Transcript Title of Presentation Date - American Society for Experimental

Early Stage Clinical Drug Development:
More than Just “Feed and Bleed”:
Opportunities for Biomarker and Experimental Medicine Studies
1 March 2012
Ajay Verma, MD, PhD
VP, Early Stage Neurology
and Experimental Medicine
Biogen Idec
American Society for Experimental NeuroTherapeutics | 15th Annual Meeting
Disclosure
Previously worked for
Merck & Co, Inc. and
Novartis Pharmaceuticals
Currently employed by
Biogen Idec
American Society for Experimental NeuroTherapeutics | 15th Annual Meeting
Learning Objectives
• To share examples of CNS pharmacodynamic biomarkers
and human experimental medicine models used in early
stage drug development
• To share lessons learned for employing CNS biomarker
and experimental medicine strategies in early stage
neuroscience drug development
American Society for Experimental NeuroTherapeutics | 15th Annual Meeting
The Need for Innovation in Early Clinical Stages
of Neuroscience Drug Development
• Drug development for CNS disorders is most lengthy, costly, and least
successful of any therapeutic area (~15 yrs and ~$1.5B/drug)
• Many trials fail to efficiently address key hypotheses
Miller G., 2010, Science, 329:502-5
Three Possible Outcomes of a
Therapeutic Efficacy Trial
“Positive”
“Negative”
“Failed”
Certainty of hypothesis
testing
Absence of certainty of
hypothesis testing
It is critical to distinguish failure of the intervention from
non-relevance of the biological target to the human disease
5
Main Hypotheses Tested During
Neurotherapeutics Development
Preclinical
Proof of
Biological Concept
Proof of
Clinical Concept
Proof of
Commercial Concept
Pharmacology can
be expressed in man
Pharmacology
translates into
patient benefit
Patient benefit is
commercially attractive
Phase 1
Phase 2
• Early Stage CNS Biomarkers
• Experimental Medicine models
Phase 3
Mitigating Risk Through Early Clinical Confirmation of Target
Engagement and Brain Pharmacodynamic Activity
7
Opportunities for Improving Hypothesis
Testing in Phase 1
Preclinical
Phase 1
Phase 2
Phase 3
• Traditional primary focus of Phase I clinical studies has been
safety, tolerability and pharmacokinetics
– Single ascending dose and multiple ascending dose
– Food effect & drug-drug interactions
– Normal volunteer and patient trials
• Focus changing to leverage maximum information from early
clinical trials
• Early use of biomarkers and surrogate endpoints can inform
efficacy, dose-response and time-effects
• Well-performed biomarker studies in early stage can provide
confidence for continued investment or an early kill (de-risking)
Biomarker Framework for Improving Early Testing of
Biological Hypothesis
RISK
Ph1 needs
Target
mechanism
TE
PD
Ph2,3 needs
Cohort
Ph2b
Ph3
enrichment
readouts
readouts
Mech
Target,
Clinical
Genetic,
Clinical scale Registerable
specific mech., or
endpoints
scale
prodromal,
population
Ligand
or
Disease
specific
imaging Pathway
Disease
epidemiology progression Biomarkers to
biomarker
or process
progression biomarker
or pathology support label
specific
or pathology
biomarker
Biobiomarker
chem
Biogen Past BIIB014 Neopterin
efforts
(A2a)
(β-IFN)
Safety
Clinical
POC
Anti-JCV
Ab
(Tysabri)
Gd+ MRI
(BG00012)
CD56+ NK
(Dac-HYP)
EDSS
Payer
readouts
Patient
Related
Outcomes
MSWS-12
CNS Biomarkers Platforms for Tracking Effects of
Neurotherapeutics
CSF:
Drug levels (PK)
ELECTROPHYSIOLOGY:
Metabolites,
Proteins,
Peptides
EEG
MEG
IMAGING:
Evoked
Potentials
FDG-PET
Regional
Blood Flow
(PET, ASL)
Transcranial
Magnetic
Stimulation
(TMS)
Functional
Imaging (fMRI)
MR Spectroscopy
Ligand PET
10
COGNITION and BEHAVIOR
Purpose Driven Biomarker Efforts in Early Development:
“Biomarker”= Readout Needed for Answering Questions
TMS
Did the drug
get in to the CNS?
Cognition
EEG
PET
ligand
Behavior
Did the dose
provide adequate
target occupancy?
Did the drug
do anything to
CNS biology?
11
Plasma
PK
Clinician
fMRI
Cerebrospinal
fluid PK
Genotype
Brain Delivery and Target Engagement Biomarkers:
PET and SPECT
John Newport
Langley,
1852-1925
“Receptive substances”
Mediate drug action
• Ligand PET and SPECT techniques allow
confirmation of brain entry and site of drug action
• SPECT is useful for longer half-life agents
such as biologics
• Displacement studies with unlabeled drug allow
for correlation of PK with % receptor occupancy
PET Imaging of small molecules provides direct drug receptor
occupancy measurements across species
Autoradiography
+ Blank
SPECT Imaging of Biologics
In vivo PET scan
+ dosing
+ dosing
24h
12
72h
120h
PET ligands have had a wider use in Psychiatry than
neurology drug development
Takano, 2010
Brain Delivery and Target Engagement Biomarkers:
CSF assays for drugs and drug effects
• In vivo human CNS pulsechase labeling assessed via
continuous CSF sampling.
• Sample collection and
measurements of labeled
amino acid and Aβ peptide
is shown following
intravenous 13C-leucine
labeling
• Allows early stage study of
drugs that alter protein
turnover (gamma-secretase
inhibitors, other
proteostasis mechanisms)
• Use of deuterated water
allows labeling of wider
range of molecules
14
Bateman, 2006
Tracking Lesions, Efficacy, and Safety with
Multimodal Imaging
Anatomy
Amyloid
plaques
Synaptic
metabolism
Blood
flow
Circuit
function
Safety
Microhemorrhage
Vasogenic edema
ROI: region of interest analysis in precuneus and entorhinal cortex
(Reisa Sperling, Keith Johnson, David Alsop)
Functional brain biomarkers employed in tracking disease
progression as well as drug effect
18FDG-PET
Progressive dementia
The brain is constantly active
and so use of functional brain
assays to detect pathology or
drug effect relies on detecting
qualitative and quantitative
change from baseline. This
applies to all functional brain
biomarkers
Pscilocybin PD effect
16
Ethanol PD effect
Brain Perfusion Based Pharmacodynamics Using
Dynamic Arterial Spin Labeling
1. ASL-MRI approach to measure rCBF
2. Experimental Medicine study for drug effect (N=16)
Legacy of Functional Brain Pharmacodynamic Biomarkers
1st functional hemodynamic expt
Angelo Mosso,
1846–1910
1st functional EEG expt
Hans Berger,
1873-1941
‘The alpha rhythm’
18
EEG and Magnetoencephalography (MEG) measure drug
pharmacodynamics
MEG Power spectra
MEG & EEG
EEG electrodes
MEG sensors
19
MEG instrument
Drug:Placebo Power spectral ratios
Lorazepa
m
Acute drug effect: relative power in frequency bands
Modafinil
Methylphenidate
Increase
20
Decrease
Fit of Purpose Pharmacodynamic
Biomarkers Examples
Fit of Purpose Pharmacodynamic Biomarkers
• Near Infrared spectroscopy (NIRS) and MR spectroscopy (31P-MRS) assay
for drug effects on in-vivo mitochondrial bioenergetics in resting and
contracted skeletal muscle.
• NIRS allows measurement of tissue oxygen utilization as an index of
Complex activity (Cytochrome c Oxidase)
•
31P-MRS
allows measurement of the kinetics of phopshocreatine (PCr)
synthesis as an index of Complex V (F1-F0 ATPase) activity
NIRS
MRS
Contraction induced PCr
Decline and recovery
Resting
Contracting
Time constant
Tc = 1/k
Cortical Function Assays: Motor Evoked Potentials
Using Transcranial Magnetic Stimulation (TMS)
•Non-invasive, painless, sensitive to glutamatergic and GABAergic drugs
•Technique can be modified to study several CNS circuits and cortical plasticity
Biomarker Validation:
Variability measures of different TMS responses
Measure
ICC
90% CI
CV
90% CI
SICI
0.53
(0.14; 0.88)
0.27
( 0.21; 0.36)
ICF
0.46
(0.11; 0.86)
0.21
( 0.17; 0.27)
SICF
0.23
(0.03; 0.73)
0.26
( 0.20; 0.33)
LICI
0.60
(0.21; 0.90)
0.41
( 0.28; 0.58)
RMT
0.93
(0.84; 0.97)
0.03
( 0.02; 0.04)
AMT
0.73
(0.41; 0.92)
0.06
( 0.05; 0.08)
CSP130AMT
0.71
(0.35; 0.92)
0.28
( 0.20; 0.39)
CSP150AMT
0.68
(0.31; 0.91)
0.23
( 0.17; 0.30)
CMCT
0.24
(0.03; 0.76)
0.20
( 0.16; 0.24)
MEP (Averaged over
all stimuli)
0.46
(0.10; 0.86)
0.22
( 0.17; 0.28)
MEP (Averaged over
stimuli>1.1 ms)
0.54
(0.15; 0.88)
0.23
( 0.18; 0.30)
MEP Input-Output Curves: Dextromethorphan (DMO) vs. PBO
Baseline
2,5
3
DMO
1.5
PBO
MEP amplitude (mV)
MEP amplitude (mV)
3
2
1,5
1
,5
0
2,5
DMO
PBO
2
1,5
1
,5
0
0.5 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.5
* P < 0.05
Stimulus Intensity (x MEP1mV)
0.5 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.5
Stimulus Intensity (x MEP1mV)
8hr post-drug
2,5
DMO
PBO
2
1,5
1
,5
0
0.5 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.5
Stimulus Intensity (x MEP1mV)
24hr post-drug
3
MEP amplitude (mV)
MEP amplitude (mV)
3
3hr post-drug
** P < 0.01
*
*
**
*
2,5
DMO
PBO
2
1,5
1
,5
0
0.5 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.5
Stimulus Intensity (x MEP1mV)
Experimental Medicine
(Clinical Pharmacodynamics) Models
• Experimental pharmacodynamic models are
simulations of naturally occurring disease states
• Symptoms are brought under laboratory control to
investigate how compounds can modulate the
elicited symptoms
• Can be applied in both healthy volunteer and
patient populations
Examples of Experimental Medicine Models
Used in Early Stage Drug Development
• Well-validated pharmacodynamic models are available
for a wide range of clinical conditions
• Examples of experimental Model
-Pain
-Anxiety
-Diabetes
-Appetite control
-Sexual dysfunction
-Cognitive impairment
-Depression
-Sleep
-Muscle atrophy
Examples of Experimental Medicine Models
Used in Drug Development
• Well-validated pharmacodynamic models are available
for a wide range of clinical conditions
• Examples of experimental Model
-Pain
-Anxiety
-Diabetes
-Appetite control
-Sexual dysfunction
-Cognitive impairment
-Depression
-Sleep
-Muscle atrophy
Challenges in Developing a Pain Model
• Many different types of pain, operating by multiple
mechanisms of action
• Pre-clinical models tend to be focused on behavioural
assessments eg tail-flick, paw-withdrawal and lack a
subjective component
• Need to deliver “significant” stimulus that can be repeated
Intradermal Capsaicin – Human Model
•
Spontaneous Pain
– Recorded on 100mm VAS
– Endpoints from “no pain” to
“worst pain imaginable”
•
Hyperalgesia
– Increased pain to a
previously painful stimuli
– Mapped using a punctate
stimulus (von Frey hair)
•
Allodynia
– Pain to a previously nonpainful stimulus
– Mapped with a paintbrush
•
Flare
– Identified by visual
assessment
– Traced on to sheet of
acetate
Intradermal Capsaicin – Human Model
Effects of 3 analgesics with
contrasting mechanisms of action
on the ID capsaicin model
Treatments:
•
–
–
–
–
•
Placebo
300mg Pregabalin
10mg Morphine
30mg Keterolac
Pain VAS
70
Pain Score (mm)
•
60
50
placebo
40
300mg Pregabalin
30
30mg Keterolac
20
10mg Morphine
10
0
0
16 healthy volunteers each
received 4 ID capsaicin injections
2
placebo
300mg Pregabalin
30mg Keterolac
10mg Morphine
10
0
0
2
4
Time (hours)
6
8
Area of Allodynia (cm2)
Area of Hyperalgesia (cm2)
50
20
8
Area of Allodynia
60
30
6
Time (hours)
Area of Hyperalgesia
40
4
50
40
placebo
30
300mg Pregabalin
20
30mg Keterolac
10mg Morphine
10
0
0
2
4
6
8
Time (hours)
John Connell, ICON
Thermal Grill
Kern, 2008
Examples of Clinical
Pharmacodynamic Models
• Well-validated pharmacodynamic models are
available for a wide range of clinical conditions
• Examples of experimental Model
-Pain
-Anxiety
-Diabetes
-Appetite control
-Sexual dysfunction
-Cognitive impairment
-Depression
-Sleep
-Muscle atrophy
CONFIDENTIAL
Presentation to the Science & Technology Committee – September 2011
Scopolamine Model of Cognitive
Impairment
• Widely used experimental model of cognitive impairment
• Non-specific muscarinic antagonist producing temporary & reversible
impairments of attention and memory in healthy subjects similar to
those seen in elderly subjects and in mild Alzheimer’s disease
Scopolamine
Cholinergic
neuron
Muscarinic drugs
Ach transmission is widespread throughout the
brain, is involved in cognition, sleep, and
dreaming, and interacts reciprocally with many
neurotransmitter systems.
• H3 inverse agonist
• Experimental drugs
acting by diverse
mechanisms
Donepezil
Scopolamine Model of Cognitive
Impairment
• Study Design
28 healthy male volunteers
5-way crossover
Placebo
0.5mg Scopolamine (iv)
0.5mg Scopolamine + 10mg Donepizil (oral)
0.5mg Scopolamine + MK3134 (H3 inverse agonist) (oral)
0.5mg Scopolamine + 10mg Donepizil (oral) + MK3134 (oral)
• Main outcome measures
CogState Early Phase computerized test battery
Groton maze learning test
One card learning
Simple reaction time
Choice reaction time
Baseline lowering of cognition with scopolamine increases
dynamic range for observing drug effects.
Mean change from baseline (+SE) in psychomotor
function for each study treatment condition
A (Placebo)
B (Scopolamine)
C (MK + Scop)
D (Dnpzil + Scop)
CogState tests show differential magnitude of
scopolamine impairment and drug reversal
E (MK + Dnpzil + Scop)
0.06
Cognitive
Domain:
detection speed
0.04
Processing Speed
0.04
0.00
0.00
Executive
function
Episodic
memory
Learning
DETECT
IDENTIFY
GMLT
GMLT-DELAY
LEARN
-0.5
-0.02
Effect size (d)
change from baseline
Attention
0
0.02
Change form baseline
Processing
speed
-0.04
-0.04
-0.06
-0.08
-0.08
-0.10
-1
-1.5
-2
-0.12
-0.12
-2.5
-0.14
-0.16
-0.16
-3
-0.18
Test:
0
1
2
3
4
5
6
7
8
9
10
11
12
13
Hours
from
baseline
hours
from baseline
Pbo, no scop
Scop vs. Pbo
Donep + Scop vs. Pbo
MK3134 + Scop vs. Pbo
MK + Donep + Scop vs. Pbo
• Transient impairment of cognition facilitates observation of pro-cognitive drug effect in normal subjects
• Both Donepezil and MK-3134 able to blunt scopolamine impairment with evidence of additive effect
37
Cho, 2010
Personal Neurometrics: Out of the lab and into people’s lives
Am I managing
my genomic risk
factors and staying
Healthy?
Can I link up my
sensor and
fitness apps
data with my
EHR?
Should I go
see a
doctor?
How’s my
driving?
Computer brain training
Body physiology
Genomics
Brain computer
interface
Real-time
Neurophysiology
Key Observations
• Biomarkers and Experimental Medicine models were
only infrequently used to make key decisions about
pipeline assets
• This stemmed in large part from a lack of familiarity
with the assays and concerns about seeing
unexplainable signals that might draw negative
attention to the pipeline asset
• Outside of PET, many project teams remain hesitant
about the use of proof of biology biomarkers with
their pipeline drugs.
• In most cases, Biomarker work and Experimental
medicine work began too late in the drug
development process to inform key decisions.
Biomarker Framework and Plans for
Clinical Development Programs
Ph1 needs
Program
X
Target
Indication TE
mechanism
bmrkr
PD
bmrkr
Safety
bmrkr
Clinical
POC
bmrkr
Neuro-repair Traumatic PET
injury
“
“
“
“
“
“
AntiNeuroinflam
“
“
“
“
“
“
AntiNeurodegen
AD
PD
HD
ALS
Y
Z
Ph2,3 needs
Cohort
Ph2b
Ph3
Enrichment Readouts
Readouts
bmrkr
(bmrkr)
(bmrkr)
Mech
Target,
Clinical
Genetic,
Clinical
Registerable
specific mech., or
scale
endpoints
scale
prodromal,
population
PET
or
specific
Process
Disease epidemiology Disease Biomarkers to
biomarker progressio biomarker progression support label
specific
Bion or
or
chem
pathology
pathology
biomarker
biomarker
MS
• Bmrkr = biomarker or readout; TE = Target engagement; PD= Pharmacodynamic; POC = proof of Concept
• Many TE and PD biomarkers to be used in man will require preclinical-clinical validation
• Some biomarker efforts can serve as platforms in that they can be applied to multiple programs
Biomarker Development Process
Biomarker review, governance, stewardship
Biomarker
Required:
…In phase 1
…In Phase 2
…In Phase 3
Drug
Candidate
Biomarker
Plan
Review
Progress
Discovery
Review
Progress
Development
Discovery
Review
Progress
Ready
Development
Ready
Validation
Discovery
Lead
Optimization
Preclinical
programs
Phase 1
Ready
Phase 2
Phase 3
Impact on Clinical Care and Practice
• Successful drug development hinges on the early
identification of compounds with a high probability of
clinical success
• Quantitative CNS pharmacodynamic assessment
techniques are needed, which have a high degree of
validation and quality control
• Investment in Biomarker and experimental medicine
techniques can help support go/no go decisions, but
needs to happen very early in the drug development
process