Recently, an increasing number of neurological symptoms are reported in different COVID-19 studies (Özdağ Acarli et al., 2020; Taquet et al., 2021). Among them, the most widely reported in humans are headache, dizziness, impaired consciousness, cognitive impairment, as well as motor disorders and neuropathic pain. Some studies suggest that the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infects the brain of COVID-19 patients. Especially, autopsies revealed SARS-CoV-2 antigens in the brain parenchyma and cortical neurons of some patients (Meinhardt et al., 2021; Song et al., 2021). The SARS-CoV-2 viral RNA has been detected in the substantia nigra. In animal studies, it is suggested that dopaminergic neurons, cortical neurons, microglia, and astrocytes are susceptible to SARS-CoV-2 infection (Chen S. et al., 2021; Ferren et al., 2021; Song et al., 2021). This infection may have a durable neurological impact (Harmer et al., 2002). Other studies established the hypothesis of a relation between a SARS-CoV-2 infection and some neurological diseases like Alzheimer’s, Parkinson’s, and dementia (Lev et al., 2003; Obulesu and Lakshmi, 2014).

Angiotensin-converting enzyme 2 (ACE2), a novel homologous of ACE expressed in renal and cardiovascular tissues, as well as the gastrointestinal system, was reported to be the main receptor for SARS-CoV-2 entry into cells (Netland et al., 2008). In the brain, several studies indicated that ACE2 was expressed in excitatory and inhibitory neurons, astrocytes, and microglia in the human and mouse brain (Chen R. et al., 2021; Mahalingam et al., 2021; Li et al., 2022). Baig et al. (2020) demonstrated that SARS-CoV-2 also uses ACE2 to target the central nervous system (CNS). Lu et al. (2020) analysed diffusion tensor imaging and 3D high-resolution T1WI sequences of several COVID-19 patients and concluded that significant structural changes happen in different regions, such as olfactory cortices, hippocampi, insula, and cingulate gyrus. These structural changes correspond to brain regions where the gene coding for ACE2 is expressed the most. Therefore, biological, cognitive, and sensorimotor functions that characterize the regions where ACE2 is overexpressed are likely to be affected by the COVID-19.

In addition to ACE2, other works (Hoffmann et al., 2020; Ren et al., 2020) have reported the transmembrane serine protease 2 (TMPRSS2) gene as another potential entry of SARS-CoV-2 in several tissues, such as kidney and lung tissue. For the brain, the potential entry ACE2/TMPRSS2 of the virus has been studied in mice and humans and has been reported by Fodoulian et al. (2020), in GSE151973. Furthermore, in addition to the role of ACE2 and TMPRSS2 as an entry of the virus through the nasal tissue to the brain, several studies have linked their brain distribution to certain neurological effects of COVID-19 (Boldrini et al., 2021; Hernández et al., 2021; Kulkarni et al., 2022).

Although the presence of SARS-CoV-2 in the CNS has been demonstrated, its mode of replication remains unclear. However, the results from modelling SARS-CoV-2 neuroinvasion and cell death using human brain organoids suggested that SARS-CoV-2 is able to use the machinery of neuronal cells to replicate (Song et al., 2021). In addition, Olivarria et al. (2022) revealed that an average of 50% of infected mice exhibited CNS infection characterised by widespread viral replication in neurons.

The aim of this study was to characterize the scope of the potential complex impact of a SARS-CoV-2 infection in the brain. It concerns different scales: the topographic, cognitive, sensorimotor, and genetic one. This characterisation concerns the cognitive and sensorimotor functions, but also the local overexpressed genes. This general understanding is crucial to establish the potential neurological manifestations of the infection (McFarland et al., 2021). Hypotheses were drawn using the genetic base of Allen Institute for Brain Science (AIBS) (Hawrylycz et al., 2012), and the topographical distance was provided by the LinkRbrain platform (Mesmoudi et al., 2015). LinkRbrain is an open-access web tool for multi-scale integration of knowledge on the brain.¹ It provides a visualisation of the literature about 460 cognitive and sensorimotor functions of each brain region and integrates this functional information with transcriptome data. The cognitive and sensorimotor database has been automatically extracted from the literature. AIBS atlas is used to cover 20,789 mRNA (messenger ribonucleic acid) expressions in the 947 human brain regions expressed by 947 Montreal Neurological Institute (MNI) coordinates. For each coordinate, the platform maps the cognitive and sensorimotor functions on a brain, as well as the expression scores of 20,789 mRNAs of genes. Based on its topographical calculation, LinkRbrain provides relational graphs of the functions and the genes. The final objective of this platform was to document the cognitive and sensorimotor functions and the genes of each brain coordinate thanks to the topographical similarities and co-expressions of mRNA.

In this work, we especially aimed to document ACE2/TMPRSS2 by identifying the cognitive and sensorimotor functions that activate, separately, the same regions where ACE2 and TMPRSS2 are overexpressed, as well as the other genes co-overexpressed with ACE2 and TMPRSS2. To do so, we first detected and mapped the regions where the mRNAs of ACE2 and TMPRSS2 are overexpressed. Then, based on the topographical distance calculation, we found the cognitive and sensorimotor functions that activate these regions. Finally, we calculate the topographical distance between a specific mRNA (of ACE2 and TMPRSS2) and the localised overexpression of the 20,789 identified mRNAs. This allowed us to identify the most expressed genes in the regions where ACE2 (or TMPRSS2) is overexpressed. The results of this topographical exploration are coherent with observed COVID-19 neurological symptoms. This method could generate new hypotheses to explore the neurological manifestations of COVID-19.

In this study, we executed several stages of the LinkRbrain process (Mesmoudi et al., 2015). First, we started by mapping the regions where the ACE2 mRNAs are overexpressed in the brain thanks to ABA H0351.2001. Then, we used the topographical distance calculated by LinkRbrain to identify the main brain functions activating the same regions where the ACE2 mRNAs are the most overexpressed. Finally, we identified the other mRNAs that could be co-overexpressed in the same regions of ACE2 mRNAs.

A correlation based on the topographical distance calculates the intersection between two groups of points. Each group represents brain functions or genes, and the points are the activations or AIBS coordinates. This intersection between the groups respects the normalisation of the group sizes; thus the higher correlation indicates a higher intersection between the groups.

The aim of this study was to characterize the scope of the potential complex impact of a SARS-CoV-2 infection in the brain. LinkRbrain contains three databases: Talairach atlas of topographical regions, 460 cognitive and sensorimotor functions, and mRNAs gene expression. Each of these databases consists of MNI coordinates related to one or several labels. Depending on the scale, the label(s) will be topographical, cognitive and sensorimotor, or genetic. We started by mapping the regions where ACE2 and MPRSS2 mRNAs are overexpressed thanks to the mRNA expression levels of genes provided by the AIBS. Then, thanks to LinkRbrain’s topographical distance calculation, we analysed the correlation between ACE2 and MPRSS2 overexpression, and either region-specific brain activation, function-specific brain activations, or regional overexpression of other genes. The results of this integration brought to light seven cognitive and sensorimotor categories and 11 genetic groups. The main regions specific to ACE2 and MPRSS2 that we found are localised in the brain stem, the subcortical, the orbitofrontal, and some occipital regions. This is coherent with the literature study of Chen R. et al. (2021) where the spatial distribution analysis demonstrated that the expression of ACE2 and MPRSS2 is relatively high in specific brain locations, such as thalamus, human middle temporal gyrus, posterior cingulate cortex, and hippocampal. Thanks to the correlation calculated by LinkRbrain, it was established that 88 cognitive and sensorimotor functions activate these regions.

The first group consists of 15 functions, all related to memory and recollection. This suggests that COVID-19 could lead to memory complications. In fact, some studies (Ragab et al., 2020; Soy et al., 2020) show that SARS-CoV-2 infection entails a cytokine storm, which can later generate cognitive and psychiatric effects (Panoskaltsis et al., 2021). Lu et al. (2020) suggest that memory and concentration problems noticed after the infection are related to the significant structural changes occurred in some specific brain regions. Patients endured decreased concentration, reduced memory, and difficulty retaining information, persisting for over 2 months. Several studies described cognitive impairment in patients infected by SARS-CoV-2 (Almeria et al., 2020; Alemanno et al., 2021). Alemanno et al. (2021) demonstrated that 80% of patients presented cognitive deficits, when Almeria et al. (2020) mentioned deterioration in attention, memory, and executive functions. The second group consists of 25 functions connected with motor functions. Meppiel et al. (2021) reported that some patients showed symptoms like motor or sensitive deficit and/or movement disorder. In addition, aggravation of specific motor functions has been reported for Parkinsonian patients (Sulzer et al., 2020). The “reward” group from our results supports the relation between some Parkinsonian symptoms, such as impulse control disorder and reward, especially monetary reward (Drew et al., 2020). Another group is represented by eight functions and concerns physical and mental suffering. Pain is established as an important symptom of COVID-19, both in the acute phase of the disease and at later stages. Potential mechanisms change nociceptor excitability, which are likely to develop pain, induce neuropathies, and worsen existing pain states (McFarland et al., 2021). Moreover, painful symptoms, as well as muscular, joint, chronic, and neuropathic pain, are clear clinical manifestations of COVID-19 infection (Kemp et al., 2020; Widyadharma et al., 2020). For the next group, 14 functions related to the lucidity filling have been grouped. The dysexecutive syndrome, with attention and orientation disorders, has been documented as a consequence of COVID-19 (Helms et al., 2020; Rogers et al., 2020; Sahoo et al., 2020). This syndrome is directly related to the listed functions “proprioceptive information,” “visuospatial task,” “alertness,” and “executive control.” Mcloughlin et al. (2020) reported an increased rate of delirium and neurological alteration in COVID-19 patients. Other studies described neurological manifestations, such as dizziness, altered consciousness, disorientation, and delirium (Rogers et al., 2020). The fifth group consists of 16 functions associated with emotional intensity and modulation. Literature reviews on coronaviruses (CoVs) reported that during acute illness, patients are suffering from depressed mood, anxiety disorders, psychosis, and irritability (Murray et al., 1992; Paniz-Mondolfi et al., 2020). Finally, cinq functions concerning sensory inputs like visual, auditory, and odor tasks have been found in our results, which are compatible with some reported SARS-CoV-2 symptoms (Helms et al., 2020; Rogers et al., 2020; Sahoo et al., 2020).

Thanks to its topographical distance correlation, the LinkRbrain platform listed the genes co-expressed with ACE2 that could be affected by SARS-CoV-2 infection. One cluster concerns olfactory receptors, but there are also genes in relation to cognitive functions like executive functions and neurodegenerative diseases like Alzheimer’s, Parkinson’s, and dementia. We identified eight genes coding for ORs, such as OR10H1, OR52B6, and OR8S1. Studies demonstrated that in the case of humans and mice (transgenic), the brain was a target organ for infection in SARS-CoV-2 (Netland et al., 2008; Meinhardt et al., 2021). Besides, these studies show that the virus could reach the brain by the olfactory bulb or by the olfactory tract of the CNS (Meinhardt et al., 2021). Furthermore, Kerslake et al. (2020) demonstrated that ORs facilitate the transport of the virus using the vagal nerve in the context of COVID-19. Butler et al. (2006) showed that in mice infected intranasally with neurovirulent strains of HCoV-OC43, virus enters the CNS via the olfactory nerves with subsequent transneuronal retrograde dissemination. Once the brain is infected by SARS-CoV-2, several neurological effects are reported (Özdağ Acarli et al., 2020; Taquet et al., 2021). It has been shown that other CoVs like human HCoV-OC43 or SARS-CoV were associated with neuronal complications, such as inflammation and encephalomyelitis (Paniz-Mondolfi et al., 2020). Groups that could be related to these neurological symptoms emerged in our results, such as inflammation, myelination, immunity, ischemia, and cancer genes. Nonetheless, Murray et al. (1992) demonstrated that murine coronavirus can cause direct lysis of oligodendrocytes and final demyelination in the human CNS. Poyiadji et al. (2020) computed tomography and MRI reports of hemorrhagic necrotising encephalopathy of a COVID-19 female. Other studies described several symptoms of inflammatory-induced nerve dysfunctions (Chen et al., 2020; Xydakis et al., 2020). Furthermore, Toscano et al. (2020) reported that several COVID-19 patients developed the autoimmune Guillain–Barré syndrome after a few days following the onset of SARS-CoV-2 infection. The other well-distinguished cluster is composed of genes related to neurodegenerative diseases, such as Alzheimer’s and Parkinson’s diseases. This group could be explained with the presence of genes related to demyelination and inflammation. In the literature, especially on transgenic mice (Sy et al., 2011), a direct causal relationship has been reported between HCoV-OC43, MHV, certain neuropathogenesis, and degeneration of specific cognitive functions. In addition, Arbour et al. (2000) reported that the genetic material of CoVs is co-localised with Aβ protein in the brain. It is likely that SARS-CoV-2 can induce the accumulation of Aβ protein and may lead to a close relationship between Alzheimer’s disease susceptibility and neuroinflammation due to SARS-CoV-2. Alzheimer’s disease does not appear to be the only one worsening neuroinflammation and decreasing cognitive function. According to our results and the literature, the hypothetical vulnerability of SARS-CoV-2 localised in basal ganglia and dopaminergic brain regions might increase Parkinsonian symptoms. This could be induced by a systemic failure of the dopamine synthesis pathway. It has been established that alterations in apoptosis-related signalling pathways are involved in Alzheimer’s and Parkinson’s diseases (Lev et al., 2003; Obulesu and Lakshmi, 2014). In our results, we found several genes related to cancer. In the literature, some of these genes, such as RBMY1A3P and TP53, are involved in apoptosis (Sutherland et al., 2005). Therefore, modulation of neuronal apoptosis by SARS-CoV-2 during latent infections could possibly be related to alterations in neuronal processes, leading to neuronal degeneration and brain damage (Patatian and Benech, 2020). Furthermore, other diseases seem to be correlated with SARS-CoV-2 like dementia. Taquet et al. (2021) released a meta-analysis on symptoms presented over 236,379 survivors of COVID-19 6 months after their contamination. It was noticed that dementia, anxiety disorder, and psychotic disorder are substantial psychiatric morbidity related to COVID-19 infection. Finally, studies found that other senses are perturbed: vision, hearing, and equilibrium (Sutherland et al., 2005; Helms et al., 2020). In our results, we found several genes related to these symptoms like AIPL1 and TMC2. Moreover, conjunctivitis has been a documented phenomenon in COVID-19, and reverse transcription polymerase chain reaction of conjunctiva and tears has tested positive for SARS-CoV-2 (Chen et al., 2020).

To compare our results with other brain data, we analysed the GSE164332 (Fullard et al., 2021) data that had focused on the frontal cortex. Intersection of a larger set of our results for ACE2/TMPRSS2 (Supplementary Table 4) with that of Gagliardi et al. (2021) (Fullard et al., 2021) provided us with five genes, namely, SLC14A1, HIF3A, RGS5, ZNF622 (instead of ZNF621), and HBA2. The HIF3A, RGS5, and HBA2 genes were validated by quantitative PCR (qPCR) method. In addition, Gagliardi et al. (2021) (Fullard et al., 2021) considered that these genes are the most deregulated (mRNA encoding the HBA hemoglobin subunits upregulated, and the SLC14A1, HIF3A, and RGS5 genes downregulated), and they are associated with hypoxia and inflammation.

Also, we found four common deregulated genes (INP5D, PRKCA, LRP1, and FCGR2A) in another study focused on the following brain regions: the dorsolateral prefrontal cortex, the medulla oblongata, and the choroid plexus GSE164485 (Gagliardi et al., 2021). We also found other representatives of the same family for others genes. For example, we obtained CD81 and CD82 instead of CD83. Similarly, we obtained DOCK1, an important paralog of DOCK2.

We were also interested in the pathways where our gene results are involved. Based on the exploration of GeneCard, the pathways that we found are implicated in the transport of glucose, glycosylation, innate immune system, GPCR, and also apoptosis, immune response IFN, interferon, and others. These results are confirmed with those obtained from the peripheral blood of convalescent people.

Furthermore, from our results, several genes appear as correlated to this infection as other works based on exploration of data from patient infected by SARS-CoV-2. From these genes, we note the following: CXCR6 (Kasela et al., 2021), CST9L⁴ (Yan et al., 2020), NOXO1 (Sindona et al., 2021), IL28B (Alothaid et al., 2020), FGF23 (George et al., 2021), and finally, the controverted vitamin D receptor VDR (Ceolin et al., 2021).

In this work, we attempted to relate the information processing performed with the cognitive and sensorimotor functions, and the genes revealed by their mRNA expression level. This type of methodology is used by Cioli et al. (2014) who concluded to a relation between the cognitive function and the genetic expression. Zhou et al. (2020) highlighted the executive functioning impairment, underlining the role of inflammatory processes. They reported a positive association between attentional deficits and inflammatory levels. In our results, both genetic and functional results are consistent. The altered functions on the movement and reward reflect the connection of certain genes with Parkinson’s disease. The functions related to memory and lucidity are directly correlated with inflammation, immunity, and neurodegenerative diseases, such as Alzheimer’s and dementia. Finally, it appears that sensory inputs are impacted, such as visual, auditory, and odor tasks. In our results, we found olfactory receptors and genes related to vision, hearing, and equilibrium disturbances.

Even though our results are coherent with observed COVID-19 neurological manifestations, and that some results are validated with infected people, the limit of this method remains in the nature of the cognitive and sensorimotor database and the lack of similar genetic data that cover the whole brain, from infected participant. The cognitive and sensorimotor database is composed with coordinates extracted from the literature. In order to complete this work, it would be necessary to conduct real experiments based on these results as they are not definitive and can only be used as a substantial field of research. This method could generate new hypotheses to explore the neurological manifestations of COVID-19. It presents results in an objective and global manner, based on existing knowledge produced by the scientific community. The technical and methodological contributions of LinkRbrain enable a functional and transcriptome characterisation of certain pathologies, including COVID-19. The creation of large open-source databases is part of the current big data scientific revolution. These techniques produce an intelligible synthesis of results and provide access to cutting-edge knowledge from related fields, which are often fundamental to successful research.

The original contributions presented in this study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

SM conceived and designed the experiments. SM, MR, and CL performed the experiments and analysed the data. All authors wrote the manuscript, contributed to the article, and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.