Increase in Pancreatic Proinsulin and Preservation of β ... · Increase in Pancreatic Proinsulin and Preservation of b-Cell Mass in Autoantibody-Positive Donors Prior to Type 1 Diabetes

Increase in Pancreatic Proinsulin and Preservation ofb-Cell Mass in Autoantibody-Positive Donors Prior toType 1 Diabetes OnsetTeresa Rodriguez-Calvo1 Jose Zapardiel-Gonzalo1 Natalie Amirian1 Ericka Castillo1 Yasaman Lajevardi1

Lars Krogvold2 Knut Dahl-Joslashrgensen2 and Matthias G von Herrath13

Diabetes 2017661334ndash1345 | DOI 102337db16-1343

Type 1 diabetes is characterized by the loss of insulinproduction caused by b-cell dysfunction andor destruc-tion The hypothesis that b-cell loss occurs early duringthe prediabetic phase has recently been challenged Herewe show for the first time in situ that in pancreas sec-tions from autoantibody-positive (Ab+) donors insulinarea and b-cell mass are maintained before diseaseonset and that production of proinsulin increases Thissuggests that b-cell destruction occurs more precipi-tously than previously assumed Indeed the pancreaticproinsulin-to-insulin area ratio was also increased inthese donors with prediabetes Using high-resolutionconfocal microscopy we found a high accumulation ofvesicles containing proinsulin in b-cells from Ab+ donorssuggesting a defect in proinsulin conversion or an accu-mulation of immature vesicles caused by an increase ininsulin demand andor a dysfunction in vesicular traffick-ing In addition islets from Ab+ donors were larger andcontained a higher number of b-cells per islet Our dataindicate that b-cell mass (and function) is maintained untilshortly before diagnosis and declines rapidly at the time ofclinical onset of disease This suggests that secondaryprevention before onset when b-cell mass is still intactcould be a successful therapeutic strategy

Type 1 diabetes is defined as an autoimmune disease inwhich clinical symptoms arise as a result of b-cell loss Ge-netic and environmental factors might render b-cells suscep-tible to attack by the immune system or could contributeto b-cell dysfunction (12) More than three decades ago

Eisenbarth and colleagues (3) described a linear loss offirst-phase insulin release after intravenous glucose adminis-tration in individuals with islet-cell antibodies who weremonitored for 10 years before diagnosis However elevationsin fasting blood glucose and peak glucose during oral glucosetolerance tests were only seen in the year before onset Thissustained loss of b-cell function in individuals with predia-betes strongly correlated with the time to overt diabetes andled to Eisenbarthrsquos (4) landmark article in which the stagesof type 1 diabetes were presented and the steady de-crease in insulin secretion was linked to a linear reductionin b-cell mass that continued after diagnosis

Although this model remained a reference for manyyears new studies have suggested that b-cell mass is notlost in a linear fashion during the prediabetic phase and adebate about the discrepancy between b-cell mass and func-tion ensued (2) Subsequent studies have also detected aloss of glucose tolerance in the months preceding diagnosis(56) b-Cell dysfunction might occur early in the diseaseprocess at the point at which the individual becomes auto-antibody positive (Ab+) but an actual decline in b-cell massmight occur later In the Diabetes Virus Detection (DiViD)study a transient b-cell dysfunction was detected in livecells obtained at diagnosis which improved in a nondiabeticculture milieu (7) Increasing dysfunction would prompt anincrease in insulin demand (89) which could eventuallycause a more cataclysmic decline in b-cell mass aroundthe clinical onset of diabetes However the cause of thedecline in function and the precise time course of eventshave remained largely undefined

1Type 1 Diabetes Center La Jolla Institute for Allergy and Immunology La JollaCA2Division of Paediatric and Adolescent Medicine Oslo University Hospital andFaculty of Medicine University of Oslo Oslo Norway3Novo Nordisk Diabetes Research amp Development Center Seattle WA

Corresponding author Matthias G von Herrath matthiasljiorg or mtvhnovonordiskcom

Received 2 November 2016 and accepted 26 January 2017

This article contains Supplementary Data online at httpdiabetesdiabetesjournalsorglookupsuppldoi102337db16-1343-DC1

copy 2017 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 httpwwwdiabetesjournalsorgcontentlicense

1334 Diabetes Volume 66 May 2017

PATHOPHYSIO

LOGY

Studies from the Network for Pancreatic Organ Donorswith Diabetes (nPOD) have recently shown that b-cellmass is not diminished in Ab+ donors and that singleb-cells and islets containing insulin can be found in do-nors with long-standing type 1 diabetes (10) The timecourse from seroconversion to onset of clinical diabeteshas been further characterized in longitudinal studiesAfter autoantibody seroconversion 145 of single Ab+and 679 of multiple Ab+ patients progressed to type1 diabetes in a 10-year follow-up study in three geograph-ically different cohorts (11) Another study also revealedthat 11 of multiple Ab+ children would progress to clin-ical disease each year (12) However the exact triggersand progression to clinical onset are not fully understood

Proinsulin is an important autoantigen in type 1 diabetesin humans and mice (13) because it shapes the autoreactiveCD8 T-cell repertoire (1415) Importantly recent studieshave shown that several epitopes within its precursor (pre-proinsulin) and proinsulin itself are recognized by islet-infiltrating CD4 andor CD8 T cells isolated from patientswith type 1 diabetes (16ndash20) suggesting a potential rolefor this antigen in disease pathogenesis Preproinsulin isprocessed into proinsulin and signal peptide (21) Only amarginal fraction of proinsulin is secreted to the circulationbut it accounts for 30ndash50 of the protein production inb-cells and increases in response to higher insulin demandBecause of this high metabolic demand b-cells are prone toendoplasmic reticulum (ER) stress and proinsulin misfold-ing which could lead to b-cell failure (22) ER stress mayalso be induced by viral infection (23) which was recentlydetected in the islets of Langerhans at diagnosis (24)Interestingly Cianciaruso et al (25) demonstrated thatcytokine-induced ER stress enhances the exosomal releaseof proinsulin Several reports present evidence of high circu-lating proinsulin and proinsulin intermediates with or with-out accompanying hyperglycemia in patients at risk fordeveloping the disease and after diagnosis (2627) Proinsulinand proinsulinndashtondashC-peptide (PI-to-C) ratios in combinationwith autoantibody concentration have been suggested as po-tential biomarkers for type 1 diabetes capable of identifying

with high sensitivity individuals at risk for developing disease1 to 40 months before clinical onset (2829) However thelink between proinsulin levels in serum and their content inthe pancreas in b-cells themselves has not been investigatedin human specimens in situ

The objective of this study was to add refinement to theclassic model of linear b-cell loss and to fill an importantgap in our understanding of the prediabetic phase in humantype 1 diabetes We studied for the first time in the hu-man pancreas the distribution of proinsulin in single- anddouble-Ab+ individuals preceding the onset of disease as wellas in patients with recent-onset type 1 diabetes and corre-lated it with loss of insulin content b-cell mass and PIareandashtondashinsulin (PI-to-INS) area ratio These studies under-line the need for development of biomarkers as well as forpreventive therapies focusing on normalizing b-cell dys-function during the prediabetic stage

RESEARCH DESIGN AND METHODS

SubjectsHuman pancreas sections were collected from cadavericorgan donors through nPOD Six micrometer formalin-fixed paraffin-embedded sections from the head bodyand tail of the pancreas were obtained from controldonors without diabetes who were single Ab+ (n = 8)double Ab+ (n = 5) and Ab2 (n = 9) and from onedonor at onset of type 1 diabetes In addition 4-mmformalin-fixed paraffin-embedded sections from living do-nors with type 1 diabetes (n = 6) were obtained throughthe DiViD study by tail resection (30) Overall 22 sectionsfrom the head 22 from the body and 29 from the tail ofthe pancreas were analyzed (n = 73 sections) Table 1summarizes the demographic information for each groupDetailed donor information can be found in Supplemen-tary Table 1 All experimental procedures were approvedby the La Jolla Institute for Allergy and ImmunologyInstitutional Review Board approved protocol numberDI3-054-0216 For the DiViD study participants providedwritten informed consent and more details can be foundin Krogvold et al (30)

Table 1mdashDonor demographic information including age percentage of males and females ethnicity BMI disease duration andC-peptide levels

Control(n = 9)

Ab+(n = 13)

Type 1 diabetes(n = 7)

Total(N = 29)

Age mean 6 SD (years) 336 6 125 361 6 135 282 6 49 326 6 41

Sex n ()Female 4 (445) 8 (615) 3 (428) 15 (517)Male 5 (555) 5 (385) 4 (572) 14 (483)

Ethnicity n ()African American 0 (0) 2 (154) 0 (0) 2 (69)Caucasian 7 (777) 8 (615) 7 (100) 22 (759)Hispanic 2 (223) 3 (231) 0 (0) 5 (172)

BMI mean 6 SD (kgm2) 282 6 6 262 6 5 25 6 32 265 6 16

Disease duration mean 6 SD (weeks) 44 6 27

C-peptide mean 6 SD (ngmL) 67 6 72 61 6 62 mdash

diabetesdiabetesjournalsorg Rodriguez-Calvo and Associates 1335

ImmunofluorescencePancreas sections were stained for insulin proinsulin andglucagon after a standard triple indirect immunofluorescencestaining After deparaffinization and rehydration in descend-ing ethanol concentrations sections were exposed to heat-based antigen retrieval (citrate buffer) Staining wasperformed using a polyclonal guinea pig anti-insulin anti-body (1500 Dako Carpinteria CA) monoclonal mouse anti-proinsulin (DSHB clone GS-9A8 150) and monoclonalmouse anti-glucagon (clone K79bB10 1300 Abcam) conju-gated in-house to Alexa Fluor 647 Secondary antibodiesincluded F(ab9)2 fragment of goat anti-guinea pig IgG conju-gated to Alexa Fluor 488 (1800 Jackson ImmunoResearch)and goat anti-mouse IgG (H+L) conjugated to AlexaFluor 555 (11000 Life Technologies Grand Island NY)incubated at room temperature for 30 min Sections werecounterstained with Hoechst (1400 Life Technologies) for10 min and then mounted with ProLong Gold AntifadeMountant (Life Technologies)

Image Acquisition and AnalysisSections were scanned with an Axio Scan Z1 slide scanner(Carl Zeiss Microscopy Thornwood NY) using a fluorescentOrca Flash 40 v2 (HXP 120 V lamp) camera with a 32008NA objective for immunofluorescence Images were acquiredwith ZEN2 software slidescan module Acquisition settingswere kept constant between specimens to allow for quantita-tive comparison between samples ZEN2 software blue editionwas used to process the images before analysis Lower andupper thresholds were defined for each channel For insulinand proinsulin similar thresholds were set for comparisonThen whole-tissue section images were exported and reducedat 30 or 45 depending on their size into tiff files forautomatic software analysis Custom macros were developedto measure tissue area (macro 1) and islet size and count(macro 2) and calculate the percentage of insulin proinsulinand glucagon areas (macro 3 only cytoplasmic staining abovebackground levels was measured) To classify and calculate thenumber of a- and b-cells the image of the whole tissue sec-tion was exported into 203 20 tiff files to improve resolutionand then processed by a different custom macro (macro 4)(Supplementary Fig 1) All of the macros were developed forFiji an image-processing software program developed forImageJ (National Institutes of Health) R software was usedto systematically analyze the data (R is available as free soft-ware under the terms of the Free Software Foundationrsquos GNUGeneral Public License in source code form) In addition threecases per group were analyzed by high-resolution confocal mi-croscopy using an LSM 880 confocal microscope with Airyscantechnology (Carl Zeiss Jena Germany) and a 363 objective

Statistical AnalysisDifferences between group pairs were analyzed with aStudent t test or Mann-Whitney test Group differenceswere analyzed using one-way ANOVA followed by a Holm-Siacutedaacutekmultiple comparisons test or Kruskal-Wallis test fol-lowed by a Dunn multiple comparisons test To assessthe plausibility of using the PI-to-INS area ratio as a

discriminator for disease status a logistic regression wasfitted to the data using R version 330 (9)

The scores for this model were calculated as thedistance from each point (plotted using their percentageof proinsulin and insulin area as coordinates) to the lineproinsulin = insulin (ie y = x) Briefly

score frac14 Insulin2Proinsulinffiffiffi

2p

The significance of the overall model was calculated bycomparing it with a model with just the intercept (ie anull model) Besides the logistic regression a receiver op-erating characteristic curve analysis using the R packageROCR version 10-7 (10) was performed to investigate thepredictive power of the model

Statistical analysis was performed using GraphPadPrism version 6 (GraphPad Software La Jolla CA) Datain graphs and tables are presented as mean 6 SD unlessotherwise indicated Findings were assumed statisticallysignificant at P 005

RESULTS

Proinsulin Area Is Significantly Increased in Ab+ DonorIslets Compared With Islets From Control DonorsWithout Diabetes Whereas Insulin Area RemainsSimilarWe systematically measured insulin proinsulin andglucagon staining from the head body and tail regionsof pancreatic tissue sections Interestingly there was asmall increase in total insulin area in Ab+ donors thatdid not reach statistical significance (Fig 1AndashC leftpanel) Conversely a significant increase in the proinsu-lin area was observed in the head (54 P = 00233)body (57 P = 00143) and tail (60 P = 00087)regions in Ab+ individuals compared with control sec-tions (Fig 1AndashC central panel) Lastly there were nomajor differences in glucagon area between both groupsfor any of the regions (Fig 1AndashC right panel) To betterunderstand whether this was caused by a shift in thesubcellular localization of proinsulin andor by an in-crease in proinsulin content a super highndashresolutionconfocal microscope with Airyscan technology wasused Proinsulin in control donors was mainly localizedclose to the nucleus with a staining pattern consistentwith the Golgi apparatus and was minimally present inother compartments In multiple Ab+ donors proinsulinwas more widely localized to the juxtanuclear region(Golgi) and vesicular compartment confirming a changein subcellular localization (Fig 2 and SupplementaryFig 2)

b-Cell Mass Is Not Reduced in Single or Double Ab+DonorsThe combination of insulin area from the head body andtail sections of the pancreas normalized by the size of therespective tissues was multiplied by the total weight of thepancreas Mean b-cell mass was almost identical in control

1336 Increased PI Area in Donors With Prediabetes Diabetes Volume 66 May 2017

and Ab+ donors (24506 605 mg control donors vs 26776802 mg Ab+) (Fig 3A) However when proinsulin area wasused as reference instead of insulin a significant increasein b-cell mass in the Ab+ donor group was seen (1874 6497 mg control donors vs 26696 997 mg Ab+) (Fig 3B)Lastly a-cell mass calculated as the adjusted percentage ofglucagon area multiplied by the total weight of the pancreaswas similar for Ab+ and control groups (1215 6 426 mgcontrol donors vs 1412 6 657 mg Ab+) (Fig 3C) b-Cellmass did not correlate with age (r = 20018 P = 09341)BMI (r = 0377 P = 00833) or time in the intensive careunit (r = 20056 P = 08250) (data not shown)

The PI-to-INS Area Ratio Is Increased in Ab+ Donors andConstitutes a Potential Indicator of b-Cell DysfunctionTo study the direct relation between insulin and proinsulinarea in the pancreas the PI-to-INS area ratio was calculatedfor each section region and donor Interestingly the ratio wasincreased for Ab+ donors compared with control donors forthe head (38 P = 00031) body (40 P = 00005) and tail(32 P = 00004) regions of the pancreas (Fig 4A) To eval-uate the use of the area ratio as a potential indicator of b-celldysfunction and to directly compare Ab+ donors with controldonors an arbitrary reference value of 11 (proinsulin areandashtondashinsulin area) was chosen and graphically represented as a line

to separate control donors from ldquoat-riskrdquo donors The distancefrom the donorrsquos area values to the line was used as a score toestimate the risk of disease and to classify and distinguishcontrol donors from Ab+ donors (Fig 4B) Then a logisticregression was performed and the area under the curve(AUC) was calculated (Fig 4C) High AUC values significantcoefficients and a good model fit were obtained for the head(AUC 085 coefficient P = 00348 model P = 000225) body(AUC 087 coefficient P = 00213 model P = 000073) andtail (AUC 09 coefficient P = 00196 model P = 000045)indicating that the pancreatic PI-to-INS area ratio couldidentify individuals at risk for developing disease

To see whether the insulin area the proinsulin area orthe PI-to-INS area ratio could correlate with the risk ofdeveloping type 1 diabetes a risk index was calculated for all ofthe donors based on their age HLA and autoantibody status(Supplementary Fig 3) To calculate the risk index a score of0 (low risk) 1 (medium risk) or 2 (high risk) was assigned asfollows based on the risk of developing type 1 diabetes

Age 40 (0) 30ndash40 (1) 0ndash30 (2) HLA no risk alleles (0) DR4 or DR3 only (1) DR4 DQ8

or DR3 DQ2 or DQ8 (2) Autoantibodies 0 (0) 1 (1) 2 (2)

Figure 1mdashProinsulin area but not insulin area is significantly increased in the pancreas of Ab+ donors compared with control donorswithout diabetes Insulin (left panel) proinsulin (middle panel) and glucagon (right panel) areas expressed as percentage of positive areawere measured in whole-tissue sections from the head (A) body (B) and tail (C) region of the pancreas obtained from control donorswithout diabetes (n = 9) and from single Ab+ (n = 8) and double Ab+ (n = 5) cadaveric organ donors without diabetes P 005 P 001


The risk index was then calculated as the sum ofthe values obtained in each category for each donor(range 0 [minimum] to 6 [maximum]) A strong positive

correlation was found between the risk index and thePI-to-INS area ratio (Supplementary Fig 3 right panel)but there was a weak correlation with proinsulin area

Figure 2mdashProinsulin accumulates in the cytoplasmic compartment in b-cells from Ab+ donors Pancreatic sections from control (upperrows) single Ab+ (middle rows) and double Ab+ (lower rows) cadaveric organ donors without diabetes were stained for insulin (green)proinsulin (red) glucagon (white) and DAPI (blue) after a standard immunofluorescence staining protocol The merged image can be seenon the right panel Images were taken using a Zeiss LSM 880 confocal microscope with Airyscan and a 363 objective Scale bar 10 mm


(central panel) and no correlation with insulin area (leftpanel)

Patients With Recent-Onset Type 1 Diabetes HaveMore Glucagon Less Insulin and Less Proinsulin Areabut Increased PI-to-INS Area RatioTo investigate the timing of the increase in proinsulinarea and the inversion of the PI-to-INS area ratio whencompared with control donors pancreas tissue from a setof recently diagnosed patients with type 1 diabetes (0ndash9 weeks postdiagnosis) was studied (Fig 5) Only the tailregion was analyzed because all samples but one wereobtained by tail resection as previously described (30)(One section from a donor with type 1 diabetes at onsetobtained from nPOD was also included for comparison)Patients with type 1 diabetes were compared with controldonors and single Ab+ and double Ab+ donors separately(Fig 5A) The insulin and proinsulin areas were lowerthan in the rest of the groups owing to a reduction inthe islets that contained insulin whereas glucagon areawas increased (Fig 5C D and F) Interestingly thePI-to-INS area ratio was increased in patients with type1 diabetes and almost identical to that of at-risk doubleAb+ donors (077 6 013 control donors vs 099 6 015single Ab+ vs 106 6 010 double Ab+ vs 107 6 018type 1 diabetes) (Fig 5E left panel)

Systematic Analysis of Islet Size RevealsHeterogeneity Within the Pancreas and SubtleDifferences Between Ab+ Donors and Control DonorsSmall morphological alterations as well as in differencesin islet number size and distribution might occur manyyears before diagnosis in individuals at risk for developingdisease First the number of islets was counted and thetotal area of the tissue section was measured (see RESEARCH

DESIGN AND METHODS for details and Supplementary Fig 1)Total islet number and tissue size were variable across thepancreas but were similar in control donors and Ab+ do-nors (data not shown) There were no major differences inmean islet density between donors without diabetes (head25 6 05 vs body 186 02 vs tail 286 07 isletsmm2)and Ab+ donors (head 266 07 vs body 216 07 vs tail31 6 09 isletsmm2) (Fig 6A) Next the islet size dis-tribution was analyzed Significant differences were foundbetween control donors without diabetes and double Ab+donors for the head body and tail (Fig 6B and Supple-mentary Fig 4) Lastly sections from patients with type1 diabetes were analyzed Larger islets were found inthese sections (Fig 5A and Supplementary Fig 4 tail re-gion only) but the islet density was lower (Fig 5B) asexpected Interestingly the tail of the pancreas presenteda distinct islet distribution compared with the head and

Figure 3mdashb-Cell mass is not reduced in single or double Ab+ donors b-Cell mass was calculated as total pancreas weight multiplied byinsulin area (A) and as total pancreas weight multiplied by proinsulin area (B) and a-cell mass was calculated as total pancreas weightmultiplied by glucagon area (C) in control donors without diabetes (n = 9) and in single Ab+ (n = 8) and double Ab+ (n = 5) donors withoutdiabetes P 005


body regions for all groups with a predominance of largeislets and fewer small islets (Fig 6C and SupplementaryFig 4) confirming important regional differences withinthe pancreas

Changes in the Number of a- and b-Cells Occur Duringthe Prediabetic Phase and After Onset of DiseaseThe number of insulin- and glucagon-expressing cellsper islet was counted and the ratio between both cellpopulations was calculated (Supplementary Fig 5A)Although no significant differences were found doubleAb+ donors presented higher b-cellndashtondasha-cell ratios in thehead body and tail regions (head median = 38 control vs

53 single Ab+ vs 64 double Ab+body median = 40 controlvs 38 single Ab+ vs 71 double Ab+tail median = 36control vs 38 single Ab+ vs 49 double Ab+) Very hetero-geneous cell distribution patterns were found among theAb+ donors with some individuals having a higher abun-dance of a-cells per islet (Supplementary Fig 5B centralpanel) and others with a clear predominance of b-cells (Sup-plementary Fig 5B right panel) compared with controlsections (Supplementary Fig 5B left panel)

Patients with type 1 diabetes had a significantly lowerb-cellndashtondasha-cell ratio (median = 11 type 1 diabetes tailregion only) (Fig 5E right panel) Lastly the percentageof islets containing only b-cells was calculated which

Figure 4mdashThe PI-to-INS area ratio is increased in Ab+ donors and constitutes a potential indicator of b-cell dysfunction A The PI-to-INSarea ratio was calculated for head (left panel) body (middle panel) and tail (right panel) regions of the pancreas from control donors (n = 9)and single Ab+ (n = 8) and double Ab+ (n = 5) cadaveric organ donors without diabetes B Proinsulin area vs insulin area xy plot atheoretical reference value of 11 (proinsulin-to-insulin) was chosen and graphically represented as a line capable of separating control fromldquoat-riskrdquo donors The area under this line represents a ratio smaller than 1 and vice versa The distance from the donorrsquos area values to theline was used as a score to estimate the risk of developing disease and to classify and distinguish control donors from Ab+ donors (singleand double combined) C Receiver operating characteristic curve for the head (left panel) body (middle panel) and tail (right panel) regionsof the pancreas The AUC was calculated for the classifier described in panel B The P values show the significance of the logisticregression model including the predictor when compared with a model with just the intercept P 001 P 0001


again was similar for the control (249 6 90) single(288 6 99) and double Ab+ group (313 6 168) andlower for the group with type 1 diabetes (35 6 49)(Fig 5G left panel) Islets with only a-cells were notcommon (control 35 6 32 vs single Ab+ 28 6 08vs double Ab+ 52 6 77) Only one double Ab+ donor(6267 189) and the donors with type 1 diabetes(522 6 224) presented evident increases in the per-centage of islets containing only a-cells compared withthe rest of the donor groups (Fig 5G right panel)

DISCUSSION

The recent access to human pancreata for research pur-poses and the subsequent histological studies have filledimportant gaps in our understanding of pancreatic pa-thology (1031ndash33) However many fundamental ques-tions remain to be answered In this study we aimed tofully characterize b- and a-cells investigating insulin pro-insulin and glucagon content and distribution across thehead body and tail regions of the human pancreas inhealthy individuals as well as in single Ab+ and multiple

Figure 5mdashHigher glucagon and lower insulin and proinsulin areas but increased PI-to-INS area ratio in patients with recent-onset type 1diabetes A Box plots represent islet size distribution for control donors without diabetes (HC n = 5109) single Ab+ (SingleAb n = 4692) anddouble Ab+ (DoubleAb n = 2859) donors and donors with type 1 diabetes (T1D n = 1748) in the tail region of the pancreas B Islet densitywas calculated as the total number of islets per section divided by the total area of the tissue for the tail region of the pancreas CRepresentative image from whole-tissue section of donor 6362 with type 1 diabetes at onset Insulin is shown in green glucagon in redand DAPI in blue Note the presence of insulin-deficient and insulin-containing islets scattered across the pancreas parenchyma Scale bar500 mm D Insulin (left panel) proinsulin (middle panel) and glucagon (right panel) areas expressed as a percentage of the positive area weremeasured in whole-tissue sections from the tail region of the pancreas E The PI-to-INS area ratio (left panel) and the b-cellndashtondasha-cell ratio(right panel) were calculated for the pancreas tail region F Representative image of an islet from a recent-onset donor (DiViD study) Insulin isshown in green proinsulin in red and glucagon in blue Scale bar 50 mm G The percentage of islets containing only b-cells (left panel) andonly a-cells (right panel) is shown All panels control donors without diabetes (n = 9) single Ab+ (n = 8) and double Ab+ (n = 5) donors withoutdiabetes and donors with type 1 diabetes (n = 7) P 005 P 001 P 0001 P 00001


Ab+ donors In addition we had access to a very uniquesubset of pancreatic tissues from living patients withrecent-onset type 1 diabetes who participated in the DiViDstudy in which a small piece of the tail of the pancreas wassurgically removed immediately in the months after diagnosis

(30) Our study is the first to show an increase in pancreaticproinsulin area in individuals with prediabetes and confirmsthe value of the PI-to-INS area ratio as an indicator of earlyb-cell dysfunction In addition we confirm previous find-ings that b-cell mass is not reduced in Ab+ individuals We

Figure 6mdashSystematic analysis of islet distribution reveals heterogeneity within the pancreas and subtle differences between Ab+ donorsand control donors A Islet density was calculated as the total number of islets per section divided by the total area of the tissue for thehead body and tail regions of the pancreas of control donors (n = 9) single Ab+ (SingleAb) donors (n = 8) and double Ab+ (DoubleAb)donors (n = 5) B Box plots represent islet size distribution for healthy control (HC) donors single Ab+ donors and double Ab+ donors inthe head (left panel) body (middle panel) and tail (right panel) region of the pancreas C Box plots represent islet size distribution for headbody and tail regions in HC donors (left panel) single Ab+ (middle panel) donors and double Ab+ (right panel) donors Number of isletshead (HC n = 4390 SingleAb n = 4021 DoubleAb n = 2067) body (HC n = 3446 SingleAb n = 4052 DoubleAb n = 2043) and tail(HC n = 5109 SingleAb n = 4692 DoubleAb n = 2859) P 005 P 001 P 0001 P 00001


therefore add critical refinement to the original model ofb-cell loss described by Eisenbarth (4) three decades agoWe did not find any significant differences in the overallinsulin-positive area between healthy control donors andAb+ donors to support a reduction in insulin content longbefore the onset of disease Moreover b-cell mass was essen-tially identical in both groups demonstrating that b-cells arepreserved in Ab+ individuals until shortly before diagnosisand in agreement with previous publications that reportedno differences in b-cell mass between donors without diabe-tes and Ab+ donors (34ndash36)

Increases in the proinsulin area were found in some ofthe single Ab+ pancreata which could explain why thefirst-phase insulin response is already abnormal in someof these individuals (4) Proinsulin primarily accumulatesin the Golgi apparatus in resting b-cells and is furtherprocessed to insulin and C-peptide in immature secretorygranules (37) Our data show significant increases in theproinsulin area in Ab+ individuals in the head body andtail regions of the pancreas Using a high-resolutionconfocal microscope to study a subset of samples weobserved that the increase in the proinsulin area waspartially caused by a shift in proinsulin subcellularlocalization from the Golgi area in control individuals(juxtanuclear) to secretory vesicles in multiple Ab+ indi-viduals (cytosolic) Although an increase in proinsulin areasuggests an increase in the amount of proinsulin proteinadditional experiments are needed to accurately measurethe protein content Nevertheless this increase in areacould be caused by an increase in proinsulin production(to cover insulin demand) or by a defect in the processingand maturation of the existing pool of proinsulin to in-sulin which is consequently released to the circulationThis is in agreement with studies that have detectedproinsulinemia in patients with prediabetes and in pa-tients with type 1 diabetes (38) In addition an increasein proinsulin synthesis could lead to ER stress proteinmisfolding and loss of glucose-stimulated insulin secretion(39) Other extrinsic factors for example recurrent auto-immune attacks or other forms of metabolic stress (4041)may affect b-cell function and proinsulin processing earlyin the disease process of prediabetes In normal b-cells upto 20 of proinsulin can be misfolded However only un-der pathological conditions and ER dysfunction and once acertain level of misfolded proinsulin has been accumulateddo b-cell toxicity and death occur (22)

In a case-control study analysis by Sims et al (42) anelevation of the PI-to-C ratio preceded disease onset inhigh-risk subjects and could be detected at least 12 monthsbefore diagnosis This is in agreement with the data pre-sented here in which the pancreatic PI-to-INS area ratiowas increased in Ab+ individuals Among the single Ab+group two donors (6170 and 6184) consistently pre-sented an elevated PI-to-INS area ratio This suggeststhat b-cells in these two individuals could have had a func-tional defect Conversely all but one double Ab+ donor(6080) presented an increased ratio in at least two of

the three regions of the pancreas supporting the notionof an early functional defect in b-cells during the predia-betic phase and before diagnosis that does not necessarilyimply b-cell loss Donor 6080 was positive for GAD andinsulin Abs but was 69 years old and therefore unlikely tohave developed the disease In conclusion our in situ re-sults correlate well with those found in serum and confirmthe potential of monitoring PI-to-C or PI-to-INS area ratiosas indicators of b-cell dysfunction

The samples obtained from living individuals with recent-onset type 1 diabetes through the DiViD study (30) showedan expected significant decrease in the insulin and proinsulinarea resulting from a reduced number of insulin-containingislets however PI-to-INS area ratios were elevated in four ofsix patients suggesting that insulin therapy at onset might notfully alleviate the dysfunctional b-cells Our findings indicatethat potential therapies should target b-cells early before on-set when they still have the ability to be functionally rescuedand when the immune system has not been fully activated

We further characterized islet distribution and composi-tion to study whether small differences in islet pathologycould be observed early in the prediabetic phase The tail ofthe pancreas contained larger islets and higher islet density inall donor groups which could point to important develop-mental and architectural differences in vascularization andinnervation of the islets in this region (4344) Interestinglysubtle differences were found in double Ab+ at-risk individu-als in whom larger islets were found in the body and tailregion of the pancreas compared with control individualsMoreover this tendency was accentuated in individualswith recently diagnosed type 1 diabetes where even largerislets could be found in the tail region many of themcontaining only a-cells This is in agreement with previousstudies in diabetic NOD mice in which small islets werepreferentially lost and a subsequent expansion of large isletswas seen (45) An increase in the size of the islet and numberof b-cells in multiple Ab+ donors could be a compensatorymechanism caused by the chronic increase in insulin demand

Our observations in the pancreas of individuals at riskfor developing type 1 diabetes point to important and veryearly (at the first sign of autoimmunity) pathological changesin b-cells without an evident loss of b-cell mass This con-firms the potential benefit of estimating the PI-to-C or PI-to-INS area ratios in Ab+ individuals to identify those patientsat high risk at a stage where b-cells might still respond topreventive therapies and to enroll patients in clinical trialsat a point in the disease when they would benefit the mostWhether a consequence of an increase in insulin demand aprimary cellular defect or a change in the orchestrated in-terplay between the immune system and the islet the higheraccumulation of proinsulin in b-cells without a reduction ininsulin content might ultimately lead to b-cell exhaustion anddeath with the subsequent release of b-cell antigens thatinitiate the autoimmune process Whether type 1 diabetesis a primary autoimmune disease or autoimmunity is second-ary to metabolic or functional defects that render b-cellssusceptible to autoimmune destruction remains unknown


Future studies on proinsulin and insulin dynamics as well asa better characterization of b-cells themselves will providethe necessary answers to fully understand the pathologicalchanges that precede the clinical onset of type 1 diabetes

Acknowledgments This research was performed with the support ofnPOD a collaborative type 1 diabetes research project sponsored by JDRF Organprocurement organizations partnering with nPOD to provide research resourcesare listed at wwwjdrfnpodorgour-partnersphp The proinsulin antibody cloneGS-9A8 developed by OD Madsen (Hagedorn Research Institute GentofteDenmark) was obtained from the Developmental Studies Hybridoma Bankcreated by the Eunice Kennedy Shriver National Institute of Child Health andHuman Development of the National Institutes of Health and maintained at TheUniversity of Iowa Department of Biology Iowa City IA The authors thank thefollowing individuals from La Jolla Institute for Allergy and Immunology ZbigniewMikulsky and Bill Kiosses for their help with image acquisition and analysis andPriscilla Colby for administrative assistanceFunding This study was supported by National Institutes of HealthNationalInstitute of Allergy and Infectious Diseases grant R01-AI-092453-03Duality of Interest MGvH is an employee of Novo Nordisk No otherpotential conflicts of interest relevant to this article were reportedAuthor Contributions TR-C performed and designed experimentsanalyzed and interpreted data and wrote the manuscript JZ-G designed thecustom macros developed for computer-assisted software analysis analyzed andinterpreted data using bioinformatics tools and helped with statistical analysisNA performed experiments EC and YL helped with analysis LK andKD-J collected patient material and revised the manuscript KD-J is PrincipalInvestigator of the DiViD study MGvH designed experiments interpreted dataand wrote the manuscript MGvH is the guarantor of this work and as such hadfull access to all the data in the study and takes responsibility for the integrity of thedata and the accuracy of the data analysisPrior Presentation Preliminary results from this study were presented atthe Immunology of Diabetes Society 14th International Congress MunichGermany 12ndash16 April 2015 the JDRF nPOD 8th Annual Scientific MeetingMiami FL 22ndash25 February 2016 the 52nd European Association for the Studyof Diabetes (EASD) Annual Meeting Munich Germany 12ndash16 September 2016and the 19th Inflammation Research Association (IRA) International Meeting SanDiego CA 20ndash21 September 2016

References1 Pugliese A The multiple origins of type 1 diabetes Diabet Med 201330135ndash1462 van Belle TL Coppieters KT von Herrath MG Type 1 diabetes etiologyimmunology and therapeutic strategies Physiol Rev 20119179ndash1183 Srikanta S Ganda OP Gleason RE Jackson RA Soeldner JS EisenbarthGS Pre-type I diabetes Linear loss of beta cell response to intravenous glucoseDiabetes 198433717ndash7204 Eisenbarth GS Type I diabetes mellitus A chronic autoimmune diseaseN Engl J Med 19863141360ndash13685 Sosenko JM Skyler JS Herold KC Palmer JP Type 1 Diabetes TrialNet andDiabetes Prevention TrialndashType 1 Study Groups The metabolic progression totype 1 diabetes as indicated by serial oral glucose tolerance testing in the Di-abetes Prevention TrialndashType 1 Diabetes 2012611331ndash13376 Helminen O Aspholm S Pokka T et al HbA1c predicts time to diagnosis oftype 1 diabetes in children at risk Diabetes 2015641719ndash17277 Krogvold L Skog O Sundstroumlm G et al Function of isolated pancreaticislets from patients at onset of type 1 diabetes insulin secretion can be restoredafter some days in a nondiabetogenic environment in vitro results from the DiViDStudy Diabetes 2015642506ndash25128 Marreacute ML James EA Piganelli JD b cell ER stress and the implications forimmunogenicity in type 1 diabetes Front Cell Dev Biol 2015367

9 Eizirik DL Cnop M ER stress in pancreatic beta cells the thin red linebetween adaptation and failure Sci Signal 20103pe710 Campbell-Thompson M Fu A Kaddis JS et al Insulitis and b-cell mass inthe natural history of type 1 diabetes Diabetes 201665719ndash73111 Ziegler AG Rewers M Simell O et al Seroconversion to multiple isletautoantibodies and risk of progression to diabetes in children JAMA 20133092473ndash247912 Bonifacio E Predicting type 1 diabetes using biomarkers Diabetes Care201538989ndash99613 Narendran P Mannering SI Harrison LC Proinsulin-a pathogenic auto-antigen in type 1 diabetes Autoimmun Rev 20032204ndash21014 Thayer TC Pearson JA De Leenheer E et al Peripheral proinsulin ex-pression controls low-avidity proinsulin-reactive CD8 T cells in type 1 diabetesDiabetes 2016653429ndash343915 Pearson JA Thayer TC McLaren JE et al Proinsulin expression shapes theTCR repertoire but fails to control the development of low-avidity insulin-reactiveCD8+ T cells Diabetes 2016651679ndash168916 Toma A Laiumlka T Haddouk S et al Recognition of human proinsulin leadersequence by class I-restricted T-cells in HLA-A0201 transgenic mice and inhuman type 1 diabetes Diabetes 200958394ndash40217 Toma A Haddouk S Briand JP et al Recognition of a subregion of humanproinsulin by class I-restricted T cells in type 1 diabetic patients Proc Natl AcadSci U S A 200510210581ndash1058618 Mallone R Martinuzzi E Blancou P et al CD8+ T-cell responsesidentify beta-cell autoimmunity in human type 1 diabetes Diabetes 200756613ndash62119 Michels AW Landry LG McDaniel KA et al Islet-derived CD4 T-cells tar-geting proinsulin in human autoimmune diabetes Diabetes 201766722ndash73420 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 2016221482ndash148721 Liu M Wright J Guo H Xiong Y Arvan P Chapter Two ndash proinsulin entryand transit through the endoplasmic reticulum in pancreatic beta cells VitamHorm 20149535ndash6222 Sun J Cui J He Q Chen Z Arvan P Liu M Proinsulin misfolding andendoplasmic reticulum stress during the development and progression of di-abetes Mol Aspects Med 201542105ndash11823 Op de Beeck A Eizirik DL Viral infections in type 1 diabetes mellitusndashwhythe b cells Nat Rev Endocrinol 201612263ndash27324 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 2015641682ndash1687

25 Cianciaruso C Phelps EA Pasquier M et al Primary human and ratbeta -cells release the intracellular autoantigens GAD65 IA-2 and proinsulinin exosomes together with cytokine-induced enhancers of immunity Diabetes201766460ndash473

26 Roslashder ME Knip M Hartling SG Karjalainen J Akerblom HK Binder C TheChildhood Diabetes in Finland Study Group Disproportionately elevated pro-insulin levels precede the onset of insulin-dependent diabetes mellitus insiblings with low first phase insulin responses J Clin Endocrinol Metab 1994791570ndash1575

27 Hartling SG Knip M Roder ME Dinesen B Akerblom HK Binder C StudyGroup on Childhood Diabetes in Finland Longitudinal study of fasting proinsulinin 148 siblings of patients with insulin-dependent diabetes mellitus Eur J En-docrinol 1997137490ndash494

28 Truyen I De Pauw P Joslashrgensen PN et al Belgian Diabetes RegistryProinsulin levels and the proinsulinc-peptide ratio complement autoantibodymeasurement for predicting type 1 diabetes Diabetologia 2005482322ndash2329

29 Watkins RA Evans-Molina C Terrell JK et al Proinsulin and heat shockprotein 90 as biomarkers of beta-cell stress in the early period after onset oftype 1 diabetes Transl Res 201616896ndash106e101


30 Krogvold L Edwin B Buanes T et al Pancreatic biopsy by minimal tailresection in live adult patients at the onset of type 1 diabetes experiences fromthe DiViD study Diabetologia 201457841ndash84331 Kaddis JS Pugliese A Atkinson MA A run on the biobank what have welearned about type 1 diabetes from the nPOD tissue repository Curr Opin En-docrinol Diabetes Obes 201522290ndash29532 Pugliese A Vendrame F Reijonen H Atkinson MA Campbell-Thompson MBurke GW New insight on human type 1 diabetes biology nPOD and nPOD-transplantation Curr Diab Rep 20141453033 Campbell-Thompson M Organ donor specimens what can they tell usabout type 1 diabetes Pediatr Diabetes 201516320ndash33034 Wagner R McNally JM Bonifacio E et al Lack of immunohistologicalchanges in the islets of nondiabetic autoimmune polyendocrine patients withbeta-selective GAD-specific islet cell antibodies Diabetes 199443851ndash85635 Inrsquot Veld P Lievens D De Grijse J et al Screening for insulitis in adultautoantibody-positive organ donors Diabetes 2007562400ndash240436 Diedisheim M Mallone R Boitard C Larger E b-cell mass in nondiabeticautoantibody-positive subjects an analysis based on the Network for PancreaticOrgan Donors Database J Clin Endocrinol Metab 20161011390ndash139737 Haataja L Snapp E Wright J et al Proinsulin intermolecular interactionsduring secretory trafficking in pancreatic b cells J Biol Chem 20132881896ndash1906

38 Hostens K Ling Z Van Schravendijk C Pipeleers D Prolonged exposure ofhuman beta-cells to high glucose increases their release of proinsulin duringacute stimulation with glucose or arginine J Clin Endocrinol Metab 1999841386ndash139039 Wang S Kaufman RJ The impact of the unfolded protein response onhuman disease J Cell Biol 2012197857ndash86740 Oresic M Simell S Sysi-Aho M et al Dysregulation of lipid and amino acidmetabolism precedes islet autoimmunity in children who later progress totype 1 diabetes J Exp Med 20082052975ndash298441 Sysi-Aho M Ermolov A Gopalacharyulu PV et al Metabolic regulation inprogression to autoimmune diabetes PLoS Comput Biol 20117e100225742 Sims EK Chaudhry Z Watkins R et al Elevations in the fasting serumproinsulinndashtondashC-peptide ratio precede the onset of type 1 diabetes Diabetes Care2016391519ndash152643 Amella C Cappello F Kahl P Fritsch H Lozanoff S Sergi C Spatial andtemporal dynamics of innervation during the development of fetal human pan-creas Neuroscience 20081541477ndash148744 Lammert E Cleaver O Melton D Induction of pancreatic differentiation bysignals from blood vessels Science 2001294564ndash56745 Alanentalo T Houmlrnblad A Mayans S et al Quantification and three-dimensional imaging of the insulitis-induced destruction of b-cells in murinetype 1 diabetes Diabetes 2010591756ndash1764


Studies from the Network for Pancreatic Organ Donorswith Diabetes (nPOD) have recently shown that b-cellmass is not diminished in Ab+ donors and that singleb-cells and islets containing insulin can be found in do-nors with long-standing type 1 diabetes (10) The timecourse from seroconversion to onset of clinical diabeteshas been further characterized in longitudinal studiesAfter autoantibody seroconversion 145 of single Ab+and 679 of multiple Ab+ patients progressed to type1 diabetes in a 10-year follow-up study in three geograph-ically different cohorts (11) Another study also revealedthat 11 of multiple Ab+ children would progress to clin-ical disease each year (12) However the exact triggersand progression to clinical onset are not fully understood

Proinsulin is an important autoantigen in type 1 diabetesin humans and mice (13) because it shapes the autoreactiveCD8 T-cell repertoire (1415) Importantly recent studieshave shown that several epitopes within its precursor (pre-proinsulin) and proinsulin itself are recognized by islet-infiltrating CD4 andor CD8 T cells isolated from patientswith type 1 diabetes (16ndash20) suggesting a potential rolefor this antigen in disease pathogenesis Preproinsulin isprocessed into proinsulin and signal peptide (21) Only amarginal fraction of proinsulin is secreted to the circulationbut it accounts for 30ndash50 of the protein production inb-cells and increases in response to higher insulin demandBecause of this high metabolic demand b-cells are prone toendoplasmic reticulum (ER) stress and proinsulin misfold-ing which could lead to b-cell failure (22) ER stress mayalso be induced by viral infection (23) which was recentlydetected in the islets of Langerhans at diagnosis (24)Interestingly Cianciaruso et al (25) demonstrated thatcytokine-induced ER stress enhances the exosomal releaseof proinsulin Several reports present evidence of high circu-lating proinsulin and proinsulin intermediates with or with-out accompanying hyperglycemia in patients at risk fordeveloping the disease and after diagnosis (2627) Proinsulinand proinsulinndashtondashC-peptide (PI-to-C) ratios in combinationwith autoantibody concentration have been suggested as po-tential biomarkers for type 1 diabetes capable of identifying

with high sensitivity individuals at risk for developing disease1 to 40 months before clinical onset (2829) However thelink between proinsulin levels in serum and their content inthe pancreas in b-cells themselves has not been investigatedin human specimens in situ

The objective of this study was to add refinement to theclassic model of linear b-cell loss and to fill an importantgap in our understanding of the prediabetic phase in humantype 1 diabetes We studied for the first time in the hu-man pancreas the distribution of proinsulin in single- anddouble-Ab+ individuals preceding the onset of disease as wellas in patients with recent-onset type 1 diabetes and corre-lated it with loss of insulin content b-cell mass and PIareandashtondashinsulin (PI-to-INS) area ratio These studies under-line the need for development of biomarkers as well as forpreventive therapies focusing on normalizing b-cell dys-function during the prediabetic stage

RESEARCH DESIGN AND METHODS

SubjectsHuman pancreas sections were collected from cadavericorgan donors through nPOD Six micrometer formalin-fixed paraffin-embedded sections from the head bodyand tail of the pancreas were obtained from controldonors without diabetes who were single Ab+ (n = 8)double Ab+ (n = 5) and Ab2 (n = 9) and from onedonor at onset of type 1 diabetes In addition 4-mmformalin-fixed paraffin-embedded sections from living do-nors with type 1 diabetes (n = 6) were obtained throughthe DiViD study by tail resection (30) Overall 22 sectionsfrom the head 22 from the body and 29 from the tail ofthe pancreas were analyzed (n = 73 sections) Table 1summarizes the demographic information for each groupDetailed donor information can be found in Supplemen-tary Table 1 All experimental procedures were approvedby the La Jolla Institute for Allergy and ImmunologyInstitutional Review Board approved protocol numberDI3-054-0216 For the DiViD study participants providedwritten informed consent and more details can be foundin Krogvold et al (30)

Table 1mdashDonor demographic information including age percentage of males and females ethnicity BMI disease duration andC-peptide levels

Control(n = 9)

Ab+(n = 13)

Type 1 diabetes(n = 7)

Total(N = 29)

Age mean 6 SD (years) 336 6 125 361 6 135 282 6 49 326 6 41

Sex n ()Female 4 (445) 8 (615) 3 (428) 15 (517)Male 5 (555) 5 (385) 4 (572) 14 (483)

Ethnicity n ()African American 0 (0) 2 (154) 0 (0) 2 (69)Caucasian 7 (777) 8 (615) 7 (100) 22 (759)Hispanic 2 (223) 3 (231) 0 (0) 5 (172)

BMI mean 6 SD (kgm2) 282 6 6 262 6 5 25 6 32 265 6 16

Disease duration mean 6 SD (weeks) 44 6 27

C-peptide mean 6 SD (ngmL) 67 6 72 61 6 62 mdash








2p



RESULTS





























DISCUSSION



































2p



RESULTS





























DISCUSSION






















































DISCUSSION














































DISCUSSION










































DISCUSSION




































DISCUSSION






























DISCUSSION



























































































Documents

Increase in Pancreatic Proinsulin and Preservation of β ... · Increase in Pancreatic Proinsulin and Preservation of b-Cell Mass in Autoantibody-Positive Donors Prior to Type 1 Diabetes