Considerable efforts have focused on developing agonists of costimulatory receptors, including OX40.
Although antibodies targeting inhibitory immune checkpoints have been central to the success of cancer immunotherapy in the past decade, evidence suggests stimulatory checkpoints are equally important in the generation of an effective antitumor response. Considerable efforts have focused on developing agonists of costimulatory receptors, including OX40.1-3
The more nuanced role of these receptors and challenges inherent to developing agonist antibodies have proved significant stumbling blocks to clinical translation,4 but many pharmaceutical companies continue their pursuit in the hopes that new drug strategies will prove successful.5
Data from ongoing clinical trials presented at recent conferences have highlighted several of these novel drugs and brought OX40 back into focus.6-8
Full activation of naïve T cells requires not only interaction between the T-cell receptor (TCR) and major histocompatibility complex–bound peptides on the surface of an antigen-presenting cell (APC) but also a secondary signal provided by costimulatory molecules on the APC that bind to their receptors on the T cell (Figure9).9-12
Figure. OX40 Mechanisms of Action9
The prototypical costimulatory receptor, CD28, is expressed on naïve T cells and provides the initial costimulatory signal when bound to its ligand, but a range of other receptors and ligands are also upregulated in the hours and days after T-cell activation. These maintain T-cell activation and deepen the T-cell response by boosting proliferation and survival and optimizing the priming, differentiation, and effector functions of various T-cell subsets.4,10,11
Among the costimulatory receptors is OX40, 1 of the 29 known members of the tumor necrosis factor receptor (TNFR) superfamily; 6 of these family members, including OX40, included are classified as immune costimulators.4 Expression of OX40 is upregulated on both CD4- and CD8-positive T cells between 12 and 24 hours after TCR engagement and is then gradually downregulated 48 to 96 hours later.10,13,14
The only known ligand is OX40L, which is similarly transiently expressed on activated APCs. OX40L is also found on other cell types, including activated natural killer and B cells, as well as nonhematopoietic cells such as smooth muscle and endothelial cells. OX40 requires clustering for efficient activation. To facilitate this, OX40L contains conserved TNF homology domains through which it forms trimers that can bind to 3 OX40 receptor molecules simultaneously.10-12
Like other TNFR superfamily members, OX40 is a transmembrane protein. Through its cytoplasmic tail, OX40 binds to members of the TNFR-associated factor (TRAF) family of adaptor proteins, specifically TRAFs 2, 3, and 5. TRAFs facilitate the activation of downstream signaling cascades, the best characterized being the nuclear factor kappa B (NFκB) pathway, ultimately eliciting a range of different effects in T cells, dependent on the specific subset.10-12
Stepping on the Gas
Despite the success of immune checkpoint inhibitors (ICIs), many patients do not respond, and the development of resistance limits durable benefit for others. Costimulatory receptors have long been pursued as an alternative and potentially complementary target.2,3
Preclinical studies have demonstrated the ability of agonistic OX40 antibodies to promote T-cell activation and enhance T-cell– mediated antitumor immunity, resulting in tumor regression and survival benefits across a variety of cancer models.15-17
The first OX40 agonist to advance to clinical testing was 9B12, a mouse monoclonal antibody developed by the Providence Portland Medical Center and AgonOx, both in Portland, Oregon. The agent was subsequently renamed MEDI6469 when AgonOx in-licensed its OX40 program to MedImmune, an AstraZeneca company.4,9,18,19
In a first-in-human study of 9B12/ MEDI6469 (NCT01644968), there were no tumor responses, although 40% of patients experienced regression of at least 1 metastatic lesion. Treatment was well tolerated, and there was significant evidence of immunomodulatory activity.20
A humanized version of this antibody, MEDI0562 (tavolimab), was subsequently developed to prevent patients from forming human anti-mouse antibodies, which limited the number of doses in the above trial.20 In a phase 1 trial in patients with advanced solid tumors (NCT02318394), there were no dose-limiting toxicities (DLTs), and 1 patient exhibited a partial response (PR), which lasted 16 weeks.21 AstraZeneca has since discontinued development of MEDI6469 and MEDI0562,5 but the latter is still being studied by academic groups.
Two notable OX40 agonists, MOXR0916 and PF-04518600, were tested22-25; however, development of the former was discontinued completely in 2019,5 and the latter is no longer listed in Pfizer’s development pipeline.26 Overall, OX40 agonists as monotherapy have failed to live up to expectations, with only modest antitumor activity.
According to experts in the field, part of the problem lies in challenges inherent to developing agonist antibodies. Because antagonist antibodies exhibit a classic dose-response relationship, they are well suited to doseescalation studies. However, searching for the highest tolerable dose of receptor agonists has the potential to exhaust T cells and negate any antitumor efficacy.4,5
Important considerations for designing optimal agonist antibodies include binding affinity, epitope selection, valency, and receptor occupancy. Furthermore, because costimulatory receptor clustering is key to the signaling activation induced by native ligand binding, an agonist antibody must be able to reproduce this effect.4,5
Conventional bivalent antibodies, which can interact with only 2 receptors per antibody molecule, have low or no intrinsic agonist activity and are dependent on secondary cross-linking of antibody-receptor complexes caused by interaction between their Fc region and the Fcγ receptor (FcγR).27,28
Capitalizing on Combination Therapy
Despite setbacks, there is still considerable interest in pursuing OX40-targeted drugs (Table), but the focus has shifted to combination therapy. The most popular strategy is to combine OX40 agonists with ICIs based on preclinical demonstration of synergy, although administering the drugs in the proper sequence can be a crucial consideration with these combinations.29,30
Table. Select Studies of OC40-Targeted Drugs in Active Clinical Development
GlaxoSmithKline is developing a humanized OX40 antibody, GSK3174998, which is being evaluated in several ongoing clinical trials in combination with different drugs. Results from the recently completed phase 1 ENGAGE-1 study (NCT02528357) were reported at the 2020 American Association for Cancer Research (AACR) Virtual Annual Meeting I.
Patients with advanced solid tumors were treated with escalating doses of GSK3174998 (0.003-10 mg/kg) intravenously every 3 weeks as monotherapy (part 1; n = 45) or in combination with 200 mg pembrolizumab (part 2; n = 96). In part 1, 1 patient had a PR and 1 achieved stable disease (SD), both at the 0.3 mg/kg dose. In part 2, there were 2 complete responses, 7 PRs, and 9 patients with SD at doses from 0.01 to 3 mg/ kg. Clinical benefit was observed in both PD-1/PD-L1 inhibitor–pretreated and treatment-naïve patients.
There were 2 DLTs in part 2 (grade 3 nonmalignant pleural effusion at 0.03 mg/kg and grade 1 myocarditis at 10 mg/kg). The most common treatment-related adverse events (TRAEs), which were predominantly grades 1 and 2, included diarrhea and fatigue in part 1 and fatigue and nausea in part 2.31
Moving Forward With New Strategies
To address the need for receptor clustering, modifying the Fc region of the antibody to enhance FcγR binding can be an effective strategy but can also boost several of the effector functions of antibodies, including antibody-dependent cellular cytotoxicity. Although this can be beneficial, providing an additional antitumor effect, it could also lead to depletion of the T cells activated by the costimulatory agonist antibody and thus diminish its antitumor immune effects.27
An alternative strategy is to increase valency. Inhibrx is developing INBRX-106, a hexavalent OX40 monoclonal antibody that can bind 6 OX40 receptors per molecule of drug. It has been shown to outperform bivalent antibodies in preclinical testing, according to the company’s website.32
Bispecific antibodies, designed to engage 2 different targets, also are under study.
ATOR-1015 targets OX40 and the coinhibitory receptor, CTLA-4. A phase 1 first-in-human study (NCT03782467) in patients with advanced solid tumors is ongoing, and preliminary results were presented at the 2020 American Society of Clinical Oncology (ASCO) Virtual Scientific Program. ATOR-1015 was administered intravenously every 2 weeks at escalating doses from 0.043 to 600 mg. Patients were heavily pretreated, having received a median of 5 prior lines of therapy.
Data from 21 patients dosed to date demonstrated a best response of SD. There were no DLTs, and TRAEs were reported in 11 patients, most frequently infusion-related reaction and rash. No severe immune-related AEs were reported.8
F-star Therapeutics is also developing FS120, a bispecific antibody that targets a second costimulatory receptor, 4-1BB (CD137), in addition to OX40. After promising preclinical trials,9 F-star announced in January that the FDA had accepted its investigational new drug application. A phase 1 dose-escalation trial is planned in patients with advanced malignancies, with anticipated enrollment of 70 patients.33
Shattuck Labs is looking to achieve an effect similar to these bispecific antibodies via a different drug design. They used their Agonist Redirected Checkpoint platform to create SL-279252, a bifunctional fusion protein that consists of the extracellular domains of OX40L and PD-1 joined by a central Fc domain.34 A first-in-human clinical trial is ongoing (NCT03894618).
OX40L Gene Therapy
Nucleic acid–based drugs hold promise for gene therapy and could be used to induce tumor cells to express OX40L. These drugs present a challenge because naked genetic material is susceptible to degradation in the circulation, but lipid-based carriers offer a possible way around this problem.35
Moderna Therapeutics is developing 2 messenger RNA (mRNA)-based drugs, mRNA-2416 and mRNA-2752, both composed of lipid nanoparticles encapsulating a synthetic mRNA encoding OX40L. mRNA-2752 also contains mRNAs encoding proinflammatory cytokines IL-23 and IL-36γ.
At the ASCO and AACR meetings, results from ongoing first-in-human studies of both drugs were highlighted. mRNA-2416 is being evaluated alone and in combination with durvalumab (Imfinzi) in patients with advanced solid tumors or lymphomas (NCT03323398). Results from the monotherapy arm were presented; as of November 2019, 39 patients with solid tumors had been treated with mRNA-2416 administered intratumorally every 2 weeks for up to 12 doses at 4 dose levels (1, 2, 4, and 8 mg). The best overall response was SD in 14 patients, including 4 of 6 patients with ovarian cancer. SD lasted for at least 14 weeks in 6 patients. Tumor shrinkage was observed in 7 patients, of whom 2 had shrinkage in injected lesions only, 3 in uninjected lesions only, and 2 in both lesion types.
There were no DLTs, and the majority of TRAEs were grade 1 or 2, most commonly fatigue and injection-related reaction. The combination of mRNA-2416 and durvalumab is being evaluated in the trial’s ongoing part B, which includes a dose-expansion portion focusing on ovarian carcinoma.6
In a separate study (NCT03739931), intratumorally administered mRNA-2752 is also being assessed as monotherapy (arm A) and in combination with durvalumab (arm B) in patients with advanced solid malignancies or lymphoma. Patients receive escalating dose of mRNA-2752 (0.25-8 mg) and a fixed 1500-mg dose of durvalumab. As of April 2020, 17 patients in arm A and 12 patients in arm B were evaluable for safety, and 15 patients in arm A and 8 in arm B were evaluable for efficacy.
In arm A, 5 patients experienced SD as best response; however, in arm B, 1 patient with PD-L1–low squamous cell bladder cancer had a PR and 4 patients achieved SD. Tumor shrinkage was observed in both injected and uninjected lesions in both study arms. TRAEs, which were predominantly grade 1 or 2, most commonly included injection site erythema, injection site pain, and pyrexia. Five patients in arm A experienced grade 3 TRAEs, but no grade 3 TRAEs occurred in arm B.7