Acknowledgements

The research leading to these results has received funding from the European Community’s 7th Framework

Program (FP7/2007-2013) COSMOS Project under grant agreement n° 266835 and from Cosmetics Europe.

Methods

www.cosmostox.eu

Toward better understanding of liver steatosis MoA:

molecular modelling study of PPAR receptor Merilin Al Sharif 1, Petko Alov 1, Mark Cronin 2, Elena Fioravanzo 3, Ivanka Tsakovska 1, Vessela Vitcheva 1,4,

Andrew Worth 5, Chihae Yang 6, Ilza Pajeva 1

1 Institute of Biophysics and Biomedical Engineering, Sofia, Bulgaria; 2 Liverpool John Moores University, Liverpool, UK; 3 S-IN Soluzioni Informatiche Srl, Vicenza, Italy; 4 US FDA CFSAN, College Park, MD, USA; 5 European Commission, Joint Research Centre, Institute for Health and Consumer Protection, Ispra, Italy; 6 Altamira LLC, Columbus, OH, USA

merilin.al@biomed.bas.bg

itsakovska@biomed.bas.bg

pajeva@biomed.bas.bg

1. Literature search – main steps to identify key studies (Fig. 3) – all studies analysed and

ranked according to an array of carefully defined criteria; the selected studies paved the

way to describe PPARγ-dependent prosteatotic MoAs.

2. Extraction of all available PPAR complexes (118) from the PDB (http://www.rcsb.org)

3. MOE software (MOE 2012.10, http://www.chemcomp.com) used to: (i) characterise binding

pockets of the complexes; (ii) identify key protein-ligand interactions; (iii) perform

pharmacophore modelling.

Key words

Peroxisome proliferator-

activated receptor gamma

AND ("liver steatosis" OR

"fatty liver“)

220 papers

analysed Criteria

MIE

Pathway

Target protein

Endpoint

Species

SEARCH RESULTS ANALYSIS and RANKING SELECTED

Fig. 3. Main steps in the literature search to identify the major AOPs from PPAR activation to liver steatosis.

Results (1) - AOPs Results (2) – PDB analysis Four AOPs were generalised that have been shown to cause fatty liver triggered by PPARγ activation – transport of fatty acids, de

novo synthesis of fatty acids; triglyceride synthesis and lipid storage (Fig. 4);

The potential of the most studied target proteins to be starting points in a MoA leading to steatosis was evaluated and CD36, FSP27 and

aP2 were selected as prosteatotic factors downstream PPARγ signalization;

A model for the toxicological MoA of PPARγ ligand-dependent activation in hepatocytes mediated by CD36, one of the cornerstones in

the metabolic disregulation leading to fatty liver, was proposed (Fig. 5).

Fig. 6. Distribution of the structures according to the type of bound ligands.

Fig. 7. Distribution of the structures according to the form of the

receptors.

118 human PPARγ structures were

extracted from PDB.

The complexes were analyzed according to

the bound ligands and to the form of the

receptors (Fig. 6 and Fig. 7).

18 papers for all

ranked target

proteins, with the

highest number of

points for CD36

Activity data (Kd, EC50, IC50) for

30 full agonists and 26 partial agonists

were found in PDB and ChEMBLdb.

Results (3) – Molecular modelling

Fig. 1. Overview of fatty acids transport, metabolism and fate (FAT/CD36 – fatty acid

translocase/CD36; FABPpm – plasma membrane fatty acid binding protein; SLC 27A2 – solute

carrier family 27 (fatty acid transporter), member 2; SLC 27A5 – solute carrier family 27 (fatty acid

transporter), member 5; FFA – free fatty acids; TG – triglycerides; VLDL - very low-density

lipoprotein; LPL – lipoprotein lipase; LD – lipid droplet).

Fig. 4. Flow diagram of the MoA from PPAR γ ligand-dependent activation to liver steatosis (fatty liver).

Dark yellow marks CD36, which PPAR γ-mediated overexpression has been confirmed to be prosteatotic by

most experimental evidence, followed by aP2 and FSP27 (light yellow ).

Fig. 5. Model of ligand-dependent PPARγ activation as a potential MIE for liver steatosis through excessive

CD36 mediated fatty acid uptake and consequent hepatic triglyceride accumulation

The PPARγ binding pocket of the full

agonists is large, has a complex form,

can accommodate more than one

ligand and allows different binding

poses of the ligands. Polar parts of the

ligands are directed to H12 – a helix

proved to play a key role in binding of

coactivators (in cyan) (Fig. 8).

The key protein interactions of the most active

agonists include hydrogen bonding to 4/5 amino

acids in the receptor pocket either directly or

through water (Fig. 9).

The pharmacophore model outlines

hydrogen bonding, hydrophobic and

aromatic structural elements as most

important for the PPARγ binding of the

full agonists (Fig. 10, rosiglitazone – in

magenta, compound 544 - in green,

farglitazar - in gray).

Fig. 9. Specific ligand protein interactions of agonists with the key amino acids residues in the PPARγ binding pocket: A) for rosiglitazone; B) for the most active agonist in the

investigated group – compound 544.

A) B)

Fig. 8. The binding poses of all full agonists within the

PPARγ binding pocket (template complex PDB ID 1FM6) .

Conclusions

Fig. 10. Pharmacophore model of PPARγ full agonists.

Agonists of hepatic PPARγ can function as a

steatogenic inducer molecules.

Four significant AoPs for liver steatosis were

summarised triggered by PPARγ ligand-

dependent activation.

Model was proposed for toxicological MoA of

PPARγ ligand-dependent activation in

hepatocytes mediated by CD36.

Pharmacophore model was derived outlining

the importance of hydrogen bonding and

hydrophobic features for agonistic activity.

Correlation between the number of

pharmacophoric points and the agonistic

effect of the ligands (with known

experimental activity) was observed.

The results can be useful in ligand- and

structure-based screening of compounds

which binding to PPARγ could serve as a

MIE for disregulation of the PPARγ activity.

Introduction and Aims

Within the mode of action/adverse outcome pathway (MoA/AOP) framework the description and characterisation of the toxicological

MoAs leading to liver toxicity are of specific interest. Liver plays a central role in free fatty acids and triglyceride metabolism (Fig. 1).

Moreover, because of its unique function in the organism, the liver, and the hepatocyte in particular, is a major target for toxicity. Non-

alcoholic fatty liver disease is one potential repeated dose toxicity adverse effect, known to encompass both steatohepatitis - the more

aggressive form of the disease, and non-alcoholic fatty liver - grouping isolated steatosis and steatosis with mild lobular inflammation

alone. There are growing evidences for the steatogenic role of hepatic peroxisome proliferator-activated receptor gamma (PPAR), a

ligand-inducible transcription factor from the nuclear receptor superfamily (Fig. 2).

In this study AoPs from PPAR

activation to liver steatosis are

identified based on a systematic

literature analysis. Further, molecular

modelling study is performed for

the molecular initiating event (MIE) –

interaction between full agonsits and the PPAR receptor. It includes: (i) analysis of the 3D structural complexes of

human PPAR published in Protein Data Bank (PDB, http://www.rcsb.org); (ii) characterisation of the binding pocket of full agonists;

(iii) identification of the ligand-receptor interactions; (iv) development of pharmacophore models of full agonists to be

used inestablishing filtering rules for effective virtual screening of compounds with potential agonistic activity towards PPAR.

Fig. 2. Schematic structure of the functional domains of the PPAR isoforms.

PPARγ complexes

PPARγ complexes

PubMed pool Final pool