The B7/CD28 pathway is much more complicated than expected. Chambers C.A. et al. Immuunol. Rev. 153:27, 1996. B7/CD28/CTLA-4. Activated T cell. AP2. AP2. TCR. CTLA-4. CD28. P. P. peptide/ MHC. B7.1 (CD80). Activated Professional APC. B7.2 (CD86).
Immuunol. Rev. 153:27, 1996.
Activated T cell
Activated Professional APC
CTLA-4 is not readily detectable innaïve T cells
but is rapidly upregulated upon T cell activation.
CTLA-4 mRNA can be readily detected within 1 h of
TCR engagementand peaks at about 24–36 h.
However, CTLA-4 is not readilydetectable at the cell surface
until 24–48 h after activation.
An accurate assessment of the kinetics of
CTLA-4protein expression is complicated
by the fact that it is notprimarily expressed at the cell surface.
Surface expression of CTLA-4 is tightly regulated as a result
of the presence of a tyrosine-based intracellular localization
motif in its cytoplasmic tail which allowsan association with
the medium subunit of the clathrin-coated
pit adaptor complex, AP2,thus providing a mechanism for
CTLA-4cellular localization. CTLA-4 expression onthe
T cell surface is stabilized and increased by tyrosine
phosphorylationof the endocytosis motif, which inhibits
AP2 association. Itis noteworthy that the intracellular portion
of CTLA-4 is 100%conserved among many different species
of animals, suggestingthat control of intracellular trafficking
may be extremely importantfor its function. This motif results
in both the rapid endocytosis of CTLA-4 from the cell surface to
endosomalcompartments, as well as the targeting of at
least some CTLA-4to the lysosomes for degradation.
It is unclear whetherendocytosed CTLA-4 can recycle back
to the surface or if thisrepresents a terminal pathway for the
As previously discussed, CD28 is constitutively expressed onT cells, whereas CTLA-4 appears after
activation. Because ofthis, and perhaps as a result of our innate appreciation forsymmetry, the idea arose
that CD28 engagement allowed initiation,while CTLA-4 provided for termination of immune responses.
Surprisingly, the majority of the in vitro data has demonstratedan inhibitory role for CTLA-4 in the early
stages of T cellactivation. IL-2 production, expression of early markers such asCD69 and CD25, and a
number of other aspects of activation areinhibited upon CTLA-4 cross-linking. These events take place
within hours of T cell activation with anti-CD3 and CD28. Infact, the inhibition of the induction of
IL-2 transcriptionwas detected 4 h after stimulation.
This suggests eitherthat there is a physiologically relevant intracellular poolof CTLA-4 present
in naïve T cells or that protein expressionis induced rapidly upon activation.
The possibility that CTLA-4 can inhibit earlystages of T cell activation has led to the
development of modelsthat stress that the dynamic interplay of costimulatory andTCR signals
depends on the activation state of the T cell aswell as the activation state of the antigen-presenting cell.
The threshold model:When B7 levelsare low and TCR signals are weak, the amount of CTLA-4 induced
is low but may be sufficient to minimize costimulation and preventactivation. Under these conditions
CTLA-4 may set a thresholdfor activation. In other words, it plays a role in setting the stimulatory
threshold required for a T cell to progress to full activation.
As in the classicaltwo-signal model,
an encounter of a naïve T cell with acell
expressing appropriate MHC/antigen complex
but lackingB7 does not result in activation of
the T cell owing to lackof costimulation.
The cells receiving a TCR signal in the absence
of CD28-mediated costimulation may be
However,engagement of the TCR can lead
to rapid induction and/or mobilizationof
small amounts of CTLA-4. Under conditions
where there isan incompletely activated
APC expressing only low amounts ofB7,
CTLA-4 could, by virtue of its higher affinity,
outcompeteCD28 for B7 and/or deliver
inhibitory signals. This could effectively
raise the threshold of CD28 and/or TCR
signals needed for fullactivation.
Chambers, C.A. et al. Ann.Rev.Immunol. 19:565, 2001
There are two scenarios in which CTLA-4 may play a role in establishinga threshold for
CD28 and/or TCR signals needed for activationof naïve T cells.
Both presume that low levels of CTLA-4pre-exist or can be rapidly induced in naïve T cells upon
engagement of the TCR and CD28.
The first scenario deals with the regulation of the responseof T cells to tonic signaling by
Continuous TCR interactions with self-peptide/MHCprovide important signals for the survival of
peripheral T cells.
Some of these tonic interactions, under conditions of lowlevels of CD28/B7 interaction, might be
sufficiently stimulatoryto lead to the activation of CD4+ T cells and the inductionand/or mobilization of
CTLA-4. Based on the analysis of theCTLA-4-/- mice, it is speculated that CTLA-4might prevent the
signals generated by these interactions fromleading to full activation of CD4+ T cells.
Thus, CTLA-4may be involved in maintaining naïve CD4+ T cells and previouslyactivated T cells
in a resting state.
This model is supported by the phenotype of CTLA-4-/- mice. The expansion of T cells that occurs in these
miceis polyclonal. Thissuggests that the expansion is not the result of a failureto terminate responses
to few environmental pathogens. Thus, the phenotype of CTLA-4-/- mice results from the activation and
expansion of T cells reactive to low-affinity self-ligands due to a decreased activation threshold.
The second scenario suggests a role for CTLA-4 in maintainingperipheral tolerance of T cells with
specificity for tissue-specificantigens that are not expressed in the thymus and have not beendeleted as
a consequence of negative selection. CTLA-4 may provide an additional level of regulation toensure
peripheral tolerance by preventing activation when aT cell encounters a normal self-antigen in the
context of lowB7 expression.
This scenario may explain the observationthat CTLA-4 blockade or deficiency accelerates the onset and
severity of insulitis and diabetes in nonobese diabetic (NOD)mice expressing a transgenic ß TCR cloned
from anislet-ß-cell-specific CD4+ T cell clone isolated froma NOD mouse. T cells bearing this
TCR are efficientlyselected, rather than deleted in the thymus, demonstrating thatcentral tolerance is not
effective for T cells with this specificity.The observation that blockade or loss of CTLA-4 dramatically
accelerates disease in this model system indicates a role for CTLA-4 in maintaining an activation threshold
for autoreactive T cells bearingTCRs specific to autoantigens.
The attenuation model maintaining:When B7 levelsare high and TCR signals are strong, the higher levels of CTLA-4
induced after activationmay be able to attenuate the response of activated cells. This model suggest that
after activation and subsequent entry into the cell cycle, CTLA-4 can limit the capacity of a T cell
Under conditions where the APC is activated maintaining
and expressing high levels of B7,
CD28 costimulation may dominate and
activation proceeds. However, this results in
induction and/or mobilization of CTLA-4
that may be proportional to the strength of the
TCR signal, resulting in differential inhibition.
CTLA-4 preferentially attenuatesthe expansion
of T cells that have been strongly activated.
This notion is supported by the effect of
CTLA-4-blockade on the proliferative capacity,
or averagenumber of daughter cells per responder,
for T cells primed withagonist or weak agonist
peptides in adjuvant.
Blockade significantlyincreased the proliferative
capacity of T cells primed withthe agonist ligand
but had a minimal effect on the cells responding
to the weak agonist peptide..
Chambers, C.A. et al. Ann.Rev.Immunol. 19:565, 2001
CTLA-4 would prevent this high-affinity population from maintainingdominating the primed pool by restricting
proliferation earlyin the response. Engagement of CTLA-4 would then serve to broadenthe pool of
T cells by limiting clonal representation of thehigh-affinity population, thus allowing more equal
representationof the cells bearing lower affinity TCRs in the early stagesof the clonal evolution of
It appears thatincreased CTLA-4 expression upon activation modulates T cellresponses
differentially and might serve to limit the burstsize of responding T cells. Overall, the results suggest
thatthe quality of the TCR signal is critical to determining ifand/or how dramatically CTLA-4 regulates
the proliferative capacityof any antigen-specific clone selected from the T cell repertoire.
While the TCR and CD28 might primarily determine the range ofT cells responding to antigen,
CTLA-4 would limit clonal representationof T cells with high-affinity TCRs.
Role played by CTLA-4 in regulating antigen-specific T cell responses. (a)
The threshold and attenuation models make different predictions about the
function of CTLA-4, depending on the stimulatory conditions encountered during
activation. Shown are graphical representations that relate the effect of CTLA-4–
mediated inhibition on T cell expansion to the degree of stimulation (TCR +
CD28 signals). Threshold model: T cells that are stimulated below a given
activation threshold (dashed lines) will not commit to entering the cell cycle.
According to the threshold model, CTLA-4 may participate in setting a threshold
for activation, thereby keeping cells that receive low amounts of stimulation from
responding (red lines). In the absence of CTLA-4 signaling, the T cell activation
threshold would be shifted, so that cells that receive weaker amounts of stimulation
could now become activated and proliferate (green lines).
Attenuation model: a responses.
certain amount of T cell stimulation may be required to up-regulate and/or mobilize
CTLA-4, thereby allowing it to participate in the regulation of a T cell response.
Thus, according to the attenuation model, CTLA-4 does not affect the threshold of
stimulation required for a T cell to enter the cell cycle; rather, it affects the extent of
subsequent expansion. Based on its expression and localization patterns, CTLA-4
may most significantly restrict the expansion of T cells that receive stronger amounts
of stimulation. (b) A diverse population of cells may respond to a given antigen, but
due to the process of antigen-driven selection, a greater number of T cells bearing
TCRs of higher affinity will be present in this population. By preferentially restricting
the expansion of cells that receive stronger TCR signals, CTLA-4 would prevent T
cells bearing higher affinity TCRs from dominating the response (red line). Thus,
in the absence of CTLA-4, the distribution of cells within the responding population
would shift toward higher affinity TCRs, resulting in a polyclonal response of reduced
diversity (green line).
The process by which CTLA-4-induced negative regulation occurs is clearly distinctfrom activation-induced
cell death. There is no evidence that CTLA-4 ligationin conjunction with TCR and CD28 cross-linking on
resting murineT cells induces apoptosis. Further, CTLA-4ligation does not alter the CD28-mediated
upregulation of survivalfactor bcl-2, arguing against apoptosis induction beingthe major mechanism of
The fact that CTLA-4 deficiency is lethal relatively early inlife and that the Fas pathway is intact
indicates that CTLA-4can play a critical role in limiting T cell expansion.
CTLA-4has effects on at least two aspects of activation that havecritical relevance to proliferation.
The first is on IL-2 production.CD28 costimulation enhances IL-2 production both at the levelof
transcription and mRNA stabilization. Extensiveligation of CTLA-4 under suboptimal conditions of
stimulationby TCR plus CD28 can result in inhibition of IL-2 production,probably at the level of
transcription. CTLA-4ligation inhibits TCR-induced productionof cdk4, cdk6, and cyclin D3, all of which
are required forG0/G1 progression.
Thus CTLA-4 can limit expansion notonly by reducing production of an important growth factor,
butalso by inhibiting TCR-mediated induction and assembly of essentialcomponents of the cell cycle
CTLA-4 induces inhibitory cytokines? responses.
It has been reported that cross-linking of CTLA-4 may enhanceproduction of TGFß by activated T cells.
This raises the possibility that CTLA-4 does not directly inhibitT cell activation but does so by the active
induction of thisinhibitory cytokine. One observation taken as evidence for anindirect role for CTLA-4
in the inhibition of T cell responseswas the failure of mixed bone marrow chimeric mice generatedwith
CTLA-4-/- and wild-type bone marrow to develop a phenotypeequivalent to the ctla-4 null mice.
It was proposed thatthere is no primary defect in CTLA-4-/- T cells but that theCTLA-4-/- phenotype is
due to a failure of T cells to secreteinhibitory cytokines such as TGFß.
Recent studies showed that CTLA-4 cross-linking resulted inthe inhibition of proliferation of T cells
from TGFß-/- mice or of T cells from mice lacking Smad3, a critical downstreamsignaling molecule in the
TGFß pathway. This suggests that neither TGFßproduction nor its signaling pathway is required for
CTLA-4-mediatedinhibition of T cell responses. These studies also failed toshow a role for
CTLA-4 in regulating TGFß production,since it was produced by CTLA-4-/- T cells.
Finally, CTLA-4ligation failed to induce production of TGFß by normalnaïve T cells.
CTLA-4 and regulatory T cells? responses.
Over the last several years, a considerable literature has documenteda role for CD25+CD45RBlowCD4+
regulatory T cells (Treg) in themaintenance of peripheral tolerance to organ-specific self-antigens.
The observation that these cells constitutively expressCTLA-4 has raised the possibility that CTLA-4
may be directlyinvolved in their function. Moreover, some studies indicate that someCTLA-4 functions
may not be necessarily T cell autonomous:(cf. bone marrow chimera with CTLA-4 -/- bone marrow cells or a
mixture of CTLA-4 -/- and CTLA-4 +/+ bone marrow cells.)
Administrationof either anti-CTLA-4 antibodies or anti-TGFß reversedthe inhibitory effects of transferred
CD25+ Treg cells on theinduction of colitis by transferred CD4+ CD25- cells in SCIDmice. This was taken
as evidence for a blocking of thesuppressive effects of Treg cells by preventing CTLA-4-mediatedinduction
of TGFß production. However, it does not exclude the possibility that theeffect of anti-CTLA-4 was a
result of enhancement of the effectorT cells. Other investigators have shown that CTLA-4 does not have a role
In the function of Treg cells. CD25+CD4+ T cells from CTLA-4-/- mice retainedinhibitory activity in in vitro
One interestingcharacteristic of Treg cells is a failure to secrete IL-2 orto proliferate in response to ligation of
the TCR and CD28. The hyporesponsiveness might be attributed to the inhibitoryproperties of CTLA-4.
However, to date, attempts to releasethe block on proliferation in response to TCR engagement by
CTLA-4 blockade have not been successful.
T cell–mediated immunotherapy represents a promising treatmentfor human malignancies.
In cancer patients, the absence of efficienttumor-specific immunity can be related to inadequate
APC functionor to T cell tolerance/ignorance towards tumor antigens.
Mice were injected intravenouslywith 5 x 104 or 105 B16-F10 melanoma cells.
Treatment using irradiatedF10/g cells (GM-CSF transfected B16-F10 cells; Presumably,GM-CSF
production at the site of vaccination might attract hostAPCs and enhance their function in vivo.) and
antibodies was started after 24 h.
These results indicatethat CTLA-4 blockade and GM-CSF–producing
vaccines actsynergistically to cause rejection of poorly immunogenic tumors.
van Elsas A. J. Exp. Med. 190:355, 1999
No treatment treatment
Depletion of Treg
Vacc. + CTLA-4 blocking
Vacc. + depletion of Treg
+ CTLA-4 blocking
+ depletion of Treg
Figure 3. CD25+ T cell depletion before vaccination enhances efficacy of treatment.
(A) Survival data of mice challenged subcutaneously with 2.5 x 103 B16-BL6 tumor cells.
Mice received either no treatment (n = 6, ), or depleting anti-CD25 on day -4 (n = 6, ) or vaccination with
GM-CSF–producing B16 on days 0, 3, and 6. The vaccinated mice were divided over three groups that
received the following Ab: CTLA-4 blocking Ab on days 0, 3, and 6 (n = 8, •); depleting anti-CD25 Ab on
day -4 (n = 8, x); or depleting anti-CD25 Ab on day -4 plus CTLA-4–blocking Ab on days 0, 3, and 6 (n = 8, ).
(B) Survival data of mice challenged subsutaneously (day 0) with 5 x 103 B16-BL6 tumor cells. Mice received
either depleting anti-CD25 Ab on day -4 (n = 6, ) or were vaccinated on days 0, 3, and 6 with GM-CSF–
producing B16. The vaccinated mice were divided over two groups that received the following Ab:
blocking anti–CTLA-4 Ab on days 0, 3, and 6 (n = 9, •); or depleting anti-CD25 Ab on day -4 combined with
blocking anti-CTLA-4 Ab on days 0, 3, and 6 (n = 9, )..
Sutmuller, R.P. et al. J. Exp. Med. 194:823, 2001
In conclusion, the data presented in this paper reveal that treatmentcombination of CTLA-4 blockade
and elimination of CD25+ Tregcells can result in more effective therapeuticantitumor immunity than
when these intervention strategies areapplied separately.
These findings support the notion that CD25+ Treg cells and CTLA-4 represent two alternative pathways
forsuppression of autoreactive T cell immunity, but do not excludethat functional overlap between these
Simultaneous intervention with both regulatorymechanisms appears to be a highly promising strategy
for theinduction of T cell immunity against tumor-associated autoantigensin the immunotherapy of cancer.
leads to tumor rejection but also induces autoimmune responses