Soluble egg antigen
Soluble worm antigen
Infection with Schistosoma mansoni (S. mansoni) or exposure to eggs from this helminth inhibits the development of type 1 diabetes in NOD mice. In this study we show that soluble extracts of S. mansoni worm or egg completely prevent onset of type 1 diabetes in these mice but only if injection is started at 4 weeks of age. T cells from diabetes-protected mice make IL-10 in recall responses to parasite antigens. These cells are furthermore impaired in their ability to transfer diabetes to NOD-SCID recipients. Bone marrow dendritic cells derived from NOD mice are found to make more IL-10 and less IL-12 following culture with S. mansoni soluble egg antigens in conjunction with lipopolysaccharides. NOD mice are deficient in NKT cells. Soluble worm and egg antigens increase the numbers of Vα14i NKT cells in NOD mice. These effects of schistosome antigens on the innate immune system provide a mechanism for their ability to prevent type 1 diabetes in NOD mice.
Type 1 diabetes is an autoimmune disease where the pancreatic β cells are destroyed by the immune system. The non-obese diabetic (NOD) mouse provides an excellent model of type 1 diabetes permitting genetic and immunological analyses of factors determining development of this autoimmune condition 1. It is known that the development of this Th1-mediated autoimmune disease is governed by both genetic and environmental influences. The concordance for diabetes development in identical twins is around 40% highlighting a role for an environmental modifier 2. The prevalence of type 1 diabetes has been increasing in the developed world but it is uncommon in tropical regions where diseases such as leprosy, leishmania, schistosomiasis and malaria are common. In rodent models of type 1 diabetes infection with virus, bacteria and helminths have been shown to inhibit the development of autoimmunity 3–6. Observations such as these in type 1 diabetes and those in asthma have lead to the "hygiene hypothesis" where it is envisaged that prevention of childhood infections may predispose to autoimmune or allergic responses 6–8. The mechanism(s) by which infections might influence autoimmune disease remain to be fully clarified. In the case of the Th1-mediated type1 diabetes it could be envisaged that helminths might deviate the islet-specific response towards a non-pathogenic Th2 response 9. This explanation cannot, however, account for the protection afforded by mycobacterial or some viral infections.
Our previous studies showed that infection with S. mansoni or injection of the parasite's eggs could completely prevent diabetes onset in NOD mice 6. To understand the mechanism by which this helminth infection might influence the spontaneous development of type 1 diabetes in NOD mice we have established the timing requirements for exposure to S. mansoni antigens. We also examined whether comparable protection from autoimmunity was afforded by soluble extracts of S. mansoni eggs (SEA) or worms (SWA). There are data showing that glycolipids known to be present in S. mansoni eggs can elicit IL-10 from peripheral blood monocytes of healthy donors 10. Our data show that SEA in conjunction with LPS increased IL-10 production from NOD bone marrow-derived dendritic cells (DC) and correspondingly diminished IL-12 production. NOD mice have a deficiency in Vα14i NKT cells and expansion of this population has been shown to prevent diabetes onset 11, 12. The adult S. mansoni worm contains the glycolipid galactosyl ceramide 13 and we have additionally explored the possibility that SWA or SEA may expand the Vα14i NKT cell population in NOD mice which is known to be reactive with this glycolipid and thus influence development of autoimmunity. At a late time point after exposure to SWA and SEA we have been able to show that T cells themselves make IL-10. IL-10 production by T cells has been shown to potently regulate diabetes development 14. The ability of infection to prevent onset of type 1 diabetes may therefore be through a range of mechanisms including those initially acting on the innate immune response but subsequently on T cell function.
2.1 Only four injections of S. mansoni eggs are necessary to prevent type 1 diabetes in NOD mice
Cohorts of 4-week-old female NOD mice were injected once a week with 10,000 freeze thawed (dead) S. mansoni eggs per mouse. Control mice were injected with PBS. We found that exposure to 20 injections of S. mansoni eggs prevented onset of type 1 diabetes in NOD mice (data not shown). To determine whether an abbreviated course of injections would provide comparable protection, NOD mice were injected either for 4 or for 10 weeks with S. mansoni eggs. Fig. 1a shows that as few as four injections were sufficient to prevent type 1 diabetes in NOD mice. None of the injected mice developed diabetes whereas the control mice developed diabetes at 70% incidence by 27 weeks of age (p<0.005).
2.2 The timing of S. mansoni egg administration influences diabetes onset in NOD mice
NOD mice develop an infiltrate of mononuclear cells in their pancreases from 5 weeks of age and develop diabetes usually from 12 weeks of age onward. To determine whether administration of S.mansoni eggs at a later time point could also prevent onset of type 1 diabetes egg injections were commenced in NOD mice at 8 weeks of age. It can be seen from Fig. 1b that neither a short nor a long course of S.mansoni eggs could significantly prevent diabetes development in NOD mice when treatment was initiated at 8 weeks of age.
2.3 Soluble worm or egg antigens from S. mansoni prevent diabetes onset in NOD mice
Soluble extracts of SEA or SWA have been used to study immune responses to S. mansoni15–17. To establish whether antigens within these extracts are able to influence the development of diabetes, groups of 4-week-old female NOD mice were injected once a week for 15 weeks with 50 μg/mouse of SWA or SEA. A control group was comparably injected with PBS. It can be seen from Fig. 2 that mice exposed to SEA or SWA are also completely protected from developing type 1 diabetes (p<0.005). Similar to the egg treatment protocol (Fig. 1), four weekly injections with SWA and SEA, effectively prevented the development of type 1 diabetes, when the injections were initiated at 4 weeks of age (data not shown). Egg and soluble antigen treatment did not completely prevent insulitis. Nevertheless if compared to the age-matched non-diabetic controls, the pancreases of S. mansoni antigen-treated mice had a higher percentage of non-infiltrated islets and a lower percentage of intra islet infiltrate (Fig. 3).
2.4 Exposure to S. mansoni antigens in vivo reduces the capacity of splenocytes from NOD mice to transfer diabetes into NOD-SCID recipient mice
At 12 weeks of age, NOD female mice are in a pre-diabetic state and by the age of 30 weeks 70-80% of the female mice in our colony have become diabetic. As expected, injection of splenocytes from diabetic NOD mice rapidly transfered type 1 diabetes to NOD-SCID recipients with 100% incidence (Fig. 4). Fig. 4 also shows that splenocytes from either 12-week-old pre-diabetic or 12-month-old non-diabetic NOD mice can transfer diabetes into NOD-SCID recipients with 100% incidence. After in vivo exposure to S. mansoni eggs, SEA, or SWA, splenocytes from injected mice had a significantly reduced capacity (p<0.05) to transfer diabetes into NOD-SCID recipients (40%, 40% and 0%, respectively).
2.5 Splenocyte proliferation and cytokine secretion in response to in vitro restimulation with SWA and SEA
Splenocytes from NOD mice that had received three injections of S. mansoni eggs and splenocytes from PBS-injected controls were restimulated in vitro with SWA and SEA. As expected there was a lymphoproliferative response to SWA and SEA (Fig. 5a) only in the group of mice that received S. mansoni eggs in vivo. Additionally, there was a diminished proliferative response to Con A in the mice that had been exposed to S. mansoni eggs in vivo (Fig. 5b). It can be seen from Fig. 5c–g that splenocytes from S.mansoni egg-injected mice secreted IL-4, IL-5, IL-10 and IL-13 after in vitro restimulation with SWA and SEA antigens. IFN-γ secretion was detected in vitro in response to SEA and Con A. Since SEA induced IFN-γ secretion in both egg-primed and unprimed splenocytes, this suggest that production of this cytokine was a direct response to in vitro SEA. Collectively. the data suggest that cytokine secretion following Con A stimulation showed a Th2 bias following egg antigen administration (p<0.05).
2.6 T cells and not B cells are the primary responding cells to S. mansoni antigens in vitro and they are responsible for IL-10 secretion
Seven-month-old NOD mice were sacrificed 12 weeks after the last of 20 S.mansoni egg injections. Fig. 6a shows that a proliferative response to SWA and SEA was still detectable in vitro from spleen cells even at this late stage after in vivo egg immunization. To establish whether the T or the B lymphocytes were proliferating and secreting high levels of cytokines in response to SWA and SEA antigen, single-cell splenocyte suspensions were depleted of either B or T cells prior to in vitro culture. From Fig. 6b it can be seen that following T cell depletion the proliferative response to SWA and SEA was lost. By contrast, when the B cells were removed from the splenocyte preparation the proliferative response to SWA and SEA remained present and was comparable with that observed with the non-depleted total splenocyte cell preparation (Fig. 6c and a, respectively). Consistent with the proliferation data, Fig. 7 shows that T cell depletion profoundly influenced secretion of all measured cytokines. IFN-γ, IL-4, IL-10 and IL-13 were present in both undepleted and the B cell-depleted splenocytes cultures although it was clear that B cell depletion had also influenced the production of IL-5 and IL-13. In the case of IL-10 there was some indication that SEA could elicit some B cell-derived IL-10 production in the in vitro recall response. This is in accord with the observation of Harn and colleagues that B cells make IL-10 in response to egg antigens 18. These data show that many weeks after egg antigen administration, a polarized cytokine response can be recalled.
2.7 Bone marrow-derived DC from NOD mice make reduced levels of IL-12 and increased levels of IL-10 following exposure to SEA but not SWA
When stimulated by LPS, NOD mouse bone marrow-derived DC secrete high levels of IL-12 and TNF-α (unpublished observations). To determine whether SEA were able to influence immune responses at an early stage through an effect on DC, we cultured bone marrow-derived DC from NOD mice with SWA and SEA, in the presence or in the absence of LPS. Fig. 8a and c show that the addition of either SEA or SWA in the absence of LPS had no direct effects on DC IL-10 and IL-12 secretion. However, in the presence of LPS SEA significantly elevated IL-10 and diminished IL-12 secretion (p<0.05). In contrast, SWA did not appear to demonstrate a significant effect on these cytokines. SEA did not appear to be simply effecting a kinetic shift in IL-10 and IL-12 secretion since the modulation appeared consistent over several time points (Fig. 8b and d). Flow cytometric analysis of DC revealed that neither SEA nor SWA had any apparent effect on the expression levels of maturation markers CD40, CD80, CD86 or class II (data not shown). LPS treatment increased the expression levels of these markers but the addition of either SEA or SWA in conjunction with LPS did not noticeably change the expression pattern (data not shown).
2.8 SWA and SEA treatment increases the number of NKT cells in NOD mice
We were unable to demonstrate an effect of SWA on DC function in vitro. In an attempt to understand how SWA might influence the immune response we explored its effect on NKT cells, another population of cells that can be involved early in an immune response and can profoundly influence its outcome. This population of lymphocytes is known to rapidly produce cytokines including IL-4 19. NKT cells have been shown to be deficient in NOD mice 19–21. The observation that transfer of NKT cells into NOD mice protected them from diabetes development suggested that their deficiency might be a predisposing factor in diabetes development 22. NKT cells recognize glycolipids presented by the class I-like molecule, CD1. One of the glycolipids recognized by NKT cells is α-galactosyl β ceramide and administration of this glycolipid to young NOD mice has been shown to prevent diabetes onset through increase in NKT cell numbers 11, 12, 23–26. Schistosome eggs and worms express a range of glycoconjugates and interestingly one of those shown to be present in the worm is galactosyl-ceramide 13. This observation prompted us to investigate whether the protection from diabetes development that we observe following SWA is due to an effect on NKT cell numbers. NKT cell numbers can be most effectively evaluated using tetramers comprising CD1d and α-galactosyl β-ceramide 27.
This is particularly relevant to NKT cell evaluation in NOD mice as this strain does not express the NK1.1 marker 28. Young NOD mice were therefore injected with SWA or SEA and NKT cell numbers evaluated in the spleen and liver. Very few α-galactosyl CD1d staining cells could be detected in the spleens (data not shown) of any of the mouse groups but they could be clearly observed in the liver (Fig. 9). Fig. 9a shows that the C57BL/6 normal mouse strain has higher numbers of liver NKT cells compared to the NOD. Injection of the NOD mice with either SWA or SEA significantly elevated the number of liver NKT cells (Fig. 9b and Table 1) (p<0.05). This suggests that one possible mechanism for SWA prevention of type 1 diabetes in NOD mice might be through increased NKT cell activity.
|Sample/treatment||% Posistive for α-gal CD1d-tetramer staining (individual mice)||Median||Comments|
|C57BL/6||10.5, 10.8, 10.37, 14.52, 14.94||10.80||ND|
|NOD||2.93, 2.93, 3.00, 3.89, 3.35, 4.45, 8.79, 7.15, 1.97, 4.25, 4.80||3.89||–|
|NOD/SWA||4.04, 6.34, 4.32, 8.95, 8.43, 10.30, 10.64, 12.51, 8.02, 2.22, 3.71||8.02||Significantly greater staining than NOD (p < 0.05)|
|NOD/SEA||2.43, 4.18, 6.38, 6.99, 8.15, 6.84, 4.62, 9.17, 3.38, 11.06, 7.60||6.84||Significantly greater staining than NOD (p < 0.05)|
The development of type 1 diabetes is influenced by both genetic and environmental factors. There is evidence suggesting that infection may be one such environmental modifier in that it plays a role in preventing onset of this autoimmune condition in rodent models and possibly also in humans 5, 8, 29. As our previous studies showed that schistosome infection prevented onset of diabetes in NOD mice 6 we carried out a series of experiments to try to identify schistosome components that mediate diabetes preventionand explored possible mechanisms of protection. The data presented here show that soluble extracts from the eggs of S mansoni are able to mediate diabetes prevention but this is only achieved if NOD mice are exposed to schistosome antigens at around 4 weeks of age. The critical time point appeared to be prior to the onset of pancreatic infiltration. Only four injections of antigen are required to prevent diabetes onset and the protection that is achieved is long lasting (at least to 12 months of age). Interestingly, lymphocytic infiltration into the pancreas still occurs but remains largely peri-islet. Furthermore transfer of splenocytes from egg-, SWA-, or SEA-protected mice into NOD-SCID recipients showed that compared to age matched diabetic or non-diabetic mice there was a failure to transfer type 1 diabetes. This suggests that there is either a loss of diabetogenic effector cells or the establishment of some dominant regulatory influence which maintains self tolerance in injected mice.
To understand the basis for this loss of diabetogenic potential, cytokine production was examined at both early and late time points after exposure to schistosome antigens. At the early time points following exposure to S. mansoni eggs there was clear evidence, as expected from studies in other mouse strains, of proliferative responses and Th2-biased cytokine production in recall responses to SEA and SWA. Interestingly, both the in vitro proliferative response and IFN-γ production elicited by Con A were reduced in schistosome egg-injected mice. At the later time point of 12 weeks after the last egg injection, the recall responses to SEA and SWA were still evident as was the diminished proliferative response to Con A (data not shown). This proliferatve response was clearly T cell dependent, as was the production of IFN-γ, IL-4 and IL-10, despite the fact that in BALB/c mice B cells have been shown to differentially secrete IL-10 in response to SEA and antigen preparations derived from other parasites 30. The production of IL-10 at this time point is of considerable interest as it may indicate the presence of regulatory T cells and a potential role for such cells in the reinforcement of regulatory circuits to prevent type 1 diabetes.
Our study of DC responses to soluble schistosome antigens suggested a potential role for IL-10 also in the early stages of exposure to S. mansoni-derived antigens. In the presence of a DC maturation signal (LPS), NOD bone marrow-derived DC make large amounts of IL-12. If SEA is additionally present the profile of cytokine production is altered and the DC make less IL-12 and correspondingly more IL-10. In agreement with others we could find no evidence of any direct effect of SEA alone on DC maturation as measured by surface-antigen expression 31 or cytokine production. SWA did not appear to influence cytokine production by DC under any culture conditions.
This effect of schistosome antigens on DC function coupled with the need to treat NOD mice early in the evolution of type 1 diabetes suggests that one effect of S mansoni eggs might be to influence presentation of autoantigen to diabetogenic effector cells. Pancreatic infiltration is not halted by early exposure of NOD mice to schistosome antigens but as diabetes fails to develop this infiltrate is clearly benign and unable to mediate β cell destruction. The long-term protection which then ensues together with T cell IL-10 production, suggests that tolerance may be maintained by regulation. The fact that SWA and SEA have differential effects on DC function furthermore suggests that SWA may exert its influence on the evolution of diabetes in NOD mice in a different way than SEA. NOD mice are deficient in NKT cells and injection of NKT cells into NOD mice prevents type 1 diabetes development 32. It is known that NKT cells are CD1 restricted. Deletion of CD1-positive cells accelerated onset of diabetes in NOD mice further supporting a role for NKT cells in diabetes control 33. It has been shown that NKT cells recognizeα-galactosyl-ceramide bound to CD1d 19. As it is known that S. mansoni worms contain galactosyl-ceramide we examined the possibility that this may be an active component of SWA in mediating diabetes protection. It has been shown that injection of α-galactosyl-ceramide into NOD mice increases the NKT population in a CD1-dependent manner and prevents onset ofdiabetes. Intriguingly, we find that SEA and SWA not only prevent diabetes in NOD mice but both significantly increase NKT cells in the liver. This strongly suggests that this may be an additional mechanism by which schistosome antigens can prevent onset of diabetes in NOD mice. Recent studies have shown that CD1d-restricted presentation of SEA is necessary for the induction of a Th2 responseand the development of egg-induced fibrosis 34. Our studies highlight an increase in NKT cell numbers following SEA or SWA injection in accord with the CD1-dependent effects noted in thispaper. It may therefore be that the range of host responses which are normally induced by infection with S. mansoni or exposure to its antigens are such that they enable peripheral tolerance to β cell antigens to be reinforced. Effects on the innate immune system may therefore not only possibly deviate T cell responses to the pancreatic beta cell towards a benign Th2 response butalso influence the development of regulatory T cells. In conclusion, our studies show that prevention of diabetes in NOD mice following exposure to schistosome antigens might arise through interaction of parasite-derived molecules with several arms of the innate immune system of the host.
4 Materials and methods
NOD and NOD-SCID mice were obtained from breeding colonies established in the Pathology Department, University of Cambridge animal facilities. C57BL/6 and TO mice were obtained from Harlan. During the experiments all the mice were maintained in the same standard conditions at the Department of Pathology with free access to food and water.
4.2 Preparation of SEA and SWA and antigen treatment
S. mansoni worms were recovered from portal venous vasculature, and parasite eggs from the livers of infected outbred mice using previously described techniques 35.SWA and SEA were prepared from the respective stages of the S. mansoni life cycle according to previously published protocols 35. NOD mice were injected i.p. with freeze-thawed S. mansoni eggs (10,000/injection) or with 50 μg/mouse of SWA or SEA. Depending on the experiment the injections were given (starting at 4 or 8 weeks of age) once a week for a period of 4, 10, 15 and 20 weeks.
4.3 Histological examination
Pancreases were processed for wax histology. Five-micrometer sections were taken at eight levels (200 μm apart) and stained with hematoxylin and eosin. Total islets per section were counted and the degree of cellular infiltration was scored. Scoring was as follows: non-infiltrated islet, peri-islet infiltrate (i.e. up to 20% of the islet is infiltrated) and intra islet infiltrate (21–100 % of the islet is infiltrated).
4.4 Adoptive transfer of type 1 diabetes into NOD-SCID mice
Splenic cells (2×107/mouse) from different groups of S. mansoni antigen-treated NOD female mice were injected intravenously into 6-week-old NOD-SCID female mice. Recipient NOD-SCID mice were tested weekly for development of diabetes.
4.5 T and B cell depletion
Spleens were taken from NOD mice injected with S. mansoni eggs or PBS at different time points after the antigen injections. T cells were depleted by complement-mediated lysis and B cells depleted using magnetic beads. For depletion of T cells, single-cell suspensions were prepared and resuspended in 0.5 ml of rat anti Thy-1 (clone YTS 148) from tissue culture supernatant plus 0.2 ml of guinea pig complement (Harlan Sera-lab LTD, Loughborough, GB). The cells were then incubated at 37°C on a rotor and after 30 min washed twice in RPMI (Gibco). This technique was able to delete more than 99% of CD3+ cells. For B cell depletion, splenocytes were resuspended in 0.5 ml of RPMI with 1.5 μg of biotin anti-mouse CD45R/B220 antibody (PharMingen, San Diego, CA). Cells were incubated at 4°C on a rotor for 30 min. Cells were washed then resuspended in 500 μl of PBS-BSA (1%)-EDTA(0.05%) in buffer containing streptavidin-coated magnetic beads (Dynal, Oslo, Norway), at a ratio of 4 beads/cell and mixed on a rotor for 30 min at 4°C before being separated using a magnet. Cells purified in this way were depleted of more than 99% of B cells as measured by CD19 expression (PharMingen).
4.6 Splenocyte culture for proliferation and cytokine assay
Single-cell suspensions were prepared and adjusted to 5×105/well (for non-depleted splenocytes) and to 2.5×105/well (for T and B cell-depleted splenocytes) in 96-well flat-bottom plates (Falcon, Cedex, France) in 200 μl of RPMI-1640 supplemented with 5% FCS (Gibco), 2 mM L-Glutamine (Gibco), 100 mg/ml streptomycin, 100 U/ml penicillin (Gibco). SEA(10 μg/ml), SWA (50 μg/ml) and Con A (2 μg/ml) (Sigma) were added in vitro and after 72 h in culture, wells were pulsed with [3H]thymidine (37 kBq/well) for the last 8 h of culture, and harvested onto glass-fiber filters and incorporation measured on a β-plate counter (LKB Wallac, Finland). Results are expressed as the ccpm (mean corrected counts per minute ± standard deviation) of triplicate wells. Similar cultures were set up in round-bottom 96-well plates and harvested at 48 h for cytokine measurement.
4.7 Bone marrow-derived DC preparation method
Seven-week-old male NOD mice were used as a source of DC. Bone marrow was eluted from the femurs and tibiae, dispersed into a single-cell suspension and washed twice in RPMI medium. DC precursors were grown in 6-well plates at 1×106/ml using 4 ml/well in conditioned medium (5% IMDM) containing 10 ng/ml of GM-CSF and 10 ng/ml IL-4 (Peprotech, UK). On day 3, the medium and nonadherent cells were removed and discarded, and fresh medium added. On day 6, 2 ml of medium was removed from each well, the cells centrifuged (300×g) and enough fresh medium was used to resuspend the pellet to exactly replace the volume removed and added back to the wells. On day 10, DC were harvested, some were stained to examine surface markers by flow cytometry and the remainder were set up in culture.
4.8 DC culture
DC were washed and adjusted to 5×104 cells/well in a flat-bottom, 96-well plate in 200-μl final volume. DC were set up in triplicates and incubated with 5 μg/ml LPS (S. enteritidis, Sigma, GB), PBS, SEA (50 μg/ml) or SWA (50μg/ml). The supernatants were collected at 24, 48 or 72 h and analyzed for cytokines by sandwich ELISA.
4.9 Detection of cytokines in culture supernatants
Concentrations of IFN-γ, IL-5, IL-10 (PharMingen), IL-12 (Biosource, Belgium) and IL-13 (R&D Systems, UK) were quantified in supernatants, using a sandwich ELISA. Protocols were followed according to the manufacturer's instructions.
4.10 Detection of IL-4 by ELISPOT assay
IL-4 secretion was quantified by ELISPOT. Briefly, Multiscreen ELISPOT plates (Millipore, MA, USA) were coated overnight with anti-IL-4 capture antibody (PharMingen). The plates were washed with sterile PBS and then blocked with milk powder (Marvel) in sterile PBS (2%) for 1 h at room temperature. After blocking, the plates were washed and rinsed with culture medium. Splenocytes were transferred onto the ELISPOT plates after 24 h pre-stimulation (the same conditions as those described above for proliferation and cytokine assays). After a second incubation period of 24 h the cells were discarded, the plates washed, and anti-IL-4-biotin (PharMingen) detection antibody was added for 2 h at room temperature. The plates were washed and streptavidin-peroxidase was added to the wells. Following 1 h of incubation the plates were washed a final time and AEC (3-amino-9ethyl-carbazole, Sigma) substrate added to the wells. After adequate color development, the plates were rinsed in tap water and the cytokine-producing cells were counted by direct observation using a dissecting microscope.
4.11 Flow cytometric analysis of α-galactosylceramide CD1d tetramer-positive cells
To characterize the α-galactosylceramide-reactive-Vα14i NKT cells, single-cell suspensions were made from liver, spleen and lymph node and re-suspended in staining buffer (PBScontaining 2% BSA and 0.05% NaN3). Cells were incubated for 30 min at 4oC with Fc receptor blocking antibody (clone 2.4G2) from tissue culture supernatant and neutravidin (Molecular Probes, Eugene, OR) to block free biotin. Cells were washed and stained with TCRβ-FITC (PharMingen) and α-galactosylceramide loaded into CD1d-PE tetramers 36 in staining buffer for 30 min at 4oC. CD1d-PE tetramers left unloaded without α-galactosylceramide were used as a negative staining control. Cells were then washed and analyzed using a FACScan flow cytometer (Becton Dickinson).
4.12 Statistical analysis
Differences in cytokine production and proliferation were analyzed using a two-sample t-test and levels of α-Gal-CD1d tetramer staining and diabetes incidence were examined using a Wilcoxon two-sample test. Data were considered significant at p<0.05.
We would like to thank Dr. Nicole Parish for critically reading this manuscript and Barry Potter for his assistance with histology. This work was funded by the Wellcome Trust and Diabetes UK.