Content uploaded by George Vavougios
Author content
All content in this area was uploaded by George Vavougios on Aug 21, 2021
Content may be subject to copyright.
Journal Pre-proof
FYN, SARS-CoV-2, and IFITM3 in the neurobiology of Alzheimer’s
disease
George D. Vavougios , Marianthi Breza , Theodore Mavridis ,
Karen Angeliki Krogfelt
PII: S2666-4593(21)00021-4
DOI: https://doi.org/10.1016/j.dscb.2021.100022
Reference: DSCB 100022
To appear in: Brain Disorders
Received date: 4 May 2021
Revised date: 15 August 2021
Accepted date: 16 August 2021
Please cite this article as: George D. Vavougios , Marianthi Breza , Theodore Mavridis ,
Karen Angeliki Krogfelt , FYN, SARS-CoV-2, and IFITM3 in the neurobiology of Alzheimer’s dis-
ease, Brain Disorders (2021), doi: https://doi.org/10.1016/j.dscb.2021.100022
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
©2021 Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
Highlights
In a recent study, IFITM3 was identified as a novel γ-secretase modulator.
Previous works from our group had identified IFITM3 networks in peripheral blood
and CNS tissue from AD patients.
In this study, we used independent datasets from the AD consensus study to validate
IFITM3-FYN networks.
We perform comparative transcriptomics with SARS-CoV-2 inducible gene
signatures, and detect significantly overlapping networks.
Our results indicate that the IFITM3-FYN endocytosis signal may mediate tau and
Αβ pathology in the setting of infection, including COVID-19.
Title Page
Title
FYN, SARS-CoV-2, and IFITM3 in the neurobiology of Alzheimer’s disease
Authors: George D. Vavougios1,2, Marianthi Breza3, Theodore Mavridis3, Karen
Angeliki Krogfelt4,5
Affiliations:
1Neuroimmunology Laboratory, Department of Neurology, Athens Naval Hospital,
P.C. 115 21, Athens, Greece
2Department of Computer Science and Telecommunications, University of Thessaly,
Papasiopoulou 2 – 4, P.C. 35 131 – Galaneika, Lamia, Greece
31st Department of Neurology, Eginition Hospital, Medical School, National and
Kapodistrian University of Athens, Athens, Greece.
4Department of Science and Environment, Roskilde University, Universitetsvej 1,
28A.1, DK-4000 Roskilde Denmark
5Molecular and Medical Biology, Roskilde University, Universitetsvej 1, 28A.1, DK-
4000 Roskilde Denmark
Corresponding Author
George D. Vavougios, M.D., Ph.D., email: dantevavougios@hotmail.com /
gvavougyios@uth.gr phone: +306936528439
Present Address
70 Deinokratous Street, Athens, Greece
Conflict of Interest Statement: None declared.
Acknowledgements: To Dr Sofia Nikou, for timely revisions on this manuscript. To
the late professor Robert Moir, a pioneer and visionary whose work inspired our own.
Author contributions: Conceptualization, analyses, first draft, final draft, revisions:
GDV. Final draft, confirmation of analyses: TM, MB; Revisions, conceptualization:
KAK. All authors have read and approved of the final draft. TM and MB contributed
equally.
Abstract
Introduction: (IFITM3) is an innate immune protein that has been identified as a
novel γ-secretase (γs) modulator. FYN is a kinase that stabilizes IFITM3 on the
membrane, primes APP for amyloidogenic γs processing and mediates tau
oligomerization. The purpose of this study is to explore the role of FYN and IFITM3
in AD and COVID-19, expanding on previous research from our group.
Methods: A 520 gene signature containing FYN and IFITM3 (termed Ia) was
extracted from a previously published meta-analysis of Alzheimer’s disease (AD)
bulk- and single nuclei sequencing data. Exploratory analyses involved meta-analysis
of bulk and single cell RNA data for IFITM3 and FYN differential expression per
CNS site and cellular type. Confirmatory analyses, gene set enrichment analysis
(GSEA) on Ia was performed to detect overlapping enriched biological networks
between COVID-19 with AD.
Results: Bulk RNA data analysis revealed that IFITM3 and FYN were overexpressed
in two CNS regions in AD vs. Controls: the temporal cortex Wilcoxon p-value=1.3e-
6) and the parahippocampal cortex Wilcoxon p-value=0.012). Correspondingly, single
cell RNA analysis of IFITM3 and FYN revealed that it was differentially expressed in
neurons, glial and endothelial cells donated b AD patients, when compared to
controls.
Discussion: IFITM3 and FYN were found as interactors within biological networks
overlapping between AD and SARS-CoV-2 infection. Within the context of SARS-
CoV-2 induced tau aggregation and interactions between tau and Ab1-42, the FYN –
IFITM3 regulome may outline an important innate immunity element responsive to
viral infection and IFN-I signalling in both AD and COVID-19.
KEYWORDS: Alzheimer’s Disease; Antimicrobial Protection Hypothesis;
COVID-19; FYN Kinase; IFITM3; Type I Interferon; Innate immunity
MAIN MANUSCRIPT
1. Introduction
Interferon-induced transmembrane protein 3 (IFITM3) belongs to a family of proteins
that act as a second line of defense against enveloped viruses, including SARS-CoV-
2. IFITM3’s role in intercepting and shuttling viral particles to the lysosomes1 was
recently complemented by its discovery as a novel γ-secretase modulator that
promotes Αβ production2. Considering the accumulating evidence on common
pathways between COVID-19 and Alzheimer’s disease (AD)3, we aimed to examine
whether FYN, a kinase regulating IFITM3’s membrane localization4,5 is accordingly
perturbed in both AD and COVID-19 transcriptomes. Current state of the art
transcriptomic evidence suggest that FYN interacts with SARS-CoV-2 during the
course of infection6, and was found to be upregulated in a recent meta-analysis of
SARS-CoV-2 expression datasets7. Expanding on our previous research on IFITM3
networks in AD8 and their overlap with COVID-199, we propose a comprehensive
model of AD pathogenesis where viral induction of the IFITM3/FYN endocytosis
signal could account for increased Aβ oligomerization via γ-secretase activation2.
Furthermore, SARS-CoV-2 induced FYN dysregulation / overactivation6,7 would
concomitantly and independently promote Tau fibrilization10, abrogate autophagy4,
and prepare APP11 for processing by the previously IFITM32-activated γ-secretase
complex.
In order to explore FYN and IFITM3’s expression patterns and networks in AD vs.
COVID-19, we applied a composite systems biology approach12. Gene expression
data from both bulk tissue and single cell RNA sequencing studies were used in order
to explore FYN and IFITM3’s expression patterns in CNS cites and cells beyond
those examined by previous research from our group8,9. Subsequently, we aimed to
investigate the overlap between FYN/IFITM3’s biological networks and SARS-CoV-
2 infectomics. Finally, we provide data on FYN/IFITM3 networks that arose in our
previous study and integrated them in a comprehensive model of AD and AD-like
manifestations of NeuroCOVID-19’s pathogenesis.
2. METHODS
2.1 Data acquisition
For this study, we utilized consensus gene module and differential gene expression
data on IFITM3 and FYN from a previously published 13 of AD brain transcriptomes.
These datasets included data generated by the Accelerating Medicines Partnership
Alzheimer’s Disease Project (AMP-AD)14as well as publicly available datasets15-17;
The AMP-AD datasets included RNA-seq provided by the Mayo Clinic Brain Bank18,
the Religious Orders Study and Memory and Aging Project (ROSMAP) 19 and the
Mount Sinai School of Medicine (MSSM)20 studies.
Consensus gene expression modules represent gene co-expression networks detected
via consensus Weighted Gene Co-expression Network Analysis (cWGCNA)21. These
networks contain highly correlated genes that are conserved across studies, and are
furthermore associated with specific sample traits, such as cell type and diagnosis, and
combined with other sources of data, such as single nuclei RNA-seq experiments and
genome wide association studies (GWAS) . For both genes of interest, module
membership and the eigengene-based connectivity (kME ; range of values: -1 to 1 ) are
reported according to primary data.
2.2 Reconstruction of the FYN – IFITM3 interactome and comparative
transcriptomics with COVID-19 datasets
Consensus gene modules containing FYN and IFITM3 were considered as candidate
interactomes, visualized via STRING (available from: https://string-db.org)22 and
used for comparative gene set enrichment analyses (GSEA) via Enrichr23 (Available
from: (Available from: https://maayanlab.cloud/Enrichr/). Enrichment analyses aimed
both to (a) determine the biological functions and pathways associated with
FYN/IFITM3 signatures and (b) determine overlap with gene signatures extracted
from COVID-19 datasets. For GSEA, adjusted p-values <0.05 were considered
statistically significant.
2.3 Determination of IFITM3-FYN’s cell- and CNS site- associations
Consensus gene modules represent gene co-expression networks that are conserved
across datasets, CNS regions and enriched for specific states (i.e. upregulated in AD)
and cell types24. In order to determine whether FYN and IFITM3’s differential
expression was region specific, we inquired each gene’s comparative expression per
study and CNS region (accessible via: http://swaruplab.bio.uci.edu:3838/bulkRNA/
)13. Expression data were reported as reported as log2 transformed hub gene
expression, with unpaired Wilcoxon test p-values<0.05 determining differential gene
expression between AD vs. Controls.
While data on asymptomatic AD are also provided, we did not consider them in the
context of determining FYN / IFITM3 associations with CNS sites in AD.
Furthermore, is should be noted that differential gene expression were complimentary
to the extraction of the FYN / IFITM3 interactome, as detailed previously.
2.4 Determination of FYN and IFITM3’s cell- and disease- associations in AD
via single cell RNA transcriptomics
For single-cell expression studies, the scREAD database (Available from:
https://bmbls.bmi.osumc.edu/scread/) was interrogated, to further characterize FYN
and IFITM3’s expression in AD-donated cells and associated CNS regions 10,11. The
scREAD platform is a unique database compiling single cell RNA sequencing
(scRNA-Seq) data from AD studies along with matched control atlases. Furthermore,
it provides differential gene expression data that account for gender, brain region, and
age. Herein, we screened gene expression data from human datasets, and examined all
cross-dataset, disease – control comparisons. For all comparisons, a gene was
considered as differentially expressed when the the independent samples Wilcoxon p-
values <0.05 (Bonferroni adjusted) and the absolute value of the logFC in the single
cell resolution was >0.2525.
2.5 Data Availability Statement
Primary data are available from their respective sources, as cited herein. All data
meta-data generated by the analyses in this manuscript are available online via
Mendeley Data (Available from: https://data.mendeley.com/datasets/5bypp2h5kj/1 )
and as supplementary files.
3. RESULTS
3.1 Reconstruction of the FYN – IFITM3 interactome, its associations with
AD and cell types
Both FYN (kME = 0.87) and IFITM3 (kME= 0.77) were detected in a consensus module
(CM9; ngenes=520) that was positively correlated with AD diagnosis, significantly
enriched in astrocytes and endothelial cells, and methylation quantitative trait loci
(mQTL)13. This FYN/IFITM3 containing module was used as a candidate interactome
(henceforth dubbed Ia ; Figure 1) and was subsequently used in GSEA).
Analysis of bulk RNA data revealed that IFITM3 was overexpressed in two regions in
the discovery datasets: the temporal cortex (TC; Mayo Clinic Study, AD vs. Controls,
Wilcoxon test p-value=1.3e-6) and the parahippocampal gyrus (PHCG; MSSM study,
AD vs. controls, Wilcoxon test p-value=0.012); (Supplementary Fig.1a). In the
validation datasets, IFITM3 was differentially expressed in Zhang et al’s study9 (AD
vs. Controls, Wilcoxon test p-value<2.2e-6; Supplementary Fig. 1b). FYN was
differentially expressed in the temporal cortex (Mayo Clinic Study, AD vs. Controls,
adj. p-value=2.9e-5), the prefrontal cortex (PFC; ROSMAP Study, AD vs. Controls,
Wilcoxon test p-value=0.004) and the parahippocampal gyrus (MSSM study, AD vs.
controls, Wilcoxon test p-value=2.9e-5); Supplementary Fig. 1c. In the validation
datasets, FYN was differentially expressed in two datasets, including Zhang et al’s
study13 (AD vs. Controls, Wilcoxon test p-value<2.2e-16; Supplementary Fig 1d).
Taken together, these results indicate that the overexpression of FYN and IFITM3 can
be localized to the PFC, the TC and the PHCG (Figure 1 and Supplementary
Materials 1a-d).
3.2 FYN and IFITM3’s cell- and disease- associations in AD via single cell
RNA transcriptomics
Cross-dataset comparisons of the scREAD database revealed that IFITM3 and FYN
were differentially expressed in AD when compared to controls (adj. p-value<0.05) in
a cell-type and CNS site-specific manner
Specifically, that IFITM3 was overexpressed (a) in astrocytes, microglia and
endothelial cells in the entorhinal cortex datasets (b) underexpressed astrocytes,
microglia and endothelial cells and oligodendrocyte precursor cells in the prefrontal
cortex datasets(d) overexpressed in microglia the superior parietal lobe datasets
AD01203, AD01204 and underexpressed in microglia from the superior parietal lobe
dataset AD01205 (Supplementary Materials 2a,b).
FYN was differentially expressed in astrocytes, microglia, oligodentrocytes,
oligodendrocyte precursors, endothelial cells, excitatory and inhibitory neurons in the
entorhinal cortex, prefrontal cortex and the superior frontal gyrus. No clear pattern of
differential expression (ubiquitously up or down) could be established between cell
types or CNS regions, indicating significant cell-to-cell variability25 (Supplementary
Materials 2a,b).
3.3 Biological pathways and functions associated with FYN / IFITM3 and
comparative transcriptomics with COVID-19 datasets.
GSEA revealed several distinct, significantly enriched pathways and ontologies for
FYN and IFITM3 (Supplementary Materials 3). Both FYN and IFITM3 were
detected in the following significantly enriched pathways / ontologies:
“Cytokine Signaling in Immune system Homo sapiens” (Reactome Identifier: R-
HSA-1280215; adjusted p-value=5.98e-04), “Immune System Homo Sapiens”
(Reactome Identifier: R-HSA-168256; adjusted p-value=0.00349), “Cytokine-
mediated signaling pathway” (Gene Ontology Identifier:0019221; adjusted p-
value=2.94 e-06).
Confirmatory GSEA of the Ia interactome indicated that FYN and IFITM3 biological
networks were significantly enriched in several COVID-19 datasets containing
SARS-CoV-2 upregulated genes (Supplementary Materials 3; adjusted p-
value<0.05). Table 1 presents the 10 first (by order of adjusted p-value) entries of
significantly enriched human COVID-19 datasets, out of a total of 447.
Finally, we include significantly enriched FYN signatures from entorhinal cortex
neurons containing neurofibrillary tangles and hippocampal neurons from our
previous study5 for comparisons (Supplementary Materials 4a-d). Notably, the
“cytokine-mediated signalling pathway” GO term, containing both FYN and
IFITM3,was significantly enriched in both Ia and our previous study (FDR<0.05).
4. DISCUSSION
In our study, we explored FYN and IFITM3’s expression patterns and networks in
AD vs. COVID-19 using a meta-analytical approach. We extracted FYN/IFITM3’s
putative interactome from a published consensus gene module in AD. GSEA revealed
that common networks involving both FYN and IFITM3 mediate cytokine signaling,
and immune processes as significantly enriched biological pathways and ontologies.
Furthermore, FYN/IFITM3’s interactome was significantly enriched in several human
ex vivo and in vitro COVID-19 datasets. Differential gene expression analysis of bulk
RNA-seq data indicated that IFITM3 was differentially expressed in non-neuronal
cells (glia and endothelial cells) in the temporal cortex, the prefrontal cortex, the
superior parietal lobe and the parahippocampal gyrus. FYN was differentially
expressed in glial, endothelial cells and neurons in the entorhinal cortex, the prefrontal
cortex and the superior frontal cortex, and its expression was characterized by high
cell-to-cell variability.
Based on our findings, we will review the state-of-the art regarding the role of FYN
and IFITM3 in the pathobiology of AD. Furthermore, we will discuss the role of
immune processes in tau and Aβ processing, and its potential perturbations. Finally,
we will synthesize our findings and the literature in a quasinfectious hypothesis on
AD pathogenesis that may be applicable to the newly emerging COVID-19 associated
dyscognitive disorder.
4.1 FYN / IFITM3 pathways in Alzheimer’s disease and innate immunity
The recent identification of IFITM3, an established antiviral factor as a gamma
secretase modulator2 has provided deeper foundations to the antimicrobial protection
hypothesis of AD pathogenesis26. Our focus on FYN is founded on its regulatory role
in stabilizing IFITM3 in the membrane via phosphorylation of a critical motif5,
enabling its potential interaction with the gamma secretase complex. Notably, FYN
has been shown to phosphorylate APP in AD neurons, a process that enhances its
amyloidogenic processing11. Notably, oligomeric Aβ can upregulate FYN by
interacting with PrPc-mGluR527 inducing FYN-mediated phosphorylation of NMDA
receptors28. Taken together, these studies indicate that FYN may regulate IFITM3 –
gamma secretase – APP amyloidogenic processing via phosphorylation, with Aβ-
PrPc-mGluR5 cascades providing a feed-forward loop at the autocrine and paracrine
milieu.
Aside from Αβ oligomerization, FYN has been shown to mediate tau aggregation in
vivo10, with process that involve both direct phosphorylation and activation of other
kinases such as GSK-3β via NMDAR activation27. Notably, GSK-3β activation can
lead to both FYN’s activation and altered subcellular localization, either (a) in the
plasma membrane, constituting a feedback loop GSK3-β-FYN-NMDAR-GSK-3β
feedback loop) or in other subcellular compartments, including the nucleus29,30.
Within the context of the antimicrobial protection hypothesis, both IFITM3 and FYN
offer unique insight in the interactions between viral modulations of the
transcriptome, innate immunity and AD pathology. We have previously shown that
IFITM3 gene signatures enriched for IFN-I signalling represent overlapping pathways
between AD and COVID-199. IFITM3’s role in SARS-CoV-2 has been shown to be
structure dependent. Specifically, mutational alterations of its YxxΦ motif, FYN’s
phosphorylation target, may maintain the balance between SARS-CoV-2’s restriction
or permission31. To our knowledge, we are the first to outline this functional
relationship in the literature. Furthermore, mutational ablation of this site indicates
that SARS-CoV-2-IFITM3 mediated entry may be mediated by hijacking of
IFITM3’s endocytic signal32, a concept that has recently been corroborated33.
FYN’s role in cytokine signalling is upstream compared to IFITM3’s, positively
regulating (i.e. proinflammatory) signal transduction following immune receptor
activation34. This modus operandi has been observed in microglia in animal models of
Parkinson’s disease35. Aside from its homeostatic role, FYN may be recruited in viral
processes that require phosphorylation36 and subversion of specific compartments
such as autophagosomes37. In SARS-CoV-2, FYN has been reported as a
differentially expressed gene via multiple omics approaches7,6 albeit the consequences
of its recruitment or perturbation have not been explored.
4.2 SARS-CoV-2-related neurocognitive deficits and their importance in AD
pathobiology: from phenotypes to genes
COVID-19 has been lately recognized as a spectrum, one that includes both
phenotypic and genomic overlap with neurodegenerative disease including
Alzheimer’s disease38. Among the more albeit easily underdiagnosed manifestations
are neurocognitive symptoms, including memory defects, even among those patients
recovered from mild disease39.
A study reporting on a 3-month follow-up of patients recovering from COVID-19
uncovered microstructural alterations in the entorhinal cortex, associated with
hyposmia, whereas memory loss was associated with hippocampal cortex
remodeling40. As a previous hypothesis from our group41, this concept has been
subsequently validated by neuropathological studies that have detected SARS-CoV-2
in the olfactory cortex and the hippocampi42,43. Notably, a primate model of of
SARS-CoV-2 neurotropism has provided further corroboration to the our previously
stated hypothesis44.
Aside from the phenotypical overlap, genetic and epigenetic mechanisms have been
shown to overlap between Alzheimer’s disease and COVID-19. Independent meta-
analyses of gene expression data have corroborated IFITM3 as a commonly perturbed
gene in both conditions45,46. In the setting of SARS-CoV-2 infection, single nuclei
RNA sequencing has identified FYN and IFITM3 perturbations spatially linked them
to blood-barrier endothelial and glial cells47. A link between SARS-CoV-2 and the
secretory phenotype of senescent endothelial cells has recently identified IFITM3
downregulation as a result of paracrine cellular communication48. Fisetin, a senolytic
that has been shown to improve cognition and soluble Αβ burden49, was used in the
previous study to upregulate IFITM3 and reduce senescent cell burden48.
Taken together, these studies outline known and interdependent connections of FYN
and IFITM3 with innate immunity, viral lifecycles and the pathophysiological
processes underlying AD.
4.3 Fitting SARS-CoV-2 in the Antimicrobial Protection Hypothesis of
Alzheimer’s Disease: a demi-infectious hypothesis
Several recent studies have bolstered the concepts underlying the antimicrobial
protection hypothesis26. Soluble Αβ has been shown to function as an opsonin-like
molecule, forming complexes with nucleic acids (both endogenous and viral) and
enhancing their recognition by glial, while upregulating IFITM3-containing IFN-I
signatures50. This non-specific mechanism of recognizing danger-associated
molecular patterns (DAMPs) represents a sentinel aspect of innate immunity that is
fundamentally perturbed in the setting of AD51. DAMP dysregulation in AD
furthermore extends to other pattern-recognition receptors (PRR), such retinoic acid-
inducible gene-I (RIG-1), and feeds back to Αβ production52.
Recognition of invading viruses, including SARS-CoV-2, by RIG-I/MDA5 is
followed by an IRF3 mediated induction of IFN-β. Autocrine and paracrine IFN-β
induction would upregulate IFITM3 and enhance its trafficking in endosomes and the
membrane53. While this is physiologically this represents an antiviral mechanism,
SARS-CoV-2’s ORF3a has been shown to block lysosomal fusion and may thus
escape into the cytosol54, 55.
furthermore, FYN-mediated enhancement of STAT3 cascades would enhance this
feedback loop and maintain proinflammatory signalling56. Up to this point, viral
invasion, global and cell-level IFN-I signalling represents a common niche between
AD and COVID-1950,57. FYN dysregulation at this point may contribute to global
IFN-I perturbations via proinflammatory signalling34 in peripheral immune cells58,59.
Alternatively, FYN has been shown to enhance neuroinflammation in AD,
Parkinson’s disease dementia (PDD) and Dementia with Lewy bodies (DLB) by
maintaining proinflammatory signalling microglia, associated with increased Αβ and
pTau and dysproteostasis60. This paradigm of peripheral and CNS immune
perturbations in AD has been recently outlined in an extended IFITM3 gene signature
including OAS1, and overlapping with transcriptome perturbations in COVID-1961.
Beyond SARS-CoV-2, interferon-related genes such as OAS162 and IFITM32 and
downstream signal modulators such as FYN34 are key players in both AD and
susceptibility to wide variety of such as HIV-1 63, WNV64 and DENV65.
IFITM3’s recent addition as a gamma secretase modulator2 further raises the question
of context, in its interaction with a HSV-166 and HSV-1’s rapid seeding by Αβ as a
potentially protective mechanism67. Notably, in silico analyses have indicated that the
previously mentioned model can be realized by SARS-CoV-2; it’s S1 has high
affinity for soluble Αβ and tau 68. This novel interaction has been shown to be isoform
specific, with Αβ1-42 enhancing rather abrogating infectivity by modulating ACE2-S1
interactions69.
Within this context, FYN dysregulation can link Αβ accumulation, Tau
hyperphosphorylation and innate immunity in the setting of COVID-19. SARS-CoV-
2 infection experiments with brain organoids have recently revealed that
neuroinvasion is followed by altered neuronal distribution of hyperphosphorylated
Tau70. Furthermore, data from SARS-CoV-2 infectomics have consistently outlined
the induction of the tau kinase pathway and impaired autophagy following SARS-
CoV-2 neuroinvasion71. Via both direct and indirect mechanisms, FYN perturbations
could account for SARS-CoV-2 -induced tau aggregation10. FYN-mediated
stabilization of IFITM3 on the membrane would serve to prime its interactions with
gamma secretase2 along with amyloidogenic APP phosphorylation11, and abort viral
entry via endosomal escape31. FYN-enhanced Aβ-autocrine signalling would promote
tau toxicity28,60, and prepare an Aβ-enriched antiviral milieu that would intercept
incoming viruses via opsonin-like interactions 71 that prime microglia and provide
feedback to IFN-I and IFITM3 specifically50.
How would this local immune stimulation translate into a network-expanding
neurodegenerative disease? Brain endothelial cells and glia cells could be the prime
culprits in SARS-CoV-2, a concept supported by both our current findings and
others72. Neuroinvasion and transolfactory spread of SARS-CoV-2 to the hippocampi,
supported by the clinicoradiological course of COVID-19 dyscognitive syndromes40
and neuropathology42,44 would then represent the clinically evident stage. As we have
previously proposed, at the neuroinvasion stage, a primary hub such as the olfactory
cortex could inform its distal connectome of an invader and glia via the efflux of Αβ
and Tau9. In the setting of the neuroimmune hypothesis, both these molecules would
have to be reconsidered as a novel class of immune mediator that functions as a
“blind” guardian in innate immunity signalling71, following DAMP recognition by
PPRs.
4.4 Limitations and context
Our current work should be considered within its limitations and context. Currently,
no study has evaluated the longitudinal development of pathologically proven AD
following exposure to COVID-19, and such a study is required to verify both our
findings and our hypotheses.
A study that combines Another important limitation is that other AD-related genes are
also implicated in COVID-19 infection, such as APOE73 and ACE274. While this
implication further bolsters SARS-CoV-2’s potential role in AD pathogenesis, genetic
variability and gene interactions between these genes and the FYN/IFITM3 switch
should be studied in detail, in order to elucidate their mechanistic effects. Cell-to-cell
variability, as it arose in our scRNA-seq scrutiny of AD vs. Controls data, is another
confounding factor in interpretating our results. This variability indicates the
departure from the concept of tissues as homogenates, and the recognition of their
heterogeneity. This paradigm has been acknowledge in microglia, along with is
potential contribution to AD75, and may furthermore indicate the need to account for
senescence and the PASP phenotype in general in tissues48. Another limitation in our
study is that the differential expression data from Morabito et al13 only provide the p-
values for pairwise comparisons and not raw data. As such, the direction of
differential expression is derived from studying the plots, rather than the output. As
both FYN and IFITM3 are a priori selected in a consensus gene module, this
limitation does not affect their implication in AD or the validity of the interactome.
5. CONCLUSIONS
FYN represents a known modulator of IFITM3’s stabilization in the membrane and its
endocytic pathway. In this analysis, we determine the function, cell-, CNS-site
specific context of their interaction within cytokine signaling in AD. Furthermore, we
are the first to propose a deminfectious model of AD pathogenesis that can be
initiated by SARS-CoV-2, building upon our previous research and contemporary
knowledge. Future studies should aim to link clinical, radiological and
neuropathological findings with genomics and basic science in order to fully
characterize both the emerging entity of COVID-19 dyscognitive syndromes, and
their implication in AD.
Declaration of interests
The authors declare that they have no known competing financial interests or personal
relationships that could have appeared to influence the work reported in this paper.
6. REFERENCES
1 Iadecola, C., Anrather, J. & Kamel, H. Effects of COVID-19 on the Nervous
System. Cell 183, 16-27.e11, doi:10.1016/j.cell.2020.08.028 (2020).
2 Hur, J.-Y. et al. The innate immunity protein IFITM3 modulates γ-secretase in
Alzheimer’s disease. Nature 586, 735-740, doi:10.1038/s41586-020-2681-2
(2020).
3 Rahman, M. A., Islam, K., Rahman, S. & Alamin, M. Neurobiochemical
Cross-talk Between COVID-19 and Alzheimer's Disease. Mol Neurobiol 58,
1017-1023, doi:10.1007/s12035-020-02177-w (2021).
4 Chesarino, N. M., McMichael, T. M., Hach, J. C. & Yount, J. S.
Phosphorylation of the antiviral protein interferon-inducible transmembrane
protein 3 (IFITM3) dually regulates its endocytosis and ubiquitination. J Biol
Chem 289, 11986-11992, doi:10.1074/jbc.M114.557694 (2014).
5 Jia, R. et al. The N-terminal region of IFITM3 modulates its antiviral activity
by regulating IFITM3 cellular localization. J Virol 86, 13697-13707,
doi:10.1128/jvi.01828-12 (2012).
6 Ahmed, S. S. S. J. et al. Regulatory Cross Talk Between SARS-CoV-2
Receptor Binding and Replication Machinery in the Human Host. Frontiers in
Physiology 11, 802 (2020).
7 Vastrad, B., Vastrad, C. & Tengli, A. Identification of potential mRNA panels
for severe acute respiratory syndrome coronavirus 2 (COVID-19) diagnosis
and treatment using microarray dataset and bioinformatics methods. 3 Biotech
10, 422, doi:10.1007/s13205-020-02406-y (2020).
8 Vavougios, G. D. et al. Double hit viral parasitism, polymicrobial CNS
residency and perturbed proteostasis in Alzheimer's disease: A data driven, in
silico analysis of gene expression data. Mol Immunol 127, 124-135,
doi:10.1016/j.molimm.2020.08.021 (2020).
9 Vavougios, G. D. et al. Outside-in induction of the IFITM3 trafficking system
by infections, including SARS-CoV-2, in the pathobiology of Alzheimer's
disease. Brain Behav Immun Health 14, 100243,
doi:10.1016/j.bbih.2021.100243 (2021).
10 Briner, A., Götz, J. & Polanco, J. C. Fyn Kinase Controls Tau Aggregation
In Vivo. Cell Rep 32, 108045, doi:10.1016/j.celrep.2020.108045 (2020).
11 Iannuzzi, F. et al. Fyn Tyrosine Kinase Elicits Amyloid Precursor Protein
Tyr682 Phosphorylation in Neurons from Alzheimer's Disease Patients. Cells
9, doi:10.3390/cells9081807 (2020).
12 Zella, D. et al. The importance of genomic analysis in cracking the
coronavirus pandemic. Expert Rev Mol Diagn 21, 547-562,
doi:10.1080/14737159.2021.1917998 (2021).
13 Morabito, S., Miyoshi, E., Michael, N. & Swarup, V. Integrative genomics
approach identifies conserved transcriptomic networks in Alzheimer’s disease.
Human Molecular Genetics 29, 2899-2919, doi:10.1093/hmg/ddaa182 (2020).
14 Hodes, R. J. & Buckholtz, N. Accelerating Medicines Partnership:
Alzheimer's Disease (AMP-AD) Knowledge Portal Aids Alzheimer's Drug
Discovery through Open Data Sharing. Expert Opin Ther Targets 20, 389-391,
doi:10.1517/14728222.2016.1135132 (2016).
15 Zhang, B. et al. Integrated systems approach identifies genetic nodes and
networks in late-onset Alzheimer's disease. Cell 153, 707-720,
doi:10.1016/j.cell.2013.03.030 (2013).
16 Berchtold, N. C. et al. Synaptic genes are extensively downregulated across
multiple brain regions in normal human aging and Alzheimer's disease.
Neurobiol Aging 34, 1653-1661, doi:10.1016/j.neurobiolaging.2012.11.024
(2013).
17 Webster, J. A. et al. Genetic control of human brain transcript expression in
Alzheimer disease. Am J Hum Genet 84, 445-458,
doi:10.1016/j.ajhg.2009.03.011 (2009).
18 Allen, M. et al. Human whole genome genotype and transcriptome data for
Alzheimer's and other neurodegenerative diseases. Sci Data 3, 160089-
160089, doi:10.1038/sdata.2016.89 (2016).
19 Mostafavi, S. et al. A molecular network of the aging human brain provides
insights into the pathology and cognitive decline of Alzheimer's disease. Nat
Neurosci 21, 811-819, doi:10.1038/s41593-018-0154-9 (2018).
20 Wang, M. et al. The Mount Sinai cohort of large-scale genomic,
transcriptomic and proteomic data in Alzheimer's disease. Sci Data 5, 180185,
doi:10.1038/sdata.2018.185 (2018).
21 Langfelder, P. & Horvath, S. WGCNA: an R package for weighted correlation
network analysis. BMC Bioinformatics 9, 559, doi:10.1186/1471-2105-9-559
(2008).
22 Szklarczyk, D. et al. The STRING database in 2021: customizable protein-
protein networks, and functional characterization of user-uploaded
gene/measurement sets. Nucleic Acids Res 49, D605-d612,
doi:10.1093/nar/gkaa1074 (2021).
23 Kuleshov, M. V. et al. Enrichr: a comprehensive gene set enrichment analysis
web server 2016 update. Nucleic Acids Res 44, W90-97,
doi:10.1093/nar/gkw377 (2016).
24 Swarup, V. et al. Identification of Conserved Proteomic Networks in
Neurodegenerative Dementia. Cell Reports 31, 107807,
doi:https://doi.org/10.1016/j.celrep.2020.107807 (2020).
25 Hoffman, J. A., Papas, B. N., Trotter, K. W. & Archer, T. K. Single-cell RNA
sequencing reveals a heterogeneous response to Glucocorticoids in breast
cancer cells. Communications Biology 3, 126, doi:10.1038/s42003-020-0837-0
(2020).
26 Moir, R. D., Lathe, R. & Tanzi, R. E. The antimicrobial protection hypothesis
of Alzheimer's disease. Alzheimers Dement 14, 1602-1614,
doi:10.1016/j.jalz.2018.06.3040 (2018).
27 Penke, B., Szűcs, M. & Bogár, F. Oligomerization and Conformational
Change Turn Monomeric β-Amyloid and Tau Proteins Toxic: Their Role in
Alzheimer's Pathogenesis. Molecules 25, doi:10.3390/molecules25071659
(2020).
28 Um, J. W. & Strittmatter, S. M. Amyloid-β induced signaling by cellular prion
protein and Fyn kinase in Alzheimer disease. Prion 7, 37-41,
doi:10.4161/pri.22212 (2013).
29 Mondragón-Rodríguez, S., Perry, G., Zhu, X. & Boehm, J. Amyloid Beta and
tau proteins as therapeutic targets for Alzheimer's disease treatment:
rethinking the current strategy. Int J Alzheimers Dis 2012, 630182,
doi:10.1155/2012/630182 (2012).
30 Pan, J. et al. Triptolide induces oxidative damage in NRK-52E cells through
facilitating Nrf2 degradation by ubiquitination via the GSK-3β/Fyn pathway.
Toxicol In Vitro 58, 187-194, doi:10.1016/j.tiv.2019.03.032 (2019).
31 Shi, G. et al. Opposing activities of IFITM proteins in SARS-CoV-2 infection.
Embo j 40, e106501, doi:10.15252/embj.2020106501 (2021).
32 Murgolo, N. et al. SARS-CoV-2 tropism, entry, replication, and propagation:
Considerations for drug discovery and development. PLoS Pathog 17,
e1009225, doi:10.1371/journal.ppat.1009225 (2021).
33 Prelli Bozzo, C. et al. IFITM proteins promote SARS-CoV-2 infection and are
targets for virus inhibition in vitro. Nature Communications 12, 4584,
doi:10.1038/s41467-021-24817-y (2021).
34 Mkaddem, S. B. et al. Lyn and Fyn function as molecular switches that control
immunoreceptors to direct homeostasis or inflammation. Nature
Communications 8, 246, doi:10.1038/s41467-017-00294-0 (2017).
35 Panicker, N. et al. Fyn Kinase Regulates Microglial Neuroinflammatory
Responses in Cell Culture and Animal Models of Parkinson's Disease. J
Neurosci 35, 10058-10077, doi:10.1523/jneurosci.0302-15.2015 (2015).
36 de Wispelaere, M., LaCroix, A. J. & Yang, P. L. The small molecules
AZD0530 and dasatinib inhibit dengue virus RNA replication via Fyn kinase.
J Virol 87, 7367-7381, doi:10.1128/jvi.00632-13 (2013).
37 Kumar, R., Agrawal, T., Khan, N. A., Nakayama, Y. & Medigeshi, G. R.
Identification and characterization of the role of c-terminal Src kinase in
dengue virus replication. Scientific Reports 6, 30490, doi:10.1038/srep30490
(2016).
38 de Erausquin, G. A. et al. The chronic neuropsychiatric sequelae of COVID-
19: The need for a prospective study of viral impact on brain functioning.
Alzheimers Dement 17, 1056-1065, doi:10.1002/alz.12255 (2021).
39 Woo, M. S. et al. Frequent neurocognitive deficits after recovery from mild
COVID-19. Brain Commun 2, fcaa205, doi:10.1093/braincomms/fcaa205
(2020).
40 Lu, Y. et al. Cerebral Micro-Structural Changes in COVID-19 Patients - An
MRI-based 3-month Follow-up Study. EClinicalMedicine 25, 100484,
doi:10.1016/j.eclinm.2020.100484 (2020).
41 Vavougios, G. D. Potentially irreversible olfactory and gustatory impairments
in COVID-19: Indolent vs. fulminant SARS-CoV-2 neuroinfection. Brain
Behav Immun 87, 107-108, doi:10.1016/j.bbi.2020.04.071 (2020).
42 Meinhardt, J. et al. Olfactory transmucosal SARS-CoV-2 invasion as a port of
central nervous system entry in individuals with COVID-19. Nature
Neuroscience 24, 168-175, doi:10.1038/s41593-020-00758-5 (2021).
43 Pacheco-Herrero, M. et al. Elucidating the Neuropathologic Mechanisms of
SARS-CoV-2 Infection. Frontiers in Neurology 12, 444 (2021).
44 Jiao, L. et al. The olfactory route is a potential way for SARS-CoV-2 to
invade the central nervous system of rhesus monkeys. Signal Transduction
and Targeted Therapy 6, 169, doi:10.1038/s41392-021-00591-7 (2021).
45 Zhou, Y. et al. Network medicine links SARS-CoV-2/COVID-19 infection to
brain microvascular injury and neuroinflammation in dementia-like cognitive
impairment. Alzheimer's Research & Therapy 13, 110, doi:10.1186/s13195-
021-00850-3 (2021).
46 Blanco-Melo, D. et al. Imbalanced Host Response to SARS-CoV-2 Drives
Development of COVID-19. Cell 181, 1036-1045.e1039,
doi:10.1016/j.cell.2020.04.026 (2020).
47 Yang, A. C. et al. Dysregulation of brain and choroid plexus cell types in
severe COVID-19. Nature 595, 565-571, doi:10.1038/s41586-021-03710-0
(2021).
48 Camell, C. D. et al. Senolytics reduce coronavirus-related mortality in old
mice. Science 373, eabe4832, doi:10.1126/science.abe4832 (2021).
49 Currais, A. et al. Modulation of p25 and inflammatory pathways by fisetin
maintains cognitive function in Alzheimer's disease transgenic mice. Aging
Cell 13, 379-390, doi:10.1111/acel.12185 (2014).
50 Roy, E. R. et al. Type I interferon response drives neuroinflammation and
synapse loss in Alzheimer disease. J Clin Invest 130, 1912-1930,
doi:10.1172/jci133737 (2020).
51 Chiarini, A., Armato, U., Hu, P. & Dal Prà, I. Danger-Sensing/Patten
Recognition Receptors and Neuroinflammation in Alzheimer's Disease. Int J
Mol Sci 21, doi:10.3390/ijms21239036 (2020).
52 de Rivero Vaccari, J. P. et al. RIG-1 receptor expression in the pathology of
Alzheimer's disease. J Neuroinflammation 11, 67, doi:10.1186/1742-2094-11-
67 (2014).
53 McMichael, T. M., Chemudupati, M. & Yount, J. S. A balancing act between
IFITM3 and IRF3. Cellular & Molecular Immunology 15, 873-874,
doi:10.1038/cmi.2017.18 (2018).
54 Zhang, Y. et al. The SARS-CoV-2 protein ORF3a inhibits fusion of
autophagosomes with lysosomes. Cell Discovery 7, 31, doi:10.1038/s41421-
021-00268-z (2021).
55 Bianchi, M., Borsetti, A., Ciccozzi, M. & Pascarella, S. SARS-Cov-2 ORF3a:
Mutability and function. Int J Biol Macromol 170, 820-826,
doi:10.1016/j.ijbiomac.2020.12.142 (2021).
56 Yu, J. et al. FYN promotes gastric cancer metastasis by activating STAT3-
mediated epithelial-mesenchymal transition. Transl Oncol 13, 100841,
doi:10.1016/j.tranon.2020.100841 (2020).
57 Lee, J. S. & Shin, E.-C. The type I interferon response in COVID-19:
implications for treatment. Nature Reviews Immunology 20, 585-586,
doi:10.1038/s41577-020-00429-3 (2020).
58 Dallari, S. et al. Src family kinases Fyn and Lyn are constitutively activated
and mediate plasmacytoid dendritic cell responses. Nature Communications 8,
14830, doi:10.1038/ncomms14830 (2017).
59 Byeon, S. E. et al. The Role of Src Kinase in Macrophage-Mediated
Inflammatory Responses. Mediators of Inflammation 2012, 512926,
doi:10.1155/2012/512926 (2012).
60 Low, C. Y. B. et al. Isoform-specific upregulation of FynT kinase expression
is associated with tauopathy and glial activation in Alzheimer's disease and
Lewy body dementias. Brain Pathol 31, 253-266, doi:10.1111/bpa.12917
(2021).
61 Magusali, N. et al. Genetic variability associated with
<em>OAS1</em> expression in myeloid cells increases the risk of
Alzheimer’s disease and severe COVID-19 outcomes. bioRxiv,
2021.2003.2016.435702, doi:10.1101/2021.03.16.435702 (2021).
62 Salih, D. A. et al. Genetic variability in response to amyloid beta deposition
influences Alzheimer’s disease risk. Brain Communications 1,
doi:10.1093/braincomms/fcz022 (2019).
63 Wang, Y. et al. The V3 Loop of HIV-1 Env Determines Viral Susceptibility to
IFITM3 Impairment of Viral Infectivity. J Virol 91, doi:10.1128/jvi.02441-16
(2017).
64 Lim, J. K. et al. Genetic variation in OAS1 is a risk factor for initial infection
with West Nile virus in man. PLoS Pathog 5, e1000321,
doi:10.1371/journal.ppat.1000321 (2009).
65 Li, M. Y. et al. Lyn kinase regulates egress of flaviviruses in autophagosome-
derived organelles. Nature Communications 11, 5189, doi:10.1038/s41467-
020-19028-w (2020).
66 Nicholl, M. J., Robinson, L. H. & Preston, C. M. Activation of cellular
interferon-responsive genes after infection of human cells with herpes simplex
virus type 1. J Gen Virol 81, 2215-2218, doi:10.1099/0022-1317-81-9-2215
(2000).
67 Eimer, W. A. et al. Alzheimer's Disease-Associated β-Amyloid Is Rapidly
Seeded by Herpesviridae to Protect against Brain Infection. Neuron 99, 56-
63.e53, doi:10.1016/j.neuron.2018.06.030 (2018).
68 Idrees, D. & Kumar, V. SARS-CoV-2 spike protein interactions with
amyloidogenic proteins: Potential clues to neurodegeneration. Biochem
Biophys Res Commun 554, 94-98, doi:10.1016/j.bbrc.2021.03.100 (2021).
69 Hsu, J. T. et al. The Effects of Aβ(1-42) Binding to the SARS-CoV-2 Spike
Protein S1 Subunit and Angiotensin-Converting Enzyme 2. Int J Mol Sci 22,
doi:10.3390/ijms22158226 (2021).
70 Ramani, A. et al. SARS-CoV-2 targets neurons of 3D human brain organoids.
The EMBO Journal 39, e106230,
doi:https://doi.org/10.15252/embj.2020106230 (2020).
71 Gordon, D. E. et al. Comparative host-coronavirus protein interaction
networks reveal pan-viral disease mechanisms. Science 370,
doi:10.1126/science.abe9403 (2020).
72 Zhou, Y. et al. Network medicine links SARS-CoV-2/COVID-19 infection to
brain microvascular injury and neuroinflammation in dementia-like cognitive
impairment. Alzheimers Res Ther 13, 110, doi:10.1186/s13195-021-00850-3
(2021).
73 Kuo, C. L. et al. APOE e4 Genotype Predicts Severe COVID-19 in the UK
Biobank Community Cohort. J Gerontol A Biol Sci Med Sci 75, 2231-2232,
doi:10.1093/gerona/glaa131 (2020).
74 Ni, W. et al. Role of angiotensin-converting enzyme 2 (ACE2) in COVID-19.
Critical Care 24, 422, doi:10.1186/s13054-020-03120-0 (2020).
75 Yang, H. S. et al. Natural genetic variation determines microglia heterogeneity
in wild-derived mouse models of Alzheimer's disease. Cell Rep 34, 108739,
doi:10.1016/j.celrep.2021.108739 (2021).
7. FIGURE LEGENDS
Figure 1. Representation of the FYN-IFITM3 interactome via STRING.
Figure 2. Schematic representation of central nervous sites associated with the
overexpression of FYN and IFITM3, i.e. the prefrontal cortex, the temporal cortex
and the parahippocampal gyrus.
Figure 3. IFITM3 and FYN in the setting of a deminfectious hypothesis of
Alzheimer’s disease pathogenesis, with SARS-CoV-2 as the pathogen. Lipid-raft
mediated SARS-CoV-2 endocytosis in lipid rafts leads to a SARS-CoV-2 – IFITM3
interaction and endosomal sequestration (1). Depending on the phosphorylation status
of IFITM3, the virus is either sequestered in IFITM3-enriched endosomes (2a) or
retained in the surface (2b). Successful clearance of an invading virus via endosome –
lysosome fusion leads to the recognition of viral fragments as DAMPs by PPRs such
as RIG-I / MDA-5, which mediate upregulate IFN-I genes. IFN-I responses can
function in an autocrine manner via their receptors (4a) and via STAT3 signaling,
provide positive feedback of IFITM3’s expression in IFN-I primed cells, or induce its
expression in naïve cells. Notably, FYN has been shown to promote STAT3 signaling
as part of proinflammatory priming of peripheral immune cells and microglia.
IFITM3 mediates IRF3 autosomal degradation (not shown here), introducing a
negative feedback loop in DAMP recognition. Viruses that escape endosomal
sequestration or capitalize mutant IFITM3 isoforms, as has been shown in SARS-
CoV-2, may escape and replicate productively (5a), recruiting host kinases. Tau
hyperphosphorylations may occur as a result of viral processes, culminating in
toxicity and oligomerization (6a). In the alternative scenario of IFITM3
phosphorylation, the latter becomes stabilized on the membrane, priming gamma
secretase (2b). Concomitant phosphorylation of APP by FYN completes
amyloidogenic priming (3b), and Aβ oligomer release in the paracrine milieu (4b).
Isoform-specific interactions between Aβ oligomers and SARS-CoV-2’s S1 have
been reported, and may alter its infectivity; Αβ1-42 in particular has been shown to
promote rather than abrogate S1-ACE2 interactions (5b), and facilitate canonical
SARS-CoV-2 entry (6b). Autocrine Aβ-PrPc interactions provide further feedback to
FYN via phosphorylation targets such as mGluR5 (3c). Physical interactions between
FYN and NMDAR link Αβ and tau pathology via the activation of multiple kinases
(indirectly, following calcium influx), the direct activation of GSK-3b, or FYN-
mediated tau oligomerization.
8. TABLES
Table 1. Significantly enriched COVID-19 datasets
Index
Name
P-
value
Adjusted
p-value
1
500 genes up-regulated by SARS-CoV-2 in
human Organoids cells from GSE154613
1.001e-
19
4.474e-17
2
Top 500 upregulated genes for SARS-CoV-2
infection in human sclera from GSE164073
1.477e-
18
3.301e-16
3
Top 500 up genes for SARS-CoV-2 infection 48
hpi in human alveolar organoids for GSE152586
9.499e-
18
1.061e-15
4
500 genes up-regulated by SARS-CoV-2 in
human Lung Organoid cells at 24 hpi from
GSE148697
6.699e-
17
4.991e-15
5
Top 500 upregulated genes for SARS-CoV-2
infection in human lung organoids from
GSE148697
6.699e-
17
4.991e-15
6
500 genes up-regulated by SARS-CoV-2 in
human Calu3 cells at 24h from GSE148729 s1
1.061e-
15
5.926e-14
Table 1. Significantly enriched COVID-19 datasets
Index
Name
P-
value
Adjusted
p-value
polyA
7
500 top upregulated genes from SARS-CoV-2
infection at 72 HPI from GSE157852
1.995e-
14
9.189e-13
8
SARS-CoV perturbation; 402 Up Genes from
GEN3VA; Human bronchial epithelial 2B4 cells;
Accession: GSE17400 Platform: GPL570; Entry
6
2.469e-
14
9.647e-13
9
500 genes up-regulated by SARS-CoV-2 in
human pancreatic organoids from GSE151803
3.327e-
13
7.435e-12
10
500 genes up-regulated by SARS-CoV-2 in
human pancreatic organoids from GSE151803
3.327e-
13
7.435e-12
