Epigenetic Drug Repositioning for Alzheimer’s Disease Based on Epigenetic Targets in Human Interactome

Abstract

Background:

Epigenetics has emerged as an important field in drug discovery. Alzheimer’s disease (AD), the leading neurodegenerative disorder throughout the world, is shown to have an epigenetic basis. Currently, there are very few effective epigenetic drugs available for AD.

Objective:

In this work, for the first time we have proposed 14 AD repositioning epigenetic drugs and identified their targets from extensive human interactome.

Methods:

Interacting partners of the AD epigenetic proteins were identified from the extensive human interactome to construct Epigenetic Protein-Protein Interaction Network (EP-PPIN). Epigenetic Drug-Target Network (EP-DTN) was constructed with the drugs associated with the proteins of EP-PPIN. Regulation of non-coding RNAs associated with the target proteins of these drugs was also studied. AD related target proteins, epigenetic targets, enriched pathways, and functional categories of the proposed repositioning drugs were also studied.

Results:

The proposed 14 AD epigenetic repositioning drugs have overlapping targets and miRs with known AD epigenetic targets and miRs. Furthermore, several shared functional categories and enriched pathways were obtained for these drugs with FDA approved epigenetic drugs and known AD drugs.

Conclusions:

The findings of our work might provide insight into future AD epigenetic-therapeutics.

Keywords

Alzheimer’s disease combined interactome epigenetic drug repositioning epigenetic drug targets epigenetic drug-target network epigenetic protein-protein interaction network epigenetics pathway enrichment analysis

INTRODUCTION

The term ‘epigenetics’ was coined by the renowned scientist Waddington in 1942, which originally described the influence of genetic processes on development. Epigenetics has emerged as an important field of study that identifies the influence of external and environmental factors on the expression of cellular genes (either active or inactive). It refers to the heritable changes in gene expression that does not involve changes to the underlying DNA sequence [1]. Recent data suggests that epigenetics plays a critical role in normal physiology, nutrition, and life experiences. It is also becoming clear that several diseases such as cancer, schizophrenia, and Alzheimer’s disease (AD) might have an epigenetic consequence/basis [2].

AD is the leading cause of neurological deficit in elderly people throughout the world [3]. Two hallmark processes characterize AD: 1) dysfunctional amyloid-β protein precursor processing, leading to the formation of plaques; and 2) hyperphosphorylation of tau proteins, resulting in the formation of neurofibrillary tangles [4, 5].

Major epigenetic changes in AD include alterations in DNA methylation, histone modifications, chromatin remodeling, and non-coding RNA–dysregulation [6]. Epigenetic modifications are reversible and can be potentially targeted by pharmacological and dietary interventions [6]. Recent era has witnessed an increasing attention of discovering new therapeutic uses for existing drugs, i.e., drug repositioning in neurodegenerative diseases. Drug repositioning, which refers to the application of known drugs and compounds to new diseases, offers more benefits than the traditional drug discovery methods [7]. Developing a brand-new drug consumes an enormous amount of time, money, and effort. Drug repositioning is one such strategy that can reduce this time frame, decrease costs, and improve success rates for drug development [7]. Since these known drugs have already passed through the thorough screening process of therapeutic drug development, the risk of failure for reasons of adverse toxicity is less for these repositioned drugs [8]. Therefore, these newly repositioned drugs would be ready for clinical trials quickly, speeding their integration into health care [8]. At present, scientists are investigating the new uses of many known drugs that function through epigenetic mechanisms. Epigenetic drugs work on epigenetic targets and reverse epigenetic changes in gene expression, and these drugs can be useful for treating many diseases. The focus lies mostly on following categories: DNA methyltransferase inhibitors (DNMTi), histone deacetylase inhibitors (HDACi), histone methyltransferase inhibitors (HMTi), histone demethylase inhibitors, non-coding RNAs, and dietary regimes [6].

Reducing the hypermethylation levels in some pathogenic genes may be an alternative therapy in AD in addition to conventional treatment with cholinesterase inhibitors and NMDA partial antagonists [6]. Examples of DNMTi include decitabine, azacitidine, and hydralazide, procainamide [9, 10], and natural products such as curcumin derivatives and tea polyphenols [11, 12]. HDACi includes several subclasses, namely valproic acid (short chain fatty acids); vorinostat (hydroxamic acids); and suramin, nicotinamide/niacinamide (sirtuin inhibitors, class III HDAC inhibitors) [6 , 13–15]. AD is shown to be associated with reduced histone acetylation which results in cognitive decline and poor memory formation. Since histone acetylation can be targeted by HDACi, the role of HDACi as the epigenetic drug is aptly considerable [6]. S-adenosylmethionine (SAMe) is one of the first HMTi that has been applied for the treatment of cancer [16]. SAMe improves memory and has also been shown to decrease the expression of PSEN1 [16]. Tranylcypromine is a histone demethylase inhibitor and is an irreversible inhibitor of mono amine oxidase A (MAO-A) and mono amine oxidase B (MAO-B) [11].

The microRNAs (miRs) exert regulatory control over mRNA stability and translation and may contribute to local and activity-dependent post-transcriptional control of synapse-associated mRNAs. The miRs are essential for normal brain development and function. However, their profiles are significantly altered in AD [17 –19]. Long non-coding RNAs (lncRNAs) have been found to act as competing for endogenous RNAs (ceRNAs) which can influence the post-transcriptional regulation by interfering with miR pathways [20, 21]. It has been found that both miR and lncRNA expression is in turn controlled by epigenetic processes such as DNA methylation and histone modification [22]. Both miRs and lncRNAs are reported to play an important role in epigenetic regulation of neurodegenerative diseases including AD [23].

It has been well established for quite some time that dietary regimes may have beneficial effects in patients with AD [6]. Methyl donors in diet, folic acid, vitamin B complex, SAMe, and L-methyl folate could be used as nutritional therapies to tackle the epigenetic alterations in AD [6].

A recent system-level study has investigated drug repositioning in AD by omics data mining and proposed drugs for AD [24]. However, our work is different from this previous work since we have solely focused on epigenetic drugs, their targeted proteins, and enriched pathways for AD. Moreover, our study is the first of its kind to incorporate human interactome for screening promising epigenetic drug-targets of AD. Information on the already known epigenetic drugs for neurodegenerative diseases can be a good measure toward the therapeutic intervention of epigenetic drug development in AD. In this work, we performed a system-level study to generate large-scale epigenetic networks from these ever increasing existing AD epigenetic data. The targets of FDA-approved epigenetic drugs for other diseases (which are also experimental epigenetic drugs for AD) were considered to screen possible repositioning drugs having maximum epigenetic connections in the human interactome. We further integrated our networks with pathway enrichment analysis, which provided ideal strategies for predicting new targets and their associated pathways for the repositioning drugs. Our predicted fourteen repositioning drugs were validated with their overlap with known AD-related epigenetic targets and miRs. Furthermore, several shared functional categories and pathways were obtained for the proposed repositioning drugs with the FDA-approved epigenetic drugs and known AD-drugs. The findings of our work might provide insight into future AD epigenetic-therapeutics.

METHODS

Figure 1 represents the workflow of this study. An AD-related epigenetic gene list was created by text mining from PubMed. Experimentally validated protein interaction databases, namely HPRD (Human Protein Reference Database, Release 9_041310) [25], BioGRID (version 3.2.112) [26], and Mentha [27] were combined. Only human interactions were considered to create an extensive human interactome (combined_interactome) from the three databases. After the removal of duplicate interactions and self-loops, the final combined_interactome contains 190,771 human protein-protein interactions. This combined_interactome was further used to identify the interacting partners of the epigenetic genes. The topological analyses of the nodes of each network (DTN and PPIN) were performed using the Network Analyzer module of Cytoscape software [28]. Information regarding the FDA-approved human drugs associated with the epigenetic genes was obtained from DrugBank database (version 5.0) [29]. The AD-related proteins were obtained from Alzgene database which consists of a collection of viable gene candidates for AD obtained from the already published genetic association studies [30]. The experimentally validated miRs associated with the target proteins were identified from TarBase module of DIANA TOOLS Database [31]. The drugs associated with the predicted miRs were obtained from the experimentally validated SM2miR Database [32]. Upon providing miRs in the search option, this database gives a list of drug(s) which can regulate that particular miR expression [32]. Experimentally validated lncRNAs associated with the miRs were obtained from DIANA-lncBase Database [33]. Functional enrichment analysis and Functional categories of the target proteins were studied from DAVID bioinformatics resources (version 6.8, Beta) [34]. In our work, the words ‘gene’ and ‘protein’ are used interchangeably throughout the manuscript.

Fig.1

The work flow of the methodology used in our study.

RESULT

Identification of epigenetic genes

To identify epigenetic genes associated with AD, we performed text-mining. For text-mining, we used the terms “Epigenetics and Alzheimer’s,” “Epigenetics proteins in Alzheimer’s,” etc. in PubMed.

In this way, we identified a list of 54 epigenetic genes from PubMed which might be potentially involved in AD progression (Table 1) [35 –43]. The information about epigenetic drugs was collected from the reported review [6] (Supplementary File 1). These drugs are FDA-approved epigenetic drugs for treating various diseases and are currently undergoing experimental procedures for their involvement in AD.

Table 1

Epigenetic protein list obtained from the literature

Gene/Protein Name	Oficial Symbol	Epigenetic functionality in Alzheimer’s Disease (AD)	References
SIRT1	SIRT1	Histone de acetylase	[35]
SIRT2	SIRT2	Histone de acetylase	[35]
SIRT3	SIRT3	Histone de acetylase	[35]
SIRT4	SIRT4	Histone de acetylase	[35]
SIRT5	SIRT5	Histone de acetylase	[35]
SIRT6	SIRT6	Histone de acetylase	[35]
SIRT7	SIRT7	Histone de acetylase	[35]
APP	APP	Hypomethylation of CpG islands of DNA	[36]
BACE1	BACE1	Hypomethylation of CpG islands of DNA	[36]
PS1	PSEN1	Change of DNA methylation pattern	[35]
PS2	PSEN2	Change of DNA methylation pattern	[35]
HDAC1	HDAC1	Histone de acetylase	[37]
HDAC2	HDAC2	Histone de acetylase	[37]
HDAC3	HDAC3	Histone de acetylase	[37]
HDAC8	HDAC8	Histone de acetylase	[37]
HDAC4	HDAC4	Histone de acetylase	[37]
HDAC5	HDAC5	Histone de acetylase	[37]
HDAC7	HDAC7	Histone de acetylase	[37]
HDAC9	HDAC9	Histone de acetylase	[37]
HDAC6	HDAC6	Histone de acetylase	[37]
HDAC10	HDAC10	Histone de acetylase	[37]
HDAC11	HDAC11	Histone de acetylase	[37]
CBP	CREBBP	Histone acetyl transferase	[35]
p300	EP300	Histone acetyl transferase	[35]
PCAF	KAT2B	Histone acetyl transferase	[35]
DNMT1	DNMT1	DNA methyl transferase	[35]
DNMT2	DNMT2	DNA methyl transferase	[35]
DNMT3a	DNMT3A	DNA methyl transferase	[35]
DNMT3b	DNMT3B	DNA methyl transferase	[35]
MeCP2	MECP2	Binds to histone deacetylase complex and methylated cytosine in CpG dinucleotides.	[35, 37]
CREB1	CREB1	Histone acetylation activity	[35]
HAT	HAT1	Histone acetyl transferase	[35]
NEP	MME	Hypermethylation and deacetylation causes downregulation of NEP	[36]
MAPT	MAPT	Hyperphosphorylation causes tau accumulation. Upon DNA binding it allows expression of silent genes immersed in heterochromatin. Thus it behaves in relation with epigenetic modifications.	[38]
RARB	RARB	Deacetylation causes activation mediated by SIRT1	[39]
ADAM10	ADAM10	Activated by SIRT1	[39]
NFKB	NFKB1	Deacetylation of its subunit is associated with neurodegeneration.	[39]
ROCK1	ROCK1	SIRT1 mediated inhibition	[39]
CPS1	CPS1	Deacetylation mediated by SIRT5	[39]
TET	TET1	Promotes DNA methylation	[37]
HTERT	TERT	Hypermethylation of this gene is implicated in AD	[40]
APOE	APOE	AD susceptibility loci for DNA methylation	[41]
MBD2	MBD2	DNA methylation maintenance factor involved in AD	[41]
PP2A	PPP2R4	Decreased methylation of this Tau phosphatase is seen in AD	[41]
ANK1	ANK1	Change of CpG methylation pattern associated with AD	[42]
RPL3	RPL3	Change of CpG methylation pattern associated with AD	[42]
RHBDF2	RHBDF2	Change of CpG methylation pattern associated with AD	[42]
CDH23	CDH23	Change of CpG methylation pattern associated with AD	[42]
MTHFR	MTHFR	Interindividual variance has been observed in DNA methylation	[43]
SORB3	SORB3	Hypermethylation of this gene is implicated in AD	[6]
EID1	EID1	Overexpression of EID1 impairs cognitive function via inhibiting CBP/p300 acetyltransferase activity and disrupting neuronal structure	[6]
BDNF	BDNF	Hyperacetylation of promoter region facilitates recovery from AD	[6]
CREB5	CREB5	Hypomethylation of this gene is implicated in AD	[6]
S100A2	S100A2	Hypomethylation of this gene is implicated in AD	[6]

Construction of AD-related epigenetic-protein-protein interaction network and epigenetic-drug target network

Interacting partners of the 54 AD-related epigenetic genes were identified from the combined_interactome, and the Epigenetic protein-protein interaction network (EP-PPIN) was constructed. EP-PPIN consisted of 8,412 protein-protein interactions. Subsequently, DrugBank drugs were identified corresponding to all the proteins from the EP-PPIN. Only FDA-approved drugs were considered in our study. There were 886 DrugBank drugs corresponding to the 419 proteins of the EP-PPIN. Using this drug and protein information, an Epigenetic Drug-Target Network (EP-DTN) consisting of 1,920 drug-target interactions was created. The target proteins associated with the known epigenetic drugs were identified from the combined_interactome, and the presence of these target proteins in the EP-PPIN was studied. The drugs associated with these target proteins in the EP-DTN were screened and subjected to an initial selection of epigenetic repositioning drugs.

Initial selection of epigenetic repositioning drugs

249 DrugBank drugs having 21 AD-related epigenetic targets in the EP-DTN were sorted. Out of these 249 drugs, 11 (acetyl salicylic acid, tamoxifen, caffeine, sorafenib, glyburide, spironolactone, methotrexate, diclofenac, lamivudine, ibuprofen, etoposide) were initially selected as candidates for AD epigenetic repositioning drugs. The Anatomic Therapeutic Classifications (ATC) of these drugs were studied to obtain the detailed properties.

Non-coding RNAs associated with the initially proposed repositioning drugs

We studied the experimentally validated non-coding RNAs (miRs and lncRNAs) associated with the target proteins of proposed eleven repositioning drugs. The AD-related targets of these drugs were studied from the Alzgene database (Table 2). Subsequently, the miRs associated with the targets of each drug were determined (Table 2). lncRNAs mediated regulation of AD-related miRs of each drug were also studied (Supplementary File 2).

Table 2

Drug-targets associated with proposed epigenetic repositioning drugs for AD

Proposed Repositioning Drug	AD-related microRNAs	Total drug-targets in interactome	AD related drug-targets	Common drug-targets with known epigenetic drugs
Acetylsalicylic acid (DB00945)	88	21	ALB, CYP2C8, TP53, HSPA5, PTGS2, PRKAB2, PRKAA1, NFKBIA	CYP2C8, SLC22A6, PTGS1
Caffeine (DB00201)	7	39	CYP2C8, CYP2D6, ABCB1, ADORA2A, CYP1A1, PDE3B	RYR1, CYP1A2, CYP2C8, CYP2C9, CYP2E1, CYP3A4, CYP3A5, CYP2D6
Sorafenib (DB00398)	2	23	ABCB1, CYP2C19, CYP2C8, ABCC2, CYP2D6	CYP1A2, CYP2C19, CYP2C8, CYP2C9, CYP3A4, CYP3A5, CYP2D6
Glyburide (DB01016)	12	20	ALB, ABCA1, ABCC2, ABCC8, ABCC9, KCNJ11	SLC22A6, SLCO2B1
Spironolactone (DB00421)	48	27	AR, CYP2C8, ABCB1, NR3C1, ABCC2, CACNB2	CYP2C8
Diclofenac (DB00586)	35	25	ALB, CYP2C8, ALOX5, TTR, PTGS2, ABCB1, CYP1A1, CYP2C19	PLA2G2A, SCN4A, CYP2C8, CYP2C18, SLC22A6, UGT2B7, CYP1A2, CYP2C19, CYP2C9, CYP2E1, PTGS1
Lamivudine (DB00709)	0	17	ABCC2	SLC22A1, SLC22A3, DCK, SLC22A2, SLC22A6
Ibuprofen (DB01050)	69	18	ALB, PPARG, PTGS2, BCL2, PLAT, FABP2	SLC22A6, SLCO2B1, PTGS1, UGT2B7
Atorvastatin (DB01076)	34	11	ABCC2, AHR, HMGCR	UGT2B7, SLCO2B1
Desipramine (DB01151)	23	30	ADRA1A, ADRA2A, ADRB1, ADRB2, CHRM1, HTR2A, HTR2C, SLC6A4	SLC22A1, SLC22A3, SLC22A4, SLC22A5, CYP2E1, SLC22A2, CYP2A6, CYP2C18
Nicotine (DB00184)	7	26	CHAT, CHRNA2, CHRNA3, CHRNA4, CHRNA7, CHRNB2, CHRNB4, CYP19A1, CYP1A1, CYP2C19, CYP2C8, MAOA, MAOB, CYP2D6	CHRNA2, SLC22A1, SLC22A3, SLC22A4, SLC22A5, CYP1A2, CYP2C19, CYP2C8, CYP2C9, CYP2E1, CYP3A4, MAOA, MAOB, SLC22A2, CYP2D6, CYP2A6
Bortezomib (DB00188)	3	14	CYP1A1, CYP2C19, CYP2C8, CYP2D6	CYP1A2, CYP2C19, CYP2C8, CYP2C9, CYP3A4, CYP2D6, PTGS1
Dexamethasone (DB01234)	32	16	ABCC2, CYP17A1, CYP19A1, CYP1A1, NOS2, NR3C1	CYP2A6
Metformin (DB00331)	0	4	Not Available	SLC22A1, SLC22A3, SLC22A2

Selection of drugs from SM2miR

The 313 miRs associated with target proteins of eleven epigenetic repositioned drugs were further used to search for their associated Drugbank Drugs contained in the SM2miR Database. 53 drugs were found to be common between known and unknown miRs. 43 out of the 53 SM2miR drugs were found to be present in the EP-DTN. Interestingly, three known AD epigenetic drugs, namely vorinostat (DB02546; degree 5), decitabine (DB01262; degree 1), and azacitidine (DB00928; degree 1), were present in this list of 43 drugs. Moreover, three initially proposed repositioning drugs, namely tamoxifen (DB00675; degree 13), sorafenib (DB00398; degree 8), and etoposide (DB00773; degree 7), were also present. The associated targets and the KEGG pathways of these drugs were studied.

Final selection of epigenetic drugs

The final selection of drugs was made based on their KEGG enrichment analysis (Supplementary File 3), AD-related targets (Table 2), and functional categories (Table 3). 14 drugs were proposed as significant repositioning drugs for AD.

Table 3

Significant functional categories associated with the drug-targets of proposed epigenetic repositioning drugs for AD^*

Proposed Repositioning Drug	Significant Functional Categories associated with these drugs (p > 0.05)
Acetylsalicylic acid (DB00945)	Polymorphism (15), Phosphoprotein (13), Metal-binding (7), Fatty acid biosynthesis (6), Disease mutation (6), Acetylation (6), UBL conjugation (5), ATP-binding (5), Kinase (5), Cholesterol Biosynthesis (2), S-nitrosylation (2), Microsome (2), Heme (2), Metalloprotein (2)
Caffeine (DB00201)	Metal-binding (4), Transferase (3), Kinase (3), ATP-binding (3)
Sorafenib (DB00398)	Polymorphism (8), ATP-binding (7), Membrane (7), Glycoprotein (6), Phosphoprotein (6), Kinase (5), Transferase (5), Disease mutation (5), Tyrosine-specific protein (3), Chromosomal re-arrangement (2)
Glyburide (DB01016)	ATP-binding (3), Polymorphism (2)
Spironolactone (DB00421)	Polymorphism (7), Phosphoprotein (6), Disease mutation (4), Metal-binding (4), Steroid binding (3), Zinc finger (3), UBL conjugation (3), Isopeptide bond (3), Transport (3), Calcium channel (2), Voltage-gated channel (2)
Diclofenac (DB00586)	Polymorphism (7), Metal-binding (5), Membrane (5), Iron (4), Oxidoreductase (4), Chromoprotein (3), Microsome (3), Heme (3), Endoplasmic reticulum (3), Monooxygenase (2), Electron transfer (2)
Lamivudine (DB00709)	ATP-binding (3), Nucleotide-binding (3)
Ibuprofen (DB01050)	Phosphoprotein (4), Metal-binding (3)
Atorvastatin (DB01076)	Glycoprotein (10), Transmembrane (10), Transport (6), ATP-binding (4), Endoplasmic reticulum (3), Microsome (2), Steroid metabolism (2)
Desipramine (DB01151)	Membrane (29), Glycoprotein (26), Polymorphism (24), Disulfide bond (21), G-protein coupled receptor (18), Phosphoprotein (16), Transport (8), Lipoprotein (7), Postsynaptic cell membrane (5), Cell junction (5), Symport (4), Sodium (4), Behavior (3), Monooxygenase (3), Microsome (3), Heme (3), Iron (3), Neurotransmitter transport (2)
Nicotine (DB00184)	Membrane (25), Polymorphism (24), Glycoprotein (15), Ion transport (13), Oxidoreductase (12), Monooxygenase (10), Heme (10), Metal-binding (10), Microsome (9), Endoplasmic reticulum (9), Postsynaptic cell membrane (8), Synapse (8), Ion channel (8), Disease mutation (8), NADP (5), Steroid metabolism (4), Epilepsy (3)
Bortezomib (DB00188)	Polymorphism (12), Membrane (9), Microsome (8), Oxidoreductase (8), Metal-binding (8), Monooxygenase (7), Acetylation (6), Proteasome (5), Lipid metabolism (5), NADP (3), Host-virus interaction (3), Protease (3)
Dexamethasone (DB01234)	Polymorphism (14), Membrane (13), Metal-binding (11), Heme (9), Iron (9), Oxidoreductase (9), Monooxygenase (8), Disease mutation (8), Microsome (5), Endoplasmic reticulum (5), NADP (4), Steroidogenesis (2)
Metformin (DB00331)	Ion transport (3), Transport (3)

^*Number of drug-targets present in each category is shown within brackets. Functional categories common with known epigenetic drugs are shown in bold.

DISCUSSION

The concept of epigenetic drug repositioning has emerged as an important field of study. Study of epigenetics is experiencing an unprecedented fast and a pervasive rise due to its implications for cancer, neurodegenerative disorders, and many other diseases. Environmental factors, stress, food, and nutrition are increasingly recognized as important epigenetic modulators. These factors act as stimuli for DNA methylation, histone modification, and non-coding RNA–mediated modifications which initiate and sustain epigenetic changes. In light of the growing interest in epigenetics, it is important to identify epigenetic target proteins from the whole interactome that will attribute to precise therapeutic development. With this view, our study proposed a novel approach which took into account the information of all the epigenetic target proteins in AD and their interactors from the whole interactome. This information guided us to screen the drugs from DrugBank having similar epigenetic targets with known epigenetic AD drugs and the formation of EP-DTN. EP-DTN contains seven known AD drugs (galantamine DB00674; choline DB00122; donepezil DB00843; ergoloidmesylate DB01049; phosphatidyl serine DB00144; rivastigmine DB00989; rosiglitazone DB00412) and all known epigenetic drugs in AD (except tranylcypromine DB00752). The initial selection of drugs was done based on their maximum number of epigenetic connections. A further selection of the epigenetic repositioning drugs was done based on their disease pathways enriched with these epigenetic target proteins. Additionally, functional categories of these targets were studied along with the known AD drugs and epigenetic drugs. Since non-coding RNAs are considered as one of the most important epigenetic modulators in biological systems, we further studied the non-coding RNAs, i.e., miRs and lncRNAs associated with these initial repositioning drugs. To incorporate more epigenetically regulated drugs, we subjected the associated miRs obtained from the initial repositioning drugs to SM2miR, which gave information about new DrugBank drugs that are regulated by these miRs. The second round of selection of drugs was done based on the association with AD-related miRs. We combined the drugs obtained from a first and second round of selection. Final selection of the epigenetic repositioning drugs was done by the disease pathways enriched with these epigenetic target proteins. Additionally, functional categories of these targets were studied along with the known and experimental AD epigenetic drugs. Furthermore, we validated these drugs by their association with known AD target proteins. Alzgene specific proteins and known epigenetic target proteins were both considered for the final selection of epigenetic repositioning drugs. In addition to the existing AD-related targets, several new unknown targets of these proposed repositioning drugs were also elucidated. Concordantly, the KEGG enrichment analysis of these targets was also performed to reveal their function in AD. Three drugs, namely tamoxifen (DB00675), etoposide (DB00773), and methotrexate (DB00563), were present in the list of initially proposed repositioning drugs. These cancer drugs were very well connected in the EP-DTN. However, these were not included in the list of 14 final repositioning drugs due to their toxic side effects.

Thus, finally, 14 drugs, (acetylsalicylic acid, caffeine, sorafenib, glyburide, spironolactone, diclofenac, lamivudine, ibuprofen, metformin, atorvastatin, desipramine, dexamethasone, bortezomib, and nicotine) were selected as AD epigenetic repositioning drugs (Table 2). It was found that cytochrome P450 family proteins (CYP2C8, CYP1A1, etc.), solute carrier 22 family proteins (SLC22A6, SLC221, etc.), and ATP binding cassette family proteins (ABCC2, ABCG2, etc.) are the maximally targeted proteins among these drugs (Table 2). Besides, serotonergic synapse, retinol metabolism, arachidonic acid metabolism, insulin signaling, metabolic pathways, ABC transporters, etc. are the over-representative pathways associated with these drugs (Fig. 2) (Supplementary File 3). Figure 3 shows the three layered regulation of proposed epigenetic repositioning drugs. Table 3 and Supplementary File 3 show the detailed properties of these drugs. These drugs are discussed in detail in the following section.

Fig.2

Schematic diagram of over-representative KEGG pathways of proposed epigenetic repositioning drugs. Drugs and targeted pathways are shown within boxes.

Fig.3

Three layered regulation of proposed epigenetic repositioning drugs. Selected AD-related miRs and their targets are shown. Drugs are shown within arrowhead; miRs are shown within rectangles. Targets are shown within octagons.

Acetylsalicylic acid (DB00945)

Acetylsalicylic acid is an anti-inflammatory drug, used in the non-steroidal anti-inflammatory drug therapy. It has anti-pyretic properties and acts as an inhibitor of cyclooxygenase which results in the inhibition of the biosynthesis of prostaglandins [29]. Salicylic acid is under investigation for its role in the treatment of different stages of AD [44]. It has been found that patients with greater neurodegeneration and cerebral inflammation have additional advantages of longer-term treatment with this drug [44].

Caffeine (DB00201)

Caffeine is a naturally occurring methylxanthine which acts as a stimulant of the central nervous system [29]. Caffeine inhibits cyclic nucleotide phosphodiesterases, modulates intracellular calcium handling, and acts as an antagonist of adenosine receptors [29]. It has been reported that caffeine lowers the risk of dementia [45, 46].

Sorafenib (DB00398)

Sorafenib is a tyrosine kinase inhibitor, used in several cancers. At present, tyrosine kinase inhibitors are widely investigated in non-oncology diseases, involving inflammatory and autoimmune processes [47]. The role of tyrosine kinases in the pathogenesis of AD has been confirmed by the results of experimental trials (masitinib, another tyrosine kinase inhibitor is currently undergoing clinical trials for AD) [47].

Glyburide (DB01016)

Glyburide is an oral anti-hyperglycemic agent used for the treatment of non-insulin dependent diabetes mellitus [29]. It belongs to the sulfonylurea category which acts by stimulating beta cells of the pancreas to release insulin. It binds to ATP-sensitive potassium channels on the pancreatic cell surface and reduces the potassium conductance and causes depolarization of the membrane [29]. It has a role in calcium channels also. It is well established that AD is closely associated with diabetes [48]. Moreover, anti-diabetic drugs have been examined in AD therapeutics [49].

Spironolactone (DB00421)

Spironolactone is a drug of the cardiovascular system, used mainly in the treatment of refractory edema in patients with congestive heart failure, nephrotic syndrome, or hepatic cirrhosis [29]. It is a mineralocorticoid receptor antagonist, working by inhibiting cJun-NK [50]. Studies have linked this drug with cognitive implications in chronic corticosterone-treated pathologies [50].

Diclofenac (DB00586)

Diclofenac is a non-steroidal anti-inflammatory agent (NSAID) with antipyretic and analgesic actions. It is primarily available as the sodium salt. The anti-inflammatory effects of diclofenac are believed to be due to inhibition of both leukocyte migration and the enzyme cyclooxygenase (COX-1 and COX-2), leading to the peripheral inhibition of prostaglandin synthesis [29]. It has been identified as an Aβ aggregation blocker. It has been shown to reduce inflammation and to abrogate amyloiddeposition.

Lamivudine (DB00709)

Lamivudine is a reverse transcriptase inhibitor and zalcitabine analog in which a sulfur atom replaces the 3’ carbon of the pentose ring [29]. It is used to treat human immunodeficiency virus type 1 (HIV-1) and hepatitis B [29]. It is well established that HIV-associated neurocognitive disorder occurs in 30% –50% of HIV-affected individuals [53]. Studies have also found a close association between AD and HIV [53]. Therefore, this drug could be a potentially important candidate epigenetic drug for treating AD.

Ibuprofen (DB01050)

Ibuprofen, a propionic acid derivative, is a prototypical nonsteroidal anti-inflammatory agent with analgesic and antipyretic properties. The anti-inflammatory effects of this drug involve the inhibition of COX-1 and COX-2 enzymes [29]. COX-2 is induced in the brain in response to excitatory synaptic activity and inflammation [29]. Studies have found that ibuprofen prevented memory impairment without producing any measurable changes in Aβ accumulation or glial inflammation [54]. This drug increases the levels of norepinephrine and dopamine [54].

Metformin (DB00331)

No Alzgene specific target was identified for this drug. The non-Alzgene specific targets were considered for further identification of miRs, and ten AD-related miRs were identified (Table 2). Metformin is a bi-guanide anti-hyperglycemic agent used for treating non-insulin-dependent diabetes mellitus [29]. It improves glycemic control by decreasing hepatic glucose production, decreasing glucose absorption and increasing insulin-mediated glucose uptake. Its effects are mediated by the initial activation by metformin of AMP-activated protein kinase (AMPK), a liver enzyme that plays an important role in insulin signaling, whole body energy balance, and the metabolism of glucose and fats [29]. Studies have found that metformin attenuates Aβ pathology mediated by nicotinic acetylcholine receptors in a C. elegans model of AD [55].

4.10 Atorvastatin (DB01076)

Atorvastatin (Lipitor) is a member of the drug class known as statins. It is used for lowering cholesterol [29]. Atorvastatin is a competitive inhibitor of hydroxymethylglutaryl-coenzyme A (HMG-CoA) reductase, the rate-determining enzyme in cholesterol biosynthesis via the mevalonate pathway [29]. Atorvastatin acts primarily in the liver. It has been reported to improve AD, through anti-inflammatory pathway [56]. In a mouse model of AD, it greatly improved the spatial cognitive deficits caused by Aβ mediated inflammatory pathways (by pro-inflammatory cytokines) [56].

4.11 Dexamethasone (DB01234)

Dexamethasone is a glucocorticoid agonist [29] which has been implicated in AD [50, 57]. Unbound dexamethasone crosses cell membranes and binds with high affinity to specific cytoplasmic glucocorticoid receptors [29]. The anti-inflammatory actions of dexamethasone are thought to involve phospholipase A₂ inhibitory proteins, lipocortins, which control the biosynthesis of potent mediators of inflammation such as prostaglandins and leukotrienes [29].

4.12 Desipramine (DB01151)

Desipramine hydrochloride is a dibenzazepine derivative tricyclic antidepressant (TCA) [29]. TCAs are potent inhibitors of serotonin and norepinephrine reuptake. The antidepressant effects of TCAs are thought to be due to an overall increase in serotonergic neurotransmission [29]. TCAs also block histamine-H₁ receptors, α₁-adrenergic receptors and muscarinic receptors, which accounts for their sedative, hypotensive, and anticholinergic effects [29]. Studies have found that 40% of AD patients have depressive symptoms [58]. It suggests a possibility that anti-depression treatment might be beneficial to cognitive impairment in AD [58].

4.13 Bortezomib (DB00188)

Bortezomib is the first therapeutic proteasome inhibitor to be tested in humans [29]. The boron atom within bortezomib catalytically binds the active site of the 26S proteasome with high affinity [29].

4.14 Nicotine (DB00184)

Nicotine is the prototypical agonist at nicotinic cholinergic receptors where it dramatically stimulates neurons and ultimately blocks synaptic transmission [29]. In the brain, nicotine binds to nicotinic acetylcholine receptors on dopaminergic neurons in the cortico-limbic pathways. This causes the channel to open and allow conductance of multiple cations including sodium, calcium, and potassium [29].

Numerous pre-clinical studies have been conducted to assess the potential of nicotine and nicotine-like compounds as therapeutic agents for conditions as diverse as AD [59]. Varenicline (DB01273), a partial agonist of α4β2 and full agonist at α7 nicotinic acetylcholine receptors, currently used for smoking cessation and over the counter drug prescribed for AD. From our combined_interactome, we identified that CHRNA7 is the only target for varenicline in interactome, and interestingly, this target is also a target of our proposed drug nicotine (DB00184). Since varenicline is a single targeted drug and nicotine is a multi-targeted drug, it may be more potent than varenicline. There have been several encouraging (albeit small) clinical trials with nicotine for AD [59].

4.15 Conclusion

Epigenetic drugs can be a promising and effective way for treating a complex disease like AD. However, currently, there are very few effective epigenetic drugs available for AD. In our desire to discover epigenetic drugs for AD, we performed a system-level study of network biology to generate large-scale epigenetic networks from ever increasing existing AD epigenetic data. The targets of known AD epigenetic drugs were considered to screen possible repositioning drugs having maximum epigenetic connections in the human interactome. We further integrated our networks with pathway enrichment analysis, which provided ideal strategies for predicting new targets and their associated pathways for the repositioning drugs. Our predicted repositioning drugs were validated with their overlap with AD-related epigenetic targets and miRs. Furthermore, several shared functional categories were obtained for the proposed repositioning drugs with the known epigenetic drugs. In addition to the existing AD-related targets, several new unknown targets of these proposed repositioning drugs were also elucidated. The information regarding the unknown targets will provide useful insights into understanding the systems level implication of such complex diseases. The fourteen epigenetic repositioning drugs identified in our study may provide novel therapeutic options for AD.

Footnotes

ACKNOWLEDGMENTS

Conceived and designed the experiments: DR; Performed the experiments: PC, DR, NR; Analyzed the data: DR, PC; Wrote the paper: DR, PC.

The authors would like to thank the Department of Biophysics, Bose Institute.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Egger

, Liang

, Aparicio

, Jones

(2004) Epigenetics in human disease and prospects for epigenetic therapy. Nature 429, 457–463.

Portela

, Esteller

(2010) Epigenetic modifications and human disease. Nat Biotechnol 28, 1057–1068.

Alzheimer’s Association (2013) 2013 Alzheimer’s disease facts and figures. Alzheimers Dement 9, 208–245.

, Li

, Jin

(2012) Epigenetics-based therapeutics for neurodegenerative disorders. Curr Transl Geriatr Exp Gerontol Rep 1, 229–236.

Balin

, Hudson

(2014) Etiology and pathogenesis of late-onset Alzheimer’s disease. Curr Allergy Asthma Rep 14, 417.

Cacabelos

, Torrellas

(2014) Epigenetic drug discovery for Alzheimer’s disease. Expert Opin Drug Discov 9, 1059–1086.

Langedijk

, Mantel-Teeuwisse

, Slijkerman

, Schutjens

(2015) Drug repositioning and repurposing: Terminology and definitions in literature. Drug Discov Today 20, 1027–1034.

DiMasi

, Hansen

, Grabowski

, Lasagna

(1991) Cost of innovation in the pharmaceutical industry. J Health Econ 10, 107–142.

Lötsch

, Schneider

, Reker

, Parnham

, Schneider

, Geisslinger

, Doehring

(2013) Common non-epigenetic drugs asepigenetic modulators. Trends Mol Med 19, 742–753.

10.

Kelly

, De Carvalho

, Jones

(2010) Epigenetic modifications as therapeutic targets. Nat Biotechnol 28, 1069–1078.

11.

Nebbioso

, Carafa

, Benedetti

, Altucci

(2012) Trials with ‘epigenetic’ drugs: An update. Mol Oncol 6, 657–682.

12.

Cuadrado-Tejedor

, Oyarzabal

, Lucas

, Franco

, García-Osta

(2013) Epigenetic drugs in Alzheimer’s disease. BioMol Concepts 4, 433–445.

13.

Nuutinen

, Suuronen

, Kauppinen

, Salminen

(2010) Valproic acid stimulates clusterin expression in human astrocytes: Implications for Alzheimer’s disease. Neurosci Lett 475, 64–68.

14.

, Ryder

, Li

, Wu

, Fox

, Solenberg

, Brune

, Paul

, Zhou

, Liu

, Ni

(2004) Lithium, a common drug for bipolar disorder treatment, regulates amyloid-precursor protein processing. Biochemistry 43, 6899–6908.

15.

Green

, Steffan

, Martinez-Coria

, Sun

, Schreiber

, Thompson

, LaFerla

(2008) Nicotinamide restores cognition in Alzheimer’s disease transgenic mice via a mechanism involving sirtuin inhibition and selective reduction of Thr231-phosphotau. J Neurosci 28, 11500–11510.

16.

Peedicayil

(2014) Epigenetic drugs in cognitive disorders. Curr Pharm Des 20, 1840–1846.

17.

Nelson

, Wang

(2010) MiR 107 is reduced in Alzheimer’s disease brain neocortex: A validation study. J Alzheimers Dis 21, 75–79.

18.

Long

, Ray

, Lahiri

(2012) MicroRNA-153 physiologically inhibits expression of amyloid-β precursor protein in cultured human fetal brain cells and is dysregulated in a subset of Alzheimer disease patients. J Biol Chem 287, 31298–31310.

19.

Femminella

, Ferrara

, Rengo

(2015) The emerging role of microRNAs in Alzheimer’s disease. Front Physiol 6, 40.

20.

Das

, Ghosal

, Sen

, Chakrabarti

(2014) lnCeDB: Database of human long noncoding RNA acting as competing endogenous RNA. PLoS One 9, e98965.

21.

Tay

, Kats

, Salmena

, Weiss

, Tan

, Ala

, Karreth

, Poliseno

, Provero

, Di Cunto

, Lieberman

, Rigoutsos

, Pandolfi

(2011) Coding-independent regulation of the tumor suppressor PTEN by competing endogenous mRNAs. Cell 147,, 344–357.

22.

, Zhang

, Li

, Lu

, Chen

, Wang

, Li

, Xu

, Li

(2015) Genome-wide DNA methylome analysis reveals epigenetically dysregulated non-coding RNAs in human breast cancer. Sci Rep 5,, 8790.

23.

Lardenoije

, Iatrou

, Kenis

, Kompotis

, Steinbusch

, Mastroeni

, Coleman

, Lemere

, Hof

, van den Hove

, Rutten

(2015) The epigenetics of aging and neurodegeneration. Prog Neurobiol 131, 21–64.

24.

Zhang

, Schmitt-Ulms

, Sato

, Xi

, Zhang

, Zhou

, St George-Hyslop

, Rogaeva

(2016) Drug repositioning for Alzheimer’s disease based on systematic ‘omics’ data mining. PLoS One 11, e0168812.

25.

Keshava Prasad

, Goel

, Kandasamy

, Keerthikumar

, Kumar

, Mathivanan

, Telikicherla

, Raju

, Shafreen

, Venugopal

, Balakrishnan

, Marimuthu

, Banerjee

, Somanathan

, Sebastian

, Rani

, Ray

, Harrys Kishore

, Kanth

, Ahmed

, Kashyap

, Mohmood

, Ramachandra

, Krishna

, Rahiman

, Mohan

, Ranganathan

, Ramabadran

, Chaerkady

, Pandey

(2009) Human protein reference database - 2009 update. (Database issue),. Nucleic Acids Res 37, D767–D772.

26.

Stark

, Breitkreutz

, Reguly

, Boucher

, Breitkreutz

, Tyers

(2006) BioGRID: A general repository for interaction datasets. (Database issue). Nucleic Acids Res 34, D535–D539.

27.

Calderone

, Castagnoli

, Cesareni

(2013) mentha: A resource for browsing integrated protein-interaction networks. Nat Methods 10, 690–691.

28.

Shannon

, Markiel

, Ozier

, Baliga

, Wang

, Ramage

, Amin

, Schwikowski

, Ideker

(2003) Cytoscape: A software environment for integrated models of biomolecular interaction networks. Genome Res 13, 2498–504.

29.

Wishart

, Knox

, Guo

, Shrivastava

, Hassanali

, Stothard

, Chang

, Woolsey

(2006) DrugBank: A comprehensive resource for in silico drug discovery and exploration. (Database issue). Nucleic Acids Res 34, D668–D672.

30.

Bertram

, McQueen

, Mullin

, Blacker

, Tanzi

(2007) Systematic meta-analyses of Alzheimer disease genetic association studies: The AlzGene database. Nat Genet 39, 17–23.

31.

Vlachos

, Paraskevopoulou

, Karagkouni

, Georgakilas

, Vergoulis

, Kanellos

, Anastasopoulos

, Maniou

, Karathanou

, Kalfakakou

, Fevgas

, Dalamagas

, Hatzigeorgiou

(2015) DIANA-TarBase v7.0: Indexing more than half a million experimentally supported miRNA:mRNA interactions. (Database issue). Nucleic Acids Res 43, D153–D159.

32.

Liu

, Wang

, Meng

, Wang

, Zhang

, Dai

, Yu

, Li

, Jiang

(2013) SM2miR: A database of the experimentally validated small molecules’ effects on microRNA expression. Bioinformatics 29, 409–411.

33.

Paraskevopoulou

, Georgakilas

, Kostoulas

, Reczko

, Maragkakis

, Dalamagas

, Hatzigeorgiou

(2013) DIANA-LncBase: Experimentally verified and computationally predicted microRNA targets on long non-coding RNAs. (Database issue). Nucleic Acids Res 41, D239–D245.

34.

Huang

, Sherman

, Lempicki

(2009) Systematic and integrative analysis of large gene lists using DAVID Bioinformatics Resources. Nat Protoc 4, 44–57.

35.

Adwan

, Zawia

(2013) Epigenetics: A novel therapeutic approach for the treatment of Alzheimer’s disease. Pharmacol Ther 139, 41–50.

36.

Liu

, Le

(2014) Epigenetic modifications of chronic hypoxia-mediated neurodegeneration in Alzheimer’s disease. Transl Neurodegener 3, 7.

37.

Coppedè

(2014) The potential of epigenetic therapies in neurodegenerative diseases. Front Genet 5, 220.

38.

Ildefonso

, Erika

, Laura

, Verónica

, Martha

, Guadalupe

, Sergio

, Juan

, María

(2015) Presence ofphosphorylated tau protein in the skin of Alzheimer’s diseasepatients. J Mol Biomark Diagn S6, 005.

39.

Huber

, Superti-Furga

(2011) After the grape rush: Sirtuins as epigenetic drug targets in neurodegenerative disorders. Bioorg Med Chem 19, 3616–3624.

40.

Lovrečić

, Maver

, Zadel

, Peterlin

(2013) The role of epigenetics in neurodegenerative diseases. In Neurodegenerative Diseases, Dr. Kishore

Uday

, ed. InTech. doi: 10.5772/54744

41.

Kwok

(2010) Role of epigenetics in Alzheimer’s and Parkinson’s disease. Epigenomics 2, 671–682.

42.

Lord

, Cruchaga

(2014) The epigenetic landscape of Alzheimer’s disease. Nat Neurosci 17, 1138–1140.

43.

Wang

, Yu

, Tan

, Jiang

, Tan

(2013) Epigenetic mechanisms in Alzheimer’s disease: Implications for pathogenesis and therapy. Ageing Res Rev 12, 1024–1041.

44.

Wattmo

, Minthon

, Wallin

ÅK

(2016) Mild versus moderatestages of Alzheimer’s disease: Three-year outcomes in routineclinical setting of cholinesterase inhibitor therapy. Alzheimers Res Ther 8, 7.

45.

Chan

, Kantham

, Rao

, Palanivelu

, Pham

, Shaw

, McGeary

, Ross

(2016) Metal chelation, radical scavenging and inhibition of Aβ₄₂ fibrillation by food constituents in relation to Alzheimer’s disease. Food Chem 199, 185–194.

46.

Nehlig

(2016) Effects of coffee/caffeine on brain health and disease: What should I tell my patients? Pract Neurol 16, 89–95.

47.

Gągało

, Rusiecka

, Kocić

(2015) Tyrosine kinase inhibitor as a new therapy for ischemic stroke and other neurologic diseases: Is there any hope for a better outcome? Curr Neuropharmacol 13, 836–844.

48.

Alam

, Islam

, Sasongko

, Gan

(2016) Type 2 diabetes mellitus and Alzheimer’s disease: Bridging the pathophysiology and management. Curr Pharm Des 22, 4430–4442.

49.

Rizvi

, Shaikh

, Waseem

, Shakil

, Abuzenadah

, Biswas

, Tabrez

, Ashraf

, Kamal

(2015) Role of anti-diabetic drugs as therapeutic agents in Alzheimer’s disease. EXCLI J 14, 684–696.

50.

Solas

, Gerenu

, Gil-Bea

, Ramírez

(2013) Mineralocorticoid receptor activation induces insulin resistancethrough c-Jun N-terminal kinases in response to chroniccorticosterone: Cognitive implications. J Neuroendocrinol 25, 350–366.

51.

Guo

, Yu

, McGeer

(2010) Simple in vitro assays to identify amyloid-beta aggregation blockers for Alzheimer’s disease therapy. J Alzheimers Dis 19, 1359–1370.

52.

Hirohata

, Ono

, Naiki

, Yamada

(2005) Non-steroidal anti-inflammatory drugs have anti-amyloidogenic effects for Alzheimer’s beta-amyloid fibrils in vitro. Neuropharmacology 49, 1088–1099.

53.

Turner

, Chadwick

, Horton

, Simon

, Jiang

, Esposito

(2016) An individual with human immunodeficiency virus, dementia, and central nervous system amyloid deposition. Alzheimers Dement (Amst) 4, 1–5.

54.

Woodling

, Colas

, Wang

, Minhas

, Panchal

, Liang

, Mhatre

, Brown

, Ko

, Zagol-Ikapitte

, van der Hart

, Khroyan

, Chuluun

, Priyam

, Milne

, Rassoulpour

, Boutaud

, Manning-Boğ

, Heller

, Andreasson

(2016) Cyclooxygenase inhibition targets neurons to prevent an early behavioral decline in Alzheimer’s disease model mice. (Pt 7). Brain 139, 2063–2081.

55.

Ahmad

, Ebert

(2017) Metformin attenuates Aβ pathology mediated through levamisole sensitive nicotinic acetylcholine receptors in a C. elegans model of Alzheimer’s disease. Mol Neurobiol 54, 5427–5439.

56.

Zhao

, Chen

, Wang

, Li

, Zhi

, Yin

, Wan

, Chen

(2016) Atorvastatin in improvement of cognitive impairments caused by amyloid β in mice: Involvement of inflammatory reaction. BMC Neurol 16, 18.

57.

Di Meco

, Joshi

, Lauretti

, Praticò

(2016) Maternal dexamethasone exposure ameliorates cognition and tau pathology in the offspring of triple transgenic AD mice. Mol Psychiatry 21, 403–410.

58.

Wang

, Li

, Yu

, Wu

, Sun

, Qi

(2016) Desipramine improves depression-like behavior and working memory by up-regulating p-CREB in Alzheimer’s disease associated mice. J Integr Neurosci 15, 247–260.

59.

Terry

AV Jr

, Plagenhoef

, Callahan

(2016) Effects of the nicotinic agonist varenicline on the performance of tasks of cognition in aged and middle-aged rhesus and pigtail monkeys. Psychopharmacology (Berl) 233, 761–771.