Use of biomarkers and models in early phase development

1
Use of biomarkers and models in early phase
development

• A critical appraisal
• Paul Rolan MD FRACP FFPM
• Medical Director
• ICON - Medeval Clinical Pharmacology
• Manchester UK

2
Biomarker types
Initial pharmacological effect
Drug
Biomarker
Biomarker
3
Biomarker types

• Type 1
• confirmation of primary pharmacology
• supports the predictions of preclinical models
• allows PK /PD relationship assessment
• but is not really closer to predicting efficacy
• drug specific
• easy to validate

• Type 2
• biological / disease response
• an effect downstream from primary pharmacology
  and which is likely to result in clinical
  benefit
• potential diagnostic or surrogate
• system not drug specific
• because of system complexity may need to be an
  experimental model
• hard to validate

May have only one or both types for a new target
4
Biomarker types - examples
5
Validity
Precision Bias Reproducibility Observer
bias Equipment variation SOPs
6
Validation

• the two aspects of validation need to be
  considered separately
• validation is a reasonable term for the
  numerical precision aspect
• predicitive utility may be a better term for
  the interpretation aspect
• predictive utility is
• time specific
• context specific
• observer dependent

7
Prospective evaluation of a biomarker - example

• tucaresol (589C80)
• designed to increase haemoglobin oxygen affinity
• intended for treatment of sickle cell anaemia
• narrow therapeutic index
• widely differing kinetics betweenanimal species
• how to dose escalate and estimatelikely
  therapeutic dosage regimen

8
New biomarker - MOD

• defined as proportion of haemoglobin molecules
  modified to high affinity form

9
Use of MOD

• between species, close correlation between MOD
  and pharmacological / toxicological effects but
  poor relationship with dose
• strong theoretical and clinical case for
  therapeutic MOD to be 15-30
• first-in-man study (including stopping rules)
  designed around achievement of defined MOD
• patient study confirmed inhibition of haemolysis
  and reduced sickle cells with achievement of
  defined MOD
• use of this novel biomarker facilitated safe,
  quick exploratory development

10
Models as biomarkers

• can simulate disease
• can be tests of the whole system
• may only be able to see drug effect on stimulated
  / challenged state
• can turn disease symptoms on and off under
  laboratory control
• may be hard to assess directly in patients due to
  comorbidities and treatments
• ethical issues in provoking disease in patients,
  especially psychiatric symptoms

11
Cold Pain Model

• Methodology
• Stirred, thermostatically-controlled water bath
  at 2C
• Subjects hand is immersed in the water for 2
  minutes
• Hand held open and submerged to the wrist
• Subject adjusts a visual analogue scale on a
  computer screen using other hand
• Scale from no pain to maximum pain

12
Cold Pain Model
Cold Pain Test
100
90
80
70
60
50
Pain score
40
30
20
10
0
0
15
30
45
60
75
90
105
120
Time (seconds)
13
Cold Pain Model - measurement aspects
Is the measurement stable and reproducible?
Yes!
N12, crossover
14
Cold Pain Model
Is the model valid to detect opioid action?
Yes!
15
Cold Pain Model
Is the model valid to rank opioid action?
Maybe not!
16
Appetite Control Model

• Method
• within-subject, placebo-controlled study
• food intake determined by subtracting weight of
  food remaining from that presented to volunteer.
• subjective ratings taken for hunger, palatability
  and satiety.
• other outcome measures energy intake,
  macronutrient intake

17
Method Development Study

• Four way cross-over study in 12 normal weight
  (BMI, 20-24.9) and 12 overweight (BMI, 25-29.9)
  male volunteers
• Treatments Placebo (twice), 30mg sibutramine
  and pizotifen 3mg

18
Appetite Control Model
19
Symptoms of Anxiety

• Autonomic
• palpitation
• sweating
• stomach churning
• dry mouth
• tremor

• Hyperventilation
• light-headed
• paraesthesia

• Psychic (subjective)
• apprehension, fear, dread
• desire to escape
• re-experiencing

20
Simulated Public Speaking Visual Analogue Mood
Scale
60.00
55.00
50.00
45.00
Placebo
VAMS Anxiety Score
40.00
35.00
30.00
25.00
20.00
Baseline
Pre-test
Anticipation
Speech
Final
Period
21
SPS Effect of 2mg lorazepam
60.00
55.00
50.00
45.00
VAMS Anxiety Score
40.00
placebo
35.00
30.00
25.00
20.00
Baseline
Pre-test
Anticipation
Speech
Final
Period
22
SPS Effect of 5mg ipsapirone
mm
65
Placebo
60
Ipsapirone
55
50


45
40
35
30
Recovery
Post-
Begin
Mid
Baseline
drug
speech
speech
23
SPS - Predictive Validity

• True positives
• lorazepam, diazepam, ipsapirone
• True negatives
• CI-988 (CCK-B antagonist)
• ritanserin
• acute clomipramine
• Other compounds
• 5-HT ag/antag e.g.SR 46349B, fenfluramine
• novel mechanisms

24
AC Model - Overview
Tone
Tone
Tone
Tone
Tone
Tone 11 noise
Tone
Tone
Tone
Tone
Tone
3
9
10
12
13
14
20
21
1
2
µmho
SF
SF
SF
Conditioning
Conditioning response to tones
Responses to tones habituate
SF spontaneous fluctuations
25
AC - Effect of ritanserin
1.0
Ritanserin
log SCR
Control
-0.4
-1.8
-3.2
-4.6
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Trials
26
AC - Predictive Validity

• True positives
• diazepam, buspirone, ritanserin
• mCPP, fenfluramine
• True negatives
• CI-988 (CCK-B antagonist)
• 5-HT3 antagonists
• Other compounds
• 5-HT ag/antag e.g.SR 46349B
• novel mechanisms

27
Biomarkers for adverse effects Suspected Sedative
First-in-man Study
600
580
560
540
520
500
placebo
480
4 units
Peak Saccadic Velocity (degrees per second)
6 units
460
8 units
440
420
400
380
360
340
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
Time (hours)
28
PK/PD response
Mean
Mean Drug
300
500
PSV
Concentration
275
480
250
225
460
200
440
175
150
420
125
400
100
75
380
50
360
25
0
340
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
Time
29
Effect of disease
30
Summary

• a new classification and terminology of
  biomarkers will help consensus on their roles and
  validation needs
• consider predictive utility rather than
  validation
• Phase 0 studies to validate Type 1 biomarkers
  important
• systems biology / in silico modelling important
  for validation of Type 2 biomarkers
• models may be useful biomarkers but validation is
  complex
• increasing debate about ethics / science /
  business of going into man without appropriate
  biomarkers
• todays biomarkers are the diagnostics of the
  future