Network-based drug repositioning: A novel strategy for discovering potential antidepressants and their mode of action

https://doi.org/10.1016/j.euroneuro.2018.07.096Get rights and content

Abstract

Current target-oriented paradigm for novel antidepressant discovery has been difficult to succeed and the failures always bring huge economic losses. Although abundant ledge of disease related genes and drug action targets has been accumulated, the successful application of the knowledge for new drug discovery is limited. Here, we predicted and validated potential antidepressants and molecular targets from DrugBank recorded drugs using a novel network-based drug repositioning approach. This approach predicted relationships between drug and targets through network-based integration of drug chemical similarity, therapeutic similarity and protein–protein interactions. We predicted genome-wide relations of drugs and targets, and then screened drugs that connect to depression-related targets of known antidepressants. Six drugs were predicted and experimentally validated to have antidepressant-like effects in the tail suspension test (TST) and forced swimming test (FST) in mice. Alverine, which is a gastrointestinal antispasmodic drug, was further validated to display antidepressant-like effects in the learned helplessness and chronic unpredictable stress models of depression. Four targets, including serotonin transporter, norepinephrine transporter, serotonin 1A receptor and serotonin 2A receptor, were included in the predictable system and confirmed as primary sites of action for alverine. The results suggest that alverine may be an effective antidepressant drug and the network-based drug repositioning may be a promising drug discovery paradigm for complex multi-genetic diseases such as depression.

Introduction

Depression is a complex disease, which affects 322 million people worldwide with high rate of morbidity, recurrence, suicide, and heavy burden (WHO, 2017). The current antidepressants are often criticized for their unsatisfactory efficacy, multiple side-effects, and late onset of effect. Moreover, most of their patents have expired. Thus, there are strong clinical and commercial requirements for new drugs. Unfortunately, Bristol-Myers Squibb, Teva, and Eli Lilly and Company recently announced that they would abandon their depression projects for BMS-820836, armodafinil, or edivoxetine, respectively, in light of their negative results in phase II or III clinical trial studies.

The target-oriented paradigm is facing significant challenge in discovering novel drugs for complex psychiatric disorders, such as depression (Harmer et al., 2017, Hendrie et al., 2013). Advances in systems biology suggest that phenotypes may undergo modification when multiple biomolecules or biological processes are regulating simultaneously (Hopkins, 2007, Li, 2007, Li, 2009). In fact, evidence showed that the first-generation antidepressants with multiple targets provide better efficacy compared with subsequent antidepressants with a single target (Anderson, 1998, Anderson, 2001, Laux, 2001, Millan, 2006).

Drug repositioning is one of the most economic strategies for new drug development and is expected to decrease the cost and reduce the failure risk associated with unwanted adverse effects (Ashburn and Thor, 2004). Traditionally, drug repositioning mainly relies on infrequent “happy accidents”, whereas network-based approaches have inspired efforts to predict drug–target associations, which provides a more fast and effective method for drug repositioning (Ashburn and Thor, 2004, Bisgin et al., 2012, Guney et al., 2016). Network pharmacology uses a network approach to elucidate complex molecular mechanisms underlying actions of drugs and explore new indications for drugs (Hopkins, 2008, Li, 2007). Network-based drug repositioning aids in the understanding of the mode of action, and drugs predicted via this method may reduce risky events induced by negative efficacy results (Hopkins, 2008, Li, 2007, Li and Zhang, 2013). In our previous study, we proposed a network-based approach, which has been successfully applied in the discovery of bioactive compounds (Qi et al., 2016) and the analysis of pharmacological activities for complex herbal formulae, such as Liu-Wei-Di-Huang Pill (Liang et al., 2014) and Qing-Luo-Yin (Zhang et al., 2013) in traditional Chinese medicine. This approach can systematically unveil pharmacological activities of drugs and promote the study of drug repositioning.

In this study, we aimed to predict potential antidepressants from known small molecular drugs in the DrugBank database using a network-based framework, drugCIPHER. This approach is regarded as a representative method of emerging network pharmacology (Barabasi et al., 2011). Six drugs were predicted and then experimentally validated to exhibit antidepressant-like effects in the tail suspension test (TST) and forced swimming test (FST) in mice. Given the satisfactory pharmacological and safety profiles, alverine was selected as a candidate, and we performed in silico drug–target interaction prediction. We further validated its antidepressant activity using oral administration in the TST, FST, the learned helpless model in mice and the chronic unpredictable stress (CUS) model in rats. More importantly, we validated some key target predictions, including serotonin transporter (SERT), norepinephrine transporter (NET), serotonin 1A receptor (5-HT1AR) and serotonin 2A receptor (5-HT2AR), which may be essential for its effects. Network-based drug repositioning provides an effective method to discover a potential antidepressant, alverine, and helps us better understand its mode of action.

Section snippets

Target prediction of drug candidates

For in silico screening of antidepressants, potential targets of drug candidates in the DrugBank database were predicted by drugCIPHER, a state-of-art network-based algorithm for global prediction of drug targets developed in our previous study (Knox et al., 2011, Zhao and Li, 2010). First, a network closeness measure that describes how close are two proteins in terms of network distance is calculated in the protein–protein interaction (PPI) network. Then, linear regression models are proposed

In silico antidepressant prediction and experimental validation of six candidates

We used clustering analysis to screen drug candidates for depression. In total, 34 existing antidepressant drugs were selected using prior knowledge as 'seeds' (Table S1). The 16 antidepressants were clustered into two major groups (Fig. 1A). The two groups included 14 other drugs that might have similar drug properties with antidepressants. We then retrieved the reported pharmaceutical, pharmacological and toxicological information of the 14 drugs. Drugs that cannot cross the blood brain

Discussion

Here we present a paradigm for predicting new antidepressants and drug–target associations based on a network-based approach. Alverine, a drug approved to treat irritable bowel syndrome, was predicted and then validated to exhibit antidepressant-like effects in the TST, FST, LH model in mice and the CUS model in rats. In vitro and in vivo experiments demonstrated that alverine had moderate affinities for hSERT, hNET, and m5-HT1AR. In addition, alverine is an antagonist of 5-HT2AR and can

Role of funding source

This work was supported by the National Natural Science Foundation of China (81630103, 81225025, 91729301, 81274117, 81302761), the National Key New Drug Creation Program (2017ZX09309020).

Contributors

Shao Li and You-Zhi Zhang contributed to research design, data analysis, and manuscript revision. Ting-Ting Zhang and Rui Xue contributed to the performance of pharmacological experiments, data analysis, and manuscript writing. Xin Wang and Shi-Wen Zhao performed computational analysis, manuscript writing and revision. Lei An participated in behavioral tests. Yun-Feng Li participated in manuscript revision. All Authors contributed to manuscript writing and approved the final version of the

Conflicts of interest

The authors declare no conflict of interest.

Acknowledgments

We thank to Dr. Zhi-qing Xu of Capital Medical University for supplying the HRK293 cells stably expressing mouse 5-HT1A receptor (m5-HT1AR).

References (53)

  • I.M. Anderson

    SSRIS versus tricyclic antidepressants in depressed inpatients: a meta-analysis of efficacy and tolerability

    Depress Anxiety

    (1998)
  • I.M. Anderson

    Meta-analytical studies on new antidepressants

    Br. Med. Bull.

    (2001)
  • L. Antkiewicz-Michaluk

    Action of antidepressant neuroleptics, chlorprothixene and levomepromazine, on the central noradrenergic system: comparison with other antidepressants

    Pol. J. Pharmacol. Pharm.

    (1985)
  • T.T. Ashburn et al.

    Drug repositioning: identifying and developing new uses for existing drugs

    Nat. Rev. Drug Discov.

    (2004)
  • A.L. Barabasi et al.

    Network medicine: a network-based approach to human disease

    Nat. Rev. Genet.

    (2011)
  • M.K. Bennett et al.

    Syntaxin: a synaptic protein implicated in docking of synaptic vesicles at presynaptic active zones

    Science

    (1992)
  • H. Bisgin et al.

    Investigating drug repositioning opportunities in FDA drug labels through topic modeling

    BMC Bioinform.

    (2012)
  • S.R. Bodnoff et al.

    The effects of chronic antidepressant treatment in an animal model of anxiety

    Psychopharmacology

    (1988)
  • A.M. Coelho et al.

    Rectal antinociceptive properties of alverine citrate are linked to antagonism at the 5-HT1A receptor subtype

    J. Pharm. Pharmacol.

    (2001)
  • S.J. Corne et al.

    A method for assessing the effects of drugs on the central actions of 5-hydroxytryptamine

    Br. J. Pharmacol. Chemother.

    (1963)
  • A.P. Davis et al.

    The comparative toxicogenomics database: update

    Nucleic Acids Res.

    (2017)
  • P. Ducrotte et al.

    On-demand treatment with alverine citrate/simeticone compared with standard treatments for irritable bowel syndrome: results of a randomised pragmatic study

    Int. J. Clin. Pract.

    (2014)
  • L.E. Eiden et al.

    The vesicular amine transporter family (SLC18): amine/proton antiporters required for vesicular accumulation and regulated exocytotic secretion of monoamines and acetylcholine

    Pflugers Arch.

    (2004)
  • W.E. Fantegrossi et al.

    Interaction of 5-HT2A and 5-HT2C receptors in R(-)-2,5-dimethoxy-4-iodoamphetamine-elicited head twitch behavior in mice

    J. Pharmacol. Exp. Ther.

    (2010)
  • A.L. Garcia-Garcia et al.

    5-HT(1A) [corrected] receptors in mood and anxiety: recent insights into autoreceptor versus heteroreceptor function

    Psychopharmacology

    (2014)
  • E. Guney et al.

    Network-based in silico drug efficacy screening

    Nat. Commun.

    (2016)
  • Cited by (14)

    • A network-biology led computational drug repurposing strategy to prioritize therapeutic options for COVID-19

      2022, Heliyon
      Citation Excerpt :

      Several studies have used different network centric method for drug repurposing e.g. network-based guilt-by-association method (GBA) for drug-disease association prediction-it made suggestions to repurpose drugs based on shared treatment profile from different disease pairs [66]. Another study integrated disease-gene expression signatures, drug targets, disease information, and molecular interaction network to prioritize drug targets for drug repositioning [67]. Some studies have tried to quantify the interplay between the drug targets and the disease-specific proteins in the human interactome by a network-based similarity measure.

    • Translational Informatics for Natural Products as Antidepressant Agents

      2022, Frontiers in Cell and Developmental Biology
    View all citing articles on Scopus
    1

    These authors contributed equally to this research.

    View full text