Elsevier

Life Sciences

Volume 146, 1 February 2016, Pages 66-72
Life Sciences

Evidence that diclofenac and celecoxib are thyroid hormone receptor beta antagonists

https://doi.org/10.1016/j.lfs.2016.01.013Get rights and content

Abstract

Long term use of NSAIDs is linked to side effects such as gastric bleeding and myocardial infarction.

Aims

Use of in silico methods and pharmacology to investigate the potential for NSAIDs diclofenac, celecoxib and naproxen to bind to nuclear receptors.

Materials and methods

In silico screening predicted that both diclofenac and celecoxib has the potential to bind to a number of different nuclear receptors; docking analysis confirmed a theoretical ability for diclofenac and celecoxib but not naproxen to bind to TRβ.

Key findings

Results from TRβ luciferase reporter assays confirmed that both diclofenac and celecoxib display TRβ antagonistic properties; celecoxib, IC50 3.6 × 10 6 M, and diclofenac IC50 5.3 × 10 6 M, comparable to the TRβ antagonist MLS (IC50 3.1 × 10 6 M). In contrast naproxen, a cardio-sparing NSAID, lacked TRβ antagonist effects. In order to determine the effects of NSAIDs in whole organ in vitro, we used isometric wire myography to measure the changes to Triiodothyronine (T3) induced vasodilation of rat mesenteric arteries. Incubation of arteries in the presence of the TRβ antagonist MLS000389544 (10 5 M), as well as diclofenac (10 5 M) and celecoxib (10 5 M) but not naproxen significantly inhibited T3 induced vasodilation compared to controls.

Significance

These results highlight the benefits of computational chemistry methods used to retrospectively analyse well known drugs for side effects. Using in silico and in vitro methods we have shown that both celecoxib and diclofenac but not naproxen exhibit off-target TRβ antagonist behaviour, which may be linked to their detrimental side effects.

Introduction

Non-steroidal anti-inflammatory drugs (NSAIDS) inhibit cyclooxygenase (COX), the enzymes that are responsible for prostaglandin production [1]. There are two isoforms, COX-1 which is constitutively expressed, and COX-2 which is inducible. NSAIDS are widely used for their analgesic, antipyretic and anti-inflammatory properties however despite their therapeutic effectiveness, their use has been widely scrutinized due to their tendency to produce side effects. Since prostaglandins protect the gastrointestinal tract and are important in platelet aggregation, NSAID reduction of prostanoid production increases the risk of gastrointestinal ulceration and bleeds. Due to the toxic effects of NSAIDs such as diclofenac on gastrointestinal mucosa, COX-2 selective drugs such as celecoxib were developed. Clinical trials revealed the side effects of both pan- and COX-1 sparing NSAIDs led to gastrointestinal damage and cardiovascular complications including myocardial infarction [2], [3].

There are currently two conflicting models that explain the cardiovascular side effects of NSAIDs. The first model put forward by Cheng et al. states that under normal physiological conditions endothelial COX-2 drives the production of prostacyclins whilst platelet COX-1 drives the production of thromboxanes [4]. The model predicts that a balance between pro-thrombotic and antithrombotic state exist under normal physiological conditions. However, when an NSAID which inhibits COX-2 in endothelial cells is introduced, the balance is disrupted and a pro-thrombotic state develops [4].

Recent evidence has emerged that provides evidence that COX-2 is not expressed in endothelial cells [5], [6], but is highly expressed in the renal medulla [7], indicating a need for a new model for what causes NSAID induced side effects to be developed. Loss or inhibition of COX-2 in mice and man leads to an increase in the production of endogenous eNOS inhibitor, asymmetric dimethyl arginine (ADMA) which suggests that specific pathways are altered by COX-2 inhibition [7].

While much debate about the side effects of NSAIDs has concentrated on the direct effects of NSAIDs on COX activity, we investigated the indirect side effects of celecoxib and diclofenac using computational chemistry methods. In silico modelling indicated a potential for both drugs to associate with thyroid hormone receptor β (TRβ), and further analysis using in vitro methods indicate that both celecoxib and diclofenac possess TRβ antagonistic properties. This nuclear receptor is of great interest, with clear relationships between hypothyroidism associated with increased heart muscle stiffness and an increased risk of myocardial infarction [8].

Section snippets

In silico methods

Open Virtual ToxLab .5,21 [9] was used to predict toxic potential by predicting binding affinities to 10 off-target nuclear receptors, 4 cytochrome P450 enzymes, a transcription factor and a potassium ion channel and forecast endocrine and metabolic disruption, some aspects of carcinogenicity and cardiotoxicity. The default values of the software for the predictions of toxic potentials for diclofenac and celecoxib were used as described previously [9].

The Pharmmapper, freely available web

In silico modelling

Using VirtualTox screening programme, structures for diclofenac and celecoxib were assessed for the potential binding to a series of target protein known to be correlated with the side effects, and a normalized toxicity potential was calculated (Table 1). The results suggest that both drugs can potentially bind all nuclear receptors, albeit with various affinities. Both drugs exhibited no affinity with CYP enzymes, arylhydrocarbon (AhR), and human Ether-à-go-go-Related Gene (hERG K). The

Discussion

Our data is the first to demonstrate that T3 induces vasodilation of rat mesenteric arteries, which can be significantly reduced by the TRβ antagonist MLS. Moreover, diclofenac and celecoxib have the capacity to bind to and antagonise TRβ and TRα receptors.

Thyroid hormones have been shown to act directly on rat aortic artery smooth muscle cells [20], rat skeletal muscle resistance arteries [21] and rat mesenteric arteries [22] to induce vasodilation in a short time frame, indicating a

Disclosure

The authors declare that there is no conflict of interest regarding the publication of this article.

Funding

The conduct of the research and preparation of the article was funded by the University of Hertfordshire, UK.

Acknowledgements

Authors thank to Biograf3R for the licence to use Open Virtual ToxLab software. We also thank Open Eye Scientific Software, Inc., for the free academic licence of the Open Eye Toolkits. We also thank Schrodinger Ltd. for their support in the final stages of the manuscript revision by providing the licence for their software.

References (36)

  • N.S. Kirkby

    LC–MS/MS confirms that COX-1 drives vascular prostacyclin whilst gene expression pattern reveals non-vascular sites of COX-2 expression

    PLoS One

    (2013)
  • B. Ahmetaj-Shala

    Evidence that links loss of cyclooxygenase-2 with increased asymmetric dimethylarginine: novel explanation of cardiovascular side effects associated with anti-inflammatory drugs

    Circulation

    (2015)
  • J. Faber et al.

    Cardiovascular disease and thyroid function

    Front. Horm. Res.

    (2014)
  • X. Liu

    PharmMapper server: a web server for potential drug target identification using pharmacophore mapping approach

    Nucleic. Acids Res.

    (2010)
  • J. Grant

    A fast method of molecular shape comparison: a simple application of a Gaussian description of molecular shape

    J. Comput. Chem

    (1996)
  • Openeye Scientific Software, Santa Fe, New Mexico,...
  • S.W. Muchmore

    The use of three-dimensional shape and electrostatic similarity searching in the identification of a melanin-concentrating hormone receptor 1 antagonist

    Chem. Biol. Drug Des.

    (2006)
  • Openeye Scientific Software, Santa Fe, New Mexico,...
  • Cited by (0)

    View full text