Methods and Findings in Experimental and Clinical Pharmacology
Vol. 24, Suppl. A, 2002, pp. 9-11
ISSN 0379-0355
Copyright 2002 Prous Science, S.A.
CCC: 0379-0355/2002
http://www.prous.com

Drug Optimization by Molecular Manipulation

M.C. Avendaño López

Department of Organic Chemistry and Pharmaceutics, Universidad Complutense de Madrid, Spain

Most biochemical regulation processes are the consequence of a molecular recognition between biomolecules. Traditional drugs are small molecules that interact with these molecular targets. However, given their great variety and, at the same time, their great similarity, it is difficult to find selective, effective compounds. Biology, in its broadest sense, must identify the problem to be solved at a molecular level and chemistry must be used as a tool that is able to add the missing piece to the system or prevent a series of events from happening. This piece is the drug which, once it has been created, must be tested to determine whether it meets the goals it has been designed for.

In this interaction, chemistry should not stop being chemistry. The problem is that we do not sufficiently understand the chemistry of life, and transferring knowledge from biology to new drug design is a far from easy task (1). In fact, in spite of the continual, almost overwhelming supply of biological knowledge, most of the organic molecules developed by the pharmaceutical industry are obtained using procedures that are not very different from those used years ago.

We need to think about drugs and their therapeutic activity from a chemical viewpoint and try to understand why two drugs that seem to be very similar and can be used to treat the same problem, such as clozapine, an atypical neuroleptic introduced by Sandoz in 1963, and olanzapine (Ryprexa^®), marketed much later by Lilly, can be distributed differently within the body, have different potencies and side effects, or are excreted at different rates. These differences in activity and behavior are caused exclusively by differences in their chemical structure, no matter how little these differences may be.

A necessary part of rational drug design is the manipulation of a "seed" in order to optimize its activity and decrease its toxicity, which often leads to an increased selectivity of action. This procedure can be dated back to the discovery of arsphenamine (Salvarsan^®) in 1910 by P. Ehrlick and S. Hata, after synthesizing about 600 analogues of atoxyl--another, much more toxic arsenic compound.

When planning structures for study, it is common to use computational molecular modeling methods that optimize their interactions with the pharmacological target, if the target's three-dimensional structure is known. When it is not known, common structural circumstances can be found in different ligands even though they may not be immediately recognizable.

The manipulation of a "seed" consists of substituting certain groups with others that have similar properties (bioisosters), either simplifying the structure (particularly in the case of naturally-occurring compounds having complex structures) in order to find the "pharmacophore group" (2) or restricting conformational freedom, although this often leads to obtaining compounds that interact with another pharmacological target. This is the case of cromakalim analogue antihypertensive drugs, which act on potassium channels and were obtained from b-blockers with an "aryloxypropanolamine" structure.

The pharmacokinetics can also be optimized by structural manipulation. This was how the quaternary ammonium salts derived from atropine or scopolamine, such as thiotropium bromide, were obtained. Acting as non-selective cholinergic antagonists, they are useful as inhaled bronchodilators because their polarity prevents them from crossing the blood-brain barrier (3).

Alternatively, prodrugs can be designed. An example of this is capecitabine, recently marketed by Roche under the name Xeloda^®. It is an orally active prodrug of
5-fluorouracil (5-FU), a classic antineoplastic drug used since 1957 whose main limitations, in addition to its lack of selectivity of action, are its low oral absorption and a high interindividual variability in its blood levels. The levels can even vary within the same subject when given by continuous infusion. Capecitabine, on the other hand, can be given orally and shows a fairly selective bioactivity in tumor tissue (4).

In the development by Bristol-Myers Squibb of antihypertensive drugs that act by inhibiting the angiotensin-converting enzyme (ACE), a model of the enzyme based on another Zn(II) metalloprotease was used (5; see 6 for a review of mechanism-based Zn(II) protease inhibitors), with a peptide contained in snake venom, teprotide, as the "seed". Manipulation of succinylproline as the "seed" led to the development of captopril (Capoten^®), used since 1979 in the treatment of arterial hypertension and congestive heart failure (7, 8).

The adverse reactions observed in this drug were similar to those caused by penicillamine. Consequently, it was speculated that these reactions were caused by the mercapto group common to the two compounds (4), and, although these reactions could be avoided by giving low doses (5, 6), this led MDS's scientists to look for ECA inhibitors with a Zn(II) ligand carboxyl group instead of the SH group. This led to enalapril and its analogues (7, 8), while the conformational restriction in enalapril led to benazepril and cilazapril, among others. Furthermore, the discovery that phosphoramidone, a phosphoramidate isolated from Streptomyces tanashiensis that inhibits the Zn(II) enzyme thermolysin, led to ACEIs with a phosphinic (fosinopril) or phosphonic (ceranapril) structure.

The ability of the mercapto group to take up free radicals has added therapeutic value to other captopril analogues which, being more lipophilic, attain high levels in heart tissue and can be used as cardioprotective drugs. This is the case with zofenopril (Zofrenil^®, Bifril^®), a prodrug with a thioester structure that activates to give zofenoprilat, a free mercapto group able to coordinate Zn(II) and act as a free radical scavenger (9).

A similar procedure has been followed in the design of HIV-1 protease inhibitors used in AIDS therapy because these enzymes' activity is needed for viral multiplication. The design of saquinavir (Inverasa^®) by researchers at Hoffman La Roche started with structures that mimicked the transitional status corresponding to the hydrolysis of these enzymes' substrate peptides, using a proline residue as P1' residue (10). Starting with a "seed" with an IC₅₀ = 750 nM, saquinavir was obtained with an IC₅₀ = 0.3 nM, although its high potency in vitro is limited by its low bioavailability in vivo, probably due to its high molecular weight and the number of amide bonds it contains.

Nelfinavir (Viracept^®) was the result of a collaboration between two pharmaceutical companies, Lilly and Agouron, taking saquinavir as their model and introducing two changes: substitution of the enzymatically labile P3-P2 residues with another group and modifying the P1 residue by inserting a sulfur atom to increase the lipophilia (11, 12).

For its part, Merck used as "seed" compounds that were active as renin inhibitors, another aspartic acid protease that catalyses conversion of angiotensinogen into angiotensin I, for the development of indinavir (Crixivan^®). Starting with a 7-residue peptidomimetic with a hydroxyethylene group isoster of the transition state of the hydrolysis reaction catalyzing the enzyme, like the two inhibitors mentioned previously, a compound was obtained that was optimized by modeling together with saquinavir, from which a structural portion was used, to finally give indinavir (13).

Lastly, we will briefly discuss the development of omeprazole (Losec^®), the top-selling antiulcer drug, and its successor esomeprazole (Nexium^®), at AstraZeneca. These drugs were introduced in 1988 and 2000, respectively. The former was obtained by manipulating a "seed" whose mechanism of action is still unknown:
2-pyridylthioacetamide (the compound CMN 131 discovered by Servier) (14).

After the failures suffered since the company began the "antacid" project in 1967 with lidocaine analogues, based on the hypothesis that these analogues inhibited gastrin release, the search for an effective drug from CMN 131 took two separate directions: the synthesis and study of heterocycles with a sulfur atom and the synthesis of sulfides or sulfoxides with an imidazole ring. The latter led to timoprazole, which, although toxic, showed that its antacid activity could be due to a new mechanism, as it blocked the response to gastrin and cAMP.

Its more active analogue picoprazole showed that the effect was due to the inhibition of the recently discovered H⁺K⁺-ATPase, but this drug was rejected because it caused vasculitis in the in vivo model, a conscious dog with a gastric fistula. By the time it was discovered in 1979 that vasculitis was a hereditary problem of the animal used, many analogues had been synthesized, including omeprazole, which was even better.

Omeprazole was registered in Sweden in 1988 and the search began for a possible replacement. Hundreds of analogues were synthesized but only one finally proved to be better than omeprazole: its S enantiomer, esomeprazole (Nexium^®). The increased activity compared with the R enantiomer in humans (the opposite happens in rats) is due to its different stability in the presence of the cytochrome P450 isoenzyme CYP2C19 (15).

In conclusion, the examples chosen seek to show that structure-based "rational drug design" is the outcome of an interdisciplinary action and may be influenced by unforeseeable circumstances.

REFERENCES

1. Stevenson, R. Strategic thinking. Chem Br 2002, 38: 30-1.

2. Avendaño, C., Menéndez, J.C. Inhibitors of multidrug resistance to antitumor agents (MDR). Curr Med Chem 2002, 9(2): 159-93.

3. Doods, H.N. Selective muscarinic antagonists as bronchodilators. Drug News Perspect 1992, 5: 345-52.

4. Budman, D.R. Capecitabine. Invest New Drugs 2000, 18: 355-63.

5. Cushmann, D.W., Cheung, H.S., Sabo, E.F., Ondetti, M.A. Design of potent competitive inhibitors of angiotensin-converting enzyme. Carboxyalkanoyl and mercaptoalkanoyl amino acids. Biochemistry 1977, 16: 5484-91.

6. Kim, D.H., Mobashery, S. Mechanism-based inhibition of zinc proteases. Curr Med Chem 2002, 8(8): 959-65.

7. Ondetti, M.A., Cushman, D.W., Rubin, B. Captopril. In: Chronicles of Drug Discovery, Vol. 2. Bindra, J.S., Lednicer, D. (Eds.). Wiley & Sons: 1983, 3-31.

8. Maxwell, R.A., Eckhardt, S.B. Captopril. In: Drug Discovery. A CaseBook and Analysis. Clifton, N.J. (Ed.). Humana Press: 1990, 19-34.

9. Heel, R.C., Brodgen, T.M., Speight, T.M., Avery, G.S. Captopril: A preliminary review of its pharmacological properties and therapeutic efficacy. Drugs 1980, 20: 409-52.

10. Cleland, J.G., Dargie, G.P., Hodsman, S.G. et al. Captopril in heart failure. A double blind controlled trial. Br Heart J 1984, 52(5):
530-5.

11. Patchett, A.A., Harris, E., Tristram, E.W. et al. A new class of angiotensin-converting enzyme inhibitors. Nature 1980, 288(5788): 280-3.

12. Patchett, A.A. The design of enalapril. In: Hypertension and the Angiotensin System: Therapeutic Approaches. Doyle, A.E., Bearn, A.G., (Eds.). Raven Press 1983, 155-66.

13. Subissi, A., Evangelista, S., Giachetti, A. Cardiovasc Drug Rev 1999, 17: 115-33.

14. Roberts, N.A., Martin, J.A., Kinchington, D. et al. Rational design of peptide-based HIV proteinase inhibitors. Science 1990, 248(4953): 358-61.

15. Kalish, V. et al. Eur J Med Chem 1995, 30: S201-14.

Methods and Findings in Experimental and Clinical Pharmacology Vol. 24, Suppl. A, 2002, pp. 9-11
ISSN 0379-0355 Copyright 2002 Prous Science, S.A. CCC: 0379-0355/2002 http://www.prous.com