Introduction
The coronavirus disease 2019 (COVID-19) pandemic, caused by the new severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has become a global health emergency. The most common manifestation in COVID-19 patients are respiratory symptoms and pulmonary infection; however, a growing body of evidence shows extrapulmonary symptoms are also common in COVID-19 patients [
1,
2]. Approximately, 30–80% of patients experience various neuropsychological symptoms such as smell and taste loss, dizziness, neuralgia, seizures, encephalitis, cognitive disorder, memory impairment, stroke, or rare Guillain–Barre syndrome [
1‐
6]. Over one third of patients recovered from the disease have lasting post-infection neuropsychological manifestations collectively referred to as post-acute sequelae of SARS-CoV-2 infection (PASC) [
7,
8]. Several studies have reported the presence of SARS-CoV-2 in brain [
9‐
11] or cerebrospinal fluid (CSF) [
12], suggesting a possible brain invasion of SARS-CoV-2. It is worthy to note, the detection of positive SARS-CoV-2 in CNS is much lower compared to that in the pulmonary system of patients with severe cardio-pulmonary manifestations, raising a caution of potential contamination during sample collecting and processing. Whereas, positive SARS-CoV-2 in CSF of certain patients with mild symptoms and/or lacking significant cardio-pulmonary manifestations implies a possible neural invasion in these patients [
13‐
19]. Postmortem examinations of COVID-19 patients reveal a neuroinflammatory scenario in the central neural system (CNS) characterized by activation of glial cells and infiltration of inflammatory cells [
9,
20,
21]. Glial cells including astrocytes and microglia have key functions in maintaining CNS homeostasis and responding to physical, infectious, and neurodegenerative disease-related insults via producing and releasing pro- and/or anti-inflammatory cytokines and chemokines, antioxidants, free radicals, and neurotrophic factors [
22,
23]. Given that the large number of functions performed by glial cells in the neuroinflammatory response, it could be expected that activation of astrocytes has a major impact on the CNS during SARS-CoV-2 infection [
23,
24].
Angiotensin converting enzyme 2 (ACE2) functions as the primary receptor for SARS-Cov-2 entry [
25]. ACE2 is a transmembrane protein, in which the extracellular peptidase domain (PD) has high affinity to the receptor-binding domain (RBD) of the spike (S) protein of SARS-CoV-2 [
26]. After binding to the PD, the S protein is primed by surface proteinases such as TMPRSS2, TMPRSS11A/B, cathepsin B/L and FURIN to facilitate SARS-CoV-2 entry into host cells followed by deregulation of membrane ACE2 [
27,
28]. In addition to membrane ACE2, intracellular ACE2 and soluble ACE2 have also been identified, although they are incapable of facilitating SARS-CoV-2 entry. ACE2 is an essential counteractor in the multifunctional renin-angiotensin system (RAS) by modulating angiotensin (Ang) peptide balance. The primary function of ACE2 in the RAS is to convert the Ang I and Ang II to Ang 1–9 and Ang 1–7, respectively [
29]. Overactive Ang II signaling promotes inflammation and injuries in multiple tissues and organs. ACE2 ameliorates inflammatory injuries via controlling the ratio of pro-inflammatory Ang II to anti-inflammatory Ang 1–7. Phase II clinical trials have shown administration of ACE2 reduces inflammation that is associated with a shift of Ang peptide balance away from Ang II toward Ang 1–7 [
30,
31]. It could be expected that deregulation of ACE2 by SARS-CoV-2 or S protein shed from SARS-CoV-2 might lead to overactive Ang II signaling. Consistent with this, a recent study shows Ang II levels in COVID-19 patients are elevated and linearly correlated to the viral load and lung injuries, suggesting an imbalance of Ang peptides might, in part, attribute to the development of inflammatory injuries in COVID-19 patients[
29,
32].
In view of the critical role of ACE2 in mediating SARS-CoV-2 entry and protecting inflammatory injuries, knowledge of its brain distribution is informative to understand the neural transmission and pathogenesis of COVID-19. However, understanding its brain distribution remains mixed. A number of research papers and comprehensive reviews have been published on the ACE2 expression in the brain. A recent study performed on the rat brain reports that ACE2 expression in brain capillaries, hindbrain pontine nucleus, pre-Bötzinger complex, nucleus of tractus solitarius, and in neurons, astrocytes, pericytes and endothelial cells. The authors conclude that ACE2-expressing neurons may play a possibly important role in the neural manifestations in COVID-19 patients [
33]. A computational analysis of human and mouse brain databases shows ACE2 is expressed in the choroid plexus, paraventricular nuclei of the thalamus and olfactory bulbs, as well as in neurons, astrocytes, oligodendrocytes, pericytes and endothelial cells. ACE2 expression has also been reported in the cultured astrocytes, microglial cells, pericytes and endothelial cells [
34]. Whereas other studies show neurons, microglial cells and endothelial cells have rare or undetectable ACE2 expression [
11,
35‐
37]. It remains to be determined which brain region(s) and cell type(s) express relatively high ACE2 levels and therefore might be vulnerable to SARS-CoV-2. In addition, it is also not clear by which route(s), such as blood brain barrier (BBB), blood-CSF barrier (BCSFB) or retrograde olfactory migration, SARS-CoV-2 might be more likely to invade into the CNS, particularly in patients without significant BBB impairment. Furthermore, smoking has been reported as a risk factor for COVID-19 severity in current smokers [
38,
39]. Previous studies have shown smoking upregulates pulmonary ACE2, which has been considered to contribute to the infection susceptibility, disease severity and treatment outcome in COVID-19 patients [
40‐
42]. Whereas it remains unclear if smoking exacerbates CNS pathogenesis in a similar or alternative manner. It is also not clear if electronic nicotine vapor (commonly known as electronic cigarette) would harm the lungs or CNS in a similar manner to tobacco cigarettes.
In the current study, we systemically examined the brain distribution of ACE2 and explored the impact of smoking on the mouse brain distribution of ACE2 and the induction of a neuroinflammatory state. Our data elucidate a spatial- and cell type-specific distribution of ACE2 and a pre-conditioned neuroinflammatory state by smoking exposure in the CNS, which might help understand the neural transmission and pathogenesis of COVID-19 in smokers and nonsmokers.
Discussion
Accumulating evidence shows that the CNS could serve as a vulnerable site for neuroinflammation and infection in COVID-19 patients. Postmortem examinations demonstrate a neuroinflammatory signature in infected brain tissues [
9,
66]. The neuroinflammation could be a consequence of direct intracranial viral infection or secondary impacts from extracranially viral infection leading to systematic hypoxia and blood–brain barrier breakdown or excessive inflammatory factors that penetrate into the CNS [
22]. In patients with severe cardio-pulmonary disorders and/or respiratory failure, neurological manifestations could be largely due to inadequate supply of oxygenated blood to the brain, particularly in patients without positive SARS-CoV-2 in CNS. However, the detection of positive SARS-CoV-2 in CSF of certain patients with mild symptoms and lacking significant cardio-pulmonary manifestations implies a possible neural invasion in these patients [
13‐
19]. The observation of positive SARS-CoV-2 signal not coinciding with immune cell infiltration in postmortem brain tissues also suggests that SARS-CoV-2-related neurological complications could be the direct consequence of the neurovirulent properties of SARS-CoV-2 [
22,
67]. The current study shows a spatial- and cell type-specific distribution of ACE2 in the brain, indicating certain brain sites and cell types are more susceptible to SARS-CoV-2 infection than the others. Deregulation of ACE2 signaling and ACE2-expressing cells in specific brain regions might impair various brain functions and therefore be associated with specific symptoms in COVID-19 patients. For example, impairment of ACE2-expressing astrocytes in the olfactory regions might be associated with the loss of sense of smell and taste (Fig.
4), while impairment in the medulla oblongata might be related to pulmonary and cardiac disorders since this region contains vital respiratory and cardiac regulatory centers (Fig.
2A.MY panel and 2B). Impairment in the hypothalamus, including tanycytes of the 3 V, might be associated with systemic disorders such as thrombosis, immune-inflammation and cardiorespiratory aberrancy since hypothalamus is responsible for sensing and regulating the whole-body fluctuation in fluid, electrolyte, metabolism and immunomodulation in addition to being connected to the olfactory/gustative and brainstem cardiorespiratory centers (Fig.
2A.HY panel, 3G and 3I) [
56,
68‐
70]. Lastly, modified ACE2 signaling in pericytes and astrocytes around microvessels might be associated with cerebral microvascular disorders in COVID19 patients, such as possible BBB damage and stroke, since these cells play essential roles in regulating microvascular integrity and function as well as maintaining CNS homeostasis (Fig.
1F and
2B). Furthermore, our study, for the first time, reveals that ACE2 might express in the radial glial cells (Fig.
3F) in the VZ and SVZ of the LV that have been known as progenitor cells in the CNS [
53‐
55]. Impairment of those progenitor cells, as well as astrocytes residing in the neurogenesis zones, might contribute to the lasting post-infection neuropsychological disorders in recovering patients. Consistent with our data, a recent study with human brain organoids reports that neural progenitor cells are vulnerable to SARS-CoV-2 infection [
71]. Furthermore, Klempin et al. report ACE2 plays an essential role in neurogenesis [
72]. Their study shows ACE2-deficent mice demonstrate impaired exercise-stimulating neurogenesis capacity through a mechanism independent of the RAS and serotonin pathways. Our finding of the intrinsic expression of ACE2 in the radial glial cells might provide a plausible explanation to their study. The ACE2 expression in the brain endothelial cells remains controversial. Although early studies show ACE2 expression in vascular endothelial cells in the human brain, more recent studies demonstrated ACE2 is not expressed in the vascular endothelial cells but instead in pericytes [
11,
35]. These inconsistencies could be due the differences in the detection methods, sample preparation, technique limitation in different labs, and possibly altered ACE2 expression by some pre-existed health issues. Our study shows that ACE2 expression is in brain microvascular pericytes instead of endothelial cells (Fig.
1D, F-G). Consistently, our data also show ACE2 is most likely distributed in the ChP stroma tissues instead of ChP capillary endothelial cells (Fig.
3R). The ChP stroma ACE2 could be secreted by local ChP epithelial cell or be filtrated from the blood flow because of the high permeability of ChP capillaries. The ChP stromal ACE2 raises a potential role of extracellular ACE2 in COVID-19, which needs further investigations.
Our studies show that astrocytes are the predominant cells that express the full-length ACE2 (Fig.
5), suggesting those cells are indeed susceptible to possible SARS-CoV-2 infection. Recent studies with postmortem brain tissues demonstrate astrocytes are the major SARS-CoV-2 infected cells, but neurons are rarely infected [
36,
52]. It has also been reported that CNS injuries are accompanied by an elevated plasma level of GFAP in moderate and severe COVID-19 patients [
73]. Our data, together with previous studies, highlight a potential role of astrocyte injury in the CNS transmission and pathogenesis of COVID-19. As one of the most abundant cell types in the CNS, astrocytes perform a large number of functions, such as maintaining CNS homeostasis via secreting neurotrophic factors and regulating intracranial fluid, ionic and metabolic balance, supporting and modulating neurons, regulating BBB integrity and function, and modulating CNS inflammatory and immune responses to physical, pathological and infectious insults [
22‐
24]. It is worthy to note, as an essential counter actor in Ang II signaling, ACE2 protects multiple tissues and organs from inflammatory injuries [
29]. SARS-CoV-2 infection of astrocytes could simultaneously impair the protective functions of ACE2 signaling and astrocytes in the CNS. The interplay of ACE2 signaling and astrocytes might play a key role in any neurovascular impairment due to COVID-19, since these cells are in close proximity to brain pericyte and endothelial cells as well as play an indispensable role in the regulation of the neurovascular unit (Fig.
2B).
Several potential brain entry routes for SARS-CoV-2 have been proposed. Crossing BBB is one of the potential routes. Our data show ACE2 expression in cerebral endothelial cells is extreme low, even lower than neurons (Fig.
5B–C). Given that neurons were rarely infected by SARS-CoV-2 [
36,
52], endothelial cells might be unsusceptible to SARS-CoV-2 infection, too. Thus, it is less likely for SARS-CoV-2 in the blood to enter the CNS in patients through an intact BBB. Crossing BCSFB located in the cerebral ventricular ChP is another potential route. ChP is formed by projected fenestrated capillaries surrounded by a layer of epithelial cells in the cerebral ventricular cavities. Fenestrated capillaries in this region lead to a higher permeability observed in the BCSFB compared to BBB [
74]. Our data suggest ACE2 expression in ChP epithelial cells and stromal tissues between the ChP epithelial cells and the ChP capillaries (Fig.
3Q–R), suggesting possible SARS-CoV-2 filtrated from the fenestrated capillaries may infect ChP epithelial cells and impact BCSFB integrity followed by movement into the CSF. Yet, ACE2 is undetectable in the ependymal cells lining the ventricular walls (Fig.
3Q), indicating it might be less likely for SARS-CoV-2 in the CSF to enter the brain parenchyma directly through the ventricular walls where there is a complete ependymal lining. However, it is possible that, in some regions of the ventricular walls, ACE2-positive cells in the SVZ regions could penetrate through the ependymal layer and contact the CSF. Also, tanycytes located at the 3 V walls in the hypothalamus (Fig.
3G and I) directly contact CSF. These cells that directly contact both the CSF and brain parenchyma might be susceptible sites for SARS-CoV-2 entry. Furthermore, the viruses may spread in the CSF and infect certain regions of meninges such as the ACE2-expressing meninges wrapping the hypothalamus and medulla oblongata to enter the brain parenchyma (Fig.
1G and
2A.MY panel). In line with this hypothesis, medulla oblongata and hypothalamus have been reported as the venerable sites for SARS-CoV-2 infection [
9,
56,
67,
75]. In addition, impaired ChP could induce BCSFB damage leading to the leakage of unwanted plasma proteins into the CSF, which has been observed in more than 40% of patients [
58,
76]. Our data together with other studies support crossing BCSFB might be a likely route for SARS-CoV-2 entry. Retrograde olfactory migration is another possible route. Previous studies have demonstrated the presentence of ACE2 and SARS-CoV-2 in the OB epithelial cells [
21,
60,
61], but it remains unclear if SARS-CoV-2 can migrate into the brain along this route. Our data show ACE2 expresses in astrocytes residing in the subependymal zone (SEZ)/RC and SEZ/RMS sites, suggesting SARS-CoV-2 could be transported into the CNS by those astrocytes along this retrograde olfactory route (Fig.
4).
Several risk factors, including smoking, have been reported to be associated with disease severity in COVID-19 patients [
38,
39]. Smoking, including tobacco smoking and electronic nicotine vaping, upregulates pulmonary ACE2 expression which could attribute to worsened outcomes observed in smokers [
40‐
42]. Smoking might worsen COVID-19 by pre-positioning an abnormal neuroinflammatory/immune state, attenuating the inflammatory regulatory capacity, reducing the pathogen clearance capacity, facilitating virus entry by impairing BBB, and/or might facilitate SARS-CoV-2 replication as it facilitates human immunodeficiency virus (HIV) replication in the host cells [
77]. However, it is not clear if smoking exacerbates neuropathogenesis in a similar or alternative manner. It is also not clear if electronic nicotine vaping would harm the lungs or CNS in a similar manner to tobacco smoking. Our data show experimental exposure to tobacco smoke or electronic nicotine vapor for 14 days does not significantly alter ACE2 expression in the mouse brain tissue (mRNA or protein) (Fig.
6A). These data suggest that unlike the pulmonary system, where smoking upregulates ACE2 levels, smoking might not directly contribute to the SARS-CoV-2/ACE2 interaction-mediated viral entry and/or pathological effects in the brain [
40‐
42]. Instead, it upregulates several cytokines such as IL-1α, IL-6 for tobacco smoking and IL-5 for electronic nicotine vaping (Fig.
6B). These cytokines have been well-documented to be closely related to COVID-19 severity [
78‐
83]. Many studies have pointed to IL-6 as a crucial signature of the cytokine storm in severe COVID-19 patients [
79‐
84]. Members of IL-1 family are cardinal mediators of inflammation [
85], among which IL-1α is one of the archetypical proinflammatory cytokines. IL-5 is a type 2 cytokine functioning as an immunomodulator, upregulation of which might compromise the early immune defense against infection [
86]. Consistent with our data, previous studies have shown smoking increases IL-1α and IL-6 in the plasma and brain tissues [
87‐
90]. Our laboratory has also shown smoking exposure (both tobacco smoke and electronic nicotine vaping) is associated with aggravated brain edema as well as reduced antioxidative molecule NRF2 and increased proinflammatory cytokine TNFα in mouse brain under normoxic and/or ischemic stroke conditions [
49,
91]. It is worthy to note, activated astrocytes are the predominant source of IL-6 and IL-1α, the key contributors to neuroinflammation in the CNS [
22,
85,
92]. Taken together, our ACE2 brain distribution data, alongside studies suggesting smoking may pre-condition a proinflammatory scenario in the brain, suggest that the CNS could be more susceptible to infectious insults with smoking, including SARS-CoV-2 infection, which could be partly due to astrocytes.
There are a few potential weaknesses in the current study design. A sub-chronic (14 days) smoking exposure was used to mimic the effect of tobacco and nicotine exposure on the brain, and the data shows this does not significantly alter brain ACE2 levels. However, the effect of long-term smoking on brain ACE2 levels remains unclear. A chronic exposure (> 1 month or longer) should be completed to address this question in the future. In addition, young and healthy mice were investigated in the current study. Since COVID-19 severity has been well-documented to be correlated with several pre-existed health issues such as diabetes, obesity and hypertension [
93], future studies are warranted to investigates the potential role of ACE2 signaling in the neural transmission and pathogenesis of COVID-19 under these pre-existing disease states. In the future, it would also be important to test ACE2 signaling in aged mice.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.