ASP5094, a humanized monoclonal antibody against integrin alpha-9, did not show efficacy in patients with rheumatoid arthritis refractory to methotrexate: results from a phase 2a, randomized, double-blind, placebo-controlled trial

Rheumatoid arthritis (RA) is a chronic, systemic, inflammatory, autoimmune disease characterized by joint synovial inflammation. This condition is associated with irreversible cartilage destruction and osteolysis, resulting in pain, disability, interference in activities of daily living, and reduced quality of life [1‐3]. To control synovitis and subsequent irreversible joint damage, the recommended course of treatment consists of disease-modifying antirheumatic drugs (DMARDs), initiated as soon as possible upon diagnosis of RA, with dosages adjusted to treat to a target of clinical remission, or if that is not reached, to a target of low disease activity [4].

Methotrexate (MTX) is a first-line therapy for the treatment of RA, and when the target is not achieved, biologic agents or Janus kinase (JAK) inhibitors can be added [5, 6]. Even though biologic DMARDs have several mechanisms of action, including inhibiting the activity of cytokines such as tumor necrosis factor-alpha (TNF-α) and interleukin (IL)-6 [7, 8], T-cell co-stimulatory pathways, and CD20 on B cells [9, 10], only about one half to two thirds of patients achieve clinical remission [11, 12], and these treatments are associated with the risk of infectious events [13‐16]. In addition, molecular signatures characteristic of RA do not achieve normal levels following treatments targeting TNF or IL-6, even if the patient achieves clinical remission [17]. Thus, the need remains for safer and more effective treatments for RA, potentially through targeting novel mechanisms of action.

The disease process of RA involves activated fibroblast-like synoviocytes (FLS) in affected joints. These cells can lead to cartilage and bone degradation through production of matrix metalloproteinases (MMPs), upregulation of receptor activator of nuclear factor κB ligand (RANKL), production of IL-6, and recruitment and activation of proinflammatory immune cells [18‐22]. Recent findings suggest distinct types of FLS may be important for chronic inflammation and bone erosion in RA and may activate lymphocytes via antigen presentation, and when present in the joint lining, be a major cause of joint destruction [23‐25].

Integrin family proteins are composed of two subunits, α and β, and they bind to extracellular matrix (ECM) components and regulate a wide range of cellular responses such as migration, survival, and proliferation [26]. Integrin alpha-9 (α9) binds to ECM proteins such as tenascin-C [27], protease-cleaved osteopontin [28], and VCAM-1 [29]; contributes to cell adhesion and migration [30, 31]; and is highly expressed in synovial tissue cells, especially FLS, in patients with RA [32, 33]. This protein is also expressed on FLS of arthritic joints [34] and is overexpressed prior to onset of arthritis in mouse models [35]. Monoclonal antibodies against α9 alleviated mouse collagen-induced arthritis (CIA) [34, 36] and significantly reduced FLS-derived biomarkers [37]. Interestingly, treatment of CIA with the α9 antibody neither altered spleen cell numbers nor decreased plasma levels of anti-type II collagen antibody. Additionally, α9 antibody did not induce a mixed lymphocyte reaction or delayed type hypersensitivity reaction [37]. Thus, α9 appears to be a promising target for new therapeutics in the treatment of RA with minimal suppression of protective immunity.

ASP5094 is a humanized immunoglobulin G1 monoclonal antibody that targets human α9. Phase 1 data from healthy adult volunteers and patients with RA showed ASP5094 to be safe and well tolerated [38]. The present study investigated the efficacy, safety, pharmacokinetics (PK), and pharmacodynamics (PD) in patients with RA refractory to MTX treatment.

The development of biologic agents targeting cytokines, T cells, and B cells marked a revolutionary change in the treatment of RA. However, not all patients achieve treatment targets, and some experience adverse events such as severe infections, highlighting the need for new therapeutic targets. Accumulating evidence supports activation and/or increased expression of α9 as a key mechanism in the RA disease process [20, 32, 42] including α9 overexpression on FLS in rheumatic joints preceding the onset of arthritis [18, 19, 35], and the activation of α9 stimulating transformation of FLS into a pathologic state that includes hyperplastic and proinflammatory activity [33]. Moreover, monoclonal antibodies against α9 suppressed arthritis development and reduced FLS-derived biomarkers while sparing systemic immune activity in mice [37]. Thus, α9 was identified as a potential target for therapeutic development in RA.

In the current study, we investigated the clinical efficacy, safety, and PK and PD effects of the monoclonal α9 blocking antibody ASP5094 in patients with moderate to severe active RA. Our results revealed that 10 mg/kg ASP5094 was not more efficacious than placebo plus MTX at improving signs or symptoms of RA. The primary efficacy endpoint of ACR50 response rate at week 12 showed no clinically meaningful or statistically significant improvement with ASP5094 versus placebo, and no meaningful differences between treatment groups were observed in any secondary endpoints. Exploratory biomarkers were highly variable and there was no clear indication of PD modulation by ASP5094. We also did not observe any clear differences or trends between treatments within subgroups classified by demographics or potential stratification biomarkers, further indicating a lack of clinical efficacy by ASP5094.

The dose of 10 mg/kg tested in this study was chosen based on the findings in a previous phase 1 study [38, 43] in which binding of ASP5094 to α9 on neutrophils was assessed via percent receptor occupancy (%RO). In that study, 3 mg/kg ASP5094 led to RO above 80% for 85 days, and above 90% for 141 days with 10 mg/kg ASP5094. RO was not assessed in the current study, yet minimum serum ASP5094 concentrations ranged from 10.6 to 13.3 μg/mL at the end of each dosing period and was 11.6 μg/mL at week 12, which was more than 100-fold higher than the mean day 85 predose level (3 mg/kg dose group) in the previous study. Therefore, the target serum ASP5094 level was likely achieved in the current study, and the lack of clinical effectiveness or PD effects occurred while serum ASP5094 concentrations were within the expected range.

There are several potential explanations for the lack of efficacy we observed. First, the favorable preclinical data for ASP5094 that supported clinical investigation was obtained using a mouse model of acute arthritis in which administration of anti-mouse α9 antibody suppressed type II collagen-induced arthritis in both prophylactic and therapeutic regimens [37]. In contrast with the mouse, increased interstitial pressure by hyperplasia and extracellular matrix deposition in humans with RA might prevent drug diffusion from blood vessels to synovial tissues. Although our PK results suggest adequate serum levels of ASP5094 were achieved, these measurements were performed in peripheral blood, which does not necessarily confirm that ASP5094 occupied and suppressed α9 in the target tissue. A discrepancy between circulating ASP5094 and target tissue ASP5094 could explain the lack of PD effects; however, these data are not feasible to obtain in patients. Second, extremely high local concentrations of multiple α9 ligands such as OPN, tenascin-C, or VCAM-1 may compete with ASP5094 for binding to α9 integrin. In our previous study, ASP5094 neutralized binding of α9 to its ligand with an IC50 value of around 1 ng/mL, indicating that the antibody had good neutralizing potency in vitro ([33], Fig. 7A). However, in another study, a higher concentration of ASP5094 was required to suppress in vitro action of FLS derived from RA patients ([33], Fig. 7C). One possible explanation for this discrepancy is that huge amounts of ligands in the system produced by FLS affected the neutralizing activity of ASP5094. Additionally, there are data that suggest OPN levels in synovial fluid are markedly elevated compared with levels in the blood in patients with RA [43], and these may out-compete α9 for ASP5094 binding. Moreover, increased levels of locally expressed α9 might absorb a significant portion of circulating ASP5094, thereby depleting the amount available for binding in the synovium. Finally, the disease model in the aforementioned preclinical studies was of acute arthritis, which may not adequately represent the complex nature of chronic human RA. It is possible that the α9 neutralization that suppressed synovium activation and yielded positive results in the mouse model is not sufficient by itself to treat human RA where direct modulation of immune-pathway activation through additional targets may also be required. Thus, α9 alone may not be an effective target in the treatment of RA.

Although the incidence of overall and drug-related TEAEs was higher in the ASP5094 group, the overall incidence of severe and/or serious TEAEs was low. Overall, no notable safety signals were observed in this study, but further safety data will be needed in order to ascertain the overall safety profile of ASP5094 in patients with RA.

Single-cell analysis data from synovial tissue samples were published in 2019 by Zhang et al. [25]. Based on the information provided by these types of innovative technologies, it might be interesting to consider testing ASP5094 in patients specifically characterized by α9 overexpression.

In this study of patients with moderate to severe active RA refractory to MTX treatment, 10 mg/kg ASP5094 was not effective for the treatment of RA. Despite the lack of efficacy, no notable safety signals were observed in this study. Results of PK and PD assessments suggested that the expected serum levels of ASP5094 were achieved, yet expected molecular target effects were not observed. These findings may suggest target site exposure to ASP5094 in RA tissue was insufficient, or that targeting integrin α9 alone may not be adequate to deactivate FLS and thus improve RA.