Monitoring of B Cell Subpopulations in Patients with Chronic Graft-Versus-Host Disease May Predict Response to Extracorporeal Photopheresis

Chronic graft-versus-host disease (cGVHD) is a major complication of allogeneic hematopoietic stem cell transplantation (HCT) requiring prolong immunosuppressive therapy and increasing non-relapse mortality. Immune mechanisms underlying cGVHD have remained elusive. Recently, we reported a severe disturbance in B cell homeostasis seen in a significant expansion of immature/transitional B lymphocytes (CD21−) and a significant decrease of non-class switched (CD19+/IgD+/CD27+) and class-switched memory B cells (CD19+/IgD−/CD27+) in patients with active chronic GVHD. In long-lasting cGVHD irreversible tissue damage cannot be distinguished from active cGVHD with certainty since no biomarkers for monitoring of cGVHD activity as well as for assessment of therapeutic response are available. We investigated serially every 3 to 6 months B cell subpopulations in the peripheral blood (PB) of 47 patients (median age 40 years, range 18–62 years) given extracorporeal photopheresis (ECP) as first line (n=13) or salvage therapy (n=34). Twenty-nine patients (64%) had more than 2 organs affected by cGVHD and in 21 patients (45%) severity of organ involvement was grade 3 according to the NIH Consensus. Duration of cGVHD before ECP was a median of 2.5 years. ECP was performed initially every two weeks on 2 consecutive days and monthly thereafter until improvement. The median duration of ECP was 30 (range, 8–50) cycles during a median of 13 (range, 3–36) months and ECP is still ongoing in 14 patients. A total of 257 samples with a median of 6 analyses per patient (range, 3 to 10) was assessed. PB leukocytes were analyzed by multiparameter flow cytometry after staining for CD19, CD27, CD21 and surface Ig. Patients were scored for cGVHD activity and response evaluation according to the NIH Consensus Development Project criteria at every sampling event. Thirty-five patients (75%) responded to ECP, including 22 with complete resolution, whereas 12 (25%) were nonresponders. Retrospective correlation of clinical response to ECP with B cell subpopulation numbers revealed that patients not responding to ECP had a significantly higher number of immature/transitional B cells (CD19+/CD21−) (mean 16.6%, range 0.8–57%) in PB samples taken prior to start of ECP compared to responders (mean 13%, range 1–54%, p 0.0001). The numbers of both non-class-switched and class-switched memory B cells were significantly higher in ECP-responders, mean 5.1% (range 1–21%) compared to a mean of 4.2% (range 0.3–14%) in nonresponders (p 0.041) The CD21−/CD27+ ratio was significantly higher in nonresponders with a mean of 8.6 (range 0.5–50) compared to a mean of 6.6 (range 0.3–89) in the responding group(p 0.0005). Three months after start of ECP patients responding to therapy had a decrease in their CD21−/CD27+ ratio (mean 3, range, 0.3–18, p=0.05) whereas in nonresponders B cell subpopulations were unchanged. After 6 months a significant decrease in immature/transitional B cell numbers (CD19+/CD21−) compared to baseline (p=0.03) values was observed in ECP responders whereas ECP nonresponders presented a further increase in immature/transitional B cell (CD19+/CD21−) numbers. One year after start of ECP 22 patients with durable complete responses had a further decrease of the CD21−/CD27+ ratio (mean 1.6, range 0.2–5.4) comparable to patients with resolved cGVHD as previously reported (Greinix et al, BBMT14:208–219, 2008). In contrast, nonresponders to ECP with persistent active cGVHD had a significantly higher CD21−/CD27+ ratio (mean 3.3, range, 0.1–11, p=0.08) in PB samples. In conclusion, determining the degree of disturbance of B cell homeostasis as seen in elevated numbers of immature/transitional B cells and decreased memory B cell subsets during active cGVHD could be used for predicting therapeutic response to ECP. Moreover, 3 to 6 months after start of ECP distribution of B cell subpopulations differed significantly between ECP responders and nonresponders. Thus, B cell subsets serially assessed could serve as possible biomarkers of response to ECP in cGVHD. Our preliminary findings should be confirmed in a larger prospective patient cohort receiving ECP.