The Role of TCR Specificity and Clonal Competition During
Reconstruction of the Peripheral T Cell Pool1
Catarina Leita ˜o, Anto ´nio A. Freitas, and Sylvie Garcia2
Survival of peripheral CD8?T cells requires TCR interactions with peptide-MHC complexes (p-MHC). In the adult mouse, in the
presence of homeostatic mechanisms that strictly control T cell numbers, it is likely that diverse T cell clones may compete for
shared patterns of p-MHC. In the present study, we investigate whether the recognition of p-MHC overlaps between different T
cell populations and what role does this process plays in the establishment of the peripheral T cell pools. Using an experimental
strategy that follows the fate of adoptively transferred polyclonal T cells into RAG0/0or different TCR transgenic RAG0/0hosts,
we demonstrate that T cells bearing different TCR specificities share identical TCR-specific requirements for survival and lym-
phopenia driven proliferation (LDP). This interclonal competition applies to both naive and activated/memory T cells and is
partially determined by the clone size of the established/resident T cells. However, clonal competition with activated/memory
resident T cells impacts differently on the fate of newly produced bone-marrow-derived T cells or adoptively transferred periph-
eral T cells. Overall, our findings indicate that p-MHC define multiple diverse resource niches that can be shared by T cells from
different compartments. The Journal of Immunology, 2009, 182: 5232–5239.
capable to face any new microbial invasion. In this context, it is
important to understand how the different processes involved in
peripheral T cell pool reconstruction may interplay with the final
composition of the T cell repertoire.
In the adult mouse, the number of peripheral T cells is main-
tained within narrow ranges and is largely independent of the num-
bers of precursor and thymus cells (1). The T cell repertoire size is
supposed to be mainly controlled by the number of T cell clones
contained in the naive pool (2). Using competitive bone marrow
(BM)3reconstitution strategies to study the establishment of the
peripheral CD8?T cell compartments, it was shown that, in the
presence of continuous new cell production, each lymphocyte had
to compete with other newly produced or resident cells for survival
(3). Several different lines of evidence suggest the important role
of TCR/self-peptide-MHC complexes (sp-MHC) interactions in
the process of T cell competition. First, TCR/sp-MHC interactions
are required for peripheral T cell survival, as demonstrated by the
findings that the absence of either the TCR (4, 5) or specific MHC-
major criterion of successful immune reconstitution fol-
lowing infection (HIV) or therapeutic-induced lym-
phopenia is the recovery of a diverse T cell repertoire,
molecules (6–10) leads to the disappearance of the peripheral
MHC-restricted mature T cells. Secondly, in monoclonal RAG-
deficient TCR transgenic (Tg) mice, the number of T cells differs
according to the TCR expressed (11), suggesting that TCR spec-
ificity for sp-MHC plays an important role to define the final clone
size (11–13). Finally, TCR recognition of sp-MHC has been
shown to be required for the expansion of mature T cells during the
peripheral T cell repopulation of immune-deficient hosts (1, 14–
16). The lower naive T cell proliferation in germfree T cell-defi-
cient hosts (17) suggests that intestinal flora drives T cell expan-
sion either by presentation of p-MHC from commensal bacteria or
by inducing the hosts APCs to increase the levels of sp-MHC
expression and the release of homeostatic cytokines.
Because of the involvement of p-MHC complexes in many ho-
meostatic T cell processes, competition among T cells in limiting
conditions for defined TCR specific ligands are likely to occur. Dif-
ferent studies have shown that a defined TCR Tg T cell population
transferred into hosts containing T cells bearing the same TCR Tg
could not proliferate, despite the relative host lymphopenia (18–21).
These observations were consistent with the existence of competition
between resident naive T cells and transferred T cells. Nevertheless,
because most of these experiments involved T cells bearing identical
TCRs, they did not allow any conclusion whether different T cell
clones could compete for identical p-MHC. Lymphopenia driven pro-
liferation (LDP) of transferred polyclonal CD4?T cells has been
shown to be inversely correlated to the repertoire diversity of resident
LDP-derived T cells (22). However, this study only addressed the role
of the number and diversity of the resident T cell clones, and provided
no information concerning the potential role of the TCR specificity of
interclonal competition extends largely beyond TCR specificity (21).
Moreover, the occurrence of clonal competition among T cells from
different compartments, which, despite sharing some cytokine re-
sources (23), are supposed to be independently regulated (24), have
never been addressed.
We have now studied the role of TCR specificity in T cell com-
petition in different homeostatic situations. Using different TCR
transgenic RAG-deficient hosts, we show that diverse polyclonal
CD8?T cell clones compete for identical p-MHC and that the
Lymphocyte Population Biology Unit, Centre National de la Recherche Scientifique,
Institut Pasteur, Paris, France
Received for publication December 4, 2008. Accepted for publication February
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1C.L. is supported by a grant from the Portuguese Foundation for Science and Technol-
ogy (FCT) and by a Pasteur-Weizmann grant. This work was supported by the Pasteur
Institute, the Centre National de la Recherche Scientifique, and grants from Agence Na-
tionale de Recherches sur le SIDA and Association pour la Recherche sur le Cancer.
2Address correspondence and reprint requests to Dr. Sylvie Garcia, Lymphocyte
Population Biology Unit, Institut Pasteur, 25 rue du Dr. Roux, Paris, France. E-mail
3Abbreviations used in this paper: BM, bone marrow; sp-MHC, self-peptide-MHC
complex; Tg, transgenic; LDP, lymphopenia driven proliferation; LCMV, lympho-
cytic choriomeningitis virus; LN, lymph node.
Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00
The Journal of Immunology
clone size of resident T cells influences the expansion and accu-
mulation of the transferred T cells in a TCR-dependent manner.
We found that this interclonal competition applied to both naive
and activated/memory T cells. In addition, we show that resident
LDP-derived activated T cells can interfere with the peripheral
establishment of newly introduced populations of monoclonal T
cells. However, while newly produced BM-derived T cells were
out-competed according to their TCR specificity, the accumulation
of a second population of transferred mature T cells occurred in a
nonspecific manner. These results indicate that p-MHC may define
TCR-specific niches shared by T cells belonging to different com-
partments and controlled by different homeostatic processes.
Materials and Methods
C57BL/6(B6).Ly5.1, B6.PL.Thy1.1, and B6.RAG20/0mice were from
the Centre de Distribution, Typage and Archivage animal (CDTA, Or-
leans, France). The TCR Tg strains on a B6 background used were: the
aHY RAG20/0mice Tg for a H-2Db-restricted TCR specific for the
HY-male Ag; the P14 RAG20/0mice Tg for a H-2Db-restricted TCR
specific for the gp33–41 epitope of the lymphocytic choriomeningitis
virus glycoprotein (LCMV); the OT-1 RAG10/0mice Tg for a H-2Kb-
restricted TCR specific for the 257–264 peptide of OVA. All Tg mice
were crossed to be Ly5.1?or Ly5.2?and were maintained in our spe-
cific pathogen free animal facilities according to the Experimental Eth-
ics Committee guidelines.
T cells were transferred into Thy1.2?RAG0/0hosts transgenic or not for the different class I-restricted TCRs: aHY, P14 and OT-1. Mice were
sacrificed 7 wk after adoptive transfer. A, CFSE profile of CD3?CD8?Thy1.1?is shown in each case. Each FACS profile is representative of one
of five mice for each type of hosts. B, Numbers of donor CFSE?(left) and CFSEhigh(right) CD8?Thy1.1?LN T cells recovered in RAG20/0(white
bars), aHY (pale gray bars), P14 (dark gray bars), and OT-1 (black bars) TCR Tg hosts. Each bar represents the mean ? SD of five mice. C and
D, Total number (C) and the number of CFSEhigh(D) donor CD8?T cells recovered at different times after the transfer of 2 ? 106polyclonal
Ly5.1?CD8?LN T cells into RAG0/0(open symbols) or OT-1 RAG0/0(closed symbols) hosts. E, Numbers of T cells in the different intact TCR
Proliferation of donor polyclonal CD8?T cells in different lymphopenic hosts. A and B, Two ? 106CFSE-labeled CD8?Thy1.1?LN
5233The Journal of Immunology
Adoptive T cell transfers
Donor lymph nodes (LN) CD8?T cells (2 ? 106), enriched after depletion
of CD19?and CD4?cells by MACS (Miltenyi Biotec), were labeled with
CFSE (Molecular Probes) as described (25) and injected i.v. into RAG0/0
hosts, transgenic or not for different TCRs. For secondary transfers, donor
CFSEhighand CFSE?CD8?T cells recovered from the first hosts were
sorted by flow cytometry (MoFlow, DakoCytomation). The CFSE?T cells
were relabeled with CFSE and ?3000–6000 cells were transferred into
secondary RAG20/0, P14 RAG20/0, or OT-1 RAG10/0hosts. In competition
experiments, sorted CFSEhighand CFSE?(relabeled with CFSE) donor
CD8?T cells were transferred alone or together with the same number of
CFSE-labeled LN OT-1 T cells into RAG20/0hosts. Hosts were killed at
different times after transfer. Spleen, inguinal, and mesenteric LN cells
were pooled and counted and analyzed by flow cytometry. For the sequen-
tial cell transfers, 2–3 ? 106P14 or OT-1 LN T cells were injected i.v. into
RAG20/0hosts and 4 wk later, a second subset of CFSE-labeled P14 or
OT-1 LN T cells (2–3 ? 106) was injected into the same hosts. The mice
were killed 4 wk after the second transfer. In all experiments, we used mice
expressing different Ly5 or Thy1 allotypes to discriminate the T cells from
different hosts and donors.
Spleen, inguinal, and mesenteric LN cells were stained with appropriate
combinations of different FITC, PE, PercP or PercPCy5.5, allophycocya-
nin, PE-Cy7 and allophycocyanin-Cy7 or allophycocyanin Alexa 750-con-
jugated anti-CD3, CD4, CD8, CD44, Ly5.1, Ly5.2, Thy1.1, Thy1.2, V?2,
V?8, and V?5 mAbs (BD Pharmingen). Acquisitions were done with LSR
or Canto (BD Biosciences) flow cytometers interfaced to the CellQuest or
RAG20/0mice were injected i.v. with 2 ? 106LN T cells from P14 or
OT-1. At the same time or 4 wk later, 2–7 ? 106T cell-depleted (by
MACS; Miltenyi Biotec) Ly5.1?BM precursors from P14 or OT-1
donors were injected i.v. into the same hosts. Please note that the mice
were not irradiated to preserve the existing first T cell population and to
avoid irradiation-induced inflammation. In these conditions, some of
the transferred precursor cells colonize the empty RAG0/0thymus and
reestablish T cell development. Chimeras were analyzed 8 wk after BM
hosts. Five to 6 wk later, CFSEhighand CFSE?Ly5.1?CD8?T cells were sorted and each population (after relabeling the CFSE?T cells with CFSE) was
retransferred into new RAG0/0, OT-1 RAG0/0, or P14 RAG0/0hosts. Six weeks later, the secondary hosts were sacrificed and the presence of donor
Ly5.1?CD3?CD8?T cells derived from the CFSE?(A) and CFSEhigh(B) cell subsets was assessed by flow cytometry. C, Shows the absolute number of
CFSEhigh(E) or CFSE?(F) derived Ly5.1?CD3?CD8?donor T cells recovered in the OT-1 RAG0/0, P14 RAG0/0, or RAG0/0secondary hosts. This
experiment is representative of four different experiments. D, Two million CFSE-labeled polyclonal Ly5.1?CD8?LN T cells were transferred into
Ly5.2?OT-1 RAG0/0hosts. The CD44 expression by the donor Ly5.1?CD3?CD8?T cells was assessed by flow cytometry before transfer (histogram) and
6 wk after transfer as a function of CFSE profile (dot plot). E, CFSE-labeled Ly5.1?CD8?polyclonal T cells were injected into Ly5.2?OT-1 RAG0/0hosts.
Six weeks later, CFSEhighLy5.1?CD8?T cells were sorted and re-injected into secondary RAG0/0hosts. Six weeks later, the transferred Ly5.1?CD8?T
cells, which have expanded extensively (see Fig. 2B), were sorted again, relabeled with CFSE, and transferred into tertiary RAG0/0or OT-1 RAG0/0hosts.
Six weeks later, the expansion of the donor T cells in RAG0/0vs OT-1 RAG0/0hosts was assessed by flow cytometry. Dot plots show Ly5.1?CD8?donor
T cells after gating on CD3?T cells. Please note that, while donor T cell were present in the tertiary RAG0/0hosts, we were unable to detect any in the
tertiary OT-1 RAG0/0hosts. This experiment is representative of two experiments.
Clonal inhibition of CFSEhighdonor CD8 T?cells. CFSE-labeled Ly5.1?CD8?polyclonal T cells were injected into Ly5.2?OT-1 RAG0/0
5234T CELL HOMEOSTASIS AND CLONAL COMPETITION
CD8?T cell LDP is regulated by interclonal competition
To evaluate the influence of resident naive T cells and their TCR
specificity on the fate of newly transferred T cells, we injected
CFSE-labeled polyclonal B6 CD8?LN T cells into different lym-
phopenic RAG0/0hosts, non-Tg or Tg for different class I-re-
stricted TCRs (aHY, P14 and OT-1Tg mice), i.e., empty hosts or
hosts containing monoclonal CD8?T cell populations. As ex-
pected, after transfer into RAG20/0hosts, the transferred T cells
expanded and the vast majority of cells recovered 7 wk later had
divided and lost their CFSE labeling (Fig. 1A). When the same
number of T cells was transferred into TCR Tg RAG20/0hosts,
an increased fraction of the transferred T cells did not divide
1?P14?aHY?RAG20/0. Symmetrically, the number of CFSE?
cells recovered in these hosts followed the opposite hierarchy
RAG20/0?aHY?P14?OT-1 (Fig. 1B). Because the proliferation
of some donor polyclonal CD8?T cells could be just delayed by
the presence of the resident T cells, we evaluated the number of
nondividing (CFSEhigh) CD8?T cells after transfer into RAG20/0
or OT-1 TCR Tg RAG10/0hosts at different time intervals. As
shown in Fig. 1C, the total number of CD8?T cells recovered
from both RAG20/0and OT-1 hosts reached a plateau ?1 mo after
transfer. At equilibrium the number of T cells recovered in the
OT-1 hosts was 2.5 ? lower than in the RAG20/0hosts (1.5 ? 106
vs 3.5 ? 106). The accumulation of CFSE?T cells followed the
same kinetics in both types of hosts (data not shown). In contrast,
while in RAG20/0hosts virtually all nondividing (CFSEhigh) CD8?
T cells end up disappearing after transfer, in OT-1 hosts some
CD8?T cells still not divided even 2 mo after transfer (Fig. 1D).
These findings indicate that the proliferation of a fraction of the
polyclonal CD8?T cells was stably prevented in OT-1 hosts.
We then asked about the mechanism that prevented division of
a fraction of the transferred CD8?T cells in the TCR Tg hosts.
The decreased T cell proliferation could be due to the reduced
availability of non-TCR related (IL-7, IL-15) and/or TCR-specific
(p-MHC ligands) resources, sequestered by the resident TCR Tg T
cells. If this inhibition was due to the consumption of non-TCR
specific resources, it would likely be proportional to the number of
peripheral T cells (Fig. 1E) and not related to their TCR specificity.
In contrast, if it was due to reduced TCR-specific resources, it
should be related to the TCR specificity of the hosts T cells, i.e.,
the proliferation of the undivided CFSEhighT cells recovered from
a defined TCR Tg host would be only specifically blocked by the
presence of T cells from the same host. To test this, we sorted
polyclonal CFSE?and CFSEhighLy5.1?CD8?T cells from pri-
mary injected OT-1 hosts and transferred each subset (?10.000
cells) into different non-Tg and Tg Ly5.2?secondary RAG0/0
hosts. The retransferred CFSE?(relabeled with CFSE) CD8?T
cells divided extensively, losing their CFSE and becoming rela-
tively easily detectable, whatever the type of secondary hosts they
were transferred into (Fig. 2, A and C). In contrast, the
CFSEhighLy5.1?CD8?T cells also divided extensively, becoming
CFSE?, after transfer into non-Tg or P14 Tg RAG20/0hosts (Fig.
2, B and C), but were not detectable in secondary OT-1 RAG10/0
hosts. This indicates that the transferred cells either died, did not
proliferate, or if so, they proliferated far less than in the RAG20/0
or P14Tg RAG20/0hosts (Fig. 2B). Thus, while these observations
exclude any intrinsic proliferation defect of these cells, they indi-
cate that their inability to expand and accumulate was dictated by
the TCR specificity of OT-1 T cells rather than due to competition
for other nonspecific factors like IL-2, IL-7, or IL-15 cytokines
previously involved in CD8?T cell LDP (23, 26). Similarly, OT-1
cells proliferate in P14Tg RAG20/0hosts, but not in OT-1 Tg
RAG10/0hosts, while P14 cells proliferate in OT-1 Tg RAG10/0
hosts, but not in P14 Tg RAG20/0hosts (supplemental Fig. S1)4
(18). Therefore, our findings confirmed that TCR-specific re-
sources, likely recognition of p-MHC (most probably self-ligands),
control CD8?T cell LDP. These resources seem identical with
those required by resident naive T cells to survive. They may de-
fine TCR-specific niches, which both control the number of CD8?
T cells that survive and proliferate.
Interclonal inhibition of activated/memory T cells during LDP
To determine whether naive and activated/memory CD8?T cells
are equally affected by interclonal competition for TCR-specific
ligands during LDP, we first characterized the phenotype of the
CFSEhighCD8?T cells, which did not proliferate after transfer into
TCR Tg hosts. We observed that the frequency of CD44low(90%
naive) and CD44high(10% activated/memory) in the nondividing
CFSEhighCD8?T cell fraction after transfer was similar to the
frequency found in the initial CD8?polyclonal T cells before
transfer (Fig. 2D). This observation suggests that the LDP of both
naive and activated/memory T cells was inhibited through TCR-
specific interactions. More precisely, we studied whether the LDP
of polyclonal activated/memory CD8?T cells was inhibited by the
presence of resident T cells. For this, we first transferred sorted
CFSEhighpolyclonal CD8?T cells, recovered from previous trans-
fer into OT-1 hosts, into RAG-deficient hosts. We tested whether
these cells, once expanded in these secondary RAG-deficient
hosts, would be then able to expand in tertiary OT-1 hosts. We
found that, even after extensive proliferation in secondary hosts,
the CD8?memory-like T cells derived from previous undivided
CFSEhighT cells were still able to proliferate and repopulate ter-
tiary RAG0/0hosts, but unable to proliferate remaining undetect-
able in new OT-1 hosts (Fig. 2E). Altogether, we concluded that
resident host naive T cells inhibit, in a clonal specific manner, the
LDP of not only naive but also activated/memory transferred
4The online version of this article contains supplemental material.
cells and OT-1 T cells. CFSE-labeled Ly5.1?CD8?polyclonal T cells
were injected into Ly5.2?OT-1 RAG0/0hosts. Six weeks later, CFSEhigh
(F, A) and CFSE?(E, B) cells were sorted and injected alone or together
with identical numbers of Ly5.2?OT-1 T cells (gray circles in both A and
B into secondary RAG0/0hosts). Six weeks after transfer, the secondary
RAG20/0hosts were sacrificed and the absolute number of each population
was calculated. p values (Student’s t test) are indicated (NS ? not signif-
icant). This experiment is representative of three experiments.
Clonal competition between polyclonal CFSEhighCD8?T
5235The Journal of Immunology
Clonal competition between different naive and activated T cell
Because resident naive OT-1 T cells could specifically prevent the
proliferation of a fraction of polyclonal CD8?T cells, we asked
whether these nonproliferating CD8?T cells could out-compete
the expansion of naive OT-1 T cells during LDP. In other words,
we wanted to study whether the TCR-specific requirements of
OT-1 naive T cells and polyclonal CD8?T cells for survival and
LDP would somehow overlap. We cotransferred OT-1 T cells to-
gether with polyclonal CFSEhighCD8?T cells, recovered from a
primary transfer into OT-1 hosts, into new RAG20/0hosts. We
found that, 6 wk after transfer, the number of OT-1 T cells recov-
ered in the new hosts was significantly lower when cotransferred
with CFSEhighCD8?T cells than when transferred alone, indicat-
ing that CFSEhighCD8?T cells could out-compete OT-1 T cells
(Fig. 3A). In contrast, the cotransfer of OT-1 cells with the
CFSE?CD8?T cells recovered from primary OT-1 hosts did not
modify the number of OT-1 T cells recovered (Fig. 3B). It should
be noted that although when transferred alone the number of
CFSE?CD8?T cells recovered is lower than their CFSEhigh
counterpart, they proliferate as much as CFSEhighCD8?T cells
when in presence of OT-1 T cells. Thus, the CFSEhighCD8?T
cells, which did not divide in the first OT-1 host mice, specifically
out-competed the OT-1 T cells during LDP, suggesting that the
OT-1 and the polyclonal CFSEhighCD8?T cells recovered from
OT-1 hosts share similar TCR-specific resources to survive and
We next asked whether an established population of LDP-
derived CD8?T cells could also affect the fate of newly intro-
duced CD8?T cells, and whether these effects would be related
to the TCR specificity of the populations involved. For this
purpose, a first population of Ly5.1?TCR Tg T cells was trans-
ferred into RAG0/0hosts, followed 4 wk later by a second pop-
ulation of Ly5.2?CFSE-labeled TCR Tg T cells. As control,
each cell population was injected alone in RAG0/0mice. Four
weeks later, we studied both the CFSE dilution pattern and the
number of cells recovered. The CFSE dilution profiles obtained
showed that the presence of a first LDP-derived T cell subset
reduced the proliferation of the second T cell subset and that
this effect was more marked when resident T cells bore the same
TCR as transferred T cells (Fig. 4, A and B). However, the
accumulation of cells from the second injected population was
significantly reduced in a nonspecific manner, i.e., indepen-
dently of the TCR expressed by the first population (Fig. 4,
C and D) following the rule “first come, first served”. It
should be noted that the number of resident T cells was not
modified by the introduction of a new cell subset (supplemental
CD8?T cell LDP in hosts containing
LDP-derived T cells. A first subset of
Ly5.1?TCR Tg P14 or TCR Tg
OT-1 LN CD8?T cells was injected
into RAG20/0hosts. Four weeks later,
a second subset of CFSE-labeled
Ly5.2?TCR Tg P14 (A and C) or
OT-1 (B and D) LN CD8?T cells
was injected into RAG20/0hosts pre-
viously injected or not with a first
subset of Ly5.1?TCR Tg P14 or
OT-1 LN CD8?T cells. Another 4
wk later, the hosts were sacrificed and
the CFSE profiles of the second sub-
set of Ly5.2?CD3?CD8?P14 (A) or
OT-1 (B) cells were analyzed after
transfer into new RAG20/0
empty (pale gray histogram), or con-
taining LDP-derived P14 (black his-
togram) or OT-1 (dark gray histo-
represents the data obtained from one
mouse and is representative of seven
to ten mice. C and D, Absolute num-
ber of cells of the second subset of
Ly5.2?TCR Tg P14 (C) or OT-1 (D)
LN CD8?T cells recovered after
transfer into RAG20/0hosts empty
(pale gray circles), or containing
LDP-derived P14 (F) or OT-1 (dark
gray circles) T cells. p values are in-
dicated. The data shown are a pool of
two different experiments.
Clonal inhibition of
5236T CELL HOMEOSTASIS AND CLONAL COMPETITION
We also studied whether a previously established population of
LDP-derived CD8?T cells could affect the fate of a newly pro-
duced population of CD8?T cells. To test this, we performed a set
of experiments in which T cell-depleted BM TCR Tg precursors
were injected into nonirradiated RAG20/0hosts either simulta-
neously or 4 wk after TCR Tg LN T cells transfer. We choose to
not irradiate the hosts to avoid any inflammatory response, which,
through the release of cytokines, may disturb homeostatic process,
and to preserve the existing first T cell population during sequen-
tial transfer. The mice were sacrificed 8 wk after BM engraftment
and the numbers of each of the T cell subsets evaluated. In these
settings, some of the transferred precursor cells colonize the empty
RAG0/0thymus and reestablish T cell development. The BM-de-
rived T cells migrated from the thymus to the periphery at a time
when “LN subset” already reached a plateau 3–4 wk after transfer.
We first assessed the fate of BM-derived OT-1 T cells in the pe-
riphery of hosts containing LDP-derived OT-1 T cells. As shown
in Fig. 5A, the number of BM-derived OT-1 T cells is decreased
more than twice (11.3 vs 4.8 ? 106cells) in the presence of LDP-
derived OT-1 T cells, while the number of LDP-derived OT-1 T
cells was unaffected by the generation of BM-derived T cells. To
determine whether this inhibition of new T cell establishment de-
pends on the TCR specificity of the preinstalled LDP-derived T
cell population, we performed another experiment in which the T
cell-depleted BM cells exhibited either the same or different TCR
specificity as the LDP-derived T cells. As exemplified for the
transfer of OT-1 BM precursors into RAG0/0containing or not
LDP-derived OT-1 T cells, the presence of the LDP-derived
Ly5.1?(gray circles) cells from OT-1 RAG0/0hosts were injected into empty RAG20/0hosts either alone or together. The number of peripheral OT-1 T
cells originated from each subset was determined 8 wk later. B–D, LN CD8?Ly5.2?T cells and BM Ly5.1?precursor cells from either TCR Tg P14 or
OT-1 RAG0/0donor mice were sequentially injected into empty RAG20/0hosts. Eight weeks after BM transfer, the hosts were sacrificed. B, The thymic
(first column) and peripheral OT-1 BM-derived populations (second column) were compared after transfer into either empty RAG20/0hosts (first row) or
those containing LDP-derived OT-1 T cells (second row), for the representation of DN, DP, and SP subsets and CD44 expression levels, respectively. Third
column, CD44 expression of LDP-derived resident OT-1 T cells upon adoptive transfer into RAG0/0(first row) or into RAG0/0in the presence of
BM-derived T cells as well (second row). C, The absolute numbers of BM-derived P14 (diamonds) and OT-1 (circles) Ly5.1?CD3?CD8?were calculated
after transfer alone (pale gray symbols) or in the presence of LN P14 (black symbols) or OT-1 (dark gray symbols) CD8?T cells. p values are indicated.
D, The absolute numbers of LN-derived P14 (left) and OT-1 (right) Ly5.1?CD3?CD8?were calculated after transfer alone (diamonds) or in the presence
of BM-derived P14 (E) or OT-1 (F) T cells. p values are indicated.
Clonal inhibition of BM-derived T cell accumulation in hosts containing LDP-derived T cells. A, LN CD8?Ly5.2?T cells (F) and BM
5237 The Journal of Immunology
resident T cells did not alter either the DN, DP, and SP thymus
subsets or the subsequent CD44 expression level of the peripheral
BM-derived T cells (Fig. 5B). Thus, resident T cells did not affect
either the thymic T cell development from the subsequently in-
jected BM precursors, or their phenotype at the periphery. How-
ever, as shown in Fig. 5C, the number of BM-derived peripheral
P14 and OT-1 T cells was significantly decreased in the presence
of LDP-derived T cells bearing the same TCR specificity, i.e.,
from the same TCR Tg donors. In contrast, the phenotype and the
number of LDP-derived T cells remained unaffected by the new
incoming BM-derived T cells (Fig. 5, B and D). These results
indicate that the peripheral accumulation of newly developing T
cells is impaired by the presence of resident T cells bearing the
same TCR specificity, i.e., in a TCR-dependent manner.
In the present study, we evaluated to which extent specific TCR/
p-MHC interactions may overlap between different CD8?T cell
clones. We found that clonal competition, determined by specific
TCR/p-MHC interactions (most probably self-ligands), affects
both naive and activated CD8?T cells and occurs during either
peripheral cell accumulation after thymic output or peripheral T
cell recovery after LDP.
LDP of monoclonal TCR Tg T cells is blocked after transfer into
hosts containing monoclonal T cells bearing the same TCR (18–
20) and our own results (supplemental Fig. 1). We now extend this
concept of “intraclonal competition” during LDP, to that of “in-
terclonal competition,” by showing that the proliferation of some
adoptively transferred polyclonal CD8?T cells is strictly and sta-
bly blocked by the presence of specific T cell clones in the lym-
phopenic hosts. This interpretation is strongly supported by the
observation that the polyclonal CD8?T cells that were unable to
proliferate in OT-1 hosts specifically out-competed the OT-1 T
cells when both T cell populations were cotransferred into
RAG20/0hosts, while other T cell clones could not. This indicates
that these cells disclose the same TCR-specific ligands or require-
ments to survive and proliferate during LDP. In addition, we found
that these requirements were shared by both naive and activated
LDP-derived CD8?T cells. It should be noted that after transfer
into RAG20/0hosts, we detected (day 9) a small fraction of
CFSElowCD8?polyclonal T cells. These cells were not observed
after transfer into OT-1 hosts and likely correspond to a “slow
proliferating” population (27). The progressive dilution of CFSE
labeling over time (day 28; all cells become CFSE?by day 60) led
to their complete disappearance, suggesting that some of these
cells proliferated in the hosts, probably in response to unlimited
access to IL-7 and MHC-ligand resources (27).
The fraction of polyclonal CD8?T cells whose proliferation
was inhibited after adoptive transfer into TCR Tg hosts varied
according to the TCR specificity of monoclonal resident T cells
and followed the hierarchy OT-1?P14?aHY hosts. It was previ-
ously shown that the LDP of naive CD4?T cells is not controlled
by the number of resident T cells but by their repertoire diversity
(22). In their study, Min et al., (22) generated RAG-deficient hosts
containing identical numbers of memory CD4?T cells, but with
different T cell repertoires complexity, by transferring different
numbers of naive CD4?T cells. They showed that the expansion
of a second polyclonal naive CD4?T cell population transferred
into these different hosts was inversely proportional to the TCR
diversity of the resident memory T cells. Because in the monoclo-
nal RAG-deficient TCR Tg mice, the number of T cells varies
according to the TCR expressed and was higher in OT-1 than in
P14 or aHY hosts, the differential expansion of polyclonal CD8?
T cells that we observed could be related to the clone size of the
resident T cells rather than to their diversity. Recent findings from
our laboratory, showing that T cell clone size is determined ac-
cording to the TCR promiscuity, i.e., cells with a broader pattern
of specificities occupying a larger space (11), reconcile our present
results with those of Min et al. (22). Indeed, one can easily pos-
tulate that promiscuous TCR-bearing resident T cells, like OT-1 T
cells, which interact with more p-MHC ligands (11) or the P14 T
cells that also interact with different self-peptides (28), will be
more numerous than the less promiscuous aHY T cells. In the case
of the transferred polyclonal T cells, the number of different T cells
that will not proliferate would be determined by the availability of
p-MHC ligands that are not sequestered by the resident T cells.
Using sequential transfers of mature T cells, we found that,
while a population of resident naive cells selectively inhibited the
proliferation and accumulation of newly transferred CD8?T cells
bearing the same TCR specificity, LDP-derived resident T cells
inhibited the expansion of all newly transferred T cells indepen-
dently of their TCR specificity. They suggest that TCR interactions
may also increase the first arrived LDP-derived T cells ability to
compete for and consume non-TCR specific trophic resources re-
quired for the survival and accumulation of newly transferred T
cells (20, 21). IL-7 or IL-15 cytokines, shown to be involved in
memory CD8?T cell survival, are the most likely candidates.
Alternatively, LDP-derived resident cells could actively inhibit or
eliminate the newly proliferating T cells through TNF-? or IFN?
secretion, by a mechanism that was previously shown to be in-
volved in memory CD8?T cell attrition (29). Altogether, these
observations strongly suggest that the survival requirements of na-
ive and LDP-derived CD8?T cells do not completely overlap.
Although they both require recognition of p-MHC, the activated
LDP-derived T cells seemed to be more capable of using other
“non-TCR specific” resources like cytokines.
By transferring BM precursors instead of mature T cells into
hosts containing LDP-derived T cells, we found that the resident
population selectively impaired the accumulation of newly pro-
duced thymus emigrants bearing the same TCR, thus in a TCR-
specific manner. Most likely, the LDP-derived long-term resident
T cells monopolize TCR-specific resources required by recent T
cells to accumulate in the peripheral pools. Despite the well-es-
tablished differences of requirements between naive and memory T
cell survival (24, 30, 31), our findings suggest that they may share
enough resources to establish some form of pre-emptive competition
as previously observed using parabiosis experiments (3). This pre-
emptive competition would favor peripheral T cell diversity, since the
new developed T cells, expressing redundant existing specificities, do
not survive. Interestingly, LDP-derived resident T cells affect differ-
ently recent T cells arising from BM precursors and adoptively trans-
different survival requirements and/or competitive fitness. Alterna-
tively, compared with the limited number of cells introduced upon
adoptive transfer, the higher number of continuously produced recent
thymic emigrants, may justify the different final outcomes.
In contrast, resident activated monoclonal T cells were unaffected
by the entry of newly thymus-produced monoclonal T cells into the
peripheral pools (Fig. 5, B and D). Other studies have shown that
polyclonal BM-derived T cells were able to replace CD4?TCR Tg
LDP-derived T cells, but not memory A1 T cells (32). These differ-
ences may suggest different behaviors for T cell populations involved
and/or in the model systems used. Replacement of peripheral T cells
should vary with the rates of production and the diversity of the T
cells produced (33). Diverse polyclonal T cells by their increased
ability to exploit different niches should in long-term replace, at least
partially, established monoclonal T cells (3).
5238T CELL HOMEOSTASIS AND CLONAL COMPETITION
Overall, our results strongly support that in the absence of in- Download full-text
tentional immunization, competition for overlapping p-MHC li-
gands controls peripheral CD8?T cell survival, proliferation, and
accumulation. This implies the existence of a complex network of
TCR/p-MHC interactions defining specific niches, which regulate
several homeostatic events and dictate the final composition of the
T cell pools (34). The fact that resident naive T cells prevent the
establishment of new T cells expressing identical TCR specificities
favors peripheral T cell repertoire diversity. In contrast, the fact
that LDP-derived activated T cells prevents, in a non-TCR specific
manner, the accumulation of other activated T cell subsets might
enhance the efficiency of an immune response by avoiding by-
stander expansion of non-Ag specific cells and favoring the ex-
pansion of the appropriate set of responding cells. These findings
may help understanding the homeostatic events leading to the es-
tablishment of a new T cell repertoire and occurring during the
reconstitution of the peripheral T pools in individuals suffering
from infection or therapeutic-induced lymphopenia and potentially
subjected to LDP processes.
We thank Anne Louise from the Cytometry Platform at the Pasteur Insti-
tute for the FACS sorting experiments.
The authors have no financial conflict of interest.
1. Almeida, A. R., J. A. Borghans, and A. A. Freitas. 2001. T cell homeostasis:
thymus regeneration and peripheral T cell restoration in mice with a reduced
fraction of competent precursors. J. Exp. Med. 194: 591–599.
2. Tanchot, C., M. M. Rosado, F. Agenes, A. A. Freitas, and B. Rocha. 1997.
Lymphocyte homeostasis. Semin. Immunol. 9: 331–337.
3. Freitas, A. A., F. Agenes, and G. C. Coutinho. 1996. Cellular competition mod-
ulates survival and selection of CD8?T cells. Eur. J. Immunol. 26: 2640–2649.
4. Polic, B., D. Kunkel, A. Scheffold, and K. Rajewsky. 2001. How ?? T cells deal
with induced TCR ? ablation. Proc. Natl. Acad. Sci. USA 98: 8744–8749.
5. Labrecque, N., L. S. Whitfield, R. Obst, C. Waltzinger, C. Benoist, and
D. Mathis. 2001. How much TCR does a T cell need? Immunity 15: 71–82.
6. Kirkberg, J., J. Berns, and H. von Boehmer. 1997. Peripheral T cell survival
requires continual ligation of the T cell receptor to major histocompatibility com-
plex-encoded molecules. J. Exp. Med. 186: 1269–1275.
7. Brocker, T. 1997. Survival of mature CD4 T lymphocytes is dependent on major
histocompatibility complex class II-expressing dendritic cells. J. Exp. Med. 186:
8. Takeda, S., H. Rodewald, H. Arakawa, H. Bluethmann, and T. Shimizu. 1996.
MHC class II molecules are not required for survival of newly generated CD4?
T cells, but affect their long-term life span. Immunity 5: 217–228.
9. Tanchot, C., F. A. Lemonnier, B. Perarnau, A. A. Freitas, and B. Rocha. 1997.
Differential requirements for survival and proliferation of CD8 naive or memory
T cells. Science 276: 2057–2062.
10. Witherden, D., N. van Oers, C. Waltzinger, A. Weiss, C. Benoist, and D. Mathis.
2000. Tetracycline-controllable selection of CD4?T cells: half-life and survival
signals in the absence of major histocompatibility complex class II molecules.
J. Exp. Med. 191: 355–364.
11. Hao, Y., N. Legrand, and A. A. Freitas. 2006. The clone size of peripheral CD8
T cells is regulated by TCR promiscuity. J. Exp. Med. 203: 1643–1649.
12. Kassiotis, G., R. Zamoyska, and B. Stockinger. 2003. Involvement of avidity for
major histocompatibility complex in homeostasis of naive and memory T cells.
J. Exp. Med. 197: 1007–1016.
13. Kieper, W. C., J. T. Burghardt, and C. D. Surh. 2004. A role for TCR affinity in
regulating naive T cell homeostasis. J. Immunol. 172: 40–44.
14. Ernst, B., D. S. Lee, J. M. Chang, J. Sprent, and C. D. Surh. 1999. The peptide
ligands mediating positive selection in the thymus control T cell survival and
homeostatic proliferation in the periphery. Immunity 11: 173–181.
15. Goldrath, A. W., and M. J. Bevan. 1999. Low-affinity ligands for the TCR drive
proliferation of mature CD8?T cells in lymphopenic hosts. Immunity 11:
16. Jameson, S. 2002. Maintaining the norm: T-cell homeostasis. Nat. Rev. 2:
17. Kieper, W. C., A. Troy, J. T. Burghardt, C. Ramsey, J. Y. Lee, H. Q. Jiang,
W. Dummer, H. Shen, J. J. Cebra, and C. D. Surh. 2005. Recent immune status
determines the source of antigens that drive homeostatic T cell expansion. J. Im-
munol. 174: 3158–3163.
18. Troy, A. E., and H. Shen. 2003. Cutting edge: homeostatic proliferation of pe-
ripheral T lymphocytes is regulated by clonal competition. J. Immunol. 170:
19. Moses, C. T., K. M. Thorstenson, S. C. Jameson, and A. Khoruts. 2003. Com-
petition for self ligands restrains homeostatic proliferation of naive CD4 T cells.
Proc. Natl. Acad. Sci. USA 100: 1185–1190.
20. Ge, Q., A. Bai, B. Jones, H. N. Eisen, and J. Chen. 2004. Competition for self-
peptide-MHC complexes and cytokines between naive and memory CD8?T
cells expressing the same or different T cell receptors. Proc. Natl. Acad. Sci. USA
21. Agenes, F., J. P. Dangy, and J. Kirberg. 2008. T cell receptor contact to restricting
MHC molecules is a prerequisite for peripheral interclonal T cell competition.
J. Exp. Med. 205: 2735–2743.
22. Min, B., G. Foucras, M. Meier-Schellersheim, and W. E. Paul. 2004. Spontane-
ous proliferation, a response of naive CD4 T cells determined by the diversity of
the memory cell repertoire. Proc. Natl. Acad. Sci. USA 101: 3874–3879.
23. Cho, J. H., O. Boyman, H. O. Kim, B. Hahm, M. P. Rubinstein, C. Ramsey,
D. M. Kim, C. D. Surh, and J. Sprent. 2007. An intense form of homeostatic
proliferation of naive CD8?cells driven by IL-2. J. Exp. Med. 204: 1787–1801.
24. Tanchot, C., and B. Rocha. 1995. The peripheral T cell repertoire: independent
homeostatic regulation of virgin and activated CD8?T cell pools. Eur. J. Im-
munol. 25: 2127–2136.
25. Garcia, S., J. DiSanto, and B. Stockinger. 1999. Following the development of a
CD4 T cell response in vivo: from activation to memory formation. Immunity
26. Tan, J. T., B. Ernst, W. C. Kieper, E. LeRoy, J. Sprent, and C. D. Surh. 2002.
Interleukin (IL)-15 and IL-7 jointly regulate homeostatic proliferation of memory
phenotype CD8?cells but are not required for memory phenotype CD4?cells.
J. Exp. Med. 195: 1523–1532.
27. Min, B., H. Yamane, J. Hu-Li, and W. E. Paul. 2005. Spontaneous and homeo-
static proliferation of CD4 T cells are regulated by different mechanisms. J. Im-
munol. 174: 6039–6044.
28. Saibil, S. D., T. Ohteki, F. M. White, M. Luscher, A. Zakarian, A. Elford,
J. Shabanowitz, H. Nishina, P. Hugo, J. Penninger, et al. 2003. Weak agonist
self-peptides promote selection and tuning of virus-specific T cells. Eur. J. Im-
munol. 33: 685–696.
29. McNally, J. M., C. C. Zarozinski, M. Y. Lin, M. A. Brehm, H. D. Chen, and
R. M. Welsh. 2001. Attrition of bystander CD8 T cells during virus-induced
T-cell and interferon responses. J. Virol. 75: 5965–5976.
30. Freitas, A. A., and B. Rocha. 1999. Peripheral T cell survival. Curr. Opin. Im-
munol. 11: 152–156.
31. Surh, C. D., and J. Sprent. 2005. Regulation of mature T cell homeostasis. Semin.
Immunol. 17: 183–191.
32. Bourgeois, C., G. Kassiotis, and B. Stockinger. 2005. A major role for memory
CD4 T cells in the control of lymphopenia-induced proliferation of naive CD4 T
cells. J. Immunol. 174: 5316–5323.
33. Tanchot, C., and B. Rocha. 1997. Peripheral selection of T cell repertoires: the
role of continuous thymus output. J. Exp. Med. 186: 1099–1106.
34. Freitas, A. A., and B. Rocha. 2000. Population biology of lymphocytes: the flight
for survival. Annu. Rev. Immunol. 18: 83–111.
5239The Journal of Immunology