Backgrounds: CD4+FoxP3+ regulatory T cells (Treg) have been shown in pre-clinical murine models and human clinical trials to protect against acute graft-versus-host (GVHD) disease following hematopoietic cell transplantation (HCT). In murine models, Treg expression of CD62L is required for in vivo protection of GVHD (Strober 2005/Blazer 2004). Certain aspects of Treg isolation and administration could affect CD62L expression, such as the use of G-CSF mobilization or freeze-and-thaw. We hypothesized that freeze-thaw of Tregs could reduce CD62L expression and prevent Tregs from protecting against GVHD. Furthermore, we hypothesized that the loss of CD62L expression is mediated at least in part by metalloproteinases (MP).

Methods and Results: We used established protocols of GVHD, in which lethal dose of fractionated total body irradiation was followed by murine bone marrow transplantation across major histocompatibility barriers. We sort-purified CD4+Foxp3+ Tregs based on CD25+(high) expression and measured baseline expression of CD62L. Cells were then gradually frozen in fetal calf serum/10%DMSO, and finally stored in liquid nitrogen for at least 7 days. After thawing Tregs showed an average of 58% of the baseline CD62L expression (CD62L expression fresh vs. frozen p=0.0009, n=5). Binding of frozen Tregs to plate bound L-Selectin ligand Mucosal addressin cell adhesion molecule-1 (MAdCAM-1) in vitrowas significantly reduced as compared to fresh Tregs (p=0.0003). To test the potential effect of freeze and thaw in vivo, we used an established murine model in which Tregs are given prior to conventional T cells (Tcon) to protect against GHVD. At the time of transplant conditioned BALB/c donors received 5x10^6 whole bone marrow and 5x10^5 CD4+CD25(high) Tregs (day 0), followed by 1x10^6 Tcon (day +2). Frozen/thawed Tregs and fresh Tregs respectively were given at the same time point at the same quantity of viable cells. Using luciferase-expressing cells (luc+) and BLI imaging we could show that homing of frozen luc+Tregs into pLN at day+5 was impaired as compared to fresh Tregs (p=0.0005). Relative weight loss as an indicator for GVHD was significantly higher and survival was significantly reduced in mice receiving frozen Treg in comparison to mice receiving fresh Treg. To confirm the importance of CD62L expression on Treg function we blocked CD62L in vivo by incubating fresh Tregs with monoclonal blocking antibody Mel14 (120ug/ml) for 1 hour prior to injection. A significantly higher proliferation of luc+ Tcon was observed in mice injected with Mel14-treated Tregs as compared to isotype control-treated Tregs. To test the potential role of metalloproteinases in cleaving CD62L following freeze and thaw, we added various metalloproteinase inhibitors (MPI) during the freezing process. We found that the addition of at least one MPI GM6001 (50nM) can diminish the loss of CD62L.

Conclusion: Augmenting or manipulating Tregs of the donor graft is a promising area of cellular therapy that could be used to prevent GVHD in HCT and in potentially many other clinical contexts. We have found that the process of freeze and thaw reduces CD62L expression on Tregs and reduces protection against GVHD. Further, MPs may play role in the loss of CD62L expression and the use of MPIs for frozen products might preserve Treg function.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.