Intranasal infection by SARS-CoV-2 Omicron variants can induce inflammatory brain damage in newly weaned hamsters

ABSTRACT Children infected by SARS-CoV-2 Omicron variant may develop neurological complications. To study the pathogenesis in the growing brain, we intranasally challenged newly-weaned or mature hamsters with SARS-CoV-2 Omicron BA.2, BA.5, or Delta variant. Omicron BA.2 and Delta infection produced a significantly lower viral load in the lung tissues of newly-weaned than mature hamsters despite comparable histopathological damages. Newly-weaned hamsters had higher brain viral load, significantly increased cerebrospinal fluid concentration of TNF-α and CXCL10 and inflammatory damages including mild meningitis and parenchymal vascular congestion, despite sparse expression of nucleocapsid antigen in brain cells. Furthermore, 63.6% (28/44) of all SARS-CoV-2 infected newly-weaned hamsters showed microgliosis in olfactory bulb (OB), cerebral cortex, and hippocampus. In infected mature hamsters, microgliosis was observed mainly in OB and olfactory cortex of 35.3% (12/34) of their brains. Neuronal degeneration was found in 75% (33/44) of newly-weaned hamsters affecting multiple regions including OB, olfactory cortex, midbrain cortex, and hippocampus, while such changes were mainly observed in the hippocampus of mature hamsters. Importantly, similar brain histopathology was also observed in Omicron BA.5-infected newly-weaned hamsters. Our study suggested that SARS-CoV-2 may affect the brain at a young age. This kind of brain involvement and histological changes are not virus variant or subvariant specific. Incidentally, a moderate amount of eosinophilic infiltration was observed in the mucosa of nasal turbinate and trachea of newly-weaned hamsters infected by Omicron BA.2 and BA.5 but not Delta variant. This histological finding is consistent with the higher incidence of laryngotracheobronchitis in young children infected by the Omicron variant. Summary Intranasal infection of newly-weaned Syrian hamsters by SARS-CoV-2 Omicron variants can lead to brain inflammation and neuron degeneration with detectable low level of viral load and sparse expression of viral nucleoprotein.


Introduction
The coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has affected over 643 million people with about 6.6 million deaths as of 26 December 2022 [1,2]. Over 40 million children under age 14 were infected with 6000 deaths, among which more than 50% were under the age of 5 [3]. Though COVID-19 is primarily a respiratory infection, neurological complications are commonly reported [4]. Up to 80% of COVID-19 patients had neurological symptoms at some point after SARS-CoV-2 infection [5].
The most common neurological manifestations at the onset of the disease are anosmia (loss of smell) due to the destruction of olfactory epithelium and sensory neurons by SARS-CoV-2 [6,7]. Other common neurological symptoms during the disease course include dizziness, headaches, myalgia, hyposmia, impaired consciousness, seizures, and ataxia [8]. Radiological and pathological evidence of brain involvement and the virologic findings in cerebrospinal fluid (CSF) suggested the possibility of direct virus invasion of the central nervous system [9]. Recent studies indicated that nearly onethird of adult COVID-19 patients, even those with mild respiratory symptoms, experienced long-term neurological symptoms which are now considered as long COVID [10,11]. They experienced malaise, debilitating fatigue, and impaired memory or executive function long after their recovery [12,13]. MRI imaging of their brains showed significantly decreased brain size or various degree of brain tissue damage, which are associated with cognitive decline [14]. Moreover, population immunity from natural infection and vaccination was associated with the emergence of immune evasive SARS-CoV-2 variants of which the Omicron variant dominated and replaced other variants by escaping from neutralizing antibodies induced by prototype virus infection or vaccination [15,16]. Since the emergence of Omicron variant, more young children have been infected with mostly mild illness similar to the disease caused by the prototype or previous variants [17,18]. However, among those hospitalized children with COVID-19, 15.0% of them were admitted for neurological complications, this percentage is significantly higher when compared with the 3.8% reported for other SARS-CoV-2 variants during the pre-vaccination period in the United Kingdom [19,20]. One study comparing Delta and Omicron infections in children reported that while Omicron caused relatively milder respiratory manifestation [18], their neurological and psychiatric complications were not reduced [10]. The neurological manifestations reported in children include seizures, encephalitis, and acute necrotizing encephalopathy [20,21]. During the Omicron surge in Hong Kong, neurological manifestations and their severity were significantly increased in children with some fatal outcomes [21]. However, the pathogenesis of Omicron variant in the respiratory and central nervous system in young children is still uncertain. To understand the impact of Omicron infection on the growing brain, we utilized the newly-weaned hamster and studied the histopathological changes after intranasal inoculation of currently circulating variants, including Omicron BA.2, BA.5, and the previous Delta variant, and compared with 8-weeks-old mature hamster. We found that all three SARS-CoV-2 variants caused more frequent inflammatory and degenerative damages in the brain of newlyweaned hamsters than mature hamsters.

Animal model
Syrian hamsters, newly-weaned at 3 weeks of age or mature at 8 weeks of age, were obtained from the Centre for Comparative Medicine Research, HKU. All the animal experimental procedures were approved by the Committee on the Use of Live Animals in Teaching and Research of HKU (CULATR #6093-22).

Intranasal inoculation of virus to hamster
Hamsters under anaesthesia of intraperitoneal injection of ketamine (150 mg/kg) and xylazine (10 mg/ kg) were inoculated intranasally with 10 5 PFU SARS-CoV-2 virus in 50 μl volume per animal as we previously reported [23,24]. Mock-infection controls were given the same volume of phosphatebuffered saline (PBS). At 2, 4, or 7 days post-infection (dpi), the hamsters were euthanized by intraperitoneal injection of pentobarbital sodium (200 mg/kg). CSF, blood, brain, lung, trachea, and nasal turbinate (NT) were harvested for analyses. CSF was collected carefully using 0.3 ml syringe through puncture at the cisterna magna to avoid any blood contamination. Brain was collected after removing calvaria.
Determination of viral load by real-time RT-PCR (RT-qPCR) and plaque assay Total RNA was extracted from homogenized brain, lung, and NT. SARS-CoV-2 RNA-dependent RNA polymerase (RdRp) and subgenomic E (Sg E) gene copies were determined by RT-qPCR [25]. The infectious virus in the tissue homogenate was determined by 50% Tissue Culture Infectious Dose (TCID 50 ) in Vero E6-TMPRSS2 cells.

Detection of SARS-CoV-2 spike RNA by RNAScope
In situ RNA hybridization to detect virus spike gene in paraffin-embedded brain tissue sections was performed using RNAScope® Probe-V-nCoV2019-S (848561, ACD) according to the manufacturer's instructions. The sections were examined using the Olympus BX53 microscope.
Histopathology, immunohistochemistry, immunofluorescence staining of tissue sections, and automated imaging analysis Formalin-fixed, paraffin-embedded brain, lung, and NT tissues were processed into 4 µm tissue sections and stained with haematoxylin and eosin (H&E) for histopathological examination as previously described [26,27]. Viral antigens and cellular surface markers were detected by immunohistochemistry or immunofluorescence with specific antibodies: home-made rabbit/mouse anti-SARS-CoV-2 nucleocapsid (N) protein antibody [7,28], anti-neuron-specific type β-III tubulin (801202, Biolegend), anti-Ionized calcium-binding adapter molecule 1 (Iba1) antibody (ab178846, abcam). Brain cell DNA fragmentation was labelled using Click-iT® Plus TUNEL assay kit (Invitrogen) according to the manufacturer's instructions as we described previously [29]. All tissue sections were examined under light or fluorescence microscopy in a blinded fashion. Images were captured with Olympus BX53 fluorescence microscope with DP80 camera and proprietary Cellsens dimension software (Olympus LS, Japan). For quantification of Iba1-positive microglia, immunofluorescence-stained brain sections were examined and 10-20 images were randomly captured in the area of olfactory bulb (OB), cerebral cortex, and hippocampus, respectively. A magnification of 400× image was applied for a correlated resolution of 0.25 μm/pixel with both DAPI and mCherry channels. Exported TIFF format files were analysed with QuPath (Version 0.3.2).
The histological changes in the brain cortex at 4 and 7 dpi after virus infection were examined systematically, recorded, and scored for comparison of the severity using criteria adopted from previous reports with modifications [30]. Briefly, scoring for meningitis: meningeal vessel congestion and increased vascular space cellularity present (1) or absent (0); brain cortex blood vessel congestion, perivascular immune cell infiltration, microgliosis and neuronal degeneration, neuronal necrosis occurred in any anatomical region: absent (0); minimal if found in one anatomical region of the brain section (1); mild if found in more than one region (2); moderate changes occurred in more than one anatomical region (3). There were no cases that can be defined as severe in this study. Lung histological changes were accessed and scored as described previously [28].
RT-qPCR determination of pro-inflammatory cytokine and chemokine mRNA expressions Total RNA was extracted from tissues with MiniBEST Universal RNA Extraction Kit (Takara) and reverse transcribed into cDNA with PrimeScript TM RT reagent kit (Takara). RT-qPCR was performed using gene-specific primers (Supplementary Tables 1 and  2) with SYBR Premix Ex Taq II Kit (Takara). The expression of cytokine and chemokine genes was analysed by ΔΔC t method, using mock-infected tissue samples as baseline.

Statistical analysis
All data were analysed with Prism 8.0 (GraphPad Software Inc). Statistical comparison among different groups was performed using Fisher's exact test, oneway ANOVA, or two-way ANOVA. p < 0.05 was considered statistically significant and highlighted in all figures.

Results
Omicron BA.2 and Delta replicate at lower levels in newly-weaned hamsters but cause severe respiratory tissue damages SARS-CoV-2 Omicron variants were reported to have attenuated pathogenesis in respiratory tissues of human and in animal models when compared to previous strains, such as Delta variant [31]. We challenged newly-weaned and mature hamsters with Omicron BA.2 or Delta variant and examined the NT and lung at 2, 4, and 7 dpi. In general, Omicron BA.2 replicated less effectively than Delta variant, which is consistent with previous reports [28].
Comparison between these two age groups challenged by SARS-CoV-2 showed that newly-weaned hamsters had more than one log lower viral load and infectious titre in both Omicron BA.2 and Delta infection with NT and lung at 4 dpi ( Figure 1(A,B)), which indicate that both viruses replicate less effectively in newly-weaned hamsters than mature hamsters.
RT-qPCR profiling for cytokine/chemokine gene expression showed significantly higher basal mRNA expression levels of IFN-α, IFN-γ, and proinflammatory cytokine/chemokine IL-6, TNF-α, CXCL10, CCL2, and CCL3 in the lung of newlyweaned hamsters than mature hamsters before virus challenge (Figure 1(C)). A rapid upregulation of IL-6, TNF-α, IL-1β, IFN-γ, CXCL10, CCL2, and CCL3 was observed at 2 dpi in newly-weaned hamster lung tissues, which were significantly higher than their age-matched mock controls. The cytokines decreased back to basal level at 4 and 7 dpi, except for IL-1β and CXCL10. However, no IFN-α upregulation was detectable after Omicron BA.2 or Delta infection in newly-weaned hamsters ( Figure  1(D)). On the contrary, mature hamsters had a significant increase in IFN-α expression together with most of the cytokine/chemokines from 2 to 7 dpi (Figure 1(D)). Serum concentration of IL-6, TNFα, and CXCL10 were significantly higher in newlyweaned hamsters challenged by Omicron BA.2 or Delta virus than their age-matched mock controls ( Figure 1(E)), indicating that systemic inflammatory responses were elicited by virus infection in newlyweaned hamster despite lower virus replication in respiratory tissues.
Histologically, Omicron BA.2-infected lung damages progressed from interstitial inflammatory infiltration at 2 dpi to diffuse alveolitis with severe alveolar space infiltration and exudation at 4 dpi same, which was similar to Delta infected lungs ( Figure 1(F)), but BA.2 infected hamster lungs had significant lower histological score ( Figure 1(H)). Importantly, some unique histopathological features were observed in Omicron BA.2 infected newlyweaned hamster, including severe submucosal oedema with eosinophil infiltration in nasal and tracheal mucosa, and severe perivascular oedema and with eosinophil infiltration in the lung tissue space around the large vessel and bronchi (Figure 1(G)). These histopathological changes were different from the dominant mononuclear cell infiltration induced by Delta variant and prototype SARS-CoV-2 infection [23].

Brain involvement by Omicron BA.2 and Delta variants in newly-weaned hamsters after intranasal inoculation
SARS-CoV-2 subgenomic E gene was detected by RT-qPCR assays in the brain samples taken at 2, 4, and 7 dpi, altogether, 6/24 (25%) Omicron BA.2 and 11/ 21 (52.4%) Delta variant-infected newly-weaned hamster brains showed positive Sg-E, respectively ( Figure 2(A)). Live virus was detected by TCID 50 assay from 5/22 (22.7%) and 2/22 (9.1%) newlyweaned hamster brains after Omicron BA.2 and Delta variant infection, respectively (Figure 2(B)). In mature hamsters, subgenomic E gene were detected at lower frequency in Delta infection (3/10) when compared with newly-weaned hamster (Figure 2(A)); 2/11 Delta virus-infected mature hamsters had live infectious virus titre, while no live virus was detectable from Omicron BA.2-infected mature hamsters ( Figure  2(B)). Brain tissues with positive viral RNA stained by SARS-CoV-2 spike RNA hybridization kit further confirmed the presence of virus gene inside the infected cells ( Figure S1). Immunohistochemistry and immunofluorescence staining both showed sporadic viral N antigen-positive cells in the OB and cerebral cortex ( Figure 2(C,D)). No signals were found in mock control brains. Double immunofluorescence staining for microglia cellular marker Iba1 and SARS-CoV-2 N antigen showed double-positive cells in the OB and cerebral cortex which indicated that microglial cells were infected (Figure 2(E)). N protein-expressing cells with positive neuronal marker Tuj1 were only rarely found (Figure 2(F)). In mature hamster brains, viral N-positive cells were also occasionally seen after Omicron BA.2 or Delta virus infection (Figure 2(G)). These findings indicated that intranasal challenge with SARS-CoV-2 Omicron BA.2 and Delta variants can infect the brain of hamsters. Live viruses were more frequently found in the brains of newly-weaned hamster infected by Omicron BA.2 than mature hamster.
Pro-inflammatory cytokine/chemokine upregulation in the brain of newly-weaned hamster after intranasal inoculation of SARS-CoV-2 Omicron BA. 2

and Delta variant
Unlike the lung, the basal expression level of innate cytokines/chemokines in uninfected newly-weaned hamster brains was significantly lower than those in mature hamster, except for a similar level of IL-1β ( Figure S2). At 4 dpi after Omicron BA.2 infection, newly-weaned hamster showed around 10-fold increase of IFN-α, IL-6, and TNF-α, and 4-fold increase of IFN-γ and CXCL10 in brain. Then all except IFN-γ decreased to the basal level at 7 dpi (Figure 3(A)). In the brains of mature hamsters, IL-1β was upregulated from 2 dpi through 7 dpi, IFN-α, IL-6, TNF-α, and CXCL10 were only upregulated at 7 dpi (Figure 3(A)). C-C motif chemokine 11 (CCL11) has been reported to be particularly associated with SARS-CoV-2-related brain damages by limiting neurogenesis [32]. As expected, we found that CCL11 together with CCL2, CCL3, and the receptors CCR2, CCR3, and CCR5 were all upregulated in newly-weaned hamster brain tissues after Delta variant infection at 2, 4, and 7 dpi, of which CCL11 and CCR2 were significantly higher at 4 dpi ( Figure 3 (B)). In Omicron BA.2 infected newly-weaned hamster brains, all these chemokines and receptors were upregulated at 4 and 7 dpi while not reaching statistically significant differences comparing with age-matched mock controls. In mature hamsters, only CCR5 was slightly upregulated at 2 and 4 dpi after Omicron BA.2 or Delta infections. Though most of these genes were upregulated at 7 dpi after Delta infection (Figure 3(B)), we did not see CCL11 upregulation in mature hamster brains after either Omicron BA.2 or Delta infection. In the CSF of newly-weaned hamsters, TNF-α concentrations were increased at 4 dpi after Omicron BA.2 infection (Figure 3(C)). Our findings suggested that higher pro-inflammatory cytokine/chemokine response occurred in the brains of newly-weaned hamsters after Omicron or Delta variant infection.
Histopathological changes in brain tissues of the newly-weaned hamsters after intranasal challenge by Omicron BA.2 or Delta variant No gross pathology such as cerebral oedema or haemorrhage was observed in any of the hamsters at 2, 4, or 7 dpi after infection. Paraffin-embedded brain tissues were cut into sagittal sections and stained by H&E for histological examination. Anatomical regions including OB, cerebral cortex, hippocampus, midbrain, thalamus, cerebellum, pons, and medulla were assessed ( Figure S3). Firstly, meningeal congestion and enlarged vascular space with some inflammatory cell infiltration suggestive of mild meningitis were frequently observed in newly-weaned hamsters at 4 dpi Data represented mean ± SD. *p < 0.05, ***p < 0.001, ****p < 0.0001, by Two-way ANOVA. by Omicron BA.2 or Delta virus challenge, but less frequently at 7 dpi (Figure 4(A)). In the brain parenchyma, congestion of the small blood vessels and capillaries with enlarged vascular space and perivascular inflammatory cells infiltration were frequently observed and small foci of haemorrhage were occasionally seen in newly-weaned hamsters challenged by Omicron BA.2 or Delta virus at 4 and 7 dpi (Figure 4(B)). Furthermore, increased microglia were observed which mainly affected OB, olfactory cortex, and midbrain cortex in most of the newlyweaned hamsters (Figure 4(C)). The brain of infected mature hamsters showed increased microglia while the meningeal and brain parenchymal vascular inflammatory histological changes were mostly absent. Comparison of the frequency and severity of the histopathological changes by semiquantitative scoring showed that Omicron BA.2 and Delta-infected newly-weaned hamsters had significantly higher scores of meningitis, parenchymal gliosis, and vascular congestion compared to mock controls (Figure 4(D), Supplementary Table 3). Furthermore, immunofluorescence staining of microglial cellular marker Iba1 and image quantification showed a significant increase of Iba1-positive cell in the newly-weaned hamster brains at 7 dpi (Figure 4(E,F)). Though infected mature hamsters also showed an increase in Iba1-positive microglial cells compared to their age-matched controls (Figure 4(F)), determination of the status of microglial cell activation by real-time RT-qPCR only showed significant upregulation of Lcn2 (Lipoclin-2), Steap4 (six transmembrane epithelial antigen of prostate-4), Timp1(TIMP metallopeptidase inhibitor 1), Osmr (Oncostatin M Receptor), Thbs2 (Thrombospondin 2), and Apoe (Apolipoprotein E) in brain tissues of newly-weaned hamster at 4 and 7 dpi, but not in mature hamsters (Figure 4(G)). These findings strongly indicated that a higher level of brain microglial cell activation was induced by SARS-CoV-2 infection of newly-weaned hamsters than mature hamster.

Neuronal degeneration, apoptosis, and necrosis in brain tissues of newly-weaned hamster after Omicron BA.2 or Delta infection
Neuronal degenerative changes in response to inflammation were common and could result in chronic neurological complications. These changes were frequently identified in H&E stained brain sections from multiple locations of newly-weaned hamsters at 4 and 7 dpi. The degenerating neurons were characterized by hypereosinophilic and ballooning cytoplasm and shrunken nucleus with nuclear dissolution ( Figure 5(A)). In the hippocampus regions, the pyramidal neurons in the Cornu Ammonis (CA) regions had eccentric shape with shrunken nuclei clustered in the CA1 to CA 3 region. Similar degenerative changes were also found in the Dentate Gyrus region (DG) (Figure 5(B)). In the OB, midbrain cortex, and occasionally cerebellum, degenerative neurons were also observed, with occasional necrotic neurons ( Figure 6). TUNEL-positive cells were more abundant in the OB but also occasionally found in the cerebral cortex, hippocampus, and cerebellum cortex and in the ependymal epithelium of ventricles (Figure 7). In the brain of mature hamsters, the degenerating neurons were found mainly confined in the hippocampus across CA1-CA3 regions at 7 dpi ( Figure S4, Supplementary Table 3). TUNEL-positive cells were rarely found in the brain of mock controls.
SARS-CoV-2 Omicron BA.5 variant causes similar histopathological changes as BA.2 in the lung and brain of newly-weaned hamsters SARS-CoV-2 Omicron subvariants continue to emerge. To investigate whether they may cause different pathology in newly-weaned hamsters, we compared their virological and histological features after intranasal challenge. The same intranasal dose of BA.5 or BA.2 caused a similar degree of nasal epithelial infection and tissue destruction (Figure 8(A)), and severe submucosal oedema with eosinophil infiltration in trachea (Figure 8(B)). In the newlyweaned hamster lungs, BA.5 infection had a significantly higher viral load than BA.2 at 4 dpi (Figure 8 (C)). Lung tissue sections showed extensive viral N protein expression in alveolar epithelium, diffuse alveolar inflammatory cell infiltration, and proteinrich fluid exudation which also filled the bronchial lumen (Figure 8(D)). Significantly increased serum concentrations of TNF-α and CXCL10 in newlyweaned hamsters indicated severe systemic inflammation (Figure 8(E)). In the brain of newly-weaned hamsters, BA.5 infection also showed a higher viral load than BA.2 infection (Figure 8(C)) with occasionally observed viral N antigen (Figure 8(F)). Increased CSF concentrations of TNF-α (Figure 8(G)), parenchymal microgliosis, and neuronal degeneration or necrosis were frequently observed (Figure 8(H)).

Discussion
The Syrian hamster SARS-CoV-2 infection model has been widely used for the studying the pathogenesis, antiviral therapy, and vaccination for SARS-CoV-2 infection [23,29,[33][34][35][36]. To simulate SARS-CoV-2 infection at young age, we performed intranasal challenge of newly-weaned hamster with SARS-CoV-2 Omicron subvariants BA.2 and BA.5, and Delta variant. We demonstrated that they commonly induced brain parenchymal inflammation and neuronal degeneration in newly-weaned hamsters compared to mature hamsters.     In lung tissues, Omicron BA.2 replicated less effectively and produced a significantly lower viral load in newly-weaned hamsters than mature hamsters. However, the histopathological damages in the lungs were similar, which is manifested as diffuse alveolitis at 4 dpi. In paediatric COVID-19 patients, lower viral load in respiratory specimens than that of adult patients was attributed partly to the higher basal expression and activity of interferon in the respiratory epithelium of children [37], and to some degree of cross-protection from natural infection by other human coronaviruses [38]. Consistent with findings in children, the basal expression levels of interferon-α and many pro-inflammatory cytokine/chemokines were significantly higher in lungs of newly-weaned hamsters than mature hamsters before infection. However, we detected no further induction of IFN-α after virus infection despite a surge of pro-inflammatory responses (IL-6, IL-1β, TNF-α, IFN-γ CCL2, CCL3, and CXCL10) compared to mature hamsters. This may suggest that the first-line antiviral response by interferon could not be further activated in the lungs of newly-weaned hamsters.
Other host factors such as the amount of receptor expression in the lungs may also contribute to the lower viral load in young hamsters [39].
Despite the lower viral load in their lung tissue, infected newly-weaned hamsters commonly developed brain inflammatory damages regardless of the SARS-CoV-2 variants used for challenge. At 4 dpi, proinflammatory cytokines/chemokines including IFN-α, were substantially upregulated in the brain tissue of newly-weaned hamsters, accompanied by histological changes of meningeal inflammation, vascular congestion in brain parenchyma, increased perivascular space and cellularity, and small foci of haemorrhage.
In mature hamster brain, only IL-1β was mildly upregulated at 2 and 4 dpi after Omicron BA.2 and Delta virus challenge. Pro-inflammatory response by gene expression of IFN-α, IL-6, and TNF-α responses were not detected until 7 dpi after Delta infection but not after Omicron infection. Histological changes of inflammation in brain tissue were limited to vascular congestion with enlarged vascular space in brain parenchyma without inflammatory cell infiltration or meningitis.
While the brain is an immune-privilege site, inflammation in response to a variety of endogenous and exogenous insults would lead to neuronal injury initiated by microglia, which are the resident macrophages of the central nervous system. Proliferation, accumulation, and activation of microglia are common features of brain parenchymal inflammation, which could be both beneficial and detrimental. In brain tissues of SARS-CoV-2 infected newly-weaned hamster, we observed increased number of microglial cells widely distributed in OB, cerebral cortex, and hippocampus. In some cases, the microglial cells clustered together to form microgliosis [40,41]. Several genes (CD68, Lcn2, Apoe, Steap4, Timp1, Osmr, and Thbs2) related to the activation status of microglial cells [42,43] were significantly upregulated as demonstrated by RT-qPCR assay. The most prominent upregulated cytokine/chemokine genes including IL-6, TNF-α, CCL2, and CCL3 in the brain tissue at 4 dpi could also be attributed to microglial activation [44,45]. Activation of microglial cells leads to further release of inflammatory mediators, increased blood-brain barrier permeability, and recruitment of systemic inflammatory cells into the brain, which could perpetuate inflammatory damages [45]. Consistent with these findings, TNF-α concentration in the CSF increased significantly in newlyweaned hamster infected by Omicron or Delta virus. In mature hamster, microgliosis was also observed in a small percentage of brain samples and mostly localized in OB and olfactory cortex without upregulation of those genes of microglial activation. These findings suggested that the olfactory mucosa may be one of the sites of virus entry and also a better immune regulatory control of brain microglial responses in the mature than newlyweaned hamsters [46]. Degenerating neurons were distributed in multiple anatomical regions including OB, olfactory cortex, cerebral cortex, hippocampus, and cerebellum which suggested different types of neurons were affected. Consistent with the minimal pro-inflammatory cytokine responses in brains of mature hamsters, the histological changes of inflammation and degeneration were confined to the OB, but in some animals also in the olfactory cortex and hippocampus.
Up to date, evidence for direct SARS-CoV-2 invasion into the brain has been controversial. This uncertainty is partly attributed to the different sensitivity and specificity of virus detection methods. To better understand the scope of SARS-CoV-2 virus infection of brain cells at young age, we studied the hamsters infected by different virus strains using RT-qPCR for viral load, innate immune gene expression profile, in situ hybridization, immunohistochemistry, and histological examination. Our RT-qPCR for subgenomic E gene confirmed the presence of replicating virus in the brain albeit at low level. RNAScope using spike gene-specific probe demonstrated viral spike RNA in brain cells of newly-weaned hamsters. Immunohistochemistry and immunofluorescence staining for viral N protein further confirmed the infection of microglia and a few neurons. Most importantly, live viruses were detected from 15.9% (7/44) of infected newly-weaned hamster brains, and 9.1% (2/22) of infected mature hamster brains. Taken together, we conclude that intranasal virus inoculation of SARS-CoV-2 Omicron and Delta variant can infect and replicate in brain tissue of newly-weaned hamsters at limited scale, but more frequently than that of mature hamsters. Consistent with our findings, previous in vitro study of brain organoids reported that SARS-CoV-2 virus can infect human microglia and neuron [47,48]. While virus entry through the olfactory mucosa to the brain is one possible route, the immature bloodbrain barrier of the developing brain in newly-weaned hamster which is weakened or damaged by the systemic inflammation induced by the SARS-CoV-2 pneumonia is another possible route [46]. Other studies on the mechanisms of SARS-CoV-2 associated neurological symptoms in mature C57BL/6 mice and rats reported that administration of spike protein S1 subunit to the brain or hippocampus area induced microglia activation and neuron cell death [49,50]. These reports tend to support the conclusion that direct virus infection of brain cells is not essential in inducing brain damage.
Though Omicron BA.2 and BA.5 both replicate less effectively in respiratory tissue of newly-weaned hamsters comparing to Delta virus, BA.5 replicated more effectively than BA.2. BA.2 and BA.5 caused diffuse alveolitis and more severe submucosal eosinophil infiltration and submucosal oedema in upper respiratory tract. This may account for, at least in part, the symptoms of croup in young children infected by the Omicron virus [51]. Infections by other respiratory viruses such as parainfluenza, influenza A and B, respiratory syncytial virus, and adenovirus can often cause croup in young children [52]. Since the subglottic structure is the narrowest part of a child's upper respiratory tract, infection of the upper airway causing inflammation, mucosal and submucosal oedema may lead to life threatening airway obstruction [52,53]. Taken together, SARS-CoV-2 Omicron subvariants can still cause respiratory or brain complications in young children. Vaccination should be given to provide protection against possible complications.
One limitation of our study is that the longest period of study was 7 days after virus infection. We are not able to provide information on whether these young hamsters could completely recover when they get mature, or how long the brain damages can persist. It has only been 3 years since the COVID-19 pandemic started, clinical studies showed that neurological symptom can persist for a period of 6 months after recovery from the infection [11]. Close follow-up studies are needed to understand the duration of CNS symptoms in human and animal models with or without specific treatments.

Disclosure statement
No potential conflict of interest was reported by the author(s).