Reactivated CD4⁺Tm Cells of T1D Patients and Siblings Display an Exaggerated Effector Phenotype With Heightened Sensitivity to Activation-Induced Cell Death

Type 1 diabetes (T1D) is an autoimmune disease that results from progressive T-cell–mediated destruction of insulin-producing β-cells in the pancreatic islets (1). Although multiple immune cell lineages participate in the pathogenesis of T1D, memory CD4⁺T cells (Tm) play a key inflammatory role in both inducing and maintaining the autoimmune response (2). The central role for CD4⁺Tm cells in driving disease was highlighted in a clinical trial showing that drug-mediated blocking of T-cell activation slowed β-cell destruction (3). Disease progression could also be prevented in the NOD mouse model by artificially guiding CD4⁺Tm cells away from an effector memory phenotype toward a regulatory memory phenotype (4). These studies clearly point toward CD4⁺Tm cells as a promising immune checkpoint target for T1D treatment.

CD4⁺Tm cells comprise functionally distinct central memory (Tcm) and effector memory (Tem) subsets, and their relative numbers in the peripheral circulation and tissues vary depending on the level of antigen exposure (5). CD4⁺Tcm cells express CD62L and CCR7, home to lymph nodes and self-renewal. In contrast, CD4⁺Tem cells lack CD62L, migrate to tissues, and differentiate into T-helper (Th)1, Th2, or Th17 lineages (6). Also contained within CD4⁺Tm cells are autoreactive cells recognizing several β-cell antigens, including GAD65, proinsulin, insulin, chromogranin A, and islet amyloid polypeptide (2). Autoreactive CD4⁺Tm cells can be found in both healthy individuals and T1D patients (7), and their activation (8,9), possibly through accidental cross-reactivity (10) with pathogen-derived antigens (11), leads to destruction of β-cells in an MHC-restricted manner.

CD4⁺Tm cells re-exposed to pathogen or self-antigen differentiate into effector cells, producing multiple cytokines that can modulate CD8⁺T-cell function, innate immunity, and antibody production by plasma cells (6). Alternatively, reactivated CD4⁺Tm cells can differentiate into induced regulatory T (iTr) cells, which are important in maintaining self-tolerance through cell surface regulatory receptors and/or inhibitory cytokines (12). After the clearance of a pathogen, a large fraction of CD4⁺Tem cells undergo contraction (13). This contraction phase involves activation-induced cell death (AICD). Two distinct, but converging pathways mediate AICD in CD4⁺Tm cells, including: 1) the extrinsic death receptor pathway (comprising Fas [CD95] or tumor necrosis factor [TNF] receptor 1) and 2) the intrinsic mitochondrial pathway (14). In healthy individuals, Fas-mediated apoptosis has been suggested to play a major role in the elimination of CD4⁺Tem cells reactivated by the T-cell receptor (TCR) activation pathway (13).

Enhanced effector responses, manifested by elevated levels of Th1 (interferon [IFN]-γ and TNF-α) and Th17 (IL-17) cytokine “signatures,” drive CD4⁺Tm-induced pathology in both T1D patients and the NOD mouse (2,12,15). The intrinsic mechanisms favoring effector versus iTr lineage commitment in CD4⁺Tm cells of T1D patients (as well as in healthy subjects) remain unclear and controversial. Loss-of-function mutations in Fas and FasL genes in mice leads to T-cell–mediated autoimmunity/lymphoproliferative pathology (16–18). Of note, mutations in the Fas system have been detected in patients with T1D (19), and collectively, these observations suggest that problems with effector immune response shutdown perhaps caused by a defect in the death receptor apoptotic pathway may be a contributing factor for pathological autoimmunity in T1D.

In line with this presumption, we asked in this study whether intrinsic differences in CD4⁺Tem responses and/or shutdown by apoptosis discriminate between T1D patients and healthy subjects. Using a unique cohort of age- and sex-matched T1D patients, nonaffected siblings, and unrelated control children, we detail an analysis of proliferation, activation, effector responses, and apoptosis in CD4⁺Tm cells reactivated through CD3/CD28 stimulation, which mimics in vivo stimulation with cognate pathogen and/or self-derived antigen. We found that reactivated CD4⁺Tm cells from both T1D patients and siblings activated easily and transitioned from central memory to effector memory cells very efficiently, leading to an “exaggerated” Tem phenotype with an elevated cytokine signature comprising IFN-γ and IL-10 in T1D patients and IFN-γ in siblings. Likewise, in both T1D patients and their siblings, we observed an exaggerated Tem shutdown effect with heightened sensitivity to AICD, which results in elimination of effector memory by apoptosis.

Thirty-one children with T1D, 45 nonaffected siblings, and 29 age- and sex-matched unrelated control children (Queensland Diabetes Centre, Mater Children’s Hospital, Brisbane, QLD, Australia) were enrolled in this study (Table 1). The study was approved by the human research ethics committee of Mater Children’s Hospital. Informed consent was obtained from parents or guardians.

The following anti-human monoclonal antibodies (mAbs) were used: purified CD8α, HLA-DR, CD45RA, CD11c, CD14, CD16, CD19, CD20, and CD56 (Coulter Immunotech, New South Wales, Australia); CD34, CD66c, CD235a, CD3-Pacific Blue, CD4-fluorescein isothiocyanate, Fas-fluorescein isothiocyanate, CD45RO-phycoerythrin (PE), CD8α-PE-Cy7, CD25-allophycocyanin (APC), CD62L-APC, CD45-APC-H7, caspase-3-PE, and Annexin-V-PE (BD Bioscience, North Ryde, NSW, Australia); and FasL-PE (BioLegend, Karrinyup, WA, Australia). Expression of surface or intracellular molecules was evaluated by flow cytometry (BD FACS LSR II) and analyzed by FlowJo v7.6.5 software (Tree Star Inc.). Absolute counts of CD3⁺T, CD4⁺T, and CD8⁺T cells in the lysed whole blood of T1D patients, siblings, and unrelated control children were calculated using multiparametric flow cytometry and Trucount beads (20).

CD4⁺Tm cells were purified from peripheral blood mononuclear cells by negative separation, excluding CD8⁺/CD45RA⁺/HLA-DR⁺/CD11c⁺/CD14⁺/CD16⁺/CD19⁺/CD20⁺/CD34⁺/CD56⁺/CD66c⁺/CD235a⁺ cells (autoMACS; Miltenyi Biotec, Macquarie Park, NSW, Australia). The pure negative fraction comprised 90–95% CD4⁺CD45RO⁺Tm cells. Unlabeled or carboxyfluorescein succinimidyl ester (CFSE)–labeled CD4⁺Tm cells (1 × 10⁵ cells/well) were maintained in RPMI medium supplemented with 2 mmol/L l-glutamine, 100 units/mL penicillin, 100 units/mL streptomycin, 10 mmol/L HEPES (Life Technologies, Thornton, NSW, Australia), and 10% volume for volume autologous serum in the presence (reactivated cells) or absence (resting cells) of plate-coated anti-CD3 mAb (2 μg/mL; eBioscience, Kensington, SA, Australia) and soluble anti-CD28 mAb (1 μg/mL; BD Bioscience) for 5 days. To address strength of CD3 signaling, in some experiments, plate-bound anti-CD3 mAb was used across a range of concentrations (1, 2, 5, and 10 μg/mL) (Supplementary Fig. 2). In cross-culture assays, autologous serum was replaced with allogeneic serum derived from siblings or unrelated control children or with FCS (Life Technologies). Proliferation dynamics of the reactivated and resting CD4⁺Tm cells were assessed by ³H-thymidine uptake (1 μCi/well, PerkinElmer, Melbourne, VIC, Australia), and kinetics of proliferation were assessed by CFSE dilution assay at days 1, 3, and 5 of culture.

Apoptosis was evaluated in reactivated and resting CD4⁺Tm cells at days 1, 3, and 5 of culture using Annexin-V and 7-aminoactinomycin D (7AAD) (Sigma-Aldrich, Castle Hill, NSW, Australia) staining to discriminate early and late stages of apoptosis (Annexin-V⁺7AAD⁻ and Annexin-V⁺7AAD⁺ cells). To block apoptosis, the antagonistic anti-Fas ZB4 mAb (2.5 μg/mL; Merck Millipore, Bayswater, VIC, Australia) or the isotype control mAb (2.5 μg/mL IgG1 NALE; BD Bioscience) was added at day 1 of culture. Efficiency of the antagonistic anti-Fas ZB4 mAb to block Fas-induced apoptosis was validated in Jurkat cells stimulated with anti-Fas CH-11 mAb (20 μg/mL, Merck Millipore) (Supplementary Fig. 3).

IFN-γ, IL-10, IL-4, IL-2, IL-6, TNF-α, and IL-17 were measured in the supernatants of CD3/CD28-stimulated CD4⁺Tm cells collected at days 1, 3, and 5 of cultures using a cytometric bead array assay per manufacturer’s instructions (BD Bioscience). Cytokine levels were determined from calculated standard curves (FCAP Array software).

CD3/CD28-stimulated CD4⁺Tm cells collected at days 1 and 5 of culture were lysed in Buffer RLT, and total RNA was extracted with an RNeasy Mini kit (QIAGEN, Chadstone Centre, VIC, Australia). cDNA was synthesized using an iScript cDNA synthesis kit and a PCR thermal cycler (Bio-Rad, Gladesville, NSW, Australia). The quantitative PCR assay was performed on the Viia 7 Real-Time PCR System (Applied Biosystems, Life Technologies) to determine the proapoptotic BIM-L, BAD, BAX, and BAK and the antiapoptotic BCL2, BCLXL (BCL2L1), and BCLW (BCL2L2) gene expressions. Expression of β-actin (ACTB) was used as a housekeeping gene for normalization. The following primers were used: BIM-L: F: 5′-AYC CYC AAG ACA GGA GCC C-3′, R: 5′-ATG GAA GCC ATT GCA CTG AGA-3′; BAD: F: 5′-GCA GCC ATC ATG GAG GCG CT-3′, R: 5′-CTG GGC TCC TCC CCC ATC CC-3′; BAX: F: 5′-GCC CTT TTG CTT CAG GGT TTC-3′, R: 5′-TGA GAC ACT CGC TCA GCT TC-3′; BAK: F: 5′-GGC TGA TCC CGT CCT CCA CTG AG-3′, R: 5′-CCT GGG CTA CCT GCT CCT CAG AAG C-3′; BCL2: F: 5′-TGG GAT GCC TTT GTG GAA CTA T-3′, R: 5′-AGA GAC AGC CAG GAG AAA TCA AAC-3′ BCLXL: F: 5′-GGT CGC ATT GTG GCC TTT-3′, R: 5′-TCC TTG TCT ACG CTT TCC ACG-3′; BCLW: F: 5′-CTG ACC CGG CTC CAC GCT GG-3′, R: 5′-TCT GCC ACC AGA GCC CGT GT-3′; and ACTB: F: 5′-CC GTA CGC CAA CAC AGT GC-3′, R: 5′-AT CTC CTG CTT GCT GAT CC-3′ (GeneWorks, Hindmarsh, SA, Australia).

ANOVA and χ² test were used to compare age and sex, respectively, between cohorts. Data within subject groups were compared by Mann-Whitney test and between groups by Kruskal-Wallis test and post hoc multiple comparisons. Spearman correlation test was used to analyze the association between clinical variables (age, age of diagnosis, disease duration, C-peptide levels, islet autoantibodies) and continuous variables (proportion of total, activated, or apoptotic CD4⁺Tm cells). Data were analyzed with GraphPad Prism 6.02 software.

Across the three cohorts, there was no statistical difference in the percentage or absolute counts of CD4⁺T cells defined within the total CD3⁺ cell compartment (Fig. 1A and B). There was also no statistical difference in the percentage or absolute counts of CD4⁺ naive T cells or percentage of CD4⁺Tm cells defined within total CD4⁺ cells. In T1D patients, the percentage or absolute counts of CD4⁺Tm cells did not correlate with disease duration or age of diagnosis (data not shown). However, there was a statistically significant reduction in the absolute counts of CD4⁺Tm cells in T1D patients compared with unrelated control children (P = 0.024) (Fig. 1B). The absolute CD4⁺Tm cell count was also reduced in siblings, but this did not reach statistical significance compared with control subjects (P = 0.07) (Fig. 1B). The lower counts of CD4⁺Tm cells in T1D patients and siblings resulted from a reduction in CD4⁺Tcm cells (Fig. 1B). The alteration in CD4⁺Tm cell counts did not occur in isolation, being associated with lymphopenia and reduced numbers of CD3⁺T, CD8⁺T, and CD8⁺Tm cells in T1D patients and siblings (Table 1 and Supplementary Fig. 1). Neutropenia and reduced monocyte counts were only evident in T1D patients (Table 1).

Using pure ex vivo CD4⁺Tm cells and anti-CD3/CD28 stimulation, we compared the proliferation dynamics of T cells from T1D patients, siblings, and unrelated control children in vitro (Fig. 2). There was no intergroup difference in the proliferation dynamics of CD4⁺Tm cells by TCR stimulation (Fig. 2A). We also examined the kinetics of proliferation in the reactivated CD4⁺Tm cells post-CD3/CD28 stimulation through CFSE labeling and multiparametric flow cytometry. Based on the level of CFSE expression at day 1, reactivated CD4⁺Tm cells underwent more than four divisions between day 3 and 5 of culture, with a proliferation index of 2.32, 2.27, and 2.59 and frequency of divided parent cells of 47.9%, 42.0%, and 54.3% in T1D patients, siblings, and control subjects, respectively (Fig. 2B). Although no statistical difference was found in the proliferation potential of CD4⁺Tm cells, an increased percentage of CD4⁺Tm cells from T1D patients and siblings were activated by CD3/CD28 stimulation (Fig. 2C). The CD4⁺Tm cells from T1D patients and siblings underwent heightened activation regardless of the concentration of the plate-bound anti-CD3 mAb in the range of 1–10 μg/mL (Supplementary Fig. 2). This increase in activated CD4⁺CD25⁺Tm cells was not due to an increase in iTr cells (CD4⁺CD25⁺Foxp3⁺) because iTr cells accounted for a similar minority (<15% of CD4⁺Tm cells) across all cohorts (data not shown). These results show that although CD4⁺Tm cells of T1D patients and siblings exhibited similar proliferation dynamics and kinetics compared with control subjects, they were acutely more sensitive to TCR-induced activation.

Given the differences in CD4⁺Tm-cell postactivation status between cohorts, we next performed an in-depth analysis of cell phenotype and function (Fig. 3). After CD3/CD28 stimulation, CD4⁺Tm cells from healthy control subjects maintained a CD62L⁺Tcm- and CD62L⁻Tem-cell compartment size similar to that at day 1 of culture (Fig. 3A). This was in contrast to CD4⁺Tm cells of T1D patients and siblings, where the percentage of CD62L⁺Tcm cells was markedly lower after 5 days of culture (Fig. 3A). Thus, in T1D patients and siblings, CD4⁺Tm cells effectively transitioned to CD62L⁻Tem cells after TCR-induced activation.

Another functional difference was related to cytokines secreted by CD4⁺Tm cells after activation (Fig. 3B). Unlike cells from healthy control subjects, CD4⁺Tm cells from T1D patients significantly increased IFN-γ and IL-10 production post-CD3/CD28 stimulation, whereas CD4⁺Tm cells from siblings elevated IFN-γ production. Although CD3/CD28 stimulation failed to increase the amount of secreted IL-4 across cohorts, IL-4 production appeared to be the best sustained in reactivated CD4⁺Tm cells from T1D patients (Fig. 3B). Kinetic studies determined that the rise in IFN-γ and IL-10 cytokine secretion occurred between days 3 and 5 of culture (data not shown). Across cohorts, there was no difference in the production of IL-6, TNF-α, and IL-17 (data not shown). Of note, CD3/CD28 stimulation reduced the amount of secreted IL-2 by CD4⁺Tm cells across all cohorts, tentatively suggesting that reactivated CD4⁺Tm cells are taking up the IL-2 that they produce through an autocrine process. Overall, the data suggest that TCR-induced stimulation promotes a Tem-cell phenotype in both T1D patients and siblings, characterized by upregulation of IFN-γ and IL-10 in T1D patients and upregulation of IFN-γ in siblings.

Given the previous observation that CD4⁺Tem cells are prone to Fas-mediated AICD, which results in their elimination by apoptosis (13,21), we next examined CD4⁺Tm apoptosis kinetics in T1D patients, siblings, and unrelated control children (Fig. 4). Across all cohorts, resting CD4⁺Tm cells underwent comparable rates of apoptosis (Fig. 4A). In contrast, after CD3/CD28 stimulation, apoptosis rates of CD4⁺Tm cells in T1D patients and siblings were markedly elevated compared with control subjects (Fig. 4A). Additionally, elevated apoptosis rates of CD4⁺Tm cells in T1D patients and siblings were evident at anti-CD3 mAb concentrations of 1–10 μg/mL (Supplementary Fig. 2). Kinetic studies determined that most apoptosis occurred between days 3 and 5 of culture (data not shown). Apoptotic cells were predominantly in the early stage of apoptosis, with a minority (<10%) in late apoptosis (data not shown). Apoptosis in reactivated CD4⁺Tm cells of T1D children and siblings was also associated with increased frequencies of cells expressing the active form of caspase-3 (Fig. 4B). Interactions with serum proteins was excluded as a possible cause of differential CD4⁺Tm apoptosis rates because comparable rates were observed regardless of the serum source (Fig. 4C).

The death receptor (Fas) was detected on the surface of the majority of reactivated CD4⁺Tm cells in all subject groups but with a higher frequency in T1D patients and siblings than in control subjects (Fig. 5A). In contrast to universally high Fas expression, FasL was expressed on a minority of reactivated CD4⁺Tm cells in all subject groups but, again, with a higher frequency in T1D patients and siblings (Fig. 5A). Treatment with the antagonistic anti-Fas ZB4 mAb, which blocks the cognate interaction between Fas and FasL, reduced Fas expression on reactivated CD4⁺Tm cells in all subject groups and blocked Fas-induced apoptosis in Jurkat cells but did not alter the levels of apoptosis of reactivated CD4⁺ Tm cells in any subject group (Fig. 5B and C and Supplementary Fig. 3). Increased Fas and FasL expression in reactivated CD4⁺Tm cells of T1D patients and siblings was coupled with an increase of proapoptotic BIM-L, BAD, BAX, and BAK transcripts compared with healthy control subjects (Fig. 5D). Reactivated CD4⁺Tm cells of T1D patients showed a statistically significant spike in BAD expression that separated T1D subjects from healthy control subjects (Fig. 5D). Of note, the antiapoptotic transcripts BCL2, BCLXL, and BCLW remained unchanged in reactivated CD4⁺Tm cells across all cohorts (Fig. 5D). These data suggest that reactivated CD4⁺Tm cells of T1D patients and siblings are acutely susceptible to AICD through a Fas-independent mechanism.

Given that CD4⁺Tm cells in T1D patients and siblings were highly susceptible to AICD, we next determined which type of memory cell was prone to AICD. In healthy control subjects, we observed that the majority of apoptosis occurred in nondividing cells (Fig. 6A). In contrast, CD4⁺Tm cells of T1D patients and siblings showed consistently high levels of apoptosis across every cell division (Fig. 6A). Given that CD4⁺Tm cells encompass functionally distinct lineages (5), we subdivided the analysis of apoptosis into Tcm (CD62L⁺) and Tem (CD62L⁻) lineages. The majority of CD4⁺Tm cells undergoing apoptosis were within the Tem lineage, which encompassed >80% of apoptotic cells across each cohort (Fig. 3B). Importantly, we observed that the apoptotic bias toward Tem cells was more pronounced in T1D patients and siblings than in control subjects, eliminating approximately one-half of the cells from the Tem lineage in T1D patients and siblings but only one-third of CD4⁺Tem cells in control subjects (Fig. 6C). In contrast to the Tem lineage, the Tcm lineage remained largely resistant to AICD; however, the Tcm lineage of T1D patients and siblings still showed greater rates of AICD compared with control subjects (Fig. 6C). These data suggest that the cells from the Tem lineage are highly susceptible to AICD in T1D patients and siblings.

T1D is one of the most heritable common diseases, with high sibling relative risk and the largest concordance rate in monozygotic twins among the autoimmune diseases (22,23). Thus, if we are to understand the mechanisms of T1D pathophysiology and intervene with preventive or curative therapies, we must look carefully at the causes of immune dysfunction in patients and their at-risk siblings. To address this gap in knowledge, we undertook a high-definition analysis of CD4⁺Tm-cell composition and function in an extremely well-characterized age- and sex-matched cohort of children comprising T1D patients, their nonaffected siblings, and unrelated healthy control subjects.

We identified a distortion in the global CD4⁺Tm-cell repertoire, with T1D patients having lower absolute counts of CD4⁺Tm cells compared with healthy control subjects. In both T1D patients and their siblings, lower absolute counts of CD4⁺Tm cells occurred at the expense of CD4⁺Tcm cells. This reduction in CD4⁺Tm cells was associated with reduced CD3⁺T-, CD8⁺T-, and CD8⁺Tm-cell counts and lymphopenia in T1D patients and their siblings. Given this distortion in global CD4⁺T-cell memory, we next asked whether a problem existed with T-cell proliferation, activation, and/or effector immune response shutdown in T1D patients and siblings. Post stimulation, no intergroup differences in cell proliferation was observed, indicating that CD4⁺Tm cells of T1D patients and siblings were not senescent, exhausted, or defective in cell replication. However, we observed a difference in cell phenotype poststimulation, with a spike in the activated CD4⁺CD25⁺Tm cells in T1D patients and siblings consistent with the reported presence of activated CD25⁺CD3⁺T cells in T1D patients and siblings (8).

The heightened activation potential of CD4⁺Tm cells was further reinforced by the observation that reactivated CD4⁺Tm cells of T1D patients and siblings were capable of exaggerated effector responses through efficient switching from central memory (CD62L⁺) to effector memory (CD62L⁻). This phenotype may provide some evolutionary benefit. For instance, compared with healthy control subjects, reactivated CD4⁺Tm cells of T1D patients and siblings can better upregulate IFN-γ. IFN-γ is critical for clearing bacteria, viruses, and cancers (24,25), and higher levels may result in more efficient T-cell surveillance. Indeed, there is some evidence of a decreased risk of particular cancers in T1D patients (26,27). In line with this idea, we recently identified a reduced rate of infection with common childhood viruses in T1D patients and siblings compared with age-, sex-, and geographically matched children (J.J.M., personal communication); however, a larger cohort is needed to categorically establish this link. Elevated levels of IL-10 production by reactivated CD4⁺Tm cells in T1D patients may also be significant given that this cytokine is known to mediate the destruction of β-cells in the NOD mouse (28). Sustained levels of the multifunctional pleiotropic cytokine IL-4 in reactivated CD4⁺Tm cells of T1D patients is also of interest given the central role of this cytokine in guiding the alternative fates of T-cell differentiation (29). Additionally, it appeared that reactivated CD4⁺Tm cells of T1D patients and siblings rapidly take up secreted IL-2 through an autocrine process. Determining whether any differences in cytokine consumption among T1D patients, siblings, and control subjects, however, will require additional experimental approaches (e.g., quantitative PCR, intracellular cytokine staining).

A striking observation in CD4⁺Tem cells of both T1D patients and siblings was a heightened sensitivity to AICD. Apoptosis in CD4⁺Tem cells occurred through TCR stimulation and might be expected to result from Fas-mediated death receptor pathways (13). Although we observed upregulation of Fas and FasL in reactivated CD4⁺Tm cells, there was upregulation of proapoptotic BIM-L, BAD, BAX, and BAK transcripts, which are associated with the mitochondrial apoptotic pathway. Additionally, we observed upregulation of the active form of caspase-3 shared by both apoptotic pathways. Of note, only in reactivated CD4⁺Tm cells of T1D patients was a significant upregulation of the proapoptotic gene BAD observed, which discriminated this disease cohort from the other subject groups. Elevated AICD rates in CD4⁺Tem cells from both T1D patients and siblings could not be prevented by blocking the Fas receptor, suggesting that a defect may exist in the Fas death receptor pathway through altered lipid raft microdomains (13) or in the assembly of the Fas-associated death-inducing signaling complex (30). Alternatively, upregulation of BIM-L, BAD, BAX, and BAK transcripts suggests that mitochondrial apoptotic pathways may be mainly responsible for the increased rate of apoptosis in CD4⁺Tem cells. Finally, although CD4⁺Tem cells of T1D patients and siblings showed similar susceptibility to apoptosis, we cannot rule out that apoptosis could differ at the clonotype level. It is possible that in T1D patients or siblings who go on to develop disease, autoreactive effector cells could be particularly resistant to apoptosis. Therefore, further subdividing the CD4⁺Tem-cell lineage with additional phenotypic/functional markers and full TCR repertoire analysis would be of key interest.

Previous investigations of CD4⁺T-cell biology in T1D have been restricted to the comparison of patients with unrelated healthy control subjects (31,32). These comparisons likely missed critical genetic and environmental cues that underpin CD4⁺Tm-cell pathomechanisms in T1D patients and siblings (33). In the current study, we identified several novel aspects of CD4⁺Tm-cell biology common to T1D patients and siblings, including reduced absolute CD4⁺Tm and CD4⁺Tcm counts, exaggerated CD4⁺Tem phenotype with an elevated IFN-γ signature, heightened sensitivity to activation, and AICD. These features of the CD4⁺Tm cells common to T1D patients and siblings associated with lymphopenia (33) but not with clinical variables such as age, age of disease, C-peptide level, and autoantibody status and are evident before the onset of autoimmunity and islet antibody detection.

In conclusion, this study reveals a novel hyperactive phenotype of reactivated CD4⁺Tm cells of T1D patients and their siblings. In these cells, we found that an exaggerated activation switch (activated CD25 receptor and Tem phenotype) is counterbalanced by an exaggerated shutdown switch (AICD). Whether an exaggerated activation switch herein induced in reactivated CD4⁺Tm cells by polyclonal TCR stimulation also operates across other T-cell lineages (e.g., CD8⁺T-cell lineages) as well as in response to antigen-specific stimulation warrants further study. Precisely how this exaggerated activation switch in the CD4⁺Tm-cell phenotype relates to β-cell destruction will require a thorough immunological dissection. Nonetheless, these data help to further define the mechanisms underlying T1D pathophysiology and illuminate new biologically relevant surface receptors, cytokines, biochemical pathways, and specific immune cell lineages for therapeutic targeting.

Acknowledgments. The authors thank the families and children for participating in this study and Tracey Baskerville and Sherrell Cardinal (Mater Children’s Hospital, Brisbane, QLD, Australia) and Stephanie Diaz-Guilas (Mater Research, Translational Research Institute, Brisbane, QLD, Australia) for recruiting the children for this study. The authors also thank Yong H. Sheng (Mater Research, Translational Research Institute) for reagents and assistance with the quantitative RT-PCR and Nigel Waterhouse and Ray Steptoe (Mater Research, Translational Research Institute) for advice on experimental design and data analysis.

Funding. This work was supported by an Australia Postgraduate Award and Mater Medical Research Institute Top-up PhD Fellowship to M.L.B. J.J.M. is a National Health and Medical Research Council Career Development Fellow.

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

Author Contributions. M.L.B. contributed to the performance of experiments, data analysis, and the preparation of the manuscript. O.H. contributed to the performance of experiments and data interpretation. D.M. contributed to the experimental design and data interpretation. M.H. and A.C. contributed to the oversight of clinical sampling, provision of clinical data, and review and editing of the manuscript. J.J.M. contributed to the research data interpretation and writing of the manuscript. S.V. contributed to the study design, data interpretation, and writing of the manuscript. S.V. 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.