Hedgehog blockade in steroid‐refractory sclerotic chronic graft‐versus‐host disease

Abstract

Sclerotic chronic graft-versus-host disease (ScGVHD) is among the most difficult to treat disease forms affecting increasing number of allogeneic haematopoietic cell transplant survivors. A profibrotic activity of Hedgehog (Hh) signalling has been described in patients and preclinical ScGVHD models, with therapeutic benefits tied to nonimmune regulation of myofibroblast activity. Hh also regulates multiple aspects of adaptive and innate immunity relevant in ScGVHD (T cell development/function; germinal centre (GC) B cell homeostasis and M2 macrophage polarization). We tested the immune correlates of effective Hh inhibition with vismodegib. Patient samples were collected during an open-label study of vismodegib in ScGVHD conducted at the University of Utah Huntsman Cancer Institute. The study was approved by the Institutional Review Board, conformed to the ethical guidelines of the 1975 Helsinki Declaration, and was registered at ClinicalTrials.gov (NCT02337517). Patients with steroid-refractory cGVHD with at least one manifestation secondary to fibrosis (primarily skin and musculoskeletal system; Fig 1A), and up to two prior therapy lines were eligible for the study. No concurrent therapy was permitted aside for continuation of corticosteroid and calcineurin inhibitor or sirolimus use. Vismodegib was administered orally daily and six patients were treated before the study was terminated by the sponsor due to slow accrual. Response was evaluated monthly for the first three months, and then every two months. Within three months, five patients achieved a partial response as determined using the National Institutes of Health cGVHD consensus response criteria. Four patients significantly decreased glucocorticoid use (>70% dose reduction; Fig 1A). While sideeffects directly attributable to the medication were frequent (muscle cramps, gastrointestinal symptoms, and dysgeusia), only one patient experienced a grade 3 toxicity (weight loss) and four patients required a toxicity-related dose reduction to every other day or twice weekly dosing regimen. Hh inhibition was associated with a significant decrease in circulating B cell subsets implicated in cGVHD, including pre-GC (CD20CD27IgDCD38) and plasmablast (PB)like (CD20CD27IgDCD38) B cells (Fig 1B,C,D) as well as with diminished CD4 T cell expression of the transcription factor gene GATA3, which is essential for T helper (Th) 2 responses and humoral pathology (Fig 1D). However, we observed no impact of Hh inhibition on other T cell subset parameters (data not shown). This included no change in cytokine production, regulatory T cell prevalence, or expression of multiple co-inhibitory molecules (PD-1, OX40, CTLA4), nor were levels of transcription factors essential for Th1 and Th17 polarization (TBX21 and RORC, respectively) affected. Furthermore, analyses of plasma for cGVHD biomarker changes showed no impact on BAFF, osteopontin, TGFb, CXCL9, MMP3, or ST2 (Fig 1E). Finally, we analysed skin for the impact of Hh inhibition on macrophage presence and found a decrease in M2-like and total macrophages in biopsies obtained six months into the treatment in two patients. While the small sample size is the major limitation for assessing fully the observed responses in patients who completed at least three months of therapy, novel biological insights uncovered here provide a rationale for the further studies of therapeutic Hh blockade in ScGVHD. Our study suggests that beyond direct reduction of collagen production by myofibroblasts, Hh signalling in ScGVHD may also influence pathogenic B cell and macrophage responses to augment disease. While B celland macrophage-intrinsic Hh effects were defined previously, we postulate that Hh-driven B cell immunopathology may promote a feed-forward