|Home | About | Journals | Submit | Contact Us | Français|
A single injection of ML29 reassortant vaccine for Lassa fever induces low, transient viremia, and low or moderate levels of ML29 replication in tissues of common marmosets depending on the dose of the vaccination. The vaccination elicits specific immune responses and completely protects marmosets against fatal disease by induction of sterilizing cell-mediated immunity. DNA array analysis of human peripheral blood mononuclear cells from healthy donors exposed to ML29 revealed that gene expression patterns in ML29-exposed PBMC and control, media-exposed PBMC, clustered together confirming safety profile of the ML29 in non-human primates. The ML29 reassortant is a promising vaccine candidate for Lassa fever.
Lassa virus (LASV), a human pathogen of the family Arenaviridae, is a rodent-borne virus that causes Lassa fever (LF) [1,2]. LF is endemic in West African countries with the highest incidence in Nigeria, Guinea, Liberia, and Sierra Leone where up to 300,000 cases and 5000 deaths occur annually [3,4]. LASV antibodies were detected in 8–52, 4–55, and 21% of the population in Sierra Leone, Guinea, and Nigeria, respectively, bringing the population at risk to 59 million with an estimated annual incidence of illness of 3 million [4-6]. Increasing international travel has resulted in importation of LF to non-endemic areas including European countries and the US . The sizeable disease burden and the possibility that LASV can be used as an agent of biological warfare make a strong case for vaccine development [8-10].
LASV, like other members of the Arenaviridae, has a bisegmented (L and S) ambisense RNA genome [11,12]. The L RNA encodes a large protein (L, or RNA-dependent RNA polymerase) , and a small zinc-binding, Z protein . The S RNA encodes the major structural proteins, nucleoprotein (NP), and glycoprotein precursor (GPC), cleaved into GP1, GP2, and signal peptide [15-17].
We previously described a live attenuated experimental vaccine for LF, clone ML29 [18-20]. The ML29 vaccine candidate encodes the NP and GPC of LASV and the Z protein and L protein of MOPV. Eighteen mutations distinguish the ML29 genome from the parental strains and likely contribute to the attenuated phenotype. The ML29 genotype was stable throughout 12 passages in tissue cultures.
Proof-of-concept studies in rodent models showed that the ML29 vaccine was attenuated and induced protective cell-mediated immune responses [19,21]. However, LASV, a rodent-borne virus, is treated differently by the immune system of rodents and non-human primates . Recently we developed a small non-human primate model of LF in the common marmoset, Callithrix jacchus . We have used this model to evaluate safety, immunogenicity, and efficacy of the ML29 vaccine candidate.
LASV (strain Josiah), MOPV (clone An 20410), MOP/LAS (clone ML29), and LCMV-WE viruses were previously described [18-20,23]. The viruses were grown in Vero cells (ATCC, CRL-1586) and cultured in Dulbecco's modified minimum Eagle's medium (DMEM, GIBCO-BRL) with 2% fetal calf serum (FCS, GIBCO-BRL), 1% penicillin–streptomycin, and l-glutamine (2 mM) at 37°C in 5% CO2 by using a multiplicity of infection (MOI) of 0.01. Supernatants were collected at 72 h post-infection, titrated on Vero E6 cells, and virus stocks ((1–5) × 107 PFU/ml) were stored at −70°C.
Eighteen marmosets (C. jacchus) were used to examine ML29 safety, and immunogenicity, and 12 animals were used in ML29 efficacy trials. Animals were 2–4 years old and ranged in weight from 380 to 425 g. In the first study, marmosets were placed into two experimental groups (8 animals/ group: 4 males and 4 females) and two animals were used as naïve controls. Animals were subcutaneously (s.c.) immunized with 1 × 103 (low dose) and 1 × 106 PFU (high dose) of ML29 in 0.5 ml. Each week, two animals from the immunized groups were euthanized for immunological and histological studies as previously described .
For FACS analysis 75 μl of whole blood were used per staining sample. Red blood cells were lysed with ACK buffer (Quality Biological, Cat. no. 118-156-061). Remaining cells were washed with PBS and stained with conjugated antibodies against the following markers: CD3ε (BD, Cat. no. 55611, clone SP34, FITC), CD4 (BD, Cat. no. 552838, clone L200, PerCP-Cy5.5), CD14 (BD, Cat. no. 555399, clone M5E2, APC), CD8 (Serotec, Cat. no. MCA1226, clone LT8, APC), CD20 (BC/IOTest, Cat. no. 6604106, clone H299, FITC), and HLA-DR (BC/IOTesr, Cat. no. IM0464, clone B8.12.2, PE). After staining (15 min at 4 °C) cells were washed and re-suspended in 100 μl of 2% paraformaldehyde. Stained cells were run in a FACScalibur (BD Bioscience, San Diego, CA) and flow cytometry data were analyzed with FowJo software (Tree Star, San Carlos, CA).
PBMC from immunized marmosets were used in TNF-α ELISPOT assay (U-CyTech B.V., Ultrecht University, The Netherlands) according to the manufacturer's recommendations with slight modifications . Briefly, 2 × 106 cells in 0.5 ml of RPMI-1640 (Invitrogen) with 5% FCS, 2 mM glutamine, 100 units/ml penicillin, 100 μg/ml streptomycin, and 25 mM Hepes buffer were stimulated by co-incubation overnight at 37 °C with 2 × 106 PFU of ML29, MOPV, or LCMV-WE. After stimulation, the cells were washed, re-suspended in the same medium, and (0.3–0.4) × 106 cells/well were added to ELISPOT 96-well plates pre-coated with mouse anti-monkey TNF-α. The plates were incubated at 37 °C for 5 h, washed, and incubated with gold-conjugated anti-biotin. The spot-forming cells (SFCs) secreting TNF-α were developed with activator solution and counted (Immunospot 3.2 Analyzer, C.T.L. Cellular Tech., Ltd.). Anti-LASV in serum samples of ML29-immunized monkeys was measured by IgG ELISA as previously described .
The challenge study was performed in the BSL4 facility of Southwest Foundation for Biomedical Research (SFBR, San Antonio, TX). Six animals were s.c. vaccinated with 1 × 103 PFU of ML29, four animals received diluted conditioned medium of Vero E6 cells in the same volume (vaccine control), and two naïve animals were used as controls for histology studies. At day 30 after immunization all animals were s.c. challenged with LASV (Josiah), 1 × 103 PFU in 0.5 ml. At predetermined time intervals, animals were sedated and blood samples were collected for hematology, chemistries, and virus titration. When animals met euthanasia criteria  they were sacrificed and tissues were harvested for histology and immunohistochemistry.
The biochemical analysis of plasma samples was performed using a mammalian liver enzyme profile rotor on a Vetscan analyzer (Abaxis, Inc., Union City, CA). Complete blood counts were performed using VetScan HMT machine (Abaxis). Histology and immunochemistry were performed as previously described . A portion of each tissue (lung, spleen, MLN, liver, stomach, ileum, kidney, heart, cerebrum, and cerebellum) was submerged in MEM medium with 10% FCS (for plaque titration) and in RNAlater (Qiagen Inc., Valencia, CA) for RNA isolation. The remaining tissue portions were fixed in 10% neutral formalin for the preparation of standard histological sections. Paraffin-embedded tissues were cut in 5-μm sections, deparaffinized, and stained with hematoxylin–eosin. For immunochemistry, deparaffinized samples were stained for CD3, CD20, HAM56, Ki67, HLA-DR, and MHC-II antigen (Dako, Carpenteria, CA) [22,24].
Plaque assay, Vero cell co-cultivation, and nested RT/PCR were used to assess viremia, viral load in tissues, and ML29 shedding. Standard plaque assay was performed on Vero E6 cell monolayer as previously described . For co-cultivation assay plasma samples or dilutions of 10% tissue suspensions were incubated on monolayer of Vero cells during 72 h and supernatants were used for infectious virus detection by standard plaque assay. To detect virus in circulating mononuclear cells, PBMC samples were tested in infectious center assays as previously described . In brief, serial 1:10 dilutions of PBMC samples were plated on sparse Vero cell monolayers and incubated 72 h. After incubation culture supernatants were removed and substituted with agarose containing overlay medium. Plaques or infectious centers produced by cells containing infectious virus particles were visualized by crystal violet staining and infectious centers were expressed per 1000 plated cells.
For detection of viral RNA in blood samples, RNA was extracted from 140 μl of plasma by using a QIAamp viral RNA minispin protocol (Qiagen, Cat. no. 52904) or from 500 μl of EDTA-treated blood using the RiboPure blood kit (Ambion, Austin, TX). RNA was extracted from tissue samples submerged in RNAlater using RNeasy mini kit (Qiagen, Cat. no. 75142), converted into cDNA, and amplified with 36E2 and 80F2 primers targeting LASV GPC . To detect ML29 shedding, oral, vaginal, and rectal swabs and urine samples (100 μl) were immersed in RNAlater and RNA was extracted using the RNeasy mini kit. RNA was converted into cDNA and amplified with 36E2 and 80F2 primers. Five microliters of PCR products was further amplified with two internal primers (reverse: 5′TGTGCAAGACCTACCACACAACAG; forward: 5′TGGTTGCGCAATTCAAGTGTCC). The expected product, a 200 bp DNA band, was visualized by agarose gel staining. The sensitivity of the nested PCR was found to the approximately 0.5 PFU/ml.
PBMCs were obtained by apheresis from normal volunteers followed by Ficoll–Hypaque isolation and re-suspension in RPMI-1640 with 10% human AB serum (Sigma). Mononuclear cells (1.2 × 107) were exposed to ML29 and MOPV (MOI of ≥ 1) for 45 min at 37 °C, plated in duplicate, incubated at 37 °C in 5% CO2, harvested at 4, 8, and 24 h post-exposure, and used for RNA extraction. As a control an exposure to diluted conditioned medium of Vero cells was used. RNA samples were treated with RNAse-free DNase and the quality and quantity of the samples was evaluated (Bioanalyzer 2100, Agilent, Santa Clara, CA). A custom-made cDNA library containing 7489 genes  was used for microarray experiments. RNA samples were labeled with Cy3 dye using the TSA labeling kit (PerkinElmer, Boston, MA) and co-hybridized with a reference RNA. Experiments were carried out in duplicate for each RNA sample. Array images were scanned using GENEPIX PRO 4100b (Axon Instruments, Inc., Sunnyvale, CA) optical scanner and the image analysis was performed using image processing tools (BioDiscovery, El Segundo, CA) as previously described [27,28].
None of the ML29-immunized marmosets had clinical mani-festations during the observation period and hematological and chemical parameters were in normal ranges. With the high dose of ML29 we observed a minor elevation of alanine aminotransferase (ALT) in the plasma in comparison with animals immunized with the low dose. However, this elevation was still in the normal range (Fig. 1). In LASV-infected marmosets ALT levels in plasma were highly elevated indicating that liver was involved in LF pathogenesis [22,29].
Immunization with ML29 at low dose did not result in viremia detectable by conventional plaque assay (Table 1) . Only in one marmoset, CJ17 (high dose), the virus was detected by plaque assay. Co-cultivation assay detected the virus in one marmoset on day 21 (CJ15), in one marmoset on day 28 (CJ7) after a low-dose immunization, and in one marmoset on day 14 (CJ12) after immunization at high dose. All PBMC samples collected during the observation period were virus-negative as judged by infectious center assay; virus was recovered by co-cultivation only from one sample collected on day 28 (CJ20) after immunization with a high dose of ML29 (Table 1).
In marmosets immunized with low-dose ML29 the virus transiently replicated at low levels (1.8–2.7 log10 PFU/g) in spleen, mesenteric lymph nodes, and lung. The virus was also detected in liver tissues of animals CJ15 and CJ7 on days 21 and 28, respectively, by co-cultivation assay (Table 1). In animals immunized with high dose the virus was found as early as day 7 and transiently replicated at moderate levels, 2–4.5 log10 PFU/g. On days 21 and 28 in three marmosets immunized with high dose, ML29 was detected in some areas of the brain and was recovered from the spinal cord.
Urine samples collected from immunized animals did not contain ML29 as judged by biological assays. Among all tested swab samples, viral RNA was detected only in oral, rectal, and vaginal samples collected on day 14 from one monkey, CJ12, immunized with high dose of ML29 (not shown).
We used FACS analysis to analyze fluctuations of blood cell subsets as well as expression of an activation marker, HLA-DR. Among cell subsets, only the distribution of CD14 and CD3 T lymphocytes was affected in immunized monkeys. As seen in Fig. 2, ML29 immunization increased the population of mononuclear CD14+ cells. In all immunized animals we have also found increased numbers of CD3+ T cells among peripheral lymphocytes (from 49.4 ± 13.0 to 73.1 ± 16.3% before and after immunization, respectively). We also saw differences in the frequency of HLA-DR+CD3+ lymphocytes before and after immunization. However, these differences were not statistically significant when all vaccinated vs. non-vaccinated subjects were compared. Among CD3-negative cells (B cells and NK cells) the frequency of activated HLA-DR+ cells was increased after immunization (from 71.1 ± 16.3 to 93.4 ± 5.2%) and these differences were statistically significant (p < 0.05).
Hematoxylin–eosin staining of tissues of ML29-immunized marmosets revealed some evidence of lymphocyte hyperplasia only in animals immunized with the high dose of ML29 and at the later time points. The liver in high-dose-immunized animals also had evidence of transient inflammation. However, morphologically these lesions differed from what we observed in animals infected with LASV (not shown).
Acute LASV infection of common marmosets resulted in multifocal hepatic necrosis with HAM56-positive infiltrates, hepatocyte proliferation, marked reduction of CD20+ and CD3+ cells, and substantial reductions in the intensity of HLA-DP, DQ, and DR staining . In contrast, in ML29-immunized animals we found over-expression of HLA-DR, P, Q, and recruitment of CD3+ cells to the hepatic parenchyma, all clearly seen in animals immunized with the high dose of the ML29 (Fig. 3). Staining for CD20+ B cells did not reveal significant differences between control and ML29-immunized marmosets. All other tissues had unremarkable histological findings within normal limits.
Previously we have shown that a single s.c. immunization of rhesus macaques with ML29-induced robust cell-mediated immune responses detectable in peripheral blood by U-CyTech IFN-γ ELISPOT as early as 7 days after immunization . Unfortunately, U-CyTech anti-macaque IFN-γ antibody pairs did not cross-react with marmoset IFN-γ and we had to use cross-reacting TNF-α antibodies in the appropriate ELISPOT assay (U-CyTech, Cat. no. CT133). PBMC from blood of vaccinated animals were stimulated by co-cultivation with ML29 and TNF-α SFC/106 were detected by ELISPOT as described in Section 2.
As seen in Table 2, after antigen stimulation cells secreting TNF-α were detected on day 14 and the frequency of SFC increased depending on immunization dose on days 21 and 28. On day 28 in marmosets immunized with the high dose the average number of TNF-α secreting cells was almost sevenfold higher than in mar- mosets immunized with low-dose ML29. Stimulation of PBMC from immunized marmosets with closely related viruses (MOPV and LCMV), but not with TACV, revealed low levels of cross-reactivity (not shown) in conformity with the previous observations in ML29-immunized rhesus macaques . Anti-LASV IgG antibodies were detected in ELISA on days 21 and 28. Their titers were low and did not differ between animals after immunization at low or high doses (Table 2). Anti-LASV neutralizing antibodies were not detectable (<1:10) in immunized marmosets (not shown).
Vaccinated and control marmosets were challenged on day 30 with 1 × 103 PFU/0.5 ml of LASV-Josiah. By days 8–10, challenged control animals became depressed, reduced stool production, became partially anorexic, and lost nearly 10% of their body weights. The disease gradually progressed and on days 17–21 all control animals met euthanasia criteria [19,22] and were sacrificed. No vaccinated animals had any clinical manifestations and all survived the observation period lasting 35 days after challenge (Fig. 4). In vaccinated and challenged animals blood chemistry and hematology data did not reveal differences from pre-challenge values (not shown). In contrast, challenge controls showed reduction of platelet numbers, elevated liver enzymes, and decreased levels of albumin in plasma as was previously described .
Plasma samples collected after challenge were tested by plaque assay and co-cultivation with Vero cells. As seen in Table 3, in all non-vaccinated marmosets LASV was detectable on day 5 after challenge and the viremia was more than 5–6 log10 PFU/ml shortly before or on the day of necropsy. Only in one ML29-vaccinated marmoset, CJ26493, LASV was detected by plaque assay on day 5. Two other LASV-positive animals, CJ26946 and CJ27008, had low transient viremia detectable by Vero co-cultivation. After day 15 no animals had LASV in plasma as judged by biological assays or by nested RT/PCR (not shown).
In tissues of non-vaccinated animals LASV load was high and varied from 3.3 to 7.6 log10 PFU/g. All tested tissues of ML29-vaccinated and LASV-challenged animals had no infectious LASV as judged by biological assays, plaque titration and Vero co-cultivation (Table 4).
Histological results from vaccination–challenge experiments are summarized in Table 5. Four animals from a non-vaccinated control group had lesions compatible with wild-type LASV infection and included: (i) wide spread hepatocellular necrosis with intracellular inclusions; (ii) severe interstitial pneumonitis; (iii) multifocal adrenal necrosis; and (iv) severe lymphocytic necrosis targeting B cell regions of the lymph nodes and spleens. All six vaccinated and LASV-challenged animals were essentially unremarkable with histological findings within normal limits.
Safety is the major concern for live attenuated vaccines. To further characterize safety features of ML29 we performed gene expression profiling in human PBMC exposed to ML29, to MOPV, and to conditioned Vero culture medium (mock-exposed control). Total RNA from infected cells was extracted at 4, 8, and 24 h after exposure and hybridized with a custom-made cDNA library .
Cluster overview of gene expression profiles is presented in Fig. 5. Genes (probe sets) with expression changes of at least twofold were designated “differentially expressed”. Exposed PBMC showed changes in 400 genes. Among them, 14 genes displayed opposite patterns (down- vs. up-regulation) after exposure with ML29 and MOPV. These genes are involved in responses to stimuli and cell–cell communication and are listed in Table 6. The ML29 expression profiles were more similar to those of mock-exposed samples than to MOPV-exposed samples. Principal component analysis (PCA) confirmed the cluster analysis data. Samples from ML29 and control groups were clustering together and maintained a significant distance from the MOPV group along the first principal component axis (x-axis: PCA1), which, incidentally, represents the highest variance between the analyzed groups.
There are several reasons to justify a replication-competent, “live”, vaccine as an attractive approach to control LF: (i) cell-mediated immunity plays the major role in LF patient recovery and in protection; (ii) a live vaccine provides the most effective natural pathway to process and present protective antigens to MHC molecules; (iii) epidemiological observations provide evidence that a single LASV exposure will induce long-term protection against disease ; and (iv) a vaccine candidate formulated to contain both LASV NP and GP antigens will induce a broad cross-reactivity and strong CD4+ memory T cells against all phylogenetic groups of LASV [19,30,31].
Safety experiments in common marmosets confirmed our previous data in rhesus macaques . Notably, the ML29 reassortant was detectable in plasma by conventional plaque assay only in one sample after immunization at high dose (Table 1). Low-dose immunization resulted in transient ML29 replication in tissues at very low levels. High-dose immunization resulted in more extensive virus replication. Still, this replication was transient, moderate (2–4 log10 PFU/g), and well controlled. However, ML29 immunization at high dose can damage the blood–brain barrier in some marmosets.
Immunization at low dose did not result in detectable virus shedding, whereas transient viral shedding was detected in one marmoset, CJ12, immunized at high dose (Table 1). Although shedding by the respiratory route after immunization at high dose cannot be excluded, transmission of vaccine virus by this route is very unlikely. Entry of arenaviruses occurred only via basolateral receptors  suggesting that ML29 can infect airway epithelia only when epithelial integrity is compromised.
To further characterize ML29 safety features we used transcriptome profiling in human PBMC. Exposure of PBMC samples to ML29 and MOPV showed changes in 400 genes but only 14 displayed opposite regulation. Interestingly, MHC-II DR beta 5 gene was significantly up-regulated in MOPV-exposed vs. ML29-exposed cells. However, immunochemistry analysis (Fig. 3) showed strong up-regulation of MHC-II class genes at the protein level suggesting that the transcriptional and translational status of the various HLA-DR loci can vary in MOPV- and ML29-exposed cells. Overall, as seen in Fig. 5, the ML29 gene expression pattern in human PBMC was closer to the pattern of mock-exposed cells and clearly maintained a significant distance from the MOPV-exposed group.
Immunization of marmosets with ML29 increased populations of CD14+ cells and CD3+ T lymphocytes in circulating blood. We also saw recruitment of CD3+ T cells and over-expression of HLA-DR, P, and Q in target tissues (Fig. 3). Taken together these data indicate that ML29 immunization resulted in antigen stimulation. In contrast, LASV infection in marmosets was associated with lymphoid depletion, marked reduction in CD3+, CD20+ cells, and down-regulation of class II MHC antigens .
Our results in rhesus macaques  and in marmosets (present study) indicate that a single s.c. inoculation of ML29 vaccine at low dose induces specific cell-mediated T cell responses assayed in IFN-γ or TNF-α ELISPOT. These responses seem to be responsible for complete protection against fatal LF disease. In confirmation of these observations target tissues of vaccinated and challenged animals collected at the end of the experiment had no histological alterations and, notably, were free from infectious LASV (Tables (Tables44 and and5).5). In ML29-immunized primates we detected very weak antibody responses measured by IgG ELISA (Table 2, see also Ref. ).
The existence of natural reassortment raises the concern that vaccination with an attenuated reassortant vaccine in the presence of the virulent strain will somehow accelerate disease spread. However, experiments with Rift Valley fever virus showed that reassortants between the wild-type virus and a live attenuated ML-12 vaccine generated only phenotypes with protective activity against the disease . Studies on molecular epidemiology of infectious bursal disease virus (IBDV) confirmed the existence of natural reassortants between vaccine and wild-type IBDV strains. However, the reassortant that exhibited a segment from a natural virulent strain and a segment from a vaccine strain induced significantly less mortality than typical wild-type IBDV . No evidence of wild-type reversion has been observed in recovered FluMist vaccine strains that have been tested. In clinical trials transmission of FluMist virus has been documented only in a single person, who remained asymptomatic .
Natural reassortment between arenaviruses has not been described so far. Recently we have performed simultaneous vaccination-challenged experiments in guinea pigs and showed that simultaneous application of ML29 and LASV attenuates wild-type infection and protects animals . All these data indicate that genetic reassortment with wild-type viruses during a vaccination process would also be expected to yield attenuated variants and to reduce the overall incidence of disease.
In summary, a ML29 reassortant is safe, immunogenic, and induces complete protection at low dose against LF in non-human primates. Based on cross-protective studies in guinea pigs , it is reasonable to expect that ML29 vaccination will also protect non-human primates against phylogenetically diverse LASV isolates. Cell-mediated immunity seems to be responsible for effective LASV clearance from blood and tissues of immunized animals . ML29 phenotype and genotype have been stable for at least 12 passages in tissue cultures. None of the existing LF vaccine candidates [9,36-38] can share these ML29 features. The recent demise of two medical doctors in Nigeria  is one more reminder that an effective vaccine for LF is urgently needed and that local health workers should be the first people offered this vaccine. In West Africa as many as a tenth of the potential vaccinees have an altered immune status. For these individuals the risk of adverse effects should be carefully evaluated and compare with vaccination benefits. To address this issue, experimental efforts are under way to determine whether ML29 is still safe and effective in the monkey model for AIDS.
This work was supported by grant RO1 AI052367 (to I.S.L.) from the National Institutes of Health and by the Regional Centers of Excellence in Biodefense and Emerging Infectious Diseases Researches, MARCE (U54 AI057168, subcontract to I.S.L.) and WRCE (U54 AI057156, subcontract to J.L.P.). Funding for microarray studies was from NIH grant AI053620 (to M.S.S.). We acknowledge the New England Primate Center (grant P51RR00168-45) for histological support.
✩This work was presented in part at the Keystone Symposia “Challenge of Global Vaccine Development”, 8–13 October 2007, Cape Town, South Africa, and at the 13th International Congress on Infectious Diseases, 19–22 June 2008, Kuala Lumpur, Malaysia.