Patent Grant
8778888
This application claims priority from U.S. Provisional Patent Application No. 61/272,816, filed Nov. 6, 2009, which is hereby incorporated by reference in its entirety.
Applicants assert that the text copy of the Sequence Listing is identical to the Sequence Listing in computer readable form found on the accompanying computer file. Applicants incorporate the contents of the sequence listing by reference in its entirety.
1. Field of the Invention
The present invention relates to the field of engineered peptides, and to the field of peptides which bind to integrins, and particularly to integrin binding peptides that bind to platelet integrins and methods of using the peptides in anti-thrombotic therapies.
2. Related Art
Development of highly specific protein ligands that selectively target a single member in a family of closely related receptors has long been a significant molecular engineering problem. Integrin receptors present a particular challenge because recognition of many family members is mediated by an Arg-Gly-Asp (RGD) consensus sequence. Integrins are a class of diverse heterodimeric (α/β) receptors that are involved in cell adhesion to the extracellular matrix and mediate signaling pathways involved in cell cycle progression. As a result, several family members have generated much interest as potential therapeutic targets in the biomedical and pharmaceutical arenas. Integrins αIIbβ3 and αvβ3 are important clinical targets for prevention of platelet-mediated thrombosis and tumor angiogenesis, respectively; however, the high degree of similarity between these two receptors, as well as the αvβ5 integrin, has made it challenging to generate RGD-containing protein ligands that selectively target only a single integrin with high affinity.
An emerging approach for developing novel protein ligands is to use a naturally-occurring protein as a framework, or scaffold, and introducing amino acid mutations that confer recognition to a specific molecular target. Directed evolution is a powerful combinatorial technique for engineering new molecular recognition properties into protein scaffolds, often with remarkable affinities and specificities. By replacing loops or domains with random or rationally designed mutations that sample a large diversity of amino acid sequence space and isolating variants that possess desired molecular recognition properties, new proteins have been generated that bind diverse targets. Ideal protein scaffolds provide a stable, well-structured core and solvent-exposed loops or domains that are highly tolerant to substitution. Protein scaffolds that have been successfully evolved to bind new targets include fibronectin, A-domains, anticalins, ankyrin repeats, and cystine knots, among others.
Cystine knot (knottin) scaffolds have been used previously in other applications that require rapid biodistribution and short in vivo half lives, such as imaging in living subjects, and offer promise as oral peptide drugs. Knottins are a diverse class of small, highly structured polypeptides with up to 60 amino acids in length and possessing a core domain of three or more interwoven disulfide bonds. The structural rigidity conferred by the disulfide-bonded knottin framework leads to exceptionally high thermal and proteolytic stability, and the solvent-exposed loops spanned by these disulfide bonds are moderately to highly tolerant of mutations. Much of the development of knottin as protein-engineering scaffolds have focused on two family members: the Ecballium elaterium trypsin inhibitor II (EETI-II) and the melanocortin receptor binding domain of the human Agouti-related protein (AgRP). EETI-II contains three disulfide bonds, while AgRP contains five; therefore several truncated versions of AgRP have been developed to simplify the scaffold. In one study, EETI-II and AgRP have been used as scaffolds to introduce entire grafted loops derived from snake venom disintegrins that contain RGD or KGD integrin-recognition sequences resulting in knottin peptides that inhibit αIIbβ3, integrin-mediated platelet aggregation with half-maximal inhibitory concentration (IC50) values in the micromolar range. This study demonstrated that the structural confirmation of the scaffold and residues flanking the RGD sequence are critically important for the biological activity of the engineered peptides and also suggested that judicious selection of these neighboring residues might be an effective strategy by which to generate knottin peptides with enhanced potency.
Previously, EETI-II and AgRP knottins that bind to integrin αvβ3 (the vitronectin receptor) with low- to sub-nanomolar affinities were engineered. In both of these prior studies, peptide mutants that bound αvβ3 integrin were identified from yeast-displayed libraries where a single knottin loop was substituted with a loop containing an RGD motif, and randomized flanking residues. Surprisingly, although the library screens were performed only against αvβ3 integrin, the two engineered scaffolds showed very different integrin specificities. In addition to binding αvβ3 integrin with relative affinities of 10-30 nM, the engineered EETI-II peptides also bound with low nanomolar affinity to the related αvβ5 integrin. Moreover, one peptide was found to bind with, high affinity to α5β1 integrin (the fibronectin receptor) as well as to αvβ3 and αvβ5 integrins. None of these engineered EETI-II peptides bound αIIbβ3 integrin with affinities stronger than the micromolar range. In contrast, the engineered AgRP peptides, bound αvβ3 integrin with high affinity (KD˜1-10 nM) but did not appreciably bind to αvβ5 or α5β1 integrins. Additionally, the AgRP peptides weakly bound αIIbβ3 integrin (KD values could not be accurately determined, due to low affinity, but were estimated at greater than several hundred nM). The specificity of the engineered AgRP peptides for αvβ3 integrin was intriguing in light of previous challenges in developing protein scaffolds that could selectively target αvβ3 integrin with high affinity over αvβ5 and αIIbβ3 integrins. For example, when phage-displayed libraries of tendamistat analogs were screened for variants that bound to αvβ3, αvβ5, or αIIbβ3 integrins, most of the proteins selected bound at least two of these integrins. A common feature in these studies and others is that the conformation of the RGD motif is critical in determining both the affinity and specificity of the ligand-integrin interaction. Accordingly, improved integrin binding affinities and specificities have been achieved with cyclic and highly structured peptides relative to linear peptides. Similarly, the residues flanking the RGD motif have a significant role in determining how the recognition sequence is presented to integrin receptors; thus higher affinities and specificities also have often been achieved when the RGD flanking residues were engineered for optimal binding, using combinational methods, as opposed to simple loop grafting of a sequence from a natural RGD-containing ligand.
To further explore the integrin specificities that can be achieved with engineered cystine-knot peptides, and to expand the repertoire of available integrin-targeting molecules, we sought to determine whether a truncated form of AgRP and AgTx (Agatoxin) which have a C-terminal portion removed, could serve as a scaffold for selectively binding integrins other than αvβ3. Engineered AgRP peptides that selectively bind αIIbβ3 integrin with high affinity versus αIIbβ3 could also provide insights into the mode of binding between integrins and conformationally-restricted scaffold loops. Moreover, such peptides would have therapeutic potential as αIIbβ3 integrin plays a critical role in the aggregation of platelets and is a clinically validated target for thrombosis. Despite the successes of three FDA-approved anti-αIIbβ3 drugs for precautionary or responsive treatment to ischemic complications, there remains an interest in improved αIIbβ3 integrin inhibitors to prevent platelet-aggregation.
Presented below is background information on certain aspects of the present invention as they may relate to technical features referred to in the detailed description, but not necessarily described in detail. That is, individual parts or methods used in the present invention may be described in greater detail in the materials discussed below, which materials may provide further guidance to those skilled in the art for making or using certain aspects of the present invention as claimed. The discussion below should not be construed as an admission as to the relevance of the information to any claims herein or the prior art effect of the material described.
US 2009/0257952 by Cochran et al., published Oct. 15, 2009, entitled “Engineered Integrin Binding Peptides,” Engineered peptides that bind with high affinity (low equilibrium dissociation constant (Kd)) to the cell surface receptors of fibronectin (α5β1) or vitronectin (αvβ3 and αvβ5 integrins) are disclosed. The molecular scaffold is preferably based on a knottin, e.g., EETI (Trypsin inhibitor 2 (Trypsin inhibitor II) (EETI-II) [Ecballium elaterium (Jumping cucumber)], AgRP (Agouti-related protein), and Agatoxin IVB, which peptides have a rigidly defined three-dimensional conformation.
Silverman et al., “Engineered cystine-knot peptides that bind αvβ3 integrin with antibody-like affinities,” J Mol Biol. 2009 Jan. 30; 385(4):1064-75. Epub 2008 Nov. 12, discloses a truncated form of the Agouti-related protein (AgRP), a 4 kDa cystine-knot peptide with four disulfide bonds and four solvent-exposed loops, as a scaffold for engineering peptides that bound to αvβ3 integrins with high affinity and specificity.
Kimura et al., “Engineered cystine knot peptides that bind αvβ3, αvβ5, and α5β1 integrins with low-nanomolar affinity,” Proteins. 2009 Nov. 1; 77(2):359-69, discloses engineered Ecballium elaterium trypsin inhibitor (EETI-II), that to bind to αvβ3/αvβ5 or αvβ3/αvβ5/α5β1 integrins.
Varga-Szabo et al., “Cell Adhesion Mechanisms in Platelets,” Arterioscler Thromb Vasc Biol. 2008:28:403-412 (March 2008) discloses that platelets express 3 different β1-integrins, namely α2β1 (collagen receptor), α5β1 (fibronectin receptor), and α6β1 (laminin receptor) as well as 2 β3-integrins, αIIbβ3 and αvβ3, the latter only being present in very low amounts. Among them, 2 are considered to be of pivotal importance for platelet adhesion and aggregation on the exposed subendothelial matrix: α2β1 and αIIbβ3.
Ruoslahti et al. have obtained a series of patents relating to RGD peptides. For example, U.S. Pat. No. 5,695,997, entitled “Tetrapeptide,” relates to a method of altering cell attachment activity of cells, comprising: contacting the cells with a substantially pure soluble peptide including RGDX where X is any amino acid and the peptide has cell attachment activity. The patent further includes an embodiment where X is any amino acid and the peptide has cell attachment activity and the peptide has less than about 31 amino acids.
Koivunen et al., “Phage Libraries Displaying Cyclic Peptides with Different Ring Sizes: Ligand Specificities of the RGD-Directed Integrins,” Bio/Technology 13:265-270 (1995) discloses selective ligands to the cell surface receptors of fibronectin (α5β1 integrin), vitronectin ((αvβ3 integrin and αvβ5 integrin and fibrinogen (αmβ3 integrin from phage libraries expressing cyclic peptides. A mixture of libraries was used that express a series of peptides flanked by a cysteine residue on each side (CX5C, CX6C, CX7C) or only on one side (CX9) of the insert.
Reiss et al., “Inhibition of platelet aggregation by grafting RGD and KGD sequences on the structural scaffold of small disulfide-rich proteins,” Platelets 17(3):153-7 (May 2006) discloses RGD and KGD containing peptide sequences with seven and 11 amino acids, respectively, which were grafted into two cystine knot microproteins, the trypsin inhibitor EETI-II and the melanocortin receptor binding domain of the human agouti-related protein AgRP, as well as into the small disintegrin obtustatin. The peptide sequences shown are modified from native scaffold sequences, having three disulfide bonds, and IC50 values in the micromolar range were reported.
Wu et al., “Stepwise in vitro affinity maturation of Vitaxin, an αvβ3-specific humanized mAb,” Proc. Nat. Acad. Sci., Vol. 95, Issue 11, 6037-6042, May 26, 1998, discloses a focused mutagenesis implemented by codon-based mutagenesis applied to Vitaxin, a humanized version of the antiangiogenic antibody LM609 directed against a conformational epitope of the αvβ3 integrin complex.
The following brief summary is not intended to include all features and aspects of the present invention, nor does it imply that the invention must include all features and aspects discussed in this summary.
The present invention provides cystine-knot peptides engineered for specificity against αIIbβ3 integrin. It is also an object of the present invention to provide a method of preparing cystine-knot peptides that are engineered for specificity against αIIbβ3 integrin. It is, moreover, an object of the present invention to provide a method of using these cystine-knot peptides in anti-thrombotic therapies.
In certain aspects, the present invention comprises a knottin peptide binding with high affinity to αiibβ3 (written out as alpha IIb beta 3) integrin, having a sequence substantially identical to a sequence having an engineered loop of 9-12 residues inserted into a knottin peptide substantially identical to AgRP or AgTx. The peptide may be selected from the sequences at least 95% identical to an AgRP-scaffold based sequence GCVRLHESCLGQQVPCCDPAATCYC X1X2RGDX6X7X8X9CYCR (SEQ ID NO: 1) or an AgTx-scaffold based sequence GCIAEDYGRCTWGGTPCCRGRGCIC X1X2RGDX6X7X8X9 CECT (SEQ ID NO: 2), where X residues are contained in a loop region and X1 is K, L, R, V, Y or N; X2 is K or G X6 is W, V, M or R; X7 is R, K or N; X8 is G, E, S, V, R, M, L, Q, or T; and X9 is K, R, M or T.
In certain aspects, the present invention comprises a knottin peptide having the above-mentioned loop and scaffold, and having essentially no binding to αvβ3 (written out as alpha v beta 3) integrin.
In certain aspects, the present invention comprises a knottin peptide as set forth in the sequences above, wherein X6 is W and X7 and X9 are each independently R or K, and being specific for αvβ3 integrin.
In certain aspects, the present invention comprises a knottin peptide binding as set forth in the sequences above, binding to both αiibβ3 integrin and αvβ3 integrin, wherein said loop region comprises sequence: X1GRGDVX7VX9 (SEQ ID NO: 134), wherein X7 and X9 are each independently R or K.
In certain aspects, the present invention comprises peptides as described above, wherein the wherein the loop portion X1X2RGDX6X7X8X9 is selected from the group consisting of (b3a) VGRGDVRRK (SEQ ID NO: 107), (b3I) RGRGDVKLR, (SEQ ID NO: 112), (2ba) LKRGDWKGK, (SEQ ID NO: 100), (2bj) NKRGDWRSK, (SEQ ID NO: 105), and (2bO), KKRGDWKER (SEQ ID NO: 106).
In certain aspects, the present invention comprises peptides as described above, where the scaffold is an AgRP sequence further comprising a native C-terminal portion. In certain aspects, the present invention comprises peptides as described above where the scaffold is an AgTx sequence substantially identical to a sequence represented by SEQ ID NO: 133.
In certain aspects, the present invention comprises a knottin peptide having an AgRP scaffold substantially identical to an agouti peptide, according to SEQ ID NO: 1, GCVRLHESCLGQQVPCCDPAATCYC X1X2RGDX6X7X8X9CYCR, wherein X1 is K, L, R, V, Y or N; X2 is K or G X6 is W, V, M or R; X7 is R, K or N; X8 is G, E, S, V, R, M, L, Q, or T; and X9 is K, R, M or T.
In certain aspects, the present invention comprises a knottin peptide having a scaffold substantially identical to an agatoxin sequence GCIAEDYGRCTWGGTPCCRGRGCICX1X2RGDX6X7X8X9CECT (SEQ ID NO: 2), wherein X1 is K, L, R, V, Y or N; X2 is K or G; X6 is W, V, M or R; X7 is R, K or N; X8 is G, E, S, V, R, M, L, Q, or T; and X9 is K, R, M or T.
In certain aspects, the present invention comprises a method of preventing platelet aggregation in a mammal which comprises administering an effective amount of a peptide having a sequence substantially identical to a sequence as described above. The peptide in this case may be contained in a pharmaceutical composition suitable for administration to a subject in need thereof, which may be a human in need of anti-thrombotic therapy. The pharmaceutical composition will comprise an engineered integrin binding knottin and a pharmaceutically-acceptable carrier.
In certain aspects, the present invention comprises a method for preparing a peptide specifically binding to an αiibβ3 integrin, and not binding to a non-αiibβ3 integrin, comprising the steps of: (a) preparing a library expressing a collection of peptides with different binding properties; (b) selecting from said collection peptides that bind to αiibβ3 integrin to obtain a positive pool; and (c) selecting and removing from said positive pool peptides that bind to said non-αiibβ3 integrin. These peptides may be expressed in recombinant yeast, and clones selected where they express high binding for αiibβ3 integrin and essentially no binding for αvβ3 integrin, or other β3 containing integrins.
or 50 nM ▪ αIIbβ3 integrin. (B) and (D) Binding to 10 nM □, 50 nM
, or 100 nM ▪ αvβ3 integrin.
αIIbβ3 integrin (also known as GPIIb-IIIa) is a glycoprotein complex found predominantly on platelets and is responsible for platelet aggregation and thrombosis. Agents that block βIIbβ3 integrin can prevent thrombosis by preventing platelets from binding fibrinogen, fibronectin, and von Willebrand factor. They find uses in prevention and acute treatment of arterial thrombosis including restenosis, drug-eluting stents, and coronary angioplasty.
Monoclonal antibodies and other agents have long circulation times and can lead to pathological bleeding; as a result suboptimal doses must be administered and the therapeutic window is very narrow. Small molecule (oral) drugs have been unsuccessful due to increased bleeding and associated mortality. It is difficult to make specific antagonists for αIIbβ3 integrin that do not also bind other integrins.
The AgRP and other peptides engineered in this work are interesting due to their novel binding specificities and for their potential as therapeutic agents to inhibit thrombosis. Platelets have ˜80,000 copies of αIIbβ3 integrin (glycoprotein IIb/IIIa) on their surface in addition to ˜1000 copies each of α2β1, α5β1, α6β1 integrin, and only ˜100 copies of αvβ3 integrin. These integrins mediate adhesion to fibrinogen as well as von-Willebrand factor, vitronectin, and fibronectin. Prior to ligand binding, αIIbβ3 integrin must be activated by inside-out signaling that leads to active site conformational changes. Fibrinogen presents multiple adhesive sites, leading to platelet crosslinking via binding to activated αIIbβ3 integrin, resulting in clot formation.
Several compounds have been used clinically to prevent αIIbβ3-mediated platelet aggregation for patients at risk of thrombosis of coronary arteries after percutaneous coronary intervention (PCI). The αIIbβ3 integrin-specific antibody fragment c7E3, abciximab (Reopro®), was approved by the FDA in 1994 as adjunctive therapy to prevent ischemic complications of coronary artery angioplasty. Abciximab led to >90% blockage of αIIbβ3 integrin and >80% inhibition of ADP-induced platelet aggregation, but in spite of a 12-hour infusion, over two-thirds of the antibody persists in the patient for several weeks. This led to bleeding issues in many patients, requiring increased red blood cell and platelet transfusions. This issue is particularly relevant in patients who need to undergo emergency coronary artery bypass graft after cessation of abciximab treatment. Additionally, abciximab also binds to αvβ3 integrin and the leukocyte integrin Mac-1. While it is unclear whether or not there is any safety issue with the associated αvβ3 integrin binding, there does not appear to be any treatment benefit to targeting both of these receptors.
Eptifibatide (Integrilin®) is an αIIbβ3 integrin-specific cyclic heptapeptide that contains a homoargenine-glycine-aspartate sequence, and is approved to reduce the risk of acute cardiac ischemic events in patients with unstable angina or undergoing PCI. While the efficacy of eptifibatide is somewhat limited compared to abciximab in that it must be used in conjunction with other anti-platelet medications such as aspirin or clopidogrel as well as heparin, its circulation time is much shorter, such that normal platelet function is restored within 2-4 hours post-treatment. As a result, no increases in bleeding have been reported for patients administered eptifibatide versus control groups.
The engineered AgRP peptides described below have properties that compare favorably to these FDA approved drugs. In addition to high specificity for αIIbβ3 (see e.g. clones identified below as 2bA, 2bJ, and 2bO) or β3 integrins (e.g. clones b3A and b3I), all of the engineered AgRP peptides performed as well or slightly better than eptifibatide in inhibiting ADP-induced platelet aggregation. It has also been shown that radiolabeled versions of the present αvβ3 integrin-binding AgRP peptides have rapid blood clearance in mice of ˜10-20 minutes, but are still able to elicit high uptake in integrin-expressing tumor xenografts models using human glioma derived U87MG cells. (See, for data, Jiang, L., Kimura, R. H., Miao, Z., Silverman, A. P., Ren, G., Liu, H. G., Li, P. Y., Gambhir, S. S., Cochran, J. R., and Cheng, Z., “Evaluation of a 64Cu-Labeled Cystine-Knot Peptide Based on Agouti Related Protein for PET Imaging of Tumors Expressing αvβ3 Integrin,” Journal of Nuclear Medicine, 51, 251-8. 2010). Therefore, it seems likely that these engineered peptides would be able to target and bind platelets in vivo, and their short half-life will be desirable for reducing the risks of bleeding and other side effects. The present engineered knottins have also shown promise for oral availability; however, thus far only injectable proteins and peptidomimetics have been proven successful, as orally administered small molecule αIIbβ3 antagonists appear to actually induce platelet aggregation in vivo. Platelet aggregation is measured in vitro by a number of known techniques, and an effective dosage may be determined in this way. Platelet aggregation may be measured by using a cone-and-plate aggregometer. Platelet response may also be monitored as a percentage reduction in adenosine 5′-diphosphate-induced platelet aggregation at selected days after beginning of therapy and compared with baseline.
Also of clinical interest is the role of αIIbβ3 integrin in tumor cell-induced platelet aggregation. Many disintegrins, cysteine-rich peptides found in snake venom, contain RGD or KGD sequences and block platelet aggregation by binding to αIIbβ3 integrin. Interestingly, two such αIIbβ3-specific disintegrins, trigramin and rhodostomin, have shown efficacy in blocking tumor cell-induced platelet aggregation, a critical component of tumor transvascular metastasis. Thus, there may be potential for using αIIbβ3-specific peptides, such as those described in this work, as therapeutic inhibitors of tumor metastasis.
As described in Yeh et al., “Rhodostomin, A Snake Venom Disintegrin, Inhibits Angiogenesis Elicited by Basic Fibroblast Growth Factor and Suppresses Tumor Growth by A Selective αvβ3 Blockade of Endothelial Cells,” Mol. Pharm. 59(5):1333-1342 (2001), disintegrins are a family of low-molecular-weight, RGD-containing peptides that bind specifically to integrins αIIbβ3, α5β1, and αvβ3 expressed on platelets and other cells including vascular endothelial cells and some tumor cells. In addition to their potent antiplatelet activity, studies of disintegrins have revealed a new use in the diagnosis of cardiovascular diseases and the design of therapeutic agents in arterial thrombosis, osteoporosis, and angiogenesis-related tumor growth and metastasis. It has been shown that disintegrin inhibited adhesion of tumor cells to extracellular matrices and in vivo experimental metastasis of B16 murine melanoma cells. Investigators have demonstrated that disintegrin inhibits either spontaneous or tumor associated angiogenesis.
As exemplified below, AgRP can be used as a scaffold for engineering peptides with different integrin specificities by mutating a single loop of the knottin peptide. These examples also indicate that creative library screening strategies can be developed and used to engineer scaffold proteins where target specificity is an issue. Additionally, AgRP peptides engineered in accordance with the present invention showed high efficacy in inhibiting platelet aggregation in vitro and have translational potential as thrombosis inhibitors.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the art to which this invention belongs. Although any methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, the preferred methods and materials are described. Generally, nomenclatures utilized in connection with, and techniques of, cell and molecular biology and chemistry are those well known and commonly used in the art. Certain experimental techniques, not specifically defined, are generally performed according to conventional methods well known in the art and as described in various general and more specific references that are cited and discussed throughout the present specification. For purposes of clarity, the following terms are defined below.
The term “knottin peptide” or “cystine knot peptide” is used herein to refer to a miniprotein with a knotted topology with three to five disulfide bridges with one disulfide penetrating through a macrocycle formed by two other disulfides and inter-connecting peptide backbones. Other terms used in the art to describe these miniproteins are “knottins” or “cystine knots”. Information about various known knottin peptides, such as primary sequences, 3D-structures and folding patterns as well as function, may be obtained by consulting the KNOTTIN database, which is freely available, at http(colon)//knottin(dot)cbs(dot)cnrs(dot)fr or http(colon)//knottin(dot)com.
Agouti-related protein (AgRP), a knottin peptide, contains a highly basic N-terminal region, a Pro-rich central domain and a C-terminal region rich in cysteine (Cys) residues. From the sequence, loop 4 of AgRP is between the 6th and 7th cysteine from the N-terminus. The same orientation may be used for other knottin peptides having the same folding pattern as AgRP. As noted below, while any of the four exposed solvent loops of the knottin peptide may be substituted with a consensus sequence of the present invention, it is generally preferred if the fourth exposed solvent loop (as defined herein) is substituted with a consensus sequence.
Miniaturized or mini as used herein for either AgRP (Agouti-related protein) or AgTx (any of the Agatoxins) means either of the AgRP or AgTx having a C-terminus portion removed. Procedures for removal of the C-terminus portion of AgRP are well-known and described in Jackson et al (2002) Biochemistry 41: 7565 and McNulty (2001) Biochemistry 40:15520. Based on the sequence for omega agatoxin, the cystine-knot domain (C-terminus removed) is approximately 38 amino acids, although this may differ slightly for other agatoxins. The present inventors have obtained miniaturized AgRP using the procedure of Jackson et al. AgTx was miniaturized (C-terminus removed) by performing a sequence alignment with AgTx and mini-AgRP, and performing the analogous truncation.
The cystine knot portion of the AgRP is residues 87-132, which is itself the C-terminal portion of the natural AgRP. What is referred to herein as mini-AgRP has been truncated to AgRP (87-120, C105A). Thus residues 121-132 are analogous to the C-terminus portion of AgTx.
The term “amino acid” includes both naturally occurring and synthetic amino acids and includes the D and L form of the acids as well as the racemic form. More specifically, amino acids contain up to ten carbon atoms. They may contain an additional carboxyl group, and heteroatoms such as nitrogen and sulfur. Preferably the amino acids are α and β-amino acids. The term α-amino acid refers to amino acids in which the amino group is attached to the carbon directly attached to the carboxyl group, which is the α-carbon. The term β-amino acid refers to amino acids in which the amino group is attached to a carbon one removed from the carboxyl group, which is the β-carbon. The amino acids described here are referred to in standard IUPAC single letter nomenclature, with “X” meaning any amino acid.
The term “AgRP” means PDB entry 1HYK. Its entry in the Knottin database is SwissProt AGRP_HUMAN. It has the sequence GCVRLHESCLGQQVPCCDPCATCYCRFFNAFCYCR-KLGTAMNPCSRT (SEQ ID NO: 132).
The dashed portion shows a C terminal fragment omitted in the “mini” version, below. The bold and underlined sequence above and in the sequences below represents a loop portion that is replaced with an integrin binding sequence that is not native to the knottin, but is created artificially as described below.
Disulfide bonds are between cysteines 1-16, 8-22, 15-33, 19-43, and 24-31 in the native sequence. The native sequence does not contain the added N terminal glycine and thus begins with Cys 1.
The term “mini” in reference to AgRP means PDB entry 1MRO. It is also SwissProt AGRP_HUMAN. It has the sequence, similar to that given above, GCVRLHESCLGQQVPCCDPAATCYCRFFNAFCYCR (SEQ ID NO: 3) where the italicized “A” represents an amino acid substitution which eliminates a possible dimer forming cystine. The native disulfide bond (cys19-cys43) is missing because of this substitution and the omission of the C-terminal fragment. (Cysteine herein refers to the single amino acid; cystine to the dimer.) The bold and underlined sequence above and in the sequences below represents a loop portion that is replaced with an integrin binding sequence that is not native to the knottin, but is created artificially as described below.
Thus, the term AgRP, as used here to refer to a scaffold for an integrin binding loop, as indicated above, can include a “native AgRP,” with 5 disulfide bonds, an engineered AgRP that has all 46 residues, plus an additional residue, e.g. G at the N terminus and a Cys replacement with a neutral residue such as Val, Ala or Gly; a “mini” AgRP that has the C-terminal portion omitted as shown above, as well as a modified N-terminus and modified Cys content.
The term “agatoxin” means omega agatoxin PDB 1OMB and the SwissProt entry in the knottin database TOG4B_AGEAP. It has the sequence EDN-CIAEDYGKCTWGGTKCCRGRPCRCSMIGTNCECT-PRLIMEGLSFA (SEQ ID NO: 133).
The dashes indicate N- and C-terminal portions of the peptide omitted for the “mini” agatoxin. As shown in Table 3, an additional glycine is added to the N-terminus of the mini-construct.
The term “substantial identity” in the context of a peptide indicates that a peptide comprises a sequence with at least 70% sequence identity to a reference sequence, preferably 80%, more preferably 85%, most preferably at least 90% or 95% sequence identity to the reference sequence over a specified comparison window, which in this case is either the entire peptide, a molecular scaffold portion, or a binding loop portion (˜9-11 residues). Preferably, optimal alignment is conducted using the homology alignment algorithm of Needleman and Wunsch (1970) J. Mol. Biol., 48:443 453. An indication that two peptide sequences are substantially identical is that one peptide is immunologically reactive with antibodies raised against the second peptide. Another indication for present purposes, that a sequence is substantially identical to a specific sequence explicitly exemplified is that the sequence in question will have an integrin binding affinity at least as high as the reference sequence. Thus, a peptide is substantially identical to a second peptide, for example, where the two peptides differ only by a conservative substitution. “Conservative substitutions” are well known, and exemplified, e.g., by the PAM 250 scoring matrix. Peptides that are “substantially identical” share sequences as noted above except that residue positions that are not identical may differ by conservative amino acid changes. As used herein, “sequence identity” or “identity” in the context of two nucleic acid or polypeptide sequences makes reference to the residues in the two sequences that are the same when aligned for maximum correspondence over a specified comparison window. When percentage of sequence identity is used in reference to proteins it is recognized that residue positions which are not identical often differ by conservative amino acid substitutions, where amino acid residues are substituted for other amino acid residues with similar chemical properties (e.g., charge or hydrophobicity) and therefore do not change the functional properties of the molecule. When sequences differ in conservative substitutions, the percent sequence identity may be adjusted upwards to correct for the conservative nature of the substitution. Sequences that differ by such conservative substitutions are said to have “sequence similarity” or “similarity.” Means for making this adjustment are well known to those of skill in the art. Typically this involves scoring a conservative substitution as a partial rather than a full mismatch, thereby increasing the percentage sequence identity. Thus, for example, where an identical amino acid is given a score of 1 and a non-conservative substitution is given a score of zero, a conservative substitution is given a score between zero and 1. The scoring of conservative substitutions is calculated, e.g., as implemented in the NIH Multiple alignment workshop (http (colon) slash slash helixweb(dot)nih (dot) gov/multi-align/). Three-dimensional tools may also be used for sequence comparison.
To determine if a peptide is substantially identical to a given sequence, one will first identify the loop region apart from the scaffold region. The loop region will be defined within 1-2 amino acids that can be varied, i.e. 1/9 residues or about 89% percentage sequence identity. Furthermore, the sequence RGD may not be varied, unless replaced by an equivalent sequence as discussed below. The scaffold sequence can tolerate greater diversity (e.g. 30/37 or about 80% percentage sequence identity) provided that the topology is maintained by intramolecular bonds. This will generally require that the cysteine residues remain unchanged in identity and sequence register, i.e. spacing within the peptide and spacing relative to the cysteine locations.
As used herein, “percentage of sequence identity” means the value determined by comparing two optimally aligned sequences over a comparison window, wherein the portion of the polynucleotide sequence in the comparison window may comprise additions or deletions (i.e., gaps) as compared to the reference sequence (which does not comprise additions or deletions) for optimal alignment of the two sequences. The percentage is calculated by determining the number of positions at which the identical nucleic acid base or amino acid residue occurs in both sequences to yield the number of matched positions, dividing the number of matched positions by the total number of positions in the window of comparison, and multiplying the result by 100 to yield the percentage of sequence identity.
A sequence identity may be specified by percentage to allow certain substitutions, with the proviso that amino acid identities in the proviso are not changed.
The term “anti-thrombotic therapy” is used as it is conventionally understood by those in the art, and refers to the use of platelet aggregation inhibitors in treating occlusive vascular disease. For example, WO 95/12412 discloses platelet-specific chimeric antibodies and methods of using the same in treating various thrombotic disorders. A prototype description of the efforts to develop this therapeutic agent and obtain approval for its use as a human therapeutic agent (generic name: abciximab, trade name: ReoPro®) was described by B. S. Coller in Circulation 92:2373 (1995).
The term “high affinity” means having an equilibrium binding constant (KD; 50% maximum binding) between the peptide and integrin ligand in soluble form of less than 100 nM. High affinity may also include KD's of less than about 60 nM or less than about 30 nM. Contrariwise, essentially no affinity would be essentially no binding at 250 nM.
1. General Considerations
The present invention provides peptide scaffolds into which are inserted engineered sequences that bind specifically to integrins such as αIIbβ3 integrin and αvβ3 integrin with high affinity. The peptides have fast clearance, but since the molecules are very high affinity (e.g. less than 100 nanomolar), they potentially have a broader therapeutic window. Exemplary features of the present peptides are:
1) peptides that are highly specific for αIIbβ3 (alpha IIb beta III) integrin over αvβ3 (alpha v beta III) integrin, or bind to both, and have high affinity, in the nanomolar range;
2) peptides that are serum-stable and are non-immunogenic;
3) peptides where a loop of a known knottin scaffold, such as AgRP (Agouti-Related Protein) or AgTx (Agatoxin) is replaced with an engineered sequence at a loop region, where the native sequence is replaced with 5-10 amino acids which include the RGD or;
4) modifications to knottin scaffolds, e.g. where a truncated-Agouti-related protein is substituted with a loop of 5-10 amino acids, which amino acids include the RGD sequence, or related specific integrin binding sequence such as KGD. U.S. Pat. No. 5,519,005, to Ofer et al., relates to certain non-peptidic compounds comprising a guanidino and a carboxyl terminal groups with a spacer sequence of 11 atoms between them, which are effective inhibitors of cellular or molecular interactions which depend on RXD or DGR recognition, wherein X is G (gly), E (glu), Y (tyr), A (ala) or F (phe). These RXD and DGR analogues are referred to as “RXD surrogates.”
2. Selection of Peptides
According to the present methods, randomized inserts such as XXRGDXXXX, where X is any amino acid, were cloned into cysteine knot peptides and expressed on the surface of yeast. The resulting library was screened using yeast surface display and FACS for peptides that bound αIIbβ3 integrin; to maximize specificity, negative selection was also performed against αvβ3 integrin. The substituted loop for the αIIbβ3 integrin-specific peptides had the consensus sequence: XKRGDWX7XX9, (SEQ ID NO: 4) where X=variable amino acids and X7,9=Arg or Lys. The substituted loop for the peptides that bind both αIIbβ3 integrin and αvβ3 integrin had the consensus sequence XGRGDV X7XX9, (SEQ ID NO: 5) where X=variable amino acids and X7,9 Arg or Lys. The consensus sequences described herein can be used with both AgRP and AgTx.
The present engineered AgRP or AgTx variants were identified from a loop-substituted yeast display library where clones were selected for binding to αIIbβ3 integrin. Fluorescence-activated cell sorting was used. Using different yeast sorting strategies, AgRP variants were isolated with specificity for either αIIbβ3 or for both integrins having a β3 subunit. When produced in soluble form, the engineered peptides bound to cellular integrins with high affinities and expected specificities. In addition, it was determined that the engineered peptides serve as potent inhibitors of platelet aggregation.
A constrained six amino acid loop in AgRP was replaced with a nine amino acid loop containing an Arg-Gly-Asp integrin recognition motif, and the neighboring residues were randomized, to create a library of 20 million AgRP mutants. These AgRP mutants were displayed as fusions on the surface of yeast and used high-throughput fluorescence-activated cell sorting (FACS) to isolate AgRP variants that bound the platelet integrin αIIbβ3. Several rounds of library screening were performed solely against αIIbβ3 integrin and it was found that the isolated AgRP peptides bound with high affinity both αIIbβ3 and the related αIIbβ3 integrin, presumably through specificity of the β3 subunit. By adopting a strategy of alternating positive sorting for binding to αIIbβ3 integrin with negative sorting against binding to αvβ3 integrin, a second set of AgRP peptides with high affinity and specificity for αIIbβ3 integrin (αIIbβ3-specific peptides) was isolated.
The specificities of the engineered AgRP peptides were measured against cell surface expressed integrin receptors. The αIIbβ3-specific peptides bound αIIbβ3 integrin with KD values from 59-87 nM and did not bind to αvβ3, αvβ5, or α5β1 integrins. In comparison, the β3-specific peptides bound αIIbβ3 and αvβ3 integrins with KD values ranging from 40-70 nM and 23-28 nM, respectively, and did not bind to αvβ5 or α5β1 integrins. Unique consensus sequences were identified within both series of peptides and were used to offer potential insight into molecular recognition events that dictate different integrin binding specificities. In addition, we showed that the engineered AgRP peptides prevented platelet aggregation as well as or slightly better than the FDA-approved cyclic peptide eptifibatide. Collectively, these data demonstrates that cystine-knot peptides can be generated with high affinity, specificity, and functional efficacy against closely-related and clinically-significant integrin receptors and suggests that this engineering approach may be extended to develop targeting peptides against other receptors of interest.
3. Peptide Sequences
The engineered AgRP and AgTx peptides were recombinantly produced in Pichia pastoris, a yeast expression system providing high yields and eukaryotic folding capability, to produce proper folding and intramolecular disulfide bonds. Using recombinant techniques, the present sequences can be varied as desired. In addition, the peptides can be produced as fusion proteins with other proteins or protein domains. For example, an engineered αIIbβ3-specific knottin peptide such as disclosed here could be prepared as part of a larger sequence attached to the N- or C-terminus encoding a protein or protein domain that is to be directed to platelets or megakaryocytes, or an adjunct protein such as the thrombolytic, tissue plasminogen activator. The present knottin peptides can be prepared as multimers, or in clusters. Also, a knottin peptide as described here could be coupled to a radiolabel for imaging.
Below is an exemplary list of peptide sequences from the cystine-knot peptides of the present invention that first binds both αIIbβ3 and second that bind both αIIbβ3 and αvβ3, i.e. in Tables 1 and 2, respectively. As described above, these peptides contain portions of the native AgRP which is CVRLHESCLGQQVPCCDPCATCYCRFFNAFCYCRKLGTAMNPCSRT (SEQ ID NO: 6).
As stated, either AgRP or AgTx may be used as a cystine-knot peptide scaffold upon which to make the described sequence substitutions. If AgTx is selected, any one of the known Agatoxins may be used, however, the C-terminus portion of each is preferably removed before use in order to preclude toxicity. The C-terminus may be removed in accordance with conventional procedures. Both AgRP and AgTx with or without the C-terminus portion may be used interchangeably in view of their identical or similar folding patterns.
4. Tertiary Structures and Variations
The present knottin scaffolds may be varied in sequence while preserving the important three dimensional structure of the scaffold. One may use, for example, AgTx variants that have homologous 3D structures to AgRP. For example, there are Agatoxin IVA and Agatoxin IVB variants that have different sequences, but they are structurally homologous. In fact, when the pattern of Cys residues and disulfide bonds is conserved, then the structure will be similar with a folding pattern identical or similar to AgRP. What is important is the 3D topology of the loop in question which is formed by the structure.
1AGG is the PDB database entry for OMEGA-AGATOXIN-IVB. This polypeptide has the sequence EDNCIAEDYGKCTWGGTKCCRGRPCRCSMIGTNCECTPRLIMEGLSFA (SEQ ID NO: 96).
Another agatoxin sequence that may be used is 1OAV. 1OAV is the PDB entry for OMEGA-AGATOXIN IVA. It has the amino acid sequence KKKCIAKDYGRCKWGGTPCCRGRGCICSIMGTNCECKPRLIMEGLGLA (SEQ ID NO: 97).
The two versions of agatoxin A and B, while very close in knottin part, differ much more in the termini, and thus have different interactions and activities with the calcium channels.
Any AgTx variant may be used as a scaffold for carrying a consensus sequence as disclosed hereinabove, provided that the AgTx variant has a folding pattern that is of the same type as the AgTx folding pattern. Protein folding is the physical process by which a polypeptide folds into its characteristic and functional three-dimensional structure from the polypeptide primary structure, i.e., amino acid sequence. Usually, folded proteins have a hydrophobic core in which side chain packing stabilizes the folded state, and charged or polar side chains occupy the solvent-exposed surface where they interact with surrounding water.
Exemplary and well-known techniques for studying protein folding include circular dichroism, dual polarization interferometry, energy landscape theory of protein folding and computational predictions of protein tertiary structure. X-ray crystallography and NMR may also be mentioned.
Additionally, attention is also directed to U.S. Pat. No. 5,680,319, entitled “Hierarchical protein folding prediction,” which, as is noted at the end of the specification, is incorporated herein by reference. This patent describes one method of predicting the three-dimensional structure of a protein fragment and could be used for varying scaffold sequences. As noted above, the loop sequences may be randomized and selected by binding affinity, according to the methods taught here.
Generally, the following guidelines may be used in preparing consensus sequence-substituted AgTx variants for use in accordance with the present invention:
Generally, any of the agatoxins may be used as scaffolds provided that it has the same or similar folding pattern as AgRP. Any of the alpha-, mu- or omega-agatoxins may be selected for further suitability. An example of an alpha variant is AG 489. Of the mu-variants, subtypes 1, 2, 3, 4, 5, and 6 may be noted having 36, 37, 38, 37, 37 and 37 amino acids, respectively. Examples of omega variants include, for example, subtypes IA, IB, IIA, IIB, IIIA, IIIB, IIIC, IIID, IVA and IVB with amino acid lengths varying from 48 to 112 amino acids. By ‘further suitability’ is meant determining the folding pattern based on known methodologies and then conducting studies to determine the preferred loci for placing any of the disclosed consensus sequences described hereinabove.
5. Formulations
The present invention also encompasses a pharmaceutical composition useful in the treatment or prevention of undesirable thrombi, e.g. in treatment of individuals who suffer from coronary artery disease or injury following myocardial infarction, atherosclerosis, arteriosclerosis, preeclampsia, embolism, platelet-associated ischemic disorders including lung ischemia, coronary ischemia, and cerebral ischemia, and for the prevention of reocclusion following thrombosis, thrombotic disorders including coronary artery thrombosis, cerebral artery thrombosis, intracardiac thrombosis, peripheral artery thrombosis, venous thrombosis, thrombotic microangiopathies including thrombotic thrombocytopenic purpura (TTP) and hemolytic uremic syndrome (HUS), essential thrombocythemia, disseminated intravascular coagulation (DIC), and thrombosis and coagulopathies associated with exposure to a foreign or injured tissue surface, in combination with angioplasty, carotid endarterectomy, anastomosis of vascular grafts, and chronic cardiovascular devices such as in-dwelling catheters or shunts. Other instances in which it would be useful to inhibit increased ADP release due to increased platelet stimulation would be in individuals at high risk for thrombus formation or reformation (severe arteriosclerosis), and inhibition of occlusion, reocclusion, stenosis and/or restenosis of blood vessels.
The present invention contemplates the administration of a therapeutically effective amount of the present peptides, with or without pharmaceutically acceptable carriers or diluents. Suitable compositions of this invention include aqueous solutions comprising compounds of this invention and pharmacologically acceptable carriers, e.g., saline, at a pH level, e.g., 7.4. The solutions may be introduced into a patient's bloodstream by local bolus injection.
When a peptide according to this invention is administered into a human subject, the daily dosage will normally be determined by the prescribing physician with the dosage generally varying according to the age, weight, and response of the individual patient, as well as the severity of the patient's symptoms.
In one exemplary application, administration occurs in an amount between about 0.1 mg/kg of body weight to about 60 mg/kg of body weight per day, preferably of between 0.5 mg/kg of body weight to about 40 mg/kg of body weight per day.
The pharmaceutical composition may be administered parenterally, topically, orally or locally. It is preferably given by parenteral, e.g., subcutaneous, intradermal or intramuscular route, preferably by subcutaneous or intradermal route, in order to reach platelets.
The composition according to the invention for parenteral administration is generally in the form of a solution or suspension of the peptide in a pharmaceutically acceptable carrier, preferably an aqueous carrier. Examples of aqueous carriers that may be used include water, buffered water, saline solution (0.4%), glycine solution (0.3%), hyaluronic acid and similar known carriers. Apart from aqueous carriers it is also possible to use solvents such as dimethylsulphoxide, propyleneglycol, dimethylformamide and mixtures thereof. The composition may also contain pharmaceutically acceptable excipients such as buffer substances and inorganic salts in order to achieve normal osmotic pressure and/or effective lyophilization. Examples of such additives are sodium and potassium salts, e.g., chlorides and phosphates, sucrose, glucose, protein hydrolysates, dextran, polyvinylpyrrolidone or polyethylene glycol. The compositions may be sterilized by conventional methods, e.g., by sterile filtration. The composition may be decanted directly in this form or lyophilized and mixed with a sterile solution before use.
The present peptides will be administered in a manner to prevent or minimize platelet aggregation in a subject at risk for thrombosis or other disorders involving unwanted platelet aggregation. A patient in need of the present platelet integrin binding peptides may receive, e.g. an 135-μg/kg-200 μg/kg intravenous loading dose and 0.5 μg/kg per minute by intravenous infusion. It should be administered in an effective dosage to cause competitive inhibition of binding of fibrinogen, Von Willebrand factor and other adhesive ligands, to resting and active GP IIb/IIIa receptors. With recommended dosage, peak plasma drug concentration should occur within about 5 minutes of injection and steady-state drug concentration is 1.5-2.2 μg/ml. The peptides are formulated for intravenous (i.v.) administration in the strength of 0.75 mg/ml and 2 mg/ml (10 ml vial) as a clear, colorless, sterile solution. The vial should be stored under refrigeration between 2-8° C. and protected from light until administration.
Library Synthesis and Screening
The AgRP loop 4 random library was prepared as described in J. Mol. Biol. 385, 1064-1075. Briefly, the AgRP gene was assembled by overlap extension PCR using Taq polymerase in the presence of DMSO. Primers were used that contained NNS degenerate codons in place of loop 4, where N=A, C, T or G and S=C or G, which codes for all 20 amino acids and the TAG stop codon. The assembled gene was PCR amplified using end primers with 40 bp overlap to the pCT vector upstream and downstream of the NheI and BamHI restriction sites for homologous recombination in yeast. Digested pCT vector (0.5-1 μg) and amplified linear gene product (5-10 μg) were electroporated into the S. cerevisiae strain EBY100, yielding a library of ˜2×107 transformants. See Nat. Biotech. 15: 553-557, regarding EBY100.
The yeast library was grown in selective media (SD-CAA, 20 g/L glucose, 6.7 g/L yeast nitrogen base without amino acids, 5.4 g/L Nα2HPO4, 8.6 g/L NaH2PO4.H2O, and 5 g/L Bacto casamino acids) and induced in selective media containing galactose (SG-CAA). For library screening, 5 to 25×106 yeast (depending on the sort round) were labeled with detergent-solubilized integrin and a 1:250 dilution of chicken anti-c-myc IgY (Invitrogen) in integrin binding buffer (IBB, 20 mM Tris pH 7.5 with 1 mM MgCl2, 1 mM MnCl2, 2 mM CaCl2, 100 mM NaCl, and 1 mg/mL bovine serum albumin) at room temperature for 2 h. The yeast cells were centrifuged at 4° C. and the supernatant was removed by aspiration. The yeast were then resuspended in ice-cold BPBS (PBS containing 1 mg/mL bovine serum albumen) containing a 1:100 dilution of phycoerythrin-conjugated goat anti-chicken IgY secondary antibody (Santa Cruz Biotechnology) and a 1:25 dilution of either fluorescein-conjugated anti-αvβ3 antibody (Biolegend) or fluorescein-conjugated anti-αIIb antibody (Biolegend), as appropriate. After incubation for 20 min on ice, the yeast were centrifuged, aspirated, and stored as pellets on ice until cell sorting. Yeast library screening was performed using a Becton Dickinson FACS Vantage SE instrument (Stanford FACS Core Facility) and CellQuest software. In positive sort rounds, 1-2% of yeast cells that best bound to αIIbβ3 integrin were collected using a diagonal sort gate to normalize binding with protein expression levels. For negative sort rounds, the 1-2% of yeast cells that expressed AgRP peptides but showed the weakest binding to αvβ3 integrin were collected. The isolated yeast clones were propagated in SD-CAA media, induced for AgRP expression in SG-CAA media, and subjected to additional rounds of FACS.
Detergent-solubilized αIIbβ3 integrin (human GPIIb/IIIa) was obtained from Enzyme Research as a Triton X-100 formulation. Detergent-solubilized αvβ3 integrin was purchased from Millipore as an octyl-β-D-glucopyranoside formulation. Seven rounds of positive-only sorts were performed using the following concentrations of αIIbβ3 integrin:—round 1: 250 nM, round 2: 250 nM, round 3: 100 nM, round 4: 50 nM, round 5: 50 nM, round 6: 25 nM, round 7: 10 nM. In a separate sorting strategy, nine rounds of varying positive and negative sorts were performed using the following concentrations of integrin: round 1: 250 nM αIIbβ3, round 2: 250 nM αIIbβ3, round 3: 100 nM αvβ3 (negative sort), round 4: 100 nM αvβ3 (negative sort), round 5: 50 nM αIIbβ3, round 6: 100 nM αvβ3 (negative sort), round 7: 100 nM αvβ3 (negative sort), round 8: 25 nM αIIbβ3, round 9: 25 nM αIIbβ3. Plasmid DNA was recovered from yeast clones using a Zymoprep kit (Zymo Research) and transformed into XL-1 blue supercompetent E. coli cells (Stratagene) for plasmid miniprep. DNA sequencing of resulting clones was performed by MCLab (South San Francisco, Calif.).
Recombinant Production of Engineered AgRP Peptides
Recombinant AgRP peptides were prepared using the Multi-Copy Pichia Expression Kit (Invitrogen) following the manufacturer's instructions. The open reading frames of the AgRP peptide genes were cloned into the pPIC9K plasmid, which also contained N-terminal FLAG and C-terminal hexahistidine epitope tags (SEQ ID NO: 136) as handles for integrin binding measurements and peptide purification. Plasmid (5-10 □g) was linearized by digestion with SacI and electroporated into the P. pastoris strain GS115. Yeast cells were then streaked onto YPD plates containing 4 mg/mL geneticin to select for cells containing multiple insertions. Individual colonies were tested for AgRP peptide expression by Western blot after probing the membrane with an antibody against the FLAG epitope tag. Scale-up was performed in 500 mL liquid cultures over 3 days while maintaining the concentration of methanol at ˜0.5% to induce AgRP peptide production. The crude yeast supernatants were adjusted to pH 8.0, filtered, and AgRP peptides were purified by affinity chromatography using Ni-NTA beads. Following concentration and buffer exchange to PBS using Amicon Ultra Centrifugation tubes (3 kDa cutoff, Millipore), the AgRP peptides were further purified by size exclusion chromatography using a Superdex-75 gel filtration column (GE Healthcare). Masses of purified peptides were confirmed by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry and concentrations were determined from absorbance readings at 270 nm and previously determined extinction coefficients. Additionally, the purity of the peptides, as well as the presence or absence of multimers was assayed with SDS-PAGE using 4-12% Bis-Tris gels that were stained with SimplyBlue Safestain (Invitrogen).
Cell Binding Assays
K562 cells stably expressing αIIbβ3, αvβ3, or αvβ5 were obtained from S. Blystone and maintained in IMDM media (Gibco) containing 10% FBS and 1.0 mg/mL geneticin. For binding assays, 105 cells were suspended in IBB, and varying concentrations of AgRP peptides were added. After 3 hour incubation at 4° C., the cells were centrifuged, aspirated, and resuspended in ice-cold BPBS containing a 1:20 dilution of fluorescein-conjugated goat anti-His antibody (Bethyl) for 20 on ice. The cells were then centrifuged, aspirated, and stored as pellets on ice until analysis by flow cytometry using a BD FACS Caliber instrument Becton Dickinson. Binding data was analyzed with FlowJo (Treestar, Inc.) and data was fit to sigmoidal plots using Kaleidagraph (Synergy Software) to calculate equilibrium dissociation constants. Experiments were performed at least 3 times and data are presented as average values with standard deviations.
Platelet Aggregation Assays
Human platelets isolated by pheresis were obtained from the Stanford Blood Center and stored in serum with 7% acid/citrate/dextrose (ACD) solution. Prior to use, the platelets were centrifuged, aspirated, washed once with Tyrode's buffer (140 mM NaCl, 3 mM KCl, 12 mM NaHCO3, 0.4 mM NaH2PO4, 2 mM MgCl2, 0.1% glucose, pH 7.4), and counted using a haemacytometer. Counted platelets were resuspended in fresh plasma to a final concentration of approximately 2.8×108 cells/mL. 180 μL of the platelet solution, or approximately 5×107 platelets, were added to separate wells in a 96-well microtiter plate. AgRP peptide or eptifibatide (Cell Sciences) were added to the wells at varying concentrations, after which the platelets were activated with 25 μM ADP. Absorbance at 365 nm was monitored over 1 hr using a Synergy 4 microplate reader (BioTek Instrumentation) with readings taken every 35 s, with 3 s of shaking between readings. The end points for each concentration of peptide tested were plotted as percentage of maximum inhibition versus log peptide concentration, and IC50 values were calculated by fitting this data to sigmoidal curves. Experiments were performed at least 3 times.
Yeast surface display was used to engineer integrin binding knottin peptides, a technique previously described in the literature in connection with other peptides. It was found that the technique provided ease of performing high-throughput, quantitative library screens using fluorescence-activated cell sorting (FACS). Using degenerate codons, loop 4 AgRP (
After the second round of FACS, the pooled yeast clones were tested for binding to 100 nM αIIbβ3 or αvβ3 integrin; this assay revealed, that our library screening approach was enriching for yeast-displayed AgRP peptides that bound to both αIIbβ3 and αvβ3 integrins. Consequentially, we revised our sorting strategy to take a two-pronged approach. In one set of experiments we sorted only against αIIbβ3, presumably leading to a set of knottin peptides enriched for binding β3 integrin-subunits. In an alternative approach, we interspersed negative sorting against αvβ3 with positive sorting against αIIbβ3 in order to improve the likelihood of isolating knottin peptides specific for αIIbβ3 integrin. We performed a total of 7 rounds of positive-only sort against αIIbβ3 integrin (
Sixteen individual clones each were sequenced from the αIIbβ3-specific and β3-specific populations, revealing 7 and 12 unique sequences, respectively. The loop 4 sequences of the engineered AgRP peptides are shown in Table 3, along with the loop 4 sequences of αvβ3-specific AgRP peptides engineered in our prior study for comparison See J. Mol. Biol. 385, 1064-1075. In
Table 3. Sequences of AgRP loop 4 from selections for αIIbβ3-specific binders or β3-specific binders. Sequences are from clones isolated after the final sort round and ordered based on multiple sequence alignment tool Multalin.
Initially the yeast-displayed AgRP loop 4 library was screened against αIIbβ3 integrin, and after two rounds of flow cytometric sorting for clones with αIIbβ3 integrin binding, the binding were nearly equally enriched as clones with αvβ3 integrin binding. This result was particularly striking considering the high degree of specificity achieved from the library when it was screened solely against αvβ3 integrin. Therefore, to isolate AgRP peptides with high specificity to αIIbβ3 integrin, a strategy of varying positive sorts against αIIbβ3 with negative sorts against αvβ3 integrin was adopted. A separate set of sorts against αIIbβ3 integrin alone was also performed. From these two library screening strategies as well as our previous work, there are engineered AgRP variants that fall into three specificity classes: αvβ3-specific (prior work), αIIbβ3-specific, and β3-specific.
The latter two classes show distinct consensus sequences in the engineered loops (
Selectivity for the αIIbβ3 integrin by eptifibatide, a cyclic heptapeptide that contains a homoarginine-glycine-aspartate sequence, is achieved in part by the ability of the homoarginine residue to reach deep into the β-propeller binding pocket of the αIIb subunit to hydrogen bond with αIIb-Asp224; the corresponding hydrogen bonding partners for the αv subunit are Asp150 and Asp218, which are in a much shallower pocket. Similarly, enhanced αIIbβ3-specificity for antibodies and peptides containing KGD versus RGD sequences may be due to the extension of the longer carbon chain of lysine into the deeper binding pocket of the αIIb subunit. The homoarginine residue in eptifibatide has a carbon chain the same length as lysine and presumably allows deeper penetration into the αIIb subunit binding pocket than normal arginine or lysine. In the future, it would interesting to determine whether replacing arginine with lysine or homoarginine in the engineered AgRP peptides would confer additional affinity to αIIbβ3 integrin. It is also worth noting that there is a Lys residue immediately preceding the RGD sequence in the αIIbβ3-specific AgRP peptides. Though it is possible that this Lys residue could extend deep into the αIIb subunit binding pocket, this might cause steric interference with the neighboring Arg residue. More likely, this Lys is restricted from the shallower αv subunit binding pocket, leading to a positioning of the RGD motif that is unfavorable for binding to αIIbβ3-integrin.
Both the αIIbβ3-specific AgRP peptides engineered in this work and eptifibatide have a Trp residue immediately following their RGD sequence. Computational and experimental studies have shown that the presence of an aromatic residue immediately after the RGD sequence in disulfide-bound cyclic peptides can confer affinity and specificity for αIIbβ3 integrin. However, in contrast to these reports, fibronectin analogs with the consensus sequence R-G-D-W-X-E (SEQ ID NO: 137) were shown to have enhanced affinity and specificity for αvβ3 over αIIbβ3 integrin. These conflicting examples illustrate the importance of the particular scaffold in determining how the binding loop is presented to its receptor. One notable structural difference between αvβ3 over αIIbβ3 integrins is that αv-binding pocket residue Asp218 is replaced in the αIIb subunit by Phe231, which increases the hydrophobicity of the pocket. Another aromatic residue on the surface of the αIIb-binding pocket, Phe160, is positioned to potentially allow π-stacking with aromatic residues immediately following the RGD sequence. These structural differences support the observation that aromatic residues contribute to specificity for αIIbβ3 integrin and may explain how the Trp residue following the RGD motif in eptifibatide and certain AgRP peptides helps confer affinity and specificity for αIIbβ3 integrin.
The β3-specific AgRP peptides described in this work also have consensus residues immediately flanking the RGD motif, namely the sequence G-R-G-D-V (SEQ ID NO: 138), while most of the αvβ3-specific peptides have the sequence G-R-G-D-(N/D) (SEQ ID NO: 139). A flexible residue such as Gly immediately preceding the RGD sequence may be necessary to allow the Arg residue to properly orient in the shallow av-binding pocket. The hydrophobic Val residue immediately following the RGD sequence (along with the C-terminal (K/R)-X-(K/R) consensus) is likely responsible for promoting αIIb subunit binding. In contrast, the more hydrophilic Asn or Asp residue and, the absence of the (K/R)-X-(K/R) consensus sequence, appears to confer AgRP peptides with a preference for αvβ3 integrin.
Clones 2bA, 2bJ, and 2bO (from the αIIbβ3-specific series), and b3A and b3I (from the β3-specific series), were chosen for further study. When displayed on yeast, these knottin peptides exhibited the best binding to detergent-solubilized αIIbβ3 or αvβ3 integrin or were representative of the observed sequence diversity (Table 3 and
To determine the integrin binding affinities and specificities of the engineered AgRP peptides, direct binding assays were performed using K562 leukemia cells that have been transfected to stably express different integrin receptors See J. Cell Biol. 127(4) (1994). Cells were incubated in the presence of varying concentrations of AgRP peptides at 4° C., to minimize receptor internalization, followed by staining with a fluorescein-conjugated anti-hexahistidine antibody and analysis by flow cytometry. Equilibrium dissociation constants (KD values) were determined by plotting AgRP peptide concentration (log scale) versus normalized mean fluorescence intensity, then fitting the data to a sigmoidal curve.
The engineered AgRP peptides were first tested against wild-type K562 cells, which naturally express α5β1 integrin. The peptides did not bind to these cells at concentrations up to 1 μM (data not shown). Next the binding of engineered AgRP peptides to K562 cells transfected to express αIIbβ3, αvβ3, or αvβ5 integrins was tested. The peptides did not bind to the K562-αvβ5 cells at concentrations up to 1 μM. Binding data for K562-αIIbβ3 and K562-αvβ3 cells are presented in
Next, the effects of the engineered AgRP peptides on αIIbβ3-mediated platelet aggregation were studied. Three αIIbβ3-specific peptides (2bA, 2bJ, and 2bO) and two β3-specific peptides (b3A and b3I) were tested again. For comparison, a αvβ3-specific AgRP peptide (peptide 7A) from a prior study that binds αvβ3 integrin with a KD ˜0.8 nM (19) and eptifibatide, an αIIbβ3-specific cyclic peptide (Mpr-Har-Gly-Asp-Trp-Pro-Cys (SEQ ID NO: 140), Mpr=mercaptopropionyl, Har=homoarginine), used clinically for inhibiting platelet aggregation, were also tested. Microtiter plates containing platelets in serum in the presence or absence of AgRP peptides were activated using ADP at a final concentration of 25 μM, and absorbance at 365 nm was monitored over 1 h. Coagulation of platelets triggered by ADP caused a decrease in the turbidity of the sample, which was quantified by measuring the absorbance at 365 nm, whereas for unactivated platelets sample turbidity was unchanged (or increased marginally). The engineered AgRP peptides inhibited platelet aggregation in dose-dependent manner (
1Could not be determined because no binding was observed at the highest concentration tested, 1 μM.
2Not tested
It can be seen below that the wild-type knottin proteins AgRP and AgTX-1 and ATx2 share sequence homology:
A Clustal W alignment of the above three sequences shows significant diversity at the amino acid level, but conserved cystine spacing, and three dimensional topology. The consensus sequence is GCIA-DYG-C-WGGTPCCRGR_C_CS_GTNCEC (SEQ ID NO: 141). BLAST 2 similarity score between AgtX-1 and AGtx2 is 77%; the BLAST 2 similarity score between agtX-1 and agrp is 34%.
Two constructs based on the closely related AgTx sequences shown above were prepared:
Loop 4 from one of our engineered AgRP-based high affinity αvβ3 integrin binders (Silverman, J. Mol. Biol. 2009) is shown below.
This loop 4 was grafted into AgTx 1 and AgTx 2 consensus sequences to yield the following:
These AgTX-1-7A and AgTx2-7A sequences were cloned into the yeast display vector and tested for binding to αvβ3 integrin. Binding was observed at 10 nM αvβ3, the lowest concentration tested, indicating that these have high affinity for αvβ3 integrin and that swapping loops from AgRP to AgTx is a viable strategy.
The above specific description is meant to exemplify and illustrate the invention and should not be seen as limiting the scope of the invention, which is defined by the literal and equivalent scope of the appended claims. Any patents or publications mentioned in this specification are intended to convey details of methods and materials useful in carrying out certain aspects of the invention which may not be explicitly set out but which would be understood by workers in the field. Such patents or publications are hereby incorporated by reference to the same extent as if each was specifically and individually incorporated by reference and contained herein, as needed for the purpose of describing and enabling the method or material referred to.
This invention was made with Government support under Contract PHS NRSA 5T32 CA09302 awarded by the National Institutes of Health. The Government has certain rights in this invention.
| Number | Name | Date | Kind |
|---|---|---|---|
| 4789734 | Pierschbacher | Dec 1988 | A |
| 4792525 | Ruoslahti et al. | Dec 1988 | A |
| 4879237 | Rudslahti et al. | Nov 1989 | A |
| 4988621 | Ruoslahti et al. | Jan 1991 | A |
| 5169930 | Ruoslahti et al. | Dec 1992 | A |
| 5519005 | Lider et al. | May 1996 | A |
| 5536814 | Ruoslahti et al. | Jul 1996 | A |
| 5695997 | Ruoslahti et al. | Dec 1997 | A |
| 5766591 | Brooks et al. | Jun 1998 | A |
| 5827821 | Pierschbacher et al. | Oct 1998 | A |
| 5880092 | Pierschbacher et al. | Mar 1999 | A |
| 5916875 | Ruoslahti et al. | Jun 1999 | A |
| 5981468 | Pierschbacher et al. | Nov 1999 | A |
| 5981478 | Ruoslahti et al. | Nov 1999 | A |
| 5994501 | Ruoslahti et al. | Nov 1999 | A |
| 6020460 | Pierschbacher et al. | Feb 2000 | A |
| 6180610 | Ruoslahti et al. | Jan 2001 | B1 |
| 6265539 | Arlinghaus | Jul 2001 | B1 |
| 6353090 | Pierschbacher et al. | Mar 2002 | B1 |
| 6451976 | Lu et al. | Sep 2002 | B1 |
| 6962974 | Kalluri | Nov 2005 | B1 |
| 20020018780 | Koenig et al. | Feb 2002 | A1 |
| 20040132659 | Markland, Jr. et al. | Jul 2004 | A1 |
| 20040146976 | Wittrup et al. | Jul 2004 | A1 |
| 20050075323 | Day et al. | Apr 2005 | A1 |
| 20050164300 | Artis et al. | Jul 2005 | A1 |
| 20050196427 | Tirrell et al. | Sep 2005 | A1 |
| 20060029544 | Sutcliffe-Goulden et al. | Feb 2006 | A1 |
| 20090257952 | Cochran et al. | Oct 2009 | A1 |
| Number | Date | Country |
|---|---|---|
| 2008045252 | Apr 2008 | WO |
| Entry |
|---|
| Kuwada, 1994, Peptide Chemistry, p. 29-32. |
| Yu, 1993, Biochemistry, 32, p. 13123-13129. |
| Heck, 1994, JACS, 116, 10426-10436. |
| Bures, 1998, Biochemistry, 37, p. 12172-12177. |
| Ana Segura, et al., “Snakin-1, a peptide from potato that is active against plant pathogens,” MPMI, 1999, vol. 12, 16-23. |
| Jean-Christophe Gelly, et al., “The KNOTTIN website and database: a new information system dedicated to the knottin scaffold,” Nucleic Acids Research, 2004, vol. 32, D156-D159. |
| Andreas Christmann, et al., “The cystine knot of a squash-type protease inhibitor as a structural scaffold for Escherichia coli cell surface display of conformationally constrained peptides,” Protein Engineering, 1999, vol. 12, 797-806. |
| FlowJo, Data analysis software for flow cytometry, User Documentation tutorial, Version 3.4, Apr. 2001. |
| Joseph C. McNulty, et al., “High-resolution NMR structure of the chemically-synthesized melanocortin receptor binding domain AGRP(87-132) of the agouti-related protein,” Biochemistry, 2001, vol. 40, 15520-15527. |
| Valerie Copie, et al., “Solution structure and dynamics of linked cell attachment modules of mouse fibronectin containing the RGD and synergy regions: comparison with the human fibronectin crystal structure,” J. Mol. Biol., 1998, vol. 277, 663-682. |
| Norelle L. Daly, et al., “Disulfide folding pathways of cystine knot proteins,” J. Biol. Chem., 2003, vol. 278, 6314-6322. |
| Aarno Hautanen, et al., “Effects of modifications of the RGD sequence and its context on recognition by the fibronectin receptor,” J. Biol. Chem., 1989, vol. 264, 1437-1442. |
| C. S. Elangbam, et al., “Cell adhesion molecules—update,” Vet Pathol, 1997, 34,61-73. |
| Leonore A. Herzenberg, et al., “Intepreting flow cytometry data: a guide for the perplexed,” Nature Immunology, 2006, vol. 7, 681-685. |
| Shuang Liu, “Radiolabeled multimeric cyclic RGD peptides as integrin αvβ3 targeted radiotracers for tumor imaging,” Molecular Pharmaceutics, 2006, vol. 3, 472-487. |
| Arne Skerra, “Engineered protein scaffolds for molecular recognition,” J. Mol. Recognition, 2000, vol. 13, 167-187. |
| Beth Wattam, et al., “Arg-Tyr-Asp (RYD) and Arg-Cys-Asp (RCD) motifs in dendroaspin promote selective inhibition of β1 and β3 integrins,” Biochem. J., 2001, vol. 356, 11-17. |
| Ralf J. Hosse, et al., “A new generation of protein display scaffolds for molecular recognition,” Protein Science, 2006, vol. 15, 14-27. |
| Geoffrey P. Smith, et al., “Small binding proteins selected from a combinatorial repertoire of knottins displayed on phage,” J. Mol. Biol., 1998, vol. 277, 317-332. |
| Robert M. Scarborough, et al., “Characterization of the integrin specificities of disintegrins isolated from american pit viper venoms,” J. Biol. Chem., 1993, vol. 268, 1058-1065. |
| Thomas J. Kunicki, et al., “The exchange of Arg-Gly-Asp (RGD) and Arg-Tyr-Asp (RYD) binding sequences in a recombinant murine Fab fragment specific for the integrin αIIbβ3 does not alter integrin recognition,” J. Biol. Chem., 1995, vol. 270, 16660-16665. |
| Herren Wu, et al., “Stepwise in vitro affinity maturation of vitaxin, an αvβ3-specific humanized mAb,” PNAS, 1998, vol. 95, 6037-6042. |
| Victor M. Garsky, et al., “Chemical synthesis of echistatin, a potent inhibitor of platelet aggregation from Echis carinatus: synthesis and biological activity of selected analogs,” PNAS, 1989, vol. 86, 4022-4026. |
| Sandy Reiss, et al., “Inhibition of platelet aggregation by grafting RGD and KGD sequences on the structural scaffold of small disulfide-rich proteins,” Platelets, 2006, vol. 17, 153-157. |
| Richard H. Kimura, et al., “Engineered knottin peptides: A new class of agents for imaging integrin expression in living subjects,” Cancer Research, 2009, vol. 69, 2435-2442. |
| Alexander Wentzel, et al., “Sequence requirements of the GPNG beta-turn of the Ecballium elaterium trypsin inhibitor II explored by combinatorial library screening,” J. Biol. Chem., 1999, vol. 274, 21037-21043. |
| Erkki Kolvunen, et al., “Phage libraries displaying cyclic peptides with different ring sizes: ligand specificities of the RGD-directed integrins,” Nature Biotechnology, 1995, vol. 13, 265-270. |
| Peter C. Brooks, et al., “Requirement of vascular integrin αvβ3 for angiogenesis,” Science, 1994, vol. 264, 569-571. |
| A. Favel, et al., “Protease inhibitors from Ecballium elaterium seeds,” Int. J. Peptide Protein Res., 1989, vol. 33, 202-208. |
| Timothy Hla, et al., “Human cyclooxygenase-2 cDNA,” PNAS, 1992, vol. 89, 7384-7388. |
| PCT International Search Report for application PCT/US 07/21218 dated May 29, 2008. |
| Final Office Action for U.S. Appl. No. 12/418,376 dated Jan. 27, 2011. |
| Adam P. Silverman, et al., Engineered Cystine-Knot Peptides That Bind αvβ3 Integrin With Antibody-Like Affinities, J. Mol. Biol., 2009, vol. 385(4), 1064-1075. |
| Number | Date | Country | |
|---|---|---|---|
| 20110136740 A1 | Jun 2011 | US |
| Number | Date | Country | |
|---|---|---|---|
| 61272816 | Nov 2009 | US |
