIGURE 1 )FIGURE 2
FIGURE 1 OX40 and CD137 are highly coexpressed on tumor-associated Tregs.A, Representative plots showing expression of OX40 and CD137 on resting and activated CD4 + Tregs, CD4 + Tconv cells, and CD8 + T cells from the spleens and tumors of mice inoculated with MC38 cells.B and C, Quantification of resting CD44 � CD62L + (left) and activated CD44 + CD62L � (right) cells within the indicated T-cell populations in the spleen (B) and tumor (C).Data in B and C were analyzed by one-way ANOVA with Tukey correction for multiple comparisons.Bars and error are mean and SEM.****, P ≤ 0.0001.

FIGURE 3 FIGURE 4
FIGURE 3 OX40/CD137 dual agonism results in decreased CD25 expression and IL2 responsiveness of Tregs.A, Representative histograms and replicate measurements showing CD25 expression on CD4 + Tregs from the spleen (left) and tumor-draining lymph node (right).B, Schema showing setup of ex vivo IL2 assay (left), replicate measurements of Foxp3-GFP expression (center), and Treg count (right) after 4 days in culture with the indicated concentrations of IL2.C, Schema representing treatment schedule of tamoxifen, FS120m, and rIL2/anti-IL2 mAb complex.D, Representative plots (left) and replicate measurements (right) of the percentage of RFP + SP exTregs (GFP � ) out of total RFP + cells in the tumors and spleens of reporter mice treated according to C. Data were analyzed by unpaired Student t test (A), unpaired Student t test with Bonferroni-Dunn correction for multiple comparisons (B), and one-way ANOVA with Tukey correction for multiple comparisons (D).Bars and error are mean and SEM.*, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.0001; ****, P ≤ 0.0001.Ctrl, control; dLN, tumor-draining lymph node.

FIGURE 5
FIGURE 5 Treg instability is induced by OX40/CD137 bispecific agonist (FS120m) and not anti-PD1 treatment.A, Representative plots showing the frequency of IFNγ + cells of CD4 + Foxp3 � Tconv cells and Foxp3 + Tregs in the blood on day 18 after MC38 tumor implantation.B, Quantification and statistical analysis of data shown in A. Data were analyzed by ordinary one-way ANOVA (B) with Tukey correction for multiple comparisons.Bars and error are mean and SEM.*, P ≤ 0.05; ***, P ≤ 0.001; ****, P ≤ 0.0001.

FIGURE 6
FIGURE 6 The antitumor efficacy of OX40/CD137 dual agonism is partially dependent upon IFNγ production by Tregs and/or their lineage-instable progeny.A, PCR genotyping of CD8 + , CD4 + Foxp3-GFP � Tconv, and CD4 + Foxp3-GFP + Tregs sorted from the spleens of mice with the indicated genotypes after treatment with tamoxifen.Tregs have cell-specific excision of IFNγ only in mice possessing both the Ifng fl/fl and Foxp3 EGFP-Cre-ERT2 alleles.B, Schema representing treatment schedule of tamoxifen and FS120m.C, Tumor measurements at indicated timepoints after MC38 implantation of Ifng fl/fl Foxp3 EGFP-Cre-ERT2 mice and controls given the indicated treatment.D, Replicate measurements of IFNγ + cells in the indicated CD4 + T-cell populations from the tumor.E, Percentages (top) and counts per gram (bottom) of the indicated cell types in the tumors of mice of the indicated genotypes given the specified treatment.Data were analyzed by two-way ANOVA with Tukey correction for multiple comparisons (C) and ordinary one-way ANOVA with Tukey correction for multiple comparisons (D and E).Bars and error are mean and SEM.*, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001.Ctrl, control.

FS120m
induces functional fragility.This aligns with previously published findings using therapeutics which target OX40 and CD137 individually.In one such study, in vivo anti-OX40 treatment-induced IFNγ and granzyme B expression in Tregs, which was associated with increased T-box expressed in T cells (T-bet) expression(58).In another, anti-CD137 treatment-induced expression of Th-characteristic markers in Tregs that had maintained Foxp3 expression, but which had also begun to express granzyme B, the transcription factor Eomesodermin and TNFα(59).Thus, both molecules targeted by FS120m have the individual capacity to induce some level of Treg fragility.However, when treatment with agonist antibodies targeting OX40 or CD137 individually were compared with FS120m in a CT26 tumor model, it was found that only FS120m was capable of inducing profound antitumor activity(45), indicating that it is only through targeting both pathways simultaneously that tumor regression can be achieved.The proinflammatory cytokine IFNγ has previously been shown to driveTreg fragility(27).Consistent with these observations, we found that blockade of IFNγ signaling suppressed the increase in IFNγ production induced in Tregs by FS120m treatment.When the Ifng gene was specifically disrupted within Tregs, FS120m treatment efficacy was markedly reduced.Importantly, current approaches do not provide the ability to specifically dissect the functional contribution of IFNγ produced by fragile Foxp3 + Tregs and their lineage-instable Foxp3 � exTreg progeny.
is possible that Tregs and Tconv cells occupy distinct anatomic or microanatomic compartments where the production of IFNγ may yield different and more potent effects.It is relevant in this context to note that Treg fragility and instability were more pronounced in systemic lymphoid tissues than in tumors upon FS120m treatment.It may be that production of IFNγ by Tregs in the systemic lymphoid compartment, tumor-draining lymph nodes, or tertiary lymphoid organs triggers more potent effects on antitumor immunity than production of IFNγ by Tregs in the tumor itself, promoting FS120m therapeutic efficacy.Indeed, experimental ablation of IFNγ production by Tregs was associated with reduced treatment-induced accumulation of CD8 + T cells within tumors, which could indicate not only that Treg fragility is required for FS120m to exert its antitumor effects but also that it does so through providing proinflammatory signals that enhance tumor infiltration by cytotoxic CD8 + T cells, themselves also shown to have enhanced cytokine production during FS120m treatment.Whether this reflects increased priming of CD8 + T cells or improved migration or expansion of CD8 + T cells within tumors is unclear.It is plausible that this signal is provided by Tregs within the lymphoid compartment.