Focused Virtual Screening Lead discovery in the Human Estrogen Receptor a presentation by Dr David Lloyd Trinity College Dublin Daylight EuroMUG 2004
1592 750 AD Biochemistry in TCD – largest Department in Country Significant research output
Molecular Design Group Centre for Structural Biology and Molecular Design Established 2004 Ireland’s first protein X-ray facility building on PRTLI and SFI investment in structural biology building on PRTLI and SFI investment in structural biology
Integrated Drug Discovery B io lo g y C h e m istry C o m p u ta tio n
Structure Based Design – looking in the ER
Structure Based Design – looking in the ER Curr Med Chem 2003, Frontiers Med Chem 2005 (in press)
� Receptor � Target: Estrogen Estrogen Receptor Target: Significance of ER • Estrogens regulate cell growth, differentiation & development of reproductive tissues in men and women. • Maintain bone density preventing osteoporosis. • Exerts anti-atherosclerotic effects which lowers Cholesterol levels. • Involved in many CNS effects (Parkinson's) and implicated in Alzheimer's. ER as a target • 60% of primary breast cancers contain ER- alpha • Estrogens are mitogenic for ER-positive breast cancer cells.
Structure Based Design – Docking in Nuclear Receptors Docking Algorithms Docking Algorithms Scoring Functions Scoring Functions
Structure Based Design – Docking in Nuclear Receptors FRED FRED PLP PLP FlexX FlexX ChemScore ChemScore Discover3 Discover3 CScore CScore eHits eHits PLC PLC In house In house
Structure Based Design – looking in the ER
Building on knowns : using receptor structural knowledge– semi-rational design
Traditional scaffold hopping –human de novo rational design J Med Chem 2001 Anti Cancer Drug Design 2001
Traditional scaffold hopping –human de novo rational design Computer-enhanced!! Benzoxepin antiestrogens OMe MeO OMe O HO Br H + n - BuLi PyHBr 3 O O O 2 3 MeO HO R 1 O Pyr.HCl aminoalkylation Heck Route O O O PhZnCl 5 7-11 6 MeO O N N N Br 8 7 9 R 1 = N O N 4 10 11 BBr 3 Suzuki Route N HO HO O Br R 2 boronic acids aminoalkylation R 2 Pd(PPh 3 ) 4 O O O 20 21-28 meta Me ( 16 ) para OMe ( 12 ) para Cl ( 17 ) ortho OMe ( 13 ) R 2 = meta NO 2 ( 18 ) meta OMe ( 14 ) para Me ( 15 ) para CN ( 19 )
Computer-enhanced human de novo rational design
Computer-enhanced human de novo rational design J Med Chem 2004 Ortho- ring substitution is tolerated - meta is not - elcectic binding mode
Let the computer decide : Virtual Screening Haystack built from 880 ‘drug-like’ compounds from WDI � 40 Cox-2 inhibitors � 40 Estrogen Receptor Modulators � 40 Histamine ‘modulators’ Active ‘needles’ introduced from a separate validated ligand set
Virtual Screening vHTS – Performance Measures – Validation hit rate observed in subset Enrichment = hit rate in database (random) Enrichment Subset Size (%) 1 5 10 15 20 Ligands 10 50 100 150 200 e.g. 1% sampled = 10 compounds. Subset - 10 actives = hit rate of 10/10 = 1.0, Hitrate in database is 40/1000 Max Actives 10 40 40 40 40 = 0.04 : enrichment = 1 / 0.04 = 25 Best Possible 25 20 10 6.7 5.0 Value
Target Database -Remove waters & Calculate PreProcessing centre of bound ligand. -Use multiple structures Docking Protocol Rescoring Generation of Hits Active set of compounds for development
In-house docking protocol • Samples search space and generates a set of binding poses for each ligand conformer. Docked positions have their respective • hydrogen bond lengths optimized to allow for refinement of the final structure. • CF (Complementarity Function) evaluates fit • Ranks these modes/ligand positions • Provides a numerical score allowing for ‘hit’ identification
Chemscore Getting it right – Scoring Functions performs best of scoring functions. Actual Hits Retrieved F_Score Accounts for: G_Score 35 Hydrogen bond PMF_Score 30 contacts, D_Score 25 Lipophilic % Hits Retrieved ChemScore contacts, 20 entropic Xscore 15 penalty. PLP_Score 10 Fresno G-score 5 focuses on Screenscore hydrogen 0 Hammerhead bonding 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Theoretical Maximum Hits % Sample Database interactions Retreived only for example.
Getting it right – Scoring Functions Sybyl6.91 Name_ID CHEMSCORE Name_ID CHEMSCORE FRED (ESTR0025) -107.73 ESTR0079 -46.568 (Drug_158) -104.85 ESTR0072 -44.427 (Drug_159) -103.42 Drug_344 -42.485 (Drug_633) -103.35 Drug_353 -42.209 (Drug_160) -103.16 Drug_440 -42.169 (Drug_474) -102.62 Drug_219 -41.647 (ESTR0046) -101.03 Drug_823 -41.635 (ESTR0034) -100.14 Drug_416 -41.031 (Drug_163) -97.1 Drug_249 -40.991 (ESTR0043) -96.64 Drug_217 -40.687 (Drug_161) -96.33 Drug_257 -40.389 (ESTR0045) -96.18 Drug_315 -40.154 (ESTR0024) -96.07 ESTR0068 -40.123 (ESTR0085) -95.72 Drug_265 -39.874 (Drug_466) -94.53 Drug_421 -39.624 (Drug_476) -94.5 Drug_259 -39.176 (Drug_751) -94.43 Drug_61 -39.112 (Drug_472) -94.24 ESTR0067 -39.083 (Drug_401) -94.16 COX20080 -38.208
Getting it right – early method validation Hit Retrieval 40 35 30 % Hit Retrieved 25 In House Protocol FlexX 20 FRED Best V alue 15 10 5 0 1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 % Database Screened
Getting it right – Ligand Pre Processing Virtual High-Throughput Screening <1 sec per compound – rigid/rigid system Corina Omega Rubicon QuacPac Corina Omega Rubicon QuacPac Stergen Stergen
Getting it right – Ligand Pre Processing X Quac-Ster-X 40_CORINA_LEVEL1 0.54 40_CORINA_LEVEL4 0.64 40_OMEGA_LEVEL1 0.48 40_OMEGA_LEVEL4 0.70 40_CATALYST_LEVEL1 0.57 40_CATALYST_LEVEL4 0.63 40_RUBICON_LEVEL1 0.47 40_RUBICON_LEVEL4 0.74 Quac-X Quac-Ster-X-10 Confs 40_CORINA_LEVEL2 0.64 40_CORINA_LEVEL5 0.69 40_OMEGA_LEVEL2 0.66 40_OMEGA_LEVEL5 0.69 40_CATALYST_LEVEL2 0.62 40_CATALYST_LEVEL5 0.73 40_RUBICON_LEVEL2 0.64 40_RUBICON_LEVEL5 0.74 Random screening – 40 actives in 1000 – Quac-X-10 Confs each active returns a score – the bigger 40_CORINA_LEVEL3 0.69 the difference between the active and 40_OMEGA_LEVEL3 0.71 inactive scores, the better the method 40_CATALYST_LEVEL3 0.69 40_RUBICON_LEVEL3 0.69 Preprocessing can increase the cutoff value for ligand consideration – reducing the subset we must consider in order to find our active ligands.
Getting it right – Ligand Pre Processing Avg Subset % 1 2 3 4 enrichment LEVEL1_CORINA 22.5 16.25 15 11.88 10 LEVEL1_OMEGA 22.5 13.75 10 8.75 7 LEVEL1_CATALYST 17.5 12.5 9.17 7.5 6 LEVEL1_RUBICON 20 16.25 12.5 8.75 6.5 LEVEL2_CORINA 25 20 18.3 11.25 7.5 LEVEL2_OMEGA 22.5 18.75 14.2 11.25 9.5 LEVEL2_CATALYST 15 8.75 6.66 5.625 5 LEVEL2_RUBICON 10 13.75 14.2 13.13 11.5 LEVEL3_CORINA 25 21.25 15 11.875 9.5 LEVEL3_OMEGA 25 21.25 15 11.25 9.5 LEVEL3_CATALYST 25 20 13.33 10.625 9 LEVEL3_RUBICON 22.5 18.75 14.16 10.625 9 LEVEL4_CORINA 25 21.25 15 11.25 9.5 LEVEL4_OMEGA 20 16.25 13.33 10.625 9.5 LEVEL4_CATALYST 17.5 10 8.33 6.875 6.5 LEVEL4_RUBICON 22.5 18.75 14.17 13.75 11 LEVEL5_CORINA 25 21.25 15 11.25 9.5 LEVEL5_OMEGA 25 20 14.17 9.375 8 LEVEL5_OMEGA 25 18.75 13.33 11.25 9 LEVEL5_RUBICON 25 22.5 16.66 13.75 12
Does it really work ?– Validate, Validate, Validate Screen ligands, prepare ranked hitlist cluster hits – 20 clusters 18 purchased and assayed 5 Hits µ m range 5 Hits µ m range µ µ µ µ µ µ Compound Number IC50 in MCF-7 MTT 4 Chemical Classes 4 Chemical Classes MDG-ER-001 8.23E-07 MDG-ER-002 8.00E-06 3 novel Chemotypes 3 novel Chemotypes MDG-ER-003 2.02E-05 MW 450- -550 550 MW 450 MDG-ER-004 5.59E-04 MDG-ER-005 6.06E-04 LOGP 4.8- -6.5 6.5 LOGP 4.8 TAMOXIFEN 5.51E-06
What else do we need? Familial scoring functions Flexible systems – dynamics in docking Chemical intelligence in fragment assembly System Simulation Tiered Discovery – integration of technologies Validation Validation Validation
Acknowledgements The ER collaborators Dr Mary Meegan – School of Pharmacy TCD Dr Vladimir Sobolev – Weismann Institute, Israel Prof James Sexton – Trinity Centre for High Performance Computing Prof Clive Williams – Biochemistry TCD Dr Daniela Zisterer – Biochemistry TCD Dr Amir Khan – Biochemistry TCD The workers The facilitators Andy Knox Dermot Frost Yidong Yang Giorgio Carta Valeria Onnis Georgia Golfis
Recommend
More recommend
