Immunology Letters 129 (2010) 57–63

Contents lists available at ScienceDirect

Immunology Letters

journa l homepage: www.e lsev ier .com/ locate /

Phenotype markers and cytokine intracellular production by CD8+ �� Tlymphocytes do not support a regulatory T profile in Behcet’s disease patientsand healthy controls

Antonio Clementea, Ana Cambraa, Iván Munoz-Saáa, Catalina Crespíb, Lucio Pallarésc,Antonio Juand, Núria Matamorosa, Maria Rosa Juliàa,∗

a Immunology Service, Son Dureta Hospital, Andrea Doria 55, 07014 Palma de Mallorca, Balearic Islands, Spainb Investigation Unit, Son Dureta Hospital, Palma de Mallorca, Balearic Islands, Spainc Autoimmune Diseases Unit of the Internal Medicine Service, Son Dureta Hospital, Balearic Islands, Spaind Rheumatology Service, Son Llàtzer Hospital, Palma de Mallorca, Balearic Islands, Spain

a r t i c l e i n f o

Article history:Received 17 November 2009Received in revised form 9 February 2010Accepted 9 February 2010Available online 16 February 2010

Keywords:�� T lymphocytesCD8T regulatory cellsIntracellular cytokinesBehcet’s disease

a b s t r a c t

�� T lymphocytes (GD) have been suggested as one of the causes of cytokine dysregulation that results inneutrophils hyperactivation in Behcet’s disease (BD) patients. In addition, GD can provoke cytotoxiclesions in autoimmune diseases by interaction with MICA (MHC class I chain-related A) molecules,through NKG2D receptor on its surface. In contrast, the CD8+ subset of �� T lymphocytes (GDCD8+)has been related to regulatory T activity.

The aim of this study was to determine the phenotype and the intracellular cytokine profile in GD fromperipheral blood, to discern if they were skewed to an effector or regulatory pattern in BD.

We performed phenotype analysis, by three-colour flow cytometry, in 28 BD, 15 healthy controls (HC)and 14 patients with recurrent bucal ulcers (RBU). We studied intracellular cytokine production in 10BD and 14 HC, after polyclonal stimulation. In addition, we analysed serum IL-15 and soluble MICA, byELISA, in 27 BD, 21 HC and 40 rheumatoid arthritis patients.

The hallmark in BD was a specific increase in CD8 expression by GD, and in GDCD8+ absolute num-bers. Most of GDCD8+ presented CD8 �� homodimers and were negative for CD103, Foxp3 and CTLA-4.GDCD8+ and GDCD8− were high IFN�−, but poor IL-2, IL-10, TGF� and IL-4-producing cells, with nodifferences between BD and HC. NKG2D expression on CD8+ T cells, serum IL-15 and soluble MICA were

not significantly increased in BD.

Our results do not suggest a T regulatory profile for GDCD8+ neither in HC, nor in BD. We cannot ruleout other suppression mechanisms or some heterogeneity within this subset that could contribute toregulatory function.

Abbreviations: BD, Behcet’s disease; GD, �� T lymphocytes; GDCD8+, CD8+ �� Tlymphocytes; ABCD8+, CD8+ �� T lymphocytes; Treg, regulatory T lymphocytes; HC,healthy controls; RBU, patients with recurrent bucal ulcerations; IEL, intraepitheliallymphocytes; Io, ionomycin; PBL, peripheral blood lymphocytes; PBMC, peripheralblood mononucleated cells; PMA, phorbol myristate esters; BDA, active Behcet’sdisease patients; BDNA, inactive Behcet’s disease patients; RA, rheumatoid arthritis;mAbs, monoclonal antibodies; ATTC, American Type Culture Collection; MFI, meanfluorescence intensity; MICA, MHC class I chain-related A.

∗ Corresponding author. Tel.: +34 971 17 56 98; fax: +34 971 17 56 98.E-mail addresses: antonio.clemente@ssib.es (A. Clemente),

ana.cambra@ssib.es (A. Cambra), fifanus@hotmail.com (I. Munoz-Saá),catalina.crespib@ssib.es (C. Crespí), lucio.pallares@ssib.es (L. Pallarés), ajuan@hsll.es(A. Juan), nuria.matamoros@ssib.es (N. Matamoros), rosa.julia@ssib.es (M.R. Julià).

0165-2478/$ – see front matter © 2010 Elsevier B.V. All rights reserved.doi:10.1016/j.imlet.2010.02.005

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

Behcet’s disease (BD) is a multisystemic vasculitis charac-terised by relapses of oral and genital ulcers, which can alsoinvolve skin, ocular, neurological, gastrointestinal and articularlesions. Evidence suggests that the basic aetiopathogenic mecha-nism consists in a hyperactivation of neutrophils but the primarytrigger is unknown [1]. Causal bacterial and viral agents havebeen implicated and experimental models, such as infected miceby herpes simplex virus, have been studied [2]. Against theinfectious hypothesis there is the lack of clinical response to

anti-viral and anti-bacterial therapies, whereas immunomodula-tory or immunosuppressive treatments are effective [3]. The mostaccepted hypothesis is that a multifactorial process induces animmune dysregulation that results in exaggerated neutrophils acti-vation [1].

5 ology

acccc

lritVpamM

tilamwcfoc

tis

TM

Ml

8 A. Clemente et al. / Immun

BD is considered a disorder that shares some features with bothutoinflammatory and autoimmune diseases. Its autoinflammatoryharacteristics are the absence of specific transferable autoreactiveells or antibodies, and the over-expression of proinflammatoryytokines [4]. The strong association with HLA-B*51 seems to impli-ate T lymphocytes in the basis of the disease [5].

�� T lymphocytes (GD) have been involved in cytokine dysregu-ation that results in neutrophils hyperactivation in BD. GD are veryapid cytokine-producing cells and they are Th1 and Th17-inducersn experimental models [6–8]. There are two main GD subpopula-ions in humans, depending on their reordered V� region: V�2+ and�1+, accounting for more than 90% of the total GD [9]. V�2+ cellsroliferate in response to bacterial non-peptidic phosphoantigensnd are predominant in peripheral blood. V�1+ lymphocytes areainly located in epithelia and respond to stress signals, such asICA/B antigens through NKG2D activating receptor on its surface.CD8+ �� T lymphocytes (GDCD8+) are responsible for many

ypes of systemic tolerance in oral, and ocular-induced exper-mental models [10–12]. GDCD8+ recirculating intraepithelialymphocytes (IEL) have also been implicated in peripheral toler-nce, in a NOD mouse model of type 1 diabetes [13]. In humans,ost IEL V�1+ bear CD8 antigen and most CD8+ �� IEL are V�1+,hereas in peripheral blood lymphocytes (PBL) about 20% of V�1+

ells present CD8 on their surface [14]. Only a few reports haveocused on human GDCD8+ as regulatory T lymphocytes (Treg); onef them on breast tumours infiltrates [15] and another on IEL from

eliac patients [16].

There are multiple descriptions of GD in BD but they are con-radictory and in some cases incomplete. Some studies found localncreases of V�2+ cells [17,18] or V�1+ [19] and others showed apecific increase of GDCD8+ [20,21]. Most of the works relate GD

able 1ain demographic and clinical characteristics of Behcet’s disease patients.

Patient Age Sex Organ involvement

OU GU SL GI

1 32 M + + +2 44 F + + +3 29 F + + +4 48 F + +5 26 F + + +6 45 F + + + +7 23 F + +8 38 M + +9 38 M + + +

10 46 F + + +11 58 F + +12 40 M +13 44 F + + +14 53 F +15 50 F + + +16 28 F + + +17 42 M +18 57 M + +19 29 M + + +20 14 M + + +21 27 F + +22 48 F + + +23 45 F + +24 38 F + +25 23 M + +26 44 F +27 42 F + + +28 35 F + +29 35 M + + + +30 42 M + + + +31 41 F + +32 61 F + + +33 47 F + + +

: male, F: female, OU: oral ulcers, GU: genital ulcers, SL: skin lesions, GI: gastrointestinaesions, and NA: not available.

Letters 129 (2010) 57–63

shifts to increased cytotoxicity or proliferative response to differentantigens [19,22].

The aim of the present work was to define GD subpopulations inPBL of BD compared with recurrent bucal ulcerations (RBU) patientsand healthy controls (HC), and to characterise the GDCD8+ sub-population. For these purposes we analysed phenotype markersrelated to MICA-induced cytotoxicity, and to regulatory T activity.In addition we determined intracytoplasmic cytokine productionby three-colour flow cytometry to explore the effector or regulatoryprofile of this lymphocyte subset.

2. Materials and methods

2.1. Patients and controls

Thirty-three Caucasian Balearic BD patients, 11 males and22 females (mean age: 39.7 years), were recruited from theAutoimmune Diseases Unit of the Internal Medicine Service atSon Dureta Hospital, and from the Rheumatology Service at SonLlàtzer Hospital. These units cover a population of more than557,000, and Son Dureta Hospital is the reference hospital for theBalearic Islands (total population: 1,006,249). The patients werediagnosed according to the published BD criteria [23] and theirmain demographic and clinical characteristics are described inTable 1.

At the time of sampling, clinical activity was assessed for activ-

ity signs and symptoms, according to the BD Current Activity Form[24]. We considered active (BDA), those patients with active oralulceration plus one more sign of clinical activity: genital ulcers,skin lesions, ocular lesions, joint symptoms, central nervous systeminvolvement, or gastrointestinal inflammation. Nine BD patients

Leukocytes (×103 cells/�l) HLA-B

CNS J OL

+ + 8.7 B35/B44+ 4.5 B35/B44+ 4.3 B13/B65+ 6.5 B08/B51+ 19.6 B35/B51+ 7.3 B50/B51+ + 10.6 B07/B49+ 10.4 B44/B51+ + 12.9 B07/B35

+ + + NA B51/B64+ 8.8 B51/B64+ + 6.5 B27/B62

+ + 6.2 B64/B64+ + NA B07/B64+ 5.3 B13/B44

7.3 B27/B51+ + + 8.3 B35/B49

5.4 B52/B577.2 B51/B648.8 B44/B51

+ + NA B62/B64+ + 3.5 B18/B35+ 12.7 B44/B49

+ NA B18/B51+ + 9.5 B44/B51+ + 8.0 B44/B49+ 11.8 B44/B51+ 5.3 B49/B58+ 15.0 B47/B62+ 9.1 B51/B65+ NA B08/B57

+ 4.0 B07/B187.3 B44/B64

l inflammation, CNS: central nervous system disease, J: joint symptoms, OL: ocular

ology

hi

wcabe

rip5mMRsaC

2

iN

E

2

hpUsm

Fnd

A. Clemente et al. / Immun

ad active disease and the remaining ones were considered as clin-cally inactive (BDNA).

At the time of analysis, 7 patients received no treatment, 11ere using colchicine alone or combined with low doses of corti-

osteroids, 1 corticosteroids alone and 14 an immunosuppressivegent, such as azathioprine, cyclosporine or methotrexate, com-ined or not with corticosteroids. For cytokine measurements, wexcluded patients receiving immunosuppressive drugs.

Fourteen patients with RBU, 8 males and 6 females, and 40heumatoid arthritis (RA) patients, 16 males and 24 females, werencluded as pathology controls in some tests. Mean age for RBUatients was 28.8 years (range: 8–50) and for RA patients was5.2 years (range: 29–81). Twenty-nine Caucasoid age- and sex-atched volunteer donors were recruited as healthy controls (HC).ean age for HC was 31.0 years (range: 16–45). Patients with

BU and HC were free of evidence of autoimmune disease. Bloodamples were collected after informed consent. This study waspproved (reference number: IB731/06PI) by the Balearic Ethicalommittee (CEIC).

.2. Measurement of serum IL-15 and soluble MICA

We measured IL-15 serum levels by ELISA, as describedn the manufacturer’s instructions (R&D Systems, Wiesbaden-ordenstadt, Germany).

We quantified MICA-soluble in serum by a MICA-sandwichLISA protocol (Immatics, Tübingen, Germany).

.3. �ı T lymphocytes phenotyping

We isolated peripheral blood mononucleated cells (PBMC) from

eparinised blood by density gradient centrifugation. We resus-ended PBMC in phosphate buffered saline (Oxoid, Hampshire,nited Kingdom) at 107 cells/ml. 100 �l of the cell suspension were

tained with appropriate amounts of fluorochrome-conjugatedonoclonal antibodies (mAbs).

ig. 1. Percentage of total TCR ��+ cells in peripheral blood lymphocytes (PBL) (A), peumbers of �� CD8+ T lymphocytes in peripheral blood (C). HC: healthy controls (n = 15isease patients (n = 19), and BDA: active Behcet’s disease patients (n = 9). *p < 0.05, **p < 0

Letters 129 (2010) 57–63 59

To evaluate surface antigens expression by GD, we gatedanti-TCR ��-FITC or anti-TCR ��-PE positive lymphoid cells (Bec-ton Dickinson PharMingen, San Diego, USA). We used CD8-PCy5mAb for GDCD8+ gating (IQtest Beckman Coulter, Fullerton,USA). Other mAbs used were anti-TCR V�2-FITC (Beckman Coul-ter Immunotech, Marseille, France), anti-TCR V�1-FITC (Endogen,Rockford. USA), anti-CD103-PE (Becton Dickinson), anti-NKG2Aand anti-NKG2D (PE-conjugated, R&D).

We measured heterodimer CD8 �� expression on GDCD8+ sep-arately using an antibody specific to �-chain (clone 2ST8.5H7PE-conjugated, Beckman Coulter Immunotech). We calculated theCD8 ��+ homodimers contribution by CD8 �+ percentage subtrac-tion from total CD8 �-positive (CD8-PCy5+) GD.

2.4. Foxp3 and CTLA-4 intracellular expression

PBMC intracellular staining to detect Foxp3 and CTLA-4 expres-sion was performed following the fixation/permeabilization setmanufacturer’s instructions (eBioscience, San Diego, USA) but using1 �g/ml of purified human IgG solution (Grifols, Barcelona, Spain),instead of mouse serum, for blocking. We used PE-conjugatedmAbs to Foxp3 (eBioscience) and to CTLA-4 (Beckman CoulterImmunotech). Unspecific intracellular staining was tested with aPE-labelled mouse IgG1 mAb (Becton Dickinson) as isotype con-trol. We chose anti-TCR ��-FITC and CD8-PCy5 as surface markersand proceeded as indicated above for surface phenotyping. Ina parallel assay we incubated cells with anti-CD4-PCy5 (IQtestBeckman Coulter) and anti-CD25-FITC (Becton Dickinson) to testCD4+CD25high cells as positive control.

2.5. �ı T lymphocytes cytokine profile

2.5.1. Culture and stimulationWe obtained PBMC from 7 BDNA, and 3 BDA, each of them tested

with a healthy control. We resuspended cells at a concentration of2 × 106 cells/ml in RPMI 1640 medium (Cansera, Ontario, Canada)with 1% penicillin–streptomycin solution (Gibco Laboratories,

rcentage of CD8+ (B), V�1+ (D) and V�2+ (E) cells in �� T lymphocytes. Absolute), RBU: patients with recurrent bucal ulcerations (n = 14), BDNA: inactive Behcet’s.01, and ***p < 0.001.

ology Letters 129 (2010) 57–63

Fig. 2. NKG2D means of fluorescence intensity (MFI) in �� CD8+, �� CD8− and�� CD8+ T lymphocytes (A). Percentage of CD4+NKG2D+ T lymphocytes in PBL(B). HC: healthy controls (n = 15), RBU: patients with recurrent bucal ulcerations(n = 14), BDNA: inactive Behcet’s disease patients (n = 19), BDA: active Behcet’s dis-

60 A. Clemente et al. / Immun

Paisley, Scotland) and supplemented with 10% heat-inactivatedfoetal calf serum (Gibco). Phorbol myristate esters (PMA) at20 ng/ml plus ionomycin (Io) at 1 �g/ml provided for polyclonalstimuli (both from Sigma). Cytokine secretion was inhibited by theaddition of 10 �g/ml of Brefeldin A (Sigma). Cells not stimulatedbut exposed to Brefeldin A served as negative controls. We keptcultures for 6 h, at 37 ◦C, in a 5% CO2 atmosphere.

2.5.2. Intracellular cytokines measurementWe permeabilized, stained and analysed PBMC as aforemen-

tioned for intracellular Foxp3 and CTLA-4 detection. After surfacestaining and fixation/permeabilization process, cells were stainedwith the appropriate amounts of PE-conjugated anti-IL-2, IL-4, IL-10, IFN� (all from Becton Dickinson PharMingen) or anti-TGF�1monoclonal antibody (IQ products, Groningen, the Netherlands).We used GDCD8+, GDCD8− and total GD gates (minimum 1000events) to analyse cytokine expression.

2.6. Cell lines

We used LS411N line, which constitutively secretes transform-ing growth factor beta 1 (TGF�1), purchased from the AmericanType Culture Collection (ATTC, Massanas, USA), as positive controlin TGF�1 staining. We chose propagation, medium renewal andsubculturing conditions for supporting cells on the basis of the rec-ommendations of the ATTC. We detached cells and cultured themin standard conditions, for 15–18 h to allow receptors regenerationafter trypsin (Sigma, St. Louis, USA) treatment.

2.7. Cytometry measurements

We acquired cells on a Coulter Epics XL MCL flow cytometer andanalysed them with Expo32 software (both from Beckman Coul-ter). Alternatively, we performed three-colour staining and analysisusing the FACScalibur system (Becton Dickinson).

2.8. Statistical analysis

We performed statistical analysis with Prism version 4.00,Graph-Pad software (San Diego, CA, USA). We used non-parametrictwo-tailed test for means comparison and Fisher’s exact test forcontingency tables analysis. When we compared means frommore than two groups we performed Kruskal–Wallis statistic test.Data are presented as mean ± standard deviation if not otherwisestated.

3. Results

3.1. �ı+, CD8+ �ı+, Vı1+ and Vı2+ T cells in peripheral bloodfrom patients and healthy controls

We detected higher percentages of total GD in PBL from BDA:8.4 ± 6.6 and RBU: 5.5 ± 5.7 compared to HC: 3.4 ± 2.4 and BDNA:3.7 ± 3.2 (Fig. 1A). Significant differences were only found inthe percentage of CD8+ cells in GD from BDA: 31.9 ± 18.9 andBDNA: 30.7 ± 13.8 versus HC: 12.7 ± 8.4 (p < 0.05 and p < 0.001,respectively) (Fig. 1B), in the absolute numbers of GDCD8+ inperipheral blood from BDA: 94.4 ± 195.0 versus HC: 6.6 ± 6.4

+

(p < 0.01) (Fig. 1C), and in the percentage of V�2 cells in GD fromRBU: 85.3 ± 16.7 compared to BDNA: 59.0 ± 30.9 (p < 0.01) (Fig. 1E).

There was a tendency to a higher percentage of V�1+ cells in GDfrom BD patients versus HC and RBU but it did not reach statisticalsignificance (Fig. 1D).

ease patients (n = 8). Soluble MICA concentration in serum (C). HC: healthy controls(n = 21), BD: Behcet’s disease patients (n = 27), and AR: rheumatoid arthritis patients(n = 40). *p < 0.05, **p < 0.01, and ***p < 0.001.

3.2. NKG2D expression by T lymphocytes and serum levels ofIL-15 and MICA

The activating receptor NKG2D is constitutively expressed byGD and a subset of CD8+ �� T lymphocytes (ABCD8+), whereas itsexpression is minimal on CD4+ �� T lymphocytes [25]. To exploreNKG2D expression we determined the mean of fluorescence inten-sity (MFI) in GDCD8+, GDCD8− and ABCD8+ and the percentage ofNKG2D-positive cells in CD4+ �� T lymphocytes.

We found NKG2D MFI values significantly higher in GDCD8+versus ABCD8+, from HC, RBU, BDNA (p < 0.001 in the three groups)and BDA (p < 0.01), and in GDCD8+ versus GDCD8−, from HC andRBU (p < 0.05 in both groups). We found no significant differences

comparing patients and controls (Fig. 2A).

Most of the healthy controls presented less than 2% NKG2D-positive CD4+ lymphocytes, in contrast means in RBU and BDAwere over 5%. Differences were significant comparing HC with RBU

A. Clemente et al. / Immunology Letters 129 (2010) 57–63 61

(top) a

(a

udep

Bn

t((vwa

3

ro(dhr

G

TC

B

Fig. 3. Intracellular Foxp3 and CTLA-4 expression by CD4+CD25high

p < 0.01), BDNA (p < 0.05) and BDA (p < 0.001) and between BDAnd BDNA (p < 0.01) (Fig. 2B).

We measured serum IL-15 that has been related to NKG2Dpregulation in T cells from patients with different autoimmuneisorders [26] and soluble MICA, which is systemic and locally overxpressed in celiac [27], diabetic [28], and rheumatoid arthritisatients [29].

IL-15 values were below the reference limit (<3.9 pg/ml) in 23D and very close to the cut-off in the remaining 4. Differences wereot significant between BD and HC (data not shown).

We studied soluble MICA in 27 BD and 40 rheumatoid arthri-is (RA) sera. According to previous reports [29] 16 of 40 AR40%) patients presented values higher than the reference value30 pg/ml). Among 27 BD patients, only 4 (14.8%) had increasedalues, one of whom was in the active phase of disease. Differencesere significant between RA and both BD (p < 0.05) and HC (p < 0.05)

nd not significant between BD and HC (Fig. 2C).

.3. CD8 �ı T lymphocytes phenotype

In GDCD8+, we studied some intracellular antigens that areelated to Treg function, as CTLA-4 and Foxp3 [30], and surfacenes, as NKG2A, linked to cytotoxicity inhibition [16] and CD103�E�7 integrin), a Treg and IEL marker [31,32]. Furthermore, we

etermined the contribution of CD8 �� homodimers and CD8 ��eterodimers, that had previously been tested with discrepantesults [33,34].

Foxp3 and CTLA-4 intracellular expression were negligible inDCD8+ from BD patients and HC (Fig. 3).

able 2D8 isoforms contribution in GDCD8+ and NKG2A expression on GD and GDCD8+.

%CD8 �� in GDCD8+ %CD8 �� in GDCD8+ %NK

HC (n = 29) 82.0 ± 17.6 18.0 ± 17.6 31.8BD (n = 16) 80.0 ± 16.8 20.0 ± 16.8 13.8

D: Behcet’s disease patients, HC: healthy controls, GD: �� T lymphocytes, and GDCD8+:* p < 0.05.

nd �� CD8+ T lymphocytes (bottom). A representative experiment.

We found no surface CD103 expression, neither in total GDnor in GDCD8+ from any group. Most of GDCD8+ expressed CD8�� homodimers, in patients and controls (Table 2), without dif-ferences, but we detected a minority that presented CD8 ��heterodimers in BD and HC.

The percentage of total GD NKG2A+ was significantly lowerin BD versus HC (p < 0.05), but it was not confirmed in GDCD8+.We found no differences between absolute numbers of GDCD8+NKG2A+ comparing patients and controls (Table 2).

3.4. Cytokine production by CD8-positive and negative �ı Tlymphocytes

We found frequencies of GD cytokine-producer cells inthe order: IL-4+ < IL-10+ < TGF�+ < IL-2+ < IFN�+. In most casesGDCD8− were better cytokine-producers than GDCD8+ (Fig. 4). Wedetected significant differences between GDCD8− and GDCD8+ inthe intracellular production of IL-4, IL-10 and TGF� (p < 0.05 in thethree cases) only in HC, and of IL-2 both in HC (p < 0.01) and BD(p < 0.05).

Means of GD IFN�-positive were higher than 35%. In most indi-viduals, they were slightly lower in GDCD8+ than in GDCD8−(Fig. 4), but differences were not significant. Comparing meansof IFN� fluorescence intensity (MFI) between both subsets, in HC

(GDCD8+: 1392 ± 1241, GDCD8−: 2922 ± 3091) and BD (GDCD8+:833 ± 1086, GDCD8−: 1759 ± 1774), the difference reached statis-tical significance (p < 0.05), only in BD (data not shown).

Furthermore, GD were better IFN�-producers than PBL (p < 0.01in HC and BD) and PBL better IL-2-producers than GD (p < 0.01

G2A in GD %NKG2A in GDCD8+ GDCD8+ NKG2A+ (cells/mm3)

± 19.0 * 26.2 ± 19.5 13.0 ± 12.6± 10.6 * 11.7 ± 10.3 12.0 ± 10.1

CD8+ �� T lymphocytes.

62 A. Clemente et al. / Immunology Letters 129 (2010) 57–63

F �: total n = 10

ic(p

cn

4

cuawptsssTtsiblnnespu

ef

ig. 4. Intracellular cytokines production in PBL: peripheral blood lymphocytes, �ymphocytes from HC: healthy controls (n = 14), and BD: Behcet’s disease patients (

n HC), in agreement with previous reports [8]. Percentages ofytokine-positive cells were higher in GD than in PBL for IL-4p < 0.05 in BD), TGF� (p < 0.01 in HC) and IL-10 (p < 0.05 in BD, and< 0.01 in HC) (Fig. 4).

We found no differences between mean percentages ofytokine-positive cells neither by comparing BD patients with HCor between BDA and BDNA.

. Discussion

In the present study we confirm that CD8-positive �� T lympho-ytes, from peripheral blood, constitute a cell subset specificallypregulated in Behcet’s disease patients. In most cases these cellslso express V�1 chains. The other main GD subpopulation (V�2+)as increased in individuals with recurrent bucal ulcers and, asreviously reported [22], it could be related to bacterial infec-ions. V�1+ T cells act as cytotoxic effectors in response to stresstimuli, such as MICA antigens, through NKG2D receptors on theirurface. NKG2D is upregulated by IL-15 in autoimmune diseases,uch as celiac disease, diabetes, and rheumatoid arthritis [26–28].he IL15–NKG2D–MICA axis has been argued to be crucial inhe aetiology of celiac disease [35]. Celiac and BD patients shareome features as the association with some MICA alleles and thencreased numbers of �� T lymphocytes in tissues and peripherallood [36]. For this reason we measured serum IL-15 and MICA

evels and determined NKG2D expression on GD, but found no sig-ificant differences between BD and controls. The relatively smallumber of BD patients used in the IL-15 assay could explain differ-nces with previous works [37], although our results coincide withome others [38]. Only the expression of NKG2D in CD4+ T lym-

hocytes was higher in BD but it was shared with recurrent bucallcers patients.

These results corroborate our previous findings about skewedxtramembrane MICA polymorphisms in BD towards a low affinityor NKG2D [39] and led us to explore markers of other functions

l �� T lymphocytes, �� CD8+: CD8+ �� T lymphocytes, and �� CD8−: CD8− �� T). Data are expressed as mean ± standard deviation.

attributed to GDCD8+. Experimental models have demonstratedthat these cells are crucial to tolerance induction [40]. Veryrecent reports have described a regulatory function in GDCD8+from breast tumour infiltrates [15] and from intraepithelial lym-phocytes in celiac patients [16]. In the former case, GDCD8+inhibited proliferative response of CD4+ T cells. In the last one, theysuppressed cytotoxicity against HLA-E-positive epithelial cells,through inhibitory receptor NKG2A, on its own surface. Indirectevidence of a regulatory role by GDCD8+ has been shown in chronicpulmonary disease patients [41].

In this work, GDCD8+ cells did not show a characteristic phe-notype of Treg: they were Foxp3, CTLA-4 and CD103-negative, andNKG2A was not upregulated in BD versus HC. The absence of CD103does not support the fact that they were recirculating IEL. Contribu-tion of CD8 isoforms to GDCD8+ was no different in Behcet’s diseasepatients than in healthy controls.

Furthermore we have focused on intracellular cytokine produc-tion by GDCD8+ in response to polyclonal stimulation. The naturallyoccurring CD4+CD25high Treg cells have been characterised by theproduction of TGF� and IL-10 and the absence of IL-2, IL-4 and IFN�secretion [42], after activation by PMA/Io. In these cells, blockage ofTGF� resulted in the inhibition of suppressive activity [43,44]. Wefound that GDCD8+, in healthy and BD individuals, present a similarcytokine profile than GDCD8−, which was distinct to the describedfor CD4+CD25high Treg. Both GD subsets were high IFN-� and poorIL-2, IL-4, IL-10, and TGF�-producing cells, being GDCD8− bettercytokine-producers.

Our results do not rule out the possibility of a suppressor activ-ity by GDCD8+. In fact, analysis of inhibitory human GDCD8+ cells,found in breast tumour infiltrates, showed a lack of IL-10 and TGF�

secretion [15]. In contrast �� CD8+ IEL in celiac patients, capable ofsuppressing cytotoxicity, were high TGF�-producers [16]. Furtherfunctional suppressive tests in vitro and cytokine analysis in cocul-ture supernatants are required to elucidate the nature of GDCD8+in BD.

ology

C

A

tf

tC(d

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

A. Clemente et al. / Immun

onflict of interest

None.

cknowledgements

We wish to thank E. Esteban, J. Rascón, A. Artigues and E. Mar-orell for their help in collecting blood samples and informationrom the patients.

Funding: This work was supported by a grant (FIS 07/0604), fromhe “Fondo de Investigaciones Sanitarias, Ministerio de Sanidad yonsumo”, Spain. The work of A. Clemente was supported by a grantCEB08/001) from the “Conselleria d’Economia i Innovació, Governe les Illes Balears”, Spain and the European Social Fund (ESF).

eferences

[1] Mendes D, Correia M, Barbedo M, Vaio T, Mota M, Goncalves O, et al. Behcet’sdisease—a contemporary review. J Autoimmun 2009;32:178–88.

[2] Seonghyang S. Etiopathology of Behcet’s disease: Herpes Simplex Virus infec-tion and animal model. Yonsei Med J 1997;38:359–64.

[3] Verity DH, Wallace GR, Vaughan RW, Stanford MR. Behcet’s disease: from hip-pocrates to the third millennium. Br J Ophthalmol 2003;87:1175–83.

[4] Gül A. Behcet’s disease as an autoinflammatory disorder. Curr Drug TargetsInflamm Allergy 2005;4:81–3.

[5] Verity DH, Marr JE, Ohno S, Wallace GR, Stanford MR. Behcet’s disease, the silkroad and HLA-B51: historical and geographical perspectives. Tissue Antigens1999;54:213–20.

[6] Ferrick DA, Schrenzel MD, Mulvania T, Hsieh B, Ferlin WG, Lepper H. Differ-ential production of interferon-� and interleukin-4 in response to Th1- andTh2-stimulating pathogens by �� T cells in vivo. Nature 1995;373:255–7.

[7] Roark CL, Simonian PL, Fontenot AP, Born WK, O’Brien RB. �� T cells: an impor-tant source of IL-17. Curr Opin Immunol 2008;20:353–7.

[8] Raga S, Julià MR, Crespí C, Figuerola J, Martínez N, Milà J, et al. GammadeltaT lymphocytes from cystic fibrosis patients and healthy donors are highTNF-alpha and IFN-gamma-producers in response to Pseudomonas aeruginosa.Respir Res 2003;4:9 [Epub 2003 September 8].

[9] Carding SR, Egan PL. �� T cells: functional plasticity and heterogeneity. Nat RevImmunol 2002;2:336–45.

10] Harrinson LC, Dempsey-Collier M, Kramer DR, Takahashi K. Aerosol insulininduces regulatory CD8 �� T cells that prevent murine insulin-dependent dia-betes. J Exp Med 1996;184:2167–74.

11] Zhou J, Appleton SE, Stadnyk A, Lee TD, Nashan BA. CD8+ �� T regulatory cellsmediate kidney allograft prolongation after oral exposure to alloantigen. Trans-plant Int 2008;21:679–87.

12] Ashour HM, Niederkorn JY. �� T cells promote anterior chamber-associatedimmune deviation and immune privilege through their production of IL-10. JImmunol 2006;177:8331–7.

13] Locke NR, Stankovic S, Funda DP, Harrison LC. TCR�� intraepithelial lympho-cytes are required for self-tolerance. J Immunol 2006;176:6553–9.

14] Deusch K, Lüling F, Reich K, Classen M, Wagner H, Pfeffer K. A major fraction ofhuman intraepithelial lymphocytes simultaneously expresses the �/� T cellreceptor, the CD8 accessory molecule and preferentially uses the V�1 genesegment. Eur J Immunol 1991;21:1053–9.

15] Peng G, Wang HY, Peng W, Kiniwa Y, Heon Seo K, Wang R. Tumor-infiltrating�� T cells suppress T and dendritic cell function via mechanisms controlled bya unique Toll-like receptor signaling pathway. Immunity 2007;27:334–48.

16] Bhagat G, Naiyer AJ, Shah JG, Harper J, Jabri B, Wang TC, et al. Small intestinalCD8+TCR��+NKG2A+ intraepithelial lymphocytes have attributes of regulatorycells in patients with celiac disease. J Clin Invest 2008;118:281–93.

17] Yamashita N, Kaneoka H, Kaneko S, Takeno M, Oneda K, Koizumi H, et al. Role

of �� T lymphocytes in the development of Behcet’s disease. Clin Exp Immunol1997;107:241–7.

18] Verjans GM, van Hagen PM, van der Kooi A, Osterhaus AD, Baarsma GS.V�9V�2 T cells recovered from eyes of patients with Behcet’s disease recog-nize non-peptide prenyl pyrophosphate antigens. J Neuroimmunol 2002;130:46–54.

[

Letters 129 (2010) 57–63 63

19] Hamzaoui K, Hamzaoui A, Hentati F, Kahan A, Ayed K, Chabbou A, et al. Pheno-type and functional profile of T cells expressing �� receptor from patients withactive Behcet’s disease. J Rheumatol 1994;21:2301–6.

20] Freysdottir J, Lau S, Fortune F. �� T cells in Behcet’s disease (BD) and recurrentaphthous stomatitis (RAS). Clin Exp Immunol 1999;118:451–7.

21] Suzuki Y, Hoshi K, Matsuda T, Mizushima Y. Increased peripheral blood �� Tcells and natural killer cells in Behcet’s disease. J Rheumatol 1992;19:588–92.

22] Bank I, Duvdevani M, Livneh A. Expansion of �� T-cells in Behcet’s disease:role of disease activity and microbial flora in oral ulcers. J Lab Clin Med2003;141:33–41.

23] International Study Group for Behcet’s Disease. Criteria for diagnosis of Behcet’sdisease. Lancet 1990;335:1078–80.

24] Bhakta BB, Brennan P, James TE, Chamberlain MA, Noble BA, Silman AJ. Behcet’sdisease: evaluation of a new instrument to measure clinical activity. Rheuma-tology 1999;38:728–33.

25] Bauer S, Groh V, Wu J, Steinle A, Phillips JH, Lanier LL, et al. Activation ofNK cells and T cells by NKG2D, a receptor for stress-inducible MICA. Science1999;285:727–9.

26] Roberts AI, Lee L, Schwarz E, Groh V, Spies T, Ebert EC, et al. Cutting edge:NKG2D receptors induced by IL-15 costimulate CD28-negative effector CTL inthe tissue microenvironment. J Immunol 2001;167:5527–30.

27] Meresse B, Chen Z, Ciszewski C, Tretiakova M, Bhagat G, Krausz TN, et al. Coor-dinated induction by IL15 of a TCR-independent NKG2D signaling pathwayconverts CTL into lymphokine-activated killer cells in celiac disease. Immunity2004;21:357–66.

28] Ogasawara K, Hamerman JA, Ehrlich LR, Bour-Jordan H, Santamaría P, Blue-stone JA, et al. NKG2D blockade prevents autoimmune diabetes in NOD mice.Immunity 2004;20:757–67.

29] Groh V, Brühl A, El-Gabalawy H, Nelson JL, Spies T. Stimulation of T cell autore-activity by anomalous expression of NKG2D and its MIC ligands in rheumatoidarthritis. Proc Natl Acad Sci USA 2003;100:9452–7.

30] O’Garra A, Vieira P. Regulatory T cells and mechanisms of immune systemcontrol. Nat Med 2004;10:801–5.

31] Lehmann J, Huehn J, de la Rosa M, Maszyna F, Kretschmer U, Krenn V, et al.Expression of the integrin �E�7 identifies unique subsets of CD25+ as well asCD25− regulatory T cells. Proc Natl Acad Sci USA 2002;99:13031–6.

32] Kilshaw PJ, Murant SJ. Expression and regulation of �7(�p) integrins onmouse lymphocytes: relevance to the mucosal immune system. Eur J Immunol1991;21:2591–7.

33] Moebius U, Kober G, Griscelli AL, Hercend T, Meuer SC. Expression of differentCD8 isoforms on distinct human lymphocyte subpopulations. Eur J Immunol1991;21:1793–800.

34] Lambert C, Genin C. CD3 bright lymphocyte population reveal �� T cells. Cytom-etry B: Clin Cytom 2004;61:45–53.

35] Hüe S, Mention JJ, Monteiro RC, Zhang S, Cellier C, Schmitz J, et al. A direct rolefor NKG2D/MICA interaction in villous atrophy during celiac disease. Immunity2004;21:367–77.

36] Setty M, Hormaza L, Guandalini S. Celiac disease: risk assessment, diagnosis,and monitoring. Mol Diagn Ther 2008;12:289–98.

37] Curnow SJ, Pryce K, Modi N, Knight B, Graham EM, Stewart JE, et al. Serumcytokine profiles in Behcet’s disease: is there a role for IL-15 in pathogenesis?Immunol Lett 2008;121:7–12.

38] Ahn JK, Yu HG, Chung H, Park YG. Intraocular cytokine environment in activeBehcet’s uveitis. Am J Ophthalmol 2006;142:429–34.

39] Munoz-Saá I, Cambra A, Pallarés L, Espinosa G, Juan A, Pujalte F, et al. Allelicdiversity and affinity variants of MICA are imbalanced in Spanish patients withBehcet’s disease. Scand J Immunol 2006;64:77–82.

40] Niederkorn JY. Emerging concepts in CD8+ T regulatory cells. Curr OpinImmunol 2008;20:327–31.

41] Pons J, Sauleda J, Ferrer JM, Barceló B, Fuster A, Regueiro V, et al. Blunted ��T-lymphocyte response in chronic obstructive pulmonary disease. Eur Respir J2005;25:441–6.

42] Dieckmann D, Plottner H, Berchtold S, Berger T, Schuler G. Ex vivo isolation andcharacterization of CD4+CD25+ T cells with regulatory properties from humanblood. J Exp Med 2001;193:1303–10.

43] Levings MK, Sangregorio R, Sartirana C, Moschin AL, Battaglia M, Orban PC, et

al. Human CD25+CD4+ T suppressor cell clones produce transforming growthfactor �, but not interleukin 10, and are distinct from type 1 T regulatory cells.J Exp Med 2002;196:1335–46.

44] Nakamura K, Kitani A, Fuss I, Pedersen A, Harada N, Nawata H, et al. TGF-�1 plays an important role in the mechanism of CD4+CD25+ regulatory T cellactivity in both humans and mice. J Immunol 2004;172:834–42.