Cortical Neurons Derived from Equine Induced Pluripotent Stem Cells Are Susceptible to Neurotropic Flavivirus Infection and Replication: An In Vitro Model for Equine Neuropathic Diseases

Abstract

Horses are susceptible to a number of neurotropic viruses, including West Nile virus (WNV), which is a pathogen of global significance in both horses and humans. However, there are no in vitro models with which to study infectious neuropathic diseases in the horse. In an effort to redress this, we have generated neurons from equine induced pluripotent stem cells (equiPSCs) that express a range of cortical neuron-specific markers, in addition to the membrane-bound ligand ephrin B3, which plays an important role in axon guidance as well as functioning as the receptor through which henipaviruses, such as Hendra virus, enter mammalian neurons. EquiPSC-derived neurons spontaneously depolarize with waves of depolarization conducted unidirectionally to adjacent neurons. We sought to confirm that equiPSC-derived neurons are a possible in vitro model for viral neuropathic diseases in the horse by examining their susceptibility to infection with flaviviruses that are known to be neurotropic in horses, including WNV and Murray Valley encephalitis virus (MVEV), and to compare these to nonpathogenic flaviviruses such as Fitzroy River virus (FRV) and Bamaga virus (BgV). All three strains of WNV tested in this study grew to high titres in the equiPSC-derived neurons, inducing a strong cytopathic effect (cpe), as did MVEV. In contrast, FRV showed restricted replication, and no cpe, which is consistent with the observation that FRV infects, but does not cause disease, in horses. BgV, which is thought to infect only marsupials, did not replicate in the equiPSC-derived neurons. Hence, our equiPSC-derived neurons display virus-specific differences in terms of viral titre and cpe that are similar to observations made in vivo, thus supporting their use as an in vitro model for neurotropic viral infection in horses.

Introduction

The study of infectious neuropathic diseases in horses has been hampered by a lack of equine primary neuronal cultures to enable disease modeling. Protocols for the induction of neuronal differentiation in human pluripotent stem cells have been successful in producing both cortical [1,2] and motor neurons [3,4], making it possible to replicate diseased neuronal phenotypes in vitro, and facilitating the development of therapeutics. While Sharma et al. have previously produced motor neurons from equine induced pluripotent stem cells (iPSCs) [5], we sought to establish an in vitro model for equine neuropathic diseases by generating cortical neurons from equine iPSCs (equiPSCs) previously produced in our laboratory [6]. To explore their potential as an in vitro model for virus-induced neuropathic diseases in the horse, we examined their susceptibility to infection with flaviviruses that are known to be neurotropic in horses, including West Nile virus (WNV) and Murray Valley encephalitis virus (MVEV).

WNV and MVEV are mosquito-borne pathogenic flaviviruses that can cause severe, potentially fatal, encephalitis in humans and horses [7]. The geographic distribution of WNV is continuing to expand with outbreaks now identified in Europe, Asia, North, Central and South America, the Middle East, and Australia, making it a pathogen of global significance. Since its first incursion into the United States in 1999, WNV has been responsible for more than 16,000 human cases of neurological disease and over 1,500 deaths [8]. In 2011, Australia experienced an unprecedented outbreak of equine encephalitis caused by a virulent strain of WNV (WNV_NSW2011) that affected more than 1,000 horses, with a mortality rate of 10%–15% [9,10]. Curiously, while highly virulent in horses, WNV_NSW2011 caused only a single case of mild disease in the human population [9,10].

MVEV is endemic to Australasia and, although uncommon, has also caused fatal encephalitis in both horses and humans [11,12]. Australia is also host to a large number of apparently nonpathogenic flaviviruses, including two recently discovered viruses belonging to the yellow fever virus group: Fitzroy River virus (FRV) and Bamaga virus (BgV). FRV is widespread in northern Australia where it infects humans and domestic animals, including horses, although to date there is no indication that it causes disease in either horses or humans [13]. In contrast, BgV appears to have a more restricted host range that may include marsupials but unlikely other mammals [14].

Materials and Methods

Maintenance of equiPSCs

Two previously characterized clones of equiPSCs (A and K) [6] were cultured on feeder layers of irradiated mouse embryonic fibroblasts (MEFs) at a density of 20,000 MEFs/cm². Maintenance medium consisted of KnockOut DMEM (Gibco), 15% (v/v) embryonic stem cell-qualified fetal bovine serum (ESC-FBS) (Hyclone), 0.1 mM nonessential amino acids (NEAAs) (Gibco), 2 mM l-glutamine (Gibco), and 0.1 mM β-mercaptoethanol (Gibco), which was supplemented with 1,000 U/mL murine leukemia inhibitory factor (LIF) (Millipore) and 10 ng/mL human basic fibroblast growth factor (bFGF) (Invitrogen).

EquiPSCs were maintained on feeder layers for 7 days with maintenance medium before undergoing neuronal differentiation (see next section). Colonies were manually dissected and ∼9 individual sections were passaged onto MEF-seeded organ culture dishes (Costar, Corning Life Sciences).

Neuronal differentiation of equiPSCs

After 7 days, the maintenance medium was replaced with neuronal medium consisting of 50% (v/v) Neurobasal medium (Gibco), 50% (v/v) KnockOut DMEM, 0.5 × N2 Supplement (Gibco), 0.5 × B27 (Gibco), 2.5 μg/mL human insulin (Invitrogen), 2 mM l-glutamine, 0.1 mM NEAAs, 0.1 mM β-mercaptoethanol, and 50 U/μg-mL penicillin/streptomycin (Invitrogen). Neuronal differentiation was induced by supplementing the neuronal medium with the dual-SMAD inhibition factors dorsomorphin (1 μM) (Miltenyi-Biotec) and SB431542 (10 μM) (Sigma-Aldrich) according to the protocol by Chambers et al. [1].

EquiPSCs were cultured in the neuronal differentiation medium for 8 days before being manually dissected and segments seeded onto the Matrigel-coated (Sigma-Aldrich) wells of a six-well plate. Passaged neuronal cultures were maintained in neuronal differentiation medium for an additional 4–6 days (total time in neuronal differentiation medium 12–14 days) before dissociation into single cells with Accutase (Invitrogen) and reseeded onto Matrigel-coated six-well plates for maintenance, or chamber slides (Nunc) for viral infections and immunocytochemistry.

Immunocytochemistry

Samples were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) (Invitrogen) at room temperature (RT) for 15 min, before blocking and permeabilization with 10% goat serum (Invitrogen) and 0.1% Triton X-100 (Ajax Finechem) in PBS for 1 h at RT. Primary antibodies were diluted in 10% goat serum in PBS and incubated overnight at 4°C. Primary antibodies and their dilutions are as follows: TUJ1 (1:500; Merck), Ephrin B3 (1:200; Abcam), SOX2 (1:150; Cell Signaling Technology), BRN2 (1:150; Cell Signaling Technology), and TBR2 (1:150; Abcam). Negative control samples were incubated with 10% goat serum in PBS without primary antibody. Samples were then incubated with isotype- and species-matched Alexa-Fluor conjugated secondary antibodies (1:1,000; Invitrogen) for 1 h at RT. Nuclei were visualized with Hoechst33342 (1:10,000; Invitrogen) or DAPI (1:10,000; Invitrogen).

RNA isolation, cDNA synthesis, and quantitative polymerase chain reaction

Total RNA was isolated with TRIzol using the Ambion PureLink RNA Mini Kit (Ambion) according to the manufacturer's instructions. Complementary DNA was synthesized with the iScript DNA Synthesis Kit (Bio-Rad Laboratories). Quantitative polymerase chain reactions (qPCRs) were performed with SsoFast EvaGreen Supermix (Bio-Rad Laboratories). All target genes were normalized to endogenous equine GAPDH. Primer pairs used are listed in Table 1. Three replicate qPCRs were performed for each transcript, and the relative changes in expression of each gene were determined via the 2^−ΔΔCT method [15].

Table 1.

Quantitative Polymerase Chain Reaction Primer Pairs

Gene	Forward—(5′–3′)	Reverse—(5′–3′)	Product size
eqTBR1	TCAGTTCATCGCCGTCACCG	GTTCTTCTCGCAGTTGCGCT	172
eqTBR2/eqEOMES	GCTCAAATTCCACCGCCACC	CGTTGTAGTGGGCACTGGGA	113
eqPAX6	TCAGAGCCCCAGTTTCGAGC	CGCCCGTTGACAAAGACACC	111
eqEMX1	GCGCCTTCGAGAAGAACCAC	CTCCTCCTCCAGCTTCTGCC	140
eqFOXG1	GAAAACGACACCGAGGCACC	GACAAGGTTGTGGCCGGATG	153
eqOTX1	GCTCTTCGCCAAGACTCGGT	ACTTCTTCTTGGCTGGGCGG	170
eqSATB2	GGAGTCGAGGCATCACCACA	GGCACAACTGTCTTCAGCCG	163
eqBRN2	GGCCCAGGAGATCACTTCCC	GAGGTGGCGTGTCCCTACTC	164
eqCUX1	TTACTGCCAAGGCCGACGAA	GAGGCCAGCTGGAGAGAGTG	139
eqPSD95/eqDLG4	AGCCGATTCGAGGCCAAGAT	TCAGGAAGCCGCAGTCCTTG	154
eqSYP	AACACCTCGGTGGTAAGCCC	CTGCCCATAGTCGCCCTGAG	135
eqVGLUT1/eqSLC17A7	GGCGAGAGCGTCAAGGTTCA	TTGGAGTGCGAGTAGCCGAC	107
eqGFAP	GATTGTAAATGGAGCCCCGCC	CCATCCCGCATCTCGACAGT	111

Fluo-4 AM calcium dye staining and imaging

Fluo-4 AM (F14201; Life Technologies) was reconstituted to a stock solution of 2 mM in DMSO. Loading solution was prepared by diluting Fluo-4 AM to a final concentration of 10 μM in Tyrode's solution (5 mM KCl, 129 mM NaCl, 1 mM MgCl₂, 2 mM CaCl₂, 30 mM glucose and 25 mM Hepes). Five week-old neuronal cultures (three each from both A and K clones) were incubated in loading solution for 40 min at 37°C followed by two washes with PBS. Cultures were then incubated in Tyrode's solution for 30 min at 37°C to allow deesterification of intracellular AM esters. Neuronal cultures were excited using a 488 nm filter and imaged using a Nikon DS-Qi2 camera mounted on an Olympus IX51 microscope. Acquired time-lapse images were processed using ImageJ to calculate mean intensities for specified regions of interest where spontaneous activity was observed. Normalized fluorescence (ΔF/F0) was determined by calculating the average of baseline fluorescence (F0) before a spontaneous calcium transient, followed by dividing the measured mean fluorescence intensities (ΔF) by the average baseline fluorescence (F0). These results were plotted as line plots with the normalized fluorescence on the y-axis and time in seconds on the x-axis.

Viruses and virus infection cultures

The following flaviviruses were assessed for replication in equiPSC-derived neurons: WNV strains NSW₂₀₁₁ [9,16], K68967 [16], and Gu0631 [16]; MVEV strain 51, BgV [14], and FRV [13]. EquiPSC-derived neurons were seeded into eight-well chamber slides (Nunc). Four chambers per slide were mock-infected, while the other four chambers each received 4 × 10³ infectious units of one of the flaviviruses listed. Due to the dense colony growth pattern of the cells, an exact multiplicity of infection (MOI) could not be determined. The chamber slides were incubated at 37°C in humidified 5% CO₂ air. At 24, 72, and 120 h postinfection (pi) one slide per virus was terminated. The culture supernatants were collected and stored at −80°C until tested (see Virus titration section), while the cells were immediately fixed in 4% neutral-buffered formalin for 15 min followed by air-drying. The slides were stored dry, in the dark at RT until subjected to immunocytochemistry.

For comparison, primary equine fibroblasts, established from the joint capsule of an adult stockhorse, and used in passage 6–7, were infected with WNV_NSW2011, MVEV, BgV, and FRV at a MOI of 2, 1, and 0.5 in duplicate wells of 96-well tissue culture plates. After 5 days, the cultures were examined visually for cytopathic effect (cpe), fixed in 20% acetone overnight at 4°C and air dried. The cells were subsequently immunolabeled with a flavivirus envelope protein-specific mouse monoclonal antibody (mAb) 4G2 [17] as described in the section below.

Immunocytochemistry for virus antigen

Detection of virus replication in cells was achieved by immunolabeling for the nonstructural protein-1 (NS1) using the flavivirus NS1-specific mAb 4G4 (all WNV strains and MVEV) or the flavivirus envelope protein-specific mAb 4G2 (BgV and FRV) using a protocol adapted from Prow et al. [18]. In brief, the formaldehyde-fixed cells were subjected to antigen retrieval by heating in EDTA, pH 9 (DAKO Antigen Retrieval solution) for 20 min at 96°C, followed by cooling to RT for 25 min. This was followed by three blocking steps with 0.3% H₂O₂ in water for 10 min, 0.15 M glycine in PBS for 10 min, and 30 min with DAKO antibody dilution buffer, with brief rinses with Tris-buffered saline with 0.1% Tween-20 (TBST) between steps. This was followed by 2 h incubation at RT with either blocking buffer (two infected and two uninfected wells) or with the mAb (neat hybridoma spent medium; two infected and two uninfected wells), followed by multiple TBST rinses over a 10–15 min period and detection of bound antibody using the DAKO Envision Kit and AEC substrate. The slides were then counterstained with Meyer's hematoxylin and mounted with DAKO Faramount medium. For each batch, a positive (tissue from a mouse experimentally infected with WNV or WNV infected C6/36 insect cells) and negative (uninfected mouse tissue or uninfected C6/36 cells) antigen control was included.

To investigate whether viral infection was specific to neurons or also occurred in neural progenitor cells (NPCs) and/or glial cells, we performed double-immunolabeling with the neuron-specific TUJ1 antibody and mAb 4G4 as described in [19]. In brief, the TUJ1 antibody (rabbit) was diluted directly in the 4G4 mAb, and the cells were incubated as described above. This was followed by sequential incubation with anti-mouse (DAKO Envision Kit with horseradish peroxidase conjugate) and anti-rabbit-IgG (Vector AB-kit with alkaline phosphatase conjugate) antibodies, followed by sequential incubation with the chromogens for the two enzymes. Controls were mouse brains as described above and wells that only received one of the primary antibodies or no primary antibody.

Virus titration

Infectious virus released into the culture supernatant was quantitated by a 50% tissue culture infective dose (TCID₅₀) assay as described [20]. In brief, spent medium from the equiPSC-derived neuron cultures was diluted in DMEM/2% FBS in a 10-fold dilution series with 8 wells of a 96-well plate per dilution. African green monkey (Vero) cells were used for titration of all WNV strains, MVEV and the mock controls, while baby hamster kidney-derived (BSR) cells were used for FRV titration. Supernatants from BgV-infected cultures were tested on C6/36 insect cells, as this virus does not replicate in Vero cells and only to a limited extent in BSR cells [14]. Cells were added at 2 × 10⁴ cells per well to a final volume of 200 μL per well and the culture plates incubated at 37°C in humidified 5% CO₂ air for 5 days. At the end of incubation, the wells were examined visually for cpe, and the cell monolayers were then fixed in 20% acetone with 0.1% BSA by overnight incubation at 4°C. This was followed by fixed-cell ELISA using the mAb 4G2 as described [17]. The final titre was calculated using the method of Reed and Muench [21].

Results

Rapid neurogenesis, cortical identity, and functionality

Within 6 days of being transferred into neuronal differentiation medium, equiPSC colonies (Fig. 1A) had formed neurospheres with numerous axonal outgrowths (Fig. 1B) that appeared to make connections with axons from adjacent neurospheres (Fig. 1C). Neurons stained positively for the neuronal marker TUJ1 (Fig. 1D, E). Neuronal cultures passaged to single cells still displayed residual clusters from the initial neurospheres; however, individual neurons could be seen to have formed connections between these clusters (Fig. 1F). Neurons also expressed the cell surface transmembrane ligand Ephrin B3, which plays a role in axon guidance (Fig. 1G).

FIG. 1.

Temporal progression of equiPSC neuronal differentiation. (A) Undifferentiated equiPSC colonies maintained on irradiated mouse embryonic fibroblasts. Magnification × 100. (B, C) Rapid neurogenesis is visible at day 12 postneuronal induction. (B) Magnification × 100; (C) Magnification × 200. (D) Immunostaining for neuron-specific tubulin TUJ1 (green); nuclei are visualized with Hoechst (blue). Magnification × 100. (E) Immunostaining for TUJ1 reveals extensive fasciculation within the neuronal networks. Magnification × 100. (F) Neuronal cultures passaged to single cells contain residual clusters from the initial neurospheres as seen in the top left and bottom right corners. Single-cell neurons are seen between these clusters forming connections. Magnification × 200. (G) Neurons show punctate immunostaining for the transmembrane receptor protein Ephrin B3 (green); nuclei are visualized with Hoechst (blue). Magnification × 630. equiPSC, equine induced pluripotent stem cell.

After 6 days in neuronal differentiation medium, cultures of clone A showed high expression levels of TBR2 (also known as EOMES) which, in human, is required for the transition from PAX6-expressing radial glia neural stem cells (NSCs) to intermediate neural progenitor cells (NPCs) [2,22] (Fig. 2A). In keeping with an increase in TBR2 expression, PAX6 was not expressed in clone A cultures (Fig. 2A). In contrast, clone K cultures retained low, but significant, levels of PAX6 expression, while TBR2 was not expressed (Fig. 2A). OTX1 expression is similarly associated with human NSCs [2,23] and significant levels of expression were seen in cultures of clone K, but not clone A (Fig. 2A). Expression of FOXG1 was not significantly upregulated in cultures of either clone A or K at day 6 (Fig. 2A). Thus, cultures from clone A, which express TBR2 but not PAX6 or OTX1, appear to contain a greater proportion of NPCs than cultures from clone K, which retain expression of the NSC markers PAX6 and OTX1 and do not express TBR2. Cultures from both clones expressed markers consistent with cortical neuron differentiation, including EMX1, CUX1, BRN2, SATB2, and TBR1 [2,22,23] (Fig. 2B). However, cultures from clone A, that appear to contain a greater proportion of NPCs than NSCs, consistently expressed higher levels of the cortical neuronal markers than the cultures from clone K (Fig. 2B). Interestingly, cultures from clone A also expressed the glial marker GFAP which, in the central nervous system, is expressed by astrocytes as well as radial glia NSCs [2,24] (Fig. 2C). Cultures from both clones A and K express the postsynaptic marker PSD95 (Fig. 2C); however, we did not detect significant levels of expression of SYNAPTOPHYSIN or VGLUT1/SLC17A7.

FIG. 2.

qPCR analysis of selected key markers for neural progenitors, cortical neurons, glia, and synapses. Neurons derived from both clones A and K were analyzed before neuronal induction (D0) and 6 days after neuronal induction (D6). Asterisks represent degrees of significance based on t-tests: ns = not significant (P > 0.05); *P < 0.05; **P < 0.01; ***P < 0.001. (A) Cultures of clone A showed high expression levels of TBR2, and no expression of PAX6, indicating a transition from PAX6-expressing NSCs to NPCs. In contrast, clone K cultures retained low, but significant, levels of PAX6 expression, while TBR2 was not detected. Expression of the NSC marker OTX1 was similarly seen in cultures of clone K, but not clone A. (B) Cultures from both clones expressed markers consistent with cortical neuron differentiation. (C) Cultures from clone A also expressed the glial marker GFAP, which is expressed by astrocytes and radial glia NSCs. Cultures from both clones express the postsynaptic marker PSD95; however, we did not detect significant levels of expression of SYNAPTOPHYSIN or VGLUT1/SLC17A7. NPCs, neural progenitor cells; NSCs, neural stem cells; qPCR, quantitative polymerase chain reaction.

At 2 weeks postneuronal induction, the majority of cells persisted as NPCs as shown by positive immunostaining for SOX2 within the residual neurosphere clusters and in the majority of the dissociated single cells (Fig. 3A). However, in cultures colabeled with the NPC marker TBR2 and the neuronal marker TUJ1, a proportion of cells were positive for both markers, indicating early neuronal differentiation. (Fig. 3B). Populations of developing neurons (TBR2⁺/TUJ1⁺) were observed adjacent to clusters of NPCs (TBR2⁺/TUJ1⁻) (Fig. 3B), demonstrating that neuronal differentiation is asynchronous, as has been described for neuronal differentiation of human pluripotent stem cells [25].

FIG. 3.

Neuronal differentiation at 14 days postinduction. (A) The majority of cells are NPCs as indicated by SOX2 immunostaining (red) within the residual neurospheres and in the dissociated single cells. Magnification × 200; scale bar 500 μm. (B) Neuronal differentiation is asynchronous as indicated by two clusters of neuronal cells, both of which contain cells that are positive for the NPC marker TBR2 (green), while only one of the clusters contains TUJ1-positive neurons (red). Magnification × 200; scale bar 500 μm. (C) Immunostaining for BRN2 identifies excitatory cortical neurons (green). Magnification × 100; scale bar 250 μm. (D) Neural rosettes (dashed white circle) contain BRN2-positive cortical neurons (green). Magnification ×200; scale bar 500 μm.

Positive immunostaining for BRN2 confirmed that the mature neurons were excitatory cortical neurons (Fig. 3C). Neural rosettes, the in vitro equivalent of neural tube formations that are commonly observed in human and mouse neuronal differentiation cultures [1,2,23,26], were observed (Fig. 3D). Neural rosettes from mouse and human pluripotent stem cells consist of NPCs that give rise to cortical neurons [1,2,23]. Similarly, neural rosettes derived from equiPSCs consisted of cells positive for BRN2 (Fig. 3D). Thus, while the majority of the cells were still NPCs, expression of the mature neuronal marker BRN2 reflects that a proportion of the cells were differentiating toward a terminal cortical fate, which is in agreement with the qPCR data.

Neuronal cultures for both clones A and K were maintained for an additional 3 weeks (total of 5 weeks of neuronal differentiation) and subsequently loaded with Fluo-4 AM calcium dye to assess neuronal functionality. Spontaneous calcium transients were observed in neurons derived from both clones A and K (Fig. 4A, B). In instances where connections between adjacent neurons were observed, calcium transients were propagated unidirectionally from one neuron to an adjacent neuron, presumably via synapses (Fig. 4A).

FIG. 4.

Spontaneous neuronal activity in equiPSC-derived neurons. (A) Selected neurons from a clone A culture loaded with Fluo-4 AM (shown in false color fire LUT—ImageJ). (a) Baseline fluorescence before a spontaneous calcium transient. (b) Spontaneous initiation of a calcium transient is observed in one of the three neurons shown (arrowhead). (c) The calcium transient is propagated from the first neuron (first arrowhead) to the soma of the second neuron (second arrowhead). (d) The calcium transient continues into the soma of the adjacent, third, neuron (arrowhead). (B) Representative calcium transients measured in neurons from clones A and K. Graphs are presented as normalized fluorescence to baseline fluorescence (ΔF/F0).

Viral infection

All three strains of WNV and MVEV grew to high titres in neuronal cells derived from both equiPSC clones A and K (Table 2). In the case of the WNV strains, this was also reflected in increasing numbers of NS1-positive cells over the 120 h of incubation (Fig. 5), all of which coexpressed the neuronal marker TUJ1, confirming that WNV infection was specific to neurons (Fig. 6). There was less concordance between virus titres and ICC results for MVEV, but this may be due to cell loss either through apoptosis/necrosis or cell detachment. In contrast, FRV showed restricted replication in the neuronal cells; that is, while virus NS1 could be detected by ICC in a small number of cells, indicating replication, the virus titre did not increase beyond input virus, but rather decreased in accordance with thermal inactivation (Table 2). This is in line with the finding that although FRV infects horses, as judged by seroepidemiological studies, the virus is not known to cause disease in this, or any other, vertebrate species [13]. Finally, there was no detectable replication of BgV in the neuronal cells, which was expected considering its very restricted growth in vertebrate cells in vitro, and its suspected mosquito–marsupial transmission cycle [14].

FIG. 5.

Examples of WNV replication in equiPSC-derived neurons demonstrated by immunocytochemistry for the flavivirus NS1. (A) Antibody isotype control. (B) WNV_NSW2011 replication in the cells at 24 h postinfection (pi). Only a single positive cell (red staining) is seen in this culture (arrow). (C) At 72 h pi with WNV_NSW2011 many cells have detached and of those remaining, >40% are expressing NS1 (red staining). (D) 4G4 labeling of mock-infected culture, demonstrating lack of nonspecific labeling despite the dense colony growth. (E) WNV_K68967 replication (red staining) at 72 h pi. (F) At 120 h pi, only scattered cells remain in the cultures, of which the majority are positive for flavivirus NS1 (red staining). NS1, nonstructural protein-1; WNV, West Nile virus.

FIG. 6.

WNV exclusively infects neurons. Neuronal cultures were immunolabeled for neuronal differentiation (anti-TUJ1; black signal) and infection with WNV (flavivirus-NS1-specific mAb 4G4; red signal). The cells were lightly counterstained with dilute Meyer's hematoxylin. (A, B) All cells that are positive for WNV are also positive for TUJ1 indicating that WNV infects only neurons and not NSCs or NPCs. Magnification × 200. (C) Positive control: WNV-infected neuronal culture immunolabeled with mAb 4G4. Magnification × 200. (D) Positive control: WNV-infected neuronal culture labeled with anti-TUJ1. Magnification × 200. (E) Negative control: noninfected neuronal culture labeled with mAb 4G4. Magnification × 200. (F) Negative control: WNV-infected neuronal culture incubated without either primary antibody. Magnification ×200. mAb, monoclonal antibody.

Table 2.

Flavivirus Replication in Equine Induced Pluripotent Stem Cell-Derived Neuronal Cells

		ICC read-out (semiquantitative ^a )			Virus titration (TCID₅₀/mL)
Clone	Virus	24 h	72 h	120 h	24 h	72 h	120 h
A	Mock	0	0	0	0	0	0
	WNV_NSW2011	0	++	+++	8.89E+00	2.11E+07	1.58E+08
	WNV_K68967	+	+++	+++^b	1.19E+06	1.19E+08	2.81E+06
	WNV_Gu0631	+	+++	+++^b	1.19E+06	1.19E+09	3.75E+07
	MVEV₅₁	0	0	+	1.58e+00	3.75E+05	2.11E+06
	BGV	0	0	0	2.81E+00	0	0
	FRV	0	+	+	2.81E+00	0	0
K	Mock	0	0	0	0	0	0
	WNV_NSW2011	+	+/++	+++	6.67E+00	2.11E+07	3.37E+06
	WNV_K68967	+/++	+++	+++	6.67E+05	2.11E+08	6.67E+06
	WNV_Gu0631	+/++	++/+++	+++	3.75E+05	3.75E+08	2.81E+07
	MVEV₅₁	0	+	+/++	2.11E+00	2.81E+00	3.75E+06
	BGV	ND	ND	ND	ND	ND	ND
	FRV	0	+	+	ND	ND	ND

On a scale of: 0 = no NS1/E signal; + = a few NS1/E-positive cells; ++ = moderate numbers of NS1/E positive cells; +++ = large numbers of NS1/E positive cells.

No colonies left, only scattered cells present, but all NS1-positive.

FRV, Fitzroy River virus; MVEV, Murray Valley encephalitis virus; ND, not done; NS1, nonstructural protein-1; TCID, tissue culture infective dose; WNV, West Nile virus.

In contrast to the high replication rates of WNV and MVEV in the equiPSC-derived neurons (Table 2) and in undifferentiated equiPSCs (data not shown), WNV and MVEV barely grew to detectable titres above input virus and no cpe was detected in the primary equine fibroblasts. There was no detectable replication of FRV and BgV in these cells (data not shown).

Discussion

Although horses are susceptible to neurotropic viruses that can infect humans, such as WNV and MVEV, there are currently no in vitro models enabling the study of virus replication or transmission. To address this deficit, we took advantage of equiPSCs previously generated in our laboratory [6] and subjected these cells to a dual SMAD-inhibition neuronal differentiation protocol widely used for human pluripotent stem cells [1]. The equiPSCs engaged in much more rapid neuronal differentiation than that observed in human pluripotent stem cells, but was similar to that observed by Sharma et al., who generated motor neurons from equine iPSCs [5]. Six days after transitioning into neuronal differentiation medium, cells expressed a range of markers associated in humans with excitatory neurons in the upper cortical layer, including EMX1, CUX1, BRN2, SATB2, and TBR1. After 2 weeks, the equiPSCs had formed extensive TUJ1-positive, fasciculated neuronal networks. Expression of BRN2 within neural rosettes and by individual neurons confirmed the differentiation of excitatory cortical neurons. After 5 weeks of neuronal differentiation, neurons displayed spontaneous calcium transients and, where connections between adjacent neurons were observed, calcium transients were propagated unidirectionally, presumably from pre- to postsynaptic neurons.

To explore the suitability of the equiPSC-derived neurons as a platform for the study of viral neuropathic diseases in the horse, we examined their susceptibility to infection with flaviviruses that are known to be neurotropic in horses, including WNV and MVEV. We observed that the three strains of WNV and MVEV grew to high titres and caused neuronal cell death, suggesting that equiPSC-derived neurons permit ready replication of these viruses. Significantly, FRV and BgV, that do not cause pathology in horses, did not replicate in the equiPSC-derived neurons. In contrast, WNV and MVEV grew poorly in primary equine fibroblasts. These results are in concordance with in vivo findings of target cell tropism: the flaviviruses target neurons, leukocytes, pericytes, and epithelial cells, but flavivirus replication has never been described in fibroblasts, regardless of virus strain or host species studied [12,18,20,27,28].

While the route(s) of neuroinvasion of the neurotropic flaviviruses remains elusive [27] and may depend on factors such as virus strain, age of host and host-species, as well as the route of virus challenge in the case of experimental infections, it appears that many different types of neurons can become infected with WNV and MVEV, including cortical neurons, and it may to some extent be a stochastic process [27]. At the time that horses show clinical signs of flavivirus encephalitis, the virus has all but ceased to replicate in the brain, as judged by both immunohistochemistry and reverse transcription (RT)-PCR [28]; however, as judged by the distribution of inflammatory leukocyte infiltrates, all parts of the brain can be involved [12,28]. Thus, the sensitivity of our equiPSC-derived neurons, which display a cortical identity, to both WNV and MVEV, appears to confirm that they are eminently suitable as an in vitro model to study neuron–virus interactions.

Considering the economic impact of the equine industry in countries such as the United States and Australia, totaling $112 billion and $8 billion, respectively, the availability of cell-based therapeutics and assays is surprisingly limited. In this study, we show that equiPSCs rapidly generate cortical neurons that permit replication of WNV and MVEV and that they also express Ephrin B3, the receptor via which henipaviruses enter mammalian neurons [29]. Given the significant impact that both flavivirus and henipavirus infections can have on equine and, in some instances, human populations, our study represents a significant advance in the establishment of in vitro horse neuronal models that can be used to study, and treat or prevent, neurotropic viral infections in horses.

Footnotes

Acknowledgments

The authors thank Cheryl Johansen for providing FRV, Caitlin O'Brien for growing stock of FRV, Agathe Colmant for BgV stock, Jessica Harrison for stock of MVEV₅₁, and Roy Hall for providing mAbs 4G2 and 4G4. This study was supported by an Alister Rogers Hendra Foundation grant (D.W.) and the Australian Research Council (ARC-LP120100686; H.B.O.).

Author Disclosure Statement

No competing financial interests exist.

References

Chambers

, Fasano

, Papapetrou

, Tomishima

, Sadelain

and Studer

. (2009). Highly efficient neural conversion of human ES and iPS cells by dual inhibition of SMAD signaling. Nat Biotechnol, 27:275–280.

Shi

, Kirwan

, Smith

, Robinson

and Livesey

. (2012). Human cerebral cortex development from pluripotent stem cells to functional excitatory synapses. Nat Neurosci, 15:477–486.

, Hu

, Jones

, Zhang

, LaVaute

, Du

and Zhang

. (2008). Directed differentiation of ventral spinal progenitors and motor neurons from human embryonic stem cells by small molecules. Stem Cells, 26:886–893.

and Zhang

. (2010). Directed differentiation of neural stem cells and subtype-specific neurons from hESCs. In: Cellular Programming and Reprogramming: Methods and Protocols. Ding

, ed. Humana Press, New York City, pp 123–137.

Sharma

, Livesey

, Wyllie

DJA

, Proudfoot

, Whitelaw

CBA

, Hay

and Donadeu

. (2014). Generation of functional neurons from feeder-free, keratinocyte-derived equine induced pluripotent stem cells. Stem Cells Dev, 23:1524–1534.

Whitworth

, Ovchinnikov

, Sun

, Fortuna

and Wolvetang

. (2014). Generation and characterization of leukemia inhibitory factor-dependent equine induced pluripotent stem cells from adult dermal fibroblasts. Stem Cells Dev, 23:1515–1523.

Mackenzie

and Williams

. (2009). The zoonotic flaviviruses of southern, south-eastern and eastern Asia, and Australasia: the potential for emergent viruses. Zoonoses Public Health, 56:338–356.

Petersen

, Carson

, Biggerstaff

, Custer

, Borchardt

and Busch MP

. (2013). Estimated cumulative incidence of West Nile virus infection in US adults, 1999–2010. Epidemiol Infect, 141:591–595.

Frost

, Zhang

, Edmonds

, Prow

, Gu

, Davis

, Hornitzky

, Arzey

, Finlaison

, et al. (2012). Characterization of virulent West Nile virus Kunjin strain, Australia, 2011. Emerg Infect Dis, 18:792–800.

10.

Roche

, Wicks

, Garner

, East

, Paskin

, Moloney

, Carr

and Kirkland

. (2013). Descriptive overview of the 2011 epidemic of arboviral disease in horses in Australia. Aust Vet J, 91:5–13.

11.

Burrow

, Whelan

, Kilburn

, Fisher

, Currie

and Smith

. (1998). Australian encephalitis in the Northern Territory: clinical and epidemiological features, 1987–1996. Aust N Z J Med, 28:590–596.

12.

Barton

, Prow

, Hall

, Kidd

and Bielefeldt-Ohmann

. (2015). A case of Murray Valley encephalitis in a 2-year-old Australian Stock Horse in south-east Queensland. Aust Vet J, 93:53–57.

13.

Johansen

, Williams

, Melville

, Nicholson

, Hall

, Bielefeldt-Ohmann

, Prow

, Chidlow

, Wong

, et al. (2017). Characterization of Fitzroy River virus and serologic evidence of human and animal infection. Emerg Infect Dis, 23:1289–1299.

14.

Colmant

, Bielefeldt-Ohmann

, Hobson-Peters

, Suen

, O'Brien

, van den Hurk

and Hall

. (2016). A newly discovered flavivirus in the yellow fever virus group displays restricted replication in vertebrates. J Gen Virol, 97:1087–1093.

15.

Livak

and Schmittgen

. (2001). Analysis of relative gene expression data using real-time quantitative PCR and the 2−ΔΔCT method. Methods, 25:402–408.

16.

Prow

, Edmonds

, Williams

, Setoh

, Bielefeldt-Ohmann

, Suen

, Hobson-Peters

, van den Hurk

, Pyke

, et al. (2016). Virulence and evolution of West Nile Virus, Australia, 1960–2012. Emerg Infect Dis, 22:1353–1362.

17.

Gentry

, Henchal

, McCown

, Brandt

and Dalrymple

. (1982). Identification of distinct antigenic determinants on dengue-2 virus using monoclonal antibodies. Amer J Trop Med Hyg, 31:548–555.

18.

Prow

, Setoh

, Biron

, Sester

, Kim

, Hobson-Peters

, Hall

and Bielefeldt-Ohmann

. (2014). The WNV-like flavivirus Koutango is highly virulent in mice due to delayed viral clearance and the induction of a poor neutralizing antibody response. J Virol, 88:9947–9962.

19.

Miura

, Travanty

, Oko

, Bielefeldt-Ohmann

, Weiss

, Beauchemin

, Holmes

. (2008). The spike glycoprotein of murine coronavirus MHV-JHM mediates receptor-independent infection and spread in the central nervous systems of Ceacam1a−/− Mice. J Virol, 82:755–763.

20.

Suen

, Uddin

, Wang

, Brown

, Adney

, Broad

, Prow

, Bowen

, Hall

and Bielefeldt-Ohmann

. (2015). Experimental West Nile Virus infection in rabbits: an alternative model for studying induction of disease and virus control. Pathogens, 4:529–558.

21.

Reed

and Muench

. (1938). A simple method of estimating fifty per cent endpoints. Am J Epidemiol, 27:493–497.

22.

Englund

, Fink

, Lau

, Pham

, Daza

, Bulfone

, Kowalczyk

and Hevner

. (2005). Pax6, Tbr2, and Tbr1 are expressed sequentially by radial glia, intermediate progenitor cells, and postmitotic neurons in developing neocortex. J Neurosci, 25:247–251.

23.

Molyneaux

, Arlotta

, Menezes

and Macklis

. (2007). Neuronal subtype specification in the cerebral cortex. Nat Rev Neurosci, 8:427–437.

24.

Sofroniew

and Vinters

. (2010). Astrocytes: biology and pathology. Acta Neuropathol, 119:7–35.

25.

Falk

, Koch

, Kesavan

, Takashima

, Ladewig

, Alexander

, Wiskow

, Tailor

, Trotter

, Pollard

and Smith

. (2012). Capture of neuroepithelial-like stem cells from pluripotent stem cells provides a versatile system for in vitro production of human neurons. PLoS One, 7:e29597.

26.

Bian

, Zheng

, Li

, Luo

and Ding

. (2016). Sequential differentiation of embryonic stem cells into neural epithelial-like stem cells and oligodendrocyte progenitor cells. PLoS One, 11:e0155227.

27.

Suen

, Prow

, Hall

and Bielefeldt-Ohmann

. (2014). Mechanism of West Nile virus neuroinvasion: a critical appraisal. Viruses, 6:2796–2825.

28.

Bielefeldt-Ohmann

, Bosco-Lauth

, Hartwig

, Uddin

, Barcelon

, Suen

, Wang

, Hall

and Bowen

. (2017). Characterization of non-lethal West Nile Virus (WNV) infection in horses: subclinical pathology and innate immune response. Microb Pathog, 103:71–79.

29.

Zeltina

, Bowden

and Lee

. 2016. Emerging paramyxoviruses: receptor tropism and zoonotic potential. PLoS Pathog, 12:e1005390.