Maturation and phenotypic heterogeneity of human CD4+ regulatory T cells from birth to adulthood and after allogeneic stem cell transplantation

Ma u a ion and Pheno ypic He e ogenei y o Human CD4+ Regula o y T Cells F om Bi h o Adul hood and A e Allogeneic S em Cell T ansplan a ion Tiago R. Ma os 1,2,3,4 , Masahi o Hi akawa 1,2 , Ana C. Alho 1,2,3 , La s Neleman 4 , Luis G aca 3 and Je ome Ri z 1,2 * 1 Di ision o Hema ologic Malignancies and Depa men o Medical Oncology, Dana-Fa be Cance Ins i u e, Bos on, MA, Uni ed S a es, 2 Ha a d Medical School, Bos on, MA, Uni ed S a es, 3 Ins i u o de Medicina Molecula , Faculdade de Medicina, Uni e sidade de Lisboa, Lisbon, Po ugal, 4 Ams e dam Uni e si y Medical Cen e s, Depa men o De ma ology, Uni e si y o Ams e dam, Ams e dam, Ne he lands CD4 + Regula o y T cells (T eg) play a c i ical ole in main aining immune homeos asis. Va ious T eg subse s ha e been iden ified, howe e he he e ogenei y o T eg subpopula ions du ing de elopmen emains uncha ac e ized. Using mass cy ome y we ob ained single cell da a on exp ession o 35 unc ional ma ke s o examine he he e ogenei y o T eg cells a bi h and in adul s. Unsupe ised clus e ing algo i hms FlowSOM and ACCENSE we e used o quan i y T eg he e ogenei y. As expec ed, T eg in umbilical co d blood we e p edomina ely naï e while T eg in adul blood we e p edomina ely cen al memo y and e ec o memo y cells. Al hough umbilical co d blood T eg a e mos ly naï e cells, we obse ed mul iple pheno ypic T eg subse s in co d blood. Ne e heless, pe iphe al blood in adul s con ained highe pe cen ages o T eg and he he e ogenei y o T eg was significan ly inc eased in adul s. We also s udied T eg he e ogenei y h oughou a 2-yea pe iod a e allogeneic hema opoie ic s em cell ansplan a ion (alloHSCT) and in pa ien s wi h ch onic g a - e sus-hos disease (cGVHD). T eg he e ogenei y eco e ed apidly a e alloHSCT and g adually inc eased in he fi s wo yea s pos - ansplan . Howe e , pa ien s wi h cGVHD had significan ly ewe dis inc T eg subpopula ions, p oposing a co ela ion be ween a dis up ed T eg he e ogenei y and cGVHD. Ou s udy is he fi s o compa e human T eg he e ogenei y a bi h, in heal hy adul s and in pa ien s a e alloHSCT wi h and wi hou cGVHD. This app oach o cha ac e ize T eg he e ogenei y based on exp ession o a la ge panel o unc ional ma ke s may enable u u e s udies o iden i y specific T eg de ec s ha con ibu e o immune dys unc ion. Keywo ds: T eg - egula o y T cell, immunology, T cell, he e ogenei y, di e si y, G HD, alloHSCT, CD4 F on ie s in Immunology | www. on ie sin.o g Janua y 2021 | Volume 11 | A icle 5705501 Edi ed by: Esen Sefik, Yale Uni e si y, Uni ed S a es Re iewed by: Hongbo Hu, Sichuan Uni e si y, China Ka ia Boni ace, Uni e si e ´de Bo deaux, F ance *Co espondence: Je ome Ri z [email p o ec ed] 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: 08 June 2020 Accep ed: 27 No embe 2020 Published: 18 Janua y 2021 Ci a ion: Ma os TR, Hi akawa M, Alho AC, Neleman L, G aca L and Ri z J (2021) Ma u a ion and Pheno ypic He e ogenei y o Human CD4+ Regula o y T Cells F om Bi h o Adul hood and A e Allogeneic S em Cell T ansplan a ion. F on . Immunol. 11:570550. doi: 10.3389/ immu.2020.570550 ORIGINAL RESEARCH published: 18 Janua y 2021 doi: 10.3389/ immu.2020.570550 HIGHLIGHTS - Dis inc subpopula ions o CD4 + T eg cells a e p esen a bi h and pheno ypic T eg he e ogenei y inc eases in adul hood - CD4 + T eg he e ogenei y inc eases a e allogeneic s em cell ansplan a ion and is educed du ing ch onic G a - e sus- Hos -Disease INTRODUCTION Regula o y T cells (T egs) a e essen ial elemen s o a heal hy immune sys em. They comp ise be ween 5–10% o he pe iphe al blood CD4 + T cell compa men in heal hy indi iduals and play a c i ical ole p o ec ing hei hos agains immunopa hological damage ollowing inflamma o y o immunological challenges (1,2). T egs a e able o supp ess a ange o e ec o cell ypes h ough se e al mechanisms, bo h di ec and indi ec , ensu ing pe iphe al ole ance and immune homeos asis (3,4). The in ica e balance be ween immune esponse and supp ession can be dis u bed by bo h a educed numbe o T egs as well as a decline in hei unc ionali y, po en ially leading o au oimmune pa hology (5). T egso igina ein he hymusandcanbeiden ified by exp ession o CD3, CD4, high le els o su ace CD25, low le els o CD127 and in acellula FoxP3 (6). T egs can be u he di ided in o h ee main ma u a ion ca ego ies: naï e, cen al memo y (CM) and e ec o memo y (EM) (7). Naï e T egs exp ess high le els o CD45RA and low le els o CD45R0 and FoxP3, whe eas memo y T egs a e FoxP3 hi and CD45RA - (8, 9). In heal hy indi iduals up o 30% o he egula o y T cells a e naï e (10). As indi iduals age, he ela i e ac ion o naï e T egs dec eases due o hymic in olu ion, while he p opo ion o memo y T egs inc eases (11). Ne e heless, ecen hymic emig an s (RTE), which co-exp ess CD31 and CD45RA, a e s ill p esen in adul s cons i u ing up o 11% o all nai e T egs. I has been epo ed ha he T eg popula ion in adul pe iphe al blood con ains up o 22 pheno ypically dis inc subpopula ions, hus o e ing new insigh s in o he he e ogenei y o hese cells (12). Ne e heless, ex ensi e compa isons be ween T eg in adul pe iphe al blood and umbilical co d blood (CB) ha e no p e iously been unde aken. P esumably, CB T eg a e mos ly comp ised o a homogeneous popula ion o naï e cells (10,13,14). Howe e , e en hough he majo i y o CB T egs a e naï e, CB also con ains small numbe s o memo y T egs, possibly due o p ena al an igen ac i a ion (15). I has also been p oposed ha ma e nal cells pass he placen a and emain in e al lymph nodes, whe e hese cells induce he de elopmen o e al T egs ha supp ess an ima e nal immuni y (16). The expansion po en ial o CB T eg has been shown o be highe han adul T eg, and expanded CB T eg a e unc ionally ac i e (17). This has led o he use o expanded CB T egs o ea men o GVHD a e allo-HSCT (17). Howe e , he ex en o which CB T egs a e a homogenous popula ion has no been s udied o whe he he ex ensi e he e ogenei y p esen in adul T eg is also p esen in CB T eg. Mass cy ome y by ime o fligh (CyTOF) allowed us o in es iga e CB and adul T eg in unp eceden ed de ail by simul aneously de ec ing and quan i ying 35 ma ke s in indi idual cells (18–20). To quan i y he e ogenei y and p o ide mo e insigh in o he pheno ype o CB T eg, we used FlowSOM and ACCENSE o high dimensional analysis o mass cy ome y da a. These analy ic ools allow us o quan i y human T eg he e ogenei y based on exp ession o a la ge se o ac i a ion, p oli e a ion, issue-homing and unc ional ma ke s in conjunc ion wi h s ages o T eg ma u a ion and di e en ia ion. These ools e ealed he e ogeneous popula ions o T eg in bo h CB and adul blood bu CB T eg we e less he e ogeneous wi h espec o ma u i y and unc ional ma ke s. A e allo- HSCT he numbe o dis inc T eg subpopula ions g adually inc eased du ing a wo-yea ollow-up pe iod. Pa ien s wi h cGVHD had significan ly ewe dis inc T eg subpopula ions based on unc ional ma ke s, p oposing a co ela ion be ween a dis up ed T eg he e ogenei y and cGVHD. METHODS Dono and Pa ien Cha ac e is ics Pe iphe al blood samples we e ob ained om 14 heal hy indi iduals (8 males and 6 emales) wi h a median age o 44 yea s ( ange, 20–69 yea s), wo child en (male o 2 yea s old and emale o 10 yea s old) and om fi e disca ded umbilical co d blood collec ions ( om wo males and h ee emales). We also s udied pe iphe al blood om 10 adul pa ien s who unde wen allogeneic HSCT a he Dana-Fa be Cance Ins i u e and B igham and Women’s Hospi al, Bos on Massachuse s. All ansplan pa ien s ecei ed educed in ensi y condi ioning wi h fluda abine plus busul an ollowed by in usion o unmodified G-CSF mobilized pe iphe al s em cell g a s. No pa ien s ecei ed an i- hymocy e globulin o GVHD p ophylaxis o low-dose in e leukin-2 (IL-2) o ea men o ch onic GVHD. F esh blood samples we e ob ained a 6 di e en ime poin s (0, 1, 3, 6, 12, and 24 mon hs) a e ansplan o du ing cGVHD (6 mon hs). Pa ien s wi h elapse we e no included. W i en in o med consen was ob ained om pa ien s and heal hy dono s p io o sample collec ion, in acco dance wi h he Decla a ion o Helsinki. P o ocol app o al was ob ained om he Human Subjec s P o ec ion Commi ee o he Dana-Fa be /Ha a d Cance Cen e . Sample P epa a ion CB mononuclea cells (CBMCs) and PBMCs we e isola ed om eshly d awn samples by densi y g adien cen i uga ion (Ficoll- Paque PLUS; GE Heal hca e). F eshlyisola edCBMCsandPBMCs om heal hy dono s we e immedia ely used o an ibody s aining. PBMCs om pa ien s we e washed and c yop ese ed in BAMBANKER (Lympho ech) be o e being analyzed. Me al-Tagged Monoclonal An ibodies A panel o 35 me al- agged monoclonal an ibodies was used o analysis o CBMCs and PBMCs. A lis o all an ibodies and Ma os e al. Ma u a ion and He e ogenei y o T egs F on ie s in Immunology | www. on ie sin.o g Janua y 2021 | Volume 11 | A icle 5705502 co esponding me al ags is p o ided in Supplemen a y Table 1. All p e-conjuga ed an ibodies we e pu chased om Fluidigm. All o he an ibodies we e pu chased in ca ie -p o ein- ee PBS and conjuga ed wi h he espec i e me al iso ope using he MaxPAR an ibody conjuga ion ki (Fluidigm) acco ding o he manu ac u e ’s ecommended p o ocol. Me al-labeled an ibodies we e dilu ed o 0.5 mg/ml in Cando PBS An ibody S abiliza ion solu ion (Cando Bioscience GmbH) o long- e m s o age a 4°C. An ibody S aining o Mass Cy ome y CBMCs and PBMCs we e washed wi h MaxPa Cell S aining Bu e (Fluidigm) and blocked wi h Human FcR Blocking Reagen (Mil enyi Bio ec) o 10 minu es a oom empe a u e. Cells we e hen incuba ed wi h all an ibodies a ge ing cell su ace ma ke s o 30 minu es a oom empe a u e and hen washed wice wi h Cell S aining Bu e . A e washing, cells we e fixed wi h Cy ofix Fixa ion Bu e (BD Biosciences) and pe meabilized wi h Phosflow Pe m Bu e III (BD Biosciences) ollowing he manu ac u e ’s ins uc ions. Fixed/pe meabilized cells we e washed wice wi h Cell S aining Bu e and incuba ed wi h all an ibodies a ge ing in acellula an igens o 30 minu es a oom empe a u e. A e s aining wi h in acellula an ibodies, cells we e washed wice wi h Cell S aining Bu e and incuba ed wi h 191/193I DNA in e cala o (Fluidigm) ollowing he manu ac u e ’s ins uc ions. P io o mass cy ome y analysis, cells we e washed wice wi h Cell S aining Bu e and wice wi h MaxPa Wa e (Fluidigm). Mass Cy ome y Cells we e analyzed on a CyTOF 2 mass cy ome e (Fluidigm) a an e en a e o app oxima ely 500 cells/s. To no malize CyTOF da a o e di e en days, EQ Fou Elemen Calib a ion Beads (Fluidigm) we e added in all samples. Resul ing da a we e analyzed wi h so wa e a ailable h ough Cy obank (www. cy obank.o g). To emo e deb is and double s, single cells we e ga ed based on cell leng h and DNA con en as desc ibed by Bendall e al. (21). To in e p e high dimensional single-cell da a p oduced by mass cy ome y, we used a isualiza ion ool based on he iSNE algo i hm ha c ea es a wo-dimensional iew o high-dimensional cy ome y da a a single-cell esolu ion, making i possible o no only isually iden i y in e es ing and a e subse s while p ese ing nonlinea i y, bu also o ga e single- cell e en s ac oss di e en samples (22). Ga ing o Popula ions T eg we e defined by CD25 + FOXP3 + co-exp ession (Figu e 1A). Nai e cells we e ga ed om he T eg popula ion wi h he exp ession o CD45RA + CD62L + , CM as CD45RA - CD62L + , A B D E F C FIGURE 1 | Regula o y T cells om co d blood a e mos ly CD31+ nai e cells. (A) Manual ga ing o T eg based on exp ession o CD4 and CD25 + FOXP3 + co- exp ession. (B) Median pe cen age o T egs in he CD4 + compa men o co d blood and adul PBMC; he e o ba s show he ange. (C) Rep esen a i e isual composi ion o T eg ma u a ion subse s in iSNE maps showing single cell ela ionships; subse s a e di ided by he black line: blue = naï e subse ; g een = Cen al Memo y (CM) subse ; o ange = E ec o Memo y (EM) subse . (D) Rela i e composi ion o Naï e, CM and EM subse s in co d blood and adul T eg. Median alues a e shown o fi e CB and 14 adul T eg samples. (E) Exp ession o CD31 isualized wi h iSNE maps wi hin naï e ac ion o CB and adul T eg. Cells a e colo ed acco ding o in ensi y o CD31 exp ession. (F) Median pe cen age o naï e T eg cells exp essing CD31 + , he e o ba s show he ange. * ep esen s s a is ical significance (p- alue < 0.05). In ViSNE maps, each poin ep esen s a single cell. Ma os e al. Ma u a ion and He e ogenei y o T egs F on ie s in Immunology | www. on ie sin.o g Janua y 2021 | Volume 11 | A icle 5705503 and EM as CD45RA - CD62L - . Recen hymic emig a ion (RTE) cells we e ga ed by co-exp ession o CD45RA and CD31 (23). Each sample was ga ed manually ia Cy obank. Clus e ing Analysis FlowSOM (R package accessible wi hin he www.bioconduc o . o g pla o m) was used o au oma ed clus e ing. FlowSOM was un in R S udio and allows o unsupe ised clus e ing and dimensionali y educ ion o da a ob ained om mass cy ome y. Subsequen subpopula ion analysis was epea ed in ACCENSE (s andalone applica ion accessible om h p://www.cellaccense. com/) o eliabili y pu poses. ACCENSE is a ool o explo a o y analysis o high-dimensional single-cell da a such as ha gene a ed by mass cy ome y (24). By combining a nonlinea dimensionali y educ ion algo i hm ( -SNE) wi h a k-means clus e ing algo i hm bo h isualiza ion o explo a o y analysis and au oma ed cell classifica ion in o subpopula ions is pe o med. Subpopula ion quan ifica ion was comple ed wice based on 26 unc ional ma ke s and 6 ma u i y ma ke s om he cy ome y panel. ACCENSE was also used o make subsequen isual figu e maps. Supplemen a y Table 2 shows which ma ke s we e used. S a is ical Analysis G aphpad P ism 7.04 was used o da a analysis. Mann-Whi ney es was used o compa e unpai ed popula ions. The Wilcoxon signed- ank es was used o compa e pai ed samples o con inuous a iables and exp ession le els o p o eins be ween subpopula ions and be ween di e en ime poin s. All es s we e 2-sided a he significance le el o 0.05 and mul iple compa isons we e no conside ed. RESULTS P e alence o T eg Ma u a ion Subse s a Bi h and Adul hood To compa e he ma u i y o he T eg cell compa men be ween co d blood and adul blood we ga ed he single cell da a biaxially in Cy obank o dis inguish T egs in all samples (Figu e 1A). T eg we e classified by he co-exp ession o in acellula FoxP3 and high-exp ession o su ace CD25 (6). Fo comple e isualiza ion o he ga ing s a egy see Supplemen a y Figu e 1. The median pe cen age o egula o y T cells in adul blood was a 4.6- old highe han co d blood (6.0% s. 1.3%; Figu e 1B)(p< 0.0001). Specific ma u a ion subse s we e defined as ollows: naï e (CD45RA + CD62L + ), CM (CD45RA - CD62L + ), and EM (CD45RA - CD62L - )(7–9). Figu e 1C iden ifies hese h ee dis inc popula ions wi hin he T eg compa men and isually ep esen s he di e ence be ween CBMC and adul PBMC. Naï e T eg cells make up he la ges subse in CB a 85.2%, ollowed by CM cells a 5.6% and EM a 4.5%. In adul PBMC, CM T eg o m he majo subse a 56.5%, ollowed by EM (25.2%) and naï e cells (15.2%) (Figu e 1D). We also quan ified he ac ion o ecen hymic emig an s (RTE) wi hin he naï e T eg subse . RTE can be ga ed om he naï e popula ion ia he exp ession o CD31, seen in Figu e 1E. Wi hin CB T eg, a median o 70.7% o ga ed CD45RA + cells exp essed CD31. In adul PB T eg, a median o 10.19% o ga ed CD45RA + cells exp essed CD31 (p = 0.0002) (Figu e 1F). CD4 T eg Cell He e ogenei y Is Es ablished in Umbilical Co d Blood and Inc eases Wi h Age We hen examined he pheno ypic he e ogenei y o T eg in CB and adul pe iphe al blood by clus e ing T egs based on he exp ession o 26 unc ional ma ke s. To quan i y he e ogenei y wi hin T eg, we used unsupe isedclus e analysiswi h FlowSOM. Webe e al, 2016. p e iously compa ed 13 flow and mass cy ome y clus e ing ools, ecommending ha FlowSOM (wi h op imal me a-clus e ing bu wi hou au oma ic selec ion o numbe o clus e s) be used as a fi s choice o analyzing new da a se s (25). We eplica ed his analysis wi h ou 26-ma ke panel and esul s we e u he alida ed by a sepa a e analysis ool, ACCENSE. Figu e 2A shows T eg cell clus e ing on unc ional ma ke s included in ou panel o 4 CB and 4 adul PB samples. Unsupe ised clus e ing based on all 26 ma ke pa ame e s e ealed a median o 15.5 dis inc clus e s in adul T eg ( ange, 10–22) and 12 T eg clus e s in CB T eg ( ange, 6– 13) (p = 0.008) (Figu e 2B). Al hough he e ogenei y o CB T eg is subs an ial his he e ogenei y inc eases significan ly in adul s. The e was no a ia ion o numbe o clus e s wi hin heal hy con ols ega ding hei age. In ac , we analyzed T eg he e ogenei y in wo child en (2 and 10 yea s old), and bo h had 16 and 18 clus e s, espec i ely (Supplemen a y Figu e 2). This finding sugges s ha T eg he e ogenei y is acqui ed e y ea ly in li e. CD4 T eg Cell Ma u a ion He e ogenei y Also Inc eases Be ween Bi h and Adul hood A e es ablishing ha he e ogenei y o T eg in umbilical co d blood inc eases in adul hood, we u he analyzed pheno ypic he e ogenei y ela ed o le els o T eg ma u i y wi hin naï e, CM and EM T eg subpopula ions. Figu es 3A, B show he ga ing s a egy using CD45RA and CD62L o iden i y he naï e/CM/ EM subse s in a iSNE map. This allowed us o isualize and quan i y he ollowing pheno ypic and unc ional ma ke s in each subse : CD31 ( ecen hymic emig an ma ke ), Ki-67 (ma ke o p oli e a ion), CD95 (ma ke o ex insic pa hway apop osis), and HLA-DR (ac i a ion ma ke ) (Figu es 3C, D). The pe cen age o CD31 + T egs wi hin he naï e subse o CB was significan ly highe han in naï e adul T egs (p = 0.0002). These esul s a e consis en wi h he p e iously es ablished highe exp ession o CD31 in CB cells compa ed o adul T egs. In con as , all 3 ma u a ion subse s o adul T egs exp ess significan ly mo e CD95 han co esponding CB T eg subpopula ions. Al hough ew naï e T egs exp ess CD95, he exp ession o his ma ke was also significan ly di e en (p = 0.01) be ween CB an adul T eg. La ge di e ences we e seen when compa ing CD95 exp ession in CM and EM T eg popula ions in CB and adul s (p = 0.005 and 0.0002, Ma os e al. Ma u a ion and He e ogenei y o T egs F on ie s in Immunology | www. on ie sin.o g Janua y 2021 | Volume 11 | A icle 5705504 espec i ely). This sugges s ha mo e memo y T egs a e suscep ible o apop osis in adul han in CB T egs. Wi h ega d o he ac i a ion ma ke HLA-DR, naï e CB T egs exp essed highe le els o HLA-DR han naï e adul T egs (p = 0.0002). This sugges s ha naï e T egs a e he mos ac i a ed subse wi hin CB. The e was li le di e ence in exp ession o HLA-DR in CM and EM T egs o newbo ns and adul s, pe haps indica ing ha hese subse s ha e simila ac i i y in bo h age g oups. Ki-67 is exp essed a highe le els in all CB T eg subse s, indica ing ha all CB T eg cells p oli e a e a a highe a e han hei adul coun e pa s. Fo each o he subg oups, significan di e ences we e ound, wi h p = 0.0003 o naï e, p = 0.0002 o CM and p = 0.0002 o EM T egs. Di e en ial Exp ession o Ma ke s in Dis inc Popula ions Re eal Func ional Cha ac e is ics E en hough ou me hod o unsupe ised clus e ing did no allow us o di ec ly compa e exp ession o ma ke s in di e en me aclus e s, we analyzed he exp ession o all unc ional ma ke s in CB e sus adul PBMC. F om he 26 unc ional ma ke s (Supplemen a y Figu e 3), eigh showed significan ly di e en le els o exp ession be ween umbilical co d blood and adul pe iphe al blood. CB T eg cells exp essed highe le els o Ki-67 (p = 0.0093) (a nuclea p oli e a ion ma ke ), PD-1 (p = 0.0485) (a ma ke o educed apop osis and exhaus ion), CCR9 (p = 0.0196) (a chemokine ecep o ha egula es lymphocy e a ficking o he small in es ine), and CCR7 (p = 0.0036) (a chemokine ecep o ha egula es lymphocy e a ficking o lymph nodes). On he o he hand, adul s showed highe exp ession le els o CCR4 (p = 0.0196) (a chemokine ecep o ha egula es lymphocy e a ficking o skin), CCR5 (p = 0.0021) and CXCR3 (p = 0.0010) (chemokine ecep o s ha egula es lymphocy e a ficking o inflamed issues), and CD95 (p = 0.0273) (a ma ke o ex insic pa hway apop osis) (Figu e 4). O he unc ional ma ke di e ences be ween CB and adul PBMC ha we e o in e es included CTLA-4 (p = 0.0624), CLA (p = 0.0800), Tbe (p = 0.0800), ICOS (p = 0.0800), and PDL-1 (p = 0.0800), which we e sligh ly lowe in adul T egs. T eg Cell He e ogenei y Inc eases A e Allogeneic Hema opoie ic S em Cell T ansplan a ion Acknowledging he impo an ole o egula o y T cells in allogeneic hema opoie ic s em cell ansplan a ion (alloHSCT) and ollowing ou p e ious s udies desc ibing he econs i u ion o T eg a e alloHSCT (23), we examined he he e ogenei y o T eg subse s a a ious imes a e alloHSCT. Using he same mass cy ome y panel, we analyzed T egs om fi e pa ien s ha ecei ed unmodified pe iphe al blood s em cell g a s a 6 ime poin s: Day 0 and days 1, 3, 6, 12, and 24 mon hs a e ansplan . None o hese pa ien s de eloped acu e o ch onic GVHD du ing his ime pe iod. The pe cen age o CD4 + T eg cells inc eased g adually in he fi s 6 mon hs a e alloHSCT and subsequen ly emained ela i e s able h oughou he 2-yea ollow up pe iod (Figu e 5A). We subsequen ly examined T eg he e ogenei y du ing his pe iod (Figu es 5B, C). Compa ed o T eg p e ansplan , he e was a dec ease in he numbe o T eg subse s 1 mon h a e ansplan . We hen obse ed an inc ease om a median o 9.5 subpopula ions o 14 subpopula ions 3 mon hs a e ansplan . This le el o T eg he e ogenei y emained ela i ely s able om 3 o 24 mon hs a e ansplan in his g oup o pa ien s who did no de elop acu e o ch onic GVHD. In e es ingly, only a ew ma ke s a y hei exp ession no iceably since he ansplan . D49a, GITR, CTLA-4, CXCR3, Tim-3, CCR7, CD28, and CCR4 a e 2- old mo eexp esseda day0compa ed oa e mon h1.F ommon h 3 onwa ds he exp ession o he ma ke s emains cons an among alloHSCT pa ien s and simila o adul T eg cells. T eg Cell He e ogenei y is Reduced Du ing cGVHD P e ious s udies ha e shown ha pa ien s wi h ac i e ch onic GVHD ha e a lowe equency o T eg cells (22,26). To A B FIGURE 2 | Analysis o T eg he e ogenei y by unsupe ised clus e ing based on exp ession o 26 unc ional ma ke s. (A) T eg me aclus e s iden ified by ACCENSE in ou ep esen a i e samples o co d blood (CB) and ou adul T eg samples. Each poin ep esen s one cell, he colo o he cells deno es a specific clus e . The numbe o me aclus e s iden ified in each sample is shown below each box. (B) Pheno ypic he e ogenei y o T eg in CB and adul PB. Fi e CB and 14 adul PB samples we e s udied. The cen e ba in he box is he median; Whiske s illus a e he minimum and maximum alues ob ained. * ep esen s s a is ical significance (p < 0.05). Ma os e al. Ma u a ion and He e ogenei y o T egs F on ie s in Immunology | www. on ie sin.o g Janua y 2021 | Volume 11 | A icle 5705505 de e mine whe he his quan i a i e T eg deficiency was also associa ed wi h abno mal T eg he e ogenei y we examined pe iphe al blood om fi e pa ien s wi h ch onic GVHD and compa ed esul s o samples ob ained 6 mon hs a e alloHSCT om fi e pa ien s wi hou GVHD ha had been p e iously analyzed. Ga ing on CD4+FoxP3+ T eg, we fi s compa ed T eg pe cen ages in pa ien s wi h and wi hou cGVHD. Median %T eg was 4.28 ( ange, 2.03–6.77) in pa ien s wi h cGVHD compa ed o 7.74 ( ange, 4.18–9.27) in pa ien s wi hou cGVHD (p = 0.055). Using FlowSOM, we compa ed T eg he e ogenei y in samples om pa ien s wi h and wi hou cGVHD. T eg om pa ien s wi h cGVHD we e ound o ha e 11 T eg me aclus e s ( ange, 10–13) compa ed o 14 me aclus e s ( ange, 12–21) in pa ien s wi hou cGVHD (p = 0.02) (Figu es 5D, E). Al hough ou me hod o unsupe ised clus e ing did no iden i y specific cha ac e is ics o missing me aclus e s in AB DC FIGURE 3 | T eg ma u a ion in co d blood and adul pe iphe al blood. (A) Exp ession o CD45RA and CD62L in co d blood and adul T eg. Colo indica es he le el o exp ession o he labeled ma ke in iSNE maps om wo ep esen a i e samples. (B) Visual ep esen a ion o T eg ma u a ion in iSNE maps based on exp ession o CD45RA and CD62L. Subse s a e di ided by he black line: blue = naï e T eg; g een = Cen al Memo y (CM) T eg; o ange = E ec o Memo y (EM) T eg. (C) Exp ession o specific unc ional ma ke s in ep esen a i e examples o co d blood and adul T eg. (D) Median pe cen age o ma ke posi i e cells wi hin naï e, CM and EM T eg subse s. Resul s a e compa ed o fi e co d blood T eg samples and 14 adul T eg samples. * ep esen s s a is ical significance (*p < 0.05, **p < 0.005, ***p < 0.0005). In iSNE maps, each poin ep esen s a single cell. Cells a e colo ed acco ding o in ensi y o exp ession o he indica ed ma ke (excluding B). Ma os e al. Ma u a ion and He e ogenei y o T egs F on ie s in Immunology | www. on ie sin.o g Janua y 2021 | Volume 11 | A icle 5705506 cGVHD pa ien s, compa ison o ma ke exp ession ound a significan di e ence in exp ession o Helios and CLA be ween people wi h and wi hou cGVHD (Supplemen a y Figu e 4). Helios was exp essed a highe le els in pa ien s wi hou cGVHD while he exp ession o CLA was highe in cGVHD pa ien s. PD- 1, CCR4 and HLA-DR also we e highly exp essed (non- A B FIGURE 4 | Co d blood and adul pe iphe al T egula o y cells di e in exp ession o a ious ma ke s. (A) iSNE maps show he exp ession o unc ional T eg ma ke s in co d blood and adul PBMC, colo indica es he le el o exp ession o he labeled ma ke . (B) Ba g aphs ma ching he iSNE ep esen a ion o hei le showing median exp ession in ensi y o each ma ke , he e o ba s show he IQ ange. * ep esen s s a is ical significance (p < 0.05). In ViSNE maps, each poin ep esen s a single cell. Cells a e colo ed acco ding o in ensi y o exp ession o he indica ed ma ke . Ma os e al. Ma u a ion and He e ogenei y o T egs F on ie s in Immunology | www. on ie sin.o g Janua y 2021 | Volume 11 | A icle 5705507 significan ) in non-cGVHD pa ien s. Fo CD49a, CCR5, CD62L, BCL-2, and CD39 he opposi e was obse ed, wi h a non- significan bu highe exp ession in cGVHD pa ien s. DISCUSSION Using mass cy ome y and ad anced compu a ional algo i hms o simul aneously measu e exp ession o 35 pheno ypic and unc ional ma ke s in indi idual cells, we unde ook a de ailed analysis o T eg he e ogenei y in humans. Ou s udy ini ially ocused on T eg in umbilical co d blood and pe iphe al blood in heal hy adul s and subsequen ly was expanded o include T egs econs i u ing a e allogeneic HSCT and in pa ien s wi h cGVHD. Unsupe ised clus e ing algo i hms based on 26 unc ional ma ke s we e used o es ablish he numbe o T eg me aclus e s in indi idual samples and he eby quan i y he e ogenei y wi hin he T eg popula ion in each sample a each ime poin . This esul ed in a unique examina ion o T eg he e ogenei y h oughou li e and h ough a c i ical pe iod whe e T eg a e known o play an impo an ole in immune econs i u ion and he es ablishmen o immune ole ance. A bi h, CB T egs a e p edomina ely comp ised o naï e cells (13,14). Genomic di e si y and p oli e a i e capaci y o CB T eg is e y high (17), and his has acili a ed he use o CB-de i ed in i o expanded T eg o ea men o p e en ion o GVHD a e alloHSCT (27). Ou s udies confi med ha CB T eg a e p edomina ely naï e cells, wi h a high ac ion o ecen hymic emig an s. In con as , adul T eg a e mo e ma u e, being p edomina ely CM and EM cells wi h a much smalle ac ions o naï e cells and ecen hymic emig an s. Ne e heless, CB T eg we e ound o be ela i ely he e ogeneous eflec ing a iable s a es o di e en ia ion, ma u a ion and ac i a ion, despi e being p edomina ely naï e cells. T eg he e ogenei y inc eased significan ly in heal hy adul s eflec ing pas exposu es and addi ional le els o di e en ia ion, ma u a ion and ac i a ion in i o. To assess he e ogenei y wi hin defined s ages o T eg ma u a ion we compa ed he exp ession o CD31 ( ecen hymus emig an ), Ki-67 (p oli e a ion), CD95 (apop osis), and HLA-DR ( unc ional ac i a ion) in naï e, CM, and EM T eg. The equency o CD31 + T egs in he naï e CB T eg was significan ly highe han in naï e adul T eg. This likely eflec s dec eased hymic unc ion and inc eased homeos a ic expansion o naï e T cells in adul s. We also obse ed inc eased exp ession o CD95 in naï e, CM and EM subpopula ions in adul T egs compa ed o CB T eg. This likely eflec s highe le els o exhaus ion and e minal di e en ia ion in adul T eg. In con as , HLA-DR was mo e highly exp essed naï e CB T eg and Ki-67 was mo e highly exp essed a all le els o di e en ia ion in CB T eg, indica ing ha all subse s o CB T eg a e ac i a ed and highly p oli e a i e. In pa ien s who unde go alloHSCT, ecipien T cells a e apidly eplaced as dono cells eng a and econs i u e a ully unc ional immune sys em in he ecipien . P ope unc ioning o he dono immune sys em equi es balanced eco e y o egula o y elemen s as well as e ec o cells and he de elopmen o cGVHD can be p edic ed by impai ed eco e y o T eg leading o an abno mally low a io o T eg o con en ional e ec o T cells (28). Enhancemen o T eg eco e y h ough T eg in usions o adminis a ion o low dose IL-2 o selec i ely induce expansion o T eg in i o can p e en o ea cGVHD p og ession (27,29–33). The abili y o manipula e T eg a e alloHSCT and in pa ien s wi h au oimmune diseases has spa ked in e es in he de elopmen o T eg-di ec ed he apies. Howe e , he e has been ela i ely li le A BD EC FIGURE 5 | Regula o y T cell he e ogenei y a e alloHSCT and du ing cGVHD. (A) The median T eg pe cen age wi hin he CD4 + compa men in heal hy dono s, be o e and a e alloHSCT and in pa ien s wi h ch onic GVHD 6 mon hs a e alloHSCT. E o ba s show he ange o alues. (B) T eg me aclus e s a e alloHSCT iden ified by ACCENSE, based on exp ession o 26 unc ional ma ke s. Each poin ep esen s one cell and colo s deno e indi idual clus e s. Resul s shown a e o a ep esen a i e pa ien and he numbe o me acluse s a each ime poin is shown below each figu e. (C) The numbe o T eg me aclus e s iden ified be o e and a e alloHSCT. Resul s shown a e median alues o fi e pa ien s wi hou acu e o ch onic GVHD a e ansplan . E o ba s show ange o alues o each ime poin . (D) T eg me aclus e s in a ep esen a i e pa ien wi h ch onic GVHD 6 mon hs a e alloHSCT. Colo s deno e indi idual clus e s iden ified by ACCENSE. (E) T eg he e ogenei y in pa ien s wi h and wi hou ch onic GVHD 6 mon hs a e alloHSCT. Box and whiske plo s show he numbe o T eg me aclus e s based on exp ession o 26 unc ional ma ke s. The cen e ba in he box is he median. Whiske s illus a e he minimum and maximum alues. Ma os e al. Ma u a ion and He e ogenei y o T egs F on ie s in Immunology | www. on ie sin.o g Janua y 2021 | Volume 11 | A icle 5705508 conside a ion o he po en ial impo ance o unc ional T eg he e ogenei y in addi ion o he abili y o simply inc ease T eg coun s in i o. In ou analysis o T eg he e ogenei y in adul pa ien s a e ansplan we ound ha T eg he e ogenei y eco e ed apidly in pa ien s wi hou acu e o ch onic GVHD. By 3–6mon hsa e alloHSCT, le els o T eg he e ogenei y we e simila o heal hy adul s. Howe e , he numbe o T eg me aclus e s in pa ien s wi h cGVHD a 6 mon hs was significan ly dec eased compa ed o pa ien s wi hou cGVHD. These findings sugges ha lack o unc ional T eg subse s may con ibu e o he de elopmen o cGVHD in addi ion o a simple nume ical deficiency in his se ing. High T cell ecep o (TCR) di e si y has been co ela ed wi h es ablishmen and main enance o sel - ole ance (34)and equi ed o op imal supp essi e unc ion o T eg cells in mu ine models o GVHD (35). Hence, u u e s udies should co ela e he TCR epe oi e di e si y o T eg subpopula ions in o de o de e mine whe he subpopula ions sha e a clonal o igin. In summa y, ou findings e eal conside able he e ogenei y o T eg subse s ha is no de ec ed in ou ine cha ac e iza ion o T eg by flow cy ome y wi h a limi ed se o ma ke s. We also show ha T eg he e ogenei y a ies conside ably among indi iduals and in pa ien s a e alloHSCT. Al hough his he e ogenei y is based on he a iable exp ession o unc ional ma ke s, u he s udies a e needed o es ablish he ex en o which his pheno ypic he e ogenei y eflec s ac ual unc ional di e ences be ween di e en T eg me aclus e s and he abili y o di e en T eg me aclus e s o egula e di e en immune cells and immune ne wo ks in i o. The main limi a ion o ou s udy was he inabili y o define each indi idual clus e and compa e hem be ween samples. Since T eg a e known o be capable o unc ional plas ici y i will also be impo an o examine he s abili y o dis inc me aclus e s in i o and in i o. As T eg di ec ed he apies a e e alua ed in pa ien s wi h au oimmune diseases as well as GVHD, i may be impo an o examine he e ec s o hese in e en ions on T eg he e ogenei y as well as he numbe o ci cula ing T eg. DATA AVAILABILITY STATEMENT The o iginal con ibu ions p esen ed in he s udy a e included in he a icle/Supplemen a y Ma e ial; u he inqui ies can be di ec ed o he co esponding au ho . ETHICS STATEMENT W i en in o med consen was ob ained om pa ien s and heal hy dono s p io o sample collec ion, in acco dance wi h he Decla a ion o Helsinki. P o ocol app o al was ob ained om he Human Subjec s P o ec ion Commi ee o he Dana- Fa be /Ha a d Cance Cen e . W i en in o med consen o pa icipa e in his s udy was p o ided by he pa icipan s’legal gua dian/nex o kin. AUTHOR CONTRIBUTIONS TM designed he esea ch s udies, conduc ed he expe imen s, acqui ed and analyzed he da a, and w o e he manusc ip . MH and AA conduc ed he expe imen s, acqui ed and analyzed he da a, and edi ed he manusc ip . LN analyzed he da a and edi ed he manusc ip . LG and JR designed he esea ch s udies, analyzed he da a, and edi ed he manusc ip . All au ho s con ibu ed o he a icle and app o ed he submi ed e sion. FUNDING This wo k was suppo ed by NIH g an P01CA229092, EADV Resea ch Fellowship, Rene-Tou aine Fellowship and a gene ous con ibu ion om he Fundacão pa a a Ciencia e a Tecnologia (FCT), FCT SFRH/BD/98980/2013. ACKNOWLEDGMENTS The au ho s a e g a e ul o all pa ien s and heal hy dono s who kindly olun ee ed o pa icipa e in his s udy. The au ho s hank he Pasqua ello Tissue Bank in Hema ologic Malignancies o p ospec i e collec ion and p ocessing o se ial blood samples. We hank John Daley, Suzan Lazo, and K is en Leone o excellen assis ance wi h mass cy ome y analysis, and Do een Hea sey and Lau en Ga ny o assis ance ob aining clinical blood samples. 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/fimmu.2020. 570550/ ull#supplemen a y-ma e ial SUPPLEMENTARY FIGURE 1 | CyTOF ga ing s a egy. (A) The ga ing s a egy shown was used on all da a p io o downs eam analysis, manual ga ing was pe o med in Cy obank. SUPPLEMENTARY FIGURE 2 | Child en ha e dis inc egula o y T cells subse s based on all 26 unc ional ma ke s. T eg subpopula ion clus e s made wi h ACCENSE showing 2 ep esen a i e samples o child en (2 yea s old and 10 yea s old) pe iphe al blood samples. Each poin ep esen s one cell, he colo o he cells deno es a specific clus e . SUPPLEMENTARY FIGURE 3 | Func ional ma ke exp ession compa ison o co d blood and adul PBMC. (A) Hea maps wi h ma ke exp ession om low (black) o high (yellow). (B) Ba g aphs showing he median in ensi y ange. * ep esen s s a is ical significance (p- alue < 0.05). SUPPLEMENTARY FIGURE 4 | Func ional ma ke disc epancies be ween AlloHSCT and cGVHD. Samples a anged by median in ensi y and compa ing alloHSCT 6 mon hs a e ansplan s and cGVHD samples 6 mon hs pos - ansplan , he e o ba s show he IQ ange. * ep esen s s a is ical significance (p- alue < 0.05). Ma os e al. Ma u a ion and He e ogenei y o T egs F on ie s in Immunology | www. on ie sin.o g Janua y 2021 | Volume 11 | A icle 5705509