Transcript Slide 1

Receptor pharmacology or animal
models for dose selection in humans?
Bart Laurijssens
Clinical Pharmacology Modelling & Simulation,
GlaxoSmithKline, UK
Satellite Meeting on Predictive Modelling in Drug Development
PAGE, St Petersburg, 23 June 2009
2
Animal Models?
Pharmacology
3
Predictive Modelling in Early Development.
 A Simulation exercise: Extrapolation!
 May include some analysis of data.
 Prediction of Dose
 Pharmacological
 Clinical
 HUMAN DOSE!
4
Why predicting Human Clinical Dose early?
TI
[adapted from: Jennifer Sims, ABPI/BIA Early Clinical Trials Taskforce, slideset]
5
Predicting Human Dose? Simple.
Bob
Barker
The Dose is Right
Pharma’s Favourite Game Show
Why is the dose “mg” not grams or “ng”?
The screening process naturally selects candidates that drive the dose range
A model can help
Parameter
Guess
Low
High
MWT
450
250
700
Kd(nM)
1
0.3
100
Clinical
10
3
100
CL(ml/min)
200
3
1000
F (%)
64%
5%
95%
Tau (hrs)
12
6
24
Dose  MWT Kd Clinical CL/F 
No prior
knowledge!
45%
35%
Probability



25%
15%
5%
-5%
0.1
0.3
1
3
10
30
100 300 1000 3000
Dose (mg)
[thanks to Daren Austin]
6
Mechanistic Classification of Biomarkers
?
Clinical Relevance of Prediction?
Ease to Predict
7
Pharmacodynamic Theory
Intrinsic
Activity
Intrinsic
Efficacy
SYSTEM
DRUG
Slope
Affinity
Disease
Age
chronic treatment
combined treatment
Tissue
species
gender
Potency
[Van der Graaf & Danhof, 1998]
8
Species differences in Receptors
9
So what about Animal Models of Disease?
 Face Validity
 Phenomenological Similarities with the disorder
 Predictive Validity




Need drugs that work
Quantitative
False positives/negatives
Mechanism specific?
 Construct Validity
 Sound theoretical rationale
 Need to understand disease and animal
10
What information to look for?
 Distribution to target(s) in Humans:
 Transporters (eg PgP)
 Extracellular vs Intracellular target
 Interaction with the Human Target(s)
 Affinity (in vitro, ex vivo)
 Efficacy (agonism vs antagonism)
 Human pharmacology
 In vitro, ex vivo
 Animal models of physiology (or Disease)
 Time course of response
 Knowledge
 Experience with mechanism in Humans
 Human physiology
 General Pharmacological Theory
11

Using Receptor Occupancy for a new target
Human PK-RO was predicted
using:
 Rat ex-vivo RO for R1
 Rat and Human in-vitro Binding
(R1 and R2)
 Rat and Human Fu, B:P
 Assumption re. PgP
12
Using Receptor Occupancy for a new target
100
100
90
90
80
Target Receptor Occupancy
70
80
NT related response (%)
Target Receptor occupancy (%, 95% CI)
2] Do notMinimal
study doses
with
“response”
during dosing interval at steady state
<80% RO
3] Doses 70
that hardly separate
based on60RO, potentially
separate 50
in efficacy
60
50
40
40
NT related response
30
30
20
1] =theoretical range of efficacy:
No suppression – No effect
Max suppression – Max effect
10
0
1
10
20
10
0
100
Compound Dose (mg)
[Page satellite meeting, Pamplona, 2005]
13
Using primary Human Pharmacology and Clinical
Knowledge
Fenoprof
en
Ketorolac
Naproxe
n
Rofecoxib
Total
Unbound
[Huntjens et al. Rheumatology 2005;44:846–859]
14
Primary Pharmacology different Human vs Animal
Gone horribly wrong
X
15
Receptor Occupancy of TGN1412 at starting Dose
[Jennifer Sims, ABPI/BIA Early Clinical Trials Taskforce, slideset]
16
Mechanism of Action of TGN1412
17
Predictive animal model
[Rocchetti et al. Eur J Cancer 43 (2007): 1862-8]
18
19
Conclusions
 It is not about animal models vs receptor occupancy, but about what
data is informative.
 Only informative data is worthy of your modelling skills and time.
 Animal Models MAY be informative
 Human Target Receptor Occupancy, or if possible, Target
(in)Activation, is always informative.
 And … nearly always available.
 HUMAN dose!
20
My Favourite Animal Model