Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02708311 2010-06-07
WO 2009/080219 PCT/EP2008/010530
-1-
PREDICTION OF GENOTOXICITY
This invention relates generally to the field of toxicology. More
particularly, the invention
relates to methods for predicting genotoxicity, and methods for screening
compounds for
potential genotoxicity.
The micronucleus test ("MNT") is a common assay in the pharmaceutical industry
routinely used to detect chromosome damage. A micronucleus forms when whole
chromosomes
or chromosome fragments do not incorporate into the daughter nuclei following
the completion
of mitosis. Aneugens and clastogens, chemicals which cause chromosomal
loss/gain and
breakage, respectively, will cause significant increases in micronuclei
formation and can be
detected using the assay. Thus, micronuclei are biomarkers of chromosome
damage and the
micronucleus assay is a sensitive method to detect chemicals which are
aneugens and/or
clastogens. The micronucleus assay is widely used in the pharmaceutical
industry as evidence of
genotoxicity (or lack thereof).
However, performing the micronucleus assay is laborious and time consuming,
false
positive results can occur when testing at cytotoxic doses, and large amounts
of supplies (cells,
reagents for cell-line maintenance, and compound) are required to perform the
assay.
Kinases are enzymes responsible for phosphorylating substrates and
disseminating inter-
and intracellular signals, including the initiation, propagation, and
termination of chromosome
replication during mitosis. Kinases are often targeted for inhibition by
pharmaceutical
companies because many signaling cascades have known roles in a variety of
diseases. Small
molecule kinase inhibitors (SMKIs) often are developed to competitively bind
to the kinase ATP
binding pocket, blocking the ability of the enzyme to phosphorylate
substrates. SMKIs often
inhibit many kinases in addition to the desired target due to the highly
conserved nature of the
ATP binding pocket within the kinome, thus toxicities associated with off-
target kinase
inhibition is a concern for this pharmaceutical class of compounds. In
particular, post-metaphase
genetic toxicity, manifested as positive micronucleus results, is a common
toxicological liability
for SMKIs.
CA 02708311 2010-06-07
WO 2009/080219 PCT/EP2008/010530
-2-
We have now invented a method for predicting which compounds will demonstrate
positive (i.e., genotoxic) results in a micronucleus assay, using a method
that is faster, uses
smaller quantities of reagents, and is easily automated.
The invention provides a method for quickly determining the likelihood that a
given
compound will exhibit genotoxicity in an MNT assay by examining the
interaction between the
compound and a number of kinases (kinase binding and/or inhibition). As kinase
inhibition
and/or binding can be determined quickly, and by using automated methods, the
method of the
invention enables high-throughput screening of compounds for genotoxicity (or
lack thereof).
In practice, binding and inhibition can be determined using methods known in
the art. See,
for example, M.A. Fabian et al., Nature Biotechnol (2005) 23:329-36,
incorporated herein by
reference in full.
One aspect of the invention is a method for predicting the genotoxicity of a
compound, the
method comprising providing a test compound; and determining the ability of
the compound to
inhibit the kinase activity of at least ten of kinases selected from the group
consisting of CDK2
(Seq. Id. 1), CLK1 (Seq. Id. 2), DYRKIB (Seq. Id. 3), ERK8 (Seq. Id. 4), GSK3A
(Seq. Id. 5),
GSK3B (Seq. Id. 6), PCTK1 (Seq. Id. 7), PCTK2 (Seq. Id. 8), STK16 (Seq. Id.
9), TTK (Seq. Id.
10), CLK2 (Seq. Id. 11), ERK3 (Seq. Id. 12), and PRKR (Seq. Id. 13), or the
group consisting of.
CDK2, CLK1, DYRKIB, ERK8 (MAPK15), GSK3A, GSK3B, PCTK1, PCTK2, STK16, TTK,
CDK7 (Seq. Id. 64), CLK4 (Seq. Id. 68), and PCTK3 (Seq. Id. 69), wherein
inhibition of at least
five of said kinases by at least 50% indicates a likelihood that said test
compound will
demonstrate genotoxicity.
In one embodiment said test compound is tested at a concentration of about 10
M.
In another embodiment the ability of the-compound to inhibit the kinase
activity of at least
twelve kinases selected from said group is tested. In another embodiment the
ability of the
compound to inhibit the kinase activity of all 13 kinases in said group is
determined.
In yet another embodiment, in addition to the groups of thirteen kinases
described above, at
least one of the following additional kinases can also be tested. High
affinity of a compound for
one or more of these additional kinases (in addition to a majority of the
thirteen identified
kinases) correlates with a higher likelihood of genotoxicity. The additional
kinases are: MKNK2
(Seq. Id. 14), SgK085 (Seq. Id. 15), PIM2 (Seq. Id. 16), TNNI3K (Seq. Id. 17),
KIT (Seq. Id. 18),
MELK (Seq. Id. 19), AURKA (Seq. Id. 20), CLK3 (Seq. Id. 21), AAK1 (Seq. Id.
22),
DCAMKL3 (Seq. Id. 23), LIMK1 (Seq. Id. 24), FLTI (Seq. Id. 25), MAP2K4 (Seq.
Id. 26),
PIM3 (Seq. Id. 27), AURKB (Seq. Id. 28), ERK2 (Seq. Id. 29), CSNKIAIL (Seq.
Id. 30),
CA 02708311 2010-06-07
WO 2009/080219 PCT/EP2008/010530
-3-
DAPK3 (Seq. Id. 31), MLCK (Seq. Id. 32), CLK3 (Seq. Id. 33), PFTK1 (Seq. Id.
34), PRKD3
(Seq. Id. 35), AURKC (Seq. Id. 36), ERK5 (Seq. Id. 37), STK17A (Seq. Id. 38),
MST4 (Seq. Id.
39), CDK3 (Seq. Id. 40), MYLK (Seq. Id. 41), CDC2L1 (Seq. Id. 42), QIK (Seq.
Id. 43), CDK11
(Seq. Id. 44), PLK1 (Seq. Id. 45), PDGFR(3 (Seq. Id. 46), PRKCM (Seq. Id. 47),
MAPK4 (Seq.
Id. 48), PIP5K2B (Seq. Id. 49), CSNKID (Seq. Id. 50), RPS6KA1 (KD1) (Seq. Id.
51), CDK5
(Seq. Id. 52), PLK3 (Seq. Id. 53), BIKE (Seq. Id. 54), PLK4 (Seq. Id. 55),
CAMK2A (Seq. Id.
56), STK3 (Seq. Id. 57), CSNK2A1 (Seq. Id. 58), STK17B (Seq. Id. 59), CDK8
(Seq. Id. 60),
MAP2K6 (Seq. Id. 61), PIM1 (Seq. Id. 62), MAP2K3 (Seq. Id. 63), CDK7 (Seq. Id.
64), IKKE
(Seq. Id. 65), TGFBR2 (Seq. Id. 66), CDK9 (Seq. Id. 67), CLK4 (Seq. Id. 68),
and PCTK3 (Seq.
Id. 69).
Another aspect of the invention is the method for screening candidate
compounds for
potential genotoxicity, comprising providing a plurality of compounds; and
determining the
ability of each compound to inhibit the kinase activity of a number of kinases
selected from the
group consisting of CDK2, CLK1, DYRKIB, ERK8, GSK3A, GSK3B, PCTK1, PCTK2,
STK16,
TTK, CLK2, ERK3, and PRKR, or the alternate group consisting of CDK2, CLK1,
DYRKIB,
ERK8 (MAPK15), GSK3A, GSK3B, PCTK1, PCTK2, STK16, TTK, CDK7, CLK4, and
PCTK3, wherein inhibition or specific binding of at least five of said kinases
by at least 50%
indicates a likelihood that said test compound will demonstrate genotoxicity.
In preferred
embodiments this method further comprises rejecting compounds that demonstrate
a likelihood
of genotoxicity.
In general, the binding affinity of a compound for a given kinase correlates
well with the
ability of the compound to inhibit the activity of that kinase, so that
binding affinity is a reliable
substitute for inhibitory activity. In preferred embodiments the ability of
the compound to inhibit
the kinase activity is hence determined by measuring the binding affinity of
the compound for
said kinases. Binding affinity may be determined by a variety of methods known
in the art; for
example by competitive assay using an immobilized kinase (or an immobilized
test compound,
or an immobilized competing ligand, any of which may be labeled). Compounds
and kinases can
be immobilized by standard methods, for example by biotinylation and capture
on a streptavidin-
coated substrate.
Thus, one can prepare a test substrate having, for example, a plurality of
immobilized
kinases, preferably comprising the thirteen kinases identified herein: CDK2,
CLK1, DYRKIB,
ERK8 (MAPK15), GSK3A, GSK3B, PCTKI, PCTK2, STK16, TTK, CLK2, ERK3, and PRKR
(or the alternate identified kinases). The kinases can be immobilized directly
(i.e., by adsorption,
CA 02708311 2010-06-07
WO 2009/080219 PCT/EP2008/010530
-4-
covalent bond, or biotin-avidin binding or the like) to the surface, or
indirectly (for example by
binding to a ligand that is tethered to the surface by adsorption, covalent
bond, biotin-avidin or
other linkage). The kinases are then contacted with the test compound(s), and
the affinity (or
enzyme inhibition) determined, for example by measuring the binding of labeled
compound or
loss of labeled competitor.
The kinase affinity of each compound is measured against at least ten of the
identified or
alternate identified kinases, and most preferably against all thirteen: use of
a larger number of
kinases (up to thirteen) results in a prediction of genotoxicity with higher
confidence. A
compound with high total activity (for example, demonstrating high affinity
for at least five of
the thirteen kinases, preferably eight or more) has a high likelihood of
genotoxicity: this
compound is predicted to test positive for genotoxicity in the MNT. A compound
having low
total activity (for example, showing only low affinity for the identified
kinases, or showing high
affinity to only 1-4 identified kinases) is predicted to test negative in the
MNT.
Candidate drugs that test positive in the assay of the invention (i.e., that
are predicted to
demonstrate genotoxicity in the MNT) are generally identified as "genotoxic"
or "potentially
genotoxic", and rejected or otherwise dropped from further development. In the
case of high-
throughput screening applications, such compounds can be flagged as toxic (for
example, by the
software managing the system in the case of an automated high-throughput
system), thus
enabling earlier decision making.
Thus, one can use the method of the invention to prioritize and select
candidate compounds
for pharmaceutical development based in part on the potential of the compound
for genotoxicity.
For example, if one has prepared a plurality of compounds (e.g., 50 or more),
having similar
activity against a selected target, and desires to prioritize or select a
subset of said compounds for
further development, one can test the entire group of compounds in the method
of the invention
and discard or reject all those compounds that exhibit positive signs of
genotoxicity. This
reduces the cost of pharmaceutical development, and the amount invested in any
compound
selected for development by identifying an important source of toxicity early
on. Because the
method of the invention is fast and easily automated, it enables the bulk
screening of compounds
that would otherwise not be possible or practical.
Environmental pollutants and the like can also be identified using the method
of the
invention, in which case such compounds are typically identified for further
study into their toxic
properties. In this application of the method of the invention, one can
fractionate an
environmental sample (for example, soil, water, or air, suspected of
contamination) by known
CA 02708311 2010-06-07
WO 2009/080219 PCT/EP2008/010530
-5-
methods (for example chromatography), and subject said fractions to the method
of the invention.
Fractions that display signs of genotoxicity can then be further fractionated,
and (using the
method of the invention), the responsible toxic agents identified.
Alternatively, one can perform
the method of the invention using pure or purified compounds that are
suspected of being
environmental pollutants to determine their potential for genotoxicity.
Because the method of the
invention is fast and easily automated, it enables the bulk screening of
samples that would
otherwise not be possible or practical.
All publications cited in this disclosure are incorporated herein by reference
in their
entirety.
Unless otherwise stated, the following terms used in this Application,
including the
specification and claims, have the definitions given below. It must be noted
that, as used in the
specification and the appended claims, the singular forms "a", "an," and "the"
include plural
referents unless the context clearly dictates otherwise.
The term "genotoxicity" as used herein refers to compounds that produce
chromosomal
aberrations, including breakage (clastogens) or abnormal copy number
(aneugens). In this
context, "genotoxicity" refers to a positive result in a micronucleus test. A
"likelihood of
genotoxicity" means specifically that the compound in question is predicted to
demonstrate
genotoxicity in an MNT with at least 75% confidence.
The term "test compound" refers to a substance which is to be tested for
genotoxicity. The
test compound can be a candidate drug or lead compound, a chemical
intermediate,
environmental pollutant, a mixture of compounds, and the like.
The term "kinase" refers to an enzyme capable of attaching and/or removing a
phosphate
group from a protein or molecule. "Inhibition of kinase activity" refers to
the ability of a
compound to reduce or interfere with such phosphatase activity. As binding
affinity of a small
molecule for a given kinase correlates well with the ability of said molecule
to inhibit the kinase
activity, binding affinity is considered synonymous with kinase activity
herein, and high binding
affinity is considered equivalent to high kinase inhibitory activity. The
correlation between
binding affinity and kinase inhibition is described by M.A. Fabian et al.,
Nature Biotechnol
(2005) 23:329-36, incorporated herein by reference in full.
The term "identified kinases" refers to the following set of kinases: CDK2
(Seq. Id. 1),
CLK1 (Seq. Id. 2), DYRKIB (Seq. Id. 3), ERK8 (Seq. Id. 4), GSK3A (Seq. Id. 5),
GSK3B (Seq.
Id. 6), PCTK1 (Seq. Id. 7), PCTK2 (Seq. Id. 8), STK16 (Seq. Id. 9), TTK (Seq.
Id. 10), CLK2
(Seq. Id. 11), ERK3 (Seq. Id. 12), and PRKR (Seq. Id. 13). "Alternate
identified kinases" refers
CA 02708311 2010-06-07
WO 2009/080219 PCT/EP2008/010530
-6-
to the set of kinases consisting of CDK2, CLK1, DYRKIB, ERK8 (MAPK15), GSK3A,
GSK3B,
PCTK1, PCTK2, STK16, TTK, CDK7, CLK4, and PCTK3. Preferred kinases are human
kinases
stated in the sequence listing. However, it is also possible to use kinases
from any other
organism in this method.
All patents and publications identified herein are incorporated herein by
reference in their
entirety.
Example
To identify the set of kinases that would indicate a likelihood that a test
compound would
demonstrate genotoxicity, the following analysis was carried out. First, 54
suitable small
molecule kinase inhibitors ("SMKIs") were selected to form a training set.
Second, for each
compound in the training set, an in vitro MNT result and single point
inhibition profiles against
290 kinases were acquired. A statistical analysis was then performed to (1)
build a model using
said single point kinase inhibition profiles to predict said MINT result and
(2) identify the kinases
correlated with MNT results. Finally, the model was validated against an
additional set of 33
SMKIs not used for training.
The in vitro micronucleus assay has been described in detail previously (M.
Fenech,
Mutation Res (2000) 455(1-2):81-95). The established permanent mouse lymphoma
cell line
L5178Y tk+i (ATCC CRL 9518), growing in suspension, was used for this
experiment. In
general, compounds were tested up to 500 .tg/mL, and at least 12 concentration
levels were
tested. The top dose for evaluation was generally selected to observe
acceptable toxicity
(decrease of the relative cell count (RCC) below 50%) or clear signs of
precipitation in the
aqueous medium. If the compound was soluble and non-toxic, a maximal dose
level of 5000
g/mL was set. For assessment of cytotoxicity, relative cell counts (RCC, as %
negative control)
were calculated. Slides were prepared by setting the cell density to
approximately 1 x 106
cells/mL and centrifuging onto clean glass slides using a cytospin (1000 rpm,
5 min). Fixation
of cells and storage was performed in ice cold methanol (-20 C, at least 4 h).
Slides were
incubated for 5 min with H 33258 (1 g/mL PBS/CMF) and mounted with 10 L
antifade for
fluorescence microscopy. A minimum of 3 concentration levels were analysed for
the presence
of micronucleated cells with the aid of an epifluorescence microscope equipped
with appropriate
filter sets. A compound is considered to possess clastogenic/aneugenic
activity if one or more
concentrations show at least a 2-fold increase in the number of micronucleated
cells in
comparison to the concurrent negative control.
CA 02708311 2010-06-07
WO 2009/080219 PCT/EP2008/010530
-7-
Fifty-four compounds were selected for inclusion in the training set, based on
a number of
criteria including selective kinase inhibition profiles, minimization of
redundancy, and chemical
diversity. From an internal database of SMKIs, only compounds that had
selective kinase
inhibition profiles were considered, where a selective compound was considered
to be one that
inhibited six or fewer kinases at single point inhibition values greater than
95%, and eleven or
fewer kinases at values greater than 85%. Kinase inhibition was determined
using the method set
forth in M.A. Fabian et al., Nature Biotechnol (2005) 23:329-36. In cases
where a number of
compounds were selective against many of the same kinases, only one of the
compounds was
selected, to minimize redundancy or over-representation of those kinases.
After these filtering
steps, a chemically diverse set was selected based on physical properties,
including ALogP,
molecular weight, number of hydrogen donors and acceptors, number of rotatable
bonds, number
of atoms, number of rings, number of aromatic rings, and number of fragments.
Diversity was
defined using the "Diverse Molecules" filter, based on a maximum dissimilarity
method, in
SciTegic's Pipeline Pilot 6Ø2.
Inhibition profiles against 290 kinases and in vitro MNT results were acquired
for each
compound in the training set (N=54). Three different readouts were obtained
for the MNT
results: negative (N=22), positive (N=26), and weakly positive (N=6). The six
weakly positive
were assigned to either negative or positive labels based on the % MN cells at
the concentration
at which the inhibition profiles were performed. This led to five of the six
compounds being re-
assigned as negative, giving a total of 27 negative and 27 positive compounds.
Pre-processing was first performed across the set of all inhibition profiles
to remove
uninformative or biased kinases. Kinases with no variance across the set of 54
compounds were
removed, as they were not informative. JNK and p38 isoforms were removed to
reduce the bias
of the large number of compounds in the training set that were developed to
target those kinases.
To ensure that the removal of JNK and p38 isoforms did not introduce a
different form of bias,
we performed an additional analysis whereby we considered only those training
set compounds
not developed for these kinase targets, and found that none of the JNK and p38
isoforms were
correlated with MNT results.
Feature selection (FS) and pattern recognition (PR) were performed in several
phases in
order to build the model. For all analyses, cross validation was used to
assess the model
performance over several trials. Each trial randomly split the initial data
into a training set and a
test set; the training set was used to build the temporary model, and the test
set was used to
predict results and then verify performance. Feature selection methods were
used to determine
CA 02708311 2010-06-07
WO 2009/080219 PCT/EP2008/010530
-8-
which kinases, or "features", were likely to correlate most with MNT result.
In each trial, the
inhibition values against the features chosen were used as input for a pattern
recognition method,
which then predicted the positive or negative result.
In the first phase, feature selection methods were divided into two groups:
methods that
could handle a large input data set (FS 1), and methods that performed better
with less data (FS2).
Different combinations of FS 1, FS2, and PR were tested over several trials
using 10 five-fold
cross-validations. The combination of methods with the lowest mean error rate
was chosen for
the next phase of the analysis. This combination includes a Kolmogorov-
Smirnov/T-test hybrid
algorithm for FS 1, Random Forests for FS2, and Support Vector Machines for PR
(T. Hastie et
al., "The Elements of Statistical Learning" (2001, Springer-Verlag); R.O. Duda
et al., "Pattern
Classification, 2nd Ed." (2000, Wiley-Interscience); and "Feature Extraction -
Foundations and
Applications" (2006, Springer-Verlag, I. Guyon et al. Eds.)).
The chosen combination of methods from the first phase were tuned for optimal
performance. Several parameters were optimized, including the number of
kinases to be used in
the model. The tuning process showed that within several trials, the mean
error rate was lowest
when the number of kinases chosen as significant after FS 1 and FS2 was 13.
Thus the model was
adjusted with the optimal parameters, then specified to choose the 13 most
significant features as
input for PR.
The accuracy of the model using this combination of feature selection and
pattern
recognition methods, number of features, and optimal tuning parameters was
then assessed by
performing 50 five-fold cross-validations. Importantly, the feature selection
and pattern
recognition was performed within each cross-validation fold. The resulting
model had an
accuracy of 80% 4%: that is, the model on average correctly predicted MNT
results 80% of the
time.
The 50 five-fold cross-validations were also used to determine the kinases
correlated with
MNT result. The selection of kinases was based on the number of times a kinase
was chosen as
significant amongst the 250 trials (50 five-fold cross-validations). 55 out of
the original 290
kinases were chosen at least once as significant. Those kinases that were
chosen with a
frequency of greater than 50% (N=13) were selected to be included in the final
model. Over
multiple runs of testing, the kinase inhibition profiles against these 13
kinases were found to be
significant in predicting actual MNT result at least 50% of the time. That is,
SMKIs with a
positive in vitro MNT result tended to have high levels of inhibition against
the thirteen kinases.
CA 02708311 2010-06-07
WO 2009/080219 PCT/EP2008/010530
-9-
For each SMKI, the model consists of single point kinase inhibition profiles
against the
following 13 kinases: CDK2, CLK1, DYRKIB, ERK8 (MAPK15), GSK3A, GSK3B, PCTK1,
PCTK2, STK16, TTK, CLK2, ERK3, and PRKR. Additionally, an in vitro MNT assay
result at
the concentration in which the kinase screen was performed is included. A
second model based
upon quantitative binding constants consisted a second (overlapping) set of
thirteen kinases:
CDK2, CLK1, DYRKIB, ERK8 (MAPK15), GSK3A, GSK3B, PCTK1, PCTK2, STK16, TTK,
CDK7, CLK4, and PCTK3. The kinases selected for the two models are highly
similar,
demonstrating the robustness of the single point kinase inhibition model.
To assess the utility of the final model, an additional set of 33 compounds
were used as a
validation set. These 33 compounds were not included in the initial set of 54,
but each compound
included a single point inhibition value against the thirteen model kinases,
plus an in vitro MNT
result. Given the validation data, the model was able to accurately predict
the MNT result of all
compounds, and thus performed with an accuracy of 76%, which lies within our
estimated
accuracy of the model based on cross-validation.
While the present invention has been described with reference to the specific
embodiments
thereof, it should be understood by those skilled in the art that various
changes may be made and
equivalents may be substituted without departing from the true spirit and
scope of the invention.
In addition, many modifications may be made to adapt a particular situation,
material,
composition of matter, process, process step or steps, to the objective spirit
and scope of the
present invention. All such modifications are intended to be within the scope
of the claims
appended hereto.
