Expansion of a restricted residual host Treg-cell repertoire is dependent on IL-2 following experimental autologous hematopoietic stem transplantation

Authors

  • Allison L. Bayer,

    1. Department of Microbiology/Immunology, University of Miami Miller, School of Medicine Miami, FL, USA
    2. Diabetes Research Institute, University of Miami Miller, School of Medicine Miami, FL, USA
    Search for more papers by this author
  • Jackeline Chirinos,

    1. Department of Microbiology/Immunology, University of Miami Miller, School of Medicine Miami, FL, USA
    Search for more papers by this author
  • Cecilia Cabello,

    1. Diabetes Research Institute, University of Miami Miller, School of Medicine Miami, FL, USA
    Search for more papers by this author
  • Jing Yang,

    1. Department of Microbiology/Immunology, University of Miami Miller, School of Medicine Miami, FL, USA
    Search for more papers by this author
  • Takaji Matsutani,

    1. Department of Microbiology/Immunology, University of Miami Miller, School of Medicine Miami, FL, USA
    Current affiliation:
    1. Laboratory of Immune Regulation, Wakayama Medical University, 105 Saito Bio Innovation Center 7-7-20, Saito-Asagi, Ibaraki-City, Osaka, 567-0085, Japan
    Search for more papers by this author
  • Thomas R. Malek,

    1. Department of Microbiology/Immunology, University of Miami Miller, School of Medicine Miami, FL, USA
    2. Diabetes Research Institute, University of Miami Miller, School of Medicine Miami, FL, USA
    Search for more papers by this author
  • Robert B. Levy

    Corresponding author
    1. Department of Microbiology/Immunology, University of Miami Miller, School of Medicine Miami, FL, USA
    • Department of Microbiology and Immunology, PO Box 016960 (R-138); University of Miami Miller School of Medicine; Miami, FL 33101, USAFax: +1–305–243–5522
    Search for more papers by this author

Abstract

We previously identified a population of residual Treg cells following autologous hematopoietic stem transplantation (HSCT), that rapidly undergoes significant expansion in lymphopenic transplant recipients prior to repopulation by donor de novo derived Treg cells. These CD4+Foxp3+ T cells provide protection from the development of autoimmune disease. Although ablative conditioning results in excess IL-7 and IL-15, IL-2 is typically not found at high levels following autologous HSCT. We therefore examined the role of these three STAT-5 signaling cytokines in the expansion of residual Treg cells after autologous HSCT. The present study found that the residual Treg cell population included surviving peripheral host Foxp3+CD4+ T cells whose expansion was critically dependent on IL-2, which could be solely provided by surviving host cells. IL-7 was found to contribute to Treg cell homeostasis, however, not as a growth factor but rather for their persistence. In conjunction with this expansion, TCR spectratype analyses revealed that the residual host Treg-cell compartment differed from that present in non-conditioned healthy mice since the residual host Treg cells exhibit a limited TCR diversity. Collectively, these data indicate that the proliferation of Treg and T effector (Teff) cells post-HSCT utilize separate pools of cytokines which has important implications regarding the development of clinical strategies to elicit the desired immune responses in patients post-transplant.

Introduction

It is well established that one major mechanism to suppress self-reacting lymphocytes is the production of a thymically-derived population of naturally occurring regulatory T (Treg) cells. We and others have shown a surviving population of residual host Treg cells following even ablative levels of total body irradiation (TBI) 1, 2. Although the numbers of these cells are initially low, residual host Treg cells exhibit high-proliferative activity so, together with slow repopulation by thymically-derived CD4+ Treg cells, residual Treg cells comprise the majority of the CD4+Foxp3+ compartment in both syngeneic or T-cell-depleted (TCD) allogeneic hematopoietic stem cell transplant (HSCT) recipients 1–2 months post-transplant 1. Several laboratories have demonstrated the functional capacity of both in situ irradiated and residual host Treg cells in vitro and in vivo 1–6. Notably, the absence of these Treg cells was shown to result in the development of autoreactive responses in HSCT recipients 1, 2. Based on these findings, we needed to determine the origin of these cells and what cytokines were required post-HSCT for the proliferation of host Treg cells and the level of diversity in the population with regard to their T-cell receptors (TCRs).

STAT-5 signaling is required for the activation and function of Treg cells 7–10. Interestingly, γc−/− mice have a more striking deficiency than IL-2−/− or IL-2R−/− mice with nearly complete thymic and peripheral absence of Foxp3+ Treg cells 8, 11, 12, consistent with the notion that, in addition to the IL-2/IL-2R interaction, other γc-cytokines must be critically important in development of CD4+Foxp3+ Treg cells. Interestingly, although IL-7- and IL-7Rα-deficient mice have severely reduced T and B cell numbers, we found apart from IL-2, IL-7/IL-7R interactions also contribute to Treg-cell development and peripheral homeostasis and that the lack of Foxp3+ cells in γc-knockout mice results from the absence of IL-2/IL-2R interaction in combination with defective IL-7R signaling 12. Evidence suggests that IL-15 can act as a Treg-cell growth factor and improve their functional capacity in vitro and in vivo 13, 14. Notably, ablative conditioning elevates IL-7 and IL-15 levels 15–17. Moreover, since IL-7 is crucial for development of both T and B lymphoid compartments, strategies employing the administration of IL-7 post-HSCT to augment thymic and peripheral immune reconstitution are being examined 18–20.

In total, the results here demonstrate that the residual host Treg-cell compartment comprises surviving peripherally derived Treg cells and the overall numbers of host Treg cells are influenced by the thymus. Importantly, IL-2 signaling is the critical pathway controlling the number of residual Treg cells during lymphopenia following HSCT. IL-2 production from the host alone was sufficient to drive this expansion. Interestingly, Vβ spectratype analyses indicated that the residual Treg cell repertoire present post-HSCT is not the same as that present pre-conditioning, as it becomes clonally restricted. Thus, IL-2 is required to drive expansion of this functional residual Treg cell population, capable of regulating autoreactivity in transplant recipients during the period of de novo natural Treg-cell generation and immune reconstitution 1, 2.

Results

The thymus is not essential for the presence of residual host Treg cells post-HSCT

To investigate the origin of host Treg cells present during the first month post-HCT, two approaches were utilized. First, C57BL/6C(B6)-Foxp3RFP recipients were transplanted with B6-Foxp3GFP TCD bone marrow (TCD-BM) and, second, thymectomized B6 mice were subjected to 900 rad TBI and given Thy1.1-B6 TCD-BM. Notably, all Treg cells present in the thymus and LNs were of host origin for the first two wk with initial de novo donor Foxp3GFP cells detected thereafter (Fig. 1A and B). As expected, the overall peripheral cell numbers in thymectomized recipients was lower versus controls 3–4 wk post-HCT, which was accompanied by lower percentage of CD4+ T cells and higher proportion of Foxp3-expressing CD4+ T cells (Supporting Information Fig. 1A and B). As previously reported, most splenic Treg cells (>95%) present 3.5 wk post-HCST were recipient (Fig. 1C) in origin 1. Importantly, peripheral host (Thy1.2+) Treg cells were present in control as well as thymectomized recipients although their overall numbers in the latter were reduced (Fig. 1D). Experiments were also performed (Supporting Information Fig. 2) exploiting Rag2p-GFP mice enabling detection of lymphocytes that have recently expressed Rag genes, identifying recent thymic emigrants (RTEs) 21. Three to four wk post-transplant, results did not support the presence of recent thymic host-derived CD25+CD4+ T-cell emigrants (Supporting Information Fig. 2C). In total, these findings indicate that residual Treg cells derived from the peripheral pool survive post-HSCT and that, although not required for their presence, the thymus clearly influences the total numbers of host Treg cells post-HCST.

Figure 1.

A residual peripheral Treg cell population exists in thymectomized HSCT recipients. (A–B) B6-FoxP3RFP reporter mice were conditioned with 9.0 Gy TBI on day 1 and transplanted with 5×106 TCD-BM cells from B6-FoxP3GFP donors. Mice were sacrificed 1, 2, or 3 wks post-HSCT and the thymus and LNs analyzed by flow cytometry. (A) Representative dot plots showing Foxp3GFP (donor BM) and Foxp3RFP (host) expression of gated single positive (SP) or CD4+ T cells from the thymus and LN, respectively. (B) Total numbers of residual host (Foxp3RFP) or BM-derived (Foxp3GFP) Treg cells from thymus (top) or LNs (bottom). Data represent mean+SEM for three mice per time point. (C–D) Adult B6 mice were thymectomized and 1 wk later subjected to 900 rad TBI and on the following day given 5×106 TCD-BM from congenic Thy1.1 B6 mice. Control mice were not thymectomized. All mice were sacrificed 3.5 wk following TBI and cells from the Spleen (SPL), LNs analyzed by flow cytometry. (C) Representative dot plots showing CD25 and Foxp3 staining of gated CD4+ T cells with corresponding dot plot showing Thy1.2 (host) and Thy1.1 (BM-derived) staining of gated CD4+Foxp3+ T cells from the Spleen of control or thymectomized mice. (D) Total numbers of residual host (Thy1.2) Treg cells in the SPL and LNs. Data represent means for three or seven mice per time point in this experiment. *p<0.0001, +P<0.001, one-way ANOVA followed by Newman–Keuls Multiple Comparison Test.

Host residual Treg cells exhibit a skewed TCR Vβ immunoscope profile

To examine host Treg cells' TCR present one month post-HSCT, RNA was initially prepared from CD4+CD25+ T cells harvested from 9.0 Gy B6 mice transplanted with TCD-BM from B6-CD4−/− donors (Fig. 2). CDR3 size distribution analysis after CD4+CD25+ enrichment from six individual recipients (three representative recipients are shown) indicated that virtually each Vβ family exhibited a non-Gaussian distribution compared with Treg-cell TCR from control, non-transplanted B6 mice (ctrl) (Fig. 2A and C, ctrl#1 and Exp#1A,B,C). To focus analysis only on Foxp3 expressing Treg cells, B6-Foxp3RFP mice were transplanted with B6-CD4−/− or B6-Foxp3GFP donor TCD-BM (Fig. 2B). Similar to results observed using enriched CD4+CD25+ Treg-cell preparations, host (i.e. RFP+) Treg cells demonstrated a narrowing of the TCR Vβ repertoire (Fig. 2C). To quantify differences, diversity (D) scores (see Materials and Methods and 22) for the residual Treg-cell population Vβ TCRs were determined and found to be higher in recipient mice of CD4−/− BM cells than Treg cells from control mice (Fig. 2C and D). When mice received Foxp3GFP BM cells, the residual Treg cell Vβ TCR also displayed a narrowing of the repertoire (Fig. 2C and D). Individual Vβ score averages, 13.6, 34.0, and 40.1 for control mice, CD4−/− and Foxp3GFP BM recipients, respectively, illustrated that the D score for residual Treg cells was higher (p<0.05) than the control Treg cells (Fig. 2D). Reconstitution is delayed post-HSCT with emergence of newly produced Treg cells from donor BM beginning at 3 wks. This “competition” from newly derived Foxp3GFP Treg cells did not impact the diminished diversity of the residual Treg-cell population, i.e. no statistical difference between this group of marrow recipients and mice that received CD4−/− BM cells (Fig. 2D), suggesting that reshaping of the residual Treg-cell repertoire occurred early after ablation and transplant. These results are consistent with the notion that clonal expansion of a subset of residual host Treg cells led to a narrowing of their TCR repertoire.

Figure 2.

Spectratype skewing of different Vβ families by the TCR expressed within the residual Treg cell pool. B6-WT or B6-Foxp3RFP reporter mice were conditioned with 9.0 Gy TBI on day −1 and transplanted with 3×106 TCD-BM cells from B6-CD4−/− or B6-Foxp3GFP donors. Recipients were sacrificed 28–30 days post-HSCT, and the Treg cells isolated for Vβ TCR spectratype analysis via (A) MACS CD25 enrichment, or (B) by high-speed cytometric sorting. Values in the representative dot plots generated prior to and post-enrichment represent the percentage positive cells in the indicated quadrants. (C) Rows 1 (Ctrl#1) and 5 (Ctrl #2) exhibit anticipated Gaussian distribution of the Vβ families examined from CD4+CD25+ Treg cells and CD4+RFP+ Treg cells isolated from non-transplanted B6-WT mice (Row 1) and B6-Foxp3 RFP KI (Row 5). Rows 2, 3, 4 represent the spectratypic analyses of B6-CD4+ CD25+ residual T cells isolated from B6-WT recipients of B6-CD4−/− donor BMCs. Rows 6, 7, and 8 represent the spectratypic analyses of CD4+RFP+ Treg cells isolated from recipients of B6-CD4−/− donor BM (row 6) and B6-Foxp3GFP donor BM (rows 7,8). Skewing is evident (D score value indicated in upper left corner of individual histograms) in residual Treg cells obtained from all transplant recipients regardless of whether CD4+CD25+ or CD4+RFP+ Treg cells were analyzed. (D) Average D score (see Materials and Methods) for control mice (untransplanted normal B6 and B6-Foxp3RFP) and B6-CD4−/− and Foxp3GFP BM recipients. *p<0.05, one-way ANOVA followed by Newman–Keuls Multiple Comparison Test. Data are representative of three independent experiments.

IL-7 supports persistence of residual host Treg cells, but contributes minimally to their expansion post-HSCT

Since IL-7 is present at markedly elevated levels following conditioning, we investigated whether this cytokine contributed to residual Treg-cell expansion following ablation and HSCT. In contrast to IL-7Rα−/− mice that possess a severe block in T-cell development, resulting in lymphopenia, thymic-directed IL-7Rα (IL-7RαTg+) expression promotes normal thymic T-cell development when expressed in IL-7Rα−/− mice 23. When IL-7RαTg+ mice (Thy1.2+) were transplanted with congenic B6-Thy1.1+ TCD-BM, the thymus was fully reconstituted with Thy1.1+ thymocytes 4 wk post-HSCT (Fig. 3A). The CD4+Foxp3+ peripheral compartment was composed of both residual host IL-7RαTg+ (Thy1.2+) and BM-derived (Thy1.1+) cells (Fig. 3A and B). Although, there was an overall slight reduction in the percentage of residual Treg cells (Thy1.2+, SPL 53.9±10.7, LNs 53.6±9.5, and peripheral blood (PB) 46.2±10.0) in IL-7RαTg+ mice compared with residual Treg cells in B6-WT mice (SPL 78.9±1.1, LNs 67.6±1.5, and PB 70.0±3.5), this reduction did not reach statistical difference (Fig. 3C). However, there was a statistically reduced number of residual Treg cells in the SPL and LNs of IL-7RαTg+ mice compared with B6-WT mice (Fig. 3E). The lack of fully functional IL-7Rα signaling did not impact proliferation by residual Treg cells as the level of Ki67 staining was similar between residual and BM-derived Treg cells in SPL, LN and PB and was comparable to that in WT mice (Fig. 3D), suggesting that IL-7 may be acting as a survival rather than a growth factor. Notably, we also observed that administration of an anti-IL-2 mAb/IL-2 complex to 9.0 Gy conditioned B6-WT and B6-IL7RαTg+ mice induced marked activation and expansion of residual Treg cells despite the IL-7R-signaling deficit (Supporting Information Fig. 3). Together, these data indicate that IL-7 does not largely contribute to the homeostatic proliferation of the residual host Treg-cell population following ablation.

Figure 3.

Signaling through IL-7Rα does not drive the proliferation of the residual host Treg cells. (A–B) B6-IL7RαTg+ (Thy1.2+) mice were subjected to 9.0 Gy TBI and on the following day transplanted with 10×106 TCD-BM cells isolated from congenic B6-Thy1.1+ mice. Mice were sacrificed 4 wk post-HSCT and the SPL, LN, and peripheral blood (PB) analyzed by flow cytometry. (A) Representative histogram showing Thy1.1 (donor BM) staining from the thymus. Representative dot plot showing CD4 and CD8 staining with the corresponding dot plot showing CD25 and Foxp3 staining of gated CD4+ population and dot plot for Thy1.1 and Thy1.2 (residual host) staining of gated CD4+Foxp3+ cells from the SPL. (B) Representative histogram showing Ki67 staining of gated Thy1.1+CD4+Foxp3+ BM-derived cells or Thy1.2+CD4+Foxp3+ residual host cells. (C-E) WT B6-CD45.1+ congenic mice were subjected to 9.0 Gy TBI and on the following day transplanted with TCD-BM isolated from B6-Thy1.1+ congenic mice. All mice were sacrificed 4 wk post-HSCT and the SPL and LN, and PB analyzed by flow cytometry. Percentage of (C) residual host or BM-derived CD4+Foxp3+ cells and (D) Ki67+CD4+Foxp3+ cells from the SPL, LN and PB from B6-IL7RαTg+ (left) and WT B6-CD45.1+ (right). (E) The numbers of residual Treg cells per million splenic or LN cells are shown. *p<0.05, unpaired t-test. (A, B) Data are representative of 5 mice per group. (C–E) Data represent mean+SEM for 3–5 mice. Data are representative of two independent experiments.

Signaling through the IL-2Rβ chain plays a major role in expanding residual host Treg cells post-HSCT

IL-2Rβ−/−mice have severely reduced numbers of functional Treg cells and succumb to lethal autoimmunity early in life; however, thymic-directed IL-2Rβ (IL-2RβTg+) restores Treg-cell production and prevents autoimmunity when expressed in IL-2Rβ−/− mice 24. These mice were transplanted with B6-Thy1.1+ TCD-BM to investigate IL-2R in residual host Treg-cell expansion. The thymus was reconstituted with Thy1.1+ BM-derived cells 4 wk post-HSCT (Fig. 4A). In contrast with WT mice, the peripheral Treg-cell compartment was largely composed of BM-derived cells with only 24–26% of CD4+Foxp3+ cells derived from the residual IL-2Rβ−/− Treg cells in the SPL, LN, or PB (Fig. 4A and B). Only a small percentage of these IL-2Rβ−/− residual Treg cells were positive for Ki67 staining, while the majority of the BM-derived cells were proliferating as assessed by Ki67 staining (Fig. 4C and D). This low proliferation is consistent with our previous work with these mice showing that these Treg cells that lack fully functional IL-2R signaling are characteristically distinct from those in their WT-counterparts, including having a lower proliferation rate and a longer life-span 25. To test the possibility that this small residual Treg-cell population survived due to signaling through the IL-7Rα chain, IL-2RβTg+ mice were transplanted with B6-Thy1.1+ TCD-BM and administered anti-IL-7Rα mAb. Blockade of IL-7Rα did not further diminish the percentage or numbers of IL-2Rβ−/− residual Treg cells or their proliferation (p>0.05, one-way ANOVA) when compared with control mice treated with IgG (Supporting Information Fig. 4A and B). We also examined a role of IL-7 using the double KOL (DKO)−2Rβ/7Rα mice that support Treg-cell development but contain peripheral Treg cells with impaired IL-2 and IL-7 receptors (2RβTG7RαTGDKO2Rβ/7Rα). When these 2RβTG7RαTGDKO2Rβ/7Rα mice (Thy1.2+) were transplanted with B6-Thy1.1 TCD-BMC, the CD4+Foxp3+ compartment was largely composed of the Thy1.1+ BM-derived Treg cells in the SPL, LN and PB (Supporting Information Fig. 5). Only few residual Thy1.2+ CD4+Foxp3+ host cells were Ki67-positive, while the normal BM-derived Treg cells were actively proliferating (Supporting Information Fig. 5). Therefore, the behavior of these residual Treg cells from 2RβTG7RαTGDKO2Rβ/7Rα mice was very similar to those in the IL-2RβTg+ mice (Fig. 4). Moreover, when IL-7RαTg mice were transplanted followed by anti-IL-2 treatment for 4 wk, the residual Treg-cell population was considerably reduced versus untreated transplant controls (p<0.01, unpaired t-test) (Supporting Information Fig. 6). Together, these data indicate that signaling through the IL-2Rβ chain largely drives the homeostasis of the residual Treg cells.

Figure 4.

Proliferation of residual host Treg cells is predominantly driven by signaling through the IL-2Rβ chain. B6-IL-2RβTg+ (Thy1.2+) mice were subjected to 9.0 Gy TBI and on the following day transplanted with 10×106 TCD-BM cells isolated from congenic B6-Thy1.1+ mice. Mice were sacrificed 4 wk post-HSCT and the thymus, SPL, LN, and PB analyzed by flow cytometry. (A) Representative histogram showing Thy1.1 (donor BM) staining from the thymus (upper left). Representative dot plot shows CD4 and CD8 staining (upper right) with the corresponding dot plot showing CD25 and Foxp3 staining of gated CD4+ population lower left) and dot plot for Thy1.1 and Thy1.2 (residual host) staining (lower right) of gated CD4+Foxp3+ cells from the SPL. (B) Percentage of residual host or BM-derived CD4+Foxp3+ cells in the SPL, LN, and PB. (C) Representative histogram showing Ki67 staining of gated Thy1.1+CD4+Foxp3+ BM-derived cells (left) or Thy1.2+CD4+Foxp3+ residual host cells (right). (D) Percentage of residual host or BM-derived Ki67+CD4+Foxp3+ cells in the SPL, LN, and PB. (A, C) Data are representative of 4 mice per group in this experiment. (B, D) Data are represented as mean+SEM of 4 mice. *p<0.001 compared with residual host Treg cells, two-way ANOVA followed by Bonferroni comparison test.

IL-15 is not required for Treg cell-compartment reconstitution following HSCT

Since the IL-2Rβ chain is part of the high affinity IL-2R and the IL-15R, these data cannot determine whether IL-2 and/or IL-15 drive residual Treg-cell expansion. Since IL-15 is also elevated post-BMT 17, its potential contribution to Treg-cell reconstitution post-HSCT was examined under different IL-15-deficient conditions. CD4+Foxp3+ LN T cell levels in both B6-IL15−/− and B6-WT recipients of B6-IL15−/− BM were comparable to those of B6-WT recipients reconstituted with B6-Thy1.1+ TCD-BM 4 wk post-HSCT (Fig. 5A and B). Splenic levels of these recipients were also not reduced versus those in B6-WT recipients. One limitation of these studies is that residual host IL-15−/− Treg cells cannot be discriminated from the B6-WT or IL-15−/− BM-derived Treg cells following reconstitution. However, the overall percentage of CD4+Foxp3+ cells was very similar compared with that in B6-WT mice that received congenic BM cells. Therefore, these observations are consistent with the notion that IL-15 is not required or likely contributory to residual host Treg cell maintenance and expansion post-HSCT.

Figure 5.

IL-15 is not required for homeostasis of residual host Treg cells following HSCT. B6-WT mice and B6-IL15−/− were conditioned with 9.0 Gy TBI on day -1 and transplanted with 3×106 TCD-BM cells from B6-IL-15−/− or B6.Thy1.1+ donors as indicated. SPL and LN cells were examined for Treg cell compartment reconstitution 28 days post-HSCT. (A) Representative dot plots from SPL and LN showing CD4 and Foxp3 expression at 4 wk post-transplant in the different recipient–donor combinations examined. (B) The percentage of CD4+ cells expressing Foxp3 in the SP and LN at 4 wk post-HSCT. Data represent mean+SEM from 2 to 3 mice in this experiment. p<0.05, IL-15−/− donor and recipient SPL value versus the other two SPL populations, one-way ANOVA followed by Newman–Keuls multiple comparison test.

Host IL-2 is the crucial cytokine for maintaining homeostasis of residual Treg cells in recipients post-HSCT

Past work showed that IL-2Rβ is expressed at nearly undetectable levels by peripheral CD4+Foxp3hi Treg cells within IL-2RβTg+, but this minimal IL-2Rβ expression lead to weak transient IL-2-dependent STAT-5 activation 25. Recent data showed that minimal IL-2Rβ signaling supports Treg-cell development and homeostasis 26. Therefore, residual Treg cells in the IL-2RβTg+ mice could result from minimal IL-2Rβ signaling. IL-2Rβ−/− mice contain a population of immature CD4+CD25negFoxp3low T cells due to IL-7Rα signaling 25. To test the possibility that the minimal IL-2Rβ signaling is maintaining the residual Treg-cell population in the IL-2RβTg+ mice, IL-2Rβ−/− mice rendered autoimmune-free by adoptive transfer of congenic B6-CD45.1+ CD4+ T cells were transplanted with congenic B6-Thy1.1+ TCD-BM. The thymus of these mice were fully reconstituted with B6-Thy1.1+ BM-derived cells 4 wk post-HSCT (Fig. 6A). In sharp contrast to the IL-2RβTg+ mice (Fig. 4), there are very few residual IL-2Rβ−/− Treg cells (Fig. 6A and B). Although this population of residual host cells is quite small, these cells maintain the immature CD25negFoxp3low phenotype (Fig. 6C), suggesting that this minimal peripheral IL-2Rβ signaling in IL-2RβTg+ mice may be supporting residual Treg cells. Nevertheless, these residual immature Treg cells are fully capable of responding to IL-7 and there are very few of these cells present 4 wk post-HSCT, suggesting IL-2 is the dominant cytokine supporting residual Treg-cell homeostasis. Thymic-restricted expression of IL-7RαTg+ in DKO2Rβ/7Rα mice (7RαTGDKO2Rβ/7Rα) also results in thymic production of immature CD4+CD25negFoxp3low cells similar to IL-2Rβ−/− mice 12, but they are genetically non-responsive to IL-2 and IL-15 and only minimally responsive to IL-7. These 7RαTGDKO2Rβ/7Rα mice resemble the cured IL-2Rβ−/− mice with very few residual Thy1.2+ host cells 4 wk post-HSCT in SPL, LN and PB (Supporting Information Fig. 7). Although the host Treg cells in IL-2Rβ−/− and 7RαTGDKO2Rβ/7Rα mice express this immature phenotype which could potentially impart differing homeostatic characteristics than the mature Treg-cell population, these data remains consistent with a dominant role for IL-2 in residual Treg-cell homeostasis.

Figure 6.

IL-2 is the dominant cytokine supporting homeostasis of residual host Treg cells following ablative conditioning and HSCT. 1-2 day old B6-IL-2Rβ−/− (Thy1.1) mice were adoptively transferred with 2×106 B6-CD45.1+ CD4+ T cells. These mice were subjected to 9.0 Gy TBI at 18 wk of age and on the following day transplanted with 10×106 TCD-BM cells isolated from congenic B6-Thy1.1+ WT mice. Mice were sacrificed 4 wk post-HSCT and the SPL, LN, and PB analyzed by flow cytometry. (A) Representative histogram (n=4 mice/group) showing Thy1.1 (donor BM) staining from the thymus. Representative dot plot shows CD4 and CD8 staining with the corresponding dot plot showing CD25 and Foxp3 staining of gated CD4+ population and dot plot for Thy1.1 and CD45.1 (donor Treg cells) staining of gated CD4+Foxp3+ cells from the SPL. (B) Percentage of residual host or BM-derived CD4+Foxp3+ cells in the SPL, LN, and PB. Data are presented as mean+SEM of 4 mice in this experiment. *p<0.001 compared with residual host Treg cells, 2-way ANOVA followed by Bonferroni comparison test. (C) Representative dot plot illustrating CD25 and Foxp3 expression of the gated residual host (Thy1.2+CD45.1) and BM-derived (Thy1.2CD45.1) CD4+Foxp3+ Treg cells.

To investigate whether the source of IL-2 comes from host or donor cells, congenic B6-CD45.1+ recipients underwent transplant with IL-2−/−Rag−/− TCD-BM and compared with B6-WT mice receiving B6-CD45.1+ TCD-BM. PB (2 and 4 wk) demonstrated that there is a comparable percentage of residual host CD4+Foxp3+ cells in mice receiving either WT or IL-2−/−Rag−/− BM (Fig. 7A and B), suggesting that the endogenously produced host IL-2 drives the homeostasis of residual Treg cells. There was also a similar population of residual Treg cells when IL-15−/− mice were given IL-2−/−Rag2−/− BM (Fig. 7A and B), supporting that IL-2 is crucial to expand Treg cells post-HSCT.

Figure 7.

Host-produced IL-2 is responsible for homeostasis of host residual Treg cells. B6-WT, congenic B6-CD45.1+, or B6-IL-15−/− mice were subjected to 9.0 Gy TBI and on the following day transplanted with TCD-BM cells isolated from either B6-CD45.1+ congenic or IL-2−/−Rag−/− mice. Mice were sacrificed 4 wk post-HSCT and the SPL, LN, and PB analyzed by flow cytometry. (A) Representative dot plot showing CD4 and CD8 staining with corresponding dot plot for CD25 and Foxp3 staining of gated CD4+ cells and dot plot for CD45.1 and Ki67 staining of gated CD4+Foxp3+ cells from the PB 2 and 4 wk post-HSCT. (B) Percentage of CD4+Foxp3+ cells in the SPL, LN, and PB 4 wk post-HSCT. Data shown are mean+SEM of 4 mice in this experiment.

Discussion

Treg cells present early post-HSCT have the capacity to undergo rapid expansion and comprise a major portion of the Treg cell compartment for several months post-transplant 1. Because these Treg cells are functional, for example, they can prevent expansion of autoreactive T cells and syngeneic GVHD in HSCT recipients, we concluded that this population may contribute an important regulatory role prior to donor engraftment and emergence of de novo derived Treg cells post-HSCT 1–3. Treg cells present post-HSCT also pose a potential obstacle to the induction of anti-pathogen and anti-tumor responses – important objectives in such recipients. Thus, understanding the cytokine signals regulating the Treg-cell compartment in post-transplant recipients should help to develop strategies to elicit desired immune responses in these individuals. We hypothesized that STAT-5-activating γc-cytokines may contribute to the homeostatic expansion of residual host Treg cells 1. Our investigation of IL-2, IL-7 and IL-15 indicated that IL-2 is crucial to Treg-cell expansion, but IL-7 also contributes to their overall numbers post-transplant. Residual Treg cells present were found to express a skewed and limited TCR repertoire.

To determine the origin of the residual Treg cells in these studies, thymectomized recipients were examined (Figs. 1 and Supporting Information Fig. 1). Since host Treg cells were present even in thymectomized recipients, we conclude that these residual cells were derived from peripheral Treg cells which survived conditioning and transplant. Notably, reduced numbers of host Treg cells were observed in thymectomized and transplanted recipients indicating thymic influence on the host Treg-cell compartment. Since the overall CD4+ T-cell numbers are dramatically reduced in thymectomized recipients (Supporting Information Fig. 1B), the level of IL-2 available to maintain Treg cells is likely diminished which could account for the reduced overall numbers in these recipients. Notably, a prior study reported that thymic progenitors can survive in lethally-conditioned recipients and can give rise to host T cells 27. It is not possible to discriminate the contribution of residual surviving peripheral Treg cells versus that from de novo derived host Treg cells amongst the increased numbers of these cells in non-thymectomized recipients (Fig. 1D). While we cannot discount such a pathway may contribute to the host Treg-cell compartment post-HCST, our kinetic analyses failed to detect any de novo derived Treg cells from healthy, transplanted donor progenitor cells (Fig. 1A–C) during the first few wk after transplant. Moreover, we observed no increase in host peripheral Treg cells at two wk despite a dramatic increase in the host thymic Treg cells (Fig. 1B), suggesting very little contribution from surviving thymic host progenitors to the overall peripheral host Treg-cell pool. Hence, we currently favor the IL-2-based explanation to account for our findings. In support of this interpretation, the addition of the anti-IL-2/IL-2 complex post-HCST markedly expanded host Treg-cell numbers (Supporting Information Fig. 3).

IL-2 is known to be crucial for peripheral homeostasis of Treg cells and essential to maintain them during steady-state conditions 12. The present study is consistent with this role, demonstrating that under transplant-induced lymphopenia, IL-2 is critical for residual Treg-cell expansion. These findings, however, contrast others who examined Treg cells transferred into Rag−/− recipients and reported the absence of an IL-2 requirement for expansion 28. Lymphopenic conditions in Rag−/− and post-HSCT recipients may differ and our studies examined Foxp3 expressing CD4+ T cells following exposure to TBI, in contrast to non-exposed CD25+ CD4+ T cells in those studies 28. Despite reduced production of IL-2 following ablation and HSCT 29, 30, residual Treg cells in our studies exhibited robust proliferation. This possibly reflects that unlike naive and memory T cells, which only express the IL-2Rβ and γc chains, Treg cells express CD25 resulting in high affinity IL-2 receptors which better compete for the limited IL-2 present. This is supported by the observations that selective compartmentalization of IL-2 in vivo is a large determinant of the non-redundant activities of IL-2R, and Treg cells require only low levels of IL-2R signaling to support their homeostasis 26, 31.

It has been well established that both IL-7 and IL-15 cytokines contribute important functions for the homeostatic proliferation and survival of naïve and memory T cells following lymphodepletive conditioning protocols 32–34. Interestingly, we found that IL-7R deficiency does not affect residual Treg-cell proliferation, but rather resulted in diminished numbers of recipient Treg cells, suggesting that IL-7 plays a role in the persistence of host Treg cells. In contrast, IL-15 does not appear to significantly contribute to residual Treg-cell homeostasis and, if so, its contribution is likely to be extremely small. Together our data support the notion that IL-2 is critical to Treg-cell expansion post-HSCT, while IL-7 contributes to their survival. Together with the fact that IL-7 and IL-15 are important for non-regulatory T-cell populations under lymphopenic conditions, the present findings may indicate that regarding expansion post-HSCT, Tconv and Treg cell populations do not principally compete for the same pool of cytokines, which has important implications for developing strategies to elicit desired anti-pathogen/tumor immune responses in such patients.

Interestingly, the results indicated that IL-2 derived from host cells was sufficient to drive Treg-cell expansion post-transplant (Fig. 7). In addition to IL-2, B7 and CD28 are required to maintain peripheral Treg-cell proliferation and survival under homeostatic conditions 7, 35. CD28 has been proposed to support Treg-cell survival by enhancing the production of IL-2 by conventional T cells as well as maintaining expression of CD25 on Treg cells 11, 35. Notably in patients following HSCT, conventional and regulatory CD4+ T cells responded very differently to the same homeostatic environment. Treg-cell homeostasis was modulated by the level of recovery of conventional CD4+ T cells 36. Since the primary source of IL-2 in vivo is derived from activated CD4+ T cells 37, this link between Treg-cell homeostasis and CD4+ T cells in patients may be related to the latter's production of IL-2 following ablation and HSCT. Furthermore, autoreactive Tconv cells themselves can represent a consistent IL-2 source for Treg cells in a model termed adaptive Treg-cell homeostasis which is driven by recognition of self-peptides 38. In addition to CD4+ T cells, other lymphocyte populations including CD8+, NK, NKT, and DC cells have also been implicated in producing IL-2 to support Treg cells in naïve mice 37, 39. CD8+ memory cells express increased levels of Bcl-2, survive TBI at greater levels than naïve CD8+ cells 40 and have been shown to contribute to post-transplant immune reconstitution following HSCT 41. Interestingly, memory CD8+ T cells, particularly central memory 42, are known producers of IL-2. Thus, it is possible this host population together with CD4+ T cells contributes to the pool of IL-2 present that drives expansion of the restricted residual Treg cells.

Treg cells inherently undergo significant proliferation as evidenced by our prior analysis which indicated ∼1 log greater rate of division by Treg cells versus Tconv cells in normal animals (∼30% versus 3.0%, respectively) 1. Thus, following aggressive TBI conditioning, considerable death must take place in the recipient Treg-cell compartment. The loss of large numbers of Treg-cell together with the IL-2-dependent proliferation occurring in the post-transplant period likely resulted in the finding that the expanded population exhibited a skewed T-cell receptor Vβ repertoire. How much narrowing of the residual Treg-cell repertoire occurred under our transplant conditions is not presently known. Although a narrow Treg-cell repertoire has been implicated in the development of multi-organ inflammatory disease in lymphopenic mice 43, we previously found that residual Treg cells were capable of inhibiting emergence of autoimmune disease by IL-2Rβ−/− BMT 1. Consistent with this finding, our laboratory has also found that adoptive transfer of Treg cells into IL-2Rβ−/− mice leads to selection of therapeutic Treg cells that display lower TCR diversity that prevented autoimmunity in these autoimmune-prone mice 22.

Lymphoid reconstitution post-HSCT is typically a slow process accompanied by sub-optimal immune function in patients 44. The relative contributions to this complex process by peripheral expansion of donor and recipient T cells and de novo thymic-derived T cells depends on several factors including the components of the transplant, conditioning and thymic function 45, 46. Treg cells function to suppress unwanted proliferation by regulating foreign antigen driven immune responses or responses by self reactive T cells which can lead to autoimmunity 47. Evidence suggests that regulation by Treg cells is also likely critical under lymphopenic conditions such as those associated with conditioning and stem cell transplant 48. Even when few Treg cells survive the following vigorous conditioning, it is possible to expand this population with appropriate anti-IL2/IL-2 complex (Supporting Information Fig. 2, 1, 49), a manipulation that could favor the balance of Treg over Tconv cells for suppression of autoimmune responses and hence may benefit some tolerance inducing protocols. Collectively, understanding the signals which control the Treg and Tconv cell compartments post-HSCT is necessary to develop optimal anti-pathogen/tumor therapies during periods of lymphopenia, reconstitution and vaccination in recipients post-transplant.

Materials and methods

Mice

IL-2Rβ−/− mice, thymic-targeted transgenic WT IL-2R expressed in IL-2R−/− mice (designated IL-2RTg+ in this study), and thymic-targeted WT IL-7R expressed in IL-7R−/− mice (designated IL-7RTg+ in this study) on the C57BL/6 genetic background have been previously described 12, 24, 50. These thymic-targeted transgenic mice were crossed to generate IL-2RTg+ and/or IL-7RTg+ in IL-2R−/− and IL-7Rα−/− double knockout mice (designated 2RTG 7RαTGDKO2R/7R or 7RαTGDKO2R/7R in this study) as described 12. B6.129S2-Cd4tm1Mak/J (B6-CD4−/−, H2b), B6.SJL-Ptprca Pep3b/BoyJ (B6-CD45.1+, H2b) and B6.PL-Thy1a/CyJ (B6-Thy1.1+, H2b) were initially obtained from the Jackson Laboratory (Bar Harbor, ME). Foxp3GFP and Foxp3RFP breeder mice were obtained from Dr. A. Y. Rudensky (University of Washington, Seattle, WA, USA) and Dr. R. A. Flavell (Yale University, New Haven, CT, USA), respectively, and are maintained at the University of Miami. C57BL/6 (B6-WT, H-2b) mice were obtained from Jackson Laboratories (Bar Harbor, ME, USA). Rag2p-GFP and WT Rag2p-GFP−/− mice (FVB, H-2q) were generously provided by Dr. B. Adkins (University of Miami, Miami, FL, USA). C57BL/6NTac-IL15tm1lmx (B6-IL15−/−, H2b) were purchased from Taconic Farms (Germantown, NY, USA). BM from B6-IL2−/− Rag−/− mice (H2b) was generously provided by Dr. M. Farrar (University of Minnesota, Minneapolis, MN, USA). Animal studies were carried out under approved protocol by the University of Miami Institutional Animal Care and Use Committee.

Antibodies and FACS analysis

Pacific Blue-conjugated CD4, allophycocyanin-Cy7-conjugated CD4, CD45.1, and streptavidin, PE-Cy7-conjugated CD25, FITC- or PE-Ki67, Biotin-CD45.1, allophycocyanin-conjugated Thy1.2, and V500-conjugated CD8 antibodies (BD Biosciences, San Diego, CA, USA); Pacific Blue-conjugated CD8 and PerCP-conjugated Thy1.1 (Biolegend, San Diego, CA, USA); Pacific Orange-conjugated CD8, (Invitrogen, Carlsbad, CA, USA); PE-conjugated Foxp3 and Foxp3 staining buffer set (e-bioscience, San Diego, CA). FACS analysis was performed as previously described 1 using a Becton-Dickinson LSRII and Diva® software. The total number of events collected for analysis was between 50 000 and 100 000 cells.

BM transplantation

BM (long bones) was TCD with anti-Thy1.2 mAb (HO13.4 ascites diluted 1:10) followed by 10% v/v Low-Tox M rabbit complement (Accurate Chemical & Scientific, Westbury, NY, USA) at 37°C for 45′. Levels were reduced to below detection by flow analysis (<0.2%). Recipient mice received 9.0 Gy TBI (Gammacell 40, Cs-137 source) at 35–40 cGy/min. TCD-BM was infused intravenously 24 h later and mice maintained on antibiotic water and food ad libitum. Groups of transplanted mice were treated with Rat IgG (Sigma, St. Louis, MO, USA) or anti-IL7Rα antibody (BioXcell, West Lebanon, NH, USA) at 20μg/g body weight every other day for 4 wk. Some mice were thymectomized as previously described 25 1 wk prior to TBI.

Treg-cell enrichment for Vβ analysis

In some experiments, Treg cells were obtained by enrichment of CD4+CD25+ cells. SPL and LN cell homogenates were depleted of B cells and CD8+ cells using antibodies as previously described 51. Enriched CD4+ T-cell preparations were labeled with anti-CD25 PE (clone PC61, BD Biosciences, San Diego, CA, USA) and positively selected using anti-PE MACS microbeads (Miltenyi Biotec, Auburn, CA, USA). Enrichment was determined by antibody labeling pre- and post-enrichment followed by flow cytometric analysis. Cells were stained with PE anti-CD25 (PC61) and PE-Cy5 anti-CD4 (clone RM4-5, BD Biosciences, San Diego, CA, USA), permeabilized and stained for Foxp3 expression using Foxp3 Staining Kit, FITC anti-Foxp3 (FJK-16 s) or FITC isotype control (eBioscience, San Diego, CA, USA). Samples were acquired and analyzed using an LSR-I or LSR-II (Becton Dickinson, San Jose, CA, USA).

In other experiments, residual CD4+Foxp3RFP cells were isolated by high-speed cell sorting on FACSAria (Becton Dickinson, San Jose, CA, USA). The SPL and LN cell homogenates were incubated with CD4 mAb. Residual Foxp3RFP cells were initially enriched by gating on viable (DAPI-negative), GFPCD4+ Foxp3RFP cell followed by a purity sort. Purity of CD4+Foxp3RFP cells was >99%.

Spectratyping

Primers for TCR Vβ spectratyping have been previously described 22. To quantify skewing of the TCR repertoire, the method of Gorochov was used and is represented as D scores as previously described 22. The D score represents the mean of these values for each of the eight Vβ segments for each experimental sample. A low D score represents a more Guassian distribution characteristic of a highly diverse TCR repertoire; a high D score represent a non-Guassian distribution characteristic of less TCR diversity.

Acknowledgements

The authors thank Dr. Michael Farrar (University of Minnesota) for the B6-IL2−/− Rag−/− mice and our colleague Dr. Rebecca Adkins (University of Miami) for the Rag2p-GFP and WT Rag2p-GFP−/− mice. They also thank Mr. Oliver Umland of the Diabetes Research Institution Flow Cytometry Core, the Sylvester Cancer Center Flow Cytometry facility, and Dr. Damaris Molano at the University of Miami Diabetes Research Institute for her surgical assistance in the thymectomy studies. This work was supported by NIH grants AI46689 and CA 120776 (RBL), AI055815 (TRM), and the American Diabetes Association Junior Faculty grant 07-09-JF-06 (ALB), and the Diabetes Research Institute Foundation (ALB).

Conflict of interest: The authors declare no financial or commercial conflict of interest.

Ancillary