A Mouse Model for Infection with Enterovirus A71 in Small Extracellular Vesicles

EV-A71 was supposed to infect the CNS through the neural pathway and the circulation of the blood in previous studies. Reverse axon transport had been confirmed as an important pathway for EV-A71 to infect the CNS; however, it is still unknown how EV-A71 infects the CNS through the circulation of the blood. Combined with the infectivity of sEVs secreted from EV-A71-infected cells and the characteristic that sEVs could cross the blood-brain barrier, we considered that sEVs may play a vital role in EV-A71 pathogenesis of the CNS.

Small extracellular vesicles (sEVs) are Ͻ200-nm particles which are released from the cells and delimited by a lipid bilayer. Exosomes are just parts of sEVs which predominantly originate from the endosomal system and have characteristic tetraspanin membrane markers (CD9, CD63, CD81).
The sEVs are now considered new extracellular functional carriers which could transport protein, nucleic acids, and lipids, etc., to neighboring cells. They also play an important role in viral spread (4,5). Viruses could transmit the infection and evade the immune system by EVs cloaking viral proteins and viral genomes (6), such as poliovirus (7), hepatitis A virus (8), and coxsackievirus B (9), and so on. The sEV-mediated extracellular communication may play an important role in viral pathogenesis and control of host immune response to infection (10).
In consideration that sEVs could transfer their cargo across the blood-brain barrier (11), it may be related to EV-A71 infection in the CNS. We separated the sEVs from EV-A71 clinical isolate-infected cells by using differential ultracentrifugation with further isopycnic gradient centrifugation and identified the sEVs by transmission electron microscopy (TEM), nanoparticle tracking analysis (NTA), and Western blotting (WB). These sEVs containing viral RNA are infectious in vitro and capable of infecting neonatal immunocompetent mice. Compared with the EV-A71 clinical isolate, these infectious sEVs showed greater neurovirulence and more fatality in the mouse model with intraperitoneal injection. This present model could be generated easily by isolated sEVs and is useful for studying the pathogenesis of enterovirus-associated diseases.
Generation of sEVs from EV-A71-infected cells. In order to separate sEVs with high specificity, we used differential ultracentrifugation with further isopycnic gradient centrifugation to isolate sEVs from the cells infected by the clinical EV-A71 isolate. The sEVs floating at 1.09 to 1.14 g cm Ϫ3 were collected and identified. Transmission electron microscopy (TEM) images showed mostly empty sEVs (Fig. 1A) and a few sEVs containing virus-like particles which were 27-to 30-nm compact particles (Fig. 1B). Nanoparticle tracking analysis (NTA) showed sEVs concentrated on the size of 118.1 nm (Fig. 1C). The positive markers and negative markers of EVs were verified by immunoblotting, and the viral capsid protein VP1 was in the sEVs (Fig. 1D). qRT-PCR analysis of sEVs showed that sEVs contained EV-A71 RNA (Fig. 1E).
To study the permeability of sEVs across the blood-brain barrier (BBB), we established the in vitro static BBB model by coculturing brain microvascular endothelial cells (BMEC) and astrocytes from neonatal mice in the transwell (Fig. 1F). To compare permeabilities between EV-A71 and sEVs, EV-A71 and sEVs with the same quantity of EV-A71 RNA were added in the BMEC of the upper side of the BBB model, respectively, and then the level of EV-A71 RNA of the bottom side was measured by qRT-PCR after 3 h. The ratio of upper/bottom viral RNA quantity was calculated as a measure of permeability. Obviously, sEVs crossed the BBB more efficiently than EV-A71 (Fig. 1G). It suggested sEVs may play an important role in CNS infection.
Infection of neonatal mouse with the sEV derived from clinical EV-A71 isolate. To explore whether sEVs released from EV-A71 clinical isolate-infected cells could effectively infect the neonatal mice, we respectively inoculated sEVs and the EV-A71 clinical isolate at a dose of 1 ϫ 10 4 PFU/mouse by intraperitoneal injection. Compared with intracranial injection, intraperitoneal injection could avoid the influence of challenge methods on the brain and simulate the pathogenesis of EV-A71 factually.
At 3 days postinfection (dpi), the mice infected with sEVs lost weight obviously whereas the mice infected with EV-A71 did not lose weight apparently ( Fig. 2A and B). Within 7 days, the mortality rate of mice infected with sEVs was 86.7% whereas the mortality rate of mice infected with EV-A71 was only 46.7% (Fig. 2C). To validate viral replication in tissues of infected mice, we collected the main infected tissues, the brain and gut, after the mice died by infection or euthanasia at 7 dpi and quantified the viral RNA. The viral RNA levels in brain and gut tissues in the sEV-infected group were significantly higher than those in the EV-A71-infected group (Fig. 2D). At the same time, histological examination was performed on the brain; consistent with the previous results, numerous histopathological changes were observed in the sEV-infected group: some nerve cells had fatty or vesicular degeneration and the number of inflammatory cells was increased obviously. However, the EV-A71-infected group exhibited mild neuropil vacuolation and neuronal loss in brainstem reticular formation (Fig. 2E).
These results suggested the neonatal mice were more susceptible to the sEVs than to the clinical isolate and that an effective animal model could be easily generated by the sEVs without adaptive mutations.
Since the first report that EV-A71 could infect newborn mice by intraperitoneal injection and cause paralysis, death, etc., in 1979, the EV-A71-infected mouse model has been optimized several times (12). Some studies investigated potential influence factors such as virus dose, challenge method, mouse strain, and mouse age. but the results showed that they cannot help to increase the virulence of EV-A71 (13).
The majority of studies tried to adapt EV-A71 in mice or mouse embryonic fibroblast NIH/3T3 cells because EV-A71 which was isolated from human samples failed to infect mice efficiently. The results showed this was an effective method; however, virus replication may be affected by the mouse immune system during this long process, and the disease in EV-A71-infected mice was readily attenuated by anti-EV-A71 antibody (2).
In our study, we increased EV-A71 susceptibility in mice by the sEVs released from EV-A71-infected NIH/3T3 cells which originated from mice. According to guidelines (14), sEVs were separated from the culture of EV-A71-infected cells by ultracentrifugation and isopycnic gradient centrifugation with high specificity. We inoculated EV-A71 in mice by intraperitoneal injection and monitored their clinical symptoms. The results showed sEVs could cause a higher rate of death and more severe CNS infection in neonatal mice, and the level of EV-A71 RNA replication in brain and gut of mice was higher than for EV-A71 virus. EV-A71 could only attach or/and enter the cells via binding EV-A71 receptors on the surface of the infected cells, while the sEVs could transport the cargo without any receptor. In consideration that sEVs could cross the blood-brain barrier by transcytosis (15), it is possible that the property of the infectious sEVs is helpful to infect the neonatal mice successfully. Absolute quantification of EV-A71 RNA in gut and brain was performed by qRT-PCR. The virus RNA copies for various tissues of each infected mouse were shown as mean for three independent experiments. The viral RNA loads of different treatment groups were compared. Statistical analyses were performed using Mann-Whitney U test. (E) Representative H&E images of brain damage. The EV-A71-infected group exhibited mild neuropil vacuolation and neuronal loss (arrows), and the sEV-infected group exhibited numerous inflammatory cells and vesicular degeneration (arrows). A representative result is shown. In all panels: *, P Յ 0.05; **, P Յ 0.01; ***, P Յ 0.001; ****, P Յ 0.0001. NC, negative control.
In summary, we have developed a stable mouse model with sEVs released from EV-A71 clinical isolate-infected cells. This new model is easier to establish than before and represents a powerful tool for investigating EV-A71 neurologic pathogenesis.
Cells and viruses. NIH/3T3 cells were cultured in Dulbecco modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS). The cell lines were obtained from ATCC. Cells were infected with EV-A71 at a multiplicity of infection (MOI) of 1 and incubated for 1 h at 37°C and 5% CO 2 . Then, cells were rinsed and further incubated with prewarmed DMEM supplemented with 5% FBS which had been ultracentrifuged to remove EVs. The cells were collected, cracked by three freeze-thawing cycles, and centrifuged at 4,000 rpm at 4°C for 10 min to remove cellular debris. Infectious virus titers were determined by 50% tissue culture infective dose (TCID 50 ) assay as described elsewhere (16).
Ethics statement. All animal procedures were approved by the ethics committee of Jiangsu University Animal Care Committee, followed the Guide for the Care and Use of Laboratory Animals published by the Chinese National Institutes of Health, and were performed in compliance with the animal behavioral guidelines, using approved protocols from the institutional animal care committee. Establishment of BBB model. Astrocytes were isolated from neonatal mouse, purified by cell passage, and then cocultured with BMEC. A schematic drawing of the in vitro BBB model is shown in Fig. 1F.
Animal experiments. For all experiments, 1-day-old ICR mice were purchased from the Laboratory Animal Centre of Jiangsu University. They were housed together in an environment of 50% humidity at 22°C under a 12-h light/dark cycle and kept with their mothers to provide food (17). Two groups of 1-day-old ICR mice were inoculated with EV-A71 and sEVs (1 ϫ 10 4 TCID 50 , 20 l per mouse), respectively, by intraperitoneal injection. One group was administered PBS (20 l per mice) as a control. The mice were monitored for 5 days for clinical symptoms and were sacrificed if the mice met the following criteria: lethargy, hind limb paralysis, hypothermia of Ͻ33.5°C, or death. Brain tissues were collected and fixed in 4% paraformaldehyde. After being embedded in paraffin and sliced, the brain sections (10 m) were stained with hematoxylin and eosin (H&E) for morphological examination. Ten sections of brain were observed per mouse in a blind manner.
qRT-PCR assay for EV-A71 RNA. Total RNA of the cells and the tissues was extracted with TRIzol reagent (Invitrogen); RNA from culture supernatants was extracted with the RNAqueous-Micro kit (Thermo Fisher Scientific). EV-A71 was determined in a SYBR green two-step qRT-PCR assay (Bio-Rad). The primer sequencing of EV-A71 was performed as described previously (18). sEV separation. Cell culture supernatant fluids were centrifuged at 1,000 ϫ g at 4°C for 10 min to remove cells and debris and then centrifuged at 10,000 ϫ g at 4°C for 30 min twice to remove subcellular fraction, and the sEVs were collected after ultracentrifugation at 100,000 ϫ g for 1 h at 4°C. The sEVs were resuspended in PBS, loaded onto an 8 to 40% iodixanol (Opti-prep) step gradient, and centrifuged at 150,000 ϫ g in an SW32TI Beckman Coulter rotor for 24 h at 4°C. One-milliliter fractions were collected, and density was determined with a refractometer (19)(20)(21).
TEM. Samples of 2.5 l were adsorbed on the surface of a glow-discharged 400 mesh carbon-coated copper grid for 5 min, and the grid was fixed in 1% glutaraldehyde with 0.15 M phosphate buffer (pH 7.4) for 1 min; afterward, it was rinsed with deionized water and stained with 3% ammonium molybdate, pH 7.0.
NTA. The size and concentration of sEVs derived from EV-A71-infected cells were analyzed using the ZetaView PMX 110 (Particle Metrix, Meerbusch, Germany) and configured with a scientific complementary metal oxide semiconductor (CMOS) camera and blue 488-nm laser. For ZetaView analysis, sEVs were diluted in 2 ml PBS, collected, and analyzed by the ZetaView 8.04.02 SP2.
Immunoblotting. Cells were lysed with radioimmunoprecipitation assay (RIPA) buffer (Kangwei Century). Immunoblot assays were performed with standard proce-dures and the indicated antibodies. Protein bands were detected by an Image Quant LAS 4000 mini (GE Healthcare).
Statistical analysis. Data were analyzed by GraphPad Prism software (version 6.01). A P value of Ͻ0.05 was considered significant.