ORIGINAL RESEARCH
published: 05 No embe 2018
doi: 10.3389/ immu.2018.02540
F on ie s in Immunology | www. on ie sin.o g 1No embe 2018 | Volume 9 | A icle 2540
Edi ed by:
Remy Bosselu ,
Na ional Cance Ins i u e (NCI),
Uni ed S a es
Re iewed by:
Sid P. Ke ka ,
Boeh inge Ingelheim, Uni ed S a es
Elizabe h Wohl e ,
Uni e si y a Bu alo, Uni ed S a es
*Co espondence:
João F. Lace da
[email p o ec ed]
†P esen Add ess:
Eka e ina Minskaia,
Ad anced The apeu ics, Sco ish
Na ional Blood T ans usion Se ice,
The Jack Copland Cen e, Edinbu gh,
Uni ed Kingdom
Ana I. S. Viei a,
Flow Cy ome y Pla o m,
Champalimaud Founda ion, Lisbon,
Po ugal
‡These au ho s sha e senio
au ho ship
Special y sec ion:
This a icle was submi ed o
T Cell Biology,
a sec ion o he jou nal
F on ie s in Immunology
Recei ed: 17 May 2018
Accep ed: 15 Oc obe 2018
Published: 05 No embe 2018
Ci a ion:
Minskaia E, Sa ai a BC, Soa es MMV,
Aze edo RI, Ribei o RM, Kuma SD,
Viei a AIS and Lace da JF (2018)
Molecula Ma ke s Dis inguishing
T Cell Sub ypes Wi h TSDR
S and-Bias Me hyla ion.
F on . Immunol. 9:2540.
doi: 10.3389/ immu.2018.02540
Molecula Ma ke s Dis inguishing
T Cell Sub ypes Wi h TSDR
S and-Bias Me hyla ion
Eka e ina Minskaia1†‡, Ba ba a C. Sa ai a1, Ma ia M. V. Soa es1, Ri a I. Aze edo1,
Ruy M. Ribei o2, Saumya D. Kuma 1, Ana I. S. Viei a1† and João F. Lace da1*‡
1Faculdade de Medicina da Uni e sidade de Lisboa, Ins i u o de Medicina Molecula –João Lobo An unes, Lisbon, Po ugal,
2Depa men o de Bioma emá ica, Faculdade de Medicina da Uni e sidade de Lisboa, Lisbon, Po ugal
Human egula o y CD4+CD25+FOXP3+T cells (T eg) play impo an oles in he
main enance o sel - ole ance and immune homeos asis in a ious disease se ings and
a e also in ol ed in he supp ession o e ec i e immune esponses. These cells a e
he e ogeneous in pheno ype and unc ion, and he abili y o eliably dis inguish be ween
a ious FOXP3-exp essing subpopula ions can a ec he de elopmen o success ul
he apies. This s udy demons a es ha hypome hyla ed CpG si es, p esen in ou
egions o he FOXP3 locus, CAMTA1 and FUT7 gene egions, can be used o dis inguish
se e al subse s o T eg om con en ional CD4+T lymphocy es (Tcon) in dono s o bo h
gende s. We desc ibe a p e iously un epo ed s and-bias hemime hyla ion pa e n in
FOXP3 p omo e and TSDR in dono s o bo h gende s, wi h he coding s and being
deme hyla ed wi hin p omo e and me hyla ed wi hin TSDR in all CD4+lymphocy e
sub ypes, whe eas he empla e s and ollows he p e iously desc ibed pa e n o
me hyla ion wi h bo h egions being mo e deme hyla ed in T eg sub ypes and mos ly
me hyla ed in Tcon. This s and-speci ic app oach wi hin he TSDR may p o e o be
ins umen al in co ec ly de ining T eg subse s in heal h and in disease.
Keywo ds: FOXP3,CAMTA,FUT7, egula o y T lymphocy es, epigene ics,s and-bias me hyla ion
INTRODUCTION
Regula o y T cells (T eg) play c ucial oles in he main enance o sel - ole ance and immune
homeos asis in diseases such as alle gy and au oimmune diso de s (AID). These cells a e also
in ol ed in he supp ession o e ec i e immune esponses agains in ading pa hogens and
au ologous cance ous cells (1,2). Human CD4+CD25+FOXP3+T cells a e he e ogeneous in
pheno ype and unc ion and include es ing ( T eg) and ac i a ed (aT eg) T eg cells, as well as
non-supp essi e T eg-like cells (3,4). Despi e his, o many yea s FOXP3 exp ession alone has
been used as a speci ic T eg ma ke un il i was deemed insu icien o iden i ica ion o supp essi e
T eg cells. The abili y o eliably dis inguish be ween a ious FOXP3-exp essing subpopula ions
and o unde s and hei oles in immunological diseases, cance and in ec ions can os e he apies
ha ei he aim o boos he supp essed immune esponses o o dampen he abno mally acu e
immune esponses.
Towa d his goal, Miya a e al. (5) used he combina ion o CD25 and CD45RA o
isola ing h ee human T eg subse s: FOXP3lowCD45RA+CD25++ supp essi e in i o T eg,
FOXP3hiCD45RA−CD25+++ e ec o (eT eg) cells and cy okine-sec e ing non-supp essi e
Minskaia e al. Epigene ics T Cell Sub ypes
FOXP3lowCD45RA−CD25++ T cells. La e , CD15s (sialyl
Lewis x) was iden i ied as a bioma ke o mos supp essi e
FOXP3high eT eg cells (6). A combina ion o CD15s and
CD45RA was ins umen al in he isola ion o dis inc
CD4+CD127lowCD25+FOXP3+T cell sub ypes: naï e
CD45RA+CD15s−T eg, highly supp essi e CD45RA−CD15s+
eT eg and a non-supp essi e CD45RA−CD15s−subse .
Toge he wi h his one ace yla ion and non-coding RNAs,
DNA me hyla ion can ei he s ably o empo a ily al e
gene exp ession depending on he immedia e physiological
equi emen s o he o ganism. Se e al egula o y egions
on FOXP3 locus a e e y impo an playe s in he T eg-
speci ic epigenome: wo conse ed non-coding sequences (CNS
1 and 3) a e in ol ed in his one ace yla ion while h ee
o he egions - ups eam enhance , p oximal p omo e and
CNS 2 (known as FOXP3 TSDR) con ibu e o FOXP3
exp ession ia deme hyla ion and we e p oposed as addi ional
molecula ma ke s ha can help dis inguish T eg om
con en ional T lymphocy es (Tcon), as well as di e en
T eg ma u a ion s ages (7–9). A he same ime, changes
in T cell DNA me hyla ion pa e ns ha e been epo ed in
diseases such as alle gies, mul iple scle osis and heuma oid
a h i is (10,11). Howe e , as FOXP3 gene is encoded on
Xp11.23, mos s udies op ed o use male dono s in o de
o a oid he a i ac s o he inac i a ion o X ch omosome
(Xi). The e o e, p ecise egula ion o FOXP3 exp ession in
emale dono s emains somewha o an enigma—ye emales
comp ise he majo i y o pa ien s wi h AID and show
a s onge esponse o in ec ions han males. CAMTA1,
encoded on ch omosome 1q3.6 and no a ec ed by Xi,
was able o di e en ia e CD25highCD45RA−CD4+T eg om
CD25−CD45RA+CD4+Tcon (7) and, he e o e, has he
po en ial o dis inguish T cell subse s in bo h male and emale
dono s.
In his s udy, we cha ac e ized he epigene ic p o ile o
hema opoie ic s em cells (CD34+) and ou popula ions o
CD4+CD25+T cells: CD45RA+CD15s−FOXP3low (naï e
nT egs), CD45RA−15s−FOXP3low (non-supp essi e T eg-like
cells), CD45RA−CD15s+FOXP3high (eT egs) and Tcons isola ed
om pe iphe al blood o heal hy male and emale dono s. As
α(1,3)- ucosyl ans e ase 7 (FUT7) media es syn hesis o CD15s
exp essed in bo h eT egs and CD34+cells (12), FUT7 p omo e
was expec ed o be deme hyla ed in hese cell popula ions o allow
o p o ein exp ession. Toge he wi h CAMTA1 in onic egion 3,
FUT7 p omo e was es ed o i s po en ial o ac as an addi ional
and/o al e na i e o FOXP3 molecula ma ke . Th ee p e iously
desc ibed egions on FOXP3 locus: ups eam enhance , p oximal
p omo e and TSDR (T eg-speci ic deme hyla ed egion), we e
also s udied oge he wi h he ou h egion, ha we now e m
p eTSDR. As DNA me hyla ion was shown o a y among
indi iduals and e en be ween wins (13,14), we a emp ed o
cha ac e ize epigene ic changes in all six gene egions om
he i e cell popula ions o each dono in o de o ob ain
comp ehensi e in o ma ion speci ic o each indi idual. Using
bisulphi e con e sion o genomic DNA (gDNA) ollowed by
sequencing o indi idual clones was ins umen al in deciphe ing
he me hyla ion s a us o indi idual CpG posi ions and he
in ica e pa e ns con olling gene exp ession in CD34+cells and
T lymphocy e subse s.
MATERIALS AND METHODS
Isola ion o Human PBMCs and Flow
Cy ome y
Pe iphe al blood samples we e ob ained om young heal hy male
(M1-6) and emale (F1-5) olun ee s. None o he dono s had
known au oimmune o gene ic condi ions.
Pe iphe al blood mononuclea cells (PBMCs) we e p epa ed
by Ficoll g adien cen i uga ion (15). CD34+cells (dono s
M4-6 and F1-5) we e i s en iched using he EasySepTM
Human CD34 Posi i e Selec ion ki (STEMCELL Technologies)
ollowing he manu ac u e ’s ins uc ions. In o de o inc ease
he pu i y o he magne ically isola ed CD34+ ac ion, he
cells we e u he s ained wi h CD34 FITC (Mil enyi Bio ec)
and so ed by luo escen ac i a ed cell so ing (FACS) on a
BD FACSA iaIII. Tcon and T eg subpopula ions we e pu i ied
om he nega i e ac ion ob ained om he EasySepTM
CD34 selec ion p o ocol as ollows: cells we e incuba ed o
25 min a oom empe a u e in PBS (2% human se um)
wi h p e- i a ed amoun s o he ollowing an ibodies: an i-
hCD3 (-Pe CP, clone OKT3, eBioscience), an i-hCD4 (-APC,
clone RPA-T4, eBioscience), an i-hCD45RA (-FITC, Mil enyl
Bio ec), an i-hCD25 (-Pe-Cy7, BD Biosciences), an i-hCD127
(-APCe780, clone eBioRDR5, eBioscience), an i-hCD15s (-
PE, BD Biosciences). Cells we e hen washed and so ed on
a BD FACSA iaIII. Cells ob ained om he EasySep CD34
nega i e ac ion we e u he used o in acellula s aining
o FOXP3. Following he su ace s aining using he same
an ibody combina ion as desc ibed abo e o cell so ing, cells
we e s ained wi h an i-hFOXP3 (eFluo 450, clone PCH101,
eBioscience) using he FOXP3 S aining Bu e Se (e-Bioscience)
acco ding o he manu ac u e ’s ins uc ions. Da a was acqui ed
on he BD FACSA iaIIu. Fo analysis o CD34+cells, whole
blood samples we e su ace s ained o 20 min a oom
empe a u e wi h he same an ibodies as abo e excep o an i-
hCD4 (Pe CPCy5.5, clone OKT4, eBioscience), an i-hCD45RA
(APC, clone T6D11, Mil enyl Bio ec), an i-hCD3 (BV510,
cloneUCHT1, BD Ho izon). In acellula s aining o FOXP3
(eFluo 450, clone PCH101, eBioscience) was pe o med using
RBC lysis, ixa ion and pe meabiliza ion eagen s (eBioscience),
acco ding o he manu ac u e ’s ins uc ions. Samples we e
acqui ed on a BD LSR Fo essa low cy ome e (BD Biosciences).
Da a we e analyzed using FlowJo R
LLC.
Isola ion o Genomic DNA
gDNA om CD34+cells and ou popula ions o CD4+T
cells we e isola ed by he Quick-gDNA MiniP ep ki (Zymo
Resea ch) ollowing manu ac u e ’s ins uc ions. B ie ly, cells
we e esuspended in 100 µl o PBS hen lysed in 400 µl
o Genomic Lysis Bu e o 10 min. Sample was loaded in o
p o ided Zymo-Spin column and cen i uged a 10,000 g o
1 min. The column was hen washed in wo s eps: (i) 200 µl o
DNA P e-Wash Bu e and (ii) 500 µl o gDNA Wash Bu e .
F on ie s in Immunology | www. on ie sin.o g 2No embe 2018 | Volume 9 | A icle 2540
Minskaia e al. Epigene ics T Cell Sub ypes
gDNA was elu ed om he column in 45 µl o DNA Elu ion
Bu e .
Bisulphi e T ea men and Me hyla ion
Analysis
De e mina ion o me hyla ion s a us o indi idual CpG si es
ac oss en i e amplicons can be achie ed by BS ea men
o gDNA ollowed by sequencing o indi idual clones, each
ep esen ing one DNA molecule om one cell in a gi en
cell popula ion. The ac ha BS ea men con e s all non-
me hyla ed cy osines (C) in o u acils (U) while me hyla ed Cs
(mC) emain unchanged means ha he wo DNA s ands in
BS DNA a e no longe complemen a y. Only one s and o
BS DNA is ampli ied by each p ime se , wi h he e e se
p ime binding he chosen a ge s and and he esul ing
ampli ied s and se ing as a empla e o he o wa d p ime .
In addi ion o s and-speci ici y, p ime s a e designed in such
a way ha biased ampli ica ion o non-me hyla ed s ands is
a oided and only BS DNA is ampli ied. To a oid any po en ial
bias a ec ing he downs eam sequencing esul s, gDNA om he
i e cell popula ions: CD34+, CD45RA+15s−, CD45RA−15s−,
CD45RA−15s+, and Tcons, was isola ed and BS ea ed a he
same ime ollowed by BS PCR and cloning o he six gene
egions.
Con e sion o gDNA om he i e cell popula ions was
pe o med by EZ DNA Me hyla ion Ligh ning Ki (Zymo
Resea ch) which esul s in o e 99.5% C o U con e sion o
non-me hyla ed esidues while o e 99.5% o mC a e p o ec ed.
B ie ly, 130 µl o Ligh ning Con e sion Reagen was added o
20 µl o gDNA and he samples we e incuba ed a 95◦C o
8 min ollowed by 54◦C o 60 min. DNA samples we e loaded
in o Zymo-Spin IC Columns con aining 600 µl o M-Binding
Bu e , mixed well and cen i uged a 10,000 g o 30 s (as all
o he cen i uga ion s eps). Following he i s wash s ep wi h
100 µl o M-Wash Bu e , he samples we e incuba ed wi h 200
µl o M-Desulphona ion Bu e a oom empe a u e o 15 min.
Columns we e washed wice wi h 200 µl o M-Wash Bu e .
Con e ed BS- ea ed DNA was elu ed in 12 µl o M-Elu ion
bu e and immedia ely used o PCR analyses. Remaining BS-
ea ed DNA was s o ed a −20◦C o la e use.
The six gene egions we e ampli ied using non-
me hyla ion-, BS- ea ed DNA-, coding s and-speci ic p ime s
(Supplemen a y Table 1). Thi y PCR eac ions o he six gene
egions we e pe o med in pa allel con aining 1.5 µl o BS DNA
and Phusion U Ho S a PCR mix (The mo Fishe Scien i ic)
in a o al olume o 25 µl. A e ini ial dena u a ion a 98◦C
o 30 s, ampli ica ion consis ed o 45 cycles a 98◦C o 10 s,
58-60◦C o 20 s, and 72◦C o 45 s. The PCR p oduc s ob ained
we e gel pu i ied using NZYGelpu e ki (NZY ech) ollowing he
manu ac u e ’s ins uc ions and cloned in o he pGEM-T Easy
ec o (P omega) ia NcoI and NsiI es ic ion si es o ensu e
di ec ional cloning. Plasmid DNAs om 22–24 clones we e
isola ed using plasmid NZYMinip ep ki (NZY ech) and 20–22
posi i e clones (unless o he wise s a ed), con i med by es ic ion
diges ion wi h he abo e-men ioned enzymes, we e sequenced
using e e se SP6 p ime : 5′-GTGACACTATAGAATACTC-3′
(NZY sequencing and S ab ida). Sequences (AB1 iles con aining
ch oma og ams) we e aligned o each gene egion’s e e ence
sequence using SeqMan so wa e (DNA S a Lase gene 8). All
non-me hyla ed cy osines we e iden i ied by he p esence o a
T nucleo ide (n ) in BS-con e ed sequences while mCs we e
iden i ied by he p esence o a C n . E iciency o bisulphi e
con e sion was con i med by con e sion o non-CpG Cs o
Ts. The pe cen age o me hyla ion in each CpG posi ion was
de e mined by de ining he p opo ion o mCs in he o al o
20 (2 mCs ou o 20=10% me hyla ion). As he six gene egions
we e ampli ied om he same BS DNA empla e, he me hyla ion
di e ences e lec he a e age me hyla ion s a us o he cell
popula ion.
S a is ical Analysis
To analyze he di e ences in me hyla ion be ween Tcon and he
h ee T eg popula ions (CD45RA+CD15s−, CD45RA−CD15s−,
and CD45RA−CD15s+) o he six genes s udied, we used
a mixed-e ec s app oach, aking in o conside a ion he pai ed
na u e o he da a. Fo each gene, he di e ence in me hyla ion
alues be ween Tcon and each o he T eg popula ions was
calcula ed and hen his di e ence was es ed o signi ican
di e ences om ze o (indica ing a di e ence in me hyla ion
be ween he wo popula ions being compa ed), wi h indi idual
and CpG posi ion conside ed as andom e ec s. This analysis
was done using he unc ion lme om he package nlme o
R (h ps://c an. -p ojec .o g/). Since in each case we used he
me hyla ion o Tcon o h ee es s, we co ec ed o mul iple
compa isons and only conside ed as signi ican hose esul s wi h
p<0.016 (equi alen o α=0.05/3). We used he same app oach
o compa e he me hyla ion o he op and bo om s ands.
Hie a chical Clus e ing
Hie a chical Agglome a i e Clus e ing was pe o med using
Euclidean dis ance ma ix and comple e linkage algo i hm.
Dis ance ma ix was calcula ed on a e age me hyla ion o each
CpG posi ion ac oss all dono s o each genomic elemen .
Hea map package in R was used o gene a e hea maps (16).
S udy App o al
E hical app o al by he ins i u ion’s E hics Commi ee was
ecei ed p io o he beginning o he s udy. W i en in o med
consen was ecei ed om pa icipan s p io o inclusion in he
s udy.
RESULTS
Isola ion o CD34+and CD4+T Cell
Subpopula ions
Cell popula ions we e isola ed om he pe iphe al blood o
heal hy dono s by FACSo ing. The mean pu i y o CD34+
cells a e FACSo was >95%, as shown in a ep esen a i e
do plo in Figu e 1A. T eg and Tcon cells we e iden i ied as
CD3+CD4+CD25B igh CD127Low and CD3+CD4+CD25Low
cells, espec i ely. T eg cells we e u he pu i ied in o he
ollowing subpopula ions: CD45RA+CD15s−(naï e T eg),
CD45RA−CD15s−(non-supp essi e T eg-like cells) and
F on ie s in Immunology | www. on ie sin.o g 3No embe 2018 | Volume 9 | A icle 2540
Minskaia e al. Epigene ics T Cell Sub ypes
FIGURE 1 | Flow cy ome y analysis and pu i ica ion s a egy o T eg, Tcon, and CD34 cell popula ions. (A) Pu i y o CD34+cells a e FACSo ing. Rep esen a i e do
plo showing he equency o CD34+cells wi hin a li e cell ga e, as de e mined by FSC and SSC analysis. (B) Rep esen a i e do plo s showing he ga ing s a egy
o he isola ion o T eg and Tcon cells by FACSo . Fi s ly, FSC and SSC analysis was used o iden i y lymphocy es and o exclude double s. T eg cells we e isola ed
as CD3+CD4+CD25B igh CD127Low and Tcon as CD3+CD4+CD25Low cells. T eg we e u he ga ed in o CD45RA+CD15s−(naï e T eg), CD45RA−CD15s−
(non-supp essi e T eg-like cells), and CD45RA−CD15s+(eT eg). (C) Rep esen a i e do plo s showing he pu i y ob ained o each T cell popula ion a e FACSo .
(D) FoxP3 exp ession le els wi hin CD45RA/CD15s T eg subpopula ions, as well as wi hin naï e Tcon cells. A ep esen a i e o e lay o Foxp3 exp ession wi hin hese
T cell popula ions is shown. (E) Median Fluo escence In ensi y (MFI) o FoxP3 wi hin T eg subpopula ions o all indi iduals analyzed. As e isks deno e s a is ically
signi ican di e ences be ween g oups (*p=0.01–0.05; ****p<0.0001).
CD45RA−CD15s+(eT eg), as p e iously desc ibed (6).
The mean pu i y o he isola ed popula ions was >95%.
Rep esen a i e do plo s illus a ing he FACSo ga ing s a egy
(Figu e 1B) and pu i y (Figu e 1C) o each subpopula ion a e
shown. These pu i ied cell popula ions we e used o subsequen
me hyla ion s udies.
F on ie s in Immunology | www. on ie sin.o g 4No embe 2018 | Volume 9 | A icle 2540
Minskaia e al. Epigene ics T Cell Sub ypes
In addi ion, a PBMC sample was s ained wi h he same
an ibody combina ion used o FACSo plus in acellula FoxP3.
When we analyzed he exp ession le els o FoxP3 wi hin he
T eg subse s de ined by CD45RA and CD15s exp ession, we
con i med p e ious obse a ions (6) ha CD45RA+CD15s−
cells exp ess he lowes , CD45RA−CD15s−cells in e media e
and CD45RA−CD15s+cells he highes le els o FoxP3. This is
illus a ed in a ep esen a i e o e lay o FoxP3 exp ession wi hin
each o hese T eg subpopula ions, as well as wi hin naï e Tcon
o compa ison (Figu e 1D). FoxP3 exp ession le els wi hin T eg
subse s om all he dono s analyzed a e shown in Figu e 1E.
Me hyla ion o CAMTA1 In onic Region
In cells o he immune sys em, including T and B cells,
Ca2+signals a e essen ial o di e se cellula unc ions
including p oli e a ion, di e en ia ion, e ec o unc ion, gene
ansc ip ion, and apop osis. The in acellula concen a ion
o Ca2+is known o ac i a e se e al signaling p o eins and
ansc ip ion ac o s (TFs) such as calcineu in and NFAT, CaMK
and CREB, MEF2, and NFkB (17). Calcineu in–NFAT pa hway is
in ol ed in T eg de elopmen and unc ion and is a ec ed by he
absence o Ca2+signals (18). I is, he e o e, no coinciden al ha
one o he egions in CAMTA 1, a Ca2+-dependen calmodulin-
binding ansc ip ion ac o (19), was p oposed as an addi ional
molecula ma ke dis inguishing T eg om Tcon (7).
The me hyla ion s a us o 13 CpG si es wi hin he 470 bp
CAMTA1 in onic egion 3 was assessed based on he op s and
o BS DNA (Supplemen a y Tables 1,2). Two me hyla ion
pa e ns we e obse ed. In he i s , ep esen ed by male
dono M4 (Figu e 2B, le panel and Supplemen a y Table 3)
and comp ising male dono s M1-4 and emale dono F2
(Figu e 2C), he i s eigh CpG si es (CpGs 1-8) we e mo e
deme hyla ed in CD34+cells and he h ee subse s o T eg as
compa ed o Tcon, as is e iden om he a e age me hyla ion
pa e n (Figu e 2B, lowe panel).The a e age me hyla ion
pe cen ages o CpGs 1-8 in hese dono s we e 37.8% (CD34+),
36.2% (CD45RA+CD15s−), 13.9% (CD45RA−CD15s−), and
35.8% (CD45RA−15s+) as compa ed o 74.1% in Tcon
(Supplemen a y Table 3). In he second pa e n, ep esen ed
by dono F4 (Figu e 2B, igh panel) and comp ising six
dono s (M5, 6, and F1, 3-5), only wo CpG posi ions we e
consis en ly mo e deme hyla ed in CD34+cells and T eg
subse s as compa ed o Tcon. While in Tcon, CpGs 2
and 11 we e 70 and 74.2% me hyla ed, a e age me hyla ion
le els o hese wo si es in he o he ou popula ions we e
signi ican ly lowe : 25.8 and 33.3% (CD34+), 13.3 and 22.5%
(CD45RA+CD15s−), 10.8 and 16.7% (CD45RA−CD15s−), and
7.5 and 11.7% (CD45RA−CD15s+), espec i ely (Figu e 2B,
igh lowe panel and Supplemen a y Table 3). O e all, he
en i e CAMTA1 egion in Tcon subse was hea ily me hyla ed
(80%) as p esen ed by he a e age me hyla ion pa e n o ele en
dono s (Figu e 2C, lowe panel and Supplemen a y Table 3),
howe e , he main di e ences we e obse ed in CpGs 1-
11 (wi h he excep ion o CpG 9). A e age me hyla ion
le els o mo e deme hyla ed CpGs 1-8 wi hin he i s
200 bp pa o CAMTA1 egion we e 56.7% (CD34+),
49% (CD45RA+CD15s−), 41% (CD45RA−CD15s−), and 48%
(CD45RA−CD15s+) as compa ed o 78.4% in Tcon. CpG
si es 2 and 11 we e mo e deme hyla ed in CD34+and T eg
subse s: 24.4, 20.5, 10, and 15% in CpG 2 and 36.3, 44.5,
27.7, and 33.6% in CpG11 in CD34+, CD45RA+CD15s−,
CD45RA−CD15s−, and CD45RA−CD15s+cells, espec i ely,
as compa ed o 66.8 and 77.3% o hese CpG posi ions in
Tcon (Supplemen a y Table 3). The di e ences in me hyla ion
alues o all CpGs we e u he analyzed using he unc ion
lme om he package nlme o R and he calcula ed di e ence
was only conside ed as signi ican o hose esul s wi h p
<0.016. This analysis demons a ed signi ican di e ences
be ween all T eg sub ypes compa ed o Tcon (p=0.008 o
CD45RA+CD15s−,p=0.0003 o CD45RA−CD15s−, and
p=0.0004 o CD45RA−CD15s+).
Me hyla ion o FUT7 P omo e
The induc ion o CD15s on human hema opoie ic cells and
lymphocy es accompanies ansc ip ional ac i a ion o FUT7
which is in ol ed in he las s ep o sialyl Lewis X syn hesis
(12,20). FUT7 is highly exp essed in eT eg compa ed o
o he FOXP3+o FOXP3−subpopula ions (5). T ansc ip ional
egula ion o FUT7 plays an impo an ole in lineage-speci ic
exp ession o CD15s among lymphocy e subpopula ions. Fo
example, he gene a ion o E-selec in ligands on T cells
unde going nai e- o-memo y ansi ion was shown o equi e
FUT7 ac i i y (21). Human FUT7 p omo e was demons a ed
o ha e binding si es o se e al TFs, six o which: T-be ,
GATA-3, Sp1, CBP/P300, HDAC-3, and HDAC-5 may o m
ansc ip ional complex in human lymphoid cells (12,20).
As FUT7 gene is encoded in he e e se s and o ch omosome
9, FUT7 p omo e egion was ampli ied wi h he e e se s and-
speci ic p ime s (Supplemen a y Table 1). Due o he leng h
o his egion and he ac ha longe amplicons canno be
success ully p oduced as a esul o BS ea men , me hyla ion
s a us o 19 indi idual CpGs was assessed wi hin wo amplicons,
500bp and 454bp in leng h, oge he comp ising FUT7 p omo e
(Supplemen a y Table 2). Me hyla ion pa e n wi hin FUT7
p omo e egion was o e all simila among dono s M1-6 and
F1-3 and is ep esen ed by dono F3 (Figu e 3B). Simila o
dono F3, CpGs 1-7 o o he dono s (wi h he excep ion o
CpG 5) did no show a ia ions in me hyla ion in all i e cell
popula ions as demons a ed by he a e age me hyla ion p o ile
o all dono s (Figu e 3C, lowe panel). CpG posi ions 8-19
we e mos ly deme hyla ed in CD34+cells (wi h he excep ion
o CpG 13) wi h a e age me hyla ion le els o he wel e
si es being 37.9% (CD34+), 54.1% (CD45RA−CD15s−), and
48.8% (CD45RA−CD15s+) as compa ed o CD45RA+CD15s−
(69.6%) and Tcon (75%). CpGs 10–12, 14, and 15 we e
mo e deme hyla ed in CD34+(38.2%), CD45RA−CD15s−
(54.5%), and CD45RA−CD15s+(45%) popula ions while
CD45RA+CD15s−and Tcon cells displayed highe me hyla ion
le els (72.8 and 77.2%, espec i ely) (Figu e 3C, lowe panel, and
Supplemen a y Table 3). O e all, nT eg (CD45RA+CD15s−)
and Tcon subse s displayed a simila me hyla ed pa e n, while
CD45RA−CD15s−, eT eg (CD45RA−CD15s+) and CD34+
cells displayed a simila deme hyla ed pa e n, which was
F on ie s in Immunology | www. on ie sin.o g 5No embe 2018 | Volume 9 | A icle 2540

Minskaia e al. Epigene ics T Cell Sub ypes
FIGURE 2 | Me hyla ion pa e n o CAMTA1 in onic egion. CAMTA 1 in onic egion is mo e deme hyla ed in T eg subse s as compa ed o CD34+cells and Tcon
wi hin he 200 bp egion con aining CpGs 1-8. CpGs 2 and 11 a e s ikingly mo e deme hyla ed in bo h T eg and CD34+cells. (A) Schema ic p esen a ion o
CAMTA1 encoded on ch omosome 1q3.6. (B) Two me hyla ion pa e ns o CAMTA1 in i e cell popula ions: CD34+(1), CD45RA+CD15s−(2), CD45RA−CD15s−
(3), CD45RA−CD15s+(4), and Tcon (5) o ep esen a i e dono s M4 (pa e n 1) and F4 (pa e n 2). A e age me hyla ion pe cen ages o dono s ollowing he wo
pa e ns p esen ed in panels below. CpGs 1-13 a e numbe ed ela i e o he 5′-3′di ec ion o he coding ( op) s and. Each ho izon al line ep esen s DNA om one
cell wi h unmodi ied C in ligh blue and mC in da k blue. Me hyla ion pe cen age o each CpG si e was calcula ed based on he numbe o mCs in a o al o 20 and
summa ized in panels below. (C) Me hyla ion pa e ns o CAMTA1 in onic egion o indi idual dono s wi h he a e age me hyla ion pa e n o all dono s p esen ed in
he panel below. See also Supplemen a y Figu es 1,3.
cha ac e ized by hypome hyla ion o he i e in e nal CpG si es:
10–12, 14, and 15 wi hin he 225 bp egion o FUT7 p omo e .
In compa ison o Tcon, he me hyla ion alues o all
CpGs wi hin FUT7 egion we e signi ican ly di e en o
CD45RA−CD15s+(p=0.0019) and CD45RA−CD15s−T eg
(p=0.0098), bu no o CD45RA+CD15s−T eg (p=0.8889).
FOXP3 Enhance
FOXP3 gene exp ession is dependen on he ac i i y o se e al
egula o y DNA elemen s and ou e o s we e concen a ed on
he ou FOXP3 gene egions: enhance , p omo e , p eTSDR
and TSDR, all o which we e ampli ied wi h he e e se
s and-speci ic p ime s. Me hyla ion da a ob ained om emale
dono s F1-5 is delibe a ely p esen ed in i s aw o ma , wi hou
accoun ing o he possible Xi.
The FOXP3 enhance is one o he h ee egula o y elemen s
in he FOXP3 locus ha was shown o be con olled ia
CpG me hyla ion (9,22). Loca ed abou 6 kb ups eam o he
ansc ip ion s a si e (TSS), his 800 bp egion is ac ually pa
o a p o ein phospha ase 1 egula o y subuni locus. Me hyla ion
pa e n o he human FOXP3 enhance was shown o be di e en
om ha o i s mu ine homolog (22), wi h bo h T eg and Tcon
popula ions o heal hy dono s displaying simila p o iles. In
pa ien s wi h a h i is, howe e , T eg subse s displayed highe
deme hyla ion le els han in heal hy dono s (9), hus accoun ing
o g ea e di e ences in me hyla ion be ween T eg and Tcon
cells in his disease se ing.
Highly CpG ich FOXP3 enhance con ains o e 50 CpG
posi ions. Me hyla ion s a us o he i s 42 CpG si es was assessed
in six dono s (M1-4 and F1, 2). Because o limi ed blood sample
ma e ial, low gDNA quan i y om a e T eg subse s and mos ly
di icul ies in cloning, de e mina ion o me hyla ion s a us
wi hin his egion was no possible o o he dono s. Despi e he
imp essi e numbe o CpG si es wi hin he 497 bp amplicon, he
F on ie s in Immunology | www. on ie sin.o g 6No embe 2018 | Volume 9 | A icle 2540
Minskaia e al. Epigene ics T Cell Sub ypes
FIGURE 3 | Me hyla ion pa e n o FUT7 p omo e . CpG si es 10–12, 14, and 15 wi hin FUT7 p omo e display highe deme hyla ion le els in CD34+and e ec o
T eg as compa ed o naï e T eg and Tcon, indica ing ha CpGs 10-15 can be used as an addi ional molecula ma ke o dis inguish hese cell popula ions.
(A) Schema ic p esen a ion o FUT7 p omo e encoded in he e e se s and o ch omosome 9. (B) Me hyla ion pa e n o FUT7 p omo e in i e cell popula ions:
CD34+(1), CD45RA+CD15s−(2), CD45RA−CD15s−(3), CD45RA−CD15s+(4), and Tcon (5) o ep esen a i e dono F3. CpGs 1-19 a e numbe ed ela i e o he
5′-3′di ec ion o he coding ( e e se) s and. Each ho izon al line ep esen s DNA om one cell wi h unmodi ied C in ligh blue and mC in da k blue. Me hyla ion
pe cen age o each CpG si e was calcula ed based on he numbe o mCs in a o al o 20 and summa ized in he panel below. (C) Me hyla ion pa e ns o FUT7
p omo e o indi idual dono s wi h he a e age me hyla ion pa e n o all dono s p esen ed in he panel below. See also Supplemen a y Figu es 1,3.
enhance egion demons a ed an in ica e me hyla ion pa e n
ep esen ed by dono M4 (Figu e 4B). A e age me hyla ion
le els o he mos deme hyla ed CpG si es 4, 9, 11, and 20 o
dono M4 we e 11.2% (CD34+), 35% (CD45RA+CD15s−),
28.7% (CD45RA−CD15s−), 32.5% (CD45RA−CD15s+),
and 81.2% (Tcon) (Figu e 4B and Supplemen a y Table 3).
Simila o his dono , CpG posi ions 1–3, 5–8, and 14–19
we e mos ly me hyla ed in all cell popula ions o all dono s
(Figu e 4C). Howe e , CpG posi ions 4, 9, 11, 20, and 24–28
demons a ed highe deg ee o deme hyla ion in CD34+cells
and T eg sub ypes compa ed o Tcon as p esen ed by he
a e age me hyla ion pa e n o his egion (Figu e 4C, lowe
panel). A e age me hyla ion le els o CpGs 4, 9, 11, and 20 in
all dono s we e 27.6% (CD34+), 36.7% (CD45RA+CD15s−),
35.5% (CD45RA−CD15s−), 28.5% (CD45RA−CD15s+),
and 72.5% (Tcon) (Figu e 4C, lowe panel and
Supplemen a y Table 3).
S a is ical analysis u he con i med signi ican di e ences
in me hyla ion alues o all CpGs wi hin FOXP3 enhance
o CD45RA+CD15s−(p<0.00005), CD45RA−CD15s−T eg
(p=0.0001), and CD45RA−CD15s+(p=0.0002), compa ed
o Tcon.
FOXP3 p eTSDR
No much is known abou his egion excep ha pa o
i was mo e deme hyla ed in CD25highCD45RA−CD4+T eg
cells compa ed o CD25−CD45RA+CD4+Tcon (Amp6/7) (7).
Conside ing he possibili y o i s hypome hyla ion in T eg
sub ypes, i was also chosen o i s p oximi y o TSDR.
Compa ed o Amp6/7, con aining 7 CpG si es, p eTSDR egion
was ex ended o con ain 10 CpG si es wi hin 700 bp leng h
(Supplemen a y Tables 1,2) and se ed as an addi ional in e nal
con ol o TSDR me hyla ion bounda ies as i is loca ed
immedia ely ups eam (hence he denomina ion we p opose,
p eTSDR).
Dono s M2, 4, 5, and F4, 5 ollowed he pa e n p esen ed
o dono M6 (Figu e 5B, le panel) whe eby CpGs 2-4 we e
hea ily me hyla ed in all i e cell popula ions while a ious
deg ees o deme hyla ion we e obse ed in CpGs 5, 8, and 9
(Figu e 5C). As p esen ed by he a e age me hyla ion p o ile
F on ie s in Immunology | www. on ie sin.o g 7No embe 2018 | Volume 9 | A icle 2540
Minskaia e al. Epigene ics T Cell Sub ypes
FIGURE 4 | Me hyla ion pa e n o FOXP3 enhance . CpGs 4, 9, 11, 20, and 24–28 wi hin FOXP3 enhance a e mo e deme hyla ed in CD34+cells and T eg sub ypes
as compa ed o Tcon. (A) Schema ic p esen a ion o he FOXP3 enhance encoded in he e e se s and o Xp11.23. (B) Me hyla ion pa e n o he FOXP3 enhance
in i e cell popula ions: CD34+(1), CD45RA+CD15s−(2), CD45RA−CD15s−(3), CD45RA−CD15s+(4), and Tcon (5), o ep esen a i e dono M4. CpGs 1-42 a e
numbe ed ela i e o he 5′-3′di ec ion o he coding ( e e se) s and. Each ho izon al line ep esen s DNA om one cell wi h unmodi ied C in ligh blue and mC in da k
blue. Me hyla ion pe cen age o each CpG si e was calcula ed based on he numbe o mCs in a o al o 20 and summa ized in panels below. (C) Me hyla ion pa e ns
o he FOXP3 enhance o indi idual dono s wi h he a e age me hyla ion pa e n o all dono s p esen ed in he panel below. See also Supplemen a y Figu es 1,4.
F on ie s in Immunology | www. on ie sin.o g 8No embe 2018 | Volume 9 | A icle 2540
Minskaia e al. Epigene ics T Cell Sub ypes
o all dono s (Figu e 5C, lowe panel), hese 3 si es we e mo e
deme hyla ed in CD34+cells (19.6%) and wo T eg sub ypes:
CD45RA-CD15s- (25.9%) and CD45RA−CD15s+(28.2%)
while Tcon (49.6%) and CD45RA+CD15s−subse s (43.3%)
displayed highe me hyla ion le els (Supplemen a y Table 3).
In dono M6, he a e age me hyla ion le els o hese h ee
CpG si es we e 10% (CD34+), 16.7% (CD45RA+CD15s−),
1.7% (CD45RA−CD15s−), 43.3% (CD45RA−CD15s+), and
76.7% (Tcon). In e es ingly, mo e CpG si es h oughou he
en i e amplicon we e deme hyla ed in CD45RA+CD15s−and
CD45RA−CD15s−popula ions o emale dono F3 (Figu e 5B,
igh panel), wi h a e age me hyla ion le els o CpGs 1-10 being
33% o CD45RA+CD15s−and 47.5% o CD45RA−15s−as
compa ed o 91.5% o Tcon. This egion-wide deme hyla ion
was also p esen in CD34+cells o dono s F1 and F2, wi h
a e age me hyla ion pe cen ages o CpGs 1-10 being 9 and
11.5%, espec i ely. Compa ed o Tcon, s a is ically signi ican
di e ences in me hyla ion alues o all CpGs wi hin FOXP3
p eTSDR we e obse ed o CD45RA−CD15s+(p=0.0087)
and CD45RA−CD15s−(p=0.0004) T eg, bu no o
CD45RA+CD15s−T eg (p=0.2103).
FOXP3 P omo e and TSDR
Two egula o y egions in he FOXP3 locus, p omo e and TSDR,
ha e been p e iously shown o be deme hyla ed in hymically
de i ed nT eg and mos ly me hyla ed in induced T eg (iT eg)
and Tcon cells, bo h in mice and humans (7,23–25). Se e al TFs,
such as NFAT, AP-1, Foxo1 and 3 bind o he FOXP3 p omo e .
Some o hese plus o he TFs, such as CREB, NF-kB, Runx1,
STAT5, Ga a3, E s1, and FOXP3 i sel in e ac wi h TSDR, wi h
CREB and NFAT, o example, in a deme hyla ion-dependen
manne (23,24,26).
Me hyla ion analysis o 10 CpG si es wi hin 451 bp
egion o FOXP3 p omo e and 15 CpG si es wi hin 700
bp egion co e ing TSDR (Supplemen a y Tables 1,2) om
he ou T lymphocy e subse s o dono M1 (Figu es 6B,E,
le panels) was in ag eemen wi h p e iously published
da a demons a ing almos comple e deme hyla ion o hese
egions in T eg sub ypes and a high deg ee o me hyla ion
in Tcon cells (Supplemen a y Table 3). Wi h he excep ion
o CD45RA+CD15s−subse , which was hea ily me hyla ed
wi hin TSDR and could accoun o indi idual di e ences,
dono M2 displayed simila me hyla ion pa e n in bo h
egions (Figu es 6C,F). A e age me hyla ion le els o CpGs
1-10 wi hin he p omo e egion o dono s M1 and M2
we e 9% (CD45RA+CD15s−), 1.7% (CD45RA−CD15s−), 7.5%
(CD45RA−CD15s+) and 58% (Tcon) (Figu e 6B, le lowe
panel), while o CpGs 1-15 wi hin TSDR egion hey we e
48.7% (CD45RA+CD15s−), 0.7% (CD45RA−CD15s−), 11.5%
(CD45RA−CD15s+), and 85.7% (Tcon) (Figu e 6E, le lowe
panel and Supplemen a y Table 3). Howe e , p omo e egion
was deme hyla ed (Figu e 6C) and TSDR was me hyla ed
(Figu e 6F) in dono s M3-6 and F1-5, wi hou signi ican
di e ences be ween cell subse s and gende s. Wi hin he
p omo e egion, he a e age me hyla ion le els o CpGs 1-
10 o hese dono s we e below 1% o CD34+cells and T eg
sub ypes and sligh ly highe (6.6%) o Tcon (Figu e 6C, lowe
panel and Supplemen a y Table 3). Wi hin TSDR, he a e age
me hyla ion le els o CpGs 1-15 we e close o 100% in all cell
popula ions (Figu e 6F and Supplemen a y Table 3). Sequences
om he 700 bp TSDR egion had an equally high le el o non-
CpG con e sion, indica ing ha mCs wi hin CpG si es we e
in insic o he analyzed egion and we e no he consequence
o incomple e con e sion.
As expec ed, no signi ican di e ences in me hyla ion wi hin
FOXP3 p omo e and TSDR we e obse ed o any o he
T eg sub ypes in ela ion o Tcon (p=0.0401 and p=0.1465
o CD45RA+CD15s−,p=0.0438 and p=0.1184 o
CD45RA−CD15s−, and p=0.0408 and p=0.1551 o
CD45RA−CD15s+T eg, espec i ely o FOXP3 p omo e and
TSDR).
Hie a chical Clus e ing
We applied an unbiased clus e ing me hod o examine
he dis ance o cell sub ypes om each o he based on
a e age CpG me hyla ion o each genomic elemen
(Supplemen a y Figu e 1). Hie a chical clus e ing shows
ha me hyla ion pa e ns o p omo e , enhance and TSDR
egion o FOXP3 and CAMTA1 in onic egion o Tcon cells a e
a hes om all o he T eg sub ypes, while o FOXP3 p eTSDR
and FUT7 p omo e me hyla ion pa e ns show clus e ing o
Tcon cells wi h CD45RA+CD15s−T eg sub ype. Clus e ing
on CpG posi ions o each genomic elemen u he highligh s
he posi ions wi h s iking me hyla ion di e ences ac oss he
cell sub ypes (Supplemen a y Figu e 1). These CpG posi ions
wi h hei a e age me hyla ion obse ed ac oss he ou cell
popula ions a e summa ized in Figu e 7.
S and-Speci ic Me hyla ion Biases Wi hin
FOXP3 TSDR and P omo e
P e iously obse ed me hyla ion s a us o FOXP3 p omo e
and TSDR o dono s M3-6 and F1-5 was puzzling o se e al
easons. Fi s , no changes we e in oduced in e ms o he wo k
low (human ac o , eagen s, p o ocols). Second, all six gene
egions we e ampli ied om he same ba ch o BS DNA om
each cell popula ion. A he same ime, he six gene egions
om he i e cell popula ions we e ampli ied om di e en
ba ches o DNA (one om each cell popula ion). Thi d, p eTSDR
egion was in en ionally in oduced as an addi ional in e nal
con ol o bo h TSDR bounda ies and e iciency o BS PCR:
he p ime s we e designed in such a way as o p oduce 700
bp amplicons (Supplemen a y Table 1). Fou h, me hyla ion
pa e ns o FOXP3 p omo e and TSDR we e simila in hund eds
o sequences analyzed (abou 110 sequences pe gene egion
om each dono , o e 1,000 sequences pe gene egion in o al).
These easons ins illed g ea con idence in ou da a and led us
o conside he possibili y ha he obse ed me hyla ion pa e ns
we e in insic p ope ies o he wo egions o dono s M3-6 and
F1-5.
To a oid any bias ela ed o BS ea men and ampli ica ion,
a new ba ch o BS- ea ed DNA om dono s M2, F1,
and F2 was used o de ec me hyla ion le els wi hin TSDR
ampli ied ei he wi h wo op s and-speci ic o bo h op and
e e se s and-speci ic p ime se s (Supplemen a y Figu e 2).
F on ie s in Immunology | www. on ie sin.o g 9No embe 2018 | Volume 9 | A icle 2540
Minskaia e al. Epigene ics T Cell Sub ypes
emale naï e and ac i a ed lymphocy es do no main ain Xi wi h
he same ideli y as o he soma ic cells and a e p edisposed
o become pa ially eac i a ed and o o e exp ess immuni y-
ela ed genes (50,52). The e o e, i is possible ha silencing
a he FOXP3 locus is incomple e and leaky Xi may ake
place. In addi ion, eigh een X-linked miRNAs we e shown o
be o e exp essed in CD4+T cells o women wi h lupus, i e o
hose—in expe imen ally deme hyla ed CD4+T cells sugges ing
ha deme hyla ion con ibu ed o he escape o Xi. Se e al o he
emale biased miRNAs po en ially egula e FOXP3 exp ession
wi h some miRNA binding si es loca ed in 3′UTR o FOXP3
ansc ip (53).
To conclude, he indings in his s udy sugges ha
hypome hyla ed CpG si es, p esen in ou egions o he FOXP3
locus, CAMTA1 and FUT7 gene egions, can po en ially be used
o dis inguish subse s o CD4+T lymphocy es in bo h sexes.
Wi h he excep ion o FUT7 p omo e , hese CpG si es also de ine
CD34+cells as ha ing a “naï e T eg-like epigenome” ha do no ,
howe e , exp ess FOXP3 p o ein.
We also desc ibe p e iously un epo ed s and-bias
hemime hyla ion pa e n wi hin he human FOXP3 p omo e
and TSDR in some dono s o bo h gende s. The coding s and
is deme hyla ed wi hin he p omo e and me hyla ed wi hin he
TSDR in all o he CD4+cell sub ypes, whils he empla e s and
ollows he “con en ional” me hyla ion pa e n.
These da a p o ide new insigh s in o he epigene ic con ol
o CD4+T lymphocy es. FOXP3 exp ession and TSDR
deme hyla ion a e classically used o con i ma ion o T eg
lineage commi men and di e en ia ion s a us, and o
p ognos ic pu poses in a ious disease se ings. While he
e ec o he desc ibed hemime hyla ion pa e n on T eg unc ion
will become he subjec o a new s udy, ou indings sugges ha
he s and-speci ic app oach can be ins umen al in disclosing
po en ial di e ences be ween subse s o T eg and Tcon. The
s and-bias me hyla ion indings challenge cu en simpli ied
in e p e a ions o TSDR me hyla ion as a T eg ma ke since op
s and-based TSDR-deme hyla ed cells may be e e se-s and
TSDR-me hyla ed.
In ligh o ou esul s, we p opose ha i is essen ial, o (i)
ob ain he TSDR coding e e se-s and me hyla ion da a i s ,
(ii) u he cla i y he p ecise TSDR me hyla ion pa e n o bo h
s ands, and /o (iii) use CAMTA1 o FUT7 le els o me hyla ion
as addi ional molecula ma ke s o clea ly dis inguish subse s o
T eg om Tcon.
Finally, ou indings will di ec ly impac bo h on T eg esea ch
and on he clinical applica ion o T eg- ela ed he apies and
p ognos ics in he ields o au oimmuni y, alle gy and cance .
ETHICS STATEMENT
This s udy was ca ied ou in acco dance wi h he
ecommenda ions o he Lisbon No h Cen al Hospi als
(CHLN) and Lisbon Academic Medical Cen e (CAML). The
p o ocol was app o ed by he E hics Commi ee o Hospi al
San a Ma ia, Lisbon, Po ugal ( e . 459/13). All subjec s ga e
w i en in o med consen in acco dance wi h he Decla a ion o
Helsinki.
AUTHOR CONTRIBUTIONS
EM co-designed he p ojec , designed and pe o med
expe imen s, analyzed da a, and w o e he manusc ip . BS
p o ided echnical assis ance. MS and RA p o ided assis ance
wi h low cy ome y analysis. RR pe o med he s a is ical
analysis. SK pe o med hie a chical analysis. AV p o ided
echnical assis ance in cell so ing. JL designed he p ojec ,
discussed p og ess and c i ically ead he manusc ip .
FUNDING
This esea ch was unded by: Fundação pa a a Ciência e
Tecnologia, Po ugal unde he Ha a d Medical School–
Po ugal P og am p ojec Induc ion o Immune Tole ance in
Human Allogeneic Hema opoie ic S em Cell T ansplan a ion
(HMSP-ICT/0001/2011), Gabine e de Apoio à In es igação
Cien í ica, Tecnológica e Ino ação, GAPIC (Faculdade de
Medicina ULisboa, GAPIC-04-2016 P ojec 20160011), LISBOA-
01-0145-FEDER-007391 (p ojec co- unded by FEDER h ough
POR Lisboa 2020—P og ama Ope acional Regional de Lisboa,
PORTUGAL 2020 and Fundação pa a a Ciência e a Tecnologia)
and ENLIGHT-TEN p ojec (Eu opean Union’s Ho izon
2020 esea ch and inno a ion p og amme unde he Ma ie
Sklodowska-Cu ie g an ag eemen 675395).
ACKNOWLEDGMENTS
We a e g a e ul o P o esso Ma in D. Ryan o Uni e si y o S .
And ews o c i ically eading he manusc ip and Sa a Buca o
echnical assis ance.
SUPPLEMENTARY MATERIAL
The Supplemen a y Ma e ial o his a icle can be ound
online a : h ps://www. on ie sin.o g/a icles/10.3389/ immu.
2018.02540/ ull#supplemen a y-ma e ial
REFERENCES
1. Miya a M, Amou a Z, Pa izo C, Badoual C, Do gham K, T ad S, e al. Global
na u al egula o y T cell deple ion in ac i e sys emic lupus e y hema osus. J
Immunol. (2005) 175:8392–400. doi: 10.4049/jimmunol.175.12.8392
2. Sakaguchi S, Yamaguchi T, Nomu a T, Ono M. Regula o y T cells and immune
ole ance. Cell (2008) 133:775–87. doi: 10.1016/j.cell.2008.05.009
3. Allan SE, C ome SQ, C ellin NK, Passe ini L, S eine TS, Bacche a R,
e al. Ac i a ion-induced FOXP3 in human T e ec o cells does no
supp ess p oli e a ion o cy okine p oduc ion. In Immunol. (2007) 19:345–
54. doi: 10.1093/in imm/dxm014
4. Wang J, Ioan-Facsinay A, an de Voo EIH, Huizinga TWJ, Toes REM.
T ansien exp ession o FOXP3 in human ac i a ed non egula o y CD4+T
cells. Eu J Immunol. (2007) 37:129–38. doi: 10.1002/eji.200636435
F on ie s in Immunology | www. on ie sin.o g 16 No embe 2018 | Volume 9 | A icle 2540

Minskaia e al. Epigene ics T Cell Sub ypes
5. Miya a M, Yoshioka Y, Ki oh A, Shima T, Wing K, Niwa A, e al.
Func ional delinea ion and di e en ia ion dynamics o human CD4+T
cells exp essing he FoxP3 ansc ip ion ac o . Immuni y (2009) 30:899–911.
doi: 10.1016/j.immuni.2009.03.019
6. Miya a M, Chade D, Sage E, Sugiyama D, Nishikawa H, Bou y D, e al. Sialyl
Lewis x (CD15s) iden i ies highly di e en ia ed and mos supp essi e FOXP3
high egula o y T cells in humans. P oc Na l Acad Sci USA. (2015) 112:7225–30.
doi: 10.1073/pnas.1508224112
7. Ba on U, Floess S, Wieczo ek G, Baumann K, G ü zkau A, Dong J, e al.
DNA deme hyla ion in he human FOXP3 locus disc imina es egula o y T
cells om ac i a ed FOXP3+con en ional T cells. Eu J Immunol. (2007)
37:2378–89. doi: 10.1002/eji.200737594
8. Zheng Y, Jose owicz S, Chaudh y A, Peng XP, Fo bush K, Rudensky AY. Role
o conse ed non-coding DNA elemen s in he Foxp3 gene in egula o y T-cell
a e. Na u e (2010) 463:808–12. doi: 10.1038/na u e08750
9. Kennedy A, Schmid EM, C ibbs AP, Penn H, Amjadi P, Syed K,
e al. A no el ups eam enhance o FOXP3, sensi i e o me hyla ion-
induced silencing, exhibi s dys egula ed me hyla ion in heuma oid a h i is
T eg cells. Eu J Immunol. (2014) 44:2968–78. doi: 10.1002/eji.2014
44453
10. G a es MC, Ben on M, Lea RA, Boyle M, Tajou i L, Maca ney-Coxson
D, e al. Me hyla ion di e ences a he HLA-DRB1 locus in CD4+T-
Cells a e associa ed wi h mul iple scle osis. Mul Scle . (2014) 20:1033–41.
doi: 10.1177/1352458513516529
11. Liu Y, A yee MJ, Padyuko L, Fallin MD, Hesselbe g E, Runa sson A,
e al. Epigenome-wide associa ion da a implica e DNA me hyla ion as an
in e media y o gene ic isk in heuma oid a h i is. Na Bio echnol. (2013)
31:142–7. doi: 10.1038/nb .2487
12. Chen G-Y, Osada H, San ama ia-Babi LF, Kannagi R. In e ac ion
o GATA-3/T-be ansc ip ion ac o s egula es exp ession o
sialyl Lewis X homing ecep o s on Th1/Th2 lymphocy es. P oc
Na l Acad Sci USA. (2006) 103:16894–9. doi: 10.1073/pnas.06079
26103
13. Zhang Y, Rohde C, Reinha d R, Voelcke -Rehage C, Jel sch A. Non-imp in ed
allele-speci ic DNA me hyla ion on human au osomes. Genome Biol. (2009)
10:R138. doi: 10.1186/gb-2009-10-12- 138
14. Kaminsky ZA, Tang T, Wang S-C, P ak C, Oh GHT, Wong AHC, e al. DNA
me hyla ion p o iles in monozygo ic and dizygo ic wins. Na Gene . (2009)
41:240–5. doi: 10.1038/ng.286
15. Böyum A. Isola ion o mononuclea cells and g anulocy es om human
blood. Isola ion o monuclea cells by one cen i uga ion, and o g anulocy es
by combining cen i uga ion and sedimen a ion a 1 g. Scand J Clin Lab In es
Suppl. (1968) 97:77–89.
16. Zepeda-Mendoza ML, Resendis-An onio O. Hie a chical Agglome a i e
Clus e ing. In: Dubi zky W, Wolkenhaue O, Cho K. H. Yoko a H, edi o s.
Encyclopedia o Sys ems Biology. New Yo k, NY: Sp inge (2013). p. 886–7.
doi: 10.1007/978-1-4419-9863-7_1371
17. Hogan PG, Chen L, Na done J, Rao A. T ansc ip ional egula ion
by calcium, calcineu in, and NFAT. Genes De . (2003) 17:2205–32.
doi: 10.1101/gad.1102703
18. Wu Y, Bo de M, Heissmeye V, Feue e M, Lapan AD, S oud JC, e al.
FOXP3 con ols egula o y T cell unc ion h ough coope a ion wi h NFAT.
Cell (2006) 126:375–87. doi: 10.1016/j.cell.2006.05.042
19. Finkle A, Ashe y-Padan R, F omm H. CAMTAs: calmodulin-binding
ansc ip ion ac i a o s om plan s o human. FEBS Le . (2007) 581:3893–8.
doi: 10.1016/j. ebsle .2007.07.051
20. Bu ows GG, Mazia z RT, Hunady K, Lehman N, Rabe A, Deans
RJ, e al. Human mul ipo en adul p ogeni o cells ansc ip ionally
egula e ucosyl ans e ase VII. Cy o he apy (2014) 16:566–75.
doi: 10.1016/j.jcy .2013.08.004
21. Takahashi R, Mizukawa Y, Yamazaki Y, Hayakawa K, Hayakawa J, Kudo
A, e al. In i o di e en ia ion om nai e o ma u e E-selec in binding
CD4 T cells: acquisi ion o skin-homing p ope ies occu s independen ly o
cu aneous lymphocy e an igen exp ession. J Immunol. (2003) 171:5769–77.
doi: 10.4049/jimmunol.171.11.5769
22. Lal G, Zhang N, an de Touw W, Ding Y, Ju W, Bo inge EP, e al. Epigene ic
egula ion o Foxp3 exp ession in egula o y T cells by DNA me hyla ion. J
Immunol. (2009) 182:259–73. doi: 10.4049/jimmunol.182.1.259
23. Kim H-P, Leona d WJ. CREB/ATF-dependen T cell ecep o –induced FoxP3
gene exp ession: a ole o DNA me hyla ion. J Exp Med. (2007) 204:1543–51.
doi: 10.1084/jem.20070109
24. Ruan Q, Kameswa an V, Tone Y, Li L, Liou H-C, G eene
MI, e al. De elopmen o Foxp3(+) egula o y cells is
d i en by he c-Rel enhanceosome. Immuni y (2009) 31:932–40.
doi: 10.1016/j.immuni.2009.10.006
25. Polansky JK, K e schme K, F eye J, Floess S, Ga be A, Ba on U, e al.
DNA me hyla ion con ols Foxp3 gene exp ession. Eu J Immunol. (2008)
38:1654–63. doi: 10.1002/eji.200838105
26. Rud a D, Egawa T, Chong MMW, T eu ing P, Li man DR, Rudensky AY.
Runx-CBFbe a complexes con ol exp ession o he ansc ip ion ac o Foxp3
in egula o y T cells. Na Immunol. (2009) 10:1170–7. doi: 10.1038/ni.1795
27. Zemmou D, P a ama A, Loughhead SM, Ma his D, Benois C. Flic , a long
noncoding RNA, modula es Foxp3 exp ession and au oimmuni y. P oc Na l
Acad Sci. (2017) 114:E3472–E3480. doi: 10.1073/pnas.1700946114
28. Mo ikawa H, Ohku a N, Vandenbon A, I oh M, Nagao-Sa o S, Kawaji H, e al.
Di e en ial oles o epigene ic changes and Foxp3 exp ession in egula o y T
cell-speci ic ansc ip ional egula ion. P oc Na l Acad Sci. (2014) 111:5289–
94. doi: 10.1073/pnas.1312717110
29. Chen Z, Ba bi J, Bu S, Yang H-Y, Li Z, Gao Y, e al. The ubiqui in ligase S ub1
nega i ely modula es egula o y T cell supp essi e ac i i y by p omo ing
deg ada ion o he ansc ip ion ac o Foxp3. Immuni y (2013) 39:272–85.
doi: 10.1016/j.immuni.2013.08.006
30. Messegue X, Escude o R, Fa é D, Núñez O, Ma ínez J, Albà
MM. PROMO: de ec ion o known ansc ip ion egula o y elemen s
using species- ailo ed sea ches. Bioin o ma ics (2002) 18:333–4.
doi: 10.1093/bioin o ma ics/18.2.333
31. Woodcock DM, Lawle CB, Linsenmeye ME, Dohe y JP, Wa en WD.
Asymme ic me hyla ion in he hype me hyla ed CpG p omo e egion o he
human L1 e o ansposon. J Biol Chem. (1997) 272:7810–6.
32. Singal R, Vanwe JM. De no o me hyla ion o an emb yonic globin gene
du ing no mal de elopmen is s and speci ic and sp eads om he p oximal
ansc ibed egion. Blood (2001) 98:3441–46. doi: 10.1182/blood.V98.12.3441
33. Luo S, P euss D. S and-biased DNA me hyla ion associa ed wi h cen ome ic
egions in A abidopsis. P oc Na l Acad Sci U S A. (2003) 100:11133–138.
doi: 10.1073/pnas.1831011100
34. Ta u a R, Zeschnigk M, Adamzik M, P obs -Keppe M, Bue J, Keh mann J.
Quan i ica ion o Regula o y T Cells in Sep ic Pa ien s by Real-Time PCR-
Based Me hyla ion Assay and Flow Cy ome y. PLoS One (2012) 7:1–9.
doi: 10.1371/jou nal.pone.0049962
35. Zhuo C, Li Z, Xu Y, Wang Y, Li Q, Peng J, e al. Highe FOXP3-
TSDR deme hyla ion a es in adjacen no mal issues in pa ien s wi h colon
cance we e associa ed wi h wo se su i al. Mol Cance (2014) 13:153.
doi: 10.1186/1476-4598-13-153
36. Baeche -Allan CM, Cos an ino CM, C e ano ich GL, Ashley CW, Be iou G,
Dominguez-Villa M, e al. CD2 cos imula ion e eals de ec i e ac i i y
by human CD4+CD25hi egula o y cells in pa ien s wi h mul iple
scle osis. J Immunol. (2011) 186:3317–26. doi: 10.4049/jimmunol.10
02502
37. Lucas S, Van Ba en N, De Sme C, Coulie PG. Deme hyla ion o he FOXP3
gene in human melanoma cells p ecludes he use o his epigene ic ma k o
quan i ica ion o T egs in unsepa a ed melanoma samples. In J Cance (2012)
130:1960–1966. doi: 10.1002/ijc.26198
38. Bailey-Buck ou SL, Ma inez-Llo della M, Zhou X, An hony B, Rosen hal
W, Luche H, e al. Sel -an igen-D i en Ac i a ion induces ins abili y o
egula o y T Cells du ing an in lamma o y au oimmune esponse. Immuni y
(2013) 39:949–62. doi: 10.1016/j.immuni.2013.10.016
39. Bending D, Pesenacke AM, U su S, Wu Q, Lom H, Thi ugnanabalan
B, e al. Hypome hyla ion a he egula o y T Cell-speci ic deme hyla ed
egion in CD25hi T Cells is decoupled om FOXP3 exp ession a he
in lamed si e in childhood a h i is. J Immunol (2014) 193:2699–708.
doi: 10.4049/jimmunol.1400599
40. Guo H, Zheng M, Zhang K, Yang F, Zhang X, Han Q, e al. Func ional de ec s
in CD4+CD25high FoxP3+ egula o y cells in ankylosing spondyli is. Sci Rep.
(2016) 6:37559. doi: 10.1038/s ep37559
41. Shimazu Y, Shimazu Y, Hishizawa M, Hamaguchi M, Nagai Y, Sugino N,
e al. Hypome hyla ion o he T eg-speci ic deme hyla ed egion in FOXP3
F on ie s in Immunology | www. on ie sin.o g 17 No embe 2018 | Volume 9 | A icle 2540
Minskaia e al. Epigene ics T Cell Sub ypes
is a hallma k o he egula o y T-cell sub ype in adul T-cell leukemia.
Cance Immunol Res. (2016) 4:136–45. doi: 10.1158/2326-6066.CIR-15-
0148
42. S ockis J, Fink W, F ancois V, Conne o e T, de Sme C, Knoops L, e al.
Compa ison o s able human T eg and Th clones by ansc ip ional
p o iling. Eu J Immunol. (2009) 39:869–82. doi: 10.1002/eji.2008
38807
43. Polansky JK, Sch eibe L, Thelemann C, Ludwig L, K üge M, Baumg ass R,
e al. Me hyla ion ma e s: Binding o E s-1 o he deme hyla ed Foxp3 gene
con ibu es o he s abiliza ion o Foxp3 exp ession in egula o y T cells. J Mol
Med. (2010) 88:1029–40. doi: 10.1007/s00109-010-0642-1
44. Lis M, Wal he D. The o ien a ion o ansc ip ion ac o binding si e mo i s
in gene p omo e egions: does i ma e ? BMC Genomics (2016) 17:185.
doi: 10.1186/s12864-016-2549-x
45. Hashimshony T, Zhang J, Keshe I, Bus in M, Ceda H. The ole o DNA
me hyla ion in se ing up ch oma in s uc u e du ing de elopmen . Na
Gene . (2003) 34:187–92. doi: 10.1038/ng1158
46. Schmidl C, Hansmann L, Lassmann T, Balwie z PJ, Kawaji H, I oh
M, e al. The enhance and p omo e landscape o human egula o y
and con en ional T-cell subpopula ions. Blood (2014) 123:e68–78.
doi: 10.1182/blood-2013-02-486944
47. Nai VS, Oh KI. Down- egula ion o Te 2 p e en s TSDR deme hyla ion
in IL2 de icien egula o y T cells. Biochem Biophys Res Commun. (2014)
450:918–24. doi: 10.1016/j.bb c.2014.06.110
48. Ichiyama K, Chen T, Wang X, Yan X, Kim B-S, Tanaka S, e al.
The me hylcy osine dioxygenase Te 2 p omo es DNA deme hyla ion and
ac i a ion o cy okine gene exp ession in T cells. Immuni y (2015) 42:613–26.
doi: 10.1016/j.immuni.2015.03.005
49. Nes o CE, Len ini A, Hägg Nilsson C, Gawel DR, Gus a sson M, Ma son
L, e al. 5-Hyd oxyme hylcy osine emodeling p ecedes lineage speci ica ion
du ing di e en ia ion o human CD4+T Cells. Cell Rep. (2016) 16:559–70.
doi: 10.1016/j.cel ep.2016.05.091
50. Ca el L, Willa d HF. X-inac i a ion p o ile e eals ex ensi e a iabili y
in X-linked gene exp ession in emales. Na u e (2005) 434:400–4.
doi: 10.1038/na u e03479
51. B ooks WH, Renaudineau Y. Epigene ics and au oimmune diseases:
The X ch omosome-nucleolus nexus. F on Gene . (2015) 6:1–20.
doi: 10.3389/ gene.2015.00022
52. Wang J, Sy e CM, K ame MC, Basu A, A chison ML, Angue a MC. Unusual
main enance o X ch omosome inac i a ion p edisposes emale lymphocy es
o inc eased exp ession om he inac i e X. P oc Na l Acad Sci. (2016)
113:E2029–E2038. doi: 10.1073/pnas.1520113113
53. Hewagama A. Role o X Ch omosome encoded miRNAs in au oimmuni y.
Rheuma ol Cu Res. (2013) 3:1–7. doi: 10.4172/2161-1149.1000114
Con lic o In e es S a emen : The au ho s decla e ha he esea ch was
conduc ed in he absence o any comme cial o inancial ela ionships ha could
be cons ued as a po en ial con lic o in e es .
Copy igh © 2018 Minskaia, Sa ai a, Soa es, Aze edo, Ribei o, Kuma , Viei a and
Lace da. This is an open-access a icle dis ibu ed unde he e ms o he C ea i e
Commons A ibu ion License (CC BY). The use, dis ibu ion o ep oduc ion in
o he o ums is pe mi ed, p o ided he o iginal au ho (s) and he copy igh owne (s)
a e c edi ed and ha he o iginal publica ion in his jou nal is ci ed, in acco dance
wi h accep ed academic p ac ice. No use, dis ibu ion o ep oduc ion is pe mi ed
which does no comply wi h hese e ms.
F on ie s in Immunology | www. on ie sin.o g 18 No embe 2018 | Volume 9 | A icle 2540