Distinct Pattern of Regulatory T Cell Reconstitution after Umbilical Cord Blood Transplantation in Pediatric Patients with Reduced Intensity Conditioning Regimen

Introduction: Compared to peripheral blood, umbilical cord blood (UCB) contains relatively higher frequencies of regulatory T cells (Tregs) with a distinct CD4+CD25bright subset. This may partially explain the less frequent and less severe acute graft-versus-host disease (aGvHD) seen in UCB transplantation (UCBT), in particular when comparing to similarly HLA-matched or mismatched living donor recipients. Treg reconstitution after UCBT is far less understood compared to living donor blood and marrow transplantation. It may also influence GvHD occurrence. Therefore, we set out to longitudinally examine the regeneration of Treg subsets and their kinetics that is impacted by proliferation and apoptosis in parallel. Tregs were defined as T cells expressing CD4+CD25brightCD127lowFoxp3+ phenotypes. Their reconstitution was examined in parallel with CD3+T cells, conventional CD4+ T cells (CD4+Tcon), sjTREC and TCRVβ repertoire complexity score. Treg reconstitution in patients with or without aGvHD was also evaluated.

Methods: Twenty pediatric patients (age range from 6m to 15y) with non-malignant diseases were transplanted and studied under an IRB approved protocol with low distant Campath+Flu/Mel/Thiotepa based reduced intensity conditioning regimen (RIC) (NCT01852370). Treg subsets were examined based on combinatorial expression of CD4, CD25, CD127, Foxp3, Helios, CD45RO, CD45RA, CD62L. Proliferation activity was determined by Ki67 antigen expression. Gating strategies are shown in Fig 1. Resting/activated Tregs were defined according to Sakaguchi publication (Immunity. 2009;30:899-911). Blood samples were collected pre-UCBT, on day 100 (D+100), at 6, 9 months, and 1 year post-UCBT. For each subset in each patient, the "reconstitution rate" (RR) was determined to represent recovery status of each population, and to normalize variations among individuals and different subsets. It was calculated as a ratio of given parameter at a specific time point post-UCBT, compared to same parameter measured pre-UCBT.

Results and Discussion:

Treg recovery compared to other T cells. After 100 days post-RIC-UCBT, the absolute # of Tregs in those patients without aGvHD recovered to 58% pre-UCBT (0.58 RR), which was comparable to TCRb repertoires regeneration (0.8 RR). However, Treg reconstitution was faster than the recovery rate in total CD3+ T cells (0.17 RR, p=0.05), CD4+Tcon (0.22 RR, p=0.6), CD8+ Tcon (0.18 RR, p=0.3) and sjTREC (0 RR, p=0.006). Tregs co-expressing central memory phenotype (CD45RO+CD62L+) had faster recovery rates (1.0 RR, p=0.001)) than naïve subsets (0.02RR). These findings point to rapid Treg recovery post-RIC-UCBT.

Treg recovery and aGvHD. At D+100, the absolute number of Tregs in patients with aGvHD, was significantly lower (Fig3a) explained by fewer of them entering the cell cycle (lower abs# of Ki67+Treg/ul, Fig3c), while there was no difference in CD4+Tcon values whether or not aGvHD occurred (Fig3b, d). This leads to an overall decreased ratio of #Treg/#CD4+Tcon and #Ki67+Treg/#Ki67+CD4+Tcon. At D+100, those with aGvHD had fewer Tregs in circulation (#/ul) displaying the central memory phenotype or Helios (Fig4). These findings point to impaired Treg homeostasis in patients with aGvHD primarily due to lower rates of proliferation, since apoptosis was not increase as measured by activated Casp3 expression. Reduced Treg numbers in those with aGvHD is likely not the consequence of poor thymopoiesis at D+100, since sjTREC production post-UCBT were similarly absent in both groups. Alternatively, it reflects differences in homeostatic expansion of the infused Treg pool.