The B1.617.2 variant is more virulent than the early strain in the brains of K18-hACE2 mice.
We inoculated 7-week-old heterozygous K18-hACE2 mice with 2.5 × 104 50% tissue culture infectious dose (TCID50)/ml SARS-CoV-2, Hu-1 (early strain), or B1.617.2 (delta variant) via the intranasal route (Fig. 1A). Clinical symptoms were monitored for 8 days post-infection (dpi). K18-hACE2 mice infected with B1.617.2 exhibited more severe weight loss (> 20%) and greater lethality than those infected with the Hu-1 strain (Fig. 1B). In addition, K18-hACE2 mice infected with B1.617.2 frequently exhibited inflammation in the entire eye, whereas inflammation was only observed in the corners of the eyes in the Hu-1 infection group (Sup. Figure 1A). The autopsy revealed damage in several organs including severe hemorrhage in the brain, lungs, and spleen (Sup. Figure 1B–D), but no distinct damage was detected in the kidneys (Sup. Figure 1E) and other organs. These clinical lesions of the brain and lungs differed between mice inoculated with B1.617.2 and Hu-1; specifically, severe brain and mild lung damage were detected in B1.617.2-infected mice, whereas severe lung and moderate brain damage was detected in Hu-1–infected mice. We determined the viral burden in lung and brain homogenates at 3–6 dpi. Lower levels of viral RNA (Fig. 1C), subgenomic RNA (Fig. 1D), and infectious SARS-CoV-2 (Fig. 1E) were detected in B1.617.2-infected lung homogenates, whereas higher levels were present in brain homogenates. At 6 dpi, viral RNA levels in other tissues, including the heart, kidneys, and spleen, were similar in B1.617.2- and Hu-1–infected lungs. By contrast, no viral RNA was detected in the liver and trachea at 6 dpi (Sup. Figure 2). The expression of hACE2, a receptor for SARS-CoV-2, was constant, supporting SARS-CoV-2 infection in the brain and other tissues (Sup. Figure 3). hACE2 expression was lower in the lungs during SARS-CoV-2 infection than in the brain (Sup. Figure 4). Similarly, expression of the nucleocapsid (N) protein declined in both Hu-1– and B1.617.2-infected lungs throughout the course of infection. By contrast, viral N protein was detected earlier in the B1.617.2-infected brain at 4 dpi (Fig. 1F). These data demonstrated that the B1.617.2 variant has high susceptibility to the lungs and brain. In addition, the replication efficiency of B1.617.2 was sustained in the brain compared to that of Hu-1, suggesting this apparent difference contributes to continuous brain damage-induced virulence and results in high lethality.
Differences Of Neuropathological Complications In The Brains Of K18-hace2 Mice After Sars-cov-2 Infection
We assessed histopathological changes in hematoxylin and eosin-stained lung (Sup. Figure 5) and brain sections (Fig. 2) from Hu-1– or B1.617.2-infected K18-hACE2 mice. The histopathological score was calculated as the average of the representative severity of the lungs according to a microscopic scoring system (Sup. Table 1). We observed similar lung pathology in lungs infected with both variants, namely moderate pulmonary edema and infiltrating inflammatory cells in the perivascular and peribronchial regions with progressive inflammation at 3 dpi. These patterns extended the rate of lung consolidation, infiltrating inflammatory cells, and partial loss of bronchiole epithelial cilia in the Hu-1–infected lung sections at 4 dpi. Hu-1– and B1.617.2-infected K18-hACE2 mice displayed consolidation in 35–50% of the lungs and blood leakage from vessels into the adjacent alveolar space with alveolar wall thickening at 6 dpi (Sup. Figure 5). Furthermore, we observed the brain pathology in both SARS-CoV-2-infected brain tissue comprising the meninges and cerebrum. At 4 dpi, B1.617.2-infected brain sections displayed distinct increases in glial and multinucleated cell counts in adjacent meningeal vessels and the perivascular region of cerebrum, as well as partial detachment of the meninges, but these changes were not present in Hu-1–infected brain sections. B1.617.2 infection continuously increased infiltrating glial cell numbers, resulting in disruption of the meninges and severe brain damage at 6 dpi (Fig. 2A). Conversely, the predominant histopathological changes of Hu-1–infected brain sections were generally weak inflammatory responses that developed as late-onset symptoms compared to the findings in B1.617.2-infected brain sections. To determine whether brain damage was correlated with the severity of SARS-CoV-2 infection, we also stained the brain sections for the N protein of SARS-CoV-2 using immunohistochemistry. At 4 dpi, N protein was distributed predominantly in perivascular neuronal cells in B1.617.2-infected cerebrum sections, but it was not detected in the Hu-1–infected cerebrum. At 5 and 6 dpi, SARS-CoV-2 infection of neuronal cells widely spread throughout the cerebrum (Fig. 2B). Furthermore, this pattern was also observed in cerebrum sections stained with GFAP, a marker of astrocytes, and the RNA and protein levels of glial fibrillary acidic protein (GFAP) were significantly increased at 4 dpi (Fig. 2C). Together, these findings provide evidence of brain damage-induced by SARS-CoV-2 infection, and neurons were more susceptible to the B1.617.2 variant than to the Hu-1 strain.
The Early Cellular Responses To Sars-cov-2 Infection
To access the transcriptional changes induced by SARS-CoV-2 infection, we performed next-sequencing generation (NGS) of SARS-CoV-2–infected brain homogenates at 0 (control), 3, and 4 dpi. The venn diagram presents the upregulated and downregulated genes in SARS-CoV-2–infected brain homogenates at 3 and 4 dpi compared to the findings in control animals (Fig. 3A). In Hu-1–infected brain homogenates, enrichment of gene signatures revealed increases in the number of upregulated (from 47 to 305 genes) and downregulated genes (from 37 to 68 genes), whereas only 49 genes overlapped at 3 and 4 dpi. Conversely, upregulated (from 62 to 586 genes), downregulated (from 31 to 52 genes), and overlapped (18 genes) gene signatures were increased in B1.617.2-infected homogenates. Gene Ontology analysis of the top upregulated genes identified various cellular responses such as immune system processes, innate immune responses, inflammation responses, defense responses to the virus, and responses to interferon (IFN, Fig. 3B). These findings highlighted the regulation of gene sets involved in type I IFN signaling, inflammatory cytokine signaling, and glial cell migration. An hiPSC-derived neuronal organoid study reported dysregulated inflammatory responses and innate immune responses coupled with the regulation of the cell death process24. At 4 dpi, we revealed that inflammatory cytokine-associated genes (Ccl5, Ccl7, Cxcl1, Il1b, and Csf3), Type I IFN-associated genes (Irf7, Stat1, and Oas2), and certain IFN-stimulated genes (Ifit1, Oas2, Mx2, and Irf7) were upregulated in B1.617.2-infected tissues compared to their expression in Hu-1–infected tissues (Fig. 3C). In addition, the cell death process was activated earlier in B1.617.2-infected brains than in Hu-1–infected brains. The expression of genes related to apoptotic (Casp8, Casp7, and Nod1) and necrotic (Tnf, Ngfr, Ripk1, Ripk3, and Pygl) processes were upregulated by SARS-CoV-2 infection. An in vitro study reported that SARS-CoV-2 limits autophagy signaling and inhibits autophagic flux25. Likewise, our results illustrated that the expression of autophagy-associated genes (Becn1 and Atg7) was not changed or decreased, indicating that an autophagy-independent cell death program was activated in the SARS-CoV-2–infected brain (Fig. 3C). These distinct transcriptional changes were determined as temporary occurrences that indicated immune-associated or cellular-regulated signatures in the early stage of SARS-CoV-2 infection in the brain.
The Necrotic Pathway Is Activated During The Early Brain Response To Sars-cov-2 Infection
We investigated the commitment of the necrotic process in the early brain response after SARS-CoV-2 infection. At 3 and 4 dpi, brain homogenates infected with SARS-CoV-2 were assessed for the expression of necrosis-related genes using the mouse necrosis RT2-profiler PCR array. Unsupervised clustering analysis illustrated that most necrosis-related genes were expressed at higher levels in B1.617.2-infected brains than in Hu-1–infected brains at 4 dpi (Fig. 4A, left). Necrosis-related genes were upregulated (Hu-1: 4 and 4; B1.617.2: 4 and 62) and downregulated (Hu-1: 1 and 2; B1.617.2: 1 and 6) after 3 and 4 days of infection, respectively. The indicated genes included necrotic markers such as receptor-interacting kinase 1 (Ripk1), Ripk3, glycogen phosphorylase L (Pygl), Poly (ADP-ribose) polymerase 1 (Parp1), and calpain-1 catalytic subunit (Capn1) (Fig. 4A, right). In addition, we identified a significant difference in Ripk3 expression between the groups with SARS-CoV-2 infection. However, Bcl2 (pro-apoptotic), Casp3 (apoptotic), and Becn1 (autophagy) expression did not differ between the groups (Fig. 4B). Taken together, these data indicated that SARS-CoV-2–infected neuronal cells can promote a Ripk3-dependent cell death program in the early brain response.