$NIGMS $FDA The Central Dogma of Molecular Pharmacology (~1985 to - - PowerPoint PPT Presentation

▶

Dec 14, 2023 94 likes •382 views

Bryan Roth Brian Kobilka, Peter Gmeiner Matt OMeara, Josh Pottel Henry Lin, Anat Levit Kate Stafford Magdalena Korzynska John Irwin $NIGMS $FDA The Central Dogma of Molecular Pharmacology (~1985 to present): Target Ligand Docking for

SLIDE 1

Matt O’Meara, Josh Pottel Henry Lin, Anat Levit Kate Stafford Magdalena Korzynska John Irwin Bryan Roth Brian Kobilka, Peter Gmeiner

$NIGMS $FDA

SLIDE 2

The Central Dogma of Molecular Pharmacology (~1985 to present): TargetàLigand

SLIDE 3

Docking for new chemotypes from large libraries

107 available compounds 105 complexes/molecule (~1012 complexes overall) Test high-scoring molecules (but which ones, exactly?)

SLIDE 4

Docking for novel agonists with new µOR biology

3.5M cmpds

Test

N OH O OH

Manglik, Lin, Aryal…Nature 2016

7/23 tested hit, 2 to 14 uM

N N H O N H N Cl

HO N N H O N H S

Ki 1 nM, EC50 4 nM

SLIDE 5

docking screens vs GPCRs: 17 to 58% hit rates, nanomolar activities, novel ligands

3x106 cmpds

Test

docking x-ray

b2AR: 25%, 0.01 to 3 uM

Kolb, PNAS 2009

A2a: 35%, 0.2 to 3 uM

Carlsson, JMC 2010

D3 model: 23% 0.2 to 3uM

Carlsson, Nat Chem Biol 2011

CXCR4:17%, 0.3 to 30uM

Mysinger, PNAS 2012

muscarinic: 58% 0.4 to 40uM

Kruze, Mol Pharm 2013

a probe for the

rphan GPR68

Haung, Nature 2015

a probe for the

rphan MRGPRX2

Lansu, unpub

SLIDE 6

(some) inaccuracies in docking

DGinteract - DGsolv, L - DGsolv, R

= DGbind

S(qi Pi + viPv)

= DGinteract

((1/D0 - 1/Dw)/2r SQ2i -Born)*aidV/S1/r4ik+ ΔHnp+ IST =DGsolv

Lennard-Jones Potential

5.000

20.000 0.00 0.50 1.00 1.50 2.00 2.50 3.00 Radius (Å) Energy (kcal/mol)

6 12

) ( r B r A x f

X = 1.50Å

PB electrostatics grid

Charges & params for 107 diverse molecules internal energies Relaxation Hydrophobicity (solvent effects) Water displacement...

NH N+ O NH

O O-

NH N+ O NH

SLIDE 7

The data are sparse, the space is big

<105 crystallographic ligand complexes
106 ligand-protein affinities measured
Many badly measured, many artifactual
Over 1062 possible drug-like molecules
Small differences in chemical structure can matter

N N

N N+

N N N HO N N HO

N H N O

SLIDE 8

Opportunities to collaborate: model systems, late state prioritization, next compound…

simplified sites, readily tested at atomic resolution 104 (…105 … 106 ) good docking hits, which to test? Optimize?

SLIDE 9

The upside down world of classical pharmacology (1930 to 1985: ligandsàtargets)

a adrenergic b adrenergic

HN OH O

O NH2 O OH HN

b2 adrenergic b1 adrenergic

H N OH HO HO

> > > >

NH HO HO HO NH2 HO HO HO

HN HO HO HO

NH HO HO HO

HN HO HO HO

NH2 HO HO HO

SLIDE 10

Relate targets by ligand similarity (~106 ligands for ~2500 targets)

SLIDE 11

3 Test at Novartis

Z-score Frequency drug vs. tA drug vs. tB

Statistical model

1 For each of 656 drugs,

2 Compare to ligands of 73 Targets

… ,

656 drugs predicted vs 73 side effect targets

E. Loukine, Nature 2012

SLIDE 12

0 to 47888 possible drug-target pairs. 1241 novel ones predicted, 1042 tested at Novartis

Lounkine, Nature 2012

48 77 348 65 478 26

in proprietary dbs unknown to SEA Confirmed <30 uM Confirmed <10 uM Confirmed < 1 uM Inactive Ambiguous

SLIDE 13

Example ADR targets (26% cross domain boundaries)

Loukine, Nature 2012

Drug Closest known SEA E- value Off- Target Known Target BLAST E Kd (nM) ADR

N O N NH N

Alosetron

N N O F

Tc 0.25 1.6e-17 5HT2b

5HT3

>>1 20 sedation

F F N N HN O

Pimozide

F O N N O NH O

Tc 0.43 3e-7 a1a 5HT7 4.8e-53 207 Ventricular arrhythmia

O O O Cl

chlorotrianisene

O O OH N O Cl

Tc 0.31 1.9e-17 COX1 Estr. Rec >>1 221 abdominal pain

N O Cl

Clemastine

NH O N HO Cl

Tc 0.32 1.1e-14 SERT hERG >>1 419 Palpitations, Insomnia

SLIDE 14

Reorganize the GPCR-ome by ligand similarity

A A B C

…

1 For each GPCR 2 ligand set 3 vs. the ligand sets for 3000 targets

4 Rank by significance

Z-score Frequency

A vs. B

A vs. C

Statistical model

? ?

Lin, Nature Methods, 2013 Keiser, Nature Biotech 2007

SLIDE 15

GPCRs by orthosteric seq. id GPCRs by ligand similarity

Bioamine Peptide Lipid Purine Adenosine Melatonin

Gloriam, JMC 2009 non-GPCRs w/GPCR- like ligands Lin, Nature Methods 2013 6 predicted crosses tested, confirmed

SLIDE 16

A 250 nM casein kinase1 inhibitor predicted to modulate CXCR2 (E-value 1.3 x10-15, EC50 780 nM)

EC50 = 0.78 uM

Lin, Nature Methods, 2013 (Bryan Roth, Flori Sassano)

SLIDE 17

An epoxide hydrolase inhibitor predicted to modulate CB2 (E-value 1.3 x10-15, EC50 2.6 uM)

Lin, Nature Methods 2013 WIN 55,212-3

CB2 Inhibition by Epox hyd inhib

SLIDE 18

How general is this?

SLIDE 19

Ligand similarity reorganizes the genome NHRs by seq. ID…by ligand similarity

Matt O’meara Sarah Barelier

SLIDE 20

Ligand similarity reorganizes the genome: LGICs by seq. ID… by ligand similarity

SLIDE 21

What is going on?

SLIDE 22

How unrelated receptors bind identical ligands

Sarah Barelier et al, ACS Chem Biol 2015

SLIDE 23

A “Metabolic Code”:

primary messengers fixed, signal in multiple time domains

GM Tomkins, Science 1975

SLIDE 24

GPCR neurotransmitter polypharmacology (the drugs mimic the neurotransmiters)

Lin, O’Meara, Barelier, unpub

DA 5HT Adrenaline neurotensin ATP adenosine ach sphingosine leukotriene prostacyclin

SLIDE 25

Nuclear hormone polypharmacology (the drugs mimic the hormones)

O’Meara, Barelier, unpub

SLIDE 26

O’Meara, Barelier, unpub

LGIC neurotransmitter polypharmacology

SLIDE 27

Opportunities : better fingerprints, bioinformatic +chemoinformatic integration, what disease?

Matt O’Meara

48 77 348 65 478 26

in proprietary dbs unknown to SEA Confirmed <30 uM Confirmed <10 uM Confirmed < 1 uM Inactive Ambiguous

New fingerprints, methods to do better than 50% predictive true positives, 4% predictive false negatives What polypharmacology exactly should be targeted?

HO HO N N H O N H O N OH OH