From Viral Infections to Alzheimer's Disease: Unveiling the Mechanistic Links Through Systems Bioinformatics

Abstract

Background

Emerging evidence suggests that viral infections may contribute to Alzheimer's disease (AD) onset and/or progression. However, the extent of their involvement and the mechanisms through which specific viruses increase AD susceptibility risk remain elusive.

Methods

We used an integrative systems bioinformatics approach to identify viral-mediated pathogenic mechanisms, by which Herpes Simplex Virus 1 (HSV-1), Human Cytomegalovirus (HCMV), Epstein-Barr virus (EBV), Kaposi Sarcoma-associated Herpesvirus (KSHV), Hepatitis B Virus (HBV), Hepatitis C Virus (HCV), Influenza A Virus (IAV) and Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) could facilitate AD pathogenesis via virus-host protein-protein interactions (PPIs). We also explored potential synergistic pathogenic effects resulting from herpesvirus reactivation (HSV-1, HCMV, and EBV) during acute SARS-CoV-2 infection, potentially increasing AD susceptibility.

Results

Herpesviridae members (HSV-1, EBV, KSHV, HCMV) impact AD-related processes like amyloid-β (Aβ) formation, neuronal death, and autophagy. Hepatitis viruses (HBV, HCV) influence processes crucial for cellular homeostasis and dysfunction, they also affect microglia activation via virus-host PPIs. Reactivation of HCMV during SARS-CoV-2 infection could potentially foster a lethal interplay of neurodegeneration, via synergistic pathogenic effects on AD-related processes like response to unfolded protein, regulation of autophagy, response to oxidative stress, and Aβ formation.

Conclusions

These findings underscore the complex link between viral infections and AD development. Viruses impact AD-related processes through shared and distinct mechanisms, potentially influencing variations in AD susceptibility.

herpesviruses, Alzheimer's disease, viral infections, hepatitis viruses, SARS-CoV-2

Alzheimer's disease (AD) is a progressive chronic neurodegenerative disease of the central nervous system characterized by memory impairment and cognitive decline [1]. The primary pathophysiological hallmarks of AD are the formation of amyloid plaques, neurofibrillary tangles, and neuronal loss in the brain [1]. Currently, there are no effective pharmacotherapies for its treatment [2]. It is the predominant form of dementia in older adults, affecting 55 million people worldwide [3]. AD is of multifactorial origin, involving the complex interaction of both genetic and environmental risk factors [4]. Viral infections have been hypothesized to be environmental factors that can increase susceptibility towards the development of AD [5]. Indeed, individuals seropositive for human Herpes Simplex Virus 1 (HSV-1) show increased risk of developing AD, and increased presence of HSV-1 DNA was found in AD postmortem brains [6–8]. The detection of infectious agents, such as HSV-1, in the brain of AD patients led to the “antimicrobial protection hypothesis of AD,” which suggests that an increased microbial burden in the brain might lead to the deposition of Aβ, due to its role in innate immune responses against pathogens, ultimately leading to sustained neuroinflammation that propagates neurodegeneration [9]. Several other viruses have also been linked to the development of AD, including Herpes Simplex Virus 2 (HSV-2), Human Cytomegalovirus (HCMV), Epstein-Barr Virus (EBV), Varicella-Zoster Virus (VZV), Human Herpesvirus 6A/B (HHV-6A/B), Human Herpesvirus 7 (HHV-7), Hepatitis C Virus (HCV), and Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) [6, 10–15]. Overall, the current evidence suggests that viral infections may increase the risk of developing AD and/or facilitate its progression. However, further investigation is warranted to understand the viral-mediated pathogenic mechanisms through which different viral species contribute to AD susceptibility.

Co-infection with 2 viruses can have synergistic pathogenic effects, potentially increasing susceptibility to AD development. For instance, reactivation of latent viruses may occur during acute infection with respiratory viruses. Evidence suggests that SARS-CoV-2 infection can trigger the reactivation of latent herpesviruses like EBV [16], HCMV [17], and HSV-1 [18]. This reactivation may contribute to Coronavirus Disease 2019 (COVID-19) severity and post–COVID-19 cognitive symptoms resembling AD, like memory impairment and brain fog [16, 19–22]. Our previous work, leveraging systems bioinformatics methodologies, also suggested an elevated risk of AD development following SARS-CoV-2 infection [23]. Hence, it is not surprising that there was an observed approximately 17% increase in AD-related deaths in 2020, possibly associated with COVID-19 infections [24]. Thus, it is important to understand the mechanisms by which the synergistic action of 2 viruses could potentially increase the risk of AD.

Computational approaches, especially network-based methodologies, have been widely employed to shed light on microbe-host interactions underlying the onset of an array of diseases [25, 26]. We have previously developed and applied innovative network-based approaches that allowed us to gain new insight on the role of existing viruses, but also emerging viruses (such as SARS-CoV-2), in the development of neurodegenerative diseases [23, 27–29]. In this study, we extend our previous research [23, 27–29] to investigate viral-mediated pathogenic mechanisms that might lead to AD through virus-host protein-protein interactions (PPIs), with a focus on specific viral species (HSV-1, SARS-CoV-2, EBV, HCV, HCMV, Kaposi Sarcoma-associated Herpesvirus [KSHV], Hepatitis B Virus [HBV], and Influenza A virus [IAV]). Figure 1 illustrates the main steps of our methodology. We also explore potential synergistic pathogenic effects that might arise from the reactivation of herpesviruses during SARS-CoV-2 infection, potentially increasing AD susceptibility.

Figure 1.

A, Reconstruction and analysis of the Alzheimer’s disease KEGG pathway-to-pathway network and infectious diseases subnetwork. B, Identification of common biological processes between groups of viruses and AD (viruses ∩ AD). Abbreviations: AD, Alzheimer’s disease; EBV, Epstein-Barr Virus; GO, Gene Ontology; HCMV, Human Cytomegalovirus; KEGG, Kyoto Encyclopedia of Genes and Genomes; PPI, protein-protein interaction; HSV-1, Herpes Simplex Virus 1; SARS-CoV-2, Severe Acute Respiratory Syndrome Coronavirus 2.

METHODS

Reconstruction and Analysis of the AD KEGG Pathway-Pathway Network

To identify pathways linked to AD development, our initial step involved utilizing the String: disease app within Cytoscape to retrieve the top 200 disease-associated proteins of AD (DOID:10652), determined by their highest disease score. Subsequently, using the ClueGO app [30] within Cytoscape we performed enrichment analysis on these proteins, utilizing the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. Specifically, we regarded pathways with an adjusted P value of ≤.01 (corrected with Bonferroni step-down) as statistically significant.

Furthermore, we reconstructed the AD KEGG pathway-to-pathway interaction network. By utilizing the KEGGREST package [31] in R, we parsed each of the 353 KEGG pathways (Homo sapiens) entries contained in the KEGG database [32] to extract information on the functional relationships of each pathway with other pathways. We then combined all the interactions obtained, which resulted in 1914 functional pathway-to-pathway interactions. Functional relationships between pathways represent the interconnectedness and communication that occurs between different pathways to accomplish complex physiological processes, such as where components from one pathway influence the activity or regulation of components in another pathway. Subsequently, we extracted the functional relationships related to the 64 KEGG pathways found through enrichment analysis associated with the top 200 disease-associated proteins of AD. This resulted in constructing an AD pathway-pathway network with 64 nodes (pathways) and 177 edges (functional interactions).

Additionally, we conducted composition analysis on the AD KEGG pathways network to determine the subclasses to which the 64 AD pathways belong. To achieve this, we used the KEGGREST package [31] in R to extract the subclass classification of each of the 64 pathways based on their classification in the KEGG database.

Furthermore, by using the igraph package in R [33], we performed topological analysis on the AD KEGG pathways network to evaluate the importance of the infection-related pathways in interacting with and influencing key AD pathways. This was achieved by measuring their centrality score within the network, calculated by summing 5 topological measures: hub score, degree, betweenness, closeness, and eigenvector centrality. The infectious diseases pathways were then ranked based on highest to lowest centrality score, with a high score indicating a high influence on AD.

Reconstruction of Integrated Virus-Host-AD PPI Networks

We initially conducted a thorough literature review analysis to identify viruses associated with the development of AD. Our investigation led us as to identify 10 specific viruses that are linked with AD, namely HSV-1, HSV-2, HCMV, EBV, VZV, HHV-6A/B, HHV-7, HCV, and SARS-CoV-2 [6, 10–15]. Next, we collected experimentally validated virus-host PPIs from the PHISTO [34] and VirHostNet 3.0 [35] databases (accessed date 1 August 2023). We also obtained virus-host PPIs for IAV, HBV, and KSHV. These viruses were found through the analysis of the AD KEGG pathway-pathway network as having the highest centrality score in influencing other AD-related pathways. Out of the 13 viruses, virus-host PPIs were available for 12 of them. Four viruses were excluded due to having less than 100 virus-host PPIs available. Duplicated entries from the 2 databases were eliminated. Table 1 provides detailed characteristics of the 8 viruses and the number of virus-host PPIs collected for further analysis from each viral species.

Table 1.

Characteristics of the Virus-Human Host Protein-Protein Interaction Networks

Virus Species	Family	Virus Genome	No. of Strains	No. of virus-host PPIs	No. of Viral Proteins	No. of Human Proteins
HSV-1	Herpesviridae	dsDNA	6	807	70	622
EBV	Herpesviridae	dsDNA	3	4735	153	1257
HCMV	Herpesviridae	dsDNA	4	3423	180	2036
HCV	Flaviviridae	+ssRNA	14	1793	174	1006
SARS-CoV-2	Coronaviridae	+ssRNA	NA	5605	16	2734
HBV	Hepadnaviridae	+ssRNA	10	240	26	134
KSHV	Herpesviridae	dsDNA	3	1748	137	739
IAV	Orthomyxoviridae	−ssRNA	62	10 866	210	2811

Abbreviations: −, negative strand; +, positive strand; ds, double strand; ss, single strand; EBV, Epstein-Barr Virus; HBV, Hepatitis B Virus; HCMV, Human Cytomegalovirus; HCV, Hepatitis C Virus; HSV-1, Herpes Simplex Virus 1; IAV, Influenza A Virus; KSHV, Kaposi Sarcoma-associated Herpesvirus; SARS-CoV-2, Severe Acute Respiratory Syndrome Coronavirus 2; PPIs, protein-to-protein interactions.

We then reconstructed 8 integrated virus-host-AD PPI networks (Supplementary Table 1) to identify the viral-mediated pathogenic mechanisms through which the 8 viral species (Table 1) could potentially contribute to the development and/or progression of AD. To generate the integrated networks, we merged the virus-host PPI network of each viral species with the AD PPI network, which consisted of the top 200 disease-associated proteins of AD. The confidence cutoff score, involving the interactions between the human proteins, was set at 0.8, as in our previous work [23, 28]. The confidence score is determined based on the nature and quality of the supporting evidence for the PPIs, with values ranging from 0 (indicating low confidence) to 1.0 (indicating high confidence). As the score increases, the likelihood of the PPIs being true positives also rises [36]. Consequently, a more rigorous cutoff of 0.8 was selected with the aim of improving reliability and minimizing the inclusion of false-positive results.

Identification of Common Gene Ontology Biological Processes Between Groups of Viruses and AD (Viruses ∩ AD)

To identify common biological processes among the 8 viral species (HSV-1, SARS-CoV-2, EBV, HCV, HCMV, KSHV, HBV, and IAV) through which they might contribute to AD, we classified them into 3 groups based on their type and disease they cause: (1) members of the Herpesviridae family (HSV-1, EBV, KSHV, and HCMV), which share genetic and structural similarities, are known for their ability to establish latent infections in their hosts, and most display a high degree of neurotropism; (2) viruses that cause hepatitis (HCV and HBV), which target the liver and can lead to various degrees of liver damage and inflammation; and (3) respiratory viruses (SARS-CoV-2 and IAV), which primarily target the respiratory tract and cause respiratory illness.

We then employed a methodology applied in our previous work [23], which first involves isolating 2 subnetworks from each of the 8 constructed integrated virus-host-AD PPIs networks. For each integrated network we extracted (1) the virus subnetwork, which includes the human protein targets of each viral species proteins and their first neighbors; and (2) the AD-related subnetwork, which includes the 200 disease-associated proteins of AD and their first neighbors.

Then, we performed enrichment analysis on the human proteins contained in each of the extracted subnetworks (Supplementary Table 1) by using the Gene Ontology Biological Processes (GO BP) database and retaining only significant processes that had an adjusted P value ≤.001 (corrected with Bonferroni step-down). For the enrichment analysis of the extracted subnetworks, we used a stricter P value than in enrichment analysis of the disease-associated proteins of AD, because the integrated virus-host-AD PPI networks contain a significant larger number of human proteins than the AD PPI network. Subsequently, Venn diagrams were used to identify the overlapping GO BP between the enriched results of the AD subnetwork and viral subnetwork associated with each of the 8 integrated virus-host-AD PPI networks (Supplementary Table 2).

Finally, for each group of viruses (members of the Herpesviridae family, viruses that cause hepatitis, respiratory viruses) we created GO BP-viruses networks. These networks included the overlapping GO BP (viruses ∩ AD) that were identified as being modulated by each viral species in AD. Subsequently, we identified the common GO biological processes that could potentially be influenced by all viruses contained in each group. We then performed functional enrichment analysis on the common GO BP to determine the functional groups that these GO BP belong to.

Potential Synergistic Viral-Mediated Pathogenic Effects in AD

To explore the potential for synergistic viral-mediated pathogenic effects in AD, we also identified common GO BP influenced by the reactivation of herpesviruses (HSV-1, HCMV, and EBV) during acute SARS-CoV-2 infection. Using the overlapping GO BP (viruses ∩ AD), we constructed 3 GO BP-viruses networks (SARS-CoV-2 and EBV, SARS-CoV-2 and HSV-1, SARS-CoV-2 and HCMV) and performed functional enrichment analysis on the common GO BP.

RESULTS

Infectious Diseases Pathways and Their Role in AD

We reconstructed and analyzed the AD KEGG pathway-pathway network (Figure 2A), consisting of 64 nodes (pathways) and 177 edges (interactions) to explore the potential involvement of viral infections in AD. This network was constructed using the 64 statistically significant pathways found through enrichment analysis, to be linked with the top 200 disease-related proteins associated to AD from various sources.

Figure 2.

A, Visualization of the AD KEGG pathway-pathway network, indicating the functional relationships between the 64 AD-related pathways. The top 10 hubs in the network are highlighted in purple color. B, Top 20 subclasses found in the AD KEGG pathway-pathway network and the number of pathways found in each subclass. C, Infection-related pathways ranked from highest-to-lowest centrality score in influencing other AD-related pathways in the network. Abbreviations: AD, Alzheimer's disease; KEGG, Kyoto Encyclopedia of Genes and Genomes.

Composition analysis of the 64 pathways within this network revealed that those belonged to 20 subclasses (Figure 2B) according to the KEGG database classification system. Among them, 18 pathways were related to infectious diseases, with 7 associated with viral infections, 7 with bacterial infections, and 4 with parasitic infections. This suggests a potential link between these infectious diseases and the pathogenesis of AD, as these pathways contained more overrepresented genes related to AD than expected by chance. This indicates that these infectious diseases modulate or influence genes associated with AD.

Furthermore, using topological analysis, we assessed the significance of each of the 18 infection-related pathways within the AD KEGG pathway-pathway network regarding their influence on other AD key pathways. We determined the significance of each infection-related pathway by calculating its centrality within the network, using a combined score based on 5 topological measures (hub score, degree, betweenness, closeness, and eigenvector centrality). Higher centrality scores indicate pathways crucial for information flow and control among network pathways. Consequently, high centrality score suggests a greater capability of the infection-related pathway to impact other AD-related pathways. This is a critical aspect to consider when comprehending the intricate interplay between infections and AD pathogenesis. Notably, the analysis revealed that members of the Herpesviridae family (including KSHV and HCMV), as well as members of the Flaviviridae family (HCV), and Hepadnaviridae family (HBV) had the highest centrality scores in influencing other AD-related pathways (Figure 2C).

To further investigate the influence of infectious disease pathways in AD, we extracted the infectious diseases subnetwork (Figure 3) from the AD KEGG pathways network. This analysis revealed that the 18 infection-related pathways functionally interact with 9 other AD pathways, of which 6 act as hubs within the network. Interaction with these hub nodes may enable pathogens to facilitate the development of AD through systemic pathogenic effects.

Figure 3.

Visualization of the Infectious Diseases Subnetwork isolated from the Alzheimer's disease KEGG pathway-pathway network. Abbreviations: MAPK, Mitogen-activated protein kinase; NOD, Nucleotide oligomerization domain; PI3K-AKT, Phosphoinositide 3-Kinase -Protein Kinase B; TNF, Tumor necrosis factor.

Common Viral-Mediated Pathogenic Mechanisms by Groups of Viruses in AD

To identify common viral-mediated pathogenic mechanisms among groups of viruses in AD we categorized the 8 viral species into 3 groups: (1) members of the Herpesviridae family (HSV-1, EBV, KSHV, and HCMV); (2) viruses that cause hepatitis (HCV and HBV); and (3) respiratory viruses (SARS-CoV-2 and IAV). Integrated GO BP-viruses networks were constructed for each group (Figure 4A, 4C, and 4E) using the overlapping GO BP (viruses ∩ AD) identified for each virus with AD. Functional enrichment analysis was performed to identify the functional groups these common processes belong to (Figure 4B, 4D, and 4F).

Figure 4.

A, C, E, Integrated GO BP-Viruses networks containing the overlapping GO BP found for each virus with Alzheimer's Disease for: (i) Members of the Herpesviridae Family (HSV-1, EBV, KSHV, and HCMV), (ii) Viruses that Cause Hepatitis (HCV and HBV), (iii) Respiratory Viruses (SARS-CoV-2 and IAV). B, D, F, Functional groups to which the common GO BP of each network are associated, as determined through functional enrichment analysis. Abbreviation: GO BP, Gene Ontology Biological Processes.

The integrated network encompassing the 4 members of the Herpesviridae family (HSV-1, EBV, KSHV, and HCMV) unveiled a total of 241 common GO BP, belonging to 14 functional groups, by which they can facilitate the development of AD (Figure 4A and 4B). Notably, 20.00% of the terms belonging to the group of “response to reactive oxygen species” (ROS). Additionally, 25.22% to “negative regulation of autophagy” and 6.09% to “regulation of autophagy.” Furthermore, 5.22% fall within “Aβ formation” group, 4.35% to “neuron death,” and 3.45% to “gliogenesis” group. These findings collectively suggest that herpesviruses have the capacity to impact diverse biological processes associated with key pathological features of AD, indicating their potential role in AD pathogenesis.

Furthermore, the integrated network containing the 2 viruses that cause hepatitis (HCV and HBV) revealed a total of 219 common GO BP that belong to 19 functional groups by which they can lead to AD (Figure 4C and 4D). Nearly one-third of the terms, 27.32%, belong to the “protein targeting” functional group, which involves the transport and precise localization of proteins to specific cellular regions, including the peroxisome and the mitochondrion. Disruptions can lead to protein mislocalization, which could potentially contribute to the accumulation of Aβ. Moreover, a considerable number of terms are linked to processes directly associated with endocytosis, with 3.66% of terms associated with “regulation of endocytosis” and 2.20% to the “receptor-mediated endocytosis” group. Furthermore, 6.34% of terms are allocated to the “autophagy to mitochondrion” classification, 4.88% to “response to hypoxia,” and 2.44% to “response to ROS.” Notably, certain terms find their place within functional groups closely associated with inflammation, with 3.66% associated with the “microglia cell activation” and 4.39% in the “leukocyte activation involved in inflammatory response.”

Moreover, the integrated network encompassing the 2 respiratory viruses (SARS-CoV-2 and IAV) unveiled a total of 172 common GO BP that belong to 9 functional groups (Figure 4E and 4F). Many of these terms are related to protein processes, with 29.25% in “protein targeting,” 20.75% in “de novo protein folding,” 4.72% in “regulation of protein stability,” and 0.94% in “regulation of protein-containing complex assembly” functional group.

Synergistic Actions Between Herpesviruses and SARS-CoV-2 in the Pathogenesis of AD

During acute SARS-CoV-2 infection, herpesvirus reactivation may synergistically enhance viral-mediated pathogenic effects through virus-host PPIs, potentially increasing AD susceptibility risk and contributing to post–COVID-19 AD-like cognitive symptoms. To determine the biological processes whose dysregulation might be amplified by the reactivation of each of the herpesviruses EBV, HSV-1, and HCMV during SARS-CoV-2 infection, we reconstructed 3 integrated GO BP-viruses networks. The integrated GO BP-viruses network involving EBV and SARS-CoV-2 revealed 186 common GO BP (Figure 5A) that belong to 8 functional groups (Figure 5B). Notably, a significant 70.21% of these processes are categorized under “protein targeting.” Additionally, 10.64% of terms involve “negative regulation of catabolic processes,” encompassing mechanisms that downregulate the breakdown of various substances, including cellular, lipid, protein, and glycolytic catabolic processes.

Figure 5.

A, B, Integrated GO BP-Viruses Networks and the Functional Groups of the common GO BP of EBV with SARS-CoV-2. C, D, Integrated GO BP-Viruses Networks and the Functional Groups of the common GO BP of HSV-1 with SARS-CoV-2. E, F, Integrated GO BP-Viruses Networks and the Functional Groups of the common GO BP of HCMV with SARS-CoV-2. Abbreviations: GO BP, Gene Ontology Biological Processes; EBV, Epstein-Barr Virus; SARS-CoV-2, Severe Acute Respiratory Syndrome Coronavirus 2; HSV-1, Herpes Simplex Virus 1; HCMV, Human Cytomegalovirus.

Moreover, the integrated GO BP-viruses network involving HSV-1 and SARS-CoV-2 revealed 152 common GO BP by which they can exert synergistic pathogenic effects, contributing to AD (Figure 5C). These processes are categorized into 10 functional groups (Figure 5D), with about one-third of the terms (27.54%) associated with the “regulation of autophagy.” Furthermore, 18.84% of the terms are associated with the “response to oxidative stress” group, and a significant 20.28% are linked to functional groups involved in the negative regulation of catabolic processes.

Lastly, the integrated GO BP-viruses network involving HCMV and SARS-CoV-2 revealed collective 169 common GO BP, which are classified into 8 functional groups (Figure 5E and 5F). Of notable significance, approximately 42.27% of the terms are associated with “response to unfolded protein,” referring to events that leads to changes in the state or behavior of a cell or organism in reaction to a stimulus arising from a misfolded protein. Furthermore, they have the potential to exert synergistic actions via processes related to “regulation of autophagy” (8.22%), “response to oxidative stress” (8.22%), “amyloid β formation” (6.85%), and “regulation of endocytosis” (2.74%).

DISCUSSION

Using an integrative network-based systems bioinformatics approach, we identified specific viral-mediated mechanisms by which 8 viral species (HSV-1, SARS-CoV-2, EBV, HCV, HCMV, KSHV, HBV, and IAV) could contribute to the development and/or progression of AD. We also isolated common viral-mediated pathogenic mechanisms that could lead to AD among groups of viral species. Additionally, we highlighted the potential for synergistic pathogenic effects during coinfection with 2 viruses, specifically the reactivation of herpesviruses (HSV-1, EBV, and HCMV) during SARS-CoV-2 infection. This synergy can amplify disruptions in AD-related processes via virus-host PPIs, potentially increasing susceptibility to AD.

Herpesviruses, particularly HSV-1, are associated with increased risk in developing AD [6–8, 37, 38]. Our analysis results indicate that members of the Herpesviridae family (HSV-1, EBV, KSHV, and HCMV) have the capacity to influence several of the primary pathological hallmarks associated with AD through virus-host PPIs. These encompass processes such as the generation of ROS, formation of Aβ, neuronal cell death, gliogenesis, and autophagy [39–41]. Increased production of ROS and dysregulation of autophagy are common pathological mechanisms found in many neurodegenerative diseases, including AD [42, 43]. Additionally, Aβ accumulation stands as a pivotal hallmark of AD [41], and the potential influence of herpesviruses on this process raises the possibility that they could have a crucial role in the initiation or progression of AD.

The antimicrobial protection hypothesis of AD proposes that Aβ could serve as a defense mechanism against brain infections, given its demonstrated antimicrobial effects against various pathogens, including viruses [9]. Importantly, there is evidence suggesting that Aβ oligomers can sequester herpesviruses within insoluble amyloid deposits [44]. Our current findings also indicate that herpesviruses have the capacity to influence processes involved in the formation of Aβ via virus-host PPIs. Through the evolutionary process, viruses have acquired various functions, including mimicking and interfering immune responses [45]. Thus, it is possible that herpesviruses interfere with the Aβ processes as a means to evade its antimicrobial properties and escape the immune system. This interaction could potentially lead to the perturbation of the host's antiviral response through Aβ, resulting in a detrimental interplay that accelerates the deposition of Aβ, contributing to AD pathogenesis. Therefore, understanding the complex interplay between the immune system, herpesviruses, and Aβ is imperative in unraveling the mechanisms through which these interactions can foster Aβ plaque formation.

Cognitive impairment has been reported in chronic infection with hepatitis viruses, HBV, and HCV; however, conflicting results exist on whether they influence AD pathogenesis [46, 47]. A recent investigation revealed that administering direct-acting antivirals for the treatment of HCV infection substantially diminishes the risk of mortality among patients afflicted with AD and associated dementia [48]. Our results indicate that hepatitis viruses (HBV and HCV) can modulate a diverse range of processes crucial for cellular homeostasis and dysfunction, including the regulation of protein targeting, endocytosis, response to hypoxia, ROS, and autophagy in the mitochondrion. Importantly, our results show that they can influence microglia activation via virus-host PPIs. Consistent with our results, it was previously shown that the mouse hepatitis virus can directly infect and subsequently activate microglia in a virus-induced mouse model of human neurodegenerative disease [49]. Additionally, evidence indicates that microglia activation correlates with cognitive dysfunction in chronic HCV infection [50]. Understanding the intricate interactions between hepatis virus and microglia activation could potentially open avenues for targeted therapeutic interventions to mitigate the impact of these viruses on neurodegeneration and cognitive decline.

The role of members of the family (KSHV and HCMV) and hepatitis viruses (HCV and HBV) in the development/progression of AD is also highlighted through the reconstruction and topological analysis of the AD KEGG pathway-to-pathway network. These viruses exhibited the highest centrality scores in influencing other AD-related pathways within the network, including several pathways that act as hubs. Pathogens often target high-centrality nodes, like hub nodes, which allows them to exert systemic pathogenic effects within the human host [51, 52]. Thus, our results suggest that these viruses may pose a higher risk of AD susceptibility as they have the ability to affect multiple other AD-related pathways, including hub nodes, hence facilitating the development of AD through systemic pathogenic effects.

Virus-virus interactions are very common in nature and have the potential to modify the trajectory of an ongoing infection, often influencing the severity of viral diseases [53]. Coinfection with 2 viruses may lead to synergistic pathogenic effects, where this collaborative impact results in the amplification of the dysregulation of biological processes beyond what each virus could achieve independently. This interaction may intensify the disruption of host biological processes associated with the pathogenesis of AD, potentially leading to more severe pathological outcomes and an increased risk of AD susceptibility. Given the high link of both members of the Herpesviridae family and SARS-CoV-2 with the development and/or progression of AD, as well as the demonstrated capability of SARS-CoV-2 to lead to reactivation of these herpesviruses [16–18], we sought to isolate the specific biological processes that would be amplified from the reactivation of each specific herpesviruses (HSV-1, HCMV, and EBV) during acute SARS-CoV-2 infection.

Our findings revealed a significant shared impact of EBV and SARS-CoV-2 on biological processes related to protein targeting. Disruption of protein targeting, especially in neuronal cells, can result in protein mislocalization [54], which could potentially contribute to the accumulation of Aβ. Dysfunction in neuronal cells can also potentially trigger neuroinflammation, an established pathological feature of AD. Based on these results, coinfection with these 2 viruses could potentially lead to more pronounced disturbances in protein localization, thereby potentially exacerbating the accumulation of mislocalized or misfolded proteins, and contributing to the molecular pathology of AD.

Furthermore, our results also indicate that HSV-1 and SARS-CoV-2 share the ability to impact processes whose dysregulation is associated with AD, including autophagy and response to oxidative stress. In the context of AD, oxidative stress triggers neuronal cell death by prompting the autophagy of accumulated Aβ, which in turn causes the permeabilization of the lysosomal membrane, ultimately resulting in neuron death [55]. Thus, the ability of coinfection with HSV-1 and SARS-CoV-2 to influence both of these crucial processes implicated in AD pathogenesis could increase susceptibility risk for its development.

Similar to coinfection with HSV-1 and SARS-CoV-2, the combination of HCMV and SARS-CoV-2 coinfection also appears to have the potential to lead to amplify pathogenic effects on processes involving autophagy and response to oxidative stress. Furthermore, both HCMV and SARS-CoV-2 have the capacity to impact processes related to response to unfolded protein and Aβ formation. These shared effects suggest that coinfection with these viruses could notably elevate the risk of AD development, as dysregulation of these processes are key pathological hallmarks of AD.

The present work enriches the scientific understanding of AD by demonstrating how specific viral infections could influence biochemical pathways integral to AD progression. While a recent study establishes a correlation between exposure to viruses and neurodegenerative conditions, such as AD, this prior research lacks mechanistic insights [56]. Through a detailed examination of virus-host PPIs using computational methods, we uncover not merely the presence but also the potential significant impact of microbial agents on critical disease mechanisms, such as Aβ accumulation and neuroinflammation. This progression from mere correlation to potential causation lays a hypothesis-driven foundation for further experimental research.

Our findings emphasize the importance of considering viral coinfections in AD research, suggesting that targeting these specific viral interactions could lead to novel therapeutic approaches. We advocate for the integration of multidimensional datasets in future research to enhance the predictive power and validity of microbial involvement in AD to advance beyond correlation. Artificial intelligence can contribute in this direction by analyzing integrated datasets that combine genomic, proteomic, imaging, and clinical data, allowing us to gain a more comprehensive understanding of how various genetic, molecular, and environmental factors, including microbial influences, contribute to the development/progression of AD. Overall, our research marks a significant advance in understanding AD's complex etiology, emphasizing the role of viral infections not just as correlated factors but as active participants in the disease's pathogenesis through specific, mechanistic pathways. This approach advocates for a shift toward more targeted and hypothesis-driven research in this field.

While our study's computational analysis is robust, we recognize that experimental validation in animal models would be ideal to verify these findings. Despite this limitation, our research sets the stage for future investigations to explore the identified pathways. By doing so, upcoming studies can build upon our computational predictions to further clarify the mechanisms by which viral infections may influence AD, thus advancing toward potential therapeutic interventions.

CONCLUSIONS

In summary, our findings highlight the potential of these 8 specific viruses to perturb crucial biological processes that are intricately linked to AD, reinforcing the connection between viral infections and the development and/or progression of AD. These perturbations can be mediated through shared and distinct mechanisms among viral species belonging to different categories, potentially contributing to variations in AD susceptibility based on the specific viral infection. Crucially, our findings suggest that reactivation of herpesviruses (HSV-1, HCMV, and EBV) during acute infection with SARS-CoV-2 could potentially foster a detrimental interplay contributing to neurodegeneration. However, coinfection of SARS-CoV-2 with different pairs of herpesviruses leads to the amplification of dysregulation in different biological processes, contributing to variations in AD susceptibility risk. Nonetheless, the final outcome in AD susceptibility will be also influenced by various other risk factors, including genetic susceptibility, other environmental factors linked to AD, and the presence of comorbidities.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

Acknowledgment. The authors extend their profound gratitude to the late Georgia Onisiforou, who is dearly missed, for her unwavering support and invaluable advice during the preparation of this article.

Author contributions. A. O. and P. Z. contributed to the conceptualization, review, and editing of the manuscript. A. O. developed the methodology, collected and analyzed the data, and wrote the original draft of the manuscript. Both authors have read and agreed to the published version of the manuscript.

Disclaimer. The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the Infectious Diseases Society of America Foundation.

Financial support. This work was support by the Infectious Diseases Society of America Foundation.

Supplement sponsorship. This article appears as part of the supplement “Advances in Identifying Microbial Pathogenesis in Alzheimer's Disease,” sponsored by the Infectious Diseases Society of America.

Data availability. This study utilized publicly available datasets for analysis. The data sources used are PHISTO (https://phisto.org/) and VirHostNet 3.0 (https://virhostnet.prabi.fr/).

Potential conflicts of interest. Both authors: No reported conflicts of interest. Both authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.