Assessment of the Potential of Immature CD19+CD21- B-Lymphocytes to Predict Response to Various Systemic Therapies in Chronic Graft-Versus-Host Disease.

Abstract 2226

Poster Board II-203

Chronic graft-versus-host disease (cGVHD) is a major cause of mortality after allogeneic hematopoietic stem cell transplantation (HCT) and deleteriously affects the quality of life in surviving patients who otherwise have been cured of their underlying disease. Corticosteroids are the mainstay of therapy for cGVHD. Patients with steroid-refractory or steroid-intolerant cGVHD have received numerous immunosuppressive agents with a wide range of responses. Due to possible side effects of long-term immunosuppression including infectious complications and severe organ toxicities, biomarkers for prediction of response would be highly desirable. Recently, we reported that proportions of immature CD19⁺CD21^- B-lymphocytes predict the response to extracorporeal photopheresis (ECP) in patients with cGVHD (Blood 114: 744-746, 2009). Whether these findings were ECP-specific had to remain speculative since no other treatment cohorts had been analyzed and the mechanisms of action of ECP are still being elucidated. Therefore, we investigated B cell subpopulations in peripheral blood (PB) of 74 patients (median age 40 years, range 20-59 years) given different therapies including cyclosporine A (CSA n=24), tacrolimus (n=13), sirolimus (n=18) and ECP (n=19) for moderate (n=29) or severe (n=45) cGVHD according to the NIH Consensus. Duration of cGVHD before assessment of B cell subpopulations was a median of 14 (range, 0.5-120) months. Organ involvement at study entry consisted of skin in 45 patients (61%), oral mucosa in 50 (68%), eyes in 28 (38%), liver in 22 (30%) and lungs in 23 (31%), respectively . Seventy-three percent of patients had more than 2 organs affected by cGVHD. PB leukocytes were analyzed by multiparameter flow cytometry after staining for CD19, CD27, CD21 and surface Ig. Patients were assessed for cGVHD activity and response to therapy according to the NIH Consensus Development Project criteria prior to start of immunosuppressive therapy and every 3 months thereafter. B cell subpopulations of responders and non-responders after 6 months of therapy were compared. Overall, 45 patients (61%) responded to therapy including 20/24 (83%) given CSA, 6/13 (46%) on tacrolimus, 7/18 (39%) on sirolimus, and 12/19 (63%) given ECP, respectively. Among responders 16 (9 on CSA, 7 on ECP) achieved a complete resolution of cGVHD after 6 months of therapy. Prior to start of immunosuppressive therapy non-responders had significantly (p=0.03) higher proportions of immature CD19⁺CD21^- B-lymphocytes with a mean of 19.4% (range, 4.17-45) compared with a mean of 13.3% (range, 1.3-54) in responders to 6 months of therapy. Highest immature CD19⁺CD21^- B-lymphocyte numbers were observed in patients not responding to 6 months of therapy with sirolimus or ECP. Comparisons revealed significantly higher proportions of immature CD19⁺CD21^- B-lymphocytes in non-responders compared to responders to either sirolimus (mean of 23% vs 9.3%, p=0.02), tacrolimus (mean of 17.9% vs 8.6%,p=0.02), or ECP (mean of 22% vs 13.7%, p=0.02), respectively. In the CSA cohort no significant difference in immature CD19⁺CD21^- B-lymphocyte numbers between responders and non-responders was observed. In addition, no significant difference in memory CD27⁺ B-lymphocytes was observed prior to start of immunosuppressive therapy in any patient cohort between responders and non-responders. After 6 months of immunosuppressive therapy all responding patients had a significant (p=0.01) decrease of percentages of immature CD19⁺CD21^- B-lymphocytes from a mean of 13.3% (range, 1.3-54) prior to therapy to a mean of 8.2% (range, 1.1-30) 6 months later. In addition, in complete responders the immature CD19⁺CD21^- B-lymphocytes significantly (p=0.04) decreased from a mean of 14.06% (range, 1.3-54.6) to a mean of 7% (range, 1.1-36) after 6 months. In all non-responder percentages of immature CD19⁺CD21^- B-lymphocytes either increased or remained unchanged after 6 months of therapy. In conclusion, relative amounts of immature CD19⁺CD21^- B-lymphocytes assessed prior to start of cGVHD therapy may predict response to ECP, tacrolimus, and sirolimus. Increased proportions of CD21^- B-lymphocytes could be part of the autoimmune pathogenesis resulting in autoreactive B-cells in cGVHD. The correlation of numbers of B cell subsets and activity of cGVHD as observed in our study supports this hypothesis. Therefore, this novel cellular biomarker should be investigated in larger treatment cohorts of cGVHD patients.