Journal of Autoimmunity - Amazon S3Short communication CD8 T-cell reactivity to islet antigens is unique to type 1 while CD4 T-cell reactivity exists in both type 1 and type 2 diabetes

Short communication

CD8 T-cell reactivity to islet antigens is unique to type 1 while CD4T-cell reactivity exists in both type 1 and type 2 diabetes

Ghanashyam Sarikonda a,1, Jeremy Pettus a,b,1, Sonal Phatak b,Sowbarnika Sachithananthama, Jacqueline F. Miller a, Johnna D. Wesley c, Eithon Cadag d,Ji Chae b, Lakshmi Ganesan b, Ronna Mallios a, Steve Edelman b, Bjoern Peters a,Matthias von Herrath a,c,*

a La Jolla Institute for Allergy and Immunology, La Jolla, CA, USAbUniversity of California San Diego, San Diego, CA, USAcNovo Nordisk Type 1 Diabetes R & D Center, Seattle, WA, USAdAyasdi, Inc., Palo Alto, CA, USA

a r t i c l e i n f o

Article history:Received 12 August 2013Received in revised form6 December 2013Accepted 8 December 2013

Keywords:Type 1 diabetesAntigen specific T-cellsCD8þ T-cellsHLA-Qdot-multimer analysisELISpot

a b s t r a c t

Previous cross-sectional analyses demonstrated that CD8þ and CD4þ T-cell reactivity to islet-specificantigens was more prevalent in T1D subjects than in healthy donors (HD). Here, we examined T1D-associated epitope-specific CD4þ T-cell cytokine production and autoreactive CD8þ T-cell frequency ona monthly basis for one year in 10 HD, 33 subjects with T1D, and 15 subjects with T2D. Autoreactive CD4þ

T-cells from both T1D and T2D subjects produced more IFN-g when stimulated than cells from HD. Incontrast, higher frequencies of islet antigen-specific CD8þ T-cells were detected only in T1D. These ob-servations support the hypothesis that general beta-cell stress drives autoreactive CD4þ T-cell activitywhile islet over-expression of MHC class I commonly seen in T1D mediates amplification of CD8þ T-cellsand more rapid beta-cell loss. In conclusion, CD4þ T-cell autoreactivity appears to be present in both T1Dand T2D while autoreactive CD8þ T-cells are unique to T1D. Thus, autoreactive CD8þ cells may serve as amore T1D-specific biomarker.

! 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Epidemiological studies indicate that the incidence of type 1(T1D) and type 2 diabetes (T2D) is increasing [1,2]. While bothconditions share the end result of hyperglycemia, they are classi-cally defined as distinct clinical entities. Generally, T1D is consid-ered an autoimmune disease [3] while T2D is thought to be ametabolic disorder driven by insulin resistance. However, recentpostmortem studies suggest that b-cell destruction occurs even inT2D subjects [4,5]. Additional studies have found evidence of isletautoimmunity in patients with T2D [6,7]. Collectively, these find-ings support the hypothesis that islet-directed autoimmunity couldplay a role in both T1D and T2D. This is evident clinically with theexistence of overlapping syndromes such as latent autoimmune

diabetes of adulthood (LADA), lean type 2 diabetics, and ketosis-prone T2D. Our ability to detect autoimmunity in diabetes iscurrently limited to determining the presence and titer of classicT1D autoantibodies. While these antibodies have proven invaluablein predicting disease onset, their ability to predict disease coursepost-onset is severely limited (discussed in Ref. [8]). Therefore, aneed exists to develop a biomarker that will indicate underlyingautoimmunity, changes as the disease progresses, and can be fol-lowed serially for evidence of therapeutic efficacy.

In T1D subjects, frequencies of islet antigen reactive CD4þ, andCD8þ T-cells [9e12] are higher compared to healthy donors (HD),and thus, have the potential to function as such a biomarker (dis-cussed in Ref. [13]). However, how these parameters differ betweenpatients with T1D and T2D is largely unknown. Further, previouscross-sectional studies have not taken into account the inherentbiological variation within a given subject that could lead tomisinterpretation of results due to stochastic fluctuations (dis-cussed in Ref. [14]). In the current report, we followed subjects withT1D or T2D and HDs over a one-year period with monthly blooddraws. At each visit, we determined IFN-g and IL-10 cytokine

* Corresponding author. La Jolla Institute for Allergy and Immunology, La Jolla,CA, USA. Tel.: þ1 858 752 6817; fax: þ1 858 752 6993.

E-mail addresses: [email protected], [email protected] (M. von Herrath).1 These authors contributed equally.

Contents lists available at ScienceDirect

Journal of Autoimmunity

journal homepage: www.elsevier .com/locate/ jaut imm

0896-8411/$ e see front matter ! 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.jaut.2013.12.003

Journal of Autoimmunity xxx (2013) 1e6

Please cite this article in press as: Sarikonda G, et al., CD8 T-cell reactivity to islet antigens is unique to type 1while CD4 T-cell reactivity exists inboth type 1 and type 2 diabetes, Journal of Autoimmunity (2013), http://dx.doi.org/10.1016/j.jaut.2013.12.003

production by CD4þ T-cells in response to diabetes-associated an-tigens, and frequencies of autoreactive CD8þ T-cells against HLA-A2-restricted antigenic epitopes. To improve the robustness of thedata and reduce the influence of inherent subject-specific variation,data from all available visits for each subject were averaged. Ourintent in this studywas to determine features of autoimmunity thatare potentially shared between both disease states and those thatare unique toT1D. If a unique feature of T1D could be identified, thiswould help to define a true biomarker for T1D. We found that CD4þ

T-cell reactivity to islet antigens was shared between T1D and T2Dwhile increased CD8þ T-cell autoreactivity was unique to subjectswith T1D.

2. Materials and methods

2.1. Subject recruitment, study design, and scheduled blood draws

We enrolled subjects clinically defined as having T1D (n¼ 33) orT2D (n ¼ 15), along with HD (n ¼ 10) as controls. Patients wereclassified as T1D or T2D by their referring practitioner usingcommonly accepted clinical criteria including age at diagnosis,initial presentation, autoantibody status, family history, depen-dence on exogenous insulin, and BMI. Diabetic donors wererecruited at the Veterans Affairs (VA) hospital in San Diego, La JollaInstitute for Allergy and Immunology (LIAI), and at Taking Control ofYour Diabetes educational conferences. Protocols were approved by

the University of California-San Diego and LIAI Institutional ReviewBoards. Informed consent, study identification numbers, clinicalcase histories, and other information were collected and recordedby clinical investigators. HDs were recruited from a normal blooddonor program at LIAI and did not require separate consent forms.To maintain the anonymity of HDs, only the age and sex of thedonors were recorded. Use of glucocorticoids at the time of studyserved as an exclusion criterion. HD and T1D cohorts in this studyhave also been described in a related paper (Sarikonda et al., sub-mitted to PLoS ONE). Each subject had w40 mL non-fasting blooddrawn once every month for 12 months.

2.2. PBMC isolation

Isolation of PBMCs using FicollePaque (Sigma) density gradientcentrifugation was performed as described [15]. PBMCs wereadjusted to a concentration of 8 # 106/mL in AIM-V medium(Invitrogen Life Sciences, San Diego, CA) and used fresh for ELISpotor cryopreserved in AIM-V medium containing 40% FCS and 10%DMSO and stored in liquid nitrogen for up to 18-months prior toanalysis.

2.3. HLA-typing

Donors were first screened using an anti-HLA-A2 antibody(eBioscience, San Diego, CA) and flow cytometry. Those who werepositive were further typed using genomic DNA. All enrolled do-nors had HLA-DR typing performed. HLA-A2 and HLA-DR typingfrom genomic DNA was performed as described [16].

2.4. ELISpot

The ELISpot assay was performed according to the originalstandardized reports [9] with minor modifications. High purity(>95%, Genscript) DR4-restricted epitopes of IA-2 (752e775) andproinsulin (C19eA3), as well as the DR3-restricted GAD65 epitope(335e352), were used to stimulate PBMCs [15]. A pool of immu-nogenic viral peptides (CEF) was used as a positive control (Mab-tech Inc., Mariemont, OH) and cells incubated only in AIM-Vmedium served as a negative controls. Development of IFN-g andIL-10 spots was performed using U-Cytech antibodies and reagentsas described [15]. Spots above 65-mM in size were counted as

Fig. 1. CD4þ T-cell autoreactivity exists in both T1D and T2D subjects. A shows response among all donors irrespective of HLA-DR4 or -DR3 positivity while B shows data only fromDR4þ donors. A: Net IFN-g and IL-10 production (spots with antigen e spots in media) against DR4-restricted epitopes of IA-2 (752e775), proinsulin (C19eA3) and a DR3-restrictedepitope of GAD65 (335e352), were pooled at each visit and averaged over all visits. Statistical significance between the groups was determined using ManneWhitney test. B: IFN-g(top panels) or IL-10 (bottom panels) production against IA-2752e775 (left half) or PIC19eA3 (right half) was assessed as net spot numbers (spots with antigen e spots in media), or asstimulation index, SI (spots with antigen O spots in media). Mean net spot numbers.

Table 1Participant information is shown including male to female ratio, age, and diseaseduration.

HD T1D T2D

Number of males (%) 1 (10) 18 (55) 10 (67)Age (Yrs) 31.3 $ 5.27 46.06 $ 15.91 60.47 $ 8.48Disease duration (Yrs) N/A 18.55 $ 14.93 14.57 $ 1.98BMI N/A 25.39 $ 3.04 35.43 $ 7.00Insulin dose (units) N/A 40.48 $ 16.6 90.09 $ 75.11TDD insulin (U/kg) N/A 0.525 $ 0.22 0.75 $ 0.49HbA1c (%) N/A 7.039 $ 1.22 7.45 $ 0.26DRB1*04 e number (%) 0 (0) 19 (58) 4 (27)DRB1*03 e number (%) 1 (14) 11 (36) 1 (7)

Mean $ SD values are shown. N/A represents not available (BMI and HbA1c for HD)or not applicable (Duration and Insulin usage for HD).

G. Sarikonda et al. / Journal of Autoimmunity xxx (2013) 1e62


cytokine spots using the KS-ELISPOT reader [16] (Zeiss, Munich,Germany). All epitopes were run in triplicate and averaged.

2.5. Qdot-HLA-A2-multimer assay

HLA-A2-restricted peptides (>95% purity, Genscript) from isletantigens and their labeling with quantum dots (Qdots) have beendescribed previously [12]: insulin (B-chain, aa 10e18), IA-2 (aa797e805), IGRP (aa 265e273), PPI (aa 15e23), GAD65 (aa 114e123), and ppIAPP (aa 5e13). Peptide-loaded HLA-A*0201 mono-mers were produced by the NIH Tetramer Core Facility (EmoryUniversity, Atlanta, GA), according to standard protocols, http://tetramer.yerkes.emory.edu/client/protocols. Streptavidin-conjugated Qdots (Qdot-585, -605, -655, -705, and -800; Invi-trogen, San Diego, CA) were added to achieve a 1:20 streptavidin-Qdot:biotinylated-pHLA ratio to generate multimeric HLA-A2-peptide complexes. Each HLA-A2-monomer was labeled with twodifferent Qdots and each peptide-multimer had its own uniqueQdot combination [12].

For each HLA-A2þ subject, samples from all available visits werelabeled, acquired and analyzed simultaneously, as described [12].Briefly, cryopreserved PBMC were thawed and 2 # 106 cells wereimmediately stained first with all Qdot-labeled multimers (0.1 mgeach multimer) followed by Alexa Flour-700-CD8 along with FITC-labeled anti-CD4, -CD14, -CD20, -CD40, and -CD16 antibodies(eBiosciences, San Diego, CA) and resuspended in PBS supple-mented with 0.5% BSA and containing 7-AAD (eBioscience, SanDiego, CA, USA). Sample data were acquired using an LSRII (BDBiosciences, San Diego, CA) with the following filter settings; for the488-nm laser: Qdot 655, 635LP, and 655/20. For the 405-nm laser:QD800, 770LP, 800/30; QD705, 685LP, 710/40; QD605, 595LP, 605/20 and QD585, blank, and 585/22.

2.6. Statistical analysis

All data were analyzed and statistical analyses performed usingGraphPad Prism 5 software (GraphPad Software, Inc.). Additionaltopological-based data analyses were conducted using the AyasdiIRIS software (Ayasdi, Inc.). Statistical significance was determinedusing Spearman’s Rho test, ManneWhitney test, unpaired t-test,and ANOVA.

3. Results and discussion

Herein we described CD4þ and CD8þ T-cell islet antigen reac-tivity in a cohort of patients with T1D compared to T2D and HDs.Our study design provided a robust amount of data in that eachpatient was followed for one year withmonthly blood draws, and toour knowledge, is the first study comparing these immunologicalparameters between disease states longitudinally. We found thatonly subjects with T1D consistently had increased numbers ofcirculating, autoreactive CD8þ T-cells yet both T1D and T2D sub-jects had similar autoreactive CD4þ T-cell-derived cytokineproduction.

We recruited 33 subjects with T1D, 15 with T2D, and 10 HD ascontrols. A summary of participants’ baseline characteristics isshown in Table 1. Detailed subject profiles with age, duration ofdisease, HLA phenotypes, and autoantibody status are listed inTable S1. A higher proportion of T1D donors were HLA-DR4þ andHLA-A2þ (58% and 49%) compared toT2D donors (27% each) or HDs(14% and 40%); this was expected given the known associationbetween HLA and T1D [17e19].

Extensive validation of the indirect ELISpot assay [9] showedlow intra- and inter-assay co-efficient of variation in assay perfor-mance (7.3% and 13%, respectively, data not shown). We then

determined IFN-g and IL-10 cytokine production from CD4þ T-cellsafter stimulation with two DR4-restricted peptides (PIC19eA3 & IA-2752e775) and one DR3-restricted epitope (GAD65335e352). As pre-vious studies have shown that these peptides are also reactive insome DR4/DR3-negative donors, we included data from all donorsregardless of their DR3/4 status in this first analysis. Net spotnumbers (spots with antigen e spots in media) against all threeepitopes were pooled at each visit and averaged over all visits.Surprisingly, T2D subjects had significantly more IFN-g producingCD4þ T-cells than controls (Fig. 1A, left panel). In contrast, T1Dsubjects (but not T2D subjects) had more IL-10 producing CD4þ

T-cells than control subjects (Fig. 1A, right panel). Importantly,while cytokine production differed between control subjects andthose with diabetes, there was no statistically significant differencebetween the T1D and T2D cohorts in either IFN-y or IL-10 pro-duction. Further, in T2D subjects, higher IFN-g production by CD4þ

T-cells suggests the existence of either general inflammation or adegree of underlying autoimmunity. In support of the latter, a

Fig. 2. Ag-specific CD8þ T-cell numbers tend to be higher in HLA-A2þ T1D donors thancontrol or T2D donors. A: Mean frequencies of each antigen-specific CD8þ T-cellsaveraged over all visits among HD, T1D, and T2D cohorts are shown. B: At each visit,CD8þ T-cell frequencies against each antigenic epitope were counted as a positiveevent when they were >0.01% of total CD8þ T-cells. Number of antigen-specific eventsdetected over all visits was pooled for each antigen. C: Antigen-specific CD8þ T-cellpositive events against all six antigenic-epitopes, detected over all visits were pooled.D: At each visit, CD8þ T-cell frequencies were counted as a positive event when theywere >0.02% of CD8þ T-cells. Cumulative number of antigen-specific CD8þ T-cellpositive events against all six antigenic-epitopes, detected over all visits is shown.Statistical values shown in C and D are from unpaired Student’s t-test.

G. Sarikonda et al. / Journal of Autoimmunity xxx (2013) 1e6 3


Fig. 3. Distinct populations comprise the T1D and T2D cohorts and highlight the existence of unique immune profiles. A: Distinct sub-groups within each disease type wereidentified when the data (ELISpot þ Clinical Parameters þ Additional Immune Response Data) were used to construct the topological network. The similarity metric and mathematicalfunctions used to generate the topological network were normalized cosine (metric) and principal and secondary principal components (embedded). B: The subgroups identified inA differed from each other in the mean total spot count, regardless of stimulating antigen.



significant proportion of T2D patients exhibited T-cell proliferationwhen exposed to human islet preparations [20].

To analyze our data more stringently, we assessed cytokineproduction against the DR4-restricted epitopes (PIC19eA3 & IA-2752e775) in only DR4þ subjects (T1D n ¼ 19, T2D n ¼ 4). As with theanalysis of entire cohort above, there were no significant differ-ences in cytokine production between T1D and T2D cohorts in thissubgroup whether we assessed the average of net spots or Stimu-lation Index (SI) values over all visits (Fig. 1B). Thus, CD4þ T-cellsfrom HLA-DR4þ T1D or T2D subjects appear to elicit similar cyto-kine production in response to islet antigens. This finding argues fora potential shared component of immune dysregulation betweenthe disease states.

Next, to determine antigen-specific CD8þ T-cell frequencies, weemployed HLA-A2-Qdot-multimer assay [12]. Mean frequencies ofeach of the six epitope-specific CD8þ T-cells averaged over all visitswere higher in the T1D cohort compared to HD or T2D cohorts(Fig. 2A and data not shown). Using a positivity cutoff of 0.01% ofCD8þ T cells [12], the T1D cohort had greater number of visits(events) positive, in all but one of the antigenic epitopes, comparedto HD and T2D cohorts (Fig. 2B and data not shown). Further, whenall antigen specificities were pooled, the distribution and meannumber of positive events were higher in T1D subjects (Fig. 2C);however, these differences were not statistically significant. Uponsetting a higher positivity threshold (>0.02% of CD8þ T-cells), theincrease in CD8þ T-cell frequencies in T1D donors became apparentas they had significantly higher number of positive eventscompared to T2D and control donors. Taken together, this datasuggests the increase in frequency of islet antigen-specific CD8þ T-cells is specific to T1D subjects, compared to HDs or T2D subjects.

While both T1D and T2D result in relative insulin deficiency, theunderlying b-cell destruction in T1D is muchmore rapid and severe(reviewed in Ref. [21]). Additional analyses using a topologicalapproach for detecting nonlinear patterns in high-dimensional datain an unsupervised fashion [22,23] revealed the possibility ofdistinct subpopulation clusters within the T1D and T2D cohorts,but not the HD group (Fig. 3A). These unique clusters were char-acterized by cytokine response, hinting at underlying differences inimmune function within both T1D and T2D populations that mayimpact disease progression (Fig. 3B).

We hypothesize that CD4þ T-cell activation is common to bothforms of diabetes resulting in a mild degree of b-cell loss and/ordysfunction. There is a range of inflammation that was easily seenin the topological analyses but has not been appreciated previously.In contrast, CD8þ T-cell activation is a feature unique to T1D andleads to the massive b-cell destruction characteristic of this disease.We propose a simplifiedmodel (Fig. S1) inwhich a primary insult tothe islet results in release of autoantigens. This initial insult may bea metabolic derangement such as obesity-induced lipotoxicity and/or dysregulated Kþ ATP channel [24e26] for T2D; whereas in T1D itmay be environmental or viral infections [27,28]. The exact mech-anisms for b-cell injury are multifactorial and complex but,regardless of the insult, the released autoantigens are taken up byantigen presenting cells (APCs) that migrate to pancreatic lymphnodes and activate T-cells. CD4þ T-cells are activated in both T1Dand T2D and result in inflammation that promotes functional b-celldecline. The initiation of CD8þ T-cell activation, in contrast, isspecific to the T1D disease process. It has been shown that themajority of patients with T1D have CD8þ T-cell infiltrates in theislets. This is most pronounced in patients close to diagnosis butalso in a portion of patients with longstanding disease [29]. Simi-larly, cadaveric pancreatic samples from patients with T1D displayupregulation of MHC class I up to 8-years post-diagnosis [29]. Wepropose that a possible secondary insult in individuals geneticallypredisposed to develop T1D leads to upregulation of MHC I on the

islets and directed CD8þ T-cell killing. The resulting release of b-cellantigens perpetuates the attack by recruiting more CD8þ T-cellswith different antigen specificities. This process could happenslowly over time in response to repeated insults as has recentlybeen proposed [30]. While a proportion of T1D subjects did nothave significant, detectable circulating autoreactive CD8þ T-cells inour study, the presence of such cells was almost exclusively foundin T1D. In those T1D subjects without detectable autoreactive CD8þ

T-cells, the frequency of these T-cells may wane as more islets aredestroyed, reducing the antigenic load. This would explain previ-ously published observation that some, but not all, pancreaticsamples of T1D patients had insulitis [29].

The amount of remaining b-cells clearly differs between type 1and type 2 diabetics, although recent investigations have shownthat some b-cells regularly survive even in type 1 patients over longperiods of time [31]. Induction of autoreactive T cells and theirmaintenance clearly has to occur after exposure to islet antigens.Two aspects of our study are, in this respect, novel and remarkable:First, CD8þ T-cells reacting with islet antigens are only found intype 1 patients, indicating over-expression of MHC class I on b-cellsin T1D [29] is instrumental for their sensitization. Second, CD4autoreactivity is present in both types of diabetes, suggesting thatmere b-cell stress can lead to presentation of autoantigens via theMHC class II pathway.

The current report indicates that CD4þ T-cell autoreactivity maybe common to both T1D and T2D, and could represent a sharedmechanism of b-cell dysfunction. Conversely, autoreactive CD8þ T-cells appear to be unique to T1D and may be responsible for theaccelerated loss of b-cell mass. This evidence gives credence to theconcept that the different forms of diabetes represent a continuumfrom autoimmunity to metabolic disturbances with some over-lapping and some unique features for each disease. Thus,measuring CD8þ autoreactivity in T1D subjects could serve as amore specific immune biomarker to study disease progression and,possibly, response to therapeutics in future clinical trials.

In conclusion, the data presented here demonstrates that thereis a shared autoimmune component between T1D and T2D. Thecomplexity of both diseases is further highlighted by topologicalanalyses; the emergence of distinct subgroups underscores theneed for better understanding of the impact of the diabetogenicimmune response on disease progression and therapeutic out-comes. Identifying and understanding immune-based differenceswithin the spectrum of diabetesdbeyond T1D and T2Ddwill ulti-mately facilitate the development and use of more targeted,personalized therapies.

Funding

GS was supported by a postdoctoral fellowship from the T1Dcenter of San Diego. This work was supported by a subawardnumber 3002074563 from University of Michigan, Ann Arbor(UMich), to MvH, issued under NIH award number 3UL1RR024986-05S1 to UMich under the direction of Dr. Thomas P. Shanley, P.I.

Conflict of interest

The authors declare that there are no conflicts of interest per-taining to this manuscript.

Acknowledgments

We thank Ken Coppieters, (NNRC, Seattle, WA) and Bart Roep(LUMC, Leiden, Netherlands) for many helpful discussions andadvice in manuscript preparation; Joana RF Abreu and Bart Roep(LUMC, Leiden, Netherlands) for assistance with the Qdot-HLA-

G. Sarikonda et al. / Journal of Autoimmunity xxx (2013) 1e6 5


multimer assay; Tobias Boettler, Darius Schneider, Renee Parkerand Mark Bouchard (LIAI) and the nurses at UCSD VA hospital fortheir assistance in obtaining blood samples for this study; DeniseBaker and Duy Le (LIAI) for their assistancewith HLA typing. Finally,we would like to thank the NIH tetramer core facility at EmoryUniversity for their help in providing HLA-A2-peptide monomers.

GS, JP, SS, SP JFM, and JC participated in obtaining samples andperforming assays. GS, JP, RM, BP, JDW, EC, and MvH performed thedata analysis. GS, JP, BP and MvH wrote the manuscript. GS, SE andMvH conceptualized and designed the project. GS, JP and MvH arethe guarantors of this work and, as such, had full access to all thedata in the study and take responsibility for the integrity of the dataand accuracy of data analysis.

Parts of the data shown in this manuscript were presented at theImmunology of Diabetes Society (IDS) conference 2012, held inVictoria, British Columbia, Canada.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.jaut.2013.12.003

References

[1] Fox CS, Pencina MJ, Meigs JB, Vasan RS, Levitzky YS, D’Agostino Sr RB. Trendsin the incidence of type 2 diabetes mellitus from the 1970s to the 1990s: theFramingham Heart Study. Circulation 2006;113:2914e8.

[2] Gale EA. The rise of childhood type 1 diabetes in the 20th century. Diabetes2002;51:3353e61.

[3] Eisenbarth GS. Type I diabetes mellitus. A chronic autoimmune disease. N EnglJ Med 1986;314:1360e8.

[4] Rahier J, Guiot Y, Goebbels RM, Sempoux C, Henquin JC. Pancreatic beta-cellmass in European subjects with type 2 diabetes. Diabetes Obes Metab2008;10(Suppl. 4):32e42.

[5] Saisho Y, Butler AE, Meier JJ, Monchamp T, Allen-Auerbach M, Rizza RA, et al.Pancreas volumes in humans from birth to age one hundred taking intoaccount sex, obesity, and presence of type-2 diabetes. Clin Anat 2007;20:933e42.

[6] Pietropaolo M, Barinas-Mitchell E, Pietropaolo SL, Kuller LH, Trucco M. Evi-dence of islet cell autoimmunity in elderly patients with type 2 diabetes.Diabetes 2000;49:32e8.

[7] Winer DA, Winer S, Shen L, Wadia PP, Yantha J, Paltser G, et al. B cells promoteinsulin resistance through modulation of T cells and production of pathogenicIgG antibodies. Nat Med 2011;17:610e7.

[8] Mallone R, Roep BO. Biomarkers for immune intervention trials in type 1diabetes. Clin Immunol 2013;149(3):286e96. http://dx.doi.org/10.1016/j.clim.2013.02.009. Epub 2013 Feb 27.

[9] Arif S, Tree TI, Astill TP, Tremble JM, Bishop AJ, Dayan CM, et al. Autoreactive Tcell responses show proinflammatory polarization in diabetes but a regulatoryphenotype in health. J Clin Invest 2004;113:451e63.

[10] Long SA, Cerosaletti K, Bollyky PL, Tatum M, Shilling H, Zhang S, et al. Defectsin IL-2R signaling contribute to diminished maintenance of FOXP3 expressionin CD4(þ)CD25(þ) regulatory T-cells of type 1 diabetic subjects. Diabetes2010;59:407e15.

[11] Penaranda C, Tang Q, Bluestone JA. Anti-CD3 therapy promotes tolerance byselectively depleting pathogenic cells while preserving regulatory T cells.J Immunol 2011;187:2015e22.

[12] Velthuis JH, Unger WW, Abreu JR, Duinkerken G, Franken K, Peakman M, et al.Simultaneous detection of circulating autoreactive CD8þ T-cells specific fordifferent islet cell-associated epitopes using combinatorial MHC-multimers.Diabetes 2010;59(7):1721e30. Published online 2010 March 31. http://dx.doi.org/10.2337/db09-1486.

[13] Roep BO, Peakman M. Surrogate end points in the design of immunotherapytrials: emerging lessons from type 1 diabetes. Nat Rev Immunol 2010;10:145e52.

[14] von Herrath M, Sanda S, Herold K. Type 1 diabetes as a relapsing-remittingdisease? Nat Rev Immunol 2007;7:988e94.

[15] Sanda S, Roep BO, von Herrath M. Islet antigen specific IL-10þ immune re-sponses but not CD4þCD25þFoxP3þ cells at diagnosis predict glycemiccontrol in type 1 diabetes. Clin Immunol 2008;127:138e43.

[16] Oseroff C, Sidney J, Vita R, Tripple V, McKinney DM, Southwood S, et al. T cellresponses to known allergen proteins are differently polarized and accountfor a variable fraction of total response to allergen extracts. J Immunol2012;189:1800e11.

[17] Fennessy M, Metcalfe K, Hitman GA, Niven M, Biro PA, Tuomilehto J, et al.A gene in the HLA class I region contributes to susceptibility to IDDM in theFinnish population. Childhood Diabetes in Finland (DiMe) Study Group. Dia-betologia 1994;37:937e44.

[18] Nerup J, Platz P, Andersen OO, Christy M, Lyngsoe J, Poulsen JE, et al. HL-Aantigens and diabetes mellitus. Lancet 1974;2:864e6.

[19] Park YS, Wang CY, Ko KW, Yang SW, Park M, Yang MC, et al. Combinations ofHLA DR and DQ molecules determine the susceptibility to insulin-dependentdiabetes mellitus in Koreans. Hum Immunol 1998;59:794e801.

[20] Brooks-Worrell BM, Reichow JL, Goel A, Ismail H, Palmer JP. Identification ofautoantibody-negative autoimmune type 2 diabetic patients. Diabetes Care2011;34:168e73.

[21] Cnop M, Welsh N, Jonas JC, Jorns A, Lenzen S, Eizirik DL. Mechanisms ofpancreatic beta-cell death in type 1 and type 2 diabetes: many differences,few similarities. Diabetes 2005;54(Suppl. 2):S97e107.

[22] Carlsson G. Topology and data. Bull Am Math Soc 2009;46:255e308.[23] Lum PY, Singh G, Lehman A, Ishkanov T, Vejdemo-Johansson M, Alagappan M,

et al. Extracting insights from the shape of complex data using topology. SciReports 2013;3:1236.

[24] Gloyn AL, Hashim Y, Ashcroft SJ, Ashfield R, Wiltshire S, Turner RC, et al.Association studies of variants in promoter and coding regions of beta-cellATP-sensitive K-channel genes SUR1 and Kir6.2 with Type 2 diabetes melli-tus (UKPDS 53). Diabet Med 2001;18:206e12.

[25] Gloyn AL, Weedon MN, Owen KR, Turner MJ, Knight BA, Hitman G, et al.Large-scale association studies of variants in genes encoding the pancreaticbeta-cell KATP channel subunits Kir6.2 (KCNJ11) and SUR1 (ABCC8) confirmthat the KCNJ11 E23K variant is associated with type 2 diabetes. Diabetes2003;52:568e72.

[26] Hani EH, Boutin P, Durand E, Inoue H, Permutt MA, Velho G, et al. Missensemutations in the pancreatic islet beta cell inwardly rectifying Kþ channelgene (KIR6.2/BIR): a meta-analysis suggests a role in the polygenic basis ofType II diabetes mellitus in Caucasians. Diabetologia 1998;41:1511e5.

[27] Filippi C, von Herrath M. How viral infections affect the autoimmune processleading to type 1 diabetes. Cell Immunol 2005;233:125e32.

[28] Filippi CM, von Herrath MG. Viral trigger for type 1 diabetes: pros and cons.Diabetes 2008;57:2863e71.

[29] Coppieters KT, Dotta F, Amirian N, Campbell PD, Kay TW, Atkinson MA, et al.Demonstration of islet-autoreactive CD8 T cells in insulitic lesions from recentonset and long-term type 1 diabetes patients. J Exp Med 2012;209:51e60.

[30] Skog O, Korsgren S, Melhus A, Korsgren O. Revisiting the notion of type 1diabetes being a T-cell-mediated autoimmune disease. Curr Opin EndocrinolDiabetes Obes 2013;20:118e23.

[31] Keenan HA, Sun JK, Levine J, Doria A, Aiello LP, Eisenbarth G, et al. Residualinsulin production and pancreatic ss-cell turnover after 50 years of diabetes:Joslin Medalist Study. Diabetes 2010;59:2846e53.



Documents

Journal of Autoimmunity - Amazon S3Short communication CD8 T-cell reactivity to islet antigens is unique to type 1 while CD4 T-cell reactivity exists in both type 1 and type 2 diabetes