Journal of
`Medicinal
`Chemistry
`
`© Copyright 2004 by the American Chemical Society
`
`Volume 47, Number 10
`
`May 6, 2004
`
`Perspective
`
`Lessons Learned from Marketed and Investigational Prodrugs
`
`Peter Ettmayer.*~l Gordon L. Amidon} Bernd Clement,§ and Bernard Testaii
`Novartis Institute for BioMedicai Research. Brunnerstrasse 5.9, A-I235 Vienna. Austria. College of Pharmacy,
`The Univeisity ofiviichigan, Ann Arbor, Michigan 48109-1065, Piiarmaceuticai Institute, Univeisity ofKie1,
`19-241 18 Kiel. Germany. and Department ofPharmacy. University Hospital’ Centre (CHUV). CI-1-1011 Lausanne. Switzerland
`
`Received August 7. 2003
`
`1. Introduction
`
`to overcome
`Prodrugs are an established concept
`barriers to a drug’s usefulness. In Germany. about 6.9%
`of all marketed medicines can be classified as prodrugs,
`an estimate based on a conservative prodrug definition
`that does not include soft drugs and limited prodrugs.
`The limited prodrugs. which comprise an estimated 4%
`of the medicines marketed in Germany, are defined as
`active agents whose metabolite(s) also contribute(s) to
`the observed therapeutic activity. Approximately 49%
`of all marketed prodrugs are activated by hydrolysis,
`and 23% are bioprecursors (i.e., lacking a promoiety)
`activated by a biosynthetic reaction.‘ Noteworthy block-
`buster prodrugs are, for example, omeprazole, simvas-
`tatin, lovastatin, enalapril. and aciclovir (Figure 1).
`Regulatory guidelines pay limited attention to issues
`specific to prodrugs?’ Thus, active drugs developed from
`prodrugs are considered as active metabolites. In me-
`dicinal chemistry, a prodrug strategy is practically never
`considered in the early phases of drug design but only
`when classical analoguing programs fail to provide the
`required drug profile. What.
`then. makes prodrug
`development so special and interesting? Prodrugs pro-
`vide a rationale and opportunities to reach target
`physicochemical, pharmacokinetic. and pharmaco-
`dynamic properties. They can be designed to overcome
`pharmaceutical, pharmacokinetic, or pharmacodynamic
`
`* To whom correspondence should be addressed. Phone: +43-1-
`86634-378.
`Fax:
`+43-1-86634-383.
`E-mail:
`peter.ettmayer@
`pha.rma.novartis.com.
`‘ Novartis Institute for BioMedical Research.
`‘The University of Michigan.
`3‘ University of Kiel.
`" University Hospital Centre.
`
`
`
`iovastatin
`
`enaiapril
`
`Figure 1. Some blockbuster prodrugs.
`
`barriers such as insufficient chemical stability, poor
`solubility, unacceptable taste or odor. irritation or pain,
`insufficient oral absorption,
`inadequate blood—brain
`barrier permeability. marked presystemic metabolism.
`and toxicity.3 A developing field of high importance is
`that of rationally designed prodrugs for tissue or cell
`targeting. However. it is worth recalling that many
`successful prodrugs in current use are in fact accidental
`prodrugs, namely, agents that were not designed as
`prodrugs and were recognized as such only late in
`development or even postmarketing. This article intends
`to provide some arguments and guidelines for the early
`
`© 2004 American Chemical Society
`lO.l02lfjm03038l2 CCC: $27.50
`Published on Web 04/13/2004
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2050 - 0001
`
`

`
`2394 Journal‘ of Medicinal Chemistry, .2004, V01. 47. No. 10
`
`Perspective
`
`recourse to, and successful application of, a prodrug
`strategy in drug discovery.
`
`2. Definitions and Classification of Prodrugs
`
`The term prodrug was first introduced in 1958 by
`Adrien Albert‘ to describe compounds that undergo
`biotransformation prior to eliciting their pharmacologi-
`cal eifects. According to his definition (which we consider
`the best available), prodrugs are “therapeutic agents
`which are inactive per se but are transformed into one
`or more active metabolites." A later definition by Bund-
`gaard5 states that "By attachment of a pro-moiety to
`the active moiety, a prodrug is formed which is designed
`to overcome the barrier that hinders the optimal use of
`the active principle." Such a definition. however,
`is
`restricted to prodrugs carrying a promoiety. This defini-
`tion is somewhat narrow because it excludes biopre-
`cursors, modular prodrugs (i.e., comprising a trigger, a
`linker, and the active agent), and metabolites produced
`by conjugation reactions.
`Albert's broader definition posits prodrugs as the
`opposite of soft drugs. The soft drugs (see section 3.3)
`are active analogues of a lead compound that are
`deactivated in a predictable and controllable way after
`having achieved their therapeutic effects But since both
`prodrugs and soft drugs find applications in local tissue
`targeting, they are sometimes confused in the literature.
`Prodrugs can be classified according to two major
`criteria, namely, (a) chemical classes (carrier—linked pro-
`drugs, bioprecursors (i.e., prodrugs lacking a promoiety),
`site—specific chemical delivery systems, macromolecular
`prodrugs, and drug-antibody conjugates) and (b) mech-
`anism of activation (enzymatic versus nonenzymatic,
`activation by oxidation. reduction or hydrolysis. cata-
`bolic versus anabolic reactions).
`The enzymatic versus nonenzymatic activation is of
`particular interest because both routes have advantages
`and disadvantages. While prodrug activation through
`bioconversion as a time— and tissue—controlled process
`has a clear benefit, inter- and intraspecies variability.
`genetic polymorphism, and the potential for drug—drug
`interactions (see section 3.1.4) pose significant chal-
`lenges for such a prodrug Strategy. When purely chemi-
`cal prodrug activation is chosen (e.g., spontaneous
`chemical cleavage at physiologic pH), interspecies vari-
`ability, genetic polymorphisms, and drug—drug interac~
`tion problems are no concern. However, there might be
`chemical stability issues (insufficient shelf life) and the
`site of prodrug activation is undefined. Interestingly,
`there are only a few reports on prodrugs designed to
`rely exclusively on a nonenzymatic activation principle7
`maybe because it is very difficulty to ascertain the
`absence of enzymatic participation.
`The complexity of this broad multidisciplinary field
`makes it difficult to cover in depth all complementary
`viewpoints. Because some topics have already been
`covered by reviews and monographs,3‘1° this report will
`focus mainly on the inherent therapeutic benefit pro-
`drugs can offer in terms of overcoming pharmaceutical,
`pharmacokinetic, or pharmacodynamic barriers and
`increased patient convenience.
`
`3. Major Objectives of Prodrug Design
`The drugs mentioned in the sections below are meant
`to illustrate various opportunities medicinal chemists
`
`
`
`celecoxib prodrugs
`
`parecoxib sodium
`
`Figure 2. Examples of promoieties for improved aqueous
`solubility.
`
`may find in a prodrug strategy. However, this presenta-
`tion should not be understood as covering all prodrugs
`in therapeutic use or in clinical development.
`3.1. Improved Bioavailability. There are a variety
`of possibilities for a prodrug to improve bioavailability
`following oral dosing. We examine in turn aqueous
`solubility, passive intestinal absorption, targeted active
`absorption, and metabolic switching as well—established
`factors limiting bioavailability.
`3.1.1. Improved Aqueous Solubility. Inadequate
`aqueous solubility is an important factor limiting
`parenteral, percutaneous, and oral bioavailability. In
`such cases, a prodrug strategy may bring great phar-
`maceutical and pharmacokinetic benefit. Charged pro-
`moieties (e.g., esters such as phosphates, hemisucci—
`nates. aminoacyl conjugates. dimethylamino acetates)
`and neutral promoieties (e.g., poly(ethylene glyco1)s,
`PEG) can be used. The latter promoieties, however,
`require PEG of high molecular weight to avoid a rapid
`clearance characteristic of low MW PEG conjugates.
`Improved aqueous solubility for better iv administration
`has been demonstrated for PEG-paclitaxel" (Figure 2).
`but we note here that 2’-PEG esters, but not 7’-PEG
`esters, are hydrolyzed enzymatically. Representative
`water solubilities of PEG-paclitaxel prodrugs are as
`follows:
`2’—PEIG5°°°, 666 mg mL‘1: PEG3°°°°, 200 mg
`mI_.‘1; PEG‘°°°°, 125 mg mL“; compared to 25 ,ug mL"
`for paclitaxel itself.”
`Fosphenytoin is another example showing how paren-
`teral delivery problems associated with a sparingly
`water—soluble drug can be overcome by a prodrug.
`Fosphenytoin” (Figure 2) is a hydrophilic phosphate
`prodrug of the anticonvulsant phenytoin and is hydro-
`lyzed rapidly by phosphatases. Another example can be
`found in the parental administration of COX-2 inhibi-
`tors of the diarylheterocyclic class (e.g., celecoxib” and
`valdecoxib'5). Intravenous treatment of acute pain and
`inflammation is hampered by their modest aqueous
`solubility (<50 yg mL‘1). A solution to the problem was
`afforded by the sodium salt of the acetylated sulfon-
`amide prodrug (parecoxib sodium”) (Figure 2), which
`exhibits a significantly improved aqueous solubility
`(~15 mg mL"). The acyl residue is rapidly cleaved in
`hepatic and intestinal preparations and in vivo to
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2050 - 0002
`
`

`
`Perspectivs
`
`Journal ofMed1'cina1 Chemistry, 2004, Vol‘. 4 7, No. 10 2395
`
`NH2
`
`H
`
`9 OH no
`OH
`
`H0
`
`HO—fPqLII3—OH
`
`NH2
`pamidronate
`
`OH
`
`NH?
`
`OH
`
`1
`l
`Ii’-OH
`HO-fl‘
`0 OHO
`alendronate
`
`.‘ N
`
`S H
`
`_ iHg N
`
`0
`
`O
`
`bacampicillin
`
`
`
`OH
`
`O
`
`HO
`
`O
`
`HO HN
`.
`OA\HNYNH
`NH,
`Zal'lal'l'llVll'
`
`benazepril
`
`fosinopril 0
`
`Figure 3. Examples of ester prodrugs showing improved
`passive intestinal absorption.
`
`liberate the active drug. Parecoxib sodium is currently
`in the preregistration phase for parental administration.
`Improved oral bioavailability and dose linearity in
`pharmacokinetics have also been reported for acylated
`celecoxib prodrugs” whose sodium salt showed im-
`proved oral bioavailability with dose linearity over a
`wider dose range compared to celecoxib itself.
`One of the potential problems in this approach is that
`solubilizing groups may sometimes generate toxic ef-
`fects. Thus, the phosphate prodrug fosphenytoin was
`designed to enhance the aqueous solubility of phenytoin
`for intravenous administration, but the high phosphate
`concentrations at the site of injection produced some
`hypocalcemjc effects leading to mild pruritus and par-
`esthesia. 17 Interestingly, such side effects were moder-
`ate compared to those of the standard injectable sodium
`phenytoin.
`3.1.2. Improved Passive Intestinal Absorption.
`Providing enhanced lipophilicity for increased passive
`intestinal absorption is the most frequent rationale
`when adopting a prodrug strategy (approximately 49%
`of all marked prodrugs are activated by hydrolysisl).
`Various esters of carboxylic acids (a frequently encoun-
`tered group of carrier-linked prodrugs) are cleaved by
`hydrolysis (enzymatic and/or chemical) to liberate the
`active carboxylic acid. There are many examples such
`as ACE inhibitors” benazepril and fosinopril (Figure
`3). statins such as simvastatin and lovastatin (Figure
`1). and some antibiotics such as pivampicillin and
`bacampicillin (Figure 3). While this article was in
`preparation. Kevin Beaumont et al.3 comprehensively
`reviewed the design of ester prodrugs to enhance oral
`absorption. We refer the reader to this in«depth analysis
`of ester prodrugs and the related challenges for the
`discovery scientists.
`A nice example of the competitive advantage provided
`by a prodrug strategy is found in the class of neuram-
`
`adefovir dipivoxil
`
`Figure 4. Examples of drugs incorporating phosphonate
`moieties and prodrugs thereof.
`
`AZT
`6
`O:g_ O
`1
`-~/ENH
`Cl)
`0
`
`AZT
`
`carboxylesterase
`—-' ‘
`
`AZT
`I
`3:3
`Q'l,°‘0
`NH
`..
`HO/£0
`
`SN?
`—p-
`
`‘(R-0
`05?
`HN
`
`O
`
`H 0
`2
`'
`
`,0-AZT
`°~‘p
`NI ‘0
`H
`
`h
`h
`o~AzT
`P “P °‘
`0,?-
`dieslerase
`‘OH
`*" -
`on
`
`HO
`
`0
`
`Figure 5. Proposed conversion of AZT-phosphoramidate to
`AZT-monophosphate.
`
`inidase inhibitors used against type A and type B
`influenza viruses. The prototypal drug is zanamivir
`(Figure 3), which shows that a very high polarity is
`incompatible with intestinal absorption and has to be
`administered via inhaler. Another neuramidase inhibi-
`tor is RO-64-0802, which like zanamivir is not absorbed
`orally because of its high hydrophilicity. However, RO+
`64-0802 is not marketed as such but as the ethyl ester
`known as oseltamivir (Figure 3). This prodrug is well
`absorbed orally, and its rapid in vivo enzymatic hy-
`drolysis provides high and sustained plasma levels of
`the active parent drug.”
`A comparable situation is found in the field of his-
`phosphonates, where drugs with improved oral bio-
`availability are badly needed.” Indeed, pamidronate,
`alendronate (Figure 4), and analogues have a poor oral
`bioavailability (<1%).“ While the methyl and ethyl
`phosphate esters are chemically and enzymatically too
`stable,
`the pivaloyloxymethyl and .S'—acetylthioethyl—
`phosphonate esters are converted enzymatically to the
`free phosphonate.” Similarly. adefovir dipivoxil. a his-
`(pivaloyloxymethyl) ester of the antiviral adenine nucleo-
`tide analogue adefovir,” was recently launched in the
`U.S. for the treatment of hepatitis B virus infection
`(Figure 4). Other conceptually successful phosphate
`prodrugs are the aryl phosphoramidates, which deliver
`the nucleoside monophosphates to cells. This has al-
`lowed the efficacy of AZT to be increased by several
`orders of magnitude (Figure 5).“ Interestingly,
`this
`promising prodrug concept has not yet been applied to
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2050 - 0003
`
`

`
`2396 Journal‘ of Medicinal Chemistry, .2004, V01. 47. No. 10
`
`Perspective
`
`
`
`O
`
`0
`
`.
`.
`O ssbraflban
`
`ximelagalran
`
`Figure B. Amidoxime prodrugs on the market or in develop-
`ment.
`
`
`
`desmopressin pivaloyl ester
`
`Figure 7. Examples of prodrugs of peptidomimetics.
`
`phosphotyrosines. It should be mentioned in this context
`that one should be cautious with prodrugs of highly
`charged drugs, since their intracellular release might
`lead to entrapment and thus to toxicity.25
`Besides carrier—linked prodrugs,
`there exist other
`approaches to reach optimal lipophilicity. Amidoximes
`can be used as bioprecursors for amidines (Figure 6).
`Amidoximes are less basic and thus unprotonated under
`physiological conditions, thereby enhancing intestinal
`absorption.” Reductases in the kidneys, liver, brain,
`lungs, and gastrointestinal tract are responsible for the
`rapid conversion of the inactive amidoximes to amidines.
`A GP IIb/IIIa-receptor—antagonist (sibrafiban) and a
`thrombin inhibitor ximelagatran (the "double” prodrug
`of melagatran) are based on the amidoxime strategy and
`are in clinical development. Ximelagatran is expected
`to become the first orally acting, direct thrombin inhibi-
`tor in medical use.“
`
`Increasing evidence is being published on prodrugs
`of peptides, a class of pharmacologically important
`compounds that almost always suffer from mucosa]
`permeation problems and rapid enzymatic degrada-
`tion.3-33 The same is true for some peptidomimetics.
`Thus, the pivaloyl ester of desmopressin (Figure 7), a
`synthetic analogue of vasopressin, showed improved
`transport across Caco—2 monolayers.” The perbutan—
`oylated prodrug glycovir showed improved bioavailabil-
`ity in rat, dog, and monkey compared to the parent
`compound SC-48334. Transient cyclization of linear
`peptides using an esterase-sensitive linker yielded pep-
`tide prodrugs with improved Caco—2 permeability and
`increased stability toward peptidases. The 0-acetyl
`
`H
`
`NH2
`E
`
`H NH,
`”°
`
`0
`
`‘—\N
`(\
`N
`
`N\ NH2
`| Y
`NH
`
`0
`
`‘W
`N
`<\
`N
`
`I
`
`NYNH2
`NH
`
`valaciclovir
`
`O
`
`valganciclovir
`
`O
`
`
`
`H
`
`“
`
`i
`HN
`/
`O5)
`
`zidovudine
`
`H
`0
`H
`
`0
`H2N;\¢J
`
`I
`
`zaicitabine
`
`Figure 8. Example of (pro)drugS utilizing oligopeptide and
`nucleoside transporters.
`
`prodrug of the E-IIV—protease inhibitor saquinavir showed
`improved oral bioavailability (Figure 7).“
`Designing carrier—linked prodrugs, one should keep
`in mind to balance the increased lipophilicity necessary
`for transcellular absorption with sufficient aqueous
`solubility. otherwise oral bioavailability will become
`dissolution limited. Such cases have been reported for
`the bifunctional prodrugs of cephalosporines“ and for
`the pivaloyloxymethyl ester of ceftizoxime.” In addition
`the toxicity potential of the promoiety should be evalu-
`ated or promoieties used that are already accepted by
`registration authorities and are known to be nontoxic.
`Recently the commonly used pivalic acid (trimethyl-
`acetic acid) was associated with some toxicity.33 Also,
`the formaldehyde released upon hydrolysis of methyl-
`ene—bridged double ester prodrugs can be a toxicological
`concern. The toxicologically more acceptable acetalde-
`hyde, on the other hand, introduces a chiral center that
`increases complexity, since the two enantiomers might
`have distinct activation profiles.
`3.1.3. Improved Transporter-Mediated Intestinal
`Absorption. Utilizing carrier-mediated transport is
`another successful prodrug strategy to actively enhance
`intestinal absorption.“ For example, oligopeptide trans-
`porters are responsible for the active uptake of ,6-lactam
`antibiotics, the ACE inhibitor enalapril35 (Figure 1),
`valaciclovir, and valganciclovir, to name a few (Figure
`8). Valaciclovir, the valine ester of aciclovir, showed an
`oral bioavailability 5-fold higher than that of aciclovir
`itself.“ The increased oral absorption is due to transport
`by the intestinal dipeptide transporter PEPTl.37 Nu-
`cleoside transporters are involved in the uptake of
`nucleoside analogues such as zidovudine. zalcitabine
`(Figure 8), cladribine, ara—C, ara-A, and fludarabine.
`Bile acid transporters are used for the uptake of
`thyroxine, chlorambucil, and crilvastatin. Further classes
`of transporters include vitamin transporters (whose
`substrates include methotrexate, nicotinic acid,
`thi-
`amine, and vitamin B12), organic cation transporters
`(e.g., choline, spermine, (+)-tubocurarine, and dopam-
`ine), organic anion transporters (e.g., methotrexate,
`cefodizime, ceftriaxone. and pravastatin), and glucose
`transporters (e.g., glucopyranosides and galactopyran-
`osides). Successful targeting of active transport depends
`on the specificity and capacity of the transporters.
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2050 - 0004
`
`

`
`Perspective
`
`Journal ofMeo'1'ci'na1 Chemistry, 2004, Vol‘. 4 7, No. 10 2397
`
` HHNN
`
`Ievormeloxifene
`
`docarpamina
`
`Figure 9. Example of prodrugs of metabolically labile drugs.
`
`Furthermore. one has to be aware of potential drug-
`drug interactions due to saturationfinhibition of intes-
`tinal transporters.
`3.1.4. Protection against Fast Metabolism (Slow-
`Release Prodrugs). The modification of metabolic
`pathways (metabolic switching) to alter the biotrans-
`formation pattern and protect against fast metabolic
`breakdown is another validated concept in prodrug
`design. While drugs with decreased metabolic stability
`(soft drugs) will be considered in section 3.3. we focus
`here on prodrugs offering increased metabolic or chemi-
`cal stability and hence intensified and prolonged activ-
`ity. In other words. the prodrug here is a protected form
`of the drug. Cases where the prodrug protects the drug
`against fast metabolism are presented first. At the end
`of the section, we discuss prodrugs that release a highly
`reactive and unstable active agent.
`Metabolically labile but important pharmacophoric
`elements can be masked or capped to avoid rapid
`metabolism. In bambuterol (Figure 9). a prodrug of
`terbutaline, the phenolic groups are masked as carbam-
`ates.“ This designed,
`long-acting bronchodilator is
`hydrolyzed by nonspecific cholinesterases. Its competi-
`tive advantage is demonstrated by once-daily dosage
`compared to 3 times daily for terbutaline.
`A wealth of other examples can be briefly mentioned.
`Thus, docarpamine (Figure 9) is a cardiotonic, orally
`bioavailable double prodrug of dopamine whose bis-
`ethylcarbonate groups are cleaved in the intestine.
`followed by amide hydrolysis and conjugation in the
`liver.“ Yet another example is levormeloxifene (Figure
`9). an 0-methylated prodrug of a selective estrogen
`receptor modulator (SERM).4” Oxidative demethylation
`in vivo leads to enhanced oral bioavailability of the
`parent drug. Estrogen deficiency is also the target of
`SCH-57050, a bis-pivaloyl derivative of a 4-hydroxy—
`phenylchromenfl-ol compound. Despite such success
`stories, one should not forget that in oxidative activation
`catalyzed by the cytochrome P450 (CYP) enzymes,
`special attention should be paid to potential drugadrug
`interactions. Concomitantly administered drugs utiliz-
`ing the same CYPs for metabolism can produce drug-
`drug interactions potentially causing unacceptable sys-
`temic exposure and hence toxicity.
`
`
`
`mitomycin
`
`O
`
`Nm
`
`c;o0cH3
`
`QOOCH3
`
`5
`
`CI
`
`5
`
`Cl
`
`c|0pidogre|
`
`2-Oxo-clopidogrel
`
`l
`
`QOOH
`
`woe
`
`CI
`
`5
`
`l
`
`QOOCH3
`
`HS
`
`CI
`
`Figure 10. Prodrugs activated to a highly unstable active
`agent. The antiaggregating agent clopidogrel undergoes ex-
`tensive hydrolysis in humans (ca. 85% of a dose) to the inactive
`acid. A smaller proportion of the dose is activated by cyto-
`chrome P450 3A to 2-oxoclopidogrel, which irreversibly an-
`tagonizes platelet ADP receptors via a covalent S-S bridge.“
`
`Groups of particular interest are prodrugs activated
`to a highly unstable active agent. Thus. a number of
`antitumor drugs are in fact reduced within tumor cells
`to a reactive metabolite. This is exemplified by mito-
`mycin, which is reduced by NADPH-cytochrome c re-
`ductase to provide the active semiquinone, which would
`not be stable enough to be effective in the cerebrospinal
`fluid (Figure 10).“ While NADPH-cytochrome creduc-
`tase is not specific for tumor cells, its activity is favored
`in such cells because of their higher reductive capacity.
`A comparable situation exists for the polyol nitrates
`marketed as coronary vasodilators (e.g., glycerol tri-
`nitrate. isosorbide dinitrate, isosorbide mononitrate),
`which can be seen as clinical release forms of nitric
`oxide.
`
`A recent and quite revealing example is that of
`clopidogrel. whose mechanism of activation and action
`has been uncovered only recently‘? (Figure 10). Clo-
`pidogrel is an antiaggregating medicine that under-
`goes extensive hydrolysis in humans (ca. 85% of a dose)
`to the inactive acid. A smaller proportion of the dose is
`activated by cytochromes P450 3A to 2-oxoclopidogrel,
`which spontaneously and rapidly hydrolyzes to a highly
`unstable thiol.43 The latter is the active agent. whose
`mechanism of action involves irreversible binding to
`platelet ADP receptors via a covalent S-S bridge.
`3.2. Tissue-Selective Delivery. Selective delivery
`to the target cells or tissues is known to the public as
`the "magic bullet” metaphor.“ While some drugs have
`a built—in capacity to accumulate in target tissues or
`cells. such a favorable pharmacokinetic feature is usu-
`ally the objective of a prodrug strategy. And indeed,
`rationally designed tissue targeting is probably the most
`exciting objective of a prodrug strategy. This can be
`achieved by (a) passive enrichment in the target tissue,
`(b) targeting specific transporters. (c) targeting tissue-
`or cell-specific enzymes. and (d) targeting surface an-
`tigens. These strategies are presented and exemplified
`in turn below.
`
`Patent Owner. UCB Pharma GmbH — Exhibit 2050 - 0005
`
`

`
`2398 Journal‘ of Medicinal Chemistry, .2004, V01. 47. No. 10
`
`Perspective
`
`d
`
`(Iii
`H
`N
`
`s
`
`.5
`
`H*
`N
`.4
`r 4?
`
`omeprazole
`
`/0
`
`I
`0
`
`n —
`@~’.+~w \
`'
`
`O
`
`H §
`O
`
`
`HO OH
`
`NH
`
`levodopa
`
`dopamine
`
`Figure 11. Example of prodrugs targeting the brain.
`
`3.2.1. Passive Enrichment in the Target Tissue.
`Cytostatic and cytotoxic agents conjugated to PEG with
`MW > 50 000 (e.g., PEG—paclitaxel, Figure 2) are
`reported to exhibit an improved therapeutic efficacy due
`to their longer half-lives and selective accumulation in
`tumor cells. This passive tumor targeting is called the
`"enhanced permeability and retention” (“EPR") effect.
`and its mechanism is not fully understood.“ Clinical
`phase II trials are ongoing using a PEG—Ala-campto-
`thecin conjugate (prothecanw) for the treatment of
`stomach tumors. The low drug load achievable by such
`macromolecular conjugates requires more potently cy-
`totoxic agents than traditional anticancer drugs.
`An interesting prodrug strategy for brain targeting
`is that of site~specif1c chemical delivery systems (CDS).’”
`After brain penetration, a lipophilic prodrug is converted
`there into a more hydrophilic molecule that remains
`"locked in". The same conversion taking place in the rest
`of the body results in increased peripheral elimination.
`The most frequently reported CDS is the 1,4—dihydro-
`trigonelline~trigonelline targetor system attached to
`the drug via an esterase-sensitive bond (Figure 11).“
`1.4-Dihydrotrigonelline is rapidly oxidized by oxidoreduc-
`tases, resulting in the formation of the inactive quater-
`nary ammonium metabolite in the periphery and the
`brain. When formed in the periphery, this polar me-
`tabolite is rapidly excreted. In contrast, its formation
`in the brain leads to entrapment followed by enzymatic
`hydrolysis. This liberates the drug in situ, whereas the
`polar quaternary pyridinium ion moiety is rapidly
`eliminated from the brain, probably by a transporter.
`Dexamethasone—CDS and estradiol—CDS entered clinical
`
`development (phase I/II), but trials were stopped in
`2001.
`
`Prodrug activation in osteoclasts and osteoblasts can
`be utilized for bone targeting. The osteotropic drug
`delivery system (ODDS) relies on the bone-homing
`properties of bisphosphonates for the targeted delivery
`of drugs to the bone and bone marrow.“ A few examples
`of ODDS delivery systems of estrogen and diclofenac are
`in exploratory stages.
`3.2.2. Targeting Specific Transporters. The num-
`ber of well-documented examples of prodrugs targeting
`tissue-specific transporters is relatively limited. Levodopa
`affords an apt example of enrichment resulting from
`active import of the prodrug into the brain followed by
`tissue entrapment of the active metabolite dopamine
`
`‘o@J»~~:« we :1
`
`...
`
`N_
`
`S
`
`I
`O 4 r__
`
`NH
`
`sulfenamidez Active specices
`
`sulfenic acid
`
`S‘oH
`
`Figure 12. Activation of omeprazole.
`
`
`
`Amtolrneiin guacil: prodrug of lolmelin
`
`Figure 13. Examples of prodrugs for colon targeting or to
`avoid gastrointestinal side effects.
`
`(Figure 11).49 Levodopa is a substrate for the neutral
`amino acid transporters present at the blood—brain
`barrier (BBB). After brain entry, levodopa is decarbox—
`ylated to dopamine, which can act locally, being no
`longer a substrate for the neutral amino acid trans-
`porter. To increase the systemic half—life of levodopa, it
`has become customary to coadminister peripherally
`acting inhibitors of DOPA decarboxylase.
`Prodrugs that target the liver utilize transporters
`predominately expressed in this organ. There are pub-
`lications on bile acid conjugates of various cytostatic
`compounds, but no clinical reports appear to have been
`published.5°
`3.2.3. Targeting Tissue- or Cell—Specific En-
`zymes. The acidic environment of the stomach is
`utilized by the proton pump inhibitor omeprazole (Fig-
`ure 12).51~52 This potent antiulcer agent is a prodrug of
`a sulfenamide that exerts its effects by covalently
`modifying cysteine residues on the luminal side of the
`proton pump. i.e., the H+fK+—ATPase of the parietal cell
`in the oxyntic mucosa of the stomach. The prodrug
`omeprazole is only activated in the acidic enviroment
`of the oxyntic mucosa of the stomach where the drug
`exercises its antisecretory effect.
`The reductive environment or the bacterial flora can
`
`be utilized for colon targeting. Balsalzide, a prodrug of
`5-aminosalicylic acid (5-ASA) and an analogue of sulfa»
`salazine. is specifically converted to 5-aminosalicilic acid
`by azo-reducing bacteria present in the colon (Figure
`l3).53 The prodrug remains intact in the gastrointestinal
`tract until it reaches the colon. where it releases 5-ASA
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2050 - 0006
`
`

`
`Perspective
`
`capacitabine H0-/_/_
`
`Journal ofMed1'cina1 Chemistry, 2004, V01. 4 7, No. 10 2399
`
`\
`
`N—§
`F
`O
`N
`/
`N
`w 0
`
`OH
`
`'
`
`carboxyiesterasa
`——u
`Liver
`
`HO
`
`0
`
`F
`cyiidin-
`f/'S,NH2 d68mi|"859
`
`I
`N
`N
`71’
`liver 5.
`
`tumor
`
`OH O
`
`be
`
`O
`
`thymidine
`phosphoryiase
`
`
`,.—
`O
`‘i_7’”7r”“
`,
`,
`tumor
`HO
`OH O
`Figure 14. Activation of capecitabine.
`
`HN
`
`F
`
`/
`
`°
`
`NH
`
`237/
`
`-
`
`to exert a local anti—inflammatory effect directly on the
`inflamed lining of the intestinal wall.
`Alternatively to colon targeting, many prodrug projects
`are directed toward avoiding gastrointestinal side ef-
`fects. A nice example of prodrug design is amtolmetin
`guacil, a lipophilic prodrug of the anti—inflammatory
`drug tolmetin created to reduce gastrointestinal side
`effects. In addition to beneficial lipophilicity,
`its hy-
`drolysis in the gastrointestinal (GI) tract releases gua-
`jacol, which in turn activates inducible gastric nitric
`oxide synthase,54 thereby reducing GI irritation (Figure
`13). It is interesting in this context to mention that the
`extensively investigated esters of nonselective COX
`inhibitors did not show the desired GI protection, since
`GI side effects are a consequence of COX-1 inhibition
`and not of local irritation.
`
`Specific targeting of virus-infected cell is elegantly
`realized by the antiviral prodrug aciclovir (Figure 1) and
`all other selective antimetabolites (a subgroup of bio-
`precursor prodrugs) used as antiviral agents. The activ-
`ity of such antimetabolites implies their intracellular
`phosphorylation by kinases to form a nucleotide ana-
`logue that inhibits DNA synthesis. The older nonselec-
`tive antiviral drugs were phosphorylated by native and
`by virus-induced thymidine kinases. In contrast, the
`selective antimetabolites are phosphorylated only by
`kinases coded by the viral genome and expressed in
`infected cells. As a result. aciclovir and all other
`selective antimetabolites are active only in infected
`cells.55
`
`The selective delivery of cytotoxic drugs to tumor cells,
`without concomitant damage to normal tissues. is a
`major challenge in cancer chemotherapy. Prodrugs for
`active tumor targeting are therefore of high interest.
`Tumor specificity can be achieved by targeting trans-
`porter systems, enzymes having a higher activity in
`tumor cells, as discussed here. or surface markers (see
`section 3.2.4) or by gene-directed enzyme prodrug
`strategy (see section 3.2.5).
`Capecitabine, an orally available triple prodrug of
`5—fluorouracil (5—FU), offers a good example for a pro-
`drug activated by tumor-specific enzymes (Figure 14).
`Following oral absorption, capecitabine undergoes three
`activation steps resulting in high tumor concentra-
`tions of the active drug, namely, (a) carboxylesterase
`mediated hydrolysis in the liver. (b) cytidine deaminase
`mediated deamination in the liver and tumor cells, and
`(c) specific liberation of 5—FU by thymidine phosphoryl-
`ase in tumor cells.“
`
` mAb-NH T
`
`Lysine attachment
`
`Figure 15. N-Acetyl-y-calicheamicin linked via a hydrolyti-
`cally labile hydrazone linker to a lysin N -atom of recombinant
`humanized anti-CD33 antibody.
`
`3.2.4. Targeting Surface Antigens. Targeting sur-
`face markers appears as a selective and promising strat-
`egy in developing tumor-selective prodrugs. In 2000, the
`FDA approved the use of CMA—676 {gemtuzumab ozo-
`gamicin, mylotarg) for the treatment of acute myelotic
`leukemia (AML).57 CMA-676 is an antibody-targeted
`chemotherapy agent consisting of a recombinant hu-
`manized anti-CD33 antibody linked via the lysin N-
`atom to N—acetyl—y-calicheamicin, a potent cytotoxic
`agent {Figure 15). The linker incorporates a hydrazone
`moiety, a site for hydrolytic release.” previously shown
`to be required for activity.“ Since this antibody recog-
`nizes AML blast cells but not hematopoietic stem cells,
`it allows selective delivery of the cytotoxic agent to
`leukemia cells. It was demonstrated that after CMA—
`
`676 infusion, the CD33 sites are quickly saturated and
`internalization of the conjugate occurs. However, elimi-
`nation of leukemia correlates with the low capacity of
`these cells to extrude the drug conjugate. This finding
`is in good agreement with the fact that CMA-676 was
`ineffective in multidrug-resistant (MDR) sublines ex-
`pressing P-glycoprotein. However, by combination of
`CMA-676 and MDR modifiers. an increased effect in
`MDR AML may be possible. While there are numerous
`preclinical reports on such carrier—linked prodrugs,
`successful clinical developments are slow to come.
`3.2.5. Enzyme—Prodrug Cancer Therapy. Selec-
`tive activation of prodrugs in tumor tissues can also be
`achieved by exogenous enzymes in a two—step approach.
`Prodrug-activating enzyme gene or functional protein
`is delivered selectively to tumor tissues, followed by
`systemic administration of a nontoxic prodrug that is
`activated by the exogenous enzyme. The net gain is a
`high local concentration of an active anticancer drug in
`tumors. Delivery of the enzyme gene is accomplished
`via viral vectors (VDEPT, virus—directed enzyme pro-
`drug therapy) and