ORIGINAL RESEARCH published: 23 February 2022 doi: 10.3389/fimmu.2022.825426 Candidate Biomarkers for the Prediction and Monitoring of Partial Remission in Pediatric Type 1 Diabetes Laia Gomez-Muñoz 1, David Perna-Barrull 1, Josep M. Caroz-Armayones 2,3, Marta Murillo 4, Silvia Rodriguez-Fernandez 1, Aina Valls 4, Federico Vazquez 5, Jacobo Perez 6, Raquel Corripio 6, Luis Castaño 7, Joan Bel 4 and Marta Vives-Pi 1* 1 Immunology Department, Germans Trias i Pujol Research Institute and University Hospital, Autonomous University of Barcelona, Badalona, Spain, 2 Department of Political and Social Sciences, Health Inequalities Research Group (GREDS- EMCONET), Pompeu Fabra University, Barcelona, Spain, 3 Johns Hopkins University–Pompeu Fabra University Public Policy Edited by: Center, Barcelona, Spain, 4 Pediatrics Department, Germans Trias i Pujol Research Institute and University Hospital, Teresa Rodriguez-Calvo, Autonomous University of Barcelona, Badalona, Spain, 5 Endocrinology Department, Germans Trias i Pujol Research Institute Helmholtz Association of German and University Hospital, Autonomous University of Barcelona, Badalona, Spain, 6 Pediatric Endocrinology Department, Parc Research Centres (HZ), Germany Taulı´ Hospital Universitari, Institut d’Investigació i Innovació Parc Taulı´ I3PT, Autonomous University of Barcelona, Sabadell, Spain, 7 Cruces University Hospital, Biocruces Bizkaia Research Institute, UPV/EHU, CIBERDEM, CIBERER, Endo-ERN, Bilbao, Spain Reviewed by: Isabelle Serr, Helmholtz Association of German The partial remission (PR) phase, a period experienced by most patients with type 1 Research Centres (HZ), Germany diabetes (T1D) soon after diagnosis, is characterized by low insulin requirements and Laura M. Jacobsen, University of Florida, United States improved glycemic control. Given the great potential of this phase as a therapeutic *Correspondence: window for immunotherapies because of its association with immunoregulatory Marta Vives-Pi mechanisms and b-cell protection, our objective was to find peripheral immunological mvives@igtp.cat biomarkers for its better characterization, monitoring, and prediction. The longitudinal Specialty section: follow-up of 17 pediatric patients with new-onset T1D over one year revealed that, during This article was submitted to the PR phase, remitter patients show increased percentages of effector memory (EM) T Autoimmune and Autoinflammatory Disorders, lymphocytes, terminally differentiated EM T lymphocytes, and neutrophils in comparison a section of the journal to non-remitter patients. On the contrary, remitter patients showed lower percentages of Frontiers in Immunology naïve T lymphocytes, regulatory T cells (TREG), and dendritic cells (DCs). After a year of Received: 30 November 2021 follow-up, these patients also presented increased levels of regulatory B cells and Accepted: 31 January 2022 Published: 23 February 2022 transitional T1 B lymphocytes. On the other hand, although none of the analyzed Citation: cytokines (IL-2, IL-6, TGF-b1, IL-17A, and IL-10) could distinguish or predict remission, Gomez-Muñoz L, Perna-Barrull D, IL-17A was increased at T1D diagnosis in comparison to control subjects, and remitter Caroz-Armayones JM, Murillo M, Rodriguez-Fernandez S, patients tended to maintain lower levels of this cytokine than non-remitters. Therefore, Valls A, Vazquez F, Perez J, these potential monitoring immunological biomarkers of PR support that this stage is Corripio R, Castaño L, Bel J governed by both metabolic and immunological factors and suggest immunoregulatory and Vives-Pi M (2022) Candidate Biomarkers for the Prediction and attempts during this phase. Furthermore, since the percentage of TREG, monocytes, and Monitoring of Partial Remission in DCs, and the total daily insulin dose at diagnosis were found to be predictors of the PR Pediatric Type 1 Diabetes. Front. Immunol. 13:825426. phase, we next created an index-based predictive model comprising those immune cell doi: 10.3389/fimmu.2022.825426 percentages that could potentially predict remission at T1D onset. Although our Frontiers in Immunology | www.frontiersin.org 1 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission preliminary study needs further validation, these candidate biomarkers could be useful for the immunological characterization of the PR phase, the stratification of patients with better disease prognosis, and a more personalized therapeutic management. Keywords: type 1 diabetes (T1D), partial remission phase, honeymoon, biomarkers, pediatrics, prediction model, immune cell subpopulations, autoimmunity INTRODUCTION insulin production (12). At a clinical level, bicarbonate concentrations >15 mg/dl, age >5 years, male sex, higher body Type 1 diabetes (T1D) is an autoimmune T-cell-mediated mass index (BMI) values, lower HbA1c levels, and <3 diabetes- disease against pancreatic b-cells, which leads to the associated autoantibodies can predispose to PR in children and insufficient production of insulin and overt hyperglycemia (1). adolescents with new-onset T1D (13, 14). Furthermore, This chronic metabolic disease affects a growing number of mechanisms of immune regulation are non-linear along the children and adolescents (2), and because of the b-cell function natural history of T1D—matching well with the proposed decay, patients rely on exogenous insulin therapy for life, relapsing-remitting character of the disease course—and the which only delays the appearance of long-term secondary PR phase is associated with these immunomodulatory changes complications. Although its exact etiology and pathogenesis are (15, 16). Consistent with this idea, studies investigating changes still elusive, the development of T1D involves complex in immune parameters have reported that patients with the interactions between b-cells and innate and adaptive immune highest frequency of CD4 +CD25 +CD127hi lymphocytes at cells, namely, T and B lymphocytes, NK cells, monocytes, disease onset experience the longest PR (17, 18), that increased dendritic cells (DCs), and neutrophils (3). Different strategies levels of regulatory T, B, and NK cells can be found during T1D capable of stopping the autoimmune process and promoting b- progression (which could reflect attempts at restoring self- cell recovery have been developed in experimental models, but tolerance) (19, 20), and that the absence of IL-4, TNF-a, IL-10, none of them has achieved total remission in humans or and IL-13 in sera correlates with the length of the remission managed to prevent or cure the disease (4, 5). This failure period (21). Interestingly, peripheral antigen-specific regulatory could be partially explained by the lack of an optimal T cells (TREG) diminish during the honeymoon phase in checkpoint for immune interventions, and the insufficient comparison to the time of diagnosis (22). Nevertheless, few biomarkers for the proper stratification of patients, an unmet studies have compared patients with and without PR to find need in clinical trials (6, 7). specific and reliable biomarkers of this phase. As for this last point, shortly after diagnosis and the initiation Non-invasive approaches like the use of peripheral blood of insulin therapy, between 50 and 80% of patients with T1D samples are essential to identify biomarkers of T1D progression experience a transient partial remission (PR) period, also called and remission. Since many studies have shown that the the honeymoon phase, which is marked by low requirements of progression of T1D is associated with changes in immune exogenous insulin and diminished glycated hemoglobin parameters, but the longitudinal data including the analysis of (HbA1c) levels (8). This stage, characterized by metabolic and the PR phase is scarce (15, 20, 21, 23–27), the present study was immunological alterations, could constitute a unique window for aimed at analyzing the percentage and absolute counts of innate therapeutic interventions. This period can last from weeks to and adaptive immune cell subsets in peripheral blood and the years with an average of 9 months of duration (9), but in some concentration of different cytokines in plasma over one year after uncommon cases, a complete or long-lasting remission has been T1D diagnosis, differentiating patients with and without PR to described (10, 11). Patients with this improved glucose control find specific monitoring and predictive biomarkers of this stage. may be at a lower risk for the short- and long-term complications These biomarkers could be of great interest to characterize the of T1D, such as chronic microvascular complications. immunological mechanisms behind the PR phase and b-cell Unfortunately, the mechanisms underlying the PR phase protection, to monitor the early course of T1D, and to stratify are poorly characterized, but this natural phenomenon has patients with better disease prognosis at the onset of the disease for been attributed to both the reduction of glucose toxicity after their selection in clinical trials. Moreover, in the case of non- the initiation of the insulin therapy and the consequent b-cell remitters, these biomarkers would allow for a more personalized rest, recovery, and regeneration with improved endogenous therapeutic management to ensure the prevention of early hyperglycemia and to reduce the risk of secondary complications. Abbreviations: AUC, Area Under the Curve; BMI, Body Mass Index; BREG, Regulatory B Cells; CM, Central Memory; DCs, Dendritic Cells; EM, Effector Memory; GAD65, Glutamic Acid Decarboxylase 65; HbA1c, glycated hemoglobin; IA-2, Insulinoma Antigen-2; IDAA1c, Insulin-Dose Adjusted HbA1c; mDCs, MATERIALS AND METHODS Myeloid Dendritic Cells; pDCs, Plasmacytoid Dendritic Cells; PR, Partial Remission; ROC, Receiver Operating Characteristic; SDS, standard deviation score; T1D, Type 1 Diabetes; TEMRA, terminally differentiated effector memory Participants T; TREG, Regulatory T Cells; VIF, Variance Inflation Factors; ZnT8, Zinc For the longitudinal study of T1D progression and the Transporter 8. characterization of the PR phase, 17 pediatric patients with Frontiers in Immunology | www.frontiersin.org 2 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission new-onset T1D (of which 16 completed the study) and 17 age- each time-point. Fasting basal C-peptide was determined by and sex-matched non-diabetic control subjects were included. ELISA (Architect i2000, Abbott, IL, USA) in both controls and For the PR predictive model, 10 additional pediatric patients only patients at each time-point. Only at T1D onset, stimulated C- at T1D onset (n = 27) were selected. peptide was measured 6 min after i.v. administration of 1 All patients fulfilled the American Diabetes Association mg glucagon. classification criteria for T1D (28), with at least one positive anti-islet autoantibody at disease onset [to insulinoma-antigen 2 (IA-2), glutamic acid decarboxylase 65 (GAD65), or zinc Flow Cytometry transporter 8 (ZnT8)]. Inclusion criteria were 4–18 years of age Phenotypic analysis of cellular subpopulations was performed in and normal BMI according to the Spanish BMI pediatric cohort control subjects and patients with T1D at each time-point. To growth chart (29). Exclusion criteria were being under that end, fresh whole blood samples of 1–2 ml were washed with immunosuppressive or anti-inflammatory treatment, type 2 15 ml of FACSFlow Sheath Fluid (ThermoFisher Scientific, diabetes, pregnancy, compromised kidney function, or liver Waltham, MA, USA), and 100 ml were stained with specific diseases. Also, candidates were excluded if they presented mAbs for 20 min at room temperature and protected from light. moderate/severe symptomatic infections or fever in the The panels of antibodies (Supplemental Tables 1, 2) were built previous 2–4 weeks before blood withdrawal (i.e., Flu or as follows: (1) T lymphocyte maturation stages panel: CD3- Covid-19). V500, CD4-PerCPCy5.5, CD8-APCH7, CD45RA-FITC, PTK7- PE, CCR7-PECy7, CD31-AF647, CD27-BV421, (2) TREG panel: CD45-FITC, CD3-V450, CD4-PerCPCy5.5, CD25-PE, CCR4- Sample Collection, Longitudinal T1D PECy7, CD127-AF647, CD45RO-APCH7, HLA-DR-V500, (3) Follow-Up, and PR TH17 lymphocytes panel: CD4-V450, CCR6-PE, CCR7-PECy7, Longitudinal T1D data collection occurred over one year in two and CCR4-AF647, (4) TCR panel: CD3-PerCP, gd TCR-PE, ab University Hospitals. Blood samples of 6 ml were obtained at TCR-FITC, CD8-APCH7, CD4-V450, (5) B lymphocyte three different time-points throughout T1D progression of each maturation stages panel: CD3-V450, CD19-V500, CD27-APC, patient: at disease onset (n = 17), at PR (n = 11) or 8 months for CD21-PE, IgD-FITC, IgM-PerCPCy5.5, (6) B lymphocyte non-remitter patients (n = 6), and at 12 months after disease subpopulations panel: CD19-PerCPCy5.5, CD24-FITC, CD38- onset (n = 16) in EDTA tubes (BD Biosciences, San Jose, CA, PE, CD27-APC, and (7) Innate cells panel: CD45-AF700, CD3- USA) and processed within 6 h. Control samples of 6 ml of blood APCH7, CD19-APCH7, CD14-V450, CD16-APC, CD11c- from non-diabetic subjects (sex and age-related, and with the PECy7, CD123-PerCPCy5.5, CD56-PE, HLA-DR-V500, Slan- same exclusion criteria) were acquired simultaneously to the FITC. After incubation, erythrocytes were lysed for 7 min diagnosis of children with T1D. In addition, non-longitudinal (Lysing Buffer, BD Biosciences). Samples were then washed blood samples of 3 ml from 10 patients only at T1D onset (6 and resuspended in FACSFlow Sheath Fluid (ThermoFisher future remitters, 4 future non-remitters) were obtained for the Scientific). Absolute counts (cells/ml) of leukocytes and generation of the PR predictive model and processed as above. At lymphocytes were obtained using Perfect Count Microspheres disease onset, all samples (n = 27) were collected between 1 and of known concentration (Cytognos SL, Salamanca, Spain). 14 days after diagnosis. After the monthly medical visit, patients A minimum of 10,000 events per sample and 5,000 beads were considered to be in PR when they fulfilled the accepted were acquired using a 3-laser FACS Canto II and a 4-laser LSR criteria of ≤9 insulin dose-adjusted HbA1c (IDAA1c), an index Fortessa Flow Cytometers (BD Biosciences) and analyzed using that is calculated as HbA1c (%) + [4 × insulin dose (U/kg/day)] FACSDiva software (BD Biosciences). Necrotic and apoptotic (30). This phase was identified between 2 and 6 months after cells were excluded from the analysis based on their FSC-A/SSC- diagnosis, and individuals who did not meet the criteria of PR A properties and doublets were excluded by FSC-A/FSC-H. The after 8 months, were defined as non-remitters. gating strategy to analyze specific leukocyte subsets was based on international consensus (31). Fluorescence minus one controls were used to define CCR7 expression on TH17 lymphocytes and Clinical and Laboratory Testing PTK7 expression on recent thymic emigrants. Furthermore, Clinical descriptors on each patient and control subject were internal reference populations were used as positivity controls collected, namely, age, sex, and BMI. BMI data were also in the analysis of CCR7 vs. CD45RA in panel 1, CD27 vs. CD24 expressed as standard deviation score (SDS) for age and sex, in panel 2, CD19 vs. CD21 in panel 3, CCR4 vs. CCR7 in panel 4, based on the Spanish BMI pediatric cohort growth chart data and CD45RO vs. CCR4 plus CD45RO vs. HLA-DR in panel 5. (29). Blood samples from patients with T1D were acquired for Absolute counts were calculated as follows: (%subset/100) × centralized measurement of HbA1c, fasting basal and stimulated counts of the main subpopulation. C-peptide, genetics, and immunology; and insulin requirements were recorded. At the time of T1D diagnosis, HLA typing of DRB1 alleles and islet autoantibodies to IA-2, GAD65, and ZnT8 Serum Cytokine Quantification were determined as previously reported (20). HbA1c was Plasma samples were obtained after centrifuging 4–5 ml of whole determined by high-performance liquid chromatography blood samples twice at 4°C within the first hour after (ADAMS A1c HA-8180V, Arkray, MN, USA) in all patients at venipuncture (1,900 G for 10 min; 16,000 G for 10 min) and Frontiers in Immunology | www.frontiersin.org 3 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission stored at −80°C until use. The BD Cytometric Bead Array (CBA) 12 months after baseline Data presented as mean ± SD, or mean (min, max). P-value calculated from Mann–Whitney test when comparing control subjects and patients with T1D (++P ≤0.01, +++P ≤0.001 and ++++P ≤0.0001), and remitter and non-remitter groups [xP ≤0.05, xxP ≤0.01 and xxxP ≤0.001 for comparison with PR; ^P ≤0.05, ^^P ≤0.01 and ^^^P ≤0.001 for comparison with 8 months after baseline (non-remitters); and °°P ≤0.01 for comparison with 12 months after baseline (remitters)]. P- value calculated from Wilcoxon test for comparisons of paired data with disease onset (*P ≤0.05 and **P ≤0.01). BMI, Body Mass Index; F, female; HbA1c, glycated hemoglobin; IDAA1c, insulin dose-adjusted HbA1c; M, male; NA, not Human Enhanced Sensitivity Flex Sets for IL-2, IL-6, IL-10, and 0.15 ± 0.06+++/xx 11.3 ± 1.2*/xxx/°° −0.5 (−1.0, 0.2) 0.85 ± 0.1xxx/°° IL-17A (BD Biosciences) were used to measure cytokine 63 ± 10.1*/x 7.9 ± 0.9*/x Non-remitter T1D patients (n = 6) 9.2 ± 2.8* concentration (detection range 274–200,000 fg/ml). Samples 17 ± 1.9 2/4 ND were acquired on an LSR Fortessa flow cytometer (BD Biosciences), and data were analyzed using CBA software. The Human TGF-b1 ELISA Kit (FineTest, Wuhan Fine Biotech, Wuhan, China) was used for the quantification of TGF-b1 (detection range 31.25–2,000 pg/ml; sensitivity 18.75 pg/ml). 8 months after baseline −0.4 (−1.1, 0.3) Statistical Analysis 8.1 ± 0.7*/xx 64.5 ± 7.4*/xx 0.3 ± 0.1+++ 0.9 ± 0.1xxx 11.5 ± 1*/xxx 9 ± 2.8 17.2 ± 2.1 Data were tested for normal distribution with the Shapiro–Wilk 2/4 ND test and are presented as mean ± SD, mean (min, max), or median and interquartile range, unless stated otherwise. For clinical parameters, a descriptive statistical analysis of the variables was performed. c² test was used for categorical variables. Differences between two groups were analyzed using the nonparametric 2- 12 months after baseline tailed Wilcoxon test for paired data and the nonparametric 2-tailed −0.1 (−1.0, 1.6) 9.2 ± 1.3**/^^ 0.3 ± 0.3++++ Mann–Whitney test for unpaired data. To test differences among 10.2 ± 4.3**,xx 7.1 ± 0.8**/^ 54.2 ± 9.1**/^ 0.5 ± 0.2^^^ longitudinal data, a mixed effects model was fit with condition as 18.6 ± 3.5* 4/6 ND fixed effect, and individual and residual random variation as Remitter T1D patients (n = 11) random effect covariates together with the Tukey’s multiple comparisons test. Kruskal–Wallis test followed by Dunn’s multiple comparisons test was used to compare data between control subjects and each time-point of T1D progression. To PR (2–6 months after baseline) TABLE 1 | Clinical features and metabolic data of control subjects and patients with T1D included in the longitudinal study. establish associations between immunological, clinical, and metabolic variables and the occurrence of PR, simple logistic −0.2 (−1.2, 1.1)* regressions were developed. Samples with missing data were 0.7 ± 0.5++ 6.9 ± 0.6** 51.5 ± 6.4** 0.4 ± 0.1** 8.4 ± 0.5** 9.1 ± 4.3 removed. The OR with the 95% CI was reported as a probability 17.7 ± 3* 5/6 ND measure. Receiver-operating characteristic (ROC) curves were generated from the logistic regressions to obtain the area under the curve (AUC), which is useful to assess the discrimination ability of the variables. G-test of goodness-of-fit was also employed. To generate a full model containing simultaneously the statistically significant immunological variables, an index was created from the T1D onset (n = 17) best cut-off values of the ROC curves of each variable. Simple and 0.3 ± 0.2++++ −0.5 (−1.6, 0.7) 101.3 ± 26.2 Baseline multiple logistic regressions (adjusted by BMI-SDS) were generated 8.7 ± 3.6 16.8 ± 2.5 11.4 ± 2.4 0.7 ± 0.2 0.6 ± 0.4 7/10 14.3 ± 3 to assess the association between the index and the event of PR. The relationship between covariates was determined with linear regressions. To check the robustness of the results, multicollinearity was examined by the variance inflation factors (VIF), R2 with other Control subjects (n = 17) variables, and Spearman’s correlations between the selected variables. applicable; ND, not determined; SDS, standard deviation score. To find statistically significant correlations between different 0.04 (−1.7, 2.7) parameters, a two-tailed Spearman’s or Pearson’s test was used. All 8.8 ± 3.4 18.3 ± 4.3 1.3 ± 0.4 Baseline 7/10 analyses were performed with GraphPad Prism 9.0 (GraphPad NA NA NA NA NA Software Inc, San Diego, CA, USA) and Stata v16.0 (Stata Corporation, College Station, TX, USA). A P-value of ≤0.05 was considered statistically significant. Stimulated C-peptide (ng/ml) RESULTS Basal C-peptide (ng/ml) Insulin dose (U/Kg/day) HbA1c (mmol/mol) Clinical Features and Metabolic Data of Pediatric Patients With T1D BMI (kg/m2) Age (years) HbA1c (%) Clinical features and metabolic data from control subjects and Sex (M/F) BMI-SDS IDAA1c patients with T1D at each time-point for the longitudinal study are summarized in Table 1. No statistically significant differences were Frontiers in Immunology | www.frontiersin.org 4 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission found in age and BMI when compared between control subjects of the maturation stages of both CD4+ and CD8+ T lymphocytes and patients at T1D onset; as expected, they were found in terms of (naïve, central memory (CM), effector memory (EM), and plasma C-peptide concentration (1.3 ± 0.4 vs 0.3 ± 0.2; P <0.0001). terminally differentiated effector memory T (T E M R A ) The 64.7% of patients with T1D presented PR between 2 and lymphocytes), revealed statistically significant changes in blood 6 months after the diagnosis, with an IDAA1c value equal to or from patients during the PR phase (Figure 2). less than 9, while the rest did not experience it after 12 months of Regarding CD4+ T lymphocytes, non-remitter patients follow-up. During the PR phase, and in comparison to disease presented higher percentages of naïve CD4+ T lymphocytes onset, patients showed increased BMI and BMI-SDS values, and both at 8 and 12 months in comparison to patients at disease a non-significant increase in basal C-peptide concentration (0.7 ± onset (P ≤0.05) (Figure 2A). While the percentage of CM CD4+ 0.5 vs 0.3 ± 0.2; ns), although the levels were still much lower than T lymphocytes tended to increase with disease progression for those of the control group (0.7 ± 0.5 vs 1.3 ± 0.4; P <0.01). After a the remitter ones—especially at 12 months after T1D onset in year follow-up, 50% of these remitter patients were still in comparison to the diagnosis and the control group (P <0.001)— remission. There were no significant differences in terms of age, the percentage of this subpopulation tended to decrease at 12 sex, and BMI between patients with and without PR at 8 months of months for non-remitter patients in comparison to the time- follow-up. Unlike patients in remission, non-remitter patients point of 8 months, where levels were also higher than those (IDAA1c >9) did not present an increase in basal C-peptide of the control group (P ≤0.05). However, no significant concentration after diagnosis (0.3 ± 0.1 vs 0.3 ± 0.2; ns), differences were observed between remitters and non-remitters indicating some b-cell recovery only in remitter patients. at 8 months (Figure 2B). After 12 months of follow-up, the group of remitter patients The main findings observed during the PR phase when showed lower levels of basal C-peptide than levels during compared with non-remitter patients were higher percentages remission; however, these doubled those of the non-remitters of EM and early EM CD4+ T lymphocytes and CD8+ TEMRA (0.3 ± 0.3 vs 0.15 ± 0.06; ns). Moreover, the group of remitter lymphocytes (Figures 2D, E, H). The percentage of CD4+ TEMRA patients still needed a lower daily dose of total insulin than the lymphocytes decreased at 12 months for remitter patients in group of non-remitter patients (0.5 ± 0.2 vs 0.85 ± 0.1; P <0.01) comparison to controls (P ≤0.05), but tended to be higher during and maintained the increased BMI values regarding the disease the PR phase in comparison to non-remitters at 8 months onset (18.6 ± 3.5 vs 16.8 ± 2.5; P ≤0.05). At 12 months, both (Figure 2C). Peripheral EM and early EM CD4 + T remitters and non-remitters maintained practically equal the lymphocytes behaved similarly (Figures 2D, E). The HbA1c values reached during the PR phase or at 8 months from percentages of both subpopulations were significantly increased diagnosis, respectively. during the PR phase in comparison to 8 months (P ≤0.05 and Clinical features and metabolic data from patients with T1D P <0.01, respectively) and 12 months (P ≤0.05) for non-remitter at disease onset for the PR predictive study are summarized in patients. The latter ones at 8 months also showed decreased Table 2. Differences at T1D onset between remitters and non- percentages in comparison to disease onset (P <0.01) and control remitters were found in terms of insulin dose (0.6 ± 0.2 vs 0.8 ± subjects (P ≤0.05 and P = 0.06, respectively), and at 12 months 0.1, P <0.01) and stimulated C-peptide (0.9 ± 0.8 vs 0.5 ± 0.5, than patients at diagnosis (P ≤0.05 and P <0.01, respectively). P ≤0.05). Concerning the absolute counts, only CD4 + T EMRA lymphocytes significantly increased during the PR phase in comparison to 8 months for non-remitters. No other Increase in Effector Memory CD4+ and significant difference was observed between patients with and CD8+ T Lymphocytes at the PR Phase without PR (Supplemental Figures 1A–E). Figure 1 shows the gating strategy followed for the antibody Regarding CD8+ T lymphocytes, patients at the PR phase panels of T lymphocyte maturation stages and TREG. The analysis presented lower percentages of naïve CD8+ T lymphocytes TABLE 2 | Clinical features and metabolic data of patients with T1D included in the PR predictive model study at disease onset separated by future remitters and non-remitters. T1D onset (future PR) (n = 17) T1D onset (future non-PR) (n = 10) P-value Age (years) 9.8 ± 5 6.4 ± 3.4 0.07 Sex (M/F) 8/9 4/6 0.72 BMI (kg/m2) 17.7 ± 3.0 15.7 ± 1.7 0.12 BMI-SDS −0.5 (−1.8, 0.7) −0.7 (−1.7, 0.5) 0.54 HbA1c (%) 11.1 ± 2.5 11.4 ± 1.7 0.61 HbA1c (mmol/mol) 98.1 ± 27.5 100.6 ± 18.6 0.61 Insulin dose (U/Kg/day) 0.6 ± 0.2 0.8 ± 0.1 0.01** Basal C-peptide (ng/ml) 0.4 ± 0.3 0.3 ± 0.2 0.15 Stimulated C-peptide (ng/ml) 0.9 ± 0.8 0.5 ± 0.5 0.03* IDAA1c 14.5 ± 3.2 14.7 ± 2.0 0.98 Data presented as mean ± SD, or mean (min, max). P-value calculated from Mann–Whitney test when comparing baseline time-points (T1D onset) between remitters and non-remitters (*P ≤0.05 and **P ≤0.01). BMI, Body Mass Index; F, female; HbA1c, glycated hemoglobin; IDAA1c, insulin dose-adjusted HbA1c; M, male; SDS, standard deviation score. Frontiers in Immunology | www.frontiersin.org 5 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission FIGURE 1 | Representative gating strategy for the T lymphocyte maturation stages and TREG panels. Gating strategy used to analyze the percentages of the different CD4+ and CD8+ T cell subsets based on the expression of the markers CD3, CD4, CD8, CD45RA and CCR7, and of CD127, CD25, CD45RO, CCR4 and HLA-DR for TREG subsets within CD4+ T cells. CD31 and PTK7 were used to characterize RTEs. Positivity controls for CCR7 vs CD45RA using lymphocytes, for CD45RO vs CCR4 using CD4- T cells and for CD45RO vs HLA-DR using CD4+ T cells as internal reference populations. FMO control for PTK7 was used. compared to patients without PR at 8 months (P ≤0.05). significant increase at disease onset compared to the control Moreover, non-remitter patients showed a higher percentage of group (P <0.01) (Figure 2G). As mentioned before, the this subpopulation than controls (P <0.01) (Figure 2F). Contrary percentage of CD8+ T EMRA lymphocytes was lower at 8 to what was observed in the CM CD4 + T lymphocyte months for non-remitter patients in comparison to patients at subpopulation, the percentage of CM CD8+ T lymphocytes the PR phase (P ≤0.05) and controls (P ≤0.05) (Figure 2H). tended to decrease with T1D progression and showed a Lastly, and regarding the percentage of EM and early EM CD8+ T Frontiers in Immunology | www.frontiersin.org 6 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission A B C D E F G H I J FIGURE 2 | Percentages of CD4+ and CD8+ T lymphocyte subsets are altered at the initial stages of T1D. For CD4+ and CD8+ T lymphocytes, respectively, the percentages (%) of (A, F) naïve T lymphocytes, (B, G) CM T lymphocytes, (C, H) TEMRA lymphocytes, (D, I) EM T lymphocytes, and (E, J) early EM T lymphocytes were determined in peripheral blood of control subjects and patients with T1D at different time-points. Squares represent controls (n = 17) (white bar), and patients are represented by light gray dots at disease onset (n = 17) (gray bar), white dots during PR (n = 11) and for remitter patients at 12 months (12 M PR) (n = 10) (blue bars), and dark gray dots for non-remitter patients at 8 months (8 M no PR) (n = 6) and 12 months (12 M no PR) (n = 6) (pink bars). Bar graphs show mean percentage values. Each symbol represents an individual patient. Lines link the same patient throughout the time-points. *P ≤ 0.05, **P < 0.01, ***P < 0.001 after mixed effects model with Tukey’s post-hoc test for longitudinal data, Kruskal–Wallis with Dunn’s post-hoc test for comparisons between control subjects and the different T1D time-points, or 2-tailed Mann–Whitney test for comparisons between two unpaired groups of data. P ≤0.05 is considered significant. lymphocytes, no differences were found between groups of follow-up (P ≤0.05) and controls (P ≤0.05), but no differences (Figures 2I, J). between remitters and non-remitters were found regarding their Concerning the absolute counts of these subsets, no absolute numbers (Supplemental Figures 2A, B). Concerning significant differences were found between patients with and the absolute counts and percentages of total CD8 + T without PR (Supplemental Figures 1F–J). lymphocytes, late EM CD4+ and CD8+ T lymphocytes, recent As for total CD4+ T lymphocytes, their percentage increased thymic emigrants, gd TCR cells, and TH17 lymphocytes, no at 8 months for non-remitter patients when compared to patients significant differences were found between patients with and during the PR phase (P <0.01), to remitter patients at 12 months without PR (Supplemental Figures 2C–N). Gating strategies for Frontiers in Immunology | www.frontiersin.org 7 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission the antibody panels of TCR and TH17 lymphocytes are depicted Increase in Regulatory B Lymphocyte in Supplemental Figures 3, 4. Subpopulations After One-Year Follow-Up for Patients That Experienced the Decrease in Total and Memory TREG at PR Phase the PR Phase Because of the described role for B lymphocytes in T1D Since TREG are expected to be implicated in the immunoregulatory pathogenesis, we next examined different naïve and memory B mechanism behind the PR phase (32), we analyzed peripheral lymphocyte subsets. Figure 4 shows the gating strategy followed TREG subsets, namely, total, memory, and activated TREG. for the antibody panel of B lymphocyte subpopulations. Furthermore, representative confirmatory staining for FoxP3 Interestingly, two B lymphocyte subsets with regulatory within the TREG population can be found in Supplemental functions—transitional T1 B lymphocytes and regulatory B Figure 5, where 74.6% of these cells do express it. cells (BREG)—were increased at 12 months for remitter patients The main findings observed in patients with T1D at the PR (Figure 5). The percentage of total transitional B lymphocytes phase when compared with patients without the PR were lower was substantially decreased at disease onset when compared to percentages of both total and memory TREG (Figures 3A, C). the control group (P ≤0.05), and their levels tended to recover Specifically, the percentage of total TREG increased at 8 months with time (Figure 5A). Within this subpopulation, transitional from diagnosis for patients without PR in comparison with both T1 B lymphocytes increased in percentage at 12 months for the control group (P ≤0.05) and patients at disease onset (P ≤0.05). patients that experienced the PR phase when compared to the Furthermore, levels at this time-point are higher than patients in diagnosis (P <0.01), while transitional T2 B lymphocytes PR, both during that phase (P ≤0.05) and 12 months after decreased (P <0.01) (Figures 5B, C). Moreover, at this time- diagnosis (P ≤0.05). After a year follow-up, patients who did not point, patients presented a higher T1/T2 ratio than patients at present PR continued to show a higher percentage of total TREG T1D onset (P <0.01) (Figure 5D). than patients who presented remission (P ≤0.05) (Figure 3A). The gating strategy followed for the antibody panel of B Regarding their absolute numbers, total TREG increased with the lymphocyte maturation stages is depicted in Supplemental disease condition in comparison to controls, finding statistically Figure 7. Concerning the percentages and absolute counts of significant differences at 8 months for patients without PR total B lymphocytes, naïve B lymphocytes, CD21−/low naïve B (P ≤0.05), and at 12 months for patients that experienced the lymphocytes, mature naïve B lymphocytes, memory B PR phase (P <0.01), whose levels were also higher than those of lymphocytes—namely, exhausted, unswitched, switched, and patients at disease onset (P <0.01) (Figure 3B). IgM memory B lymphocytes—(Supplemental Figures 8, 9) Concerning memory TREG, they increased in percentage and plasmablasts (Supplemental Figure 10), no differences compared to the control group at 8 months for patients without were found between patients with and without PR. remission (P <0.01), and at 12 months for patients who experienced PR (P ≤0.05). As mentioned above, patients without remission at 8 months present a higher percentage of memory TREG than patients Innate Immune Cells Are Altered at the during PR (P ≤0.05) (Figure 3C). Their absolute numbers behaved Initial Stages of T1D and During the PR similarly, increasing with the disease condition both at 12 months Phase after diagnosis for remitter patients (P <0.001) and at 8 months for We have previously shown the potential that subsets of NK cells non-remitters (P ≤0.05) when compared to controls. Also, their have as biomarkers of T1D progression and PR (20). Thus, numbers were higher at 12 months for remitter patients than at changes in neutrophils and different subsets of monocytes and T1D onset (P ≤0.05) (Figure 3D). DCs were investigated. Figure 6 shows the gating strategy As for activated TREG, no differences were found between followed for the antibody panel of innate cells. Although no patients with and without PR (Supplemental Figure 6). significant differences were found between groups regarding the A different tendency on the percentage of TREG from disease percentage of total monocytes (Figure 7A), we found an increase onset (baseline) to PR was observed for 4 children, who increased in the percentage of classical CD16− monocytes during the PR this cell subpopulation in contrast to the other remitter patients, phase in comparison to disease onset (P ≤0.05), and a subsequent and as non-remitter patients did (Figure 3E). Since those 4 decrease in the percentage of non-classical CD16+ monocytes children in PR present the same behavior as the non-remitters at (P ≤0.05) (Figures 7B, C). 8 months, differences between the two groups of remitter The percentage of total DCs increased after 8 months from patients were investigated. Interestingly, these 4 children with diagnosis for non-remitter patients in comparison to patients at increasing percentages of TREG from baseline were all males, and PR (P ≤0.05) and to patients at disease onset (P ≤0.05). Those they were the youngest (mean 5.5 vs 11.1 years, P ≤0.05). significant increases can be also observed 12 months after Moreover, they presented lower basal C-peptide concentrations diagnosis for remitter patients (Figure 7D). When analyzing (mean 0.28 vs 1.03, P ≤0.05) and BMI values (mean 15.6 vs 18.97, the two main subsets of DCs, the myeloid (mDCs) and the P ≤0.05), required lower doses of insulin (mean 0.28 vs 0.44, plasmacytoid (pDCs) ones, no significant differences were found P ≤0.05), and presented a tendency of a higher percentage of between groups regarding mDCs (Figure 7E), but pDCs HbA1c in comparison to the other 7 remitters with decreasing increased at 12 months for non-remitter patients when percentages of TREG from baseline (Supplemental Table 3). compared to diagnosis (P ≤0.05) (Figure 7F). Frontiers in Immunology | www.frontiersin.org 8 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission A B C D E FIGURE 3 | Peripheral blood TREG are quantitatively altered at the initial stages of T1D. Percentages (%) and concentrations (cells/µl) of (A, B) TREG and of (C, D) memory TREG were determined in peripheral blood of controls and patients with T1D at different time-points. Squares represent controls (n = 17) (white bar), and patients are represented by light gray dots at disease onset (n = 16) (gray bar), white dots during PR (n = 11) and for remitter patients at 12 months (12 M PR) (n = 10) (blue bars), and dark gray dots for non-remitter patients at 8 months (8 M no PR) (n = 6) and 12 months (12 M no PR) (n = 6) (pink bars). (E) Representative plots for the difference in the TREG percentage between PR and non-PR (8 M), and changes in this cell population from the disease onset to the PR phase or non-PR (8 M). Bar graphs show mean percentage or absolute count values. Each symbol represents an individual patient. Lines link the same patient throughout the time- points. *P ≤0.05, **P < 0.01, ***P < 0.001 after mixed effects model with Tukey’s post-hoc test for longitudinal data, Kruskal–Wallis with Dunn’s post-hoc test for comparisons between control subjects and the different T1D time-points, or 2-tailed Mann–Whitney test for comparisons between two unpaired groups of data. P ≤0.05 is considered significant. Frontiers in Immunology | www.frontiersin.org 9 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission FIGURE 4 | Representative gating strategy for the B cell subpopulations panel. Gating strategy used to analyze the percentages of transitional B cells (total, T1, and T2), BREG and plasmablasts based on the expression of the markers CD19, CD27, CD38, and CD24. Positivity control for CD27 vs CD24 using lymphocytes as an internal reference population. Concerning the absolute numbers of monocytes and their P ≤0.05, respectively; for absolute numbers, P <0.01 and P ≤0.05, subsets and mDCs and pDCs, no differences were found between respectively). For 4 out of 10 remitter patients after a year of follow- patients with and without PR (Supplemental Figures 11A–C, up, there was also a decrease in the percentage of CD16hi E, F). However, one year after onset, non-remitters presented neutrophils when compared to controls (P ≤0.05) (Figures 7G, H). lower counts of total DCs than remitters (P ≤0.05) (Supplemental Figure 11D). When further subdividing the mDC subset Increase in the Concentration of Cytokines according to the expression of Slan and CD16, no differences at T1D Onset were found between patients with and without PR (Supplemental Different cytokines involved in T1D, namely, IL-2, IL-6, TGF-b1, Figures 12A–F). IL-17A, and IL-10 were prospectively analyzed in the plasma of Finally, we detected that the percentage and counts of CD16hi patients as potential biomarkers of PR (Figure 8). Although no neutrophils remained constant between the control group, patients statistically significant differences were found between patients at disease onset, and patients at PR. Interestingly, the percentage and with and without PR, a tendency for higher levels of all cytokines absolute numbers of CD16+ neutrophils decreased in 3 out of 6 non- at T1D onset compared to controls was observed. No differences remitter patients both at 8 months and 12 months from diagnosis in were noticed regarding the concentration of IL-2 between the comparison to patients at the PR phase (for percentage, P = 0.07 and different groups (Figure 8A). Concerning IL-6, patients at Frontiers in Immunology | www.frontiersin.org 10 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission A B C D E F FIGURE 5 | Two regulatory subsets of B lymphocytes are quantitatively altered at the initial stages of T1D. Percentages (%) of (A) total transitional B lymphocytes and of their subdivision into (B) T1, and (C) T2, (D) the T1/T2 ratio, and (E, F) the percentage (%) and concentration (cells/µl) of BREG were determined in peripheral blood of controls and patients with T1D at different time-points. Squares represent controls (n = 17) (white bar), and patients are represented by light gray dots at disease onset (n = 16) (gray bar), white dots during PR phase (n = 11) and for remitter patients at 12 months (12 M PR) (n = 10) (blue bars), and dark gray dots for non-remitter patients at 8 months (8 M no PR) (n = 6) and 12 months (12 M no PR) (n = 6) (pink bars). Bar graphs show mean percentage or absolute count values. Each symbol represents an individual patient. Lines link the same patient throughout the time-points. *P ≤ 0.05, **P < 0.01 after mixed effects model with Tukey’s post-hoc test for longitudinal data, or Kruskal–Wallis with Dunn’s post-hoc test for comparisons between control subjects and the different T1D time-points. P ≤0.05 is considered significant. disease onset presented a significant increase in comparison to showed higher IL-17A levels than the control group (P ≤0.05) non-remitters at 12 months (P ≤0.05) (Figure 8B). Moreover, (Figure 8D). Finally, IL-10 remained constant throughout the the higher concentration of TGF-b1 found at diagnosis tended to different groups (Figure 8E). decrease with disease progression, especially during the PR phase and at 12 months for patients without PR (P ≤0.05) (Figure 8C). Percentages of TREG, DCs, and Regarding IL-17A, a statistically significant increase at T1D onset Monocytes, and the Total Daily Insulin was found when compared to controls (P <0.01). Interestingly, Dose at T1D Onset Could Serve as remitter patients tended to maintain lower levels of this cytokine Predictive Biomarkers of PR in comparison to non-remitter patients, even at a time-point of We next tested the predictive capacity of the PR phase of all the 12 months. After a year follow-up, non-remitter patients also immune cell subsets, cytokines, and metabolic and clinical Frontiers in Immunology | www.frontiersin.org 11 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission FIGURE 6 | Representative gating strategy for the innate cells panel. Gating strategy used to analyze the percentages of DCs, monocytes, and neutrophils. Cells were first gated for singlets, PBMCs, and CD3-CD19- cells. With the use of CD56 and CD14, cells were divided into NK cells and monocytes, which were both further divided into their subsets with the use of CD16, and DCs (lineage negative and HLA-DR+). Different subsets of DCs (mDC and pDC) were analyzed using CD123 and CD11c, and mDCs were further divided into their subsets with the use of CD16 and Slan. parameters analyzed at T1D onset. First, simple logistic 0.46 to 0.94, P ≤0.05), and DCs (OR = 1.39, 95% CI: 1.05 to 2.07, regression analyses were used to determine the association P ≤0.05) did, meaning that having increasing percentages of between the event of PR and each of the parameters (Table 3). TREG and monocytes would correspond with lower odds of being While none of the different cytokine concentrations in plasma in remission. Contrarily, the likelihood of being in PR is 39% could predict the PR event, the percentage of TREG (OR = 0.45, higher for each increase in the percentile of DCs. As for the 95% CI: 0.17 to 0.83, P ≤0.05), monocytes (OR = 0.71, 95% CI: metabolic parameters, only the insulin dose could predict the Frontiers in Immunology | www.frontiersin.org 12 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission A B C D E F G H FIGURE 7 | Innate cells are quantitatively altered at the initial stages of T1D. Percentages (%) of (A) monocytes and their subsets (B) classical, and (C) non- classical, of (D) DCs and their subsets (E) mDCs, and (F) pDCs, and (G, H) the percentage (%) and concentration (cells/µl) of CD16hi neutrophils were determined in peripheral blood of control subjects and patients with T1D at different time-points. Squares represent controls (n ≥16) (white bar), and patients are represented by light gray dots at disease onset (n = 17) (gray bar), white dots during PR (n = 11) and for remitter patients at 12 months (12 M PR) (n = 10) (blue bars), and dark gray dots for non-remitter patients at 8 months (8 M no PR) (n = 6) and 12 months (12 M no PR) (n = 6) (pink bars). Bar graphs show mean percentage or absolute count values. Each symbol represents an individual patient. Lines link the same patient throughout the time-points. *P ≤ 0.05, **P < 0.01 after mixed effects model with Tukey’s post-hoc test for longitudinal data, Kruskal–Wallis with Dunn’s post-hoc test for comparisons between control subjects and the different T1D time-points, or 2-tailed Mann–Whitney test for comparisons between two unpaired groups of data. P ≤ 0.05 is considered significant. event of PR (OR = 7.3e−4, 95% CI: 4.18e−007 to 0.16, P ≤0.05), dose, and immunological variables) was determined with linear meaning that the requirement of higher doses of exogenous regressions, finding that none of the parameters were predictive of insulin would correspond with lower odds of being in each other, except for insulin dose and TREG percentage remission. In fact, these cell percentages and insulin dose values (Supplemental Table 4). Although the results were not at T1D onset, separated by remitters and non-remitters, showed significant, age and BMI at T1D onset could also serve as how those remitter patients were diagnosed with lower potential predictors of PR, with a 20% increase in the odds of percentages of TREG (P <0.01) and monocytes (P ≤0.05), a being in PR for each year added (P = 0.08) and with a 42% increase higher percentage of DCs (P ≤0.05), and lower doses of in the odds for each BMI unit increase (P = 0.08). No other exogenous insulin (P <0.01) compared to those non-remitter metabolic parameter or cell subpopulation was able to predict the patients (Figures 9A–D). Moreover, the G-test was 6.982 for event of PR (Table 3, Supplemental Table 5). TREG (P <0.01), 6.152 for monocytes (P ≤0.05), 5.554 for DCs Figures 9E–H show the ROC curves depicting the sensitivity (P ≤0.05), and 7.501 for the insulin dose (P <0.01), having those by 1-specificity and the AUC values for the distinction of the PR parameters the highest G-test scores of all the analyzed parameters phase by the percentage of TREG (AUC = 0.82, 95% CI: 0.65 to (Table 3). Then, the relationship between covariates (age, sex, 0.98, P = 0.01), DCs (AUC = 0.7983, 95% CI: 0.62 to 0.97, P = BMI, HbA1c, basal and stimulated C-peptide, IDAA1c, insulin 0.0242), and monocytes (AUC = 0.7899, 95% CI: 0.60 to 0.98, P = Frontiers in Immunology | www.frontiersin.org 13 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission A B C D E FIGURE 8 | Circulating cytokine concentrations at different stages of T1D. Plasma from control subjects and patients with T1D at different time-points was obtained to quantitatively determine the concentrations of (A) IL-2, (B) IL-6, (D) IL-17A, and (E) IL-10 by CBA and of (C) TGF-b1 by ELISA. Uncolored boxes represent controls (n = 17), and patients are represented by gray boxes at disease onset (n = 17), light-colored boxes at PR (n = 11) and for remitter patients at 12 months (12 M PR) (n = 10), and dark-colored boxes for non-remitter patients at 8 months (8 M no PR) (n = 6) and 12 months (12 M no PR) (n = 6). Data are presented as box- and-whisker plots. Boxes indicate the first and third quartiles. The horizontal bar in the box indicates the median. Whiskers are drawn using Tukey’s criteria of 1.5× the interquartile range. Outliers beyond the whiskers are shown. *P ≤ 0.05, **P <0.01 after mixed effects model with Tukey’s post-hoc test for longitudinal data, or Kruskal–Wallis with Dunn’s post-hoc test for comparisons between control subjects and the different T1D time-points. P ≤ 0.05 is considered significant. 0.0284), and the insulin dose (AUC = 0.7971, 95% CI: 0.63 to of 0.77; Insulin dose with a median of 0.79 and mean of 0.72) 0.96, P = 0.0112). To visually examine how well these (Figures 9I–L). The predicted probability of PR for the observed independent variables do at predicting PR, we plot the non-PR group was significantly lower than those of the observed distribution of predicted probabilities for both the observed PR PR group in all cases (P <0.01 for %TREG; P ≤0.05 for %DCs; and non-PR groups. Looking at the violin plots for the group that P ≤0.05 for %Monocytes; and P <0.01 for Insulin dose). experienced the PR phase, the majority of patients had predicted However, the variables did not perform as well with classifying probabilities of being in PR above 0.5 in all cases (%TREG with a the group of observed non-remitters, as the predicted median of 0.82 and mean of 0.73; %DCs with a median of 0.82 probabilities are more uniformly distributed (%TREG with a and mean of 0.77; %Monocytes with a median of 0.79 and mean median of 0.57 and mean of 0.48; %DCs with a median of 0.58 Frontiers in Immunology | www.frontiersin.org 14 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission TABLE 3 | Simple logistic regressions for determinants of PR at T1D onset. Variable Coefficient SE OR [95% CI] |Z| P-valueG G test P-valueH Immunological variables Peripheral immune cell subsets (%) TREGA −0.798 0.377 0.45 [0.17 to 0.83] 2.116 0.03* 6.982 0.01* MonocytesB −0.332 0.172 0.71 [0.46 to 0.94] 1.936 0.05* 6.152 0.01* DCsB 0.330 0.168 1.39 [1.05 to 2.07] 1.959 0.05* 5.554 0.02* Cytokine concentration (in plasma)C IL-17A (fg/ml) 0.009 0.008 1.00 [0.99 to 1.02] 1.153 0.25 1.696 0.19 IL-10 (fg/ml) 0.0004 0.002 1.00 [0.99 to 1.00] 0.169 0.87 0.029 0.86 IL-2 (fg/ml) 0.001 0.004 1.00 [0.99 to 1.01] 0.285 0.78 0.088 0.77 IL-6 (fg/ml) 4.823e−005 6.575e−005 1.00 [1.00 to 1.00] 0.733 0.46 0.888 0.35 TGF-b (ng/ml) −4.324e−005 0.0002 1.00 [0.99 to 1.00] 0.180 0.86 0.032 0.86 Clinical and metabolic variables Age (years) 0.189 0.108 1.20 [0.99 to 1.55] 1.747 0.08 3.779 0.05* SexD 0.288 0.808 1.33 [0.28 to 6.90] 0.36 0.72 0.128 0.72 BMI (kg/m2) 0.352 0.204 1.42 [1.00 to 2.28] 1.726 0.08 3.928 0.05* BMI-SDS 0.492 0.638 1.64 [0.48 to 6.26] 0.771 0.441 0.610 0.44 HbA1c (%) −0.049 0.183 0.95 [0.65 to 1.37] 0.265 0.79 0.070 0.79 C-peptide (ng/ml) 2.436 1.983 11.43 [0.41 to 1170] 1.228 0.22 1.926 0.17 Stimulated C-peptide (ng/ml)B 2.191 1.427 8.94 [0.96 to 248.8] 1.535 0.12 3.687 0.05* IDAA1c −0.031 0.149 0.96 [0.71 to 1.30] 0.21 0.83 0.044 0.83 Insulin dose (U/kg/day) −7.217 3.174 7.3e−4 [4.18e−007 to 0.16] 2.274 0.02* 7.501 0.01* a-IA-2 autoAbE,F 0.470 1.524 1.60 [0.08 to 31.77] 0.31 0.756 0.09 0.76 a-GAD65 autoAbE,F 0.470 1.524 1.60 [0.08 to 31.77] 0.31 0.76 0.09 0.76 a-ZnT8 autoAbC,F −0.133 1.069 0.87 [0.10 to 7.11] −0.12 0.90 0.02 0.90 A n = 25; Bn = 24; Cn = 17; DRef.: female; En = 15; FCategorical variable; GP-value from OR; HP-value from G test. BMI, Body Mass Index; HbA1c, glycated hemoglobin; IDAA1c, insulin dose-adjusted HbA1c; SDS, standard deviation score; SE, standard error. *P ≤0.05. and mean of 0.57; %Monocytes with a median of 0.63 and mean conditions or cut-offs the patients meet for entering the PR of 0.55; Insulin dose with a median of 0.49 and mean of 0.47). phase, reducing in that way three variables to one. In that sense, higher values represent more conditions met (0, no condition; 3, An Index-Based Model Comprising the all conditions) (Figure 10C). Simple logistic regression analysis Percentage of TREG, DCs, and Monocytes was used to determine the association between the event of PR at T1D Onset Could Predict the PR and this index, finding 12-fold increased odds for being in PR To create a full model containing simultaneously the statistically with each condition met (OR = 12.49, 95% CI: 2.350 to 283.5, P = significant variables, we tested if they were monotonically related 0.03) (Table 4). The G-test score for this model was 12.75 (P = through Spearman’s tests. The percentage of TREG positively 0.0004). Then, to evaluate the discriminatory ability of this correlated with the insulin dose (Spearman’s r = 0.43, P ≤0.05) model among patients who will or will not be in PR, a ROC (Figure 10A). No significant correlations were found between curve was performed, presenting an AUC of 0.9219, 95% CI: 0.77 the percentage of TREG, monocytes, and DCs in the correlation to 1.00 (P = 0.0028) (Figure 10D). Moreover, in the violin plot, matrix plot in Figure 10B. Thus, a model containing only the distribution of predicted probabilities for the observed PR independent immune-related variables was created. group was far above 0.5 in all cases, being most of them around 1 First, from the ROC curves (Figures 9E–G) we selected the (median of 0.97 and mean of 0.89). On the contrary, the best cut-off values to discriminate between remitters and non- distribution of predicted probabilities for the observed non-PR remitters for each of the three independent variables, considering group was below 0.5 in all but one patient (median of 0.17 and both the sensitivity and the specificity. For the percentage of mean of 0.30) (P <0.001) (Figure 10E). TREG, the best cut-off value to determine PR was <8.35 with a Female sex, being younger than 5 years old, and higher BMI sensitivity of 68.75% (95% CI: 44.40 to 85.84%) and a specificity values have been related to the risk for non-remission (13, 33). of 88.89% (95% CI: 56.50 to 99.43%); for the percentage of DCs, Since age and BMI are highly correlated, we decided to adjust our the best cut-off value was >16.45 with a sensitivity of 64.71% index for BMI-SDS, which is standardized for child age and sex. (95% CI: 41.30 to 82.69%) and a specificity of 100% (95% CI: Multiple logistic regressions revealed 14-fold increased odds for 64.57 to 100.0%); and for the percentage of monocytes, the best being in PR with each condition met (OR = 14.43, 95% CI: 2.371 cut-off value was <14.35 with a sensitivity of 64.71% (95% CI: to 568.1, P ≤0.05), while BMI-SDS was not statistically significant 41.30 to 82.69%) and a specificity of 85.71% (95% CI: 48.69 to (Table 4). Therefore, by including this clinical parameter, the 99.27%) (Figure 10C). positive predictive effect of the simple logistic regression is Since the sample size is limited and to avoid problems of maintained at a significant level. Finally, multicollinearity was overfitting of the multiple variables, an index including the cut- evaluated for the variables index and BMI-SDS using VIF and R2 off’s values was created to determine on a 0–3 scale how many with other variables. Regarding both the index and BMI-SDS, Frontiers in Immunology | www.frontiersin.org 15 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission A B C D E F G H I J K L FIGURE 9 | The percentage of TREG, DCs, and monocytes, and the total daily insulin dose at T1D onset can discriminate between future remitters and non- remitters. (A–D) Box-and-whisker plots showing the percentage levels of (A) TREG, (B) DCs, and (C) monocytes, and (D) insulin dose at T1D onset separated by future remitters (blue boxes) and non-remitters (pink boxes). Boxes indicate the first and third quartiles and whiskers range from minimum to maximum values. The horizontal bar in the box indicates the median. (E–H) ROC curves plotting sensitivity and 1-specificity for detecting children with PR using percentages of (E) TREG, (F) DCs, and (G) monocytes, and (H) insulin dose. The AUC is indicated, being a measure of how well a quantitative test can distinguish between patients with and without PR. (I–L) Violin plots showing the frequency distribution of predicted probabilities for both the observed PR and non-PR groups regarding the percentage of (I) TREG, (J) DCs, and (K) monocytes, and (L) insulin dose. Each dot represents an individual patient. Violin plots show the median (thick dashed line) and first and third quartiles (thin dashed lines). *P ≤ 0.05, **P < 0.01 after 2-tailed Mann–Whitney test. P ≤ 0.05 is considered significant. N = 24–27. VIF was 1.002 and R2 was 0.0017 (Supplemental Table 6). T ERMA lymphocytes, T REG , neutrophils, DCs, B REG , and Consequently, multicollinearity is not a problem in our data. transitional T1 B lymphocytes. In addition, we have created a In summary, a model comprising the percentages of TREG, prediction model of PR that is based on an index that considers DCs, and monocytes all together in an index could predict the PR the percentages of TREG, monocytes, and DCs, and that could phase in children and adolescents at T1D diagnosis. distinguish between remitters and non-remitters at diagnosis. Despite needing independent validation, these candidate immunological biomarkers to monitor and predict the PR DISCUSSION corroborate that this stage is governed by both metabolic and immunological factors. The PR phase is still a poorly characterized stage of T1D natural Although some studies investigating the PR phase have been history but is of great interest given its association with better reported, only a few have performed a longitudinal follow-up glycemic control and the consequent reduction in secondary together with comparisons between remitter and non-remitter complications (8). In this study, we have analyzed up to 52 patients at an immunological level. To the best of our knowledge, peripheral immune cell subpopulations and different cytokines and based on the IDAA1c index to define remission, this is the in plasma during one year from T1D onset, focusing mainly on first study that prospectively monitors immune cell subsets and the stage of PR. Specific alterations of this phase have been cytokines in children and adolescents with T1D from disease discovered in terms of the percentage of EM T lymphocytes, onset to the first year after diagnosis, focusing on the PR phase Frontiers in Immunology | www.frontiersin.org 16 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission A B C D E FIGURE 10 | An index-based model comprising the percentage of TREG, DCs, and monocytes could predict PR at T1D diagnosis. (A) Spearman’s correlation coefficient matrix between insulin dose and the percentage of TREG (n = 25), and (B) between the percentages of TREG, DCs, and monocytes (n = 22) are shown in the form of a heatmap for r values (blue; positive correlation; red, negative correlation). (C) Strategy followed to create an index comprising all the independent significant variables. From ROC curves, the best cut-off values were selected to discriminate between the remitter and the non-remitter groups for the percentage of TREG (<8.35), DCs (>16.45), and monocytes (<14.35), considering both the sensitivity and the specificity with 95% CI. An index including the selected cut-off’s values was created to determine for each patient how many conditions they met for entering PR on a 0–3 scale (0, no condition; 3, all conditions), being the cut-offs the conditions that should be fulfilled. (D) Graphical representation of the AUC for the created index. (E) Violin plot showing the frequency distribution of predicted probabilities for both the observed PR (n = 16) and non-PR (n = 6) groups regarding the index. Each dot represents an individual patient. Violin plots show the median (thick dashed line) and first and third quartiles (thin dashed lines). *P ≤0.05 after Spearman’s correlation test, ***P <0.001 after 2-tailed Mann–Whitney test. P ≤0.05 is considered significant. and comparing between remitters and non-remitters to find (34, 35). The here reported increase in the percentages of reliable specific biomarkers. peripheral blood CD4+ and CD8+ TERMA lymphocytes and EM CD8+ T cells are the predominant component of insulitis in CD4+ T lymphocytes at PR could reflect their lower migration to recent-onset T1D followed by macrophages and CD4+ T cells the pancreas, resulting in reduced percentages of peripheral Frontiers in Immunology | www.frontiersin.org 17 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission TABLE 4 | Simple logistic regression of index as a determinant of PR in pediatric patients at T1D onset and its adjustment for BMI-SDS. Simple logistic regression Variable Coefficient SE OR [95% CI] |Z| P-valueC G test P-valueD IndexA 2.525 1.159 12.49 [2.350 to 283.5] 2.18 0.03* 12.75 0.0004*** Multiple logistic regression IndexA 2.669 1.335 14.43 [2.371 to 568.1] 1.999 0.05* 12.83 0.0016** BMI-SDSC −0.374 1.303 0.69 [0.043 to 10.69] 0.287 0.77 A TREG (%) + Monocytes (%) + DCs (%) cut-off values from independent ROC curves as conditions to be fulfilled on a 0–3 scale. CP-value from OR; DP-value from G test. BMI, Body Mass Index; SDS, standard deviation score; SE, standard error. *P ≤0.05, **P ≤ 0.01, ***P ≤ 0.001. naïve T lymphocytes. According to these findings, increased memory T R E G are decreased in percentage at PR in percentages of circulating cytotoxic T lymphocytes positive for comparison to patients without PR. One possible explanation IFN-g, TH1 lymphocytes, and TH17 lymphocytes have been is that these cells could recover their impaired function with the found during the PR phase in comparison to the disease onset rapid rectification of hyperglycemia after T1D onset, being more (21), and also after the first year from diagnosis for CD4+ and active in secondary lymphoid organs and the target tissue. Using CD8+ TERMA lymphocytes, EM CD4+ T lymphocytes, and TH17 the same markers, Fitas et al. (21) found only a decrease in TREG lymphocytes (19). Of note, islets from patients with T1D express after a year from T1D onset, but not during the PR phase, while CXCL10, a chemokine involved in autoreactive T lymphocyte we previously found an increase in activated TREG one year after recruitment, while in controls, neither CXCL10+ endocrine cells diagnosis (19). Other studies investigating the association of T nor CXCR3+ lymphocytes were detected (36). That could lymphocytes and PR found that its length positively correlates partially explain the lower percentages of EM CD8+ T cells with the high frequency of activated T R E G (18) and observed at T1D onset in the periphery (37). Also, b-cells from CD4 + CD25 + CD127 h i T cells (17) at disease onset. patients with T1D hyperexpress HLA class I, thus increasing Furthermore, IL-10-dependent regulatory CD4+ T lymphocyte autoantigen presentation to autoreactive CD8+ T cells (38), pathways are involved in long-term remission of T1D (15), and which were found within the islets showing an antigen- islet-specific IL-10+ immune responses but not experienced phenotype (CD45RA−) (39, 40). Since the main CD4+CD25+FoxP3+ cells at diagnosis predict glycemic control. mediators of b-cell destruction in T1D are autoreactive effector In fact, peripheral antigen-specific TREG were diminished during CD4+ and CD8+ T cells, and supposing that EM T lymphocytes PR in comparison to diagnosis (22). We found this trend for total are acting less in situ, their increase in the periphery during PR TREG cells in most of our remitter patients except four. Those could reflect an attempt at immunoregulation and b-cell patients presented a similar behavior to non-remitters, who recovery. Interestingly, a recent longitudinal study (24) presented increased percentages of TREG cells from disease described an association of the PR phase with the restoration onset. Interestingly, they are the youngest within the PR group. of the programmed cell death-1/programmed death-ligand 1 axis Since pediatric patients diagnosed before age 7 may have a more on T cells, suggesting a mechanism of immunoregulation that aggressive form of T1D, that could be related to the observed did not occur in non-remitter patients. changes in T R E G cells (47). In conclusion, further On the other hand, homeostasis between TREG and effector T characterization of TREG subsets in terms of phenotype and lymphocytes is crucial for the induction and maintenance of function needs to be addressed to dissect their role during PR. peripheral tolerance. The prediabetic phase in the natural history Innate immune cells are crucial players in the development of T1D is indeed very heterogeneous; it can last from months to and progression of T1D and multiple interactions occur between years, and different factors influence the fact of presenting overt them and lymphocytes (48). Here, neutrophils and DCs were T1D, like the number and titers of autoantibodies. Evidence found to be altered during PR. Neutrophils are present in the supports that having higher frequencies of insulin-specific TREG insulitis before diagnosis, and they continue infiltrating the is associated with a slow progression to clinically symptomatic pancreas as the disease progresses, having a direct pathogenic T1D (41), but that at this time-point, TREG are dysfunctional, role (49). Previous results showed a reduction in neutrophil thus contributing directly to disease development [reviewed in counts at diagnosis that is associated with a poor endocrine (42)]. However, the dominance of TREG over effector T cells may pancreatic function (50, 51). Here, we did find a reduction in contribute to PR occurrence (32). In fact, islet-specific CD8+ T CD16+ neutrophil absolute counts in non-remitter patients both cells with an exhaustion-like profile identify patients with slow at 8 and 12 months after the diagnosis when compared to T1D progression after onset (43). Until recently, it was generally remitter patients. Peripheral neutrophils mainly express a accepted that the overall frequency of peripheral blood rather atypical CD16, the FcgRIIIB, which can trigger CD4+FoxP3+ TREG is unaltered in patients with T1D (44–46). neutrophil activation. The possible explanations for the Nonetheless, contradictory data on TREG have been reported, decreased CD16+ neutrophil are 1) the greater activation or probably because of the different ways that these cells can be apoptosis that would involve FcgRIIIB proteolytic cleavage, 2) identified. Here, we found that CD4+CD25+CD127−/low and tissue detainment, or 3) abnormal maturation (51–53). As Frontiers in Immunology | www.frontiersin.org 18 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission hyperglycemia is linked to neutrophil dysfunction, non-remitters cytokines were predictive of PR, in a prospective study of T1D, are more susceptible to present those impairments. On the other lower levels of IL-10, IFN-g, and IL-1R1 at diagnosis were hand, DCs play a pivotal role in modulating T cell responses by associated with remission (26). In summary, the use of altering the balance between tolerance and autoimmunity (54). cytokines as T1D progression markers does not seem to have We found that patients at PR present lower percentages of DCs enough power to correctly distinguish PR. than non-remitter patients. Previous research in pediatric Because the efficacy of immunotherapies depends on the time patients found impaired functionality, percentages, and/or of intervention, mathematical models for predicting disease numbers in total DCs, mDCs, and pDCs both at T1D onset progression (68) and PR are important tools to improve and one year after diagnosis (55–57). One explanation could be therapeutic strategies. Regarding metabolism, previous studies the previously described reduced production of DCs from have shown associations between metabolic and clinical monocytes in T1D (58), although our data on the DC numbers parameters and the PR phase. In our bivariate analysis, only do not fit well with those results. Also, inflammation causes the lower insulin dose at T1D onset could predict remission. CCL2 release from islets, prompting translocation of CCR2+ DCs Although not significant, and according to previous data, we from circulation to inflamed tissues, thus causing a reduction in found that patients with higher BMI and older age at diagnosis DCs counts and their CCR2 expression in peripheral blood (56). are more prone to experience remission (69). In general, children Since the amelioration of hyperglycemia could contribute to the experience remission to a higher extend than adults, and within recovery of the immunosuppressive properties of DCs during the the first group, the probability of remaining in remission is PR phase (59), tolerogenic DCs expressing CCR2 could be more greater as the onset age increased (70). Different regression active in the target tissue, diminishing in peripheral blood and analyses showed that higher pH and bicarbonate levels, higher prompting TREG responses. BMI, male sex, and lower HbA1c, insulin dose, and numbers In addition, after 12 months of follow-up, only patients that of islet antibodies at T1D onset are predictors of remission experienced PR showed increased levels of BREG and transitional (13, 14, 70–72). However, the prediction of remission using T1 B lymphocytes. According to our results, other studies in immunological data is rather scarce. Unlike IFN-g concentration, new-onset T1D have not found alterations in other B cell subsets elevated levels of IL-10 and IL-6 in serum are positively (37, 60), while transitional B cells and BREG increased in pediatric associated with remission (22, 23), and the absence of IL-4, patients one year after diagnosis (19). Both BREG and transitional TNF-a, IL-10, and IL-13 detection is positively correlated with B lymphocytes have immunoregulatory properties and can its length (21). Also, CD4+CD45RO+ T cells, activated TREG, inhibit effector T cell proliferation (61, 62). Some results CD4+CD25+CD127hi T cells, and apoptotic TREG are reportedly suggest that B cell subsets could play a role in T1D associated with remission (17, 18, 73). Here, we found that the pathogenesis, for instance, both the percentage of IL-10+ BREG percentages of TREG, monocytes, and DCs at T1D onset are and IL-10+ immature transitional B cells were significantly lower independent predictors of the PR phase. TREG were also in patients with T1D at diagnosis than in controls and patients positively correlated with the insulin dose, highlighting the with worse glycemic control (63, 64). Regarding transitional B crucial role of immunometabolism interactions during T1D cells, these are classified in T1 and T2 subsets, having T1 cells and the mechanisms rectifying glucose toxicity in the process greater immunosuppressive properties. Therefore, the here of b-cell protection. The bivariate analysis of the generated index, reported higher ratio of T1/T2 cells and the increased counts which includes the cell percentages, showed an even higher of BREG after a year of follow-up for remitter patients could ability to distinguish remitters from non-remitters. When the reflect an immunoregulatory attempt. index was adjusted for possible confounders, only it remained a Cytokines orchestrate multiple interactions between b cells significant predictor of PR. In summary, we propose a novel and immune cells. Here, none of the analyzed cytokines showed immune-based predictive model of remission, although it must significant differences between remitters and non-remitters, be further validated in a larger cohort of patients. which would limit their use as biomarkers of remission or T1D We are aware of the limitations of the presented study. First, progression. However, we found increased levels of IL-17A, a the number of individuals studied lowers the statistical power; relevant cytokine in T1D (65), at disease onset, and a trend to however, such data are very hard to obtain. Another weak point reduce this cytokine in remitter patients. IL-17A expression is of our cohort is the wide range of ages (from 4 to 18) and the upregulated in the pancreas of both humans and animal models variability that this entails regarding the physiological changes (44, 66), and among other effects, it recruits and activates that the subjects undergo (i.e., puberty). Whereas variations in neutrophils (67). Because IL-23 and/or TGF-b plus IL-6 drive the levels of immune cell subpopulations over one year in healthy the production of IL-17 by T cells, the here reported IL-17A pediatric or adult subjects are subtle or inexistent (74, 75), increase together with the higher levels of TGF-b and IL-6 could changes are observed within longer tracking periods (76). In reflect the inflammatory state at diagnosis, which is less evident that sense, we performed correlations between age and the at PR. In fact, we found a positive correlation between IL-6 and percentage or absolute counts of the different analyzed IL-17A at T1D onset, whereas this correlation was negative in immune cell subsets both in controls and patients at T1D controls (data not shown). In a previous non-longitudinal study, onset. We observed that the levels of transitional B cells and low TGF-b levels were a feature of PR (19), which has not been CD8+ TEMRA cells are negatively correlated with age and that the confirmed in this longitudinal study. While none of the analyzed levels of EM and CM CD4+ T cells are positively correlated (data Frontiers in Immunology | www.frontiersin.org 19 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission not shown). Since our longitudinal data showed an increase in University Hospital. Written informed consent to participate transitional B cells after one year of follow-up and a decrease in in this study was provided by the participants’ legal guardian/ the percentage of EM CD4+ T cells, these changes are likely to be next of kin. a consequence of T1D progression. Although the usage of peripheral blood as a source of biomarkers is a non-invasive technique, immune cells are tightly regulated by their generation in primary lymphoid organs and their migration to tissues, their AUTHOR CONTRIBUTIONS changes in the periphery being difficult to interpret. The analysis of the target organ and further studies on the functionality of LGM, DPB, SRF, and MVP designed the experiments. FV, MM, immune cells (i.e., cytokine production)—including antigen- AV, JP, RC, LC, and JB selected the patients, obtained the specific T cells—are needed to correlate changes in the samples, and/or determined autoantibodies. LGM and DPB periphery with the events occurring in the pancreas. Moreover, performed the experiments. LGM, DPB, and JMCA analyzed a larger sample size in future studies is needed to validate the the data. LGM and MVP wrote the manuscript. All authors listed here proposed biomarkers for the PR phase. Since limited have made a substantial, direct, and intellectual contribution to amount of blood was obtained from pediatric patients, the work and approved it for publication. autoantibodies to insulin (IAA) have not been explored. However, due to its b-cell specificity, the baseline titer of IAA could be of interest for the study of the PR phase. Still focusing on this stage, we are not evaluating its length, which could be FUNDING related to the persistence and activity of certain immune cells. Funding for this study was provided by the Spanish Government The strengths of the study include the longitudinal picture of (FIS PI18/00436) co-financed with the European Regional T1D early stages, the characterization of the PR phase versus Development funds (FEDER), and by DiabetesCero non-remission by a wide range of immune subpopulations and Foundation. LGM is supported by the Generalitat de cytokines, and the use of IDAA1c of ≤9 to define PR, an index Catalunya (PERIS PIF-Salut Grant No. SLT017/20/000049). that includes glycemic control and insulin doses, that correlates This work has been supported by positive discussion through with C-peptide levels, and that has been validated in a large Consolidated Research Group #2017 SGR 103, AGAUR, cohort of young patients with T1D (77). Generalitat de Catalunya. In conclusion, the PR phase is not only accompanied by changes in different metabolic parameters, but also by changes in immune cells and molecules. These alterations could be potential monitoring and predictive biomarkers useful for patient ACKNOWLEDGMENTS stratification in clinical trials and the identification of patients with better glycemic control that will enable a more personalized We are grateful to the Lilly Foundation for sponsoring the award therapeutic management. given to this work during the 43rd Congress of the Spanish Society of Pediatric Endocrinology (2021). Special thanks to Ms. N. Real from the Pediatrics Dept. of the HUGTiP for obtaining the samples. We appreciate the advice from Dr. E. Martı́nez- DATA AVAILABILITY STATEMENT Cá c eres, Dr. A. Teniente, and Ms. B. Quirant from the The original contributions presented in the study are included in Immunology Dept. (HUGTiP). We are grateful to Mr. M. the article/Supplementary Material. Further inquiries can be Ferná ndez and G. Requena from the Flow Cytometry Platform directed to the corresponding author. (IGTP) for their support. ETHICS STATEMENT SUPPLEMENTARY MATERIAL The studies involving human participants were reviewed and The Supplementary Material for this article can be found online approved by the Committee on the Ethics of Research of the at: https://www.frontiersin.org/articles/10.3389/fimmu.2022. Germans Trias i Pujol University Hospital and Parc Taulı́ 825426/full#supplementary-material 2. Patterson CC, Harjutsalo V, Rosenbauer J, Neu A, Cinek O, Skrivarhaug T, REFERENCES et al. Trends and Cyclical Variation in the Incidence of Childhood Type 1 1. Katsarou A, Gudbjörnsdottir S, Rawshani A, Dabelea D, Bonifacio E, Diabetes in 26 European Centres in the 25 Year Period 1989–2013: A Anderson BJ, et al. Type 1 Diabetes Mellitus. Nat Rev Dis Prim (2017) Multicentre Prospective Registration Study. Diabetologia (2019) 62(3):408– 3:17016. doi: 10.1038/nrdp.2017.16 17. doi: 10.1007/s00125-018-4763-3 Frontiers in Immunology | www.frontiersin.org 20 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission 3. Lehuen A, Diana J, Zaccone P, Cooke A. Immune Cell Crosstalk in Type 1 22. Sanda S, Roep BO, von Herrath M. Islet Antigen Specific IL-10+ Immune Diabetes. Nat Rev Immunol (2010) 10(7):501–13. doi: 10.1038/nri2787 Responses But Not CD4+CD25+FoxP3+ Cells at Diagnosis Predict Glycemic 4. Atkinson MA, Roep BO, Posgai A, Wheeler DCS, Peakman M. The Challenge Control in Type 1 Diabetes. Clin Immunol (2008) 127(2):138–43. of Modulating b-Cell Autoimmunity in Type 1 Diabetes. Lancet Diabetes doi: 10.1016/j.clim.2007.12.003 Endocrinol (2019) 7(1):52–64. doi: 10.1016/S2213-8587(18)30112-8 23. Pfleger C, Mortensen HB, Hansen L, Herder C, Roep BO, Hoey H, et al. 5. Roep BO, Wheeler DCS, Peakman M. Antigen-Based Immune Modulation Association of IL-1ra and Adiponectin With C-Peptide and Remission in Therapy for Type 1 Diabetes: The Era of Precision Medicine. Lancet Diabetes Patients With Type 1 Diabetes. Diabetes (2008) 57(4):929–37. doi: 10.2337/ Endocrinol (2019) 7(1):65–74. doi: 10.1016/S2213-8587(18)30109-8 db07-1697 6. Culina S, Mallone R. Immune Biomarkers in Immunotherapeutic Trials for 24. Li X, Zhong T, Tang R, Wu C, Xie Y, Liu F, et al. PD-1 and PD-L1 Expression Type 1 Diabetes: Cui Prodest? Diabetes Metab (2012) 38(5):379–85. in Peripheral CD4/CD8+ T Cells Is Restored in the Partial Remission Phase in doi: 10.1016/j.diabet.2012.05.005 Type 1 Diabetes. J Clin Endocrinol Metab (2020) 105(6):1947–56. 7. von Herrath M, Peakman M, Roep B. Progress in Immune-Based Therapies doi: 10.1210/clinem/dgaa130 for Type 1 Diabetes. Clin Exp Immunol (2013) 172(2):186–202. doi: 10.1111/ 25. Alizadeh BZ, Hanifi-Moghaddam P, Eerligh P, van der Slik AR, Kolb H, cei.12085 Kharagjitsingh AV, et al. Association of Interferon-g and Interleukin 10 8. Fonolleda M, Murillo M, Vá zquez F, Bel J, Vives-Pi M. Remission Phase in Genotypes and Serum Levels With Partial Clinical Remission in Type 1 Paediatric Type 1 Diabetes: New Understanding and Emerging Biomarkers. Diabetes. Clin Exp Immunol (2006) 145(3):480–4. doi: 10.1111/j.1365- Horm Res Paediatr (2017) 88(5):307–15. doi: 10.1159/000479030 2249.2006.03172.x 9. Abdul-Rasoul M, Habib H, Al-Khouly M. “The Honeymoon Phase” in 26. Schloot NC, Hanifi-Moghaddam P, Aabenhus-Andersen N, Alizadeh BZ, Children With Type 1 Diabetes Mellitus: Frequency, Duration, and Saha MT, Knip M, et al. Association of Immune Mediators at Diagnosis of Influential Factors. Pediatr Diabetes (2006) 7(2):101–7. doi: 10.1111/j.1399- Type 1 Diabetes With Later Clinical Remission. Diabetes Med (2007) 24 543X.2006.00155.x (5):512–20. doi: 10.1111/j.1464-5491.2007.02096.x 10. Murillo M, Fonolleda M, Bosch L, Rodrı́guez-Ferná ndez S, Vá zquez F, Bel J, 27. Kaas A, Pfleger C, Hansen L, Buschard K, Schloot NC, Roep BO, et al. et al. Sustained Spontaneous Partial Remission in a Pediatric Patient With Association of Adiponectin, Interleukin (IL)-1ra, Inducible Protein 10, IL-6 Type 1 Diabetes. J Clin Transl Endocrinol Case Rep (2017) 6:11–3. and Number of Islet Autoantibodies With Progression Patterns of Type 1 doi: 10.1016/j.jecr.2017.09.001 Diabetes the First Year After Diagnosis. Clin Exp Immunol (2010) 161(3):444– 11. Karges B, Durinovic-Bello I, Heinze E, Boehm BO, Debatin K-M, Karges W. 52. doi: 10.1111/j.1365-2249.2010.04193.x Complete Long-Term Recovery of -Cell Function in Autoimmune Type 1 28. ADA. 2. Classification and Diagnosis of Diabetes: Standards of Medical Care Diabetes After Insulin Treatment. Diabetes Care (2004) 27(5):1207–8. in Diabetes—2018. Diabetes Care (2018) 41:S13–27. doi: 10.2337/dc18-S002 doi: 10.2337/diacare.27.5.1207 29. Carrascosa Lezcano A, Ferná ndez Garcı́a JM, Ferná ndez Ramos C, Ferrá ndez 12. Zhong T, Tang R, Gong S, Li J, Li X, Zhou Z. The Remission Phase in Type 1 Longá s A, Ló pez-Siguero JP, Sá nchez Gonzá lez E, et al. Estudio Transversal Diabetes: Changing Epidemiology, Definitions, and Emerging Immuno- Español De Crecimiento 2008. Parte II: Valores De Talla, Peso E Índice De Metabolic Mechanisms. Diabetes Metab Res Rev (2020) 36(2):e3207. Masa Corporal Desde El Nacimiento a La Talla Adulta. Pediatrı́a (2008) 68 doi: 10.1002/dmrr.3207 (6):552–69. doi: 10.1157/13123287 13. Marino KR, Lundberg RL, Jasrotia A, Maranda LS, Thompson MJ, Barton BA, 30. Mortensen HB, Hougaard P, Swift P, Hansen L, Holl RW, Hoey H, et al. New et al. A Predictive Model for Lack of Partial Clinical Remission in New-Onset Definition for the Partial Remission Period in Children and Adolescents With Pediatric Type 1 Diabetes. PloS One (2017) 12(5):e0176860. doi: 10.1371/ Type 1 Diabetes. Diabetes Care (2009) 32(8):1384–90. doi: 10.2337/dc08-1987 journal.pone.0176860 31. Robb MA, McInnes PM, Califf RM. Biomarkers and Surrogate Endpoints: 14. Wong TWC, Wong MYS, But WMB. Features of Partial Remission in Developing Common Terminology and Definitions. JAMA (2016) 315 Children With Type 1 Diabetes Using the Insulin Dose-Adjusted A1c (11):1107. doi: 10.1001/jama.2016.2240 Definition and Risk Factors Associated With Nonremission. Ann Pediatr 32. Jaberi-Douraki M, Pietropaolo M, Khadra A. Continuum Model of T-Cell Endocrinol Metab (2021) 26(2):118–25. doi: 10.6065/apem.2040202.101 Avidity: Understanding Autoreactive and Regulatory T-Cell Responses in 15. Karges B, Durinovic-Belló I, Heinze E, Debatin K-M, Boehm B, Karges W. Type 1 Diabetes. J Theor Biol (2015) 383:93–105. doi: 10.1016/ Immunological Mechanisms Associated With Long-Term Remission of j.jtbi.2015.07.032 Human Type 1 Diabetes. Diabetes Metab Res Rev (2006) 22(3):184–9. 33. Pyziak A, Zmyslowska A, Bobeff K, Malachowska B, Fendler W, Wyka K, et al. doi: 10.1002/dmrr.600 Markers Influencing the Presence of Partial Clinical Remission in Patients 16. Tang R, Zhong T, Wu C, Zhou Z, Li X. The Remission Phase in Type 1 With Newly Diagnosed Type 1 Diabetes. J Pediatr Endocrinol Metab (2017) 30 Diabetes: Role of Hyperglycemia Rectification in Immune Modulation. Front (11):1147–53. doi: 10.1515/jpem-2017-0100 Endocrinol (Lausanne) (2019) 10:824. doi: 10.3389/fendo.2019.00824 34. Nigi L, Maccora C, Dotta F, Sebastiani G. From Immunohistological to 17. Narsale A, Lam B, Moya R, Lu T, Mandelli A, Gotuzzo I, et al. CD4+CD25 Anatomical Alterations of Human Pancreas in Type 1 Diabetes: New +CD127hi Cell Frequency Predicts Disease Progression in Type 1 Diabetes. Concepts on the Stage. Diabetes Metab Res Rev (2020) 36(4):e3264. JCI Insight (2021) 6(2):e136114. doi: 10.1172/jci.insight.136114 doi: 10.1002/dmrr.3264 18. Moya R, Robertson HK, Payne D, Narsale A, Koziol J, Davies JD. A Pilot 35. Planas R, Carrillo J, Sanchez A, Ruiz de Villa MC, Nuñez F, Verdaguer J, et al. Study Showing Associations Between Frequency of CD4+ Memory Cell Gene Expression Profiles for the Human Pancreas and Purified Islets in Type Subsets at Diagnosis and Duration of Partial Remission in Type 1 Diabetes. 1 Diabetes: New Findings at Clinical Onset and in Long-Standing Diabetes. Clin Immunol (2016) 166–167:72–80. doi: 10.1016/j.clim.2016.04.012 Clin Exp Immunol (2010) 159(1):23–44. doi: 10.1111/j.1365- 19. Villalba A, Fonolleda M, Murillo M, Rodriguez-Fernandez S, Ampudia R-M, 2249.2009.04053.x Perna-Barrull D, et al. Partial Remission and Early Stages of Pediatric Type 1 36. Nigi L, Brusco N, Grieco GE, Licata G, Krogvold L, Marselli L, et al. Pancreatic Diabetes Display Immunoregulatory Changes. A Pilot Study. Transl Res Alpha-Cells Contribute Together With Beta-Cells to CXCL10 Expression in (2019) 210:8–25. doi: 10.1016/j.trsl.2019.03.002 Type 1 Diabetes. Front Endocrinol (Lausanne) (2020) 11:630. doi: 10.3389/ 20. Gomez-Muñoz L, Perna-Barrull D, Villalba A, Rodriguez-Fernandez S, fendo.2020.00630 Ampudia R-M, Teniente-Serra A, et al. NK Cell Subsets Changes in Partial 37. Oras A, Peet A, Giese T, Tillmann V. Uibo R. A Study of 51 Subtypes of Remission and Early Stages of Pediatric Type 1 Diabetes. Front Immunol Peripheral Blood Immune Cells in Newly Diagnosed Young Type 1 Diabetes (2021) 11:611522. doi: 10.3389/fimmu.2020.611522 Patients. Clin Exp Immunol (2019) 198(1):57–70. doi: 10.1111/cei.13332 21. Fitas AL, Martins C, Borrego LM, Lopes L, Jörns A, Lenzen S, et al. Immune 38. Richardson SJ, Rodriguez-Calvo T, Gerling IC, Mathews CE, Kaddis JS, Cell and Cytokine Patterns in Children With Type 1 Diabetes Mellitus Russell MA, et al. Islet Cell Hyperexpression of HLA Class I Antigens: A Undergoing a Remission Phase: A Longitudinal Study. Pediatr Diabetes Defining Feature in Type 1 Diabetes. Diabetologia (2016) 59(11):2448–58. (2018) 19(5):963–71. doi: 10.1111/pedi.12671 doi: 10.1007/s00125-016-4067-4 Frontiers in Immunology | www.frontiersin.org 21 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission 39. Coppieters KT, Dotta F, Amirian N, Campbell PD, Kay TWH, Atkinson MA, 59. Thomas AM, Dong Y, Beskid NM, Garcı́a AJ, Adams AB, Babensee JE. Brief et al. Demonstration of Islet-Autoreactive CD8 T Cells in Insulitic Lesions Exposure to Hyperglycemia Activates Dendritic Cells In Vitro and In Vivo. From Recent Onset and Long-Term Type 1 Diabetes Patients. J Exp Med J Cell Physiol (2020) 235(6):5120–9. doi: 10.1002/jcp.29380 (2012) 209(1):51–60. doi: 10.1084/jem.20111187 60. Thompson WS, Pekalski ML, Simons HZ, Smyth DJ, Castro-Dopico X, Guo 40. Gonzalez-Duque S, Azoury ME, Colli ML, Afonso G, Turatsinze J-V, Nigi L, H, et al. Multi-Parametric Flow Cytometric and Genetic Investigation of the et al. Conventional and Neo-Antigenic Peptides Presented by b Cells Are Peripheral B-Cell Compartment in Human Type 1 Diabetes. Clin Exp Targeted by Circulating Naïve CD8+ T Cells in Type 1 Diabetic and Healthy Immunol (2014) 177(3):571–85. doi: 10.1111/cei.12362 Donors. Cell Metab (2018) 28(6):946–960.e6. doi: 10.1016/j.cmet.2018.07.007 61. Catalá n D, Mansilla MA, Ferrier A, Soto L, Oleinika K, Aguilló n JC, et al. 41. Serr I, Fürst RW, Achenbach P, Scherm MG, Gökmen F, Haupt F, et al. Type 1 Immunosuppressive Mechanisms of Regulatory B Cells. Front Immunol Diabetes Vaccine Candidates Promote Human Foxp3+ Treg Induction in (2021) 12:611795. doi: 10.3389/fimmu.2021.611795 Humanized Mice. Nat Commun (2016) 7(1):10991. doi: 10.1038/ncomms10991 62. Simon Q, Pers J-O, Cornec D, Le Pottier L, Mageed RA, Hillion S. In-Depth 42. Hull CM, Peakman M, Tree TIM. Regulatory T Cell Dysfunction in Type 1 Characterization of CD24 High CD38 High Transitional Human B Cells Diabetes: What’s Broken and How Can We Fix it? Diabetologia (2017) 60 Reveals Different Regulatory Profiles. J Allergy Clin Immunol (2016) 137 (10):1839–50. doi: 10.1007/s00125-017-4377-1 (5):1577–1584.e10. doi: 10.1016/j.jaci.2015.09.014 43. Wiedeman AE, Muir VS, Rosasco MG, DeBerg HA, Presnell S, Haas B, et al. 63. El-Mokhtar MA, Elsherbiny NM, Sayed D, Raafat DM, Askar E, Hussein A, Autoreactive CD8+ T Cell Exhaustion Distinguishes Subjects With Slow Type et al. Altered Regulatory B Cell Subsets in Children With Type 1 Diabetes 1 Diabetes Progression. J Clin Invest (2020) 130(1):480–90. doi: 10.1172/ Mellitus. J Immunol Res (2020) 2020:8935694. doi: 10.1155/2020/8935694 JCI126595 64. Wang Y, Qin Y, Wang X, Zhang L, Wang J, Xu X, et al. Decrease in the 44. Ferraro A, Socci C, Stabilini A, Valle A, Monti P, Piemonti L, et al. Expansion Proportion of CD24 Hi CD38 Hi B Cells and Impairment of Their Regulatory of Th17 Cells and Functional Defects in T Regulatory Cells Are Key Features Capacity in Type 1 Diabetes Patients. Clin Exp Immunol (2020) 200(1):22–32. of the Pancreatic Lymph Nodes in Patients With Type 1 Diabetes. Diabetes doi: 10.1111/cei.13408 (2011) 60(11):2903–13. doi: 10.2337/db11-0090 65. Arif S, Moore F, Marks K, Bouckenooghe T, Dayan CM, Planas R, et al. 45. Hamari S, Kirveskoski T, Glumoff V, Kulmala P, Simell O, Knip M, et al. Peripheral and Islet Interleukin-17 Pathway Activation Characterizes Human Analyses of Regulatory CD4+ CD25+ FOXP3+ T Cells and Observations Autoimmune Diabetes and Promotes Cytokine-Mediated -Cell Death. From Peripheral T Cell Subpopulation Markers During the Development of Diabetes (2011) 60(8):2112–9. doi: 10.2337/db10-1643 Type 1 Diabetes in Children. Scand J Immunol (2016) 83(4):279–87. 66. Vukkadapu SS, Belli JM, Ishii K, Jegga AG, Hutton JJ, Aronow BJ, et al. doi: 10.1111/sji.12418 Dynamic Interaction Between T Cell-Mediated b-Cell Damage and b-Cell 46. Brusko T, Wasserfall C, McGrail K, Schatz R, Viener HL, Schatz D, et al. No Repair in the Run Up to Autoimmune Diabetes of the NOD Mouse. Physiol Alterations in the Frequency of FOXP3+ Regulatory T-Cells in Type 1 Genomics (2005) 21(2):201–11. doi: 10.1152/physiolgenomics.00173.2004 Diabetes. Diabetes (2007) 56(3):604–12. doi: 10.2337/db06-1248 67. Pelletier M, Maggi L, Micheletti A, Lazzeri E, Tamassia N, Costantini C, et al. 47. Leete P, Willcox A, Krogvold L, Dahl-Jørgensen K, Foulis AK, Richardson SJ, Evidence for a Cross-Talk Between Human Neutrophils and Th17 Cells. Blood et al. Differential Insulitic Profiles Determine the Extent of b-Cell Destruction (2010) 115(2):335–43. doi: 10.1182/blood-2009-04-216085 and the Age at Onset of Type 1 Diabetes. Diabetes (2016) 65(5):1362–9. 68. Jaberi-Douraki M, Pietropaolo M, Khadra A. Predictive Models of Type 1 doi: 10.2337/db15-1615 Diabetes Progression: Understanding T-Cell Cycles and Their Implications on 48. Diana J, Gahzarian L, Simoni Y, Lehuen A. Innate Immunity in Type 1 Autoantibody Release. PloS One (2014) 9(4):e93326. doi: 10.1371/ Diabetes. Discov Med (2011) 11(61):513–20. journal.pone.0093326 49. Huang J, Xiao Y, Xu A, Zhou Z. Neutrophils in Type 1 Diabetes. J Diabetes 69. Passanisi S, Salzano G, Gasbarro A, Urzì Brancati V, Mondio M, Pajno GB, Investig (2016) 7(5):652–63. doi: 10.1111/jdi.12469 et al. Influence of Age on Partial Clinical Remission Among Children With 50. Qin J, Fu S, Speake C, Greenbaum CJ, Odegard JM. NETosis-Associated Serum Newly Diagnosed Type 1 Diabetes. Int J Environ Res Public Health (2020) 17 Biomarkers Are Reduced in Type 1 Diabetes in Association With Neutrophil (13):4801. doi: 10.1515/jpem-2017-0100 Count. Clin Exp Immunol (2016) 184(3):318–22. doi: 10.1111/cei.12783 70. Zhong T, Tang R, Xie Y, Liu F, Li X, Zhou Z. Frequency, Clinical 51. Vecchio F, Lo Buono N, Stabilini A, Nigi L, Dufort MJ, Geyer S, et al. Characteristics, and Determinants of Partial Remission in Type 1 Diabetes: Abnormal Neutrophil Signature in the Blood and Pancreas of Different Patterns in Children and Adults. J Diabetes (2020) 12(10):761–8. Presymptomatic and Symptomatic Type 1 Diabetes. JCI Insight (2018) 3 doi: 10.1111/1753-0407.13044 (18):e122146. doi: 10.1172/jci.insight.122146 71. Chobot A, Stompó r J, Szyda K, Sokołowska M, Deja G, Polań ska J, et al. 52. Wang Y, Jönsson F. Expression, Role, and Regulation of Neutrophil Fcg Remission Phase in Children Diagnosed With Type 1 Diabetes in Years 2012 Receptors. Front Immunol (2019) 10:1958. doi: 10.3389/fimmu.2019.01958 to 2013 in Silesia, Poland: An Observational Study. Pediatr Diabetes (2019) 20 53. Silvestre-Roig C, Fridlender ZG, Glogauer M, Scapini P. Neutrophil Diversity (3):286–92. doi: 10.1111/pedi.12824 in Health and Disease. Trends Immunol (2019) 40(7):565–83. doi: 10.1016/ 72. Nagl K, Hermann JM, Plamper M, Schröder C, Dost A, Kordonouri O, et al. j.it.2019.04.012 Factors Contributing to Partial Remission in Type 1 Diabetes: Analysis Based 54. Audiger C, Rahman MJ, Yun TJ, Tarbell KV, Lesage S. The Importance of on the Insulin Dose-Adjusted HbA1c in 3657 Children and Adolescents From Dendritic Cells in Maintaining Immune Tolerance. J Immunol (2017) 198 Germany and Austria. Pediatr Diabetes (2017) 18(6):428–34. doi: 10.1111/ (6):2223–31. doi: 10.4049/jimmunol.1601629 pedi.12413 55. Rodriguez-Fernandez S, Murillo M, Villalba A, Perna-Barrull D, Cano-Sarabia 73. Glisic-Milosavljevic S, Wang T, Koppen M, Kramer J, Ehlenbach S, Waukau J, M, Gomez-Muñoz L, et al. Impaired Phagocytosis in Dendritic Cells From et al. Dynamic Changes in CD4+ CD25+ High T Cell Apoptosis After the Pediatric Patients With Type 1 Diabetes Does Not Hamper Their Tolerogenic Diagnosis of Type 1 Diabetes. Clin Exp Immunol (2007) 150(1):75–82. Potential. Front Immunol (2019) 10:2811. doi: 10.3389/fimmu.2019.02811 doi: 10.1111/j.1365-2249.2007.03475.x 56. Nieminen JK, Vakkila J, Salo HM, Ekstrom N, Harkonen T, Ilonen J, et al. 74. Garcia-Prat M, Á lvarez-Sierra D, Aguiló -Cucurull A, Salgado-Perandré s S, Altered Phenotype of Peripheral Blood Dendritic Cells in Pediatric Type 1 Briongos-Sebastian S, Franco-Jarava C, et al. Extended Immunophenotyping Diabetes. Diabetes Care (2012) 35(11):2303–10. doi: 10.2337/dc11-2460 Reference Values in a Healthy Pediatric Population. Cytom Part B Clin Cytom 57. Klocperk A, Petruzelkova L, Pavlikova M, Rataj M, Kayserova J, Pruhova S, (2019) 96(3):223–33. doi: 10.1002/cyto.b.21728 et al. Changes in Innate and Adaptive Immunity Over the First Year After the 75. Alpert A, Pickman Y, Leipold M, Rosenberg-Hasson Y, Ji X, Gaujoux R, et al. Onset of Type 1 Diabetes. Acta Diabetol (2020) 57(3):297–307. doi: 10.1007/ A Clinically Meaningful Metric of Immune Age Derived From High- s00592-019-01427-1 Dimensional Longitudinal Monitoring. Nat Med (2019) 25(3):487–95. 58. Jansen A, van Hagen M, Drexhage HA. Defective Maturation and Function of doi: 10.1038/s41591-019-0381-y Antigen-Presenting Cells in Type 1 Diabetes. Lancet (1995) 345(8948):491–2. 76. Shearer WT, Rosenblatt HM, Gelman RS, Oyomopito R, Plaeger S, Stiehm ER, doi: 10.1016/S0140-6736(95)90586-3 et al. Lymphocyte Subsets in Healthy Children From Birth Through 18 Years Frontiers in Immunology | www.frontiersin.org 22 February 2022 | Volume 13 | Article 825426 Gomez-Muñoz et al. Biomarkers for Autoimmune Diabetes Remission of Age. J Allergy Clin Immunol (2003) 112(5):973–80. doi: 10.1016/ Publisher’s Note: All claims expressed in this article are solely those of the authors j.jaci.2003.07.003 and do not necessarily represent those of their affiliated organizations, or those of 77. Max Andersen MLC, Hougaard P, Pörksen S, Nielsen LB, Fredheim S, the publisher, the editors and the reviewers. Any product that may be evaluated in Svensson J, et al. Partial Remission Definition: Validation Based on the this article, or claim that may be made by its manufacturer, is not guaranteed or Insulin Dose-Adjusted HbA1c (IDAA1C) in 129 Danish Children With endorsed by the publisher. New-Onset Type 1 Diabetes. Pediatr Diabetes (2014) 15(7):469–76. doi: 10.1111/pedi.12208 Copyright © 2022 Gomez-Muñoz, Perna-Barrull, Caroz-Armayones, Murillo, Rodriguez-Fernandez, Valls, Vazquez, Perez, Corripio, Castaño, Bel and Vives-Pi. This is an open-access article distributed under the terms of the Creative Commons Conflict of Interest: Unrelated to the work herein presented, MVP holds a patent Attribution License (CC BY). The use, distribution or reproduction in other forums is that relates to immunotherapy for T1D and is co-founder of Ahead Therapeutics permitted, provided the original author(s) and the copyright owner(s) are credited and SL, and SRF is part-time employed at Ahead Therapeutics S.L. The other authors that the original publication in this journal is cited, in accordance with accepted declare that the research was conducted in the absence of any commercial or academic practice. No use, distribution or reproduction is permitted which does not financial relationships that could be construed as a potential conflict of interest. comply with these terms. Frontiers in Immunology | www.frontiersin.org 23 February 2022 | Volume 13 | Article 825426