Available online at www.sciencedirect.com
naDrug discovery projects are in danger of failure shouldpharmacodynamic (PD) and pharmacokinetic (PK) pre-requisites prove chemically incompatible that is imposs-ible to combine in a single molecule. Because prodrugsand their active metabolite have different chemical struc-tures, they must differ in their biological behavior, a factthat opens the door to dissociating PD and PK properties.
PK objectives are currently the most important ones inprodrug research. Foremost among these is firstly, a needto improve oral bioavailability, be it by improving the oralabsorption of the drug and/or by decreasing its presyste-mic metabolism. Other objectives are secondly, toimprove absorption by parenteral (nonenteral, e.g. dermaland ocular) routes; thirdly, to lengthen the duration ofaction of the drug by slowmetabolic release; and finally to
Current Opinion in Chemical Biology 2009, 13:338344 www.sciencedirect.comProdrugs: bridging pharmacodyBernard Testa
In this mini review, prodrugs are discussed with a focus on their
pharmaceutical, pharmacokinetic, and pharmacodynamic
objectives, as well as on the resulting therapeutic benefits.
Carrier-linked prodrugs remain the most extensively
investigated and receive due attention here with recent
successes highlighted. A clear trend is apparent in modern
prodrug research, namely the increased attention given to the
knowledge-based design of bioprecursors, namely prodrugs
devoid of a detachable promoiety. In most cases, such
prodrugs are activated by in situ reduction, hence their
designation as bioreductive prodrugs. This is a particularly
active field in the design of more selective, small-molecule
antitumor agents. New antimicrobial agents are also in the
pipeline. In addition, biooxidative bioprecursors offer a
promising strategy in specific cases, as illustrated by the
successful antiaggregating agent clopidogrel.
Pharmacy Department, University Hospital Centre, CHUV/BH-04, 46
Rue du Bugnon, CH-1011 Lausanne, Switzerland
Corresponding author: Testa, Bernard (Bernard.Testa@chuv.ch)
Current Opinion in Chemical Biology 2009, 13:338344
This review comes from a themed issue on
Edited by KarlHeinz Altmann and Dario Neri
Available online 25th May 2009
1367-5931/$ see front matter
# 2009 Elsevier Ltd. All rights reserved.
IntroductionWhat makes prodrugs different from other drugs is thefact that they are devoid of intrinsic pharmacologicalactivity [1,2,315]. Thus, the simplest and clearestdefinition, in this writers view, is that given by Albert, who coined the term. In modified form, the defi-nition reads: Prodrugs are chemicals with little or nopharmacological activity, undergoing biotransformationto a therapeutically active metabolite.mic/pharmacokinetic gaps
Why prodrugs?The prodrug concept has found a number of usefulapplications in drug research and development. Thesecorrespond to a number of objectives, as presented inFigure 1. However, it should be clear that such a view istoo schematic, these objectives being often intertwined.Thus, an improved solubility can greatly facilitate oralabsorption, while improving the chemical stability of anactive agent can allow tissue-selective delivery and evenlead to its in situ activation.
Pharmaceutical scientists are often faced with seriousformulation problems resulting from poor solubility,insufficient chemical stability or poor organoleptic prop-erties. While pharmaceutical technology can solve suchproblems in favorable cases (e.g. by improving the solu-bility of cyclosporine), success is not guaranteed and maybe time-consuming to achieve.
An example of chemical interest is that of isotaxel (1 inFigure 2), a prodrug of the potent antitumor agent pacli-taxel (2 in Figure 2) [17,18]. Paclitaxel has a very poorwater solubility (about 0.25 mg/L) which necessitatescoinjection of a detergent. Taking advantage of thewell-known mechanism of intramolecular nucleophilictransacylation, an analog of paclitaxel was prepared hav-ing the benzoyl group located on the a-hydroxy grouprather than on the amido group. What is an amido func-tion in paclitaxel thus became a basic amino group inisotaxel, and solubility was greatly improved by protona-tion (water solubility of 1HCl is 0.45 g/L). O,N-Migrationof the benzoyl moiety to the active paclitaxel was fastunder physiological conditions of pH and temperature,but much slower under slightly acidic, pharmaceuticallycompatible conditions.
Note that increasing solubility is not only a pharmaceu-tical but also a PK objective. Indeed, and as made explicitby the biopharmaceutics classification scheme (BCS) ,solubility is one of the main factors influencing oralabsorption.
Bridging pharmacodynamic/pharmacokinetic gaps Testa 339achieve the organ/tissue-selective delivery of an activeagent. Some of these objectives are exemplified belowwith clinically or potentially useful prodrugs.
Achieving improved oral absorption by a prodrug strategyis a frequent rationale in marketed prodrugs, as aptlyillustrated by neuraminidase inhibitors of therapeuticvalue against type A and B influenza in humans .Target-oriented rational design has led to highly hydro-philic agents which are not absorbed orally, as exempli-
A schematic classification of objectives in prodrug research. Note that
many of these objectives are in fact intertwined (modified from ).fied by the zwitterion zanamivir which is administered inaerosol form. Another active agent is Ro-64-0802, whichalso shows very high in vitro inhibitory efficacy toward theenzyme but low oral bioavailability, also being a (too)hydrophilic zwitterion by virtue of basic amino and acidiccarboxylate groups. To circumvent this problem, theactive agent was derivatized to its ethyl ester known asoseltamivir (3 in Figure 2). This prodrug is well absorbedorally and undergoes rapid enzymatic hydrolysis to pro-duce high and sustained plasma levels of the active agent(structure marked in red).
Another PK objective is the organ-selective or tissue-selective delivery of a given drug, also known as thesearch for the magic bullet. Targeting a prodrug toenzymes highly expressed in the target organ or tissue isa useful strategy for selective delivery. Given the highconcentration of cytochromes P450 (CYPs) in the liver,it comes as no surprise that CYP-activated prodrugsshould have been designed with a view of selectivelydelivering antihepatitis virus agents to the liver. A goodexample is provided by pradefovir (4 in Figure 2), aprodrug currently in clinical trials as the cis-(2R,4S)-isomer shown here . When examined in rats, this 1-aryl-1,3-propanyl prodrug showed a 12-fold improve-
www.sciencedirect.comment in the liver/kidney ratio of adefovir (structuremarked in red) over adefovir dipivoxil, the prodrug incurrent use.
Targeting prodrugs to enzymes overexpressed in cancertissues is a topic of major interest and promise in drugdiscovery . A clinically significant example is that ofcapecitabine (5 in Figure 2), a multistep, orally activeprodrug of 5-fluorouracil (structure in red) . Capeci-tabine is well absorbed orally and undergoes three acti-vation steps resulting in high tumor concentrations of theactive drug. It is first hydrolyzed by liver carboxylesterase,the resulting metabolite being a carbamic acid whichspontaneously decarboxylates to 50-deoxy-5-fluorocyti-dine. The enzyme cytidine deaminase present in theliver and tumors then transforms 50-deoxy-5-fluorocyti-dine into 50-deoxy-5-fluorouridine. Transformation into5-FU is catalyzed by thymidine phosphorylase and occursselectively in tumor cells.
Capecitabine is of great interest in the context of thisreview as it affords an impressive gain in therapeuticbenefit compared to 5-FU due to its oral bioavailabilityand a relatively selective activation in and delivery totumors. Nevertheless, most of the current attention inantitumor prodrug design is focused on reductive biopre-cursors, namely prodrugs reduced in situ to a cytotoxicmetabolite.
PD objectives can be understood as being synonymouswith decreasing systemic toxicity. Two major cases areillustrated here, namely the masking of a reactive agent toimprove its therapeutic index, and the in situ activation ofa chemotherapeutic agent.
The masking of a reactive agent to improve its thera-peutic index is aptly exemplified by the successful anti-aggregating agent clopidogrel (6 in Figure 3). Thiscompound, whose molecular mechanism of action waspoorly understood for years, is now known to be a prodrug[24,25,26]. It is of interest among prodrugs in that itsmajor metabolic route in humans (about 80% of a dose) isone of hydrolysis which leads to the inactive acid. Incontrast, much of the remaining dose is activated in athree-step sequence. First, CYP3A oxidizes clopidogrel to2-oxo-clopidogrel (7 with the corresponding substituentsR and R0). This is followed by a CYP-catalyzed sulfox-idation to the cyclic sulfoxide intermediate. The latterhydrolyzes spontaneously to a highly reactive sulfenicacid metabolite (8 with the corresponding substituents Rand R0) which forms a covalent SS bridge with a thiolgroup in platelet ADP receptors, thereby causing theirlong-lasting inactivation. The sulfenic intermediate can
also be reduced by glutathione, leading to the correspond-ing thiol which for some years was thought to be the activemetabolite.
Current Opinion in Chemical Biology 2009, 13:338344
340 Next-generation therapeuticsFigure 2It is interesting to compare the activation of clopidogrelwith that of prasugrel (9 in Figure 3), a more recent andsmartly designed analog . In contrast to the for-mer, the latter is activated by carboxylesterases 1 and 2 inhumans. Indeed, C(2) in prasugrel carries an acetyl esterfunction whose hydrolysis yields the correspondinghydroxy-thiophene which tautomerizes to the thiolactone(7 with the corresponding substituents R and R0), a closeanalog of 2-oxoclopidogrel. Sulfoxidation followed byhydrolytic ring opening yields the (yet to be discovered)active sulfenic acid metabolite (8 with the correspondingsubstituents R and R0).
In situ activation to toxic agents is a part of the well-knownmechanism of action of the antibacterial and antiparasiticnitroarenes such as metronidazol. This strategy is findinga recent and promising revival with the discovery of
Some prodrugs discussed in the text. An illustrative prodrug developed for
paclitaxel (2). Examples of carrier-linker prodrugs (active drug shown in red
agents pradefovir (4), and capecitabine (5), a well-known pro-prodrug of 5-f
Current Opinion in Chemical Biology 2009, 13:338344potent bioreductive prodrugs. These are designed tobe activated by a metabolic reaction of reduction in targettissues or organisms with a low oxygen concentration.Such a state of hypoxia may favor reduction reactions, butanother factor accounting for the shift toward reduction isthe hypoxia-increased expression of genes coding forreductive enzymes. A recent and promising illustrationof bioreductive antimicrobial prodrugs is provided by theantitubercular nitroimidazooxazine PA-824 (10 inFigure 3) now in clinical trials . There is evidenceto indicate that the active species in antimicrobial nitroar-enes is their nitrenium cation, a strong and highly reactiveelectrophile .
The case of antitumor bioprecursors is best illustratedwith tirapazamine (11 in Figure 3), since this muchstudied agent and its analogs have revealed inspiring
improved water solubility is namely isotaxel (1), a promising prodrug of
) are the anti-influenza virus agent oseltamivir (3), the antihepatitis virus
Bridging pharmacodynamic/pharmacokinetic gaps Testa 341Figure 3information on their molecular mechanism of action. Tirapazamine is inactivated by two-electronreduction steps catalyzed by quinone reductase, butactivated to a cytotoxic radical (12 in Figure 3) by aone-electron reduction catalyzed by NADPH-cyto-chrome P450 reductase and multiple reductases in thenucleus. The strong redox potential of this benzotriazi-nyl radical allows it to abstract a hydrogen radical fromDNA, thereby causing single-strand and double-strandbreaks.
How prodrugs?Given the stated objectives, how can medicinal chemistsdesign clinically useful prodrugs? As implicit above, twomain classes of prodrugs exist, namely the carrier-linkedprodrugs and the bioprecursors .
A comparison of the activation pathways of the antiaggregant prodrugs clop
CYP-catalyzed reactions, first to the thiolactone 7, then to its sulfoxide (not
sulfenic acid (8). The latter antagonizes platelet ADP receptors via a covalen
carboxylesterase-catalyzed deacetylation, followed by the same subsequen
antitubercular nitroimidazooxazine PA-824 (10) whose activation is by nitror
tirapazamine (11) to the radical 12.
www.sciencedirect.comIn carrier-linked prodrugs, the active agent (the drug) islinked to a carrier (also known as a promoiety). In manycases, carrier-linked prodrugs are esters activated byenzymatic hydrolysis, although nonenzymatic hydrolysismay contribute . The variety of ester prodrugs in theliterature is immense and has served to mask carboxylategroups (e.g. oseltamivir in Figure 2), phosphate groups(e.g. pradefovir Figure 2), acidic tetrazolate groups, oralcoholic or phenolic functions. Examples of promoietiesused to derivatize such active agents include simple orfunctionalized alkyl groups, aryl groups, simple or com-plex carboxylic acids, amino acids, and fractionable moi-eties (e.g. carbonic acid esters).
Amides form another major group of carrier-linked pro-drugs, the active agent usually being an amine. Here,
idogrel (6) and prasugrel (9). The former is activated by two sequential
shown), followed by spontaneous hydrolytic ring opening to the reactive
t SS bridge. The first activation step of prasugrel (9) is by
t steps as clopidogrel. The figure also shows the structure of the
eduction, and the activation of the hypoxia-selective antitumor agent
Current Opinion in Chemical Biology 2009, 13:338344
342 Next-generation therapeuticsFigure 4useful promoieties include carboxylic acids, amino acids,and carbonic acids to yield carbamates. Mannich basesand even imines have also been prepared as potentialprodrugs of amines. Extensive compilations of promoi-eties have been published [5,6,11].
In contrast to carrier-linked prodrugs, bioprecursors donot contain a promoiety yetmay be activated by hydration(e.g. lactones such as some statins), oxidation (e.g. prasu-grel in Figure 3), or reduction (e.g. PA-824 and tirapaza-mine in Figure 3). This is currently a highly active field inthe search for more selective chemotherapeutic agents(antibacterial, antiparasitic, or anticancer prodrugs) acti-vated by reductases . Such prodrugs includenitroarenes, quinones...