It is an undisputed fact that coronaviruses can infect the CNS. CoV RNA has been detected in the CNS of patients with numerous neurological diseases.7,8 CNS infection has caused symptoms of encephalitis in children.9 Cases of meningitis, Guillain-Barré syndrome, and other neuroimmune disorders have also been reported in the context of CoV infection.10–16 HCoV-OC43 is considered the coronavirus with the greatest neuroinvasive potential, as it has been shown to invade, replicate, and remain in the mouse CNS, causing direct damage to neurons.17,18 The virus has also been found to exploit axonal transport.19 HCoV-OC43 can induce cell degeneration and death.20–22 In humans, HCoV-OC43 has been detected in brain tissue from patients with a wide range of neurological diseases, including Alzheimer disease, Parkinson's disease, and multiple sclerosis, as well as in the brains of healthy individuals. Likewise, HCoV-229E has been associated with febrile seizures.23 The literature also includes a report of an immunocompromised child who died due to HCoV-OC43–associated encephalitis.24 Both HCoV-OC43 and HCoV-229E can infect innate immune cells.25,26

MERS-CoV and SARS-CoV can cause severe lower respiratory tract infections, characterised by acute breathing difficulties and such extrapulmonary manifestations as diarrhoea, lymphocytopaenia, hepatic and renal dysfunction, and multiple organ dysfunction syndrome, both in immunocompetent and immunocompromised individuals. The associated mortality rate is higher than 10%, although some cases are asymptomatic.27 MERS-CoV and SARS-CoV have caused 2 epidemics, with a large number of people infected. Presence of neurological symptoms in these epidemics was rare according to the literature, although the available studies only analyse the initial phase of each epidemic. A retrospective study conducted in Saudi Arabia reported confusion in 25.7% of patients with Middle East respiratory syndrome (MERS), and seizures in 8.6%.28 Only 4 cases of CNS involvement (acute disseminated encephalomyelitis, stroke, and encephalitis) and one case of critical illness polyneuropathy have been reported in the context of MERS.29,30 Cases of Guillain-Barré syndrome and delayed-onset peripheral neuropathy have also been reported, and one article even presents the timeline of neurological symptoms in a patient with MERS.31 The literature also includes a case of MERS-CoV infection associated with intracerebral haemorrhage; according to the author, MERS-CoV infection was the only possible explanation for the stroke, with no other risk factors recorded.32 An in vitro study has shown that certain cell lines, such as human neuronal cell lines, are particularly susceptible to MERS-CoV.33 The delayed onset of neurological complications of MERS-CoV infection, the absence of the virus in cerebrospinal fluid, and the unclear association between MERS-CoV infection and Guillain-Barré syndrome suggest that the manifestations of MERS-CoV infection involve an immune component.34 According to some authors, however, the failure to detect the virus in biological samples may be explained by methodological issues.35 During the SARS-CoV outbreak of 2002–2003, reports of neurological symptoms were anecdotal.36,37 Some patients developed axonal polyneuropathy a month after the onset of severe acute respiratory syndrome (SARS), and improved during follow-up. Other patients presented such neuromuscular disorders as myopathy and rhabdomyolysis.38,39 SARS is associated with more severe symptoms than non-SARS viral respiratory tract infection.40 Therefore, muscular disorders may be the result of the patients’ critical situation rather than the virus itself. Stroke has also been reported in the context of SARS-CoV infection.41,42

CoV infection has been proposed as a contributing factor in the pathogenesis of multiple sclerosis (MS).43 This hypothesis is supported by findings from several studies. Coronavirus-like particles have been detected in autopsied brain tissue from a patient with MS.44 CoV isolates were identified in autopsied brain tissue from 2 patients with MS,45 and other researchers have detected CoV RNA in brain tissue from patients with MS.46–48 Another study reports intrathecal synthesis of antibodies to human CoV, which suggests CNS infection.49 Human CoV RNA has also been detected in the cerebrospinal fluid of patients with MS.50 Experimental studies have found that human CoV can infect neurons, astrocytes, and microglia in primary cultures,51 as well as immortalised human microglial cells.52 Interestingly, human CoV can also infect oligodendrocytic cell lines.53 The hypothesis is further supported by the fact that viral upper respiratory tract infections, which may be associated with human coronaviruses, constitute an important trigger of MS relapses.54,55

Mouse hepatitis virus (MHV) is a coronavirus that infects mice but not humans.56 As it induces neurological disorders,57,58 it constitutes an excellent experimental model for studying the effects of coronaviruses on the CNS. Neurotropic MHV strains can infect the CNS by intranasal or intravenous inoculation both in mice59 and in primates.60 Viral entry through the blood-brain barrier has been associated with downregulation of interferon beta production in brain microvascular endothelial cells.61 MHV infection causes acute encephalomyelitis associated with focal areas of demyelination; the role of the immune response during viral infection has been studied.62 Microglial activation and inflammatory mediator expression in the CNS contribute to a local microenvironment that regulates viral replication, which may promote demyelination.63 The demyelinating lesions associated with MHV infection suggest that this model may be useful in multiple sclerosis research.

Although SARS-CoV and SARS-CoV-2 are similar in many respects, they present genetic and structural differences.64 The novel virus is characterised by the ease with which it spreads65; it is more contagious than SARS-CoV,66 spreading through respiratory droplets and contact with infected individuals and objects. It can also be spread by asymptomatic infected individuals.67–69 Several observational studies have analysed the symptoms of the disease, but few have addressed neurological symptoms, with the exception of headache and vestibular symptoms in observational studies70–75 and a specific study of these symptoms in hospitalised patients,76 and isolated case reports77. However, it has been suggested that the virus may affect the CNS, similarly to SARS-CoV, which was detected in the brains of some patients.78 Clinical data reveal differential characteristics, including olfactory alterations and hallucinations; these symptoms have not been reported in patients with SARS-CoV or MERS-CoV infection. Several features of SARS-CoV-2 support the hypothesis of a particular affinity for the CNS.

The main structural differences between SARS-CoV and SARS-CoV-2 are observed in the fusion protein and in accessory proteins ORF3b and ORF8. The coronavirus genome encodes 4 main structural proteins, namely spike, envelope, membrane, and nucleocapsid proteins.79 The viral surface glycoprotein (S) may induce neurodegeneration.80 After infecting host cells, the viral genome is translated into 2 large precursor polyproteins, which are processed by ORF1a-encoded viral proteinases into 16 mature nonstructural proteins (nsp1-nsp16). Nonstructural proteins play an essential role in viral RNA replication and transcription,81 whereas ORF3b has been associated with the immune response.82 As occurs with SARS-CoV, the SARS-CoV-2 spike glycoprotein S1 subunit receptor-binding domain binds to ACE2 receptors, the site of viral entry into the host cell; this has given rise to the possibility of creating an experimental model of SARS-CoV-2 infection.83 Interestingly, the ACE2 receptor is widely expressed in the brain,84 which supports the hypothesis that CNS involvement may be common in the SARS-CoV-2 epidemic.85 In vitro studies report a positive correlation between ACE2 expression and SARS-CoV infection.86 Viruses can enter the CNS via the haematogenous route (in which the virus infects the endothelial cells of the blood-brain barrier) or the neuronal route. It is unlikely that SARS-CoV-2 is able to cross the blood-brain barrier87 due to its large size; the most likely entry route is through the olfactory or trigeminal nerves. Thus, the high incidence of olfactory alterations in SARS-CoV-2–infected individuals may indicate viral entry into the CNS. Findings from studies of the MHV model and the detection of CoV in patients’ brains suggest that the virus may remain in the host's CNS for long periods of time without causing neurological symptoms and that the brain may be a viral reservoir.88

Coronaviruses can enter the CNS, where they may either damage CNS cells or remain latent. Although neurological symptoms are not frequent, they may indicate viral entry into the CNS, which may cause early- or delayed-onset neurological symptoms. Follow-up of patients with SARS-CoV-2 infection should include careful assessment of the CNS.