Development of an Expedient Process for the Multi-Kilogram Synthesis of Chk1 Inhibitor GDC-0425

Article abstract of DOI:10.1021/acs.oprd.5b00105

A process leading to the multikilogram GMP synthesis of Chk1 inhibitor GDC-0425 (1) was developed. Highlights of the synthesis include protection of the pyrrole ring of a 1,7-diazacarbazole as propyl ethyl ether, an efficient Pd catalyzed cyanation of an aryl chloride, aryl ether formation by SNAr fluoride displacement, and development of a controlled crystallization providing the API with the required polymorphic form. The process delivered high-quality GDC-0425 with low levels of impurities and residual metals in five steps and 31% overall yield.

Full text of DOI:10.1021/acs.oprd.5b00105

Subscriber access provided by NEW YORK UNIV
Full Paper
Development of an Expedient Process for the Multi-
Kilogram Synthesis of Chk1 Inhibitor GDC-0425
Andreas Stumpf, Zhigang Ken Cheng, Brian Wong, Mark Reynolds, Remy
Angelaud, James Ken Girotti, Alan Deese, Christine Gu, and Lewis Gazzard
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.5b00105 • Publication Date (Web): 13 May 2015
Downloaded from http://pubs.acs.org on May 15, 2015
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.
Organic Process Research & Development 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 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Development of an Expedient Process for the Multiꢀ
Kilogram Synthesis of Chk1 Inhibitor GDCꢀ0425
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
,
†,a
†,a
†,a
†,a
†,a
Andreas Stumpf,* Zhigang Ken Cheng, Brian Wong, Mark Reynolds, R†,ec my Angelaud,
†
,b
†,b
†,b
James Girotti, Alan Deese, Christine Gu, Lewis Gazzard
†
a
b
c
Small Molecule Process Chemistry, Small Molecule Analytical Chemistry, Department of
Discovery Chemistry, Genentech, A member of the Roche Group, 1 DNA Way,
South San Francisco, CA 94080, USA
*Corresponding Author:
stumpf.andreas@gene.com.
1
ACS Paragon Plus Environment

Organic Process Research & Development
Page 2 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table of Contents Graphic
OH
N
N
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N
O
O
F
O
NC
Cl
NC
5
1) protection
3 steps
N
N
N
2) cyanation
3) fluoride
N
N
N
N
N
H
displacement
N
H
GDC-0425 (1)
1% overall from 2
99 A% HPLC
2
6e
3
>
2
ACS Paragon Plus Environment

Page 3 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
ABSTRACT
A process leading to the multiꢀkilogram GMP synthesis of Chk1 inhibitor GDC-0425 (1) was
developed. Highlights of the synthesis include protection of the pyrrole ring of a 1,7ꢀ
diazacarbazole as propyl ethyl ether, an efficient Pd catalyzed cyanation of an aryl chloride, aryl
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
ether formation by S
providing the API with the required polymorphic form. The process delivered high quality GDC-
425 with low levels of impurities and residual metals in 5 steps and 31 % overall yield.
NAr fluoride displacement and development of a controlled crystallization
0
KEYWORDS: Carbazole, Cyanation, Propyl Vinyl Ether, Aromatic Fluoride Displacement.
3
ACS Paragon Plus Environment

Organic Process Research & Development
Page 4 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
INTRODUCTION.
GDC-0425 is an orally bioꢀavailable small molecule inhibitor of Chk1 that has been developed
as a treatment for various cancerous malignancies. As part of a recent drug development
program, we required multiꢀkilogram amounts of API to support human clinical studies. The
1
aꢀd
initial discovery chemistry synthesis
involved a multi–step conversion from 2 to GDC-0425.
This route (Scheme 1) provided initial quantities of API but could not be scaled up because it
used undesirable reagents (SEMꢀchloride, NaH, TBAF), high catalyst loadings and required
tedious workup and isolation procedures. We needed to develop a phaseꢀappropriate fitꢀforꢀ
st
purpose 1 generation process that would overcome these issues and allow expeditious scaleꢀup
to multiꢀkilograms of GDC-0425. In addition to chemistry optimization, the removal of residual
heavy metals and the development of a crystallization process for the isolation of the penultimate
API also needed to be addressed. The process development of the end game chemistry will be
discussed in detail.
4
ACS Paragon Plus Environment

Page 5 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Scheme 1. Discovery Route to GDC-0425
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
RESULTS AND DISCUSSION.
Manufacture of the GMP starting materials 2 and 5. GMP starting material 2 was made on
30 kg scale via the synthetic route shown in Scheme 2 with an overall yield of 50 % (see
Supporting Information). Because of time constraints at this stage of development, no major
process development was performed on the synthesis of carbazole 2 and the discovery chemistry
route was implemented with only minor modifications for scale up.
5
ACS Paragon Plus Environment

Organic Process Research & Development
Page 6 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 2. Manufacture of carbazole 2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Amination of bromoꢀpyridine 7 (59 kg) via Buchwald coupling with tꢀbutyl carbamate in
dioxane with 2 mol % Pd(OAc) /Xantphos, provided the Boc protected amino pyridine 8. The
2
major issue for this step was the purification of 8. When crude 8 was used directly in the next
step, a severe emulsion was observed during work up. It was assumed that the remaining
Pd/xantphos, catalyst which was difficult to remove via crystallization, affected the workup.
Thus, a silica plug was applied on scale and the product 8 was eluted with 10 % ethyl acetate in
cyclohexane. 59 kg of 8 were isolated after concentration of the eluents to 2 volumes and
crystallization from heptane in 78 % yield and 91 HPLC A % purity (~9 A % of residual tꢀbutyl
carbamate). Compound 8 was then first deꢀprotonated with nꢀBuLi in THF/TMEDA at ꢀ78 °C,
and then treated with iodine to obtain the iodocarbamate 9. The reaction mixture was acidified
with aqueous 1M HCl and extracted with MTBE, followed by an aqueous sodium thiosulfate
solution wash to remove excess iodine. Iodocarbamate 9 could be isolated after concentration
6
ACS Paragon Plus Environment

Page 7 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
and trituration with ethanol, but 9 % of material was lost in the mother liquors. To increase the
yield, the ethanol trituration was eliminated and the crude suspension of 9 was telescoped
directly into the deꢀprotection of the amino group. Initial deꢀprotection conditions with 37 %
aqueous HCl in MeOH were tried and the HCl salt of 10 could be isolated, however, we
observed that a significant amount of the iodide of 10 was exchanged with chloride. Thus
alternate deꢀprotection conditions were investigated and we were pleased to find that TFA in
DCM at ambient temperature led to clean reaction. Solvent exchange from DCM to dioxane was
then performed, and the HCl salt of 10 was crystallized by addition of 4N HCl in dioxane
followed by trituration from heptane. 59 kg of 10 HCl salt were isolated in 88 % yield over 2
steps and 96 HPLC A % purity. The following Suzuki coupling with 2ꢀfluoroꢀ3ꢀpyridine boronic
acid using Pd(tꢀBu PC H NMe ) Cl (AMPHOS) and excess KF in MeCN to give the biꢀ
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
2
6
4
2 2
2
pyridine intermediate 11 was initially tried directly with the HCl salt of 10 but the reaction
.
worked very slowly. Thus, 10 HCl salt was free based with aqueous K CO and extracted into
2
3
MTBE, followed by a solvent exchange from MTBE to MeCN. Lowering the catalyst loading
o
o
from 5 mol % to 2 mol %, and lowering reaction temperature from the original 90 C to 65 C
resulted in complete conversion of 10 to 11. 40 kg of 11 were isolated by crystallization after
extraction into EtOAc followed by solvent exchange to heptane and addition of MTBE in 90 %
yield and 93 HPLC A %. The amount of loss to the mother liquors was around 4 %.
Final ring closure of 11 to GMP starting material 2 was easily accomplished using 2.5 equiv
LiHMDS in THF. After quench with water, the pH was adjusted to 2~3 with aqueous 1N HCl,
and crude 2 was isolated by filtration, followed by recrystallization from MeCN. The
recrystallized 2 showed high residue on ignition (ROI) (~8 wt %) and necessitated a rework to be
suitable for use in the subsequent GMP step to GDC-0425. Therefore, 2 was reslurried with
7
ACS Paragon Plus Environment

Organic Process Research & Development
Page 8 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
water to remove the inorganic salt residue (presumably potassium carbonate), and finally 30 kg
of 2 were obtained in 81 % yield and 99 HPLC A % purity. We also explored an alternative route
2
to 2 to shorten the lengthy synthetic sequence, but this effort was not successful .
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
GMP starting material 5 was manufactured in two batches via the two synthetic routes shown in
Scheme 3.
Scheme 3. Manufacture of alcohol 5
Initially 5 kg of 5 was made in 83 % yield and 98 % purity using route A. Reduction of ketone
14 to 5 was accomplished with sodium borohydride in ethanol. Route A produced small amounts
1
3
of impurity 15, which was carried through downstream chemistry and later detected in the API .
The impurities formed from 15 were hard to purge, so in order to avoid formation of 15, route B
was implemented where 14 was hydrogenated with Raney Ni in ethanol to form alcohol 5, which
was then converted to the HCl salt. After a solvent exchange from ethanol to acetonitrile and
distillation, the product was isolated by filtration and 9 kg of 5 HCl salt was obtained in 76 %
yield and > 98 % purity. See supporting information for further experimental detail for the
synthesis of GMP starting materials 2 and 5.
8
ACS Paragon Plus Environment

Page 9 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Process development and optimization of the GMP steps.
Scheme 4. Process Development of 1,7-Diazacarbazole 2 Protection.
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
F
F
Cl
Cl
N
N
N
N
N
R
N
H
3
a = SEM (2-(trimethylsilyl)-ethoxy)-methyl)
3b = EM (ethoxymethyl)
c = THP (tetrahydropyran)
d = EE (ethoxyethyl)
3e = PE (propoxyethyl)
2
3
3
We initiated our process development by reꢀexamining the chemistry developed for the
protection of 1,7ꢀdiazacarbazole 2 in which the 2ꢀ(trimethylsilyl)ꢀethoxy)ꢀmethyl (SEM) group
was initially selected in order to avoid dimerization of the 1,7ꢀdiazacarbazole motif during the
3
later fluoride displacement in stage 3 of the synthesis . We reasoned that the use of high cost 2ꢀ
4
(
trimethylsilyl)ꢀethoxymethylchloride and the use of hazardous and unnecessarily strong base
5
(
pKa = ~35) sodium hydride, with subsequent hydrogen gas evolution and the formation of
formaldehyde, a volatile, toxic and carcinogenic side product, was a serious cause of concern in a
scaleꢀup operation. Also, multiple washes, trituration and chromatography were required for the
isolation and purification of the protected product. We initially considered protecting groups that
were similar to SEM in terms of electronegativity and steric environment yet have minimal
potential safety risks.
Table 1. Protection Group Screen for Stage 1
Conversion
Entry Protective Group
Conditions
a
(A%)
Ethoxymethyl
1
1.3 equiv KOtBu, 1.3 equiv
chloroꢀmethylethyl ether, rt, THF
64
(
EM)
Tetrahydropyran
THP)
3.0 equiv dihydropyran, 5 mol%
2
o
0
(
PPTS, 60 C, THF
9
ACS Paragon Plus Environment

Organic Process Research & Development
Page 10 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
8 equiv ethyl vinyl ether, 5 mol
3
4
Ethoxyethyl (EE)
Propoxyethyl (PE)
o
95
86
%
PPTS, 70 C, THF
3.2 equiv propyl vinyl ether, 5
mol % PPTS, 66 °C, THF
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Conversion of 2 to 3(b-e) after 3 h as measured by HPLC area%.
First, ethoxymethyl ether (EM) protection of 2 was examined (Table 1, entry 1). Reacting 2 with
chloroꢀmethylethyl ether led to the major desired product 3b alongside with the formation of a
major impurity (36 A % by HPLC), where alkylation on the pyridine nitrogen had occurred.
Next it was attempted to install the THP group (Table 1, entry 2) by reacting 2 with DHP and
pyridinium pꢀtoluenesulfonic acid (PPTS) in THF but no conversion to the desired protected
product was observed. We then turned our attention to the ethoxy ethyl (EE) protecting group,
6
which traditionally is hydrolyzed by HCl at elevated temperature . Reacting 2 with ethyl vinyl
ether and 5 mol % PPTS in THF at 70 °C in a closed system resulted in up to 95 % conversion to
3d (Table 1, entry 3). However, the degree of conversion was inconsistent, presumably because
of loss of some of the ethyl vinyl ether due to its low boiling point (33 °C). Continuously dosing
a minimum of 18 equiv of ethyl vinyl ether via a syringe pump helped to drive the reaction to
completion, but it was reasoned that using the higher boiling propyl vinyl ether (PVE, 65 °C)
would avoid this problem altogether. Indeed, we found that the use of PVE in the protection
reaction enabled us to lower the amount of ether from 18 to 3 equiv. A solvent screen was then
performed on the reaction (Table 2).
10
ACS Paragon Plus Environment

Page 11 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
a
Table 2 . Selected Solvent Screen for Reaction of 2 with PVE
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
Entry
Solvent
3e
A%)
84
17
(A%)
12
a
(
1
2
3
4
5
6
7
Tetrahydrofuran
2ꢀmethyl THF
Dichloromethane
Ethyl acetate
Acetone
68
53
40
37
36
22
9
8
6
9
4
7
Toluene
Dimethylformamide
a
Conversion of 2 to 3e as measured by HPLC area%. All reactions were
run with the following conditions: 1.4 mmol of 2, 5 mol% PPTS, 3 equiv
propyl vinyl ether, 70°C, 10 vol solvent for 3 h.
Results from this screen indicated that THF was the best solvent and led to complete conversion
(
Table 2, entry 1) of 2 to 3e although with concomitant formation of 12% of an unknown side
7
product which was subsequently isolated and identified as impurity 17 .
Next, the influence of various acids on the activation of PVE and conversion of 2 to 3e was
examined. It was found that pꢀtoluenesulfonic acid (PTSA) was superior to all other acids
examined and resulted in conversion of 2 to 3e even at room temperature while the reaction with
PPTS proceeded only at 70 °C. On the other end of the reactivity spectrum, acetic and citric acid
proved to be completely ineffective catalysts. Further lowering the temperature from 25 °C to 0
°
C with PTSA as catalyst caused the reaction to not proceed at all. We therefore established a
reaction protocol with a reaction temperature of 25 °C.
11
ACS Paragon Plus Environment

Organic Process Research & Development
Page 12 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
We also wanted to understand the impact of the PVE dosing rate on the formation of impurity
1
7. Using a syringe pump, PVE was charged to 2 in THF over 2 hours, but reaction was similar
to the semi batch mode addition, and no impact on the amount of 17 generated was observed.
While a second charge of 0.5 equiv PVE was usually necessary in the lab setting, at the 5
kilogram scale, the reaction went to completion when 3.5 equiv was added in one portion,
possibly due to more effective agitation. Quenching the reaction mixture with aqueous NaHCO3
was necessary to prevent deꢀprotection of 3e caused by residual acid. To avoid a work up
operation, direct isolation of 3e from the reaction mixture by addition of water was attempted but
resulted in the product oiling out. A solvent switch from THF to acetone, methanol or
acetonitrile followed by addition of water also resulted in the formation of oily material. We next
tried to extract 3e from the reaction mixture using EtOAc, but this resulted in an emulsion. At
this point it was hypothesized that polymeric impurities generated from PVE were the cause for
the oiling out and emulsion formation and to remove these impurities, the THF solution was
solvent switched to EtOAc and then treated with a combination of 20 wt. % of charcoal and 1 wt.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
%
of celite. After filtration, the EtOAc solution was solvent exchanged to MeOH, from which,
after further concentration by distillation, product 3e crystallized. At multiꢀkilogram scale, we
substituted MeOH for higher boiling EtOH to more effectively remove EtOAc during the solvent
switch. After further concentration of the reaction mixture by distillation and cooling to 25 °C,
the product 3e crystallized. Addition of water as an antiꢀsolvent was necessary to increase
recovery and 3e was isolated after drying as a free flowing solid in 97 % yield and 80 % purity.
A further recrystallization from MeOH resulted in 5.32 kg of 3e in a 71 % overall yield and with
a purity of 96 A % by HPLC.
12
ACS Paragon Plus Environment

Page 13 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Scheme 5. Process Development of Pd catalyzed Cyanation of 3e.
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
With acceptable protection conditions in hand, we turned our attention to the cyanation of 3e.
The initial chemistry for the palladium catalyzed cyanation of 3a used a high loading of tetrakis
triphenylphosphine palladium(0) (15 wt %) and zinc cyanide (2.4 equiv) adding to overall cost,
waste disposal and heavy metals removal issues. In the work up phase, multiple washes,
trituration of the crude product, and a column chromatography were necessary to provide the
product in acceptable purity. The first step in our optimization was to seek a more efficient
catalyst system that could be applied towards the cyanation of 3e, reducing catalyst loading and
offering better workꢀup conditions.
Several examples of cyanation of aryl halides have been described in the literature, some of
8
which have been performed on multiꢀkilogram scale . Notably, aryl bromides have been
subjected to Pd catalyzed cyanation on large scale using catalytic Pd (dba) , dppf, and Zn(CN) .
2
3
2
This protocol has been applied to the cyanation of aryl chlorides with the incorporation of Zn
1
0f
II
0
powder . In this case Zn metal takes on the role of reducing Pd to Pd rendering the reaction
more robust. This example was of particular interest to us since it describes the cyanation of a
less active aryl chloride as with our substrate 3e. We first performed a solvent screen for the
Pd(PPh ) catalyzed cyanation of 3e to 4e using Zn(CN) and Zn powder (Table 3, entry 1ꢀ5) and
3
4
2
found DMA to be the best solvent in terms of reaction rate and minimal side product formation.
The main side product, 18, resulted from deꢀprotection of 4e during prolonged heating of the
9
reaction mixture . A Pd catalyst screen was then performed in DMA with screening results for
13
ACS Paragon Plus Environment

Organic Process Research & Development
Page 14 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the effects of different Pd precursors, ligands, and catalyst loadings on conversion of 3e to 4e are
shown in Table 5. The Pd(OAc) , dppf, Zn catalyst system (Table 3, entry 15) was found to be
2
superior to all other catalysts examined in terms of conversion and minimal side product
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
10
formation .
Table 3. Pd Catalyst Screen for Cyanation of 3e to 4e
Entry
Catalyst
mol%
catalyst
15
Ligand
Solvent
(vol)
DMA (20)
Time
(h)
18
4e
(A%)
83
18 a
a
(A%)
b,e
1
Pd(PPh3)4
Pd(PPh3)4
Pd(PPh3)4
Pd(PPh3)4
Pd(PPh3)4
N/A
N/A
25
43
56
N/A
N/A
23
21
21
6
b,e
2
15
15
15
15
5
NMP (20)
DMF (20)
Toluene (20)
Dioxane (20)
DMA (20)
DMA (20)
DMA (20)
DMA (20)
DMA (10)
DMA (5)
18
18
18
18
21
18
21
3
57
44
5
b,e
3
N/A
b,e
4
N/A
b,e
5
N/A
3
b,f
g
6
Pd (dba)
Xantphos
Dppf
Dppf
Dppf
Dppf
Dppf
Dppf
77
79
79
94
93
88
96
2
3
b,f
7
Pd (dba)
5
2
3
b,f
8
Pd(OAc)2
Pd(OAc)2
Pd(OAc)2
Pd(OAc)2
Pd(OAc)2
5
c,f
9
5
c,f
10
5
3
7
c,f
11
5
3
12
4
c,f
12
1
DMA (10)
3
a
Conversion of 3e to 4e and 20 as measured by HPLC (area%). Reactions were run with the
b
c
d
following conditions: 0.3 mmol scale. 3 mmol scale. Ratio of ligand to catalyst = 2:1.
e
f
g
Zn(CN) 1.2 equiv, zinc 10 mol %. 120 °C. Zn(CN) 0.6 equiv, zinc 10 mol %. 120 °C. 4,5ꢀ
2
2
bis(diphenylphosphino)ꢀ9,9ꢀdimethylxanthene
14
ACS Paragon Plus Environment

Page 15 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
In order to minimize decomposition of 4e to 18 we investigated different parameters. Reducing
the volume of DMA from 20 vol to 5 vol resulted in an increase of 18 from 6 to 12 A % (Table
3, entries 9ꢀ11). Reduction of catalyst loading from 5% to 1% resulted in a slight decrease of 18
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
9
from 7 to 4 A % (Table 3, entries 10 and 12) . At the several hundred gram scale, complete
conversions (> 99 %) were consistently obtained with 4 mol % of Pd. Based upon these results,
the optimal parameters to ensure a complete conversion were set at 4 mol % Pd(OAc)2, 8 mol %
dppf, 10 mol % Zn and 8 vol DMA. Next, a workꢀup procedure was developed. After reaction
completion, the mixture was filtered over celite to remove residual Zn and water was added to
precipitate the product. The filterability of the solids was markedly improved when water was
added at 40 °C as opposed to addition at 0 °C which resulted in very slow filtration. These solids
were usually obtained with ~90 A % purity by HPLC and required further purification.
Enrichment of impurities in the mother liquors was only found in toluene and MTBE at elevated
temperature (50 °C), but purification of the crude product 4e by hot trituration in MTBE and hot
filtration would be difficult to implement on kilogram scale. Earlier solubility studies showed
poor solubility of product in various solvents (toluene, MTBE, IPAC, MeCN and 2ꢀbutanol), but
good solubility in DMF. For this reason crystallization in DMF was further examined. A good
purity upgrade (90 % to 98 %) in 75 % yield was observed when 4e was crystallized from DMF
o
(2 vol) and isolated by filtration at 5 C. In order to increase the recovery of 4e, the isolation
o
o
temperature was lowered to ꢀ5 C and ꢀ10 C, with isolated yields increased to 90 % and no
negative impact on product purity (98 A %). We also tried to remove residual Pd, Zn and Fe at
this step, but treatment with various metal scavengers (florisil, quadrasil MP and SiꢀThiol; 20 wt
%
each) resulted in significant degradation of product and loss of yield. Therefore, metal
removal was postponed to the next step. The optimized conditions were then applied to the
15
ACS Paragon Plus Environment

Organic Process Research & Development
Page 16 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
o
kilogramꢀscale production, 4e was recrystallized from DMF at ꢀ10 C resulting in 3.5 kg of 4e
being obtained after filtration and drying in 72 % yield and > 99 A % purity by HPLC. The
impurity 18 was completely purged during crystallization from DMF.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 6. Process Development of Stage 3a/3b Penultimate Coupling Step
A major safety concern in the discovery route was the use of NaH to deprotonate the
piperidine alcohol HCl 5. We set out to explore alternative bases and solvents to optimize
fluoride displacement of the electron deficient 4e, which presumably proceeds via a S Ar
N
11
12
mechanism . As a first option we explored K CO and DMF. We found no conversion of 4e to
2
3
6
e even after heating at higher temperatures (75 °C and 120 °C) (Table 4, entries 1 and 2).
However, 70 % conversion was observed when 4 equiv of potassium tꢀbutoxide was used at 25
C (Table 4, entry 3). In order to drive the reaction to completion, the amount of base was
doubled from 4 to 8 equiv, but this caused the formation of a significant amount of impurities
°
(Table 4, entry 4).
Table 4. Base Screen For S Ar Fluoride Displacement of 4e with Piperidine Alcohol HCl 5
N
Entry
Base
CO
CO
KOtBu
Base
equiv)
4
Temp Solvent
6e
(A%)
0
Impurities
o
a
(
( C)
(A%)
0
1
2
3
K
2
3
75
120
25
DMF
DMF
DMF
K
4
4
0
0
0
2
3
70
16
ACS Paragon Plus Environment

Page 17 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
4
5
6
7
8
9
KOtBu
KOtBu
KOtBu
8
25
40
75
75
75
75
0
DMF
DMF
DMF
DMF
THF
NMP
THF
THF
THF
THF
47
77
71
86
28
30
0
32
20
27
12
70
70
85
17
15
4
6
6
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
K
2
CO
3
/KOtBu
2.5/3
KOtBu
6
6
6
6
3
3
KOtBu
LDA
1
0
1
1
NaHMDS
NaHMDS
NaHMDS
25
25
0
83
85
96
b
1
2
3
b
1
a
Conversion to 6e as measured by HPLC. All reactions were run with the
following conditions except noted otherwise: 2.5 equiv 5, reaction time 1 h.
b
Reactions were run with the following conditions: 1.5 equiv 5, reaction time 1 h.
The conversion of 4e to 6e could be improved to 77 % by raising the reaction temperature to >
4
0 °C and simultaneously increasing the amount of potassium tꢀbutoxide from 4 to 6 equiv
(
Table 4, entries 5 and 6), but the reaction still generated a high level of impurities. The loading
of potassium tꢀbutoxide was then further decreased by freeꢀbasing the piperidine alcohol HCl 5
with K CO (2.5 equiv) (Table 4, entry 7), followed by addition of potassium tꢀbutoxide (3
2
3
equiv). Using this approach it was possible to drive the reaction to higher conversion (86 A %),
yet, a considerable amount of impurities continued to be formed (12 A % total). Investigation of
THF and NMP as alternate reaction solvents (Table 4, entries 8 and 9) gave worse results than
DMF. We next examined LDA but to our dismay no product 6e was formed and the reaction
gave a complex mixture. A breakthrough occurred with NaHMDS. When a mixture of 4e and 5
in THF was reacted with 6 equiv of a 1 M solution of NaHMDS in THF at 25 °C (Table 4, entry
1
1) we observed a high conversion (83 %) to 6e. It was further elaborated that complete
17
ACS Paragon Plus Environment

Organic Process Research & Development
Page 18 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
conversion could be obtained with 1.5 equiv of 5 and 3 equiv of NaHMDS (Table 4, entry 12).
While this still resulted in the formation of a large unknown impurity (15 A % by HPLC), by
lowering the reaction temperature from 25 °C to 0 °C we were pleased to find that the impurity
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
level could be reduced to < 4 A %. The excess base was quenched with aqueous NH Cl,
4
followed by a solvent exchange from THF to EtOAc. In order to remove residual Zn, Pd and Fe
from the stage 2 cyanation, an aqueous extraction with trisodium EDTA was employed, followed
by a florisil and thiosilicycle treatment. All attempts to isolate 6e directly from the reaction
mixture failed and therefore it was decided to concentrate the product solution and telescope it
into the deꢀprotection step.
Scheme 7. Process Development of Deprotection of 6e
In the discovery chemistry route, TBAF was needed for removal of the SEM group. Switching
from SEM to the PE group facilitated acidic hydrolysis at elevated temperatures. First,
deprotection of 6e was investigated in 2ꢀpropanol using 37% aqueous HCl at 3 different
strengths (3, 2, 1.5 equiv) (Table 5, entries 1ꢀ3) at 50 °C and it was found that the reaction went
to completion with 3 equiv of HCl after 24 h. In order to increase the rate and decrease reaction
time the temperature was increased to 60 °C and a solvent screen was performed in EtOH,
MeOH, MeCN, MeTHF and THF using 2 equiv of HCl. No crystalline material could be
obtained from THF and MeTHF and they were abandoned. The highest rate of deꢀprotection was
found in MeCN, which was complete after 1 h (Table 5, entry 6). Conversion of 6e to GDC-
18
ACS Paragon Plus Environment

Page 19 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
0
425 in EtOH and MeOH (Table 5, entries 4ꢀ5) proceeded markedly slower than in MeCN. From
these experiments it was also found that deꢀprotection of 6e using HCl in MeCN resulted in a
higher isolated yield and higher HPLC purity of GDC-0425 than in EtOH. Interestingly, the
dihydrochloride salt of GDC-0425 was obtained from the reaction mixture.
Table 5. Hydrolysis of 6e to GDC-0425
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
Entry
Conditions
Acid Time Conversion Impurities
a
(
equiv) (h)
(A%)
(A%)
1
2
3
4
5
2ꢀpropanol, 50 °C
2ꢀpropanol, 50 °C
2ꢀpropanol, 50 °C
EtOH, 60 °C
3
24
24
24
5
100
13
2
91
85
13
12
13
13
1.5
2
100
100
MeOH, 60 °C
2
5
2
1
6
MeCN, 60 °C
100
11
a
Conversion of 6e as measured by HPLC. All reactions
were run with 37% aqueous HCl in 10 vol of solvent.
The optimized protocol was then applied to the kilogramꢀscale hydrolysis of 6e using 37%
aqueous HCl in MeCN. In order to optimize yield, 50 % of the solvent was distilled off and, after
cooling to 25 °C, GDC-0425·2HCl was isolated by filtration in 80 % yield and 97 A % purity.
Process Development of GDC-0425 Freebase. The solubility of GDC-0425 has been
evaluated and found to be poor in most organic solvents. The solubility of GDC-0425 (Figure 1)
in water is pH dependent, with good solubility observed at low pH, but GDC-0425 is practically
insoluble at neutral pH. We made use of the pH solubility dependence of the GDC-0425 in water
for isolation purposes.
19
ACS Paragon Plus Environment

Organic Process Research & Development
Page 20 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 1: pH Solubility of GDC-0425
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
HPLC concentration assays for filtrates from aqueous saturated mixtures
of GDC-0425 at various pH after trituration for 48 h at 37 °C.
.
Because GDC-0425 2HCl is soluble in water, we started out by exploring two solvent systems,
IPAꢀwater and THFꢀwater mixtures. Two inorganic bases, sodium hydroxide and sodium
.
bicarbonate were chosen for further evaluation. First, GDC-0425 2HCl was suspended in a
mixture of water and organic solvent, heated to 55 °C to completely dissolve the salt. The base
was then slowly added while the free base GDC-0425 crystallized out. Filterability of GDC-
0
425 was best for the THFꢀNaHCO combination. Finally, we addressed the issue of particulate
3
filtration and came to the conclusion that the reaction mixture could be filtered after addition of 1
mol equiv of base. A purity upgrade from 99.2 % to > 99.9 % could be achieved. For GMP
production, NaHCO was exchanged with NaOH, in order to avoid foaming of the reaction
3
1
3
mixture from CO evolution. Due to the occurrence of the dimer impurity 16 in the stage 3
2
product during the GMP run, a DMF trituration was performed which purged the impurity to <
.
.
0
.25 % and converted the GDC-0425 2HCl to GDC-0425 HCl. The mono HCl salt required only
20
ACS Paragon Plus Environment

Page 21 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
1
mol equiv of base to convert to freeꢀbase and NaOH was added as a 0.6 M solution in water.
On kilogramꢀscale crude GDC-0425 was isolated in 77.7 % yield and >99 A % purity.
Process Development of Stage 5 Recrystallization of GDC-0425. A polymorph study on the
free base of GDC-0425 was performed and only one crystal form was obtained from multiple
solvents (MeOH, EtOH, 1ꢀBuOH, EtOAc, MIBK). From earlier studies we knew that GDC-
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
0
425 had poor solubility in most solvents at room temperature (Figure 2).
Figure 2: GDC-0425 Solvent Solubility
We evaluated solubility of GDC-0425 in various alcohols at reflux, and 1ꢀBuOH was found to
o
be best in terms of volumetric efficiency (50 vol) due to its high boiling point (118 C). The
o
metastable zone for GDC-0425 in 1ꢀBuOH was determined and 110 C was selected as the
temperature at which seeding would be performed.
Seeding in the metastable zone and crystal growth was then evaluated. For this purpose, GDC-
0
425 was dissolved in 1ꢀBuOH (50 vol) and seed crystals (0.9 % slurried in 10 vol 1ꢀBuOH for 1
o
14
h) were introduced at 105 C. Crystal growth was monitored by a FBRM probe while the
21
ACS Paragon Plus Environment

Organic Process Research & Development
Page 22 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
mixture was slowly cooled to room temperature. Maximal crystal growth with particle size range
o
from 46 nm to 158 nm occurred at 97 C. Crystals with smaller particle size solubilize at this
temperature, opening up the possibility of growing larger crystals at this temperature and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
o
improving filterability. No further change in crystal size was observed below 47 C. Next, drying
time was evaluated in order to ensure drying would be feasible on large scale. The amount of 1ꢀ
o
BuOH after 14 h drying in a vacuum oven at 60 C was determined by GC analysis as 1700 ppm.
1
5
The same sample contained 20 ppm THF, which was well below ICH guidelines . To help us
determine what amount of residual DMF might be trapped during 1ꢀBuOH crystallization, a run
was performed with spiking at 1 % DMF. After crystallization and drying in a vacuum oven at
o
6
0 C for 15 h, no residual DMF was detected by GC analysis. The amount of seed crystals was
evaluated in two experiments. In one run 20 % of seed crystals was added resulting in 600 ppm
of 1ꢀBuOH as opposed to 1700 ppm of 1ꢀBuOH for the other run with 0.9 % seeds (both samples
o
were dried at 60 C for 14 h). Powder XRD analysis of both samples showed better crystallinity
with a higher seed loading. At multiꢀkilogram scale, GDC-0425 was dissolved in 1ꢀbutanol (50
o
o
vol) at 115 C, the solution was then cooled to 110 C, where seeding and aging were performed,
o
followed by slow cooling to 20 C (8 h). 1.95 kg of high purity GDC-0425 (>99 %, with no
single impurity ≥ 0.25 %) was isolated after filtration and drying in 97 % yield.
Conclusion
The optimized kilogram route to GDC-0425 is shown in Scheme 8. Highlights of the synthesis
include (1) an efficient Pd catalyzed cyanation of the aryl chloride 3e; (2) fluoride displacement
of 4e with 1ꢀethylpiperidinꢀ4ꢀol under mild reaction conditions using NaHMDS; and (3)
recrystallization of crude GDC-0425 from 1ꢀbutanol. The process delivered 1.95 kg of the
22
ACS Paragon Plus Environment

Page 23 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
desired crystal form of GDC-0425 as highly pure material (> 99 A% by HPLC) in five steps
with an overall yield of 31 %.
Scheme 8. Optimized Kilogram Synthetic Route to GDC-0425
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
Experimental Section
General. In process methods for (Stages 1ꢀ4). Diluent: Acetonitrile. Mobile Phase A: 0.05 %
Formic Acid/Water. Mobile Phase B: 0.05 % Formic Acid/Acetonitrile. Column: Phenomenex
Onyx Monolithic C18 column, 2x50 mm (CV = 0.157 mL) (Part No. CH0ꢀ8373). Column
Temperature: 35˚C. Detector Wavelength: 220 nm. Injection Volume: 2ꢁL. Flow Rate: 0.785
mL/minute (5 CV/min). Program: 0.0 min 5.0 % B, 0.4 min 5.0% B, 2.4 min 65.0 % B, 2.8 min
95 % B, 3.0 min 95 % B, 3.2 min 5.0 % B, 4.0 min 5.0 % B. Typical Retention Times: 2 (RT
2
.62 min), 3e (RT 3.50 min), 4e (RT 3.30 min), 6e (RT 2.36 min).
-chloro-5-fluoro-9-(1-propoxyethyl)-9H-pyrrolo[2,3-b:5,4-c']dipyridine 3e. To a 100 L
reactor was charged 6ꢀchloroꢀ5ꢀfluoroꢀ9Hꢀpyrrolo[2,3ꢀb:5,4ꢀc']dipyridine 2 (5.00 kg, 22.6 mol,
equiv.), pꢀtoluenesulfonic acid monohydrate (0.22 kg, 1.1 mol, 0.05 equiv.) and
6
1
23
ACS Paragon Plus Environment

Organic Process Research & Development
Page 24 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
tetrahydrofuran (50 L, 10 vol). Propyl vinyl ether (8.80 L, 79.0 mol, 3.5 equiv.) was charged to
o
the reactor over 21 min while maintaining the temperature at 30 C. The reaction was
o
maintained at 25 ± 5 C until HPLC IPC showed 2 to be ≤ 4.0 A % (21 h). The mixture was
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
quenched by addition of aqueous NaHCO (0.25 kg, 3.0 mol, dissolved in 4.75 L purified water)
3
o
o
over ~ 3 min at 25 ± 5 C. The organics were distilled under vacuum at ≤ 55 C (internal
temperature) to ~ 20 L and then solvent exchanged to EtOAc (30 L in 3 x 10 L portions) at ≤ 55
o
C and concentrated to ~ 20 L. To the solution was added EtOAc (20.0 L), charcoal (Darco Gꢀ
60, 100 mesh, 1.00 kg, 20 % wt) and celite 545 (5.00 kg, 1 wt) and the slurry was aged for 19 h.
The mixture was filtered through celite 545 (5.00 kg, 1 wt) on a filter with the filtrate transferred
to a 100 L reactor and the cake washed with EtOAc (40.0 L, 8 vol). The combined organics
were washed twice with aqueous NaHCO (2 x 0.75 kg, 9.0 mol each, dissolved in 14.25 L
3
purified water) and the batch was concentrated by distillation under vacuum to ~ 20 L and then
o
solvent exchanged to EtOH (40 L in portions: 10.0 L, 10.0 L, 20.0 L) at ≤ 55 C (internal
o
temperature) and concentrated to ~ 20 L. The mixture was heated to 60 ± 5 C to ensure
o
dissolution of any solids, cooled to 25 ± 5 C over ~ 5 h and aged for 11 h to precipitate the crude
3e. Purified water (20.0 L, 4 vol) was added to aid recovery and the suspension was aged for 30
min. The crude product was filtered and washed with purified water (10.0 L, 2 vol) (MLs assay:
total volume 55 L, product concentration 1.93 g/L, 1.5 % product loss). The cake was dried in
o
the filter at 50 ± 5 C (jacket temperature) under vacuum for 42 h. The crude product was
o
transferred to a 20 L reactor and MeOH (15.0 L, 3 vol) was added and heated to 60 ± 5 C to
o
ensure dissolution. The mixture was cooled to 0 ± 5 C over 3.5 h, aged for 2 h, and a small
sample of the reaction mixture was filtered. Analysis of the supernatant was used as an IPC to
determine that the product concentration in the supernatant was ≤ 15 wt %. The mixture was
24
ACS Paragon Plus Environment

Page 25 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
o
o
filtered at 0 ± 5 C and washed with cold MeOH (0 ± 5 C; 5.0 L, 1 vol) (MLs assay: 16 product
o
mol%). The cake was dried on the filter under vacuum at 50 ± 5 C (jacket temperature) until
GC showed the MeOH content to be ≤ 1.0 wt % (63 h). The process gave 5.32 kg of 6-chloro-5-
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
fluoro-9-(1-propoxyethyl)-9H-pyrrolo[2,3-b:5,4-c']dipyridine 3e (70.9 % yield; 96 A %) as an
1
offꢀwhite solid. H NMR (600 MHz, CDCl ) δ 8.91 (d, J = 1.3 Hz, 1H), 8.67 (dd, J = 4.8, 1.7 Hz,
3
1H), 8.51 (dd, J = 7.8, 1.7 Hz, 1H), 7.33 (dd, J = 7.8, 4.8 Hz, 1H), 6.58 (q, J = 6.1 Hz, 1H), 3.45
(dt, J = 9.2, 6.4 Hz, 1H), 3.19 (dt, J = 9.3, 6.8 Hz, 1H), 1.82 (d, J = 6.2 Hz, 3H), 1.58 (dtd, J =
13.9, 7.4, 6.5 Hz, 2H), 0.85 (t, J = 7.5 Hz, 3H).
13
C NMR (150 MHz, CDCl ) δ 151.61, 150.58, 149.71, 148.84, 135.04, 135.01, 132.64, 132.62,
3
1
29.94, 129.91, 127.51, 127.40, 117.62, 117.49, 117.06, 111.36, 111.35, 81.57, 70.49, 22.59,
20.80, 10.58.
HRMS calcd for C H ClFN O 308.0960; found [M+H] 308.0962.
+
1
5
15
3
5-fluoro-9-(1-propoxyethyl)-9H-pyrrolo[2,3-b:5,4-c']dipyridine-6-carbonitrile 4e. To a 100
L reactor under nitrogen was charged 6-chloro-5-fluoro-9-(1-propoxyethyl)-9H-pyrrolo[2,3-
b:5,4-c']dipyridine 3e (5.28 kg, 17.2 mol, 1 equiv), zinc cyanide (1.20 kg, 10.3 mol, 0.6 equiv),
zinc powder (0.11 kg, 1.7 mol, 0.1 equiv), palladium acetate (0.15 kg, 1.0 mol, 0.04 equiv), dppf
(0.76 kg, 2.1 mol, 0.08 equiv) and DMA (42.6 L, 8 vol). The mixture was degassed by 3 cycles
o
of vacuum/nitrogen and heated to 120 ± 5 C. HPLC showed 3e to be ≤ 2.0 A % after 4.5 h.
o
The reaction was cooled to 40 ± 5 C, celite 545 (1.59 kg, 0.3 wt) was added and the mixture was
aged for 4 h. The mixture was filtered through celite 545 (1.59 kg, 0.3 wt) on a filter with the
filtrate transferred to a 100 L reactor. The filter cake was washed with DMA (11.0 L, 2 vol) and
o
the combined organics were cooled to 20 ± 5 C. Purified water (38.0 L, 7.1 vol) was added over
o
5
0 min while maintaining the temperature at ≤ 40 C to precipitate the crude product. The slurry
25
ACS Paragon Plus Environment

Organic Process Research & Development
Page 26 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
was aged for 4 h, filtered and the cake washed with purified water (16.0 L, 3 vol). The crude
o
product was dried in the filter at 60 ± 5 C (jacket temperature) under vacuum until Karl Fischer
analysis showed the water content to be 1.0 wt % (88 h). The crude product (5.85 kg) and DMF
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
o
(
10.25 L, 2 vol) were charged to a 20 L reactor, heated to 70 ± 5 C to ensure dissolution, aged
o
o
for 30 min, cooled to 45 ± 5 C over 2 h, aged for 32 min, cooled to 25 ± 5 C over 3 h, and
o
cooled to –10 ± 5 C over 3.5 h. The slurry was aged for 3 h and a small sample of the reaction
mixture was filtered and analysis of the supernatant was used to determine that the product
o
concentration in the supernatant was ≤ 30 wt %. The mixture was filtered at ꢀ10 ± 5 C and the
o
cake was washed with cold DMF (0 ± 5 C; 5.0 L, 0.5 vol) (MLs assay: 26.7 product mol%).
After no more filtrate could be collected from the filter, the cake was dried on the filter at 60 ± 5
o
C (jacket temperature) under vacuum until Karl Fischer analysis showed the water content to be
0.02 wt % (94 h). The process gave 3.49 kg of 5ꢀfluoroꢀ9ꢀ(1ꢀpropoxyethyl)ꢀ9Hꢀpyrrolo[2,3ꢀ
b:5,4ꢀc']dipyridineꢀ6ꢀcarbonitrile 4e (72 % yield, > 99 A %) as a light yellow solid.
1
H NMR (600 MHz, CDCl ) δ 9.22 (d, J = 2.1 Hz, 1H), 8.74 (dd, J = 4.9, 1.6 Hz, 1H), 8.59 (dd,
3
J = 7.8, 1.6 Hz, 1H), 7.44 (dd, J = 7.8, 4.8 Hz, 1H), 6.63 (q, J = 6.1 Hz, 1H), 3.46 (dt, J = 9.3, 6.5
Hz, 1H), 3.22 – 3.14 (m, 1H), 1.84 (d, J = 6.2 Hz, 3H), 1.58 (ddd, J = 14.1, 7.6, 6.7 Hz, 2H),
0.85 (td, J = 7.4, 0.9 Hz, 3H).
1
3
C NMR (150 MHz, CDCl ) δ 158.60, 156.79, 151.03, 150.22, 137.23, 137.19, 133.29, 133.26,
3
132.71, 132.70, 118.15, 115.13, 115.02, 114.25, 114.22, 111.27, 111.18, 82.17, 70.84, 22.55,
1.21, 10.53.
HRMS calcd for C H FN O 299.1303; found [M+H] 299.1302.
2
+
16 15
4
Crude
5-(1-ethyl-piperidin-4-yloxy)-9H-dipyrido[2,3-b-4',3'-d]pyrrole-6-carbonitrile
.
dihydro-chloride crude GDC-0425 2HCl. To a 100 L reactor inerted with nitrogen was charged
26
ACS Paragon Plus Environment

Page 27 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
5
ꢀfluoroꢀ9ꢀ(1ꢀpropoxyethyl)ꢀ9Hꢀpyrrolo[2,3ꢀb:5,4ꢀc']dipyridineꢀ6ꢀcarbonitrile 4e (3.44 kg, 11.5
mol, 1 equiv.), 1ꢀethylpiperidinꢀ4ꢀol hydrochloride (2.93 kg, 17.7 mol, 1.5 equiv.) and THF (17.7
o
L, 5 vol). The mixture was cooled to – 5 ± 5 C and 1.0 M sodium bis(trimethylsilyl)amide in
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
THF (35.2 L, 35.2 mol, 3.1 equiv.) was added over 3.5 h, taking care to maintain the temperature
o
o
at ≤ 5 C. The reaction mixture was warmed to 0 ± 5 C and aged over ~ 15 min and HPLC
showed 4e to be ≤ 2.0 A %. The reaction was quenched by addition of aqueous ammonium
chloride (1.82 kg, 34.0 mol, dissolved in 7.0 L water, 20.6 wt%), agitated for 16 min,
o
concentrated by distillation vacuum at ≤ 55 C (internal temperature) to ~ 20 L, solvent
o
exchanged to EtOAc (28 L in 2 x 10 L and 8 L portions) and concentrated at ≤ 55 C to ~ 20 L.
EtOAc (3.5 L) and purified water (28.0 L) were added and the mixture was agitated for 16 min.
The layers were allowed to separate and the aqueous layer was discarded. The organics were
filtered through celite 545 (2.0 kg, 0.3 wt) on a filter with the filtrate transferred to a 100 L
reactor and the filter cake washed with EtOAc (3.5 L, 1 vol). The combined organics were
washed twice with an aqueous EDTA solution (2 x 1.16 kg EDTA, 2.9 mol each, dissolved in 7.0
L purified water). Florisil (7.0 kg, 2 wt) and Silicycle (SiꢀThiol, 1.4 kg, 0.4 wt) were added to
the organics and the slurry was aged for 16 h. The mixture was filtered on a filter with the
filtrate transferred to a 100 L reactor and the cake washed with EtOAc (14 L, 4 vol). A 30 mL
sample of the combined organics was concentrated to an oil 6e.
1
H NMR (600 MHz, CDCl ) δ 9.11 (s, 1H), 8.66 (dd, J = 4.8, 1.6 Hz, 1H), 8.61 (dd, J = 7.8, 1.6
3
Hz, 1H), 7.34 (dd, J = 7.8, 4.8 Hz, 1H), 6.60 (q, J = 6.1 Hz, 1H), 5.01 (dp, J = 9.1, 3.9 Hz, 1H),
3.44 (dt, J = 9.3, 6.5 Hz, 1H), 3.18 (dt, J = 9.2, 6.8 Hz, 1H), 2.99 (dt, J = 9.8, 5.9 Hz, 2H), 2.49
(q, J = 7.2 Hz, 2H), 2.25 (td, J = 11.3, 10.3, 5.2 Hz, 4H), 2.15 – 2.04 (m, 2H), 1.81 (d, J = 6.1
Hz, 3H), 1.57 (p, J = 7.1 Hz, 2H), 1.12 (t, J = 7.2 Hz, 3H), 0.85 (t, J = 7.4 Hz, 3H).
27
ACS Paragon Plus Environment

Organic Process Research & Development
Page 28 of 37
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
13
C NMR (150 MHz, CDCl ) δ 154.33, 150.98, 149.13, 136.32, 132.62, 132.04, 119.96, 117.55,
3
1
17.46, 114.72, 112.93, 81.80, 80.84, 70.65, 52.05, 50.63(2C), 32.13(2C), 22.56, 21.17, 12.20,
0.55.
1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
IPC of the oil by ICPꢀAES was used to determine the content of Pd, Zn and Fe. After it was
determined that all of the specified metals were ≤ 20 ppm (Fe = 8 ppm, Pd ≤ 3 ppm, Zn ≤ 3
o
ppm), the solution was concentrated by distillation under vacuum at ≤ 55 C to ~ 20 L, followed
o
by solvent exchange to MeCN (35 L in 2 x 10 L and 15 L portions) at ≤ 55 C and concentrated
to ~ 20 L. Additional MeCN (21.0 L, 6 vol) and hydrochloric acid (37 % aqueous) (2.32 L, 23
o
mol, 2.0 equiv) were added, maintaining the temperature at 55 ± 5 C. HPLC showed 6e to
be ≤ 1.0 A % after 3 h. The reaction mixture was concentrated by distillation under vacuum at
o
o
≤ 55 C to ~ 23 L. After cooling to 25 ± 5 C over 3 h and aging for 13 h, the slurry was filtered
on a filter and the cake was washed with MeCN (20.0 ±, 5.7 vol) (MLs assay: total volume 40 L,
product concentration 0.285 g/L, 0.25 product mol %). The filter cake was transferred to drying
o
trays and dried in a vacuum oven at 60 ± 5 C for 46 h giving 3.66 kg of crude 5ꢀ(1ꢀethylꢀ
piperidinꢀ4ꢀyloxy)ꢀ9Hꢀdipyrido[2,3ꢀbꢀ4',3'ꢀd]pyrroleꢀ6ꢀcarbonitrile
dihydrochloride
GDC-
.
0425 2HCl (80 % yield, 97 A %) as a light yellow solid.
Crude 5-(1-ethyl-piperidin-4-yloxy)-9H-dipyrido[2,3-b-4',3'-d]pyrrole-6-carbonitrile crude
GDC-0425. To a 100 L reactor inerted with nitrogen was charged crude 5ꢀ(1ꢀethylꢀpiperidinꢀ4ꢀ
.
yloxy)ꢀ9Hꢀdipyrido[2,3ꢀbꢀ4',3'ꢀd]pyrroleꢀ6ꢀcarbonitrile dihydrochloride GDC-0425 2HCl (3.58
o
kg, 9.1 mol, 1 equiv) and DMF (17.5 L, 4.9 vol). The mixture was heated to 100 ± 5 C and
o
agitated for 2 h. It was then cooled to 25 ± 5 C over 6 h and aged for 13 h. The slurry was
filtered on a filter (ID 457 mm) and the cake was washed with DMF (7.0 L, 2 vol) (MLs assay:
total volume 40 L, product concentration 7.46 g/L, 10.5 % loss). The filtered solid was
28
ACS Paragon Plus Environment

Page 29 of 37
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
transferred to a 100 L reactor and DMF (28.0 L, 7.8 vol) was added. The mixture was heated to
o
o
1
00 ± 5 C, aged for 3 h, then cooled to 25 ± 5 C over 6 h and aged for 12 h. The slurry was
filtered on a filter and the filter cake was washed with DMF (7.0 L, 2 vol) and THF (7.0 L, 2 vol)
MLs assay: 1 product mol%). HPLC analysis showed the purity of the cake to be 99.9 A %,
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
(
with no single impurity ≥ 0.25 A %. The cake was transferred to drying trays and dried in a
o
vacuum oven at 60 ± 5 C until a GC analysis showed the DMF content to be 0.05 wt % (65 h).
The solid was transferred to a 100 L reactor (reactor A), dissolved in purified water (17.2 L, 4.8
vol), and filtered through an inline filter (0.5 micron) into a 100 L reactor (reactor B). Reactor A
was rinsed with purified water (3.5 L, 1 vol), followed by THF (17.0 L, 4.8 vol), with both
solutions transferred to reactor B via the inline filter. The water/THF solution was heated to 55 ±
o
5 C and aged for 20 min. Sodium hydroxide (50 % in H O, 1.2 L, 22.7 mol, 2.5 equiv) was
2
diluted with purified water (28.0 L) in reactor A and added in portions through the inline filter to
reactor B. pH (by pH paper) was 10.5, outside the desired range of 8ꢀ10, so hydrochloric acid (37
%
aq) (166 mL) was diluted with purified water (3.84 L) and added in portions (2.5 L: 0.5 L, 1.0
L, 1.0 L) through the inline filter to reactor B. The pH value of the contents in reactor B was
o
measured by pH paper to be 8.3. The mixture in reactor B was cooled to 25 ± 5 C over 2.5 h
and aged for 16 h. It was then filtered and washed with purified water (14.0 L, 4 vol) (MLs
o
assay: 4.6 product mol%). The filter cake was dried in the filter at 60 ± 5 C (jacket temperature)
under vacuum until Karl Fischer analysis showed the water content to be 0.02 wt % (91 h). This
process gave 2.22 kg of 5ꢀ(1ꢀethylꢀpiperidinꢀ4ꢀyloxy)ꢀ9Hꢀdipyrido[2,3ꢀbꢀ4',3'ꢀd]pyrroleꢀ6ꢀ
carbonitrile crude GDC-0425 (77.6 % yield; 99 A %) as an offꢀwhite solid.
5-(1-ethyl-piperidin-4-yloxy)-9H-dipyrido[2,3-b-4',3'-d]pyrrole-6-carbonitrile GDC-0425. To
a 100 L reactor inerted with nitrogen was charged 5ꢀ(1ꢀethylꢀpiperidinꢀ4ꢀyloxy)ꢀ9Hꢀ
29
ACS Paragon Plus Environment