Relevance of Plasma Matrix Metalloproteinase-9 for Bronchiolitis Obliterans Syndrome after Allogeneic Hematopoietic Cell Transplantation.

Bronchiolitis obliterans syndrome (BOS) is a highly morbid form of chronic graft-versus-host disease (GVHD) after allogeneic hematopoietic cell transplantation (HCT). Several plasma proteins have been identified as biomarkers for BOS after lung transplantation. The relevance of these biomarkers in BOS patients after allogeneic HCT has not been examined. We hypothesized that biomarkers associated with BOS after lung transplantation are also associated with BOS after allogeneic HCT. We tested plasma samples from 33 adult HCT patients who participated in a phase II multicenter study of fluticasone, azithromycin, and montelukast (FAM) treatment for new-onset BOS (NCT01307462), and matched control samples of HCT patients who had non-BOS chronic GVHD (n = 31) and those who never experienced chronic GVHD (n = 29) (NCT00637689 and NCT01902576). Candidate biomarkers included matrix metalloproteinase-9 (MMP-9), MMP-3, and chitinase-3-like-1 glycoprotein (YKL-40). MMP-9 concentrations were higher in the patients with BOS compared with those with non-BOS chronic GVHD (P = .04) or no chronic GVHD (P < .001). MMP-3 concentrations were higher in patients with BOS (P < .001) or non-BOS chronic GVHD (P < .001) compared with those with no chronic GVHD. YKL-40 concentrations did not differ statistically among the 3 groups. MMP-9 concentrations before starting FAM therapy were higher in patients who experienced treatment failure within 6 months compared with those with treatment success (P = .006), whereas MMP-3 or YKL-40 concentrations did not differ statistically between these 2 groups. Patients with an MMP-9 concentration ≥200,000 pg/mL before starting FAM therapy had worse overall survival compared with those with lower MMP-9 concentrations. Our data suggest that plasma MMP-9 concentration could serve as a relevant biomarker at diagnosis of BOS after allogeneic HCT for prognostication of survival and for prediction of treatment response. Further validation is needed to confirm our findings.