Stereoselective Syntheses of the Antihistaminic Drug Olopatadine and Its E-Isomer

Practical stereoselective synthetic routes to the antihistaminic drug olopatadine and its E-isomer have been developed, the key steps being a trans stereoselective Wittig olefination using a nonstabilized phosphorus ylide and a stereoselective Heck cyclization. The stereoselectivity of the Wittig reaction depends on both the phosphonium salt anion and the cation present in the base used to generate the ylide. O is a nonsedating histamine H1-receptor antagonist with mast cell stabilizing properties. This dual action allows seasonal allergic conjunctivitis and rhinitis signs and symptoms to be controlled. Structurally, olopatadine is a dibenzo[b,e]oxepin derivative bearing a Z-configurated (dimethylamino)propylidene substituent and an acetic acid chain at the C-11 and C-2 positions, respectively, of the tricyclic system. Previous synthetic approaches to olopatadine involve the generation of the trisubstituted exocyclic double bond from a tricyclic dibenz[b,e]oxepin ketone, which represents a drawback in terms of Z/E stereoselectivity. We herein report short and practical stereoselective syntheses of olopatadine and its E-isomer (trans-olopatadine), based on a common strategy involving successive stereoselective Wittig and intramolecular Heck reactions as the key steps. After a Williamson reaction to assemble the benzyl aryl ether moiety, the Wittig olefination would lead to the key intermediates 1 or 2, depending on the aromatic aldehyde and aryl iodide derivatives used as the starting materials (Scheme 1). Taking into account that in particular cases, by choosing the appropriate reaction conditions, high selectivity for (E)and/or (Z)-alkene can be attained in the Wittig reaction of nonstabilized phosphorus ylides with aldehydes and that the Heck reaction usually involves a syn-addition/syn-elimination mechanism, the above Williamson−Wittig−Heck strategy could provide stereoselective access to both olopatadine and its E-isomer. The starting aldehyde 5 was prepared from 2-iodobenzyl chloride 3 and 3-formyl-4-hydroxyphenylacetic ester 4 under the reaction conditions indicated in Scheme 2. The Wittig reaction from 5 was initially performed using the ylide generated from the commercially available phosphonium bromide 6 (X = Br) by treatment with KHMDS in toluene (Table 1, entry 1), giving a mixture of E/Z alkenes in a 1:3 ratio. Similar results were obtained starting from the iodide salt 6 (X = I) (entry 2). Notably, the use of the lithium base LHMDS to generate the ylide from phosphonium iodide 6 (X = I) produced a dramatic change in the stereoselectivity of the Wittig reaction, leading to a 9:1 E/Z ratio of alkenes 1 in 73% yield (entry 3). A similar effect, although less pronounced (4:1 E/Z ratio), was observed from the bromide salt 6 (X = Br) (entry 4). In contrast, the use of phosphonium chloride 6 (X = Received: May 14, 2012 Published: June 25, 2012 Scheme 1. Synthetic Strategy: A Wittig/Heck Approach to Olopatadine and Its E-Isomer Note

O lopatadine is a nonsedating histamine H1-receptor antagonist with mast cell stabilizing properties. This dual action allows seasonal allergic conjunctivitis and rhinitis signs and symptoms to be controlled. 1 Structurally, olopatadine is a dibenzo [b,e]oxepin derivative bearing a Z-configurated (dimethylamino)propylidene substituent and an acetic acid chain at the C-11 and C-2 positions, respectively, of the tricyclic system. Previous synthetic approaches to olopatadine 2 involve the generation of the trisubstituted exocyclic double bond from a tricyclic dibenz[b,e]oxepin ketone, which represents a drawback in terms of Z/E stereoselectivity. 3 We herein report short and practical stereoselective syntheses of olopatadine and its E-isomer (trans-olopatadine), based on a common strategy involving successive stereoselective Wittig and intramolecular Heck reactions as the key steps. 4 After a Williamson reaction to assemble the benzyl aryl ether moiety, the Wittig olefination would lead to the key intermediates 1 or 2, depending on the aromatic aldehyde and aryl iodide derivatives used as the starting materials (Scheme 1). Taking into account that in particular cases, by choosing the appropriate reaction conditions, high selectivity for (E)-and/or (Z)-alkene can be attained in the Wittig reaction of nonstabilized phosphorus ylides with aldehydes 5 and that the Heck reaction usually involves a syn-addition/syn-elimination mechanism, 6 the above Williamson−Wittig−Heck strategy could provide stereoselective access to both olopatadine and its E-isomer.
The starting aldehyde 5 was prepared from 2-iodobenzyl chloride 3 and 3-formyl-4-hydroxyphenylacetic ester 4 7 under the reaction conditions indicated in Scheme 2. The Wittig reaction from 5 was initially performed using the ylide generated from the commercially available phosphonium bromide 6 (X = Br) by treatment with KHMDS in toluene ( Table 1, entry 1), 8 giving a mixture of E/Z alkenes in a 1:3 ratio. Similar results were obtained starting from the iodide salt 6 (X = I) 9 (entry 2). Notably, the use of the lithium base LHMDS to generate the ylide from phosphonium iodide 6 (X = I) produced a dramatic change in the stereoselectivity of the Wittig reaction, leading to a 9:1 E/Z ratio of alkenes 1 in 73% yield (entry 3). 10 A similar effect, although less pronounced (4:1 E/Z ratio), was observed from the bromide salt 6 (X = Br) (entry 4). In contrast, the use of phosphonium chloride 6 (X = Cl) 11 as the ylide precursor gave poor results in terms of both selectivity and chemical yield (entry 5).
The above trans selectivity is worthy of comment, as Wittig olefination reactions of nonstabilized phosphorus ylides with aromatic aldehydes usually produce the thermodynamically less stable cis isomer as the major product. 5 The presence of a nucleophilic group (for instance, amino) in the side chain of these ylides causes a shift in the stereochemistry of the alkene product toward the trans isomer, 12 whose production is also increased by lithium ions. 13 The pronounced effect of the lithium cation on alkene stereochemistry was also observed in the Wittig reaction from benzaldehyde 9, which was prepared from 2-formylbenzyl bromide 7 14 and 4-hydroxy-3-iodophenylacetic ester 8 15 (Scheme 3). Thus, alkenes 2 were produced in an E/Z ratio of 1:3 when the ylide was generated in toluene solution from phosphonium iodide 6 (X =I) using KHMDS as the base ( Table 2, entry 1). Neither the use of THF as the solvent (entry 2) nor starting from the bromide salt (entry 3) substantially modified the E/Z ratio. However, when using the lithium base LHMDS to generate the ylide from iodide 6 (X = I), the stereoselectivity was reversed, with a significant enhancement of E stereochemistry, leading to an E/Z ratio of 3.5:1 (entry 4).
From the synthetic standpoint, with four different alkenes now in hand, prepared in acceptable yield and good (E-1) to acceptable (Z-1, E-2, Z-2) stereoselectivity, we were ready to tackle the key Heck cyclization. The intramolecular Heck reaction from iodo alkene E-1 was performed under solid− liquid phase transfer conditions, using a stoichiometric quantity of Bu 4 NCl as the transfer agent 16 and K 2 CO 3 as the base, 17 in the presence of a catalytic amount of Pd(OAc) 2 (without phosphine ligands 18 ) in an acetonitrile−water mixture. 19 A single dibenzoxepin derivative 10, bearing a Z-configurated double bond, was formed with complete stereoselectivity in 60% yield. 20 A similar Pd-catalyzed cyclization from iodo alkene E-2 stereoselectively afforded E dibenzoxepin 11 in 55% yield. The above results were not unexpected and are consistent with a syn-addition of the initially formed arylpalladium intermediate to the alkene, with a subsequent syn β-elimination of a hydridopalladium halide, as outlined in Scheme 4.
However, in contrast with the high stereoselectivity observed in the above Heck arylations from trans alkenes E-1 and E-2, the Heck cyclization from cis alkenes Z-1 and Z-2 was not stereoselective, leading to mixtures of the cyclized products 10 and 11 (approximate ratios 3:2), probably as a consequence of a competitive Pd(II)-promoted isomerization of the starting Zconfigurated alkenes to the more stable E-isomers.
Finally, alkaline hydrolysis of the cyclized products 10 and 11 furnished the target drug olopatadine and its E-isomer, respectively.
In summary, we have developed practical synthetic routes to the antihistaminic drug olopatadine and its E-isomer based on a common strategy involving successive highly stereoselective Wittig olefination and Heck cyclization reactions as the key steps.  ■ EXPERIMENTAL SECTION triphenylphosphonium Iodide (6, X = I). A mixture of 3-(dimethylamino)-1-propyl chloride hydrochloride (20 g, 126 mmol), PPh 3 (33 g, 126 mmol), and NaI (19 g, 126 mol) in acetonitrile (80 mL) was heated at reflux for 5.5 days. After cooling at rt, H 2 O−acetonitrile (2:1, 360 mL) was added, and the mixture was stirred at 45°C for 0.5 h. The resulting suspension was filtered, solid K 2 CO 3 was added to the filtrate until pH 9−10, and the solution was extracted with CH 2 Cl 2 (3 × 150 mL). The combined organic extracts were dried and concentrated, and the resulting oil was dissolved in CH 2 Cl 2 (25 mL). AcOEt (120 mL) was added, and the mixture was stirred until the formation of a white solid, which was collected by filtration ( ,58.11;H,5.73;N,2.95;I,26.70. Found: C,58.13;H,5.66;N,2.90;I,26.58. (3-Dimethylaminopropyl)triphenylphosphonium Chloride (6, X = Cl). A mixture of 3-(dimethylamino)-1-propyl chloride hydrochloride (10 g, 63 mmol) and PPh 3 (16.5 g, 63 mmol) in 1butanol (40 mL) was heated at reflux for 4 days. The mixture was cooled at rt, toluene−Et 2 O (1:1, 50 mL) was added, and the solution was kept in the refrigerator for a night. The solid formed was filtered and washed with toluene−Et 2 O and Et 2 O to give pure 6·HCl (X = Cl) (15.6 g, 59%) as a white powder. A solution of this hydrochloride in 2 N aqueous K 2 CO 3 (50 mL) was extracted with CH 2 Cl 2 (3 × 30 mL).
Data for E-1: