HO OH Immucillin H
HO OH Immucillin H
FIGURE 11.6 The structures of PNP reaction substrate inosine, transition state, and reaction product (top panel) and transition state-based inhibitor Immucillin H (bottom panel).
Immucillin H was a 56 pM inhibitor of human PNP with good potency against cultured human T-cell lines in the presence of deoxyguanosine. Currently, Immucillin H is in phase II clinical trials for the treatment of leukemia.
In the preceding section of this chapter we established the fundamental importance to drug discovery of a deep, mechanistic understanding of the reaction mechanism of an enzyme target. While this can be accomplished by the application of mechanistic enzymology, it can be facilitated greatly by the knowledge of the three-dimensional structure of the protein, obtained via biostructure-based technologies such as computational biochemistry, NMR spectroscopy, and x-ray crystallography. Visualization of the detailed architecture of an enzyme's active site, in complex with a small-molecule inhibitor, can be an important driver in the optimization of a medicinal chemistry effort. The structural insights thus obtained allow for improvements in target potency, selectivity, and inhibitor physicochemical properties, all of which are paramount in establishing inhibitor SAR.
The term "rational drug design" is often used to describe the application of structure-guided drug discovery approaches. Over the past two decades, several drugs have been made available to patients as a result of advances in protein crystallography and other structural methods. For example, more than 40 compounds have entered clinical trials whose discovery was reliant upon a biostructure-based approach, and as of 2007, at least 10 of these have been approved by regulatory agencies. In this section, we exemplify how a detailed understanding of the topography of a ligand-enzyme complex can provide a basis for the design of better inhibitors and can complement enzymological studies to rationalize their biochemical mode of action.
11.5.1 Structure-Based Design of Protein Kinase Inhibitors
Owing to their central roles in mediating cellular signaling pathways, protein kinases are increasingly important targets for treating a number of diseases. In particular, many of the over 500 kinases encoded by the human genome function to regulate tumor cell proliferation, migration, and survival, rendering them attractive targets for chemotherapeutic intervention in the treatment of cancer. Despite their diversity, all protein kinases catalyze the transfer of the y-phosphate of ATP to the hydroxyl group of serine, threonine, or tyrosine residues on specific proteins. Their catalytic domains reflect this singular function in that they share a common feature called the protein kinase fold, which includes a highly conserved ATP-binding pocket, flanked by N-terminal and C-terminal lobes. The ATP-binding site has been the major focus of inhibitor design; owing to its high degree of conservation, however, selectivity has been a major challenge for inhibitors that target this binding site of protein kinases. The use of biostructure-based approaches has therefore been of great importance in the optimization of targeted anticancer therapies.
X-ray crystallographic studies have indicated that the catalytic activity in most kinases is controlled by an "activation loop," which adopts different conformations depending upon the phosphorylation state of serine, threonine, or tyrosine residues within the loop. In kinases that are fully active, the loop is thought to be stabilized in an open conformation as a result of phosphorylation, allowing a P-strand within the loop to serve as a platform for substrate binding. While the "active" conformation of the loop is very similar in all known structures of activated kinases, there is great variability in the loop conformation in the inactive state of kinases. In this inactive-like conformation, the loop places steric constraints, which preclude substrate binding.
One of the first protein kinase inhibitors developed as a targeted cancer therapy is imatinib (Gleevec®; Novartis Pharmaceuticals, Basel, Switzerland). Imatinib has been used with remarkable success to treat patients with chronic myeloid leukemia (CML), which is a malignancy resulting from the deregulated activity of Abl due to a chromosomal translocation that gives rise to the breakpoint cluster region-abelson tyrosine kinase oncogene (BCR-ABL). Imatinib inhibits the tyrosine kinase activity of Bcr-Abl and it is considered as a frontline treatment for CML by virtue of its high degree of efficacy and selectivity. Together with biochemical studies, crystallographic studies of the interaction of imatinib with the Abl kinase domain have revealed that imatinib binds to the Bcr-Abl ATP-binding site preferentially when the centrally located activation loop is not phosphorylated, thus stabilizing the protein in an inactive conformation (Figure 11.7). In addition, imatinib's interactions with the NH2-terminal lobe of the kinase appear to involve an induced-fit mechanism, further adding to the unique structural requirements for optimal inhibition. One of the most interesting aspects of this interaction is that the specificity of inhibition is achieved despite the fact that residues that contact imatinib in Abl kinase are either identical or very highly conserved in other Src-family tyrosine kinases. Thus, despite targeting the relatively well conserved nucleotide-binding pocket of Abl, studies have shown that imatinib achieves its high specificity by recognizing the distinctive inactive conformation of the Abl activation loop. Biostructure-based methods have had a further impact on more recent efforts to design second-generation therapies targeting imatinib-resistant mutations in Bcr-Abl kinase that have been identified in CML patients. It is very likely that these new inhibitors will have substantial clinical utility in the treatment of imatinib-resistant CML; continued exploration of the structural details of the interactions between these compounds and the mutant kinase are still necessary, as resistance remains an inevitable consequence of such drug treatment regimens.
The three catalytically active receptor tyrosine kinases (RTKs) of the ErbB family represent another attractive target for the treatment of a variety of cancers: epidermal growth factor receptor (EGFR, also known as ErbB1), ErbB2 (also known as HER2/neu), and ErbB4. These RTKs are large, multidomain proteins that contain an extracellular ligand-binding domain, a single transmembrane domain, and a cytoplasmic domain responsible for the tyrosine kinase activity. Ligand binding to the extracellular domain induces the formation of receptor homo- and heterodimers, which leads to activation of the tyrosine kinase activity and subsequent phosphorylation of the cytoplasmic tail.
A number of ErbB-targeted molecules have already reached the market, with a number of others in various stages of clinical investigation. Two of these molecules, erlotinib (Tarceva®, Genetech, Inc. and OSI Pharmaceuticals, Inc. San Francisco, CA) and lapatinib (Tykerb®, GlaxoSmithKline plc. Brentford, U.K.), share a common 4-anilinoquinazoline core, yet their ErbB inhibition profiles and mechanisms of action are clearly differentiated on the basis of biochemical and crystallographic studies. For example, while erlotinib is a potent and selective inhibitor of EGFR only = 0.4 nM), lapatinib exhibits potent activity against both EGFR and ErbB2, with estimated Kiapp values of 3 and 13 nM, respectively. In addition to its dual kinase activity profile, lapatinib can be distinguished further from erlotinib in that it has a prolonged off-rate from its kinase targets compared to the very fast off-rate from EGFR of erlotinib. This translates to a half-life of dissociation of 300 min for the lapatinib-EGFR complex. Importantly, in cellular washout experiments this slow off-rate correlates with a prolonged inhibition of receptor tyrosine phosphorylation in tumor cells (see Section 11.6 for additional information).
An evaluation of the binding mode of lapatinib, based on the crystal structure of the compound in complex with EGFR, suggests a rationale for its long target residence time compared to other 4-anilinoquinazoline inhibitors. Not surprisingly, the quinazoline ring was observed to be hydrogen-bonded to the flexible hinge region between the NH2 and COOH-terminal lobes of the kinase, but there are variations in the key H-bonding interactions compared to those revealed in the erlotinib-EGFR structure. These differences indicate that lapatinib binds to a relatively closed form of this binding site, whereas erlotinib binds to a more open form. In addition, the ATP-binding pocket of the lapatinib-EGFR complex has a larger back pocket than the apo-EGFR or erlotinib-EGFR structures owing to a shift in one end of the C-helix. This enlarged back pocket accommodates
the 3-fluorobenzyloxy group of lapatinib (Figure 11.8). The structural change is significant because it results in the loss of a highly conserved Glu738-Lys721 salt bridge, which is an important regulatory mechanism of kinases, functioning to ligate the phosphate groups of ATP. The net result of these structural differences is that the activation loop in the lapatinib-EGFR structure adopts a conformation that is reminiscent of that found in inactive kinases, while the erlotinib-EGFR structure displays the activation loop in an active conformation. These effects provide a potential molecular rationale for the prolonged residence time of lapatinib on its target, which in turn may result in the observed duration of drug activity in cells. In total, these elegant structural and biochemical studies have important implications for the discovery of novel, targeted signal transduction inhibitors, and suggest that subtle differences in kinase inhibitor structure can have a profound impact on the binding mode, kinetics, and cellular activity.
11.5.2 Structure-Based Design of HIV Protease Inhibitors
Perhaps the greatest impact of structure-based design on the identification of novel medicines has been in the treatment of AIDS, the etiologic agents of which are human immunodeficiency virus type 1 and type 2 (HIV-1 and HIV-2). These retroviruses encode relatively simple genomes consisting of three open reading frames (ORFs), gag, pol, and env. The gag gene encodes the structural capsid, nucleocapsid, and matrix proteins, while the env gene is processed by multiple alternative splicing events to yield regulatory proteins. The pol ORF encodes the essential viral enzymes necessary for viral replication: RT, integrase, and protease (PR). HIV-1 PR is an aspartyl protease that is required for proteolytic processing of the Gag and Gag-Pol polyprotein precursors to yield
FIGURE 11.9 Clinically approved HIV-1 protease inhibitors for the treatment of AIDS.
FIGURE 11.9 Clinically approved HIV-1 protease inhibitors for the treatment of AIDS.
the viral enzymes and structural proteins, and it is absolutely indispensable for proper virion assembly and maturation. For this reason it has been an important target for the discovery of anti-HIV therapeutics, and indeed there are at least eight drugs in current clinical use whose antiviral mode of action is by potent inhibition of the HIV protease (Figure 11.9).
One of the major driving forces behind the rapid progress in the identification of HIV protease inhibitors to combat AIDS has been the intense investigation of the structure of the enzyme, particularly in complex with a number of different inhibitors. HIV-1 PR is a dimer comprised of two polypeptide chains of 99 amino acids, each contributing a single catalytic aspartate residue within the active site that lies at the dimer interface. This active site is covered by two symmetric flaps whose dynamic motions allow entry and exit of polypeptide substrates. For each of the different substrates, three to four amino acids on each side of the scissile bond are thought to be involved in binding to the substrate cavity. Since there is little similarity in the primary sequence of the cleavage sites of each of the protease substrates, binding specificity is thought to be driven by the conservation in the secondary structure surrounding the cleavage sites. All of the inhibitors currently used to treat HIV infection are competitive in nature and bind to the protease active site.
Saquinavir was the first HIV protease inhibitor available for the treatment of AIDS, and its design was based on a strategy using a transition state mimetic. A distinguishing feature of HIV PR is its ability to cleave Tyr-Pro and Phe-Pro sequences found in the viral substrates, as mammalian endopeptidases are unable to cleave peptide bonds followed by a proline. A rational inhibitor design approach based on this property offered hopes of identifying inhibitors selective for the viral enzyme. Since reduced amides and hydroxyethylamine isosteres most readily accommodate the amino acid moiety of Tyr-Pro and Phe-Pro in the HIV substrates, they were chosen for further interrogation. Systematic substitutions were explored on a minimum peptidic pharmacophore, and one compound containing an (S,S,S)-decahydro-isoquinoline-3-carbonyl (DIQ) replacement for proline exhibited a Ki value of 0.12 nM at pH 5.5 for HIV-1 PR and <0.1 nM for HIV-2 PR. The interactions of this compound, later named saquinavir, with HIV-1 PR were studied crystallographically (Figure 11.10). The compound was shown to bind to the enzyme in an extended conformation with the carbonyl of the DIQ group binding to a water molecule that connects the inhibitor with the flap regions. These studies shed much light on the binding mode of the first HIV PR inhibitors and set the stage for further exploration of novel compounds with improved properties.
The availability of new HIV protease inhibitors represented a great triumph in the fight against AIDS, but it was only a matter of time before the selective pressure of antiretroviral therapy led to the emergence of HIV strains harboring drug-resistant mutations against protease inhibitors. One of the primary mutations first noted in protease inhibitor-resistant strains was in Val82 of HIV-1 PR. Crystallographic and modeling studies suggested that the binding of protease inhibitors like ritonavir might be compromised due to the loss of hydrophobic interactions between the isopropyl side chain of Val 82 of the enzyme and the isopropyl substituent projecting from the 2 position of the P3 thiazolyl group of ritonavir. This functionality was substituted to identify an inhibitor whose activity was less dependent on interaction with Val 82, and the optimization that was supported by modeling studies led to the identification of ABT-378, later named lopinavir. This novel inhibitor had extraordinary potency against wild-type and mutant HIV PR (Ki = 1.3-3.6 pM) in vitro, and maintained activity against ritonavir-resistant mutants of HIV-1. Lopinavir, as a combination drug
with ritonavir, is known as Kaletra, and is an important salvage drug for patients who have failed primary therapy with other protease inhibitors.
The structural details of drug-binding pocket interactions, gleaned from crystallographic and other biophysical methods, can provide a rich source of information for inhibitor optimization. Historically, SAR through structure-based methods has been limited by the time and protein demands of x-ray crystallography. These limitations, however, are rapidly diminishing due to significant advances in the technologies of protein expression and protein crystallography. In particular, robotic methods for crystallization trials have significantly reduced the time required for obtaining crystals of a target protein in complex with multiple inhibitory compounds. Likewise, the more routine use of high-energy beam sources has facilitated structure determinations from crystals that would otherwise not be sufficient for diffraction studies. These advances provide the basis for greater reliance on structural biology as a common tool during the iterative process of lead optimization.
Conformational dynamics within the drug-binding pockets of enzymes is a common feature, dictated by the chemistry of enzyme catalysis. Hence, binding pocket structures are not static; rather, they often change in response to encounters with inhibitory molecules. Thus, as described in this chapter, there can often be a temporal component to enzyme-inhibitor affinity. Advances in kinetic methodologies and instrumentation have made the determination of inhibition kinetics more facile, so that such measurements can be a routine part of the SAR of lead optimization. It is therefore no longer necessary to rely solely on equilibrium measures of inhibitor-binding affinity, such as IC50 and Ki values, for lead optimization. Instead, routine measurements of enzyme-inhibitor association and dissociation rates are becoming practical, with throughput that makes these measurements germane to drug discovery. Hence, increased attention is being paid to the importance of understanding these kinetic components of drug-target interactions, and their potential impact on clinical efficacy. For example, the duration of drug efficacy in vivo has been suggested to depend in part on the duration of the drug-target complex; this is experimentally measured as the residence time, which is the reciprocal of the dissociation rate constant for the drug-target complex. Drugs that demonstrate long residence times, especially when this exceeds the pharmacokinetic half-life of the drug, may significantly extend the pharmacodynamic efficacy of a drug in vivo, and may also ameliorate the potential for adverse events. Future drug discovery efforts may thus be focused not merely on optimization of inhibitor affinity, but also on the extension of residence time (see Further Readings).
Copeland, R. A. (2000) Enzymes: A Practical Introduction to Structure, Mechanism and Data Analysis, 2nd edn. Wiley, Hoboken, NJ.
Copeland, R. A. (2005) Evaluation of Enzyme Inhibitors in Drug Discovery: A Guide for Medicinal Chemists and Pharmacologists. Wiley, Hoboken, NJ. Copeland, R. A., Pompliano, D. L., and Meek, T. D. (2006) Drug-target residence time and its implications for lead optimization. Nat. Rev. Drug Discov., 5: 730-739. Nagar, B., Bornmann, W. G., Pellicena, P., et al. (2002) Crystal structures of the kinase domain of c-Abl in complex with the small molecule inhibitors PD173955 and imatinib (STI-571). Cancer Res. 62: 4236-4243. Roberts, N. A., Martin, J. A., Kinchington, D., et al. (1990) Rational design of peptide-based HIV proteinase inhibitors. Science 248: 358-361. Robertson, J. G. (2005) Mechanistic basis of enzyme-targeted drugs. Biochemistry 44, 5561-5571. Schramm, V. L. (2005) Enzymatic transition states: Thermodynamics, dynamics and analogue design. Arch.
Biochem. Biophys. 433, 13-26. Wood, E. R., Truesdale, A. T., McDonald, O. B., et al. (2004) A unique structure for epidermal growth factor receptor bound to GW572016 (Lapatinib): Relationships among protein conformation, inhibitor off-rate, and receptor activity in tumor cells. Cancer Res., 64: 6652-6659.
Was this article helpful?