Dipeptidyl peptidase-1 inhibition with brensocatib reduces the activity of all major neutrophil serine proteases in patients with bronchiectasis: results from the WILLOW trial

Non-cystic fibrosis bronchiectasis (NCFBE) is a chronic, progressive respiratory disease defined by permanent dilatation of the bronchi [1]. Pathogenesis in NCFBE is complex, involving not only inflammatory components and lung tissue damage, but also infection and airway dysfunction [2‐4]. These four features have been termed the vicious cycle [1] or more recently, the vicious vortex [3]. Neutrophils accumulate in the airways resulting in excess neutrophil serine protease (NSP) activity that can lead to destruction of lung tissue [2‐4]. The three main NSPs include neutrophil elastase (NE), proteinase 3 (PR3), and cathepsin G (CatG). Elevations in NE activity in sputum in NCFBE are associated with increased exacerbations and declines in lung function [5]. Excess PR3 and CatG activity may also play a role in NCFBE disease progression [6, 7]. Thus, reduction of sputum NSP activity, as opposed to inhibition of NE exclusively, represents a novel therapeutic approach to attenuate this vicious vortex in NCFBE [2, 3].

Dipeptidyl peptidase-1 (DPP-1, EC 3.4.14.1), also known as cathepsin C, is an enzyme responsible for activating NE, PR3, and CatG zymogens in promyelocytes during neutrophil differentiation in the bone marrow [8]. Collectively these proteolytic enzymes are stored within neutrophil granules and released during degranulation and neutrophil extracellular trap release (NETosis) [9]. Human loss-of-function mutations in DPP-1 in Papillon-Lefèvre syndrome subjects markedly reduce NSP protein and proteolytic activities (active NE/PR3 < 5%; active CatG < 1%) in mature blood neutrophils [10], demonstrating that DPP-1 is a key regulator of NSP activation in humans [8]. Importantly, patients with Papillon-Lefèvre syndrome are not typically at increased risk for infections [8].

Brensocatib is an investigational, oral, selective, reversible inhibitor of DPP-1 [11, 12]. Evaluation of once-daily brensocatib (10, 25 or 40 mg) in healthy subjects for 21 or 28 days resulted in a dose-dependent reduction in NE activity in neutrophils in the blood, achieving maximal reductions of 30, 49 and 59%, respectively, after 28 days [12]. The one-month time frame to achieve maximal and steady-state effect on blood NE levels is consistent with a 15-day maturation time for human neutrophils in the bone marrow. This phase 1 study did not report effects of brensocatib on PR3 and CatG but established the potential for DPP-1 inhibitors like brensocatib to attenuate NSP activity in neutrophil-driven inflammatory diseases.

In a phase 2b, randomized, double-blind, placebo-controlled trial in 256 patients with NCFBE (WILLOW), 24 weeks of treatment with 10 or 25 mg brensocatib was associated with improvements in clinical outcomes including prolongation in the time to first exacerbation and reduction in the frequency of exacerbations [11]. Of the 256 patients, 87 were assigned to receive placebo, 82 to receive 10 mg of brensocatib, and 87 to receive 25 mg of brensocatib. Additionally, brensocatib reduced NE activity in sputum reaching a nadir at the 4-week timepoint that was maintained throughout the 24-week treatment duration, with a return to baseline levels at the 28-week timepoint, four weeks after termination of brensocatib treatment [13]. The achievement of sputum NE activity below the limit of quantitation (BLQ) post baseline was also associated with a statistically significant reduction in pulmonary exacerbations [14].

In this manuscript, we conducted an exploratory analysis to evaluate the effect of brensocatib on PR3 and CatG activity in sputum, and NE activity in white blood cell (WBC) extracts, which were not previously reported in the phase 2b trial in NCFBE [13]. We also compared the extent of reduction of NSP activity in both WBC extracts and sputum compartments and identified correlations in all three NSP activities at baseline and on treatment. These data may inform on the effect of a reversible DPP-1 inhibitor like brensocatib on the relevant pharmacodynamic (PD) biomarkers and their contribution to morbidity in NCFBE.

Bronchiectasis is an inflammatory disorder and neutrophils dominate the airway inflammatory profile in the majority of bronchiectasis patients. NSPs such as NE, PR3 and CatG are released from neutrophil granules when neutrophils are activated and during the formation of neutrophil extracellular traps [1, 9]. The resulting protease damage to the airways has been linked to impaired host defense through cleavage of host receptors and antimicrobial peptides leading to increased susceptibility to infection and disease progression through extracellular matrix degradation. Thus, elevated NE activity has been shown to predict risk of exacerbation, lung function decline over time and mortality in bronchiectasis [5]. The association between CatG and PR3 sputum activity and outcomes in bronchiectasis has not been thoroughly investigated. However, in CF patients, PR3 and CatG are found at high concentrations in the sputum and bronchial alveolar lavage fluid so they should not be discounted as contributors to lung disease in NCFBE [6]. This study shows strong correlations among NE, CatG and PR3 activities in sputum suggesting the potential for elevated CatG and PR3 activity to be associated with increased exacerbation risk.

There are currently no approved treatments for bronchiectasis, and there is a need for non-antibiotic treatments that can reduce lung inflammation. To that end, inhibition of DPP-1 represents a novel therapeutic strategy to reduce NSP activity and lung inflammation [13]. While brensocatib’s direct mechanism of action is to reduce active NSPs in circulating neutrophils, the pharmacologic effect in bronchiectasis relies on a reduction of NSP activity in the lung and these reductions were indeed observed in sputum samples. NSP activity in sputum is ultimately a consequence of circulating blood neutrophils migrating to the lung and releasing their NSP cargo. Thus, it is not unexpected that the reduction in circulating neutrophil NSP activity directly translates to reduced NSP activity in the lung.

Interestingly, a greater reduction of NE activity was observed in sputum than in WBCs (Table 1), suggesting that it may not be necessary to completely inhibit DPP-1 in circulating neutrophils to have a meaningful impact on lung markers of inflammation. Biomarker analysis planned for the ongoing Phase 3 brensocatib ASPEN trial (ClinicalTrials.gov Identifier: NCT04594369) and Phase 2 trial in CF (NCT05090904) may inform on whether CatG and PR3 activities are also reduced in sputum to a greater extent than in WBCs.

There are several plausible explanations for a greater effect in the lung than in the circulation. Reduced NSP activity may directly or indirectly reduce neutrophil recruitment to the lung, or the inflammatory stimulus for NSP release including potentially impacting bacterial infection through restoration of antimicrobial peptides. This theoretically generates a “virtuous cycle”, the counterpoint to the vicious cycle, whereby reductions in lung inflammation improve other aspects of airway disease, leading to larger and more sustained reductions in airway inflammation. These reductions in NSP activity in the lung may allow for a return to homeostatic balance of NSPs with their inhibitors: alpha-1 antitrypsin (A1AT), elafin and secretory leukocyte peptidase inhibitor (SLPI).

The positing of a virtuous cycle in response to brensocatib suggests that once brensocatib treatment is removed, NSP activities in the lung may take longer to return to baseline levels than for those in the circulation (i.e., WBC extracts). Indeed, four weeks after termination of brensocatib treatment, NE activity in the WBCs returned to baseline levels, while NE, PR3 and CatG activities in sputum did not.

Sputum NE, PR3 and CatG activities were strongly correlated at baseline and positive correlations were observed on study treatment at all timepoints, suggesting that brensocatib’s effect on all three NSP activities is interrelated. This does not appear to be a surprising result given that the mechanism of action of DPP-1 is to activate all three NSPs. However, while the NSP correlations remained positive on treatment, the strength of those correlations decreased somewhat for those on brensocatib, and more so for those on the higher 25 mg dose of brensocatib. The explanation may be related to the differential reduction in the NSP activities to brensocatib treatment, with CatG being reduced to the greatest extent, followed closely by NE, with a lower reduction in PR3 activity in sputum. Thus, it may not be unexpected that the very strong correlation between NE and CatG at baseline showed the least change in response to brensocatib treatment, since both biomarkers were more likely to have been reduced comparably. In contrast, the positive correlations between PR3 and both NE and CatG declined on brensocatib treatment (but not on placebo), which may be explained by the lower reduction in PR3 activity relative to both NE and CatG. The residual reductions in the NSP correlations that cannot be explained by differential impact of brensocatib on the individual NSP biomarkers may be related to the increased appearance of NSP activity values below the quantification limit. Notably, four weeks after the end of brensocatib treatment, the sputum NSP correlations all recovered close to their initial values.

DPP-1 inhibition with brensocatib had the greatest effect in reducing CatG activity in sputum and the least effect on PR3 activity. There may be multiple factors accounting for the differential response in NSP activity to DPP-1 inhibition. In the absence of a DPP-1 inhibitor, nearly all of the NSPs zymogens are converted to active forms during neutrophil maturation [16]. Studies with DPP-1 knockout mice (dpp1-/-) suggest that there is a “DPP-1-like” protease that activates a portion of NE and PR3 zymogens (app. 10%) but < 1% of CatG zymogen is converted to its active form [17]. This DPP-1-like protease termed as NSPs-Alternative Activating Protease (NSPs-AAP) is also present in humans and maturates < 5% of NE/PR3 zymogens and < 1% of CatG zymogen [10]. Thus, if NSPs-AAP is not inhibited by brensocatib, then CatG activity would be reduced to a greater extent than NE or PR3, as was observed. The greater reduction in NE activity in sputum compared to PR3 may be a consequence of differences in binding affinity of their zymogens to either DPP-1 or the NSPs-AAP, as well as the presence of protease inhibitors in the sputum like A1AT that has preferential binding affinity to NE [18]. The presence of bound A1AT to the active site of NE (or PR3 or CatG) may lead to it not being quantified in our NSP activity assay.

Brensocatib reduced all three NSP biomarkers in sputum illustrating the unique mechanism of action of DPP-1 inhibitors that simultaneously reduce several key inflammatory mediators linked to exacerbation. In contrast, specific inhibitors to NE ([19] for review, [20, 21]), while showing promising preclinical data, have failed to achieve their primary endpoints in cystic fibrosis, NCFBE or chronic obstructive pulmonary disease, even though NE is associated with worsening disease [5]. One possible explanation for the failure of NE inhibitors alone is the challenge with achieving adequate local inhibition in the immediate vicinity of the neutrophils in the lung. Thus, the NE inhibitor may be unable to rapidly quench the extracellular release of high concentrations of NE from a single azurophil granule before they effect tissue damage prior to diffusion to a more dilute concentration that can be effectively inhibited [22]. This tissue injury is a consequence of the quantum proteolysis effect [22]. The failure of NE inhibitors alone, and the promising data for brensocatib in the Phase 2 WILLOW trial [13], suggests that combined inhibition of NE, PR3 and CatG, as well as inhibition prior to extracellular release by the neutrophils, may be required for observation of efficacy in NCFBE.

The thesis that PR3 and CatG are also important regulators of the local inflammatory response is supported by the literature [6, 7, 9, 23] and thus their inhibition may be required to attenuate disease progression in the lung for those with NCFBE. PR3 and CatG, like NE, also process chemokines and stimulate the production of cytokines leading to the activation and mobilization of immune cells to the site of tissue damage [23]. PR3 is not restricted to the azurophilic granules, like NE, and is present in easily mobilizable secretory vesicles [23, 24]. CatG activates metalloproteases and cleaves extracellular matrix proteins, which, in the lung, can contribute to tissue damage. Additionally, in CF bronchial alveolar lavage fluid, surfactant protein A, a peptide that facilities microbial clearance by macrophages, is inhibited to a greater extent by CatG than either NE or PR3, and this can lead to reductions in macrophage phagocytic activity and increased bacterial survival [6]. Furthermore, CatG also inhibits the ability of macrophages to clear apoptotic cells from CF airways, which can lead to increased neutrophil necrosis and an uncontrolled release of proteases into the lung [7]. Finally, in preclinical studies in NE and CatG knock-out mice with Pseudomonas endobronchitis infections, examination of bacterial clearance and airway inflammation led to the conclusion that CatG interferes with bacterial clearance and escalates the pulmonary inflammatory response [25]. Thus, all three NSPs may directly contribute to the dysregulation of the local inflammatory environment in the lung in NCFBE [9]. The ongoing Phase 3 ASPEN trial, when complete, may inform on the association of PR3 and CatG with increased exacerbation risk.

Brensocatib demonstrated dose dependent reductions in blood NE and sputum NSP activity in the WILLOW study. These data suggest a broad anti-inflammatory effect of brensocatib underlying its clinical efficacy observed in NCFBE patients.