| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Immunology, University of Toronto, Toronto, Ontario, Canada
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials MethodsResultsDiscussionReferences |
|---|
48–72 h after in vitro T cell activation
(reviewed in Ref. 7). Several members of the TNFR family,
including 4-1BB(CD137), OX40 (CD134), and CD27 have been shown to play
a role in T cell activation subsequent to initial activation events
(8, 9, 10). Given the recent evidence that once activated, T
cells continue their activation program through several divisions in
the absence of a further antigenic signal, the question arises as to
when and how these activation-induced costimulatory receptors come
into play. 4-1BB is a costimulatory member of the TNFR family expressed on activated T cells (8). Upon engagement with 4-1BB ligand (4-1BBL)3 or aggregation with anti-4-1BB, 4-1BB can provide a CD28-independent costimulatory signal leading to CD4 and CD8 T cell expansion, cytokine production, development of CTL effector function, and prevention of activation-induced death (11, 12, 13, 14, 15, 16, 17, 18). In vitro, resting CD28-/- T cells can up-regulate 4-1BB quickly enough to respond to 4-1BBL within 24 h of anti-CD3 treatment (15). In mixed lymphocyte reactions, 4-1BBL was found to increase T cell numbers, particularly at days 5–7 of stimulation, consistent with a role for 4-1BBL in T cell survival (18). 4-1BBL (19) is expressed on activated APC only after 2–3 days of activation and is likely the limiting factor in 4-1BB-mediated costimulatory responses (11). Thus, although 4-1BB can provide a CD28-independent costimulatory signal for resting T cells in vitro, in vivo it is likely to act later in the response. Initial analysis of 4-1BBL-/- mice showed that 4-1BBL plays a role in augmenting suboptimal anti-viral CTL responses, skin allograft rejection as well as MHC I- and MHC II-restricted graft-vs-host disease (16, 20, 21, 22). For influenza-specific responses, 4-1BBL-/- mice showed a partial defect in the in vitro secondary CTL response to influenza virus. In the case of lymphocytic choriomeningitis virus (LCMV), there was no defect in viral clearance or in the primary T cell response in 4-1BBL-/- mice in one study (16). However, in another study, it was found that there was a 2-fold decrease in Ag-specific T cell numbers, measured at day 8 after infection (21). Under conditions of suboptimal priming using a lipidated LCMV peptide, 4-1BBL-/-, but not wild-type, mice succumbed to challenge with a more virulent strain of LCMV (21). In vitro studies using LCMV-specific P14 transgenic T cells showed 4-1BBL-dependence of the LCMV-specific response only when a weak agonist peptide was used (16). Thus, there is accumulating evidence that 4-1BBL can influence both primary and secondary virus-specific CTL responses.
To pinpoint the time at which 4-1BBL plays a role during the response to influenza infection, in this report we have followed specific T cell expansion in wild-type, CD28-/-, and 4-1BBL-/- mice following primary and secondary infection. The results presented show that CD28 is critical for primary expansion of influenza-specific CD8 T cells, whereas 4-1BBL has an impact only late in the primary response, thereby influencing the size of the memory pool and the subsequent secondary CD8 T cell response.
|
Materials Methods |
|---|
|
TopAbstractIntroductionMaterials MethodsResultsDiscussionReferences |
|---|
C57BL/6 mice were bred in our facility from breeder pairs obtained from Charles River Breeding Laboratories (St.-Constant, Quebec, Canada). 4-1BBL-/- mice were also bred in our facility after backcrossing with C57BL/6 breeders (n = 6 backcrosses) and being screened as previously described (16). Following the final backcross, 4-1BBL-/- mice, as well as wild-type siblings from the same F2 cross, were used. No differences were observed between C57BL/6 mice from Charles River Breeding Laboratories and 4-1BBL+/+ litter mates bred in our facility, with respect to the number of tetramer-positive T cells observed 7 days after influenza infection (data not shown). CD28-/- mice backcrossed onto the C57BL/6 background were provided by Dr. T. Mak (Amgen Institute, Toronto, Canada) and were bred in our facility.
Influenza virus infection
Most studies of influenza virus infection of mice have used intranasal infection of naive mice or mice primed i.p. with virus. Low numbers of Db/NP366-374-specific CD8 T cells are detected in the spleen and bronchoalveolar lavage of mice during primary intranasal HKx31 infections, with the peak of the response at day 8 (23). Pilot experiments using i.p. infection of C57BL/6 mice with influenza A HKx31 of wild-type mice resulted in highly reproducible numbers of Db/NP366-374-specific CD8+ T cells responding in the spleen, with the response peaking at day 7 after primary infection (average, 7% of CD8+ T cells). This allowed us to assess subtle differences in the 4-IBBL-/- vs wild-type mice during the primary response to influenza infection. Therefore, we continued with i.p. infections throughout the study. Seven- to 10-wk-old mice were infected with 200 hemagglutinin units (HAU) of influenza A HKx31 (H3N2) produced as described (24). This virus does not replicate extensively when given to mice i.p. At 3 wk postinfection, some mice were challenged with the same strain (HKx31) or with the serologically distinct A/PR8/34 (PR8, H1N1) which shares the NP gene with HKx31. Mice were sacrificed at the indicated time points and their spleens were harvested for single cell suspensions. Although the CD28-/- mice showed a greatly reduced NP366-374-specific CD8 T cell response to influenza virus, none of the mice showed any obvious ill effects of virus infection.
Flow cytometry
Tetramer staining.
Spleen cell suspensions were prepared in PBS/2% FCS/0.01% sodium
azide on ice. Cells were surface-stained with PE-conjugated
anti-mouse CD8
, FITC-conjugated anti-mouse CD62L
(eBioscience, San Diego, CA), and APC-labeled tetramers consisting of
murine class I MHC molecule H-2Db,
2-microglobulin, and influenza nucleoprotein
peptide, NP366-374 (National Institute of Allergy and Infectious
Disease MHC Tetramer Core Facility, Atlanta, GA). For each
experiment, appropriate isotype control mAbs were used.
Intracellular IFN-
staining.
Spleen cell suspensions were restimulated in culture medium (RPMI/10%
FCS with antibiotics and 2-ME) for 6 h at 37°C with 1 µM of
NP366-374 peptide and GolgiStop (BD PharMingen, San Diego, CA).
Cells were harvested, resuspended in PBS/2% FCS/azide, and
surface-stained with PE-anti-CD8 and FITC-anti-CD62L as
described above. Following surface-staining, cells were fixed in
Cytofix/Cytoperm solution (BD PharMingen) and then stained with
APC-conjugated anti-mouse IFN- diluted in 1x perm/wash solution
(BD PharMingen). Samples were analyzed using a FACSCaliber and
CellQuest software (BD Biosciences, Mountain View, CA).
Cytotoxicity assay
Wild-type siblings, 4-IBBL-/-, and CD28-/- mice were infected with 200 HAU influenza A HKx31 as described above. Splenocytes were harvested after 3 wk and restimulated in vitro by the addition of 100 nm of the H-2Db-restricted peptide NP366-374 to 1-ml cultures containing 5 x 106 spleen cells. On day 5, cells were resuspended to 0.5 ml and serial 3-fold dilutions of effectors were performed (referred to as dilution of standard culture) and assayed for anti-influenza NP-specific CTL activity against 51Cr-labeled EL4 cells pulsed with 50 µM NP366-374 peptide. After 5 h, 70 µl of supernatant was harvested onto 96-well harvest plates (Canberra Packard, Mississauga, Ontario, Canada) and counted on a Topcount scintillation counter (Canberra Packard). Maximum and spontaneous release was determined from wells that contained 1% SDS or medium alone. Spontaneous release was <10% for each assay. The percentage of specific lysis was calculated from the equation: (experimental 51Cr release - spontaneous 51Cr release)/(maximum 51Cr release - spontaneous 51Cr release) x 100% = % specific lysis.
Detection of influenza-specific Abs
Neutralizing Abs were measured by serial 2-fold dilutions of serum with 500 HAU/ml influenza HKx31 in 96-well round-bottom plates at 37°C for 45 min and compared with normal mouse serum. Samples were then added to 0.5% washed chicken red blood cells in 96-well round bottom plates. The titer of neutralizing Ab was the last dilution which blocked hemagglutination. Results are presented as the number of 2-fold dilutions above normal mouse serum which block hemagglutination.
Influenza A-specific Abs were determined for IgM, IgG1, and IgG2a isotypes. Each well of the 96-well plates were coated with 100 µl of 500 HAU/ml influenza A HKx31 in PBS for 1 h at 37°C followed by incubation at 4°C overnight. Plates were blocked with 5% skim milk in PBS/0.1% Tween 20 for 2 h at 37°C. Five-fold serial dilutions of serum in PBS/0.1% Tween 20/5% skim milk were added to wells overnight at 4°C. After washing in PBS/0.1% Tween 20, HRP-conjugated anti-isotype Abs (Caltag Laboratories, Burlingame, CA) were added for 2 h at 37°C. Following washing, H2O2 and ABTS (Sigma-Aldrich, St. Louis, MO) were added in citrate phosphate buffer (pH 5.0) and color development was measured after 20 min at OD405.
Cytokine assays
Spleen cells (5 x 106 cells) were
incubated with 250 HAU/ml of heat-killed (56°C, 30 min) influenza A
Hkx31 for 48 h. Supernatants were taken and the levels of IL-2 and
IFN- were measured. IL-2 was detected using the indicator cell line
CTLL ligand as described (12). Serial dilutions of
culture supernatant were prepared in duplicate and incubated with
1 x 104 indicator cells in 100 µl in
96-well plates for 24 h. During the final 6 h, the cells were
labeled with [3H]thymidine (Amersham, Baie
d’Urfe, Quebec, Canada). Cultures were harvested and analyzed on the
Topcount 96-well liquid scintillation counter (Canberra Packard). ELISA
was performed on diluted supernatants from cultures using pairs of
anti-murine IFN- mAbs purchased from BD PharMingen according to
the manufacturer’s instructions.
|
Results |
|---|
|
TopAbstractIntroductionMaterials MethodsResultsDiscussionReferences |
|---|
The immune response of mice infected with influenza HKx31 has been
well studied (23, 25). The most important immune effectors
during the primary response are the virus-specific CD8 T cells, with
the NP366-374 epitope being immunodominant in C57BL/6 mice
(26). To assess the role of the 4-1BB/4-1BBL vs CD28/B7
costimulatory pathways during the immune response to influenza,
wild-type siblings, 4-1BBL-/-, or
CD28-/- mice were infected i.p. with 200 HAU of
influenza virus A strain HKx31. Mice were sacrificed at different times
following infection and spleen cells were analyzed for the number of
Ag-specific cells using fluorescent-labeled
Db/NP366-374 tetramers. Initial experiments with
heterozygous 4-1BBL+/- mice showed no difference
from their wild-type littermates, so subsequent experiments were done
only with +/+ or -/- mice. The primary expansion of
Db/NP366-374-specific CD8 T cells was readily
detected in wild-type mice following infection with influenza HKx31,
with a peak response at 7 days postinfection (Fig. 1
A). The responding CD8 T
cells were CD62Llow, consistent with cells being
of the activated/memory phenotype (Fig. 1C).
4-1BBL-/- mice showed a similar rate and
magnitude of expansion of Db/NP366-374-specific
CD8 T cells during the primary response to influenza virus, indicating
that there is no defect in initial T cell expansion in the absence of
4-1BB/4-1BBL interaction. The results in Fig. 1 are expressed as the
percentage of CD8 T cells staining with tetramer. Conversion of these
numbers to total tetramer-positive T cells per spleen, based on cell
recoveries from spleen, gave identical results. There was no overall
change in CD8 T cell recoveries from infected vs wild-type mice, likely
due to the fact that influenza does not replicate extensively following
i.p. infection of mice. Seven percent of CD8 T cells at day 7
corresponded to 106 cells in the spleen for both
wild-type and 4-1BBL-/- mice. In contrast,
CD28-/- mice responded poorly to influenza
virus, showing a substantial decrease in the number of
Db/NP366-374-specific CD8 T cells detected at
days 5 and 7 following infection. At day 7, 0.8% of CD8 T cells
corresponded to 1.2 x 105 cells/spleen.
|
A were also analyzed by
intracellular staining for the production of IFN-, after a 6 h
restimulation with NP366-374 peptide (Fig. 1B). Although the
number of IFN--producing CD8 T cells detected was consistently less
than the number of tetramer-positive cells observed at each time point,
the results were qualitatively similar to those obtained with
Db/NP366-374 tetramer staining. Thus,
4-1BBL-/- mice show no defect in the number of
IFN--producing NP366-374-specific CD8 T cells following primary
influenza infection. As was seen using tetramer staining,
CD28-/- mice also show a large decrease in the
number of NP366-374-specific IFN--secreting CD8 T cells following
influenza infection.
By day 21 postinfection, the number of
Db/NP366-374-specific CD8 T cells declines to
1.5% of the CD8 T cell population which corresponded to 2.2 x
105 cells/spleen in wild-type mice (Fig. 2A). This population of
virus-specific CD8 T cells is thought to represent the pool of
"memory" T cells that have avoided the cell death that follows the
initial wave of T cell expansion. Although the differences detected
were small, we consistently observed one-third lower numbers of
Db/NP366-374-specific CD8 T cells in
4-1BBL-/- mice which had declined to 1.0% of
the CD8 T cell population or 1.4 x 105
cells/spleen. 4-1BBL-/- mice also had slightly
lower numbers of IFN--producing cells compared with wild-type mice
(0.81 ± 0.18% vs 1.27 ± 0.22%), consistent with the
difference seen in tetramer staining at that time point (Fig. 2B).
|
, A and C). Comparing the results at day 38 with
those at day 21, it can be seen that in wild-type mice, the T cell
numbers remain higher than those seen in
4-IBBL-/- mice (1.04 vs 0.59%). Although
small, the differences between wild-type and
4-IBBL-/- mice at days 21 and 38 were found to
be statistically significant (p < 0.01). These
data suggest that 4-1BBL influences the size of the memory T cell pool
late in the primary response, but has no apparent effect on initial T
cell expansion. Role of 4-IBBL vs CD28 on CTL responses following in vitro restimulation with NP366-374
To correlate the number of
Db/NP366-374-specific T cells observed in vivo
with the acquisition of CTL effector function, as measured in vitro,
splenocytes from mice infected 21 days previously with influenza virus
HKx31 were restimulated with NP366-374 for 5 days. At the end of
restimulation, splenocytes were assayed for their ability to lyse
NP366-374-labeled syngeneic target cells (Fig. 3A). At limiting E:T ratios,
3-fold more CTL effectors were required from the
4-IBBL-/- mice to give equivalent killing
compared with effectors from wild-type litter mates. As previously
observed (16), CD28-/- mice were
severely impaired in their recall CTL response to influenza virus and
this correlated with the failure to expand CD8+ T
cells in the primary response, resulting in a limited in vitro
secondary response. At the same time as the CTL assay, the number of
Db/NP366-374-specific tetramer-positive and
IFN--producing CD8+ T cells were measured
(Fig. 3, B and C). Restimulated cultures from
influenza-infected 4-IBBL-/- mice had
approximately two-thirds the number of tetramer-positive or IFN-
producing CD8+ T cells as observed in cultures
from wild-type mice. Thus, although there had been a large expansion of
Db/NP366-374-specific cells in cultures from
influenza-infected wild-type and 4-IBBL-/-
mice, the relative defect observed at day 21 following primary
infection is maintained following in vitro expansion of splenocytes
from the mice.
|
To monitor the effect of 4-IBB/4-1BBL vs CD28/B7-mediated
costimulation during the in vivo secondary response to influenza virus,
mice which had been infected with influenza A strain HKx31 21 days
previously were given a second dose of influenza A. T cell expansion
was measured at days 5 and 7 postinfection using
Db/NP366-374 tetramer staining of splenocytes
directly ex vivo (Fig. 4A) or
by intracellular staining for IFN- following restimulation with
peptide (Fig. 4B).
|
-producing CD8 T cells at days 5 and 7 postinfection
(Fig. 4B). The expansion of
Db/NP366-374-specific CD8 T cells upon secondary
infection was more rapid during the secondary response as indicated by
the greater number of tetramer-binding cells at the 5-day time point.
In contrast to the results observed in the primary response, during the
secondary response to influenza, 4-1BBL-/- mice
showed a substantial decrease in the number of
Db/NP366-374-specific T cells at days 5 and 7
postsecondary infection. Wild-type mice contained an average of 16% of
CD8 T cells which corresponded to 2.3 x 106
cells/spleen at day 7 following secondary PR8 infection, whereas
4-1BBL-/- mice only had 7% of
CD8+ T cells which corresponded to
106 cells per spleen (this difference was
statistically significant, p < 0.001). The primary
response to influenza PR8 is similar to the primary response to
influenza x31, resulting in 7% of CD8 T cells staining with the
Db/NP366-374 tetramer (data not shown). Thus, the
secondary response to influenza in 4-1BBL-/-
mice is reduced to the level of a primary response. The defect in cell
numbers in 4-1BBL-/- mice was also observed by
monitoring intracellular IFN- staining, although as seen in the
primary response, the numbers were lower overall.
Fig. 5 summarizes the time course of the
immune response to influenza virus in wild-type siblings,
4-1BBL-/- or CD28-/-
mice during the in vivo immune response to influenza virus, using
tetramer staining measured at days 5, 7, 10, 14, 21, and 38 after
primary infection and days 5 and 7 after secondary infection. The
results show a delay in the effect of 4-1BB/4-1BBL-mediated
costimulation, with CD28 influencing cell numbers early, and 4-1BBL
influencing cell numbers only late in the primary response. As
expected, mice given a second dose of the same influenza strain HKx31
showed a reduced number of tetramer and IFN--staining cells compared
with PR8, presumably due to the neutralizing Ab reducing the infectious
viral load (Fig. 5). Again, 4-1BBL-/- mice
showed reduced responses compared with wild-type mice. Thus under
conditions of both a high and low viral load,
4-1BBL-/- mice show a reduced CD8 T cell memory
response to influenza virus. CD28-/- mice
responded poorly to secondary infection with either PR8 or HKx31
consistent with their greatly diminished primary response to
influenza.
|
The effects of 4-1BBL could be due to a direct effect on CD8 T
cells or perhaps indirect through CD4 T cells or by changing the levels
of neutralizing Ab which would alter the availability of infective
virus and the CTL response. To address this issue with CD4 T cells, we
analyzed both late primary (day 38) and secondary CD4 T cell responses
to influenza by restimulating T cells from immunized mice with
heat-inactivated virus and measuring IL-2 and IFN- production (Fig. 6). In contrast to the results observed
late in the primary and in the secondary CD8 T cell responses to
influenza, there was no defect in CD4 T cell responses at either time
point. In the same experiment, CD28-/- mice
showed a complete lack of responsiveness in terms of IL-2 and IFN-
production.
|
Previous studies on 4-1BBL-/- mice showed
no defect in the neutralizing Ab response to vesicular stomatitis virus
(16) or in levels of LCMV-specific Abs in serum
(20). However, treatment with agonistic anti-4-1BB Abs
has been shown to interfere with humoral immunity, suggesting a role
for 4-1BB in regulation of Ab responses (27). To address
this issue in the context of influenza virus infection, we measured the
neutralizing Ab response as well as the levels of influenza-specific
Abs in wild-type vs CD28-/- or
4-1BBL-/- mice (Fig. 7). Both wild-type and
4-1BBL-/- mice produced similar amounts of
neutralizing Ab to influenza A HKx31 during the infection (Fig. 7A). The initial response was predominantly IgM at day 7
with class switching to IgG1 and IgG2a late in the primary and in the
secondary response to influenza Hkx31 (Fig. 7B).
CD28-/- mice initially produced neutralizing Ab
at levels similar to wild-type mice at day 7 (Fig. 7A) and
as expected this was predominantly IgM (Fig. 7B). However,
late in the primary response, less neutralizing Ab was produced in
CD28-/- mice compared with wild-type mice,
correlating with low levels of influenza-specific IgG1 and IgG2a (Fig. 7, A and B). Thus, there is no defect in
influenza-specific Ab responses in 4-1BBL-/-
mice under conditions where a defect is observed in the
CD28-/- mice. Taken together, the data suggest
that the defect in the 4-1BBL-/- mice is
restricted to CD8 T cells.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials MethodsResultsDiscussionReferences |
|---|
).
Previous work had demonstrated that 4-1BBL-/-
mice are impaired in secondary responses to influenza virus, as
measured by assaying cytotoxic T cell function upon restimulation in
vitro. However, it was not clear at what point in the response the
defect in the 4-1BBL-/- mice emerged
(16). In the present study, by following Ag-specific T
cell numbers using MHC I tetramers or IFN- staining, we found that
4-1BBL-/- mice have no defect in initial T cell
expansion in response to influenza virus, rather the defect is observed
much later in the response.
In vitro, 4-1BB is up-regulated on T cells with peak expression at 48–72 h after peptide-specific T cell activation and the expression is transient (17, 18). In vivo, immunization of mice with the superantigen staphylococcal enterotoxin A leads to peak expression of 4-1BB on CD8 T cells at 12 h with declining expression by 24 h (28). We have been unable to detect 4-1BB expression on the tetramer-positive T cells during primary or secondary infection with influenza likely due to the low level and transient nature of its expression (data not shown). However, the available evidence suggests that 4-1BB is expressed only early and transiently during the immune response in vivo. These data invoke a model in which signals delivered through 4-1BB early in the response impact on cell numbers much later in the primary response. Such an effect would be observed if 4-1BB impacted on signals involved in long-term T cell survival or differentiation into a "memory" cell. This model is conceptually similar to the work of Huang et al. (29) in which short-term IL-4 exposure during TCR stimulation results in induction of long-term memory of CD8+ T cells. Similarly, as mentioned in the introduction, several groups have shown that a short-term exposure to Ag initiates a program of events, resulting in several cycles of cell division (3, 4, 5, 6). These types of experiments imply that signals given early during T cell activation can impact on the long-term fate of T cells. Based on the kinetic data shown here, we suggest that 4-1BB engagement early in T cell activation may impact on the efficiency with which the T cells enter or survive in the memory pool. The cell numbers, at days 21 and 38 of the primary response, show small but statistically significant differences (p < 0.01) between wild-type and 4-1BBL-/- mice and are consistent with a loss of memory T cells late in the primary response.
The number of Ag-specific T cells present at a given time point after viral infection reflects the net effects of expansion and death. Thus, 4-1BB could influence the size of the memory T cell pool by influencing the amount of cell division, the number of cells which undergo activation induced cell death following initial expansion, or could impact on the gradual loss of cells from the memory T cell pool over time. Hurtado et al. (13) have shown that anti-4-1BB treatment in vitro can prevent activation-induced cell death of repetitively activated T cells. However, in the present study examination of the loss of tetramer-positive cells immediately after the peak of the response shows no evidence for a major impact of 4-1BBL at days 10–14 after infection, a time when the activated effector cells are being rapidly lost from the T cell pool. Furthermore, we did not observe differences in annexin V staining on tetramer-positive cells from influenza-infected wild-type or 4-1BBL-/- mice (data not shown). Differences in T cell numbers between wild-type and 4-1BBL-/- mice were statistically significant at days 21 and 38 after influenza infection. This might reflect a gradual loss of the memory T cells that have survived the initial activation-induced cell death that follows the expansion of the effector T cell pool.
The loss of tetramer+
CD62Llow memory CD8 T cells late in the primary
response could be due to a direct effect of 4-1BBL on long-term
survival of the CD8+ influenza-specific T cells
or it could be due to an indirect effect of CD4 help on the maintenance
of the CD8 T cell response. CD4 T cells have been shown to respond to
4-1BBL in vitro (12, 13, 17, 18), and an effect of 4-1BB
has been implied during MHC II-restricted graft-vs-host disease in vivo
(22). However, we did not detect any decrease in secondary
CD4 T cell responses to influenza in these mice (Fig. 6), arguing
against a major role for 4-1BBL in the CD4 T cell response to influenza
virus. Similarly, Tan et al. (20) observed little or no
difference in the CD4 T cell response to LCMV in
4-1BBL-/- vs wild-type mice. Thus, although
4-1BBL clearly can participate in activation of CD4 T cells in some
experimental models, in the virus infection models examined to date,
defects in the immune response in 4-1BBL-/-
mice are limited to the CD8 T cell subset (Refs. 16, 20 and this report).
It has been reported that treatment with agonistic anti-4-1BB Ab during immunization results in a block in humoral immunity (27). In contrast, examination of specific Ab production during LCMV or VSV infection of 4-1BBL-/- mice revealed no differences in neutralizing Ab responses or the amounts of specific Ab isotypes produced (16, 20). The level of influenza-specific Ab produced could influence the size of the CTL response by altering the availability of infective virus. To address this possibility, we analyzed the Ab response to influenza and found no difference between 4-1BBL-/- and wild-type mice with respect to neutralizing Ab or isotypes produced at days 7 and 38 following a primary infection. In addition, following secondary infection with influenza virus, under conditions where there was a significant difference in the percentage of influenza-specific CD8 T cells between 4-1BBL-/- and wild-type mice, no differences were seen in the levels of neutralizing Ab or in the specific isotypes produced. However, CD28-/- mice clearly showed a reduced neutralizing Ab response and a reduction in the amounts of influenza-specific Ab for all isotopes tested.
In contrast to the effects of 4-1BBL on influenza virus, the response to LCMV in 4-1BBL-/- mice was initially thought to be relatively unimpaired. DeBenedette et al. (16) showed no defect in the primary or secondary CTL response in 4-1BBL-/- or 4-1BBL-/-CD28-/- mice and no defect in viral clearance. Tan et al. (20) also found that 4-1BBL-/- mice cleared virus normally. However, when they analyzed the 4-1BBL-/- mice during the immune response to LCMV using tetramers specific for several viral epitopes, they found a small change in the percentage of CD8 T cells responding to LCMV at day 8 following infection (for example, 17% of CD8 T cells stained with GP33–41/Db tetramers at day 8 following infection, compared with 12% in 4-1BBL-/- mice). Furthermore, there seemed to be a general defect in expansion of the CD44high subset of CD8 T cells in the LCMV-infected mice, such that there were 2-fold more CD44high CD8 T cells in wild-type vs 4-1BBL-/- mice at day 8 following LCMV infection. In a subsequent study in which a lipidated peptide was used to immunize mice, Tan et al. (21) found a severe defect in the memory response to LCMV in 4-1BBL-/- mice such that mice succumbed to lethal viral challenge. These data support a role for 4-1BBL in memory CD8 T cell responses to LCMV.
Intraperitoneal delivery of influenza virus results in a fairly strong
immune response in the spleen with 7% of CD8 cells specific for the
major CTL epitope NP366-374 at the peak of the response. However,
influenza does not replicate extensively in the mouse. Thus, we did not
observe increases in the size of the spleen of the influenza-infected
mice. In contrast, LCMV replicates extensively in mice leading to
significantly higher expansion of viral-specific
CD8+ T cells in the spleen than seen in influenza
infection. For example, 107
LCMV-NP396–404-specific CD8+ T cells are seen in
the spleen at day 8 following LCMV infection (20) whereas
only 106 influenza-NP366-374-specific
CD8+ T cells are found in the spleen at day 7
following influenza infection (in this report).
These differences in the amount of viral replication and the tremendous
increases in numbers of CD8 T cells during the response to LCMV might
explain the differences in the kinetics of the effect of 4-1BBL that we
see in this study compared with the studies of LCMV infection of
4-1BBL-/- mice (20). During the
LCMV response, defects were observed at day 8 in both
CD28-/- mice (30) and
4-1BBL-/- mice (20) although it is
difficult to evaluate kinetic differences in the effects of these
costimulatory pathways, as only the day 8 time point was reported. In
the present study, we observed a clear segregation of the effects of
CD28 vs 4-1BBL costimulation, where CD28-/-
mice have defects much earlier in the response to infection than the
4-1BBL-/- mice. Furthermore, in the influenza
model, the effects of CD28 were much more dramatic than the effects of
4-1BBL, whereas in the LCMV model the effects of CD28 were similar in
magnitude to the effects of 4-1BBL, that is 2-fold when converted to
total numbers of tetramer-positive CD8 T cells per spleen (20, 30). We do not think our failure to detect an effect of 4-1BBL
on primary T cell expansion is due to insensitivity of the assay or due
to high Ag load. In the case of LCMV, the response was larger and yet
both CD28-/- and
4-1BBL-/- mice showed defects at day 8 after
infection. In our studies, we could readily detect a defect in the
CD28-/- mice early in the response, but only
detected the effect on the 4-1BBL-/- mice later
in the response. The very large expansion of CD8 T cells due to
extensive LCMV replication early during the infection may accelerate
the timing of the effects of costimulation. In contrast, the more
modest expansion of CD8 T cells during the immune response to influenza
has allowed us to visualize a temporal segregation of the effects of
the CD28 vs 4-1BB costimulatory pathways, with CD28 clearly acting very
early in the response and 4-1BBL only impacting on T cell numbers much
later in the response.
Members of the TNFR family, including 4-1BB are known to influence cell
survival by activating the NF-
B pathway (31), which in
turn can lead to up-regulation of Bcl-xL
(32) as well as cellular inhibitors of apoptosis 1 and 2
(33). Upon aggregation of 4-1BB on a T cell, 4-1BB
recruits TNFR-associated factors 1 and 2 which in turn links
4-1BB to the NF-B pathway as well as the p38 and stress-activated
protein kinase/c-Jun N-terminal kinase mitogen-activated protein
kinase pathways (15, 31, 34, 35, 36). Both CD28 as well as
members of the TNFR family, including OX40, influence
Bcl-xL expression (37, 38, 39). Thus, it
is conceivable that several pathways impact on regulation of T cell
survival perhaps acting to sustain or reinforce
Bcl-xL expression as well as other survival
signals. We used intracellular staining for
Bcl-xL to try to detect differences in
Bcl-xL expression on tetramer-positive
influenza-specific T cells. However, the staining with
anti-Bcl-xL was too weak to draw conclusions
(data not shown). In addition to members of the Bcl-2 family, long-term
T cell survival can also be regulated by cytokines. For example,
pretreatment with IL-4 can affect long-term CD8 T cell survival
(29). Furthermore, CD8 memory T cell survival has been
shown to be dependent on IL-15 (40, 41, 42). Thus, another
possible mechanism of 4-1BB-induced survival of the memory pool would
be to influence the responsiveness of the memory T cell pool to
cytokines.
Other TNFR family members have also been suggested to be important in sustaining T cell activation. For example, CD27-/- mice showed a reduced number of CD4 and CD8 T cells infiltrating the lungs during secondary influenza virus infection (43), although there was no decrease in CTL effector function as a result of the CD27 deficiency. OX40 has also been shown to affect the immune response to influenza virus (44). In this case, the effects of OX40 deficiency were limited to the CD4 T cell subset (44, 45, 46, 47), with no detectable effect on the CD8 T cell response (44). Recently, it was demonstrated using in vitro analysis of TCR transgenic T cells lacking CD28 or OX40 that CD28 regulates Bcl-2 family levels at days 2–4 after activation and that OX40 is responsible for prolonging the life of anti-OX40-treated TCR transgenic T cells by regulating Bcl-xL and Bcl-2 later in the immune response. These data provide evidence for sequential action of CD28 and OX40 on T cell survival in vitro and are entirely consistent with our data on the temporal segregation of the effects of 4-1BBL from those of CD28 in vivo. Thus, CD28 appears to be responsible for a first wave of survival signals and this can then be sustained by inducible costimulatory pathways such as OX40 or 4-1BB. It is likely that these survival signals are multifaceted and include anti-apoptotic as well as prosurvival signals. The precise molecular signals induced by each of the costimulatory pathways remains to be fully elucidated.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Tania H. Watts, Department of Immunology, University of Toronto, Room 5263 Medical Sciences Building, 1 King’s College Circle, Toronto, Ontario, Canada M5S 1A8. E-mail address: tania.watts{at}utoronto.ca
3 Abbreviations used in this paper: 4-1BBL, 4-1BB ligand; LCMV, lymphocytic choriomeningitis virus; HAU, hemagluttinin unit; CD62L, CD62 ligand.
Received for publication September 10, 2001. Accepted for publication February 6, 2002.
|
References |
|---|
|
TopAbstractIntroductionMaterials MethodsResultsDiscussionReferences |
|---|
2-microglobulin constructs. J. Immunol. 160:1598.B. Mol. Cell. Biol. 18:558.B control. Proc. Natl. Acad. Sci. USA 94:10057.B. Biochem. Biophys. Res. Comm. 242:613.[Medline]
B cascade is important in Bcl-xL expression and for the anti-apoptotic effects of the CD28 receptor in primary human CD4+ lymphocytes. J. Immunol. 165:1743.This article has been cited by other articles:
|
|
 |
|
  L. Sabbagh, G. Pulle, Y. Liu, E. N. Tsitsikov, and T. H. Watts ERK-Dependent Bim Modulation Downstream of the 4-1BB-TRAF1 Signaling Axis Is a Critical Mediator of CD8 T Cell Survival In Vivo J. Immunol., June 15, 2008; 180(12): 8093 - 8101. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Fuse, W. Zhang, and E. J. Usherwood Control of Memory CD8+ T Cell Differentiation by CD80/CD86-CD28 Costimulation and Restoration by IL-2 during the Recall Response J. Immunol., January 15, 2008; 180(2): 1148 - 1157. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. C. P. Waller, N. McKinney, R. Hicks, A. J. Carmichael, J. G. P. Sissons, and M. R. Wills Differential costimulation through CD137 (4 1BB) restores proliferation of human virus-specific "effector memory" (CD28 CD45RAHI) CD8+ T cells Blood, December 15, 2007; 110(13): 4360 - 4366. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Habib-Agahi, T. T. Phan, and P. F. Searle Co-stimulation with 4-1BB ligand allows extended T-cell proliferation, synergizes with CD80/CD86 and can reactivate anergic T cells Int. Immunol., December 1, 2007; 19(12): 1383 - 1394. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. B. Borowski, A. C. Boesteanu, Y. M. Mueller, C. Carafides, D. J. Topham, J. D. Altman, S. R. Jennings, and P. D. Katsikis Memory CD8+ T Cells Require CD28 Costimulation J. Immunol., November 15, 2007; 179(10): 6494 - 6503. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. P. McAleer, D. J. Zammit, L. Lefrancois, R. J. Rossi, and A. T. Vella The Lipopolysaccharide Adjuvant Effect on T Cells Relies on Nonoverlapping Contributions from the MyD88 Pathway and CD11c+ Cells J. Immunol., November 15, 2007; 179(10): 6524 - 6535. [Abstract] [Full Text] [PDF]
|
|
|
|
