Identification of the Clinical Development Candidate MRTX849, a Covalent KRASG12CInhibitor for the Treatment of Cancer

Article abstract of DOI:10.1021/acs.jmedchem.9b02052

Capping off an era marred by drug development failures and punctuated by waning interest and presumed intractability toward direct targeting of KRAS, new technologies and strategies are aiding in the target's resurgence. As previously reported, the tetrahydropyridopyrimidines were identified as irreversible covalent inhibitors of KRASG12C that bind in the switch-II pocket of KRAS and make a covalent bond to cysteine 12. Using structure-based drug design in conjunction with a focused in vitro absorption, distribution, metabolism and excretion screening approach, analogues were synthesized to increase the potency and reduce metabolic liabilities of this series. The discovery of the clinical development candidate MRTX849 as a potent, selective covalent inhibitor of KRASG12C is described.

Full text of DOI:10.1021/acs.jmedchem.9b02052

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
pubs.acs.org/jmc
Article
Identification of the Clinical Development Candidate MRTX849, a
Covalent KRASG12C Inhibitor for the Treatment of Cancer
Jay B. Fell,* John P. Fischer, Brian R. Baer, James F. Blake, Karyn Bouhana, David M. Briere,
Karin D. Brown, Laurence E. Burgess, Aaron C. Burns, Michael R. Burkard, Harrah Chiang,
Mark J. Chicarelli, Adam W. Cook, John J. Gaudino, Jill Hallin, Lauren Hanson, Dylan P. Hartley,
Erik J. Hicken, Gary P. Hingorani, Ronald J. Hinklin, Macedonio J. Mejia, Peter Olson, Jennifer N. Otten,
Susan P. Rhodes, Martha E. Rodriguez, Pavel Savechenkov, Darin J. Smith, Niranjan Sudhakar,
Francis X. Sullivan, Tony P. Tang, Guy P. Vigers, Lance Wollenberg, James G. Christensen,
and Matthew A. Marx*
Cite This: https://dx.doi.org/10.1021/acs.jmedchem.9b02052
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Capping off an era marred by drug development
failures and punctuated by waning interest and presumed
intractability toward direct targeting of KRAS, new technologies
and strategies are aiding in the target’s resurgence. As previously
reported, the tetrahydropyridopyrimidines were identified as
irreversible covalent inhibitors of KRASG12C that bind in the
switch-II pocket of KRAS and make a covalent bond to cysteine 12.
Using structure-based drug design in conjunction with a focused in
vitro absorption, distribution, metabolism and excretion screening
approach, analogues were synthesized to increase the potency and
reduce metabolic liabilities of this series. The discovery of the clinical development candidate MRTX849 as a potent, selective
covalent inhibitor of KRASG12C is described.
(∼14,000 new US cases annually) and 5% of colorectal
adenocarcinomas (∼5000 new US cases annually) and is
present in smaller fractions of other cancers. Collectively,
KRASG12C mutations comprise a patient population with a
worldwide annual incidence of greater than 100,000 individ-
uals. A novel scientific basis for targeting KRASG12C was
recently described in a breakthrough article by Shokat and
Wells, in which they identified a previously unknown allosteric
pocket near the nucleotide binding site.9 Small molecules
binding to this pocket can inhibit signaling through KRAS by
locking the protein in its inactive GDP-bound state.10−13
Although targeting this site with covalent, small-molecule
inhibitors was a clear advancement, it was also evident that the
molecules identified required significant optimization to
achieve drug-like potency and pharmaceutical properties.
As previously reported, compound 1 (Figure 1) is an
irreversible covalent inhibitor of KRASG12C that binds in the
INTRODUCTION
■
KRAS is the single most frequently mutated oncogene and the
first of more than 700 genes to be causally implicated in human
cancer (COSMIC).1 Mutations in KRAS are prevalent
amongst the top three most deadly cancer types in the United
States: pancreatic (95%), colorectal (45%), and lung (35%).2
Its frequent mutation across a spectrum of aggressive cancers
has stimulated an intensive drug discovery effort to develop
therapeutic strategies that block KRAS function for cancer
treatment. Despite nearly four decades of research, a clinically
viable KRAS cancer therapy has remained elusive largely
because of its high affinity for GTP and lack of a well-defined
binding pocket.3−6 However, recent findings have stimulated a
new wave of activities to develop KRAS-targeted therapies.
Capping off an era marred by drug development failures and
punctuated by waning interest and presumed intractability
toward direct targeting of KRAS, new technologies and
strategies are aiding in the target’s resurgence.7,8 Central to
this renewal is a single mutation: KRASG12C, a well-validated
driver mutation and the most frequent individual KRAS
mutation in lung cancer.4 Associated with poor prognosis and
resistance to treatment, KRASG12C represents both an
extraordinary unmet clinical need and opportunity. This
mutation has a causal role in 14% of lung adenocarcinomas
Received: December 10, 2019
https://dx.doi.org/10.1021/acs.jmedchem.9b02052
© XXXX American Chemical Society
J. Med. Chem. XXXX, XXX, XXX−XXX
A

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
O-sulfate (M3). Glutathione (GSH) conjugation of the
acrylamide also appeared to play an important role, 50% of
the metabolites (M1, M2, M4, and M5) contained this
modification.
Given the metabolite profile, it was predicted that removal of
the hydroxyl moiety from compound 1 may decrease overall
CL by reducing conjugative metabolism. As previously
described,14 the naphthol ring of this class of inhibitors
binds in a well-defined hydrophobic pocket formed by Val9,
Met72, Phe78, Tyr96, Ile100, and Val103. In addition, the
naphthyl hydroxyl forms a hydrogen bond to the carboxylate of
Asp69. Table 2 describes a series of substituted indazoles that
Figure 1. Structural comparison of compound 1 to clinical
development candidate MRTX849.
switch-II pocket of KRAS and makes a covalent bond to
cysteine 12.14 Inhibitor 1 showed robust target engagement in
a PK/PD experiment using MIA PaCa-2 tumor bearing mice
when dosed IP. In a related tumor growth inhibition (TGI)
experiment, compound 1 demonstrated rapid regressions with
cures when dosed as low as 30 mg/kg IP once a day. While
providing a valuable tool compound, inhibitor 1 demonstrated
clearance (CL) of 46 mL/min/kg following a 3 mg/kg IV dose
to CD-1 mice and oral bioavailability of 2.4% with a 100 mg/
kg PO dose. The pharmacokinetic limitations of 1 necessitated
several iterations of optimization to identify a tractable clinical
candidate. Described in this paper are the insights that led to
the discovery of the clinical candidate MRTX849.
Table 2. Exploration of Substituted Indazoles as Hydroxy
Naphthyl Replacements
a
RESULTS AND DISCUSSION
■
In order to advance the tetrahydropyridopyrimidine series of
covalent KRASG12C inhibitors, it was necessary to develop an
understanding of the metabolic liabilities of 1 that contributed
to high CL and low bioavailability in CD-1 mice. Compound 1
was incubated with mouse hepatocytes, and the metabolites
were identified by liquid chromatography mass spectrometry
(LCMS) with relative concentrations being estimated based on
peak area in the corresponding A290nm chromatogram (Table
1). The metabolism of naphthol is well-documented,15 and it
was not surprising that the hydroxyl moiety was the major
metabolic liability for compound 1 with 76% of the identified
metabolites being O-glucuronides (M1, M2, M6, and M7) or
a
NCI-H358 cellular pERK inhibition at 3 h.
Table 1. Metabolites of Compound 1 Identified in
Incubations with Mouse Hepatocytes
were designed as naphthol replacements that fill the hydro-
phobic pocket while maintaining a hydrogen bonding
interaction to Asp69. The activity of these compounds,
reported as IC50 (nM), was determined in NCI-H358 cells
by measuring inhibition of phospho-ERK after a 3 h incubation
with the inhibitor. Because the intracellular drug disposition
was unknown, the liabilities of conducting SAR studies using
cellular data were recognized. The 5-methylindazole 2
displayed 1/10th of the cellular potency of compound 1; this
loss in potency was attributed to inadequate interactions with
the protein. The 5-ethylindazole 3 and the 5-isopropylindazole
4 are equipotent to analogue 2 within variance of the assay.
Through improved contact with the protein, both the 5-
trifluoromethylindazole 5 and the 5,6-dimethylindazole 6
showed cellular IC50’s approximately five fold more potent
than compound 2. As the most active of the naphthol
replacements, the substituted indazoles 5 and 6 were profiled
in a series of metabolic stability and permeability assays. While
the human hepatocyte and microsome stability data were
promising, with a hepatic extraction ratio (ER) of <55%, these
analogues exhibited low permeability and were potent P-gp
metabolite
M1
M2
biotransformation(s)
% peak area
dealkylation; glucuronide; GSH
glucuronide; GSH
sulfate
8
18
9
M3
M4
M5
dealkylation; GSH
GSH
6
14
14
27
4
M6
M7
dealkylation; glucuronide
glucuronide
Compound 1
B
https://dx.doi.org/10.1021/acs.jmedchem.9b02052
J. Med. Chem. XXXX, XXX, XXX−XXX

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
efflux substrates with permeability efflux (PE) ratios of 75 and
88, respectively. These results prompted the exploration of an
alternate strategy for progressing the project.
Because the indazole analogues designed as naphthol
replacements did not achieve acceptable cellular permeability,
the hydroxyl moiety was targeted for removal in an effort to
improve both permeability as well as metabolic stability. The
deshydroxy analogue 7 was synthesized (Figure 2), and the PK
Figure 3. X-ray crystal structure of compound 7 bound to KRASG12C
with 2.27 Å resolution, hydrogens added for clarity (PDB code
6USX).
Table 4. Substituted Piperazines Designed To Displace
a
KRASG12C Gly10-Bound Water
Figure 2. Compound 1 and the deshydroxy analogue 7.
was evaluated in CD-1 mice. Following a 3 mg/kg IV dose of
compound 7 to CD-1 mice (n = 3), CL was 25.6 mL/min/kg,
the t1/2 was 0.908 h, and the volume of distribution (VSS) was
0.996 L/kg (Table 3). Oral administration of a 10 mg/kg dose
Table 3. PK of Compound 7 Dosed in CD-1 Mice
dose
parameter
value
3 mg/kg IV
CL (mL/min/kg)
t1/2 (h)
VSS (L/kg)
Cmax (μg/mL)
Tmax (h)
25.6
0.908
0.996
0.131
0.500
0.302
13.9
10 mg/kg PO
AUCinf (h·μg/mL)
F (%)
of 7 to mice (n = 3) resulted in a Cmax of 0.131 μg/mL and
AUCinf of 0.302 h·μg/mL, with 13.9% bioavailability. The
lower CL and higher bioavailability of 7 relative to 1, albeit at a
different dose, validated the decision to remove the hydroxyl
metabolic liability. The weak cellular activity of compound 7
toward KRASG12C, IC50 = 4400 73 nM compared to 142
12 nM for compound 1, prompted a search for additional
interactions with the protein to move the series into a
progressable activity range.
a
NCI-H358 cellular pERK inhibition at 3 h.
Figure 3 illustrates the X-ray crystal structure of compound 7
bound to KRASG12C. Notable in this structure is a bound water
molecule complexed to Gly10 and Thr58 which is near the
piperazine ring of compound 7. This water molecule forms
hydrogen bonds with the side chain hydroxyl of Thr58 (ca. 2.7
Å) and the carbonyl of Gly10 (3.0 Å), highlighted in Figure 3.
The water is also ca. 3.0 Å from the backbone NH of Gly10.
Analysis of the proximal hydrogen bonding network suggested
that displacement of this water could lead to a large potency
increase.16 An appropriately substituted piperazine ring would
have the correct trajectory to displace this bound water.
Table 4 describes a series of analogues that was synthesized
to probe the space occupied by the Gly10-bound water.
Substituents were placed at either R1 or R2 while keeping the
other positions on the piperazine unsubstituted. Compounds
8−11 are a set of R1- and R2-substituted hydroxy methyl and
hydroxy ethyl analogues that were designed to displace the
bound water and replace the hydrogen bonding interactions in
the pocket. These compounds were equal or less potent than
the unsubstituted compound 7 in the NCI-H358 cell assay.
Analogue 12, substituted at the R1 position with CH2CN,
showed a 300-fold increase in potency with a cellular IC50 = 14
nM. The diastereomers of 12 were synthesized, and the S-
diastereomer 12a was determined to be ∼100× more potent
than the R-diastereomer 12b. In addition to binding, another
factor that could contribute to increased potency of a covalent
inhibitor is electrophile reactivity. For example, compound 7
was incubated with GSH at 37 °C and the percent parent
remaining at 1 and 4 h was measured. From these data, the
half-life (t1/2), a relative measure of electrophile reactivity, was
extrapolated to be 17 h. Compound 12a, on the other hand,
had a GSH t1/2 of 4 h, indicating that the electron withdrawing
cyanomethyl substituent had the effect of increasing the
C
https://dx.doi.org/10.1021/acs.jmedchem.9b02052
J. Med. Chem. XXXX, XXX, XXX−XXX

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
a
reactivity of the acrylamide toward GSH. Interestingly, the
opposite enantiomer 12b also showed increased reactivity,
GSH t1/2 = 1 h. However, unlike 12a which was 400× more
potent than 7, 12b with a cell IC50 = 1020 nM was only 4 fold
more potent than compound 7. No increase in potency was
observed with the cyanomethyl regioisomer 13.
Table 5. Exploration of 8-Naphthyl Position
In addition to the potent cell activity, compound 12a
displayed favorable in vitro absorption, distribution, metabo-
lism, and excretion (ADME) properties. This compound had
95% plasma protein binding such that the free fraction
adjusted cell IC50 was calculated to be 200 nM. Compound
12a had measured solubility of 800 μg/mL in FaSSIF at pH
6.5. CL in microsome preparations was modest, with an ER of
45% and high in hepatocytes, with an ER of 90%. This
compound displayed medium permeability with efflux (Papp (A
to B) = 2.2 × 10−6 cm/s, PE = 28) in the MDR LLCPK assay.
Taken together, these data suggested that compound 12a
represented an interesting lead compound that warranted
further investigation. Optimization of compound 12a
proceeded with substitution of the naphthalene moiety. Figure
4 shows a surface view of 12a bound to KRASG12C that
compound
R
cellular IC50 (nM)
14
15
16
17
18
19
methoxy
cyano
22
64
6
7
7
1
ethyl
trifluoromethyl
chloro
methyl
4
1
1
1
0.1
0.3
a
NCI-H358 cellular pERK inhibition at 3 h.
Table 6. PK of Compounds 18 and 19 Dosed in CD-1 Mice
dose
parameter
18
19
3 mg/kg IV
CL (mL/min/kg)
t1/2 (h)
VSS (L/kg)
Cmax (μg/mL)
Tmax (h)
37.3
0.496
0.867
2.46
0.250
2.27
31.6
1.33
1.91
1.66
0.500
4.91
31.1
30 mg/kg PO
AUCinf (h·μg/mL)
F (%)
16.9
PK/PD studies were conducted to evaluate the pharmaco-
dynamic response and to correlate drug exposure with target
inhibition (Figure 5). Compounds 18 and 19 were
Figure 4. X-ray crystal structure of compound 12a bound to
KRASG12C with 2.03 Å resolution, surface pose (PDB code 6USZ).
revealed a small pocket in the protein oriented from the 8-
position of the naphthyl, highlighted with a blue arrow. A
number of small hydrophobic groups were designed to fill this
narrow hydrophobic cleft formed by residues Val9, Thr58,
Met72, and Tyr96.
Following the synthetic route described in the Chemistry
section, a series of analogues was synthesized to probe the
vector from the naphthyl 8-position. The 8-substituted
naphthyl compounds and their corresponding NCI-H358
cellular IC50 values are shown in Table 5. The methoxy and
cyano analogues 14 and 15 were less potent compared to the
proto analogue 12a. 8-Ethyl naphthyl 16 and 8-trifluoromethyl
naphthyl 17 displayed similar potency to the parent with single
digit cellular IC50 values. Finally, both the chloro and methyl
substituted compounds 18 and 19 were more potent with
cellular IC50’s of 1 nM.17
Pharmacokinetics of the chloro and the methyl substituted
compounds 18 and 19 were evaluated in CD-1 mice (Table 6).
Oral administration of a 30 mg/kg dose of compound 18 to
mice (n = 3) resulted in a Cmax of 2.46 μg/mL and AUCinf of
2.27 h·μg/mL, with 16.9% bioavailability. Oral administration
of a 30 mg/kg dose of compound 19 to mice (n = 3) resulted
in a Cmax of 1.66 μg/mL and AUCinf of 4.91 h·μg/mL, with
31.1% bioavailability. Given the cellular potency and favorable
oral exposure observed in mice, compounds 18 and 19 were
progressed into efficacy studies.
Figure 5. PK/PD of Compounds 18 and 19.
administered via oral gavage over a range of dose levels to
NCI-H358 xenograft-bearing mice. Tumors and plasma were
collected 6 h after a single dose. Compound 18 showed 70%
KRASG12C modification at 10 mg/kg, 89% protein modification
at 30 mg/kg, and >90% modification at 100 mg/kg as
determined by LCMS analysis of the tumor lysate. In
comparison, compound 19 exhibited 20, 60, and 80%
modification at the same doses. Given the similarities between
these two compounds, there was no clear explanation for the
differences seen with in vivo target engagement. However, this
study demonstrated a dose-dependent increase in covalent
modification of KRASG12C that was proportional to the plasma
concentration of the drug for the individual compounds.17
D
https://dx.doi.org/10.1021/acs.jmedchem.9b02052
J. Med. Chem. XXXX, XXX, XXX−XXX