Human brain microvascular endothelial cells hCMEC/D3 (purchased from Meisen CTCC, Zhejiang, China) and microglial cells HCM3 (purchased from Meisen CTCC, Zhejiang, China) were cultured in DMEM medium (Gibco, Invitrogen, California, USA) containing 10% fetal bovine serum (FBS) (Gibco), 100 U/mL penicillin and 100 μg/mL streptomycin. Human mast cells LAD2 were cultured in RPMI 1640 medium (Gibco), containing 10% fatal FBS with 100 U/mL penicillin and 100 μg/mL streptomycin. For degranulation, LAD2 cells (purchased from Huzhen Company, Shanghai, China) were grown in StemPro-34 medium (Gibco) supplemented with 100 μg/ml stem cell factor (Novoprotein, Suzhou, China), 100 μg/ml IL-6 (Novoprotein), nutrient supplement (Gibco), 100 U/ml penicillin (Invitrogen, California, USA), 100 µg/ml of streptomycin (Invitrogen), and 2 mM L-Glutamine (Gibco).

HIV-NL4-3/spike pseudotyped virus was generated by Lipo 2000 transfection reagent (Invitrogen)-mediated co-transfection of HEK293T cells, with the spike-expressing plasmid pcDNA3.1-2019-nCoV-S-IRES (strain 2019-nCoV WIV04) and pNL4-3. Luc. ΔR ΔE. These two plasmids and HEK293T cells are provided by Dr. Lu Lu (Fudan University, Shanghai, China). Harvested HIV-NL4-3/spike pseudovirus were used to infect hCMEC/D3 and HMC3 cells for 48h, viral infection was measured by detecting luciferase activity (Promega, Madison, WI, USA).

For detecting the expression of HLA-DR and ACE2, cells were incubated with anti-human HLA-DR allophycocyanin (APC) (Tonbo Biosciences, San Diego, USA, 20-9952) or anti-human ACE2-phycoerythrin (PE) (Bioss, Beijing, China, bs-1004R) at 4°C for 30 min and detected with flow cytometry (BD Accuri C6). For detecting the expression of ZO-1, JAM2, and OCLN in hCMEC/D3 cells, cells (1 × 10⁵) were treated with spike-RBD proteins (2 μg/mL) derived from the wild type of SARS-CoV-2 (GenScript, Nanjing, China, Z03483), or co-cultured with LAD2 cells (1 × 10⁵) in the presence or absence of spike-RBD proteins (2 μg/mL), or stimulated with RBD-treated LAD2 cell culture supernatants (300 μL), or histamine (5 μg/mL), tryptase (5 μg/mL, or chymase (5 μg/mL) for 48h. In some samples, loratadine (Lor.). (2.5 μg/mL, Selleck, Houston, Texas, USA), ebastine (Eba.) (1.5 μg/mL, Selleck), or ketotifen fumarate (Ket.) (4 μg/mL, Yuanye Biology, China, S46226), was used to prior-treat cells for 2h before the next treatment. Cells were harvested and blocked with 5% BSA for 1h at room temperature and incubated with primary antibody for 2h. Primary antibodies against ZO-1 (Invitrogen, 402200), OCLN (Invitrogen, OC-3F10), and JAM-2 (Abcam, Cambridge, UK, EPR2489) were used. Cells were washed with FACS buffer and incubated with Alexa Flour 488–labeled goat anti-rabbit (Invitrogen, A11034) for 30 min and were analyzed with flow cytometry (BD Accuri C6).

HMC3 cells were pre-labeled with CFSE (carboxyfluorescein succinimidyl ester) (Invitrogen) for 10 min at 37°C; the reaction was terminated with 40% FBS and washed three times with phosphate buffered saline (PBS), then HMC3 cells were co-cultured with LAD2/RBD or treated with histamine (5 μg/mL), tryptase (5 μg/mL), or chymase (5 μg/mL) for 24h at 37°C. Cells were analyzed with flow cytometry (BD Accuri C6).

For Western blotting, cells were lysed for 30 min at 4°C in lysis buffer (Beyotime, Shanghai, China). After centrifugation for 10 min at 12,000 g, the supernatant was boiled in reducing SDS sample loading buffer and analyzed by SDS-PAGE. The anti-ZO-1 antibody (Invitrogen, 402200), anti-OCLN antibody (Invitrogen, OC-3F10), anti-JAM-2 antibody (Abcam, EPR2489), and anti-GAPDH antibody (Abcam, ab82633), and the horseradish peroxidase–conjugated secondary antibody were used in Western blotting.

Specific pathogen-free 6-week-old female wild-type BALB/c mice were infected intranasally with the 501Y.V2 SARS-CoV-2 variant (B.1.351, GISAID: EPI_ISL_2423556) at a 50% cell culture infective dose (CCID₅₀) of 7 × 10(4). The 501Y.V2 strain was provided by Guangzhou Medical University, Guangdong, China. Five mice were in each group and three mice were in control group. The brains were collected at 3 days post-infection (dpi) for pathological, virological, and immunological analyses. The brains of mice were harvested and fixed in zinc formalin. For routine histology, tissue sections (~4 μm each) were stained with hematoxylin and eosin (H&E) or toluidine blue (T. blue) (Sigma-Aldrich, USA). In brief, the tissue sections were stained for 1h at room temperature, and then washed in distilled water three times, anhydrous ethanol two times, and covered with a coverslip. The sections were analyzed using a Pannoramic MIDI slice scanning apparatus (3DHISTECH, USA). The fixed tissue sections were stained at the Pathology Center of the GIBH, CAS.

Total cellular RNAs were extracted by using TRIzol Reagent (Invitrogen) and then reverse transcribed into cDNA with synthesis Kit (TOYOBO, Shanghai, China, FSQ-301), according to the manufacturer’s instructions. Real-time polymerase chain reaction (PCR) was carried out by using the SYBR qPCR Mix (Genestar, Beijing, China, A33-101) with the following thermal cycling conditions: initial denaturation at 95°C for 2 min, amplification with 40 cycles of denaturation at 95°C for 15 s, primer annealing at 60°C for 15 s, and extension at 72°C for 30s. The data were analyzed by SYBR green-based semi-quantification and normalized with GAPDH. Real-time PCR was performed on the Bio-Rad CFX96 real-time PCR system. The primers and probes for (RT-) PCR were listed in Supplementary Table S1.

hCMEC/D3 cells were treated with spike RBD–treated LAD2 cell culture supernatants or LAD2 normal culture supernatants for 24h. Total RNAs were extracted using Trizol (Invitrogen) according to the manufacturer’s protocol, and ribosomal RNA removed using QIAseq FastSelect-rRNA HMR Kits (QIAGEN, Germany). Fragmented RNAs (average length approximately 200 bp) were subjected to first strand and second strand cDNA synthesis, followed by adaptor ligation and enrichment with a low-cycle according to the instructions of NEBNext UltraTM RNA Library Prep Kit for Illumina (NEB, USA). The purified library products were evaluated using the Agilent 2200 TapeStation and Qubit2.0 (Life Technologies, USA). The libraries were paired-end sequenced (PE150, sequencing reads were 150 bp) at Genewiz (Jiangsu, China) using Illumina HiSeq 3000 platform. The data analysis was according to the reference (Wu et al., 2021).

Graphpad Prism 8.0 was used for statistical analysis. Student’s unpaired two-tailed t-test was performed to analyze significant difference. Significance levels are indicated as *p < 0.05, **p < 0.01, ***p < 0.001.

The autopsies of COVID-19 patients reveal brain tissue lesions, including hemorrhage, perivascular inflammatory cell infiltration, intravascular platelet aggregates, vascular congestion, and so forth (Matschke et al., 2020). We recapitulated these observations in experimentally SARS-CoV-2–infected mice. BALB/c mice were intranasally infected with the SARS-CoV-2 (501Y.V2, 7 × 10⁴ CCID₅₀), then euthanized at 3 days post-infection (dpi) to harvest brain tissues for histopathology analysis. Mock infection was performed using the same amount of PBS (Figure 1A). Compared to the mock-infection control, the H&E staining of brain sections showed the vascular congestion, dilated perivascular space, inflammation cell infiltration, neurofibrillary degeneration, and gliocyte hyperplasia in SARS-CoV-2–infected mice (Figure 1B).

Because neurons, astrocytes, oligodendrocytes, and glial cells can be infected with SARS-CoV-2 through viral binding with ACE2 and TMPRSS2 (McQuaid et al., 2021; Pacheco-Herrero et al., 2021), we next investigated the susceptibility of microvascular endothelial cells and microglial cells to SARS-CoV-2 infection. Both human brain microvascular endothelial cells hCMEC/D3 and the microglial cells HCM3 expressed ACE2 receptor (Figure 1C). The HMC3 cells, but not hCMEC/D3 cells, showed susceptibility to infection with SARS-CoV-2 spike-pseudotyped lentivirus (NL4-3/spike) (Figure 1D).

To investigate the potential role of MCs in SARS-CoV-2 infection–caused brain lesion, the brain tissue sections of SARS-CoV-2–infected mice were stained with T. blue to indicate MCs (Wu et al., 2021, 2022). Compared to mock infection (Figure 1E), the accumulation of MCs around cerebrovascular region was observed (Figure 1F), demonstrating that SARS-CoV-2 infection induced MC accumulation around cerebrovascular region.

MC-degranulated mediators could further activate microglia and other brain cells to release inflammatory and neurotoxic mediators that could cause neuroinflammation and exacerbate neuro-COVID disease severity (Theoharides and Kempuraj, 2023). We thus investigated the role of MC degranulation in inducing inflammatory factors in human microglial cells and brain microvascular endothelial cells (Figure 2A). The MC cell line LAD2 can be triggered by SARS-CoV-2 spike/RBD protein for rapid degranulation (Wu et al., 2021, 2022). The spike/RBD protein-triggered MC degranulation led to release of factors into cell culture supernatants (Deg. Supern.), which were used to treat hCMEC/D3 and HCM3 cells for 24h. Compared to the treatment with normal cell culture supernatant of LAD2 (supern.), the treatment with MC Deg. Supern. significantly stimulated the expression of inflammatory factor IL-6, IL-8, TNF-α, CCL20, and CXCL5 in HCM3 cells, and elevated the expression of CXCL5, SAA2, SAA4, SERPINA1, and SERPINA3 in hCMEC/D3 (Figure 2B). Whereas the direct stimulation with SARS-CoV-2 spike/RBD protein for 24h did not induce the expression of these inflammatory factor in either cell line (Figure 2B).

MC granules include multiple biologic mediators including histamine, serotonin, heparin, cytokine/chemokines, and enzymes such as chymase and tryptase (Elieh Ali Komi et al., 2020). We have previously detected the rapid release of histamine, chymase and tryptase in SARS-CoV-2 spike/RBD protein-treated LAD2 cells (Wu et al., 2021, 2022). To profile the mediators-induced expression of inflammatory factors, cells were treated with histamine, tryptase or chymase. The treatment with histamine, but not with tryptase or chymase, induced the significant expression of inflammatory factors (Figure 2C).

Microglial activation and proliferation are the common pathology found in the brain autopsy samples of COVID-19 individuals (Matschke et al., 2020; Theoharides and Kempuraj, 2023). We investigated the stimulation of MC degranulation on microglial activation and proliferation. Microglial activation can be determined by the elevated immunostaining of markers HLA-DR, IBA1 (ionized calcium-binding adaptor molecule 1), and CD68 (Al-Dalahmah et al., 2020; Kantonen et al., 2020; Matschke et al., 2020; Rutkai et al., 2022). The treatment of HCM3 cells with MC Deg. Supern. significantly elevated HLA-DR expression (Figure 2D). To confirm this, the LAD2/HMC3 cell co-culture was stimulated with spike/RBD protein, and elevated HLA-DR expression was observed (Figures 2E, F). In comparison, the direct treatment with spike/RBD protein did not induce HLA-DR expression in HMC3 cells (Figures 2D–F). The released histamine might account for the induction of HLA-DR, as the treatment with histamine, but not with tryptase or chymase, elevated HLA-DR expression (Figure 2G). Spike/RBD protein-triggered degranulation during the LAD2/HMC3 cell co-culture induced HMC3 proliferation, as detected with CFSE staining and FACS analysis (Figure 2H). Taken together, these results demonstrate that MC degranulation induces inflammatory factors in microglia and brain microvascular endothelial cells, and induces microglial activation and proliferation.

To globally monitor MC degranulation-induced alteration of cellular signaling in human brain microvascular endothelial cells, the transcriptome was analyzed. hCMEC/D3 cells were treated with spike/RBD protein-triggered LAD2 cell Deg. Supern. for 24h. Transcriptomes were analyzed using standard protocols. Compared with the treatment with normal LAD2 cell culture supernatant (Supern.), the treatment with Deg. Supern. upregulated 1,300 and downregulated 1,816 genes (Figure 3A). Gene ontology functional enrichment analysis of differentially expressed genes (DEGs) showed obvious upregulation of gene sets involved in inflammation response and downregulation of gene sets involved in cell junction organization and regulation of cell adhesion (Figure 3B). The gene set enrichment analysis (GSEA) linked the upregulated genes to the regulation of inflammation response and the downregulated genes to the regulation of cell adhesion and cell junction (Figure 3C). The core DEGs involved in the regulation of cell adhesion/junction and inflammation were catalogized. The upregulated core DEGs linked to inflammation response mainly included CCL20, CXCL5, CX3CL1, SERPINA3, SERPINA1, SAA2 and SAA1, and the downregulated core DEGs linked to cell adhesion/junction mainly included MYH9, VCL, CLDN11, and OCLN (Figure 3D).

The core DEGs were confirmed with real-time PCR. The addition of SARS-CoV-2 spike/RBD protein in the co-culture of hCMEC/D3-LAD2 cells significantly elevated the expression of SERPINA3, SERPINA1, CXCL5, and SAA2 and downregulated the expression of VCL, OCLN, and CLDN11 in hCMEC/D3 cells (Figure 3E). The direct stimulation of hCMEC/D3 with spike/RBD protein did not alter these expressions (Figure 3E).

We have above demonstrated that the released histamine was a key factor for triggering the production of inflammatory cytokines. To verify this, the antihistamines Lor., Eba., and Ket. were used in the next set of experiments. hCMEC/D3-LAD2 cells were treated with these antihistamines prior to co-culture, and this led to reduced expression of inflammatory factors in hCMEC/D3-LAD2 cells during subsequent co-culture (Figure 3E). Taken together, this transcriptome analysis reveals the alteration of cellular signaling triggered by MC degranulation in human brain microvascular endothelial cells.

We have recently demonstrated that spike/RBD protein-induced MC degranulation disrupted barrier integrity of alveolar epithelial and pulmonary microvascular endothelial cells (Wu et al., 2021; Wu et al., 2022), and MC accumulation can be found around cerebrovascular region in SARS-CoV-2–infected mice. In this study, we further examined the effect of MC degranulation on the integrity of human brain microvascular endothelial cells, by detecting the expression of tight junction proteins (Figure 4A). Spike/RBD protein-treated LAD2 cell Deg. Supern. was used to treat hCMEC/D3 cells for 48h, the expression of JAM2, OCLN, and ZO-1 were detected. The treatment with Deg. Supern. reduced these proteins, as detected with flow cytometry (Figure 4B). To confirm this, the LAD2-hCMEC/D3 cell co-culture was used. The addition of spike/RBD protein to the co-culture reduced the expression of these proteins in hCMEC/D3 cells (Figures 4C–E). As expected, the direct treatment of hCMEC/D3 cells with spike/RBD protein for 48h did not downregulate the expression of these tight junction proteins (Figure 4E). Moreover, the prior treatment of hCMEC/D3 cells with antihistamines Lor., Eba., and Ket. rescued the reduction of tight junction proteins (Figures 4D, E).

We have previously demonstrated that the matrix metallopeptidase could be induced by MC degranulation in alveolar epithelial cells to disrupt the barrier integrity (Wu et al., 2021). The addition of spike/RBD protein to the co-culture of LAD2/hCMEC/D3 induced the MMP11 expression in hCMEC/D3 cells, and prior treatment of hCMEC/D3 cells with antihistamines reduced MMP11 induction (Figure 4F). Taken together, these data demonstrate that MC degranulation disrupts the tight junction proteins in brain microvascular endothelial cells.