Memory precursor phenotype of CD8+ T cells reflects early antigenic experience rather than memory numbers in a model of localized acute influenza infection.

Abstract

The mechanistic basis of memory T-cell development is poorly defined. Phenotypic markers that define precursors at effector stages have been characterized for acute systemic infections with high antigen load. We asked whether such markers can identify memory precursors from early effectors (d6) to late memory (>d500) for two immunodominant CD8(+) responses during the course of a localized low-load influenza infection in mice. CD8(+) T cells stained with the D(b) NP(366) and D(b) PA(224) tetramers were characterized as IL-7Rα(hi) , IL-7Rα(hi) CD62L(hi) or IL-7Rα(hi) KLRG1(lo) . While the D(b) NP(366) - and D(b) PA(224) -specific responses were comparable in size, decay kinetics and memory precursor frequency, their expansion characteristics differed. This correlated with a divergence in the IL-7Rα(hi) , IL-7Rα(hi) CD62L(hi) and IL-7Rα(hi) KLRG1(lo) phenotypes on effector, but not naïve, CD8(+) populations. That effect was abrogated by priming with viruses engineered to present equivalent levels of NP(366) and PA(224) peptides, indicating that memory phenotypes reflect early antigenic experience rather than memory potential. Thus, the IL-7Rα(hi) KLRG1(lo) phenotype had a poor predictive value in identifying memory precursors in the spleen and at the site of infection. Greater consistency in influenza-specific IL-7Rα(hi) KLRG1(lo) CD8(+) T-cell numbers was found in draining lymph nodes, suggesting that this may be the preferential site for memory establishment and maintenance following localized virus infections.

Full text

Page 1

Memory precursor phenotype of CD81T cells reflects
early antigenic experience rather than memory numbers
in a model of localized acute influenza infection
Hayley A. Croom1, Alice E. Denton1, Sophie A. Valkenburg1,
Natasha G. Swan1, Matthew R. Olson1, Stephen J. Turner1,
Peter C. Doherty1,2and Katherine Kedzierska1
1Department of Microbiology and Immunology, University of Melbourne, Parkville, Melbourne,
Australia
2Department of Immunology, St Jude Children’s Research Hospital, Memphis, TN, USA
The mechanistic basis of memory T-cell development is poorly defined. Phenotypic
markers that define precursors at effector stages have been characterized for acute
systemic infections with high antigen load. We asked whether such markers can identify
memory precursors from early effectors (d6) to late memory (4d500) for two immuno-
dominant CD81responses during the course of a localized low-load influenza infection in
mice. CD81T cells stained with the DbNP366and DbPA224tetramers were characterized as
IL-7Rahi, IL-7RahiCD62Lhior IL-7RahiKLRG1lo. While the DbNP366- and DbPA224-specific
responses were comparable in size, decay kinetics and memory precursor frequency, their
expansion characteristics differed. This correlated with a divergence in the IL-7Rahi,
IL-7RahiCD62Lhiand IL-7RahiKLRG1lophenotypes on effector, but not naı ¨ve, CD81popu-
lations. That effect was abrogated by priming with viruses engineered to present
equivalent levels of NP366and PA224peptides, indicating that memory phenotypes reflect
early antigenic experience rather than memory potential. Thus, the IL-7RahiKLRG1lo
phenotype had a poor predictive value in identifying memory precursors in the spleen and
at the site of infection. Greater consistency in influenza-specific IL-7RahiKLRG1loCD81
T-cell numbers was found in draining lymph nodes, suggesting that this may be
the preferential site for memory establishment and maintenance following localized
virus infections.
Keywords: Cytotoxic T cells.Repertoire.T cells.T-cell receptors.Viral infection
Supporting Information available online
Introduction
Memory CD81T cells provide protection against intracellular
pathogens. Although priming CD81T-cell memory has the
potential to enhance any vaccine strategy, our understanding
of development pathways for memory is far from clear.
One proposal is that subsequent to a ten-fold contraction in
numbers, memory T cells are a surviving subset of the
acute effector population. Recent evidence suggests, however,
that this idea is simplistic and the full expansion of effector
T cells is not a pre-requisite for the establishment of long-term
memory [1–6].
Correspondence: Dr. Katherine Kedzierska
e-mail: kkedz@unimelb.edu.au
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji-journal.eu
DOI 10.1002/eji.201040625 Eur. J. Immunol. 2011. 41: 682–693 Hayley A. Croom et al.
682

Page 2

Understanding early memory T-cell development has been
limited by a lack of specific phenotypic markers. Recently, the
expression of IL-7Ra [7] and IL-7RahiKLRG1lo[8] has been
considered to define the memory precursors within effector T-cell
populations. Following LCMV infection, the memory precursor
effector cells (MPEC) are IL-7RahiKLRG1lowhile the short-lived
effector cells (SLEC) are IL-7RaloKLRG1hi. Acquisition of this
memory IL-7RahiKLRG1lophenotype in models of rapid, systemic
infection (LCMV and Listeria monocytogenes) depends on the
spectrum of inflammatory cytokines present during the acute
phase of the disease process [8, 9].
The idea that the memory precursor phenotype is determined
by the nature of inflammatory stimuli raises the question
of what happens during localized infections. The present analysis
uses the un-manipulated non-TCR-transgenic model of localized
transient influenza virus infection in C57Bl/6J mice to char-
acterize what drives the emergence of the IL-7RahiKLRG1loCD81
T cells. The most prominent CD81populations emerging after
initial priming are specific for H2Db-bound peptides from the
viralnucleoprotein(DbNP366–374)
(DbPA224–233) proteins [10]. These two responses are, however,
far from identical. Although the primary infection with influenza
virus results in similar DbNP1
ses, the DbPA1
polyfunctionality [11].
Longitudinal analysis of TCR CDR3b profiles for the DbNP1
CD81and DbPA1
TCRb clonotypes within effector and memory populations [5, 12,
13]. Additionally, influenza-specific CD81cells detected within
the draining lymph nodes (mediastinal LN, MLN) as soon as d3.5
after infection can survive into memory and be recalled after
subsequent challenge [6]. This suggests that stable CD81
T-cell memory is established early in the antigen-driven
phase of influenza virus infection. Being able to track memory
precursors phenotypically at the early stages of infection would
greatly advance our understanding of anti-viral T cell-mediated
immunity.
Our study analyses recently proposed memory phenotypes
(IL-7Rahi, IL-7RahiCD62Lhi, IL-7RahiKLRG1lo) for their predictive
value in identifying memory precursors, from the early effector
(d6) to the late memory (4d500) stage of influenza virus
infection. Although the IL-7RahiKLRG1loCD81sets recovered
from the spleen and the infected respiratory tract show
evidence of enhanced survival, this phenotype does not define
memory numbers. We show that more stable IL-7RahiKLRG1lo
numbers are observed in the regional MLN providing further
evidence that the draining LN offer the optimal anatomical
niche for memory establishment and maintenance after influenza
infection.Furthermore, the
IL-7RahiKLRG1lophenotypes can be manipulated by altering
antigen dose, character of antigen presentation, extent of T-cell
division and CD81T-cell precursor numbers, suggesting that cell-
surface molecule expression on CD81T cells reflects early
antigenic experience rather than the capacity to generate CD81
T-cell memory.
andacid polymerase
366CD81and DbPA1
224CD81respon-
224CD81population is of higher pMHCI avidity and
366
224CD81sets showed an identical distribution of
IL-7Rahi,IL-7RahiCD62Lhi,
Results
Early IL-7Ra expression is not a good predictor of long-
term survival
Analysis of influenza-specific tetramer1CD81T cells over a very
long time-course (d6–d575; Fig. 1A) showed that both the
CD81DbNP1
224counts peak between d8 and
d10 after i.n. challenge. While the DbPA1
more prominent initially (d6/d7), the DbNP1
a ?2? higher maximum between d9 and d15 (Fig. 1A). These
differential kinetics result from a larger naı ¨ve DbPA1
precursor frequency and TCR repertoire allowing that response to
emerge earlier [14–16]. Conversely, the more abundant viral NP
induces higher expression of the DbNP366epitope, which leads to
greater proliferation of DbNP1
virus clearance (?d9), a contraction of DbNP1
PA1
cells results in 5–8% surviving into long-term
memory (Fig. 1A).
The levels of IL-7Ra expression on these DbNP1
DbNP1
infection(Fig.1B). Irrespective
of DbPA1
the DbPA1
sets displayed higher IL-7Ra prevalence
than DbNP1
populations. This difference was most
obvious at the acute time-points, with an unexpectedly high
proportion of DbPA1
on d8, 58.4% on d9 and 70.2% on d10) in comparison to
DbNP1
d10). These data indicate that the extent of IL-7Ra expression at
the effector phase does not correlate with the 5–8% of cells that,
for both epitopes, survive into long-term memory. Furthermore,
given that the T-cell ‘‘diaspora’’ to other sites is already estab-
lished early after influenza virus infection [17], it is likely that
much of the IL-7Rahipopulation dies rather than localizes to
other tissues.
Differential IL-7Ra expression on responding DbPA1
and DbNP1
regulates IL-7Ra expression. To assess whether differential
IL-7Ra expression at the acute phase relates to the level of IL-7Ra
staining on naı ¨ve DbNP1
precursors were identified with tetramer-based magnetic enrich-
ment [16, 18] and co-stained with anti-IL-7Ra and CD44 (Fig. 2).
Naı ¨ve DbNP1
able phenotype and MFI of IL-7Ra and CD44 staining, indicating
that differential IL-7Ra expression on effectors does not reflect
distinct IL-7Ra profiles on naı ¨ve populations. Furthermore, both
IL-7Rahiand IL-7Ralosubsets within DbNP1
Information Table 1) and DbPA1
lations had common TCRb clonotypes. This suggests that IL-7Rahi
and IL-7Ralosets for both DbNP1
populations were established in a similar fashion from common
progenitors. Similar conclusion of common progenitors was
recently also reached for IL-7RahiKLRG1loand IL-7RaloKLRG1hi
subsets [19].
366and CD81DbNP1
224CD81population is
366CD81set achieves
224CD81
366CD81cells on d9–d10. Following
366CD81and Db
224CD81
224CD81and
366CD81cells (Fig. 1A) varied from d8 until d115 after
of therelativenumbers
224CD81and DbNP1
224CD81
366CD81
366CD81cells surviving into memory,
224CD81cells expressing IL-7Ra (51.5%
366CD81sets (33.7% on d8, 26.0% on d9 and 38.2% on
224CD81
366CD81cells raises the intriguing question of what
366CD81and DbPA1
224CD81cells, naı ¨ve
366CD81and DbPA1
224CD81sets displayed compar-
366CD81(Supporting
224CD81(data not shown) popu-
366CD81and DbPA1
224CD81
Eur. J. Immunol. 2011. 41: 682–693
Immunity to infection
683
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji-journal.eu

Page 3

To determine whether distinct IL-7Ra profiles correlate with a
measure of activation, the sets were co-stained for IL-7Ra and
CD62L (Fig. 1C). At acute time-points (d8–d10), the prevalence
of IL-7RahiCD62Lhiphenotype was about 2? higher for DbPA1
CD81cells from the spleen (Fig. 1C), draining MLN and the site
of infection sampled by bronchoalvolar lavage (BAL) (data not
shown).The DbPA1
set remained
(CD62Lloor CD62Lhi) from d15 to d575 while DbNP1
T cells were more likely to be IL-7RaloCD62Llountil d60 after
infection (Fig. 1C), indicative of a higher activation status of
these cells. Furthermore, although the DbPA1
progressed to a memory IL-7RahiCD62Lhiphenotype at a faster
rate than DbNP1
rates (Fig. 1A). Following secondary virus challenge, the staining
profiles for IL-7Ra and CD62L retained the same differential for
the acute time-points (Fig. 1D).
Analysis of early activation markers established that DbNP1
and DbPA1
224CTL have different activation states, with the
DbNP1
CD69 was up-regulated earlier on DbNP1
(Fig. 3A), spleen (Fig. 3B) and lung (Fig. 3C), suggesting
224
224CD81
mostly IL-7Rahi
366CD81
224CD81population
366CD81set, both decayed numerically at similar
366
366CD81effectors displayed a more activated phenotype.
366CD81in MLNs
increased antigen priming/exposure for DbNP1
tion. Marked differences were found on d7 and d10 after influ-
enza infection. The MFI of CD44 staining was also higher for
DbNP1
increased CD44 expression in MLN (Fig. 3D) but not spleen at
different stages of infection (Fig. 3E and Supporting Information
Fig. 1B). Conversely, MFI of tetramer staining (Supporting
Information Fig. 1A), TCRb and CD3e [20] were comparable for
both DbNP1
indicate that the expression of IL-7Ra on effector CTLs is epitope-
dependent and reflects decreased activation of CD81T cells,
rather than predicting survival to memory.
366CD81popula-
366CD81cells in lungs (Fig. 3F), with a trend toward
366CD81and DbPA1
224CD81cells. Thus, our data
Differential prevalence of the MPEC IL-7RahiKLRG1lo
phenotype
Based on the recent definition [8] of the memory-precursor MPEC
IL-7RahiKLRG1lophenotype, we asked whether further subsetting
of IL-7RahiCD81cells based on KLRG1 expression might predict
the relative numbers of DbPA1
224CD81and DbNP1
366CD81cells
Figure 1. Distinct IL-7Rahiand IL-7RahiCD62Lhiprofiles for CD81DbNP1
DbNP1
mice with X31 virus; (B) relative proportion of IL-7Rahiwithin the tetramer1populations. (C, D) Profiles of IL-7Ra and CD62L staining. Lymphocytes
were obtained from the spleens of individual mice (n53–34 mice per time-point) at (A–C) 6-575d after primary infection or (D) 6-420d after
challenge. Secondary challenge with the X31 (H3N2) was performed 6 wk after mice were primed with PR8 (H1N1) influenza. Cells were stained
with the DbNP366or DbPA224tetramers-APC, anti-CD8a-PerCPCy5.5 and anti-CD127-PE. (A, B) Statistically significant (?p o 0.01;]p o 0.05) differences
are shown for DbNP1
Experiments were performed twice for d8–d60. (C, D) Representative FACS plots are shown for (C) primary and (D) secondary responses. For the
gating strategy to FACS data presented in all the figures, refer to Supporting Information Fig. 3.
366and CD81DbPA1
224cells following influenza infection. (A) Numbers of
366CD81and DbPA1
224CD81cells from the early acute (d6) phase until long-term memory (d575) following primary i.n. challenge of naı ¨ve B6
366CD81and DbPA1
224CD81cells at various time-points. A two-tailed paired t-test was performed. Error bars represent SD.
Eur. J. Immunol. 2011. 41: 682–693Hayley A. Croom et al.
684
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji-journal.eu

Page 4

transiting into memory. Assessment of IL-7Ra and KLRG1
staining at the acute (d8–d10) and early to long-term memory
(d35–d508) time-points (Fig. 4) established that this phenotype
is differentially expressed on DbPA1
sets. In the long-term, the majority of memory DbPA1
DbNP1
IL-7RahiKLRG1loMPEC profile (Fig. 4A, B, E and F). However,
marked differences were observed at the earlier time-points
(d8–d10), with the DbPA1
phenotype more rapidly than the DbNP1
spleen (Fig. 4A and B), MLN (Fig. 4E and F) and BAL (Fig. 4I
and J). Interestingly, rather than showing the IL-7RaloKLRG1hi
[8] phenotype of SLEC, the main subset of influenza-specific
DbNP1
224CD81and DbNP1
366CD81
224CD81and
366CD81cells in spleen and MLN display the characteristic
224CD81set progressing to the MPEC
366CD81population in
366CD81cells in spleen and MLN on d8–d10 was KLRG1lo
IL-7Ralo(Fig. 4A and E). This is in contrast to the situation for
the DbPA1
MLN displayed IL-7RahiKLRG1lo(Fig. 4B and F), though a
IL-7RaloKLRG1lopopulation dominated in BAL. How does such
differential IL-7RahiKLRG1loexpression on DbNP1
DbPA1
CD81T cells into long-term memory?
To determine whether IL-7RahiKLRG1losubsets of DbNP1
and DbPA1
population from the effector phase (d8) through to long-term
memory (d508), we calculated the numbers for all four subsets
identified by IL-7Ra and KLRG1 staining. In the spleen, it was
evident that every population (IL-7RahiKLRG1lo, IL-7RahiKLRG1hi,
IL-7RaloKLRG1loand IL-7RaloKLRG1hi) expanded between d8 and
d10 and then contracted at essentially equivalent rates (Fig. 4C and
D). Overall, none of these four phenotypically defined sets can be
considered to characterize a stable memory precursor population, at
least in the context of influenza virus infection.
DbNP1
DbPA1
This suggests that the accelerated progression of DbPA1
responders to the MPEC phenotype occurs irrespective of the
extent of IL-7RahiKLRG1loclonal expansion. Assessment of MPEC
IL-7RahiKLRG1lonumbers between d10 and d35 showed a
decrease in both DbNP1
(6.4-fold) cells in spleen. The decay rates for the remaining three
(IL-7RahiKLRG1hi, IL-7RaloKLRG1loand IL-7RaloKLRG1hi) subsets
were more dramatic for both populations (DbNP1
9.9–78.4?; DbPA1
subsets characterized in spleen by differential IL-7Ra and KLRG1
staining profiles declined in numbers during the progression to
memory, with the DbNP1
lation displaying the lowest decay rate.
Analysis of DbNP1
MLN led, however, to a slightly different conclusion. All the
subsets were more stable in the MLN between d10 and d35 for
both epitope-specific sets (especially the MPEC IL-7RahiKLRG1lo)
with no numerical changes for DbNP1
decrease for DbPA1
our earlier studies showing superior memory survival and recall
potential for influenza-specific CD81T cells from the MLN versus
spleen [6]. Subsequent functional analysis of DbPA1
DbNP1
higher proportion of spleen-derived CD81sets (of both total
tetramer1and MPEC populations) produced granzyme B when
compared to populations recovered from MLNs (Fig. 5). These
data suggest preferential accumulation of memory (rather than
effector) CD81T cells in the MLN.
Assessment of the DbNP1
within the BAL showed that all four subsets decreased in numbers
at the same rate, which likely reflects the kinetics of antigen
clearance at the site of infection (Fig. 4K and L). The decay for
MPEC IL-7RahiKLRG1locells was between two- and threefold for
both influenza-specific CD81populations. It appears that with
the exception of the DbNP1
224CD81set, where most of effectors in spleen and
366CD81and
224CD81cells relate to the progression of influenza-specific
366CD81
224CD81cells represent a stable memory precursor
366CD81cells tended to be IL-7RaloKLRG1lowhile the
224CD81set was more likely to be IL-7RahiKLRG1loat d10.
224CD81
366CD81(2.8-fold) and DbPA1
224CD81
366CD81, range:
224CD81, range: 14.3–45). Thus, all four
366CD81MPEC IL-7RahiKLRG1lopopu-
366CD81and DbPA1
224CD81cells from the
366CD81and 2.3-fold
224CD81(Fig. 4G and H). These results support
224CD81and
366CD81cells obtained from spleen and MLN showed that a
366CD81and DbPA1
224CD81cells
366CD81set in MLN, the MPEC
Figure 2. Comparable IL-7Ra and CD44 profiles for naı ¨ve DbNP1
and DbPA1
within lymph nodes and spleens were identified with the DbNP366-PE
and DbPA224-PE tetramers using an established protocol [16]. Naı ¨ve
DbNP1
profiles by staining with anti-CD127-APC and anti-CD44-PECy7 mAbs.
(A) Dot plots of memory DbNP1
naı ¨ve DbPA1
and (C) MFI of IL-7Ra and CD44 staining for infected and naı ¨ve CD81
T cells are for groups of four mice. Data shown are mean1SD.
Experiment was performed twice.
366CD81
224CD81cells. Naı ¨ve DbNP1
366CD81and DbPA1
224CD81precursors
366CD81and DbPA1
224CD81cells were assessed for IL-7Ra and CD44
366CD81(control), naı ¨ve DbNP1
366CD81and
224CD81cells are shown for IL-7Ra and CD44. (B) Phenotype
Eur. J. Immunol. 2011. 41: 682–693
Immunity to infection
685
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji-journal.eu

Page 5

IL-7RahiKLRG1lopopulations decrease significantly in numbers at
the different anatomical sites. Thus, the MPEC IL-7RahiKLRG1lo
phenotype in the context of influenza infection is a poor predictor
for identifying the memory precursors.
To determine whether any of the IL-7RahiKLRG1loMPEC sets
within DbNP1
to exhaustion,weassessed
PD-1 expression [21]. Our data show that both IL-7RahiKLRG1lo
and IL-7RaloKLRG1hisubsets of DbPA1
more IL-2 in comparison to the DbNP1
supports our previous work [11] demonstrating that overall
DbPA1
As the proportion of IL-2-producing DbPA1
CD81cells markedly increases with progression to memory [11],
366CD81or DbPA1
224CD81populations are more prone
IL-2production,CD25and
224CD81populations produce
366CD81cells (Fig. 6A). This
224CD81T cells produce more IL-2 than DbNP1
366CD81sets.
224CD81and DbNP1
366
these IL-2 data confirm our findings of decreased activation of
DbPA1
Expression ofCD25 and
IL-7RahiKLRG1losets within DbNP1
populations (Fig. 6B and C), suggesting that none of those two
MPEC subsets is more prone to exhaustion. Interestingly, both
CD25 and PD-1 levels were elevated on the SLEC IL-7Ralo
KLRG1hisubsets in comparison to IL-7RahiKLRG1losets, with
CD25 levels being the highest for DbPA1
cells. These observations might explain a low survival rate of
SLEC IL-7RaloKLRG1hipopulations.
224CD81T cells rather than their exhaustion.
PD-1
366CD81and DbPA1
wascomparable for
224CD81
224CD81IL-7RaloKLRG1hi
MPEC IL-7RahiKLRG1loT-cell numbers are increased in
the recall response
Following secondary challenge, the acute phase of DbNP1
response is massively immunodominant [22], constituting up to
80% of the total virus-specific CD81pool (Fig. 7A). Might such
differencesin recall response
epitope-related divergence in IL-7Ra and KLRG1 profiles? Similar
to the differential progression to the MPEC IL-7RahiKLRG1lo
phenotype in the primary infection, the secondary DbPA1
effectors displayed higher level of IL-7RahiKLRG1lothan DbNP1
CD81CTLs (Fig. 7B). However, especially in MLN and BAL, the
extent of MPEC IL-7RahiKLRG1loexpression during the acute
phase of the secondary response was markedly higher (?60%)
than found for the primary DbNP1
sets (Fig. 4). These results suggest that either the MPEC
IL-7RahiKLRG1lo
populationis
expanded after secondary infection or that there is a reversion
from the remaining subsets following challenge. Thus, despite
significant differences in recall numbers (Fig. 7A), the MPEC
IL-7RahiKLRG1lophenotype for the recall DbPA1
DbNP1
primary co-dominant responses.
366CD81
magnitude contribute to
224CD81
366
366CD81and DbPA1
224CD81
preferentially recalledand
224CD81and
366CD811 cells (Fig. 7B) appears to be similar to the
Early antigenic stimulation determines the phenotypes
of effectors
Considering that the load of NP antigen is considered to be much
higher than for PA antigen [14], it is possible that the antigen
exposure impacts on the development of CD81phenotypes.
Engineered neuramindase (NA)NPand NAPAinfluenza viruses
were used to probe the spectrum of IL-7Rahi, IL-7RahiCD62Lhi
and MPEC IL-7RahiKLRG1loexpression for these two dominant,
primary influenza-specific CD81sets. These recombinant viruses
express NP366and PA224peptides in the viral NA stalk, in the
context of single amino acid mutations (N5Q) that disrupt
presentation from the endogenous site [14]. This NA substitution
strategy modifies both antigen dose and the APC presentation
profile for the NP366or PA224peptides by making the level of
antigen presentation equivalent for the DbNP366 and DbPA224
epitopes [14].
Figure 3. DbNP1
than DbPA1
and DbPA1
shown in (A) MLN, (B) spleen and (C) lungs. (D-F) MFI of CD44 staining
of tetramer1CD81cells was assessed on d7 for (D) MLN, (E) spleen and
(F) lungs. Lymphocytes were obtained from the individual mice
(n54–5). Cells were stained with the DbNP366or DbPA224tetramers-
PE, anti-CD8a-PerCPCy5.5, anti-CD69-FITC and anti-CD44-APC. Experi-
ments were performed twice for days 7, 10 and 14. Data are mean7SD.
?p o 0.01 by a two-tailed paired t-test.
366CD81effectors display a more activated phenotype
224CD81cells. (A–C) Profiles of CD69 staining for DbNP1
224CD81cells on d7–d42 after i.n. infection with X31 are
366CD81
Eur. J. Immunol. 2011. 41: 682–693 Hayley A. Croom et al.
686
& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji-journal.eu