Discovery & Optimisation of
Novel Agonists of GPR119:
17th RSC/SCI Medicinal Chemistry
Symposium
James S. Scott
AstraZeneca
Alderley Park, UK
[email protected]
Outline of talk
•
Introduction to target and start point
•
Cyanopyridyl as a replacement for sulphone
•
Improving solubility
•
Seizure assay and a return to the sulphone…..
•
A hERG beating methylene
•
Summary and learning
Jamie Scott| 04 2012
R&D | CVGI | AP
GPR119 as a target for diabetes
•
Class A GPCR (deorphanised)
•
Expression reported in pancreas, gut, (brain)
•
Targeting an Agonist
[Shah, U. RSC Drug Discovery
Series, 2012, Chapter 7]
•
Enhanced incretin (GLP-1) secretion and direct effect on pancreas
•
‘Oral GLP-1’ concept – combination opportunity with DPPIV inhibitors
Jamie Scott| 04 2012
R&D | CVGI | AP
When something is too good to be true....
F3C
N
N
hGPR119 pEC50
hERG IC50
logD
N
S
O
S
O O
6.3 (42%)*
2.5 mM
2.9
N
N
F
O
hGPR119 pEC50
hERG IC50
logD
S
O O
7.6 (44%)
16 mM
2.1
• Robust Glucose lowering
• But also in GPR119 knock-out mice.....
• Off target activity
Stop Series
[Bioorg. Med. Chem. Lett. 2011, 21, 7310.]
*cAMP potency EC50 with intrinsic activity (% effect) relative to 50 mM oleoylethanol amide (OEA; endogenous ligand)
Jamie Scott| 04 2012
R&D | CVGI | AP
Plan B - A second series.....
No hERG and on target efficacy
A) Sulphone
O
N
O
pEC50
Sol
logD
hERG
N
N
O
N
S O
O
7.2 (83%)*
0.03 mM [c]
3.2
>33 mM
3x
increase
1000x needed
Potency in mouse similar
pEC50
7.3 (99%)
*cAMP potency EC50 with intrinsic activity (% effect) relative to 50 mM oleoylethanol amide (OEA; endogenous ligand)
Jamie Scott| 04 2012
R&D | CVGI | AP
Plan B - A second series.....
No hERG and on target efficacy
B) Pyridine
A) Sulphone
O
N
O
pEC50
Sol
logD
hERG
N
N
O
O
N
N
S O
O
7.2 (83%)
0.03 mM [c]
3.2
>33 mM
3x
O
increase
1000x needed
Potency in mouse similar
pEC50
7.3 (99%)
pEC50
Sol
logD
hERG
N
N
O
N
6.2 (71%)
23 mM [c]
3.4
>33 mM
N
30x
3x
increase
needed
Pyridyl N acting as acceptor
(but less optimal than SO2Me)
• Pharmacophoric elements are lipophile (often carbamate) and acceptor
• Struggled to move from neutrals and retain potency
Key issues to address:
Jamie Scott| 04 2012
Solubility / Potency / logD / Boc group
R&D | CVGI | AP
A) Understanding the insolubility of the sulphone
O
N
O
N
N
O
N
Sol
logD
0.03 mM [c]
3.2
S O
O
• Molecules pack – end to end via
sulphone – sulphone interactions
And in ladder interactions
• Seen in 1/4 of sulphones in CSD
End-End
Ladder
Both
None
Jamie Scott| 04 2012
18%
18%
23%
41%
R&D | CVGI | AP
Acceptor SAR – around pyridyl
B) Optimisation of pyridine
O
N
O
N
N
O
N
N
N
pEC50
Sol
logD
LLE
N
6.2 (71%) pEC50 <5.7
23 [c]
Sol
66 [c]
3.4
logD 3.7
2.8
LLE=5
• 2-substitution  potency
N
pEC50
Sol
logD
LLE
7.1(101%)
4 [c]
3.8
3.3
• 3-substitution  potency
GPR119 pEC50
LLE=4
O
LLE=3
LLE=2
N
O
N
N
R
O
N
N
• Summary of pyridine SAR with lines of
constant pEC50-logD
Ligand Lipophilic Efficiency (LLE)
Sized by intrinsic activity (10-155%)
• CN represents a breakthrough!
logD
Jamie Scott| 04 2012
R&D | CVGI | AP
Bringing it all together
O
N
O
O
N
O
N
N
O
N
pEC50
Sol
logD
LLE
6.2 (71%)
23 mM
3.4
2.8
N
N
N
7.7 (155%)
32 mM
2.8
4.9
O
N
pEC50
Sol
logD
LLE
O
N
pEC50
Sol
logD
LLE
O
N
N
N
N
O
N
N
6.7 (68%)
45 mM
3.8
2.9
 Me adds potency (& logD)
 CN adds potency (& lowers logD)
Jamie Scott| 04 2012
R&D | CVGI | AP
Bringing it all together
O
N
O
O
N
O
N
O
N
pEC50
Sol
logD
LLE
N
6.2 (71%)
23 mM
3.4
2.8
N
N
pEC50
Sol
logD
LLE
N
7.7 (155%)
32 mM
2.8
4.9
O
O
O
N
pEC50
Sol
logD
LLE
N
N
N
O
N
O
N
N
O
N
N
6.7 (68%)
45 mM
3.8
2.9
pEC50
Sol
logD
LLE
N
N
N
O
N
8.2 (171%)
21 mM
3.3
4.9
 100x potency improvement
 LogD neutral
 Solubility maintained
 Me adds potency (& logD)
 CN adds potency (& lowers logD)
Key issues to address:
Jamie Scott| 04 2012
N
Solubility / Potency / logD / Boc group
R&D | CVGI | AP
Oxadiazole as an isostere for Boc
O
N
O
N
N
N
O
N
pEC50
Sol
logD
LLE
O N
N
8.2 (171%)
21 mM
3.3
4.9
Key issues to address:
N
N
N
N
N
pEC50
Sol
logD
LLE
O N
O
N
8.1 (199%)
7uM
3.3
4.8
N
F 3C
N
N
N
N
N
O
N
pEC50
Sol
logD
LLE
N
8.7 (243%)
3 mM
4.0
4.7
Solubility / Potency / logD / Boc group
• Compound is active in vivo
in wild-type BUT NOT knock-out mouse
• Glucose & GLP-1 endpoints
On target efficacy
Jamie Scott| 04 2012
R&D | CVGI | AP
Improving Solubility #1 - An alternative to Boc.
O
N
O
N
N
N
N
O
N
pEC50
Sol
logD
LLE
O
N
O
8.2 (171%)
21 mM
3.3
4.9
O
N
N
N
O
N
pEC50
Sol
logD
LLE
O
N
5.9 (95%)
690uM [c]
1.9
4.0
O
CF3
O
N
N
N
N
O
N
pEC50
Sol
logD
LLE
7.7 (171%)
110uM
2.5
5.2
Strong ether acceptor not
satisfied by any H-bond donors
Key issues to address:
Jamie Scott| 04 2012
Solubility / Potency / logD / Boc group
R&D | CVGI | AP
N
Improving Solubility #2 - Not all oxadiazoles are equal...
Acceptor scales - logKb
[Abraham et al. J. Chem. Soc., Perkin Trans. 2, 1989, 1355.]
Subsequently shown to correlate with electrostatic potential by
[Kenny, P.; J. Chem. Soc., Perkin Trans. 2, 1994, 199.]
2.1
0.4
O N
O
N
Ph O
N O
N
N
F3C
0.2
F3C
-0.8
[Boström, J et al J. Med. Chem., 2011, 1817]
O N
F 3C
N
N
N
N
N
O
N
pEC50
Sol
logD
LLE
N O
N
F 3C
8.7 (243%)
3 mM
4.0
4.7
Key issues to address:
Jamie Scott| 04 2012
N
N
N
N
F3C
0.4
N
N
N
8.3 (236%)
5uM
3.1
5.2
1.4
N N
N
F 3C
N
O
N N
O
N
pEC50
Sol
logD
LLE
1.0
O
pEC50
Sol
logD
LLE
N
N
N
N
O
N
7.6 (220%)
120 mM
2.7
4.9
Solubility / Potency / logD / Boc group
R&D | CVGI | AP
N
Riding the line of LLE down to improve properties
Potency vs logD
N
GPR119 pEC50
N
F 3C
LLE=5
Hu
EC50
(uM)
1
O
N
F3C
LLE Mouse hERG
pEC50 PPB
IC50
LogD -logD
(%)
(uM)
N
Sol
(uM)
N
N
0.002
4.0
4.7
1.2
7
3.1 [c]
N
0.005
3.1
5.2
3.8
6
4.8 [c]
N
0.026
2.7
4.9
11
18
120 [c]
O
N
F3C
3
O
N
N
F3C
2
N
N
N
N
O
logD
• Control of lipophilicity by oxadiazole isomer
• Big impact in terms of solubility
• Benefits in terms of hERG
• Loss in potency offset by increased free levels
[MedChemCommun., 2013, 4, 95]
14
Jamie Scott| 04 2012
R&D | CVGI | AP
Problems seen in mouse toxicity studies
O N
N
N
N
N
N
O
N
N
Clonic convulsions in 2 mice (day 10 @ 300mpk) whilst undergoing an IRWIN assessment
• Seizure: synchronized abnormal neuronal activity that may result in a number of behavioral
symptoms including tonic-clonic convulsion
• Convulsion: the behavioral consequence of a seizure – typically muscle rigidity followed by rhythmic
jerking movements
• Convulsion ≠ seizure
• Abnormal motor events, syncopal events, etc. may be incorrectly mistaken for convulsions
How can we find a back-up compound?
Jamie Scott| 04 2012
R&D | CVGI | AP

In vitro Brain Slice
• Evoked population spike: Summated firing of CA1
pyramidal neurones (measure of cell excitability and
underlying synaptic activity)
• Potentially seizurogenic compounds can induce
excitatory effect by variety of potential mechanisms
[Easter A. et al., J Pharmacol Toxicol Methods. 2007, 56, 223]
Jamie Scott| 04 2012
R&D | CVGI | AP

In vitro Brain Slice
• Evoked population spike: Summated firing of CA1
pyramidal neurones (measure of cell excitability and
underlying synaptic activity)
• Potentially seizurogenic compounds can induce
excitatory effect by variety of potential mechanisms
[Easter A. et al., J Pharmacol Toxicol Methods. 2007, 56, 223]
Compound caused a
concentration-dependent
increase in excitability,
consistent with a
convulsant compound
Jamie Scott| 04 2012
R&D | CVGI | AP
Results from Brain Slice Assay
O N
N
N
N
N
N
N N
O
N
N
F 3C
O
N
N
N
N
O
N
N
+ve Brain slice
logD
2.7
+ve Brain slice
logD
3.3
pEC50
logD
Sol
hERG
7.6 (220%)
2.7
140uM
18uM
pEC50
logD
Sol
hERG
7.6 (121%)
3.0
1uM
7uM
Matched Pair
O N
F 3C
N
N
N
N
N
N N
O
N
+ve Brain slice
logD
4.0
N
F3C
O
N
N
N
F
O
N
O
S
O
-ve Brain slice
logD
3.0
• Future compounds should not contain a CN pyridyl to avoid this issue
• Impact is that achieving high solubility will be challenging.....
• & compounds may have a hERG issue.....
Jamie Scott| 04 2012
R&D | CVGI | AP
hERG can be circumvented – the CH2SO2Me!
Circumventing hERG
hERG can be difficult to predict
GPR119
hERG
- Structure driven rather than logD
- Small changes can have a big effect
‘Unlucky’ here – can we fix it?
N N
F3C
O
N
N
N
F
O
N
pEC50
logD
hERG
O
S O
7.6 (121%)
3.0
7uM
Jamie Scott| 04 2012
R&D | CVGI | AP
hERG can be circumvented – the CH2SO2Me!
Circumventing hERG
GPR119
hERG can be difficult to predict
hERG
- Structure driven rather than logD
- Small changes can have a big effect
‘Unlucky’ here – can we fix it?
N N
F3C
N N
F3C
O
N
N
N
F
O
N
pEC50
logD
hERG
O
S O
N O
F3C
Jamie Scott| 04 2012
N
N
N
CH2 spacer to polar sulphone
removes hERG liability
(potency and logD neutral)
F
O
N
pEC50
logD
hERG
7.6 (121%)
3.0
7uM
Oxadiazole isomer
raises logD (increases
potency & hERG)
O
N
N
pEC50
logD
hERG
7.6 (100%)
2.9
>33uM
N
N
F
O
N
S
O O
O
S O
8.4 (89%)
3.6
4uM
R&D | CVGI | AP
hERG can be circumvented – the CH2SO2Me!
Circumventing hERG
GPR119
hERG can be difficult to predict
hERG
- Structure driven rather than logD
- Small changes can have a big effect
‘Unlucky’ here – can we fix it?
N N
F3C
N N
F3C
O
N
N
N
F
O
N
pEC50
logD
hERG
O
S O
N O
F3C
Jamie Scott| 04 2012
N
N
N
CH2 spacer to polar sulphone
removes hERG liability
(potency and logD neutral)
F
O
N
pEC50
logD
hERG
7.6 (121%)
3.0
7uM
Oxadiazole isomer
raises logD (increases
potency & hERG)
O
N
N
pEC50
logD
hERG
7.6 (100%)
2.9
>33uM
N
N
F
O
N
8.4 (89%)
3.6
4uM
S
O O
N O
F3C
O
S O
N
N
pEC50
logD
hERG
N
N
F
O
N
8.1 (84%)
3.4
>33uM
Combination gives
good potency & hERG
R&D | CVGI | AP
S
O O
Further results from Brain Slice Assay
O N
N
N
N
N
N
N N
O
N
N
F 3C
O
N
N
N
N
O
N
O
N
O
CF3
O
N
N
F
O
N
S
O O
-ve Brain slice
logD
2.9
+ve Brain slice
logD
2.7
+ve Brain slice
logD
3.3
N
Matched Pairs
O N
F 3C
N
N
N
N
N
N N
O
N
+ve Brain slice
logD
4.0
N
F3C
O
N
N
N
F
O
N
O
S
N N
O
F3C
O
-ve Brain slice
logD
3.0
N
N
N
F
O
N
S
O O
-ve Brain slice
logD
2.9
• CH2sulphone also clean in brain slice assay
N O
F3C
N
N
N
N
F
O
N
S
O O
-ve Brain slice
logD
3.4
Sol
1.8 uM
Jamie Scott| 04 2012
R&D | CVGI | AP
Understanding the solubility of the CH2 sulphone
N O
F3C
N
N
N
N
• End to end interaction is absent
F
O
N
Sol
logD
1.8 mM [c]
3.4
S
O O
• Ladder interactions observed
(Both a-CH interacting with O)
N O
F 3C
N
O O
S
N
Jamie Scott| 04 2012
N
N
H
O
N
N O
F 3C
H
N
N
F
O O
S
H
N
N
H
O
N
F
R&D | CVGI | AP
Overall Profile looks good
Hu pEC50
Mu pEC50
logD
LLE
8.1 (84%)
7.2 (71%)
3.4
4.7
hERG
Sol
N O
F3C
>33uM
1.8 uM
N
N
N
N
F
O
N
S
O O
On-target, glucose lowering & GLP-1 release
Good Suspension PK (m,r,d) with dose linearity
Clean in Brain Slice assay
2
0
voltage (mV)
0
10
20
time (ms)
30
-2
0.1% DMSO
1 µM AZ13457680
-4
3 µM AZ13457680
10 µM AZ13457680
-6
30 µM AZ13457680
100 µM 4-AP
PSl area (% vehicle)
140
130
40
50
120
110
100
90
80
0.1
1
10
100
[AZ13457680] μM
-8
Jamie Scott| 04 2012
R&D | CVGI | AP
Synthesis – Oxadiazole isomers from N-CN
N
OR
N
HN
N
CNBr, NaHCO3,
CH2Cl2/H2O, 59%
N
N
N
OR
N
N
O N
X
N
OR
N
N
N
i) NH2OH.HCl, Na2CO3, DMF;
ii) (XCO)2O, pyridine, toluene, D
•
•
25
N O
N
Jamie Scott| 04 2012
X
N
N
N
N N
OR
N
N
i) XC(=NOH)NH2, ZnCl2,
THF/EtOAc;
ii) HCl, EtOH, D
X
N
N
O
OR
N
N
i) NaN3, Et3N.HCl, toluene, D;
ii) (iPr)2NEt, (XCO)2O,
Chlorobenzene, D
All 3 isomers available from common building block
Chemistry works on phenol (R=H) or with protection (R=Bn)
R&D | CVGI | AP
Synthesis II
N O
CF3
N
CF3
N
N
Cl
N O
CF3
O
HN
N
N O
O
N
N
CF3
O
N
Br
N
N
N O
NH
CF3
N
N
N
N
Br
N
O
N
N
O
N
N
O
N O
CF3
N
N
N
N
OH
N
N O
N
CF3
F
F
HO
MeO
MeO
O
O
O
O
O
S
O
O
N
S
OO
S
O
Convergent synthesis of 10 steps.
Longest linear sequence 6 steps
Run on >100g scale
• CH2SO2Me building block introduced as last step
• NaSO2Me displacement of bromo avoids need for S-oxidation
• Pd catalysed Br to OH conversion works in presence of heterocycle
26
Jamie Scott| 04 2012
F
O
HO
MeO
Br
N
F
F
F
N
N
R&D | CVGI | AP
Summary
O
N
N
N
O
O
N
O N
N
N
S O
O
N
F3C
N
pEC50
Sol
logD
hERG
Jamie Scott| 04 2012
N N
O
N
N
7.2 (83%)
0.03 mM [c]
3.2
>33 mM
N
N
pEC50
Sol
logD
LLE
N O
pEC50
Sol
logD
hERG
N
8.1 (199%)
7uM
3.3
4.8
N
N
F 3C
N
F
N N
N
S
O O
F3C
O
pEC50
Sol
logD
hERG
N
O
N
pEC50
Sol
logD
LLE
O
8.1 (84%)
1.8uM
3.4
>33 mM
O
N
N
N
7.6 (220%)
120 mM
2.7
4.9
N
N
N
F
O
N
O
S O
7.6 (121%)
1uM
3.0
7uM
R&D | CVGI | AP
Conclusions & Learning
Startpoint
- Biggest difference to where you end up!
- Potency lead
confidence that contradiction
- Solubility lead
could be resolved
LLE
- pEC50-logD
- good guide to progress (cpd quality)
- keeps phys props at the forefront
- Use with caution (agonists)
‘CH2’
- a small group makes a big difference
- solved hERG
Oxadiazoles
- Not all bio-isosteres are equal!
- Subtle changes have big impact
X-ray structures
- Can really assist design teams
- “Beauty is truth, truth beauty”
Jamie Scott| 04 2012
R&D | CVGI | AP
Acknowledgements
Lead Generation (Mö)
Anders Broo
Öjvind Davidsson
Udo Bauer
Linda Sundström
Birgitta Svalstedt Karlsson
Walter Lindberg
Ann-Charlotte Egnell
Chemistry
Roger Butlin
Jamie Scott
Paul Schofield
David Clarke
Jules Hudson
Sam Groombridge
Kristen Goldberg
Alan Birch
Ruth Poultney
Sue Bowker
Phys / Comp Chem
Andrew Leach
Matt Wood
Phil MacFaul
Jamie Scott| 04 2012
Bioscience
Katy Brocklehurst
Carina Ämmälä
Joanne Teague
David Laber
David Baker
Rob Garcia
Simon Poucher
Steve Bloor
DMPK
Charles O’Donnell
Hayley Brown
Pablo Morentin Gutierrez
Teresa Collins
Project Leader
Darren McKerrecher
Large Scale Chemistry
Neil Hawkins
David Rudge
Pete McLachlan
Statistics
Liz Mills
Legal
Tom Miller
Informatics
Jasen Chooramun
Animal Science & Welfare
Bill Brown
Neil Reavey
Diane Tibbs
Ruth Storer
Adrienne Sanders
Paul Sheridan
Lisa Shawcross
John Burton
Gary Slade
Laura Hayward
X-Ray
Per Svensson
Anne Ertan
Pharmaceutical Development
Claire Patterson
R&D | CVGI | AP