Opposite kinetics of TCRαβ⁺ effector and regulatory T-cell frequencies separate patients with therapy-refractory aGVHD from those with therapy-responsive aGVHD. (A) Heatmap displaying phenotypically different CD4- or CD8-expressing T-cell subclusters, including CD4⁺ Tregs. Effector T-cell subclusters were separated on the combined presence (chemokineR^high) or absence (chemokineR^low) of CXCR3, CCR9, and CCR10. Matching color codes on the y-axis identify the markers used for subcluster annotation. (B) Frequency of effector T-cell and Treg populations with differential expression of chemokine receptors (as defined in [A]) in 3 different patient groups over time (purple indicates HSCT patients with steroid-responsive aGVHD [Steroid-CR]; green indicates HSCT patients with steroid-refractory aGVHD responsive to MSC-based second-line therapy [GVHD-CR]; red indicates HSCT patients with steroid-refractory aGVHD nonresponsive to MSC-based second-line therapy [GVHD-NR]). Differences in cluster frequencies before (t = 1) and after initiation of immune suppressive therapy (steroids only or steroids plus MSC) were compared using the Wilcoxon signed-rank test. (C) Radial plots showing differences in individual T-cell subclusters found in HSCT patients with steroid-refractory aGVHD responsive to MSC-based second-line therapy (GVHD-CR in green) and HSCT patients with steroid-refractory aGVHD nonresponsive to MSC-based second-line therapy (GVHD-NR in red) before (left) and 1 week (center) or 4 weeks (right) after initiation of MSC treatment. Cluster numbers correspond to heatmap annotation in (A). Significantly more prevalent T-cell subclusters are shown in bold. Scale of radial plots represent cell frequency (×10²) as percentage of the major CD4 or CD8 lineage. ∗P < .05; ∗∗P < .01.