of June 13, 2013.
This information is current as
Heterologous Viral Serotypes
Their Response to Variant Epitopes of
Quantitative and Qualitative Differences in
Dengue Virus-Reactive CD8+ T Cells Display
Hema S. Bashyam, Sharone Green and Alan L. Rothman
2006; 176:2817-2824; ;
, 21 of which you can access for free at:
cites 36 articles
is online at:
The Journal of Immunology
Information about subscribing to
Submit copyright permission requests at:
Receive free email-alerts when new articles cite this article. Sign up at:
Print ISSN: 0022-1767 Online ISSN: 1550-6606.
Immunologists All rights reserved.
Copyright © 2006 by The American Association of
9650 Rockville Pike, Bethesda, MD 20814-3994.
The American Association of Immunologists, Inc.,
is published twice each month by
The Journal of Immunology
by guest on June 13, 2013
Dengue Virus-Reactive CD8?T Cells Display Quantitative and
Qualitative Differences in Their Response to Variant Epitopes
of Heterologous Viral Serotypes1
Hema S. Bashyam, Sharone Green, and Alan L. Rothman2
Reactivation of serotype cross-reactive CD8?memory T lymphocytes is thought to contribute to the immunopathogenesis of
dengue disease during secondary infection by a heterologous serotype. Using cytokine flow cytometry, we have defined four novel
HLA-A*02-restricted dengue viral epitopes recognized by up to 1.5% of circulating CD8?T cells in four donors after primary
vaccination. All four donors had the highest cytokine response to the epitope NS4b 2353. We also studied the effect of sequence
differences in heterologous dengue serotypes on dengue-reactive CD8?memory T cell cytokine and proliferative responses. The
D3 variant of a different NS4b epitope 2423 and the D2 variant of the NS4a epitope 2148 induced the largest cytokine response,
compared with their respective heterologous sequences in all donors regardless of the primary vaccination serotype. Stimulation
with variant peptides also altered the relative frequencies of the various subsets of cells that expressed IFN-?, TNF-?, MIP-1?,
and combinations of these cytokines. These results indicate that the prior infection history of the individual as well as the serotypes
of the primary and heterologous secondary viruses influence the nature of the secondary response. These differences in the effector
functions of serotype cross-reactive memory T cells induced by heterologous variant epitopes, which are both quantitative and
qualitative, may contribute to the clinical outcome of secondary dengue infection. The Journal of Immunology, 2006, 176: 2817–
activation of the cross-reactive T cell component of the memory
pool by unrelated viruses resulted in altered T cell epitope hierar-
chies (1) and reshaped T cell repertoires (2). These changes influ-
enced the ultimate outcome of sequential infection by improving
protective immunity (3) or, in some instances, by enhancing im-
munopathology (4). The mechanisms of heterologous immunity
that may lead to a negative outcome have been inadequately stud-
ied in humans.
Dengue virus infections offer a clear paradigm for investigating
the role of cross-reactive memory T cells during an immune re-
sponse to sequential infections with heterologous viruses. The four
natural variants (serotypes) of dengue virus (1–4) can cause an
acute infection in humans that presents clinically either as classical
dengue fever (DF)3or as the more severe dengue hemorrhagic
fever (DHF) (5, 6). Primary infection with any of the four sero-
types results in lifelong immunity to the same serotype but leaves
the individual susceptible to sequential secondary infections by
everal recent studies have highlighted the importance of T
cell cross-reactivity in mediating immune responses dur-
ing infections with heterologous Ags. In these studies, re-
heterologous serotypes (7). Epidemiological studies in Cuba and
Southeast Asia have shown a strong association between second-
ary dengue infection and DHF (8, 9). These data indicate that
previous immunity to dengue may be a risk factor in the develop-
ment of severe dengue disease (10, 11). The mechanisms proposed
for this effect involve serotype cross-reactive, non-neutralizing
Abs, which can enhance infection, and serotype cross-reactive
memory T lymphocytes.
We and others have postulated that serotype cross-reactive T
cells in the memory repertoire are preferentially activated and ex-
panded during secondary infection and contribute to DHF immu-
nopathogenesis through an enhanced output of effector function
(12, 13). In support of this hypothesis, CD8?T cells from patients
with secondary dengue infection show higher binding to tetramers
with heterologous epitope sequences than to tetramers with se-
quences of the secondary serotype, revealing the higher affinity of
these cells for other presumably previously encountered dengue
serotypes (13). Plasma levels of IFN-?, TNF-?, soluble TNFRs
(sTNFRs), sIL-2R, and sCD8 have been found to be higher in
patients with DHF than in those individuals with DF (14–16).
PBMC from patients with DHF have a higher percentage of
CD69?CD8?cells as well as a higher frequency of dengue
epitope-specific CD8?lymphocytes, compared with PBMC from
patients with DF (17, 18).
In in vitro experiments, heterologous restimulation of dengue-
immune PBMC induced only a partial activation of a subset of
serotype cross-reactive cells specific for a dengue epitope. This
epitope was characterized as a partial agonist in its heterologous
form because it induced poor proliferation and IFN-? production
in a CTL clone while able to sensitize target cells for lysis (19).
Consequently, we have hypothesized that the cross-reactive
epitopes of the secondary heterologous serotypes may act as al-
tered peptide ligands for dengue-specific cross-reactive CD8?T
cells and effect changes in the overall functional response of this
Center for Infectious Disease and Vaccine Research, University of Massachusetts
Medical School, Worcester, MA 01655
Received for publication March 31, 2005. Accepted for publication December
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1This work was supported by National Institutes of Health Grants U19 AI57319 and
K08 AI01729 (to S. G).
2Address correspondence and reprint requests to Dr. Alan L. Rothman, Center for
Infectious Disease and Vaccine Research, University of Massachusetts Medical
School, Room S5-326, 55 Lake Avenue North, Worcester, MA 01655. E-mail ad-
3Abbreviations used in this paper: DF, dengue fever; DHF, dengue hemorrhagic
fever; s, soluble.
The Journal of Immunology
Copyright © 2006 by The American Association of Immunologists, Inc.0022-1767/06/$02.00
by guest on June 13, 2013
population by varying T cell proliferation, cytokine/chemokine
production, and cytotoxicity.
Several studies that examined the functional potential of human
memory CD8?T cells in systems such as HIV, CMV, and Ep-
stein–Barr virus have reported a polyclonal population of Ag-spe-
cific cells that include subsets with a diverse array of partial and
full effector functions (20–24). These subsets include cytokine-
competent but cytolysis-defective cells, proliferating but cytokine-
deficient cells, and cell populations that secrete a limited range of
cytokines. This functional heterogeneity of the memory population
suggests that the quality of the recall response will depend on
which of these different subsets undergo selective expansion and
dominate the secondary response.
The present study was designed to determine whether stimula-
tion of dengue epitope-specific memory CD8?T cells with variant
sequences from heterologous serotypes quantitatively or qualita-
tively alter the range of their effector functions. We used PBMC
from individuals immunized with experimental monovalent live-
attenuated dengue vaccines. Four new HLA-A*02-restricted
CD8?T cell epitopes were identified, and the responses to the
variants from each serotype were studied by staining for intracel-
lular IFN-?, TNF-?, MIP-1?, and analyzing proliferation by
CFSE. Stimulation with epitope variants resulted in quantitative
and qualitative changes in proliferative and cytokine responses.
The presence of such epitope-specific cross-reactive cells with dif-
ferent (serotype-dependent) effector function potential during in
vitro secondary response offers a mechanistic insight into in vivo
immunopathogenesis during secondary dengue infection.
Materials and Methods
Peripheral blood samples were obtained from four volunteers who had
been immunized with experimental live-attenuated dengue vaccines (Table
I) (25–28). A flavivirus-naive Massachusetts resident also was included in
the study. PBMC were isolated by Ficoll-Hypaque and density gradient
centrifugation. They were resuspended at 107/ml in RPMI 1640 containing
20% FBS and 10% DMSO and cryopreserved until use. HLA typing of
these donors done at the tissue typing center at the University of Massa-
chusetts Medical Center (Worcester, MA) identified all the donors as
The amino acid sequence of serotype 3 H87 strain (GenBank accession no.
NC 001475) was initially chosen to generate a list of nonamer epitopes
based on the peptide binding motif for HLA-A*0201 using the epitope
prediction algorithm (available online at ?http://bimas.dcrt.nih.gov/molbio/
hla?bind?). Twenty-five sequences with the highest binding scores were
chosen, and the peptides were synthesized at the University of Massachu-
setts Medical School Peptide Core Facility. The four identified epitopes
were identical in D3 CH53489 (GenBank accession no. AF017733, non-
structural; and GenBank accession no. M86733, structural). Variant pep-
tides were synthesized based on the D1 45AZ5 (GenBank accession no.
NC?001477), D2 NGC (GenBank accession no. M29095), and D4 814669
(GenBank accession no. AF326573).
Intracellular cytokine staining and flow cytometry
PBMC were thawed and washed with RPMI 1640 containing 10% heat-
inactivated human AB serum. Cells (106) were stimulated with 10 ?g/ml
peptide and incubated at 37°C for a total of 6 h with Golgi Plug (Brefeldin
A; BD Pharmingen) added during the last 5 h. PBMC were washed with
FACS buffer (2% FBS and 0.1% sodium azide in PBS) and incubated with
50 ?g/ml mouse IgG for 15 min at 4oC to block nonspecific binding to IgG
Fc receptors. Cells were then stained for surface markers such as CD3
(PE-Cy7 Ab; Caltag Laboratories) and CD8 (PerCP Ab; BD Biosciences)
at 4°C for 30 min. After washing with FACS buffer, cells were fixed and
permeabilized with the Cytofix/Cytoperm reagent (BD Pharmingen) for 20
min at 4°C. This was followed by a wash in PermWash buffer (BD Pharm-
ingen) and intracellular staining with Abs for CD69 (PE), IFN-? (FITC),
TNF-? (allophycocyanin), MIP-1? (PE) at 4°C for 30 min. All Abs were
obtained from BD Pharmingen. Data were acquired on a FACSAria ma-
chine (BD Biosciences) and analyzed using versions 4.2, 4.5, or 6.1 of
FlowJo software (Tree Star). The lymphocyte gate included side-scatter
and forward-scatter low populations that were selected for CD3?and
CD8?cells. The number of events collected for each sample varied be-
tween 150,000 and 250,000, depending on the donor. The number of
CD3?CD8?analyzed in each of the donors are as follows: donor 1,
20,000 ? 2,000; donor 2, 27,000 ? 2,000; donor 3, 44,000 ? 3,000; donor
4, 60,000 ? 3,000. Each experiment included isotype controls for each
cytokine-Ab conjugate. Based on these, gates were first set on the medium-
stimulated panel, and the same gating was then applied to all other samples
within each donor.
CFSE staining and bulk culture of PBMC
PBMC were thawed and resuspended in RPMI 1640 medium at 10 ?
106/ml. Cells were stained with CFSE at 1 ?g/ml for 25 min at 37°C in the
dark. Cells were washed three times with cold RPMI 1640 medium con-
taining 10% FBS, stimulated with 10 ?g/ml peptide and 25 ng/ml IL-7, and
incubated in the dark. IL-2 (50 U/ml) was added on day 3. A total of 3 ?
105bulk culture cells were mixed with 7 ? 105autologous B lymphoblas-
toid cell lines and restimulated with 10 ?g/ml peptides on day 5 and day
14 and stained for surface markers (CD3 and CD8) and intracellular cy-
tokines (IFN-?-allophycocyanin, TNF-?-AlexaFluor 610, MIP-1?-PE) as
Identification of HLA-A*0201-restricted epitopes from PBMC of
To identify dengue virus-specific T cell epitopes for our study, we
generated a panel of 25 9-mer peptides that fit the HLA-A*0201
binding motif and used these to stimulate PBMC from a dengue
3-immune HLA-A*02 donor. Frequencies of ?0.1% of CD69?
IFN-?-producing cells were detected in the CD3?CD8?T cell
population in response to stimulation by four of the peptides, with
the highest response being 0.45% to the NS4b 2353 peptide (Fig.
1A). The frequency of cells specific for NS4b 2353, NS4b 2423,
and NS4a 2148 were higher than the frequency of cells specific for
two previously identified HLA-B*62 epitopes in this donor (19).
Responses to these four peptides also were detected in three other
donors who were vaccinated with other dengue serotypes (Table
II). No responses were detected in a HLA-A*02-negative dengue-
immune donor or in a HLA-A*02-positive flavivirus-naive donor
(data not shown).
In the dengue 3-immune donor, all four peptides induced a
TNF-? response that was higher than the corresponding IFN-?
response, suggesting that the epitope-specific population was com-
prised of subpopulations of cells that secreted TNF-? but not
IFN-?. In addition to IFN-?, TNF-?, and CD69 up-regulation,
three of the four peptides also induced proliferation responses as
measured by CFSE dilution (Fig. 1B). Cells that underwent more
than six divisions over 5 days ranged from 5 to 16% of
CD3?CD8?T cells remaining in the culture, which indicates that
not all cytokine-producing cells respond by proliferating.
Table I. Vaccination history and class I HLA alleles of study subjects
Class I HLA Type
2818ALTERED FUNCTIONAL RESPONSE TO DENGUE VIRUS EPITOPE VARIANTS
by guest on June 13, 2013
Altered patterns of IFN-?, TNF-?, and MIP-1? coexpression
following heterologous stimulation by variant epitopes
Because of variation in amino acid sequences between the four
dengue serotypes (Table III) and the evidence from initial studies
that vaccinated individuals responded to heterologous serotypes,
we next investigated the range of cytokine production among each
epitope-specific T cell population and determining whether these
expression profiles were skewed following stimulation with vari-
ant sequences of these epitopes from heterologous serotypes. Of
the four epitopes, the E493 and the NS4b 2423 epitopes are closely
conserved within all four serotypes, whereas the NS4a 2148 and
NS4b 2353 epitopes are poorly conserved (Table III). PBMC were
stimulated with each of the peptide variants and double-stained for
IFN-?/TNF-? (Fig. 2A), and IFN-?/MIP-1? (Fig. 2B). Responses
to NS4b 2353, which was the most immunogenic epitope in all
four donors, are shown in Fig. 2. The cytokine response in each
donor was composed of IFN-??, TNF-??, IFN-??TNF-??, and
IFN-??MIP-1??cells (Fig. 2, A and B). In three of the four do-
nors, the frequencies of MIP-1??cells were higher than the cor-
responding frequencies of IFN-??and TNF-??cells. In the same
three donors, the D2 and D3 variants induced the highest frequen-
cies of IFN-?/TNF-? and IFN-?/MIP-1? cells.
The epitope sequence that was homologous to the serotype of
the donor’s vaccination did not always induce the highest re-
sponse. Although PBMC from donors 2 and 4 showed the highest
cytokine response when stimulated with their homologous pep-
tides, stimulation of PBMC from donors 1 and 3 with heterologous
peptides induced higher cytokine responses. Changes also were seen
in the pattern of cytokine expression in response to heterologous
stimulation. In donor 1, the D3 variant induced 0.76% IFN-??/MIP-
1??and 0.37% MIP-1??cells, whereas the D2 and D4 variants
skewed the response, with most of the responding cells producing
MIP-1??alone. The skewing of the cytokine profile also was appar-
ent in donor 4 where the prevalent populations in response to D2
peptide stimulation were TNF-??, IFN-??/TNF-??, and IFN-??/
MIP-1??cells. In contrast, the prevalent populations in response to
D3 peptide stimulation were IFN-??and IFN-??TNF-??cells.
Triple staining for IFN-?, TNF-?, and MIP-1?: variant peptides
alter both the size of the cytokine response as well as the relative
frequencies of the various heterogeneous cytokine subsets
Because the initial double-staining experiments revealed the exis-
tence of distinct subsets of epitope-specific cells that were either
single-positive for each cytokine or double-positive for IFN-?/
TNF-? and IFN-?/MIP-1?, we wanted to further define the het-
erogeneity of these populations by checking for coexpression of
TNF-? and MIP-1?. An example of the gating analysis of staph-
ylococcal enterotoxin B-stimulated samples stained simultaneously
with IFN-?, TNF-?, and MIP-1? is shown in Fig. 3. Seven dif-
ferent populations of cytokine-producing cells (IFN-??, TNF-??,
epitopes. PBMC from donor 2 (D3 immune) were stimulated for 6 h and
analyzed by FACS as described in Materials and Methods. A, Contour
plots show intracellular staining for IFN-?, TNF-?, and CD69 in the
CD3?CD8?lymphocyte population. B, PBMC from donor 2 were stained
with 1?M CFSE and stimulated by peptides for 5 days. Panels represent
cell division in CD3?CD8?T cells, which are further gated into cells that
have divided two to five times and cells that have undergone more than five
Identification of four novel HLA-A*0201-restricted T cell
Table II. Frequencies of IFN?-secreting epitope-specific CD8?/?T cells in PBMC of dengue-immune
% IFN-?-Producing CD69?Cellsb
aScores indicate the predicted half-time of dissociation from class I HLA molecules.
bPBMC were first gated on CD3?CD8?lymphocytes.
cDonors (Table I).
dData are representative of one of several independent experiments.
Table III. Sequences of epitopes and their heterologous variants
2423 NS4b (181–189)
aNumbers indicate the position of the epitope within each viral protein based on
the dengue 3 serotype.
2819 The Journal of Immunology
by guest on June 13, 2013
MIP-1??, IFN-??TNF-??, IFN-??MIP-1??, TNF-??MIP-1??,
and IFN-??TNF-??MIP-1??) were identified (Fig. 3, D–I). Re-
sults of such analyses were highly reproducible. In one individual
(donor 2) with large numbers of PBMC available, the analysis was
repeated in three separate experiments, and in almost all cases, the
SD was ?0.03%; the pattern of responses across serotypes was
consistent in all repeated experiments. However, not all of the
epitopes and their variants induced all seven cytokine subsets (Fig.
4). For example, stimulation of D1-immune PBMC (donor 1) with
the D3 NS4b 2353 resulted in all seven functional populations,
with the IFN-??MIP-1??subset being the largest. In contrast, the
D1 NS4b 2353 peptide induced IFN-??, TNF-??, MIP-1??, and
a very small population of TNF-??MIP-1??cells (Fig. 4A). Sim-
ilarly, in donor 2, although the homologous D3 NS4b 2423 peptide
induced six different subsets of cytokine-producing cells, stimula-
tion of D3 immune PBMC with the other variants of this epitope
only gave rise to mainly IFN-??and MIP-1??cells (Fig. 4B).
Thus, heterologous peptide stimulation changed not only the size
of the response but also altered its qualitative nature.
Other donor-specific patterns of cytokine production were also
observed. For example, in donor 2, except for D3 variants of the
NS4b peptides and D2 NS4a 2148, the response to the other vari-
ants of all four epitopes are mediated either by IFN-??or MIP-
1??cells or both. In contrast, the responses to all variants of all the
epitopes in donor 3 were mediated by all seven subsets and the
frequency of cells producing MIP-1? were lower than the frequen-
cies of cells producing IFN-? and TNF-? for all epitope variants.
When we looked at total cytokine responses induced by each
variant, we found that D3 NS4b 2423 and D2 NS4a 2148 variants
induced the highest responses in all four donors irrespective of
their primary immunization serotype (Fig. 4, B and C). However,
the profile of the response to each variant differed in each donor.
Thus, the heterologous response to D3 NS4b 2423 in donors 1, 3,
and 4 were dominated by MIP-1? or TNF-?, while the homolo-
gous response to this peptide in donor 2 was dominated by MIP-1?
and IFN-?. Although the heterologous response to D2 NS4a 2148
in donors 1 and 2 were dominated by IFN-?-producing cells, the
lowing stimulation with variant sequences of NS4b
2353. A, PBMC from all four donors were stimulated
with all four variants of the NS4b 2353 epitope for 6 h.
Contour plots show simultaneous intracellular staining
for IFN-? and TNF-? in the CD3?CD8?population. B,
Double intracellular staining for IFN-? and MIP-1?.
Heterogeneous cytokine expression fol-
2820 ALTERED FUNCTIONAL RESPONSE TO DENGUE VIRUS EPITOPE VARIANTS
by guest on June 13, 2013