Introduction
Influenza virus infection is often associated with neurological symptoms; i.e., febrile seizure, meningitis, transverse myelitis, Guillain-Barré syndrome, and encephalopathy/encephalitis [
20,
22,
100]. Of these, influenza-associated encephalopathy (IAE) is one of the most serious neurological complications, often resulting in severe sequelae with high mortality [
11,
63]. Acute encephalopathies, including IAE, had started to be reported in the 1990s mainly in East Asia with pediatric cases [
58,
92]. Recently, the number of reports has increased, especially during and after the 2009 pandemic, from all over the world, not only of children, but also of adults to the elderly [
6,
8,
22,
31,
57,
64,
72,
73,
83].
Various pathogens, mainly viruses, cause this severe neurological complication [
1]. Such cases have been reported as Reye syndrome, influenza encephalitis, and recent neurological complications of coronavirus disease 2019 (COVID-19), which included the broader disease entity of acute encephalopathy and encephalitis related to infection [
14,
16,
18,
30,
78,
91,
94,
96]. The most recent national surveillance in Japan pointed out the changes in responsible pathogens; a decrease in influenza virus along with a wider range of pathogens [
78].
With the increasing availability of magnetic resonance imaging (MRI), acute encephalopathy has been subclassified into several clinico-radiological syndromes such as acute necrotizing encephalopathy (ANE) [
58], hemorrhagic shock and encephalopathy syndrome (HSES) [
31,
50], acute encephalopathy with biphasic seizures and late reduced diffusion (AESD) [
89], and clinically mild encephalitis/encephalopathy with a reversible splenial lesion (MERS) [
87]. Although overall mortality nowadays improved compared to that of twentieth century, there was still no fundamental treatment but supportive care and only 30–50% recovered without sequelae, including mild cases [
1,
11,
22,
42].
The pathology of IAE has been investigated in clinical studies and animal experiments, focusing on host factors and biomarkers. Risk genetic polymorphisms have been also reported in East Asian populations. These include
CPT-2,
RANBP2, SCN1A, and
TLR-3 genes [
7,
28,
33,
38,
47,
71,
79,
81]. Markers of disease severity and treatment efficacy have been examined using physiological, radiological, and biochemical analysis [
3,
27,
32,
34,
35,
39,
43,
51,
61]. Several mechanistic studies were performed using IAE related animal models; Reye syndrome-related animal models, congenital metabolic disease-related models, administration of lipopolysaccharides (LPS) or cytokines, models with neurotropic influenza virus, and others [
5,
10,
12,
15,
36,
52,
66,
90,
93,
95,
98,
101]. Although these models showed brain edema and increased cytokines, the dynamics and pathogenesis of the influenza virus in IAE are still unknown.
Autopsy research has identified IAE-related pathological features [
40,
54,
58,
63,
86,
92,
102]. Brain edema is severe and is sometimes accompanied by herniation. The inflammatory response is inactive, while multiple vascular lesions, such as hemorrhage, bleedings, fibrin thrombosis, and leakage of plasma components, are observed in the brain parenchyma [
44,
49,
67]. It is noteworthy that the virus antigen was detected in some cases in the brain parenchyma and spinal fluid [
19,
23,
88,
99]. The detection of viral antigen or RNA from edematous brains has been also reported with the influenza virus-inoculated mouse models [
80,
101]. Autopsy studies suggested that the virus itself should have a pathological role in IAE, although the virus has hardly been detected in human IAE patients probably due to its rapid and progressive course.
Patients with severe types of IAE develop critical neurological symptoms such as seizures and coma within 10 h after the onset of fever and die within 48 h due to the fatal brain edema [
63,
69]. Brain autopsies revealed fewer inflammatory cells and less detection of the virus, as noted above, in contrast to the severe edema and high fever. In the clinical course of IAE, there is a time window between the viral infection and the brain edema, that is, between the infection and the onset of neurological symptoms. This time window is an important indicator of IAE pathogenesis since it must include the key viral dynamics to trigger the brain edema.
To investigate viral dynamics in this time window, we developed an IAE mouse model and analyzed its pathophysiology over time. Furthermore, we compared and verified the results with the involved cell lines and IAE autopsy brains, to confirm the pathogenesis of the infection-related acute brain edema. Through this analysis, we identified the interventional methods to prevent the onset of IAE.
Materials and methods
Further information can be found in Supplemental Methods.
Virus
Influenza A/Puerto Rico/8/34 (PR8, H1N1) was obtained from ATCC (VR95; Manassas, VA, USA). To propagate IAV, Madin-Darby canine kidney (MDCK) cells (JCRB9029, NIBIOHN) was inoculated with the virus at an MOI of 0.001 and then cultured in virus culture medium (VCM; minimum essential medium (MEM) containing 0.01 M HEPES, 0.21% albumin, and 0.25% trypsin) in a cell culture incubator at 37 °C and 5% CO2 for 72 h. The crude viral lysates of MDCK cells were spun down at 4 °C for 15 min at 4000×g. The viral supernatants were passed through a 0.22 μm filter and centrifuged at 4 °C for 3 h at 45,000×g. The viral pellet was re-suspended in 10 ml of phosphate buffered saline (PBS) and was stored at −80 °C as stock and used for experimental infection in mice and in vitro. The titer of the IAE-PR8 stocks was determined using a plaque titration assay, as described below. For viral inactivation, IAV-PR8 was boiled for 20 min and it was confirmed that there was no infectivity in the boiled viral solution by plaque titration assay.
Animal experiment
All animal experiments were approved by the Animal Research Committee of NIBIOHN and Osaka University and all experiments were carried out following the Guidelines for the Japanese Animal Protection and Management Law, and with the approved standard operating procedure of the biosafety level facility.
Male C57BL/6 mice (CREA Japan Inc.) were randomly assigned into two groups; one group with live PR8 inoculation and another with boiled-inactivated PR8. PR8 inoculation experiments were performed as described previously with modifications [
46]. To induce severe and acute brain edema mimicking IAE, 3- or 4-week-old mice were inoculated with several routes and several doses of PR8, from 1 × 10
4 to 2 × 10
4 median tissue culture infectious dose (TCID
50) of PR8 intranasal inoculation, until the apparent neurological symptoms occurred 3 days post-inoculation [72 h post inoculation (hpi)]. Once the appropriate route and dose were fixed, the appropriate dose of the chosen route was modified according to the body weight of each mouse for further experiments, adjusting the variability in the body weight of developing young mice. After PR8 inoculation, body weight, neuromuscular function and neurological symptoms of either live or inactivated virus-inoculated mice were measured daily till 72 hpi. Neurological performance was assessed using clinical scoring of neuromuscular function [
77]. Neurological symptoms were also assessed by observing seizures.
Mice were sacrificed under anesthesia at 3, 6, 24, 48, and 72 hpi to collect whole blood and brain tissue. Brain tissue was sectioned and weighted before use in each experiment. Evans blue dye (1%) was injected intravenously 1-h before sacrifice and perfused transcardially with PBS, for the vascular permeability study to confirm brain edema.
Cell lines
HUEhT-1 cells (JCRB1458), U-251 MG cells (IFO50288) and MDCK cells (JCRB9029) were obtained from the JCRB Cell Bank of NIBIOHN. HUEhT-1 cells were grown in Molecular, Cellular, and Development Biology (MCDB) 131, including 10 mM of L-glutamine supplemented with 10% heat-inactivated fetal bovine serum (FBS), 30 μg/ml of endothelial cell growth supplement (ECGS), 5 μg/ml of heparin-Na on a collagen coated dish. U-251 MG was grown in Eagle’s minimum essential medium (EMEM) with 10% FBS. MDCK was grown in Dulbecco’s modified Eagle medium (DMEM) with 10% FBS and used mainly for viral proliferation and viral titration as mentioned at corresponding sections.
Cells with 90–95% confluency in 6-well plates were inoculated with PR8 at MOI = 1 (MOI = 0.01 for the viral quantitative assay) and cultured for 24, 48, and 72 h. The supernatant and inoculated cells were prepared to use for viral quantitative analysis, immunocytochemistry, real-time RT-PCR and Western blot, respectively.
Human autopsy tissue
Human brain tissue samples were obtained from the Osaka General Medical Center and Osaka Women’s and Children’s Hospital. The brains of 3 patients with IAE (died in 1998–2004); CHARGE association as co-morbidity of a patient and tuberous sclerosis of another patient, and 1 control (died in 2014) were examined (Supplementary Table 1). Diagnoses of IAE with Reye’s syndrome or HSES were made by direct detecting of the IAV antigen and a typical clinical course for 2 cases. The diagnosis of IAE with Reye’s syndrome for 1 case was made by a typical clinical course and concurrence with an influenza epidemic. Autopsies were performed after obtaining written informed consents from the parents. This study was approved by the Central Ethics Review Boards of the Osaka University Hospital (23,194) for all facilities. This study was conducted following the Declaration of Helsinki and the Ethics Guidelines for Medical Research Involving Human Subjects.
Tissue obtained during the autopsy were immediately fixed in 10% formalin and processed into paraffin blocks. 5-μm-thick paraffin sections of the frontal cortex, temporal cortex, and cerebellum were used for immunohistochemical studies. The tissue was used after confirming that they did not contain tubers or areas of calcification.
Antiviral drugs and chemicals
Baloxavir acid (BXA) was purchased from Shionogi Co. Ltd., and favipiravir (FVP) and peramivir trihydrate (PER) from Tokyo Chemical Industry. The drugs including cycloheximide (CHX) were prepared according to the manufacture’s data sheets and stored at −80 ℃. They were diluted with PBS upon use, and the concentration was adjusted with a series of dilutions for the LDH assay.
For the intervention study of the mouse models, 4 mg/kg of BXA was subcutaneously injected [
53] one day before and 8 h after viral inoculation, for pretreatment and 8 hpi treatment, respectively. 25 mg/kg of PER was also intramuscularly injected [
41] one day before and 8 h after inoculation, as with BXA treatment (Fig.
7a).
Cytotoxicity LDH assay
The LDH cytotoxicity release cell death assay was performed to analyze the effectiveness of antiviral drugs and chemicals for each cell line using the cytotoxicity LDH assay Kit-WST. The assay was performed following the manufacturer’s instructions. Adherent cells (MDCK cells, HUEhT-1 cells and U-251 MG cells) that were previously cultured for 3–4 days were grown to approximately 90% confluency. Each cell was washed, detached with trypsin, and pelleted. Cells were then counted using a Neubauer counting chamber, diluted with each cell culture medium, and seeded in 96-well plates at the number that reached 90% confluence at 24 h after seeding. 24 h later, the cell growth in the wells of the plates was confirmed confluency at 90–95% and the cell culture medium was removed, and the plates were washed thrice. 100 μl of VCM containing IAV-PR8 (MOI = 10), three-staged dilution series of each antiviral drugs/chemical (BXA, FVP, CHX, and PER) and controls were added to each well, and incubated for further 48 h in the incubator. A provided lysis buffer was added to the high control wells 30 min before the applications of the working solution, and then stop solution to all wells. The LDH absorbance released from each well was measured at 490 μm by a plate reader.
Immunohistochemical/immunocytochemical analysis
Histological analyses of the animal model were performed as previously described [
45]. In brief, the mice were anesthetized (isoflurane) and transcardially perfused with 4% paraformaldehyde (PFA) in PBS or saline. The brains were removed, equilibrated with 4% PFA/PBS, cryoprotected with 30% sucrose, and embedded in the OCT compound (Sakura Finetek) using 2-methylbutane cooled in liquid nitrogen. The brain tissue was then sectioned to an 8-μm thickness.
Histological analyses of the autopsy brains were performed with Hematoxylin and Eosin (H&E) staining following standard protocols. For immunohistochemistry, the brain sections of the animal model and autopsy were stained using the antibodies listed in the Supplementary Methods.
Areas of brain endothelial cells (EC) and IAV-nucleoprotein (NP)-stained areas from both live and inactivated group animals were calculated in pixels, using immunohistochemical lower magnification images (100×) of CD31 (EC) and IAV-NP. For the immunocytochemical analysis of inoculated cells, cells were washed with 1 × PBS and fixed with 10% formaldehyde/PBS for 30 min. After washed with 1 × PBS, cells were treated using the same procedures as in immunohistochemical studies.
All immunohistochemical slides were viewed on a Zeiss Axio Observer for bright field and fluorescence microscope (Carl Zeiss Co., Ltd). Dual or triple immunofluorescence slides were also imaged with Leica TCS SP8 Microscope capable of 3D (motorized XY stage and Z focus).
Viral quantitative analysis with animal brain tissue and infected cells
A plaque titration assay was performed as described previously [
46]. Viral quantitative analysis of the animal brains, whole blood, and infected cells was performed according to the plaque titration assay with some modifications. For brain tissue and whole blood of mice, the collected tissue and whole blood were homogenized and sonicated in VCM, and the supernatant was titrated. Details were described in the Supplementary Methods.
Real-time RT-PCR
RNA from brain tissue and cells was purified using TRIzol ™ Reagent (Thermo Fisher Scientific) according to the manufacturer’s protocol. Viral RNA was quantified using the QuantiTect® Probe RT-PCR kit (QIAGEN). IAV-NP cycle threshold (Ct) value was analyzed using the Applied Biosystems 7900HT Fast Real-Time PCR system (Applied Biosystems Japan Ltd.) under the following conditions: 50 ℃ for 30 min., one cycle; 95 ℃ for 15 min., one cycle; 94 ℃ for 15 s, and 56 ℃ for 75 s, 45 cycles. The IAV RNA for the NP gene was detected using primers; a forward (5′-GARRTYATAARRATGATGG-3′), reverse (5′-ATTGTCTCCGAAGAAATAAG-3′), and TaqMan probes (5′-FAM-CGTCYGAGAGCTCRAARACTCCCC-MGB-3′) [
65].
Western blot analysis
The lysates including proteins from brain tissue and cells, were purified after the RNA and DNA were removed using TRIzol. The extracted lysates were dissolved in 1% SDS solution and sonicated. The supernatants of the lysates were harvested, loaded onto 10% SDS-PAGE gels (50 μg total protein per well) and transferred to a polyvinylidene difluoride membranes. The membranes were blocked with 5% dry milk for 1 h and incubated with an anti-IAV-NP antibody (flu A-NP; 1:1600, Gene Tex), anti-CD11b antibody (1:1000, Abcam), or an anti-β-actin antibody (1:500, Sigma Aldrich) in TBS-T (20 mM Tris–HCl pH7.6, 137 mM NaCl and 0.2% Tween 20®) at 4℃ overnight. Then, the membranes were washed with TBS-T three times each for 5 min and then incubated with horseradish peroxidase (HRP)-conjugated secondary-antibodies in TBS-T for 1 h at room temperature (RT). Protein signals were detected using a chemiluminescent substrate (Chemi-Lumi One, Nacalai Tesque Inc.) and the image was obtained using the ImageQunat™ LAS 500 (Cytiva).
Statistics
A two-tailed Student’s t-test was used to determine the significance of the differences between the two groups, while one-way ANOVA was used for multiple different groups. The occurrence of death was assessed by Kaplan–Meier analysis and differences were determined using the log-rank test. Data are presented as mean ± SEM. Statistical significance was set at P < 0.05. GraphPad Prism 9.0 were used for statistical analyses and data visualization.
Discussion
This study showed that IAV invaded the brain from brain EC and that viral proteins accumulated mainly in brain EC, resulting in brain edema or equivalently IAE using the IVE mouse model. The influenza virus was proved to enter the brain EC in IVE mice and patients with IAE. Although IAV did not show prominent proliferation in mouse brains, IAV-NP accumulated in the brains of the IVE mice and the IAE patients and spread from EC to outside parenchyma or cells including astrocytes. The accumulation of viral protein should be responsible for the severe clinical course and lethal prognosis in IVE mice. Consistent with this, an antiviral drug that blocked viral transcription and translation showed the complete effects in the treatment of IVE mice.
Several IAE related animal models have been reported as mentioned: encephalitis like models using none-seasonal influenza virus, animal models using KO mice and animal models using drugs which act to exaggerate brain edema [
5,
10,
36,
55,
60,
66,
80,
90,
93,
95,
98,
101]. Among these models, ‘mouse models for Reye’s syndrome’ and the suckling period mouse models using juvenile visceral steatosis (JVS) mice had presented severe brain edema by inoculating epidemic influenza virus without adding any other drugs, as well as our IVE mice [
80,
101]. Both models showed influenza virus antigen from brain endothelial cells, although the concise viral dynamics were unknown. Here we partially showed the viral dynamics and the possibility of treatment in IAE pathophysiology with IVE models and IAE autopsy brains.
The IVE model mice recapitulated many salient characteristics of human IAE: an acute and aggressive course, brain edema, increased inflammatory cytokines observed in two-thirds of animals, and typical pathological features; leakage of plasma components to the brain parenchyma, bleeding/hemorrhage, lack or least accumulation of inflammatory cells, and astrocytic clasmatodendrosis. Increased inflammatory cytokines was very variable in IVE mice, which was also observed in patients with IAE [
26,
62], indicating that the increase in cytokines in IAE would be one of sufficient conditions causing the disease. These characteristics were especially observed in ANE, HSES, or other severe types of IAE.
The detection of influenza viral antigen in the brain parenchyma of IVE mice and patients with IAE was a key finding of this study. In the present study, we detected live virus and viral RNA in the IVE mouse brains and viral protein accumulation in the brains of both the IVE mice and IAE patients. Previous studies have also reported the viral detection from neurons, EC, astrocytes, and CD8-positive cells in IAE autopsy brains [
19,
23,
37,
88,
99]. The routes to the extra-respiratory organs, especially central nervous system (CNS), of influenza virus invasion has long been a debate [
49]. Although hematogenous spreading is a popular route, the low amounts and frequencies of viral antigens detected in the CNS or other extra-pulmonary organs are the weak points of this hematogenous route theory. Although our present IVE mouse model was established based on this hematogenous route, mice with intranasal inoculation also showed IAV antigen in the brains without apparent neurological symptoms, though the level was low (Supplementary Fig. 2B). Several studies have reported that IAV (H5N1) live virus, antigen, or RNA was detected in serum or autopsied brain [
9,
82], and detection of IAV RNA was also reported in the 2009 pandemic (H1N1) and IAV (H3N2) infection [
17,
75]. These detection of the viral antigen in brain EC and in vasculature were also reported with intranasally inoculated mice [
13,
59].
The present study detected IAV live virus, viral RNA and viral protein in the brains of IVE model mice. In IVE mice, live virus disappeared from the vasculature within 24 h after IAV intravenous inoculation and live virus in the brains did not increase with time. These results indicated that IAV proliferated neither in the blood nor in the brains in IVE mice, and therefore, the viral loads were scarce in IVE mice brains. This is consistent with the paucity of viral antigens detected in the brains of patients with IAE. Despite the paucity of viruses in the IVE brains, viral existence in the brain itself seems pathological. In cultured EC, IAV decreased the number of daughter virus less than one tenth (Fig.
2g) consistent with earlier studies [
29,
84]. In another cultured EC, IAV has been reported to enter human lung microvascular endothelial cells (HMVEC) and promote vascular leakage without apparent proliferation [
2,
24,
68]. As seen in vitro and in vivo, IAV caused denature of EC without viral proliferation. After denaturing EC, IAV spreads to the brain parenchyma and a decrease of brain EC and increased vascular permeability led to brain edema. In this way, the temporal and spatial shifts of brain lesions should be linked to the onset and formation of IAE. Given these perspectives, unlike the IAV respiratory infection via airway epithelial cells, IAV seems to enter brain via EC and causes severe pathogenesis without regular proliferation. How viral protein spreads from EC to brain parenchyma and other CNS cells is yet to be solved.
Our data suggests that IAV should induce severe brain lesions despite less proliferative characteristics. We observed the key discrepancy between the amount of IAV protein and viral load. The amounts of viral protein NP increased over time, while the viral loads did not. Based on these observations, the direct pathogenesis of IAV proteins in IAE was tested using anti-influenza drugs in the present study.
Anti-influenza drugs were effective in MDCK cells in vitro. In contrast, only drugs that inhibit transcription and/or translation were effective in HUVEC, and all drugs tested showed little or no effect in cultured astrocytes. Notably, CHX; an inhibitor of translation, was highly effective in HUVEC, with only limited effects in MDCK cells. This indicates that HUVEC cell death can be potentially explained by synthesis of viral proteins rather than by viral proliferation. This was evident compared to the mechanisms of antiviral drugs in MDCK cells, in which any blockade of the IAV proliferation cascades was effective. Drugs that were effective in HUVEC in vitro were also effective in the IVE mice, suggesting that inhibition of transcription/translation, but not viral replication or proliferation, could prevent the progression of IAE pathogenesis. IAV virulence in brain EC was revealed to be the most important in IAE pathogenesis, since BXA was effective in IVE mice despite the limited effects in astrocytes in vitro.
Although influenza virus infects EC or astrocytes [
29,
97], anti-influenza drugs has rarely been tested in these cells. The mechanisms of action of anti-influenza drugs are variable [
21,
48,
85], and it is important which drug to be selected for each cell line. In fact, neuraminidase inhibitors and RNA-dependent RNA polymerase (RdRp) inhibitors were effective for MDCK, an epithelial cell line. However, these drugs had a limited effect on EC and IVE mice. The discrepancy of the trigger or entry point, between the IAE and the regular respiratory influenza infections, was responsible for the difference in the efficacy of antiviral drugs. Our study showed the importance of identifying a key process of each IAV-related pathogenesis and that different pathogeneses led to the different efficacies of drugs.
Despite propensity to induce viral resistance [
74], only BXA is currently available as a drug that inhibits viral transcription and/or translation. Even under the restricted conditions, combined use of BXA and other anti-influenza drugs should be effective for IAE while reducing the risk of viral resistance.
This study showed the temporal-spatial dynamics of IAV in the pathogenesis of IAE. IAV entry into brain EC triggers the pathogenesis of IAE and promotes the accumulation of IAV protein in EC and brain parenchyma without apparent viral proliferation. The resultant denaturation of EC followed by astrocytic damage leads to brain edema and severe neurological symptoms in IVE mice. Inhibitors of viral transcription and/or translation are effective in rescuing EC denaturation and improving the lethality of IAE models.
The IAE pathogenesis, with severe brain edema caused by IAV, could be established without viral proliferation. The regulation of viral transcription and translation is the key to treatment of IAE.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.