3094
`
`J. Med. Chem. 1999, 42, 3094-3100
`
`Novel Prodrug Approach for Tertiary Amines: Synthesis and Preliminary
`Evaluation of N-Phosphonooxymethyl Prodrugs
`
`Jeffrey P. Krise,† Jan Zygmunt,‡ Gunda I. Georg,‡ and Valentino J. Stella*,†
`
`Department of Pharmaceutical Chemistry, The University of Kansas, 2095 Constant Avenue, Lawrence, Kansas 66047, and
`Department of Medicinal Chemistry, The University of Kansas, Malott Hall, Lawrence, Kansas 66045
`
`Received September 25, 1998
`
`The synthesis and preliminary evaluation of a novel prodrug approach for improving the water
`solubility of drugs containing a tertiary amine group are reported. The prodrug synthesis
`involves a nucleophilic substitution reaction between the parent tertiary amine and a novel
`derivatizing reagent, di-tert-butyl chloromethyl phosphate, resulting in formation of the
`quaternary salt. The tertiary butyl groups are easily removed under acidic conditions with
`trifluoroacetic acid giving the N-phosphonooxymethyl prodrug in the free phosphoric acid form,
`which can subsequently be converted to the desired salt form. The synthesis was successfully
`applied to a model compound (quinuclidine) and to three tertiary amine-containing drugs
`(cinnarizine, loxapine, and amiodarone). The prodrugs were designed to undergo a two-step
`bioreversion process. The first step was an enzyme-catalyzed rate-determining dephosphory-
`lation followed by spontaneous chemical breakdown of the N-hydroxymethyl intermediate to
`give the parent drug. Selected prodrugs were shown to be substrates for alkaline phosphatase
`in vitro. A preliminary in vivo study confirmed the ability of the cinnarizine prodrug to be
`rapidly and completely converted to cinnarizine in a beagle dog following iv administration.
`
`Introduction
`The synthesis and preliminary evaluation of a novel
`prodrug strategy for improving the water solubility of
`tertiary amine-containing drugs is described. A repre-
`sentation of the prodrug strategy is shown in Scheme
`1. The tertiary amine 1 is chemically derivatized to
`produce the polar, water-soluble prodrug 2. The prodrug
`is designed to release the parent tertiary amine in vivo
`through a two-step bioreversion process. The first step
`(k1, rate-determining step) in prodrug bioreversion
`involves a phosphatase-catalyzed dephosphorylation to
`give the resultant hydroxymethyl quaternary ammo-
`nium intermediate 4 and inorganic phosphate (3). The
`intermediate 4, at physiological pH, is highly unstable
`and spontaneously breaks down (k2) to give the parent
`tertiary amine 1 plus formaldehyde (5).
`Many drugs, including tertiary amines, have poor
`aqueous solubilities, which can create obstacles to their
`safe and effective delivery. Formulation-related toxici-
`ties can occur when a drug is given parenterally.1,2 In
`addition, poor and/or erratic bioavailability can occur
`when a drug is given orally.3 Prodrugs of tertiary amines
`have received little attention in the literature. Bodor
`et al. has previously described acyloxyalkyl quaternary
`ammonium derivatives of tertiary amines as “soft
`drugs”,4-7 while others, albeit few, have explored qua-
`ternary ammonium derivatives as water-soluble pro-
`drugs for tertiary amines.8-11 Quaternary amine pro-
`drugs resulting from N-phosphonooxymethyl derivitiza-
`tion of the tertiary amine functionality of problematic
`drugs represent a novel approach for improving the
`water solubility.
`
`* To whom correspondence should be addressed.
`† Department of Pharmaceutical Chemistry.
`‡ Department of Medicinal Chemistry.
`
`Scheme 1. Illustration of the Prodrug Strategy
`
`Three drugs (loxapine, cinnarizine, and amiodarone)
`and one model compound (quinuclidine), each containing
`a tertiary amine group, were chosen to demonstrate the
`synthetic feasibility of this prodrug strategy. Cinnariz-
`ine is a calcium channel blocker that is clinically used
`as a vasodilator and antihistaminic and antiallergic
`agent.12 Cinnarizine has a pKa1 of 1.9513 and pKa2 of
`7.47.14 The poor, pH-dependent, aqueous solubility of
`cinnarizine is believed to be responsible for the erratic
`oral bioavailability observed in dogs and in humans.15,16
`Loxapine is a tricyclic dibenzoxazepine antipsychotic
`agent, which is used in the treatment of schizophrenic
`and psychotic disorders.17 The loxapine pKa2 is 7.5, and
`the intrinsic solubility of the free base is 12.6 μg/mL.18
`The drug is currently formulated for im administration
`
`10.1021/jm980539w CCC: $18.00 © 1999 American Chemical Society
`Published on Web 07/17/1999
`
`Patent Owner, UCB Pharma GmbH – Exhibit 2013 - 0001
`
`

`
`Novel Prodrug Approach for Tertiary Amines
`
`Journal of Medicinal Chemistry, 1999, Vol. 42, No. 16 3095
`
`Chart 1. N-Phosphonooxymethyl Derivatives of
`Quinuclidine (6), Cinnarizine (7), Loxapine (8), and
`Amiodarone (9)
`
`Scheme 2. General Synthetic Scheme for
`N-Phosphonooxymethyl Prodrugs of Tertiary Amines
`
`in a cosolvent consisting of propylene glycol (70% v/v)
`and polysorbate 80 (5% v/v) which has been shown to
`cause muscle damage.19 Amiodarone is a widely used
`antiarrhythmic agent which is currently approved for
`life-threatening ventricular arrhythmias.20-22 Amio-
`darone has a pKa of 6.56 and also has poor aqueous
`solubility.23 This necessitates an iv formulation to have
`lowered pH and nonaqueous cosolvent addition.24 The
`development of venous phlebitis has been reported with
`the iv infusion of this formulation which is thought to
`be caused through precipitation of amiodarone at the
`injection site.25 When given orally, the absolute bio-
`availability of amiodarone ranges from 22-86% which
`has been partially attributed to the poor water solubil-
`ity.26
`The purpose of the current studies was to evaluate
`the potential usefulness of the novel prodrug strategy
`in improving the water solubility of these drugs. To meet
`this goal the prodrug must have ample water solubility,
`sufficient chemical stability, and a rapid and complete
`bioreversion process. These issues were topics of sepa-
`rate reports.18,27 At pH 7.4, the loxapine prodrug had
`over a 15 000-fold improvement in solubility relative to
`loxapine, and the predicted shelf life of an aqueous
`formulation was nearly 2 years. Both the loxapine and
`cinnarizine prodrugs appeared to be rapidly and quan-
`titatively converted to their respective parent compound
`following iv dosing. The specific aim of this work was
`to describe the method of synthesis of the prodrugs. Also
`included are preliminary in vitro reversion kinetics of
`7 and 8 in the presence of isolated human placental
`alkaline phosphatase and an in vivo evaluation of 7 in
`a beagle dog.
`
`Results
`
`Synthesis. The structures of the quaternary N-
`phosphonooxymethyl prodrugs of quinuclidine (6), cin-
`narizine (7), loxapine (8), and amiodarone (9) are shown
`in Chart 1. The general mechanism for the synthesis of
`prodrugs is depicted in Scheme 2. The parent tertiary
`amine 10 undergoes a nucleophilic substitution reaction
`
`with di-tert-butyl chloromethyl phosphate (11), which
`results in the formation of the quaternary ammonium
`phosphate-protected prodrug 12. The solvent selection
`for this quaternization reaction was important. It was
`found that a relatively polar aprotic solvent such as
`acetonitrile worked best. The derivatizing reagent, di-
`tert-butyl chloromethyl phosphate (11), was not com-
`pletely stable under the elevated temperature conditions
`required by less reactive amines, such as cinnarizine,
`and would slowly degrade to produce hydrochloric acid
`(HCl). The generation of HCl protonated any unreacted
`tertiary amine and catalyzed the removal of tertiary
`butyl protecting groups from both the product (12) and
`the unreacted derivatizing reagent (11). This led to the
`formation of multiple polar products, which made the
`purification of some prodrugs difficult. To limit this
`occurrence, a nonnucleophilic proton scavenger (1,2,2,6,6-
`pentamethylpiperidine) was used in excess in all reac-
`tions, except for quinuclidine which was sufficiently
`reactive to proceed at 37 °C in a relatively short time.
`Even in the presence of the proton scavenger, cinnariz-
`ine and loxapine reaction products were isolated as the
`monotertiary butyl-protected derivatives as opposed to
`the ditertiary butyl-protected intermediates that were
`isolated with quinuclidine and amiodarone. The mono-
`or diester prodrugs 12 were then treated with trifluo-
`roacetic acid in benzene at room temperature to remove
`the tertiary butyl group(s). The resulting prodrugs in
`their free acid form, 13, were isolated and characterized.
`In the cases of cinnarizine and loxapine, more than one
`molecule of trifluroacetic acid might have been associ-
`ated with the products since they contain other basic
`centers. The prodrugs can be converted to their sodium
`salt by neutralization; however, the compounds were
`somewhat hygroscopic, so materials were stored in the
`quaternary, free acid form and all initial studies were
`on these materials.
`The synthesis of derivatives of 11 with an improved
`leaving group, compared to chlorine (i.e., p-toluene-
`sulfonate, triflate), were attempted; however, these
`compounds were too unstable for isolation. Derivatives
`of 11 with phosphate protecting groups other than tert-
`butyl may allow for these substitutions. A derivatizing
`reagent with improved reactivity could potentially lead
`to enhanced yields and less difficult purifications (par-
`ticularly important when derivatizing weakly basic
`tertiary amines).
`In Vitro and in Vivo Evaluation. In order for the
`derivatives to behave as prodrugs they must undergo a
`chemical or enzymatic bioreversion process. Alkaline
`
`Patent Owner, UCB Pharma GmbH – Exhibit 2013 - 0002
`
`

`
`3096 Journal of Medicinal Chemistry, 1999, Vol. 42, No. 16
`
`Krise et al.
`
`Figure 1. In vitro alkaline phosphatase-catalyzed reversion
`of 7 (b) to cinnarizine free base (O) at 37 °C and pH 10.4. As
`a control, levels of 7 (2) are plotted as a function of time under
`the same conditions in which the buffer is devoid of alkaline
`phosphatase.
`
`Figure 2. In vitro alkaline phosphatase-catalyzed reversion
`of 8 (b) to loxapine free base (O) at 37 °C and pH 10.4. As a
`control, levels of 8 (2) are plotted as a function of time under
`the same conditions in which the buffer is devoid of alkaline
`phosphatase.
`
`phosphatase is an enzyme that is found throughout
`various tissues, membranes, etc.28 The described pro-
`drugs were designed to be substrates for this enzyme,
`and the ability of the prodrugs to generate parent drug
`in the presence of the enzyme was assessed.
`Figures 1 and 2 show the depletion of 7 and 8 in the
`presence of alkaline phosphatase along with the simul-
`taneous formation of cinnarizine and loxapine, respec-
`tively. The closed triangles represent the prodrug levels
`under identical conditions without the addition of
`enzyme, which serves to demonstrate the chemical
`stability of the prodrugs under the conditions. As seen
`in Figure 1, complete disappearance of 7 led to ap-
`proximately 88% apparent molar recovery of cinnarizine
`free base. The 12% not recovered could be partially
`accounted for by a 3.7% water content (Karl Fisher
`
`Figure 3. Plot of cinnarizine plasma concentration versus
`time following equimolar (3 μmol/kg) iv doses of cinnarizine
`and 7 to a beagle dog:
`(b) cinnarizine levels following
`cinnarizine administration; (O) cinnarizine levels following
`administration of 7.
`
`titration) and some minor impurities based on relative
`peak area HPLC analysis of the sample. Similarly, in
`Figure 2, the disappearance of 8 led to near-quantitative
`94% molar recovery of loxapine. The 6% not recovered
`could be accounted for by a 4.2% water content and
`2-3% impurities based on relative peak area HPLC
`analysis of the sample. NMR spectra (hydrogen, carbon,
`and phosphorus) and MS as well as HPLC chromatog-
`raphy all suggested that the prodrugs were quite pure;
`however, in hindsight the less than quantitative recov-
`ery from the alkaline phosphatase study suggested the
`presence of some unaccounted for impurities in the
`samples. The most likely impurities are small quantities
`of chloride ion and trifluroacetic acid species.
`The ability of the prodrugs to revert to parent
`compound in the presence of alkaline phosphatase
`clearly indicates that these prodrugs are substrates for
`the enzyme. The promising in vitro reversion profiles
`of the prodrugs warranted further investigation in
`animals. In vivo, ideally, these prodrugs should be
`nontoxic and be rapidly and quantitatively converted
`to the parent compound. The ability of 7 to revert to
`cinnarizine was evaluated in a single beagle dog,
`crossover study. Neither injection (cinnarizine or 7)
`caused observable signs of discomfort or toxicity to the
`dog. Figure 3 shows the concentration versus time
`profile for cinnarizine following equimolar (3 μmol/kg)
`iv doses of cinnarizine and 7. A comparison of the
`cinnarizine area under the plasma concentration versus
`time curve was used to assess the extent of bioreversion
`in the dog. The [AUC]0-8h values were 456.0 and 449.5
`ngh/mL after cinnarizine and 7, respectively. Dividing
`the resulting cinnarizine [AUC]0-8h after administration
`of 7 by the [AUC]0-8h following cinnarizine administra-
`tion gives a value of 0.98, which represents the extent
`of reversion of the prodrug to parent drug. This, along
`with the superimposability of the profiles, suggests that
`the prodrug is both rapidly and quantitatively converted
`to parent drug in dog.
`
`Patent Owner, UCB Pharma GmbH – Exhibit 2013 - 0003
`
`

`
`Novel Prodrug Approach for Tertiary Amines
`
`Journal of Medicinal Chemistry, 1999, Vol. 42, No. 16 3097
`
`Conclusion
`
`The reaction of selected tertiary amine-containing
`drugs with di-tert-butyl chloromethyl phosphate fol-
`lowed by deprotection led to the formation of the
`N-phosphonooxymethyl prodrugs. When the parent
`tertiary amine is highly nucleophilic (e.g., quinuclidine)
`the prodrug can be obtained in high yield with little
`purification necessary. When the amine is less nucleo-
`philic (e.g., cinnarizine), the required elevated temper-
`ature and long reaction times led to complex product
`formations, difficult purification, and reduced yields.
`Prodrugs 7 and 8 were shown to be substrates for
`alkaline phosphatase, an enzyme ubiquitous to the
`human body. In a preliminary in vivo dog study, 7 was
`shown to produce cinnarizine rapidly and quantitatively
`following intravenous administration. Further in-depth
`studies evaluating this prodrug concept are reported in
`separate papers.18,27 Ongoing chemical studies include
`improved synthesis and purification procedures for
`these polar molecules.
`
`Experimental Section
`
`Synthetic Materials. Anhydrous acetonitrile, chloro-
`iodomethane, anhydrous dimethoxyethane, 1,2,2,6,6-penta-
`methylpiperidine, quinuclidine, tetramethylammonium hy-
`droxide 10% (w/v) aqueous solution, trifluoroacetic acid,
`magnesium chloride hexahydrate, tetrabutylammonium di-
`hydrogen phosphate, and zinc chloride were obtained from
`Aldrich Chemical Co. (Milwaukee, WI). Cinnarizine and amio-
`darone hydrochloride were obtained from Sigma Chemical Co.
`(St. Louis, MO). Loxapine succinate was obtained from Re-
`search Biochemicals Inc. (Natick, MA). Di-tert-butyl phosphite
`was obtained from Lancaster Synthesis, Inc. (Windham, NH),
`while potassium bicarbonate and glycine were obtained from
`Fischer Scientific (Pittsburgh, PA). Potassium permanganate
`was purchased from Mallinckrodt Chemical Works (St. Louis,
`MO). Normal phase silica gel, particle size 32-63 μm, was
`obtained from Selecto Scientific (Norcross, GA), while prepara-
`tive TLC plates (1000-μm thickness, 20 × 20 cm) were obtained
`from Alltech Associates, Inc. (Deerfield, IL). Amiodarone
`hydrochloride and loxapine succinate were converted to their
`free base form prior to use. All water was distilled in an all-
`glass still prior to use. All other chemicals and solvents were
`of reagent grade obtained from conventional sources and used
`without further purification.
`Analytical Equipment. Nuclear magnetic resonance spec-
`tra were obtained using either a Bruker AM 500 or a Varian
`XL 300 instrument. Chemical shifts are reported as parts per
`million downfield from tetramethylsilane or 3-(trimethylsilyl)-
`propanesulfonic acid as internal standards for 1H NMR spectra
`and from 85% H3PO4 as an external standard in 31P NMR
`spectra. J values are given in hertz (Hz). Elemental mi-
`croanalyses were performed by Desert Analytics, Tuscon, AZ.
`Melting points obtained were determined in capillary tubes
`on a Mel-Temp II apparatus. Mass spectral analyses were
`performed by the Mass Spectral Laboratory at The University
`of Kansas (Lawrence, KS). Water content was determined on
`a Metrohm AG Karl Fischer coulometer (Herisam, Switzer-
`land).
`Di-tert-butyl Chloromethyl Phosphate (11). The conver-
`sion of di-tert-butyl phosphite into the corresponding phosphate
`was performed using a modification of the method reported
`by Zwierzak and Kluba.29 To a stirred solution, in an ice bath,
`of di-tert-butyl phosphite (7.84 g, 40.36 mmol) and potassium
`bicarbonate (2.42 g, 24.20 mmol) in water (35 mL) was added,
`in three equal portions, finely powdered potassium perman-
`ganate (4.46 g, 28.30 mmol) over 1 h followed by continued
`stirring at room temperature for an additional 30 min.
`Decolorizing carbon (0.6 g) was added, and the resulting
`
`mixture was stirred at 60 °C for 15 min and filtered. The filter
`cake was washed three times with water (5 mL); all filtrates
`were combined, mixed with decolorizing carbon (1.0 g), and
`again stirred at 60 °C for 20 min. After filtration, the resulting
`colorless solution was cooled to 0 °C (ice bath) and carefully
`acidified with concentrated hydrochloric acid (7 mL) with
`stirring. The precipitated di-tert-butyl phosphate was filtered,
`washed with ice-cold water (10 mL), and dissolved in acetone
`(100 mL). To this stirred solution in an ice bath was added 1
`mol equiv of a 10% water solution of tetramethylammonium
`hydroxide (24.2 mmol), and the resulting, homogeneous solu-
`tion was evaporated under reduced pressure to give 7.16 g of
`solid. Crystallization from refluxing dimethoxyethane gave
`6.52 g of the pure tetramethylammonium di-tert-butyl phos-
`phate (6.52 g, 57%) as white hygroscopic crystals. To a
`refluxing solution of this phosphate (3.59 g, 12.70 mmol) in
`dimethoxyethane (70 mL) was added chloroiodomethane (25
`g, 141.74 mmol), and the resulting reaction mixture was
`refluxed for 1.5 h. The mixture was filtered, and the filtrate
`was placed under reduced pressure to remove excess chlor-
`oiodomethane and solvent to yield a viscous yellow oil, which
`was a mixture of 11 and bis-di-tert-butyl methyl phosphate.
`The bis-di-tert-butyl methyl phosphate product was identified
`by NMR and mass spectral analysis (data not shown). The two
`products were easily separated via flash column chromatog-
`raphy (eluent 30% ethyl acetate and 70% hexane). 11 was
`isolated as a pale-gold oil (2.07 g, 35%): 1H NMR (CDCl3, 300
`MHz) δ 1.51 (s, 12H), 5.63 (d, 2H, J ) 15); 31P NMR (CDCl3,
`300 MHz) δ -13.9 (s); MS (FAB+, GLY) m/z 259 (M + 1). Anal.
`(C9H20ClO4P) C, H: calcd 41.79, 7.79; found 41.60, 7.68.
`N-(Phosphonooxymethyl)quinuclidinium Trifluoro-
`acetate (6). To a solution of 11 (0.16 g, 0.64 mmol) in
`anhydrous acetonitrile (5 mL) maintained under argon was
`added quinuclidine (0.07 g, 0.636 mmol). The reaction mixture
`was capped with a rubber septum and stirred at 37 °C for 12
`h. The mixture was then placed under reduced pressure to
`remove solvent. To the residue was added anhydrous ethyl
`ether (5 mL) with stirring to create a suspension that was
`centrifuged, and the supernatant was removed. This process
`was repeated three times. The remaining solid was dried under
`vacuum to give di-tert-butyl N-(phosphonooxymethyl)quinu-
`clidinium chloride (0.18 g, 78%) isolated as a white amorphous
`solid: mp 88-105 °C; 1H NMR (CDCl3, 300 MHz) δ 1.54 (s,
`18H), 2.07 (m, 6H), 2.27 (m, 1H), 3.86 (t, 6H, J ) 7.9), 5.36 (d,
`2H, J ) 8.4); 31P NMR (CDCl3, 500 MHz) δ -14.97 (t, J )
`19.4); MS (FAB+, GLY) m/z 334 (M+).
`Deprotection of di-tert-butyl N-(phosphonooxymethyl)qui-
`nuclidinium chloride (0.17 g, 0.46 mmol) was by the addition
`of trifluoroacetic acid (40 μL, 0.52 mmol) in benzene (5 mL)
`with stirring at room temperature for 24 h. The reaction vessel
`was then placed under reduced pressure to remove excess
`trifluoroacetic acid, HCl, and benzene to afford the title
`compound (0.14 g, 94% yield) as a white amorphous solid: 1H
`NMR (D2O, 300 MHz) δ 1.97 (m, 6H), 2.18 (m, 1H), 3.40 (t,
`6H, J ) 7.9), 4.68 (d, 2H, J ) 6.8); 19F NMR (D2O, 500 MHz)
`δ -77.98 (s); 31P NMR (D2O, 500 MHz) δ -2.6 (s); MS (FAB+,
`GLY) 222 (M+).
`N-(Phosphonooxymethyl)cinnarizinium Trifluoroac-
`etate (7). To a solution of 11 (0.178 g, 0.74 mmol) in anhydrous
`acetonitrile (4 mL) maintained under argon were added
`cinnarizine (0.227 g, 0.62 mmol) and 1,2,2,6,6-pentamethylpi-
`peridine (125 μL, 0.74 mmol). The reaction mixture was capped
`with a rubber septum and stirred at 70 °C for 6 days. The
`mixture was then placed under reduced pressure to remove
`the solvent. To the residue was added anhydrous ethyl ether
`(5 mL) with stirring to create a suspension, which was then
`centrifuged, and the supernatant was removed. This process
`was repeated three times. The product was then purified using
`preparative thin-layer chromatography (eluent 75% methylene
`chloride and 25% methanol). The Rf value of the intermediate
`was 0.7. The mono-tert-butyl N-(phosphonooxymethyl)cin-
`narizinium chloride (0.033 g, 8%) was isolated as a white
`amorphous solid: 1H NMR (acetonitrile-d3, 300 MHz) δ 1.35
`
`Patent Owner, UCB Pharma GmbH – Exhibit 2013 - 0004
`
`

`
`3098 Journal of Medicinal Chemistry, 1999, Vol. 42, No. 16
`
`Krise et al.
`
`(s, 9H), 2.70 (m, 4H), 3.39 (m, 2H), 3.56 (m, 2H), 4.12 (d, 2H,
`J ) 7.8), 4.46 (s, 1H), 5.01 (d, 2H, J ) 8.4), 6.4 (m, 1H), 6.95
`(d, 1H, J ) 16), 7.3 (m, 15H); 31P NMR (acetonitrile-d3, 500
`MHz) δ -4.9 (s); MS (FAB+, GLY) m/z 535 (M+).
`This product (0.027 g, 0.048 mmol) was mixed with trifluo-
`roacetic acid (20 μL, 0.26 mmol) in benzene (1 mL) with
`stirring at room temperature for 24 h to remove the tert-butyl
`protecting group. The reaction was then placed under reduced
`pressure to remove excess trifluoroacetic acid, HCl, and
`benzene to yield 7 (0.025 g, 87%) as a white amorphous solid
`(sample had 3.7% residual water): mp 140-146 °C; 1H NMR
`(D2O, 300 MHz) δ 2.98 (m, 4H), 3.58 (m, 4H), 4.23 (d, 2H, J )
`7.7), 4.72 (s, 1H), 4.98 (d, 2H, J ) 6.2), 6.3 (m, 1H), 7.01 (d,
`1H, J ) 15) 7.2-7.6 (m, 15H); 31P NMR (acetonitrile-d3, 500
`MHz) δ 2.1 (s); MS (FAB+, GLY) m/z 479 (M+). HPLC analysis
`showed a single major peak accounting for the majority of the
`total peak peak. There was a small peak corresponding to
`cinnarizine which accounted for <1% of the total peak area.
`N-(Phosphonooxymethyl)loxapinium Trifluoroace-
`tate (8). To a solution of 11 (0.24 g, 0.91 mmol) maintained
`under argon in anhydrous acetonitrile (1 mL) were added
`loxapine (0.20 g, 0.61 mmol) and 1,2,2,6,6-pentamethylpiperi-
`dine (500 μL, 2.76 mmol), and the reaction mixture was capped
`with a rubber septum and stirred at 50 °C for 64 h. The
`mixture was then placed under reduced pressure to remove
`the solvent. To the residue was added anhydrous ethyl ether
`(5 mL) with stirring to create a suspension that was centri-
`fuged, and the supernatant was removed. This process was
`repeated three times. The product was then purified using
`preparative thin-layer chromatography (eluent 90% methylene
`chloride and 10% methanol). The product had an Rf value of
`0.3. The mono-tert-butyl N-(phosphonooxymethyl)loxapinium
`chloride was isolated as a white solid (0.08 g, 25%): 1H NMR
`(D2O, 300 MHz) δ 1.41 (s, 9H), 3.22 (s, 3H), 3.4-3.8 (m, 6H),
`3.85 (m, 2H), 5.02 (d, 2H, J ) 7.6) 7.20 (m, 5H), 7.42 (m, 2 H);
`31P NMR (D2O, 500 MHz) δ -5.7 (t, J ) 17.5); MS (FAB+,
`NBA) m/z 494 (M+).
`The mono-tert-butyl-protected N-(phosphonooxymethyl)loxa-
`pinium chloride (0.08 g, 0.153 mmol) was treated with a
`solution of trifluoroacetic acid (60 μL, 0.78 mmol) in benzene
`(4 mL) at room temperature for 24 h to remove the remaining
`tert-butyl protecting group. The reaction was then placed under
`reduced pressure to remove excess trifluoroacetic acid, HCl,
`and benzene to yield 8 (0.066 g, 76%) as a white amorphous
`solid (sample had 4.2% residual water): mp 114-145 °C; 1H
`NMR (D2O, 300 MHz) δ 3.27 (s, 3H) 3.4-4.2 (m, 8H), 5.08 (d,
`2H, J ) 7.2), 7.10-7.45 (m, 7H); 31P NMR (D2O, 500 MHz) δ
`-1.77 (m); MS (FAB+, TG) m/z 438 (M+). HPLC analysis
`showed a single major peak accounting for 97% of the total
`peak area.
`N-(Phosphonooxymethyl)amiodaronium Trifluoroac-
`etate (9). To a solution of amiodarone (0.275 g, 0.42 mmol),
`11 (0.217 g, 0.84 mmol), and 1,2,2,6,6-pentamethylpiperidine
`(152 μL, 0.84 mmol) maintained under argon in anhydrous
`acetonitrile (3 mL) was added sodium iodide (5 mg, 33 μmol),
`and the resulting reaction mixture was capped with a rubber
`septum and stirred at 40 °C for 24 h with protection from light.
`The mixture was then placed under reduced pressure to
`remove the solvent. To the residue was added anhydrous ethyl
`ether (5 mL) with stirring to create a suspension that was
`centrifuged, and the supernatant was removed. This process
`was repeated three times. After vacuum-drying the di-tert-
`butyl-protected N-(phosphonooxymethyl)amiodaronium chlo-
`ride was obtained as a white amorphous solid (0.180 g, 47%):
`1H NMR (CDCl3, 300 MHz) δ 0.92 (t, 3H, J ) 7.3), 1.30-1.85
`(m, 28H), 2.89 (t, 2H, J ) 7.7), 3.88 (q, 4H, J ) 4.3), 4.4-4.6
`(c, 4H), 5.47 (d, 2H, J ) 7.4), 7.3 (m, 2H), 7.49 (d, 2H, J )
`8.1), 8.21 (s, 2H); 31P NMR (CDCl3, 500 MHz) δ -12.34 (t, J )
`17.2); MS (FAB+, NBA) m/z 868 (M+).
`The di-tert-butyl-protected N-(phosphonooxymethyl)amio-
`daronium chloride (0.154 g, 0.17 mmol) was treated with
`trifluoroacetic acid (60 μL, 0.78 mmol) in benzene (5 mL) at
`room temperature for 24 h to remove the tert-butyl groups.
`
`The reaction was then placed under reduced pressure to
`remove excess trifluoroacetic acid, HCl, and benzene to yield
`9 as a yellow oil. This was dissolved in water (5 mL) containing
`a two molar excess of sodium bicarbonate (0.026 g, 0.31 mmol)
`to form the disodium salt. The aqueous solution was then
`lyophilized to remove the water to yield N-(phosphonooxy-
`methyl)amiodaronium sodium salt (0.16 g, 94%) as a white
`hygroscopic solid. Residual water content was variable due to
`the hygroscopic nature of the sample but corresponded ap-
`proximately to a pentahydrate: 1H NMR (DMSO-d6, 300 MHz)
`δ 0.84 (t, 3H, J ) 7.2), 0.98 (m, 2H), 1.31 (m, 6H), 1.68 (m,
`2H), 2.74 (m, 2H), 3.54 (m, 4H), 3.84 (m, 2H), 4.36 (m, 2H),
`4.95 (d, 2H, J ) 8.7), 7.22-7.65 (m, 4H), 8.18 (s, 2H); 31P NMR
`(D2O, 500 MHz) δ 4.77 (s); MS (FAB+, NBA) m/z 756 (M+).
`In Vitro/in Vivo Evaluation. Materials: 7 and 8 were
`synthesized using the previously described procedure. Human
`placental alkaline phosphatase, type XVII (14 units/mg), and
`meclizine were obtained from Sigma Chemical Co. (St. Louis,
`MO). The synthesis and characterization for (SBE)4M-(cid:3)-CD has
`been previously described.30 Magnesium chloride hexahydrate,
`tetrabutylammonium dihydrogen phosphate, and zinc chloride
`were obtained from Aldrich Chemical Co. (Milwaukee, WI).
`Glycine was obtained from Fisher Scientific (Pittsburgh, PA).
`All other chemicals and solvents were of reagent grade and
`were used without further purification.
`In Vitro Enzymatic Evaluation. 1. Procedure: All
`experiments involving alkaline phosphatase were performed
`in a pH 10.4 glycine buffer at 37 °C. The glycine buffer solution
`contained 1 mM ZnCl2, 1 mM MgCl2, and 0.1 M glycine. The
`final pH of the buffer was adjusted to pH 10.4 with additions
`of a 2 N NaOH solution. An alkaline phosphatase stock
`solution was prepared at a concentration of 1.54 units/mL in
`the glycine buffer. The enzyme solution was brought to 37 °C
`and spiked with a prodrug stock solution to give an initial
`prodrug concentration of 264 and 230 μM for 8 and 7,
`respectively. Aliquots (0.2 mL) were separated into individual
`4-mL glass tubes, capped, and placed in a 37 °C shaking water
`bath. Samples were removed from the bath at predetermined
`time points (0, 5, 10, 15, 30, 45, 60, 75, and 90 min) and spiked
`with acetonitrile (0.2 mL) which served to quench the enzy-
`matic reaction and to solubilize the reaction products. As a
`test for the chemical stability of the prodrugs under these
`experimental conditions, the previously described procedure
`was repeated with the removal of enzyme from the media. All
`samples were analyzed by HPLC for both parent and prodrug
`concentrations.
`2. Analytical: The HPLC system for all compounds con-
`sisted of a Waters model 510 pump (Milford, MA), a Waters
`717 autosampler (Milford, MA), an LDC analytical spec-
`tromonitor 3100 variable wavelength detector, a Waters C18
`symmetry column (3.9 × 150 mm; Milford, MA), a Shimadzu
`CR6A integrator (Kyoto, Japan), and a column heater (Tim-
`berline Instruments Inc., Boulder, CO).
`For the analysis of cinnarizine and 7, the mobile phase
`consisted of acetonitrile (55% v/v) and a 10 mM ammonium
`dihydrogen phosphate buffer adjusted to pH 3 with phosphoric
`acid (45% v/v) and was pumped at a flow rate of 0.9 mL/min.
`The injection volume was 20 μL, and the detection was
`ultraviolet using a wavelength of 254 nm. Under these
`conditions, the retention times were 5 and 9 min for 7 and
`cinnarizine, respectively.
`For the analysis of loxapine and 8, the mobile phase
`consisted of acetonitrile (28% v/v) and a 10 mM ammonium
`dihydrogen phosphate buffer adjusted to pH 3 with phosphoric
`acid (72% v/v) and was pumped at a flow rate of 0.9 mL/min.
`The injection volume was 20 μL, and the detection was
`ultraviolet using a wavelength of 251 nm. Under these
`conditions, the retention times were 8 and 12 min for 8 and
`loxapine, respectively.
`In Vivo Dog Study. 1. Preparation of formulations:
`Cinnarizine for iv injection was prepared in a 10 mM phos-
`phate buffer solution at pH 4.5 at a concentration of 12.5 mg/
`10 mL (3.39 mM) along with 37.5 mM (SBE)4M-(cid:3)-CD as a
`
`Patent Owner, UCB Pharma GmbH – Exhibit 2013 - 0005
`
`

`
`Novel Prodrug Approach for Tertiary Amines
`
`Journal of Medicinal Chemistry, 1999, Vol. 42, No. 16 3099
`
`solubilizing excipient. The cyclodextrin solution (10 mL) was
`prepared first and the pH adjusted to 3.5 with HCl. Cinnariz-
`ine was then added, and the solution was then sonicated for 3
`h and subsequently stirred overnight. The pH was then
`adjusted to 4.5 with NaOH along with the addition of 29 mg
`of NaCl to adjust for isotonicity. The solution was subsequently
`passed through a 0.22-μm nylon membrane filter just prior to
`iv administration. The prodrug injection was prepared by
`dissolving 16.97 mg of 7 in 10 mL of 0.9% NaCl sterile solution
`for injection (3.39 mM). The solution was passed through a
`0.22-μm nylon membrane filter just prior to injection.
`2. Experimental procedure: The evaluation of the cin-
`narizine plasma concentration versus time was conducted
`using a male beagle dog weighing 11.1 kg. The dog received
`the cinnarizine injection followed by a 2-week washout period
`and then received the prodrug injection in an equal molar
`quantity. Samples were taken from the dog prior to dosing (10-
`mL blank plasma) and 2, 6, 10, 20, and 40 min and 1, 2, 4, 6,
`and 8 h postdosing (3 mL each). Blood samples were drawn
`from either cephalic, saphenous, or jugular vein. The samples
`were centrifuged for 10 min, and 1 mL of plasma was
`separated, collected, and frozen at -20 °C prior to extraction.
`The dog received a regular diet between experiments and was
`fasted the day of the experiment. The trapezoidal calculation
`was used to approximate the area under the plasma concen-
`tration versus time profile from 0 to 8 h.
`3. Analytical procedure: The procedure for the analysis
`of cinnarizine from dog plasma was adapted from a procedure
`described by Ja¨rvinen et al.31 To each 1 mL of plasma was
`added 100 μL of acetonitrile-H2O (70:30, v/v), 100 μL of
`acetonitrile-H2O (70:30,v/v) containing 40 μg of meclizine
`(internal standard), and 100 μL of 0.5 M HCl. The plasma
`sample was then vortexed for 30 s; 1 mL of carbon tetrachlo-
`ride was then combined with the sample followed by an
`additional minute of vortexing. Samples were then centrifuged
`for 10 min after which the carbon tetrachloride was removed
`and evaporated to dryness under a stream of dry nitrogen. The
`residue was then redissolved in 200 μL of acetonitrile-H2O
`(70:30,v/v) prior to sampling by HPLC. A six-point standard
`curve was obtained using the above procedure with addition
`of cinnarizine in known amounts to give plasma concentrations
`in the range of 5.5-550 ng/mL. The calibration curve was
`calculated using peak area ratios (cinnarizine peak area
`divided by meclizine peak area). A t