Brain Disease Network Analysis to Elucidate the Neurological Manifestations of COVID-19

Abstract

Although COVID-19 largely causes respiratory complications, it can also lead to various extrapulmonary manifestations resulting in higher mortality and these comorbidities are posing a challenge to the health care system. Reports indicate that 30–60% of patients with COVID-19 suffer from neurological symptoms. To understand the molecular basis of the neurologic comorbidity in COVID-19 patients, we have investigated the genetic association between COVID-19 and various brain disorders through a systems biology-based network approach and observed a remarkable resemblance. Our results showed 123 brain-related disorders associated with COVID-19 and form a high-density disease-disease network. The brain-disease-gene network revealed five highly clustered modules demonstrating a greater complexity of COVID-19 infection. Moreover, we have identified 35 hub proteins of the network which were largely involved in the protein catabolic process, cell cycle, RNA metabolic process, and nuclear transport. Perturbing these hub proteins by drug repurposing will improve the clinical conditions in comorbidity. In the near future, we assumed that in COVID-19 patients, many other neurological manifestations will likely surface. Thus, understanding the infection mechanisms of SARS-CoV-2 and associated comorbidity is a high priority to contain its short- and long-term effects on human health. Our network-based analysis strengthens the understanding of the molecular basis of the neurological manifestations observed in COVID-19 and also suggests drug for repurposing.

Full text

ORIGINAL ARTICLE
Brain Disease Network Analysis to Elucidate the Neurological
Manifestations of COVID-19
Kartikay Prasad
1
&Suliman Yousef AlOmar
2
&Saeed Awad M. Alqahtani
3
&Md. Zubbair Malik
4
&Vijay Kumar
1
Received: 23 October 2020 / Accepted: 16 December 2020
#The Author(s), under exclusive licence to Springer Science+Business Media, LLC part of Springer Nature 2021
Abstract
Although COVID-19 largely causes respiratory complications, it can also lead to various extrapulmonary manifestations
resulting in higher mortality and these comorbidities are posing a challenge to the health care system. Reports indicate that
30–60% of patients with COVID-19 suffer from neurological symptoms. To understand the molecular basis of the neurologic
comorbidity in COVID-19 patients, we have investigated the genetic association between COVID-19 and various brain disorders
through a systems biology-based network approach and observed a remarkable resemblance. Our results showed 123 brain-
related disorders associated with COVID-19 and form a high-density disease-disease network. The brain-disease-gene network
revealed five highly clustered modules demonstrating a greater complexity of COVID-19 infection. Moreover, we have identified
35 hub proteins of the network which were largely involved in the protein catabolic process, cell cycle, RNA metabolic process,
and nuclear transport. Perturbing these hub proteins by drug repurposing will improve the clinical conditions in comorbidity. In
the near future, we assumed that in COVID-19 patients, many other neurological manifestations will likely surface. Thus,
understanding the infection mechanisms of SARS-CoV-2 and associated comorbidity is a high priority to contain its short-
and long-term effects on human health. Our network-based analysis strengthens the understanding of the molecular basis of the
neurological manifestations observed in COVID-19 and also suggests drug for repurposing.
Keywords COVID-19 .Neurologic comorbidity .Network biology .Disease-gene interaction .Hub genes .Drug repurposing
Introduction
As of the end of October 2020, the novel Coronavirus Disease
2019 (COVID-19) caused by severe acute respiratory syn-
drome coronavirus 2 (SARS-CoV-2) has crossed over 50 mil-
lion cases globally with more than 1 million deaths. While
SARS-CoV-2 is known to cause substantial complications in
respiratory and pulmonary systems, causing pneumonia and
acute respiratory distress syndrome (ARDS), several COVID-
19 cases present various extrapulmonary manifestations of
COVID-19. These include manifestations of cardiovascular,
renal, hematologic, gastrointestinal, hepatobiliary, endocrino-
logical, ophthalmological dermatological, and neurological
systems [1–4]. To date, increasing evidences are indicating
the neurological manifestations of the central and peripheral
nervous system in COVID-19 [5]. These neurological compli-
cations include headaches, dizziness, nausea, loss of con-
sciousness, seizures, anorexia, anosmia, ageusia, encephalop-
athy, and meningo-encephalitis [6–11]. In addition, ischemic
stroke [12], acute necrotizing encephalopathy (ANE) [8], and
acute inflammatory demyelinating polyneuropathy (Guillain-
Barré syndrome, GBS) [13] have also been reported to be
associated with COVID-19. Moreover, COVID-19 patient au-
topsy studies have also identified viral RNA transcripts and
viral proteins in brain tissues [14,15].
The potential mechanisms underpinning the various neuro-
logical syndromes include direct or indirect viral neuronal
injury [16], a secondary hyperinflammation syndrome related
*Md. Zubbair Malik
zubbairmalik@jnu.ac.in
*Vijay Kumar
vkumar33@amity.edu
1
Amity Institute of Neuropsychology & Neurosciences, Amity
University, Noida 201303, India
2
Doping research chair, Department of Zoology, College of Science,
King Saud University, Riyadh 11451, Saudi Arabia
3
Department of Physiology, Faculty of Medicine, Taibah University,
Medina, Saudi Arabia
4
School of Computational & Integrative Sciences, Jawaharlal Nehru
University, New Delhi 110067, India
Molecular Neurobiology
https://doi.org/10.1007/s12035-020-02266-w

to cytokine storm [17], post-infectious immune-mediated dis-
orders, or the effects of a severe systemic disorder with the
neurological consequences of sepsis, hyperpyrexia, hypoxia,
vasculopathy, and/or coagulopathy.
Recently, two studies utilized the 3D human brain
organoids to study the neuroinvasive potentials of SARS-
CoV-2. Ramani et al. [18]haverevealedthatSARS-CoV-2
readily targets cortical neuronal cells and induces tauopathies
and neuronal cell death. The other study by Song et al. [19]
models the SARS-CoV-2 infection of neuronal cells in hiPSC-
derived brain 3D organoids. These comorbidities thus lead to
higher risk of disease development and higher mortality asso-
ciated with COVID-19. It is largely assumed that other neu-
rological manifestations will likely surface in the near future.
Thus, understanding the infection mechanisms of SARS-
CoV-2 and associated comorbidity is a high priority to contain
its short- and long-term effects on human health.
SARS-CoV-2 causes the disease by hijacking the host cell
machinery and perturbs the highly organized cellular net-
works. Moreover, the highly coordinated interactions between
molecules observed in a healthy cell are also gets altered in a
disease condition [20]. This suggests that SARS-CoV-2 inter-
action with host human cells will be different in healthy and
diseased cells, and thus could lead to different impacts of
COVID-19 infection. Therefore, pre-existing clinical condi-
tions can facilitate the appearance of another disease if they
share the same or functionally related genes [21]. As SARS-
CoV-2 has been shown to be associated with several neuro-
logical manifestations, we here predicted the risk of COVID-
19 infection in patients with various neurological disorders.
Over the last 20 years, numerous human-viral interactomes
have been constructed to understand the mechanisms of viral
entry, infection, and disease progression [22–26].
Investigating such interactomes has led to the discovery of
shared and distinct molecular pathways associated with viral
pathogenicity. In the present work, we have utilized a
network-based system biological framework (Fig. 1)toinves-
tigate the molecular interplay between COVID-19 and various
human neurological disorders. Here, we have constructed a
brain-specific protein-protein interaction network of the 332
genes of human’stargetedbySARS-CoV-2reportedby
Gordon et al. [27], with their neighboring genes and named
the network as COVID-19 target network (CTN). The genes
in the network were then further used to identify the neuro-
logical disease associated with them. Based on the shared
genes, we have integrated the CTN with brain diseases and
generated a disease-gene network of the brain (BDGN). This
human brain disease-gene network consists of a total of 123
various brain disorders including COVID-19 interacting with
653 genes of CTN. Out of these 123 diseases, 28 diseases
were directly linked with the COVID-19, indicating the co-
morbidity and complexity of COVID-19. Next, we have iden-
tified the functional modules and hub genes of the CTN
network that can be considered as the hotspot for comorbidity
and could be the target for drug repurposing. We therefore
emphasized that targeting these functional modules will inhib-
it the viral replication and growth and thus will improve the
medical conditions in comorbidity associated with COVID-
19.
Material and Methods
Brain-Specific Protein-Protein Interaction (PPI)
Network for COVID-19 Target Genes
Interactome data of the human brain was retrieved from the
TissueNet v.2 database with 165,240 interactions [28].
Quantitative tissue association for human PPI is provided by
TissueNet. TissueNet gathers RNA-Seq and protein-based as-
say profiles from the genotype-tissue expression project
(GTEX) and human protein atlas (HPA) for preparing exten-
sive interaction networks. Experimentally validated protein
interaction information extracted from four major databases
(DIP, BioGrid, MINT, and IntAct) was also included for pre-
paring PPI networks. A list of 332 genes of humans known to
interact with COVID-19 was obtained from Gordon et.al [27]
and was used for creating a subnetwork having interaction
between the COVID-19 target genes and their neighboring
genes. The subnetwork was called the COVID-19 target net-
work (CTN). Out of 332 genes, 4 genes (GDF15, INHBE,
SBNO1, and CEP43) did not show any interactions in the
brain.
Assembling Brain-Specific Disease-Gene and Disease-
Disease Interaction Network
After obtaining the list of COVID-19 target genes and their
neighboring genes from CTN, Gene ORGANizer [29]tool
was used to identify the brain-related disorders linked with
these genes. Gene ORGANizer database allows us to analyze
the relationship between the query genes and the organs af-
fected by them. The database provides organ-specific disease-
gene relation from highly curated DisGeNET and human phe-
notype ontology (HPO) tools. Disease-gene interactions hav-
ing valid HPO identifiers were considered for further study. A
total of 2002 disease-gene interactions related to brain disor-
der was obtained, which included 127 various brain disorders
interacting with 653 genes. Forty-three out of 653 genes were
the direct target of the COVID-19. Finally, a disease-gene
interaction network was prepared which consists of 780 nodes
(i.e., diseases/genes) and 2046 interactions. Disease-Disease
interaction network was then prepared using the disease-gene
interaction network in which two diseases are considered to be
associated only if they share a common gene known to cause
those diseases. A network of 123 diseases connected with
Mol Neurobiol

each other was prepared, in which 28 diseases were directly
connected to the COVID-19. The calculation of the topolog-
ical coefficient of the disease-gene interaction network also
indicated the tendency of nodes in the network to have shared
neighbors [30]. The values of topological coefficient also sug-
gest that CTN may have functional modules within the net-
work, as the genes in the modules are more connected in-
between rather than the genes present outside the module [31].
Identification of Functional Protein Modules of CTN
Organization of biological function is believed to be in a mod-
ular and hierarchal manner [32]. It is believed that cellular func-
tion modular organization is reflected as modular structure in
molecular network [33]. Protein modules are groups of highly
connected proteins that tend to be co-localized and co-
expressed. Proteins in protein-modules form a single multimo-
lecular machine by interacting with each other at the same time
and place [34,35]. The MCODE plugin of the Cytoscape tool
was used for identifying protein modules in CTN with degree
cut-off value equals to 2, node score cut-off value equals to 0.2
and k-core value equals to 2. MCODE algorithm operates on
three stages, i.e., vertex weighing, complex prediction, and op-
tionally post-processing step, for identifying the locally highly
connected region in the interaction network [36]. A total of 32
modules were obtained, from which the five largest modules in
terms of their sizes have been selected for further studies. These
selected five modules have the maximum number of genes and
also have multiple COVID-19 target genes in it. These five
functional modules thus cover a major part of the COVID-19
target network and most likely will play an important role in the
progression of the disease. The functional modules identified in
CTN are large, so it is important toidentifywhichgenesineach
module best explains its behavior. A widely used approach is to
identify highly connected genes (also known as hub genes) in
the modules [37].
Identification of Hub Genes of CTN
Hub genes in the network are the genes that are highly con-
nected with other genes in the network on a direct basis. Any
change in the expression or activity of the hub gene has the
potential to influence the working of the network. Hubs are
frequently more relevant to the functionality of the network
than other genes. We calculated the topological properties like
the degree of connectivity (K), betweenness centrality value,
and closeness centrality for all the 5 functional modules, sim-
ilar to our previous study [26]. All these network topology
parameters were calculated using the network analyzer plugin
of the Cytoscape tool [38]. Briefly, degree (k) signifies the
number of interactions made by nodes in a network and is
expressed as:
Degree of connectivity kðÞ¼∑vεKuwu;vðÞ
where K
u
is the node-set containing all the neighbors of
node u, and w(u,v) is the edge weight connecting node uwith
node v.
Fig. 1 A strategic workflow adopted in this study with self-explained legends
Mol Neurobiol

Betweenness centrality (C
b
) represents the degree to which
nodes stand between each other based on the shortest paths. A
node with higher betweenness centrality represents more con-
trol over the network. It is expressed as:
CbuðÞ¼∑k≠u≠f
pk;u;fðÞ
pk;fðÞ
where p(k,u,f) is the number of interactions from kto fthat
passes through u,andp(k,f) denotes the total number of
shortest interactions between node kand f.
Closeness centrality (C
c
) is a measure of how fast
information is traveled from one node to other nodes in
the network. Closeness centrality value ranges from 0 to
1, and isolated genes have closeness centrality value
equal to zero.
CczðÞ¼ 1
avg L z;mðÞðÞ
where zis the node for which the closeness value is calcu-
lated and L(z,m) is the length of the shortest path between two
nodes zand m. It has been seen that genes having a high
degree of connectivity also have high closeness centrality
score.
Eccentricity (e
G
) is defined as shortest distance between
one node from all the other nodes present in the network.
The shortest the distance between the nodes, the faster the
travel of information among them. If a node is an isolated
node, its eccentricity value will be zero.
eGvðÞ¼max distGv;uðÞ:uϵVGðÞ
fg
where vis the central node in a connected network Gfor
which we are calculating the eccentricity and uare the nodes
presented in the network.
Topological coefficients (T
f
) indicate the tendency of the
nodes in the network to have shared neighbors. Nodes having
no or one neighbor are assigned a topological coefficient of
zero. The topological coefficient of a node, nwith k
f
neighbors
is computed as:
Tf¼avg j f ;pðÞðÞ
kf
where j(f,p) is the number of shared neighbors between f
and p, plus 1 if there is an edge between f and p.
The top 35 genes having a high degree of connectivity
and betweenness values were considered as hub genes of
the CTN. Apart from topological parameters, we have
also calculated the Dyadicity, which measures the con-
nectedness of the nodes belonging to the same groups,
and the Jaccard similarity coefficient to check the extent
of molecular overlapping among COVID-19 and other
neurological disorders.
Identification of Key Genes and Key Regulators of the
Protein Modules
After identifying the hub genes of the CTN network, we have
identified one key gene in each of the functional modules. A
key gene is among the hub genes of the modules which also
plays role in connecting the functional modules with each
other and provides the idea about how these modules shares
information with each other. Furthermore, for identifying the
key regulator or driver genes which control the regulation of
the modules, tracing of hub genes of each module up to motif
level was performed. Key regulators are the genes that are
highly connected in the network not only when seen from
the top of the network but its connection reaches the lower
levels in the network making the gene an important and influ-
ential part of the network. Apart from identifying the impor-
tant genes in the modules, it is equally important to identify
the processes and the pathways in which these modules play
role, to better understand how the alterations in the function of
these modules affects the host condition.
Gene Ontology and Pathway Enrichment Analysis
For enrichment analysis of protein-modules, Database for
Annotation Visualization and Integrated Discovery (David
v.6.7) tool was used [39]. Gene Ontology (GO) enrichment
analysis includes annotation at biological, cellular, and at mo-
lecular levels. DAVID uses the GO and Kyoto encyclopaedia
of genes and genomes (KEGG) database for the enrichment
analysis of the genes. Pathways and functions having Pvalue
less than 0.05 were considered significantly enriched.
Network-Based Drug Repurposing
For identifying the drug targets for hub genes, key genes, and
key regulators, databases such as DrugBank, Clue.io [40],
ChemblInteraction [41], and DGIdb [42]werescreenedout.
A final drug-gene interaction network was prepared using the
STITCH database [43]. STITCH is a database known to pre-
dict the physical and functional interaction between the query
genes and drugs. The interactions in the STITCH database are
derived from five sources namely, automated text mining,
high throughput lab experiment data, co-expression interac-
tion data, interaction prediction by genomic context, and by
previous knowledge from other databases. For each interac-
tion, STITCH calculated a combined score. A combined score
is calculated by combining the corrected probability of ob-
serving an interaction randomly and probabilities of interac-
tion from different evidence channels. The drug-gene interac-
tions having a combined score value > 0.7 were considered
high confidence interactions.
Apart from chemical molecules as drugs, we also identified
the microRNAs (miRNAs) as a potential drug candidate, the
Mol Neurobiol

miRNAs interacting with the 35 hub genes were retrieved
from the miRTarBase database [44]. mirTarBase database
provides miRNA interaction with several hosts organism in-
cluding humans. This database uses several validation
methods such as reporter assay, western blot, qPCR, microar-
ray, NGS, and pSILAC to validate the interaction between
host genes and miRNAs. For 35 hub genes, 1997 miRNA-
gene interactions were retrieved. Interactions having at least
one strong evidence of validation methods were used for fur-
ther analysis.
Results
Protein-Protein Interaction Network Between COVID-
19 and Human Host in the Brain
For constructing COVID-19 and its human interaction
network, we retrieved the protein-protein interaction net-
work of the brain from TissuevNet2.0 [28] database. A
network consisting of 12,968 number of nodes and
165,241 number of edges was constructed using the
brain’s PPI data (Supplementary Fig. 1). A list of 332
human target genes of COVID-19 was retrieved from
Gordan et.al [27]. A subnetwork of these 332 genes with
their neighboring genes was constructed from the brain’s
PPI network and was named as the COVID-19 target
network (CTN) (Fig. 2a, Supplementary Table 1). Out
of the 332 COVID-19 target genes, 327 genes were part
of the subnetwork. The subnetwork has 5061 number of
nodes and 95,802 number of edges. More than 50% of
interactions from the main network become the part of a
subnetwork which clearly depicts how deeply the
COVID-19 target genes are connected in the brain. The
degree distribution of CTN indicates that the network has
a scale-free property (Fig. 2b). A scale-free network is
one that follows a power-law distribution and is indepen-
dent of the size of the network (i.e., the number of nodes
in the network), which means the basic structural foun-
dation of the network remains the same even when the
network grows [45,46]. Next, we also calculated the
dyadicity (D) between the COVID-19 target genes in
CTN and obtain a dyadicity value as 24.2. Dyadicity
refers to the connectedness of the nodes belonging to
the same group in a network. In a dyadic network, nodes
belonging to the same group are more connected with
each other than in a random network. A network is
called dyadic when D > 1 [47,48]. A value of 24.2 in
our analysis indicates that COVID-19 target genes are
very well connected and making a complex in the net-
work, which can hijack the host cellular machinery. The
complexes in a network are likely to contribute to the
progression of disease and comorbidity at the molecular
level [49]. Proteins present in a community have a higher
probability to play role in comorbidity as compared to
the other proteins which are not part of the complex.
Therefore, to understand the high risks of comorbidities
associated with covid19, we have prepared and briefly
Fig. 2 aCOVID-19 target genes (in red) interaction network in the brain with neighboring genes (in green). bScatter plot showing the distribution of
degree (k) in the COVID-19 target network (CTN)
Mol Neurobiol

analyzed the disease-gene and disease-disease interaction
network of CTN.
Disease-Gene and Disease-Disease Interaction
Network in the Brain
A disease-gene interaction network and disease-disease inter-
action network from CTN were prepared to understand the
links of COVID-19 with neurological comorbidities. A
disease-gene interaction map of CTN was constructed from
the disease-gene interaction data of the brain obtained from
the Gene ORGANizer database. A total of 313 brain diseases,
715 genes, and 2965 disease-gene pairs were considered for
network construction. The disease-gene association map in
CTN will then be constructed by connecting the associated
node (genes) and brain disorder. The resulting network re-
vealed 653 genes linked to a total of 127 disorders, which also
includes COVID-19 (Fig. 3a, Supplementary Table 2).
Figure 3a represents the disease-gene network map compris-
ing of 780 nodes and 2002 disease-gene interactions and is
termed as the brain disease-gene network (BDGN).
The disease-gene interaction network showed that several
disorders were connected with more than one gene in the
network such as ataxia (k= 245), dementia (k= 91), autism
(k= 69), and COVID-19 (k=43) (Fig. 3b). Similarly, the
disease-gene interaction network also showed that many of
the disorder share the common genotype. For example,
SARS-CoV-2 targets, TBK1 (k= 16), BCS1L (k=7),
DNMT1 (k= 7), FBN1 (k= 7), WFS1 (k= 7) and neighbor-
hood nodes, CDKL5 (k= 17), MECP2 (k=16),VCP(k=16),
TARDBP (k= 14), and ATXN2 (k= 12) are linked to multiple
disorders (Fig. 3c, Supplementary Table 3). Also, the calcula-
tion of the topological coefficient of the BDGN network indi-
cated that several nodes in the network have shared neighbors
(Supplementary Fig. 2). Thus, the resulting BDGN reveals the
molecular connection of COVID-19 with various brain disor-
ders and also the close association of COVID-19 targets with
the genes causing the brain disorders (Supplementary Fig. 3).
To further demonstrate the association between COVID-19
and brain disorders, a disease-disease association network was
then constructed, where two diseases were considered to be
related if they share one common gene. The resulting disease-
disease interaction network contains a total of 123 diseases
(nodes) and 436 edges, representing a higher clustering be-
tween diseases (Fig. 3d and Supplementary Table 4). Out of
123 disease nodes in the network, 28 nodes (in red) were
directly linked with COVID-19 (yellow node) (Fig. 3d).
Jackard similarity coefficient was also calculated to check
Fig. 3 aDisease gene interaction network: the figure represents the
interaction of CTN network genes with their related brain disorder.
COVID-19 target genes are represented in red, neighboring genes of
COVID-19 target genes are represented in green, diseases are represented
by blue color, and COVID-19 disease is represented by yellow color. b
Dot plot of highly connected diseases and the number of genes associated
with disease in the brain’s gene-disease interaction network. cBar plot of
genes highly connected to multiple diseases in the brain’s gene-disease
interaction network. dBrain’s disease-disease interaction network. Red
color nodes represent the disease directly connected to COVID-19
Mol Neurobiol

the extent of molecular overlapping between COVID-19 and
other brain-related disorders. Several diseases like ataxia, dys-
arthria, spasticity, cerebral atrophy, autism, dementia, and
stroke were closely related to the COVID-19 and are also
connected with multiple genes in the brain (Supplementary
Fig. 4Aand4B). Thus, because of these molecular overlap-
ping, patients having these diseases are more prone to
COVID-19 and vice-versa. The higher molecular similarities
between brain diseases create a highly complex high-density
comorbidity cluster which contributes to higher mortality in
COVID-19 patients.
To further strengthen the above statement of the high-
density comorbidity, we calculated the eccentricity value of
the network.The eccentricity of a node in a biological network
can be interpreted as the easiness of the node to be function-
ally reached by all other nodes in the network [50]. A node
with a high eccentricity value compared to the average eccen-
tricity value of the network will be more easily influenced by
the activity of the other nodes and conversely could also easily
influence the other nodes in the network by its activity. We
observed that more than 80% of the nodes in the disease-
disease network shows the eccentricity value greater than or
equal to the network’s average eccentricity value (i.e., 4),
which represents that the functionality of the nodes in the
network is highly linked to each other making the network a
highly complex cluster (Supplementary Table 5).
Recent studies evidently showed the extra pulmonary as-
sociations in COVID-19 that contribute to higher mortality in
COVID-19 patients. Thus, it is of utmost necessity to develop
effective drugs to target the patient-specific risks of comorbid-
ity during COVID-19 infection. However, presence of several
overlapping molecular connection makes it difficult to identi-
fy and prioritize the targets for the treatment of the COVID-19
infection. We therefore focused next to target the host func-
tional protein modules linked with diverse brain diseases.
A Broad Range of Disorders Are Linked to Functional
Protein-Modules of Host Interaction Network
It is generally accepted that biological networks are not ran-
domly connected and follow a structural pattern which gives
rise to the modular structure and hierarchical organization.
Modularity suggests nodes that are highly connected in a com-
munity are most likely to have the same biological functions
and play a role in similar pathways [51]. Many complex net-
works exhibit modular structures, where in-between modules
interactions are less dense as compared to interaction within
modules. These modules reflect the organization of the func-
tional unit in a network with relative independence [52].
Generally, diseases sharing similar genes are more
predisposed to form disease modules and comorbidity.
Similarly, genes related to similar diseases are likely to highly
interact with each other [53]. For identifying protein modules,
we used the MCODE module of the Cytoscape tool. A total of
32 protein modules in CTN were identified by MCODE. The
top 5 modules having a larger number of nodes and edges
were selected for further studies (Supplementary Table 6).
Module-1 (Fig. 4a)had 257 nodes and 1542 edges including
15 COVID-19 target genes (in red). Module-2 (Fig. 5a)had
253 nodes and 921 edges including 17 COVID-19 target
genes. Module-3 (Fig. 6a) had 183 nodes and 619 edges in-
cluding 14 COVID-19 target genes. Module-4 (Fig. 7a)had
175 nodes and 899 edges including 10 COVID-19 target
genes and module-5 (Fig. 8a) had 155 nodes and 340 edges
including 5 COVID-19 target genes. The presence of a high
number of COVID-19 target genes in these protein modules
indicated that during the infection, these modules might be
hijacked and strongly altered as compared to other modules
in the network and will eventually disrupt the majority of the
network function. DAVID tool was used for gene ontology
analysis and KEGG pathway analysis.
Biological functional enrichment analysis of module-1 re-
vealed its role in RNA splicing, mRNA processing, protein
complex biogenesis and assembly, regulation of programmed
cell death and pathway analysis reveal module-1 role in pro-
teolysis, ribosome pathway, spliceosome pathway, in
Huntington’s disease pathway, and cancer pathways (Fig.
4b). The nodes of module1 were associated with disorders like
ataxia, dysarthria, spasticity, encephalopathy, coma, and de-
layed speech and language development along with many
other disorders (Fig. 4c).
Similarly, biological functional analysis indicates that
module-2 plays a role in macromolecular complex assembly,
negative regulation of gene expression, RNA transport, and
localization, and play role in pathways like spliceosome, ribo-
some, cell cycle, gliomas, pathways in cancer, and RIG-I like
receptor signaling pathways (Fig. 5b). The nodes are associ-
ated with disorders like hydrocephalus, dementia, autism,
muscular dystrophy, language impairment, and
frontotemporal dementia (Fig. 5c).
Enrichment analysis of module-3 reveals its role in RNA
splicing, cell division, mRNA processing, spindle assembly,
and nuclear division-related biological process, whereas path-
way analysis reveals roles in Notch signaling pathways, Toll-
like receptor signaling pathways, and SNARE interaction in
vesicular transport (Fig. 6b). Nodes of the module-3 are asso-
ciated with diseases like ataxia, sleep apnea, leukodystrophy,
cerebral ataxia, and Joubert syndrome (Fig. 6c).
Module-4 plays a role in biological functions like RNA
processing, translational elongation, ncRNA processing, and
in the viral infection cycle, and also plays a role in pathways
like NOD-like receptor signaling pathways, Parkinson’sdis-
ease, Huntington’s disease, and Cytosolic DNA sensing path-
ways (Fig. 7b). Module-4 is associated with diseases like atax-
ia, spasticity, amyotrophic lateral sclerosis, chorea, and audi-
tory neuropathy (Fig. 7c).
Mol Neurobiol

Module-5 biological functional analysis reveals its role in
the regulation of transcription from RNA polymerase II
promoter, chromatin modification, regulation of transcription,
and in cellular protein catabolic process, pathways enrichment
Fig. 4 aModule1 with COVID-19 target genes (in red). bBiological functions and KEGG pathways related to module-1.cDot plot of module-1 related
disease with their gene counts
Fig. 5 aModule-2 with COVID-19 target genes (in red). bBiological functions and KEGG pathways related to module-2. cDot plotof module-2 related
disease with their gene counts
Mol Neurobiol

analysis reveal a role in Wnt signaling pathways, viral repro-
duction, MAPK signaling pathway, neurotrophin signaling
pathway, TGF-beta signaling pathway, and ErbB signaling
pathway (Fig. 8b). Module 5 is associated with diseases such
Fig. 6 aModule-3 with COVID-19 target genes (in red). bBiological functions and KEGG pathways related to module-3. cDotplotofmodule-3-
related disease with their gene counts
Fig. 7 aModule4 with COVID-19 target genes (in red). bBiological functions and KEGG pathways related to module-4. cDot plot of module-4-related
disease with their gene counts
Mol Neurobiol

as ataxia, myopathy, Chiari malformation, GAIT ataxia, limb
ataxia, and febrile seizures (Fig. 8c).
Interestingly, during the early stage of infection, SARS-
CoV-2 hijack these modules and the concerned biological
processes and synthesize its RNA. Many clinical conditions
such as ataxia, spasticity, encephalopathy, and dementia are
associated with all modules, revealing a greater risk of the
severe illness of COVID-19 patients. The presence of com-
mon disease, the pathway, and the biological functions in the
modules indicated that these modules are connected with each
other to some extent (Supplementary Table 7). Besides, we
also observed a different spectrum of disorders like cancer,
myopia, dysarthria, sleep apnea, and thromboembolic stroke
that were associated with these modules. Therefore, besides
neurologic comorbidity, disorders in various other organs can
also be a potential threat for COVID-19 patients as also dem-
onstrated by Gysi et al. [54]. Our network-based results will
thus add to strengthen these observations.
Core Regulatory Hubs of the Protein Network
Several virus-host network studies have indicated that viral
proteins target the hub proteins which have a high degree of
connectivity in the network [22,24–26,55]. Here, we have
identified 35 candidate hub proteins that exhibit a high degree
of connectivity and betweenness value, using the Cytoscape’s
Network analyzer tool (Supplementary Table 8). Out of these
35 hub proteins, 16 proteins bind to SARS-CoV-2 [27].
Biological process enrichment analysis of these 35 hub pro-
teins using GeneMania webserver and GO biological process
analysis indicated their role in proteasome-mediated ubiqui-
tin-dependent protein catabolic process, G1/S transition of the
mitotic cell cycle, Notch signaling pathway, ERBB signaling
pathway, and response to hypoxia (Fig. 9and Supplementary
Table 9). The GO function analysis reveals the roles of hub
proteins in nitric-oxide synthase regulator activity, ubiquitin-
protein ligase binding, transcription regulator activity, RNA
binding, and cytoskeletal protein binding (Supplementary
Table 9).
Strikingly, 26 of the SARS-COV-2 viral proteins inter-
act with the above-identified 16 hub genes (Table 1). As
can be seen, SARS-CoV-2’s NSP8, NSP13, ORF3a, and
ORF9c target the most of these hub proteins (i.e., eight in
total), SARS-CoV-2 NSP2 and ORF7a targets seven hub
proteins, while SARS-CoV-2 N, M, NSP6, and NSP7 have
six hub protein targets (Table S3). Other SARS-CoV-2
proteins like NSP’s (4, 9, 12) and SARS-CoV-2 ORFs’
(3b, 8, 10) have five targets. Intriguingly, all the 16 hub
genes are targets of more than one SARS-CoV-2 protein
(Table 1), out of which NPM1, SNW1, GRB2,
HSP90AA1, and ELAVL1 are the target of several
SARS-CoV-2 proteins. These findings are consistent with
previous findings which indicates that an individual viral
protein can target multiple host proteins and several viral
Fig. 8 aModule-5 with COVID-19 target genes (in red). bBiological functions and KEGG pathways related to module-5. cDotplotofmodule-5-
related disease with their gene counts
Mol Neurobiol

proteins can interact with the same host protein [25,26,
55–57].
Moreover, the spatio-temporal expression analysis of these
35 hub genes was studied through the BEST (Brain
Expression Spatio-Temporal) webserver (http://best.psych.
ac.cn/#) [58]. This server utilizes eight human brain
expression datasets obtained from BrainSpan Atlas, Allen
brainmap, GTEx, and other sources. The expression pattern
of selected genes in different brain regions (spatial pattern)
and age stages (temporal pattern) were analyzed using RNA-
seq Data from Brainspan and RNA-seq Data from GTEx and
has been shown in Supplementary Fig. 5. As can be seen in the
figure that most of the genes were upregulated from the neo-
natal stages to adulthood, these genes were downregulated in
the hypothalamus, olfactory bulb, substantia nigra, insula, and
parahippocampal region (Supplementary Fig. 5A). However,
in older age, the genes were moderately upregulated in the
hypothalamus and substantia nigra, whereas they largely get
downregulated in the olfactory bulb, thalamus, and in the cor-
tical region (Supplementary Fig. 5B). The downregulation of
these genes in the olfactory bulb, thalamus, substantia nigra,
and the cortical regions of the brain indicates the impairment
of sensory systems, memory, and cognition.
Network-Based Drug Repurposing
Because of high connectivity and the ability to rapidly transfer
information in the network, hub genes could be the most ap-
propriate target in any network for drug identification [26].
We propose to target these hub proteins, hijacked by SARS-
CoV-2, through drug repositioning. Initially, we identified
drugs for the 35 hub genes by screening multiple databases
related to drug-genes interaction [42,59,60]. A total of 3286
drug-gene interactions for 35 hub genes were identified
(Supplementary Table 10). For 16 hub genes showing inter-
action with COVID-19 viral protein, 1477 drug-gene interac-
tions were identified. Out of these 16 genes, 12 genes showed
significant interactions with 41 drugs having a combined
score > 0.7 in the STITCH database (Supplementary Fig. 6).
Considering the severity and complexity of COVID-19, we
also target key genes connecting the functional modules using
drug repurposing. Targeting these key genes makes much
more sense, asinformation transmitted by these key genes will
further provide instructions to other modules controlling the
network. We have identified five key genes (ESR1, TP53,
UBC, HSP90AA1, and MYC) that were linking the modules
in-between and can act as a suitable candidate for drug
Fig. 9 Protein-protein interaction network of 35 hub genes derived from GeneMANIA along with functional enrichment
Mol Neurobiol

repurposing (Supplementary Fig. 7A). A total of 138 number
of FDA approved drugs were identified for the 5 key genes. A
drug-gene interaction network was prepared, and the STITCH
database was used for the final categorization of interactions.
Interactions having a combined score value > 0.7 were con-
sidered as high confidence interactions (Supplementary
Fig. 7B). Moreover, earlier clinical studies indicated that
many of the drugs used for the treatment of viral CNS infec-
tions have poor CNS penetration abilities and thus were un-
able to inhibit viral RNA replication [61]. These drugs were
shown to have lower log Pvalues (CNS penetration ability)
and hence were unable to cross the blood-brain barrier.
Ideally, according to Lipinski’s rule of 5, a log P value less
than 5 has been considered as a pharmacologically effective
drug [62]. The log Pvalue of the identified drugs has been
shown in parentheses and this information would be highly
useful for selecting the drugs for drug repurposing against
COVID-19 induced CNS infection.
ESR1 gene showed high confidence interactions with FDA
approved drugs, fulvestrant (logP, 6.54), tamoxifen (logP,
5.93), raloxifene (logP, 5.45), estradiol (logP, 3.57), and
ethinylestradiol (logP, 3.63). TP53 gene showed interaction
with etoposide (logP, 0.73), 5-fluorouracil (logP, −0.89),
mitomycin-c (logP, −0.40), bortezomib (logP, 0.89), and
doxorubicin (logP, 1.41) drugs. MYC gene showed interac-
tion with imatinib (logP, 3.47), calcitriol (logP, 5.51), and
amifostine (logP, −1.4) drugs and HSP90AA1 gene showed
interaction with rifabutin (logP, 4.25) drug.
Key regulators in a network are the proteins that are deeply
rooted in the network/modules and serve as the backbone of
the network. Since networks serve as hierarchical characteris-
tics, removal of key regulators from the network may cause
alarming changes in the whole network and especially in the
deeper levels of the network [63]. Thus, these key regulators
will be proved as a good target gene in the modules. A total of
12 key regulators in 5 modules were identified by tracing the
hub proteins of each module up to the motif level (Fig. 10 a).
For module 1, YWHAQ, CUL4B, CUL2, and HSPA8 genes
were identified as key regulators. Similarly, for module 2, the
HDAC1 gene was identified, for module 3, the KAT2B gene,
for module 4, the MYH9 gene, and for module 5, SMARCA4,
SMARCC2, MYC, H2AX, and GAN genes were identified as
Table 1 Interactionof16hub
genesoftheCTNwithSARS-
CoV-2 proteins
Proteins The hub genes of the CTN
SARS-COV2 NSP1 NPM1, SNW1
SARS-COV2 NSP9 ELAVL1, HSP90AA1, NPM1, SNW1, TP53
SARS-COV2 NSP4 HSP90AA1, NPM1, SNW1, VCP, XPO1
SARS-COV2 SPIKE GRB2, NPM1, SNW1
SARS-COV2 ORF6 CDK2, ELAVL1, NPM1, SNW1, VCP
SARS-COV2 NSP14 GRB2, SNW1
SARS-COV2 NSP7 CAND1, CDK2, GRB2, HSP90AA1, NPM1, XPO1
SARS-COV2 ORF9C CAND1, CCDC8, ELAVL1, GRB2, HSP90AA1, NPM1, VCP, XPO1
SARS-COV2 ORF9B ELAVL1, NPM1, SNW1, XPO1
SARS-COV2 ORF3B ELAVL1, GRB2, HSP90AA1, NPM1, SNW1
SARS-COV2 M CAND1, CDK2, HSP90AA1, NPM1, SNW1, XPO1
SARS-COV2 ORF10 CUL2, NPM1, SNW1, VCP, XPO1
SARS-COV2 ORF8 APP, CCDC8, NPM1, SNW1, XPO1
SARS-COV2 N CDC5L, ELAVL1, HSP90AA1, MOV10, NPM1, VCP
SARS-COV2 NSP11 CUL2, ELAVL1, HSP90AA1, NPM1, SNW1
SARS-COV2 NSP15 NPM1, SNW1
SARS-COV2 NSP2 CDC5L, CDK2, ELAVL1, GRB2, HSP90AA1, NPM1, SNW1
SARS-COV2 NSP5 GRB2, NPM1, SNW1
SARS-COV2 NSP8 CCDC8, CDC5L, ELAVL1, GRB2, MOV10, NPM1, RNF2, SNW1
SARS-COV2 NSP13 CDC5L, CDK2, GRB2, HSP90AA1, NPM1, SNW1, TP53, VCP
SARS-COV2 E CDC5L, ELAVL1, NPM1, SNW1
SARS-COV2 NSP6 CAND1, GRB2, NPM1, SNW1, VCP, XPO1
SARS-COV2 ORF7A CAND1, CDC5L, HSP90AA1, NPM1, SNW1, VCP, XPO1
SARS-COV2 NSP12 APP, GRB2, HSP90AA1, NPM1, SNW1
SARS-COV2 NSP10 ELAVL1, NPM1, SNW1
SARS-COV2 ORF3A CDC5L, ELAVL1, GRB2, HSP90AA1, NPM1, RNF2, SNW1, TP53
Mol Neurobiol

key regulators. Drug targets against these key regulators were
identified. A total of 107 FDA approved drugs were identified
for these key regulators. Out of these 12 key regulators, 6 key
genes showed significant interaction with drugs (combined
value of > 0.7) using the STITCH database (Fig. 10b). For
the HSPA8 protein of module-1, significant interactions were
obtained with compounds such as hydrogen peroxide (logP, −
0.45), ibuprofen (logP, 3.97), arsenite, and DB07045 (logP,
1.39). Similarly, the HDAC1 protein of module-2 showed
interactions with trapoxin B (logP, 3.1), trichostatin A (logP,
2.36), and vorinostat (logP, 1.88) drug molecules. Module-3
protein, KAT2B showed interaction with R4368 (N-(3-
AMINOPROPYL)-N-(2-NITROPHENYL)) (logP, 1.6) drug
molecule. MYH9 protein of module-4 showed interaction
with blebbistatin (logP, 2) and module 5 protein, MYC
showed interactions with trichostatin A (logP, 2.36),
troglitazone, tamoxifen, etoposide, doxorubicin, and arsenite.
Another protein of module-5, H2AX showed interaction with
vorinostat, etoposide, and doxorubicin. Many of the drugs
targeting these key regulators can be used either individually
or in combination.
Finally, we searched the miRNAs target for the 35 hub
proteins that may be considered as a potential drug to target
the comorbidity. Among these 35 proteins, 19 proteins were
targeted by 157 miRNAs. Of which, 11 miRNAs including,
miR-125a-5p, miR-125b-5p, let-7a-5p, miR-130a-3p, miR-
34a-5p, miR-29b-3p, miR-27a-3p, miR-24-3p, miR-145-5p,
miR-200a-3p, and miR-200c-3p were noticed to target more
than three interconnected proteins (Fig. 11). These miRNAs
thus may show its utility as a drug for therapeutic intervention.
Discussion
The most alarming situation in COVID-19 infection is the
associated comorbidity in the aged patients which increases
the severe health risks worldwide. The present work has
shown the risk of COVID-19 infection on the onset of various
neurological disorders and the molecular basis of this comor-
bidity through the principle of system-based network biology.
We here showed the complexity of COVID-19 and wide
range of SARS-CoV-2 targets in the host cell, which estab-
lishes the molecular connection with various brain-related dis-
orders. The disease-gene and disease-disease network map
has revealed the overlapping molecular connections and high
clustering of diseases within the vicinity of the same network,
thus demonstrating a close pathobiological similarity.
Moreover, the formation of a scale-free network among brain
disorders and COVID-19 indicates the presence of a high-
density comorbidity cluster. These results increase our under-
standing of the molecular basis of the neurological comorbid-
ity associated with COVID-19 patients.
Another significant finding is the existence of functional
modules that exhibit greater connectivity among nodes within
amodule(Figs.4,5,6,7,8). Therefore, targeting the func-
tional protein modules which are primarily hijacked by
SARS-CoV-2 and are the origin of many brain disorders will
prevent the virus replication and growth. The most highly
connected module relates to the ubiquitin-proteasome system
(UPS), ribosomal proteins, cell signaling, and cellular export
proteins that can be considered as critical targets for reducing
the SARS-CoV-2 infection. Furthermore, in-depth network
analyses revealed 35 hub proteins present in different func-
tional modules and are involved in more than one function
(Supplementary Table 8). To provide the system-wide impor-
tance of these hub proteins in COVID-19 and associated co-
morbidity,we categorized these hub proteins into three groups
based on their functionality. The proteins in the first group
have been largely relevant to the UPS, cell cycle, and cell
death. Similarly, the proteins in the other two groups are in-
volved in transcription regulator activity, RNA metabolic
pathway, signaling pathway, nuclear transport, and cytoskel-
etal protein binding.
The group 1 proteins are generally multifunctional and
contain genes responsible for highly represented UPS. These
are CUL1, CUL2, CUL3, CUL7, CAND1, OBSL1, VCP,
FBXO6, UBC, and RNF2. Four of these genes (CUL1,
CUL2, CUL3, and CUL7) belong to the core component of
E3 ubiquitin-protein ligase complexes, which mediate the
ubiquitination of target proteins. Three other UPS-related pro-
teins, VCP, FBXO6, and UBC are the key mediators of the
ER-associated degradation (ERAD) pathway. The UPS is
largely involved in the maintenance of cellular homeostasis
and plays a fundamental role in viral replication [64,65].
Several viruses, including coronaviruses, often modulate
ubiquitin and ubiquitin-related pathways for their survival
[66,67]. In this study, we find that module 1 and module 2
includes several members of ubiquitin-ligase and ER-
associated proteins (Fig. 4a and 5a). Since the loss of
ubiquitin-proteasome activity results in aggregation of pro-
teins and disturbs cellular functions, we suggested that
SARS-CoV-2 targets these modules to hinder the UPS-
mediated cellular responses. The other genes of this group
are involved in the regulation of cell cycle and cell death. It
was recently suggested that COVID-19 infection may cause
neurodegenerative disease and leads to neuronal death. The
study revealed that SARS-CoV-2 infection-induced patholog-
ical effects closely resembling tauopathies and neuronal cell
death [18]. In our study, we found genes like APP, TP53,
MYC1, VCP, and UBC as hub genes that are involved in
protein misfolding and aggregation, which ultimately leads
to cell death.
Moreover, the second largest group of hub genes is in-
volved in transcription regulation and RNA binding. The
genes in this group include MOV10, XPO1, ESR1, SIRT7,
Mol Neurobiol

Mol Neurobiol

NPM1, ELAVL1, NXF1, TP53, CDC5L, SNW1, RNF2, and
COPS5. A recent network study identified XPO1, NPM1,
HNRNPA1, and JUN, as “hub proteins”with the highest
number of functional connections within 119 host proteins
that interact with the human SARS-CoV-2 [56]. Of note, both
NPM1 and HNRNPA1 interact directly with XPO1 in normal
cells, and JUN is involved in inflammatory responses, includ-
ing those associated with viral infections.
A recent single-cell RNA sequencing of the SARS-CoV-2
infected human brain organoid showed that SARS-CoV-2 in-
fection induces a locally hypoxic environment in neurons
[19]. Also, post-mortem studies of the brain of COVID-19
patients indicated the acute hypoxic-ischemic damage in the
brain [14]. Consistent with these results, our findings indicate
the enrichment of the hypoxia-inducing genes in the BDGN,
including CUL2, TP53, UBC, and MDM2, suggesting a pos-
sible mechanism of SARS-CoV-2 induced neuropathology.
We then mapped the known drug-target network to search
for druggable, cellular targets. We identified 41 approved tar-
gets for the 16 hub genes. For example, TP53, CDK2,
HSP90AA, HDAC1, NPM1, and ESR1 were the most target-
able proteins. One of the important drug targets is the export
protein, XPO1, which enables the transport of viral proteins
from the nucleus to the cytoplasm. The FDA-approved drug,
Selinexor (logP, 2.85), (Fig. 12a) is a selective inhibitor of
nuclear export (SINE) compound which blocks the cellular
protein XPO1 and is now entered into a randomized clinical
trial in COVID-19 patients as announced by Karyopharm
Therapeutics Inc. [68]. Selinexor is FDA-approved drug
against multiple myeloma. SINE compounds have been used
to block the replication of several viruses and act as anti-
inflammatory drugs [69]. In a study, verdinexor (logP, 3.7),
which is highly similar in structure and biological activity to
selinexor, was identified as a candidate drug inhibiting the
interactions of several of the SARS-CoV-2 proteins to their
Fig. 10 Identification of key regulators of CTN. aTracing of hub genes
through different levels in the modules. The red arrow represents the
transfer of hub genes to the next level. bInteraction network of key
regulators with FDA-approved drugs. The green color interaction be-
tween the drug and genes was a significant interaction with a combined
score > 0.7
Fig. 11 miRNA network. The network shows the miRNA (red node) targeted 19 hub genes (green node). The miRNA (yellow highlighted node) that
binds to three or more than three hub genes are shown only
Mol Neurobiol

human targets and was recommended for further evaluation to
target COVID-19 [27].
TheotherkeydrugtargetgeneisMYH9whichhasbeen
traced to a candidate gene of module- 4 largely involved in viral
infection. The MYH9 gene is involved in synthesizing a protein
called myosin-9 which is one subunit of the myosin IIA protein.
The MYH9-related disorder includes thrombocytopenia, hear-
ing loss, kidney disease, and cataracts [70]. Mutations in the
MYH9 gene lead to a genetic condition known as MYH9-
related thrombocytopenia (MYH9RD) which is characterized
by the presence of large but lesser platelets, results in bleeding
in the body or the skin [71]. Thrombocytopenia is now consid-
ered as a risk factor for increased morbidity and mortality in
COVID-19 patients. According to a recent study, COVID-19
patients showed blood clots in small vessels, clots in the lungs,
and stroke-causing clots in cerebral arteries [72]. In some cases,
a stroke may also lead to a ruptured blood vessel, and cause
bleeding in the brain [73]. We here found that the chemical
compounds bind to MYH9 includes guanosine triphosphate,
MgADP, MgATP, and blebbistatin (logP, 2) (Fig. 12 b). The
drug blebbistatin is a myosin-2 inhibitor and has been common-
ly used drugs in several research disciplines including neuro-
science, muscle physiology, cell migration, cell differentiation,
cell death, and cancer research [74].Ithasalsobeenshownto
reduce the viral infection in many of the viruses by inhibiting its
entry [75] and reducing the viral loads [76].Here, we also sug-
gests eleven micro-RNAs (miRNAs) which shows the binding
with hub proteins and also regulates hub proteins. The miRNA
has been considered as drug molecules for targeting the SARS-
CoV-2 proteins [77]. It is largely accepted that host cellular
miRNA can directly target both the coding region of the viral
genome and 3’UTR to induce the antiviral effect. Very recently,
Fulzele et al. [78] identified over 800 miRNAs that target the
COVID-19 genome. Interestingly, many of the miRNAs iden-
tified in this study (e.g., miR-130a-3p, miR-29b-3p, miR-27a-
3p, miR-145-5p, and miR-200a-3p) are found to be unique to
the SARS-CoV-2 genome according to the results of the
Fulzele et al. [78]. Moreover, miRNAs including miR-34a-5p,
miR-200a-3p, and let-7a-5p have been shown to strongly inter-
act with the SARS-CoV-2 [79]. Besides, polyphenols are also
considered as a potential and valuable natural compound for
designing new drugs that could be used effectively in the com-
bat against COVID-19 [80–82]. Overall, our BDGN analysis
opens up avenues for multiple candidate repurposable drugs
that target diverse cellular pathways involved in COVID-19
and associated neurologic comorbidity.
Conclusions
In conclusion, we constructed a human-SARS-CoV-2 BDGN
interactome, demonstrating the risk of COVID-19 infection
on multiple brain-related disorders. The study of disease-
gene and disease-disease association networks reveals
COVID-19-related structural and functional modules that es-
tablish the molecular connection with brain disorders. Also,
Fig. 12 Drug-gene interaction network of two important key genes and
its drugs using the STITCH database. aThe network shows selinexor,
verdinexor, and guanosine triphosphate as drug candidates for XPO1. b
The network shows blebbistatin as a candidate drug that binds to MYH9
and could be considered for drug repurposing
Mol Neurobiol

the highly clustered network of the diseases suggests close
pathobiological similarity and large comorbidity. We also
identified crucial functional hubs in the network that showed
close links with brain-related disorders including dementia,
ataxia, encephalopathy, stroke, and many other disorders that
indicate a high degree of comorbidity and the severity of
COVID-19. We here also propose candidate drugs and
miRNAs, targeting the hub proteins interconnecting the mo-
lecular pathways linked to virus infection that can be used to
treat SARS-CoV-2 infection. We believe our results will help
to understand the molecular pathobiology of neurological co-
morbidities linked to COVID-19. Given the severity and com-
plexity of SARS-CoV-2 infection, we further suggest that
drugs in combinations or drugs targeting multiple proteins will
be the choice to improve clinical outcomes.
Supplementary Information The online version contains supplementary
material available at https://doi.org/10.1007/s12035-020-02266-w.
Acknowledgments S.Y.A is grateful to the Deanship of Scientific
Research, King Saud University, for funding through Vice Deanship of
Scientific Research Chairs. This project was supported by the health
ministry, Project number 895, Saudi Arabia. K.P. and V.K. sincerely
thank the Amity University, Noida, for providing facilities. MZM is fi-
nancially supported by the Department of Health and Research, Ministry
of Health and Family Welfare, Government of India under young scientist
FTS No. 3146887.
Authors’Contributions V.K. designed the study. K.P. and M.Z.M per-
formed the experiments and calculations. K.P., M.Z.M, and V. K pre-
pared figures for the results. V.K., K.P., S.Y.AO, and S.A.M.A analyzed
the data and wrote the main text.
Funding This work was supported by the funding from Department of
Health and Research, Ministry of Health and Family Welfare,
Government of India, under young scientist FTS No. 3146887. This
project was supported by the health ministry, Project number 895,
Saudi Arabia.
Data Availability The data generated and analyzed during this study are
included in this article and its supplementary information files.
Compliance with Ethical Standards Not applicable.
Conflict of Interest The authors declare that they have no conflict of
interest.
Consent to Participate Not applicable.
Consent for Publication An informed consent and a consent to publish
were obtained from each of the participants.
References
1. Gupta A, Madhavan MV, Sehgal K, Nair N, Mahajan S, Sehrawat
TS, Bikdeli B, Ahluwalia N et al (2020) Extrapulmonary
manifestations of COVID-19. Nat Med 26(7):1017–1032. https://
doi.org/10.1038/s41591-020-0968-3
2. Guan WJ, Ni ZY, Hu Y, Liang WH, Ou CQ, He JX, Liu L, Shan H
et al (2020) Clinical characteristics of coronavirus disease 2019 in
China. N Engl J Med 382(18):1708–1720. https://doi.org/10.1056/
NEJMoa2002032
3. Zhou F, Yu T, Du R, Fan G, Liu Y, Liu Z, Xiang J, Wang Y et al
(2020) Clinical course and risk factors for mortality of adult inpa-
tients with COVID-19 in Wuhan, China: a retrospective cohort
study. Lancet 395(10229):1054–1062. S0140-6736(20)30566-3.
https://doi.org/10.1016/S0140-6736(20)30566-3
4. Zhou P, Yang XL, Wang XG, Hu B, Zhang L, Zhang W, Si HR,
Zhu Y et al (2020) A pneumonia outbreak associated with a new
coronavirus of probable bat origin. Nature 579(7798):270–273.
https://doi.org/10.1038/s41586-020-2012-7
5. Khatoon F, Prasad K, Kumar V (2020) Neurological manifestations
of COVID-19: available evidences and a new paradigm. J
NeuroVirol 26(5):619–630 1–12
6. Paterson RW, Brown RL, Benjamin L, Nortley R, Wiethoff S,
Bharucha T, Jayaseelan DL, Kumar G et al (2020) The emerging
spectrum of COVID-19 neurology: clinical, radiological and labo-
ratory findings. Brain. 143:3104–3120. https://doi.org/10.1093/
brain/awaa240
7. Mao L, Jin H, Wang M, Hu Y, Chen S, He Q, Chang J, Hong C et al
(2020) Neurologic manifestations of hospitalized patients with co-
ronavirus disease 2019 in Wuhan, China. JAMA Neurol. https://
doi.org/10.1001/jamaneurol.2020.1127
8. Poyiadji N, Shahin G, Noujaim D, Stone M, Patel S, Griffith B
(2020) COVID-19–associated acute hemorrhagic necrotizing en-
cephalopathy: CT and MRI features. Radiology 296(2):E119–
E120 201187
9. Wu Y, Xu X, Chen Z, Duan J, Hashimoto K, Yang L, Liu C, Yang
C (2020) Nervous system involvement after infection with COVID-
19 and other coronaviruses. Brain Behav Immun 87:18–22. S0889-
1591(20)30357-3. https://doi.org/10.1016/j.bbi.2020.03.031
10. Cheng Q, Yang Y, Gao J (2020) Infectivity of human coronavirus
in the brain. EBioMedicine 56:102799. https://doi.org/10.1016/j.
ebiom.2020.102799
11. Herman C, Mayer K, Sarwal A (2020) Scoping review of preva-
lence of neurologic comorbidities in patients hospitalized for
COVID-19. Neurology 95(2):77–84. https://doi.org/10.1212/
WNL.0000000000009673
12. Beyrouti R, Adams ME, Benjamin L, Cohen H, Farmer SF, Goh
YY, Humphries F, Jager HR et al (2020) Characteristics of ischae-
mic stroke associated with COVID-19. J Neurol Neurosurg
Psychiatry 91(8):889–891. https://doi.org/10.1136/jnnp-2020-
323586
13. Toscano G, Palmerini F, Ravaglia S, Ruiz L, Invernizzi P, Cuzzoni
MG, Franciotta D, Baldanti F et al (2020) Guillain-Barre syndrome
associated with SARS-CoV-2. N Engl J Med 382(26):2574–2576.
https://doi.org/10.1056/NEJMc2009191
14. Solomon IH, Normandin E, Bhattacharyya S, Mukerji SS, Keller K,
Ali AS, Adams G, Hornick JL et al (2020) Neuropathological fea-
tures of Covid-19. N Engl J Med 383:989–992. https://doi.org/10.
1056/NEJMc2019373
15. Puelles VG, Lutgehetmann M, Lindenmeyer MT, Sperhake JP,
Wong MN, Allweiss L, Chilla S, Heinemann A et al (2020)
Multiorgan and renal tropism of SARS-CoV-2. N Engl J Med
383:590–592. https://doi.org/10.1056/NEJMc2011400
16. Zubair AS, McAlpine LS, Gardin T, Farhadian S, Kuruvilla DE,
Spudich S (2020) Neuropathogenesis and neurologic manifesta-
tions of the coronaviruses in the age of coronavirus disease 2019:
a review. JAMA Neurol. https://doi.org/10.1001/jamaneurol.2020.
2065
17. Mehta P, McAuley DF, Brown M, Sanchez E, Tattersall RS,
Manson JJ (2020) COVID-19: consider cytokine storm syndromes
Mol Neurobiol

and immunosuppression. Lancet 395(10229):1033–1034. https://
doi.org/10.1016/S0140-6736(20)30628-0
18. Ramani A, Müller L, Ostermann PN, Gabriel E, Abida-Islam P,
Müller-Schiffmann A, Mariappan A, Goureau O et al (2020)
SARS-CoV-2 targets cortical neurons of 3D human brain organoids
and shows neurodegeneration-like effects. bioRxiv. https://doi.org/
10.1101/2020.05.20.106575
19. Song E, Zhang C, Israelow B, Lu P, Weizman O-E, Liu F, Dai Y,
Louvi A et al (2020) Neuroinvasive potential of SARS-CoV-2 re-
vealed in a human brain organoid model. bioRxiv. https://doi.org/
10.1101/2020.06.25.169946
20. Goh KI, Cusick ME, Valle D, Childs B, Vidal M, Barabasi AL
(2007) The human disease network. Proc Natl Acad Sci U S A
104(21):8685–8690. https://doi.org/10.1073/pnas.0701361104
21. Zheng C, Xu R (2018) Large-scale mining disease comorbidity
relationships from post-market drug adverse events surveillance
data. BMC Bioinformatics 19(Suppl 17):500. https://doi.org/10.
1186/s12859-018-2468-8
22. Pfefferle S, Schopf J, Kogl M, Friedel CC, Muller MA, Carbajo-
Lozoya J, Stellberger T, von Dall'Armi E et al (2011) The SARS-
coronavirus-host interactome: identification of cyclophilins as tar-
get for pan-coronavirus inhibitors. PLoS Pathog 7(10):e1002331.
https://doi.org/10.1371/journal.ppat.1002331
23. Gulbahce N, Yan H, Dricot A, Padi M, Byrdsong D,Franchi R, Lee
DS, Rozenblatt-Rosen O et al (2012) Viral perturbations of host
networks reflect disease etiology. PLoS Comput Biol 8(6):
e1002531. https://doi.org/10.1371/journal.pcbi.1002531
24. Shapira SD, Gat-Viks I, Shum BO, Dricot A, de Grace MM, Wu L,
Gupta PB, Hao T et al (2009) A physical and regulatory map of
host-influenza interactions reveals pathways in H1N1 infection.
Cell 139(7):1255–1267. https://doi.org/10.1016/j.cell.2009.12.018
25. Ackerman EE, Kawakami E, Katoh M, Watanabe T, Watanabe S,
Tomita Y, Lopes TJ, Matsuoka Y et al (2018) Network-guided
discovery of influenza virus replication host factors. mBio 9(6):
e02002–e02018. https://doi.org/10.1128/mBio.02002-18
26. Prasad K, Khatoon F, Rashid S, Ali N, AlAsmari AF, Ahmed MZ,
Alqahtani AS, Alqahtani MS et al (2020) Targeting hub genes and
pathways of innate immune response in COVID-19: a network
biology perspective. Int J Biol Macromol. https://doi.org/10.1016/
j.ijbiomac.2020.06.228
27. Gordon DE, Jang GM, Bouhaddou M, Xu J, Obernier K, White
KM, O'Meara MJ, Rezelj VV et al (2020) A SARS-CoV-2 protein
interaction map reveals targets for drug repurposing. Nature. 583:
459–468. https://doi.org/10.1038/s41586-020-2286-9
28. Basha O, Barshir R, Sharon M, Lerman E, Kirson BF, Hekselman I,
Yeger-Lotem E (2017) The TissueNet v.2 database: a quantitative
view of protein-protein interactions across human tissues. Nucleic
Acids Res 45(D1):D427–D431. https://doi.org/10.1093/nar/
gkw1088
29. Gokhman D, Kelman G, Amartely A, Gershon G, Tsur S, Carmel L
(2017) Gene ORGANizer: linking genes to the organs they affect.
Nucleic Acids Res 45(W1):W138–W145
30. Doncheva NT, Assenov Y, Domingues FS, Albrecht M (2012)
Topological analysis and interactive visualization of biological net-
works and protein structures. Nat Protoc 7(4):670–685
31. Vella D, Marini S, Vitali F, Di Silvestre D, Mauri G, Bellazzi R
(2018) MTGO: PPI network analysis via topological and functional
module identification. Sci Rep 8(1):1–13
32. Barabasi A-L, Oltvai ZN (2004) Network biology: understanding
the cell’s functional organization. Nat Rev Genet 5(2):101–113
33. Pinkert S, Schultz J, Reichardt J (2010) Protein interaction
networks—more than mere modules. PLoS Comput Biol 6(1):
e1000659
34. Spirin V, Mirny LA (2003) Protein complexes and functional mod-
ules in molecular networks. Proc Natl Acad Sci 100(21):12123–
12128
35. Terentiev A, Moldogazieva N, Shaitan K (2009) Dynamic proteo-
mics in modeling of the living cell. Protein-protein interactions.
Biochem Mosc 74(13):1586–1607
36. Bader GD, Hogue CW (2003) An automated method for finding
molecular complexes in large protein interaction networks. BMC
bioinformatics 4(1):2
37. van Dam S, Vosa U, van der Graaf A, Franke L, de Magalhaes JP
(2018) Gene co-expression analysis for functional classification and
gene–disease predictions. Brief Bioinform 19(4):575–592
38. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D,
Amin N, Schwikowski B et al (2003) Cytoscape: a software envi-
ronment for integrated models of biomolecular interaction net-
works. Genome Res 13(11):2498–2504
39. Sherman BT, Lempicki RA (2009) Systematic and integrative anal-
ysis of large gene lists using DAVID bioinformatics resources. Nat
Protoc 4(1):44
40. Subramanian A, Narayan R, Corsello SM, Peck DD, Natoli TE, Lu
X, Gould J, Davis JF et al (2017) A next generation connectivity
map: L1000 platform and the first 1,000,000 profiles. Cell 171(6):
1437–1452 e1417. https://doi.org/10.1016/j.cell.2017.10.049
41. Gaulton A, Bellis LJ, Bento AP, Chambers J, Davies M, Hersey A,
Light Y, McGlinchey S et al (2012) ChEMBL: a large-scale bioac-
tivity database for drug discovery. Nucleic Acids Res 40(Database
issue):D1100–D1107. https://doi.org/10.1093/nar/gkr777
42. Cotto KC, Wagner AH, Feng YY, Kiwala S, Coffman AC, Spies G,
Wollam A, Spies NC et al (2018) DGIdb 3.0: a redesign and ex-
pansion of the drug-gene interaction database. Nucleic Acids Res
46(D1):D1068–D1073. https://doi.org/10.1093/nar/gkx1143
43. Kuhn M, von Mering C, Campillos M, Jensen LJ, Bork P (2008)
STITCH: interaction networks of chemicals and proteins. Nucleic
Acids Res 36(Database issue):D684–D688. https://doi.org/10.
1093/nar/gkm795
44. Huang HY, Lin YC, Li J, Huang KY, Shrestha S, Hong HC, Tang
Y, Chen YG et al (2020) miRTarBase 2020: updates to the exper-
imentally validated microRNA-target interaction database. Nucleic
Acids Res 48(D1):D148–D154. https://doi.org/10.1093/nar/gkz896
45. Watts DJ, Strogatz SH (1998) Collective dynamics of ‘small-
world’networks. Nature 393(6684):440–442
46. Groningen Uo Scale-free networks.
47. Hu T, Andrew AS, Karagas MR, Moore JH (2015) Functional
dyadicity and heterophilicity of gene-gene interactions in statistical
epistasis networks. BioData mining 8(1):1–11
48. Park J, Barabási A-L (2007) Distribution of node characteristics in
complex networks. Proc Natl Acad Sci 104(46):17916–17920
49. Wang Q, Liu W, Ning S, Ye J, Huang T, Li Y, Wang P, Shi H et al
(2012) Community of protein complexes impacts disease associa-
tion. Eur J Hum Genet 20(11):1162–1167
50. Hage P, Harary F (1995) Eccentricity and centrality in networks.
Soc Networks 17(1):57–63
51. Tripathi S, Moutari S, Dehmer M, Emmert-Streib F (2016)
Comparison of module detection algorithms in protein networks
and investigation of the biological meaning of predicted modules.
BMC bioinformatics 17(1):129
52. Wang Z, Zhang J (2007) In search of the biological significance of
modular structures in protein networks. PLoS Comput Biol 3(6):
e107
53. Barabási A-L, Gulbahce N, Loscalzo J (2011) Network medicine: a
network-based approach to human disease. Nat Rev Genet 12(1):
56–68
54. Gysi DM, Do Valle I, Zitnik M, Ameli A, Gan X, Varol O, Sanchez
H, Baron RM et al (2020) Network medicine framework for iden-
tifying drug repurposing opportunities for COVID-19. ArXiv
arXiv:2004.07229v1
55. Bosl K, Ianevski A, Than TT, Andersen PI, Kuivanen S, Teppor M,
Zusinaite E, Dumpis U et al (2019) Common nodes of virus-host
Mol Neurobiol

interaction revealed through an integrated network analysis. Front
Immunol 10:2186. https://doi.org/10.3389/fimmu.2019.02186
56. Zhou Y, Hou Y, Shen J, Huang Y, Martin W, Cheng F (2020)
Network-based drug repurposing for novel coronavirus 2019-
nCoV/SARS-CoV-2. Cell Discov 6:14. https://doi.org/10.1038/
s41421-020-0153-3
57. Kumar N, Mishra B, Mehmood A, Athar M, Mukhtar MS (2020)
Integrative network biology framework elucidates molecular mech-
anisms of SARS-CoV-2 pathogenesis. Science:SSRN 23(9):
101526. 3581857. https://doi.org/10.2139/ssrn.3581857
58. Guo L, Lin W, Zhang Y, Li W, Wang J (2019) BEST: a web server
for brain expression Spatio-temporal pattern analysis. BMC
Bioinformatics 20(1):632. https://doi.org/10.1186/s12859-019-
3222-6
59. Wishart DS, Feunang YD, Guo AC, Lo EJ, Marcu A, Grant JR,
Sajed T, Johnson D et al (2018) DrugBank 5.0: a major update to
the DrugBank database for 2018. Nucleic Acids Res 46(D1):
D1074–D1082
60. Davis AP, Grondin CJ, Johnson RJ, Sciaky D, McMorran R,
Wiegers J, Wiegers TC, Mattingly CJ (2019) The comparative
toxicogenomics database: update 2019. Nucleic Acids Res
47(D1):D948–D954
61. Nightingale S, Winston A, Letendre S, Michael BD, McArthur JC,
Khoo S, Solomon T (2014) Controversies in HIV-associated
neurocognitive disorders. Lancet Neurol 13(11):1139–1151.
S1474-4422(14)70137-1. https://doi.org/10.1016/S1474-4422(14)
70137-1
62. Liu C, Zhou Q, Li Y, Garner LV, Watkins SP, Carter LJ, Smoot J,
Gregg AC et al (2020) Research and development on therapeutic
agents and vaccines for COVID-19 and related human coronavirus
diseases. ACS Cent Sci 6(3):315–331. https://doi.org/10.1021/
acscentsci.0c00272
63. Farooqui A, Tazyeen S, Ahmed MM, Alam A, Ali S, Malik MZ,
Ali S, Ishrat R (2018) Assessment of the key regulatory genes and
their Interologs for Turner syndrome employing network approach.
Sci Rep 8(1):1–13
64. Raaben M, Posthuma CC, Verheije MH, te Lintelo EG, Kikkert M,
Drijfhout JW, Snijder EJ, Rottier PJ et al (2010) The ubiquitin-
proteasome system plays an important role during various stages
of the coronavirus infection cycle. J Virol 84(15):7869–7879.
https://doi.org/10.1128/JVI.00485-10
65. Tang Q, Wu P, Chen H, Li G (2018) Pleiotropic roles of the
ubiquitin-proteasome system during viral propagation. Life Sci
207:350–354. S0024-3205(18)30355-2. https://doi.org/10.1016/j.
lfs.2018.06.014
66. Lin D, Zhong B (2015) Regulation of cellular innate antiviral sig-
naling by ubiquitin modification. Acta Biochim Biophys Sin
Shanghai 47(3):149–155. https://doi.org/10.1093/abbs/gmu133
67. Longhitano L, Tibullo D, Giallongo C, Lazzarino G, Tartaglia N,
Galimberti S, Li Volti G, Palumbo GA et al (2020) Proteasome
inhibitors as a possible therapy for SARS-CoV-2. Int J Mol Sci
21(10):3622. https://doi.org/10.3390/ijms21103622
68. Karyopharm to evaluate low dose selinexor as a potential treatment
for hospitalized patients with COVID-19. [news release]. Newton,
MA: Karyopharm Therapeutics Inc.; April 7, 2020. (2020). https://
bitly/2VftoVw Accessed July15, 2020
69. Widman DG, Gornisiewicz S, Shacham S, Tamir S (2018) In vitro
toxicity and efficacy of verdinexor, an exportin 1 inhibitor, on op-
portunistic viruses affecting immunocompromised individuals.
PLoS One 13(10):e0200043. https://doi.org/10.1371/journal.pone.
0200043
70. Fernandez-Prado R, Carriazo-Julio SM, Torra R, Ortiz A, Perez-
Gomez MV (2019) MYH9-related disease: it does exist, may be
more frequent than you think and requires specific therapy. Clin
Kidney J 12(4):488–493. https://doi.org/10.1093/ckj/sfz103
71. Ban BH, Shah V (2017) MYH9-related thrombocytopenia. Mayo
Clin Proc 92(7):1169–1170. S0025-6196(17)30323-3. https://doi.
org/10.1016/j.mayocp.2017.05.003
72. Lippi G, Plebani M, Henry BM (2020) Thrombocytopenia is asso-
ciated with severe coronavirus disease 2019 (COVID-19) infec-
tions: a meta-analysis. Clin Chim Acta 506:145–148. S0009-
8981(20)30124-8. https://doi.org/10.1016/j.cca.2020.03.022
73. Xu P, Zhou Q, Xu J (2020) Mechanism of thrombocytopenia in
COVID-19 patients. Ann Hematol 99(6):1205–1208. https://doi.
org/10.1007/s00277-020-04019-0
74. Rauscher AA, Gyimesi M, Kovacs M, Malnasi-Csizmadia A
(2018) Targeting myosin by Blebbistatin derivatives: optimization
and pharmacological potential. Trends Biochem Sci 43(9):700–
713. S0968-0004(18)30123-3. https://doi.org/10.1016/j.tibs.2018.
06.006
75. Sanchez EG, Quintas A, Perez-Nunez D, Nogal M, Barroso S,
Carrascosa AL, Revilla Y (2012) African swine fever virus uses
macropinocytosis to enter host cells. PLoS Pathog 8(6):e1002754.
https://doi.org/10.1371/journal.ppat.1002754
76. Kumakura M, Kawaguchi A, Nagata K (2015) Actin-myosin net-
work is required for proper assembly of influenza virus particles.
Virology 476:141–150. S0042-6822(14)00553-4. https://doi.org/
10.1016/j.virol.2014.12.016
77. Wang X, Dhindsa R, Povysil G, Zoghbi A, Motelow J, Hostyk J,
Goldstein D (2020) Transcriptional inhibition of host viral entry
proteins as a therapeutic strategy for SARS-CoV-2. Preprints.
https://doi.org/10.20944/preprints202003.0360.v1
78. Fulzele S, Sahay B, Yusufu I, Lee TJ, Sharma A, Kolhe R, Isales
CM (2020) COVID-19 virulence in aged patients might be impact-
ed by the host cellular MicroRNAs abundance/profile. Aging Dis
11(3):509–522. https://doi.org/10.14336/AD.2020.0428
79. Bartoszewski R, Dabrowski M, Jakiela B, Matalon S, Harrod KS,
Sanak M, Collawn JF (2020) SARS-CoV-2 may regulate cellular
responses through depletion of specific host miRNAs. Am J Phys
Lung Cell Mol Phys 319(3):L444–L455. https://doi.org/10.1152/
ajplung.00252.2020
80. Fusco R, Scuto M, Cordaro M, D'Amico R, Gugliandolo E,
Siracusa R, Peritore AF, Crupi R et al (2019) N-
palmitoylethanolamide-oxazoline protects against middle cerebral
artery occlusion injury in diabetic rats by regulating the SIRT1
pathway. Int J Mol Sci 20(19). https://doi.org/10.3390/
ijms20194845
81. Brunetti G, Di Rosa G, Scuto M, Leri M, Stefani M, Schmitz-
Linneweber C, Calabrese V, Saul N (2020) Healthspan mainte-
nance and prevention of Parkinson’s-like phenotypes with
hydroxytyrosol and oleuropein aglycone in C. elegans. Int J Mol
Sci 21(7):2588. https://doi.org/10.3390/ijms21072588
82. Calabrese EJ, Calabrese V, Tsatsakis A, Giordano JJ (2020)
Hormesis and Ginkgo biloba (GB): Numerous biological effects
of GB are mediated via hormesis. Ageing Res Rev 64:101019.
https://doi.org/10.1016/j.arr.2020.101019
Publisher’sNoteSpringer Nature remains neutral with regard to jurisdic-
tional claims in published maps and institutional affiliations.
Mol Neurobiol