Mechanisms of COVID-19-related neurologic diseases

Abstract

Purpose of review

As of January 8, 2022, a global pandemic caused by infection with severe acute respiratory syndrome coronavirus (SARS-CoV)-2, a new RNA virus, has resulted in 304,896,785 cases in over 222 countries and regions, with over 5,500,683 deaths (www.worldometers.info/coronavirus/). Reports of neurological and psychiatric symptoms in the context of coronavirus infectious disease 2019 (COVID-19) range from headache, anosmia, and dysgeusia, to depression, fatigue, psychosis, seizures, delirium, suicide, and meningitis, encephalitis, inflammatory demyelination, infarction, and acute hemorrhagic necrotizing encephalopathy. Moreover, 30–50% of COVID-19 survivors develop long-lasting neurologic symptoms, including a dysexecutive syndrome, with inattention and disorientation, and/or poor movement coordination. Detection of SARS-CoV-2 RNA within the central nervous system (CNS) of patients is rare, and mechanisms of neurological damage and ongoing neurologic diseases in COVID-19 patients are unknown. However, studies demonstrating viral glycoprotein effects on coagulation and cerebral vasculature, and hypoxia- and cytokine-mediated coagulopathy and CNS immunopathology suggest both virus-specific and neuroimmune responses may be involved. This review explores potential mechanistic insights that could contribute to COVID-19-related neurologic disease.

Recent findings

While the development of neurologic diseases during acute COVID-19 is rarely associated with evidence of viral neuroinvasion, new evidence suggests SARS-CoV-2 Spike (S) protein exhibits direct inflammatory and pro-coagulation effects. This, in conjunction with immune dysregulation resulting in cytokine release syndrome (CRS) may result in acute cerebrovascular or neuroinflammatory diseases. Additionally, CRS-mediated loss of blood-brain barrier integrity in specific brain regions may contribute to expression of proinflammatory mediators by neural cells that may impact brain function long after resolution of acute infection. Importantly, host co-morbid diseases that affect vascular, pulmonary or CNS function may contribute to the type of neurologic disease triggered by SARS-COV-2 infection.

Summary

Distinct effects of SARS-CoV-2 S protein and CNS compartment- and region-specific responses to CRS may underlie acute and chronic neuroinflammatory diseases associated with COVID-19.

Introduction

The SARS-CoV2 beta-coronavirus causes a severe systemic infection, coronavirus infectious disease 2019 (COVID-19), with profound fatigue, respiratory, gastrointestinal, and neuropsychiatric symptoms [1–4]. During acute COVID-19 common neurological symptoms include headache, anosmia, and dysgeusia, but even patients without a prior history of psychiatric diseases may present with trouble concentrating, poor sleep, fatigue, hallucinations, delusions, and behavioural changes [5]. In addition, cases of meningitis, encephalitis, inflammatory demyelination, infarction, and acute hemorrhagic necrotizing encephalopathy have all been reported as presentations of COVID-19 [6–14]. Moreover, survivors of mild or severe COVID-19 may develop neuropsychological sequelae, including a dysexecutive syndrome consisting of inattention, disorientation, and poor movement coordination [15–21], or mood disturbances, personality changes or psychosis [4]. Clinical studies indicate that multi-organ dysfunction may be due to virus induced systemic inflammatory response syndrome (SIRS) with high levels of cytokine release (cytokine release syndrome; CRS) [22], in addition to direct infection of cells with SARS-CoV-2. SARS-CoV-2 Spike (S) protein has been shown to induce venous and arterial endothelial cell activation and endotheliitis, and to contribute to coagulopathy [23–25]. In addition, accumulating evidence indicates that SARS-CoV-2 may productively infect meningeal, but not parenchymal, CNS compartments [26–30]. Accordingly, mechanisms of CNS dysfunction and neuroinflammation, especially at the vasculature, have been attributed to CRS within the serum or meninges, in addition to hypoxia due to severe respiratory distress. The mechanisms that contribute to neuropsychiatric defects post-COVID-19 are unknown; neuroimaging and cerebrospinal fluid (CSF) analyses in COVID-19 patients suggest persistent inflammatory processes triggered in the acute setting are likely involved, however, few animal models have been developed to define all the neuroimmune correlates involved.

Determinants of SARS-CoV-2 tropism and infection

Viral infections require binding of viral particles to host surface cell receptors. SARS-CoV-2 has a linear, positive-sense, single-stranded RNA genome encoding four structural proteins, known as the S (spike), E (envelope), M (membrane), and N (nucleocapsid) proteins [31]. The N protein contains the RNA genome, and the S, E, and M proteins together create the viral envelope [32]. Coronavirus S proteins are glycoproteins and also type I membrane proteins divided into two functional parts (S1 and S2). The SARS-CoV-2 spike protein S1 catalyzes attachment to, while S2 promotes fusion with the membrane of a host cell. SARS-CoV-2 has high affinity to the receptor angiotensin converting enzyme 2 (ACE2), which is expressed by epithelial and endothelial cells [33]. SARS-CoV-2 spike (S) glycoprotein binds to ACE2 via its receptor binding doman (RBD) [34]. During binding, host serine proteases, including TMPRSS2, TMPRSS4, furin, and endosomal cathepsins, promotes membrane fusion via proteolysis of the S protein, allowing entry of the virus into the cytosol [35,36]. Neuropilin-1, a membrane-bound coreceptor to a tyrosine kinase receptor for both vascular endothelial growth factor (VEGF) and semaphorin family members, is suggested to serve as another host factor for SARS-CoV-2 infection [37,38]. ACE2-positive cells are found in the lung, small intestine, gallbladder, kidneys, testes, thyroid, adipose tissue, heart muscle, vagina, breast, ovary, and pancreas [33], consistent with the multi-organ disease reported in COVID-19 patients. TMPRSS2 is expressed at highest levels by epithelial cells, especially within the gastrointestinal tract and prostate [39], while cathepsin L is expressed in all tissues and cell types. Neuropilin-1, a membrane-bound coreceptor to a tyrosine kinase receptor for both vascular endothelial growth factor (VEGF) and semaphorin family members, exhibits high expression in multiple tissues including the brain, with highest levels found in the hippocampal formation [40].

Studies of ACE2 expression in the CNS using RNA sequencing and RNAscope approaches indicate very low levels may be detected in 0.39–3.6% of human samples, depending on the region examined. ACE2 was relatively highly expressed in the choroid plexus and paraventricular nuclei of the thalamus, with nuclear expression in both excitatory and inhibitory neurons and in astrocytes, oligodendrocytes, and endothelial cells in the human middle temporal gyrus and posterior cingulate cortex [41,42]. A few ACE2-expressing nuclei were also reportedly found in a hippocampal dataset [41]. In addition to these CNS regions, murine datasets indicate expression within the olfactory bulb, and in endothelial cells and pericytes [41–43]. Importantly, TMPRSS2 expression is essentially undetectable in the CNS parenchyma [44], supporting numerous studies that have failed to detect viral RNA within neural cells of the brain. Although neither olfactory sensory neurons (OSN) of the olfactory neuroepithelium (ONE), nor olfactory bulb (OB) neurons express ACE2 and TMPRSS2, these receptors have been detected on OB pericytes and sustentacular cells [42], which maintain the integrity of olfactory sensory neurons [45], but do not provide an olfactory route of entry into the CNS. Immunohistochemical (IHC) studies of human intestinal and choroid plexus tissues detect ACE2 and TMPRSS2 in neuronal perikarya and glial cells of the enteric nervous system, and by epithelial cells of the choroid plexus [46]. ACE2 and TMPRSS2 are also expressed by epithelial cells within the choroid plexus [41], consistent with reports of SARS-CoV-2 transcripts detected at this site in patients with COVID-19 [26].

Does SARS-CoV-2 invade the CNS?

Despite the many neurological manifestations reported in patients with acute COVID-19, multiple studies examining brain tissues from patients with documented CNS diseases have not demonstrated SARS-CoV-2 infection of parenchymal neural cells [47,48]. SARS-CoV-2 nucleoprotein has been detected within brain endothelial cells via IHC [49], consistent with reports of cerebrovascular disease and endothelial cell injury. However, rare cases of meningitis/encephalitis have been reported with SARS-CoV-2 RNA detectable within cerebrospinal fluid (CSF) of some patients [9,50]. ACE2 and TMPRSS2 expression within the choroid plexus [41] is consistent with reports of SARS-CoV-2 transcripts detected at this site in patients with COVID-19 [26]. Infection of the ventricular compartment may contribute to parenchymal inflammation via high levels of cytokines, including interleukin (IL)-1, −6, −8 and tumor necrosis factor (TNF), which, along with β2-microglobulin and glial markers, have been detected within the CSF of patients with COVID-19-related encephalitis [51]. SARS-CoV-2 infection of sustentacular cells has been reported in humans and in animal models of COVID-19 infection [42,52,53]. As mentioned above, there is little evidence that these cells provide a route for entry into the brain, however their infection may lead to anosmia via severe ONE damage or type I interferon-mediated loss of OSN expression of odorant receptors [53–55] Studies in appropriate animal models of SARS-CoV-2 infection, such as use of the Syrian Golden hamster, which may be infected with and transmit SARS-CoV-2 as in humans [56], or variants able to infect mice (strains with B.1.351, B.1.1.28, and B.1.617.1 spike proteins)[57], have similarly not demonstrated SARS-CoV-2 neuroinvasion [30].

Acute SARS-CoV-2 infection and acute neuroinflammation

SARS-CoV-2 infections in humans present with various levels of severity with the most severe cases associated with an exaggerated cytokine release syndrome (CRS), with excessive levels of IL-1β, IL-6, IL-8, IL-17 and TNF. The mechanisms that underlie this effect are not completely understood. Innate immune responses to viral infections are initiated by the recognition of pathogen-associated molecular patterns (PAMPs) by pattern recognition receptors (PRRs), the latter of which induce upregulation of type I and III interferons (IFNs), which act in an autocrine or paracrine fashion to promote the expression of hundreds of interferon stimulated genes (ISGs) [58]. ISGs include both antiviral proteins that disrupt viral replication and packaging within infected cells, and proinflammatory molecules (chemokines and cytokines), which promote both humoral and cell-mediated immune responses that neutralize and clear virus. SARS-CoV-2 initially replicates in epithelial cells of the upper respiratory tract and in pulmonary myeloid cells, during which ssRNA intermediates are detected by primary RNA cytosolic PRRs, including RIG-I-like receptors (RLRs), MDA5, LGP2, and NOD1, and by endosomal toll-like receptors (TLR)-7/8. RIG-I, whose interaction with the SARS-CoV-2 genome does not induce IFN production, directly blocks viral polymerase-mediated transcription of negative-strand RNA [59]. A blunted IFN response has been recognized as a feature of SARS-CoV-2 infection, and may contribute to the observed CRS due to a delay in viral clearance or an imbalance in pro-inflammatory versus anti-inflammatory mechanisms [60]. In addition, recent studies indicate that myeloid cells, which express little to no ACE2, may bind S protein via C-type lectins, including DC-SIGN, L-SIGN, LSECtin, ASGR1, and CLEC10A, or Tweety family member 2, a chloride channel that is expressed within the CNS and PNS [61]. While these interactions do not lead to SARS-CoV-2 infection, they elicit hyperinflammatory responses that correlate with disease severity in COVID-19 patients [62].

Endothelial cell expression of ACE2, which is increased in the context of comorbid conditions such as diabetes, dementia or hypertension [63,64], may bind the S1 subunit of the SARS-CoV-2 Spike protein. Rhea et al. demonstrate that intravenously injected or intranasal radioiodinated S1 is taken up by brain capillaries and may cross the blood–brain barrier (BBB) via adsoptive transcytosis [65]. In vitro studies using human brain microvascular endothelial cells indicate that binding to S1 is associated with upregulation of vascular cellular adhesion molecules and BBB destabilizing cytokines and matrix metalloproteinases, which all contributed to increased paracellular permeability [63]. Studies in hamsters show that S protein downregulates expression of ACE2 in endothelial cells, leading to impairment of mitochondrial function and endothelial NO synthase (eNOS) [66]. Loss of ACE2 has been also suggested to dysregulate the renin-angiotensin system, exacerbating endothelial dysfunction, leading to endotheliitis, which has been observed in basilar and vertebral arteries in COVID-19 patients with encephalopathy, but without ischemia or microbleeds [67]. SARS-CoV-2-induced CRS or S protein-induced endothelial cell injury might underlie the increase in BBB permeability observed in patients with COVID-19-induced encephalopathy [68]. IL-1β- or TNF-mediated activation of RhoA GTPase within brain endothelial cells destabilizes tight junctions via contraction of stress fibers [69]. Studies in hamsters similarly detect increased BBB permeability, particularly in the medulla oblongata and forebrain [30].

Elevated cytokines, such as IL-1, IL-6 and TNF, also promote hypercoagulability via monocyte and/or endothelial cell release of tissue factor (TF), a high-affinity receptor and cofactor for factor (F)VII/VIIa, which together activate the extrinsic coagulation pathway culminating in the formation of fibrin clots [70]. Transcriptomic and proteomic studies of serum in SARS-CoV-2-infected macaques also demonstrate increased complement activation and coagulation cascade pathways [71]. Acute cerebrovascular diseases, such as cerebral hemorrhage, ischemic stroke, and cerebral venous sinus thrombosis were reported early in the course of the pandemic [72], particularly in older individuals with severe COVID-19 and acute respiratory distress syndrome (ARDS). Cerebrovascular diseases may also occur as a consequence of virus-mediated lung inflammation, which has been shown to induce anaerobic metabolism within the CNS with consequent alterations in cerebral blood flow, cerebral vasodilation, interstitial edema, swelling of brain cells, and even ischemia [73]. Thus, persistent hypoxia, cytokine-mediated endothelial cell activation and induction of coagulation may all contribute to acute ischemic stroke during severe COVID-19 infections, especially in those with comorbid conditions.

Systemic cytokines can also cause glial activation leading to increased CNS levels of chemoattractants for peripheral immune cells, promoting inflammation [74]. Studies of post-mortem specimens from patients who succumbed to acute COVID-19 reveal hypoxic damage, microglial activation, astrogliosis, leukocytic infiltration and microhemorrhages [29,75,76]. These data suggest the CNS may develop neuropathology associated with hypoxia and inflammation. Indeed, neuroimaging studies in post-acute COVID-19 patients report disruption of fractional anisotropy and diffusivity, suggesting micro-structural and functional alterations, within the hippocampus [77], a brain region critical for memory formation, that also participates in a subcortical network that regulates anxiety and stress responses. Consistent with this, studies of post-mortem CNS tissues derived from acutely infected patients reveal microglial and neuronal expression of interleukin(IL)-1B and IL-6, respectively, within this brain region. These cytokines are known to regulate the generation of new neurons within the hippocampal subgranular zone (SGZ), which is essential for the formation of episodic and spatial memories. Notably, the SGZ of hamsters and humans with acute COVID-19 exhibit a dramatic decrease in numbers of neuroblasts, which, in hamsters, recovers as SARS-CoV-2 is cleared in the periphery. It is not known whether loss of hippocampal neurogenesis recovers in all or only some humans.

SARS-CoV-2 and chronic neurologic diseases

Long COVID-19 is defined as post-infectious symptoms lasting beyond 3 weeks that may include fatigue, loss of taste and smell, headache, confusion, ‘brain fog’, autonomic neuropathy, muscle weakness, pain, physical disability, dyspnea, chest pain, myocarditis, and postorthostatic tachycardia syndrome [78]. While the mechanisms underlying the neurological manifestations of these chronic symptoms are unknown, emerging data indicate that vascular abnormalities and endothelial cell damage, occurring as a result of hypercoagulability, cytokine-mediated injury, microvascular thrombosis, and ischemia, may produce long-lasting effects of COVID-19 that depend on the brain region affected. For example, post-mortem and MRI studies of COVID-19 patients with persistent anosmia revealed microvascular injury and abnormal enhancement, respectively, within the olfactory bulbs [79].

Acute neurologic syndromes are rare in COVID-19 patients, studies indicate that up to 70% of COVID-19 survivors of severe or moderate disease may also develop ongoing motor, dysexecutive function and memory impairments that worsen over time. Mental skills that are significantly impacted include psychomotor speed, memory, and executive function, the latter of which encompasses defects in working memory, flexible thinking, and self-control. COVID-19 survivors of severe or moderate disease develop ongoing motor, dysexecutive function and memory impairments that worsen over time. Anxiety, depression, post-traumatic stress disorder, insomnia, obsessive-compulsive symptomatology and suicide are also common in COVID-19 survivors, particularly in females [80,81]. Patients with a prior history of psychiatric comorbidities exhibit worsened scores of psychiatric symptoms at one month after hospital treatment [80]. As mentioned above, COVID-19 is associated with CNS exposure to high levels of IL-1β, which has been shown to promote microglial and glial activation, inappropriate neuronal excitability, and to decrease adult neurogenesis within the hippocampus, [82–88] It is possible that certain patients recover slowly or not at all from these effects. Persistent microglial activation has been observed in rodent models of post-viral cognitive dysfunction, which, together with IL-1-mediated effects on neural stem cells, leads to ongoing synapse elimination with lack of repair [84,89–91]. Further studies in human tissues and appropriate animal models are needed to determine the CNS region- and neural cell-specific effects of SARS-CoV-2 infection.

Conclusions

Since the start of the 2019 pandemic caused by SARS-CoV-2, severe neurological diseases have been reported in patients with acute COVID-19, albeit in rare numbers. Considerable evidence suggests that CRS with S glycoprotein-mediated endothelial cell activation or injury contributes to coagulopathy and cerebrovascular diseases. Evidence of SARS-CoV-2 RNA and BBB destabilizing cytokines within the CSF may indicate direct infection of the choroid plexus, with initiation of innate immune responses that contribute to microglial and astrocyte activation, and recruitment of lymphocytes to perivascular spaces. High levels of cytokine exposure at blood and CSF barriers in the setting of hypoxia may similarly promote neuroinflammation, leading to CNS vascular injury. Expression of cytokines by neural cells may underlie neurocognitive impairments or exacerbate neuropsychiatric symptoms in patients with these comorbidities via effects on hippocampal neurogenesis, alterations in neuronal circuitry, or impairment of mitochondrial function. Use of anti-cytokine therapies to treat or prevent neurologic diseases in COVID-19 are currently under investigation in animal models or in clinical trials; assessment of psychoneurologic symptoms should be included in assessments of these symptoms.

Highlights.

• Acute SARS-CoV-2 infection may present with rare neurologic manifestations of diseases.

• One third of Patients hospitalized with COVID-19 develop persistent neurological symptoms.

• While little evidence indicates SARS-CoV-2 invades the CNS parenchyma, rare cases of SARS-CoV-2 RNA detection in CSF suggest infection of choroid plexus may occur.