COVID-19 has been associated with a wide range of clinical manifestations, including neurological and psychiatric alterations. In this study, we investigated whether MHV-3, can directly affect the CNS and serve as a suitable platform to study COVID-19-related neurobiological mechanisms and identify novel therapeutic targets. The main findings of this study are: (i) MHV-3 delivered intranasally is capable of infecting and replicating in the CNS, causing mild histopathological changes; (ii) MHV-3 infection results in increased glutamate release, intracellular calcium levels and pro-inflammatory mediators in the cortex; (iii) MHV-3 infection decreases the production of neuroprotective mediators such as BDNF and CX3CL1 in the brain; and (iv) MHV-3 infection induces anxiety-like, depressive-like behaviors, and motor dysfunction, especially in female mice.
Models that assess behavioral changes in the context of betacoronavirus infections in the CNS are still scarce. We have demonstrated the applicability of a murine betacoronavirus model as a platform to recapitulate the acute neurological and psychiatric symptoms observed in patients with COVID-19. All analyses were performed up to the 5th day post infection, i.e., at the peak of systemic disease [14]. First, the neurotropism of MHV-3 was confirmed by recovering viable virus particles in the brain of mice. In addition, histopathological analysis of the brain of MHV-3-infected mice revealed mild changes, such as dilated blood vessels in the meninges, an inflammatory infiltrate, and hyperemic vessels in the cerebral cortex. These findings support human studies showing that SARS-CoV-2 was detected in the post-mortem brain of half of COVID-19 patients along with mild neuropathological lesions [3]. Although many studies have not found SARS-CoV-2 in the brain, it is already known that the immune response triggered by the virus can culminate in a “cytokine storm”, which reaches and compromises the homeostasis of the CNS, as well as other tissues [9].
In addition, we conducted a quantitative analysis of neurons (NeuN), microglia/macrophage (IBA1), and astrocytes (S100B), as well as evaluated apoptosis (cleaved caspase 3). None of these markers were found to be altered in the cortex of infected animals when compared to mock. However, this finding does not necessarily indicate that glia and neurons were not affected. In fact, brain tissues from individuals who died with COVID-19 showed that microglia abundantly expressed the lysosomal marker CD68, which is a marker of phagocytic activity, especially in the olfactory bulb and cerebellum [74]. Boroujeni et al. (2021) also demonstrated that the inflammatory response to COVID-19 caused neuronal death in the cerebral cortex (post-mortem) of critically ill patients. NeuN quantification revealed a similar number of marker-positive neurons when we compared the infected and control groups. Nonetheless, the analysis of the ultrastructure of cerebral cortex samples from animals infected with MHV-3 revealed the presence of darkened neurons, known as Dark Neurons (DNs), indicating ongoing degeneration. Golgi complexes and dilated endoplasmic reticulum and mitochondria with ruptured cristae were also observed in these cases. Although the NeuN quantification showed no change, about 90–99% of DNs recover with time, while a small number of neurons die [76]. Electron microscopy studies have shown that these DNs have a low functional activity. However, the origin and mechanism of the emergence of these dark neurons remain undefined [76].
In the present study, we also measured the levels of the excitatory neurotransmitter glutamate and calcium (Ca2+) released at presynaptic terminals in the cerebral cortex. Our results showed that MHV-3 infection led to massive glutamate release and increased intracellular Ca2+ levels. Other viruses such as HIV, ZIKV, and H1N1 have been reported to impair glutamatergic transmission, which can result in excitotoxicity and impaired cell signaling [19, 51, 82, 83]. Additionally, HCoV-OC43, a human coronavirus, can infect human CNS neuronal and glial cells and activate neuroinflammatory mechanisms [52]. Infection of mice with HCoV-OC43 has been shown to induce neuronal dysfunction and decrease the expression of the glutamate transporter GLT-1 in astrocytes, potentially leading to increased central glutamate levels. We also observed an increase in IL-6 levels in the PFC of animals after MHV-3 infection. IL-6 can be produced within the CNS by various cell types and infiltrating inflammatory cells under neuroinflammatory conditions [54]. It has been reported that membrane depolarization is one of the primary mechanisms for the upregulation of IL-6 in neurons, which is dependent on glutamatergic activation of N-methyl-D-aspartate receptors (NMDA-R), an increase in Ca2+ levels, and activation of Ca2+/dependent kinases, such as calmodulin [55]. In a ZIKV infection model, we demonstrated that neuronal cell death could be prevented when infected animals were treated with a non-competitive inhibitor of NMDA-R. Neurodegeneration and microgliosis induced by the infection were also inhibited in vitro and in vivo [9, 83].
It is interesting to note that increased levels of IL-6 have been associated with anxiety and depressive symptoms [56, 57], and plays a role in neurogenesis [58]. Additionally, mounting evidence suggests that inflammation and alterations in glutamate neurotransmission may contribute to the pathophysiology of mood disorders [59]. There is evidence indicating that fractalkine has protective effects against glutamate-mediated excitotoxicity, once fractalkine increases BDNF levels and TrkB phosphorylation [60]. In our study, we observed a decrease in cortical levels of fractalkine and BDNF, which leads us to speculate that neuroprotection mechanisms are reduced, contributing to depressive-like behaviors in mice. While IFN-γ contributes to virus clearance [61], we observed a decrease in IFN-γ levels after MHV-3 infection. Interestingly, at low physiological glutamate concentrations, glutamate can directly activate naïve T cells via AMPA-R. However, when glutamate concentration markedly increases, this neurotransmitter usually inhibits T cell function [62].
In order to evaluate the behavioral consequences of MHV-3 infection, we conducted several specific tests. In the elevated plus maze test, we observed anxious-like behavior in the infected animals. This finding is consistent with a previous study on DENV-3 encephalitis, which showed that infected mice also displayed anxiety-like behavior and increased mRNA expression of pro-inflammatory cytokines in the hippocampus associated with neuronal loss [77]. BDNF has been shown to have an antidepressant function in the PFC [66] and hippocampus [67]. In addition, another study demonstrated that administration of BDNF into the hippocampus of rats reduced anxiety-like behavior in the elevated plus maze test [68], suggesting that the reduction in BDNF levels may be associated with the development of anxiety disorders. Moreover, impaired neuron-microglia communication, specifically the CX3CL1/CX3CR1 pathways, has been increasingly linked to the development of psychiatric conditions [69].
Studies performed in transgenic mice (K18-hACE2) that overexpress human ACE-2 and are infected with SARS-CoV-2 strongly support the acute CNS impairments observed in our MHV-3 model. Kumari and colleagues (2021) reported that intranasal inoculation of SARS-CoV-2 in K18-hACE2 mice was associated with neuroinvasion and encephalitis. These findings were corroborated by Oladunni et al. (2020), who demonstrated that intranasal infection of K18-hACE2 mice resulted in viable viral titers in various organs, including the nostril, lungs, and brain. Moreover, the infection was linked to local and systemic chemokine storm, with increased levels of CXCL-2, CXCL-10, and CCL-3 in the brain of SARS-CoV-2-infected mice. Finally, another recent study showed that primary neurons isolated from K18-hACE2 mice are susceptible to SARS-CoV-2 infection, and that the infection upregulates the expression of genes involved in innate immunity and inflammation, including IFN-α, ISG-15, CXCL-10, CCL-2, IL-6, and TNF, in the neurons [85].
According to Andrade et al. [14], there is no gender difference in the lung disease caused by MHV-3 in mice. This finding is supported by Oladunni et al. (2020) who found no significant difference in chemokine levels in the lung between male and female K18-hACE2 transgenic-infected mice, except for CXCL-10 at early time points of infection, which was significantly higher in female K18- hACE2 mice. However, our results indicate significant gender differences in the central nervous system (CNS) of MHV-3 infected mice, particularly in terms of the inflammatory response. Female mice infected with MHV-3 had increased levels of IL-6 in the analyzed brain regions (PFC, hippocampus, and striatum), while this cytokine did not change in male mice. On the other hand, infected male mice showed an increase in CX3CL1 in the PFC and BDNF in the hippocampus compared to the control group, while females infected had reduced levels of these mediators in the PFC. Previous studies have also reported the role of BDNF in the pathogenesis of depression, and its action is directly related to the brain region [65].
The Brazilian COVID Registry Project, a multicenter study that collected data from 39 Brazilian hospitals in 17 cities, observed that patients admitted with COVID-19 and clinically diagnosed with neurological syndromes had a higher incidence of septic shock, ICU admission and death compared to controls [70]. Furthermore, the study separated neurological manifestations presented at hospital admission according to incidence by sex and age group and found that women were more likely to have headaches and less likely to have seizures [71]. In MHV-3 infection in females there was a reduction of BDNF and CX3CL1, which may result in behavioral changes and loss of protection mechanisms. On the other hand, infected males show an increase in CX3CL1 and maintain BDNF levels, which may help to mitigate possible behavioral repercussions.
The MHV-3 model, like other models used to study the neurological effects of COVID-19, has several limitations. Firstly, it is not suitable for investigating the viral entry step, as SARS-CoV-2 uses a different receptor, ACE-2, while MHV-3 enters the cell through the CEACAM-1 adhesion molecule. This limits the usefulness of the model in studying drugs that target viral entry. Additionally, MHV-3 model is very acute and lethal, with mice succumbing to the infection within a short period of time (around 6–7 dpi). As a result, it may be challenging to detect tissue damage in refractory organs such as the brain and heart [86] using regular techniques. However, despite these limitations, the MHV-3 model can still be valuable in mimicking a severe betacoronavirus infection and may serve as a useful platform for further studies on acute neuropsychiatric changes caused by COVID-19, as well as testing potential novel therapeutic strategies.