| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Theoretical Biology and Biophysics, Los Alamos National Laboratory, Los Alamos, NM 87545;
University of New South Wales, Kensington, New South Wales, Australia; and
Institute of Molecular Medicine, John Radcliffe Hospital, Oxford, United Kingdom
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Analysis of the evolution of CD8+ T cell
responses in humans has relied on studies of natural infection. Our
work has focused on the response to EBV and has involved an analysis of
the Ag specificity and clonal composition of the
CD8+ T cell response to the virus during the
primary phase of infection (the primary response) and 1 year later,
during the persistent phase of infection with the virus (the memory
response; in the context of EBV infection, we use this term to describe
the response that persists following resolution of primary infection).
The results uphold the concepts that epitope selection and clonal
dominance may be a feature of primary responses (9, 10)
and that the primary clonal burst is followed by decay in the response
(9, 11). However, comparison of the epitope specificity
and clonal composition of the CD8+ T cell
response to EBV during the primary and memory responses reveals
surprising changes in the hierarchies of dominance both at the level of
the epitope and, in some responses, at the level of the T cell clones
responding to these epitopes (12). At the level of the
epitope, we find that the most dominant responses are most extensively
down-regulated during the resolution of the primary response and these
responses actually become relatively less dominant in memory (12, 13). At the level of the T cell clones responding to the
epitopes, we also find that, in many instances, the clones that
dominate the primary response to an epitope are most heavily
down-regulated, leaving other clones to dominate the memory response.
Thus, ongoing focusing of the response to EBV does not occur. In this
study, we report an analysis of the data and show that the extent of
down-regulation of the responses is related to the logarithm of the
number of cells that mediate the primary response. We suggest an
explanation for the findings based on the assumption that individual T
cell clones do not have an infinite life span and that clonal survival
may diminish with increasing turnover (14). We propose
that, if a T cell clone has divided extensively during the primary
immune response, many of the progeny will die, leaving relatively few
cells to contribute to the memory response (12). To test
this hypothesis, we developed models of T cell replication and death
(see Fig. 1
). The first, basic model is
adapted from mathematical models that have been used previously to
analyze CD8+ T cell responses to virus infection
(15, 16). Thus, we explicitly incorporate 1) variation in
avidity of different responding T cell clones and 2) a dependence of
cell division rate on the avidity of the cell for Ag and on the
abundance of Ag. We compare this basic model with a second model that
includes 1) an explicit accounting for cell division number and 2) a
decline in survival with increased cell division number. We apply these
models to our previously published data and to some new data and find
that it is the outcome of the second model that most closely resembles
the observed relationship between the primary and memory responses to
this virus. This result strongly implies that T cell senescence is an
important process that prevents excessive immune focusing over time,
and hence helps to mold optimal memory T cell responses.
|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Six patients positive for the HLA-A2 allele, four patients positive for the HLA-B8 allele, and one patient positive for both alleles were identified during the acute stage of infectious mononucleosis due to primary EBV infection. Samples of peripheral blood were taken from the patients immediately after diagnosis. In general, patients had had symptoms for a period of 1–3 wk before the diagnosis was made. Further samples of peripheral blood were taken 1 year later, at which time all the patients had made a complete clinical recovery from infectious mononucleosis. PBMC were separated using Lymphoprep (Axis-Shield, Oslo, Norway) density gradient centrifugation.
Staining of PBLs with HLA-peptide tetrameric complexes
Soluble PE-labeled HLA-peptide tetrameric complexes were produced as previously described (9). For the purposes of this study, we constructed tetramers of HLA-A2 complexed with the GLCTLVAML epitope from the EBV lytic protein BMLF1 (17), of HLA-B8 complexed with theRAKFKQLL epitope from the EBV lytic protein BZLF1 (18), and of HLA-B8 complexed with the FLRGRAYGL epitope from the EBV latent protein EBV-encoded nuclear Ag 3A (19). PBMC were incubated with 5 µl tetramer in solution at a concentration of 1 µmol/L on ice for 30 min. The cells were washed and stained with a tricolor-conjugated anti-CD8 mAb (Caltag Laboratories, Burlingame, CA) and were analyzed on a FACSCalibur using CellQuest software (BD Biosciences, Oxford, U.K.).
Analysis of TCR V chain usage by T cells specific for EBV epitopes
Samples of PBMC were stained with one of a panel of mAbs
specific for the human TCR V chains, washed, and then stained with a
FITC-conjugated anti-mouse or anti-rat mAb (DAKO, Carpinteria,
CA). Following this, the cells were washed, stained sequentially with
the relevant HLA-peptide tetrameric complex and anti-CD8 mAb, and
analyzed, as above. The V
2 mAb was found to interfere with the
binding of the HLA-A2/GLCTLVAML tetramer in some patients, and
therefore results obtained using this mAb were omitted from the
analysis of GLCTLVAML-specific T cells.
Estimation of the avidity and dissociation kinetics of the interaction between the HLA-B8/RAKFKQLL tetramer and the RAKFKQLL-specific CD8+ T cells
Samples of PBMC from blood taken during the primary and persistent phases of EBV infection were stained with saturating concentrations of PE-labeled HLA-B8/RAKFKQLL tetramer at 4°C. The PBMC were washed, and an aliquot of cells was removed for FACS analysis. The remaining PBMC were incubated with a 100-fold excess of unlabeled HLA-B8/RAKFKQLL tetramer at 4°C. This unlabeled reagent effectively blocked binding/rebinding of PE-labeled tetramer. Further aliquots of cells were removed and analyzed by FACS at seven time points. The total fluorescence within the PE-positive gate was plotted against time to give the dissociation curve. This total fluorescence was calculated as the sum of the fluorescence intensities of the tetramer-positive cells normalized per lymphocyte. This was then normalized to percentage of the total fluorescence at the initial time point and plotted on a logarithmic scale. An initial experiment was performed using an HLA-B8-restricted RAKFKQLL-specific T cell clone. Results confirmed that the plot of the logarithm of normalized fluorescence against time was linear and revealed that the t1/2 of dissociation was 46.8 min at 4°C.
To estimate apparent Kd values for the interaction between RAKFKQLL-specific T cells and the HLA-B8/RAKFKQLL tetramer, samples of PBMC were stained at room temperature with the HLA-B8/RAKFKQLL tetramer at a range of subsaturating concentrations. A graph of tetramer concentration plotted against bound tetramer (total fluorescence, calculated as above) was obtained and, in addition, Scatchard plots of bound tetramer (total fluorescence) vs bound tetramer divided by free tetramer were also drawn. The absolute values for total fluorescence were larger for PBMC from primary infection because of the higher frequency of Ag-specific T cells in these samples. This, however, does not affect the calculation of the apparent Kd. The tetramer was in vast excess, and therefore the concentration of free tetramer was taken as being equal to the concentration of the tetramer in the staining solution. The apparent Kd was derived by the nonlinear fitting method from the graph of tetramer concentration plotted against bound tetramer.
Mathematical modeling of the T cell response
The basic model of the T cell response to EBV is shown in Fig. 1
a and was adapted from published mathematical models of the
CD8+ T cell response to virus infection in vivo
(15, 16). It uses a number of ordinary differential
equations to simulate the T cell response to three viral epitopes
(denoted A, B, and C). The naive repertoire
(Ni) recognizing each epitope consists of
10 clones that vary in their avidity for Ag
(Ai). The avidity of each T cell is
assigned as a randomly generated number between 0 and 10. It is assumed
that T cells are stimulated by increasing doses of Ag and that, for any
given level of avidity, the level of stimulation increases with
increasing doses of Ag and is saturable. It is assumed that the
different viral epitopes (A – C) are processed and
presented by an infected cell with different efficiencies
(EA – EC).
This has the effect of producing an immunodominance hierarchy between
the epitopes (17). The level of stimulation of an
individual clone (Si) is calculated by:
 |
), the rate of proliferation of
effector cells (SiC), the rate of
reactivation of memory cells (Si
), and
the rate of death of effector cells ((1 -
Si)bE). Thus,
increasing levels of antigenic stimulation increase the number of
effector cells and decrease the death rate of effector cells.
Conversely, decreasing levels of antigenic stimulation decrease the
number of effector cells and increase the death rate of effector cells
(12, 20, 21, 22).
Therefore, the numbers of naive effector and memory cells are described
by the following differential equations:
 |
 |
 |
, consistent with the observation that differentiation to memory
cells appears to be stochastic and dependent on clonal burst size of
the effector population (5, 8). Finally, memory cell death
and division occur at a rate bM and
p, respectively. Importantly, this model allows for infinite
growth of individual T cell clones. Incorporation of a limited life span of T cell clones
To introduce the concept of replicative senescence into the
model, it is necessary to keep track of the number of cell divisions
that an individual T cell has undergone and then apply a maximum limit
to the life span of individual T cells. The model described above was
modified by the addition of different compartments of effector and
memory cells that reflect the number of cell divisions the cell has
undergone. The modified model is presented pictorially in Fig. 1b. The level of stimulation of T cells in this model is
dependent on both the level of Ag and the number of cell divisions
(d) the CTL have undergone. Thus, the responsiveness of T
cells to Ag declines as they senesce (as d increases), such
that at division cycle dsen the level of
antigenic stimulation for a given level of Ag has declined to half and
the level of effector death is half maximal. Thus, at low antigenic
stimulation, effector cells die rapidly regardless of how many cell
divisions they have undergone. When antigenic stimulation is maximal,
cells in early division cycles have low death rates until they
approach a high number of cell divisions
(dsen). The level of antigenic stimulation
is calculated by:
 |
Mathematical modeling of viral replication
A simple mathematical model of virus replication was developed,
based upon our understanding of the life cycle of EBV (Fig. 1c). Briefly, EBV virus infects target cells (X)
to create productively infected cells (Y) that divide
rapidly (at a rate "r") and produce free virus
(V) at a rate "k." Productively infected
cells may become latently infected (L) at a slow rate (
).
Latently infected cells may also reactivate to become productively
infected at a low rate (
). Productively infected cells express both
lytic and latent Ags. Latently infected cells in this context refer to
cells characterized by the latency I program of EBV gene expression,
whereby the expression of virtually all EBV genes is down-regulated and
the three Ags under consideration in this study are not expressed.
The numbers of uninfected, infected, latently infected cells, and free
virus are determined by the following equations:
 |
 |
 |
 |
), latently infected
cells (
), and free virus (µ) are constant. However, the death rate
of productively infected cells (
) is determined by two factors, the
natural death rate of infected cells (
) and the rate of T cell
killing. The overall level of cytotoxic T cell-mediated killing of the
virus is determined by the total number ofeffector cytotoxic T cells.
Therefore, the total death rate () of productively infected cells is
determined by:
 |
) plus the sum of all
effector T cells from each clone (i = 1–10) specific for each epitope
(A – C) times the rate of T cell killing (
). |
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We used fluorescent-labeled HLA-peptide tetramers to analyze the
frequency of CD8+ T cells specific for three
epitopes from EBV at two different time points (on diagnosis of primary
EBV infection and 1 year later) in HLA-A2+ and
HLA-B8+ patients. Consistent with previous
studies, we found a striking CD8+ T cell
lymphocytosis during primary infection. Within this expanded lymphocyte
population, we found populations of T cells that stained with the
HLA-B8/RAKFKQLL tetramer (4.1–40% CD8+ T cells)
and populations of T cells that stained with the HLA-A2/GLCTLVAML
tetramer (3–11% CD8+ T cells). We also found
small populations of CD8+ T cells specific for
the HLA-B8-restricted FLRGRAYGL epitope (up to 2.2%
CD8+ T cells) (12). The burst sizes
of these primary responses are striking, suggesting that large numbers
of naive T cells are recruited into the responses and/or that the cells
divide multiple times during the primary response (23, 24). The initial blood samples were taken as early as possible
during the primary immune response. Despite this, patients had had
symptoms for 1–3 wk at the time of sampling. Work in murine models of
infection suggests that the immune response may vary substantially over
a short period of time during a primary infection (25).
However, in two individuals we were able to analyze the T cell response
at two time points, 3 wk apart, during primary infection and found that
the magnitude of the responses were stable over this time period. One
year later, the absolute lymphocyte counts had fallen to normal levels,
but we were able to detect populations of CD8+ T
cells specific for the RAKFKQLL epitope (2.7–7.3%
CD8+ T cells), the GLCTLVAML epitope (0.5–5.5%
CD8+ T cells), and the FLRGRAYGL epitope (up to
1.8% CD8+ T cells). In terms of absolute numbers
of cells responding to a given epitope, the size of the response was
always larger at the time of primary infection than it was 1 year
later. Thus, the primary clonal burst was followed by a period of
decay, leaving a smaller population of memory cells (24).
In terms of the frequency of T cells specific for a given epitope as a
proportion of the CD8+ T cell population, we
occasionally found that the contribution of T cells specific for a
given epitope increased at 1 year. That is, the memory response was not
simply a reduced version of the primary response, as has been reported
in studies from other groups. Such a result would have suggested that
selection of the memory pool of T cells specific for a range of
epitopes was simply stochastic. We refer to this as the stochastic
entry into memory hypothesis. Instead, we found that very large
responses tended to decrease more than would be expected by the
stochastic entry into memory model, and many smaller responses
increased more than expected. We hypothesized that this may be related
to senescence of the responding T cells, and therefore related to the
number of cell divisions they had undergone (proportional to the
logarithm of the cell number). Comparing the relationship predicted by
the stochastic entry into memory hypothesis (memory = M
x primary) with a relationship that allows for senescence
(memory = primary x (K - B
log2(primary))), we found that the latter model
fitted the data better (F test, p = 0.0056;
R2 = 0.43). We then quantified the
percentage of change in contribution of T cells specific for a given
epitope to the CD8+ T cell repertoire as follows:
 |
 |
 |
Contrary to the prediction of the stochastic entry into memory
hypothesis that the memory response would be a fixed proportion of the
primary response, we found a range from -91 to +250%, and that a
higher number of cell divisions during the primary immune response is
associated with a larger decrease in representation in the memory
response (Fig. 2a).
|
The longitudinal analysis of T cell responses to different epitopes from a persistent virus may be complicated by differences in levels of expression of different viral proteins over time. This may be particularly relevant to a comparison of responses to epitopes from latent vs lytic cycle proteins. We therefore analyzed the T cell responses to the two lytic epitopes in more detail, focusing on the dynamics of the clones that contributed to these responses.
We combined staining with the HLA-A2/GLCTLVAML and HLA-B8/RAKFKQLL
tetramers with staining using a panel of Abs specific for TCR V
chains and with RT-PCR to analyze the TCR use of the Ag-specific T
cells. The analysis was performed on samples of peripheral blood
taken as early as possible during primary infection and again on the
samples taken 1 year subsequently. In two individuals, we were able to
obtain samples of blood on two occasions, 3 wk apart, during primary
infection. In both patients, the TCR repertoire of the Ag-specific T
cells was stable over this time period. However, as previously
reported, we found that the clonal and oligoclonal populations of
CD8+ T cells that dominated the response during
primary infection were often relatively underrepresented in memory. We
quantitated the percentage of change in contribution of a particular
V chain to an epitope-specific response from primary infection to 1
year as follows:
|
|
 |
 |
We analyzed the relationship between the percentage of change and
the number of epitope-specific T cells that expressed that V chain
during the primary response. The relationship is more complex than that
observed for the overall epitope-specific populations. The results
again suggested that clones that had proliferated most extensively were
less likely to contribute to memory (Fig. 2b). Again, if
selection into the memory pool had been purely stochastic, then the
relationship between percentage of change and
log2 (cell number of epitope-specific cells using
that V in primary infection) would have been depicted by a
horizontal line (the stochastic entry into memory hypothesis).
Alternatively, if selection into the memory pool had been characterized
by ongoing affinity maturation, there would have been a positive
correlation between the parameters. Interestingly, the inverse
relationship between percentage of change and
log2 (cell number of epitope-specific cells using
that V in primary infection) did not hold true when an analysis of
V chains that were used by very small numbers of tetramer-reactive T
cells was undertaken. These cells appeared to contribute less to the
memory response than a simple semilogarithmic relationship would
predict. This observation suggests that the clones that expand little
during the primary response are not subject to the same constraints as
those that expand most extensively. This point is discussed in further
detail below.
Longitudinal analysis of the avidity and kinetics of tetramer binding to Ag-specific T cells
One explanation for a change in the TCR repertoire of Ag-specific
T cells over time is that ongoing selection occurs for T cells
expressing receptors with optimal kinetics of interaction with the
MHC/peptide ligand (6). In two
HLA-B8+ individuals, we analyzed the avidity and
the dissociation kinetics of the interaction between the T cells and
the HLA-B8/RAKFKQLL tetramer at the two time points sampled. We found a
small increase in the apparent Kd of
the T cells for the tetramer (Kd = 114
nmol in primary and Kd = 134 nmol 1
year subsequently in one donor, and Kd
= 110 nmol in primary and Kd = 120
nmol 1 year subsequently in the second donor) (Fig. 3, a and b, and
data not shown). This small rise in Kd
equates to a small fall in affinity, although it is not statistically
significant at a p value of <0.05. The
Kd were of a similar order of
magnitude to those that have been described for the interaction between
tetramers of MCC/I-Ek and the 2B4 TCR
(Kd = 60 ± 9 nmol)
(6). The affinity of the 2B4 TCR for MCC/I-Ek has
previously been estimated as 40–90 µmol at 25 degrees (26, 27). Thus, the estimates of apparent
Kd we obtained are consistent with a
relatively high affinity interaction, supporting the idea that, in this
viral infection, selection on the basis of affinity has already
occurred by the time of sampling of the primary response. Furthermore,
the dissociation kinetics of the interaction between the T cells and
the tetramer was also found to be very similar when T cells mediating
primary vs memory responses were analyzed (Fig. 3c and data
not shown). These results argue strongly against affinity maturation as
an explanation for the observed changes in the clonal composition of
the EBV-specific T cell response over time.
|
We went on to use simple mathematical models to simulate the
response to infection in 20 patients. In the basic model (Fig. 1a), individuals differed by the random generation of
avidities for each T cell clone (28). This resulted in
variations in peak viremia and in dominance of individual T cell clones
during the response. The data for primary infection were taken on the
day of peak viral load, and those for chronic infection at day 300 in
the model. This basic model produced an Ag-specific T cell repertoire
in acute infection that was dominated by the highest avidity clones
(Fig. 4a). High-avidity T
cells were stimulated earlier in the response, because as the viral
load gradually increased, they were sensitive to lower levels of virus.
At the peak of the response, the high viral load stimulated both low
and high-avidity T cells, although high-avidity T cells continued to
dominate. During progression to chronic infection, low avidity clones
were greatly reduced in number due to the lower viral loads. Over time,
the repertoire became increasingly focused, with the high-avidity
clones that dominated in acute infection coming to dominate the
response to chronic infection. The picture clearly differs from that
observed in the context of natural EBV infection.
|
b and 4b). Although this did not
significantly alter the outcome of primary infection, it did change the
predicted outcome of chronic infection. Excessive expansion during the
primary response led to senescence of high-avidity T cells, resulting
in extensive down-regulation of these populations (Fig. 5a). The model predicts that T
cells that were greatly expanded during the primary response will be
preferentially down-regulated in the response at 1 year, and that
smaller clones will be more likely to expand and replace them. We
further analyzed the model output for the contribution of individual T
cell clones to an epitope-specific response (Fig. 5b).
Interestingly, the relationship between the log2
(cell number) of individual clones and the percentage of change is
predicted to be complex and again closely resembles that found by
experiment. Clones that expand dramatically during primary infection
are more prone to die, whereas those that expand less dramatically may
contribute relatively more to the response at 1 year. However, clones
that are very poorly represented in the primary response are often
actually further underrepresented at 1 year. This would appear at first
to be counterintuitive and contrary to the previous observations that
the largest clones were most heavily down-regulated. However, the model
predicts that at peak viremia, low avidity clones receive some
stimulation and proliferate. During chronic infection, the lower levels
of virus are insufficient to stimulate these clones, and thus they
decline to very low levels. Therefore, the mathematical model provides
an insight into why the experimental results suggest that both the
largest and the smallest clones in the primary response will be
relatively underrepresented in memory.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our analysis was, inevitably, confined to the Ag-specific CD8+ T cell populations within peripheral blood. Both primary and memory populations of T cells may traffic to different sites within the body and may be present at differing frequencies in spleen, lymph nodes, liver, bone marrow, and other tissues (29). Experiments performed in murine models of infection do show that the overall fall in numbers of Ag-specific T cells following the peak primary response may be less dramatic when all these other sites are also taken into account (29). Nevertheless, differences in patterns of recruitment seem unlikely to account for the differences in the extent of down-regulation of different populations of T cells that we have observed.
Changes in relative levels of expression of different viral proteins during the primary and persistent phases of infection may be one factor that influences shifts in immunodominance between primary and memory T cell responses. However, the observation that there are also changes in the clonal composition of the responses to individual epitopes suggests that other factors are likely to be involved in the evolution of T cell memory. In mouse models of primary and secondary responses, one such factor is ongoing affinity maturation (2, 3, 4, 6, 7). We have not found evidence for this in our study, and indeed we find that T cells have a relatively high avidity for the relevant tetrameric HLA-peptide complex at the time of first sampling during the primary response. The extent of exposure to Ag during primary EBV infection is likely to be high and more prolonged than during a single immunization, and hence the process of affinity maturation is likely to be advanced in our patients when they first donate blood. The differences observed between primary and memory responses must therefore reflect another biological process.
Statistical analysis revealed that it was the logarithm of the number
of responding cells during the primary response that correlated with
the extent of down-regulation of the response. Such a relationship
provided a better fit to our data than a simple stochastic entry into
memory model, in which the memory response would be linearly related to
the primary response (p = 0.0056, Fig. 2a). An intuitive explanation for the findings was that T
cells that have proliferated most extensively are more prone to die and
are therefore less likely to be represented within the memory pool. We
used computer models of T cell proliferation and death to test this
idea formally. The initial model was adapted from conventional models
of T cell proliferation and did not include the possibility of clonal
senescence. Analysis of the numbers of cells responding to different
epitopes in a simulated prime-boost scenario reproduced the enhanced
immunodominance or epitope focusing seen in murine models of the
response (2, 3, 4, 6, 7). However, analysis of the response
at later time points in a simulated chronic infection showed a
progressive focusing of the response. The model was then modified to
incorporate the concept of cellular senescence. This required keeping
track of cell division number in both the memory and effector
compartment for each clone. The model was able to reproduce the
observed effects of prime-boost in mice. However, the large Ag load and
chronic Ag stimulation produced by simulation of EBV infection led to a
different outcome from the previous model. The simple inclusion of a
factor of cellular senescence in a model of affinity-driven clonal
selection was sufficient to reproduce the relationship observed in
natural infection. Furthermore, in the presence of very high and
sustained viral loads, this model predicts that all responding T cells
may senesce (data not shown). This phenomenon of T cell exhaustion has
been observed in the context of high dose lymphocytic choriomeningitis
virus infection of mice (30, 31, 32).
We propose that the situation in natural EBV infection lies somewhere between the two extremes of acute low dose challenge (in which cells do not experience senescence phenomena because they do not undergo sufficient expansion) and immune exhaustion (in which responsiveness is lost due to senescence of all responding CTL) that have been reported in mouse models. Thus, EBV infection results in senescence and loss of some of the T cells that respond most vigorously in acute infection. However, T cells that have expanded less vigorously do not senesce and are able to maintain the response to virus. Because the term "exhaustion" has been associated with the failure of senescent CTL to be replaced (30, 31, 32), we prefer the term "clonal succession" to describe the sequential changes in dominance of the responding clones.
The mechanisms underlying senescence remain to be explored in more detail. They may involve an inability of the expanded populations of cells to modulate expression of survival factors (12, 33) or they may reflect other constraints such as telomere shortening.
Our experiments are performed in the setting of a natural infection in which Ag load is likely to be high during the primary phase of infection and very much lower during the persistent phase of infection. We are not able to perform experiments that address the question as to what would happen if further high dose challenge with Ag occurred in the context of EBV infection. Work on primary, memory, and secondary responses to influenza in mice has suggested that, while the frequency of the response that was immunodominant during primary infection fell to below that of a subdominant response in the resting memory phase, on secondary challenge the original hierarchy of immunodominance was reestablished (29).
It will, in the future, be interesting to analyze the response to different infections using similar methods to those described in this study and to ask whether cellular senescence might also constrain the response to other persistent virus infections such as HIV or to nonpersistent, but recurrent virus infections such as influenza. Certainly, any immune stimulus that can cause cells to undergo many rounds of cell division may be capable of causing immune senescence.
It is interesting to consider teleological arguments as to why T cell senescence may be beneficial to the host, despite the observed effect of the loss of high responder T cell clones. Cellular senescence is often thought of as a protection mechanism against neoplastic proliferation of cells. T cell senescence may also provide some protection from host death due to overwhelming infection and an excessive immune response or due to autoimmunity (32, 34). In this study, we suggest that cellular senescence within the cells of the immune system is an important adaptation that allows for a long-term effective T cell response to a persistent Ag. In the absence of senescence, modeling predicts that the T cell response to a pathogen would become progressively more focused, both at the level of the epitopes recognized and at the level of the clones responding to those epitopes. Ultimately, in the case of a persistent infection, the host response would depend on a single expanded clone recognizing a single epitope. Such a focused response renders the host extremely vulnerable to immunological escape by the pathogen. This study shows how T cell senescence prevents such extreme focusing and allows the immune response to remain relatively broad. A broad immune response comprising many different T cell clones recognizing several different epitopes renders the host less susceptible to virus escape mechanisms, and therefore represents a better biological response to a persistent infection.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Margaret F. C. Callan, Institute of Molecular Medicine, John Radcliffe Hospital, Oxford, OX4 1DQ, U.K. E-mail address: mcallan{at}molbiol.ox.ac.uk
Received for publication September 28, 2001. Accepted for publication February 1, 2002.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
diversity during the initiation of a CD4+ T cell response in vivo. J. Exp. Med. 192:1719.This article has been cited by other articles:
|
|
 |
|
  D. A. Price, T. E. Asher, N. A. Wilson, M. C. Nason, J. M. Brenchley, I. S. Metzler, V. Venturi, E. Gostick, P. K. Chattopadhyay, M. Roederer, et al. Public clonotype usage identifies protective Gag-specific CD8+ T cell responses in SIV infection J. Exp. Med., April 13, 2009; 206(4): 923 - 936. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Venturi, H. Y. Chin, T. E. Asher, K. Ladell, P. Scheinberg, E. Bornstein, D. van Bockel, A. D. Kelleher, D. C. Douek, D. A. Price, et al. TCR {beta}-Chain Sharing in Human CD8+ T Cell Responses to Cytomegalovirus and EBV J. Immunol., December 1, 2008; 181(11): 7853 - 7862. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Yerly, D. Heckerman, T. M. Allen, J. V. Chisholm III, K. Faircloth, C. H. Linde, N. Frahm, J. Timm, W. J. Pichler, A. Cerny, et al. Increased Cytotoxic T-Lymphocyte Epitope Variant Cross-Recognition and Functional Avidity Are Associated with Hepatitis C Virus Clearance J. Virol., March 15, 2008; 82(6): 3147 - 3153. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. A. Price, A. D. Bitmansour, J. B. Edgar, J. M. Walker, M. K. Axthelm, D. C. Douek, and L. J. Picker Induction and Evolution of Cytomegalovirus-Specific CD4+ T Cell Clonotypes in Rhesus Macaques J. Immunol., January 1, 2008; 180(1): 269 - 280. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. R. Almeida, D. A. Price, L. Papagno, Z. A. Arkoub, D. Sauce, E. Bornstein, T. E. Asher, A. Samri, A. Schnuriger, I. Theodorou, et al. Superior control of HIV-1 replication by CD8+ T cells is reflected by their avidity, polyfunctionality, and clonal turnover J. Exp. Med., October 1, 2007; 204(10): 2473 - 2485. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. K. Day, A. J. Carmichael, I. J. M. ten Berge, E. C. P. Waller, J. G. P. Sissons, and M. R. Wills Rapid CD8+ T Cell Repertoire Focusing and Selection of High-Affinity Clones into Memory Following Primary Infection with a Persistent Human Virus: Human Cytomegalovirus J. Immunol., September 1, 2007; 179(5): 3203 - 3213. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. S. Cush, K. M. Anderson, D. H. Ravneberg, J. L. Weslow-Schmidt, and E. Flano Memory Generation and Maintenance of CD8+ T Cell Function during Viral Persistence J. Immunol., July 1, 2007; 179(1): 141 - 153. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. N. Akbar and M. Vukmanovic-Stejic Telomerase in T Lymphocytes: Use It and Lose It? J. Immunol., June 1, 2007; 178(11): 6689 - 6694. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Castiglione, K. Duca, A. Jarrah, R. Laubenbacher, D. Hochberg, and D. Thorley-Lawson Simulating Epstein-Barr virus infection with C-ImmSim Bioinformatics, June 1, 2007; 23(11): 1371 - 1377. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Lichterfeld, X. G. Yu, S. K. Mui, K. L. Williams, A. Trocha, M. A. Brockman, R. L. Allgaier, M. T. Waring, T. Koibuchi, M. N. Johnston, et al. Selective Depletion of High-Avidity Human Immunodeficiency Virus Type 1 (HIV-1)-Specific CD8+ T Cells after Early HIV-1 Infection J. Virol., April 15, 2007; 81(8): 4199 - 4214. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. M. M. van Leeuwen, E. B. M. Remmerswaal, M. H. M. Heemskerk, I. J. M. ten Berge, and R. A. W. van Lier Strong selection of virus-specific cytotoxic CD4+ T-cell clones during primary human cytomegalovirus infection Blood, November 1, 2006; 108(9): 3121 - 3127. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Singh and Y. Paterson Vaccination Strategy Determines the Emergence and Dominance of CD8+ T-Cell Epitopes in a FVB/N Rat HER-2/neu Mouse Model of Breast Cancer. Cancer Res., August 1, 2006; 66(15): 7748 - 7757. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Fletcher, M. Vukmanovic-Stejic, P. J. Dunne, K. E. Birch, J. E. Cook, S. E. Jackson, M. Salmon, M. H. Rustin, and A. N. Akbar Cytomegalovirus-Specific CD4+ T Cells in Healthy Carriers Are Continuously Driven to Replicative Exhaustion J. Immunol., December 15, 2005; 175(12): 8218 - 8225. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. A. Price, J. M. Brenchley, L. E. Ruff, M. R. Betts, B. J. Hill, M. Roederer, R. A. Koup, S. A. Migueles, E. Gostick, L. Wooldridge, et al. Avidity for antigen shapes clonal dominance in CD8+ T cell populations specific for persistent DNA viruses J. Exp. Med., November 21, 2005; 202(10): 1349 - 1361. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. E. Tebo, M. J. Fuller, D. E. Gaddis, K. Kojima, K. Rehani, and A. J. Zajac Rapid Recruitment of Virus-Specific CD8 T Cells Restructures Immunodominance during Protective Secondary Responses J. Virol., October 15, 2005; 79(20): 12703 - 12713. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. E. Miller, J. R. Bonczyk, Y. Nakayama, and M. Suresh Role of Thymic Output in Regulating CD8 T-Cell Homeostasis during Acute and Chronic Viral Infection J. Virol., August 1, 2005; 79(15): 9419 - 9429. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Chen, J. Hou, X. Jiang, S. Ma, M. Meng, B. Wang, M. Zhang, M. Zhang, X. Tang, F. Zhang, et al. Response of Memory CD8+ T Cells to Severe Acute Respiratory Syndrome (SARS) Coronavirus in Recovered SARS Patients and Healthy Individuals J. Immunol., July 1, 2005; 175(1): 591 - 598. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. V. D. Soares, F. J. Plunkett, C. S. Verbeke, J. E. Cook, J. M. Faint, L. L. Belaramani, J. M. Fletcher, N. Hammerschmitt, M. Rustin, W. Bergler, et al. Integration of apoptosis and telomere erosion in virus-specific CD8+ T cells from blood and tonsils during primary infection Blood, January 1, 2004; 103(1): 162 - 167. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. M.-L. Choi, J.-L. Chen, L. Wooldridge, M. Salio, A. Lissina, N. Lissin, I. F. Hermans, J. D. Silk, F. Mirza, M. J. Palmowski, et al. High Avidity Antigen-Specific CTL Identified by CD8-Independent Tetramer Staining J. Immunol., November 15, 2003; 171(10): 5116 - 5123. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. N. Naumov, E. N. Naumova, K. T. Hogan, L. K. Selin, and J. Gorski A Fractal Clonotype Distribution in the CD8+ Memory T Cell Repertoire Could Optimize Potential for Immune Responses J. Immunol., April 15, 2003; 170(8): 3994 - 4001. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Fuller and A. J. Zajac Ablation of CD8 and CD4 T Cell Responses by High Viral Loads J. Immunol., January 1, 2003; 170(1): 477 - 486. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
= 2 x
105,