FITC anti-mouse CD3; PE anti-human/mouse Integrin β7 (Biolegend, San Diego, CA, USA); APC anti-mouse CD45R/B220; anti-Mouse/Rat Foxp3 PE; anti-Mouse CD4 FITC; anti-Mouse CD25 APC; rat IgG2a Isotype Control PE (eBioscience, Carlsbad, CA, USA); Alexa 594 anti-mouse; Alexa 488 anti-rabbit; anti-glucagon; anti-insulin (Innovex Biosciences, Richmond, CA, USA); recombinant mouse MAdCAM-1 Fc chimera (R&D BioSystems, Minneapolis, MN, USA); murine TNF-α (Peprotech, Rocky Hill, NJ, USA); mouse IL-1β, IFNγ, and IL-17 ELISA kits (R&D Biosystems). AS101 and SAS were supplied by M. Albeck, Bar-Ilan University, Ramat-Gan, Israel, and were synthesized at Bar-Ilan University.

Three to Four week old NOD/ShiLtJ (strain 001976) female mice were purchased from the Jackson Laboratory (Bar Harbor, ME), bred, and housed in individual metabolic cages with unlimited access to food and water. Experiments conformed to approved institutional protocols and were approved by the Institutional Animal Care and Use Committee.

SVEC-4 murine endothelial cells were obtained from the American Type Culture Collection. Cells were cultured in DMEM containing 10% fetal calf serum at 37°C with 5% CO₂ and 95% air.

CD4⁺CD25⁺ Treg were purified from 3 months old NOD mice using a CD4⁺CD25⁺ Treg purification kit (Miltenyi Biotec) according to the manufacturer's instructions, with slight modifications. The CS columns were used for depleting non-T cells, and LS columns were employed to positively select CD4⁺CD25⁺ cells. Non-CD4 cells were magnetically labeled using a cocktail of biotin-conjugated antibodies recognizing CD8, CD11b, CD45R, CD49b, Ter-119, followed by incubation with anti-biotin micro beads. The labeled cells were then depleted over a column. To attain maximum purity, the CD25 positive selection was performed twice. CD4⁺CD25⁻ T cells were isolated from the CD25-depleted fraction using CD4 microbeads and LS columns.

Cells (2.5 × 10⁵) were incubated for 1 h in the presence of mobilized MadCAM-1 (100 μg/ml) medium containing MnCl2 (0.25 mM) and various concentrations of SAS or AS101. The cells were washed twice with PBS, and equal numbers of cells were loaded onto 8 mm polycarbonate membrane inserts. The bottom chambers were filled with RPMI-1640 (800 μl) with FBS (20%), containing MadCAM-1 (100 μg/ml) as a chemoattractant. Cell migration was quantified after 24 h by Trypan Blue exclusion test using a hemacytometer.

For evaluation of integrin expression on lymphocyte suspensions, single cells from spleens, peripheral LNs, and PLNs from 8 week old NOD mice were stained with FITC-conjugated anti CD3, PE-conjugated anti β7, and APC-conjugated anti B220. Cells were then analyzed on a BD FACSCalibur flow cytometer. The proportion of B and T cells that express the integrin was evaluated on 1 × 10⁴ B220 B or CD3 T cells in the lymphocyte forward scatter/side scatter gate. For evaluation of the proportion of Tregs, lymphocytes from PLNs from 20 w old treated mice were isolated and stained for membranal CD4 (FITC) and CD25 (APC) and nuclear Foxp3 (PE). Cells positive for both CD4 and CD25 were gated, and the expression of Foxp3 on this subpopulation was evaluated. The percentage of CD4⁺CD25⁺Foxp3⁺ cells from total lymphocytes was evaluated.

Isolated pancreatic extracts from treated mice were prepared in Tris/acetate buffer (pH 7.5) at 30°C. The extracts were centrifuged at 12,000 × g for 10 min, and the supernatant was collected. A volume of supernatant equivalent to 100 μg of protein was assayed for caspase-1 or caspase-3 activity using colorimetric caspase-1 and−3 assay kits (R&D Biosystems).

IL-1β, IFNγ, and IL-17 ELISA kits were used for the quantitative measurement of these cytokines in pancreas extracts of treated mice.

For testing attachment to MadCAM-1, 96-well plates were coated with 80 μL of MAdCAM1 (100 μg/ml), or BSA. Cells were incubated in the wells for 2 h in the presence or absence of AS101 or SAS followed by extensive (3x) washing. The proportion of cells that remained attached to the wells was determined by the colorimetric XTT (2,3-bis[2-methoxy-4-nitro-S-sulfophenynl]H-tetrazolium-5-carboxanilide) assay at 450 nm.

For testing attachment to endothelial cells, SVEC-4 endothelial cells were cultured on 6 well plates until 80% confluency. Next, 10 μl of rTNFα (20 ng/ml) was added to the wells to induce MadCAM-1 expression, and then washed. Lymphocytes at 1 × 10⁶/well were then added in the presence or absence of AS101 or SAS and incubated o.n. Wells were washed with PBS. The remaining attached cells were stained with anti CD3 and anti B220. Endothelial cells are not stained by either antibody (Supplementary Figure 1). The proportion of T or B cells in the total cell population was recorded.

Resected pancreas heads were fixed in 4% formalin and then paraffinized. To prepare histological sections, 5 μM sections were cut from each paraffin block, and stained with hematoxylin-eosin for detection of insulitis.

The level of insulitis was determined according to the following scale: Grade 0 - no insulitis (no infiltration); Grade 1- pre-insulitis (<25% of the islet area infiltrated); Grade 2- mild insulitis (<50% of the islet area infiltrated); Grade 3 – severe insulitis (>50% of the islet area infiltrated).

Fixed paraffin embedded pancreas head sections were deparaffinized and rehydrated. After antigen retrieval, slides were incubated with anti-mouse insulin Ab and anti-glucagon Ab o.n. at 4°C, followed by secondary Alexa 594-conjugated and Alexa 488-conjugated secondary antibodies for 1 h at RT. This was followed by Hoechst staining. Ten slides/mouse were visualized by fluorescence microscopy.

After 12 h fast, mice were administered an i.p bolus of 2 g/kg glucose Blood glucose concentrations were measured by tail bleed at the indicated time points before and after glucose administration. All glucose measurements were performed using a Free Style Freedom (Alameda, CA, USA) glucometer.

NOD mice (5 w old) were treated with PBS or with 1 mg/kg AS101 every other day. Mice were mated, and treatment continued during gestation and during lactation, until the offspring reached an age of 3 weeks. At that time treatment stopped. Some of the male offspring of treated mothers were irradiated at 6w of age with 550 rads and were injected 24 h later with 20 × 10⁶ splenocytes from diabetic female NOD mice.

Data are presented as means ± SE. For comparisons between groups in the in vitro and in some in vivo studies, we used the one- or two-way ANOVA. For in vivo studies (weight, blood glucose, glucose tolerance test, insulitis), ANOVA multiple comparisons with repeated measures with Bonferroni corrections, were applied. For analysis of incidence of diabetes and survival, the Kaplan Meier survival analysis tests were applied. The software used for all statistical analysis was IBM SPSS Statistics 21. p < 0.05 was considered statistically significant.

The α4β7 integrin, MAdCAM-1 pathway is essential for infiltration of lymphocytes into pancreatic islets, and for the initiation of T1D in NOD mice (21).Therefore, regulation of this pathway has the potential to ameliorate the course of the disease. For this purpose, we first analyzed β7 integrin expression on T and B cells from various organs of NOD mice. The results shown in Figure 1 demonstrate that most spleen (Figure 1A) and peripheral lymph node (Figure 1B) T and B cells express the α4β7 integrin, while about 50% of pancreatic T cells (Figure 1C), but most of B cells in this organ still express this integrin. We then analyzed the ability of AS101 and SAS to inhibit the activity of α4β7 on T and B cells from various organs of NOD mice. This was performed by testing the level of attached cells to the α4β7 specific ligand, MADCAM-1. Figure 2 shows a dose dependent and significant inhibition of the α4β7 integrin activity by both compounds. This was evident in both total spleen cells (Figure 2A), peripheral lymph node cells (Figure 2B), and pancreatic lymph node cells (Figure 2C). Moreover, SAS significantly inhibited the activity of the integrin in sorted T (Figure 3A), and B (Figure 3B) splenocytes (Figure 3).

To further substantiate the inhibitory effect of SAS on integrins, we examined the extent of lymphocyte adhesion, to high endothelial venular ('HEV') cells. MAdCAM-1 is expressed on approximately 95% of HEVs on the lymph nodes of young NOD mice (6). The interaction between MAdCAM-1 on HEVs in PLN and α4β7 on lymphocytes enables their trafficking to PLN. For these experiments, we used the SVEC-4 endothelial cell line which represents HEV cells and highly expresses MADCAM-1 after exposure to TNFα (22). Figure 4 shows that treatment with SAS significantly inhibits attachment of both T and B cells from NOD mice to endothelial cells, and that the effect is dose–dependent. Collectively, these results suggest that by inhibiting the interaction between α4β7 on T or B lymphocytes from NOD mice, and MADCAM-1 on endothelial cells, AS101 and SAS might affect attachment and penetration of autoreactive T and B cells through pancreatic high endothelial venules, potentially affecting the course of T1D. For this purpose, we used the NOD mouse model of T1D using AS101 and SAS in various concentrations and different timing of administration.

Figure 5A shows that extended treatment of NOD mice with an optimal dose of AS101 starting at 5–6 weeks (before the development of insulitis), resulted in the delay and prevention of diabetes incidence in a dose-dependent fashion, showing a narrow effective dose range (Figure 5). The optimal concentration of AS101 (0.5 mg/kg/injection) prevented disease in 90% of mice. Moreover, sub or supra-optimal concentrations of AS101 marginally delayed the onset of the disease, though these effects did not reach statistical significance This effect was accompanied by increased body weight (Figure 5B), and increased survival (Figure 5C) in mice receiving the optimal dose of 0.5 mg/kg. When treatment with AS101 at 0.5 mg/kg/mouse was stopped at 40 weeks of age, no change occurred with respect to percent incidence of diabetes, percent survival, or body weight until at least 52 weeks, suggesting a lasting and significant remission in the course of the disease. The average weight of mice at 52 weeks was 30 ± 77 gr, and 100% of mice surviving at 40 weeks, were also surviving with no diabetes at 52 weeks. None of the mice in the PBS control group survived to 52 weeks of age (data not shown). Importantly, even when AS101 was administered at an advanced stage of insulitis (12 weeks), the incidence of diabetes in treated mice until 40 weeks of age was only 10%, whereas that of control mice reached 90% at that time point (Figure 6A). This timing of AS101 administration resulted in preserved body weight (Figure 6B) and increased survival (Figure 6C). Importantly, even when treatment with AS101 at 0.5 mg/kg/mouse was started at 12 weeks and stopped at 40 weeks, all mice were remained alive, diabetes-free, and maintained their body weight at 52 weeks (average body weight of 26.88 ± 1.01 grams at 52 weeks). None of the mice in the control group were still alive at that time (data not shown). Moreover, not only could AS101 prevent development of diabetes when administered to pre-diabetic mice with different degrees of insulitis, but AS101 increased survival and body weight as well when administered to 14 w old mice that already developed symptoms of diabetes (all untreated NOD mice were diabetic at this time point, not shown) (Figures 6D,E).

The encouraging results obtained regarding the beneficial effects of AS101 treatment on the course of T1D led us to investigate the potential effect of the new generation tellurium compound, SAS, shown here to inhibit the activity of α4β7 (Figures 2–4), on the intensity of disease. Treatment with SAS was compared to that with AS101 (the optimal dose of the specific batch of AS101 at that time was 1 mg/kg/injection).

Figure 7 shows that extended treatment with the tellurium compounds, AS101 and SAS of NOD mice starting at 5–6 weeks (Figures 7A–D) or at 3–4 weeks (Figures 7E–H) [initiation of the autoimmune process (23)], decreased levels of blood glucose (Figures 7A,E), decreased incidence of diabetes (Figures 7B,F), decreased weight loss (Figures 7C,G) and eventually increased survival (Figures 7D,H) (Figure 7). The results show that treatment with the tellurium compounds delay the onset of the disease, and even result in its prevention with increased efficiency when treatment is started earlier.

Pretreatment with AS101 or SAS preserved optimal glucose tolerance (Figure 8). Treatment with AS101 and SAS resulted in normal levels of glucose in intraperitoneal glucose tolerance test (IPGTT), either when the test was performed at 12 w of age (Figure 8A), or at 20 w (Figure 8B), suggesting that preservation of pancreatic islets may have occurred in AS101 or SAS-treated mice early after treatment The fact that in treated mice, 1 h after glucose loading, glucose levels amounted to only ~100 mg/dL, considered to be normal levels, suggests that treatment with tellurium compounds prevents islet destruction, enabling normal function of β cells. Indeed, the results show that the level of insulitis was dramatically and significantly decreased in treated mice (Figures 8C–H) both with respect to the percentage of infiltrated islets (Figure 8G) and the percentage of cells with severe insulitis score (Figure 8H). Moreover, the function of the islets was preserved in treated mice, producing normal insulin in pancreatic islets (Figure 9A), as compared to mice in a pre-diabetic state. More importantly, in addition to the normal morphology and size of islets of treated mice (Figure 9B), caspase-3 activity, the classical hallmark of apoptosis, was significantly decreased compared to control mice, and not significantly different from that of pre-diabetic mice (Figure 9C).

In order to gain insight into the mechanism of the beneficial effects of both tellurium compounds in T1D, we first examined whether integrin α4β7 activity in PAN lymph nodes was inhibited following treatment. As can be seen in Figure 10A, the integrin activity in PLN lymphocytes from NOD mice treated with the compounds from 3 to 20 w was significantly inhibited. This was manifested by diminished attachment of the cells to the α4β7 specific ligand, MADCAM-1. This inactivation might prevent the infiltration of autoreactive cells to the pancreas.

We next examined protein expression of IL-1β, a key proinflammatory cytokine, which drives several autoimmune diseases. IL-1 is a pleiotropic cytokine, which attenuates CD4⁺CD25⁺FoxP3⁺ regulatory T cell function, and allows the CD4⁺CD25⁻ autoreactive effector subset to escape from suppression (24). IL-1β has been previously shown to be inhibited by AS101 in various pre-clinical studies (14, 16, 25). Figure 10B shows that both AS101 and SAS significantly inhibited pancreatic IL-1 protein levels at 20 weeks. Furthermore, caspase-1 activity, previously shown to be inhibited by AS101 (16, 25) resulting in decreased IL-1β, was also inhibited in this study in PAN LN by AS101 and SAS-treated mice (Figure 10C). Importantly, the pro-inflammatory cytokine IL-17, which is implicated as essential in the development of T1D (26), was also inhibited in treated mice (Figure 10D). Furthermore, the Th1 cytokine IFNγ, playing a dominant role in the pathogenesis of autoimmune diabetes (27), was also significantly decreased in SAS and AS101-treated mice (Figure 10E). These results prompted us to evaluate the proportion of CD4⁺CD25⁺FoxP3⁺ regulatory T cells, known to be affected by these two inflammatory cytokines (IL-1β and IL-17). Figures 11A–D shows a significant increase in the proportion of Tregs among PAN LN cells.

The increase in Tregs following treatment by SAS or AS101 prompted us to evaluate if the tellurium compounds affect the α4β7 integrin activity in Tregs vs. T effector cells and to explore the effects and if this property affects the migration of each subset of cells. Figures 12A,B shows that SAS and AS101 inhibits the activity of the α4β7 integrin in both isolated T effector cells (Figure 12A) and regulatory cells (Figure 12B); nevertheless, the extent of inhibition is higher in T effector cells. Moreover, the α4β7 integrin-dependent migration of both T cell subsets was also inhibited by both compounds, however that of T effector cells was more intense. Furthermore, The extent of control T regulatory cells migration was significantly lower than that of T effector cells (Figure 12C). Interestingly, the expression of a4b7 in T regulstoty cells was lower than that of T effector cells (Figure 12D). This might explain the lower a4b7-dependent of Tregs and their decreased sensitivity to tellurium compound inhibition.

Finally, we wished to evaluate if offspring of AS101-treated NOD mothers acquire some resistance to T1D development. Female NOD mice were treated with AS101 starting 2 weeks prior to mating, during pregnancy and throughout breast feeding. When the pups were analyzed, they exhibited decreased incidence of diabetes (Supplementary Figure 1A) and increased survival (Supplementary Figure 1B). Moreover, offspring of treated mothers were relatively more resistant to adoptive transfer of autoreactive cells as expressed by decreased incidence of diabetes (Supplementary Figure 1C) and increased survival (Supplementary Figure 1D). These results suggest that AS101 can significantly protect the fetus from future diabetes development when the mother is treated during pregnancy. Furthermore, the fetus might acquire both diminished autoreactive cells and increased regulatory mechanisms.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.