Regulatory T cells (Treg) play a major role in controlling GvHD. Therefore many clinical studies focus on prophylactic or therapeutic Treg approaches. We focus on Treg reconstitution and their function in GvHD. Our aim is to identify patients with altered Treg function which may benefit from Treg DLI.


We analysed peripheral blood samples of 38 patients with acute and 11 patients with chronic GVHD as well as 18 patients without GVHD after Alemtuzumab based conditioning, 10 patients with ATG for TCD and 10 healthy individuals. To identify Treg we stained for CD3, CD4, CD25, CD127, FoxP3. As activation markers GARP, TGF-beta, HLA-DR, CD45RA, CD39, CD44, CD62L, Granzyme A, CCR5, ILT3, CTLA4, CXCR3 and GITR were used. To investigate the influence of the CD52-AB Alemtuzumab, we also stained for CD52.

Treg function was evaluated in CSFE suppression assays and TGF beta secretion. To investigate patient derived natural occurring Treg we sorted Treg from patients with GVHD or healthy controls with FACS and performed a CFSE proliferation assay with healthy PBMCs as effector T cells. To rule out the influence of systemic immunosuppression we analysed the effect of corticosteroids and CSA A on Treg marker expression and activation in healthy Treg. Treg were activated in vitro with IL2 (100U/ml AIM V+ human serum 10%) and OKT3 (30ng/ml).


Like in other publications in our cohort there were differences in absolute and relative Treg numbers in patients with acute, chronic or no GVHD compared with healthy controls: patients with aGVHD (mean 2,2%) had less Treg in comparison to patients with chronic (mean 4,5% p=0.0076) or no GVHD (mean 4,4% p=0.0045).

CSA A or corticosteroids ex vivo (and in vivo) did not influence Treg marker expression significantly.

Especially in patients after Alemtuzumab CD52 expression is the most prominent marker: Frequency of CD52pos Treg differed highly significant between patients with aGVHD (mean 15,94%) in comparison to patients with chronic (mean 75,85%) or no GVHD (mean 92,09%; p<0,0001). For all patients with acute or chronic GVHD FACS sorted Treg showed reduced capacity to suppress T cell proliferation in the proliferation assay.

To determine the suppressive capacity of CD52neg Treg in comparison with CD52pos Treg of the same patient we divided Treg from one patient by FACS-Sort in CD52pos/neg subpopulations, which were used in a CFSE-proliferation assay again.

We could show that CD52neg Tregwere less able to suppress effector T cell proliferation in comparison to CD52pos Treg (tested in 3 patients). Independent of the conditioning regimen Treg from GVHD patients were less able to suppress effector T cells in comparison with Treg from patients without GVHD or healthy individuals.

We could not show any difference in activation markers ex vivo except GARP/TGFbeta. CD52neg Treg showed a reduced GARP/TGFbeta expression (mean 4,6% vs 1,65% p=0,012). After stimulation in vitro GARP/TGFbeta healthy controls showed the highest GARP/TGF-ß expression compared to patients with any GVHD, patients without GVHD exhibited higher amounts of GARP/TGFbeta compared to patients with aGVHD.

We also analyzed patients after Alemtuzmab weekly during the first 200 days after HSCT to monitor the reconstitution of CD52neg Treg as a predictive marker for the development of aGVHD (n= 10). Patients who reconstituted with more than 50% of CD52pos Treg before d50 never developed any aGVHD. Patients with aGVHD and persistence of high amounts of CD52neg Treg showed a steroid refractory aGVHD with high mortality.


Patients with acute GVHD showed different marker expression and decreased Treg function. Especially CD52neg Treg derived from patients after Alemtuzumab showed a reduced suppressive capacity and lower expression of activation markers. Higher GARP/TGFbeta expression seems to be associated with less development of acute GVHD. CD52 and GARP expression/TGFbeta secretion could be helpful biomarkers to identify a patient cohort with functional altered Treg and high risk of severe acute GVHD. In particular this patient cohort may benefit from Treg-DLI.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.