Deficiency in NOD Antigen-Presenting Cell Function May Be Responsible for Suboptimal CD4⁺CD25⁺ T-Cell–Mediated Regulation and Type 1 Diabetes Development in NOD Mice

C57BL/6 and NOD female mice from 3 to 35 weeks of age were in our barrier facility under SPF conditions and obtained from The Jackson Laboratory (Bar Harbor, ME). Mice were maintained according to institutional animal care and use committee guidelines.

We labeled 1 ×10⁶ lymph node cells with APC–anti-CD25, peridinin chlorophyll protein (PerCP)-anti-CD4, fluorescein isothiocyanate–anti-CD62L, phycoethrin (PE)–anti-CTLA-4 (Pharmingen, San Diego, CA), or PE–anti-GITR (R&D, Minneapolis, MN) antibodies for 20 min in Dulbecco’s PBS, 1% FCS, and 0.1% NaNO₃ and washed them twice. For CTLA-4 labeling, cells were fixed and permeabilized according to the manufacturer’s instructions.

mRNA was extracted, using a Picopure RNA isolation kit (Arcturus, Mountain View, CA), and reverse transcribed, using a Taq Man reverse transcriptase kit (Applied Biosystems, Foster City, CA). cDNA was amplified in duplicate by real-time PCR, using a SYBR Green PCR kit (Applied Biosystems) with primers for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and Foxp3. Relative Foxp3 mRNA amounts were normalized with respect to GAPDH mRNA amounts. The fold increase by comparison to CD4⁺CD25⁻ T-cells is represented (fold change = 1 for CD4⁺CD25⁻ T-cells). Primer sequences are: GAPDH: 5′-GGA-GCG-AGA-CCC-CAC-TAA-CA-3′ and 5′-ACA-TAC-TCA-GCA-CCG-GCC-TC-3′; and Foxp3: 5′-CCC-ACC-TAC-AGG-CCC-TTC-TC-3′ and 5′-GGC-ATG-GGC-ATC-CAC-AGT-3′.

Lymphocytes were obtained from lymph nodes and spleens and processed in Hank’s balanced salt solution containing 2% heat-inactivated FCS (Hyclone, Logan, UT). CD4 cells were enriched on a CD4 cell-enrichment column (R&D) labeled with PE–anti-CD25 antibody, incubated with anti-PE beads (Miltenyi Biotech, Auburn, CA), and then passed through a magnetic column according to the manufacturer’s instructions. CD4⁺CD25⁺ T-cell purity was consistently >90%.

CD4⁺CD25⁻ T-cells (2.5 × 10⁴ per well in 96-well round-bottom plates) were cultured for 3 days at 37°C in 5% CO₂ with irradiated spleen cells as APCs (2 × 10⁵ per well) and anti-CD3 antibody at 10 μg/ml, with or without CD4⁺CD25⁺ T-cells (≤2.5 × 10⁴ cells per well) at an optimal 1:1 regulatory-to-responder cell ratio, as shown in previous publications (26,27,31,34). The cell cultures were pulsed on day 3 with 0.5 μCi [³H]thymidine for the last 18 h.

Spleen cells or pancreatic lymph node cells were collected from sick or nonsick NOD mice and pooled. Splenic CD90⁺ and CD90⁻ were purified, using CD90 beads and MS columns (Miltenyi Biotech). In some experiments, spleen cells were depleted of CD25⁺ cells by incubation with 7D4 cell (American Type Culture Collection) supernatant and rabbit complement (Cedarlane, Hornby, Canada). We intravenously injected 40 ×10⁶ CD90⁻ cells or 10 × 10⁶ CD90⁺ splenic cells into 4-week-old female NOD mice. Blood glucose was monitored weekly until 30 weeks of age, and diabetes incidence was determined when the glucose measurement was >33 mmol/l for 2 consecutive weeks.

Data were analyzed using either Student’s t test, ANOVA and the Tukey-Kramer multiple comparisons test, or the Mann-Whitney U test. Experiments were performed at least twice with reproducible results. One representative experiment is shown in each figure.

Controversial data have been reported concerning the proportion of CD4⁺CD25⁺ regulatory T-cells in the lymphoid organs of NOD mice. One study has shown that the percentage of CD4⁺CD25⁺ regulatory T-cells is normal in the periphery of NOD mice (35). In contrast, lower levels of CD4⁺CD25⁺ regulatory T-cells have been found in autoimmune disease–prone mice, including NOD mice, and could be associated with autoimmune disease development (8,9). To evaluate CD4⁺CD25⁺ T-cells in our colony, lymph nodes and spleens from NOD or B6 mice were analyzed for the presence of CD4⁺CD25⁺ T-cells. Because timing could have been a reason for the conflicting findings reported above, female NOD and B6 mice of various ages were used for this study. Lymph nodes and spleens were collected from individual mice, and cells were labeled with PerCP–anti-CD4 and PE–anti-CD25 antibodies. We have also found that NOD mice exhibit a significant decrease in the percentage of CD4⁺CD25⁺ T-cells in lymph node and spleen by 9 weeks of age in comparison to B6 mice (Fig. 1B). However, in 3-week-old mice, only a slight difference could be found in the lymph node but not in the spleen (Fig. 1A).

Various molecules, including CTLA-4 and GITR, have been shown to be critical for the development and/or regulation of CD4⁺CD25⁺ regulatory T-cells (30). In addition, CD62L, a homing receptor for naive T-cell migration into the lymph node, is expressed by a good portion of CD4⁺CD25⁺ regulatory T-cells and can help distinguish between regulatory and activated cells because its downregulation generally indicates an activated or memory cell phenotype. We examined whether CD4⁺CD25⁺ regulatory T-cells from adult NOD mice at different stages of the disease (i.e., pre-diabetic or sick) expressed different levels of these molecules compared with CD4⁺CD25⁺ regulatory T-cells from age-matched B6 mice. Lymph node and spleens were collected from individual NOD or B6 mice and labeled with anti-CD4 and anti-CD25 antibodies and either anti-GITR, anti–CTLA-4, or anti-CD62L antibodies. Normal levels of CTLA-4, GITR, and CD62L were expressed by CD4⁺CD25⁺ regulatory T-cells harvested from either lymph nodes (Fig. 2) or spleens (online appendix Fig. 1A [available at http://diabetes.diabetesjournals.org]) of pre-diabetic NOD mice. The percentage of CD4⁺CD25⁺ cells expressing CD62L, however, was reduced in sick mice (>26 weeks) (Fig. 3 and online appendix Fig. 1B). These data indicate that although CD4⁺CD25⁺ regulatory T-cells from NOD mice express normal levels of GITR and CTLA-4, they express decreased levels of CD62L as they become sick.

A recent study has suggested that CD4⁺CD25⁺ regulatory T-cells from NOD mice may be not effective in in vitro assays that test for regulatory function (36). Because these experiments were performed in the presence of NOD APCs, and APCs from adult NOD mice have been shown to be abnormal at various levels (10,13–15), including aberrant cytokine production, defective antigen presentation, and differentiation, we examined whether defective regulatory cell function in vitro could be caused by a deficiency in the ability of NOD APCs to activate CD4⁺CD25⁺ regulatory T-cells. To test this possibility, we compared the ability of B6 APCs and NOD APCs to stimulate NOD CD4⁺CD25⁺ regulatory function in vitro. NOD CD4⁺CD25⁻ responder and CD4⁺CD25⁺ regulatory T-cells isolated and pooled from adult B6 or NOD mice were cultured with pooled B6 or NOD APCs and soluble anti-CD3 antibody. Figure 4A shows that CD4⁺CD25⁺ regulatory T-cells from NOD mice that were stimulated with NOD APCs (right section, fourth column) were much less efficient at suppressing CD4⁺CD25⁻ cells than B6 CD4⁺CD25⁺ regulatory T-cells stimulated with B6 APCs (left section, second column). Interestingly, when NOD CD4⁺CD25⁺ cells were stimulated with B6 APCs in vitro in the same experiment, they became as efficient as B6 CD4⁺CD25⁺ cells at suppressing responder cell proliferation (Fig. 4A, left section, fourth column). For comparison, we determined whether B6 CD4⁺CD25⁺ regulatory cells could be activated by NOD APCs in vitro. We found hat NOD APCs were also not able to effectively activate B6 CD4⁺CD25⁺ regulatory T-cells (Fig. 4A, right section, second column). Percent inhibition of proliferation was represented in Fig. 4B to aid in the interpretation of the data. Altogether, these data suggest that CD4⁺CD25⁺ regulatory T-cells from NOD mice are not intrinsically deficient. However, NOD APCs appear to deliver a nonoptimal activation signal to CD4⁺CD25⁺ regulatory T-cells, thereby preventing optimal regulation. Although NOD APCs have been shown previously to be impaired in their ability to stimulate responder T-cells and a cell population referred to as suppressor T-cells (37), the current report shows for the first time that NOD APCs lack the ability to stimulate natural CD4⁺CD25⁺ regulatory cell function.

Numerous reports have shown that APCs from adult NOD mice are deficient at various levels (10,13–15), including aberrant cytokine production, defective antigen presentation, and differentiation. APCs from NOD mice could become deficient as the disease progresses, or they could be deficient from birth. APCs from individual 3-week-old NOD mice (i.e., before development of insulitis) and from individual 9-week-old NOD mice (i.e., after insulitis but before diabetes onset) were tested for their ability to induce 1) B6 responder cell proliferation and 2) CD4⁺CD25⁺ regulatory T-cell function. Responder cells stimulated with 3-week-old NOD APCs proliferated much less than when stimulated with 3-week-old B6 APCs (Fig. 5A), suggesting that NOD APCs have an impaired ability to activate T-cells in general. Because of this dramatic difference in proliferation, the regulation experiments performed with APCs from individual mice (10 per group) are presented in Fig. 5B as a percent inhibition of proliferation. As shown in Fig. 5B, the regulatory cell function (i.e., percent inhibition of proliferation) induced by 3-week-old NOD APCs was significantly reduced in comparison to that induced by APCs from age-matched B6 mice, and this was observed using either B6 or NOD CD4⁺CD25⁺ regulatory T-cells. These results suggest that APCs from NOD mice are deficient early in life at stimulating both responder and regulatory cells and that CD4⁺CD25⁺ regulatory T-cells may not be optimally activated even before insulitis develops. Similarly, APCs from 9-week-old NOD mice (pre-diabetic) (Fig. 5C) as well as APCs from 30-week-old sick NOD mice (Fig. 6B) induced reduced regulation in comparison to APCs from age-matched B6 mice.

The reason for which a small percentage of NOD mice (<10% in our facility) never develop diabetes is still not clear. We hypothesized that APCs from 30-week-old disease-free NOD mice may not be deficient in their ability to activate CD4⁺CD25⁺ regulatory T-cells and may thereby induce optimal regulation. To test this hypothesis, we collected spleens from individual NOD mice that were still disease free after 30 weeks, and we tested their ability to activate responder cell proliferation (Fig. 6A) and, more importantly, CD4⁺CD25⁺ regulatory T-cells (Fig. 6B). APCs from B6 or disease-free NOD mice could induce similar levels of CD4⁺CD25⁻ responder cell proliferation (Fig. 6A). However, proliferation was significantly decreased when responder cells were stimulated with APCs from sick NOD mice (Fig. 6A). Moreover, the inhibition of proliferation of responder cells by CD4⁺CD25⁺ regulatory T-cells was similar when cells were stimulated with APCs from B6 or disease-free NOD mice, but it was significantly less when cells were stimulated with APCs from diabetic NOD mice (Fig. 6B). Therefore, the APCs from disease-free NOD mice, in contrast to APCs from NOD mice that develop disease, appear to have the ability to optimally activate CD4⁺CD25⁺ regulatory T-cells and responder CD4⁺CD25⁻ cells.

Normal APC function in disease-free NOD mice could impact regulatory cells at several different levels, including the proper activation of CD4⁺CD25⁺ regulatory T-cells that can control pathogenic T-cells in vivo, as well as increasing the percentage of CD4⁺CD25⁺ regulatory T-cells present in the lymphoid organs. We first examined whether cells harvested from lymphoid organs of disease-free NOD mice (>30 weeks of age) had the ability to prevent diabetes development directly in NOD mice and whether these cells were CD4⁺CD25⁺ regulatory T-cells. As shown in Fig. 7A (line labeled NS CD90⁺), CD90⁺ cells purified from the spleens of disease-free NOD mice (nonsick) could prevent type 1 diabetes when transferred into 4-week-old NOD mice. However, when CD90⁺ cells were depleted of CD25 cells, they were incapable of preventing diabetes (Fig. 7A, line labeled NS CD90⁺CD25⁻) and in fact dramatically accelerated disease development. We found that 95% of CD90⁺ cells injected were CD3⁺, and 99% of CD90⁺ cells were CD25⁻ (online appendix Fig. 2A and B). In addition, splenic CD90⁻ cells from neither sick nor nonsick NOD mice could prevent disease when transferred into 4-week-old NOD mice (Fig. 7B).

Next, we examined whether disease-free NOD mice exhibit an increase in the percentage of CD4⁺CD25⁺ T-cells. Lymph nodes and spleens from individual disease-free and sick NOD mice (>30 weeks of age) were collected and cells were labeled with PerCP–anti-CD4 and PE–anti-CD25 antibodies as well as fluorescein isothiocyanate–anti-CD62L antibody because expression of CD62L by CD4⁺CD25⁺ cells was reduced in the spleens of sick NOD mice (Fig. 3). Flow cytometric analysis revealed that the percentage of CD4⁺CD25⁺ T-cells expressing CD62L^high was also significantly decreased in the spleens (Fig. 8A) of nonsick NOD mice. The percentage of CD4⁺CD25⁺CD62L^high cells was still significantly decreased in the lymph nodes (data not shown) and in the spleens (Fig. 8B) of nonsick NOD mice. Moreover, Foxp3 mRNA expression in unpurified spleen cells from age-matched nonsick and sick NOD mice showed no significant difference (Fig. 8C).

These data suggest that spleens of disease-free NOD mice possess pathogenic cells that are kept under control by functional CD4⁺CD25⁺ regulatory T-cells. Although disease-free NOD mice still exhibit a decrease in the percentage of CD⁺CD25⁺ cells, these cells have regained an optimal regulatory activity.

NOD mice have major defects in immunoregulation (3,4,6,8,9) and APCs (10,13–15), which could have a direct impact on the development of diabetes. In the current study, we show that the CD4⁺CD25⁺ regulatory T-cells from NOD mice do not have an intrinsic defect but that the NOD APCs are defective in their ability to stimulate or activate the CD4⁺CD25⁺ regulatory T-cells.

CD4⁺CD25⁺ regulatory T-cells were found in lower percentages in the peripheral lymphoid organs of NOD mice, as has been reported previously in adult mice (8,9). CD4⁺CD25⁺ regulatory T-cells from NOD mice expressed normal levels of some critical molecules associated with CD4⁺CD25⁺ regulatory T-cell development and regulation, including CTLA-4 and GITR. Moreover, NOD CD4⁺CD25⁺ regulatory T-cells downregulated the proliferation of responder CD4⁺CD25⁻ T-cells in vitro as efficiently as control CD4⁺CD25⁺ regulatory T-cells when stimulated with B6 APCs. These data suggest that the immunoregulatory defect may not be within the CD4⁺CD25⁺ regulatory T-cells themselves. Recent studies have reported that NOD CD4⁺CD25⁺ regulatory T-cells that have been activated in vitro or in vivo are capable of preventing diabetes development much more efficiently (32,36), suggesting that stimulation of CD4⁺CD25⁺ regulatory T-cells in NOD mice is not optimal. In contrast, transfer of CD4⁺CD25⁺ regulatory T-cells controls gastritis and oophoritis development in BALB/c and (A/JxB6)F1 mice, respectively, without requiring preactivation (21,38). Taken together, these data suggest that CD4⁺CD25⁺ regulatory T-cells require extra stimulation to control diabetes in NOD mice.

NOD mice have abnormalities that affect antigen presentation, including defects in the protein kinase C activation pathway (15) as well as low expression levels of CD80 and CD86 (39), which affect costimulation. Moreover, the decreased expression of CD86 on NOD APCs has been shown to impair T-cell activation (11). Our data show that the proliferation of CD4⁺CD25⁻ responder T-cells is dramatically impaired when stimulated with NOD versus B6 APCs, confirming that NOD APCs are impaired in their ability to stimulate the T-cell response. Because activation of CD4⁺CD25⁺ T-cells is absolutely required for regulation to take place in vitro (26,27), we hypothesized that NOD APCs may not be capable of activating CD4⁺CD25⁺ T-cells properly. The data presented in this study show that from as early as 3 weeks of age, APCs from NOD mice are impaired in their ability to activate CD4⁺CD25⁺ T-cell–mediated regulation. A recent study reported that the ability of diabetogenic effector cells to be counterregulated decreases with age in NOD mice (40). We did not find a decrease in the ability of CD25⁻ cells from 30-week-old NOD mice to be suppressed, as shown in Fig. 6. This discrepancy is likely caused by the fact that the 30-week-old mice used in the current study were disease free. Moreover, the percent inhibition shown in that recent report (40) is low (∼40–60%), corroborating our findings that regulation in vitro is not optimal when NOD APCs are used. The identification of the deficient APC type, as well as the deficient phenotype, is currently under evaluation. Our data indicate that NOD APCs do not optimally stimulate either CD4⁺CD25⁻ responder T-cells or CD4⁺CD25⁺ T-cells. This result could perhaps be explained by the decreased expression of costimulatory molecules (e.g., B7) on the surface of NOD APCs (11,39). When compared with intact NOD mice, CD28^−/− and B7^−/− NOD mice develop a higher incidence and earlier onset of diabetes that appears to be caused by a reduced number of CD4⁺CD25⁺ T-cells in the periphery (8). Moreover, B7 costimulation plays an important role in the maintenance of self-tolerance by sustaining CD4⁺CD25⁺ regulatory T-cells (41,42). Finally, conversion of CD4⁺CD25⁻ cells into CD4⁺CD25⁺ regulatory T-cells requires B7 costimulation (34). Altogether, these results suggest that deficiency in B7 expression by NOD APCs could be involved in abnormal induction of this cell population, and this may explain the impaired level of peripheral CD4⁺CD25⁺ T-cells in NOD mice and the development of disease.

A small percentage of NOD mice in our colony do not develop diabetes (5–10%). In contrast to NOD mice that develop disease, our data show that these disease-free NOD mice possess not only potent CD4⁺CD25⁺ T-cells but also functional APCs. Splenic APCs from disease-free NOD mice stimulated CD4⁺CD25⁻ responder T-cells and CD4⁺CD25⁺ mediated regulation as effectively as B6 APCs. Furthermore, splenic T-cells from disease-free NOD mice, after being transferred into 4-week-old NOD mice, were able to prevent disease in a significant percentage of mice. These data corroborate previous studies showing that CD4⁺ T-cells that reside in the spleens of young pre-diabetic NOD mice can prevent diabetes in a transfer model (43–45). We show in the current study that the cells responsible for disease protection appear to be CD4⁺CD25⁺ T-cells because depletion of CD25⁺ cells from splenic T-cells abolishes the ability of these T-cells to prevent diabetes development. Thus, our results demonstrate the presence, in the spleens of 30-week-old disease-free NOD mice, of potent APCs and CD4⁺CD25⁺ regulatory T-cells that are capable of controlling diabetogenic cells. There is a possibility that the impaired functional capacity of APCs in NOD mice with overt disease is a consequence (i.e., the result of exposure to high levels of glucose), not a cause, of their resistance to diabetes development. However, two of our observations suggest that this APC defect is unlikely to be a consequence of the disease: 1) APCs from 3-week-old NOD mice without disease symptoms (i.e., destruction of islet and high blood glucose) already exhibit impaired function (Fig. 5), and 2) APCs from 8-week-old transgenic (OVE26 on an FVB background) mice that express calmodulin under the rat insulin II promoter and consequently exhibit hyperglycemia (>600 mg/dl) for 8 consecutive weeks (46) exhibit normal function (data not shown). In a recent study, anti-CD3 antibody treatment has been shown to induce a potent population of CD4⁺CD25⁺ regulatory T-cells that could inhibit transfer of disease and gain the ability to optimally suppress the proliferation of CD4⁺CD25⁻ responder T-cells in vitro in the presence of NOD APCs (36). In the light of our findings, these data suggest that strong preactivation of CD4⁺CD25⁺ regulatory T-cells bypasses the need for optimal APC stimulation. We have not yet identified the APCs responsible for the increased CD4⁺CD25⁺ T-cell regulatory activity. Many studies have shown that NOD mice do not produce normal myeloid dendritic cells in vivo (13,38) and that mature dendritic cells transferred into NOD mice can prevent diabetes development (47–49). Moreover, mature dendritic cells are the only APCs that can trigger in vitro proliferation of CD4⁺CD25⁺ T-cells (50). For these reasons, there is a possibility that disease-free NOD mice possess a particular population of dendritic cells capable of activating CD4⁺CD25⁺ T-cells.

Together, the data presented in this study suggest that NOD mice that develop diabetes do not exhibit an inherent defect in CD4⁺CD25⁺ regulatory T-cells, but they possess a deficient APC compartment that does not allow for efficient activation of these regulatory T-cells. In contrast, APCs from NOD mice that do not develop diabetes are capable of activating CD4⁺CD25⁺ regulatory T-cells and induce potent CD4⁺CD25⁺ regulatory T-cells that keep pathogenic cells under control. Therefore, targeting the APC population may provide a therapeutic approach for inducing and maintaining peripheral tolerance in autoimmune disease–prone individuals.

This research was supported by a junior faculty award from the American Diabetes Association and a basic grant from the University of Louisville School of Medicine.

The authors thank Mike Myers for technical assistance, Chris Worth for cell sorting, and the staff of the animal facility for animal care.