A fit for purpose synthesis of Bruton's tyrosine kinase inhibitor GDC-0852

Article abstract of DOI:10.1016/j.tetlet.2020.152447

The development of an expedient synthesis to GDC-0852 (1), a reversible BTK inhibitor drug candidate, is described. The key starting material tricyclic lactam 5 was prepared by an annulation reaction of unprotected piperidine-2-carbaldehyde HCl salt (20) and N-Boc piperidine-2,4-dione 21 in a safe and scalable manner. A highly selective Pd-catalyzed C[sbnd]N coupling of lactam 5 and linker 2a, followed by Suzuki?Miyaura coupling to fragment 8 subsequently provided a direct and convergent access to the penultimate 17. A simple NaBH₄ aldehyde reduction completed the synthesis to GDC-0852 (1) in high yield (54% over 3 steps from 5) and purity (99.0 A% HPLC).

Full text of DOI:10.1016/j.tetlet.2020.152447

Tetrahedron Letters 61 (2020) 152447
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
A fit for purpose synthesis of Bruton’s tyrosine kinase inhibitor
GDC-0852
⇑
Ngiap-Kie Lim , Haiming Zhang, C. Gregory Sowell, Francis Gosselin
Small Molecule Process Chemistry, Genentech Inc., 1 DNA Way, South San Francisco, CA 94080, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The development of an expedient synthesis to GDC-0852 (1), a reversible BTK inhibitor drug candidate, is
described. The key starting material tricyclic lactam 5 was prepared by an annulation reaction of unpro-
tected piperidine-2-carbaldehyde HCl salt (20) and N-Boc piperidine-2,4-dione 21 in a safe and scalable
Received 18 August 2020
Accepted 8 September 2020
Available online 11 September 2020
manner.
A highly selective Pd-catalyzed CAN coupling of lactam 5 and linker 2a, followed by
SuzukiꢀMiyaura coupling to fragment 8 subsequently provided a direct and convergent access to the
penultimate 17. A simple NaBH4 aldehyde reduction completed the synthesis to GDC-0852 (1) in high
yield (54% over 3 steps from 5) and purity (99.0 A% HPLC).
Keywords:
BTK inhibitor
Tetrahydroindolizine
Palladium-catalyzed
Regioselective
CAN coupling
Ó 2020 Elsevier Ltd. All rights reserved.
Introduction
Bruton’s tyrosine kinase (BTK) plays a crucial role in regulating
the development, activation and signaling of B-cells, which are a
subtype of lymphocytes critical to our body’s adaptive immune
system and its defense against pathogens [1,2]. Conversely an
enhanced or overexpression of B-cell receptor signaling can lead
to harmful lymphocytic malignancies or the pathogenesis of an
autoimmune or inflammatory disorder [3]. Recent developments
in this field have led to the discovery of ibrutinib [4] and acalabru-
tinib [5], approved as treatment for chronic lymphocytic leukemia
(CLL) and mantle-cell lymphoma (MCL), as well as several new
clinical drug candidates, such as fenebrutinib [6], evobrutinib [7],
spebrutinib [8], and tirabrutinib [9], to address these aberrant B-
cell development or signaling function.
GDC-0852 (1) is a small molecule drug candidate which targets
inhibition of BTK in a reversible (non-covalent) manner to improve
intended kinase target selectivity (Fig. 1) [2]. Herein we describe an
expedient synthesis of GDC-0852 (1) to enable further develop-
ment of this drug candidate.
The discovery synthesis of GDC-0852 (1) utilized successive Pd-
catalyzed couplings to assemble fragment 5 and then 7 onto a linker
molecule 4 to quickly furnish penultimate substrate 9 (Scheme 1). A
simple saponification of the ester protecting group then reveals
GDC-0852 (1), the target drug substance. This synthetic strategy
serves as a divergent as well as streamlined protocol to multiple
Fig. 1. GDC-0852 (1), a reversible BTK inhibitor drug candidate.
analogs around central linker 4 to support structureꢀactivity rela-
tionship (SAR) studies. However, this approach suffers from pro-
tecting group manipulation, high reactant stoichiometry (3.0
equiv of 4), poor yields in the endgame steps, and the requirement
of silica gel column chromatographic purifications, making it
unsuitable for longer term development.
Another major drawback of the discovery synthesis is the long
and challenging synthesis of fragment 5 [2,10]. The original route
encompasses acetic anhydride mediated cyclization of Morita-Bay-
lis-Hillman product 11 to generate indolizine 12, followed by
selective hydrogenation of the pyridine ring to form 13 [11], instal-
lation of ethylamine chain via Minisci type alkylation/nitrile reduc-
tion and finally base mediated lactam ring closure (Scheme 2).
However, the combination of poor reproducibility, lack of robust-
ness, safety concerns encountered in steps leading to 12 and 14
during scale up and tedious purifications resulted in <4% overall
yield of tricyclic lactam 5.
⇑
Corresponding author.
E-mail address: lim.ngiapkie@gene.com (N.-K. Lim).
https://doi.org/10.1016/j.tetlet.2020.152447
0040-4039/Ó 2020 Elsevier Ltd. All rights reserved.

N.-K. Lim et al.
Tetrahedron Letters 61 (2020) 152447
Scheme 3. Retrosynthetic analysis for assembly of GDC-0852.
Practical supply routes to fragments 2 [2,12] and 7 [13] are
known and were adopted without changes to enable faster devel-
opment for early phase delivery of GDC-0852 (1). We turned our
attention to developing a concise synthesis of heterocycle 5 via
the annulation of unprotected a-amino aldehyde 20 and diketone
21 [14,15]. This approach drastically reduced the number of key
bond forming steps to tricyclic lactam 5 since a more advanced
starting material piperidine-2,4-dione (21) could be utilized to
form ring B convergently (Scheme 4).
Due to the unavailability of piperidine-2-carbaldehyde 20 on
scale, we commenced our synthesis with the oxidation of commer-
cially available 18 using TEMPO/NaOCl condition in DCM/water to
give 19, which was carried directly into the subsequent step. The
Boc-deprotection was performed with HCl in i-PrOAc to facilitate
the isolation of 20 as a crystalline HCl salt in high yield of 87%.
Treatment of 20 with key fragment 21 in DCM and pyrrolidine fur-
nished the desired tricyclic lactam 22, which was unfortunately
not stable under work-up and isolation conditions. The crude reac-
tion mixture was therefore telescoped to the next step, in which
TFA was employed to complete the Boc-deprotection. A crystalliza-
tion condition in ethanol was developed to purify and isolate 5 in
modest yield of 48% over two steps. This four-step synthesis pro-
vided a fit-for-purpose means to prepare multiple kilogram quan-
tities of 5 safely and rapidly to support the material need for the
endgame process.
Scheme 1. Discovery synthetic approach to GDC-0852 (1).
With the key fragments (2, 5 and 7) in hand, we began to eval-
uate the endgame chemistry to assemble GDC-0852 (1). A prelim-
inary proof-of-concept for the direct CAN coupling of 5 to linker 2
was first achieved. The reaction employing 5 mol% Pd2(dba)3/Xant-
phos (1:2) as the catalyst and K2CO3 as the base in 1,4-dioxane pro-
ceeded encouragingly in 99% conversion at 95 °C (Fig. 2). However,
despite high excess of bromide 2 (2.5 equiv), the selectivity of the
Scheme 2. Original synthesis to fragment 5.
Results and discussion
After evaluating of the structure of GDC-0852 (1), we proposed
a retrosynthesis shown in Scheme 3, which would bypass ineffi-
cient alcohol protection/deprotection steps in the discovery route
and shorten the overall synthesis sequence. In this alternate route,
the active pharmaceutical ingredient (API) would be derived from
penultimate aldehyde 16, which could be assembled through a
SuzukiꢀMiyaura coupling of the boronic ester (8) of fragment 7
with bromide 17. Intermediate 17 would be directly generated
through palladium-catalyzed CAN coupling of fragment 5 with lin-
ker 2.
Scheme 4. Large scale preparation of 5 via new synthesis route.
2

N.-K. Lim et al.
Tetrahedron Letters 61 (2020) 152447
conversions (46–47%) were obtained in toluene as well as MeTHF.
On the other hand, the reaction in toluene at 80 °C with 1.2 equiv
linker 2 and 5 mol% catalyst achieved high conversion (97%), but
bis-coupling impurity 23 (>15%) remained stubbornly high
(Table 1, entry 6).
In order to achieve higher selectivity, a greater bias between the
reactivity of the halides on linker 2 and the CAN coupling product
was believed to be essential [19]. This hypothesis was tested and
later validated by the reaction using 2-bromo-6-chloro-4-fluo-
robenzaldehyde 2a. Even when the linker stoichiometry was
reduced to 1.2 equiv, only 0.5% of 23 was observed. Additionally,
the reaction produced excellent isolated yield (93%) on 10.0 g scale
(Table 1, entry 7). Preferential amidation of bromide over chloride
was evident. To our pleasant surprise, 2,6-dichloro-4-fluoroben-
zaldehyde (2b), a symmetrical substrate, could also provide high
selectivity (97:3) and yield (90%) in favor of mono amidation pro-
duct 17a (Table 1, entry 8).
Due to the cost advantage and broader commercial availability,
linker 2b was selected for further development. On implementa-
tion to kilogram scale, the CAN coupling of 5 and 2b provided
17a in excellent selectivity (99:1) and clean reaction profile. Fur-
thermore, crystallization in toluene purged side product 23 to
<0.5% and afforded 17a in 92% isolated yield and 98.0 A% HPLC pur-
ity (Eq. (1)).
Fig. 2. Proof of concept for direct CAN coupling of 5 to linker 2.
desired product 17 versus bis-coupling impurity 23 was poor
(82:18) [16]. As the concentration of 17 grew during the reaction,
the side reaction forming the bis-coupling impurity (<1% initially)
turned into a major competitor due to limited bias of reactivity
between bromides 2 and 17.
We evaluated the Pd catalyzed CAN coupling through a micro-
scale high-throughput experimentation (HTE) approach using a
moderate stoichiometry of 2 (1.5 equiv) and K2CO3 (3.0 equiv) as
base at 70 °C [17] confirmed Xantphos as one of the best ligands
for this transformation (Fig. 3) [18]. Validation experiments on lab-
oratory scale (0.50 g) showed that lower catalyst loading (5 mol%)
in toluene at 80 °C could provide 17 in 97% conversion. A slightly
improved selectivity (13% of bis-coupling impurity 23) was
observed when compared to the starting 1,4-dioxane conditions
(Table 1, entries 1 and 2). However, reduction of linker stoichiom-
etry (to 1.5 equiv) unfortunately led to higher bis-coupling impu-
rity (19% of 23, Table 1, entry 3). Further fine-tuning of
temperature (70 °C) and linker stoichiometry (1.2 equiv) did not
provide better outcome (Table 1, entries 4 and 5). Instead, poor
ð1Þ
Next, we switched our focus to the SuzukiꢀMiyaura coupling to
form penultimate intermediate 16 (Scheme 5). In order to further
streamline the process, we explored and quickly adapted a tele-
scoped borylation/SuzukiꢀMiyaura cross-coupling strategy using
7 and 17a. This protocol circumvented the practice of preparing
and isolating pinacol boronate 8 in a separate step. An active,
dual-purpose catalyst PdCl2[P(t-Bu)2Ph]2 was identified to accom-
modate both the borylation and SuzukiꢀMiyaura reactions in the
same reactor. The telescoped reactions proceeded smoothly on
kilogram scale with only minor impurities being observed. Subse-
quent aqueous work up and crystallization of SuzukiꢀMiyaura
product in ethanol gave 16 in 71% yield and 98.0 A% HPLC purity.
It should be noted that the purifications of 16 and 17a by crystal-
lization prior to the final API step were critical to control the resid-
ual Pd in the isolated API, which also enabled removal of
chromatographic purifications from the endgame process.
The final aldehyde reduction to convert 16 to GDC-0852 API
was subsequently achieved by performing the reaction in metha-
nol with a slow addition of a solution of NaBH4 in aqueous 1.0 N
NaOH. The crude product, after work up, was then treated with
Si-Thiol/Si-TMT to control residual Pd to <20 ppm, followed by
recrystallization in IPA/water to furnish GDC-0852 (1) in 81% yield
and 99.0 A% HPLC purity.
In conclusion, we have developed an enabling process for the
delivery of kilogram quantities of GDC-0852 (1). Key starting mate-
rial 5 was successfully prepared on scale via an efficient annulation
of piperidine-2-carbaldehyde HCl salt (20) and N-Boc piperidine-
2,4-dione (21). The endgame chemistry features a Pd-catalyzed
CAN coupling of tricyclic lactam 5 and dichloride 2b to assemble
key intermediate 17a and a telescoped borylation/SuzukiꢀMiyaura
reaction to furnish the penultimate aldehyde intermediate 16.
NaBH4 reduction of aldehyde 16 completed the synthesis of GDC-
0852 (1) and afforded the API in 54% over 3 steps from 5 and
99.0 A% HPLC purity.
Fig. 3. Subset of HTE results for CAN coupling.
3

N.-K. Lim et al.
Tetrahedron Letters 61 (2020) 152447
Table 1
CAN coupling selectivity of different linker substrates.
entrya
linker (equiv)
solvent / temp (°C)
% conv
23 (%)b
1c
2
3
2 (2.5)
2 (2.5)
2 (1.5)
2 (1.2)
2 (1.2)
2 (1.2)
2a (1.2)
2b (1.2)
dioxane, 95
toluene, 80
toluene, 80
toluene, 70
MeTHF, 70
toluene, 80
toluene, 80
toluene, 80
99
97
100
47
46
97
100 (93)
99 (90)
18
13
19
11
10
19
0.5
2.9
4d
5d
6e
7e
8e
f
f
a
All reactions were conducted by heating 5 (0.50 g, 2.6 mmol), 120, 150 or 250 mol% linker 2, 200 mol% K2CO3, 5 mol% Pd(OAc)2 and 10 mol% Xantphos in solvent (4.0 mL)
at 80 or 95 °C; bNormalized percent value to product by HPLC at 220 nm; cPd2(dba)3 (5 mol%) and 5.0 mL solvent employed; d3.0 mL solvent employed; e150 mol% K2CO3
employed; 1solated yield for 10.0 g reaction.
f
Scheme 5. Improved endgame synthesis to GDC-0852.
[2] J.J. Crawford, W.B. Young, PCT International Application WO 2013/067277,
Declaration of Competing Interest
May 10, 2013.
[3] K.D. Puri, J.A. Di Paolo, M.R. Gold, Int. Rev. Immunol. 32 (2013) 397–427.
The authors declare that they have no known competing finan-
cial interests or personal relationships that could have appeared
to influence the work reported in this paper.
[4] Z. Pan, H. Scheerens, S. Li, B.E. Schultz, P.A. Sprengeler, L.C. Burrill, R.V.
Mendonca, M.D. Sweeney, K.C.K. Scott, P.G. Grothaus, D.A. Jeffery, J.M. Spoerke,
L.A. Honigberg, P.R. Young, S.A. Dalrymple, J.T. Palmer, ChemMedChem 2
(2007) 58–61.
[5] J.C. Byrd, B. Harrington, S. O’Brien, J.A. Jones, A. Schuh, S. Devereux, J. Chaves,
W.G. Wierda, F.T. Awan, J.R. Brown, P. Hillmen, D.M. Stephens, P. Ghia, J.C.
Barrientos, J.M. Pagel, J. Woyach, D. Johnson, J. Huang, X. Wang, A. Kaptein, B.J.
Lannutti, T. Covey, M. Fardis, J. McGreivy, A. Hamdy, W. Rothbaum, R. Izumi, T.
G. Diacovo, A.J. Johnson, R.R. Furman, N. Engl. J. Med. 374 (2016) 323–332.
[6] J.J. Crawford, A.R. Johnson, D.L. Misner, L.D. Belmont, G. Castanedo, R. Choy, M.
Coraggio, L. Dong, C. Eigenbrot, R. Erickson, N. Ghilardi, J. Hau, A. Katewa, P.B.
Kohli, W. Lee, J.W. Lubach, B.S. McKenzie, D.F. Ortwine, L. Schutt, S. Tay, B. Wei,
K. Reif, L. Liu, H. Wong, W.B. Young, J. Med. Chem. 61 (2018) 2227–2245.
[7] R.D. Caldwell, H. Qiu, B.C. Askew, A.T. Bender, N. Brugger, M. Camps, M.
Dhanabal, V. Dutt, T. Eichhorn, A.S. Gardberg, A. Goutopoulos, R. Grenningloh, J.
Head, B. Healey, B.L. Hodous, B.R. Huck, T.L. Johnson, C. Jones, R.C. Jones, I.
Mochalkin, F. Morandi, N. Nguyen, M. Meyring, J.R. Potnick, D.C. Santos, R.
Schmidt, B. Sherer, A. Shutes, K. Urbahns, A.V. Follis, A.A. Wegener, S.C.
Zimmerli, L. Liu-Bujalski, J. Med. Chem. 62 (2019) 7643–7655.
Acknowledgments
The authors thank Anny Tangkilisan, Colin Masui, Theresa Crav-
illion and Scott Savage for assistance with high-throughput exper-
imentations, Tina Nguyen and Drs. Sam Yang, Kenji Kurita and
Chris Crittenden for analytical support.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.tetlet.2020.152447.
[8] E.K. Evans, R. Tester, S. Aslanian, R. Karp, M. Sheets, M.T. Labenski, S.R.
Witowski, H. Lounsbury, P. Chaturvedi, H. Mazdiyasni, Z. Zhu, M. Nacht, M.I.
Freed, R.C. Petter, A. Dubrovsky, J. Singh, W.F. Westlin, J. Pharmacol. Exp. Ther.
346 (2013) 219–228.
References
[1] A.B. Satterthwaite, O.N. Witte, Immunol. Rev. 175 (2000) 120–127.
4

N.-K. Lim et al.
Tetrahedron Letters 61 (2020) 152447
[9] H.S. Walter, S.A. Rule, M.J.S. Dyer, L. Karlin, C. Jones, B. Cazin, P. Quittet, N. Shah,
C.V. Hutchinson, H. Honda, K. Duffy, J. Birkett, V. Jamieson, N. Courtenay-Luck,
T. Yoshizawa, J. Sharpe, T. Ohno, S. Abe, A. Nishimura, G. Cartron, F.
Morschhauser, C. Fegan, G. Salles, Blood 127 (2016) 411–419.
[10] A.J.M. Barbosa, P.A. Blomgren, K.S. Currie, R. Krishnamoorthy, J.E. Kropf, S.H.
Lee, S.A. Mitchell, D. Ortwine, A.C. Schmitt, X. Wang, J. Xu, W. Young, H. Zhang,
Z. Zhao, P.E. Zhichkin, PCT International Application WO 2011/140488,
November 10, 2011.
[11] For other reported synthesis of this intermediate, see: (a) E. Vedejs, F.G. West,
J. Org. Chem. 48 (1983) 4773ꢀ4774. (b) H. Ardill, R. Grigg, J. Malone, V.
Sridharan, W.A. Thomas, Tetrahedron 50 (1994) 5067ꢀ5082. (c) B.V.M.
Teodoro, J.T.M. Correia, F. Coelho, J. Org. Chem. 80 (2015) 2529ꢀ2538. (d) Y.
Xue, X. He, T. Yang, Y. Wang, Z. Liu, G. Zhang, Y. Wang, K. Wang, L. Zhang, L.
Zhang, Eur. J. Med. Chem. 182 (2019) 111618.
[13] H. Zhang, T. Cravillion, N.-K. Lim, Q. Tian, D. Beaudry, J.L. Defreese, A. Fettes, P.
James, D. Linder, S. Malhotra, C. Han, R. Angelaud, F. Gosselin, Org. Process Res.
Dev. 22 (2018) 978–990.
[14] S.S. Capomolla, N.-K. Lim, H. Zhang, Org. Lett. 17 (2015) 3564–3567.
[15] N.-K. Lim, H. Zhang, C. Han, PCT International Application WO 2016/066490,
May 6, 2016.
[16] Previous studies had shown a high excess of linker (>5 equiv) was required to
obtain reasonable selectivity (>95:5) for Buchwald–Hartwig amidation with
similar symmetrical substrates: J. Hong, J.P. Davidson, Q. Jin, G.R. Lee, M.
Matchett, E. O’Brien, M. Welch, B. Bingenheimer, K. Sarma, Org. Process Res.
Dev. 18 (2014) 228ꢀ238.
[17] HTE was performed at 70 °C to accommodate lower boiling solvents.
[18] Other ligands, such as (S)-BINAP, (R)-MeO-BIPHEP, dppf, XPhos, Josiphos and
SPANphos afforded inferior conversion and/or selectivity.
[12] F. Gosselin, S. Lau, C. Nadeau, T. Trinh, P.D. O’Shea, I.W. Davies, J. Org. Chem. 74
(2009) 7790–7797.
[19] The approach to use a differentiated linker was also described and successfully
implemented on scale by Hong and co-workers (ref. 16).
5