Challenges and opportunities in the development of combination immunotherapy with OX40 agonists

Abstract

Introduction:

Costimulatory members of the tumor necrosis factor receptor family, such as OX40 (CD134), provide essential survival and differentiation signals that enhance T cell function. Specifically, OX40 (CD134) agonists stimulate potent anti-tumor immunity in a variety of preclinical models but their therapeutic impact in patients with advanced malignancies has been limited thus far.

Areas covered:

In this review, we discuss the current state of combination immunotherapy with OX40 agonists including preclinical studies and recent clinical trials. We also discuss the strengths and limitations of these approaches and provide insight into alternatives that may help enhance the efficacy of combination OX40 agonist immunotherapy.

Expert opinion:

OX40 agonist immunotherapy has not yet demonstrated significant clinical activity as a monotherapy or in combination with immune checkpoint blockade (ICB), likely due to several factors including the timing of administration, drug potency, and selection of agents for combination therapy clinical trials. We believe that careful consideration of the biological mechanisms regulating OX40 expression and function may help inform new approaches, particularly in combination with novel agents, capable of increasing the therapeutic efficacy of this approach.

1. Introduction

Productive T cell activation requires the T cell receptor (TCR) mediated recognition of cognate peptide-major histocompatibility complexes (pMHC) on antigen presenting cells (APCs) (signal one) in the presence of costimulation, primarily through CD28 engagement by B7 ligands (signal two) [1–3]. While these two signals help initiate T cell proliferation, additional costimulatory signals (signal three) are needed to drive optimal T cell activation, differentiation, and survival [4,5]. Specifically, our group and others have explored the mechanisms by which agonist therapies targeting members of the tumor necrosis factor receptor superfamily (TNFRSF) including OX40 (CD134), 4–1BB (CD137), and GITR (CD357) augment T cell-specific effector function and subsequent anti-tumor immunity. In the current review, we discuss the clinical development of OX40 agonists with a focus on the challenges and opportunities associated with combination immunotherapy using OX40 agonists.

2. OX40 expression and the biological effects of OX40 engagement

OX40-mediated costimulation of activated CD4⁺ and CD8⁺ T cells significantly boosts their proliferation, cytokine production (e.g, IL-2, IFN-γ, TNF-α), effector function (granzyme B, cytolytic activity), and generation of long-lived memory cells (Figure 1) [6–8]. OX40 is also expressed on regulatory Foxp3⁺ CD4⁺ T cells (Tregs). Agonist OX40 therapy can have differential effects on Tregs as some preclinical models have shown that it can alter Foxp3 expression, alleviate Treg suppressive activity and/or induce Treg depletion [9–14], while others have not observed any change in Treg function or even shown Treg expansion depending upon the local cytokine milieu [15–17]. OX40 expression has also been observed on natural killer (NK) cells and OX40 ligation can boost NK cell effector function [18,19].

Figure 1. Effects of OX40 agonist immunotherapy within the TME.

OX40 agonist therapy has pleiotropic effects within the TME including: A) enhanced effector CD4⁺ T cell function (e.g., IL-2, IL-2R expression); B) reduced Foxp3 expression and/or Treg function; and C) increased CD8⁺ T cell effector function (e.g., IL-2, IFN-γ, granzymes) resulting in cancer cell death. D) OX40 expression and signaling is coregulated by TCR and IL-2/IL-2R signaling. E) The effects of OX40 agonist therapy in the presence of aCTLA-4 ICB are further augmented by ITK inhibition, leading to the generation of memory-like effector Eomes⁺ CD8⁺ T cells. F) Abrogation of galectin-3-mediated suppression of TCR signaling can also enhance the efficacy of OX40 agonist immunotherapy. Created with BioRender.com. TME: tumor microenvironment; IL-2: interleukin-2; IL-2R: interleukin-2 receptor; IFN-γ: interferon gamma; TCR: T cell receptor; CTLA-4: cytotoxic T-lymphocyte associated protein 4; ICB: immune checkpoint blockade; ITK: IL-2 inducible T cell kinase; Eomes: Eomesodermin.

Given the potent functional consequences of OX40 ligation, OX40 expression is tightly regulated to prevent autoimmunity. OX40 is controlled through the integration of multiple signals including TCR engagement and cytokine signaling [20–22]. TCR signaling drives the initial upregulation of OX40, which peaks approximately 48–72 hours post-TCR engagement and is more prominent on CD4⁺ T cells as compared to CD8⁺ T cells [23]. While the magnitude of OX40 expression is related to the strength of TCR stimulation [24], TCR stimulation alone is not sufficient to induce OX40. Instead, there is a coordinated cascade of signals needed to induce OX40 that starts with TCR ligation, which leads to increased interleukin-2 receptor alpha (IL-2Rα; CD25) expression, thereby forming the high-affinity trimeric IL-2R complex. The subsequent combination of TCR and IL-2/IL-2R signaling drives maximal OX40 expression through a JAK3-STAT5-dependent pathway in both murine and human T cells. Other common gamma chain cytokines (e.g., IL-4, IL-7, and IL-21) also contribute to increased OX40 expression, albeit less efficiently than IL-2 [21,22,25,26].

These signals (TCR+IL-2R+OX40) initiate a cascade of downstream signaling events including TRAF2-mediated canonical NF-κβ activation and induction of the non-canonical NF-κβ induced kinase (NIK) [27–29]. OX40 and IL-2R signaling also promote PI3K/AKT signaling, which supports T cell survival through the upregulation of anti-apoptotic molecules including Bcl-2, Bcl-xL, and survivin [25,30–32]. OX40 signaling also initiates a downstream program associated with the acquisition of critical T cell effector functions including cytokine production (IFN-γ, IL-2, IL-4, IL-9), granzyme B expression, and increased cytolytic function [29,33–37]. Since OX40 ligation leads to the generation of potent T cell effector functions, multiple studies have explored the ability of an agonist anti-OX40 (aOX40) mAb to boost anti-tumor immunity in preclinical models. These studies demonstrated that aOX40 therapy can induce a significant expansion of tumor-reactive T cells and remodeling of the tumor microenvironment (TME), including altered activity of tumor-associated macrophages (TAMs), which leads to enhanced tumor regression and improved survival of tumor-bearing mice [38–43].

3. Combination immunotherapy with OX40 agonists

Several groups have explored the safety and efficacy of OX40 agonists in patients with metastatic cancer, however, thus far limited clinical responses have been observed [44–50]. The reasons for the low clinical efficacy of aOX40 are unknown but may be related to unique limitations of the agents themselves such as suboptimal dosing or timing of drug administration, or perhaps a broader issue that plagues immunotherapy in general, the multitude of suppressive mechanisms through which tumors can inhibit and/or evade immune-mediated clearance. Indeed, preclinical data revealed that aOX40 monotherapy was not sufficient to rescue the proliferative response of Ag-specific CD8⁺ T cells rendered anergic, or non-responsive, following chronic exposure to self-Ag [51–53]. Instead, the provision of exogenous Ag (through vaccination) was necessary to provide sufficient TCR stimulation to enable responsiveness to aOX40 therapy in these models. Due to the limited clinical efficacy of aOX40 monotherapy, most ongoing clinical trials with OX40 agonists aim to explore their activity in combination with other immune modulating agents, including other T cell agonists, ICB, and/or radiation therapy (RT). Here, we will focus on the current state of combination therapy clinical trials with OX40 agonists and discuss other novel combinations that have yet to be evaluated in clinical trials.

3.1. Agonist aOX40 and ICB

Several groups have explored the efficacy of aOX40 therapy in conjunction with ICB. The underlying hypothesis is that combining a T cell agonist with blockade of co-inhibitory signals would synergize to augment anti-tumor immunity as these are complementary, but not necessarily overlapping, pathways regulating T cell activation. Of the ICB agents that have been examined in combination with OX40 agonists, aCTLA-4 ICB is thought to enhance the initial priming of tumor-reactive T cells in the draining lymph nodes and elsewhere, while aPD-1 ICB likely restores the function of exhausted T cells within the TME [54]. Given the differences in certain aspects of aCTLA-1 and aPD-1 ICB, one might expect that combining either with an OX40 agonist will also have some shared and unique impacts on T cell activation, differentiation, and/or anti-tumor efficacy. For example, combining aOX40 therapy with aCTLA-4 or aPD-1 ICB was sufficient to drive the expansion of IFN-g⁺ CD4⁺ and CD8⁺ T cells and support tumor regression in various tumor models [13,15,55–60]. However, OX40 plus aCTLA-4, but not aPD-1, uniquely induced a population of Eomesodermin⁺ (Eomes⁺) effector CD8⁺ T cells [15,60], which were critical for supporting the anti-tumor efficacy of this particular combination in tumor-bearing mice [60]. In addition, aOX40 therapy increased the frequency of CTLA-4⁺ CD8⁺ T cells in mice, which likely facilitated the synergistic effects of combination aOX40+aCTLA-4 therapy on the generation of effector and memory CD8⁺ T cells [61]. This was unexpected as previous preclinical studies had demonstrated the primary effects of CTLA-4 ICB monotherapy on CD8⁺ T cells occurred indirectly through IL-2/IL-2R signaling via IL-2-producing effector CD4⁺ T cells and not from aCTLA-4 blockade directly on CD8⁺ T cells [62]. With respect to the PD-1/PD-L1 axis, aOX40 therapy promotes PD-L1 expression in the TME, likely through its ability to enhance effector T cell IFN-γ production, which drives PD-L1 expression [15,37,58,63]. Together, these studies provided a strong rationale for exploring the clinical efficacy of aOX40+aCTLA-4 and aOX40+aPD-1 therapy.

Initially, several clinical trials explored the safety and efficacy of OX40 agonists plus CTLA-4 ICB with tremelimumab (NCT02705482; aOX40 Q2W and aCTLA-4 Q4W) or ipilimumab (NCT02737475; given together Q3W) (Table 1) in patients with advanced solid tumors. Unfortunately, minimal clinical responses were observed, and none were greater than expected from aCTLA-4 monotherapy [48,64]. Another study is exploring the safety and efficacy of a bispecific aOX40+aCTLA-4 mAb in patients with advanced solid malignancies (NCT03782467; Q2W), but efficacy data has yet to be reported. Similarly, the delivery of OX40 agonists in conjunction with aPD-L1 ICB (durvalumab; NCT02705482; aOX40 Q2W and aPD-L1 Q4W) or aPD-1 ICB with nivolumab (NCT02737475; given together Q2W) or pembrolizumab (NCT02528357; given together Q3W) provided little or no additional benefit beyond that seen with aPD-L1 or aPD-1 monotherapy [48,64,65] (Table 1).

Table 1.

Completed or terminated clinical trials with OX40 agonists and combination therapy.

Phase	OX40 agonist	Combination therapy	Disease	Lead Organization	NCT Identifier (Ref)
Ib	MEDI6469	Radiation therapy	Metastatic prostate cancer	Providence Health & Services	NCT01303705
I	MEDI6469	Radiation therapy	Metastatic breast cancer	Providence Health & Services	NCT01862900
I	PF-04518600	Utomilumab (a4-1BB)	Advanced solid tumors	Pfizer	NCT02315066 [101]
I	MOXR0916	Atezolizumab (aPD-L1), bevacizumab (aVEGF)	Advanced solid tumors	Genentech	NCT02410512
I	GSK3174998	Pembrolizumab (aPD-1)	Advanced solid tumors	GlaxoSmithKline	NCT02528357 [65]
Ib/II	PF-04518600	Avelumab (aPD-L1)	Advanced solid tumors	Pfizer	NCT02554812
I	MEDI0562	Durvalumab (aPD-L1), tremelimumab (aCTLA-4)	Advanced solid tumors	AstraZeneca	NCT02705482 [64]
I/IIa	BMS-986178	Nivolumab (aPD-1), ipilimumab (aCTLA-4)	Advanced solid tumors	Bristol-Myers Squibb	NCT02737475 [48]
II	MOXR0916	Atezolizumab (aPD-L1)	Urothelial carcinoma	Genentech	NCT03029832
I/II	PF-04518600	Radiation therapy, avelumab (aPD-L1), utomilumab (a4-1BB)	Advanced solid tumors	MD Anderson Cancer Center	NCT03217747
I/II	mRNA-2416 (OX40L)	IL-23, IL-36γ, Durvalumab	Advanced solid tumors	Moderna	NCT03323398
Ib/II	PF04518600	Avelumab (aPD-L1), Azacitidine	Acute myeloid leukemia	MD Anderson Cancer Center	NCT03390296 [147]
I	HS-130 (OX40L)	HS-110 (allogeneic tumor cell vaccine)	Advanced solid tumors	Heat Biologics	NCT04116710
I	ABBV-368	Tilsotolimod (TLR9 agonist), nab-paclitaxel, ABBV-181 (aPD-1)	Recurrent/ metastatic oral cancer	Abbvie	NCT04196283
Ib/II	INCAGN01949	CMD-001 (virus like particles)	Metastatic pancreatic cancer	USC/Norris Comprehensive Cancer Center	NCT04387071

One hypothesis to explain the apparent lack of clinical activity of OX40 agonists is some inherent limitation of the agents themselves. Researchers are addressing this issue by developing novel OX40 agonists that may exhibit enhanced functional properties as compared to the traditional bivalent mAbs. For example, one group generated a novel hexameric OX40 agonist that is thought to provide more robust costimulatory signals than its bispecific aOX40 mAb counterpart, which is being tested alone or in combination with aPD-1 in patients with advanced solid tumors (NCT04198766; Table 2). Others are evaluating the safety and efficacy of a PD1-Fc-OX40L fusion protein (NCT03894618). It is possible that the increased strength of OX40 signaling (with the hexameric aOX40 mAb) or physical proximity of OX40 ligation plus PD-1 ICB with a fusion protein may lead to improved T cell function and subsequent tumor regression. However, it remains to be seen whether these newer OX40 agonist combination therapies will be more successful clinically as compared to previous clinical trials with aOX40 plus ICB.

Table 2.

Active clinical trials with OX40 agonists and combination therapy.

Phase	OX40 agonist	Combination therapy	Disease	Lead Organization	NCT Identifier	Active/enrolling?
II	PF-04518600	Axitinib (tyrosine kinase inhibitor)	Renal cell carcinoma	Pfizer	NCT03092856	Active, not recruiting
I	BMS986178	SD-101 (TLR9 agonist), radiation therapy	Low grade B cell non-Hodgkin lymphoma	Stanford University	NCT03410901	Active, not recruiting
I/II	mRNA-2752 (triplet)	OX40L, IL-23, and IL-36γ	Advanced solid tumors	Moderna	NCT03739931	Recruiting
I	BMS986178	SD101 (TLR9 agonist)	Advanced solid tumors	Stanford University	NCT03831295	Active, not recruiting
I	ABBV-368	ABBV-927 (aCD40), ABBV-181 (aPD1), chemotherapy	Advanced solid tumors	Abbvie	NCT03893955	Active, not recruiting
I	SL-279252	PD1-Fc-OX40L	Advanced solid tumors	Shattuck Lab	NCT03894618	Active, not recruiting
I/II	GSK3174998	Belantamab mafodotin	Multiple myeloma	GlaxoSmithKline	NCT04126200	Recruiting
I/II	INBRX-106	Pembrolizumab (aPD-1)	Advanced solid tumors	Inhibrx Inc.	NCT04198766	Recruiting
Ia/Ib	BGB-A445	Tislelizumab (aPD-1)	Advanced solid tumors	BeiGene	NCT04215978	Recruiting
I	FS120	OX40-CD137 bispecific antibody	Advanced solid tumors	F-star Beta Limited	NCT04648202	Recruiting
I	ES102	Toripalimab (aPD-1)	Advanced solid tumors	Elpiscience Biopharma, Ltd.	NCT04991506	Recruiting

The reasons underlying the apparent lack of clinical benefit of OX40 agonist therapy plus ICB remain unclear but may reflect suboptimal timing of aOX40 administration in relation to ICB (i.e., concurrent versus sequential ICB blockade) and/or differences in the biological activity of specific OX40 agonists. The first issue – the impact of timing of ICB with respect to aOX40 therapy, was raised in several recent publications demonstrating that sequential (aOX40 prior to ICB), but not concurrent, aOX40+ICB therapy was necessary for delaying tumor growth and enhancing survival in mice with mammary carcinoma or human papilloma virus E7-expressing epithelial cancer (Figure 2) [58,66]. In these models, delayed ICB was given the week after the initiation of aOX40 therapy as the rapid rate of tumor growth in these models limits the therapeutic window of opportunity for delayed treatment. Mechanistic studies suggested that increased T cell apoptosis and/or expression of inhibitory receptors likely limited the ability of tumor-reactive T cells to mount a sufficient anti-tumor response when these agents were given concurrently. In contrast, several other groups have demonstrated robust therapeutic efficacy of concurrent aOX40+aCTLA-4 or aOX40+aPD-1/L1 ICB in tumor-bearing mice [15,55,57,59,61]. In these studies, concurrent therapy elicited significant expansion of effector T cells, tumor regression, and improved survival without any apparent reduction in therapeutic efficacy. The reasons for these discrepancies are unknown but may be related to unique aspects of the specific tumor models utilized in the different studies. Most importantly, whether altered sequencing of aOX40 in the context of ICB therapy plays a pivotal role in its subsequent therapeutic efficacy in humans remains unknown.

Figure 2. Model of concurrent versus sequential OX40 agonist immunotherapy regimens.

Preclinical studies have explored concurrent (top) or sequential (middle) treatment with OX40 agonists plus aPD-1/L1 immune checkpoint blockade and/or tumor-specific vaccination in multiple tumor models. Other studies utilitzed a triplet approach (bottom) combining aOX40+aCTLA-4 therapy plus an ITK inhibitor. Assessing outcomes including tumor growth, survival, and T cell responses using flow cytometry and/or single-cell RNA-sequencing provides insight into the similarities and differences among these approaches. Whether a specific treatment regimen is more effective in patients has yet to be determined. Created with BioRender.com. PD-1: programmed cell death protein 1; PD-L1: programmed death ligand 1; CTLA-4: cytotoxic T-lymphocyte associated protein 4; ITK: IL-2 inducible T cell kinase.

The second issue relates to the biological activity of various OX40 agonists. While each of these agents were evaluated in preclinical assays for potency and costimulatory activity, there is likely some variability with respect to their in vivo activity. For example, monotherapy treatment with certain OX40 agonists was shown to drive T cell proliferation in patients [46,59,67] but this was not observed with others [44,47,65]. In addition, there is little evidence for synergistic effects on T cell expansion and/or differentiation following combined aOX40+ICB therapy in humans, for reasons that remain unclear. One possibility is that the frequency of dosing may not be optimal as several OX40 agonists were administered on the same cycle as ICB (e.g., every 2–4 weeks), which may not align with the optimal biological activity of aOX40. Indeed, preclinical studies have shown that OX40 agonists should be administered at the time of vaccination (day one) and again the next day (day two) to support optimal T cell activation and expansion, rather than spaced further apart [68–70]. In contrast, most clinical trials have tested aOX40 given once every 2–4 weeks, often in conjunction with the approved schedule of ICB [48,64,65]. This schedule is practical and convenient for patients as they receive the ICB infusion every 2–3 weeks, which avoids extra visits for patients and helps to minimize costs associated with the trial. It may be the case that altering the schedule of aOX40 therapy to administer multiple doses of the drug over the course of a few days or week followed by a rest period of 4–6 weeks may favor more robust responses as compared to dosing once every 2–4 weeks. Of note, one ongoing clinical trial is exploring the safety and potential efficacy of providing a priming dose of a hexameric OX40 agonist followed by combined aOX40+aPD-1 therapy (NCT04198766). Future studies will be needed to determine whether this or other alternative dosing regimens can enhance the clinical activity of OX40 agonists.

3.2. aOX40 and innate agonists

Several groups have evaluated the therapeutic efficacy of combining OX40 agonists with stimulators of innate immunity such as IL-12 or Toll-like receptor (TLR) agonists. IL-12 is a potent proinflammatory cytokine that provides maturation signals to antigen-presenting cells, while also supporting M1 macrophage polarization and OX40-mediated T cell survival [71,72]. Several preclinical studies have demonstrated increased anti-tumor immunity following OX40 agonist treatment plus recombinant IL-12 [34,73,74]. However, IL-12-mediated toxicity has limited its clinical development and it has yet to be tested with an OX40 agonist in a clinical trial [75]. Treatment with TLR agonists, particularly the TLR9 agonist CpG oligodeoxynucleotides, can also improve anti-tumor immunity, in part through the induction of type I interferons (IFN-α/β) that enhance APC maturation and subsequent T cell priming [76–79]. Combination OX40 agonist plus CpG therapy enhanced T cell-mediated anti-tumor immunity across a variety of preclinical models [80–84]. In some of these studies, the therapeutic benefits were further enhanced by the addition of RT, which has led to clinical trials exploring the combination of OX40 agonists, RT, and CpG (discussed below).

3.3. aOX40 and RT

The anti-tumor effects of OX40 agonists have also been explored in conjunction with RT. RT induces tumor cell death and the subsequent release of tumor-associated antigens have the potential to be processed and presented in the draining lymph nodes by professional APCs to then support T cell priming. RT can also activate the cGAS/STING pathway leading to the production of type I IFNs [85,86]. Preclinical studies have shown that OX40 ligation plus RT in the adjuvant setting boosts the therapeutic efficacy versus RT alone, particularly when given within 48 hours post-RT [87–89]. These results recapitulate the observed benefits of aOX40 given immediately post-vaccination, supporting the general hypothesis that aOX40 therapy is most effective when there is a source of tumor-associated antigens. RT plus aOX40 therapy also led to an increased proportion of antigen-specific CD8⁺ T cells in the peripheral blood of tumor-bearing mice [90], suggesting that this combination elicits an increased pool of tumor-reactive CD8⁺ T cells capable of mediating tumor regression. Further, several groups have explored the impact of CpG in conjunction with aOX40+RT [91]. Several reports previously highlighted the ability of each of the respective dual combinations, aOX40+CpG or RT+CpG, to enhance tumor regression and improve tumor-free survival in preclinical models [80,81,92]. In the case of RT+CpG, favorable responses were also observed in a small cohort of patients with previously untreated indolent lymphoma [93].

Based upon these preclinical studies, several clinical trials were initiated to explore the safety and efficacy of aOX40+RT and/or TLR9 ligation. One study tested the effects of aOX40 plus RT (800 cGy × 1 to up to 3 bone metastatic sites) in men with metastatic prostate cancer. In this study, three doses of aOX40 were given two days apart starting the day of RT (NCT01303705). An additional trial evaluated three different doses of RT (day 1; 15 Gy × 1; 20 Gy × 1; or 20 Gy ×2) plus aOX40 (day 1, 3, 5) in breast cancer patients with liver or lung metastases and progressive disease (NCT01862900). Another study is testing low-dose RT (day 1, 2) plus intratumoral aOX40+TLR9 agonist (days 2, 9, 16, 23, and 30) and aOX40 (IV; days 2, 30, 58, 86, 114, and 142) in low-grade B-cell non-Hodgkin lymphoma (NCT03410901). To date, no results have been published from these studies, suggesting a lack of clinical efficacy above that expected from RT alone. Future clinical trials will be needed to address questions regarding the optimal schedule and dosing of OX40 agonists with RT as well as the optimal dose and schedule (e.g., hypofractionated) of RT most likely to synergize with aOX40.

3.4. Combination therapy with multiple T cell agonists

The anti-tumor efficacy of OX40 agonists have also been tested in combination with other T cells agonists. For example, aOX40 was combined with a4-1BB (CD137), in part because aOX40 can have more pronounced effects on CD4⁺ T cells, while a4-1BB primarily stimulates CD8⁺ T cell and NK cell responses [8,49,50,94,95]. Combined aOX40+a4-1BB was shown to promote more effective anti-tumor immunity than either monotherapy through enhanced T cell function and the induction of more stem-like CD8⁺ T cells in multiple preclinical models [96–100]. Results from a recent clinical trial (NCT02315066) demonstrated that agonist aOX40 (Q2W) plus a4-1BB (Q4W) therapy was safe and well-tolerated in a cohort of 57 patients with advanced malignancies with some evidence of clinical response, albeit in a small subset of patients [101]. It isn’t known whether this combination was more effective than might be expected with T cell agonist therapy plus ICB as these were not tested in a randomized comparison study. It’s also unknown whether an altered treatment schedule would have improved outcomes of this trial.

4. Novel combinations with OX40 agonists

Thus far, clinical trials with aOX40 in combination with ICB and/or RT have not yielded significant improvements in therapeutic efficacy as compared to ICB or RT alone. This limited clinical activity has spurred the investigation of alternative combinatorial strategies, beyond simply adding ICB, that might be capable of synergizing with the costimulatory activity of OX40 agonists. Here, we highlight some of these novel approaches.

4.1. OX40 agonist therapy plus cytokines

Given the coordinated relationship between OX40, IL-2R expression, and IL-2 production, one combination examined in preclinical studies was aOX40+IL-2 agonist therapy. The IL-2 agonist used in this study was an IL-2/aIL-2 mAb complex, which serves to prolong significantly the half-life of the cytokine and directs its binding to the IL-2Rβ subunit (CD122), which is preferentially expressed on CD8⁺ T cells and NK cells [102]. Combination therapy synergized to boost tumor regression in MCA-205 (sarcoma) tumor-bearing mice and rescued the function of anergic tumor-reactive CD8⁺ T cells rendered non-responsive by chronic antigen exposure [21]. Presumably, increased signaling through the IL-2R with IL-2 agonists in the presence of aOX40 supports enhanced T cell survival through upregulation of anti-apoptotic proteins while simultaneously driving effector T cell function. Numerous trials are examining the safety and efficacy of various IL-2 agonists that have the additional potential benefit of reduced toxicity as compared to the only currently FDA-approved IL-2 therapy (for patients with metastatic melanoma or renal cell carcinoma) is high-dose recombinant IL-2 (HD IL-2), which has to be administered in the ICU due to its severe toxicity [103–106]. Whether the addition of aOX40 to these next generation IL-2 agonists will improve upon the ~30% overall response rate observed with HD IL-2 monotherapy remains to be determined [106].

A different study is testing the impact of an mRNA-based intratumoral therapy that incorporates an agonist OX40L with two proinflammatory cytokines, IL-23 and IL-36γ, with the goal of stimulating dendritic cells in the presence of an OX40 agonist to drive more robust anti-tumor responses [107]. The treatment is being given intratumorally with the expectation that local delivery might reduce the frequency and/or severity of any potential adverse effects. This triple therapy is currently being tested in a phase I clinical trial alone and in combination with aPD-L1 in patients with advanced malignancies (NCT03739931).

4.2. aOX40+aCTLA-4 ICB+ITK inhibition

Preclinical studies investigating the mechanisms contributing to the efficacy of combination immunotherapy revealed a significant induction of Eomes⁺ CD8⁺ T cells following aOX40+aCTLA-4, but not aOX40+aPD-1, treatment [15,60]. Eomes is a transcription factor that, along with Tbx21 (T-bet), promotes the acquisition of CD8⁺ T cell effector function, including IFN-γ and granzyme B production, and helps regulate memory CD8⁺ T cell generation [108–111]. Whether Eomes played a functional role in driving the activity of aOX40+aCTLA-4 therapy was examined by removing, or conversely, increasing Eomes expression. Eomes was removed using a CD8⁺ T cell-specific Eomes knockout mouse model, which abrogated the therapeutic efficacy of aOX40+aCTLA-4 therapy [60]. In contrast, Eomes expression was increased by taking advantage of insights regarding its transcriptional regulation. Eomes is transcriptionally repressed by the upstream molecular regulator interferon regulatory factor 4 (IRF4), which is promoted by the interleukin-2-inducible T cell kinase (ITK) [112,113]. Thus, pharmacological inhibition of ITK will reduce IRF4, thereby alleviating the transcriptional repression of Eomes. As hypothesized based on the Eomes-deficient CD8⁺ T cell experiments, preclinical models demonstrated that the addition of an FDA-approved small molecule Bruton’s tyrosine kinase (BTK)/ITK inhibitor, ibrutinib, to aOX40+aCTLA-4 therapy significantly enhanced the induction of cytolytic proliferating (Ki-67⁺) effector (granzyme⁺, IFN-γ⁺) Eomes⁺ CD8⁺ T cells and subsequent tumor regression as compared to aOX40+aCTLA-4. [60]. The safety and potential clinical efficacy of aOX40+aCTLA-4+ibrutinib triple therapy has not yet been evaluated in clinical trials.

4.3. OX40 agonists and arginine depletion

OX40-directed therapy has also been combined with a novel engineered pegylated recombinant arginase-1 enzyme (pegzilarginase) in preclinical murine tumor models. Pegzilarginase was engineered to mediate depletion in arginine auxotrophic tumors, which are unable to synthesize their own arginine and thus require exogenous arginine to sustain their growth and survival [114]. By starving the tumors of this essential nutrient, pegzilarginase directly promotes tumor cell death through increased autophagy and increase MHC I and II expression, which likely increase the supply of tumor-associated antigens available for priming naïve CD8⁺ T cells, while also potentially making tumor cells more susceptible to effector T cell-mediated killing [115]. Thus, one would hypothesize that pegzilarginase treatment may provide a more favorable TME to enhance the activity of aOX40 therapy. In murine tumor models, the immune-stimulatory effects of aOX40+pegzilarginase therapy enhanced CD8⁺ T cell proliferation, effector function (granzyme B expression), and tumor-free survival as compared to either monotherapy. Combination therapy also led to remodeling of the TME to favor M1 macrophage polarization across multiple preclinical tumor models [115]. These data were somewhat paradoxical as one might hypothesize that arginine depletion also affects the tumor-reactive T cells, thereby limiting the efficacy of T cell-directed therapies. However, recent work demonstrated that when arginine is limited, activated T cells can use a salvage pathway of citrulline uptake via L-type amino acid transporter 1 (LAT1) to generate arginine that then supports T cell proliferation [116]. Whether this mechanism contributes to the therapeutic efficacy of aOX40+pegzilarginase remains to be determined. With respect clinical translation of this combination, pegzilarginase has been tested in conjunction with aPD-1 ICB (pembrolizumab) in patients with small cell lung cancer in a recently completed phase I clinical trial (NCT03371979), but no results have been reported. However, the safety and efficacy of aOX40+pegzilarginase has not yet been tested in patients.

4.4. aOX40 and galectin-3 inhibition

Another novel approach is aOX40 therapy in conjunction with blockade of the immune suppressive molecule and lectin family member, galectin-3 (Gal-3). Gal-3 is a β-galactoside binding lectin that is expressed in both intracellular and extracellular compartments [117–119]. In tumors, increased intracellular Gal-3 is associated with enhanced tumor cell proliferation, metastasis, and poor outcomes in patients. Extracellularly, Gal-3 promotes tumor seeding at distant sites and is also highly immunosuppressive, inhibiting TCR signaling, sequestering proinflammatory cytokines within the TME, and supporting M2 macrophage polarization [118,120–130]. Based upon the pleiotropic role for Gal-3-mediated immune suppression, several groups tested the hypothesis that treatment with a Gal-3 inhibitor may synergize with aOX40 or ICB therapy by targeted these related, but non-overlapping immune regulatory pathways. Preclinical studies revealed therapeutic benefit from combining Gal-3 blockade with aPD-1/L1 ICB and a recent phase I clinical trial demonstrated favorable outcomes following Gal-3 inhibition plus aPD-1 (pembrolizumab) therapy in patients with metastatic melanoma or head and neck squamous cell carcinoma (HNSCC) [131–133]. Our data demonstrated that aOX40 plus concurrent treatment with a Gal-3 inhibitor significantly improved tumor-free survival and reduced metastasis in preclinical models as compared to either monotherapy [132]. These effects were associated with increased CD8⁺ T cell infiltration within the tumor and reduced suppressive activity of monocytic myeloid-derived suppressor cells (Mo-MDSCs). However, clinical trials to examine the safety and/or efficacy of combination OX40 agonist therapy and Gal-3 inhibition have not been conducted.

4.5. OX40 agonists and cancer vaccines

Given the ability of TCR ligation to elicit OX40 expression, there has been considerable interest in the ability of OX40 agonists to synergize with cancer vaccines. Ideally, these vaccines would provide an exogenous source of tumor-associated antigens (TAA) capable of inducing OX40 expression on TAA-specific T cells and thereby sensitizing them to the costimulatory effects of OX40 ligation. Numerous studies have highlighted the ability of aOX40+vaccination or OX40L-expressing vaccines to significantly improve tumor regression across a wide variety of tumor models. For example, aOX40+aCTLA-4 ICB plus a dendritic cell HER2-directed vaccine led to increased tumor-free survival in a model of HER2⁺ mammary carcinoma [61]. Other studies revealed potent effects of OX40 agonists plus vaccination alone or in the presence of additional immune modulating signals (e.g., IL-12, aPD-1/L1 ICB, GM-CSF, aCD40, STING, IDO inhibitor, etc.) in models of glioblastoma, sarcoma, mammary carcinoma, prostate cancer, lung cancer, colon cancer, pancreatic cancer, and more [34,97,134–141]. Despite the wide variety of tumor types demonstrating therapeutic benefit from combination aOX40+vaccination, clinical translation of this approach has been hampered by the lack of effective therapeutic cancer vaccines available for testing in a phase I trial. Thus, the potential clinical impact of aOX40+vaccination remains to be determined.

5. Conclusion

In summary, OX40 agonist immunotherapy has shown great potential for enhancing anti-tumor immunity in a wide variety of preclinical tumor models, an effect typically amplified when they are given in combination with ICB or other novel agents. Recent work detailing the mechanisms by which combination aOX40 therapy drives productive tumor-specific T cell responses has informed novel therapeutic strategies that may eventually improve outcomes. Thus, while initial clinical trials with OX40 agonists have not yielded significantly improved outcomes, the development of next-generation OX40-directed therapeutics and novel combinations may alter the landscape of T cell agonist immunotherapy to provide new options and renewed hope for patients with advanced malignancies.

6. Expert opinion

Despite the potent activity of OX40 agonists in preclinical models, OX40 agonists have yet to achieve similar results in patients, either as monotherapy or in combination with ICB. The reasons for this remain obscure but are likely related to a multitude of factors including the schedule of administration, potency of the OX40 agonist being tested, and the availability of TAAs capable of driving OX40 expression on tumor-specific T cells. The first two issues (scheduling and the potency of current OX40 agonists) have been discussed above. The role of TAAs are of particular importance given data demonstrating that exogenous vaccination was needed to drive TCR stimulation and thus support subsequent responsiveness to aOX40 therapy in models of CD8⁺ T cell anergy [51–53]. Thus, while aOX40 monotherapy or in combination with ICB can drive robust T cell activation and expansion in the peripheral blood of patients with advanced malignancies, it may not be sufficient to restore the function of TAA-specific T cells required to mediate tumor regression unless given with a source of exogenous antigen via vaccination. Numerous clinical trials are currently exploring a variety of novel vaccine approaches including personalized vaccine platforms targeting neoantigens that appear promising [142–144]. However, whether OX40-directed therapy in conjunction with tumor-specific vaccination will boost the generation of TAA-specific T cells, their effector differentiation, and/or improve clinical outcomes remains to be determined.

With the maturation of the immuno-oncology (I-O) field in recent years, it has become clear that simply combining I-O agents with ICB is not sufficient to cure metastatic cancer in most patients. Thus, considerable interest has arisen in looking at alternative approaches such as: 1) moving I-O therapies into the neoadjuvant setting to reduce and/or prevent recurrence; and 2) moving beyond ICB to exploring novel therapeutics beyond ICB capable rendering the TME more receptive to the anti-tumor effects of OX40 agonists. Regarding neoadjuvant therapy, preclinical data demonstrated that neoadjuvant therapy with intratumoral aOX40+CpG (TLR7) significantly improved the survival of mammary (4T1) or colorectal carcinoma (CT26) tumor-bearing mice. These effects were further enhanced in the presence of systemic aPD-1 ICB through a CD8⁺ T cell-dependent mechanism [82]. While neoadjuvant aOX40+CpG has not been tested in patients, two other recent phase I trials explored the safety and immunological effects of neoadjuvant aOX40 monotherapy given prior to surgery in patients with HNSCC (NCT02274155) or metastatic melanoma or HNSCC (NCT03336606). In patients with HNSCC, neoadjuvant aOX40 was sufficient to promote T cell proliferation in the peripheral blood and tumor within several weeks post-treatment and did not adversely affect subsequent definitive surgery [67]. Moreover, those patients identified as having an expansion of putative tumor-reactive CD39⁺CD103⁺ CD8⁺ T cells within the tumor tissue did not develop recurrent disease up to 48 months post-treatment [67,145], suggesting that neoadjuvant aOX40 may help reduce recurrence in these patients.

The second approach of evaluating novel combinations beyond ICB (including IL-2, IL-12, galectin-3 inhibition, arginine depletion, CD40 agonists, etc.) has been studied extensively by our group and others in multiple preclinical models. Furthermore, a recent study tested an even more complex combination comprised of sequential therapies seeking to prime TAA-specific T cells via tumor-specific vaccination plus an IL-15 agonist, support their expansion through aOX40+a4-1BB therapy, while also reducing immune suppression with systemic aPD-L1 [146]. This treatment did have therapeutic benefits but would also be extremely challenging to translate to patients given the highly complex nature of the therapeutic regimen, most of which are not FDA-approved agents, and potential for adverse effects. Indeed, one of the major barriers to testing novel combinations is the reluctance of industry partners and/or the FDA to conduct phase I trials with two investigational agents, neither of which are FDA-approved. While the FDA has provided guidance for these types of studies, most companies are unwilling to combine their drug with another investigational agent in a phase I trial, even if the initial monotherapy phase I safety studies are complete. Thus, phase I trials testing the combination of OX40 agonists with IL-2 agonists, pegzilarginase, or Gal-3 inhibitors have not been conducted despite the strong biological rationale and robust preclinical activity of these specific combinations. Given the critical need for more effective therapies for patients with advanced disease, we hope that future studies will test these combinations, perhaps with financial support from a variety of sources including philanthropy, industry, and government agencies (NIH).

Article highlights.

• OX40-mediated costimulation can induce potent T cell activation and differentiation.

• Combination therapy with OX40 agonists and immune checkpoint blockade (ICB) or other novel agents stimulates robust anti-tumor immunity in a variety of preclinical models.

• However, clinical translation of OX40 agonists alone or in combination with ICB has not yielded similar efficacy in patients. This may be due to suboptimal dosing or timing of drug administration, differences in the biological activity of various OX40 agonists, and/or the presence of other suppressive mechanisms within the tumor microenvironment.

• Novel approaches are needed to drive more effective responses and fully realize the therapeutic potential of combination immunotherapy with OX40 agonists.