Immunology - 2008 - Montcuquet - Regulatory T Cell Expansion and Function Do Not Account For The Impaired Alloreactivity of

IMMUNOLOGY ORIGINAL ARTICLE
Regulatory T-cell expansion and function do not account for the

impaired alloreactivity of ex vivo-expanded T cells
Nicolas Montcuquet,1,2,3 Patricia Summary

Mercier-Letondal,1,2,3 Sylvain
CD3- and CD28-activated T cells expanded for 12 days ex vivo to produce
Perruche,1,2,3 Anne Duperrier,1,2,3
suicide gene-modified T cells are hyporesponsive to alloantigens. To investi-
Mélanie Couturier,1,2,3 Abdelghani
gate whether this impaired alloreactivity is a result of preferential expansion
Bouchekioua,1,2,3 Mark Bonyhadi,4
of regulatory T (Treg) cells, we compared peripheral blood mononuclear
Christophe Ferrand,1,2,3 Pierre
cells (PBMC) activated with CD3 and CD28 antibodies co-immobilized on
Tiberghien1,2,3 and Eric
beads and expanded for 12 days with interleukin (IL)-2 (CoCD3/CD28 cells)
Robinet1,2,3*
1
to the respective unactivated PBMC in terms of proliferation, cytokine pro-
INSERM, Besançon, France, 2Université de
duction, and expression of Treg markers [cytotoxic T-lymphocyte antigen 4
Franche-Comte, Besançon, France, 3EFS,
Bourgogne Franche-Comté, Besançon, France, (CTLA4), glucocorticoid-induced tumour necrosis factor receptor (GITR)
and 4Invitrogen Corporation, Carlsbad, CA, and forkhead box P3 (FoxP3)] after allostimulation. Alloreactive cells were
USA identified by carboxyfluoresceine succinimidyl ester staining dilution. Allo-
reactive cells in CoCD3/CD28 cells had a lower proliferative response and a
lower potential for IL-2 and interferon-c secretion than did those in PBMC,
demonstrating a functional impairment of alloreactive cells during ex vivo
expansion. Expression of Treg markers transiently increased during ex vivo
expansion and was unaffected by depletion of CD25+ cells (containing Treg
cells) before ex vivo PBMC expansion. Such prior CD25+ depletion did not
doi:10.1111/j.1365-2567.2008.02843.x restore the alloreactivity of CoCD3/CD28 cells. After allostimulation, expres-
Received 30 October 2007; revised 20 sion of Treg markers was restricted to proliferative (alloreactive) cells
February 2008, 5 March 2008; accepted 6 among PBMC or CoCD3/CD28 cells. Lastly, CD4+ CD25+ cells purified from
March 2008.
CoCD3/CD28 cells lacked suppressive activity when used as a third party, in
*Present address: INSERM, U748, 67000
Strasbourg, France. contrast to CD4+ CD25+ cells purified from PBMC. In conclusion, the
Correspondence: E. Robinet, PhD, INSERM impaired alloreactivity of T cells expanded ex vivo is not a result of preferen-
U748, Institut de Virologie, 3 rue Koeberlé, tial Treg cell expansion and/or enhanced suppressive Treg activity.
67000 Strasbourg, France. Email:
eric.robinet@viro-ulp.u-strasbg.fr Keywords: adoptive cellular immunotherapy; alloreactivity; cellular ther-
Senior author: Eric Robinet apy; cultured cells; regulatory T cells
HSV-tk. Thus, GvHD can be resolved while preserving

Introduction
other immune cells, such as non-alloreactive GMC and
T cells in an allogeneic haematopoietic stem cell (HSC) immune cells that are not gene modified.
graft provide benefits such as antileukemic properties During retroviral-mediated gene transfer in T cells, the
[graft-versus-leukemia (GvL) effect], antiviral potential, transduction of target cells requires induction of T-cell
and facilitation of engraftment. However, these T cells proliferation. Depending on the clinical setting, gene
also introduce a major complication to HSC transplanta- transfer may target antigen-specific cells, using antigen-
tion, graft-versus-host disease (GvHD), which results specific stimulation,3 or polyclonal T cells, using mitogens
from the recognition of host antigens by donor allore- or CD3 antibody-mediated activation.4 We5–7 and
active T cells. Retroviral-mediated ex vivo transfer of the others8–14 have shown both in vitro and in vivo, using
herpes simplex virus thymidine kinase (HSV-tk) gene into murine and canine models of HSC transplantation, that
donor T cells enables control of alloreactivity after HSC T-cell expansion following stimulation with phytohaema-
transplantation,1,2 providing a means to specifically glutinin (PHA), CD3 or CD3/CD28 is associated with
deplete alloreactive gene-modified cells (GMC) expressing decreased alloreactivity. In particular, we reported that
320 2008 The Authors Journal compilation 2008 Blackwell Publishing Ltd, Immunology, 125, 320–330
13652567, 2008, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-2567.2008.02843.x by Nat Prov Indonesia, Wiley Online Library on [22/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Alloreactivity of ex vivo-expanded T cells
human peripheral blood mononuclear cells (PBMC)

Materials and methods
expanded ex vivo for 12 days with interleukin (IL)-2 after
soluble CD3 monoclonal antibody (mAb) stimulation
T-cell expansion
(CoCD3 cells) have decreased alloreactivity compared
with fresh PBMC, as determined by [3H]dT incorporation After obtaining informed consent, blood was harvested
during mixed lymphocyte reactions (MLRs),6,7 pre-T from normal volunteer platelet donors using cytapheresis
helper (Th) limiting dilution assay (LDA), pre-cytotoxic kit filters, and PBMC were isolated by centrifugation over
T-lymphocyte (CTL) LDA,6 or interferon-c (IFN-c) Histopaque-1077 (Sigma, Saint-Louis, MO) and resus-
enzyme-linked immunosorbent spot-forming cell assay pended in complete medium consisting of RPMI 1640 cul-
(ELISPOT).6 Assessments performed in murine GvHD ture medium (Cambrex BioSciences, Clermont-Ferrand,
models in vivo demonstrate that CD3/CD28 costimulation France) supplemented with 10% human serum (EFS
prevents alterations of the T-cell receptor Vb repertoire Bourgogne/Franche-Comté, Besançon, France). The cells
during ex vivo expansion7,15 as well as decreases in (106 cells/ml) were incubated at 37 with 5% CO2 in a
Epstein–Barr virus (EBV) reactivity16,17 and alloreactivi- humidified incubator, activated by CD3 and CD28
ty.11,13 Thus, CD3/CD28 costimulation is preferred to monoclonal antibody (mAb)-coated beads (Dynabe-
CD3 or mitogen activation for producing GMC.8,13,18 adsClinExVivo CD3/CD28, formerly XcyteDynabe-
However, cells expanded with interleukin (IL)-2 after ads; 1 bead/cell; Invitrogen, Cergy Pontoise, France) in
CD3/CD28 costimulation (CoCD3/CD28 cells) still have the presence of 500 IU/ml recombinant human IL-2 (Pro-
decreased alloreactivity when assessed in vitro in MLRs.7 leukin; Novartis Pharma Canada, Dorval, Québec), and
These defects are largely caused by the culture, as the divided every 3 days until day 12 of culture. We refer to
longer the culture duration14 and the greater the expan- cells expanded for 12 days after CD3/CD28 activation as
sion,8 the greater the impairment of alloreactivity. CoCD3/CD28 cells.
The mechanisms leading to impaired alloreactivity of
cells expanded ex vivo remain unclear; a better under-
Isolation of T cells
standing of such mechanisms may help in maintaining
the alloreactivity of GMC. Indeed, as the deleterious CD4+ T cells, CD8+ T cells, and total CD3+ T cells were
effects of donor GMC-mediated alloreactivity can be con- negatively selected according to the manufacturer’s
trolled with a pro-drug that is otherwise non-toxic to instructions (Miltenyi Biotec, Bergish Gladbach, Ger-
other cells, ganciclovir [an antiviral drug used to treat many) by incubating PBMC or CoCD3/CD28 cells with bio-
cytomegalovirus (CMV) infections], it is highly desirable tinylated mAbs directed against non-target cells. The cells
to maintain the highest possible GMC alloreactivity in were washed, antibiotin-conjugated magnetic microbeads
order to provide a powerful GvL effect. Alloreactivity may were added and, after incubation, the cells were washed
be reduced as a result of quantitative defects, such as loss again, and magnetic separation was performed on a
of alloreactive cells, during the ex vivo expansion or MLR, VarioMacs cell sorter (Miltenyi Biotec).
for example, by activation-induced cell death of allore-
active cells upon alloantigen recognition. Alternatively,
MLR
alloreactivity may be reduced as a result of qualitative
defects such as functional impairment resulting from Triplicates of 2 · 105 responder cells (PBMC or CoCD3/
induction of anergy or suppression. CD28 cells) were cultured in complete medium in the
In the present study, we demonstrate that the decreased absence or presence of irradiated (75 Gy) allogeneic
alloreactivity of CoCD3/CD28 cells is caused, at least in part, stimulating EBV-transformed B-cell lines (B-EBVallo) at a
by functional impairment of alloreactive cells. We also 2 : 1 cell ratio (responder:stimulator). Cultures were per-
exclude the possibility that it results from expansion of formed in round-bottomed 96-well plates in a final vol-
regulatory T (Treg) cells in the final product. We demon- ume of 200 ll. The proliferative response in the MLR was
strate an increased expression of Treg markers during evaluated by dilution of carboxyfluoresceine succinimidyl
ex vivo expansion and show that it does not result from ester (CFSE) staining as previously described.19 Cells with
expansion of Treg cells but instead from de novo expres- reduced CFSE content (CFSElow) were identified as allo-
sion of such markers by non-regulatory T cells. Further- reactive, while cells that did not lose CFSE staining
more, we demonstrate that CoCD3/CD28 cells lack intrinsic (CFSEhigh) were identified as non-alloreactive. Data are
suppressive activity and that depleting Treg cells before expressed as the percentage of CFSElow cells or as the
expansion does not restore alloreactivity to the Co cells. ratio of the median fluorescence intensity (MFI) of CFSE
Therefore, the reduced alloreactivity of CoCD3/CD28 cells is for CFSEhigh cells to the MFI of CFSE for CFSElow cells;
a result of intrinsic functional impairment of alloreactive the greater the number of cell divisions, the higher the
cells, such as exhaustion, rather than exogenous effects ratio. In some cases, the proliferative response in the
such as suppression by Treg cells. MLR was evaluated by determining bromodeoxyuridine
2008 The Authors Journal compilation 2008 Blackwell Publishing Ltd, Immunology, 125, 320–330 321
N. Montcuquet et al.
(BrdU) uptake during the last 8 hr of culture using a Del- were sorted with a FACS Aria (Becton Dickinson) to a
fia Proliferation kit (Perkin Elmer, Wellesley, MA), purity > 98%.
according to the manufacturer’s instructions.
Results
Enzyme-linked immunosorbent assay (ELISA) and
ELISPOT The differential proliferative responses of uncultured
and cultured responder T cells reflect functional
Responder cells were allostimulated as described above
impairment of alloreactive T cells after ex vivo
in flat-bottomed 96-well plates in a final volume of
expansion
200 ll. Human IFN-c or IL-2 secretion and the frequen-
cies of cells secreting human IFN-c or IL-2 during the We previously reported that the extent of impairment of
MLR were quantified respectively with ELISA and ELI- alloreactivity correlates with a reversal of the CD4/CD8
SPOT kits (Diaclone, Besançon, France), according to ratio during ex vivo expansion of PBMC activated by
the manufacturer’s instructions. The number of spots CD3 mAb or CD3/CD28 microbeads (Mercier-Letondal
per well was determined using a KS ELISPOT reader et al., in press). In order to compare homogeneous cell
(Zeiss, Le Pecq, France). The ratio of cytokine produc- suspensions in a preliminary experiment, CD4+ or CD8+
tion (measured by ELISA and expressed as pg/ml) to T cells were purified from PBMC and CoCD3/CD28 cells by
the number of cytokine-secreting cells (measured by negative immunomagnetic sorting (> 95% purity) and
ELISPOT and expressed as spot-forming cells/5 · 105 allostimulated. In agreement with our previous results
cells) allowed calculation of the amount of cytokine pro- using PBMC and bulk Co cells,6 the maximal prolifera-
duced per secreting cell: tion was higher in MLRs of CD4+ and CD8+ T cells puri-
Pa Pu fied from PBMC (BrdU incorporation) than in MLRs of
CD4+ and CD8+ T cells purified from CoCD3/CD28 cells
Fa Fu
(Fig. 1). Thus, in addition to the reversal of the CD4/
where Pa is the mean cytokine production in allostimulat- CD8 ratio accounting for the impairment of alloreactivity
ed wells, Pu is the mean cytokine production in control in Co cells, decreased alloreactivity can also be observed
unstimulated wells, Fa is the mean frequency in allostimu- in purified CD4+ and CD8+ responder cells. This suggests
lated wells and Fu is the mean frequency in control that intrinsic functional alterations of CD4+ and CD8+ T
unstimulated wells. The results are expressed as picograms cells might occur during ex vivo expansion.
of cytokine per cytokine-secreting cell. The level of BrdU incorporation by CoCD3/CD28 cells in
MLRs may be lower than that of PBMC as a result of
lower frequencies of alloreactive cells in the Co cells (i.e.
Flow cytometry
quantitative defects) and/or an impaired proliferative
For membrane staining, 05–1 · 106 cells were stained potential of alloreactive cells (i.e. qualitative defects). To
with CD3-phycoerythrin (PE)-Cy7 (Beckman Coulter
Immunotech, Marseille, France), CD4 or CD8-PE-Cy5 60
(Beckman Coulter), CD25-allophycocyanin (APC) (Becton CD4 CD8

50
Dickinson, San Diego, CA), or anti-glucocorticoid-
(BrdU incorporation x 10–3)
PBMC
Proliferative response
induced tumour necrosis factor receptor (GITR)-PE (R&D CoCD3/CD28

40
Systems, Abingdon, UK) mAbs, according to the manufac-
turer’s instructions. Adequate isotypic controls were used 30
as negative controls.
After membrane staining, cells were fixed and permeabi- 20
lized using a Fix and Perm Caltag kit (Invitrogen) accord-
ing to the manufacturer’s instructions. Cells were then 10
stained intracellularly with anti-IFN-c-APC, anti-IL-2-

0
PE (Becton Dickinson), or anti-CTLA4-PE (eBioscience, 1 2 3 4 5 6
San Diego, CA). Adequate isotypic controls were used in Time of MLR (day)
parallel as negative controls. Forkhead box P3 (FoxP3)
Figure 1. Impaired alloreactivity of CoCD3/CD28 cells with purified
staining was performed with a detection kit (eBioscience),
CD4+ or CD8+ responder T cells. CD3+ CD4+ T cells (triangles) and
according to the manufacturer’s instructions. Data were CD3+ CD8+ T cells (squares) were purified from peripheral blood
acquired using a Cyan ADP flow cytometer (Dako, Fort mononuclear cells (PBMC; closed symbols, solid lines) or CoCD3/CD28
Collins, CO). When indicated, CD4+ T cells isolated by cells (open symbols, dashed lines) and used as responders in the
negative selection were stained with APC-conjugated mixed lymphocyte reaction (MLR). The kinetics of proliferation were
CD25 (Becton Dickinson) antibody and CD4+ CD25+ cells evaluated by bromodeoxyuridine (BrdU) incorporation.
further investigate the proliferation kinetics after allo- CD4+ and CD8+ T cells from CFSElow CoCD3/CD28 cells at
stimulation, PBMC and CoCD3/CD28 cells were stained the end of the MLR (Fig. 2b), no difference in the
with CFSE and stimulated in MLR by allogeneic percentages of CFSElow CD4+ and CD8+ T cells was
B-EBVallo. Cells with reduced CFSE content (CFSElow) observed between PBMC and CoCD3/CD28 cells (Fig. 2c).
were identified as alloreactive, while cells that did not lose This confirms that, independent of the possible variations
CFSE staining (CFSEhigh) were identified as non-allore- in the frequencies of alloreactive precursors, alloreactive
active. When PBMC were used as responder cells, CD4+ cells underwent more cell divisions in PBMC than in
and CD8+ T cells started to divide on day 4 of the MLR. CoCD3/CD28 cells. Therefore, qualitative defects, at least,
Then, as a result of active cell division, the cells rapidly are induced in alloreactive cells during ex vivo expansion,
lost CFSE content between days 4 and 6, preventing eval- resulting in the impairment of their proliferative response
uation of the number of cell divisions at day 6 of the during the MLR.
MLR (Fig. 2a). In contrast, CD4+ and CD8+ T cells from
CoCD3/CD28 responder cells began dividing earlier, on day
The decreased alloreactivity of Co cells is associated
2 of the MLR, indicating the presence of more ‘primable’
with decreased potential for cytokine production
cells in the CoCD3/CD28 arm. However, only eight cell divi-
sions were quantifiable after 6 days of MLR (Fig. 2a). To further explore the functional impairment of allo-
While CD4+ and CD8+ T cells from CFSElow PBMC dem- reactive T cells after ex vivo expansion, PBMC and
onstrated a more pronounced reduction of CFSE content CoCD3/CD28 cells were stained with CFSE and then stimu-
(evaluated by determining the MFI of CFSElow cells, nor- lated by allogeneic B-EBVallo. Cells secreting IFN-c and
malized to the MFI of CFSEhigh cells) compared with IL-2 were identified by intracellular staining on day 3 of
(a) PBMC CD4 Co CD4 PBMC CD8 Co CD8

2·6% (5·9) 52·4% (2·1) 4·0% (4·7) 80·5% (2·7)
Day 3
13·9% (26·7) 82·0% (3·5) 10·7% (15·0) 87·4% (4·6)
Day 4
61·3% (105·6) 73·4% (7·8) 38·1% (68·5) 83·2% (6·8)
Day 5
79·8% (216·9) 81·7% (12·1) 68·5% (209·4) 91·2% (14·4)
Day 6
CFSE
(b) (c)
200 100
Proliferating (CFSE-) cells (%)
P = 0·0014 P = 0·0019
Level of CFSE dilution
80
150
60
100
40
50
20
0 0
PBMC Co PBMC Co PBMC Co PBMC Co
CD4 CD8 CD4 CD8
Figure 2. Alloreactive T cells in CoCD3/CD28 cells demonstrate slower proliferation during the mixed lymphocyte reaction (MLR) than do allo-
reactive T cells in peripheral blood mononuclear cells (PBMC). (a) PBMC or CoCD3/CD28 cells were stained with carboxyfluoresceine succinimidyl
ester (CFSE) and stimulated with allogeneic Epstein–Barr virus (EBV)-transformed B cells (B-EBVallo) during the MLR. CFSE staining is shown
from days 3–6 of the MLR for CD3+ CD8) (CD4+ T cells) and CD3+ CD8+ T cells. The percentage of CFSElow cells and the ratio of the median
fluorescence intensity (MFI) of CFSEhigh cells to the MFI of CFSElow cells are shown. Data are from one experiment representative of five experi-
ments. (b, c) MFI of CFSE staining (b) and the percentage of dividing cells (c) on day 6 of MLR. The MFI of CFSElow cells from 10 independent
experiments was normalized to the MFI of CFSEhigh cells, as indicated in the legend for Fig. 2a. Data are expressed as the median ± SE of 10 in-
dependent experiments. P-values are for a paired t-test (b and c).
the MLR (i.e. at the time of maximum cytokine detec- alloreactive (CFSElow) cells present in CoCD3/CD28 cells
tion; data not shown) and analysed in the CFSElow and have a strongly reduced potential for secretion of IFN-c
CFSEhigh populations. Most cytokine-producing cells were and IL-2 after ex vivo expansion as compared with allo-
present in the CFSElow population, indicating that cyto- reactive cells present in PBMC. However, reduced CFSE
kine production was restricted to the cells responding to staining characterized only some of the alloreactive cells
allogeneic stimulation (Fig. 3a). The percentages of on day 3 of the MLR, especially in PBMC. Therefore, we
IFN-c+ and IL-2+ cells among the CD4+ CFSElow cells analysed cytokine production in the secondary MLR.
were higher when PBMC were used as responder cells CFSE-stained PBMC or CoCD3/CD28 cells stimulated by
than when CoCD3/CD28 cells were used as responder allogeneic B-EBVallo for 5 days (i.e. until completion of
cells (Fig. 3a). Similar results were obtained for the CD8+ the primary MLR) were restimulated in the secondary
CFSElow cells, albeit with lower frequencies of cytokine- MLR by the same allogeneic B-EBVallo cell line for 8 hr.
positive cells (data not shown). IFN-c and IL-2 secretion was assessed in the CFSElow and
As gating on CFSElow cells excluded the non-allore- CFSEhigh CD4+ T-cell compartments and the CD8+ T-cell
active cells from analysis and only alloreactive cells were compartments. As reported during the primary MLR,
considered, the possible quantitative loss of alloreactive during the secondary MLR the frequencies of IFN-c+ or
cells during ex vivo expansion or initiation of the MLR IL-2+ cells among CFSElow CD4+ T cells (Fig. 3b) or
(for example through hypothetical induction of activa- CD8+ T cells (data not shown) were higher when PBMC
tion-induced cell death upon alloantigen recognition) was were used as responder cells than when CoCD3/CD28 cells
not taken into account. Therefore, our results suggest that were used as responder cells. This finding confirms that
(a) 50 (c)
IFN-γ+ IL-2– 60 CD4 CD8
Cytokine-positive cells (%)
(pg/IL-2-secreting cells)
40 IFN-γ+ IL-2+ 50 PBMC

IFN-γ– IL-2+
IL-2 production
CoCD3/CD28
30 40
30
20
20
10
10
0 0
PBMC Co PBMC Co 1 2 3
CFSElow CFSEhigh Time of MLR (day)
(b) 20 (d)
IFN-γ+ IL-2– 800 CD4 CD8
Cytokine-positive cells (%)
(pg/IFN-γ-secreting cells)
IFN-γ+ IL-2+ 700 PBMC

15
IFN-γ production
600 CoCD3/CD28
IFN-γ– IL-2+
500
10 400
300
5 200
100
0 0
PBMC Co PBMC Co 1 2 3
CFSElow CFSEhigh Time of MLR (day)
Figure 3. Impaired production of interferon (IFN)-c and interleukin (IL)-2 in alloreactive CoCD3/CD28 cells. Peripheral blood mononuclear cells
(PBMC) or CoCD3/CD28 cells were stained with carboxyfluoresceine succinimidyl ester (CFSE) and allostimulated for 3 days (a) or stimulated for
5 days and then restimulated with the same alloantigen for 8 hr (b) before labelling with CD3 and CD8 monoclonal antibodies (mAbs) and
intracellular anti-IFN-c and anti-IL-2 mAbs. The percentages of CD3+ CD8) cells (CD4+ T cells) positive for IFN-c+, IL-2+ or both are indicated
for the CFSEhigh and CFSElow populations. Data are the median ± standard error (SE) of three (a) and four (b) experiments. (c, d) Impaired
production of IL-2 and IFN-c by alloreactive CD4+ and CD8+ CoCD3/CD28 cells. CD4+ (triangles) or CD8+ (squares) T cells (105) were purified
from PBMC (closed symbols and solid lines) or CoCD3/CD28 cells (open symbols and dashed lines) by negative immunomagnetic sorting and
allostimulated in anti-IL-2 (c) or anti-IFN-c (d) mAb-coated [enzyme-linked immunosorbent spot-forming cell assay (ELISPOT)] or uncoated
[enzyme-linked immunosorbent assay (ELISA)] 96-well plates in a final volume of 200 ll. At the indicated times, cytokine production was
quantified by ELISA, and the frequencies of cytokine-secreting cells were evaluated by ELISPOT. The quantities of secreted cytokines (pg/ml)
were divided by the frequency of cytokine-secreting cells (expressed as spot-forming cells/5 · 105 cells) in order to evaluate cytokine production
per secreting cell (pg of cytokine/cytokine-secreting cells). Data are from one experiment representative of three experiments with similar results.
MLR, mixed lymphocyte reaction.
qualitative defects occur in alloreactive cells, with the Table 1. Depletion of CD25+ cells before ex vivo expansion by CD3/
culture decreasing the potential of these cells to produce CD28 does not prevent expression of cytotoxic T-lymphocyte antigen
cytokines. 4 (CTLA4), glucocorticoid-induced tumour necrosis factor receptor
(GITR) or forkhead box P3 (FoxP3)
To further characterize the impairment of cytokine
production by CoCD3/CD28 cells, IL-2 and IFN-c levels were CD25 depletion
quantified in the supernatant of the MLR by ELISA using Day of
CD4+ T cells and CD8+ T cells purified from either PBMC Marker culture ) +
or CoCD3/CD28 cells as responder cells. The frequencies of
CTLA4 0 12 (05–34) 02 (00–09)
cells secreting IL-2 and IFN-c were evaluated in parallel by
6 714 (337–830) 734 (374–871)
ELISPOT. Alloreactive cells were identified by their ability GITR 0 095 (02–18) 01 (00–01)
to secrete cytokines. The ratio of cytokine production 6 364 (290–537) 308 (28–519)
(measured by ELISA) to the frequency of cytokine-secret- FoxP3 0 25 (13–69) 07 (02–07)
ing cells (measured by ELISPOT) allowed determination 6 137 (49–645) 37 (26–697)
of the quantity of cytokine production per secreting cell,
i.e. the amount of cytokine produced per alloreactive Peripheral blood mononuclear cells (PBMC) underwent expansion
cell. Per cell, alloreactive CD4+ and CD8+ PBMC produced by CD3/CD28 and interleukin (IL)-2 with or without prior depletion
of CD25+ cells. The percentage of CD25+ cells in the total popula-
more IL-2 (Fig. 3c) and IFN-c (Fig. 3d), respectively,
tion of CD4+ T cells was 62% (18–127%) before depletion and
than did alloreactive CD4+ and CD8+ CoCD3/CD28
08% (01–41%) after depletion. Data are expressed as the median
cells, confirming the functional impairment of Co cells in (min–max) (n = 3–5) percentage, in the total population of CD4+ T
terms of cytokine production. cells, of CD4+ CD25+ T cells coexpressing CTLA4, GITR or FoxP3 at
the indicated time of culture.
Expression of regulatory T-cell markers increases
that CD25+ cell depletion did not prevent the expansion
during ex vivo expansion
of CD3+ CD4+ FoxP3+ cells: starting with 106 PBMC
One possible mechanism accounting for the qualitative seeded on day 0, the absolute CD4+ CD25+ FoxP3+ cell
defects of CoCD3/CD28 cells could be expansion of Treg cells number at day 0 was 20 (04–49) · 103 in the absence
during ex vivo expansion. We first analysed the expression versus 5 (04–6) · 103 [i.e. 56 (11–98)-fold lower] in
of Treg markers (CTLA4, GITR and FoxP3) and CD25 by the presence of CD25+ cell depletion (P = 0078 versus
CD3+ CD8) (considered to be CD3+ CD4+) and CD3+ no depletion), while at day 12 the absolute CD4+
CD8+ cells during ex vivo expansion using flow cytometry. CD25+ FoxP3+ cell number seeded was 335 (182–
After activation with CD3/CD28 and IL-2, the percent- 2111) · 103 in the absence versus 479 (127–1392) · 103
age of CD4+ CD25+ T cells coexpressing CTLA4, GITR or [i.e. only 12 (06–15)-fold lower] in the presence of
FoxP3 in the total population of CD4+ T cells increased CD25+ cell depletion (P = 0608 versus no depletion).
during the first 6 days of culture and then decreased pro- Because removal of CD25+ Treg cells at the start of
gressively until day 12 (supplementary data Fig. S1). Simi- expansion did not prevent the expansion of CD3+ CD4+
larly, after activation with CD3/CD28 and IL-2, CTLA4, FoxP3+ cells, we conclude that most CD3+ CD4+ FoxP3+
GITR and FoxP3 were expressed by CD8+ CD25+ T cells, cells observed during the ex vivo expansion were not
with similar kinetics as for CD4+ T cells, albeit at a lower derived from Treg cells, but instead resulted from the
level (supplementary data Fig. S1). These results show acquisition of FoxP3 expression during expansion. More-
that following activation with CD3/CD28 and IL-2 the over, CoCD3/CD28 cells expanded with and without deple-
percentage of cells with a Treg phenotype increased tion of CD25+ cells exhibited similar proliferative
during culture, peaking on days 3–6. responses in the MLR, demonstrating that depletion of
Next, we sought to determine whether the expression Treg cells before ex vivo expansion did not prevent the
of Treg markers during ex vivo culture resulted from reduction of alloreactivity in CoCD3/CD28 cells (Fig. 4).
expansion of Treg cells present at the start of the culture
or from acquisition of these markers by the cultured cells.
Treg cell markers are expressed by alloreactive cells
CD25-depleted or non-depleted PBMC were stimulated
during the MLR
with CD3/CD28 beads, and expression of Treg markers
was assessed at day 6, i.e. at the time of maximum We next analysed the expression of CTLA4, GITR and
expression. The expression of Treg markers was slightly, FoxP3 after allostimulation. CTLA4, GITR or FoxP3 was
but not significantly, reduced after CD25 depletion mainly expressed by alloreactive (CFSElow) cells (Fig. 5),
(Table 1), suggesting that the expression of CTLA4, GITR not by non-alloreactive (CFSEhigh) cells, where the
or FoxP3 resulted mainly from acquisition of these mark- percentage of CTLA4+, GITR+ or FoxP3+ cells was < 10%
ers by the cultured cells rather than from expansion of at all time-points (data not shown). Interestingly, the
Treg cells. This is further supported by the observation percentage of CD4+ CD25+ cells coexpressing CTLA4,
150 and FoxP3 were present in CD8+ T cells from PBMC or

Expt 1
Expt 2 Co cells, also with preferential expression in the CFSElow
gate and with similar kinetics as for the respective CD4+ T
Proliferative response 100 cells, but at lower levels than for CD4+ T cells (data not
shown). These data show that expression of CTLA4, GITR
or FoxP3 was not associated with an impaired proliferative
response in alloreactive cells, as expression of these
50
markers was higher in alloreactive (CFSElow) cells than in
non-alloreactive (CFSEhigh) cells. Moreover, the maximal
levels of expression of CTLA4, GITR and FoxP3 were
NT NT
0 higher in PBMC than in CoCD3/CD28 cells.
CD25 depletion: – + – + – + – +
PBMC Co PBMC Co
CD4 CD8 Expression of Treg cell markers does not correlate
Figure 4. Depletion of CD25+ cells before CD3/CD28-induced expan-
with the impaired proliferative response in the MLR
sion does not prevent the impaired proliferative response of Treg cells are unresponsive in the MLR.20 Thus, if the
CoCD3/CD28 cells in the mixed lymphocyte reaction (MLR). CoCD3/CD28 CD4+ CD25+ cells expressing Treg cell markers (CTLA4,
cells generated from CD25-depleted or non-depleted peripheral blood
GITR or FoxP3) during ex vivo expansion were derived
mononuclear cells (PBMC) were stained with carboxyfluoresceine
from Treg cells, one should expect that they would not
succinimidyl ester (CFSE) and used as responder cells in the MLR.
CD25-depleted or non-depleted PBMC were used as controls. Data are
proliferate upon allogeneic stimulation. To investigate this
from two independent experiments and are expressed as the normal- possibility, CFSE staining was used to assess the level of
ized median fluorescence intensity (MFI) of CFSElow CD4+ or CD8+ T proliferation in CD4+ T cells negative or positive for
cells on day 6 of the MLR, as indicated in Fig. 2. CD25-depleted PBMC CD25 and/or Treg cell markers at day 5 of the MLR.
were not tested (NT) in experiment 2 (black bars). When PBMC were used as the responder cells, the per-
centage of CFSElow cells was similar between
PBMC CoCD3/CD28 CD4+ CD25+ cells that did and did not express CTLA4,
100 CTLA4 + GITR or FoxP3 (Fig. 6). In contrast, the percentage of
CFSElow cells was much lower among CD4+ CD25) cells,
GITR+
regardless of expression of Treg cell markers (Fig. 6).
FoxP3+
75 Therefore, expression of Treg cell markers during the
Positive cells (%)
MLR was not associated per se with decreased prolifera-

tion; CD25 expression, but not CTLA4, GITR or FoxP3
50
expression, correlated with proliferation during the MLR.
Similar results were observed when CoCD3/CD28 cells were
25 used as the responder cells (data not shown) and when
analysis was gated on CD8+ cells instead of CD4+ cells
(data not shown).
0 Treg markers were weakly expressed at the time of
3 4 5 6
Time of MLR (day) MLR initiation when PBMC or CoCD3/CD28 cells were
used as responder cells (on days 0 and 12 of expansion,
Figure 5. Cytotoxic T-lymphocyte antigen 4 (CTLA4), glucocorticoid- respectively); expression was greatest between days 3 and
induced tumour necrosis factor receptor (GITR) and forkhead box P3 6 of expansion (supplementary data Fig. S1). We sorted
(FoxP3) are expressed by alloreactive cells during the mixed lympho-
CD4+ T cells at day 0, 5 or 12 of expansion with CD3/
cyte reaction (MLR). Peripheral blood mononuclear cells (PBMC;
CD28 and IL-2, stained them with CFSE, and stimulated
closed symbols and solid lines) or CoCD3/CD28 cells (open symbols and
dashed lines) were stained with carboxyfluoresceine succinimidyl ester
them with B-EBVallo cell lines. Although the percentage of
(CFSE) and allostimulated before staining at the indicated time of the CD4+ T cells coexpressing CTLA4, GITR or FoxP3 was
MLR with CD3, CD8 and CD25 monoclonal antibodies (mAbs) and higher in CoCD3/CD28 cells at day 5 than in CD4+ T cells
with CTLA4 (diamonds), GITR (squares) or FoxP3 (triangles) mAbs. at day 0 or 12 (see supplementary data Fig. S1), the pro-
Data are from one experiment representative of three experiments and liferative response in the MLR of CD4+ T cells sorted at
are expressed as the percentage of CD3+ CD8) CD25+ T cells coex- day 5 of expansion was not greatly diminished but rather
pressing CTLA4, GITR or FoxP3 gated in CFSElow CD3+ CD8) T cells. was intermediate between that of CD4+ T cells sorted at
days 0 and 12 (data not shown). Thus, the level of
GITR or FoxP3 was higher in CFSElow CD4+ cells when expression of Treg cell markers at the initiation of the
PBMC were used as responder cells than when CoCD3/CD28 MLR does not directly correlate with impaired prolifera-
cells were used as responder cells (Fig. 5). CTLA-4, GITR tion during the MLR. Similar results were obtained when
CD25– CD25+ 100

4
10
22·6 % 84·6 %
CTLA4+
Proliferative response (%)

3 75
10
CTLA4
2
10
50
2·6 % 79·1 %
CTLA4–
1
10
25
0
104
10
1·2 % 59·5 %
0
GITR+
10
3 Responder: 3rd party 1:1 1:0·5 1:0·125 1:1 1:0·5 1:0·125
Source of 3rd party PBMC CoCD3/CD28
GITR
2
10
3·8 % 92·1 % Figure 7. CD4+ CD25+ CoCD3/CD28 cells are not immunosuppressive.
GITR–
10
1
CD4+ CD25+ or CD4+ CD25) cells were sorted by flow cytometry
from peripheral blood mononuclear cells (PBMC) or CoCD3/CD28
0
104 cells on day 12 of expansion and used as irradiated third-party cells
10
58·2 % 87·5 % FoxP3+ in the mixed lymphocyte reaction (MLR) at the indicated ratio with
10
3 autologous CD4+ CD25) responder cells. Irradiated allogeneic
Epstein–Barr virus (EBV)-transformed B cells (B-EBVallo) were used
10
2 as stimulating cells. Data are expressed as the per cent proliferative
FoxP3
16·1 % 93·1 % response in the absence of third-party cells (mean ± standard error
FoxP3–
10
1 of three determinations).
0
10 0 1 2 3 4
10 10 10 10 10
CD25 B-EBVallo cells in the absence of third-party cells were used
CFSE
as the proliferative culture control. BrdU incorporation by
Figure 6. Expression of regulatory T (Treg) cell markers during the allostimulated responder cells in the presence of
mixed lymphocyte reaction (MLR) is not associated with a decreased CD4+ CD25+ CoCD3/CD28 third-party cells was 80% and
proliferative response. Peripheral blood mononuclear cells (PBMC) 100% of that for the control MLR at responder:third-party
were stained with carboxyfluoresceine succinimidyl ester (CFSE) and ratios of 1 : 1 and 1 : 0125, respectively (Fig. 7). This
allostimulated. On day 5 of the MLR, coexpression of CD25 and cyto- demonstrates that these cells were not or only slightly sup-
toxic T-lymphocyte antigen 4 (CTLA4), glucocorticoid-induced
pressive. In contrast, CD4+ CD25+ T cells sorted from fresh
tumour necrosis factor receptor (GITR) or forkhead box P3 (FoxP3)
PBMC (containing Treg cells) inhibited the MLR, the level
was assessed in CD3+ CD8) cells (left panels). The percentages of
CFSElow cells are indicated between parentheses for CD3+ CD8)
of BrdU incorporation being 50–70% of that for the con-
CD25) cells (middle panels) and CD3+ CD8) CD25+ cells (right trol MLR in the presence of fresh CD4+ CD25+ third-party
panels). The percentage of CFSElow cells was higher among CD25+ cells. Similar results were obtained when CD4+ CD25+ cells
responder cells than among CD25) responder cells, irrespective of the were immunomagnetically sorted from CoCD3/CD28 cells at
expression of CTLA4, GITR or FoxP3. Data are from one experiment day 5 of expansion (i.e. at a time where expression of Treg
representative of three experiments with similar results. cell markers is maximum) or when unsorted CoCD3/CD28
cells cultured for 12 days were used as a third party (data
not shown).
CD4+ CD25+ cells (instead of total CD4+ T cells) were
sorted at day 5 or 12 (data not shown).
Discussion
During our phase I/II clinical trial of the administration
Cells with a Treg phenotype are less suppressive in
of GMC,1 production of GMC required a 12-day ex vivo
expanded cells than in PBMC
culture period, including T-cell activation with a soluble
We next investigated whether the decreased alloreactivity of CD3 mAb in the presence of IL-2, retrovirus-mediated
CoCD3/CD28 cells results from the presence of suppressive gene transfer from day 3 to day 4, and selection of the
cells. We analysed the suppressive activity of CoCD3/CD28 GMC from day 5 to day 12.4 We previously showed that
cells using CD4+ CD25+ T cells sorted by flow cytometry CoCD3 cells, cultured according to the same procedure as
from CoCD3/CD28 cells at day 12 of expansion as a third GMC, exhibit reduced alloreactivity compared with fresh
party in a 6-day MLR using autologous CD4+ CD25) PBMC, as determined based on 3HdT incorporation dur-
PBMC as the responder cells and B-EBVallo cells as the ing the MLR,6,7 pre-Th LDA, pre-CTL LDA, or IFN-c
stimulating cells. CD4+ CD25) PBMC stimulated by ELISPOT.6 CoCD3/CD28 cells also have reduced alloreactivity
when assessed in MLR in vitro.7 However, the mecha- Co cells expanded for 12 days. The subsequent prolifera-
nisms accounting for such a decrease in alloreactivity tive response in the MLR was highest in PBMC, inter-
have been little explored. It is possible that they may mediate in Co cells expanded for 6 days (data not
result from quantitative defects, such as loss of allo- shown), and lowest in Co cells expanded for 12 days. The
reactive cells during the ex vivo expansion or MLR as a discordance between the pattern of expression of Treg cell
result of, for example, the induction of activation-induced markers by responder cells at the start of the MLR and
cell death of alloreactive cells upon alloantigen recogni- the pattern of the subsequent proliferative response does
tion. We have shown that cultured cells have a decreased not support the hypothesis that expression of Treg cell
frequency of alloreactive cells in LDA,6 but this is not markers is a prerequisite for suppressive activity and
demonstrative of a quantitative defect. Indeed, alloreactive inhibits proliferation during the MLR.
cells can be detected only based on their functional Thirdly, depletion of CD25+ cells before CD3/CD28
reactivity (proliferative response, cytotoxic activity, cyto- costimulation did not prevent the expansion of cells
kine secretion, expression of membrane activation mark- expressing Treg markers, especially CD4+ FoxP3+ T cells.
ers, etc). In the absence of a phenotypic method that Therefore, most cells expressing Treg cell markers are
could enable detection of alloreactive cells without alloge- effector cells that transitionally express Treg cell markers,
neic stimulation, such as human leucocyte antigen (HLA) rather than Treg cells expanded upon CD3/CD28 costi-
tetramer staining, quantitative loss of alloreactive mulation.
cells remains difficult to demonstrate. Fourthly, depletion of CD25+ cells before CD3/CD28
An alternative, but not exclusive, hypothesis to explain costimulation did not improve the alloreactivity of
the reduced alloreactivity is the occurrence of qualitative expanded cells. CoCD3/CD28 cells generated after CD25+
defects during cell expansion. By analysing alloreactive cell depletion exhibited impairment of alloreactivity simi-
cells at the single-cell level, we have provided evidence of lar to that of normal CoCD3/CD28 cells.
such qualitative defects, demonstrating that alloreactive Fifthly, a low suppressive activity was detected when
CoCD3/CD28 cells, identified by their reduced CFSE stain- CD4+ CD25+ cells purified at day 6 (i.e. the time of max-
ing during the MLR, underwent fewer cell divisions and imal expression of Treg cell markers) or day 12 of ex vivo
secreted less IL-2 and IFN-c during the MLR than did al- expansion were used as a suppressive third party in the
loreactive T cells present in uncultured cells (PBMC). MLR, as well as when using non-separated Co cells at day
Several hypotheses may account for the impaired respon- 12 of expansion (data not shown). These latter observa-
siveness, including T-cell exhaustion, anergy or suppres- tions are at odds with the findings of Mesel-Lemoine
sion. During expansion, we observed expression of et al.14 They reported that depletion of CD25+ T cells
CTLA4, GITR and FoxP3 Treg cell markers in CD4+ T before induction of expansion with CD3 and IL-2 or
cells and to a lesser extent in CD8+ T cells. Such expres- CD3/CD28 and IL-2 preserved the allogeneic response of
sion reached a maximal level between days 3 and 6 of cultured cells and that dividing cells purified at day 6 of
culture, and then decreased until day 12 of culture, in ex vivo expansion were able to inhibit the MLR when
accordance with previous reports.21 CD8+ T-cell subsets used as a non-irradiated third party, suggesting that
have also been shown to express Treg cell markers such stimulation with CD3 and IL-2 or CD3/CD28 and IL-2
as FoxP3.22–24 Therefore, we tested the hypothesis that preferentially expanded Treg cells. However, in this previ-
expansion of Treg cells accounts for the functional ous study, expanded cells used as a third party were not
impairment of the alloreactivity of Co cells. Our study irradiated, raising the possibility that they inhibited the
provides several results invalidating this hypothesis. alloproliferative response by IL-2 consumption.
First, when PBMC or Co cells were used as responder In this study, as in a previous study,21 depletion of
cells in the MLR, Treg cell markers were preferentially CD25+ T cells before expansion led to a lower percentage
expressed by cells proliferating in response to alloantigen of FoxP3+ cells during expansion compared with non-
(CFSElow), rather than by cells that were non-alloreactive depleted control cultures. The authors interpreted these
(non-proliferating, CFSEhigh). Expression of Treg cell results as an expansion of pre-existing FoxP3+ Treg
markers by alloreactive cells is inconsistent with the cells.14 However, in these two studies, the decrease in the
hypothesis that these markers are expressed by Treg cells, percentage of FoxP3+ cells in CD25-depleted cultures was
which are hyporesponsive in the MLR.20 Such transient less dramatic on days 3–6 of expansion than on day 0 of
expression of Treg cell markers did not confer any hypo- expansion, as compared with non-depleted controls. This
responsiveness to the alloreactive cells that acquired them. indicates that most of the FoxP3 expression at days 3–6
Secondly, the level of expression of the Treg cell mark- of expansion results from induction of FoxP3 expression
ers at the start of the MLR did not correlate with the sub- rather than from expansion of pre-existing FoxP3+ cells.
sequent proliferative response in the MLR. Indeed, these Indeed, if FoxP3+ cells observed during ex vivo culture
levels were lowest in T cells from PBMC, highest in Co result only from the expansion of pre-existing FoxP3+
cells expanded for 6 days, and intermediate, but low, in cells, the extent of the decrease in CD25-depleted cells
should be the same at days 3–6 of culture and at day 0, 2 Bonini C, Ferrari G, Verzeletti S et al. HSV-TK gene transfer
which is not the case (Table 1). into donor lymphocytes for control of allogeneic graft-versus-
Additional findings also contraindicate the preferential leukemia. Science 1997; 276:1719–24.
3 Koehne G, Gallardo HF, Sadelain M, O’Reilly RJ. Rapid selec-
expansion of Treg cells in our conditions. First, it is known
tion of antigen-specific T lymphocytes by retroviral transduction.
that ex vivo T-cell expansion after activation with CD3 and
Blood 2000; 96:109–17.
IL-2 or CD3/CD28 and IL-2 leads to impaired in vitro
4 Robinet E, Certoux JM, Ferrand C et al. A closed culture system
alloreactivity and is associated with a reversed CD4/CD8 for the ex vivo transduction and expansion of human T lym-
ratio, resulting from abundant CD8+ T-cell expan- phocytes. J Hematother 1998; 7:205–15.
sion.6,25,26 We previously reported that the conditions of 5 Contassot E, Murphy W, Angonin R et al. In vivo alloreactive
culture allowing preservation of alloreactivity (i.e. use of potential of ex vivo-expanded primary T lymphocytes. Trans-
lower IL-2 concentrations or replacement of IL-2 with plantation 1998; 65:1365–70.
IL-7) are associated with preservation of the CD4/CD8 6 Sauce D, Tonnelier N, Duperrier A et al. Influence of ex vivo
ratio,25 which is incompatible with preferential expansion expansion and retrovirus-mediated gene transfer on primary T
of CD4+ CD25+ Treg cells. Secondly, ex vivo expansion of lymphocyte phenotype and functions. J Hematother Stem Cell
Treg cells after activation with CD3/CD28 and IL-2 has Res 2002; 11:929–40.
7 Coito S, Sauce D, Duperrier A et al. Retrovirus-mediated gene
been reported, but requires a preliminary purification
transfer in human primary T lymphocytes induces an activation-
step.20,27–29 Indeed, Treg cells proliferate at a slower rate
and transduction/selection-dependent TCR-B variable chain rep-
than do normal T cells; using PBMC as a source of cells, as ertoire skewing of gene-modified cells. Stem Cells Dev 2004;
we did, leads to overgrowth of normal T cells, which are 13:71–81.
rapidly dividing, relative to Treg cells.20,27,29 8 Duarte RF, Chen FE, Lowdell MW, Potter MN, Lamana ML,
Overall, our data are in agreement with those of Gavin Prentice HG, Madrigal JA. Functional impairment of human
et al.21 and Allan et al.30 who reported that FoxP3 expres- T-lymphocytes following PHA-induced expansion and retroviral
sion can be induced in non-Treg cells upon activation with transduction: implications for gene therapy. Gene Ther 2002;
CD3/CD28 and IL-2, but, because such FoxP3 expression is 9:1359–68.
transient, it may not be sufficiently sustained and/or reach 9 Drobyski WR, Majewski D, Ozker K, Hanson G. Ex vivo anti-
CD3 antibody-activated donor T cells have a reduced ability to
a sufficient intensity to induce suppressive properties in
cause lethal murine graft-versus-host disease but retain their
FoxP3+ cells generated during expansion. Indeed, our
ability to facilitate alloengraftment. J Immunol 1998; 161:2610–9.
results show that despite the expression of Treg cell markers
10 Lamana ML, Bueren JA, Vicario JL, Balas A. Functional and
such as CTLA4, GITR or FoxP3, in association with CD25 phenotypic variations in human T cells subjected to retroviral-
expression by most of the cultured cells, these cells are not mediated gene transfer. Gene Ther 2004; 11:474–82.
suppressive or are only poorly suppressive. Therefore, our 11 van Rijn RS, Simonetti ER, Hagenbeek A, Bonyhadi M, Storm
data do not support the hypothesis that the impaired func- G, Martens AC, Ebeling SB. Quantitative assessment of human T
tionality of cells expanded ex vivo results from preferential lymphocytes in RAG2(-/-)gammac(-/-) mice: the impact of ex vivo
expansion of Treg cells. As IL-2 is present during the cul- manipulation on in vivo functionality. Exp Hematol 2007;
ture period, it is also likely that induction of IL-2-sensitive 35:117–27.
anergy is not involved in reduction of alloreactivity. Alter- 12 Weijtens M, van Spronsen A, Hagenbeek A, Braakman E, Mar-
tens A. Reduced graft-versus-host disease-inducing capacity of T
natively, CoCD3/CD28 cells, especially alloreactive cells, may
cells after activation, culturing, and magnetic cell sorting selec-
simply be exhausted by ex vivo expansion, and alloreactive
tion in an allogeneic bone marrow transplantation model in rats.
cells may be intrinsically hyporesponsive to alloantigens,
Hum Gene Ther 2002; 13:187–98.
independent of any suppressive activity. 13 Bondanza A, Valtolina V, Magnani Z et al. Suicide gene therapy
of graft-versus-host disease induced by central memory human
Acknowledgements T lymphocytes. Blood 2006; 107:1828–36.
14 Mesel-Lemoine M, Cherai M, Le Gouvello S, Guillot M, Leclercq
This work was supported by the Ligue contre le Cancer, V, Klatzmann D, Thomas-Vaslin V, Lemoine FM. Initial deple-
comité du Doubs and by the Scientific Board of the Eta- tion of regulatory T cells: the missing solution to preserve the
blissement Français du Sang. NM and PML have received immune functions of T lymphocytes designed for cell therapy.
fellowships from the Ligue Contre le Cancer, Comité du Blood 2006; 107:381–8.
Doubs. NM is currently a recipient of a fellowship from 15 Ferrand C, Robinet E, Contassot E, Certoux JM, Lim A, Herve
P, Tiberghien P. Retrovirus-mediated gene transfer in primary
the Etablissement Français du Sang.
T lymphocytes: influence of the transduction/selection process
and of ex vivo expansion on the T cell receptor beta chain hy-
References pervariable region repertoire. Hum Gene Ther 2000; 11:1151–
64.
1 Tiberghien P, Ferrand C, Lioure B et al. Administration of 16 Sauce D, Bodinier M, Garin M et al. Retrovirus-mediated gene
herpes simplex-thymidine kinase-expressing donor T cells with a transfer in primary T lymphocytes impairs their anti-Epstein-
T-cell-depleted allogeneic marrow graft. Blood 2001; 97:63–72.
Barr virus potential through both culture-dependent and selec- 27 Hoffmann P, Eder R, Kunz-Schughart LA, Andreesen R, Edinger
tion process-dependent mechanisms. Blood 2002; 99:1165–73. M. Large-scale in vitro expansion of polyclonal human
17 Sauce D, Rufer N, Mercier P et al. Retrovirus-mediated gene CD4(+)CD25high regulatory T cells. Blood 2004; 104:895–903.
transfer in polyclonal T cells results in lower apoptosis and 28 Godfrey WR, Ge YG, Spoden DJ, Levine BL, June CH, Blazar
enhanced ex vivo cell expansion of CMV-reactive CD8 T cells as BR, Porter SB. In vitro-expanded human CD4(+)CD25(+)
compared with EBV-reactive CD8 T cells. Blood 2003; 102:1241– T-regulatory cells can markedly inhibit allogeneic dendritic
8. cell-stimulated MLR cultures. Blood 2004; 104:453–61.
18 Pollok KE, van der Loo JC, Cooper RJ, Kennedy L, Williams 29 Karakhanova S, Munder M, Schneider M, Bonyhadi M, Ho AD,
DA. Costimulation of transduced T lymphocytes via T cell Goerner M. Highly efficient expansion of human CD4+CD25+
receptor-CD3 complex and CD28 leads to increased transcrip- regulatory T cells for cellular immunotherapy in patients with
tion of integrated retrovirus. Hum Gene Ther 1999; 10:2221–36. graft-versus-host disease. J Immunother 2006; 29:336–49.
19 Wells AD, Gudmundsdottir H, Turka LA. Following the fate of 30 Allan SE, Crome SQ, Crellin NK, Passerini L, Steiner TS, Bacch-
individual T cells throughout activation and clonal expansion. etta R, Roncarolo MG, Levings MK. Activation-induced FOXP3
Signals from T cell receptor and CD28 differentially regulate the in human T effector cells does not suppress proliferation or
induction and duration of a proliferative response. J Clin Invest cytokine production. Int Immunol 2007; 19:345–54.
1997; 100:3173–83.
20 Earle KE, Tang Q, Zhou X, Liu W, Zhu S, Bonyhadi ML, Blue-
stone JA. In vitro expanded human CD4+CD25+ regulatory T Supplementary material
cells suppress effector T cell proliferation. Clin Immunol 2005; The following supplementary material is available for this
115:3–9.
article online:
21 Gavin MA, Torgerson TR, Houston E et al. Single-cell analysis
Figure S1. Transient expression of cytotoxic T-lympho-
of normal and FOXP3-mutant human T cells: FOXP3 expression
without regulatory T cell development. Proc Natl Acad Sci USA
cyte antigen 4 (CTLA4), glucocorticoid-induced tumour
2006; 103:6659–64. necrosis factor receptor (GITR) and forkhead box P3
22 Bisikirska B, Colgan J, Luban J, Bluestone JA, Herold KC. TCR (FoxP3) during CD3/CD28-induced expansion of CD4+
stimulation with modified anti-CD3 mAb expands CD8+ T cell and CD8+ T cells after CD3/CD28 costimulation. Data
population and induces CD8+CD25+ Tregs. J Clin Invest 2005; are expressed as the percentage of CD3+ CD8) CD25+ T
115:2904–13. cells (CD4+ T cells; closed symbols and solid lines) in the
23 Ahmadzadeh M, Antony PA, Rosenberg SA. IL-2 and IL-15 each total population of CD3+ CD8) T cells coexpressing
mediate de novo induction of FOXP3 expression in human CTLA4 (diamonds), GITR (squares), or FoxP3 (triangles)
tumor antigen-specific CD8 T cells. J Immunother 2007; 30:294– or the percentage of CD3+ CD8+ CD25+ T cells (open
302.
symbols and dashed lines) in the total population of
24 Jarvis LB, Matyszak MK, Duggleby RC, Goodall JC, Hall FC,
CD3+ CD8+ T cells coexpressing CTLA4, GITR or FoxP3.
Gaston JS. Autoreactive human peripheral blood CD8+ T cells
with a regulatory phenotype and function. Eur J Immunol 2005;
Each time-point is the median of two to six experiments.
35:2896–908. This material is available as part of the online article
25 Mercier-Letondal P, Montcuquet N, Sauce D et al. Alloreactivity from http://www.blackwell-synergy.com.
of ex vivo-cultured T cells is correlated with expansion and Please note: Blackwell Publishing are not responsible
CD4/CD8 ratio. Cytotherapy 2008; 10:275–88. for the content or functionality of any supplementary
26 Marini FC 3rd, Kornblau SM. Production and culture of HSVtk materials supplied by the authors. Any queries (other
transduced suicidal lymphocytes induces variable changes in the than missing material) should be directed to the corre-
lymphocyte subset composition. Bone Marrow Transplant 1999; sponding author for the article.
23:907–15.

Immunology - 2008 - Montcuquet - Regulatory T Cell Expansion and Function Do Not Account For The Impaired Alloreactivity of

Uploaded by

Copyright:

Available Formats

Immunology - 2008 - Montcuquet - Regulatory T Cell Expansion and Function Do Not Account For The Impaired Alloreactivity of

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Immunology - 2008 - Montcuquet - Regulatory T Cell Expansion and Function Do Not Account For The Impaired Alloreactivity of

Uploaded by

Copyright:

Available Formats

IMMUNOLOGY ORIGINAL ARTICLE

Regulatory T-cell expansion and function do not account for the

Nicolas Montcuquet,1,2,3 Patricia Summary

HSV-tk. Thus, GvHD can be resolved while preserving

human peripheral blood mononuclear cells (PBMC)

(Beckman Coulter), CD25-allophycocyanin (APC) (Becton CD4 CD8

induced tumour necrosis factor receptor (GITR)-PE (R&D CoCD3/CD28

stained intracellularly with anti-IFN-c-APC, anti-IL-2-

(a) PBMC CD4 Co CD4 PBMC CD8 Co CD8

13·9% (26·7) 82·0% (3·5) 10·7% (15·0) 87·4% (4·6)

61·3% (105·6) 73·4% (7·8) 38·1% (68·5) 83·2% (6·8)

79·8% (216·9) 81·7% (12·1) 68·5% (209·4) 91·2% (14·4)

40 IFN-γ+ IL-2+ 50 PBMC

IFN-γ+ IL-2+ 700 PBMC

150 and FoxP3 were present in CD8+ T cells from PBMC or

MLR was not associated per se with decreased prolifera-

CD25– CD25+ 100

Proliferative response (%)

You might also like