Autophagy is induced and supports virus replication in Enterovirus A71-infected human primary neuronal cells

Enterovirus A71 (EV-A71), which belongs to the family Picornaviridae, can invade the central nervous system (CNS) and cause severe CNS complications or death. The EV-A71 antigen has been detected in the neurons in the brains of humans who died from EV-A71 infection. However, the effect of EV-A71 infection on human neuronal cells remains poorly understood. Human neural stem cells (NSCs) and IMR-32 neuroblastoma cells were differentiated into neuronal cells for this study. Although the neuronal cells were permissive to EV-A71 infection, EV-A71 infection did not induce an obvious cytopathic effect on the neuronal cells. EV-A71 infection did not induce apoptosis in neuronal cells. However, autophagy and autophagic flux were induced in EV-A71-infected neuronal cells. The production of autophagosomes was shown to be important for EV-A71 viral RNA (vRNA) replication in neuronal cells.


Results
The characteristics of human NSCs and neuroblastoma IMR-32-derived neuronal cells. To understand the effect of EV-A71 infection on neuronal cells, human NSCs were used to perform neuronal differentiation. The human NSCs used in this study were derived from NIH-approved H9 human embryonic stem cells and can differentiate into neurons. The human NSCs were maintained in conditional culture medium without the growth factors EGF and FGF-β for 7 days for neuronal differentiation (Fig. 1A). The neurons formed from human NSC differentiation exhibited long neurites (Fig. 1B). Immunofluorescence assay (IFA) was performed to confirm the differentiation efficiency: neuron-specific class III β-tubulin (Tuj1), which is expressed primarily in the early stage of neuronal differentiation 29 , and microtubule-associated protein 2 (MAP2), which is expressed mainly in neurons, were used as specific markers of neuronal differentiation 30 . Neurons derived from human NSCs expressed MAP2 and Tuj1 (Fig. 1C). In addition, the neurons derived from human NSCs also expressed GAD67, which is a specific marker of GABAergic neurons (Fig. 1C). We also used the other cells to perform neuronal differentiation. The protocol for differentiated neuronal cells from neuroblastoma IMR-32 cells is shown in Fig. 1D, and these neuronal cells also exhibited long neurites (Fig. 1E). The neuronal-specific markers Tuj1 and MAP2 and the GABAergic neuronal marker GAD67 were detected in differentiated IMR-32 cells (Fig. 1F). We also stained neuronal-specific markers in undifferentiated human NSCs and IMR-32 to confirm that these neuronal markers were only expressed in differentiated neuronal cells (Figure S1A andS1B).

Neuronal cells are permissive to EV-A71 infection.
To study the susceptibility of human neuronal cells to EV-A71, human NSC-derived neurons were examined for EV-A71 infection. At 12, 24 and 48 h postinfection, the typical cytopathic effect (CPE) of EV-A71 infection was not observed. Few neurons had detached from the culture plates ( Fig. 2A). The survival rate of mock-or EV-A71-infected neurons was quantified by a trypan blue assay, which showed no significant difference in cell survival between mock-and EV-A71-infected cells (Fig. 2B). To determine whether EV-A71 could infect neurons, an IFA was performed to evaluate the expression of EV-A71 3D pol in EV-A71-infected neurons. The IFA results revealed the expression of EV-A71 3D pol in MAP2 positive neurons over the course of EV-A71 infection (Fig. 2C). The EV-A71 3A protein was also detected in GAD67-positive neurons (Fig. 2D). EV-A71 replication was also demonstrated by vRNA and viral protein expression and by the viral growth curve. Total RNA was extracted from the neurons infected with EV-A71, and the expression of EV-A71 vRNA was evaluated with RT-qPCR. The relative level of EV-A71 5′ UTR expression increased significantly with infection time (Fig. 2E). Proteins were extracted from EV-A71-infected neurons harvested at various time points, and viral protein expression was measured using western blotting. The expression of the EV-A71 3D pol protein was first detected at 8 h postinfection and was further increased at 16 h postinfection (Fig. 2F). The total lysates and supernatants of the EV-A71-infected neurons were harvested at various time points, and a plaque assay was performed to measure the production of progeny virus. The virus of total lysate (supernatant and cells) increased significantly from 8 to 48 h postinfection. The viral titer of the supernatant also increased with increasing viral infection time (Fig. 2G). The findings indicated that the lack of a CPE was not due to the inability of EV-A71 to enter and replicate in neuronal cells.
Caspase 3, caspase 8 and caspase 9 are not activated in EV-A71-infected neuronal cells. Our results revealed that EV-A71 infection did not cause an obvious CPE in neurons derived from human NSCs. To test whether EV-A71 could induce apoptosis in neuronal cells, neurons derived from human NSCs were infected with EV-A71 at an MOI of 2. The neurons were then stained with antibodies against active caspase 3 and Tuj1 at 12, 24 and 48 h postinfection. The IFA results revealed that only a few Tuj1-positive cells also expressed active caspase 3. The expression of active caspase 3 was not increased over the course of EV-A71 infection (Fig. 3A). Western blot assays were performed to measure the expression of the precleaved and active forms of caspase 3. The protein level of active caspase 3 in EV-A71-infected neurons did not increase with infection time (Fig. 3B). Similar results were observed in differentiated IMR-32 cells (Fig. 3C). The expression of active caspase 8 and caspase 9 was also not induced in the EV-A71-infected differentiated IMR-32 cells (Fig. 3D). The inhibition of caspase 3 inhibits EV-A71 replication in rhabdomyosarcoma (RD) cells 31 . In our previous report, treatment with the caspase 3 inhibitor Z-VAD decreased EV-A71 replication in RD cells 32  www.nature.com/scientificreports/ 3 inhibitor Z-VAD did not inhibit viral growth in differentiated IMR-32 cells ( Figure S2). EV-A71 infection did not induce apoptosis in the neuronal cells.

EV-A71 induces autophagosome formation in neuronal cells and differentiated IMR-32 cells.
Autophagy is important in the regulation of neuronal physiology 33 . EV-A71 has been reported to induce autophagy in many different cell lines, including neuroblastoma cell lines, but the effect of EV-A71 on autophagy in primary neuronal cells is unclear. Neuronal cells derived from human NSCs were infected with EV-A71, and an IFA was performed to measure the expression of LC3 positive puncta and EV-A71 3D pol at 12 and 24 h postinfection. EV-A71 infection induced the formation of LC3-positive puncta in neurons, and these LC3-positive puncta colocalized with the EV-A71 3D pol protein-expressed cells. (Fig. 4A). The LC3-positive puncta were quantified and significantly increased with increasing infection time in EV-A71-infected neurons (Fig. 4B). During the formation of autophagosomes, LC3-I is conjugated with phosphatidylethanolamine www.nature.com/scientificreports/  www.nature.com/scientificreports/ (PE) to form LC3-II, so the level of LC3-II correlates with the number of autophagosomes 34 . Proteins were extracted from EV-A71-infected neurons derived from human NSCs, and the levels of LC3-I and LC3-II were analyzed using western blotting. The protein level of LC3-II significantly increased as infection time increased (Fig. 4C,D). Next, differentiated IMR-32 cells were infected with EV-A71 at an MOI of 2, and proteins were harvested at various time points. Western blot analysis revealed that the protein level of LC3-II was significantly increased with increasing infection time (Fig. 4E,F). The protein level of LC3-II was also increased with increasing viral infection titer ( Figure S3A). Different EV-A71 strains also increased the expression level of LC3-II in differentiated IMR-32 cells ( Figure S3B). One of the autophagy initiation steps is the activation of the class III phosphatidylinositol 3-kinase (PI3K-III) complex consisting of Beclin1, VSP15, VSP34 and ATG14. To regulate autophagy initiation, Beclin1 must be phosphorylated at several residues, including S15, which is phosphorylated by ULK. Beclin1 S15 phosphorylation is required for PI3K-III complex activation in PtdIns-3P production, which is required for autophagosome formation 35,36 . We measured the expression of phospho-Beclin1 S15 in EV-A71-infected differentiated IMR-32 cells and found that its expression was induced at 4 h postinfection (Fig. 4G). EV-A71 infection induced the formation of LC3 puncta and the protein expression level of LC3-II was increased over the course of EV-A71 infection in the neuronal cells. The results revealed that EV-A71 infection induced autophagy in neurons and differentiated IMR-32, and the initiation of autophagosomes was associated with phospho-Beclin1 S15.

EV-A71-infected differentiated IMR-32 cells exhibit complete autophagic flux. Autophagy can
be divided into multiple stages, and complete autophagic flux refers to the formation of autolysosomes and their degradation. To test whether autolysosomes are formed in EV-A71-infected differentiated IMR-32 cells, the autophagy tandem sensor RFP-GFP-LC3B was added to EV-A71-infected differentiated IMR-32 cells to monitor autophagy flux. GFP-LC3 is acid sensitive, and RFP-LC3 is acid insensitive. Both GFP-LC3 and RFP-LC3 are visible in autophagosomes. GFP-LC3 is quenched in autolysosomes, which have an acidic pH, but RFP-LC3 is still visible in autolysosomes. RFP expression was higher than GFP expression in EV-A71-infected differentiated IMR-32 cells at 40 h post infection (Fig. 5A). The amount of GRP-and RFP-LC3 puncta was quantified and the RFP-LC3 puncta was significantly more than GFP-LC3 puncta at 40 h post infection (Fig. 5B). These findings indicated that EV-A71 infection induces autolysosome formation in differentiated IMR-32 cells. When the outer membranes of autophagosomes fuse with lysosomes to form autolysosomes, the inner membrane and luminal content of autophagic vacuoles, such as LC3-II, are degraded by lysosomal enzymes 17 . p62 is an ubiquitinbinding scaffold protein that can bind LC3, leading to the degradation of ubiquitinated proteins and its own degradation in autolysosomes 37 . Therefore, a decreased level of p62 is observed in complete autophagic flux. Proteins were extracted from differentiated IMR-32 cells infected with EV-A71 at various time points. Western blot analysis revealed that the level of p62 was significantly decreased in EV-A71-infected cells compared with mockinfected cells over the course of EV-A71 infection (Fig. 5C,D). The expression of p62 was also decreased in the mock-infection. Because neuronal cells have basal autophagy to maintain cellular metabolism and homeostasis, the decreased p62 may be caused by the spontaneous autophagy. In addition, a turnover assay 38 was performed to investigate the induction of complete autophagic flux. Differentiated IMR-32 cells were treated with bafilomycin A1, which can inhibit the fusion of autophagosomes and lysosomes 39 , and then subjected to EV-A71 or mock infection. Bafilomycin A1 significantly increased the expression of LC3-II in mock-or EV-A71-infected differentiated IMR-32 cells compared with untreated cells (Fig. 5E,F). In addition, differentiated IMR-32 cells were treated with pepstatin A and E-64d, which can inhibit the activity of lysosome protease 40 , and then subjected to EV-A71 or mock infection. Among the mock-and EV-A71-infected differentiated IMR-32 cells, the level of LC3-II was significantly higher in cells treated with pepstatin A and E-64d than in untreated cells (Fig. 5G,H). These results show that EV-A71 infection could induce complete autophagic flux in differentiated IMR-32 cells.

The induction of autophagy increases EV-A71 vRNA replication in differentiated IMR-32 cells.
To determine whether autophagy affects EV-A71 replication in differentiated IMR-32 cells, the cells were treated with two autophagy inducers, rapamycin and metformin, and then infected with EV-A71. Rapamycin induces autophagy by inhibiting the activity of mTOR 41 . The expression level of LC3-II was more significantly upregulated in rapamycin-treated differentiated IMR-32 cells than in DMSO-treated IMR-32 cells over the course of EV-A71 infection (Fig. 6A,B). Metformin, which induces autophagy by activating the AMPK pathway 42 , also increased LC3-II levels in differentiated IMR-32 cells compared with untreated IMR-32 cells during virus infection (Fig. 6C,D). The expression levels of the EV-A71 5′ UTR vRNA were measured with RT-qPCR. The relative EV-A71 5′UTR vRNA levels were significantly upregulated in differentiated IMR-32 cells treated with both autophagy inducers compared with untreated cells (Fig. 6E). To knock down the expression of ATG5 in differentiated IMR-32 cells, small hairpin RNA (shRNA) was delivered by a lentivirus expression system. The cells were transduced with lentivirus expressing shRNA targeting ATG5 or negative control (NC) shRNA, and puromycin was used to select stable clones. Western blotting was performed to confirm the ATG5 knockdown efficiency (Fig. 6F). The expression of LC3-II was also decreased significantly in ATG5-knockdown cells compared to shNC-transfected cells (Fig. 6F,G). This indicated that the formation of autophagosomes was inhibited under this condition. The expression of EV-A71 5′ UTR vRNA was significantly decreased in ATG5knockdown differentiated IMR-32 cells compared to shNC-transfected differentiated IMR-32 cells (Fig. 6H). This result revealed that the inhibition of autophagy decreased EV-A71 vRNA replication in differentiated IMR-32 cells. www.nature.com/scientificreports/ of autophagy is important for this phenomenon, differentiated IMR-32 cells were treated with bafilomycin A1 or pepstatin A and E-64d and then infected with EV-A71. We performed the IFA to detect the expression of LC3 positive-puncta in bafilomycin A1-treated or pepstatin A and E-64d-treated differentiated IMR-32 cells infected with EV-A71. Bafilomycin A1 inhibits the fusion of autophagosomes and lysosomes, so the expression levels of LC3-positive puncta were increased significantly in bafilomycin A1-treated cells compared to DMSO-treated controls. The treatment of pepstatin A and E-64d inhibit the protease activity of autolysosomes and prevent the degradation of autolysosome components, but not inhibit the fusion of autophogosomes and lysosomes. Therefore, the expression of LC3-positive puncta was not increased as much as bafilomycin A1 treated-condition (Fig. 7A,B). Total RNA was harvested from cells at various time points, and RT-qPCR was performed to measure the expression of EV-A71 vRNA. The relative amount of EV-A71 vRNA was significantly increased in the bafilomycin A1-treated group compared with the DMSO-treated control (Fig. 7C). In contrast, EV-A71 vRNA was not increased in pepstatin A and E-64d-treated cells (Fig. 7D). Thus, blockade of the fusion of autophagosomes and lysosomes could induce more autophagosome formation and support EV-A71 vRNA replication in differentiated IMR-32 cells. The EV-A71 viral proteins 3D and 3A colocalized with the neuronal-specific markers MAP2 and GAD67, which indicated that EV-A71 could infect neurons. EV-A71 vRNA and viral protein expression were detected in EV-A71-infected neurons and viral growth curve was also increased with infection time. These results revealed that EV-A71 could replicate in neurons. However, EV-A71 infection did not induce significant neuronal cell death, and caspase 3 was not activated in neuronal cells derived from human NSCs and IMR-32 cells. Furthermore, caspase 8 and caspase 9 were not activated in EV-A71-infected differentiated IMR-32 cells. These results revealed that EV-A71 infection did not induce caspase-dependent apoptosis in neuronal cells, in contrast to results in diverse EV-A71-infected human cell lines. EV-A71 infection activates the intrinsic apoptosis pathway in neural lineage cells and the extrinsic apoptosis pathway in nonneuronal cells 47 . PV has been indicated to infect primary hippocampal neurons and delay the apoptosis of PV-infected neurons compared to that of nonneuronal cells 48 . VZV can infect neurons derived from human embryonic stem cells and cause persistent nonproductive infection for weeks but does not cause an obvious CPE or apoptosis in virus-infected neurons 45,49,50 . The activation of caspase 3 was shown to promote EV-A71 viral protein translation and viral production in RD cells 31 . However, we found that the inhibition of caspase3 does not decrease EV-A71 growth in differentiated IMR-32 cells. EV-A71 replication in neuronal cells may not be dependent on the activation of caspase 3.

Discussion
Basal autophagy is important for cellular metabolism and homeostasis, especially in postmitotic cells, such as neuronal cells. Because neuronal cells cannot dilute wastes such as protein aggregates, nonfunctional organelles or organelles damaged by cell division, autophagy is important for neuronal cells to maintain the quality of their proteome and organelles to support their long-term viability 51 . Autophagy also plays an important role in the mechanism by which cells remove invading pathogens. Virus infection, including enterovirus infection, has been shown to induce autophagy 52 . In our study, EV-A71-infected neuronal cells showed more LC3-positive puncta in the cytosol than did mock-infected neuronal cells. LC3-positive puncta are formed in the process of autophagy and indicate the formation of autophagosomes. The number of LC3-positive puncta in neuronal cells increased with infection time. Western blot assays revealed that the expression of LC3-II was increased in neuronal cells over the course of EV-A71 infection. www.nature.com/scientificreports/ We found that the expression of LC3-II continuously increased in EV-A71-infected neurons and in differentiated IMR-32 cells. However, the expression of LC3-II was decreased in EV-A71-infected neuroblastoma SK-N-SH cells in a late stage of infection 28 . This increased expression of LC3-II may be due to increased autophagy induction or may indicate the inhibition of autophagosome and lysosome fusion. The autophagy tandem sensor GFP-RFP-LC3B was added to EV-A71-infected differentiated IMR-32 cells to confirm autophagic flux. GFP-LC3 is acid sensitive, and RFP-LC3 is acid insensitive. Both GFP-LC3 and RFP-LC3 are visible in neutral pH autophagosomes. GFP-LC3 is quenched in autolysosomes, which have an acidic pH, but RFP-LC3 is still visible in autolysosomes. RFP-LC3 expression was higher than GFP-LC3 expression in EV-A71-infected differentiated IMR-32 cells at a late stage of infection. That revealed the autolysosomes formation. We used bafilomycin A1 to inhibit the fusion of autophagosomes and lysosomes and used pepstatin A1 and E-64d to inhibit the protease activity of autolysosomes. Both conditions can prevent the degradation of autophagosomal components, such as LC3-II, in autolysosomes. If the autophagic flux is complete, the expression level of LC3-II is increased with drug treatment. In contrast, if the expression level of LC3-II is not changed with drug treatment, complete autophagic flux does not occur. In our results, the expression of LC3-II was significantly increased under both conditions compared to the untreated control condition, indicating that complete autophagic flux occurred in EV-A71-infected neuronal cells. Autophagy can deliver viral components, viral particles or the host factors required for viral replication to lysosomes for degradation and can suppress virus replication 53 . However, viruses have developed molecular strategies to avoid their recognition and degradation via autophagy and can even exploit autophagy for their benefit 54 . CVB3 has been indicated to inhibit the transcription and translation of syntaxin 17 and SNARE to block the fusion of autophagosomes and lysosomes, preventing degradation via autophagy 55 . The CVB3 2A protease can cleave p62 to evade detection and autophagy-mediated degradation. CVA16 infection also causes incomplete autophagic flux in infected cells 56 . PV is detected by the host protein galectin 8, which subsequently initiates autophagy to degrade the viral genome; however, PV can use the host protein HRAS-like suppressor 3 (PLA2G16) to avoid autophagy detection 57 . However, our results indicated that EV-A71 infection did not inhibit the degradative function of autophagy in differentiated IMR-32 cells.
In this study, we found that EV-A71 vRNA replication was increased in differentiated IMR-32 cells treated with autophagy inducers. In contrast, EV-A71 vRNA replication was decreased in differentiated IMR-32 cells in which autophagy was inhibited. RNA viruses can exploit the double-membrane structure of autophagosomes to proceed with their replication because autophagosomes supply a membrane-bound and protective environment 54 . Autophagy also supplies metabolites and energy for virus replication. Autophagy can support EV-A71 replication in human cell lines. There is no direct evidence to indicate which stage of autophagy is involved in this mechanism. Therefore, differentiated IMR-32 cells were treated with bafilomycin-A1 to block the fusion of autophagosomes and lysosomes, after which additional autophagosomes accumulated in the cells. Under this condition, EV-A71 vRNA replication was significantly increased compared with that under the DMSO-treated control condition. The EV-A71 3D pol protein level was also increased in bafilomycin-A1-treated differentiated IMR-32 cells compared with control differentiated IMR-32 cells. In EV-A71-infected RD cells, the blockade of autolysosome fusion by chloroquine and bafilomycin A1 was shown to inhibit viral replication 58 . The fusion of autophagosomes and lysosomes requires v-ATPases to form an acidic environment. The acidic environment in vesicles is critical for the activation of proteases and hydrolases. PVs require vesicular acidification for the cleavage of VP0 to form VP2 and VP4 58 . The acidic environment in vesicles is important for viral replication in EV-A71-infected RD cells. However, vesicular acidification may not be the major mechanism that regulates EV-A71 replication in neuronal cells. Pepstatin A and E-64d were used to inhibit the activity of lysosomal enzymes, but not autolysosomes formation. The LC3 positive puncta expression of pepstatin A and E-64d-treated cell was not as much as bafilomycin A1-treated cells. Under this condition, EV-A71 vRNA replication and the viral protein EV-A71 3D pol were not increased. The results in PV-and EV-A71-infected RD cells were similar 59 and indicate that autolysosomes cannot degrade the factors required for EV-A71 replication.
In this study, we demonstrate that differentiated neuronal cells are permissive to EV-A71 infection. However, apoptosis was not induced, while autophagy was induced in EV-A71-infected differentiated neuronal cells, which regulated EV-A71 vRNA replication. Blockade of the fusion between autophagosomes and lysosomes increased the expression of EV-A71 vRNA in differentiated neuronal cells. This finding indicates that autophagosome accumulation can increase EV-A71 vRNA replication in differentiated neuronal cells. The band intensities of LC3-II were measured using ImageJ software and normalized to the β-actin. (E) The total RNA was extracted, and RT-qPCR was performed to measure the relative levels of EV-A71 5′ UTR vRNA. (D) Differentiated IMR-32 cells transduced with shNC or shATG5 lentivirus were infected with EV-A71 (MOI = 2) or mock infection. Western blot assays were performed to measure the levels of ATG5, LC3-I, LC3-II, and EV-A71 3D pol . (F) The band intensities of LC3-II were measured using ImageJ software and normalized to the β-actin. (G) Total RNA was also extracted for RT-qPCR to measure the relative levels of the EV-A71 5′ UTR. β-actin was used as an internal control in all western blot. The experiments were performed in triplicate, and the error bars represent the SD. Student's t test. *p < 0.05; **p < 0.01; ***p < 0.001. www.nature.com/scientificreports/ Methods NSC culture. The human NSCs used in this study were derived from NIH-approved H9 human embryonic stem cells (Thermo Fisher Scientific, MA, USA). The cells were handled following a kit's protocol. The complete medium used to maintain the human NSCs consisted of 1 × KnockOut D-MEM/F12, 2 mM GlutaMAX-I supplement, 20 ng/ml FGF-β, 20 ng/ml EGF and 2% StemPro neural supplement (all from Thermo Fisher Scientific, MA, USA). Adherent human NSCs were thawed and subcultured in plates precoated with a matrix consisting of CELLStart substrate and D-PBS containing calcium and magnesium (both from Thermo Fisher Scientific, MA, USA). The cells were incubated at 37 °C with 5% CO 2, and the medium was replaced with new complete medium every 2 days. When the NSCs were approximately 90% confluent, they were ready to be passaged. The spent medium was removed from the cells, and the cellular surface was rinsed with 1 × D-PBS. To detach the cells, they were treated with Accutase and incubated at 37 °C in a 5% CO 2 incubator for 1-2 min, after which complete medium was added to stop the dissociation reaction. The cells were gently pipetted several times, transferred to a 15 ml tube, and centrifuged at 200 × g for 5 min. The supernatant was aspirated, and the cells were resuspended in complete medium and seeded in precoated plates for maintenance.  EV-A71 infection. The differentiation media of neuronal cells were removed, and the cells were washed once with PBS. Virus stock was diluted in neuronal differentiation medium and added to the cells for adsorption at 37 °C in a 5% CO 2 incubator. After 1 h, the viral solution was removed, and the cells were washed with PBS. Fresh culture medium was added to cells, which were incubated at 37 °C in a 5% CO 2 incubator.

Transduction of GFP-RFP-LC3. A Premo autophagy tandem sensor RFP-GFP-LC3 kit (Thermo Fisher
Scientific, MA, USA) was used to monitor the autophagic flux in EV-A71-infected neuronal cells. A total of 10,000 cells were added with 2 μl of the tandem sensor LC3-FP. The reagent was mixed gently with fresh culture medium and added to cells. The cells were incubated at 37 °C in a 5% CO 2 incubator overnight (> 16 h). Cells were imaged with an Olympus BX51 fluorescence microscope (Olympus, Tokyo, Japan). RNA extraction and RT-qPCR. Total RNA was collected with a TRI reagent solution (Thermo Fisher Scientific, MA, USA) at various time points. The cells were homogenized in TRI reagent solution and mixed with chloroform. The homogenate was incubated for 5 min at room temperature and centrifuged at 12,000 × g for 15 min at 4 °C. The aqueous phase, which contained RNA, was transferred to fresh tubes, and an equal amount of isopropanol was added before incubation at − 80 °C overnight. The mixture was centrifuged at 12,000 × g for 10 min at 4 °C, and the supernatant was removed. Then, the RNA pellet was washed with 1 ml of 75% ethanol and centrifuged at 7,000 × g for 5 min at 4 °C. The 75% ethanol was removed, and the RNA pellet was air dried at room temperature, dissolved in sterile water and incubated at 55 °C to fully dissolve. This RNA solution was stored at − 80 °C. cDNA synthesis was performed with a RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, MA, USA). RNA (1-0.5 μg) was used for cDNA synthesis, and the first step was treatment with DNase I at 37 °C for 30 min to remove genomic DNA. EDTA (50 mM) was added to the mixture to inactivate DNase I at 65 °C for 10 min. A random hexamer primer was used for cDNA synthesis, and the protocol was as follows: 65 °C, 5 min; 4 °C, 2 min; 25 °C, 5 min; 42 °C, 80 min; 70 °C, 5 min; and incubation at 4 °C. One microliter of cDNA sample was subjected to qPCR with 5 μM primers, and SYBR green was used to quantify expression. qPCR was carried out in a 384-well plate, and the results were analyzed with a Roche LightCycler 480 system. Each sample was assayed in triplicate, and 18S rRNA was used as a reference gene. We analyzed the relative expression level of each gene by the 2 -∆∆CT method. The following primers were used in this study: EV-A71 5′ untranslated region (UTR), forward (5′ CCC TGA ATG CGG CTA ATC C 3′) and reverse (5′ ATT GTC ACC ATA AGC AGC CA 3′) and 18S rRNA, forward (5′ GTA ACC CGT TGA ACC CCA TT 3′) and reverse (5′ CCA TCC AAT CGG TAG TAG CG 3′).

Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.