Antigen kinetics determines immune reactivity.
ABSTRACT A current paradigm in immunology is that the strength of T cell responses is governed by antigen dose, localization, and costimulatory signals. This study investigates the influence of antigen kinetics on CD8 T cell responses in mice. A fixed cumulative antigen dose was administered by different schedules to produce distinct dose-kinetics. Antigenic stimulation increasing exponentially over days was a stronger stimulus for CD8 T cells and antiviral immunity than a single dose or multiple dosing with daily equal doses. The same was observed for dendritic cell vaccination, with regard to T cell and anti-tumor responses, and for T cells stimulated in vitro. In conclusion, stimulation kinetics per se was shown to be a separate parameter of immunogenicity. These findings warrant a revision of current immunization models and have implications for vaccine development and immunotherapy.
- [Show abstract] [Hide abstract]
ABSTRACT: Vaccines aim to protect against or treat diseases through manipulation of the immune response, promoting either immunity or tolerance. In the former case, vaccines generate antibodies and T cells poised to protect against future pathogen encounter or attack diseased cells such as tumours; in the latter case, which is far less developed, vaccines block pathogenic autoreactive T cells and autoantibodies that target self tissue. Enormous challenges remain, however, as a consequence of our incomplete understanding of human immunity. A rapidly growing field of research is the design of vaccines based on synthetic materials to target organs, tissues, cells or intracellular compartments; to co-deliver immunomodulatory signals that control the quality of the immune response; or to act directly as immune regulators. There exists great potential for well-defined materials to further our understanding of immunity. Here we describe recent advances in the design of synthetic materials to direct immune responses, highlighting successes and challenges in prophylactic, therapeutic and tolerance-inducing vaccines.Nature Material 10/2013; 12(11):978-90. · 36.43 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Many physicochemical characteristics significantly influence the adjuvant effect of micro/nanoparticles; one critical factor is the kinetics of antigen exposure to the immune system by particle-adjuvanted vaccines. Here, we investigated how various antigen-nanoparticle formulations impacted antigen exposure to the immune system and the resultant antigen-specific immune responses. We formulated antigen with poly(lactic-co-glycolic acid) (PLGA) nanoparticles by encapsulating antigen within nanoparticles or by simply mixing soluble antigen with the nanoparticles. Our results indicated that the combined formulation (composed of antigen encapsulated in nanoparticles and antigen mixed with nanoparticles) induced more powerful antigen-specific immune responses than each single-component formulation. Mice immunized with the combined vaccine formulation displayed enhanced induction of antigen-specific IgG antibodies with high avidity, increased cytokine secretion by splenocytes, and improved generation of memory T cell. Enhanced immune responses elicited by the combined vaccine formulation might be attributed to the antigen-depot effect at the injection site, effective provision of both adequate initial antigen exposure and long-term antigen persistence, and efficient induction of dendritic cell (DC) activation and follicular helper T cell differentiation in draining lymph nodes. Understanding the effect of antigen-nanoparticle formulations on the resultant immune responses might have significant implications for rational vaccine design.Biomaterials 04/2014; · 8.31 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Generation of adaptive immune response relies on efficient drainage or trafficking of antigen to lymph nodes for processing and presentation of these foreign molecules to T and B lymphocytes. Lymph nodes have thus become critical targets for new vaccines and immunotherapies. A recent strategy for targeting these tissues is direct lymph node injection of soluble vaccine components, and clinical trials involving this technique have been promising. Several biomaterial strategies have also been investigated to improve lymph node targeting, for example, tuning particle size for optimal drainage of biomaterial vaccine particles. In this paper we present a new method that combines direct lymph node injection with biodegradable polymer particles that can be laden with antigen, adjuvant, or other vaccine components. In this method polymeric microparticles or nanoparticles are synthesized by a modified double emulsion protocol incorporating lipid stabilizers. Particle properties (e.g. size, cargo loading) are confirmed by laser diffraction and fluorescent microscopy, respectively. Mouse lymph nodes are then identified by peripheral injection of a nontoxic tracer dye that allows visualization of the target injection site and subsequent deposition of polymer particles in lymph nodes. This technique allows direct control over the doses and combinations of biomaterials and vaccine components delivered to lymph nodes and could be harnessed in the development of new biomaterial-based vaccines.Journal of Visualized Experiments 01/2014;
Antigen kinetics determines immune reactivity
Pål Johansen*, Tazio Storni*, Lorna Rettig*, Zhiyong Qiu†, Ani Der-Sarkissian†, Kent A. Smith†, Vania Manolova‡,
Karl S. Lang§, Gabriela Senti*¶, Beat Mu ¨llhaupt?, Tilman Gerlach?, Roberto F. Speck**, Adrian Bot†,
and Thomas M. Ku ¨ndig*††
*Unit Experimental Immunotherapy, Department of Dermatology, University Hospital Zurich, Gloriastrasse 31, CH-8091 Zurich, Switzerland;†MannKind
Corporation, 28903 North Avenue Paine, Valencia, CA 91355;‡Cytos Biotechnology, Wagistrasse 25, CH-8952 Schlieren, Switzerland;§Institute of
Experimental Immunology, University Hospital Zurich, Schmelzbergstrasse 12, CH-8091 Zurich, Switzerland;¶Clinical Trials Center, University
Hospital Zurich, Gloriastrasse 31, CH-8091 Zurich, Switzerland;?Department of Gastroenterology and Hepatology, University Hospital Zurich,
Raemistrasse 100, CH-8091 Zurich, Switzerland; and **Division of Infectious Diseases and Hospital Epidemiology, University Hospital Zurich,
Raemistrasse 100, CH-8091 Zurich, Switzerland
Edited by Tak Wah Mak, University of Toronto, Toronto, ON, Canada, and approved January 15, 2008 (received for review July 5, 2007)
A current paradigm in immunology is that the strength of T cell
responses is governed by antigen dose, localization, and costimu-
latory signals. This study investigates the influence of antigen
kinetics on CD8 T cell responses in mice. A fixed cumulative antigen
dose was administered by different schedules to produce distinct
dose-kinetics. Antigenic stimulation increasing exponentially over
days was a stronger stimulus for CD8 T cells and antiviral immunity
than a single dose or multiple dosing with daily equal doses. The
same was observed for dendritic cell vaccination, with regard to T
cell and anti-tumor responses, and for T cells stimulated in vitro. In
conclusion, stimulation kinetics per se was shown to be a separate
current immunization models and have implications for vaccine
development and immunotherapy.
antigen presentation ? antiviral immunity ? CD8 T cell responses ?
adopted to enhance vaccine efficacy. For example, vaccines are
often delivered as emulsions or particles with comparable di-
mensions to pathogens (4), and pathogen-associated molecular
patterns (PAMPs) stimulating the immune system via Toll-like
receptors (TLR) are used as adjuvants to activate antigen-
presenting cells (4, 5). Another hallmark of pathogens is that
replication exposes the immune system to increasing amounts of
antigen and PAMPs.
The present study investigates whether the dose-kinetics of
antigen is a separate parameter of immunogenicity. Mice were
immunized with a fixed cumulative dose of antigenic peptides
and CpG ODN or with antigen-pulsed dendritic cells (DCs), but
following different kinetics. MHC class-I binding peptides were
chosen, because their short in vivo half-life allows the production
of sharp kinetics (6, 7). When administered in a dose-escalating
fashion, the vaccines stimulated much stronger CD8 T cell
responses than when administered as a single shot or at uniform
daily doses. Thus, the immune system seems to interpret expo-
nentially increasing antigenic stimulation per se as a signal
related to pathogen replication, and consequently enhances T
he immune system has evolved to respond to pathogens
(1–3). Therefore, pathogens’ characteristics have been
Exponentially Increasing Antigenic Stimulation Enhances CD8 T Cell
Responses. Transgenic TCR318 T cells (106cells) were trans-
ferred into C57BL/6 mice, which were immunized with 125 ?g
of gp33 peptide and 12.5 nmol of CpG using different protocols
[supporting information (SI) Table 1 and Fig. 1C]. Mice infected
with 250 pfu of lymphocytic choriomeningitis virus (LCMV)
served as a positive control. On day 6 (Fig. 1A), CD8 T cell
responses were quantified by intracellular IFN-? staining in
blood. Exponentially increasing immunizations produced a
stronger response than uniform daily doses of gp33 and CpG
(P ? 0.0001). If either one of the vaccine components was given
as a single dose, the efficacy was weak but still significant
compared with the naive control (2.23 ? 0.84% vs. 0.19 ? 0.12%
P ? 0.02); day 6 represents the peak of the immune, and the
response retracted after day 12 (data not shown). Also in
of IFN-?-producing CD8 T cells (2.1 ? 0.4%) compared with
other vaccination protocols (P ? 0.008), which barely induced
detectable frequencies. Similar observations were made in HLA
A2.1 transgenic mice immunized with the influenza matrix
peptide (data not shown).
Compared with single bolus injection, exponential vaccination
over 4–8 days (Fig. 2A), significantly enhanced the immune
response at its peak, which was 4–7 days after the last injection
(Fig. 2B). Four weeks later, there was no significant difference
in the number of CD44-positive resting memory cells between
the different groups (Fig. 2C). T helper (Th) epitopes can be
crucial for functional CD8 T cell immunity (8–10), but expo-
nential vaccination with a mixture of the class-I gp33 and
the class-II binding gp61 peptides did not affect the outcome
(data not shown), suggesting that the above findings were Th
Exponentially Increasing Stimulation Prolongs T Cell Proliferation. To
test how the kinetics of immunization affected the proliferation
of carboxyfluorescein diacetate-succinimidyl ester (CFSE)-
labeled TCR318 CD8 T cells, mice were injected with a single
dose [schedule 1 (s1)], uniform daily doses (s2), or exponentially
increasing doses (s4) of gp33 peptide and CpG. A bolus injection
of gp33 peptide and CpG triggered the cells to divide after 3 days
(Fig. 3). On day 5, precursor cells still entered division although
to a lower extent than at day 3. By day 7, the CFSE-labeled cells
had ceased to enter new divisions (data not shown). In contrast,
four equally sized or four dose-escalating vaccine doses pro-
duced much stronger stimuli for proliferation, even at day 3 after
priming, despite not yet having received the whole immunization
regime. Moreover, the division index, i.e., the average number of
division undergone, was significantly higher (P ? 0.05 by Mann–
Whitney) for s4 than for s2.
Exponentially Increasing Immunization Enhances Antiviral Immunity.
To evaluate the functional relevance of the above observations,
immunized C57BL/6 mice were challenged with LCMV or a
Author contributions: P.J. and T.S. contributed equally to this work; P.J., T.S., L.R., K.A.S.,
G.S., B.M., T.G., R.F.S., A.B., and T.M.K. designed research; P.J., T.S., L.R., Z.Q., A.D.-S., V.M.,
and K.S.L. performed research; K.A.S. and A.B. contributed new reagents/analytic tools;
P.J., T.S., L.R., Z.Q., A.D.-S., V.M., K.S.L., and T.M.K. analyzed data; and P.J., T.S., L.R., K.A.S.,
R.F.S., A.B., and T.M.K. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
††To whom correspondence should be addressed. E-mail: firstname.lastname@example.org.
This article contains supporting information online at www.pnas.org/cgi/content/full/
© 2008 by The National Academy of Sciences of the USA
April 1, 2008 ?
vol. 105 ?
no. 13 ?
LCMV glycoprotein-expressing vaccinia virus (vacc-gp) (11);
protection against both viruses depends on CD8 T cells (12).
Exponentially increasing doses of gp33 and CpG induced sig-
nificantly higher frequencies of gp33-tetramer-positive memory
(CD44hi) cells (Fig. 4A) and of IFN-?-producing effector and
memory cells (Fig. 4B) than did a bolus vaccination. On day 30,
all mice were challenged with 250 pfu of LCMV. Whereas
bolus-vaccinated mice were not significantly protected against
viral replication (Fig. 4C), exponential vaccination induced
strong protection with LCMV titers 10- to 20-fold higher than
naive or bolus-vaccinated mice (P ? 0.009). Similarly, a chal-
lenge with 1.5 ? 106pfu of vacc-gp 8 (Fig. 4D) or 24 (Fig. 4E)
days after priming revealed significantly inhibited viral replica-
tion in mice immunized with the dose-escalating protocol as
compared with s1 and s2.
Exponential Immunization Delays Maximal Activation of DCs. To test
whether the immunization kinetics affected the numbers and
activation status of DCs, mice were immunized with exponen-
tially increasing doses (s4) or by bolus injection (s1). The kinetics
type did not crucially affect the numbers of DCs in the draining
lymph nodes, or their activation status as monitored by MHC-
class II (I-Ab) and CD86 expression (Fig. 5A), but the peaks of
DC numbers and activation were separated by 2–3 days.
Because DC activation reached its maximum 1 day after the
maximal CpG dose, independent of the type of immunization
kinetics, we tested whether exponentially vaccination was ben-
eficial for CD8 T cell induction, simply because the peptide is
delivered when DCs are mostly activated. If so, administration of
a high peptide dose 1 day after a bolus CpG injection or after the
last dose of exponentially given CpG should result in a T cell
response comparable with giving both peptide and CpG in a
dose-escalating fashion. However, prestimulation with CpG
( s l l e
s3s4 s5 s6
Naive s1s2s3 s4s5s6
( s l l e
(A) or C57BL/6 WT mice (B) were s.c. immunized with identical cumulative
vaccine doses, but using the immunization schedules s1–s6 illustrated in C. s1,
of gp33 peptide and CpG; s3, exponentially decreasing doses of gp33 peptide
and CpG; s4, exponentially increasing doses of gp33 peptide and CpG; s5,
exponentially increasing doses of gp33 peptide and an initial single dose of
of CpG; naive, untreated mice; LCMV, mice immunized with 250 pfu of LCMV
i.v. on day 0. CD8 T cells were analyzed for IFN-? production after in vitro
restimulation of blood lymphocytes with gp33 peptide on day 6 (A), or day 8
(B). Values represent the means and SEM of four mice per group. The exper-
iment was repeated twice.
Exponentially increasing doses of both gp33 peptide and CpG
µ ) g
5 10 1520
cell induction. C57BL/6 mice were immunized with fixed cumulative dose of
(A) with peaks at day 0 (bolus), day 3, day 5, or day 7. At different time points
thereafter, mice were bled, and the CD44 expression and the IFN-? secretion
of CD8 T cells after restimulation in vitro with gp33 peptide were analyzed (B
and C). The FACS density blots illustrate the frequencies of CD44hiand IFN-?-
producing CD8-positive lymphocytes as measured by FACS at the peak of the
immune response, and the numbers show the mean percentage of IFN-?-
producing CD44hiCD8?T cells. The mean percentage of IFN-?-producing
CD44hiCD8?T cells is also illustrated as a function of time (C). One of two
similar experiments is shown (n ? 3–4).
Four days of antigenic stimulation is necessary for optimal for CD8-T
C57BL/6 mice received by i.v. adoptive transfer 1.5 ? 106CFSE-labeled and
magnetic cell separation (MACS) selected CD8 cells from TCR318 spleens and
LNs and 1 day later were immunized s.c. with fixed cumulative vaccine doses
of gp33 peptide and CpG according to the immunization protocol s1, s2, s4 or
and analyzed for CD8 expression and CFSE staining by flow cytometry.
Exponential immunization favors prolonged T cell proliferation.
www.pnas.org?cgi?doi?10.1073?pnas.0706296105Johansen et al.
resulted in immune responses that were significantly lower than
those produced by concomitant dose-escalating gp33 and CpG
administration (Fig. 5B; P ? 0.016).
Exponentially Increasing Numbers of Peptide-Pulsed DCs Enhances
Tumor Protection. To further investigate the contribution of
APCs, C57BL/6 mice were immunized with 1.11 ? 105peptide-
pulsed DCs by different kinetics (SI Materials and Methods).
Bone-marrow derived DCs were loaded with the HPV E7
peptide, and then injected as a bolus on day 1 (s1?), or in an
increasing (s4?) pattern on days 1 (103cells), 3 (104cells), and 6
(105cells). The vaccines were administered directly into a s.c.
lymph node (LN) to ensure that all DCs were available for T cell
priming. Again, DC dose-escalation (s4?) induced higher num-
bers of antigen-specific CD8 T cells than a bolus injection of DCs
(s1?) as shown for both tetramer and IFN-? enzyme-linked
immunospot assays (ELISPOTs) on days 17 and 22, respectively
(Fig. 6A). Importantly, mice vaccinated with the dose-escalating
protocol, but not the bolus, rejected a challenge with the
HPV-transformed tumor cell line C3.43 (Fig. 6B). By the same
token, mice immunized with the s4?-protocol of DCs loaded with
the VSV np52 peptide showed improved survival after a chal-
lenge with mouse lymphoma cells EL-4 transfected to express
the VSV nucleoprotein (Fig. 6C). The survival of s4? immunized
was also significant better than mice immunized according to the
s2? protocol with DCs given in uniform numbers on all 3 days
(P ? 0.0084).
Exponentially Increasing Antigenic Stimulation Enhanced IL-2 Produc-
tion of T Cells: The Impact of Antigen Kinetics Occurs at the Level of
T Cell Clones. We next investigated whether the above observa-
tions could be explained at the level of a T cell clone, or whether
they were the result of in vivo T cell selection processes, involving
T cell clonotypes of different affinity, avidity and functionality.
Splenocytes from TCR318 mice were stimulated in vitro with
10?9M gp33, either as a bolus on day 0, with exponentially
increasing or decreasing doses over four days, or with a uniform
dose every day; the splenocytes also contain macrophages, DCs
and B cells. IL-2, IL-10, and IFN-? were determined daily in
supernatants, and on day 6, cytotoxic T lymphocyte (CTL)
activity was determined in a51Cr release assay (Fig. 7). Expo-
nentially increasing antigen doses induced the strongest CTL
response, followed by daily administrations of a uniform dose,
whereas antigen given as a bolus or in a decreasing dose profile
generated weaker responses. The difference was even greater,
when cells were stimulated with one tenth of the peptide doses
(data not shown). The cytotoxicity correlated with the produc-
tion of IL-2. Uniform daily stimulation induced high amounts of
IL-10 with an earlier onset as compared with exponentially
increasing stimulation. IFN-? was transiently produced at an
amounts of antigen. In contrast, daily stimulation by uniform or
exponentially increasing doses induced higher amounts of
Naive s1 s4 LCMV
( s l l e
c T 8
Days after immunization
( i h
Naive s1 s2 s3s4LCMV
Naive s1 s4 LCMV
u f p
e t i t
u f p 0
o l ( r
e t i t -
and LCMV). On day 30 (C), the mice were challenged i.p. with 250 pfu of LCMV (C). Alternatively, the mice were challenged on day 8 (D) or 24 (E) with 1.5 ? 106pfu
of vacc-gp. Four or five days later, spleens or ovaries were harvested for determination of LCMV or vaccinia titers, respectively.
I I -
s r e
0 2 4 6 80 2 4 6 8
( s l l e
c T 8
Vaccination schedule (4 days)
s1 (bolus injection) and s4 (exponentially increasing doses) as described in Fig. 1.
cytometry for the expression of the DC marker CD11c, as well as CD86 and the
MHC class II marker I-Ab (A). The results in the Top are illustrated as the relative
frequency of cells expressing both CD11c and I-Ab. The Middle illustrates the
relative MFI of CD86 expression and the Bottom the relative MFI of I-Ab expres-
sion on DCs. In all cases, the results are normalized to that of naive controls (day
0) for which reason the starting point is always 100. Mice were also immunized
with gp33 peptide and CpG according to modified protocols as illustrated (B).
group received exponentially increasing CpG doses on days 0–3 followed by a
gp33 peptide bolus on day 4. The last group received exponentially increasing
doses of gp33 peptide and CpG on days 1–4 as described above (s4). The fre-
Exponential immunization delays DC recruitment. C57BL/6 mice were
Johansen et al.PNAS ?
April 1, 2008 ?
vol. 105 ?
no. 13 ?
This study demonstrates that the kinetics of antigenic stimula-
tion is a key parameter of immunogenicity. Exponentially in-
creasing antigen doses stimulated stronger immune responses
than constant stimulation with uniform doses or immunization
with a bolus. Such dose escalation reflects the increasing
amounts of antigen associated with highly virulent pathogens
pathogens represent less danger, require less T cell activity and
are well controlled by innate immunity. Likewise, maximal T cell
responses were induced by antigenic stimulation over 4 days,
whereas rapid exponential growth of antigenic stimulation over
1–3 days induced weaker T cell responses. This observation
reflects the fact that proliferation and differentiation of T cells
takes several days, and it is impossible for the immune system to
race a pathogen that overwhelmingly infects the host within
Likely candidates for mediating this perceptive faculty are the
DCs. Whereas the different vaccination protocols did not induce
different numbers and activation levels of DCs in the lymph
node, they differed in the time required to produce the peak of
DC activation and numbers, which by dose-escalating vaccina-
tion was delayed by 3 days as compared with bolus vaccination.
Because both vaccination regimes induced peak DC activation 1
day after the maximal vaccine dose was administered, it could
imply that an optimal vaccination schedule would be a single
injection of antigen 1 day after the adjuvant injection. However,
our data demonstrates that the latter was inferior to doses of
adjuvant and antigen that escalated exponentially in parallel.
This result illustrates the importance for antigen and adjuvant to
be administered together and not separated in time (1, 3, 13).
That the observed effect was not exclusively mediated by the
number of antigen presenting DCs nor by their activation status,
was demonstrated by immunizing mice with DCs that were
pulsed with MHC class-I binding peptides ex vivo. Also here,
ules. In these experiments, DC activation was kept at the same
the same, demonstrating that the strength of the immune
response is most probably enhanced by the synchronization of
antigen-presenting cell numbers and the frequency of available
T cell precursors. Whereas the low frequency of specific T cells
during the early response can be efficiently stimulated with a low
number of antigen pulsed DCs, it seems important to restimulate
the high frequency of specific T cells during the later primary
response with a high number of DCs. Moreover, because pro-
liferation is an exponential process, it is likely that exponential
vaccination may match this process by allowing quantitative
synapses generation on available progenitor CD8 T cells (14).
Indeed, whereas dividing cells were observed for only 2–3 days
after bolus immunization, exponentially increasing stimulation
05 10 1520 25
( l a
v i v r u
Time post challenge (days)
u l o
v r o
x ( T
( r e
a r t e t 7
intralymphatic injection of bone-marrow-derived DCs loaded with HPV17 E7 peptide (A and B). The DCs were given as a bolus on day 1 (s1?) or equally (s2?) or
exponentially increasingly distributed on day 1, day 3, and day 6 (s4?). Naive mice were used as negative controls. (A) On day 17 and on day 22, the frequency
(means and SEM; n ? 7). (B) On day 21, three vaccinated mice and 10 naive mice were challenged with the HPV-transformed tumor cell line C3.43. Tumor
loaded with the VSV np53 peptide (n ? 4). DCs (1.11 ? 105) were given as a bolus on day 1 (s1?) or as equal (s2?) or dose-escalating doses (s4?) on days 1, 3, and
s2?? s1?: P ? 0.0082; s1? ? Naive, P ? 0.401.
Exponential immunization with peptide-loaded DCs induces strong T cell and anti-tumor responses. Groups of 10 C57BL/6 mice were immunized by
Specific lysis of target cells (%)
Dilution of standard culture
Cytokine concentration (pg / ml)
Cytokine concentration (pg / ml)
tion and cytotoxicity. TCR318 T cells (1 ? 105) were cocultured with 2 ? 106
irradiated syngeneic splenocytes serving as feeder cells. Cultures were stimu-
lated with the same total doses of gp33, but following different kinetics: ■, a
single dose of 10?9M at day 0; Œ, four equal doses of 0.25 ? 10?9M during 4
days; F, four exponentially decreasing doses of 10?9, 10?10, 10?11, and 10?12
M at days 0, 1, 2, and 3, respectively; and ?, four exponentially increasing
doses of 10?12, 10?11, 10?10, and 10?9M at days 0, 1, 2, and 3, respectively. ?,
control culture without gp33 peptide. (A) After 6 days, CTL activity was
measured by using gp33-pulsed EL-4 target cells in a 5-h51Cr-release assay.
Values represent means of duplicate cultures. One representative of two
similar experiments is shown. (B) Supernatants were analyzed for IL-2, IL-10,
and IFN-?. Values represent means of triplicate culture wells, and one repre-
sentative of two experiments is shown.
Exponential in vitro stimulation of CD8 T cells enhances IL-2 produc-
www.pnas.org?cgi?doi?10.1073?pnas.0706296105Johansen et al.
Furthermore, antigen persistency has been accepted to pro-
persists for hours, it has a cumulative effect that is necessary for
the maintenance of the immunological synapse, for T cell
proliferation and for IL-2 production (14). To what extent it
applies for exponential vaccination schedules is currently not
known. However, our work does not necessarily imply that
effector T cells require daylong stimulation in vivo. Because
vaccination with increasing antigen kinetics also enhances clonal
expansion of specific T cells, stimulation of the single T cell must
not be prolonged. Hence, our work is also compatible with the
temporal summation model that offers an explanation for how
signals originating from serially triggered TCRs are accumulated
and integrated over the period required for T cell activation
The production of IL-2 is a hallmark of T cell activation and
plays a key role in regulating several stages of the T cell
response. Engagement of the TCR (signal 1) and costimula-
tory molecules (signal 2) induces only limited clonal expansion
of T cells. Extensive amplification of T cells as well as
differentiation into effector cells requires signaling via the
IL-2R (signal 3) (19), and autocrine IL-2 production by CD8
T cells subsequently triggers in vivo CD8 T cell expansion (19,
20). However, IL-10 is a main inhibitor of T cell proliferation
partly via modulation of DCs (21). Consistent with these
findings, we observed that exponentially increasing antigen
doses stimulated IL-2 production in vitro more efficiently than
other regimens and that production of immunosuppressive
IL-10 occurred at a later time point compared with constant
antigenic stimulation. These phenomena have furthermore
shown to be accompanied by higher T cell avidity, which again
is crucial for efficient interaction between T cells and DCs
(22). Thus, these in vitro observations may explain our in vivo
findings and also suggest that, at a clonal level, T cells are
capable of decoding the kinetics of antigen exposure.
The more an antigen resembles a virulent pathogen, the more
immunogenic it is likely to be. High antigen doses (23–25) and
the presence of antigen in lymphoid organs (26, 27), both
mimicking widespread replication of a virulent pathogen, induce
strong immune responses. Particulate antigens that resemble the
structure of viruses or bacteria induce stronger immune re-
sponses than soluble antigens (4, 28, 29). Finally, presentation of
an antigen together with pathogen components such as bacterial
DNA, lipopolysaccharide or viral RNA strongly enhances the
immune response. Based on this study, it seems that exponen-
tially increasing antigenic stimulation (a further hallmark of
virulent pathogens) is also recognized by the immune system as
a pattern associated with pathogens, driving strong immune
The presented findings offer an additional explanation for the
observation that live attenuated vaccines usually induce strong
and long lasting immune responses after only one injection, such
that many viral vaccines of this type have efficacies ?90% (30).
In contrast, vaccines consisting of killed microorganisms, toxins,
subunit vaccines including peptide vaccines, or naked DNA
vaccines, are of considerably lower efficacy and boosting immu-
nizations are essential. Whereas live vaccines produce increasing
antigen doses that call for strong immune responses, nonrepli-
cating vaccines produce a decreasing antigen profile, which we
demonstrate here to be a weak stimulus of T cells. Because the
trend in vaccine development goes toward subunit vaccines that
are safer than live vaccines, more consideration must be given to
the dose-kinetics of antigenic stimulation.
Current strategies to improve the efficiency of vaccination aim
at increasing the duration of antigen presentation (31–40). As
also shown in this study, constant daily antigenic stimulation
enhanced the CD8 T cell response when compared with single
shot administration. However, our findings show that optimal T
cell induction requires an exponentially increasing immunization
regimen. Therefore, a vaccine should not be administered as one
single bolus (a ‘‘shot’’) or in a depot formulation, but rather in
a dose escalating fashion over several consecutive days. This
strategy could be used for immunotherapeutic approaches to
enhance T cell responses against chronic infectious diseases or
Materials and Methods
Mice. Female C57BL/6 mice were purchased from Harlan or The Jackson
Laboratory at 6–12 weeks of age. TCR318 transgenic mice expressing a T cell
receptor specific for the lymphocytic choriomeningitis virus (LCMV) glycopro-
tein H-2Dbepitope aa33-41 (gp33) on a C57BL/6 background were obtained
from M. F. Bachmann and R. M. Zinkernagel (41). HHD transgenic mice
expressing HLA A2.1 were originally obtained from the MannKind Corpora-
tion (42) and were bred and kept in a SPF facility at the University Hospital
Zurich according to Swiss guidelines.
Viruses, Peptides, and Oligodeoxynucleotides. LCMV isolate WE titers were
determined by using a focus-forming assay on MC57 fibroblasts (43).
Recombinant vaccinia virus expressing the LCMV glycoprotein (vacc-gp)
(44) was grown and plaqued on BSC40 cells (45). LCMV peptides gp33
(KAVYNFATM) and gp61 (GLNGPDIYKGVYQFKSVEFD) and VSV peptide
np52 (SDLRGYVYQGLKSG) were purchased from EMC Microcollections
(Tu ¨bingen, Germany). Influenza matrix peptide (GILGFVFTL) was obtained
from Neosystems (Strasbourg, France). The HPV16 E7 (aa49-57; RA-
HYNIVTF) peptide used was synthesized at MannKind Corporation (Valen-
cia, CA) to ?99% purity. Phosphorothioate-modified CG-rich oligode-
oxynucleotide CpG 1668 (5?-TCC ATG ACG TTC CTG AAT AAT-3?) was
synthesized by Microsynth (Balgach, Switzerland). CpG ODN was chosen as
adjuvant because it strongly enhances CD8 T cell responses (46, 47) and is
cleared from plasma with a half-life of 30–60 min (48). In tissue, CpG ODN
is relatively stable with a half-life of 48 h (49).
to s6 were designed to deliver the same cumulative dose of 125 ?g of peptide
and 12.5 nmol of CpG over a time frame of 1–4 days (SI Table 1). Schedules 3
(s3) and 4 (s4) follow an exponentially decreasing or increasing pattern at
5-fold dilution steps. In some experiments, C57BL/6 mice received 106sex-
matched TCR318 cells to increase precursor T cell frequencies and facilitate
assessment of the immune response. One day later, the recipients were s.c.
vaccinated in the neck region with peptide and CpG at the indicated doses.
Assessment of Antiviral Immunity in Vivo. Vaccinated mice were infected with
titers were determined on BSC 40 cells (45). Alternatively, the mice were
on MC57 cells (43).
Cytotoxicity Assay and Cytokine Secretion Analysis. TCR318 transgenic T cells
(105) were cultured with syngeneic irradiated feeder cells (2 ? 106cells per
well; 2,000 rad) for 6 days in 24-well plates and pulsed with the indicated
amounts of gp33 peptide. Effector cells were then resuspended in 300 ?l of
Radioactivity in cell culture supernatants was measured with a Cobra II
Counter (Canberra Packard, Downers Growe, IL). Nonradioactive culture su-
pernatants were assessed daily for IFN-?, IL-2, and IL-10 concentrations using
bead-multiplex-assays and flow cytometry.
Statistical Analysis. Nonparametric or nonnormally distributed data were
comparison of Kaplan Meier survival curves was performed by using log rank
ACKNOWLEDGMENTS.We thank R. M. Zinkernagel and M. F. Bachmann for
comments and for providing virus, gp33 tetramers, and TCR318 mice. The
HPV transformed tumor cell line C3.43 was provided by W. Martin Kast
(Loyola University, Chicago, IL). We also thank Adrian Urwyler and Anna
Flace for technical assistance and Nicole Graf for help with statistics. This
project was supported by Swiss National Science Foundation Grant SNF
Johansen et al. PNAS ?
April 1, 2008 ?
vol. 105 ?
no. 13 ?
1. Janeway CA, Jr (1989) Cold Spring Harb Symp Quant Biol 54:1–13.
2. Zinkernagel RM (1996) Science 271:173–178.
3. Germain RN (2004) Nat Med 10:1307–1320.
4. O’Hagan DT, Valiante NM (2003) Nat Rev Drug Discov 2:727–735.
5. Johansen P, Senti G, Martinez Gomez JM, Storni T, von Beust BR, Wuthrich B, Bot A,
Kundig TM (2005) Clin Exp Allergy 35:1591–1598.
6. Falo LD, Jr, Colarusso LJ, Benacerraf B, Rock KL (1992) Proc Natl Acad Sci USA
7. Widmann C, Maryanski JL, Romero P, Corradin G (1991) J Immunol 147:3745–3751.
8. Johansen P, Stamou P, Tascon RE, Lowrie DB, Stockinger B (2004) Eur J Immunol
9. Shedlock DJ, Shen H (2003) Science 300:337–339.
10. Sun JC, Bevan MJ (2003) Science 300:339–342.
11. Kaech SM, Wherry EJ, Ahmed R (2002) Nat Rev Immunol 2:251–262.
12. Binder D, Kundig TM (1991) J Immunol 146:4301–4307.
13. Schwartz RH (1992) Cell 71:1065–1068.
14. Huppa JB, Gleimer M, Sumen C, Davis MM (2003) Nat Immunol 4:749–755.
15. Gray D (2002) Nat Rev Immunol 2:60–65.
16. Iezzi G, Karjalainen K, Lanzavecchia A (1998) Immunity 8:89–95.
17. Rachmilewitz J, Lanzavecchia A (2002) Trends Immunol 23:592–595.
18. van Stipdonk MJ, Lemmens EE, Schoenberger SP (2001) Nat Immunol 2:423–429.
19. Malek TR, Bayer AL (2004) Nat Rev Immunol 4:665–674.
20. D’Souza WN, Schluns KS, Masopust D, Lefrancois L (2002) J Immunol 168:5566–5572.
21. Moore KW, de Waal Malefyt R, Coffman RL, O’Garra A (2001) Annu Rev Immunol
22. Bousso P, Robey E (2003) Nat Immunol 4:579–585.
23. Mitchison NA (1964) Proc R Soc London B Biol Sci 161:275–292.
24. Weigle WO (1973) Adv Immunol 16:61–122.
25. Nossal GJ (1983) Annu Rev Immunol 1:33–62.
26. Zinkernagel RM (2000) Semin Immunol 12:163–171.
27. Zinkernagel RM, Hengartner H (2001) Science 293:251–253.
28. Bachmann MF, Rohrer UH, Kundig TM, Burki K, Hengartner H, Zinkernagel RM (1993)
29. Storni T, Kundig TM, Senti G, Johansen P (2005) Adv Drug Deliv Rev 57:333–355.
30. Nossal G (1999) in Fundamental Immunology, ed Paul WE (Lippincot-Raven, Philadel-
phia), pp 1387–1425.
31. Lofthouse S (2002) Adv Drug Deliv Rev 54:863–870.
32. Ehrenhofer C, Opdebeeck JP (1995) Vet Parasitol 59:263–273.
33. Guery JC, Galbiati F, Smiroldo S, Adorini L (1996) J Exp Med 183:485–497.
34. Zhu G, Mallery SR, Schwendeman SP (2000) Nat Biotechnol 18:52–57.
35. Borbulevych OY, Baxter TK, Yu Z, Restifo NP, Baker BM (2005) J Immunol 174:4812–
36. Levitsky V, Zhang QJ, Levitskaya J, Masucci MG (1996) J Exp Med 183:915–926.
37. Chen JL, Stewart-Jones G, Bossi G, Lissin NM, Wooldridge L, Choi EM, Held G, Dunbar
PR, Esnouf RM, Sami M, et al. (2005) J Exp Med 201:1243–1255.
38. Rivoltini L, Squarcina P, Loftus DJ, Castelli C, Tarsini P, Mazzocchi A, Rini F, Viggiano V,
Belli F, Parmiani G (1999) Cancer Res 59:301–306.
HS, Engelhard VH, Slingluff CL, Jr (1999) Int J Cancer 83:326–334.
40. Ayyoub M, Mazarguil H, Monsarrat B, Van den Eynde B, Gairin JE (1999) J Biol Chem
41. Pircher H, Burki K, Lang R, Hengartner H, Zinkernagel RM (1989) Nature 342:559–561.
42. Pascolo S, Bervas N, Ure JM, Smith AG, Lemonnier FA, Perarnau B (1997) J Exp Med
43. Battegay M, Cooper S, Althage A, Banziger J, Hengartner H, Zinkernagel RM (1991)
J Virol Methods 33:191–198.
44. Bachmann MF, Kundig TM, Freer G, Li Y, Kang CY, Bishop DH, Hengartner H, Zinker-
nagel RM (1994) Eur J Immunol 24:2228–2236.
45. Kundig TM, Castelmur I, Bachmann MF, Abraham D, Binder D, Hengartner H, Zinker-
nagel RM (1993) J Virol 67:3680–3683.
MF (2003) Eur J Immunol 33:1465–1470.
47. Krieg AM (2002) Annu Rev Immunol 20:709–760.
48. Farman CA, Kornbrust DJ (2003) Toxicol Pathol 31(Suppl):119–122.
49. Mutwiri GK, Nichani AK, Babiuk S, Babiuk LA (2004) J Control Release 97:1–17.
(1993) J Immunol 150:4450–4456.
www.pnas.org?cgi?doi?10.1073?pnas.0706296105 Johansen et al.