6614
`
`J. Med, Chem. 2006, 49, 6614—6620
`
`Synthesis and Pharmacokinetics of Valopicitabine (NM283), an Efficient Prodrug of the Potent
`Anti-HCV Agent 2’-C-Methylcytidine
`
`Claire Pierra,t Agnés Amador,' Samira Benzaria,t Erika Cretton-Scott,! Mare D’Amours,’ John Mao,! Steven Mathieu,?
`Adel Moussa,! Edward G. Bridges,! David N. Standring,' Jean-Pierre Sommadossi,! Richard Storer,’ and Gilles Gosselin*+
`Laboratoire Cooperatif Idenix-CNRS— Université Montpellier Il, Case Courrier 008, Université Montpellier If, Place Eugene Bataillon, 34095
`Montpellier Cedex 5, France, and Idenix Pharmaceuticals Inc., 60 Hampshire Street, Cambridge, Massachusetts 02139, and Laboratoires
`Idenix SARL, Laboratoire de Chimie Médicinale, Cap Gamma, 1682 Rue de la Valsiere, BP 50001, 34189 Montpellier Cedex 4, France
`
`Received March 28, 2006
`
`In our search for new therapeutic agents against chronic hepatitis C, a ribonucleoside analogue,
`2’-C-methylcytidine, was discovered to be a potent and selective inhibitor in cell culture of a number of
`RNAviruses, including the pestivirus bovine viral diarrhea virus, a surrogate model for hepatitis C virus
`(HCV), and three flaviviruses, namely, yellow fever virus, West Nile virus, and dengue-2 virus. However,
`pharmacokinetic studies revealed that 2’-C-methylcytidine suffers from a low oral bioavailability. To overcome
`this limitation, we havesynthesized the 3’-O-L-valinyl ester derivative (dihydrochloride form, valopicitabine,
`NM283)of 2’-C-methylcytidine. Wedetail herein for the first time the chemical synthesis and physicochemical
`characteristics of this anti-HCV prodrug candidate, as well as a comparative study of its pharmacokinetic
`parameters with those of its parent nucleoside analogue, 2’-C-methylcytidine.
`
`®O
`O°
`NH,
`Cl
`q
`SN
`NH
`;
`C
`LA
`wo
`Homo. N°
`20=-/
`Hts
`“Fay
`HO
`OH
`Oxe3
`OH
`2'-C-methyluridine(2) ot=i
`
`HO
`
`3
`
`3-O-valinyi ester of
`2'-C-methylcytidine (3)
`(dihydrochloride salt)
`
`HOF
`
`NH
`>
`‘e
`a
`on
`L Hic?
`HO
`OH
`2'-C-methylcytidine(1)
`
`Introduction
`
`Hepatitis C virus (HCV) hasinfected an estimated 170 million
`individuals, 3% of the world’s population.’ The virus establishes
`a persistent infection in the majority of cases, leading to chronic
`hepatitis that often develops into cirrhosis and, in many cases,
`causes hepatocellular carcinoma.! There is no vaccine available
`against HCV, and current
`therapies, namely, pegylated or
`nonpegylated interferon-c (IFN-c) monotherapy and combina-
`tion of IFN-c with oral ribavirin, are expensive, often poorly
`tolerated, and effective only in half of the patient population.27
`Therefore, there is an urgent need to develop new and more
`effective therapies in response to this important unmet medical
`need,*
`
`In the course of our HCV program, we recently discovered
`that 2’-C-methylcytidine (1, Figure 1) is a potent and selective
`inhibitor of Flaviviridae virus replication in cell culture.More
`particularly, 1 inhibited the replication of the bovine viral
`diarrhea virus (BVDV,a pestivirus surrogate model for HCV),’
`eliminated persistent BVDV infection at nontoxic concentra-
`tions, and was synergistic in combination with interferon-o,
`but not with ribavirin.’ Compound 1 has no activity against
`human immunodeficiency virus (HIV)or against DNA viruses.
`In contrast to ribavirin, which is not effective alone in reducing
`vital RNAlevels but stimulates the immune boosting capacity
`of interferon-a when used in combination, 1 is the first example
`of an anti-RNA virus agent that is active via a nucleoside
`analogue mode of action. Thus, in primary human hepatocyte
`cultures, in a human hepatomacell line (HepG2), and in a bovine
`kidney cell
`line (MDBK), 1 is converted into its major
`metabolite, 2’-C-methylcytidine-5’-triphosphate, along with
`smaller amounts of 2’-C-methyluridine-5‘-triphosphate, resulting
`from deamination. The active metabolite 2’-C-methylcytidine
`
`
`* Corresponding author. Tel: + 33-4-67143855. Fax: + 33-4-67549610.
`E-mail: gosselin@univ-montp2.fr.
`¥ Université Montpellier I.
`1 Tdenix Pharmaceuticals Inc.
`§ Laboratoires Idenix SARL.
`¥ Present address: VASToxplc, 9} Milton Park, Abingdon, Oxfordshire,
`OX14 4RY, U.K.
`
`Figure 1. Structures of compounds 1—3.
`
`triphosphate is a competitive inhibitor of purified BVDV RNA
`polymerase in vitro (Kj = 160 nM).°
`Preliminary pharmacokinetic studies in animals revealed that
`further development of 1 would be hampered byits low oral
`bioavailability. To overcome this limitation, we devoted our
`efforts to the design of 2’-C-methylcytidine prodrugs with more
`favorable oral absorption profiles. From a literature survey, it
`was found that a broad variety of amino acid ester derivatives
`have been studied and successfully employed as nucleoside
`prodrug forms, Moreparticularly, efficacy of such derivatives
`has been proved in the case of valacyclovir, the L-valiny] ester
`of acyclovir (Figure 2).!%'! After active absorption via peptide
`transport mechanism, valacyclovir is rapidly and almost com-
`pletely converted into acyclovir by enzymatic hydrolysis,
`increasing considerably the oral bioavailability and celular
`uptake of the parent drug.’°" The efficacy of L-valinyl
`derivatives has been also demonstrated in the case of ganciclovir
`(GCV), since valganciclovir (L-valinyl ester of GCV, Figure 2)
`has an oral bioavailability 10-fold higher than the parent
`nucleoside.'23 Recently, as part of our hepatitis B program,
`we have synthesized and studied several L-valiny! ester prodrugs
`of 2’-deoxy-f-L-cytidine (1-dC) in order to improve the oral
`bioavailability of L-dC. Among them, the 3’-O-L-valinyl ester
`of L-dC (Val-L-dC,valtorcitabine, Figure 2) emerged as the most
`attractive L-dC prodrug! and is currently in phaseII clinical
`studies.'5
`Onthe basis of these considerations, we decided to synthesize
`and study the 3’-O-.-valinyl ester derivative (NM283, valop-
`
`10.102 1/jm0603623 CCC: $33.50
`© 2006 American Chemical Society”
`Published on Web 10/06/2006
`s
`
`CLARK EXHIBIT 2125
`Sommadossi v. Clark
`Contested Case 105,871
`
`GIL2018
`I-MAK, INC. V GILEAD PHARMASSETLLC
`IPR2018-00120
`
`1
`
`GIL2018
`I-MAK, INC. V GILEAD PHARMASSET LLC
`IPR2018-00120
`
`

`

`Synthesis and Pharmacokinetics of Valopicitabine
`°
`
`qe®
`
`N
`e
`O
`N7
`HgN. tl
`L.~ loo
`:
`Valaciciovir
`
`| ae
`SN
`>NH2,
`
`o
`
`® 0
`NH,
`Ci
`>
`Rea ,
`=
`Nagy _OH
`j
`
`oO
`
`NH
`
`Ox
`
`fe)
`
`.
`
`N
`if
`
`HO
`
`eo,
`9
`ll
`
`
`Cl HNg5og ON v7 20 iy
`E
`~/
`crm
`H3N"
`Val-L-dc
`
`Journal of Medicinal Chemistry, 2006, Vol. 49, No. 22 6615
`
`fluoride in tetrahydrofuran (TBAF/THF), followed by acidic
`hydrolysis using a saturated solution of hydrogen chloride
`in ethyl acetate,”* led to the target prodrug 3 as a dihydro-
`chloride salt.
`
`Although this conventional synthetic route afforded the
`desired prodrug in a satisfactory overall yield of 52%, several
`chemical difficulties (instability of the formamidine intermediate
`4 and partial racemization of the L-valine during the coupling
`step), as well as cost issues (use of expensive feri-butyldiphe-
`nylsilyl group), made scale-up using this initial route unattrac-
`tive. In our search for a more direct and scalable synthetic
`process, we then developed a new synthetic strategy based on
`the direct condensation of 1 with L-Boc-valine, followed by
`removal of the N-zert-butyloxycarbony! group, thus reducing
`the number ofthe synthetic steps from 5 to 2 (Scheme 2). We
`decided to explore this direct coupling approach after examining
`a report by McCormick et al, who described the room-
`temperature reaction of unprotected guanosine with Boc-
`anhydride in the presenceoftriethylamine, DMAP, and DMSO.-
`This reaction selectively gave 3’-Boc-guanosinein 55% yield.23
`In our first attempt, we reacted compound 1 in DMSO with
`carbonyldiimidazole(CDD-activated L-Boc-valine in the presence
`of triethylamine and DMAP. Thereaction did not progress at
`all when kept at room temperature. The reaction produced the
`desired product and several byproducts when heated at 50 °C
`using 4 equiv of CDI-activated L-Boc-valine. Optimum selectiv-
`ity was achieved by using 1.1 equiv of CDI-activated L-Boc-
`valine and runningthe reaction for only 1 h at 80 °C. We used
`only 0.1 equiv of DMAPin order to avoid the racemization of
`the L-valine. Other coupling reagents such as EDC and N,N’-
`dicyclohexylcarbodiimide (DCC). did not produce desired
`product, and DMF was difficult to change due to the insolubility
`of compound 1 in most organic solvents, HPLC analysis showed
`68% of the desired compound, 11% of starting nucleoside 1,
`and the two 3’,5’/3’,N*-divalinyl ester byproducts. Pure product
`3 was obtained in 54% yield and 99% purity by using simple
`acid—baseextraction, eliminating the extensive chromatographic
`step. Thus, this tume- and labor-saving alternative process could
`be accomplished without
`involvement of chromatographic
`
`Valganciclovir
`
`~
`
`Figure 2, Structures of valacyclovir, valganciclovir, and val-L-dC.
`
`in order to improve the oral
`icitabine, 3; Figure 1) of 1
`bioavailability of the parent nucleoside.
`:
`Results and Discussion
`
`Synthesis. The 3’-O-valinyl ester prodrug of 2’-C-methyl-
`cytidine (3) was first prepared by the route described in Scheme
`1. Our multistep sequence involved successive protections of
`the exocyclic amino function and of the 5’-hydroxyl group of
`1, followed by condensation with N-tert-butyloxycarbonyl-L-
`valine (L-Boc-valine) and,
`finally,
`total deprotection. N,N-
`Dialkylformamidinesand especially N,N-dimethylformamidine
`have been widely used as protective groups for exocyclic amino
`function of nucleosides.'®!? In the present work, the synthesis
`of N*-[(dimethylamino)methylene]-2’-C-methyleytidine (4) was
`carried out following a procedure described by Kerr et al for
`the protection of 1 using N,N-dimethylformamide dimethylac-
`etal.'® Selective-silylation of the 5’ primary hydroxy! group using
`a reported method! led to the protected key intermediate 5.
`Condensation of 5 with L-Boc-valine using the coupling agent
`N’-(3-dimethylaminopropyl)-N-ethylcarbodiimide hydrochloride
`(EDC),?° with 4-(dimethylamino)pyridine (DMAP) asa cata-
`lyst,!! provided the N-Boc-protected ester derivative 6. Finally,
`desilylation of 6 with excess ammonium fluoride (NH4F) in
`methanol,?! an economical alternative to tetrabutylammonium
`
`Scheme 1. A Conventional Route for the Synthesis of 37
`NH,
`(oy
`HO 0 Nv
`ars
`HO
`OH
`1
`
`Oo
`
`athe
`
`HO
`
`ae
`"
`
`o
`
`=e: ;
`Hat
`HO
`4
`(80% first crop
`
`crystatlization)
`
`WONG
`.
`oN
`CO
`ls, INTO
`pao:
`Co , AG
`|
`~" HO
`OH
`5
`
`ii
`
`| iii
`
`®
`NH,
`‘en
`SN
`HOY 9. NO
`Hgts
`—/
`Oxy®
`OH
`oe, ’ i
`;
`3'-O-valinyl ester of 1
`(hydrachloride farm, 3, 81%)
`
`Ct
`
`NH3
`CY
`SN
`HO 9. NO
`
`c
`v
`7
`9 0x0
`oon i
`
`iv
`
`NANG
`t
`“Sy
`0. no
`-+si-o
`Hay
`(
`O
`6 O9 OH
`os my
`x NH
`
`7
`(78% for 3 steps)
`(i) Mez2NCH(OMe)2, DMF,rt, 1.5 h; (ii) TBDPSCI, imidazole, pyridine,rt, 6 h; (iii) N-Boc-L-valine, EDC, DMAP, CH3CN/
`@ Reagents and conditions:
`DMF,tt, 2 d; (iv) NHaF, McOH,reflux, 3 h; (v) HCI/EtOAc, EtOAc,rt.
`
`2
`
`€
`

`

`6616
`
`Journal of Medicinal Chemistry, 2006, Vol. 49, No. 22
`
`Pierra et al.
`
`Scheme 2. Alternative Route for the Synthesis of 3¢
`
`NH
`
`NH»
`
`be
`Ce
`Hoy gp NO Hoya NO
`ean
`—iil
`Tas
`HO
`OH
`9 On 0
`OH
`J. Aly
`x? NH}
`7
`(54%)
`
`1
`
`®
`NH;
`
`Co
`INO
`
`it
`
`HOsuo.
`Age
`0.0 OH
`aoe
`i 4
`cl
`NH3”
`|
`3'-O-valinyl ester of4
`(dihydrochloride sait, 3,
`
`Cl
`
`JTable 1. Pharmacokinetic Parameters of Compounds 1 and 3,
`Following a Single Oral Administration of 3 to Sprague-Dawley Rats*
`prodrug3
`compd 1
`dose
`equiv dose
`AUC
`Tin
`Cmax
`Tmax
`SF
`
`(mg/kg)
`(mg/kg)
`drug
`(wgb/mL)
`(bh)
`(agin)
`(bh)
`(%)
`100
`72
`3
`8.95
`0.64
`3.62
`i100
`-
`30.0 7.10 6.12 2.0] 33.6
`
`
`
`
`
`
`# AUCy-, area under the plasma concentration~time curve from time
`0 to 4 Tro:
`terminal elimination half-life. Cyax: maximum plasma
`concentration. Tinax;
`time to maximum plasma concentration. F: apparent
`oral bioavailability calculated on the basis of a dose-normalized AUC value
`of 1.24 (ug b/mL) from a single iv dose of 1 to the rat.
`
`@ Reagents and conditions:
`THF; (Gi) HCVEtOH.
`
`(i) N-Boc-L-valine, CDI, DMAP, TEA, DMF/
`
`At pH 4.5
`
`
`
` Relativeconcentration
`
`0
`
`at
`“O-~
`
`3in=3)
`4(n=3)
`
`
`
`
`Z
`
`2s
`
`s 5a Ga
`
`|
`|
`S
`“160
`gO]
`
`{
`
`4
`
`
`+
`' ag
`8
`is
`ee
`32
`40
`48
`Time (h)
`
`Figure 4, Mean plasma concentration of compounds 1 and3, following
`a single oral administration of 3 to Sprague-Dawleyrats.
`
`human liver cytosol, and 89 fractions are not surprising and
`depend on the activity/quantity of enzyme(s) present in these
`samples. In all cases, no further metabolism of 1, such as
`deamination to its corresponding uridine metabolite 2, has been
`observed.
`Protein Binding of Compounds 1 and 3 to Rat Plasma.
`Thein vitro binding of compounds 1 and 3 to rat plasmaproteins
`was investigated bytheultrafiltration method using radiolabeled
`test articles ((3H]-2’-C-methylcytidine and [°H]-3’-O-valinyl
`ester of 2’-C-methyleytidine) at a concentration of 20 “M for
`each compound.As expected for nucleosides, the protein binding
`for the two compounds was low, with 7% and 5% boundfor 3
`and 1, respectively.
`Oral Bioavailability and Pharmacokinetic Studies of
`Compounds 1 and 3. The pharmacokinetics (PK) of prodrug
`3 were evaluated in Sprague—Dawley rats in a mass balance
`using radiolabeled 3. Both intact and bile duct cannulated (BDC)
`rats were used in this study. PK parameters of 1 and its 3’-O-
`valinyl ester 3 are presented in Table 1, Plasma concentration—
`time profiles of 1 and 3 are shown in Figure 4.
`Following a single oral dose of ['4C]-3’-O-valinyl ester of
`2’-C-methylcytidine (3) at 100 mg/kg (as free base), urine, feces,
`and cage rinses were collected for a period of 72—168 h
`postdoseto evaluate the excretion of radioactivity and the overall
`mass balance. Plasma samples were collected over a 48-h period
`postdose to study the PK of1 and3.In intactrats, fecal excretion
`was also the predominant route of elimination of total radio-
`activity, accounting for 63.5% of the administered dose.
`Excretion of radioactivity occurred largely within the first 24 h
`(averaging 60% of the dose). Urinary excretion accounted for
`31.9% of the administered dose. The mean total recovery of
`radioactivity, in urine, feces, and cage rinses, was 96.6%. Bilary
`excretion oftotal radioactivity in BDC rats accounted for a mean
`
`At pH 7.2
`
`10
`
`
`
`Relativeconcentration o >
`
`0
`
`§00 1000 1600 2000
`
`(
`
`Incubation time (min)
`®o
`s
`NH,
`Cl
`N
`
`HOq of
`Hats
`Hi
`OH
`A= 2-C-methylcytidine (1)
`Kinax = 273.6 am
`
`5000 10000 15000 20000
`tncubation time (min)
`ee
`NH,
`Cl
`Shy
`
`aa ae
`Test
`OH
`
`:
`
`HOS
`on?
`ee
`cl
`HGNT TT
`O = 3-0-valinyl ester of 2'-C-methylcytidine (3)
`A max 5 273.5 1M
`
`Tage 6.1 days at pH 4.5
`Tae 3.9 hours atpH7.2
`
`Figure 3. Kinetic curves ofhydrolysis of prodrug 3 at pH 4.5 and 7.2.
`
`purifications, and compound 3 was obtained in high purity using
`simple crystallization.
`AqueousSolubility, Lipophilicity, and Chemical Stability
`of Prodrug 3. To be considered as a suitable prodrug for oral
`administration, the 3’-O-valinyl ester of 2’-C-methylcytidine (3)
`should possess adequate solubility in aqueous media in order
`to dissolve in the small
`intestine,
`thus being available for
`absorption. The aqueous solubility of 3 has been determined in
`comparison with the solubility of the parent nucleoside 1. Thé
`concentration of saturated solutions of 1 and 3 in water were
`32 and 423 g/L, respectively. Both compounds were found to
`be highly soluble in aqueous media, which has been cor-
`roborated by their low distribution coefficient (log P) values:
`—0.965 and —1.34 for 1 and 3, respectively.
`The main aim ofstability studies was to determine whether
`the prodrug of 2’-C-methylcytidine would be sufficiently
`chemically stable in the gastrointestinal tract before its absorp-
`tion. Prodrug 3 appearedto be fully stable at pH 1.2 but was
`hydrolyzed at pH 4.5 and 7.2 into the parent drug 1 following
`first-order kinetics. The half-lives of 3 at pH 4.5 and 7.2 were
`6.1 days and 3.9 h,respectively (Figure 3). It is noteworthy
`that 1 was fully stable atall the acidic and neutral studied pHs.
`Stability in Human Blood, Plasma, and Liver Cytosol.
`Prodrug 3 was rapidly converted into 1 in both human plasma
`and whole blood, exhibiting in vitro half-life values of 130 and
`40 min, respectively. This rapid conversion is primarily due to
`the presenceof esterases in blood and plasma. Conversion of 3
`into 1 was also observed in humanliver cytosol and $9fractions,
`with approximately 30% of 3 converted into 1 within | h.
`Differences in half-life for prodrug 3 in plasma, whole blood,
`
`3
`
`

`

`Synthesis and Pharmacokinetics of Valopicitabine
`
`Journal of Medicinal Chemistry, 2006, Vol. 49, No. 22 6617
`
`of 0.3% of the dose over the 72-h collection period. In this
`group, fecal excretion was the predominantroute of elimination,
`accounting for a mean 65.9% of the dose. A mean 29.8% of
`the dose was recovered in urine. Excretion of radioactivity was
`largely completed by 24 h in BDC rats. On the basis of the
`combined mean recovery of radioactivity in bile and urine, it
`was apparentthat at least 30.0% ofthe [!4C]-3’-O-valiny] ester
`of 2’-C-methylcytidine (3) oral dose was absorbed. The mean
`recovery of total radioactivity in this group was 97.0% of the
`administered dose, In the group following a single dose of '4C-3
`at 100 mg/kg (as free base), maximum plasma concentrations
`of total
`radioactivity (averaging 11.9 wg equiv/mL) were
`achieved 1—2 h after the oral dose. Plasma concentrations of
`3, determined by LC/MS/MS, were greatest 0.5—1 h after
`dosing, and Cmax averaged 3.62 ug/mL. Plasma 3 concentrations
`decayed with a half-life of 0.64 h, and no prodrug 3 was detected
`in plasma by 12 h, Plasma concentrations of 1, determined by
`LC/MS/MS, were greatest 1-2 h after dosing, and Cinax
`averaged 6.12 zg/mL. Plasma concentrations of 1 decayed with
`a half-life of 7.1 h. The AUC of 1 was morethan 3-fold greater
`than the AUC of 3, 2’-C-Methyluridine (2) was not detected in
`rat plasma (<0.1 wg/mL).
`The pharmacokinetics of 1 were also studied in Sprague—
`Dawleyrats in a toxicokinetic (TK) study in which compound
`1 was administered intravenously daily for 15 days. There were
`three dose groups (60, 120, and 300 mg/kg/day) in the study,
`and TK. sampling was conducted on day | following the first
`dose and on day 15 after the last dose. For the scope ofthis
`paper, only the TK. parameters of the 60 mg/kg/day dose group
`on day | are presented. Followinga single iv dose of compound
`1 to rats at 60 mg/kg, the mean plasma Cmax of 1 was 75.6
`#e/mL and the mean AUC was 74.4 ug h/mL. The mean AUC
`value of 1 from this dose group was compared to the AUC value
`of 1 from the prodrug 3 rat study above, and an apparentoral
`bioavailability was calculated for 1. Thus, the apparent oral
`bioavailability of 1 following oral administrationof the valiny]
`ester derivative 3 at 100 mg/kg (as free base) was calculated to
`be 33.6%.
`
`Conclusion
`
`3’-O-Valinyl ester of 2’-C-methylcytidine (dihydrochloride
`salt, NM283, valopicitabine, 3) has been synthesized in order
`to improve the oral bioavailability of the parent compound 2’-
`C-methylcytidine (1). For that purpose, two different strategies
`have been developed, both starting from 1. The first one is a
`conventional route that involves successive protection steps, and
`the second one is more appropriate for large-scale synthesis and
`is based on selective 3’-O-esterification. Physicochemical,
`pharmacokinetic, and toxicokinetic studies have shown that
`compound 3 is an acid-stable prodrug of 1 with excellent
`pharmacokinetic and toxicokinetic profiles, Prodrug 3 is rapidly
`converted into compound 1 in both human plasma and whole
`blood, probably due to the presence of esterases. The apparent
`oral bioavailability of the parent drug 1 following oral admin-
`istration of the prodrug 3 is 34%.
`Thus,on thebasis ofits ease of synthesis, its physicochemical
`properties, and pharmacokineticprofile, 3’-O-valinyl ester 3 has
`emerged as a promising prodrug of 2’-C-methyleytidine (1). The
`. 3’-O-valinyl ester of 2’-C-methyleytidine (3, NM283, valopic-
`itabine) is currently undergoing phase IIb clinicaltrials for the
`treatment of chronic HCV infection.”
`
`Experimental Section
`'H NMR spectra were
`General Methods for Chemistry.
`recorded at ambient temperature on a Bruker AC 200 MHz, 250,
`
`'H NMR chemical shifts (4) are
`300, or 400 MHz spectrometer.
`quoted in parts per million (ppm) referencedto the residual solvent
`peak [dimethyl sulfoxide (DMSO-d,)] set at 2.49 ppm. The accepied
`abbreviations are as follows: s, singlet; d, doublet; t, triplet; m,
`multiplet. FAB mass spectra were recorded in the positive-ion or
`negative-ion mode on a JEOL DX 300 mass spectrometer operating
`with a JMA-DA 5000 mass data system and using a mixture of
`glycerol and thioglycerol (1/1, v/v, G--T) as the matrix. Melting
`points were determined in open capillary tubes with a Biicht B-545
`apparatus and are uncorrected, UV spectra were recorded on an
`Uvikon XS spectrophotometer. Elemental analyses werecarried out
`by the Service de Microanalyses du CNRS, Division de Vernaison
`(France). Thin-layer chromatography (TLC) was performed on
`precoated aluminum sheets of silica gel 60 Fos, (Merck, Art. 5554),
`visualization of products being accomplished by UV absorbance
`and by charring with 10% ethanolic sulfuric acid with heating or
`with a 0.2% cthanolic ninhydrin solution for compounds bearing
`an amide function.
`i
`Column chromatography was carried outon silica gel 60 (Merck,
`Art. 9385). Evaporation of solvents was carried out in a rotary
`evaporator under reduced pressure. All moisture-sensitive reactions
`were carried out under rigorous anhydrous conditions under an
`argon atmosphere using oven-dried glassware. Solvents were dried
`and distilled prior to use and solids were dried over POs under
`reduced pressure. Analytical high-performance liquid chromatog-
`raphy (HPLC) studies were carried out on a Waters Associates unit
`(multisolvent delivery system, 717 autosampler injector, 996
`photodiode array detector and a Millenium data workstation) using
`a reverse-phase analytical column (Nova-Pak Silica 60 A, 4 ym,
`Cis, 150. x3.9 mm). The compound to be analyzed was eluted
`using a linear gradient of 0-50% acetonitrile in 20 mM triethy-
`lammonium acetate buffer (TEAC, pH 7) programmed over a 30-
`min period with a flow rate of 1 mL/min.
`General Precedure for Aqueous Solubility Studies. Aqueous
`solubilities were determinedin distilled water at room temperature.25
`An excess of studied compounds was added to aqueoussolution,
`and the suspension was shaken and centrifugated. Samples of
`supernatant were analyzed by HPLC (same conditions as described
`in General Methods for Chemistry), and the concentration of
`saturatedsolution was determined accordingto a calibration curve.
`The experiment was repeated three times.
`General Procedure for Distribution Coefficient Studies.
`Distribution coefficients between 1-octanol and an aqueous phase
`(phosphate buffer solution 0.02 M) were determined at room
`temperature using a shake-flask procedure.An aliquot of a 107?
`M aqueous solution of studied compounds was diluted to 1 mL
`with the aqueous phasepreviously saturated with octanol. An equal
`volume of octanol, previously saturated with the aqueous phase,
`was added to give a total volume of 2 mL. The mixture was shaken
`vigorously. The two phases were centrifugated and separated.
`Samples of each phase were collected and analyzed by HPLC (same
`conditions as described in General Methods for Chemistry). UV
`absorbances for both phases were measured at the respective
`maximum wavelength. The distribution coefficient was calculated
`from theratio of the area of the signal detected in the octanol and
`aqueous phases. Each experiment was repeated twice.
`GeneralProcedure for Chemical Stability Studies. Chemical
`hydrolysis rates were determined in KCI—-HCI buffer 0.135 M
`solution (pH 1.2, 37 °C), acetate buffer 0.02 M solution (pH 4.5,
`37 °C), and phosphate buffer 0.02 M solution (pH 7.2, 37 °C).
`Solutions (10-4 M) of 3 have been incubated at 37 °C in buffer
`solutions. Aliquot samples were collected at different timeintervals
`and analyzed by HPLC(the sameconditions as described in General
`Methods for Chemistry), Rates of decomposition have beeneasily
`determined using a method developed in our laboratory and based
`on pseudo-first-order kinetic models.??
`General Procedure for in Vitro Blood and Plasma Stability
`Studies. A stock solution of 3 was aliquoted (1 mg/mL stored in
`MeOH at —20 °C), dried down, and resuspended to 20 zg/mL with
`either human plasma or whole blood (Bioreclamation Inc., Hicks-
`ville, NY) preheated to 37 °C. Aliquot samples were collected at
`
`4
`
`

`

`6618
`
`Journal of Medicinal Chemistry, 2006, Vol. 49, No, 22
`
`Pierra et al.
`
`3’-O-valinyl ester of 2’-C-methylcytidine, and 1 and 3 were
`different time intervals 0, 2, 5, 10, 20, 40, and 60 min for blood
`evaluated. Blood samples were collected from group 3 male rats
`studies and 0, 30, 70, 100, and 140 minfor plasmastudies. Samples
`predose and 0.25, 0.5, 1, 2, 3, 4, 8, 12, 16, 24, and 48 h postdose
`were quenched.in 5 volumes ofice-cold solvent consisting of 50%
`(3 rats/time point), and the separated plasma was analyzed by LSC
`acetonitrile/50% methanol and centrifuged at 16 000 ref for 3 min.
`and by LC/MS/MS. The LC/MS/MS method was validated ac-
`Supernatant was removed, dried downina centrifugal concentrator,
`cordingto ICH guidelinesforbioanalytical method validation prior
`resuspended in mobile phase (5% MeOH~95% potassium phos-
`to sample analysis. A daily cage rinse, a cage wash, and a cage
`phate, monobasic, pH 5.0), filtered (0.2 4m nylon Spin X tubes),
`wipe at termination were also performed for groups 1 and 2. All
`and analyzed by reverse-phase HPLC using an Agilent model 1100
`postdose samplesofbile, urine, feces, plasma, and cage residues
`instrament with automatic injection and a diode-array spectropho-
`(cage rinse, cage wash, and cage wipe) were analyzed for
`tometric detector. The mobile phase consisted of buffer A (25 mM
`radioactivity by LSC. Mean plasma concentrations of1 andits 3’-
`potassium phosphate, pH 5) and buffer B (methanol), A linear
`O-valinyl ester were uscdto calculate descriptive pharmacokinctic
`gradient from 5% to 100% buffer B was run over 15 min. The
`parameters by noncompartmental analysis using WinNonlin 4.1
`HPLC system used a Cjg Columbus column (5 zm, 250 x 4,6 mm,
`(Pharsight).
`Phenomenex, Torrance, CA). Authentic standards were used to
`The 15-day toxicokinetic study of nucleoside analogue 1 was
`identify the peaks in HPLC. This assay was done onetime (blood
`conducted to evaluate the pharmacokinetics of 1
`following iv
`and plasma from one donor) in duplicate.
`administration to Sprague— Dawley rats. Compound 1 was admin-
`General Procedure for Metabolic Stability Studies. Theability
`of derivative 3 to function as substrate for relevant metabolic
`istered once daily to rats intravenously for 15 days at dose levels
`of 60, 120, and 300 mg/kg/day. Each dose group consisted of eight
`enzymes was examined using human liver subcellular fractions
`animals/gender. On days 1 and 15, blood samples were collected
`obtained from BD Gentest (Wobum, MA). The prodrug was
`from thefirst set of four rats/sex/group at predose and approximately
`incubated for 1 h at 50 wg/mL with | mg/mL liver 89 orliver
`20 min and 2 h postdose and from the secondset of four rats/sex/
`cytosol. Prior to HPLC analysis, samples were quenched in 5
`group at approximately 5 min and 1 and 4 h postdose, Plasma
`volumes of ice-cold solvent (50% acetonitrile/50% methanol) and
`samples were analyzed for 2’-C-methylcytidine and 2’-C-methy-
`centrifuged (16 000 ref for 3 min). Supernatant was removed,dried
`luridine by a validated bioanalytical method (LC/MS/MS), and
`down in a centrifugal concentrator, resuspended in mobile phase
`composite PK parameters were calculated by noncompartmental
`(5% McOH—95% potassium phosphate, monobasic, pH 5.0),
`analysis using WinNonlin 4.1 (Pharsight).
`filtered (0.2 4m nylon Spin X tubes), and analyzed via HPLC (same
`conditions as described in General Procedure for in vitro Blood
`For LC/MS/MSanalysis, compound 1,its 3’-O-valinyl ester, and
`added internal standards were extracted from plasma (50 #L) using
`and Plasma Stability Studies).
`protein precipitation with acetonitrile (0.5 mL). The supernatant
`General Procedure for Protein Binding Studies. The protem
`wasdried, reconstituted, and analyzed by LC/MS/MS using HPLC
`binding of prodrug 3 and parent nucleoside 1 in rat plasma was
`coupled to a PE Sciex API3000 tandem mass spectrometer with a
`investigated using the ultrafiltration centrifugation method at a
`Turbo IonSpray interface. The HPLC system used an Alltech
`concentration of 20 uM for each. The studies used both nonradio-
`Platinum Cyg column (@ ym, 53 x 7 mm, Alltech Associates,
`labeled and radiolabeled test articles. The [*H]-3’-O-valinyl ester
`Deerfield, IL), HPLC elution was carried out using a gradient of
`of 2’-C-methylcytidine and [?H]-2’-C-methylcytidine, provided by
`ammonium acetate (10 mM,pH 4) andacetonitrile. The flow rate
`Moravek Biochemicals, Inc. (Brea, CA), were prepared by tritium
`was 2.5. mL/min with a split of 0.5 mL/min into the mass
`exchange (catalyst and labile tritium were removed during purifica-
`spectrometer. Each analyte was detected by multiple reaction
`tion). To achieve the desired concentrations, concentrated stock
`solutions for each test article containing the appropriate amounts
`monitoring (MRM)underthe positive ion mode using the precursor
`to product ion pair of m/z 357 to 112 (collision energy 27 eV) and
`of the radiolabeled and nonradiolabeledarticles were prepared in
`m/z 258 to 112 (collision energy 20 eV), for 3 and 1, respectively.
`HPLC grade water. Rat plasma (1 mL) was then added to each’
`Quantitation was achieved by constructing a calibration curve by
`tube and vortexed. The sample tubes were incubated at 37 °C for
`Lh.
`weighted linear regression of the ratio of the analyte peak area to
`that of the addedinternal standard. Calibration standards were in
`Centrifee micropartition units (1-mL capacity, Amicon Inc.) with
`blank rat plasma. The calibration range was 0.01-5 yg/mL in
`a molecular weight cutoff of 30 000 Da were used to separate the
`plasma for 3 and 0.05~25 yg/mL in plasma for the parent
`unbound from the bound test article by membrane filtration.
`nucleoside 1.
`,
`Fortified plasma (0.6 mL) was added to the sample reservoir portion
`N‘.[(imethylamino)methylene]-2-C-methyl-f-p-cytidine (4).
`of the ultrafiltration device. Plasma samples were centrifuge for
`A solution of 15628 (1.65 g, 6.43 mmol) in V,N-dimethylformamide
`10 min at 37 °C and. 1800g so that the amount of sample filtered
`(DMF, 32 mL) wastreated with dimethylformamide dimethylacetal
`ranged between 20 and 50% ofthe total volume. After centrifuga-
`(8.2 mL, 61.73 mmol) andstirred for 1.5 h at room temperature."
`tion, the ultrafiltrate of each sample was collected, and 10 wL
`The solution was evaporated under reduced pressure and coevapo-
`aliquots were analyzed for radioactivity using liquid scintillation
`rated with ethanol. Crystallization from ethanol/ether yielded the
`counting (LSC) on a Beckman Coulter LS6500. Protein binding
`hitherto unknown title compound4(first crop = 1.21 g, 60%,
`determinations were done in duplicate for cach test article.
`second crop slightly impure on TLC = 0.46 g, 23%) as crystals.
`General Procedure for Pharmacokinetic Studies. The phar-
`All
`the following physicochemical characteristics have been
`macokinetic and massbalance study of the [!C]-3’-O-valinylester
`determined on the crystals issued from the first crop crystalliza-
`of 2’-C-methylcytidine ('4C label at C-2, Moravek Biochemicals,
`tion: mp 169-172 °C; 'H NMR6 8.62 (s, 1H, N=CH), 8.17 (d,
`Inc.) was conducted in three groups of male Sprague—Dawleyrats.
`1H, H-6, Js—s = 7.3 Hz), 5.91 (m, 2H, A-1’, H-5), 5.16 (m, 1H,
`A single oral dose of the [!4C]-3’-O-valinyl ester of 1 was
`OH-5’, D2O exchangeable), 5.06 (s, 1H, OH-2’, D,O exchangeable),
`administered as a solution in 0.01 N HCIat a target dose of 100
`3.8—3.5 (m, 4H, H-3’, H-4’, H-5’, and H-5”), 3.15 and 3.02 (2s,
`mg/kg (as free base) to the three groups ofmale rats after fasting
`6H, N(CHs)2), 0.92 (s, 3H, CH3); FAB>0 m/z 625 (2M + Hyt,
`overnight. Group | was used to characterize the excretion of the
`313 (M + H)*, 167 (B + 2H)'; FAB<0 m/z 419 (M+ T — H),
`radioactivity derived from the [!4C]-3’-O-valinyl ester of 2’-C-
`403 (M + G — H)-, 311 (M — H)-, 165 (B); HPLC tp = 5.96
`methylcytidine in urine and feces. Urine and feces were collected
`min; Amex = 316.1 nm; UV (H20): Amax = 313 nm (€ 30 200), Amin
`up to 168 h postdose from group 1 rats. Group 2 comprised bile
`= 242 nm (€ 3600);
`fa]*®p = +134.7 (c = 0.95, H20). Anal.
`duct cannulated (BDC) male rats and was used to characterize the
`(Cy3H29N405°0.1H20) Gs H, N.
`excretion of the radioactivity derived from the [‘4C]-3’-O-valinyl
`3’-O-[N-(tert-Butoxycarbonyl)-L-valinyl]-2’-C-methyl-