The Nucleoprotein Is Required for Lymphocytic Choriomeningitis Virus-Based Vaccine Vector Immunogenicity

Abstract

Recombinant glycoprotein-deficient lymphocytic choriomeningitis virus-based vaccine vectors (rLCMV/ΔGP) are potent CD8⁺ T cell inducers. To investigate the underlying molecular requirements, we generated a nucleoprotein-deficient vector counterpart (rLCMV/ΔNP). NP but not GP is a minimal trans-acting factor for viral transcription and genome replication. We found that, unlike rLCMV/ΔGP, rLCMV/ΔNP failed to elicit detectable CD8⁺ T cell responses unless NP was trans complemented in a transgenic host. Hence, NP-dependent intracellular gene expression is essential for LCMV vector immunogenicity.

TEXT

Lymphocytic choriomeningitis virus (LCMV), the prototypic arenavirus, has a negative-strand RNA genome with two segments, designated S and L, each of which carries two viral genes in ambisense orientation (wild-type LCMV [LCMV wt]) (Fig. 1A) (1). Viral particles contain S and L ribonucleoproteins (RNP) wherein the genomic RNAs are encapsidated by the nucleoprotein (NP) in association with the RNA-dependent RNA polymerase (RdRp) L. NP and L represent the minimal viral trans-acting factors for viral RNA transcription and replication (Fig. 1B), whereas the glycoprotein (GP) mediates receptor binding and cell entry (2, 3). Upon RNP release to the cytoplasm, the virion-contained RdRp initiates NP and L mRNA transcription from the 3′ untranslated region (UTR) promoters of the NP-encapsidated S and L segments (Fig. 1B), respectively, a process that is independent of de novo viral protein synthesis. Conversely, de novo NP synthesis is required for encapsidation of the nascent antigenome RNA species for them to serve as the templates for GP and Z mRNA transcription, respectively, and genome replication.

FIG 1.

NP-deficient LCMV vector genomes, predicted transcription-replication steps in noncomplementing cells, and growth in complementing and noncomplementing cells. (A) Schematic of the wild-type LCMV genome (LCMVwt; clone 13 strain[Cl.13]) and derived NP- and GP-deficient vector genomes used in this study. Arrows indicate ORFs, and inverted writing indicates antisense polarity. (B) mRNA transcription (dashed or dotted bent arrows) and genome replication steps (solid bent arrows) as they are predicted to occur in noncomplementing (not NP-expressing) cells. Dotted arrows, as opposed to dashed arrows, indicate very low levels of expected transcription due to the lack of template RNA amplification and an almost complete absence of viral protein expression. Red crosses indicate the steps that are predicted not to occur due to NP deficiency of the respective vector. (C) BHK-NP cells (clone 6) and wt BHK-21 cells were infected with rLCMV/ΔNP-GFP3′ at MOI = 0.01, and infectious virus or vector in supernatant was quantified by immunofocus assay (using GP-2-specific monoclonal antibody 83.6 [17]) on BHK-NP cells as the substrate. (D) BHK-NP cells were infected with the indicated virus and vectors, and infectivity was determined as described for panel C. Symbols in panels C and D indicate the mean results ± standard deviations (SD) from three tissue culture wells. Data from one representative experiment of two similar experiments are shown.

We have previously shown that deletion of the viral glycoprotein gene (GP) and its replacement by sequences encoding vaccine antigens offers a strategy to generate single-round infectious vaccine vectors (recombinant GP-deficient LCMV [rLCMV/ΔGP]) (Fig. 1A) (4–6). The production of infectious rLCMV/ΔGP by pseudotyping requires GP-expressing cell lines (e.g., stably GP-expressing BHK-21 cells [BHK-GP]). In the vaccinees, the lack of de novo GP synthesis prevents vector spread. Nevertheless, rLCMV/ΔGP induces polyfunctional and protective CD8⁺ T cell responses that reach a sizeable magnitude at doses as low as 300 PFU (4).

Here, we studied the viral molecular requirements for immunogenicity of single-round infectious LCMV vectors. Contrary to the notion that CD8⁺ T cell-inducing vaccines should produce antigens in cells of the vaccinee for direct presentation on major histocompatibility complex (MHC) class I (7), hydrogen peroxide-inactivated LCMV has recently been shown to induce protective CD8⁺ T cell immunity (8–10). Whether this reflects residual translation of viral proteins in the absence of productive replication or cross-presentation of particle-derived proteins remains undefined (9). Here, we tested specifically whether rLCMV vector immunogenicity depends on de novo synthesis of NP in vivo, which is required for viral genome replication and robust intracellular gene expression. To generate NP-deficient LCMV vectors (rLCMV/ΔNP), we engineered an S segment wherein the NP gene was replaced by a green fluorescent protein (GFP) reporter gene. The corresponding vector, designated rLCMV/ΔNP-GFP3′ (Fig. 1A), consisted of the engineered S segment together with a wt L segment and was rescued from cDNA using polymerase I- and II-driven plasmids (11). As a trans-complementing cell substrate, we utilized stably transfected NP-expressing BHK-21 cells (BHK-NP). rLCMV/ΔNP-GFP3′ reached titers of >10⁵ PFU/ml in BHK-NP cells but failed to amplify in normal BHK-21 cells, as expected (Fig. 1C and D). To differentiate between mRNA transcription from the virion-derived, NP-encapsidated S segment (reading from the 3′ UTR) and genome replication with subsequent transcription/replication from the antigenomic promoter (5′ UTR), we also generated rLCMV/ΔNP-GFP5′ (Fig. 1A and B). In this vector, GP was artificially expressed from the 3′ UTR and GFP from the 5′ UTR. Finally, to address potential attenuating effects on rLCMV/ΔNP-GFP3′ that might have resulted from deletion of cis-acting RNA elements in the NP open reading frame (ORF), we mutated the NP start codon of the wt S segment to a stop codon (rLCMV/NPΔAUG) (Fig. 1A and B). rLCMV/ΔNP-GFP5′, expressing GP from the 3′ UTR promoter, grew more slowly than rLCMV/ΔNP-GFP3′ or rLCMV/NPΔAUG. Still, all NP-deficient vectors reached comparable end titers, which were >10-fold lower than the end titer of LCMV wt.

Flow cytometric measurements indicated that the intracellular NP levels of BHK-NP cells were equivalent or higher than the levels in LCMV wt-infected BHK-21 cells (Fig. 2A). Additionally, a comparison of six individual BHK-NP transfectant clones did not reveal a clear dose relationship between cellular NP levels and rLCMV/ΔNP-GFP3′ endpoint titers (Fig. 2B). These observations suggested that insufficient NP was unlikely to be the sole explanation for the reduced growth of NP-deficient rLCMV vectors in BHK-NP cells (compare Fig. 1D). Furthermore, NP or GP transgenesis did not exert a detectable negative influence on the permissiveness of BHK-21 cells to LCMV wt replication (Fig. 2C), indicating that impaired growth of NP-deficient LCMV vectors was unlikely to be attributable to a dominant-negative effect of transgenic NP.

FIG 2.

Transgenic NP expression levels do not correlate with rLCMV/ΔNP-GFP3′ growth, and transgenic NP does not inhibit LCMV wt production. (A) Comparison of NP expression by BHK-NP cells (clone 6) and BHK-21 cells infected for 40 h with LCMV Cl.13 (MOI of 0.01) or left uninfected. The staining index was calculated as the mean fluorescence intensity (MFI) of anti-NP antibody (VL4)-stained cells divided by the MFI of isotype control antibody-stained cells. The results of one of two similar experiments are shown. (B) We infected six independent BHK-NP transfectant cell lines (clones 1, 2, 5, 6, 7, and 8) and nontransfected BHK-21 wt control cells with rLCMV/ΔNP-GFP3′ at an MOI of 0.01. After 48 h, we quantified infectious vector in the supernatant by immunofocus assay (detecting GP), using BHK-NP cells as the substrate. The results for three individual cell culture wells are shown for each cell line tested. Individual symbols are shown but may be invisible when superimposed. (C) Growth curves of LCMV wt (MOI = 0.01) in three independent BHK-NP transfectant cell lines (clones 2, 5, and 6), in GP-expressing BHK-GP cells (4), and in noncomplementing BHK-21 cells. Symbols represent the results for individual cell culture wells from one of two similar experiments.

Both rLCMV/ΔNP-GFP5′ and rLCMV/ΔNP-GFP3′ expressed GFP in complementing BHK-NP, but rLCMV/ΔNP-GFP3′ yielded more cells expressing GFP at high levels (GFP^high), putatively due to differential promoter activities of the 3′ UTR and 5′ UTR (Fig. 3) (12). rLCMV/ΔNP-GFP3′ infection of noncomplementing BHK-21 cells yielded a population of cells expressing GFP at low levels (GFP intermediate [GFP^int]), a signal which was consistently found by flow cytometry but was too faint to be reliably detected by fluorescence microscopy. Conversely, no such population was observed upon rLCMV/ΔNP-GFP5′ infection. This finding was in line with the predicted low-level GFP mRNA transcription from the 3′ UTR (rLCMV/ΔNP-GFP3′) but not from the 5′ UTR (rLCMV/ΔNP-GFP5′) of the viral S segment RNP when NP protein de novo synthesis was absent (compare Fig. 1B).

FIG 3.

GFP expression by rLCMV/ΔNP-GFP3′ and rLCMV/ΔNP-GFP5′ in complementing BHK-NP and noncomplementing BHK-21 cells. We infected BHK-NP (clone 6) and noncomplementing BHK-21 cells at an MOI of 0.5 with rLCMV/ΔNP-GFP3′ or rLCMV/ΔNP-GFP5′ vector, respectively. After 48 h, we assessed GFP expression by flow cytometry, discriminating GFP^high (upper gate), GFP^int (lower gate), and GFP-negative cells (below the GFP^int gate). Numbers indicate the percentage of cells falling into the respective gate. Gates were set such that no fluorescent cells were recorded in uninfected cells. Results from one representative experiment out of four similar experiments are shown.

To assess the impact of NP expression on LCMV vector immunogenicity, we exploited transgenic mice (ST-NP) (13) that constitutively express NP from the β-actin promoter. The H-2^b-restricted CD8⁺ T cell response to LCMV wt targets several protective epitopes in GP, NP, and L (14), and the immunodominant glycoprotein-derived epitope GP33 was monitored as a representative specificity in our experiments. Eight days after immunization with either rLCMV/ΔNP-GFP3′ or rLCMV/ΔNP-GFP5′, ST-NP mice but not C57BL/6 wild-type mice exhibited a strong GP33-specific CD8⁺ T cell response (Fig. 4A). Thus, NP was essential for immunogenicity of single-round infectious rLCMV vectors, irrespective of the positioning of the gene of interest in the vector's S segment. To also directly assess rLCMV vector-induced CD8⁺ T cell immunity, we relied on the LCMV clone 13 (Cl.13) challenge model (15). Specific CD8⁺ T cell immunity protects against chronic infection with Cl.13, which persists in the blood of unimmunized mice for several weeks after infection. We challenged all groups of animals with 2 × 10⁶ PFU of LCMV Cl.13 intravenously (i.v.). GP33-specific CD8⁺ T cells in rLCMV/ΔNP-GFP3′- and rLCMV/ΔNP-GFP5′-immunized C57BL/6 wt mice remained in the same range as in the nonimmunized control on day 4 and were only clearly detected on day 7 after challenge, thus following the kinetics of nonimmunized control groups (Fig. 4A). Furthermore, rLCMV/ΔNP-GFP3′- or rLCMV/ΔNP-GFP5′-immunized ST-NP mice cleared Cl.13 from the blood by day 11, whereas unchecked viremia persisted in identically immunized C57BL/6 mice (Fig. 4B). These observations on the kinetics of GP33-specific T cell responses upon Cl.13 challenge, as well as on Cl.13 control, corroborated the conclusion that NP was essential for rLCMV vector immunogenicity. rLCMV/ΔNP-GFP3′ and rLCMV/ΔNP-GFP5′ could have spread in trans-complementing ST-NP mice. Repeated attempts at reisolating virus from rLCMV/ΔNP-GFP3′-infected ST-NP mice have failed (data not shown), suggesting that NP trans complementation in these mice did not allow viral dissemination at LCMV wt levels, but low-level spread could have contributed to the efficacy of the vaccine. To test whether vector spread was required for protection against Cl.13, we assessed the efficacy of rLCMV/ΔGP, which cannot spread but expresses NP. Both rLCMV/ΔGP-vaccinated wt and ST-NP mice resolved viremia by day 11 after challenge.

FIG 4.

NP-deficient LCMV vectors are immunogenic and protective in NP-transgenic mice but not in wt control animals. (A) We immunized NP-transgenic (ST-NP) and syngeneic wt (C57BL/6) control mice with 1 × 10⁵ PFU of rLCMV/ΔNP-GFP3′ or rLCMV/ΔNP-GFP5′ i.v. On day 8 after immunization, we measured vaccination-induced CD8⁺ T cells in peripheral blood using MHC class I tetramers loaded with the immunodominant LCMV glycoprotein-derived epitope GP33 (Day 8 after vaccination). On day 42 after immunization, all immunized animals and naive control ST-NP mice were challenged with 2 × 10⁶ PFU of LCMV wt (clone 13 strain), which establishes chronic infection in unimmunized mice. The CD8⁺ T cell recall response in peripheral blood was determined on day 46 and day 49 (Day 4 after challenge and Day 7 after challenge, respectively). Mean results ± standard errors of the means (SEM) for three to five mice per group and time point are shown. Results from one representative experiment of two similar experiments are shown. (B) ST-NP and wt control mice were immunized with the vectors indicated in the chart. Twenty days later, they were challenged with LCMV wt, and viremia was determined by standard immunofocus assay (18). Symbols represent the mean results ± SD for four to five mice per group and time point. Results from one of two similar experiments are shown.

In summary, the following conclusions can be drawn. (i) Transgenes under 3′ UTR control can be expressed in the absence of de novo NP synthesis, albeit at very low levels. (ii) Still, in the absence of NP synthesis in vivo, LCMV vectors fail to induce detectable CD8⁺ T cell responses or protective immunity. (iii) With respect to the underlying mechanisms, the present results for single-cycle rLCMV/ΔGP vectors are in line with earlier data (4–6) and suggest that the lack of rLCMV/ΔNP immunogenicity cannot be accredited to its failure to spread in the vaccinees. The exquisite immunogenicity of rLCMV/ΔGP vectors has been linked to their capacity to target and activate dendritic cells in vivo for direct antigen presentation (4). It therefore seems likely that the antigenic mass in infected dendritic cells and/or the duration of antigen expression in these cells (4, 16) represent rate-limiting factors for T cell induction. In such a scenario, the immunogenicity of single-cycle infectious arenavirus vectors depends on robust viral mRNA transcription and genome replication and, thus, on de novo synthesis of the viral trans-acting factor NP. In return, the lack thereof seems the most likely explanation for the failure of rLCMV/ΔNP vectors to induce immunity. These conclusions have potential implications for the refinement of arenavirus-based vaccine vector technology and may add to our understanding of CD8⁺ T cell induction by inactivated virus vaccines in general (8–10).