Session 1 part 3

1

An overview of the drug discovery process

“Hit to Lead”

Nature Review Drug Discovery,8, 892 2009.

2

From Hit to Lead

Hits from HTS screening- may have many potential scaffolds

Hit-to-lead involves synthesis of many compounds to determine what is important

Need to see if there is room to improve the compound

NN

N

N NH

Synthesis

HTS HIT/Natural Product

Essential scaffold

Synthesis

Potential lead compound

3

Hit to lead – fragment evolution

Nature Reviews Drug Discovery 3, 660-672 (August 2004)

Fragment evolution – aided by structure of fragment in the protein

Essential fragment

Synthesis to increase potencyPotential lead

compound

N

N

N

Hits from Fragment based screening- may have many potential scaffolds

Hit-to-lead involves synthesis to expand the core to move from binding to activity

Most efficient when aided by structure-based methods

4

From Hit to Lead

For a hit to become a lead it must:

Show structure-activity relationships (SAR)

Activity should be sensitive to structure

Losing activity is NOT a negative result!

The compound should have handles for reactivity

Able to modify

Most scaffolds are retained during optimization

Compounds should be simple

Stereocenters = cost

Should show activity in a cellular assay (or in vivo)

Can your hits get into a cell or a target tissue?

Should show lead-like molecular properties

Expedite and simplify further optimization

5

Lead Optimization

Nat Rev Drug Disc 2, 369-78, 2003

Medicinal chemistMedicinal chemist

In vivo efficacy is key

6


7

Medicinal Chemistry Refinement

Synthesis of compounds

Screen for activity AND/OR

Screen against activity AND/OR

Screen for ADME

Data Analysis (SAR trends)

Refinement of criteria

Planning

Many compounds must be made! What are the strategies used for efficient synthesis? What tools

are in the chemists’ synthetic toolbox?

8

Approaches to synthesis - discovery

Compounds are made in bunches, not as single efforts

The more molecules made at once, the better to understand trends in efficacy, physicochemical properties, etc.

If one compound fails to show the expected in vivo pharmacology, others are there to fall back on-

Is it the scaffold?

Is it the target?

Without a variety of lead compounds, you won’t know!

Compounds may show similar activity, but vary greatly in selectivity, or ADME properties

Making series of compounds helps to spot trends to guide future research

Parallel synthesis of groups of compounds made by facile reactions from a common intermediate

Allows response to biological data with the shortest turnaround time possible

9

A case study for library design

R. J. Gillespie et al. / Bioorg. Med. Chem. 17 (2009) 6590–6605

N

Cl

ClHO

O

A diversifiable scaffold with three synthetic handles

N

Cl

ClHO

O

N

Cl

ClN

O

R1

R

Facile coupling reactions with commercially available

amines create a library to explore space around this

position

N

R2

ClN

O

R1

R

The more reactive chloride can be

replaced with various groups through carbon-carbon bond formation

N

R2

XN

O

R1

R

The chloride can be substituted with various heteroatoms and groups

Straightforward chemistries and commercial reagents allow for rapid diversification

Prioritization is necessary

10


11

Synthesis of an active pharmaceutical ingredient (API)

Syntheses that are scalable from gms to kgs

Syntheses that avoids metals, such as Pd

Metal impurities must be minimal in the final compound

Removal of metals can be very expensive

Syntheses that can be purified easily

Salt forms are often used as APIs due to their greater stability and solubility

As the focus of chemistry efforts shift from making a library of many compounds to making large amounts of one compound, strategies change

12

Discovery synthesis vs API synthesis: A case study

The chosen compound 5 has a methyl group added in the last step via a Pd catalyzed reaction as part of a parallel chemistry scheme

13

Synthetic scheme for compound 5 as an API

W. Hu et al. / Bioorg. Med. Chem. Lett. 17 (2007) 414–418

Methyl group is set early in the synthesis via a

cyclization reaction

“Green chemistry”

14

Summary

The path to drug discovery begins with the selection of the library picked for screening

Libraries should be chosen for the same reasons that compounds are chosen later in development

There are a variety of complimentary ways to get hits

Optimization of hits toward clinical candidates

Increase of potency and selectivity

Increase of in vivo efficacy

Maintenance of potency and selectivity; optimization of other factors

Incorporation of drug-like molecular property filters in the front end of discovery facilitates this process

Chemists use standard tools in drug discovery regardless of the therapeutic area

Pattern recognition

Parallel chemistries

15

Conclusions

Many factors influence all steps of drug discovery, from choosing how to find a hit to choosing a clinical candidate

Drug discovery chemistry works to find compounds that are potent and selective with ADME properties that forecast in vivo efficacy in the clinic

Discovery synthesis and design should be efficient and make the best compounds possible to guarantee success

Chemistry efforts are led by biological results

Constant communication and feedback between team members of different disciplines gives the best chance to overcome the many obstacles and to succeed in the discovery of an efficacious drug

16

Thank you for your attention!

17

18

A structure – toxicity study - A2A antagonists

N N

N

N

NH2

O

N

N

O

H3CO

A2A binding: 2.8 nm A1 binding: 601 nm

3mg/kg p.o. efficacious in vivo for anti-cataleptic

activity

Molecular Weight: 449.51log P: 3.33

tPSA: 100.51

hERG inhibition of 81%

Maintain potency and selectivity while decreasing hERG % inhibition

Molecular Weight: 447.53Log P: 2.83tPSA: 91.28

J. J. Matasi et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3670–3674J. J. Matasi et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3675–3678

Natural Products as Drug Starting Points

Frank E. Koehn6th Drug Discovery for NeurodegenerationFebruary 13th , 2012New York, NY

Confidential and Internal to Pfizer - subject to works council and/or union consultations and other legal requirements.

HO

O

NOOO

O

HOO

O

O

OH

R

O

Just What in Fact, is a Natural Product?

~300,000 distinct compounds from microbes, plants, and other organisms

FK-506- fujimycinStreptomyces tsukubaensis

palytoxinPalythoa tuberculosum

Cl

OH OHO O

CONH2

OH

OH MeH

NMe2

OH

aureomycinStreptomyces sp.

N

N

H

nicotineNicotiana tabacum


Natural products- A major impact on drug discovery

Liberal analysis - 47% of New Chemical Entities 1940-2006 are “Natural Product Derived”

Conservative analysis - 1970-2012 - 58 approved NCE’s came directly from natural products

10% of all drugs over last 10 years (19 of 200)

Native molecules- 27, analogues- 31

Sources: microorganisms> Plants>> marine sources

Unique Challenges with NPs. Accessibility- synthetic manipulation

NP extracts- Isolation is slow, resource-intensive

Pure NP libraries- difficult to enable

J. Med Chem. 2009, 52 1953-1962, Curr. Opin. in Chem. Biol, 2008, 12:306-317 21


Targets, Libraries and Screening Strategies

• Chemical Space - Exceeds 1060 compounds with less than 500 MW

• Not all chemical space is biologically relevant!

• To screen effectively- screen the biologically relevant part of chemical space

• Natural products are privileged (biased to occupy biologically relevant chemical space) Predicted score plot of NP and

medicinally active WOMBAT compounds.

Rosen, et. al.,J. Med. Chem. 2009, 52, 1953–1962


Screening for Lead Generation

Compounds

Biochemical HTS(Single target)

Target-compound binding

Phenotypic Screening(many targets)

NP chemical Library

Phenotypic response

New target & mechanism

Cell

Target


minutes

AB

SO

RB

AN

CE

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34.

Media componentspolar metabolites& biopolymers

Lipids, fatty acidsnon-polar biopolymers

Crude Extract Library

Fractions/extract

Library size per culture

Low

Assay interferences

High

Sample prep Low

Redundancy High

Hit identification Slow

Sensitivity 10X

Pre-fractionated Library

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 0

Moderate

Moderate

Moderate

Moderate

Moderate

100X

2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34

Pure Compound Library

Moderate

Low

High

Very Low

Rapid

10 Liter Fermentation

100 Liter Fermentation

optimized

Screening and Natural Products Library Design


The Challenge- Tougher Targets

The Rule of Five (Ro5) has guided the design of compounds into privileged ADME space

MW <500 Da

ClogP <5

HBD <5

HBA (N, O) <10

Good Fraction Absorbed

(Solubility, Permeability)

Low Clearance

Oral

Bioavailability

Excellent strategy for many targets…..

But not for targets involving protein-protein interactions

25


The “Druggable” Genome - Hopkins

26

Highly “Druggable” targets, Ro5

leads

Disease relevant

“Undruggable” biological targets,

Beyond Ro5 leads

Very Limited Overlap

Hopkins, A.L., Groom, C.R. “The druggable genome” Nat. Rev. Drug Discov., 2002, 1(9) 727-30.

Confidential and Internal to Pfizer - subject to works council and/or union consultations and other legal requirements. 27

NP Lead, year NCEs Indication/MOA MW ClogP HBD HBA Oral Bioavaila

bilityDose

Validamycin, 1970

Acarbose, 1990Voglibose, 1994

Anti-diabetic/glucosidase inhibitor 498 -6.2 13 14 25 mg

Midecamycin, 1971 Miocamycin, 1985 Antibacterial/protein

synthesis inhibitors 815 3.5 4 16 100% 600 mg

Rapamycin, 1974

Sirolimus, 1999Everolimus, 2004Zotarolimus, 2005Temsirolimus, 2007

Immune suppression/mTOR 914 7.0 3 14 20% 2 mg

Cyclosporine A, 1975 Cyclosporine, 1983 Immune suppression /IL-2

inhibitor 1203 14.4 5 23 30% 25 mg

Lipstatin, 1975 Orlistat, 1987 Obesity/Lipase inhibitor 492 7.6 1 6 120 mgAvermectin B1a, 1979 Ivermectin, 1987 Antiparasitic/Glutamate-

gated chloride channel 873 5.1 3 14 100% 3 mg

FK506, 1984 Tacrolimus, 1993Immune suppression/T-lymphocyte activation inhibitor

804 5.8 3 13 20% 1 mg

Myriocin Gilenya, 2010 Multiple sclerosis/S1P1 inhibitor 402 2.8 6 7 93% 0.5 mg

Natural Products are Successful Therapeutics in the Beyond Ro5 Space

Selected Orally Active BRo5 Natural Product Drugs

Confidential and Internal to Pfizer - subject to works council and/or union consultations and other legal requirements. 28

Recent Synthetic Natural Product Derived Drugs

OH

NH2HO

HO

H2N

O

OH

O

HO

HO

O

O

OO

OO

O

O

O

O

O

O

OO

HOOH

O

O

OH

MyriocinMycelia sterilia

Fingolimod

Halichondrin BHalichondria okadai

Eribulin


HO

HOO

OHStarter acid

HO

HOO

OH

HO

Shikimic acid

rapK

O

OH

OH

O

OH

O

OH

ORHO

S

OH

OH

O

OH

O

OH

ORHO

OSS

OH

O

OH

O

OH

ORHO

O

rapC

Module 11

Module 12Module 13

Module 14

SO

O

OH

OH

O

OH

O

ORHO

OH

ACPACP KSAT AT

KRDH

ER

ACPKS AT

KR

ACPKS AT

KR

KS

RAPS 3

S

O

OH

O

OH

ORHO

OS

O

OH

O

OH

ORHO

O

O

OH

S

O

OH

ORHO

OS

O

OH

O

OH

ORHO

OS

O

OH

O

OH

ORHO

OS

O

OH

O

OH

ORHO

rapB

Module 5

Module 6

Module 7

Module 8

Module 9

Module 10

ACP ACP ACP ACP ACP ACPKS KS KS KS KS KS ATATATATAT

KR KR KR KR KR KRDHDHDHDHDH

ER

RAPS 2

rapA

RAPS 1

ORHO

O

SH

Loadingdomain

CoL ER ACP

S

ORHO

O

Module 1

ER

ACPKS AT

DH KR

SO

OH

ORHO

Module 2

KS AT ACP

KR

O

OH

ORHO

OS

Module 3

ACP

KR

ER

DH

ATKS

Module 4

O

OH

ORHO

SO

KS AT ACP

KRDH

X

X

HOO

AlternativeStarter unit

R2R1

methylation and oxidation

Pipecolate Incorporating Enzyme

OHNH

O

O OCH3

O

OHOON

OHO

O

OH

OH3CO

OCH3

1) Gregory, M.A. and Leadlay, P.F. et al., Angew. Chem. Int. Ed. 2005, 44, 4757-4760. 2) Gregory, M. A. and Leadlay, P.F. et al., Org. & Biomol. Chem. 2006, 4, 3565-3568.

PKS Engineering of Rapamycin

rapamycin


Rationale for NP Biological Bias is Based on Protein Fold Space

Properties

30

Protein sequence space is essentially infinite- at 300 aa, possible sequences = 20300 >>> than particles in known universe (1080)

Total complement of estimated world proteome 1010

Most proteins resemble other proteins - built by amplification, recombination, divergence from a basic set of folding units- domains

Around 100 domain families have been recognized by sequence

Only ca. 1000 folds are populated in nature

Subdomain level - recurrent local arrangements of secondary structures

Biophysical constraints limit the number of folded conformations


Characteristics of Protein folds

• Distinct sequences often adopt very similar folds

• Highly similar sequences can adopt very different folds

• Identical peptide sequences can have different conformations in different proteins

• A single protein chain may encode for more than one structural domain.

• Similar domains are formed via different “methods”

Structure is conserved far more than sequence.

31


Distinct Sequences Often Adopt Very Similar Folds

32

Superposition of 3 proteins of similar structure but distinct sequences.

1-Isomerase from Rhodopseudomonas palustris

2- B chain of limonene-1,2-epoxide hydrolase from Rhodococcus erythropolis

3- Putative polyketide cyclase from Acidithiobacillus ferrooxidans

a) 1 and 2b) 2 and 3c) 1 and 3

<20% sequence identity in aligned regionsRegions of overlap in protein 1Regions of overlap in protein 2

A- Proteins with virtually identical structure and little or no sequence similarity

Current Opinion in Structural Biology 2009, 19:312–320, J Biol Chem 2009, 284:992-999

B- Proteins with high sequence similarity and no structure similarity

Arl2 (BART) from Homo sapiens and ADP-ribosylation factor-like protein 2-binding protein from Danio rerio – 72%


Domains in Related Enzymes can be Formed in Distinctly Different Ways

33

(a)Dimerization domain of GDP-mannose dehydogenase from P. aeruginosa

(b) Central dimerization domain of UDP-glucose dehydrogenase from S. pyogenes

(c) Single chain domain of ovine 6-phosphogluconate dehydrogenase The blue and yellow fragments highlight the correspondence with the chains shown in (b).

Current Opinion in Structural Biology 2009, 19:312–320


Natural Products Bind Proteins

As substrates for via PKS, NRPS, tailoring enzymes, etc.

Outcome of selective pressure to binding protein and cellular targets

Domains of these fold targets are conserved in the “protein foldome”

Natural product ligands leverage these properties in their mechanism and properties

Natural products, by virtue their origin, are within or at least proximal to, biologically relevant chemical

space.

34


Polyketide Immunophilin Ligand Family

HO

NO

O

HOO

O

OH

OH

OH

OH

HO

O

meridamycinnormeridamycin

rapamycin

HO

O

NOOO

O

HOO

O

O

OH

R

O

OCH3

NO

OHO O

CH3OHO

O

O

O

OH

O

OCH3

fujimycin (FK-506): R = allylascomycin (FK-520): R = ethyl

HO

HO

NO

OOHO

O

O

OO

antascomicin BOH

Immunosuppressive

Non-immunosuppressive

n= 1,0

Salituro, G. et. al., Tet. Lett., 1995, 36(7), 997-1000Summers, M.Y.; Leighton, M.; Liu, D.; Pong, K.; Graziani, E.I., J. Antibiot., 2006, 59(3), 184-189.


OCH3

NO

OHO O

HOO

O

OH

O

OCH3

H3COO

Natural Products lead to Unanticipated Drug Targets and Mechanisms

FKBP bindingdomain

mTOR effectordomain

Sehgal, S.N.; Baker, H.; Vézina, C., J. Antibiot., 1975, 28(10), 721-726.Choi, J.; Chen, J.; Schrieber, S.L.; Clardy, J., Science, 1996, 273, 239-241.

1. Rapamycin binds tightly to FKPB12 via FKBP binding domain2. Rapa-FKBP12 complex binds mTOR, disrupting TORC1 complex

mTOR

FKBP-12

RAPAMYCIN

36

Natural products, by virtue their origin, are within or at least proximal to, biologically relevant chemical space!

Technology

Session 1 part 3