Development of a model of Saint Louis encephalitis infection and disease in mice

Eight- to 12-week-old wild type (WT) C57BL/6 or Balb/c mice were purchased from Centro de Bioterismo of UFMG (Belo Horizonte, Brazil). All animals were kept in the laboratory animal facility under controlled temperature (23 °C) with a strict 12-h light/dark cycle, food, and water available ad libitum. All experimental procedures were approved by and complied with the regulations of Universidade Federal de Minas Gerais (UFMG) Committee for Ethics in Animal Use (CEUA), under protocol number 349/2012.

SLEV strain BeH 355964 was provided by Prof. Luis Tadeu Moraes Figueiredo (Universidade de São Paulo, SP, Brazil). BeH 355964 stocks were generated by passage in C6/36 mosquito cell monolayers, cultivated in Leibovitz-15 supplemented with 10% v/v fetal bovine serum (FBS) (Cultilab, Brazil) and antibiotics. Clarified supernatants containing virus were titrated by plaque assay in Vero cells, and viral titers were expressed in plaque forming units (PFU)/mL of supernatant. SLEV BeH 355964 complete genome sequence is available at GenBank under accession number KM267635 [5].

Mice were inoculated intracranially (i.c.) or intraperitoneally (i.p.) with different inocula of SLEV or saline, as mock-infected control. Inocula were prepared by diluting viral stocks (as L-15 clarified supernatants) in saline. Viral dilutions in saline were at least 1000-fold, resulting in a solution that contained negligible amounts of the original L-15 culture supernatant. Injected volumes were 20 and 100 μL for i.c. and i.p. routes, respectively. For the i.c. inoculation, mice were anesthetized using Isoflurane (Biochimico, Brazil) 5% v/v inhalation. A syringe containing the inoculum was positioned perpendicularly to the head on the intersections of medial and sagittal planes, following insertion of the needle into the cranial cavity, injection and perpendicular removal from within the cranial cavity. Mice were observed twice a day for 14 days or up to determined time points for sample collection, at which mice were anesthetized with ketamine/xylazine (Syntec, Brazil) before collection of blood and organs, or euthanized by CO₂ inhalation. All tissue samples were stored at −80 °C until analysis. On survival experiments, mice presenting with severe disease signs (such as complete paralysis) were euthanized by CO₂ inhalation and considered dead in data analysis.

SLEV load in cell culture and tissue samples was determined by plaque assay and/or reverse transcriptase quantitative PCR (RT-qPCR). Tissue samples were processed into 10% w/v homogenates in DMEM prior to analysis by both techniques. Briefly, the plaque assay consisted in the serial dilution of samples for adsorption in Vero cell monolayers, for an hour. Samples were removed, following the addition of an overlay media containing 1.5% w/v carboxymethylcellulose (Synth, SP, Brazil) in 2% fetal bovine serum (FBS) v/v DMEM. After 7 days, plates were fixed with formaldehyde, washed and stained with methylene blue (Synth, SP, Brazil) 1% w/v. Results were expressed as plaque forming units (PFU)/mL of supernatant or PFU/100 mg of tissue. For the RT-qPCR reaction, samples were submitted to RNA extraction (QIAamp viral RNA extraction kit, QIAgen) and cDNA synthesis using random primers (Promega) and SuperScript Reverse Transcriptase III (Invitrogen), according to the companies’ specifications. The qPCR reaction was performed in the 7500 Fast platform using SYBR green reagents (Applied Biosystems) and primers targeting SLEV NS5 gene (primer forward FG1 TCAAGGAACTCCACACATGAGATGTACT, primer reverse nSLE ATTCTTCTCTC AATCTCCGT), as described elsewhere [31]. All PCR reactions were accompanied by a standard curve of the 232 bp NS5 amplicon. Results were expressed as relative number of genome copies of SLEV per sample.

Concentrations of Interferon (IFN) γ, CCL5, CXCL-1, IL-6, IL-1β, IL-10, IL-17, and TNF-α were quantified in cell culture or processed tissue samples by ELISA (R&D Systems, USA), following the manufacturer instructions. The detection limit of quantitative ELISA was in the range of 4–8 pg/mL or picogram per 100 mg of tissue. Results are expressed as picogram per 100 mg of tissue, picogram per milliliter of supernatant or by absorbance at 490 nm. Type I IFN (IFNα and IFNβ) levels were determined in processed samples by RT-qPCR using specific primers (IDT, USA). Cycle threshold (Ct) values of target genes were normalized to the housekeeping gene 18S and analyzed according to the ΔΔCt method: 2^{−ΔCt (sample)−ΔCt (housekeeping)}. Results are expressed as fold increase over the mock-infected WT group.

Assays to detect the activity of myeloperoxidase (MPO), eosinophil peroxidase (EPO), and N-acetil-β-d-glucosaminidase (NAG) were performed in tissue samples as a measure of leukocyte recruitment into target organs. To measure MPO and EPO activity, tissue homogenates were prepared in 1 mL of PBS containing 0.5% hexadecyltrimethyl ammonium bromide (HTAB) and 5 mM EDTA. Saline/Triton X-100 0.1% v/v was used to process tissues for NAG activity measurement. Test protocols were performed as already described [32]. Briefly, samples were evaluated for their ability to convert the substrates p-nitrophenyl-β-glycosamine (for NAG), 3,3′ 5,5′ tetramethylbenzidine (for MPO) or o-phenylenediamine (for EPO) to generate a colored solution proportional to the amount of enzyme in the sample. Plates for each assay were read at 405, 450, and 495 nm, respectively, and results expressed as absorbance.

Blood was collected from the brachial plexus of anesthetized mice in heparinized tubes for total and differential leukocyte count, platelet count, measurement of the hematocrit index, and, separately, for serum. Platelets and leukocytes were quantified in an optical microscope (Zeiss ICS Standard 25) using a Neubauer chamber or mounted microscope slides, containing Diff-quik-stained samples (Laboclin, Brazil). The hematocrit index was determined on centrifuged blood samples in heparinized glass capillaries. Results are as counts/mm³, per milliliter of blood (leukocytes), or percentage (hematocrit).

Mice were euthanized and perfused with 15 mL of PBS. The brains were collected, homogenized individually, and centrifuged against Percoll gradients (35 and 70%) for separation of leukocytes/microglia. Collected cells were washed, counted, and stained with antibodies against leukocyte surface markers (CD3, CD4, CD8, CD19, CD69, GR-1, F4/80, NK1.1) (all purchased from BD Biosciences). Cells were fixed in buffered 4% v/v formaldehyde and acquired in a FACS Canto II cytometer (BD Biosciences). Analyzed cell populations included lymphocytes, granulocytes, and macrophages/microglia, initially gated by size/granularity and subsequently by the expression of surface markers, and compared to negative and isotype-stained controls. Data were analyzed using FlowJo (Tree Star), and results expressed as total cell number of positive cells per brain/animal.

Mice were euthanized and perfused with 15 mL of PBS and 15 mL of buffered 4% v/v formaldehyde, for the collection of brains. Sections (5 μm thick) were made in a rostral to caudal fashion, mounted and stained in H&E. Sections were then analyzed for signs of brain pathology and inflammation, focusing on the regions of the hippocampus and the cortex/meninges. Images were obtained at ×200–400 magnification using an Olympus BX51 optical microscope equipped with a camera.

Behavioral changes induced in mice by SLEV infection were first assessed using the SHIRPA battery of tests [33]. Briefly, mice are submitted to 40 quick tests individually, observed and evaluated semi-quantitatively on their performance. The SHIRPA battery was performed in the laboratory animal facility, and results were expressed as a score for each animal in experimental groups. Mice were also assessed for alterations on the spontaneous locomotor activity, using the open-field method. Mice are placed individually in a transparent acrylic cage for 20 min. Their movement was recorded with a video camera and analyzed for total traveled distance, using the software Any-Maze (Stoelting Company). Results are expressed as traveled distance per mouse in a given experimental group.

N1E-115 mouse neuroblast cell line was kindly provided by Josiane Piedade (FUNED, Belo Horizonte, Brazil) and maintained in DMEM supplemented with 5 mM HEPES, 10% v/v SFB, and antibiotics. Primary neuronal cell cultures were obtained from the cortex and striatum areas of C57BL/6 embryos at 15 days post coitum. After dissection, the tissue was digested with trypsin and dissociated. Cells were plated into poli-l-ornitine-coated plaques with Neurobasal medium supplemented with 2nM Glutamax, B-27, and antibiotics (all purchased from Gibco, Thermo Fisher Scientific, USA). Cell cultures were incubated in 37 °C at 5% CO₂ for 5–7 days until use. Primary mixed glia cultures were obtained by collecting the brains of newborn mice (1–2 days old). Brains were minced and digested in 0.1% w/v trypsin (Gibco) for 15 min under agitation. Brain homogenates were washed in DMEM twice and plated in cell culture flasks containing DMEM supplemented with 10% FBS. Mixed glial cell cultures were harvested after 9–14 days. Briefly, cell cultures were infected with SLEV at a MOI of 0.1 for 1 h, washed in DMEM and incubated at 37 °C 5% CO₂ until sample collection (24–96 h post infection). Collected samples included cell culture supernatants for measurement of cytokine levels and viral load. Neuronal death was assessed in primary cultures using the LIVE/DEAD Viability/Cytotoxicity kit for mammalian cells (Molecular probes, Thermo Fisher Scientific, USA) and observed using a FLoid Cell imaging station (Life Technologies).

Results are expressed as mean plus standard error of mean, unless otherwise stated. Raw data were first analyzed for the presence of outliers (GraphPad quickCalcs) and checked for Gaussian distribution. Data sets were compared using ANOVA, followed by Tukey or Sidak post-tests. Differences between survival curves were analyzed using the log-rank test. Results with P < 0.05 were considered significant. All data are representative of at least two experiments (n = 4 to n = 12 replicates or n = 7 to n = 17 mice).

Our studies were initiated by the inoculation of SLEV BeH 355964, henceforth referred to as SLEV, in adult (8–12 weeks) wild-type (WT) BALB/c mice. Mice were inoculated with 10⁵ PFU of SLEV via the intraperitoneal (i.p.) or intracranial (i.c.) routes and followed for signs of disease and mortality for 14 days (Fig. 1a). The majority of mice inoculated with SLEV via the i.c. route presented disease signs such as ruffled fur and hunched back at 6 to 7 days after infection (p.i.), which developed into complete paralysis and death. Mice inoculated with SLEV via the i.p. route or injected with saline (Fig. 1a, Mock i.c., Mock i.p.) did not present with disease signs or death. We also performed an experiment in which mice were inoculated with SLEV via intraplantar or subcutaneous (s.c.) routes, in addition to i.p. and i.c. routes already tested. We observed that mice infected by peripheral routes (intraplantar, s.c. and i.p.) did not present disease signs or mortality, whereas mice inoculated with SLEV via the i.c. route manifested disease and mortality at 6-7 days p.i.  (Additional file 1: Figure S1).

To investigate if SLEV-induced mortality was inoculum-dependent, we infected groups of BALB/c (Fig. 1b) with different inocula of SLEV and followed for signs of disease and need for euthanasia. This experiment was performed in parallel with other commonly used laboratory mouse strains (SV129, Fig. 1c) (C57BL/6, Fig. 1d) to test SLEV infection reproducibility. Infected mice of all strains manifested severe disease and death in an inoculum-dependent fashion, although strains had different degrees of susceptibility to SLEV infection. SV129 mice (Fig. 1c) were found to be more resistant to SLEV i.c. infection, followed by BALB/c mice (Fig. 1b) and finally by the more susceptible C57BL/6 mice (Fig. 1d). All mock-infected mice (injected with saline) had no disease signs. In order to minimize variation and inconsistencies, only female mice were used in the following experiments.

Our data indicate that SLEV can cause disease and death in immunocompetent mice when injected directly into the CNS but not when injected systemically. SLEV-induced disease and mortality were inoculum-dependent and reproducible across different commonly used mouse strains.

In order to characterize the disease induced by the i.c. injection of SLEV in mice and ultimately the events leading to death in this model, we inoculated female adult C57BL/6 mice intracranially with 10³ PFU of SLEV. This inoculum was selected as it was equivalent to a lethal dose (LD₁₀₀) of SLEV in C57BL/6 mice (Fig. 1d). Mice were euthanized for sample collection at days 3, 5, and 7 after infection, in time points that precede and include the peak of mortality observed in previous experiments. Brain samples were collected, processed, and assessed for viral load by plaque assay (Fig. 2a) and RT-qPCR (Fig. 2b). Our results indicated that the number of plaque forming units (PFU) and SLEV genome copies increase exponentially in the brains of infected mice during the evaluated time points (Fig. 2a, b), which are both undetectable in mock-infected controls. SLEV accumulation in the brain peaked at day 7 p.i., as indicated by both techniques. Thus, our results indicate that, upon i.c. inoculation, SLEV infects and replicates in the mouse brain.

SLEV replication in the brain was accompanied by the local production of cytokines, chemokines, and interferons (IFNs). Tissue samples from mock-infected and SLEV-infected mice euthanized at days 3, 5, and 7 p.i. were collected, processed, and assessed by ELISA or by RT-qPCR (Fig. 3). The choice of cytokines was based on the known role of these molecules in the context of neuropathology, either by mediating inflammation (IL-6, IL-1β, TNFα, CCL5, CXCL1) and/or by their association to viral infections (IFNs). All cytokines measured were increased in infected brains as infection progressed to day 7 p.i., when compared to levels observed in the mock-infected group. The cytokines IL-6 (Fig. 3a), CCL5 (Fig. 3d), and CXCL1 (Fig. 3e) were already increased in the brains of SLEV-infected mice at day 5 p.i. In contrast to IL-6, levels of CCL5 and CXCL1 continued to increase and reached peak levels at day 7 p.i.. The cytokines IL-1β (Fig. 3b), TNFα (Fig. 3c), and IFNs (Fig. 3f, g, h) were increased only at day 7 p.i.. In summary, SLEV replication in the brain is associated to the expression of proinflammatory cytokines and IFNs, which reached peak levels at day 7 p.i.

The observation that SLEV caused the production of chemokines led us to investigate whether there would be influx of circulating leukocytes into the brain after infection. Blood samples collected from mice inoculated with saline (Mock) or 1 LD₁₀₀ of i.c. inoculated SLEV, at days 3, 5, and 7 p.i., were used for total and differential counting of leukocytes (Fig. 4a, b). Our results showed that SLEV-infected mice present reduced numbers of circulating leukocytes already at day 3 p.i., which is maintained during day 5 and further reduced at day 7 p.i. (Fig. 4a). The differential leukocyte count showed that the leukopenia presented by infected mice was due to lymphopenia, as infected mice had a significant decrease in lymphocyte numbers at all time points evaluated, when compared to uninfected controls (Mock) (Fig. 4b). In addition to leukocyte counts, blood samples were also evaluated for platelet counts and for the hematocrit index. SLEV infection could not alter the numbers of platelets or the hematocrit index, which remained similar to levels presented by saline-injected controls (Mock) (Additional file 2: Figure S2).

To continue our analysis, brain homogenates from infected and control mice were tested for the presence of infiltrating leukocytes through detection of the enzymatic activity of MPO, NAG, and EPO, present in neutrophils, macrophages, and eosinophils, respectively (Fig. 4c, d, e). Our results showed an increase in MPO activity (Fig. 4c) and a minor increase in EPO activity (Fig. 4e) in the brains of SLEV-infected mice at day 7 p.i., relative to the respective Mock controls. No differences were observed for the activity of NAG in mock and SLEV-infected groups (Fig. 4e). These data indicate that neutrophils, and to lesser extent eosinophils, are recruited to SLEV-infected mice brains.

In order to confirm our previous results and to investigate the recruitment of lymphocytes in SLEV in vivo infection, flow cytometry experiments were performed (Fig. 5). Groups of mice were injected with saline i.c. or inoculated i.c. with 1 LD₁₀₀ of SLEV. Brains were perfused, collected at days 5 and 7 p.i., and processed for the isolation of leukocytes. Recovered cells were composed mainly of neutrophils and lymphocytes (Fig. 5a) and increased in number as infection progressed (Fig. 5b). Neutrophils were the most prevalent leukocyte population in SLEV infected brains, with an average of 1.5 million cells recruited at day 7 p.i. (Fig. 5c). Macrophages and microglia were both evaluated according to the expression of F4/80 and found to be increased in SLEV-infected groups, although mock-infected mice present a significant amount of F4/80⁺ tissue-resident cells (Fig. 5d). Activated T CD4⁺ (Fig. 5e) and T CD8⁺ lymphocytes (Fig. 5f) were increased in the brains of infected mice at both days 5 and 7 p.i., reaching peak numbers at day 7 p.i.. Activated NK cells were also present in infected brains and also peaked at day 7 after infection (Fig. 5g). B lymphocytes were increased in the brains of infected mice only at day 7 p.i. (Fig. 5h). Mock-injected mice showed no evidence of leukocyte recruitment above basal levels.

Altogether, data presented in this section indicate that SLEV infection leads to lymphopenia and to significant leukocyte recruitment to the brain, composed of neutrophils and lymphocytes. Leukocyte recruitment in SLEV experimental infection follows chemokine production and viral replication.

Next, we examine whether i.c. SLEV infection would also induce systemic inflammation, as it inflamed the brain. Spleen and sera were collected from SLEV-infected and mock-infected mice at different time points p.i. and assessed for viral load and or signs of inflammation (Additional file 3: Figure S3). Spleens of both mock and SLEV-infected mice were negative for SLEV, as assessed by plaque assay (Additional file 3: Figure S3A), and showed no evidence for neutrophil or macrophage recruitment, as measured by the activity of MPO and NAG in tissue samples, respectively (Additional file 3: Figure S3B, C). Levels of CCL5, TNFα, and IFNγ, which are increased in the brains of infected mice, were similar between infected and non-infected mice (mock) in the spleen (Additional file 3: Figure S3D, E, F). Accordingly, CCL5, TNFα, and IFNγ could not be detected in the sera infected or non-infected mice throughout the evaluated time points (Additional file 3: Figure S3G, H, I). In summary, we found no evidence suggesting that SLEV i.c. infection becomes systemic in this model or induces systemic inflammation.

Because i.c. SLEV infection in mice caused brain inflammation and because human infection with SLEV is associated with significant morbidity, we assessed whether SLEV infection induced brain damage and/or functional alterations. Our results showed that SLEV causes progressive pathological alterations in the mouse brain, as compared to normal saline-injected brains (Fig. 6a, b). On day 3, a discrete sign of meningitis is observed (Fig. 6c, asterisk), which evolves to severe meningoencephalitis at days 5 (Fig. 6e) and 7 p.i. (Fig. 6g). The hippocampus was progressively damaged by SLEV infection throughout the evaluated time points (Fig. 6d, f, h, arrows), which may be associated with neuronal death, and presented infiltrating leukocytes at day 7 p.i. (Fig. 6h, asterisk). At day 7 p.i., microgliosis is observed in both cortex and hippocampus (Fig. 6g, h). Histological slides generated in the experiment were also subjected to semi-quantitative analysis (Additional file 4: Figure S4). Our data show that SLEV-infected mice present alterations consistent with meningitis and with damage to the cerebrum and hippocampus (Additional file 4: Figure S4A, B, C). In addition, we found that SLEV infection also caused significant alterations in the brainstem (Additional file 4: Figure S4D).

In order to characterize the behavioral alterations caused by SLEV infection, we used the SHIRPA battery of tests and the open-field test (Fig. 7). For the SHIRPA test, mock-infected and SLEV-infected mice at day 6 p.i. were subjected to a series of quick tests, to measure aspects of murine neurological function. The time point chosen (day 6 p.i.) is due to the fact that infected mice at day 7 p.i. are mostly incapacitated and thus, not able to perform any tests. Our results showed that SLEV-infected mice present a reduction in neuropsychiatric (Fig. 7a) and motor function (Fig. 7b) scores when compared to mock-infected mice, indicating that these neurological functions are affected by SLEV infection.

The open-field test measures the spontaneous movement of a mouse in a defined area. For this experiment, groups of mice were inoculated with saline (Mock), 1 LD₁₀₀ or 1 LD₅₀ of SLEV i.c., to compare mice that receive lethal or sublethal inocula of SLEV, and to study mice that survive the infection among those receiving a sublethal inoculum. We observed that at day 6 p.i., mice from control and infected groups traveled similar distances (Fig. 7c), indicating that these mice had the same rate of spontaneous movement. At day 12 p.i., 4–5 days after the onset of severe disease/mortality, surviving mice present reduced spontaneous movement, as observed by reduction in the distances traveled during the open-field test in comparison to the mock control group (Fig. 7d).

In summary, we observed that SLEV i.c. infection causes progressive brain damage in mice, likely including neuronal loss, that is characterized by meningoencephalitis. The extent of tissue damage correlates SLEV load and inflammation, suggesting that these processes are associated. The onset of experimental St. Louis encephalitis is preceded and followed by neurological alterations, especially in motor function.

In order to identify the main cell types infected by SLEV, we performed a series of in vitro infections of cells representing the main cell populations present in the CNS: neurons and glia (Fig. 8). The murine neuroblast cell line N1E-115 was first tested and found to be permissive to SLEV infection and replication (Fig. 8a). When infected with SLEV at a multiplicity of infection (MOI) of 0.1, N1E-115 cells released infective SLEV in the culture supernatant at all time points evaluated (24–96 h p.i.), reaching peak levels at 72 h p.i.. We next performed a primary culture of murine neurons isolated from C57BL/6 mice, which were infected with SLEV at a MOI of 0.1 and assessed for neuronal death (Fig. 8b). We observed that primary murine neurons are very susceptible to SLEV infection, as SLEV-infected cultures presented 50% cell death already at 24 h p.i.. Neuronal death increased to 80% at 48 and 72 h p.i. and resulted in complete death at 96 h p.i., when compared to mock-infected controls, which received culture medium.

Primary glial cultures were obtained from newborn C57BL/6 mice and consisted in a confluent mixed culture of astrocytes, oligodendrocytes, and microglia. Cultures were again infected with SLEV at a MOI of 0.1 and maintained for up to 120 h for supernatant collection (Fig. 8c, d). Samples were assessed for their viral content and results showed that SLEV replicates in glial cells, reaching peak levels in cultures at 72 h p.i. (Fig. 8c). SLEV replication was associated to the release of proinflammatory cytokines in the culture supernatant (Fig. 8d–g). Cytokines CCL5, IL-6, and CXCL1 were detected in glia culture supernatants and increased along the time points evaluated in SLEV-infected cultures. CCL5 was detected earlier in glial cultures, increasing in infected cultures at 48 h p.i. in comparison to Mock, and increasing further at 72, 96, and 120 h p.i. (Fig. 8d). IL-6 was increased in cell cultures at 72 and 96 h p.i. and reached peak levels at 120 h p.i. (Fig. 8f). CXCL1 was increased in comparison to Mock controls only at 120 h p.i. (Fig. 8g). Notably, IFNγ (Fig. 8e) was not detected in control or SLEV-infected cultures at any of the evaluated time points.

In conclusion, we suggest that SLEV is able to infect and replicate in cultures of both neuronal and glial cells of mice. In vitro SLEV infection is pathogenic, capable of inducing neuronal cell death and cytokine release by glial cells, which correlates with SLEV pathogenicity in vivo and suggests that SLEV may interact with more than one cell type in the brain.