C'urrem‘ Drug Metabolism. 2003. 4. 461-485
`
`461
`
`Design of Ester Prodrugs to Enhance Oral Absorption of Poorly Permeable
`Compounds: Challenges to the Discovery Scientist
`
`Kevin Beaumont”, Robert Webster‘, lain Gardner] and Kevin Dackz
`
`iDepartmem Qf'Pharmacokt'net:'cs. Dynamics and Metabolism, 3Department qfDi'scove.vjv Chemistry, Fffizer Global
`Research and Deveiopmem, Sandwich Laboratories, Ramsgare Road, Sandwich, Kent, UK CT} 3 9M}
`
`Abstract: Many drugs are administered at sites that are remote from their site of action. The most common route of drug
`delivery is the oral route. The optimal physicochemical properties to allow high transcellular absorption following oral
`administration are well established and include a limit on molecular size, hydrogen bonding potential and adequate
`lipophilicity.
`
`For many drug targets, synthetic strategies can be devised to balance the physicochemical properties required for high
`transcellular absorption and the SAR for the drug target. However, there are drug targets where the SAR requires
`properties at odds with good membrane permeability. These include a requirement for significant polarity and groups that
`exhibit high hydrogen bonding potential such as carboxylic acids and alcohols. In such cases, prodrug strategies have been
`employed.
`
`The rationale behind the prodrug strategy is to introduce lipophilicity and mask hydrogen bonding groups of an active
`compound by the addition of another moiety, most commonly an ester. An ideal ester prodrug should exhibit the
`following properties:
`
`1) Weak (or no) activity against any pharmacological target,
`2)
`Chemical stability across a pH range,
`
`3)
`4)
`5)
`6)
`
`High aqueous solubility,
`Good transcellular absorption,
`Resistance to hydrolysis during the absorption phase,
`Rapid and quantitative breakdown to yield high circulating concentrations of the active component post absorption.
`
`review the literature around marketed prodrugs and determine the most appropriate prodrug
`This paper will
`characteristics. In addition, it will examine potential Discovery approaches to optimising prodrug delivery and recommend
`a strategy for prosecuting an oral proclrug approach.
`
`INTRODUCTION
`
`For patient convenience, most drugs are administered by
`the oral
`route. However,
`there are
`significant hurdles
`confronting the delivery of a drug from the oral route, which
`often means that all of the administered compound does not
`reach its intended site of action. The extent to which the
`
`compound can overcome the hurdles to oral drug delivery
`and reach the systemic circulation is quantified by the term
`oral bioavailability. Optimising the oral bioavailability of a
`candidate molecule is a key objective of an oral drug
`discovery program. Clearly, compounds exhibiting low oral
`bioavailability are likely to require high doses to achieve the
`desired effects, since systemic exposure to the active com-
`pound will be limited. In addition, low oral bioavailability
`agents are prone to exhibit greater variation in exposure than
`higher bioavailability agents. This is particularly an issue in
`
`of
`the Depanment
`at
`author
`this
`to
`correspondence
`‘Address
`Pliarmacokiiietics. Dynamics and Metabolism, Pfizer Global Research and
`Development, Sandwich Laboratories, Ramsgate Road. Sandwich, Kent,
`UK. CTI3 9NJ, USA; Tel: (0[304) 616161; Fax: (01304) 65l8l7';
`E-mail: kevin_beaumont@sandwich.pfizer.con1
`
`the area of drug-drug interactions where concomitantly
`administered drugs can lead to unacceptable systemic expo-
`sure to a drug.
`
`The factors limiting the oral bioavailability of drug
`molecules are well established. They fall
`into 2 broad
`categories: absorption from the g.i.
`tract and first-pass
`extraction by the gut wall and liver.
`Incomplete oral
`absorption can be a major cause of poor oral bioavailability
`and much research has focused on the physicochemical
`properties required to ensure high oral absorption.
`
`The major physicochemical detenninants of extensive
`passive transcellular oral absorption have been extensively
`studied. The Lipinski ‘rule of 5’ mnemonic [1] suggests that
`in order to exhibit good oral absorption in humans, a drug
`should possess a molecular weight lower than 500, less than
`5 hydrogen bond donors,
`less than 10 hydrogen bond
`acceptors and a clog P less than 5. In addition,
`it is well
`known that in order to be transcellularly absorbed, a drug
`must possess a degree of lipophilicity (in general log Dm,
`greater
`than 0). These physicochemical constraints are
`related to the requirement of the drug to pass through the
`lipophilic environment ofthe gut wall cell membrane.
`
`l 389-26021-‘I33 $4 1 .80-i-.00
`
`© 2003 Bentham Science Publishers Ltd.
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2048 - 0001
`
`

`
`462 Current Drug Metabolism. 2003, Vol. 4, Na. 6
`
`Beaumont et at‘.
`
`However, there are pharmacological targets (eg Gpllb/Ila
`antagonists, ACE inhibitors, viral reverse transcriptase inhi-
`bitors etc.) for which some chemical series require physico-
`chemical properties that are at odds with good transcellular
`permeation. Thus, drug intervention at
`these targets can
`require compounds that are highly polar, have high molecular
`weight and/or high hydrogen bonding functionality. In order
`to make such compounds amenable to oral delivery,
`it
`is
`necessary to temporarily address these physicochemical cons-
`traints in a manner that can be readily reversed post absorp-
`tion. This is the rationale behind the prodrug approach.
`
`The major aim of a prodrug approach is to mask polar or
`ionisable groups within a molecule. This increases the
`overall lipophilicity of the molecule and promotes membrane
`permeability and oral absorption. However,
`increases in
`lipophilicity produced by a prodrug approach do not inevit-
`ably lead to major improvements in oral bioavailability. This
`
`is due to the multiplicity of the barriers facing oral delivery.
`These are outlined in Fig. (1).
`
`In order to be absorbed, the prodrug must be in solution
`in the contents of the gut. Thus, it must exhibit sufficient
`aqueous solubility to dissolve the entire dose. Once in
`solution, the prodrug needs to avoid extensive chemical and
`enzymic degradation, and must pass through the membrane
`of the gut wall cells (enterocyte). It is well known that efflux
`transporters are present
`in the enterocyte membrane [2],
`which are capable of intercepting drugs during membrane
`passage and placing them back into the gut lumen. Thus,
`successful prodrugs may need to avoid affinity for these
`transport proteins. Enterocytes are metabolically competent
`cells that express a wide range of drug metabolising enzymes
`including esterases [3], cytochrome P450 isoforrns [3-5] and
`UDP-glucuronyl transferases [3,6]. Prodrug ester hydrolysis
`in the enterocyte can be productive, if the active principle
`
`HEPATOCYTE
`
`SYSTEMIC
`CIRCULATION
`
`Biliary Excretion
`it (I)
`
`
`
`-COOR T ‘COOH
`
`(h)
`
`
`PORTAL BLOOD
`(93
` °‘ -COOR _-COOH
`I
`I
`
`
`
`
`
`
`
`M
`(bl
`(C)
`GUT
`(0)
`l-UMEN ‘C002 -COOR ::> j-coon
`
`Apical
`membrane
`
`(a) Chemical instability or degradation by gut lumen esterases,
`(b) Eftlux of ester from lhe enterocyle membrane by a transporter mechanism (such as P—glycoprotein),
`(c)
`Permeation of ester through the apical membrane of the enterocyte,
`(d) Breakdown of the ester to the acid by enterocyte esterases,
`(e) Return of acid to gut lumen (possibly by an efflux system).
`(I) Crossing of the basolateral membrane of the enterocyte by the acid,
`-(g) Crossing of the basolateral membrane of the enteroeyte by the ester,
`(h) Hydrolysis of the ester by esterases in the portal vein blood,
`(i) Hepatic extraction of the ester,
`(j) Hydrolysis ufthc ester by hepatocytes csterascs.
`(k) Non-esterase metabolism ofthe ester (such as CYP-mediated metabolism).
`(1)
`Biliary excretion of the acid,
`(I11) Return of the liberated acid to the systemic circulation (probably transporter mediated).
`
`Fig. (1). Barriers confronting the oral delivery of a protirug.
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2048 - 0002
`
`

`
`Design ofEsrer Prodrugs to Enhance Oral Absarplimr
`
`C'urren.r Drug Metabolism. 2003. Vol. 4. N0. 6
`
`463
`
`can pass through the basolateral membrane of the enterocyte
`and into the blood. However, if the active principle is a
`substrate for apical membrane transporters, it is most likely
`to be returned to the gut lumen. Any cytochrome P450 or
`conjugation metabolism on passage through the enterocyte
`will also be observed as incomplete oral bioavailability of
`the active principle.
`
`the ideal prodrug should rapidly and
`Post absorption,
`quantitatively be converted to the active principle. Our
`analysis suggests that
`this is rarely the case, due to the
`multiplicity of enzyme systems available to metabolise drugs
`on first-pass through the liver. Esterase enzymes present in
`the blood are capable of hydrolysing ester prodrugs. How-
`ever, these esterase enzymes have a strict structure activity
`relationship and do not metabolise all esters. Therefore some
`ester prodrugs do require hepatic ester hydrolysis and this
`poses
`several
`issues
`for quantitative release of active
`principle. Following hepatic ester hydrolysis,
`the active
`principle will be released in the hepatocyte. Since the aim of
`a prodrug strategy is to improve membrane permeation, there
`is potential for the active principle to be trapped within the
`hepatocyte and require active efflux from the cell. Should the
`active principle be a preferential substrate for eanalicular
`transporters, it will be extensively excreted into the bile. In
`order to be returned to the blood, it may require eftlux by
`sinusoidal
`transporters.
`in addition,
`the hepatocyte also
`expresses high levels of Phase I and I1 metabolising enzymes
`that may metabolise the prodrug or the active principle in a
`non-productive manner. Thus, any metabolism or excretion
`of either the prodrug or the active principle, which does not
`lead to the transfer of the active principle to the blood, will
`be non-productive and lead to a lowering of potential oral
`bioavailability.
`
`Overall, the barriers confronting the oral delivery of pro-
`drugs are considerable. In addition, to improving membrane
`permeability of a polar active principle, a prodrug should
`avoid transporter mediated efflux and be designed to yield
`the active principle into the systemic circulation. Incomplete
`membrane permeation, efflux, non-esterase metabolism or
`biliary elimination will lead to a reduction in the potential
`oral bioavailability of the active principle. Thus, in order to
`be successful, a prodrug approach must consider the balance
`of all these issues.
`
`This paper examines many literature examples illustra-
`ting the issues facing the prodrug approach. It will attempt to
`review the oral bioavailability of marketed prodrugs and
`outline examples of prodrug approaches that have failed.
`Finally, an attempt will be made to outline a strategy for the
`Discovery scientist
`to successfully prosecute a prodrug
`approach.
`
`REVIEW OF MARKETED ESTER PRODRUGS
`
`A number of examples of marketed ester prodrugs are
`listed in Tables I to 6. The prodrugs show a wide diversity in
`chemical structure, molecular weight and lipophilicity. How-
`ever, relatively few different classes of ester groups have
`been employed as the prodrug moiety. In general, the highest
`oral bioavailability values that can be achieved clinically
`using ester prodrugs are 40-60%,
`indicating that complete
`oral bioavailability in a prodrug programme is very unlikely.
`
`ALKYL ESTERS
`
`The most common prodrug moiety in marketed drugs is
`the esterification of an acid group with a simple alkyl
`alcohol. Major examples are illustrated in Table 1. One of
`the most successful classes of alkyl ester prodrugs is the
`diacidic angiotensin converting enzyme (ACE)
`inhibitors,
`where one of the acidic functions is masked as an ethyl ester.
`The oral bioavailability values for the diacidic active princi-
`ples (eg benazeprilat, quinaprilat etc) are not widely reported
`but are likely to be low. For instance, the oral absorption of
`enalaprilat is only 3% [7] and the bioavailability of cilaza-
`prilat is 19% [8]. In contrast, the oral bioavailability values
`of most of the monoacid prodrug agents (enalapril, bena-
`zepril, quinapril etc.) are significantly higher than 20%.
`However, even within this class the oral bioavailability
`values exhibit a significant range, from 17.5 % for moexipril
`to 57% for cilazapril.
`
`There are a number of reasons why the monoacid ester
`pro-drugs have higher bioavailability than their respective di-
`acid active principle and why there is a large variability in
`bioavailability between different ACE inhibitors. Firstly,
`changing one of the acid groups to an ethyl ester increases
`lipophilicity with a
`resultant
`increase
`in transcellular
`absorption. However, a number of the monoester ACE
`inhibitors exhibit greater oral absorption than would be
`expected purely from their lipophilicity values. For example,
`enalapril
`is relatively hydrophilic but
`is well absorbed in
`humans.
`It appears that as well as increasing lipophilicity,
`the monocarboxylic acid ester ACE inhibitors are substrates
`for intestinal uptake transporters [9-15]. In contrast although
`a number of di-acid ACE inhibitors can interact with the
`
`intestinal uptake transporters the second negative charge of
`these compounds makes their transport by these proteins
`unfavourable [l2,l6], and this in part explains their lower
`bioavailabilty.
`
`The ACE inhibitor ethyl esters tend to be stable to
`hydrolysis in human blood and require the action of hepatic
`(and possibly intestinal and/or kidney) esterases to liberate
`the active principle [17-20]. Thus, for enalapril the prodrug
`is 53-74% absorbed following oral dosing in man and the
`bioavailability of enalaprilat is 36-44%. This is consistent
`with the release of diacid from ester after
`intravenous
`administration. which is calculated to be 60%. The relatively
`slow hydrolysis of the ethyl ester prodrugs illustrates the
`potential benefit offered by blood labile prodrugs, which
`would give even higher concentrations of enalaprilat per mg
`dose of prodrug. Thus,
`the oral bioavailability values for
`individual ACE inhibitors will be dependant on several
`factors such as
`intrinsic lipophilicity of the monoester,
`affinity for the absorptive transporter and the rate of release
`of the active principles by cellular esterases.
`
`Ximelagattan is a more recent example of an ethyl ester
`prodrug. It contains two modifications to the direct acting
`thrombin compound melagatran. in addition to the ethyl ester
`group. the amidino group of melagatran is hydroxylated in
`the prodrug. Thus, two metabolic reactions are needed to
`liberate
`the
`active principle
`[21]. The modifications
`incorporated into ximelagatran increase the octanol: water
`partition by I70-fold compared to melagatran. This increased
`lipophilicity is reflected by an 80-fold greater flux through a
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2048 - 0003
`
`

`
`464 Current Drug Mcrabrniistn. 2003. Vol. 4'. No. 6
`
`Hmltrvrant er al.
`
`Table 1. Human Absorption and Bioavailability Data for Selected Alkyl and Aryl Ester Prndrugs
`
`Prodrug
`|references|
`
`Benazepri]
`
`[31]
`
`Cilazupril
`
`I5]
`
`Moexipril
`
`H37‘ 22]
`
`Pcrindopri I
`
`[92.l(i5]
`
`Absorption of
`prodrug {"/n)
`
`Oral hiuavailahiiity :2!‘ active principle
`when given as prodrug (°/n)
`
`zl7'.5
`
`estimate from urine moexiprilat
`Cxcnzlion Llfici‘ iv and oral mncxipril
`doses
`
`I‘-J—3 5
`
`higher values seen in elderly p:-uicnts
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2048 - 0004
`
`

`
`Design ofEster Prortrugs to Enimnce 0ru!Ab.s'arp1:'mr
`
`Current Drug Membufism, 2003. Vol. 4. No. 6
`
`465
`
`(Table 1) contd....
`
`Prodrug
`|references|
`
`Structure
`
`Absorption of
`prodrug 0%)
`
`Oral binavailability nf active principle
`when given as prodrug ("/n)
`
`Trandnlapril
`
`[93]
`
`Dclupril
`
`[91.m7.1u:<]
`
`Tcmocapril
`
`[I09]
`
`Famciclovir
`
`[110]
`
`Oseltamivir
`
`[Ill]
`
`Xllliclugatran
`
`[21,22]
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2048 - 0005
`
`

`
`466 Current Drug Metabolism. 2003, Vol’. 4‘. No. 6
`
`Beamrmnt et at‘.
`
`Structure
`
`Absorption of
`prodrug (%)
`
`Oral hioavailability of active principle
`when given as prodrug ("/n)
`
`48-69
`
`Dose-dependent
`
`(Table 1) contd....
`
`Prodrug
`|references|
`
`Candoxatril
`
`E23]
`
`Carbenecillin
`
`indanyl ester
`
`[24]
`
`Carbenicillin
`Phenyl ester
`
`E24]
`
`ND = no data
`
`for ximelagatran compared with
`Caco-2 cell monolayer
`melagatran [22]. Furthermore,
`the oral bioavailability of
`melagatran is 18-24% when dosed as ximelagatran compared
`to 3—7% when dosed as melagatran [22].
`In addition.
`ximelagatran
`is unaffected by
`food whilst
`the
`oral
`bioavailability of melagatran was decreased to 0.9% after
`food.
`
`ARYL ESTERS
`
`In some cases aryl esters have been preferred to simple
`alkyl esters (Table 1). Successful prodrugs of this type
`include candoxatril, and the indanyl and phenyl esters of
`carbenicillin. The
`oral bioavailability of candoxatrilat
`following oral administration of candoxatril
`is 32% [23].
`Similarly,
`two aryl esters of carbenicillin have been
`described. The bioavailability of carbenicillin indanyl ester
`and carbenicillin phenyl sodium in man is appreciable (>
`40%) [24]. Whilst the bioavailability of carbenicillin phenyl
`sodium is fairly linear with dose the bioavailability of
`carbenicillin indanyl ester shows clear dose-dependence that
`is probably due to saturation of an active absorption
`mechanism.
`Indeed studies in the rat have shown that
`
`carbenicillin indanyl ester is a substrate for an intestinal
`monoearboxylie acid transport system [25].
`
`DOUBLE ESTERS
`
`Double esters (Table 2) have been the preferred choice
`for producing oral prodrugs of the B~lactam antibiotics,
`nucleoside monophosphate anti-viral drugs. for the ACE
`
`inhibitor fosinopril, and for a number of angiotensin II
`receptor antagonists such as candesartan cilexetil.
`In most
`cases where double ester prodrugs have been used clinically.
`the esters are readily hydrolysed in vivo and circulating
`levels of the intact prodrug are usually not observed.
`
`Morrison et at’. [26] have conducted a detailed study of
`the contribution of various organs to the hydrolysis of
`fosinopril in the dog. In tissue hornogenate, the rank order of
`hydrolysis activity was liver 2 kidney > small
`intestine >
`lung. In blood, the hydrolysis of fosinopril was slow. In vivo
`both the small intestine and liver were shown to have high
`extraction ratios for fosinopril but due to the intestine having
`first exposure to the drug it was calculated that 75% of
`absorbed fosinopril was hydrolysed by the intestine and 25%
`by the liver.
`
`Sultamicillin is an interesting example ofa double ester
`prodrug where each molecule of pro-drug contains one
`molecule of ampicillin (an antibiotic) and one molecule of
`sulbactam (a [3-Iactamase inhibitor). During the absorption of
`the prodrug the double ester is cleaved and ampicillin and
`sulbactam are
`efficiently
`delivered
`to
`the
`systemic
`circulation[27].
`
`CYCLIC CARBONATE PRODRUGS
`
`A further class of clinically precedented prodrugs are the
`cyclic esters. exemplified by olmesartan and lenampicillin
`(Table 3). These prodrugs can be considered a subset of the
`double ester prodrugs They are designed to be labile in
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2048 - 0006
`
`

`
`Design of.E'srer Prod‘rng.¢ to Enhrmce Ora! Absarprimr
`
`Current Drug Membafism, 2003. Vol. 4. N0. 6
`
`467
`
`Table 2. Human Oral Bioavailability Data for a Number of Double Ester Prodrugs. Binavailability Data of the Active Principle,
`when Administered as Such, is Given for Comparison where Available
`
`Prodrug
`
`Structure
`
`Oral bioavailability of active
`principle when given as prodrug (“/o)
`
`Oral bio-availability of active
`principle ("/n)
`
`Adefbvir dipivoxil
`
`[60,] [2]
`
`Tenofovir disoproxil
`
`[136]
`
`Candcsaraan Cilcxitll
`
`[113.114]
`
`Fusinopril
`
`[xx,x9,1 15,1 lb]
`
`Bacampicillin
`
`[43,117.11x]
`
`Pivarnpicillin
`
`[43.l l7]
`
`Sultamicillin
`
`{211 l9.l20]
`
`25-39
`
`the higher values are in fed subjects
`
`> 80 arnpicillin
`> 63 sulbactam
`
`32-55 ampicillin
`sulbactam has low oral F"/o
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2048 - 0007
`
`

`
`468 Current Drug Metabolism. 2003. Vol. 4‘. No. 6
`
`Beaumont er at‘.
`
`Oral bioavailability of active
`principle when given as prndrug (“/o)
`
`Oral hioavailahility of active
`principle ("/n)
`
`31-58
`
`lower values are from fasted subjects
`
`32-43
`
`lower values are from fasted subjects
`
`(Table 2) mntd....
`
`Prodrug
`
`cefetamel pivoxil
`
`[l2l.l22]
`
`eefpodoxime pmxetil
`
`[122-124]
`
`Cefuroxirne axetil
`
`[122.l25.l2(>]
`
`Bacmccillinum
`
`[127]
`
`Pivmecillinam
`
`{[27]
`
`human plasma. avoiding the need for metabolism by cellular
`esterases for active principle release. The mechanism for the
`active principle release is discussed in more detail later in
`this review.
`
`active HMG-CoA inhibitors tend to be low due to high first-
`pass extraction by the liver. However, since these agents
`target the liver for their efficacy, achieving high systemic
`bioavailability after oral dosing is not necessary.
`
`LACTON ES
`
`AMINO ACID ESTERS
`
`The favoured strategy for increasing the oral absorption
`ofthe HMG-CoA reduetase inhibitors has been to cyclize the
`acid to produce a lactone (Table 4). Conversion of the
`lactone to the active acid occurs in the liver. The use ofthe
`
`ring closed lactone form results in significant oral absorption
`allowing the compounds to reach the site of action (the liver)
`at high concentration. The oral bioavailability values of
`
`The use of amino acid (valine) esters of the anti-viral
`drugs aciclovir and ganciclovir has been shown to improve
`oral bioavailability of the
`active principle (Table 5).
`Valganciclovir exhibits an oral bioavailability of ganciclovir
`of 61%, compared with a value of 6% when ganeiclovir is
`itself dosed orally [28]. Similarly the oral bioavailability
`of acielovir
`is 54% following dosing as valaeiclovir
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2048 - 0008
`
`

`
`D;-sign of.Esrer Prorlriegs to Enhrmce 0ru!Absarp1i:m
`
`Current Drug Membmlism. 2003. Vol’. 4. Na. 6
`
`469
`
`Table 3. Human Oral Bioavailability Data for Examples of Marketed Cyclic Ester Prodrugs
`
`Oral binavailahility nf active principle when
`given as prodrug (°/n)
`
`(Jlmcsarlan Mcdoximil
`
`[l2B.l2‘)j
`
`Lcnampicil Iin
`
`Table 4. Human Oral Absorption Data for Lactone Prudrugs
`
`144,130]
`[l3l.l32]
`[39]
`
`Structure
`
`Oral absorption of prodrug (°/n)
`
`Prodrug
`
`Luvaslalin
`
`[13l.l32j
`
`Simvaslatin
`
`Table 5. Human Oral Binavailability of Amino Acid Prudrugs. Bioavailahility of the Active Principle when Dosed Orally is
`Included for Comparison
`
`Oral bioavailability of active principle
`when given as prndrug (%)
`
`Oral bioavailability of
`active principle ("/n)
`
`Prodrug
`
`Valganciclovir
`
`[2S.l33]
`
`Valaciclovir
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2048 - 0009
`
`

`
`470 Current Drug Metabolism. 2003, Vol. 4‘. No. 6
`
`Table 6.
`
`Human Oral Bioavailability of other Marketed Prodrugs
`
`Beamrmnt er at‘.
`
`when given as prodrug ("/n)
`
`Oral absorption of
`prodrug {%)
`
`Oral hinavailability of active principle
`
`Mycophcnolatc
`Mofetil
`
`[72]
`
`compared to l2 — 20% when aciclovir itself is dosed [29].
`The increased oral absorption of these compounds is due to
`transport by the intestinal dipeptide transporter PEPTI
`[30,31]. This transport appears to be stereoselective with the
`l-valine ester of aciclovir being transported while the d-
`valine ester is not [29]. Valaciclovir is rapidly converted to
`aciclovir with the Cmax of valaciclovir being about 10% and
`the AUC less than 1% of that of aciclovir [29]. In addition,
`valaciclovir is undetectable in the plasma within 3 hours of
`an oral dose [32]. The enzyme responsible for hydrolysis of
`valaciclovir in the rat is located in both the intestine and liver
`
`[33].
`
`OTHER ESTERS
`
`Other examples of ester prodrugs used clinically include
`the morpholinoethyl ester of mycophenolic acid. This
`prodrug is completely de-esterified with no detectable levels
`of intact prodrug being observed. Data from intravenous
`infusion studies suggest that the half-life of the pro-drug is
`short (< 2 iriin) with no prodrug being detected within 10
`minutes of the end of infusion [34].
`
`ORAL ABSORPTION OF ESTER PRODRUGS
`
`is clear that
`it
`From our review of marketed prodrugs,
`one of the most successful prodrug approaches, the addition
`of simple esters to polar molecules to improve oral bioavail—
`ability, is evident in the ACE inhibitor area. The ACE enzyme
`is partially responsible for the control of blood pressure and
`ACE inhibitors have been used in the control of hyperten-
`sion. ACE is a zinc-dependant di-peptidase that metabolises
`angiotensin l to the potent vasoconstrictor angiotensin ll.
`Inhibition of ACE decreases the concentrations of angio-
`tensin ll at the same time as increasing angiotensin I and
`renin concentrations. This leads to blood pressure reductions.
`ACE recognises and subsequently hydrolyses the amide to
`liberate C-terminal dipeptide of angiotensin I, and inhibitors
`therefore tend to mimic this polar dipeptide fragment and
`contain a C-terminal carboxylic acid and a zinc binding
`group which is a 2nd polar fragment,
`typically a second
`carboxylic acid (e.g. enalaprilat), a thiol (e.g. captopril) or a
`phosphonate (e.g. fosinoprilat). Thus, potent ACE inhibition
`requires
`a compound that
`exhibits high polarity and
`hydrogen bonding capacity. This
`is
`incompatible with
`extensive transccllular absorption and consequently,
`the
`prodrug approach has been employed in order to deliver
`several ACE inhibitors by the oral
`route. These include
`enalaprilat and ramiprilat (Table 1), which are given as the
`ester prodrugs enalapril and ramipril. Ranadive er till. [35]
`have studied the physicochemical properties of these ACE
`
`inhibitors and their prodrugs. The dicarboxylic acid ACE
`inhibitors
`(enalaprilat and ramiprilat) are highly polar
`molecules (log D90, values of <-3 and -1.96, respectively)
`and as such are unlikely to be well absorbed by the trans
`cellular route. The ethyl esters, enalapril and ramipril, exhibit
`significant
`increases in lipophilicity (log DUO, values of
`-1.15 and +0.14, respectively) due to the masking of one of
`the ionisable carboxylic acid functions. For enalapril,
`this
`more than 100-fold increase in lipophilicity leads to a human
`oral absorption value of 53 to 74% (Table 1). This oral
`absorption value is unexpected since enalapril is still a polar
`molecule. However, there is evidence that this ethyl ester is
`actively absorbed by a carrier mechanism [12,16]. Rami-
`prilat exhibits higher intrinsic lipophilicity than enalaprilat
`due to the addition of an extra 3 methylene units. Thus, the
`oral absorption value for ramipril (S4 to 65%) may be driven
`by increased passive transccllular absorption relative to
`enalapril.
`
`The phosphorous containing ACE inhibitor, fosinoprilat,
`is administered as a double ester prodmg (fosinopril), which
`is attached via the phosphate functionality (Table 2). The log
`Dm,., values for fosinoprilat and fosinopril are —0.48 and
`+2.7 respectively. driven by the masking of the hydroxyl
`function and the significant addition of the lipophilicity
`associated with the ester functionality. This approximate
`1000-fold increase in overall lipophilicity leads to a human
`oral absorption value of 32 to 36% for fosinopril and oral
`bioavailability values of25 to 29%.
`
`is
`it
`the ACE inhibitor story suggests that
`Overall,
`possible to improve the oral absorption potential of polar
`compounds by the prodrug approach. In addition, it provides
`an
`insight
`into the potential problems
`that may be
`encountered by the prodrug approach since oral absorption
`of enalapril, ramipril and fosinopril is not complete. The log
`Dm” values of the ethyl ester derivatives are approximately
`0 which suggests equal partition between lipid and aqueous.
`This leads to significant oral absorption values with the more
`lipophilic esters exhibiting greater absorption. One could
`speculate that esters with greater lipophilicity would be
`likely to exhibit higher oral absorption values in humans.
`However, simply increasing lipophilicity at the expense of
`other
`favourable physicochemical properties
`is not an
`attractive option. This is
`illustrated by fosinopril, which
`exhibits greater lipophilicity than either of the ethyl esters,
`but a lower oral absorption value than ramipril. This may be
`due to the nature of the ester functionality increasing the
`molecular weight of fosinoprilat towards 500 or the lipo-
`philic nature of fosinopril may lead to a reduction in aqueous
`solubility and thus potential
`for solubility limited oral
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2048 - 0010
`
`

`
`Design ofEsrer Prodrugs to Enhance Oral Absarplimr
`
`C'urren.r Drug Metabolism. 2003. Vol. 4. N0. 6
`
`«WI
`
`absorption. Alternatively, the lower oral bioavailability exhi-
`bited by fosinopril may be due to non-productive clearance
`or to a lower contribution of active uptake in the absorption
`of fosinopril (see later section) Overall however, despite the
`incomplete oral absorption of these ACE inhibitor prodrugs,
`they are successful oral antihypertensive agents.
`
`A further example of the prodrug approach to increasing
`lipophilicity and hence transcellular absorption potential has
`been provided by Hovgaard et (Ii. [36]. These authors studied
`the effect of the addition of a cyclopropane ester on the
`lipophilicity and Caco-2 cell flux of a series of [3-blocker
`agents. The results are summarised in Table 7.
`
`Overall, esterification of these [3-blocking agents with
`cyclopropane carboxylic acid increased the log D94) by over
`I
`log unit. This lead to increases in Caco-2 cell permeabi-
`lities. Since Caco-2 cell flux is a good indicator of trans-
`cellular absorption potential
`[37], oral absorption of the
`esters is likely to be improved over the B-blocking agents,
`especially for the most polar examples.
`
`The GPlIbfIlla fibrinogen receptor antagonists provide
`further good examples of the use of the prodrug strategy. The
`fibrinogen-platelet interaction is modulated by a 3 amino
`acid sequence of arginine-glycine-aspartic acid. Antagonists
`of this receptor need to mimic this amino acid sequence and
`as
`such require both acidic and basic centres. Thus,
`GPIlb/llla antagonists are zwitterionic, highly polar and tend
`to exhibit poor oral absorption.
`
`A prodrug strategy with the GPIlb/Illa antagonist EMD
`80200 Fig. (2) has been reported [38]. The calculated log
`D(7_4, of EMD 80200 is —3.2, due to its basic guanidine and
`carboxylic acid moieties. This
`leads
`to a Caco-2 cell
`absorptive flux of 0.275 x 10"’ cm/sec, which is incompatible
`with significant
`transcellular
`absorption.
`in this case,
`esterification of the carboxylic acid did not
`lead to an
`increase in lipophilicity (calculated log Dn.4) ~3.55) or
`transcellular flux (Papp 0.277 x 10'“ cm/sec) due to the
`unmasked ionisation potential of the guanidine function. In
`EMD 122347, this guanidine function was masked with a
`methyl
`carbamate moiety,
`leading to
`an increase
`in
`calculated log D94, (-1.88 vs -3.2 for EMD 80200). At a
`concentration of 300].1M, the Caco-2 cell absorptive flux of
`EMD 122347 was 7.3 x 10'" cm/sec, some 25-fold greater
`than EMD 80200. However,
`flux in the non-absorptive
`(secretory) direction was 13.4 x 10'“ cm/sec.
`In addition,
`absorptive flux was shown to be saturable with increasing
`
`concentration of EMD 122347 and the efflux ratio could be
`
`abolished with verapamil. This profile of Caco-2 cell
`monolayer flux has been described previously [39, 40] and is
`ascribed to saturable efflux by drug transporters (such as P-
`glycoprotein). Interestingly, the asymmetry of flux was not
`observed with more lipophilic prodrugs of EMD 80200,
`where the ethyl ester was retained and greater lipophilicity
`was added at the guanidine end of the molecule. Overall, the
`increased Caco-2 cell flux of EMD 122347 over EMD 80200
`was shown to translate to increased oral absorption as judged
`by oral bioavailability determination in cynomologous
`monkeys (EMD 80200 3 to 10% and EMD 122347 40%).
`However,
`the more lipophilic esters, although showing
`greater Caco-2 cell
`flux, did not exhibit greater oral
`bioavailability values in the monkey. This may have been
`due to non-productive clearance of the prodrug.
`
`A further example of efflux transporters limiting the
`absorptive flux of prodrugs in the Caco-2 cell flux model of
`oral absorption has been described [4]]. The fibrinogen
`receptor antagonist, LY-767,679,
`is a polar (log DIM <-3)
`and
`zwitterionic
`compound which exhibits
`low oral
`bioavailability in animals. Caco-2 cell flux of LY-767,679
`was low (0.25 to 0.5%/h in both directions) indicating that
`this low oral bioavailability was due to poor absorption. The
`benzylic prodrug (LY—775,3l8, Fig. (2)) was significantly
`more lipophilic (log D94, 0.7) than LY-767,679. However,
`this did not lead to a large increase in absorptive Caco-2 flux
`(0.75%/h). In addition, flux in the secretory direction was
`3%/h and this asymmetry c