Bo Kyung Koo,1,2 Sehyun Chae,3 Kristine M. Kim,4 Min Jueng Kang,5 Eunhee G. Kim,4

Soo Heon Kwak,1 Hye Seung Jung,1 Young Min Cho,1 Sung Hee Choi,1 Young Joo Park,1

Choong Ho Shin,6 Hak C. Jang,1 Chan Soo Shin,1 Daehee Hwang,3,7 Eugene C. Yi,5 andKyong Soo Park1,5

Identification of NovelAutoantibodies in Type 1Diabetic Patients Using a High-Density Protein MicroarrayDiabetes 2014;63:3022–3032 | DOI: 10.2337/db13-1566

Autoantibodies can facilitate diagnostic and therapeuticmeans for type 1 diabetes (T1DM). We profiled autoanti-bodies from serum samples of 16 T1DM patients, 16 type 2diabetic (T2DM) patients, and 27 healthy control subjectswith normal glucose tolerance (NGT) by using proteinmicroarrays containing 9,480 proteins. Two novel autoanti-bodies, anti-EEF1A1 and anti-UBE2L3, were selected frommicroarrays followed by immunofluorescence staining ofpancreas. We then tested the validity of the candidates byELISA in two independent test cohorts: 1) 95 adults withT1DM, 49 with T2DM, 11 with latent autoimmune diabetesin adults (LADA), 20 with Graves disease, and 66 with NGTand 2) 33 children with T1DM and 34 healthy children. Con-centrations of these autoantibodies were significantlyhigher in T1DM patients than in NGT and T2DM subjects(P < 0.01), which was also confirmed in the test cohort ofchildren (P < 0.05). Prevalence of anti-EEF1A1 and anti-UBE2L3 antibodies was 29.5% and 35.8% in T1DM, re-spectively. Of note, 40.9% of T1DM patients who lackanti-GAD antibodies (GADA) had anti-EEF1A1 and/or anti-UBE2L3 antibodies. These were also detected in patientswith fulminant T1DM but not LADA. Our approach identi-fied autoantibodies that can provide a new dimension ofinformation indicative of T1DM independent of GADA andnew insights into diagnosis and classification of T1DM.

Type 1 diabetes (T1DM) results from immune-mediatedpancreatic b-cell destruction. Several autoantibodies, suchas GAD65 antibody (GADA) (1), islet cell antibody (ICA)(2), protein tyrosine phosphatase antibody (IA-2 antibod-ies [IA-2A]) (3), and zinc transporter antibody (ZnT8A)(4), were identified and used for diagnosis and predictionof T1DM.

However, the nature and mechanism of b-cell destruc-tion in humans seem to be variable, which results in a broadspectrum of T1DM spanning from fulminant T1DM (5) tolatent autoimmune diabetes in adults (LADA) (6). Fulmi-nant T1DM is a rapidly progressing form of T1DM withoutevidence of autoimmunity, whereas LADA is autoimmunediabetes not requiring insulin at diagnosis. In most cases ofLADA, which account for 10% of incident cases of diabetesin adults, b-cell function is severely impaired within 6 years,leading to insulin dependency (6). Fulminant T1DM is animportant subtype of T1DM in Asian adults (7), accountingfor 15%–20% of T1DM with ketosis or ketoacidosis inJapan (5). Moreover, T1DM in adults often lacks evidenceof autoimmunity. The prevalence of autoantibodies associ-ated with T1DM declines as the onset age increases, espe-cially in the case of IA-2A (8) and ZnT8A (4,9). ZnT8A wasobserved in 80% of individuals in late adolescence (4,9),

1Department of Internal Medicine, Seoul National University College of Medicine,Seoul, Korea2Department of Internal Medicine, Boramae Medical Center, Seoul, Korea3School of Interdisciplinary Bioscience and Bioengineering, Pohang University ofScience and Technology, Pohang, Korea4Department of Systems Immunology, College of Biomedical Science, and Instituteof Bioscience and Biotechnology, Kangwon National University, Chuncheon, Korea5Department of Molecular Medicine and Biopharmaceutical Sciences, GraduateSchool of Convergence Science and Technology, and College of Medicine orCollege of Pharmacy, Seoul National University, Seoul, Korea6Department of Pediatrics, Seoul National University College of Medicine, Seoul,Korea

7Center for Systems Biology of Plant Aging Research, Institute for Basic Science,Daegu Gyeongbuk Institute of Science and Technology, Daegu, Korea

Corresponding author: Kyong Soo Park, kspark@snu.ac.kr; Eugene C. Yi, euyi@snu.ac.kr; or Daehee Hwang, dhwang@dgist.ac.kr.

Received 14 October 2013 and accepted 11 April 2014.

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db13-1566/-/DC1.

B.K.K., S.C., K.M.K., and M.J.K. contributed equally to this work.

© 2014 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, andthe work is not altered.

3022 Diabetes Volume 63, September 2014

IMMUNOLOGY

AND

TRANSPLANTATIO

N

12% of adults ,50 years of age, and very rarely for adults.60 years of age (9). Consequently, there has been a needfor novel markers that can reflect b-cell destruction in sucha broad spectrum of T1DM, which in turn can be used todifferentiate T1DM from other classes of diabetes.

Thus, we systematically explored autoantibodies asso-ciated with T1DM in a large-scale screening study ofautoantibody repertoires in subjects with T1DM by usinga high-density, fluorescence-based protein microarray con-taining duplicate spots of 9,480 human proteins derivedfrom the Ultimate ORF Clone Collection (Life Technolo-gies) to immunologically characterize a broad spectrumof T1DM. Among 9,480 proteins, we selected two candi-date autoantibodies—anti–eukaryote translation elonga-tion factor 1 a 1 (EEF1A1) autoantibody (EEF1A1-AAb)and anti–ubiquitin-conjugating enzyme 2L3 (UBE2L3) au-toantibody (UBE2L3-AAb)—based on multivariate partialleast squares–discriminant analysis (PLS-DA) and immuno-fluorescence staining and validated them in two indepen-dent cohorts.

RESEARCH DESIGN AND METHODS

Sample PreparationFor autoantibody profiling, we carefully selected the firstcohort of 16 Korean patients with recent-onset T1DM, 16patients with type 2 diabetes (T2DM), and 27 healthycontrol subjects with normal glucose tolerance (NGT).This cohort was used to screen candidate autoantibodiesusing protein microarrays. Serum samples were drawnfrom T1DM patients with 1) a fasting C-peptide level,0.3 nmol/L or serum C-peptide ,0.6 nmol/L after glu-cagon loading (10), 2) an initiation of insulin treatmentwithin 6 months after diagnosis, and 3) a duration ofdiabetes #12 months. Mean age of patients with T1DMin the first cohort was 42 6 16 years (Table 1). Thecontrol serum samples were obtained from T2DMpatients who were treated only with oral antidiabeticdrugs for at least 5 years and from subjects with NGTwho had no history of diabetes, no first-degree relativeswith diabetes, a fasting plasma glucose concentration of,6.1mmol/L, and an HbA1c value of ,5.8% (40 mmol/mol)(Table 1).

The second cohort to test the validity of the autoanti-body candidates comprised 95 T1DM, 49 T2DM, 11 LADA,and 66 NGT subjects, where T1DM patients met the firstand second selection criteria of the first cohort. Amongthem, 11 patients with fulminant T1DM, as definedaccording to a previous report (11), were included. Meanage of patients with T1DM in the second cohort was 32 611 years, and median duration of diabetes was 3 (range 0–14) years (Table 1). The same criteria for T2DM and NGTin the first cohort were used. Patients with LADA definedas previously reported (6) (n = 11) and patients withGraves disease who had no history of diabetes as the au-toimmune disease control (12) (n = 20) were also includedin the second cohort. The third cohort, comprising 33 chil-dren with T1DM and 34 healthy children, was used as an

Tab

le1—Clinicalcharacteristics

ofsub

jects

Firstcohort

Second

cohortThird

cohort

T1DM

T2DM

NGT

T1DM*

T2DM

LADA

NGT

Graves

T1DM

NGT

Num

ber

ofsub

jects16

1627

9549

1166

2033

34

Age

(years)42

616

686

368

62

326

1150

615

486

1166

62

306

886

410

61

Female

sex6(37.5)

7(43.8)

14(51.9)

47(49.5)

20(40.8)

6(54.5)

33(50.0)

10(50.0)

21(63.6)

17(50.0)

Duration

(years)0(0–1)

15(1–21)

––

3(0–14)

6(0–21)

3(0–9)

——

0(0–7)

—

BMI(kg/m

2)21.5

63.0

24.46

3.523.7

63.3

21.36

3.225.7

63.5

20.56

3.024.1

62.9

—16.8

62.5

17.76

2.5

HbA1c(%

)8.9

62.4

8.16

1.55.2

60.3

9.46

2.98.1

61.7

9.16

2.05.2

60.3

—9.6

62.2

5.46

0.5

FastingC-pep

tide(ng/m

L)0.4

60.4

1.86

0.71.8

60.7

0.36

0.32.5

60.9

0.56

0.51.6

60.6

—0.4

60.3

—

Data

aremean

6SD,n(%

),or

med

ian(range)

unlessotherw

iseind

icated.*T1D

Mexclud

ingLA

DA.

diabetes.diabetesjournals.org Koo and Associates 3023

independent test set. The same criteria for T1DM in thesecond cohort were used for the third cohort. Age ofpatients with T1DM in the third cohort was 8 6 4 years,and median duration of diabetes was 0 (range 0–7) years.Mean age of healthy children was 10 6 1 years (Table 1).

For patients with T1DM, serum GADA concentrationwas measured by a radioimmunoassay (RSR Ltd.). GADAvalues .1.45 International Units/mL were consideredpositive.

This study was conducted in accordance with theprovisions of the Declaration of Helsinki for participationof subjects in human research. The institutional reviewboard of the Clinical Research Institute at Seoul NationalUniversity Hospital approved the study protocol, andwritten informed consent was obtained from each subject.

Protein Microarray ExperimentsProtein microarray platform version 5.0 (ProtoArray;Invitrogen) contains duplicate spots of 9,480 N-terminalglutathione S-transferase–tagged human proteins printedon a nitrocellulose slide. In this array, the antigens havetheir native cellular enzymatic activities and conforma-tions. Assays were performed according to the instruc-tions of the manufacturer (13) as follows: Serum diluted1:500 in PBS with Tween was incubated on microarrayslides for 90 min, and Alexa Fluor 647–conjugated anti-human IgG was used to detect the reaction. The arrayslides were scanned with a GenePix 4000B fluorescentscanner (Molecular Devices).

Analysis of ProtoArray DataSignal intensities were acquired using ProtoArray Pros-pector, a software tool developed for ProtoArray data byInvitrogen (14). We converted fluorescent intensities forthe spot proteins (or autoantigens) to log2 intensities andthen normalized across the total 59 data sets by using thequantile normalization method (15). Protein microarraydata used in this study are available from the Gene Ex-pression Omnibus database (http://www.ncbi.nlm.nih.gov/gds) under submission number GSE 50866. To deter-mine the presence of autoantibodies for each probe, wecalculated a z score defined as follows: (spot intensity –the averaged intensity of all probes) / SD of all probes.The autoantibody with z score .1.64 (95th percentile ofall probe intensities) was considered present. To identifyautoantibodies showing a significant difference betweentwo groups (T1DM vs. NGT, T1DM vs. T2DM, or T1DMvs. T2DM + NGT), the M-statistic test in ProtoArray Pros-pector was applied to the normalized intensities asreported previously (16–20). For each probe, M test 1)produced the number of samples (M-statistic counts)with the intensity greater than or equal to the M-statisticthreshold in two groups and 2) computed the P value (thesignificance of the difference between two groups) fromthe sample sizes and the M-statistic counts in the twogroups. As an initial set of autoantibody candidates, theproteins with P , 0.05 and z score .1.64 in at least oneof T1DM samples were selected.

We then applied PLS-DA to compute the variableimportance in projection (VIP) representing the relativecontribution of each autoantibody to the separationamong T1DM, T2DM, and NGT samples (21). To deter-mine a cutoff for significant VIPs, we estimated an em-pirical distribution of VIP for the null hypothesis (i.e., anautoantibody has no contribution to the separation) byapplying PLS-DA followed by Gaussian kernel density es-timation method (22) to randomly permutated experi-ments. We then determined a VIP cutoff of 1.59 as the90th percentile of the empirical VIP distribution. Thus,the autoantibodies with VIP .1.59 were selected as thosesignificant contributing to the separation among T1DM,T2DM, and NGT samples.

Immunofluorescence Staining andImmunohistochemistryPrimary anti-EEF1A1 (Merck Millipore), anti-UBE2L3(Santa Cruz), and anti-chromosome 7 open reading frame53 (c7orf53) (Santa Cruz) antibodies were loaded ontoslides of pancreas tissue of a subject who underwentpartial pancreatectomy due to serous adenoma. AlexaFluor 647– (Merck Millipore), 594– (Invitrogen), or 488–(Invitrogen) conjugated secondary antibodies were usedto visualize insulin, glucagon, and candidate autoantigens,respectively. The cell nuclei were counterstained withDAPI (Invitrogen). An Olympus FluoView FV1000 confo-cal laser scanning microscope was used to acquire data. Toinvestigate the expression of candidate autoantigens inother tissues, including liver, kidney, stomach, intestine,muscle, brain, and thyroid gland, immunohistochemistryusing tissue array (SuperBioChips) was performed withEnVision System-HRP–labeled polymer (Dako) accordingto the manufacturer’s instructions.

ELISA and ImmunoprecipitationThe differential expression of autoantibodies identified byProtoArray in T1DM patients was confirmed by ELISAusing corresponding recombinant proteins. Plates coatedwith EEF1A1 or UBE2L3 (Origene) were incubated withdiluted serum samples (1:1,000–1:10,000) in PBS. Allexperiments were performed in triplicate for each patientand control sample. The recombinant proteins and theircorresponding antibody, anti-EEF1A1, and anti-UBE2L3were used as the positive control for the assay. Autoanti-body bound to the protein was detected with goat anti-human IgG-HRP (Jackson ImmunoResearch). The cutoffpoint for the presence or absence of each autoantibodywas determined according to the mean absorbance + 3 3SD of NGT, which corresponded to ;99% of normal con-trol. Because titers of autoantibodies may differ betweenthe solid and liquid phases as a result of the difference inexposure of binding epitopes, the relative levels ofEEF1A1-AAb and UBE2L3-AAb were analyzed using animmunoprecipitation–Western blot method. T1DM pa-tient samples categorized as high, medium, and lowwere selected from the second cohort based on ELISAabsorbance values at 450 nm. Autoantibodies present in

3024 Novel Autoantibodies in Type 1 Diabetes Diabetes Volume 63, September 2014

these groups of T1DM and NGT samples were immuno-precipitated using EEF1A1-flag or UBE2L3-flag. Autoanti-body:antigen immune complex captured with antiflagM2-antibody–conjugated magnetic beads was eluted fromthe beads following the manufacturer’s protocol (Sigma)and analyzed by Western blot using anti-human IgG-HRP. Antigens coeluted from the beads were detectedusing antibodies for EEF1A1 and UBE2L3. An irrelevanthuman IgG antibody (ForteBio) was used as a control forthe immunoprecipitation assay.

Statistical AnalysisStatistical analysis for the microarray studies was as justdescribed. Results of ELISA experiments were evaluatedby means of independent t test. The x2 test was used toevaluate differences in the proportion of antibody-positivesubjects per group. The number of test sets for ELISA wasdetermined from the results of ELISA in the second cohort.That is, to reliably identify a significant group difference bythe two-sample t test with a type I error rate of 0.05 andstatistical power of 0.8, 32 subjects in disease and controlgroups were needed for the test sets.

RESULTS

Serum Autoantibody ProfilingWe profiled autoantibodies in the sera from 16 recent-onset T1DM, 16 T2DM, and 27 NGT subjects by using theProtoArray human protein microarray version 5 contain-ing 9,480 different proteins spotted in duplicate. These9,480 proteins correspond to 6,867 of the 19,050annotated proteins (36.0%) (Swiss-Prot 2013_05), whichinclude 1,632 of 4,673 genes (34.9%) encoding proteinslocalized in plasma membrane or extracellular regionsbased on their gene ontology cellular component annota-tions and 809 of 1,929 proteins (41.9%) previouslydetected in human serum and plasma samples accordingto the human plasma proteome project (23). The autoan-tibodies against 3,060 among the 9,480 proteins werefound to be present in at least one of the samples. Toassess the data quality, we examined whether the signalintensity of a known autoantigen, GAD65, obtained fromProtoArray reflects a serum concentration of GADA thatwas independently measured by a conventional radioim-munoassay and confirmed a significant correlation be-tween them (Spearman r = 0.863, P , 0.001). On theother hand, none of IA-2A, ICA, or ZnT8 is spotted inthe ProtoArray platform (Supplementary Table 1).

Identification of Autoantibodies PredominantlyExpressed in T1DMTo identify autoantibodies predominantly expressed inT1DM samples, the relative abundance of 3,060 autoanti-bodies detected in the profiling were compared among 1)T1DM versus NGT, 2) T1DM versus T2DM, and 3) T1DMversus NGT + T2DM by using M test. From these com-parisons, we identified 103, 27, and 79 autoantibodies,respectively, whose abundance increased in T1DM withM test P , 0.05 and z . 1.64 in at least one of the

T1DM samples (Fig. 1A). Of these, we focused on the69 autoantibodies (Fig. 1B and Supplementary Table 2)identified from T1DM versus NGT + T2DM, which over-lapped with those in either T1DM versus NGT or T1DMversus T2DM.

Selection of Autoantibodies for Validation in a Large-Scale CohortTo identify a set of biomarkers predictive of T1DM, it isimportant to evaluate the relative contribution of the 69proteins to the separation between T1DM and T2DM/NGT samples. To this end, we applied multivariate PLS-DA (21) to the abundance of the 69 autoantibodies toevaluate their collective contribution to the separation(Fig. 2A). The PLS-DA result showed that a clear separa-tion between T1DM and T2DM/NGT samples can beachieved by the differential expression of the 69 proteins(Fig. 2B and Supplementary Table 3). The relative col-lective contribution of individual autoantibodies to the

Figure 1—Identification of T1DM-associated autoantibodies usingserum autoantibody profiling. A: T1DM-associated autoantibodycandidates whose abundance increased in T1DM. The Venn dia-gram shows the numbers of these candidates identified from thethree comparisons: 1) T1DM vs. NGT, 2) T1DM vs. T2DM, and 3)T1DM vs. NGT + T2DM. B: Increased abundance of the 69 autoanti-bodies in T1DM revealed by hierarchical clustering analysis. Thecolumns and rows in the heat map represent samples and autoanti-bodies, respectively. In each row, the intensities were divided by theSD after subtracting the mean intensity from them (i.e., autoscaling).Yellow and blue represent the increase and decrease in abundance,respectively. The color bar represents the gradient of the auto-scaled log2 intensities.

diabetes.diabetesjournals.org Koo and Associates 3025

separation achieved by PLS-DA was estimated as their VIPvalues (21) (Supplementary Table 4). A large VIP valueindicates a high contribution to the separation. Amongthe 69 autoantibodies, we selected 50 with VIP values.1.59 (Fig. 2C) (see RESEARCH DESIGN AND METHODS). We

also evaluated the contribution of individual samples tothe separation captured by the PLS-DA model by usingthe sample contribution analysis. This analysis showed nosignificant correlation between disease duration and thedistances to the PLS-DA model (r = –0.465 and P = 0.070)

Figure 2—Selection of three autoantibodies associated with T1DM. A: The overall scheme for selection of two novel autoantibodies testedin large-scale cohorts. See RESEARCHDESIGN ANDMETHODS for detailed procedures. B: Discrimination of T1DM from T2DM and NGT using PLS-DA. The two-dimensional PLS-DA score plot shows a separation among T1DM, T2DM, and NGT samples. The line indicates a decisionfunction between T1DM and T2DM + NGT samples. C: Selection of a second list of autoantibodies with significant collective contribution tothe separation-based VIP values. The cutoff of 1.59 (line) was determined as described in RESEARCH DESIGN AND METHODS. Among the 50autoantibodies, four (GADA, EEF1A1-AAb, c7orf53-AAb, and UBE2L3-AAb) were selected (see RESEARCH DESIGN ANDMETHODS). D: A heat mapshowing the increased abundance of GADA, EEF1A1-AAb, c7orf53-AAb, and UBE2L3-AAb in T1DM patients compared with T2DMpatients and NGT subjects. See Fig. 1B legend for the descriptions of the heat map.

3026 Novel Autoantibodies in Type 1 Diabetes Diabetes Volume 63, September 2014

(Supplementary Fig. 1). We further selected the followingthree autoantibodies that were detected in none of NGTsamples (i.e., z , 1.64 in all NGT samples) (Fig. 2D): 1)EEF1A1-AAb, 2) UBE2L3-AAb, and 3) c7orf53-AAb. Forthe selection of the autoantibodies targeting b-cells, weperformed immunofluorescence staining of the pancreasusing antibodies with specificity for EEF1A1, UBE2L3,and c7orf53. EEF1A1 and UBE2L3 were predominantlyexpressed in b-cells, whereas c7orf53 was exclusivelyfound in a-cells (Fig. 3). On the basis of these results,EEF1A1-AAb and UBE2L3-AAb were selected as the can-didates for the T1DM-associated autoantibodies for thesubsequent studies. Both EEF1A1 and UBE2L3 wereexpressed in skin, stomach, colon, kidney, lung, and ad-renal gland tissues (Supplementary Fig. 2), which wasconsistent with previous reports showing extrapancreaticexpression of EEF1A1 and UBE2L3 (24–27).

Validation of Novel AutoantibodiesWe measured the amount of EEF1A1-AAb and UBE2L3-AAb in serum samples by ELISA in a large cohort, whichcomprised 95 T1DM, 49 T2DM, 11 LADA, and 66 NGTsubjects as well as 20 patients with Graves disease.EEF1A1-AAb and UBE2L3-AAb were significantly higherin their abundance in T1DM samples than in NGT (P ,0.001 in both) and T2DM (P , 0.001 in both) samples

(Fig. 4A). Of note, EEF1A1-AAb showed a significantlyhigher abundance in T1DM compared with Graves disease(P = 0.003). On the other hand, UBE2L3-AAb was notsignificantly different between T1DM and Graves disease.The prevalence of EEF1A1-AAb and UBE2L3-AAb was29.5% and 35.8% in T1DM, respectively, which was sig-nificantly higher than T2DM and NGT (P , 0.001 inboth). These antibodies were detected in only 2.0% ofT2DM and 1.5% of NGT subjects (Table 2). To quantita-tively assess the accuracy for prediction of T1DM, wegenerated receiver operating characteristic curves for theautoantibodies and computed the areas under the curve(AUCs). The AUCs (0.73 for EEF1A1-AAb, 0.78 forUBE2L3-AAb) revealed high accuracy of the autoantibodiesin predicting T1DM (Fig. 4B). The comparison of EEF1A1-AAb and UBE2L3-AAb in the third cohort of childrenbetween T1DM and NGT showed a significantly higherabundance of these autoantibodies in T1DM than in NGTsubjects (P = 0.007 and 0.020, respectively) (Fig. 4C), whichis consistent with the results in the adult T1DM cohort.

Western blot analysis of the pull-down autoantibodiesshowed that the relative abundance of autoantibodies inthe T1DM samples measured by the immunoprecipitationmethod reasonably agrees with the titers of autoanti-bodies measured by ELISA (Supplementary Fig. 3), andthey were well correlated with each other (r = 0.779,

Figure 3—Localization of candidate autoantigens in pancreatic b-cells. Triple immunofluorescence staining of the pancreas using specificantibodies against EEF1A1 (A), UBE2L3 (F ), and c7orf53 (K) (green) along with anti-insulin (B, G, and L) (white) and anti-glucagon (C, H, andM) (red) antibodies. The cell nuclei were counterstained with DAPI (D, I, and N) (blue). EEF1A1 and UBE2L3 were mainly enriched in theb-cells in the islets (E and J), whereas c7orf53 was expressed exclusively in a-cells (O). Original magnification 3400, scale bar = 50 mm.

diabetes.diabetesjournals.org Koo and Associates 3027

P = 0.039, for EEF1A1-AAb and r = 0.940, P = 0.002, forUBE2L3-AAb) (Supplementary Fig. 4).

Clinical Implications of Novel AutoantibodiesClinical characteristics of the novel autoantibodies wereevaluated in the second cohort. The prevalence ofEEF1A1-AAb and UBE2L3-AAb in T1DM was 29.5% and35.8%, respectively, which was comparable to that of

ZnT8A in adult T1DM (28). The most prevalent autoan-tibody, GADA in T1DM (8), was detected in 76.3% ofT1DM patients (71 of 93). To further examine the clinicalimportance of this finding, the prevalence of the novelautoantibodies in the T1DM patients without GADAwas analyzed. Of the 22 GADA-negative T1DM patients,an equal distribution (31.8%) of autoantibody betweenEEF1A1-AAb and UBE2L3-AAb was observed, and a total

Figure 4—Testing the validity of the selected autoantibodies in independent cohorts. A: Relative titers of EEF1A1-AAb and UBE2L3-AAbmeasured by ELISA absorbance at 450 nm in T1DM, T2DM, and NGT subjects in the second cohorts. The signal above (mean + 33 SD) ofhealthy controls (dotted line) was considered positive. The solid line for each group represents the mean value of ELISA measurements. B:Receiver operating characteristic analysis representing the accuracy (AUCs) of the novel autoantibodies in predicting T1DM. C: Relativetiters of EEF1A1-AAb and UBE2L3-AAb measured by ELISA absorbance at 450 nm in T1DM, T2DM, and NGT subjects in the third cohorts.The solid line of each group represents the mean value of ELISA measurements. AAb, autoantibody. *P < 0.05, **P < 0.01.

Table 2—Prevalence of anti-EEF1A1 and anti-UBE2L3 antibodies from ELISA

Prevalence of positivity* P values†

T1DM(n = 95)

T2DM(n = 49)

LADA(n = 11)

Gravesdisease(n = 20)

NGT(n = 66)

T1DM vs.NGT

T1DM vs.T2DM

T1DM vs.LADA

T1DM vs.Graves disease

EEF1A1 28 (29.5) 1 (2.0) 0 2 (10.0) 1 (1.5) ,0.001 ,0.001 0.035 0.094

UBE2L3 34 (35.8) 1 (2.0) 0 4 (20.0) 1 (1.5) ,0.001 ,0.001 0.015 0.201

Data are n (%). *The signal above (mean + 3 3 SD) healthy controls was considered as positive. †P values from x2 test.

3028 Novel Autoantibodies in Type 1 Diabetes Diabetes Volume 63, September 2014

of 40.9% (9 of 22) had either EEF1A1-AAb or UBE2L3-AAb (Fig. 5A), which was higher than that of ZnT8A in anAsian population (29). Additionally, it is worth notingthat the proportion of patients with evidence of autoim-munity in T1DM was increased from 76.3% to 86.0%when two newly identified novel autoantibodies wereadded to GADA.

We also observed the trend of negative correlationbetween the age of onset and the prevalence of EEF1A1-AAb. The age of disease onset was significantly younger inT1DM patients with EEF1A1-AAb than in those withoutEEF1A1-AAb (P , 0.001) (Table 3). Categorical analysisshowed that the prevalence of EEF1A1-AAb in T1DM in-creased when the age of disease onset was young, whichwas not observed for GADA (P , 0.001 for trend) (Fig.5B), which was maintained even in the stratified analysisaccording to duration of diabetes (Fig. 5C). Duration ofdiabetes did not affect the prevalence of EEF1A1-AAb(Table 3), and EEF1A1-AAb titer was not associatedwith disease duration (r = 0.153, P = 0.139), whereasZnT8A has been shown to decline rapidly in the first yearsafter disease onset and is less persistent than IA-2A orGADA in the longer term (30). Of note, the prevalence ofEEF1A1-AAb increased as the C-peptide level decreased(P = 0.03 for trend), which was not observed for GADA(Fig. 5E). In addition, EEF1A1-AAb was detected in 27.3%

of subjects with fulminant T1DM (3 of 11), whereasGADA was not detected in fulminant T1DM patients(Fig. 5F and Table 3). There was no difference in theprevalence of other autoimmune diseases or history ofdiabetic ketoacidosis between T1DM patients with orwithout EEF1A1-AAb. However, EEF1A1-AAb notablywas not detected among the patients with LADA (Table2). Furthermore, the characteristics found in patientswith EEF1A1-AAb were similar to those with UBE2L3-AAb (Table 3 and Fig. 5B, D, and G). UBE2L3-AAb wasdetected in 18.2% of patients with fulminant T1DM,whereas UBE2L3-AAb was not detected in patients withLADA.

DISCUSSION

Novel autoantibodies indicative of the autoimmune de-struction in the serum of T1DM patients can be im-portant in developing new diagnostic and therapeuticmeans for T1DM patients. In this study, we extensivelyprofiled autoantibodies by using a protein-antigen micro-array in the sera of 16 T1DM, 16 T2DM, and 27 NGTsubjects. Among 9,480 different proteins in the array,we initially selected three novel autoantibody candi-dates (EEF1A1-AAb, c7orf53-AAb, and UBE2L3-AAb) byM test coupled with PLS-DA. Considering potential auto-antigens to be expressed in pancreatic b-cells, EEF1A1-AAb

Figure 5—Clinical implications of the selected autoantibodies. A: The Venn diagram shows the prevalence of EEF1A1-AAb and UBE2L3-AAb in T1DM with and without GADA. Nine of 22 GADA-negative T1DM patients had either EEF1A1-AAb or UBE2L3-AAb. B: Theprevalence of EEF1A1-AAb in T1DM was 45.6% (26 of 57), 8.3% (2 of 24), and 0% (0 of 14) in patients aged <30, 30–39, and $40 yearsat onset, respectively (P< 0.001 for trend). The negative correlation trend between the age of onset and the prevalence of EEF1A1-AAb (C)and UBE2L3-AAb (D) autoantibodies was maintained even in the analysis stratified according to duration of diabetes (0, 1–2, 3–5, and >5years). E: Prevalence of EEF1A1-AAb and UBE2L3-AAb according to fasting C-peptide level. Relative titers of EEF1A1-AAb (F ) andUBE2L3-AAb (G) measured by ELISA absorbance at 450 nm among typical patients with T1DM, fulminant T1DM, and LADA in the secondcohort. Dotted lines represent the signal above (mean + 3 3 SD) of healthy controls, and the solid lines are the mean value of ELISAmeasurements in each group. AAb, autoantibody.

diabetes.diabetesjournals.org Koo and Associates 3029

and UBE2L3-AAb were selected as autoantibody candidatesbased on immunofluorescence staining for further valida-tion studies. The increase in abundance of the selectednovel autoantibodies in T1DM was confirmed by ELISAin a second cohort and a third test set of children.

The prevalence of EEF1A1-AAb and UBE2L3-AAb was;30% for T1DM patients and 1.5% for NGT subjects inthe present study, which is comparable to that of ZnT8Ain adult T1DM patients (9). The two identified autoanti-bodies EEF1A1-AAb and UBE2L3-AAb have several clini-cal advantages compared with ZnT8A. First, the diseaseduration did not affect the prevalence of these autoanti-bodies. By contrast, ZnT8A was shown to decline rapidlyin the first years after disease onset and to be less persis-tent than IA-2A or GADA in the longer term (30). Thismay imply that both EEF1A1-AAb and UBE2L3-AAb areefficient diagnostic markers of T1DM for patients witha long duration of diabetes. Second, the prevalence ofEEF1A1-AAb in GADA-negative T1DM was 31.8%, whichis higher than that of ZnT8A in an Asian population (29).The prevalence further increased to 40.9% when UBE2L3-AAb was added. Considering that the prevalence of theseautoantibodies increased for young T1DM patients(45.6% in T1DM patients with an age of onset ,30years), the use of both autoantibodies should be moreefficient for the young patients without GADA. Third,T1DM patients with and without GADA showed no sig-nificant difference in the prevalence of EEF1A1-AAb andUBE2L3-AAb (38.0 and 28.2% in GADA-positive T1DMand 31.8 and 27.3% in GADA-negative T1DM forEEF1A1-AAb and UBE2L3-AAb, respectively). In addition,

they were detected in 27% of patients with fulminantT1DM without GADA but were not detected in LADApatients. Thus, the two autoantibodies can provide noveldiagnostic information and new insights into the classifi-cation of and therapy for T1DM as well as into T1DMpathogenesis.

EEF1A1 and UBE2L3 are involved in protein trans-lation (31) and degradation (24), respectively. We previ-ously reported aminoacyl-transfer RNA synthetase asa novel marker of T1DM (32), which is also involved inprotein translation. EEF1A1 is associated with Felty syn-drome, an autoimmune disease (33), and UBE2L3 is as-sociated with the risk of such autoimmune diseases asrheumatoid arthritis (34), Crohn disease (35), and sys-temic lupus erythematosus (36). These data indicatethat protein homeostasis represented by these autoanti-bodies can reflect the autoimmunity in T1DM. Of note,the two autoantibodies correlated well in their levels (Sup-plementary Fig. 5). Additionally, the immunofluorescencestaining of the corresponding two autoantigens in thepancreas using antibodies specific for EEF1A1 andUBE2L3 showed that both autoantigens were expressedat higher levels in islets of Langerhans in the pancreas,more specifically in b-cells, than in the acinar cells anda-cells. The tissue array data showed that they areexpressed in many tissues, suggesting that autoantibodiestargeting EEF1A1 and UBE2L3 may come from epitopespreading, which is a typical phenomenon occurring dur-ing the preclinical and clinical disease process in T1DM(37). Nonetheless, all these data suggest that the proteinhomeostasis represented by these proteins can be used to

Table 3—Clinical characteristics of T1DM patients with anti-EEF1A1 antibody, anti-UBE2L3 antibody, and anti-GADA

Anti-GADA Anti-EEF1A1 antibody Anti-UBE2L3 antibody

Positive(n = 71)

Negative(n = 22)

Positive(n = 28)

Negative(n = 67)

Positive(n = 34)

Negative(n = 61)

Female sex 36 (50.7) 9 (40.9) 12 (42.9) 35 (52.2) 12 (35.3) 35 (57.4)

Age (years) 29 (17–69) 32 (19–58) 25 (17–33)† 34 (17–69) 23 (17–37)† 34 (17–69)

Age of onset (years) 25 (13–66) 29 (10–58) 21 (13–33)† 30 (10–66) 19 (14–36)† 32 (10–66)

Disease duration (years) 3.5 6 3.0 2.6 6 3.2 3.3 6 2.7 3.2 6 3.7 2.9 6 2.7 3.4 6 3.7

Insulin requirement(International Units/day/kg) 0.68 6 0.30 0.75 6 0.21 0.73 6 0.28 0.69 6 0.28 0.73 6 0.27 0.69 6 0.29

History of DKA 20 (28.2)‡ 16 (72.7) 11 (35.3) 27 (40.3) 14 (41.2) 24 (39.3)

Fulminant T1DM 0‡ 11 (50.0) 3 (10.7) 8 (11.9) 2 (5.9) 9 (14.8)

Other autoimmune disease 15 (21.4) 3 (14.3) 5 (18.5) 14 (21.2) 7 (21.1) 12 (20.0)

BMI (kg/m2) 21.1 6 3.1 21.6 6 3.3 20.9 6 3.2 21.4 6 3.2 20.5 6 3.1 21.4 6 3.3

Fasting C-peptide (ng/mL) 0.33 6 0.32 0.22 6 0.24 0.22 6 0.23 0.33 6 0.33 0.25 6 0.26 0.33 6 0.33

HbA1c (%) 9.7 6 3.1 8.5 6 2.0 9.7 6 3.6 9.4 6 2.9 9.8 6 3.3 9.3 6 3.0

Prevalence of GADA — — 21 (75.0) 50 (76.9) 27 (79.4) 44 (74.6)

Prevalence of anti-UBE2L3 27 (38.0) 7 (31.8) 26 (92.9)‡ 8 (11.9) — —

Prevalence of anti-EEF1A1 21 (29.6) 7 (31.8) — — 26 (76.5)‡ 2 (3.3)

Data are n (%), median (range), or mean6 SD. Anti-EEF1A1 and anti-UBE2L3 antibodies were measured by ELISA, and anti-GADA wasmeasured by radioimmunoassay. DKA, diabetic ketoacidosis. †P , 0.05 from independent t test comparing the subjects withoutrespective autoantibody. ‡P , 0.05 from x2 test comparing the subjects without respective autoantibody.

3030 Novel Autoantibodies in Type 1 Diabetes Diabetes Volume 63, September 2014

show the autoimmune destruction of pancreatic islet b-cells.Detailed functional studies can be designed to elucidate themechanisms underlying the links of the protein homeostasisto the T1DM-related autoimmunity in the pancreas.

Furthermore, our approach involving autoantibodyprofiling provided a comprehensive list of autoantibodycandidates that extends extensively the current list ofautoantibodies identified by previous approaches, suchas protein expression profiling of pancreas (4) and mea-surements of molecular targets associated with insulinsecretion in b-cells (3,38) or pathologically proven auto-antigens (2,39). Other than the two autoantibodies testedin study, an additional 66 potential novel autoantibodies,including fms-like tyrosine kinase receptor 3 ligand(FLT3LG), were previously shown to be associated withdiabetes (40) and identified in the current study (Fig. 1Band Supplementary Table 2). FLT3LG treatment wasreported to delay diabetes onset in the NOD mouse modelof autoimmune diabetes (41). Protein kinase C (PRKCA),also shown in our list, is associated with multiple sclerosis(42), which is known to be initiated by a misdirectedimmune response against myelin autoantigens (43). Inaddition, our list includes the molecules associated withT1DM-related pathophysiology in pancreatic b-cells (Sup-plementary Fig. 6), such as insulin secretion [PRCKA (44),ACVR2B (45), GLUL (46), and PSMD9 (47)] and glucosemetabolism [GPI (48)], indicating their potential roles inthe autoimmune destruction of the pancreatic b-cellsin T1DM. Thus, the list of autoantibody candidates weidentified in this study can be a comprehensive basis forunderstanding the links of autoantibodies to T1DM di-agnosis and pathogenesis.

The clinical implications of the two novel autoantibodiescan be further tested with a larger number of GADA-negative and fulminant T1DM patients. In addition,longitudinal studies can be designed to further demon-strate the nature of dynamic changes of these autoanti-bodies during the course of disease progression, and newsubtypes of T1DM patients might be further characterizedbased on the dimension of protein homeostasis repre-sented by these autoantibodies when larger numbers ofadult T1DM patients are used. Nevertheless, our approachsuccessfully demonstrated the possibility of the use of thetwo autoantibodies as diagnostic markers and/or thera-peutic targets. Finally, our autoantibody profiling approachcan be applied to other autoimmune diseases for whichautoantibody markers are unknown, and the newly iden-tified autoantibodies can contribute to the understandingof the mechanisms underlying autoimmune destruction.In summary, our approach successfully identified two novelautoantibodies, EEF1A1-AAb and UBE2L3-AAb, that canadd a new dimension to the diagnosis, classification, therapy,and pathogenesis of T1DM.

Funding. This work was supported by the Seoul National University HospitalResearch Fund (to B.K.K.), the National Research Foundation of Korea (to K.M.K.),

a 2014 Research Grant from Kangwon National University (to K.M.K.), the Pro-teogenomic Research and Priority Research Centers Program (to D.H. and E.C.Y.),the World Class University program (to E.C.Y. and K.S.P.), Institute for BasicScience grant CA1308 (to D.H.), and the 21C Frontier Functional ProteomicsProject (to D.H., E.C.Y., and K.S.P.).Duality of Interest. No potential conflicts of interest relevant to this articlewere reported.Author Contributions. B.K.K. contributed resources and to the researchdesign, experiments, and writing of the manuscript. S.C. contributed to the dataanalysis and writing of the manuscript. K.M.K. contributed resources and to theresearch design, experiments, data analysis, and writing of the manuscript. M.J.K.contributed to the experiments, data analysis, and writing of the manuscript. E.G.K.contributed to the experiments and data analysis. S.H.K., H.S.J., Y.M.C., S.H.C.,Y.J.P., C.H.S., and H.C.J. contributed to the experiments. C.S.S. contributed tothe data analysis. D.H., E.C.Y., and K.S.P. contributed resources and to theresearch design, data analysis, and writing of the manuscript. D.H., E.C.Y.,and K.M.K. are the guarantors of this work and, as such, had full access to allthe data in the study and take responsibility for the integrity of the data and theaccuracy of the data analysis.Prior Presentation. Parts of this study were presented in abstract form atthe 74th Scientific Sessions of the American Diabetes Association, San Francisco,CA, 13–17 June 2014.

References1. Baekkeskov S, Nielsen JH, Marner B, Bilde T, Ludvigsson J, Lernmark A.Autoantibodies in newly diagnosed diabetic children immunoprecipitate humanpancreatic islet cell proteins. Nature 1982;298:167–1692. Bottazzo GF, Florin-Christensen A, Doniach D. Islet-cell antibodies in di-abetes mellitus with autoimmune polyendocrine deficiencies. Lancet 1974;2:1279–12833. Rabin DU, Pleasic SM, Shapiro JA, et al. Islet cell antigen 512 is a diabetes-specific islet autoantigen related to protein tyrosine phosphatases. J Immunol1994;152:3183–31884. Wenzlau JM, Juhl K, Yu L, et al. The cation efflux transporter ZnT8(Slc30A8) is a major autoantigen in human type 1 diabetes. Proc Natl Acad SciU S A 2007;104:17040–170455. Imagawa A, Hanafusa T, Uchigata Y, et al. Fulminant type 1 diabetes:a nationwide survey in Japan. Diabetes Care 2003;26:2345–23526. Stenström G, Gottsäter A, Bakhtadze E, Berger B, Sundkvist G. Latent au-toimmune diabetes in adults: definition, prevalence, beta-cell function, andtreatment. Diabetes 2005;54(Suppl. 2):S68–S727. Imagawa A, Hanafusa T. Fulminant type 1 diabetes—an important subtypein East Asia. Diabetes Metab Res Rev 2011;27:959–9648. Vermeulen I, Weets I, Asanghanwa M, et al.; Belgian Diabetes Registry.Contribution of antibodies against IA-2b and zinc transporter 8 to classificationof diabetes diagnosed under 40 years of age. Diabetes Care 2011;34:1760–17659. Lampasona V, Petrone A, Tiberti C, et al.; Non Insulin Requiring Autoim-mune Diabetes (NIRAD) Study Group. Zinc transporter 8 antibodies complementGAD and IA-2 antibodies in the identification and characterization of adult-onsetautoimmune diabetes: Non Insulin Requiring Autoimmune Diabetes (NIRAD) 4.Diabetes Care 2010;33:104–10810. Hother-Nielsen O, Faber O, Sørensen NS, Beck-Nielsen H. Classification ofnewly diagnosed diabetic patients as insulin-requiring or non-insulin-requiringbased on clinical and biochemical variables. Diabetes Care 1988;11:531–53711. Imagawa A, Hanafusa T. Fulminant type 1 diabetes mellitus. Endocr J 2006;53:577–58412. Burman KD, Baker JR Jr. Immune mechanisms in Graves’ disease. EndocrRev 1985;6:183–23213. Invitrogen. ProtoArray Human Protein Microarray v5.0 for Immune ResponseBioMarker Profiling (IRBP): IRBP Experienced Users Guide. Grand Island, NY, LifeTechnologies, 2011

diabetes.diabetesjournals.org Koo and Associates 3031

14. Vermeulen N, de Béeck KO, Vermeire S, et al. Identification of a novelautoantigen in inflammatory bowel disease by protein microarray. Inflamm BowelDis 2011;17:1291–130015. Bolstad BM, Irizarry RA, Astrand M, Speed TP. A comparison of normali-zation methods for high density oligonucleotide array data based on variance andbias. Bioinformatics 2003;19:185–19316. Le Roux S, Devys A, Girard C, Harb J, Hourmant M. Biomarkers for thediagnosis of the stable kidney transplant and chronic transplant injury using theProtoArray� technology. Transplant Proc 2010;42:3475–348117. Hu S, Vissink A, Arellano M, et al. Identification of autoantibody biomarkersfor primary Sjögren’s syndrome using protein microarrays. Proteomics 2011;11:1499–150718. Babel I, Barderas R, Díaz-Uriarte R, Martínez-Torrecuadrada JL, Sánchez-Carbayo M, Casal JI. Identification of tumor-associated autoantigens for the di-agnosis of colorectal cancer in serum using high density protein microarrays. MolCell Proteomics 2009;8:2382–239519. Pallier A, Hillion S, Danger R, et al. Patients with drug-free long-term graftfunction display increased numbers of peripheral B cells with a memory andinhibitory phenotype. Kidney Int 2010;78:503–51320. Sboner A, Karpikov A, Chen G, et al. Robust-linear-model normalization toreduce technical variability in functional protein microarrays [published correctionappears in J Proteome Res 2010;9:636]. J Proteome Res 2009;8:5451–546421. Hyung SW, Lee MY, Yu JH, et al. A serum protein profile predictive of theresistance to neoadjuvant chemotherapy in advanced breast cancers. Mol CellProteomics 2011;10:M111.01102322. Bowman AW, Azzalini A. Applied Smoothing Techniques for Data Analysis:The Kernel Approach With S-Plus Illustrations (Oxford Statistical Sciences Series).New York, Clarendon Press, Oxford University Press, 199723. Farrah T, Deutsch EW, Omenn GS, et al. A high-confidence human plasmaproteome reference set with estimated concentrations in PeptideAtlas. Mol CellProteomics 2011;10:M110.00635324. Ciechanover A. The ubiquitin-proteasome pathway: on protein death andcell life. EMBO J 1998;17:7151–716025. Knudsen SM, Frydenberg J, Clark BF, Leffers H. Tissue-dependent variationin the expression of elongation factor-1 alpha isoforms: isolation and charac-terisation of a cDNA encoding a novel variant of human elongation-factor 1 alpha.Eur J Biochem 1993;215:549–55426. Chambers DM, Peters J, Abbott CM. The lethal mutation of the mousewasted (wst) is a deletion that abolishes expression of a tissue-specific isoform oftranslation elongation factor 1alpha, encoded by the Eef1a2 gene. Proc Natl AcadSci U S A 1998;95:4463–446827. Lee S, Francoeur AM, Liu S, Wang E. Tissue-specific expression in mam-malian brain, heart, and muscle of S1, a member of the elongation factor-1 alphagene family. J Biol Chem 1992;267:24064–2406828. Kawasaki E, Uga M, Nakamura K, et al. Association between anti-ZnT8autoantibody specificities and SLC30A8 Arg325Trp variant in Japanese patientswith type 1 diabetes. Diabetologia 2008;51:2299–230229. Yang L, Luo S, Huang G, et al. The diagnostic value of zinc transporter8 autoantibody (ZnT8A) for type 1 diabetes in Chinese. Diabetes Metab Res Rev2010;26:579–58430. Wenzlau JM, Walter M, Gardner TJ, et al. Kinetics of the post-onset declinein zinc transporter 8 autoantibodies in type 1 diabetic human subjects. J ClinEndocrinol Metab 2010;95:4712–471931. Merrick WC. The Protein Biosynthesis, Elongation Cycle. New York, ColdSpring Harbor Laboratory Press, 2009

32. Park SG, Park HS, Jeong IK, et al. Autoantibodies against aminoacyl-tRNAsynthetase: novel diagnostic marker for type 1 diabetes mellitus. Biomarkers2010;15:358–36633. Ditzel HJ, Masaki Y, Nielsen H, Farnaes L, Burton DR. Cloning and ex-pression of a novel human antibody-antigen pair associated with Felty’s syn-drome. Proc Natl Acad Sci U S A 2000;97:9234–923934. Orozco G, Eyre S, Hinks A, et al.; UKRAG consortium. Study of the commongenetic background for rheumatoid arthritis and systemic lupus erythematosus.Ann Rheum Dis 2011;70:463–46835. Fransen K, Visschedijk MC, van Sommeren S, et al. Analysis of SNPs withan effect on gene expression identifies UBE2L3 and BCL3 as potential new riskgenes for Crohn’s disease. Hum Mol Genet 2010;19:3482–348836. Harley JB, Alarcón-Riquelme ME, Criswell LA, et al.; International Consor-tium for Systemic Lupus Erythematosus Genetics (SLEGEN). Genome-wide as-sociation scan in women with systemic lupus erythematosus identifiessusceptibility variants in ITGAM, PXK, KIAA1542 and other loci. Nat Genet 2008;40:204–21037. Bonifacio E, Lampasona V, Bernasconi L, Ziegler AG. Maturation of thehumoral autoimmune response to epitopes of GAD in preclinical childhood type 1diabetes. Diabetes 2000;49:202–20838. Kawasaki E, Hutton JC, Eisenbarth GS. Molecular cloning and character-ization of the human transmembrane protein tyrosine phosphatase homologue,phogrin, an autoantigen of type 1 diabetes. Biochem Biophys Res Commun 1996;227:440–44739. Maclaren NK, Huang SW, Fogh J. Antibody to cultured human insulinomacells in insulin-dependent diabetes. Lancet 1975;1:997–100040. Vasu C, Dogan RN, Holterman MJ, Prabhakar BS. Selective induction ofdendritic cells using granulocyte macrophage-colony stimulating factor, butnot fms-like tyrosine kinase receptor 3-ligand, activates thyroglobulin-specificCD4+/CD25+ T cells and suppresses experimental autoimmune thyroiditis.J Immunol 2003;170:5511–552241. Van Belle TL, Juntti T, Liao J, von Herrath MG. Pre-existing autoimmunitydetermines type 1 diabetes outcome after Flt3-ligand treatment. J Autoimmun2010;34:445–45242. Barton A, Woolmore JA, Ward D, et al. Association of protein kinase C alpha(PRKCA) gene with multiple sclerosis in a UK population. Brain 2004;127:1717–172243. Wang H, Munger KL, Reindl M, et al. Myelin oligodendrocyte glycoproteinantibodies and multiple sclerosis in healthy young adults. Neurology 2008;71:1142–114644. Leitges M, Plomann M, Standaert ML, et al. Knockout of PKC alpha en-hances insulin signaling through PI3K [published correction appears in Mol En-docrinol 2002;16:1203]. Mol Endocrinol 2002;16:847–85845. Yamaoka T, Idehara C, Yano M, et al. Hypoplasia of pancreatic islets intransgenic mice expressing activin receptor mutants. J Clin Invest 1998;102:294–30146. Tholey G, Sena AH, Ledig M. Specific insulin-mediated regulation of glu-tamine synthetase in cultured chick astroglial cells. J Neurochem 1986;47:1490–149247. Mahtani MM, Widén E, Lehto M, et al. Mapping of a gene for type 2 diabetesassociated with an insulin secretion defect by a genome scan in Finnish families.Nat Genet 1996;14:90–9448. Farese RV, Standaert ML, Yamada K, et al. Insulin-induced activation ofglycerol-3-phosphate acyltransferase by a chiro-inositol-containing insulin me-diator is defective in adipocytes of insulin-resistant, type II diabetic, Goto-Kakizakirats. Proc Natl Acad Sci U S A 1994;91:11040–11044

3032 Novel Autoantibodies in Type 1 Diabetes Diabetes Volume 63, September 2014