Research ArticleImmunologyTransplantation Open Access | 10.1172/jci.insight.162978

Xiaoqian Ma,1,2 Lu Cao,1,2 Martina Raneri,1 Hannah Wang,1 Qi Cao,1 Yuanfei Zhao,1 Naiara G. Bediaga,3 Gaetano Naselli,3 Leonard C. Harrison,3 Wayne J. Hawthorne,1 Min Hu,1 Shounan Yi,1 and Philip J. O’Connell1

1Centre for Transplantation and Renal Research, Westmead Institute for Medical Research, University of Sydney, Sydney, New South Wales, Australia.

2Cell Transplantation and Gene Therapy Institute, The Third Xiangya Hospital, Central South University, Changsha, China.

3Walter and Eliza Hall Institute of Medical Research, University of Melbourne, Melbourne, Victoria, Australia.

Address correspondence to: Philip J. O’Connell, Min Hu, or Shounan Yi, Westmead Institute for Medical Research, 176 Hawkesbury Road, Westmead, New South Wales 2145, Australia. Phone: 61.2.8627.3005; Email: philip.oconnell@sydney.edu.au (PJO). Phone: 61.2.8627.3049; Email: min.hu@sydney.edu.au (MH). Phone: 61.2.8627.3085; Email: shounan.yi@sydney.edu.au (SY).

Authorship note: XM, LC, and MR are co–first authors. MH, SY, and PJO are co–corresponding authors.

Find articles by Ma, X. in: JCI | PubMed | Google Scholar

1Centre for Transplantation and Renal Research, Westmead Institute for Medical Research, University of Sydney, Sydney, New South Wales, Australia.

2Cell Transplantation and Gene Therapy Institute, The Third Xiangya Hospital, Central South University, Changsha, China.

3Walter and Eliza Hall Institute of Medical Research, University of Melbourne, Melbourne, Victoria, Australia.

Address correspondence to: Philip J. O’Connell, Min Hu, or Shounan Yi, Westmead Institute for Medical Research, 176 Hawkesbury Road, Westmead, New South Wales 2145, Australia. Phone: 61.2.8627.3005; Email: philip.oconnell@sydney.edu.au (PJO). Phone: 61.2.8627.3049; Email: min.hu@sydney.edu.au (MH). Phone: 61.2.8627.3085; Email: shounan.yi@sydney.edu.au (SY).

Authorship note: XM, LC, and MR are co–first authors. MH, SY, and PJO are co–corresponding authors.

Find articles by Cao, L. in: JCI | PubMed | Google Scholar

1Centre for Transplantation and Renal Research, Westmead Institute for Medical Research, University of Sydney, Sydney, New South Wales, Australia.

2Cell Transplantation and Gene Therapy Institute, The Third Xiangya Hospital, Central South University, Changsha, China.

3Walter and Eliza Hall Institute of Medical Research, University of Melbourne, Melbourne, Victoria, Australia.

Address correspondence to: Philip J. O’Connell, Min Hu, or Shounan Yi, Westmead Institute for Medical Research, 176 Hawkesbury Road, Westmead, New South Wales 2145, Australia. Phone: 61.2.8627.3005; Email: philip.oconnell@sydney.edu.au (PJO). Phone: 61.2.8627.3049; Email: min.hu@sydney.edu.au (MH). Phone: 61.2.8627.3085; Email: shounan.yi@sydney.edu.au (SY).

Authorship note: XM, LC, and MR are co–first authors. MH, SY, and PJO are co–corresponding authors.

Find articles by Raneri, M. in: JCI | PubMed | Google Scholar

1Centre for Transplantation and Renal Research, Westmead Institute for Medical Research, University of Sydney, Sydney, New South Wales, Australia.

2Cell Transplantation and Gene Therapy Institute, The Third Xiangya Hospital, Central South University, Changsha, China.

3Walter and Eliza Hall Institute of Medical Research, University of Melbourne, Melbourne, Victoria, Australia.

Address correspondence to: Philip J. O’Connell, Min Hu, or Shounan Yi, Westmead Institute for Medical Research, 176 Hawkesbury Road, Westmead, New South Wales 2145, Australia. Phone: 61.2.8627.3005; Email: philip.oconnell@sydney.edu.au (PJO). Phone: 61.2.8627.3049; Email: min.hu@sydney.edu.au (MH). Phone: 61.2.8627.3085; Email: shounan.yi@sydney.edu.au (SY).

Authorship note: XM, LC, and MR are co–first authors. MH, SY, and PJO are co–corresponding authors.

Find articles by Wang, H. in: JCI | PubMed | Google Scholar |

1Centre for Transplantation and Renal Research, Westmead Institute for Medical Research, University of Sydney, Sydney, New South Wales, Australia.

2Cell Transplantation and Gene Therapy Institute, The Third Xiangya Hospital, Central South University, Changsha, China.

3Walter and Eliza Hall Institute of Medical Research, University of Melbourne, Melbourne, Victoria, Australia.

Address correspondence to: Philip J. O’Connell, Min Hu, or Shounan Yi, Westmead Institute for Medical Research, 176 Hawkesbury Road, Westmead, New South Wales 2145, Australia. Phone: 61.2.8627.3005; Email: philip.oconnell@sydney.edu.au (PJO). Phone: 61.2.8627.3049; Email: min.hu@sydney.edu.au (MH). Phone: 61.2.8627.3085; Email: shounan.yi@sydney.edu.au (SY).

Authorship note: XM, LC, and MR are co–first authors. MH, SY, and PJO are co–corresponding authors.

Find articles by Cao, Q. in: JCI | PubMed | Google Scholar

1Centre for Transplantation and Renal Research, Westmead Institute for Medical Research, University of Sydney, Sydney, New South Wales, Australia.

2Cell Transplantation and Gene Therapy Institute, The Third Xiangya Hospital, Central South University, Changsha, China.

3Walter and Eliza Hall Institute of Medical Research, University of Melbourne, Melbourne, Victoria, Australia.

Address correspondence to: Philip J. O’Connell, Min Hu, or Shounan Yi, Westmead Institute for Medical Research, 176 Hawkesbury Road, Westmead, New South Wales 2145, Australia. Phone: 61.2.8627.3005; Email: philip.oconnell@sydney.edu.au (PJO). Phone: 61.2.8627.3049; Email: min.hu@sydney.edu.au (MH). Phone: 61.2.8627.3085; Email: shounan.yi@sydney.edu.au (SY).

Authorship note: XM, LC, and MR are co–first authors. MH, SY, and PJO are co–corresponding authors.

Find articles by Zhao, Y. in: JCI | PubMed | Google Scholar

1Centre for Transplantation and Renal Research, Westmead Institute for Medical Research, University of Sydney, Sydney, New South Wales, Australia.

2Cell Transplantation and Gene Therapy Institute, The Third Xiangya Hospital, Central South University, Changsha, China.

3Walter and Eliza Hall Institute of Medical Research, University of Melbourne, Melbourne, Victoria, Australia.

Address correspondence to: Philip J. O’Connell, Min Hu, or Shounan Yi, Westmead Institute for Medical Research, 176 Hawkesbury Road, Westmead, New South Wales 2145, Australia. Phone: 61.2.8627.3005; Email: philip.oconnell@sydney.edu.au (PJO). Phone: 61.2.8627.3049; Email: min.hu@sydney.edu.au (MH). Phone: 61.2.8627.3085; Email: shounan.yi@sydney.edu.au (SY).

Authorship note: XM, LC, and MR are co–first authors. MH, SY, and PJO are co–corresponding authors.

Find articles by Bediaga, N. in: JCI | PubMed | Google Scholar

1Centre for Transplantation and Renal Research, Westmead Institute for Medical Research, University of Sydney, Sydney, New South Wales, Australia.

2Cell Transplantation and Gene Therapy Institute, The Third Xiangya Hospital, Central South University, Changsha, China.

3Walter and Eliza Hall Institute of Medical Research, University of Melbourne, Melbourne, Victoria, Australia.

Address correspondence to: Philip J. O’Connell, Min Hu, or Shounan Yi, Westmead Institute for Medical Research, 176 Hawkesbury Road, Westmead, New South Wales 2145, Australia. Phone: 61.2.8627.3005; Email: philip.oconnell@sydney.edu.au (PJO). Phone: 61.2.8627.3049; Email: min.hu@sydney.edu.au (MH). Phone: 61.2.8627.3085; Email: shounan.yi@sydney.edu.au (SY).

Authorship note: XM, LC, and MR are co–first authors. MH, SY, and PJO are co–corresponding authors.

Find articles by Naselli, G. in: JCI | PubMed | Google Scholar

1Centre for Transplantation and Renal Research, Westmead Institute for Medical Research, University of Sydney, Sydney, New South Wales, Australia.

2Cell Transplantation and Gene Therapy Institute, The Third Xiangya Hospital, Central South University, Changsha, China.

3Walter and Eliza Hall Institute of Medical Research, University of Melbourne, Melbourne, Victoria, Australia.

Address correspondence to: Philip J. O’Connell, Min Hu, or Shounan Yi, Westmead Institute for Medical Research, 176 Hawkesbury Road, Westmead, New South Wales 2145, Australia. Phone: 61.2.8627.3005; Email: philip.oconnell@sydney.edu.au (PJO). Phone: 61.2.8627.3049; Email: min.hu@sydney.edu.au (MH). Phone: 61.2.8627.3085; Email: shounan.yi@sydney.edu.au (SY).

Authorship note: XM, LC, and MR are co–first authors. MH, SY, and PJO are co–corresponding authors.

Find articles by Harrison, L. in: JCI | PubMed | Google Scholar

1Centre for Transplantation and Renal Research, Westmead Institute for Medical Research, University of Sydney, Sydney, New South Wales, Australia.

2Cell Transplantation and Gene Therapy Institute, The Third Xiangya Hospital, Central South University, Changsha, China.

3Walter and Eliza Hall Institute of Medical Research, University of Melbourne, Melbourne, Victoria, Australia.

Address correspondence to: Philip J. O’Connell, Min Hu, or Shounan Yi, Westmead Institute for Medical Research, 176 Hawkesbury Road, Westmead, New South Wales 2145, Australia. Phone: 61.2.8627.3005; Email: philip.oconnell@sydney.edu.au (PJO). Phone: 61.2.8627.3049; Email: min.hu@sydney.edu.au (MH). Phone: 61.2.8627.3085; Email: shounan.yi@sydney.edu.au (SY).

Authorship note: XM, LC, and MR are co–first authors. MH, SY, and PJO are co–corresponding authors.

Find articles by Hawthorne, W. in: JCI | PubMed | Google Scholar |

1Centre for Transplantation and Renal Research, Westmead Institute for Medical Research, University of Sydney, Sydney, New South Wales, Australia.

2Cell Transplantation and Gene Therapy Institute, The Third Xiangya Hospital, Central South University, Changsha, China.

3Walter and Eliza Hall Institute of Medical Research, University of Melbourne, Melbourne, Victoria, Australia.

Address correspondence to: Philip J. O’Connell, Min Hu, or Shounan Yi, Westmead Institute for Medical Research, 176 Hawkesbury Road, Westmead, New South Wales 2145, Australia. Phone: 61.2.8627.3005; Email: philip.oconnell@sydney.edu.au (PJO). Phone: 61.2.8627.3049; Email: min.hu@sydney.edu.au (MH). Phone: 61.2.8627.3085; Email: shounan.yi@sydney.edu.au (SY).

Authorship note: XM, LC, and MR are co–first authors. MH, SY, and PJO are co–corresponding authors.

Find articles by Hu, M. in: JCI | PubMed | Google Scholar

1Centre for Transplantation and Renal Research, Westmead Institute for Medical Research, University of Sydney, Sydney, New South Wales, Australia.

2Cell Transplantation and Gene Therapy Institute, The Third Xiangya Hospital, Central South University, Changsha, China.

3Walter and Eliza Hall Institute of Medical Research, University of Melbourne, Melbourne, Victoria, Australia.

Address correspondence to: Philip J. O’Connell, Min Hu, or Shounan Yi, Westmead Institute for Medical Research, 176 Hawkesbury Road, Westmead, New South Wales 2145, Australia. Phone: 61.2.8627.3005; Email: philip.oconnell@sydney.edu.au (PJO). Phone: 61.2.8627.3049; Email: min.hu@sydney.edu.au (MH). Phone: 61.2.8627.3085; Email: shounan.yi@sydney.edu.au (SY).

Authorship note: XM, LC, and MR are co–first authors. MH, SY, and PJO are co–corresponding authors.

Find articles by Yi, S. in: JCI | PubMed | Google Scholar

1Centre for Transplantation and Renal Research, Westmead Institute for Medical Research, University of Sydney, Sydney, New South Wales, Australia.

2Cell Transplantation and Gene Therapy Institute, The Third Xiangya Hospital, Central South University, Changsha, China.

3Walter and Eliza Hall Institute of Medical Research, University of Melbourne, Melbourne, Victoria, Australia.

Address correspondence to: Philip J. O’Connell, Min Hu, or Shounan Yi, Westmead Institute for Medical Research, 176 Hawkesbury Road, Westmead, New South Wales 2145, Australia. Phone: 61.2.8627.3005; Email: philip.oconnell@sydney.edu.au (PJO). Phone: 61.2.8627.3049; Email: min.hu@sydney.edu.au (MH). Phone: 61.2.8627.3085; Email: shounan.yi@sydney.edu.au (SY).

Authorship note: XM, LC, and MR are co–first authors. MH, SY, and PJO are co–corresponding authors.

Find articles by O’Connell, P. in: JCI | PubMed | Google Scholar |

Authorship note: XM, LC, and MR are co–first authors. MH, SY, and PJO are co–corresponding authors.

Published October 24, 2023 - More info

Published in Volume 8, Issue 23 on December 8, 2023
JCI Insight. 2023;8(23):e162978. https://doi.org/10.1172/jci.insight.162978.
© 2023 Ma et al. This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Published October 24, 2023 - Version history
Received: June 23, 2022; Accepted: October 17, 2023 View PDF Abstract

Regulatory T cells (Tregs) have potential for the treatment of autoimmune diseases and graft rejection. Antigen specificity and functional stability are considered critical for their therapeutic efficacy. In this study, expansion of human Tregs in the presence of porcine PBMCs (xenoantigen-expanded Tregs, Xn-Treg) allowed the selection of a distinct Treg subset, coexpressing the activation/memory surface markers HLA-DR and CD27 with enhanced proportion of FOXP3+Helios+ Tregs. Compared with their unsorted and HLA-DR+CD27+ double-positive (DP) cell–depleted Xn-Treg counterparts, HLA-DR+CD27+ DP-enriched Xn-Tregs expressed upregulated Treg function markers CD95 and ICOS with enhanced suppression of xenogeneic but not polyclonal mixed lymphocyte reaction. They also had less Treg-specific demethylation in the region of FOXP3 and were more resistant to conversion to effector cells under inflammatory conditions. Adoptive transfer of porcine islet recipient NOD/SCID IL2 receptor γ–/– mice with HLA-DR+CD27+ DP-enriched Xn-Tregs in a humanized mouse model inhibited porcine islet graft rejection mediated by 25-fold more human effector cells. The prolonged graft survival was associated with enhanced accumulation of FOXP3+ Tregs and upregulated expression of Treg functional genes, IL10 and cytotoxic T lymphocyte antigen 4, but downregulated expression of effector Th1, Th2, and Th17 cytokine genes, within surviving grafts. Collectively, human HLA-DR+CD27+ DP-enriched Xn-Tregs expressed a specific regulatory signature that enabled identification and isolation of antigen-specific and functionally stable Tregs with potential as a Treg-based therapy.

Graphical Abstract Introduction

Regulatory T cells (Tregs) are an essential component of immune homeostasis. Tregs were initially characterized as CD4+CD25+ T cells in mouse models of autoimmune disease and make up 5% to 10% of peripheral CD4+ T cells (1, 2). FOXP3 is the key transcription factor that characterizes this population of thymically derived Tregs (2, 3). In humans, mutation of the FOXP3 gene leads to severe autoimmune disease, demonstrating the importance of this cell subset in suppressing unwanted inflammatory responses to self-antigens (4). In addition to their role in preventing autoimmunity, CD4+CD25+ T cells that express FOXP3 have been shown to be important for suppressing alloimmunity and for inducing and maintaining allograft tolerance (5, 6). In models of transplantation, dominant tolerance is characterized by indefinite graft survival after an initial brief period of induction therapy in the presence of an otherwise intact immune system. In other words, this nonresponsiveness is antigen specific. Given their potential to reduce the requirement for immunosuppression, CD4+CD25+FOXP3+ Tregs have been pursued for their therapeutic potential to suppress autoimmunity and to reduce or eliminate the requirement for immunosuppression after transplantation. Trials of naive Tregs in kidney or liver transplantation have been disappointing in that the results have been modest and the relatively large number of Tregs required has provided regulatory, cost, and production challenges (7–9).

Animal studies of transplant tolerance have shown that both the induction and maintenance of tolerance depend on the development of antigen-specific Tregs, as deletion of Foxp3+ Tregs leads to prompt rejection (10, 11). Also, a relatively small number of cells can transfer graft-specific tolerance to a naive host, as has been shown by numerous adoptive transfer studies (12–14). The observations are equally true for autoimmunity as they are for all immunity. For example, in models of autoimmune type 1 diabetes, antigen-specific Tregs, which were isolated from pancreatic lymph nodes or pulsed with islet antigen, were superior to polyclonal Tregs at preventing or curing the disease (15, 16). Similarly, alloantigen-specific Tregs, enriched by alloantigen-stimulated expansion in vitro or engineered to express a T cell receptor (TCR) transgene, were more effective than polyclonal Tregs at preventing rejection of organ and tissue grafts. Studies in humanized mouse models have shown similar results: alloantigen-expanded human Tregs were more potent suppressors of skin graft rejection than were their polyclonal counterparts (14), and human alloantigen-specific Tregs generated with an HLA-A2–specific chimeric antigen receptor were superior to polyclonal Tregs at preventing xenogeneic graft versus host disease (GVHD) caused by HLA-A2+ T cells (17). An important advantage of enhanced Treg antigen specificity is that the suppression is targeted to the graft, hence avoiding potential opportunistic infection and malignancy that may result from nonspecific suppression as a result of the application of polyclonal Tregs.

One of the problems in developing in vitro–expanded Tregs has been identifying markers that would facilitate their selection after stimulation. Intracellular FOXP3 expression, enhanced Helios expression, and demethylation of a Treg-specific demethylated region (TSDR) within the FOXP3 locus represent the gold standard for estimating the fraction of stable Tregs within a population. However, it does not allow for sorting a specific subset that could be used therapeutically. The lack of discriminative markers also affects systematic functional optimization of in vitro–generated Tregs, such as genetically engineered Tregs with transgenic TCR or chimeric antigen receptor (CAR) constructs. To identify the characteristics of an antigen-activated subset, we have used a model of Treg development to xenoantigens such as porcine PBMC or neonatal porcine islet cell clusters (NICCs) (18, 19). Due to the phylogenetic distance between pig and humans, these provide a large antigen load and hence provide a larger pool of antigen-specific Tregs available for study. For instance, we have also reported that human Tregs expanded ex vivo with xenoantigen stimulation are more potent than polyclonal Tregs at suppressing xenoreactive effector cell proliferation in a xenogeneic mixed lymphocyte reaction (MLR), although they are equally suppressive as polyclonal Tregs in an allogeneic or polyclonal MLR xenogeneic response, indicating acquisition of xenoantigen specificity after xenoantigen stimulation (18).

A number of Treg activation-induced surface markers, such as HLA-DR (20), CD27 (21), CD45RO (22), and ICOS (23), have been described to identify activated and/or memory Tregs (24). Among those, HLA-DR+ Tregs have been reported to express higher levels of Treg-associated activation markers and produce lower levels of effector cytokines (20, 24). HLA-DR+ Tregs are present in human peripheral blood, thymus, and umbilical cord blood and are more suppressive than HLA-DR– Tregs in vitro (20). CD27 expression has also been associated with increased Treg suppressive function (21, 25). It has been shown that a CD4+CD127−/loCD25+CD45RA− Treg subpopulation cultured in vitro with tacrolimus lost their Treg TSDR demethylation phenotype, which correlated with a reduction of CD27 expression, suggesting an association of CD27 expression with Treg stability (26). Thus, given that human Tregs expanded with xenoantigen have been demonstrated already to express upregulated levels of activated/effector markers, HLA-DR, ICOS, and CD45RO, which were associated with their enhanced suppression (18), it is feasible that one or a combination of surface activation markers could be identified as a specific signature for the isolation of a stable, antigen-specific Treg subset that could be suitable for clinical immunotherapy.

Based on previous findings, we wished to test the hypothesis that human Tregs expanded in the presence of antigen and expressing the surface markers of HLA-DR and CD27 would represent a stable antigen-specific Treg population that is more potent than their unsorted counterparts at suppressing a graft immune response. To test this hypothesis, we used the human T cell response to porcine PBMCs in vitro and NICCs in vivo as a proof-of-concept study, where a human HLA-DR+CD27+ double-positive (DP) Treg subset from xenoantigen-expanded Tregs was compared in efficacy and stability with their unsorted, HLA-DR+CD27+ DP-depleted Treg and polyclonal Treg counterparts at protecting islet xenografts from rejection mediated by human T cells.

Results

Xenoantigen-expanded Tregs express upregulated levels of Treg activation/memory markers HLA-DR and CD27. Tregs (CD4+CD25+CD127−/lo) were isolated from human PBMCs and stimulated either with anti-CD3/CD28 beads alone (polyclonally stimulated Tregs, Pc-Treg) or with anti-CD3/CD28 dynabeads combined with irradiated porcine PBMCs (xenoantigen-expanded Tregs, Xn-Treg) for 3 rounds (7 d/round). Consistent with our previous study (18), Xn-Treg demonstrated similar levels of purity to their freshly isolated Treg (Fresh-Treg) and Pc-Treg counterparts (96% vs. 96.2% vs. 89.6% of CD4+CD25+ cells of Xn-Treg vs. Pc-Treg vs. Fresh-Treg), with high-level expression of FOXP3, cytotoxic T lymphocyte antigen 4 (CTLA4), glucocorticoid-induced tumor necrosis factor receptor (GITR), and CD62L (Figure 1A). In addition to their upregulated expression of the activation marker CD62L (98.0% vs. 62.0% of CD4+CD25+CD62L+ of Xn-Treg vs. Fresh-Treg), there was no change in CD127 expression between the 2 ex vivo–expanded Treg subsets and Fresh-Treg (Figure 1A). Next, the transcription factor Helios, a marker for Treg stability (27), and coexpression of FOXP3 were assessed. A similar proportion of FOXP3+Helios+ Tregs was seen in Xn-Treg (61.3% ± 24.3%) compared to Fresh-Treg (62.5% ± 14.7%), while Pc-Treg showed a decreased proportion of FOXP3+Helios+ Tregs (41.1% ± 26.3%) (Figure 1B and Supplemental Figure 1; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.162978DS1), suggesting that Xn-Treg was a more stable Treg subset than Pc-Treg. The Xn-Treg also expressed higher levels of the Treg activation/memory surface markers HLA-DR and CD27 than PC-Treg, as evidenced by mean fluorescence intensity (MFI) (HLA-DR: 5,848 vs. 3,242.8 and CD27: 6,447.8 vs. 2,415.4 of Xn-Treg vs. Pc-Treg), while both types of expanded Tregs had similar levels of CD25 (MFI: 6,996 vs. 8,513.4 of Xn-Treg vs. Pc-Treg) and FOXP3 expression (MFI: 2,489.8 vs. 2,589 of Xn-Treg vs. Pc-Treg) (Supplemental Figure 2). Therefore, Xn-Treg showed a greater proportion of cells coexpressing HLA-DR and CD27 when compared with freshly isolated and polyclonally expanded Tregs (52% ± 11.0% vs. 24.3% ± 13.8% vs. 9.6% ± 8.0% of Xn-Treg vs. Pc-Treg vs. Fresh-Treg) (Figure 1C), suggesting that a Treg subset coexpressing HLA-DR and CD27 was selectively enriched in Xn-Treg.

Figure 1

Phenotypical characterization of ex vivo–expanded human Tregs. Representative flow cytometric plots of CD4+CD25+CD127–/lo Treg phenotypes isolated from human PBMCs (Fresh-Treg), Tregs expanded with anti-CD3/CD28 dynabeads (Pc-Treg), and stimulation in presence of irradiated porcine PBMCs (Xn-Treg) after 3 cycles (weeks) of stimulation. (A) Gates were set on CD4+ T cells. FOXP3 and other cell surface marker expression shown as the percentage of CD4+ T cells coexpressing individual Treg markers (CD4+CD25+, CD25+FOXP3+, CD25+CTLA4+, CD127–CD25+, CD62L+CD25+, CD25+GITR+). (B) The proportion of Tregs coexpressing FOXP3 and Helios on Fresh-Tregs, Pc-Tregs, Xn-Tregs and negative control effector T cells. The gating strategies are shown in Supplemental Figure 1. (C) The proportion of Tregs coexpressing HLA-DR and CD27 after gating on CD4+CD25+ cells. (D) Phenotyping of Xn-Tregs. Representative histograms of CD95 expression (surface and intracellular staining), ICOS (surface staining), CTLA4 (surface staining), FOXP3 (intracellular staining), and Helios (intracellular staining) on HLA-DR+CD27+ double-positive enriched Xn-Treg (DP-enriched; red line), total Xn-Treg (Total; green line), and Xn-Treg depleted of HLA-DR+CD27+ double positive cells (DP-depleted; blue line). Expression of CD95, ICOS, CTLA4, FOXP3, and Helios in different Treg subsets is also shown by the mean fluorescence intensity (MFI). Numbers in brackets in each plot are the ranges of the percentage of individual Treg markers detected in 4 independent experiments with Tregs from 4 individual donors (A) and 5 individual donors (B) and 7 independent experiments with 9 individual donors (C). Data represent 4 independent experiments with Xn-Tregs from 5 individual donors (D). Error bars indicate the mean ± SD (A–C) and mean ± SEM (D). One-way ANOVA: *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

We further analyzed the phenotype of the HLA-DR+CD27+ DP Treg subset of Xn-Treg (DP-enriched Xn-Treg) by flow cytometry, demonstrating that significantly enhanced intracellular expression of Treg function molecule CD95 (Fas) (28) was detected in HLA-DR+CD27+ DP-enriched Xn-Treg compared with either HLA-DR+CD27+ DP-depleted Xn-Treg or total Xn-Treg, as evidenced by MFI (83,708.2 ± 5,328.8 vs. 44,843.2 ± 5,879.9 vs. 56,270.4 ± 6,725.8 of DP-enriched Xn-Treg vs. DP-depleted Xn-Treg vs. total Xn-Treg) (Figure 1D). The enhanced CD95 expression was associated with enhanced HLA-DR and CD27 expression (Supplemental Figure 3A). Moreover, HLA-DR+CD27+ DP-enriched Xn-Treg expressed a significantly higher level of the Treg activation marker ICOS than DP-depleted Xn-Treg (MFI: 6,854.4 ± 498.5 vs. 4,790 ± 393.6 of DP-enriched Xn-Treg vs. DP-depleted Xn-Treg), and there was an increasing trend of ICOS expression when compared with total Xn-Treg (Figure 1D). The enhanced ICOS was also associated with HLA-DR expression (Supplemental Figure 3B). Although not significant, higher expression levels of the Treg activation marker CTLA4 (24), and the transcription factors, Helios and FOXP3, were seen in HLA-DR+CD27+ DP-enriched Xn-Treg, when compared with DP-depleted Xn-Treg or total Xn-Treg (Figure 1D). Furthermore, HLA-DR+CD27+ DP-enriched Xn-Treg had significantly increased FOXP3 expression when compared with either HLA-DR–CD27– or HLA-DR+CD27– Xn-Treg subsets, indicating enhanced FOXP3 expression tracked with CD27 expression (Supplemental Figure 3C).

Next, we explored the expression of CD27 and HLA-DR within Xn-Treg across different stages of expansion. For the first 2 rounds of stimulation, the proportion of CD27+ Xn-Treg was decreased slightly when compared with Fresh-Treg. However, by round 3 of stimulation, the proportion of CD27+ Xn-Treg was increased (compared with round 2) (Figure 2A). By contrast, the proportion of HLA-DR+ Xn-Treg increased incrementally after each round of stimulation (Figure 2A). As a consequence, coexpression of CD27 and HLA-DR increased from round 1 to round 3 of stimulation, while the other 3 subsets either decreased (HLA-DR–CD27+ subset) or remained unchanged (HLA-DR+CD27– and HLA-DR–CD27– subsets) (Figure 2B). Consistent with this observation, CD27+HLA-DR+ DP-enriched Xn-Treg had a higher proportion of FOXP3+Helios+ cells than total Xn-Treg or Fresh-Treg, and this higher proportion of FOXP3+Helios+ cells was observed across different expansion stages, suggesting stability of the HLA-DR+CD27+ DP-enriched Xn-Treg subset (Figure 2C). Collectively, these findings suggested that 3 rounds of stimulation and expansion in the presence of antigen resulted in a highly stable HLA-DR+CD27+ DP-enriched Xn-Treg subset.

Figure 2

HLA-DR and/or CD27 expression within Xn-Tregs at different stimulation times and their FOXP3 and Helios expression. (A) The proportion of Xn-Tregs expressing HLA-DR or CD27 after gating on CD4+ cells after round (week) 1, 2, and 3 of stimulation. (B) The proportions of HLA-DR+CD27–, HLA-DR+CD27+, HLA-DR–CD27+, HLA-DR–CD27– subsets within Xn-Tregs (after gating on CD4+ cells) following round 1, 2, and 3 of stimulation. The numbers in the corners represent the percentage of cells in each quadrant. (C) The representative flow cytometric plots and the percentage of FOXP3+Helios+ cells on HLA-DR+CD27+ DP-enriched Xn-Tregs following rounds 1, 2, and 3 of stimulation (after gating on HLA-DR+CD27+ cells) and the proportion of FOXP3+Helios+ cells in Fresh-Tregs, total Xn-Tregs, and HLA-DR+CD27+ DP-enriched Xn-Tregs. Data represent 3 independent experiments with Treg from 5 individual donors. Error bars indicate the mean ± SD. One-way ANOVA: *P < 0.05, **P < 0.01, and ***P < 0.001.

The HLA-DR+CD27+ DP-enriched Xn-Treg subset is more suppressive and xenoantigen specific. To test this hypothesis, we undertook cell sorting to isolate HLA-DR+CD27+ Tregs from Xn-Tregs. Xn-Tregs were sorted into HLA-DR+CD27+ DP-enriched Xn-Treg and DP-depleted Xn-Treg subsets after undergoing a series of sequential cell gating (Supplemental Figure 4). Treg suppressive capacity was then assessed by MLR. Effector cells were stimulated with irradiated xenogeneic or allogeneic PBMCs or with polyclonally anti-CD3/CD28 dynabeads for xenogeneic, allogeneic, or polyclonal MLR assay, respectively. Tregs were added to the assay at predetermined suppressor-to-effector ratios to determine their efficacy at suppressing the respective MLRs. All Treg subsets tested, including Pc-Treg, unsorted Xn-Treg, DP-depleted Xn-Treg, and HLA-DR+CD27+ DP-enriched Xn-Treg, showed similar potency in inhibition of polyclonal or allostimulated MLRs in a Treg number-dependent manner (Figure 3A). However, consistent with our previous study (18), unsorted Xn-Tregs showed stronger suppressive capacity in the xenostimulated (Xeno) MLR than Pc-Tregs at lower Treg/responder ratios of 1:16 through 1:256, and this stronger potency in suppressing the xenogeneic but not polyclonal or allogeneic response was further enhanced by replacing unsorted Xn-Tregs with the sorted HLA-DR+CD27+ DP-enriched Xn-Treg subset in the Xeno MLR, showing that even at the lowest Treg/responder ratio tested (1:256), a 43.5% suppression of xenoreactive cell proliferation was still detected, which was not seen with other Treg subsets (43.5% vs. 1.97% vs. 15.1% vs. 6% of suppression by HLA-DR+CD27+ DP-enriched Xn-Treg vs. Pc-Treg vs. total Xn-Treg vs. DP-depleted Xn-Treg) (Figure 3A). The higher suppressive potency and xenoantigen stimulation-dependent suppression by HLA-DR+CD27+ DP-enriched Xn-Tregs was verified by depletion of HLA-DR+CD27+ DP cells from Xn-Tregs resulting in impaired Treg suppressive capacity and xenoantigen specificity as assessed by the Xeno MLR (Figure 3A). Using the same protocol of Xn-Treg expansion, alloantigen-expanded Tregs (Al-Treg) were generated from CD4+CD25+CD127–/lo Tregs isolated from PBMCs stimulated with irradiated alloantigen PBMCs for 3 rounds. HLA-DR+CD27+ DP-enriched Al-Treg was equally suppressive of the Xeno MLR as HLA-DR+CD27+ DP-depleted Al-Treg or total Al-Treg, thereby suggesting that the antigen specificity and enhanced suppression were features of their xenoantigen stimulation and not solely the result of HLA-DR+ and CD27 coexpression (Figure 3B).

Figure 3

In vitro suppression assay of HLA-DR+CD27+ DP-enriched Xn-Tregs. (A) MLR assay of Xn-Treg suppressive capacity compared with Pc-Treg. Carboxyfluorescein diacetate succinimidyl ester–labeled (CFSE-labeled) autologous human PBMCs (CD4+CD25+CD127–/lo depleted) were stimulated with irradiated xenogeneic porcine (Xn-MLR) or allogeneic (Allo MLR) PBMCs or anti-CD3/CD28 dynabeads (Poly MLR), in the presence or absence of serial dilutions of unsorted Xn-Treg or HLA-DR+CD27+ DP-enriched Xn-Treg or HLA-DR+CD27+ DP-depleted Xn-Treg or anti-CD3/CD28 dynabead–expanded Pc-Treg for 7 days, prior to measurement of PBMC proliferation by CFSE dilution. (B) Alloantigen-expanded Treg (Al-Treg) suppressive capacity in Xn-MLR. CFSE-labeled autologous human PBMCs were stimulated with irradiated xenogeneic porcine PBMCs, in the presence or absence of serial dilutions of unsorted Al-Treg or HLA-DR+CD27+ DP-enriched Al-Treg or HLA-DR+CD27+ DP-depleted Al-Treg or Pc-Treg for 7 days, prior to measurement of PBMC proliferation by CFSE dilution. (C) Assessment of IFN-γ concentration in supernatants of the Xn-MLR assay of Xn-Treg suppression. IFN-γ secretion in supernatants collected from xenogeneic MLR assay as described in A was measured by ELISA. Data are presented as mean ± SD of 4 independent experiments with Tregs from 4 individual donors (A) except in Poly MLR with 6 individual donors for Xn-Treg, DP-depleted Xn-Treg, and DP-enriched Xn-Treg, from 3 individual donors (C) and 6 independent experiments with Treg from 6 individual donors (B). Two-way ANOVA: *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

This finding was consistent with ICOS expression in the different expanded Treg populations. ICOS expression in HLA-DR+CD27+ DP-enriched Xn-Treg was significantly higher than that seen in DP-depleted Xn-Treg (P < 0.05) (Figure 1D), whereas there was no significant difference in ICOS expression in HLA-DR+CD27+ DP-enriched Pc-Treg (2,248 ± 772.9), HLA-DR+CD27+ DP-depleted Pc-Treg (1,686.4 ± 288.8), and total Pc-Treg (1,960.8 ± 551.2) (Supplemental Figure 5). This suggests that enhanced ICOS expression in HLA-DR+CD27+ DP-enriched Xn-Treg was a feature of xenoantigen simulation.

In addition to their higher capacity to specifically suppress the proliferating xenoreactive effector cells, HLA-DR+CD27+ DP-enriched Xn-Tregs were also more capable of suppressing IFN-γ secretion in the Xeno MLR cultures, where the biggest reduction in IFN-γ secretion was in the presence of HLA-DR+CD27+ DP-enriched Xn-Tregs even at the lower Treg/responder ratios of 1:16 through 1:256 (Figure 3C). Together, these results demonstrated that the HLA-DR+CD27+ DP-enriched Xn-Treg subset led to increased suppression of the xenogeneic response, including xenoreactive cell proliferation and effector cytokine secretion, and this suppression was antigen specific. Alternatively, depletion of HLA-DR+CD27+ cells from Xn-Tregs impaired their suppressive capacity and antigen specificity in vitro.

HLA-DR+CD27+ DP-enriched Xn-Tregs are functionally stable in vitro. Since stable FOXP3 expression reflected by demethylation of TSDR within the FOXP3 gene is a prerequisite for the suppressive function, the functional stability of the HLA-DR+CD27+ DP-enriched Xn-Treg subset was evaluated by TSDR assay. The results showed that HLA-DR+CD27+ DP-enriched Xn-Treg retained a demethylation phenotype with no significant difference when compared to Fresh-Treg (10.6% ± 4.9% vs. 14.4% ± 3.7% of methylation in HLA-DR+CD27+ DP-enriched Xn-Treg vs. Fresh-Treg) and were also less methylated within their FOXP3 gene than total Xn-Treg and DP-depleted Xn-Treg counterparts, (10.6% ± 4.9% vs. 17.1% ± 6.1% vs. 24.9% ± 11.9% of methylation in HLA-DR+CD27+ DP-enriched Xn-Treg vs. total Xn-Treg vs. DP-depleted Xn-Treg) (Figure 4A). This verifies that the demethylation of the FOXP3 gene in expanded HLA-DR+CD27+ DP-enriched Xn-Tregs was similar to that of fresh, naive/rested, unexpanded Treg. HLA-DR+CD27+ DP-enriched Xn-Tregs are functionally stable and unlikely to revert to T effector cells. The functional stability of HLA-DR+CD27+ DP-enriched Xn-Tregs was further assessed under pro-inflammatory conditions to test their plasticity toward an effector Th17 or Th1 cell phenotype. Tregs were stimulated with a combination of pro-inflammatory cytokines for 6 days prior to detecting the proportion of CD4+FOXP3+ Tregs coexpressing IL-17 or IFN-γ. After stimulation, no significant change in proportion of IL-17–coexpressing cells was observed within both HLA-DR+CD27+ DP-enriched Xn-Tregs and Pc-Tregs (Figure 4B and Supplemental Figure 6). In contrast, a considerably increased proportion of IL-17–coexpressing cells was detected after inflammatory stimulation within total Xn-Tregs or DP-depleted Xn-Tregs, showing a significant difference from that seen within HLA-DR+CD27+DP-enriched Xn-Tregs (Figure 4B). Moreover, while all other Treg subsets tested demonstrated substantially increased IFN-γ–coexpressing cells upon inflammatory stimulation, HLA-DR+CD27+ DP-enriched Xn-Tregs showed a significantly reduced response to the stimulation, with a slight increase in IFN-γ–coexpressing cells (Figure 4B and Supplemental Figure 6). Taken together, these results indicated that HLA-DR+CD27+ DP-enriched Xn-Tregs were functionally stable in association with the expression of both HLA-DR and CD27 activation/memory markers.

Figure 4

Evaluation of Treg functional stability. (A) TSDR assay. The stability of Treg master function marker FOXP3 was evaluated by measurement of the status of demethylation of TSDR within FOXP3 in all Treg subsets examined. Data are mean ± SD of independent experiments with Tregs from 5 individual donors. (B) Test of Treg plasticity and stability under pro-inflammatory conditions. The multiple types of Tregs were stimulated with a combination of pro-inflammatory cytokines (IL1β, IL6, IL21, IL23, TGF-β) and IL2 for 6 days (Stimulated) (details in Methods) prior to flow cytometric analysis of proportions of these cultured Tregs coexpressing IL17 or IFN-γ. Control samples were Treg subsets with IL2 only (nonstimulated). Data presented are mean ± SD of 5 independent experiments with Tregs from 5 individual donors. Paired t test comparison (2 tailed) between DP-enriched Xn-Treg and DP-depleted Xn-Treg or Xn-Treg (A) and 2-way ANOVA (B): *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

HLA-DR+CD27+ DP-enriched Xn-Tregs were more capable of suppressing islet xenograft rejection. To study HLA-DR+CD27+ DP-enriched Xn-Treg function in vivo, NOD/SCID IL-2 receptor γ–/– (NSG) mice were transplanted with NICC xenografts and, 3 days after transplantation, reconstituted with or wit