THE
JOURNAL •RESEARCH •www. asebj.o g
Na u ally p esen ed HLA class I– es ic ed epi opes
om he neu o ophic ac o S100-ba e a ge s o
he au oimmune esponse in ype 1 diabe es
C is ina Cal iño-Samped o,*
,1
I ia Gomez-Tou ino,*
,1,2
Osca J. Co de o,* Ped o A. Reche,
†
Ma a G´
omez-Pe osanz,
†
Jose Luis S´
anchez-T incado,
†
Miguel ´
Angel Rod ´
ıguez,
‡
Au elio M. Suei o,
§
Juan E. Viñuela,
{
and Rub´
en Va ela Cal iño*
,3
*Depa amen o de Bioqu´
ımica y Biolog´
ıa Molecula , Facul ad de Fa macia, and
‡
Depa amen o de Biolog´
ıa Funcional, Cen o de In es igaci´
on en
Biolox´
ıa (CIBUS), Uni e sidade de San iago de Compos ela, San iago de Compos ela, Spain;
†
Depa amen o de Inmunolog´
ıa, Facul ad de Medicina,
Uni e sidad Complu ense de Mad id, Mad id, Spain; and
§
Se icio de Endoc inolog´
ıayNu ici
´
on, Hospi al de Conxo, and
{
Se icio de Inmunolog´
ıa,
Hospi al Cl´
ınico Uni e si a io, Complejo Hospi ala io Uni e si a io de San iago de Compos ela (CHUS), San iago de Compos ela, Spain
ABSTRACT: Type 1 diabe es (T1D) esul s om he des uc ion o panc ea ic b-cells by he immune sys em, and CD8
+
T lymphocy es a e c i ical ac o s in his au oimmune esponse. Panc ea ic isle s a e su ounded by a mesh o ne ous cells,
he pe i-insula Schwann cells, which a e also a ge ed by au o eac i e T lymphocy es and exp ess speci ic an igens, such
as he neu o ophic ac o S100-b. P e ious wo k has shown inc eased p oli e a i e esponses o whole S100-bin bo h
human T1D pa ien s and he nonobese diabe ic (NOD) mouse model. We desc ibe o he i s ime na u ally p ocessed
and p esen ed epi opes (NPPEs) p esen ed by class I human leukocy e an igen–A*02:01 (A2.1) molecules de i ed om
S100-b. These NPPEs igge ed IFN-g esponses mo e equen ly in bo h newly diagnosed and long- e m T1D pa ien s
compa ed wi h heal hy dono s. Fu he mo e, he same NPPEs a e ecognized du ing he au oimmune esponse leading o
diabe es in A2.1- ansgenic NOD mice as ea ly as 4 wk o age. In e es ingly, when hese NPPEs a e used o p e en
diabe es in his animal model, an accele a ion o he disease is obse ed oge he wi h an exace ba ion in insuli is and an
inc ease in S100-b–speci ic cy o oxici y in accina ed animals. Whe he hese can be used in diabe es p e en ion needs o
be ca e ully e alua ed in animal models be o e use in u u e clinical assays.—Cal iño-Samped o, C., Gomez-Tou ino, I.,
Co de o,O.J.,Reche,P.A.,G´
omez-Pe osanz, M., S´
anchez-T incado, J. L., Rod ´
ıguez, M. ´
A.,Suei o,A.M.,Viñuela,J.E.,
Cal iño, R. V. Na u ally p esen ed HLA class I– es ic ed epi opes om he neu o ophic ac o S100-ba e a ge s o he
au oimmune esponse in ype 1 diabe es. FASEB J. 33, 000–000 (2019). www. asebj.o g
KEY WORDS: cy o oxic lymphocy es •au oan igen •S100bpep ide epi opes •immuno he apy •pe i-insula
Schwann cells
One o he hallma ks o ype 1 diabe es (T1D) in bo h hu-
man pa ien s and nonobese diabe ic (NOD) mice is he
des uc ion o he insulin-p oducing b-cells in he pan-
c ea ic isle s. The au oimmune o igin o his a ack comes
om se e al lines o e idence, including he de ec ion o
au oan ibodies yea s be o e clinical disease onse (1) and
he de ec ion o T lymphocy es speci ic o he e y same
an igens (2–5).
Du ing his p ocess, panc ea ic isle s a e in il a ed by
di e en immune cell ypes, including CD4
+
and CD8
+
T lymphocy es (4, 6, 7). Howe e , CD8
+
lymphocy es a e
c i ical o diabe es de elopmen . Fi s , nei he insuli is
ABBREVIATIONS: A2.1, A*02:01; BB7.2, an i-human HLA-A*02:01–FITC mAb; CTL, cy o oxic T lymphocy e; ELISPOT, enzyme-linked immunospo ; GAD
65
,
glu amic acid deca boxylase 65; GFAP, glial ib illa y acidic p o ein; HD, heal hy dono ; HLA, human leukocy e an igen; HLA-DR4, HLA-DRB1*04:01; IA-2,
insulinoma-associa ed p o ein 2; IFA, incomple e F eund’s adju an ; K562/A2.1, HLA-A*02:01–exp essing K562 cell line; LS, long s anding; MHC, majo
his ocompa ibili y complex; MS, mass spec ome y; ND, newly diagnosed; NOD, nonobese diabe ic; NPPE, na u ally p ocessed and p esen ed epi ope;
PBMC, pe iphe al blood mononuclea cell; pSC, pe i-insula Schwann; ROC, ecei e -ope a ing cha ac e is ic; SI, s imula ion index; TBS, T is-bu e ed saline
1
These au ho s con ibu ed equally o his wo k.
2
Cu en a ilia ion: BioFa ma Resea ch G oup, Depa men o Pha macology, Pha macy and Pha maceu ical Technology. CIMUS Resea ch Cen e .
Uni e si y o San iago de Compos ela, San iago de Compos ela, Spain.
3
Co espondence: Depa amen o de Bioqu´
ımica y Biolog´
ıa Molecula , Facul ad de Fa macia, Campus Vida s/n, Uni e sidade de San iago de Com-
pos ela, 15782 San iago de Compos ela, La Co uña, Spain. E-mail: [email p o ec ed]
This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion 4.0 In e na ional (CC BY 4.0) (h p://c ea i ecommons.
o g/licenses/by/4.0/) which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed.
doi: 10.1096/ j.201802270R
This a icle includes supplemen al da a. Please isi h p://www. asebj.o g o ob ain his in o ma ion.
0892-6638/19/0033-0001 © The Au ho (s) 1
no diabe es de elops in NOD mice lacking majo his o-
compa ibili y complex (MHC) class I molecules and CD8
+
T cells (8). Second, in human T1D, he e is a linkage
o suscep ibili y o pa icula human leukocy e an igen
(HLA) class I molecules, such as A*2401 and A*0201, and
also o p o ec ion, such asA*01 and A*1101 (9, 10). Finally,
in human T1D pa ien s, CD8
+
T lymphocy es a e he
p edominan cell subse seen ininsuli ic lesionslinked o a
hype exp ession o HLA class I molecules (2, 11, 12).
Se e al a ge so hisin lamma o y esponseha ebeen
iden i ied, among which p ep oinsulin, glu amic acid
deca boxylase 65 (GAD
65
), and insulinoma-associa ed
p o ein 2 (IA-2)–bha e been epea edly ci ed in he li e -
a u e (13). Recen esul s in bo h he NOD mice and human
T1D pa ien s ha e led o he iden i ica ion o o he im-
po an a ge s o hose au o eac i e T lymphocy es,
such as zinc anspo e 8 o isle -speci ic glucose-6-
phospha ase– ela ed p o ein (14, 15).
Howe e , du ing diabe es de elopmen , pe i-insula
Schwann (pSC) cells a e also a ge ed by au o eac i e
lymphocy es (16). pSC cells o ganize a ne wo k, o ming a
mesh su ounding he isle s (16, 17). In NOD mice, lym-
phocy ic in il a ion leads o he b eakdown o he pSC
ne wo k and he de elopmen o insuli is (17, 18). pSC cells
exp ess speci ic an igens, such as glial ib illa y acidic p o-
ein (GFAP) and he neu o h ophic ac o S100-b.Those
an igens ha e been shown o be a ge ed by au oan ibodies
(19, 20) and T lymphocy e esponses (16, 21–25). Lym-
phocy e p oli e a ion agains GFAP and S100-ban igens
has been demons a ed in NOD mice and human T1D
pa ien s (16, 22). In he case o GFAP, I-A
g7
and K
d
pep ide
epi opes ha e been desc ibed in NOD mice (21), and
HLA-A2.1– es ic ed pep ide epi opes ha e been desc ibed
in humans (24). Recen ly, na u ally p ocessed and p e-
sen ed epi opes (NPPEs) de i ed om S100-b, es ic ed by
he HLA-DRB1*04:01 (HLA-DR4) class II molecule, and
a ge ed by he au oimmune esponse in T1D pa ien s ha e
been iden i ied (23). In NOD mice, pep ide epi opes de i ed
om S100-band ecognized by CD4
+
T lymphocy es ha e
also been desc ibed (25); in e es ingly, dominan S100-b–
de i ed epi opes in NOD mice lie wi hin he same egions
as hose a ge ed by human T esponses.
Immuno he apies employing ei he whole GFAP o
S100-bp o ec NOD mice om diabe es de elopmen (16).
Mo eo e , immuno he apy wi h class II I-A
g7
–and class I
K
d
– es ic ed pep ide epi opes de i ed om GFAP p o ec
NOD mice om diabe es de elopmen (21). These da a
indica e ha immuno he apy wi h pSC-de i ed an igens
could p e en he de elopmen o clinical symp oms.
Despi e hese esul s, no S100-bpep ide epi opes a -
ge ed by CD8
+
cy o oxic T lymphocy es (CTLs) in T1D
pa ien s ha e been iden i ied, no has hei po en ial e -
ec i eness in p e en ing diabe es de elopmen been e-
po ed. In he p esen s udy, we ha e iden i ied se e al
NPPEs de i ed om S100-band p esen ed by he class I
HLA molecule A*02:01 (A2.1). We hypo hesized ha e-
sponses agains hese pep ide epi opes migh be p esen in
bo h human T1D pa ien s and A2.1- ansgenic NOD mice.
Ou esul s suppo his hypo hesis: highe esponses
agains some S100-b–de i ed NPPEs can be de ec ed mo e
equen ly in bo h newly diagnosed (ND) and long- e m
T1D pa ien s. Fu he mo e, T1D pa ien s espond mo e
equen ly and wi h highe in ensi y o mo e han 1 epi-
ope. Simila esponses can be de ec ed in A2.1- ansgenic
NOD mice as ea ly as 4–5 wk o age. When hese pep ide
epi opes we e used o immuno he apy, an appa en ac-
cele a ion o disease de elopmen was obse ed oge he
wi h a wo sening in he insuli ic lesions and inc eased cy-
o oxici y agains S100-b–posi i e a ge s by CTLs. Ou
esul s highligh he need o ca e ul e alua ion o di e en
adminis a ion ou es and dosage p o ocols using his an-
igen o amelio a e he au oimmune esponse in T1D.
MATERIALS AND METHODS
Cloning o S100-band gene a ion o an S100-
b–exp essing su oga e an igen-p esen ing cell
S100-bcDNA (GenBank accession no. BC001766.1) was ob ained
om Sou ceBioscience (No ingham, Uni ed Kingdom) and
subcloned in o he BamHI si e o he pcDNA3.1/Zeo(+) ec o
(The mo Fishe Scien i ic, Wal ham, MA, USA).
A su oga e an igen-p esen ing cell was gene a ed by ans-
ec ion o an HLA-A2.1–exp essing K562 cell line (K562/A2.1)
wi h S100-b/pcDNA3.1-Zeo using Lipo ec amine (The mo
Fishe Scien i ic) ollowing he manu ac u e ’sins uc ions.
Clones exp essing high le els o S100-bwe e selec ed using
Gene icin and Zeocin (The mo Fishe Scien i ic).
Exp ession o bo h A2.1 and S100-bwas e i ied by low
cy ome y and immuno luo escence. Fo low cy ome y, cells
we e su ace s ained wi h an an i-human HLA-A2.1–FITC mAb
(BB7.2 clone) (BD Biosciences, San Jose, CA). A e washing, cells
we e ixed and pe meabilized (Cy o ix/Cy ope m; BD Biosci-
ences) and s ained in acellula ly wi h a mouse an i–S100-bmAb
(Abcam, Camb idge, MA, USA) ollowed by an an i-mouse IgG
phycoe y h in-labeled goa pAb (Abcam) and analyzed using a
BD FACScalibu low cy ome e (BD Biosciences).
Fo immuno luo escence analysis, cells we e ixed on o poly-L-
Lysine–coa ed co e slips; s ained wi h ei he he BB7.2-FITC
an ibody (su ace) o , a e ixa ion/pe meabiliza ion, he S100-b–
FITC an ibody (Abcam) (in acellula ) and DAPI; and imaged in an
Olympus BX51 luo escence mic oscope(Olympus,Tokyo,Japan).
Iden i ica ion o HLA-A2.1– es ic ed NPPEs de i ed
om S100-b
Pep ides bound o A2.1 in he cell su ace o ei he K562/A2.1 o
K562/A2.1–S100-bwe e elu ed by a b ie incuba ion in acid ci -
a e bu e (pH 3.3) and sequen ially en iched using a 3 kDa cu o
Amicon Ul a Fil e (Millipo eSigma, Bu ling on, MA, USA) and
a Disco e y DSC-18 i unc ional C18 silica esin column (Milli-
po eSigma). Pep ide ac iona ion was done using a 150 mm 3
2.1 mm BioBasic 18 column (The mo Fishe Scien i ic), and
pep ide-con aining ac ions we e s o ed a 280°C un il analysis
by mass spec ome y (MS).
MS was ca ied ou in he Spec ome y Se ice [Ins i u o de
In es igaciones Sani a ias (IDIS), San iago de Compos ela] on a
ma ix-assis ed lase deso p ion/ionisa ion– ime o ligh mass
spec ome y (MALDI-TOF) MS Analyze (The mo Fishe Sci-
en i ic). The analysis was ca ied ou wi h he 4000 Se ies Ex-
plo e so wa e .3.5 (The mo Fishe Scien i ic) and Masco .2.1
(Ma ix Science, Bos on, MA, USA) o sea ch agains a Na ional
Cen e o Bio echnology In o ma ion non edundan (NCBIn )
p o ein da abase o in an S100-b–speci ic da abase. Unique m/z
alues we e iden i ied using Findpep (h p://web.expasy.o g/
indpep /) o selec hose ha could be de i ed om S100-b.Po-
en ial unique pep ide epi opes o 8–10 aa long we e chosen.
2 Vol. 33 May 2019 The FASEB Jou nal x www. asebj.o g CALVIÑO-SAMPEDRO ET AL.
In i o p o easome diges ion and in silico
p o easomal clea age analysis o pu i ied
human S100-b
Pu i ied human S100-b(23) was incuba ed wi h pu i ied 20S
p o easome (Enzo Li e Sciences, Fa mingdale, NY, USA) (mola
a io 250:1) in diges ion bu e [30 mM T is-HCl (pH 8.0), 10 mM
NaCl, 2 mM MgCl
2
, 1 mM DTT, 0.01% sodium dodecyl sul a e)
o 16 h a 37°C. As a con ol, he same diges ion was se up
wi h ei he no p o easome o wi h ace ic acid (1%)–inac i a ed
p o easome. Gene a ed pep ides we e pu i ied and concen a ed
using a ca ionic esin (ZipTip wi h s ong ca ion exchange; Mil-
lipo eSigma) ollowing he manu ac u e ’sins uc ions.Re ained
pep ides we e elu ed om he esin and analyzed by MS.
The S100-bhuman p o ein sequence was analyzed in silico o
po en ial p o easome and immunop o easome clea age si es
using se e al published algo i hms (26, 27).
Pep ides
S100-b–de i ed NPPEs S100
10–18
(ALIDVFHQY) and S100
20–28
(GREGDKHKL) we e syn hesized by ChinaPep ide (Hangzhou,
China) o .90% pu i y. A s ock o 100 mg/ml in DMSO was
p epa ed o each pep ide and s o ed a 220°C un il use.
HLA s abiliza ion assay
Binding o he S100-bNPPE candida es was examined by a
con en ional HLA s abiliza ion assay as p e iously desc ibed
(28). B ie ly, T2 cells we e washed wice wi h se um- ee AIM V
medium (The mo Fishe Scien i ic) and incuba ed in medium
con aining b2-mic oglobulin (Millipo eSigma) and each o he
es pep ides a a ious concen a ions (100–0.78 mM). Su ace
A2.1 exp ession was de e mined using he BB7.2 an ibody in a
FACSCalibu low cy ome e .
Human dono s
In o al, 35 subjec s we e s udied: 18 nondiabe ic heal hy dono s
(HDs) wi hou amily his o y o disease and 7 ND (,1 y ) and 10
long-s anding (LS) (.1 y ) T1D pa ien s. Blood was d awn wi h
he in o med consen o all subjec s and app op ia e pe mission
was ob ained om he Ins i u ional E hics Commi ee (Comi ´
e
´
E ico de In es igaci´
on Cl´
ınica de Galicia, CEIC). T1D pa ien s
we e en olled om hose a ending he diabe ic clinic a he En-
doc inology Se ice (Hospi al de Conxo, San iago de Compos-
ela, Spain). Heal hy nondiabe ic dono s we e ec ui ed om
labo a o y and hospi al s a du ing he same pe iod. HLA-DR4
and HLA-A2.1 posi i i y was de e mined by PCR as p e iously
desc ibed by Bunce (29). GAD
65
and IA-2 au oan ibodies we e
de e mined by ELISA (Palex Medical, Ba celona, Spain).
Mice
T ansgenic NOD.B6-Tg(HLA-A2.1)1Enge/D sJ mice we e pu -
chased om The Jackson Labo a o y (Ba Ha bo , ME, USA) and
main ained in he speci ic-pa hogen– ee acili ies a he Molec-
ula Medicine and Ch onic Diseases Resea ch Cen e (San iago
de Compos ela). Glucosu ia was de e mined weekly om 8 wk
o age onwa d using Medi-Tes Glucose S ips (Mache ey-Nagel,
D¨
u en, Ge many). Diabe es was diagnosed when glucose le els
we e $500 mg/dl (27.8 mM) in 2 consecu i e measu es. All ex-
pe imen al p ocedu es we e app o ed by he co esponding
e hics commi ee. Splenocy es we e s ained wi h he BB7.2 an i-
body o con i m A2.1 exp ession.
To ob ain single-cell suspensions, spleens we e passed
h ough a 70-mm po e size il e (Co ning cell s aine ; Co ning,
Co ning, NY, USA) and cen i uged o e a Ficoll g adien .
S100-bNPPE immuno he apy
An immuno he apy p o ocol was pe o med as p e iously desc ibed
by Han e al. (30). B ie ly, emale mice we e immunized in ape i o-
neally wi h a 1:1 mix u e o S100
10–18
and S100
20–28
in saline. Im-
muniza ions s a ed a 4 wk o age and we e pe o med e e y 2 wk
un il 8 wk o age o 3 wk he ea e . Mice we e ollowed un il hey
became diabe ic o eached 30 wk o age. High (100 mg/pep ide) and
low (10 mg/pep ide) pep ide doses we e e alua ed. As a con ol,
emale mice we e inocula ed wi h saline con aining DMSO (0.2%).
Insuli is and immunohis ochemis y
Bouin- ixedpanc easeswe eincludedinpa a in,and3pa allel
se ies wi h 16 mm hick sec ions we e ob ained. Sec ions we e
s ained wi h hema oxylin-eosin, and be ween 30 and 50 isle s pe
mouse we e sco ed in a blinded ashion. Quan i ica ion o insu-
li is wi h age was done using 5–6 animals pe age g oup. Insuli is
sco e in ea ed animals was de e mined a e diabe es diagnosis
o a 30 wk o age. Each isle was sco ed acco ding o he scale
indica ed in he supplemen al da a.
Fo immunohis ochemis y, panc eases we e ixed in 4% pa a-
o maldehyde. Sec ions we e p e ea ed wi h P o einase K (10 mg/
ml) o 5 min a oom empe a u e. Sec ions we e insed in 0.05 M
T is-bu e ed saline (TBS) (pH 7.4) o 5 min and ea ed wi h 10%
H
2
O
2
in TBS (30 min, oom empe a u e). A e insing in 0.05 M TBS
(pH 7.4) o 5 min, sec ions we e incuba ed o e nigh a oom
empe a u ewi hanan i-GFAPmousemAb(clone2A5;Abcam).
Sec ions we e insed 3 imes in TBS (10 min each) and incuba ed wi h
an an i-mouse ho se adish pe oxidase–coupled seconda y an ibody
(1 h, oom empe a u e). Sec ions we e de eloped wi h 0.25 mg/ml
diaminobenzidine e ahyd ochlo ide (Millipo eSigma) in TBS (pH
7.4) and 0.00075% H
2
O
2
and coun e s ained wi h hema oxylin-
eosin. All dilu ions we e made wi h TBS con aining 15% no mal goa
se um (Millipo eSigma) and 0.2% T i on X-100 (Millipo eSigma). All
incuba ions we e ca ied ou in a humidchambe .Finally, hesec-
ions we e dehyd a ed and moun ed.
Enzyme-linked immunospo assays
IFN-genzyme-linked immunospo (ELISPOT) assays (U-CyTech
Biosciences, U ech , The Ne he lands) we e used o de ec S100-
b–speci ic esponses acco ding o he manu ac u e ’s ins uc-
ions. Mouse splenocy es we e used in all assays, and an indi ec
ELISPOT assay was pe o med wi h a 6-h p es imula ion s ep.
When human esponses we e analyzed, a di ec ELISPOT assay
was pe o med (3 310
5
pe iphe al blood mononuclea cells
(PBMCs)/well in iplica es pe an igen) wi h a 24-h s imula ion
pe iod wi h he co esponding pep ide/an igen.
Pep ides we e added o a inal concen a ion o 10 mM. Bo h
cul u e medium alone and DMSO (0.5%) we e used as nega i e
con ols. CEF Pep ide Pool (Mab ech AB, Nacka, Sweden) o
concana alin A (25 ng/ml) we e used as posi i e con ols in
human and mouse ELISPOTs, espec i ely.
Da a a e exp essed as a s imula ion index (SI) (SI = o al spo s
wi h an igen/ o al spo s wi h DMSO). In human ELISPOTs, a
esponse was conside ed posi i e when he SI was abo e he
h eshold de e mined using ecei e -ope a ing cha ac e is ic
(ROC) cu es. In mouse ELISPOTs, a esponse was conside ed
posi i e when he SI was .2.
Cy o oxici y assays
Lysis o pep ide-loaded a ge s was done as p e iously indica ed
by Ne i e al. (31). B ie ly, splenocy es om young nondiabe ic
emale mice we e used as a ge s, whe eas splenocy es om ei-
he saline- (n= 5) o S100-b–(n=13) accina edmicewe eused
NOVEL S100-bEPITOPES IN TYPE 1 DIABETES 3
as e ec o s. Ta ge cells we e incuba ed 2 h wi h S100-bNPPEs
(bo h a 10 nM) o DMSO and s ained wi h calcein (15 mM)
(Millipo eSigma). Cells we e washed, and 1 310
4
cells we e
pla ed in U-bo om, 96-well pla es. E ec o cells we e added in
quad uplica es a di e en e ec o : a ge a ios (1:1 o 40:1).
Maximum lysis was ob ained by adding 2% T i on X-100 o
a ge cells, and spon aneous lysis was ob ained by adding cul-
u e medium o a ge cells. A e 4 h o incuba ion, luo escence
(530 nM) in cul u e medium was de e mined in a Syne gy H1M
pla e eade (BioTek Ins umen s, Winooski, VT, USA). The
pe cen age o cy o oxici y was calcula ed as (MF
sample
2
MF
spon aneous
)/(MF
maximum
2MF
spon aneous
)3100, whe e MF is
mean luo escence.
S a is ical analysis
Di e ences in he pe cen age o highly in il a ed isle s and he
equency o posi i e esponses among di e en g oups we e
analyzed using Fishe ’s exac es . Compa ison o median e-
sponses in ELISPOT assays be ween g oups was done using he
Mann-Whi ney U es o he K uskal-Wallis es o mul iple
compa isons. Su i al was analyzed using he log- ank Man el-
Cox es . A alue o P#0.05 was conside ed signi ican . All
s a is ical analysis was done using SPSS .24 (IBM SPSS, Chi-
cago, IL, USA) and P ism (G aphPad, La Jolla, CA, USA).
RESULTS
Gene a ion o su oga e an igen-p esen ing
cells o he iden i ica ion o A2.1- es ic ed
S100-b–de i ed NPPEs
Se e al clones we e ob ained om he o iginal K562-
A2.1/S100-bcell line, and 4 we e selec ed based on hei
exp ession o bo h A2.1 and S100-b(Supplemen al Fig.
S1A,B), con i ming he gene a ion o a su oga e an igen-
p esen ing cell.
No signi ican di e ences we e no ed be ween ch o-
ma og ams om K562/A2.1 and K562/A2.1–S100-b–
elu ed pep ides (Supplemen al Fig. S1C). In o al, 44
pep ide-con aining ac ions we e analyzed by MS, iden-
i ying 7298 and 7027 m/z masses om K562/A2.1 and
K562/A2.1–S100-bcells, espec i ely. Up o 3403 m/z
could be de i ed om S100-b,allowinga1Dae o ,wi h
jus 2035 m/z emaining when only pep ides be ween 8 and
11 aa long we e kep . These m/z masses we e compa ed
wi h he equi alen ac ion de i ed om K562/A2.1 cells
as well as he immedia ely p eceding and pos e io ac-
ions looking o unique masses. This analysis led o he
iden i ica ion o 4 unique candida e masses (Table 1).
Mos p edic ions o each expe imen al m/z co espond
ei he o o e lapping pep ides o pep ides lying wi hin he
same sho egion o S100-b(Table 1).
P o easome clea age and A2.1 binding
a ini y o candida e S100-bNPPEs
F om all he sequences shown in Table 1, po en ial
p o easome clea age si es a e p edic ed by se e al algo-
i hms o 5 pep ide sequences (10–18 ALIDVFHQY; 20–28
GREGDKHKL; 38–45 NNELSHFL; 54–61 VDKVMETL;
67–75 GECFQEFM). We nex de e mined he in i o
p o easome clea age pa e n o S100-bby diges ion wi h
pu i ied 20S p o easome. The use o ac i e p o easome
gene a es mul iple unique agmen s (Fig. 1A, B,bo om
panels), which we e no p esen when S100-bwas in-
cuba ed alone (Fig. 1A, B, op panels) o wi h inac i a ed
p o easome (Fig. 1A, B,middlepanels).Uniquem/z p esen
in samples o S100-bincuba ed wi h ac i e p o easome
we e analyzed by andem MS and Masco , and 9 o hem
we e iden i ied as being de i ed om S100-b(Fig. 1A, B,
bo om panels, a ows), wi h many o hese co esponding
o i s amino- e minal end (Fig. 1C). Compa ison o
p o easome-gene a ed pep ides wi h A2.1-elu ed pep ide
MS da a (Table 1) indica es ha he ca boxy- e minal
end o pep ides S100
10–18
(ALIDVFHQY) and S100
20–28
TABLE 1. NPPEs de i ed om S100-band p esen ed by he A2.1 molecule
Expe imen al m/z Theo e ical m/z ΔDa Sequence Posi ion
935.26 934.491 20.934 VDKVMTEL 54–61
935.359 20.067 ETLDNDGDG 59–67
935.359 20.067 TLDNDGDGE 60–68
936.398 0.971 VMETLDND 57–64
974.290 973.474 20.816 INNELSHF 37–44
973.474 20.816 NNELSHFL 38–45
973.520 20.769 EIKEQEVV 47–54
973.640 20.649 KLKKSELK 27–34
974.356 0.065 GECDFQEF 67–74
974.515 0.225 KEQEWDK 49–56
974.588 0.298 LKKSELKE 28–35
974.588 0.298 KKSELKEL 29–36
1039.565 1039.480 20.084 QYSGREGDK 17–25
1039.564 0.000 GREGDKHKL 20–28
1105.506 1105.396 20.109 GECDFQEFM 67–75
1105.568 0.061 ALIDVFHQY 10–18
ΔDa, di e ence be ween expe imen al and heo e ical m/z (in Da); Expe imen al m/z, unique m/z
p esen only in he pep ide mix de i ed om K562/A2.1–S100-bcells; Posi ion, amino acid posi ion in
he S100-bp o ein o he p edic ed pep ide; Sequence, amino acid sequence o he p edic ed pep ide;
Theo e ical m/z,m/z p edic ed by Findpep .
4 Vol. 33 May 2019 The FASEB Jou nal x www. asebj.o g CALVIÑO-SAMPEDRO ET AL.
(GREGDKHKL) is gene a ed by he p o easome, making
hese sequences po en ial HLA class I– es ic ed epi opes.
MS da a and p o easome clea age we e he 2 main c i e ia
o selec he NPPEs o be e alua ed in he ELISPOT assays.
O he 5 pep ides wi h he po en ial p o easome clea -
age si es indica ed abo e, only pep ide S100
10–18
(ALIDVFNQY) demons a ed some weak binding a ini y
o A2.1, whe eas o he o he 4 S100-bpep ides, no
binding a ini y could be demons a ed (Supplemen al
Fig. S2A). In e es ingly, se e al algo i hms (32–35) indi-
ca e ha S100
10–18
is one o he pep ides wi h he highes
binding a ini y o A2.1 om all o he po en ial 9-me
pep ides de i ed om S100-b(Supplemen al Fig. S2B). In
ac , when combining p edic ions o p ocessing by he
p o easome, anspo e associa ed wi h an igen p o-
cessing binding, and A2.1 binding a ini y (36), S100
10–18
is once again he pep ide showing he highes combined
p edic ion sco e, ollowed by S100
67–75
and S100
20–28
(Supplemen al Fig. S2B). All o hese da a combined
sugges ha pep ide epi opes S100
10–18
(ALIDVFHQY)
and S100
20–28
(GREGDKHKL) could cons i u e po en ial
candida es o be ecognized by au o eac i e CD8
+
T
lymphocy es in T1D pa ien s.
T1D pa ien s show speci ic T-cell esponses
agains A2.1- es ic ed S100-b–de i ed NPPEs
We in es iga ed whe he lymphocy e esponses agains
S100
10–18
and S100
20–28
could be de ec ed in T1D pa ien s
by IFN-gELISPOT (see Table 2 o demog aphic da a).
ROC analysis gi es an op imal SI cu o o 1.85 and
1.75 o S100
10–18
and S100
20–28
, espec i ely (Fig. 2A);
he e o e, an SI $2 was chosen o classi y a esponse as
posi i e.
Figu e 1. MS analysis o in i o p o easome diges ion o pu ified S100-b.A,B) Spec a co esponding o m/z ange o 917–1564
(A) and 1556–3145 (B) a e shown. Da a o S100-bincuba ed wi h no p o easome ( op panels), inac i e p o easome (middle
panels), o ac i e p o easome (bo om panels) a e shown. Unique m/z p esen only in samples o S100-bincuba ed wi h ac i e
p o easome we e subsequen ly iden ified by andem MS and analyzed by Masco (a ows). C) Sequences o he 9 mos in ense
agmen s gene a ed by in i o diges ion o S100-bwi h pu ified 20S p o easome a e aligned wi h he amino acid sequence o he
an igen. The m/z o each agmen is shown on he le . The ca boxy- e minal amino acid is unde lined. The sequences o
po en ial pep ide epi opes elu ed om A2.1 a e shown a he bo om (ca boxy- e minal amino acid is unde lined only i he
agmen has been gene a ed by he p o easome and iden ified by MS). The ca boxy- e minal ends o pep ides S100
10–18
(ALIDVFHQY) and S100
20–28
(GREGDKHKL) a e gene a ed by he p o easome, making hem po en ial class I– es ic ed
pep ide epi opes.
NOVEL S100-bEPITOPES IN TYPE 1 DIABETES 5
Clea esponses agains a leas 1 o he 2 S100-bpep-
ides could be de ec ed (Fig. 2B), and hese esponses we e
signi ican ly mo e equen among T1D pa ien s han HDs
o bo h S100-bpep ide epi opes [S100
10–18
:64.7%T1D
(11/17) s. 11.1% HD (2/18). P= 0.0016; S100
20–28
:70.6%
T1D (12/17) s. 16.7% HD (3/18). P= 0.002. Two- ailed
Fishe ’s exac es ]. Posi i e esponses o S100
10–18
could be
de ec ed in bo h HLA-A2.1–posi i e (n=6;Fig.2C, ed
squa es) and HLA-A2.1–nega i e (n=5;Fig.2C,black
squa es) T1D pa ien s. A simila esul is seen o S100
20–28
in HLA-A2.1–posi i e (n=7;Fig.2D, ed squa es) and
HLA-A2.1–nega i e (n=6;Fig.2D,blacksqua es)T1D
pa ien s. Conside ing only HLA-A2.1–posi i e subjec s,
esponses in A2.1+ T1D pa ien s a e s ill signi ican ly
highe compa edwi hHDA2.1+subjec s o S100
20–28
(P=
0.04; 2- ailed Fishe ’sexac es )bu no o S100
10–18
(P=
0.30; 2- ailed Fishe ’s exac es ).
The magni ude o he esponses (median SIs) is also
highe among T1D pa ien s [S100
10–18
: 4.8 (T1D) s. 1.0
(HD). P= 0.004. S100
20–28
: 4.1 (T1D) s. 0.9 (HD). P=
0.0006. Mann-Whi ney U es ] (Fig. 2C, D). Conside ing
only HLA-A2.1–posi i e subjec s, esponses in A2.1+ T1D
pa ien s a e s ill signi ican ly highe compa ed wi h HD
A2.1+ subjec s o S100
20–28
(P= 0.02; Mann-Whi ney U
es ) and almos signi ican o S100
10–18
(P= 0.09; Mann-
Whi ney U es ).
We nex decided o compa e he esponses be ween ND
and LS T1D pa ien s. In he case o S100
10–18
, he equency
o posi i e esponses among ND (4/7; 57.1%) and LS (8/
10; 80.8%) T1D pa ien s was highe compa ed wi h ha
seen in HDs (2/10; 20.0%), eaching s a is ical signi icance
only o LS T1D pa ien s (P= 0.002; 2- ailed Fishe ’sexac
es ) (Fig. 2E). In he case o S100
20–28
, equencyo posi i e
esponses among ND (6/7; 85.7%) and LS T1D pa ien s
(6/10;60.0%) wasagainhighe compa ed wi h ha seenin
HD (2/10; 20%), eaching s a is ical signi icance only o
ND T1D pa ien s (P= 0.015; 2- ailed Fishe ’sexac es )
(Fig. 2F).
Fu he mo e, he magni ude o he esponses (median
SIs) agains S100
10–18
we e signi ican ly highe in LS T1D
pa ien s (13.7) s. HD (1.0) (Fig. 2E)(P= 0.0024; K uskal-
Wallis, Dunn’spos hoc es ). In he case o S100
20–28
,me-
dian esponses in ND (4.1) and LS (3.6) T1D pa ien s we e
signi ican ly highe compa ed wi h HD (0.9) (Fig. 2F)(P=
0.01 and P= 0.001, espec i ely; K uskal-Wallis, Dunn’s
pos hoc es ).
In e es ingly, dual posi i i y (i.e., esponse o bo h
pep ides) is mo e equen ly de ec ed in T1D pa ien s (9/
17; 52.9%) compa ed wi h HDs (1/18; 5.5%) (P= 0.0027; 2-
ailed Fishe ’s exac es ) (Fig. 2G).
In summa y, all hese da a indica e ha p oin-
lamma o y lymphocy e esponses agains he 2 no el
S100-bNPPEsa emo e equen in T1D pa ien s, bo h ND
and LS.
Immune esponses agains S100-b–de i ed
NPPEs in he A2.1- ansgenic NOD mice
We examined whe he simila esponses agains he
S100
10–18
and S100
20–28
pep ide epi opes could be de ec ed
in he A2.1- ansgenic NOD mouse p eclinical model (37).
Diabe es de elopmen kine ics and insuli is in bo h males
and emales in ou colony a e shown in Supplemen al Fig.
S3.
Posi i e esponses by splenocy es agains S100
10–18
and
S100
20–28
can be de ec ed in male [8/22 (36.4%) and 8/22
(36.4%), espec i ely] (Fig. 3A) and emale mice [19/31
(61.3%) and 15/31 (48.4%), espec i ely] (Fig. 3A). The e
a e no s a is ical di e ences in he pe cen age o esponses
be ween male and emale mice (P= 0.09, P= 0.41; 2- ailed
Fishe ’s exac es ). Howe e , he magni ude o he e-
sponses (median SIs) is signi ican ly highe among emales
(2.6) when compa ed wi h males (1.5) o S100
10–18
(P=
0.035; Mann-Whi ney U es ) (Fig. 3A).
When hose esponses we e examined acco ding o age
in emale mice, esponses o bo h S100-b–de i ed NPPEs
can be de ec ed as ea ly as 4–5wko age(Fig.3B, C).
These da a con i m ha S100-bisa a ge o heau o-
immune esponse in T1D and ha epi opes S100
10–18
and
S100
20–28
a e a ge ed by T lymphocy es.
Immuno he apy o A2.1- ansgenic NOD mice
wi h S100-b–de i ed pep ides
We nex e alua ed he capaci y o hese NPPEs o p e en
disease de elopmen by pe iodic immuniza ion o young
emale mice. No signi ican educ ion in disease de elop-
men equency is seen compa ed wi h saline-inocula ed
mice (Fig. 4A)(100% s. 84.6%; P= 1.0; 2- ailed Fishe ’s
exac es ). In ac , S100-b accina ion seems o sligh ly
accele a e T1D de elopmen , al hough his kine ic does
TABLE 2. HD and T1D pa ien demog aphics
Va iable HD (n= 19) ND (n=7) LS(n= 11)
Age [mean 6SD (y )] 21 61.1 34.6 611.8 43.4 615.5
Sex (M/F) 5/14 3/4 5/6
Time since diagnosis [mean 6SD (mo)] N/A 2.6 61.2 53.1 675.2
Time since diagnosis [ ange (mo)] N/A 1.0–4.5 10–267
GADA posi i e (%) N/A 71.4 72.7
IA-2A posi i e (%) N/A 42.9 36.4
HLA-DR4 posi i e (%) 15.8 0 50
HLA-A2.1 posi i e (%) 31.6 40 58.4
F, emale; IA-2A, isle an igen 2 au oan ibodies; GADA, GAD
65
au oan ibodies; M, male; N/A, no
applicable.
6 Vol. 33 May 2019 The FASEB Jou nal x www. asebj.o g CALVIÑO-SAMPEDRO ET AL.
no each s a is ical signi icance (Fig. 4A)(P= 0.290; log-
ank, Man el-Cox). Howe e , accina ion does ha e an
e ec on hecellula au oimmune esponse, as he numbe
o highly in il a ed isle s is signi ican ly highe in S100-b–
accina ed animals (Fig. 4B)(P= 0.0054; Fishe ’sexac es ).
Mo eo e , nume ous GFAP-immuno eac i e cells a e ob-
se ed a ound he isle s o young nondiabe ic con ol emale
mice (Supplemen al Fig. S4A–D) compa ed wi h diabe ic
S100-b–immunized emale mice, in which hese cells a e
sca ce and show a ain GFAP immuno eac i i y (Supple-
men al Fig. S4E–H, see a ows in panels Fand H). Some
immuno eac i e in a-isle cells a e also seen in he panc ea ic
isle s o con ol animals (Supplemen al Fig. S4, a owheads in
panels B and D), which a e absen in he immunized animals.
These esul s show a dec ease in he numbe o pSC cells in
he S100-b–immunized animals compa ed wi h young
nondiabe ic insuli is- ee emale mice.
Despi e his inc ease in isle in il a ion se e i y, posi-
i e esponses agains S100
10–18
(70%, 7/10) (Fig. 4C, S100-
b) o S100
20–28
(50%, 5/10) (Fig. 4D, S100-b)a eno
signi ican ly highe han ha obse ed in nonmanipula ed
animals (S100
10–18
: 53.1%, 17/32; S100
20–28
: 48.4%, 15/31)
(Fig. 4C, D, spon aneous) o inocula ed wi h saline
(S100
10–18
: 66.7%, 8/12; S100
20–28
: 66.7%, 8/12) (Fig. 4C, D,
saline) (P.0.05; Fishe ’s exac es ). Mo eo e , accina-
ion wi h S100-b–de i ed NPPEs does no signi ican ly
modi y he in ensi y o he esponse measu ed as he me-
dian SI (P= 0.4; K uskal-Wallis es ).
To add ess whe he he unexpec ed inding o high isle
in il a ion in S100-b– ea ed animals was due o he
Figu e 2. IFN-gELISPOT analysis o esponses agains S100
10–18
and S100
20–28
pep ide epi opes. A) ROC cu es we e pe o med
o disc imina e op imal cu o alues be ween esponde s and non esponde s, gi ing an SI cu o o 1.85 and 1.75 o S100
10–18
and S100
20–28
, espec i ely. An SI $2 was used as a cu o . B) A ep esen a i e esponse om a T1D pa ien is shown (in
pa en heses, spo mean numbe 6SD o iplica e wells) o S100
10–18
and S100
20–28
. A pep ide mix u e con aining known
pep ide epi opes de i ed om i al p o eins (CEF) was included as a posi i e con ol. Spon aneous esponses we e de e mined
by PBMC cul u e in medium alone (cul u e medium) o medium con aining 0.5% DMSO. C–F) IFN-gsec e ion in esponse o
S100
10–18
(C, E) and S100
20–28
(D, F) by PBMCs om HDs, all T1D pa ien s (T1D), ND T1D pa ien s (T1D-ND), and LS T1D
pa ien s (T1D-LS). Black symbols: HLA-A2.1–nega i e subjec s; ed symbols: HLA-A2.1–posi i e subjec s. K uskal-Wallis es wi h
Dunn’spos hoc es (C,D). *P,0.05, **P,0.01, ***P,0.001. G) Responses agains bo h S100
10–18
and S100
20–28
we e mo e
equen ly de ec ed among T1D pa ien s (ND T1D: 57.1%; LS T1D: 60%) compa ed wi h HDs (11.1%). Two- ailed Fishe ’s exac
es , P= 0.0045. The e is also a posi i e co ela ion be ween he esponses o bo h pep ides ( = 0.66, Spea man’s co ela ion, P,
0.0001). The dashed line (C–G) ep esen s he h eshold.
NOVEL S100-bEPITOPES IN TYPE 1 DIABETES 7
pep ide doses, mice we e accina ed using doses 10 imes
lowe . Su p isingly, disease kine ics we e accele a ed,
eaching s a is ical signi icance when compa ed wi h bo h
saline- ea ed (P= 0.048; log- ank, Man el-Cox) and non-
manipula ed emales (Spon aneous. P=0.02;log- ank,
Man el-Cox) (Fig. 5A). Vaccina ion wi h he NPPEs also
inc eased he numbe o highly in il a ed isle s (Fig. 5B)
(P= 0.011; Fishe ’s exac es ).
Simila o he esul s seen using high doses, IFN-g e-
sponses in mice accina ed wi h S100
10–18
and S100
20–28
[6/13 (46.1%) and 7/13 (53.8%), espec i ely] a e no sig-
ni ican ly highe compa ed wi h hose seen in animals
inocula ed wi h saline [S100
10–18
: 66.7% (8/12); S100
20–28
:
66.7% (8/12)] o hose de eloping T1D spon aneously
[S100
10–18
: 61.3% (19/31); S100
20–28
: 48.4% (15/31)] (Fig.
5C, D)(P.0.05; Fishe ’s exac es ). Mo eo e , accina ion
wi h S100-b–de i ed NPPEs does no signi ican ly modi y
he in ensi y o he esponse (P=0.24 o S100
10–18
and P=
0.097 o S100
20–28
; K uskal-Wallis es ). Howe e , he
immuno he apy egime seems o ac i a e he cellula au-
oimmune esponse agains S100-bbecause a signi ican
inc ease in he cy o oxic ac i i y agains splenocy es in-
cuba ed wi h he pep ide epi opes is seen (Fig. 5E)(P,
0.05 in all e ec o : a ge a ios; Mann-Whi ney U es ),
sugges ing ha accina ion wi h S100-b–de i ed NPPEs
ac i a es an ongoing au oimmune esponse a he han
igge ing a egula o y one.
DISCUSSION
Isle in il a ion by bo h CD4
+
and CD8
+
T cells is obse ed
in T1D, bu se e al s udies ha e es ablished an impo an
ole o he la e . In ac , he cy o oxic e ec o unc ion o
hese cells is hough o be one o he mechanisms o b-cell
des uc ion in i o (38).
Du ing diabe es de elopmen , pSC cells, a ne wo k o
ne ous cells su ounding he isle mass, a e also a ge ed
by he immune sys em. In NOD mice, as insuli is p og-
esses, pSC cells a e g adually elimina ed (16–18). pSC
cells exp ess an igens ha a e ei he speci ic, such as
GFAP, o sha ed wi h he b-cells, such as he neu o ophic
ac o S100-b(16). In e es ingly, many au oan igens a -
ge ed by he immune esponse in T1D a e exp essed by
di e en ne ous and neu oendoc ine issues, such as
GAD
65
, glu amic acid deca boxylase 67, isle cell an igen
512 (also e e ed o as IA-2), phog in (also e e ed o as
IA-2–b), isle cell an igen 69, o ch omog anin A (39).
Ou g oup was he i s o de ine NPPEs de i ed om
S100-band es ic edby heHLA-DRB1*04:01(HLA-DR4)
class II molecule. T cells om bo h ND and long- e m T1D
pa ien s ecognized he S100-b–de i ed NPPEs and se-
c e ed IFN-g(23). In he p esen s udy, we de ined
S100
10–18
and S100
20–28
as new HLA-A2.1– es ic ed epi-
opes. Those epi opes we e ecognized mo e equen ly by
T cells om bo h ND and long- e m T1D pa ien s. Mo e-
o e , he equency o dual esponde s (i.e., hose
esponding o bo h S100-bpep ides) was highe in T1D
pa ien s, pa icula ly long- e m ones. P e ious s udies
ha e shown ha di e en ypes o au oimmune esponses
could be aking place in T1D pa ien s. Di e en ial insuli ic
p o iles de e mine he ex en o b-cell des uc ion and he
age a onse (40); he e o e, i emains possible ha youn-
ge T1D pa ien s han he ones shown in ou s udy show
mo e obus S100-b–speci ic cy o oxic esponses, as p e-
iously shown by Banwell e al. o whole an igen in chil-
d en using a p oli e a ion assay (22).
Figu e 3. Spon aneous au oimmune esponses agains S100
10-
18
and S100
20-28
in A2.1- ansgenic NOD mice. A) Spon aneous
au oimmune esponses agains S100
10–18
and S100
20–28
we e
de e mined in splenocy es by IFN-gELISPOT in male (n= 22;
10.32 64.76 wk o age, mean 6SD) and emale (n= 31; 9.83 6
3.97, mean 6SD) A2.1- ansgenic NOD mice. Da a shown a e
he SIs (yaxis) calcula ed as he a io be ween he numbe o
spo s wi h pep ide and he numbe o spo s in cul u e medium
only. SIs $2 we e conside ed posi i e esponses. Median
esponses we e highe in emales compa ed wi h males o
S100
10–18.
B,C) Mann-Whi ney U es . P= 0.035. Spon aneous
esponses agains S100
10–18
(B) and S100
20–28
(C) acco ding o
age we e de e mined in emale mice (n= 31; 5–10 animals/age
g oup). Posi i e esponses (SI $2) can be de ec ed as ea ly as
5 wk o age o bo h pep ide epi opes. The dashed line
ep esen s he h eshold.
8 Vol. 33 May 2019 The FASEB Jou nal x www. asebj.o g CALVIÑO-SAMPEDRO ET AL.
S100
10–18
lies in he same egion as he HLA-DR4 pep ide
epi ope S100
6–25
(KAMVALIDVFHQYSGREGDK) and
S100
20–28
in he egion o S100
21–36
(REGDKHKLKKSELKEL).
Fo o he well-cha ac e ized au oan igens in T1D, MHC class
I–and II–binding pep ides ha e been iden i ied in o e lapping
egions o in close p oximi y (21, 24, 41). Recen da a in NOD
mice iden i ied dominan class II– es ic ed epi opes (S100
1-15
and S100
78–92
) (25) in egions p e iously iden i ied by us in
human T1D (23). These da a poin o he ac ha hose egions
a e immunodominan in bo h NOD mice and human T1D
pa ien s and cons i u e impo an a ge s o he au oimmune
esponse in diabe es.
P oli e a i e lymphocy e esponses agains GFAP and
S100-bwhole an igens ha e been de ec ed in bo h NOD
mice and human T1D pa ien s (16, 22). Two GFAP-de i ed
K
d
- es ic ed epi opes ha e been iden i ied in he NOD
mice. Howe e , when e alua ed o hei po en ial in
p e en ing diabe es, only he one wi h he highe MHC
binding a ini y was able o signi ican ly delay T1D de-
elopmen (21). Ou esul s show ha esponses agains
HLA-A2.1– es ic ed S100-bNPPEs can be de ec ed as
ea ly as 4 wk o age; howe e , when used o p e en dis-
ease de elopmen , an accele a ion was obse ed, mo e
agg essi e insuli is was de ec ed, no educ ion o IFN-g
sec e ion was seen, and an inc ease in cy o oxic ac i i y
agains S100-b–loaded a ge s was obse ed.
Se e aldi e encesbe ween hep esen s udyand hose
using o he pep ide an igens o p e en T1D could explain
he di e ences. Fi s , GFAP-de i ed pep ide epi opes
we e iden i ied using algo i hms p edic ing high binde s,
whe eas ou S100-b–de i ed pep ide epi opes a e ue
NPPEs wi h low o no de ec able a ini y o HLA-A2.1
using an HLA s abiliza ion assay. The e is a co ela ion
be ween class I binding a ini y and immunogenici y o
CTL pep ide epi opes (42). I is belie ed ha pep ide epi-
opes a ge ed by CTLs adminis e ed sys emically in high
doses and in aqueous solu ions induce CTL pe iphe al
ole ance, p obably due o he dele ion o pep ide-speci ic
CTLs (43, 44). Fo example, a single subcu aneous acci-
na ion using a pep ide de i ed om he adeno i us ype
5 ea ly egion 1A oncogene induce a speci ic CTL
ole iza ion agains adeno i us ype 5 ea ly egion 1A
oncogene–exp essing umo cells; howe e , accina ion
wi h a di e en pep ide (human papilloma i us ype 16
E7
49–57
) in simila condi ions induces p o ec ion in
mice challenged wi h human papilloma i us ype 16–
ans o med umo cells (44). An impo an di e ence
be ween bo h pep ides is hei a ini y o he es ic ion
elemen (H-2D
b
), wi h a highe binding a ini y showing
o he i s pep ide. Fo class II– es ic ed pep ide epi-
opes, pep ides wi h highe MHC binding a ini y showed
a highe capaci y o p e en expe imen al au oimmune
encephalomyeli is and induce ole ance when adminis-
e ed in condi ions simila o hose shown in he p esen
s udy (45).
Second, MHC class I– es ic ed GFAP pep ides (21)
we e adminis e ed in incomple e F eund’sadju an (IFA)
compa ed wi h saline in he p esen s udy, and his may
con ibu e o hei in i o s abili y. In cance animal
models, pep ide/IFA accina ion p imed umo -speci ic
Figu e 4. T1D de elopmen in
A2.1- ansgenic NOD mice ac-
cina ed in ape i oneally wi h
S100
10–18
and S100
20–28
.A)
Animals we e accina ed wi h
100 mg each o S100
10–18
and
S100
20–28
pe dose (n= 10;
con inuous line) s a ing a
4 wk o age and hen e e y
2–3 wk un il hey de eloped
diabe es o eached 30 wk o
age. Animals inocula ed wi h
saline (n= 12; do ed line) we e
used as con ols. Vaccina ion
wi h S100
10–18
and S100
20–28
does no seem o significan ly
modi y ei he he incidence o
he kine ics o diabe es de el-
opmen in A2.1- ansgenic
NOD mice. Log- ank, Man el-
Cox, P= 0.29. B) Insuli is in
A2.1- ansgenic NOD mice ac-
cina ed wi h S100-b–de i ed
NPPEs o inocula ed wi h saline.
The pe cen age o isle s hea ily
infil a ed (sco es 3 and 4) is
significan ly highe in S100-b–
accina ed animals compa ed
wi h hose inocula ed wi h saline. Fishe ’s exac es ; P=0.0054.C,D)IFN-gsec e ion in esponse o S100
10–18
(C)and
S100
20–28
(D) in A2.1- ansgenic NOD mice accina ed wi h he S100-b–de i ed NPPEs (S100-b), saline (saline), o nonmanipula ed
(spon aneous). In all cases, splenocy es we e ha es ed a disease diagnosis o a 30 wk o age. Nei he equency o posi i e
esponses (Fishe ’s exac es , P.0.05) no median SI esponses (K uskal-Wallis es , P=0.4)we esignifican ly di e en in S100-
b–immunized mice compa ed wi h spon aneous and saline-inocula ed mice.
NOVEL S100-bEPITOPES IN TYPE 1 DIABETES 9
