Vaccines inducing immunity to Lassa virus glycoprotein and nucleoprotein protect macaques after a single shot

See allHide authors and affiliations

Science Translational Medicine  02 Oct 2019:
Vol. 11, Issue 512, eaaw3163
DOI: 10.1126/scitranslmed.aaw3163

A vaccine to fight Lassa fever

Thousands of people in Western Africa succumb to Lassa fever each year. Lassa virus infections can in some cases respond to antiviral treatment, but there is no vaccine. Mateo et al. tested three different viral-vectored Lassa virus vaccines in a nonhuman primate model. They compared transcriptomic and proteomic signatures as well as B cell and T cell responses after vaccination. One vaccine using a measles virus vector with Lassa virus glycoprotein and nucleoprotein provided almost sterilizing protection upon a lethal Lassa virus challenge. This vaccine will soon be tested in humans and has the potential to make a major impact on the incidence of Lassa fever.


Lassa fever is a major threat in Western Africa. The large number of people living at risk for this disease calls for the development of a vaccine against Lassa virus (LASV). We generated live-attenuated LASV vaccines based on measles virus and Mopeia virus platforms and expressing different LASV antigens, with the aim to develop a vaccine able to protect after a single shot. We compared the efficacy of these vaccines against LASV in cynomolgus monkeys. The vaccines were well tolerated and protected the animals from LASV infection and disease after a single immunization but with varying efficacy. Analysis of the immune responses showed that complete protection was associated with robust secondary T cell and antibody responses against LASV. Transcriptomic and proteomic analyses showed an early activation of innate immunity and T cell priming after immunization with the most effective vaccines, with changes detectable as early as 2 days after immunization. The most efficacious vaccine candidate, a measles vector simultaneously expressing LASV glycoprotein and nucleoprotein, has been selected for further clinical evaluation.


Lassa virus (LASV) infects tens of thousands people each year in Western Africa, causing thousands of deaths. Its main host is Mastomys natalensis, a rodent living in proximity to human housing and responsible for frequent transmission of the virus to humans (1). Human-to-human transmission then occurs through mucosal/cutaneous contacts with the body fluids of patients with Lassa fever (LF). Outbreaks in previously unaffected countries, such as Benin, have shown mortality rates of up to 40%, and Nigeria faced an unprecedented outbreak in 2018 (2). Moreover, one-third of LF survivors develop severe sequelae, such as hearing deficits (3). There is neither treatment with demonstrated efficiency nor a licensed vaccine to fight against this deadly emerging virus. Thus, LF is a major public health issue and has been classified as an epidemic threat requiring urgent preparedness by the World Health Organization.

Vaccine studies in nonhuman primates (NHPs) have suggested that an LF vaccine should be live-attenuated and induce robust T cell immunity not only against the glycoprotein (GPC) but also the nucleoprotein (NP), if possible. The presence of LASV GPC is required for protection using a vaccinia vector and adding NP improved vaccine efficacy (4). The LASV Z protein could also be of interest for a vaccine because it is able to induce the release of virus-like particles (VLPs) that could increase the immunogenicity (5). The endemicity of LASV and its extensive area of circulation, as well as the frequent transmission of the virus from wildlife to humans, argue in favor of mass vaccination for the estimated 180 million people living at risk of the disease (6). An efficient vaccine for a ring immunization strategy during outbreaks would also be valuable. Despite advances in LF vaccine research over the past few years, development of a vaccine has been hampered by insufficient safety, immunogenicity, and efficacy data (7). The most advanced candidates include vesicular stomatitis virus (VSV)–LASV and ML29, a live reassortant of LASV and Mopeia virus (MOPV), an arenavirus closely related to LASV but nonpathogenic for NHPs. Both have shown promising results in NHPs (8, 9). Previous vaccine trials in NHPs suggest that although generating LASV GPC–specific T cell responses is required to protect animals (1013), presenting that NP is also crucial for protecting against divergent LASV strains (10). The high genetic diversity of LASV strains is a major issue for vaccination, and the use of two antigens is a means to address it. We recently developed a vaccine platform called MOPEVAC, based on a recombinant MOPV that is further attenuated and can serve as a vector for all arenaviruses by swapping the MOPV GPC with those of both Old World and New World arenaviruses (14). The LASV GPC–expressing MOPEVAC (MOPEVACLAS) expresses MOPV NP, which is very close to LASV NP, with 74% amino acid homology. Moreover, MOPV can induce efficient cross-protection against LASV challenge in NHP (15). Here, we present an approach using the Schwarz measles (MeV) vaccine platform (16). The MeV vaccine presents that all the required features for an efficient LASV vaccine and MeV vaccine strains that express antigens from a wide range of pathogens have been developed and shown safety, stability, immunogenicity, and efficacy in several animal models (17, 18). In addition, clinical trials have demonstrated that preexisting immunity to MeV does not affect the immunogenicity of a MeV–chikungunya (CHIK) vaccine against CHIK virus in humans (19, 20). MeV-based vectors are attractive because they can be produced at low cost, administered by various routes, and are efficient inducers of T cell responses and long-term memory. We generated MeV expressing both GPC and Z protein to allow the release of GPC-expressing VLP because this approach has been successful for the MeV-CHICK vaccine (21). We also generated vaccines containing both GPC and NP because these proteins have been shown to induce potent immunity against LASV, particularly divergent strains (10). Here, we tested MeV-based vaccines and MOPEVACLAS for their immunogenicity and efficacy after a single dose in cynomolgus monkeys, a relevant animal model for LF, with the aim to select the best candidate to bring to clinical trials.


Development and characterization of MeV-based vaccines against LASV

We generated MeV expressing GPC alone or in combination with NP or Z protein (Fig. 1A). All vectors presented similar growth kinetics in VeroE6 cells (Fig. 1B), expressed the antigens (Fig. 1C), and were genetically stable over 10 passages (table S1). At a multiplicity of infection (MOI) of >0.1, Z-expressing MeV vaccines grew to lower titers than the other vaccines (fig. S1), likely due to Z-induced cytotoxicity (22). At an MOI of <0.01, MeV-Z replicated to similar titers as the other constructs. The large amount of GPC detected in the supernatant of the MeV-Z + GPCLASV construct relative to the others suggests efficient release of VLPs (Fig. 1C). Z protein is sufficient to induce VLP budding (5).

Fig. 1 Generation and characterization of MeV-based vaccine candidates for Lassa virus.

(A) Scheme of the MeV-based vaccines generated using the Schwarz strain. ATU, additional transcription unit. (B) Growth kinetics of vaccines in Vero NK cells. The results are presented as the means ± SEM of three independent experiments. (C) LASV NP, GPC, Z protein, and MeV F antigen expression in supernatants and cell extracts (cells) of infected Vero NK cells analyzed by Western blot. Actin was used to normalize the amount of loaded proteins. (D) Induction of type I IFN mRNA expression in primary human monocyte-derived dendritic cells (left graph) and macrophages (right graph) 1 day after infection by the MeV-based constructs. (E) Cell surface expression of coactivation molecules in primary human monocyte-derived dendritic cells (left graph) and macrophages (right graph) at day 2 after infection. Results (D and E) are presented as the means ± SEM of three or more independent experiments using different donors. (F) Infection of CHO-K1 or CHO-hCD46 cells by MOPEVACLAS or MeV-NP using an MOI of 1. LASV GPC (green) was detected by immunofluorescence 72 hours after infection. Nuclei are stained with DAPI (blue).

All vectors induced type I interferon (IFN) synthesis and expression of activation molecules in human primary antigen-presenting cells (APCs) (macrophages and dendritic cells), except vectors expressing the wild-type LASV NP (Fig. 1, D and E). Arenavirus NP contains an exonuclease domain that inhibits type I IFN induction (23). We thus introduced two mutations in the NP exonuclease domain to abrogate its IFN antagonist activity (2325) and obtained a vector (MeV-NPEXON + GPCLASV) that induced the same quantities of type I IFN as the other vectors. We also confirmed that the insertion of LASV GPC into the MeV backbone did not change its tropism, because this vector was only able to enter Chinese hamster ovary (CHO)–K1 cells expressing CD46, a known receptor of the MeV vaccine strain, in contrast to MOPEVACLAS, which entered both CHO-K1 and CHO-CD46 cell lines efficiently through α-dystroglycan (Fig. 1F). On the basis of these results, we selected MeV-NPEXON + GPCLASV (MeV-NP) and MeV-Z + GPCLASV (MeV-Z) for further evaluation.

Single-shot immunization of NHPs

We evaluated the efficacy of MeV-NP, MeV-Z, and MOPEVACLAS in cynomolgus monkeys, using a single-shot strategy. Three groups of four macaques were immunized using the subcutaneous route with 2 × 106 tissue culture infective doses (TCID50) of MeV-NP or MeV-Z, or intramuscularly with 6 × 106 focus-forming units (FFU) of MOPEVACLAS (Fig. 2A). Three control macaques were immunized with 2 × 106 TCID50 of MeV Schwarz strain. We then followed the biological parameters of the animals (body temperature, weight, and respiratory rate) and noted no adverse effects. The body temperature of animals, continuously monitored by intraperitoneal devices, was unaltered, except for a transient increase a few hours after MOPEVACLAS injection (fig. S2). We also assessed viremia after immunization and did not detect any viral RNA by reverse transcriptase quantitative polymerase chain reaction (RT-qPCR), either in plasma or peripheral blood mononuclear cells (PBMCs) (table S2). Similarly, we did not detect viral RNA in the nasal or oral secretions or in urine of vaccinated animals (table S2). Thus, the vaccines appeared safe in monkeys, and no virus shedding was observed after immunization.

Fig. 2 Immunogenicity of vaccine candidates in cynomolgus monkeys.

(A) Scheme presenting the time of immunization, sampling, challenge, and necropsy for each group. (B) Detection of MeV-specific IgM and IgG in plasma according to the time after immunization. MeV control animals were only analyzed at days 0 and 23. Each bar represents the means ± SEM of four samples, except for MeV controls (n = 3). (C) Quantification of CD8+ (left graphs) and CD4+ (right graphs) T cells specific for LASV GPC (top graphs), NP (middle graphs), and Z (bottom graphs) peptides. The percentage of T cells producing TNFα and/or IFNγ in response to peptide stimulation among total CD8+ or CD4+ T cells is presented according to the time after immunization after subtraction of the respective value measured in unstimulated T cells. Each bar represents the means ± SEM of four samples. Statistical analysis was performed using a Kruskal-Wallis one-way ANOVA on ranks, but differences were not significant. The proportion of each subpopulation of cytokine-positive T cells is presented in the pie charts. (D) Changes in the expression of 51 genes linked to the response to type I IFN in monkeys according to the time after immunization. Each time point represents the mean of four individual samples. The color indicates the average standardized (centered and scaled) gene expression, and the intensity of standardized expression is indicated in the scale. The standard average expression (Standard aver. expr.) of the genes was plotted according to the time after immunization. Within each vaccine, a one-way ANOVA adjusted on the day effect was fitted on the standardized (centered and scaled) genes expression. Comparisons of average gene expression between day 0 and other days were performed using contrasts of this linear model (Student statistics). The red asterisks indicate significant differences from the expression at day 0 (P < 0.05).

We assessed the induction of LASV-specific immunoglobulin M (IgM) and IgG after immunization. There was no detectable IgM response, but we observed low titers of IgG in all MOPEVACLAS-immunized animals and in two of four MeV-NP animals 37 days after immunization (table S3). We also detected low titers of neutralizing antibodies (nAbs) in three of four MOPEVACLAS-immunized animals and in one of four animals in the other groups at that time point. All MeV-vaccinated monkeys developed similar and robust IgM and IgG MeV-specific responses from 2 weeks after immunization (Fig. 2B). We monitored the induction of LASV-specific CD8+ and CD4+ T cells by quantifying T cells producing IFNγ and/or tumor necrosis factor–α (TNFα) in response to overlapping peptides covering the entire LASV GPC, NP, and Z protein (Fig. 2C). There was an increase in the percentage of GPC-specific cytokine-producing CD8+ and CD4+ T cells 23 days after immunization with MeV-NP and MOPEVACLAS but not MeV-Z (Fig. 2C, top). NP-specific cytokine-producing T cells appeared after 14 days in MeV-NP– and MOPEVACLAS-immunized animals and were still present after 23 days (Fig. 2C, middle). TNFα was the main cytokine detected in the response. Last, only weak T cell responses were observed in response to Z peptides in MeV-Z– and MOPEVACLAS-immunized animals (Fig. 2C, bottom).

We characterized the transcriptomic signature associated with the vaccines by RNA sequencing of PBMC samples collected after immunization (Fig. 2D). We performed pairwise comparisons to identify differentially expressed (DE) genes between samples corresponding to the day just before immunization (day 0) and the other time points. Overall, immunization with the three vaccines led to marked changes in the transcriptomic profiles, as early as 2 days after vaccination with MeV-NP. Changes were evident from day 4 for the MeV-Z animals. We observed marked transcriptomic changes 2 days after immunization in the MOPEVACLAS group; changes were the highest 2 weeks after vaccination (fig. S3). We investigated the expression of 51 genes reported to be expressed in response to type I IFN (26), of which 49 were significantly DE for at least one of the pairwise comparisons (P < 0.05) (Fig. 2D). In MeV-NP–immunized animals, the average expression of type I IFN–related genes was higher at day 2 than at day 0, and this trend persisted up to day 7. In MeV-Z monkeys, the average overexpression of these genes was induced between 4 and 7 days after immunization. Overexpression of these genes was even greater at day 2 in MOPEVACLAS monkeys than in the MeV-NP animals but did not persist over time. Many DE genes in the MOPEVACLAS animals were composed of IFN-stimulated genes (those of the IFIT, OAS, IFI, MX, TRIM, PARP, HERC, and DDX families and ISG15, PLSCR1, CMPK2, and RSAD2). Some genes involved in IFN-signaling pathways (IRF7, STAT1, and TLR7), chemoattraction (CXCL10), and ubiquitination (UBE2L6 and FBXO6) were instead induced 7 days after immunization in both the MeV-NP and MeV-Z monkeys. These transcriptomic changes had major consequences on the activation of cellular pathways (figs. S4 to S6 and Fig. 3). Pathways involved in immunity, such as type I and II IFN responses, the inflammatory response, antigen presentation response, and T cell responses, were activated more highly and earlier in PBMC from the MeV-NP animals than the others. A larger number of pathways were activated in the MOPEVACLAS than MeV-Z monkeys. In contrast, the pathways linked to TNFα signaling via nuclear factor κB and E2F targets were most highly up-regulated in the MeV-Z animals and more highly in the MOPEVACLAS than MeV-NP animals.

Fig. 3 Transcriptomic changes after immunization.

Network representation of genes differentially expressed (P < 0.05) using the data obtained from RNA-seq analysis (performed using samples from all immunized animals for each time point) grouped by Hallmark gene sets for comparisons among vaccines for days 2 to 14 (top graphs) and comparisons between days 2 and 0 for each vaccine (bottom graphs). Each node represents one Hallmark gene set with a size proportional to the number of genes belonging to the set. Edges represent shared genes among the gene sets (nodes). The thickness of the edges is proportional to the number of shared genes between two connected nodes. Nodes are colored according to the overexpression and underexpression of the gene sets. Nodes represent gene sets with the circle size proportional to the number of genes belonging to the sets.

We also evaluated the proteomic profiles after immunization. All animals showed substantial changes in individual protein abundance as early as 2 days after immunization (fig. S7A). However, these changes resulted in only a few pathway modifications, such as a significant drop in the activation of integrin pathways observed the second week after immunization in MeV-Z animals and some changes in pathways involved in complement activation and innate immunity in the MeV-NP and MOPEVACLAS animals (P < 0.05) (fig. S7B). Each vaccine induced a profile that was distinguishable from the others, as shown by principal components analysis, with the MeV-Z group showing more divergence than the MeV-NP and MOPEVACLAS groups (fig. S8, top). The profiles of each group were also different depending on the time after immunization, as soon as after 2 days (fig. S8, bottom). Together with the transcriptomic study, these results show that specific signatures characterize each vaccine candidate and may correlate with their efficacy.

A single shot of MeV-NP, MeV-Z, and MOPEVACLAS protects monkeys against LASV

All monkeys were challenged 37 days after immunization with a subcutaneous injection of 1,500 FFU of LASV, Josiah strain, the same strain used to generate the vaccines. The animals were then monitored for up to 30 days and attributed clinical scores based on their body temperature, weight, ability to feed and hydrate normally, behavior, and clinical signs, with a score of 15 being the end point for euthanasia. The three control animals had continuously increasing scores from day 3 and had to be euthanized within 2 weeks after challenge (Fig. 4A). All three animals presented severe symptoms, such as high fever (Fig. 4B), asthenia, anorexia, prostration, and weight loss. Animal MeV-1 was euthanized with a score of 12 because it presented severe clinical signs and was hypothermic, a sign of terminal phase of LF. All vaccinated animals survived the LASV infection, but the clinical outcomes were different depending on the vaccine. MeV-NP animals showed a slight increase in their clinical score by day 5 due to a transient increase in body temperature. In contrast, MeV-Z animals experienced severe symptoms, such as a high fever between days 3 and 12. Although animals Z-2 and Z-3 completely recovered by day 12, animals Z-1 and Z-4 had difficulties feeding and drinking and showed prostration. Animal Z-4 had difficulties balancing and lost more than 7.5% of its body weight by day 30, reaching a score of 14. In the MOPEVACLAS group, three animals showed a slight increase in their score between days 6 and 12, also mainly due to elevated body temperature, whereas animal MOP-2 did not show any clinical sign or fever.

Fig. 4 Clinical monitoring of immunized animals after a LASV challenge.

(A) Clinical score of individual animals after the challenge. (B) Monitoring of body temperature during the course of infection. Individual data are presented for each animal, except for one MeV animal for which the data were missing (failure of the device). (C) Analysis of biological parameters during LASV infection. Individual data of ALT, AST, LDH, and CRP concentrations are presented for each animal according to the time after challenge.

The blood counts showed a modest and transient drop in lymphocyte numbers for all animals 3 days after challenge (fig. S9). Lymphocytosis was then observed until 15 days after challenge for all monkeys, except the MeV control animals. The number of monocytes and neutrophils increased continuously in animals until the day of euthanasia for controls or day 15 for the other animals and then returned to baseline values, except for MeV-NP animals, in which the number remained elevated.

In all MeV control monkeys, the concentrations of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) continuously increased from day 6 (Fig. 4C). The quantities of liver enzymes remained low in all immunized animals. High and increasing concentrations of lactate dehydrogenase (LDH) were present in the plasma of MeV controls from day 6. Elevated concentrations of LDH were also found in two MeV-Z animals 9 to 12 days after infection. In contrast, LDH remained unchanged in MeV-NP and MOPEVACLAS animals. The amounts of C-reactive protein (CRP) peaked at the time of euthanasia in control animals. In the MeV-NP group, CRP values remained low, except for a transient increase at day 6 for one animal. We observed a similar increase in animals MOP-1 and MOP-4. In the MeV-Z group, CRP concentrations peaked at day 9 for all animals but remained elevated for animal Z-4. The concentrations of albumin (ALB) also dropped in control animals before euthanasia, together with elevated urea and uric acid (UA) concentrations (fig. S10). There was also a transient decrease in ALB values in the MeV-Z, and, to a lesser extent, MOPEVACLAS, animals. The other analyzed parameters did not change.

Control of LASV replication in immunized animals

The infectious titers in the plasma of MeV controls increased continuously from day 3 (Fig. 5A). We detected no infectious virus in MeV-NP animals, despite a low amount of viral RNA detected in plasma at day 6 (Fig. 5B), suggesting almost sterilizing immunity in these animals. In the MeV-Z group, we detected moderate LASV titers in all animals at day 6 but up to day 15 for animal Z-4 (Fig. 5A). The viremia in animal Z-4 dropped between days 9 and 12 but rebounded by day 15. We also detected very low transient infectious titers in most MOPEVACLAS animals, except animal MOP-2, for which neither infectious particles nor viral RNA was detected in plasma (Fig. 5, A and B). We detected infectious LASV and viral RNA in the oral and nasal swabs of all control animals, but not in the secretions of MeV-NP or MOPEVACLAS animals, despite the detection of low amounts of viral RNA in the secretions of some animals (Fig. 5A and fig. S11A). In contrast, we detected infectious particles from days 6 to 12 in the oral and nasal secretions of the MeV-Z animals. No infectious virus was detected in the urine of animals, except Z-1 (250 FFU/ml), and we only detected viral RNA after day 9 in controls and after day 12 in the MOP-4 and most MeV-Z animals (fig. S11A). Viral RNA and infectious particles were present in all tested organs collected at the time of necropsy of the MeV controls (Fig. 5C and fig. S11B), whereas we found viral RNA in the lymphoid organs of at least one animal in each group of immunized monkeys, with animals in the MeV-Z group showing the highest quantities. We found infectious virus in the inguinal lymph nodes of two of the MeV-Z animals and in the spleen of one of them.

Fig. 5 LASV replication in cynomolgus monkeys.

(A) Quantification (in FFU per milliliter) of LASV infectious particles in plasma (top), oral swabs (middle), or nasal swabs (bottom) according to the time after challenge. Individual data are presented for each animal. (B) Quantification of viral load by RT-qPCR in the plasma of animals. Individual values of each animal are presented as the number of viral RNA copies per milliliter according to the time after challenge. (C) Quantification in (FFU per milligram of tissue) of viral load in various organs, including inguinal lymph nodes (ILN), mesenteric lymph nodes (MLN), and cerebellum (Cereb.). Individual values are only presented for the three MeV controls and two MeV-Z animals because no infectious particles were detected in the organs of the other animals.

Immune responses induced after LASV challenge

In MeV control animals, the titers of LASV-specific IgG remained low until euthanasia, and no nAb was detected (Fig. 6, A and B). We detected a substantial increase in LASV-specific IgG titers 9 days after infection in all immunized animals, with lower titers in the MeV-Z group than the others until day 15 (Fig. 6B). LASV NP–specific IgG was detected as soon as 12 days after challenge in all MeV-NP and three of four MOPEVACLAS animals, whereas IgG was mainly directed against glycoprotein 2 (GP2) and Z protein in the MeV-Z group at the same time (Fig. 6A). At necropsy, anti-LASV GP2 and NP IgG were present in all samples, but the amounts of NP IgG were lower in three of four MeV-Z animals. In these animals, the IgG also weakly recognized glycoprotein 1 (GP1) and Z protein (Fig. 6A). NAb titers increased in vaccinated animals between 6 and 15 days after infection. All MeV-Z animals showed high nAb titers, whereas only half of those of the MeV-NP and MOPEVACLAS groups showed such titers. Most animals showed a drop in nAb titers 28 days after infection.

Fig. 6 Analysis of humoral and cellular immune responses after LASV infection.

(A) Western blot analysis of humoral responses after LASV infection. Plasma samples obtained 12 and 28 days after LASV challenge were analyzed against LASV antigens. Polyclonal and monoclonal antibodies specific for GP1, GP2, NP, or Z protein were used as positive controls. (B) Detection of LASV-specific IgG in animals after LASV infection by ELISA (top graph). Data are presented as the mean titers ± SEM of four samples, except for the MeV group (n = 3). Quantification of nAb in the plasma of animals after LASV infection. Individual values of the animals represent the highest dilution of plasma positive in a 50% neutralizing assay. (C) Quantification of soluble mediators in the plasma according to the time after LASV infection. Data represent the means ± SEM of four samples, except for the MeV group (n = 3). Statistical differences between one or several groups and others have been determined with a Kruskal-Wallis one-way ANOVA on ranks test and are specified above the respective time point (c, MeV; n, MeV-NP; z, MeV-Z; m, MOPEVACLAS) using the following code: n-cz means that the MeV-NP group was significantly different (P < 0.05) from the MeV and MeV-Z groups. (D) The percentage of CD8+ (top) and CD4+ (bottom) T cells expressing KI67, granzyme B (GrzB), perforin (perfo), or HLA-DR (DR) is presented according to the time after LASV infection. Each bar represents the means ± SEM of four samples, except for the MeV group (n = 3). Significant differences between groups have been calculated using a one-way ANOVA test and are indicated by *P < 0.05 or **P < 0.01.

We detected a transient release of IFNγ in the plasma of all animals, peaking at 6 and 9 days after infection (Fig. 6C). However, lower IFNγ quantities were found in the MeV-NP animals than in the others. The concentrations of perforin rose until day 9 or 12 for all animals and then returned to low values by day 22. Once again, the quantities observed in MeV-NP animals were lower. We detected elevated concentrations of soluble CD137 from 9 days after infection in the MeV control and MeV-Z animals, whereas we detected only moderate concentrations in the MeV-NP and MOPEVACLAS monkeys. Interleukin-6 (IL-6) appeared in the plasma of all animals by day 6 and was still present 3 days later except for the MeV-NP group. IL-6 concentrations were still rising at the terminal stages in the MeV controls. We also detected elevated quantities of IL-8 from day 6 in the MeV control and MeV-Z animals, whereas only low concentrations were detected at days 9 and 12 in the MeV-NP and MOPEVACLAS monkeys. We detected low quantities of IL-18 during the course of the infection in the MeV-Z animals, whereas IL-18 concentrations rose until death in the MeV controls. Monocyte chemoattractant protein 1 (MCP1) remained at basal values in the MeV-NP animals and only moderately and transiently increased in the MOPEVACLAS group. In contrast, we detected high concentrations from 6 to 9 days after infection in the MeV control and MeV-Z animals, respectively. Concentrations of IL-10 and the IL-1 receptor antagonist (IL-1RA) increased until day 9 in the MOPEVACLAS and MeV-Z animals but was still rising at the terminal stages in MeV controls. Except for small amounts detected 6 days after infection, IL-10 and IL-1RA were not released into the plasma of MeV-NP monkeys. Last, IFNα was only detected in MeV animals 6 days after infection.

The percentage of proliferating Ki67+ CD8+ T cells increased until euthanasia in the MeV monkeys and until 15 days after infection in the other animals (Fig. 6D). We observed a greater percentage of Ki67+ CD4+ T cells in the MOPEVACLAS monkeys at the day of challenge than in the other groups and observed robust proliferation of CD8+ T cells 9 days after challenge among the MeV-NP monkeys, which was less pronounced in the other animals. Last, we detected a higher number of Ki67+ CD8+ T cells in the MeV-Z monkeys and, to a lesser extent, MOPEVACLAS animals 22 days after challenge. In addition, we observed an increase in the percentage of proliferating CD4+ T cells in the vaccinated animals by day 15 and on the day of the challenge and 9 days later for the MOPEVACLAS and MeV-NP monkeys.

The percentage of CD8+ T cells expressing granzyme B (GrzB) increased after the challenge in all vaccinated monkeys, but not in the MeV control animals. The MeV-NP, MOPEVACLAS, and MeV-Z monkeys showed a peak of expression 9, 12, and 15 days after infection, respectively. GrzB expression was significantly higher 9 and 12 days after infection in the MeV-NP monkeys and 12 days after infection in the MOPEVACLAS animals than in the MeV controls (P < 0.01). CD4+ T cells displayed similar kinetics for GrzB expression, except for the MeV-Z monkeys, which only showed a significant increase at 15 days after infection (P < 0.05). Perforin-expressing CD8+ and CD4+ T cells in the immunized monkeys peaked 22 days after infection. Last, we analyzed the expression of human leucocyte antigen (HLA)-DR, which is expressed by activated T cells. We observed a higher percentage of HLA-DR+ CD8+ T cells 9 to 12 days after challenge in the MeV-NP animals than the others, whereas HLA-DR+ CD4+ T cells were detected only after 15 to 22 days in all vaccinated animals.

Analysis of the T cell responses

We assessed CD45RA, CD27, and CD28 surface expression to characterize the various T cell effector/memory phenotypes (Fig. 7, A and B, and fig. S12). However, we were unable to obtain reliable CCR7 staining and were consequently unable to discriminate between some subpopulations, such as naïve/pre-effector 1 (pE1) (CD45RA+CD28+CD27+) and central memory/effector memory 1 (EM1) (CD45RACD28+CD27+) T cells. An enrichment in naïve/pE1 CD8+ (Fig. 7A and fig. S12A) and CD4+ (Fig. 7B and fig. S12B) T cells was observed 3 to 9 days after challenge in MeV-NP and MeV-Z animals and 9 days after in MeV control and MOPEVACLAS animals. An increase in the proportion of terminally differentiated effector memory (EMRA) CD8+ T cells (CD45RA+CD28CD27) was observed between 3 and 6 days after challenge. Then, we observed in immunized animals an increase in the percentage of intermediate EM2 CD8+ T cells (CD45RACD28CD27+) from 9 to 15 days after challenge and of EM3 CD8+ T cells (CD45RACD28CD27) between days 9 and 22. The proportion of pE2 CD4+ T cells (CD45RA+CD28CD27+) increased 6 to 15 days after challenge in all animals. A similar enrichment was observed for EM2 CD4+ T cells 12 and 15 days after challenge but not in MeV-Z animals. Then, EM3 CD4+ T cells (CD45RACD28CD27) began to rise 22 days after challenge in immunized animals. Last, a transitory increase in the proportion of EMRA CD8+ and CD4+ T cells was observed 22 days after challenge.

Fig. 7 T cell responses after LASV infection.

The proportion of memory CD8+ (A) and CD4+ (B) T cell subpopulations in the blood according to CD45RA, CD28, and CD27 expression is presented according to the time after infection. The subpopulations are naïve, pre-effector 1/2 (pE1/2), central memory (CM), effector memory 1/2/3/4 (EM1/2/3/4), and terminally differentiated effector memory (EMRA) T cells. Data represent the means ± SEM of four samples, except for the MeV group (n = 3). Significant differences between groups have been calculated using a one-way ANOVA test and are indicated as for Fig. 4. Quantification of CD4+ (left graphs) and CD8+ (right graphs) T cells specific for LASV GPC (C) and NP (D) peptides. The percentage of T cells producing TNFα and/or IFNγ in response to peptide stimulation among total CD8+ or CD4+ T cells is presented (bar graphs) according to the time after immunization after subtraction of the respective value measured in unstimulated T cells. Each bar represents the means ± SEM of four samples, except for the MeV group (n = 3). Statistical differences between the different groups have been calculated using a Kruskal-Wallis one-way ANOVA on ranks test and are indicated by *P < 0.05. The proportion of each subpopulation of cytokine-positive T cells among cytokine-producing T cells is presented in the pie charts. PBMCs from MeV-Z animals were not stimulated with NP peptides.

Cytokine production was detected in a very low percentage of T cells after stimulation of PBMC with LASV Z peptides after the challenge (fig. S13) and in the MeV controls in response to GPC and NP peptides (Fig. 7, C and D). The percentage of CD8+ and CD4+ T cells from MeV-NP PBMCs producing cytokines in response to GPC peptides rose to 2 and 0.6% at 9 days after infection, respectively, and then returned to basal values by day 22 (Fig. 7C). Most of the T cells only produced IFNγ, but the proportion of polyfunctional T cells producing IFNγ and TNFα increased from days 9 to 22. In MeV-Z animals, a moderate number of cytokine-producing CD8+ and CD4+ T cells were only detectable from days 12 to 9 after infection, respectively. Most T cells produced only IFNγ and the frequency of polyfunctional T cells remained at about 20% for CD8+ T cells and rose to 40% for CD4+ T cells. T cells from the MOPEVACLAS animals responded to GPC peptides from 9 days after infection, but the proportion of CD8+ T cells was moderate relative to that of MeV-NP animals. IFNγ-producing cells mainly accounted for the CD8+ response, and the percentage of polyfunctional T cells was between 31 and 44. The percentage of IFNγ-producing CD4+ T cells decreased until day 22, whereas the opposite occurred for polyfunctional T cells, which rose to 59%. A notable proportion of T cells of all immunized monkeys produced only TNFα at 22 days after infection.

Trace amounts of cytokine-producing T cells from the MeV controls were only detected 12 days after infection in response to LASV NP peptides, whereas responding T cells were detected in the MeV-NP animals as soon as 3 days after infection, with a peak response at day 9 (Fig. 7D). The phenotype of T cells varied at day 3 after infection and then mainly produced only TNFα by day 6. At day 9, IFNγ-producing cells were dominant, whereas the proportion of polyfunctional T cells increased until day 22. For the MOPEVACLAS animals, TNFα+ T cells were detected at day 6 after infection, and the percentage of responding T cells increased to day 9 but were of various phenotypes. Then, TNFα+ T cells, IFNγ+ T cells, and polyfunctional T cells became dominant by day 22.

Immune correlate of the protection against LF

We identified the genes (Fig. 8A) and proteins (Fig. 8B) for which the changes in expression after immunization were the most predictive of the clinical score by Lasso analysis. Some changes inversely correlated with the clinical scores and were thus associated with protective immunity, whereas others correlated with the clinical scores and were associated with the absence of efficacy. Among these genes and proteins correlated with protection or on the contrary with low vaccine efficacy, we identified transcription factors (ZNF, POU6F1, and MYBBP1A) and enzymes (SRGAP1, PAK4, NIT2, MINDY4, AK6, hepatocyte growth factor (HGF) activator, inter–α-trypsin inhibitor, and FETUIN B). Genes and proteins with roles in inflammation/immunity (FCN2/Ficolin-2–like, CHST1, CADM1, SERPIN, Igκ constant, CD109, CRP, and bikunin precursors), complement pathways (properdin and C1q C chain), and coagulation (SERPIN, hyaluronan binding protein, and thrombin) were also identified.

Fig. 8 Immune correlates of protection.

Bar plots of Pearson correlations between the clinical score and gene expression (using data obtained from RNA-seq analysis) (A) or protein abundance (using data obtained from proteomics MS/MS analysis) (B) selected by the Lasso method. RNA-seq and proteomic analyses were performed using samples from all immunized animals. This variable selection method, based on penalized linear regression, identifies the genes/proteins, which are the most related to the clinical score (see Supplementary Materials and Methods). The x axis indicates the value of the correlation between each gene/protein selected by the Lasso and the clinical score. The color of the bars indicates the sign of the correlation (negative in blue and positive in red).


Here, we compared different platforms for a LASV vaccine. Because arenavirus NPs contain an exonuclease domain with a type I IFN antagonist activity, we removed the exonuclease function in both MOPEVACLAS and MeV-NP to avoid production of an immunosuppressive protein by the vaccines. We demonstrated by comparing the responses induced by MeV producing wild-type or mutated NP in APCs that this function is detrimental to immunogenicity. The three candidates were stable, and neither adverse effects nor viremia or shedding was observed after immunization. MOPEVACLAS, due to removal of the NP exonuclease activity (14), is a hyperattenuated form of MOPV, which is nonpathogenic for NHPs. The lack of altered tropism of MeV-based vaccines argues in favor of the same safety profile as that described for similar constructs (19).

The MeV-NP candidate provided the most efficient protection against LASV. All four monkeys remained healthy, and the acquired protection was almost sterilizing, with only traces of viral RNA detected. The concentrations of cytokines associated with severe LF did not or only modestly increased after the challenge of MeV-NP–immunized animals. MOPEVACLAS was also highly potent, allowing three of four animals to control LASV with no clinical signs, except for a transient fever and slight alterations of biological parameters. We observed a low and fleeting infectious viremia, as well as a transient viral RNA load in swabs and organs. The fourth animal experienced complete sterilizing immunity, with neither viral RNA detection nor elevation of body temperature. Clinical signs were not observed in any animals from these groups, demonstrating the efficacy of both vaccines. In contrast, despite survival, the MeV-Z animals were seriously ill and viremic. One presented lasting infectious viremia and viral RNA in organs 30 days after challenge, including in the brain, cerebellum, testis, and bladder. Viral RNA was also detected in the urine, concomitant with dizziness and weight loss. We detected LASV in the urine and aqueous humor, together with an elevated clinical score in another MeV-Z animal 1 month after challenge. This suggests incomplete control of replication and viral persistence. It confirms that a LASV vaccine needs to provide more than simply survival and allow the complete control of viral replication to avoid the risk of long-term viral persistence and potential further transmission through fluids from immunologically privileged organs. All immunized monkeys showed transient lymphopenia, followed by lymphocytosis. Lymphopenia during viral infections is due to the secretion of type I IFNs and, to a lesser extent, TNFα (27). Type I IFNs induce marked changes in the distribution of T cells and B cells and their retention within lymphoid organs. Type I IFNs are also crucial for subsequent pathogen-specific T cell proliferation, which is consistent with our observations (28). We did not detect IFNα in plasma, but the release probably occurred early after infection, before vanishing by day 3.

The humoral response elicited after immunization was low. No LASV-specific IgG was detected in the MeV-Z group, and only two of four MeV-NP animals were positive at low titers. MOPEVACLAS was more efficient because plasma from all animals contained detectable LASV IgG on the day of challenge and the LASV IgG was neutralizing for three of them. In contrast, only one of four MeV-NP and MeV-Z monkeys presented such a neutralizing titer. However, the finding that nAb- and LASV-specific IgG titers rose substantially after challenge in immunized animals, but not MeV monkeys, is suggestive of a potent secondary humoral response. The robust humoral responses induced against MeV suggest an efficient vaccine take and the induction of protective immunity against MeV. MeV-IgG titers correlate with protective immunity against measles (29). This indicates that the MeV-based candidates are bivalent vaccines, either inducing primary immunity to MeV in nonvaccinated individuals or boosting an anamnestic response in others. Preexisting MeV immunity does not impede immunity against heterologous antigens expressed by the MeV platform (19, 20). Although they probably play a role in vaccine efficacy, nAbs alone are not sufficient to protect animals against LF. MOPEVACLAS was not the most efficient despite the presence of LASV nAbs in most animals on the day of challenge. Similarly, MeV-Z animals were the most severely affected despite their high nAb titers after the challenge. Various patterns of nonneutralizing LASV IgG responses were observed. Nonneutralizing IgG appeared earlier and at higher titers in the MeV-NP and MOPEVACLAS than MeV-Z and MeV animals, suggesting that they played an important role in protection. Consistent with this observation, the kinetics of the appearance of high titers of antibodies, from day 6 (MeV-NP and MOPEVACLAS) to day 15 (MeV-Z) matched those of the respective control of viremia. IgG specificity also varied, with protection associated with early production of LASV NP–specific IgG, whereas IgG from the MeV-Z animals was rather directed against GP2 and Z. Unexpectedly, IgG against LASV GP1 was detected only in MeV-Z animals, and at low titers and late in the course of the disease. Early recognition of NP and Z protein reflects the antigenic content of the vaccines; in this study, we did not determine whether the humoral response against NP is important for the control of LASV infection. Overall, these data are consistent with the previous demonstration that nonneutralizing LASV-specific IgG is associated with protection in guinea pigs immunized with bivalent rabies–LASV vaccine (30). However, human monoclonal antibodies able to neutralize four LASV clades have been shown to fully protect macaques against LF, suggesting that high titers of potent nAb could be sufficient to control LASV (31).

Our vaccine candidates allowed evaluation of the immunogenicity of several antigen combinations with GPC as a common denominator. The presence of Z protein was not protective, despite the probable release of GPC-expressing VLPs. These structures were immunogenic in the MeV-CHIK vaccine, and this approach was successfully used for another LASV vaccine expressing the Z protein (19, 22). The lack of GPC-specific T cells after immunization with MeV-Z is expected because high titers of GP2-specific IgG were induced. This low T cell immunogenicity could result from the attenuation of MeV-Z relative to the other two candidates and/or different immunogenicity favoring humoral rather than cellular responses due to pseudoparticle release. The attenuation of LASV Z at high MOI may be due to cytotoxicity (22), because the similar growth kinetics of MeV-Z and MeV-NP at low MOI of 0.01 suggest that the addition of Z protein at the 3′ extremity of the genome did not alter its replication. It is thus possible that replication after administration of the high inoculum was lower, limiting the immunogenicity of the MeV-Z candidate. However, we cannot draw conclusions concerning the negative effect of LASV Z, because we did not evaluate the same construct without the Z protein. The presence of LASV or MOPV NP in the vaccines was associated with potent and protective immunity. LASV NP–specific T cells were present before and after the challenge in both MeV-NP and MOPEVACLAS monkeys. A low number of NP-specific CD8+ T cells were present in the blood of MeV-NP animals as soon as 3 days after challenge, and specific CD4+ T cells were also detected earlier in these animals than in the others. The lower protective efficiency of MOPEVACLAS relative to that of MeV-NP could be due to the presence of a heterologous NP rather than lower immunogenicity of the MOPEVAC platform, because it was the most immunogenic candidate before the challenge. Because one MOPEVACLAS animal experienced complete sterilizing immunity, this vaccine can be highly efficacious. The combination of GPC and NP appears to be the best choice for the LASV vaccine, because it provides complete protection and is efficacious against divergent LASV strains (10, 32). We postulate that removal of the NP exonuclease is important to avoid inhibition of innate and, consequently, adaptive immunity. Our vaccine candidates are the only replicating NP-based vaccines containing this feature to date. It has been reported that NP exonuclease removal does not improve the immunogenicity of LASV replicon particles, which can be explained by the absence of propagation of this single-cycle vaccine (33).

Previous studies have shown that GPC is essential in LASV vaccines (10, 11, 13). Modest GPC-specific T cell responses were induced after immunization in MOPEVACLAS and MeV-NP animals. The percentage of specific T cells was modest but consistent with that observed in humans after immunization with the 17D yellow fever virus (YFV) (34). The disappearance of responding T cells in PBMCs 30 days after immunization has previously been described in PBMC from MeV-infected macaques (35) and can be explained by the fact that activated lymphocytes transiently circulate in the blood stream before localization into secondary lymphoid organs or peripheral tissues (36). Moreover, the contraction of specific T cell populations after the initial proliferation that occurs during memory T cell development may also contribute to the decrease in the percentage of the responding T cells. Robust T cell responses against GPC peptides were observed after challenge in MeV-NP and MOPEVACLAS animals. In MeV control and MeV-Z animals, GPC-specific T cells were absent or detected at very low concentrations, respectively. The intensity of CD8+ and, to a lesser extent, CD4+ T cell responses specific for GPC appeared to be associated with low clinical scores and viremia, suggesting a contribution to LASV control, regardless of the vaccine used. Most CD8+ T cells and a large proportion of CD4+ T cells from the MeV-NP monkeys showed cytotoxic properties a few days after challenge, suggesting that efficient control of LASV after immunization is associated with cytotoxic T cell induction, most likely specific against GPC. The main subpopulations during the first 10 days after challenge were mostly composed of uniquely IFNγ-producing and, to a lesser extent, TNFα-producing T cells, after which polyfunctional T cells expressing both TNFα and IFNγ became dominant. CD8+ T cells from the vaccinated monkeys acquired EM2 and EM3 phenotypes after the challenge, but it occurred earlier for the MeV-NP animals. Among CD4+ T cells from the vaccinated monkeys, pE2 and then EM3 T cells became enriched. Last, the proportion of CD8+ and CD4+ EMRA T cells also increased in immunized monkeys. These observations show that, together with activation, proliferation, and cytokine production, the specific T cells acquired phenotypes consistent with long-term protective memory responses. We did not perform experiments with T cell depletion or the passive transfer of antibodies to identify the arm(s) of the immune system responsible for the observed acquired protection.

We identified early profiles associated with vaccine efficacy. Genes associated with the response to type I IFN were strongly induced 2 days after immunization in the MeV-NP and MOPEVACLAS groups, with the induction lasting longer for the MeV-NP animals. In the MeV-Z animals, we observed only modest expression of these genes at days 4 and 7. Such delayed expression may be due to attenuation of the MeV-Z candidate. This suggests that type I IFN and more generally inflammatory responses were induced very early after immunization with the protective vaccines, as previously described for the vesicular stomatitis virus-Ebola virus vaccine and 17D YFV (26, 37). This response probably plays a crucial role in the induction of the observed specific T cell responses, because type I IFN is an essential costimulatory factor during of T cell activation (28). Consistent with this hypothesis, pathways related to these transcriptomic changes and involved in innate and adaptive immunity were activated earlier and more strongly in the MeV-NP animals than the others. However, we also identified several pathways that were more highly activated in the MeV-Z than other groups. Similarly, the proteomic profiles of animals were modified from day 2 after immunization and thereafter. Each vaccine induced a specific profile, distinct from the others, but those of the MeV-NP and MOPEVACLAS animals were relatively similar and deviated substantially from that of the MeV-Z animals. We identified the genes and proteins for which the expression before challenge was the most predictive of vaccine efficacy. Several mediators potentially involved in inflammation/immune processes were highly predictive of efficacy. Helicase lymphoid-specific (HELLS) expression is required for the proliferation of mature T lymphocytes (38). Ficolin-2 (FCN2)/Ficolin-2–like is involved in the lectin pathway of complement activation and innate immunity (39). Carbohydrate sulfotransferase 1 (CHST1) expression by vascular endothelial cells regulates shear-resistant leukocyte rolling via l-selectin, thus promoting the flood of lymphocytes toward the site of inflammation (40). SERPIN A1 and A3 and intelectin-1 are acute phase proteins released during the inflammatory response (41, 42). Igκ constant chain is associated with the humoral response. FETUINB is a negative acute-phase protein (43). Properdin is a positive regulator of the alternative pathway that contributes to innate and adaptive immunity (44). Last, the C1q C chain is the initiation component of the classical complement pathway. Upon validation, these modulated genes and proteins could be used as predictive markers of vaccine immunogenicity and efficacy.

There are some limitations in this study. First, the number of animals per group was low due to the capacity of our biosafety level 4 (BSL-4) animal facilities. This low number is in agreement with the need to reduce animal use and nevertheless allowed us to evaluate vaccine efficacy. There were also some limitations concerning our analysis of the GPC-specific humoral responses. Protective LASV GPC–specific IgG often bind to the conformationally intact GP1/GP2 heterodimer, and the methods used in this study did not allow their detection. Last, the lack of identification of the immune mechanisms involved in the control of LASV infection did not substantially improve our knowledge of the immune responses necessary to control LASV relative to the previous LASV vaccine candidates evaluated in NHPs (911). However, this study represents a detailed description of the innate and adaptive immune responses induced in NHPs by LASV vaccines.

MeV-NP and MOPEVACLAS induced efficient protection after a single injection, and MeV-NP was the most efficient. Mass vaccination would be the ideal strategy to control LF, given the extensive number of people at risk for the disease and the frequent transmission of LASV from rodents (6). A single-shot vaccine provides an advantage because most targeted populations live in remote areas and the global cost of vaccination could be an issue. Furthermore, in case of an outbreak, ring vaccination with a one-shot vaccine represents an efficient strategy, as demonstrated for EBOV (45). MeV-NP could be used in a prime-boost regimen, as for the other MeV-based vaccines, to further improve its efficacy. The MeV-NP candidate has been selected by the Coalition for Epidemic Preparedness and Innovation and should enter into clinical development in the near future.


Study design

Our objective was to evaluate the immunogenicity and efficacy of various LASV vaccines. Groups of four male cynomolgus monkeys (Macaca fascicularis; 32 to 39 months old, 3 to 4 kg) were immunized in A2 facilities (SILABE, France) by subcutaneous injection of 2 × 106 TCID50 of MeV-NP, or MeV-Z, or by intramuscular injection of 6 × 106 FFU of MOPEVACLAS. Another group of three monkeys was immunized with the MeV vaccine strain Schwarz (2 × 106 TCID50, subcutaneously). Blood draws, oral and nasal swabs, and urine sampling were performed every 2 to 3 days during the first 2 weeks and then once a week up to day 37. After 37 days, the monkeys were transported to BSL-4 facilities (Laboratoire P4 INSERM–Jean Mérieux, France) and were challenged subcutaneously with 1,500 FFU of LASV (Josiah strain). Animals were followed for clinical signs and euthanized according to scoring based on body temperature, body weight, feeding, dehydration, behavior, and clinical signs. Blood draws, oral and nasal swabs, and urine sampling were performed every 2 to 3 days during the first 2 weeks and then once a week up to day 28. All animals reaching a clinical score of 15 were euthanized before the end of the experimentation (28 to 30 days after challenge) to limit animal suffering. All procedures were approved by the Comité Régional d’Ethique en Matière d’Expérimentation Animale de Strasbourg (20160826144775) and the Comité Régional d’Ethique pour l’Expérimentation Animale Rhône Alpes (20161110143954). Primary data are reported in data file S1.

Cell cultures

293 T7/N/P cells stably expressing the T7 polymerase and the MeV N and P proteins were used to rescue recombinant MeV and were maintained as previously described (16). Vero NK and Vero E6 cells were grown in GlutaMAX Dulbecco’s modified Eagle’s medium (DMEM; Life Technologies) supplemented with 5% fetal bovine serum (FBS) and 0.5% penicillin-streptomycin. Baby hamster kidney fibroblasts (BHK) T7 cells were grown in Glasgow’s minimum essential medium (Life Technologies) supplemented with 5% FBS, 10% tryptose phosphate broth, 1% nonessential amino acids, and 0.5% penicillin-streptomycin under hygromycin selection. Monocytes were isolated from the blood of consenting healthy donors provided by the Etablissement Français du Sang (Lyon, France) using a monocyte isolation kit (Miltenyi) and then differentiated into macrophages or dendritic cells as previously described (14).

Viruses and infections

To generate MeV/LASV vectors, codon-optimized open reading frames (ORFs) of LASV GPC, NP, and Z (Josiah strain) were cloned in the pTM1-MVSchwarz vector in additional transcriptional units (ATUs) placed upstream of N (ATU1 for Z) or between the P and M genes (ATU2, GPC alone, or NP + GPC). MeV expressing LASV antigens were rescued as previously described (16). Briefly, 293 T7/N/P cells were transfected with plasmids encoding the MeV L polymerase and the antigenomic segment of the MeV vector. Clonal syncytia were picked and used to infect Vero NK cells. Cells were then detached and overlaid on Vero NK cells to produce passage 1 (P1) stocks. Vero NK cells were then infected at an MOI of 0.01 and then incubated at 37°C for 2 to 3 days. To harvest virus, cells were scraped into Opti-MEM I–reduced serum medium and freeze-thawed twice. Titers were determined by TCID50 titration on Vero NK cells. The kinetics on Vero NK cells and infection of macrophages and dendritic cells were performed at an MOI of 1. Vector stocks for animal immunization were prepared in 1% DMEM.

The generation of MOPEVAC has been described elsewhere (14). Briefly, a recombinant MOPV virus mutated in the NP and expressing the LASV GPC (Josiah strain) was rescued in BHKT7 cells and amplified in Vero E6 cells.

LASV Josiah strain was produced on Vero E6 cells and then diluted in phosphate-buffered saline (PBS). Animals received 1,500 FFU by the subcutaneous route. Viral titers in plasma and organs were determined by titration on Vero E6 cells (14, 24).

For plaque reduction neutralization assays, plasma samples were diluted and mixed with 150 FFU of LASV. After 1 hour, the mixtures were added to Vero E6 cells and incubated for 1 hour before addition of carboxymethylcellulose diluted in DMEM. After 7 days, the number of foci was calculated by focus-forming immunodetection using anti-LASV antibodies.

Quantitative RNA analysis

For RT-qPCR experiments, total RNA was isolated using the RNeasy Mini Kit (QIAGEN), and a supplementary deoxyribonuclease step added using the Turbo DNA free kit Ambion (Thermo Fisher Scientific). Synthesis of complementary DNA was performed using SuperScript III, and amplification was performed using the Gene Expression Master Mix Kit (Applied Biosystems). For type I IFN, the primer/probe mix was developed in house (46). Runs of qPCR assays were performed in a LightCycler 480 (Roche Diagnostics). The expression of genes was standardized to that of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene and expressed as the fold induction relative to GAPDH. qPCR for viral RNA was performed with the EuroBioGreen qPCR Mix Lo-ROX (Eurobio) using LASV-specific primers and a quantified LASV NP–specific synthetic RNA standard developed in house (46).

Flow cytometry

Cells were detached by incubation with Versene solution (Life Technologies) or resuspended 48 hours after infection, saturated with human IgG (Fc receptor blocking solution, Miltenyi), and stained with antibodies to human CD40, CD83, CD80, CD86, CD14, CD1a, HLA-ABC, and HLA-DR (BD Biosciences, BioLegend) before fixation in PBS with 1% paraformaldehyde (PFA).

For intracellular staining, cells were stained with surface antibodies, permeabilized using the FoxP3 staining buffer set (Miltenyi), and stained using antibodies to Ki67 or GrzB (BD Biosciences). LASV-specific T cells were analyzed from 200 μl of fresh whole blood. Cells were incubated with overlapping GPC, NP, or Z peptides and anti-human CD28 and anti-human CD49d antibodies (2 μg/ml; BD Biosciences) and Brefeldin A (10 μg/ml; Sigma-Aldrich) for 6 hours at 37°C. Staphylococcus enterotoxin A (1 μg/ml; Sigma-Aldrich) and PBS were used as positive and negative controls of activation, respectively. The peptides (1 μg/ml each) were 15 amino acids long with an overlap of 11 residues and spanned the complete LASV GPC, NP, or Z ORF (Josiah strain). PBS-EDTA (2 mM final concentration) was added to samples before staining for CD3, CD4, and CD8 (BD Biosciences). Red blood cells were then lysed using PharmLyse (BD Biosciences). Cells were fixed and permeabilized using the FoxP3 staining buffer set (Miltenyi) before intracellular staining with antibodies to IFNγ and TNFα (BioLegend). Cells were analyzed by flow cytometry using an LSRFortessa cytometer (BD Biosciences) or a 10-color Gallios cytometer (Beckman Coulter). Data were analyzed using Kaluza 2.0 software (Beckman Coulter).


Thirty seven analytes were measured in plasma samples from EDTA tubes using magnetic-bead assays: NHP cytokine/chemokine magnetic bead panel I [granulocyte-macrophage colony-stimulating factor, transforming growth factor–α, granulocyte colony-stimulating factor, IFNγ, IL-2, IL-10, IL-15, sCD40L, IL-17, IL-1RA, IL-13, IL-1β, IL-4, IL-5, IL-6, IL-8, macrophage inflammatory protein–1α (MIP-1α), MCP1, TNFα, MIP-1β, IL12/23p40, vascular endothelial growth factor, and IL-18] and panel II (fibroblast growth factor–2, Eotaxin, IL-16, Fractalkine, sCD137, Granzyme A and B, IL-1α, IL-23, interferon γ-induced protein 10 (IP-10), sFasL, TNFβ, IL-28A, and Perforin) (Merck). Plates were prepared according to the manufacturer’s recommendations and analyzed on a Magpix instrument (Merck).

Hematology and biochemistry

The number of circulating CD4 and CD8 T cells, B cells, NK cells, monocytes, and neutrophils was determined by flow cytometry using cell-specific panels. Plasma concentrations of ALT and AST, LDH, creatine kinase, amylase, total bilirubin, direct bilirubin, creatinine, urea, iron, UA, total proteins, ALB, and CRP were measured using a Pentra C200 analyzer (Horiba Medicals).

Transcriptomic analysis

Total RNA was extracted from PBMCs using the RNeasy Extraction Kit (QIAGEN). RNA samples were then quantified using the Quantifluor RNA system (Promega) and qualified using Pico RNA Chips on a Bionanalyzer 2100 (Agilent). External RNA Controls Consortium (ERCC) RNA Spike-in Mix 1 (Thermo Fisher Scientific) was added to all samples to limit the variability of samples in multiple batches and mRNA was poly(A)-captured using NEXTflex Poly(A) beads (Bioo Scientific). The libraries were prepared using the NEXTflex Rapid Directional RNA-seq (RNA sequencing) Kit (Bioo Scientific) and were quantified and qualified using a Quantus quantification kit (Promega) and a fragment analyzer advanced analytical (AATI). Sequencing was performed on a NextSeq 500 Flow Cell High Output SR75 instrument with nine samples per flow cell. Bioinformatic and statistical analyses were performed using the Sequana RNA-seq pipeline and DESeq2 R package (47) and are described in Supplementary Materials and Methods.

Proteomic analysis

Plasma samples were first depleted of ALB and immunoglobulins using High-Select HSA/Immunoglobulin depletion columns (Thermo Fisher Scientific). The protein concentration of the depleted plasma was measured, and the samples were diluted with Laemmli 4× buffer at 1 mg/ml. Samples were heated for 5 min at 99°C. Proteins (20 μg) were loaded onto 4 to 12% gradient 10-well NuPAGE gels (Invitrogen) and subjected to short electrophoresis (5 min) in 2-(N-morpholino)ethanesulfonic acid buffer (Invitrogen). After Coomassie Blue Safe staining (Invitrogen) for 10 min and washing, the polyacrylamide band corresponding to the whole proteome of each sample was sliced out. The polyacrylamide bands were then bleached with water, reduced with dithiothreitol, and treated with iodoacetamide, before proteolysis with Trypsin Gold, Mass Spectrometry Grade (Promega) in the presence of 0.01% ProteaseMAX surfactant (Promega), as previously described (48). Fifty microliters of peptides were extracted per sample. The process for high-resolution tandem mass spectrometry (MS/MS), protein identification, and spectral count quantification is described in Supplementary Materials and Methods. Statistical analysis was performed using the limma R package (49) for differential analysis and camera method from the same package (50) for gene set analysis (50) (see Supplementary Materials and Methods for details).

Enzyme-linked immunosorbent assay

IFNα2 concentrations were detected in plasma using a human IFNα2-specific enzyme-linked immunosorbent assay (ELISA) (IFNα human-matched antibody pair, Life Technologies) according to the manufacturer’s instructions.

For the determination of LASV-specific IgG antibody titers, polysorp plates (Nunc) were coated overnight with lysate of LASV- or mock-infected Vero E6 cells as positive and negative antigens, respectively. After blocking with PBS 5% bovine serum albumin for 2 hours, plasma samples were then incubated for 2 hours on coated plates at dilution of 1:250, 1:1000, 1:4000, and 1:16,000 dilution. Plates were further incubated with polyclonal peroxidase–conjugated antibodies directed against NHP IgG γ-chain (Kirkegaard and Perry Laboratories). Tetramethylbenzidine (TMB) (Eurobio) was used for detection. The titer of positivity represented the highest dilution, which gave a positive signal. The threshold of positivity was calculated as 2 × means of three negative plasma samples + 2 SDs of the mean.

For the detection of MeV-specific IgM and IgG, Maxisorp plates (Nunc) were coated with inactivated MeV antigens (PR-BA 102, Jena Biosciences) at 1 μg/ml overnight at +4°C. Coated plates were then incubated with diluted plasma (1:100, 1:400, and 1:1600 for IgM; 1:250, 1:1000, 1:4000, 1:16,000, and 1:64,000 for IgG) or positive and negative control sera for 1 hour at 37°C. Plates were further incubated with polyclonal peroxidase–conjugated antibodies directed against NHP IgG γ-chain or IgM μ-chain for 1 hour at 37°C before detection using TMB. All samples were analyzed in duplicates, and the results are presented as the means of duplicates (dilution 1:400 for IgM and 1:1000 for IgG).

Detection of LASV-specific antibodies by Western blot.

LASV GP1 and GP2 [from GPC-expressing pseudoparticles (51)], NP (from purified LASV particles), or Z protein (IBT BioServices) was resolved on 4 to 15% criterion TGX gels and transferred to polyvinylidene fluoride membranes using the TransBlot Turbo transfer system (Bio-Rad). Membranes were blocked, and GP, NP, or Z strips were then incubated for 1 hour with plasma diluted 1:100 in PBS with 1% milk and 0.1% Tween 20 and then for 45 min with polyclonal peroxidase–conjugated goat anti-monkey IgG (Merck). Revelation was performed using the Opti-4-CN substrate kit (Bio-Rad). Monoclonal mouse anti-GP1 (clone 52-134-23A), anti-GP2 (clone L52-272-7), anti-NP (polyclonal mouse ascitic fluid 243-03), and polyclonal rabbit anti-Z (Agrobio) were used as positive controls with polyclonal peroxidase-conjugated antibodies anti-mouse or rabbit IgG (Sigma-Aldrich). Plasma from a naïve monkey served as a negative control.

Statistical analysis

Individual sample sizes and details of statistical analyses are specified in each figure caption or in Supplementary Materials and Methods. Statistical analyses were performed using Sigma Plot 14 (Systat Software Inc.), except for transcriptomics and proteomics. Data were analyzed using a one-way or two-way analysis of variance (ANOVA) if the dataset passed the normality test (Shapiro-Wilk) and the equal variance test (Brown-Forsythe). If not, a Kruskal-Wallis one-way ANOVA on ranks was used. The statistical analyses performed on transcriptomic and proteomic datasets are described in Supplementary Materials and Methods. The P value used to determine that there were significant differences in the gene expression, in the protein content, or in the activation of pathways was P < 0.05.


Materials and Methods

Fig. S1. Growth kinetics of MeV-NP and MeV-Z in Vero NK cells.

Fig. S2. Monitoring of body temperature in animals after immunization.

Fig. S3. Transcriptomic changes after immunization.

Fig. S4. Cellular pathway changes after immunization determined using the “Hallmark” gene set.

Fig. S5. Cellular pathway changes after immunization determined using the Kyoto encyclopedia of genes and genomes (KEGG) gene set.

Fig. S6. Cellular pathway changes after immunization determined using the “Reactome” gene set.

Fig. S7. Proteomic changes after immunization.

Fig. S8. Principal components analyses of the proteomic changes after immunization.

Fig. S9. Dynamics of white blood cell counts after challenge with LASV.

Fig. S10. Blood biochemistry of LASV-challenged animals.

Fig. S11. Viral RNA load in monkeys after LASV challenge.

Fig. S12. Memory T cell differentiation after LASV challenge.

Fig. S13. Specific T cell responses against Z protein peptides after LASV challenge.

Table S1. Genetic stability of the LASV genes introduced in measles-based vaccines.

Table S2. Detection of vaccine RNA by RT-PCR.

Table S3. Humoral responses after immunization.

Data file S1. Primary data.


Acknowledgments: We thank C. Combredet and N. Escriou (Viral Genomics and Vaccination Unit, Institut Pasteur) for advice on the generation of the MeV vaccines. We thank S. Mundweiller, A. Duthey, M. Dirheimer, B. Labrosse, D. Pannetier, S. Mély, and C. Léculier (P4 INSERM–Jean Mérieux, US003, INSERM) for assistance in conducting the BSL-4 experiments. We also thank J. Hortion, C. Germain, and M. Jourdain (UBIVE) for technical help and L. Bellanger (CEA, Marcoule) for the proteomics analyses. We thank S. Becker (Institut of Virology, Marburg, Germany) for providing us with the LASV; T. G. Ksiazek, P. E. Rollin, and P. Jahrling (Special Pathogens Branch, Center for Disease Control, Atlanta, GA) for the LASV monoclonal antibodies; and the Etablissement Français du Sang for the human blood. We are grateful to C. Gerke (Institut Pasteur) and the “Grand Programme Fédérateur–Vaccinologie” of the Institut Pasteur for invaluable support. We also thank S. Dussurgey and T. Andrieu from the ANIRA–flow cytometry platform (SFR BioSciences, Lyon). Funding: This study was supported by the Grand Programme Fédérateur–Vaccinologie of the Institut Pasteur and the Fondation pour la Recherche Médicale (FRM team grant 2013). The MOPEVAC vaccine was supported by the LABEX ECOFECT (ANR-11-LABX-0048, Lyon University), within the program “Investissements d’Avenir” (ANR-11-IDEX-0007, French National Research Agency). The MeV-based LASV vaccines received funds from the Fondation pour l’Innovation en Infectiologie (FINOVI). M.M. was supported by a Pasteur Roux fellowship. Author contributions: M.M., F.T., and S. Baize conceived the project. F.T. designed the measles vaccine platform. M.M., S.R., and S. Baize designed protocols. X.C. designed and generated MOPEVACLAS. M.M. designed, generated, and characterized measles recombinant vaccines. M.M., X.C., S.R., A.J., N.B., J.S., and C.P. performed RNA extractions, PBMC isolation, and flow cytometry sample preparation. L.B., A.V., and S. Barron took care of the animals and performed LASV challenge, scoring, sampling, hematological and biochemical analyses, and necropsies. L.B. supervised animal experiments. C.C. and H.R. supervised and validated BSL-4 protocols. S.R. and A.J. performed ELISA and Luminex assays. S.R. and S. Baize performed flow cytometry analyses. M.M. performed RT-PCR and qPCR analyses. M.M. and S.R. performed viral titrations, seroneutralization assays, and Western blots. M.M. prepared samples for transcriptomic and proteomic analyses. C.L.-L. and R.A. performed the RNA-seq. L.F. was in charge of the immunization of animals and clinical monitoring before the challenge. J.A. and J.-C.G. performed proteomics analyses. E.P., K.-H.H., N.P., and M.-A.D. analyzed transcriptomic and proteomic data. M.M. and S. Baize analyzed data and wrote the paper. S. Baize supervised the project. Competing interests: The authors declare that they have no competing interests. The MeV-based vaccine candidates described in this study are protected by the U.S. patent no. 62/609,155 and the MOPEVAC platform by the U.S. patent no. 62/245,631. Data and materials availability: All data associated with this study are present in the paper or the Supplementary Materials. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD014277 and 10.6019/PXD014277. The RNA sequences have been deposited in the Gene Expression Omnibus National Center for Biotechnology Information database with the accession number GSE132998.

Stay Connected to Science Translational Medicine

Navigate This Article