Development of an Efficient Manufacturing Process for Reversible

Subscriber access provided by UNIV OF DURHAM

Full Paper

Development of an Efficient Manufacturing Process for Reversible BTK Inhibitor GDC-0853 Haiming Zhang, Theresa Cravillion, Ngiap-Kie Lim, Qingping Tian, Danial Beaudry, Jessica Defreese, Alec Fettes, Philippe James, David Linder, Sushant Malhotra, Chong Han, Remy Angelaud, and Francis Gosselin Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.8b00134 • Publication Date (Web): 02 Jul 2018 Downloaded from http://pubs.acs.org on July 2, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a 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 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.

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

Organic Process Research & Development

Development of an Efficient Manufacturing Process for Reversible BTK Inhibitor GDC-0853 Haiming Zhang,*,† Theresa Cravillion,† Ngiap-Kie Lim,† Qingping Tian,† Danial Beaudry,† Jessica L. Defreese,† Alec Fettes,‡ Philippe James,¶ David Linder,¶ Sushant Malhotra,† Chong Han,† Remy Angelaud† and Francis Gosselin† †

Department of Small Molecule Process Chemistry, Genentech, Inc., 1 DNA Way, South San

Francisco, California 94080, United States ‡

Department of Process Chemistry and Catalysis, ¶Department of Drug Substance Scale-up and

Supply, F. Hoffmann-La Roche AG, Grenzacherstrasse 124, 4070 Basel, Switzerland

ACS Paragon Plus Environment

1

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

Page 2 of 42

TOC Graphic


2

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


ABSTRACT: Efforts toward the process development of reversible Bruton's Tyrosine Kinase (BTK) inhibitor GDC-0853 (1) are described. A practical synthesis of GDC-0853 was accomplished via a key highly regioselective Pd-catalyzed C–N coupling of tricyclic lactam 5 with 2,4-dichloronicotinaldehyde (6) to afford the C–N coupling product 3, a Suzuki–Miyaura cross-coupling of intermediate 3 with boronic ester 4 derived from a Pd-catalyzed borylation of tetracyclic bromide 7, to generate penultimate aldehyde intermediate 2 and subsequent aldehyde reduction and recrystallization. Process development of starting materials 5, 6 and 7 is also discussed.

Keywords: BTK inhibitor, regioselective, palladium-catalyzed, C–N coupling, borylation, Suzuki–Miyaura coupling


3


Page 4 of 42

INTRODUCTION Bruton tyrosine kinase (BTK) plays a significant role in all aspects of B cell development, including cell growth, maturation, migration and apoptosis, and is also crucial in the initiation, survival, and progression of B cell lymphoproliferative disorders.1 Therefore, BTK inhibitors have been actively investigated as potential treatments for several lymphoproliferative malignancies and autoimmune diseases.2 GDC-0853 (1, Figure 1), is an orally bioavailable reversible BTK inhibitor currently in clinical studies for the treatment of autoimmune diseases such as rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE).3 We describe herein our efforts toward the development of a practical and efficient manufacturing process for GDC0853 to support clinical development of this drug candidate, suitable for manufacturing up to hundreds of kilogram quantities of active pharmaceutical ingredient (API). O Me

N N Me Me

N OH

N N O

NH O N

Me

N

GDC-0853 (1)

Figure 1. GDC-0853, a reversible BTK inhibitor currently in clinical development RESULTS AND DISCUSSION Retrosynthetically, we envisioned that GDC-0853 (1) could be derived by a late-stage reduction from the penultimate aldehyde 2, which could be synthesized by a Suzuki‒Miyaura coupling4 of advanced intermediate 3 and boronate 4. Compounds 3 and 4 could be generated by a


4

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


regioselective Pd-catalyzed C‒N coupling5 of tricyclic lactam 5 and 2,4-dichloronicotinaldehyde (6) and a Pd-catalyzed borylation6 of tetracyclic bromide 7, respectively (Scheme 1). Scheme 1. Retrosynthesis of GDC-0853

Synthesis of Tricyclic Lactam 5. We embarked our process development of GDC-0853 (1) with the specific goal of developing a scalable and efficient process of tricyclic lactam 5. A careful evaluation of the structure of compound 5 led us to propose a retrosynthesis shown in Scheme 2. We envisioned that compound 5 could be synthesized by a base-facilitated 1,4-aza-Michael addition7 and cyclization of piperazin-2-one (5b) and 2-chloro-4,4-dimethylcyclopent-1-ene-1carbaldehyde (5c), which could be prepared from bulk starting material 3-methylcyclopent-2-en1-one (5e) via a conjugate Grignard addition and a Vilsmeier-Haack8 chloroaldehyde formation. Scheme 2. Retrosynthetic Analysis of Tricyclic Lactam 5


5


Page 6 of 42

Our synthetic efforts started with the Grignard conjugate addition to 3-methylcyclopent-2-en-1one (5e). We found that in the presence of catalytic amount of CuCl (20 mol %), conjugate addition of MeMgCl to 5e in THF at ‒20 °C proceeded smoothly, producing 43% yield of the desired 3,3-dimethylcyclopentanone (5d) after isolation by vacuum distillation (Scheme 3). The subsequent Vilsmeier-Haack reaction was achieved by employing 5d, POCl3 / DMF in DCM at 40 °C.9 Due to relative instability of the Vilsmeier-Haack product 5c, the reaction mixture, after aqueous work up, was telescoped directly to the subsequent base-promoted 1,4-aza-Michael addition and cyclization reaction employing N-methylmorpholine (NMM, 1.2 equiv) with piperazin-2-one (0.8 equiv) in NMP at 115 °C to afford the product tricyclic lactam 5 in 51% yield (Scheme 3). Although still low in overall yield (ca. 22%) and using an environmentally unfriendly solvent DCM, this process allowed us to quickly access large quantities (>100 kg) of 5 in only 3 steps in nearly 3× improvement in yield relative to the discovery synthesis.3 Scheme 3. Synthesis of Tricyclic Lactam 5.


6

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


During the Grignard addition reaction, two major impurities (5f and 5g) were observed based on GC-MS analysis (Scheme 4). We hypothesized that Grignard adduct 5h could undergo further reaction with either the starting material 5e or the product 5d to generate these two impurities, and thus, an in situ trapping of the adduct 5h could minimize the formation of the impurities. Indeed, addition of TMSCl to the Grignard reaction in THF not only accelerated the reaction but also trapped the intermediate enolate as the corresponding silyl enol ether 5i. We found 5i stable under the reaction conditions and did not dimerize or react with the starting material 5e. Furthermore, 5i has a higher boiling point that led to diminished losses during solvent exchange distillations. The catalyst loading was further optimized to 5 mol % CuCl with the addition of 10 mol % LiCl.10After the completion of the Grignard adduct trapping reaction and aqueous work up, the silyl enol ether product 5i (ca. 76% assay yield) was solvent-switched to PhMe, then cleaved in situ using sub-stoichiometric amount of HCl generated from POCl3 / H2O, followed by the addition of DMF and POCl3 to furnish the Vilsmeier-Haack reaction.11 The PhMe solution of 5c was then subjected to i-Pr2NEt promoted annulation reaction with piperazin-2-one (5b, 0.8 equiv) at 115 °C, affording 45% yield of the desired tricyclic lactam product 5 over four steps (Scheme 5). Overall, with this improved process, we were able to manufacture >300 kg of tricyclic lactam product 5 which not only eliminated the use of DCM from the process, but also improved the yield from 22% to 45% and generated the tricyclic lactam product 5 in excellent quality (>99.5A% HPLC purity). Scheme 4. Hypothesis to Eliminate Impurities 5f and 5g


7


Page 8 of 42

Scheme 5. Improved Process of Tricyclic Lactam 5

Synthesis of 2,4-Dichloronicotinaldehyde (6). 2,4-Dichloronicotinaldehyde12 (6) was readily synthesized by treating commercially available 2,4-dichloropyridine (6a) with LDA at ‒70 °C to generate the corresponding 3-pyridyllithium, which was then quenched with DMF to afford the desired aldehyde 6 in 70% yield at kilogram scale (eq 1). However, this chemistry suffered from cryogenic conditions, unstable organolithium intermediate and relatively low yield (eq 1). Thus,

Cl

Cl

1. LDA, THF, −70 °C 2. DMF

O Cl

Cl

(1)

N 70% 6a

N 6

we investigated a continuous process for further development of 3,4-dichloronicotinaldehyde (6) to address these drawbacks (Scheme 6).


8

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


Scheme 6. Continuous Process to 3,4-Dichloronicotinaldehyde (6) O Cl

continuous

Cl

Cl N

30 to 20 °C 6a

Feed 1 i-Pr2NH Feed 2 n-BuLi Feed 3 6a Feed 4 DMF

Cl N

87%

6 Thermostat ( 30 to 20 °C)

Mixer 1 Mixer 2 tR = 20 30 s tR = 20 30 s

Mixer 3

Quench Reactor

tR = 20 30 s

Lithium diisopropylamide (LDA) solution was prepared in situ when solutions of diisopropylamine (1.5 equiv) in THF and n-BuLi (1.1 equiv) in hexanes were pumped into a tubular reactor with a residence time (tR) of 20‒30 seconds at Tr = ‒20 to 0 °C. The LDA solution and a solution of 2,4-dichloropyridine (6a) in THF were then mixed into a second tubular reactor with a residence time of 20‒30 seconds at Tr = ‒30 to ‒20 °C to complete the lithiation reaction. The resulting 3-pyridyllithium solution and DMF (2.8 equiv) were then flowed into a third tubular reactor with the same residence time and Tr, producing the aldehyde product 6 after quenching in aqueous HCl. Aldehyde 6 was isolated in 87% yield after crystallization from PhMe / heptane in >99.5% purity on >200 kg scale. The continuous process not only eliminated cryogenic reaction conditions, but also overcame the instability issue of the organolithium intermediate by setting very short residence time of its generation and quenching, and resulted in a 17% increase in yield relative to the corresponding batch process. Synthesis of Tetracyclic Bromide 7. We initiated the synthesis of tetracyclic bromide 7 by examining various possible bond formation sequences from four building blocks, namely (S)-2methylpiperazine (7a), 5-chloro-2-nitropyridine (7b),13 oxetan-3-one (7c), and 3,5-dibromo-1methylpyridin-2-one (7d). The most promising route was then selected for further process


9


Page 10 of 42

optimization and development (Scheme 7). In short, the selective Boc protection of (S)-2methylpiperazine (7a) to generate 7e was readily achieved by employing 1.0 equiv of Boc2O in MeOH / H2O in 86% yield. A palladium-catalyzed Buchwald‒Hartwig amination5 of 7e and 5chloro-2-nitropyridine (7b) under optimized conditions employing 1.5 mol % Pd(OAc)2, 1.5 mol % BINAP and 2.5 equiv K3PO4 proceeded smoothly in PhMe at 85 °C. The C‒N coupling product 7f was then telescoped to a direct Boc deprotection using aqueous HCl to afford 7g in 73% yield over two steps. A reductive amination of compound 7g employing 1.3 equiv of oxetan-3-one (7c), 1.5 equiv of NaBH(OAc)3 and 0.5 equiv HOAc successfully produced 92% yield of the reductive amination product 7h. Catalytic hydrogenation of 7h in MeOH / PhMe gave product 7i, which was obtained as a solution in MeOH / PhMe and telescoped into a Buchwald‒Hartwig amination with 3,5-dibromo-1-methylpyridin-2-one (7d) employing 2 mol % Pd(OAc)2, 4 mol % Xantphos and 1.5 equiv of K2CO3 at 105 °C to produce tetracyclic bromide 7 in 76% yield over two steps with >99.5 A% HPLC purity. Scheme 7. Process Development of Tetracyclic Bromide 7


10

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


Concurrently, 3,5-dibromo-1-methylpyridin-2-one (7d) was prepared in 66% yield in two steps via NBS bromination / N-methylation of commercially available 2-hydroxypyridine (7j) (Scheme 8). Overall, the convergent synthesis of tetracyclic bromide 7 was accomplished in 8 steps with 6 longest linear steps in ca. 44% yield (Scheme 7). The process was successfully implemented in multiple batches and produced >300 kg of tetracyclic bromide 7 for downstream process development and manufacturing of GDC-0853 (1) API to support clinical demands. Scheme 8. Synthesis of 3,5-Dibromo-1-methylpyridin-2-one (7d)

Highly Regioselective C‒N Coupling of 5 and 6. We started the Pd-catalyzed Buchwald‒ Hartwig amination of tricyclic lactam 5 and 2,4-dichloronicotinaldehyde (6) with the goal of developing a highly regioselective C‒N coupling reaction. Initially, the Pd-catalyzed C‒N coupling of tricyclic lactam 5 and 1.5 equiv of 2,4-dichloronicotinaldehye (6) was performed in the presence of Pd2(dba)3 (2.5 mol%), Xantphos (5 mol%) and K2CO3 in MeCN at 80 °C, producing >95% conversion of the reaction, albeit with in situ formation of 5 A% of the C‒N coupling regioisomer 3a (Scheme 9, A). Direct addition of H2O to the reaction mixture allowed for isolation of crude product 3. Reslurrying crude 3 in THF / H2O at 65 °C afforded product 3 in 65% yield on multiple kilograms scale in >98 A% HPLC purity and regioisomer 3a was purged down to 0.8 A% by HPLC (Scheme 9, A). Scheme 9. Process Development of C‒N Coupling to Produce 3


11


Page 12 of 42

1. 1.5 equiv 6 2.5 mol % Pd2(dba)3 5 mol % xantphos 1.5 equiv K2CO3 MeCN, 80 °C A 2. THF/H2O, 65 °C 65%

Me Me

O N NH

B

Cl

+ Cl

1.5 equiv 6 2 mol % Pd(OAc)2 4 mol % dppf 1.5 equiv K2CO3 THF (20v), 68 °C

Me Me N

N

O

O 5

O

N

84%

6

C

Cl N

3

1.1 equiv 6 2 mol % Pd(OAc)2 4 mol % dppf 1.5 equiv K2CO3 THF (10v), 68 °C 85%

Me Me

O Cl

N

Me Me N

N N

O 3a

Me Me

O

N

O

N N

N

O 3b

The key challenge of the C‒N coupling reaction of 5 and 6 is the control of the regioisomer impurity 3a, which has similar physicochemical property to that of the desired C‒N coupling product 3. The regioisomer was very difficult to purge due to the overall poor solubility of 3 in many solvents with 1000 ppm). Thus, reduction and control of the regioisomer impurity 3a and residual Pd became the main goals for further process improvement.


12

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


Figure 2. Solubility of C‒N coupling product 3 at 20 °C. We performed an extensive catalyst screening for the Pd-catalyzed C‒N coupling of 5 and 6 via high-throughput experimentation (HTE). The result of a subset of HTE screening using Pd(OAc)2 as the catalyst and different ligands with K2CO3 as the base in THF is shown in Figure 3, which identified Pd(OAc)2 and 1,1'-bis(diphenylphosphino)ferrocene (dppf) as a more regioselective and cost-effective catalyst system. Further validation of the HTE conditions and optimization led to the optimal conditions for the C‒N coupling of 5 and 6 which employed 2 mol % Pd(OAc)2 as the pre-catalyst, 4 mol % dppf as the ligand, 1.5 equiv K2CO3 as the base in THF (20v, v = L/kg) at reflux. Under these conditions, we were able to successfully reduce the formation of regioisomer 3a to ca. 2 A% at reaction completion. This improvement hence introduced the possibility of single stage purification / isolation to achieve ≤0.8 A% 3a in the isolated product 3. In fact, the direct isolation of 3 was achieved in 84% yield by adding H2O (10v) to the reaction mixture, followed by aging, filtration and wash (Scheme 9, B). Although the loss in mother liquor was relatively high (ca. 10%), the final regioisomer 3a in the isolated 3 was markedly lower (0.56 A%) than 0.8 A%, allowing certain buffer room for experimental deviations. Residual Pd was also purged down to ca. 100 ppm.


13


Page 14 of 42

Figure 3. A sub-set of HTE screening using Pd(OAc)2 as the catalyst The impact of the stoichiometry of 2,4-dichloronicotinaldehyde (6) to the C‒N coupling reaction was also investigated. It was discovered that lowering 6 to 1.1 equiv (from 1.5 equiv) surprisingly generated less regioisomer 3a. We also observed a sharp reduction of the regioisomer 3a towards the end of the reaction (Figure 4) that was likely caused by the competitive consumption of 3a by starting material 5 to generate bis-coupling impurity 3b. Since 3b has significantly different structure to that of product 3, it was readily purged to 100 kg scale) of the desired C‒N coupling product 3 (Scheme 9, C). The new process not only greatly reduced the overall number of unit processing operations, but also generated highly crystalline 3, thus significantly increasing the filtration rate during isolation, and was successfully scaled up to >100 kg scale.


14

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


Figure 4. C‒N coupling reaction of 5 and 6 (1.1 equiv) Borylation of Tetracyclic Bromide 7. Next, we investigated the synthesis of boronate 4 by Pdcatalyzed borylation of tetracyclic bromide 7 with bis(pinacolato)diboron (B2pin2). We quickly identified the combination of Pd2(dba)3 and XPhos as an effective catalyst system to fulfill the borylation reaction in the presence of KOAc in THF at 65 °C. Further optimization of the reaction conditions by evaluating the stoichiometry of catalyst / ligand, B2pin2 and KOAc, and solvent volume was conducted. The optimal conditions we obtained thus employed 0.25 mol % Pd2(dba)3, 0.6 mol % XPhos, 1.5 equiv B2pin2 and 2 equiv KOAc in THF (10v) at 65 °C (Scheme 10). Under this set of optimal conditions, the reaction afforded >99% conversion in 12‒ 20 h with typically 100 kg scale and minimized the mother liquor loss to 99.5 A% with

Development of an Efficient Manufacturing Process for Reversible

Recommend Documents