Abstract
Chronic GVHD (cGVHD) is characterized by a dysregulation of the adaptive immune system with an aberrant B cell homeostasis after allogeneic hematopoietic stem cell transplantation (allo-SCT). It is unclear, whether B cell dysregulation is solely a consequence of cGVHD or a predictive indicator of cGvHD development. Several studies have shown an increase in CD21low B cells in cGVHD patients. CD21low B cells are heterogeneous comprising three distinct B cell populations, CD10+CD38hi immature B cells, CD10- CD11c+ B cells expressing T-bet, and CD38hi CD27hi CD20- plasmablasts.
To elucidate the development of B cell dysregulation before the manifestation of cGVHD a prospective multicenter study exploring B cell phenotype and frequencies starting at d56 after allo-SCT, was initiated, a time point when first B cells are detectable in peripheral blood. We included 172 allo-SCT patients and excluded patients with EBV reactivation >10,000 copies/ml, post-transplant treatment with rituximab. After completion of follow-up, study patients were assigned retrospectively with respect to onset of GVHD based on Glucksberg criteria for acute GVHD (aGVHD) and NIH consensus criteria 2014 for cGVHD. For GVHD grouping only patients with a follow-up of at least 180 days after allo-SCT were evaluated. Patients without any form of GVHD or with aGVHD alone at day 180 or day 365 required at least 3 months of follow-up to exclude subsequent onset of cGVHD. Thus, three groups were evaluated: 1) no GVHD (n=17), 2) acute GVHD without subsequent cGVHD (n=32) and 3) cGVHD with a median onset of 180 days after allo-SCT (n=59, maximum severity: mild n = 17, moderate n= 30, severe n = 12).
Multiparameter flow cytometry of the three subpopulations of CD21low B cells revealed that already 56 days after allo-SCT the frequency of immature B cells was significantly elevated and remained increased until day 365 after allo-SCT in all patients, whereas CD11c+ B cells were significantly reduced at day 56 and raised subsequently until these cells were found expanded at day 180 and day 365. In contrast, plasmablasts showed a transient significant increased frequency 90 days and 180 days after allo-SCT. The frequency of plasmablasts was significantly elevated already 90 days after allo-SCT in patients developing subsequent cGVHD, as compared to patients without GVHD or aGVHD patients (median 5.9% vs 2.2% vs 2.2% of CD19+ cells; p=0.0016 and p=0.0304, respectively). The proportion of CD11c+ B cells in patients with cGVHD increased later in the course of time but was not significantly different between cGVHD patients at d+365 and patients without GVHD or aGVHD patients (median 4.4% vs 2.2% vs 4.0% of CD19+ cells, respectively). Frequencies of immature B cells did not correlate with the occurrence of aGVHD or cGvHD at any time point.
Multivariate analysis with clinical transplant parameters revealed that elevated frequencies of plasmablasts (>3% of CD19+) at 90 days after allo-SCT are an independent risk factor for subsequent cGVHD in a Cox proportional hazards model excluding HLA-matching as covariable (hazard ratio, 4.08; 95% CI, 1.4 to 11.86).
To obtain insight into the origin of plasmablast expansions we performed single-cell B cell receptor (BCR) sequencing of plasmablasts (n=5) from cGVHD patients. These analyses revealed prominent clonal expansions, a high frequency of somatic mutations, a low Shannon index of the BCR repertoire and overrepresentation of certain selected VHgenes. Most plasmablasts were found class switched to IgG and even more frequently to IgA. Recombinant expression of plasmablast-derived IgA antibodies revealed recognition of bacterial antigens in most cases of expanded clones. Regarding IgG antibodies analysis of sera from four cGvHD patients on protein arrays covering more than 7,000 human proteins established binding of a hitherto unknown spectrum of multiple autoantigens including RNA-binding proteins overlapping between patients.
In summary, our prospective study demonstrates an early oligoclonal expansion of hypermutated plasmablasts already 90d after allo-SCT in patients subsequently developing cGvHD, in a manner similar to systemic autoimmune diseases. Our findings strongly suggest an antigen-driven process induced by auto- or allo-antigens and indicate a pathogenic contribution of plasmablasts to cGvHD development.
Disclosures
Mackensen:Miltenyi Biomedicine: Honoraria; BMS/Celgene: Honoraria; Kite/Gilead: Honoraria; Novartis: Honoraria. Wolff:Incyte Corporation: Honoraria; Sanofi: Honoraria; Behring: Honoraria; Novartis: Honoraria, Research Funding.
Author notes
Asterisk with author names denotes non-ASH members.