|
Q: What information and experiments are most useful for setting up the best assay for HTS
and lead selection? What can and should be known about the
time-course of the biochemistry or cell biology being measured?
A: The time courses illustrated show a few of the wide variety of possibilities for patterns frequently observed in biological assays. The major point is that there is extremely valuable information contained in time course measurements that may be applied for the design of information-rich assays useful in HTS and/or follow-up assays, including as the basis for dose-response analyses. Carefully designed cell assays can be as informative and understandable as biochemical reactions, although of course more work is involved in the characterization, since cell systems are usually encumbered by multiple processes occurring simultaneously. Put your cursor over the box associated with each curve to see an interpretation of the time course.
Q: What happens if the wrong time-course is chosen as the measurement for screening or characterization?
A: The following graph shows what can happen when a dose-response is generated without paying sufficient attention to the time interval used in the dose-response analysis. All three dose-response curves are from experiments using the same enzyme and the same inhibitor, but the inhibition appears to differ significantly. The only difference among the three curves is that a time of 10 min was used in the left-most curve, a time of 30 min in the center curve, and a time of 80 min in the right-most curve. There are two effects: a shift in IC50 of about 4-fold and - more subtle - a change in the shape of the dose-response curves. The shape-change is readily evident, if non-linear least squares analysis is used to analyze the data. The practical effect is to make the assay less sensitive to detection of inhibitors, which can be a critical issue in running a screen. The problem is strictly because a less-than-optimal time interval was chosen when designing the experiment and reading the data for the 30 min and 80 min time intervals.
A: Typical screening results for an
inhibition assay using 30 x 384-well plates and 32 control wells per plate.
The results from control wells are not
included in this view, so there are 10560 individual readings shown on a
nominal inhibition scale on the left.
Note that it is NOT a percentage scale.
See elsewhere on this website for a discussion of the problems
associated with converting to percentages for many data.
Questions include: How should a cut-off be chosen for separating
compounds with enough activity to warrant follow-up from those of no further
interest?
Using typical 50% cut-off yields 51 "hits".
Using 3 SD from control -> 257 "hits", and using 3 SD from the data
shown, which is all the compound-containing wells without the controls ->
159 "hits". But which is correct?
What information is really in the screening results?
Data from the previous graph are redrawn as a
probability density function, i.e. the results on the x-axis are the
same as the Inhibition results shown on the y-axis of the previous graph, and
the number of results in a narrow band of inhibitions is represented on the
y-axis across the range of results to generate the histogram.
The dark blue symbols and the light blue line show the histogram for
the pooled results. Within the pooled
results are two classes: results from compounds that do not affect the assay
and result in a normal or Gaussian distribution with a mean of 0 inhibition
and results from compounds that inhibit the assay with a wide range of
efficacies. A useful distribution to
describe the latter class of results is a Gamma Distribution.
Together the pooled results are distributed as a particular Mixture
Distribution, which in this case is a "mixture" of the normal and gamma
distributions. The pooled results are
deconvoluted to show the underlying normal distribution (dark blue) and the
gamma distribution (red). The
difficulty technically has been finding a method to deconvolute the mixture
distribution. This problem can be
solved (submitted for publication) for cases like the one shown and provides a
new tool to understand screening data.
Note that the analysis is independent of the controls. The Normally
distributed data serve as an internal control and, in fact, should mimic the
uninhibited controls in this example. Reasons for differences between the
uninhibited controls and the Normally distributed part of the Mixture
Distribution should be investigated.
A close-up view of the
right-hand tail of the mixture distribution and its deconvoluted parts.
The mixture distribution and the normal part of the mixture
distribution right-hand tails are shown in light blue and dark blue,
respectively. The red line shows the
gamma distribution for compounds inhibiting the assay.
Using the deconvoluted mixture distribution, statistically defined
cut-offs can be determined based on probabilities of a result from a compound
coming from the normal distribution i.e. not inhibiting the reaction
or from the gamma distribution i.e. inhibiting the reaction.
The parameters for the deconvoluted distributions are: 90% of the total
results are ~N(0,95) [Normal distribution with mean = 0 and standard deviation
= 95] and
10% of the total results
are ~γ(200,150).
The parameters for the mixture are ~Mix(21,119).
Be aware that apparent high potency for compounds
detected in screening does not necessarily correlate with drug-like or
drugable properties. In fact, many of
the compounds with apparent high potency are really just compounds that
interfere in the assay, including such problems as insolubility or other
complicating physical chemical properties.
This serves to emphasize the importance of broadening the range of hits
to detect apparently lower potency compounds that may not have such
limitations.
Based on results from analysis of the Mixture
Distribution, statistically-defined cut-off values may be determined, such as
the "Optimized" cut-off shown, which is a value where the maximum number of
points (maximum area) of the Gamma Distribution are included for the minimum
area of the Normal Distribution. In
this example, the area to the right of the "optimized" cut-off includes, for
the total 10560 inhibition readings, 360 and 78 readings from the Gamma and
the Normal Distributions, respectively.
Another interesting value that may be used for a cut-off is where 99%
of the results are from the Gamma Distribution, which in this example includes
the inhibition readings from 170 compounds.
Basing cut-offs on the deconvoluted Mixture Distribution provides a
firm statistical basis for setting cut-off levels in screening in contrast to
the more arbitrary methods usually used.
In many assays, the results with low-level
inhibition do not verify on replicate determinations in follow-up assays to
the initial screening.
Insight into the failure-to-verify "problem" depends on
application of fundamental statistics. The difficulty in verification assays
is the problem of comparing two mean values to determine if they are
significantly different or if the results are attributable simply to
variability in the assay. In this
case, one of the mean values is from the uninhibited controls or is -
probably most appropriately - the mean of the Normal Distribution values
determined from the analysis of the Mixture Distribution.
The second mean for the comparison is a mean of results from a single
compound. We need to determine how many values should be in the verification
sample, i.e. the number of replicate measurements for each compound,
in order to have a reasonable probability of observing a statistically
significant difference from the background of results with a mean of 0 and a
standard deviation of 95. To address
this question, we must decide how certain we want to be that we have
observed the true result in the face of the variability inherent in the
assay. A full description of the
necessary calculations includes consideration of Type I and Type II errors,
and many standard statistics text books may be used to guide the reader to
the necessary level of background and understanding.
For the present, we will assume that we want to be sure to have
gathered enough samples to observe a statistically significant difference
90% of the time (Type II error or beta), if the difference really exists,
and to do so with a statistical significance (Type I error or alpha) = 0.05.
On the inhibition scale shown on the preceding graphs, the
"optimized" difference has a value of 228.
In order to be able to detect a difference 90% of the time, as
specified above, requires at least 5 replicates in the verification test,
and to detect the difference 99% of the time would require at least 7
replicate measurements of the test compound.
In order to detect the difference at the 99% Gamma level requires
at least 4 replicates in the verification test for the 90% probability of
detection and at least 5 replicate measurements in order to have a 99%
probability of detecting a difference with a magnitude of 338
inhibition units in the example given.
In both cases, the mean and standard deviation of the results needs
to be incorporated into a t-test or a similar statistical test in order to
test the verification results. The
common practice of using only two or three replicate measurements in
verifying activity will frequently not be sensitive enough to detect small
but statistically significant differences, such as between the mean control
value of 0 versus the cut-off values of 228 and 338 in this example.
Sometimes, a criterion applied to "verification" is 3SD from the
control mean for all replicates in the verification assay, which is
considerably more stringent than a statistically significant difference
based on the mean of verification replicates. In many cases the application
of such criteria account for failures in verifying activity.
Q: In screening, what are the benefits and hazards of cell assays in
contrast to assays based on isolated proteins or other isolated cell
constituents, such as enzymes and receptors?
What is the relationship between "high content" screening and
cell-based assays?
Use of cellular assays in HTS
Cell assays are essential
for measuring certain types of biological responses, such as ion channel
activity and for testing gene expression. Cell assays can also be very useful
in examining intracellular signal transduction pathways. However, in order to
be most useful in screening, cell assays come with price, just like all other
assay formats. The general rule is that the more information we have about an
assay, the more we can learn about particular responses. This will help to
determine the best characteristics for design and interpretation of a screening
assay. Many operators tout cell assays as being high throughput and high
information. While sometimes true, cell assays are not a panacea for high
content screening. In fact, there are numerous complications to the
interpretation of cell assays. Some examples are described in the following
paragraphs, although this is certainly not an exhaustive description of the
topic.
Unless the desired response
is mediated by a cell surface receptor, ion channel, or similar cell surface
molecule, only compounds that can already enter cells will be detected as
"hits".
Problem: Compounds
that may be very specific for an intracellular target, but deficient in cell
permeation, will not be detected. Although solving the problem of cell
permeability for many compounds is not trivial, is it wise to eliminate all such
compounds at the first step of a screening process? This question should be
carefully addressed before embarking on a cell-based screening assay. Only
compounds active in the whole cell environment will be detected.
Problem: If the goal
of screening is to find a drug, this is a reasonable stipulation. However,
leads that suffer from diminished activity in a cellular environment may not be
detected at all. Is it better to detect compounds that are active against a
target, even in less than ideal - i.e. cellular - conditions first and
then discover that lack of activity is due to cellular conditions that differ
from the original assay conditions including problems of penetration into the
target cellular compartment or is it most important to have a compound
obtained directly from primary screening with a list of apparently desirable
properties? The answers vary with the target and with the knowledge obtained
and obtainable from the biological system containing the target. Nevertheless,
it seems important to point out that a cellular screening assay is not
inherently superior to an assay based on an isolated receptor, enzyme or other
target molecule.
Isn't it obvious that
screening seeks to find compounds that can get into a cell? Ultimately, yes, a
compound must be able to reach its target. However, the target does not even
have to be intracellular to be difficult to reach. An obvious example is
targeting cell surface receptors in the CNS, which must cross the blood-brain
barrier before reaching their cell surface targets. With very rare exceptions,
all compounds must be able to penetrate, not just a single cell membrane at the
target cell, but also must be able to traverse one or more cell layers, that is,
they must pass into AND out of cell membranes on the path from, e.g., the
gut to a target tissue and target cells. Activity in a cell culture assay does
not guarantee that a compound will be able to traverse all cell and membrane
barriers needed to reach both its tissue and its intracellular target. This is
why a cellular assay is not necessarily superior to an assay of isolated target
molecules.
What information can be
obtained from a cellular assay that cannot be obtained from an assay based on
isolated molecules? Obviously, a cell assay is dependent on action in a
system. That is, multiple steps must be completed successfully for activity to
be detected. However, one or more of the steps may present a barrier that must
be identified and specifically overcome for a compound with activity at the
target to be identified. To understand the cellular system requires further
experiments, and if the compounds chosen from a screen were already restricted
to a very few with a long list of physical properties that are difficult to
expect in an initial screening hit, the game may be over before it is started.
Often, a screening effort starting with a cellular assay will turn up no usable
leads, and then efforts will have to be redirected to a more targeted and
non-cellular assay system in order to allow biologists and chemists to break the
problem down into more manageable segments, such as specificity and efficacy at
one or more target molecules, then penetration of the compounds into appropriate
cellular compartment, and optimizing activity for intracellular conditions.
Problem: Many
compounds apparently acting in a cell assay are really operating via a process
that is not useful for drug or marker development, such as via aggregation
leading to changes including disruption of the cell membrane. Such effects
almost always require the use of two or often multiple cell lines tested
simultaneously or sequentially in order to detect compounds that act on cells in
a manner inconsistent with the desired biology. This increases the workload in
screening, because it is equivalent to running a couple or several screens.
The bottom line is that in
order for a cell-based assay to be useful for high-content screening, there is
usually a considerable amount of work needed in order to be able to design an
effective, efficient and sensitive HTS assay and the required follow-up assays.
If the work is not done before screening is begun, the interpretation of
screening data will be difficult or impossible. Because cellular assays almost
always require more biological steps, including mandatory penetration into the
proper cell compartment, in generating the response the screener wants to
measure, the information required to interpret the results from the assay are
almost always more complicated than the problems faced in using isolated
proteins or other cell fractions. In order for a cell-based HTS assay to be
truly "high content" requires that significant details of the system be worked
out and applied to the design of the screen and interpretation of the data
resulting from the screen.
Q:
What are
pharmacokinetics and ADME and why are they important?
A:
The function
of a drug is dependent on its efficacy at a specific site of action, how much
gets to the site, and how long it stays at the site. Pharmacology, the science
of drugs, encompasses two areas: pharmacodynamics (PD) & pharmacokinetics (PK).
PD is how a drug acts on a therapeutic target. PK encompasses what is
normally referred to as ADME, which are absorption, distribution, metabolism and
elimination. In layman terms, PD is the study of what the drug does to the body
and PK is the study of what happens to the drug in the body. Understanding PK
and PD among drug leads and candidates is critical to choosing the compounds
most-likely-to-succeed through the process of drug development. Early
assessment of interesting compounds is essential for holding down drug
development costs and maximizing speed to market.
Q:
When should ADME
characteristics be determined in the drug discovery funnel ?
A:
There is a
natural overlap between drug discovery research and preclinical development.
For example, the determination of whether a compound inhibits human CYP 3A4 is
an important experiment to conduct in either domain, since CYP 3A4 inhibition
affects both candidate selection following high-throughput screening (HTS) and
is a necessary part of preclinical development for satisfying regulatory
requirements. Therefore, a number of ADME estimations should be made early in
drug discovery, as soon as possible following selection of compounds from HTS (see Figure below). That way the most attention
and expensive resources can be focused on compounds with the greatest potential
for success from early in the drug development process. In addition, coupling ADME measurements with medicinal chemistry efforts can help guide the chemistry
in preserving and enhancing the best possible drug-like properties. This way,
fewer compounds with poor drug characteristics are taken to the stage of
engaging considerably more expensive GLP studies prior to IND and NDA regulatory
filings. Modern ADME research involves a determination of when to do particular
studies and when to incorporate data acquisition into a high-throughput format
so that factors determining if a candidate is dropped or moved further along the
development funnel can be incorporated early in the discovery and development
funnel to hold down the costs of the whole process. Ask PharmOptima to help you
in determining which tests should be done when in your drug development
process.
The
graph below is a representative drug discovery & development funnel of
activities from the ADME perspective. Activities range from discovery HTS
through preparation for IND regulatory filing.
Q: Why are ADME studies so
important to my company's success?
A: Since
the determination of human ADME drug characteristics can only follow in very
expensive Phase I, II and III clinical trials, estimation of human ADME
characteristics using animal systems is an integral part of preclinical
scientists responsibilities. To put it another way, the identification of
compounds that will ultimately fail in man is the job of the preclinical
scientist. "We kill compounds!" but at a much reduced cost than is encumbered
by carrying forward compounds that are doomed to failure.
Q:
What is the experience
level of the Drug Metabolism & Pharmacokinetic Research scientists?
A:
The PharmOptima Drug Metabolism &
Pharmacokinetic Research staff is highly experienced in conducting ADME studies.
We function as part of PharmOptima's First-in-Animal capabilities, as well as
operating with separate capabilities to support customer needs. PharmOptima
scientists offer our customers a historical record of successful preclinical ADME research characterized by pairing the identification and application of
emerging technologies with solid scientific expertise and experience.
PharmOptima has
four scientists with over 100 years of combined pharmaceutical industry
experience who have written more than 1000 scientific reports covering an
extraordinary range of compounds and drugs. PharmOptima offers extensive
expertise in non-GLP studies that should be part of your early preclinical
development assessment. These services include such important determinations as
early assessment of bioavailability, drug-drug interaction liability, in
vitro metabolic stability, allometric scaling and in vivo
metabolic
profiling. In addition to direct consulting and direct laboratory application
in non-GLP studies, we can help you determine and plan studies so that your
costs in the expensive arena of GLP studies are minimized.
Selected Peer-Reviewed
Publications of ADME Staff
With ADME and Analytical
Emphasis
1. M.A.
Wynalda, F. Lincoln, and
F.A. Fitzpatrick, "High Performance Liquid Chromatographic Analysis of
Prostaglandin I2 (Prostacyclin)," Journal of Chromatography 176,413‑417 (1979).
2.
Stryd RP,
Gilbertson TJ. Some problems in development of a high-performance liquid
chromatographic assay to measure 25-hydroxyvitamin D2 and 25-hydroxyvitamin D3
simultaneously in human serum," Clin. Chem. 1978;24:927-30.
3.
Stryd RP,
Gilbertson TJ, Brunden, MN. A seasonal variation study of 25-hydroxyvitamin D3
serum levels in normal humans. J. Clin. Endocrinol. Metab. 1979;48:771-5.
4.
M.A. Wynalda
and F.A. Fitzpatrick, "Albumin Stabilizes Prostaglandin I2," Prostaglandins 20,
853‑861 (1980).
5.
M.A. Wynalda
and F.A. Fitzpatrick, "High Performance Liquid Chromatographic Determination of
5‑Halo‑Pyrimidinone Interferon Inducers," Analytical Chemistry 52, 1931‑1934 (1980).
6. M.A.
Wynalda, D.R. Morton, R.C.
Kelly, and F.A. Fitzpatrick, "Liquid Chromatographic Analysis of Intact
Leukotriene A4", Analytical Chemistry 54, 1079‑1082 (1982).
7.
Vrbanac, J.J.,
W.E. Braselton, J.F. Holland and C.C. Sweeley. Automated qualitative and
quantitative metabolic profiling analysis of urinary steroids by a gas
chromatography-mass spectrometry-computer system. J. Chromatogr. 239:
265-276, 1982.
8.
Vrbanac, J.J.,
W.E. Braselton, J.D. Pinkston, and C.C. Sweeley. Automated metabolic profiling
analysis of urinary steroids by a gas chromatography-mass spectrometry-data
system. Biomed. Mass Spectrom. 10: 155-161, 1983.
9.
Gilbertson TJ, Ruwart MJ, Stryd RP,
Brunden MN, Friedle NM, Rush BD, Christianson CA. Partial characterization of
the gastrointestinal weight changes produced in the female rat by 16,16-dimethyl
E2. Prostaglandins 1983;26(5):745-59.
10.
Peng GW, Sood VK. Liquid
Chromatographic Assay of Arbaprostil. J. Liquid Chromatography, 6 (8),
1499‑1511 (1983).
11.
Peng GW, Sood VK and Rykert
UM. Quantitative Liquid Chromatographic Determination of Bromadoline
and its N‑Demethylated Metabolites in Blood, Plasma, Serum and Urine Samples. J.
Pharm. Sci., 74, 304‑307 (1985).
12.
M.A. Wynalda,
J.R. Brashler, M.K. Bach, D.R. Morton, and F.A. Fitzpatrick, "Determination of
Leukotriene C4 by Radioimmunoassay with a Specific Antiserum Generated from a
Synthetic Hapten Mimic," Analytical Chemistry 56, 1862‑1865 (1984).
13.
Lakings DB, Stryd RP,
Gilbertson TJ. Quantitative determination of
N-(trans-2-dimethylamino-cyclopentyl)-N-(3',4'-dichloro-phenyl)-propanamide and
its N-dimethyl metabolite in dog serum by GC-EC. J. Pharm. Sci.
1984;73(3):317-20.
14.
Hubbard, L.H., T.D. Eller, P.V.
Halushka, J.J. Vrbanac, and D.R. Knapp. Extraction of thromboxane B2
from urine using an immobilized antibody column for subsequent analysis by gas
chromatograph-mass spectrometry. Prostaglandins, 33: 149-160,
1987.
15.
Vrbanac, J.J.,
T.D. Eller, and D.R. Knapp. Quantitative analysis of 6-keto-prostaglandins F1"
using immunoaffinity purification and gas chromatography-mass spectrometry. J.
Chromatogr., 425, 1-9, 1988.
16.
Cox JW, Larson PG, Wynalda MA,
Sood VK, Verburg MT, Pullen RH, and Daniels EG, "Pharmacokinetics and
Excretion of the 21‑Aminosteroid Antioxidant U‑74006F in Rat and Perfused Rat
Liver". Drug Metab Disp. 17(4), 373-379 (1989).
17.
Vrbanac, J.J.,
Cox, J.W., Eller, T.D. and Knapp, D.R. Immunoaffinity
Purification-Chromatographic Quantitative Analysis of Arachidonic Acid
Metabolites. Meth Enzymol. 187, 62-70, 1990.
18.
Greenfield JC, Loux SJ, Sood VK,
Jenkins KM and Davio SR,. (1991) In Vitro Evaluation of the Plasma and
Blood Compatibility of a Parenteral Formulation for Ditekiren, A Novel Renin
Inhibitor Pseudopeptide. Pharmaceutical Research 8: 475-479.
19. Vrbanac,
JJ, O'Leary, IA,
Baczynskyj, L. Utility of the parent-neutral loss scan screening technique:
Partial characterization of urinary metabolites of U-78875 in monkey urine.
Biol. Mass Spectrum, 21, 517-522, 1992.
20.
Wynalda M.A
.and Wienkers L.C. Assessment of Potential Interactions Between Dopamine
Receptor Agonists and Various Human Cytochrome P450 Enzymes Using a Simple In
Vitro Inhibition Screen. Drug Metabolism and Disposition, 25(10):1213-1217,
1997.
21.
Chang M, Sood VK, Wilson GJ,
Kloosterman DA, Sanders PE, Hauer MJ, Zhang W, Branstetter DG. Metabolism of
the HIV‑1 Reverse Transcriptase Inhibitor Delavirdine in Mice. Drug Metab
Dispos 1997; 25: 828‑839.
22.
Chang M, Sood VK, Kloosterman DA,
Hauer MJ, Fagerness PE, Sanders PE, Vrbanac JJ. Identification of the
metabolites of the HIV‑1 reverse transcriptase inhibitor delavirdine in
monkeys. Drug Metab Dispos 1997; 25: 814‑827.
23.
Chang M, Sood VK, Wilson GJ,
Kloosterman DA, Sanders PE, Hauer MJ, Fagerness PE. Metabolism of the HIV‑1
reverse transcriptase inhibitor delavirdine in rats. Drug Metab Dispos 1997;
25: 228‑242.
24. Wynalda
MA, Hauer MJ and Wienkers
LC. Human Biotransformation of Bropirimine. Characterization of the Major
Bropirimine Oxidative Metabolites Formed In Vitro. Drug Metabolism and
Disposition, 26(10):1048-1051, 1998.
25.
Wienkers LC, Allievi C, Hauer MJ and
Wynalda MA. Cytochrome P
-450 Mediated Metabolism of the Individual Enantiomers of the Antidepressant Agent Reboxetine in Human Liver Microsomes.
Drug Metabolism and Disposition 27(11):1334-1340, 1999.
26.
Peng, GW, Stryd RP, Murata S,
Igarachi M, Chiba K, Aoyama H, Aoyama M, Zenki T, Ozawa N. Determination of
linezolid in plasma by reversed-phase high-performance liquid chromatography.
Journal of Pharmaceutical and Biomedical Analysis 20, 65-73 (1999).
27.
Wynalda MA
and Wienkers LC. Assessment of Potential Interactions Between Dopamine Receptor
Agonists and Various Human Cytochrome P450 Enzymes Using a Simple In Vitro
Inhibition Screen. Drug Metabolism and Disposition, 25(10):1213-1217, (1997).
28.
T. Streeper, P. G. Pearson, Z. Zhao,
S. A. Mizsak, P. E. Sanders, J. J. Vrbanac. In vitro Metabolic
Transformations of 2,4-Dipyrrolidinylpyrimidine: A Chemical Probe for P-450
mediated Oxidation of Tirilazad Mesylate. Xenobiotica, 27,
1131-1145, 1997.
29.
Ballard KD, Orkiszewski RS, Vachet RW,
Glich GL, Vrbanac JJ, Gaskell SJ.. Parent ion resolution in linked scans
for dissociations occurring in the first field‑free region of sector mass
spectrometers. J. Am. Soc. Mass Spectrom, 8, 545-553, 1997.
30.
Chang M, Sood VK, Kloosterman
DA, Hauer MJ, Fagerness PE, Sanders PE, Vrbanac JJ. Identification of
metabolites of the HIV-1 reverse transcriptase inhibitor delavirdine in monkeys.
Dr
ug Metabolism and Disposition, 25, 814-827, 1997.
31.
Streeper, P. G. Pearson, Z. Zhao, S.
A. Mizsak and J. J. Vrbanac. Synthesis of Deuterium Labeled
2,4-Dipyrrolidinylpyrimidine as a Chemical Probe for P450 Mediated Oxidation of
Tirilazad Mesylate. J. Label. Coupounds Radio. Pharmaceut., XLI, 577-584,
1998.
32.
Chang M, Sood VK, Wilson GJ,
Kloosterman DA, Sanders PE, Scuette MR, Judy RW, Slatter JG. Absorption,
distribution, excretion, and metabolism of atevirdine in rats. Drug Metab Dispos
1998; 26: 1008‑1018
33.
Steenwyk
RC, Pearson PG, Vrbanac JJ. Metabolism of tirilazad mesylate in humans:
Application of stable isotopic labeling and a novel bile cannulation technique
in human volunteers. J. Label. Coupounds Radio. Pharmaceut, 42, 903-904,
1999.
34.
Wynalda MA,
Hauer MJ and Wienkers LC. Oxidation of the Novel Oxazolidinone Antibiotic
Linezolid in Human Liver Microsomes. Drug Metabolism and Disposition,
28(9): 1014-1017, (2000).
35.
Humphrey SJ,
Curry JT, Turman CN, Stryd RP. Cardiovascular Sympathomimetic Amine
Interactions in Rats Treated with Monoamine Oxidase Inhibitors and the Novel
Oxazolidinone Antibiotic Linezolid. Journal of Cardiovascular Pharmacology 37,
No. 5, 548-563, 2001.
36.
Slatter JG, Stalker DJ, Feenstra KL,
Welchman IR, Fagerness JB, Stryd RP, Peng GW, and Shobe EM.
Pharmacokinetics, Metabolism, and Excretion of Linezolid following an Oral Dose of
[C14]Linezolid to Healthy Human Subjects, Drug Metabolism and Disposition, 29, No.
8, 1136-1145, 2001.
37.
Wienkers LC and
Wynalda MA.
Multiple Cytochrome P450 Enzymes Responsible for the Oxidative Metabolism of
(-)-OSU6162 in Human Liver Microsomes. Drug Metabolism and Disposition
30(12):1372-1377 (2002).
38.
Wynalda KM, Amore BM, and
Wienkers LC. Characterization of the Major Bropirimine Glucuronide Metabolite
Formed In Vitro. Xenobiotica, Sept, 2003.
39.
Wynalda MA, Hutzler MJ, Koets
MA, Podoll T, and Wienkers LC. In Vitro Metabolism of Clindamycin in Human Liver
and Intestinal Microsomes. Drug Metabolism and Disposition August,
2003:31(7);878-887.
40. L.C. Wienkers, M.A. Wynalda, In
vitro metabolism of chlorzoxazone is mediated by human microsomal CYP1A2 and
CYP2E1E Pharmacogenetics, submitted
May, 2003.
Q: Now that we've identified some promising leads from our high throughput screen (HTS), what are the next steps in using these compounds to advance our drug discovery efforts?
A: Active compounds identified from an HTS assay (so-called "hits") must be carefully evaluated in silico and in vitro before they can be considered legitimate "leads" that act on the intended biochemical, receptor, or membrane target. Even after clearing this mechanistic hurdle, HTS leads rarely if ever qualify as true drug candidates. Rather, they represent key starting points for synthesizing unique patentable compounds having sufficient in vitro potency and specificity and in vivo safety and efficacy to warrant full-scale drug development. An essential step in deciding which of your preliminary leads offer the best chances of ultimately leading to a useful drug candidate is to compare their in vivo effects in a relevant whole animal model. Specific questions that should be addressed with each potential lead are:
- Can the compound be formulated and given in a therapeutically relevant manner?
- When given in vivo, do drug levels in the target tissues and/or fluid compartments reach concentrations that should exert biological activity?
- Do these agents demonstrate in vivo pharmacologic activity consistent with the intended disease target?
- Do the compounds have significant untoward effects or interactions that preclude their further exploration, or could such possible side effects be eliminated synthetically?
The more completely these in vivo questions can be answered, the better the chances of successfully narrowing the list of potential leads and finding the best template for targeted analog synthesis. While it is impossible to predict what in vivo activities might ultimately eliminate a lead from further consideration, the theoretical results tabulated below demonstrate several key aspects of formulation, stability, and in vivo pharmacology and pharmacokinetics (PK) that can be enormously helpful in selecting leads for further pursuit:
|
HTS Lead |
Aqueous Formulation |
Stability |
Total Dosage |
Drug Tolerance |
Pharmacol.
Activity |
PK
Profile |
Worth pursuit? |
|
A |
Very insoluble |
Good |
10 mg/kg |
Hypotension, bradycardia |
None seen |
Short
half-life |
No |
|
B |
Moderately soluble |
Good |
10 mg/kg |
No observed side effects |
Dose
dependence |
Good |
Primary lead |
|
C |
Slightly soluble |
Fair |
10 mg/kg |
CNS depression |
Weakly
Active |
Good |
No |
|
D |
Good solubility |
Poor |
50 mg/kg |
Diarrhea |
Inactive at all doses |
Short
half-life |
No |
|
E |
Low solubility |
Good |
30 mg/kg |
No observed side effects |
Active at high dose |
Low clearance |
Secondary lead |
|
F |
Very insoluble |
Poor |
10 mg/kg |
Hemolysis, hematouria |
None seen |
Low Cmax |
No |
Q: What animal models and experimental endpoints are deemed most important for evaluating my HTS screening leads?
A: In most cases rats are the best species for "First in Animal" drug evaluations. Ideally an experimental protocol should be customized to match the drug target and anticipated pharmacologic effects. At PharmOptima, four different rat protocols can be adapted to evaluate your HTS leads, namely, anesthetized or conscious rats during acute or chronic drug administration.
Basic endpoints include circulating and tissue drug levels, drug tolerance, pharmacologic indices of efficacy, and changes in primary cardiovascular parameters such as blood pressure and heart rate. Other more specialized cardiovascular, CNS behavioral, gastrointestinal, and renal endpoints can also be monitored. If the drug supply is limited, in most cases these types of studies can be scaled down to conscious or anesthetized mice. Once profiled in one or more rodent species, and if warranted by the drug target, PharmOptima can also devise efficient protocols to test your best leads in other small laboratory species such as guinea pigs, gerbils, or rabbits. PharmOptima can likewise work with you to design and implement pharmacologic studies in larger animal species such cats, dogs, or pigs working through our industry contacts. More specific disease models can also be implemented.
In all cases, PharmOptima scientists are experienced in formulating novel agents for in vivo administration, designing incisive protocols, appraising drug actions and interactions, and correlating pharmacodynamic effects with drug exposure (PD/PK relationships). By accurately characterizing your HTS leads in sensitive whole animal models and determining their attractive and limiting features, PharmOptima insures that you have the optimal template for future analog synthesis, thereby expediting your search for viable drug candidates.
Typical Conscious Rat Preparation for Conducting a "First in Animal" Study:
Q: How much drug is needed to conduct a meaningful "First in Animal" evaluation of an HTS lead for biological activity and pharmacokinetics?
A: Drug requirements for PharmOptima's First in Animal services depend on the species, group sizes, dosages, route and frequency of administration, stability of the drug, and ease with which it can be formulated and administered. An agent that results in highly variable responses will also require more material to accurately determine its profile. Another consideration is the level of assurance sought from the study, as protocols designed to generate larger and thus more accurate data will require considerably more drug. With these qualifiers in mind, as a general rule it is generally possible to profile a readily formulated agent in two adult rats at 30 mg/kg using only 20 mg of drug. The table below further estimates how much drug would be needed for a relatively straightforward rat or mouse study at selected group sizes and doses.
|
Parameter
|
Rats |
Mice |
|
Group size: |
n = 2 |
n = 4 |
n =6 |
n =2 |
n =4 |
n = 6 |
|
Total dose: |
15 mg/kg |
50 mg/kg |
15 mg/kg |
50 m/kg |
15 mg/kg |
50 mg/kg |
15 mg/kg |
50 mg/kg |
15 mg/kg |
50 m/kg |
15 mg/kg |
50 mg/kg |
|
mg for dosing: |
9.0 |
30 |
18 |
60 |
27 |
90 |
0.9 |
3.0 |
1.8 |
6.0 |
2.7 |
9.0 |
|
mg for formulation: |
5.0 |
5.0 |
5.0 |
5.0 |
5.0 |
5.0 |
5.0 |
5.0 |
5.0 |
5.0 |
5.0 |
5.0 |
|
Total mg needed: |
14 |
35 |
23 |
65 |
32 |
95 |
5.9 |
8.0 |
6.8 |
11 |
7.7 |
14 |
|
