Early Detection of Peripheral Blood Cell Signature inChildren Developing b-Cell Autoimmunity at a Young AgeHenna Kallionpää,1 Juhi Somani,2 Soile Tuomela,1 Ubaid Ullah,1 Rafael de Albuquerque,1 Tapio Lönnberg,1

Elina Komsi,1 Heli Siljander,3,4 Jarno Honkanen,3,4 Taina Härkönen,3,4 Aleksandr Peet,5,6 Vallo Tillmann,5,6

Vikash Chandra,3,7 Mahesh Kumar Anagandula,8 Gun Frisk,8 Timo Otonkoski,3,7 Omid Rasool,1

Riikka Lund,1 Harri Lähdesmäki,2 Mikael Knip,3,4,9,10 and Riitta Lahesmaa1

Diabetes 2019;68:2024–2034 | https://doi.org/10.2337/db19-0287

The appearance of type 1 diabetes (T1D)-associatedautoantibodies is the first and only measurable param-eter to predict progression toward T1D in geneticallysusceptible individuals. However, autoantibodies indi-cate an active autoimmune reaction, wherein the im-mune tolerance is already broken. Therefore, there isa clear and urgent need for new biomarkers that predictthe onset of the autoimmune reaction preceding auto-antibody positivity or reflect progressive b-cell destruc-tion. Here we report the mRNA sequencing–basedanalysis of 306 samples including fractionated samplesof CD4+ and CD8+ T cells as well as CD42CD82 cellfractions and unfractionated peripheral blood mono-nuclear cell samples longitudinally collected fromseven children who developed b-cell autoimmunity(case subjects) at a young age and matched controlsubjects. We identified transcripts, including interleukin32 (IL32), that were upregulated before T1D-associatedautoantibodies appeared. Single-cell RNA sequencingstudies revealed that high IL32 in case samples wascontributed mainly by activated T cells and NK cells.Further, we showed that IL32 expression can be in-duced by a virus and cytokines in pancreatic isletsand b-cells, respectively. The results provide a basisfor early detection of aberrations in the immune system

function before T1D and suggest a potential role for IL32in the pathogenesis of T1D.

Family and sibling studies in type 1 diabetes (T1D) haveimplicated a firm genetic predisposition to a locus con-taining HLA class I and class II genes on chromosome6 suggesting a role for CD4+ as well as CD8+ T cells in T1Dpathogenesis (1–3). As much as 30–50% of the genetic riskis conferred by HLA class II molecules, which are crucial inantigen presentation to CD4+ T cells. Further, CD4+ cellsreactive to b-cell antigen peptides are found in peripheralblood and the pancreas and typically secrete the cytokineIFNg (4,5). CD4+ cells orchestrate adaptive immuneresponses, including that of antibody-secreting B cells aswell as cytotoxic CD8+ T cells. Indeed, circulating autoanti-bodies against b-cell antigens may appear years before theclinical onset. Further, a cytolytic CD4+ subtype mightdirectly contribute to target cell killing (6).

Although HLA class II is associated with the develop-ment of autoantibodies, HLA class I seems to be morestrongly linked to disease progression (7). Histologicalanalysis of pancreatic sections of cadaveric donors withT1D revealed that HLA class I is highly expressed in islets

1Turku Bioscience Centre, University of Turku and Åbo Akademi University, Turku,Finland2Department of Computer Science, Aalto University School of Science, Espoo,Finland3Children’s Hospital, University of Helsinki, and Helsinki University Hospital,Helsinki, Finland4Research Programs Unit, Diabetes and Obesity, University of Helsinki, Helsinki,Finland5Department of Pediatrics, University of Tartu, Tartu, Estonia6Children’s Clinic of Tartu, Tartu University Hospital, Tartu, Estonia7Research Programs Unit, Molecular Neurology and Biomedicum Stem Cell Centre,Faculty of Medicine, University of Helsinki, Helsinki, Finland8Department of Immunology, Genetics and Pathology, Uppsala University, Sweden9Folkhälsan Research Center, Helsinki, Finland

10Tampere Center for Child Health Research, Tampere University Hospital,Tampere, Finland

Corresponding author: Riitta Lahesmaa, riitta.lahesmaa@utu.fi

Received 17 April 2019 and accepted 10 July 2019

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db19-0287/-/DC1.

H.K., J.S., S.T., and U.U. contributed equally to this work.

H.K., M.K., and R.L. share senior authorship.

© 2019 by the American Diabetes Association. Readers may use this article aslong as the work is properly cited, the use is educational and not for profit, and thework is not altered. More information is available at http://www.diabetesjournals.org/content/license.

2024 Diabetes Volume 68, October 2019

GENETIC

S/G

ENOMES/P

ROTEOMIC

S/M

ETABOLOMIC

S

(8,9). Moreover, CD8+ cells are the most abundant cell typeduring insulitis (10), and the islets contain CD8+ cellsspecific for T1D autoantigens (11). Thus, the autoimmunecascade in T1D might be initiated by self-reactive CD4+

cells that activate B cells to produce autoantibodies thattarget the b-cells and unleash the cytotoxic activity of theautoreactive CD8+ cells. The environmental factors trig-gering and driving the autoimmunity in T1D are poorlydefined, but the disease has been associated with viralinfections (12), diet in early childhood (13), and reduceddiversity of gut microbiota (14).

Currently, the appearance of T1D-associated autoanti-bodies is the first and only measurable parameter topredict progression toward T1D in genetically susceptibleindividuals. Although the disease progression rate variesconsiderably, children with genetic HLA risk expressing atleast two T1D autoantibodies will very likely progress toclinical disease during the next 15 years (15). However,autoantibodies are poor prognostic markers for the timingof the clinical presentation of T1D. The appearance ofautoantibodies indicates an active autoimmune reaction,wherein the immune tolerance is already broken. There-fore, there is a clear and urgent need for new biomarkersthat predict the onset of the autoimmune reaction pre-ceding autoantibody positivity or reflect progressive b-celldestruction. Such markers would present a window forearly intervention aimed at complete disease prevention.Previously, we reported changes in whole-blood transcriptsand serum proteins before the detection of diabetes-associated antibodies in children who later progressed to T1D(16,17). Therefore, we hypothesized that a comprehensiveanalysis of the transcriptome of longitudinal cellular sam-ples including CD4+ and CD8+ T cells will lead to theidentification of new early biomarkers.

RESEARCH DESIGN AND METHODS

Study CohortSamples were collected as part of the DIABIMMUNE studyfrom Finnish (n = 10) and Estonian (n = 4) participants(Supplementary Table 1). The HLA-DR-DQ genotypes wereanalyzed as previously described (18). A total of 836 childrenwith HLA-DR-DQ risk allele were monitored and sampled at3, 6, 12, 18, 24, and 36 months of age. The study protocolswere approved by the ethics committees of the participatinghospitals, and the parents gave written informed consent.Autoantibodies against insulin (IAA), glutamic acid decar-boxylase (GADA), islet antigen-2 (IA-2A), and zinc trans-porter 8 (ZnT8A) were measured from serum with specificradiobinding assays (19). Islet cell antibodies (ICAs) wereanalyzed with immunofluorescence in autoantibody-positivesubjects. The cutoff values were based on the 99th per-centile in children without diabetes, which were 2.80relative units (RU) for IAA, 5.36 RU for GADA, 0.78 RU forIA-2A, and 0.61 RU for ZnT8A. The detection limit in theICA assay was 2.5 Juvenile Diabetes Foundation units(JDFU). A sample was considered seropositive when anyof the autoantibodies exceeded the thresholds.

Sample CollectionsAt each study visit, 8 mL blood was drawn in sodium-heparin tubes (368480, Vacutainer; BD Biosciences). Pe-ripheral blood mononuclear cells (PBMCs) were isolated byFicoll-Paque centrifugation (17-1440-03; GE Healthcare)and were suspended in RPMI-1640 medium (42401-018;Gibco) supplemented with 10% DMSO (cat. no. 0231,500 mL, Thermo Fisher Scientific), 5% Human AB Serum(cat. no. IPLA-SERAB-OTC; Innovative Research), 2mmol/L L-glutamine (G7513; Sigma-Aldrich), and 25 mmol/Lgentamicin (G-1397; Sigma-Aldrich). After overnightincubation at 280°C, samples were stored in liquid nitro-gen (2180°C). For fractionation, PBMC samples werethawed quickly in a 37°C water bath and quantitatedfor cell numbers and viability. On average, 90% of cellswere viable. Magnetic antibody-coupled beads were usedfor sequential positive enrichment of CD4+ and CD8+ cells(11331D and 11333D; Invitrogen). RNA was isolatedfrom the samples with an AllPrep kit (80224; QIAGEN),and quantity and quality were determined using a QubitRNA assay (Q32852; Invitrogen) and Bioanalyzer 2100(Agilent), respectively.

Bulk RNA Sequencing of PBMCs and Other FractionsAt least 80 ng total RNA was processed for RNA sequenc-ing (RNA-seq) with the TruSeq Stranded mRNA LibraryPrep kit (RS-122-2101; Illumina). The sequencing wascarried out with the Illumina HiSeq2500 instrument usingTruSeq v3 (23 100 base pairs [bp] chemistry). The averagesequencing depth was ;51 million reads. Quality controlwas performed using FastQC (version 0.10.0). All thesamples passed the quality criteria. The reads were alignedto the human reference transcriptome, GRCh37 assemblyversion 75, using TopHat (version 2.0.10) (20). Averagemapping percentage was 93. The concordant pairs per-centage was ;89. The aligned reads were counted withhtseq-count (HTSEq. 0.6.1; overlap mode of “intersection-strict”) (21). The read counts of genes were normalizedusing the trimmed means of the M values (TMM imple-mented in edgeR [22]). Coding, noncoding informationwere taken from Ensembl. Differential expression analyseswere conducted separately for coding and noncodinggenes, using edgeR (22). The variance of the data wasestimated using the trended dispersion method. A furtherfiltering step retained only those genes as differentiallyexpressed (DE) that had |median log2 fold change (FC)|.0.5 and had .65% samples across all individuals regu-lated in the same direction (i.e., up- or downregulated).These filtering steps were added to discard false positivesthat may arise due to the heterogeneity of the samplesresulting from normal variation, which is unrelated to T1Das well as to discard the outliers. A flowchart of the schemeof analysis has been shown in Supplementary Fig. 1.

Single-Cell RNA-seqThe concentrations of the PBMC samples varied from 0.55to 1.803 106 cells/mL. From each sample, we aimed at the

diabetes.diabetesjournals.org Kallionpää and Associates 2025

recovery of 5,000 single cells, loading ;9,000 cells on theChromium Controller using Single Cell 39 Solution v2reagents and following the manufacturer’s instructions(CG00052, Rev B; 10x Genomics). Single-cell RNA-seq(scRNA-seq) sample processing was carried out in threebatches on consecutive days using the same lot of reagentsand chips for all samples. The cDNA was further amplifiedusing a Veriti Thermal Cycler (Applied Biosystems/ThermoFisher Scientific), followed by cleanup (SPRIselect kit;Beckman Coulter). Finally, enzymatic fragmentation, endrepair, A-tailing, adaptor ligation, and PCR were performedto produce indexed libraries, which were sequenced withIllumina HiSeq 3000 (one sample per lane) using pairedend sequencing and 26 + 98 bp read-length configuration.The data were processed using the Cell Ranger pipeline,version 2.0.0, yielding on average 2,546 viable cells persample and 114,309 reads per cell.

The reads were aligned to the human reference genome(hg19) using STAR (23). The mean raw reads per cell variedfrom 57,000 to 200,000. Quality control analysis andfurther exploration were done using Seurat (24). Afterfiltering steps, 18,396 cells expressing 20,830 genes wereretained. For details on the filtering steps please see Sup-plementary Data. The data were normalized using Seurat’sdefault. Highly variable genes were selected for principalcomponent analysis. The top 20 PCs were used in the graph-based clustering. To identify marker genes for each cluster,cells of a single cluster were compared with the cells of allother clusters combined. A gene was considered a marker ofa cluster if it was expressed in at least 25% of the cells ofeither of the two groups and the logFC between the clusterand all other clusters was at least 0.25.

For trajectory analysis, the pooled cells were ordered inpseudotime (i.e., placed along a trajectory correspondingwith a type of biological transition, such as differentiation)using Monocle 2 (25). The analysis was performed on cellsspecifically from CD4+ and CD8+ T-cell clusters. For thedetails on the trajectory analyses, please see Supplemen-tary Data.

RT-PCR AnalysisFor PBMC samples, 50 ng total RNA was treated withDNaseI (Invitrogen), and cDNA was synthesized with theTranscriptor First Strand cDNA Synthesis Kit (Roche LifeScience). For isoform-specific (IL32a, b, and g) assay,quantitative PCR (qPCR) analysis was performed in tripli-cate runs using SYBR Select Master Mix (Applied Biosys-tems). DCt values was calculated relative to EF1a. For CD4+

T cells and pancreatic islets, RNA was isolated using theRNeasy Mini Kit (74106; QIAGEN) and RNeasy Plus MiniKit (74134; QIAGEN), respectively. Purified RNA was trea-ted with DNaseI and cDNA was synthesized with Super-Script II Reverse Transcriptase (18064014; Invitrogen).For the detection of global IL32, qPCR reactions wererun using a custom TaqMan Gene Expression Assay reagent(AJ5IQA9; Thermo Fisher Scientific) in duplicate and intwo separate runs. DCt values were calculated relative to

GAPDH. The amplification was monitored with the Quant-Studio 12K Flex Real-Time PCR System under the followingPCR conditions: 10 min at 95°C, followed by 40 cycles of15 s at 95°C and 60 s at 60°C, and analysis with QuantStudiosoftware on Thermo Fisher Cloud.

For EndoC-bH1 cell data, cDNA was synthesized usingthe Maxima First Strand cDNA Synthesis Kit (ThermoFisher Scientific) as per the manufacturer’s recommenda-tions. All reactions were performed in duplicates on atleast three biological replicates. Cyclophilin-A was used asan endogenous control. Primer sequences are presented inSupplementary Table 2.

ELISA

To measure secreted IL32 levels, we used IL-32 DuoSetELISA (DY3040-05 and DY008; R&D Systems) followingthe manufacturer’s instructions.

Intracellular Staining and Flow CytometryThe cells were fixed for 10 min in Fix Buffer I (557870; BDBiosciences), followed by 45 min permeabilization usingice-cold permeabilization buffer III (558050; BD Bioscien-ces). The cells were stained using allophycocyanin-conjugated IL32⍺ antibody (IC30402A; R&D Systems)and FITC-conjugated IFNg antibody (MHCIFG01; Invitro-gen) in PBS containing 0.5% FCS. The data were acquiredin BD Fortessa and analyzed using FlowJo (version 10.4.2).

EndoC-bH1 Cell CultureThe EndoC-bH1 human b-cell line was obtained fromUnivercell Biosolution S.A.S., Toulouse, France. The cellswere cultured as previously described (26). EndoC-bH1cells were stimulated with IL32g either alone (100 ng/mL;R&D Systems) or in combination with a cocktail of IL1b(5 ng/mL; R&D Systems) and IFNg (50 ng/mL; R&DSystems) for 24 h. RNA samples were collected at theend of each treatment and analyzed by RT-qPCR.

Human CD4 T-Cell Isolation and CulturingCD4+ T cells were isolated from cord blood collected fromneonates born in Turku University Hospital and werecultured in Iscove’s modified Dulbecco’s medium contain-ing 1% AB serum in absence (Th0) or presence (Th1) of2.5 ng/mL IL12 (R&D Systems). Cells were activated withplate-bound CD3 (0.5 mg/well of a 24 well-plate) andsoluble CD28 (0.5 mg/mL), both from Immunotech,with or without 50 ng/mL rIL32g (R&D Systems).12 ng/mL IL2 was added at 48 h. For IFNg neutralization,anti-IFNg antibody (MAB285, 10 mg/mL; R&D Systems)was used. For reactivation, cells were treated with 5 ng/mLphorbol myristic acid (Calbiochem) and 0.5 pg/mL iono-mycin (Sigma-Aldrich) for 5 h.

Human Pancreatic Islets and Their Infection WithCoxsackie B VirusHuman islets were isolated from pancreases obtained frombrain dead organ donors and purified by handpicking to

2026 Early Signs of T1D Autoimmune Reaction Diabetes Volume 68, October 2019

a purity of .90%. Islet culturing and virus infection withCoxsackie B virus-1 (CBV-1-7-10796 [CBV-1-7]) was per-formed as previously described (27). Islets were collected atthe day 4 time point, and RNA was extracted using theRNeasy Plus Mini Kit or the AllPrep DNA/RNA Mini Kit(QIAGEN). For RNA-seq, 100 ng total RNA from threedonors was used for library preparation according to Illu-mina TruSeq RNA Sample Preparation v2 Guide (part no.15026495). The high quality of the libraries was confirmedwith the Agilent Bioanalyzer 2100 and Qubit FluorometricQuantitation (Life Technologies). The libraries were pooledin two pools and run in two lanes on the Illumina HiSeq2500 instrument using 2 3 100 bp.

Data and Resource AvailabilityAll the raw data will be deposited to European Genome-phenome Archive for access. The study does not involveany noncommerical reagents or tools.

RESULTS

Fractionation of PBMC Sample Into CD4+, CD8+, andCD42CD82 Cellular Subsets Reveals Distinct andOverlapping Gene Expression SignaturesWe performed RNA-seq of 306 longitudinal samples in-cluding unfractionated PBMCs, as well as CD4-enriched(CD4+), CD8-enriched (CD8+), and CD4 and CD8 cell–depleted (CD42CD82) cell fractions from seven case-controlpairs (Table 1). The seven case children who developedT1D-related autoantibodies (Aab+) were selected from theDIABIMMUNE birth cohort (18), where HLA-susceptiblechildren are sampled at 3–36 months of age (Fig. 1A). Allseven children developed T1D-associated autoantibodiesby the age of 2 years (Table 1), and four of them developedclinical T1D between the ages of 2.4 and 3.7 years. For eachcase subject, an autoantibody-negative control child wasmatched for sex, date and place of birth, and HLA-conferred risk category.

The samples clustered according to the cell fraction (Fig.1B), and the clustering was not affected by case-controlstatus or sampling age, indicating that cell fraction–specific differences dominated over variation derived fromother factors (Supplementary Fig. 2A and B). When CD4+,CD8+, and CD42CD82 samples from control subjects werecompared with the unfractionated PBMC samples (also

referred to as a fraction henceforth), 889, 399, and 1,002genes were DE specifically in CD4+ (e.g., CD28, CTLA4),CD8+ (e.g., CD8A, CD8B, KLRK1), and CD42CD82 (e.g.,IL1A, IL1B, IL6) fractions, respectively (Fig. 1C and Sup-plementary Table 3). CD4+ and CD8+ fractions shared1,815 DE genes, of which 1,803 genes (99%) were con-cordant (either up or down in both fractions) (Supple-mentary Fig. 2C and Supplementary Table 3). In summary,fractionation of the PBMC population based on the T-cellphenotype allowed improved detection of DE genes andenabled identification of cell subset–specific gene expres-sion signatures.

RNA-seq Analysis Identifies Transcriptomic ChangesAssociated With b-Cell AutoimmunityComparison of case samples with their respective controlsidentified 51, 69, 143, and 85 genes as DE (false discoveryrate ,0.05) in CD4+, CD8+, CD42CD82, and PBMC frac-tions, respectively (Supplementary Table 4), with a total of278 unique DE genes in one or more fractions (Fig. 2A). Sixgenes, AMICA1, BTN3A2, IL32, RPSAP15, RPSAP58, andWASH7P, were upregulated in the case subjects in all fourfractions (Fig. 2A). Only 16% of the DE genes have pre-viously been reported as DE in genetically susceptiblechildren with prediabetes, using microarrays (16,28,29)or RT-PCR (30–32), confirming dysregulation of these genesin children progressing to T1D. Besides protein-codinggenes, 54 noncoding genes, including 3 antisense, 2 senseintronic, 7 enhancer, and 18 promoter-associated lncRNAs,were DE. To our knowledge, none of these lncRNAs havebeen linked to the etiology of T1D (16,28–32).

Hierarchical Clustering Identifies Coregulated GeneExpression Clusters Associated With T1DAutoimmunityGene- and sample-wise hierarchical clustering for each cellfraction, including PBMCs, identified a cluster upregulatedin the case samples in all four fractions (Fig. 2B andSupplementary Fig. 3A–D). Interestingly, this cluster con-sistently contained IL32 and BTN3A2, along with otherfraction-specific genes (Fig. 2C). In the CD8+ fraction,expression of a distinct cluster, including IFNG, was lowerin most of the case samples than in control samples(Supplementary Fig. 3B). Surprisingly, in the PBMC

Table 1—Summary of the case and control children sampled at the age of 3–36 months

Case no. Sex Seroconversion age* First autoantibodies Age at T1D diagnosis Matched control no.

Case 1 Female 12 months IAA, GADA 3.2 years Control 1

Case 2 Male 12 months IAA — Control 2

Case 3 Male 18 months IAA, ICA 3.7 years Control 3

Case 5 Female 24 months IAA, IA-2A, ZnT8A, ICA 2.6 years Control 5

Case 9 Male 18 months IAA, GADA, ICA — Control 9

Case 10 Male 12 months IAA, GADA — Control 10.1, control 10.2

Case 11 Female 18 months GADA 2.4 years Control 11

For further details, see Supplementary Table 1. *First detection of T1D-associated autoantibodies.

diabetes.diabetesjournals.org Kallionpää and Associates 2027

fraction, we detected case-specific upregulation of a cluster,including insulin (INS), glucagon (CGC), and regulin 1a(REG1A), transcripts (Supplementary Fig. 3D), which arepredominantly expressed in the pancreas.

To explicitly define coregulated genes in these clusters,we calculated Euclidean distances for IL32 (in each frac-tion), IFNG (in CD8+ fraction), and INS (in PBMC fraction)and considered the genes with a median Euclidean dis-tance ,2.5 across all case-control pairs to be cocluster-ing with the gene of interest (Supplementary Table 5).In three of the four fractions, the IL32 cluster includedBTN3A2, AMICA1, LARS, and RSU1 (Fig. 2C). IL32,AMICA1, and BNT3A2 show concerted gene expression

profiles in CD4+ samples (Fig. 2D). In at least two offour fractions, this cluster also comprised TRBV4-1,TMEM14C, UROS, WASH7P, BTN3A3, CARD8, CCDC167,and LINC01184. The profile of these and other interestinggenes is shown in Supplementary Fig. 4. Upon examinationof the overrepresented transcription factor binding siteson the promoters of IL32 cluster genes, the V$IK_Q5_01motif bound by Ikaros (IKZF1) was revealed to be amongthe enriched transcription factor binding sites shared inboth the CD4+ and PBMC fractions (Supplementary Table5). IKZF1 has been genetically associated with T1D (33).The T1D-associated risk allele rs10272724 (T) increasesIKZF1 transcript level (34).

Figure 1—Fractionation of PBMC sample into CD4+, CD8+, and CD42CD82 cellular subsets reveals distinct and overlapping geneexpression signatures. A: Outline of the sample collection and cell fractionation. B: t-SNE (t-distributed stochastic neighbor embedding)visualization of the log2-transformed expression data (without any filtering steps) colored according to cell fraction information.C: Number ofDE genes when CD4+, CD8+, and CD42CD82 fractionated samples were compared with their original PBMC aliquots. The functionallyimportant fraction-specific upregulated genes are highlighted in red. Analysis was restricted to healthy control subjects only. For the genelists, see Supplementary Table 3.

2028 Early Signs of T1D Autoimmune Reaction Diabetes Volume 68, October 2019

IFNG cluster of the CD8+ cells included TBX21 (codesfor TBET), BHLHE40, and ZEB2, transcription factorsexpressed in CD8+ T cells (35), as well as NKG7, OASL,and KLRD1 (Supplementary Table 5). ZEB2 has beenreported to drive terminal effector CD8+ cell differentia-tion together with T-bet (36). In the PBMC fraction, GCGand REG1A were coregulated with INS (SupplementaryTable 4 and Supplementary Fig. 5).

Transcriptional Changes Preceding the Appearance ofT1D-Related Autoantibodies Are Enriched in the CD8+

T-Cell FractionTo identify changes that occur immediately before the firstdetection of T1D-related autoantibodies (i.e., seroconver-sion), we performed a separate differential expressionanalysis for the samples drawn at most 12 months beforeseroconversion. Altogether, 121 coding and noncodinggenes were DE in case subjects compared with matchedcontrol subjects (Supplementary Table 4 and Supplemen-tary Fig. 6). Notably, more than half of these (58%) weredetected only in the CD8+ fraction. Besides IL32, only twoother genes were common to all fractions, RPSAP58 andRPSAP15—both being the pseudogenes with unknown

functions with very similar expression profiles (Supple-mentary Fig. 4M–T).

Higher IL32 expression in case subjects was validatedusing qRT-PCR. Interestingly, genes encoding all threemajor isoforms (IL32 a, b, and g) were upregulated inPBMC samples in all the case children at each of the timepoints, including 3 months (Fig. 3A and Supplementary Fig.7). Among these isoforms, the gene encoding IL32g wasexpressed at the highest level, followed by IL32b and IL32a.

scRNA-seq Identifies T and NK Cells as the IL32-HighPopulationTo specify the cell populations responsible for the IL32 andINS signatures, we performed scRNA-seq on four selectedcase and their nearest matched control PBMC sampleswhere the expression of IL32 or INS was high (or low)based on the bulk RNA-seq data (Supplementary Table 6).Unsupervised clustering of 18,396 single cells from alleight PBMC scRNA-seq runs identified 13 clusters (Fig.3B and Supplementary Fig. 8). The 2 largest clustersexpressing high CCR7 were merged as one cluster of naiveT cells, reducing the number of clusters to 12. Clustersnamed as “RGCC+ T cells,” “CD62L+ T cells,” and “activated

Figure 2—RNA-seq analysis identifies transcriptomic changes associated with b-cell autoimmunity. A: Number and overlap of DE genesbetween case and control subjects identified in cell fractions analyzed. Genes shared between all four fractions are highlighted. B: Heat mapof the genes DE in CD4+ T cells between the case and control subjects. Values are presented as log2FC (truncated between [22, 2]) betweeneach case-control pair at each time point (3–36 months) and standardized to the mean of each gene. Genes coregulated with IL32 (,2.5Euclidean distance) are marked with red box and text. Sample collection time with respect to (w.r.t) seroconversion, sample pairinginformation, and clinical status have been indicated with colors on top of the heat map. “Before/After SC” informs whether the case samplewas collected before or after seroconversion. “Pair Info” provides the case-control pair information. The “SC / T1D” annotation indicateswhether the case subject has progressed to clinical T1D diagnosis (T1D) or not (SC). C: Number and overlap of IL32 coclustered genes inindicated cell fractions. Genes regulated at least in two fractions are highlighted. D: Profiles of IL32, AMICA1, and BNT3A2 in CD4+ samples,presented in log2 RPKM (reads per kilo base per million mapped reads) scale. For individual profiles, see Supplementary Fig. 4. The case-control pairs are grouped according to the diagnosis of the case subjects. T1D, case subject has been diagnosed with clinical T1D; SC, casehas seroconverted to autoantibody positivity.

diabetes.diabetesjournals.org Kallionpää and Associates 2029

Th cells” expressed lower levels of CCR7. Activated CD8+

T cells cluster expressed high levels of CD8A and CD8B aswell as NKG7, and two separate clusters of CD8+ T cellsexpressing either granulysin or granzyme A were observed(“activated GNLY+ CD8+ T cells” and “activated GZMA+

CD8+ T cells,” respectively). A subcluster of activatedGZMA+ CD8+ cells had higher expression of cell-cycle genes(e.g., STMN1, TUBA1B) and was named “activated pro-liferating GZMA+ CD8+ T cells.” An NK cell cluster waspositive for expression of CD56, NKG7, and GNLY andnegative for CD8A and CD3E. A B-cell cluster was identifiedby the expression of MS4A1, CD79A, and CD79B, whereasthe monocyte/dendritic cells (DC) cluster was composed of

cells expressing CD14 or FCGR3A, LYZ, and TYROBP. In-terestingly, the expression of many HLA class II moleculeswas as high in B cells as in monocytes, suggesting highantigen-presentation potential.

The contribution of different case or control samples tothe cells in a given cellular population (cluster) varied fromcluster to cluster (Supplementary Figs. 9 and 10A and B).The naive T-cell cluster was dominated by the cells from thecontrol samples (P , 0.05) whereas the monocyte/DCcluster had more cells from case subjects (P , 0.005)(Supplementary Fig. 10B). Case 9, with the highest IL32expression levels in the bulk RNA-seq data, dominated theCD62L+ T-cell cluster, activated NK cell cluster, and, most

Figure 3—scRNA-seq of PBMCs identifies T and NK cells as IL32 high populations. A: Expression of IL32g isoform in longitudinal PBMCsamples of case subjects and their control subjects (n = 7 + 7), assayed by qRT-PCR (for a and b isoforms, see Supplementary Fig. 7). B:t-SNE (t-distributed stochastic neighbor embedding) clusters from the pooled data from all scRNA-seq samples (4 case and 4 controlsubjects [in total 18,396 cells]). Clusters are named according to the expression of classical marker genes, such as CD8A (for details andmarker gene list, see Supplementary Fig. 8; for contribution of each sample per cluster, refer to Supplementary Figs. 9 and 10).C: Expressionof IL32 in the 12 cell clusters (natural logarithm transformation with addition of 1). For case-control comparison, please see SupplementaryFig. 11. D–F: Trajectories emerging when using the data from CD4+ cells and the precursor cells. G-I: Trajectories emerging when using thedata from CD8+ and the precursor cells. Here, “precursor cells” refer to cells from the naive and RGCC+ T-cell clusters. For the trajectoryanalysis of all the cells from all clusters as well as the breakdown of each individual cluster, see Supplementary Fig. 12. In D and G, cells arecolored based on the contributions from different t-SNE clusters. In E and H, cells are colored by case (orange) or control (gray) status. In Fand I, cells are colored by the intensity of IL32 expression (log10 transformation with addition of 0.1). Act., activated; prolif., proliferating.

2030 Early Signs of T1D Autoimmune Reaction Diabetes Volume 68, October 2019

clearly, activated and proliferating GZMA+ CD8+ T cellclusters (Supplementary Fig. 10B). Conversely, control chil-dren 5 and 9 seemed to dominate the cluster of developingT cells expressing pre–T-cell receptor PTCRA, suggesting thepresence of immature T cells in those samples.

Insulin, glucagon, and REG1A expression was notdetected even in the INS-high samples of cases 5 and 9,leaving the origin of these transcripts in bulk RNA-seq asan open question. In contrast, IL32 expression was clear,and as expected, it was explicitly overexpressed in the casesamples (Supplementary Fig. 11). IL32 was expressed ata very low level in monocyte/DC, B cell, and developingT cell clusters; however, it was expressed at higher levels byboth the T cells and the NK cells (Fig. 3C).

To further define the relationship of IL32 expression andT-cell activation status, we performed separate trajectoryanalyses for the CD4+ and CD8+ T cells. The less activatedprecursor populations (naive and RGCC+ T cells), whichdetect CD4 and CD8 encoding transcripts in low abundances,were used as a starting point for the trajectory analyses.The results revealed three major cellular branches (I–III) inthe data in both CD4+ and CD8+ T cells (Fig. 3D–I). BranchI consisted mainly of naive T cells, among which cells from

the control samples were enriched (Fig. 3E and H andSupplementary Fig. 12). In contrast, the highest levels ofIL32 were expressed by cells close to the end points ofbranches II and III, corresponding to more advanced stagesof differentiation (Fig. 3F and I and Supplementary Fig. 12).

IL32 and IFNg Are Coexpressed by Th1 CellsTo further study IL32 expression, we measured intracel-lular IL32 expression at the protein level in CD4+ T cellsisolated from human umbilical cord blood. Cells wereeither activated through CD3/CD28 in the absence ofcytokine (Th0) or were differentiated toward a Th1 celllineage for 72 h. IL32 was induced upon activation and,unlike IFNg, was expressed in both Th0 and Th1 cells (Fig.4A). Interestingly, in Th1 cells, most IFNg-producing cellswere also positive for IL32 (Fig. 4A and Supplementary Fig.13A) and the proportions of IL32-positive cells and the percell IL32 levels were higher in IFNg-producing Th1 cellsthan in Th0 cells (Fig. 4B and C). Furthermore, neutral-ization of IFNg significantly reduced IL32 secretion by Th1cells (Fig. 4D), confirming that IFNg positively regulatesIL32 expression. IL32 expression was also induced by IL32itself in Th1 cells, both at the RNA level (Fig. 4E) and in the

Figure 4—Virus- and cytokine-induced IL32 expression by pancreatic b-cells. A: Representative FACS dot plots showing IFNg and IL32double staining in Th0 and Th1 polarizedCD4+ cells. Staining controls and two other replicates are shown in Supplementary Fig. 13A. PercentIL32-positive cells as well as median fluorescence intensity (MFI) data (mean 6 SD) from all the three replicates are shown in B and C,respectively. Statistical significance was determined by paired two-tailed t test. D: IL32 secretion in culture supernatant as measured byELISA. Cells were cultured in Th0/1 condition for 72 h in the presence (+) or absence (2) of anti-IFNg. The expression plotted is relative to Th0(2). Statistical significance was determined by paired two-tailed t test. E: IL32 expression in nonpolarized Th0 cells and cells differentiated toTh1 for 72 h in the presence (+) or absence (2) of IL32g as measured by the Taqman assay. The expression is calculated relative to EEF1A.Statistical significance was determined by unpaired two-tailed t test. F: IL32 secretion in culture supernatant as measured by ELISA. Cellswere cultured in Th0/1 condition for 7 days in the presence (+) or absence (2) of IL32g, followed by washing and restimulation by phorbolmyristic acid and ionomycin for 48 h. The expression plotted is relative to Th0 (2). Statistical significancewas determined by paired two-tailedt test.G: Expression of the TNFA and IL6 or IL8 and IL32 geneswhen the EndoC-bH1 cells were stimulated with IL32g alone or in combinationwith other inflammatory cytokines for 24 h. The fold change is calculated compared with nontreated cells. The results shown here are fromfour independent biological replicates (mean6 SD). Statistical significance for the effects on IL32 expression was determined by paired two-tailed t test. ns, not significant.H: IL32 expression asmeasured in an RNA-seq experiment where pancreatic islets were infectedwith CBV1-7.Statistical significance was determined by edgeR. I: IL32 expression in virus-infected pancreatic islets as measured by RT-qPCR Taqmanassay. The expression is calculated as 2^2(dCt). The statistical significance is determined by paired two-tailed t test. RPKM, reads per kilobase per million mapped reads. *P , 0.05; **P , 0.01; ***false discovery rate ,0.001.

diabetes.diabetesjournals.org Kallionpää and Associates 2031

culture supernatant upon 48-h restimulation after 7 daysof polarization in Th1 condition (Fig. 4F).

Pancreatic b-Cells Can Express IL32 in Response toCytokine Stimulation and Viral InfectionTo study how the elevated IL32 levels may influence b-cellfunction, we treated human EndoC-bH1 b-cell line for24 h with recombinant IL32g either alone or in combina-tion with the proinflammatory cytokines IL1b and IFNg.In agreement with earlier published data on pancreaticductal cancer cell lines (37), IL1b and IFNg significantlyinduced IL32 expression in human EndoC-bH1 cells (Fig.4G). However, addition of IL32g did not further enhance1) the IL1b- and IFNg-induced IL32 expression; 2) theexpression of inflammatory cytokines TNFA, IL6, and IL8(Fig. 4G); or 3) the expression of ER stress marker genes(ATF3, ATF4, ATF6, HSPA5, CHOP, and sXBP1) (Supple-mentary Fig. 13B) in EndoC-bH1 cells. Furthermore, theIL32g treatment did not affect the expression of b-cell–specific genes, such as INS, MAFA, and PDX1 (Supplemen-tary Fig. 13C). These results suggest that, while IL32 doesnot appear to directly affect the survival or the differen-tiation status of the b-cells, b-cells actively contribute toinflammation in the islets by secreting IL32 upon stimu-lation by cytokines.

Coxsackie B viruses are b-cell trophic viruses that havebeen linked to the development of T1D (38–43). To studythe possible trigger of IL32 expression in b-cells, weinfected purified human pancreatic islets of three cadavericdonors with CBV1-7 strain. Infection by the virus led tothe induction of IL32 expression in the islets (Fig. 4H).We further validated this finding in the three islet sam-ples used for RNA-seq as well as one additional isletsample using qRT-PCR assays and found a consistentincrease in the IL32 expression upon CBV1-7 infection(Fig. 4I). Taken together, these results suggest that upona viral infection (Fig. 4H and I) or a cytokine rush (Fig.4H), b-cells may upregulate IL32 secretion, contributingto inflammation.

DISCUSSION

We identified a panel of novel molecular players detectedearly in children who developed T1D-associated autoanti-bodies or even the clinical disease at a young age. Since theimmunological changes related to T1D are known to bestrongest among the T1D cases diagnosed at an early age(44), focusing on this age-group should enhance the possi-bility of detecting aberrations in the immune system pre-disposing to the disease. In this study, unbiased RNA-seqof CD4+ and CD8+ cells revealed many T1D-associated DEtranscripts not previously reported. Analysis of the PBMCpopulation offers an excellent overview of stable gene ex-pression patterns but, at the same time, appears to masksome of the subtle fraction-specific changes. Such changesincluded upregulation of CD52 detected only in the CD4+ cellfraction and downregulation of the IFNG and associatedtranscription factors ZEB2, TBX21, and ZNF683 detected

specifically in the CD8+ cells. Further studies are needed tounderstand whether at-risk children have defects in formu-lating effector CD8+ response or their effector CD8+ cellshave homed to the sites of inflammation in the pancreas.

We selected IL32 as our candidate for functional studiesbecause it has not been linked to seroconversion before, itis easy to measure with available assays from clinicalsamples, and as a secreted molecule it can potentiallyaffect the function of several cell types in paracrine andsystemic fashion. Increased expression of IL32 in casesubjects across many cell types before seroconversionsuggest that IL32 is a critical member of the immunologicalsignature characteristic for children developing b-cellautoimmunity.

IL32 is expressed by many immune and epithelial cellsand has been described to be proinflammatory (45). How-ever, to our knowledge, it has not been associated withhuman b-cell autoimmunity. In contrast, IL32 is down-regulated in CD4+ T cells from recently diagnosed adultT1D patients (46), which, along with our findings, suggestsdynamic changes in immune cell signaling during thepathogenesis of the disease. On the other hand, IL32overexpression was observed in synovial biopsies ofpatients with rheumatoid arthritis (47), in inflamed mu-cosa of inflammatory bowel disease patients (48), and inthe serum of myasthenia gravis patients (49), indicatinga connection between IL32 and autoimmunity in general.In T cells, IL32 is induced by T-cell activation, and itmodulates human CD4+ T-cell effector function by pro-moting Th1 and Th17 responses (50). Both Th1 and Th17cells have been linked to T1D pathogenesis in both humanand mouse (51). The IL32 gene has been identified only inhigher mammals, excluding rodents. Nonetheless, humanIL32g transgenic mice exhibit impaired glucose toleranceand increased levels of IFNg and other proinflammatorycytokines in the pancreas, as well as accelerated streptozotocin-induced experimental T1D (52). No specific cell-surfacereceptor for IL32 has been identified, but it may actthrough cell-surface integrins or proteinase-3 (53).

Our results showed that IL32 was often coregulatedwith genes previously linked to autoimmunity. For exam-ple, the BTN3 gene cluster resides in the extended MHCclass I locus. Further, BTN3 genes have been associatedwith T1D in a genetic screen, especially in the case ofBTN3A2 (54). AMICA1 is a plasma membrane proteininvolved in lymphocyte migration through its interactionwith Coxsackie-adenovirus receptor (CAR) expressed inepithelial cells and has been associated with multiplesclerosis (55). An analogous scenario could be envisagedfor T1D: CAR is expressed by the pancreatic islet cells,including b-cells (42), and its expression is elevated inautoantibody-positive individuals and patients with T1D(56), suggesting that it might help recruit T cells to theislets. Interestingly, the findings point to human-specificphenomena not detectable in mouse models, as IL32 andthe BTN3 protein family are not encoded by the mousegenome.

2032 Early Signs of T1D Autoimmune Reaction Diabetes Volume 68, October 2019

The strength of our study is that the children studiedhere comprise a homogeneous population with the earlyappearance of T1D-associated autoantibodies. Increasingevidence suggests that T1D can be subdivided into differ-ent phenotypes, e.g., characterized by age-dependent B-cellinfiltration in the pancreas (57), defect in Coxsackievirus-induced antibody response in children with early insulinautoimmunity (58), or rapid versus slow progression toclinical disease (59). Thus, our results may not apply to“late progressors,” adolescents, and adults. Although theanalysis of the global transcriptome of T-cell subsets ofchildren with prediabetes over the period of seroconver-sion is unique, a limitation of the current study is theanalysis of only seven Aab+ children. The results of thisstudy need to be validated and expanded in a larger cohortof children with prediabetes but serve as a starting pointfor better understanding of immunological changes pre-ceding the clinical onset of the disease. In the future, weare interested in addressing whether our findings ona cellular level are reflected also in IL32 levels in plasmaas well as studying whether IL32 alone or in combinationwith other identified molecules would have sufficientsensitivity and specificity as an early indicator for T1D.

Acknowledgments. The authors are grateful to the families for theirparticipation in the DIABIMMUNE study. The DIABIMMUNE study group is ac-knowledged for excellent collaborations, work with the families, and collection ofthe samples for the study. Marjo Hakkarainen, Sarita Heinonen, Päivi Junni, andElina Louramo (Turku Bioscience Centre, University of Turku and Åbo AkademiUniversity, Turku, Finland) are acknowledged for skillful assistance in thelaboratory. Next-generation sequencing was performed at the Finnish FunctionalGenomics Centre (FFGC), Turku, Finland, part of the Biocenter Finland network.The authors thank Satu Mustjoki and her team at University of Helsinki for advice indesigning scRNA-seq experiments and Riina Kaukonen at the FFGC for the samplepreparation.Funding. This work was financially supported by the JDRF; the Academy ofFinland (AoF) Centre of Excellence in Molecular Systems Immunology andPhysiology Research (SyMMyS) 2012–2017 (grant no. 250114); the AoF Person-alized Medicine Program (grant no. 292482); AoF grants 294337, 292335,319280, and 314444; the Sigrid Jusélius Foundation; the Diabetes ResearchFoundation (Diabetestutkimussäätiö); the Novo Nordisk Foundation InnovativeMedicines Initiative 2 Joint Undertaking under grant agreement no. 115797(INNODIA). This Joint Undertaking receives support from the Union’s Horizon2020 research and innovation programme and EFPIA, JDRF, and The Leona M. andHarry B. Helmsley Charitable Trust. The DIABIMMUNE study was supported by theEuropean Union Seventh Framework Programme (grant no. 202063). T.L. wassupported by the AoF (311081).Duality of Interest. No potential conflicts of interest relevant to this articlewere reported.Author Contributions. H.K., S.T., and U.U. were responsible for theinterpretation of the results. J.S. conducted bioinformatic analyses. H.K., J.S., S.T.,and U.U. drafted the manuscript. H.K., J.S., and U.U. prepared the figures. H.K.was responsible for supervising E.K. R.d.A., E.K., and O.R. were responsible for theisoform-specific IL32 RT-PCR assay and the intracellular IL32 staining in T cellsand interpretation of the results. T.L. provided expertise in scRNA-seq studydesign, sample and data analysis, and interpretation of the results. H.S., J.H., T.H.,A.P., and V.T. were responsible for sample collection, sample storage, and furtherclinical information of the children. V.C. and T.O. carried out the experiments andinterpreted the results of the studies in pancreatic b-cells. M.K.A. and G.F. were

responsible for experiments on virus-infected pancreatic islets. R.Lu. was re-sponsible for study design, cell fractionation, sample analysis, and data pro-duction. H.L. was responsible for computational data analysis, interpretation of theresults, editing the manuscript, and supervising J.S. M.K. was responsible for theDIABIMMUNE study design, sample collection, sample storage, clinical informationfor the children, directing of the clinical study, interpreting the results, and editingthe manuscript. R.La. was responsible for study design, sample and data analysis,interpretation of the results, writing the manuscript, and supervision of the study.All authors contributed to the final version of the manuscript. H.L. and R.La. are theguarantors of this work and, as such, had full access to all the data in the study andtake responsibility for the integrity of the data and the accuracy of the dataanalysis.

References1. Todd JA, Bell JI, McDevitt HO. HLA-DQ beta gene contributes to suscep-tibility and resistance to insulin-dependent diabetes mellitus. Nature 1987;329:599–6042. Nejentsev S, Howson JM, Walker NM, et al.; Wellcome Trust Case ControlConsortium. Localization of type 1 diabetes susceptibility to the MHC class I genesHLA-B and HLA-A. Nature 2007;450:887–8923. Todd JA. Etiology of type 1 diabetes. Immunity 2010;32:457–4674. Babon JA, DeNicola ME, Blodgett DM, et al. Analysis of self-antigenspecificity of islet-infiltrating T cells from human donors with type 1 diabetes. NatMed 2016;22:1482–14875. Delong T, Wiles TA, Baker RL, et al. Pathogenic CD4 T cells in type 1 diabetesrecognize epitopes formed by peptide fusion. Science 2016;351:711–7146. Takeuchi A, Saito T. CD4 CTL, a cytotoxic subset of CD4+ T cells, theirdifferentiation and function. Front Immunol 2017;8:1947. Lipponen K, Gombos Z, Kiviniemi M, et al. Effect of HLA class I and class IIalleles on progression from autoantibody positivity to overt type 1 diabetes inchildren with risk-associated class II genotypes. Diabetes 2010;59:3253–32568. Foulis AK, Farquharson MA, Hardman R. Aberrant expression of class II majorhistocompatibility complex molecules by B cells and hyperexpression of class Imajor histocompatibility complex molecules by insulin containing islets in type1 (insulin-dependent) diabetes mellitus. Diabetologia 1987;30:333–3439. Richardson SJ, Rodriguez-Calvo T, Gerling IC, et al. Islet cell hyperexpressionof HLA class I antigens: a defining feature in type 1 diabetes. Diabetologia 2016;59:2448–245810. Willcox A, Richardson SJ, Bone AJ, Foulis AK, Morgan NG. Analysis of isletinflammation in human type 1 diabetes. Clin Exp Immunol 2009;155:173–18111. Coppieters KT, Dotta F, Amirian N, et al. Demonstration of islet-autoreactiveCD8 T cells in insulitic lesions from recent onset and long-term type 1 diabetespatients. J Exp Med 2012;209:51–6012. Rodriguez-Calvo T, Sabouri S, Anquetil F, von Herrath MG. The viral paradigmin type 1 diabetes: who are the main suspects? Autoimmun Rev 2016;15:964–96913. Virtanen SM. Dietary factors in the development of type 1 diabetes. PediatrDiabetes 2016;17(Suppl. 22):49–5514. Knip M, Siljander H. The role of the intestinal microbiota in type 1 diabetesmellitus. Nat Rev Endocrinol 2016;12:154–16715. Ziegler AG, Rewers M, Simell O, et al. Seroconversion to multiple isletautoantibodies and risk of progression to diabetes in children. JAMA 2013;309:2473–247916. Kallionpää H, Elo LL, Laajala E, et al. Innate immune activity is detected priorto seroconversion in children with HLA-conferred type 1 diabetes susceptibility.Diabetes 2014;63:2402–241417. Moulder R, Bhosale SD, Erkkilä T, et al. Serum proteomes distinguishchildren developing type 1 diabetes in a cohort with HLA-conferred susceptibility.Diabetes 2015;64:2265–227818. Peet A, Kool P, Ilonen J, Knip M, Tillmann V; DIABIMMUNE Study Group. Birthweight in newborn infants with different diabetes-associated HLA genotypes inthree neighbouring countries: Finland, Estonia and Russian Karelia. DiabetesMetab Res Rev 2012;28:455–461

diabetes.diabetesjournals.org Kallionpää and Associates 2033

19. Cianciaruso C, Phelps EA, Pasquier M, et al. Primary human and rat b-cellsrelease the intracellular autoantigens GAD65, IA-2, and proinsulin in exosomestogether with cytokine-induced enhancers of immunity. Diabetes 2017;66:460–47320. Kim D, Pertea G, Trapnell C, Pimentel H, Kelley R, Salzberg SL. TopHat2:accurate alignment of transcriptomes in the presence of insertions, deletions andgene fusions. Genome Biol 2013;14:R3621. Anders S, Pyl PT, Huber W. HTSeq–a Python framework to work with high-throughput sequencing data. Bioinformatics 2015;31:166–16922. Robinson MD, McCarthy DJ, Smyth GK. edgeR: a Bioconductor package fordifferential expression analysis of digital gene expression data. Bioinformatics2010;26:139–14023. Dobin A, Davis CA, Schlesinger F, et al. STAR: ultrafast universal RNA-seqaligner. Bioinformatics 2013;29:15–2124. Butler A, Hoffman P, Smibert P, Papalexi E, Satija R. Integrating single-celltranscriptomic data across different conditions, technologies, and species. NatBiotechnol 2018;36:411–42025. Qiu X, Mao Q, Tang Y, et al. Reversed graph embedding resolves complexsingle-cell trajectories. Nat Methods 2017;14:979–98226. Ravassard P, Hazhouz Y, Pechberty S, et al. A genetically engineered humanpancreatic b cell line exhibiting glucose-inducible insulin secretion. J Clin Invest2011;121:3589–359727. Anagandula M, Richardson SJ, Oberste MS, et al. Infection of human islets ofLangerhans with two strains of Coxsackie B virus serotype 1: assessment of virusreplication, degree of cell death and induction of genes involved in the innateimmunity pathway. J Med Virol 2014;86:1402–141128. Reynier F, Pachot A, Paye M, et al. Specific gene expression signatureassociated with development of autoimmune type-I diabetes using whole-bloodmicroarray analysis. Genes Immun 2010;11:269–27829. Ferreira RC, Guo H, Coulson RM, et al. A type I interferon transcriptionalsignature precedes autoimmunity in children genetically at risk for type 1 diabetes.Diabetes 2014;63:2538–255030. Jin Y, Sharma A, Bai S, et al. Risk of type 1 diabetes progression in isletautoantibody-positive children can be further stratified using expression patternsof multiple genes implicated in peripheral blood lymphocyte activation andfunction. Diabetes 2014;63:2506–251531. Reinert-Hartwall L, Honkanen J, Salo HM, et al.; DIABIMMUNE Study Group;DIABIMMUNE Study Group. Th1/Th17 plasticity is a marker of advanced b cellautoimmunity and impaired glucose tolerance in humans. J Immunol 2015;194:68–7532. Heninger AK, Eugster A, Kuehn D, et al. A divergent population ofautoantigen-responsive CD4+ T cells in infants prior to b cell autoimmunity. SciTransl Med 2017;9:eaaf884833. Swafford AD, Howson JM, Davison LJ, et al. An allele of IKZF1 (Ikaros)conferring susceptibility to childhood acute lymphoblastic leukemia protectsagainst type 1 diabetes. Diabetes 2011;60:1041–104434. Ram R, Mehta M, Nguyen QT, et al. Systematic evaluation of genes andgenetic variants associated with type 1 diabetes susceptibility. J Immunol 2016;196:3043–305335. Arsenio J, Kakaradov B, Metz PJ, Kim SH, Yeo GW, Chang JT. Earlyspecification of CD8+ T lymphocyte fates during adaptive immunity revealed bysingle-cell gene-expression analyses. Nat Immunol 2014;15:365–37236. Dominguez CX, Amezquita RA, Guan T, et al. The transcription factors ZEB2and T-bet cooperate to program cytotoxic T cell terminal differentiation in responseto LCMV viral infection. J Exp Med 2015;212:2041–205637. Nishida A, Andoh A, Inatomi O, Fujiyama Y. Interleukin-32 expression in thepancreas. J Biol Chem 2009;284:17868–17876

38. Dotta F, Censini S, van Halteren AG, et al. Coxsackie B4 virus infection of betacells and natural killer cell insulitis in recent-onset type 1 diabetic patients. ProcNatl Acad Sci U S A 2007;104:5115–512039. Krogvold L, Edwin B, Buanes T, et al. Detection of a low-grade enteroviralinfection in the islets of Langerhans of living patients newly diagnosed with type1 diabetes. Diabetes 2015;64:1682–168740. Laitinen OH, Honkanen H, Pakkanen O, et al. Coxsackievirus B1 is associatedwith induction of b-cell autoimmunity that portends type 1 diabetes. Diabetes2014;63:446–45541. Richardson SJ, Willcox A, Bone AJ, Foulis AK, Morgan NG. The prevalence ofenteroviral capsid protein vp1 immunostaining in pancreatic islets in human type1 diabetes. Diabetologia 2009;52:1143–115142. Ylipaasto P, Klingel K, Lindberg AM, et al. Enterovirus infection in humanpancreatic islet cells, islet tropism in vivo and receptor involvement in cultured isletbeta cells. Diabetologia 2004;47:225–23943. Oikarinen S, Tauriainen S, Hober D, et al.; VirDiab Study Group. Virus antibodysurvey in different European populations indicates risk association betweencoxsackievirus B1 and type 1 diabetes. Diabetes 2014;63:655–66244. Shields BM, McDonald TJ, Oram R, et al.; TIGI Consortium. C-peptide declinein type 1 diabetes has two phases: an initial exponential fall and a subsequentstable phase. Diabetes Care 2018;41:1486–149245. Kim SH, Han SY, Azam T, Yoon DY, Dinarello CA. Interleukin-32: a cytokineand inducer of TNFalpha. Immunity 2005;22:131–14246. Orban T, Kis J, Szereday L, et al. Reduced CD4+ T-cell-specific gene ex-pression in human type 1 diabetes mellitus. J Autoimmun 2007;28:177–18747. Joosten LA, Netea MG, Kim SH, et al. IL-32, a proinflammatory cytokine inrheumatoid arthritis. Proc Natl Acad Sci USA 2006;103:3298–330348. Shioya M, Nishida A, Yagi Y, et al. Epithelial overexpression of interleukin-32alpha in inflammatory bowel disease. Clin Exp Immunol 2007;149:480–48649. Na SJ, So SH, Lee KO, Choi YC. Elevated serum level of interleukin-32a in thepatients with myasthenia gravis. J Neurol 2011;258:1865–187050. Jung MY, Son MH, Kim SH, Cho D, Kim TS. IL-32g induces the maturation ofdendritic cells with Th1- and Th17-polarizing ability through enhanced IL-12 andIL-6 production. J Immunol 2011;186:6848–685951. Walker LS, von Herrath M. CD4 T cell differentiation in type 1 diabetes. ClinExp Immunol 2016;183:16–2952. Jhun H, Choi J, Hong J, et al. IL-32g overexpression accelerates strepto-zotocin (STZ)-induced type 1 diabetes. Cytokine 2014;69:1–553. Xin T, Chen M, Duan L, Xu Y, Gao P. Interleukin-32: its role in asthma andpotential as a therapeutic agent. Respir Res 2018;19:12454. Viken MK, Blomhoff A, Olsson M, et al. Reproducible association with type1 diabetes in the extended class I region of the major histocompatibility complex.Genes Immun 2009;10:323–33355. Alvarez JI, Kébir H, Cheslow L, et al. JAML mediates monocyte and CD8 T cellmigration across the brain endothelium. Ann Clin Transl Neurol 2015;2:1032–103756. Hodik M, Anagandula M, Fuxe J, et al.; POD-V Consortium. Coxsackie-adenovirus receptor expression is enhanced in pancreas from patients with type1 diabetes. BMJ Open Diabetes Res Care 2016;4:e00021957. Leete P, Willcox A, Krogvold L, et al. Differential insulitic profiles determinethe extent of b-cell destruction and the age at onset of type 1 diabetes. Diabetes2016;65:1362–136958. von Toerne C, Laimighofer M, Achenbach P, et al. Peptide serum markers inislet autoantibody-positive children. Diabetologia 2017;60:287–29559. Achenbach P, Hummel M, Thümer L, Boerschmann H, Höfelmann D, ZieglerAG. Characteristics of rapid vs slow progression to type 1 diabetes in multiple isletautoantibody-positive children. Diabetologia 2013;56:1615–1622

2034 Early Signs of T1D Autoimmune Reaction Diabetes Volume 68, October 2019