Moving the needle on chronic GVHD of the lung Open Access

In this issue of Blood Advances, Sharifi et al¹ share the results of their phase 1 STOP-BOS trial evaluating the antifibrotic oral small molecule pirfenidone for bronchiolitis obliterans syndrome (BOS), the principal form of pulmonary chronic graft-versus-host disease (cGVHD) following hematopoietic cell transplantation (HCT). They demonstrate in patients with established BOS that pirfenidone is well tolerated and slows disease progression or even improves lung function.

BOS is hallmarked by peribronchiolar inflammation and intraluminal fibroproliferation causing small airway constriction, leading to airflow obstruction, progressive dyspnea, and cough. BOS is defined by post-HCT obstructive decline: forced expiratory volume in 1 second (FEV1) <75% or <5th percentile predicted, FEV1/forced vital capacity <70%, and air trapping per imaging or residual volume >120% without concomitant respiratory infection.² Although relatively rare, affecting 2% to 10% of total HCT survivors, or up to 15% of patients with cGVHD, BOS is cGVHD’s most dismal form, with 40% 5-year survival.³ 

The goal of BOS treatment is to halt progression and, optimally, reverse manifestations of disease. First-line therapy often includes combinations of inhaled or systemic corticosteroids, azithromycin, montelukast, and inhaled long-acting beta agonists. For contemporary immune pathway–targeting drugs used as second- or further-line therapy in refractory cGVHD, pulmonary function improves in ∼25% to 35% of patients, making the lung one of the least treatment-responsive cGVHD organs. Potentially reflecting “fixed” tissue fibrosis, severe BOS (FEV1 ≤39%) is often unresponsive.⁴ These observations are nuanced by few BOS-specific studies, heterogeneous response definitions, and variable alignment with National Institutes of Health (NIH) BOS diagnostic consensus criteria.² Effective BOS therapy (and dedicated BOS clinical trials) is thus majorly needed.

Enter pirfenidone. Pirfenidone was US Food and Drug Administration–approved in 2014 for idiopathic pulmonary fibrosis (IPF), which shares aspects of BOS pathophysiology. Pirfenidone acts on multiple relevant cell types to leverage production of and responsiveness to transforming growth factor (TGF)-β, the prototypical cytokine of fibrosis, and mitigate oxidative stress and inflammasome activity.⁵ Sharifi et al show in their single-center trial of 30 patients with BOS (90% fulfilling NIH consensus criteria) that 63% tolerated 56 weeks of full-dose pirfenidone, with only 6% completely discontinuing due to adverse events (primarily transaminitis), comparable to rates in patients with IPF. Moreover, 41.3% of evaluable patients experienced a clinical improvement, defined as improved FEV1 slope compared to FEV1 trend up to 24 months pretrial. Notably, FEV1 overtly rose in 5 responders rather than merely remaining steady throughout treatment (shown in the Sharifi et al. supplemental Material). Considering these responses in a disease with historically poor outcomes and few proven therapies, responders were offered extended pirfenidone treatment for up to 56 months. This cohort maintained responses, with most remaining within 10% of their pretrial FEV1s, if not otherwise improving. These findings mirror mouse models wherein pirfenidone ameliorated established bronchiolitis obliterans, resulting in improved lung function along with decreased macrophage infiltration, collagen production, and TGF-β expression.⁶ 

These observations raise provocative possibilities. Considering that these patients overwhelmingly had longstanding BOS, it is probable that many pretrial pulmonary function declines occurred despite standard-of-care frontline therapy, indicative of treatment-refractory disease. However, once on pirfenidone, at least 80% of all patients (responders and nonresponders) did not exceed 10% further FEV1 decline (a threshold which some prior trials have defined as treatment failure), and they reported improved respiratory quality on validated inventories. Might this translate into longer survival or other meaningful clinical and quality-of-life outcomes even when numerical FEV1 trends do not overtly improve? These questions will require large-scale, randomized trials.

A novel aspect of this study addresses whether BOS therapy responses can be predicted, making a point that not all BOS is created equal. Strikingly, baseline spirometry values (which conventionally define BOS severity), preceding BOS duration or timing of onset, extrapulmonary cGVHD profiles, and systemic immunosuppression status did not differ between pirfenidone responders and nonresponders. However, response correlated with less-impaired baseline quantitative computed tomography (qCT), which measures proportions of normal lung parenchyma, degree of air trapping, and volume expansion and heterogeneity. Worse qCT indices likely indicate high degree of fixed fibrotic lung tissue (“functionally advanced” disease), which may be less capable of remodeling. Whether baseline qCT could inform treatment response or the need for therapy intensification remains to be fully elucidated. Along these lines, might incorporating qCT into the early workup for BOS provide an opportunity to optimize frontline therapy to prevent further, rapid fibrosis?

This cohort largely had cGVHD, affecting multiple organs, representing a high cGVHD burden. Although STOP-BOS was not structured to formally evaluate extrapulmonary cGVHD, there were clinically meaningful patient-reported outcome (PRO) improvements in the eye and mouth (and additional quantitative skin improvement) per the Lee Symptom Scale, plus overall improved physical and emotional functioning. Intriguingly, these extrapulmonary improvements did not necessarily represent prototypically fibrotic cGVHD organs nor did they strictly align with pirfenidone respiratory response status. May this indicate utility of pirfenidone in other cGVHD tissues, a hitherto unexplored therapeutic consideration?

Mechanistically, pirfenidone could inhibit multiple potential actors in BOS pathophysiology. Mouse models demonstrate pathogenic roles of colony stimulating factor-1–dependent alternatively activated tissue macrophages, type-17 T cells, and auto-/alloantibodies, pathways potentially involved in extrapulmonary organ cGVHD^6-8 and post-lung allograft BOS.⁹ TGF-β derived from macrophages and injured lung epithelium then targets fibroblasts, which activate into myofibroblasts producing excessive and disorganized collagen and extracellular matrix, causing pathologic airway remodeling. Furthermore, TGF-β–stimulated alveolar macrophages expand local tissue resident memory T cells, spurring a vicious cycle of further epithelial injury.¹⁰ Pirfenidone is believed to reduce fibroblast-intrinsic TGF-β responsiveness, macrophage alternative activation and TGF-β responsiveness, and CD4⁺ and CD8⁺ T-cell proliferation,⁵ which are not wholly fibrosis-directed phenomena. This may speak to potential pirfenidone utility in nonfibrotic cGVHD pathophysiology, as hinted by STOP-BOS oral and ocular PROs.

In sum, developing effective strategies to treat BOS remains critical. In this trial, prolonged treatment with pirfenidone was well tolerated and associated with airflow obstruction stabilization, and in some subjects, improvement. Moreover, patient-reported quality-of-life and other organs improved regardless of lung response. Potentially incorporating into future trials, correlates including qCT may inform therapy responsiveness and/or be proxies for degree of lung impairment. Randomized trials will be crucial to (1) fully evaluate the effectiveness of agents like pirfenidone in mitigating or reversing bronchiolar inflammation characteristic of BOS, (2) determine whether frontline pirfenidone may have an advantage over existing standard-of-care options, and (3) identify the benefit that it may have in extrapulmonary cGVHD.

Conflict-of-interest disclosure: The authors declare no competing financial interests.