Chemoinformatics in Drug Design - DTU Bioinformatics · Chemoinformatics in Drug Design Biological Sequence Analysis, ... In silico Databases Canonical SMILES: ... Nat Rev Drug Discov

Chemoinformatics in Drug Design

Biological Sequence Analysis, June 8, 2010

Irene Kouskoumvekaki, Associate Professor, Computational Chemical Biology, CBS, DTU-Systems Biology

2 CBS, Department of Systems Biology

Computational Chemical Biology group

Irene Kouskoumvekaki

Associate Professor

Olivier Taboureau

Associate Professor

Sonny Kim Nielsen

PhD student Jens Eric Pontoppidan Larsen

PhD student

Honey Polur

MSc student


Competences


A small molecule drug ... is a (ligand) compound that binds to a biological

target (protein, enzyme, receptor, ...) and in this way either initiates a process (agonist) or inhibits the natural signal transmitters in binding (antagonist)

The structure/conformation of the ligand is complementary to the space defined by the protein’s active site

The binding is caused by favorable interactions between the ligand and the side chains of the amino acids in the active site. (electrostatic interactions, hydrogen bonds, hydrophobic contacts...)


Drug discovery process

Screening collection

HTS

Actives

103 actives 106 cmp.




HTS

Actives


High rate of false positives !!!

High throughput is not enough … to get high output…..




HTS

Actives


Follow-up Chemical structure Purity Mechanism Activity value




HTS

Actives


Follow-up

Hits

1-10 hits




HTS

Actives


Follow-up

Hits

1-10 hits

SAR

Analogues synthesis and tesiting




HTS

Actives


Follow-up

Hits

1-10 hits

Lead series

0-3 lead series

Hit-to-lead




HTS

Actives


Follow-up

Hits

1-10 hits

Lead series

0-3 lead series

Hit-to-lead

Drug candidate

0-1

Lead-to-drug




HTS

Actives


Follow-up

Hits

1-10 hits

Lead series

0-3 lead series

Hit-to-lead

Shift on time requirements

Drug candidate

0-1

Lead-to-drug

In vivo experiments

ADMET properties

Selectivity profile Safety


Failures

Clinical efficacy (53%)

Side effects and toxicity (35%)

Pharmacokinetics (4%)

Portfolio (4%) Other (4%)

Phase III failures 1992 – 2002

Schuster et al, Curr. Pharm. Des. 2005, 11, 3545-3559


Failures


Drug discovery

• Diverse set of molecules

HTS Virtual Screening • Computational methods to select

subsets based on prediction of drug-likeness, solubility, binding, pharmacokinetics, toxicity, side effects, ...


The Lipinski ‘rule of five’ for drug-likeness prediction Octanol-water partition coefficient (logP) ≤ 5 Molecular weight ≤ 500 # hydrogen bond acceptors (HBA) ≤ 10 # hydrogen bond donors (HBD) ≤ 5 If two or more of these rules are violated, the compound might

have problems with oral bioavailability. (Lipinski et al., Adv. Drug Delivery Rev., 23, 1997, 3.)

Rules have always exceptions.

(antibiotics, antibacterial and antimicrobials,…)


Quantitative Structure Activity Relationships (QSAR)



In silico Databases

Canonical SMILES:

C1=CC=C2C(=C1)C(=O)NS2(=O)=O -OEChem-06071012303D

17 18 0 0 0 0 0 0 0999 V2000

-1.6163 -0.8146 0.0000 S 0 0 0 0

-2.0135 -1.3702 1.2749 O 0 0 0 0

-2.0120 -1.3733 -1.2741 O 0 0 0 0

-0.5478 2.8682 0.0022 O 0 0 0 0


BioAssay: Bioactivity screens of chemical substances

Compound: Unique chemical structures

Substances: sample descriptions

Yesterdays count: 433,863 BioAssays 27,111,073 Compounds (R05: 18,917,923)


Descriptors


fingerprints

1D descriptors:

MW, number of features,…

2D descriptors:

Topological, physichochemical,

BCUT,…

2D/3D pharmacophores

Descriptors


Choosing the right descriptors can be tricky…

Wolfgang Sauer, SMI 2004


Virtual Screening

• Computational techniques for a rapid assessment of large libraries of chemical structures in order to guide the selection of likely drug candidates.

• Exploit knowledge of target(s) and/or active ligand(s) and/or target family.

Similarity-based /

Pharmacophore-based LIGAND INFORMATION

Docking PROTEIN STRUCTURE


Similarity-based VS


Identification by VS of new Generation of Bacterial Biofilm Inhibitors > 60% of all infections are related to biofilm formation

Scanning Electron Micrograph of a Pseudomonas aeruginosa Biofilm found on

a Daily Wear Soft Contact Lens Dürig A., et al. Appl Microbiol Biotechnol, (2010) EPub


Green tea, Camellia sinensis

Has been used for over 5000 years

Green tea is known to have health enhancing qualities


Virtual Screening of Natural Compound Database for new anti-biofilm compounds

Fisetin inhibits biofilm formation of both S. aureus and Strep. dysgalactiae at ~ 10 fold lower concentrations than the 1st and 2nd generation queries


Pharmacophore-based VS


Discovery of selective PPAR-γ ligands

Kouskoumvekaki I. et al. , Petersen RK. et al. In preparation


Docking

• Explore the binding sites of new crystal structures

• Explore the effect of mutations on the binding affinity

• Search of new ligands


Docking analysis of identified PPAR-γ ligands

Kouskoumvekaki I. et al. , Petersen RK. et al. In preparation


From Chemoinformatics to…


Systems Chemical Biology

S. Berger and R. Iyengar, Bioinformatics, 2009, 25(19), 2466-72




• drug repurposing

• side effect, toxicity

• new druggable targets

• effective drug combinations





Disease – Target – Drug Network


From Network Visualization to Biological Activity Prediction

Comp 1 Comp 2

Target


Chemical-Protein Network www.cbs.dtu.dk/services/Chem-Prot-1.0

Chemical libraries

600,000 chemicals annotated to 15,800 proteins

Inweb

(941629 ppi)

Small compound – protein – disease associations

Taboureau O. et al. , In preparation


Recent publications using Systems Chemical Biology Approaches

• Yildirim MA et al., Nature Biotechnology (2007), 25, 1119-1126 (Drug-Target-Disease Networks)

• Campillos M et al, Science (2008) 321, 263-266 (Side Effect similarity – Target Identification)

• Keiser MJ et al. Nature, (2009), 12(462), 175-81 (Drug-Target Networks)

• Hansen NT et al. Clinical Pharm & Therap. (2009), 86(2), 183-9 (PGx Gene Prediction)

• Adams JC et al. PLOS Comp Biol (2009), 5(8) (Drug-Metabolite Networks)

• Chen B et al. J Chem Inf Mod (2009), 49(9), 2044-55 (BioAssay Networks - Target Prediction)

• Iorio F et al., J Comput Biol (2009), 16(2), 241-251 (Gene Expression Profiling of Drug Action)

• Qu XA et al., BMC Bioinformatics (2009), 10(Suppl 5):S4 (Drug-Disease Networks)

• Schadt EE, et al, Nat Rev Drug Discov (2009), 8, 286-295 (Perspective on Drug-Disease Networks)

• Xie L et al., PLOS Comp Biol (2009), 5(5) (Drug-Target Networks – Side Effects)

• Yamanishi Y et al., Bioinformatics, (2010), 26, i246-254 (Drug-Target Predictions)

• Audouze K et al., PLOS Comp Biol (2010), 6(5) (Toxicogenomics Networks)

• Suthram S et al., PLOS Comp Biol (2010), 6(2) (Human Disease Network)


Questions?

Chemoinformatics in Drug Design - DTU Bioinformatics · Chemoinformatics in Drug Design Biological Sequence Analysis, ... In silico Databases Canonical SMILES: ... Nat Rev Drug Discov

Documents