A hype mo phic N kbid allele con ibu es o impai ed hymic
dele ion o au o eac i e diabe ogenic CD8+ T-cells in NOD mice1,,2
Maximiliano P esa*, Je emy J. Racine*, Jenni e R. Dwye *, Deanna J. Lamon *, Je emy J.
Ra iu*, Vishal Kuma Sa sani*, Yi-Guang Chen†, A on Geu s†, Ingo Schmi z‡,§, Timo hy
S ea ns*, Jenni e Allocco*, Ha old D. Chapman*, and D . Da id V. Se eze, PhD*
* The Jackson Labo a o y, Ba Ha bo , Maine USA.
† Medical College o Wisconsin, Milwaukee, Wisconsin USA
‡ Sys ems-O ien ed Immunology and In lamma ion Resea ch G oup, Helmhol z Cen e o
In ec ion Resea ch, B aunschweig Ge many
§ Ins i u e o Molecula and Clinical Immunology, O o- on-Gue icke Uni e si y, Magdebu g
Ge many
Abs ac
In bo h NOD mice and humans, he de elopmen o ype 1 diabe es (T1D) is dependen in pa on
au o eac i e CD8+ T-cells ecognizing panc ea ic ß-cell pep ides p esen ed by o en qui e common
MHC class I a ian s. S udies in NOD mice p e iously e ealed he common H2-Kd and/o H2-Db
class I molecules exp essed by his s ain abe an ly lose he abili y o media e he hymic dele ion
o pa hogenic CD8+ T-cell esponses h ough in e ac ions wi h T1D suscep ibili y (
Idd
) genes
ou side he MHC. A gene(s) mapping o p oximal Ch omosome 7 was p e iously shown o be an
impo an con ibu o o he ailu e o he common class I molecules exp essed by NOD mice o
media e he no mal hymic nega i e selec ion o diabe ogenic CD8+ T-cells. Using an inducible
model o hymic nega i e selec ion and mRNA ansc ip analyses we ini ially iden i ied an
ele a ed
N kbid
exp ession a ian as a likely NOD p oximal Ch omosome 7 egion gene
con ibu ing o impai ed hymic dele ion o diabe ogenic CD8+ T-cells. CRISPR/Cas9-media ed
gene ic a enua ion o
N kbid
exp ession in NOD mice esul ed in imp o ed nega i e selec ion o
au o eac i e diabe ogenic AI4 and NY8.3 CD8+ T-cells. These esul s indica ed allelic a ian s o
N kbid
con ibu e o he e iciency o in a hymic dele ion o diabe ogenic CD8+ T-cells. Howe e ,
while enhancing hymic dele ion o pa hogenic CD8+ T-cells, abla ing
N kbid
exp ession
1MP was suppo ed by JDRF Fellowship 3-PDF-2014–219-A-N. Fo pa s o his wo k, JJR1 was suppo ed by ei he NIH Fellowship
1F32DK111078 o JDRF Fellowship 3-PDF-2017–372-A-N. He was also suppo ed by a g an om he Diabe es Resea ch
Connec ion DRC 006887 JR. DVS is suppo ed by NIH g an s DK-46266, DK-95735, and OD-020351–5022, as well as by Ju enile
Diabe es Resea ch Founda ion g an 2018–568. This wo k was also pa ly suppo ed by Cance Cen e Suppo G an CA34196. IS
was suppo ed by g an s o he Deu sche Fo schunggemeinscha (SCHM1586/6–1 and p ojec A23 o SFB854).
2MP designed and conduc ed expe imen a ion, in e p e ed da a, and w o e he manusc ip . JJR1, designed and conduc ed
expe imen a ion, con ibu ed o da a in e p e a ion, and w i ing o he manusc ip . JRD designed and conduc ed expe imen s,
in e p e ed da a and con ibu ed o w i ing he manusc ip . JJR2, DJL, JA and HDC conduc ed expe imen a ion. VKS and TS
con ibu ed o s a is ical analyses. YGC, AG and IS con ibu ed o expe imen al design. DVS con ibu ed o s udy concep ion,
supe ised expe imen al e o , and w i ing o he manusc ip .
Co espondence: The Jackson Labo a o y, 600 Main S ., Ba Ha bo , Maine 04609, da [email p o ec ed], (207) 288-6403.
HHS Public Access
Au ho manusc ip
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Published in inal edi ed o m as:
J Immunol
. 2018 Oc obe 01; 201(7): 1907–1917. doi:10.4049/jimmunol.1800465.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip
su p isingly accele a ed T1D onse in NOD mice ha was associa ed wi h nume ic dec eases in
bo h egula o y T- and B-lymphocy es (T egs/B egs).
In oduc ion:
In bo h he NOD mouse model and humans, ype 1 diabe es (T1D3) esul s om he
au oimmune des uc ion o insulin-p oducing panc ea ic β-cells media ed by he combined
ac i i y o CD4+ and CD8+ T-cells, as well as B-lymphocy es (1). T1D is highly polygenic
in na u e, as mo e han 50 loci ha e been associa ed wi h disease suscep ibili y o esis ance
in bo h humans and NOD mice (2, 3). While i s comple e pa hogenic e iology emains
unsol ed, he e is a wide accep ance ha disease de elops as a consequence o in e ac ions
be ween T1D suscep ibili y (
Idd
) genes esul ing in b eakdowns in mechanisms con olling
induc ion o immunological ole ance o sel -p o eins (4, 5). Some
Idd
genes appea o
con ibu e o de ec s in cen al ole ance mechanisms ha no mally ep esen a i s
checkpoin engende ing he hymic dele ion o au o eac i e CD4+ and CD8+ T-cells du ing
ea ly s ages o hei de elopmen (6–8). Howe e , e en in heal hy indi iduals, some
au o eac i e T-cells escape cen al ole ance mechanisms (9). Such au o eac i e e ec o s a e
no mally p e en ed om media ing pa hogenic e ec s by a second checkpoin , p o ided by
mul iple mechanisms o pe iphe al ole ance in ol ing egula o y lymphocy e popula ions
(T egs/B egs), wi h signi ican e idence o some o hese p ocesses also being dis up ed by
a ious
Idd
genes (5).
Imma u e CD4+CD8+ double posi i e (DP) hymocy es exp essing TCR molecules allowing
o high a idi y in e ac ions wi h APC displaying sel -pep ide-MHC complexes a e no mally
dele ed by apop osis o di e ed o a CD4+Foxp3+ T eg lineage (9).
Idd
gene-elici ed
de iciencies in his p ocess con ibu e o he hymic su i al and pe iphe al seeding o bo h
au o eac i e CD4+ and CD8+ T-cells ha o e whelm he supp essi e capaci y o sel -
an igen-speci ic T egs (4, 5). MHC class II- es ic ed au o eac i e CD4+ T-cells clea ly play
a c i ical ole in T1D de elopmen . Howe e , hey canno do so by di ec ly engaging insulin-
p oducing panc ea ic β-cells ha only exp ess MHC class I molecules. Hence, h ough an
abili y o di ec ly engage and des oy panc ea ic β cells, MHC class I es ic ed au o eac i e
CD8+ T-cells a e likely he ul ima e media o s o T1D de elopmen in bo h humans and
NOD mice (10, 11). Indeed, epidemiological s udies ha e shown ha in addi ion o class II
e ec s, pa icula HLA class I a ian s p o ide an independen T1D isk ac o in humans
(12, 13). In e es ingly, his includes some qui e common MHC class I a ian s ha in bo h
humans and NOD mice only acqui e an abe an abili y o con ibu e o T1D when exp essed
in he con ex o o he
Idd
genes.
3C57BL/6J, (B6); CD4+CD8+ double posi i e, (DP); CD4+Foxp3+ egula o y T-cell, (T eg); ch omosome, (Ch .); dend i ic cell,
(DC); Lymphocy ic Cho iomeningi is Vi us, (LCMV); H2-Db/MimA2 e ame capable o binding he AI4 TCR, (Te -AI4); H2-Kd/
NRPV7 e ame capable o binding he NY8.3 TCR, (Te -NRPV7); isle -speci ic glucose-6-phospha ase ca aly ic subuni - ela ed
p o ein, (IGRP); NOD/ShiL D s, (NOD); NOD/ShiL D s-
N kbid
<em3D s>/D s, (NOD-
N kbid−/−);
NOD.B6-(Gpi1-D7Mi 346)/
L J, (NOD.
Ch 7B6
FL); NOD.Cg-
P kdcscid
Em 30b/D s, (NOD.
scid
); panc ea ic lymph node, (PLN); p opidium iodide, (PI);
egula o y B-cell, (B eg); single guide RNA, (sgRNA); ype 1 diabe es, (T1D); spleen, (SPL); hymus, (Thy); ype 1 diabe es
suscep ibili y, (
Idd
).
P esa e al. Page 2
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip
P e ious s udies ound in e ac i e con ibu ions om some polymo phic genes ou side o he
MHC de e mine he ex en o which pa icula class I molecules can allow he de elopmen
and unc ional ac i a ion o diabe ogenic CD8+ T-cells. This was ini ially demons a ed by
congenically ans e ing a ansgenic Vα8+Vβ2+ TCR om he H2g7 class I es ic ed AI4
diabe ogenic CD8+ T-cell clone om a NOD gene ic backg ound s ain (NOD-AI4) o a
T1D esis an C57BL/6-s ock (B6) congenic o he
H2g7
haplo ype (B6.
H2g7
). Au o eac i e
AI4 TCR ansgenic T-cells we e dele ed o a signi ican ly g ea e ex en a he DP s age o
hymic de elopmen in B6.
H2g7
han NOD mice (7). This esul indica ed non-MHC genes
allelically di e ing in NOD and B6 mice egula e he ex en o which au o eac i e
diabe ogenic CD8+ T-cells unde go hymic dele ion. Such non-MHC genes con olling
di e ing hymic nega i e selec ion e iciency in NOD and B6 backg ound mice we e ound
o unc ion in a T-cell in insic manne (7). Subsequen linkage analyses mapped a gene(s)
on he p oximal egion o Ch omosome (Ch .) 7 s ongly con ibu ing o he di e en ial
le els o AI4 DP hymocy es in NOD and B6.
H2g7
mice (8). This inding was con i med by
subsequen analyses o an NOD-AI4 TCR ansgenic s ock ca ying a B6 o igin p oximal
Ch . 7 congenic in e al delinea ed by he lanking ma ke s
D7Mi 267
(29.5 Mb) and
D7Mi 346
(58.7Mb) (14) (s ain now designa ed NOD.
Ch 7B6
FL-AI4). The
Ch 7B6
FL
congenic in e al was ound o con ain a gene(s) in insically elici ing a signi ican dec ease
in numbe s o AI4 DP hymocy es han in o he wise ully NOD gene ic backg ound mice
(8).
A locus e med
Idd7
locus had also been p e iously mapped o p oximal Ch . 7 (15, 16).
Howe e , he
Idd7
gene ic locus has been enigma ic in ha he NOD a ian was o iginally
iden i ied in seg ega ion s udies as con ibu ing o T1D esis ance a he han suscep ibili y
(15, 16). This o iginal inding became a bi mo e pe plexing by he abo e desc ibed inding
ha he
Ch 7B6
FL congenic in e al exe ed an e ec esul ing in signi ican ly ewe
numbe s o AI4 DP hymocy es han in o he wise ully NOD gene ic backg ound mice (8).
Thus, i is unclea i he p oximal Ch . 7 egion gene(s) con olling he nega i e selec ion
e iciency o diabe ogenic AI4 CD8 T-cells is he same as ha esponsible o he
Idd7
e ec
on disease de elopmen .
In his s udy, we iden i y a di e en ial exp ession-le el a ian s o
N kbid
(aka
IκBNS
) ha
can modula e NF-κB ac i i y (17) as a likely p oximal Ch omosome 7 egion gene
egula ing he e iciency o diabe ogenic CD8+ T-cell hymic nega i e selec ion. Using TCR
ansgenic mouse models we show ha a highe
N kbid
exp ession a ian in NOD
compa ed o B6 mice con ibu es o wha is likely diminished nega i e selec ion in he
o me s ain o bo h he AI4 and NY8.3 MHC class I es ic ed au o eac i e diabe ogenic
CD8+ T-cell clono ypes. While gene ic a enua ion o
N kbid
exp ession o he lowe le els
obse ed in B6 backg ound mice esul s in imp o ed hymic dele ion o pa hogenic CD8+ T-
cells, o al abla ion o his gene also nume ically diminishes le els o pe iphe al T egs and
egula o y B-lymphocy es (B egs) associa ed wi h a mo e apid onse o T1D in he mu an
s ock han in s anda d NOD mice.
P esa e al. Page 3
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip
Ma e ials and Me hods
Mouse s ains
NOD/ShiL D s (he ea e NOD), C57BL/6J (B6), and NOD.Cg-
P kdcscid
Em 30b/D s
(NOD.
scid
) mice a e main ained in a speci ic pa hogen ee esea ch colony a The Jackson
Labo a o y. NOD mice ca ying a ansgenic Vα8.3+Vβ2+ TCR de i ed om he AI4
diabe ogenic CD8+ T-cell clone (NOD-AI4) ha e also been p e iously desc ibed (18).
P e iously gene a ed NOD.LCMV mice ansgenically exp ess a Vα2+Vβ8+ TCR om a
CD8+ T-cell clone ecognizing he H2-Db es ic ed gp33 pep ide (KAVYNFATM) de i ed
om Lymphocy ic Cho iomeningi is Vi us (LCMV) (19, 20). NOD.NY8.3 mice
ansgenically exp essing he TCR de i ed om he H2-Kd es ic ed diabe ogenic NY8.3
CD8+ T-cell clone (21) a e main ained in an he e ozygous s a e. A NOD s ock congenic o
a segmen o B6-de i ed Ch . 7 delinea ed by he lanking ma ke s
D7Mi 276
(29.5 Mb) and
D7Mi 346
(58.7 Mb) (14) (he ea e abb e ia ed NOD.
Ch 7B6
FL) was used o in oduce he
congenic in e al in o NOD-AI4 mice (8).
Congenic unca ion analysis
The NOD.
Ch 7B6
FL s ock was in e c ossed wi h NOD-AI4. Resul an F1 p ogeny we e
in e c ossed and all F2 o sp ing we e analyzed o ecombina ion e en s wi hin he o iginal
Ch . 7 FL congenic egion. Genomic ail DNA samples we e sc eened o he mic osa elli e
ma ke s
D7Mi 267
,
D7Mi 117, D7Mi 155, Gpi1b, D7Mi 79, D7Mi 225, D7Mi 78,
D7Mi 247, D7Mi 270, D7Mi 230,
and
D7Mi 346
by PCR. Iden i ied ecombinan mice
we e backc ossed o NOD-AI4 and hen in e c ossed o gene a e p ogeny homozygous o
he sub-congenic egions. The homozygous s a e o each sub-congenic line was e i ied by
PCR analyses o he ma ke s lis ed abo e. Two lines esul ed om his p ocess, Ln82 wi h a
sub-congenic egion
D7Mi 117
–
D7Mi 247
and Ln16, wi h a sub-congenic egion
D7Mi 79
–
D7Mi 247
.
Gene a ion o NOD N kbid-de icien mice
CRISPR/Cas9 echnology was u ilized o di ec ly abla e he
N kbid
gene in NOD mice.
NOD/ShiL D s emb yos we e mic oinjec ed wi h 3pl o a solu ion con aining Cas9 mRNA
and single guide RNA (sgRNA) a espec i e concen a ions o 100ng/μl and 50ng/μl. The
sgRNA sequence (5’-AGGCCCATTTCCCCTGGTGA-3’) was designed o dele e he
ansc ip ional s a si e o
N kbid
in exon 3. Genomic ail DNA was sc eened by a ge ed
Sange sequencing: he genomic egion a ound exon 3 was ampli ied by PCR wi h p ime s
N kbid-KO-F1 (5’-TGCTTGAGATCCAGTAG-3’) and N kbid-KO-R1 (5’-
CCCTGACATCTCAGAATA-3’). The esul ing PCR p oduc was pu i ied and sequenced
using an ABI 3730 DNA analyze (Applied Biosys ems-The mo Fishe Scien i ic, Wal ham,
MA, USA). Analyses o Sange sequencing da a was done using he so wa e Poly Peak
Pa se (22) which allows iden i ica ion o he di e en alleles p esen in he e ozygous
mu an mice. Du ing he sc eening o N1 mu an s, we iden i ied an allele cha ac e ized by an
8-nucleo ide dele ion and an inse ion o 154-nucleo ides o igina ed by copy and in e sion
o a DNA segmen om
N kbid
exon 2. This p oduced a majo dis up ion o he
N kbid
gene and dele ion o he e e ence ansla ion s a si e (Supplemen a y Fig. 1). N1F1 NOD-
N kbid+/−
mu an mice we e in e c ossed and he dis up ed allele ixed o homozygosi y.
P esa e al. Page 4
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip
The new s ock, o mally designa ed NOD/ShiL D s-
N kbid
<em3D s>/D s (he ea e
abb e ia ed NOD-
N kbid−/−
), was main ained by b o he -sis e ma ing.
N kbid gene exp ession analysis
To al RNA om whole hymus o so ed DP cell lysa es was ex ac ed using he RNeasy
Mini ki (Qiagen, Ge man own, MD, USA). Fo whole hymus issue RNA, he comple e
o gan was pu in 2ml o RNAla e s abiliza ion solu ion (The mo Fishe Scien i ic),
incuba ed 24 h a oom empe a u e and hen s o ed a −20˚C un il p ocessing. In o he
expe imen s 2×105 DP cells we e so ed in 100% FBS, washed 2 imes in cold PBS and
esuspended in RLT lysis bu e (Qiagen) and s o ed a −20˚C un il p ocessing. Ex ac ed
o al RNA was quan i ied by NanoD op (The mo Fishe Scien i ic) and quali y assessed in
Bioanalyze (Agilen Technologies, San a Cla a, CA, USA). Samples wi h a RNA in eg i y
index (RIN index) o 7 o highe we e used o gene exp ession analyses. Fo cDNA
syn hesis 500ng o o al RNA was dilu ed o 5μl in DEPC ea ed wa e and mixed wi h 5μl
o Supe Sc ip IV VILO Mas e Mix (The mo Fishe Scien i ic) ollowing endo
ins uc ions.
Exp ession o
N kbid
was assessed by a p edesigned and alida ed TaqMan qPCR assay
using Mm.PT.58.12759232 (IDT In eg a ed Technologies, Co al ille, Iowa, USA) a ge ing
he exon 3–4 egion and able o iden i y all ou possible ansc ip s om wild ype alleles.
The assay consis s o a FAM-labeled and double-quenched p obe (5’-/56-FAM-
CCTGGTGAT-/ZEN/-GGAGGACTCTCTGGAT-/3IABkFQ/−3’) ha spans he 8-
nucleo ide dele ion si e and he p ime s p1 (5’-GACAGGGAAGGCTCAGGATA-3’) and p2
(5’-GCTTCCTGACTCCTGATTTCTAC-3’). The assay was used a a inal concen a ion o
500nM p ime s and 250nM p obe. As an endogenous con ol, we used a p edesigned and
alida ed TaqMan assay o
Gapdh
(ID: Mm99999915_g1) (VIC-labeled) (Applied
Biosys ems – The mo Fishe Scien i c). The qPCR analyses we e done using a ViiA-7 eal
ime PCR sys em (Applied Biosys ems - The mo Fishe Scien i ic). Rela i e gene
exp ession was de e mined by he ddC me hod using he applica ion RQ in The mo Fishe
Cloud So wa e, Ve sion 1.0 (The mo Fishe Scien i ic).
Mic oa ay analysis4
Fi e-week-old NOD.LCMV and NOD.Ln82-LCMV emale mice we e i. . injec ed wi h
0.5μg o gp33 o Db binding con ol pep ide (ASNENMETM) (n=3–5 biological eplica es
pe s ain and ea men ype). Two hou s pos -injec ion, whole hymus issue was p ocessed
as desc ibed abo e o o al RNA ex ac ion. RNA samples we e hyb idized o A yme ix
Mouse Gene 1.0 ST A ays (A yme ix – The mo Fishe Scien i ic). A e age signal
in ensi ies o each p obe se wi hin a ays we e calcula ed by and expo ed om
A yme ix’s Exp ession Console (Ve sion 1.1) so wa e using he RMA me hod. Two
pai wise compa isons we e used o s a is ically esol e gene exp ession di e ences be ween
expe imen al g oups using he R/maano a analysis package (23). Speci ically, di e en ially
exp essed genes we e de ec ed by using F1, he classical F-s a is ic ha only uses da a om
indi idual genes. S a is ical signi icance le els o he pai wise compa isons we e calcula ed
4Mic oa ay da a has been deposi ed a [email p o ec ed]. Accession # GSE115754.
P esa e al. Page 5
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip

by pe mu a ion analysis (1000 pe mu a ions) and adjus ed o mul iple es ing using he alse
disco e y a e (FDR), q- alue, me hod (24). Di e en ially exp essed genes a e decla ed a
an FDR q- alue h eshold o 0.05. Two con as g oups we e es ablished: con ol Db pep ide
(NOD.LCMV s NOD.Ln82-LCMV), and gp33 pep ide (NOD.LCMV s NOD.Ln82-
LCMV). Da a we e il e ed o an exp ession old change (EFC) > 2 and an exp ession FDR
(q- alue) < 0.05. Genes mee ing hese c i e ia a e iden i ied in Supplemen a y Table 1, and
NF-κB a ge genes a e u he anno a ed o unc ion as documen ed in D . Thomas
Gilmo e’s cu a ed da abase (h p://www.bu.edu/n -kb/gene- esou ces/ a ge -genes/, accessed
3/26/2018).
Gene Exp ession analysis o NF-κB amily membe s and NF-κB egula ed genes
Li e DP hymocy es we e so ed om 5–6-week-old NOD-AI4 and NOD-AI4-
N kbid−/−
mice di ec ly in o e al bo ine se um using a FACSA ia II so e (BD Biosciences San Jose,
CA, USA). Cells we e pelle ed, esuspended in TRIzol Reagen (The mo Fishe Scien i ic)
and ozen un il use. Following phenol-chlo o o m phase sepa a ion, RNA was pu i ied om
he aqueous phase using a Quick-RNA MiniP ep Plus column (Zymo Resea ch, I ine, CA).
Following RNA quan i a ion and e e se ansc ip ion as desc ibed abo e, qPCR was
conduc ed using Powe SYBR G een PCR mas e mix (Applied Biosys ems – The mo
Fishe Scien i ic) wi h 250nM p ime s and 6ng o cDNA empla e. P ime s a e documen ed
in Supplemen a y Table 2. S anda d cu es o each gene we e pe o med using pooled
cDNA, gene exp ession was no malized o
Gapdh
and exp essed as old change compa ed o
NOD-AI4 con ol samples.
Flow cy ome y analysis
Single cell suspensions o hymus (Thy), spleen (SPL) and panc ea ic lymph nodes (PLN)
om 5–11-week-old mice we e p epa ed. Red blood cells in splenocy e samples we e lysed
wi h Gey’s bu e (25). Aliquo s o 2–4×106 cells we e s ained wi h he ollowing
monoclonal an ibodies: CD4-BV650, -BV785, o FITC (GK1.5), CD8-PE-Cy7, -BV480 o -
APC (53–6.72), TCRβ-BV711 (H57–597), TCRVα8.3-FITC o -PE (B21.14), TCRVβ2-
A647 (B20.6), TCRVβ8.1,2,3-FITC (F23.1) CD19-A700 o -RG710 (ID3), B220-BUV496
(RA3–6B2), CD11b-PB (M1/70), CD11c-PB (N418), CD21-FITC (7G6), CD23-PE (B3B4),
CD45.1-BV421 (A201), CD5-PE (53–7.313), IL-10-APC (JES5–16E3), Foxp3-AF488
(FJK-16s) acqui ed om BD Bioscience, In i ogen - The mo Fishe Scien i ic, o
BioLegend (San Diego, CA, USA), MHC class I e ame s Te -AI4-PE (H2-Db/MimA2,
sequence YAIENYLEL) (26) and Te -NRPV7-PE (H2-Kd/NRPV7, sequence
KYNKANVFL) acqui ed om he Na ional Ins i u es o Heal h e ame co e acili y
(A lan a, GA, USA). Dead cells we e excluded by DAPI o p opidium iodide (PI) s aining
o un ixed samples, o by Fixable Viabili y Dye eFluo 780 (eBioscience - The mo Fishe
Scien i ic) s aining o samples ixed wi h Mouse Foxp3 Bu e Se (BD Bioscience).
S ained cells we e acqui ed using a BD LSRII (BD Bioscience), BD FACSymphony A5 (BD
Bioscience), o A une 1s Gene a ion Flow Cy ome e (The mo Fishe Scien i ic). All low
cy ome ic da a we e analyzed wi h FlowJo (FlowJo LLC, Ashland, OR, USA).
P esa e al. Page 6
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip
T-cell P oli e a ion assay
To al splenic T-cells we e pu i ied om 5-week-old NOD and NOD-
N kbid
−/− emale mice
by nega i e deple ion o CD11c+, CD11b+, B220+, Te 19+, CD49b+ cells wi h bio in-
conjuga ed an ibodies and s ep a idin magne ic beads o e MACS columns (Mil enyi
Bio ech, Cologne, Ge many). Pu i ied T-cells we e washed wi h cold PBS, coun ed and
adjus ed o a concen a ion o 2×107/ml o labeling wi h 10μM cell ace eFluo 670
(eBioscience - The mo Fishe Scien i ic). Labeled cells we e adjus ed o 4×106/ml. Whole
collagenase diges ed splenocy es om NOD.
scid
mice we e used as APCs. Labeled T-cells
we e seeded in iplica e in o a 96 well pla e a a densi y o 2×105/well wi h 4×105 APC/
well. An i-CD3 monoclonal an ibody (145–2C11) was added a inal concen a ions anging
om 4–0.125μg/ml. The cells we e incuba ed a 37˚C, 5% CO2 o 48h. T-cell p oli e a ion
was assessed by low cy ome ic analysis o cell ace eFluo 670 dilu ion. P oli e a ion
pa ame e s we e ob ained using he p oli e a ion pla o m analysis in FlowJo (FlowJo LLC).
Cy okine sec e ion analysis
IL-2 and IFN-γ sec e ion was assessed in supe na an s o T-cell p oli e a ion cul u es by
ELISA using BD Op EIA ki s (BD Bioscience).
Wes e n blo
Whole hymus lysa es we e p epa ed in RIPA bu e (Cell Signaling Technologies, Dan e s,
MA, USA) con aining Hal ™ P o ease Inhibi o Cock ail (The mo Fishe Scien i ic). To al
p o ein con en was quan i ied by B ad o d assay (The mo Fishe Scien i ic) and adjus ed o
4mg/ml in RIPA bu e . P o ein lysa es we e p epa ed o au oma ed wes e n blo using he
Simple Wes sys em (P o einSimple, San Jose, CA, USA). B ie ly, samples we e mixed a a
4:1 a io wi h 5x luo escen mas e mix and hea ed a 95°C o 5 min. Samples plus
bio inyla ed molecula weigh s anda ds (P o einSimple) we e loaded along wi h blocking
solu ion, bio in labeling eagen , wash bu e s, p ima y an ibodies, ho se adish-pe oxidase
conjuga ed seconda y an ibodies and chemiluminescen subs a e in o a pla e p e illed wi h
s acking and sepa a ion ma ices. In o de o pe o m o al p o ein no maliza ion, each
sample was loaded in duplica e; one capilla y was used o immunode ec ion o N kbid and
he second o o al p o ein quan i ica ion. A 1:100 dilu ion o abbi polyclonal an i-N kbid
(27) wi h a 60 min incuba ion ime was used o immunode ec ion. Chemiluminescen signal
was cap u ed and he esul ing image was analyzed by he Compass o Simple Wes e n
so wa e package (P o einSimple). N kbid peak a ea was no malized o o al p o ein a ea
ollowing P o einSimple ecommenda ions and s anda d guidelines (28, 29). Speci ically,
he N kbid a ea o each sample was mul iplied by a no maliza ion ac o (calcula ed as he
o al p o ein o he B6 sample di ided by he o al p o ein o each indi idual sample).
Diabe es incidence
Diabe es was moni o ed once weekly using u ine glucose s ips (Dias ix, Baye , Le e kusen,
Ge many). Mice wi h wo consecu i e eadings >250mg/dl (co esponding o a blood
glucose o >300mg/dl) we e conside ed diabe ic.
P esa e al. Page 7
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip
S a is ical analysis
Da a analysis and g aphs we e made using P ism 6 so wa e (G aphPad, San Diego, CA,
USA). S a is ical analyses a e de ailed in he co esponding Figu e Legends.
Resul s
Mapping an Idd7 egion gene(s) egula ing hymic numbe s o diabe ogenic AI4 CD8+ T-
cells o a 5.4 Mb egion on Ch . 7.
Among he mul iple epi opes ecognized by he p omiscuous AI4 TCR is an insulin de i ed
pep ide (30). Insulin is hymically exp essed in NOD mice (30). Thus, AI4 T-cells
de eloping in he hymus a e an igenically exposed. Such hymic an igenic a ailabili y
indica es a gene(s) allelically a ying be ween NOD and B6 mice on p oximal Ch . 7 likely
plays a s ong ole in con olling he ex en o diabe ogenic AI4 CD8+ T-cell nega i e
selec ion. We subsequen ly de i ed om he o iginal NOD.
Ch 7B6
FL-AI4 s ock wo sub-
congenic lines designa ed NOD.Ln82-AI4 (maximum leng h de ined by he ma ke s
D7Mi 117
–
D7Mi 247
) and NOD.Ln16-AI4 (maximum leng h de ined by he ma ke s
D7Mi 79
–
D7Mi 247
) (Fig. 1A). T1D de elopmen in he o iginal NOD.
Ch 7B6
FL-AI4
congenic s ock did no di e om s anda d NOD con ols (8), indica ing sha ed loci
be ween he wo s ains likely con ibu ed o au o eac i e T-cell ac i a ion, expansion, and
pa hogenici y sepa a e om he obse ed hymic selec ion e ec s o he congenic egion.
The e o e, hymic selec ion, and no o e T1D de elopmen , was e alua ed in he Ln82 and
Ln16 sub-congenic s ocks. Flow cy ome ic analyses ound ha compa ed o NOD-AI4
con ols, he yield o DP Vα8.3+ AI4 hymocy es was signi ican ly dec eased in he
NOD.Ln82-AI4, bu no he NOD.Ln16-AI4 sub-s ain (Fig. 1B, C). These di e ences
indica ed a polymo phic gene(s) in luencing numbe s o DP AI4 hymocy es esides wi hin a
5.4 Mb egion on Ch .7 maximally delinea ed by he ma ke s
D7Mi 117
-
D7Mi 225
.
Exp ession le els o he clono ypic Vα8.3 TCR elemen was highe in he NOD.Ln82-AI4
s ock compa ed o NOD-AI4 con ols (Fig. 1D), a ai p e iously obse ed in
NOD.
Ch 7B6
FL-AI4 mice (8). These esul s indica ed a p oximal Ch . 7 egion gene(s)
modula ing TCR exp ession le els maps o he 5.4 Mb egion be ween
D7Mi 117
and
D7Mi 225
.
N kbid is he only di e en ially exp essed p oximal Ch . 7 egion gene in hymocy es
unde going low e sus high le els o dele ion.
Due o di e ing cellula composi ion, we did no eel i app op ia e o u ilize hymii om
NOD-AI4 and NOD.Ln82-AI4 mice o ca y ou ini ial mRNA ansc ip analyses o
iden i y genes possibly con ibu ing o hei po en ially di e ing nega i e selec ion o
diabe ogenic CD8+ T-cells. Ins ead, we easoned a mo e app op ia e pla o m o ca y ou
such analyses would be p o ided by a p e iously p oduced NOD backg ound s ock
ansgenically exp essing a Vα2+Vβ8+ TCR ecognizing he H2-Db class I es ic ed gp33
pep ide (KAVYNFATM) de i ed om Lymphocy ic Cho iomeningi is Vi us (LCMV) (19,
20). The basis o choosing his NOD.LCMV s ock was ha LCMV eac i e CD8+ T-cells
de eloping in he hymus would no be unde any nega i e selec ion p essu e un il such ime
hei cogna e an igen is exogenously in oduced. We gene a ed an addi ional NOD.LCMV
s ock homozygous o he Ln82 congenic in e al. A baseline, p opo ions o DP
P esa e al. Page 8
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip
hymocy es did no di e in NOD.LCMV and NOD.Ln82-LCMV mice (Figu e 2A). This
indica ed a p oximal Ch . 7 egion gene(s) does no con ol he posi i e selec ion o
expansion o DP hymocy es. NOD.LCMV and NOD.Ln82-LCMV mice we e i. . injec ed
wi h 0.5μg o he gp33 o an H2-Db binding con ol pep ide (ASNENMETM). A 24 hou s
pos -injec ion, we assessed he p opo ion o Vα2+ TCR LCMV-speci ic DP hymocy es
p esen in NOD.LCMV and NOD.Ln82-LCMV mice ha had ecei ed he gp33 pep ide
ela i e o con ol coho s. Compa ed o hose ecei ing con ol pep ide, ea men wi h gp33
elici ed a g ea e educ ion o DP hymocy es in NOD.Ln82-LCMV han NOD.LCMV mice
(Fig. 2B). The inding ha he g ea e diminu ion o DP hymocy es in NOD.Ln82-LCMV
han NOD.LCMV mice was only obse ed unde an igen s imula ed condi ions indica es his
pheno ype likely ep esen s a p oximal Ch . 7 egion gene(s) con olled s ain di e ence in
he induc ion o nega i e a he han posi i e selec ion.
The LCMV TCR ansgenic sys em p o ided an oppo uni y o compa e he in luence o he
p oximal Ch . 7 egion o e he ansc ip ional p o ile o hymocy es unde going di e en
le els o an igen-induced nega i e selec ion. Time-cou se expe imen s indica ed ha a 2 and
4 hou s pos -injec ion wi h he cogna e an igen o pep ide con ol he e was no signi ican
educ ion in he numbe o Vα2+ DP hymocy es (Fig. 2C). Howe e , we hypo hesized ha
a 2-hou s pos -an igen injec ion, a di e en ial gene exp ession p o ile may ha e al eady
been induced in NOD.LCMV and NOD.Ln82-LCMV hymocy es con ibu ing o hei
subsequen a ying le els o dele ion seen a 24 hou s. The e o e, we conduc ed mic oa ay
analysis o whole hymus compa ing gp33 o con ol pep ide ea ed NOD.LCMV and
NOD.Ln82-LCMV mice a his ime poin . Based on cu o c i e ia o a q- alue <0.05 and
minimum wo- old s ain a ia ion, 16 and 212 genes we e ound o be di e en ially
exp essed in hymocy es om NOD.LCMV and NOD.Ln82-LCMV mice injec ed wi h he
Db binding con ol o gp33 pep ide espec i ely (Fig. 2D, Supplemen a y Table 1). This
included 12 genes ha we e commonly di e en ially exp essed in hymocy es om
NOD.LCMV and NOD.Ln82-LCMV mice ea ed wi h ei he he con ol o gp33 pep ide
(Fig. 2D, Supplemen a y Table 1). None o he 16 genes di e en ially exp essed in
hymocy es om NOD.LCMV and NOD.Ln82-LCMV mice ea ed wi h he con ol pep ide
mapped o he ea lie de ined 5.4 Mb p oximal Ch . 7 suppo in e al (genomic coo dina es
7:29,926,306–35,345,930). In iguingly, o he 212 genes ma ching ou dual-cu o
di e en ial exp ession c i e ia unde gp33 s imula ed condi ions,
N kbid
was he sole one
mapping wi hin he abo e de ined 5.4 Mb p oximal Ch . 7 suppo in e al (Fig. 2E).
N kbid
was exp essed a a 3.6- old highe le el in hymocy es om gp33 ea ed NOD.LCMV han
NOD.Ln82-LCMV mice (Fig. 2E). Typing o he in agenic s3142493 single nucleo ide
polymo phism (SNP) con i med he Ln82 congenic in e al con ained he B6 o igin
N kbid
allele (Fig. 1A). We also subsequen ly ound by qPCR analyses ha
N kbid
is exp essed a
~2- old highe le els in low cy ome ically so ed equalized numbe s o DP hymocy es
om NOD-AI4 han NOD.Ln82-AI4 mice (Fig. 2F).
N kbid
(aka
IκBNS
) encodes a non-con en ional modula o o he ansc ip ion ac o NF-
κB (17) and i onically, was i s disco e ed as a gene exp essed in hymocy es unde going
nega i e, bu no posi i e selec ion (31). Depending on cell ype and physiological con ex ,
N kbid can s imula e o inhibi NF-κB pa hway ac i i y (17). Low o mode a e le els o NF-
κB ac i i y (p ima ily he p50/p65 he e odime ic o m) epo edly inhibi s hymic nega i e
P esa e al. Page 9
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip
P ep oinsulin-Speci ic CD8 T Cells Is P edominan ly G anule-Media ed Wi h he Po ency
Dependen Upon T-Cell Recep o A idi y. Diabe es 62: 205–213. [PubMed: 22936177]
12. Nejen se S, Howson JM, Walke NM, Szeszko J, Field SF, S e ens HE, Reynolds P, Ha dy M,
King E, Mas e s J, Hulme J, Maie LM, Smy h D, Bailey R, Coope JD, Ribas G, Campbell RD,
Clay on DG, and Todd JA. 2007 Localiza ion o ype 1 diabe es suscep ibili y o he MHC class I
genes HLA-B and HLA-A. Na u e 450: 887–892. [PubMed: 18004301]
13. Noble JA, Valdes AM, Va ney MD, Ca lson JA, Moonsamy P, Fea AL, Lane JA, La an E,
Rappne R, Louey A, Concannon P, Mychaleckyj JC, E lich HA, and . . T. D. G. Conso ium.
2010. HLA Class I and Gene ic Suscep ibili y o Type 1 Diabe es: Resul s F om he Type 1
Diabe es Gene ics Conso ium. Diabe es 59: 2972–2979. [PubMed: 20798335]
14. Chen J, Rei snyde P, Scheuplein F, Scho W, Mileiko sky M, Soodeen-Ka ama h S, Nagy A,
Dosch M, Ellis J, Koch-Nol e F, and Lei e E. 2005 “Agou i NOD”: iden i ica ion o a CBA-
de i ed Idd locus on Ch omosome 7 and i s use o chime a p oduc ion wi h NOD emb yonic s em
cells. Mammalian Genome 16: 775–783. [PubMed: 16261419]
15. Ghosh S, Palme SM, Rod igues NR, Co dell HJ, Hea ne CM, Co nall RJ, P ins JB, McShane P,
La h op GM, Pe e son LB, and e al. 1993 Polygenic con ol o au oimmune diabe es in nonobese
diabe ic mice. Na Gene 4: 404–409. [PubMed: 8401590]
16. McAlee MA, Rei snyde P, Palme SM, P ochazka M, Lo e JM, Copeman JB, Powell EE,
Rod igues NR, P ins J-B, Se eze DV, DeLa a o NH, Wicke LS, Pe e son LB, Scho k NJ, Todd
JA, and Lei e EH. 1995 C osses o NOD Mice Wi h he Rela ed NON S ain: A Polygenic Model
o IDDM. Diabe es 44: 1186–1195. [PubMed: 7556956]
17. Schus e M, Annemann M, Plaza-Si en C, and Schmi z I. 2013 A ypical IkappaB p o eins -
nuclea modula o s o NF-kappaB signaling. Cell Communica ion and Signaling 11: 23. [PubMed:
23578005]
18. G ase RT, DiLo enzo TP, Wang F, Ch is ianson GJ, Chapman HD, Roopenian DC, Na henson SG,
and Se eze DV. 2000 Iden i ica ion o a CD8 T cell ha can independen ly media e au oimmune
diabe es de elopmen in he comple e absence o CD4 T cell helpe unc ions. J Immunol 164:
3913–3918. [PubMed: 10725754]
19. Pi che H, Bu ki K, Lang R, Henga ne H, and Zinke nagel RM. 1989 Tole ance induc ion in
double speci ic T-cell ecep o ansgenic mice a ies wi h an igen. Na u e 342: 559–561.
[PubMed: 2573841]
20. Se eze DV, Johnson EA, Chapman HD, G ase RT, Ma on MP, DiLo enzo TP, Sil ei a P,
Yoshimu a Y, Na henson SG, and Joyce S. 2001 Au o eac i e Diabe ogenic T-Cells in NOD Mice
Can E icien ly Expand F om a G ea ly Reduced P ecu so Pool. Diabe es 50: 1992–2000.
[PubMed: 11522664]
21. Ve dague J, Schmid D, Am ani A, Ande son B, A e ill N, and San ama ia P. 1997 Spon aneous
Au oimmune Diabe es in Monoclonal T Cell Nonobese Diabe ic Mice. The Jou nal o
Expe imen al Medicine 186: 1663–1676. [PubMed: 9362527]
22. Hill JT, Dema es BL, Bisg o e BW, Su Y-C, Smi h M, and Yos HJ. 2014 Poly peak pa se :
Me hod and so wa e o iden i ica ion o unknown indels using sange sequencing o polyme ase
chain eac ion p oduc s. De elopmen al Dynamics 243: 1632–1636. [PubMed: 25160973]
23. Wu H, Ke MK, Cui X, and Chu chill GA. 2003 MAANOVA: A So wa e Package o he
Analysis o Spo ed cDNA Mic oa ay Expe imen s In The Analysis o Gene Exp ession Da a:
Me hods and So wa e Pa migiani G, Ga e ES, I iza y RA, and Zege SL, eds. Sp inge New
Yo k, New Yo k, NY 313–341.
24. S o ey JD 2002 A di ec app oach o alse disco e y a es. Jou nal o he Royal S a is ical Socie y:
Se ies B (S a is ical Me hodology) 64: 479–498.
25. Julius MH, and He zenbe g LA. 1974 ISOLATION OF ANTIGEN-BINDING CELLS FROM
UNPRIMED MICE. Demons a ion o An ibody-Fo ming Cell P ecu so Ac i i y and Co ela ion
be ween P ecu so and Sec e ed An ibody A idi ies 140: 904–920.
26. Liebe man SM, Takaki T, Han B, San ama ia P, Se eze DV, and DiLo enzo TP. 2004 Indi idual
Nonobese Diabe ic Mice Exhibi Unique Pa e ns o CD8+ T Cell Reac i i y o Th ee Isle
An igens, Including he Newly Iden i ied Widely Exp essed Dys ophia Myo onica Kinase. The
Jou nal o Immunology 173: 6727–6734. [PubMed: 15557165]
P esa e al. Page 16
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip

27. Annemann M, Wang Z, Plaza-Si en C, Glauben R, Schus e M, Ewald Sande F, Mama eli P,
Kühl AA, Siegmund B, Lochne M, and Schmi z I. 2015 IκBNS Regula es Mu ine Th17
Di e en ia ion du ing Gu In lamma ion and In ec ion. The Jou nal o Immunology 194: 2888–
2898. [PubMed: 25694610]
28. Taylo SC, and Posch A. 2014 The Design o a Quan i a i e Wes e n Blo Expe imen . BioMed
Resea ch In e na ional 2014: 361590. [PubMed: 24738055]
29. Fosang AJ, and Colb an RJ. 2015 T anspa ency Is he Key o Quali y. Jou nal o Biological
Chemis y 290: 29692–29694. [PubMed: 26657753]
30. Lamon D, Mukhe jee G, Kuma PR, Saman a D, McPhee CG, Kay TWH, Almo SC, DiLo enzo
TP, and Se eze DV. 2014 Compensa o y Mechanisms Allow Unde sized Ancho -De icien Class I
MHC Ligands To Media e Pa hogenic Au o eac i e T Cell Responses. The Jou nal o Immunology
193: 2135–2146. [PubMed: 25063871]
31. Fio ini E, Schmi z I, Ma issen WE, Osbo n SL, Touma M, Sasada T, Reche PA, Tibaldi EV,
Hussey RE, K uisbeek AM, Reinhe z EL, and Clay on LK. 2002 Pep ide-Induced Nega i e
Selec ion o Thymocy es Ac i a es T ansc ip ion o an NF-kB Inhibi o . Molecula cell 9: 637–
648. [PubMed: 11931770]
32. Jimi E, S ickland I, Voll RE, Long M, and Ghosh S. 2008 Di e en ial Role o he T ansc ip ion
Fac o NF-κB in Selec ion and Su i al o CD4+ and CD8+ Thymocy es. Immuni y 29: 523–537.
[PubMed: 18957265]
33. Aken BL, Ayling S, Ba ell D, Cla ke L, Cu wen V, Fai ley S, Fe nandez Bane J, Billis K, Ga cía
Gi ón C, Hou lie T, Howe K, Kähä i A, Kokocinski F, Ma in FJ, Mu phy DN, Nag R, Ru ie M,
Schus e M, Tang YA, Vogel J-H, Whi e S, Zadissa A, Flicek P, and Sea le SMJ. 2016. The
Ensembl gene anno a ion sys em. Da abase 2016: baw093-baw093.
34. Touma M, An onini V, Kuma M, Osbo n SL, Bobenchik AM, Keskin DB, Connolly JE, G usby
MJ, Reinhe z EL, and Clay on LK. 2007 Func ional Role o IκBNS in T Cell Cy okine
Regula ion As Re ealed by Ta ge ed Gene Dis up ion. The Jou nal o Immunology 179: 1681–
1692. [PubMed: 17641034]
35. Schus e M, Glauben R, Plaza-Si en C, Sch eibe L, Annemann M, Floess S, Kühl Anja A.,
Clay on Linda K., Spa wasse T, Schulze-Os ho K, P e e K, Huehn J, Siegmund B, and
Schmi z I. 2012 IκBNS P o ein Media es Regula o y T Cell De elopmen ia Induc ion o he
Foxp3 T ansc ip ion Fac o . Immuni y 37: 998–1008. [PubMed: 23200824]
36. Am ani A, Ve dague J, Se a P, Ta u o S, Tan R, and San ama ia P. 2000 P og ession o
au oimmune diabe es d i en by a idi y ma u a ion o a T-cell popula ion. Na u e 406: 739–742.
[PubMed: 10963600]
37. Kuwa a H, Ma sumo o M, A a ashi K, Mo ishi a H, Hi o ani T, Koga R, and Takeda K. 2006
IκBNS Inhibi s Induc ion o a Subse o Toll-like Recep o -Dependen Genes and Limi s
In lamma ion. Immuni y 24: 41–51. [PubMed: 16413922]
38. Touma M, Keskin DB, Shi oki F, Sai o I, Koyasu S, Reinhe z EL, and Clay on LK. 2011 Impai ed
B Cell De elopmen and Func ion in he Absence o IκBNS. The Jou nal o Immunology 187:
3942–3952. [PubMed: 21900180]
39. Ra iu JJ, Racine JJ, Hasham MG, Wang Q, B anca JA, Chapman HD, Zhu J, Donghia N, Philip V,
Scho WH, Wasse all C, A kinson MA, Mills KD, Lee h CM, and Se eze DV. 2017 Gene ic and
Small Molecule Dis up ion o he AID/RAD51 Axis Simila ly P o ec s Nonobese Diabe ic Mice
om Type 1 Diabe es h ough Expansion o Regula o y B Lymphocy es. The Jou nal o
Immunology.
40. Wang Q, Racine JJ, Ra iu JJ, Wang S, E inge R, Wasse all C, A kinson MA, and Se eze DV.
2017 T ansien BAFF Blockade Inhibi s Type 1 Diabe es De elopmen in Nonobese Diabe ic Mice
by En iching Immuno egula o y B Lymphocy es Sensi i e o Dele ion by An i-CD20 Co he apy.
The Jou nal o Immunology 199: 3757–3770. [PubMed: 29055002]
41. Ramak ishnan P, Yui MA, Tomalka JA, Majumda D, Pa ameswa an R, and Bal imo e D. 2016
De iciency o Nuclea Fac o -κB c-Rel Accele a es he De elopmen o Au oimmune Diabe es in
NOD Mice. Diabe es 65: 2367–2379. [PubMed: 27217485]
42. A nold CN, Pi ie E, Doseno ic P, McIne ney GM, Xia Y, Wang N, Li X, Siggs OM, Ka lsson
Hedes am GB, and Beu le B. 2012 A o wa d gene ic sc een e eals oles o N kbid, Zeb1, and
P esa e al. Page 17
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip
Ru bl2 in humo al immuni y. P oceedings o he Na ional Academy o Sciences 109: 12286–
12293.
43. Pede sen GK, Àdo i M, Khoenkhoen S, Doseno ic P, Beu le B, and Ka lsson Hedes am GB. 2014
B-1a ansi ional cells a e pheno ypically dis inc and a e lacking in mice de icien in IκBNS.
P oceedings o he Na ional Academy o Sciences 111: E4119–E4126.
44. Pede sen GK, Ádo i M, S a k j. M., Khoenkhoen S, A nold C, Beu le B, and Ka lsson Hedes am
GB. 2016 He e ozygous mu a ion in IκBNS leads o educed le els o na u al IgM an ibodies and
impai ed esponses o T-independen ype 2 an igens. F on ie s in Immunology 7.
45. Miu a M, Hasegawa N, Noguchi M, Sugimo o K, and Touma M. 2016 The a ypical IκB p o ein
IκBNS is impo an o Toll-like ecep o -induced in e leukin-10 p oduc ion in B cells.
Immunology 147: 453–463. [PubMed: 26749055]
P esa e al. Page 18
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip
Figu e 1: Mapping a p oximal Ch . 7 egion gene(s) con olling he hymic nega i e selec ion
e iciency o diabe ogenic AI4 CD8+ T-cells o a 5.4 Mb egion.
(A) Schema ic diag am o B6 de i ed Ch .7 congenic egions in NOD-AI4 mouse s ains.
Ma ke posi ions a e indica ed in Mb based on genome assembly elease GRCm38.p5 (33)
(B) Flow cy ome ic ga ing s a egy o enume a e AI4 DP hymocy es. Thymocy es om 5-
week-old emale NOD-AI4 (n=32), NOD.Ln82-AI4 (n=20) and NOD.Ln16-AI4 (n=14)
we e s ained wi h CD4, CD8 and Vα8.3 speci ic an ibodies. (C) Numbe s o li e (PI
nega i e) CD4+CD8+ (DP) AI4 (Vα8.3+) hymocy es in he indica ed s ains. (D) Mean
luo escence in ensi y (MFI) o s aining o DP hymocy es om he indica ed s ains by he
Vα8.3-PE monoclonal an ibody B21.14. All alues ep esen mean±SEM o mo e han 3
independen expe imen s, s a is ical signi icance was de e mined by Log- ank (Man el-Cox)
- es (ns p>0.05, ** p<0.001, **** p<0.00001).
P esa e al. Page 19
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip
Figu e 2: N kbid is he only di e en ially exp essed p oximal Ch . 7 egion gene in diabe ogenic
CD8+ T-cells unde going low e sus high le els o hymic dele ion.
(A). Simila p opo ions o DP hymocy es in un ea ed NOD.LCMV and NOD.Ln82-
LCMV mice. (B) Ei he 0.5 μg o gp33 (cogna e an igen) o Db binding con ol pep ide we e
i. . injec ed in o i e-week-old emale NOD.LCMV (n=18 o con ols and n=6 o gp33
ea ed) o NOD.Ln82-LCMV (n=22 o con ols and n=6 o gp33 ea ed) mice.
Remaining DP Vα2+ hymocy es we e quan i ied a 24 hou s pos -injec ion in each g oup.
The e iciency o dele ion was calcula ed as a a io o he numbe DP Vα2+ hymocy es
emaining in gp33 ea ed mice o he a e age numbe s in con ols and exp essed as a
pe cen age (x gp33 DP Va2 +
yDb DP Va2 + × 100). S a is ical signi icance was de e mined by Log- ank
(Man el-cox) - es (ns p>0.05, * p<0.05). (C) A 2–4 hou s pos -i. . injec ion wi h gp33 o
Db binding con ol pep ide numbe s o DP Vα2+ hymocy es do no di e in i e-week-old
NOD.LCMV (n=18 o con ols and n=7 o gp33 ea ed) o NOD.Ln82-LCMV (n=22 o
con ols and n=6 o gp33 ea ed) emale mice. Calcula ion o dele ion e iciencies and
s a is ical compa isons as in panel B. (D) As assessed by mic oa ay analyses numbe o
genes de e en ially exp essed decla ed a a alse disco e y a e (q- alue) <0.05 and a old
change ≥2 in hymocy es om NOD.LCMV and NOD.Ln82-LCMV mice 2 hou s a e i. .
injec ion wi h Db con ol o gp33 pep ide (n=4–5 biological eplica es pe s ain and
ea men ype). (E) Volcano plo showing mic oa ay assessed ela ionship be ween old
change and le el o signi icance (F1 pe mu a ion p- alue) in hymic gene exp ession p o iles
in NOD.LCMV (n=5) and NOD.Ln82-LCMV (n=3) mice 2 hou s a e i. . injec ion wi h
gp33. Da a poin s o he 212 p obe se s wi h a old-change > 2 (log2=1) a e shown le o
igh o he dashed e ical lines. The a ow depic s di e en ial
N kbid
exp ession. The o he
211 genes a e iden i ied in Supplemen a y Table 1. (F) To al RNA was ex ac ed om 2×105
DP Vß8+ hymocy es so ed om 6-week-old emale NOD-AI4 and NOD.Ln82-AI4 mice.
N kbid
mRNA ela i e quan i ica ion was done using he ddC me hod using
Gapdh
as an
endogenous con ol and NOD.Ln82-AI4 mice as e e ence, mean ela i e old change (RFC)
± SEM is indica ed. * p<0.05 unpai ed - es .
P esa e al. Page 20
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip
Figu e 3: Ta ge ing N kbid by CRISPR/Cas9 in NOD mice.
(A) Rep esen a ion o mouse
N kbid
gene showing he ansla ion s a si e a he end o
exon 3 and he sgRNA (unde lined) designed o a ge his egion by CRISPR/Cas9. The
PAM sequence is indica ed in lowe case. (B) Schema ic o he esul ing a ge ed mu a ion
consis ing o an 8-nucleo ide dele ion and an inse ion o 154 nucleo ides o igina ed by copy
and in e sion o a segmen om
N kbid
exon 2 (Fo mo e de ail, see Supplemen a y Figu e
1). (C) N kbid p o ein abundance was assessed in o al hymus lysa es om he indica ed
s ains by Simple Wes au oma ed wes e n blo analysis. B6.
N kbid
−/− hymus samples we e
p o ided by D . Ingo Schmi z. N kbid was de ec ed as a 55 kDa band. The di e ence in
molecula weigh compa ed o he heo e ical 35 kDa, could be due o pos ansla ional
modi ica ions. Lowe panel indica es he co esponding o al p o ein s aining o each
sample. Righ panel shows he N kbid-no malized a ea o NOD (n=6), NOD-
N kbid
+/−
(n=9), NOD-
N kbid
−/− (n=8), B6 (n=6), and B6.
N kbid
−/− (n=4) hymus samples. N kbid-
peak a ea was no malized as ollows: (Sample N kbid a ea) × B6 o al p o ein a ea
(Sample o al p o ein a ea). Ba s
ep esen he mean ± SEM. ns p>0.05, ** p<0.01, Mann-Whi ney analysis.
P esa e al. Page 21
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip

Figu e 4: Gene ic a enua ion o N kbid exp ession con ols he numbe s o AI4 DP hymocy es.
(A) Wild ype
N kbid
gene exp ession was analyzed by qPCR o o al hymus RNA om 6-
week-old B6 (n=2), NOD-AI4 (n=4), NOD-AI4-
N kbid+/−
(n=4) and NOD-AI4-
N kbid−/−
(n=3) mice.
Gapdh
was used as endogenous con ol and B6 as he e e ence g oup. G aph
shows he mean ± SEM o ela i e old change o
N kbid
exp ession no malized o
Gapdh
.
S a is ical signi icance was de e mined by one-way ANOVA and Bon e oni’s mul iple
compa ison es . The p- alue o samples compa ed o he e e ence g oup is indica ed. No e:
We did no use B6.
H2g7
-AI4 mice as a e e ence in his s udy because he numbe s o
hymocy es is ex emely low and would no p o ide a ai compa ison o gene exp ession
analysis. (B) Illus a ion o ga ing o enume a e DP and CD8+ SP hymocy es s aining wi h
he AI4 TCR speci ic e ame . Depic ed p o iles a e o hymocy es om an NOD-AI4
con ol. (C) Numbe s o AI4 TCR exp essing DP and CD8+ SP hymocy es om 6-week-old
NOD-AI4, NOD-AI4-
N kbid+/−
, and NOD-AI4-
N kbid−/−
emale mice. Da a ep esen he
mean ± SEM o DP o CD8+ SP hymocy es s aining posi i e wi h Te -AI4. (D) Geome ic
P esa e al. Page 22
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip
MFI o Te -AI4 s aining o he same DP and CD8+ SP hymocy es e alua ed in panel B. (E)
Quan i ica ion o mRNA old change exp ession o NF-κB amily membe s in so ed DP
NOD-AI4 s NOD-AI4-
N kbid−/−
hymocy es. Indi idual gene exp ession was i s
no malized o
Gapdh
, and hen old change in compa ison o exp ession in NOD-AI4 was
calcula ed. Da a is combined om 13 emale mice, 5–6 weeks o age combined om wo
coho s. (F) NF-κB a ge gene exp ession analysis was pe o med wi h he same samples as
in (E). Ba s ep esen he mean ± SEM. S a is ical signi icance o A-D was analyzed by 2-
way ANOVA and Bon e oni’s mul iple compa ison es . S a is ical signi icance o E and F
we e analyzed by Mann-Whi ney. ns p>0.05, * p<0.05, ** p<0.01, **** p<0.0001.
P esa e al. Page 23
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip
Figu e 5: N kbid exp ession le els modula e hymic nega i e selec ion e iciency o diabe ogenic
NY8.3 clono ypic CD8+ T-cells.
(A) Illus a ion o ga ing o enume a e DP and CD8+ hymocy es s aining wi h he NY8.3
TCR speci ic Te -NRPV7. Depic ed p o iles a e o hymocy es om an NOD.NY8.3
con ol. (B) Numbe s o NY8.3 TCR exp essing DP and CD8+ hymocy es om 6-week-old
emale NOD.NY8.3 and NOD.NY8.3-
N kbid
+/− he e ozygous mice. Da a ep esen he
mean±SEM o DP o CD8+ SP hymocy es s aining posi i e wi h he NY8.3 speci ic Te -
NRPV7 eagen . (C) Numbe s o NY8.3 TCR exp essing DP and CD8+ SP hymocy es om
4-week-old emale NOD.NY8.3 and NOD.NY8.3-
N kbid
−/− homozygous mice. Ba s
ep esen he mean±SEM o 3 independen expe imen s. (D) Geome ic MFI o Te -NRPV7
s aining o DP and CD8+ SP hymocy es om NOD.NY8.3, NOD.NY8.3-
N kbid
+/−, and
NOD.NY8.3-
N kbid
−/− mice. Ba s ep esen mean±SEM. S a is ical signi icance was
analyzed by 2-way ANOVA and Bon e oni’s mul iple compa ison es . ns p>0.05, **
p<0.01.
P esa e al. Page 24
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip
Figu e 6: N kbid de iciency accele a es T1D onse in NOD mice associa ed wi h impai ed
de elopmen o egula o y lymphocy e popula ions.
(A) T1D de elopmen was analyzed in a coho o NOD, NOD-
N kbid+/−
and NOD-
N kbid
−/−
emale mice. Su i al cu e compa isons we e analyzed h ough Log- ank (Man el-Cox)
es . (B) Numbe s o splenic AI4 T-cells in 6-week-old NOD (n=31), NOD-
N kbid+/−
(n=12) and NOD-
N kbid−/−
(n=16) emale mice. (C) Splenocy es om 7-week-old emale
NOD (n=3) and NOD-
N kbid−/−(
n=5) mice we e labeled wi h an ibody o CD19, CD21
and CD23 and analyzed by low cy ome y. Single , DAPI- CD19+ cells we e iden i ied as
o al B-cells. Based on exp ession o CD23 and CD21, B-cells we e classi ied as ollicula
(FO: CD23+CD21lo), ma ginal zone (MZ: CD23-CD21hi) ansi ional 1 (T1: CD23-CD21-)
and ansi ional 2 (T2: CD23in CD21hi) (D) Pe i oneal ca i y B-cells we e isola ed by
la age wi h cold HBSS om he same g oup o mice desc ibed in (C), and s ained o CD19,
B220, CD5 and analyzed by low cy ome y. To al B-cells (Single , DAPI- CD19+) we e
classi ied based in he exp ession o B220 and CD5 as B1-a (CD5+ B220in /lo) and B1-b
(CD5- B220lo) (E). The p oli e a ion capaci y o pu i ied o al T-cells om NOD and NOD-
N kbid−/−
mice was assessed as desc ibed in me hods sec ion. B ie ly, MACS pu i ied o al
T-cells we e labeled wi h cell ace eFluo 670 and s imula ed
in i o
wi h inc easing
amoun s o an i-CD3 monoclonal an ibody. A e 48 hou s o cul u e he cells we e labeled
wi h an ibody o CD4 and CD8. P oli e a ion in CD4+ and CD8+ T-cell subse s was
analyzed by low cy ome ic analyses o cell ace eFluo 670 dilu ion. P oli e a ion index
was de e mined using he Flow Jo cell p oli e a ion ool. Da a ep esen he mean±SEM o 3
echnical eplica es. (F) Cell cul u e supe na an s o pu i ied T-cells used in he p oli e a ion
P esa e al. Page 25
J Immunol
. Au ho manusc ip ; a ailable in PMC 2019 Oc obe 01.
Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip