Regulatory T cell phenotype and function 4 years after GAD–alum treatment in children with type 1 diabetes

Abstract

Glutamic acid decarboxylase (GAD)₆₅ formulated with aluminium hydroxide (GAD-alum) was effective in preserving insulin secretion in a Phase II clinical trial in children and adolescents with recent-onset type 1 diabetes. In addition, GAD-alum treated patients increased CD4⁺CD25^hi forkhead box protein 3⁺ (FoxP3⁺) cell numbers in response to in-vitro GAD₆₅ stimulation. We have carried out a 4-year follow-up study of 59 of the original 70 patients to investigate long-term effects on the frequency and function of regulatory T cells after GAD-alum treatment. Peripheral blood mononuclear cells were stimulated in vitro with GAD₆₅ for 7 days and expression of regulatory T cell markers was measured by flow cytometry. Regulatory T cells (CD4⁺CD25^hiCD127^lo) and effector T cells (CD4⁺CD25^–CD127⁺) were further sorted, expanded and used in suppression assays to assess regulatory T cell function after GAD-alum treatment. GAD-alum-treated patients displayed higher frequencies of in-vitro GAD₆₅-induced CD4⁺CD25⁺CD127⁺ as well as CD4⁺CD25^hiCD127^lo and CD4⁺FoxP3⁺ cells compared to placebo. Moreover, GAD₆₅ stimulation induced a population of CD4^hi cells consisting mainly of CD25⁺CD127⁺, which was specific of GAD-alum-treated patients (16 of 25 versus one of 25 in placebo). Assessment of suppressive function in expanded regulatory T cells revealed no difference between GAD-alum- and placebo-treated individuals. Regulatory T cell frequency did not correlate with C-peptide secretion throughout the study. In conclusion, GAD-alum treatment induced both GAD₆₅-reactive CD25⁺CD127⁺ and CD25^hiCD127^lo cells, but no difference in regulatory T cell function 4 years after GAD-alum treatment.

Introduction

Type 1 diabetes (T1D) is a consequence of an autoimmune reaction towards insulin-producing β cells of the pancreas. Immunomodulatory approaches to prevent or treat T1D have been developed and tested, with variable results 1–4. Autoantigens may be used to induce immunologic tolerance as an alternative to immunosuppression 5. Glutamic acid decarboxylase 65 (GAD₆₅) is one of the main antigens to which patients with T1D mount a destructive immune response, and autoantibodies directed against GAD₆₅ (GADA) and T cells specific for GAD₆₅ epitopes are common in T1D patients 6–8. We have shown previously preservation of residual insulin secretion by GAD-alum treatment in a Phase II clinical trial in children with recent-onset T1D 3. In addition, trial participants treated with GAD-alum up-regulated CD4⁺CD25^hiforkhead box P3⁺ (FoxP3⁺) cells in response to GAD₆₅ stimulation in vitro and had a predominant T helper type 2 (Th2) immune response 9,10. Preservation of C-peptide secretion was still detectable after 4 years in patients with <6 months T1D duration at baseline in the same trial 11, and the residual C-peptide secretion was accompanied by sustained high levels of GADA, increased memory T cell frequencies and T cell activation upon in-vitro GAD₆₅ stimulation 12. Recently, additional Phases II and III clinical trials of GAD-alum have been conducted both in Europe and the United States, neither finding an effect on preservation of insulin secretion 13,14. The present Phase II trial included patients with a T1D duration of <18 months, whereas the European Phase III trial included patients with a duration of <3 months, which may contribute to the discrepancy in outcome.

Self-tolerance is maintained physiologically by regulatory T cells (T_reg) in the periphery 15, and defects in T_reg function have been hypothesized to be involved in the pathogenesis of autoimmune disease 16. Because tolerance in the periphery is maintained by T_regs, induction of active tolerance has long been a proposed mechanism of action of antigen-based therapies such as GAD-alum treatment 17. T_regs typically express high levels of the interleukin (IL)-2 receptor α-chain CD25, the transcription factor FoxP3 and low levels of the IL-7 receptor CD127 18–22. However, both FoxP3 and CD25 can also be expressed by activated non-regulatory T cells. CD39 has also been suggested to be involved in T_reg function through the removal of adenosine triphosphate (ATP) and has thus been used to identify subsets of T_regs 23. T_regs can suppress proliferation and cytokine secretion in a broad range of cell types, including CD4⁺ and CD8⁺ T cells, and their dysfunction leads to immunopathology 24. It has been reported recently that rather than there being a deficiency in T_reg numbers, effector T cells (T_eff) from patients with T1D are resistant to T_reg-mediated suppression 25,26.

The aim of this work was to investigate whether an increase in cells with a T_reg phenotype persisted at 4 years after GAD-alum treatment. In addition, we tested whether GAD-alum treatment affected the suppressive capacity of T_regs.

Materials and methods

Ethics statement

This study was approved by the Research Ethics Committee at the Faculty of Health Sciences, Linköping University, Sweden. Written informed consent was obtained from participating individuals, and for those aged <18 years also their parents, in accordance with the Declaration of Helsinki.

Population

The design and characteristics of the Phase II trial have been described elsewhere 3. Briefly, 70 T1D children between 10 and 18 years of age with fewer than 18 months of disease duration were recruited at eight Swedish paediatric centres. Participants had a fasting serum C-peptide level above 0·1 nmol/l and detectable GADA at inclusion. They were randomized to subcutaneous injections of 20 μg GAD-alum (n = 35) or placebo (n = 35) at day 0 and a booster injection 4 weeks later in a double-blind setting. After 4 years, patients and their parents were asked whether they were willing to participate in a follow-up study. Fifty-nine patients, of whom 29 had been treated with GAD-alum and 30 received placebo, agreed to participate.

Antibodies

Fluorescein isothiocyanate (FITC)-conjugated anti-CD39 (clone A1; Biolegend, San Diego, CA, USA), phycoerythrin (PE)-conjugated anti-FoxP3 (clone PCH101), allophycocyanin (APC)-conjugated anti-CD25 (clone BC96) and FITC- and PE-cyanine 7 (PE-Cy7)-conjugated anti-CD127 (clone eBioRDR5; eBioscience, San Diego, CA, USA), Alexa 700- and Pacific Blue-conjugated anti-CD4 (clone RPA-T4), APC-Cy7-conjugated anti-CD25 (clone M-A251; BD Pharmingen, Franklin Lakes, NJ, USA), and relevant isotype- and fluorochrome-matched control antibodies were used in this study. In addition, 7-amino-actinomycin D (7-AAD; BD Pharmingen) was used to measure cell viability.

Flow cytometry

Peripheral blood mononuclear cells (PBMC) from GAD-alum-treated (n = 24) and placebo-treated (n = 25) patients were isolated from whole blood by Ficoll-Paque (Pharmacia Biotech, Piscataway, NJ, USA) density gradient centrifugation within 24 h after drawing. Three of the GAD-alum-treated patients were classified as responders, 14 as intermediate responders and seven were non-responders, while the placebo group contained 10 intermediate responders and 14 non-responders (one patient unclassified due to missing serum sample). PBMC were incubated in AIM-V medium (Invitrogen, Carlsbad, CA, USA) with β-mercaptoethanol at 37°C, 5% CO₂ for 7 days, with or without 5 μg/ml recombinant GAD₆₅ (Diamyd Medical, Stockholm, Sweden). One million cells were washed in 2 ml phosphate-buffered saline (PBS) containing 0·1% bovine serum albumin (BSA; Sigma-Aldrich, St Louis, MO, USA) and subsequently stained with anti-CD4, CD39, CD127 and CD25 antibodies. Cells were then fixed and permeabilized using a FoxP3 staining kit (eBioscience), according to the manufacturer's instructions. After washing, cells were stained with PE anti-FoxP3, reconstituted in PBS, acquired on a fluorescence activated cell sorter (FACS) (BD FACSAria) and analysed using Kaluza software version 1·1 (Beckman Coulter, Indianapolis, IN, USA). The FoxP3⁺ gate was set using the negative population, as the negative population had a higher median fluorescence intensity (MFI) than the isotype control.

Cell sorting

Cells were sorted and expanded when sufficient cell numbers were available. Cryopreserved cells from GAD-alum- (n = 4) and placebo- (n = 3) treated patients were stained with Pacific Blue conjugated anti-CD4, FITC-conjugated anti-CD127 and APC-conjugated anti-CD25 and sorted into T_reg and T_eff subsets based on CD4⁺CD25^hiCD127^lo and CD4⁺CD25^–CD127⁺ phenotype, respectively. After sorting, cells were pelleted by centrifugation at 400 g for 10 min, resuspended in AIM-V 10% human serum (HS) and allowed to rest for 2 h at 37°C, 5% CO₂ before expansion was initiated. Aliquots of sorted cells were re-acquired to assess purity. The average T_eff contaminant in sorted T_regs was 0·1%. PBMC from one single freshly drawn healthy donor were stained, sorted as above and stored frozen to serve as interassay control.

T_eff and T_reg expansion

T_regs were distributed at 4 × 10⁴ cells per well in 125 μl AIM-V 10% HS into 96-well U-bottomed plates, and stimulated with anti-CD3/CD28 Dynabeads (Invitrogen) at a 1:1 bead-to-cell ratio. T_effs were plated at 5 × 10⁵ cells per 500 μl medium, into 96-well flat-bottomed plates precoated overnight with 10 μg/ml anti-CD3 (OKT3; eBioscience) at 4°C. Cultures also contained 1 μg/ml soluble anti-CD28 antibody (CD28·2; eBioscience). Culture volume was doubled the following day, and 30 and 300 U/ml of recombinant human IL-2 (R&D Systems, Abingdon, UK) were added to T_eff and T_reg cultures, respectively. T_regs were washed and supplemented with fresh IL-2 every 2 days. T_regs and T_effs were restimulated as above on the ninth day of culture, and frozen down after 15 days of expansion. To verify post-expansion phenotype, cryopreserved T_regs and T_effs were cultured for 24 h in AIM-V 10% HS and 5 U/ml IL-2, and subsequently stained and acquired as described above.

Suppression assay

T_reg-mediated suppression was assessed using expanded T_regs and T_effs from placebo- (n = 3) and GAD-alum- (n = 4) treated patients. Expanded T_regs and T_effs were thawed and incubated in AIM-V 10% HS at 37°C, 5% CO₂ overnight, then resuspended at 0·5 × 10⁵ cells/ml. T_effs were plated into 96-well U-bottomed plates at a density of 5 × 10⁴ cells per well, while T_regs were plated into T_eff-containing wells at T_reg-to-T_eff ratios of 1:1, 1:2, 1:4, 1:8 and 1:16. T_reg/T_eff cultures were stimulated with 5 μg/ml soluble anti-CD3 and 1 μg/ml soluble anti-CD28 antibodies. Unstimulated wells were included as negative controls, both from patients and interassay control healthy T_effs. IL-2 (1 U/ml) was added to all wells. Supernatants were collected after 3 days of culture and cells were incubated with 0·2 μCi [³H]-thymidine (PerkinElmer, Waltham, MA, USA) for 18 h before harvesting. Thymidine incorporation was measured using a 1450 Wallac MicroBeta counter (PerkinElmer).

C-peptide measurements

C-peptide levels were measured in serum samples with a time-resolved fluoroimmunoassay (AutoDELFIA™ C-peptide kit, Wallac; PerkinElmer), as described 3. Stimulated C-peptide was measured during a mixed meal tolerance test (MMTT) in GAD-alum- (n = 21) and placebo- (n = 10) treated patients who had a maximal C-peptide response of >0·20 nmol/l at the 30-month follow-up. Clinical effect of treatment was defined by changes in stimulated C-peptide measured as area under the curve (AUC) from baseline to 48 months.

Statistical analysis

Statistically significant differences were determined using the Mann–Whitney two-tailed U-test for unpaired observations, as data were determined to be significantly different from a Gaussian distribution. Wilcoxon's signed-rank test was used to compare paired samples. Linear regression was used to compare slope and Y-intercept of suppression curves, and correlations were determined with Spearman's rank correlation coefficient test. A probability level of <0·05 was considered statistically significant. All statistical analyses were performed using GraphPad Prism software, version 5·04 (GraphPad Software, Inc., La Jolla, CA, USA).

Results

GAD₆₅ stimulation increases the proportion of both CD25^hiCD127^lo and CD25⁺CD127^+/hi cells in GAD-alum-treated patients

We have demonstrated previously that in-vitro stimulation with GAD₆₅ induced CD4⁺CD25^hi FoxP3⁺ cells in PBMC from GAD-alum-treated patients 9. To determine whether this effect persisted 4 years after treatment, we analysed CD25^hiCD127^lo cells and used FoxP3 and CD39 as additional markers to discriminate T_regs from activated T cells more accurately. Thus, the expression of CD25, CD127, FoxP3 and CD39 on CD4⁺ lymphocytes was analysed in PBMC after 7 days of incubation with or without GAD₆₅. Gates used for analysis and representative PBMC samples describing the expression of CD4, CD25 and CD127 are shown in Fig. 1a,b.

Fig. 1.

Gating strategy and glutamic acid decarboxylase (GAD₆₅)-induced T cell populations. (a,b) Gating strategy and representative peripheral blood mononuclear cell (PBMC) samples from a GAD-alum-treated subject at resting conditions and after in-vitro GAD₆₅ stimulation, respectively. Upper left panel describes the division of lymphocytes into small and large. Upper right panel describes CD4 expression. The lower panels describe the distribution of CD4⁺ cells between the CD25^hiCD127^lo and the CD25⁺CD127⁺ populations, gates indicated above plots. (c,d) Expression of CD25 and CD127 on CD4⁺ cells (placebo n = 25, GAD-alum n = 23). (e) Changes in CD4 median fluorescence intensity (MFI) and (f) frequency of forward-scatter (FSC)^hiside-scatter (SSC)^hi cells were calculated by subtracting MFI and frequencies at resting conditions from those obtained after GAD₆₅ stimulation and are represented as delta values (placebo n = 25, GAD-alum n = 24). The cut-off for high CD4 expression was defined as an increase of 10 units of CD4 MFI in CD4⁺ cells following GAD₆₅ stimulation and is indicated by the dashed line in (e). (g) Frequencies of CD25^hiCD127^lo, CD25⁺CD127⁺ and forkhead box protein 3⁺ (FoxP3⁺) cells among FSC^hiSSC^hi (red squares) and small (blue circles) CD4⁺ cells from GAD-alum-treated patients (n = 15) are shown as %. Each point represents an individual and median values are indicated with horizontal lines. Statistical significances are shown as P-values.

The frequency of CD25^hiCD127^lo cells in the CD4⁺ population was increased significantly upon GAD₆₅ stimulation in GAD-alum-treated patients compared to unstimulated cells (7·4% and 4·5%, respectively), but not in the placebo group (Fig. 1c). The frequency of CD4⁺CD25⁺CD127⁺ cells was also increased following GAD₆₅ stimulation in the GAD-alum-treated patients compared to resting cells (13·1% and 2·6%, respectively; Fig. 1d). In contrast, GAD₆₅ stimulation did not induce expression of CD25^hiCD127^lo or CD25⁺CD127⁺ compared to resting cells in the placebo group (Fig. 1c,d). The frequencies of CD4⁺CD25^hiCD127^lo and CD4⁺CD25⁺CD127⁺ cells were also significantly higher in the GAD-alum-treated group compared to placebo individuals after stimulation with GAD₆₅ (Fig. 1c,d).

GAD₆₅ stimulation induces a population of activated T cells in GAD-alum-treated subjects

Stimulation with GAD₆₅ in GAD-alum-treated patients induced a population of forward-scatter (FSC)^hiside-scatter (SSC)^hi cells, consisting mainly of CD4⁺ memory T cells, as we have reported previously 12. These FSC^hiSSC^hi cells are illustrated in Fig. 1a,b and are characterized by high CD4 expression (Fig. 1e). The FSC^hiSSC^hi population was observed in 16/24 GAD-alum patients and in one of 25 placebo individuals. In line with the GAD₆₅ recall response induced in GAD-immunized individuals, GAD₆₅ stimulation induced higher CD4 MFI (Fig. 1e) and higher percentages of FSC^hiSSC^hi cells (Fig. 1f) among CD4⁺ cells from GAD-alum patients compared to the placebo group. Next, we analysed the expression of T_reg-associated markers among FSC^hiSSC^hi CD4⁺ cells from the GAD-alum group, and found that 25% were CD25^hiCD127^lo, 46·2% were CD25⁺CD127^+/hi and 74% were FoxP3⁺ (Fig. 1g).

GAD₆₅ stimulation induces cells with a T_reg phenotype in GAD-alum-treated subjects

FoxP3 expression on CD4⁺ and CD4⁺FSC^hiSSC^hi cells was enhanced significantly by GAD₆₅ stimulation in the GAD-alum group (Fig. 2a–c), while GAD₆₅ stimulation did not induce any change compared to resting cells in the placebo group (Fig. 2c). To define further whether the increased CD25⁺CD127^lo population in GAD₆₅ stimulated PBMC from GAD-alum-treated patients corresponded to a T_reg population, CD39 and FoxP3 were added as additional T_reg markers. Indeed, CD4⁺CD25^hiCD127^lo FoxP3⁺CD39⁺ cells were also found to be increased selectively in these patients following in-vitro GAD₆₅ stimulation (Fig. 2d). Thus, in-vitro GAD recall leads to expansion of both T_regs and activated CD25⁺CD127⁺ T effector cells, which is observed only in patients treated previously with GAD-alum. There were no significant differences in expression of any measured marker on resting cells between the two treatment arms (Figs 1 and 2).

Fig. 2.

Expression of regulatory T cell (T_reg)-associated markers. (a) Expression of forkhead box protein 3 (FoxP3) on CD4⁺ resting cells from a representative glutamic acid decarboxylase (GAD)-alum-treated subject; (b) expression of FoxP3 in GAD-stimulated cells, gates indicated above plots. (c) Expression of FoxP3 on CD4⁺ cells (placebo n = 25, GAD-alum n = 24); (d) the fraction of CD4⁺ cells co-expressing CD25^hiCD127^loFoxP3⁺CD39⁺ from placebo (n = 24) and GAD-alum (n = 22) patients. Each point represents an individual and median values are indicated with horizontal lines. Statistical significances are shown as P-values.

The post-expansion phenotype of T_regs and T_effs remains lineage-specific

T_regs (CD4⁺CD25^hiCD127^lo) from GAD-alum-treated patients expanded approximately 900-fold, to a similar extent as T_regs from placebo-treated patients (800-fold; Table 1). T_effs (CD4⁺CD25^–CD127⁺) from both GAD-alum- and placebo-treated patients expanded approximately 100-fold. To verify the phenotype of sorted and expanded T_regs and T_effs after cryopreservation, we analysed the expression of T_reg markers on thawed cells by flow cytometry. T_regs maintained predominant expression of CD25, FoxP3, cytotoxic T lymphocyte antigen-4 (CTLA-4) and low expression of CD127 and CD45RA, and roughly 50% were CD39⁺. Meanwhile, expression of FoxP3, CTLA-4, CD45RA and CD25 was scarce on T_eff, but the majority expressed CD127 and expressed CD39 in the same proportion as T_regs. Measurement of cell viability by 7-AAD staining 24 h after thawing demonstrated that T_effs had a viability of 90%, whereas 70% of T_regs were viable (data not shown).

Table 1.

Characteristics of expanded cells after freeze/thaw.

	Treg	Teff
	CD25hiCD127lo	CD25−CD127+
In-vitro expansion	Fold increase (range)	Fold increase (range)
GAD-alum	896 (706–1389)	117 (75–291)
Placebo	800 (667–1039)	80 (45–180)

Protein expression phenotype	% (range)	% (range)
CD4+ out of total CD3+	98·8 (92·3–99·3)	97·9 (67·6–99·3)
CD25hiCD127lo out of total CD4+	94·5 (88·8–98·5)	5·3 (2·7–32·3)
FoxP3+ out of total CD4+	81·4 (51·7–89·7)	9·6 (4·9–16·2)
CTLA-4+ out of total CD4+	74·7 (30·7–96·5)	14·4 (3·8–26·4)

Expression of regulatory T cell (T_reg) markers was measured by flow cytometry 24 h after thawing. Data are pooled from both placebo and glutamic acid decarboxylase (GAD)-alum-treated type 1 diabetes (T1D) patients. CTLA-4: cytotoxic T lymphocyte antigen-4; FoxP3: forkhead box protein 3; T_eff: effector T cells.

Expression of T_reg markers is not related to clinical outcome

We tested whether the expression of any of the markers affected by GAD₆₅ stimulation was related to clinical outcome of treatment. We found no significant correlation between expression of T_reg markers used in this study and changes in stimulated C-peptide measured as ΔAUC or AUC 4 years after treatment. C-peptide secretion was not significantly different in patients where an FSC^hiSSC^hi population was induced by GAD₆₅ stimulation compared to those who did not respond in this way (data not shown).

Suppression of polyclonally activated T_effs is not affected by GAD-alum treatment

To test whether the function of T_regs in T1D children included in the Phase II trial was affected by GAD-alum or placebo administration, suppression assays using sorted and expanded T_regs (CD4⁺CD25^hiCD127^lo) and T_effs (CD4⁺CD25^–CD127⁺) were performed. Gates used to sort T_regs and T_effs are illustrated in Fig. 3a. Expanded T_regs from patients treated with GAD-alum suppressed proliferation of autologous T_effs to the same extent as T_regs from placebo patients (Fig. 3b). T_regs from both groups of patients displayed dose-dependent suppression of proliferation. As reported previously 25,27,28, we further found that suppression in autologous cultures of T_regs and T_effs was reduced in all patients (n = 7, placebo and GAD-alum combined) compared to a healthy control (seven repeated measurements, Fig. 3c, P < 0·0001). To determine whether this attenuated suppression was intrinsic to T_regs or T_effs, we tested the suppression of T_regs from T1D patients (either GAD-alum- or placebo-treated, with similar results; Fig. 3b,d), and from a healthy control in autologous and cross-over culture suppression assays. As shown in Fig. 3c, T1D T_regs exerted the same level of suppression on T_effs coming from either T1D or healthy subjects. In the reverse experiment, healthy T_regs were able to suppress T_effs from healthy or T1D subjects to a similar degree. Taken together, these results suggest that attenuated suppression from T_regs of T1D patients is due to reduced T_reg efficacy rather than to increased T_eff resistance to suppression.

Fig. 3.

Dose-dependent regulatory T cell (T_reg)-mediated suppression of effector T cells (T_eff). Proliferation was assessed by [³H]-thymidine incorporation. Percentage of suppression of proliferation in co-cultures of T_regs and T_effs was calculated by dividing the mean counts per minute (cpm) of co-culture wells by the mean cpm obtained from T_effs cultured alone. (a) Left panel illustrates the gate discriminating CD4⁺ cells and the right panel shows the gates for sorting CD25^hiCD127^lo (T_regs) and CD25^–CD127^+/hi cells (T_effs). (b) Suppression exerted by T_regs from patients treated with placebo (n = 3, square) and glutamic acid decarboxylase (GAD)-alum (n = 4, circle) is shown as median for each group. (c) Suppression exerted by T_regs from type 1 diabetes (T1D) patients on autologous T_effs (red downward triangle; n = 7), by healthy T_regs on autologous T_effs (green circle), by T1D T_regs in co-culture with healthy T_effs (blue square) and by healthy T_regs in co-culture with T1D T_effs (purple upward triangle). The suppression is shown as percentage of decrease in T_eff proliferation in the presence of T_regs compared to T_effs cultured in the absence of T_regs. Healthy represents repeated measurements (n = 7) of one healthy reference. (d) Left panel illustrates suppression exerted by healthy T_regs in cross-culture with patient T_effs and the right panel shows suppression exerted by patient T_regs in cross-culture with healthy T_effs. Statistical significances are shown as P-values: *P < 0·05; **P < 0·01; ***P < 0·001. Data points indicate medians, error bars designate interquartile range.

To determine whether there was a difference in reduced T_reg-mediated suppression due to treatment, we tested if the suppression exerted in cross-over cultures of T1D T_regs versus healthy T_effs and healthy T_regs versus T1D T_effs was different between treatment arms. There was no difference in suppression exerted by T_regs from GAD-alum-treated patients compared to placebo T_regs in cross-over cultures with healthy T_effs, nor in the suppression exerted by healthy T_regs cultured with T_effs from GAD-alum-treated patients and T_effs from placebo subjects (Fig. 3d).

Discussion

Previous findings from the Phase II GAD-alum trial indicated an ability of GAD-alum to up-regulate CD25 and FoxP3 following in-vitro stimulation with GAD₆₅, in addition to increased secretion of several pro- and anti-inflammatory cytokines from PBMC 9,10. Preservation of C-peptide secretion was still present in a 4-year follow-up to the Phase II trial 11, and induction of a T cell subset with memory phenotype was observed upon GAD₆₅ stimulation 12.

Here we demonstrate that a great majority of lymphocytes in this T cell subset with memory phenotype expressed FoxP3 and high levels of CD25. Although some differences in the experimental setup were introduced in the present study, the main difference being that PBMC were cultured for 72 h at 21 and 30 months and for 7 days at the 4-year follow-up, the increased frequencies of CD25^hi and FoxP3⁺ cells detected in this 4-year follow-up of the study are in agreement with our previous findings at 21 and 30 months after treatment 9. In the present study, the CD127 and CD39 markers were included to further define T_regs. Both CD4⁺CD25^hiCD127^lo and CD4⁺CD25⁺CD127⁺ cells were expanded by GAD₆₅ stimulation, but a higher proportion of FSC^hiSSC^hi CD4⁺ cells were CD127⁺ than CD127^lo/–, suggesting that the frequency of T cells with both T_reg and activated-non-T_reg phenotype increased following GAD₆₅ stimulation.

Expression of CD39, an ectonucleotidase expressed on a subset of T_regs which hydrolyzes ATP into adenosine monophosphate (AMP) 23,29, was also increased upon antigen recall in GAD-alum-treated patients. It has been postulated that removal of proinflammatory ATP could be a suppressive mechanism mediated by CD39 on T_regs. In a recent study, CD39⁺ but not CD39^–CD4⁺CD25^hi cells were able to suppress IL-17 production 30. As the levels of IL-17 were undetectable in the supernatants of both expanded T_effs and T_eff/T_reg cultures, we cannot draw any conclusion on the ability of T_regs to suppress production of this cytokine in our settings. However, we have shown previously that secretion of IL-17, along with that of several other cytokines, was increased by GAD₆₅ stimulation in PBMC supernatants 12. Although the current study does not include healthy subjects, the expression of CD39 on resting CD4⁺CD25^hiCD127^lo cells detected by us in these T1D patients seems to be lower than what has been reported in healthy individuals by others using the same anti-CD39 clone and fluorochrome 30.

In line with previous findings 31, expanded CD25⁺CD127^lo T_regs were suppressive and retained their phenotype after expansion and cryopreservation. Although we were able to sort, expand and assess suppression in a limited number of individuals, there was no readily evident difference in the suppressive capacity of T_regs between placebo and GAD-alum-treated patients 4 years after administration of the treatment. Cross-over culture experiments revealed that T_regs isolated from patients with T1D participating in the GAD-alum trial had an impaired suppressive effect on autologous T_effs and also on T_effs from a healthy individual. In addition, T_regs from T1D patients were not capable of suppressing the proliferation of T_effs more readily from the healthy control than they did with their own T_effs. Further, T_effs from T1D patients were suppressed to a greater extent by T_regs from the healthy control than by their own T_regs. Taken together, these findings suggest that the reduced regulation observed in autologous co-cultures of cells isolated from T1D patients was due to reduced T_reg-mediated suppression intrinsic to the T_reg population. Our results are in contrast with previous findings, showing that responder T cells from T1D were more resistant to suppression 25,26. This could be explained by differences in the definition of cellular phenotypes and expansion conditions. While Schneider et al. used adaptive T_regs generated in vitro from CD4⁺CD25^– cells 25, the T_regs used by us in this study were expanded from the CD4⁺CD25^hiCD127^lo population. In the study by Lawson et al., sorted CD4⁺CD25^hi cells without in-vitro expansion from patients with long-standing T1D were used, and CD127 was not included to discriminate T_regs 26. Although we have identified a deficient T_reg-mediated suppression of polyclonal T cell stimulation in T1D patients who participated in the GAD-alum Phase II trial, treatment with GAD-alum did not affect the suppressive activity of T_regs. It should be kept in mind that samples included in the current study were drawn 4 years after treatment, and that an effect on suppression shortly after treatment cannot be excluded. Furthermore, due to the random selection of patients based on the availability of samples, none of the GAD-alum-treated patients classified as responders to treatment were included in suppression assays 10, and we were thus unable to relate suppression to clinical outcome. Because our assay measures suppression of polyclonal activation, an effect on the specific suppression in response to GAD₆₅ stimulation cannot be excluded. In fact, changes in the frequency of T cells with a T_reg phenotype during the trial have been observed only upon GAD₆₅ stimulation 9, while the frequency of T_regs after culture in medium alone has been similar in GAD-alum and placebo-treated patients throughout the study. Proliferative responses of PBMC from GAD-alum-treated patients in response to GAD₆₅ stimulation were significantly stronger compared to placebo in a thymidine incorporation assay, as we have reported previously 12, suggesting that the GAD₆₅-specific responses initiated by in-vitro antigen recall are not anergic.

In conclusion, we demonstrate GAD₆₅ recall-induced populations of CD4⁺CD25^hiCD127^lo T_regs as well as FSC^hiSSC^hiCD4⁺CD25⁺CD127⁺ activated T cells, detectable 4 years after treatment. A deficiency in T_reg-mediated suppression detected in T1D patients was intrinsic to the T_reg population, but was not affected by GAD-alum treatment. Overall, these results suggest that higher numbers of T_reg cells readily expandable upon GAD₆₅ stimulation, rather than increased T_reg function, may be responsible for the persistent clinical benefit of GAD-alum treatment after 4 years 11.

Disclosure

The authors declare that they have no conflicts of interest.