Vision of a Pharmaceutical Scientist:
Drug Development and Regulation
Leslie Z. Benet, PhD
Department of Bioengineering and Therapeutic Sciences
Schools of Pharmacy and Medicine
University of California San Francisco
New Frontiers in Manufacturing
Technology, Regulatory Sciences and
Pharmaceutical Quality Systems
Brasilia June 26, 2012
Kola and Landis, 2004
Was there a major scientific advance
between 1991 and 2000 that led to
this significant decrease in failure
rates due to PK/Bioavailability?
In fact, No.
The advance was academicians
being able to convince chemists
that drug exposure, i.e., systemic
concentrations, were important.
Historical New Chemical Entity (NCE)
Success Rate (Slide I first made about 18 years ago for FDA lectures at UCSF)
Traditional Med Chem
Rational Drug Design
Combinatorial Chemistry & High Throughput Screening
NCEs prepared
500
5,000
5,000,000
Evaluate PK/Metab/Tox
50
500
5,000
Clinical
3
3
3
Marketed Drugs
1
1
1 (an overestimate)
Although combinatorial chemistry and high
throughput screening of drug targets enabled
many new chemical entities to be identified as
potential therapeutic agents, these techniques
did not address the rate limiting step in drug
development, preclinical in vivo studies of
metabolism, pharmacokinetics and toxicity.
These are the areas of academic science that
I and my colleagues have tried to address
over the past 40 years.
What about the new “combinatorial
chemistries” that drug discovery and
development scientists are enamoured
with today and with which the regulatory
agencies must grapple?
1. In silico drug discovery and development
2. Pharmacogenomics
3. Biomarkers
4. Systems biology
5. Transporters
6. Nanomedicine and targeted drug delivery
7. Preclinical alternatives to animal studies
High Solubility Low Solubility H
igh
P
erm
eab
ilit
y
Lo
w
Pe
rmeab
ilit
y
Amidon et al., Pharm Res 12: 413-420, 1995
Class 2 Low Solubility
High Permeability
Class 1 High Solubility
High Permeability
Rapid Dissolution
Class 3 High Solubility
Low Permeability
Class 4 Low Solubility
Low Permeability
This morning, Dr. Shah described BCS
Biopharmaceutics Classification System
High Solubility Low Solubility H
igh
P
erm
eab
ilit
y
Rate
Lo
w
Pe
rmeab
ilit
y
Rate
Class 1
Metabolism
Class 3 Renal & Biliary
Elimination of
Unchanged Drug
Class 4 Renal & Biliary
Elimination of
Unchanged Drug
Major Routes of Drug Elimination
Class 2
Metabolism
Wu and Benet, Pharm. Res. 22: 11-23 (2005)
Biopharmaceutics Drug Disposition Classification System
BDDCS
High Solubility Low SolubilityE
xte
ns
ive
Me
tab
oli
sm
Po
or
Me
tab
oli
sm
Class 2Low SolubilityExtensive Metabolism
Class 1High SolubilityExtensive Metabolism
(Rapid Dissolution and ≥70% Metabolism for Biowaiver)
Class 3High SolubilityPoor Metabolism
Class 4Low SolubilityPoor Metabolism
Wu and Benet, Pharm. Res. 22: 11-23 (2005)
Major Differences Between
BDDCS and BCS
Purpose: BCS – Biowaivers of in vivo
bioequivalence studies.
BDDCS – Prediction of drug disposition
and potential DDIs in the intestine & liver.
BDDCS– Predictions based on intestinal
permeability rate
BCS – Biowaivers based on extent of
absorption, which in a number of cases
does not correlate with jejunal permeability
rates
BDDCS Applied to
Over 900 Drugs
Leslie Z. Benet, Fabio Broccatelli,
and Tudor I. Oprea
AAPS Journal
13: 519-547 (2011)
Here we compile the BDDCS classification for 927 drugs, which
include 30 active metabolites. Of the 897 parent drugs, 78.8%
(707) are administered orally. Where the lowest measured
solubility is found this value is reported for 72.7% (513) of
these orally administered drugs and a Dose Number is
recorded. Measured values are reported for percent excreted
unchanged in urine, Log P and Log D7.4 when available. For
all 927 compounds, the in silico parameters for predicted Log
solubility in water, calculated Log P, Polar Surface Area and
the number of hydrogen bond acceptors and hydrogen bond
donors for the active moiety are also provided, thereby
allowing comparison analyses for both in silico and
experimentally measured values. We discuss the potential use
of BDDCS to estimate disposition characteristics of novel
chemicals (new molecular entities, NMEs) in the early stages of
drug discovery and development.
High Solubility Low Solubility H
igh
P
erm
eab
ilit
y
Lo
w
Pe
rmeab
ilit
y Class 1 Marketed Drugs
40%
NMEs: 18%
Class 3 Marketed Drugs
21%
NMEs: 22%
Class 4 Marketed Drugs
6%
NMEs: 6%
Distribution of Drugs on the Market
(698 oral IR) vs. Small Molecule NMEs
Class 2 Marketed Drugs
33%
NMEs: 54%
NME percentages
from a data set of
28,912 medicinal
chemistry
compounds tested
for at least one target
and having affinities
at μM or less
concentrations.
[LZ Benet, F Broccatelli and TI Oprea, AAPS J. 13: 519-547 (2011)]
[Broccatelli et al., Mol. Pharmaceut. 9: 570-580 (2012)]
High Solubility Low Solubility H
igh
P
erm
eab
ilit
y
Lo
w
Pe
rmeab
ilit
y Class 1 Marketed Drugs
40%
NMEs: 18% 79.3% (55.8%)
Class 3 Marketed Drugs
21%
NMEs: 22% 90.1% (74.8%)
Class 4 Marketed Drugs
6%
NMEs: 6% 39.5% (82.7%)
Distribution of Drugs on the Market
(698 oral IR) vs. Small Molecule NMEs [LZ Benet, F Broccatelli and TI Oprea, AAPS J. 13: 519-547 (2011)]
Class 2 Marketed Drugs
33%
NMEs: 54% 78.1% (85.1%)
Best in silico
solubility vs
measured
VolSurf+
r2=0.33
(ALOGPS
r2 = 0.24)
High Solubility Low Solubility H
igh
P
erm
eab
ilit
y
Lo
w
Pe
rmeab
ilit
y Class 1 Marketed Drugs
40%
NMEs: 18%
Class 3 Marketed Drugs
21%
NMEs: 22%
Class 4 Marketed Drugs
6%
NMEs: 6%
Distribution of Drugs on the Market
(698 oral IR) vs. Small Molecule NMEs [LZ Benet, F Broccatelli and TI Oprea, AAPS J. 13: 519-547 (2011)]
Class 2 Marketed Drugs
33%
NMEs: 54%
For either
measured or
calculated
Log P >2, the
probability of
extensive
metabolism is
79.9 and 81.0 %,
respectively.
For Log P
values < 0 the
probability of
poor metabolism
is 84.0 % for
M Log P and
83.4 % for
CLogP.
Figure 7 When Log P
values range
from 0-2
(31.4% of
drugs for
M Log P
and 27.5%
of drugs for
CLogP)
What about the new “combinatorial
chemistries” that drug discovery and
development scientists are enamoured
with today and with which the regulatory
agencies must grapple?
1. In silico drug discovery and development
2. Pharmacogenomics--Reviewed yesterday
3. Biomarkers
4. Systems biology
5. Transporters
6. Nanomedicine and targeted drug delivery
7. Preclinical alternatives to animal studies
What about the new “combinatorial
chemistries” that drug discovery and
development scientists are enamoured
with today and with which the regulatory
agencies must grapple?
1. In silico drug discovery and development
2. Pharmacogenomics
3. Biomarkers
4. Systems biology
5. Transporters
6. Nanomedicine and targeted drug delivery
7. Preclinical alternatives to animal studies
What about the new “combinatorial
chemistries” that drug discovery and
development scientists are enamoured
with today and with which the regulatory
agencies must grapple?
1. In silico drug discovery and development
2. Pharmacogenomics
3. Biomarkers
4. Systems biology
5. Transporters
6. Nanomedicine and targeted drug delivery
7. Preclinical alternatives to animal studies
High Solubility Low Solubility H
igh
P
erm
eab
ilit
y/
Me
tab
oli
sm
Lo
w
Perm
eab
ilit
y/
Me
tab
oli
sm
Class 1 Transporter
effects minimal in
gut and liver
Class 3 Absorptive
transporter effects
predominate (but can
be modulated by efflux
transporters)
Class 4 Absorptive and
efflux transporter
effects could be
important
Oral Dosing Transporter Effects
Class 2 Efflux transporter
effects predominate in
gut, but both uptake &
efflux transporters
can affect liver
S. Shugarts and L. Z. Benet. Pharm. Res. 26, 2039-2054 (2009).
Class 1
highly soluble, high permeability,
extensively metabolized drugs
• Transporter effects will be minimal
in the intestine and the liver
• Even compounds like verapamil that
can be shown in certain cellular
systems (MDR1-MDCK) to be a
substrate of P-gp will exhibit no
clinically significant P-gp substrate
effects in the gut and liver
Class 2
poorly soluble, highly permeable,
extensively metabolized drugs • Efflux transporter effects will be important
in the intestine and the liver
• In the intestine efflux transporter –enzyme
(CYP 3A4 and UGTs) interplay can markedly
affect oral bioavailability
• In the liver the efflux transporter-enzyme
interplay will yield counteractive effects to
that seen in the intestine.
• Uptake transporters can be important for the
liver but not the intestine.
Class 3
highly soluble, low permeability,
poorly metabolized drugs
• Uptake transporters will be
important for intestinal absorption
and liver entry for these poor
permeability drugs
• However, once these poorly
permeable drugs get into the
enterocyte or the hepatocyte efflux
transporter effects can occur.
Recent Transporter Interplay
Reviews from the Benet Lab
The Role of Transporters in the Pharmacokinetics
of Orally Administered Drugs. S. Shugarts and
L. Z. Benet. Pharm. Res. 26, 2039-2054 (2009).
The Drug Transporter-Metabolism Alliance:
Uncovering and Defining the Interplay. L. Z. Benet.
Mol. Pharmaceut. 6, 1631-1643 (2009).
Predicting Drug Disposition via Application of a
Biopharmaceutics Drug Disposition Classification
System. L. Z. Benet. Basic Clin. Pharmacol. Toxicol.
106, 162-167 (2010).
Improving the Prediction
of the Brain Disposition
of Orally Administered Drugs
Using BDDCS
F. Broccatelli, C.A. Larregieu, G. Cruciani,
T.I. Oprea and L.Z. Benet
Advanced Drug Delivery Reviews
64: 95-109 (2012)
Prior to our paper earlier this year, the best
methodologies for predicting central effects of
drug candidates had an accuracy of ~80%,
based on lipophilicity, molecular structure and
susceptibility to transport by brain efflux
transporters (P-glycoprotein and BCRP).
With our recognition that BDDCS class 1
compounds would exert central effects
even if they are substrates for efflux
transporters, we have been able to cut the
lack of predictability in half.
Class 1 Drugs A major proposition of
BDDCS is that Class 1,
P450/UGT metabolized drugs
are not substrates of clinical
relevance for transporters
in the intestine, liver,
kidney and brain.
Another Implication
Class 1 compounds will
achieve brain concentrations
whether this is desired or not
for an NME, which could be
the rationale for not always
wanting Class 1 NMEs.
How can the concepts presented be
used in predicting DMPK of an NME?
• In silico methodology, at present, is not sufficient (except possibly for Vss). We cannot predict clearance & bioavailability
• Permeability measures in Caco-2, MDCK or PAMPA vs metoprolol or labetalol will predict Class 1 & 2 vs. 3 & 4 and thus major route of elimination in humans.
• Is solubility over the pH range 1-7.5 more than 0.2 mg/ml ( i.e., 50 mg highest dose strength) as proposed by Pfizer scientists, defining Class 1 & 3 vs. 2 & 4?
Potential DDIs Predicted by BDDCS
• Class 1: Only metabolic in the intestine and liver
• Class 2: Metabolic, efflux transporter and efflux transporter-enzyme interplay in the intestine. Metabolic, uptake transporter, efflux transporter and transporter-enzyme interplay in the liver.
• Class 3 and 4: Uptake transporter, efflux transporter and uptake-efflux transporter interplay
What about the new “combinatorial
chemistries” that drug discovery and
development scientists are enamoured
with today and with which the regulatory
agencies must grapple?
1. In silico drug discovery and development
2. Pharmacogenomics
3. Biomarkers
4. Systems biology
5. Transporters
6. Nanomedicine and targeted drug delivery
7. Preclinical alternatives to animal studies
Yesterday, we heard Dr. Stavchansky describe
the exciting potential of nanotechnololgy in drug
delivery, diagnostics and nanomedicine, as well
as our ambition to achieve targeted drug
delivery. However, it is important to recognize
that these techniques must be far more
successful than even proposed by Dr.
Stavchansky if we are to solve major drug
delivery issues.
This leads me to comment on the very
unrealistic projections being made for
macromolecular drugs.
We are all well aware of the concern
with the limited number of new drug
approvals, and with the statistics
suggesting that of the approvals in the
last few years, macromolecular drug
products range from 25 to 40%.
This has led some observers, particularly
venture capitalists, to predict that the future
of drug development is in macromolecular
products and to estimate that 60-70% of future
drug approvals will be macromolecules.
But in my opinion this is a poor
prediction and can only be true if we can
discover how to deliver macromolecules
orally and relatively inexpensively.
The monoclonal antibody products being
approved now are the low hanging fruit that
address primarily cancer treatments where we
can tolerate expensive parenteral or complex
drug delivery processes.
In fact, I predict that the fraction of new drug
approvals for macromolecules will decrease
rather than increase.
What about the new “combinatorial
chemistries” that drug discovery and
development scientists are enamoured
with today and with which the regulatory
agencies must grapple?
1. In silico drug discovery and development
2. Pharmacogenomics
3. Biomarkers
4. Systems biology
5. Transporters
6. Nanomedicine and targeted drug delivery
7. Preclinical alternatives to animal studies
Regulatory pressure, public
clamor against animal testing,
especially in Europe:
“REACH” initiative: Comprehensive EU legislation regulating
chemical substances has brought controversy because it
portends massive increases in animal testing unless suitable
non-animal alternatives are implemented.
EU also passed sanctions against products relying on animal
testing
Sanctions become effective in tranches: 2009, 2013
I did not provide a Financial Disclosure
but I want to discuss some recent
advancements by Hμrel Corporation,
a company that I co-founded and for
which I serve as Chair of the SAB
PETA honored Hμrel with its “Proggy”
Award for Best Scientific Achievement 2010
Liver
Lung Gas Exchange
Other Tissues (Non-metabolizing, non-
accumulating)
Adipose
In
Out
The original vision: an “human on a chip”
(i.e., an in vitro, multi-tissue, microfluidic, cell-
based assay platform for improved
pharmacological / toxicological prediction)
Working together with bioengineers and
experts in liver microstructure we
developed microfluidic, cell-based biochips
Individual compartments contain cultures of living cells of different organs
Heterogeneous cell types mimic different organs or tissues of an animal (and humans)
Compartments fluidically interconnected
Fluid and compounds recirculate
as in a living system
(See Nature, 435: 12-13, May 5, 2005;
Forbes, August 15, 2005, pp. 53-54;
The Observer, September 25, 2005, p.7;
Newsweek, October 10, 2005, p.59
Nature, 471: 661-665, March 31, 2011)
Photo of early prototype silicon biochip
From the web site: www.hurelcorp.com
The metabolic competency of Hμrel®’s patented flow-based biochips
and instrumentation enables parent compound clearance and
metabolite generation significantly greater than that afforded by
hepatocytes cultured under static conditions, when presented
to compounds that represent a broad range of Phase 1 and
Phase 2 enzymes. Hμrel®’s flow-based metabolic competency
has been shown stable for 14 days and longer.
Before I close, I would like to present our
laboratories latest efforts to improve and
shorten the drug development process.
There has been a marked increase in the utilization
of pharmacokinetic-pharmacodynamic modeling in
the drug development process, with a strong
interest in this area from the regulatory
authorities.
Historically, such PKPD modeling has considered
the hysteresis between measured concentrations
and PD measures, as characterizing
“direct” and “indirect” relationships.
Direct Indirect
Pharmacokinetic- Pharmacodynamic Modeling Direct vs. Indirect PKPD
distribution to the site of action
slow turnover and transduction
processes
site of action is the central circulation
rapid receptor binding, turnover, and transduction
mechanisms
increasing t1/2,ke0
increasing t1/2,kout
Can a Single Framework Describe Benefit-Toxicity
Relationships for Many Diverse Drugs? A. Grover & L.Z. Benet, J. Clin. Pharmacol., in revision
Today each NME, particularly a first-in-class therapeutic, is investigated de novo with respect to choosing the appropriate dose and dosing regimen. We reasoned that for drugs showing a direct and rapid response to drug levels (both benefit and adverse effects) there should be a general relationship between drug levels above an effective (or toxic) concentration measure (e.g., EC50) and the appropriate dosing interval and dose. Many drugs will not be direct effect, but we reasoned that a continuum should exist between direct and indirect response models.
Therapeutic index concepts are exclusively
Can a Single Framework Describe
Benefit-Toxicity Relationships
for Many Diverse Drugs? A. Grover & L.Z. Benet, J. Clin. Pharmacol., in revision
Here we show that for 17 diverse drugs (19 evaluations) that a ED50 Relevance Parameter defined as the fraction of the dosing interval (τ) when drug levels are not above EC50:
(τ - Time concentrations over EC50) / τ
can be related to keo or kout
Results
5 10 150.0
0.5
1.0
Prednisolone
LorazepamEtodolac
LevodopaTerazosin
Dexamethasone
Terbutaline
AtropineRocuronium
Dexamethasone
Rosuvastatin
Ibuprofen
PrednisoloneEtodolac
Bilastine
Mizolastine
Ranitidine
Tocilizumab
kout, kout regression
ke0, ke0 regression
ke0 & kout combined regression
[EC50RP = 1.1 exp(-0.46 (ke0 or kout)) - 0.17]
ke0 or kout (hr-1)
EC
50 R
ele
va
nce
Pa
ram
ete
r
(Ta
u -
tim
e o
ve
r E
C5
0)/
Ta
u
Results
0.001 0.01 0.1 10 100
0.5
1.0
kout, kout regressionke0, ke0 regression
Prednisolone
LorazepamEtodolac
LevodopaTerazosin
Dexamethasone
Terbutaline
AtropineRocuronium
Dexamethasone
Rosuvastatin
Ibuprofen
PrednisoloneEtodolac
Bilastine
Mizolastine
Ranitidine
Tocilizumab
ke0 & kout combined regression
[EC50RP = 1.1 exp(-0.44 (ke0 or kout)) - 0.17]
ke0 or kout (hr-1)
EC
50 R
ele
va
nce
Pa
ram
ete
r
(Ta
u -
tim
e o
ve
r E
C5
0)/
Ta
u
Can a Single Framework Describe Benefit-
Toxicity Relationships for Many Diverse
Drugs? A. Grover & L.Z. Benet, J. Clin. Pharmacol., in revision
Here we show that for 17 diverse drugs (19 evaluations) that a
ED50 Relevance Parameter defined as the fraction of the dosing interval (τ) when drug levels are not above EC50:
(τ - Time concentrations over EC50) / τ
can be related to keo or kout
But we also have identified 3 drugs that do not fit the relationship. In two of those cases we believe that an inappropriate PD measure was chosen, but the third drug may indicate that the relationship is not general. We need others to confirm or invalidate the relationship with company data.
Benefit-Toxicity Relationships
Dosing regimens that fall along the log-linear regression balance between over-dosing [below the regression line] and under-dosing [above the regression line]
0.001 0.01 0.1 10 100
1.0
over-dosed:increased potential for
adverse eventsassociated with highdrug concentrations
under-dosed: effects(therapeutic or toxic) are not
evident because concentrationsare not above the EC50 for
sufficient time
ke0 or kout (hr-1)
EC
50 R
ele
va
nce
Pa
ram
ete
r
(Ta
u -
tim
e o
ve
r E
C5
0)/
Ta
u