`
`ELSEVIER
`
`advanced
`drug delivery
`reviews
`Advanced Drug Delivery Reviews. 19 (1996) 287—310
`
`
`
`Review
`
`Prodrugs of phosphates, phosphonates, and phosphinates
`
`Jeffrey P. Krise, Valentino J. Stella*
`Department of Pharmaceutical Chemistry and the Center for Drug Delivery Research. The University of Kansas.
`2095 Constant Avenue. Lawrence. KS 6604?. USA
`
`Received 1 August 1995: accepted I October 1995
`
`
`
`Abstract
`
`The objective of this paper is to review the literature on the use of prodrugs to overcome the drug delivery
`obstacles associated with phosphate. phosphonate and phosphinate functional group-containing drugs. This is an
`important area of research because. although we have been successful at identifying numerous phosphate and
`phosphonate functional groupcontaining drugs as antiviral and anticancer agents, as well as for other uses. our
`ability to orally deliver these drugs and to target
`them to desired sites has led to limited success. Various
`acyloxyrnethyl— and aryl-ester prodrugs have shown promise. Alternative and imaginative approaches may be
`necessary before complete success is realized. It
`is our—hope that this review will stimulate further innovative
`prodrug research into overcoming the barriers to the delivery of these important drugs.
`
`Keywords: Nucleotide analog; Nucleotide prodrug; Bioavailability; Ester; Permeability
`
`
`Contents
`
`288
`2. Why prodrugs ofphosphates. phosphonales and phosphinates'?
`288
`3. How can prodrugs overcome these problems?
`
`290
`4. Specificcxamples“
`290
`4]. Nucleotide analog prCidi'Lig3"
`290
`4.1 I. Simple and substituted alkylaiidaryimester prodrugsof phosphates and phosphonates
`4.1.2. Acyloxyalkylcstcrs 295
`4.1.3. Phospholipid derivatives.
`29?
`
`4.1.4. (.‘yclic prodrugs..................
`298
`4.1.5. Nucleotide analog conjugate systems.
`299
`
`4.1.6. Carbohydratederivatives 301
`4.1.3". ‘SATE‘ andDTE’
`30]
`
`4.2. Miscellaneous prodrugapplications... 301
`
`4..2.l Fosiuopn'l: An ACE inhibitor...
`....
`301
`4.2.2. CBS 24562: A neutral cndopeptidase inhibitor
`302
`4.23 L690330: A inhibitor of Inosnol monophosphatase:
`303
`.
`......
`
`4.2.4. (Hydroxy—2—naphthalenylmethyl)phosphonic acid. an lRiK InthItor
`..............
`304
`
`304
`5. Conclusions.
`
`References 305
`
`*Corresponding author. Fax: + l 913 8425612.
`
`© 1996 Elsevier Science B.V. All rights reserved
`0169-409X!96i$32.00
`SSDI 0169-409X{95)00111—S
`
`1
`
`G|L2019
`I-MAK, INC. V GILEAD PHARMASSET LLC
`|PR201 8-00123
`
`1
`
`GIL2019
`I-MAK, INC. V GILEAD PHARMASSET LLC
`IPR2018-00123
`
`

`

`288
`
`JP. Kris-e. VJ. Stella l Advanced Drug Delivery Reviews l9 {I996} 287—310
`
`0
`RO-—i=|’—-OR'
`([)R'
`Phosphate
`
`0
`R——i=|‘—OR'
`(in
`t’hosphmmic
`
`fi
`R—tfuR
`0‘"
`Phosphinutc
`
`Drug:
`
`R = organic residue: R' = H or [-3
`
`Prodrug;
`
`R 2 organic residue: R' = promoeity
`
`I. General structure of phosphate. phosphonate. and
`Fig.
`phosphinate drugs. R represents an organic residue while R'
`represents H or anionic charge for the parent drug or the
`promoiety for the prodrug.
`
`1. Introduction
`
`to review the
`The intent of this paper is
`literature on the use of prodrugs to overcome
`drug delivery obstacles associated with phos-
`phate, phosphonate and phosphinate functional
`group containing drugs. The general structures of
`phosphate, phosphonate, and phosphinate-con—
`taining drugs are shown in Fig.
`1 where R
`represents an organic residue, and R’ represents
`either an anion charge or a hydrogen atom in the
`case of the parent compound. or R’ is a neutral
`ester in the case of the prodrug.
`Initial
`research on the problems associated
`with the delivery of phosphate functional group-
`containing drugs began in the early 19605 with
`phosphate—containing nucleoside analogs USed in
`cancer and viral chemotherapy. Effort continued
`to expand and has recently increased in interest
`with the advent of the variety of drug classes
`utilizing the ph0sphate, phosphonate or phosphi-
`nate functionalities, especially antiviral agents
`and various molecules designed to alter cell
`signaling processes.
`
`2. Why prodrugs of phosphates, phosphonates
`and phosphinates?
`
`Although the range of drugs containing either
`a phosphate, phosphonate or phosphinatc—con—
`taining functional has changed,
`the underlying
`problem with the delivery of such drugs has not.
`The shortcomings in the delivery of these drugs
`can be broken down into two basic problems:
`
`and phosphinate
`1. Phosphate, phosphonate
`groups impart an anionic charge (mono- or
`di—) at nearly all physiological pH values
`making them very polar. This high polarity
`can be the basis for many deficiencies in the
`efficacy of drug delivery. Specifically. highly
`ionized species do not readily undergo passive
`diffusion across cellular membranes.
`
`2. BecauSe of the increased polarity, these agents
`often exhibit a low volume of distribution and
`
`therefore tend to be subject to efficient renal
`clearance as well as possibly biliary excretion.
`In addition to renal clearance. phosphates.
`particularly those of primary alcohols and
`phenols, are known to be substrates for many
`phosphorylases present
`in the body which
`readily clip the phosphate group from the
`drug. The rapid dephosphorylation results in a
`short duration of action. Phosphonates and
`phosphinates have the advantage of being
`more chemically stable and showing essential—
`ly no enzymatic lability.
`
`3. How can prodrugs overcome these
`problems?
`
`In an attempt to overcome these shortcomings.
`the ionizable phosphate, phosphonate and phos—
`phinate groups have been neutralized via chemi—
`cal derivatization. This generally involves de—
`rivatization of the phosphorus—coupled oxygen(s)
`to form neutral ester(s). If the intention of the
`ester is for it to breakdown in the body to release
`the parent drug, then such a derivative would be
`a prodrug. The neutralization of the charge(s)
`has been proposed to serve a number of pur—
`poses:
`
`l. The first is to decrease the polarity by increas—
`ing the lipophilicity of the drug molecule thus
`allowing access to cells and tissues that might
`not be available to the non-modified specie
`and possibly altering the distributionlelimina-
`tion pattern of the parent drug.
`2. Another use of the neutral ester prodrugs is
`particularly important for phosphates which
`are substrates for nonspecific serum phos-
`phohydrolases, such as alkaline and acid phos—
`
`2
`
`

`

`JJ’. Krise. VJ. Stella t’ Advanced Drug Delivery Reviews 19 (£996) 287——3'IU
`
`289
`
`it has been pro-
`phatases. With these drugs,
`posed that the neutral ester serves to disguise
`the phosphate from the enzymes thereby
`altering the apparent elimination and half-life
`as illustrated in Scheme 1. A flaw in this
`
`thinking is that once deprotected. the intrinsic
`properties of the drug should not be dramati—
`cally altered. Nevertheless, entrapment of the
`parent drug in tissues in which it
`is not
`normally accessible could lead to apparent
`changes in its pharmacokineticipharmacody
`namic properties.
`
`Although alterations in apparent clearance
`rates may be important, the principal goals of
`most prodrug modification efforts on phosphate,
`phosphonate and phosphinate drugs is alteration
`of membrane permeability to improve oral (GI
`permeability), brain, tumor and cellular delivery
`(mainly to virally infected cells) of these agents.
`When these prodrugs are used for improving
`oral bioavailability, various issues dealing with
`G1 absorption of drugs must be considered. The
`ability to address these issues will ultimately
`determine the proper selection of the prodrug
`system and its
`likely success. The optimal
`scenario for enhanced systemic delivery of pro-
`drugs after oral dosing is as follows:
`
`1. The prodrug must display adequate chemical
`stability for formulation purposes as well as
`stability in the variable pH environment of the
`GI tract.
`
`2. The prodrug should have adequate solubility
`in the GI
`tract environment
`to allow for
`dissolution.
`
` 0 Phosphatase - II—-
`
`
`(DO We
`09
`CFUS
`
`CDC“
`mm
`
`+ He—a—m
`.
`0”
`
`Proclrug
`Neutralization
`
`0
`o
`I
`II
`Ph
`so
`—P-—
`_
`C30 . on 92?;- OOH m sea
`0-H
`0-1
`m
`NETAKJLITE
`
`3. Once dissolved, the prodrug should also dis-
`play enzymatic stability to lumenal contents as
`well as the enzymes found in the brush border
`membrane.
`
`4. The prodrug should have properties that allow
`for good permeability (generally associated
`with an adequate log P value).
`5. After permeation of the lumenal membrane,
`the prodrug could revert to the parent drug
`either in the enterocyte or once absorbed into
`systemic circulation. Post—enterocyte reversion
`is desired because conversion in the en-
`
`terocyte would also allow for back diffusion
`into the GI lumen, a problem which is not
`generally recognized.
`
`When the prodrug is formulated to increase
`cellular permeability into viral-infected cells,
`tumor cells or across barriers like the blood brain
`
`barrier, the desired characteristics might change.
`Replacing the desire for complete and rapid post
`absorption reversion.
`is a need for balance in
`lability. The most optimal scenario, however
`unrealistic, would be for the prodrug to have
`complete enzymatic and chemical stability during
`the absorption process and in blood but readily
`revert
`to the parent compound once it has
`permeated the targeted cell,
`thereby ‘trapping’
`the drug in the cell (Scheme 2).
`Considering both of these scenarios, prodrugs
`for
`improved oral delivery and prodrugs for
`improved cell
`targeted delivery.
`the rate of
`bioreversion is a very important process that
`
`SYSTEM”mm
`
`O
`(Do—see
`
`PM
`
`Newngizalmn
`
`l
`
`|
`mediating
`
`r"
`
`\
`
`\
`
`\
`
`TAHGET CEIJ.
`
`0
`'
`
`V -2{”0"}
`
`9
`CEO-e0“
`I
`0'“
`m RI
`
`ii
`Q—ewm
`on
`mo
`
`Scheme 1. Scheme showing the possible advantages of pro-
`drugs in altering the phosphatase cleavage of phosphate—
`containing drugs.
`
`Scheme 2. Proposed scheme illustrates the potential advan—
`tagescf prodrugs over the parent molecule for intracellular
`targeting.
`
`3
`
`

`

`290
`
`.IJ’. Krffl’. VJ. Stella l Advanced Drug Delivery Reviews .1”) (”'96) 287—3.”)
`
`must be considered in detail when designing
`prodrug systems. For example, if bioreversion is
`very fast and non-specific, prodrug reversion may
`take place before the limiting barrier is over—
`come. On the other hand. if reversion is slow and
`
`inefficient at all sites. the prodrug may readily
`reach the site of action but never release enough
`parent drug to elicit a pharmacological response.
`With these factors in mind. choosing a suitable
`hioreversible protective group for phosphates.
`phosphonates and phosphinates presents a major
`challenge.
`
`4. Specific examples
`
`4.1. Nucleotide analog prodrugs
`
`Purine and pyrimidine nucleoside analogs have
`found great utility in treatment of neoplastic and
`viral diseases [1—8]. Please note, the word nu»
`cleoside will be equated with nucleoside analog
`for the remainder of this manuscript for simplici-
`ty purposes. These nucleosidic drugs mostly rely
`on viral or kinase—mediated (i.e.. thymidine ki-
`nase) activation step(s)
`to produce the phos-
`phorylated nucleotide. necessary to display bio-
`logical activity i.e.,
`the nucleosides are them~
`selves prodrugs. Unfortunately. dependency on
`kinase mediated synthesis can lead to the de-
`velopment of resistance [9—11]. The first step in
`the phosphorylation of a nucleoside to the mono—
`phosphate is known to be highly specific and
`often causes the development of resistance [12;
`
`15]. Therefore. it can be argued that an approach
`to circumventing the resistance development
`problem is
`to administer the monophosphate—
`containing nucleoside drug. This strategy has two
`flaws:
`
`l. The highly polar monophosphate has limited
`passive absorption properties and therefore,
`transcellular transport is very restricted [16—
`19]. This flaw is supported by the work of
`Leibman ct al. [18] who demonstrated the lack
`of passive as well as the absence of an active
`transport mechanism for ara-CMP (the mono
`phosphate derivative of ara C; Fig. 2).
`Rapid in vivo dephosphorylation of the mono—
`phosphate is observed with this class of drugs
`[20-122].
`
`go
`
`Could prodrugs be used to overcome these
`problems? The aim of this portion of the review
`is to focus on phosphorus—coupled oxygen pro~
`drug esters of nucleotides, namely the mono—
`phosphates and their phosphonate analogs. An
`effort has been made to group the various
`prodrug strategies according to the type of neu—
`tral ester chosen. The interested reader may wish
`to refer to other reviews on this subject that take
`a little different approach from those taken here
`[23-25].
`
`4.1.]. Simple and substituted alkyi and aryi ester
`pmdrags of phosphates and phosphonates
`Rosowsky et al.
`[26] have examined several
`mono-5’—(alkyl phosphate) esters of ara-C (Fig.
`
`N”;
`
`Nll_.
`
`IlU
`
`N/
`o‘J‘\
`
`N
`
`o
`
`no
`
`on
`
`ara-C
`
`|
`
`N/
`o‘j'\
`
`N
`
`|
`
`o
`
`in.
`
`Rape—o
`OR,
`
`0
`
`u
`
`on
`
`ara—(IMP {RJ = H. R2 = H)
`
`ill R1. R2 =C2H6» H
`{2} R]. R2 = n'Crle‘ H
`[3] RI- R2 2 n-C‘SHU. H
`[4] R.. R2 = n-Can. H
`(5] R.. R2 = n-CmHn‘ H
`[6} RI- R1 = Cszr Csz
`or) R]. R2 =n-C.H-;. D‘CiHst
`(R) R.. R2 : n-Cslln. I'I-Can
`(91 R.. R2 = n—C16H33.n'C|fiHJJ
`
`Fig. 2. Structures of ara-C. ara—(3MP. and araCMP alkyl prodrugs. Structures (1)—(4) are mono-alkyl prodrugs [26] while (5)—(9)
`are dialkyl prodrugs [30].
`
`4
`
`

`

`LP. Krfse, VJ. Stella 1' Advanced Drug Delivery Reviews l9 {I996} 287—310
`
`291
`
`2, 1—5) in an effort to deliver ara—CMP to cancer
`cells. The cytotoxicity of the proposed prodrugs
`toward cultured L1210 leukemia
`and B16
`
`melanoma cells appeared to obey an inverse
`structure—activity relationship with respect
`to
`alkyl chain length. The n-butyl and n—hexyl
`prodrugs were approximately half as active as the
`ethyl prodrug. The structure—activity relationship
`seemed to plateau as chain length became longer
`as exemplified by n-octyl and n-Clfi H33 esters
`having nearly the same lDSO values. Similarly,
`Mullah and coworkers [27] produced a 5'-O
`methyl and 5’—0 phenyl esters of 2’,3’-didehydro-
`2',3'—dideoxyadenosine and 2',3’-didehydro-2',3’~
`dideoxycytosine which displayed similar in vitro
`results to the parent nucleosidcs. When incuw
`bated in serum containing medium, the phenyl
`prodrug produced parent nucleoside along with
`nucleoside monophoSphate; the methyl prodrug
`was not evaluated in this manner. In an in vivo
`
`experiment, assessing activity against adenocar-
`cinoma 755 in mice, Montgomery et al. [28] has
`evaluated the mono- and diethyl. butyl, and
`phenyl
`esters
`of 6-mercaptopurine
`ribonu—
`cleotide. The monoester prodrugs were of the
`same order of effectiveness as (Hnercaptopurinc
`ribonucleotide, whilst the diesters were markedly
`less effective.
`
`In general, the mono-alkylfaryl ester analogs
`of phosphates failed to act as efficient prodrugs
`for the delivery of nucleoside—monophosphate
`analogs intracellularly. The poor activity of the
`monoalkyllaryl esters can be attributed to:
`
`1. High degree of polarity. Due to presence of
`the mono—anionic
`charge,
`limited passive
`transport across cells can be expected.
`2. The relative ease of in vivo conversion of the
`
`monoalkyl esters in vivo back to the parent
`nucleoside before the cell barrier has been
`
`overcome. This could be a possible explana-
`tion for the monoalkyl prodrugs and parent
`nucleoside sharing similar activities.
`
`McGuigan et al. [29] recently synthesized and
`evaluated, in vitro, 3 series of alkyl prodrugs of a
`hydrogen-phosphonate derivative of AZT (Fig.
`3) in an attempt to increase its antiviral activity.
`They showed that the short chain (Cl—C7) alkyl
`
`NH2
`
`n / "‘9
`
`l
`
`J\
`
`o
`
`N
`
`R= Me
`E :E-ll-lcpl
`R 2 "am“
`
`(III)
`RO-Ir—O
`H
`
`0
`
`"3
`
`AZ'I‘ Hephosphonate analog (R = H)
`
`Fig. 3. Structures of an AZT H-phosphonate analog and its
`prodrugs [29].
`
`hydrogen phosphonates to be more active than
`the long-chain (C18) ester prodrugs in HIV-1—
`infected C8166 cells. The short chain phospho-
`nate prodrugs were 5—10 times more potent than
`the parent phosphonate; however, all prodrugs
`were Poorly active in an infected JM cell
`line,
`which is thought
`to lack the kinase mediated
`phosphorylation to produce the active nucleotide
`analog. This demonstrated lack of activity in the
`J M cell line suggested that these prodrugs were
`acting as depot forms of the nucleoside rather
`than the nucleotide.
`
`To cause a further reduction in polarity, inves-
`tigators have also synthesized diesters of phos-
`phatesiphosphonates and evaluated their effec-
`tiveness as prodrugs in in vitro and in in vivo
`tests. The results were not consistent with the
`
`observation with the monoesters. For example,
`Colin et al.
`[30]
`shOWed a clear
`relationship
`between inhibition of thymidine incorporation by
`mammalian epithelial cells and lipophilicity (in-
`creased alkyl chain length) of diester prodrugs of
`ara-CMP utilizing a similar series of protecting
`groups to those studied by Rosowsky et al. (Fig.
`2, 6—9). The in vitro activity was lowest for the
`ethyl ester and highest
`for
`the hexy] ester.
`Similar work performed on ara-AMP gave a
`correlation between an in vitro inhibition of
`
`DNA synthesis and lipophilicity; however, no
`anti—viral activity at concentrations up to 100
`,ugfml was detected against a range of viruses
`[31,32]. AZTMP alkyl esters were also employed
`to improve membrane permeability [33,34].
`In
`vitro.
`the tricsters showed a complete lack of
`
`5
`
`

`

`292
`
`LP. Kris-e. VJ. Stella a“ Advanced Drug Delivery Reviews 1'9 (119%) 287—3“)
`
`inhibition of HIV, which was attributed to the
`
`high stability of the simple alkyl esters with little
`or no conversion to the active 5'—phosphate.
`
`</N
`0
`0
`R 0 III?
`N
`1 _
`R25 \/ \/\0/
`
`NH;
`
`\ N
`N)
`
`9-[2-[Phosphonomedmxykthoxyladeninc (R.. R2 = H}
`
`Fig. 4. Structure of 9-[2-(phosPhonomethonyethoxyladeninc
`(R1. R2 = H). Refer to Table l for representations of R1 and
`R2 [35].
`
`Ta blc l
`
`Serafinowska and co-workers [35] have made a
`series of dialkyl (among others) prodrugs of 9-[2-
`(phosphonomethoxy)ethoxy}adenine
`(Fig.
`4,
`Table 'l)
`in an effort
`to improve the poor
`bioavailability, 7% in rats [36], < 1% in monkeys
`[37].
`After oral administration to mice, the inves-
`
`tigators monitored blood concentrations of the
`diester, monoester, and free acid (1—5, Table 1).
`In all cases,
`the diester was well absorbed:
`
`however, subsequent conversion to the monoes—
`ter and free acid was more dependent upon the
`nature of the alkyl ester. Short chain diesters,
`being chemically (and probably enzymatically)
`
`Concentration of 9—[2—{phosphonomethoxy)ethoxy]adenine 1 (see Fig. 4) and its mono— and dialkyl esters in the blood following
`oral administration of esters to mice (Reprinted with permission from [35] American Chemical Society.)
`Number
`R!
`R2
`Total AUC.‘ 15,180 min (nM)
`
`l
`Monoestcr
`Diester
`Biowailabilit}I
`
`of l (%}“
`
`l
`2
`3
`4
`5
`
`H
`Me
`Et
`i—Pr
`n—Bu
`
`H
`Me
`Et
`i—Pr
`n‘Bu
`
`—
`U
`4
`5
`[J
`
`,
`9
`35
`2‘)
`69
`
`—
`39
`9
`10
`2
`
`2
`0
`8
`It}
`t]
`
`3t]
`0
`U
`[S
`NleJJCO_.CH2
`Me‘CCOZCH,
`6
`o
`o
`u
`n
`Me
`Me‘CCOZCH:
`?
`{l
`t)
`40
`0
`Et
`MefiCOfiH2
`8
`I)
`0
`8
`[1
`i—Pr
`lVle_1(.“.CtC)_,Cl-l2
`9
`[l
`n
`I l
`[l
`n—Bu
`Me‘(jt’leCH2
`It]
`[J
`0
`I}
`I)
`l'vle3C(Cl-l_,CI)CO,C'H2
`Me,(I(CI-I2CI)CO:CH3
`1]
`24
`0
`0
`12
`MeJCHC02CH(Me)
`Mc2CHCOZCH(Me}
`12
`74
`t]
`0
`31'
`MeJCCOICH(Me)
`Mc_.CC02CH(Me)
`13
`4
`t]
`18
`2
`(“hHfiCHJD
`ChflfiCHPO
`I4
`2
`9
`3
`I
`4—BrCuH‘CHEO
`Lurchiiqcup
`:5
`8
`3|
`12
`4
`4—C1C6H4CH30
`4-CIC,,H,CH2(J
`If:
`1
`I)
`0
`(1.5
`4-Me2CHC03Cfil-I4CHEO
`4—Me2CHC0?CfiH¢CH20
`l?
`s
`o
`o
`4
`apacocfifldcm
`4—AcOCfiH4CH3
`la
`8
`I)
`0
`4
`Br(CH3)20
`Br(CH_,_)30
`19
`6
`IS
`[3
`3
`CIECHCHzC}
`C12CHCH,O
`20
`o
`[3
`22
`o
`EtO{CH,}:O
`Et0(CH?)3O
`21
`26
`{J
`l
`13
`C,,H,O
`ChHSO
`22
`50
`U
`0
`25
`CnHFO
`CfiHSO
`23
`6
`0
`o
`3
`4-BrC6HdU
`4—BrChH40
`24
`22
`0
`t]
`l
`Z-McCfiH‘O
`2—MeCth0
`25
`It}
`0
`16
`5
`2—Ac0CbH‘O
`2—AcOCbH40
`26
`8
`0
`5
`4
`4-AcOCfiH40
`4-Ac0C,_I-[,O
`2'?
`0
`[l
`I
`0
`4»ACOCH_,C6H‘O
`4-AcOCl-IECbH40
`28
`m
`0
`l4
`5
`4AMcOCfiHJO
`cmeocth
`29
`
`30 I} 4-MeJCC02C6HJO 4-Me‘CCOJChH40 0 U 0
`
`
`
`
`
`"The bioavailability of 1 after oral administration of prodrugs was calculated from the equation: % bioavailability = (AUC l)!{iv
`AUG 1) X 100. where iv AUG 1 = 50 pM I1.
`
`I
`
`6
`
`

`

`J'.P. Krise. VJ. Stella ! Advanced Drug Delivery Reviews 19(1'996) 287—310
`
`293
`
`stable, were predominantly detected unchanged
`in the serum after oral administration. As the
`
`alkyl diester size increased the prodrugs tended
`to break down more efficiently to the monocster;
`however, the monoester tended to build up in
`the blood and failed to be converted to the
`
`[38] also
`parent phosphonate. Starrett et al.
`explored the effectiveness of dialkyl prodrugs to
`improve
`the
`bioavailability
`of
`9—[2—(phos-
`phonomcthoxy)ethyl]adenine (PMEA) and ob—
`tained similar results.
`
`Investigators also described the utility of the
`aryl functionality in prodrug design, both with
`unsubstituted and substituted aryl phosphate
`esters. When these were designed to increase
`cellular permeability, the ability of the prodrug
`to deliver
`the phosphate-containing drug in-
`tracellularly was minimal [39—43] probably due
`to removal of one or both of the esters prior to
`reaching the target cell. However, when the aryl
`functionality was used to increase GI tract per-
`meability several authors reported dramatic in—
`creases in bioavailability [35.38,44.45] (also refer
`to section 4.3.2 of this review). Table 1 contains
`oral bioavailability data for a number of aryl
`prodrugs of 9-[2—(phosphonomethoxy)ethoxy]-
`adenine (22-30). The most promising aryl pro-
`drug was the phenyl prodrug 23 (23 is a hydro—
`chloride salt of 22) which gave 50% oral bio—
`availability of the parent phosphonate.
`In an effort
`to produce more chemically!
`metabolically labile prodrugs, Mitchell et al. [46]
`synthesized the dibenzyl ester of methoxycar-
`bonylphosphonate. The hydrolysis
`in pH 7.4
`buffers (36.4°C) was rapid and resulted in a
`complex product profile. The major complication
`with this potential prodrug was that P—C cleavage
`occurred, yielding the dibenzyl phosphitc. In an
`attempt to shift the hydrolysis more toward P-O,
`or 0—bcnzyl group bond cleavage possibly yield-
`ing parent monophosphate-containing drug after
`hydrolysis,
`para-substituents
`on
`the
`benzyl
`groups were employed [4?]. As expected, para—
`electron withdrawing substituents yielded esters
`that hydrolyzed to the monobenzyl ester but
`little or no parent drug was formed. Only hydro—
`Iytic data was provided with these studies. A
`potential draw back in the clinical usefulness of
`benzyl esters with porn electron donating sub-
`
`stituents is the potential formation of quinone
`methide intermediates which are known to be
`toxic.
`
`By what mechanism do these alkyl esters of
`monophosphates exert their potential biological
`effect? The prodrugs could either act as depot
`for the nucleoside in which the P—O-nucleoside
`
`link would be broken to provide the nucleoside
`(pathway 1, Scheme 3). or act as an intracellular
`source of
`the nucleoside monophosphate by
`cleavage of
`the P-O-alkyl bond (pathway 2.
`Scheme 3).
`The first possibility (pathway 1), although
`potentially useful [48]. would fail to overcome
`the resistance seen with many of these drugs i.e.,
`it still delivers the nucleoside rather than its
`
`monophosphate.
`[31]
`To assess this issue, McGuigan ct al.
`prepared compounds analogous to conventional
`dialkyl phosphate prodrugs through the use of
`methylene bridges from the phosphorus to the
`alkyl chain (phosphinate linkage). The phosphi-
`nate (P-C)
`linkage is well known to be bio-
`logically stable [49]. The activity of di-n-butyl
`phosphinate ((C4H9)2PO—nucleoside) and dim-
`propyl
`phosphate
`((C3H,O)2PO—nucleoside)
`were compared (assumes similar Van der Waals
`radii for methyl and oxygen). Similar biological
`activities of di-n-butyl phosphinate and dim—pro-
`pyl phosphate could be expected if the mecha—
`nism of action was that both acted as depot
`prodrugs of the parent nucleoside. However, the
`phosphate was nearly twice as effective as the
`phosphinate. The fact that the phosphinate dis-
`played activity (albeil
`low) provides some evi-
`dence that
`the prodrugs may act,
`in part, as
`depots for the nuclebside. On the other hand.
`since the phosphates were consistently more
`active than their phosphinate counterparts, it was
`possible to infer that an increase in the phos-
`phate-to-alkyl link lability would produce more
`efficient prodrugs.
`In an effort
`to increase the [ability of the
`phosphate-promoiety bond (pathway 2, Scheme
`3), several halo alkyl prodrug esters have been
`made and evaluated [50,51]. The halo substituv
`tion was thought
`to favor
`the formation of
`nucleoside monophosphate and increase biologi-
`cal activity. Bis(2,2,2~trihaloethyl) phosphate de-
`
`7
`
`

`

`294
`
`LP. Krist’. VJ. Swim it Advanced Drug Delivery Reviews I9 ([996) 287—31’0
`
`./”“‘~.
`
`I
`
`1H:
`
`i
`\N/
`o\l
`"\.__________.9-
`
`no
`
`+
`
`it
`RO-P—OH
`I
`OR
`
`
`
`{gang}
`
`0
`\N/
`ll
`Ho—f—o—l/Oq
`a\.....
`5_J
`OH
`\_
`
`t
`
`[120le
`
`Scheme 3. illustration of the possible decomposition pathways for a generic nucleotide prodrug. Pathway | represents cleavage of
`the P-O—nucleosidc link and results in nuclcosidc release. Pathway 2 represents Cleavage of the P—O—alkyl linkage and results in
`accumulation of nucleotide inside the cell.
`
`2’,3’-dideoxycytidine
`of AZT and
`rivatives
`(Fig. 5) have been synthesized and
`(ddCD)
`evaluated [52].
`The in vitro anti—HIV activities of these com—
`
`pounds were evaluated and it was found that
`none of the prodrugs were as active as the parent
`nucleosides. The trichloroethyl moiety was found
`to be more efficacious than the trifluoroethyl
`moiety for ddCDMP: however, the two halogen
`prodrugs were
`nearly equally effective
`as
`AZTMP. When this promoeity was subsequently
`applied to ara-AMP and ara-CMP [53],
`in vitro
`testing showed the trichloro-containing prodrugs
`were consistently more active than the trifluoro—
`
`containing prodrugs. For the ara-CMP prodrug,
`the activity was found to be greater than that of
`the parent nucleoside. These observations were
`not consistent with the expected lability of the
`prodrug esters, as one would expect the trifluoro
`analogs to be the most labile. The observed trend
`in activity, however. was found to correlate with
`lip0philicity as exemplified by the larger OCIaI'IOU
`water partition coefficient for the tfichloro—con—
`taining prodrug. Subsequently, experiments were
`conducted to discover the potential utility of
`other related halo—alkyl prodtugs [54]. Mono—,
`di-. and trichloro substituted alkyl prodrugs of
`AZTMP (Fig. 5) were synthesized and com—
`
`0
`
`n
`RU-Fr'o
`0“
`
`NH2
`
`Me
`
`N/ I
`0J\
`
`N
`
`0
`
`H
`
`N3
`
`AZTMP (R = H}
`
`R= Hit?!l;
`R -—. CI3CCH;
`
`R : Elect-[CHE
`
`R = CICHZCH;
`
`OOJU
`
`N
`
`ll
`RO-FI’—O
`OR
`
`NH2
`
`N/
`
`R : cheer]2
`
`a rjccn1
`
`0
`
`ddCDMP(R = H]
`
`Fig. 5. Structures of AZ'I‘ mon0phosphate and ddCD monophosphate along with their corresponding haloalkyl prodrugs {52.54}.
`
`8
`
`

`

`LP. Krise. VJ. Stella 1" Advanced Drug Delivery Reviews [9 ([996) 287—310
`
`295
`
`pared. The expected lability of these prodrugs
`would be in the order
`trichloro>dichloro>
`
`monochloro. The observed activity. however, was
`in
`the
`order
`of
`triehloro>monoehloro>
`
`dichloro. When the haloalkyl prodrugs were used
`to increase bioavailability, only marginal
`in—
`creases were observed (see Table 1) [35,38].
`The results on alkyl, haloalkyl, benzyl and aryl
`prodrugs tend to imply that ability of
`these
`prodrugs to biorevert from the diester to mono--
`ester prodrug to the parent phosphate or phos-
`phonate does not solely rely on their chemical
`lability. For example, prodrugs with good leaving
`groups were not always the most efficient pro-
`drugs. What mechanism(s) are involved in the in
`vivo reversion of these prodrugs? The conversion
`of the diester prodrug to the monoester prodrug
`is not well understood and the in vivo rate of
`
`conversion of the dies‘rer prodrug to the monoes-
`ter prodrug is too rapid to be explained solely by
`chemical hydrolysis. On the other hand, no one,
`to our knowledge. has identified a true phosphot—
`riesterase in the mammalian system. One could
`speculate either non-specific enzymes or micro
`environmental pH changes are responsible for
`the conversion. Conversion of the monoester
`
`prodrug to the parent phosphatetphoSphonate
`drug can be explained by the presence of phos—
`phodiesterases (and phosphonodiesterases) [55—
`5?].
`In addition, phosphonate monoesters and
`phosphate diesters with systematically varied
`leaving groups were tested as substrates for 5’—
`nucleotide phosphodiesterase [58]. Although ali—
`phatic and benzyl esters were poor substrates,
`the aryl ester derivatives were shown to be good
`substrates for their diesterases. It was suggested
`that the geometry of the aryl group may be more
`important than the inductive nature of the aro—
`matic group.
`
`4.1.2. Acyloxyalkyt esters
`One approach to overcome the hydrolytic
`resistance observed with many of the prodrugs
`previously reviewed would be to utilize an en-
`zymatically labile promoiety; acyloxymethyl pro—
`drugs could serve as neutral lipophilic prodrugs.
`In theory, these prodrugs could transverse cell
`membranes by passive diffusion and revert, in-
`tracellularly.
`to the parent
`ionic phosphate or
`phosphonale after cleavage of the acyl group by
`esterases and rapid elimination of formaldehyde
`spacer group (Scheme 4).
`
`O
`ll /\ II
`2 a—c—o
`o P—O
`
`{/J MN“:
`
`\N/
`o\|
`1/
`\......-x
`
`Ester-rise
`——~—
`
`/""‘-. K
`
`0
`\N/
`li
`ll
`o
`R—vc—-—~0/\0-P—0
`. 1/ a
`<
`‘-._.....J
`
`OH
`
`{1120
`Fast
`
`/"\_:
`
`...___ +¢—~—
`
`li
`R—C—O
`
`\N/
`ii
`o
`O-P--0
`a. 1/ \.|
`L4
`
`{Ml/"Ra‘s
`
`ii
`\N/
`“04.4)
`0\J
`5.. 1/
`.
`no;
`
`Scheme 4. General mechanism for the bioreversion of acyloxymethyl phosphate prodrugs. The acyl group is cleaved through
`esterase activity to yield the hydroxymethyl analog. This analog quickly decomposes to formaldehyde and the monoester prodrug.
`The second group is subsequently cleaved by the same mechanism but possibly a different enzyme.
`
`9
`
`

`

`296
`
`JJ’. Krise. VJ. Stella it Advanced Drug Delivery Reviews [9 {F996} 287—3”)
`
`Investigators have utilized this strategy and
`evaluated the acyloxymethyl ester prodrugs for a
`variety of antiviral agents [59—61,l39] and CAMP
`[60]. Starrett et al.
`[62] have synthesized the
`bis(pivaloyloxymethyl) prodrug of the antiviral
`agent PMEA to increase antiviral activity. Star-
`rett’s studies showed the prodrug to have sub—
`stantially increased anti—herpes
`simplex virus
`activity in vitro compared with PMEA. Similarly.
`Sastry et
`al.
`[63] have produced the bis-
`(pivaloyloxymethyl) prodrug of 2’,3'-dideoxy-
`uridine 5'-mon0ph05phate (ddUMP). Addition-
`ally, metabolism studies in two human T cell
`lines lead to the formation of 5’-mono-. di~. and
`
`triphosphates of ddU after exposure to bis—
`(pivaloyloxymethyl) ddUMP.
`In contrast.
`these
`phosphorylated metabolites were not observed in
`cells treated with ddU or ddUMP alone,
`sug—
`gesting that this prodrug was able to intracellu-
`larIy release ddUMP. Freed et al.
`[64] utilized
`this same prodrug strategy for 5—fluorodcoxy—
`uridine monophosphate (5dUMP) and obtained
`results consistent with intracellular delivery of
`the monophosphate. Bis(pivaloyloxymethyl) pro—
`drugs of 9-(2—phosphonylmethoxycthyl)adenine
`( PMEA)‘
`9—( 2—phosphonylmethoxypropyl )ad—
`eninc (PMPA). and 9-(2—phosphonylmethoxy-
`propyl)diaminopurine
`{PMPDAP} were
`also
`synthesized and compared with the unmodified
`analogs. The bis(pivaloyloxymethyl) derivatives
`were biologically more active and showed en—
`hanced antiviral activities in vitro [65], most
`probably due to increased permeability of the
`prodrugs compared to the parent compound. In
`order to gain an understanding of the metabo~
`lism
`of
`this
`prodrug
`system,
`the
`[3H]bis(pivaloyloxymethyl) PMEA was
`incu-
`bated in an in vitro cell culture system. It was
`observed that
`the
`["H[bis(pivaloyloxymethy|}
`PMEA was rapidly hydrolyzed to the parent
`compound (PMEA) within the cells and was
`further metabolized to the mono- and sub-
`
`sequently to the diphosphate analog. It was also
`found that the diester prodrug releases a substan-
`tial quantity of the rnono(pivaloyloxymethyl)
`PMEA extracellularly. which could be a limita—
`tion
`of
`these
`prodrugs
`since
`the mono-
`(pivaloyloxymethyl)
`specie would not
`likely
`penetrate biological membranes.
`
`In addition to increasing antiviral activity,
`investigators have also evaluated acyloxyalkyl
`ester
`prodrugs of PMEA and
`9-[2-(phos-
`phonomethoxy)ethoxy] adenine in order to in—
`crease their poor oral bioavailabilities [35.38].
`Table 1, compounds 6—13, showed dramatically
`increased oral bioavailability for some acyloxy
`alkyl prodrugs. The bis(pivaloyloxymethyl) pro—
`drug (6) had reached an oral bioavailability of
`30%: however, methyl substitution at the alpha
`carbon of 6. led to an oral bioavailability of 74%
`for 13. Shaw et al. [66] tested alkyl. aryl, alkox-
`yalkyl. and alkyl amino acid phosphoramidate
`esters. Based upon an in vitro stability and
`transport
`screen.
`the
`bis(pivaloyloxymethyl)
`PMEA was selected as a potential oral prodrug
`for further in vivo animal studies. The bioavail-
`
`ability of bis(pivaloyloxymethyIJPMEA utilizing
`three different
`formulations was subsequently
`evaluated in monkeys [67]. The three formula-
`tions (hydroxypropyl—fi—cyclodextrin. PEG, and
`aqueous suspension) were used to explore any
`dissolution limitations with