Emerging Immune Checkpoint Research Focuses on CD137

During the past decade, a growing number of PD-1/PD-L1 inhibitors gained FDA approval to treat a wide range of cancer types. Their stimulatory counterparts also emerged as sought-after anticancer targets but have proved much more challenging to manipulate therapeutically.^1,2

Based on the same premise of reigniting an antitumor immune response, development of agonist antibodies targeting stimulatory T-cell receptors has lagged, hampered by severe toxicity and limited efficacy.²

Nevertheless, efforts to develop drugs that target these receptors continue. One facet of this research focuses on drugs targeting CD137 (4-1BB), an inducible costimulatory receptor expressed on the surface of activated T cells but absent from resting T cells.²

In early clinical trials, the first generation of CD137 agonists exemplified the difficulty of targeting this receptor. Urelumab (BMS-663513) caused liver inflammation at effective doses, and utomilumab (PF-05082566) elicited few responses.² Both drugs remain under study as part of combination regimens.

Meanwhile, as investigators seek to understand toxicity issues with this therapeutic approach, a new generation of agents is emerging (Table). Investigators are pursuing novel designs to target CD137 agonism conditionally to avoid systemic toxicity.³ Bispecific antibodies that simultaneously target tumor-associated antigens (TAAs) in addition to CD137 represent a prominent alternative strategy. These include agents such as cinrebafusp alfa (PRS-343), a firstin-class CD137 x HER2 bispecific fusion protein, and drugs targeting CD137 plus PD-L1. Another approach involves trispecific antibodies.

These novel agents are in the early stages of clinical development, but promising initial data demonstrate favorable safety profiles while hinting at potential antitumor efficacy.³

Challenges of Targeting CD137

As the main effectors of the antitumor immune response, T cells have been a central focus of cancer immunotherapy.¹ CD137, a member of the tumor necrosis factor receptor superfamily, generates stimulatory signals that help regulate T-cell activation, thus promoting T-cell proliferation, survival, and effector functions.²

Investigators designing chimeric antigen receptor (CAR) T-cell therapies have successfully developed constructs that incorporate the intracellular domain of CD137. CARs with a 4-1BB domain, as it is frequently described in CAR therapies, differentiate into central memory T cells and are associated with fewer toxicities than agents that use CD28 costimulatory domains.⁴

The story has been different for attempts to leverage CD137 as a drug target in its own right. Agonist antibodies targeting CD137 have been pursued for nearly 20 years, culminating in the development of the first-generation drugs urelumab and utomilumab.^2,3

Urelumab is a fully human IgG4 agonist antibody5 for which clinical development began in 2005 with 2 trials evaluating its safety and efficacy as monotherapy (NCT00309023, NCT00612664) in patients with advanced solid tumors and melanoma, respectively. All trials were halted in 2008 following 2 hepatotoxicity-related deaths.⁶

Subsequent analyses demonstrated that liver toxicity was dose dependent—severe transaminitis was associated with doses of 1 mg/kg or greater⁶—and was likely a result of on-target effects of urelumab on liver Kupffer cells.⁷

Clinical development resumed in 2012, with trials evaluating urelumab monotherapy at reduced doses below 1 mg/kg and as part of various combination regimens.^6,8 Unfortunately, the reduced doses abolished singleagent activity,⁹ with the exception of limited responses among patients with non-Hodgkin lymphoma (NHL). Objective response rates (ORRs) of 6%, 12%, and 17% were observed in patients with diffuse large B-cell lymphoma, follicular lymphoma (FL), and other B-cell lymphomas, respectively (NCT01471210). The ORR was 35% among patients with FL treated with a combination of urelumab and rituximab (Rituxan) and 10% for participants with diffuse large B-cell lymphoma (NCT01775631).⁵

Utomilumab is a fully human IgG2-based antibody with weaker CD137 agonist activity than urelumab.¹⁰ As a result, it is well tolerated up to a dose of 10 mg/kg; however, the trade-off is limited activity both as a single agent and in combination with other anticancer agents. In trials treating patients with solid tumors, utomilumab monotherapy led to an ORR of 3.8% (NCT01307267)¹¹ and utomilumab in combination with the CCR4 antagonist mogamulizumab-kpkc (Poteligeo) produced an ORR of 4.2% (NCT02444793).¹⁰ The combination of utomilumab and OX40 agonist PF-8600 in 30 patients with melanoma and non–small cell lung cancer (NSCLC) elicited only 1 partial response (NCT02315066).¹² The combination of utomilumab and pembrolizumab (Keytruda) was more promising, with an ORR of 26%, including 2 complete responses (CRs), in patients with advanced solid tumors (NCT02179918).¹³

Table. Ongoing Clinical Development of CD137-Targeted Drugsa

Clinical trials of both drugs are ongoing, with a focus on combination therapies. In results from another cohort of the phase 1 clinical trial (NCT01307267), the combination of utomilumab and rituximab in patients with relapsed/refractory NHL demonstrated an ORR of 21.2%, including 4 CRs, with a median duration of response of 20.3 months. Most utomilumab treatment-related adverse events (TRAEs) were grade 1 or 2 in severity; most common was fatigue. There was 1 grade 3 increase in alanine aminotransferase at a dose of 2.4 mg/kg.¹⁴

Meanwhile, urelumab in combina-tion with nivolumab (Opdivo) and GVAX (an allogeneic cancer vaccine comprising pancreatic tumor cells that express granulocyte-macrophage colony-stimulating factor) is being evaluated as neoadjuvant or adjuvant therapy in patients with resectable pancreatic ductal adenocarcinoma in 1 arm of a phase 1/2 trial (NCT02451982).

Preliminary results presented at the 35th Annual Meeting of the Society for Immunotherapy of Cancer in 2020 demonstrated a pathologic complete response rate of 30% among 10 evaluable patients, with 9 of them remaining disease free after a median follow-up of 1 year. The combination was well tolerated, with nausea the most common AE attributed to urelumab.¹⁵

Conditional Agonists

A swath of next-generation CD137 agonists that recently entered clinical development is designed to overcome the limitations of the first-generation drugs.³ Intrinsic agonistic strength, immunoglobulin isotype, and interactions with Fcγ receptors (FcγRs) have been found to be critical to the relative safety and efficacy of CD137 antibodies. Thus, investigators have been tinkering with these properties to optimize CD137 agonist design.^3,16

Notably, conditional CD137 agonists have been designed to be dependent on FcγR cross-linking so that they will stimulate CD137 signaling only in the context of ongoing immune cell activation. The theory is that high concentrations of endogenous IgG circulating in blood and in highly vascularized tissues such as the liver will block the agonist function of these antibodies by competing for binding to FcγRs. The agonist activity of the antibodies will be targeted to the tumor microenvironment, where both CD137 and FcγRs are strongly expressed.^16-21

AGEN2373 is a fully human, non–ligand-blocking, IgG1-based antibody that binds to cysteine-rich domain (CRD) IV on CD137 as opposed to urelumab, which binds to CRD I, and utomilumab, to CRD II and III.^19,22 Results from an ongoing first-in-human study in patients with advanced solid tumors (NCT04121676) were recently presented at the 2021 American Society of Clinical Oncology Annual Meeting (ASCO 2021).

Patients (N = 19) were treated with AGEN2373 monotherapy at escalating doses from 0.03 to 2 mg/kg every 4 weeks. The best response, prolonged disease stabilization, was seen in 5 patients. There were no dose-limiting toxicities and no grade 2 or higher drug-related increases in transaminase or bilirubin levels. The most common TRAEs were fatigue and nausea.²²

ADG106 is a fully human, ligand-blocking, IgG4-based antibody, which was engineered on Adagene’s NEObody platform.¹⁸ Results from ongoing phase 1 clinical trials (single-center, NCT03802955; multicenter, NCT03707093) in patients with advanced solid tumors and/or relapsed/ refractory NHL were presented at ASCO 2021. As of November 2020, the disease control rate (DCR) across these trials was 56%, and ADG106 displayed a favorable safety profile.²³ According to an earlier analysis of the single-center trial, the most common treatment-emergent AEs were arthralgia, decreased appetite, diarrhea, hypothyroidism, and vomiting, which were predominantly grade 1 or 2.²⁴

Investigators also identified a potential biomarker that correlated with response to ADG106; 3 of the 4 biomarker-positive patients experienced greater than 30% tumor shrinkage at doses of 3 mg/kg or 5 mg/kg.²³ A phase 2 trial of ADG106 being planned will select for patients with this biomarker, the identity of which has not been publicly revealed.²⁵

Alligator Bioscience is developing ATOR-1017, a fully human, IgG4-based ligand-blocking antibody that binds to CRD II, the natural ligand-binding domain of CD137.^21,26 In an ongoing first-in-human trial of ATOR-1017 monotherapy in patients with advanced solid tumors (NCT04144842), 12 patients had been enrolled as of January 2021 (at doses of 0.38, 1.5, 5, 15, 40, and 100 mg). Preliminary results demonstrated a best response of stable disease (SD). TRAEs were predominantly grade 1 or 2 and included chest pain, headache, pyrexia, upper abdominal pain, mouth ulceration, nausea, leukopenia, neutropenia, cytokine release syndrome, arthralgia, neck pain, and rash.²⁷

LVGN6051, another conditional agonist of CD137 that Lyvgen Biopharma is developing, is unique in that its Fc domain has been engineered to selectively bind to FcγRIIB, the sole inhibitory FcγR.¹⁶ FcγR engagement by a therapeutic antibody triggers dual effects: clustering of the antibody’s receptor target and activation of the antibody’s cytotoxic effector functions. Although the latter outcome can be advantageous in certain circumstances, with agonist antibodies these effector functions can eliminate the T cells that express the costimulatory receptor, inadvertently reducing the antibody’s antitumor efficacy. However, FcγRIIB mediates CD137 clustering without inducing cytotoxicity and thus avoids this off-target effect.¹⁶

As of January 2021, 16 patients had been enrolled in an ongoing first-in-human clinical trial of LVGN6051 as monotherapy or in combination with pembrolizumab for the treatment of advanced cancers (NCT04130542). Seven patients treated with monotherapy achieved SD, with the longest duration greater than 8 months. More than 10% tumor reduction was observed in patients with melanoma and neuroendocrine tumors. One patient with metastatic head and neck cancer treated with combination therapy had an immune partial response (PR) that lasted more than 6 months.²⁰

Multispecific Antibodies

A prominent alternative strategy for leveraging CD137 is the development of bi- and trispecific antibodies, which engage a TAA in addition to CD137. This is another way to potentially achieve conditional stimulation of CD137 at the tumor site, as the drug should be activated only where CD137 and the TAA are simultaneously expressed.^2,3

An array of multispecific antibodies targeting numerous TAAs is in the early stages of clinical development. Pieris Pharmaceuticals is developing cinrebafusp alfa, a bispecific fusion protein composed of a HER2 antibody and a CD137-binding Anticalin. Anticalins are recombinant human proteins derived from a group of transport proteins called lipocalins. Small, stable, and modular, Anticalins are being tested as an alternative to antibodies.^28,29

In a phase 1 trial in patients with HER2-positive solid tumors (NCT03330561), cinrebafusp alfa was administered at doses ranging from 0.0005 to 8 mg/kg every 2 or 3 weeks. Among 19 response-evaluable patients treated at the active dose range of 2.5 mg/kg and above, the DCR was 58%, with confirmed PRs in 11% of patients. The most common TRAEs were fatigue, chills, and diarrhea.³⁰

Genmab and BioNTech are jointly developing GEN1042 and GEN1046, CD137 agonist antibodies that simultaneously target CD40 and PD-L1, respectively.³¹ According to preliminary data from the ongoing firstin-human trial of GEN1046 (NCT03917381) in patients with advanced solid tumors (N = 61 as of June 2020), the DCR was 65.6%, with SD or PRs observed in patients with triple-negative breast cancer, ovarian cancer, and NSCLC. The most common TRAEs included transaminase elevation, hypothyroidism, and fatigue.³²

Roche is testing its bispecific antibodylike fusion proteins RO7122290 and RO7227166 in phase 1 clinical trials. Both agents target CD137; in addition, RO7122290 binds to fibroblast activation protein alpha, which is highly expressed on cancer-associated fibroblasts in many tumors, whereas RO7227166’s secondary target is CD19.

RO7122290 is being evaluated in a European-based 3-part first-in-human clinical trial as a single agent and in combination with atezolizumab (Tecentriq) in patients with advanced solid tumors (Eudra CT number: 2017-003961-83). Results from the dose-escalation portion of the trial (parts A and B) were presented at the European Society for Medical Oncology Virtual Congress 2020. In part A, 62 patients were treated with monotherapy doses ranging from 5 to 2000 mg; in part B, 39 patients received RO7122290 at doses ranging from 45 to 2000 mg in combination with atezolizumab 1200 mg.

The ORRs were 3.6% in part A and 18.4% in part B. AEs were generally mild to moderate in both parts. Grade 3 and higher AEs included asthenia, increased aspartate aminotransferase levels, and pneumonia in part A and pneumonia, pneumonitis, neutropenia, and lymphocytopenia in part B. The maximum- tolerated dose was not reached.³³