Human CD4+ regulatory T cells express lower levels of the IL-7
receptor alpha chain (CD127), allowing consistent
identification and sorting of live cells
Dennis J. Hartigan-O'Connora, Chungkee Poona,b,
Elizabeth Sinclairb, Joseph M. McCunea,⁎
aDivision of Experimental Medicine, Department of Medicine, University of California, San Francisco, San Francisco, CA 94143, United States
bCore Immunology Laboratory, Department of Medicine, University of California, San Francisco, San Francisco, CA 94143, United States
Received 7 June 2006; received in revised form 11 September 2006; accepted 2 October 2006
Available online 3 November 2006
would greatly facilitate the analysis of these cells in disease states. In an attempt to identify markers that are sensitive and specific for
human T-regs, we analyzed the expression of fourteen intracellular and cell surface markers on human CD4+cells. Many markers were
partiallyselective forCD25hiT-regs,but consistentand specificdiscriminationoffunctionalT-regswas onlymadepossible by focus on
CD127, the alpha chain ofthe IL-7 receptor.AlthoughmostCD4+humanTcellsexpressCD127,T-regsexhibiting suppressive activity
in vitro display distinctly lower surface expression of this marker, irrespective of their level of CD25 expression. Sorted cells with the
surface phenotype CD4+CD25+CD127lowhad higher levels of intracellular FOXP3 and CTLA-4 and, as determined by functional
assays, were suppressive, hypoproliferative, and poorly responsive to TCR signaling. The CD4+CD25+CD127lowphenotype was also
regulatory Tcells in disease states as well as for enrichment of live regulatory Tcells for functional analyses and/or expansion in vitro.
© 2006 Published by Elsevier B.V.
Keywords: Human; T cells; Cell surface molecules; Transcription factors
to be important for generation and maintenance of the
cells), contain intracellular FOXP3/scurfin (a transcription
factor encoded by the foxp3 gene), and suppress the
proliferation of other cells in a contact-dependent manner
(Mills, 2004). The responsiveness of other cells to sup-
pression depends on the strength of stimulus delivered,
with especially strong stimuli able to overcome the influ-
ence of T-regs (Baecher-Allan et al., 2001). Most murine
natural T-regs emigrate from the thymus after day three of
life; day three thymectomy therefore induces an autoin-
flammatory state predisposing to autoimmune disease
(Sakaguchi et al., 1995; Asano et al., 1996). Some have
Journal of Immunological Methods 319 (2007) 41–52
Abbreviations: T-regs, regulatory T cells; HIV, human immunode-
ficiency virus; PBMC, peripheral blood mononuclear cells; TCR, Tcell
receptor; CV, coefficient of variation.
⁎Corresponding author. Tel.: +1 415 206 8101; fax: +1 415 826 8449.
E-mail address: email@example.com (J.M. McCune).
0022-1759/$ - see front matter © 2006 Published by Elsevier B.V.
Fig. 1. Relationship of fourteen potential T-reg markers to CD25 expression among CD3+CD4+cells. The vertical red line in each cytogram is placed
at the 95th percentile for CD25 intensity; cells to the right of the line are considered CD25hi. The percentage of CD3+CD4+CD25hicells positive for
each marker (i.e., above the horizontal red line indicating maximal intensity observed in the FMO control) is indicated. The percent positive
expression of individual markers was derived by comparison of fully stained samples to “fluorescence minus one” (FMO) controls for each marker.
Fully stained and FMO samples were gated on CD4+CD25hicells and a “positive” gate drawn so that fewer than 0.1% of cells in the FMO control
were included. The percentageof CD4+CD25hicellsin the resulting positivegate is indicated on the figure. Eachmarkerwas testedbetween three and
ten times on cells of at least three different subjects; representative examples are shown in this figure and data from all experiments are summarized in
Table 1. A. Nine markers with poor ability to distinguish CD25hicells from other CD3+CD4+cells in the peripheral blood. Expression of some
markers (e.g., GITR, CCR6, CD27, and CD28) was nearly uniform on CD4+Tcells. Expression of other markers (e.g., CD45RO, CD45RA, CD62L,
and β7-integrin) was more variable but did not effectively distinguish CD25hicells from CD25−/lowcells. B. CD38 was variable across the spectrum
of CD25 expression, while HLA-DR expression was specific to a small fraction of CD25hicells. C. Markers distinguishing CD25hicells with
reasonable sensitivity and specificity: FOXP3, intracellular CTLA-4, and CD127low. Note that the horizontal red line in the cytogram depicting
CD127 expression separates CD127hicells from CD127lowcells. D. Distribution of markers on gated CD25hi cells. The vertical red lines in these
histograms separate marker-negative cells (to the left, with intensity comparable to that of FMO controls) from marker-positive cells (to the right).
(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
42D.J. Hartigan-O'Connor et al. / Journal of Immunological Methods 319 (2007) 41–52
toward self-antigens (Cozzo et al., 2003), though self-
2004). TCR signaling greatly increases the regulatory
activity of these cells.
Natural T-regs share many characteristics of activated
cells and, indeed, activated human (but not murine)
CD4+T cells can up-regulate FOXP3 and acquire regu-
latory characteristics (Walker et al., 2003; Aandahl et al.,
2004). It has been shown, however, that expression of
FOXP3 alone is insufficient to confer regulatory
activity. Accordingly, it is not clear whether all T cells
that express FOXP3 are suppressive (Allan et al., 2005).
The cells may also play a “gateway” role in the adaptive
response to microbes, as Toll-like receptor signaling in
antigen-presenting cells renders responding T cells
resistant to suppression (Pasare and Medzhitov, 2003).
There is tremendous interest in elucidating the role of
T-regs in human disease, including autoimmune disease,
cancer, and infectious disease. For example, several
investigators have hypothesized that this cell population
is important in the pathogenesis of HIV disease. The
available data suggest that there is a similar frequency of
CD25-expressing cells in the peripheral blood of HIV-
infected and-uninfected patients, though marginally more
CD25hicells are found in peripheral blood from some
HIV+ patients (Aandahl et al., 2004; Kinter et al., 2004).
The biologic significance of such increased numbers
remains obscure. On one hand, the ability of CD25+
T cells to suppress proliferative and cytokine responses
against HIV antigens may lead to accelerated disease
progression (Aandahl et al., 2004; Kinter et al., 2004;
Rouse and Suvas, 2004). On the other hand, suppression
of immune activation by T-regs may slow or prevent
disease progression because inflammation is dampened
(Kinter et al., 2004; Eggena et al., 2005). One reason for
the disparate conclusions among these investigators isthe
lack of a validated, reproducible measurement of human
T-regs. Clarification of the unresolved issues regarding
T-regs in this and other settings will depend upon the
ability to isolate phenotypically and functionally homoge-
neous populations of T-regsfrom humanperipheralblood.
Human natural T-regs exist at only low frequency in
adult peripheral blood, usually estimated at 1–2% of
CD4+T cells (Baecher-Allan et al., 2001), whereas
murine T-regs represent 5–10% of splenic or thymic
CD4+CD8−cells (Yagi et al., 2004). It is possible that
humans contain fewer natural T-regs than mice or that T-
regs are more common in the lymphoid organs than in
peripheral blood. The frequency of natural T-regs may
also vary in young and old animals. Indeed, fetal human
tissues contain extraordinarily high numbers of regula-
tory T cells (Cupedo et al., 2005; Darrasse-Jeze et al.,
2005; Michaelsson et al., 2006).
To explore the role of T-regs in HIV disease and other
human diseases, reliable quantitative and functional
assays are required. Unfortunately, the relative rarity of
T-regs, their extensive phenotypic overlap with activat-
ed cells, and the lack of reliable surface markers have
prevented development of such assays. The CD25hi
phenotype may under- or over-estimate T-reg content in
clinical samples, depending on the clinical situation as
well as on the way in which the gate for high expression
of CD25 is determined. The expression of intracellular
FOXP3 is likely more specific than CD25 but is also an
insensitive measure, often yielding fewer cells than the
CD25higate. Furthermore, this intracellular marker
cannot be used to sort live cells for functional analyses.
To facilitate the study of T-regs in human disease, we
searched for surface markers that would facilitate
consistent quantitative detection and enrichment of
viable T-regs from samples of human peripheral blood.
Distribution of cells with candidate T-reg phenotypes within CD3+CD4+
Phenotype% of CD3+
% of CD3+CD4+
aValues shown are a mean of three to ten measurements made using
“fluorescence minus one” controls. The table indicates the following
three characteristics of the populations shown:
bPercentageof CD3+CD4+cells with the indicatedphenotype. Markers
that are specific for T-regs should be displayed on a reasonably small
fraction of CD3+CD4+cells.
cPercentage of cells with the indicated phenotype that exceed the 95th
percentile for CD25 expression among CD3+CD4+cells. Markers that
are specific for T-regs should delimit populations containing a large
fraction of CD25hicells.
dPercentage of all CD3+CD4+CD25hicells that display the indicated
phenotype. Markers that are sensitive for T-regs should be found on a
large fraction of CD25hicells.
43 D.J. Hartigan-O'Connor et al. / Journal of Immunological Methods 319 (2007) 41–52
2. Materials and methods
2.1. Human blood and cell samples
HIV-negative blood donors were adult volunteers.
Whole blood samples were used for most flow
cytometric analysis. When required for sorting experi-
ments, PBMC were purified on ficoll step gradients
according to standard methods (Coligan, 2001).
2.2. Antibodies and flow cytometry
these studies: CD3 (clone UCHT1, Pacific Blue-labeled,
44 D.J. Hartigan-O'Connor et al. / Journal of Immunological Methods 319 (2007) 41–52
from Pharmingen), CD4 (clone SK3, custom AmCyan-
labeled, from BD Biosciences), CD8 (clone RPA-T8,
Alexa Fluor 700-labeled, from Pharmingen), CD25
(clone MA-251, PE- or PE-Cy7- or APC-Cy7-labeled,
from Pharmingen), CD27 (clone M-T271, FITC-labeled,
from Pharmingen), CD28 (clone CD28.2, custom APC-
Cy7-labeled, from BD Biosciences), CD38 (clone HB7,
PE-Cy7-labeled, from BD Biosciences), CD45RA (clone
2H4, ECD-labeled, from Immunotech), CD45RO (clone
UCHL1, ECD-labeled, from Immunotech), CD62L
(clone Dreq56, APC-labeled, from Pharmingen; and
clone SK11, FITC-labeled, from BD Bioscience),
CD127 (clone R34.34, PE-labeled, from Beckman–
Coulter; and clone hIL-7R-M21, PE-labeled, from BD
Biosciences), CCR5 (clone 2D7/CCR5, APC-labeled,
from Pharmingen), CCR6 (clone 53103.111, from R&D
Systems, conjugated to AlexaFluor 488 using a kit from
Molecular Probes), CTLA-4 (clone BNI3, PE- or PE-
Cy5-labeled, from Pharmingen), GITR (clone EB11,
FITC-labeled, from eBiosciences), HLA-DR (clone
L243, PerCP-Cy5.5-labeled, from Pharmingen), β7-
integrin (clone FIB504, PE-Cy5-labeled, from Pharmin-
gen; and clone FIB504, APC-labeled, from Pharmingen),
and FOXP3 (clone PCH101, APC-labeled, from
eBioscience; and clone 206D, AlexaFluor 647-labeled,
from BioLegend). When staining for CD25 and CD127
simultaneously, PE-conjugated anti-CD127 was used in
combination with APC-Cy7 or PE-Cy7-conjugated anti-
CD25. These combinations yielded comparable results,
although CD25 staining was somewhat brighter when
using the PE-Cy7 conjugate. For whole blood staining, a
mixture of antibodies against surface markers [complete
or “fluorescence minus one” (FMO)] was placed into
12×75 mm polystyrene tubes (BD Biosciences, San
Jose, CA). 50 μl of whole blood were then added and the
mixture was incubated for 15 min at room temperature.
1–2 ml of FACSlyse (BD Biosciences) were added and
lysis allowed to proceed for 10 min at room temperature.
The tubes were filled to the top with calcium- and
magnesium-free PBS containing 2% FBS (stain buffer)
and centrifuged to collect the cells. Following centrifu-
gation, if necessary, intracellular staining was performed
according toinstructionsprovidedbythe manufacturerof
the antibody. Purified peripheral blood mononuclear cells
(PBMC) were stained according to a similar protocol,
except that the antibody mixture was added to cells
suspended in 50 μl of stain buffer. Data were collected on
a BD LSRII and analyzed using FlowJo software. The
percent positive expression of individual markers was
derived by comparison of fully stained samples to
“fluorescence minus one” (FMO) controls for each
marker. Each of these two samples was gated on CD4+
CD25hicells (defined as the brightest 5% of CD4+ cells)
and a “positive” gate was drawn so that fewer than 0.1%
of cells in the FMO control were included. The
percentage of cells from the fully stained sample in the
resulting positive gate was taken to represent the
proportion of positive cells.
2.3. Flow sorting
Four-color sorts were performed using a BD Vantage
sorter with the DIVA option. Cells were labeled using the
anti-CD3, anti-CD4, anti-CD25, and anti-CD127 anti-
set to encompass an obviously distinct population, as
shown in Fig. 2A. The CD25hisubset of this gate was
expression than found on any CD127hicell. Therefore,
cells in the CD25+CD127hiand CD25+CD127lowgates
had an equivalent level of CD25 expression but different
levels of CD127 expression.
Fig. 2. CD127lowcells suppress T cell proliferation in vitro. A. The CD25+/hiCD127lowpopulation is clearly separable from other cells (left). This
population wasdividedinto four sub-populations sothatT-regfunctioncouldbetestedinCD25+CD127hiand CD25+CD127lowcells(right). The CD25hi
subset was chosen to include all cells with a higher intensity of CD25 expression than found on any CD127hicell. Therefore, cells in the CD25+CD127hi
cells served as negative and positive controls for T-reg characteristics, respectively. B. CD127lowpopulations contain more intracellular FOXP3 and
CTLA-4 than do CD127hipopulations. FOXP3 and CTLA-4 stains are of low intensity and do not allow clear separation of negative from positive cells.
atleftshowthepercentage of CFSE-dim cells inharvestedcultures, exclusiveof CFSE-negativecells(candidateregulators added backtothe culturesare
CFSE-negative).UsingFlowJo'sproliferationplatformtoanalyze proliferationofCD8+cellsyielded equivalentresults,asshownintherightmostgraph.
proliferation by sorted CD127lowcells from a second subject. The histograms show dilution of CFSE in cultures with candidate T-regs added to CD25-
of CFSE-negative, CD3+CD4+cells recovered after the five-day culture period.
45D.J. Hartigan-O'Connor et al. / Journal of Immunological Methods 319 (2007) 41–52
2.4. Suppression assays
2, BD Biosciences) was added to wells of 96-well, U-
bottom plates. The plates were incubated for at least 4 h
and then washed three times with PBS before use.
Responder cells were prepared by depletion of purified
PBMC using anti-CD25-labeled paramagnetic microbe-
ads and MS columns, according to the instructions of the
manufacturer (Miltenyi Biotec, Auburn, CA). The cells
were resuspended in PBS containing 5% FBS at con-
centrations from 0.5–10×106per ml. One-ninth volume
of diluted CFSE stock (50 μM) was then added and the
cell suspension inverted several times to mix. The cells
washed at least three times in the same solution. Finally,
labeled responder cells were resuspended in medium
(RPMI 1640 supplemented with 2 nM L-glutamine,
5 mM HEPES, and 100 U/ml penicillin/mg/ml strepto-
mycin). Candidate T-regs were prepared either by flow
sorting, as described above, or by positive selection on
MS columns. These cells were not labeled with CFSE.
1.5×105responder cells were then added to antibody-
Fig. 3. ResponseofT-regpopulationstoTCRsignaling.TheindicatedpopulationswereseparatedbyflowsortingusingthegatesillustratedinFig.2A.Twenty
thousand cells of each type were then cultured in the presence of plate-bound anti-CD3 and soluble anti-CD28 antibodies. Cells were recovered and surface-
line on the cytograms separates sorted CD25+from CD25hicells at time zero, while the horizontal black line separates 127hifrom 127lowcells at time zero.
46 D.J. Hartigan-O'Connor et al. / Journal of Immunological Methods 319 (2007) 41–52
coated wells of a 96-well plate alone or together with 5.0,
3.75, 3.0, or 1.5×104candidate regulatory cells. Plates
were placed in the incubator for 5 days before harvesting,
staining, and analysis using the LSRII and FlowJo
software as described above.
2.5. Stimulation of sorted cells
Sorted regulatory populations were stimulated in 96-
well plates using techniques similar to those described
for suppression assays. Plates were coated with anti-
CD3 antibody as described above, then 20,000 sorted
cells were added to each well in 100 μl of medium.
2.5 ng of anti-CD28 antibody (clone CD28.2, BD
Biosciences) were added. Cells were collected and
stained immediately or after 4, 16, or 64 h.
Female scurfy carrier mice (B6.Cg-Foxp3sf/J) were
obtained from The Jackson Laboratory (Bar Harbor,
ME). These carrier mice were bred to C57BL/6 males in
a barrier facility. One affected male and three unaffected
male siblings were chosen for analysis at 2 weeks of age.
Genotypes were confirmed by PCR analysis according
to the protocol suggested by The Jackson Laboratory.
2.7. Rhesus blood samples
Rhesus macaque blood was obtained from animals
Rhesus macaques (Macaca mulatta) were housed in
accordance with the regulations of the American Associ-
ation for Accreditation of Laboratory Animal Care
Standards at the California National Primate Research
12 ml/kg/month in accordance with CNPRC SOPs) were
obtained from each animal by femoral venipuncture under
ketamine anesthesia (10 mg/kg, given intramuscularly).
3.1. Specificity of putative T-reg cell markers on human
To allow identification and isolation of T-regs from
human subjects, we searched for cell surface markers
that would discriminate CD3+CD4+CD25hiT-cells from
all other CD4+cells (i.e., markers that were specific for
the CD25hiphenotype) and that were expressed on a
large fraction of CD25hiT cells (i.e., markers that were
sensitive for the CD25hiphenotype).
Markers chosen for study included CD27 (Ruprecht
et al., 2005), CD28 (Bour-Jordan et al., 2004), CD38,
CD127, CCR5 (Wysockietal.,2005), CCR6 (Earle etal.,
2005), CTLA-4 (intracellular), GITR (McHugh et al.,
Fig. 4. Assay characteristics. A. CD25+/hiCD127lowcells detected in
seven normal volunteersover time. The CD25+/hiCD127lowgate drawn
is similar to the example shown in the left panel of Fig. 2A and
includes both CD25+CD127lowand CD25hiCD127lowcells. Each data
point represents a sample separated from the others by approximately
1 week. B. CD25+/hiCD127lowcells detected in whole blood (WB),
ficoll-separated PBMC, and cryopreserved (cryo) PBMC from three
volunteers. Single samples drawn from these volunteers were analyzed
successively after preparation in three different ways. Cryopreserved
cells were frozen for 1 week before analysis. C. Intra-assay variability
measured in whole blood samples. Five aliquots of blood were taken
fromsingletubesdrawn fromeach subjectandthenassayedin parallel.
47 D.J. Hartigan-O'Connor et al. / Journal of Immunological Methods 319 (2007) 41–52
2002), HLA-DR, β7-integrin (Stassen et al., 2004), and
FOXP3. The expression of these markers on CD3+CD4+
peripheral blood T cells was analyzed in relation to the
expression of CD25 (Fig. 1). We expected that markers
sensitive for regulatory cells should be uniformly present
on the brightest of CD25hicells, which are generally
considered to be those with the most potent regulatory
activity (Baecher-Allan et al., 2001). In addition, we
expected that specific markers for regulatory cells should
be mostly absent from CD25lowand CD25−cells, since
the majority of these cells presumably do not have
48 D.J. Hartigan-O'Connor et al. / Journal of Immunological Methods 319 (2007) 41–52
Nine of fourteen markers were not useful for dis-
crimination of CD25hicells (Fig. 1A). Expression levels
of GITR, CCR6, CD27, and CD28 were nearly uniform
on CD3+CD4+T cells. In occasional subjects, CCR5
expression was confined to the CD25hipopulation, but
this finding was not consistent. The surface phenotypes
CD45RO+, CD45RA−, CD62L+, and β7-integrin−were
observed on a large fraction of CD25hicells, but also on
many CD25−/lowT cells (Fig. 1A and Table 1, second
Expression of CD38 was variable within the CD25hi
population and might identify CD38−and CD38+
of HLA-DR was observed on a small fraction (10%) of
CD25hiTcells and demarcates a subpopulation of T-regs
(Table 1, last column; Baecher-Allan et al., 2006).
In contrast to the above cell surface markers, intra-
cellular FOXP3, intracellular CTLA-4, and surface
CD127 consistently distinguished the highest of CD25hi
As has been reported by others (Roncador et al., 2005),
the expression of intracellular CTLA-4 and FOXP3 was
positively correlated with surface CD25 expression.
ing that use of these markers for regulatory cell detection
may be insensitive (Table 1, last column). Plotted against
distribution and require drawing of a subjective positive
gate. Surface CD127, on the other hand, offered promise
for accurately enumerating and even sorting live human
regulatory cells: 80% of CD3+CD4+cells express high
levels of CD127 while 20% are CD127-negative or -low
(Table 1). 84% of CD25hicells, however, are CD127low
(Table 1, last column).
3.2. Lower levels of CD127 separate regulatory T cells
from other CD4+CD25+cells in humans
PBMCs from healthy human subjects contained a
mean of 8% (range 5–12%, n=10) CD3+CD4+CD25+/hi
CD127lowcells (percentage of CD3+CD4+cells; Fig. 2A
whetherthese CD127lowTcellscomprisedpopulations of
T-regs. Four subpopulations were first purified by cell
sorting: CD25lowCD127hi, CD25+CD127hi, CD25+-
CD127low, and CD25hiCD127low(Fig. 2A). These pop-
ulations were then tested for phenotypic and functional
characteristics commonly associated with T-regs. We hy-
pothesized that both of the CD127lowpopulations would
exhibit the classic characteristics of regulatory cells, de-
spite their different levels of CD25 expression.
Based on the mean fluorescence intensity of an
intracellular stain, CD25+CD127lowcells were found to
contain more FOXP3 than CD25+CD127hicells but
marginally less than the CD25hiCD127lowpopulation
(Fig. 2B). CD25+CD127lowcells were also intermediate
in CTLA-4 expression, with levels above those ob-
served in CD25+CD127hicells but below those ob-
served in the CD25hiCD127lowpopulation (Fig. 2B).
Virtually all CD25+CD127lowcells were in the
CD45RA−CD62L+memory subset (data not shown).
These candidate regulatory populations were tested
directly for suppressive activity in proliferation assays
(Fig. 2C, D). CD25−responder cells were prepared by
bead depletion, labeled with CFSE, and added to plates
coated with anti-CD3 antibodies. Sort-purified, candidate
analyzed by flow cytometry. CD25+CD127lowcells were
found to be as suppressive as their CD25hicounterparts at
all ratios tested. CD127hicells, by contrast, showed no
suppressive activity, regardless of their level of CD25
expression. Similar results were obtained in four consec-
utive sorting experiments, of which two are illustrated in
the figure. Note that, in Fig. 2, the populations labeled
CD25+CD127hiand CD25+CD127lowdisplayed equiva-
lent levels of surface CD25. These two populations, one
suppressiveandthe other not,couldonlybedistinguished
from one another on the basis of CD127 expression.
CD25+CD127lowcells as well as CD25hiCD127low
cells were hypoproliferative, as they failed to proliferate
in response to the anti-CD3 antibodies used in our assay.
In contrast, CD127hicells divided readily (Fig. 2E).
CD25+CD127lowcells therefore displayed all the
classic characteristics of regulatory cells: expression of
FOXP3 and CTLA-4, the ability to suppress proliferation
Fig. 5. T-regsofmiceandmacaquesareCD127low.A.DistributionofCD25andCD127onCD4+cellsofwild-typemice(leftmostpanel)andscurfymice
to have lower levels of CD127. Scurfy mice (at right) lack Foxp3, T-regs, and CD25+CD127lowcells. However, these mice contain a population of
CD25+CD127+cells that are largely absent from healthy mice. B. Distribution of CD25 and CD127 on CD4+cells of eight rhesus macaques. Staining of
CD25-negative gates, which encompass the CD127+population. C. Suppression by CD25+cells from macaque peripheral blood. The histograms show
after addition of CD25+cells to CD25-depleted PBMC at a ratio of 1:3. The division index (DI) is indicated on each histogram.
49 D.J. Hartigan-O'Connor et al. / Journal of Immunological Methods 319 (2007) 41–52
of other Tcells, and hypoproliferation in response to anti-
CD3 antibodies. Indeed, this population is clearly
contiguous with the CD25hiCD127lowpopulation and a
division between them could only be made arbitrarily.
Nevertheless, we observed a trend in our assays toward
more pronounced regulatory characteristics in the subset
that had higher expression of CD25.
3.3. Reduced response of T-reg populations to
TCR signaling and cell activation seem to be required
for regulatory T cells to exert their suppressive effects
(Thornton and Shevach, 1998). Given the tendency of
CD25+CD127lowcells to exert less suppression than
CD25hiCD127lowcells, this population may contain
inactive or partially active precursors to fully activated
To test this possibility, CD25+CD127lowand CD25hi
CD127lowcells were sort purified and activated by
plate-bound anti-CD3 and soluble anti-CD28 antibo-
dies. As expected, both CD127lowpopulations were
relativelyresistant to activation through the TCR (Fig. 3,
second row). It is clear, however, that CD25+CD127low
cells up-regulate CD25 and FOXP3 in response to
sustained stimulation (16–64 h; Fig. 3, third and fourth
rows), suggesting that CD25+T-regs are not fundamen-
tally different from their CD25hicounterparts.
It has been reported that human CD4+CD25−cells
can up-regulate FOXP3 and acquire suppressive capac-
ity under some stimulatory conditions (Walker et al.,
2003; Aandahl et al., 2004). We found that CD25+
CD127hicells did not down-regulate CD127 or up-
regulate FOXP3 after 4 h of stimulation, although
expression of both molecules changed detectably after
longer periods. Eleven percent of purified CD25low
CD127hicells expressed FOXP3 and most down-
regulated CD127 after 64 h of stimulation (Fig. 3).
Others have demonstrated that these stimulated cells
acquire in vitro suppressive ability (Walker et al., 2003;
Aandahl et al., 2004). Therefore, CD127 is down-
regulated coordinately with the increase in FOXP3
expression and acquisition of a regulatory phenotype.
After 4 h of stimulation, however, many activated cells
retained high levels of CD127, showing that a reduction
in surface CD127 is not associated with the early stages
of T cell activation per se.
3.4. Assay characteristics
canbeusedasa reliableandrobustdescriptorofT-regs in
human peripheral blood, a variety of assay characteristics
in healthy volunteers revealed that the fraction of CD25+
CD127lowT cells is stable over time in individual
subjects, although there is substantial variability between
subjects (Fig. 4A). The low variability of the assay over
these repeated measures (largest inter-assay CV, 1.2%)
may be an upper limit on the true inter-assay variability.
The assay also gave very similar results in samples of
whole blood, fresh PBMC, and frozen PBMC (Fig. 4B).
Finally, we tested the intra-assay variability by quintupli-
cate measurements on whole blood samples, which
demonstrated an intra-assay CVof 6.3% (Fig. 4C).
3.5. Lower levels of CD127 on regulatory T cells of
mice and macaques
Since many observations made in humans may be
further explored by experiments in murine and non-
human primate models, it was of interest to determine
whether CD127 is a useful marker for the viable dis-
crimination of T-regs in these species. In wild-type mice,
CD3+CD4+CD25+cells were uniformly CD127low
(Fig. 5A). Low expression of CD127 was also observed
in all Foxp3-expressing T cells, whether CD25-positive
or-negative. Although scurfy (sf−/−) mice lacked CD25+
CD127lowcells, they did have an expanded population of
CD25+CD127hicells, which have been demonstrated to
lack regulatory activity (Fontenot et al., 2003).
of CD3+CD4+cells in whole blood) were uniformly
CD127lowand FOXP3+(Fig. 5B). CD25 expression on
macaque CD4+T cells is unstable, being detectable on
only 3–5% of ficoll-purified PBMC and disappearing
completely after overnight storage. Scattered FOXP3+
to that observed in mice but rather different from that in
humans. We used bead sorting to test the suppressive
ability of CD25+CD127lowcells from rhesus macaque.
The cells are indeed suppressive (Fig. 5C), although the
suppression mediated by these bead-sorted rhesus
macaque cells tended to be less complete than that
mediated by FACS-sorted human CD25+CD127lowcells
(Figs. 5C, 2D).
Quantitative identification and isolation of viable T-
regs is particularly difficult in samples from human
patients. The best available surface marker, CD25, is
expressed not only on T-regs but also on activated cells
50 D.J. Hartigan-O'Connor et al. / Journal of Immunological Methods 319 (2007) 41–52
FOXP3, is necessary but not sufficient for regulatory
function (Allan et al., 2005), is expressed by many ac-
tivatedcells thatmay lacksuppressive function, isa weak
stain that fails to clearly separate positive from negative
cells, and can only be used to isolate cells that are
permeabilized and fixed.
We report here that human CD4+T-regs are
distinguished by a low level of surface CD127. The
CD4+CD25+/hiCD127lowpopulation comprises a clear-
ly distinct group of cells that can be gated consistently
over time, using fresh or frozen cells. T-regs identified
by this phenotype amount to 5–12% of CD3+CD4+
cells in normal subjects. This level of T-regs is similar to
that observed in mice and also to that required to exert
detectable suppression in vitro (about 10%). By
contrast, measurement of T-reg content using the
CD25hiphenotype yields estimates of 3–5% T-regs, a
level surprisingly lower than that observed in mice and
seemingly too low to exert detectable suppression in
vivo. Most of these CD25hicells express L-selectin,
suggesting that they might be present at a higher
frequency in lymphoid tissues, but we find that CD25hi
cells of macaques and humans are present at roughly
equal frequency in peripheral blood and lymph nodes
(data not shown and Michaelsson et al., 2006). Our
data suggest that T-regs are substantially more common
than CD4+CD25hicells and approach the frequencies
required for observation of suppression in vitro.
Cells found in the CD127lowpopulation are suppres-
sive regardless of their level of CD25 staining. CD25+
CD127lowcells are suppressive, hypoproliferative, and
hyporesponsive to TCR stimulation, while CD25+
CD127hicells (displaying the same intensity of CD25
staining) are not. It is perhaps surprising that CD25+
CD127lowcells are robustly suppressive despite lower
expression of FOXP3 (Figs. 2B, 3). However, CD25+
CD127lowcells rapidly up-regulate FOXP3 upon stim-
ulation, while CD25+CD127hicells do not (Fig. 3). It
seems likely that this activation step is required for
suppression by CD25+CD127lowcells. The time re-
quired for activation and up-regulation of FOXP3 might
account for the subtly less potent in vitro regulatory
activity of CD25+CD127lowcells as compared to CD25hi
T-regs of mice and macaques—including murine
CD25−Foxp3+cells that have been proven regulatory in
function—are also CD127low. Murine regulatory cells
have been identified as CD25+or CD25hi; however, it
was previously unclear whether all murine CD4+CD25+
cells are regulatory. In Fig. 5, we show that scurfy mice,
which lack regulatory cells, nonetheless harbor CD25+
CD127hicells. Therefore, CD127 can be used in mice as
CD25+cells. Healthy adult rhesus macaques have few
and sorting of cells with low-intensity CD25 expression
is difficult and resultsin isolation offew functional cells.
Fig. 4 shows that CD127 can distinguish cells that are
FOXP3+but bear little CD25 on their cell surface.
Our finding that the CD25+/hiCD127lowphenotype is
applicable in three species, and that down-regulation of
CD127 is associated not with T cell activation but rather
expression of the IL-7 receptor is an intrinsic quality of T-
be controlled independently by these two cytokines.
It has previously been reported that CD4+CD25+
cells of mice, most of which are regulatory cells, are
CD127low(Cozzo et al., 2003). We report here that
murine CD25−Foxp3+cells, which are knownto be true
T-regs with suppressive ability (Fontenot et al., 2005),
are also CD127low. This finding suggests that, in mice, a
lower level of CD127 expression characterizes cells
with the potential, after activation, to become active
CD25+CD127lowT-regs. It would be interesting to
know whether murine CD25−Foxp3−CD127lowcells
also have suppressive potential. Perhaps these cells can
up-regulate Foxp3 and become suppressive on activa-
tion. CD127 may prove useful in identifying similar
inactive T-regs in humans. Interestingly, healthy human
subjects have very few CD25−CD127lowcells, while
HIV-infected subjects often have a substantial number
of these cells (data not shown).
Dysfunctional or hyperactive T-regs may contribute
to the pathogenesis of autoimmune diseases, cancer, and
chronic infectious diseases. Surface markers are needed
to allow consistent identification and functional testing
of T-regs. Given the close association of T-regs with
activated Tcells, identification of T-regs in disease states
is particularly problematic. Use of the CD3+CD4+
CD127lowphenotype to discriminate T-regs may permit
better understanding of the role of regulatory T cells in
these pathologic situations.
We thank Jeffrey E. Mold and Jakob Michaelsson for
recommending CD127 as a possible marker for T-regs;
Kristina Abel and Christopher Miller of the Virology
and Immunology Section at the California National
Primate Research Center for provision of samples of
51 D.J. Hartigan-O'Connor et al. / Journal of Immunological Methods 319 (2007) 41–52
rhesus macaque peripheral blood and for the helpful Download full-text
discussions; and Brinda Emu, Jeff Critchfield, and
Steven G. Deeks for their helpful discussions.
This work was supported in part by NIH award R37
AI40312 to JMM, NIH award F32 AI067088 to DJH-
O'C, and an AIDS Research Institute at UCSF/
California AIDS Research Center grant (#CC02-SF-
002) funded by the University wide AIDS Research
Program to DJH-O'C. JMM is a recipient of the
Burroughs Wellcome Fund Clinical Scientist Award in
Translational Research and the NIH Director's Pioneer
Award Program, part of the NIH Roadmap for Medical
Research, through grant number DPI OD00329.
2004. Human CD4+ CD25+ regulatory T cells control T-cell
responses to human immunodeficiency virus and cytomegalovirus
antigens. J. Virol. 78, 2454.
Ziegler, S.F., Roncarolo, M.G., Levings, M.K., 2005. The role of 2
FOXP3 isoforms in the generation of human CD4+ Tregs. J. Clin.
Invest. 115, 3276.
Asano, M., Toda, M., Sakaguchi, N., Sakaguchi, S., 1996. Autoimmune
disease as a consequence of developmental abnormality of a T cell
subpopulation. J. Exp. Med. 184, 387.
CD25 high regulatory cells in human peripheral blood. J. Immunol.
identifies functionally distinct human regulatory Tcells. J. Immunol.
Bour-Jordan, H.l.n., Salomon, B.Æ.L., Thompson, H.L., Szot, G.L.,
Bernhard, M.R., Bluestone, J.A., 2004. Costimulation controls
diabetes by altering the balance of pathogenic and regulatory Tcells.
J. Clin. Invest. 114, 979.
Cozzo, C., Larkin III, J., Caton, A.J., 2003. Cutting edge: self-peptides
drive the peripheral expansion of CD4+CD25+ regulatory T cells.
J. Immunol. 171, 5678.
Cupedo, T., Nagasawa, M., Weijer, K., Blom, B., Spits, H., 2005.
Development and activation of regulatory T cells in the human
fetus. Eur. J. Immunol. 35, 383.
Darrasse-Jeze, G., Marodon, G., Salomon, B.L., Catala, M.,
Klatzmann, D., 2005. Ontogeny of CD4+CD25+ regulatory/
suppressor T cells in human fetuses. Blood 105, 4715.
Earle, K.E., Tang, Q., Zhou, X., Liu, W., Zhu, S., Bonyhadi, M.L.,
Bluestone, J.A., 2005. In vitro expanded human CD4+CD25+
regulatory T cells suppress effector T cell proliferation. Clin.
Immunol. 115, 3.
Eggena, M.P., Barugahare, B., Jones, N., Okello, M., Mutalya, S.,
Kityo, C., Mugyenyi, P., Cao, H., 2005. Depletion of regulatory
Tcells in HIV infection is associated with immune activation.
J. Immunol. 174, 4407.
Fontenot, J.D., Gavin, M.A., Rudensky, A.Y., 2003. Foxp3 programs
the development and function of CD4+CD25+ regulatory T cells.
Nat. Immunol. 4, 330.
Rudensky, A.Y., 2005. Regulatory Tcell lineage specification by the
forkhead transcription factor foxp3. Immunity 22, 329.
Kinter, A.L., Hennessey, M., Bell, A., Kern, S., Lin, Y., Daucher, M.,
Planta, M., McGlaughlin, M., Jackson, R., Ziegler, S.F., Fauci, A.S.,
favorable clinical markers of disease status. J. Exp. Med. 200, 331.
McHugh, R.S., Whitters, M.J., Piccirillo, C.A., Young, D.A., Shevach,
E.M., Collins, M., Byrne, M.C., 2002. CD4(+)CD25(+) immuno-
regulatory Tcells: Gene expression analysis reveals a functional role
for the glucocorticoid-induced TNF receptor. Immunity 16, 311.
of Tcell responses in the developing human fetus. J. Immunol. 176,
Mills, K.H.G., 2004. Regulatory T cells: friend or foe in immunity to
infection? Nat. Rev., Immunol. 4, 841.
Pasare, C., Medzhitov, R., 2003. Toll pathway-dependent blockade of
CD4+CD25+ T cell-mediated suppression by dendritic cells.
Science 299, 1033.
Roncador, G., Brown, P.J., Maestre, L., Hue, S., Martinez-Torrecua-
drada, J.L., Ling, K.L., Pratap, S., Toms, C., Fox, B.C., Cerundolo,
V., Powrie, F., Banham, A.H., 2005. Analysis of FOXP3 protein
expression in human CD4+CD25+ regulatory Tcells at the single-
cell level. Eur. J. Immunol. 35, 1681.
Rouse, B.T., Suvas, S., 2004. Regulatory cells and infectious agents:
detentes cordiale and contraire. J. Immunol. 173, 2211.
Ruprecht, C.R., Gattorno, M., Ferlito, F., Gregorio, A., Martini, A.,
Lanzavecchia, A., Sallusto, F., 2005. Coexpression of CD25 and
Med. 201, 1793.
Sakaguchi, S., Sakaguchi, N., Asano, M., Itoh, M., Toda, M., 1995.
of self-tolerance causes various autoimmune diseases. J. Immunol.
Stassen, M., Fondel, S., Bopp, T., Richter, C., Muller, C., Kubach, J.,
2004. Human CD25+ regulatory T cells: two subsets defined by the
integrins alpha 4 beta 7 or alpha 4 beta 1 confer distinct suppressive
properties upon CD4+ T helper cells. Eur. J. Immunol. 34, 1303.
E.L., McDevitt, H., Bonyhadi, M., Bluestone, J.A., 2004. In vitro-
expanded antigen-specific regulatory T cells suppress autoimmune
diabetes. J. Exp. Med. 199, 1455.
Thornton, A.M., Shevach, E.M., 1998. CD4+CD25+ immunoregula-
tory T cells suppress polyclonal T cell activation in vitro by
inhibiting interleukin 2 production. J. Exp. Med. 188, 287.
Walker, M.R., Kasprowicz, D.J., Gersuk, V.H., Benard, A., Van
Landeghen, M., Buckner, J.H., Ziegler, S.F., 2003. Induction of
FoxP3 and acquisition of T regulatory activity by stimulated
human CD4+CD25- T cells. J. Clin. Invest. 112, 1437.
Wysocki, C.A., Jiang, Q., Panoskaltsis-Mortari, A., Taylor, P.A.,
McKinnon, K.P., Su, L., Blazar, B.R., Serody, J.S., 2005.
Critical role for CCR5 in the function of donor CD4+CD25+
regulatory T cells during acute graft-versus-host disease. Blood
Yagi, H., Nomura, T., Nakamura, K., Yamazaki, S., Kitawaki, T., Hori,
S., Maeda, M., Onodera, M., Uchiyama, T., Fujii, S., Sakaguchi,
S., 2004. Crucial role of FOXP3 in the development and function
of human CD25+CD4+ regulatory Tcells. Int. Immunol. 16, 1643.
52D.J. Hartigan-O'Connor et al. / Journal of Immunological Methods 319 (2007) 41–52