Patent Application
20080305041
The present invention generally relates to compositions and methods for modulating PF4 activity and, more specifically, to compositions and methods for modulating such PF4-mediated processes as angiogenesis, cell proliferation, cell migration and immune system processes. In particular, the invention relates to pharmacophore molecules that emulate the three-dimensional structure of a pharmacophore on the mature wild-type human PF4 molecule and to mutants or variants of such pharmacophore molecules, as well as to mimetic compounds (for example, peptidomimetics or small molecules) that have a pharmacophore or pharmacophore-like three-dimensional structure that is substantially the same as that of a PF4 ligand, or that differs in a function-determining aspect from a PF4 ligand and are capable of modulating PF4 activity. The invention also relates to methods of using such mimetic compounds to modulate PF4 activity, as well as to screening methods for identifying further mimetic compounds, including small molecules.
Chemokines are a superfamily of structurally related, secreted, chemotactic peptides primarily affecting leukocyte migration during the inflammatory response. Their sequences are similar and are characterized by a 4-cysteine motif at the N-terminus. Structurally, all family members have a flexible N-terminal region followed by a loop, then three antiparallel beta strands and a single C-terminal alpha helix. One sub-class of chemokines, designated CXC, contain an intervening residue between the first two N-terminal cysteines. IL-8 is the most well-characterized CXC chemokine, but others include Gro-α and Gro-β, platelet factor-4 (PF4) and IL-10. CXC chemokines signal through receptors designated CXCR, where R designates an integer selected from the group of 1-6. All known CXCR are G-protein-coupled receptors having seven transmembrane-spanning alpha-helix domains.
The CXC chemokines have been implicated in human acute and chronic inflammatory diseases such as arthritis, respiratory diseases, and arteriosclerosis, and also in some acute disorders such as heparin-induced thrombocytopenia. Several CXC chemokines function as agonists of platelet function and stimulators of neutrophils. Recently, some chemokines have been shown to regulate endothelial cell migration and proliferation, suggesting a role in angiogenesis (Murdoch et al., Cytokine 1999; 9: 704-712).
Platelet factor 4 (PF4), which is also known as CXCL4, is a member of the CXC sub-family of chemokines derived from platelets. A preferred PF4 amino acid sequence has been described (see, e.g., Poncz et al, Blood 1987, 69:219-223) and is available from the GeneBank Database (Accession No. P02776). This full-length PF4 amino acid sequence is also provided here, in
Other “variant” PF4 polypeptides are also known. For example, one preferred variant, referred to as PF4var1, has been described by Green et al (Mol. Cell. Biol. 1989, 9:1445-1451) and is available from the GeneBank Database (Accession No. P10720). This full-length PF4var1 sequence is also provided in
PF4 is released from platelets during platelet aggregation, stimulates neutrophil adhesion to endothelial cells, and in the presence of co-stimulatory cytokines such as TNF, induces neutrophil degranulation in response to injury (Kasper et al, Blood 2003, 103:1602-1610). In addition, PF4 induces human natural killer cells to synthesize and release the related CXCL molecule IL-8, a potent neutrophil chemoattractant and activator (Marti et al., J Leukoc Biol. 2002; 72(3):590-7). PF4 also binds heparin with high affinity, resulting in the formation of immune complexes comprising PF4, heparin and IgG. These complexes lead to further platelet activation via binding of the IgG Fc to FcγRIIa receptors on platelets, resulting in thrombocytopenia and/or thrombosis in individuals receiving heparin.
Recently, PF4 was shown to bind directly to activated T cells and to inhibit their proliferation as well as the release of IFN gamma (Fleischer et al., J Immunol. 2002; 169(2):770-7). In addition, a peptide comprising amino acid residues 34-58 of PF4 produced a 30-40% inhibition of proliferation of murine hematopoietic progenitors (Lecompte-Raclet et al., Biochemistry. 2000; 39(31):9612-22). This activity has been attributed to the alpha helical motif at positions 34-58 of PF4, allowing a DLQ motif at position 54-56 to bind to the progenitor cells. Inhibition of human leukemic/megakaryocyte cell lines by PF4 was also dependent on certain C-terminal residues (residues 1-24 and 13-24 but not residues 16-24) (Lebeurier et al., J Lab Clin Med. 1996; 127(2):179-85). Abrogation or enhancement of PF4 inhibitory activity could be altered by mutations at specific residues within the 13-24 region.
Another important inhibitory activity of PF4, in particular of a C-terminal fragment comprising amino acid residues 47-70, is its anti-angiogenic activity. PF4 inhibits angiogenesis by binding to fibroblast growth factor 2 (FGF2) and preventing FGF-2 binding to vascular endothelial cells (Hagedorn et al., FASEB J. 200; 15(3):550-2). PF4 also disrupts binding of vascular endothelial cell growth factor, a mitogen for endothelial cells, thereby inhibiting its activity (Gengriniovitch et al., J. Biol. Chem. 1995; 270(25):15059-65). Modified C-terminal fragments of PF4 containing the sequence ELR (or the related modified motif DLR) had several times greater anti-angiogenic activity than the unmodified peptide (Hagedorn et al., Cancer Res. 2002; 62(23):6884-90). A single amino acid residue mutation at residue 52 (Cys52Ser) abolished all inhibitory activities (Hagedorn et al., 2001, supra). The conformation of the C-terminal inhibitory fragment in solution has been determined and has been found to be composed of two helical subdomains which interact with FGF in a specific 1:1 complex. Both subdomains are likely required for inhibition of fibroblast growth factor-driven mitogenesis (Lozano et al., J. Biol. Chem. 2001; 276(38):35723).
Recently, a splice variant of a previously known CXC receptor, CXCR3, was shown to bind PF4 with high affinity and act as a functional receptor for PF4 (Lasagni et al., J. Exp. Med. 2003; 197: 153749). Overexpression of this variant, designated CXCR3-B, in a human microvascular endothelial cell line, resulted in reduced DNA synthesis and in increased apoptosis.
NMR and crystal structures of PF4 demonstrate that the molecule exists as a homotetramer (Mayo et al., Biochemistry. 1995; 34(36):11399-409; and Zhang et al., Biochemistry. 1994; 33(27):8361-6). As described above, different residues from distinct structural motifs in the monomeric form of PF4 have been identified that confer specific activities to the molecule. However, there remains a need in the art for peptidomimetics, as well as for small molecule analogues that can mimic or preserve the functional groups on the amino acid residues within these motifs, for use as specific modulators of the immune response and angiogenesis.
Studies using a fluorescently labeled human recombinant PF4 purportedly show that the molecule preferentially binds at regions of active angiogenesis in vivo. Hansell et al., Am. J. Physiol. (1995) 269 (3 Pt 2):H829-836. This has led to the suggestion that PF4 might be useful as an imaging marker for angiogenesis in certain types of tumors, particularly in breast cancer tumors. Borgstrom et al. (Anticancer Res. (1998) 18(6A):4035-4041) and Moyer et al. (J. Nucl. Med. (1996) 37(4):673-679) describe using a 99mTc-labeled polypeptide that is said to contain the heparin-binding region of PF4. This peptide, which Moyer et al. refer to as P483H, is said to provide high contrast images of infection in vivo. The utility of PF4 as an imaging agent is limited, however, by the molecule's short half-life in blood plasma. Hence, there is a need for molecules, including peptidomimetics and small molecules, that can mimic or preserve the binding activity of PF4, and serve as useful imaging agents in vivo.
The development of such molecules requires elucidation of the entire pharmacophore structure for PF4 and/or PF4 variants (for example, PF4var1), and precise identification of essential and non-essential functional groups for a given activity.
The citation and/or discussion of a reference in this section and throughout the specification is provided merely to clarify the description of the present invention and is not an admission that any such reference is “prior art” to the invention described herein.
In response to one or more of the foregoing needs for PF4 activity modulation, the present invention provides novel pharmacophores that are useful, inter alia, for identifying novel compounds, such as novel peptidomimetics or small molecules, that are PF4 agonists or, alternatively, PF4 inhibitors. In particular, the invention provides a PF4 pharmacophore having at least 7 and preferably 10 functional groups, as set forth in Table 1, infra, and arranged in three-dimensional space in a manner that is substantially identical to the arrangement of corresponding functional groups in a PF4 polypeptide (see, for example,
In a preferred aspect, the invention provides methods for identifying novel or existing compounds interacting with PF4 and/or having PF4-like or PF4 antagonistic activities. Such compounds include peptidomimetics and small molecules. Entities identified according to these methods can be either designed (e.g., in silico) and synthesized, or they can be selected from an existing compound library, e.g., by screening in silico. Entities identified according to these methods will modulate PF4 activity as agonists, antagonists, or inhibitors. In some embodiments, these methods comprise comparing a three-dimensional structure for a candidate compound to a three-dimensional structure of a PF4 pharmacophore (preferably a PF4 pharmacophore as substantially described herein). The three-dimensional structures for many compounds that can be screened according to these methods have already been elucidated and can be obtained, e.g., from publicly available databases or other sources. Alternatively, where the three-dimensional structure of a candidate compound has not yet been elucidated, its structure can often be determined using routine techniques (for example, X-Ray diffraction or NMR spectroscopy). Similarity between these three-dimensional structures and associated intramolecular characteristics (such as hydrogen bond forming properties as proton donors or acceptors, hydrophobic interactions, sulfide bond forming properties and electrostatic interactions) would predict that the candidate compound is a compound that modulates PF4 activity. In particular, the root-mean square deviation (RMSD) between the two three-dimensional structures is preferably not greater than about 1.0. The preselected compounds can then be tested as to whether they have the desired activity, in the presence of the pharmacophore molecule or in the presence of native PF4, the latter in vitro or in vivo. Alternatively, a PF4 mimic displaying the PF4 pharmacophore could be a “stand-in” for PF4 in in vitro screening libraries of compounds for those, if any, that have PF4 modulating activity.
In other embodiments, the invention provides PF4 mimetics, which can be mutant PF4 polypeptides that modulate (enhance or impede) PF4 activity in cells. The mutant PF4 polypeptides of the present invention preferably comprise the mature PF4 amino acid sequence set forth in
In other embodiments of PF4 mutants, the amino acid substitutions include substitutions in the DLQ sequence motif, such as one or more of the amino acid substitutions Gln9→Arg, Gln9→Ala, and Asp7→Ala. Other preferred amino acid substitutions include one or more of Leu11→Ser, Val13→Gln, Thr16→Ala, Gln18→Ala, Val19→Ser and His23→Ala. It should be noted that mimetics of these PF4 mutants are also within the invention, as long as the three-dimensional structure and intramolecular properties of the original and mutated key residues (including the modifications thereof) are preserved. There is also considerable freedom in linker structures present between key residues of the PF4 mutant or of its mimetic, again as long as the three-dimensional structure is preserved. For example, the invention additionally provides, within its scope, mutant PF4 polypeptides that comprise one or more amino acid additions or deletions, in addition to any of the key residue substitutions described above. Preferred mutant PF4 amino acid sequences of the invention comprise an amino acid sequence as set forth in any of SEQ ID NOS:2-30. See also, Table 3, infra.
Mutants used for validation of the pharmacophore are not active since the point of such mutagenesis is to replace one or more residues that are believed to be important for activity, with other residues that are believed to be unimportant for activity (i.e., the replacement of such residues is expected to abolish or modulate activity). If the mutant is deprived of all (or even some) biological activity compared to the wild type molecule, this means that the residue is crucial for biological activity and should be included in the pharmacophore definition.
The nature of the mutation can also be crucial. For example, it may not be beneficial to replace a hydrophilic residue with one that is hydrophobic (for example, alanine) since both will typically lead to the same type of interaction. The environment of the residue selected for mutation can also be crucial. For example, a mutation may give misleading positive or negative results because neighboring residues compensate (e.g., by conformational change) for the constraints imposed or released by the mutation. This can lead to erroneous interpretation of the results. In addition, the nature of the mutation is preferably chosen to avoid a shift of activity of PF4 toward IL8. Otherwise, the resulting mutant may have IL8-like properties.
The coordinates of the validation mutants described here are not important since the mutants have no interesting biological activity. The mimetics of PF4 can be readily determined with the pharmacophore. If the “candidate mimetic” fits on (i.e., is three-dimensionally superimposable with) the pharmacophore, it is a real mimetic. If the candidate contains only a part of the pharmacophore it can be an antagonist, capable of binding the protein target and competing with PF4 but not capable of activating the target. At least one such mimetic is provided in the present invention, and discussed in detail below.
In preferred embodiments, the present invention provides novel compositions that modulate PF4 activity, e.g., as PF4 agonists and/or antagonists. For example, the invention provides a compound having the following chemical formula:
Still other compounds provided by the invention are set forth in Formulas II through VIII illustrated in FIGS. 8 and 9A-9B. In addition, peptide based compounds are provided that can be used, e.g., as PF4 agonists and/or antagonists in accordance with the invention. These include the peptides referred to in the Examples, infra, as P34-56 (SEQ ID NO:157), P37-56 (SEQ ID NO:158), P34-53 (SEQ ID NO:159) and P35-53 (SEQ ID NO:160). A particularly preferred PF4 agonist is the peptide moiety P34-56 (SEQ ID NO:157), whereas the peptide moiety P34-53 (SEQ ID NO:159) is a particularly preferred PF4 antagonist.
In other embodiments, the invention provides detectable markers that are useful for detecting PF4 binding sites, such as PF4 receptors. These detectable markers generally comprise a PF4 antagonist of the invention with a detectable label conjugated thereto. Generally speaking, these detectable markers can be used to detect PF4 binding sites in an individual (for example, in a medical imaging technique such as MRI) by (a) administering the detectable marker to an individual; and (b) detecting the detectable marker's presence in the individual. Previous reports have indicated that PF4 preferably binds to sites of infection and/or angiogenesis in individuals, and can be used to detect certain tumors such as breast cancer tumors. Hence, the methods of this invention can also be used to detect sites of infection and/or angiogenesis in an individual.
These and other aspects of the present invention are described in detail in the following sections.
The present invention pertains to pharmacophore molecules for a cytokine that is referred to here as Platelet Factor 4 or “PF4”. The PF4 cytokine is also known as CXCL4. The PF4 amino acid sequence has been previously described (see, for example, Deuel et al, Proc. Natl. Acad. Sci. U.S.A. 1977, 74:2256-2258; Walz et al, Thromb. Res. 1977, 11:893-898; and Poncz et al., Blood 1987, 69:219-223). The sequence is also available, e.g., on the GenBank databases (Benson et al., Nucleic Acids Research 2003, 31:23-27) under the Accession No. P02776 (GI No. 130304).
For convenience, the invention is described here primarily in terms of the mature PF4 polypeptide whose amino acid sequence is set forth in
In the case of WTPF4, the full length PF4 cytokine (SEQ ID NO:32) is expressed as a polypeptide chain of 101 amino acid residues. The first 31 amino acid residues of this “full length” PF4 amino acid sequence correspond to a domain that is generally referred to as the “signal sequence domain,” whereas the remaining amino acid residues (i.e., residues 32-101 of SEQ ID NO:32) correspond to what is generally referred to as the “mature” PF4 amino acid sequence. On processing, the PF4 signal sequence domain is cleaved and the “mature” PF4 polypeptide, which exhibits PF4 cytokine activity, is secreted by cells. Hence, pharmacophore molecules of the present invention contain the pharmacophoric structure of the mature PF4. For convenience, a mature wild-type human PF4 amino acid sequence is provided in
The three-dimensional structure of PF4 has also been determined by both X-ray crystallography (Zhang et al, Biochemistry 1994, 33:8361-8366) and NMR spectroscopy (Mayo et al, Biochemistry 1995, 34:11399-11409). The coordinates of these structures are available on the Protein Data Bank (Berman et al., Nucleic Acids Research 2000, 28:235-242) under the Accession Numbers 1RHP and 1PFM, respectively. For convenience, a list of coordinates from a preferred three-dimensional structure for mature human PF4 is also provided here in a PDB file format, as an Appendix, infra.
5.1. PF4 Pharmacophores
The term “pharmacophore,” as it is used to describe the present invention, refers to a compound or molecule having a particular collection of functional groups (e.g., atoms) in a particular three-dimensional configuration. More specifically, the term pharmacophore refers to compounds possessing this collection of functional groups in a three-dimensional configuration that is substantially identical to their three-dimensional arrangement on a protein or other compound of interest (referred to here as the “prototype” protein or compound). The present invention concerns the prototype protein PF4. Hence, pharmacophores of the present invention preferably possess a collection of functional groups in a three-dimensional configuration that is substantially identical to their three-dimensional arrangement on PF4. For example, the RMSD between functional groups in a prototype compound of interest and in a pharmacophore should preferably be less than or equal to about one angstrom as calculated, e.g., using the Molecular Similarity module within a molecular modeling program such as QUANTA (available from Molecular Simulations, Inc., San Diego, Calif.).
Preferred pharmacophores are derived from the three-dimensional structure of the protein (preferably the mature or active form of the protein) or other prototype compound of interest that is experimentally determined, e.g., by X-ray crystallography or by nuclear magnetic resonance (NMR) spectroscopy. However, suitable pharmacophores can also be derived, e.g., from homology models based on the structures of related compounds, or from three-dimensional structure-activity relationships. For example, preferred pharmacophores of the present invention are derived from the analysis of point mutations in a PF4 polypeptide, and evaluation of the effects those mutations have on PF4 activity. Suitable PF4 pharmacophores can then be deduced or derived, e.g., by correlating the effects of such mutations to three-dimensional, homology models of a mature PF4.
In preferred embodiments of the invention, PF4 antagonists can be used to detect PF4 receptor molecules, or other PF4 binding sites. The usefulness of detecting such PF4 binding sites is well known in the art. For example, Moyer et al., (J. Nucl. Med. (1996) 37(4):673-679) have described a polypeptide, which they call P483H, that purportedly contains a heparin-binding domain of PF4. 99mTc-labeled versions of this polypeptide are said to provide high contrast images of infection in vivo. Others have suggested that PF4 might be useful as an imaging marker for angiogenesis in certain types of tumors—particularly in breast cancer tumors. Borgstrom et al., Anticancer Res. (1998) 18(6A):4035-4041.
Accordingly, the present invention also provides detectable markers that can be used to detect PF4 binding molecules (for example, PF4 receptor molecules) and PF4 binding. Such detectable markers generally comprise a PF4 antagonist having a detectable label conjugated thereto. The PF4 antagonist can be any compound that binds to a PF4 receptor or binding site without activating the receptor or otherwise inducing PF4-mediated activity. An example of one small molecule antagonist is illustrated in
While
The PF4 antagonist moiety can be readily conjugated to a detectable label according to any technique that is well known and routine to a person having ordinary skill in the art. In preferred embodiments, the detectable marker is used to detect PF4 binding sites in vivo, for example in a medical diagnostic or imaging assay such as magnetic resonance imaging (MRI) or computer assisted tomography (CAT). The PF4 antagonist can be conjugated to any of a variety of contrast or detection agents for such uses, including metals, radioactive isotopes, and radioopaque agents (e.g., gallium, technetium, indium, strontium, iodine, barium, bromine and phosphorus-containing compounds), radiolucent agents, contrast agents, dyes (e.g., fluorescent dyes and chromophores) and enzymes that catalyze a calorimetric or fluorometric reaction. In general, such agents can be attached using any of a variety of techniques known in the art, and in any orientation. See, for example, U.S. Pat. Nos. 5,330,742; 5,384,108; 5,618,513; 5,804,157; 5,952,464; and 6,797,255. One or more water soluble polymer moieties, such as poly-ethylene glycol or “PEG,” can also be conjugated to the PF4 antagonist, e.g., to increase solubility and/or bioavailability of the detectable marker.
As mentioned above, such detectable markers can be used to detect or identify the presence of PF4 binding sites, including the presence of PF4 receptors, in an individual. Generally, such methods comprise steps of administering the detectable marker to the individual, and detecting its presence, e.g. by detecting the presence of the detectable label. Previous reports have indicated that PF4 will preferably bind to sites of angiogenesis and/or infection in an individual. Hence, these methods can also be used to detect sites of angiogenesis and/or infection in individuals. The methods of detecting angiogenesis are particularly useful for detecting the sites of tumors or other cancers in individuals.
In preferred embodiments, these methods detect PF4 binding sites using known methods of medical imaging, such as magnetic resonance imaging (MRI). However, the methods can be practiced using any technique available to a person of ordinary skill for detecting the presence of the detectable label. For example, the methods can also be practiced by detecting the presence of the detectable label in situ (e.g., in a tissue sample from an individual), using, for example, a fluorescent moiety for the detectable label.
Pharmacophores of the present invention are particularly useful for identifying compounds, such as peptidomimetics or small molecules (i.e., organic or inorganic molecules that are preferably less than about 2 kDa in molecular weight, and are more preferably less than about 1 kDa in molecular weight), that modulate PF4 activity in cells (either in vitro or in vivo). For example, in certain embodiments pharmacophores of the present invention can be used to identify compounds that mimic the natural activity of PF4, e.g., by binding to a PF4 receptor. Such compounds, which are capable of increasing or enhancing PF4 activity, are referred to here as PF4 “agonists” or “agonist compounds.” In other embodiments, pharmacophores of the invention can be used to identify compounds that compete with PF4, e.g., for binding to a PF4 receptor, but do not themselves generate any PF4 activity. Such compounds therefore effectively inhibit or decrease PF4 activity, and are referred to here as PF4 “antagonists” or “antagonist compounds.”
Pharmacophore molecules of the present invention are generally more effective, and hence preferable, when the molecule consists essentially of those unique functional groups or elements that are necessary for PF4 activity, while having few if any functional groups or elements that do not affect such activity. Such pharmacophores thereby simplify the search for PF4 agonists and antagonists since the number of functional groups that must be compared between candidate compounds and the pharmacophore is greatly reduced. Accordingly, the present invention provides, in preferred embodiments, a PF4 pharmacophore that consists essentially of at least seven and not more than ten functional groups or “pharmacophore points” bearing the aforementioned spatial relationship Preferred pharmacophore points are given numbers and are set forth in Table I below. Each of these points corresponds to a particular amino acid side chain in the mature PF4 polypeptide sequence set forth in
For consistency, the atoms and functional groups in Table 1 use the same notation that is used in the PDB file set forth as an Appendix, infra.
Preferably, a pharmacophore in the present invention is described using a coordinate system in which each point of the pharmacophore is described by a set of at least three coordinates representing and/or indicating its position in three-dimensional space. In this way, the arrangement of key points in the pharmacophore can be readily modeled and/or visualized (e.g. using various programs and algorithms for modeling molecular structure, such as INSIGHT II described infra). The coordinates of the pharmacophore can also be readily used to compare the pharmacophore structure, as described below, with points in a peptidomimetic or other candidate compound.
Additional parameters can and preferably are also used to describe other properties of the individual pharmacophore points. These can include, in the case of pharmacophore points that are hydrogen bond donors or acceptors, parameters indicating the preferred direction, orientation, size and/or distance of the hydrogen bond. Other parameters that can be used include, for hydrophobic pharmacophore points, a parameter indicating the size (e.g., the distance or volume) of the preferred hydrophobic interaction.
An example of a particularly preferred coordinate system and its use to describe the preferred PF4 pharmacophore is set forth in Example 6.2.5, below. This system can use either Cartesian or spherical coordinates to indicate the position of each pharmacophore point. Those skilled in the art will appreciate that the Cartesian coordinates for a given point can be readily converted into a set of spherical coordinates, and vice-versa, using well-known mathematical relationships between those two coordinate systems that are also set forth in the Example. To describe the preferred size and orientation of hydrogen bonds, the Example also provides, for each hydrogen bond donor and acceptor, coordinates for a hydrogen-bond vector, A, pointing in the direction of the preferred hydrogen bond. The surface area, S, of a preferred hydrogen bonding potential is also provided for each hydrogen bond donor and acceptor in the pharmacophore. This parameter defines the surface of a sphere cap around the hydrogen bonding vector, A, corresponding to the surface where hydrogen bond formation is preferable. For each hydrophobic pharmacophore point, the Example provides a point, m, indicating a point at the closest distance to the pharmacophore point at which undesirable interactions (e.g., interactions with hydrophilic or polar residues, or with polar solvent) should be avoided.
5.2. Peptidomimetics
As noted above, PF4 pharmacophores of the present invention are particularly useful as peptidomimetics and other compounds that are agonists and/or antagonists of PF4 activity. Accordingly, the invention also provides peptidomimetics that are agonists or antagonists of PF4 activity.
Peptidomimetics are described generally, e.g., in International Patent publication no. WO 01/5331 A2 by Gour et al. Such compounds can be, for example, peptides and peptide analogues that comprise a portion of a PF4 amino acid sequence (or an analogue thereof) which contain pharmacophore points substantially similar in configuration to the configuration of functional groups in a mature PF4 pharmacophore. However, one or more pharmacophore points in a peptidomimetic can be modified in a manner that affects PF4 activity (either as an agonist or antagonist), such as by replacement of an amino acid residue displaying that particular pharmacophore point. Alternatively, at least a portion of the peptidomimetics may be replaced by one or more non-peptide structures, such that the three-dimensional structure of functional groups in the pharmacophore is retained at least in part. In other words, one, two, three or more amino acid residues within a PF4 peptide may be replaced by a non-peptide structure. In addition, at least one key amino acid residue can be replaced by another having different characteristics (for example, different properties of hydrophobicity, hydrophilicity, proton donor or acceptor properties, electrostatic properties, etc.). Other portions of a peptide or peptidomimetic can also be replaced by a non-peptide structure.
Typically, peptidomimetics (both peptide and non-peptidyl analogues) may have improved properties (e.g., decreased proteolysis, increased retention or increased bioavailability) that make them more suitable for pharmaceutical compositions than a PF4 peptide. Peptidomimetics may also have improved oral availability. It should be noted that peptidomimetics of the invention may or may not have similar two-dimensional structures, such as sequences and structural formulas. However, all peptidomimetics within the invention with the same activity will share common three-dimensional structural features and geometry with one another, and all will be close to the three-dimensional structure of the pharmacophore of the native human PF4. Each peptidomimetic of the invention may further have one or more unique additional binding elements. The present invention provides methods (described infra) for identifying peptidomimetics.
All peptidomimetics provided herein have a three-dimensional structure that is substantially similar to a three-dimensional structure of a pharmacophore displayed on the native molecule as described above. Generally, the three-dimensional structure of a compound is considered substantially similar to that of a pharmacophore if the two structures have RMSD less than or equal to about one angstrom, as calculated, e.g., using the Molecular Similarity module with the QUANTA program (Biopolymer module of INSIGHT II program available from Accelrys, Inc., San Diego, Calif.) or using other molecular modeling programs and algorithms that are available to those skilled in the art. In preferred embodiments, compounds of the invention have a RMSD less than or equal to about 1.0 Angstrom. More preferably, compounds of the invention have an RMSD that is less than or equal to about 0.5 Angstrom, and still more preferably about 0.1 Angstroms. In particular, a peptidomimetic of the invention will have at least one low-energy three-dimensional structure that is or is predicted to be (e.g. by ab-initio modeling) substantially similar to the three-dimensional structure of a PP4 pharmacophore.
Lower energy conformations can be identified by conformational energy calculations using, for example, the CHARMM program (Brooks et al., J. Comput. Chem. 1983, 4:187-217). The energy terms include bonded and non-bonded terms, including bond length energy. It will be apparent that the conformational energy of a compound can also be calculated using any of a variety of other commercially available quantum mechanic or molecular mechanic programs. Generally, a low energy structure has a conformational energy that is within 50 kcal/mol of the global energy minimum.
As an example, and not by way of limitation, low energy conformations can be identified using combinations of two procedures. The first procedure involves a simulated annealing molecular dynamics approach. In this procedure, the system (which includes the designed peptidomimetics and water molecules) is heated up to above room temperature, preferably to around 600 degrees Kelvin (i.e., 600 K), and is simulated for a period for about 50 to 100 ps (e.g., for 70 ps) or longer. Gradually, the temperature of the system is reduced, e.g., to about 500 K and simulated for a period of about 100 ps or longer, then gradually reduced to 400 K and simulated for a period of 100 ps or longer. The system temperature is then reduced, again, to about 300 K and simulated for a period of about 500 ps or longer. During this analysis, the atom trajectories are recorded. Such simulated annealing procedures are well known in the art and are particularly advantageous, e.g., for their ability to efficiently search the conformational “space” of a protein or other compound. That is to say, using such procedures, it is possible to sample a large variety of possible conformations for a compound and rapidly identify those conformations having the lowest energy.
A second procedure involves the use of self-guided molecular dynamics (SGMD), as described by Wu & Wang, J. Physical Chem. 1998, 102:7238-7250. The SGMD method has been demonstrated to have an extremely enhanced conformational searching capability. Using the SGMD method, therefore, simulation may be performed at 300 K for 1000 ps or longer, and the atom trajectories recorded for analysis.
Conformational analysis of peptidomimetics and other compounds can also be carried out using the INSIGHT II molecular modeling package. First, cluster analysis may be performed using the trajectories generated from molecular dynamics simulations (as described above). From each cluster, the lowest energy conformation may be selected as the representative conformation for this cluster and can be compared to other conformational clusters. Upon cluster analysis, major conformational clusters may be identified and compared to the solution conformations of the cyclic peptide(s). Specifically, a peptidomimetic or other agonist/antagonist compound is optimally superimposed on the pharmacophore model using computational methods well known to those of skill in the art as implemented in, e.g., CATALYST™ (Molecular Simulations, Inc., San Diego, Calif.). A superposition of structures and the pharmacophore model is defined as a minimization of the root mean square distances between the centroids of the corresponding features of the molecule and the pharmacophore. A van der Waals surface is then calculated around the superimposed structures using a computer program such as CERIUS2™ (Molecular Simulations, Inca, San Diego, Calif.). The conformational comparison may also be carried out by using the Molecular Similarity module within the program INSIGHT II.
Similarity in structure can also be evaluated by visual comparison of the three-dimensional structures in graphical format, or by any of a variety of computational comparisons. For example, an atom equivalency may be defined in the peptidomimetic and pharmacophore three-dimensional structures, and a fitting operation used to establish the level of similarity. As used herein, an “atom equivalency” is a set of conserved atoms in the two structures. A “fitting operation” may be any process by which a candidate compound structure is translated and rotated to obtain an optimum fit with the cyclic peptide structure. A fitting operation may be a rigid fitting operation (e.g., the pharmacophore structure can be kept rigid and the three dimensional structure of the peptidomimetic can be translated and rotated to obtain an optimum fit with the pharmacophore structure). Alternatively, the fitting operation may use a least squares fitting algorithm that computes the optimum translation and rotation to be applied to the moving compound structure, such that the root mean square difference of the fit over the specified pairs of equivalent atoms is a minimum. Preferably, atom equivalencies may be established by the user and the fitting operation is performed using any of a variety of available software applications (e.g., INSIGHT II (available from Accelrys Inc. in San Diego, Calif.) or QUANTA, (available from Molecular Simulations)). Three-dimensional structures of candidate compounds for use in establishing substantial similarity can be determined experimentally (e.g., using NMR or X-ray crystallography techniques) or may be computer generated ab initio using, for example, methods provided herein. The use of such modeling and experimental methods to compare and identify peptidomimetics is well known in the art. See, for example, International Patent Publication Nos. WO 01/5331 and WO 98/02452, which are incorporated herein by reference in their entireties (see, Section 7 below).
As one example, and not by way of limitation, chemical libraries (containing, e.g., hydantoin and/or oxopiperazine compounds) may be made using combinatorial chemical techniques and initially screened, in silico, to identify compounds having elements of a PF4 pharmacophore of the invention, which are therefore likely to be either PF4 agonists or antagonists. Combinatorial chemical technology enables the parallel synthesis of organic compounds through the systematic addition of defined chemical components using highly reliable chemical reactions and robotic instrumentation. Large libraries of compounds result from the combination of all possible reactions that can be done at one site with all the possible reactions that can be done at a second, third or greater number of sites. Such methods have the potential to generate tens to hundreds of millions of new chemical compounds, either as mixtures attached to a solid support, or as individual, isolated compounds.
PF4 pharmacophores of the present invention can be used to greatly simplify and facilitate the screening of such chemical libraries to identify those compounds that are most likely to be effective agonists or antagonists of PF4. As a result, library synthesis can focus on those library members with the greatest likelihood of interacting with the target (e.g., a PF4 receptor or the PF4 polypeptide itself), and eliminate the need for synthesizing every possible member of a library (which often results in an unwieldy number of compounds). The integrated application of structure-based design and combinatorial chemical technologies can produce synergistic improvements in the efficiency of drug discovery. By way of example, hydantoin and oxopiperazine libraries may be limited to those compounds that involve only the addition of histidine and valine surrogates to a hydantoin or oxopiperazine backbone.
Peptidomimetic compounds of the present invention also include compounds that are or appear to be unrelated to the original PF4 peptide, but contain functional groups positioned on a nonpeptide scaffold that serve as topographical mimics. Such peptiomimetics are referred to here as “non-peptidyl analogues.” Non-peptidyl analogues can be identified, e.g., using library screens of large chemical databases. Such screens use the three-dimensional conformation of a pharmacophore to search such databases in three-dimensional space. A single three-dimensional structure can be used as a pharmacophore model in such a search. Alternatively, a pharmacophore model may be generated by considering the crucial chemical structural features present within multiple three-dimensional structures.
Any of a variety of databases of three-dimensional structures can be used for such searches. A database of three-dimensional structures can also be prepared by generating three-dimensional structures of compounds, and storing the three-dimensional structures in the form of data storage material encoded with machine-readable data. The three-dimensional structures can be displayed on a machine capable of displaying a graphical three-dimensional representation and programmed with instructions for using the data. Within preferred embodiments, three-dimensional structures are supplied as a set of coordinates that define the three-dimensional structure.
Preferably, the three-dimensional (3D) structure database contains at least 100,000 compounds, with small, non-peptidyl molecules having relatively simple chemical structures particularly preferred. It is also important that the 3D coordinates of compounds in the database be accurately and correctly represented. The National Cancer Institute (NCI) 3D-database (Milne et al., J. Chem. Inf. Comput. Sci. 1994, 34:1219-1224) and the Available Chemicals DIrector (ACD; available from MDL Information Systems, San Leandro, Calif.) are two exemplary databases that can be used to generate a database of three-dimensional structures, using molecular modeling methods such as those described, supra. For flexible molecules, which can have several low-energy conformations, it is desirable to store and search multiple conformations. The Chem-X program (Oxford Molecular Group PLC, Oxford, United Kingdom) is capable of searching thousands or even millions of conformations for a flexible compound. This capability of Chem-X provides a real advantage in dealing with compounds that can adopt multiple conformations. Using this approach, hundreds of millions of conformations can be searched in a 3D-pharmacophore searching process.
Typically, a pharmacophore search will involve at least three steps. The first of these is generation of a pharmacophore query. Such queries can be developed from an evaluation of distances in the three-dimensional structure of the pharmacophore. For example,
Using the pharmacophore query, a distance bit screening is preferably performed on a database to identify compounds that fulfill the required geometrical constraints. First, the candidate compounds are scanned in order to determine their important physical points (i.e., hydrogen bond donors, hydrogen bond acceptors, hydrophobic volumes, etc.) and important geometric parameters (i.e., relative distances between important physical points). Chemical groups (i.e., hydrophobic, NH4+, carbonyl, carboxylate) are used to map the surface of each candidate compound, while interaction fields are utilized to extract the number and nature of key-points within candidate molecules. There are a number of well-known techniques in the art, such as the GRID program (Molecular Discovery Ltd., London, United Kingdom; Goodford, 1985), which automatically extract important physical points and geometric parameters from the candidate molecules.
Once key-points are extracted from candidate molecules, the candidate compounds and the pharmacophores of the present invention are superimposed or aligned. The degree of similarity between the pharmacophore points and the corresponding key-points of the candidate compound is calculated and utilized to determine a degree of similarity between the two molecules. Details of the superposition method that can be utilized to compare the candidate molecules and the pharmacophores of the present invention are found in the following publications, De Esch et al., J Med. Chem. 2001 24:1666-74 and Lemmen et al., J Med. Chem. 1998 41(23):4502-20. Fitting of a compound to the pharmacophore volume can be done using other computational methods well known in the art. Visual inspection and manual docking of compounds into the active site volume can be done using such programs as QUANTA (Molecular Simulations, Burlington, Mass., 1992), SYBYL (Molecular Modeling Software, Tripos Associates, Inc., St. Louis, Mo., 1992), AMBER (Weiner et al., J. Am. Chem. Soc., 106: 765-784, 1984), or CHARMM (Brooks et al., J. Comp. Chem., 4: 187-217, 1983). This modeling step may be followed by energy minimization using standard force fields, such as CHARMM or AMBER. Other more specialized modeling programs include GRID (Goodford et al., J. Med. Chem., 28: 849-857, 1985), MCSS (Miranker & Karplus, Function and Genetics, 11: 29-34, 1991), AUTODOCK (Goodsell & Olsen, Proteins: Structure, Function and Genetics, 8: 195-202, 1990), and DOCK (Kuntz et al., J. Mol. Biol., 161:269-288 (1982)). In addition, compounds may be constructed de novo in an empty active site or in an active site including some portions of a known inhibitor using computer programs such as LUDI (Bohm, J. Comp. Aid. Molec. Design, 6: 61-78, 1992), LEGEND (Nishibata & Itai, Tetrahedron, 47: 8985, 1991), and LeapFrog (Tripos Associates, St. Louis, Mo.).
After the superposition procedure, molecules with a high matching score or high degree of similarity are selected for further verification of their similarity. Programs, such as ANOVA (performed, for example, with Minitab Statistical Software (Minitab, State College, Pa.)), extract differences that are statistically significant for a defined p value (preferably p values are less than 0.05) between the pharmacophore of the present invention and the candidate molecule. Candidate molecules with a p value below the defined p value are rejected.
A number of different mathematical indices can be utilized to measure the similarity between pharmacophore and candidate molecules. The mathematical indices of interest for the present invention are generally incorporated in the software packages. The choice of mathematical indices will depend on a number of factors, such as the pharmacophore of interest, the library of candidate molecules, and the functional groups identified as essential for activity. For a review on this topic see, Frederique et al., Current Topics in Medicinal Chem. 2004, 4: 589-600.
Compounds that have at least one low energy conformation satisfying the geometric requirement can be considered “hits,” and are candidate compounds for PF4 agonists or antagonists. In a specific embodiment of the invention, compounds of the invention are not PF4, PF4 mutants, IL-8, or a peptide having the amino acid sequence selected from the group consisting of: PHSPTAQLIA TLKNGQKISL DLQAP (SEQ ID NO:34); PHSPTVQLIA TLKNGQKISL DLQAP (SEQ ID NO:35); PYSPTAQLIA TLKNGQKISL DLQEP (SEQ ID NO:36); PHSPQTELUV KLKNGQKISL DLQAP (SEQ ID NO:37); PHSPTAQLIA TLKNGQKISV DLQAP (SEQ ID NO:38); AHSPTAQLIA TLKNGQKISL DLQAP (SEQ ID NO:39); AHSPTVQLIA TLKNGQQISL DLQAP (SEQ ID NO:40); AYSPTAQLIA TLKNGQKISL DLQEP (SEQ ID NO:41); AHSPQTELIV KLKNGQKISL DLQAP (SEQ ID NO:42); AHSPTAQLIA TLKNGQKISV DLQAP (SEQ ID NO:43); PHSATAQLIA TLKNGQKISL DLQAP (SEQ ID NO:44); PHSATVQLIA TLKNGQKISL DLQAP (SEQ ID NO:45); PYSATAQLIA TLKNGQKISL DLQEP (SEQ ID NO:46); PHSAQTELIV KLKNGQKISL DLQAP (SEQ ID NO:47); PHSATAQLIA TLKNGQKISV DLQAP (SEQ ID NO:48); AHSATAQLIA TLKNGQKISL DLQAP (SEQ ID NO:49); AHSATVQLIA TLKNGQQISL DLQAP (SEQ ID NO:50); AYSATAQLIA TLKNGQKISL DLQEP (SEQ ID NO:51); AHSAQTELIV KLKNGQKISL DLQAP (SEQ ID NO:52); AHSATAQLIA TLKNGQKISV DLQAP (SEQ ID NO:53); PHSPTAQLIA TLKNGQKISL DLQAPLY (SEQ ID NO:54); PHSPTVQLIA TLKNGQKISL DLQAPLY (SEQ ID NO:55); AHSATAQLIA TLKNGQKISL DLQAPLY (SEQ ID NO:56); PHSPQTELIV KLKNGQKISL DLQAPRY (SEQ ID NO:57); PHSPTAQLIA TLKNGQKISL DLQAPRY (SEQ ID NO:58); PHSTAAQLIA TLKNGQKISL DLQAPLY (SEQ ID NO:59); PHCPTAQLIA TLKNGRKICL DLQAP (SEQ ID NO:60); PHSPTPQLIA TLKNGQKISL DLQAP (SEQ ID NO:61); PHSTAPQLIA TLKNGQKISL DLQAPLY (SEQ ID NO:62); PHSPTAQLIA TLKNGQKISL (SEQ ID NO:63); PHSPTVQLIA TLKNGQKISL (SEQ ID NO:64); PYSPTAQLIA TLKNGQKISL (SEQ ID NO:65); PHSPQTELIV KLKNGQKISL (SEQ ID NO:66); PHSPTAQLIA TLKNGQKISV (SEQ ID NO:67); AHSPTAQLIA TLKNGQKISL (SEQ ID NO:68); AHSPTVQLIA TLKNGQQISL (SEQ ID NO:69); AYSPTAQLIA TLKNGQKISL (SEQ ID NO:70); AHSPQTELIV KLKNGQKISL (SEQ ID NO:71); AHSPTAQLIA TLKNGQKISV (SEQ ID NO:72); PHSATAQLIA TLKNGQKISL (SEQ ID NO:73); PHSATVQLIA TLKNGQKISL (SEQ ID NO:74); PYSATAQLIA TLKNGQKISL (SEQ ID NO:75); PHSAQTELIV KLKNGQKISL (SEQ ID NO:76); PHSATAQLIA TLKNGQKISV (SEQ ID NO:77); AHSATAQLIA TLKNGQKISL (SEQ ID NO:78); AHSATVQLIA TLKNGQQISL (SEQ ID NO:79); AYSATAQLIA TLKNGQKISL (SEQ ID NO:80); AHSAQTELIV KLKNGQKISL (SEQ ID NO:81); AHSATAQLIA TLKNGQKISV (SEQ ID NO:82); PHSPTAQLIA TLKNGRKISL (SEQ ID NO:83); PHSPTVQLIA TLKNGRKISL (SEQ ID NO:84); PYSPTAQLIA TLKNGRKISL (SEQ ID NO:85); PHSPQTELIV KLKNGRKISL (SEQ ID NO:86); PHSPTAQLIA TLKNGRKISV (SEQ ID NO:87); AHSPTAQLIA TLKNGRKISL (SEQ ID NO:88); AHSPTVQLIA TLKNGRQISL (SEQ ID NO:89); AYSPTAQLIA TLKNGRKISL (SEQ ID NO:90); AHSPQTELIV KLKNGRKISL (SEQ ID NO:91); AHSPTAQLIA TLKNGRKISV (SEQ ID NO:92); PHSATAQLIA TLKNGRKISL (SEQ ID NO:93); PHSATVQLIA TLKNGRKISL (SEQ ID NO:94); PYSATAQLIA TLKNGRKISL (SEQ ID NO:95); PHSAQTELIV KLKNGRKISL (SEQ ID NO:96); PHSATAQLIA TLKNGRKISV (SEQ ID NO:97); AHSATAQLIA TLKNGRKISL (SEQ ID NO:98); AHSATVQLIA TLKNGRQISL (SEQ ID NO:99); AYSATAQLIA TLKNGRKISL (SEQ ID NO:100); AHSAQTELIV KLKNGRKISL (SEQ ID NO:101); AHSATAQLIA TLKNGRKISV (SEQ ID NO:102); PHSPTAQLIA TLKNGQKISL ELR (SEQ ID NO:103); PHSPTVQLIA TLKNGQKISL ELR (SEQ ID NO:104); PYSPTAQLIA TLKNGQKISL ELR (SEQ ID NO:105); PHSPQTELIV KLKNGQKISL ELR (SEQ ID NO:106); PHSPTAQLIA TLKNGQKISV ELR (SEQ ID NO:107); AHSPTAQLIA TLKNGQKISL ELR (SEQ ID NO:108); AHSPTVQLIA TLKNGQQISL ELR (SEQ ID NO:109); AYSPTAQLIA TLKNGQKISL ELR (SEQ ID NO:110); AHSPQTELIV KLKNGQKISL ELR (SEQ ID NO:111); AHSPTAQLIA TLKNGQKISV ELR (SEQ ID NO:112); PHSATAQLIA TLKNGQKISL ELR (SEQ ID NO:113); PHSATVQLIA TLKNGQKISL ELR (SEQ ID NO:114); PYSATAQLIA TLKNGQKISL ELR (SEQ ID NO:115); PHSAQTELIV KLKNGQKISL ELR (SEQ ID NO:116); PHSATAQLIA TLKNGQKISV ELR (SEQ ID NO:117); AHSATAQLIA TLKNGQKISL ELR (SEQ ID NO:118); AHSATVQLIA TLKNGQQISL ELR (SEQ ID NO:119); AYSATAQLIA TLKNGQKISL ELR (SEQ ID NO:120); AHSAQTELIV KLKNGQKISL ELR (SEQ ID NO:121); AHSATAQLIA TLKNGQKISV ELR (SEQ ID NO:122); PHSPTAQLIA TLKNGRKISL ELR (SEQ ID NO:123); PHSPTVQLIA TLKNGRKISL ELR (SEQ ID NO:124); DYSPTAQLIA TLKNGRKISL ELR (SEQ ID NO:125); PHSPQTELIV KLKNGRKISL ELR (SEQ ID NO:126); PHSPTAQLIA TLKNGRKISV ELR (SEQ ID NO:127); AHSPTAQLIA TLKNGRKISL ELR (SEQ ID NO:128); AHSPTVQLIA TLKNGRQISL ELR (SEQ ID NO:129); AYSPTAQLIA TLKNGRKISL ELR (SEQ ID NO:130); AHSPQTELIV KLKNGRKISL ELR (SEQ ID NO:131); AHSPTAQLIA TLKNGRKISV ELR (SEQ ID NO:132); PHSATAQLIA TLKNGRKISL ELR (SEQ ID NO:133); PHSATVQLIA TLKNGRKISL ELR (SEQ ID NO:134); PYSATAQLIA TLKNGRKISL ELR (SEQ ID NO:135); PHSAQTELIV KLKNGRKISL ELR (SEQ ID NO:136); PHSATAQLIA TLKNGRKISV ELR (SEQ ID NO:137); AHSATAQLIA TLKNGRKISL ELR (SEQ ID NO:138); AHSATVQLIA TLKNGRQISL ELR (SEQ ID NO:139); AYSATAQLIA TLKNGRKISL ELR (SEQ ID NO:140); AHSAQTELIV KLKNORKISL ELR (SEQ ID NO:141); AHSATAQLIA TLKNGRKISV ELR (SEQ ID NO:142); PHSPTAQLIA TLKNGQKISL ELRAPLY (SEQ. ID NO:143); PHSPTVQLIA TLKNGQKISL ELRAPLY (SEQ ID NO:144); AHSATAQLIA TLKNGQKISL ELRAPLY (SEQ ID NO:145); PHSPQTELIV KLKNGQKISL ELRAPRY (SEQ ID NO:146); PHSPTAQLIA TLKNGQKISL ELRAPRY (SEQ ID NO:147); PHSATAQLIA TLKNGQKISL ELRAPLY (SEQ ID NO:148); PHSPTAQLIA TLKNGRKISL ELRAPLY (SEQ ID NO:149); PHSPTVQLIA TLKNGRKISL ELRAPLY (SEQ ID NO:150); AHSATAQLIA TLKNGRKISL ELRAPLY (SEQ ID NO:151); PHSPQTELIV KLKNGRKISL ELRAPRY (SEQ ID NO:152); PHCPTAQLIA TLKNGRKICL DLQAP (SEQ ID NO:153); PHSPTPQLIA TLKNGQKISL DLQAP (SEQ ID NO:154); PHSTAPQLIA TLKNGQKISL ELRAPLY (SEQ ID NO:155) or PHSPTAQLIA TLKNGQKISL DLQAP (SEQ ID NO:156).
Those skilled in the art will appreciate that a compound structure may be optimized, e.g., using screens as provided herein. Within such screens, the effect of specific alterations of a candidate compound on three-dimensional structure may be evaluated, e.g., to optimize three-dimensional similarity to a PF4 pharmacophore. Such alterations include, for example, changes in hydrophobicity, steric bulk, electrostatic properties, size and bond angle. Biological testing of candidate agonists and antagonists identified by these methods is also preferably used to confirm their activity.
Once an active peptidomimetic has been identified, related analogues can also be identified, e.g., by two-dimensional similarity searching. Such searching can be performed, for example, using the program ISIS Base (Molecular Design Limited). Two-dimensional similarity searching permits the identification of other available, closely related compounds which may be readily screened to optimize biological activity.
The present invention is also described and demonstrated by way of the following examples. However, the use of these and other examples anywhere in the specification is illustrative only and in no way limits the scope and meaning of the invention or of any exemplified term. Likewise, the invention is not limited to any particular preferred embodiments described here. Indeed, many modifications and variations of the invention may be apparent to those skilled in the art upon reading this specification, and such variations can be made without departing from the invention in spirit or in scope. The invention is therefore to be limited only by the terms of the appended claims along with the full scope of equivalents to which those claims are entitled.
6.1. Experimental Procedures
6.1.1 Recombinant PF4 Production
Recombinant PF4 was produced in E. coli as a protein containing a unique methionine residue immediately preceding the PF4 portion. More specifically, expression plasmids were constructed by cloning a synthetic gene encoding native sequence PF4 between the NcoI and XhoI sites in the multiple restriction site region of plasmid pET-15b (available from Novagen, Fontenay-sous-Bois, France). Mutant PF4 genes were generated using standard PCR amplification of synthetic oligonucleotide primers and the wild-type construct as template. All constructs were independently sequenced and verified (Génome Express, Grenoble, France).
BL21(DE) bacteria (available from Novagen, Fontenay-sous-Bois, France) carrying the PF4 plasmids were cultured at 37° C. in EZmix 2×YT medium containing 1 M glucose and appropriate antibiotics. Protein expression was induced in these cell cultures with 1 mM IPTG for 4 hours. Bacterial cells were harvested by centrifugation and were subjected to lysozyme treatment (1 mg/ml) and sonication. The resultant fusion protein was extracted from the lysis pellet with 6 M Urea in 50 mM Tris-HCl, pH 7.4, 5 mM EDTA, and 10 mM DTT. The extracts were then purified using ion-exchange chromatography, and the PF4 proteins were eluted with a gradient of 0-1 M NaCl followed by dialysis into PBS containing 0.5 NaCl. The final protein concentration was determined by use of a BCA Protein Assay Reagent. The homogeneity of recombinant PF4 proteins thus produced was verified by SDS-PAGE and Western blotting with polyclonal antibody against PF4.
6.1.2 Endothelial Cell Cultures
Human umbilical vein endothelial cells (HUVEC) were isolated by collagenase (Roche Diagnostics) digestion as described previously (Jaffe et al., J Clin Invest. 1973; 85(11):2745-56).
Cells were grown in M199 medium containing 15% fetal calf serum (FCS), 5% human serum, 2 mM glutamine, 50 U/ml penicillin, 50 μg/ml streptomycin, 2.5 μg/ml amphotericin B, and 15 mM HEPES. Cultures were maintained at 37° C. amd 5% CO2 in humidified atmosphere. Every 3-4 days, the cultures were harvested by trypsin treatment, diluted, replated and grown to confluence. HUVEC grown until confluence from the second or third passage are preferably used for experiments described here.
6.1.3 Endothelial Cell Proliferation Assays
Inhibition of DNA synthesis was measured by [3H]-thymidine incorporation assay. Cells were plated at 15,000 cells per well in a 24 well-plate in 0.5 ml medium containing 2.5% FCS and allowed to attach for 4 hours at 37° C. Proliferation was then induced by addition of 10 ng/ml of FGF-2, VEGF165 or VEGF121. Increasing concentrations of purified recombinant PF4 proteins were added to some wells and HUVEC were further incubated for 48 hours. [3H]-thymidine (1 μCi/well) was added during the last 20 hours of incubation. Cells were washed twice with PBS and treated with ice-cold 10% (w/v) trichloroacetic acid for 30 minutes. The resulting precipitates were solubilized with 1 M NaOH and incorporated radioactivity was measured in a Beckman LS-6500 multi-purpose scintillation counter.
6.1.4 HUVEC Migration Assay
HUVEC migration was evaluated in a modified Boyden chamber assay. Transwell cell culture chamber inserts with porous polycarbonate filters (8 μM pore size) were coated with 0.2% gelatin. HUVEC suspended in medium supplemented with 2.5% FCS were added to the inserts at 4×104 cells per well. The inserts were placed over chambers containing a chemotactic stimulus (10 ng/ml VEGF165), and cells were allowed to migrate for 4 hours at 37° C. in a CO2 incubator. For inhibition experiments, recombinant PF4 proteins were added to both the lower and upper chambers. After incubation, filters were rinsed with PBS, fixed with 1% paraformaldehyde and stained with hematoxyline of Harris (EMD Chemicals Inc. Gibbstown, N.J.).
The upper surfaces of the filters was scraped with a cotton swab to remove the nonmigrant cells. The upper surfaces of the filters were viewed in a optical microscope at high powered (×200) magnification, and the number of cells within the microscope visualization field was recorded. Each experimental point was performed in triplicate, and 20 visual fields were analyzed per filter.
6.1.5 Molecular Modeling
IL8 and PF4 polypeptide molecules were modeled in a molecular dynamics simulation that ran for 700 ps at 300 degrees Kelvin (i.e., 300 K). The molecules were modeled with periodic boundary conditions in a 62 Å×62 Å×62 Å box with approximately 8,000 water molecules. Seven Cl− ions were included in simulations of the PF4 molecule and 4 Cl− ions in simulations of the IL8 molecule, to neutralize electrostatic charges.
Molecular dynamics simulation of peptides ran for 700 ps at 900 K. The peptides were modeled with periodic boundary conditions in a 62 Å×62 Å×62 Å box with approximately 7680 water molecules and 720 trifluoroethanol molecules. Data from NMR analysis of the peptides were included in the molecular dynamics simulation. Harmonic distance constraints with coupling constants and velocities were adjusted to obtain a conformity between NMR experiments and simulation protocol when comparing coupling constants, relative population of different conformers of the same molecule, chemical shift anisotropy, dipole-dipole relaxation rates and other experimental factors to theoretical data.
Virtual peptides were modeled using Langevin dynamics, or other fast technique that avoids using periodic boundary condition with explicit water solvent, to increase the diversity of test peptides. Virtual peptides were randomly mutated at biologically active residues via computer manipulations. After molecular dynamics, virtual peptides were selected for probable activity using a QSAR filter and synthesized and tested on cell cultures (Grassy G, Calas B, Yasri A, Lahana R, Woo J, Iyer S, Kaczorek M, Floc'h R, Buelow R. Computer-assisted rational design of immunosuppressive compounds. Nat. Biotechnol. 1998; 16(8): 748-52).
Langevin dynamics simulations ran for 700 ps at 900 K under harmonic constraints on the peptide backbone. Quenched dynamics of certain density systems were used along with a distance-dependent dielectric constant (∈) to cool the simulated system to 300 K for re-equilibration. The last conformation obtained at the end of the quenched dynamics was finally submitted to 500 ps of molecular dynamics at 300 K.
6.1.6 Statistical Analysis
Triplicate determinations per experimental point were performed for most experiments, and the results are expressed as the mean±one standard deviation (SD) for the data combined from separate experiments. The significance of differences between groups was determined by a standard Student t-test for unpaired data.
6.2. Results
Peptide fragments of the mature PP4 polypeptide sequence depicted in
Molecular dynamics calculations of full length PF4 and the related IL8 polypeptides were also performed. Active peptides were found to have a triad of amino acid residues Asp-Leu-Gln (DLQ) near the N-terminus with the same conformation as the Asp-Leu-Gln triad in full length PF4.
Next, the PF4 surface was mapped using site-directed mutagenesis. In particular, a series of mutant PF4 polypeptides was generated, and their angiogenic activity in HUVEC cells was investigated using assays such as those described in Section 6.1, above. Table 3 below lists amino acid sequences of the PF4 polypeptides generated, along with each polypeptide designation and sequence identification number (SEQ ID NO.). The first sequence, which is designated WTPF4, corresponds to the wild-type, mature PF4 amino acid sequence that is also depicted in
A
C LCVKTTSQVR PRHITSLEVI KAGPHCPTAQ
A
C LCVKTTSQVR PRHITSLEVI KAGPHCPTAQ
A
C SCQKTASQVR PRHITSLEVI KAGPHCPTAQ
A
C SCQKTASQVR PRHITSLEVI KAGPHCPTAQ
The significance of each mutation described in Table 3, above, is summarized in Table 4, infra, along with a description of the mutation's expected effect on PF4 activity.
By characterizing the activity of these mutations and correlating the results with a three-dimensional structure of PF4, more complete pharmacophore structures for that molecule have been identified. In particular, this PF4 pharmacophore consists essentially of at least seven and up to ten key functional groups and of their spatial relationships that are believed to be critical for specific interactions of PF4 with a PF4-receptor. Each point in this pharmacophore structure corresponds to a particular, unique atom or functional group on an amino acid side chain of the mature PF4 sequence set forth in
6.3. Coordinate System Visualization and Bonding Potentials
The PF4 pharmacophore of the invention was further visualized to elucidate bonding and hydrophobic potential around each of the pharmacophore points. As described above, each pharmacophore point is classified as either a hydrogen bond acceptor, a hydrogen bond donor, or as participating in a hydrophobic interaction. By visualizing these points onto a coordinate system, the hydrophobic volumes and hydrogen bonding spherical surface caps can be better understood for the purposes of agonist/antagonist design.
An origin was chosen and defined as 0 from which both Cartesian and spherical coordinate systems were drawn. The three dimensional figure from
Likewise, the cartesian coordinates, x, y and z, can be readily determined from given spherical coordinates, r, θ and φ, using the relationships:
x=r sin θ cos φ
y=r sin θ sin φ
z=r cos θ
For convenience, preferred Cartesian and spherical coordinates for the pharmacophore points are set fort below in Table 7.
A point, M, was defined as the closest point to a hydrophobic pharmacophore point at which an undesirable interaction could be avoided. The hydrophobic volume around the pharmacophore point is defined as 4/3π(rhy)3 wherein rhy is the distance between the pharmacophore point and point M on the surface of the hydrophobic volume.
Furthermore, one or more hydrogen bond vectors, A, were calculated for each of the polar pharmacophore points using standard electronegativity data.
The surface area of the hydrogen bond cap is defined as 2πRcaph wherein Rcap is the radius of the sphere and h is concave depth of the spherical cap. For convenience, preferred Cartesian and spherical coordinates for the hydrogen bond vector points (A points) for this pharmacophore are set forth below in Table 8. Similarly, the hydrophobic volumes (“Vol”) and hydrogen bonding cap surface areas (“S”) for this pharmacophore are set forth below in Table 9.
6.4. Use of the Pharmacophore in Compound Design
This example demonstrates how a pharmacophore of this invention can be used to identify, design and synthesize compounds that can be either agonists or antagonists of the PF4 receptor. In particular, a lead compound, referred to here as BQ-A01104 (Formula I), is disclosed.
BQ-A01104 is a neutral molecule with one anionic group (a carboxylic acid group) and a cationic group (a quaternary amine in the piperidinium ring). The compound is soluble in an aqueous solution of sodium chloride. The compound comprises all ten of the PF4 pharmacophore points listed in Table 5, supra, held structurally rigid by a scaffold that, for convenience, can be conceptualized a seven distinct subunits or “zones.” The chemical structure of BQ-A01104 is illustrated in
Zone 1 (
Zone 2 (
Zone 3 (
Zone 4 (
Zone 5 (
Zone 6 (
Zone 7 (
Pharmacophore points I, II, V, VI and VIII are connected to backbone subunits in BQ-A011004 via flexible aliphatic chains. By contrast, pharmacophore points III, IV, VII, IX and X are connected to the backbone subunits of BQ-A011004 by chains that are relatively rigid and constrained. These latter pharmacophore points are therefore relatively constrained compared to the former. This reflects the relative flexibility of different pharmacophore points in the PF4 polypeptide itself. For example, restrained flexibility of pharmacophore points X and IX, which are located on the Ala43 and Leu45 amino acid residues of PF4 (SEQ ID NO:1), is imposed by the existence of an α-helix that is necessary for PF4 activity. The stability of this helix is maintained by a capping box present at its N-terminal end. In the PF4 polypeptide (SEQ ID NO:1), therefore, the movements of residues Val13 and Leu11 are restrained due to the rigidity of the PF4 skeleton imposed by two disulfide bridges.
6.4.1 Preparation of BQ-A01104
BQ-A01104 and other compounds identified and designed as either agonists or antagonists of the PF4 receptor can be obtained via standard, well-known synthetic methodology.
Various compounds identified and designed as either agonists or antagonists of the PF4 receptor contain one or more chiral centers, and can exist as racemic mixtures of enantiomers or mixtures of diastereomers. These isomers may be asymmetrically synthesized or resolved using standard techniques such as chiral columns or chiral resolving agents. See, e.g., Jacques, J., et al., Enantiomers, Racemates and Resolutions (Wiley-Interscience, New York, 1981); Wilen, S. H., et al., Tetrahedron 33:2725 (1977); Eliel, E. L., Stereochemistry of Carbon Compounds (McGraw-Hill, NY, 1962); and Wilen, S. H., Tables of Resolving Agents and Optical Resolutions p. 268 (E. L. Eliel, Ed., Univ. of Notre Dame Press, Notre Dame, Ind., 1972).
Some convenient methods are illustrated in Schemes 14. These schemes are merely meant to be illustrative of one synthetic pathway, however, these synthetic pathways can be modified in ways that will be obvious to those skilled in the art to create a variety of compounds. Starting materials useful for preparing the compounds of the invention and intermediates therefore, are commercially available or can be prepared from commercially available materials using known synthetic methods and reagents.
Methods of synthesizing the compounds of the present invention are illustrated in the following schemes. Because of possible discrepancies in using chemical nomenclature, where structures are provided for compounds or moieties the structure, and not the chemical name, controls the definition of the compound or moiety.
In scheme 1, intermediate 5 is produced by first alkylating 4-phenylbutylamine (1) (Aldrich Chemical Co.) with aluminum chloride in water with chloroacetic acid to produce phenylacetic acid compound 2. Compound 2 is reacted with thionyl chloride to produce the acid chloride which is reacted with the benzimidazol-5-yl-methylamine to form the amide compound 3. Benzimidazol-5-yl-methylamine is made in 3 steps from commercially available benzimidazole carboxylic acid (Aldrich Chemical Co.); (1) treatment of the carboxylic acid with thionyl chloride to form the acid chloride, (2) reaction of the acid chloride with ammonia to form the corresponding primary amide (See Beckwith et al. in Zabicky The Chemistry of Amides Wiley, NY, 1970, pg. 73), and (3) reduction of the amide with lithium aluminum hydride in THF to form the desired methyl amine (See Challis et al. in Zabicky The Chemistry of Amides Wiley, NY, 1970, pg. 795). Compound 3 is then alkylated again with 3-chloropropionic acid and aluminum chloride in water to produce the trisubstituted phenyl compound 4. Finally compound 4 is reacted with thionyl chloride and ammonia to convert the carboxylic acid to the amide intermediate 5.
In scheme 2, intermediate 12 is produced by converting the cylcopentenyl amide compound (6) to the 1,3-dicarbonyl compound (7) with osmium tetroxide followed by treatment with sodium periodate and then treatment with water and a mild reducing agent such as NaHSO3. Compound 6 is formed in 3 steps from commercially available cyclopentanone (Aldrich Chemical Co.); (1) an aldol reaction of cyclopentanone with the enolate of ethyl acetate, (2) dehydration of the resultant alcohol by treatment with acid, and (3) conversion of the resultant α,β-unsaturated ester to its corresponding amide upon reaction with the sodium or lithium salt of aniline (Majetich et al. Tetrahedron Lett. 1994, 35, 8727). Compound 7 is oxidized using standard techniques, for example treatment with KMnO4, to the carboxylic acid compound 8. Compound 8 is treated with vinylmagnesium chloride and the resultant alcohol subsequently dehydrated with acid to produce the diene compound 9. The vinyl alkene of compound 9 is brominated with hydrogen bromide followed by hydrogenation of the heptenyl olefin with hydrogen gas in the presence of a catalytic amount of palladium on carbon. Finally the 1-bromoalkane is reacted with magnesium to produce the alkyl grignard reagent 10. Compound 10 is then reacted with 3-aminopropanal in ether to produce alcohol compound 11. Finally compound 11 is reacted with a base, followed by ethylbromide and then acid to form ethyl ether intermediate 12.
In scheme 3, intermediate 19 is produced in three steps from commercially available 3-butenal diethyl acetal (Aldrich Chemical Co.); (1) hydroboration with BH3 followed by oxidation with NaOH/H2O2, 2) conversion of the diethyl acetal to the aldehyde with treatment of catalytic p-toluene sulfonic acid, and (3) protection of the alcohol of 4-hydroxy-butanal to form compound 13. The choice of appropriate protecting groups in this and other steps of the synthesis will be readily determined by one of ordinary skill in the art. Suitable protecting groups and standard techniques for choosing and synthesizing protecting groups can be found in T. W. Greene and P. G. M. Wuts, Protective Groups in Organic Synthesis (Wiley-Interscience, New York, 1999). Compound 13 is then reacted with the grignard reagent (14; forms in two steps, (1) addition of HBr to 4-phenyl-1-butyne (Aldrich Chemical Co.) under Markovnikov conditions, and (2) reaction of the resultant vinylic bromide with magnesium) in ether to produce alcohol compound 15. Compound 15 is reacted with tosyl chloride, or other suitable leaving group precursor, in the presence of base, e.g., NEt3, to produce tosylate compound 16. Compound 16 is reacted with the Gringard reagent formed by protecting 4-bromobutanal (4-bromobutanal is made from 4-hydroxy-butanal (supra) upon treatment with 2,4,6-trichloro[1,3,5]triazine, NaBr and N,N-dimethylformamide in methylene chloride; de Luca et al. Org. Lett., 2002, 4, 553-555) with a protecting group that is orthogonal to P1 and reacting the protected compound with magnesium to form compound 17. Compound 17 is deprotected to remove the original protecting group P1 and the free alcohol is subsequently oxidized to the carboxylic acid with, for example, CrO3. After oxidation, the intermediate is brominated with tribromophospine and bromine gas to form the α-bromo carboxylic acid. The carboxylic acid is then treated with thionyl chloride and the resultant acid chloride is treated with ammonia to produce amide compound 18. The olefin of compound 18 is brominated with hydrogen bromide to afford the primary bromide and the second protecting group (P2) is removed from the intermediate and the resultant alcohol oxidized to the aldehyde using standard methods, e.g. treatment with the Swern or Dess-Martin reagent, to form intermediate 19.
Finally, in scheme 4, the dibromo intermediate 19 is coupled with the amine intermediate 12 in the presence of a base and tert butyl-ammonium iodide (TBAI) to give the piperidine intermediate 20. The carboxylic acid of intermediate 20 is coupled with the amine of intermediate 5 in the presence of DCC and catalytic DMAP followed by oxidation of the remaining aldehyde with, for example KMnO4, to afford title compound I, BQ-A01104.
6.5. Optimization of Activity
Using routine techniques of chemical synthesis and modification, it is possible to further optimize both the activity and Absorption, Distribution, Metabolism and Excretion (ADME) properties of compounds that are designed and/or identified using the pharmacophores of this invention. For example, candidate PF4 agonist or antagonist compounds can be modified either by modifying one or more functional groups that correspond to pharmacophore points, by modifying the scaffolding (e.g. the subunits or “zones” described, supra, for BQ-A011004), or both.
The preparation of the compound of Formula II is illustrated in schemes 5-6. The key modifications to the BQ-A011004 scaffold are the substitution of a cyclohexyl ring for the piperazine ring, and the substitution of an isopropylamide group for the phenylamide group.
The substitution of the cyclohexyl fragment for the piperazine ring is accomplished according to Schemes 5a, 5b. First, γ-caprolactone (Aldrich Chemical Co.) is converted to the corresponding ring-opened ester with treatment of sodium ethoxide (Scheme 5A). The intermediate ester alcohol is oxidized under Swern conditions to the corresponding ester aldehyde, and the aldehyde protected as its dioxolane with ethylene glycol in the presence of catalytic p-TSA. The ester is then reduced to the alcohol with LiAlH4 to provide compound 22. Treatment of compound 22 with N-bromosuccinimide and methylsulfide converts the allylic alcohol to its corresponding allyl bromide (Corey et al. Tetrahedron Lett. 1972, 4339). Reaction of the allyl bromide with the lithium enolate of acetaldehyde (formed from acetaldehyde and lithium diisopropylamide) provides alkene aldehyde compound 23.
As illustrated in Scheme 5B, 3-phenyl-1-propanol (24, Aldrich Chemical Co.) is first oxidized under Swern conditions to the aldehyde and the aldehyde is reacted with vinylmagnesium bromide which, upon reaction workup, affords the corresponding allylic alcohol. The allylic alcohol is first reacted with NBS and DMS to afford the allyl bromide and the bromide is converted to the corresponding Grignard reagent (25) with magnesium. Compound 25 is then added to aldehyde 23 and the resultant alcohol is converted to the corresponding tosylate (26) with tosyl chloride in the presence of base (e.g., NEt3). The tosylate is displaced by treatment with a protected 4-hydroxybutyl Grignard reagent to form diene 27. Ring closing metathesis of compound 27 with a catalytic amount of Hoveyda-Grubbs 2nd generation catalyst (28) (Hoveyda et al. J. Am. Chem. Soc. 121, 791, 1999) followed by a Wacker oxidation with PdCl2 (Tsuji, J. Synthesis 1990, 739) affords cyclic ketone 29.
The remainder of the compound of Formula II is constructed as illustrated in Scheme 6. Compound 10C (See Scheme 9, infra) is reacted with 3-bromopropionaldehyde to afford alcohol 30. The alcohol is then converted to its corresponding ethyl ether upon treatment with ethyl bromide in the presence of base (e.g., NEt3) and the intermediate compound is converted to its corresponding Grignard reagent (31) with magnesium metal. Grignard reagent 31 is then added to compound 29 and the resultant alcohol dehydrated to its corresponding alkene (32) upon treatment with acid (the dioxolane group is also removed in this step). Compound 32 is then hydrogenated in the presence of hydrogen and catalytic palladium on carbon and the aldehyde converted to its corresponding amide by 1) oxidation to the acid with KMnO4, 2) conversion of the acid to the acid chloride with thionyl chloride, and 3) reaction of the acid chloride with ammonia. The resultant amide 33 is then coupled with compound 5 (See Scheme 1) in the presence of DCC and catalytic DMAP. Finally, the compound of Formula II is completed when the protecting group P2 is removed and the resultant alcohol oxidized to its corresponding acid with KMnO4.
The preparation of the compound of Formula III is illustrated in scheme 7-8. The key modifications to the BQ-A011004 scaffold are the substitution of an aminocarbonyl ethyl group for the aminocarbonyl group substituted on the piperazine ring, and the substitution of a 4-[4-aminobutyl]-4,5-dihydropyrazole for the aminomethylbenzimidazole fragment.
The substitution of an 4-[4-aminobutyl]-4,5-dihydropyrazole group is achieved through the synthesis of fragment 5A as illustrated in Scheme 7. Reaction of 6-amino heptyne (IA) with diazomethane under Pechmann conditions (T. L. Jacobs in R. C. Elderfield, Heterocyclic Compounds 5, 70 (New York, 1957)) affords 4-[4-aminobutyl]pyrazole which is reduced to the corresponding 4,5-dihydropyrazole (2A) with hydrogen in the presence of catalytic palladium on carbon. Dihydropyrazole 2A is then coupled with the acid chloride of compound 2 (i.e., reaction of compound 2 from Scheme 1 with thionyl chloride) to form amide 3A. Compound 3A is then alkylated again with 3-chloropropionic acid and aluminum chloride in water to produce the trisubstituted phenyl compound 4A. Finally compound 4A is reacted with thionyl chloride and ammonia to convert the carboxylic acid to the amide intermediate 5A.
The substitution of an aminocarbonyl ethyl group is achieved from compound 17 (prepared in Scheme 3, above). As shown in Scheme 8, Compound 17 is deprotected to remove the original protecting group P1 and the free alcohol is subsequently oxidized to the aldehyde under Swern conditions. After oxidation, the intermediate is brominated with tribromophospine and bromine gas to form the α-bromo aldehyde 18B. The α-bromo aldehyde is reacted with α-(p-nitrophenoxycarbonyl)methyldiethylphosphonate (prepared from the p-nitrophenyl ester of acetic acid and diethylchlorophosphonate in the presence of, for example, NEt3) under Horner Wadworth Emmons conditions to form the corresponding α,β-unsaturated γ-bromo ester. The activated ester is then converted to the corresponding amide 19B by treatment with ammonia (See Beckwith, A. L. J., in Zabicky The Chemistry of Amides; Wiley: NY, 1970, p. 96). The olefin of compound 19B is brominated with hydrogen bromide to afford the primary bromide and the second protecting group (P2) is removed from the intermediate and the resultant alcohol oxidized to the aldehyde using standard methods, e.g., treatment with the Swern or Dess-Martin reagent Finally, the α,β-unsaturated amide is hydrogenated with hydrogen in the presence of catalytic palladium on carbon to afford fragment 20B. To complete the synthesis of the compound of Formula III, compound 20A is coupled with compound 12 under the conditions described in Scheme 4 above. The resultant product is then coupled with compound 5A (See Scheme 5) in the presence of DCC and catalytic DMAP and the aldehyde oxidized to the corresponding carboxylic acid with, for example, KMnO4.
The preparation of the compound of Formula IV is illustrated in Scheme 9. The key modifications to the BQ-A011004 scaffold are the substitution of a 2-methylbutyl group for the ethoxy group y to the piperazine ring, and the substitution of an isopropoyl amide group for the phenyl amide group. The synthesis of a compound with these two modifications can be achieved via the synthesis of modified fragment 13C (Scheme 9). Fragment 13C is produced by converting the cyclopentenyl isopropylamide compound (6C) to the 1,3-dicarbonyl compound (7C) with osmium tetroxide followed by treatment with sodium periodate and then treatment with water and a mild reducing agent such as NaHSO3. Compound 6C is formed in 3 steps form commercially available cyclopentanone (Aldrich Chemical Co.); (1) an aldol reaction of cyclopentanone with the enolate of ethyl acetate, (2) dehydration of the resultant alcohol by treatment with acid, and (3) conversion of the resultant α,β-unsaturated ester to its corresponding amide upon reaction with the lithium isopropylamide (Majetich et al. Tetrahedron Lett. 1994, 35, 8727). Compound 7C is oxidized using standard techniques, for example treatment with KMnO4, to the carboxylic acid compound 8C. Compound 8C is treated with vinylmagnesium chloride and the resultant alcohol subsequently dehydrated with acid to produce the diene compound 9C. The vinyl alkene of compound 9C is brominated with hydrogen bromide followed by hydrogenation of the heptenyl olefin with hydrogen gas in the presence of a catalytic amount of palladium on carbon. Finally the 1-bromoalkane is reacted with magnesium to produce the alkyl grignard reagent 10C. Compound 10C is then reacted with 3-aminopropanal in ether, followed by treatment with mild acid to produce alcohol compound 11C. Compound 11C is converted to its corresponding tosylate 12C by treatment with tosyl chloride in the presence of base, for example NEt3. Finally, treatment of tosylate 12C with 3-methylbutyl magnesium bromide (produced from 3-methyl-1-bromobutane and magnesium in ether) affords fragment 13C. The remainder of the synthesis of the compound of Formula IV can be achieved by substituting compound 13C for compound 12 in Scheme 4 (supra) and carrying out the appropriate coupling reactions with compounds 19 and 5.
The preparation of the compound of Formula V is illustrated in Scheme 10. The key modification to the BQ-A011004 scaffold is the substitution of an isopropoyl amide group for the phenyl amide group. The synthesis of a compound with these two modifications can be achieved via the synthesis of modified fragment 12D. In scheme 10, intermediate 12D is produced by converting the cyclopentenyl isopropylamide compound (6C) to the 1,3-dicarbonyl compound (7C) with osmium tetroxide followed by treatment with sodium periodate and then treatment with water and a mild reducing agent such as NaHSO3. Compound 7 is oxidized using standard techniques, for example treatment with KMnO4, to the carboxylic acid compound 8C. Compound 8C is treated with vinylmagnesium chloride in the resultant alcohol subsequently dehydrated with acid to produce the diene compound 9C. The vinyl alkene of compound 9C is brominated with hydrogen bromide followed by hydrogenation of the heptenyl olefin with hydrogen gas in the presence of a catalytic amount of palladium on carbon. Finally the 1-bromoalkane is reacted with magnesium to produce the alkyl Grignard reagent 10C. Compound 10C is then reacted with 3-aminopropanal in ether, followed by treatment with mild acid to produce alcohol compound 11C. Finally compound 11C is reacted with a base, followed by ethylbromide and then acid to form ethyl ether intermediate 12D. The remainder of the synthesis of the compound of Formula IV can be achieved by substituting compound 12D for compound 12 in Scheme 4 (supra) and carrying out the appropriate coupling reactions with compounds 19 and 5.
The key modification to the BQ-A011004 scaffold for the compound of Formula VI is the substitution of a 4-[4-aminobutyl]-4,5-dihydropyrazole for the aminomethylbenzimidazole fragment. The synthesis is achieved by the coupling of compound 5A (See Scheme 7) with compound 20 (See Scheme 4) with DCC in the presence of catalytic DMAP followed by oxidation of the aldehyde to the corresponding carboxylic acid with, for example, KMnO4.
Pharmacophore molecules of the invention can also be selected or modified by selecting or modifying molecules so that they include certain points of the PF4 pharmacophore while selectively excluding others. For example, without being limited to any particular theory or mechanism of action, lead PF4 antagonists (which bind to but do not activate PF4 receptor) can be selected and/or identified by identifying compounds that include certain pharmacophore points required and/or preferred for binding to the PF4 receptor, while selectively excluding other points that may be required or preferred for target (in this example PF4 receptor) activation. See also, Section 5.1, above.
The chemical structure of one such compound is illustrated in
In preferred embodiments, such PF4 agonist and/or antagonist compounds can be used to detect PF4 receptor polypeptides or fragments thereof. For example, a PF4 agonist or antagonist can be conjugated to a detectable label, and binding of the agonist molecule to PF4 receptor can be detected by detecting the detectable label. In particular embodiments, the PF4 agonist is conjugated to a contrasting agent, for detecting in a medical imaging application such as magnetic resonance imaging (MRI). For imaging purposes, any of a variety of diagnostic agents may be incorporated into a pharmaceutical composition, either linked to a modulating agent or free within the composition. Diagnostic agents include any substance administered to illuminate a physiological function within a patient, while leaving other physiological functions generally unaffected. Diagnostic agents include metals, radioactive isotopes and radioopaque agents (e.g., gallium, technetium, indium, strontium, iodine, barium, bromine and phosphorus-containing compounds), radiolucent agents, contrast agents, dyes (e.g., fluorescent dyes and chromophores) and enzymes that catalyze a calorimetric or fluorometric reaction. In general, such agents may be attached using a variety of techniques as described above, and may be present in any orientation. In such embodiments, one or more water soluble polymers (for example, polyethylene glycol or “PEG”) can also be conjugated to the PF4 agonist or antagonist.
One preferred, exemplary embodiment is illustrated in
6.6. Additional PF4-Derived Polypeptides
In still other embodiments, the present invention provides still other peptides that are derived from the amino acid sequence of PF4, and are useful, e.g. as PF4 agonists and/or antagonists according to methods described here. Particularly preferred polypeptides of these other embodiments include polypeptides having any one or more of the following amino acid sequences:
The peptide designated P34-56 (SEQ ID NO:157), above, is so named because it is derived from the sequence of amino acids corresponding to residues 34-56 in the full-length, mature PF4 amino acid sequence set forth in
The peptide designated P34-53 (SEQ ID NO:159) is likewise named because its sequence is derived from the sequence of amino acids corresponding to residues 34-53 of the full-length, mature PF4 amino acid sequence depicted in
The peptide designated P35-53 (SEQ ID NO:160) is identical to P34-53 (SEQ ID NO:159), except that the His2 residue of P34-53 (SEQ ID NO:159) has been removed. This modification is understood to abolish PF4 binding activity, so that the P35-53 peptide (SEQ ID NO:160) does not bind to or activate PF4 receptor.
Without being limited to any particular theory or mechanism of action, it is believed that the activities of these peptides can be attributed to configurations of certain amino acid residues corresponding to some, but not necessarily all, of the PF4 pharmacophore points described, supra, in this application. This can be more readily seen by comparing three dimensional structures of the different peptides to the PF4 pharmacophore configuration. Two such exemplary comparisons are provided herein, in
Specifically, the bottom half of
The PF4 pharmacophore is partially present in the P34-56 peptide. Specifically, Gln23 in P34-56 (SEQ ID NO:157) mimics the position and orientation of Gln9 in wild-type, mature PF4 (SEQ ID NO:1) and, hence, provides functional groups corresponding to PF4 pharmacophore points III and IV listed in Table 1, supra. Leu22 in P34-56 (SEQ ID NO:157) mimics the position and orientation of Leu8 in WTPF4 (SEQ ID NO:1) and, hence, provides functional groups corresponding to PF4 pharmacophore point VIII. Asp21 in P34-56 (SEQ ID NO:157) mimics the position and orientation of Asp7 in wild-type PF4 (SEQ ID NO:1), and provides functional groups corresponding to PF4 pharmacophore points I and II. The P34-56 peptide (SEQ ID NO:157) residue Leu12 mimics the position and orientation of the Leu11 amino acid residue in WTPF4 (SEQ ID NO:1), and provides a functional group corresponding to pharmacophore point X. P34-56 peptide (SEQ ID NO:157) amino acid residue Ile9 mimics WTPF4 (SEQ ID NO:1) residue Val 13 and provides PF4 pharmacophore point IX. Finally, the His2 amino acid residue of P34-56 (SEQ ID NO:157) mimics Gln18 of WTPF4 (SEQ ID NO:1). This amino acid residue therefore provides a functional group corresponding to PF4 pharmacophore VI. Unlike glutamine, however, the histidine side chain does not comprise an oxygen. Hence, His2 and, by extension, the P34-56 peptide itself (SEQ ID NO:157) do not comprise a functional group corresponding to PF4 pharmacophore point V. A functional group corresponding to PF4 pharmacophore point VII also is not present in the P34-56 peptide (SEQ ID NO:157).
As explained, supra, the P34-56 peptide (SEQ ID NO:157) is derived from and corresponds to the sequence of amino acid residues 34-56 in the WTPF4 amino acid sequence set forth at SEQ ID NO:1. A person skilled in the art will therefore appreciate that amino acid residues His2, Ile9, Leu12, Asp21, Leu22 and Gln23 in that peptide (SEQ ID NO:157) correspond to residues His35, Ile42, Leu45, Asp54, Leu55 and Gln56, respectively, in SEQ ID NO:1. These residues are therefore identified in the bottom half of
Further inspection of
Numerous references, including patents, patent applications and various publications, are cited and discussed in the description of this invention. The citation and/or discussion of such references is provided merely to clarify the description of the present invention and is not an admission that any such reference is “prior art” to the invention described here. All references cited and/or discussed in this specification (including references, e.g., to biological sequences or structures in the GenBank, PDB or other public databases) are incorporated herein by reference in their entirety and to the same extent as if each reference was individually incorporated by reference.
See also, Zhang et al., Biochemistry 1994, 33:8361-8366; and Accession No. 1RHP of the Protein Data Bank (both of which are hereby incorporated by reference and in their entireties).
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/US05/42386 | 11/21/2005 | WO | 00 | 8/1/2007 |
| Number | Date | Country | |
|---|---|---|---|
| 60629883 | Nov 2004 | US | |
| 60634433 | Dec 2004 | US | |
| 60637428 | Dec 2004 | US |
