Drug repurposing and therapeutic anti-miRNA predictions in - - PowerPoint PPT Presentation

Drug repurposing and therapeutic anti-miRNA predictions in oxLDL-induced the proliferation of vascular smooth muscle cell associated diseases

Shun-Tsung Chen, Chien-Hung Huang, Victor C. Kok, Chi-Ying F Huang, Jin-Shuei Ciou, Jeffrey J. P. Tsai, Nilubon Kurubanjerdjit and Ka-Lok Ng *

* Department of Bioinformatics & Medical Engineering

Asia University, Taiwan

2

Contents

• 1. Introduction – human cardiovascular disease (CVD), drug repurposing,

Connectivity Map (cMap), antimiR ,

• 2. Methods - The eBayes algorithms for identifying DEGs, ClusterONE,

cMap, DAVID, DrugBank, NCBI, miRTarBase and TarBase resources for identifying potential drugs and therapeutic antimiR

• 3. Results - supported by the literature, in-vitro IC50 and clinical trials
• 4. Discussion - obtained certain successive results, there are still some

deficiencies or limits

• 5. Conclusions - drug-finding pipeline is effective for drug repurposing and

therapeutic antimiR discovery

Introduction

Ø Cardiovascular disease (CVD) is the most important cause of

morbidity and mortality worldwide. According to the World Health Organization (WHO) report, CVD affects tens of millions of human beings each year. Therefore, how to improve the diagnosis, treatment and prevention of CVD is an urgent and important issue.

Ø The vessel is subjected to oxidized low-density lipoprotein (oxLDL)

→ initiate signaling pathways, then trigger differential gene expression, next induce vascular smooth muscle cell (VSMC) dedifferentiation → results in VSMC phenotypic transition → VSMC phenotypic transition is the major cause of ATs, restenosis and hypertension.

3

4

Blood artery and cellular structure

VSMC

5

VSMC proliferation-associated diseases Heart and blood artery

6

Atherosclerosis (ATs) and VSMC

Hypothesis ~ ATs and cancer formation

Ø Studies have suggested that ATs and cancer formation involve similar

cellular processes: proliferation, inflammation and genomic instability.

Ø Both

types

• f

diseases possess common pathways

• r

signal transduction networks, such as PI3k/Akt, can mediate several functional and morphological alterations of VSMC after being activated to develop vascular diseases, as well as affect the growth, apoptosis and cell cycle regulation of various cell types to induce cancer progression.

Ø It is also known that the MAPK pathway, involved in VSMC

proliferation, hypertrophy, and migration, → The concept of similar therapeutic strategies and drugs is proposed to target both conditions via identifying potential drugs.

7

From disease to drug discovery

Nature Reviews Cancer 7(1), 54-60 (2007)

The main objective of biomedical research is to connect human diseases with the genetic elements and identify drugs that treat them

9

Traditional drug discovery process

Reference : http://www.slidegeeks.com/shapes/product/business-steps-powerpoint- templates-marketing-drug-discovery-process-ppt-slides

Drug repurposing

Ø It is a recently developed approach that endeavors

to identify new uses for existing drugs and has achieved certain successes.

Ø It has the potential to accelerate the development

and bring down the exploitation costs for drugs, as well as reducing side effects.

Ø The drug-gene interaction database, cMap, is used

to find potential drugs for treating the proliferation

• f VSMC associated diseases.

Nature Reviews Cancer 7(1), 54-60 (2007) Science 313(5795), 1929-1935 (2006)

cMap

Using Gene Expression Signatures to connect Drugs, Genes, and Diseases.

cMap (continued)

Ø The cMap is a robust tool to address questions in different cells. → four cell line experiments, i.e. MCF7, PC3, HeL60 and SKMEL5 → dosage ( ~ 2 to 20 micro-molar), duration ( ~ 6 hours) Ø This diversity of cell types provides an opportunity to assess the extent to which results are context dependent. Ø Even though it is impossible to meet all conditions, results from different cell lines are in general acceptable to most of the users.

query

Negative signature

Disease Normal

If a drug signature could reverse, at least in part, the gene expression signature of disease, this drug might have the potential to inhibit disease pathways and thereby treat VSMC proliferation-associated diseases.

Hypothesis

Search for the proliferation of VSMC associated diseases drugs

14

Central dogma of molecular biology miRNA & antimiR (suppress miRNA)

15

miRNA & antimiR

Ø MiRNAs may modulate gene expression and

protein synthesis through negatively regulating. → regulate many biological processes (BPs), such as cell development, differentiation, signaling, metabolism and etc.

Ø MiR-21, 143/145 and 221/222

→ actively involved in regulation of VSMC

Ø MiR-145 has an anti-proliferative effect when

• ver-expressed.

16

miRNA & antimiR (continued)

¡ The above results indicated miRNAs play a key role in

the regulation of the VSMC phenotype transition and causing VSMC associated diseases. → as biomarkers in the proliferation of VSMC associated diseases diagnosis and prevention → development of miRNA-based therapeutic treatments, through using antimiRs to inhibit miRNA expression

Ø MiRNA-based therapy (gene therapy)

→ FDA (the treatment of hyperlipidemia)

17

miRNA & antimiR (continued)

Ø AntimiR mediated pharmacological inhibition

• f disease-associated miRNAs, shows great

promise in the development of novel miRNA- based therapeutics.

Ø three in vivo studies suggested that the inhibition of

miR-33 by antimiR- 33 → raise HDL cholesterol levels → reverse cholesterol transport and regresses ATs → antimiR-based miRNA therapeutics could be a useful strategy for treating the proliferation of VSMC associated diseases

18

Microarray data analysis

Ø The microarray experiment enables us to filter

differentially expressed genes (DEGs)

Ø Microarray data analysis resource, Bioconductor,

was used to identify DEGs.

Ø DEGs were divided into two groups ~

→ up-regulated and down-regulated

19

The treatment concept of oxLDL-induced the proliferation of VSMC associated diseases

• Fig. 1. The concept of treatment for oxLDL-induced VSMC associated diseases. ‘Up’ and ‘down’ denote the

up- and down-regulated DEGs, respectively. The dotted line denotes the interactive relationships among antimiR, miRNA and down-regulated DEGs. The arrows and the blunts denote activation and suppression respectively.

20

Methods - workflow of this study

Drug repurposing and therapeutic anti-miRNA predictions (oxLDL- induced the proliferation of VSMC associated diseases): (1) DEGs identification, (2) ClusterONE analysis, (3) GO enrichment analysis, (4) KS-test, (5) drug repurposing and effectiveness verification (IC50), (6) drug target, miRNAs and antimiRs identification

21

MTT™ cell viability test & Clonogenic assay

Ø MTT™ cell viability test To determine the effective cytotoxicity of screening drugs, MTT assay was used for cell viability and proliferation. In general, all incubated NSCLC cell lines (A549 and H460) were seeded in a 96-well microplate for up to 24 hrs dependent on the baseline growth rate. Ø Clonogenic assay We use two different high clonogenic lung cancer cell lines, A549 and H460 to perform the clonogenic assay.

22

Ø NCBI GEO microarray experiment, GSE13139

• made use of human aortic samples with gene

expression measured at five different time course

• f 0, 2, 6, 12 and 24 hours.

Ø Each time point measurement was repeated twice to examine the temporal patterns of the gene expression in response to oxLDL.

Datasets

23

Identification of DEGs

Ø Microarray data analysis resource, Bioconductor, was adopted in this study. Ø eBayes algorithm, an intrinsic function of the limma package, computes moderated t-statistics

• f

differential expression by eBayes shrinkage of the standard errors towards a common value. moderated t-statistics is defined by

24

Identification of DEGs (continued)

Ø eBayes is adopted for identifying top ranked DEGs. → DEGs with an adjusted p-values < 5% Ø up- and down-regulated DEGs were employed for searching anti-correlated drug-induced signatures → a drug found in the cMap may reverse the disease signature if the drug-induced gene expression profile is significantly negative correlated with the disease-induced gene expression profile. Ø up- or down-regulated DEGs whose expressions must satisfy below condition: Consecutively up- or down-regulated in the four time intervals, i.e. 0~2, 2~6, 6~12 and 12~24 hr.

25

ClusterONE analysis

Ø To improve the drug prediction accuracy according to IC50 measurements. → ClusterONE, clustering with overlapping neighborhood expansion, is applied for generating clusters. Ø Protein-protein interaction (PPI) information from BioGRID, ClusterONE identify both the up- and down-regulated clusters.

26

Gene Ontology (GO) enrichment analysis

Ø Functional annotation of the DEGs is given by DA VID → One can obtain the most relevant BPs and pathways terms associated with a given gene list. Ø The list of VSMC DEGs was submitted to DA VID → enriched BPs and KEGG pathways were

• btained

27

Kolmogorov–Smirnov test (KS-test)

Ø The KS-test seeks differences between two datasets, and it is non-parametric and distribution free. Ø It is conjectured that the set of consecutively up- and down-regulated DEGs may tend to cluster among the top-ranked DEGs. → we performed the KS-test for testing if this conjecture is significant or not.

28

Drug repurposing

Ø It has the potential to accelerate the development and bring down the exploitation costs for drugs, as well as reducing side effects. Ø Using cMap to find potential drugs for treating the proliferation of VSMC associated diseases. → Using both up- and down-regulated DEG sets to query the cMap database. → Potential drugs with p-values < 0.05, both enrichment scores and connective scores < 0. → Potential drugs are supported by cell viability (the IC50 values)

29

Identify drug targets, miRNAs and antimiRs

Ø Over-expressed miRNAs may down-regulated certain genes VSMC phenotype transition Ø Certain drugs may target down-regulated DEGs and reverse their expression level. → Drug-targeted are obtained by submitting the potential drugs to both DrugBank and PubChem →MiRNA targeted gene data are obtained from miRTarBase and TarBase. Ø Identify antimiRs that could hybridize the group of miRNAs which targeted down-regulated DEGs for inhibiting aberrant miRNA expression.

30

Results

DEGs and their related functional terms

Ø Obtained 7,245 DEGs, 86 of them are relevant for the formation of CVDs, 48 of them are associated with the proliferation of VSMC associated diseases Ø A website provides the following information: (i) gene list information (SABiosciences and literature review), and (ii) the 48 DEGs with their supporting references, see http://ppi.bioinfo.asia.edu.tw/vsmc/. Ø Table 1 listed these 48 genes with certain functional terms.

31

32

Results

Gene Ontology (GO) enrichment analysis

Ø DA VID can only process a limited number of input genes.

• 1. top 3,000 DEGs (adjusted p-values < 5.48e-03)
• 2. 449 up- and 452 down-regulated genes←7,245
• 3. 283 up- and 278 down-regulated genes←top 3,000

→ 1~3 were submitted to DA VID, enriched BP- and pathway-related terms were obtained. → BP or pathways with a p-value < 0.05 are kept. Ø Top two clusters of BPs and pathways information can be found in http://ppi.bioinfo.asia.edu.tw/vsmc/ .

33

Results

GO enrichment analysis (continued)

Ø Results show:

• 1. Enriched BPs or pathways are associated with

cancer- or cell regulation-related events.

• 2. Many DEGs involved in the proliferation of

VSMC associated diseases or cancer.

34

Results

Kolmogorov-Smirnov (KS) test

Ø Table 2 shows the percentages of the set of up- and down-regulated DEGs. Ø KS-test: p-value = 0.313 ← D is 0.4 → does not suggest that the consecutively up- and down-regulated DEGs tends to cluster in top-ranked DEGs

35

Results

KS-test (continued)

• Fig. 4. The comparison plots of the KS-test:

(a) KS-test comparison percentile plot (b) KS-test comparison cumulative fraction plot. Dotted lines denote the up- and down-regulated DEGs, and solid lines denote the not up- and down-regulated DEGs.

36

Results

Potential repurposed drugs

Ø 298 and 322 up- and down-regulated DEGs were analyzed → 37 drugs were identified and their information were presented in Table 3 → GW-8510, 8-azaguanine and camptothecin (8.1%) have determined IC50 activities → none of above three drugs are under clinical trials → there are ten drugs (27%) which are under clinical trials for CVD-associated diseases (Table 4). Ø Identified drugs are potential repurposed drugs for inhibiting VSMC proliferation-associated diseases.

37

38

Results

Drugs derived from ClusterONE

Ø One up-regulated cluster with 3 genes (p-value < 0.03) and three down-regulated clusters with 6 genes (p-value < 0.002), 3 genes (< 0.03) and 3 genes ( < 0.08) were combined for further analysis. Ø Three combinations, the up-regulated cluster with ~ the down-regulated cluster with p-value < 0.03 → IC50 measurement hit ratio is 22.2% (8/36). ~ the cluster with p-value < 0.08 → the hit ratio is 14.3% (6/42). ~ the cluster with p-value < 0.002 → the hit ratio is 12.5% (2/16).

39

Results

Drugs derived from ClusterONE (continued)

Ø Hit ratios of three combinations >= 12.5%, which are higher than the 8.1% without using ClusterONE. Ø 8-azaguanine had been reported for four times (three combinations, and also reported without the use

• f ClusterONE.

Ø Camptothecin and GW-8510 had been identified for three times. Ø Thioguanosine, verteporfin and azacitidine for two times. Ø Camptothecin, vorinostat and mitoxantrone showed a good IC50 in MTT assay (< 1 uM).

40

Results

Common drug targets

Ø HRH1 and DNMT1 are targeted by cinnarizine and flucytosine, respectively. ~ HRH1 is a drug target recorded in both databases. Ø SCN9A, a target of lidocaine, was obtained from both databases. Ø DNMT1 is a down-regulated DEG; HRH1 and SCN9A are up-regulated DEGs. Ø HRH1, DNMT1 and SCN9A play important roles in the proliferation of VSMC associated diseases. Ø Drugs derived from ClusterONE ~ no drug target was found (very few DEGs)

41

Results

Common drug targets (continued)

Ø Table 5 summarized the results of the identified drug targets.

Table 5. List of drug targets obtained from DrugBank and PubChem, and the common DEGs cMap Drug Target Gene from DrugBank Target Gene from NCBI PubChem DEG cinnarizine CACNA1C, CACNA1D, CACNA1F. CACNA1G, CACNA1H, CACNA1I, CACNA1S, DRD1, DRD2, DRD5, CHRM1, CHRM2, CHRM3, CHRM4, CHRM5, HRH1 ADRA2A, ADRA2B, ADRA2C, ADRA1D, CHRM1, CHRM3, CHRM4, CHRM5, DRD2, DRD3, HRH1, HTR2A, HTR2B, OPRM1, SLC6A4, SIGMAR1, SLC6A2. SLC6A3 HRH1 flucytosine DNMT1 Not available DNMT1 lidocaine SCN9A, SCN10A, EGFR SCN3A, SCN4A, SCN9A SCN9A

42

Results

miRNA-target interactions

Ø miRTarBase and TarBase → obtain miRNA – drug-target regulatory relations Ø Three miRNA-regulated module are found (Fig.3). Ø Table 6 shows, miRNAs target HRH1, DNMT1 and SCN9A, may resulted in the proliferation of VSMC associated diseases.

43

44

Results

miRNA target candidates for antimiRs

Ø We obtained one miRNA - down-regulated drug target

• regulated cluster.

Ø These 17 miRNAs represent a potential discovery

• f targets for developing antimiRs (Table 6).

→ a new therapeutic treatment concept → the therapeutic potential of miRNA regulation → key candidate role for developing the miRNA inhibitor. Ø AntimiRs as anti-disease drugs in treating CVDs.

Discussion and Conclusions

Ø By processing the query gene sets with the ClusterONE, we achieved a

higher hit ratio of drugs with measured IC50 activities.

Ø Some of the identified drugs are under clinical trial testing. Ø Several miRNA-regulated DEGs, whose miRNAs are potential targets

for developing antimiR therapy.

Ø The findings are supported by in vitro, clinical trial data and literature ,

shows the feasibility of the proposed approach.

45

Discussion and Conclusions

Ø Deficiencies or limits as follows:

• 1. Using ClusterONE to improve drug prediction

→ very few numbers of input genes → no drug target was identified

• 2. we have not benchmarked the ClusterONE algorithm with
• ther clustering algorithms.
• 3. The results are well supported; however, it is not certain whether the

cMap drugs, tested for cancer cells, may be toxic to VSMC.

• 4. At the present time, the antimiRs therapeutics have not yet been

tested in a clinical setting.

46

Discussion and Conclusions

Ø To examine how VSMC react in response to oxLDL stimulation, a

systematic study was proposed → use time-course microarray data for identifying DEGs → perform enrichment analysis for identifying the most relevant BP and pathway terms → cMap was used to identify potential drugs → use DrugBank and NCBI PubChem Compound to identify drug targets (DEGs) → miRTarBase and TarBase were used to investigate miRNA-target interactions, which regulated drug targets

47

Discussion and Conclusions

Ø In conclusion,

~ Biological networks are composed of functional related modules, which play an essential role in many BPs (cooperative interaction and induce response). ~ potential drugs were identified, and certain drugs have been tested for effectiveness by in vitro anti-CVD effects and clinical trials ~ certain miRNAs have been identified as targets for developing antimiRand they may be used as an alternative therapeutic strategy for anti-CVD.

48

Discussion and Conclusions

Ø Consequently,

~ The findings may be helpful in drug repurposing discovery and alternative therapeutic strategy. ~ Further studies are warranted to explore the effects of the potential drugs and antimiRs for the proliferation of VSMC associated diseases

49

Acknowledgement

Ø Professor Chien-Hung Huang,

Department of Computer Engineering, National Formosa University , Taiwan.

Ø Professor Chi-Ying F Huang,

Institute of Biopharmaceutical Sciences, National Yang-Ming University, Taiwan.

Ø Professor Jeffrey J. P. Tsai and Victor C. Kok,

Department of Bioinformatics and Medical Engineering, Asia University, Taiwan.

Ø Professor Nilubon Kurubanjerdjit,

School of Information Technology, Mae Fah Luang University, Thailand.

Ø Ministry of Science and Technology of Taiwan (MOST)

50

51