Natalizumab for GVHD: too little or too late?

Graft-versus-host disease (GVHD) after allogeneic hematopoietic cell transplantation is a major cause of nonrelapse morbidity and mortality. Although ruxolitinib is now approved for the treatment of steroid-refractory GVHD, to date, no agent added to corticosteroids has been shown to improve outcomes compared with corticosteroids alone. In this issue of Blood Advances, Al Malki et al presented the results of a multicenter phase 2 study that tested whether the addition of natalizumab, a humanized antibody against the α4 subunit of α4β7 integrin, would improve the outcomes of new-onset acute GVHD.¹ The primary end point was a complete response after 28 days, defined as the clinical resolution of GVHD in the target organs.

There were several notable aspects of the study design. First, the trial aimed to focus on patients at a high risk of treatment failure, as predicted by a previously validated risk stratification system based on serum levels of ST2 and REG3α (Ann Arbor scores of 2 or 3).² A second aspect was that the outcomes were compared with those of a synthetic control cohort treated only with corticosteroids, extracted from a patient database assembled by the Mount Sinai Acute GVHD International Consortium, the same group that tested the Ann Arbor biomarker scoring system.³ 

In total, 75 evaluable patients were enrolled. The study participants could have received any conditioning regimen, stem cell source, or GVHD prophylaxis, and biopsy specimen–proven GVHD was not required. The subjects could have received up to 3 days of corticosteroids while awaiting biomarker screening, but no other systemic therapy was permitted. Generally, study patients and controls were well-matched with respect to comorbidity index, donor type, allograft source, transplantation indication, conditioning regimen, and GVHD prophylaxis, with the exception that study patients received less antithymocyte globulin and had a longer median time to GVHD onset.

Unfortunately, the study failed to meet its primary end point and was closed after the interim futility analysis. There were also no significant differences between natalizumab recipients and controls when comparing the secondary outcomes of overall response rate, nonrelapse mortality, overall survival, and biomarker scores on day 28, and this was true across all subgroups analyzed.

It is worth considering the scientific rationale for interference with T-cell trafficking as a method for treating GVHD. The α4β7 integrin expressed by T cells contributes to their migration into the gastrointestinal tract through its interaction with mucosal vascular addressin cell adhesion molecule 1. Studies from mouse models using both antibody-mediated blockade and donor T cells lacking key chemokine receptors or integrins have, in part, supported the idea that restricting T-cell trafficking could prevent GVHD.^4-6 In a nonrandomized trial, the addition of the CCR5 antagonist maraviroc, which is also predicted to restrict T-cell entry into the gut, to standard immunosuppression reduced GVHD relative to what might have been expected based on historical data,⁷ although it failed in a larger multicenter phase 2 trial.⁸ Vendolizumab, an antibody targeting α4β7, was found to be safe when added to standard GVHD prophylaxis⁹ and is now being tested in a multicenter phase 3 trial (NCT03657160). However, these data do not address the question of whether an anti-α4 antibody can effectively treat GVHD of the bowel after it is established.

Anti-α4 antibodies have had some efficacy in treating inflammatory bowel disease. Although natalizumab failed to meet the primary study end point in treating Crohn's disease, subgroup analyses suggested that patients with the most inflammation might benefit.¹⁰ This benefit, albeit small, was demonstrated in a subsequent randomized trial.¹¹ Natalizumab was also shown to be modestly effective as maintenance therapy in patients with Crohn's disease who responded to induction therapy.^10,12 Vedolizumab has been approved for the treatment of inflammatory bowel disease based on randomized trials, although it only has a relatively modest benefit.^13,14 There are retrospective case series of patients with steroid-refractory bowel GVHD treated with vedolizumab; however, a phase 2a trial testing vedolizumab in this setting was closed after futility analysis.^15,16 Taken together, targeting α4-expressing integrins has, at best, only had a small benefit in treating established tissue-inflammatory bowel diseases.

With this context, why did natalizumab fail in this study? A trivial explanation would be that it was inadequately dosed. However, the authors used a regimen based on pharmacokinetics in subjects with inflammatory bowel disease, and although no pharmacokinetic studies were performed, there is no obvious reason why the dose used would have been inadequate. An alternative explanation for the clinical results could be that GVHD is not reliant on the entry of new T cells from the blood into the affected tissues; instead, once GVHD is established, it is largely maintained locally by intratissue T cells. Such a model would predict that treating GVHD with agents that impair T-cell entry into tissues would not be especially effective. Our recent work in mouse models of GVHD supports these ideas.¹⁷ Through analysis of the tissue distribution of alloreactive T-cell clones and parabiosis of mice with GVHD, we found that after GVHD was established, it was largely maintained locally by tissue-resident T cells. Our data also suggest that the local GVHD response is dependent on a Tcf7-expressing subset of intratissue T cells, which transcriptionally and phenotypically resemble so called "T exhausted progenitors," shown to be critical for maintaining responses against chronic viral infections and tumors and is a key subset of cells that respond to immune checkpoint inhibition.¹⁸ The fact that vedolizumab and natalizumab have been so disappointing in treating inflammatory bowel diseases suggests that these diseases may also be mostly locally maintained, largely independent of new blood-derived T cells.

Despite the negative results, the authors should be commended for testing a rational hypothesis through the conduct of a trial with a robust design, leveraging biomarker-based risk stratification. This approach enables the conduct of clinical trials for GVHD treatment with smaller cohorts while maintaining adequate power. Such cost-effective designs could accelerate clinical GVHD research regardless of whether future trials are positive or negative.

Conflict-of-interest disclosure: W.D.S. holds stock options and is a paid consultant for Bluesphere Bio and holds stock options in Orca Bio. The remaining authors declare no competing financial interests.