Process Development and Good Manufacturing ... - ACS Publications

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES

Full Paper

Process Development and GMP Production of a Tyrosinase Inhibitor via Titanium Mediated Coupling Between Unprotected Resorcinols and Ketones Thibaud Gerfaud, Cedric Martin, Karinne Bouquet, Sandrine Talano, Corinne Millois-Barbuis, Branislav Musicki, Jean-Guy Boiteau, and Isabelle Cardinaud Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.7b00036 • Publication Date (Web): 20 Mar 2017 Downloaded from http://pubs.acs.org on March 20, 2017

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 30

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

Process Development and GMP Production of a Tyrosinase Inhibitor via Titanium Mediated Coupling Between Unprotected Resorcinols and Ketones Thibaud Gerfaud*, Cédric Martin, Karinne Bouquet, Sandrine Talano, Corinne Millois-Barbuis, Branislav Musicki, Jean-Guy Boiteau* and Isabelle Cardinaud Nestlé Skin Health R&D, 2400 Route des colles BP 87, 06902 Sophia-Antipolis Cedex, France

ACS Paragon Plus Environment

1


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 30

TABLE OF CONTENTS:


2

Page 3 of 30

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: A concise and economically attractive process for the synthesis of a novel tyrosinase inhibitor has been developed and implemented on a multi-kilogram scale under GMP. Major achievement to the success of the process is the development of a direct coupling between free resorcinol and ketone. First developed under basic conditions, this coupling has been turned to a novel titanium (IV) mediated process allowing good selectivity, easy isolation and high atom efficiency. Other key steps feature an alkene reduction by palladium catalyzed transfer hydrogenation and a urea formation using N,N’-Disuccinimidyl carbonate as the carbonyl source. This route allowed us to produce kilogram batches of the candidate to support preclinical and clinical studies.

KEYWORDS: Tyrosinase Inhibitors, Resorcinol, Titanium mediated coupling

INTRODUCTION: Hyperpigmentation disorders such as melasma are characterized by an increase in melanin synthesis which accumulates in the epidermis and is responsible for a darkening of the skin.1 Melanogenesis occurs in the basal layer of the epidermis into specific organelles of the melanocytes called melanosomes.2 A detailed analysis of the biosynthetic pathway reveals that tyrosinase is a key enzyme in melanogenesis3 and is responsible for the oxidation of tyrosine into DOPA (3,4-dihydroxyphenylalanine) and DOPAquinone.4 1 (Figure 1) is a melanogenesis inhibitor working through inhibition of tyrosinase (IC50 = 0.1 µM on normal


3


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 4 of 30

human epidermal melanocytes) currently under development at Nestlé Skin Health R&D for the topical treatment of hyperpigmentation disorders.5 Figure 1. Structure of Tyrosinase Inhibitor Clinical Candidate 1

Clinical candidate 1 is a small molecule (MW = 340.42 g.mol-1) designed against the following parameters: activity, skin penetration properties, topical formulability, low systemic exposure and high safety. Key structural features are a resorcinol moeity functionalized at position 4, a urea and one chiral center. Discovery chemistry synthesis6 (Scheme 1) started from 4-bromoresorcinol, an expensive raw material, difficult to source on large scale. Double benzyl protection followed by lithium-halogen exchange under cryogenic conditions (-70 °C) and consecutive addition on 1-carbethoxy-4piperidone yielded the tertiary alcohol 3 (62% yield after recrystallization). Hydrogenation of 3 led to the reduced and deprotected intermediate 4: re-protection of the 2 phenol moieties was necessary to avoid further side reactions and was accomplished with benzyl bromide to give 5. Carbamate hydrolysis, followed by condensation with (S)-(-)-α-methylbenzyl isocyanate and final benzyl deprotection gave 1 in 16% overall yield and 7 chemical steps. Scheme 1. Discovery Chemistry Route to 1


4

Page 5 of 30

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


While well suited for library synthesis with an advanced intermediate 6 that could be easily functionalized at the N position, this synthesis required modifications to be amenable to large scale GMP production. Not only chromatographic purifications nor cryogenic conditions had to be avoided but material costs and atom efficiency were major issues to address for further development. Our long-term retrosynthetic strategy was thus driven mainly by the use of cheap and easily available reagents (Scheme 2). Scheme 2. Retrosynthetic Analysis of 1

Route selection Given some important time constraints and relatively small amounts of drug substance required for early toxicological studies (100 g), preliminary investigations focused on reducing the overall number of steps and limiting protection/deprotection sequences. Doubly Obenzylated bromoresorcinol 2 was kept as first intermediate and efforts were turned toward coupling of the piperidine moeity and formation of the urea (Scheme 3). Chirality was set up using cheap and readily available (S)-(-)-α-methylbenzylamine 8 and urea formation


5


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 6 of 30

accomplished by pre-activating 8 with N,N′-carbonyldiimidazole (CDI), avoiding the use of the unstable isocyanate on scale. Suzuki coupling on bromoresorcinol 2 was considered as an alternative to ketone 1,2-addition and various options were investigated, most relevant ones being depicted in Scheme 3. Enol triflate formation from N-Boc-4-piperidone 9 followed by Miyaura borylation with bis(pinacolato)diboron (B2pin2) furnished vinylboronate 10 (57% yield on 10 g scale, isolated as a solid). Subsequent Suzuki coupling with 2 gave 11 in good yield and good overall purity (oil). Boc deprotection, urea formation and final double bond reduction/benzyl deprotection furnished 1 in 28% overall yield. An alternative approach using commercially available 4-pyridineboronic acid gave pyridine 14 (63% yield on 100 g scale, isolated as a solid), which was hydrogenated using a mixture of palladium and platinum oxide catalyst with concomitant benzyl deprotection to deliver amine 15-AcOH as the acetate salt. Interestingly urea formation with 13 proceeded smoothly and delivered 1 in 20% overall yield with very low amounts of O-acylation byproduct observed. Other approaches involving direct Suzuki couplings on unprotected 4-bromoresorcinol were unsuccessful.7 Scheme 3. Alternative Strategies Featuring a Suzuki Coupling


6

Page 7 of 30

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


While with these routes overall yield were improved and protection/deprotection sequences reduced, major limitations still remained for a successful scale-up campaign: prices of raw materials (4-bromoresorcinol, 4-pyridineboronic acid), cryogenic temperature (enol formation), harsh conditions (high catalyst loading, catalyst poisoning during pyridine hydrogenation)… It was anticipated that even with strong optimization efforts, final material costs of 1 would remain high using these synthetic pathways. The high selectivity observed when conducting the urea formation with activated benzylamine 13 on fully unprotected amine 15 prompted us to investigate original synthetic pathways to prepare this intermediate. Resorcinol and other polyphenols are good nucleophiles toward C-alkylation, especially under Friedel-Craft conditions but unprotected resorcinols (free OH) remain challenging substrates because O-alkylation can be predominant with certain electrophiles (e.g. acyl chlorides). Calkylation can be achieved but selectivity and over alkylation issues are often encountered. We were intrigued by the possibility to form the requisite C-C bond directly from resorcinol (or one of its cheap derivatives) and preliminary investigations started with dimethoxy resorcinol which condensed with piperidone in acetic acid and conc. HCl to yield alkene intermediate 16 (Scheme 4).8 Subsequent reduction furnished 17 in 39% yield over 2 steps and demethylation was accomplished under harsh conditions (Pyr-HCl, 160 °C) yielding targeted 15-HCl in 90% yield. This approach afforded 1 in only 5 steps and 23% overall yield using inexpensive raw material. Scheme 4. Alternative Approach with Dimethoxyresorcinol


7


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 8 of 30

Unfortunately, high amounts of impurities were formed during the first step and were impossible to separate from 16 by non-chromatographic methods. The very harsh conditions required for the demethylation were another important drawback of this approach. To overcome these issues, we decided to focus on the development of a method to couple unprotected resorcinol with N-Boc-4-piperidone, both starting materials being cheap and readily available.

RESULTS AND DISCUSSION: Development of coupling conditions with unprotected resorcinol Under acidic conditions, unprotected resorcinols are known to undergo polycondensations with aldehydes and ketones providing resorcinarenes and polymeric species (with aldehydes)9 or chroman derivatives (with ketones)10 (Scheme 5). Scheme 5. Known Direct Condensation of Unprotected Resorcinol with Ketones11

With piperidone derivatives however, aldolization has been reported to be much slower enabling a direct 1,2-addition on ketone and isolation of the corresponding alcohol.11 These conditions were evaluated using N-Boc or N-benzyl piperidone leading in our case to the formation of mixtures of alcohols and corresponding alkenes albeit always in low yields and accompanied by significant amounts of bis-addition side-products, chroman and other impurities


8

Page 9 of 30

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). A rapid screening of the reaction conditions regarding to resorcinol stoichiometry, acid source (AcOH, H2SO4, p-TSA), solvent and temperature did not permit to optimize yield and ratio of targeted products 18 and 19 vs bis-addition side-products to a level sufficient for scale-up. Scheme 6. Products Formed During Direct Resorcinol Alkyation Under Acidic Conditions

Given the poor preliminary results obtained under acidic conditions, we decided to evaluate basic ones, aiming that deprotonation of resorcinol would enhance its nucleophilicity toward Calkylation and allow it to add on N-Boc piperidone. Interestingly, first experiments (LiOH in THF/water mixture) yielded the desired alcohol intermediate 18 in 35% isolated yield along with 27% of bis-addition adduct 20 (HPLC A%). An optimization of these conditions was performed against the following parameters: type of base (NaOH, MeONa, EtONa, t-BuOK), solvent, resorcinol stoichiometry, temperature… but minimization of the amount of bis-addition adduct formed during the reaction and isolation of the desired product free of this impurity proved to be a significant challenge12. The best balance was obtained using sodium ethoxide in ethanol at 23


9


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 10 of 30

°C, with 4 equiv of resorcinol and a slow addition of N-Boc piperidone over the deprotonated resorcinol (Scheme 7). In this case, only 8-10% of 20 were formed. Direct isolation of 18 out of the reaction mixture was not possible, probably because the large excess of resorcinol used prevented its crystallization. Upon completion of the reaction, pH of the solution was adjusted to 6.3 using NaH2PO4 and the reaction mixture warmed to reflux to dehydrate the alcohol into corresponding alkene 19. Subsequent solvent exchange from EtOH to water led to crystallization of 19 in 40% yield contaminated by 4-5% of impurity 22. This procedure was scaled up to deliver >300 g of 19. Unfortunately, despite an extensive screening of recrystallization conditions, a process able to fully remove the bis-addition impurity at this stage was not found.13 Best compromise proved to be a trituration of the solid in 3 vol of toluene which improved the impurity profile, reducing the amount of 22 to less than 3%. Further scale-up of this approach was impossible since product crystallization was capricious leading alternatively to gummy residues or solid material with variable particle size, sometimes very difficult to drain out of the reaction vessel. Lack of robustness of the crystallization was attributed to the large excess of resorcinol used in the process. First 100 g batches of 1 were produced using this route but bisaddition impurity was not depleted in the downstream chemistry (alkene reduction, Boc deprotection and coupling with activated urea). Triturations and chromatographic purifications of intermediates were required to control the level of bis-addition impurity below 1% in the final product. Scheme 7. Optimized Base Promoted Condensation of Unprotected Resorcinol with N-Boc piperidone


10

Page 11 of 30

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


Intrigued by the difficulty encountered in the optimization of the coupling process (regarding resorcinol stoichiometry and amount of bis-addition impurity), the product of the condensation (pure alcohol 18) was isolated by chromatography and re-submitted to the reaction conditions (Table 1). Surprisingly, after 12 h only 50% (HPLC A%) of 18 remained in solution along with 35% resorcinol and 14% of bis-addition impurity 20. These results clearly indicated that a retroaddition occurred with cleavage of a C-C bond and thus that an equilibrium was reached under the reaction conditions. Table 1. Reversibility Observed for the Base Promoted Processa

Time (h)

18 (A%)b

Resorcinol (A%)b

20 (A%)b

0

100

0

0

24c

50

35

14

a

Reaction conducted with 3.2 mmol of 18 in 7 vol of EtOH and 4 equiv of EtONa (2.7 M solution in EtOH) at 23 °C for 24 h; bConversion measured by HPLC (A%); cN-Boc-4-piperidone formation was confirmed by TLC analysis (DCM/EtOAc 9:1).

In light of this new information, it became clear that further attempts to reduce the amount of resorcinol and bis-addition by-product would be useless. To solve these issues (reversibility, resorcinol amount, robustness) we re-examined the acidic conditions aiming to find an oxophilic


11


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 12 of 30

Lewis acid able to play multiple roles: 1) activate the ketone with a decent turn-over, 2) orientate resorcinol attack to minimize bis-addition adducts, 3) chelate hydroxyl group generated to avoid retro-addition reaction and 4) keep the Boc protection untouched. An extensive screening of Lewis acids and solvents was performed14 and amongst all acids tested we were pleased to find that only titanium alkoxides were able to promote the reaction to complete conversion with limited amounts of side-products under mild conditions. Various solvents were compatible with the transformation and 2-MeTHF15 was selected for its process compatibility. At 23 °C a clean reaction profile was observed with Ti(OiPr)4, providing a promising starting point for further optimization (Table 2, entry 1). At higher temperature, higher levels of bis-addition adducts were observed. Lowering the reaction temperature to 0 °C or the amount of resorcinol to one equivalent had a detrimental impact on reaction’s rate or levels of bis-addition impurity (Table 2, entry 2 and 4). We postulated that the reaction proceeds first through a ligand exchange between resorcinol and titanium (IV) alkoxide (a precipitate was observed at the beginning of the Ti addition), followed by activation of the ketone leading to titanium complex 24 and C-C bond formation as shown in Scheme 8.16 Formation of 24 was deemed to be responsible for the fast reaction rate as it would bring the 2 reagents into close spatial proximity, leading to fast intramolecular addition. Scheme 8. Proposed Mechanism for the Ti(IV)-Mediated Condensation


12

Page 13 of 30

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


This assumption was reinforced by the fact that running the same reaction with 3-methoxyphenol gave only one isomer (28) with no evidence of bis-addition impurity (Scheme 9). Scheme 9. Ti(IV)-Mediated Condensation with 3-methoxyphenola

a

Reaction conducted on 10 mmol scale with 2 equiv of 3-methoxyphenol and 0.5 equiv of Ti(OiPr)4 in 7 vol of solvent We hypothesized that formation of the bis-addition impurity proceeds via complexes 26 and/or 27 and that a key role is played by the ligands surrounding the Ti atom. Indeed, higher levels of bis-addition impurity 22 and longer reaction time were observed with less hindered Ti(OEt)4 (Table 2, entry 6). In this case, a thick and difficult to stir precipitate17 formed and remained in suspension whereas with Ti(OiPr)4 the precipitate dissolved rapidly (15-20 minutes) and less than 4 h were necessary to reach full conversion (Table 2, entry 5). With more hindered Ti(OtBu)4, levels of bis-addition impurity could be reduced down to 3-4% with a shorter reaction time and no precipitate was observed (Table 2, entry 7). We assume that formation of complex 23 (or


13


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 14 of 30

a structurally close polymeric species) is responsible of the observed precipitate. With small ethoxide ligand, the Ti atom can accommodate more resorcinol ligands (x≥ 2) leading to precipitation of 23 whereas with more hindered isopropoxide and tert-butoxide ligands, complex 23 is more soluble (x = 1) and less stable, explaining the higher reactivity and the faster dissolution of the precipitate. On the other hand, complex 27 is probably less prone to form with bulkier ligands, explaining the lower amounts of impurity 22 observed with Ti(Ot-Bu)4. Table 2. Conditions Screening for the Ti(IV)-Mediated Couplinga

entr y 1

conditions

conversionb

iPr

Ti equiv 1.0

100%

product yieldc 79%

2-MeTHF, 23 °C, 1 h

6%

2

iPr

0.5

2-MeTHF, 0 °C, >6 h

Process Development and Good Manufacturing ... - ACS Publications

Recommend Documents