Hugo Kubinyi, www.kubinyi.

de

Changing Paradigms
in Drug Discovery

Hugo Kubinyi

Germany

E-Mail kubinyi@t-online.de
HomePage www.kubinyi.de

Hugo Kubinyi, www.kubinyi.de

Drug Discovery - The Ancient Times

Folk medicine (mainly plants)

pro: thousands years of human

experience

con: lack of reproducibility

(varying doses)

Experiments in humans
(J. Lind, 1747, treatment of scurvy)

pro: the “right” object

con: toxicity
Hugo Kubinyi, www.kubinyi.de

Drug Discovery - The Early Times

Natural products and their analogs

pro: high percentage of actives

O
large chemical diversity O
R
con: availability may pose problems
most often difficult chemistry OH
Sugar-O
H

Animal experiments

pro: ADMET included

con: slow, expensive

ethical issues

Hugo Kubinyi, www.kubinyi.de

H
S
RHN „Penicillin
N happened,
O COOH it came out
of the blue.“

A. Fleming,
1930
Hugo Kubinyi, www.kubinyi.de

All NCE‘s, 01/1981 - 06/2006, By Source

(without biologicals and vaccines)

(n = 974)

N = natural products ND = derived from natural products

S = synthetic products NM = natural product mimics
S* = synthetics but pharmacophore derived from natural product
D. J. Newman and G. M. Cragg, J. Nat. Prod. 70, 461-477 (2007)

Hugo Kubinyi, www.kubinyi.de

Drug Discovery - The Golden Age

OH H
Endogenous transmitters and hormones HO N
R
HO OH

pro: active lead structures O CH3 H

+
involved in many different diseases H3C O
N CH3
CH3
H H
HO
I I
con: limited number of lead structures
HO O CH2CH(NH2)COOH

Isolated organs as test models

pro: includes membrane permeability

con: slow, expensive

no ADMET
ethical issues
Hugo Kubinyi, www.kubinyi.de

Parkinson‘s Disease L-dopa

dopa
caused by degeneration decarboxylase
dopamine presynaptic
of dopamine-producing Ca2+
storage membrane
cells in Substantia nigra
vesicals
ion channel

presynaptic
nerve cell
Ca2+

MAO dopamine
inactive transporter
metabolites
synaptic gap
postsynaptic
nerve cell postsynaptic
membrane
dopamine receptor
with G-protein complex

Hugo Kubinyi, www.kubinyi.de

A Rational Therapy of Parkinson‘s Disease

healthy sick Therapy

ACh + + ACh ! or
dopamine + - dopamine "

Problems oral L-DOPA,

dopamine is not bio- peripheral DOPA
available, peripheral decarboxylase
side effects, MAO blocker, central
MAO blocker
Hugo Kubinyi, www.kubinyi.de

Drug Discovery - “Rational” Approaches

Structure-based and computer-aided design

pro: straightforward approach

con: only targets with known 3D structure

only ligand design - no ADMET
risk of failure

In vitro test models

pro: very fast (up to 100,000’s / day)

target-oriented

con: no ADMET
single target

Hugo Kubinyi, www.kubinyi.de

Drug Discovery - Nowadays

Combinatorial chemistry, chemical libraries

pro: generate a multitude of compounds

con: limited chemical diversity

chemistry-driven libraries most often
outside of biological space

Chemical biology

pro: fast screening in biological systems

membran permeability included

con: no ADMET in cellular models

target/s remain/s unknown
Hugo Kubinyi, www.kubinyi.de

Drug Discovery - Nowadays

Virtual screening and fragment-based design

pro: straightforward approach

saves time and resources

con: only ligand design

still some risk of failure

Chemogenomics

pro: multitarget-oriented
quick information on
selectivity

con: no ADMET

Hugo Kubinyi, www.kubinyi.de

Estimation on Druggable Targets

G. Müller, Drug Discov. today 8, 681-691 (2003)

 Hugo Kubinyi, www.kubinyi.de

Is there really a „druggable genome“ ?

Alternative splicing and posttranslational modification
generate a multitude of proteins
→ the „druggable proteome“ ?
Protein complexes (nAChR, GABA-R, integrins, hetero-
dimeric GPCRs, cross-talking)
→ the „druggable targetome“ ?

Balanced activity against a series of targets

→ the „druggable physiome“

H. Kubinyi, Drug Research: Myths, Hype and Reality,

Nature Rev. Drug Discov. 2 (8), 665-668 (2003)

Hugo Kubinyi, www.kubinyi.de

Is Target Focus the a) b)

Best Strategy? Ki 5-HT2A = 4 nM 2.5 nM

H Ki 5-HT2B = 12 nM
N S Me Ki 5-HT2C = 11 nM 2.5 nM
Ki 5-HT3 = 57 nM
N Ki dop D1 = 31 nM 119 nM
N Ki dop D2 = 11 nM
Ki dop D4 = 27 nM
N Ki musc M1 = 1.9 nM 2.5 nM
Me Ki musc M2 = 18 nM
Olanzapine, a clozapine-like Ki musc M3 = 25 nM 13 nM
„atypical“ neuroleptic with Ki musc M4 = 13 nM 10 nM
a promiscuous binding pattern
Ki musc M5 = 6 nM
a) F. P. Bymaster et al., Neuropsycho-
pharmacology 14, 87-96 (1996)
Ki adr α1 = 19 nM
b) F. P. Bymaster et al., Schizophrenia Ki adr α2 = 230 nM
Research 37, 107-122 (1999) Ki hist H1 = 7 nM
Hugo Kubinyi, www.kubinyi.de

“Discouraging Data on the New Antidepressant”

Hugo Kubinyi, www.kubinyi.de

 Hugo Kubinyi, www.kubinyi.de

The Generation of Scaffold Diversity

O OH NH2
N H2NHN
NH

N NC
NH O OMe

NH2 CHO
HN
O
N N
HN NC
(Me)2N
O N
MeO O
OMe O COSH

L. Weber, Drug Discov. today 7, 143-147 (2002)

Hugo Kubinyi, www.kubinyi.de

Drug Research is ....

the Search for a Needle in a Haystack

 Hugo Kubinyi, www.kubinyi.de

The Productivity Gap in Pharmaceutical Industry

T.T. Ashburn and K.B. Thor, Nature Rev. Drug Discov. 3, 673-683 (2004)

Hugo Kubinyi, www.kubinyi.de

FDA-Approved NCEs Over the Last Years

B. Hughes, Nature Rev. Drug Discov. 7, 107–108 (February 2008)

Hugo Kubinyi, www.kubinyi.de

The Medicinal Chemistry Space

C. Lipinski and A. Hopkins, Nature 432, 855-861 (2004)

Hugo Kubinyi, www.kubinyi.de

Success in Drug Research

# A compound hundred thousands

is no hit thousands

# is no lead dozens
optimization: thousands
is no candidate some

is no drug 1
Hugo Kubinyi, www.kubinyi.de

by courtesy of Dr. Simona Ceccarelli, Hoffmann-La Roche

Hugo Kubinyi, www.kubinyi.de

A. Cressy Morrison

Man in a Chemical World

Medicinal The Service of Chemical Industry
Chemist
Ch. Scribner‘s Sons, NY, 1937

„Chemical Industry, Upheld

by Pure Science, Sustains
Modeller the Production of Man‘s
Necessities“
Hugo Kubinyi, www.kubinyi.de

Strategies in Drug Design

no protein protein
3D structure, 3D structure,
no ligands no ligands

combichem, HTS, de novo design

virtual screening (protein flexibility !)

no protein 3D protein 3D structure,

structure, ligands ligands

pharmacophores, structure-based
(3D) similarity, design
(3D) QSAR
LBLD SBLD

Hugo Kubinyi, www.kubinyi.de

Thrombin and Thrombin Inhibitor Complexes

Thrombin-Thrombstop Complex
 Hugo Kubinyi, www.kubinyi.de

Structure-Based Design of Captopril

Zn2+ Carboxy-
Zn2+
peptidase

O O
H - -
N O O
{ N + HO +
H NH2 NH2
O O
R
H2N N H2 N N
substrate H
inhibitor H

CH3
HOOC N HS N
O COOH O COOH
IC50 = 330 µM Captopril IC50 = 23 nM

Hugo Kubinyi, www.kubinyi.de

Success Stories and Failures of

Structure-based Design

HO OH
phospholipase
inhibitor HO OH
CH3 CH2 N N
O
MeO HIV protease
inhibitor
O OH
O O N COOH
O COOH
S HO
N OH
H AcNH
O N HN NH2
H
N
neuraminidase NH
thrombin inhibitor HN NH2 inhibitor
Hugo Kubinyi, www.kubinyi.de

Influenza
In 1918/19, the „Spanish Flu“
killed about 20-40 mio people.
Especially young and very old
people died from influenza.
The heavy death toll of this
pandemic disease has to be
compared to the number of
11 mio victims of World War I.
Egon Schiele prepared this
drawing of his wife, one day
before her death and four days
before he died himself, only
28 years old.

Hugo Kubinyi, www.kubinyi.de

Influenza Virus
schematic view
and electron micro-
scopic picture
 Hugo Kubinyi, www.kubinyi.de

hemagglutinine + sialic acid (green) neuraminidase + DANA (red)

Hugo Kubinyi, www.kubinyi.de

Design of Neuraminidase Inhibitors

O Neu5Ac2en
OH OH
HO O OH
Ki = 1 000 nM
AcNH 4
HO
Glu-276
Arg-371
O
HO OH Arg-292
HO O O
-
Arg-118
AcNH 4
HO

+
Glu-119 Glu-227
 Hugo Kubinyi, www.kubinyi.de

Design of Neuraminidase Inhibitors

O Neu5Ac2en
OH OH
HO O OH
Ki = 1 000 nM
AcNH 4
HO
Glu-276
Arg-371 (Et)2CHO COOR
O
HO OH Arg-292
HO O O
-
Arg-118 AcNH
AcNH 4 NH2
HN
+ GS 4071, R = H
H2N NH2
IC50 = 1 nM
Glu-119 Glu-227
GS 4104, R = Et
4-Guanidino-Neu5Ac2en, Ki = 0.1-0.2 nM Oseltamivir
Zanamivir (Relenza, Glaxo-Wellcome) (Tamiflu, Roche)

Hugo Kubinyi, www.kubinyi.de

Design of Selective ERα and ERβ Ligands

blue: hERαα LBD
(crystallography)

green: hERß LBD

(homology model)
estradiol

α $ hERß
hERα
„upper“ side:
Leu384 $ Met336

„lower“ side:
Met421 $ Ile373
A. Hillisch et al., Ernst Schering Res. Found. Workshop 46, 47-62
(2004); A. Hillisch et al., Mol. Endocrinol. 18, 1599-1609 (2004)
Hugo Kubinyi, www.kubinyi.de

Design of Selective ERα and ERβ Ligands

Potency ERα α : 40 % of E2 Potency ERβ β : 50 % of E2

Selectivity: 300 fold Selectivity: 190 fold

Hugo Kubinyi, www.kubinyi.de

Stepwise Virtual Screening (Virtual) Library

Property Filters
(MW, rule of 5, rot, drug-like, ...)

1D Pharmacophore and
3D Pharmacophore Searches

Docking and Scoring

Selection (Diversity, Similarity,

Inspection) Leads / Candidates
Hugo Kubinyi, www.kubinyi.de

Stepwise Virtual Screening

1.8 mio compounds (Aventis in-house repository)
MW, rot-bond filter, 3D pharmacophore search
22,950 compounds
docking into an α 1A receptor model
(GOLD, PMF)
300 top-scoring compounds
clustering, diversity selection
80 compounds tested, 37 hits with Ki < 10 µM

N F
H N
N
N
N
O S
α 1A adrenergic receptor antagonist, Ki = 1.4 nM

A. Evers and T. Klabunde, J. Med. Chem. 48, 1088-1097 (2005)

Hugo Kubinyi, www.kubinyi.de

Fragment-Based Approach (Astex)

a) complex
binding
site
b) fragments
in different
pockets
c) assembly
of different
fragments
d) growing
out from a
fragment

T. L. Blundell et al., Nature Rev. Drug Discov. 1, 45-54 (2002)

Hugo Kubinyi, www.kubinyi.de

Fragment-
based Design
of a Thrombin
Inhibitor
OMe
O
O S
OMe HN
N N
O
N N HO
O S
+ NH
HN N N
Cl N N
HO IC50 = 100 µM
NH2 Cl
IC50 = 12 µM IC50 = 1.4 nM

A. Ciulli and C. Abell, Curr. Opin. Biotechnol. 18, 489-496 (2007)

Hugo Kubinyi, www.kubinyi.de

Reasons for Failure in Drug Development

(n = 198)
5% 5%
10%
39%

11%

30% Pharmacokinetics
R. A. Prentis et al., Lack of efficacy
Br. J. Clin. Pharmacol. Animal toxicity
25, 387-396 (1988); Adverse effects in man
T. Kennedy, Drug Discov. Commercial reasons
today 2, 436-444 (1997) Miscellaneous
Hugo Kubinyi, www.kubinyi.de

Human Absorption and Polar Surface Area

FDp = fraction of dose
absorbed to the portal vein

red triangles: data from

K. Palm et al., Pharm. Res.
14, 568-571 (1997)

Hugo Kubinyi, www.kubinyi.de

Human Absorption and Polar Surface Area

FDp = fraction of dose
absorbed to the portal vein

triangles and circles:

G. M. Grass and P. J. Sinko,
Drug Discov. Today 6,
S54-S61 (2001)

data from
W. K. Sietsema, Int.
J. Clin. Pharmacol.
27, 179-211 (1989)
Hugo Kubinyi, www.kubinyi.de

Reasons for Failure in Drug Development

(n = 121; without
7% 7% 7% antiinfectives)
16%

17%
R. A. Prentis et al.,
Br. J. Clin. Pharmacol.
25, 387-396 (1988);
T. Kennedy, Drug Discov.
today 2, 436-444 (1997)
46%
Pharmacokinetics
Lack of efficacy
Animal toxicity
Adverse effects in man
Commercial reasons
Miscellaneous

Hugo Kubinyi, www.kubinyi.de

Rodent, Dog, Primate and Human Bioavailability

Rodents
green circles

Dog
red triangles

Primates
blue squares

data from:
W. K. Sietsema,
Int. J. Clin.
Pharmacol. 27,
179-211 (1989)

G. M. Grass and P. J. Sinko, Drug Discov. Today 6, S54-S61 (2001)

Hugo Kubinyi, www.kubinyi.de

The Role of
Transporters in
Drug Uptake
and Elimination

H. Gleaser et al.,
in R. J. Vaz and
T. Klabunde,
Antitargets,
Wiley-VCH, 2008,
pp. 341-366

Hugo Kubinyi, www.kubinyi.de

 Hugo Kubinyi, www.kubinyi.de

Gene Technology in Drug Research

% Identification of a therapeutically relevant protein:
Identification of a gene and determination of its
sequence yield the protein sequence. Elucidation
of its function and 3D-structure prediction
% Proof of therapeutic concept: Introduction, amplifi-
cation or knock-out of the gene in animals
% Development of a molecular test system: Screening
with human protein, reduction of animal experiments
% Production of the protein: finally leads to the three-
dimensional structure and to structure-based design.

Hugo Kubinyi, www.kubinyi.de

Species Specificity of Renin Inhibitors

Remikiren

O O OH
IC50 = 0.8 nM (human)
O
H 1.0-1.7 nM (monkeys)
S N
N 107 nM (dog)
H 3 600 nM (rat)
O OH

N NH

IC50 = 0.6 nM (human)

O 2 nM (marmoset)
OH
H 7 nM (dog)
H 2N N NH2 11 nM (rabbit)
O 63 nM (guinea pig)
O
80 nM (rat)
150 nM (pig)
MeO Aliskiren 8 500 nM (cat)

J. M. Wood et al., Biochem. Biophys. Res. Comm. 308, 698-705 (2003)

Hugo Kubinyi, www.kubinyi.de

„All things are poison

and nothing without
poison; only the dose
determines, whether
a thing be no poison“

Salt, Fat, Alcohol ...

Aspirin, Corticoids ...
Phenacetin, Phenphen,
Cerivastatin ...

Hugo Kubinyi, www.kubinyi.de

Acute Toxicity of Tetrachlorodibenzodioxin

Cl O Cl
2,3,7,8-Tetrachloro-
dibenzodioxin
Cl O Cl
Species LD 50 in µg/kg
Mouse 114-280
Rat 22-320
Hamster 1,150-5,000
Guinea Pig 0.5-2.5
Mink 4
Rabbit 115-275
Dog > 100 < 3,000
Monkey < 70
Man ??
Hugo Kubinyi, www.kubinyi.de

Acute Toxicity of Lysergic Acid Diethylamide

in Animals and Maximum Tolerated Dose in Man
LD50 in
Species
H CO-N(Et)2 mg/kg
Mouse 50-60

N Rat 16.5
CH3
H Rabbit 0.3

Elephant « 0.06
HN
LSD
Man » 0.003

Albert Hofmann, LSD - My Problem Child, McGraw Hill, 1980

Hugo Kubinyi, www.kubinyi.de

An Early Clinical Study - Coffee or Tea ?

In late 18th century Gustav III, King of Sweden,
performed a “clinical study” to confirm the
negative effects of coffee drinking on health.
One convicted murder had to drink only coffee,
another one tea, instead.
Two physicians supervised the study.

First, one physician died.

Then the other physician died.
Then the king was murdered.
The tea drinker died in the age of 83.
The coffee drinker survived all others.

Nevertheless, in 1794 coffee drinking was forbidden in Sweden

and later again, in 1822.

An early clinical trial, Ann. Int. Med. 117, 1, 30 (1992)

Hugo Kubinyi, www.kubinyi.de

Clinical Studies - the Typical Volunteer

Phase I

healthy volunteers,
age 18-55 years,
males and females
(however, no females
who could be or could
become pregnant),
normal weight, no
smokers, no alcohol
(ab)use, standard
food, drug taken with
150 ml water, no other
therapy, no intake of
fruit juices or illegal
drugs.

Hugo Kubinyi, www.kubinyi.de

The Patients

plus other disease(s)

Hugo Kubinyi, www.kubinyi.de

The Past

Voltaire (1694-1778):

Doctors
pour drugs of which
they know little,
to cure diseases of which
they know less,
into human beings
of whom
they know nothing.
Voltaire, by J. A. Houdon

Hugo Kubinyi, www.kubinyi.de

The Future: Pharmacogenomics -

New Opportunities from Personalized Medicine

Genotyping of drug targets and metabolic enzymes

enables
- cost savings in drug development through better
design of clinical trials
- selection of the „best drug“ for a certain patient
- individual dose ranges (variance in target sensitivity,
reduced or increased metabolism)
- fewer toxic side effects
- fewer unexpected drug-drug interactions