Recombinant Listeria monocytogenes expressing a single immune-dominant peptide confers protective immunity to Herpes Simplex Virus-1 infection

Abstract

The vast majority of the world's population is infected with herpes simplex virus (HSV). Although antiviral therapy can reduce the incidence of reactivation and asymptomatic viral shedding, and limit morbidity and mortality from active disease, it cannot cure infection. Therefore, the development of an effective vaccine is an important global health priority. In this study, we demonstrate that recombinant Listeria monocytogenes (Lm) expressing the H-2K^b gB_498−505 peptide from HSV-1 triggers a robust CD8 T cell response to this antigen resulting in protective immunity to herpes simplex virus infection. Following challenge with HSV-1, immune competent mice primed with recombinant Lm expressing gB_498−505 antigen were protected from HSV-induced paralysis. Protection was associated with dramatic reductions in recoverable virus, and early expansion of HSV-1 specific CD8 T cells in the regional lymph nodes. Thus, recombinant Lm expressing antigen from HSV represents a promising new class of vaccines against HSV infection.

INTRODUCTION

Herpes simplex virus (HSV) types 1 and 2 are ubiquitous human pathogens. Depending on the population examined, between 50 to 100% of adults have seroevidence of infection with HSV-1 or HSV-2 (1). After primary infection in immune competent hosts, the virus establishes life-long latency within the sensory ganglia that is associated with sporadic recurrences of clinical lesions and asymptomatic virus shedding. However, infection in neonates or other immune compromised hosts commonly causes disseminated infection resulting in a high rate of morbidity and mortality. While antiviral therapy can reduce morbidity and mortality in disseminated infection, and long-term suppressive therapy can reduce the frequency and severity of viral reactivation, antivirals cannot not “cure” infection (2, 3). Furthermore, the increasing incidence of HSV resistant to common antiviral medications emphasizes the current need for an effective vaccine to prevent HSV infection (4, 5).

CD8 T cells contribute to protective immunity to HSV infection. In biopsy specimens from humans with recurrent HSV infection, viral clearance is associated with a high concentration of local CD8 T cells with cytolytic activity against infected cells (6). In animal models, depletion of CD8 T cells impairs clearance of virus from the central nervous system, while TCR transgenic CD8 T cells specific for the immune dominant H-2K^b restricted peptide in HSV-1 glycoprotein B gB_498−505 transferred into mice lacking other components of adaptive immunity results in viral clearance (7, 8). These studies demonstrate that HSV-specific CD8 T cells play a protective role in HSV infection.

Infection with the Gram-positive intracellular bacterium Listeria monocytogenes (Lm) is a well-characterized experimental model in which Lm-specific CD8 T cells can confer protective immunity to secondary Lm infection (9, 10). Moreover, recombinant Lm expressing defined antigens from other intracellular viral pathogens such as LCMV, influenza, HIV, simian immune deficiency virus, or feline immune deficiency virus primes CD8 T cells specific for these heterologous antigens that protect against subsequent viral challenge (11-17). Accordingly, in this study we examined the ability of recombinant Lm expressing a single immune dominant antigen from HSV-1 to prime HSV-specific CD8 T cells and confer protective immunity to HSV challenge.

MATERIALS AND METHODS

Bacteria

Lm ΔactA strain DPL1942 and recombinant Lm secreting the ovalbumin protein behind the Lm hly promoter (Lm-OVA) have been described (18, 19). Transformation of Lm was performed by penicillin treatment as described (20). For infections, Lm was grown and sub-cultured in BHI media containing chloramphenicol (20μg/ml) to early log-phase (OD₆₀₀ 0.1), washed and diluted in PBS to a final concentration of 1 × 10⁶ CFUs per 200 μl and inoculated intravenously into mice.

Expression constructs

The DNA fragment encoding ovalbumin, HA tag, Lm hly promoter and signal sequence was PCR amplified from Lm-OVA (19) using the following primers: 5' tctagattaacatttgttaacgacgac 3' and 5' ggatccttaaggggaaacacatctgcc 3', and cloned into the Xba1 and BamH1 sites in the low copy vector pAM401 containing resistance to chloramphenicol (21) (Figure 1A). This construct was then cut with Pst1 and Stu1, and ligated with the overlapping primer sets for pHSVgB (coding strand: 5'- accacctcctccatcgagttcgcccggctgcagtttacagg-3', non-coding strand 5'- cctgtaaactgcagccgggcgaactcgatggaggaggtggttgca-3') (Figure 1A); and pCONTROL (coding strand: 5' atgacagagcagcagtggaatttcgcgggtatcgaggccgcggcaagcgcaatccagggaaatgtagg-3', non-coding strand 5'-cctacatttccctggattgcgcttgccgcggcctcgatacccgcgaaattccactgctgctctgtcattgca-3'). The relevant portions of these constructs were verified by DNA sequencing.

Figure 1.

Generation of Lm ΔactA pHSVgB or Lm ΔactA pCONTROL A. Construct map for creating recombinant HA-tagged fusion proteins containing HSV-1 gB_498−505 and mTB ESAT6₁₋₂₀ peptides that are expressed and secreted under the Lm hly promoter and signal sequence within the pAM401 vector (cat, chloroamphenicol acetyltransferase). B. Western blot of supernatant protein from Lm ΔactA pHSVgB (lane 1), Lm ΔactA pCONTROL (lane 2), and Lm-OVA (19) with anti-HA Ab.

Western blotting

Supernatant protein was prepared by filtering (0.2 μm syringe filter) Lm culture media (BHI) with bacteria in log-phase growth (2 to 4 hours after 1:100 back-dilution of stationary phase cultures to OD₆₀₀ 0.4 to 0.6), trichloroacetic acid precipitation (10%), and separation by SDS gel electrophoresis. Proteins where transferred to nitrocellulose and probed with rabbit anti-HA (clone HA-11, Covance, Princeton, NJ).

Herpes Simplex Virus

HSV-1 (KOS strain) viral stocks were prepared for hind footpad infection (2.5 × 10⁶ PFU per footpad) following dermal abrasion, as described (22). After infection, mice were monitored twice daily for 14 days for HSV central nervous system disease manifested as ataxia and/or hind limb paralysis. Our previous studies have demonstrated that > 80% of mice that develop these symptoms later succumb to infection, and accordingly paralyzed mice were euthanized in accordance with our IACUC protocol. For determining tissue HSV titers, the hind footpads and spinal cord were harvested, snap frozen, homogenized, and titered on Vero cells (22).

Mice

Female C57Bl/6 (H-2^b) and IFN-γ-deficient mice on the C57Bl/6 background were purchased from the Jackson Laboratory. MyD88-deficient mice were a gift from Dr. S. Akira (Osaka University) and were backcrossed onto the C57Bl/6 background for at least 10 generations. All mice were maintained in the University of Washington specific pathogen free facility. For gB peptide immunization, 100μg of purified peptide in 100μl saline was mixed with 100 μl alum (Imject Alum, Pierce) and inoculated i.p. All studies were approved by the University of Washington Institutional Animal Care and Use Committee.

Quantification of CD8 T cell response

HSV-1 gB_498−505-specific CD8 T cells were analyzed in peripheral blood by staining with H-2K^b Dimer × loaded with gB_498−505 peptide according to the manufacturer's instructions (BD Biosciences). For intracellular cytokine staining, single cell suspensions of splenocytes or cells from the draining lymph nodes were incubated in the presence of the indicated peptides (10⁻⁶ M) and monensin (GolgiStop reagent, BD Bioscience) for 5 h, surface stained, permeabilized (Cytoperm solution, BD Biosciences), stained for intracellular IFN-γ.

Cytokine production

The concentration of IFN-γ in splenocyte culture supernatants (5 × 10⁶ cells/ml) was quantified 72 hours after peptide stimulation by ELISA using reagents from R&D Systems (Minneapolis, MN).

Statistics

The differences in mean viral PFUs, the percentages and numbers of cells, and cytokine concentrations were determined using the Student t test. The difference in survival between groups of mice were determined using Log Rank Chi square test (Prism, GraphPad software).

RESULTS

Generation of Lm expressing HSV gB_498−505

To test the ability of recombinant Lm expressing HSV antigen to trigger HSV-specific CD8 T cells, we engineered an expression construct (designated pHSVgB) containing the HSV-1 peptide gB_498−505 secreted as a HA-tagged recombinant protein expressed under the Lm hly promoter and signal sequence, within the Gram positive bacterial vector pAM401 (21), (Figure 1A). In a similar fashion, we inserted the coding sequence for a peptide antigen from an irrelevant pathogen (Mycobacterium tuberculosis) into the same Lm expression construct that was used as a control, pCONTROL (Figure 1A). Since the ultimate goal of these studies was to evaluate the potential of using recombinant Lm as live attenuated vaccine vectors, we transformed either pHSVgB or pCONTROL into a ΔactA Lm strain, DPL-1942. We and others have shown that the ΔactA Lm mutant primes a robust Lm specific CD8 and CD4 T cell response, and clearance of this strain does not require components of innate immunity such as MyD88 or IFN-γ that are normally critical for protection from WT Lm infection (23, 24). Lm ΔactA transformed with either pHSVgB (Lm ΔactA pHSVgB) or pCONTROL (Lm ΔactA pCONTROL) each secreted a protein of predicted size (∼19kD) into culture supernatants as detected by blotting with anti-HA antibody (Figure 1B).

Lm ΔactA pHSVgB triggers antigen-specific CD8 T cell expansion

We next examined the gB-specific immune response triggered by Lm ΔactA pHSVgB in comparison with the immune response triggered by gB peptide administered in alum. By day 6 after inoculation with 10⁶ CFUs of Lm ΔactA pHSVgB, gB-specific CD8 T cells were detectable in the peripheral blood. The magnitude of gB-specific CD8 T cell expansion peaked at day 8, began to contract by day 13, and reached levels ∼20% of day 8 levels at day 28 post-infection (Figure 2A). For comparison, mice administered gB peptide in alum or mice infected with Lm ΔactA pCONTROL had only background level of gB-specific CD8 T cells at each of these time points.

Figure 2.

Induction of HSV-specific CD8 T cells after infection with Lm ΔactA pHSVgB A. The percentage of HSV-gB_498−505 specific CD8 T cells in the peripheral blood determined by staining with H-2K^b dimer × loaded with gB_498−505 peptide at the indicated time points after inoculation with 10⁶ CFUs of either Lm ΔactA pHSVgB or Lm ΔactA pCONTROL, or gB peptide (100μg) in alum. B. IFN-γ production by CD8 and CD4 T splenocytes from mice day 8 after infection with 10⁶ CFUs of either Lm ΔactA pHSVgB or Lm ΔactA pCONTROL after re-stimulation with gB_498−505 peptide and LLO_189−201 peptide, respectively. Numbers in the upper right hand quadrant indicate the mean percentage (± standard error) of IFN-γ+ cells of total CD8+ or CD4+ T cells for five mice per group from two separate experiments. C. Absolute number of IFN-γ producing CD8 T cells per spleen day 8 after infection with Lm ΔactA pHSVgB or Lm ΔactA pCONTROL determined by intracellular cytokine staining. D. IFN-γ concentration in splenocyte culture supernatants from day 8 Lm ΔactA pHSVgB or Lm ΔactA pCONTROL infected mice after 72-hour stimulation with gB_498−505, LLO_189−201, or no peptide.

To further evaluate the HSV-specific CD8 T cell response triggered by Lm ΔactA pHSVgB infection, we examined the antigen-specific response in splenocytes at the peak of the T cell response (day 8). At this time point, CD8+ splenocytes from Lm ΔactA pHSVgB-infected mice readily produced IFN-γ in response to stimulation with gB_498−505 peptide as determined by both intracellular cytokine staining and ELISA, while splenocytes from Lm ΔactA pCONTROL infected mice produced only background amounts of cytokine (Figure 2B-D). Thus, infection with Lm ΔactA pHSVgB primes a robust CD8 T cell response to gB_498−505 in B6 mice. To confirm that mice infected with Lm ΔactA pHSVgB and Lm ΔactA pCONTROL only differed by the CD8 T cell response to gB_498−505, we examined the response to the endogenous listeriolysin-O peptide, LLO_189−201, presented by MHC class II (Figure 2B,D). Similar frequencies of IFN-γ producing CD4 T cells and total IFN-γ production were found in both infection groups after stimulation with this peptide.

Infection with Lm ΔactA pHSVgB confers protective immunity to HSV-1 infection

To examine if the gB-specific CD8 T cell response triggered by Lm ΔactA pHSVgB infection confers protection to HSV-1 infection, groups of mice infected with either Lm ΔactA pHSVgB or Lm ΔactA pCONTROL were infected in the hindfoot pads 28 days later with an inoculum of HSV-1 that normally causes ataxia and/or hindlimb paralysis in naïve mice. Many features of disease pathogenesis in human infection are represented in this acute infection model. After infection virus travels anterograde through the enervating sciatic nerve to the dorsal root ganglia, replicates in the ganglia, and then returns to the site of infection via retrograde axonal transport resulting in a primary lesion of the footpad (22). Virus in the dorsal root ganglia can also cross the synapse, enter the spinal cord, and ascend to the brain causing paralysis. In the first seven to nine days after HSV infection, 85% (17 of 20) of mice primed with Lm ΔactA pCONTROL developed hind limb paralysis compared with only 25% (5 of 20) of mice primed with Lm ΔactA pHSVgB (P = 0.0002) (Figure 3A). These mice were monitored for up to 14 days after infection, and no additional paralysis developed for any mice beyond day 9 after HSV infection.

Figure 3.

Lm ΔactA pHSVgB protects mice from lethal challenge with HSV-1. A. Paralysis-free survival after HSV-1 footpad inoculation (2.5 × 10⁶ PFUs) in mice primed 28 days previously with Lm ΔactA pHSVgB (squares) or Lm ΔactA pCONTROL (triangles). Difference in survival between these two groups, P = 0.0002. These data represent 20 mice per experimental group pooled from two independent experiments with similar results. B. Recoverable HSV-1 titers in the footpad and spinal cord on day 6 after HSV infection in naïve mice, or mice primed 28 days previously with Lm ΔactA pHSVgB, Lm ΔactA pCONTROL, or gB peptide in alum. Percentage of CD8 T cells in draining popliteal lymph nodes producing IFN-γ after stimulation with gB_498−505 peptide (black bars) or no peptide (white bars) on day 3 (C) and day 6 (D) after footpad inoculation with HSV-1 (2.5 × 10⁶ PFUs) in naïve mice, or mice primed 28 days previously with Lm ΔactA pHSVgB or Lm ΔactA pCONTROL. E. Recoverable HSV-1 titers in the footpad and spinal cord on day 6 after HSV infection in IFN-γ-deficient or MyD88-deficient mice primed 28 days previously with Lm ΔactA pHSVgB or Lm ΔactA pCONTROL.

To determine if protection from paralysis was directly related to reductions in viral burden, we quantified the amount of recoverable virus just before mice develop paralysis (day 6). Mice primed with Lm ΔactA pHSVgB, when compared with mice primed with Lm ΔactA pCONTROL, gB peptide in alum, or naïve mice, had ∼10-fold and ∼200-fold reductions in HSV-1 titers in the footpad and spinal cord, respectively (Figure 3B). Additionally, these protective effects of prior Lm ΔactA pHSVgB infection were associated with a rapid and robust expansion of gB-specific CD8 T cells in the draining popliteal lymph nodes in response to HSV-1 challenge. By day 3 after HSV-1 infection, 1.2% of CD8 T cells in the lymph nodes from Lm ΔactA pHSVgB primed mice produced IFN-γ in response to gB_498−505 peptide stimulation, while lymph node cells from Lm ΔactA pCONTROL primed or naïve mice had no detectable antigen-specific response (Figure 3C). By 6 days post-infection, the gB-specific response in popliteal lymph node cells reached ∼15% of total CD8 T cells in Lm ΔactA pHSVgB primed mice, compared with a ∼7% response in Lm ΔactA pCONTROL or naïve mice (Figure 3D). The delayed and dampened response in control mice represented the primary CD8 T cell response to HSV, which was not sufficient to protect them; the majority of these mice developed HSV-induced paralysis and ataxia and had markedly increased amounts of virus in both the central nervous system and peripheral tissues.

Lastly, we examined the ability of Lm ΔactA pHSVgB to trigger protective immunity in immune-deficient mice that are more susceptible to HSV-1 infection. In both IFN-γ-deficient and MyD88-deficient mice, an gB-specific immune response was readily detected by dimer staining after Lm ΔactA pHSVgB inoculation (data not shown), however no significant protection from HSV-1 challenge could be detected for these mice (Figure 3E). Taken together, these data demonstrate that recombinant Lm expressing a single peptide from HSV-1 confers protection to HSV-1 infection in immune competent mice.

DISCUSSION

Numerous properties make recombinant Lm an attractive vaccine vector candidate for priming antigen-specific CD8 T cells. First and foremost, Lm infection is a strong adjuvant, and the resultant antigen-specific CD8 T cell response to Lm or recombinant antigen is protective and long-lasting (13, 25). These properties are directly related to the ability of the bacterium to gain access to the cytoplasmic compartment of infected cells delivering antigens to the MHC class I antigen presentation pathway. Second, pre-existing immunity to Lm does not diminish the therapeutic efficacy of recombinant Lm strains(26, 27). Thirdly, genetic manipulation allowing for expression of heterologous antigen and targeted disruption of virulence factors is readily accomplished. Accordingly, numerous attenuated Lm strains that cause minimal disease yet maintain immunogenicity have been described (28, 29). Together, these qualities make attenuated Lm expressing recombinant antigen promising vaccine candidates for priming antigen-specific T cells.

The widespread prevalence of HSV infection combined with the lack of “curative” therapy and increasing resistance to standard antiviral therapy emphasizes the need for developing a vaccine that can prevent HSV infection (30, 31). In this study, we examined the potential for recombinant Lm expressing a single immune dominant MHC class I-restricted peptide from HSV-1 to prime HSV-specific CD8 T cells to reduce disease. We demonstrate that Lm ΔactA pHSVgB induces a robust CD8 T cell response to HSV gB with ∼2% of peripheral CD8 T cells specific for this heterologous antigen at the peak of primary expansion in immune competent and IFN-γ-deficient and MyD88-deficient mice. In response to HSV-1 challenge, immune competent mice were protected from HSV-1 induced disease, had marked reductions in the amount of recoverable virus, and a more rapid and robust expansion of antigen-specific CD8 T cells. Although recombinant LmΔactA pHSVgB triggered a robust gB-specific CD8 T cell response in IFN-γ-deficient or MyD88-deficient mice, these response were associated with little or protection against subsequent HSV-1 infection. These results are inconsistent with the readily achievable protective immunity to subsequent Lm infection in these mice after priming with Lm ΔactA (23, 24), and may reflect differences in CD8 T cell effectors, or other MyD88-or IFN-γ-dependent immune mechanisms required for adaptive immunity to HSV-1 compared with Lm.

To our knowledge, this is the first study demonstrating protection from infection-associated disease in addition to reduction in viral burden conferred by recombinant Lm. Ideally a HSV vaccine would both prevent new infection and be curative for established latent infection thereby preventing recurrent disease. However HSV rarely reactivates from latency in mice, preventing assessment of this aspect of the vaccine. Nevertheless, the data presented here represent an important first step in the development of recombinant Listeria as a novel class of vaccine vectors against HSV infection.