The Journal of Immunology
MHC Class II-Associated Invariant Chain Linkage of Antigen Dramatically Improves Cell-Mediated Immunity Induced by Adenovirus Vaccines1 Peter Johannes Holst,* Maria Rathmann Sorensen,* Camilla Maria Mandrup Jensen,* Cathrine Orskov,† Allan Randrup Thomsen,2* and Jan Pravsgaard Christensen* The ideal vaccine induces a potent protective immune response, which should be rapidly induced, long-standing, and of broad specificity. Recombinant adenoviral vectors induce potent Ab and CD8ⴙ T cell responses against transgenic Ags within weeks of administration, and they are among the most potent and versatile Ag delivery vehicles available. However, the impact of chronic infections like HIV and hepatitis C virus underscore the need for further improvements. In this study, we show that the protective immune response to an adenovirus-encoded vaccine Ag can be accelerated, enhanced, broadened, and prolonged by tethering of the rAg to the MHC class II-associated invariant chain (Ii). Thus, adenovirus-vectored vaccines expressing lymphocytic choriomeningitis virus (LCMV)-derived glycoprotein linked to Ii increased the CD4ⴙ and CD8ⴙ T cell stimulatory capacity in vitro and in vivo. Furthermore, mice vaccinated with a single dose of adenovirus-expressing LCMV-derived glycoprotein linked to Ii were protected against lethal virus-induced choriomeningitis, lethal challenge with strains mutated in immunodominant T cell epitopes, and systemic infection with a highly invasive strain. In therapeutic tumor vaccination, the vaccine was as efficient as live LCMV. In comparison, animals vaccinated with a conventional adenovirus vaccine expressing unmodified glycoprotein were protected against systemic infection, but only temporarily against lethal choriomeningitis, and this vaccine was less efficient in tumor therapy. The Journal of Immunology, 2008, 180: 3339 –3346.
T
he world depends critically on improved vaccine systems to combat intractable infectious diseases and tumors. So far, adenoviral vectors have emerged as potent vehicles for delivery of vaccine Ags and for induction of CD8⫹ T cell and Ab responses, but the challenges of chronic infections, such as those caused by HIV and hepatitis C virus (HCV),3 have highlighted the need for further improvements (1, 2). Additionally, problems remain concerning high levels of pre-existing immunity to adenoviral vectors derived from previous encounters with wildtype (WT) adenoviruses (3). DNA priming before boosting with adenovirus, shifting to different serotypes of adenovirus or modifications of the adenoviral capsid proteins, have addressed these *Institute of International Health, Immunology and Microbiology, The Panum Institute, Copenhagen, Denmark; and †Institute of Medical Anatomy, The Panum Institute, Copenhagen, Denmark Received for publication August 16, 2007. Accepted for publication December 27, 2007. 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. 1 This work was supported in part by the Novo Nordisk Foundation, the Danish Cancer Society, the Lundbeck Foundation, the Sophus C. E. Friis Foundation, the Foundation for the Advancement of Medical Sciences, Aase and Ejnar Danielsen Foundation, The Hede Nielsen Family Foundation, The Leo Nielsen Foundation, Merchant Foght’s Foundation, the Christian and Ellen Larsen Foundation, and The Leo Pharma Research Foundation. P.J.H. and M.R.S. are the recipients of Research Fellowships from the Faculty of Health Science, University of Copenhagen. 2 Address correspondence and reprint requests to Dr. Allan Randrup Thomsen, Institute of International Health, Immunology, and Microbiology. The Panum Institute, Building 22.5.18, Blegdamsvej 3c, DK-2200 Copenhagen, Denmark. E-mail address:
[email protected]
challenges (4, 5). However, the use of boosted immunizations, though capable of inducing high numbers of virus-specific T cells, tends to drive the differentiation of responding lymphocytes toward a committed effector memory phenotype that may be inappropriate for protection against chronic infections, and may result in a narrowing of the ensuing T cell response (6 – 8). Accordingly, there is a need to generate broad and qualitatively improved memory from one or a limited number of immunizations. We have recently found that immune responses to Ags expressed by the use of adenovirus vectors could benefit from modification of the encoded Ag (9). In the present study, we attempted to increase Ag presentation to CD4⫹ T cells, as their activation is normally limited after adenoviral vaccination and triggering of this subset is critical for the control of chronic infections and required for an optimal response to unmodified adenoviral Ag (9, 10). To increase CD4⫹ T cell responses, we chose to link Ag to the MHC class II-associated invariant chain (Ii) as the ability of Ii to increase presentation in the context of MHC class II is well-established (11–15). This strategy was found to accelerate, enhance, and prolong vaccine-induced protection against acute or chronic viral infection. Additionally, this vaccine provides efficient protection against viral escape variants due to an increased breadth of the vaccine-induced CD8⫹ T cell response. In therapeutic vaccination against melanomas expressing a viral neoantigen, the modified vaccine was as efficient as live lymphocytic choriomeningitis virus (LCMV).
Materials and Methods Mice
3
Abbreviations used in this paper: HCV, hepatitis C virus; Ii, invariant chain; LCMV, lymphocytic choriomeningitis virus; NP, nucleoprotein; WT, wild type; DC, dendritic cell; BMDC, bone marrow-derived DC; MOI, multiplicity of infection; i.c., intracerebral(ly). Copyright © 2008 by The American Association of Immunologists, Inc. 0022-1767/08/$2.00 www.jimmunol.org
C57BL/6 and matched MHC class II-deficient mice were obtained from Taconic Farms. C57BL/6.SJL mice were bred locally from breeder pairs originally obtained from The Jackson Laboratory. Transgenic mice (TCR318 and SMARTA) expressing a TCR for LCMV gp33– 41 or gp61– 80, respectively, were the progeny of breeding pairs provided by H.
3340 Pircher, A. Oxenius, and R. M. Zinkernagel (University of Zu¨rich, Zu¨rich, Switzerland). All mice used in this study were between 7 and 10 wk old and housed in a specific pathogen-free facility. All experimental procedures were approved by the local animal ethics council and performed according to local experimental guidelines.
Adenoviral vectors and vaccination LCMV glycoprotein, Ii glycoprotein, LCMV nucleoprotein (NP), IiNP, and LCMV glycoprotein fused with the C-terminal LAMP-1-sorting motif and transmembrane region (glycoprotein Lamp-1) were amplified by single or overlapping PCR from plasmid or cDNA templates and cloned into a pacCMV-based shuttle vector (Ii plasmid was a gift from N. Koch (University of Bonn, Bonn, Germany); plasmid expressing the LCMV glycoprotein and NP was a gift from M. B. Oldstone (The Scripps Research Institute, La Jolla, CA); LAMP-1 was amplified from splenic cDNA). From all shuttle plasmids, human type 5 recombinant adenovirus vectors were then produced from homologous recombination by standard methods (16). After purification, adenoviral stocks were immediately aliquoted and frozen at ⫺80°C in 10% glycerol, and the infectivity of adenovirus stocks was determined with the Adeno-X Rapid Titer kit (BD Clontech). Mice to be vaccinated were anesthetized and injected with 2 ⫻ 107 HEK293 infectious units in the right footpad.
Virus infection Mice were infected intracerebrally (i.c.) with 20 PFU of LCMV Armstrong clone 53b, Armstrong gp33 nil, Armstrong gp276 nil, Armstrong gp33/ gp276 nil, or Armstrong NP396 nil (17); for i.v. infection, mice were inoculated with 105 PFU of highly invasive LCMV Armstrong clone 13. For infection of bone marrow-derived dendritic cells (BMDCs) with adenovirus, the infectability of the cells was first determined using Ad-GFP, and an optimal multiplicity of infection (MOI) of 250 was determined as nontoxic and capable of transducing ⬃60% of the DCs (data not shown). All viruses were produced, quantified, and stored as previously described (17, 18).
Culture of BMDCs Single-cell suspensions of bone marrow cells were prepared from femurs, and the cells were cultured in the presence of recombinant murine GMCSF for 10 days as described previously (19).
T cell stimulation assay Two-fold dilutions of DCs transduced with adenoviral vectors 24 h previously were cultured in 96-well plates with whole splenocytes (106 cells/ well) or enriched CD8⫹ T cells (105 cells/well) from TCR318, SMARTA or C57BL/6 mice; 68 h later, all cultures were pulsed with 1 Ci of [3H]thymidine for 6 h.
CD8⫹ T cell enrichment CD8⫹ T cells were enriched through negative selection of CD4⫹, Ig⫹, and MHC class II⫹ cells using mAbs and magnetic beads (Dynabeads; Dynal).
Enumeration of Ag-specific T cells Single-cell suspensions of splenocytes were obtained by pressing the organs through a fine steel mesh, followed by centrifugation and resuspension in RPMI 1640 cell culture medium. Epitope-specific CD8⫹ T cell responses were enumerated by intracellular cytokine staining after a 5-h incubation with 1 g/ml relevant peptide as described (20).
CFSE labeling, transfer of TCR-transgenic splenocytes, and CFSE dilution analysis Spleen cells from TCR318-transgenic or SMARTA (both CD45.2⫹) mice were mixed with CFSE at a concentration of 1 M and incubated 10 min at 37°C. The reaction was stopped with a one-tenth volume of FCS; the cells were subsequently washed three times and finally resuspended in PBS and counted. A total of 9 ⫻ 107 CFSE-labeled cells were adoptively transferred into B6.SJL (CD45.1⫹) recipients. The following day, recipients were vaccinated with 107 infectious units of the indicated virus in both hind footpads and, 68 h later, the popliteal lymph nodes and spleens were harvested and subjected to flow cytometric analysis. Cells were stained with Abs against CD45.2, CD4, or CD8, and TCR V␣2 (CD4⫹ TCR transgenic) or TCR V8.1/2 (CD8⫹ TCR transgenic).
Survival study Mortality was recorded for 14 days after i.c. inoculation.
IMPROVED IMMUNITY BY ADENOVIRUS-BASED VACCINES Organ virus titers Organs were homogenized as a 10% organ suspension, and viral titers were determined using an immune focus assay as previously described (21).
Assessment of liver steatosis Frozen sections were fixed in paraformaldehyde and stained with oil Red-O solution and counterstained with hematoxylin. The relative area of oil Red-O staining was measured in four randomly selected visual fields from each slide using Image Pro, and the mean percentage ⫾ SEM was then calculated for each slide.
Tumor cell inoculation Mice were injected s.c. in the flank with 106 B16.F10 melanoma cells expressing the LCMV gp33– 41 epitope; these cells were donated by H. Pircher (22). Five days later, when tumors were palpable, the mice were vaccinated and tumor growth was measured every 2–3 days. Mice bearing tumors ⬎12 mm or showing ulceration were sacrificed in accordance with local ethical rules for animal experiments.
Statistical evaluation Quantitative results were compared using the Mann-Whitney U test. Survival curves after tumor challenge were compared using Mantel-Cox statistical analysis.
Results
Ii-linked Ag induces enhanced proliferation of naive CD4⫹ and CD8⫹ T cells To study the effect of increased Ag presentation to CD4⫹ T cells, we made adenoviral vectors expressing the LCMV glycoprotein (Ad-glycoprotein) and vectors expressing the LCMV glycoprotein fused to the C terminus of the murine Ii (Ad-Ii glycoprotein) (9). Previously, this strategy has been shown to increase presentation to CD4⫹ T cells (12). However, the majority of studies that have used Ii-coupled Ag for T cell stimulation only evaluated ex vivo priming of naive cells, and either report enhancement of CD4⫹ T cell responses only, or—if CD8⫹ T cell activation was actually evaluated—reveal modest if any effect (11–15). To address whether Ii-linked LCMV glycoprotein functioned as previously reported for Ii-linked Ags, we studied the early activation of naive CD4⫹ and CD8⫹ TCR-transgenic cells in vivo. To this end, we transferred CFSE-labeled splenocytes from C57BL/6 SMARTA (CD4⫹ T cells) or C57BL/6 TCR318 (CD8⫹ T cells) mice into C57BL/6.SJL mice. The C57BL/6 SMARTA mice express a V␣2/V8.3-transgenic TCR directed against a MHC class II-restricted epitope of LCMV glycoprotein (gp61– 80) on ⬃90% of their CD4⫹ T cells whereas the C57BL/6 TCR318 mice express a V␣2/V8.1-transgenic TCR directed against a MHC class I-restricted epitope of the LCMV glycoprotein (gp33– 41) on ⬃60% of their CD8⫹ T cells. Twenty-four hours after receipt of donor cells, the C57BL/6.SJL mice were vaccinated with Ad-Ii glycoprotein or Ad-glycoprotein in both hind paws, and popliteal lymph nodes were harvested 68 h after vaccination (Fig. 1A). As expected, a robust induction of CD4⫹ T cell proliferation could be observed in the draining lymph node of Ad-Ii glycoprotein-vaccinated mice. In contrast, vaccination with the Ad-glycoprotein vector induced only limited proliferation of CD4⫹ T cells in the draining lymph nodes. When looking at the proliferation of TCR-transgenic CD8⫹ T cells in the draining lymph nodes, it was found that Ad-Ii glycoprotein induced a higher fraction of the recruited cells to proliferate, but also that a substantial fraction of CD8⫹ T cells in the lymph nodes of animals primed with Ad-glycoprotein had begun to divide (Fig. 1A). These observations suggest a moderate enhancement of direct CD8⫹ T cell priming and a more pronounced enhancement of CD4⫹ T cell priming by Ii linkage.
The Journal of Immunology
FIGURE 1. Capacity of adenoviral vectors to stimulate CD4⫹ and CD8⫹ T cell responses in vivo and in vitro. A, Splenocytes from CD4⫹ or CD8⫹ TCR-transgenic mice (CD45.2) were labeled with CFSE and transferred into C57BL/6.SJL (CD45.1) recipients on day ⫺1. On day 0, mice were vaccinated with the indicated adenoviral vaccines, and popliteal lymph nodes were removed for analysis 68 h later. The histograms show the dilution of CFSE on TCR-transgenic cells from popliteal lymph nodes at 68 h postvaccination; cells were gated for CD45.2, CD4, and TCR V␣2 (CD4⫹ TCR) or CD45.2, CD8 and TCR V8.1/2 (CD8⫹ TCR). R1 depicts the percentage of cells that have undergone at least one cell division; R2 depicts percentage of cells that have not divided (average ⫾ SD, n ⫽ 4 mice in each group). B, Splenocytes from CD4⫹ TCR-transgenic, CD8⫹ TCR-transgenic, or WT (data not shown) mice were cocultured with 2-fold dilutions of adenovirus-transduced BMDCs starting at splenocyte-DC ratio of 10:1. Shown is the induced [3H]thymidine incorporation (cpm, average of duplicate cultures) during a 6-h pulse after 68 h of coculture.
The enhanced induction of CD8⫹ T cell proliferation contrasts with previous reports in the literature that showed similar proliferation of OT-I cells in vitro and in vivo (11, 12). To address this issue, we expanded BMDCs in GM-CSF ex vivo and infected these cells with Ad-Ii glycoprotein, Ad-glycoprotein, or control vectors (Ad-IiOVA and Ad-GFP) at a predetermined MOI of 250. Twenty-four hours later, 2-fold dilutions of these cells were mixed with splenocytes from TCR318, SMARTA, or WT mice, and the induced proliferation was measured 68 h later by a 6-h [3H]thymidine pulse. As can be seen in Fig. 1B, Ad-Ii glycoprotein transduced DCs were on a cell-for-cell basis markedly more efficient in stimulating the proliferation of both SMARTA and TCR318 cells— but not WT cells (data not shown)—than were matched Ad-glycoprotein-transduced cells; little stimulation was noted with Ad-IiOVA- or Ad-GFP-transduced DCs used for control. The augmented stimulation of CD8⫹ T cells by DCs infected with Ad-Ii glycoprotein could reflect increased costimulation as well as increased Ag expression. To address the first possibility,
3341
FIGURE 2. Phenotypic analysis of adenovirus-transduced BMDCs. A, BMDCs were transduced with increasing amounts of AD-II glycoprotein or Ad-glycoprotein and surface stained 24 h later with antiMHC class II and anti-CD40. Mean fluorescence intensity (MFI) for expression of CD40 on mature (MHC class II high (MHC high)), cf part B) and immature (MHC class II intermediate (MHC int)) DCs are presented as a function of either particle-to-cell ratio or infectious unitto-cell ratio; an arbitrary scale is applied. B, BMDCs were transduced with Ad-Ii glycoprotein or Ad-glycoprotein at a MOI of 250 and, 24 h later, the cells were surface stained with a combination of Abs against MHC class II and CD40, CD80, CD86, or MHC class I; untransduced and LPS stimulated BMDCs were included for comparison. Plots are representative of three independent experiments.
we compared the capacity of Ad-Ii glycoprotein and Ad-glycoprotein to induce expression of CD40 on BMDCs over a wide range of particle-to-cell ratios. As can be seen in Fig. 2A, incubation with both constructs led to increased DC activation with increased particle-to-cell ratio, highlighting the capacity of adenovirus particles to induce DC activation (23). However, no marked difference in the dose-response curves was observed using these constructs, and this was the case also if the constructs were compared based on the content of infectious virus. To further look for differences in the expression of cell surface molecules known to play a role in T cell activation, we subjected a sample of DCs (MOI of 250), similar to those used for T cell stimulation to an extended phenotypic analysis using Abs to a range of relevant surface molecules (MHC class I and II, CD80, CD86, and CD40). It is evident from this analysis (Fig. 2B) that no differences in the expression of these molecules could be detected, when cells infected with the two adenovirus constructs were compared. Although the above results strongly suggested that the augmented CD8⫹ T cell response following stimulation with Ad-Ii glycoprotein-infected DCs resulted from increased expression of Ag in association with MHC class I, we could not formally exclude the possibility that the increased proliferation somehow reflected increased CD4⫹ T cell help. To explore this possibility experimentally, we repeated the in vitro proliferation experiment using CD8⫹ T cells as responders (enriched by negative selective of CD4⫹, Ig⫹, and MHC class II⫹ cells) and included DCs derived from MHC class II-deficient mice for comparison (Fig. 3). Similar to whole spleen cell responders, CD8⫹ T cells clearly responded more vigorously to stimulation with Ad-Ii
3342
IMPROVED IMMUNITY BY ADENOVIRUS-BASED VACCINES
FIGURE 3. Ad-Ii glycoprotein-induced augmentation of CD8⫹ T cell proliferation does not require CD4⫹ T cell help. CD8⫹ TCR-transgenic cells (enriched through negative selection of CD4⫹, Ig⫹, and MHC class II⫹ cells) were cocultured with 2-fold dilutions of adenovirus-transduced BMDCs from WT and MHC class II-deficient mice. Shown is the induced [3H]thymidine incorporation (cpm, average of duplicate cultures) during a 6 h pulse after 68 h of coculture.
glycoprotein-infected DCs, and this pattern was observed irrespective of whether MHC class II was expressed on the DCs or not. Linkage of vaccine-encoded Ag to the HLA-associated Ii induces enhanced, sustained, and polyfunctional T cell responses Prompted by the ability of Ii-linked Ag to augment the proliferation of both CD4⫹ and CD8⫹ TCR-transgenic T cells, we next studied the effects of immunization with this construct on the polyclonal T cell response of nontransgenic mice. For that purpose, cohorts of C57BL/6 mice were vaccinated and the ensuing functional CD4⫹ and CD8⫹ T cell immune responses were evaluated through staining for intracellular IFN-␥ following brief ex vivo stimulation with relevant peptides. We found that Ad-Ii glycoprotein induced an augmented and accelerated CD4⫹ T cell response to gp61– 80, whereas the response following vaccination with the unmodified Ad-glycoprotein vector was barely discernible (Fig. 4A). With regard to CD8⫹ T cells, we evaluated the response to four MHC class I-restricted epitopes, and found accelerated, enhanced and protracted responses (Fig. 4B). This was particularly apparent for the normally subdominant epitopes gp276 –284, gp92–100, and gp118 –126, against which Ad-glycoprotein induced only minimal responses. Indeed, gp276 –284 became a codominant epitope as a result of vaccination with the linked construct. Some studies have indicated that long-term protection against chronic viral infection requires the existence of CD8⫹ T cells capable of coproducing several effector cytokines, in particular IFN-␥, IL-2, and TNF-␣ (24 –26). To address whether such cells were maintained in Ad-Ii glycoprotein-vaccinated mice, we evaluated CD8⫹ T cell responses in mice vaccinated 1 year previously and found that a substantial fraction of the cells responding to peptide stimulation with IFN-␥ production ex vivo also produced IL-2 and TNF-␣ (Fig. 4C). Similar analysis could not be per-
FIGURE 4. Immune responses induced by adenoviral vaccine vectors. A, Total number of splenic CD4⫹ T cells responding with IFN-␥ production to the gp61– 80 peptide at the indicated times after vaccination with either Ad-glycoprotein or Ad-IiGP. B, Total number of splenic CD8⫹ T cells responding with IFN-␥ production to the immunodominant gp33– 41 and gp276 –286 peptides (top) or subdominant gp92–101 and gp118 –125 peptides (bottom) at the indicated times after vaccination. C, Splenocytes from animals vaccinated 1 year previously with Ad-Ii glycoprotein were stimulated ex vivo with a pool of peptides (gp33– 41, gp92–101, gp118 – 125, and gp276 –284) and stained for surface CD8, intracellular IFN-␥, and TNF-␣ or IL-2. Shown are the percentages of the IFN-␥-producing CD8⫹ T cells that also produce either TNF-␣ or IL-2. For Ad-glycoprotein-vaccinated mice similar analyses could not be performed due to too low numbers of Ag-specific cells.
formed on cells from Ad-glycoprotein-vaccinated mice as the frequencies of virus-specific cells were too low (cf. Fig. 4B). Ad-Ii glycoprotein induces fast and prolonged protection against LCMV infection Next, we compared the protective capacity of the responses induced by the Ad-glycoprotein and Ad-Ii glycoprotein vectors. To this end, vaccinated mice were challenged i.v. with 105 PFU of LCMV clone 13 either 3 wk or 9 mo after vaccination, and spleens,
The Journal of Immunology
3343 Table I. Protection against lethal LCMV challenge % Surviving Animals after i.c. Challengea Days p.v.
1 3 5 7 14 21 60 90 180 270 360
Ad-GP
Ad-IiGP b
0% (5) 0% (10) 0% (5) 6% (18) 53% (15) 20% (5) 0% (5) 0% (5) 0% (5) 0% (5) 0% (10)
20% (5) 100% (10)c 100% (5) 100% (5) ND 100% (5) 100% (5) 100% (5) 100% (5) 100% (9) 90% (10)d
a Mice were vaccinated with Ad-GP or Ad-IiGP, and on the indicated days after vaccination, some of the mice were challenged with a lethal dose (20 PFU) of LCMV Armstrong clone 53. b Number of animals in each group. c Five of five mice vaccinated with an adenovector encoding an irrelevant Ag linked to Ii died. d One animal died without symptoms of LCMV.
FIGURE 5. Protective efficacy of adenoviral vaccines. A, Spleen and lung virus titers 5 days after infection with 105 PFU of LCMV clone 13 in mice vaccinated with Ad-glycoprotein, Ad-Ii glycoprotein, or a control vector (Ad-nil) 21 days previously. B, Spleen and lung virus titers 5 days after infection with 105 PFU of LCMV clone 13 in mice vaccinated with Ad-glycoprotein, Ad-Ii glycoprotein, or Ad-nil 9 mo previously. Points represent individual mice; dotted line denotes limit of detection. C, Hepatic steatosis in animals, 5 days after infection with 105 PFU of LCMV clone 13 in mice vaccinated with Ad-glycoprotein, Ad-Ii glycoprotein, or Ad-nil 9 mo previously; one representative tissue section from each group is shown with the average percentages ⫾ SEMs of liver area stained for fat droplets inserted. D, Mice were vaccinated with Ad-glycoprotein, Ad-Ii glycoprotein, or Ad-nil and challenged i.c. 60 days later with 20 PFU of LCMV Armstrong clone 53b. Five animals from each group of mice were sacrificed on days 3, 5, and 7 after infection, and brain virus titers were determined by plaque assay; a line connects median viral titers. The indicated percentages show the mortalities of control groups in which the infection was allowed to follow its own course. Points represent individual mice; dotted line denotes limit of detection.
lungs, and livers were harvested 5 days later. In this setting, both vaccine vectors induced accelerated immune responses that could control the infection early after vaccination (Fig. 5A). However, by 9 mo postvaccination, the Ad-glycoprotein-vaccinated mice showed a reduction in virus titers when compared with controls vaccinated with an empty vector, but infectious virus was readily detectable. In contrast, hardly any infectious virus was recovered from Ad-Ii glycoprotein-vaccinated animals even when challenged this long after vaccination (Fig. 5B). Furthermore, it was found that Ad-glycoprotein-vaccinated mice, but not Ad-Ii glycoprotein-vaccinated mice (no virus left) or control vaccinated mice (T cells not yet generated) exhibited enhanced immunopathology as evidenced by increased hepatic steatosis (Fig. 5C) (27). To determine the ability of the vaccine to protect mice against infection of a solid, nonlymphoid organ, vaccinated mice were challenged i.c. with a lethal dose of 20 PFU of LCMV Armstrong clone 53b at various time points between 1 day and 1 year after vaccination. In this setting, the outcome of infection is decided by a race between virus replication in the CNS and the local accumulation of antiviral CD8⫹ T cells, and naive mice invariably die (28). When tested under these conditions, Ad-glycoprotein-vaccinated mice were never more than partially protected. In contrast, the Ad-Ii glycoprotein vaccine induced virtually complete protection as early as 3 days and as late as 1 year after vaccination (Table I). As expected, protection correlated with the ability of Ad-Ii glycoprotein-vaccinated mice to rapidly abort the infection in the brain (Fig. 5D). Notably, Ii-linked irrelevant Ag did not induce any protection demonstrating that the rapid protection was Ag specific. Ad-Ii glycoprotein is as efficient as LCMV in protecting against melanomas expressing LCMV-derived Ag We also wanted to evaluate the ability of our vaccine construct to provide protection against a nonviral challenge. To this end, mice were injected s.c. with 106 LCMV glycoprotein-expressing melanoma cells and 5 days later, when tumors were palpable, the mice were vaccinated with adenovirus-encoding Ii-linked irrelevant Ag (Ad-IiOVA), Ad-glycoprotein, Ad-Ii glycoprotein, or live LCMV, and tumor growth was recorded. Vaccination with Ad-glycoprotein was found to be able to delay tumor growth significantly as compared with Ad-IiOVA, this delay could be further significantly increased by Ad-Ii glycoprotein and, remarkably, no significant
3344
IMPROVED IMMUNITY BY ADENOVIRUS-BASED VACCINES
FIGURE 6. Vaccination protects against growing melanomas. Shown are survival of animals injected with 106 gp33– 41 expressing B16.F10 melanoma cells and vaccinated 5 days later—when tumors were palpable—with Ad-Ii glycoprotein, Ad-glycoprotein, or Ad-IiOVA or 103 PFU LCMV Armstrong. Mean survival times are shown next to the graphs; n ⫽ 7– 8 mice/group. Bars denoting statistical significance between Ad-IiOVA and Ad-glycoprotein, as well as between Ad-Ii glycoprotein and Ad-glycoprotein are inserted (Mantel-Cox test).
difference could be seen when Ad-Ii glycoprotein was compared with live LCMV (Fig. 6). Broad CD8⫹ T cell response induced by the Ad-Ii glycoprotein vaccine protects against viral escape variants Apart from magnitude and quality of a vaccine-induced immune response, a critical issue is the breadth of the induced response. To functionally evaluate the breadth of the induced protection, we infected naive, Ad-glycoprotein-vaccinated, or Ad-Ii glycoprotein-vaccinated mice i.c. with LCMV variants carrying mutations in the gp33– 41 (gp33 nil) epitope, the gp276 –284 (gp276 nil) epitope, or both (gp33/gp276 nil) (17) and monitored survival. The Ad-Ii glycoprotein-vaccinated mice were completely protected against challenge with either the gp33 nil or the gp276 nil variant, and 70% of Ad-Ii glycoprotein-vaccinated mice were protected against challenge with the gp33/gp276 nil variant (Fig. 7A). Analysis of the secondary immune response in the surviving Ad-Ii glycoprotein-vaccinated mice disclosed that survival was associated with robust expansion of CD8⫹ T cells directed against epitopes shared between the vaccine and the challenge virus (Fig. 7A). That Ii linkage of Ag increases the protective efficacy of the broadened response is perhaps the most important finding in a real world vaccination scenario and, accordingly, we wanted to determine whether this could be generalized by using another Ag. Given the accessibility of a mutated variant for challenge, we focused on the LCMV NP and constructed adenoviral vectors expressing NP with (Ad-IiNP) and without (Ad-NP) linkage to Ii. Groups of animals were then vaccinated with one of these constructs and challenged with either LCMV Armstrong 53b or the NP396 nil variant (17) 90 days later. Five of five animals survived challenge with LCMV 53b if they had been vaccinated with either Ad-NP or Ad-IiNP. This is perhaps not surprising as NP has previously been reported to be markedly superior to glycoprotein as a viral Ag with respect to affinity of the dominant epitope, affinity of the responding T cells for bound peptide, amount of bound peptide needed for lysis of target cells, and the protective capacity of the induced T cells and, finally, as has become relevant during challenge, the kinetics of protein synthesis in LCMV-infected cells (29 –32). However, the response toward NP is also highly focused on the dominant NP396 – 404 epitope and consequently, 5 of 10 of the Ad-NP-vaccinated animals succumbed after challenge with the LCMV NP396-nil variant. In contrast, 10 of 10 of the Ad-IiNP-
FIGURE 7. Survival and anamnestic CD8⫹ T cell responses after challenge with virus escape variants. A, Survival of animals that were vaccinated with Ad-glycoprotein, Ad-Ii glycoprotein, or sham-vaccinated and challenged i.c. 90 days later with 20 PFU LCMV Armstrong variants that carry mutations in gp33– 41 (gp33-nil, upper panel), gp276 –286 (gp276nil, middle panel) or both epitopes (gp33/gp276-nil, lower panel). B, Animals were vaccinated with Ad-NP, Ad-IiNP, or sham-vaccinated and challenged i.c. 90 days later with 20 PFU LCMV Armstrong variants that carry mutations in the NP396 – 404 epitope (NP396-nil). For all panels, ⴱ, statistical significance relative to animals vaccinated with relevant Ag without Ii. The total numbers of splenic T cells responding to the epitopes of the vaccine Ag, or the immunodominant epitope not included in the vaccine (NP396 for Ad-Ii glycoprotein-vaccinated, gp33 for Ad-NP, and Ad-IiNPvaccinated) were determined in the surviving mice 14 days after challenge and are shown to the right of the respective survival curves.
vaccinated animals survived lethal challenge with this variant ( p ⬍ 0.05, Fig. 7B). As expected, analysis of the surviving animals revealed a much more prominent recall response against
The Journal of Immunology the subdominant epitopes NP205–212, NP163–175, and NP238 – 243 in Ad-IiNP-vaccinated mice as compared with Ad-NP-vaccinated mice (Fig. 7B).
Discussion Based on the results presented in this report, we conclude that adenoviral vectors can benefit profoundly from Ag engineering, thus improving breadth, kinetics, potency, quality, and stability of ensuing CD8⫹ T cell responses. We found these improvements to be functionally relevant both in a peripheral and a systemic challenge model, in which effector memory and central memory T cells, respectively, would be expected to play the most important roles. We also found an effect in a therapeutic tumor vaccination model that rivals live LCMV. In this therapeutic tumor vaccination model system, protection has previously been demonstrated to depend on CD8⫹ T cell cytokine production (22). Such suggested cytokine-producing competence may also be important in vaccination against chronic infections such as HIV where long-term nonprogression has been associated with the frequency of polyfunctional CD8⫹ T cells capable of coproducing cytokines such as IFN-␥, TNF-␣, and in particular IL-2 (24). In mice vaccinated 1 year previously, we found that a substantial fraction of IFN-␥producing Ag-specific CD8⫹ T cells also produced TNF-␣ and IL-2 whereas similar responses could not be detected in Ad-glycoprotein-vaccinated mice. Even though it remains to be determined to what extent other Ags will benefit from Ag engineering, we have observed accelerated, broadened, and augmented CD8⫹ T cell responses for a clear majority of the Ags that we have evaluated so far (data not shown). Thus, it seems safe to conclude that adenovirus vaccines expressing Ii-linked Ags may be used successfully in situations where unmodified vaccines have failed. In particular, the ability to induce protective immunity from subdominant epitopes, which we have verified for both the LCMV glycoprotein and NP, could be relevant for mutating and highly variable viruses like HCV and HIV (33, 34). Even if a protective HIV or HCV vaccine remains elusive, at the very least, the strategy described should be applicable to make improved protective vaccines for the related arenaviruses Lassa, Junin, Sabia, Machupo, and Guanorito which infect humans with high case-fatality rates, or the filoviruses Ebola and Marburg toward which unmodified adenovirus vaccines have already proved effective (35–38). Apart from the ability to provide protection in endemic areas, the marked acceleration of protective efficacy provided by Ii fusion vaccines would be particularly useful if some of these viruses were introduced as bioterrorist weapons in a previously unvaccinated population, as is indeed speculated by the U.S. Centers for Disease Control and Prevention (www.bt.cdc.gov/agent/agentlist-category. asp). In such a setting, there would be a need for immediately protective vaccines to protect local health care workers and exposed individuals. Although we have not yet established the special property of the Ii sequence that confers superior vaccine protection to the glycoprotein Ag, some lessons can be drawn from this study. First, most of the previous observations with Ii-linked Ags have focused on the ability of Ii to increase stimulation of transgenic CD4⫹ T cells in vitro. We have now verified this mechanism in vivo. We also observed that Ii-linked glycoprotein stimulates CD8⫹ T cells more efficiently both in vitro and in vivo. Although this is consistent with the accelerated and increased vaccine efficacy, it is not as easy to explain. Indeed, the in vitro effect contrast to what has previously been shown studying the SIINFEKL epitope of OVA (12). The reason for this discrepancy is not entirely clear, but may derive from differences in TCR or epitope affinity, or perhaps reflect saturation in the previous experiments, as a titration of
3345 APCs was not presented. What we can say from preliminary experiments is that the sorting signal from LAMP-1 cannot substitute for Ii (data not shown). We also find that the effect can be recapitulated in vitro, which suggests a direct interaction between the transduced APCs and Ag-specific T cells. Thus, on a per-cell basis Ad-Ii glycoprotein-transduced DCs were more efficient stimulators of naive CD8⫹ T cells. Combined with our results demonstrating that key costimulatory molecules are equally up-regulated on Ad-Ii glycoprotein and Ad-glycoprotein-transduced cells, and that CD4⫹ T cell help is not required for a better stimulation in vitro, the mechanism appears to be related to a more efficient pathway for presentation of MHC class I-restricted antigenic peptides. Getting closer to a mechanistic understanding is challenging, as Ii has multiple domains with a possible involvement in T cell stimulation. These include domains involved in endosomal sorting, MHC class I and II association, and intramembrane as well as exoplasmic trimerization, binding, and signaling through CD44 and MIF in surrounding cells and through NF-B p65/RelA homodimers in Ii-expressing cells (39 – 43). Given this multitude of options, it will probably require extensive mutagenesis mapping of domains within the Ii sequence to identify regions important for increased CD8⫹ T cell stimulation to guide experimental studies pinpointing the function. Such analysis is beyond the scope of this study, and we are currently in the initial phases of such studies. In conclusion, our results point to Ii linkage of the Ag expressed from adenoviral vectors as a simple way to improve critical aspects of T cell-mediated immunity against both viruses and tumors.
Acknowledgments We thank G. Tho¨rner, L. Malthe, P. Rasmussen, B. Jensen, and D. Bardenfleth for expert technical assistance, C. Deacon for comments to the manuscript, and Norbert Koch, H. Pircher, M. B. Oldstone, M. Whitt, and D. Pinschewer for reagents.
Disclosures The University of Copenhagen has filed a patent application regarding the findings described herein. The authors Peter Johannes Holst, Allan Randrup Thomsen and Jan Pravsgaard Christensen are entitled to a fraction of any net income derived from this patent application.
References 1. McMichael, A. J. 2006. HIV vaccines. Annu. Rev. Immunol. 24: 227–255. 2. Maeda, K., K. West, D. Hayasaka, F. A. Ennis, and M. Terajima. 2005. Recombinant adenovirus vector vaccine induces stronger cytotoxic T-cell responses than recombinant vaccinia virus vector, plasmid DNA, or a combination of these. Viral Immunol. 18: 657– 667. 3. Sumida, S. M., D. M. Truitt, M. G. Kishko, J. C. Arthur, S. S. Jackson, D. A. Gorgone, M. A. Lifton, W. Koudstaal, M. G. Pau, S. Kostense, et al. 2004. Neutralizing antibodies and CD8⫹ T lymphocytes both contribute to immunity to adenovirus serotype 5 vaccine vectors. J. Virol. 78: 2666 –2673. 4. Bangari, D. S., and S. K. Mittal. 2006. Current strategies and future directions for eluding adenoviral vector immunity. Curr. Gene Ther. 6: 215–226. 5. Roberts, D. M., A. Nanda, M. J. Havenga, P. Abbink, D. M. Lynch, B. A. Ewald, J. Liu, A. R. Thorner, P. E. Swanson, D. A. Gorgone, et al. 2006. Hexon-chimaeric adenovirus serotype 5 vectors circumvent pre-existing anti-vector immunity. Nature 441: 239 –243. 6. Masopust, D., S. J. Ha, V. Vezys, and R. Ahmed. 2006. Stimulation history dictates memory CD8 T cell phenotype: implications for prime-boost vaccination. J. Immunol. 177: 831– 839. 7. Vaccari, M., C. J. Trindade, D. Venzon, M. Zanetti, and G. Franchini. 2005. Vaccine-induced CD8⫹ central memory T cells in protection from simian AIDS. J. Immunol. 175: 3502–3507. 8. Palmowski, M. J., E. M. Choi, I. F. Hermans, S. C. Gilbert, J. L. Chen, U. Gileadi, M. Salio, A. Van Pel, S. Man, E. Bonin, et al. 2002. Competition between CTL narrows the immune response induced by prime-boost vaccination protocols. J. Immunol. 168: 4391– 4398. 9. Holst, P. J., C. Bartholdy, A. Stryhn, A. R. Thomsen, and J. P. Christensen. 2007. Rapid and sustained CD4⫹ T-cell independent immunity from adenovirus encoded vaccine antigens. J. Gen. Virol. 88: 1708 –1716. 10. Thomsen, A. R., J. Johansen, O. Marker, and J. P. Christensen. 1996. Exhaustion of CTL memory and recrudescence of viremia in lymphocytic choriomeningitis virus-infected MHC class II-deficient mice and B cell-deficient mice. J. Immunol. 157: 3074 –3080.
3346 11. Rowe, H. M., L. Lopes, Y. Ikeda, R. Bailey, I. Barde, M. Zenke, B. M. Chain, and M. K. Collins. 2006. Immunization with a lentiviral vector stimulates both CD4 and CD8 T cell responses to an ovalbumin transgene. Mol. Ther. 13: 310 –319. 12. Diebold, S. S., M. Cotten, N. Koch, and M. Zenke. 2001. MHC class II presentation of endogenously expressed antigens by transfected dendritic cells. Gene Ther. 8: 487– 493. 13. Sanderson, S., K. Frauwirth, and N. Shastri. 1995. Expression of endogenous peptide-major histocompatibility complex class II complexes derived from invariant chain-antigen fusion proteins. Proc. Natl. Acad. Sci. USA 92: 7217–7221. 14. Malcherek, G., C. Wirblich, N. Willcox, H. G. Rammensee, J. Trowsdale, and A. Melms. 1998. MHC class II-associated invariant chain peptide replacement by T cell epitopes: engineered invariant chain as a vehicle for directed and enhanced MHC class II antigen processing and presentation. Eur. J. Immunol. 28: 1524 –1533. 15. Sponaas, A., C. Carstens, and N. Koch. 1999. C-terminal extension of the MHC class II-associated invariant chain by an antigenic sequence triggers activation of naive T cells. Gene Ther. 6: 1826 –1834. 16. Becker, T. C., R. J. Noel, W. S. Coats, A. M. Gomez-Foix, T. Alam, R. D. Gerard, and C. B. Newgard. 1994. Use of recombinant adenovirus for metabolic engineering of mammalian cells. Methods Cell Biol. 43(Pt. A): 161–189. 17. Lewicki, H. A., M. G. von Herrath, C. F. Evans, J. L. Whitton, and M. B. Oldstone. 1995. CTL escape viral variants. II. Biologic activity in vivo. Virology 211: 443– 450. 18. Thomsen, A. R., A. Nansen, C. Andersen, J. Johansen, O. Marker, and J. P. Christensen. 1997. Cooperation of B cells and T cells is required for survival of mice infected with vesicular stomatitis virus. Int. Immunol. 9: 1757–1766. 19. Lutz, M. B., N. Kukutsch, A. L. Ogilvie, S. Rossner, F. Koch, N. Romani, and G. Schuler. 1999. An advanced culture method for generating large quantities of highly pure dendritic cells from mouse bone marrow. J. Immunol. Methods 223: 77–92. 20. Christensen, J. P., S. O. Kauffmann, and A. R. Thomsen. 2003. Deficient CD4⫹ T cell priming and regression of CD8⫹ T cell functionality in virus-infected mice lacking a normal B cell compartment. J. Immunol. 171: 4733– 4741. 21. Battegay, M., S. Cooper, A. Althage, J. Banziger, H. Hengartner, and R. M. Zinkernagel. 1991. Quantification of lymphocytic choriomeningitis virus with an immunological focus assay in 24- or 96-well plates. J. Virol. Methods 33: 191–198. 22. Prevost-Blondel, A., M. Neuenhahn, M. Rawiel, and H. Pircher. 2000. Differential requirement of perforin and IFN-␥ in CD8 T cell-mediated immune responses against B16.F10 melanoma cells expressing a viral antigen. Eur. J. Immunol. 30: 2507–2515. 23. Molinier-Frenkel, V., A. Prevost-Blondel, S. S. Hong, R. Lengagne, S. Boudaly, M. K. Magnusson, P. Boulanger, and J. G. Guillet. 2003. The maturation of murine dendritic cells induced by human adenovirus is mediated by the fiber knob domain. J. Biol. Chem. 278: 37175–37182. 24. Betts, M. R., M. C. Nason, S. M. West, S. C. De Rosa, S. A. Migueles, J. Abraham, M. M. Lederman, J. M. Benito, P. A. Goepfert, M. Connors, et al. 2006. HIV nonprogressors preferentially maintain highly functional HIV-specific CD8⫹ T cells. Blood 107: 4781– 4789. 25. Kristensen, N. N., J. P. Christensen, and A. R. Thomsen. 2002. High numbers of IL-2-producing CD8⫹ T cells during viral infection: correlation with stable memory development. J. Gen. Virol. 83: 2123–2133. 26. Wherry, E. J., and R. Ahmed. 2004. Memory CD8 T-cell differentiation during viral infection. J. Virol. 78: 5535–5545.
IMPROVED IMMUNITY BY ADENOVIRUS-BASED VACCINES 27. Lohler, J., J. Gossmann, T. Kratzberg, and F. Lehmann-Grube. 1994. Murine hepatitis caused by lymphocytic choriomeningitis virus. I. The hepatic lesions. Lab. Invest. 70: 263–278. 28. Thomsen, A. R., M. Volkert, and O. Marker. 1979. The timing of the immune response in relation to virus growth determines the outcome of the LCM infection. Acta Pathol. Microbiol. Scand. C. 87C: 47–54. 29. Gairin, J. E., H. Mazarguil, D. Hudrisier, and M. B. Oldstone. 1995. Optimal lymphocytic choriomeningitis virus sequences restricted by H-2Db major histocompatibility complex class I molecules and presented to cytotoxic T lymphocytes. J. Virol. 69: 2297–2305. 30. Hudrisier, D., H. Mazarguil, F. Laval, M. B. Oldstone, and J. E. Gairin. 1996. Binding of viral antigens to major histocompatibility complex class I H-2Db molecules is controlled by dominant negative elements at peptide non-anchor residues: implications for peptide selection and presentation. J. Biol. Chem. 271: 17829 –17836. 31. Gallimore, A., T. Dumrese, H. Hengartner, R. M. Zinkernagel, and H. G. Rammensee. 1998. Protective immunity does not correlate with the hierarchy of virusspecific cytotoxic T cell responses to naturally processed peptides. J. Exp. Med. 187: 1647–1657. 32. Probst, H. C., K. Tschannen, A. Gallimore, M. Martinic, M. Basler, T. Dumrese, E. Jones, and M. F. van den Broek. 2003. Immunodominance of an antiviral cytotoxic T cell response is shaped by the kinetics of viral protein expression. J. Immunol. 171: 5415–5422. 33. Frahm, N., P. Kiepiela, S. Adams, C. H. Linde, H. S. Hewitt, K. Sango, M. E. Feeney, M. M. Addo, M. Lichterfeld, M. P. Lahaie, et al. 2006. Control of human immunodeficiency virus replication by cytotoxic T lymphocytes targeting subdominant epitopes. Nat. Immunol. 7: 173–178. 34. Turnbull, E. L., A. R. Lopes, N. A. Jones, D. Cornforth, P. Newton, D. Aldam, P. Pellegrino, J. Turner, I. Williams, C. M. Wilson, et al. 2006. HIV-1 epitopespecific CD8⫹ T cell responses strongly associated with delayed disease progression cross-recognize epitope variants efficiently. J. Immunol. 176: 6130 – 6146. 35. Charrel, R. N., and X. L. de Lamballerie. 2003. Arenaviruses other than Lassa virus. Antiviral Res. 57: 89 –100. 36. Gunther, S., and O. Lenz. 2004. Lassa virus. Crit. Rev. Clin. Lab. Sci. 41: 339 –390. 37. Sullivan, N. J., T. W. Geisbert, J. B. Geisbert, L. Xu, Z. Y. Yang, M. Roederer, R. A. Koup, P. B. Jahrling, and G. J. Nabel. 2003. Accelerated vaccination for Ebola virus haemorrhagic fever in non-human primates. Nature 424: 681– 684. 38. Wang, D., M. Hevey, L. Y. Juompan, C. M. Trubey, N. U. Raja, S. B. Deitz, J. Woraratanadharm, M. Luo, H. Yu, B. M. Swain, et al. 2006. Complex adenovirus-vectored vaccine protects guinea pigs from three strains of Marburg virus challenges. Virology 353: 324 –332. 39. Bakke, O., and B. Dobberstein. 1990. MHC class II-associated invariant chain contains a sorting signal for endosomal compartments. Cell 63: 707–716. 40. Powis, S. J. 2006. CLIP-region mediated interaction of invariant chain with MHC class I molecules. FEBS Lett. 580: 3112–3116. 41. Dixon, A. M., B. J. Stanley, E. E. Matthews, J. P. Dawson, and D. M. Engelman. 2006. Invariant chain transmembrane domain trimerization: a step in MHC class II assembly. Biochemistry 45: 5228 –5234. 42. Starlets, D., Y. Gore, I. Binsky, M. Haran, N. Harpaz, L. Shvidel, S. Becker-Herman, A. Berrebi, and I. Shachar. 2006. Cell-surface CD74 initiates a signaling cascade leading to cell proliferation and survival. Blood 107: 4807– 4816. 43. Shi, X., L. Leng, T. Wang, W. Wang, X. Du, J. Li, C. McDonald, Z. Chen, J. W. Murphy, E. Lolis, et al. 2006. CD44 Is the signaling component of the macrophage migration inhibitory factor-CD74 receptor complex. Immunity 25: 595– 606.
