Anti-lymphocyte-depleting antibodies have increasingly been utilized in the clinic as induction therapy aiming to improve transplantation outcomes by reducing the need for long-term immunosuppression. However, maintenance immunosuppression is still required as lymphocyte reconstitution through homeostatic proliferation, partially driven by IL-7, continues to replenish tolerance-refractory immune cells capable of rejection. In murine models of MHC mismatched skin grafting, we investigated whether it is feasible to control the lymphocyte reconstitution process to delay rejection and favor tolerance processes. We found that a short course of anti-IL-7 receptor blocking antibody following T cell depletion, combined with the mammalian target of rapamycin inhibitor Rapamycin, could significantly delay graft rejection in one mouse strain, and achieve transplantation tolerance in another. The combination treatment was found to delay T cell reconstitution and, in the short term, enriched for Foxp3+ regulatory T cells (Tregs), at the expense of effector cells. Extended graft survival and tolerance were dependent on TGF-ß, indicating a role for induced Tregs. These findings point to the feasibility of building on lympholytic induction by guiding early lymphocyte reconstitution to favor endogenous regulatory mechanisms.
aglycosylated variant of αIL-7R
blocking anti-IL-7R antibody
fluorescence-activated cell sorting
median survival time
mammalian target of rapamycin
peripherally induced Treg
regulatory T cell
Anti-lymphocyte-depleting antibodies are widely adopted as induction therapy in transplantation to safeguard the graft during healing, after which the requirements for immunosuppression are less intense. Lymphocyte depletion alone, albeit profound and sustained, does not induce graft acceptance  as rebound lymphocyte reconstitution is antagonistic to tolerance [2, 3]. In part, this rebound is mediated through feedback control by IL-7 [4, 5], in a process often referred to as homeostatic proliferation . This results in preferential emergence of effector/memory T cells [7-10], with the lymphocyte population acquiring resistance to induction of tolerance [3, 11]. As an infusion of regulatory T cells (Tregs) could overcome this tolerance-refractory state , we asked whether one could therapeutically guide reconstitution to favor the endogenous Tregs. We targeted the IL-7 receptor-α (IL-7R) and mammalian target of rapamycin (mTOR) as both are involved in homeostasis and activation of T cells, while mTOR inhibition also favors the induction of Foxp3+ Tregs [12, 13]. Moreover, as IL-7R is poorly expressed in naïve Tregs , we expected operational selectivity in the quality of reconstitution following IL-7R blockade.
We demonstrate that a short course of an mAb blocking anti-IL-7R (αIL-7R) , when used after lymphocyte depletion in association with the mTOR inhibitor Rapamycin, can substantially delay graft rejection, and in one mouse strain, induce tolerance to fully MHC mismatched skin grafts. This protocol favors early Treg reconstitution over other T cells, and is dependent on TGF-ß, implicating a role for peripherally induced Tregs (pTregs).
Materials and Methods
CBA/Ca (H-2k), CBA.Rag1−/− (H-2k), C57BL/10 (B10) (H-2b), (B10 × BALB/k)F1 (H-2b,k), C57BL/6 (B6) (H-2b), (B6 × CBA/Ca)F1 (H-2b,k), D90.1 CBA/Ca (Thy1.1) (H-2k), homozygous CBA/Ca mice transgenic for hCD52 expression in T cells (hCD52-Tg) (H-2k)  and heterozygous (B6xhCD52-Tg)F1 mice were bred and maintained under specific pathogen-free conditions in the animal facility at the Sir William Dunn School of Pathology (Oxford, UK). Heterozygous (hCD52-Tg × CBA/Ca)F1 were used in some experiments instead of hCD52-Tg, because of homozygotes breeding shortage.
Skin grafting and T cell transfer were performed between sex-matched mice. All procedures were conducted in accordance with the Home Office Animals (Scientific Procedures) Act of 1986.
mAbs and immunosuppressants
For T cell depletion mice were treated with campath, a rat IgG2b anti-hCD52 , 1 mg intraperitoneally (i.p.) on days −3 and −1, such depletion being equivalent in both homozygous and heterozygous hCD52-Tg mice. For B cell depletion mice received a mouse IgG2a anti-murine CD20 (αCD20) from Genentech (San Francisco, CA), 250 µg i.p. on days −3 and −1. For co-receptor blockade, mice were given YTS177.9, a rat IgG2a anti-mouse CD4 (αCD4), and YTS105.18, a rat IgG2a anti-mouse CD8 (αCD8), 1 mg of each i.p. on days 0, 2 and 5. For IL-7R blockade, mice received A7R34, a rat IgG1 αIL-7R , or its aglycosylated variant (agly-αIL-7R), seven i.p. injections of 0.4 mg from days −3 to 14. This antibody was validated by blockade of B cell development  (Figure S1). Campath, αCD4, αCD8 and αIL-7R were produced in our laboratory as previously described. For TGF-ß neutralization, mice were given 1D11, a mouse IgG1 αTGF-ß  (or the relevant isotype 13C4) seven i.p. injections of 2 mg from days −3 to 21; antibodies provided by Genzyme (Cambridge, MA). Rapamycin (Rapa) (Calbiochem, Nottingham, Nottinghamshire, UK) was dissolved in 100% ethanol and stored at −20°C; on the day of the injection a stock aliquot was diluted with 0.2% carboxymethycellulose sodium salt (Sigma-Aldrich, Gillingham, Dorset, UK) and administered i.p.; unless otherwise specified, mice received seven injections of 1.5 mg/kg from days −3 to 14 .
Preparation of chimeric aglycosyl human IgG1 αIL-7R
DNA encoding VH and VL of the rat αIL-7R  was prepared using the SMART cDNA kit (Clontech, Saint-Germain-en-Laye, France) applying SMART 11 A oligo AAGCAGTGGTATCAACGCAGAGTACGCGGG and kappa SMART primer GGTCTAACACTCATTCCTGTTGAA for the light chain and rat2ACH2 SMART primer ATCTTTGGTCTTTGGGGGGAAGAT for the heavy chain. Second strand RACE was then performed using primer nest A, AAGCAGTGGTATCAACGCAGAGT and kappa nest primer CTAACACTCATTCCTGTTGAAGCTCTT for the light chain and rat2AHinge nest primer TGTACATCCACAAGGATTGCATTC for the heavy chain. The amplified light chain fragment was cloned into EcoR1 site of PEE12 vector (Lonza, Slough, Berkshire, UK) and heavy chain fragment into Spe1 and BamH1 site of PEE12/CD5L/HuIgG1 aglycosyl vector. The resulting plasmids were co-transfected by electroporation into the cell line NSO, and stable transfectants were selected in glutamine-free IMDM (Invitrogen, Paisley, Renfrewshire, UK). Positive clones were selected by ELISA for binding to recombinant mouse IL-7RFc chimera (R&D Systems, Abingdon, Oxfordshire, UK). Antibody was purified by ProteinA affinity-binding (GE Healthcare, Little Chalfont, Buckinghamshire, UK) and quality-controlled for endotoxin.
Full thickness tail-skin (0.5–1 cm2) was transplanted on recipients thoracic flank. When performing re-grafts, these were transplanted contra-laterally. Grafts were generally observed 3 times/week after the removal of the bandage at day 7, and considered rejected when no viable donor skin was present.
T cell enrichment and adoptive transfer
A single-cell suspension was obtained from pooled spleens and nodes, and erythrocytes were lysed with Tris-buffered ammonium-chloride. Cells were incubated with M5/114 and 187.1, rat antibodies anti-mouse H2-I-A/I-E and kappa chain, then with sheep anti-rat IgG magnetic beads (Dynabeads; Invitrogen), and negatively sorted. To investigate the effect of αIL-7R and Rapa in lymphocyte-deficient mice, T cell preparations from CBA/Ca donors were transferred into syngeneic CBA.Rag1−/− recipients at the dose of 6.5–9.5 × 106 per mouse, 3 days before B6 skin transplantation. To assess proliferation, 107 CFSE-labeled T cells from Thy1.1+ donors were transferred into syngeneic hCD52-Tg mice.
Analysis of T cell proliferation in vivo according to CFSE dilution
Thy1.1+ T cell enriched preparations were labeled with carboxyfluorescein diacetate succinimidyl ester 3 µM solution (CFSE; Molecular Probes, Paisley, Renfrewshire, UK), and injected into hCD52-Tg recipients on day 0 after depletion with campath; 7 days later spleens were collected for fluorescence-activated cell sorting (FACS) and proliferation analyses.
Analysis of T cell proliferation in vivo by EdU labeling
5-ethynyl-2′-deoxyuridine (EdU; Invitrogen) was dissolved in drinking water to 0.8 mg/mL and made available ad libitum to hCD52-Tg mice for 6 days from day 1 after depletion. Spleens were then collected and EdU-labeled T cells were detected with Click-iT EdU Flow-Cytometry Assay Kit (Invitrogen).
Spleens were collected in RPMI containing 10% fetal bovine serum; single-cell suspensions and erythrocytes lysis were obtained as above; vital cells were counted with Trypan Blue (Sigma-Aldrich). Blood samples were collected in heparin and erythrocytes water lysis performed. Nonspecific labeling was blocked with an anti-mouse Fc gamma receptor antibody (2.4G2). Cells were stained for membrane markers with primary antibodies, then, where required, with streptavidin conjugates (BD Pharmingen, Oxford, Oxfordshire, UK), finally fixed with formaldehyde. For Foxp3 staining, cells were fixed and permeabilized with Foxp3 Staining Buffer Set (eBioscience, Hatfield, Hertforshire, UK), then stained with anti-Foxp3 antibodies. The following anti-mouse antibodies and relevant isotype controls were used: anti-CD19 (1D3), anti-CD45R/B220 (RA3-6B2), anti-CD3 (145-2C11), anti-CD4 (H129.19 and RM4-5), anti-CD8 (53-6.7), anti-CD44 (IM7), anti-CD62L (MEL-14), anti-CD90.1 (OX-7) (from BD Pharmingen); anti-CD25 (PC61.5) and anti-Foxp3 (FJK-16s) (from eBioscience); goat-anti-mouse IgG (Sigma); anti-kappa chain (187.1, produced in our laboratory). Flow-cytometry was performed with a four-color FACSCalibur flow-cytometer (BD Biosciences). Data were analyzed using FlowJo software, version 9.5 (Tree Star, Ashland, OR).
Statistical analysis was performed using Graphpad Prism software, version 5.0d (San Diego, CA). Log-rank Mantel-Cox test was used to compare graft survival curves and, unless otherwise specified, one-way analysis of variance followed by Bonferroni's posttests for comparison among the groups. p-Values < 0.05 were considered statistically significant, p > 0.05 not significant (NS), *p < 0.05, **p < 0.01; ***p < 0.001 and ****p < 0.0001.
Lymphocyte depletion is insufficient for transplantation tolerance, but does not prevent its induction through co-receptor blockade
In order to study the effect of lymphocyte depletion and subsequent reconstitution on graft survival, we utilized, as recipients of MHC mismatched skin grafts, hCD52-Tg CBA/Ca mice , in which treatment with the anti-hCD52 antibody (campath) was particularly effective at depleting T cells (Figure 1A, Figure S2). T cell depletion in mice delayed, but never prevented, graft rejection (Figure 1B). Additional B cell depletion could not provide any further benefit (Figure S2). However, as anticipated , tolerance induction through co-receptor blockade was not hindered by T cell depletion (Figure 1B and C).
Short-course αIL-7R therapy following T cell depletion with campath can induce dominant transplantation tolerance in hCD52-Tg CBA/Ca mice
The addition of short-course αIL-7R to campath significantly prolonged graft survival, and in some experiments promoted long-term acceptance in more than 50% of the mice (p = 0.0003 vs. campath) (Figure 2A). The use of agly-αIL-7R achieved the same outcome (p < 0.0001 vs. campath) (Figure 2B and C), indicating that this benefit derives wholly from blockade of IL-7 binding to its receptor. Tolerant mice exhibited linked-suppression  to third-party (multiple minor) antigens incorporated in a second graft bearing the tolerated antigens (Figure 2D) so providing some evidence for active regulation. Moreover, αIL-7R prevented production of donor-specific antibodies (Figure S3), although this was not the overriding mechanism for enabling tolerance, as some mice making no anti-donor antibody rejected their grafts.
The combination of αIL-7R and Rapa promotes long-term graft survival
However, the benefit of αIL-7R after T cell depletion was not always as impressive, and variable results were observed across different experiments: We obtained indefinite median survival time (MST) (>150 days) in five experiments, MST 55 days in one experiment, MST 39 days in two experiments using hCD52-Tg homozygous CBA/Ca recipients and MST 21 days in one experiment using hCD52-Tg heterozygous mice (Figure 3A). The addition of Rapa to campath did not prevent rejection (MST 31 days), while the combination of Rapa with campath and αIL-7R achieved extended graft survival in virtually all the recipients (MST 56 days, p = 0.001 vs. campath + Rapa), and, in a separate experiment with a prolonged course of Rapa, complete graft survival in two-thirds of the mice (MST > 100 days) (Figure 3A). The synergy between αIL-7R and Rapa was also observed when small numbers of T cells were transferred into lymphocyte-deficient mice, where campath had not been used to purge T cells (Figure 3B), and in campath-treated (B6xhCD52-Tg)F1 mice (Figure 3C), although in this strain tolerance was not achieved.
Treatment with αIL-7R and Rapa delays T cell reconstitution and enriches for Tregs
In order to establish the extent to which αIL-7R and Rapa could influence T cell reconstitution, we treated hCD52-Tg mice with campath alone, or in combination with either αIL-7R or Rapa, or both drugs. We found that splenic T cell counts reached the nadir 1 week following depletion, irrespective of any additional drug (p < 0.001 vs. untreated group). Thereafter, T cell reconstitution started and was complete by week 4 in mice treated with campath alone, and proceeded similarly in mice that received αIL-7R. In contrast, reconstitution was delayed at week 4 in mice that received Rapa, alone or in combination with αIL-7R (p < 0.01 vs. untreated group). Nevertheless, even in these mice T cell counts returned to basal levels by weeks 12–14, 10–12 weeks after Rapa was stopped (Figure 4A, C and D). The addition of αIL-7R to Rapa prevented the imbalance between reconstituting CD8+ and CD4+ cells at weeks 1 and 4 (Figure 4B). Following depletion, irrespective of any additional drug, the emerging pattern of T cell phenotypes was significantly skewed toward an effector/memory composition characterized by more CD44+ cells and fewer CD62L+. This skewing was transient and disappeared once T cell reconstitution was complete (by week 4 for mice treated with campath ± αIL-7R, and by week 12–14 for those treated with campath + Rapa ± αIL-7R) (Figure 4E–H). Finally the addition of αIL-7R to campath induced a significant increase in the proportion of Foxp3+ Tregs at week 1; Rapa had a similar effect, which extended through to week 4; the combination of αIL-7R and Rapa further increased the proportion of reconstituting Tregs, which however, returned to basal levels by weeks 12–14. These data suggest that the treatment with αIL-7R and Rapa following depletion with campath might promote graft acceptance by delaying T cell reconstitution, and enriching for Tregs in the early reconstitution period. In this way the control of homeostatic expansion might provide a suitable context on which an allograft can impose tolerance.
Treatment with αIL-7R and Rapa selectively reduces the expansion of conventional T cells, but not regulatory T cells Tregs
We studied the effects of combined treatment on T cell proliferation in vivo by monitoring EdU incorporation in T cells. The proportion of EdU-labeled cells among splenic T cells was significantly higher in all the groups treated with campath compared to untreated controls (Figure 5A–C). While EdU incorporation into CD3+ cells seemed unaffected by αIL-7R and/or Rapa (Figure 5A), there was a significant reduction in the proportion of CD8+ EdU+ cells (46.6% ± 5.0 vs. 63.7% ± 9.2; p < 0.05) (Figure 5B) and an increase in CD4+ EdU+ cells (65.7% ± 3.0 vs. 47.6% ± 5.6; p < 0.01) (Figure 5C) with combined αIL-7R and Rapa, compared with campath alone. These findings suggest that treatment with αIL-7R and Rapa preferentially blocked CD8+ cell expansion.
The study above was conducted in circumstances where potentially lytic campath antibody remained in the circulation. In order to get a clear window on reconstitution, we studied proliferation by transferring CFSE-labeled T cells from Thy1.1+ donors into syngenic hCD52-Tg recipients. Consistent with the data on reconstitution (Figure 4), Thy1.1+ T cells transferred into hosts treated with campath acquired an effector/memory phenotype characterized by higher CD44 expression and lower CD62L, irrespective of any additional drug (Figure 6A and B). A significantly higher proportion of Thy1.1+ CD4+ Foxp3+ cells was found in mice treated with campath compared with controls (23.4% ± 5.3 vs. 6.7% ± 1.7, p < 0.001); the addition of αIL-7R or Rapa further enriched for Tregs (35.6% ± 4.0 and 39.0% ± 4.5, p < 0.01 vs. campath); and the combination of the three drugs was particularly effective at favoring Treg emergence (52.4% ± 3.7, p < 0.01 vs. campath + αIL-7R or campath + Rapa) (Figure 6C).
We studied the extent of proliferation of T cell subsets by analyzing CFSE dilution with the FlowJo proliferation platform. Two parameters were followed: the division index, the mean number of divisions of the whole cell population; and the proliferation index, the mean number of divisions of the cells that have divided at least once . The division index of Thy1.1+ CD8+ and CD4+ cells was increased when the cells were transferred into mice depleted with campath compared to that into replete controls (1.5 ± 0.3 vs. 0.1 ± 0.1, p < 0.001 for CD8+; 1.2 ± 0.1 vs. 0.1 ± 0.09, p < 0.001 for CD4+). The addition of αIL-7R or Rapa significantly decreased this index, and the combination of the two drugs further reduced it to levels similar to those found in controls (0.3 ± 0.1, p > 0.5 for CD8+; 0.2 ± 0.05, p > 0.5 for CD4+ vs. controls) (Figure 7A and B, left diagrams). The proliferation index of Thy1.1+ CD8+ cells was not affected by any treatment, whereas the proliferation index of Thy1.1+ CD4+ cells was increased in depleted mice treated with αIL-7R and Rapa (3.4 ± 0.5 vs. 1.8 ± 0.2 in mice treated with campath; p < 0.001) (Figure 7A and B, right diagram). The division index of Thy1.1+ CD4+ Foxp3+ and Foxp3− cells was similar following treatment with campath (1.0 ± 0.2 vs. 1.2 ± 0.2, p > 0.05), whereas the addition of αIL-7R and Rapa selectively reduced the division index of Foxp3− cells (0.8 ± 0.1 vs. 0.1 ± 0.01, p < 0.001) (Figure 7C). Taken together these results indicate that the combination of αIL-7R and Rapa reduces the homeostatic expansion of CD8+ and CD4+ T cells but does not hinder the accumulation of Tregs, which emerge preferentially under this treatment.
TGF-ß neutralization reduces graft survival in T cell–depleted mice treated with αIL-7R and Rapa
We asked whether TGF-ß, a cytokine crucial for the induction of pTreg  might play a role in this setting. In mice receiving the combined drugs TGF-ß neutralization substantially diminished graft survival time (MST 28 days vs. >40 days in mice that received the isotype control; p = 0.03) (Figure 8A), although the proportion of reconstituting Foxp3+ T cells favored by treatment with αIL-7R and Rapa was not reduced by TGF-ß blockade (Figure 8B).
Extended graft survival and even transplantation tolerance to MHC mismatched skin grafts were achieved following T cell depletion by use of a protocol based on guiding lymphocyte reconstitution. Clinical experience with alemtuzumab has indicated preferential survival of CD4+ CD25+ Tregs in patients [22, 23], and also in hCD52-Tr mice . Clearly, that differential impact of lymphocyte ablation was insufficient to prevent graft rejection . Previous studies in lymphopenic mice have indicated that sufficient naïve Tregs made available from the outset were able to control rejection responses mediated through homeostatic expansion [2, 25].
Antibody blockade of IL-7R has proven effective at preventing and treating certain experimental autoimmune diseases , and graft-versus-host disease . We found that short-course αIL-7R could, albeit variably, prolong the survival of fully mismatched skin grafts in hCD52-Tg mice following T cell depletion with campath, and induce transplantation tolerance in up to 50% of CBA mice carrying the CD52 transgene (Figure 2). This extended survival was associated with dominant tolerance because of the evidence of linked suppression toward third-party antigens presented together with the tolerated set . This variation between strains might be explained by the strain-specific persistence of effector/memory T cells, which are more resistant to antibody lysis than naïve T cells . The addition of Rapa, known to block proliferation of activated T cells , and to selectively enhance Treg expansion  provided a level of synergy in hCD52-Tg heterozygous mice (Figure 3A). The same combination could also delay rejection in lymphocyte-deficient mice injected with syngeneic T cells (Figure 3B), a system that provides a more stringent test of a tolerizing protocol  as transferred T cells rapidly repopulate by homeostatic expansion and a substantial proportion acquire a functional memory-like phenotype .
This protocol, despite enriching for Foxp3+ T cells early on, did not prevent eventual normalization of T cell counts and acquisition by some T cells of typical memory-like phenotypes [2, 9, 25] (Figure 4). Moreover, the addition of αIL-7R was capable of restraining the expansion of CD8+ T cells and promotion of CD8+ memory cells by Rapa when given alone.
Measures of proliferation in vivo with two different approaches (Figures 5 and 7) allowed us to confirm that treatment with αIL-7R and Rapa constrained CD8+ T cell proliferation. The effect on CD4+ T cells is more complex: αIL-7R and Rapa reduced the proportion of CD4+ cells entering division, but those cells that escaped this block proliferated more than reconstituting CD4+ cells in mice treated with campath alone. Consistent with our working hypothesis, we found that Tregs are included within this set of CD4+ cells whose proliferative capacity is unaffected by αIL-7R and Rapa.
A further and unexpected benefit of αIL-7R relates to its ability to inhibit generation of anti-donor antibodies (Figure S3); nonetheless, additional B cell ablation did not provide extended graft survival beyond T cell depletion alone (Figure S2).
Finally, neutralization of TGF-ß abrogated the benefit of our short maintenance protocol (Figure 8); as Tregs are essential for tolerance  and are peripherally induced by TGF-ß , enhanced graft loss following TGF-ß blockade might be due to decreased conversion to pTregs.
Collectively, these results suggest that one might build on the benefits of lymphocyte depletion as induction therapy, by additional short-term therapies, which could skew the lymphocyte reconstitution pattern in favor of regulation. This is not to say that the drug combination we have used should be adopted in the clinic, but rather emphasizes that future research should seek drug combinations targeting the character of lymphocyte reconstitution.
The work was supported by grants from the ERC Framework Programmes 6 “RISET” and 7 “Betacell therapy,” ERC Advanced Investigator Grant “PARIS,” and the Medical Research Council. GP was partially supported by a grant by Italian Society of Nephrology.
The authors of this manuscript have conflicts of interest to disclose as described by the American Journal of Transplantation. Drs. Waldmann and Cobbold are recipients of royalties for the commercialization of alemtuzumab, and the work reported here was partially supported by Genzyme Corporation.