Combined Administration of a Mutant TGF-β1/Fc and Rapamycin Promotes Induction of Regulatory T Cells and Islet Allograft Tolerance

Allograft outcome, whether rejection or tolerance, may depend on the balance between the function of effector and regulatory T cells (Tregs) (1–3). Strategies that can promote Treg expansion, while at the same time inhibiting effector T cells (Teffs), offer considerable therapeutic potential in transplantation and other immune-mediated disorders. CD4⁺CD25⁺ Tregs that express the Foxp3 transcription factor have emerged as critically important for the control of autoimmunity and for the maintenance of allograft tolerance (4, 5). Recent studies demonstrate that naive peripheral CD4⁺CD25⁻Foxp3⁻ T cells can be converted into CD4⁺CD25⁺Foxp3⁺ Tregs by TGF-β in the context of TCR signaling and costimulation (6). Indeed, there is a consensus that TGF-β may be indispensable for the development and maintenance of Tregs in the periphery (7). TGF-β is a multifunctional cytokine that plays a crucial role in fundamental cellular functions, including differentiation proliferation, migration, and survival (8). In the context of an inflammatory cytokine milieu, TGF-β supports the de novo differentiation of naive CD4⁺ T cells toward pathogenic IL-17–producing Th cells (Th17) (9, 10). Thus, control of proinflammatory cytokines plays a key role in the induction of Tregs.

Rapamycin is a macrolide antibiotic with potent anti-inflammatory, immunosuppressive, and immunoregulatory properties. The mammalian target of rapamycin has an important role in modulating innate and adaptive immunity (11). It has been observed that rapamycin can selectively expand murine and human CD4⁺CD25⁺Foxp3⁺ Tregs, while at the same time inhibiting Teffs or at least preventing their expansion (12–14). Moreover, binding of rapamycin to the immunophilin FK506 binding protein-12 (FKBP-12) abolishes FKBP-12–induced blockade of TGF-βR–mediated signaling events, suggesting a possible signaling link between rapamycin and TGF-β (15). Rapamycin can also program activated lymphocytes to produce TGF-β (16). Thus, the immunosuppressive effects of rapamycin may be mediated in part by TGF-β. We hypothesized that adjunctive treatment of allograft recipients with rapamycin and TGF-β would 1) block the production of proinflammatory cytokines (such as IL-6), inhibit the differentiation of Th17 cells, and suppress proliferative responses of activated Teff, and 2) promote the generation of Foxp3⁺ Tregs and thus strengthen the cadre of immuno-Tregs, shifting the balance from immunity to tolerance.

Most cell types secrete TGF-β in a biologically inactive form that is composed of a TGF-β dimer in association with the latency-associated protein (17, 18). rTGF-β1 is also secreted in a latent form that can be activated by acidification (19). Active TGF-β1 is a cleaved 24-kDa homodimer. Such small proteins are usually cleared rapidly from the circulation, with a t_1/2 of minutes to hours (19–22). Several studies have shown that three cysteine (Cys) residues located in the pro region of the TGF-β1 precursor can form interchain disulfide bonds that prevent the release of the mature, active TGF-β1 (19, 23, 24). To produce a long-lasting, biologically active form of TGF-β1, we changed three Cys codons in the pro region of human TGF-β1 precursor into serine (Ser) codons. We then genetically fused the mutant TGF-β1 cDNA with human IgG4 Fc to produce an autoactive TGF-β1/Fc immunoligand that does not depend on acidification for activation. The Ig component extends the circulating t_1/2 to 32 h following systemic administration. In this study, we report on the construction and use of the mutant TGF-β1/Fc fusion protein in combination with rapamycin for the induction of Tregs and the promotion of pancreatic islet transplant tolerance.

Male C57BL/6 (B6; H-2^b), B6.Foxp3^GFP knock-in (H-2^b), B6.congenic for CD45.1(H-2^b; CD45.1⁺), B6D2F1 (H-2^b,d; CD45.2⁺), B6AF1 (H-2^{b/d. k}), DBA/2 (H-2^d), and C3H (H-2^K) mice were purchased from The Jackson Laboratory (Bar Harbor, ME) and were maintained under the pathogen-free conditions at the University of Pittsburgh Central Animal Facility.

Human TGF-β1 cDNA was amplified by PCR from the cDNA library of activated human PBMC using TGF-β1–specific synthetic oligonucleotide primers. The primers have a NotI site incorporated into the 5′ end and a BglII site at the 3′ end, and their sequences are as follows: 5′-GTTAGCGGCCGCCACCATGGCCGCCCTCCGGGCTGC-3′ (forward) and 5-′TACTAGATCTCTGCACTTGCAGGAGCGCACGAT-3′ (reverse). Human Fcγ4 cDNA was generated from an IgG4-secreting hybridoma (HB-8636; American Type Culture Collection, Manassas, VA) using RT MMLV-RT and synthetic oligo (dT) oligonucleotides (Life Technologies, Grand Island, NY). The primers are as follows: 5′-AATCGGATCCCAAATATGGTCCCCCATGCCCA-3′ (forward) and 5-′AACTCTAGACTCATTTACCCAGAGACAGGGAGA-3′ (reverse). To generate an autoactive TGF-β1/Fc construct, oligonucleotide site-directed mutagenesis was used to substitute three Cys residues (Cys-33, Cys-223, and Cys-225) in the pro region of the TGF-β1 precursor with serine residues to prevent the formation of disulfide bonds (23). Ligation of mutant TGF-β1 and Fcγ4 components in the correct translational reading frame at the unique BglII site yielded a 1887-bp long open reading frame encoding a 613-aa polypeptide. The mature secreted TGF-β1/Fc was predicted to possess the mutated TGF-β1 precursor complex and a molecular mass of 190 kDa, exclusive of glycosylation (Fig. 1).

The appropriate genetic construction of TGF-β1/Fc sequences was confirmed by DNA sequence analysis after cloning of the fusion genes as NotI-XbaI cassettes into the eukaryotic expression plasmid pRc/CMV (Invitrogen, San Diego, CA). The plasmid was transfected into Chinese hamster ovary (CHO) cells by electroporation and selected by G418. High-yield clones were selected and cultured in serum-free medium. TGF-β1/Fc fusion protein was then purified from culture supernatants by protein A-sepharose affinity chromatography, followed by dialysis against PBS and 0.22-μm filter sterilization. Purified protein was stored at −20°C before use (25). Supernatants of the transfected CHO cells yielded ∼0.5 μg/ml TGF-β1/Fc, and the endotoxin level was <0.01 endotoxin units/μg of fusion protein.

Western blot analysis was performed following SDS-PAGE under reducing (with DTT) and nonreducing (without DTT) conditions using anti-human TGF-β1 mAb and anti-human IgG Fc polyclonal Ab (Pierce, Rockford, IL).

TGF-β1/Fc biological activity was accessed using the IL-4–dependent HT-2 cell (CRL-1841; American Type Culture Collection) growth inhibition assay. HT-2 cells (1 × 10⁵ cells/well) were seeded in 96-well flat-bottom plates in complete medium (RPMI 1640 supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, 10% FBS, 2 mM l-glutamine and 50 μM 2-ME) with recombinant human IL-4 (rIL-4, 10 ng/ml; BD Pharmingen, San Diego, CA). After 30 min, varying concentrations of TGF-β1/Fc (0.16–2.56 nM) or commercially supplied rTGF-β1 (0.05–0.8 nM, R&D Systems, Minneapolis, MN), were added for an additional further 24 h incubation. The cells were pulsed with [³H]thymidine (1 μCi/well) for the final 6 h and [³H]thymidine incorporation measured as cpm in a liquid scintillation counter (Perkin Elmer, Waltham, MA). Percent inhibition of proliferation was determined using the following formula: ([1 – mean cpm of the treated group ÷ mean cpm of the untreated control] × 100%).

The ability of TGF-β1/Fc to activate the Smads pathway was analyzed using Western blot. Aliquots of IL-4–stimulated HT-2 cells were seeded at 2 × 10⁶ cells per 100-mm dish and treated with TGF-β1/Fc (2 μg/ml) or rTGF-β (2 ng/ml), or human IgG4 (2 μg/ml) for 24 h. The levels of phosphorylated Smad2 (pSmad2) and total Smad2 were determined in protein extracts from cell lysates with phospho-Smad2 (Ser465/467) Abs and Smad2 Abs (Cell Signaling, Boston, MA). Thirty micrograms of proteins were loaded in each lane; protein content was determined by the BCA method (Pierce).

The serum concentration of TGF-β1/Fc was determined over time after a single bolus i.v. injection of the fusion protein (0.1 mg) to four 10-wk-old C57BL/6 mice. Serial 100-μl retro-orbital blood samples were obtained at 5 min, 1 h, 5 h, 8 h, 24 h, 48 h, 72 h, and 96 h after injection. A sandwich ELISA was performed using mouse anti-human TGF-β1 mAb as the capture Ab, and a biotin-conjugated mouse anti-human IgG4 mAb as the detection Ab (BD Pharmingen), thus ensuring that the assay was specific for the TGF-β1/Fc fusion protein and not for TGF-β1 or IgG4.

Single-cell suspensions were prepared from spleens and lymph nodes, and RBCs were removed using RBC lysis buffer (Qiagen, Valencia, CA). CD4⁺ or CD8⁺ T cells were isolated by CD4⁺ or CD8⁺ T cell enrichment columns according to the manufacturer’s instructions (Miltenyi Biotec, Auburn, CA). CD4⁺CD25⁻ and CD4⁺CD25^hi populations were sorted using a fluorescence-activated cell sorter (FACS Aria; BD Biosciences, San Diego, CA). To assess T cell proliferation, CD4⁺ or CD8⁺ T cells from C57BL/6 mice (2 × 10⁵/well) were stimulated with plate-bound anti-CD3 (2 μg/ml) and soluble anti-CD28 (2 μg/ml) mAb in 96-well flat-bottom plates for 72 h. Different concentrations of rapamycin (5, 10, 25, 50, and 100 ng/ml; Sigma-Aldrich, St. Louis, MO) and/or TGF-β1/Fc (1, 5, and 10 μg/ml) were added at the start of culture. To quantify suppressive cell function, flow-sorted CD4⁺CD25⁻ naive B6AF1 responder T cells (1 × 10⁵ per well) were cultured with various numbers of CD4⁺CD25^hi T cells from naive or tolerant B6AF1 mice in round-bottom 96-well plates, using anti-CD3 and anti-CD28 mAb (each 2 μg/ml) or irradiated (3000 rad) splenocytes (5 × 10⁴ per well) from donor (DBA/2) or the third-party (C3H) as stimulators. During 72 h or 6 d MLR, the cells were pulsed with [³H]thymidine (1 μCi/well) for the final 8 h and [³H]thymidine incorporation measured as described above.

For in vitro Foxp3⁺ induction, flow-sorted CD4⁺GFP⁻ cells from naive B6.Foxp3^GFP knock-in mice were cultured in 96-well plates (2 × 10⁵ per well), and stimulated with anti-CD3 and anti-CD28 mAb (each 2 μg/ml) for 72 h or with LPS-matured DBA/2 bone marrow-derived dendritic cells (DCs) for 7 d, in the presence of TGF-β1/Fc (5 μg/ml), rapamycin (5 ng/ml), or TGF-β1/Fc plus rapamycin. The cells were analyzed for GFP (Foxp3) expression by flow cytometry. In some experiments, IL-6 and IL-17 levels in culture supernatant were measured by ELISA using a mouse IL-6 and IL-17 ELISA kit (R&D Systems).

For in vivo Treg conversion, flow-sorted CD4⁺CD25^– T cells (1 × 10⁷) from naive congenic B6.CD45.1 mice were injected i.v. into semiallogeneic nonirradiated B6D2F1 mice. After cell transfer, B6D2F1 hosts were treated on days 0, 1, and 2 with TGF-β1/Fc (0.1 mg, i.p.) or rapamycin (0.6 mg/kg, i.p.) alone or with both agents. On day 4, lymph node and spleen cells harvested from B6D2F1 hosts were stained for surface markers (CD45.1 and CD4) and intracellular Foxp3 and then analyzed by flow cytometry, gating on CD45.1⁺CD4⁺cells.

B6AF1 mice were rendered diabetic by a single i.p. injection of streptozotocin (225 mg/kg), 1 wk prior to transplantation. Approximately 400 DBA/2 islets were transplanted under the renal capsule of each diabetic recipient mouse as previously described (26). Recipients were treated i.p. with TGF-β1/Fc and/or rapamycin as follows: 0.1 mg TGF-β1/Fc every other day for 10 d, 0.3 mg/kg rapamycin daily for the first 7 d after transplantation, and then every other day for 7 d. The recipients with long-term islet allograft survival (150 d) received a second DBA/2 or C3H islet allograft in the contralateral kidney 1 wk after the removal of kidney bearing islet allograft. Allograft function was assessed by monitoring blood glucose levels daily for the first 14 d, then twice weekly, with rejection defined as blood glucose levels of >300 mg/dl on two consecutive measurements.

Cells harvested from lymph nodes, spleens, or peripheral blood of host mice, as well as cultured cells, were analyzed for the expression of various cell surface or intracellular markers using an LSRII flow cytometer (BD Biosciences). Data were analyzed using FlowJo software (Tree Star, Ashland, OR).

Total RNA was extracted from draining lymph nodes or islet allografts using RNeasy mini-Kit (Qiagen, Valencia, CA) and reverse transcribed to cDNA using SuperScript III RT-kit (Invitrogen). Specific mRNA levels were quantified by real-time PCR using the ABI 7500 Sequence Detection System (Applied Biosystems, Foster City, CA). Gene-specific primers and probes for Foxp3, IL-6, and IL-17 were purchased from Applied Biosystems. Data are expressed using the C_T method and normalized to the housekeeping gene GPDH.

The kidney containing the islet graft was removed at day 8 or 150 after transplantation and embedded in OCT compound, snap frozen in liquid nitrogen, and stored at −80°C until sectioning. For morphologic evaluation, cryostat sections were fixed in methanol and stained with H&E. For immunohistochemical staining, sections were stained by a four-layer peroxidase–antiperoxidase method involving overnight incubation with rat anti-mouse mAb to CD4 (H129.19; BD Pharmingen) or α-smooth muscle actin (α-SMA, 1A4; Sigma-Aldrich). Samples were evaluated in a blinded fashion, using two to three different levels of sectioning.

Values are expressed as mean ± SD. Analyses for statistically significant differences were performed using the Student t test and one-way ANOVA test. The influence of different treatments on islet allograft survival was analyzed using a Log-Rank test; p < 0.05 was considered significant.

To confirm the molecular size and the cytokine–isotype specificity of TGF-β1/Fc (Fig. 1), the affinity-purified fusion protein was characterized by Western blot analysis following SDS-PAGE. As shown in Fig. 2, the TGF-β1/Fc migrated under reducing conditions as two species of molecular size 95 kDa and 45 kDa (Fig. 2, lane 2), indicating that there are two forms of TGF-β1 fused to IgG4 Fc components, the mutant pro–TGF-β1 (large form) and the mature chain of TGF-β1 (small form), which corresponds with a previous report that the rTGF-β1 is secreted as a precursor complex (24). Under nonreducing conditions, the TGF-β1/Fc ran as three species with sizes of 190 kDa, 140 kDa, and 90 kDa (Fig. 2 a–c, lane 3) indicating a mixture of three forms of TGF-β1/Fc fusion proteins. The 190-kDa or 90-kDa species represent the homodimeric fusion proteins containing mutant pro–TGF-β1 or mature TGF-β1, respectively. The 140-kDa species represents a homodimer composed of a pro–TGF-β1 and a mature TGF-β1. Moreover, the TGF-β1/Fc fusion proteins were bound by both anti–TGF-β1 mAb and anti-IgG Fc polyclonal Abs, confirming the cytokine and isotype specificity of the TGF-β1 moiety and Fcγ4 domain, respectively.

The biologic activity of the TGF-β1/Fc was determined by growth inhibition of HT-2 cells. As shown in Fig. 3A, TGF-β1/Fc inhibited IL-4–stimulated HT-2 cell proliferation in a dose-dependent manner, similar to the function of active rTGF-β1, although the molecules were not compared on a mole-for-mole basis. We further investigated the influence of TGF-β1/Fc on Smad pathway activation using Western blot with specific Abs. In HT-2 cells, incubation for 24 h with either TGF-β1/Fc or rTGF-β1, but not with IgG4, significantly increased the expression of pSmad2 at the protein level (Fig. 3B), supporting an essential role for TGF-β/Smad signaling cascade activation in TGF-β1/Fc-mediated cellular responses.

The circulating t_1/2 of TGF-β1/Fc after a single i.v. bolus was 32 h, whereas the t_1/2 of active TGF-β1 is 3 min. The TGF-β1/Fc concentration decreased in a biphasic manner, with an initial clearance of 48 h, followed by a slower, terminal component of 96 h (Fig. 4). The serum TGF-β1/Fc concentration was 0.9 ± 0.2 μg/ml at 96 h. Thus, the fusion protein exhibits a protracted in vivo t_1/2.

To examine the influence of TGF-β1/Fc combined with rapamycin on the proliferation of alloreactive T cells, CD4⁺ or CD8⁺ T cells were stimulated with anti-CD3 and anti-CD28 mAbs. Rapamycin exerted a dose-dependent inhibitory effect on T cell proliferation (Fig. 5A). A concentration of rapamycin (5 ng/ml) that reduced proliferation by ∼50% was used in subsequent in vitro experiments. TGF-β1/Fc alone at 1, 5, or 10 μg/ml did not significantly inhibit T cell proliferation. However, combination of TGF-β1/Fc and rapamycin resulted in stronger suppression of CD4⁺ and CD8⁺ T cell proliferation than that achieved with rapamycin or TGF-β1/Fc alone (Fig. 5B, 5C), suggesting an interactive inhibitory effect of TGF-β1/Fc and rapamycin on T cell proliferation.

Recent studies have shown that both TGF-β and rapamycin can induce CD4⁺Foxp3⁺ Tregs from naive T cells (6, 27), but the influence of combined TGF-β1/Fc and rapamycin exposure on Treg conversion remains unclear. To investigate its influence on de novo induction of peripheral Foxp3⁺ T cells in vitro, flow-sorted naive CD4⁺GFP⁻ T cells (>99% pure) were stimulated with anti-CD3 and anti-CD28 mAbs for 72 h. TGF-β1/Fc alone induced 16 ± 4.2% GFP⁺(Foxp3⁺) T cells from naive CD4⁺GFP⁻ T cells. A similar conversion rate was observed when rTGF-β1 was added to the cultures (15–20%; data not shown). Rapamycin alone also promoted the induction of ∼7 ± 2.4% Foxp3⁺ T cells. When TGF-β1/Fc and rapamycin were added together to the cultures, the conversion rate increased up to 30 ± 4.5% (Fig. 6A).

There is evidence that TGF-β mediates the reciprocal differentiation of naive T cells to either Tregs or Th17 cells depending on the cytokine milieu and that IL-6 plays an important role in the induction of Th17 cells (10). We investigated whether the combination of TGF-β1/Fc and rapamycin might suppress inflammatory conditions and promote Treg differentiation. Purified naive CD4⁺GFP⁻ T cells were cultured with LPS-matured DCs for 7 d to generate an alloimmune response and an inflammatory cytokine milieu. IL-6 levels in the culture supernatants were reduced by TGF-β1/Fc or rapamycin treatment alone and more markedly by combined treatment (p < 0.01) compared with untreated cultures. In contrast, IL-17 production was increased by TGF-β1/Fc alone, whereas combined treatment reversed this effect and reduced the level of IL-17 significantly (p < 0.01; Fig. 6C). In these cultures, untreated naive CD4⁺ T cells expressed little GFP(Foxp3). The addition of rapamycin slightly increased Foxp3 expression to 7 ± 1.2%. With TGF-β1/Fc alone, the proportion of CD4⁺Foxp3⁺ cells increased to 25 ± 4.5%, whereas combined treatment further enhanced the induction of CD4⁺Foxp3⁺ cells to 41 ± 3.2% (Fig. 6B).

We compared the influence of TGF-β1/Fc and rapamycin on in vivo conversion of naive CD4⁺ T cells into Tregs. Flow-sorted CD4⁺CD25⁻ T cells (>99% pure) from congeneic B6.CD45.1 mice were adoptively transferred into semiallogeneic B6D2F1 mice, and the recipients were treated with TGF-β1/Fc, rapamycin, or both for 3 d. In hosts treated with rapamycin alone, the incidence of CD4⁺Foxp3⁺ among the CD45.1⁺CD4⁺ T cells in the spleen was 20 ± 3.1%. TGF-β1/Fc treatment alone resulted in 12 ± 2.7% Foxp3⁺ T cells. However, combination of TGF-β1/Fc and rapamycin resulted in ∼25 ± 2.3% of CD45.1⁺CD4⁺ T cells expressing Foxp3 (Fig. 6D). These data indicate that TGF-β1/Fc and rapamycin exert an additive effect on the induction of CD4⁺Foxp3⁺ Tregs during naive T cell differentiation in the context of their inhibitory effect on inflammatory cytokine production.

To determine the influence of TGF-β1/Fc and rapamycin on allograft survival, MHC-mismatched DBA/2 islet allografts were transplanted into diabetic B6AF1 mice. As shown in Fig. 7A, untreated recipients rejected their grafts in a mean survival time (MST) of 19 d (n = 6). Administration of TGF-β1/Fc resulted in delayed islet allograft rejection (MST, 61 ± 21 d; n = 5; p = 0.008). Engraftment was prolonged in recipients treated with rapamycin (MST, 85 ± 15 d; n = 9; p = 0.001). In contrast, combined treatment with TGF-β1/Fc and rapamycin resulted in indefinite allograft survival in 90% of recipients (MST, >120 d; n = 11; p < 0.001) and proved more effective in promoting engraftment than the two reagents used separately (p < 0.05). Surgical removal of kidneys bearing islet allografts from three combined treatment recipients 150 d after transplantation led to hyperglycemia, demonstrating that the euglycemic state was maintained by the islet allograft. Two nephrectomized recipients received a second DBA/2 islet allograft under the contralateral renal capsule in the absence of further immunosuppressive treatment, and both accepted the second transplant for >100 d, demonstrating a state of tolerance. One recipient received a third-party C3H islet allograft that was rejected at day 11 after transplant, thereby confirming that the unresponsiveness was donor specific and that the host immune response was intact (Fig. 7B).

To determine whether short-term combined TGF-β1/Fc and rapamycin treatment resulted in long-term, stable expansion of CD4⁺Foxp3⁺ Tregs in graft recipients, we analyzed the expression of Foxp3. The results revealed that a population of CD4⁺Foxp3⁺ Tregs in PBMCs expanded (15.8 ± 2.1% versus 6.4 ± 0.2%; p < 0.05) and that Foxp3mRNA expression in draining lymph nodes and islet allografts was also enhanced significantly (p < 0.05) in tolerant recipients given combined treatment compared with naive mice (Fig. 8A, 8B). To ascertain whether the expanded Tregs retained the ability to suppress proliferation of Teffs, we tested the ability of flow-sorted CD4⁺CD25^hi cells (>99% pure) from the lymph nodes and spleens of tolerant B6AF1 mice to suppress the function of naive B6AF1 CD4⁺CD25⁻ cells stimulated polyclonally with anti-CD3 and anti-CD28 mAb. CD4⁺CD25^hi cells alone failed to proliferate after stimulation, consistent with the anergic properties ascribed to Tregs, whereas CD4⁺CD25^hi Tregs from tolerant mice markedly inhibited the naive Teff proliferation at a suppressor/responder ratio of 1:4 and exhibited much more potent ability to suppress proliferation than naive CD4⁺CD25^hi Tregs (63.4 ± 2.3% versus 9.8 ± 1.8%; p < 0.001). We further examined whether the regulatory function of the induced CD4⁺CD25^hi Tregs was related to Ag-specific triggering. Naive CD4⁺CD25⁻ T cells were stimulated with irradiated splenocytes from either the same donor (DBA/2) or third party (C3H) mice. The CD4⁺CD25^hi Tregs from tolerant mice significantly inhibited the proliferation of autologous CD4⁺CD25⁻ T cells activated by the same donor Ag in MLR, but could not suppress third party Ag-induced alloreactivity (67.7 ± 2.2% versus 35.5 ± 9.7%; p < 0.05; Fig. 8C).

Foxp3, IL-6, and IL-17 mRNA expression was analyzed in islet allografts and draining lymph nodes by quantitative real-time PCR, and serum levels of IL-6 and IL-17 were measured by ELISA on day 8 after transplantation. As depicted in Fig. 9A–C, combined TGF-β1/Fc and rapamycin markedly upregulated Foxp3 gene expression and profoundly downregulated IL-6 and IL-17 mRNA expression in the islet grafts compared with the untreated or TGF-β1/Fc-treated groups (p < 0.01). Likewise, in the lymph nodes draining the grafts, combined treatment significantly enhanced Foxp3 gene transcripts and decreased IL-6 mRNA expression (p < 0.01). IL-17 mRNA could not be detected in lymph node samples. These changes were accompanied by a reduction in serum levels of IL-6 and IL-17. Rapamycin alone reduced IL-6 expression in both the grafts and draining lymph nodes, and it enhanced Foxp3 gene expression in draining lymph nodes, but not in the grafts.

Histopathologic changes were assessed in H&E-stained allograft sections at day 8 after transplant. A dense mononuclear cell infiltrate of the islet grafts was observed in untreated recipients (Fig. 10A) with less intense infiltration in TGF-β1/Fc- or rapamycin-treated recipients (Fig. 10B, 10C). In the combined treatment group, cellular infiltration was dramatically reduced and islet architecture was well preserved (Fig. 10D). As judged by immunohistochemical analysis, abundant CD4⁺ cells were present in allografts of untreated hosts (Fig. 10E), whereas a significant decrease in CD4⁺ cell infiltration was noted in recipients given TGF-β1/Fc and rapamycin combination therapy (Fig. 10F). This result is in keeping with the hypothesis that TGF-β1/Fc and rapamycin act in concert to inhibit the alloreactivity of Teffs. To evaluate whether treatment strategies including short-term use of TGF-β1/Fc promoted fibrosis in allografts of combined treatment long-term graft recipients, we examined immunoreactivity for α-SMA, a valuable marker in determining activated myofibroblasts, which has been shown to be well correlated with progressive fibrosis in various organs (28–30). Only a few dim α-SMA⁺ cells were observed in the vessel walls and interstitial areas of the kidney, and no α-SMA⁺ cells were detected in grafted islets at day 150 after transplant (Fig. 10G), similar to that observed in untreated mice (Fig. 10H).

CD4⁺Foxp3⁺ Tregs have emerged as important factors in promoting tolerance of solid organ transplants (5, 31, 32). Since the elucidation of the immunoregulatory properties of TGF-β1 in Treg-mediated immune tolerance (6, 7), it has been anticipated that TGF-β1 might be a potentially important therapeutic adjunct for the treatment of transplant rejection and autoimmune disease. Nevertheless, the latent form of most cell-secreted precursor TGF-β1 and the short-circulating t_1/2 of biologically active mature TGF-β1 are major limitations for its in vivo application (18, 22). Previous studies (19, 24) have revealed that CHO cells express wild type simian TGF-β1 in precursor complex form, consisting of pro–TGF-β1, the pro region of the precursor, and mature TGF-β1. Three Cyss located in the pro region of the TGF-β1 precursor influence the maturation and activation of TGF-β1. It has been demonstrated that substitution of Cys 33 with a serine residue results in the generation of a more mature TGF-β1 dimer, and that substitution of both Cys-223 and Cys-225 results in the production of only monomeric precursor forms, yet mature TGF-β1 is still able to form a bioactive dimer (23, 24). In this study, we constructed a mutant human TGF-β1/Fc immunoligand using genetic engineering. Although the purified mutant TGF-β1/Fc was found to be a mixture of three forms of fusion proteins, both the mutant pro–TGF-β1– and the mature TGF-β1– constituted TGF-β1/Fc fusion proteins could yield biologic activity as in previous studies (23). Our results indicate that the mutant TGF-β1/Fc inhibits IL-4–dependent HT-2 cell proliferation in a dose-dependent manner and induces activation of the receptor-regulated Smad2 pathway (33), confirming the biologic function, intracellular signal transduction, and specificity of the TGF-β1 moiety in the TGF-β1/Fc fusion protein. Moreover, the Fc fragment ensured TGF-β1/Fc a greatly prolonged plasma t_1/2 of 32 h. Therefore, TGF-β1/Fc fusion protein can circulate and exert its activity for a greatly extended period following a single injected dose. Although human IgG4 has been demonstrated to be an isotype with relatively ineffective complement-dependent cytotoxicity and Ab-dependent cellular cytotoxicity (34, 35), the neutral effect status of IgG4 may vary because of population FcγR polymorphisms (36). Thus the specific mutations in Fcγ4 region may further eliminate Fc effector functions to guarantee the lack of cytolytic capacity of TGF-β1/Fc (37).

Both TGF-β and rapamycin inhibit T cell proliferation (38, 39). Our results show that TGF-β1/Fc and rapamycin can act together to inhibit the proliferation of both CD4⁺ and CD8⁺ T cells more effectively in a dose-dependent manner. This suggests potential interaction between TGF-β1/Fc- and rapamycin-sensitive mammalian target of rapamycin signaling pathways in T cell cycle progression. Evidence has accumulated that the control of Treg and Th17 cells may be interlinked (6, 9, 10). Thus, we examined the combined effects of TGF-β1/Fc and rapamycin on Treg and Th17 differentiation. In either mature DC- or anti-CD3/CD28 mAb-stimulated in vitro culture systems, both TGF-β1/Fc and rapamycin independently induced Foxp3⁺ Treg differentiation from naive CD4⁺Foxp3⁻ cells, although the ability of rapamycin was lower than that of TGF-β1/Fc. However, a combination of the two agents promoted significantly more Foxp3 expression, demonstrating the combined influence of TGF-β1/Fc and rapamycin on de novo induction of Tregs. This effect may reflect the ability of rapamycin to stimulate TGF-β production (16, 40) and to suppress cell cycle progression, as Teffs undergoing extensive proliferation fail to express Foxp3 (41). Moreover, in the mature DC–T cell culture environment, IL-6 (a major cytokine produced by LPS-stimulated DCs) (42), can inhibit Treg generation and divert T cell differentiation to Th17 cells (10), an effect attributed to the activation by IL-6 of the transcription factor STAT3 (43). Because rapamycin can inhibit IL-6 signal transduction (44), blockade of IL-6 activity may be another important mechanism by which rapamycin favors Treg induction. Our results show that rapamycin decreases levels of IL-6 and IL-17 under these culture conditions. Although TGF-β1/Fc also inhibits IL-6, it enhances IL-17 production under proinflammatory conditions. We propose that the regulatory influence of TGF-β1/Fc on immune responses is complex and depends on the extent of inflammatory stimulation. If an excess of IL-6 overrides the immunosuppressive action of TGF-β1/Fc, then the cytokine milieu will support the predominance of TGF-β–mediated Th17 cell differentiation. Therefore, our observation that combined treatment with TGF-β1/Fc and rapamycin was more effective in decreasing IL-6 and IL-17 production suggests that the cooperative inhibitory effects on inflammatory responses may ultimately modulate the balance between Treg and Th17 cell generation. In this study, we have also demonstrated that TGF-β1/Fc and rapamycin act in concert in vivo to promote the conversion of naive CD4⁺CD25⁻ T cells into CD4⁺Foxp3⁺ Tregs in a model similar to graft versus host disease, in which transferred alloreactive CD4⁺Foxp3⁻ T cells respond to host alloantigen and proliferate (27); however, TGF-β1/Fc by itself does not cause significant Treg conversion in vivo, which was different from that shown in vitro. This finding may reflect the increase in endogenous IL-6 under acute in vivo immuno-inflammatory conditions; however, concomitant therapy with rapamycin may abrogate the increase in IL-6 (45) and favor skewing toward TGF-β–driven Treg induction.

On the basis of our observations, we assessed the ability of combined TGF-β1/Fc and rapamycin to promote transplantation tolerance in a murine MHC-mismatched pancreatic islet allograft model. A 14-d course of combined treatment achieved permanent survival of the islet allograft in most treated recipients (90%) in addition to donor- or Ag-specific immune tolerance exhibited 150 d after transplantation. These mice accepted second donor but not third-party islet allografts. Furthermore, combined treatment yields an improvement on the effect of TGF-β1/Fc or rapamycin monotherapy, by which only slight to moderate prolongation of islet allograft survival was observed in recipients, indicating that the combined strategy was tolerogenic and that each of the components was required for optimal therapeutic effect. Considering that TGF-β1 is required more for the induction phase of Treg generation than for the maintenance phase (46), and that an excess of TGF-β1 could increase the risk of tissue fibrosis (28), we used a regimen comprising only five doses of TGF-β1/Fc over the first 10 d after transplantation, in combination with rapamycin, to initiate Foxp3⁺ T cell differentiation and allow Treg-mediated tolerance to take place (47). Importantly, this protocol did not result in unwanted tissue fibrosis, which we investigated by α-SMA staining in allograft sections 150 d after transplantation, and it might be attributed to the potential antifibrotic effect of low-dose rapamycin (48, 49). Peripheral blood, draining lymph nodes, and islet allografts of long-term tolerant recipients exhibited an elevated expression of Foxp3 relative to naive animals. Moreover, the CD4⁺CD25^hi Tregs purified from tolerant recipients exerted a more potent suppressive effect on naive Teff proliferation following polyclonal stimulation. This finding suggests that expansion of Foxp3⁺ Tregs and maintenance of their suppressive potency may at least in part underlie allograft tolerance induced by the combined therapy. Furthermore, these CD4⁺CD25^hi Tregs potently suppressed the proliferation of naive Teffs against donor Ag, not third party alloantigen in MLR, indicating their alloantigen-specificity and their role in the observed donor-specific tolerance.

Active immune suppression by Foxp3⁺ Treg is a pivotal mechanism underlying peripheral T cell tolerance, whereas inflammation of local tissue during transplantation not only limits Treg suppression but also promotes proinflammatory Th17 responses (50, 51). Our data obtained on day 8 after islet transplantation revealed that the combined therapy markedly increased Foxp3 and decreased IL-6/IL-17 gene expression in both islet allografts and draining lymph nodes, accompanied by a reduction in serum IL-6 and IL-17 levels and a decrease in islet allograft CD4⁺ T cell infiltration. This finding strongly suggests that TGF-β1/Fc acts in concert with rapamycin to inhibit IL-6–mediated Th17 responses and to provide conditions that favor Foxp3⁺ Treg generation in the “battlefield” of the alloimmune response. The potential benefit of this combined therapy is also indicated by the recruitment of Foxp3⁺ Tregs to the allografts and draining lymph nodes, in which the continuing presence of Tregs may ensure effective control of antidonor reactivity and promote allograft tolerance.

This immunoligand is as effective biologically as active rTGF-β1, has negligible potential for immunogenicity, and acts in conjunction with rapamycin to inhibit both CD4⁺ and CD8⁺ T cell proliferation. We also demonstrate for the first time that TGF-β1/Fc and rapamycin, especially in combination, promote the de novo generation of Foxp3⁺ Tregs in vitro and in vivo, while simultaneously inhibiting IL-6–mediated Th17 cell differentiation. Furthermore, short-term combined treatment with TGF-β1/Fc plus rapamycin achieves donor-specific tolerance in a mouse model of islet transplantation, accompanied by Foxp3⁺ Treg expansion and their improved alloantigen-specific suppressive function. The concerted influence of these two agents on Foxp3, IL-6, and IL-17 transcripts and the trafficking of Tregs after transplantation suggest an important underlying mechanism of tolerance induction.

Disclosures The authors have no financial conflicts of interest.