Critical role for CCR5 in the function of donor CD4⁺CD25⁺ regulatory T cells during acute graft-versus-host disease

Acute graft-versus-host disease (GVHD) is a severe and potentially fatal complication of allogeneic bone marrow transplantation (allo-BMT). Acute GVHD is caused by mature donor T cells that recognize alloantigens presented initially by host antigen presenting cells (APCs).^1,2  Our group has demonstrated that the accumulation of donor T cells during the peritransplantation period takes place primarily in lymphoid tissues, followed by recruitment to parenchymal tissues such as the gastrointestinal (GI) tract, liver, lung, and skin.³  We and others have recently demonstrated that eliminating expression of the chemokine receptor CCR5 from donor T cells in an experimental GVHD model resulted in exacerbated GVHD and increased T-cell infiltration of the liver and lung in lethally irradiated recipient animals.^4,5  These data suggested that the primary role for CCR5 during GVHD in conditioned transplant recipients is not to direct effector-cell recruitment as originally hypothesized, but to down-modulate target organ inflammation.

Chemokines are predominantly 8-kDa to 12-kDa chemotactic proteins, which bind a family of 7-transmembrane–spanning G protein–coupled receptors, and function primarily in leukocyte migration (reviewed in Moser et al⁶ ). The chemokine receptor CCR5 is expressed on activated T helper-1/T cytotoxic-1 (T_H1/T_C1) T cells, natural killer (NK) cells, macrophages, and dendritic cells.⁷  The ligands for this receptor, CCL3, CCL4, and CCL5, are expressed at sites of inflammation during acute GVHD.^4,8-11  CCR5 may play a role in directing effector cells to sites of inflammation.^12-15  Interestingly however, multiple studies in CCR5-deficient mice, in addition to our GVHD studies, demonstrate enhanced cell-mediated immune responses during pathogen infection,^16-18  delayed-type hypersensitivity,¹⁹  and tumor vaccination,²⁰  suggesting an immunoregulatory role for CCR5.

Within the past decade, CD4⁺CD25⁺ regulatory T cells (T_regs) have been recognized as important mediators of peripheral tolerance, and deficiency of this population is associated with autoimmune inflammation, including type I diabetes,^21-23  gastritis,²⁴  inflammatory bowel disease,²⁵  and thyroiditis²⁴  in several murine models. T_regs inhibit T-cell activation, proliferation, and effector function (reviewed in Piccirillo and Shevach²⁶ ). T_regs are also capable of inhibiting allogeneic T-cell responses, such as skin and solid organ allograft rejection.^27-30  T_reg-mediated inhibition of acute GVHD has been demonstrated by our group and others.^31-37  The gene FoxP3, encoding the forkhead family transcription factor Scurfin, is constitutively expressed by T_regs, and is crucial in the development of their suppressive phenotype.^38-44 

Our group and others have demonstrated the suppression of T-cell expansion within lymphoid tissues during GVHD by T_regs.^45,46  The presence of T_regs has been demonstrated within parenchymal organs such as the pancreas^21,47  and lung,⁴¹  at sites of infection,^48-51  and within allografts³⁰  and tumors,^52-54  and their presence at these sites has been shown to impact local inflammation. Although a number of studies have demonstrated chemokine receptor expression and responsiveness of T_regs to chemokines,^22,53,55-59  no study has definitively demonstrated a role for proinflammatory chemokine receptors in T_reg function. Accumulation of FoxP3-expressing cells within cardiac allografts was recently shown to be dependent on CCR4 expression by recipient mice.³⁰  However, a direct requirement for CCR4 expression by T_regs was not demonstrated.

Here we show that the expression of CCR5 by donor T_regs is critical for their ability to suppress lethality due to acute GVHD. The inability of CCR5^-/- T_regs to suppress GVHD lethality correlates with reduced accumulation of CCR5^-/- as compared with wild-type (WT) T_regs in GVHD target organs more than one week after transplantation.

Mice used for transplantation were described previously.^4,60  Within each experiment, all recipient mice were the same sex and age, which ranged from 6 to 10 weeks. Donor mice were gender-matched to recipients, and ranged from 6 to 13 weeks of age. Scurfy mice (Sf^-/-), which have been described,⁶¹  were a kind gift from Dr Virginia Godfrey. All animal experiments were performed in accordance with protocols approved by the University of North Carolina Institutional Animal Care and Use Committee.

T-cell–depleted bone marrow (TCD-BM) and whole splenic T cells were prepared from WT donors as described.⁴  CD25^- splenic T cells were prepared from WT or CCR5^-/- donors as follows: splenic T cells were isolated using Mouse T-cell Immunocolumns (Cedarlane Laboratories, Hornby, ON, Canada), according to the manufacturer's protocol, and stained with anti-CD25-biotin (clone 7D4; BD Biosciences Pharmingen, San Diego, CA), followed by streptavidin–phycoerythrin (PE). Cells were labeled with anti-PE microbeads (Miltenyi Biotech, Gladbach, Germany), and CD25⁺ cells depleted following methods from the manufacturer (Miltenyi Biotech). The depletion of CD25⁺ T cells was more than 90%. In some experiments, CD4⁺CD25⁺ T_regs were purified by isolating CD25⁺ cells using the approach described above, staining with anti-CD4-Cy5 (clone RM4-5; BD Biosciences Pharmingen), and sorting the CD4⁺CD25⁺ cells to 90% to 97% purity using a MoFlo high speed cell sorter (Cytomation, Fort Collins, CO). In most experiments, T_regs were purified from the spleens of donor mice, to 90% to 95% purity, using a previously described column-purification protocol.⁴⁵ 

Irradiation, transplantation, and housing of mice were performed as described.⁴  Transplant recipients received 3 × 10⁶ TCD WT bone marrow cells and 5 × 10⁶ CD25-depleted (CD25^-) T cells (WT or CCR5^-/-). For survival studies, some groups received transplants supplemented with T_regs from WT or CCR5^-/- donors. The dose of T_regs administered for survival and GVHD grading studies was 1.5 × 10⁶ in experiments using T_regs sorted by fluorescence-activated cell sorting (FACS), and 2.75 × 10⁶ in experiments using column-purified T_regs. For in vivo migration studies, recipients were given transplants containing WT TCD-BM and either WT whole splenic T cells or WT CD25^- T cells, alone or supplemented with 0.5 × 10⁶ to 1.5 × 10⁶ T_regs purified from WT or CCR5^-/- donors. In most migration studies, T_reg donors were WT/enhanced green fluorescent protein (eGFP) and CCR5^-/-/eGFP.

Mice were observed twice weekly and GVHD evaluated using a published clinical scoring system.⁶² 

Isolation of total infiltrating leukocytes from target organs has been described previously.⁴  To purify donor cell infiltrates, the total organ leukocytes were stained with anti-H2K^d-PE (clone SF1-1.1; BD Biosciences Pharmingen), followed by incubation with anti-PE microbeads (Miltenyi Biotech), and depletion of H2K^d-expressing host cells according to the manufacturer's instructions. The remaining cells were more than 95% donor.

Monoclonal antibodies used for FACS were obtained from BD Biosciences Pharmingen and included anti-H2K^d-PE (or a biotinylated version) (SF1-1.1–mouse IgG2a-kappa), anti-mCD4-PerCP (RM4-5–rat IgG2a-kappa), anti-CD25-biotin (clone 7D4–rat IgM). Isotype controls were obtained from BD Biosciences Pharmingen. To generate polyclonal antimurine Scurfin antibody, rabbits were immunized with the peptide CLLGTRGSGGPFQGRDLRSGAH from the N-terminus of the murine Scurfin protein (Proteintech Group, Chicago, IL). Polyclonal antibody was affinity-purified using the immunizing peptide via the SulfoLink Kit (Pierce Biotechology, Rockford, IL). Specificity of the purified antibody was confirmed by Western blot and FACS analysis (Figure S1, available on the Blood website; see the Supplemental Figure link at the top of the online article). For intracellular staining, BD FACS permeabilizing solution 2 (BD Biosciences Pharmingen) was used according to the manufacturer's protocol. Cells were stained with antimouse Scurfin antibody or normal rabbit IgG (Caltag Laboratories, Burlingame, CA), followed by goat anti–rabbit IgG (H+L)–PE (Caltag). Flow cytometry and analysis were performed as previously described.⁴ 

T_regs and CD4⁺CD25^- T cells were isolated from WT and CCR5^-/- donors, and 1 × 10⁶ cells were activated and cultured using immobilized anti-CD3ϵ antibody and interleukin 2 (IL-2).³³  After 5 days in culture, cells were counted and chemotaxis assays performed to assess migration to 1, 10, 100, and 1000 ng/mL recombinant murine CCL3, CCL4, or CCL5 (Peprotech, Rocky Hill, NJ) as previously described.⁴ 

Total RNA isolation, synthesis of cDNA, and quantitative real-time RT-PCR (QPCR) were performed using a previously described method.⁴  Primers and probe for murine CCR5 were as follows (5′ to 3′): CCR5 forward: GACTCTGGCTCTTGCAGGAT; CCR5 reverse: GCCGCAATTTGTTTCACAT; probe: TCAAGGGTCAGTTCCGACCTATAGC.

Murine FoxP3 and 18S rRNA primers and probes have been previously described.^4,40  Copies of CCR5 or FoxP3 mRNA in each sample were calculated using standard curves, and were normalized by dividing by the copies of 18S rRNA and multiplying by 10⁴.

Allogeneic stimulators were prepared from spleen cell suspensions of B6D2 mice, depleted of T cells using anti-CD90 magnetic beads (Miltenyi Biotech), and irradiated with 2100 cGy from a ¹³⁷Cs source. Stimulators (1 × 10⁵) were incubated with responders (WT CD4⁺CD25^- T cells, 1 × 10⁵) and various ratios of T_regs from WT or CCR5^-/- donors. Cultures were incubated for 5 days at 37°C and 1 μCi (0.037 MBq) ³H-thymidine was added for the last 16 hours of culture. Incorporation of ³H-thymidine was assessed as described.⁴  In other experiments, a previously described protocol was used to measure T_reg-mediated suppression.⁶³  To measure suppression by T_regs sorted from GVHD target organs, cultures were prepared as above, using allogeneic stimulators (2.5 × 10⁴), and WT CD4⁺CD25^- responders (2.5 × 10⁴). Infiltrating leukocytes were isolated from GVHD target organs and the GFP-expressing CD4⁺ T_regs or CD4⁺CD25^- T cells were sorted using a MoFlo high-speed cell sorter (Cytomation). T_regs or CD4⁺CD25^- T cells were added to cultures at the indicated ratios. Anti-CD3ϵ antibody (0.5 μg/mL) was added, and ³H-thymidine incorporation was assessed after 72 hours.

T_regs were column-purified from WT or CCR5^-/- donors, activated, and expanded in culture. Aliquots of cells were taken on days 3 and 5 of culture, and assessed for apoptosis using an Annexin V Apoptosis Detection kit (BD Biosciences Pharmingen), according to the manufacturer's instructions.

Estimates of the probability of survival for all groups were determined using the method of Kaplan and Meier.⁶⁴  Mean survival times were compared using the rank sum test. Cells migrating to chemokines were log transformed and compared by Student t test. Clinical scores, cell infiltration, and FoxP3 expression were compared by Student t test. For all tests, P values less than or equal to .05 were considered significant.

Initially, we assessed the expression of CCR5 on T_regs, and the ability of those cells to migrate in response to CCR5-binding chemokines in vitro. We found that CD4⁺CD25⁺ T cells isolated from the spleens of unmanipulated C57BL/6 (WT) mice did not express significant levels of CCR5 mRNA. However, after polyclonal activation by culturing on immobilized anti-CD3ϵ monoclonal antibody in the presence of IL-2, expression of CCR5 mRNA was strongly upregulated on WT T_regs (Figure 1A). We observed significant migration of WT T_regs in response to the CCR5-binding chemokines CCL3, CCL4, and CCL5. Responsiveness of activated T_regs to a physiologic concentration (10 ng/mL) of CCL4 and CCL5 was significantly greater than that of similarly activated CD4⁺CD25^- T cells, with a strong trend for increased responsiveness to CCL3 (Figure 1B). Similar trends were observed at chemokine doses of 1 ng/mL and 100 ng/mL (data not shown). T_regs isolated from CCR5^-/- mice did not migrate in response to the CCR5-specific chemokine CCL4 (Figure 1C), confirming that the responsiveness of WT T_regs to this chemokine was CCR5 specific. Activated WT T_regs that migrated in response to CCL4 expressed similar levels of FoxP3 mRNA, compared with T_regs evaluated prior to the migration assay (Figure 1D), demonstrating that CCL4 induces the migration of FoxP3-expressing T_regs.

To address the role of CCR5 expression in the function of donor T_regs in vivo during acute GVHD, we used a parent-to-F1 (first filial generation) murine GVHD model similar to that used in our previous study.⁴  Lethally irradiated B6D2 recipients were given transplants consisting of 3 × 10⁶ WT TCD-BM cells plus 5 × 10⁶ CD25-depleted (CD25^-) T cells from WT or CCR5^-/- mice. All of these recipients suffered GVHD-related mortality. However, unlike our previous findings using whole (T_reg-replete) splenic T cells,⁴  we did not observe earlier mortality in recipients of CCR5^-/- CD25^- T cells (Figure 2A). Median survival of recipients of WT or CCR5^-/- CD25^- T cells was 17 and 19 days, respectively (P = .23), suggesting that the enhanced GVHD found previously using CCR5^-/- T cells may be due to impaired in vivo function of CCR5^-/- T_regs.

To more specifically assess differences in the ability of WT and CCR5^-/- T_regs to inhibit GVHD, we added back WT or CCR5^-/- T_regs to transplants containing WT CD25^- T cells, and compared their ability to prevent GVHD lethality and reduce signs of GVHD-related morbidity (Figure 2A-B). Adding WT T_regs to transplants was highly effective in preventing GVHD lethality, as median survival increased from 17 days in mice receiving WT CD25^- T cells alone, to 59 days (P = .002), with 50% of those mice surviving until termination of the experiment on day 75. This correlated with a significant decrease in clinical GVHD score compared with mice receiving WT CD25^- T cells alone, starting one week after transplantation and continuing throughout the course of the experiment. In contrast, adding CCR5^-/- T_regs did not prevent lethality, as only 12.5% of recipients survived through day 75. Furthermore, adding CCR5^-/- T_regs was significantly less effective in prolonging median survival time compared with adding WT T_regs (P = .02). Adding CCR5^-/- T_regs decreased the clinical score of GVHD in the first 4 weeks after transplantation, compared with control mice receiving WT CD25^- T cells alone. However, by day 30 and for the remainder of the experiment, clinical scores in mice receiving CCR5^-/- T_regs were similar to the peak scores found in the control group. These data demonstrated that CCR5^-/- T_regs were significantly impaired in their ability to inhibit GVHD in vivo.

We next sought to address the mechanism for the impaired in vivo function of CCR5^-/- T_regs during GVHD. We compared the suppressive abilities of T_regs isolated from the spleens of unmanipulated WT and CCR5^-/- mice using previously published in vitro assays.^34,63  CD4⁺CD25⁺ T cells were present in comparable proportions in the spleens and peripheral lymph nodes of unmanipulated CCR5^-/- and WT mice (Figure 3A). T _regs from CCR5^-/- mice were comparable to those from WT mice in their ability to suppress the proliferation of WT CD4⁺CD25^- responder cells in response to irradiated B6D2 allogeneic stimulators at all T_reg-to-responder ratios (Figure 3B). FoxP3 mRNA expression in CCR5^-/- T_regs was comparable to that in WT T_regs (Figure 3C). We also assessed the ability of WT and CCR5^-/- T_regs to proliferate and undergo apoptosis when activated and expanded in culture. A similar net expansion of CCR5^-/- and WT T_regs was found over 5 days in vitro, and similar proportions of WT and CCR5^-/- T_regs underwent apoptosis at days 3 and 5 in culture (data not shown). Thus, CCR5 does not play a critical role in regulating the development, suppressive function, or FoxP3 gene expression of T_regs, and the absence of CCR5 does not impair the proliferation or survival of T_regs in vitro.

Previous studies have demonstrated that T_reg function during GVHD was dependent on the expression of l-selectin (CD62L) by T_regs.^45,46 l-selectin expression on the surface of WT and CCR5^-/- T_regs and conventional CD4 T cells was similar both in freshly isolated cells and cells isolated during the first week after transplantation (data not shown).

We assessed the accumulation of WT and CCR5^-/- T_regs in lymphoid tissues after transplantation by purifying CD4⁺CD25⁺ T_regs from WT/eGFP or CCR5^-/-/eGFP transgenic mice and adding eGFP-expressing T_regs to transplants containing WT TCD-BM and WT whole splenic T cells. On day 5 after transplantation, no significant differences were observed in the numbers of GFP-positive CD4 T cells in spleen or mesenteric lymph node (MLN) of mice receiving WT or CCR5^-/-/eGFP T_regs (Figure 4). Additionally, no significant differences were observed in the numbers of WT/eGFP and CCR5^-/-/eGFP T_regs infiltrating the liver and lung. Similar results were found when assessing the accumulation of WT or CCR5^-/- T_regs in the spleen, MLN, liver, and lung on day 7 after transplantation (data not shown). Thus, the absence of CCR5 on T_regs did not affect the accumulation of these cells in either lymphoid tissues or GVHD target organs in the first week after transplantation.

The accumulation of T_regs in GVHD target organs in the second week after transplantation was dependent on CCR5. For these experiments, recipients underwent transplantation as described above, except that CD25-depleted T cells as opposed to whole splenic T cells were given. CCR5^-/-/eGFP T_reg numbers were lower than WT/eGFP T_reg numbers in both the liver and spleen on day 10 after transplantation (Figure 5A). No significant differences in CCR5^-/- versus WT GFP⁺CD4 T-cell accumulation were observed in the MLN at this time (data not shown).

To address the concern that the GFP⁺ cells evaluated could be contaminated with rapidly proliferating non-T_regs,^41,43  we quantified the expression of FoxP3 in donor cell infiltrates from specific tissues isolated from mice at day 10. We found significant levels of FoxP3 mRNA expression in donor cells isolated from target organs only when T_regs were added to the transplants, confirming the utility of this approach to assess T_reg accumulation. FoxP3 mRNA expression in donor cell infiltrates isolated from the livers of mice receiving CCR5^-/-/eGFP T_regs was lower than that in mice receiving WT T_regs. Decreased FoxP3 expression correlated directly with the reduction in CCR5^-/- GFP⁺CD4 T-cell numbers in livers, demonstrating that the accumulation of CCR5^-/-/eGFP T_regs at this site was significantly less than that of WT/eGFP T_regs (Figure 5B). FoxP3 mRNA levels were also reduced in donor cell infiltrates from the lungs of mice receiving CCR5^-/- T_regs as compared with WT T_regs (Figure 5C). Interestingly, although we had observed a significant reduction in GFP⁺CD4 T-cell numbers in the spleens of mice receiving CCR5^-/-/eGFP T_regs as compared with WT/eGFP T_regs, FoxP3 mRNA levels were not significantly different between these 2 groups (Figure 5B).

We assessed FoxP3 expression in donor cells infiltrating the liver, spleen, and MLN at days 13 and 16 after transplantation (Figure 5D). At days 13 and 16, FoxP3 expression in donor cell infiltrates isolated from liver, as well as spleen and MLN, was lower when mice received CCR5^-/-/eGFP T_regs, as compared with WT/eGFP T_regs. Importantly, this did not correlate with an overall reduction in GFP-expressing cells in GVHD target tissues over time, as the number of both WT and CCR5^-/- GFP⁺CD4 T cells infiltrating liver (CCR5^-/- 167% increase; WT 235%) and spleen (CCR5^-/- 135%; WT 91%) increased between days 10 and 13. Thus, although CCR5 expression by T_regs was not required for their accumulation in lymphoid tissues during the first week after transplantation, accumulation of CCR5^-/- T_regs in lymphoid tissues was lower than WT T_regs after day 10. The discrepancy between GFP⁺CD4⁺ T-cell numbers and FoxP3 mRNA expression in donor cells isolated from spleens at day 10 suggested that a proportion of the GFP⁺CD4⁺ T cells infiltrating the spleens were non-T_reg.

To resolve the discrepancy between GFP⁺ T_reg numbers and FoxP3 expression data, we used a novel anti-Scurfin polyclonal antibody generated by our group (Figure S1) to determine the total number of GFP⁺CD4⁺Scurfin⁺ cells infiltrating the spleen and liver 10 days after transplantation (Figure 5E). We found no significant difference in GFP⁺CD4⁺Scurfin⁺ cell numbers in the spleens of mice receiving WT/eGFP or CCR5^-/-/eGFP T_regs, but a significant 3-fold reduction, which was similar to the difference in FoxP3 mRNA expression, in the number of these cells in the livers of mice receiving CCR5^-/-/eGFP compared with WT/eGFP T_regs. Thus, measurement of FoxP3 mRNA expression in donor-cell infiltrates isolated from target organs, or flow cytometric determination of the frequency of FoxP3⁺ cells, are more rigorous methods of quantifying T_reg accumulation during GVHD than use of labeled CD4⁺CD25⁺ T cells. This observation is consistent with recent data demonstrating that approximately 25% of CD4⁺CD25⁺ cells from normal mice do not express FoxP3 protein, and in fact proliferate when stimulated in vitro.⁴¹ 

We sought to determine whether the FoxP3-expressing T_regs residing in target tissues during GVHD maintain their suppressive phenotype. To accomplish this, we isolated GFP-expressing T_regs from the spleens and livers of mice having received transplants containing WT TCD-BM, WT CD25^- T cells, and either WT/eGFP or CCR5^-/-/eGFP T_regs, 14 days earlier. As a control, a group of recipient mice received transplants containing CD4⁺CD25^- T cells from WT/eGFP donors. Both WT/eGFP and CCR5^-/-/eGFP T_regs effectively suppressed responder cell proliferation (Figure 6), when sorted from either the spleen or the liver. GFP⁺ CD4⁺CD25^- T cells isolated from the spleens of control mice did not suppress responder cell proliferation. Thus, donor T_regs infiltrating both lymphoid and extra-lymphoid GVHD target organs maintained their suppressive phenotype through the first 2 weeks after transplantation. Furthermore, these data indicate that the lack of function of CCR5^-/- T_regs in suppressing GVHD lethality was not due to loss of the suppressive phenotype during GVHD in vivo.

The function of naturally occurring CD4⁺CD25⁺ T_regs in modulating GVHD has been well documented.^31-36,46  Our group and others have shown previously that the ability of l-selectin^Hi T_regs to inhibit effector T-cell expansion within secondary lymphoid tissues early after allo-BMT correlated with their ability to suppress GVHD.^45,46  Interestingly, here we demonstrate that recruitment of T_regs to sites of inflammation and lymphoid tissues later after transplantation via the chemokine receptor CCR5 is also crucial for the function of these cells. This study provides the first definitive demonstration of a requirement for a chemokine receptor in the function of these cells in vivo.

Previous studies have demonstrated that CCR5 expression by T_regs is specific to a subset of T_regs, which preferentially infiltrate extralymphoid sites and sites of inflammation.^21,41  This T_reg subset has been shown to express a more activated phenotype. Our data support the notion that CCR5 expression by T_regs requires an activation stimulus. It is interesting to note that a number of studies have demonstrated enhanced cell-mediated immune responses in CCR5^-/- mice, including increased GVHD severity,^4,5  tumor immunity,²⁰  responses to pathogens,^16-18  and delayed-type hypersensitivity.¹⁹  Additionally, humans with the CCR5Δ32 mutation, which reduces CCR5 expression on leukocytes,⁶⁵  have a propensity for autoimmune diseases of the liver and lung.^66-68  Thus, CCR5 expression by T_regs may be important in suppression of inflammation in a specific subset of organs.

Our data suggest that CCR5 is not required for accumulation of T_regs in lymphoid tissues during the first 10 days after transplantation. Although we have demonstrated CCR5-ligand expression within the spleen beginning early after transplantation in this model,⁴  the similarity in WT and CCR5^-/- T_reg numbers in spleen and MLN during the first 10 days after transplantation demonstrates that the presence of T_regs in lymphoid tissues does not initially depend on their expression of CCR5. Based on previous studies, l-selectin, perhaps acting in concert with ligands that bind to CCR7, directs the early homing of T_regs to lymphoid tissues after transplantation.^22,45,46 

Previously, we demonstrated expression of the CCR5-ligands CCL3 and CCL4 in the liver and lung during GVHD, which reached peak levels between days 7 and 14.⁴  Here, although not affected during the first week, by 10 days after transplantation, CCR5^-/- T_reg infiltrates were significantly lower than WT T_regs. Lower CCR5^-/- T_reg infiltration of the liver was also evident at later time points. Thus, the time points at which target organ infiltration by T_regs was affected by CCR5-deficiency correlated with the kinetics of CCR5-ligand expression at these sites. Decreased infiltration of CCR5^-/- T_regs in the liver and lung correlated with increased total CCR5^-/- CD4⁺ and CD8⁺ T cells, and increased tissue pathology found in our previous study at these sites.⁴  Interestingly, on days 13 and 16, lower CCR5^-/- T_reg infiltration was observed in spleen and MLN as well. Thus, although CCR5 does not direct the initial entry of T_regs into lymphoid tissues after transplantation, at later time points activated T_regs may recirculate to lymphoid tissues via CCL3, CCL4, or CCL5 produced at these sites.

Reduced infiltration of CCR5^-/- T_regs, as compared with WT T_regs, most likely reflects impairment in their migration in response to ligands that bind to CCR5. Conceivably, defects in proliferation or survival could be responsible for these findings, although we were unable to demonstrate these in vitro. Comparable suppressive function of CCR5^-/- and WT T_regs sorted from target organ infiltrates on day 14 after transplantation is not consistent with a greater fraction of CCR5^-/- T_regs undergoing apoptosis in target tissues. Additionally, we found that the number of eGFP-expressing cells in target organs increased over time in mice receiving either WT/eGFP or CCR5^-/-/eGFP T_regs, strongly suggesting that differential expansion or survival of CCR5^-/-/eGFP T_regs is not responsible for the differences observed.

We propose a bimodal model for the function of T_regs in suppressing lethal acute GVHD. Early after transplantation, T_regs limit the expansion and differentiation of alloantigen-specific T cells within lymphoid tissues. Activation of T_regs within lymphoid tissues induces expression of CCR5. The presence of ligands for CCR5 in target tissues, which are produced by infiltrating donor T cells,⁶⁹  would induce T_reg migration to those sites and further limit the local expansion and/or cytolytic effector function of alloreactive T cells. Later, the expression of CCR5 may be crucial in the migration of activated T_regs back to inflamed lymphoid tissues.

In conclusion, we have found that the expression of CCR5 on donor T_regs is critical for their infiltration of GVHD target organs later than one week after transplantation and eliminating expression of CCR5 from T_reg-exacerbated GVHD in lethally conditioned, MHC-mismatched recipients. Our findings suggest that the CCR5Δ32 allele may be a predisposing factor for the occurrence of lethal GVHD after allogeneic human stem cell transplantation in recipients of myeloablative conditioning. Additionally, inhibition of T_reg migration may be clinically important in tumor immunotherapy, as migration of T_regs to the tumor may represent a barrier to therapeutic efficacy.⁵³  We are currently investigating the importance of CCR5 in T_reg-mediated tolerance in tumor models.

The authors thank Christin Raimondo, Ivana Ferrer, Nicholas Siefers, V. McNeil Coffield, Shannon Pop, and William Reed for technical assistance, and Larry Arnold and the University of North Carolina Flow Cytometry Facility for FACS-based cell sorting.