Abstract 1949


Chronic graft-versus-host disease (cGVHD) is the most common long-term complication of allogeneic stem cell transplant (allo-SCT), affecting 30–70% of patients who survive beyond the first 100 days. While there is evidence to support a role for activated B cells, the exact cause of cGVHD remains unknown. Thus, we evaluated the influence of lymphocyte reconstitution at days 30 and 100 following allo-SCT on subsequent development of cGVHD.


An extensive immune reconstitution flow cytometric “immunome” assay was developed at The Ohio State University, allowing for monitoring of changes in cell activation markers, memory T cell status, Treg subsets, NK cell subsets and Th1 vs Th2 cell subsets. To evaluate the effect of immune reconstitution on development of cGVHD, we have correlated samples collected at days 30 and 100 following allo-SCT with an IRB-approved, clinical database of patients who have received an SCT. Day 30 samples were collected on 70 patients, 66 patients had day 100 samples, and 40 patients had both. Logistic regression was used to evaluate the influence of both absolute numbers and percentage of lymphocyte subsets on the subsequent development of cGVHD. Because so few subsets were associated at both d30 and d100, we evaluated the influence of change over time using a Wilcoxon rank sum test.


All patients received either a peripheral blood or bone marrow graft, and 62% had unrelated donors. The median age was 55, and the median time to development of cGVHD was 155 days (range 110–389). On univariate analysis for the d30 samples, increased odds of developing cGVHD were associated with increased absolute numbers of naïve CD4+ and CD8+ T cells (p=0.05 and 0.02, respectively), as well as CD3-/CD56+16+/CD117- NK cells (p=0.05). These were not associated with cGVHD on univariate analysis for the d100 samples, but CD4+/CD193+ (p=0.018) and CD4+/CD183+ (p=0.015) were significantly associated at that time point. A higher percentage of activated NK cells (p=0.002) at d30, as well as a higher percentage of CD3+/CD69+cells (p=0.04), appears negatively associated with subsequent development of cGVHD. Higher percentages of both activated and naïve CD4+ cells at d100 were associated with cGVHD (p=0.03–0.04). Additionally, higher percentages of CD4+/CD29+ (p=0.03), CD4+/CD193+ (p=0.016), and CD4+/CD183+ cells (p=0.04) were positively associated. A higher percentage of CD8+/CD27- (p=0.02), CD8+/CD29- (p=0.03), andCD27-/CDRO+ (p=0.045) cells appeared to have a negative association with cGVHD. Percent of CD4+/CD27+ cells was the only marker to have an association at both d30 and d100 (p=0.045 and 0.01 respectively). From d30 to d100, there were statistically significant larger decreases in the absolute number of CD8+/CD45RO- T cells (p<0.0001) and CD3-/CD56+16+/CD117- NK cells (p<0.0001) among patients who developed cGVHD. A larger decrease in the percentage of CD3-/CD56+16+/CD117- NK cells was seen, as well (p<0.0001). A larger increase in the percentages of CD19+/CD86+ B cells (p=0.023) and naïve CD4+ T cells were associated with cGVHD. cGVHD was also associated with an increase in both the number and percentage of CD4+/CD193+ cells (p=0.006 and 0.0001 respectively) over time, compared with a decrease among patients who did not develop cGVHD. Multivariate analysis is ongoing.


Patients who developed cGVHD had a larger increase in CD4+ T cells and a smaller increase in CD8+ T cells compared with patients who did not develop cGVHD over time, suggesting a selective expansion of CD4+ T cells. Further, a significant decrease in NK cells and concomitant increase in percentages of activated B cells was noted. An increase in CD4+ cells is associated with an inflammatory phenotype, and a Th2-skewed proinflammatory response may contribute to B cell activation, which has previously been associated with cGVHD. The presence of a Th2-skewed phenotype is supported by the presence of increased CD4+/CD193+ cells among patients with cGVHD, as CCR3 is preferentially expressed on Th2 cells. While a Th2-skewed phenotype has been demonstrated in several mouse models of cGVHD, this has not been established in humans. Further analysis will be needed before conclusions can be reached, but routine use of the “immunome” and clinical correlation may allow for more insight into the pathogenesis of cGVHD in humans and contribute to identification of targets for therapeutic intervention.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.