βig-h3/TGF-βi is a secreted protein capable of binding to both extracellular matrix and cells. Human genetic studies recently revealed that in the tgfbi gene encoding for βig-h3, three single nucleotide polymorphisms were significantly associated with type 1 diabetes (T1D) risk. Pancreatic islets express βig-h3 in physiological conditions, but this expression is reduced in β-cell insult in T1D. Since the integrity of islets is destroyed by autoimmune T lymphocytes, we thought to investigate the impact of βig-h3 on T-cell activation. We show here that βig-h3 inhibits T-cell activation markers as well as cytotoxic molecule production as granzyme B and IFN-γ. Furthermore, βig-h3 inhibits early T-cell receptor signaling by repressing the activation of the early kinase protein Lck. Moreover, βig-h3–treated T cells are unable to induce T1D upon transfer in Rag2 knockout mice. Our study demonstrates for the first time that T-cell activation is modulated by βig-h3, an islet extracellular protein, in order to efficiently avoid autoimmune response.

Type 1 diabetes (T1D) is a polygenic autoimmune disease characterized by smoldering inflammatory response directed against the insulin-producing β-islet cells of the pancreas. Both CD4+ and CD8+ T cells are involved in the destruction of β-cells (shed by a prior insult). CD4+ T cells are insulin reactive, and CD8+ T cells play a major role as β-cell killers (1,2). Although both in vitro and in vivo evidence point to a role for regulatory T cells in T-cell–mediated regulation controlling diabetes in mouse models (3), the influence of the inflammatory environment in which they function remains poorly understood. The cross talk between the inflammatory milieu and the immune response during insulitis is likely to be communicated in large part through the extracellular matrix (ECM). βig-h3 (also known as TGF-βi) is a secreted protein found in the ECM and it has an N-terminal secretory signal (aa 1–23), four FAS1 homologous internal domains, and a cell attachment site (RGD) at its C terminus (4). βig-h3 binds to the ECM through interaction of the YH motif in its FAS1 domains with collagens I, II, IV, and VI (5,6). Its FAS1 domain interacts with its α3β1, αVβ3, and αVβ5 integrins on the cell surface. Previous studies demonstrated that recombinant βig-h3 could preserve the integrity and enhance the function of cultured pancreatic islet cells (7). Conversely, βig-h3 knockout (KO) islet function was compromised in vivo after transplantation in recipients with diabetes (8). These studies point out that βig-h3 might play an active role in the protection of β-cells against T-cell cytotoxic insult in T1D. The effect of βig-h3 on T-cell function and activation relative to its roles in islet β-cell–ECM destruction in T1D in mice has never been investigated and it is the focus of our project.

Mice

Female C57Bl/6 NOD mice were bred and housed in specific pathogen-free conditions. T1D was chemically induced in-C57Bl/6 using a standard low-dose (35 mg/kg) streptozotocin (STZ) protocol (9). Alternatively, for antibody depletion experiments, five consecutive intraperitoneal injections of 30 mg/kg STZ were used. Fourteen days after STZ treatment, 300 μg/kg of anti–TGF-βi neutralizing or control monoclonal antibody (Bio X Cell, West Lebanon, NH) was subsequently intraperitoneally injected. Blood glucose was monitored from day 14 until sacrifice. Animals with glycemia superior to 200 mg/dL were considered diabetic. The local animal experimentation ethics committee approved the experimental protocols.

Patients

Samples (3 μm) were obtained from the National Institute for Diabetes, Bucharest, Romania, from cadaveric donor pancreata from patients with and without T1D and T2D. All experimental procedures were approved by the Romanian National Ethics Committee.

Cell Culture

Cells extracted from draining lymph nodes were cultured in 96 U-shaped well plates (300,000 cells/well). TGF-βi treatment (5 μg/mL; R&D Systems) was done for 24 h. Cells were activated with different doses of anti-mouse CD3ε antibody (eBioscience) for the indicated times. Anti-mouse CD3ε antibody (eBioscience) was used for cell activation as indicated. Carboxyfluorescein succinimidyl ester (CFSE) staining was used for T-cell proliferation assays according to the manufacturer’s protocol (Invitrogen). Islet isolation was obtained after collagenase perfusion as reported previously (10) and handpicked before experimentation.

Flow Cytometry Analysis

Cells isolated from draining lymph node or perfused pancreata were stained using the following antibodies: CD8-V450, CD4-V500, CD44 A700 (BD Biosciences), and CD69 FITC (ImmunoTools). Intracellular staining was performed after phorbol myristic acid and ionomycin (1 μg/mL) activation for 4 h under normal culture conditions in the presence of GolgiPlug (BD Pharmingen). After activation, cells were permeabilized using Cytofix/Cytoperm (BD Pharmingen) prior to staining with anti-interferon (IFN) (clone XMG1.2; BD Pharmingen) and anti–granzyme B (clone GB12; Invitrogen). Flow cytometry was performed with a BD LSRFortessa flow cytometer (BD Biosciences) and analyzed with BD FACSDiva software v5.0.1 (BD Biosciences) and FlowJo (Tree Star, Inc.).

Western Blot

Proteins were extracted with a RIPA buffer; 30 µg of protein was separated using SDS-PAGE and transferred to a nitrocellulose membrane (Amersham). Primary antibodies (anti-Lck [Cell Signaling]; anti-pSrc pY416, which recognizes Lck on pY394 [Cell Signaling]; anti-pLck pY505 [Cell Signaling]; mouse anti-actin [Sigma-Aldrich]) were incubated overnight, followed by horseradish peroxidase–conjugated secondary antibodies (Amersham) prior to detection using an enhanced chemiluminescence solution (Pierce). Image Lab (Bio-Rad) was used for subsequent quantification.

Immunofluorescence and Confocal Microscopy

Triple immunofluorescent staining was performed on a 3-μm-thick paraffin pancreas section. Sections were rehydrated and subjected to antigen retrieval in an unmasking solution (Vector H-3300). Tissue sections were incubated overnight with anti–TGF-βi (rabbit; Sigma-Aldrich), anti-insulin (guinea pig; Dako), and anti-glucagon (mouse; Sigma-Aldrich) primary antibodies. Secondary fluorescence-conjugated antibody (Jackson ImmunoResearch) was incubated 1 h prior to mounting using a DAPI-VECTASHIELD Mounting Medium (Vector Laboratories). Cell immunofluorescent staining was performed on cytospined cells fixed in 4% paraformaldehyde prior to permeabilization. Primary antibodies, anti–T-cell receptor (TCR)-β (Becton Dickinson), and phospho-Lck pY505 (Cell Signaling) were incubated overnight and subsequently detected using specific anti–Fab’2-Alexa 488, 647, and 555 (Molecular Probes) secondary antibodies prior to mounting. All staining was acquired on an LSM 780 Zeiss confocal microscope and subsequently analyzed using Zen software (Zeiss). The method used for colocalization quantifications was previously described (11).

Statistical Analysis

P values were calculated with Student t test (GraphPad Prism), as specified in figure legends (*P < 0.05; **P < 0.01; ***P < 0.001).

βig-h3 Inhibits T-Cell Activation

To study a direct role of βig-h3 on T cells, we isolated T cells from wild-type (WT) mouse lymph nodes and treated them using a human recombinant βig-h3 for 24 h prior to their activation with different doses of anti-CD3 antibody (Fig. 1). Since TCR avidity modulates T-cell activation (12) and autoreactive T cells respond to low-affinity and avidity peptides, we compared the ability of CD8+ T cells to respond to high (1,000 ng/mL), intermediate (100 ng/mL), and low concentrations (10 ng/mL) of anti-CD3 after being treated (or not) with recombinant βig-h3. We show that at high and intermediate concentrations of anti-CD3 βig-h3 did not modulate the activation of CD8+ T cells after TCR engagement, as shown by the activation markers CD69 and CD44. In contrast, βig-h3 was able to inhibit subsequent low-dose CD8+ (as well as CD4+) T-cell activation, as shown by the significant reduction of CD69 and CD44 expression (Fig. 1A and B and Supplementary Fig. 1A). Furthermore, both CD8+ and CD4+ T cells treated with βig-h3 for 24 h displayed reduced production of intracellular IFN-γ and granzyme B (Fig. 1C and D and Supplementary Fig. 1A). In order to explore the role of βig-h3 in the pathological condition of T1D, we extracted T cells from 6-month-old diabetic NOD mice and treated them with recombinant βig-h3 for 24 h before assessing their production of IFN-γ. We found βig-h3–pretreated pancreatic T cells from diabetic NOD mice to produce significantly lower intracellular IFN-γ (Supplementary Fig. 1B). Therefore, βig-h3 was able to reduce the production of inflammatory cytokines by diabetogenic T cells ex vivo. Thus, together these results point out an important immunomodulatory role for βig-h3 protein in the activation and production of inflammatory cytokines by T cells.

Figure 1

βig-h3 inhibits T-cell activation. A and B: Draining lymph node cells from WT C57Bl/6 (3 × 105 cells/condition) were treated or not with βig-h3 (5 μg/mL) for 24 h. Anti-CD3 Ab was further added at the indicated concentration for another 24 h. Representative histograms are illustrating the percentage of CD8+ T cells expressing CD69 (A) and CD44 (B) at their surface and IFN-γ (C) and granzyme B (D) in their cytoplasm. Graphs are quantifying the percentage of CD8+ T cells expressing CD69 (A) and CD44 (B) and the percentage of CD8+ T cells expressing IFN-γ (C) and granzyme B (GrzB) (D) in three individual mice. The results are representative of three independent experiments with three mice per group. *P < 0.05; ***P < 0.001.

Figure 1

βig-h3 inhibits T-cell activation. A and B: Draining lymph node cells from WT C57Bl/6 (3 × 105 cells/condition) were treated or not with βig-h3 (5 μg/mL) for 24 h. Anti-CD3 Ab was further added at the indicated concentration for another 24 h. Representative histograms are illustrating the percentage of CD8+ T cells expressing CD69 (A) and CD44 (B) at their surface and IFN-γ (C) and granzyme B (D) in their cytoplasm. Graphs are quantifying the percentage of CD8+ T cells expressing CD69 (A) and CD44 (B) and the percentage of CD8+ T cells expressing IFN-γ (C) and granzyme B (GrzB) (D) in three individual mice. The results are representative of three independent experiments with three mice per group. *P < 0.05; ***P < 0.001.

Close modal

βig-h3 Inhibits Early TCR Signaling

βig-h3 has been reported to signal from the extracellular environment through αvβ3 integrins (13). Interestingly, integrin signaling has been reported to share early TCR signaling molecules (14,15). Therefore, we investigated the effect of βig-h3 signaling on the activation of early actors of TCR signaling. Initiation of the TCR signaling pathway relies on the induction of tyrosine protein phosphorylation of the kinase Lck (16,17). The regulation of the enzymatic activity of this Src family tyrosine kinase is tightly controlled by conformational changes mainly relying on phosphorylation and dephosphorylation on two regulatory tyrosine residues, Y505 and Y394 (18,19). In resting T cells, as previously reported (20), a larger percentage of Lck was phosphorylated on Y394 and thus was present under its activated conformation (Fig. 2A). However, in clear contrast, βig-h3–treated CD8+ T cells were mainly phosphorylated on Y505, implying that βig-h3 signaling influenced the Lck conformation equilibrium toward its inactive state (Fig. 2A and B). After TCR engagement, with a low dose of anti-CD3, the ratio of Y505 to Y394 was higher in βig-h3–treated T cells than in nontreated cells (Fig. 2A and B) at all time points (2 and 10 min). These data suggest that βig-h3 blocks the TCR signaling pathway by influencing the equilibrium of Lck conformation toward an inactive state, characterized by Y505 phosphorylation. Next, we searched for the localization of Lck phosphorylated on inhibitory Y505 in resting and stimulated T cells by confocal microscopy. Lck phosphorylated on Y505 was colocalized with the TCR in high proportion (∼60%) (Supplementary Fig. 2) in resting untreated T cells (Fig. 2C). In contrast, after TCR engagement with anti-CD3, pY505 colocalization with the TCR was significantly diminished (Fig. 2C and Supplementary Fig. 2). In βig-h3–treated cells, the colocalization of Lck Y505 with TCR was similar in resting cells and after TCR engagement (∼58%) (Supplementary Fig. 2 and Fig. 2C), suggesting that βig-h3 signaling interferes with T-cell activation.

Figure 2

βig-h3 signaling affects the Lck phosphorylation equilibrium. T cells were treated or not with βig-h3 (5 μg/mL) for 24 h and further stimulated with 10 ng/mL of anti-CD3 for the indicated times. A: Western blot analysis of expression levels of pLck Y505, pLck Y394, total Lck, and β-actin is indicated. B: Quantification of the ratio of pLck Y505 to pLck Y394 to total Lck mean intensity in different conditions. C: Confocal immunofluorescence of resting (0) or stimulated for 2 min (α-CD3 2 min) T cells stained with anti–TCR-β, pLck Y505, and DAPI. Results are representative of three independent experiments.

Figure 2

βig-h3 signaling affects the Lck phosphorylation equilibrium. T cells were treated or not with βig-h3 (5 μg/mL) for 24 h and further stimulated with 10 ng/mL of anti-CD3 for the indicated times. A: Western blot analysis of expression levels of pLck Y505, pLck Y394, total Lck, and β-actin is indicated. B: Quantification of the ratio of pLck Y505 to pLck Y394 to total Lck mean intensity in different conditions. C: Confocal immunofluorescence of resting (0) or stimulated for 2 min (α-CD3 2 min) T cells stained with anti–TCR-β, pLck Y505, and DAPI. Results are representative of three independent experiments.

Close modal

βig-h3 Expression Is Reduced in Pancreatic Islets in T1D

To investigate the modulation of βig-h3 expression in pathological conditions, we used chemical-induced (STZ) and genetic (NOD) mouse models of T1D. In the genetic model, ∼80% of NOD female mice develop diabetes at 6 months (21). Diabetic NOD females showed similar immune infiltration as NOD normoglycemic nondiabetic age-matched females (Fig. 3A), suggesting that a local immunosuppression mechanism might be involved. Immunofluorescence analysis showed that βig-h3 was expressed in pancreatic islets in normal mice as well as in normoglycemic nondiabetic NOD mice (Fig. 3A and B). We found βig-h3 expressed in the cytoplasm of both α- and β-cells as well as in the extracellular islet compartment (Fig. 3A and B, arrows). In contrast, in T1D islets, βig-h3 expression was reduced at both the mRNA (Supplementary Fig. 3) and protein level (Fig. 3A and B). Only α-cells retained expression of βig-h3 in diabetic mice (Fig. 3A and B). We also analyzed human pancreas sections from patients with T1D (n = 2) and T2D (n = 2) and patients without diabetes (n = 2). In human pancreas, βig-h3 was expressed only by β-cells and in the ECM in patients with T2D and patients without diabetes. Patients with T1D show a much reduced βig-h3 expression (Fig. 3C). Together, these data suggest that pancreatic islet expression of βig-h3 might play a role in the susceptibility to develop T1D.

Figure 3

Altered expression of βig-h3 in T1D insult in pancreatic islets. A and B: Immunofluorescent costaining for insulin, glucagon, and βig-h3 within pancreatic islets in control mice (WT CIT), STZ-treated diabetic mice (T1D STZ), 6-month-old normoglycemic nondiabetic NOD mice (NOD), and diabetic age-matched mice (T1D NOD). The results are representative of three independent experiments with three mice per group. Insets are magnified views of merge. Arrows indicate extracellular staining for βig-h3. C: Representative immunofluorescent stainings of a pancreas section from two cadaveric donors with T1D, two with T2D, and two without diabetes (ctrl).

Figure 3

Altered expression of βig-h3 in T1D insult in pancreatic islets. A and B: Immunofluorescent costaining for insulin, glucagon, and βig-h3 within pancreatic islets in control mice (WT CIT), STZ-treated diabetic mice (T1D STZ), 6-month-old normoglycemic nondiabetic NOD mice (NOD), and diabetic age-matched mice (T1D NOD). The results are representative of three independent experiments with three mice per group. Insets are magnified views of merge. Arrows indicate extracellular staining for βig-h3. C: Representative immunofluorescent stainings of a pancreas section from two cadaveric donors with T1D, two with T2D, and two without diabetes (ctrl).

Close modal

βig-h3 Affects Activation of Diabetogenic T Cells In Vivo

C57Bl/6 mice treated with low-dose STZ (35 mg/kg) for 5 consecutive days develop hyperglycemia and T1D starting from day 14 (Supplementary Fig. 4A). To assess the role of βig-h3 in diabetes development in vivo, we titrated STZ to a suboptimal dose (30 mg/kg for 5 consecutive days) that does not induce diabetes in C57Bl/6 mice from day 14. Mice treated with suboptimal STZ doses were injected on day 14 with neutralizing anti–βig-h3 and assessed for glycemia on day 16 (Fig. 4A). In contrast to control Ab–treated mice, βig-h3–depleted animals develop diabetes from suboptimal doses of STZ, as measured by a significant increase of glycemia (Fig. 4B). Next, we isolated islets from WT C57Bl/6 mice and cultured them in the presence of a neutralizing anti–βig-h3 antibody prior to their coculture with STZ-diabetic isolated T cells. As shown in Supplementary Fig. 4B, we found that 24-h treatment with anti–βig-h3 significantly decreased the amount of βig-h3 protein detected in the supernatant of cultured islets. T cells extracted from STZ-diabetic pancreas were CFSE stained and added to the islet culture for 48 h prior to the quantification of the intracellular production of granzyme B and IFN-γ by flow cytometry (Fig. 4C). Whereas the proliferation of T cells was not affected by the βig-h3 antibody treatment of islet cells (Fig. 4D), we found that the production of granzyme B and IFN-γ was significantly increased in βig-h3–depleted islet conditions (Fig. 4E and F), suggesting that physiological release of the protein by the islets was able to inhibit T-cell cytotoxic cytokine production. In order to investigate the physiological relevance of the repression of the T-cell response in vitro, we used an in vivo diabetes transfer model. T cells from the pancreas of diabetic mice (STZ-induced T1D) were treated with recombinant βig-h3 and then intravenously injected in immunodeficient Rag2 KO congenic mice. Mice were killed 2 days after T-cell injection, and islet infiltration by granzyme B+ cells was assessed by immunofluorescence (Fig. 4I and Supplementary Fig. 4C). Measurements of glycemia showed that βig-h3–pretreated T cells were unable to induce diabetes as compared with untreated T cells (Fig. 4H). Although equivalent percentages of CD45+ cells were detected by FACS in the recipient’s pancreas 10 days postinjection (Fig. 4J), T cells pretreated with βig-h3 showed lower levels of CD69 (Fig. 4K) and less production of IFN-γ and CD107a (Fig. 4L and M). Thus, these findings reveal an important role of βig-h3 in inhibiting diabetogenic T-cell activity in vitro and in vivo.

Figure 4

βig-h3 controls activation of T cells in vivo. A: Mice were treated with a suboptimal dose of STZ (30 mg/kg) for 5 consecutive days and intraperitoneally treated on day 14 with neutralizing anti–βig-h3 or control Ab (at a concentration of 300 μg/kg) and assessed for glycemia. B: Glycemia measurement on day 16, representative of two independent experiments with five mice per group. ctrl, control. C: T cells extracted from diabetic STZ mouse pancreas were CFSE stained and added to the islet culture (βig-h3 depleted of ctrl Ab treated) for 48 h, and the intracellular production of granzyme B (GrzB) (E) and IFN-γ (F) was assessed by flow cytometry. D: Quantification of the percentage of undivided (CFSE high) and divided cells (CFSE inter and CFSE low). G: Purified T cells from diabetic STZ-treated pancreas were incubated (treated) or not (control) with βig-h3 (5 μg/mL) for 24 h and intravenously transferred in Rag2 KO recipient mice (5 × 105 cells/animal). H: Graph shows glycemia measurements. I: Immunofluorescent staining for insulin, granzyme B, and DAPI within the pancreas of Rag2 KO recipient mice transferred with untreated T cells (control) or βig-h3–treated T cells (treated) at day 2 after transfer. The results are representative of two independent experiments with three mice per group. Quantification of the percentage of CD45+ cells 14 days after cell transfer in mouse pancreas (J) and CD8+ T cells expressing CD69 (K), IFN-γ (L), and CD107a (M). *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 4

βig-h3 controls activation of T cells in vivo. A: Mice were treated with a suboptimal dose of STZ (30 mg/kg) for 5 consecutive days and intraperitoneally treated on day 14 with neutralizing anti–βig-h3 or control Ab (at a concentration of 300 μg/kg) and assessed for glycemia. B: Glycemia measurement on day 16, representative of two independent experiments with five mice per group. ctrl, control. C: T cells extracted from diabetic STZ mouse pancreas were CFSE stained and added to the islet culture (βig-h3 depleted of ctrl Ab treated) for 48 h, and the intracellular production of granzyme B (GrzB) (E) and IFN-γ (F) was assessed by flow cytometry. D: Quantification of the percentage of undivided (CFSE high) and divided cells (CFSE inter and CFSE low). G: Purified T cells from diabetic STZ-treated pancreas were incubated (treated) or not (control) with βig-h3 (5 μg/mL) for 24 h and intravenously transferred in Rag2 KO recipient mice (5 × 105 cells/animal). H: Graph shows glycemia measurements. I: Immunofluorescent staining for insulin, granzyme B, and DAPI within the pancreas of Rag2 KO recipient mice transferred with untreated T cells (control) or βig-h3–treated T cells (treated) at day 2 after transfer. The results are representative of two independent experiments with three mice per group. Quantification of the percentage of CD45+ cells 14 days after cell transfer in mouse pancreas (J) and CD8+ T cells expressing CD69 (K), IFN-γ (L), and CD107a (M). *P < 0.05; **P < 0.01; ***P < 0.001.

Close modal

A major goal of T1D research is to restore β-cell function while eliminating diabetogenic T cells by immunotherapy in the hope of achieving insulin independence. To develop an effective immunotherapy, there must be a continued effort to define the molecular basis that underlies the T-cell response to pancreatic islet antigens in T1D. Therefore, identifying new targets involved in the protection against cytotoxic T-lymphocyte attack might be important for defining new therapy hints. We show here that βig-h3, a protein of the islet ECM, is able to repress diabetogenic T-cell activation by interfering with early actors of the TCR signaling pathway, i.e., Lck. This alteration has major consequences on T-cell expression of activation markers (CD69 and CD44) upon TCR engagement, as well as production of IFN-γ and granzyme B. Previous findings identified βig-h3 as a vital trophic factor promoting islet survival, function, and regeneration (7). We demonstrate that in addition to its trophic support to pancreatic islets, βig-h3 plays an active role in inhibiting the activation T-cell activation to low concentrations of antigen or autoreactive T cells. Signaling through integrins has been described to only facilitate T-cell activation by lowering an activation threshold without qualitatively changing the T-cell response (22). Our results show that βig-h3 signaling through T-cell surface integrins was able to inhibit TCR activation by phosphorylating Lck on inhibitory tyrosine Y505. Previous studies showed that Lck plays an important role in regulating autoimmunity since a specific deficiency of this kinase was described in patients with T1D (23). A key role of Lck in the CTL function of CD8+ T cells was also reported in mice (24). Nika et al. (20) showed that the ratio of Y394/Y505 was approximately 1.5 in normal cells. After βig-h3 treatment, Lck was mostly phosphorylated on inhibiting tyrosine Y505 (ratio of Y505 to Y394 above 2) (Fig. 2A and B) and colocalized with the TCR (Fig. 2C). As a consequence, engagement of the TCR in those βig-h3–pretreated cells did not lead to delocalization of Y505 away from the TCR compared with untreated cells. Therefore transfer of βig-h3–treated diabetogenic T cells into WT animals led to blunted activation and cytotoxic capacities in recipient pancreas. An interesting finding is that NOD female mice that do not develop diabetes at 6 months (∼20%) express βig-h3 in the pancreatic islets (Fig. 3B) and that the expression is lost in hyperglycemic age-matched diabetic NOD mice. Since the immune infiltration was similar, this data suggest that βig-h3 expression might contribute as a protective shield against cytotoxic T-cell attack. Our experimental findings reconsider the conventional understanding of “cytotoxic” T-cell–induced β-cell death in T1D in order to take into account β-cell “matrix” as a critical component of T1D pathogenesis. Studies of the βig-h3 impact on islet-specific T cells in from human peripheral blood T1D patients will further elucidate whether this protein is also a key target of immune tolerance in human T1D.

Acknowledgments. The authors thank ANICAN (Cancer Research Center of Lyon) animal staffs for maintenance of mouse strains. The authors thank Christophe Vanbelle for helpful assistance with confocal microscopy.

Funding. R.T. and D.R. were supported by fellowships from the French Ministry of Education and Research. This work was supported by a grant from La Ligue Contre le Cancer (P.B. and A.H.).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. M.P. performed the experiments and analyzed the data. R.T., D.G., and D.R. performed the experiments. C.Z. provided and analyzed the human samples. I.-S.K. provided the depleting βig-h3 antibody and designed the depletion experiments. P.B. wrote the manuscript. A.H. designed the experiments, analyzed the data, and wrote the manuscript. A.H. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

1.
DiLorenzo
TP
,
Lieberman
SM
,
Takaki
T
, et al
.
During the early prediabetic period in NOD mice, the pathogenic CD8(+) T-cell population comprises multiple antigenic specificities
.
Clin Immunol
2002
;
105
:
332
341
[PubMed]
2.
Tsai
S
,
Shameli
A
,
Santamaria
P
.
CD8+ T cells in type 1 diabetes
.
Adv Immunol
2008
;
100
:
79
124
[PubMed]
3.
Zhao
N
,
Hao
J
,
Ni
Y
, et al
.
Vγ4 γδ T cell-derived IL-17A negatively regulates NKT cell function in Con A-induced fulminant hepatitis
.
J Immunol
2011
;
187
:
5007
5014
[PubMed]
4.
Skonier
J
,
Bennett
K
,
Rothwell
V
, et al
.
Beta ig-h3: a transforming growth factor-beta-responsive gene encoding a secreted protein that inhibits cell attachment in vitro and suppresses the growth of CHO cells in nude mice
.
DNA Cell Biol
1994
;
13
:
571
584
[PubMed]
5.
Hashimoto
K
,
Noshiro
M
,
Ohno
S
, et al
.
Characterization of a cartilage-derived 66-kDa protein (RGD-CAP/beta ig-h3) that binds to collagen
.
Biochim Biophys Acta
1997
;
1355
:
303
314
[PubMed]
6.
Choi
Q
,
Hong
KH
,
Kim
JE
,
Kim
HK
.
Changes in plasma levels of natural anticoagulants in disseminated intravascular coagulation: high prognostic value of antithrombin and protein C in patients with underlying sepsis or severe infection
.
Ann Lab Med
2014
;
34
:
85
91
[PubMed]
7.
Han
B
,
Qi
S
,
Hu
B
,
Luo
H
,
Wu
J
.
TGF-beta i promotes islet beta-cell function and regeneration
.
J Immunol
2011
;
186
:
5833
5844
[PubMed]
8.
Han
KJ
,
Lee
SY
,
Kim
NH
, et al
.
Increased risk of diabetes development in subjects with the hypertriglyceridemic waist phenotype: a 4-year longitudinal study
.
Endocrinol Metab (Seoul)
2014
;
29
:
514
521
[PubMed]
9.
Like
AA
,
Rossini
AA
.
Streptozotocin-induced pancreatic insulitis: new model of diabetes mellitus
.
Science
1976
;
193
:
415
417
[PubMed]
10.
Wu
Y
,
Han
B
,
Luo
H
, et al
.
DcR3/TR6 effectively prevents islet primary nonfunction after transplantation
.
Diabetes
2003
;
52
:
2279
2286
[PubMed]
11.
Manders
EMM
,
Verbeek
FJ
,
Aten
JA
.
Measurement of co-localization of objects in dual-colour confocal images
.
J Microsc
1993
;
169
:
375
382
12.
Savage
PA
,
Boniface
JJ
,
Davis
MM
.
A kinetic basis for T cell receptor repertoire selection during an immune response
.
Immunity
1999
;
10
:
485
492
[PubMed]
13.
Tumbarello
DA
,
Temple
J
,
Brenton
JD
.
β3 integrin modulates transforming growth factor beta induced (TGFBI) function and paclitaxel response in ovarian cancer cells
.
Mol Cancer
2012
;
11
:
36
[PubMed]
14.
Goda
S
,
Quale
AC
,
Woods
ML
,
Felthauser
A
,
Shimizu
Y
.
Control of TCR-mediated activation of beta 1 integrins by the ZAP-70 tyrosine kinase interdomain B region and the linker for activation of T cells adapter protein
.
J Immunol
2004
;
172
:
5379
5387
[PubMed]
15.
Cimo
AM
,
Ahmed
Z
,
McIntyre
BW
,
Lewis
DE
,
Ladbury
JE
.
CD25 and CD69 induction by α4β1 outside-in signalling requires TCR early signalling complex proteins
.
Biochem J
2013
;
454
:
109
121
[PubMed]
16.
Davis
SJ
,
van der Merwe
PA
.
Lck and the nature of the T cell receptor trigger
.
Trends Immunol
2011
;
32
:
1
5
[PubMed]
17.
Rhee
I
,
Veillette
A
.
Protein tyrosine phosphatases in lymphocyte activation and autoimmunity
.
Nat Immunol
2012
;
13
:
439
447
[PubMed]
18.
Yamaguchi
H
,
Hendrickson
WA
.
Structural basis for activation of human lymphocyte kinase Lck upon tyrosine phosphorylation
.
Nature
1996
;
384
:
484
489
[PubMed]
19.
Boggon
TJ
,
Eck
MJ
.
Structure and regulation of Src family kinases
.
Oncogene
2004
;
23
:
7918
7927
[PubMed]
20.
Nika
K
,
Soldani
C
,
Salek
M
, et al
.
Constitutively active Lck kinase in T cells drives antigen receptor signal transduction
.
Immunity
2010
;
32
:
766
777
[PubMed]
21.
Makino
S
,
Kunimoto
K
,
Muraoka
Y
,
Mizushima
Y
,
Katagiri
K
,
Tochino
Y
.
Breeding of a non-obese, diabetic strain of mice
.
Jikken Dobutsu
1980
;
29
:
1
13
[PubMed]
22.
Bachmann
MF
,
McKall-Faienza
K
,
Schmits
R
, et al
.
Distinct roles for LFA-1 and CD28 during activation of naive T cells: adhesion versus costimulation
.
Immunity
1997
;
7
:
549
557
[PubMed]
23.
Nervi
S
,
Atlan-Gepner
C
,
Kahn-Perles
B
, et al
.
Specific deficiency of p56lck expression in T lymphocytes from type 1 diabetic patients
.
J Immunol
2000
;
165
:
5874
5883
[PubMed]
24.
Milstein
O
,
Hagin
D
,
Lask
A
, et al
.
CTLs respond with activation and granule secretion when serving as targets for T-cell recognition
.
Blood
2011
;
117
:
1042
1052
[PubMed]

Supplementary data