Intramolecular hydrogen bonding to improve membrane permeability and absorption in beyond rule of five chemical space

Alexander Alex; David S. Millan; Manuel Perez; Florian Wakenhut; Gavin A. Whitlock

doi:10.1039/C1MD00093D

View PDF VersionPrevious ArticleNext Article

DOI: 10.1039/C1MD00093D (Concise Article) Med. Chem. Commun., 2011, 2, 669-674

Show CompoundsShow Chemical Terms

Intramolecular hydrogen bonding to improve membrane permeability and absorption in beyond rule of five chemical space†

Alexander Alex , David S. Millan *, Manuel Perez , Florian Wakenhut and Gavin A. Whitlock
Worldwide Medicinal Chemistry, Pfizer Global Research and Development, Sandwich Laboratories, Ramsgate Road, Kent, CT13 9NJ, UK. E-mail: david.millan@pfizer.com

Received 31st March 2011 , Accepted 22nd April 2011

First published on 20th May 2011

Abstract

Utilising ‘beyond rule of five’ chemical space is becoming increasingly important in drug design, but is usually at odds with good oral absorption. The formation of intramolecular hydrogen bonds in drug molecules is hypothesised to shield polarity facilitating improved membrane permeability and intestinal absorption. NMR based evidence for intramolecular hydrogen bonding in several ‘beyond rule of five’ oral drugs is described. Furthermore, the propensity for these drugs to form intramolecular hydrogen bonds could be predicted for through modelling the lowest energy conformation in the gas phase. The modulation of apparent lipophilicity through intramolecular hydrogen bonding in these molecules is supported by intrinsic cell permeability and intestinal absorption data in rat and human.

Introduction

Intermolecular hydrogen bonding between drug molecules and biological receptors can be an important interaction for driving potent binding.¹ A requirement for achieving potency through this interaction is for the two heavy atoms and the hydrogen involved to adopt the optimal angle and intermolecular distance to achieve maximal binding potency and compensate for the energy penalty required to desolvate the molecule from water. The specificity of this interaction is one that can drive selectivity for binding of one target over another and therefore is an interaction that can be considered important for minimising the overall promiscuity profile of a drug and in turn the in vivo pharmacological and toxicological profile.²

However, introduction of too many hydrogen bonding motifs into drug molecules can have a deleterious effect on membrane penetration, presumably due to a high water de-solvation penalty as well as unfavourable interactions of the polarised hydrogen with the aliphatic fatty acid side chains in the interior of the membrane.³ Therefore a drug with optimal binding potency and selectivity, as well as favourable intestinal absorption, requires a balance of lipophilic and hydrogen bonding groups. Lipinski's ‘rules of five’ (Ro5) was instrumental in guiding the medicinal chemist towards an improved balance of properties within a drug molecule.⁴ The Ro5 suggests poor oral absorption is more likely when molecular weight (M_w) is >500 Da, the hydrogen bond (H-bond) acceptor count (N/O count) is >10, the H-bond donor count (NH/OH) is >5, and the clogP is >5.⁴ The hypothesis follows that not failing more than one of these guidelines should lead to reduced pharmacokinetics related attrition in early development. In fact, recent data suggest that candidate attrition due to poor pharmacokinetics and/or bioavailability has reduced significantly in the post Ro5 era, with lack of efficacy and safety/toxicology emerging as the major cause of clinical attrition.⁵

However, recent examination of the patent literature, as well as analyses of property trends of oral drugs, have shown that Ro5 properties are inflating with time.^6–10 It has been suggested that changes in Ro5 properties over time are largely driven by the changing makeup of target portfolios within pharmaceutical companies.^6,11 Historically, aminergic G-protein coupled receptors (GPCRs) have been a relatively rich source of oral drugs due to the ligand physicochemistry being consistent with the Ro5.¹² There has, however, been a reduced investment in this gene family in recent years which could, in part, explain the lack of success over the past decade. The pharmaceutical industry has now expanded its interest into gene families such as peptidic GPCRs, protein kinases and proteases.^6,11 Indeed other target classes, such as inhibition of protein–protein interactions, are requiring molecules which are often at the edge of, or beyond, Ro5 space.¹³ It appears much more difficult to design inhibitors with the potency and selectivity alongside acceptable Ro5 properties against these targets. This is evidenced by Vieth's analysis which suggests that drugs that reside in the peptidic GPCR and proteases gene families have the lowest Ro5 compliance.¹¹ This is also supported by the global pharmacophore mapping work of Paolini and co-workers which showed that literature compounds with <100 nM potency against peptide GPCRs, phosphodiesterases, protein kinases and serine proteases all showed relatively higher numbers of hydrogen bond donors and acceptors.¹⁴ Successfully designing compounds within this chemical space that make it to and survive human trials will rely on an improved understanding of this physicochemical space and how it can be managed.

Designing a compound with hydrogen bond donors and acceptors, such as those described above, creates the opportunity to form an intramolecular hydrogen bond (IHB). A relatively stable temporary ring system relies on the donor and acceptor being able to form a conformation in which the two can adopt a suitable geometry and distance. As has been pointed out in the literature previously, this conformation can shield polarity relative to the open conformation and therefore the IHB stabilised conformation will be relatively lipophilic and may permeate membranes more freely (Fig. 1).^15–18


	Fig. 1 Drug compound containing a hydrogen bond acceptor (X) and donor (YH) that is capable of forming an IHB to facilitate membrane permeability.

An elegant analysis of the Cambridge structural database (CSD) and the protein data bank (PDB) by Kuhn and co-workers suggested a relatively high propensity for certain atom topologies to form an IHB in 5 to 8 membered rings.¹⁵ This work also included the investigation of physicochemical properties of closely matched pairs of low M_w compounds capable of forming IHBs, or not, but it was concluded that “designed model systems did not reveal a uniform behavior with respect to membrane permeability, solubility, and lipophilicity.”

Evidence for increased membrane permeability through IHB ring systems is supported by higher M_wcyclic compounds such as cyclosporine A (CsA)¹⁹ and other non-natural cyclic peptides.¹⁷ CsA is reported to exhibit oral bioavailability of about 30% which is remarkable given its significant deviation from Ro5 space.²⁰ It is now known that CsA exhibits very different conformations depending on the nature of media in which it is contained.^19,21 We hypothesise that a conformation of CsA in non-polar media, such as deuterated chloroform or carbon tetrachloride, closely represents a conformation that permeates through the lipid interior of the membrane, as these solvents have a similar dielectric constant to that estimated to exist in the interior of a phospholipid bilayer.^17,22

Forming an IHB in a non-cyclic compound which has led to improved oral absorption has been reported in a few cases. For example, an antagonist of the human luteinizing hormone releasing receptor, a peptidic GPCR, was reported to have improved oral absorption in monkeys due to the presence of an IHB.²³ Similarly for antagonists of the endothelin receptor A, several compounds were predicted, by molecular modelling, to exhibit IHBs which was used to rationalise the unexpectedly higher oral bioavailability in rats.¹⁶ The formation of an IHB can not only potentially increase the permeability of molecules across the intestinal membrane but also across the blood brain barrier. A noteworthy example includes compounds designed as antagonists against NK1.¹⁸ These researchers utilised a small molecule X-ray structure, alongside molecular modelling to design molecules that could potentially form a stable IHB. This was observed by variable temperature NMR and led to improved in vivo efficacy which relied on central nervous system penetration.

These examples illustrate the potential general utility of IHBs in the design of membrane permeable small molecules that would otherwise be more polar and less penetrant. This paper aims to expand on these few examples and provide a rational way to design, predict and demonstrate the importance of IHBs to improving membrane permeability. Particular attention will be applied to small molecules that reside at the edge or ‘beyond Ro5’ (BRo5) property space that are typically required for some of the more challenging gene families.

Results and discussion

As mentioned above, several recent datasets suggest that Ro5 properties of oral drugs are inflating with time and so we chose to construct a dataset of approved (FDA-New Chemical Entities) oral drugs since the year 2000 (2000–2010 inclusive), and use this to look at very recent trends in Ro5 properties, and other physicochemical descriptors considered important for oral absorption and permeability. This dataset overlaps with that of Vieth¹¹ and is not intended to be a comprehensive survey of recent trends in Ro5 properties, but rather as an up to date snapshot. During this period 109 oral drugs were approved with a mean M_w of 385 Da, and interestingly with 17% having M_w > 500 Da (Table 1). This is an increase on a mean M_w of 333 Da observed in the earlier but larger dataset of Proudfoot,⁹ but which contained only 7% of drugs with M_w > 500. This increase in average M_w, and in particular the proportion of drugs with M_w > 500 Da, suggests that the impact of higher M_w on oral bioavailability may not be as pronounced as the Ro5 suggests. However, other properties including the mean clogP (2.4), mean H-bond donors (2.0), and to a lesser extent mean N/O count (6.2), are remarkably consistent with analyses on older datasets^4,7–9 suggesting a close correlation with clinical success and reaffirming their importance in the Ro5. It is also interesting to note that the average topological polar surface area (TPSA) and clogP of this dataset fall within the recommendations set out by Hughes and co-workers²⁴ on physicochemical properties associated with toxicological outcomes.

Table 1 Comparison of drug datasets

Dataset	Mean M_w/Da	% >500 Da	Mean clogP	% >5	Mean N/O count	% >10	Mean NH/OH count	% >5	Mean TPSA/Å²
a See ref. 9. b See ref. 8. c See ref. 7.
Current set	385	17	2.4	13	6.2	9	2.0	1	85
Proudfoot^a	333	7	2.5	8.5	5.1	4.8	1.5	1.1
Vieth^b	344		2.3		5.5		1.8		78
Wenlock^c	337		2.5		4.9		2.1

Focussing more closely on the higher M_w compounds within this dataset (M_w > 500 Da), most (15/19) formally fail the Ro5 on 2 or more counts (see ESI†). Alternative physicochemical descriptors such as TPSA and rotatable bond count (nROT) are also higher for this subset of drug compounds (Fig. 2). It is widely accepted that limiting TPSA²⁵ to below 140 Å² and the nROTs to less than 10²⁶ will increase the likelihood of good oral absorption. The current dataset suggests that the vast majority of drugs with M_w < 500 Da do indeed meet these criteria. However, about half of the higher M_w compounds (M_w > 500 Da) also have TPSA > 140 Å². It is also clear that many of the same compounds within this subset also have >10 nROTs. It is these intriguing BRo5 outliers that are of particular interest. What enables these molecules to be orally absorbed and can we learn from them?


	Fig. 2 Topological polar surface area of drugs approved by the FDA (2000–2010) according to M_w. Coloured by nROTs (green: ≤5, yellow: 5 < x ≤ 10, red: >10).

We hypothesise that as molecules increase in heavy atom count, and become more flexible through increased nROTs then greater opportunity exists for the formation of IHB stabilised conformations. These IHB stabilised conformations are potentially less polar, and therefore more lipid soluble, due to the fact that an hydrogen bond donor and hydrogen bond acceptor are no longer available to form hydrogen bonds to water, since they are interacting with each other in the IHB. If the IHB stabilised conformation is able to exist on the timescale of intestinal membrane permeation then this apparent higher lipophilicity may enable absorption of the compound, but would retain the aqueous solubility when in the open conformation due to the presence of hydrogen bonding motifs that could interact with water. Assuming this is an important factor in the absorption of some of the BRo5 drugs, it is of course not predicted by the TPSA calculation based on atom type contributions. Therefore, an opportunity exists to develop methods which take into account the three dimensional structure of compounds to prospectively design for relatively stable IHB conformations to enable membrane permeation and intestinal absorption of BRo5 compounds.

As mentioned above, CsA (Table 2) is a well studied molecule with regard to its ability to form different conformations depending on its environment, and has been suggested to harbour some of the properties described above. A single crystal X-ray structure of CsA derived from carbon tetrachloride illustrates, that in the solid state, the methylated nitrogens of CsA point out towards solvent and even more interestingly that 4 IHBs are formed from internalisation of the exchangeable protons and several carbonyl groups (Fig. 3).

Table 2 Calculated physicochemical properties

Drug	M _w/Da	clogP	N/O count	NH/OH count	TPSA/Å²
CsA	1202	14.4	23	5	279
Atazanavir	705	5.9	13	5	171
Aliskiren	552	3.5	9	6	146


	Fig. 3 X-Ray crystal structure of cyclosporine A from carbon tetrachloride with IHB indicated (CSD DEKSAN).

Several single crystal X-ray structures of CsA, this time derived from water, have been reported that generally demonstrate more open conformations where the exchangeable protons are pointing more towards the solvent and hydrogen bonding with it, rather than forming IHBs. Moreover, some of the N-methylated amides are now pointing away from the polar solvent to the interior of the cyclic structure (Fig. 4). The X-ray derived structures are of course of interest but ultimately the adopted conformation will be influenced by molecular stacking and intermolecular interactions, and hence may not be a true representation of the solution conformations. NMR derived solvent structures of CsA in deuterated chloroform have been reported and do support the evidence for IHB formation provided for by the carbon tetrachloride derived X-ray structures.²¹ Key to these experiments is the choice of solvent as the aim is to understand the conformational preference for a molecule in the interior of a membrane. In accordance with the work of Kessler,²¹ we chose to conduct NMR temperature dependency experiments in lipophilic solvents such as deuterated chloroform, or deuterated toluene, as these are believed to have a similar dielectric (chloroformε_r = ∼4.8, tolueneε_r = ∼2.3) to that of the centre of the phospholipid bilayer.^17,22


	Fig. 4 X-Ray crystal structure of cyclosporine A from water with IHB indicated (CSD GABXEN).

Our NMR temperature dependency experiments on CsA in deuterated chloroform clearly indicated a preference for a single conformation on the NMR timescale (see ESI†). Moreover, evidence suggests this relatively rigid structure is stabilised by three strong IHBs and one weak IHB. These results are similar to those observed by Kessler²¹ and suggest that CsA does internalise its polarity in a low dielectric medium, and therefore has the potential to do so in a phospholipid bilayer. With evidence for 4 IHBs formed in a biologically relevant solvent we can now revise the Ro5 count for CsA from an NH/OH count of 5 to 1, and an N/O count from 23 to 19. Even with a revised polar atom count CsA fails the Ro5, but clearly IHB formation and hitherto internalisation of polarity improves its physicochemical properties in low dielectric media. The dependence of clogP and TPSA on conformation is likely to be important also, particularly for larger more flexible molecules. Therefore, IHB facilitated membrane diffusion could help explain the remarkable oral bioavailability, and absorption, observed in rats and humans for CsA (Table 3).

Table 3 ADME and pharmacokinetics

Drug	Papp AB^a/×10⁻⁶ cm s⁻¹	F rat (%)	Fabs rat (%)	F human (%)
a Determined in MDCK cell line. b Analysis failed.
Cyclosporine	NA^b	24	26	30
Atazanavir	35	10	54	60
Aliskiren	1	0.1	0.5	3

Demonstration of IHB formation by NMR temperature dependency is clearly a useful technique to provide evidence for this phenomenon in solution but of course it relies upon the synthesis of the compound to carry out the experiment. What would be more useful is a predictive method that could be used in design, in a prospective sense. With this goal in mind we began looking at in silico modelling of low energy conformations calculated in the gas phase and then related this to our NMR experiments (see ESI†). As the conformational search calculations are carried out in the gas phase and hence at low dielectric (ε_r = 1), we reasoned that conformations identified would be a good approximation of a conformation that exists in solution in our NMR experiments, as well as conformations adopted in the membrane, due to similarities in the dielectric constant.

The lowest energy conformation calculated in the gas phase for CsA suggests the cyclic structure can form up to four IHBs (Fig. 5). It is also interesting to note the N-methylated amides are predicted to point to solvent in this conformation while the exchangeable nitrogen and oxygen protons are predicted to reside internally, facilitated by IHB formation. These predictions are in line with the observed X-ray crystal structure from carbon tetrachloride, as well as the NMR experiments described above. The gross similarity between predicted and observed IHB formation and conformational preference highlights the potential for this computational method to be used in a predictive sense.


	Fig. 5 Lowest energy conformation of cyclosporine A calculated in the gas phase with IHB indicated.

Returning to the dataset of recently approved oral drugs described above, we hypothesised that some of the non-cyclic BRo5 drugs within the higher M_w subset (M_w > 500 Da) may in fact be enabled by IHB facilitated membrane diffusion, in a similar manner to CsA. One of the most intriguing oral drugs within this set is the protease inhibitor atazanavir. This compound fails the Ro5 on three counts, M_w, clogP, and N/O count, and has a significant number of hydrogen bond donors (Table 2). Other physicochemical descriptors such as TPSA and nROT (16) are clearly also high, and therefore it is remarkable that this compound has such a high reported oral bioavailability of 60% in humans.²⁷

To understand the potential for IHB formation we first submitted atazanavir to a conformational search in the gas phase. The lowest energy conformation predicts the formation of three IHBs, with two amide carbonyls involved in H-bonding to two amide NHs, as well as the secondary alcohol forming an IHB with an amide carbonyl (Fig. 6). We next carried out NMR temperature dependency experiments in deuterated toluene to investigate how well the gas phase model predicts the observed solution conformation (see ESI†). The NMR data suggest the formation of one conformer (and one rotamer) as well as two strong and two weak IHBs on the timescale of the NMR. If we now revise the Ro5 count for this compound based on the formation of 4 IHBs, then the compound is much closer now to acceptable physicochemical properties for membrane permeation and oral absorption (N/O count 9, NH/OH count of 1).


	Fig. 6 Lowest energy conformation of atazanavir calculated in the gas phase with IHB indicated.

Again, the reported clogP and TPSA are likely to be conformation dependent and should be used with caution when considering large and flexible molecules, such as atazanavir. Of course molecular volume, for which M_w is a surrogate, is also dependent on conformation and it is tempting to suggest that the M_w of atazanavir is a misleading parameter when the NMR and gas phase modelling data suggest the compound exists in a compact conformation, with relatively low molecular volume given its M_w.

In line with the above thoughts and observations, atazanavir has high intrinsic cell permeability in an MDCK cell line suggesting the apparent polarity of the compound is lower than would be expected given its calculated physicochemistry. Rat pharmacokinetics also suggests the compound is well absorbed, but has low bioavailability due to high first pass clearance (Table 3). These data support the notion that atazanavir does indeed form an IHB stabilised conformation that hides its polarity and enables passive membrane permeation and therefore intestinal absorption to occur in rats and in human.

Aliskiren is another peptide mimetic BRo5 drug within the higher M_w (M_w > 500 Da) group described above. The compound breaks the Ro5 due to its combined high M_w and H-bond donors (Table 2), but unlike atazanavir, aliskiren is poorly absorbed in humans with a very low reported oral bioavailability (F ≈ 3%).²⁸ We were interested to investigate whether or not the modelled lowest energy conformation calculated in the gas phase would predict for IHB formation. Interestingly, the lowest energy conformation predicts for two IHBs, leaving four exposed exchangeable protons (Fig. 7). Clearly, four exposed hydrogen bond donors would be considered deleterious for good membrane permeation, and indeed the compound suffers from poor intrinsic cell permeability, as observed in the MDCK cell line, and also poor absorption in rat and human (Table 3). Importantly, this observation demonstrates that the gas phase conformational searching does not always find a conformation that is capable of IHB formation and therefore compounds with poor membrane permeability could be predicted for using this technique. Efforts to determine the number of IHBs by NMR temperature dependency experiments were hampered by the low solubility of aliskiren in deuterated chloroform, and deuterated toluene, and therefore we were unable to obtain useful data.


	Fig. 7 Lowest energy conformation of aliskiren calculated in the gas phase with IHB indicated.

The gas phase calculations were applied to the remaining drug compounds from the higher M_w (M_w > 500 Da) group described above (see ESI†). When IHBs were taken into account for each lowest energy conformation, and the NH/OH count revised on this basis, the NH/OH count was reduced to two or fewer.

Conclusions

BRo5 chemical space is becoming increasingly important to access due to the physicochemical requirements of the shifting targets and modalities being explored in the pharmaceutical industry. The retrospective analysis within this paper suggests that BRo5 drugs are increasingly being discovered but perhaps without a clear understanding of how oral absorption is being achieved in this chemical space. Structural evidence on the few examples cited in this paper suggests that IHB formation, in biologically relevant solvents, can improve intrinsic membrane permeability and in turn, intestinal absorption of cyclic as well as non-cyclic BRo5 compounds. Moreover, propensity to form IHBs can be modelled by determining low energy conformations in the gas phase, thereby providing a tool to prospectively design and predict for improved membrane permeability, and absorption, in novel compounds.

Notes and references

C. Laurence and M. Berthelot, Perspect. Drug Discovery Des., 2000, 18, 39 Search PubMed.
(a) E. Freire, Drug Discovery Today, 2008, 13, 869 Search PubMed; (b) J. E. Ladbury, G. Klebe and E. Freire, Nat. Rev. Drug Discovery, 2010, 9, 23 CrossRef CAS.
O. A. Raevsky and K.-J. Schaper, Eur. J. Med. Chem., 1998, 33, 799 CrossRef CAS.
C. A. Lipinski, F. Lombardo, B. W. Dominy and P. J. Feeney, Adv. Drug Delivery Rev., 1997, 23, 3 CrossRef.
I. Kola and J. Landis, Nat. Rev. Drug Discovery, 2004, 3, 711 CrossRef CAS.
P. D. Leeson and B. Springthorpe, Nat. Rev. Drug Discovery, 2007, 6, 881 CrossRef CAS.
M. C. Wenlock, R. P. Austin, P. Barton, A. M. Davis and P. D. Leeson, J. Med. Chem., 2003, 46, 1250 CrossRef CAS.
M. Vieth, M. G. Siegel, R. E. Higgs, I. A. Watson, D. H. Robertson, K. A. Savin, G. L. Durst and P. A. Hipskind, J. Med. Chem., 2004, 47, 224 CrossRef CAS.
J. R. Proudfoot, Bioorg. Med. Chem. Lett., 2005, 15, 1087 CrossRef CAS.
P. D. Leeson and A. M. Davis, J. Med. Chem., 2004, 47, 6338 CrossRef CAS.
M. Vieth and J. J. Sutherland, J. Med. Chem., 2006, 49, 3451 CrossRef CAS.
K. Beaumont, E. Schmid and D. A. Smith, Bioorg. Med. Chem. Lett., 2005, 15, 3658 Search PubMed.
(a) M. R. Arkin and J. A. Wells, Nat. Rev. Drug Discovery, 2004, 3, 301 CrossRef CAS; (b) P. Chene, ChemMedChem, 2006, 1, 400 CrossRef CAS.
G. V. Paolini, R. H. B. Shapland, W. P. van Hoorn, J. S. Mason and A. L. Hopkins, Nat. Biotechnol., 2006, 24, 805 CrossRef CAS.
B. Kuhn, P. Mohr and M. Stahl, J. Med. Chem., 2010, 53, 2601 CrossRef CAS.
C. Wu, E. R. Decker, N. Blok, J. Li, A. R. Bourgoyne, H. Bui, K. M. Keller, V. Knowles, W. Li, F. D. Stavros, G. W. Holland, T. A. Brock and R. A. F. Dixon, J. Med. Chem., 2001, 44, 1211 Search PubMed.
T. Rezai, B. Yu, G. L. Millhauser, M. P. Jacobson and R. S. Lokey, J. Am. Chem. Soc., 2006, 128, 2510 CrossRef CAS.
V. A. Ashwood, M. J. Field, D. C. Horwell, C. Julien-Larose, R. A. Lewthwaite, S. McCleary, M. C. Pritchard, J. Raphy and L. Singh, J. Med. Chem., 2001, 44, 2276 Search PubMed.
N. El Tayar, A. E. Mark, P. Vallat, R. M. Brunne, B. Testa and W. F. van Gunsteren, J. Med. Chem., 1993, 36, 3757 Search PubMed.
Absolute human oral bioavailability value taken from the Sandimmune® drug label http://www.fda.gov/ .
(a) H.-R. Loosli, H. Kessler, H. Oschkinatb, H.-P. Weber, T. J. Petchera and A. Widmera, Helv. Chim. Acta, 1985, 68, 682 CAS; (b) H. Kessler, M. Kock, T. Wein and M. Gehrke, Helv. Chim. Acta, 1990, 73, 1818 CAS.
(a) R. B. M. Koehorst, R. B. Spruijt, F. J. Vergeldt and M. A. Hemminga, Biophys. J., 2004, 87, 1445 Search PubMed; (b) U. V. Mardolcar, C. A. Nieto de Castro and F. J. V. Santos, Fluid Phase Equilib., 1992, 79, 255 Search PubMed.
S. Sasaki, N. Cho, Y. Nara, M. Harada, S. Endo, N. Suzuki, S. Furuya and M. Fujino, J. Med. Chem., 2003, 46, 113 Search PubMed.
J. D. Hughes, J. Blagg, D. A. Price, S. Bailey, G. A. DeCrescenzo, R. V. Devraj, E. Ellsworth, Y. M. Fobian, M. E. Gibbs, R. W. Gilles, N. Greene, E. Huang, T. Krieger-Burke, J. Loesel, T. Wager, L. Whiteley and Y. Zhang, Bioorg. Med. Chem. Lett., 2008, 18, 4872 CrossRef CAS.
(a) K. Palm, P. Stenberg, K. Luthman and P. Artursson, Pharm. Res., 1997, 14, 568 Search PubMed; (b) P. Ertl, B. Rohde and P. Selzer, J. Med. Chem., 2000, 43, 3714 CrossRef CAS.
D. F. Veber, S. R. Johnson, H.-Y. Cheng, B. R. Smith, K. W. Ward and K. D. Kopple, J. Med. Chem., 2002, 45, 2615 CrossRef CAS.
P. J. Piliero, Drugs Today, 2004, 40, 901 Search PubMed.
Absolute human oral bioavailability value taken from the Tekturna® drug label http://www.fda.gov/ .

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c1md00093d

Click here to see how this site uses Cookies. View our privacy policy here.