Urelumab alone or in combination with rituximab in patients with relapsed or refractory B‐cell lymphoma

Abstract Urelumab, a fully human, non‐ligand binding, CD137 agonist IgG4 monoclonal antibody, enhances T‐cell and natural killer‐cell antitumor activity in preclinical models, and may enhance cytotoxic activity of rituximab. Here we report results in patients with relapsed or refractory diffuse large B‐cell lymphoma (DLBCL), follicular lymphoma (FL), and other B‐cell lymphomas, in phase 1 studies evaluating urelumab alone (NCT01471210) or combined with rituximab (NCT01775631). Sixty patients received urelumab (0.3 mg/kg IV Q3W, 8 mg IV Q3W, or 8 mg IV Q6W); 46 received urelumab (0.1 mg/kg, 0.3 mg/kg, or 8 mg IV Q3W) plus rituximab 375 mg/m2 IV QW. The maximum tolerated dose (MTD) of urelumab was determined to be 0.1 mg/kg or 8 mg Q3W after a single event of potential drug‐induced liver injury occurred with urelumab 0.3 mg/kg. Treatment‐related AEs were reported in 52% (urelumab: grade 3/4, 15%) and 72% (urelumab + rituximab: grade 3/4, 28%); three led to discontinuation (grade 3 increased AST, grade 4 acute hepatitis [urelumab]; one death from sepsis syndrome [urelumab plus rituximab]). Objective response rates/disease control rates were 6%/19% (DLBCL, n = 31), 12%/35% (FL, n = 17), and 17%/42% (other B‐cell lymphomas, n = 12) with urelumab and 10%/24% (DLBCL, n = 29) and 35%/71% (FL, n = 17) with urelumab plus rituximab. Durable remissions in heavily pretreated patients were achieved; however, many were observed at doses exceeding the MTD. These data show that urelumab alone or in combination with rituximab demonstrated manageable safety in B‐cell lymphoma, but the combination did not enhance clinical activity relative to rituximab alone or other current standard of care.


| INTRODUCTION
Diffuse large B-cell lymphoma (DLBCL) and follicular lymphoma (FL) comprise approximately half of all cases of non-Hodgkin lymphoma (NHL). 1 DLBCL, the most common type of NHL (≈30% of cases), is a heterogeneous, aggressive lymphoma, 1,2 whereas FL is an indolent lymphoma accounting for approximately 22% of NHL cases. 1 The introduction of chemoimmunotherapy, including high-dose chemotherapy regimens in combination with the CD20-directed monoclonal antibody rituximab, has improved outcomes in patients with DLBCL and FL. 1, 3,4 The majority of patients with DLBCL can be cured with first-line therapy; however, approximately one-third of all patients are refractory to treatment or relapse afterward. 2,3 In contrast, most patients with FL experience disease progression (PD) after treatment, with recurrent relapses characterized by shorter remissions with each successive line of therapy. 5 Patients with FL who progress within 24 months of diagnosis after first-line chemoimmunotherapy have significantly shorter overall survival. 6 Patients with relapsed FL may also become refractory to chemoimmunotherapy or undergo histological transformation to a more aggressive NHL subtype. 5,7 Prognosis remains poor for patients with relapsed or refractory (R/R) DLBCL or FL 2,5,6 ; therefore, novel, more effective regimens are needed for these R/R populations. Therapeutic blockade of checkpoint pathway inhibitory receptors has demonstrated efficacy in multiple malignancies, including in patients with R/R classic Hodgkin lymphoma. 8,9 However, an unmet need remains in patients with R/R B-cell lymphomas, as variable clinical benefit has been observed with single-agent checkpoint pathway blockade. [10][11][12] Additional immunotherapy approaches targeting immunoregulatory receptors, including agonist antibodies against costimulatory molecules such as CD137 (4-1BB), may enhance antitumor immunity in patients with cancer. [13][14][15] Signaling via CD137, a costimulatory member of the tumor necrosis factor receptor (TNFR) superfamily, can lead to induction of cytokines, protection from activation-induced cell death, and upregulation of cytotoxic T-cell activity and may also reduce the infiltration of regulatory T cells into tumors. [14][15][16][17][18][19] In murine lymphoma models, agonist anti-CD137 treatment led to long-lasting antitumor activity mediated by natural killer and CD8 T cells. 19 Urelumab is a fully human, non-ligand binding, CD137 agonist immunoglobulin-γ 4 (IgG4) monoclonal antibody, that was evaluated as monotherapy or in combination with other immunotherapies or targeted agents in multiple phase 1/2 clinical trials. 16,[20][21][22] In an integrated safety analysis of three urelumab monotherapy studies (NCT00309023, NCT00612664, and NCT01471210), urelumab doses ≥1 mg/kg every 3 weeks (Q3W) were shown to be associated with more frequent transaminitis. 16 Therefore, in these and all subsequent studies of urelumab, a lower dose range (<1 mg/kg Q3W) was evaluated, and liver toxicities were closely monitored. Results from urelumab monotherapy and combination studies suggested limited clinical activity in patients with advanced solid tumors; however, preliminary activity was observed in patients with lymphoma. 20 Figure S1).
In both studies, patients were treated until PD, unacceptable toxicity, or withdrawal of consent. Treatment beyond PD, defined by the International Working Group (IWG) Response Criteria for NHL, 23

| Patients
In CA186-011, eligible patients with B-cell NHL had R/R disease after ≥1 prior line of standard therapy per IWG Response Criteria for NHL. 23 In CA186-017, eligible patients had CD20 + B-cell NHL with measurable disease per IWG Response Criteria for NHL 23 that was refractory to or had relapsed after ≥1 prior line of standard therapy.
Patients in the expansion phase must have received ≥1 prior multiagent chemotherapy regimen and must have had R/R disease after prior rituximab alone or in combination with chemotherapy. In both studies, patients had to be aged ≥18 years with an ECOG performance status of 0 or 1 and were required to provide pre-and ontreatment biopsies or fine-needle aspirates.
Adverse events (AEs) were assessed during treatment, for

| Statistical analyses
Descriptive statistics were used to characterize baseline demographics, safety, PK, immunogenicity, and pharmacodynamics.
Clopper-Pearson 95% two-sided confidence intervals were used to estimate objective response rate (ORR; defined as best response of complete remission or partial remission) and disease control rate (DCR; defined as best response of complete remission, partial remission, or stable disease). Kaplan-Meier methodology was used to esti-    (Table S2).

| Safety
The MTD was determined to be urelumab 0.1 mg/kg or 8 mg Q3W after a single event of potential drug-induced liver injury (pDILI) occurred in a patient treated with urelumab 0.3 mg/kg in study CA186-011. Following this event, enrollment into the 0.3 mg/kg dose level was halted, and all subsequent patients enrolled into the CA186-011 and CA186-017 studies were treated with lower doses.
In CA186-011, dosing was reduced to 0.1 mg/kg for all ongoing patients and to 8 mg Q3W or Q6W for all subsequent patients. In CA186-017, dosing was reduced to 0.1 mg/kg or 8 mg Q3W.

| Pharmacokinetics
Urelumab PK parameters increased proportionately with dose and were not altered substantially when urelumab was combined with rituximab (Table S3). Following a 1-hour IV infusion, maximum con-

| Immunogenicity
Sixteen percent of patients treated with urelumab 8 mg Q3W monotherapy, 16% of patients treated with urelumab 8 mg Q6W monotherapy, and 30% of patients treated with urelumab 0.3 mg/kg monotherapy were antidrug antibody (ADA) positive after treatment (Table S4). Overall, ADA positivity did not appear to affect urelumab safety. No patients were ADA positive after treatment with urelumab plus rituximab.

| Efficacy
With urelumab monotherapy, ORR and DCR were respectively 6% and 19% in patients with DLBCL (n = 31), 12% and 35% in patients with FL (n = 17), and 17% and 42% in patients with other types of Bcell NHL (n = 12;  however, correlative analyses were limited by a small sample size.

| DISCUSSION
The rationale for evaluation of urelumab in hematologic malignancies was based on preclinical analyses of human primary lymphomas, including DLBCL and FL. 19 Bulk tumor samples from patients with lymphoma showed overexpression of CD137 mRNA compared with nonlymphoma samples and were infiltrated with CD137 + T cells, while tumor B cells were uniformly negative for CD137. 19 The population of CD137 + tumor-infiltrating T cells may be a source of tumorreactive cells that could be stimulated by CD137 agonism. 19 Moreover, in murine lymphoma models, anti-CD137 treatment led to durable antitumor activity as monotherapy. 19 In the CA186-011 and CA186-017 phase 1 studies, the safety and antitumor activity of urelumab alone or in combination with rituximab were evaluated in patients with solid tumors (CA186-011 only) and B-cell NHL. These studies were designed to assess a lower urelumab dose range, with a focus on liver toxicities, due to liver injury and drug-related deaths observed in prior studies that evaluated higher doses of urelumab. 16 While the clinical mechanism is unclear, previously published preclinical data suggest that anti-CD137-induced liver toxicity may be partially due to infiltration of S100A4 + macrophages into the liver, following activation of CD8 + T cells and secretion of IFN-γ. 24,25 The TRAEs leading to discontinuation included grade 3 increased AST in a patient treated with urelumab 8 mg Q3W, grade 4 acute hepatitis in a patient treated with urelumab 0.3 mg/kg, and one death from sepsis syndrome in a patient treated with urelumab 0.3 mg/kg plus rituximab. The sepsis syndrome was the only treatment-related death reported in either study.
Overall, the MTD was established as urelumab 0.1 mg/kg or 8 mg Q3W based on a pDILI reported in one patient treated with urelumab 0.3 mg/kg (CA186-011). Despite this single event, liver toxicity was less frequent and severe in these studies than previously observed with higher urelumab doses. 16 In the CA186-017 study, urelumab in combination with rituximab did not enhance clinical activity relative to rituximab alone 26,27  In the CA186-011 and CA186-017 studies, antitumor activity, including several durable remissions, was observed with urelumab as monotherapy or in combination with rituximab. However, many of these responses were observed at doses that exceeded the MTD, suggesting that the limited clinical activity observed may be due to suboptimal CD137 agonism. Future studies evaluating nextgeneration therapies that target CD137 with unique approaches to safely increase the dose/exposure of CD137 delivery are underway. 14,29 Specific approaches include but are not limited to bispecific antibodies engaging 4-1BB and a tumor antigen/stromal component, intratumoral delivery, local nanoparticle-anchored antibodies, and/or unique combination therapies. 14,29-32 These strategies may lead to more efficacious CD137 therapy for patients with R/R B-cell lymphoma, a population with a high unmet need.

ACKNOWLEDGMENTS
This study was supported by Bristol-Myers Squibb, Princeton, NJ. We thank the patients and their families and the investigators and