Human adenovirus type 7 (HAdV-7) infection induces pulmonary vascular endothelial injury through the activation of endothelial autophagy

Abstract

Background

HAdV-7 is a prevalent pathogen that can cause severe pneumonia in children. Previous studies have shown a significant increase in serum levels of vascular permeability factor (VPF/VEGF) and viral load in pediatric patients with fatal HAdV-7 infection, suggesting potential damage to the pulmonary vascular endothelium. Further research is necessary to elucidate the underlying mechanism.

Methods

The human lung microvascular endothelial cell line-5a and human CD46 mice were used for in vitro and in vivo experiments, respectively. RNA-seq was employed for correlative omics analysis. Viral infection and copy status were examined using transmission electron microscopy to observe virus particles, immunofluorescence to detect the viral protein Hexon, and qPCR to assess HAdV-7 fiber gene copies. Various methods, including ELISAs for VEGF and other injury markers, the CCK8 assay for cell viability, and flow cytometry for endothelium numbers, were employed to evaluate endothelial damage. Acute lung injury severity was evaluated by scoring pathological inflammation and measuring pulmonary vascular permeability. Autophagy activation was assessed by observing autophagosomes and validating marker proteins.

Results

GSEA analysis showed significant enrichment of gene sets related to endothelial functions (barrier, defense, and regeneration) and ALI in the HAdV-7-infected group. GO analysis indicated an enrichment of autophagy-related pathways linked to cell death. Subsequently, successful signs of HAdV-7 infection and replication were observed in the endothelium, including cytopathic effects, intracellular virions, and increased HAdV-7 fiber gene copies. Endothelial injury, including mitochondrial damage, decreased endothelium, and elevated levels of endothelial injury markers such as VEGF, sICAM-1, sVCAM-1, E-selectin, ESM1, MCP1, and IL1β were observed after HAdV-7 infection. Additionally, evidence of leaky lung blood vessels and ALI was observed, including progressive weight loss, elevated pulmonary vascular permeability, and severe lung consolidation. Furthermore, HAdV-7 infection induced autophagosome formation in the endothelium and triggered complete cell autophagy. Importantly, inhibiting autophagic flux reduced VEGF levels and other endothelial injury markers, decreased viral load, improved cell survival rate, alleviated pulmonary vessel leakage, and mitigated lung inflammation.

Conclusions

HAdV-7 successfully infects pulmonary vascular endothelium and replicates effectively, causing injury to the endothelium, high VEGF expression and viral load in the serum, as well as ALI/ARDS. Autophagy inhibitors can alleviate endothelial injury, inhibit viral replication, relieve leakage from the vasculature, and reduce lung inflammation.

Introduction

Human adenovirus type 7 (HAdV-7) is a prevalent pathogen that causes severe community-acquired pneumonia in children, significantly contributing to infant mortality associated with pneumonia [1]. Children are highly vulnerable and contagious, often leading to serious complications such as acute lung injury/acute respiratory distress syndrome (ALI/ARDS) and viremia [2, 3], posing a significant threat to their life and health. Previous studies have mainly focused on the impact of HAdV-7 infection on alveolar epithelial cells, with limited knowledge about its effect on pulmonary vascular endothelial cells. Particularly during fatal HAdV-7 infections, the virus not only targets alveolar epithelial cells but also damages pulmonary vascular endothelial cells, compromising lung vascular endothelial cell function [4, 5].

The pulmonary microvascular endothelium is a metabolically active continuous monolayer of squamous endothelial cells responsible for substance exchange between tissues and blood. It plays crucial roles in maintaining barrier integrity, anti-inflammatory, antioxidant, and antithrombotic properties, regulating vascular tone, and controlling hemostasis. Damage to these cells can result in glycocalyx/barrier disruption, increased inflammation, cell death, thrombus formation, and oxidative stress [6]. Studies have shown that single-cell sequencing of mouse lung tissue infected with Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) and influenza viruses has identified endothelial cell damage in the pulmonary vasculature [7, 8]. Moreover, severe cases of SARS-CoV-2 infection and autopsy findings have indicated disruptions in extra pulmonary endothelial barriers like the blood–brain barrier, gut-blood barrier, and glomerular filtration membrane [9, 10]. Severe SARS-CoV-2 infection can cause endothelial damage, allowing the virus to enter the bloodstream from lung tissue and result in viral sepsis, which is currently a focus of research [10, 11]. Some drugs targeting endothelial cells have shown reduced mortality rates in patients with Corona Virus Disease-19 (COVID-19) [12]. Targeting endothelial cell injury is crucial for improving outcomes of severe respiratory virus infections. However, there is a lack of comprehensive data on pulmonary endothelial damage caused by severe HAdV-7 infection, emphasizing the necessity for further investigation.

Our previous study discovered a notable increase in serum VEGF and viral load in children with fatal HAdV-7 infection [13], indicating that severe HAdV-7 infection may cause damage to vascular endothelial cells and result in heightened vascular permeability. To further investigate this phenomenon, we employed in vitro models using HULEC-5a cells and in vivo models using hCD46 mice. Our aim was to explore how pulmonary endothelial cell damage influences the upregulation of VEGF expression, elevated serum viral load, and the development of ALI/ARDS following severe HAdV-7 infection. Furthermore, we will delve into the underlying mechanism behind these observed effects, providing novel insights to improve adverse outcomes related to fatal HAdV-7 infections.

Materials and methods

Cell line

HULEC-5a cells were acquired from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). The cells were cultured in a specialized medium (Procell, CM-0565) designed for optimal growth of HULEC-5a cells at 37 °C with 5% CO₂.

HAdV-7 strain

The HAdV-7 strain (CQ45_2019, MT113943) was obtained from nasopharyngeal aspirates of a pediatric patient with adenoviral pneumonia. The strain was cultured in A549 cells, purified [14], and quantified [15, 16] using standard procedures, resulting in a stock concentration of 3 × 10¹¹ viral particles per milliliter.

Animal model

The hCD46 mice (C57BL/6 J) were obtained from Cyagen Biotechnology Co., LTD (Suzhou, China) and housed in a specific pathogen-free facility at the Experimental Animal Center of Chongqing Medical University. The mice were maintained under a 12-h light–dark cycle, with a humidity of 55 ± 5% and a temperature of 23 ± 1 °C, and provided with ad libitum access to food and water. Prior to infection with HAdV-7 (80 μl/mouse, nasal drip) [17], hCD46 homozygous mice (6–8 weeks old) were pretreated with intraperitoneal injections of rapamycin (RAPA, MedChemExpress, HY-17589A), 3-methyladenine (3MA, MedChemExpress, HY-19312), and chloroquine (CQ, MedChemExpress, HY-17589A) one hour. Mock-infected mice received PBS treatment for comparison. RAPA was dissolved in a mixture containing 2% DMSO (Sigma-Aldrich) and PEG300 and Tween at a concentration of 2.5 mg/mL and injected at a dose of 6 mg/kg [18]. 3MA was dissolved in PBS at a concentration of 4 mg/mL and administered at a dose of 30 mg/kg [19], while CQ was dissolved in a mixture containing 2% DMSO (Sigma-Aldrich), PEG300, and Tween at a concentration of 2.5 mg/mL and injected at a dose of 60 mg/kg [18]. Regarding the timing of sample collection, during the previous study on constructing an acute lung injury model in hCD46 mice infected with HAdV-7, we observed that inflammation, weight loss, and reduced activity were evident on the first day post-infection. By the third day, severe lung inflammation, decreased body weight, hypothermia, and mortality were observed. From the 5th to 7th day post-infection, a gradual resolution of lung inflammation along with recovery in body weight and mouse activity was noted. Based on these empirical findings, we opted to collect samples on day 3 post-infection [17, 20].

RNA extraction library construction and sequencing

Total RNA was extracted using Trizol reagent (Thermos Fisher, 15596018) following the manufacturer’s protocol. The samples were then flash-frozen in liquid nitrogen for 30 min and shipped on dry ice to LC-Bio Technology Co., Ltd. (Hangzhou, China) for transcriptome sequencing analysis. The total RNA quantity and purity were analysis of Bioanalyzer 2100 and RNA 6000 Nano LabChip Kit (Agilent, CA, USA), high-quality RNA samples with RIN number > 7.0 were used to construct sequencing library. mRNA was purified from 5 μg of total RNA using Dynabeads Oligo (dT) (Thermos Fisher, CA, USA) with two rounds of purification. The purified mRNA was then fragmented into short pieces using divalent cations at 94 ℃ for 5–7 min with Magnesium RNA Fragmentation Module (NEB, MA, USA). These fragmented RNA pieces were reverse-transcribed into cDNA using SuperScript™ II Reverse Transcriptase (Invitrogen, CA, USA). The average insert size of the final cDNA libraries was 300 ± 50 bp. Finally, 2 × 150 bp paired-end sequencing (PE150) was carried out on an Illumina NovaSeq™ 6000 by LC-Bio Technology Co., Ltd. (Hangzhou, China) following the vendor’s recommended protocol.

Differential expression genes (DEGs) and GO enrichment analysis

DEGs analysis was performed by DESeq2 software (version 1.26.0) between two different groups (and by edgeR between two samples). The genes with the parameter of false discovery rate (FDR) below 0.05 and absolute fold change ≥ 2 were considered differentially expressed genes. GO analysis was conducted using Omic Studio tools (https://www.omicstudio.cn/tool), with terms considered significantly enriched at a cutoff of P < 0.05.

Gene set enrichment analysis

We performed gene set enrichment analysis using software GSEA (version 4.3.2) and MSigDB to identify whether a set of genes in specific GO terms shows significant differences in two groups. Briefly, we input gene expression matrix and rank genes by Signal2Noise normalization method. Enrichment scores and p value was calculated in default parameters. GO terms meeting this condition with |NES|> 1, NOM P-val < 0.05, FDRq-val < 0.25 were considered to be different in two groups.

Optical microscope and transmission electron microscopy assays

Morphological characteristics of cells were observed using an optical microscope. Images were captured with a Nikon microscope (Nikon, Japan) and processed for analysis using Image Viewer software (version 4.5). For the observation of mitochondrial ultrastructure and viral particles, the treated HULEC-5a cells were fixed with 3% glutaraldehyde at 4 °C overnight, followed by fixation with 1% osmium tetroxide at 4 °C for 1 h. Dehydration steps using acetone and embedding in Epon-812 were then performed. Ultra-thin sections were prepared using a Leica ultramicrotome (LEICA, Buffalo Grove, IL, USA) and stained with uranyl acetate and lead citrate. TEM images were acquired with a JEM-1400FLASH transmission electron microscope (JEOL Ltd., Tokyo, Japan).

ELISA

Endothelial injury markers in the cell supernatant and bronchoalveolar lavage fluid (BALF)/blood were quantified using ELISA. BALF was obtained by rinsing the right lung three times with 0.5 mL of PBS, resulting in a 90% recovery rate. Levels of VEGF (NeoBioscience, EHC108.96, EMC103.96.2), sICAM-1 (4A Biotech, CHE0052), sVCAM-1 (finetest, EM1382), E-selectin (finetest, EM0229), ESM1 (finetest, EH0125, EM0075), MCP1 (NeoBioscience, EHC113.96, EMC105.96), and IL-1β (NeoBioscience, EMC113.96) were determined using ELISA kits according to the manufacturer’s instructions. Absorbance was read at 450 nm with a microplate reader (Biotek Epoch, VT, USA), and protein concentrations were calculated using ELISA calc software (version 0.1).

Cell viability assays

HULEC-5a cells were pre-treated with RAPA (20 nM), 3MA (5 mM), CQ (20 μM), and Z-VAD-FMK (40 μM, Selleck, S7023) for one hour before infection with HAdV-7(MOI = 10⁵). Subsequently, the cells were incubated with 10% Cell Counting Kit-8 (CCK8) reagent (MedChemExpress, HY-K0301) for 0.5 to 2 h. Absorbance was read at 450 nm with a microplate reader (Biotek Epoch, VT, USA).

Immunofluorescence staining

For the Immunofluorescence assay, HULEC-5a cells were cultured on confocal dishes, while Paraffin-embedded tissue sections underwent deparaffinization and antigen retrieval treatment. The cells and lung tissues were then fixed with 4% paraformaldehyde and permeabilized with 0.2% Triton X-100 (Solarbio, 9002-93-1). Subsequently, after blocking with 5% Bovine Serum Albumin (BSA, Sigma-Aldrich, V900933), the cells and lung tissues were incubated overnight at 4 °C with primary antibodies LC3B (Cell Signaling Technology, 83506S, 1:500), SQSTM1 (Cell Signaling Technology, 16177S, 1:500), Anti-Adenovirus antibody (Abcam, ab7428, 1:1000) and CD31 (Abcam, ab124432, 1:500). Following this, secondary antibodies Cy3-labeled goat anti-mouse/rabbit IgG (Beyotime, A0521/ A0516, 1:500), FITC-labeled goat anti-rabbit IgG (Beyotime, A0562, 1:500), and Alexa Fluor488 labeled goat anti-mouse IgG (Beyotime, A0428, 1:500) were applied for 1 h at room temperature, followed by DAPI staining (Beyotime, C1005). Fluorescent images were captured using a Nikon A1R + /A1 confocal microscope (Nikon, Tokyo, Japan), and fluorescence intensity analysis was performed using NIS ELEMENTS Viewer software (version 4.5).

Flow cytometry

The lungs were digested in Roswell Park Memorial Institute 1640 culture medium (Gibco) at 37 °C for 30 min with collagenase IV (Sigma, C0130) at a concentration of 2 mg/mL and DNase I (Roche, 10104159001) at a concentration of 30 mg/mL. The samples were then filtered through a 40-μm cell strainer to obtain single-cell suspension. Subsequently, the cells were blocked with rat serum for 30 min and stained in the dark for one hour with fixable viability dye eFluor 660 (Invitrogen, 65-0864-14), PE-Cy7-conjugated anti-CD45 antibody (eBioscience, 17-0451-82), and FITC-conjugated anti-CD31 antibody (eBioscience, 11-0311-82). The samples were analyzed using BD FACSCanto plus equipped with 50-mW 405-nm, 50-mW 488-nm, and 20-mW 633-nm lasers and an ND1.0 filter in front of the forward scatter photodiode.FlowJo software (version 10, TreeStar, USA) was used for further data analysis.

Gene mouse identification

The genotyping procedure is as follows: 98 μl lysis buffer and 2 μl proteinase K (20 mg/mL) are added to the toe of each tested mouse and incubated at 56 ℃ overnight. Then, they are placed in a metal bath at 98 ℃ for 15 min, centrifuged at 12,000 rpm at 4 ℃ for 15 min, and the supernatant contains DNA. Next, PCR amplification detection is performed using Premix Taq enzyme (TAKALA, RR901Q). The PCR cycle consists of heating at 94 ℃ for 3 min, denaturation at 94 ℃ for 30 s, annealing at 60 ℃ for 35 s, and extension at 72 ℃ for 35 s with a total of 35 cycles. CD46 primer sequences used are Forward1: 5′-GCTAAGTCTGCAGCCATTACTAAAC-3′, Reverse1: 5′-GAAATCAGGCTGCAAATCTCAGC-3′; Forward2: 5′-GGTTGGCTATAAAGAGGTCATCAG-3′, Reverse2: 5′-GAAATCAGGCTGCAAATCTCAGC-3′. Gel electrophoresis is then conducted by preparing a 1.5% agarose gel supplemented with GoldView (3–5 μl), which is cooled and solidified before adding the test DNA and DNA ladder(both in amounts of 5 μl). Electrophoresis takes place under 120v voltage for 25 min, followed by visualization under ultraviolet light to determine genotype based on band appearance.

Quantification of HAdV-7 fiber gene copy number

The total viral nucleic acid was extracted using a QIAamp mini-viral DNA extraction kit (Qiagen, 51306) following the manufacturer's instructions. Polymerase Chain Reaction (PCR) amplification assays were performed with TaqMan Universal Master MixII (Applied Biosystems, 4440040) on a Real-Time PCR Detection System (Bio-Rad, CFX96 Touch). The HAdV-7 fiber gene copies were amplified using the forward primer 5′-CATAAGTGCCACCACACCAC-3′, reverse primer 5′-GTGCGCTTAACTCCTGTCCA-3′, and probe sequence 5′-FAM-5′-TGTCGCTAGGACCCGGATTAGAAACAA-3-BHQ1-3′. PCR cycles included 50 °C for 2 min, 95 °C for 10 min, followed by 40 cycles at 95 °C for 15 s and 60 °C for 1 min. Absolute quantification of HAdV-7 gene copies was done using a plasmid DNA standard curve.

Western blotting assays

Total protein extraction from HULEC-5a cells was conducted using the Total Protein Extraction Kit (Key GEN, KGP2100). Protein concentrations in the cell lysates were quantified with the NanoDrop One instrument (Thermo Fisher, 840-317400). Following separation by SDS-PAGE, proteins were transferred to polyvinylidene fluoride (PVDF) membranes (Millipore, ISEQ00010). The membranes were then blocked with Quick Blocking Buffer (Beyotime, P0252) and probed with specific primary antibodies, including LC3B (Cell Signaling Technology, 83506S, 1:1000), SQSTM1 (Cell Signaling Technology,16177S, 1:1000), GAPDH (Zenbio, 380626, 1:10,000), and β-actin (Zenbio, 700063, 1:10,000). The membranes were then incubated with the secondary antibody (Zenbio, 511103, 511203, 1:10,000) at room temperature for 1 h. Signal detection employed a super sensitive ECL luminescence reagent (MeilunBio, MA0186). Densitometry analysis was performed using Image Lab software (version 5.0).

Histopathology

The lung of hCD46 mice was fixed with 4% paraformaldehyde, dehydrated, embedded, sectioned (4um thickness), and stained with hematoxylin and eosin (H&E). The staining method and pulmonary pathology evaluation followed a previously established protocol [17]. A Slide Scan System (SQS-40P, Shengqiang) and image Viewer G software (version G1.1.7.3107) were utilized for scanning and analysis, respectively.

Detection of pulmonary vascular permeability

hCD46 mice were intravenously injected with 0.5% Evans blue (EB, MedChemExpress, HY-B1102) at a dose of 3 mL/kg 1 h prior to euthanasia. Lung tissue weighing approximately 100 mg was then homogenized and mixed with dimethylformamide (1 mL/100 mg). The mixture was incubated at 60 °C for 24 h. The absorbance at a wavelength of 620 nm was measured using a microplate reader (Biotek Epoch, VT, USA), and the concentration of EB content was calculated using Curve Expert software (version 1.4.0.0). Total protein concentration in bronchoalveolar lavage fluid was determined following the instructions provided by the Bicinchoninic acid (BCA) assay kit (Beyotime, P0010).

Statistical analysis

The quantitative data were reported as mean ± standard error of the mean. Unpaired Student’s t-test was used for comparing two groups, while One-way analysis of variance, two-way analysis of variance, or Kruskal–Wallis test were employed for multiple group comparisons. Data analysis was conducted using GraphPad Prism (version 9.4.1). Statistical significance was defined as a P-value less than 0.05.

Results

Gene set enrichment analysis

Based on previous clinical observations, a series of validations were conducted to investigate the impact of HAdV-7 infection on lung vascular endothelial cells, starting with omics analysis. Gene sets related to vascular endothelial cell function and ALI/ARDS were identified in the MsigDB database and analyzed through gene set enrichment analysis. The HAdV-7 group exhibited significant enrichment of gene sets associated with endothelial barrier (including VEGF, CD31, IL1β) (Fig. 1a, b), defense (ICAM1, VCAM1) (Fig. 1c, d), endothelium/vessel regeneration (ESM1, MCP1) (Fig. 1e–h), andALI (Fig. 1i, j).These findings suggest that HAdV-7 infection could potentially impair endothelial barrier, defense, and regeneration functions, ultimately leading to ALI. The key contributing genes identified in this study may have a critical role in the impairment of endothelial function.

Fig. 1

Gene set enrichment analysis in HULEC-5a cells infected with or without HAdV-7 (MOI = 10⁵) for 24, 48, and 72 h. a Enrichment plot b and heatmap of gene expression levels of establishment of endothelial barrier gene set. c Enrichment plot d and heatmap of gene expression levels of biocarta lair pathway gene set. e Enrichment plot f and heatmap of gene expression levels of endothelial cell proliferation gene set. g Enrichment plot h and heatmap of gene expression levels of sprouting angiogenesis gene set. i Enrichment plot j and heatmap of gene expression levels of acute lung injury gene set. Each column represents a sample and each row represents a gene. The color gradient from blue to red signifies different levels of gene expression

HAdV-7 infection induces damage to pulmonary vascular endothelial cells

To validate the GSEA findings, we initially investigated the infectivity of HAdV-7 on pulmonary vascular endothelial cells. An in vitro model was established using HULEC-5a cells infected with HAdV-7. Morphological observations showed time-dependent cytopathic effects, including cell aggregation and fusion into clusters (Fig. 2a). TEM confirmed the presence of viral particles in the nuclei of infected cells, indicating successful infection and replication (Fig. 2b). Subsequently, we searched for evidence of cellular damage. Ultrastructural examination revealed significant mitochondrial damage in infected cells, characterized by mitochondrial shrinkage and shortened or disappeared mitochondrial ridges (Fig. 2c, d). Furthermore, ELISA detection indicated elevated expression of sICAM-1 (an endothelial activation/dysfunction marker), ESM1 (an glycocalyx damage marker), MCP1 (an endothelial inflammatory injury marker), and IL1β (an endothelial inflammatory injury marker) (Fig. 2e–h). CCK8 experiments demonstrated reduced cell viability at 24, 48, and 72 h post-infection (Fig. 2i). These results suggest that HAdV-7 effectively infects HULEC-5a cells and results in cellular damage.

Fig. 2

HAdV-7 infected and induced damage to HULEC-5a cells. a Representative optical microscope images of HULEC-5a cells infected with or without HAdV-7 (MOI = 10⁵) for 24, 48, and 72 h (scale bar: 10 μm). b–d Representative TEM images of (b) viral particles within the nucleus (c, d) and mitochondria structure (scale bar: 2 μm & 500 nm) in HULEC-5a cells infected with or without HAdV-7 (MOI = 10⁵) for 72 h. Red arrow indicates viral particles. e–h Relative protein expression levels of indicated endothelial injury biomarkers in HULEC-5a cells infected with or without HAdV-7 (MOI = 10⁵) for 72 h were determined. i CCK8 analysis of HULEC-5a cells infected with HAdV-7 (MOI = 10⁵) for 24, 48, and 72 h were determined. The data presented are from three independent experiments, with values expressed as standard errors of the means (SEM). *P < 0.05, **P < 0.01, ****P < 0.0001. Compared with the indicated group. HPI: hour post-infection

Subsequently, an in vivo model was established using hCD46 mice infected with HAdV-7. Immunofluorescence co-localization demonstrated the presence of viral protein Hexon in both lung tissue and pulmonary vessels in the infected group (Fig. 3a). Flow cytometry results indicated a decrease in the population of CD45⁻CD31⁺ endothelial cells within lung tissue (Fig. 3b–d). ELISA detection revealed increased levels of sVCAM-1 (an endothelial inflammatory injury marker), E-selectin (an endothelial inflammatory injury marker), MCP1, and IL1β as markers for pulmonary endothelial cells, suggesting endothelial inflammatory injury (Fig. 3e–h). These findings suggest that HAdV-7 effectively infects pulmonary vascular endothelial cells of hCD46 mice, leading to cellular injury.

Fig. 3

HAdV-7 infected and induced damage to pulmonary vascular endothelial cells in hCD46 mice. a hCD46 mice were infected with or without HAdV-7 (80 μl/mouse, nasal drip) for 72 h. Confocal images were obtained after paraformaldehyde fixation of lung tissue (scale bar: 50 μm). Red signal indicated CD31, orange signal indicated Hexon, and nuclei were counterstained with DAPI (blue). b–d Flow cytometry analysis of CD45⁻CD31⁺ cells in lung tissue single-cell suspensions e–h and relative protein expression levels of indicated endothelial injury biomarkers in BALF from hCD46 mice infected with or without HAdV-7 (80 μl/mouse, nasal drip) for 72 h were determined. The data presented are from three independent experiments, with values expressed as standard errors of the means (SEM). *P < 0.05, **P < 0.01, compared with the control group

HAdV-7 infection induces an increase in the expression of VEGF and a high viral load

The VEGF protein, also known as vascular permeability factor, plays a crucial role in enhancing blood vessel permeability and promoting the migration, proliferation, and survival of endothelial cells. It is considered a marker for both endothelial hyper permeability and disease severity during infections [21]. Previous studies have shown an upregulation of VEGF expression in patients with severe COVID-19 infections [22, 23]. Does HAdV-7 infection similarly increase VEGF levels? ELISA results showed a significant rise in VEGF expression in infected cell supernatant (Fig. 4a), hCD46 mice blood (Fig. 4b), and BALF (Fig. 4c). Interestingly, VEGF expression was higher in hCD46 mouse BALF compared to blood samples, supporting the idea that HAdV-7 infection results in endothelial cell damage and increased pulmonary vascular permeability. Subsequent QPCR results revealed a notable increase in fiber gene copy number in infected cell supernatant (Fig. 4d), hCD46 mice blood (Fig. 4e), and lung tissue (Fig. 4f). These findings are consistent with clinical observations suggesting that HAdV-7 infection leads to elevated VEGF expression and a high viral load.

Fig. 4

HAdV-7 infection induced increased expression of VEGF and high viral load. A–c VEGF expression levels (a) in supernatant of HULEC-5a cells (b) in blood (c) and in BALF of hCD46mice infected with or without HAdV-7 (MOI = 10⁵, 80 μl/mouse, nasal drip) for 72 h were determined. d–e HAdV-7 fiber gene copy number (d) in supernatant of HULEC-5a cells (e) in blood (f) and in lung tissue of hCD46 mice infected with or without HAdV-7 (MOI = 10⁵, 80 μl/mouse, nasal drip) for 72 h were determined. The data presented are from three independent experiments, with values expressed as standard errors of the means (SEM). *P < 0.05, **P < 0.01. Compared with the control group

It is worth noting that both MCP1 and IL1β (pro-inflammatory cytokines) were elevated in the cellular supernatant (Fig. 2g, h) and mice BALF (Fig. 3g, h) after HAdV-7 infection. Our previous clinical observation also found an increase in serum MCP1 levels in children with severe HAdV-7 infection [13], suggesting that MCP1 and IL1β may play a significant role in the high expression of VEGF and viral load. Studies have shown that immune cells produce various pro-inflammatory cytokines (such as MCP, IL1β, IL6, IL8) during phagocytosis of pathogenic microorganisms. Specifically, MCP1 is responsible for recruiting monocytes to the site of infection, amplifying the expression of other inflammatory factors and cells, influencing macrophage adhesion and chemotaxis, and disrupting the expression of tight junction proteins and endothelial adhesion molecules to regulate endothelial permeability [24]. Furthermore, IL1β triggers MCP1 expression, which facilitates the recruitment of lymphocytes and the activation of adhesion molecule expression. This process enhances the attachment of white blood cells to the endothelial surface, resulting in endothelial activation and increased vascular permeability [25]. Consequently, inflammatory cells and factors can more easily breach the endothelium, allowing pathogens easier entry into the bloodstream.

HAdV-7 infection induces damage to endothelial cells by activating autophagy in pulmonary vascular endothelial cells

Next, we delved into the mechanism by which HAdV-7 infection affects endothelial cells. Analysis of DEGs identified 16,167,546 significant DEGs at 24, 48, and 72 h post HAdV-7 infection (Fig. 5a). Subsequent GO enrichment analysis highlighted the enrichment of autophagy regulatory pathways associated with cell death in the DEGs at 72 h (Fig. 5b, c), suggesting a potential involvement of autophagy in lung vascular endothelial cell injury induced by HAdV-7 infection.

Fig. 5

DEGs screening and GO enrichment analysis in HULEC-5a cells infected with or without HAdV-7 (MOI = 10⁵) for 24, 48, and 72 h. a Distribution of up- and down-regulated DEGs in HULEC-5a cells at various time points. b The top 20 enriched GO terms identified from the DEGs at the 72-h time point. c Autophagy-related GO terms from the DEGs at the 72-h time point. An adjusted P-value < 0.05 was considered significant in GO analysis. Selected boxes show interest GO terms

Therefore, we initially need to determine if autophagy is activated by HAdV-7. We observed a significant increase in autophagy vacuole structures (Fig. 6a, d) and elevated expression of LC3BII (a marker protein for autophagy) (Fig. 6g) in HULEC-5a cells following HAdV-7 infection. To determine whether this increase was due to autophagy activation or impaired lysosomal degradation, we further examined LC3B II after intervention with autophagy inhibitor. The results showed that treatment with CQ (an inhibitor of autophagosome and lysosome fusion) enhanced the expression of LC3BII (Fig. 6g) and LC3 (Fig. 6b, e), along with an increased number of autophagic vesicle structures (Fig. 6a, d). Conversely, treatment with 3MA (an inhibitor of autophagosome formation) resulted in decreased expression of LC3B II (Fig. 6g) and LC3 (Fig. 6b, e), as well as a reduction in the presence of autophagic structures (Fig. 6a, d). These findings suggest that HAdV-7 activates autophagy in endothelial cells.

Fig. 6

HAdV-7 infection induced autophagy activation in HULEC-5a cells. a–g HULEC-5a cells treated with either 3MA (5 mM), CQ (20 μm), HAdV-7 (MOI = 10⁵), or 3MA and CQ plus with HAdV-7. a Representative TEM images of autophagic vacuoles in HULEC-5a cells (treated for 24 h) (Scale bar: 1 μm). Red arrows indicate autophagic vacuole. b, c Confocal images were obtained after paraformaldehyde fixation of cells, red signal indicated endogenous LC3 (scale bar: 25 μm), green signal indicated SQSTM1 (scale bar: 100 μm), nuclei were counterstained with DAPI (blue) (treated for 72 h). d The autophagic vacuoles e the LC3 f and SQSTM1 mean intensity per cell were calculated. g Western blotting analysis LC3B-II and SQSTM1 expression levels in HULEC-5a cells (treated for 72 h). The data presented are from three independent experiments, with values expressed as standard errors of the means (SEM). *P < 0.05, **P < 0.01, ****P < 0.0001. Compared with the control group, N.S., not significant

It is worth noting that LC3BII expression increased, while the autophagy substrate SQSTM1 decreased (Fig. 6c, f, g). Treatment with CQ (an inhibitor of autophagosome and lysosome fusion) resulted in an increase in SQSTM1 levels (Fig. 6c, f, g). Conversely, inhibition of autophagosome formation with 3MA led to a decrease in SQSTM1 expression, suggesting complete degradation of the autophagy substrate (Fig. 6c, f, g). These findings suggest that HAdV-7 infection triggers autophagy in HULEC-5a cells and maintains a steady autophagic flux.

HAdV-7 infection successfully activates autophagy in endothelial cells.What is the role of autophagy in endothelial cell damage caused by HAdV-7 infection? ELISA results demonstrated an increase in the expression of endothelial cell injury markers (sICAM1, sVCAM1, ESM1, E-selectin, MCP1, IL-1β) after HAdV-7 infection (Fig. 7a–d, e–h). Intervention with the autophagy agonist RAPA enhanced the expression of certain markers in the cell supernatant (Fig. 7c), while the autophagy inhibitor CQ/3MA reduced their expression both in the cell supernatant (Fig. 7a–d) and BALF (Fig. 7e–h). The CCK8 assay revealed that RAPA aggravated cell death (Fig. 7i), whereas CQ significantly prevented it after HAdV-7 infection (Fig. 7j). These results suggest that increased endothelial autophagy can exacerbate endothelial injury and promote cell death, whereas inhibiting autophagy can alleviate injury and inhibit cell death. Additionally, considering concurrent apoptosis, we employed an apoptosis inhibitor Z-VAD-FMK for intervention, however, it did not significantly rescue cell death (Fig. 7 l), indicating that HAdV-7 infection induces autophagic cell death.

Fig. 7

HAdV-7 infection induced endothelial cell injury and death by activating autophagy. a–d Relative protein expression levels of indicated endothelial injury biomarkers in HULEC-5a cells treated with either RAPA (20 nM), 3MA (5 mM), CQ (20 μm), HAdV-7 (MOI = 10⁵), or RAPA, 3MA and CQ plus with HAdV-7 for 72 h were determined. e–h Relative protein expression levels of indicated endothelial injury biomarkers in hCD46 mice treated with either RAPA (6 mg/kg, ip), 3MA (30 mg/kg, ip), CQ (60 mg/kg, ip), HAdV-7 (80 μl/mouse, nasal drip), or RAPA, 3MA, CQ plus with HAdV-7 for 72 h were determined. i–l CCK8 analysis of HULEC-5a cells treated with either RAPA (20 nM), 3MA (5 mM), CQ (20 μm), Z-VAD-FMK (40μM), HAdV-7 (MOI = 10⁵), or RAPA, 3MA, CQ and Z-VAD-FMK plus with HAdV-7 for 24, 48, and 72 h were determined. The data presented are from three independent experiments, with values expressed as standard errors of the means (SEM). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Compared with the control group, N.S., not significant.

HAdV-7 infection induces damage to endothelial cells by activating autophagy, leading to subsequent upregulation of VEGF expression and high viral load

Next, we will explore the relationship between endothelial cell injury and high VEGF expression as well as a high viral load. Since we observed a significant effect of autophagy inhibitor on saving endothelial cell damage in both cell (Fig. 7a–d) and mouse models (Fig. 7e–h), we initially utilized the autophagy inhibitor CQ/3MA to establish a model for reducing endothelial damage to assess VEGF expression levels and viral load. ELISA results showed a significant decrease in VEGF expression in the cell supernatant (Fig. 8a), mouse blood (Fig. 8b), and BALF (Fig. 8c) when treated with CQ/3MA. Additionally, qPCR results demonstrated a notable reduction in the copy number ofthe fiber gene in the cell supernatant (Fig. 8d), mouse blood (Fig. 8e), and lung tissue (Fig. 8f) with CQ/3MA treatment (Fig. 8e,f), suggesting that alleviating endothelial cell damage can lead to a reduction in both VEGF expression and HAdV-7 viral load.

Fig. 8

HAdV-7 infection induced increased expression of VEGF and high viral load by activating autophagy. a–f HULEC-5a cells/hCD46 mice treated with either RAPA (20 nM)/ (6 mg/kg, ip), 3MA (5 mM) /(30 mg/kg, ip), CQ (20 μm)/(60 mg/kg, ip), HAdV-7 (MOI = 10⁵)/(80 μl/mouse nasal drip) or RAPA, 3MA and CQ plus with HAdV-7 for 72h, (a-c) VEGF expression levels (a) in supernatant of HULEC-5a cells (b) in blood (c) and in BALF of hCD46 mice were determined. (d–f) HAdV-7 fiber gene copy number (d) in supernatant of HULEC-5a cells (e) in blood (f) and in lung tissue of hCD46 mice were determined. The data presented are from three independent experiments, with values expressed as standard errors of the means (SEM). *P < 0.05, **P < 0.01, compared with the indicated groups

Although the autophagy agonist RAPA did not further increase endothelial injury markers in mice, suggesting insufficient evidence for the aggravated mouse endothelial injury model constructed by RAPA, some endothelial injury markers increased after RAPA intervention in the cell model (Fig. 7c). Especially, CCK8 results demonstrated that RAPA significantly exacerbated endothelial cell death (Fig. 7i), providing acceptable evidence for constructing an in vitro model of aggravated endothelial injury using RAPA. Therefore, we primarily utilized the in vitro model to detect VEGF expression levels and viral load. The results showed an increase in VEGF expression (Fig. 8a) and HAdV-7 fiber gene copy number (Fig. 8d). From a certain perspective, exacerbating endothelial injury could elevate VEGF expression and HAdV-7 viral load. In summary, our findings indicate that HAdV-7 infection activates endothelial autophagy leading to cellular damage, resulting in elevated VEGF expression and viral load. The failure of RAPA to construct a mouse model of aggravated endothelial injury may be attributed to either the complexity of the mouse environment compared to the cell environment or an off-target reaction of the drug.

HAdV-7 infection activates pulmonary vascular endothelial cell autophagy in hCD46 mice resulting in ALI

Blocking endothelial autophagy has been shown to alleviate cell injury and death, raising the question of whether it could also alleviate ALI. The results of body weight monitoring showed a continuous decrease in mice body weight between 24 and 72 h following HAdV-7 infection. Specifically, mice infected with HAdV-7 and treated with RAPA experienced an average weight loss of 7.11% at 24 h, 12.03% at 48 h, and an average decrease of 11.94% by the end of the experiment (Fig. 9a). In contrast, mice infected with HAdV-7 and treated with autophagy inhibitor CQ/3MA exhibited minimal weight loss, showing a trend towards recovery at 48 h post-infection. These findings suggest that autophagy inhibitor could mitigate the weight loss associated with HAdV-7 infection.

Fig. 9

HAdV-7 infection induced ALI in hCD46 mice by activating autophagy in pulmonary endothelial cells. a Weight changes in hCD46 mice treated with either RAPA (6 mg/kg, ip), 3MA (30 mg/kg, ip), CQ (60 mg/kg, ip), HAdV-7 (80 μl/mouse, nasal drip), or RAPA, 3MA, CQ plus with HAdV-7 for 24, 48, and 72 h were determined. b–g hCD46 mice treated with either RAPA (6 mg/kg, ip), 3MA (30 mg/kg, ip), CQ (60 mg/kg, ip), HAdV-7 (80 μl/mouse, nasal drip), or RAPA, 3MA, CQ plus with HAdV-7 for 72h. (b) Confocal images were obtained after paraformaldehyde fixation of lung tissues (scale bar: 100 μm). Red signal indicated CD31, green signal indicated LC3B, and nuclei were counterstained with DAPI (blue). (c) Representative H&E images of lung tissue in hCD46 mice. (d) Pathological score (e) Injury score (f) Evans blue concentration (g) BALF total protein lever of lung tissue in hCD46 mice were determined. The data presented are from three independent experiments, with values expressed as standard errors of the means (SEM). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Compared with the control group, N.S., not significant

To investigate autophagy activation in pulmonary vascular endothelium, we used immunofluorescence to detect the autophagy marker protein LC3. Our results demonstrated increased LC3 expression in pulmonary vascular endothelium following HAdV-7 infection (Fig. 9b). Treatment with the autophagy inhibitor 3MA reduced LC3 expression, while treatment with CQ augmented LC3 production. These results align with our in vitro findings and indicate effective autophagy activation in pulmonary vascular endothelial cells triggered by HAdV-7 infection.

Subsequently, the lung tissue inflammation score was assessed. The findings indicated that HAdV-7 infection induced significant inflammatory cells infiltration in various areas of the lungs in mice, including bronchial tubes, bronchioles, perivascular regions, and alveolar septum. This was accompanied by notable alveolar congestion, structural destruction, and lung consolidation (Fig. 9c). Additionally, both the inflammatory pathological score (Fig. 9d) and lung injury score (Fig. 9e) showed a marked increase. The autophagy activator RAPA exacerbated lung inflammation in HAdV-7 infected mice (Fig. 9c–e), while the autophagy inhibitor CQ demonstrated remarkable protective effects. CQ significantly reduced inflammatory cells infiltration, alleviated lung congestion, and prevented lung consolidation (Fig. 9c), resulting in decreased scores for inflammation and injury (Fig. 9d, e). Conversely, 3MA did not show a significant protective effect (Fig. 9c–e). These findings indicate that the autophagy inhibitor CQ may effectively alleviate inflammation in lung tissue.

Next, the levels of Evans blue in lung tissue and the concentrations of total proteinin BALF were measured to assess pulmonary vascular permeability. The results demonstrated that CQ effectively reduced both Evans blue levels (Fig. 9f) and total protein concentrations (Fig. 9g), indicating its potential for reducing pulmonary vascular permeability. In summary, our study emphasizes the significance of autophagy inhibition in alleviating lung vascular endothelial cell injury, reducing pulmonary vascular permeability, and mitigating lung tissue inflammation in hCD46 mice.

Discussion

Adenovirus is a significant pathogen responsible for community-acquired pneumonia in children under 5 years old. Specifically, HAdV-7 is characterized by a higher viral load and longer detoxification time [26], increasing the risk of severe conditions such as ALI/ARDS and viremia. Limited research has been conducted on the impact of severe HAdV-7 infection on lung endothelial cells, and the connection between lung endothelial cell damage and negative outcomes of HAdV-7 infection remains unclear. Our findings indicate that HAdV-7 infection damages lung vascular endothelial cells, leading to increased VEGF expression, elevated viral load, and the development of ALI/ARDS. Activation of autophagy in pulmonary endothelial cells may significantly contribute to these phenomena.

The discovery that HAdV-7 infection leads to damage in the pulmonary endothelium aligns with previous research on endothelial damage induced by RSV, H1N1, and SARS-CoV-2 [27, 28]. During the process of adsorption, penetration, dehulling, DNA synthesis, and release of progeny viruses, viruses disrupt the synthesis of proteins and DNA in host cells, as well as destroy organelles, causing direct damage to cells. Moreover, the viruses recruit numerous inflammatory cells/factors resulting in indirect harm[29]. Our study identified elevated levels of markers for endothelial cell damage following HAdV-7 infection, including sICAM-1, sVCAM-1, E-selectin, ESM1, MCP1, and IL1β (Fig. 7a-h). These markers were also found in patients with severe COVID-19 infection [30]. Adhesion molecules like sICAM-1, sVCAM-1, and E-selectin serve as biomarkers for activated endothelial cells [31], facilitating recognition/adhesion between white blood cells and endothelial cells to promote penetration of inflammatory factors/cells through the endothelium and compromise its barrier function. ESM1 serves as a marker for glycocalyx damage, regulating vascular homeostasis by controlling leukocyte adhesion to the endothelium and maintaining vascular permeability [32]. Disruption of glycocalyx structure heightens inflammation/oxidative stress, ultimately breaking down the endothelial barrier [33, 34]. Furthermore, the activation of autophagy inendothelial cells can lead to cellular injury, which is consistent with previous findings [35, 36]. Autophagy plays a crucial role in maintaining cellular homeostasis by removing damaged organelles, abnormal proteins, and excess metabolites [37]. In response to viral infection, autophagy is triggered to degrade viral components and initiate immune responses. However, excessive autophagy may result in autophagic cell death [37].Our research suggests that autophagy agonists and inhibitors play a crucial role in exacerbating or mitigating endothelial cell injury. Specifically, the autophagy inhibitor CQ has been shown to effectively rescue endothelial cell injury (Fig. 7a–h) and prevent cell death (Fig. 7j), highlighting the importance of inhibiting autophagy in saving endothelial injury. It is important to note that while the autophagy agonist RAPA does not elevate markers for endothelial injury in mice (Fig. 7e–h), evidence from cell models suggests that RAPA actually exacerbates endothelial injury (Fig. 7c) and cell death (Fig. 7i). Therefore, there is a basis for believing that RAPA stimulates endothelial autophagy and worsens endothelial cell injury.

HAdV-7 infection triggers elevated VEGF expression, aligning with our prior research on severe cases in children [13]. This finding is consistent with increased VEGF levels observed in severe cases of novel coronavirus [22]. Our study detected heightened VEGF levels in cell supernatant, mouse blood, and BALF after HAdV-7 infection (Fig. 8a–c), with BALF showing higher VEGF levels than serum VEGF (Fig. 8b, c), indicating enhanced pulmonary vascular permeability. As previously mentioned, the up-regulation of VEGF may be linked to pro-inflammatory cytokines MCP1 and IL-1β impacting endothelial cell activation and injury. Additionally, we also observed an increase in EB in lung tissue (Fig. 9f) and total protein content in BALF (Fig. 9g) in HAdV-7 infected mice, suggesting vascular leakage. Treatment with the autophagy agonist RAPA further boosted VEGF levels in cell supernatant (Fig. 8a), and BALF (Fig. 8c), while the autophagy inhibitors CQ/3MA notably reduced VEGF levels in these samples (Fig. 8a-c), as well as EB in lung tissue (Fig. 9f) and protein concentrations in BALF (Fig. 9g), indicating a significant role of endothelial autophagy in influencing VEGF levels and vascular leakage.

The serum viral load in HAdV-7 infection is notably high, which is consistent with our previous research on severe cases in children [13]. This observation aligns with studies that have shown an increased serum viral load in children infected with other respiratory viruses [11, 38]. Not only did we detect a rise in the copy number of the HAdV-7 fiber gene in mouse serum (Fig. 8e), but also in mouse lung tissue (Fig. 8f) and cellmodels (Fig. 8d). This could be attributed to damage to alveolar epithelium and pulmonary vascular endothelial cells caused by HAdV-7 infection, leading to the disruption of tight connections, breakdown of the glycocalyx/barrier, and direct entry of the virusinto the bloodstream [28]. Our ELISA results also indicate heightened expression of EMS1 (a marker for glycocalyx damage) (Fig. 7b), providing further evidence to support this hypothesis. Additionally, HAdV-7 infection can induce endothelial cell autophagy to enhance viral replication. The presence of HAdV-7 fiber gene copies in the cell supernatant notably increased when treated with the autophagy agonist RAPA(Fig. 8d), while it decreased significantly in the cell supernatant, blood, and lung tissue when exposed to the autophagy inhibitor 3MA/CQ (Fig. 8d–f).This suggests that inhibition of autophagy can inhibit HAdV-7 replication, consistent with recent research on influenza virus [39], novel coronavirus [40], and RSV infection [41] inducing autophagy to affect viral replication. Although the autophagy agonist RAPA did not elevate gene copy numbers in mice and was only effective in cell models, the autophagy inhibitor notably reduced viral load in both mice and cell models. Therefore, there are valid reasons to believe that inhibiting autophagy and reducing its flux can effectively decrease viral load. The precise mechanism through which autophagy impacts viral replication is not yet fully understood; it may involve inducing double-membrane vesicles formation [41] or interfering with autophagosome maturation [42]. It may also involve hijacking autophagosomes or lysosomes to impact viral replication [43]. Further research is necessary to explore how HAdV-7-induced endothelial autophagy influences viral replication.

Our study also demonstrates that HAdV-7 infection induces ALI/ARDS by activating autophagy in pulmonary vascular endothelial cells. Specifically, the autophagy inhibitor CQ was found to significantly reduce lung tissue inflammation (Fig. 9c–e) and pulmonary vascular leakage (Fig. 9f,g). It is essential to emphasize that our study only focuses on the involvement of autophagy in pulmonary vascular endothelial cells during ALI/ARDS development, while the impact of alveolar epithelial and macrophage autophagy on lung injury cannot be disregarded [44, 45]. The observed reduction in lung tissue inflammation may be attributed to the collective inhibition of autophagy in these cell types. Additionally, it should also be noted that this study solely focuses on intervention through autophagy regulators and does not conduct relevant verifications at the level of autophagy genes, which could potentially result in off-target drug effects. Furthermore, a thorough examination of the defense and regeneration functions of endothelial cells is necessary for further advancements.

Conclusions

Our study revealed that HAdV-7 infection induces damage to pulmonary endothelial cells, resulting in elevated VEGF levels, a high viral load, and ALI/ARDS. Activation of endothelial autophagy may be an important mechanism underlying these observations. These findings indicate that targeting autophagy could be a novel approach to rescue pulmonary vascular endothelial cell damage and improve poor prognosis in severe HAdV-7 infection.