Heterobicyclic inhibitors of transforming growth factor beta receptor I (TGFβRI)

Article abstract of DOI:10.1016/j.bmc.2018.01.014

The TGFβ-TGFβR signaling pathway has been reported to play a protective role in the later stages of tumorigenesis via increasing immunosuppressive Treg cells and facilitating the epithelial to mesenchymal transition (EMT). Therefore, inhibition of TGFβR has the potential to enhance antitumor immunity. Herein we disclose the identification and optimization of novel heterobicyclic inhibitors of TGFβRI that demonstrate potent inhibition of SMAD phosphorylation. Application of structure-based drug design to the novel pyrrolotriazine chemotype resulted in improved binding affinity (Ki apparent = 0.14 nM), long residence time (T_1/2 > 120 min) and significantly improved potency in the PSMAD cellular assay (IC₅₀ = 24 nM). Several analogs inhibited phosphorylation of SMAD both in vitro and in vivo. Additionally, inhibition of TGFβ-stimulated phospho-SMAD was observed in primary human T cells.

Full text of DOI:10.1016/j.bmc.2018.01.014

Accepted Manuscript
Heterobicyclic inhibitors of transforming growth factor beta receptor I
(TGFβRI)
Lalgudi S. Harikrishnan, Jayakumar Warrier, Andrew J. Tebben, Gopikishan
Tonukunuru, Sudhakara R. Madduri, Vishweshwaraiah Baligar, Raju
Mannoori, Balaji Seshadri, Hasibur Rahaman, P.N. Arunachalam, Amol
Dikundwar, Brian E. Fink, Joseph Fargnoli, Mark Fereshteh, Yi Fan, Jonathan
Lippy, Ching-Ping Ho, Barri Wautlet, Steven Sheriff, Max Ruzanov, Robert M.
Borzilleri
PII:
S0968-0896(17)32023-0
https://doi.org/10.1016/j.bmc.2018.01.014
BMC 14168
DOI:
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
15 October 2017
10 January 2018
20 January 2018
Please cite this article as: Harikrishnan, L.S., Warrier, J., Tebben, A.J., Tonukunuru, G., Madduri, S.R., Baligar, V.,
Mannoori, R., Seshadri, B., Rahaman, H., Arunachalam, P.N., Dikundwar, A., Fink, B.E., Fargnoli, J., Fereshteh,
M., Fan, Y., Lippy, J., Ho, C-P., Wautlet, B., Sheriff, S., Ruzanov, M., Borzilleri, R.M., Heterobicyclic inhibitors
of transforming growth factor beta receptor I (TGFβRI), Bioorganic & Medicinal Chemistry (2018), doi: https://
doi.org/10.1016/j.bmc.2018.01.014
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Bioorganic & Medicinal Chemistry
Heterobicyclic inhibitors of transforming growth factor beta receptor I (TGFRI)
Lalgudi S. Harikrishnana*, Jayakumar Warrierb, Andrew J. Tebbenc, Gopikishan Tonukunurub, Sudhakara
R. Maddurib, Vishweshwaraiah Baligarb, Raju Mannoorib, Balaji Seshadrib, Hasibur Rahamanb, P. N.
Arunachalamb, Amol Dikundwarb, Brian E. Finka, Joseph Fargnolid, Mark Fereshtehe, Yi Fane, Jonathan
Lippye, Ching-Ping Hod, Barri Wautletd, Steven Sheriffc, Max Ruzanovc and Robert M. Borzilleria.
a Department of Chemistry, Bristol-Myers Squibb Company, P.O. Box 4000, Princeton, NJ, USA
b BBRC, Special Economic Zone, Biocon Park, Jigani Link Road, Bangalore, India
c Molecular Structure & Design, Bristol-Myers Squibb Company, P.O. Box 4000, Princeton, NJ, USA
d Immuno-Oncology Small Molecule Biology, Bristol-Myers Squibb Company, P.O. Box 4000, Princeton, NJ, USA
e Lead Discovery & Optimization,, Bristol-Myers Squibb Company, P.O. Box 4000, Princeton, NJ, USA
ARTICLE INFO
ABSTRACT
Article history:
Received
Received in revised form
Accepted
The TGF-TGFR signaling pathway has been reported to play a protective role in the later
stages of tumorigenesis via increasing immunosuppressive Treg cells and facilitating the
epithelial to mesenchymal transition (EMT). Therefore, inhibition of TGFR has the potential to
enhance antitumor immunity. Herein we disclose the identification and optimization of novel
heterobicyclic inhibitors of TGFRI that demonstrate potent inhibition of SMAD
phosphorylation. Application of structure-based drug design to the novel pyrrolotriazine
chemotype resulted in improved binding affinity (Ki apparent = 0.14 nM), long residence time
(T1/2 > 120 min) and significantly improved potency in the PSMAD cellular assay (IC50 = 24
nM). Several analogs inhibited phosphorylation of SMAD both in vitro and in vivo.
Additionally, inhibition of TGF-stimulated phospho-SMAD was observed in primary human T
cells.
Available online
Keywords:
TGF
TGF-beta
TGF
ALK
2009 Elsevier Ltd. All rights reserved.
TGFBR
1. Introduction
the recruitment of TGFRI leading to the formation of a
tetrameric complex comprised of two units each of Receptor II
TGF-signaling mediates protumorigenic changes in the
tumor microenvironment (TME) and promotes epithelial to
mesenchymal transition (EMT), both of which aid in tumor
progression and invasiveness.1 In multiple mouse models, TGF
pathway inhibition has been reported to increase the activity of
anti-tumoral adaptive immune cells.2,3 In humans, beneficial
effects of TGF pathway inhibition have been reported in early
stage clinical trials.4 Whereas the reported approaches target the
tumor directly, our approach was distinct in that we targeted the
immune system. 1, 5
(RII) and Receptor I (RI). The formation of this complex allows
RII to cross phosphorylate RI, which results in its activation and
subsequent recruitment and phosphorylation of SMAD proteins.
The phosphorylation of SMAD results in the dimerization of
SMAD, translocation to the nucleus followed by gene
transcription.6 We set out to identify a potent TGFR inhibitor
that would inhibit SMAD phosphorylation in a dose dependent
manner, both in vitro and in vivo. Moreover, such an inhibitor
would be expected to inhibit TGF-stimulated Phospho-SMAD
levels in primary human T cells.
TGFRI and TGFRII are transmembrane serine-threonine
kinases that are involved in signal transduction in the TGF
canonical pathway. The binding of TGF to TGFRII results in
2. Structure Activity Relationship

Figure 1. Comparison of 1,3-disubstituted six membered ring compound 3 with 1,2-disubstituted five membered ring analogs 1 and 2.
Compounds were evaluated by a homogenous time resolved
fluorescence (HTRF) binding assay7 to monitor their ability to
bind to TGFRI and RII. Potent analogs were then evaluated for
their ability to inhibit phosphorylation of SMAD protein in a
cellular assay.8
1,3-disubstituted 6-membered ring.9 Comparison of the predicted
binding modes of a model 1,3-disubstituted system, 2,4-
disubstituted pyrrolotriazines with a quinoline at the 4-position,
to 2 revealed that the key hydrogen bond between the quinoline
nitrogen and the hinge residue histidine 283 was not present in
the 2,4-model (Figure 2, panel A) when the quinoline is linked
directly through its 4-position, as it is in 2. However, modeling
suggested that an additional nitrogen between the pyrrolotriazine
core and the quinoline would restore the hydrogen bond while
maintaining good overlap of the core and 2-(6-methylpyridyl)
substituent (Figure 2, panel B).
Several reported TGFR inhibitors such as pyrrolopyrazoles 1
and 2 have a 1,2-disubstituted five membered ring as a common
structural motif.4 Several X-ray co-crystal structures have been
published detailing the favorable interactions between such small
molecule inhibitors and the TGF receptor kinase domain. We
envisaged accessing the same set of favorable interactions with a
Figure 2: Comparison of the modeled binding modes of 2 (blue) and 2,4-disubstituted pyrrolotriazines (orange). Hydrogen bonds are
indicated by the dashed lines. A: Directly linked 4-quinoline. B: Amino linked 4-quinoline.
Table 1. SAR at 2- and 4-positions of pyrrolotriazine

#LM Stability %
Remaining (h, r, m)
IC50 (uM)
TGFRI
0.0013
0.0025
0.019
Cmpd #
R1
R2
R3
TGFRII
0.55
0.077
>15
PSMAD*
0.099
0.35
3.9
4
Me
H
H
H
H
H
H
H
F
11, 28, 1
33, 33, 1
12, 36, 2
24, 48, 13
11, 35, 6
15, 39, 7
65, 67, 26
54, 54, 19
5
H
6
OMe
CF3
Me
Me
Me
H
H
7
H
0.0068
0.00096
0.0034
0.062
8.5
0.71
0.18
1.2
8
F
0.57
>15
9
F
10
CONH(iPr)
H
>15
>10
11
F
H
0.0017
0.11
0.50
*Inhibition of phosphorylation and subsequent nuclear translocation of SMAD in mink lung epithelial cells.8 #Liver microsomal stability assays were carried out
as described earlier.10
To test the model, pyrrolotriazine 3 was synthesized, which
demonstrated potencies comparable to pyrazole 2. While several
close analogs of 3 with substituted aminoquinolines at the 4-
position demonstrated very good potency, these compounds
exhibited very poor solubility even in DMSO. Therefore, the
corresponding pyridine analog 4 was synthesized (Table 1) and
observed to demonstrate potency comparable to 3 and
significantly improved solubility in organic solvents, making
purification and isolation more convenient.
microsomes, a small amount (13%) of trifluoro compound 7 was
detected in the assay.9 It was hypothesized that electron
withdrawing groups on the pyridyl ring might deplete electron
density on pyridyl nitrogen, making it a poorer ligand for the iron
atom in the heme, and consequently a poorer substrate for CYP
mediated metabolism.11 Therefore, electron withdrawing
substituents on 4-pyridyl ring were evaluated. Small
improvements in liver microsomal stability were observed with
fluoro substituted compounds 8 and 9 (6% and 7% remaining
after 10 minutes).9 A more pronounced improvement in
microsomal stability was observed with amide compound 10
(26% remaining after 10 minutes). Unfortunately, amide
substituted compound 10 did not inhibit SMAD phosphorylation
in the Mv1Lu lung epithelial cell line.
With the unsubstituted aminopyridine at 4-position of
pyrrolotriazine, SAR at 2-position was investigated. Removing
the methyl group on 2-pyridyl retained potency against TGFRI
but eroded selectivity versus RII (4 vs. 5). Since inhibition of RII
in addition to RI did not offer any improvement in cellular
potency, we focused on RI selective inhibitors to selectively
inhibit the SMAD dependent canonical pathway. When the
methyl at R1 was replaced by a sterically more demanding
methoxy or trifluoromethyl group, a further improvement in
selectivity for RI was observed (6 and 7). In all cases the mouse
liver microsomal (mLM) stability was poor, which precluded in
vivo investigation. While no measurable amounts of compounds
4-6 were remaining after 10 minute incubation with mouse liver
In an effort to understand the site(s) of metabolism, one of the
metabolically more stable compounds 11 was evaluated in
biotransformation studies.12 Multiple fast-forming metabolites
were observed in these studies as outlined in Figure 3. Rather
than selectively blocking a specific metabolic soft-spot, we
introduced polar groups on the molecule with the aim of
improving microsomal stability.
Figure 3. Biotransformation studies of 11. The main soft spot in both human and mouse liver microsomes was the pyrrolotriazine-
fluoropyridyl region, which was subject to oxidation (M1) and bis oxidation (M3). Another soft spot in both human and mouse liver
microsomes was the pyridine moiety, which was subject to oxidation (M2).
The positions likely to be tolerant of polar substitution were
identified from a model of 11 in TGFR1 (Figure 4). The
predicted binding mode showed that pyrrolotriazine positions 1,
3 and 7 are buried within the binding site, but 5 and 6 are solvent
accessible. In addition, the amino pyridine at the 4-position
presented additional valencies that are solvent exposed. From this
set, the 5-position was the least sterically encumbered and would
project substituents toward the ribose pocket, making it the
optimal site for further substitution.

Figure 4. Binding model of 11 in TGFRI. A: Surface representation of the predicted binding mode of 11 (blue sticks) in the TGFRI
ATP binding site (Fluorine atom in 11 is omitted for clarity). For reference, the binding mode of AMP-PNP from the TGFRII crystal
structure (PDB id 5E92) is shown in thin white sticks. B: Ligand interaction diagram of 11 in TGFRI. Interacting residues are shown
as circles with hydrogen bonds denoted as straight dashed lines. The protein contour is indicated by the curved dashed lines.
Substituents enclosed by this line are buried within the protein. Ligand atoms highlighted by the purple clouds are solvent exposed and
the degree of solvent exposure is proportional to the size of the circle.
Table 2. SAR at 5-position of pyrrolotriazine
IC50 (uM)
LM Stability %
Remaining (h, r, m)
Cmpd #
R =
H
TGFRI
TGFRII
PSMAD
0.50
11
0.0017
0.11
54, 54, 19
12
0.0043
2.0
0.076
86, 72, 48
13
14
0.0039
0.0044
0.81
3.0
0.39
0.31
85, 90, 82
93, 92, 100
In accordance with proposed binding model, SAR was
investigated at 5- position of pyrrolotriazine (Table 2). Indeed,
polar substituents at the 5- position of pyrrolotriazine resulted in
potent analogs with significantly improved microsomal stability,
but poor selectivity across a panel of over 100 kinases (data not
shown). Compounds that lacked substituents at the 5- position
demonstrated >100 fold selectivity versus an in-house panel of
over 60 kinases and minimal to no cytotoxicity in an in-house

cytotoxicity panel. Since polar analogs (e.g., 12-14) interact with
the ribose binding pocket, many analogs bound to several other
kinases (such as CK1, CRIK and Haspin) and displayed single
digit micromolar cytotoxicity in multiple tumor cell lines (such
as Colo678, DU4475 and MDA-MB-468). Therefore polar
substituents near the hinge binding region were evaluated (Table
3). Introduction of small hydrogen bond donors (such as NHAc
in compounds 15 and 16, that cannot reach the ribose binding
pocket) near the hinge region did result in analogs with improved
microsomal stability while retaining potency and selectivity
(cytotoxicity was not observed across a panel of cancer cell
lines). However, sterically more demanding piperazine
compound 17 resulted in cytotoxicity across multiple cell lines.
This is perhaps due to favorable binding interactions with the
polar (ribose binding) pocket that is present in all kinases.
Overall, compound 15 demonstrated comparable microsomal
stability and improved potency in PSMAD and human primary T
cell assays compared to published inhibitor 1.13 Having achieved
modest microsomal stability with pyrrolotriazine 15, we aspired
to improve the binding potency (to increase target coverage) by
designing new analogs via structure based drug design.
Table 3. SAR of hydrogen bond donors near hinge region of pyrrolotriazine
IC50 (uM)
*TGF stim.
PSMAD inh.
in hu Primary
T cells IC50
(uM)
LM
Stability %
Remaining
(h, r, m)
Cmpd #
R1
R2
R3
TGFRI
TGFRII
PSMAD
7
CF3
CF3
CF3
H
H
F
H
0.0068
0.0014
0.025
8.5
0.71
0.18
NT
24, 48, 13
75, 35, 49
78, 68, 66
15
16
NHAc
NHAc
NT
>15
0.12
17
CF3
H
0.0021
NT
0.22
70, NT, 77
1
N/A
N/A
N/A
0.0037
0.19
0.75
58, 46, 36
0.37
*TGF stimulated PSMAD inhibition assay is described in the experimental section 5.1; NT = Not tested; N/A = Not applicable
The proximity of the 7-position to Asp-351 in the model of 11
highlighted the potential for an additional hydrogen bond (Figure
4). This hydrogen bond is accessible via replacement of CH at 7-
position of pyrrolotriazine with a hydrogen bond donor such as
NH. Indeed, introduction of a hydrogen bond donor at 7-position
of 8, resulted in purine 18 and pyrrolopyrimidine 19 that
demonstrated improved binding affinities (Ki apparent for 15 =
1.3 nM), long residence times in the TGFRI binding pocket
(T1/2 for 18 & 19 > 120 min), translating to superior potency in
the cellular assay (Table 4). Replacing the methyl group in 18
and 19 with trifluoromethyl group, resulted in potent analogs
with improved microsomal stability (20 & 21). Our structural
hypothesis was confirmed by an X-ray crystal structure of a
related pyrrolopyrimidine analog, 22 which clearly shows a
hydrogen bond between the pyrrole NH and side chain carbonyl
of Asp-351 (Figure 5).
Figure 5. X-ray crystal structure of 22 in TGFRI (PDB id
6B8Y).

Table 4. Chemotype hopping to purine and pyrrolopyrimidine
IC50 (uM)
Ki
LM
TGF stim.
Cmpd
#
apparent
TGFRI
(nM)
Stability %
Remaining
(h, r, m)
PSMAD inh. in
hu Prim. T cells
IC50 (uM)
X
Y
R1
R2
R3
TGFRI
TGFRII
PSMAD
18
19
20
21
22
NH
NH
NH
NH
NH
N
Me
Me
CF3
CF3
CF3
F
H
0.0052
0.0070
0.0011
0.0010
0.00055
1.1
0.15
0.14
0.14
NT
38, 47, 4
0.026
0.0051
0.050
0.012
NT
CH
N
F
H
0.31
NT
0.024
0.13
25, 38, 2
H
H
F
NHAc
NHAc
H
85, 77, 59
81, 80, 27
36, NT, 28
CH
CH
0.94
>15
0.043
0.047
0.15
0.22
However, these new chemotypes suffered from poor aqueous
solubility compared to their pyrrolotriazine counterparts. The
poor aqueous solubility can be rationalized based on small
molecule X-ray structure of a related compound 27, which
revealed a network of intermolecular hydrogen bonds (Figure 2S
in supporting information). Consequently, purine and
pyrrolopyrimidine analogs were deprioritized relative to the
120
pyrrolotriazines due to poor physiochemical properties.
100
80
60
40
20
0
Owing to acceptable physiochemical properties,
pyrrolotriazine chemotype was selected for in vivo evaluation
despite being less potent compared to pyrrolopyrimidines in
cellular assays. Pyrrolotriazine 15 was evaluated in
pharmacodynamic (PD) studies using 4T1 mouse mammary
carcinoma model at a single oral dose of 100, 50 and 25 mg/kg
each. The serum exposures at 7 h after first dose were greater
than PSMAD cellular IC50 for all three doses (Figure 5).
Control
125 mpk 100 mpk
Cmpd. 1
50 mpk
Cmpd. 15
25 mpk
Figure 6. In vivo PD evaluation of 15 in the 4T1 mouse tumor model
revealed dose-dependent reduction in the levels of PSMAD that was
comparable to published inhibitor 1.
Figure 5. Exposures of cmpds. 1 and 15 in serum 7 h after forst dose.
3. Chemistry
A general route to access 2,4-disubstituted pyrrolotriazine
analogs is outlined in Scheme 1. Dichloropyrrolotriazine 2314
was converted to the 4-phenoxy derivative 24 using sodium
phenoxide in tetrahydrofuran. Palladium mediated coupling of
chloroaryl compound 24 with metallated pyridine afforded 2-
pyridyl pyrrolotriazine 25. Nucleophilic aromatic displacement of
phenoxide with aminopyridine gave amonipyridyl compound 26.
Similar chemistry was employed to access purine and
Compound 15 demonstrated dose-dependent PSMAD
inhibition in mice (Figure 6). The magnitude of PSMAD
inhibition was comparable to published inhibitor 1.13 (see
Experimental section 5.1 for details of PD study)
pyrrolopyrimidine compounds. Details of synthetic route to
access purine and pyrrolopyrimidine analogs are described in the
Supporting Information section (Scheme 1S).

Scheme 1: Conditions: a. NaOPh, THF, 1 h; b. ArSnBu3, [PdCl2(PPh3)2], toluene, 100 oC, 18 h; c. ArB(OH)2, [Pd(dppf)Cl2], dioxane, K3PO4, 90 oC, 18 h; d.
ArBr, Bu3SnSnBu3, [Pd(PPh3)4], dioxane, 110 oC, 18 h; e. ArBr, Bu3SnSnBu3, [Pd(PPh3)4], dioxane, 140 oC, microwave radiation; f. ArNH2, NaH, DMF, 1 h.
4. Conclusion
unstimulated cells and the total was the 2.5 ng/mL TGF
stimulation without compound inhibition.
Based on published 1,2-disubstituted TGFRI inhibitors, 1,3-
disubstituted series was designed and evaluated. SAR
TGFβ stimulated PSMAD inhibition assay in primary human
T cells: Primary human T cells were starved by centrifuging and
resuspending in RPMI media (Corning Cellgro, Cat# 10-040-
CM) + 0.5% FBS (Sigma, Cat# F4135-500mls). Cells were
incubated overnight at 37 °C. Compounds were diluted from 5
mM, 1:3 in DMSO (Sigma, Cat # D2650), as a 10 point series in
v bottom plates (Nunc, Cat # 442581). 400 l of RPMI + 0.5%
FBS were added to each well of a 1 mL deep well block
(Axygen, Cat# PDW11CS). 2 l of the compound in DMSO
were transferred to the media in the deep well blocks (1:200) and
mixed. 80 l, in duplicate rows, of the compound + media
solution were added to 96 well RB plates (Corning Costar, Cat #
3799). Starved cells were counted, centrifuged, and cell pellets
were re-suspended at 1.5x106 cells/mL with RPMI + 0.5% FBS.
100 l of cells (1.5 x 105 cells/well) were added to each well of
the plate, then incubated for 1 h at 37 °C. Stock of TGFβ (rh
TGFβ- R&D Systems, Cat # 240-B/CF; 33.3 ug/mL) was diluted
33,000x to 1 ng/mL in RPMI + 0.5% FBS (10 x). Each well of
columns 1-11 was stimulated with 20 µl/well of 1.0 ng/mL (10x)
TGF. (0.1 ng/mL Final). 20 µl of RPMI + 0.5% FBS were
added to each well of column 12. Plates were incubated at 37 °C
for 90 minutes, centrifuged, flicked off media, and the cells were
lysed with 50 µl 0.2% Triton X-100 (Sigma, Cat# 93443-500`1)
in PBS (Corning Cellgro, Cat# 21-040-CM) containing protease
inhibitors (Roche, Cat# 11-836-153-001 ) and phosphatase
inhibitors (Roche, Cat# 04-906-837-001). PSMAD AlphaLisa
was run according to manufacturer instructions (Surefire
PSMAD Alpha Screen-Perkin Elmer, Cat# TGRSM3S10K;
Protein A Alphalisa Acceptor Beads- Perkin Elmer, Cat#
AL101M; Streptavidin Donor Beads- Perkin Elmer, Cat#
6760002); incubated overnight at room temperature and read on
appropriate Alpha screen reader.
In vivo PD assay: 4T1 cells (1x106 cells per mouse) were
subcutaneously implanted into Balb/c female mice. Six mice per
cohort with tumor volume around 150 mm3 were chosen for this
study. Single oral doses of 25, 50,and 100 mpk of compound 15
was administered to tumored mice except the control group. At 7
hours and 24 hours post dose, 3 mice from each group were
euthanized and whole blood (serum) was collected for compound
PK. At 24 h, tumors were excised and immediately snap frozen
in liquid nitrogen. Frozen tumors were homogenized in RIPA
buffer (Sigma-R0278) containing protease and phosphatase
inhibitors. Western Blotting was performed with SMAD
antibodies (Cell Signaling- 3103 and 3108) in PeggySue
(ProteinSimpleTM). Western Blotting results were analyzed using
CompassTM software from ProteinSimple Inc.
investigation led to the identification of several analogs with
excellent potency and modest metabolic stability. Introduction of
a hydrogen bonding donor (NH) led to the identification of two
new chemical series with improved potencies in cellular PSMAD
assay. Additionally, inhibition of TGF-stimulated PSMAD
levels in human primary T cells was also observed.
Pyrrolotriazine derivative 15 that demonstrated a good balance
between potency and physiochemical properties was evaluated in
vivo. Compound 15 demonstrated dose-dependent inhibition of
PSMAD in vivo, which confirmed target engagement. Efforts to
design chemotypes that capitalize on hydrogen bonding with
Asp-351, thereby increasing the residence time and improving
cellular potency, without compromising on physiochemical
properties is forthcoming.
5. Experimental Section
5.1 Protocols for biology assays:
HTRF Assays: Assays were conducted in 1536-well plates
and 2 mL reactions were prepared from addition of HIS-TGFRI
T204D kinase domain15 or HIS-TGFRII WT kinase domain,
anti-HIS detection antibody, a labeled small molecule probe (Kd
= <100 nM; koff = <0.001 s-1) and test compounds in assay buffer
(20 mM HEPES pH 7.4, 10 mM MgCl2, 0.015% Brij35, 4 mM
DTT, and 0.05 mg/mL BSA). The reaction was incubated for 1
hour at room temperature and the HTRF signal was measured on
an Envision plate reader (Ex: 340 nm; Em: 520 nm / 495 nm).
Inhibition data were calculated by comparison to no enzyme
control reactions for 100% inhibition and vehicle-only reactions
for 0% inhibition. The final concentration of reagents in the
assay were 1 nM HIS-TGFRI T204D or HIS-TGFRII WT, 0.2
nM anti-HIS detection antibody, labeled small molecule probe (at
Kd) and 0.5% DMSO. Dose response curves were generated to
determine the concentration required inhibiting 50% of kinase
activity (IC50). Compounds were dissolved at 10 mM in dimethyl
sulfoxide (DMSO) and evaluated at eleven concentrations. IC50
values were derived by non-linear regression analysis.
PSMAD2 Assay: Mv1Lu mink lung epithelial cells were
plated 4000 cells/well in RPMI+10% FBS in 384 well PDL
coated plate overnight and then starved in RPMI +0.1% FBS for
20 hours. The cells were echo dispensed with compounds and
incubated for 1 hour. 2.5 ng/mL TGF was applied to stimulate
the cells for 15 min. The reaction was stopped with 4%
formaldehyde fixation buffer for 20 min. The cells were
permeabilized with 0.1% triton, and stained for DNA and total
SMAD. The plate was read using Cellomics ArrayScan Health
Profiling BioApplication, 10 x objectives. The data was analyzed
in Toolset and Curvemaster. The background was from
5.2 General Procedure and Analytical Data:
5.2.1 General procedure:

Phenoxy intermediate 24: To a 100 mL flask was added 2,4-
dichloropyrrolo[2,1-f][1,2,4]triazine 2314 (15.96 mmol), and
tetrahydrofuran (40 mL). To the resulting solution was added
sodium phenolate (17.55 mmol) portion wise. After 1 h, an
aliquot of the reaction mixture was diluted with methanol and
analyzed by LCMS to ensure complete conversion. The reaction
mixture was concentrated. The residue was stirred with water,
filtered, washed extensively with water and dried to obtain 2-
chloro-4-phenoxypyrrolo[2,1-f][1,2,4]triazine 24.
N-[2-(6-Methoxypyridin-2-yl)pyrrolo[2,1-f][1,2,4]triazin-4-
yl]pyridin-4-amine 6
2.2 mg; 8% yield; LCMS m/z 319.2 (M+H); rt 1.16 min;
1
Conditions A. H NMR (500 MHz, DMSO-d6)  8.74 (d, J=5.4
Hz, 4H), 8.13 (s, 1H), 7.98-7.86 (m, 2H), 7.40 (d, J=3.7 Hz, 1H),
7.06-6.97 (m, 2H), 4.08 (s, 3H).
N-{2-[6-(Trifluoromethyl)pyridin-2-yl]pyrrolo[2,1-
f][1,2,4]triazin-4-yl}pyridin-4-amine 7
Pyridyl substituted intermediate 25: To a scintillation vial was
added 2-chloro-4-phenoxypyrrolo[2,1-f][1,2,4]triazine 24 (2
mmol), 2-substituted-6-(tributylstannyl)pyridine (2.2 mmol),
bis(triphenylphosphine)palladium(II) dichloride (0.10 mmol) and
toluene (5 mL). The resulting reaction mixture was degassed by
bubbling nitrogen through the reaction mixture. The vial was
capped with a pressure-safe septum cap and heated at 90 °C for
18 h. An aliquot was concentrated and the residue was dissolved
in methanol and analyzed by LCMS to ensure completion of
reaction. The reaction mixture was concentrated and the residue
was purified by silica gel chromatography using hexanes:ethyl
acetate to obtain 2-pyridyl intermediate 25. Please see supporting
information for a representative method to access compounds 18-
22.
18.0 mg; 40% yield; LCMS m/z 357.2 (M+H); rt 1.27 min;
1
Conditions C. H NMR (400 MHz, DMSO-d6)  6.90-6.98 (m,
1H), 7.33-7.39 (m, 1H), 8.01-8.12 (m, 2H), 8.15-8.21 (m, 2H),
8.27-8.36 (m, 1H), 8.49-8.59 (m, 3H), 10.31-10.39 (m, 1H). 13
C
NMR (100 MHz, DMSO-d6)  153.86, 151.95, 151.45, 150.09,
146.71, 146.03, 139.64, 125.84, 122.93, 121.35, 120.59, 120.20,
114.92, 113.12, 112.73, 102.58.
3-Fluoro-N-[2-(6-methylpyridin-2-yl)pyrrolo[2,1-
f][1,2,4]triazin-4-yl]pyridin-4-amine 8
18.8 mg; 50% yield; LCMS m/z 321 (M+H); rt 1.32 min;
1
Conditions F. H NMR (400 MHz, DMSO-d6)  8.76 (d, J=3.0
Disubstituted compound 26: To a 1 dram vial was added 2-
substituted-4-phenoxy intermediate 25 (0.066 mmol), substituted
aminopyridine (0.165 mmol), DMF (1 mL) and a 60% dispersion
of sodium hydride (0.132 mmol) in mineral oil. The resulting
reaction mixture was stirred at room temperature for 1 h. An
aliquot was diluted with wet methanol and analyzed by LCMS.
The reaction mixture was quenched with wet DMF and purified
by reverse phase HPLC to obtain compound 26.
Hz, 1H), 8.53 (d, J=5.4 Hz, 1H), 8.34-8.46 (m, 1H), 7.93-8.13
(m, 3H), 7.52 (d, J=7.4 Hz, 1H), 7.39 (br. s, 1H), 6.89-6.99 (m,
1H), 2.64 (s, 3H).
3,5-Difluoro-N-[2-(6-methylpyridin-2-yl)pyrrolo[2,1-
f][1,2,4]triazin-4-yl]pyridin-4-amine 9
10.4 mg; 60% yield; LCMS m/z 339 (M+H); rt 1.57 min;
1
Conditions F. H NMR (400 MHz, DMSO-d6) Broad peaks were
observed.  8.70 (br. s, 2H), 8.06 (br. s, 1H), 7.75 (br. s, 2H),
7.16-7.41 (m, 2H), 6.91 (br. s, 1H), 3.39 (s, 1H), 2.54 (br. s, 3H).
5.2.2 Analytical data for final compounds:
N-[2-(6-Methylpyridin-2-yl)pyrrolo[2,1-f][1,2,4]triazin-4-
yl]quinolin-4-amine 3
3.3 mg; 14% yield; LCMS m/z 353 (M+H); rt 1.67 min;
Conditions F; 1H NMR (400 MHz, DMSO-d6) d: 8.96 (br. s.,
1H), 8.30 (d, J=8.1 Hz, 1H), 8.18 (br. s., 1H), 8.09 (d, J=8.4 Hz,
1H), 8.03 (br. s., 1H), 7.81-7.90 (m, 2H), 7.61-7.81 (m, 2H),
7.22-7.55 (m, 2H), 6.90 (br. s., 1H), 2.55 (s, 3H).
4-{[2-(6-Methylpyridin-2-yl)pyrrolo[2,1-f][1,2,4]triazin-4-
yl]amino}-N-(propan-2-yl)pyridine-3-carboxamide 10
9.0 mg; 25% yield; LCMS m/z 388 (M+H); rt 1.70 min;
1
Conditions F. H NMR (400 MHz, DMSO-d6)  9.16 (br. s, 1H),
8.93-9.07 (m, 2H), 8.73 (d, J=5.8 Hz, 1H), 8.15 (d, J=7.7 Hz,
1H), 8.09 (br. s, 1H), 7.93 (t, J=7.7 Hz, 1H), 7.45 (d, J=7.7 Hz,
1H), 6.89-7.02 (m, 2H), 4.12-4.29 (m, J=13.3, 6.6, 6.6, 6.6 Hz,
1H), 2.62 (s, 3H), 1.22 (d, J=6.6 Hz, 6H).
N-[2-(6-Methylpyridin-2-yl)pyrrolo[2,1-f][1,2,4]triazin-4-
yl]pyridin-4-amine 4
12.8 mg; 52% yield; LCMS m/z 303 (M+H); rt 1.22 min;
1
Conditions B. H NMR (400 MHz, DMSO-d6)  8.46-8.64 (m,
3-Fluoro-N-[2-(pyridin-2-yl)pyrrolo[2,1-f][1,2,4]triazin-4-
yl]pyridin-4-amine 11
J=6.4 Hz, 2H), 8.07-8.22 (m, 3H), 7.93-8.06 (m, 2H), 7.89 (t,
J=7.7 Hz, 1H), 7.40 (d, J=7.4 Hz, 1H), 7.32 (d, J=3.7 Hz, 1H),
6.91 (dd, J=4.4, 2.7 Hz, 1H), 2.61 (s, 3H).
33.0 mg; 39% yield; LCMS m/z 307 (M+H); rt 1.64 min;
1
Conditions E. H NMR: (400 MHz, DMSO-d6)  8.73-8.70 (m,
1H), 8.65 (d, J=3.0 Hz, 1H), 8.46 (d, J=5.5 Hz, 1H), 8.30-8.23
(m, 1H), 8.14 (s, 1H), 8.01-7.92 (m, 2H), 7.51-7.47 (m, 1H),
7.35-7.31 (m, 1H), 6.91-6.86 (m, 1H).
N-[2-(Pyridin-2-yl)pyrrolo[2,1-f][1,2,4]triazin-4-yl]pyridin-4-
amine 5
13.4 mg; 89% yield; LCMS m/z 289 (M+H); rt 1.16 min;
Conditions B. 1H NMR (400 MHz, DMSO-d6)  10.31 (s, 1H),
8.76 (d, J=4.4 Hz, 1H), 8.54 (d, J=5.7 Hz, 2H), 8.28 (d, J=7.7
Hz, 1H), 8.13 (d, J=6.1 Hz, 2H), 7.94-8.06 (m, 2H), 7.54 (dd,
J=7.1, 5.0 Hz, 1H), 7.33 (d, J=3.4 Hz, 1H), 6.91 (dd, J=4.0, 2.7
Hz, 1H).
(1R,4R)-1-N-({4-[(3-Fluoropyridin-4-yl)amino]-2-(pyridin-2-
yl)pyrrolo[2,1-f][1,2,4]triazin-5-yl}methyl)cyclohexane-1,4-
diamine 12
3.0 mg; 14% yield; LCMS m/z 433.2 (M+H); rt 1.32 min;
1
Conditions C. H NMR (400 MHz, DMSO-d6)  9.17-9.09 (m,
1H), 9.02-8.97 (m, 1H), 8.84-8.76 (m, 1H), 8.66-8.59 (m, 1H),
8.52-8.45 (m, 1H), 8.31-8.19 (m, 1H), 8.13-8.06 (m, 1H), 7.84-

7.72 (m, 1H), 7.10-7.00 (m, 1H), 4.42-4.32 (m, 2H), 2.22-2.13
(m, 2H), 2.09 (s, 4H), 2.05-1.93 (m, 2H), 1.56-1.40 (m, 3H),
1.32-1.16 (m, 2H).
NMR (100 MHz, DMSO-d6) 168.84, 153.85, 151.92, 151.72,
151.34, 148.44 (F), 147.81 (F), 139.63, 126.69, 122.94, 121.40,
120.68, 113.13 (F), 112.75 (F), 111.40, 103.50 (F), 102.80 (F),
61.11, 53.13, 52.49.
(3S,4S)-4-Amino-1-({4-[(3-fluoropyridin-4-yl)amino]-2-
(pyridin-2-yl)pyrrolo[2,1-f][1,2,4]triazin-5-yl}methyl)piperidin-
3-ol 13
N-(3-Fluoropyridin-4-yl)-2-(6-methylpyridin-2-yl)-9H-purin-
6-amine 18
4.0 mg; 14% yield; LCMS m/z 435.2 (M+H); rt 1.21 min;
Conditions C. H NMR (400 MHz, DMSO-d6)  12.23 (br. s.,
26.3 mg; 67% yield; LCMS: m/z 322.2 (M+H); rt 1.22 min;
Conditions E. H NMR (400 MHz, DMSO-d6)  13.33 (br. s.,
1
1
1H), 8.83 (dd, J=7.0, 5.5 Hz, 1H), 8.75 (ddd, J=4.8, 1.83, 0.98
Hz, 1H), 8.63 (d, J=2.9 Hz, 1H), 8.46 (d, J=5.4 Hz, 1H), 8.23 (dt,
J=7.9, 1.07 Hz, 1H), 7.99 (td, J=7.7, 1.7 Hz, 1H), 7.95 (d, J=2.7
Hz, 1H), 7.52 (ddd, J=7.5, 4.83, 1.1 Hz, 1H), 6.84 (d, J=2.7 Hz,
1H), 4.92 (br. s., 1H), 3.93 (d, J=13.7 Hz, 1H), 3.82 (d, J=13.7
Hz, 1H), 3.16 (d, J=5.9 Hz, 1H), 2.97 (d, J=11.7 Hz, 2H), 2.42
(br. s., 1H), 2.12 (s, 1H), 1.87 (br. s, 1H), 1.75 (d, J=10.3 Hz,
1H), 1.30-1.42 (m, 1H).
1H), 9.64 (br. s., 1H), 8.72 (br. s., 1H), 8.57 (d, J=3.0 Hz, 1H),
8.47 (s, 1H), 8.39-8.43 (m, 1H), 8.15 (d, J=7.5 Hz, 1H), 7.83 (t,
J=7.5 Hz, 1H), 7.35 (d, J=7.6 Hz, 1H), 2.60 (s, 3H). 13C NMR
(100 MHz, DMSO-d6) 158.125, 155.849, 150.979, 149.774,
147.652, 146.856, 146.638, 145.810, 141.75 (F), 141.66(F),
139.620, 130.65 (F), 130.33(F), 128.567, 123.284, 116.695,
116.108, 20.246 (tautomerism).
3-Fluoro-N-[2-(6-methylpyridin-2-yl)-7H-pyrrolo[2,3-
d]pyrimidin-4-yl]pyridin-4-amine 19
(3R,4R)-4-Amino-1-({4-[(3-fluoropyridin-4-yl)amino]-2-
(pyridin-2-yl)pyrrolo[2,1-f][1,2,4]triazin-5-yl}methyl)piperidin-
3-ol 14
110 mg; 37% yield; LCMS m/z 321.2,(M+H); rt 2.26 min;
Conditions E; 1H NMR (400 MHz, DMSO-d6)  12.13 (br. s.,
1H), 9.63-9.55 (m, 1H), 8.80 (s, 1H), 8.58-8.52 (m, 1H), 8.43-
8.32 (m, 1H), 8.14 (s, 1H), 7.81 (t, J=7.8 Hz, 1H), 7.47-7.39 (m,
1H), 7.32 (s, 1H), 7.02-6.95 (m, 1H), 2.59 (s, 3H).
8.5 mg; 6% yield; LCMS m/z 435.2 (M+H); rt 1.68 min;
1
Conditions E. H NMR (400 MHz, DMSO-d6)  12.34-12.16 (m,
1H), 8.91-8.81 (m, 1H), 8.75-8.71 (m, 1H), 8.63 (s, 1H), 8.53-
8.40 (m, 1H), 8.34-8.19 (m, 1H), 8.02-7.90 (m, 2H), 7.65-7.48
(m, 1H), 6.88-6.80 (m, 1H), 4.96-4.82 (m, 1H), 4.00-3.89 (m,
1H), 3.84-3.76 (m, 1H), 3.17-3.08 (m, 2H), 3.01-2.92 (m, 2H),
2.73-2.64 (m, 1H), 2.35-2.28 (m, 1H), 2.21-1.98 (m, 1H), 1.88 (s,
1H), 1.80-1.68 (m, 1H), 1.46-1.29 (m, 1H). 13C NMR (100 MHz,
DMSO-d6) 150.11, 137.58, 125.24, 123.28, 119.78, 117.76,
114.91, 114.63, 111.78, 21.74.
N-(4-((2-(6-(Trifluoromethyl)pyridin-2-yl)-9H-purin-6-
yl)amino)pyridin-2-yl)acetamide 20
18.0 mg; 42% yield; LCMS: m/z 415.0 (M+H); rt 1.19 min;
1
Conditions D. H NMR (400 MHz, DMSO-d6)  13.58 (br. s.,
1H), 10.47 (s, 1H), 10.34 (s, 1H), (9.25, m, 2H), 8.47 (s, 1H),
8.23 (t, J= 8.0 Hz, 1H), 8.16 (d, J=5.4 Hz, 1H), 8.0 (d, J=7.2 Hz,
1H), 7.87 (d, J=4 Hz, 1H), 2.40 (s, 3H).
N-(4-((2-(6-(Trifluoromethyl)pyridin-2-yl)pyrrolo[2,1-
f][1,2,4]triazin-4-yl)amino)pyridin-2-yl)acetamide 15
325 mg; 30% yield; LCMS m/z 414.2 (M+H); rt 2.40 min;
Conditions E. 1H NMR (400 MHz, DMSO-d6)  10.45 (s, 1H),
10.37 (s, 1H), 8.73-8.84 (m, 2H), 8.22-8.31 (m, 2H), 8.07-8.12
(m, 2H), 8.00-8.05 (m, 1H), 7.39-7.44 (m, 1H), 6.90-6.94 (m,
1H), 2.15 (s, 3H).
N-[4-({2-[6-(Trifluoromethyl)pyridin-2-yl]-7H-pyrrolo[2,3-
d]pyrimidin-4-yl}amino)pyridin-2-yl]acetamide 21
46.8 mg; 34% yield; LCMS m/z 414.2 (M+H); rt 2.02 min;
Conditions E; 1H NMR NMR (400 MHz, DMSO-d6) δ 12.28 (br.
s., 1H), 10.36 (s, 1H), 9.99 (s, 1H), 8.87 (d, J=8.0 Hz, 1H), 8.67
(d, J=2.0 Hz, 1H), 8.28-8.14 (m, 3H), 7.97 (d, J=7.0 Hz, 1H),
7.47 (d, J=3.5 Hz, 1H), 7.01 (d, J=3.5 Hz, 1H), 2.14 (s, 3H).
N-[5-Fluoro-4-({2-[6-(trifluoromethyl)pyridin-2-
yl]pyrrolo[2,1-f][1,2,4]triazin-4-yl}amino)pyridin-2-yl]acetamide
16
N-(3-Fluoropyridin-4-yl)-2-(6-(trifluoromethyl)pyridin-2-yl)-
7H-pyrrolo[2,3-d]pyrimidin-4-amine 22
47.0 mg; 27% yield; LCMS: m/z, 432.0 (M+H); rt 2.56 min;
Conditions E. 1H NMR (400 MHz, DMSO-d6)  10.61 (s, 1H),
10.30 (br. s., 1H), 8.93 (d, J=6.0 Hz, 1H), 8.68 (d, J=8.0 Hz, 1H),
8.40 (d, J=2.5 Hz, 1H), 8.21 (t, J=8.0 Hz, 1H), 8.13 (d, J=1.5 Hz,
1H), 8.05-7.96 (m, 1H), 7.38 (d, J=3.0 Hz, 1H), 6.93 (dd, J=4.3,
2.8 Hz, 1H), 2.17-2.12 (m, 3H).
45.7 mg; 43% yield; LCMS m/z 375.2 (M+H); rt 1.26 min;
Conditions E;1H NMR (400 MHz, DMSO-d6)  12.38 (br. s.,
1H), 9.97 (br. s., 1H), 8.95 (br. s., 1H), 8.68 (br. s., 2H), 8.40
(d, J=6.0 Hz, 1H), 8.26 (s, 1H), 7.98 (s, 1H), 7.53 (s, 1H), 7.29-
6.71 (m, 1H). 13C NMR (100 MHz, DMSO-d6) 155.94, 153.73,
152.85, 151.74, 151.16 (F), 148.66 (F), 146.70 (F), 146.36 (F),
143.03, 139.40, 138.27 (F), 138.19 (F), 134.48 (F), 134.22 (F),
125.94 (F), 125.82 (F), 123.04, 120.78, 120.31, 116.52, 105.23,
99.76.
2-[4-(2-Hydroxyethyl)piperazin-1-yl]-N-[4-({2-[6-
(trifluoromethyl)pyridin-2-yl]pyrrolo[2,1-f][1,2,4]triazin-4-
yl}amino)pyridin-2-yl]acetamide 17
20.0 mg; 21% yield; LCMS m/z 542.2 (M+H); rt 1.80 min;
Conditions C. 1H NMR (400 MHz, DMSO-d6) δ 9.89 (s, 1H),
8.82-8.88 (m, 2H), 8.26-8.30 (m, 2H), 8.03-8.11 (m, 3H), 7.41-
7.43 (m, 1H), 6.92-6.94 (m, 1H), 3.51 (t, J=16.0 Hz, 2H), 3.23 (s,
5.3 Crystallographic Methods:
Protein for TGFRI-T204D crystals in complex with inhibitor
22 was prepared as previously described.15 Data for complex with
3H), 2.58-2.61 (m, 4H), 2.41-2.49 (m, 2H), 1.90 (s, 4H). 13
C

10. Kieltyka, K.; Zhang, J.; Li, S.; Vath, M.; Baglieri, C.; Ferraro, C.;
Zvyaga, T. A.; Drexler, D. M.; Weller, H. N.; Shou, W. Z. Rapid
Commun. Mass Spectrom. 2009, 11, 1579-1591.
11. Shah, M. B.; Pascual, J.; Zhang, Q.; Stout, C. D.; Halpert, J. R.
Molecular Pharmacology 2011, 80, 1047-1055.
12. Paiva, A.A.; Klakouski, C.; Li, S. et al. Bioanalysis 2017, 9, 541-
552.
13. Li, H.; McMillen, W. T.; Charles R. Heap, C. R.; Denis J.
McCann, D. J. et al., J. Med. Chem. 2008, 51, 2302-2306.
14. Mastalerz, H.; Wittman, M. D.; Zimmermann, K. et al. PCT Int.
Appl. (2008), WO 2008005956 A2 20080110
inhibitor 22 were collected in the laboratory using a Rigaku FR-E
rotating anode generator, MicroMax Confocal optics, and a
Saturn92 detector. Raw data were processed with XDS16,17 and
scaled with AIMLESS.18 The atomic coordinates of human
TGFRI kinase domain (PDB ID 3TZM) were used as a search
model for rigid body only refinement in AmoRe.19-20 Refinement
was carried out with BUSTER/TNT21 (GlobalPhasing, Ltd.,
Cambridge, UK), GRADE (GlobalPhasing, Ltd., Cambridge,
UK) for the inhibitor restraint dictionary, and COOT22 for
modeling into electron density.
15. Tebben, A. J.; Ruzanov, M.; Gao, M.; Xie, D.; Kiefer, S. E.; Yan,
C.; Newitt, J. A.; Zhang, L.; Kim, K.; Lu, H.; Kopcho, L. M.;
Sheriff, S. Acta Crystallogr. Sect. D. Struct. Biol. 2016, 72, 658-
674.
16. Kabsch, W. Acta Crystallogr. Sect. D Biol. Crystallogr. 2010, 66,
125-132.
17. Kabsch, W. Acta Crystallogr. Sect. D Biol. Crystallogr. 2010, 66,
133-144.
Acknowledgments
18. Evans, P.R.; Murshudov, G.N. Acta Crystallogr. Sect. D Biol.
Crystallogr. 2013, 69, 1204-1214.
19. Navaza, J. Acta Crystallogr. Sect. A 1994, 50, 157-163.
20. Collaboration Computational Project, Number 4. The CCP4
Suite: programs for protein crystallography. Acta Crystallogr
Sect. D 1994, 50, 760-763.
21. Blanc, E.; Roversi, P.; Vonrhein, C.; Flensburg, C.; Lea, S. M.;
Bricogne, G. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2004,
60, 2210-2221.
22. Emsley, P.; Lokhamp, B.; Scott, W. G.; Cowtan, K. Features and
development of Coot. Acta Crystallogr., Sect. D: Biol.
Crystallogr. 2010, 66, 486-501.
The authors acknowledge Karen Augustine, Karen Parrish,
Simone Oppenheimer, Ernest Tai, Marina Gelman, Wilson Shou,
Anthony Paiva, Cheryl Klakouski, Tatyana Zvyaga, John Herbst,
Singamsetti Kumar, Sreekanth Midde, Amrita Roy, Manoranjan
Panda, Sylwia Stachura, Cullen Cavallaro, Hao Lu, Ramakanth
Sarabu, Richard Rampulla, Murugesan Natesan, Arvind Mathur,
Chunhong Yan, Susan E. Kiefer, Dianlin Xie, Single Compound
Purification group, in-vivo group and Department of Chemical
Synthesis for their contributions toward the generation of data
disclosed in this manuscript.
23. Harikrishnan, L. S.; Fink, B. E.; Borzilleri, R. M. et al. PCT Int.
Appl. 2017, WO 2017019757 A1 20170202.
24. Harikrishnan, L. S.; Fink, B. E.; Borzilleri, R. M. et al. PCT Int.
Appl. 2017, WO 2017015425 A1 20170126.
References and notes
1.
2.
3.
4.
5.
Pickup, M.; Novitskiy, S.; Moses, H. L. Nature Rev Cancer 2013,
13, 788-799.
Uhl, M.; Aulwurm, S.; Wischhusen, J.; Weiler, M. et al. Cancer
Res. 2004, 64, 7954-7961.
Kim, S.; Buchlis, G.; Fridlender, Z. G.; Sun, J. et al. Cancer Res.
2008, 68, 10247-10256.
Akhurst, R. J.; Hata, A. Nature Rev. Drug Discov. 2012, 11, 790-
811.
Li, H.; Wang, Y.; Heap, C. R, et al. J. Med. Chem. 2006, 49,
2138-2142.
Shi, Y.; Massague, J. Cell 2003, 113, 685-700.
See experimental section for a description of HTRF assays.
See experimental section for a description of SMAD cellular
assay.
For examples of 1,3-disubstituted six membered rings, see: Dugar,
S.; Chakravarty, S.; Murphy, A.; Mcenroe, G. et al. PCT Int. Appl.
2005, WO 2005032481 A2 20050414.
25. Borzilleri, R. M.; Fink, B. E.; Harikrishnan, L. S. et al. PCT Int.
Appl. 2017, WO 2017040448 A1 20170309.
26. Borzilleri, R. M.; Fink, B. E.; Harikrishnan, L. S. et al. PCT Int.
Appl. 2017, WO 2017035118 A1 20170302.
Supplementary Material
Supplementary material for this article can be found in the
online version at http://dx.doi.org/ (BMC to fill up this
information). X-ray cocrystal structure of compound 22 in
TGFRI can be found in the PDB (pdb id 6B8Y). Detailed
experimental procedures and characterization data can also be
found in the recently published patent applications.23-26
6.
7.
8.
9.

Supporting Information:
Title: Heterobicyclic inhibitors of transforming
growth factor beta receptor I (TGFRI)
Contents:
1. Analytical HPLC Conditions A – F (page 2)
2. Figure 1S and 2S: Small molecule X-ray structure of a purine compound related to
analogs 18 - 22 showing intermolecular hydrogen bonds that might rationalize the poor
solubility of this series (CCDC deposition number 1579603) (pages 3-4)
3. Crystallographic statistics for compound 22 (page 5)
4. Scheme 1S: A representative procedure for the synthesis of purines and
pyrrolopyrimidines 18 to 22 (pages 6-8)
12

1. Analytical HPLC Conditions A – F:
A. Waters Acquity UPLC BEH C18 (2.1 x 50 mm), 1.7 micron; Solvent A: 5:95
acetonitrile:water with 0.1% trifluoroacetic acid; Solvent B: 95:5 acetonitrile:water with 0.1%
trifluoroacetic acid; Temperature: 50 °C; Gradient: 0-100% B over 3 minutes, then a 0.75-minute
hold at 100% B; Flow: 1.11 mL/min; Detection: UV at 220 nm.
B: Waters Acquity BEH C18 (2.1 x 50 mm) 1.7 micron; Buffer:5 mM AmmoniumAcetate pH
3.5, Solvent A:Buffer:ACN (95:5), Solvent B:Buffer:ACN (5:95), Method:%B: 0 min-
5%:1.1min -95%:1.7min-95%, Flow: 0.8 mL/min.
C: Ascentis Express C18 (2.1x50 mm), 2.7 micron; Solvent A: 5:95 acetonitrile: water with 10
mM NH4OAc; Solvent B: 95:5 acetonitrile: water with 10 mM NH4OAc; Temperature: 50 °C;
Gradient:0-100% B over 3 minutes; Flow rate: 1.1 mL/min; Detection: UV at 220 nm.
D: Column: Ascentis Express C18 (50x2.1) mm, 2.7 micron; Solvent A: 5:95
Acetonitrile:water with 0.1% TFA; Solvent B: 95:5 Acetonitrile:water with 0.1% TFA;
Temperature: 50 °C; Gradient:0-100%B over 3 minutes; Flow:1.1 mL/min.
E: Kinetex XB-C18 (75x3 mm) 2.6 micron; Solvent A: 10 mM ammonium formate in water:
acetonitrile (98:02); Mobile Phase B: 10 mM ammonium formate in water:acetonitrile (02:98);
Temperature: 50 °C; Gradient: 0-100% B over 3 minutes; Flow rate: 1.1 mL/min; Detection: UV
at 220 nm.
F: Waters Acquity UPLC BEH C18 (2.1 x 50 mm), 1.7 micron; Solvent A: 5:95
acetonitrile:water with 10 mM ammonium acetate; Solvent B: 95:5 acetonitrile:water with 10
mM ammonium acetate; Temperature: 50 °C; Gradient: 0-100% B over 3 minutes, then a 0.75-
minute hold at 100% B; Flow: 1.11 mL/min; Detection: UV at 220 nm.
13

2. Crystal structure data for compound 27:
Experimental
Single crystals of compound 27 (C16H10F3N7), structurally related to compounds 18 - 22 in the manuscript,
were obtained from DMSO solution by solvent evaporation at 60 °C. A suitable crystal was selected and mounted
on a nylon cryoloop with paratone oil on Bruker APEX-II CCD diffractometer. Data collection was carried out at
room temperature, 296 (2) K. Using Olex2 [1], the structure was solved with the ShelXS [2] structure solution
program using Direct Methods and refined with the XL [3] refinement package using Least Squares minimization.
1. Dolomanov, O.V.; Bourhis, L.J.; Gildea, R.J.; Howard, J.A.K.; Puschmann, H. (2009), J. Appl. Cryst. 42,
339-341.
2. Sheldrick, G.M. (2008). Acta Cryst. A64, 112-122.
3. Sheldrick, G.M. (2008). Acta Cryst. A64, 112-122.
Crystal data for compound 27: Mol. Formula: C16H10F3N7 (M =357.31 g/mol); monoclinic, space group
P21/n (no. 14), a = 4.4115(19) Å, b = 26.644(11) Å, c = 12.556(5) Å, β = 92.861(18)°, V = 1474.0(10) Å3, Z = 4, T =
296 K, μ(MoKα) = 0.131 mm-1, Dcalc = 1.610 g/cm3, 3243 reflections measured (3.58° ≤ 2Θ ≤ 49.98°), 2052
unique (Rint = 0.1490, Rsigma = 0.2825) which were used in all calculations. The final R1 was 0.0926 (>2sigma(I)) and
wR2 was 0.2532 (all data).
14

Figure 1S: Molecular structure of a purine compound 27 (ORTEP drawn at 50% probability
ellipsoids)
Figure 2S: Intramolecular and intermolecular interactions observed in crystal structure of a
purine compound 27 showing presence of multiple strong (N-H…N/N-H…F) and weak (C-
H…F) hydrogen bonds that might rationalize the poor solubility of this series of compounds
(CCDC deposition number 1579603).
15

1. Crystallographic Statistics for compound 22:
Compound
Space group
22
P212121
Cell dimensions, Å
Resolution, Å
Unique reflections
Redundancy
Completeness (%)
Rsym
42.1, 77.0, 90.4
42.11-1.65 (1.90-1.65)
36153 (12474)
6.5 (6.5)
100 (100)
0.070 (0.354)
14.6 (4.6)
I/I
CC1/2
0.998 (0.946)
Refinement
Resolution, Å
R-work
R-free
22.92-1.65 (1.70-1.65)
0.179 (0.224)
0.208 (0.237)
0.010
r.m.s.d. bond lengths, Å
1.0
r.m.s.d. bond angles, 
Ramachandran favored and allowed* (%) 100
Ramachandran disallowed* (%)
PDB ID
0
Numbers in parentheses represent the highest resolution shell of data.
* As defined by Laskowski, R. A., MacArthur, M. W., Moss, D. S. & Thornton, J. M. (1993)
PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl.
Crystallogr. 26, 283-291.
16

2. A representative procedure to access purines and pyrrolopyrimidines 18-
22:
Scheme 1S:
2,6-Dichloro-9-(4-methoxybenzyl)-9H-purine 1B
To a solution of commercially available 2,6-dichloro-9H-purine (1.0 g, 5.29 mmol) in
DMF (10 mL) was added potassium carbonate (0.804 g, 5.82 mmol) and 1-(chloromethyl)-4-
methoxybenzene (0.829 g, 5.29 mmol). The resultant reaction mixture was stirred at room
temperature for 16 h. An aliquot of the reaction mixture was analyzed by LCMS to ensure
complete conversion (TLC and LC-MS showed regioisomer formation). The reaction mixture
was quenched with water (20 mL) and extracted with ethyl acetate (100 mL x 2). The organic
phase was washed with brine (25 mL x 2), dried over anhydrous sodium sulphate and evaporated
under reduced pressure to obtain a yellow semisolid, which was purified by silica gel
chromatography using 20-100% ethyl acetate in hexanes to afford the desired compound 2,6-
dichloro-9-(4-methoxybenzyl)-9H-purine (600 mg, 1.94 mmol, 73 % yield) and the regioisomer
2,6-dichloro-7-(4-methoxybenzyl)-7H-purine (150 mg, 0.49 mmol, 18 % yield). LCMS: m/z
309.2 (M+H); rt 2.28 min; conditions E.
2-Chloro-9-(4-methoxybenzyl)-6-phenoxy-9H-purine 1C
To a 100 mL round bottom flask was added of 2,6-dichloro-9-(4-methoxybenzyl)-9H-
purine (4.0 g, 12.94 mmol) in tetrahydrofuran (100 mL) and stirred. To the resulting solution
17

was portionwise added sodium phenolate (2.038 g, 17.55 mmol) and the resulting mixture was
heated at 80 °C for 15 h. An aliquot of the reaction mixture was diluted with methanol and
analyzed by LCMS to ensure complete conversion. The reaction mixture was concentrated and
the residue was suspended in DCM (250 mL), washed with water (25 mL) and brine (25 mL).
The organic phase was dried over anhydrous sodium sulfate and concentrated under reduced
pressure to get 2-chloro-9-(4-methoxybenzyl)-6-phenoxy-9H-purine (4.0 g, 10.91 mmol, 84 %
yield) as a brown solid. LCMS: m/z 367.2; rt 3.59 min; conditions E.
9-(4-Methoxybenzyl)-2-(6-methylpyridin-2-yl)-6-phenoxy-9H-purine 1D
To a 50 mL scintillation vial was added 2-chloro-9-(4-methoxybenzyl)-6-phenoxy-9H-
purine (400 mg, 1.091 mmol), 2-methyl-6-(tributylstannyl)pyridine (417 mg, 1.091 mmol),
tetrakis(triphenylphosphine)palladium(0) (0.126 g, 0.109 mmol) and dioxane (10 mL). The
resulting reaction mixture was degassed by bubbling nitrogen gas through the solution. The vial
was capped with a pressure-safe septum cap and heated at 110 °C for 18 h. An aliquot of the
reaction mixture was analyzed by LCMS to ensure completion of reaction. The reaction mixture
was concentrated and the residue was purified by silica gel chromatography using 30-100% ethyl
acetate in hexanes to afford 9-(4-methoxybenzyl)-2-(6-methylpyridin-2-yl)-6-phenoxy-9H-
purine (330 mg, 0.779 mmol, 71 % yield) as an off white solid. LCMS: m/z 424.2 (M+H); rt
3.86 min; conditions E.
N-(3-Fluoropyridin-4-yl)-9-(4-methoxybenzyl)-2-(6-methylpyridin-2-yl)-9H-purin-6-amine 1E
To a solution of 9-(4-methoxybenzyl)-2-(6-methylpyridin-2-yl)-6-phenoxy-9H-Purine
(500 mg, 1.181 mmol) and 3-fluoropyridin-4-amine (529 mg, 4.72 mmol) in DMF (5 mL) was
added a 60% dispersion of sodium hydride (236 mg, 5.90 mmol) in mineral oil and the mixture
was stirred for 3 h. LCMS indicated completion of reaction. The reaction mixture was quenched
carefully with water (25 mL) and allowed to stand for two hours. The resulting brown
precipitate was filtered and washed with water followed by petroleum ether and dried to afford
N-(3-fluoropyridin-4-yl)-9-(4-methoxybenzyl)-2-(6-methylpyridin-2-yl)-9H-purin-6-amine (400
mg, 0.634 mmol, 53.7 % yield) as a brown solid LCMS: m/z 442.2 (M+H); rt 2.28 min;
conditions E.
N-(3-Fluoropyridin-4-yl)-2-(6-methylpyridin-2-yl)-9H-purin-6-amine 18
18

A solution of N-(3-fluoropyridin-4-yl)-9-(4-methoxybenzyl)-2-(6-methylpyridin-2-yl)-
9H-purin-6-amine (400 mg, 0.906 mmol) in TFA (10 mL) was heated at 80 °C for 15 h. The
reaction mixture was concentrated. The residue was dissolved in methanol and purified by
reverse phase HPLC to obtain purine 19 (110 mg, 0.906 mmol, 37 % yield) as an off white solid.
1
LCMS: m/z 322.2 (M+H); rt 1.22 min; conditions E. H NMR (400 MHz, DMSO-d6)  13.33
(br. s., 1H), 9.64 (br. s., 1H), 8.72 (br. s., 1H), 8.57 (d, J=3.01 Hz, 1H), 8.47 (s, 1H), 8.39 - 8.43
(m, 1H), 8.15 (d, J=7.53 Hz, 1H), 7.83 (t, J=7.53 Hz, 1H), 7.35 (d, J=7.6 Hz, 1H), 2.60 (s, 3H).
Details of synthetic routes to access pyrrolotriazine, purine and pyrrolopyrimidine analogs are
described in WO17015425, WO17019757, WO17035118 and WO17040448 (references 23-26
in manuscript).
19

Graphical abstract
Heterobicyclic inhibitors of transforming
Leave this area blank for abstract info.
growth factor beta receptor I (TGFRI)
Lalgudi S. Harikrishnan*, Jayakumar Warrier, Andrew J. Tebben, Gopikishan Tonukunuru, Sudhakara
R. Madduri, Vishweshwaraiah Baligar, Raju Mannoori, Balaji Seshadri, Hasibur Rahaman, P. N.
Arunachalam, Amol Dikundwar, Brian E. Fink, Joseph Fargnoli, Mark Fereshteh, Yi Fan, Jonathan
Lippy, Ching-Ping Ho, Barri Wautlet, Steven Sheriff, Max Ruzanov and Robert M. Borzilleri.
Department of Chemistry, Bristol-Myers Squibb Company, P.O. Box 4000, Princeton, NJ 08543-4000, USA.
20