Neurological consequences of COVID-19: what have we learned and where do we go from here?

A simple explanation for the neurological effects of COVID-19 is direct viral entry and infection of the CNS. Epidemiological data show a latency of up to 1 week between the initial infection and hospital admittance for COVID-19 patients [9, 21], providing a window for potential viral entry into the CNS. Neurotropism is commonly observed in coronaviruses, with neuro-invasive properties well documented in SARS-CoV, MERS-CoV, HCoV-229E, HCoV-OC43, and porcine hemagglutinating encephalomyelitis coronavirus (PHE) [11, 13, 27, 28]. The SARS-CoV-2 spike protein also alters barrier function in human models of the blood-brain barrier, providing an additional mechanism of potential CNS entry [29]. Given the genetic similarity and conserved viral structure with SARS-CoV, it appears likely that SARS-CoV-2 may also exhibit neurotropic properties [30, 31].

Tissue distribution of host receptors is generally believed to decide the tropisms of viruses [32‐34]. In contrast to MERS-CoV, which exploits dipeptidyl peptidase 4 (DPP4) to evade host cells [35, 36], the densely glycosylated spike protein of SARS-CoV-2 virus binds with high affinity to the type I transmembrane metallocarboxypeptidase, angiotensin-converting enzyme 2 (ACE2), providing a mechanism of viral entry into human cells that mirrors the entry point for SARS-CoV [8, 37‐41]. ACE2, which negatively regulates the renin-angiotensin-aldosterone system by degrading angiotensin II to generate angiotensin 1-7, is required to lower blood pressure and as such, is a frequent target for anti-hypertensive drug development [40, 42‐46]. Other functions of ACE2 include the metabolism of apelin-13, neurotensin, kinetensin, dynorphin, [des-Arg⁹]-bradykinin, and [Lys-des-Arg⁹]-bradykinin [47]. ACE2 is widely expressed in airway epithelium, lung parenchyma, vasculature, kidney, heart, and the gastrointestinal tract [48, 49], primary sites of infection by SARS-CoV and SARS-CoV-2; however, it is interesting to note that ACE2-expressing endothelial cells and human intestinal cells were unaffected by SARS-CoV [50, 51], while ACE2 negative hepatocytes were susceptible to SARS-CoV infection [32]. Thus, the expression of ACE2 alone may not be sufficient for host cell infection by SARS-CoV-2.

Initial studies failed to observe ACE2 expression in the brain [48, 49]; yet, RT-PCR studies detected low levels of ACE2 mRNA expression in the human brain while subsequent studies revealed that ACE2 immunoreactivity was exclusively within brain endothelial and smooth muscle cells [52]. ACE2 expression is also reported in both neurons and glia [53, 54], suggesting the brain may be a potential target of SARS-CoV-2. Consistent with the possibility of direct CNS infection, SARS-CoV-2 RNA was detectable in the cerebrospinal fluid (CSF), but not in a nasopharyngeal swab from a 24-year-old COVID-19 patient presenting with seizures, hippocampal atrophy, and pan-paranasal sinusitis that was subsequently diagnosed with viral meningitis [18]. Similarly, a 56-year-old encephalitis patient exhibiting reduced consciousness had detectable SARS-CoV-2 in the CSF. These findings in COVID-19 patients are in agreement with reports showing HCoV-OC43 RNA in the CSF of a 15-year-old acute demyelinating encephalomyelitis patient [55], whereas SARS-CoV was detected in the serum and CSF from SARS patients with persistent epilepsy [56]. Therefore, the capacity to leave the respiratory tract and potentially infect other tissues may be a defining feature of CoVs.

CoVs may enter the CNS via retrograde neuronal diffusion, potentially via the cribriform plate of the ethmoid bone [57]. In mice, ACE2 and TMPRSS2, a protease that contributes toward the spread of CoVs [58], were expressed in sustentacular cells of the olfactory epithelium, with a more pronounced expression in aged mice [59]. SARS-CoV and MERS-CoV were observed within the CNS, raising the possibility of trans-synaptic viral spread via peripheral nerve terminals as a possible mechanism whereby CoVs may gain access to the CNS [27, 28, 60‐62]. SARS-CoV particles were observed in CNS neurons and brain samples from patients diagnosed with SARS [62‐64] while other CoVs, including HEV67 and avian bronchitis virus utilized trans-synaptic transfer [27, 28, 61]. Transgenic mice expressing human DPP4 (hDPP4) under the control of the surfactant protein C or cytokeratin-18 promoter showed a progressive fatal course that was paralleled by high viral titers in thalamus and brain stem within 2–6 days after intranasal administration of MERS-CoV [65]. Similarly, lethal intranasal inoculation of SARS-CoV in transgenic mice expressing human ACE2 (hACE2) in the airway and other epithelia resulted in pro-inflammatory activation and the presence of the virus within the olfactory bulb, thalamus, and brain stem via postulated spread through the olfactory nerves [66]. Following the viral entry into the CNS, infection rapidly spread via a trans-neuronal route to other connected brain regions, culminating in mortality due to neuronal loss in the cardiorespiratory centers within the medulla [67]. Finally, SARS was associated with delayed olfactory neuropathy while the loss of olfactory function is an internationally reported symptom of COVID-19, with some patients showing bilateral obstructive inflammation of the olfactory clefts correlating with impaired olfaction [68‐71]. Thus, retrograde trans-synaptic transport from the lung and lower respiratory airways to the medullary cardiorespiratory centers of the brain and the olfactory centers may mediate the progressive acute respiratory failure and anosmia in COVID-19 patients.

Beyond trans-synaptic spread from the respiratory system, another possibility is movement via the brain-gut axis. The gastrointestinal (GI) tract is directly infected by SARS-CoV-2 and up to a quarter of COVID-19 patients display GI issues, including nausea, anorexia, vomiting, and diarrhea [72, 73]. A temporal correlation exists between GI and neurological symptoms, and it is postulated that anorexia and nausea may be caused, at least in part, by infection of the lateral hypothalamic nuclei [73, 74]. Toward this end, SARS-CoV-2 may enter the CNS via the vagus nerve, a cranial nerve that regulates parasympathetic control of the heart, lungs, and GI tract.

In addition to direct neuronal entry, SARS-CoV-2 may infect non-neuronal cell types to produce neurological complications. The SARS-CoV-2 entry genes, ACEs and TMPRSS2, are detectable in non-neuronal cell types in the olfactory epithelium and olfactory bulb [75]. Thus, infection of glia and vascular cells could contribute toward hypoperfusion, local inflammation, and cytokine release, loss of function of supporting cells, or damage to sustentacular and Bowman’s gland cells to induce olfactory neuronal dysfunction or death [68, 76‐78]. Future studies surely will shed new light on these mechanisms of CNS spread, which may have significant implications for the future treatment of CoV-infected patients.

Critically ill COVID-19 patients exhibited an increased ratio of white blood cells/lymphocytes and higher plasma levels of C-reactive protein (CRP), IL-2, IL-7, IL-10, GSCF, IP10 (CXCL10), MCP-1 (CCL2), MIP-1α (CCL3), and TNF-α, as compared to non-ICU patients [9, 87]. Inflammatory cytokines, such as IL-6, IL-10, and TNF-α, are elevated following infection with SARS-CoV-2 and are believed to orchestrate a cytokine storm [84]. Given these appreciated detrimental effects, a number of clinical trials using tocilizumab, an IL-6 receptor antagonist (NCT04306705, NCT04322773); sarilumab, a IL-6 receptor antagonist (NCT04322773, NCT04315298); or clazakizumab, an IL-6 neutralizing antibody (NCT04343989; NCT04348500), were initiated as potential therapies to limit the cytokine storm in COVID-19 patients.

In contrast to the established association between the cytokine storm and respiratory distress in COVID-19 patients, relatively less is known about the lasting neurological effects of these events. The CNS is regarded as an immune-privileged organ, yet the brain is highly vulnerable to inflammatory mediators and tissue hypoxia [88‐91]. Infectious encephalitis is an inflammation of the brain that may develop in bacteria- or virus-infected children, elderly, and immuno-compromised individuals. While mild encephalitis produces transient flu-like symptoms, including fever, headache, seizures, light sensitivity, neck stiffness, and loss of consciousness, more severe cases can produce confusion, psychosis, limb weakness, double vision, cognitive impairments, speech and hearing deficits, coma, and increased fatality. During the course of COVID-19 infection, reports of a rare condition, acute necrotizing hemorrhagic encephalopathy, emerged in patients showing intracranial cytokine storm syndrome without direct viral invasion [92]. Radiological imaging of acute necrotizing hemorrhagic encephalopathy indicates lesions within the thalamus, brain stem, and cerebral white matter [93], suggesting the likely need for neurological assessments of COVID-19 patients. In addition, cytokine-induced pulmonary injury during ARDS may adversely affect brain function due to the intimate association between the lungs and the respiratory centers in the medulla and pons of the brain stem [94‐97]. Thus, the neurological manifestations of COVID-19 may be secondary to the consequences of ARDS-mediated inflammation and hypoxemia/hypoxia [94, 95]. As clinical data becomes more widely available regarding the link between the nervous and respiratory systems, this knowledge will greatly shape further pre-clinical efforts.

Understanding the immune dysregulation in patients with COVID-19 will provide a greater understanding of SARS-CoV-2 pathogenesis. The detrimental impact of unrestrained immune activation and the cytokine storm are clearly evident, but therapeutic targets beyond anti-viral drugs remain a major obstacle to limiting neurological injury secondary to COVID-19. As a significant member of the pattern recognition receptor (PRR) family, Toll-like receptors (TLRs) play a crucial role in the initiation of immune responses against viral infections. In addition to initiating the intracellular response to viral RNA, TLRs induce signaling cascades and activate transcription factors that shape the cellular response to infection. Along these lines, activation of TLRs mobilize and recruit innate immune cells (e.g., neutrophils, monocytes, innate lymphoid cells) and induce cytokines and chemokines that limit viral progression and activate acquired immunity [98]. Of the TLRs, TLR3, which is expressed in both immune and non-immune cells, recognizes double-stranded RNA (CoVs are double-stranded RNA viruses). Upon activation, TLR3 induces interferon regulatory transcription factor 3 (IRF3) to stimulate the production of type I interferons as a host defense mechanism against viruses [99]. Importantly, mounting evidence suggests that TLR3 may initiate the cytokine storm and drive systemic inflammatory responses [100‐102]. Thus, TLR3 may represent a target for immunotherapeutic modulation to limit neurological dysfunction in COVID-19 patients [103].

COVID-19 patients frequently exhibit complications associated with coagulopathy, including venous thromboembolism, acute coronary syndrome, myocardial infarction, and cerebral infarction [104‐107]. SARS-CoV-2 infection was associated with prolonged prothrombin time, platelet abnormalities, elevated levels of D-dimer, increased fibrinogen/fibrin degradation products, and sepsis-induced coagulopathy (SIC), a form of disseminated intravascular coagulation (DIC), which was observed in the majority of COVID-19-related deaths [108, 109]. Severe COVID-19 patients exhibit hypoxia, a risk factor that increases thrombosis via activation of hypoxia-inducible transcriptional regulation and by increasing blood viscosity [110]. Given the role of coagulopathy, administration of anticoagulants were postulated as a treatment for severe COVID-19 patients [106, 109]; however, anticoagulation did not reduce life-threatening thrombotic complications in a recent multi-center prospective cohort study of 150 COVID-19 patients with ARDS [105], suggesting the need for extensive research to identify alternative targets for therapeutic intervention.

With respect to the CNS, cytokine release, encephalopathy, and onset of ischemic stroke symptoms are correlated in COVID-19 patients [111, 112]. Inflammation and coagulation are inextricably linked processes that exhibit reciprocal cross-talk [113]. Systemic inflammation activates coagulation mechanisms by driving tissue factor-mediated thrombin generation and inhibiting endogenous fibrinolysis. In turn, activation of the coagulation system may influence inflammatory activity and contribute toward the development of hemorrhagic fever and thrombotic microangiopathy. While a clear association exists between SARS-CoV-2 and stroke incidence, it remains unanswered whether coagulation, secondary to COVID-19 infection, is an initiating factor for ischemic stroke or whether the immune response in response to the viral infection worsens the severity of a stroke. In support of the former possibility, elevated inflammation may heighten the risk of developing an acute ischemic stroke in the elderly, potentially via modulation of the coagulation cascade, whereas the latter possibility may be explained by exacerbation of the post-stroke inflammatory response [114‐117]. While it is clear that COVID-19 patients exhibiting pro-thrombotic and/or pro-inflammatory activation may require neurological evaluation, further clinical data and pre-clinical research are needed to define the mechanistic link between SARS-CoV-2 and stroke outcomes.

Given the limited efficacy of broad anticoagulants in COVID-19 patients, alternative therapeutic targets are needed to reduce the detrimental effects of coagulopathies. Neutrophils are circulating innate immune cells that rapidly mobilize to phagocytose pathogens as a mechanism of host protection after an infection. An elevated neutrophil-to-lymphocyte ratio was an independent risk factor for mortality in hospitalized COVID-19 patients [118‐120]. Recent evidence suggests that activated neutrophils also may extrude a meshwork of chromatin fibers into the extracellular space to form cloud-like neutrophil extracellular traps (NETs), which may function as a mechanism of pathogen trapping. Extensive infiltration of neutrophils into the pulmonary capillaries of COVID-19 patients was associated with fibrin deposition and vascular lesions in the absence of sepsis while elevated neutrophil counts were associated with ocular dysfunction during SARS-CoV-2 infection [121‐125]. Moreover, NETs, which stimulate pro-inflammatory responses in human airway epithelial cells [126], are present in many pulmonary diseases, including asthma, chronic obstructive pulmonary disease (COPD), cystic fibrosis, respiratory syncytial virus bronchiolitis, influenza infection, bacterial pneumonia, ARDS, and tuberculosis [127‐132]. While the extent of neutrophil priming and NET formation in ARDS patients correlated with disease severity and mortality [130, 133‐136], the clinical significance of NETs in the pathophysiology of COVID-19 remains undefined.

Sera from COVID-19 patients displayed elevated levels of cell-free DNA, myeloperoxidase-DNA complexes, and citrullinated histone H3, suggesting NET formation and raising the possibility that NETs may provide a potential target for intervention in COVID-19 patients [121, 137]. Interestingly, in addition to roles in host defense against viruses and bacteria, NETs also provide a scaffold for thrombogenesis [138, 139]. Indeed, impaired degradation of NETs is clinically associated with acute thrombotic microangiopathies [140], while the presence of citrullinated histone H3, a biomarker of NET formation, within thrombi retrieved from acute ischemic stroke patients was independently associated with patient mortality [141, 142]. Of interest, we recently reported that elevated NET formation was associated with microvascular occlusion and cerebral hypoperfusion after acute brain injury in both mice and humans [143]. Conversely, administration of recombinant human DNase-I, an FDA-approved drug under investigation for the management of COVID-19-induced ARDS [144], improved blood flow and outcomes after both experimental stroke and traumatic brain injury [143, 145‐147]. Thus, the widespread generation of NETs after SARS-CoV-2 may provide a potential target to reduce acute and chronic neurological consequences, including headache, elevated stroke risk, and potential cognitive issues due to COVID-19.