2-Formylpyridyl Ureas as Highly Selective Reversible-Covalent Inhibitors of Fibroblast Growth Factor Receptor 4

Article abstract of DOI:10.1021/acsmedchemlett.7b00485

As part of a project to identify FGFR4 selective inhibitors, scaffold morphing of a 2-formylquinoline amide hit identified series of 2-formylpyridine ureas (2-FPUs) with improved potency and physicochemical properties. In particular, tetrahydronaphthyridi

Full text of DOI:10.1021/acsmedchemlett.7b00485

Subscriber access provided by READING UNIV
Letter
2-Formylpyridyl-ureas as highly-selective reversible-
covalent inhibitors of fibroblast growth factor receptor 4
Thomas Knoepfel, Pascal Furet, Robert Mah, Nicole Buschmann, Catherine Leblanc, Sebastien
Ripoche, Diana Graus-Porta, Markus Wartmann, Inga Galuba, and Robin A. Fairhurst
ACS Med. Chem. Lett., Just Accepted Manuscript • DOI: 10.1021/acsmedchemlett.7b00485 • Publication Date (Web): 01 Feb 2018
Downloaded from http://pubs.acs.org on February 2, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Medicinal Chemistry Letters is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 8
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
9
2-Formylpyridyl-ureas as highly-selective reversible-covalent inhibi-
tors of fibroblast growth factor receptor 4
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Thomas Knoepfel, Pascal Furet, Robert Mah, Nicole Buschmann, Catherine Leblanc, Sebastien
Ripoche, Diana GrausꢀPorta, Markus Wartmann, Inga Galuba, Robin A. Fairhurst*
Novartis Institutes for BioMedical Research, CHꢀ4002 Basel, Switzerland.
ABSTRACT: As part of a project to identify FGFR4 selective inhibitors, scaffold morphing of a 2ꢀformylquinoline amide hit idenꢀ
tified series of 2ꢀformylpyridine ureas (2ꢀFPUs) with improved potency and physicochemical properties. In particular, tetrahyꢀ
dronaphthyridine urea analogues with cellular activities below 30 nM have been identified. Consistent with the hypothesised reꢀ
versibleꢀcovalent mechanism of inhibition the 2ꢀFPUs exhibited slow binding kinetics and the aldehyde, as the putative electroꢀ
phile, could be demonstrated to be a key structural element for activity.
KEYWORDS: 2-formylpyridines, fibroblast growth factor receptor 4, kinase inhibitors, reversible-covalent binders, binding kinet-
ics, hepatocellular carcinoma.
Fibroblast growth factor 4 (FGFR4) is one of a family of four
receptor tyrosine kinases which are regulated by their interacꢀ
tions with a much larger number of fibroblast growth factors
(FGFs). FGFR signaling is involved in the control of multiple
developmental and physiological processes, as well as being
implicated as a driver in certain forms of cancer.1 A number of
panꢀFGFR inhibitors are currently undergoing clinical investiꢀ
gation for the treatment of a range of FGFR 1ꢀ3 driven canꢀ
cers.2 More recently FGF19 signaling through FGFR4, which
also requires the presence of the coꢀreceptor βꢀklotho, has
been highlighted as a driver in a sub set of solid tumors, in
particular hepatocellular carcinoma (HCC).3 As a result a
number of groups have sought to identify selective inhibitors
of FGFR4 signaling as a way to treat these cancers whilst
avoiding the onꢀtarget sideꢀeffects associated with FGFR1ꢀ3
inhibition,4 and in particular FGF23ꢀmediated hyperphosphaꢀ
temia.5 Selective agents have the potential to not only be better
tolerated than panꢀFGFR inhibitors, but also to enable the imꢀ
pact of more robust inhibition of FGFR4 to be investigated.
To explore this possibility, we, and others, have identified a
poorly conserved cysteine residue within the middleꢀhinge
region of the FGFR4 ATP binding pocket that can be targeted
to gain selectivity.4 Cys552 is situated two positions beyond
the gateꢀkeeper (GK+2) residue in FGFR4 and presents an
opportunity for selectivity based upon the relatively small size
of the side chain, or as a nucleophile for covalent binding. The
other FGFR family members possess a tyrosine at the GK+2
position, and within the human kinome only twelve other kiꢀ
nases have an equivalent sized, or smaller residue, at this posiꢀ
tion, of which only four contain a cysteine.6 Figure 1 shows
the structures of selective FGFR4 inhibitors that have been
described to date. Compounds 1 and 2 achieve FGFR4 selecꢀ
tivity by exploiting the size difference in the middleꢀhinge
Figure 1. Selective inhibitors of the kinase activity of FGFR4.
region, and the acrylamides H3Bꢀ6527, BLU9931, BLU554, 3
and 4 by making a covalent bond with Cys552.4 In this manuꢀ
script we report the first part of our work which ultimately led to
the identification of FGF401, a highlyꢀselective inhibitor of
FGFR4 which is currently undergoing clinical evaluation for
the treatment of FGFR4 and βꢀklotho positive solid tumors.7ꢀ9
ACS Paragon Plus Environment

ACS Medicinal Chemistry Letters
conformational restraint for FGFR4 activity.4 The introduction
Page 2 of 8
of sp3ꢀhybridised atoms into the quinoline moiety was exꢀ
plored as a way to reduce the molecules coꢀplanar aromatic
nature, and as a result the efficiency of intermolecular interacꢀ
tions within the solid state.11 Morphing the quinolone amide
moiety to the corresponding tetrahydronaphthyridine urea
(THNU) generated the analogue 6a. The melting point for 6a
(133 °C) was found to be lowered by 42 °C when compared to
5 (175 °C), consistent with the above hypothesis.12 Although
high throughput equilibrium solubility (HTES) remained outꢀ
side the measured range (< 4 ꢁM),12 more precise measureꢀ
ments indicated the thermodynamic solubility of 6a to be imꢀ
proved relative to 5 (aqueous pH 6.8 solubilities: 5 < 2.9 ꢁM
and 6a 4.9 ꢁM). More importantly, consistent results could be
obtained with 6a in inꢀvitro assays, with no evidence for arteꢀ
facts. In addition, a 4ꢀfold increase in biochemical potency
was observed with the THNU modification, Table 1, further
supporting a scaffold morphing strategy in this direction.
1
2
3
4
5
6
7
8
CF3
CF3
N
N
N
N
N
H
H
Scaffold
O
H
N
O
H
O
N
O
Morphing
5
6a
Figure 2. Scaffold morphing to the 2ꢀformyl THNU series 6.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We have previously described aspects of our FGFR4 inhibitor
hitꢀfinding activities which included the identification of the 2ꢀ
formylquinoline amide series, exemplified by 5, Figure 2, as a
highly selective starting point.4 Before committing to an exꢀ
pansive optimisation effort with this series, a number of key
questions were raised. Of these a greater understanding of the
mechanism of FGFR4 inhibition was considered important to
guide the optimisation. In particular exploring the role of the
aldehyde, as the putative electrophile, reacting in a reversibleꢀ
covalent manner through the formation of a hemithioacetal
with Cys552 of FGFR4.10 However, the low solubility of 5
made in vitro testing problematic, and hindered answering the
above question. Therefore, identifying opportunities to imꢀ
prove physical properties, whilst also being compatible with
improving the biological profile, became the first priority.
The THNU analogue 6a was anticipated to bind within the
ATPꢀpocket of FGFR4 in an analogous manner to that hyꢀ
pothesized previously for 5.4 Reaction between the Cys552
thiol and the aldehyde of 6a leads to the formation of the heꢀ
mithioacetal adduct, shown in Figure 3, which is supported by
the aminopyridyl moiety hydrogen bonding to the hinge resiꢀ
dues Glu551 and Ala553. Based upon this model, the opporꢀ
tunity to further increase potency, and to regulate physical
properties, was anticipated to be possible through the modifiꢀ
cation of the hinge binding element within 6a.
The poor solubility of 5 was apparent across a broad range of
solvents, including aqueous media. As a result in vitro pharꢀ
macology at concentrations of > 1 ꢁM proved difficult to inꢀ
terpret with a high level of confidence. To overcome this limiꢀ
tation, one strategy was to reduce the coꢀplanar aromatic naꢀ
ture of 5, which is reinforced by the intramolecular Hꢀbond
between amide NꢀH and quinoline nitrogen. However, the
intramolecular Hꢀbond was also believed to be an important
As a first step towards exploring the hingeꢀbinding SAR surꢀ
rounding 6a, a number of alternative heterocyclic groups were
evaluated, Figure 4. Of these, only the 2ꢀpyrimidyl and 2ꢀ
pyrazinyl derivatives 7a and 7b, and the extended 2ꢀ
pyrazoylderivative 7f, led to activities approaching the 2ꢀ
pyridyl analogues. For 7f the 3,5ꢀdimethoxyphenyl moiety is
anticipated to make the key backꢀpocket interactions described
previously for several FGFR inhibitors.13 The 2ꢀpyridyl hingeꢀ
binder was taken for further exploring the SAR, in particular
through substitution at the 4ꢀ and 5ꢀpyridyl positions based
upon the modeled interaction shown in Figure 3. In the 5ꢀ
pyridyl position (R1, Figure 5) trifluoromethyl 6a, cyano 6b
and chloro 6c all increased the potency to a similar extent
when compared to the unsubstituted pyridine ring 6d. The
cyano group in the 5ꢀposition was of particular interest based
upon the lower lipophilicity (logD7.4 6a 4.1; 6b 2.9; 6c 3.7).12
In the 4ꢀpyridyl position (R2, Figure 5) a range of Cꢀ, Nꢀ and
Oꢀsubstituents 6eꢀk, were designed to fill a small hydrophobic
cleft, formed between Gly474 and Val481, and to make van
der Waals contacts with Leu619.12 Encouragingly these modiꢀ
fications were found to further increase potency. In particular,
the larger 5ꢀpyridyl ether analogues 6f and 6g, and the 5ꢀ
pyridyl Nꢀ(4ꢀhydroxyꢀ4ꢀmethyl)ꢀpiperidinyl analogue 6k reꢀ
sulted in FGFR4 biochemical activities below 2 nM. These
interactions are highlighted for 6k in Figure 6.
Figure 3. Model of 6a bound within the ATP binding site of
FGFR4. The compound forms hydrogen bonds with residues
Ala553 and Val500, a pseudo hydrogen bond with Glu551, in
addition to the intramolecular hydrogen bond. Hydrogen bonds
are indicated by the dotted violet lines.
CN
H
N
N
H
Heterocycle
O
OMe
N
N
N
N
N
OMe
O
HN
N
CN
N
N
H
N
g
c
e
a
N
N
N
N
N
N
OMe
7
b
d
f
Figure 4. Hingeꢀbinding variants 7aꢀg prepared in the 2ꢀformyl THNU series.
ACS Paragon Plus Environment

Page 3 of 8
ACS Medicinal Chemistry Letters
Table 1. Biochemical and cellular FGFR4 data.
R1
1
2
3
4
5
6
7
8
N
HN
N
R2
IC50 (nM)
Cellular Ba/F3
H
O
Compound
Biochemical
FGFR4
N
O
pFGFR4
162 ± 52
ꢀ
2.2 ± 0.9
14 ± 10
57 ± 10
13 ± 9
16 ± 5
21
1
BLU9931
5
Compound
R1
CF3
CN
Cl
R2
2755 ± 177
ꢀ
H
6a
6b
6c
6d
6e
6f
6a
9
H
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
108 ± 33
916 ± 657
ꢀ
6b
H
H
6c
H
300
6d
CN
CN
CN
CN
OMe
OiPr
6.3
72 ± 16
29
6e
0.7
6f
OCH2CH2OMe
OCH2CH2NMe2
6g
6h
1.3 ± 0.5
11 ± 5
17
18 ± 8
88 ± 8
150 ± 112
224
6g
6h
CN
CN
CN
6i
6j
6i
27
6j
1.2
24
6k
45 ± 8
33 ± 4
5200
131 ± 24
ꢀ
7a
6k
7b
Figure 5. Structures of the THNU analogues 6aꢀk.
ꢀ
7c
9500
ꢀ
7d
The introduction of a basic amine in 6h led to an improvement
in aqueous solubility (HTES 300 ꢁM), albeit with a slight reꢀ
duction in potency. In addition, high permeabilities were deꢀ
termined for all the compounds, and were consistent with the
potential for high levels of absorption from the gut following
oral administration.12 Taken together, these data supported the
hypothesized binding mode and indicated that the potency and
physicochemical properties of the series could be modulated in
a favorable way through modification of the hingeꢀbinding
moiety.
790
> 3000
2800 ± 283
> 3000
ꢀ
7e
23 ± 17
140
7f
7g
> 10000
42 ± 14
3.2 ± 1.1
1.1 ± 0.4
660 ± 57
175 ± 21
6.5 ± 2.5
7.4 ± 2.0
> 10000
4400
8
531 ± 57
750 ± 136
55 ± 3
> 3000
741 ± 77
41 ± 35
91 ± 23
ꢀ
9
10a
10b
11a
11b
12a
12b
13
Based upon the improvements derived from switching the
quinoline amide to the THNU scaffold, further 2ꢀformylꢀ
pyridine urea (2ꢀFPU) variants were prepared to explore the
SAR of the coreꢀstructure linking the hinge binder to the
ꢀ
14
data are from single experiments, or expressed as mean ± SD,
where 2 to 5 repeated experiments were performed.
aldehyde electrophile, Figure 7. In each case the new scaffold
morphs were designed to preserve the intramolecular hydroꢀ
genꢀbond, which was considered to be essential for the optimal
alignment of the two key structural elements. Support for this
was derived from the FGFR4ꢀinactive Nꢀmethylated analogue
8, Figure 8, in which the intramolecular hydrogenꢀbond is
absent, and the required electrophile/hingeꢀbinding alignment
is highly disfavoured.12 These data also indicate that the aldeꢀ
hyde functionality alone is not sufficient for strong FGFR4
inhibitory activity.
Introduction of oxygen into the saturated THN ring gave the
analogue 9 which showed a 3ꢀfold decrease in FGFR4 activity.
Reducing the size of the saturated THN to the 5ꢀmembered
Figure 6. Model of 6k bound within the FGFR4 ATP pocket
highlighting the hydrophobic interactions of the R2 group.
ACS Paragon Plus Environment

ACS Medicinal Chemistry Letters
Page 4 of 8
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. 2ꢀFPU analogues 9ꢀ12 prepared as scaffold morphs of the quinolineꢀamide hit 5.
analogues retained a similar level of activity for both the aroꢀ
matic 10a and dihydro 10b analogues. However, although
isolatable, the pyrrolopyridine urea 10a was very susceptible
to hydrolysis and was not considered to be a viable option for
further evaluation. Increasing the ring size to the 7ꢀmembered
analogues 11a and 11b led to a > 50ꢀfold decrease in potency.
This decrease in activity was rationalised by modeling to arise
from an unfavorable interaction with a hydrophobic slot,
formed between residues Leu473 and Gly556, at the solvent
exposed side of the FGFR4 ATP binding site.12 This slot was
considered too narrow to readily accommodate the larger ring
size.
fusion with the translocation, ets, leukemia transcription factor
(TEL). Inhibition of FGFR4 autophosphorylation was measꢀ
ured following a 1 hour incubation with the test compounds,
Table 1.12 For the majority of the 2ꢀFPU analogues a conꢀ
sistent 5ꢀ to 50–fold IC50 shift from the biochemical to cellular
assay was observed, which is slightly lower than the 74–fold
shift observed for the nonꢀcovalent FGFR4ꢀselective reference
compound 1. Exceptions being the pyrrolopyridine urea 10a
and extended 2ꢀpyrazoyl derivative 7f where greater than 100–
fold shifts were observed. For 10a this is likely as a result of
the chemical instability associated with the urea linkage deꢀ
scribed above. Overall, the Ba/F3 cellular assay data showed
the potential of the THNU series, with three analogues disꢀ
playing IC50 values below 30 nM.
Due to the presence of the intramolecular hydrogenꢀbond in
the above compounds they can be considered to be pseudo
tricyclic structures. Therefore, the opening up of the saturated
ring of the THN was not anticipated to impact upon the overall
alignment of the key structural elements. Consistent with this
analysis the ring opened analogues 12a and 12b retained a
similar level of potency to the THN starting point 6. Additionꢀ
ally, modeling highlighted that a range of urea Nꢀsubstituents
would have the potential to make favorable interactions.12 In
particular, within the hydrophobic slot formed between
Leu473 and Gly556, described above. Not only providing the
opportunity to improve the FGFR4 activity, but also to further
regulate physicochemical properties (HTES 12b > 1 mM).
With potent, soluble and highlyꢀselective compounds identiꢀ
fied, we next sought to investigate further the mechanism of
FGFR4 inhibition. One consequence of the proposed reversiꢀ
bleꢀcovalent interaction was anticipated to be extended bindꢀ
ing kinetics.14 To gain an understanding of the inhibitor /
FGFR4 interaction kinetics, competition binding experiments
were conducted with three of the 2ꢀFPU analogues from Table
1.12 All three 2ꢀFPU analogues exhibited classical monoꢀ
exponential binding, from which kinetic parameters could be
derived, Table 2. In contrast, the nonꢀcovalent reference comꢀ
pound 1 dissociated from the FGFR4 kinase domain faster
than the labelled probe used in the assay, and as a result no
kinetic parameters could be assigned. The 2ꢀFPU analogues
exhibited slowꢀbinding kinetics, with 1.5 – 4.5 hour residence
times, that are at least 75x longer than for 1, whilst falling
within a similar affinity range. Interestingly, inhibitors posꢀ
sessing longer target residence times have been associated
with enhanced selectivity and higher levels of in vivo activiꢀ
ty.15
Assessing FGFR family selectivity, 2ꢀFPUs 5ꢀ12 all exhibited
excellent selectivity versus FGFR2, and where measured verꢀ
sus FGFR1 and FGFR3.12 Additionally, selectivity against a
diverse panel of 52 kinases was excellent, the majority of kiꢀ
nases tested exhibited IC50 values > 10 ꢁM, with no IC50 valꢀ
ues < 1 ꢁM.12 As described previously for 5, biochemical acꢀ
tivity was completely lost using a variant of the FGFR4 kinase
domain in which Cys552 was exchanged for alanine, indicatꢀ
ing the reversibleꢀcovalent interaction remains valid for the 2ꢀ
FPU compounds.4,12
Having determined slow binding kinetics with the 2ꢀFPUs, the
biochemical assay was evaluated to understand which condiꢀ
tions would reflect the maximum equilibrium IC50 values. The
approach taken was analogous to that which has been used to
Cellular activity was assessed using a Ba/F3 cell model, in
which the FGFR4 kinase domain is constitutively activated by
characterise
a number of slowꢀbinding typeꢀII kinaseꢀ
inhibitors.16 Starting with a nonꢀphosphorylated FGFR4 kinase
domain construct, the inhibitor preincubation times were seꢀ
quentially increased from 0 to 4 hours prior to initiating the
phosphorylation reaction (1 hour incubation).12 No change up
to a modest 5ꢀfold increase in potency was observed when the
preincubation period was increased to 1 hour. However, no
further increases were noted with longer preincubations for the
majority of compounds.
Figure 8. Structures of the Nꢀmethyl urea 8, benzylic alcohol
13 and desformyl analogue 14.
The biochemical data shown in Table 1 were generated using
ACS Paragon Plus Environment

Page 5 of 8
ACS Medicinal Chemistry Letters
1
2
Table 2. Binding kinetic data with the FGFR4 kinase domain.
3
4
5
6
7
8
9
Compound
Kd (nM)
20
kon (M-1min-1)
ꢀ
koff (min-1)
ꢀ
3.7 x 10ꢀ3
4.3 x 10ꢀ3
9.6 x 10ꢀ3
Residence time (min)
< 1.4
272
235
105
1*
6a
6b
9
28
2.2 x 103
1.4 x 104
4.5 x 103
5.0
35
* binding kinetics are too fast to measure, and the residence time is below the measurable lower limit for the assay of 1.4 minutes.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
the ‘standard’ screening conditions with a longer FGFR4 kiꢀ
naseꢀdomain construct of mixed phosphorylation states, and
were derived without preincubation (1 hour phosphorylation
reaction).4 Assuming the same temporal relationship to the
data in Table 3: the biochemical data with the ‘standard’ assay
in Table 1 would then also be anticipated to be within 5ꢀfold
of the maximum equilibrium IC50 values, for the majority of
the 2ꢀFPUs. Therefore, with the possibility for only a small
IC50 shift, we positioned the ‘standard’ biochemical assay as
the primary measure of FGFR4 activity going forwards, and
planned only to spot check the binding kinetics for key comꢀ
pounds to monitor for any significant change in profile.
situation to a number of acrylamideꢀbased irreversibleꢀ
covalent kinase inhibitors, where the electrophile has typically
been introduced into an already low nM hingeꢀbinding scafꢀ
fold.17
When preparing the 2ꢀFPU derivatives a number of commonly
applied approaches for forming the central ureaꢀlinkage were
found to be inefficient, which was attributed to the low nucleꢀ
ophilicity of the 2ꢀaminopyridines. Two methods that were
discovered to reliably provide access to a wide range of bisꢀ
aminopyridyl ureas in good yield are exemplified by the synꢀ
theses of 6b and 11a, in Schemes 1 and 2.
Preparation of 6b started with the acylation of Nꢀlithiated 2ꢀ
(dimethoxymethyl)ꢀtetrahydroꢀnaphthyridine 15 with diphenyl
carbonate, to prepare the corresponding phenyl carbamate
intermediate 16. Further reaction of 16 with Nꢀlithiated 2ꢀ
aminoꢀ5ꢀcyanopyridine 17 formed the urea, and subsequent
acid hydrolysis of the acetal gave 6b in 70% overall yield. To
prepare 11a the phenyl carbamate of the 2ꢀaminopyridine 18
was heated with the 2ꢀ(dimethoxymethyl)ꢀdihydroazepine 19,
in a reaction that likely proceeds via the highly reactive and
labile 2ꢀpyridylisocyanate.18 The acetal group in the resulting
urea was then hydrolysed to give 11a in 43% overall yield.
The limiting kinetic factor in HCC was anticipated to be the
relatively rapid FGFR4 resyntheis rate (< 2 h),4 suggesting no
obvious benefit for inhibitors with residences times beyond
this value. In fact, a faster target engagement was anticipated
to be more advantageous in this situation, to rapidly engage
the newly synthesised protein, and ensure the maximum level
of inhibition can be maintained without the need for high inꢀ
hibitor concentrations. This target kinetic understanding was
pivotal at the onset of the project to prioritise the 2ꢀFPU series
ahead of other covalent hitꢀseries interacting by what would be
considered irreversibleꢀcovalent interactions, and ultimately
differentiated the project from the acrylamideꢀbased inhibitors
described by other groups.4 With such a rapid target turnover,
and the need for a continuous high level of pathway inhibition
to achieve maximum efficacy, sustained exposure to a reversiꢀ
bleꢀcovalent inhibitor was considered to be safer, and to potenꢀ
tially offer a simpler pharmacokinetic opportunity.4
In conclusion, scaffold morphing from a quinoloneꢀamide hit
Scheme 1. Synthesis of the THNU analogue 6b.
The slowꢀbinding kinetics provided further support for the
hypothesized reversibleꢀcovalent interaction for the 2ꢀFPU
series with FGFR4. However, we sought additional evidence
for this mechanism of action as a way to increase our underꢀ
standing of the inhibitor/protein interaction, and to assist with
the further optimisation of the series. Consistent with our preꢀ
vious observations the benzylic alcohols derived by reduction
of the aldehyde, as exemplified by 13 in Figure 8, showed no
FGFR4 inhibitory activity.4 However, in the absence of covaꢀ
lent binding an unfavorable steric interaction was anticipated
to occur between the side chain of Cys552 and the benzylic
alcohol of 13.12 Therefore, to better understand the strength of
the hingeꢀ binding interaction, and the contribution of the alꢀ
dehyde, the desformyl analogue of the more soluble 6h was
prepared, compound 14 in Figure 8 (HTES 173 ꢁM). This
compound inhibited FGFR4 with a biochemical activity 400x
weaker than the parent 6h, and highlighted the key contribuꢀ
tion from the aldehyde, consistent with the proposed mechaꢀ
nism of action. Interestingly, this makes for a very different
Reagents and conditions: (a) LiHMDS, (PhO)2CO, THF, ꢀ15
°C, 1.5 h, 78%; (b) LiHMDS, 2ꢀaminoꢀ5ꢀcyanopyridine, THF,
ꢀ15 °C, 1.5 h, 94%; (c) HCl(aq), THF, 25 °C, 1 h, > 95%.
Scheme 2. Synthesis of the 2ꢀFPU analogue 11a.
Reagents and conditions: (a) PhO(CO)Cl, pyridine, THF, 25
°C, 5 h, 84%; (b) 19, DMAP, THF, reflux, 23 h, 54%; (c)
HCl(aq), THF, 25 °C, 1 h, > 95%.
ACS Paragon Plus Environment

ACS Medicinal Chemistry Letters
Page 6 of 8
factor receptor 4: a putative key driver for the aggressive phenoꢀ
type of hepatocellular carcinoma. Carcinogenesis 2014, 35, 2331ꢀ
2338.
identified a number of 2ꢀFPU variants with increased FGFR4
activity and improved physicochemical properties. In particuꢀ
lar, potent and selective THNU 6, dihydropyrrolopyridine urea
10, and ringꢀopened urea 12 analogues were identified for
further optimisation, with examples with cellular FGFR4 acꢀ
tivities below 50 nM. Characterisation of the binding kinetics
and preliminary SAR surrounding the 2ꢀFPUs remained conꢀ
sistent with the hypothesized reversibleꢀcovalent interaction.
This in combination with an understanding of the FGFR4
turnover rate in HCC provided insights into how best to evaluꢀ
ate and optimise the series going forward. Future disclosures
will focus upon how the leads and mechanistic understanding,
described herein, were used to identify the clinical candidate
FGF401.
1
2
3
4
5
6
7
8
4. Fairhurst, R. A.; Knoepfel, T.; Leblanc, C.; Buschmann, N.; Gaul,
C.; Blank, J.; Galuba, I.; Trappe, J.; Zou, C.; Voshol, J.; Genick,
C.; BrunetꢀLefeuvre, P.; Bitsch, F.; GrausꢀPorta, D.; Furet, P. Apꢀ
proaches to selective fibroblast growth factor receptor 4 inhibition
through targeting the ATPꢀpocket middleꢀhinge region. Med.
Chem. Commun. 2017, 8, 1604ꢀ1613.
5. Hierro, C.; Rodon, J.; Tabernero, J. Fibroblast growth factor
(FGF) receptor/FGF inhibitors: novel targets and strategies for opꢀ
timization of response of solid tumors. Sem. Oncol. 2015, 42,
801ꢀ819.
6. Leproult, E.; Barluenga, S.; Moras, D.; Wurtz, J.ꢀM.; Winssinger,
N. Cysteine mapping in conformationally distinct kinase nucleoꢀ
tide binding sites: application to the design of selective covalent
inhibitors. J. Med. Chem. 2011, 54, 1347–1355.
7. Fairhurst, R. A.; Knoepfel, T.; Furet, P.; Buschmann, N.; Leblanc,
C.; Mah, R.; Kiffe, M.; GrausꢀPorta, D.; Weiss, A.; Kinyamuꢀ
Akunda, J.; Wartmann, M.; Trappe, J.; Gabriel, T.; Hofmann F.;
Sellers, W. R. FGF401: A reversible-covalent inhibitor of FGFR4
for the treatment of hepatocellular carcinoma. Presented at the
253rd National Meeting, San Francisco, April 5th, 2017, American
Chemical Society, MEDI 353.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website.
Details for: the synthesis of the compounds and characterizing
data; the physicochemical measurements; the biochemical and
cellular kinase assays (PDF)
8. GrausꢀPorta, D.; Weiss, A.; Stamm, C.; Fairhurst, R. A.; Wartꢀ
mann, M.; Knoepfel, T.; Leblanc, C.; Buschmann, N.; Blank, J.;
Furet, P.; Galuba, I.; Trappe, J.; Mignault, A.; Ramkumar, T.;
Kauffmann, A.; Engelman, J. A.; Hofmann, F.; Sellers, W. R.
NVP-FGF401, a first-in-class highly selective and potent FGFR4
inhibitor for the treatment of HCC. Presented at the AACR Annuꢀ
al Meeting, Washington, April 1–5, 2017, American Association
for Cancer Research, abstract 2098.
AUTHOR INFORMATION
Corresponding Author
*Eꢀmail: robin.fairhurst@novartis.com
9. Weiss, A.; GrausꢀPorta, D.; Reimann, F.; Buhles, A.; Stamm, C.;
Fairhurst, R. A.; KinyamuꢀAkunda, J.; Sterker, D.; Kiffe, M.; Muꢀ
rakami, M.; Wartmann, M.; Wang, Y.; Engelman, J. A.; Hofꢀ
mann, F.; Sellers, W. R. NVP-FGF401: Cellular and in vivo pro-
file of a novel highly potent and selective FGFR4 inhibitor for the
treatment of FGF19/FGFR4/KLB+ tumors. Presented at the
AACR Annual Meeting, Washington, April 1–5, 2017, American
Association for Cancer Research, abstract 2103.
10. Caraballo, R.; Dong, H.; Ribeiro, J. P.; JiménezꢀBarbero J.; Romꢀ
ström, O. Direct STD NMR identification of β–galactose inhibiꢀ
tors from a virtual dynamic hemithioacetal system. Angew.
Chem., Int. Ed. 2010, 49, 589–593.
11. Lovering, F.; Bikker, J.; Humblet, C. Escape from flatland: inꢀ
creasing saturation as an approach to improving clinical success.
J. Med. Chem. 2009, 52, 6752ꢀ6756.
12. Further details are included in the Supporting Information.
13. Tucker, J. A.; Klein, T.; Breed, J.; Overman, R.; Phillips, C.;
Norman, R. A. Structural insights into FGFR kinase isoform seꢀ
lectivity: diverse binding modes of AZD4547 and ponatinib in
complexes with FGFR1 and FGFR4. Structure 2014, 22, 1764–
1774.
14. Bernetti, M.; Cavalli, A.; Mollica, L. Proteinꢀligand (un)binding
kinetics as a new paradigm for drug discovery at the crossroad beꢀ
tween experiments and modelling. Med. Chem. Commun. 2017, 8,
534ꢀ550.
15. Copeland, R. A. The drugꢀtarget residence time model: a 10ꢀyear
retrospective. Nat. Rev. Drug Discovery 2016, 15, 87ꢀ95.
16. Grütter, C.; Simard, J. R.; MayerꢀWrangowski, S. C.; Schreier, P.
H.; PérezꢀMartín, J.; Richters, A.; Getlik, M.; Gutbrod, O.; Braun,
C. A.; Beck, M. E.; Rauh, D. Targeting GSK3 from ustilago
maydis: typeꢀII kinase inhibitors as potential antifungals. ACS
Chem. Biol. 2012, 7, 1257ꢀ1267.
17. Wang, Y.; Chen, Z.; Dai, M.; Sun, P.; Wang, C.; Gao, Y.; Zhao,
H.; Zeng, W.; Shen, L.; Mao, W.; Wang, T.; Hu, G.; Li, J.; Chen,
S.; Long, C.; Chen, X.; Liu, J.; Zhang, Y. Discovery and optimiꢀ
zation of selective FGFR4 inhibitors via scaffold hopping. Bioorg.
Med. Chem. Lett. 2017, 27, 2420ꢀ2423.
Author Contributions
All authors have given approval to the final version of the manuꢀ
script. The authors declare the following financial interest(s):
Authors include employees of Novartis AG and/or stockholders of
Novartis AG.
ACKNOWLEDGMENT
The authors would like to thank Damien Hubert, Van Huy Luu,
Elvira Masso, Michel Niklaus, Pierre Nimsgern, Milen Todorov,
and Jasmin Wirth for technical assistance in the preparation of the
compounds; Stephane Rodde for making the physicochemical
measurements; Mario Centeleghe and Jacqueline Loretan for deꢀ
velopment and execution of the cellꢀbased assays; Proteros GmbH
for carrying out the kinetic binding assays; Ina Dix and Philippe
Piechon for performing small molecule Xꢀray crystallography.
ABBREVIATIONS
FGFR4, fibroblast growth receptor 4: FGFs, fibroblast growth
factors: HCC, hepatocellular carcinoma: GK+2, gateꢀkeeper resiꢀ
due plus two: THNU, tetrahydronaphthyridine urea: HTES, high
throughput equilibrium solubility: 2ꢀFPU, 2ꢀformylpyridine urea:
SAR structure activity relationship: TEL, translocation ets leukeꢀ
mia: LiHMDS, lithium hexamethyldisilazide: DMAP, 2,6ꢀ
dimethylaminopyridine.
REFERENCES
1. Carter, E. P.; Fearon, A. E.; Grose, R. P. Careless talk cost lives:
fibroblast growth factor signaling and the consequences of pathꢀ
way malfunction. Trends Cell Biol. 2015, 25, 221–233.
2. Lemieux, S.; Hadden, M. K. Targeting the fibroblast growth facꢀ
tor receptors for the treatment of cancer. Anti-Cancer Agents Med.
Chem. 2013, 13, 748–761.
3. Gauglhofer, C.; Paur, J.; Schrottmaier, W. C.; Wingelhofer, B.;
Huber, D.; Naegelen, I.; Pirker, C.; Mohr, T.; Heinzle, C.;
Holzmann, K.; Marian, B.; SchulteꢀHermann, R.; Berger, W.;
Krupitza, G.; Grusch, M.; GraslꢀKraupp, B. Fibroblast growth
18. Holt, J.; Andreassen, T.; Bakke, J. M.; Fiksdahl, A. Nitropyridyl
isocyanates. J. Heterocyclic Chem. 2005, 42, 259ꢀ264.
ACS Paragon Plus Environment

Page 7 of 8
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
7
ACS Paragon Plus Environment

ACS Medicinal Chemistry Letters
Page 8 of 8
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
136x55mm (300 x 300 DPI)
ACS Paragon Plus Environment