Proportions of immature CD19⁺CD21⁻ B lymphocytes predict the response to extracorporeal photopheresis in patients with chronic graft-versus-host disease

To the editor:

B lymphocytes are increasingly assigned an important role in the pathogenesis of auto-and alloimmune diseases including chronic graft-versus-host disease (cGVHD)^1-4  where abnormalities of B-cell homeostasis in patients with active cGVHD have been observed.^1,4  Extracorporeal photopheresis (ECP) achieves high response rates in steroid-refractory cGVHD.⁵  At present, no reliable biomarkers are available for prediction of response to ECP and monitoring of patient outcome. We thus investigated the distribution of immature CD19⁺CD21⁻ B cells and memory CD19⁺CD27⁺ B cells in peripheral blood (PB)¹  before ECP (n = 19), 6 (n = 28), 12 (n = 34), and 21 months (n = 33) after start of ECP in 49 patients (28 male, 21 female) with moderate (n = 25) or severe (n = 24) cGVHD⁶  and correlated results with clinical response.⁷  Informed consent was obtained in accordance with the Declaration of Helsinki. Eighteen patients were included in a prior report.¹ 

Complete (CR) and partial response (PR) to ECP was observed in 25 of 34 patients (74%) after 12 months. ECP nonresponders at 6 months had a significantly higher (P = .02) percentage of immature CD19⁺CD21⁻ B lymphocytes (mean 22%, range 4%-52%) in PB before start of ECP compared with CR (mean 8%; range 1.3%-29%) and PR (mean 16%; range 1.6%-55%) patients (Figure 1A). Percentages of memory CD19⁺CD27⁺ B cells before first ECP were not significantly different (P = .12) between the groups. The CD21⁻/CD27⁺ ratio was significantly higher (P = .03) in ECP nonresponders (mean 17.4; range 0.5-50) compared with CR patients (mean 1.6; range 0.3-5).

Percentages of immature CD19⁺CD21⁻ B lymphocytes were significantly (P < .001) lower in CR patients compared with ECP nonresponders 6 (mean 5% vs 25%; range 1%-15% vs 14%-43%), 12 (mean 6% vs 24%, range 1.2%-16% vs 1.5%-58%), and 21 (mean 6% vs 26%; range 1%-21% vs 2%-60%) months after start of ECP (Figure 1B). In addition, in ECP responders compared with ECP nonresponders the CD21⁻/CD27⁺ ratio was significantly lower 6 months (mean 2.2 vs 9.6; range 0.1-6.5 vs 1.2-20; P = .001), 12 (mean 2.7 vs 6; range 0.2-15 vs 0.6-16; P = .006), and 21 months (mean 2 vs 17; range 0.2-9 vs 0.6-76; P = .001) after start of ECP.

To our knowledge, relative amounts of immature CD19⁺CD21⁻ B lymphocytes represent the first reported cellular biomarker able to predict response to ECP. Whether our findings are ECP-specific has to remain speculative because no other treatment cohorts were analyzed and the mechanisms of action of ECP are still subject to further research.⁵  CD19⁺CD21⁻ B lymphocytes could also serve as a novel biomarker for measuring disease activity of cGVHD quantitatively and objectively, which would be highly desirable.⁸  Of note, CD21⁻ B lymphocytes are increased in proportion in autoimmune diseases such as systemic lupus erythematosus and active cGVHD.^1,10  Increased proportions of CD21⁻ B lymphocytes could be part of the autoimmune pathogenesis compatible with inefficient censoring of autoreactive B cells in cGVHD.⁹  Disrupted B-cell homeostasis and relative rather than absolute preponderance of an activated B-cell pool has recently been reported in cGVHD.⁴  Our findings warrant larger prospective studies analyzing the predictive value of B-cell subsets for successful ECP treatment of cGVHD.