One of the BB rat diabetes (diabetes mellitus (DM)) susceptibility genes is an Ian5 mutation resulting in premature apoptosis of naive T cells. Impaired differentiation of regulatory T cells has been suggested as one possible mechanism through which this mutation contributes to antipancreatic autoimmunity. Using Ian5 congenic inbred rats (wild-type (non-lyp BB) and mutated (BB)), we assessed the development of BB regulatory CD8−4+25+T cells and their role in the pathogenesis of DM. BB rats have normal numbers of functional CD8−4+25+Foxp3+ thymocytes. The proportion of CD25+ cells among CD8−4+ recent thymic emigrants is also normal while it is increased among more mature CD8−4+ T cells. However, BB CD8−4+25+Foxp3+ thymocytes fail to undergo homeostatic expansion and survive upon transfer to nude BB rats while Foxp3 expression is reduced in mature CD8−4+25+ T cells suggesting that these cells are mostly activated cells. Consistent with this interpretation, peripheral BB CD8−4+25+ T cells do not suppress anti-TCR-mediated activation of non-lyp BB CD8−4+25− T cells but rather stimulate it. Furthermore, adoptive transfer of unfractionated T cells from diabetic BB donors induces DM in 71% of the recipients while no DM occurred when donor T cells are depleted of CD8−4+25+ cells. Adoptive transfer of 106 regulatory non-lyp BB CD8−4+25+ T cells to young BB rats protects the recipients from DM. Taken together, these results demonstrate that the BB rat Ian5 mutation alters the survival and function of regulatory CD8−4+25+ T cells at the post-thymic level, resulting in clonal expansion of diabetogenic T cells among peripheral CD8−4+25+ cells.
The BB rat spontaneously develops a type 1a diabetic syndrome very similar to that observed in humans (1). The disease is polygenic, T cell-mediated, and affects most males and females around puberty (1, 2, 3, 4, 5, 6, 7). Two groups have recently reported the positional cloning of the BB rat diabetes susceptibility gene Iddm2/Lyp located on chromosome 4 (8, 9). Both groups identified a frameshift mutation in the third exon of the immune-associated nucleotide-binding gene Ian5 (named Ian4 by Hornum et al. (9)), a member of an evolutionary conserved gene family involved in immune mechanisms and the regulation of apoptosis in plants, rodents and humans (8, 9).
The initial identification of human Ian5, the ortholog of rat Iddm2/Lyp resulted from the screening of a cDNA library to search for genes conferring resistance to apoptosis induced by gamma-radiation and the protein phosphatase inhibitor, okadaic acid in Jurkat T cells (10). It was also identified by another group through an in silico search for novel members of the human Ian gene family (11). Human Ian5 is predominantly expressed in the thymus, peripheral CD4+ and CD8+ T cells as well as monocytes but not in normal B cells. However, high levels of expression are observed in neoplasms of the B cell lineage (11). The subcellular localization of the human 35-kDa Ian5 protein is still controversial because one study provided evidence for its anchoring to the mitochondrial membrane through its hydrophobic C terminus while the other reported a centrosomal/Golgi/endoplasmic reticulum localization (10, 11).
The rat ortholog of human Ian5 is present in mitochondria derived from thymocytes and peripheral T cells, and is required for the maintenance of mitochondrial integrity (12). Its spontaneous mutation in the BB rat results in the loss of mitochondrial membrane potential and subsequent fragmentation of DNA in recent thymic emigrants (RTE)3 (12). This characterization of normal and mutated rat Iddm2/Lyp/Ian5 at the molecular level is consistent with the immune abnormalities that were previously reported in the BB rat (13).
We and others have shown that the BB rat Iddm2/Lyp/Ian5 mutation manifests itself at the latest stages of intrathymic T cell development (14, 15, 16). Although BB rat thymocytes appear phenotypically normal, there is a 5- to 10-fold reduction in the number of peripheral T cells, mostly due to the lack of long-lived, naive CD90−, ART2+, CD45RC+ T cells (15, 17, 18). Consequently, the pool of recirculating T cells of this animal is mostly comprised of CD90+, CD45RC− RTE that are destined to die rapidly after thymic exit unless they are rescued from premature apoptosis by Ag activation (15, 16).
The BB rat Iddm2/Lyp/Ian5 mutation is necessary though not sufficient for the development of autoimmunity, and its contribution to the diabetogenic process remains poorly understood (3). One proposed mechanism is an impaired development of regulatory T cells, as suggested by the results of depletion and reconstitution experiments (19). Specifically, in vivo depletion of peripheral ART2+ T cells that are absent in BB rats (17) precipitates the onset of diabetes in nonlymphopenic, diabetes-resistant (DR) rats maintained in an open environment (20). Conversely, adoptive transfer of ART2+ T cells to young diabetes-prone BB rats prevents the development of disease (21). It is important to note, however, that the phenotype of the regulatory ART2+ T cells that could prevent diabetes in young BB rats could not be inferred from these depletion and reconstitution experiments since large numbers of unfractionated T cells were transferred, and ART2+ T cells account for most mature T cells in normal rats (18, 21, 22).
Regulatory T cells with various phenotypes have been well-characterized in murine experimental systems. The type of regulatory cells that predominates in a particular autoimmune response may depend on the nature of the self-reactive effector T cells, and the circumstances in which these effector T cells are activated. The best characterized subset of regulatory T cells has the surface phenotype of CD4+25+ (23, 24). These cells account for ∼5–10% of CD4+8− thymocytes and peripheral T cells in rodents and humans (25, 26, 27, 28). Regulatory CD4+25+ T cells can prevent autoimmunity that develops after day 3 thymectomy, adoptive transfer of CD4+25− T cells to nude mice (29), as well as in nonobese diabetic (NOD) mice (30). Even more striking, it has recently been shown that regulatory CD4+25+ T cells can cure spontaneously diabetic NOD mice, and mice affected with advanced, experimentally induced colitis (31, 32). Regulatory CD4+25+ T cells do not proliferate nor produce IL-2 following antigenic stimulation in vitro, but require TCR-mediated activation to mediate their suppressive activity (23, 24). In vitro, CD4+25+ T cells suppress the proliferative response and the cytokine secretion of effector T cells when the two cell populations are cocultured. This suppression is cell contact-dependent but appears to be cytokine-independent. None of the surface markers (CD45RB, CTLA-4, glucocorticoid-induced TNF receptor family-related gene) used to identify CD4+25+ regulatory T cells could completely discriminate these cells from activated, effector, or memory T cells until the demonstration that Foxp3 is specifically expressed in CD4+25+ regulatory T cells (33).
In rat models of autoimmunity, including experimentally induced type 1 diabetes, regulatory cells have been found among both CD4+25− and CD4+25+ T cells (34, 35, 36) as well as CD4−8+45RClow T cells (37). The CD4+8− regulatory T cell subsets express ART2 and a low level of CD45RC on their surface, and appear to function, at least in part, through the release of IL-4 and TGF-β (38).
In the present study, we have assessed the effects of the BB rat Iddm2/Lyp/Ian5 mutation on the development of regulatory CD8−4+25+ T cells. Our results demonstrate that the differentiation of this T cell subset is altered at the post-thymic level, and this impaired maturation plays a major role in the pathogenesis of antipancreatic autoimmunity in this animal.
Materials and Methods
Three congenic lines of BB rats were used in this study. Lymphopenic and diabetes-prone BB rats were obtained from Biomedical Research Models. Nonlymphopenic and DR BB rats (non-lyp BB) were developed in our laboratory through introgression of the wild-type Ian5 locus derived from DR (DR-BB/W) BB rats (Biomedical Research Models) into BB rats. Briefly, BB and DR-BB/W rats were crossed, and the resulting F1 animals were backcrossed to BB rats. This step was followed by 10 more, marker-assisted backcrosses of Ian5 heterozygous progeny to BB rats. After the eleventh backcross, nonlymphopenic rats were intercrossed, and their progeny homozygous for wild-type Ian5 were selected for establishing the congenic non-lyp BB line. A similar approach was taken to develop BB rats congenic for the rnu mutation (nude BB) introgressed from inbred rnu/rnu WAG rats (Biomedical Research Models). Briefly, BB and WAG rnu/rnu rats that share the same MHC (RT1u) haplotype, were intercrossed, and this intercross was followed by 11 marker-assisted backcrosses to BB rats. After the first backcross, animals that were lymphopenic and heterozygous for the rnu mutation were selected for subsequent breeding. Consequently, the genotyping of animals to be used in the following backcrosses was limited to the rnu locus. After the eleventh backcross, rnu heterozygous animals were intercrossed to establish the congenic nude BB line that has been maintained through mating of rnu homozygous brothers with rnu heterozygous sisters.
All animals were housed in specific pathogen-free conditions and all the sentinels were negative for circulating Abs specific for the diabetogenic Kilham’s virus. Diabetes-prone rats were tested three times a week for the presence of glycosuria and ketonuria. Once animals became glycosuric, the diagnosis of diabetes was established on the basis of hyperglycemia (blood glucose >16.7 mM) for 2 consecutive days.
mAbs, multicolor immunofluorescence, and FACS analysis
The mAbs used in this study were affinity purified from hybridoma culture supernatants on rat anti-mouse Ig-, and mouse anti-rat Ig-Sepharose, and then conjugated with FITC, biotin, PE, or allophycocyanin using standard procedures. These mAbs included anti-rat Ig κ (MARK1), anti-CD8α (OX8), anti-CD8β (341), anti-CD4 (W3/25), anti-CD45RC (OX22), anti-CD90 (OX7), anti-TCRαβ (R73), anti-ART2.1 (DS4.23), and anti-CD3ε (G4.18). The hybridomas OX8, W3/25, OX22, and OX7 were kindly provided by Dr. D. Mason (University of Oxford, Oxford, U.K.), R73 and 341 by Dr. T. Hünig (University of Würzburg, Würzburg, Germany), DS4.23 by Dr. D. Greiner (University of Massachusetts, Worcester, MA), and G4.18 by Dr. G. W. Butcher (The Babraham Institute, Cambridge, U.K.) with the permission of Dr. B. M. Hall (University of New South Wales, Sydney, Australia). Streptavidin PE/Texas Red Tandem was purchased from Southern Biotechnology Associates. Affinity purified normal rat and mouse Ig, as well as rat antiserum specific for mouse Ig, all reagents required for rosetting, were purchased from Biocan (Cedarlane Laboratories).
Suspensions of mononuclear cells (MNC) were incubated with biotinylated mAb, followed by streptavidin PE/Texas Red Tandem. PE-, allophycocyanin-, and FITC-conjugated mAbs were then added simultaneously, as previously described (39). Viable cells were gated using forward and side angle scatter and analyzed flow cytometrically with a FACSCalibur (BD Biosciences).
Purification of cell subsets
CD25+ and CD25− subsets of lymph node CD8−4+ T cells were purified by a two-step procedure. In a first step, CD8−4+ T cells were purified by negative selection using a rat CD4 T cell subset column kit from R&D Systems (Cedarlane Laboratories). The resulting cell suspension was then incubated with fluorescent anti-CD25, anti-CD8β, and anti-CD4 mAbs, and the two CD25+ and CD25− subsets of CD8β−4+ T cells were sorted using a FACSARIA (BD Biosciences). CD8−4+ thymocytes were purified from unfractionated thymocytes by FACS using fluorescent anti-CD4 and anti-CD8α mAbs. CD25+ and CD25− subsets of single-positive CD8−4+ thymocytes were purified by a two-step procedure. In a first step, CD8− thymocytes were enriched from unfractionated thymocytes by negative selection using a rosetting technique, as previously described (2). The resulting cell suspension was then incubated with fluorescent anti-CD25, anti-CD8β, and anti-CD4 mAbs, and the two CD25+ and CD25− subsets of CD8β−4+ thymocytes were purified by FACS. T cell-depleted APC were prepared from non-lyp BB splenocytes by negative selection using biotinylated anti-CD3 mAb, followed by streptavidin microbeads (Miltenyi Biotec), and magnetic depletion using an AutoMacs (Miltenyi Biotec). Unfractionated and CD25− splenic T cells from recently diagnosed diabetic BB rats were purified in two steps as follows. In a first step, T cells were enriched by negative selection using a rat T cell column kit from R&D Systems (Cedarlane Laboratories). The resulting cell suspension was then incubated with fluorescent anti-TCRαβ and anti-CD25 mAbs and unfractionated T cells as well as CD25− T cells were purified by FACS. The purity of cell subsets was always >98%.
Adoptive transfers of T cell subsets
CD8−4+25+ thymocytes (2.0 × 106/recipient) purified from 6-wk-old BB and non-lyp BB donors were labeled with CFSE (Molecular Probes) and injected i.v. into 6- to 8-wk-old nude BB rats. Three weeks later, recipient lymph nodes and spleen were collected for phenotypic analysis through cell surface immunofluorescence and FACS analysis.
CD8−4+25+ and CD8−4+25− T cell subsets (1.0 × 106/recipient) purified from lymph nodes of 6-wk-old, non-lyp BB donors were injected i.v. into 3-wk-old, diabetes-prone BB rats. Recipients were followed for the development of diabetes up to 4 mo of age as previously described (2).
Unfractionated and CD25− T cells purified from splenocytes of recently diagnosed diabetic BB rats were activated in vitro with PMA (20 ng/ml), ionomycin (25 μM), and 20 U/ml rIL-2 at 37°C for 72 h, as previously described (2). PMA and ionomycin were purchased from Sigma-Aldrich. At the end of the culture, an aliquot of the cells was kept for phenotypic analysis, and the remaining cells (2.0 × 106 cells/recipient) were injected i.v. into 3-wk-old BB rats. Recipients were followed for the development of diabetes within a month posttransfer.
T cell proliferation and IL-2 secretion assays
Cell cultures were performed in triplicate in 96-well plates containing DMEM medium, 10 mM HEPES, 1% penicillin-streptomycin, 5.5 × 10−5 M 2-ME, and 10% FCS for 48 h as previously described (39). Culture reagents were purchased from Invitrogen Life Technologies. For proliferation assay, CD8−4+25− T cells (5 × 104/well) purified from lymph nodes of non-lyp BB rats were cultured with soluble anti-TCR mAb (10 μg/ml) in the presence of gamma-irradiated (2000 R), T-depleted splenocytes (4 × 105/well) from the same donors. Indicated numbers of purified CD8−4+25+ T cells derived from the thymus and lymph nodes of BB and non-lyp BB rats were added to the cultures. Cultures were pulsed with [3H]TdR for the last 6 h of culture, and incorporation of [3H]TdR into DNA was assessed through liquid scintillation spectroscopy. For cytokine secretion assay, cells were cultured in the same conditions, and supernatants were collected after 48 h. IL-2 levels were measured in culture supernatants by specific ELISA (R&D Systems). To obtain recently activated CD8−4+ T cells expressing CD25, sorted non-lyp BB CD8−4+25− T cells (5 × 105 cells/ml) were activated by plastic-bound anti-TCR (1 μg/ml) and anti-CD28 (10 μg/ml) mAbs for 18 h, collected, counted, and used for subsequent coculture experiments.
Semiquantitative and real-time quantitative RT-PCR
Total RNA was extracted from freshly purified T cell subsets using TRIzol Reagents (Invitrogen Life Technologies) and cDNA was reverse-transcribed using Superscript II (Invitrogen Life Technologies) as per manufacturer’s instructions. Detection of Foxp3 transcripts in T cell subsets was assessed by semiquantitative RT-PCR. All semiquantitative PCRs used the same serially diluted cDNA batches shown for β-actin. The Foxp3- and β-actin-specific primers, expected amplicon lengths, and annealing temperatures were: for Foxp3: forward 5′-TGCATCAGCTCTCCACTGTAGACGCA-3′, reverse 5′-CTGTCTTTCCTGGGTGTACCTGAGCG-3′; 234 bp amplicon; annealing temperature 55°C; for β-actin: forward 5′-CGACGAGGCCCAGAGCAAGAGAGG-3′, reverse 5′-CGTCAGGCAGCTCATAGCTCTTCTCCAGGG-3′; 568 bp amplicon; annealing temperature 55°C. PCR products were separated by agarose gel electrophoresis and visualized by ethidium bromide staining. PCR products shown correspond to the expected molecular sizes.
Foxp3 and GAPDH mRNA levels were quantified in T cell subsets by real-time quantitative RT-PCR using gene-specific, fluorescent TaqMan probes and an ABI Prism 7000 (Applied Biosystems). The nucleotide sequences of Foxp3 primers were: forward 5′-CCA TTG GTT CAC ACG CAT GT-3′, reverse 5′-TGG CGG ATG GCA TTC TTC-3′, while the sequence of the Foxp3-specific TaqMan probe was: 5′-FAM CGC CTA CTT CAG AAA CCA CCC CGC-3′. Quantification of GAPDH transcript levels used TaqMan rodent GAPDH control reagents from Applied Biosystems. Standard curves with cDNAs from non-lyp BB CD8−4+25+ single-positive thymocytes were used to calibrate the threshold cycle to relative quantities of Foxp3 and GAPDH cDNAs in each sample. Normalized value for Foxp3 mRNA expression in each sample was calculated as the relative quantity of Foxp3 divided by the relative quantity of GAPDH. All sample were run in triplicates.
Results are expressed as the mean ± 1 SD and overall differences between variables were evaluated by Mann-Whitney U test. When indicated, Fisher’s exact test was used.
Intrathymic development of functional CD8−4+25+ T cell precursors in BB rats
Studies in humans and rodents have shown that regulatory CD8−4+25+ T cells develop in the thymus where they become detectable at the transition from double-positive CD8+4+ to single-positive CD8−4+ stages (24). As illustrated in Fig. 1, the proportion of CD25+ cells among CD8−4+ thymocytes was similar in 6-wk-old BB and non-lyp BB animals (5.1 ± 0.3% and 5.6 ± 0.2%, respectively, n = 5). Because the total number of thymocytes was not significantly different between these congenic lines, this result strongly suggests that normal numbers of CD8−4+25+ thymocytes are produced in BB rats. The question follows as to whether this subset of BB cells exhibits a normal regulatory function in vitro.
First, we assessed the DNA synthesis and IL-2 secretion of CD8−4+25+ thymocytes from BB and non-lyp BB donors in response to activation by soluble anti-TCR mAb. The anti-TCR-induced incorporation of [3H]TdR into DNA of CD8−4+25+ thymocytes was very low in both strains (256 ± 41 cpm in BB rats vs 200 ± 42 cpm in non-lyp BB animals, n = 5), and similar to that observed in unstimulated, lymph node CD8−4+25− T cells (98 ± 18 cpm) of non-lyp BB origin (Fig. 2,A). This lack of stimulation of DNA synthesis in response to TCR-mediated activation of CD8−4+25+ thymocytes from BB and non-lyp BB donors was associated with a very low secretion of IL-2 (16.1 ± 1.8 and 10.4 ± 2.1 pg/ml, respectively) by these cells in the same stimulatory conditions (Fig. 3,A). Furthermore, CD8−4+25+ thymocytes from both BB and non-lyp BB donors could suppress the proliferative response of non-lyp CD8−4+25− T cells. Specifically, the incorporation of [3H]TdR into DNA of CD8−4+25− T cells induced by soluble anti-TCR (17,123 ± 1,806 cpm) was reduced 2- to 3-fold when these cells were activated in the presence of CD8−4+25+ thymocytes from non-lyp BB (10,405 ± 405 cpm) and BB (7,637 ± 609 cpm) donors (Fig. 2,A). As illustrated in Fig. 3 A, CD8−4+25+ thymocytes regulated the proliferative response of non-lyp BB CD8−4+25− T cells through inhibition of IL-2 expression (40). Thus, the secretion of IL-2 by CD8−4+25− T cells in response to anti-TCR stimulation (1,097 ± 10.2 pg/ml) was reduced >50-fold when these cells were cocultured with CD8−4+25+ thymocytes of BB (21.9 ± 4.7 pg/ml) and non-lyp BB (16.1 ± 0.3 pg/ml) origin. Taken together, these results demonstrate that the BB rat Ian5 mutation does not alter the intrathymic development of regulatory CD8−4+25+ T cell precursors.
Altered post-thymic development of regulatory CD8−4+25+ T cells in BB rats
Although the proportion of CD25+ cells among BB rat CD8−4+ thymocytes was normal, this was not the case among the progeny of these CD8−4+ T cell precursors. Specifically, the proportion of peripheral CD8−4+ T cells expressing CD25 on their surface was significantly (p < 0.05) higher in 6-wk-old BB rats (10.7 ± 1.5%, n = 5) than in non-lyp BB animals (7.9 ± 0.5%, n = 5) (Fig. 1). This increased proportion of peripheral CD8−4+25+ T cells in BB rats suggested that these cells could have a selective advantage over their CD25− counterparts either at the level of their thymic output and/or their subsequent survival. This led us to assess the proportion of CD25+ cells among CD8−4+ RTE and more mature T cells. These two subsets of T cells can be distinguished in rats through their differential expression of CD90 that is lost by RTE within a week of their peripheral migration (18). As illustrated in Fig. 4, the proportion of CD25+ cells among CD8−4+ thymocytes (histograms in A and B) and CD90+8−4+ RTE (histograms in C and D) was similar in BB and non-lyp BB rats. This observation strongly suggested that a differential thymic output of CD25+ and CD25− T cells in BB rats could not explain the increased proportion of peripheral CD8−4+25+ T cells in this strain. In contrast, there was a >3-fold increase (p < 0.01) in the proportion of CD25+ cells among BB rat CD90−8−4+ T cells (27.0 ± 3.7%, n = 5, histogram in F) when compared with non-lyp BB mature T cells (7.9 ± 1.4%, n = 5, histogram in E). It must be noted (Fig. 1), however, that the proportion of CD25+ T cells among unfractionated mononuclear cells of BB rat lymph nodes (1.4 ± 0.1%, n = 5) was significantly (p < 0.01) lower than that observed in non-lyp BB animals (4.2 ± 0.1%, n = 5) due to the severe, peripheral T lymphopenia of the former strain.
The enrichment of BB rat CD90−8−4+ T cells in CD25+ cells could reflect a selective advantage of regulatory T cells over naive T cells, possibly as a consequence of the higher avidity of the former subset for self-peptide-MHC (41). Alternatively, because it has been shown that Ag activation rescues naive T cells from premature apoptosis in BB rats, the elevated proportion of CD90−8−4+25+ T cells could simply result from the differential life span of naive and recently activated T cells in this animal (16). This led us to assess the regulatory function of peripheral CD8−4+25+ T cells in Ian5 congenic BB rats.
Mature CD8−4+25+ T cells are mostly activated cells in BB rats
As expected, CD8−4+25+ T cells purified from lymph nodes of non-lyp BB rats failed to proliferate and secrete IL-2 in response to activation by soluble anti-TCR (Figs. 2,B and 3,B). [3H]TdR incorporation into DNA of 2.5 × 104 activated CD8−4+25+ T cells (350 ± 57 cpm) was very low and comparable to that (184 ± 19 cpm) observed in 5.0 × 104 unstimulated CD8−4+25− T cells of the same origin (Fig. 2,B). Similarly, the secretion of IL-2 by activated CD8−4+25+ T cells of non-lyp BB rats (17.2 ± 2.5 pg/ml) was significantly lower than that (119.1 ± 14.8 pg/ml) observed in cultures of unstimulated CD8−4+25− T cells of the same origin. Therefore, this subset of non-lyp BB CD8−4+25+ T cells responded to TCR-mediated activation in a manner that was expected from regulatory T cells. This was not the case for BB CD8−4+25+ T cells which proliferated vigorously following activation by soluble anti-TCR (Fig. 2,B). Specifically, [3H]TdR incorporation into DNA of 2.5 × 104 BB rat CD8−4+25+ T cells (12,408 ± 459 cpm) was only 3-fold lower than that (39,452 ± 1,066 cpm) observed in 5.0 × 104 activated CD8−4+25− T cells of non-lyp BB origin (Fig. 2,B). This robust proliferation of BB rat CD8−4+25+ T cells in response to anti-TCR activation was associated with a paradoxically low concentration of IL-2 (21.4 ± 3.7 pg/ml) in culture supernatants (Fig. 3 B). We interpret this observation as the likely consequence of IL-2 consumption by the proliferating BB rat CD8−4+25+ T cells.
The opposite responses of CD8−4+25+ T cells from Ian5 congenic BB donors to anti-TCR stimulation correlated with a differential ability of these subsets to suppress the TCR-induced proliferation of non-lyp BB CD8−4+25− T cells. As illustrated in Figs. 2,B and 3,B, CD8−4+25+ T cells from non-lyp BB donors inhibited both the anti-TCR-mediated proliferation and IL-2 secretion of syngeneic CD8−4+25− T cells in a dose-dependent manner. Thus, the [3H]TdR incorporation into DNA of 5.0 × 104 activated CD8−4+25− T cells (39,452 ± 1,066 cpm) was reduced to 4,376 ± 329 cpm and 27,504 ± 912 cpm when these cells were cocultured with 2.5 × 104 and 6.2 × 103 CD8−4+25+ T cells from non-lyp BB donors, respectively (Fig. 2,B). Simultaneously, there was a profound inhibition of the IL-2 secretion by CD8−4+25− T cells from 552 ± 16 pg/ml, when activated alone, to 42.1 ± 14 pg/ml when these cells were stimulated in the presence of 2.5 × 104 syngeneic CD8−4+25+ T cells (Fig. 3,B). In contrast, CD8−4+25+ T cells of BB rats failed to inhibit the TCR-mediated proliferation of CD8−4+25− T cells from non-lyp BB donors. In fact, there was a >2-fold increase in the [3H]TdR incorporation into DNA of 5.0 × 104 non-lyp BB CD8−4+25− T cells (91,939 ± 2,300 cpm) when these cells were activated in the presence of 2.5 × 104 BB CD8−4+25+ T cells (Fig. 2 B). As expected, this stimulation of DNA synthesis when the two subsets of T cells were cocultured was associated with a lack of inhibition of IL-2 secretion (413 ± 18.4 pg/ml).
The proliferation of BB rat CD8−4+25+ T cells in response to anti-TCR stimulation and the inability of these cells to exhibit suppressive activity in vitro strongly suggested that they represent recently activated T cells rather than regulatory T cells. To determine whether this is the case, we first examined the expression of Foxp3, a marker of regulatory T cells, in thymocytes and peripheral T cells of Ian5 congenic BB rats by RT-PCR (Fig. 5). In the thymus of both strains, Foxp3 transcription was restricted to single-positive CD8−4+ cells expressing CD25. Similarly, CD25+ subsets of lymph node CD90−8−4+ (Fig. 5) and CD90+8−4+ T cells (not shown) from both congenic lines transcribed Foxp3 while CD8−4+25− and CD8+4− T cell subsets did not (Fig. 5). Real-time quantitative PCR analyses, however, revealed differential levels of Foxp3 transcripts in T cell subsets of Ian5 congenic BB rats that were consistent with the results and interpretation of our functional analyses. In the thymus of both strains, Foxp3 mRNA levels were 100- to 200-fold higher in CD8−4+25+ cells than in CD8−4+25− and CD8+4+ subsets (Fig. 6 A). The similar levels of Foxp3 transcription in thymocyte subsets of the two Ian5 congenic strains, combined with the results of our phenotypic and functional analyses confirmed that the BB rat Ian5 mutation does not significantly compromise the intrathymic differentiation of regulatory CD8−4+25+ T cell precursors.
Among CD8−4+ RTE, levels of Foxp3 transcripts were >200-fold higher in cells expressing CD25 than in CD25− cells of both BB and non-lyp BB donors (Fig. 6,B). Foxp3 mRNA levels were almost undetectable in more mature CD90−8−4+25− T cells of both strains (Fig. 6 B). However, transcription levels of Foxp3 by mature CD90−8−4+25+ T cells were 4- to 5-fold higher in non-lyp BB rats than in BB donors. This differential level of Foxp3 transcription and the results of our functional analyses strongly suggested that the CD90−8−4+25+ T cell subset of BB rats was comprised of a high proportion of recently activated T cells and a low proportion of regulatory T cells while it was the reverse in non-lyp BB animals. However, the relatively stable levels of Foxp3 transcription during the maturation of CD8−4+25+ T cells in BB rats raised the possibility that, in contrast to most naive T cells, regulatory CD8−4+25+ T cells do not die precipitously by apoptosis in BB rats though they are not functional.
The life span of CD8−4+25+ thymocytes and RTE purified from Ian5 congenic BB donors was first assessed in vitro. As illustrated in Fig. 7, while 65 ± 11% of non-lyp BB CD8−4+25+ thymocytes still excluded trypan blue after 20 h of culture, only 32 ± 6% of their BB counterparts did. Similarly, the proportion of RTE expressing CD25 that survived after the same period was >3-fold higher among cells of non-lyp BB origin (75 ± 4%) than among BB cells (20 ± 4%). These results strongly suggested that the life span of thymocytes and RTE expressing CD25 was also compromised by the BB rat Ian5 mutation. This was confirmed by our in vivo analyses. CD8−4+25+ thymocytes were purified from Ian5 congenic BB rats by FACS sorting (>99% pure), CFSE-labeled and, then, adoptively transferred (2.0 × 106 thymocytes/recipient) into nude BB rats. We chose to transfer thymocytes rather than RTE for two reasons. Our results show that thymocytes expressing CD25 are phenotypically and functionally very similar in Ian5 congenic BB rats. Moreover, a differential proportion of activated cells among RTE between the two strains could possibly influence the survival and homeostatic expansion of adoptively transferred cells. CD8−4+ T cells account for <1% of pooled lymph node and splenic MNC in nonreconstituted nude BB rats (Fig. 8,A). Consequently, the progeny of non-lyp BB thymocytes was easily detected in nude recipients 3 wk after transfer, and ∼25–33% of these cells expressed CD25 on their surface (Fig. 8, B and C). Importantly, the number of recovered cells (4.1 ± 0.4 × 106 cells/recipients, n = 3) was larger than the injected number, and all the cells were CFSE−, demonstrating that normal CD8−4+25+ thymocytes could survive and undergo homeostatic expansion in nude BB recipients. In contrast, very few CD8−4+25+ or − T cells were recovered from recipients of BB CD8−4+25+ thymocytes (Fig. 8, D and E). Combined with the decrease in absolute number of peripheral CD8−4+25+ T cells in unmanipulated BB rats, the adoptive transfer experiments demonstrate that newly developed, regulatory CD8−4+25+ T cell precursors have a very short life span and cannot undergo homeostatic expansion.
Having established that the BB rat Ian5 mutation compromises the survival of regulatory CD8−4+25+ T cells, we sought to determine what ratio of recently activated T cells to regulatory T cells was required to mimic the lack of suppressive activity observed in CD8−4+25+ T cells of BB origin. Naive CD90+8−4+25− T cells from non-lyp BB donors were sorted, and activated overnight by plastic-bound anti-TCR and anti-CD28 mAbs to generate recently activated T cells. At the end of this activation, >95% of the cells expressed a level of CD25 on their surface similar to that observed in freshly isolated, syngeneic CD8−4+25+ T cells (data not shown). These in vitro-activated T cells were collected and mixed in various ratios with ex vivo, syngeneic CD8−4+25+ T cells. These mixtures of cells were then assessed for their suppressive activity on syngeneic CD8−4+25− T cells activated by soluble anti-TCR. As illustrated in Fig. 9, the suppressive activity of freshly isolated regulatory CD8−4+25+ T cells was progressively lost when they were mixed with increasing numbers of recently activated T cells. When the ratio of regulatory to recently activated T cells was 1:5, i.e., similar to that of Foxp3 transcripts in mature BB and non-lyp BB CD90−8−4+25+ T cells, we observed a 2-fold stimulation of the proliferative response of syngeneic CD8−4+25− T cells. This increased proliferation was strikingly similar to that observed when non-lyp BB CD8−4+25− T cells were cocultured with BB CD90−8−4+25+ T cells (compare Figs. 2,B and 9). These results strongly suggested that there was a ∼80% decrease in the proportion of regulatory cells among BB CD90−8−4+25+ T cells when compared with their non-lyp BB counterparts.
Evidence that impaired differentiation of BB rat CD8−4+25+ T cells plays a central role in the development of diabetes in these animals
Having established that the BB rat Ian5 mutation compromises the differentiation of regulatory CD8−4+25+ T cells at the post-thymic level, we next determined whether this altered maturation plays a role in the pathogenesis of diabetes.
CD8−4+25− and regulatory CD8−4+25+ T cells were purified from non-lyp BB rats and adoptively transferred (106 cells/recipient) to 3-wk-old, diabetes-prone BB rats. Recipients were followed for the development of diabetes thereafter. Seventy percent of the PBS-injected animals developed diabetes by the age of 4 mo, a cumulative incidence that was higher but not significantly different from that (45%) observed in recipients of CD8−4+25− T cells (Table I). In contrast, animals that received regulatory CD8−4+25+ T cells were protected from disease because only 13% of these rats became diabetic (p = 0.009 and 0.09 vs PBS-injected and CD8−4+25−-injected rats, respectively). Although cells of donor and recipient origin could not be distinguished within the splenic and lymph node CD8−4+25+ subsets at sacrifice, the proportion of CD8−4+ T cells expressing CD25 was significantly higher in recipients of CD8−4+25+ T cells than in the two other experimental groups (data not shown). Thus, adoptive transfer of non-lyp BB regulatory CD8−4+25+ T cells to diabetes-prone BB rats reconstitute this T cell subset in recipients, and protects them from antipancreatic autoimmunity.
|.||CD8−4+25+ .||CD8−4+25− .||PBS .|
|.||CD8−4+25+ .||CD8−4+25− .||PBS .|
Cumulative incidence of diabetes in BB rats 3 mos after adoptive transfer of 106 CD8−4+25+ or − T cells from wild-type Ian5 congenic BB donors.
p = 0.009 vs PBS injection.
p = 0.09 vs CD8−4+25− injection.
p = NS vs PBS injection (two-tailed Fisher’s test).
The transcription levels of Foxp3 in CD8−4+25+ T cells of Ian5 congenic BB rats, and the functional analyses of these cells indicated that, in BB rats, these cells were comprised mostly of activated T cells. The question followed as to whether these activated T cells were enriched in diabetogenic T cells. Unfractionated and CD25− splenic T cells were purified from recently diabetic BB rats, preactivated in vitro, and then transferred into 3-wk-old syngeneic rats. Recipients were followed for the development of diabetes within 1 mo posttransfer. Seventy-one percent of the recipients of unfractionated T cells developed diabetes (Table II) while none of the uninjected and CD25− T cell-injected animals did (p = 0.035). These results demonstrate that the impaired development of regulatory T cells in BB rats allows the expansion of diabetogenic T cells that are present within the CD25+ subset of these animals.
|.||Unfractionated T Cells .||T Cells Depleted of CD8−4+25+ Cells .||No T Cell Transfer .|
|.||Unfractionated T Cells .||T Cells Depleted of CD8−4+25+ Cells .||No T Cell Transfer .|
Cumulative incidence of diabetes 1 mo after adoptive transfer of T cell subsets (2 × 106 T cells/recipient) derived from diabetic BB donors to 3-wk-old BB recipients.
p = 0.035 vs the two other groups (two-tailed Fisher’s test).
The exact mechanism(s) through which the BB rat Ian5 mutation contributes to the development of diabetes has remained elusive. The current study demonstrates that this mutation compromises the development of regulatory CD8−4+25+ T cells, and this altered differentiation plays a central role in the pathogenesis of the disease. Furthermore, we show that while the BB rat CD8−4+25+ T cell subset is deprived of regulatory cells, it is enriched in diabetogenic T cells.
Several lines of evidence had previously implicated an imbalance between regulatory and autoreactive T cells as one of the diabetogenic mechanisms resulting from the BB rat Ian5 mutation (19). First, BB rats lack mature, peripheral ART2+, CD45RC− T cells, i.e., the population that contains the three subsets of regulatory T cells, ART2+CD8−4+45RC−25+, ART2+CD8−4+45RC−25−,and CD8+4−45RC− that have been characterized in normal rats (17). The second line of evidence comes from reconstitution experiments. Specifically, adoptive transfer of large numbers of normal, histocompatible T cells enriched in ART2+CD8−4+ cells to diabetes-prone BB rats results in the partial correction of the peripheral T lymphopenia, and the prevention of both insulitis and diabetes (21). Conversely, administration of depleting anti-ART2 mAb to DR rats genetically related to BB rats but not lymphopenic, results in the development of type 1 diabetes when the recipients are simultaneously treated with inducers of IFN-α (6, 7). Of note, the suppressive activity of the cells used in these reconstitution experiments, or depleted by Ab treatment was not assessed in vitro. Furthermore, because virtually all mature T cells express ART2 on their surface in normal rats, the subset(s) of regulatory cells that could prevent antipancreatic autoimmunity in BB rats could not be inferred from these reconstitution and depletion experiments. Our adoptive transfer experiments demonstrate that low numbers of regulatory CD8−4+25+ T cells derived from wild-type Ian5 donors can inhibit the development of diabetes in most BB recipients, a protection that could not be afforded by similar numbers of CD8−4+25− T cells. This observation is consistent with the results of previous studies that compared the regulatory function of rat CD8−4+25+ and CD8−4+25− T cells in models of diabetes and thyroiditis induced by thymectomy and split-dose gamma-irradiation (42). These studies showed that the regulatory function of CD8−4+25− T cells is confined to a subset of memory CD45RC− cells but indicated also that the suppressive activity of CD8−4+45RC−25− T cells is considerably weaker than that of CD8−4+25+ T cells on a cell-per-cell basis (34). The BB rat Ian5 mutation affects the development of CD8+4− T cells more severely than that of CD8−4+ T cells. It is therefore not implausible that a lack of regulatory CD8+4−45RC− T cells, recently shown to suppress Th1 T cells, also contributes to the development of diabetes in BB rats (37). However, this contribution must be modest given the efficient protection afforded by low numbers of CD8−4+25+ T cells.
The Ian 5 mutation of the BB strain is responsible for the premature apoptotic death of RTE and the lack of long-lived naive T cells (15, 16). Consequently, BB rat peripheral T cells with a given Ag specificity are either T cells that emigrated from the thymus in the previous 24–48 h, or T cells that were rescued from premature apoptosis through Ag activation (16). There is evidence that, in normal animals, the avidity of CD8−4+25+ thymocytes for self-peptide-MHC is intermediate between that required for the positive selection of CD8−4+25− thymocytes and that leading to negative selection of T cells expressing TCRs of high affinity for self (41). The question followed as to whether thymus-derived CD8−4+25+ T cells produced in normal number in the BB rat would escape early apoptotic death as a consequence of their higher avidity for self-peptide-MHC. The low number and lack of suppressive activity of these cells in unmanipulated BB rats, combined with the inability of their precursors to undergo homeostatic expansion and survive upon adoptive transfer to nude BB rats, demonstrate that the fate of CD8−4+25− and CD8−4+25+ RTE is similar in BB rats in the absence of Ag activation. The only minor difference appears to be a slight delay in the apoptotic death of CD8−4+25+ T cells as suggested by the presence of CD90−8−4+25+ T cells expressing Foxp3 in unmanipulated BB rats while naive CD90−8−4+25− are absent in the same animals (16). Although the death of BB rat CD8−4+25+ RTE could occur more slowly than that of CD8−4+25− RTE, it is important to note, however, that these cells have lost the suppressive activity displayed by their thymic precursors in vitro. This functional loss must occur in vivo very early after emigration from the thymus because the normal proportion of CD8−4+ RTE expressing CD25 and Foxp3 in unmanipulated BB rats is insufficient to inhibit the clonal expansion and differentiation of diabetogenic T cells in these animals.
Adoptive transfer of the BB rat diabetic syndrome by CD8−4+ and CD8+4− T cells isolated from diabetic donors requires polyclonal preactivation of donor cells in vitro (43, 44). Here we show that depletion of CD25+ T cells before in vitro activation prevents the subsequent transfer of disease. Although this observation indirectly confirms the lack of functional, regulatory T cells within the CD25+ subset, it also provides some insights into the autoreactive T cell repertoire of BB rats. Impaired negative selection of diabetogenic T cell precursors in BB rats had been suggested based on their susceptibility to organ-specific autoimmunity, the presence of abnormalities in the thymic stroma of these animals (45) and the ability of their thymocytes to induce diabetes following their adoptive transfer to nude rats treated with anti-ART2 mAb (46). Our inability to transfer diabetes with CD25− T cells from diabetic donors show that the proportion of diabetogenic precursors in this T cell population is low, and remains low after in vitro expansion because these cells cannot precipitate diabetes in young, diabetes-prone recipients. In contrast, the high incidence of diabetes after adoptive transfer of unfractionated T cells demonstrate that the CD25+ subset lacks regulatory cells but is enriched in clonally expanded, diabetogenic T cells. Therefore, these adoptive transfer experiments strongly suggest that negative selection of β cell-specific T cell precursors is not impaired in BB rats. We and others had previously shown that BB rat RTE could be rescued from apoptosis by (self) Ag activation. This observation led us to propose that one of the mechanisms through which the Ian 5 mutation of the BB rat contributes to the development of diabetes is the progressive enrichment of the surviving T cell repertoire in diabetogenic T cells (16). Our adoptive transfer experiments further support this conclusion.
The authors have no financial conflict of interest.
We thank Gisele Knowles (Sunnybrook and Women’s College Health Science Centre) for expert assistance with flow cytometry.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by grants from the Canadian Institutes of Health Research and Genome Canada.
Abbreviations used in this paper: RTE, recent thymic emigrant; NOD, nonobese diabetic; DR, diabetes-resistant; MNC, mononuclear cell.