Adenovirus-Based Vaccine against Listeria monocytogenes: Extending the Concept of Invariant Chain Linkage

Vaccination is the most advantageous medical intervention available for combating microbial infections. Although most vaccines in current use mediate protection primarily by means of specific Abs, this strategy will not suffice against some of the major microbial threats of our time (e.g., HIV, hepatitis C virus, tuberculosis, and malaria) (1, 2). The requirement for sustained protective cellular immunity has led to studies of the use of replication-deficient adenoviruses as vaccine vectors (3). These vectors induce strong CD8⁺ T cell responses but relatively weak CD4⁺ T cell responses. We previously demonstrated greatly enhanced responses from both CD4⁺ and CD8⁺ T cells by coupling the Ag of interest to MHC class II–associated invariant chain (Ii) (4). This vaccination principle has shown its worth against several viral infections, but it has not been investigated against infection with intracellular bacteria. For the purpose of extending the vaccination concept to include intracellular bacterial infections, the well-characterized model organism Listeria monocytogenes was used in this study.

L. monocytogenes is a Gram-positive, facultative intracellular bacterium, and it is listed as a Category B agent on the National Institute of Allergy and Infectious Diseases’ list of priority pathogens (http://www.niaid.nih.gov/topics/BiodefenseRelated/Biodefense/Pages/CatA.aspx). L. monocytogenes is an enteroinvasive, food-borne human pathogen capable of causing chorioamnionitis in pregnant women, possibly leading to severe infection of the fetus. In immunocompromised individuals, infection may lead to septicemia and meningitis (5, 6).

To effectively clear a L. monocytogenes infection, the host must elicit an adaptive immune response involving both CD8⁺ and CD4⁺ T cells (7, 8). Although both CD8⁺ and CD4⁺ T cells contribute to protective immunity against L. monocytogenes, studies demonstrated that CD8⁺ T cells play the main role in conveying long-term protective immunity (9–11). CD8⁺ T cells mediate protection in a number of ways, including cytolysis of infected cells via the granule exocytosis pathway, which is dependent on perforin and granzymes, as well as via upregulation of the expression of FasL (CD95L) for interaction with Fas (CD95), resulting in the aggregation of these molecules on target cells for initiation of programmed cell death (12–14). Additionally, secretion of cytokines, such as IFN-γ, TNF-α, and CCL3, by CD8⁺ T cells results in recruitment and activation of phagocytes (15, 16). Studies revealed IFN-γ secretion and perforin-dependent degranulation to be the primary effector mechanisms underlying protection in the fight against L. monocytogenes (14, 17–19).

To determine whether replication-deficient adenovirus vaccines will confer protection against L. monocytogenes, we first tested two adenovirus serotype 5 (Ad5) constructs expressing the glycoprotein (GP) of lymphocytic choriomeningitis virus (LCMV) with and without Ii linkage; Ad-Ii-GP and Ad-GP, respectively. These constructs have a previously documented in vivo protective capacity against viral infection, regardless of whether perforin or IFN-γ represents the major antiviral effector mechanism (20, 21). Through challenge of vaccinated mice with a recombinant strain of L. monocytogenes modified to secrete part of LCMV GP (Lm-GP33), we obtained basic proof-of-concept for the hypothesis that the adenovirus-based vaccine strategy will confer protection against an intracellular bacterial infection like L. monocytogenes and that protection might be improved by tethering of the target Ag to Ii. We then extended the concept to natural L. monocytogenes infection, vaccinating mice with adenoviral constructs encoding truncated versions of either of two well-known L. monocytogenes Ags: listeriolysin O (LLO) and p60. Vaccination with natural Ags of L. monocytogenes confirmed the impact of Ii linkage on CD8⁺ T cell responses to weaker Ags, such as p60 in BALB/c mice and LLO in C57BL/6 mice. The strong Ag, LLO in BALB/c mice, did not benefit significantly from linkage during the early memory phase (30 d postvaccination; dpv); however, the protection induced by LLO vaccination was prolonged in the spleen (60 dpv).

Female C57BL/6 and BALB/c mice were obtained from Taconic Farms and housed in a specific pathogen–free facility. IFN-γ–knockout (IFNγ^−/−) mice, perforin-knockout (Prf^−/−) mice, and IFN-γ–perforin double-knockout (IFNγ^−/−Prf^−/−) mice, all on a C57BL/6 background, were bred locally from breeder pairs originally obtained from The Jackson Laboratory (Bar Harbor, ME). All mice were acclimatized for ≥1 wk after arrival at the facility before being used in experiments; by this time they were between 6 and 12 wk old. All experiments were approved by the local animal ethics council and were performed in accordance with the national guidelines for animal experiments.

Replication-deficient E1-deleted Ad5 vectors with a nonfunctional E3 gene, expressing either GP or nucleoprotein (NP) of LCMV linked to Ii (designated Ad-Ii-GP and Ad-Ii-NP); LLO or p60 of L. monocytogenes linked to Ii (designated Ad-Ii-LLO and Ad-Ii-p60); or GP, LLO, or p60 alone (designated Ad-GP, Ad-LLO, and Ad-p60) were produced as described previously (22). Adenoviral particles were purified using standard methods, aliquoted, and frozen at −80°C in 10% glycerol. Inserts were verified by sequencing and restriction enzyme digestion. Infectivity of the adenovirus stocks was determined using Adeno-X Rapid Titer Kit (Clontech Laboratories, Mountain View, CA). Mice to be vaccinated with adenoviral constructs were anesthetized with isoflurane and immunized s.c. in the right hind footpad with 2 × 10⁷ PFU in 30 μl PBS (23).

For L. monocytogenes infection, wild-type L. monocytogenes serovar 1/2a EGD (Lm-wt) and Lm-GP33 were used (24). The production and validation of the recombinant strain were described previously (25), and the LCMV gp33–41 epitope is expressed within a secreted dihydrofolate reductase fusion protein (26). Bacteria were grown to logarithmic phase in brain–heart infusion (BHI) broth with shaking at 37°C overnight. The next morning, 1 ml this culture was transferred to 50 ml fresh BHI broth and incubated for 1 h. OD₆₀₀ was measured and converted to an estimate of the bacterial concentration via a predetermined standard curve. The culture was diluted in PBS to the relevant concentration for i.v. injection of 0.3 ml; in most experiments 1 × 10⁵ or 3 × 10⁴ CFU was injected into C57BL/6 (L. monocytogenes–resistant) or BALB/c (L. monocytogenes–susceptible) mice, respectively. For each experiment, the bacterial dose used for infection was controlled by plating, as described below. Mice were monitored three times daily following challenge and sacrificed at 3 d postinfection (dpi), unless otherwise stated.

For determination of bacterial loads in internal organs, livers and spleens were removed aseptically and frozen at −80°C. On the day of assaying, organs were ground using a tissue grinder (Tenbroeck 432-5000; Wheaton), and serial dilutions were plated using BHI agar. Following 24 h of incubation, numbers of bacterial colonies were counted to determine CFU; results are presented as CFU/g organ.

To obtain splenocytes in single-cell suspensions for intracellular cytokine staining (ICS), spleens were aseptically removed and pressed through a fine steel mesh (70 μm), followed by centrifugation and resuspension in RPMI 1640 cell culture medium containing 10% FCS supplemented with 2-ME, l-glutamine, and penicillin-streptomycin. For enumeration of Ag-specific T cells, splenocytes were incubated at 37°C and 5% CO₂ for 5 h in the presence of 1 μg/ml relevant peptide, 50 IU/ml IL-2, and 3 μM monensin. Cells were stained for cell surface markers using FITC-CD44, PerCP–Cy5.5–CD8, and PerCP–Cy5.5–CD4. Subsequent to surface staining, the cells were washed, permeabilized, and stained for intracellular IFN-γ using PE–IFN-γ. For evaluation of degranulation, Alexa Fluor 488–CD107a was present throughout the 5-h incubation period. Samples were analyzed using a FACSCalibur (BD Biosciences), and data analysis was conducted using FlowJo software (TreeStar).

To deplete mice of CD8⁺ T cells, depleting anti-CD8 Abs (YTS156 and YTS169; each at 100 μg/d) were injected i.p. into recipient mice on days −1 and 0, with day 0 representing the day of bacterial challenge. To validate the depletion of CD8⁺ T cells, splenocytes from recipient mice were surface stained with anti-CD8, anti-CD4, and anti-CD3 mAbs on day 3 postchallenge and analyzed by flow cytometry.

Statistical analysis was performed using GraphPad Prism, and a nonparametric Mann–Whitney U test was used to evaluate statistical significance.

To ascertain the magnitude of the CD8⁺ T cell responses and the effector functions induced by the GP-encoding vaccines, Ad-Ii-GP and Ad-GP, C57BL/6 mice were vaccinated with each of the two adenoviral constructs. Fourteen days later, splenocytes were harvested and analyzed by ICS and flow cytometry to determine the numbers of GP_33–41-specific CD8⁺ T cells secreting IFN-γ and undergoing degranulation upon peptide stimulation. Degranulation was evaluated by incubation with anti-CD107a Abs during the 5 h of peptide stimulation, because CD107a molecules (also known as LAMP-1) are exposed on the cell surface following exocytosis of granules and then reinternalized by the cell, resulting in a continuous uptake of anti-CD107a Abs by the degranulating cells (27).

As shown in Fig. 1, both Ad-Ii-GP and Ad-GP vaccination resulted in significant induction of CD8⁺ T cells double positive for IFN-γ secretion and degranulation (IFNγ⁺CD107a⁺). Furthermore, linkage of Ii significantly increased the number of IFNγ⁺CD107a⁺ CD8⁺ T cells compared with the unlinked vaccine (p < 0.01).

To study the timeline of in vivo protection, challenge studies were conducted in which the bacterial load was evaluated on day 3 postinfection. For challenge studies, C57BL/6 mice were anesthetized on day 0 for vaccination with Ad-Ii-GP, Ad-GP, or Ad-Ii-NP (irrelevant Ag); mice were challenged i.v. with ≈10⁵ CFU Lm-GP33 at 3, 5, 35, 60, 100, and 210 dpv. The results of these challenge studies are presented in Fig. 2. Early after vaccination, at 3 dpv, no in vivo protection was elicited from either vaccine (Fig. 2A). This changed over the next 2 d, resulting in a reduction in spleen and liver titers from ≈10⁸ CFU/g in unvaccinated mice to ≈10²–10⁴ CFU/g in Ad-Ii-GP–vaccinated animals. In contrast, vaccination with Ad-GP was not associated with significant protection when assessed at 5 dpv (Fig. 2B). As evident from the challenge carried out, at 35 dpv the Ad-GP vaccine eventually induced significant protection compared with naive mice, reducing the bacterial burden ≈100-fold. This protection was further enhanced by Ii linkage of GP, resulting in an additional ≈100-fold decrease in the bacterial load at this time point (Fig. 2C). Notably, protection associated with the unlinked vaccine was lost again at 60 dpv (Fig. 2D). In contrast, mice vaccinated with the Ii-linked vaccine retained significant protection at 60 dpv, although the degree of this protection had faded compared with that seen after 35 d (Fig. 2C). At 100–120 dpv, significant protection could still be observed in Ad-Ii-GP–vaccinated mice in some experiments (e.g., Fig. 2E), whereas, in others (see later discussion), no protection could be demonstrated by day 3 after challenge, indicating that there is a critical limit to the longevity of clinical protection as evaluated in the short-term protection format used in the majority of our experiments. If, in contrast, the challenged mice were followed until day 5 postinfection, the impact of Ad-Ii-GP vaccination became more marked. At this point, mice had either died of bacterial infection or had reduced bacterial loads compared with those at 3 dpi. Although all unvaccinated mice and mice vaccinated with the unlinked vaccine died on day 3–4 postinfection, this was the case for only one mouse vaccinated with the Ii-linked vaccine (Fig. 2E). At 210 dpv, a single vaccination using either construct was no longer sufficient to afford significant protection (Fig. 2F).

In an attempt to correlate the kinetics of in vivo protection with parameters of CD8⁺ T cell–mediated immunity, mice were vaccinated as described; 30, 60, and 120 d later, mice were sacrificed, and splenocytes were harvested and analyzed by ICS and flow cytometry to determine the numbers and phenotype of GP_33–41-specific CD8⁺ T cells secreting IFN-γ and undergoing degranulation upon peptide stimulation (Fig. 3).

When splenocytes were analyzed at 30 dpv, the population of cytokine-secreting GP_33–41-specific CD8⁺ T cells had already contracted compared with the size of this population at the peak of the primary response (Fig. 1). A further significant decrease in cell numbers was observed between 30 and 60 dpv, whereas no additional contraction was observed between 60 and 120 dpv. With regard to phenotype, GP_33–41-specific CD8⁺ T cells were classified according to their expression of CD127 (IL-7R) and/or KLRG-1; these markers were used previously to divide Ag-specific CD8 T cells into four major subsets (28): double-negative cells, representing very early effector cells; CD127⁻ KLRG-1⁺ cells, representing short-lived effector cells; CD127⁺KLRG-1⁻ cells, representing long-lived memory precursor cells; and, finally, double-positive (CD127⁺KLRG-1⁺) cells, which, despite expressing a marker of terminal differentiation (KLRG-1), tend to survive into the memory phase, particularly under conditions of prolonged Ag stimulation (28–30). As can be seen in Fig. 3B, there is a relative decrease in short-lived effector cells and a similar relative increase in memory precursor cells between 30 and 60 dpv, as would be expected with transition into a memory state (28); no further change in phenotype distribution was observed between 60 and 120 dpv, and the fraction of Ag-specific, cytokine-producing CD8⁺ T cells, which degranulate upon Ag stimulation, remained stable throughout the time frame studied (Fig. 3C).

To determine the effector mechanisms underlying in vivo protection, challenge studies were conducted in which IFNγ^−/−, Prf^−/−, IFNγ^−/−Prf^−/−, and WT mice were vaccinated in parallel with Ad-Ii-GP. At 60 dpv, half of the WT mice were depleted of CD8⁺ T cells, and all mice were challenged according to our standard protocol (Fig. 4). In both liver and spleen, CD8⁺ T cell–depleted and Prf^−/− mice exhibited significantly higher bacterial loads compared with immunological intact, vaccinated controls. However, these bacterial loads were still significantly lower than were those of unvaccinated mice, with the exception of bacterial loads in the liver of CD8⁺ T cell–depleted mice. This residual protection may reflect the vaccine-induced protection mediated by Ag-specific CD4⁺ T cells secreting IFN-γ for the activation of infected macrophages, as well as IFN-γ secretion of CD8⁺ T cells in the case of Prf^−/− mice. Knockout of IFN-γ and double knockout of IFN-γ and perforin abolished vaccine-induced protection, as well as resulted in higher bacterial titers in the liver compared with unvaccinated controls, in addition to increased mortality. This likely reflects the crucial role of IFN-γ secretion from NK cells as part of the innate immune response toward L. monocytogenes (31).

Revaccination is commonly used to prolong protection against human pathogens. In an attempt to induce prolonged protection from vaccine-induced Ag-specific T cells, C57BL/6 mice were vaccinated with Ad-Ii-GP on day 0, revaccinated with the same vector 60 d later, and challenged with the normal dose of bacteria 120 d later (180 dpv) (Fig. 5). Except for a single outlier, protection at 180 dpv was very efficient following boosting of the Ag-specific CD8⁺ T cell response. In contrast, mice receiving a single vaccination on either day 0 or day 60 showed no significant protection compared with unvaccinated mice. As an alternative to homologous boosting, we also tested boosting with vaccinia virus expressing LCMV GP and observed at least comparable protection (data not shown); together, these observations clearly show that Ad5-induced protection against L. monocytogenes infection may be prolonged through relevant boosting.

We next wanted to extend our vaccine strategy to the setting of a natural L. monocytogenes infection (Lm-wt) (i.e., avoid exogenous Ag like GP). Based on analysis of the available information regarding known L. monocytogenes epitopes in the soluble L. monocytogenes Ags LLO and p60 (Table I, targeted epitopes have been highlighted), we designed truncated versions of these molecules to be expressed from our Ad5 vector, either alone or tethered to Ii (Table II contains the nucleic acid sequences of the inserts; LLO_1–351 and p60_198–482). These Ags were selected based on their ability to induce strong CD8⁺ and CD4⁺ T cell responses during natural L. monocytogenes infection in both H-2^d and H-2^b mice (32) and because they are both secreted Ags, a factor that was found to be important in conveying protection against L. monocytogenes infection (33, 34).

To validate the immunogenicity of these L. monocytogenes–specific vaccines, we used ICS and flow cytometry to determine the frequency and numbers of Ag-specific IFN-γ–secreting CD4⁺ and CD8⁺ T cells. Groups of BALB/c mice were vaccinated with each of the four constructs made (Ad-LLO^+/− Ii, Ad-p60^+/− Ii), whereas C57BL/6 mice were vaccinated with only the two LLO-containing constructs (Ad-LLO^+/− Ii), because no CD8⁺ epitopes and only weak CD4⁺ epitopes have been detected for p60 in this mouse strain (Table I). T cell responses were evaluated 13 dpv, and representative results for CD8⁺ and CD4⁺ T cells are shown in Figs. 6 and 7, respectively. Total numbers of Ag-specific cells in spleens, as calculated from the attained frequencies, are shown in Fig. 8. As expected, the well-known epitope LLO_91–99 induced a high frequency of activated CD8⁺ T cells in BALB/c mice. This response had only little to gain from Ii linkage of the Ag (Figs. 6A, 8). In contrast, CD8⁺ T cell responses to the two epitopes—p60_217–225 and LLO_296–304—benefitted greatly from Ii linkage (Figs. 8, and 6B, 6C, respectively). Strong CD4⁺ T cell responses were absent, and only two epitopes (LLO_189–200 and LLO_318–329) seemed to elicit fair responses when the Ags were Ii coupled (Fig. 7).

Having established that relevant immune responses were induced from our L. monocytogenes–specific vaccines, challenge studies were conducted in which all four L. monocytogenes–specific vaccines were used for immunization of BALB/c mice, and vaccines encoding LLO were used for immunization of C57BL/6 mice. Mice were challenged with Lm-wt at 6, 30, and 60 dpv (Fig. 9). Because BALB/c mice are rather susceptible to L. monocytogenes infection, these mice were challenged with a reduced dose (≈3 × 10⁴ CFU) of Lm-wt compared with that used in C57BL/6 mice (≈10⁵ CFU).

Early after vaccination, at 6 dpv, both LLO-encoding vaccines provided marked protection in BALB/c mice, with Ii linkage improving upon the already significant effect of vaccination with the unlinked vaccine. In contrast, the unlinked p60 vaccine was associated with limited protection this early after vaccination, and marked improvement was obtained through linkage of this Ag to Ii; indeed, vaccination with the linked vaccine brought the bacterial loads in p60-vaccinated mice down to the level seen in mice given the stronger Ag, LLO. In the case of vaccination against LLO in C57BL/6 mice, Ii linkage also significantly improved the early protection induced by adenovector immunization (Fig. 9A).

At 30 dpv (Fig. 9B), significant protection was seen in LLO-vaccinated BALB/c mice, regardless of Ii linkage, which did not result in significantly lower bacterial loads in the spleen, whereas protection in the liver was significantly improved. Ad-Ii-p60 vaccination of BALB/c mice resulted in significant protection in both spleen and liver, whereas vaccination with the Ad-p60 vaccine resulted in significantly lower bacterial loads only in the liver. In C57BL/6 mice, the unlinked LLO vaccine induced protection, which was significantly enhanced by Ii linkage. For this experiment, we included a control group of mice vaccinated with the Ad-Ii-GP vaccine to exclude peptide-unspecific protection; as expected, bacterial loads in the organs did not differ between these mice and unvaccinated controls.

At 60 dpv (Fig. 9C), we found induced protection from all vaccines, with the exception of Ad-p60, in BALB/c mice. Overall, the vaccine-induced protection had decreased compared with that observed at 30 dpv, with the exception of the protection in the spleen following Ad-Ii-LLO vaccination in BALB/c mice; in this organ, Ii linkage led to sustained protection matching that observed on 30 dpv. This was not the case in the liver: both Ad-Ii-LLO– and Ad-LLO–induced protection had faded relative to day 30, and a significant effect of Ii linkage was no longer observed. Similarly, p60 vaccination resulted in fading of protective efficiency for both the linked and the unlinked vaccine, with a significant advantage associated with Ii linkage of the target Ag. In C57BL/6 mice, reduced, but significant, protection was sustained for mice given either vaccine variant (Ad-Ii-LLO or Ad-LLO), with no significant difference observed between the two groups.

To get a better biological context for evaluating the efficiency of the protective responses elicited through adenoviral vaccination, we compared the in vivo protection in C57BL/6 mice vaccinated with either Ad-Ii-LLO or a sublethal dose (≈10³ CFU) of live Lm-wt 30 d earlier. In this experiment, we also included a group of mice receiving Ad-Ii-LLO in the right hind footpad combined with Ad-Ii-p60 vaccination in the left hind footpad. As shown in Fig. 10, we found no significant differences in in vivo protection among the three intervention groups.

Considering that no p60 CD8⁺ epitopes and only weak p60 CD4⁺ epitopes have been reported in C57BL/6 mice, it is perhaps not surprising to find that added vaccination with a vaccine targeting this molecule would not improve the overall clinical protection in this mouse strain. Therefore, to more formally investigate whether combinations of vaccine vectors represent a way to improve in vivo protection, we vaccinated BALB/c mice with Ad-Ii-LLO, Ad-Ii-p60, or both and challenged the mice 60 d later with a lethal dose of Listeria. Organs were removed 3 d later, and the bacterial loads were determined (Fig. 11). In this case, both vaccines should have induced significant protection on their own, which was confirmed; nevertheless, even under these conditions no additive effect of vaccine combination was demonstrated.

In this study, we demonstrated that adenovirus-based vaccines targeting endogenous Ags can be used to induce protection against the intracellular bacterium, L. monocytogenes, and moreover, that linkage of a relevant Ag to MHC class II–associated Ii may markedly augment vaccine-induced protection, particularly when subdominant epitopes are targeted by the vaccine.

CD8⁺ T cell–mediated clearance of L. monocytogenes is known to rely upon two mechanisms working in synergy: perforin-dependent cytolysis of infected cells via the granule exocytose pathway and IFN-γ secretion for activation of macrophages to ingest and kill L. monocytogenes (14, 17–19). Expanding on published data, adenovirus vaccines were shown to induce cytotoxic CD8⁺ T cells capable of secreting IFN-γ and undergoing degranulation upon antigenic peptide stimulation (30), and total numbers of primed Ag-specific cells in the spleen were significantly increased from Ii linkage (Fig. 1). Experimental analysis using CD8⁺ T cell depletion and various knockout mice identified CD8⁺ T cells capable of secreting IFN-γ and performing perforin-dependent cytolysis as being the main contributors to adenovirus-induced protective immunity against listerial infection. IFN-γ–secreting CD4⁺ T cells may also contribute to protection, because significant protection was observed in the vaccinated mice, which were depleted of CD8⁺ T cells ahead of Lm-GP33 challenge (Fig. 4); however, as also indicated by the very limited protection in Prf^−/− mice, the impact of these cells must be marginal.

Regarding the impact of tethering to Ii, we observed accelerated, augmented, and prolonged protection, similar to what was seen previously in the context of viral infections (e.g., LCMV) (4). Unfortunately, the vaccine-induced protection faded with time, and ∼4 mo after vaccination, little protection was observed using our standard challenge set-up. This does not mean that protection is completely gone, and it is very likely that we would still be able to demonstrate significant in vivo protection using lower challenge doses and a prolonged observation period (Fig. 2E). Nevertheless, this observation points to a serious limitation of T cell–based vaccines, which was noted before (35), but is often ignored; in the majority of experimental vaccine studies protection is not evaluated beyond 2 mo postvaccination, at which time serious deficiencies may not yet be clearly evident. Interestingly, confirming analyses performed earlier in our laboratory, we found that numbers of GP-specific CD8⁺ T cells following vaccination with the Ad-Ii-GP construct decreased between 30 and 60 dpv but then remained fairly constant in the interval from 60 to 120 dpv (4, 30), despite the marked change in in vivo protection during that interval. This implies that a change in the functional state of the vaccine-primed cells is the major cause of the fading in in vivo protection, although we have not been able to identify any marker(s) clearly linked with this functional decline (Fig. 3, M.A. Steffensen and A.R. Thomsen, unpublished observations).

Reducing this early fading of the protective qualities of the Ag-specific cells is crucial for a future vaccine to be clinically relevant. To this end, we explored homologous boosting of the induced T cell response and observed a level of protection at 180 dpv that matched the level of acute protection following a single vaccination (Fig. 2C). This result represents proof-of-concept regarding the substantial boosting effect of Ii-linked adenovirus vaccination, despite previous encounter with the vector and resultant pre-existing vector immunity. These results support and extend our previously published evidence indicating that adenoviral vectors can induce efficient CD8⁺ T cell memory, even in individuals with pre-existing vector immunity (36). Boosting of the primary Ad-Ii-GP–induced response at 60 dpv may also be obtained using a heterologous vector, vaccinia virus encoding GP, providing similarly low bacterial titers in challenged mice (data not shown). Taken together, these results suggest that there is ample opportunity to boost adenovirus-induced immunity and, thus, obtain prolonged protection.

To further underpin the potential clinical relevance of our results obtained in the GP vaccination setting, we produced novel L. monocytogenes–specific vaccines encoding the two well-known listerial Ags—LLO and p60—with and without Ii linkage. ICS studies validated the immunogenicity of these vaccines targeting real bacterial Ags. Furthermore, the results obtained strengthen our view that the concept of augmented immunity resulting from Ii linkage of the target Ag in an adenovirus vaccine is epitope specific, rather than Ag/protein specific. Thus, we found a marked increase in the CD8⁺ T cell response to LLO_296–304, whereas little increase was noted with regard to the response to another epitope from the same molecule: LLO_91–99. Strong induction of IFNγ⁺CD44^high CD8⁺ T cell responses was further translated into low bacterial titers assayed in the liver and spleen 3 d post–Lm-wt challenge. Hence, impressive protection was noted in LLO-vaccinated BALB/c mice, benefitting from Ii linkage both during the acute response (6 and 35 dpv, Fig. 9A, 9B) and when the memory phase is reached (60 dpv, Fig. 9C). In addition, greatly augmented protection resulting from Ii linkage was observed with regard to the weaker Ags: p60- and LLO-vaccinated BALB/c and C57BL/c mice, respectively.

Clearly, the comparison of the Ii-linked adenovirus vaccines with matched unlinked vaccines reveals the remarkable benefit of Ii linkage, in terms of both bacterial loads and overall survival. Nevertheless, to really assess the degree of protection resulting from adenoviral vaccination, we also chose to compare the induced protection with that induced by a previous nonlethal L. monocytogenes infection. Comparing the protection induced by Ad-Ii-LLO vaccination in C57BL/6 mice with what we perceive as a kind of gold standard, the Ad-Ii-LLO construct induced protection at the same level as did previous infection with live bacteria (Fig. 10).

A future human vaccine, whether based on the adenovirus vector or any other delivery system, would undoubtedly benefit from the inclusion of more than one pathogen-specific Ag. This could be required to obtain a maximal T cell response involving as many pathogen-specific cells as possible. However, when tested in our model system, we could not document improved protection when the Ad-Ii-p60 and Ad-Ii-LLO vaccines were coadministered (Figs. 10, 11). In C57BL/6 mice, the lack of improved protection could readily be explained by the fact that no p60 CD8⁺ epitopes and only weak p60 CD4⁺ epitopes were reported in this mouse strain (Table I). However, even in the BALB/c setting, in which the individual vaccines induced significant protection, no significant advantage from the Ag combination could be documented (Fig. 11). At first glance, these observations might seem to argue against an advantage of including several Ags in the same vaccine; however, these results are representative of only inbred populations, which are atypical in that it can be ascertained beforehand whether the chosen Ag will induce a strong host response. This cannot be done in an outbred population (i.e., humans); therefore, the main reasoning behind the use of multiple Ags is to avoid vaccine nonresponders rather than to improve the immune response in an individual vaccinee. Consequently, a mixture of Ags may still be relevant in the clinical setting.

In a natural setting, listerial infection is typically acquired as a food-borne infection; for that reason, a weakness of our study is that we have not tested vaccine-induced protection in the context of an oral challenge. Although our results strongly indicate that systemic spreading of the infection is inhibited by prior vaccination, we cannot rule out that infection via a mucosal route and i.v. infection may occur through entirely different pathways involving different cell types and routes of dissemination. For this reason, future studies on the resistance to oral infection, as well as the possibility of applying the vaccine orally, will be necessary to fully evaluate the relevance of these results in a human vaccine setting.

Developing effective human vaccines capable of eliciting protective, broad, and long-lasting cellular immune responses are crucial in the combat against the most resilient microbial threats. Vaccines based on attenuated virus vectors pose many advantages in this respect, mimicking natural infections. In this study, we extended the concept of Ii linkage of a target Ag as a way to augment and prolong protective T cell responses. In particular, we find that, upon Ii linkage, CD8⁺ T cell responses to subdominant epitopes may be elevated from being rather irrelevant to gaining biological weight in terms of total numbers of activated IFNγ⁺CD44^high CD8⁺ T cells and, more importantly, in terms of in vivo protection. This feature is of great value in a fight against pathogens capable of mutating to avoid T cell recognition. In the present setting involving protection against L. monocytogenes infection, we found impressive protection from Ii-coupled constructs, which match previous immunization with live L. monocytogenes (Fig. 10); although the protective capacity faded over time (Fig. 2), strong protection could be retained for ≥6 mo by boosting the T cell response on day 60 postinitial vaccination.

We thank Peter Overbeck Rasmussen and Annika Elisabeth Lindkvist for expert technical assistance.

The University of Copenhagen holds a patent regarding the Ii vaccine strategy. A.R.T. and J.P.C. are entitled to a fraction of any net income that may derive from the commercialization of this patent. The other authors have no financial conflicts of interest.