Rabbit Polyclonal Mouse Antithymocyte Globulin Administration Alters Dendritic Cell Profile and Function in NOD Mice to Suppress Diabetogenic Responses ¹

Type 1 diabetes is an autoimmune disorder caused by Th1-mediated cellular destruction of insulin-producing pancreatic beta cells (1, 2). Dendritic cells (DC)³ are the most potent APCs of the immune system, with the unique capacity to initiate and to modulate effector T cell responses. DC prime a Th1- vs Th2-biased T cell repertoire, depending on complex factors such as the DC subset, costimulatory molecule expression, and cytokine milieu, during both Ag presentation and subsequent T cell activation (3, 4). Alternatively, DC can induce regulatory T (Treg) cells and cross-present beta cell Ags in a tolerogenic fashion to autoreactive T cells in the pancreatic lymph nodes (PLN) (5, 6, 7). Abnormalities in DC function, including defects in maturation and ability to activate immunoregulatory T cells, have been demonstrated in NOD mice, and may play an important role in eliciting the characteristic autoimmune reaction to beta cells (8, 9, 10). Hence, manipulation of DC profile or function, with the purpose of reshaping the repertoire of T cells, represents a potential therapeutic option for the treatment of type 1 diabetes (11, 12, 13, 14, 15, 16, 17, 18).

Polyclonal rabbit antithymocyte globulin (ATG) has been used in a wide variety of clinical settings, including renal transplantation, graft-versus-host disease, and aplastic anemia (19, 20). ATG binds to multiple epitopes on the surface of T cells and induces a rapid lymphocytopenia by several mechanisms including complement-dependent cytolysis, cell-dependent phagocytosis, and apoptosis (21, 22). Recent clinical studies demonstrating the reversal of type 1 diabetes following therapy with more “restrictive” anti-CD3 mAbs have provided the impetus for planned clinical studies testing the efficacy of ATG in this disorder (23, 24). Moreover, the recent development of rabbit polyclonal mouse ATG (mATG) has allowed for more precise study of the potential mechanisms of action for this agent. Recently, we demonstrated the ability of mATG to attenuate the development of type 1 diabetes through a process that mechanistically was associated with an increased frequency of CD4⁺CD25⁺FoxP3⁺ T cells and enhanced immunoregulatory function as determined through in vitro suppression assay of effector T cell responses (25). Polyclonal mATG is prepared by immunization with pooled thymus cells, which undoubtedly contain DC. Furthermore, DC and T cells share several common cell surface Ags. Thus, Abs against DC likely exist within the mATG preparation. Given the pivotal role DC play in the generation of T cell responses responsible for type 1 diabetes, we examined the impact of mATG on DC profile and function. We report a novel mechanism that contributes to the protection from type 1 diabetes with this agent and may be active in other immune-based disorders subject to similar immunomodulatory based treatment.

Female NOD/LtJ, NOD.RAG^−/−, NOD.BDC2.5 TCR transgenic mice were purchased from The Jackson Laboratory and housed under specific pathogen-free conditions, according to institutional guidelines. Protocols were approved by the Institutional Animal Care and Use Committees of The University of Florida and Johns Hopkins University.

The mATG preparation was purified after immunizing rabbits with pooled thymus cells prepared from NOD, C3H/He, DBA/2, and C57BL/6 mice (Genzyme). Tests for quality control and quality assurance for functional activities were performed in accordance with standard procedures by the manufacturer. Twelve-week-old female NOD mice were administered i.p. injections of 1.0 mg of mATG or rabbit IgG (Jackson Immunologicals) diluted into saline or saline alone, and sacrificed 1, 7, or 14 days later. This amount of mATG was considered optimal, as it represented the minimal dose of mATG providing protection from type 1 diabetes in NOD mice in our previous studies (25).

Cells harvested from spleen, PLN, inguinal lymph node (ILN), and thymus were subjected to flow cytometric analysis using a FACSCalibur flow cytometer (BD Biosciences). Data were analyzed using the FCS Express analysis software (De Novo). The following Abs (as well as relevant isotype controls) were used for staining (murine monoclonal from BD Pharmingen, unless otherwise noted): FITC anti-CD8α (53-6.7), FITC anti-CD8α (KT15; Serotec), allophycocyanin anti-CD11c, PerCP-Cy5.5 anti-CD11b, PE anti-plasmacytoid DC Ag-1 (PDCA1; Miltenyi Biotec), FITC anti-IA^k, PE or biotin anti-CD86, and PE anti-CCR7 (eBioscience). Absolute cell numbers were calculated by multiplying the frequency of DC subset by the total cell number per organ.

Splenocytes were harvested from mice within each treatment group by collagenase D (Roche Diagnostics) digestion, with DC purified using a CD11c positive selection magnetic bead separation technique (Miltenyi Biotec) according to the manufacturer’s protocol. The purity of splenic DC was >90%. Cells were then cultured in RPMI 1640 with 10% FBS, l-glutamine, penicillin-streptomycin-neomycin, 500 U/ml murine GM-CSF, and stimulated with TLR agonists, including LPS (1 μg/ml) and CpG (ODN 1826; 5 μg/ml), or anti-CD40 (5 μg/ml) for 24 h. Supernatants were harvested for cytokine measurement (described below). To measure intracellular cytokine production, purified splenic DC from ATG-treated mice were stimulated with CpG and LPS for 24 h in the presence of GolgiPlug for the last 4 h. IL-10/IL-12 production by CD8⁺ and CD8⁻ DC was then measured by flow cytometry.

CD4⁺ T cells from NOD.BDC 2.5 TCR transgenic mice were purified by negative selection using the CD4 magnetic bead separation technique (Miltenyi Biotec). CD4⁺ T cells (1 × 10⁵) were cultured in 96-well round-bottom plates for 5 days with purified splenic DC from IgG- or mATG-treated mice at varying DC to T cell ratios (1:10, 1:20, 1:40, 1:80) in the presence of a mimetope peptide (termed 1040-55; 500 ng/ml), having the sequence RVRPLWVRME, as previously described (26). On day 4, 0.5 μCi [³H]thymidine was added to each well. Following 16 h of incubation, cells were lysed and the ³H incorporation was determined using a 1450 Microbeta Trilux Beta Scintillation counter (Wallac). In addition, 2 × 10⁵ CD4⁺ T cells were cultured with 2 × 10⁴ purified splenic DC in the presence of peptide 1040-55. Supernatants were harvested at days 2 or 5 for cytokine measurement. On day 5, cells were stimulated with 5 ng/ml PMA and 500 ng/ml ionomycin (Sigma-Aldrich) for 4 h in the presence of GolgiPlug, followed by surface staining with mAb anti-CD4 and intracellular staining with FITC anti-IFN-γ and PE anti-IL-4 using BD Cytofix/Cytoperm Plus (BD Pharmingen), per the manufacturer’s protocol. In separate experiments, splenic CD8⁺ and CD8⁻ DC were purified from naive NOD mice using flow sorting (BD VantageSE), and cultured with purified CD4⁺ T cells (with or without CFSE labeling) at a ratio of 1:10. Cytokine production was assayed by Bio-Plex. T cell proliferation was determined as a percentage of cells exhibiting CFSE fluorescence dilution.

CD4⁺ T cells from NOD.BDC 2.5 TCR transgenic mice were purified and cocultured with splenic DC from mATG- or IgG-treated NOD mice in the presence of the mimetope peptide (DC to T cell ratio, 1:5), as described. After 5 days, DC were depleted from these cultures using CD11c magnetic bead separation (positive selection), and remaining T cells (3 × 10⁶) were cotransferred with spleen cells (3 × 10⁶) from newly diabetic NOD mice by i.p. injection into NOD.RAG^−/− mice. All mice were followed daily for onset of diabetes, defined as three consecutive blood glucose levels higher than 250 mg/dl measured by blood glucose meter (Ascensia Contour).

Bone marrow-derived DC were generated as described (10). Briefly, bone marrow cells from 8- to 12-wk-old NOD mice were cultured for 5 days in RPMI 1640 medium with 10% FBS, l-glutamine, 2-ME, penicillin-streptomycin-neomycin, 500 U/ml murine GM-CSF, and 100 U/ml murine IL-4. On day 3, one-half of the supernatant was replaced with fresh cytokine-containing medium, and 1–100 μg/ml mATG was added. On day 4, 1.0 μg/ml LPS was added to stimulate DC activation/maturation. On day 5, supernatants were collected for cytokine measurement, and cells were harvested for flow cytometric analysis.

DC endocytosis was detected as previously described (27). Briefly, bone marrow-derived DC were treated with mATG on day 3 and harvested on day 5. A total of 5 × 10⁵ cells were incubated with 5 μg/ml FITC-albumin (Sigma-Aldrich) or 0.1 mg/ml FITC-dextran (m.w. 40,000; Molecular Probes) at either 37°C or 4°C (negative control) for 1 h. Endocytosis was stopped by cold wash in 0.1% sodium azide/1% FBS/PBS. Cells were stained with anti-CD11c and 7-aminoactinomycin D (7-AAD), followed by flow cytometric analysis.

Splenic DC were isolated from 8- to 12-wk-old NOD mice and cultured for 6 h with RPMI 1640 containing 10% FBS, 500 U/ml GM-CSF, l-glutamine, and penicillin-streptomycin-neomycin, in the presence of 0.01 mg/ml mATG or IgG. Cells were harvested, stained with anti-CD11c, anti-CD8, anti-PDCA1, Annexin V, and 7-AAD and analyzed by flow cytometry. Cells (gated on CD8⁺, CD8⁻, or PDCA1⁺ DC) that were Annexin V-positive and 7-AAD-negative stained were considered apoptotic, whereas double positive staining indicated necrosis.

Culture supernatants were subjected to cytokine analysis, by ELISA using OptEIA Set Mouse IL-12 and IL-10 kits (BD Biosciences), Luminex using Lincoplex platform (Linco) or Bio-Plex multiplex cytokine detection system (Bio-Rad), according to the manufacturer’s protocol.

Cells freshly isolated from NOD mice or bone marrow-derived DC were incubated with 0.01 mg/ml mATG or IgG for 20 min and washed. Those cells, as well as cells directly isolated from in vivo mATG- or IgG-treated mice, were stained with FITC-labeled donkey anti-rabbit IgG (Amersham) and allophycocyanin-labeled anti-CD11c, followed by flow cytometric analysis.

Statistical analysis was performed using Student’s t test and log rank test, where appropriate. Data are presented as mean ± SD. A value for p < 0.05 was deemed significant.

To confirm the ability of mATG to bind DC, lymphoid tissues from mATG-treated mice, as well as freshly isolated splenic DC and bone marrow-derived DC treated with mATG in vitro, were investigated. Intensive binding was observed on splenic and bone marrow DC obtained from mATG-treated mice (i.e., 1 day following injection) with less binding demonstrated on thymic DC (Fig. 1,A). No binding was detected on any DC from mATG-treated mice 14 days after injection. In vitro, mATG administration resulted in strong binding to both splenic and thymic DC, as well as bone marrow-derived DC (Fig. 1 B). These findings are consistent with our previous studies demonstrating in vivo binding of rabbit polyclonal human ATG to peripheral blood DC in human subjects (28).

Because trafficking of DC from peripheral tissue to regional lymph nodes is important to their tolerogenic function, we analyzed the DC frequency (i.e., the percentage of CD11c⁺ cells) in lymphoid organs of NOD mice 1 day following initiation of the various treatments. Administration of mATG significantly increased the DC frequency in spleen, PLN, and ILN (Fig. 2,A) compared with the IgG-treated group. To address whether this increase simply represented the relative reduction of T cells characteristic of mATG administration, the absolute DC number was calculated by multiplying the percentage of DC by the total number of cells from one organ. Results showed that the absolute number of DC was significantly increased only in PLN following mATG treatment (Fig. 2,B). In separate experiments, NOD mice were administered mATG or IgG and sacrificed 7 or 14 days later. The frequency of DC in PLN and ILN in the mATG treatment group remained significantly higher compared with the IgG group at both time points (Fig. 2,A). Although there was no significant increase in the absolute DC number in any of the lymphoid organs at these later time points, a nonsignificant trend in PLN did emerge (Fig. 2 B).

Because a particular DC subset can influence the ultimate T cell response, the three recognized mouse DC subsets were characterized in spleen, thymus, PLN, and ILN 1 day following the noted injection schemes: CD11c⁺CD8⁺CD11b^−∼lowPDCA1⁻ conventional DC (CD8⁺ DC); CD11c⁺CD8⁻CD11b⁺PDCA1⁻ conventional DC (CD8⁻ DC); CD11c⁺CD11b⁻PDCA1⁺ plasmacytoid DC (Fig. 3, A and B). Injection of mATG resulted in a significant reduction of CD8⁺ DC and increase of CD8⁻ DC frequency within the spleen, PLN, and ILN compared with the IgG-treated group (Fig. 3,C). The frequency of plasmacytoid DC was also significantly reduced in PLN and ILN of mATG-treated mice (Fig. 3,C). Analysis of absolute PLN DC subset numbers revealed that although all three subsets were significantly increased, the large majority of this increase represented CD8⁻ DC (Fig. 3,D). To determine whether mATG-induced cell death could contribute to the relative reduction of CD8⁺ DC and plasmacytoid DC frequency, purified splenic DC were treated with mATG or IgG control in vitro for 6 h, and analyzed by flow cytometry. mATG induced an increase in both apoptosis and necrosis of CD8⁺ DC but not CD8⁻ DC compared with control wells (Fig. 4). An increase in plasmacytoid DC necrosis was also demonstrated. Thus, although an absolute increase in CD8⁻ DC explains the relative reduction of CD8⁺ DC, at least in the PLN, increased cell death of other DC subsets may play a role.

Because defects in DC maturation are implicated in the pathogenesis of type 1 diabetes, the maturational state of DC among lymphoid organs in different treatment groups was assessed. mATG induced DC maturation characterized by higher CD86 and MHC class II expression in each of the aforementioned lymphoid organs except thymus 1 day following injection compared with IgG (Fig. 5). These findings were consistent for all three DC subsets (data not shown). Similar up-regulation of CCR7 expression was also demonstrated for total DC (Fig. 5). However, subgroup analysis revealed that CCR7 expression was up-regulated on CD8⁻ DC (15 ± 2.7% vs 10 ± 2.4%, p < 0.01) but not on CD8⁺ DC in PLN, supporting migration of CD8⁻ DC from the pancreas as an explanation for the increase in absolute number. No difference in CD86, MHC class II, or CCR7 expression was found between IgG- and mATG-treated groups at 7 or 14 days after injection.

To assess the cytokine production of splenic DC from mATG- or IgG-treated mice, DC were stimulated ex vivo with various activation agents. mATG treatment induced more IL-10 production from splenic DC following stimulation with CpG, LPS, and anti-CD40, and slightly reduced IL-12 production after stimulation with CpG (Fig. 6,A). No IL-12 production was detected in LPS- or anti-CD40-stimulated groups. DC subtype analysis by intracellular staining revealed significantly higher IL-10 and lower IL-12 production by CD8⁻ DC compared with CD8⁺ DC (Fig. 6 B). This cytokine pattern is consistent with a tolerogenic DC phenotype.

To address whether mATG has direct effects on DC in vitro, bone marrow-derived DC obtained from NOD mice were treated with escalating doses of mATG. No effect on DC generation (percentage of CD11c⁺ cells in live bone marrow cells) or viability (measured by 7-AAD staining) was demonstrated when added early (day 0) or late (day 3) and with or without LPS stimulation (data not shown). When added late at day 3 of culture, however, mATG did inhibit LPS-induced DC maturation, as evidenced by reduction of CD86 and MHC class II expression (Fig. 7,A). No effect on endocytosis of albumin or dextran was observed by mATG treatment (data not shown). Regarding cytokine production, mATG significantly increased LPS-stimulated IL-10 production while reducing IL-12 production (Fig. 7 B).

To investigate the ability of mATG-treated DC to modulate T cell responses, CD4⁺ T cells from BDC2.5 TCR transgenic mice were cocultured ex vivo with splenic DC from different treatment groups in the presence of BDC2.5-specific peptide. Surprisingly, splenic DC from mATG-treated mice induced more T cell proliferation compared with control DC from IgG-treated mice (Fig. 8,A). Analysis of this response by intracellular staining revealed a large shift toward induction of IL-4-producing Th2 cells (Fig. 8,B). Analysis of cytokine secretion in supernatants from these cultures confirmed the predominant Th2 cytokine pattern characterized by increased IL-4, IL-5, and IL-10 production, as well as reduced IFN-γ production (Fig. 8 C). No difference in IL-2 production was observed after culture between the two treatment groups (IgG vs mATG: 1612.64 ± 278.16 pg/ml vs 1699.53 ± 189.18 pg/ml; p = NS). Addition of IL-10-neutralizing Ab to mATG-treated DC/T culture resulted in a slight but significant increase in IFN-γ (3240.7 ± 431.8 pg/ml vs 2549.8 ± 659.6 pg/ml; p < 0.05) but no change in IL-4 or IL-5 levels (data not shown).

To determine whether alteration of the DC subset ratio following mATG treatment might contribute to the observed T cell response, purified CD8⁺ or CD8⁻ DC from naive NOD mice were cultured with CD4⁺ T cells from BDC2.5 TCR transgenic mice, as previously described. A trend toward greater proliferation was noted after stimulation with CD8⁻ DC (Fig. 9,A). Analysis of cytokine production revealed a Th2-biased repertoire, consistent with the experiments using CD8⁻ predominant mATG-treated DC (Fig. 9 B).

Because type 1 diabetes is regarded as a Th1-mediated disease, and mATG-treated splenic DC induced a Th2 response ex vivo in BDC2.5 T cell culture, we sought to determine whether this immune deviation would alter diabetogenic activities in vivo. For these experiments, BDC2.5 T cells were isolated after culture with in vivo mATG- vs IgG-treated splenic DC plus peptide and cotransferred with spleen cells from newly diabetic NOD mice by i.p. injection into NOD.RAG^−/− mice. Type 1 diabetes onset was significantly delayed in mice that received T cells stimulated by mATG- vs IgG-treated splenic DC, respectively (Fig. 10). These in vivo data support our hypothesis that mATG modulation of DC phenotype and function alters the T cell response to suppress diabetogenic responses.

Similar to earlier reports using antilymphocyte serum (29, 30), our previous study demonstrated the ability of mATG to inhibit the development of type 1 diabetes in NOD mice, however, not only by a transient reduction of lymphocytes but also through the enhancement of CD4⁺CD25⁺ Treg functional activity (25). Our current study provides in vitro and in vivo evidence for an additional novel mechanism, whereby mATG alters DC phenotype and function to shape the subsequent T cell response.

DC comprise a heterogeneous population of cells with divergent effects on the immune response. In the mouse, at least three subtypes are identified in the spleen: CD8⁺CD11b⁻PDCA1⁻ conventional DC (CD8⁺ DC), CD8⁻CD11b⁺PDCA1⁻ conventional DC (CD8⁻ DC), and CD11b⁻PDCA1⁺B220⁺ plasmacytoid DC. The CD8⁺ DC subset is considered the major source of IL-12 secretion, with the capacity to promote Th1-biased immune responses and to cross-prime CTL, whereas CD8⁻ DC produce more IL-10 and preferentially induce Th0- or Th2-like responses (31). DC subset function, however, exhibits considerable plasticity. For example, treatment of DC with IL-10 in vitro induces a selective decrease in CD8⁺ DC viability, whereas incubation with IFN-γ down-regulates the Th2-promoting capacities of CD8⁻ DC and increases the Th1 skewing properties of both subsets (3).

In the current study, in vivo mATG administration resulted in a relative reduction of the CD8⁺ DC population in various lymphoid tissues. Although the precise mechanisms by which this reduction occurred are unclear, in vitro mATG treatment of isolated splenic DC induced necrosis and apoptosis of CD8⁺ but not CD8⁻ DC, suggesting a possible direct cellular effect. However, when cultured with various activation stimuli, splenic DC from mATG-treated mice produced higher levels of IL-10, with greater production from the CD8⁻ DC subset. Similarly, bone marrow-derived DC cultured with mATG also produced increased IL-10. Thus, DC IL-10 production after mATG administration may have contributed further to selective reduction of the CD8⁺ DC subset in vivo (3). Reduction of the CD8⁺ DC population also likely underlies the ensuing Th2 response after mATG-treated splenic DC were cocultured with BDC2.5 TCR transgenic CD4⁺ cells, as purified CD8⁻ DC from naive NOD mice in this system yielded a similar Th2 deviation. Finally, the absolute DC number was increased in PLN, with the majority of this increase identified as CD8⁻ DC. Other studies have demonstrated a similar increase of this DC phenotype in PLN of NOD mice after treatment with α-galactosylceramide, which also conferred protection from type 1 diabetes. Those studies suggested that the increase in PLN CD8⁻ DC represented an increased migration from the pancreas (16, 17). However, ascribing tolerogenic function solely to one individual DC subtype is still controversial because some groups have reported that CD8⁺ DC can also induce tolerance to tissue-associated Ags (32, 33).

Traditionally, immature DC with low level expression of MHC class II and costimulatory molecules are considered important for maintenance of tolerance under basal conditions. However, recent studies have shown that tolerance induction in NOD mice may require DC to express high levels of these surface molecules (i.e., mature phenotype) for cross-tolerization of Ag-specific T cells (9, 12, 34, 35). In addition, DC migration to draining LN is essential to generate or activate Treg cells where these T cells first encounter cognate Ag. Previous studies have demonstrated that mature DC migration to the draining lymph nodes is up to 8-fold higher than immature DC (36). A recent report suggests that disruption of E-cadherin adhesions between peripheral tissue DC and neighboring cells, which occurs concomitant with tissue emigration, may be responsible for inducing DC maturation during steady-state conditions (37). DC maturation in this study was characterized by typical up-regulation of costimulatory molecules, MHC class II, and chemokine receptors. However, unlike maturation induced by microbial products or inflammation, no immunostimulatory cytokine release occurred and T cells with a regulatory as opposed to an effector phenotype were generated. In NOD mice, DC maturation is reportedly impaired and may limit DC migration to PLN, leading to insufficient immunoregulation induction (9, 10). We report the accumulation of more mature CD8⁻ predominant DC in PLN of mATG-treated NOD mice. Furthermore, we noted a significant increase in the percentage of CD8⁻ DC expressing CCR7, the molecule directing the migration of DC to lymphoid tissues (38). Taken together, these findings suggest that mATG administration induces the maturation and subsequent migration of a predominant CD8⁻ DC population to the PLN of NOD mice to activate immunoregulatory mechanisms. Our findings are consistent with the α-galactosylceramide studies mentioned, which demonstrated protection from type 1 diabetes associated with NKT cell-induced maturation and migration of DC to PLN (16, 17). Whether mATG-induced DC maturation represents an indirect effect, as with α-galactosylceramide, is presently unclear. Our in vitro bone marrow-derived DC culture experiments, which lack the DC interactions with other cellular elements present in vivo, failed to demonstrate maturation after mATG treatment. This discrepancy may indicate involvement of an intermediary cell type in vivo, such as NKT cells, or perhaps disruption of E-cadherin adhesions not present in vitro.

In our previous report, we demonstrated a relative increase in CD4⁺CD25⁺FoxP3⁺ Treg cell frequency and suppressive function following in vivo administration of mATG (25). Ruzek et al. (39) demonstrated the induction of Treg cells after treatment of splenocytes from C57BL/6 mice in vitro with mATG, although the addition of exogenous IL-2 was required. Contrary to our initial expectations, a Th2 and not a Treg response ensued when in vivo mATG-treated NOD DC were cultured ex vivo with BDC2.5 cells. Luo et al. (40) were able to generate Treg cells by beta cell peptide-pulsed DC using a similar culture system. However, in their report, the addition of TGF-β1 to culture was required. Furthermore, as mentioned previously, the mATG-treated splenic DC used in our experiments represent a CD8⁻ DC predominant population, which in most reports skews toward a Th2 response. Our adoptive cotransfer studies demonstrated the ability of these Th2 cells to delay the onset of diabetes, confirming the in vivo relevance of the altered DC phenotype in shaping the T cell repertoire toward suppression of diabetogenic activity. Interestingly, Lopez et al. (41) found that induction of Treg cells from human peripheral blood lymphocytes during in vitro treatment with human ATG was dependent on the presence of Th2 cytokines. Thus, our findings do not rule out the possibility that mATG alteration of DC profile and function promoted the relative sparing of Treg cells observed in vivo and augmentation of their suppressive function following administration of mATG in our previous study (25). Finally, we noted higher proliferation in T cell cultures stimulated with mATG-treated DC (CD8⁻ predominant) or purified CD8⁻ DC from naive NOD. These findings are consistent with reports by Shortman and colleagues (32, 33), which noted a reduced capacity of quiescent splenic CD8⁺DC vs CD8⁻ DC to stimulate T cell proliferation.

In summary, we demonstrate that in vivo mATG administration transiently recruited a mature CD8⁻ predominant DC population into PLN of NOD mice. In addition, splenic DC from mATG-treated mice exhibited a more mature phenotype with increased IL-10 production and the potent capacity to induce a Th2 response ex vivo that upon adoptive transfer significantly delayed the onset of diabetes in vivo. These studies suggest the alteration of DC profile and function by mATG may similarly skew the Th1/Th2 balance in vivo and through such actions, represent an additional mechanism by which this agent provides its beneficial effect.

We thank Neal Benson and the staff of the Flow Cytometry Core Laboratory at the University of Florida Interdisciplinary Center for Biotechnology Research for assistance with flow cytometry. The FACSCalibur flow cytometer was purchased by the Center for Immunology.

J.W. serves as Vice President of Immune Mediated Diseases Research, Genzyme Corporation, which is the company providing the rabbit polyclonal mouse antithymocyte globulin and research funding for these studies. K.L.W. is recipient of research support from Genzyme. The remaining authors have no financial conflict of interest.