Formulation technologies to overcome poor drug-like propertiesOvercoming poor permeability – the role of prodrugs for oral drug delivery
Introduction
The initial goal of a lead optimization program is to increase the affinity of the compounds at the target with an eye on maintaining the ligand efficiency as high as possible. This initial phase often leads to candidates with less than optimal physical or pharmacokinetic properties and commonly, a great deal of effort is devoted to improve the metabolic stability and the pharmacokinetic characteristics, such as in vivo half life. Unfortunately, this results frequently in pre-clinical leads with either poor solubility or suboptimal permeability or both, in other words in compounds belonging to the Biopharmaceutics Classification System (BCS) classes II–IV [1]. The medicinal chemists is at this point are searching for the ideal compound, one with the desired potency, the proper physical and pharmacokinetic properties. For compounds with poor solubility and/or low permeability, solutions are being sought with experts in formulations and options for prodrugs are being explored. Fortunately, it is now common that these alternate approaches are being investigated early on and form an integral part of late lead optimization programs. This review focuses on detailing the principles applied to the identification and optimization of prodrugs to either improve solubility or permeability, or both, and to employ recent examples as a guide through the discussion [2].
Before embarking on a prodrug effort it is imperative to fully understand the problem and to have a clear target in terms of dose and route of administration. With these prerequisites in place, an assessment is possible if a prodrug approach will likely be successful and which kind of prodrug should be investigated first – one to increase solubility or one to improve permeability or one to improve both.
A typical evaluation cascade in such an effort is shown in Fig. 1. The first tier consists of in vitro evaluations and will allow for an immediate assessment if progress toward improving solubility and lipophilicity is being made. A renewed search for novel prodrugs is initiated should an insufficient improvement in the measured property be observed at any step in the cascade. This iterative process will continue until a prodrug with acceptable properties is obtained to advance it to pre-clinical safety assessments. The branching indicated between prodrugs targeted at improving solubility from the ones improving permeability is shown to illustrate the point that not all tests need to be conducted at every step with every compound. An efficient screening paradigm will be flexible and concentrate on the most critical aspects at the given moment.
In Fig. 1, the Caco2 experiment stands for any test to assess permeability, including the assessment of the involvement of transporters; either influx, such as the human intestinal peptide transporter PEPT1, or efflux transporters, primarily PgP. Thus, such a test needs to be conducted with the appropriate controls to ensure the proper expression levels of transporter proteins. In Caco2 cells, this is routinely achieved at the level of 21 day cells, but not with cells after seven days of incubation.
All prodrugs, independent of route of administration, will need to fulfil as much as possible the following typical requirements:
- (a)
Prodrug attachment should add negligibly to the cost of goods; that is be based on easy synthesis and inexpensive prodrug moieties.
- (b)
The prodrug moiety should be readily cleaved in the desired tissue or cellular compartment. In most cases, the cleavage should be rapid and complete. Prolonged, slow cleavage may be desirable as a special form of a controlled release mechanism.
- (c)
The prodrug or the prodrug moiety should not add an additional side effect or toxicity burden to the parent drug; that is being readily excreted from the body and not possessing any inherent toxic potential.
In addition to pre-systemic metabolism, a low oral bioavailability can be a result of either low solubility or low permeability or both. For an oral drug, a minimum solubility in the gastric environment is required. This is dependent on the dose necessary to achieve the desired therapeutic effect. Permeability has to be sufficient to deliver a therapeutically effective amount of drug into the portal vein. If the permeability is low, the greater the solubility of the drug has to be to result in the desired systemic exposure level. Thus solubility and permeability of oral drugs are necessarily linked and cannot be considered individually in isolation.
Many therapeutic agents are of a highly polar nature or are charged at physiological pH, which impacts the permeability significantly. Prominent examples among such compounds are carboxylic acids, strongly basic compounds such as amidines and guanidines, nucleosides and nucleotides, for the latter mainly analogs of nucleoside-monophosphates. For a comprehensive review, a number of recent reviews should be consulted [2]. Here, just a few recent examples are presented, with which the principles of such prodrugs are discussed.
There are a great many examples in the literature for the use of ester derivatives to enhance the bioavailability of carboxylic acids and the area is well reviewed [3]. Here included is just one recent example to illustrate the benefit of esters in enhancing the permeability and thus oral bioavailability of an ester. The calcium receptor antagonist 1 is a zwitter-ionic acid with a molecular weight of 447 and a barely measurable bioavailability of 0.3% in rats (Fig. 2). The ethyl ester 2 boosted the bioavailability as measured by the acid 1 in the same species 30-fold [4]. The prodrug could not be detected in the systemic circulation, an observation which is typical for ester prodrugs. The enzymes involved in these conversions have been studied extensively providing strong guidance to which ester should be tried first as a prodrug derivative [5].
The example chosen is the inhibitor of the Heat Shock Protein 90 named SNX-2112 and its prodrug SNX-5422 (Fig. 2). In a typical medicinal chemistry optimization program, SNX-2112 emerged as the optimal compound from an activity spectrum [6]. The amorphous form had a reasonable solubility and the compound in this form showed an acceptable bioavailability of about 40% in mice. Once a crystalline form was identified, the solubility at physiological pH (7.4, buffer) was reduced significantly (25-fold, to about 3 μg/mL) and with it the oral bioavailability. Thus, a prodrug modification had to serve a dual purpose – increasing solubility and permeability. Conveniently, the molecule offered a suitable handle for prodrug attachment in the secondary alcohol. The glycine derivative SNX-5422 fulfilled the requirements, demonstrating a solubility of 10 mg/mL and a bioavailability of about 80% in mice as measured by the parent SNX-2112. The moderate pKa of the amino group of the glycine pro-moiety (pKa approx. 8 [7]) makes this molecule uncharged in the small intestine and this is likely the reason that the permeability is so enhanced while also providing good solubility to the compound in the acidic environment of the stomach. A more strongly basic amine would have retained the good solubility profile while likely displaying reduced permeability.
The classical prodrug modification of the very basic benzamidine functional group (pKa of approx. 11.6) is the N-hydroxy modification, which lowers the pKa drastically (to about 5.0) [8] with Ximelagatran, an anticoagulant evaluated extensively as a replacement to warfarin, is a well reviewed, commercially successful example. Ximelagatran has a bioavailability in humans of 20%, an improvement over that of the non-hydroxylated compound, melagatran, with a bioavailability of 6% [9] (Fig. 3a). Recently, the bis-hydroxylation of benzamidine has been disclosed as an improved prodrug of benzamidines [10]. The oral bioavailability in the pig of this model compound 3 was 91%, while that of the mono-hydroxylated analog 4 was 74% (Fig. 3b). Both forms are converted by reduction to the benzamidine. The lower bioavailability of the mono-hydroxylated compound is believed due to partial pre-systemic reduction to benzamidine, which is less likely with the bis-hydroxylated compound.
An alternative approach to increasing the lipophilicity through a reduction in the pKa is a prodrug modification which targets one of the uptake transporters on the endothelial cells in the small intestine. The most often targeted transporter is the human peptide transporter 1 (hPEPT1) due to its large capacity and a relative broad substrate specificity [11]. The general recognition elements have been described. The application of such an approach to a model benz-guanidine 5 has been detailed, including the clarification of the cleavage mechanism [12] (Fig. 3c). The valine derivative 6 is a good substrate not only for the hPEPT1 transporter but also for the human valacyclovirase (hVACVase) present in the same Caco2 cells. Indeed it has been shown for several substrates of the hPEPT1 transporter that they are also substrates for the associated hydrolase, among them many nucleosides (see below).
Another interesting example is the sulfenamide prodrug 7 of metformin, an important agent for the treatment of diabetes [13] (Fig. 3d). This highly polar bisguanidine metformin (log D (pH 7.4) = −3.37) has a remarkable bioavailability in the rat of 43%, likely facilitated by its small size. In humans, the drug is highly effective but suffers from serious gastrointestinal side effects at therapeutic doses. Thus it is critical to increase the amount absorbed to reduce the potential for such side-effects. While the formation of amides or carbamates was tried, this was not successful as a result of these metformin prodrugs to undergo rearrangement or cyclization reactions. Two sulfonamide prodrugs were characterized in detail, the phenyl- and cyclohexyl-sulfenamides, 8 (log D (pH 7.4) = −0.76) and 7 (log D (pH 7.4) = 0.49, rat bioavailability = 64%), respectively. The bioconversion is believed to occur by endogenous thiols with the formation of a bisulfide, which can break down further to the individual thiols. The phenyl analog 8 releases thiophenol, which was toxic to the rats in the initial in vivo experiment and thus was removed from further considerations. Recently, a patent application disclosed the cysteine prodrug of metformin 9, however details regarding bioavailability were not made public [14].
Flavonoids are poly-hydroxylated aromatic compounds with interesting biological properties, however the compounds suffer from low bioavailabilities [15]. A great number of prodrug approaches have been described, including simple phenolic esters and ethers. In a recent paper a focused investigation into prodrugs targeting the hPEPT1 transporter of the flavonoid tricin is described [16]. Tricin has anti human cytomegalovirus (HCMV) activity, but poor bioavailability typical of this compound class. A systematic search for the proper amino acid attached at the sterically hindered phenol group was undertaken, with the permeability of the compounds evaluated in Madin-Darby canine kidney (MDCK) cells; a test system which is similar to the Caco2 cell line. The results showed that the single amino acids gave a modest two- to threefold enhancement in permeability, the dipeptide alanine-glutamic acid gave a sixfold enhancement in permeability in MDCK cells and a 45-fold exposure of the parent tricin in rats after oral administration of the prodrug (Fig. 4). This example shows, that it is possible to direct a prodrug effort toward a facilitated uptake mechanism by utilizing the data from related compounds in the literature, in this case the Ala-Glu prodrug of quercetin [17].
The nucleoside class is characterized by highly polar compounds with typically good aqueous solubilities. Prodrugs are employed to enhance the permeability through either passive diffusion or by targeting an active uptake transporter, typically hPEPT1 [18]. Classical examples for such an approach are the two commercially successful antiviral compounds, Valacyclovir and Valganciclovir. That either passive or active uptake mechanism can be employed for the same nucleoside is shown here by the two Gemcitabine derivatives, 10 and 11 (Fig. 5). The 5′-O-l-valyl prodrug 10 was shown to be a good substrate of the hPEPT1 transporter, along with the 5′-O-l-isoleucyl analog, achieving a high apical to basolateral flux in Caco2 cells [19]. Concomitant with transport, most of the prodrug moiety is cleaved during passage through the cell monolayer. Thus, the dominant species reaching the liver is the parent nucleoside Gemcitabine. Mainly in the liver, Gemcitabine is being deaminated to the inactive uracil metabolite. To avoid this first pass metabolism to the uracil metabolite, the N-4-valproyl amide was investigated by a group at Lilly [20]. Indeed, 11, has good permeability and is sufficiently stable to reach the portal vein unchanged. Gemcitabine is liberated in the liver leading to prolonged Gemcitabine exposure after an oral dose. It is believed that this prodrug is currently in a phase 1 clinical study by Lilly.
In many cases, the simple esterification of the hydroxyl groups in a nucleoside is sufficient to enhance the permeability to suitable levels. The example of 2′-deoxy-2′-spirocyclopropylcytidine serves to illustrates this point [21]. The bis-isobutyrate ester 12 demonstrated a 24-fold higher bioavailability, as measured by the parent) in rats and a dose proportional increase in exposure between 10 and 90 mg/kg. Cleavage of the esters is rapid, with none of the bis-ester and only 0.2% of the 3′-O-monoester detected besides parent in the systemic circulation.
With the introduction of phosphonates, such as Tenofovir in the form of its prodrug Tenofovir diisoproxil fumarate [22] (Fig. 6a), into clinical practice along with the recognition that several nucleosides are inactive due to a lack of a metabolic activation to the corresponding 5′-monophosphate, an intense effort has resulted in a number of 5′-monophosphate prodrugs. Most prominent among these is the series of phosphoramidates championed by McQuigan [23], recently applied to the uridine analog of the cytidine PSI-6130 [24]. In vivo, the cytidine is phosphorylated by deoxy-cytidine-kinase to the corresponding 5′-monophosphate, while the corresponding uridine is not. At the tri-phosphate level, both are active as inhibitors of the hepatitis C virus polymerase (NS5B). The phosphoramidate PSI-7977, being developed as a single diastereomer, was selected based on a detailed analysis of its stability and its activation mechanism to the triphosphate in hepatocytes. PSI-7977 is soon entering phase 3 clinical trials for the treatment of hepatitis C (Fig. 6b).
While the original McQuigan phosphoramidates are chiral at phosphorous, the most recently disclosed phosphorodiamidates are not [25]. In a detailed analysis applied to the 6-OMe-2′-CMe-guanosine, it was demonstrated that small amino acids are preferred in this symmetrical phosphordiamidate. The bis-alanine derivate 13, gave the identical liver triphosphate levels as the unsymmetrical phosphoramidate 14, thus offering a potentially easier to develop monophosphate prodrug (Fig. 6c).
Perzinfotel is a potent and competitive N-methyl-d-aspartate (NMDA) receptor antagonist and not surprisingly, this negatively charged phosphonate demonstrated low oral bioavailability. Through a fairly comprehensive investigation of potential prodrugs, the bis-methylene linked benzoic acid ester 15 (Fig. 7a) emerged as the compound with the best overall properties in terms of bioavailability and stability in plasma, simulated gastric fluid (SGF), simulated intestinal bile fluid (SIBLM) and simulated intestinal fluid (SIF) [26]. Liberation of the parent drug releases formaldehyde and benzoic acid. The formaldehyde issue has been discussed extensively in the past and it is now generally accepted that the amounts of formaldehyde released from prodrugs do not pose a safety concern [27].
Fructose-1,6-bisphophatase (FBPase) is critical in the control of glucose levels and thus an interesting target for the control of type 2 diabetes mellitus (T2DM). Designed as AMP mimics as inhibitors of this enzyme, the phosphonate 16 emerged as a lead (IC50 = 90 nm) [28]. Attempts to prepare prodrugs, including phosphoramidates and phosphordiamidates were unsuccessful in providing useful systemic exposures of the parent. This was attributed to the high molecular weight of the inhibitor, with the potential prodrugs having molecular weights above 600. A renewed search led to a smaller scaffold, which could be successfully derivatised to the symmetrical prodrug 17, having a molecular weight of 500. Following detailed in vivo studies in animals, this compound was advanced into the clinic as the first orally active FBPase inhibitor (Fig. 7b).
Section snippets
Conclusions
Prodrugs are an effective way to rescue poorly permeable compounds and to convert them into useful drugs. In this brief review, a number of recent examples were used to discuss the approaches taken resulting in successful prodrugs. Through the analysis of the examples a number of guiding principles emerged to facilitate the design and identification of prodrugs:
- (1)
The issue(s) with the parent drug candidate needs to be understood in detail:
- (a)
Is it a solubility problem?
- (b)
Is it a permeability issue?
- (c)
Need
- (a)
Conflict of interest
The author declares no conflict of interest
References (28)
- et al.
Enzymes involved in the bioconversion of ester-based prodrugs
J. Pharm. Sci.
(2006) - et al.
Your prodrug releases formaldehyde: Should you be concerned? No!
J. Pharm. Sci.
(2008) A theoretical basis for a biopharmaceutics drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability
Pharm. Res.
(1995)- A number of comprehensive reviews have been published: a. Rautio, J., ed. (2011) Prodrugs and Targeted Delivery, Volume...
Prodrugs of carboxylic acids
Antagonists of the calcium receptor. 2. Amino alcohol-based parathyroid hormone secretagogues
J. Med. Chem.
(2009)Discovery of novel 2-aminobenzamide inhibitors of heat shock protein 90 as potent, selective and orally active antitumor agents
J. Med. Chem.
(2009)Protonation equilibria and alkaline hydrolysis of glycine ethyl ester
Nature
(1965)Prodrugs of benzamidines
Melagatran and ximelagatran
Drugs Future
(2001)
N,N′-Dihydroxyamidines: a new prodrug principle to improve the oral bioavailability of amidines
J. Med. Chem.
How to make drugs orally active: a substrate template for peptide transporter PepT1
Angew. Chem. Int. Ed.
Enhancing the intestinal absorption of molecules containing the polar guanidino functionality: a double-targeted prodrug approach
J. Med. Chem.
The first bioreversible prodrug of metformin with improved lipophilicity and enhanced intestinal absorption
J. Med. Chem.
Cited by (14)
Biocatalytic approaches applied to the synthesis of nucleoside prodrugs
2015, Biotechnology AdvancesCitation Excerpt :For this purpose, excellent reviews are available (Fresco-Taboada et al., 2013; Lewkowicz and Iribarren, 2006; Mikhailopulo, 2007; Mikhailopulo and Miroshnikov, 2011). The only exceptions included in this paper are those nucleoside analogues that are in vivo activated by routes different to phosphorylation (Maag, 2012; Zhang et al., 2014). For organisative purposes, the chemo-enzymatic preparation of these compounds is divided in nucleoside and nucleobase prodrugs.
Chemistry-enabled drug delivery (prodrugs): Recent progress and challenges
2014, Drug Discovery TodayCitation Excerpt :By avoiding metabolism by CYP enzymes, dabigatran etexilate is reported to have reduced adverse events compared with ximelagatran [24]. The first-line approach for increasing permeation via an addition of a promoiety that neutralizes charge(s) and/or increases lipophilicity has been recently reviewed by Maag [31]. Using prodrug conjugates that target membrane transporters or enzymes also affords improved drug bioavailability, selectivity and bioconversion of poorly absorbed molecules [4,32].
The expanding role of prodrugs in contemporary drug design and development
2018, Nature Reviews Drug Discovery