Targeting a deubiquitinase blocks GVHD

In this issue of Blood, Cheng et al¹ have identified that targeting of a deubiquitinase (DUB), ovarian tumor deubiquitinase 1 (OTUD1), leads to inhibition of Notch2 signaling and abrogation of acute graft-versus-host disease (GVHD).

In the past few years, we have seen the approval of several drugs to treat GVHD in humans. Despite the advancements in our understanding of the disease and the treatment or prevention of acute GVHD, it remains a problem after allogeneic transplant, affecting up to one-half of recipients, and new approaches are needed.

One novel approach in treating GVHD is targeting the Notch pathway. The role of Notch signaling in T cells involved in acute and chronic GVHD has been elucidated. Notch signaling is not only critical to immune cell development but regulates mature T-cell activation, differentiation, and function during acute GVHD.² Notch inhibition has been reported to be most effective when targeting both Notch1 and Notch2 isoforms.³ Attempts to target Notch have included γ-secretase inhibitors (GSIs), anti-Notch antibodies, anti-Notch ligand antibodies, or an inhibitor of the transcription complex downstream of Notch.⁴ Broadly, targeting the Notch pathway with GSIs has proven difficult because of Notch’s role in many cellular processes outside of the T cell compartment, including the gut epithelium. Systemic pan-Notch inhibition with GSIs was poorly tolerated with gut toxicity made worse by transplant.³ Targeted antibodies against Notch ligands can also control GVHD when used transiently,⁵ but it is unclear what the long-term effects of pan-Notch inhibition are and whether this could affect T-cell development in the thymus and immune reconstitution after transplant.

Notch, like many cellular proteins, is subject to degradation through ubiquitination. Regulation of ubiquitination has been recognized as a means to fine-tune immune responses in T cells. The first drug approved targeting ubiquitination was bortezomib, and inhibition of the proteasome, and therefore destruction of ubiquitinated proteins, by bortezomib attenuated acute GVHD.⁶ Ubiquitination is also reversible through isopeptidases called DUBs. There are over 100 DUBs, but their role in allogeneic T-cell responses and GVHD is poorly understood, with only 1 DUB previously being implicated.⁷ No specific DUB has been identified for Notch2 prior to this work. OTUD1 is a DUB implicated in the inflammatory immune response, antiviral and antifungal immunity, and cancer. Loss-of-function mutations of OTUD1 have been identified in patients with autoimmune diseases.⁸ In addition, OTUD1 in hematopoietic cells inhibited colonic inflammation in a dextran sulfate sodium–induced colitis model.⁹ On the other hand, deficiency of OTUD1 has been reported to inhibit pro-inflammatory cytokine production following fungal infection, making deficient animals more susceptible to infection.¹⁰ No specific inhibitors to OTUD1 have been identified to date.

In the current study, Cheng et al explored, through a series of experiments, the role of OTUD1 and Notch signaling in alloreactive T cells. They show that OTUD1 is upregulated in T cells of human patients with acute GVHD. They provide evidence that OTUD1 enhances alloreactive Th1 and Th17 T cells by increasing the activity of intracellular Notch. Absence of OTUD1 in T cells led to decreased T cell activation, decreased Th1- and Th17-polarized T cells, and increased regulatory T cells, while impairing proliferation in activated CD4⁺ T cells. Donor OTUD1 deficiency inhibited T-cell activation, differentiation, and proinflammatory cytokine production. Through additional experiments, they demonstrated that Notch2 and Stat1 are downregulated in OTUD1 deficiency. Interestingly, Notch2, but not other isoforms of Notch (1, 3, or 4), is degraded in vitro in activated OTUD1-deficient CD4⁺ T cells. Activation of T cells induced the OTUD1-Notch2 interaction. OTUD1 acts to deubiquitinate Notch2 intracellular domain (NICD), leading to its protection from degradation, so blocking this DUB leads to Notch2 signaling degradation. Using a computer-aided drug screen, they identified a potential inhibitor of OTUD1, dapagliflozin, that down-regulated protein levels of NICD through increasing its ubiquitination. Dapagliflozin is a US Food and Drug Administration–approved oral hypoglycemic drug known to block sodium glucose transporter 2 (SGLT2) in the kidney. Cheng et al identify an unknown off-target effect of dapagliflozin in inhibiting OTUD1 and go on to show dapagliflozin inhibits the OTUD1/NICD axis to reduce GVHD in mice by promoting ubiquitination of Notch2-ICD and its degradation.

Through their work, Cheng et al demonstrate a novel pathway to exploit in selectively targeting Notch2 for the treatment of acute GVHD. The data suggest that targeting OTUD1 could be a new treatment for GVHD. Further studies are needed to understand the impact on graft-versus-tumor effects and to optimize OTUD1 targeting and inhibition given the modest effects of dapagliflozin. Notably, evaluation of an alternative SGLT2 inhibitor, empagliflozin, decreased levels of NICD but had no effect on ubiquitination of the NICD of Notch2, so this does not appear to be a class effect of SGLT2 drugs, and small molecule inhibitors could be refined to more specifically target OTUD1.

Several questions remain. Given the critical role Notch plays in immune cell development, will continual inhibition of this pathway lead to impaired T-cell reconstitution posttransplant? Or is the approach of targeting inducible expression of OTUD1 in activated T cells and only affecting Notch2 while preserving Notch1 a way to limit impaired T-cell reconstitution? Prior data suggests that both Notch1 and 2 are critical in GVHD development.³ So will selective targeting of Notch2 degradation through OTUD1 be adequate to control GVHD? Finally, given the identified role of OTUD1 in autoimmune disease and response to infections, will its targeting be tolerated? Despite these unanswered questions, this study suggests selectively targeting and exploiting DUBs may prove to be a ubiquitous way to control alloreactive responses.

Conflict-of-interest disclosure: M.A.S. declares no competing financial interests.