Intrinsic Defects in the T-Cell Lineage Results in Natural Killer T-Cell Deficiency and the Development of Diabetes in the Nonobese Diabetic Mouse

All mice used in this study were purchased from the Jackson Laboratory (Bar Harbor, ME), and all antibodies were purchased from Pharmingen Canada (Mississauga, Ontario, Canada).

Thymocytes from 4-week-old donor mice were incubated with antibodies to CD4 (RM4-4), CD8 (53.6.7), and CD3 (145.2C11), washed, and then incubated with baby rabbit complement. Enriched triple-negative thymocytes were then stained with FITC-labeled antibodies to CD4, CD8, CD44, TCRβ, and PE (R-phycoerythrin)-labeled antibodies to CD25. The CD25⁺CD44⁻ thymic precursor cells were purified by cell sorting. The purified CD25⁺CD44⁻ cells (>95% pure) were suspended in a phosphate-buffered saline (PBS) solution at a concentration of 2.5 × 10⁸ cells/ml. Two- to 3-day-old NOD.scid mice were used as recipients of thymic precursor cells. Intrathymic injection was performed as described previously (29). Briefly, the recipient mice were anesthetized and immobilized, and a midline upper thoracic incision was made followed by a “V” excision of the upper sternum/ribcage under a stereomicroscope. Purified CD25⁻CD44⁺ cells (0.5 × 10⁶ cells/mouse, resuspended in 2 μl of PBS) were injected into both lobes of the thymus of each recipient under the stereomicroscope. As controls, littermates of the recipients received an injection of PBS solution alone.

Four weeks after injection of the precursor cells, T-cell reconstitution of the recipient mice was monitored weekly by fluorescence-activated cell sorter (FACS) analysis of peripheral blood mononuclear cells (PBMCs). PBMCs were stained with antibodies to TCRβ, Thy1.1, Thy1.2, CD4, CD8, and B220 (Pharmingen Canada). Lymph node cells and splenocytes of the NOD.scid recipient mice were also examined when the mice were killed. A small number of Thy1.2⁺ T-cells were detected in two recipients of AKR thymic precursor cells. These cells were likely of host origin that developed as a result of a “leaky” scid mutation. These mice were removed from the study. The expression of TCR Vβ chains in the thymocytes of recipient mice was examined by reverse transcriptase–polymerase chain reaction using a common primer for the C fragment and 17 primers specific for individual Vβ chains (30).

B-cells were isolated from splenocytes of 5- to 8-week-old NOD mice using B220 antibody–coated beads and a Mini MACS system (Miltenyi Biotec, Auburn, CA). The purity of B220⁺ cells was higher than 95%. The B-cells were resuspended in PBS (1 × 10⁸ cells/ml) and transferred into recipient mice by intraperitoneal injection.

In mixed lymphocyte reactions, 1 × 10⁶ responder splenocytes were mixed with 0.5 × 10⁶ irradiated stimulator cells and incubated for 3 days and pulsed for an additional 16 h with [³H]thymidine. To measure NKT cell function, the NOD.scid recipient mice received an injection of antibody to CD3 (10 μg/mouse i.p.). One hour after injection, spleens were removed and splenocytes were resuspended at 5 × 10⁶ cells/ml. The culture supernatant was collected at 2 and 4 h, and the concentration of IFN-γ and IL-4 in the supernatant was determined by enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, MN).

Blood glucose levels of nonfasting mice were measured as described elsewhere (31). The mean blood glucose level of nonfasting AKR mice was 7 mmol/l. In this study, nonfasting animals with blood glucose levels greater than 16 mmol/l for 3 consecutive days were scored as diabetic.

Lymph node cells from NOD.scid or BALB.scid recipients were preincubated with antibody against CD16/32 (2.4G2; Pharmingen Canada) in PBS containing 2% fetal calf serum. The cells were then incubated with biotinylated DX5 antibody on ice for 20 min. Control cells were incubated with dilution buffer. After washes, cells were incubated with streptavidin-perCP and anti–TCR-β–FITC antibody. The cells were resuspended and analyzed by FACS.

Diabetes-resistant NOR mice share an identical major histocompatibility complex (MHC) haplotype and most of the genome with NOD mice, and NOR mice may possess many defects in their immune system similar to those in NOD mice, including defective thymic education. To determine whether NOR mice develop an NKT cell deficiency as do NOD mice, we examined the DX5⁺TCR⁺ cell population in the thymus and periphery of NOR and NOD mice, because DX5⁺TCR⁺ cells represent the majority of NKT cells in NK1.1-negative strains (Y.Y., B.M., J.-W.Y., manuscript submitted). The DX5⁺TCR⁺ cell population in NOR and NOD mice was compared with that in AKR mice (Fig. 1) because the NKT cell population develops well in AKR mice. We found very few DX5⁺ T-cells in the thymus and periphery of NOR and NOD mice. In fact, the NKT cell population in NOD and NOR mice was similar to that found in AKR mice with a β2 m gene deficiency (AKR-β2 m mice; Fig. 1). NKT cell development in AKR-β2 m mice was defective because of a failure of expression of the CD1/β2 m complex. Analysis of the DX5⁺TCR⁺ cell population in various organs clearly showed that the NKT cell population is deficient in both NOD and NOR mice.

To determine whether the NKT cell deficiency in NOD and NOR mice is associated with the thymic microenvironment, we reconstituted the T-cell compartment of NOD.scid mice with thymic precursor cells isolated from NOD, NOR, and AKR mice (Fig. 2A). The purified thymic precursor cells were injected into the thymus of NOD.scid neonates at 3 days of age. Thymic precursor cells isolated from syngeneic and allogeneic donor mice developed equally well into mature T-cells in these NOD.scid recipients. T-cell populations were detected in the circulation of the recipients within 3–4 weeks after thymic injection. When the lymph node cells were analyzed, it was found that >80% of the lymph node cells in the reconstituted NOD.scid mice were T-cells (Fig. 2B). T-cells in the recipients of NOD and NOR thymic precursor cells expressed Thy1.2 antigen, whereas all of T-cells in the recipients of AKR thymic precursor cells expressed Thy1.1 antigen (Fig. 2B). No T-cells were detected in control NOD.scid mice. These results show that all T-cells in the recipients originated from donor precursor cells.

The spleens of the reconstituted mice contained a small proportion (5–20%) of T-cells, and most of the thymocytes were CD4 or CD8 single-positive cells; only a small number (<5%) were CD4⁺CD8⁺ double-positive cells (Fig. 2C), indicating that all of the donor thymic precursor cells developed rapidly into mature thymocytes in the recipients. The donor-derived T-cell populations remained for >24 weeks in the NOD.scid recipients. B-cells were virtually undetectable in these recipient mice.

We then analyzed the TCR Vβ chain repertoire in the reconstituted NOD.scid mice and found a diversified TCR Vβ chain usage in both the splenocytes and thymocytes of all reconstituted mice (Fig. 3a). As was previously found in NOD mice (32), both Vβ3- and Vβ17-expressing T-cells were missing in all of the NOD.scid recipients, indicating that the TCR repertoire of the NOD.scid recipients was shaped by NOD thymic education. Furthermore, the lymphocytes from NOD.scid recipients of AKR thymic precursor cells responded to irradiated splenocytes of AKR mice in a mixed lymphocyte reaction but not to NOD splenocytes (Fig. 3B). This result indicates that the mature T-cell populations that developed in the NOD.scid recipients were restricted by the host MHC expressed in the thymus.

We next compared the development of the DX5⁺ NKT cell population among the NOD.scid recipients of NOD, NOR, and AKR thymic precursor cells. Very few DX5⁺ NKT cells were found in the lymph node cells of NOD.scid recipients of NOD or NOR thymic precursor cells. In contrast, a significant number of DX5⁺ NKT cells were detected in the NOD.scid recipients of AKR thymic precursor cells (Fig. 4A). In these mice, the DX5⁺ NKT cells expressed Thy1.1 antigen but not Thy1.2, indicating that the DX5⁺ NKT cells were of AKR donor origin. In contrast, DX5⁺ NK cells did not express Thy1.1, but some of these cells expressed a low level of Thy1.2, indicating that the DX5⁺TCR⁻ NK cells were derived from the NOD.scid host.

Cytokines produced by NKT cells favor the Th2 immune response in TCR-mediated primary activation (21,22,33). To determine whether the NKT cell cytokine profiles differ among the reconstituted NOD.scid mice, we injected antibody to CD3 antigen into the NOD.scid recipients of NOD or AKR precursor cells and measured the production of IL-4 and IFN-γ in the splenocytes. Within 4 h of antibody injection, splenocytes from the recipients of AKR precursor cells produced a much greater amount of IL-4 (P < 0.001) and a similar amount of IFN-γ compared with the splenocytes from the recipients of NOD precursor cells (Fig. 4B). This result indicates that the NOD T-cell populations differ from the AKR T-cell populations in the ability to release IL-4 quickly after activation. Because the stimulation protocol used in these experiments largely targets NKT cells (21), this difference is likely to be due to the defective development of NKT cells from NOD thymic precursor cells.

To determine whether the reconstituted NOD.scid mice would develop insulitis and diabetes, we transferred purified B-cells from young NOD mice (4–5 weeks) into these recipient mice (4 × 10⁷/recipient), because B-cells are required for the development of T-cell–mediated autoimmunity in NOD mice (7,8,34,35). The transfer of NOD B-cells alone did not induce insulitis and diabetes in control NOD.scid mice. However, mild to severe insulitis developed in 10 of 12 NOD.scid recipients of NOD thymic precursor cells along with NOD B-cells (Tables 1 and 2), and five of these mice developed diabetes within 12 weeks. The total number of T-cells in the NOD.scid recipients was <20% (8 × 10⁶/mouse) of that found in normal NOD mice, and a very few transferred B-cells remained in these recipients, likely as a result of CD8⁺ T-cell–mediated rejection (35). The low number of T- and B-cells in NOD.scid recipients probably resulted in the relatively low incidence of overt diabetes. Mild to severe insulitis also developed in all nine NOD.scid recipients of NOR thymic precursor cells, and three of these NOD.scid recipients developed diabetes. However, insulitis was not detected in any of the eight NOD.scid recipient mice of AKR thymic precursor cells at 25 weeks of age in the presence of NOD B-cells (Tables 1 and 2).

Thus, insulitis and diabetes developed in almost all of the NOD.scid mice reconstituted with NOD and NOR thymic precursor cells, and these mice developed deficient NKT cell populations. In contrast, NOD.scid mice reconstituted with AKR thymic precursor cells contained a normal NKT cell population and were resistant to the development of insulitis and diabetes. To determine whether the resistance to insulitis and diabetes in the NOD.scid mice reconstituted with T-cells of AKR origin was associated with the development of the NKT cell population, we further reconstituted NOD.scid mice with thymic precursor cells from β2 m gene–deficient AKR mice. NKT cells are positively selected by CD1 antigen expressed on double-positive thymocytes (36), but the double-positive thymocytes of β2 m gene–deficient AKR mice do not express CD1 antigen (Fig. 5A). Therefore, the NKT cell development in these β2 m gene–deficient AKR mice was defective (Fig. 1). In addition, the CD8⁺ T-cell population did not develop in these β2 m gene–deficient AKR mice because of lack of the expression of class I MHC molecules (Fig. 5B). However, when thymic precursor cells from β2 m gene–deficient AKR mice were introduced into NOD.scid recipients, both CD8⁺ and CD4⁺ T-cell populations developed (Fig. 5B). The restoration of the CD8⁺ T-cell population was evidently due to positive selection by host MHC class I molecules. However, the NKT cell population was deficient in these recipient mice (Fig. 5C) because the CD4⁺CD8⁺ double-positive thymocytes were of donor origin, and these cells were unable to express CD1d for positive selection of NKT cells. In addition, these mice did not produce detectable IFN-γ or IL-4 after anti-CD3 antibody injection (data not shown). Examination of the pancreatic islets of these recipient mice revealed that all mice had developed insulitis (Table 2) and 6 of 10 had developed diabetes (Table 1). These results indicate that T-cells of AKR origin can induce diabetes in NOD.scid recipients if the NKT cell population does not fully develop and that thymic precursor–derived NKT cells play a critical role in the prevention of autoimmune diabetes in the reconstituted NOD.scid mice.

Because the NKT cell population in the reconstituted NOD.scid mice developed from donor thymic precursor cells, we asked whether the thymic precursor cells express DX5 antigen. We enriched CD25⁺ thymocytes from AKR mice and found that CD25⁺ thymocytes did not express DX5 antigen and DX5 antigen–expressing cells were present only in CD25⁻ thymocytes (Fig. 6A). We then examined the expression of CD1d antigen in NOD and AKR mice and found that the expression levels of CD1d antigen in CD25⁺CD44⁻ and CD4⁺CD8⁺ thymocytes, as well as in B-cells, were identical between the two strains (data not shown). This result indicates that the DX5-expressing NKT cells differentiated from late-stage premature T-cells after the injection of CD25⁺CD44⁻ thymic precursor cells into the thymus of recipient mice and that the defective development of the NKT cell population in NOD mice is not due to a low level of CD1d expression.

To determine whether a different thymic microenvironment has any effect on the development of the NKT cell population, we injected AKR or NOD thymic precursor cells (CD25⁺CD44⁻) into the thymus of neonatal BALB.scid mice and examined the development of the NKT cell population in the adult recipients. The numbers of NKT cells were very low in the BALB.scid recipients of NOD precursor cells. However, a well-developed NKT cell population was found in the recipients of AKR precursor cells (Fig. 6B). These results were similar to those obtained from the reconstituted NOD.scid mice. Therefore, defective development of the NKT cell population in these reconstituted mice resulted from the thymic precursor cells but not from the thymic microenvironment.

We next tested whether the defective differentiation of NOD thymic precursor cells can be corrected in the presence of AKR precursor cells by introducing mixed (1:1 ratio) NOD and AKR thymic precursor cells into NOD.scid recipients. T-cells of both AKR and NOD origin developed in these recipient mice, and there was a moderate increase in the number of NKT cells as compared with the number in recipients of NOD precursor cells alone. However, when the thymic precursor cells of AKR donors were irradiated and then mixed with NOD thymic precursor cells, only T-cells of NOD origin developed, and the NKT cell population was deficient (Fig. 6C). We also injected NOD thymic precursor cells mixed with irradiated CD4⁺CD8⁺ AKR thymocytes (1:2) into NOD.scid recipients and found only T-cells of NOD origin with a deficient NKT cell population (data not shown). Thus, cell-cell interaction with AKR thymocytes could not rescue the defective development of NKT cells from NOD thymic precursor cells.

A triple-negative thymocyte population (CD25⁺CD44⁻, CD117⁻, and CD3⁻CD4⁻CD8⁻) has been identified as the earliest precursor cell type irreversibly committed to the T-cell lineage; these cells do not develop into any other cells of hematopoietic origin (37,38,39,40). If CD25⁺CD44⁻ thymocytes are isolated from a desired donor and introduced into the thymus of NOD.scid recipient mice, then not only should T-cell populations developed in the recipient mice be of donor origin, but more importantly, the thymic education of the T-cell populations should be determined entirely by the thymic microenvironment of the recipient. Consequently, the mature T-cells will be restricted by the recipient’s MHC, and their immune response will be carried out in conjunction with the antigen-MHC complex of the recipient’s antigen-presenting cells. On the basis of this concept, we have developed a unique animal model in which the genetic background of the T-cell populations can be different from that of all of the other cell populations in the animal.

The analysis of reconstituted NOD.scid mice showed that both syngeneic and allogeneic CD25⁺CD44⁻ thymic precursor cells developed equally well into mature T-cell populations under an identical NOD thymic education. More importantly, the development of T-cell populations in the recipient mice was shaped by NOD thymic education. As a consequence of the host thymic education, the T-cells of the NOD.scid recipients, regardless of donor genetic background, recognized the host MHC but not the donor MHC as self. Therefore, using our unique animal model would allow us to distinguish the different pathogenic roles of thymic education and intrinsic defects within the T-cell lineage of NOD mice when T-cells of different origins were educated by an identical thymic microenvironment.

To determine whether the defective thymic education contributes to the NKT cell deficiency in NOD mice, we first examined the development of NKT cells in NOR mice, because NOR mice share an identical MHC locus and most of the genome with NOD mice. We found that NKT cell development is defective in both NOD and NOR mice. Because most NKT cells develop from the thymus, the defective development of the NKT cell population in NOD and NOR mice may be due to their MHC or other defective factors in their thymic microenvironment. It has been shown that IL-7, which is a product of thymic stromal cells, is required for the development and maturation of NKT cells (41,42). Exogenous IL-7 improved the IL-4 production of NOD thymocytes (43), suggesting defective NOD stromal cell functions.

To determine the origin of the NKT cell deficiency in NOD and NOR mice, we reconstituted both NOD.scid and BALB.scid mice with thymic precursor cells from NOD, NOR, and AKR donors. If the NKT cell deficiency is a result of defects in NOD thymic education, then the reconstituted NOD.scid mice would develop deficient NKT cells, whereas the BALB.scid recipient mice would have normal NKT cell development regardless of the origin of thymic precursor cells. In contrast, if intrinsic defects within the T-cell lineage of NOD mice result in the NKT cell deficiency, then this deficiency will be associated only with T-cells of NOD origin. We found that a normal NKT cell population developed from AKR thymic precursor cells. In contrast, a deficient NKT cell population developed from NOD and NOR thymic precursor cells in both NOD.scid and BALB.scid mice. These results show that the genetic defects in the thymic precursor cells of NOD and NOR mice are the major cause of the NKT cell deficiency, whereas the thymic microenvironment of NOD mice does not block NKT cell development.

We further asked whether the NKT cell deficiency is due to the defective expression of CD1 and other selecting factors on the surface of NOD thymocytes, because it has been shown that NKT cells are positively selected by CD1/β2 m complex expressed on CD4⁺CD8⁺ thymocytes (44). However, we found that the expression level of CD1d on CD4⁺CD8⁺ thymocytes was identical in NOD and AKR mice. In addition, we found that the defective NKT cell development cannot be rescued when the NOD thymic precursor cells mixed with irradiated AKR thymic precursor cells or CD4⁺CD8⁺ thymocytes before injection into NOD.scid recipients. Thus, the failure in NKT cell development is unlikely to be due to defects of cell surface markers. Instead, defects in the intracellular signals required for NKT cell differentiation may be responsible for the deficiency.

Consistent with our results, both Src family tyrosine kinase Fyn and transcription factor Ets1 have been shown to be required for the development of the NKT cell population (45,46). Both Fyn and Ets1 are involved in the intracellular signal pathway for NKT cell development. Our preliminary results showed that the T-cell lineage of NOD mice expresses the Ets1 gene at a level identical to that in T-cells in AKR mice and that there is no mutation in the Ets1 gene coding frame (unpublished data). However, it was reported that polymorphism in the 3′ nontranslating region of Ets1 may be associated with autoimmune syndromes (47). The mechanisms of the NKT cell deficiency in NOD mice remain to be elucidated. However, our results provide invaluable information to facilitate identification of the gene(s) responsible for the NKT cell deficiency. In addition, the results of the current study provide direct evidence that NKT cells develop from T-cell–committed precursor cells. It is known that NKT cells have a close ontogenic relationship with the T-cell lineage (48,49), but it was unclear from which precursor cells the NKT cells differentiate. We found that NKT cells develop from CD25⁺CD44⁻ thymic precursor cells, in which a TCR gene rearrangement has taken place (38). However, these thymic precursor cells do not express the NK cell marker DX5 antigen (Fig. 6A). These thymocytes are precursor cells for both T and NKT cells, and our results further indicated that during development, TCR expression occurs before the expression of NK cell markers. The expression of NK cell markers may be an indicator of NKT cell maturation (48).

In NOD.scid recipient mice reconstituted with T-cells of different origins, a sharp contrast among groups of recipient mice was observed with respect to the development of insulitis and diabetes. The NOD.scid recipients of NOD and NOR thymic precursor cells developed insulitis, and some of these mice became diabetic, but none of the NOD.scid recipients of AKR precursor cells developed insulitis. Because the donor thymic precursor cells experienced an identical thymic education mediated by NOD MHC molecules, all of the three groups of NOD.scid recipient mice should contain a similarly high frequency of autoreactive T-cells against β-cell autoantigens (50,51,52). However, these autoreactive T-cells were unable to attack pancreatic β-cells in the NOD.scid recipients of AKR thymic precursor cells. This result provides direct evidence that the intrinsic defects of the NOD T-cell lineage play a critical role in the pathogenesis of autoimmune diabetes, and the replacement of the T-cell lineage of NOD mice with the one of different origin can prevent autoimmune diabetes. Conversely, the development of insulitis and diabetes in the NOD.scid mice reconstituted with thymic precursor cells from β2 m gene–deficient AKR mice indicated that T-cells of different origin can be diabetogenic under NOD thymic education when the regulatory factors are removed. Furthermore, our results showed the NKT cell deficiency is strongly associated with the development of autoimmunity in the reconstituted NOD.scid mice. The deficient NKT cell population developed from NOD and NOR thymic precursor cells in the NOD.scid recipients correlated with the development of autoimmune diabetes, and the resistance to diabetes in the NOD.scid reconstituted with AKR thymic precursor cells was abolished when the development of NKT cells was defective as a result of the β2 m gene deficiency.

Therefore, consistent with other studies (27,28), our results showed that NKT cells can suppress the development of autoreactive T-cells and that the NKT cell deficiency contributes to the pathogenesis of autoimmune diabetes in NOD mice. However, the underlying mechanisms by which NKT cells control autoimmunity are still not well understood. The defective production of both IL-4 and IFN-γ by NKT cells has been indicated as a pathogenic role (19,27). We detected a higher IL-4 production in splenocytes of NOD.scid mice containing T-cells of AKR origin. However, we also found that NKT cells purified from spleens of AKR mice produce a large dose of IFN-γ in culture (unpublished observation). More detailed analysis will be conducted using thymic precursor cells from IL-4 or IFN-γ gene-deficient donors to understand which cytokine is important or whether the cytokine profile is key in NKT cell–mediated immunoregulation.

NOR and NOD mice share ∼88% of the genome, including the MHC locus that confers susceptibility to autoimmune diabetes (53,54). Although NOR mice are resistant to insulitis and diabetes, the NOD.scid recipients of NOR thymic precursor cells developed insulitis and diabetes, as well as a deficient NKT cell population. It is very likely that the T-cell lineage in NOR mice shares genetic defects with those of NOD T-cells and that NOR T-cells are diabetogenic when they are developed and function in the immune system of NOD mice. Thus, diabetes resistance in NOR mice is not directly related to the T-cell lineage but to other immunocytes, such as antigen-presenting cells. This finding is consistent with other studies showing differences in the expression and regulation of the MHC genes and the function of antigen-presenting cells between NOD and NOR mice (54,55,56,57).

Autoimmune diabetes in NOD mice is a polygenic disease, and multiple genetic defects have been associated with its pathogenesis. However, the biological effects of the genetic defects and the cell population expressing a specific genetic defect remain largely unknown. By replacing the T-cell lineage in NOD mice with lineages from different genetic backgrounds, we were able to identify the pathogenic roles played by the intrinsic defects of T-cells in NOD mice. The results of this study will be useful for further identification of the genes involved in NKT cell development and NKT cell deficiency in NOD mice. Furthermore, the unique animal model developed for this study, in which specific cell lineages of the immune system can be made genetically distinct from each other, will be invaluable for studies on the development of the immune system and for the determination of genetic factors involved in immune regulation and dysfunction, such as those that occur in autoimmune diseases and inflammatory responses.

We are grateful to Dr. Merle Olsen for advice and help with animal care, Laurie Robertson for assistance with flow cytometry, and Dr. Ann Kyle and Karen Clarke for editorial assistance. This work was supported by a grant from the Canadian Institutes of Health Research (MA 9584). Y.Y. is a recipient of the Career Development of Award from the Juvenile Diabetes Research Foundation International. J.-W.Y. holds a Canada Research Chair in Diabetes.