OECAC

LPSRN

IImh

63

H©P

verlap Between Molecular Markersxpressed by Naturally OccurringD4�CD25� Regulatory T Cells andntigen Specific CD4�CD25� andD8�CD28� T Suppressor Cells

uigi Scotto, Afzal Jamal Naiyer, Sara Galluzzo,aola Rossi, John Sanil Manavalan,eunghee Kim-Schulze, Jianshe Fang,iccardo Dalla Favera, Raffaello Cortesini, and

icole Suciu-Foca

Cctiwfrcuic1p

ABSTRACT: Alloantigen specific CD8�CD28� T sup-pressor (TS) cells differ from naturally occurringCD4�CD25� T-regulatory (natural TR) cells not only bytheir phenotype but also by their mechanism of action.Natural TR have been extensively studied, leading to theidentification of characteristic “molecular markers” suchas Forkhead box P3 (FOXP3), glucocorticoid-inducedtumor necrosis factor receptor–related protein (GITR) andcytotoxic T lymphocyte–associated antigen 4 (CTLA-4).We have investigated the expression of these genes inalloantigen specific TS and CD4�CD25� T regulatory(TR) cells and found that they are expressed at levelssimilar to those observed in natural T . Furthermore,

Rsimilar to natural CD4�CD25� TR, antigen-specific E

TC cytotoxic T cells

TTTTT

ttcsito

a05-6941; Fax: (212) 305-3429; E-mail: ns20@columbia.edu.

Received August 25, 2004; accepted September 9, 2004.

uman Immunology 65, 1297–1306 (2004)American Society for Histocompatibility and Immunogenetics, 2004

ublished by Elsevier Inc.

D8�CD28�CD62L� TS cells have more suppressiveapacity than CD8�CD28�CD62L� TS cells. In spite ofhese similarities, natural TR are not antigen-specific andnhibit other T cells by T cell–to–T cell interaction,hereas TS are antigen-specific and exert their inhibitory

unction by interacting with antigen-presenting cells andender them tolerogenic to other T cells. The molecularharacterization of TS cells may contribute to a betternderstanding of mechanisms involved in inhibition ofmmune responses in autoimmunity, transplantation, andhronic viral infection. Human Immunology 65,297–1306 (2004). © American Society for Histocom-atibility and Immunogenetics, 2004. Published by

lsevier Inc.

ABBREVIATIONSAPC antigen-presenting cellsCTL cytolytic T-lymphocyteDC dendritic cellsEC endothelial cellsILT immunoglobin-like transcript

CL T cell lineCR T cell receptorH T-helper cellsR regulatory T cells

T suppressor

S

NTRODUCTIONn recent years, the concept that peripheral tolerance isaintained by T cells with immunoregulatory function

as reemerged. Research in this area is driven by the need

From the Department of Pathology, Columbia University, New York, NY.Address reprint requests to: Dr. Nicole Suciu-Foca, Columbia University,

30 West 168 Street–P&&S 14-401, New York, NY 10032; Tel: (212)

o develop strategies that can prevent autoimmunity,ransplant rejection, and progression of malignant andhronic viral diseases. It is apparent that although, inome pathologic conditions, the induction of efficientmmune responses is crucial to survival, in others, induc-ion of antigen-specific tolerance prevents developmentf the disease.

In both humans and rodents the most extensively char-

cterized regulatory T (TR) cells are the thymus-derived

0198-8859/04/$–see front matterdoi:10.1016/j.humimm.2004.09.004

Cbapnedattlid[C[hfttctfddhTg

tlNItn

stacb(c

rr2hrimDip

MFtT

ahWoCm

MGPwdac[pCuaCCbC

PUppCcDlB1rT3

CEmAlbs9w

1298 L. Scotto et al.

D4�CD25� natural TR. Depletion of these cells in new-orn mice leads to spontaneous development of variousutoimmune diseases, whereas reconstitution of the de-leted population prevents the development of autoimmu-ity [1–6]. Similarly, elimination of CD4�CD25�T cellsnhances allograft rejection and tumor immunity in ro-ents, whereas expansion of CD4�CD25� TR suppresseslloreactivity [1–6]. CD4�CD25� natural TR were showno be anergic, yet they require T cell receptor (TCR)riggering to suppress antigen-induced activation and pro-iferation of other CD4� or CD8� T cells. Natural TR actn an antigen-nonspecific manner by a cell contact–depen-ent mechanism that involves T cell–to–T cell interaction1–7]. The existence of a similar subset of anergicD4�CD25� TR has also been documented in humans

8–10]. Similar to the situation encountered in mice,uman CD4�CD25� natural TR have no antigen speci-icity and inhibit T-cell proliferation in response to alloan-igens, phytohaemagglutinin (PHA), and immobilized an-i-CD3 monoclonal antibodies (mAb) [8–10]. The mostharacteristic feature of murine and human natural TRishe expression of the forkhead/winged helix transcriptionactor FOXP3, a key regulatory gene involved in theevelopment of natural TR [11–13]. It has been recentlyemonstrated that unprimed CD4�CD25� T cells fromuman peripheral blood can be converted into FOXP3�

R by TCR triggering and exposure to transformingrowth factor-beta (TGF-�) [14, 15].

Other nonantigen-specific regulatory T-cell subtypeshat have been described include: CD4�CD25� inter-eukin (IL)-10 producing TR1 cells, gamma-delta T cells,K1.1 T cells, CD8�CD25�, and CD4�CD8� T cells.

t is still unknown whether these regulatory T-cell sub-ypes share molecular markers, such as FOXP3, withatural TR [16–21].

Regulatory T-cell populations that display antigenpecificity and exert their function by conditioning an-igen presenting cells (APC) to become tolerogenic havelso been described [22–30]. Included in this latterategory are human CD8�CD28�, major histocompati-ility complex (MHC) class I restricted, T suppressorTS) cells [22–29] and CD4�CD25�CD45RO�, MHClass II restricted, TR cells [30].

In previous studies, we have shown that MHC-allo-estricted TS and TR can be generated in vitro by multipleounds of T-cell stimulations with allogeneic APC [22,5, 26, 28]. Evidence that TS and TR also develop in vivoas emerged from studies of rejection-free organ allograftecipients [28–30]. TS and TR act directly on APC,nducing the upregulation of inhibitory receptors, im-unoglobulin-like transcript (ILT)3 and ILT4 [28–30].endritic cells (DC) and endothelial cells (EC), which

nteract with allospecific TS, become tolerogenic to

rimed and unprimed T-helper (TH) cells that recognize e

HC class II antigens on their membrane [28, 30, 31].urthermore, ILT3high ILT4high DC and EC were showno promote the differentiation of CD8� TS and CD4�

Rcells [28, 30, 31].To gain insight into the common denominators of

ntigen specific and nonspecific regulatory T cells, weave analyzed their characteristics at the molecular level.e now demonstrate that there is an almost complete

verlap between natural TR and alloantigen-specificD8�CD28� TS in terms of expression of moleculararkers considered to be specific for natural TR.

ATERIALS AND METHODSeneration of Allospecific T-Cell Lineseripheral blood mononuclear cells from healthy donorsere primed with either gamma-irradiated CD2-epleted APC from an allogeneic donor or with irradi-ted monomyelocytic DC-like cells from the KG1ell line (ATCC, Manassas, VA) as previously described22, 28, 30]. After multiple stimulations with allogeneicriming cells, CD8�CD28�, CD8�CD28�,D4�CD25�, and CD4�CD25� T cells were isolatedsing magnetic beads (Miltenyi Biotech, Auburn, CA),s previously described [22, 28, 30]. CD62L� andD62L� T-cell subsets were separated from allospecificD8�CD28� T cells using goat–anti-mouse magneticeads (Dynal, Lake Success, NY) coupled with mAb toD62L (BD Pharmingen, San Diego, CA).

roliferation Assaysnprimed CD4�CD25� T cells (5 � 104) from fresheripheral blood were stimulated in 5-day mixed lym-hocyte culture (MLC) with allogeneic DC (1.25 � 104).D8�CD28� TS cells (5 � 104) or CD4�CD25� TRells (5 � 104) from T cell line (TCL) primed to the sameC were added to triplicate reactions. Cultures were

abeled with 3H thymidine (1 �Ci/well) (Amershamiosciences, Piscataway, NJ) on day 5 and harvested after8 hours. Percent inhibition of 3H thymidine incorpo-ation in TH primed to allogeneic DC in the presence ofS and TR was calculated as previously described [22, 28,0].

ytolytic T-Lymphocyte Assayspstein-Barr virus–transformed B cell lines, sharing hu-an leukocyte antigen (HLA) class I antigens with thePC used for TCL priming, served as targets for cyto-

ytic T-lymphocyte (CTL) assays. Target cells were la-eled with 100 �Ci of Na2

51CrO4, (Amersham Bio-ciences) and plated at a concentration of 5 � 103/well in6-well plates. Serial twofold dilutions of effector cellsere added to a constant number of target cells to obtain

ffector to target cell ratios of 100:1, 50:1, 25:1, and

1ata[

G

TwiprwmttCiwTeclqmpt(eDtCLwRLf(tcs(fra1

MWWt5ww

TrsvaaCfcddissctgctpGf

STCTaCisFAriviffpCA#GbCApC2ATC

1299Molecular Markers in Regulatory T Cells

2.5:1. After 4 hours of incubation at 37 °C in a 5% CO2

tmosphere, the amount of 51Cr released in the superna-ant was measured using a gamma scintillation counternd percent lysis was calculated as previously described22].

eneration of Gene Expression Profiles

cells from the peripheral blood of different respondersere stimulated two times (at 7-day intervals) with

rradiated KG1 cells. At the end of the second round ofriming, CD8� T cells were isolated from these lines andestimulated for 24 hours with KG1. The KG1 cellsere then removed using mAb to CD34 conjugated toagnetic beads (Dynal). The CD28� and CD28� frac-

ions were isolated and tested for suppressive and cyto-oxic activity as previously described [22]. CD8�

D28� T cells, which had no cytotoxic activity, yetnhibited CD4� TH proliferation to KG1 by �80%,ere used as TS. CD8�CD28� T cells from the sameCL, which exhibited �30% cytotoxic activity at a 25:1ffector to target cell ratio, were used as cytotoxic T (TC)ells. CD8�CD28� TS and CD8�CD28� TC cells iso-ated from five independent lines and fulfilling the re-uirements discussed, were subjected to oligonucleotideicroarray analysis [32]. Briefly, RNA samples were

repared using the TRIzol reagent (Invitrogen Corpora-ion, Carlsbad, CA) and the Rneasy RNA extraction kitQIAGEN Inc, Valencia, CA) following the manufactur-rs’ recommendations. Double-stranded complementaryNA was synthesized from 6 �g of total RNA according

o Affymetrix methodology (Affymetrix, Santa Clara,A) and correspondent cDNA was purified with Phaseock Gels (Eppendorf, Hamburg, Germany). cDNAere then used as template to synthesize biotin-labeledNA with the BioArray High Yield RNA Transcriptabeling Kit (Enzo, New York, NY). Hybridizationrom biotinylated cRNA to human genome GeneChipsU95A Affymetrix) was performed in accordance withhe manufacturer’s instructions. The hybridized genehips were stained with streptavidin-phycoerythrin andcanned using the Hewlett-Packard gene array scannerPalo Alto, CA). MicroArray Suite, version 5.0, (Af-ymetrix) was used to calculate average differences, logatio, and absolute call from the scanned images. Theverage difference of each experiment was normalized to00 to allow comparison among multiple arrays.

icroarray Data Analysise used the software platform Genes@Work (IBM, TJatson Research Center, Yorktown, NY) to analyze

he 10 data sets (5 sets of CD8� CD28� TS cells andsets of CD8�CD28� TC cells; each pair of TS and TC

ere from the same TCL). An unsupervised analysis

as first performed to determine whether the two G

-cell subsets (TS and TC) clustered separately withegard to their global gene expression, followed by aupervised analysis to identify candidate genes. Super-ised clustering allows the identification of genes thatre differentially expressed in two cell types definedpriori according to a given criterion (the expression ofD28 in our study). Data are presented in a matrix

ormat. Each row represents a single gene, and eacholumn an experimental sample. The ratio of the abun-ance of transcripts of each gene to the median abun-ance of the gene’s transcript is represented by a colorn the corresponding sample in the matrix. Greenquares indicate transcript levels below the median; redquares indicate transcripts levels above the median. Arucial requirement for the identification of differen-ially expressed genes is the consistent expression of aiven gene across all five samples of a given subset. Toonfirm the microarray data, semiquantitative reverseranscriptase–polymerase chain reaction (RT-PCR) waserformed using a Mastercycler (Eppendorf, Hamburg,ermany) and specific primers to amplify fragments

rom the different genes analyzed.

emiquantitative RT-PCRotal RNA was extracted from CD8�CD28�TS,D8�CD28� TC, CD4�CD25� TR, and CD4�CD25�

cells isolated from allospecific TCL. Total RNA waslso extracted from CD4�CD25� natural TR,D4�CD25�, CD8�CD28�, and CD8�CD28� T cells

solated from fresh peripheral blood using the corre-ponding magnetic beads as previously described [22].or RNA extraction, the RNAqueous kit (Ambion, Inc.,ustin, TX) was used according to the manufacturer’s

ecommendation. First-strand cDNA was synthesized us-ng SuperScript II RNase H� Reverse Transcriptase (In-itrogen Corporation). The following primers were usedn PCR reactions: glucocorticoid-induced tumor necrosisactor receptor (TNFR)–related protein (GITR) primersrom R&D System (Minneapolis, MN); FOXP3: 5=rimer #1 TTGGACAAGGACCCGATGCCCAAC-CC, 3= primer #2 CCCTGGCAGGCAAGACAGTGG-AACCTC (expected size, 1350 or 1450 bp); 5= primer3 TCCCAGAGTTCCTCCACAAC, 3= primer #4CAAGACAGTGGAAACCTCAC (expected size, 465p); LYN: 5= primer AAGGCAGAAGAGAGAC-AACGTTT, 3= AGAATAGATGTTTTCTTCTCA-ACGGC (expected size, 383 bp); KIR2DL3: 5=rimer CTGGAATCTGAAGGCGTGAGTC, 3= primerACGTGTCTAAGTGCCGTGTTAA (expected size,87 bp); KIR3DL1: 5= primer AGCCCTGTCTCA-AACCGAGTT, 3= primer TGGGTAAGTGCCACG-CAAGA (expected size, 323 bp); KIR3DL2: 5= primerTCTTCCTCACACCACGAATC, 3= primer CAA-

TGTGTAAGTGCCGTGTT (expected size, 250 bp);

C3sCGG3TGTppsCGTT(GCb(vs

FF(DcwmcdCcC

ch1LCtePie

dC(Io

mssu

RCCIleo

Ceskrt3CT

CcKHiC1Cc[

Veiap

Cfo(

CCCTng

1300 L. Scotto et al.

TLA4: 5= primer CATAGCAGTGCTTGATTGCGT,= primer TAAGAATTGGGCCCATCGAAC (expectedize, 217 bp); OX40: 5= primer GTGGGGCCCAG-ATAACATAC, 3= primer GAACAGCTTAAGTGCA-ACGAGA (expected size, 282 bp); 4-1BB: 5= primerCTTTGGGACATTTAACGATCAGA, (expected size69 bp) 3= primer GAGTTTCTTTCTGCCCCGTT-AAC; CD103: 5= primer GAACCACAGAACTAA-ATCACTGTCGT, 3= primer AGCAGGTCCTAAT-CTCTTCTTCGAG (expected size, 231 bp); CD25: 5=rimer AATCAAAGGTGCTAAATGGTCGC, 3=rimer TTTATTAGGCAACGTGAACGGG (expectedize 459 bp); CD62L: 5= primer GATCCTTTAAATC-TTCCATGAAACG, 3= primer CTGGAGTTAGAAA-AAAGGAGAGCGTA (expected size, 463 bp);NFR2: 5= primer GTGTGTGTAGCCAAGGTCGG-AA, 3= primer CACAGAGTCTCCAAATTCATCGC

expected size, 673 bp); and GAPDH: 5= primer CG-AGTCAACGGATTTGGTCGTAT, 3= primer AGC-TTCTCCSTGGTGGTGAAGAC (expected size, 362p). PCR reactions were performed in a MastercyclerEppendorf, Hamburg, Germany) at 30 cycles in a 20 �lolume; PCR products were analyzed in agarose geltained with ethidium bromide.

low Cytometrylow cytometry studies were performed with a FACScanBecton Dickinson). CaliBRITE beads from Bectonickinson were run under the FACSComp program to

alibrate the instrument [30]. CD8� T cells were stainedith the following antibodies: phycoerythrin-conjugatedAbs to IL-10, TGF-�, and interferon (IFN)-�, fluores-

ein isothiocyanate–conjugated mAb to CD62L, per-inin chlorophyll protein to CD28 and allophycocymin-D8 all from BD Pharmingen. Other mAbs used toheck for purity include phycoerythrin-CD16, CD56,D14, and CD19.Intracellular levels of IL-10, TGF-�, or IFN-� in T

ells from alloantigen-specific TCL were measured after 6ours of incubation with 25 ng/ml phorbol 12-myristate3 acetate and 1 �M ionomycin, both from Sigma (St.ouis, MO). Cells were stained with allophycocymin-D8, perdinin chlorophyll protein-CD28 fluorescein iso-

hiocyanate conjugated- CD62L (BD Pharmingen), andither phycoerythrin-, IFN-�, �GF-�, or IL-10 (BDharmingen) using cytoperm/cytofix intracellular stain-

ng kit (BD Pharmingen) according to the manufactur-r’s instructions.

ILT3 and ILT4 expression on CD14� monocytes wasetermined using fluorescein isothiocyanate conjugated-D14 (BD Pharmingen), R-phycoerythrin-cyanine

PC5) conjugated anti-human ILT3 (Beckman-Coulternc., Miami, FL). ILT4 staining was performed as previ-

usly described [28, 30]. 3

For each cell surface marker, a corresponding isotype-atched control antibody that was conjugated with the

ame fluorescent dye was used. Six parameters (forwardcatter, side scatter, and four fluorescence channels) weresed for list mode data analysis.

ESULTSomparison of CD8�CD28� With CD8�CD28� Tells From Allospecific TCL

n an attempt to identify markers characteristic for al-ospecific CD8�CD28� TS, we first compared their genexpression profile (mRNA microarray profiles) with thatf allospecific CD8�CD28� TC cells from the same TCL.

Affymetrix gene chip analysis of CD8�CD28� TS andD8�CD28� TC from five different TCL showed differ-nces in the level of expression of genes that encode cellurface molecules, signal transduction molecules, chemo-ines, cytokines, apoptosis-related proteins, cell growthegulators, and metabolic enzymes. Of 12,000 genesested, 72 were differentially expressed. Of these 72 genes,4 had higher and 38 had lower expression inD8�CD28� TS cells compared with the CD8�CD28�

C cells.Among the genes with higher expression in the

D8�CD28� TS cells, three were members of the killerell immunoglobulin-related receptors (KIR) family:IR3DL1 (NKAT3, CD158E2), which is specific forLA-B antigens; KIR3DL2 (NKAT4, CD158K), which

s specific for HLA-A antigens; and KIR2DL3 (NKAT2,D158B2), which is specific for HLA-C antigens (FigureA) [33]. The higher expression of KIR genes inD8�CD28� TS compared with CD8�CD28� TC isonsistent with the fact that TS have no cytolytic activity19].

The oncogene LYN (V-YES-1 Yamaguchi Sarcomairal Related Oncogene Homolog), a tyrosine kinase

ssential for establishing immunoreceptor tyrosine-basednhibitory motif (ITIM)–dependent signaling [34], waslso expressed at higher levels in CD8�CD28� TS com-ared with CD8�CD28� TC cells.

The increased expression of these genes inD8�CD28� TS compared with CD8�CD28� TC was

urther confirmed by semiquantitative RT-PCR analysisf independent TS and TC samples from other TCLFigures 1A, B).

omparison of Different Markers in NaturalD4�CD25� TR, Alloantigen-SpecificD4�CD25� TR, and CD8�CD28� TS

he intensive search for molecular markers specific foratural TR has generated a growing list of candidateenes that includes CTLA4 [35, 36], FOXP3 [11, 13,

7], L-selectin (CD62L) [38], �7 integrin (CD103)

FppCchsSyKTf

1301Molecular Markers in Regulatory T Cells

IGURE 1 Gene expressionrofiles in CD8�CD28� T sup-ressor (TS) cells andD8�CD28� cytotoxic T (TC)ells. (A) A total of 34 genes hadigher and 38 had lower expres-ion in TS compared to TC. (B)emiquantitative RT-PCR anal-sis of KIR2DL3, KIR3DL1,IR3DL2, and Lyn, in TS andC. GAPDH was used as control

or sample loading.

[G(

4Ca(pcCifaCdecTlam

eTicFdtdspaF

FA

IoFgc(yiFaTCeCe

FUrTft

1302 L. Scotto et al.

39], and members of the TNFR family such as theITR [40, 41], TNFRSF4 (OX-40) [42], TNFRSF9

4-1BB) [43], and TNFR2 [44].The mRNA expression level of GITR, CTLA-4, OX-

0, CD103, FOXP3, and TNFR2 was similar in naturalD4�CD25� TR, alloantigen-induced CD4�CD25�,nd CD8�CD28� regulatory T-cell subsets, respectivelyFigure 2). TS showed lower expression of CD25 com-ared with natural and alloantigen-induced TR. Allospe-ific CD4�CD25� TR showed lower expression ofD62L compared with natural CD4� TR and antigen-

nduced CD8� TS. Unprimed CD8�CD28� T cellsrom fresh peripheral blood, which have no regulatoryctivity, do not express FOXP3, GITR, OX40, CD25,D62L, and 4-1BB. However, these unprimed T cellsisplay low levels of CTLA-4 and TNFR2. CD103 wasxpressed at the same level in unprimed CD8�CD28� Tells and in primed regulatory cells. Thus CD8�CD28�

S differ from natural CD4�CD25� TR only by theower mRNA expression level of CD25. CD4�CD25�

ntigen-induced TR differ from natural TR by the lowerRNA expression level of CD62L.Because unprimed CD8�CD28� T cells from periph-

ral blood do not express FOXP3, whereas CD8�CD28�

S from allostimulated TCL are FOXP3-positive, wenvestigated the kinetics of FOXP3 expression after T-ell allostimulation in cultures. Daily monitoring ofOXP3 expression revealed its absence during the first 7ays after the primary stimulation. FOXP3 became de-ectable 2 days after the second round of stimulation, onay 9, and its expression peaked on day 14 (data nothown). Stimulation of CD8�CD28� T cells from fresheripheral blood with phorbol 12-myristate 13 acetatend ionomycin for 3 days did not result in inducedOXP3 expression in the stimulated T cells.

OXP3� Expression in Natural TR andntigen-Specific TS and TR

n previous studies, we have described the identificationf a new isoform of FOXP3, which we referred to asOXP3 [31]. The new isoform, FOXP3, lacks a re-ion of 105 nucleotides from position 213 to 317, whichorresponds to the entire exon 3 of the FOXP3 geneGenBank Accession #014009). FOXP3 expression anal-sis, by RT-PCR, in unprimed and primed T-cell subsetsndicated that FOXP3 was expressed together withOXP3 in unprimed CD4�CD25� natural TR andlloantigen-specific CD8�CD28� TS and CD4�CD25�

R. However, CD4�CD25�, CD8�CD28�, andD8�CD28� T-cell subsets isolated from fresh periph-ral blood as well as CD4�CD25� T cells andD8�CD28� TC isolated from TCL did not express

ither FOXP3 or FOXP3 (Figure 3). c

f

IGURE 2 Analysis of gene expression in T-cell subsets.nprimed CD8�CD28�, naturally occurring CD4� CD25�

egulatory T cells (TR), and alloantigen-specific CD8�CD28�

suppressor cells and CD4� CD25� TR cells were examinedor the expression of candidate regulatory genes by semiquan-itative reverse transcriptase–polymerase chain reaction. Toontrol for sample loading, GAPDH cDNA was amplified

rom the same RNA sample.

SCRwpeCeqfsCwaa1smCdsthnCCF

CiTCfbtC

ra

TitwCACfmrIu

DTc

FsCcssl

Fpirss

1303Molecular Markers in Regulatory T Cells

tudy of the Suppressor Activity of CD8� CD28�

D62L� TS

ecent studies have shown that CD4�CD25� natural TRhich express CD62L display a significantly stronger sup-ressor activity than CD4�CD25�CD62L� TR [38]. Wexplored the possibility that a similar association betweenD62L expression and increased suppressive activity mayxist within the CD8�CD28� Ts population. The fre-uency of CD62L�T cells within CD8�CD28� Ts fromive different TCL ranged from 64% to 77% as mea-ured by flow cytometric analysis (data not shown).D8�CD28� CD62L� and CD8�CD28� CD62L� TSere sorted from the same TCL (A primed to B) and tested

t different concentrations for their effect on the prolifer-tive response of autologous CD4�CD25� T cells (5 �04) to allogeneic dendritic cells (1 � 104/well) fromtimulator B used for priming. CD4� TH proliferation waseasured after 5 days. Although both the CD62L� andD62L� subsets inhibited CD4� TH alloreactivity in aose-dependent manner, CD62L� TS were more potentuppressors on a per-cell basis (Figure 4A). Addition tohe cultures of various concentrations of mAb to CD62Lad no effect on suppression indicating that CD62L isot directly involved in the inhibitory activity ofD8�CD28� TS cells. RT-PCR studies of CD62L� andD62L� TS showed no differences in mRNA expression ofOXP3, FOXP3, GITR, and CTLA4 (data not shown).

Flow cytometric analysis of the CD8�CD28�

D62L� and CD8�CD28� CD62L� subsets showed nontracellular expression of IL-10, TGF-�, or IFN-�.ranswell assays in which mixtures of CD62L� orD62L� TS and APC (upper chamber) were separated

rom the mixtures of CD4�, TH, and APC (lower cham-er) showed no inhibition of TH proliferation by either ofhese subsets, indicating that both CD62L� and

IGURE 3 Expression analysis of FOXP3 in unprimed andrimed T-cell subsets. Fractionated T-cell subsets were exam-ned for FOXP3 and FOXP3 expression by semiquantitativeeverse transcriptase–polymerase chain reaction. To control forample loading, GAPDH was amplified from the same RNAample.

D62L� TS subsets require cell contact to suppress. This T

esult is consistent with previous studies of unfraction-ted CD8�CD28� allospecific TS [23].

Because we have previously shown that CD8�CD28�

S act directly on APC, inducing the upregulation of thenhibitory receptors ILT3 and ILT4 and the downregula-ion of costimulatory molecules (such as CD80 and CD86),e compared the capacity of CD8�CD28�CD62L� andD8�CD28�CD62L� cells to induce such changes inPC (monocytes and DCs). When CD8�CD28�

D62L� TS were coincubated with CD14� monocytesrom the allostimulating donor, they were consistentlyore potent inducers of tolerogenic changes such as down-

egulation of CD86 (data not shown) and upregulation ofLT3 and ILT4 than CD8�CD28� CD62L� T cells (Fig-re 4B).

ISCUSSIONhe resurgence of interest in regulation of antigen-spe-ific immune responses has led to the identification of

IGURE 4 CD8�CD28�FOXP3�CD62L� T cells displaytronger suppressor activity compared with CD8�

D28�FOXP3�CD62L� T cells. (A) Inhibition of T-helperell proliferation by CD62L� and CD62L�CD8�CD28� T-uppressor cells. (B) Induction of immunoglobulin-like tran-cript (ILT)3 and ILT4 on allostimulatory monocytes by al-oantigen-specific CD8�CD28�CD62L� and CD62L�

-suppressor cells.

Tcopi

acCCdCeAcuogmsshw3i3Cc

CafitcbaopocKcpoccadtvtit

eTd

rttttaIptCttfst

A

T1t

R

1304 L. Scotto et al.

-cell subsets that have suppressive properties. The bestharacterized populations of regulatory cells are naturallyccurring CD4�CD25� TR from which a growing list ofhenotypic markers, presumed to allow the unequivocaldentification of regulatory cells, has been derived [45].

In the present study, we demonstrate that these genesre expressed at relatively equal levels in naturally oc-urring CD4�CD25� TR, alloantigen-specificD8�CD28� TS, and CD4�CD25� TR cells. Of notice,D4�CD25� natural TR exert their function by actingirectly on other T cells, whereas antigen-specificD8�CD28� TS and CD4�CD25� TR acquire andxert their function via direct interaction with APC.ntigen-specific TS and TR recognize MHC peptide

omplexes on the cell surface of APC and trigger thepregulation of inhibitory receptors and downregulationf costimulatory molecules, rendering the APC tolero-enic [28, 30]. This implies that, regardless of theirechanism of action, T cells with contact-dependent

uppressor activity show elevated expression of the sameet of genes. This set of genes include CTLA-4 a coin-ibitory molecule that belongs to the B7-CD28 family;hich is involved in the control of T-cell immunity [35,6]; FOXP3, which encodes a transcription repressornvolved in the generation and function of TR [11, 13,7]; and genes encoding cell adhesion molecules such asD62L [38] and CD103, which bind to epithelialadherins [39].

Comparison of global gene expression inD8�CD28� and CD8�CD28� T-cell subsets suggestsrole for some members of the KIR family in the

unction of the CD8� TS subset. The KIRs regulate thenhibition and activation of natural killer cell responseshrough recognition of HLA class I molecules on targetells. Some of the KIRs receptors (i.e., KIR2DL3) haveeen previously found on a subset of CD8� T cells withmemory phenotype [46–48]. The biologic significancef KIRs has been elucidated only recently. In vitro ex-eriments using T-cell clones showed that the expressionf KIRs inhibits the effector functions of cytotoxic Tells. In vivo studies suggest that the biologic function ofIRs resides in downregulation of activation-induced

ell death and promotion of cell survival [47, 48]. It isossible, therefore, that the specific upregulation of somef the KIRs in allospecific CD8�CD28� TS blocks theirytotoxic capacity, accounting for the inability of theseells to kill their target cells. It is also possible that Lyn,tyrosine kinase that is essential for establishing ITIM-ependent signaling and for activation of specific proteinyrosine phosphatases within myeloid cells, is also in-olved in this phenomenon [34]. As with other inhibi-ory KIRs, KIR3DL1, KIR3DL2, and KIR2DL3 harborntracytoplasmic ITIMs that recruit and activate protein

yrosine phosphatases SHP-1 or SHP-2 [34]. The higher

xpression of the tyrosine kinase Lyn in CD8� CD28�

S cells may be required for establishing ITIM-depen-ent signaling through KIRs.

Recent findings suggest that the CD4�CD25� TRepresent a heterogeneous population with respect toheir proliferative and suppressive capacity [38]. Func-ional analysis of CD62L� and CD62L� T cells withinhe CD8�CD28� T-cell population indicates the exis-ence of a similar heterogeneity. Although both CD62L�

nd CD62L� subsets do not produce IL-10, TGF-�, andFN-�, the CD8�CD28� CD62L� TS cells are moreotent suppressors of THproliferation and inducers ofolerogenic changes in APC compared withD8�CD28�CD62L� TS cells. This finding suggests

hat CD62L, an adhesion molecule, strengthens the in-eraction between TS and APC. This study, which is theirst report on molecular markers expressed by antigen-pecific TS and TR, emphasizes the need to further dissecthe mechanisms of immunosuppression.

CKNOWLEDGMENT

his work was supported by grants from the NIH (AI25210-8, AI55234-02) and the Interuniversitary Organ Transplan-ation Consortium, Rome, Italy.

EFERENCES

1. Sakaguchi S, Sakaguchi N, Asano M, Itoh M, Toda M:Immunologic self-tolerance maintained by activated Tcells expressing IL-2 receptor alpha-chains (CD25). Break-down of a single mechanism of self-tolerance causes vari-ous autoimmune diseases. J Immunol 155:1151, 1995.

2. Suri-Payer E, Amar AZ, Thornton AM, Shevach EM:CD4�CD25� T cells inhibit both the induction andeffector function of autoreactive T cells and represent aunique lineage of immunoregulatory cells. J Immunol160:1212, 1998.

3. Sakaguchi S, Sakaguchi N, Shimizu J, Yamazaki S,Sakihama T, Itoh M, Kuniyasu Y, Nomura T, Toda M,Takahashi T: Immunologic tolerance maintained byCD25� CD4� regulatory T cells: their common role incontrolling autoimmunity, tumor immunity, and trans-plantation tolerance. Immunol Rev 182:18, 2001.

4. Itoh M, Takahashi T, Sakaguchi N, Kuniyasu Y, Shimizu J,Otsuka F, Sakaguchi S: Thymus and autoimmunity: pro-duction of CD25�CD4� naturally anergic and suppressiveT cells as a key function of the thymus in maintainingimmunologic self-tolerance. J Immunol 162:5317, 1999.

5. Shevach EM: Regulatory T cells in autoimmunity*. AnnuRev Immunol 18:423, 2000.

6. Asano M, Toda M, Sakaguchi N, Sakaguchi S: Autoimmunedisease as a consequence of developmental abnormality of aT cell subpopulation. J Exp Med 184:387, 1996.

7. Thornton AM, Shevach EM: Suppressor effector function

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

1305Molecular Markers in Regulatory T Cells

of CD4�CD25� immunoregulatory T cells is antigennonspecific. J Immunol 164:183, 2000.

8. Stephens LA, Mottet C, Mason D, Powrie F: HumanCD4(�)CD25(�) thymocytes and peripheral T cells haveimmune suppressive activity in vitro. Eur J Immunol31:1247, 2001.

9. Levings MK, Sangregorio R, Roncarolo MG: Humancd25(�)cd4(�) t regulatory cells suppress naive andmemory T cell proliferation and can be expanded in vitrowithout loss of function. J Exp Med 193:1295, 2001.

0. Jonuleit H, Schmitt E, Stassen M, Tuettenberg A, KnopJ, Enk AH: Identification and functional characterizationof human CD4(�)CD25(�) T cells with regulatory prop-erties isolated from peripheral blood. J Exp Med 193:1285, 2001.

1. Fontenot JD, Gavin MA, Rudensky AY: Foxp3 programsthe development and function of CD4�CD25� regula-tory T cells. Nat Immunol 4:330, 2003.

2. Hori S, Nomura T, Sakaguchi S: Control of regulatory Tcell development by the transcription factor Foxp3. Sci-ence 299:1057, 2003.

3. Khattri R, Cox T, Yasayko SA, Ramsdell F: An essentialrole for Scurfin in CD4�CD25� T regulatory cells. NatImmunol 4:337, 2003.

4. Walker MR, Kasprowicz DJ, Gersuk VH, Benard A, VanLandeghen M, Buckner JH, Ziegler SF: Induction ofFoxP3 and acquisition of T regulatory activity by stimu-lated human CD4�CD25- T cells. J Clin Invest 112:1437, 2003.

5. Chen W, Jin W, Hardegen N, Lei KJ, Li L, Marinos N,McGrady G, Wahl SM: Conversion of peripheralCD4�CD25- naive T cells to CD4�CD25� regulatoryT cells by TGF-beta induction of transcription factorFoxp3. J Exp Med 198:1875, 2003.

6. Groux H, O’Garra A, Bigler M, Rouleau M, Antonenko S,de Vries JE, Roncarolo MG: A CD4� T-cell subset in-hibits antigen-specific T-cell responses and prevents coli-tis. Nature 389:737, 1997.

7. Hayday A, Tigelaar R: Immunoregulation in the tissuesby gammadelta T cells. Nat Rev Immunol 3:233, 2003.

8. Zeng D, Lewis D, Dejbakhsh-Jones S, Lan F, Garcia-Ojeda M, Sibley R, Strober S: Bone marrow NK1.1(-) andNK1.1(�) T cells reciprocally regulate acute graft versushost disease. J Exp Med 189:1073, 1999.

9. Cosmi L, Liotta F, Lazzeri E, Francalanci M, Angeli R,Mazzinghi B, Santarlasci V, Manetti R, Vanini V, Ro-magnani P, Maggi E, Romagnani S, Annunziato F: Hu-man CD8�CD25� thymocytes share phenotypic andfunctional features with CD4�CD25� regulatory thy-mocytes. Blood 102:4107, 2003.

0. Zhang ZX, Yang L, Young KJ, DuTemple B, Zhang L:Identification of a previously unknown antigen-specificregulatory T cell and its mechanism of suppression. NatMed 6:782, 2000.

1. Sutmuller RP, Offringa R, Melief CJ: Revival of the

regulatory T cell: new targets for drug development. DrugDiscov Today 9:310, 2004.

2. Liu Z, Tugulea S, Cortesini R, Suciu-Foca N: Specificsuppression of T helper alloreactivity by allo-MHC classI-restricted CD8�CD28� T cells. Int Immunol 10:775,1998.

3. Ciubotariu R, Colovai AI, Pennesi G, Liu Z, Smith D,Berlocco P, Cortesini R, Suciu-Foca N: Specific suppres-sion of human CD4� Th cell responses to pig MHCantigens by CD8�CD28� regulatory T cells. J Immunol161:5193, 1998.

4. Jiang S, Tugulea S, Pennesi G, Liu Z, Mulder A, LedermanS, Harris P, Cortesini R, Suciu-Foca N: Induction of MHC-class I restricted human suppressor T cells by peptide prim-ing in vitro. Hum Immunol 59:690, 1998.

5. Liu Z, Tugulea S, Cortesini R, Lederman S, Suciu-Foca N:Inhibition of CD40 signaling pathway in antigen present-ing cells by T suppressor cells. Hum Immunol 60:568,1999.

6. Li J, Liu Z, Jiang S, Cortesini R, Lederman S, Suciu-Foca N:T suppressor lymphocytes inhibit NF-kappa B-mediatedtranscription of CD86 gene in APC. J Immunol 163:6386,1999.

7. Colovai AI, Liu Z, Ciubotariu R, Lederman S, Cortesini R,Suciu-Foca N: Induction of xenoreactive CD4� T-cellanergy by suppressor CD8�CD28� T cells. Transplan-tation 69:1304, 2000.

8. Chang CC, Ciubotariu R, Manavalan JS, Yuan J, ColovaiAI, Piazza F, Lederman S, Colonna M, Cortesini R, Dalla-Favera R, Suciu-Foca N: Tolerization of dendritic cells byT(S) cells: the crucial role of inhibitory receptors ILT3 andILT4. Nat Immunol 3:237, 2002.

9. Colovai AI, Mirza M, Vlad G, Wang S, Ho E, Cortesini R,Suciu-Foca N: Regulatory CD8�CD28� T cells in hearttransplant recipients. Hum Immunol 64:31, 2003.

0. Manavalan JS, Rossi PC, Vlad G, Piazza F, Yarilina A,Cortesini R, Mancini D, Suciu-Foca N: High expression ofILT3 and ILT4 is a general feature of tolerogenic dendriticcells. Transpl Immunol 11:245, 2003.

1. Manavalan JS, Kim-Schulze S, Scotto L, Naiyer AJ, Vlad G,Colombo PC, Marboe C, Mancini D, Cortesini R, Suciu-Foca N: Alloantigen specific CD8�CD28� FOXP3� Tsuppressor cells induce ILT3� ILT4� tolerogenic endothe-lial cells, inhibiting alloreactivity. Int Immunol 16:1055,2004.

2. Suciu-Foca Cortesini N, Piazza F, Ho E, Ciubotariu R,LeMaoult J, Dalla-Favera R, Cortesini R: Distinct mRNAmicroarray profiles of tolerogenic dendritic cells. HumImmunol 62:1065, 2001.

3. Moretta L, Biassoni R, Bottino C, Mingari MC, MorettaA: Human NK-cell receptors. Immunol Today 21:420,2000.

4. Harder KW, Parsons LM, Armes J, Evans N, KountouriN, Clark R, Quilici C, Grail D, Hodgson GS, Dunn AR,

Hibbs ML: Gain- and loss-of-function Lyn mutant mice

3

3

3

3

3

4

4

4

4

4

4

4

4

4

1306 L. Scotto et al.

define a critical inhibitory role for Lyn in the myeloidlineage. Immunity 15:603, 2001.

5. Takahashi T, Tagami T, Yamazaki S, Uede T, Shimizu J,Sakaguchi N, Mak TW, Sakaguchi S: Immunologic self-tolerance maintained by CD25(�)CD4(�) regulatory Tcells constitutively expressing cytotoxic T lymphocyte-associated antigen 4. J Exp Med 192:303, 2000.

6. Read S, Malmstrom V, Powrie F: Cytotoxic T lymphocyte-associated antigen 4 plays an essential role in the function ofCD25(�)CD4(�) regulatory cells that control intestinalinflammation. J Exp Med 192:295, 2000.

7. Brunkow ME, Jeffery EW, Hjerrild KA, Paeper B, ClarkLB, Yasayko SA, Wilkinson JE, Galas D, Ziegler SF,Ramsdell F: Disruption of a new forkhead/winged-helixprotein, scurfin, results in the fatal lymphoproliferativedisorder of the scurfy mouse. Nat Genet 27:68, 2001.

8. Fu S, Yopp AC, Mao X, Chen D, Zhang N, Mao M, DingY, Bromberg JS: CD4� CD25� CD62� T-regulatorycell subset has optimal suppressive and proliferative po-tential. Am J Transplant 4:65, 2004.

9. Lehmann J, Huehn J, de la Rosa M, Maszyna F, KretschmerU, Krenn V, Brunner M, Scheffold A, Hamann A: Expres-sion of the integrin alpha Ebeta 7 identifies unique subsetsof CD25� as well as CD25- regulatory T cells. Proc NatlAcad Sci U S A 99:13031, 2002.

0. Shimizu J, Yamazaki S, Takahashi T, Ishida Y, SakaguchiS: Stimulation of CD25(�)CD4(�) regulatory T cellsthrough GITR breaks immunological self-tolerance. NatImmunol 3:135, 2002.

1. McHugh RS, Whitters MJ, Piccirillo CA, Young DA,Shevach EM, Collins M, Byrne MC: CD4(�)CD25(�)

immunoregulatory T cells: gene expression analysis re-

veals a functional role for the glucocorticoid-induced TNFreceptor. Immunity 16:311, 2002.

2. Takeda I, Ine S, Killeen N, Ndhlovu LC, Murata K,Satomi S, Sugamura K, Ishii N: Distinct roles for theOX40-OX40 ligand interaction in regulatory and non-regulatory T cells. J Immunol 172:3580, 2004.

3. Choi BK, Bae JS, Choi EM, Kang WJ, Sakaguchi S, VinayDS, Kwon BS: 4-1BB-dependent inhibition of immuno-suppression by activated CD4�CD25� T cells. J LeukocBiol 75:785, 2004.

4. Annunziato F, Cosmi L, Liotta F, Lazzeri E, Manetti R,Vanini V, Romagnani P, Maggi E, Romagnani S: Pheno-type, localization, and mechanism of suppression ofCD4(�)CD25(�) human thymocytes. J Exp Med 196:379, 2002.

5. Wood KJ, Sakaguchi S: Regulatory T cells in transplan-tation tolerance. Nat Rev Immunol 3:199, 2003.

6. Mingari MC, Vitale C, Cambiaggi A, Schiavetti F, MelioliG, Ferrini S, Poggi A: Cytolytic T lymphocytes displayingnatural killer (NK)-like activity: expression of NK-relatedfunctional receptors for HLA class I molecules (p58 andCD94) and inhibitory effect on the TCR-mediated targetcell lysis or lymphokine production. Int Immunol 7:697,1995.

7. Young NT, Uhrberg M: KIR expression shapes cytotoxicrepertoires: a developmental program of survival. TrendsImmunol 23:71, 2002.

8. Ugolini S, Arpin C, Anfossi N, Walzer T, Cambiaggi A,Forster R, Lipp M, Toes RE, Melief CJ, Marvel J, VivierE: Involvement of inhibitory NKRs in the survival of asubset of memory-phenotype CD8� T cells. Nat Immu-

nol 2:430, 2001.