Encephalomyocarditis virus impairs the blood-brain barrier by degrading tight junction proteins via AKT3-dependent autophagic and apoptotic pathways

ABSTRACT

Encephalomyocarditis virus (EMCV) infection causes viral encephalitis; however, the mechanisms underlying blood-brain barrier (BBB) disruption remain poorly understood. Here, we demonstrate that EMCV actively replicates in mouse brain tissue, induces robust neuroinflammation characterized by elevated proinflammatory cytokines and chemokines, and markedly increases BBB permeability as evidenced by Evans blue and sodium fluorescein extravasation. Importantly, tight junction (TJ) proteins ZO-1 and Occludin are selectively degraded at the post-transcriptional level, whereas Claudin-5 expression remains stable. Consistently, in vitro BBB models confirmed EMCV traversal, reduced transendothelial electrical resistance, and TJ disruption. Mechanistically, EMCV induces biphasic PI3K/AKT modulation and specifically downregulates AKT3. Notably, AKT3 knockdown exacerbates both autophagy and apoptosis, thereby accelerating ZO-1 and Occludin degradation while promoting viral replication. Furthermore, pharmacological inhibition of autophagy (chloroquine) or apoptosis (Z-VAD-FMK) effectively rescues TJ proteins and reduces viral load. Interestingly, the Caspase-8 inhibitor Z-IETD-FMK provides the most robust protection, implicating the extrinsic apoptotic pathway as the dominant route. Collectively, EMCV sequentially activates non-redundant AKT3-dependent autophagic and apoptotic pathways to degrade TJ proteins, ultimately enabling viral traversal across the compromised BBB and offering therapeutic targets for viral encephalitis.

Introduction

Viral encephalitis is a severe infectious disease of the central nervous system (CNS), frequently leading to high mortality and neurological sequelae [1]. The blood-brain barrier (BBB) serves as a critical structural barrier separating the peripheral circulation from the brain parenchyma, and pathogen crossing of the BBB represents a common pathway for numerous neurotropic pathogens to invade the CNS [2,3]. Encephalomyocarditis virus (EMCV), a pathogen belonging to the family Picornaviridae, is capable of cross-species infection in multiple mammalian species, including humans, pigs, and rodents [4]. Following host infection, EMCV can induce acute and fatal encephalitis or myocarditis. Notably, in mouse models, the lethality rate of EMCV infection reaches up to 100% with distinct pathological features [5]. Given its unique etiological and epidemiological characteristics, EMCV holds great potential as a research model for viral encephalitis. Nevertheless, the molecular mechanisms by which EMCV impairs BBB function remain to be fully elucidated.

The BBB serves as a crucial structure for safeguarding the central nervous system, with its integrity largely dependent on the tight junctions (TJs) formed between brain microvascular endothelial cells. This dynamic complex consists of multiple proteins, wherein transmembrane proteins, including members of the Occludin and Claudin families, interact with intracellular scaffold proteins such as ZO-1, thereby establishing an effective physical barrier adjacent to the cells [2,3]. Numerous neuroinvasive viruses have developed sophisticated strategies to breach this barrier. For example, the Japanese encephalitis virus (JEV) and West Nile virus (WNV), both belonging to the Flaviviridae family, can replicate within brain microvascular endothelial cells, evade the antiviral effects of the host IFITM1 protein, and induce the down-regulation of TJ proteins, which increases barrier permeability [6–9]. Additionally, enteroviruses such as Coxsackievirus B3 (CVB3) and enterovirus 71 (EV71) can induce the internalization and degradation of TJ proteins like Occludin following their infection of endothelial cells, resulting in cytoskeletal rearrangement and barrier disruption [10–12]. These viruses manipulate the degradation of TJ proteins by directly infecting endothelial cells and antagonizing signaling pathways involved in inflammatory responses, ultimately leading to the disintegration of TJ structures [13–15].

A serine/threonine-specific protein kinase AKT family comprises three functionally non-redundant isoforms- AKT1, AKT2, and AKT3 [16,17]. While AKT1 broadly regulates cell survival and AKT2 controls glucose homeostasis, AKT3 is the predominant isoform in the brain, accounting for ~50% of total AKT protein, and is indispensable for normal brain development and cognition [16,17]. Importantly, AKT3 loss cannot be compensated by other isoforms, underscoring its unique role as the principal AKT regulator in the CNS [16]. Recent studies further implicate AKT3 as a critical guardian of BBB integrity: endothelial N-cadherin stabilizes TJs via the PI3K-AKT3 axis [18], and R-Ras activates AKT3 to enforce endothelial quiescence and barrier stability through Jagged1-Notch signaling [19]. Under physiological conditions, AKT preserves TJs structure and BBB function [20–22], whereas under pathological insults such as inflammatory or ischemia-hypoxia, diminished AKT phosphorylation triggers BBB breakdown [23,24]. In addition to noninfectious insults, neurotropic viruses exploit the PI3K/AKT pathway to disrupt BBB integrity: EV71 and human immunodeficiency virus (HIV) have been shown to compromise barrier function through AKT modulation [25,26]. Moreover, AKT3 uniquely couples to mTORC2-mediated Ser473 phosphorylation, positioning it to balance mTORC1-dependent autophagy against apoptotic priming [16,17]. This distinctive regulatory capacity makes AKT3 a compelling candidate for investigating how neurotropic viruses coordinate degradative pathways to compromise barrier integrity.

Here, we aim to clarify the damaging effect of EMCV on the BBB and its capacity to traverse the BBB using three approaches: in vivo animal models, in vitro BBB models, and immortalized human brain microvascular endothelial cells (hCMEC/D3). We also seek to reveal the critical role of AKT in this process. This study provides new insights into EMCV pathogenesis and identifies a potential therapeutic target for viral encephalitis.

Materials and methods

Animals and ethics statements

Six- to eight-week-old, SPF-grade BALB/c mice (18–22 g) used in this study were obtained from the Laboratory Animal Center of Lanzhou Veterinary Research Institute, Chinese Academy of Agricultural Sciences. All experimental procedures were reviewed and approved by the Animal Ethics Committee and the Animal Protection and Utilization Committee of Northwest Minzu University (Approval No. xbmu-sm-2024101). During the study, we conducted in accordance with the ARRIVE guidelines to ensure transparency and reproducibility and meanwhile, strictly followed the animal welfare and experimental operation requirements set forth in the “Guidelines for the Management and Use of Laboratory Animals” issued by the Ministry of Science and Technology of China (2006).

Animal experiment

Randomization was performed using the RAND function in Microsoft Excel (Microsoft 365) to generate the allocation sequence. Based on this sequence, 57 BALB/c mice were randomly allocated to two experimental groups. To ensure allocation concealment, the researcher who generated the randomization sequence was not involved in the subsequent conduct of the study or outcome assessment. The control group (MOCK, n = 16) received an intramuscular injection of 500 μL of sterile PBS. The EMCV-infected group (n = 41) received an intramuscular injection of 500 μL of EMCV (100 TCID₅₀). The latter was subdivided by days post-infection (dpi) into the 1st day (n = 9), the 3rd day (n = 16), and the 5th day (n = 16) subgroups. Mice from different groups were housed under controlled conditions in separate cages with ad libitum access to food and water. The housing environment was maintained on a 12/12-hour light/dark cycle, with temperature and relative humidity set at 22 ± 1°C and 26 ± 2%, respectively. At the end of the experiment, mice were administered under isoflurane anesthesia, blood was collected from the ocular veins, and brain tissues were dissected to analyze the level of viral infection and the damage to the BBB structure. All outcome assessments, including viral loading, histological analysis, and image quantification, were performed by investigators blinded to the group allocation. Statistical data are expressed and analyzed as the mean ± SD from at least three independent experiments using Student’s t-test in GraphPad Prism software (*p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, ns represents no significant differences). These experiments have adhered to the ARRIVE guidelines.

Virus and cell lines

The EMCV (PV21) strain was supplied and propagated by the Biomedical Research Center of Northwest Minzu University. Baby hamster kidney cells (BHK-21) and human glioma cells (U251) were maintained in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% neonatal bovine serum (NBS). Immortalized human brain microvascular endothelial cell line D3 (hCMEC/D3) cells were obtained from BNCC Bio (Henan, China) and cultured in Complement Endothelial Cell Medium (ECM; Cat.#1001) from ScienCell Research Laboratories, which contains 5% fetal bovine serum, 1% Endothelial Cell Growth Supplement (ECGS), and 1% penicillin-streptomycin. All cell lines were incubated at 37°C with 5% CO₂.

Antibodies and reagents

The mouse monoclonal antibody (Mab) targeting the EMCV VP1 protein was prepared by GeneCreate (Wuhan, China). The Mab (TA503129) targeting β-tubulin was purchased from OriGene Technologies (Maryland, USA). Polyclonal antibodies against ZO-1 (21773-1-AP), Occludin (27260-1-AP), AKT2 (28113-1-AP), AKT3 (21641-1-AP), Caspase-3 (19677-1-AP) and LC3 (14600–1-AP) were all purchased from Proteintech Biotechnology (Wuhan, China). Claudin 5 Mab (clone 4C3C2) (35-2500) was bought from Invitrogen (Carlsbad, CA, USA). Antibodies against AKT1 (342529), phospho-AKT (R381555), PI3K (R22768), and phospho-PI3K (341468) were provided by ZEN-Bioscience (Chengdu, China). The SQSTM1/p62 antibody (A11483) and Beclin-1 antibody (Abp57192) were purchased from ABclonal (Wuhan, China) and Abbkine (Wuhan, China), respectively. Peroxidase-labeled and Cy3-labeled goat anti-mouse IgG (H+L) and 488-labeled goat anti-rabbit IgG (H+L) secondary antibodies, were all purchased from Jackson ImmunoResearch Laboratories (Pennsylvania, USA).

Reagent LY294002 (HY-10108), obtained from MCE (Shanghai, China), is a broad-spectrum inhibitor of PI3K/AKT that also inhibits phospho-AKT [27,28]. Afuresertib (A136023), an ATP-competitive pan-Akt inhibitor acting on Akt1, Akt2, and Akt3, was sourced from AmBeed Corporation (Shanghai, China). The Caspase-3 inhibitor Ac-DEVD-CHO (HY-P1001) and the Caspase-8 inhibitor Z-IETD-FMK (HY-101297) were acquired from MedChemExpress (MCE) (Shanghai, China). The Caspase-9 inhibitor Z-LEHD-FMK TFA (S7313) was procured from Selleck (Houston, TX, USA). Animal anesthetic, isoflurane (792632) was sourced from Sigma-Aldrich. Cytotoxicity of these inhibitors was assessed using the Cell Proliferation and Cytotoxicity Assay Kit (CCK-8) (Enhanced type, MA0218), purchased from Meilun Bio (Dalian, China).

Virus titration

For the in vivo experiment, 100 mg of mouse brain tissue (4 mice each group) was placed in a clean homogenizer and 1 mL of PBS was added. After thorough homogenization, the homogenate was centrifuged, and the supernatant was passed through a 0.22 μm Millipore sterilizing filter to obtain the test virus solution. Viral titer was then determined by the TCID₅₀ method. Briefly, collected samples were subjected to tenfold serial dilution in DMEM. Each dilution was inoculated onto 96-well plates containing monolayers of BHK-21 cells, with eight replicate wells per dilution and control wells of uninfected cells. Plates were incubated at 37 °C in 5% CO₂ and examined daily by microscopy for cytopathic effect (CPE) until the CPE stabilized. Titer was calculated using the Spearman-Kärber method. For the in vitro experiment, EMCV‑infected cells were freeze‑thawed three times, filtered for sterilization, and titrated by the same procedure.

Hematoxylin and eosin (H&E) staining

Brain tissues were harvested from mice at 3, and 5 dpi, along with a mock-infected control (MOCK) group (3 mice each group), and fixed in 4% paraformaldehyde. The fixed samples were trimmed into blocks measuring 5 mm × 5 mm × 3 mm and processed through dehydration, clearing, paraffin infiltration, embedding, and sectioning according to standard protocols to obtain 3 µm sections. After H&E staining, the sections were visualized under an optical microscope.

Evans blue (EB) staining in vivo

To assess the BBB permeability in mice, a 2% EB solution (MCE, Shanghai, China) was administered via the tail vein at 4 mL/kg body weight under isoflurane anesthesia. Mice were euthanized 1 hour later, and brains were removed and photographed. Each weighed brain was minced and incubated in formamide (Aladdin, Shanghai, China) at 37°C for 48 hours to extract the dye. Samples were then centrifuged at 3000 rpm for 15 minutes, and the supernatant was transferred to a 96-well microplate. Absorbance at 620 nm was measured with a Multiskan Mk3 microplate reader (Thermo Fisher Scientific Instruments, Shanghai, China). Using formamide as the blank, EB formamide standards of 8.0, 4.0, 2.0, 1.0, 0.5, and 0.25 mg/L were prepared to generate a standard curve, and EB concentrations of brain from mice at 1, 3, and 5 dpi, along with a mock-infected control group (3 mice each group) were calculated from that curve. EB content (µg/g) = [EB brain tissue content (µg/L) × formamide volume (mL)] / [brain weight (g)].

In vivo fluorescein sodium permeability experiment

To assess BBB permeability in model animals, we administered a sodium fluorescein solution (MCE, Shanghai, China) via the tail vein at a dosage of 0.5 mg/g body weight. After allowing 20 minutes for circulation, we euthanized the mice from different groups (MOCK, 1, 3, and 5 dpi) (3 mice each group) and collected brain tissues. Imaging was performed with an AniView Phoenix full‑spectrum in vivo imaging system (Guangzhou Boluteng Biotechnology Co., LTD., Guangzhou, China). During acquisition, excitation was 465 nm and emission was 520 nm.

Enzyme-linked immunosorbent assay (ELISA)

To measure serum levels of inflammation-related factors (IL-1β, IL-6, CXCL10, and CCL5) after EMCV infection, mice from different groups (MOCK, 1, 3, and 5 dpi) (4 mice/group) were anesthetized and blood was collected from the ocular veins for serum isolation. ELISA kits from Lianke Biology (Hangzhou, China) were used according to the manufacturer’s instructions: Mouse CCL5/RANTES ELISA Kit, Mouse CXCL10/IP-10 ELISA Kit, Mouse IL-6 ELISA Kit, and Mouse IL-1β ELISA Kit. Standards and samples were applied to plates, which were washed and blocked before sequential incubation with biotinylated detection antibodies and streptavidin-HRP. Color was developed with TMB substrate, halted with stop solution, and absorbance readings were taken at 450 nm and 630 nm on a Thermo Scientific microplate reader.

In vitro BBB permeability assays

Transwell cell chambers with a 3 μm pore size and PET membrane (FTW040-12Ins) (Beyotime) were used to construct an in vitro co-culture model of immortalized hCMEC/D3 and U251 cells to mimic the BBB. U251 cells were seeded on the outer surface of the filter membrane two days before the experiment began in the inverted chamber. On the designated date (defined as day 0 of co-culture), the chamber was returned to a 12-well plate to form a two-chamber system, and hCMEC/D3 cells were seeded on the inner side of the upper chamber’s filter membrane. Seeding density for both cell types was 1.5 × 10⁴ cells/cm². After 3 days of co-culture, transendothelial electrical resistance (TEER) was measured with an R/V Meter of Epithelium (Kingtech technology co ltd., Beijing, China) to assess barrier integrity.

To assess permeability of the in vitro BBB model, we measured transwell flux of FITC-Dextran. After 3 days of culture, FITC-Dextran (MW 4000, Beyotime) at a final concentration of 1 mg/mL was added to the upper chamber. Following one hour incubation, culture media from the upper and lower chambers were collected, and fluorescence was measured with a GloMax Explorer Multimode Microplate Reader (Promega, USA). The apparent permeability coefficient (P_app) was calculated using the following formula [29]: Cb, fluorescein concentration in the lower chamber; Vb, volume of the lower chamber (cm³); Ca, fluorescein concentration in the upper chamber; A, cell growth area (cm²); t, time (s).

$$P_{\mathit{app}}=\frac{C_b\times V_b}{C_a\times \mathit{A}\times \mathit{t}}$$

In vitro EMCV infection assay

To examine how different inoculation times and doses of EMCV affect host proteins, hCMEC/D3 cells were plated in six-well plates. When cultures reached 80–90% confluence, cells were infected with EMCV at 1 (or 0.01, 0.1) multiplicity of infection (MOI). After a 1-hour adsorption period, the viral inoculum was removed and cells were returned to maintenance medium (ECM medium with 1.5% FBS). Cells were then incubated for 24 hours or at a series of time points (1, 2, 4, 6, 8, 12, and 24 hours post-infection (hpi)) and subsequently harvested for analysis of viral and host mRNA and protein levels.

EMCV infection of the BBB cell model was performed as follows: upon reaching the specified resistance level, EMCV was added to the apical side of the Transwell at a concentration of 1 MOI or at varying MOIs (0.01, 0.1, 1). After a one-hour incubation, the viral solution was aspirated and replaced with maintenance medium (ECM medium supplemented with 1.5% FBS). Cells were then harvested at defined time points post-infection for Western blot analysis of TJ proteins expression. At the same time, medium from the lower (basolateral) chamber was collected to determine virus titer by TCID₅₀ assay.

Western blot

Approximately 50 mg of animal brain tissue from each mice in MOCK group or 5 dpi group (3 mice/group) was rapidly ground to a powder in liquid nitrogen, then 400 μL of high-efficiency RIPA lysis solution (Solarbio, Beijing, China) with protease inhibitors was added and the sample was incubated on ice for 30 minutes. Treated cell samples were lysed on ice with an appropriate volume of lysis buffer. After lysis, tissue or cell lysates were centrifuged at 12,000 g and 4 °C for 20 minutes, the supernatant was collected, and protein concentration was determined by the BCA method. Protein samples were separated by SDS-PAGE and transferred to PVDF membranes (Millipore, Bedford, Massachusetts, USA). Membranes were blocked with 5% skim milk in TBST, then probed sequentially with primary antibodies against target proteins, such as TJ proteins ZO-1 and Occludin, followed by HRP-conjugated secondary antibodies. Thermo Scientific^TM PageRuler Prestained Protein Ladder (Cat#. 26617) was used as molecular marker here. Signal development was performed with ECL chemiluminescence reagent (Bio-Rad, California, USA), and band intensities were quantified with ImageJ software.

Small interfering RNAs and transfection

Small interfering RNAs (siRNAs) targeting human AKT1, AKT2, and AKT3, as described in the literature [30–32], were selected. The sequences are listed below: siAKT1: 5’-GGAGATCATGCAGCATCGC (dTdT)-3’, siAKT1: 5’-GCUCCUUCAUUGGGUACAATT (dTdT)-3’, and siAKT3: 5’-UGCAAGUGGACGAGAAUAAUU (dTdT)-3’. All three specific siRNAs were synthesized by Accurate Biology Company, and the negative control siRNAs (siNCs) were supplied by the same company (Accurate Biology Company, Hunan, China).

hCMEC/D3 cells were seeded in six-well plates and grown to 70–80% confluence. siRNA transfections were then performed using the AccuFect RNAi Transfection Kit (AG51018) from Accurate Biology (Hunan, China). Transfections were carried out at 100 nM per well according to the manufacturer’s instructions. Experiments were performed 24–48 hours after transfection.

RT-qPCR and TaqMan probe RT-qPCR

Approximately 50 mg of animal brain tissue was homogenized, and total RNA was extracted using the TransZol-Up Enhanced RNA Extraction Kit (ET111-01-V2, Transgene, Beijing, China). RNA from cell samples was extracted with the same kit. cDNA was synthesized by reverse transcription using the PrimeScriptTM RT kit (RR047A; Takara Biomedical Technology, Beijing). RT-qPCR was performed with TB Green® Premix Ex Taq™ II (Tli RNaseH Plus) (Takara, Beijing, China). GAPDH was used as the reference gene, and relative gene expression was calculated by the ^2−ΔΔCT method. The primers used for RT-qPCR are listed in Table 1 [6,33–39]. For precise quantification of viral copy number in infected cells, TaqMan probe RT-qPCR was performed with the EMCV probe 5’-CACTTCGATCACTATGCTTGCCGTT-3’ [40] and the EMCV-3D gene primers listed in Table 1.

Table 1.

Primers used in this study.

	Gene	Forward (5′-3′)	Reverse (5′-3′)
Mouse
	IL-1β	CAGCTCTCTTTAGGAAGACAC	CAAGGAGAAGAAAGTAATGAC
	IL-6	AATGAGGAGACTTGCCTGGT	GCAGGAACTGGATCAGGACT
	CXCL10	GGCCATAGGGAAGCTTGAAAT	TCGTGGCAATGATCTCAACAC
	CCL5	TGCCCACGTCAAGGAGTATTTC	AACCCACTTCTTCTCTGGGTTG
	ZO-1	TGAACGCTCTCATAAGCTTCGTAA	ACCGTACCAACCATCATTCATTG
	Occludin	TGTGGGATAAGGAACACATTTATGA	CAGACACATTTTTAACCCACTCTTCA
	Claudin-5	TCTGCTGGTTCGCCAACAT	CGGCACCGTCGGATCA
	GAPDH	AAGGCCATCACCATCTTCCA	GCCAGTAGACTCCACAACATAC
	EMCV-3D	GTCATACTATCGTCCAGGGACTCTAT	CATCTGTACTCCACACTCTCGAATG
Cell
	IL-1β	CAAAGGCGGCCAGGATATAA	CTAGGGATTGAGTCCACATTCAG
	IL-6	GGAGACTTGCCTGGTGAAA	CTGGCTTGTTCCTCACTACTC
	CXCL10	ACCAAATCAGCTGCTACTACTC	CAGGGTCAGAACATCCACTAAG
	CCL5	CTCCGTCACAACAACAACAAC	AGAGCTCAGAACCTAGAGACTT
	ZO-1	GAAGGCGGGGCCTACACTGAT	GTGGGCTCCTCCAGTCTGACATT
	Occludin	AGACACCCCAAGGTTCCATCCGA	GTGTCGGCCTCCTCCCTCGGT
	Claudin-5	TTAACAGACGGAATGAAGTT	AAGCGAAATCCTCAGTCT
	AKT1	GCTGACGGCCTCAAGAAGCA	ACCTTGCCGAAAGTGCCCTT
	AKT2	CCGCCTGTGCTTTGTGATGG	TTTCCAGCTTGATGTCGCGG
	AKT3	ACAGATGGCTCATTCATAGG	CCCTTTCCTCTGGAGTATCT
	GAPDH	GTCTCCTCTGACTTCAACAGCG	ACCACCCTGTTGCTGTAGCCAA
	EMCV-3D	GTCATACTATCGTCCAGGGACTCTAT	CATCTGTACTCCACACTCTCGAATG

Immunofluorescence assays (IFA)

hCMEC/D3 cells were plated on slides, fixed with 4% paraformaldehyde, and permeabilized with 0.25% Triton X-100. Blocking was carried out in TBST containing 5% bovine serum albumin (BSA). Cells were then incubated sequentially with a primary antibody against the target protein and a fluorescently labeled secondary antibody, each for 1 hour at 37 °C. After incubations, cells were washed with PBS, counterstained with DAPI to label nuclei, and mounted in a Mounting Medium, anti-fading (with DAPI) (S2110, Solarbio, Beijing, China). Imaging was performed on a confocal microscope (LSM 900 with Airyscan 2).

Statistical analysis

Statistical analyses were performed using GraphPad Prism 10.4 (GraphPad Software, San Diego, California, USA). Student’s t-test was used for comparisons between two groups, and one-way analysis of variance was applied for comparisons among multiple groups according to the experimental design. Statistical significance in figures is indicated by asterisks (*** indicates p < 0.001, ** indicates p < 0.01, and * indicates p < 0.05), with p < 0.05 considered the threshold for significance.

Results

EMCV infection leads to encephalitis in mice

To assess EMCV infection and its pathological effects on brain tissue, we infected BALB/c mice with EMCV and collected samples at 3 and 5 dpi (Figure 1(A)). We assessed viral loads, brain inflammatory responses, and the expression and secretion of inflammation-related factors. Viral RNA copy number and viral titer increased progressively over time, confirming active EMCV replication and proliferation in mouse brain tissue (Figure 1(B,C)). Histopathological analysis of brain sections at 5 dpi revealed meningeal thickening with inflammatory exudate (Figure 1(D), c and d), venular lymphocytic cuffing (Figure 1(D), g and h), multiple hippocampal hemorrhagic foci (Figure 1(D), k, l, and o), and neurophagy (Figure 1(D), d and p), as indicated by arrows with different colors.

Figure 2.

EMCV infection impairs the integrity of the BBB in mice.

(A) Evaluation of BBB permeability in mice using EB extravasation assay. Mice were anesthetized at different days (0, 1, 3, 5) post EMCV inoculation, followed by tail vein injection of EB solution and subsequent execution of the EB extravasation assay. After euthanasia through excessive inhalation of anesthetic, mouse brain tissues were harvested and photographed (left panel). Subsequently, the brain tissues were minced and incubated with formamide for dye extraction, and the absorbance of different samples at the wavelength of 620 nm was measured using a microplate reader. Using standard curves and formulas, the content of EB in brain tissue was quantified and statistical analysis was conducted (right panel). (B) Evaluation of BBB permeability in mice using sodium fluorescein permeability assay. Mice in different groups were anesthetized at 0 dpi, 1 dpi, 3 dpi, 5 dpi, and administered sodium fluorescein solution via tail vein injection. Twenty minutes later, Imaging was performed with an AniView Phoenix full‑spectrum in vivo imaging system in vivo imaging was performed using a full-spectrum in vivo imaging system. (C) Detection of the transcriptional levels of TJ proteins in mouse brains via RT-qPCR. Total RNA was extracted from mouse brain tissues and subjected to reverse transcription. Quantitative PCR was then conducted using specific primers targeting Claudin-5, Occludin and ZO-1, with GAPDH serving as the reference gene for data normalization. (D) Detection of the expression levels of TJ proteins in mouse brains before and after EMCV infection. Mouse brain tissues were rapidly homogenized and lysed, followed by SDS-PAGE. Western blot analysis was performed using specific antibodies against Claudin 5, ZO-1 and Occludin. Tubulin was used as the loading control for gray value analysis. Data are presented as mean ± SEM. Statistical analysis was performed using Student’s t-test. *p < 0.05, **p < 0.01, ns means Not Significant.

Figure 1.

EMCV infects mouse brain tissue and induces encephalitis.

(A) Schematic diagram of the animal experimental design. Mice were inoculated with EMCV at a dose of 100 TCID₅₀. Blood samples and brains were harvested after mice were euthanized through excessive inhalation of anesthetic. (B) Detection of viral RNA copies in mouse brain tissues post EMCV inoculation. Mouse brains were homogenized for viral RNA extraction, and TaqMan probe-based RT-qPCR was performed to quantify the copies of the EMCV 3D gene. (C) Viral titer analysis in mouse brain tissues post EMCV inoculation. 100 mg of mouse brain tissue was suspended in 1 mL of PBS, subjected to homogenization, centrifugation and filtration sequentially, and the viral titer was determined using the TCID₅₀ method. (D) Pathological changes in mouse brain tissues at 5 dpi. Mouse brain tissues were fixed with 4% paraformaldehyde, followed by sectioning and H&E staining. Black arrows: Meningeal thickening and inflammatory cell infiltration; Red arrows: “Perivascular cuffing” phenomenon; Yellow arrows: Hemorrhagic foci in the hippocampus; Blue arrows: Neuronophagia. Scale bar = 50 µm. (E) Detection of mRNA levels of cytokines including IL-1β, IL-6, CXCL10 and CCL5 in mouse brain tissues. Total RNA was extracted from brain tissues, and RT-qPCR was conducted to detect the mRNA levels of the four cytokines at 0, 3 and 5 dpi, with GAPDH as the reference gene for data normalization. (F) Quantitative analysis of IL-1β, IL-6, CXCL10 and CCL5 in mouse serum. Commercial ELISA kits were used to measure the concentrations of IL-1β, IL-6, CXCL10 and CCL5 in serum samples collected before EMCV infection and at 5 dpi. Data are presented as mean ± SEM (n = 4). Statistical analysis was performed using Student’s t-test. *p < 0.05, **p < 0.01, ***p < 0.001, ns means Not Significant.

Both interleukin-1β (IL-1β) and interleukin-6 (IL-6) are proinflammatory cytokines whose expression levels reflect the severity of central inflammation after viral infection and the degree of tight-junction disruption in the BBB [41]. Chemokine C-X-C motif ligand 10 (CXCL10) and chemokine C-C motif ligand 5 (CCL5) are chemokines that recruit immune cells and contribute to BBB impairment [42,43]. Therefore, we measured IL-1β, IL-6, CXCL10, and CCL5 mRNA in brain tissue. On 3 dpi, IL-6 and CXCL10 mRNA were significantly elevated relative to the MOCK group, whereas IL-1β and CCL5 exhibited no significant change. By 5 dpi, all four transcripts were dramaticlly upregulated (Figure 1(E)). ELISA quantification confirmed a marked increase in serum concentrations of all four factors at 5 dpi (Figure 1(F)). These results support that EMCV effectively invades mouse brain tissue and induces viral encephalitis.

EMCV damages the integrity and permeability of the BBB in mice

To characterize dynamic BBB disruption in mice after EMCV infection, we anesthetized animals at 1, 3, and 5 dpi and then administered EB or sodium fluorescein by tail-vein injection to assess BBB permeability (Figure 2(A,B), above). Under normal conditions, EB, which binds albumin, cannot cross the BBB; when the barrier is compromised, EB traverses into the brain parenchyma and produces staining [44]. We observed progressively stronger EB staining of brain tissue with increasing infection duration. Compared with the MOCK group, EB extravasation in infected brains was significantly elevated at 3 and 5 dpi (Figure 2(A)). The sodium fluorescein assay displayed a similar pattern: relative to the MOCK and 1 dpi groups, fluorescein signal intensity in the brain parenchyma was markedly increased at 3 and 5 dpi (Figure 2(B)). Together, these data indicate that EMCV infection increases BBB permeability in mice.

To determine whether EMCV infection compromises BBB integrity, we characterized the expression of key TJ components. RT-qPCR and western blot were used to quantify mRNA and protein levels of Claudin-5, Occludin, and ZO-1. The analyses detect no significant differences in mRNA abundance of the three TJproteins between EMCV-infected and MOCK groups (Figure 2(C)). In contrast, western blot analysis demonstrated EMCV infection induced a marked reduction of ZO-1 and Occludin at protein level in mouse brains, whereas Claudin-5 remained stable (Figure 2(D)). This selective targeting pattern is consistent with documented viral pathogenesis mechanisms of neurotropic viruses (JEV and ZIKV) [45,46]. These results indicate that EMCV disrupts BBB integrity by decreasing translational or post-translational levels of TJ proteins rather than their transcription, a mechanism consistent with paracellular viral traversal of the BBB.

EMCV disrupts the integrity of the BBB model and induces inflammatory responses

Viruses are thought to enter the central nervous system primarily by crossing the BBB via paracellular, transcellular, and “Trojan horse” routes [1]. Our data illustrate that EMCV replicates and proliferates in hCMEC/D3 and U251 (Figure 3(A)). Combined with the results in Figure 2, this finding implies that EMCV may traverse the BBB by multiple routes, including transcellular and paracellular pathways. To probe this possibility, we established an in vitro BBB model in a Transwell system using hCMEC/D3 and U251 cells for subsequent experiments (Figure 3(B)). The barrier properties of this model were assessed by measuring TEER and FITC-Dextran permeability. Compared with the blank control, the BBB model showed a substantial TEER (Figure 3(C)) and selective permeability (Figure 3(D)), meeting the barrier-function requirements for the subsequent experiments [47].

Figure 3.

EMCV crosses the in vitro BBB cell model and upregulates the expression of inflammatory cytokines.

(A) Proliferation analysis of EMCV in hCMEC/D3 and U251 cells. hCMEC/D3 and U251 cells were respectively infected with EMCV at a MOI of 0.0001. At 12, 24, 36, 48, 60 and 72 hpi, viral titers were determined using the TCID₅₀ method. (B) Schematic diagram of the dual-culture BBB cell model. Two days prior to the initiation of co-culture, U251 cells were seeded onto the outer surface of the filter membrane of inverted transwell inserts at a density of 1.5 × 10⁴ cells/cm². After 48 hours of culture, the inserts were placed back into 12-well plates to form a two-chamber system. Subsequently, hCMEC/D3 cells were seeded onto the inner surface of the filter membrane in the upper chamber at a density of 1.5 × 10⁴ cells/cm², and the co-culture was maintained for 3 days. Meanwhile, transwell inserts without cell seeding were set as blank controls. (C) Measurement of TEER of the BBB cell model. At 3 days of co-culture, the TEER values of the Transwell system were measured using a sterilized R/V Meter of Epithelium. (D) Fluorescein permeability assay of the BBB cell model. FITC-Dextran was added to the apical chamber of the Transwell to a final concentration of 1 mg/mL. After incubation, the culture media in both the apical and basal chambers were collected separately, and the fluorescence intensity was measured for the calculation of apparent permeability coefficient (P_app) and subsequent statistical analysis. (E) Effect of EMCV infection on the TEER values of the BBB cell model. The BBB cell model was infected with EMCV at an MOI of 1, and the TEER values were detected and statistically analyzed at 0, 12 and 24 hpi. (F) Effect of EMCV infection on the permeability of the BBB cell model. At 24 hpi with EMCV (MOI = 1), the changes in the P_app of FITC-Dextran across the BBB cell model were analyzed. (G and H) Analysis of the BBB-crossing capacity of EMCV in thein vitromodel. The apical chamber was inoculated with EMCV at different MOIs (0.01, 0.1, 1). At 24 hpi, the viral copy numbers in the basal chamber were detected using TaqMan probe-based RT-qPCR (G). When the apical chamber was inoculated with EMCV at an MOI of 0.1, the viral titers in the basal chamber were determined using the TCID₅₀ method at different time points (24, 48, 72 hpi) (H). (I) EMCV upregulates the expression of inflammatory cytokines in hCMEC/D3 cells. EMCV-infected hCMEC/D3 cells were collected from BBB cell model. RT-qPCR was performed to detect the mRNA expression levels of IL-1β, IL-6, CXCL10 and CCL5 in the mock-infected group and EMCV-infected groups with different MOIs. The data were normalized against the reference gene GAPDH. (J) EMCV downregulates the expression of TJ proteins in the BBB cell model. Western blot analysis was used to detect the protein expression levels of ZO-1 and Occludin in the BBB cell model at 12, 24 and 36 hpi with EMCV, with β-tubulin serving as the loading control. Data are presented as mean ± SEM. Statistical analysis was performed using Student’s t-test. *p < 0.05, **p < 0.01, ***p < 0.001, ns means Not Significant.

Subsequently, we assessed EMCV’s impact on BBB model by measuring TEER and conducting FITC-dextran permeability assays. Compared with the MOCK group, TEER decreased significantly and FITC-dextran permeability increased markedly in the EMCV group (Figure 3(E,F)), indicating that EMCV infection alters barrier permeability. To confirm viral transit across the BBB, we quantified viral copy number and titer on the apical and basal sides after inoculation. As shown in Figure 3(G), viral copy number on the basal side correlated positively with the dosage of the virus inoculation. Besides, the basal-side viral titer rose progressively with infection duration (when at 0.1 MOI) (Figure 3(H)), demonstrating efficient traversal of EMCV across the BBB cell model.

After EMCV infection in BBB cell model, hCMEC/D3 cells were collected for RT-qPCR assa. The mRNA levels of the proinflammatory cytokines IL-1β and IL-6 and the chemokines CXCL10 and CCL5 were significantly increased in the infection model (Figure 3(I)). Western blot analysis indicated that, during EMCV infection, the TJ proteins ZO-1 and Occludin in hCMEC/D3 cells progressively declined (Figure 3(J)). Notably, Claudin-5 was undetectable under identical experimental conditions, which is likely attributable to the inherent characteristics of this cell line [48].

EMCV infection regulates the expression of TJ proteins

Because brain microvascular endothelial cells form the primary structural component of the BBB [2], we explored how EMCV modulates TJ proteins expression in the hCMEC/D3 cell line. Initially, hCMEC/D3 cells were infected with EMCV at 0.01, 0.1, and 1 MOI, and the expression levels of TJ proteins were measured by RT-qPCR and Western blot, respectively. The mRNA levels of Claudin-5, Occludin, and ZO-1 did not change across infection doses (Figure 4(A)). However, as the MOI increases, the protein levels of Occludin and ZO-1 gradually decrease in a dose-dependent manner (Figure 4(B)). This finding aligns with previous in vivo observations and indicates that EMCV affects TJ proteins post-transcriptionally rather than at the mRNA level. Meanwhile, as infection duration increased, the expressions of ZO-1 and Occludin proteins gradually decreased from 2 hpi (Figure 4(C)). IFA assay further revealed a pronounced loss of ZO-1 signal in EMCV-positive cells (red), as indicated by reduced green fluorescence intensity (Figure 4(D)), providing visual evidence that viral infection compromises BBB structural integrity.

Figure 4.

EMCV infection regulates the expression of TJ proteins in hCMEC/D3 cells.

(A) Effect of EMCV infection on the mRNA expression of TJ proteins. hCMEC/D3 cells were infected with EMCV at varying MOI (0.01, 0.1, 1). RT-qPCR was performed to detect the mRNA levels of Claudin-5, Occludin and ZO-1, with GAPDH as the reference gene for data normalization. (B) Effect of different doses of EMCV on the expression of TJ proteins in hCMEC/D3 cells. At 24 hpi with EMCV at varying MOIs (0.01, 0.1, 1), Western blot analysis was conducted using specific antibodies to determine the protein expression levels of VP1, Occludin and ZO-1, respectively. Tubulin was used as the loading control for gray value analysis. (C) Regulatory effect of EMCV infection on TJ proteins expression at different time points. hCMEC/D3 cells were inoculated with EMCV at an MOI of 1. Cells were harvested at 0, 1, 2, 4, 8, 12 and 24 hpi, subjected to lysis, and the expression levels of relevant proteins were detected by Western blot. (D) Analysis of the effect of EMCV infection on TJ protein expression by IFA assay. Immunofluorescence staining was performed using specific primary antibodies against viral protein VP1 (red) and ZO-1 (green) as well as the corresponding fluorescent secondary antibodies. Cell nuclei were labeled with DAPI (blue). Fluorescent signals were captured using a laser scanning confocal microscopy system. Scale bar is 10 µm. Data are presented as mean ± SEM. Statistical analysis was performed using Student’s t-test. ns means Not Significant.

The degradation of TJ proteins depends on AKT during EMCV infection

The PI3K/AKT pathway is essential for brain development and regulates neuronal proliferation, differentiation, autophagy, and apoptosis [49,50]. Dysregulation of this pathway is closely linked to the pathogenesis of several central nervous system disorders, including cerebral hemorrhage and stroke [49,50]. Prior studies indicate that PI3K/AKT signaling can affect TJ proteins expression, but the precise molecular mechanisms remain insufficiently defined. We therefore investigated whether PI3K/AKT signaling participates in EMCV infection and in the regulation of TJ proteins. First, we examined temporal changes in related proteins after EMCV infection of hCMEC/D3 cells. As shown in Figure 5(A), total PI3K levels remained stable after infection, while phosphorylated PI3K (pPI3K) increased at a certain degree. Time-course analysis revealed that EMCV infection induced a biphasic modulation of AKT phosphorylation: an early transient activation (2 to 4 hpi) followed by marked suppression (8 to 24 hpi) (Figure 5(A)). These findings indicate that EMCV infection actively modulates PI3K/AKT signaling, characterized by an initial surge and subsequent decline in AKT phosphorylation.

Figure 5.

EMCV promotes TJ protein degradation by inhibiting the AKT signaling pathway.

(A) Regulatory effect of EMCV infection on the PI3K/AKT signaling pathway. hCMEC/D3 cells were seeded in 6-well plates and infected with EMCV (1 MOI) when the cell confluence reached 90%. Samples were collected at 0, 1, 2, 4, 8, 12 and 24 hpi. Western blot was performed to detect the expression levels of PI3K, phosphorylated PI3K (pPI3K), AKT, phosphorylated AKT (pAKT) and VP1, with β-tubulin serving as the loading control. (B) Evaluation of the effect of LY294002 at different concentrations (1, 5, 10 and 20 μM) on the viability of hCMEC/D3 cells using the CCK-8 kit. DMSO, the solvent of the drug, was used as the negative control. (C) Effects of LY294002 on AKT activation, TJ proteins expression and EMCV replication. hCMEC/D3 cells were pretreated with 10 μM LY294002 for 2 hours prior to infection with EMCV (1 MOI). At 8, 12 and 24 hpi, Western blot was conducted to detect the expression levels of AKT, pAKT, VP1, ZO-1 and Occludin, respectively. (D) Detection of the effect of Afuresertib at different concentrations (0.1, 1, 5, 10, 20 and 30 μM) on the viability of hCMEC/D3 cells using the CCK-8 kit. (E) Effect of Afuresertib on TJ proteins expression. After cells were treated with Afuresertib at different concentrations, Western blot was performed to measure the protein levels of AKT, ZO-1 and Occludin, with β-tubulin as the loading control. Band gray value analysis was carried out using ImageJ software. (F and G) Role of AKT isoforms in regulating TJs. hCMEC/D3 cells were seeded in 6-well plates and then transfected with siRNAs targeting AKT1, AKT2 and AKT3, respectively. RT-qPCR (F) and Western blot (G) were used to verify the knockdown efficiency, and Western blot was simultaneously performed to detect the protein levels of ZO-1 and Occludin. (H) Analysis of the degradation pathway of TJ proteins. hCMEC/D3 cells were infected with EMCV (1 MOI) and then treated with MG132 (20 μM), CQ (20 μM) and Z-VAD-FMK (20 μM), respectively. Western blot was used to detect the protein levels of ZO-1 and Occludin, with β-tubulin as the loading control. (I) Effect of inhibitors targeting different degradation pathways on Afuresertib-induced TJ proteins degradation. hCMEC/D3 cells were treated with 20 μM Afuresertib alone or in combination with MG132/CQ/Z-VAD-FMK. Western blot was performed to detect the protein levels of ZO-1 and Occludin. Data are presented as mean ± SEM. Statistical analysis was performed using Student’s t-test. *p < 0.05, **p < 0.01, ***p < 0.001, ns means Not Significant.

To explore whether the AKT family participates in regulating TJ protein expression during EMCV infection, hCMEC/D3 cells were pretreated with the broad‑spectrum PI3K/AKT pathway inhibitor LY294002 or DMSO control prior to EMCV infection and harvested at 0, 8, 12, and 24 hpi for Western blot analysis. The non‑cytotoxic concentration (10 μM) established by CCK‑8 assay in Figure 5(B). As shown in Figure 5(C), LY294002 effectively suppressed AKT phosphorylation stimulated by EMCV. Compared with DMSO controls, LY294002‑treated cells exhibited markedly accelerated and exacerbated degradation of ZO‑1 and Occludin. Concomitantly, viral capsid protein VP1 expression was significantly elevated in LY294002‑treated cells at 12, and 24 hpi (Figure 5(C)). Further, after determining a safe working concentration for the pan‑AKT inhibitor Afuresertib (Figure 5(D)), hCMEC/D3 cells were exposed to 0.1, 1, 5, 10, and 20 μM Afuresertib. Analysis of AKT, ZO‑1, and Occludin protein levels revealed a dose‑dependent decline in total AKT accompanied by corresponding dose‑dependent decreases in ZO‑1 and Occludin (Figure 5(E)). These findings indicate that AKT signaling controls the protein levels of TJs components.

The siRNAs targeting genes of AKT1, AKT2, and AKT3, were transfected into hCMEC/D3 cells individually. The results demonstrated that siAKT1, siAKT2, and siAKT3 significantly down-regulated the mRNA levels and protein expression of their respective target genes (Figure 5(F,G)). Notably, compared to siNC, the knockdown of AKT3 led to a significant reduction in the expression levels of ZO-1 and Occludin (Figure 5(G)), indicating that AKT3 may play a predominant role in the regulation of TJs in hCMEC/D3 cells.

To define how EMCV induces TJ proteins loss, we treated hCMEC/D3 cells with inhibitors of three major protein-degradation pathways: the proteasome inhibitor MG132, the autophagy-lysosome inhibitor chloroquine (CQ), and the pan‑caspase inhibitor Z‑VAD‑FMK to assessed TJ proteins levels in EMCV-infected cells. Figure 5(H) illustrates that the expression of ZO‑1 and occludin were restored by CQ and Z‑VAD‑FMK, but not MG132. Previous studies have shown that AKT can suppress both autophagy and apoptosis under certain conditions [51–53]. Here, Afuresertib treatment caused a marked reduction in occludin and ZO‑1 compared with the blank and DMSO controls, while CQ and Z‑VAD‑FMK rescued occludin and ZO‑1 to levels comparable to the blank group (Figure 5(I)). Together, these data indicate that AKT regulates TJ proteins stability in hCMEC/D3 cells via autophagy- and apoptosis-dependent mechanisms rather than the proteasomal pathway.

AKT3 facilitates the degradation of TJ proteins via the autophagy pathway

Autophagy and apoptosis represent distinct yet intricately interconnected forms of programmed cell death. To clarify their regulatory relationship during EMCV infection, we examined the temporal dynamics of both pathways and performed pharmacological intervention studies. Western blot analysis of EMCV-infected hCMEC/D3 cells revealed time-dependent activation of autophagy, apoptosis, and viral replication. Autophagy initiation was evidenced from 1 hpi onward, as shown by progressive upregulation of Beclin-1 and transient accumulation followed by degradation of p62 (Figure 6(A)). Apoptotic signaling emerged subsequently, with Bax upregulation apparent from 4 hpi and marked consumption of total caspase-3 from 8 hpi (Figure 6(A)), consistent with extensive zymogen activation. Concomitantly, AKT3 protein levels declined progressively from 4 hpi and were markedly reduced from 8 to 12 hpi. Viral capsid protein VP1 became detectable at 8 hpi and accumulated substantially at 12 and 24 hpi (Figure 6(A)), indicating that active viral replication followed the establishment of autophagic and apoptotic stress responses.

Figure 6.

EMCV degrades TJ proteins via the AKT3-dependent autophagic pathway.

(A) Temporal dynamics of EMCV-induced changes of AKT3 and cellular stress responses. hCMEC/D3 cells were inoculated with EMCV at an MOI of 1. Cells were harvested at 0, 1, 2, 4, 8, 12 and 24 hpi, subjected to lysis, and the expression levels of relevant proteins were detected by Western blot. (B) Effect of CQ on apoptosis during EMCV infection. Western blot analysis of the protein levels of VP1, AKT3, P62, Beclin-1, caspase-3, Bax, with β-tubulin as the loading control. (C) Effect of AKT3 knockdown on EMCV-induced autophagic flux. hCMEC/D3 cells were transfected with siNC or siAKT3 prior to infection with EMCV at a MOI of 1. Cells were harvested at 24 hpi, and Western blot was performed to detect the protein levels of Beclin-1, p62, ZO-1 and Occludin. β-tubulin was used as the loading control. (D to F) Effect of CQ on EMCV-induced BBB dysfunction. Thein vitroBBB cell model was treated with CQ upon EMCV infection. At 24 hpi, TEER measurement (D) and FITC-dextran permeability assay (E) were performed respectively, and the P_appof FITC-dextran across the BBB model was analyzed. Meanwhile, mock-infected and DMSO-treated groups were set as controls. (F) Western blot analysis of the protein levels of VP1, Occludin and ZO-1, with β-tubulin as the loading control for gray value analysis. (G) Determination of viral titers in the basal chamber of the BBB model using the TCID₅₀ method. Data are presented as mean ± SEM. Statistical analysis was performed using Student’s t-test. *p < 0.05, **p < 0.01, ***p < 0.001, ns means Not Significant.

To determine whether autophagy modulates apoptotic execution during EMCV infection, hCMEC/D3 cells were infected with EMCV in the presence or absence of CQ. As shown in Figure 6(B), EMCV infection upregulated Bax and downregulated Caspase-3, consistent with Figure 6(A). Compared with EMCV-only and DMSO-treated controls, CQ-treated EMCV-infected hCMEC/D3 cells exhibited further enhanced Bax upregulation and more pronounced Caspase-3 downregulation, accompanied by severely diminished VP1 expression (Figure 6(B)). These results suggest that CQ treatment exacerbates apoptotic activation while suppression of autophagy inhibits viral replication.

AKT3 is predominantly expressed in brain and testis and is essential for neuronal survival, brain development, and reproductive function. Dysregulation of AKT3 is closely linked to neurological disorders such as Alzheimer’s disease and glioblastoma [18,51,54,55]. To determine whether AKT3 modulates EMCV-activated autophagy to regulate TJ protein expression during infection, hCMEC/D3 cells were transfected with siAKT3 or siNC and subsequently infected with EMCV. As presented in Figure 6(C), EMCV infection markedly activated autophagy, while AKT3 knockdown further enhanced EMCV-induced autophagic activity, manifested by a more pronounced reduction in p62 and elevation of Beclin-1. Concurrently, AKT3 silencing exacerbated the degradation of ZO-1 and occludin, and the expression of viral capsid protein VP1 (Figure 6(C)). Collectively, these results indicate that AKT3 negatively regulates EMCV-induced autophagy, and loss of AKT3 aggravates autophagy-mediated TJ protein degradation while facilitating viral replication.

To assess how autophagy activation affects EMCV-induced BBB disruption, CQ was added to the in vitro BBB model after EMCV infection. TEER measurements revealed that CQ substantially attenuated the EMCV-induced decline in barrier resistance (Figure 6(D)), and FITC-dextran permeability was reduced to a level not significantly different from the MOCK group (Figure 6(E)). Western blot demonstrated that CQ inhibited EMCV-triggered degradation of Occludin and ZO-1 and correspondingly reduced EMCV VP1 expression (Figure 6(F)). Consistent with these results, the viral titer on the basal side of the Transwell chamber was significantly decreased (Figure 6(G)). Together, these data indicate that EMCV may suppress the autophagy pathway by downregulating AKT3 expression and phosphorylation, thereby accelerating TJ proteins degradation and impairing BBB integrity.

EMCV degrades TJ proteins via the AKT3-mediated apoptotic pathway

hCMEC/D3 cells were transfected with siAKT3 or siNC and infected with EMCV. Western blot analysis showed that siAKT3 alone reduced Bcl-2 and increased Bax expression, while EMCV infection further amplified this pro-apoptotic shift (Figure 7(A)). The combined treatment (siAKT3 + EMCV) produced the most pronounced elevation in the Bax/Bcl-2 ratio and near-complete depletion of total Caspase-3 (Figure 7(A)). Concomitantly, ZO-1 and Occludin were markedly degraded, and VP1 expression was significantly enhanced, indicating that AKT3 suppression creates a cellular environment permissive for both apoptotic execution and viral amplification.

Figure 7.

EMCV degrades TJ proteins via the AKT3-dependent apoptotic pathway.

(A) Effect of AKT3 knockdown on apoptotic proteins. hCMEC/D3 cells were transfected with siNC or siAKT3prior to infection with EMCV at a MOI of 1. Cells were harvested at 24 hpi, and Western blot was performed to detect the protein levels of Caspase-3, Caspase-8, Caspase-9, ZO-1 and Occludin. β-tubulin was used as the loading control, and band intensity analysis was conducted using ImageJ software. (B) Evaluation of the effect of inhibitors at different concentrations on the viability of hCMEC/D3 cells using the CCK-8 assay kit, with the DMSO-treated group set as the control. (C) Effect of different Caspase inhibitors on EMCV-induced TJ proteins degradation. hCMEC/D3 cells were treated with Ac-DEVD-CHO, Z-IETD-FMK and Z-LEHD-FMK TFA respectively upon infection with EMCV (MOI = 1). Cells were harvested at 24 hpi for Western blot analysis. (D) Effect of different Caspase inhibitors on LPS-induced TJ proteins degradation. hCMEC/D3 cells were sequentially treated with LPS and/or Ac-DEVD-CHO, Z-IETD-FMK and Z-LEHD-FMK TFA, followed by Western blot to detect the protein expression levels of ZO-1 and Occludin. (E and F) Effect of Z-VAD-FMK on EMCV-induced BBB dysfunction. Thein vitroBBB cell model was treated with Z-VAD-FMK upon EMCV infection. At 24 hpi, TEER measurement (E, left) and FITC-dextran permeability assay (E, right) were performed respectively, and the P_app of FITC-dextran across the BBB model was analyzed. Meanwhile, mock-infected and DMSO-treated groups were set as controls. Western blot was performed to analyze the protein levels of VP1, Occludin and ZO-1, with β-tubulin serving as the loading control (F). (G) Determination of viral titers in the basal chamber of the BBB model using the TCID₅₀ method. Data are presented as mean ± SEM. Statistical analysis was performed using Student’s t-test. *p < 0.05, **p < 0.01, ***p < 0.001, ns means Not Significant.

To further identify the principal caspases mediating this response, we first established optimal concentrations of three inhibitors Ac-DEVD-CHO, Z-IETD-FMK, and Z-LEHD-FMK TFA, targeting Caspase-3, -8, and -9, respectively, using CCK-8 assay kit (Figure 7(B)). We then treated EMCV-infected hCMEC/D3 cells with each inhibitor separately and analyzed TJ proteins by Western blot. All three inhibitors rescued TJ proteins expression, with Z-IETD-FMK producing the strongest restoration (Figure 7(C)). To test whether this effect generalizes beyond EMCV, we treated hCMEC/D3 cells with the broad-spectrum apoptosis inducer Lipopolysaccharide (LPS). As shown in Figure 7(D), LPS markedly induced degradation of ZO-1 and occludin, and all three inhibitors attenuated this degradation. Notably, the Caspase-8 inhibitor Z-IETD-FMK provided the most robust protection of TJ proteins, implicating the extrinsic apoptotic pathway as the dominant route for TJ proteins degradation.

Subsequently, we treated the in vitro BBB model with Z-VAD-FMK and observed that it significantly mitigated the EMCV-induced drop in TEER value (Figure 7(E), left) and the rise in FITC-dextran permeability (Figure 7(E), right). At the molecular level, Z-VAD-FMK restored Occludin and ZO-1 protein levels and reduced expression of the viral capsid protein VP1 (Figure 7(F)). Viral titer in the Transwell basal chamber was also significantly decreased (Figure 7(G)), indicating that the ability of EMCV to cross BBB was weakened after apoptosis was inhibited.

Overall, EMCV infection downregulates AKT3, sequentially relieving its suppressive effects on autophagy and apoptosis. Autophagy activation precedes apoptotic execution, as evidenced by time-course marker analysis. The additive protective effect of chloroquine and Z-VAD-FMK on tight junction protein levels demonstrates that these pathways are functionally non-redundant and coordinately contribute to BBB disruption. Ultimately, convergent pathway activation enables EMCV to cross the compromised blood-brain barrier (Figure 8).

Figure 8.

A schematic diagram depicts the mechanism of the EMCV disrupt BBB integrity.

EMCV infection triggers sequential cellular responses: autophagy is activated at approximately 2 hpi, followed by AKT3 downregulation and apoptotic activation at approximately 4 hpi. AKT3 suppression relieves its inhibitory constraints on both pathways, exacerbating tight junction protein (ZO-1 and occludin) degradation, compromising BBB integrity, and ultimately facilitating EMCV transmigration across the blood-brain barrier. Green arrows indicate EMCV-mediated effects. Chloroquine (CQ) and Z-VAD-FMK (red inhibitory lines) block autophagy and apoptosis, respectively, thereby preserving tight junction protein levels, attenuating BBB disruption, and impeding EMCV neuroinvasion.

Discussion

BBB protects the central nervous system from pathogens, including neurotropic viruses. Although human EMCV infection is typically mild, the virus induces encephalitis-related pathology in experimental animals [5]. As a member of the Picornaviridae, EMCV serves as a valuable model for studying viral encephalitis. This research investigates how EMCV breaches the BBB by disrupting TJ integrity. Our findings demonstrate that EMCV suppresses AKT signaling during late infection, thereby relieving inhibition on autophagy and apoptosis. The resulting activation of these pathways degrades TJ proteins and facilitates viral neuroinvasion.

This study systematically identified EMCV replication in the mouse brain and analyzed the inflammatory response of brain tissue through pathological observations and cytokines detection. Additionally, the study confirmed the virus’s detrimental effect on BBB permeability in mice using EB staining and the sodium fluorescein permeability assay. EMCV markedly reduced TJ proteins (ZO-1 and Occludin) at the post-transcriptional level, indicating that viral disruption of BBB integrity occurs primarily through translational or degradative mechanisms rather than transcriptional repression. An important observation was the differential vulnerability of TJ proteins to EMCV infection: ZO-1 and Occludin were consistently downregulated in both mouse brain tissues and hCMEC/D3 cells, whereas Claudin-5 remained unchanged in vivo and was undetectable in vitro. This selective targeting pattern aligns with established mechanisms of neurotropic viral pathogenesis. Chen et al. demonstrated that JEV selectively disrupts ZO-1 in brain microvascular endothelial cells while sparing claudin-1, claudin-5, and occludin [45]. Similarly, Zika virus (ZIKV) differentially regulates TJ proteins in a strain-dependent manner, downregulating occludin and claudin-5 while paradoxically upregulating ZO-1 under certain conditions [46]. Although Claudin-5 is not a primary target of EMCV, our study revealed that this protein is hard to detected in hCMEC/D3 cells—consistent with published reports showing >4-fold lower abundance compared with primary brain microvessels [56,57]. Collectively, these findings highlight that neurotropic viruses have evolved sophisticated strategies to selectively dismantle barrier integrity by targeting specific junctional components while sparing others.

EMCV can replicate in hCMEC/D3 and U251 cells, suggesting its capability, akin to members of the Flaviviridae family [8], to breach the BBB and access the CNS through a transcellular pathway. EMCV could migrate from the apical to the basal side of the BBB cell model, accompanied by a significant decrease in TJ proteins (ZO-1 and Occludin) in hCMEC/D3 cells. These results are consistent with the outcomes of animal experiments, indicating that this model could partially substitute for animal trials, thereby cutting costs and shortening the experimental timeline. Moreover, the reduction of TJ proteins in hCMEC/D3 cells by EMCV is dependent on the dosage, and the impairment to TJ proteins worsens over time with prolonged infection.

The AKT signaling pathway is critical in viral infection and the regulation of the BBB [58–60]. Viruses frequently “antagonize” or “exploit” the AKT pathway to establish a conducive environment for their replication. For example, influenza viruses and porcine coronaviruses activate the AKT pathway to enhance their entry and replication [26,61]. Meanwhile, some research indicates that AKT contributes to stabilizing the TJs structure of the BBB and preserving its barrier function [53]. However, during EMCV infection, the AKT signaling pathway presented as “early-stage activation and later-stage down-regulation.” AKT pathway inhibitor experimentation revealed a significant intensification in the degradation of TJ proteins (occludin and ZO-1) following AKT inhibitor treatment. Furthermore, experiments on the knockdown of the three AKT family subtypes discovered the pivotal role of AKT3 in TJs protein degradation in hCMEC/D3 cells. This subtype-specific functionality offers a fresh perspective on the involvement of the AKT pathway in virus-host interactions.

Autophagy and apoptosis represent vital yet interconnected cellular responses to viral infections [62,63], and accumulating evidence indicates that these pathways can be co-regulated by a common upstream signaling node to execute coordinated cellular outcomes [64–68]. In this study, EMCV infection downregulated AKT3 and sequentially activated both pathways: early autophagy initiation (2–8 hpi) was evidenced by increased Beclin-1 and reduced p62, followed by apoptotic priming and execution (>8 hpi) as shown by elevated Bax/Bcl-2 ratios and diminished full-length Caspase-3. Rather than prolonged suppression of cell death, EMCV appears to induce transient early autophagy supporting viral RNA translation and progeny assembly, followed by AKT3-dependent apoptotic execution to facilitate progeny release. To clarify the virological logic, EMCV suppression of AKT3 is not a uniform pro-replication or pro-death signal, but a temporally partitioned antagonism. In the early phase (2–8 hpi), AKT3 downregulation relieves mTOR-mediated suppression of autophagy, generating membrane scaffolds and metabolic substrates for viral RNA translation. In the late phase (>8 hpi), continued AKT3 suppression permits apoptotic execution, which—while terminating replication in individual cells—facilitates progeny release and neuroinvasion across the BBB. Thus, EMCV sequentially desuppresses AKT3-controlled pathways to first exploit autophagy for replicative advantage, then tolerate apoptosis for dissemination. This trade-off is consistent with the lytic life cycle of picornaviruses, where cell death is the inevitable cost of progeny escape.

Consistently, AKT3 knockdown and pan-AKT inhibitor (Afuresertib) augmented these sequential changes and exacerbated TJ protein loss, indicating that AKT3 normally suppresses autophagy and apoptosis to maintain barrier integrity. This dual regulatory capacity of AKT3 is exemplified in diverse pathological contexts: miR-20b-5p targets AKT3 to activate autophagy and modulate apoptosis in acute pancreatitis [64]; SARS-CoV-2 spike protein suppresses PI3K/AKT signaling to trigger both pathways [65]; and miRNA-145 activates AKT3 to inhibit autophagy and apoptosis following myocardial infarction [67]. In our study, time-course analysis revealed that autophagy activation preceded apoptotic execution and viral replication, suggesting that EMCV-induced autophagy may initially serve a proviral function. This interpretation is supported by CQ treatment data: blockade of autophagic flux not only exacerbated Caspase-3 cleavage but also markedly reduced VP1 expression, indicating that autophagy contributes to viral protein synthesis or progeny maturation. The concomitant enhancement of apoptosis upon autophagy inhibition may reflect cellular stress responses to impaired lysosomal function or compensatory activation of alternative death pathways [69]. Collectively, these findings suggest that EMCV orchestrates a temporally ordered exploitation of autophagy and apoptosis—utilizing early autophagy for replicative advantage while tolerating late apoptotic execution for dissemination—coordinated through AKT3 suppression.

An unexpected observation was that Afuresertib, an ATP-competitive AKT kinase inhibitor, dose-dependently reduced total AKT protein. Although AKT inhibitors primarily target kinase activity, effects on total AKT abundance have been reported [70,71]. Figure 5(I) demonstrated that MG132, CQ, and Z-VAD-FMK all failed to rescue total AKT levels, effectively ruling out enhanced protein degradation. We posit that sustained AKT inhibition attenuates mTORC1-dependent cap-dependent translation, thereby reducing de novo AKT synthesis [72–75].

Following EMCV infection, IL-1β, IL-6, CXCL10, and CCL5 expressions were significantly increased in mouse brain tissues and cells, aligning with the inflammatory response seen in neurotropic viral infections [13–15]. These well-known pro-inflammatory factors can reduce TJ proteins expression by activating pathways like NF-κB and JAK/STAT3, potentially worsening BBB damage through immune cell chemotaxis [38,42]. This implies that EMCV might indirectly impact BBB integrity via cytokine activity. However, this study did not extensively explore this aspect.

We acknowledge that the precise phosphorylation substrates of AKT3 in brain endothelial cells—whether Bcl-2, Beclin-1, or occludin—remain unidentified. Furthermore, the absence of phospho-AKT3–specific antibodies and the undetectability of LC3B in hCMEC/D3 cells preclude direct mechanistic atomism at this stage. Nevertheless, this study establishes AKT3 as a non-redundant, isoform-specific temporal coordinator of autophagy-apoptosis crosstalk at the BBB. Future phosphoproteomic and proximity labeling studies will map the direct molecular targets mediating this compartment-specific regulation. Besides, AKT3 genetic manipulation was not validated at the animal level. While our in vivo data demonstrate EMCV-induced BBB breakdown and TJ protein loss in mouse brains, these observations do not establish that endothelial AKT3 mediates these effects cell-autonomously. AKT3 global knockout is associated with developmental phenotypes, necessitating the use of endothelial-specific AKT3 conditional knockout mice in future studies to bridge this gap.

The near-absence of LC3B signal in hCMEC/D3 cells aligns with established evidence that normal brain endothelial cells poorly express this isoform [75,76]. While we employed p62 and Beclin-1 as alternative autophagy readouts, we acknowledge that the potential contribution of LC3A to autophagy in this cell line was not assessed. Future studies utilizing LC3A-specific antibodies will be necessary to fully delineate the LC3 isoform dependency of brain endothelial autophagy.

In conclusion, our findings delineate an EMCV-AKT3-autophagy/apoptosis-TJ degradation-BBB disruption axis that governs viral neuroinvasion. We identify AKT3 as an isoform-specific suppressor of autophagy and apoptosis in EMCV-infected hCMEC/D3 cell, whose sequential downregulation first enhances autophagy activation, then apoptotic execution, resulting in coordinated TJ protein degradation. This study provides an in vitro conceptual framework for EMCV-induced barrier injury, with direct in vivo validation of endothelial AKT3 function representing a critical future direction .

Funding Statement

This research was supported by Gansu Province Innovation Fund Project for College Teachers [No. 2025B-035], Science and Technology Program of Gansu Province [No. 25YFWA023], National Natural Science Foundation of China [No. 32260037], Talent Introduction Research Projects of Northwest Minzu University [No. xbmuyjrc2020021], and Fundamental Research Funds for the Central Universities [No. 31920250028].

Disclosure statement

Adi Idris is a Director of Research for Intelligene Pty Ltd and holds shares in the company. The other authors declare that they have no competing interests.

Data availability statement

The data that support the findings of this study are openly available in figshare (https://doi.org/10.6084/m9.figshare.30984970).