Construction of SARS-CoV-2 N conditional knock-in mouse

Based on the principles of a constructed SARS-CoV-2 trans-complementary system, we aimed to generate a mouse model expressing hACE2 and the SARS-CoV-2 N protein. Several mouse models expressing hACE2 have been reported for authentic SARS-CoV-2 infection, and we particularly preferred the K18-hACE2 KI mouse among these models.15 Consequently, we designed and constructed a SARS-CoV-2 N conditional knock-in mouse based on this strain. A CAG-loxP-stop-loxP-Kozak-SARS-CoV-2 N-WPRE-polyA cassette was inserted into the Tigre locus on mouse chromosome 9 via CRISPR-Cas9 technology (Fig. 1a). The resulting F0 founder mice were genotyped (Fig. 1b), and the target mice were subsequently bred with Cre driver mice (Rosa26-SA-CreERT2 mice or Sftpc-IRES-iCre mice) to generate Cre-N-hACE2 mice (SA-N-hACE2 or Sftpc-N-hACE2 mice). All primers used for genotyping the mice are listed in Supplementary Table 1.

Fig. 1

Generation of the SARS-CoV-2 N conditional knock-in mouse model. a Schematic diagrams illustrating the knock-in strategy, in which the CAG-loxP-Stop-loxP-SARS-CoV-2 N-WPRE-PolyA sequence was inserted into the Tigre locus on chromosome 9, using the K18-hACE2 KI mouse background (N-hACE2 mouse). The stop sequence was flanked by two loxP sites, and upon expression of Cre recombinase, the stop sequence located between these loxP sites was excised. N-hACE2 mice were crossed with Rosa26 SA-CreERT2 or Sftpc-IRES-iCre mice, resulting in the generation of SA-N-hACE2 (TAM-inducible systemic Cre expression) and Sftpc-N-hACE2 (lung-specific Cre expression) mice. b N-hACE2 mouse genotyping by PCR showed the presence of N gene and hACE2 gene fragments. M: GimiRun DM5000 DNA Marker. c PCR confirmed excision of the stop sequence between loxP sites in SA-N-hACE2 mice after 5 days of TAM injection. Western blot analysis of SARS-CoV-2 N protein expression in multiple tissues from SA-N-hACE2 mice after TAM induction (d) and from Sftpc-N-hACE2 mice (e). f Immunofluorescence staining of lung sections was conducted using 4,6-diamidino-2-phenylindole (DAPI, blue), an anti-SARS-CoV-2 N antibody (red), and an anti-hACE2 antibody (green) to evaluate the expression of SARS-CoV-2 N and hACE2 in lung cells

SA-N-hACE2 mice were administered 75 mg/kg tamoxifen (TAM) every 24 h for five consecutive days, followed by a seven-day feeding period to induce the expression of the SARS-CoV-2 N protein.41 Tail clipping was performed to identify the knockout status of the loxP-stop-loxP (LSL) sequence in the mice (Fig. 1c and Supplementary Table 1). The SARS-CoV-2 N protein was detected in multiple tissues of SA-N-hACE2 mice after TAM induction (Fig. 1d), whereas it was restricted to the lungs of Sftpc-N-hACE2 mice (Fig. 1e). Immunofluorescence (IF) analysis further confirmed the expression of the N and hACE2 proteins in the lungs (Fig. 1f).

SA-N-hACE2 mice are susceptible to SARS-CoV-2 ΔN/GFP-HiBiT

TAM-induced SA-N-hACE2 mice were intranasally inoculated with 5 × 104 TCID50 (low dose) or 1 × 106 TCID50 (high dose) of SARS-CoV-2 ΔN/GFP-HiBiT. Mice were observed daily for clinical symptoms, body weight changes, and survival (Fig. 2a). Tissue samples from the brain, eyes, trachea, lungs, heart, liver, kidneys, spleen, intestines, and testes were harvested at 2, 4, and 7 days post-infection (dpi). The mice in both the low-dose and high-dose groups exhibited slight body weight loss (Fig. 2b), and viral loads were detected in the lung and brain tissues (Fig. 2c). Viral loads in the lungs showed an increase during the first 4 dpi and subsequently decreased from 4 to 7 dpi, peaking at 4 dpi (Fig. 2d, f). This observation was consistent with the clinical manifestations of SARS-CoV-2 infection.42 Furthermore, mice that received a high dose demonstrated higher viral loads in their lungs than did mice in the low-dose group (Fig. 2d). Additionally, the HiBiT detection data reflected the same trend as the viral load results (Fig. 2e). We subsequently conducted histopathological analyses on the infected lung and brain tissues from the high-dose group. The analysis indicated a gradual progression of pneumonia in the lungs, ranging from mild to severe (Fig. 2g, h). At 2 dpi, mild damage was observed in the lungs of the mice. By 4 and 7 dpi, the lungs displayed severe tissue damage characterized by increased infiltration of immune cells and thickening of the alveolar walls. Additionally, no pathological changes were observed in the brains of the high-dose-infected mice compared with mock-infected mice (Fig. 2i, j).

Fig. 2

SA-N-hACE2 mice are susceptible to SARS-CoV-2 ΔN/GFP-HiBiT infection. a Experimental design for intranasal infection. After TAM treatment, SA-N-hACE2 mice were infected either with 5 × 104 or 1 × 106 TCID50 of SARS-CoV-2 ΔN/GFP-HiBiT. Tissue samples were collected at the indicated dpi. b Changes in the weights of the mice are shown (n = 4 per group). c qRT‒PCR was used to quantify viral loads in tissues at 7 dpi (n = 4 per group). The viral loads (d) and luminescence (e) were measured in the lungs collected at 2, 4, and 7 dpi (n = 4 per group). f Immunofluorescence staining of lung sections at 2, 4, and 7 dpi was conducted using DAPI (blue) and an anti-GFP antibody (green) specific to GFP in SARS-CoV-2 ΔN/GFP-HiBiT. g–j Pathological changes observed using H&E staining in lung (g) and brain (i) tissues from SA-N-hACE2 mice challenged with 1 × 106 TCID50 at 0, 2, 4, and 7 dpi. Pathology scores for the lungs (h) and brain (j) were calculated (n = 4 per group). Data are presented as means ± SD. Statistical significance was determined using one-way ANOVA with Dunnett’s multiple comparisons test (h). ns not significant; *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. ND not detected

Sftpc-N-hACE2 mice exhibit lung-specific susceptibility to SARS-CoV-2 ΔN/GFP-HiBiT

Sftpc-N-hACE2 mice were intranasally inoculated with 5 × 104 TCID50 (low dose) or 1 × 106 TCID50 (high dose) of SARS-CoV-2 ΔN/GFP-HiBiT (Fig. 3a). At 2, 4, and 7 dpi, tissue samples were harvested from the brain, eyes, trachea, lungs, heart, liver, kidneys, spleen, intestines, and testes. The data indicated that mice in both the low-dose and high-dose groups exhibited slight weight loss (Fig. 3b). Furthermore, viral loads were detectable only in the lung tissues (Fig. 3c), suggesting a correlation between viral loads and the expression of the N protein. The viral loads in the lungs increased during the first 4 dpi, peaked at 4 dpi, and then declined between 4 and 7 dpi, showing a trend similar to that observed in the lungs of SA-N-hACE2 mice (Fig. 3d, f). HiBiT luminescence served as a reference (Fig. 3e). Additionally, high-dose-infected mice presented higher viral loads in the lungs than low-dose-infected mice (Fig. 3d). Histopathological analyses were performed on lung and brain tissues from the high-dose group. In the lungs, pneumonia progressed from mild to severe (Fig. 3g, h). At 2 dpi, only mild damage was observed. By 4 and 7 dpi, lung damage was more pronounced, with increased immune cell infiltration and alveolar wall thickening. Additionally, no pathological changes were observed in the brains of the infected mice compared with mock-infected mice (Fig. 3i, j).

Fig. 3

Sftpc-N-hACE2 mice are susceptible to SARS-CoV-2 ΔN/GFP-HiBiT infection. a Illustration of the protocol for intranasal infection with tissue samples collected at the indicated dpi. Sftpc-N-hACE2 mice were infected with 5 × 104 or 1 × 106 TCID50 of SARS-CoV-2 ΔN/GFP-HiBiT. b Changes in the body weights of the mice are shown (n = 4 per group). c Viral loads in the tissues obtained at 7 dpi were quantified via qRT‒PCR (n = 4 per group). The viral loads (d) and luminescence (e) were measured in the lungs collected at 2, 4, and 7 dpi (n = 4 per group). f Immunofluorescence staining of lung sections at 2, 4, and 7 dpi was conducted using DAPI (blue) and an anti-GFP antibody (green) specific to GFP in SARS-CoV-2 ΔN/GFP-HiBiT. g–j Pathological changes in the lungs (g) and brains (i) of Sftpc-N-hACE2 mice challenged with 1 × 106 TCID50 at 0, 2, 4, and 7 dpi were assessed using H&E staining. Pathology scores for the lungs (h) and brains (j) are shown (n = 4 per group). Data are presented as means ± SD. Statistical significance was determined using one-way ANOVA with Dunnett’s multiple comparisons test (h). ns not significant; *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. ND not detected

Considering that no SARS-CoV-2 ΔN/GFP-HiBiT was detected in the brain in this infection model, whereas the lungs exhibited more severe damage, we proposed to increase the dosage of infection to determine whether lung infection could lead to the death of the mice. Sftpc-N-hACE2 mice were then intranasally inoculated with 5 × 106 TCID50 of SARS-CoV-2 ΔN/GFP-HiBiT (Fig. 4a). The results revealed that infected mice exhibited significant weight loss starting from 4 dpi, with 25% of the mice reaching the euthanasia criteria or dying directly (Fig. 4b, c). Body weight subsequently began to recover at 8 dpi. The viral loads in the lungs decreased from 7 dpi and were undetectable by 21 dpi (Fig. 4d). Moreover, no viral load was detected in the brain (Fig. 4e). The pathological examination results further demonstrated that lung damage was most severe at 7 dpi, with gradual recovery beginning at this point, and essentially returning to normal by 21 dpi, which is consistent with the observed clinical symptoms (Fig. 4f, h). However, no pathological damage was observed in the brain (Fig. 4g, i), and no infection with SARS-CoV-2 ΔN/GFP-HiBiT was detected by immunofluorescence (Fig. 4j). As assessed by colabeling IBA1 and CD68, the microglia/macrophages in the brains of the Sftpc-N-hACE2 mice showed no significant activation, indicating the absence of evident neuroinflammation (Fig. 4k). Furthermore, iDisco+ clearing and light-sheet imaging of lung tissue from infected mice at 7 dpi confirmed the distribution of SARS-CoV-2 N in lung epithelial cells, as well as the infection pattern of SARS-CoV-2 ΔN/GFP-HiBiT (Fig. 4l).

Fig. 4

SARS-CoV-2 ΔN/GFP-HiBiT infection in Sftpc-N-hACE2 mice can lead to lethal disease. a The schematic outlines the procedure for infecting the mice with 5 × 106 TCID50 of SARS-CoV-2 ΔN/GFP-HiBiT. At 0, 7, 10, 14, and 21 dpi, four mice were euthanized at each time point for sample collection. Mice that experienced a loss of more than 20% of their initial body weight were euthanized as a humane endpoint. Mice were monitored for body weight changes (b) and survival (c) (n = 16). E gene copies in lung (d) and brain (e) tissues were quantified via qRT‒PCR. f–i Pathological changes were assessed using H&E staining in the lungs (f) and brains (g). Pathology scores for the lungs (h) and brains (i) were recorded (n = 4 per group). j Immunofluorescence staining of brain sections was performed using DAPI (blue) and an anti-GFP antibody (green) specific to GFP in SARS-CoV-2 ΔN/GFP-HiBiT. k Immunofluorescence analysis was performed on brain sections from mock-infected mice, K18-hACE2 KI mice challenged with live SARS-CoV-2, and both SA-N-hACE2 and Sftpc-N-hACE2 mice exposed to SARS-CoV-2 ΔN/GFP-HiBiT. The sections were stained with anti-IBA1 (red) and anti-CD68 (green) antibodies. l Light-sheet imaging of cleared lung tissues from Sftpc-N-hACE2 mice at 7 dpi confirmed the distribution of SARS-CoV-2 N in lung epithelial cells and SARS-CoV-2 ΔN/GFP-HiBiT infection. Data are presented as means ± SD. Statistical significance was determined using one-way ANOVA with Dunnett’s multiple comparisons test (h). ns not significant; *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. ND not detected

Immune responses to SARS-CoV-2 ΔN/GFP-HiBiT infection

We analyzed the lungs of SA-N-hACE2 and Sftpc-N-hACE2 mice infected with SARS-CoV-2 ΔN/GFP-HiBiT RDPs through immunofluorescence staining to investigate the temporal dynamics of immune cells (Supplementary Fig. 1a–d). At 2 dpi, there was an early and significant increase in neutrophils (Ly6G+) and macrophages (CD68+). At 4 dpi, neutrophil infiltration peaked, while macrophages continued to accumulate. Concurrently, the numbers of B cells (CD19+), CD4+ T cells (CD3+CD4+), and CD8+ T cells (CD3+CD8+) markedly increased. At 7 dpi, the infiltration of T and B lymphocytes continued to increase, and the number of macrophages remained elevated.

We conducted transcriptome sequencing on the infected mice (Fig. 5a and Supplementary Fig. 3a). The hierarchical clustering heatmap illustrates the Pearson correlation coefficients among all RNA-seq samples (Supplementary Fig. 2a–d). RNA-seq analysis further confirmed that key innate immune cell populations, including neutrophils, plasmacytoid dendritic cells (pDCs), and macrophages, were activated at 2 dpi. At 4 dpi, T cells differentiated into multiple effector subsets and acted synergistically with activated B cells. By 7 dpi, both the effector and memory lymphocyte populations significantly expanded, accompanied by a marked increase in regulatory T cells (Tregs) and myeloid-derived suppressor cells (MDSCs) (Supplementary Fig. 1e).

Fig. 5

Transcriptomic profiling of the lungs of SA-N-hACE2 and Sftpc-N-hACE2 mice revealed distinct molecular features. a Scheme illustrating the transcriptome sequencing. Gene expression heatmaps of IFN-I-related genes (b) and cytokines and chemokines (c) in the lungs of the mice. Mock: PBS-treated mice; the SA-N-hACE2 and Sftpc-N-hACE2 mice were infected with 1 × 106 TCID50 of SARS-CoV-2 ΔN/GFP-HiBiT. d Bubble plot showing gene set enrichment analysis (GSEA) results for pathways enriched in the lungs of the infected mice. The color scale indicates normalized enrichment scores (NES), while the size of the bubbles corresponds to −log10 (p values)

RNA-seq analysis of IFN-I-related genes in mouse lung tissue (Fig. 5b) revealed that at 2 dpi, canonical interferon-stimulated genes (ISGs), such as Ifit1/2/3, Oas2/3, Mx1/2, Isg15, Rsad2, and Bst2, were significantly upregulated, indicating that the host initiated an acute antiviral program primarily aimed at restricting viral replication during the early stage of infection.43,44 From 4 to 7 dpi, genes involved in signaling and inflammatory regulation, including Ifnar1/2, Stat1/2, Tyk2, and Irf1/5/8, were persistently upregulated, suggesting that the IFN-I system may have transitioned from the acute antiviral phase to a stage of signal maintenance and immune regulation, accompanied by sustained upregulation of inflammation-related factors (Fig. 5c).

Gene set enrichment analysis (GSEA) of the lungs revealed significant enrichment of multiple immune-related pathways following infection (Fig. 5d). The RIG-I-like receptor signaling pathway and the JAK–STAT signaling pathway were broadly upregulated, while the chemokine signaling pathway, cytokine‒cytokine receptor interaction, and viral protein interaction with cytokines and cytokine receptors were also markedly increased. These findings indicate that viral infection triggers an immune response centered on cytokines and chemokines through RIG-I-mediated pathogen recognition and JAK–STAT signal amplification, promoting immune cell recruitment and inflammatory reactions. Concurrently, the activation of pattern recognition receptors and pro-inflammatory pathways, including the NOD-like receptor (NLR) signaling pathway, Toll-like receptor (TLR) signaling pathway, NF-κB signaling pathway, and TNF signaling pathway, further suggested that the innate immune system recognized viral pathogen-associated molecular patterns (PAMPs) and stimulated the release of downstream inflammatory factors (such as IL-17 and TNF), which may lead to cytokine storms or tissue damage. Conversely, pathways related to anti-inflammatory responses and tissue repair, including the TGF-β signaling pathway and the Wnt signaling pathway, were significantly downregulated. Suppression of the TGF-β pathway may impair anti-inflammatory mechanisms and hinder alveolar epithelial repair, whereas downregulation of the Wnt pathway may inhibit lung tissue regeneration. Collectively, these effects could exacerbate inflammatory dysregulation and tissue damage.

RNA-seq analysis of mouse brains (Supplementary Fig. 3a) revealed mild upregulation of IFN-I-related genes in SA-N-hACE2 mice post-infection (Supplementary Fig. 3b), accompanied by slight increases in cytokines and inflammatory factors (Supplementary Fig. 3c). Individual mice displayed significant upregulation of IFN-I-related genes as well as cytokines and inflammatory factors. In contrast, Sftpc-N-hACE2 mice presented no observable changes in the expression of IFN-I-related genes, cytokines, or inflammatory factors (Supplementary Fig. 3b, c). Furthermore, several pathway components enriched in the GSEA plot were either undetected or showed no significant alterations (Supplementary Fig. 3d).

Cre-N-hACE2 mice are applicable for antiviral drug and neutralizing antibody evaluation

To evaluate the potential of SA-N-hACE2 and Sftpc-N-hACE2 mice in the development of anti-SARS-CoV-2 drugs, we tested the antiviral drug nirmatrelvir (NMV) and the neutralizing antibody 7B345 in these mouse models. We then compared the results with those observed in K18-hACE2 KI mice infected with SARS-CoV-2 WT viruses. SA-N-hACE2 and Sftpc-N-hACE2 mice were intranasally infected with 1 × 106 TCID50 of SARS-CoV-2 ΔN/GFP-HiBiT RDPs at 0 dpi, whereas K18-hACE2 KI mice were intranasally infected with 3.57 × 102 TCID50 of SARS-CoV-2 WT at 0 dpi. In the NMV group, the mice received 300 mg/kg NMV daily via oral gavage from 0 to 3 dpi. In the 7B3 group, the mice were intraperitoneally injected with 10 mg/kg mAb 7B3 at 1 dpi. All the mice were euthanized for sample collection at 4 dpi (Fig. 6a).

Fig. 6

Assessment of anti-SARS-CoV-2 therapeutics using SA-N-hACE2 and Sftpc-N-hACE2 mouse models. a This schematic diagram illustrates the procedure for evaluating the neutralizing antibody 7B3 and NMV. SA-N-hACE2, Sftpc-N-hACE2, and K18-hACE2 KI mice received 10 mg/kg of 7B3 at 1 dpi or 300 mg/kg NMV via oral gavage from 0–3 dpi. SA-N-hACE2 and Sftpc-N-hACE2 mice were challenged with 1 × 106 TCID50 of SARS-CoV-2 ΔN/GFP-HiBiT, whereas K18-hACE2 KI mice were infected with 3.57 × 102 TCID50 of SARS-CoV-2 WT at 0 dpi. At 4 dpi, all mice were sacrificed, and lung samples were collected. The expression of the viral E gene in lung tissues from mock-, NMV- and 7B3-treated SA-N-hACE2 (b), Sftpc-N-hACE2 (c) and K18-hACE2 KI mice (d) was quantified via qRT‒PCR (n = 4 per group). e–j Lung tissues from mock-, NMV-, and 7B3-treated SA-N-hACE2 (e), Sftpc-N-hACE2 (f), and K18-hACE2 KI (g) mice were analyzed using H&E staining to assess pathological changes, with histological scores documented for each group (n = 4 per group) (h–j). Data are presented as means ± SD. Statistical significance was determined using one-way ANOVA with Dunnett’s multiple comparisons test (b–d, h–j). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001

We measured lung viral loads in these mice (Fig. 6b–d). Both NMV and 7B3 markedly decreased lung viral loads relative to the control group. Lung pathology analysis revealed severe interstitial pneumonia in the control groups of all three mouse models, characterized by perivascular inflammatory cell infiltration, thickened alveolar walls, disrupted alveolar structure, and epithelial cell collapse. In contrast, mice treated with NMV or 7B3 presented milder lung lesions (Fig. 6e–j).

Our results demonstrate that SA-N-hACE2 and Sftpc-N-hACE2 mice can be utilized for SARS-CoV-2 antiviral drug screening, with pathophysiological indicators of infection closely resembling observations from live SARS-CoV-2 virus infections.