The coronavirus disease 2019 (COVID-19) pandemic, caused by severe acute respiratory syndrome coronavirus 2019 (SARS-CoV-2) (1), has been responsible for more than 650,000,000 infections and more than 6.6 million deaths worldwide (2). Respiratory failure is a leading cause of death in patients with COVID-19 (3, 4). SARS-CoV-2 triggers a hyperinflammatory response through the innate Toll-receptor signaling pathway by alveolar type II pneumocytes that bind to SARS-CoV-2 (5). This is amplified by resident macrophage and other immune cell activation through Toll-receptor signaling contributing to coronavirus-related pulmonary failure (6–10). Bruton tyrosine kinase (BTK) is a member of the Toll-receptor signaling pathway that is activated in response to viral and bacterial pathogens, including coronaviruses (11). Patients with SARS-CoV-2 infection exhibit higher levels of activated BTK in peripheral blood monocytes and release of the inflammatory cytokine interleukin (IL)-6 versus uninfected controls (12). The potential for BTK inhibitors to block inflammatory cytokine release was observed in patients with chronic lymphocytic leukemia (CLL), Waldenström macroglobulinemia (WM), and chronic graft-versus-host disease, and involved many of the proinflammatory and chemoattractant cytokines found elevated in the plasma of patients with SARS-CoV-2 (13–17).

Clinically, as reported by us and other investigators, patients with COVID-19–related respiratory distress have shown improved oxygenation following treatment with a BTK inhibitor (12, 18–22). These findings suggest that BTK inhibitors may block the hyperinflammatory response related to SARS-CoV-2. Zanubrutinib is a highly potent and selective covalent BTK inhibitor approved for the treatment of mantle cell lymphoma, marginal zone lymphoma, CLL/small lymphocytic lymphoma, and WM (23). Zanubrutinib demonstrates an improved safety profile, including a lower incidence of atrial fibrillation, in comparison to the first-generation BTK inhibitor ibrutinib (24, 25). This is particularly relevant because atrial fibrillation has been observed in up to 20% to 40% of patients with severe COVID-19 infections (26, 27). As such, we conducted a two-cohort, phase II study to evaluate the activity and safety of zanubrutinib in patients hospitalized with COVID-19 infection requiring supplemental oxygen due to respiratory distress. One cohort was a prospective, randomized, multicenter, double-blind, placebo-controlled trial, and the other was a single-arm study (BGB-3111-219).

Between July 6, 2020, and February 1, 2021, a total of 97 patients were screened. Of these, 63 were randomly assigned to receive zanubrutinib (n=30) or matching placebo (n=33) in cohort 1; four of the patients randomized to placebo did not receive study drug due to investigator decision and were excluded from the efficacy analysis. Therefore, in the efficacy analysis set, there were 29 patients in the placebo arm, and a total of 59 patients for cohort 1. Four patients were enrolled in cohort 2 ( Figure 1 ). The baseline demographics for enrolled patients showed similar characteristics in both arms of cohort 1 and in cohort 2 as shown in Table 1 .

No patient in either cohort who received protocol therapy received a COVID-19 vaccination prior to study entry or during hospitalization. In cohort 1, among the 59 who received treatment, five (8.5%) patients received either an influenza or pneumococcal vaccine during hospitalization. Nearly half of these patients (44.4%) had three or more co-existing medical conditions at baseline that included hypertension (58.7%), diabetes mellitus (41.3%), asthma (14.3%), and chronic cardiac disease (12.7%); these were generally balanced between the treatment arms. All patients in cohort 1 were receiving supplemental oxygen at time of randomization, 47 (74.6%) with World Health Organization (WHO) grade 4 (oxygen by mask or nasal canula) and 16 (25.4%) with WHO grade 5 (non-invasive ventilation or high-flow oxygen) requirements. The median duration of treatment was 8.5 days in the zanubrutinib arm and 7.0 days in the placebo arm of cohort 1; and 13 days in cohort 2. Information on exposure and analysis sets can be found in the Supplementary Data .

We investigated the activity of zanubrutinib in a prospective study that included a randomized, double-blind cohort, to add to the evidence from preclinical and clinical studies that supports a potential role for BTK inhibitors to block SARS-CoV-2–related respiratory distress by suppressing pulmonary inflammatory responses (8, 10, 12, 22, 33). Our findings showed that zanubrutinib did not statistically differ from placebo for either respiratory failure-free survival rate or time to return to breathing room air at day 28.

A lack of meaningful clinical benefit was also reported in a prospective, placebo-controlled, randomized study of ibrutinib in 46 patients with COVID-19 who were hospitalized with hypoxia (34). Similar to our study, most patients received concomitant remdesivir and/or dexamethasone. The primary endpoint for the proportion of patients alive and without respiratory failure at day 28 was not met in that study. The addition of the BTK inhibitor acalabrutinib to best supportive care in patients with SARS-CoV-2 respiratory symptoms also failed to increase the proportion of patients who remained alive and free of respiratory failure in the two phase II CALAVI trials (NCT04380688 and NCT04346199); the formal publications of these two studies are awaited. Importantly, the TEAEs for zanubrutinib were similar to placebo in our study, thereby providing continued evidence for the safety profile of zanubrutinib. One patient in cohort 2 died of intracranial hemorrhage that was assessed as possibly related to the study drug, SARS-CoV-2 infection, and/or concomitant anticoagulation. Intracranial hemorrhage is a known complication of COVID-19, with an incidence of 8% among hospitalized patients (35, 36).

Most patients in the current study and the ibrutinib study (34) received steroids and antiviral therapy, which may have confounded the ability to observe differences between the active treatment and placebo arms with respect to the primary and secondary endpoints. Regarding biomarker findings, no differences in SARS-CoV-2 IgM or IgG antibody response were observed between patients in the zanubrutinib and placebo arms of cohort 1, whereas SARS-CoV-2 IgG antibody levels were higher in cohort 2 versus cohort 1 patients. Similar findings were reported by Roltgen et al. (37) who observed higher levels of SARS-CoV-2 IgG antibody response among those patients with more severe illness that required intensive care. A poor response to the COVID-19 vaccination has been observed in patients with CLL and WM receiving BTK inhibitors, although as many as half of patients with CLL and WM who were treatment-naive also lacked serological response to vaccination (38, 39). Many of these patients show a paucity of immune effector response. Placed into context with the findings from the current study, the contribution of BTK inhibitors alone is unlikely to be responsible for a blunted vaccine response in patients with CLL or WM. Treatment with a BTK inhibitor before SARS-CoV-2 vaccination, as well as qualitative differences for immune stimulation between vaccination and infection, may also have contributed to the blunted vaccination responses observed in patients with CLL or WM on BTK inhibitors.

The key circulating cytokines implicated in COVID-19 pathology were also noteworthy (9, 10, 40, 41). Patients who received zanubrutinib displayed lower levels for G-CSF, IL-10, and MCP-1 on days 2 and 7, as well as IL-4 and IL-13 on day 7. However, MCP-1 levels in zanubrutinib-treated patients in cohort 1 were significantly lower at baseline. Surprisingly, other cytokines elevated in COVID-19 patients such as C-X-C motif chemokine ligand 10, IL-6, MIP-1α, and TNF-α, which are downregulated in patients with certain B-cell malignancies and graft-versus-host disease treated with BTK inhibitors, showed no differences between zanubrutinib- and placebo-treated patients in cytokine studies (13–16). The lack of impact of zanubrutinib- versus placebo-treated patients on these cytokines may have reflected the concurrent use of steroids, which most patients in this study received. In addition, the cell of origin and signaling cascades for C-X-C motif chemokine ligand 10, IL-6, MIP-1α, and TNF-α production in COVID-19 patients may differ from those with B-cell malignancies and graft-versus-host disease; therefore, the effects of BTK inhibitors may differ in these patient groups.

Particularly relevant were single-cell RNA studies that depicted the impact of zanubrutinib treatment on cytokines, chemokines, and relevant receptor signaling pathways in circulating immune cell populations. These studies, though exploratory in nature, revealed novel insights into the activity of zanubrutinib that may also be relevant in other disease settings. CD14-expressing monocytes showed significant downregulation in multiple inflammatory cytokines including M-CSF, MIP-1α, MIP-3α, IL-1β, IL-8, and multiple members of the JAK-STAT pathway that contribute to the inflammatory response.

Microenvironmental support of malignant cells by monocytes through release of inflammatory proteins has been reported in CLL and WM, and their inhibition by zanubrutinib and other BTK inhibitors may contribute to their mechanism of action (42, 43). Downregulation of JAK-STAT pathway members was also observed among various CD4 and CD8 T-cell populations. Notable was the downregulation of CXCR4 in response to zanubrutinib across many immune cell populations, a known consequence of BTK inhibitors in B-cell malignancies (44). CXCR4 has been implicated in acute lung injury by triggering inflammatory response through MAPK and nuclear factor-kB (45).

An unexpected finding of our study was the upregulation of IFN-γ in γδ-T cells, denoting activation of this specialized population of immune effector cells. γδ-T cells mediate direct pathogen cytotoxic activity, but also have antitumor activity that is independent of human leukocyte antigen recognition (46). Although studies addressing the impact of BTK inhibitors on γδ-T cells are limited, an inhibitory effect of ibrutinib on γδ-T cell activity has been reported, which contrasts with the findings in this study with zanubrutinib (46, 47). Such differences may be related to the off-target inhibition of ITK by ibrutinib, which is not shared by zanubrutinib (48, 49).

Our findings may also be relevant to the potential use of zanubrutinib as an adjunct to cellular immune-based treatments. Cytokine release syndrome (CRS) is a common adverse effect associated with cellular immunotherapies including chimeric antigen receptor T-cell and bispecific antibody therapies (50). Many of the primary and secondary inflammatory cytokines and signaling pathways involved in CRS were markedly abrogated by zanubrutinib, including IL-1β, IL-6, IL-8, G-CSF, and MIP-1α (50–52). Tocilizumab is used to block IL-6 in the prophylaxis and treatment of CRS, demonstrating the importance of agents directed at cytokine suppression with cellular immunotherapies (52–55). Zanubrutinib may offer an advantage over IL-6 antibodies by more broadly blocking inflammatory pathways and signaling cascades relevant to CRS. In addition, myeloid derived suppressor cells, including monocytic myeloid derived suppressor cells, play a fundamental role in tumor propagated T-cell exhaustion through elaboration of inflammatory mediators that include IL-6 (56). Lastly, our unexpected finding of γδ-T cell activation following zanubrutinib treatment represents the first report of such activity, and may be relevant to the development of chimeric antigen receptor γδ-T cells and/or bispecific antibodies.

In summary, we conducted a prospective, randomized, multicenter, double-blind study to examine the impact of zanubrutinib versus placebo in hospitalized COVID-19 patients in respiratory distress. We observed a marked reduction in inflammatory cytokine signaling along with preserved serological response in those receiving zanubrutinib. Despite the immunological impact, zanubrutinib did not show improvement over placebo in clinical recovery from respiratory distress. Concurrent administration of steroids and antiviral therapy to most patients in this study may have contributed to these results. The data from this prospective, double-blind study may support the investigation of zanubrutinib in other settings where cytokine release and immune cell exhaustion are important.