CD4dimCD25bright Treg Cell Frequencies Abovea Standardized Gating Threshold are Similar

in Asthmatics and Controls

Hans J€urgen Hoffmann,1* Tine M. Malling,2 Ayfer Topcu,1 Lars P. Ryder,3 Kaspar R. Nielsen,4

Kim Varming,4 Ronald Dahl,1 Øyvind Omland,2,5 and Torben Sigsgaard5

1Department of Respiratory Medicine, Aarhus University Hospital, Aarhus, Denmark2Department of Occupational Medicine, Aalborg Hospital, Aarhus University Hospital, Aalborg, Denmark

3Department of Clinical Immunology, Tissue Typing Lab, National University Hospital, Copenhagen, Denmark4Department of Clinical Immunology, Aalborg Hospital, Aarhus University Hospital, Aalborg, Denmark

5Department of Occupational and Environmental Medicine, Institute for Public Health, Aarhus University, Aarhus, Denmark

Received 26 September 2006; Revision Received 23 December 2006; Accepted 28 December 2006

Background: Thymus selected CD41CD25bright naturalregulatory Treg cells expressing FOXP3 may contribute tocontrol of immune responses. No unique markers havebeen available to identify and characterize Treg. We pres-ent a gating strategy that allows enumeration of Treg onthe basis of CD4 and CD25 and investigate whether asth-matics have fewer Treg than controls.Methods: Asthmatics and controls were selected fromresponses to a mailed questionnaire. CD25, CD4, HLA DR,and appropriate isotypes were recorded by flow cytometry.Results: The CD4 T cells expressing most CD25 are aseparate population expressing FOXP3 and lower levels ofCD4 and CD127. On a CD4 CD25 dot-plot, the CD4 MFIof Treg for 152 participants was calculated to be 0.83 60.043*MFI of CD25bright T-cells. CD4dimCD25bright T cells

in a rectangular gate with a CD4 MFI � 0.9 (0.83 1[2*0.043])*MFI of CD251 T cells were enumerated andshown to be similar for controls (median 8.34%) and asth-matics (median 10.1%). HLA DR expression on Treg corre-lated with CD25 expression.Conclusions: A standardized two color gating methoddefines Treg. It may be applied in most clinical scenariosand is useful for sorting viable Treg. Asthmatics and con-trols have similar numbers of Treg. q 2007 International

Society for Analytical Cytology

Key terms: regulatory T cells; FOXP3; CD25; CD4;CD127; HLA DR; standardized flow cytometry; allergy;ECRHS

Control of immune responses is vital for immune func-tion. Natural regulatory T cells (Treg) are thought to beone of the subsets of regulatory T cells that limit or con-trol T cell mediated immunity (reviewed in (1)). They arepositively selected in the thymus and not derived fromrecent activation by a limited array of antigens (2,3). Treglimit proliferation and cytokine secretion of other CD4 Tcells by a signal depending on intercellular contact (4,5).In contrast, Tr1 or TH3 T cells can be induced in aCD41CD25dim or CD41CD252 population by Treg, andsecrete IL-10 or TGF-b (6). Treg are characterized byexpression of the forkhead transcription factor FOXP3,CD4, and high levels of CD25 (7), and in man have beenshown to also express GITR, CD152 (CTLA-4), and HLADR (8). The identification of Treg by surface markers hasbeen hindered by the lack of unique markers for this phe-notype (7) as activated T cells share the expression ofCD25 and CD4 with Treg. Recently, low expression of theIl-7 receptor CD127 (9,10) and expression of GPR83 andseveral other candidate markers of Treg have beenreported (11,12). We noted in our own work and in the lit-

erature (7,13–15) that the brightest population ofCD41CD251 T cells had a slightly lower expression ofCD4 than did the bulk of the population, and we havedeveloped a gating algorithm to use this in the identifica-tion of Treg.Flow cytometry is well suited for the study of T cells

(16). Small populations of cells can be identified as it isnow possible to analyze more than 500,000 events onstandard instruments. When attempting to separate popu-lations that are marginally different, the precise locationof regions for gating is important. To identify the popula-

*Correspondence to: Hans J€urgen Hoffmann, PhD, Department of Re-

spiratory Diseases, Aarhus University Hospital, Building 2b, Nørrebrogade

44, DK 8000 Aarhus C, Denmark.

E-mail: hjhof@as.aaa.dk

Grant sponsors: Velux and Aarhus University Foundations; Danish Min-

istry of the Interior and Health; Research Centre for Environmental

Health’s Fund.

Published online 27 April 2007 in Wiley InterScience (www.interscience.

wiley.com).

DOI: 10.1002/cyto.a.20389

q 2007 International Society for Analytical Cytology Cytometry Part A 71A:371–378 (2007)

tion containing Treg cells by two color flow cytometry,we defined a ratio of CD4 expression of Treg andCD41CD251 T cells, respectively, to set the thresholdbetween Treg and activated CD4 T cells.

We applied two color flow cytometry in a population-based study on risk factors for adult asthma. Asthma is aninflammatory disease of the airways characterized by epi-sodes of broncho-constriction that can be life threatening(17), and that often has an immunologic component,atopy, involving TH2 T cells (18). Similar to other inflam-matory diseases like the TH1-mediated rheumatoid arthri-tis (19) or Crohn’s disease (20), it is thus a disease whichmight be limited by regulatory T cells. The purpose of thestudy was to identify in a reproducible way the populationof cells containing Treg and measure the number of thesecells in asthmatics and controls. We defined a ratio of CD4expression to set the threshold between Treg and acti-vated CD4 T cells.

METHODSIdentification of FOX P31 and

CD41CD25brightCD127dim T cells as CD4dimCD25bright

For method development, six persons were recruited;three patients with known, but currently untreated aller-gic asthma (one with GINA Stage 1 and two with GINAStage 2 severity of asthma) and three nonatopic controlswithout respiratory or other disease. Staining for FOXP3with the antibody PCH101 (eBioscience, San Diego, CA)was done according to instructions from the supplierusing PCH101 PE, CD4 FITC, and CD25 APC and wascompared with staining with CD25 PE and CD4 APC(Serotec, Kidlington, UK). Eight normal hospital workersdonated blood to analyze the expression of CD127 Alexa647 (BD Biosciences, San Jose, CA) on hemolyzed bloodsamples also stained with CD4 FITC and CD25 PE.Appropriate compensation controls for each experimentwere generated with BD Comp beads (BD, San Jose, CA).Data (�250,000 events) was acquired uncompensatedwith a four color FACS Calibur flow cytometer (BD Bios-ciences, San Jose, CA) and was compensated off-line andanalyzed with FlowJo v 5.4.3 and v 6.4.1 (Treestar, Ash-land, OR).

Patient Selection and Clinical Evaluation

Participants were recruited from a cross-sectional studyof adult asthma in a Danish county, designed in accord-ance with the protocol of the European Community Respi-ratory Health survey (ECRHS) (21). A short questionnaireregarding current asthma, asthma medication, and age forasthma debut were sent out to 2,000 subjects aged 20–44years. All persons with current asthma and adult onset ofasthma, and a random sample of nonasthmatic subjectswere invited for further examinations. Subjects were clas-sified as having current asthma (n 5 43) if they answered‘‘yes’’ to at least one of the following questions: ‘‘Did youwake up at night with dyspnea within the last 12months?’’, ‘‘Have you had an attack of asthma during thelast 12 months?’’, ‘‘Do you currently use asthma medica-tion, inclusive inhalators, aerosols, or tablets?’’, ‘‘Have youused an inhaler within the last 24 hours?’’ and ‘‘Have youused any kind of inhaled medicine to help your breath atany time within the last 12 months?’’ Controls (n 5 109)were selected among those that answered no to all ques-tions on respiratory symptoms. Asthmatics and controlsdid not differ with regard to age, sex, and smoking (Table1). When the blood sample was taken approximately oneyear after the survey, the classification was re-evaluated bythe attending physician (Table 1).

Phenotyping of Peripheral Blood Leukocytes

Aliquots of heparinized blood (100 ll) were labeledwith conjugated antibodies to CD4 PE Cy5, CD25 PE or anisotype, HLA DR FITC or an isotype for characterization ofthe regulatory T cells. Antibodies were from Serotec(Kidlington, UK). Data (100,000 events for the first 10 par-ticipants, and subsequently 30,000 events) was acquiredthe next day on a FACS Scan (BD, San Jose, CA) withouthardware compensation. Compensation controls weregenerated on each day of analysis, with Comp beads (BD,San Jose, CA) labeled with the relevant antibodies (CD4,CD25, and HLA DR).

Data Analysis

Data files were compensated in batch with FlowJo(Treestar, OR) and regions were placed to identify firstlymphocytes and then CD41 T cells. A region was set to

Table 1Demographic Parameters

Controls

Questionaire identified asthma

All Untreated b2 agonist ICS

Number 109 43 20 15 8Current asthma 0:1 98:10 [1] 14:28 [1]a 8:12 5:10 1:6 [1]Atopy 0:1 62:43 [4] 14:27 [2]a 12:8 2:13 0:6 [2]Smoking ever 0:1 54:54 [1] 21:21 [1]b 7:13 9:6 5:2[1]Smoking current 0:1 72:36 [1] 30:12 [1]b 11:9 14:1 5:2 [1]Age (mean 6 SD) 34 (76 6) 32 (9 6 7)b 33 (66 7) 32 (7 6 7) 31 (3 6 10)Sex F/M 59/49 [1] 22/20 [1]b 10/10 8/7 4/3 [1]

Numbers in square brackets indicate missing data.aP < 0.01.bNot significantly different from controls.

372 HOFFMANN ET AL.

Cytometry Part A DOI 10.1002/cyto.a

include 1% of isotype for CD25 PE, and was copied on thefile labeled with CD25 to define the CD41CD251 popula-tion. A region was set to include the top 2% of the CD251

population to define the CD41CD252% population. Athird region was set such that the CD4 MFI of this regionwas 0.90*CD4 MFI of the CD251 population to define theCD4dimCD25bright population. Cells in the CD4 andCD4dimCD25bright regions were counted. The CD41

CD252% region, the CD4dimCD25bright region, and theCD41CD251 region set above were copied onto data filesacquired of cells labeled with CD4 Pe Cy5, CD25 PE, andHLA DR FITC, or an isotype control. Cells in the threeregions that were positive for HLA DR were counted, andthe fraction of cells positive for the isotype (~2%) was sub-tracted. The fraction positive cells was plotted againstCD25 or compared for the different regions defined onthe CD25:CD4 plot (Fig. 2) as had been done previously(8,22).

RESULTSFOXP31 T Cells are CD4dimCD25bright

We confirmed (7) that human CD4 T cells that expresshigh levels of CD25 express FOX P3 (Fig. 1d). Only CD4 Tcells expressed FOX P3 as a coherent population (notshown). To demonstrate that FOX P3 is expressed only inCD25 bright cells, cells were stained with CD4 FITC andCD25 APC, and FOX P3 PE or an isotype. FOXP31 T cellscomprised 6% 6 1% of the CD41 T cells (n 5 6). Therewas no difference between atopic patients and controls.Two regions of CD41 T cells were defined on the basis ofexpression of CD25; the top 2% expressing CD25, as wellas a CD252 population. Staining with FOXP3 or Isotypewas evaluated on these populations. FOXP3 binding wasnot detected on CD252 cells (Fig. 1e), but was detected inCD25 bright cells. The top 2% of CD25-positive CD4 Tcells contained only 33% 6 4.9% of FOXP31 T cells (n 56). FOXP31 T cells expressed less CD4 than FOXP32 Tcells (Fig. 1f). CD25bright cells did not bind isotype forFOXP3, demonstrating that FOXP3 binding was specific(Fig. 1d).

An Objective Gating Strategy to IdentifyCD4dimCD25bright T Cells

To identify FOX P31 T cells only on the basis of surfacemarkers CD4 and CD25, we defined an objective andthereby reproducible gating strategy to determine the fre-quency of CD4dimCD25bright T cells (Fig. 2). Instead of fo-cusing exclusively on the expression of CD25, we definedthese cells as CD4dimCD25bright. We proposed that a gateenriched for FOX P31 T cells can be defined asCD4dimCD25bright, a region that contains predominantlyFOXP31 T cells (Figs. 1 and 2). The top 2% of theCD41CD251 population is FOXP31 (Fig. 1, (7)) and hadregulatory properties (22). As we had collected data filesof 30,000 events with CD4 and CD25, which containedapproximately 4,000 CD4 T cells in each, we testedwhether a meaningful measure of fluorescence intensitycan be determined from the approximately 80 cells in the

CD41CD252% region by comparing spread of mean andmedian of consecutive time slices of one large data file toresample the same dataset repeatedly (Fig. 3a). The stand-ard deviation for CD4 MFI was similar over the range ofcell numbers investigated (10–40,000), and the SD of 16individual measurements was 3.1% and 2.7% of the MFIdetermined of 58 cells and 118 cells, respectively, suggest-ing that it was possible to obtain a precise measurementof the fluorescence of CD4 the CD41CD252% populationfrom 80 cells.We determined the MFI for CD4 of the CD252% T cells

as a fraction of the MFI of CD251 T cells and calculatedthe mean and standard deviation of this ratio. In thecohort of 152 participants of this study, the CD4 MFI ofthe CD252% T cells was 0.83 6 0.046 (mean 6 SD) of theCD4 MFI for all CD41CD251 cells. The CD4 MFI of theCD252% T cells of controls (n 5 109), asthmatics not trea-ted with ICS (n 5 35) and asthmatics treated with ICS(n 5 8) were 0.83 6 0.045, 0.83 6 0.051, and 0.83 60.057, respectively. With this information, we constructedan algorithm to obtain a metric proportional to the num-ber of FOX P3 expressing T cells on the basis of CD4 andCD25 expression; it is the number of CD4dimCD25bright Tcells in the region that has a CD4 MFI of � 0.9 (@0.83 1(2*0.046)) the CD4 MFI of CD251CD4 T cells.To place the region correctly, one starts with two rectan-

gular regions on a CD25 CD4 dot-plot of CD41 lymphocyteswide enough to cover the range of CD4 expression; oneregion (CD41CD251) contains all CD251 cells above iso-type, and the other region (CD4dimCD252%) contains the 2%cells that express most CD25 (Fig. 2). If the MFI for CD4 ofCD4dimCD252% � CD41CD251, the sample may recentlyhave been activated, and cannot be evaluated. If the MFI forCD4 of CD4dimCD252% is less than the CD4 MFI of theCD41CD251 region, move the CD4dimCD252% towardthe CD4 axis until the MFI for CD4 of this region is 0.9 3the CD4 MFI of the CD41CD251 region. The region set bythis procedure is termed CD4dimCD25bright. We investigatedthe effect of CD4 T cell number on the precision of the mea-surement of CD4dimCD25bright on the same data file as usedto determine the effect of cell number on CD4 MFI measure-ment above. The number of CD4 T cells evaluated here(4,000 CD4 T cells) was on the threshold between samplessize with high precision (>6,000 CD4 T cells) and largespread (<3,000 CD4 T cells, Fig. 3b) and may close to thelower limit of the cell number required for this analysis.This gating strategy, defining CD4dimCD25bright T cells,

identified 98% 6 19% of FOXP31 CD41 T cells (n 5 6).When comparing FOXP3 expression of the CD4dim

CD25bright and the CD41CD252%, the CD4dimCD25bright

population contained markedly more FOXP31 cells thanthe CD252% population, but contained also a proportionof FOXP32 T cells (Fig. 1e). The CD4dimCD25bright T cellsand the CD252% T cells could be further characterized bya reduced expression of CD4, as well as low forward andlow side scatter (Fig. 1j and 1k). The CD4dimCD25bright his-tograms have a shape similar to the histograms for theCD252% population, but include markedly more FOXP31

T cells.

373TREG IDENTIFIED AS CD4dimCD25bright T CELLS

Cytometry Part A DOI 10.1002/cyto.a

FIG. 1. Characterization of FOX P3 expression. A CD4 T cell preparation purified by negative selection with Miltenyi reagents (anti-CD8, CD14, CD16,CD19, CD36, CD56, CD123, TCR g/d, and CD235a, Miltenyi Biotech, Bergisch-Gladbach, DE) was permeabilized and stained with anti-CD4 FITC, anti-CD25APC, and PCH101 PE (panels a–f), or stained with CD4 APC and CD25 FICT (panels g–l). For clarity, 500,000 events were collected to display clearly sepa-rable subpopulations. (a–c) The gating strategy for identifying CD25bright CD4 T cells is shown. In Panel c, gates for the top 2% of CD251 cells are in dottedred, the CD4dimCD25bright gate is in solid blue and the CD251 gate is in stippled orange. A CD252 gate is in (dash-dot) green. In Panel d, coexpression ofCD25 and FOXP3 is shown in contrast to an isotype. In Panel e, distribution of FOXP31 to FORP32 cells is shown for the CD252% (dotted red),CD4dimCD25bright (solid blue), CD251 (stippled orange), and CD252 (green; dash dot) gates. Panel f illustrates that the FOXP31 population expresses lessCD4 than the FOXP32 population. Panels g–i illustrate the gating strategy on the same buffy coat, without permeabilization. Panel j illustrates that blasts(maroon dot dot dash region in Panel g) have elevated CD25 as well as CD4 expression. In Panels k and l, forward and side scatter of cells in the gatesCD252%, CD4dimCD25bright and CD251 as well as in the blast gate are compared. [Color figure can be viewed in the online issue, which is available atwww.interscience.wiley.com.]

374 HOFFMANN ET AL.

Cytometry Part A DOI 10.1002/cyto.a

CD41CD25brightCD127dim and CD4dimCD25bright

Populations have Similar Frequencies

Expression of the IL-7 receptor CD127 with CD25 wascompared with the present algorithm (Fig. 4). Using thethree markers CD4, CD25, and CD127, it is possible toidentify a population (9.1% 6 1.6%, n 5 8) of T cells thatare CD4dimCD25brightCD127dim. By comparison, when thepresent algorithm is applied to the same data files, a popu-lation of 9.2% 6 1.3% is determined. The triple markerregion contained 71.0% 6 4.9% of the cells identified withthe Treg algorithm, and conversely, 82.7% 6 3.2% of cellsidentified by the Treg algorithm were also identified bythe triple marker combination.

Asthmatics and Controls Have a Similar Frequencyof CD4dimCD25bright Treg Cells

The gating algorithm was applied to data files collectedon blood leukocytes of participants of a cross-sectionalstudy of risk factors for adult asthma in Denmark. Thenumber of T cells in the CD4dimCD25bright region was

FIG. 2. Gating strategy to objectively identify CD4dimCD25bright Treg cells between the CD252% and CD251 populations. (a) the CD251 population (redbox region) was defined such that it contained 1% of events binding an antibody isotype (blue). (b) a region containing the brightest 2% of CD25+ popula-tion (red region, CD4 MFI 5 215) that definitely is FOXP31 is contrasted with the region defined in (a) which is CD251 (CD4 MFI 5 247). This region setto contain the top 2% of the CD251 T cells in (b) was extended down to include more CD251 T cells until the MFI for CD4 of this population had increasedto 0.9 times the CD4 MFI for the region defined by the isotype in (a) (222.3). (c) This region was adjusted individually for each participant, and the cells init were enumerated. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

FIG. 3. Precision of measuring fluorescence intensity for CD4 (a) andcell number of a population of CD4dimCD25bright T cells (b). A large data-file was divided into successively shorter time intervals and CD4 mean(top line) and median (bottom lines) fluorescence intensity and the frac-tion of CD4dimCD25bright cells were plotted against the number of CD4 Tcells evaluated. [Color figure can be viewed in the online issue, which isavailable at www.interscience.wiley.com.]

FIG. 4. Expression of the Il-7 receptor CD127 on CD4dimCD25bright

Treg. Comparison of identification procedures for CD4dimCD25bright Tregusing CD4 and CD25 only (a), or in combination with CD127 (b). TheCD4dimCD25bright algorithm described here identifies 83% ofCD4dimCD25brightCD127dim cells (colored blue) and includes 29% of CD4T cells that were not CD127dim (n 5 8). [Color figure can be viewed inthe online issue, which is available at www.interscience.wiley.com.]

375TREG IDENTIFIED AS CD4dimCD25bright T CELLS

Cytometry Part A DOI 10.1002/cyto.a

determined after individual optimization of the regions.The median frequency of Treg by the method describedabove was 9.41%, with an interquartile range of (IQR)7.44%–12.85%. Asthmatics using ICS had a marginallylower number of CD4dimCD25bright cells as a fraction ofCD4 T cells (median 8.34%, IQR 7.06%–11.82%, n 5 8)than asthmatics treated with b2 agonists (10.17%, IQR8.70%–12.94%, n 5 15), asthmatics not receiving treat-ment for asthma (10.17%, IQR 8.03%–12.51%, n 5 20)and the same as controls without asthma (8.99%, IQR7.32%–13.70%, n5 109, Fig. 5).

HLA DR Expression Correlates withCD25 Expression as a Marker of

CD4dimCD25bright Treg Cells

HLA DR expression is highly significantly co-ordinatedwith CD25 expression (8) (Fig. 6). Regression lines for theCD4dimCD25bright populations and for the CD252% popula-tions were similar and separate from the regression linesfor all CD251 T cells and the CD25dim T cells. There were6.64% HLA DR positive cells in the CD41CD251gate, 19%in the CD4dimCD25bright gate, and 33.4% HLA DR1 cells inthe CD41CD252% gate (P < 0.001; n 5 144). There wasno difference between asthmatic and control populations(P > 0.117).

DISCUSSION

We have shown that FOXP31 Treg do have the pheno-type we suspected, CD4dimCD25bright. In fact, the FOXP31

T cell population is CD4dim (Fig. 1f), and this allows fordiscrimination of activated CD4 T cells that are

CD41CD251, and Treg that are CD4dimCD25bright. The dif-ference in expression of CD4 could be determined pre-cisely from the approximately 80 cells in theCD4dimCD252% gate of 4,000 CD41 T cells. It has recentlybeen shown for populations >70 that the median itself isthe best estimate for the mean (23), which is graphicallysupported by Figure 3a. Treg have low CD4 expression,side-, and forward-scatter, which is consistent with the of-ten reported inability to proliferate in the absence of IL2and IL15 (14). In fact, the reduced level of CD4 expressionmay be due to a change in cell size, and activated cells, incontrast to Treg, have high CD4 expression, as well ashigh forward and side scatters, which may relate to theiractivation and a commonly used measure for activation,imminent cell division.We reasoned that as there are almost double as many

FOXP31 T cells in the CD4dimCD25bright region as in theCD252% region, while the scatter parameters and CD4were comparable in the two regions, this region was auseful compromise between the ideal (staining for intra-cellular FOXP3) and the possible (two-color flow cytome-try to enumerate CD4dimCD25bright T cells). As the num-ber of activated CD4brightCD25bright T cells in the regiondefining CD4dimCD25bright T cells is limited by the criter-ion that the CD4 MFI of this region is 2 standard devia-tions above the mean of the CD4 MFI for the CD41CD251

T cells, the number of false positive cells is proportionalto the number of true positive cells, and thus will not dis-tort the result. This method of identification is usefulbecause it does not require time consuming intracellularstaining for FOXP3, requires only two detectors of theflow cytometer, and can be used with as little as 30,000

FIG. 5. Cells in the CD4dimCD25bright region defined in Figure 2 werecounted for each participant and were expressed as a fraction of the num-ber of CD4 T cells classified by asthma status.

FIG. 6. Correlation of MFI for HLA DR vs CD25 for CD41CD252,CD41CD251, CD4dimCD25bright, and CD4dim CD252% populations of eachpatient were plotted, and linear regression lines were drawn through thedata sets to illustrate that CD4dimCD25bright (blue) and CD4dimCD252%

(red) are very similar with regard to HLA DR and CD25 expression anddifferent from CD41CD252 (green) CD41CD251 (orange) populations.[Color figure can be viewed in the online issue, which is available atwww.interscience.wiley.com.]

376 HOFFMANN ET AL.

Cytometry Part A DOI 10.1002/cyto.a

events (~4,000 CD41 T-cells). The objective method pro-posed here to set the threshold between Treg and CD251

T cells does include some non-Treg T cells in the Treggate, but as it is technically less demanding than the tech-nique used previously (purification of CD4dimCD25bright Tcells by magnetic separation), we suggest that this draw-back is outweighed by its advantages. The collection ofcohort material was done by operators, in a different city,trained for one day on their instrument (FACS STAR) byone of the authors who was responsible for identificationof FOXP31 cells on a FACS Calibur.

Treg are also characterized by lack of expression ofthe IL-7 receptor, CD127 (9,10). The IL-7 receptor isexpressed on non-Treg T cells, and is shedded after activa-tion. We confirm that adding this marker may improvedetection of Treg (Fig. 4) and show that we find 9.1% 61.6% Treg in a small sample of healthy adults. This ishigher than the 6.35% reported elsewhere (9) and mayreflect subtle differences in gating strategy. As theCD4dimCD25bright algorithm presented here is objective, itmay prove to be more reproducible by independent inves-tigators than gating on CD41CD25brightCD127dim cells.

The fraction of CD4dimCD25bright T cells detected here(10.88% 6 4.82%) is similar to the number reported byothers (1). Only a fraction of these express FOXP3 (6% 61%, n 5 6), but the number of FOXP31 T cells is basedon a smaller number of observations obtained by a longintracellular staining protocol, whereas the number ofCD4dimCD25bright T cells was determined by surface stain-ing. Untreated mild asthmatics had marginally more Tregthan asthmatics treated with ICS and controls. This con-curs with other studies of Treg (5) and may be becauseTreg exert their effect by ‘‘infectious tolerance’’ where theIL10 that is part of their phenotype is not made by Tregthemselves, but by T cells in physical contact with Treg(6,24). Asthmatics treated with ICS express increasedlevels of FOXP3, and treatment of severe asthmaticswith systemic glucocorticoid leads to an increase ofCD41CD25brightCD45RO1CD62L1 T cells (25). This studywas done on severe asthmatics who may be considered tohave a considerable chronic element of inflammation intheir disease. In the present study, cases had mild, inter-mittent asthma, and only nine were in ICS treatment.

In previous work on regulatory T cells in allergic asthma(reviewed in (26,27)), the focus was on either function of Tcells (cytokine secretion or inhibition of T cell proliferation),which typically were evaluated on a per cell basis (4,28,29),or on the total CD41CD251 population (5,30). We havedesigned an algorithm to more precisely identify Treg asCD4dimCD25bright T cells in circulation on the basis of sur-face markers and have shown that asthmatics and nonasth-matics have similar fractions of CD4dimCD25bright T cells.

Recently, we showed that CD41CD45RO1 T cells,expressing Treg associated markers CD25 and CD152, aswell as plasmacytoid dendritic cells, were recruited fromcirculation when wheat allergics with respiratory allergywere challenged with wheat in a double blind placebocontrolled food challenge. Although there were clear signsof allergic inflammation, none of the patients reported

symptoms, suggesting that we observed potent immuneregulation in vivo (31). The regulation of the inflammatoryresponse may have been mediated by FOXP31 T cells inthe CD41CD25bright T cell population and may have beenamplified by the plasmacytoid dendritic cells recruitedfrom circulation at the same time.HLA DR is thought to be an activation marker rather

than a marker of Treg, but it has repeatedly been shownto be associated with Treg (8,22). As Treg are known toinhibit proliferation of T cells by contact (14,32), HLA DRmay be the part of the receptor complex (with moleculeslike CD152 and PD1) that restricts inhibition to cellsexpressing CD4, as the TCR associated with CD4 is theligand for HLA Class II. There appears to be a continuumof CD41 cells that express increasing amounts of CD25,HLA DR, and FoxP3 that become increasingly regulatory.We can show that there is coordinated expression ofCD25 and HLA DR, and it has been shown elsewhere thatCD25 and FOXP3 are coordinate expressed (7). The factthat there was no difference in expression of HLA DRbetween asthmatic patients and the control populationconcurs with published data (4,5). Although question-naire defined diagnosis of asthma, like the ECRHS-basedsurvey applied here, have a high Youden index, data arelikely to improve once the clinical diagnosis is used toassign persons to either asthma or control on the basis ofobjective clinical measures (33). A more detailed analysisof CD4dimCD25bright T cell populations will be made oncemore clinical data is available.

CONCLUSION

We have shown that FOXP31 T cells areCD4dimCD25bright, have characteristics associated withquiescent cells, and that these cells accumulate in aregion defined as CD25bright and CD4dim�0.9*CD4 MFI.These cells express HLA DR co-ordinately with CD25and do not express CD127. In an ECRHS-like survey, wewere able to show that asthmatics and a control popula-tion have similar frequencies of Treg.

ACKNOWLEDGMENTS

We thank Anne Marie Toft, Anna-Louise Fjordside Kristen-sen, and Anitta Holmager for excellent technical assistanceand Dan Knudsen of IDWorks for help with the artwork.

LITERATURE CITED1. Piccirillo CA, Shevach EM. Naturally-occurring CD41CD251 immu-

noregulatory T cells: Central players in the arena of peripheral toler-ance. Semin Immunol 2004;16:81–88.

2. Fujishima M, Hirokawa M, Fujishima N, Sawada K. TCRab repertoirediversity of human naturally occurring CD41CD251 regulatory Tcells. Immunol Lett 2005;99:193–197.

3. Hoffmann P, Eder R, Boeld TJ, Doser K, Piseshka B, Andreesen R, Edin-ger M. Only the CD45RA1 subpopulation of CD41CD25high T cellsgives rise to homogeneous regulatory T-cell lines upon in vitro expan-sion. Blood 2006;108:4260–4267.

4. Smith TRF, Alexander C, Kay AB, Larche M, Robinson DS. Cat allergenpeptide immunotherapy reduces CD4(1) T cell responses to cat aller-gen but does not alter suppression by CD4(1) CD25(1) T cells: Adouble-blind placebo-controlled study. Allergy 2004;59:1097–1101.

5. Shi H, Li S, Xie Z, Qin X, Qin X, Zhong X. Regulatory CD41CD251 Tlymphocytes in peripheral blood from patients with atopic asthma.Clin Immunol 2004;113:172–178.

377TREG IDENTIFIED AS CD4dimCD25bright T CELLS

Cytometry Part A DOI 10.1002/cyto.a

6. Jonuleit H, Schmitt E, Kakirman H, Stassen M, Knop J, Enk AH. Infec-tious tolerance: Human CD25(1) regulatory T cells convey suppres-sor activity to conventional CD4(1) T helper cells. J Exp Med2002;196:255–260.

7. Roncador G, Brown PJ, Maestre L, Hue S, Martinez-Torrecuadrada JL,Ling K, Pratap S, Toms C, Fox BC, Cerundolo V, Powrie F, Banham AH.Analysis of FOXP3 protein expression in human CD41CD251 regula-tory T cells at the single-cell level. Eur J Immunol 2005;35:1681–1691.

8. Baecher-Allan C, Brown JA, Freeman GJ, Hafler DA. CD41CD25highregulatory cells in human peripheral blood. J Immunol 2001;167:1245–1253.

9. Seddiki N, Santner-Nanan B, Martinson J, Zaunders J, Sasson S, LandayA, Solomon M, Selby W, Alexander SI, Nanan R, Kelleher A, Fazekasde St Groth B. Expression of interleukin (IL)-2 and IL-7 receptors dis-criminates between human regulatory and activated T cells. J ExpMed 2006;203:1693–1700.

10. Liu W, Putnam AL, Xu-Yu Z, Szot GL, Lee MR, Zhu S, Gottlieb PA,Kapranov P, Gingeras TR, Fazekas de St Groth B, Clayberger C, SoperDM, Ziegler SF, Bluestone JA. CD127 expression inversely correlateswith FoxP3 and suppressive function of human CD41 T reg cells. JExp Med 2006;203:1701–1711.

11. Hansen W, Loser K, Westendorf AM, Bruder D, Pfoertner S, Siewert C,Huehn J, Beissert S, Buer J. G protein-coupled receptor 83 overexpres-sion in naive CD41CD25- T cells leads to the induction of Foxp31regulatory T cells in vivo. J Immunol 2006;177:209–215.

12. Sugimoto N, Oida T, Hirota K, Nakamura K, Nomura T, Uchiyama T,Sakaguchi S. Foxp3-dependent and -independent molecules specificfor CD251CD41 natural regulatory T cells revealed by DNA microar-ray analysis. Int Immunol 2006;18:1197–1209.

13. Levings MK, Sangregorio R, Sartirana C, Moschin AL, Battaglia M,Orban PC, Roncarolo M. Human CD251CD41 T suppressor cellclones produce transforming growth factor b, but not interleukin 10,and are distinct from type 1 T regulatory cells. J Exp Med 2002;196:1335–1346.

14. Dieckmann D, Plottner H, Berchtold S, Berger T, Schuler G. Exvivo isolation and characterization of CD4(1)CD25(1) T cells withregulatory properties from human blood. J Exp Med 2001;193:1303–1310.

15. Shimizu J, Yamazaki S, Takahashi T, Ishida Y, Sakaguchi S. Stimulationof CD25(1)CD4(1) regulatory T cells through GITR breaks immuno-logical self-tolerance. Nat Immunol 2002;3:135–142.

16. Janossy G. Clinical flow cytometry, a hypothesis-driven discipline ofmodern cytomics. Cytometry Part A 2004;58A:87–97.

17. Standards for the diagnosis and care of patients with chronic obstruc-tive pulmonary disease (COPD) and asthma. This official statement ofthe American Thoracic Society was adopted by the ATS Board ofDirectors, November 1986. Am Rev Respir Dis 1987;136:225–244.

18. Agrawal DK, Bharadwaj A. Allergic airway inflammation. Curr AllergyAsthma Rep 2005;5:142–148.

19. Leipe J, Skapenko A, Lipsky PE, Schulze-Koops H. Regulatory T cellsin rheumatoid arthritis. Arthritis Res Ther 2005;7:93.

20. Kelsen J, Agnholt J, Hoffmann HJ, Romer JL, Hvas CL, Dahlerup JF.FoxP3(1)CD4(1)CD25(1) T cells with regulatory properties can becultured from colonic mucosa of patients with Crohn’s disease. ClinExp Immunol 2005;141:549–557.

21. Burney PG, Luczynska C, Chinn S, Jarvis D. The European communityrespiratory health survey. Eur Respir J 1994;7:954–960.

22. Baecher-Allan C, Wolf E, Hafler DA. Functional analysis of highlydefined, FACS-isolated populations of human regulatory CD41CD251 T cells. Clin Immunol 2005;115:10–18.

23. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variancefrom the median, range, and the size of a sample. BMC Med Res Meth-odol 2005;5:13.

24. Kearley J, Barker JE, Robinson DS, Lloyd CM. Resolution of airwayinflammation and hyperreactivity after in vivo transfer ofCD41CD251 regulatory T cells is interleukin 10 dependent. J ExpMed 2005;202:1539–1547.

25. Karagiannidis C, Akdis M, Holopainen P, Woolley NJ, Hense G, Ruck-ert B, Mantel P, Menz G, Akdis CA, Blaser K, Schmidt-Weber CB. Glu-cocorticoids upregulate FOXP3 expression and regulatory T cells inasthma. J Allergy Clin Immunol 2004;114:1425–1433.

26. Shi H, Qin X. CD4CD25 regulatory T lymphocytes in allergy andasthma. Allergy 2005;60:986–995.

27. Akdis M, Blaser K, Akdis CA. T regulatory cells in allergy. Chem Immu-nol Allergy 2006;91:159–173.

28. Ling EM, Smith T, Nguyen XD, Pridgeon C, Dallman M, Arbery J, CarrVA, Robinson DS. Relation of CD41CD251 regulatory T-cell suppres-sion of allergen-driven T-cell activation to atopic status and expressionof allergic disease. Lancet 2004;363:608–615.

29. Akdis CA, Blesken T, Akdis M, Wuthrich B, Blaser K. Role of inter-leukin 10 in specific immunotherapy. J Clin Invest 1998;102:98–106.

30. Lara-Marquez ML, Moan MJ, Cartwright S, Listman J, Israel E, PerkinsDL, Christiani DC, Finn PW. Atopic asthma: Differential activationphenotypes among memory T helper cells. Clin Exp Allergy 2001;31:1232–1241.

31. Hoffmann HJ, Skjold T, Raithel M, Adolf K, Hilberg O, Dahl R.Response of respiratory flour allergics in an ingested flour challengemay involve plasmacytoid dendritic cells, CD251 and CD1521 Tcells. Int Arch Allergy Immunol 2006;140:252–260.

32. Jonuleit H, Schmitt E, Stassen M, Tuettenberg A, Knop J, Enk AH.Identification and functional characterization of human CD4(1)CD25(1) T cells with regulatory properties isolated from peripheralblood. J Exp Med 2001;193:1285–1294.

33. Sistek D, Tschopp JM, Schindler C, Brutsche M, Ackermann-LiebrichU, Perruchoud AP, Leuenberger P. Clinical diagnosis of currentasthma: Predictive value of respiratory symptoms in the SAPALDIAstudy. Swiss Study on Air Pollution and Lung Diseases in Adults. EurRespir J 2001;17:214–219.

378 HOFFMANN ET AL.

Cytometry Part A DOI 10.1002/cyto.a