2.1 Mouse-adapted SARS-CoV-2 infection causes fatal pneumonia in the mid-aged mice

The BALB/c mice (female, young: 6 w/o, mid-aged: 30–50 w/o) were inoculated intranasally with 1 × 10⁴ pfu of the mouse-adapted SARS-CoV-2 (MA-P10). The mid-aged mice lost their body weight from 2 days postinfection (dpi) and died within 7 dpi without any weight recovery (Figure 1a,b). By contrast, the young mice showed slight body weight loss; however, their body weight recovered without lethal infections (Figure 1a,b). The lungs excised from the virus-infected mice had red color, thus suggesting the existence of pulmonary inflammation (Figure 1c). The lungs of the mid-aged mice specifically showed hepatization and increased body weights, indicating severe lung congestion (Figure 1d). The levels of viral RNA and infectious virus titer were significantly higher in the infected mid-aged mice than in the infected young mice (Figure 1e,f).

2.2 Histopathological analysis of the mouse lungs

Histopathological analysis of the lungs of young and mid-aged mice had been carried out at 4 dpi to examine the differences in pulmonary pathology of the infected mice and of the human cases of severe COVID-19, as previously reported (Paul et al., 2022). Pulmonary hemorrhages, infiltration of the inflammatory cells, and vasculitis with perivascular accumulation of the inflammatory cells were observed only throughout the lungs of infected mid-aged mice (Figure 2a). Notably, many thrombi were also observed in the mid-aged group of mice (Figure 2a).

Immunohistochemistry showed that the SARS-CoV-2 nucleocapsid (N) signals in the whole lungs of the mice were significantly higher in the mid-aged mice group than in the young mice group. Additionally, there was some difference in the alveolar area between the young and the mid-aged mice groups (Figure 2b), which is consistent with the difference in viral loads recorded between the young and the mid-aged mice groups (Figure 1e,f).

The CD68-positive area observed in the whole lungs of the mice tended to be larger; however, it was smaller in the N protein signal-rich area in the infected mid-aged mice than in the young mice (Figure 2c). Infiltration of the MPO-positive neutrophils was more evident in the mid-aged mice (Figure 2d). The ratio of CD11b⁺Ly6G⁺ cells (Figure S1A) and CD11b⁺Ly6C⁺ cells (Figure S1B) in the whole lungs of the mice tended to increase in the infected mid-aged mice group than in the young mice group, suggesting that more neutrophils and macrophages tended to infiltrate into the lung tissues of the mice upon viral infection. To detect the presence of thrombi in the virus-infected lungs, a CD41 (platelet marker) immunostaining was performed. The number of vessels containing CD41-positive platelet aggregates was significantly higher in the mid-aged mice group than in the young mice group (Figure 2e,f), suggesting that thrombus formation had occurred in the infected mid-aged mice, as it occurs in the human COVID-19 cases. Additionally, immunohistochemistry of the mice lungs showed more N protein in the alveolar–capillary area of the mid-aged mice than in the young mice. N protein signals were larger in mid-aged mice than in young mice, suggesting increased infection of Type II alveolar epithelial cells or macrophages (Figure 2g). Furthermore, fluorescent double immunostaining of CD31 and SARS-CoV-2 spike (S) showed the presence of S protein signal surrounded by CD31 signal, which is expressed on the endothelial cell membrane, supporting the uptake of virus by pulmonary endothelial cells as shown in Figures 1, 2h, and Video S1.

2.3 Isolation of the pulmonary ECs from the SARS-CoV-2-infected mice and transcriptome analysis

To address the pathology of the pulmonary ECs in severe COVID-19 infection, the gene expression profile of ECs in the young and the mid-aged mice that were infected with MA-P10 was investigated. For RNA-Seq analysis, the pulmonary ECs were isolated from the lungs of the young and the mid-aged mice with/without the viral infection.

The ECs were isolated as CD31⁺CD45⁻ cells from the lungs of 10 mice in each group, through a combination of a magnetic cell separation system (MACS) and a flow cytometric cell sorting, according to our previous method of EC isolation (Hida et al., 2004; Tsumita et al., 2022; Figure S3A). The recovered ECs from each group were 93%–96% CD31⁺CD45⁻, indicating high purity of the isolated ECs. Additionally, approximately 67%–84% of the isolated ECs were PI-negative, demonstrating that the viability of the recovered ECs was enough to be used for subsequent RNA-Seq (Figures 3a and S3B–D). The levels of viral RNA in ECs that were isolated from the lung tissues of the infected mice were quantified using quantitative reverse transcription polymerase chain reaction (qRT-PCR). The results showed that the viral RNA was 14-fold higher in ECs of the mid-aged mice lungs than in ECs of the lungs of the young mice, suggesting that more viral uptake had occurred in ECs of the mid-aged mice lungs (Figure 3b). The expression of the cluster of differentiation 147 (CD147) and neuropilin 1 (NRP1) among the molecules that were involved in SARS-CoV-2 entry was observed in all ECs, whereas the expression levels of ACE2 and TMPRSS2 (an EC surface protein) that represent the key molecules involved in viral uptake were low (fragments per kilobase million: FPKM < 10). Additionally, TMEM106B, as an alternative receptor for SARS-CoV-2 entry into ACE2-negative cells, was expressed at very low levels (Figure 3c). These findings revealed that viral entry into the pulmonary ECs was independent on the ACE2 expression, which was different from the canonical route of the SARS-CoV-2 entry. In the pathological analysis, more viral antigens around ECs in the mid-aged mice group were observed; thus, we investigated whether the cultured human umbilical vein ECs (HUVECs) were susceptible to viral infection and uptake or not. The levels of viral RNA in ECs increased in an MOI-dependent manner, indicating that viral uptake had occurred in ECs in the form of an increase in the amount of viral exposure (Figure 3d). The results of RNA-Seq of the mouse pulmonary vascular endothelium revealed that the expression levels of DDX58, a gene encoding the nucleic acid recognition sensors RIG-I and IRF7 that was downstream of DDX58, were significantly upregulated in the infected mid-aged mice group (Figure 3e). These findings suggest that SARS-CoV-2 was uptaken and stimulated an innate immune response in the vascular endothelium.

2.4 Upregulation of the inflammation-related genes in the pulmonary ECs upon SARS-CoV-2 infection

Differential expression analysis of individual genes was performed between pulmonary ECs from young mice and those from mid-aged mice treated with PBS or SARS-CoV-2 using the DESeq2 package, following the criteria of an adjusted p-value < 0.05 (Figure 4a). Comparing the effect of SARS-CoV-2 infection and the control, significant gene expression changes were found in the mid-aged mice (upregulated: n = 910 genes, downregulated: n = 491 genes) compared with those in young mice (upregulated: n = 352 genes, downregulated: n = 241 genes). Comparing the mid-aged and young mice, 485 significantly upregulated and 330 significantly downregulated genes were observed in the SARS-CoV-2-infected groups, whereas 169 significantly upregulated and 178 significantly downregulated genes were detected in the uninfected group (Figure 4a). To recognize the canonical pathways and networks caused by SARS-CoV-2 infection of the pulmonary ECs in both the young and mid-aged mice, the ingenuity pathway analysis (IPA) was applied. The graphical summary illustrated an outline of the connections between the key biological themes and the molecules in the IPA core analysis, which were regulated through SARS-CoV-2 infection. “Role of hypercytokinemia/chemokinemia in the pathogenesis of influenza,” is a common key of biological themes, which was upregulated by SARS-CoV-2 infection between the young and mid-aged mice groups. This was accompanied by an enhancement of the Interferon-regulated factor 3 (IRF3) and IRF7, which represent the two pivotal transcription factors that manage the production of IFN-α and IFN-β (Figure S4). The top five significant canonical pathways are listed for each of the young and mid-aged mice groups (Figure 4b). Moreover, “role of hypercytokinemia/chemokinemia in the pathogenesis of influenza,” was the most significantly affected pathway by the viral infection in both young and mid-aged mice groups, followed by the activation of the IRF through the cytosolic pattern recognition receptors (Figure 4b). Heatmaps of the differences in the expression levels of the genes listed as “role of hypercytokinemia/chemokinemia in the pathogenesis of influenza” between the four groups of ECs demonstrated that the expressions of CXCL10 (C-X-C motif chemokine ligand 10), EIF2AK2, DDX58, and IRF7 that were involved in the viral response were also highly expressed in the mid-aged mice lung ECs (Figure 4c). The bubble chart expressed the predicted differences in the activity of the canonical pathway in ECs, which were recovered from the young or mid-aged mice infected with SARS-CoV-2 (Figure 4d). These pathways were sorted according to the statistical significances and were colored according to the predicted activation status, which was indicated by a z-score. The number of genes that overlapped each pathway was indicated by the sizes of the bubbles. Additionally, the bubble chart showed the inflammation-related pathways. The role of hypercytokinemia/hyperchemokinemia in the pathogenesis of influenza, the pathogen-induced cytokine storm signaling pathway, and the toll-like receptor signaling were more activated in the mid-aged mice than in the young mice. Besides the inflammation-related pathways, the hypoxia-related pathways, including HIF-1a signaling, VEGF signaling, and hepatic fibrosis signaling, were also more enhanced in the mid-aged mice than in the young mice (Figure 4d). The gene set enrichment analysis (GSEA) demonstrated that the genes in HALLMARK inflammatory response were more highly expressed in the mid-aged mice group than in the young mice group (Figure 5a). The FPKM of CCL2, CXCL2, and CXCL10 in ECs was comparable among the four groups. The CXCL10 expression was significantly upregulated upon viral infection in both of the young and mid-aged mice groups. However, the CXCL2 were significantly upregulated on viral infection in the mid-aged mouse ECs. The proinflammatory cytokines such as interleukin-1β (IL-1β), which are related to the severity of COVID-19, were not significantly different among the four groups. Conversely, the signal transducer and activator of transcription 3 (Stat3) that is activated by the JAK–STAT pathway downstream of IL-6 was upregulated in the infected mid-aged mice group (Figure 5b). The inflamed ECs are known to increase the expression of the adhesion molecules to the immune cells, including the leukocytes (Dayang et al., 2019). The GSEA showed that ECs in the mid-aged mice group had more active “leukocyte adhesion to vascular endothelial cells” signaling than the young mice ECs (Figure 5c). In this regard, the expression levels of VCAM1, vWF (von Willebrand factor), P-selectin (a type-1 trans-membrane protein), and E-selectin (an adhesion receptor) increased significantly upon viral infection in the mid-aged mice ECs. Alternatively, there was no significant difference in ICAM-1 (intercellular adhesion molecule 1), although there was a tendency of upregulation of ICAM-1 in the mid-aged mice ECs (Figure 5d).

2.5 Upregulation of the coagulation-related genes in ECs of mid-aged mice on SARS-CoV-2 infection

The GSEA showed that “coagulation-related genes” were more upregulated in the mid-aged mice ECs than in the young mice ECs (Figure 6a). Using IPA, the expression of coronavirus network and the blood coagulation pathway genes were also recorded to be upregulated by a viral infection in the mid-aged mice ECs (Figure S5). A similar result was obtained on comparative analysis between the young and the mid-aged mice groups (Figure 6b). Thus, the IPA network involved in blood coagulation demonstrated that more molecules were activated in the virus-infected mid-aged mice group. The heatmap of the coagulation-related genes represented the difference in the levels of gene expression among the four groups of mice ECs. The genes that were upregulated in the virus-infected mid-aged mice group included F3, PLAT, and others (Figure 6b). The FPKM levels of Fbn1, F3, PLAT, Pf4, IL-1β, Tlr4, PAI-1, vWF, and P-selectin, which are the well-known thrombogenic molecules, were higher in ECs of the virus-infected mid-aged mice group (Figure 6c). Notably, the expression levels of PLAT were significantly upregulated upon viral infection in the mid-aged mice ECs. Moreover, the PLAT expression was significantly higher in the infected mid-aged mice ECs than in the infected young mice ECs. Meanwhile, SERPINE1 (PAI-1) and P-selectin showed a significant increase on viral infection of the mid-aged mice ECs only (Figure 6c).

The P-selection and vWF were expressed in the initial phase of coagulopathy, unlike the PLAT, PAI-1, and TF, which were upregulated in the late phase (fibrinolytic system). Moreover, the P-selection and vWF have been suggested as biomarkers for severe COVID-19 (Gorog et al., 2022). The results of enzyme-linked immunosorbent assay (ELISA) analysis of the blood levels of the virus-infected mice expressed higher levels in the mid-aged mice group than in the young mice one (Figure 6d). Immunohistochemistry demonstrated P-selectin-positive ECs, which were identified in the N protein signal-rich area of the mid-aged mice lungs, thus suggesting that the virus-uptaking ECs had activated the platelets in the mid-aged mice group (Figure 6e,f).

4.1 Experimental design

This study was designed to address whether the endothelial cells are infected by SARS-CoV-2 and endothelial pathophysiology using a mouse-adapted SARS-CoV-2-infected mouse model. Pulmonary endothelial cells were isolated and purified to compare their transcriptome between nonsevere/severe (young/mid-aged) mice. Pathophysiology of the pulmonary endothelial cells upon virus infection was comprehensively analyzed using RNA sequencing.

4.2 Cell cultures

The African green monkey kidney-derived Vero E6 cells (ATCC) were maintained in Dulbecco's modified Eagle's medium (DMEM) that was supplemented with 10% heat-inactivated fetal bovine serum (FBS) (v/v) and then incubated at 37°C in a humidified atmosphere containing 5% CO₂. The Vero E6 cells stably express the human type II transmembrane serine protease (VeroE6/TMPRSS2) cells, which were described previously by Sasaki et al. (2021). The HUVECs (Lonza) were maintained in EC growth medium-2 (EGM-2; Lonza). Finally, the mouse oral squamous cells (MOC1; Kerafast) were maintained in Iscove's modified Dulbecco's medium.

4.3 siRNA transfection

Three stealth small interfering RNA (siRNA) for the mouse ACE2 (mACE2) receptor was obtained through the BLOCK-iT RNAi design program (Invitrogen). Transfection was accomplished using the lipofectamine RNAiMAX reagent (Invitrogen) for the MOC1 cells as previously reported (Torii et al., 2021). For siRNA targeting the mACE2, the used sequences were 5′-CCAGGCAACCUUUCCUGCUAAGAAA-3′ and 5′-UUUCUUAGCAGGAAAGGUUGCCUGG-3′. The stealth control siRNA (Invitrogen, Medium GC, sequence not available) was used as a control.

4.4 Animal experiments

All the applied animal experiments were approved by the Ethical Committee for Experimental Animal Care of Hokkaido University, Japan (approval number: 20-0113). Approximately 6-week-old or 36-week-old female BALB/c mice that were obtained from CLEA Japan (Tokyo, Japan) or from SLC Japan (Shizuoka, Japan) were used in the present experiments. Challenge SARS-CoV-2 virus [1.0  ×  10⁴ pfu (plaque-forming unit) 50 μL⁻¹] was injected through intranasal instillation, which was diluted in a phosphate-buffered saline (PBS), as previously described by Leist et al. (2020). At 4 dpi, the lungs were excised using a sterile scalpel after collecting the blood samples through a cardiac puncture following euthanasia. The treated mice that reached the endpoint before 4 dpi were included in Figure 1b only.

4.5 Viral preparation

The SARS-CoV-2 WK-521 (EPI_ISL_408667) viral strain was kindly provided by Dr. Saijo (National Institute of Infectious Diseases, Tokyo, Japan). The mouse-adapted SARS-CoV-2 strain (MA-P10) was established from this SARS-CoV-2 WK-521 virus stock as previously described (Uemura et al., 2023). All experiments involving the use of SARS-CoV-2 were performed in the Biosafety Level-3 (BSL-3) facility of Hokkaido University, in accordance with the institutional guidelines.

4.6 RNA isolation and amplification using qRT-PCR

The total RNA was extracted from the inoculated mouse cells or lung tissues using TRIzol Reagent (Thermo Fisher Scientific) and RNeasy Mini Kit (QIAGEN). The extracted RNAs were subjected to quantification using qRT-PCR analysis; with the THUNDERBIRD Probe One-step RT-qPCR Kit (TOYOBO). A primer probe sets for N2 (Takara Bio) was used. The endogenous expression level of β-actin was quantified as an endogenous control with primer and probe sets for the nonhuman primate β-actin (Overbergh et al., 2005) and the mouse β-actin (Mouse ACTB Endogenous Control, 4352933E, Thermo Fisher Scientific). The levels of SARS-CoV-2 N gene were normalized to that of β-actin. All qRT-PCR assays were performed using the CFX96 Real-Time PCR System (BioRad).

4.7 Viral entry assay

The HUVECs were inoculated with SARS-CoV-2 WK-521 at different multiplicity of infection (MOI = 0, 0.1, 1, and 10). After incubation at 37°C for 24 h, the cells were washed with 1 × PBS, trypsinized, quenched in DMEM supplemented with 2% FBS, centrifuged (300 × g, 3 min, 4°C), washed again with 1 × PBS, and centrifuged, and finally, the cell pellet was lysed in a TRIzol Reagent (Thermo Fisher Scientific). The intracellular virus was analyzed quantitatively using real-time PCR (RT-PCR).

4.8 The viral plaque assay

The mouse lungs were homogenized in 3 mL of Hank's balanced salt solution by TissueRuptor II (QIAGEN), and centrifuged at 1000 × g for 5 min at 4°C. The supernatants were filtrated using 0.45-μm Millipore syringe filter (Sartorius) and then serially diluted with DMEM containing 2% FBS. The diluted mixture was inoculated into the VeroE6/TMPRSS2 cells and incubated at 37°C for 1 h with agitation. After incubation, the cells were overlaid with 2% FBS DMEM containing 0.4% Bacto agar (Becton Dickinson). After 48 h of incubation at 37°C, the cells were fixed with 3.7% buffered formaldehyde overnight and then stained with 1% crystal violet. The number of plaques was enumerated, and the viral titer was calculated in pfu mL⁻¹.

4.9 Histopathological analysis

The formalin-fixed paraffin-embedded lung tissue sections (4.25 μm in thickness) were stained with hematoxylin and eosin (H&E). An immunohistochemistry assay was employed using several primary antibodies against SARS-CoV-2 spike protein (mouse, GTX632604, GeneTex), SARS-CoV-2 nucleocapsid protein (rabbit, GTX635679, GeneTex), CD31 (rabbit, ab28364, Abcam), CD41 (rabbit, ab134131, Abcam), CD45 (rat, #103202, Biolegend), MPO (rabbit, ab9535, Abcam), CD68 (rabbit, ab125212, Abcam), and P-selectin/CD62P (rabbit, ab255822, Abcam). The used secondary antibodies include Goat Anti-Rabbit Immunoglobulins/HRP (P0448, Dako), Goat Anti-Mouse Immunoglobulins/HRP (P0447, Dako), and Rabbit Anti-Rat Immunoglobulins/HRP (P0450, Dako). The liquid DAB⁺ Substrate Chromogen System (K3468, Dako) was used for color development. DAB signal-positive areas of CD41, CD45, MPO, and CD68 were quantified using an ImageJ software (NIH). The area with a particularly high (N) protein positive signal was defined as “N protein signal-rich area,” and the positive areas of CD41, CD45, MPO, and CD68 were quantified using an ImageJ software (NIH).

4.10 Immunofluorescence imaging of mouse lung ECs

The formalin-fixed paraffin-embedded lung tissue sections (8 μm in thickness) were dewaxed, rehydrated, and heat-treated for 30 min in 10 mM Tris/1 mM EDTA buffer (pH 9.0) and cooled to room temperature. After cooling, slides were washed in PBS and incubated with 0.5% H₂O₂ for 10 min at RT to inhibit endogenous peroxidases. After PBS wash, sections were incubated with 5% Goat serum in 1 × PBS for 1 h at room temperature and incubated overnight at 4°C with primary antibodies against SARS-CoV-2 spike protein (mouse, GTX632604, GeneTex) and CD31 (rabbit, ab28364, Abcam) in PBS containing 5% goat serum. For negative controls, sections were incubated without the primary antibody or mouse and rabbit isotype antibody controls. Sections were rinsed and incubated with Goat anti-Mouse IgG (H + L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 594 (Invitrogen, A-11032) and Goat anti-Rabbit IgG (H + L) Highly Cross-Adsorbed Secondary Antibody (Invitrogen, A32731) for 1 h at RT. Nuclei were counterstained with DAPI (Dojin chemical). TrueVIEW Reagent (TrueVIEWTM Autofluorescence Quenching Kit, Vector Laboratories) was used according to manufacturer instructions to reduce autofluorescence. Sections stained with secondary antibodies alone showed no specific staining.

4.11 Confocal fluorescence imaging and three-dimensional (3D) reconstruction

Confocal imaging of the lung tissue sections stained with immunofluorescence described above was performed using a Nikon confocal laser microscopy system (A1-ECLIPSE Ti2; Nikon). An objective was used: PLAN APO l D (×40/NA = 0.95). Three laser lines at 407, 488 and 561 nm for excitation and three filter cubes at 400 nm/480 nm, 480 nm/560 nm, and 560 nm/640 nm for detection were used. The images with 0.31 μm/pixel of 1024 × 1024 pixels, and a 12-bit color depth were acquired. For detailed 3D fluorescence morphometry, confocal images with × 5 optical zoom were taken with 0.5 μm step sizes for voxel sampling. The image size of x-y plane was 0.06 μm/pixel of 1024 × 1024 pixels. The 3D fluorescence images were constructed from z-series images using the IMARIS software program (Bitplane) as described previously (Sci Rep. 2022 Oct 7;12(1):16799. doi: 10.1038/s41598-022-20,793-5).

4.12 Isolation of the mouse lung ECs

The ECs were isolated from the mouse lung tissues following the protocol described previously (Cong et al., 2021; Hida et al., 2004; Kikuchi et al., 2020), with slight modifications.

Briefly, the excised lung tissues were minced using a sterile scalpel and then digested with collagenase II (Thermo Fisher Scientific) for 1 h at 37°C. After incubation, the blood cells were removed using a lysing buffer (BD Biosciences). The lung cell suspensions were then incubated at 4°C for 10 min with FcR blocking reagent (Miltenyi Biotec), followed by CD31 microbeads (Miltenyi Biotec). The CD31-positive cells were isolated as lung ECs using a magnetic cell separation system (MACS; Miltenyi Biotec), according to the manufacturer's instructions. To improve the purity of the lung ECs, the cells were further sorted from a CD31⁺CD45⁻ population with Alexa Fluor 488-conjugated anti-mouse CD31 rat antibody (#102414, Biolegend) and Pacific Blue-conjugated anti-mouse CD45 rat antibody (#103125, Biolegend), through a flow cytometric cell sorter (FACS Melody; BD Biosciences). The data were acquired using a FACSMelody™ Cell Sorter and a BD FACSChorus™ Software (BD Biosciences) and then analyzed using a FlowJo software (Treestar).

4.13 The flow cytometric analysis of the mouse lungs

The flow cytometric analysis was performed on the lung cell suspensions, as indicated in “isolation of the mouse lung ECs.” The used antibodies included Brilliant Violet 510™ anti-mouse/human CD11b rat antibody (Clone M1/70; #101245, Biolegend), Alexa Fluor® 488 anti-mouse Ly-6G rat antibody (#127626, Biolegend), and Alexa Fluor® 488 anti-mouse Ly-6C rat antibody (#128022, Biolegend).

4.14 Extraction and quality assessment of the RNA

The purified lung ECs were centrifuged (800 × g, 20 min, at 4°C), and the total RNA was extracted using a TRIzol™ LS (Thermo Fisher Scientific) and a RNeasy Mini Kit (QIAGEN). The quality of the extracted RNA was assessed using RNA 6000 Nano Assay kit and Agilent 2100 Bioanalyzer (Agilent Technologies). The RNA integrity number was generated and analyzed using a Bioanalyzer 2100 Expert Software (Agilent Technologies).

4.15 RNA-Seq

RNA-Seq was conducted three times independently (Exp. 1, 2, and 3). For Exp. 1, the sequencing library for RNA-Seq was prepared using a Smartseq v4 Ultra Low Input Kit (Takara Bio) and Nextera XT DNA Library Prep Kit (Illumina) and then subjected to deep-sequencing with NovaSeq 6000 (Illumina). For Exp. 2 and 3, the sequencing libraries were prepared with a SmartSeq Stranded Kit (Takara Bio) and then subjected to deep-sequencing with NovaSeq 6000 (Illumina).

Read data were aligned to the mouse GRCm39 reference genome using STAR v2.7.10b (Dobin et al., 2013) with—quantMode Transcriptome SAM and—outSAM type BAM SortedByCoordinate option parameters. Assembling transcripts, merging assembled transcripts, and generating read coverage tables were performed using StringTie v2.1.5 (Pertea et al., 2015). Read coverage tables were converted to a DESeq2-compatible form using prepDE.py3. Differential expression gene analysis was performed using DESeq2 v1.38.3 (Love et al., 2014) following standard protocols. The StringTie output file was used to count the data, which were tagged with sample group names and batch numbers. The DESeqDataSet object was constructed with ~batch + sample_group option in the design argument to remove batch effects. Differential expression gene analysis between each sample group was performed. The log-fold change shrinkage function was applied for gene ranking and data visualization. The EnhancedVolcano package v1.16.0 (Blighe et al., 2023) was used to generate volcano plots. We constructed custom reference genomes for SARS-CoV-2 viral titer measurements by concatenating the SARS-CoV-2 WK-521 (EPI_ISL_408667) sequence with the mouse genome (GRCm39). Before the FPKM calculation, a modified StringTie merged.gtf file was obtained with unique_gene_id.py (Boyle, 2020) to avoid gene ID conflict. The FPKM calculation was performed using Ballgown v2.30.0 (Frazee et al., 2015).

4.16 Pathway and network analysis

The IPA (QIAGEN) is a web-based software application that was used to assess whether the molecules or the pathways of coagulation and inflammation (generated through use of QIAGEN IPA by QIAGEN Inc., Hilden, Germany, https://digitalinsights.qiagen.com/IPA) were activated or inhibited in the lung ECs of the infected mid-aged mice, which were related to SARS-CoV-2 severity. The Ensemble ID number, the absolute TPM values from RNA-Seq data (Exp. 1), and the log fold change value were imported into the IPA. The IPA core analysis was first carried out to identify the significantly affected candidate pathways. The significance of association between the gene expression data and the canonical pathways was determined by estimating the ratio of the number of overlapping genes from our dataset to the total number of IPA dataset genes in a particular pathway (Kramer et al., 2014). Afterward, the significance of association between the genes and the canonical pathway was statistically determined using the Fisher's exact test, where p < 0.05 was considered significant (QIAGEN IPA by QIAGEN Inc., https://digitalinsights.qiagen.com/IPA). The result of canonical pathway analysis was illustrated as a graphical abstract, the top five pathways list, and the bubble charts, respectively, to visualize the pathways that were related to the severity of the SARS-CoV-2 infection. A log2 fold change and absolute TPM counts were used for filtering the dataset.

Besides the IPA analysis and in order to identify and provide a visual representation of the deferentially expressed genes in the determined canonical pathways that were affected by SARS-CoV-2 infection, the hierarchically clustering was carried out using the MeV tool (Howe et al., 2011). The results were exported as a heatmap. For the hierarchically clustering analysis, the log Ratio calculated from TPM values were used (Makondi et al., 2018). The list of gene set that was used for the hierarchically clustering was derived from those in the IPA canonical pathway.

4.17 GSEA

The GSEA was employed using the GSEA software version 2.0.13, which was available from the Broad institute website (www.broadinstitute.org/gsea/). The hallmark gene sets were downloaded from the MSigDB database (Subramanian et al., 2005). All gene set files for this analysis were obtained from the GSEA website (www.broadinstitute.org/gsea/). An enrichment map was used for visualization of the GSEA results.

4.18 ELISA

The mouse plasma was obtained from the supernatant of the centrifuged heparinized whole blood, which was collected under isoflurane anesthesia. The concentration of P-selectin (CD62P) in the plasma of each mouse was determined using a sandwich ELISA through the mouse P-Selectin ELISA Kit (CD62P; ab200014, Abcam), according to manufacturer's instruction.

4.19 Statistics

The statistical analyses were performed using GraphPad Prism v9 (GraphPad Software Inc). Data are presented as the mean values ± standard deviation (SD) of the biological assays, which were carried out in triplicates. Statistical analysis of the data with a parametric distribution was carried out using one-way analysis of variance. This was followed by a Dunnett's multiple comparison test for the multiple comparisons and or a two-sided Student's t test for comparison between each two groups. However, the data with a nonparametric distribution was analyzed using a Mann–Whitney U test between each two groups. For all these data sets, a p value of less than 0.05 was considered significant. Meanwhile, ****p < 0.0001, ***p = 0.0001–0.001, **p = 0.001–0.01, and *p = 0.01–0.05 were considered significant.

The RNA-Seq data analysis was performed using DESeq2, as described above. The Wald test was performed to identify differentially expressed genes between the two samples, and the Benjamini–Hochberg false discovery rate (FDR) was calculated for multiple test correction as adjusted p-values. For the RNA-Seq data analysis, ***adjusted p < 0.005, **adjusted p = 0.005–0.01, and *adjusted p = 0.01–0.05 were considered significant.