Patent Grant
10662235
The present invention relates to binding molecules engineered with cysteine residues and methods of making and using the same. More particularly, the invention is directed to fibronectin type III (FN3) domain molecules that may bind to EGFR and/or c-Met that are cysteine engineered.
Epidermal growth factor receptor (EGFR or ErbB1 or HER1) is a transmembrane glycoprotein of 170 kDa that is encoded by the c-erbB1 proto-oncogene. EGFR is a member of the human epidermal growth factor receptor (HER) family of receptor tyrosine kinases (RTK) which includes HER2 (ErbB2), HER3 (ErbB3) and HER4 (ErbB4). These RTKs share a homologous structure that consists of a ligand-binding extracellular domain (ECD), a single span transmembrane domain and an intracellular domain that contain catalytic kinase domain and a C-terminal tail. EGFR signaling is initiated by ligand binding followed by induction of conformational change, dimerization and trans-autophosphorylation of the receptor (Ferguson et al., Annu Rev Biophys, 37: 353-73, 2008) which initiates a signal transduction cascade that ultimately affects a wide variety of cellular functions, including cell proliferation and survival. Increases in expression or kinase activity of EGFR have been linked with a range of human cancers, making EGFR an attractive target for therapeutic intervention (Mendelsohn et al., Oncogene 19: 6550-6565, 2000; Grünwald et al., J Natl Cancer Inst 95: 851-67, 2003; Mendelsohn et al., Semin Oncol 33: 369-85, 2006). Furthermore, increases in both the EGFR gene copy number and protein expression have been associated with favorable responses to the EGFR tyrosine kinase inhibitor, IRESSA™ (gefitinib), in non-small cell lung cancer (Hirsch et al., Ann Oncol 18:752-60, 2007).
EGFR therapies include both small molecules and anti-EGFR antibodies, approved for treatment of colorectal cancer, pancreatic cancer, head and neck cancer, and non-small cell lung cancer (NSCLC) (Baselga and Arteaga, J Clin Oncol 23:2445-2459 (20005; Gill et al., J Biol Chem, 259:7755-7760, 1984; Goldstein et al., Clin Cancer Res, 1:131 1-1318; 1995; Prewett et al., Clin Cancer Res, 4:2957-2966, 1998).
Efficacy of anti-EGFR therapies may depend on tumor type and EFGR mutation/amplification status in the tumor, and may result in skin toxicity (De Roock et al., Lancet Oncol 11:753-762, 2010; Linardou et al., Nat Rev Clin Oncol, 6: 352-366, 2009; Li and Perez-Soler, Targ Oncol 4: 107-119, 2009). EGFR tyrosine kinase inhibitors (TKI) are commonly used as 2nd line therapies for non small cell lung cancer (NSCLC), but often stop working within twelve months due to resistance pathways (Riely et al., Clin Cancer Res 12: 839-44, 2006).
c-Met encodes a tyrosine kinase receptor. It was first identified as a proto-oncogene in 1984 after it was found that treatment with a carcinogen resulted in a constitutively active fusion protein TPR-MET (Cooper et al., Nature 311:29-33, 1984). Activation of c-Met through its ligand HGF stimulates a plethora of cell processes including growth, motility, invasion, metastasis, epithelial-mesenchymal transition, angiogenesis/wound healing, and tissue regeneration (Christensen et al., Cancer Lett 225:1-26, 2005; Peters and Adjei, Nat Rev Clin Oncol 9:314-26, 2012). c-Met is synthesized as a single chain protein that is proteolytically cleaved into a 50 kDa alpha- and 140 kDa beta-subunit linked by a disulphide bond (Ma et al., Cancer and Metastasis Reviews, 22: 309-325, 2003). c-Met is structurally similar to other membrane receptors such as Ron and Sea and is comprised of an extracellular ligand-binding domain, a transmembrane domain, and a cytoplasmic domain (containing the tyrosine kinase domain and a C-terminal tail region). The exact stoichiometry of HGF:c-Met binding is unclear, but it is generally believed that two HGF molecules bind to two c-Met molecules leading to receptor dimerization and autophosphorylation at tyrosines 1230, 1234, and 1235 (Stamos et al., The EMBO Journal 23: 2325-2335, 2004). Ligand-independent c-Met autophosphorylation can also occur due to gene amplification, mutation or receptor overexpression.
c-Met is frequently amplified, mutated or over-expressed in many types of cancer including gastric, lung, colon, breast, bladder, head and neck, ovarian, prostate, thyroid, pancreatic, and CNS. Missense mutations typically localized to the kinase domain are commonly found in hereditary papillary renal carcinomas (PRCC) and in 13% of sporadic PRCCs (Schmidt et al., Oncogene 18: 2343-2350, 1999). In contrast, c-Met mutations localized to the semaphorin or juxtamembrane domains of c-Met are frequently found in gastric, head and neck, liver, ovarian, NSCLC and thyroid cancers (Ma et al., Cancer and Metastasis Reviews, 22: 309-325, 2003; Sakakura et al., Chromosomes and Cancer, 1999. 24:299-305). c-Met amplification has been detected in brain, colorectal, gastric, and lung cancers, often correlating with disease progression (Ma et al., Cancer and Metastasis Reviews, 22: 309-325, 2003). Up to 4% and 20% of non-small cell lung cancer (NSCLC) and gastric cancers, respectively, exhibit c-Met amplification (Sakakura et al., Chromosomes and Cancer, 1999. 24:299-305: Sierra and Tsao, Therapeutic Advances in Medical Oncology, 3:S21-35, 2011). Even in the absence of gene amplification, c-Met overexpression is frequently observed in lung cancer (Ichimura et al., Jpn J Cancer Res, 87:1063-9, 1996). Moreover, in clinical samples, nearly half of lung adenocarcinomas exhibited high levels of c-Met and HGF, both of which correlated with enhanced tumor growth rate, metastasis and poor prognosis (Sierra and Tsao, Therapeutic Advances in Medical Oncology, 3:S21-35, 2011; Siegfried et al., Ann Thorac Surg 66: 1915-8, 1998).
Nearly 60% of all tumors that become resistant to EGFR tyrosine kinase inhibitors increase c-Met expression, amplify c-Met, or increase its only known ligand, HGF (Turke et al., Cancer Cell, 17:77-88, 2010), suggesting the existence of a compensatory pathway for EGFR through c-Met. c-Met amplification was first identified in cultured cells that became resistant to gefinitib, an EGFR kinase inhibitor, and exhibited enhanced survival through the Her3 pathway (Engelman et al., Science, 316:1039-43, 2007). This was further validated in clinical samples where nine of 43 patients with acquired resistance to either erlotinib or gefitinib exhibited c-Met amplification, compared to only two of 62 untreated patients. Interestingly, four of the nine treated patients also acquired the EGFR activating mutation, T790M, demonstrating simultaneous resistance pathways (Beat et al., Proc Natl Acad Sci USA, 104:20932-7, 2007).
The individual roles of both EGFR and c-Met in cancer is now well established, making these targets attractive for combination therapy. Both receptors signal through the same survival and anti-apoptotic pathways (ERK and AKT); thus, inhibiting the pair in combination may limit the potential for compensatory pathway activation thereby improving overall efficacy. Combination therapies targeting EGFR and c-Met are tested in clinical trials with Tarceva (erlotinib) in combination with anti-c-Met monovalent antibody for NSCL (Spigel et al., 2011 ASCO Annual Meeting Proceedings 2011, Journal of Clinical Oncology: Chicago, Ill. p. 7505) and Tarceva (erlotinib) in combination with ARQ-197, a small molecule inhibitor of c-Met (Adjei et al., Oncologist, 16:788-99, 2011). Combination therapies or bispecific anti-EGFR/c-Met molecules have been disclosed for example in: Int. Pat. Publ. No. WO2008/127710, U.S. Pat. Publ. No. US2009/0042906, Int. Pat. Publ. No. WO2009/111691, Int. Pat. Publ. No. WO2009/126834, Int. Pat. Publ. No. WO2010/039248, Int. Pat. Publ. No. WO2010/115551.
Current small molecule and large molecule (i.e. antibody) approaches to antagonize EGFR and/or c-Met signaling pathways for therapy may be sub-optimal due to possible lack of specificity with small molecules and therefore potential off-target activity and dose-limiting toxicity encountered with small molecule inhibitors. Typical bivalent antibodies may result in clustering of membrane bound receptors and unwanted activation of the downstream signaling pathways, and monovalent antibodies (half arms) pose significant complexity and cost to the manufacturing process.
Accordingly, the need exists for additional monospecific and bispecific EGFR and/or c-Met inhibitors that also have the additional capability of conjugating cytotoxic drugs thus targeting these potent compounds to the EGFR/c-met-expressing tumor cells, enhancing the anti-tumor activity of these EGFR/c-Met inhibitors. While antibody drug conjugates exist in the art, conventional means of attaching a drug moiety generally leads to a heterogeneous mixture of molecules where the drug moieties are attached at a number of sites on the antibody. For example, cytotoxic drugs have typically been conjugated to antibodies through the often-numerous lysine residues of an antibody, generating a heterogeneous antibody-drug conjugate mixture. Depending on reaction conditions, the heterogeneous mixture typically contains a distribution of antibodies with from 0 to about 8, or more, attached drug moieties. In addition, within each subgroup of conjugates with a particular integer ratio of drug moieties to antibodies, is a potentially heterogeneous mixture where the drug moiety is attached at various sites on the antibody. Analytical and preparative methods are inadequate to separate and characterize the antibody-drug conjugate species molecules within the heterogeneous mixture resulting from a conjugation reaction. Antibodies are large, complex and structurally diverse biomolecules, often with many reactive functional groups. Their reactivities with linker reagents and drug-linker intermediates are dependent on factors, such as pH, concentration, salt concentration, and co-solvents. Furthermore, the multistep conjugation process may be non-reproducible due to difficulties in controlling the reaction conditions and characterizing reactants and intermediates.
Chemical conjugation via cysteines present in antibodies has also been demonstrated. However, engineering in cysteine thiol groups by the mutation of various amino acid residues of a protein to cysteine amino acids is potentially problematic, particularly in the case of unpaired (free Cys) residues or those that are relatively accessible for reaction or oxidation. Unpaired Cys residues on the surface of the protein can pair and oxidize to form intermolecular disulfides, and hence protein dimers or multimers. Disulfide dimer formation renders the new Cys unreactive for conjugation to a drug, ligand, or other label. Furthermore, if the protein oxidatively forms an intramolecular disulfide bond between the newly engineered Cys and an existing Cys residue, both Cys groups are unavailable for active site participation and interactions. In addition, the protein may be rendered inactive or nonspecific, by misfolding or loss of tertiary structure (Zhang et al (2002) Anal. Biochem. 311: 1-9).
Thus, a need exists for a molecule that can undergo homogeneous chemical conjugation and avoid the issues faced by antibody conjugates.
The present invention provides an isolated cysteine engineered fibronectin type III (FN3) domain comprising at least one cysteine substitution at a position selected from the group consisting of residues 6, 8, 10, 11, 14, 15, 16, 20, 30, 34, 38, 40, 41, 45, 47, 48, 53, 54, 59, 60, 62, 64, 70, 88, 89, 90, 91, and 93 of the FN3 domain based on SEQ ID NO: 27, and the equivalent positions in related FN3 domains. A cysteine substitution at a position in the domain or protein comprises a replacement of the existing amino acid residue with a cysteine residue.
The present invention also provides an isolated cysteine engineered fibronectin type III (FN3) domain comprising the amino acid sequence of SEQ ID NO: 27 with at least one cysteine substitution from the amino acid sequence of SEQ ID NO: 27 and specifically binds epidermal growth factor receptor (EGFR) and blocks binding of epidermal growth factor (EGF) to EGFR.
The present invention further provides an isolated cysteine engineered fibronectin type III (FN3) domain comprising the amino acid sequence of SEQ ID NO: 114 with at least one cysteine substitution from the amino acid sequence of SEQ ID NO: 114, and specifically binds hepatocyte growth factor receptor (c-Met) and blocks binding of hepatocyte growth factor (HGF) to c-Met.
The present invention provides novel positions at which cysteine substitutions may be made to generate the cysteine engineered FN3 domains. Said positions include one or more of residues 6, 8, 10, 11, 14, 15, 16, 20, 30, 34, 38, 40, 41, 45, 47, 48, 53, 54, 59, 60, 62, 64, 70, 88, 89, 90, 91, or 93 of SEQ ID NOS: 11-114 and/or 122-137.
An aspect of the invention is a process to prepare the isolated cysteine engineered FN3 domains by mutagenizing a nucleic acid sequence of a parent FN3 domain by replacing one or more amino acid residues with a cysteine residue to encode the cysteine engineered FN3 domain; expressing the cysteine engineered FN3 domain; and isolating the cysteine engineered FN3 domain.
Another aspect of the invention is a chemically-conjugated, isolated cysteine engineered FN3 domain wherein the FN3 domain is covalently attached to a chemical reagent comprising a maleimide moiety.
Another embodiment of the invention is a chemically-conjugated, isolated cysteine engineered FN3 domain that can inhibit the growth of EGFR-overexpressing and/or c-Met-expressing tumor cell lines.
The present application also provides an isolated cysteine engineered bispecific FN3 molecule comprising a first fibronectin type III (FN3) domain and a second FN3 domain, wherein the first and second FN3 domains comprise cysteine substitutions at a position selected from the group consisting of residues 6, 8, 10, 11, 14, 15, 16, 20, 30, 34, 38, 40, 41, 45, 47, 48, 53, 54, 59, 60, 62, 64, 70, 88, 89, 90, 91, and 93, specifically binds epidermal growth factor receptor (EGFR) and blocks binding of epidermal growth factor (EGF) to EGFR, and the second FN3 domain specifically binds hepatocyte growth factor receptor (c-Met), and blocks binding of hepatocyte growth factor (HGF) to c-Met.
Another aspect of the invention is a chemically-conjugated, isolated cysteine engineered bispecific molecule wherein the bispecific molecule is covalently attached to a chemical reagent comprising a maleimide moiety.
A further aspect of the invention is a process to prepare the isolated cysteine engineered bispecific FN3 by mutagenizing a nucleic acid sequence of a parent FN3 bispecific molecule by replacing one or more amino acid residues with cysteine residues to encode the cysteine engineered bispecific molecule; expressing the cysteine engineered molecule; and isolating the cysteine engineered bispecific molecule.
The term “fibronectin type III (FN3) domain” (FN3 domain) as used herein refers to a domain occurring frequently in proteins including fibronectins, tenascin, intracellular cytoskeletal proteins, cytokine receptors and prokaryotic enzymes (Bork and Doolittle, Proc Nat Acad Sci USA 89:8990-8994, 1992; Meinke et al., J Bacteriol 175:1910-1918, 1993; Watanabe et al., J Biol Chem 265:15659-15665, 1990). Exemplary FN3 domains are the 15 different FN3 domains present in human tenascin C, the 15 different FN3 domains present in human fibronectin (FN), and non-natural synthetic FN3 domains as described for example in U.S. Pat. Publ. No. 2010/0216708. Individual FN3 domains are referred to by domain number and protein name, e.g., the 3rd FN3 domain of tenascin (TN3), or the 10th FN3 domain of fibronectin (FN10).
The term “substituting” or “substituted” or “mutating” or “mutated” as used herein refers to altering, deleting of inserting one or more amino acids or nucleotides in a polypeptide or polynucleotide sequence to generate a variant of that sequence.
The term “randomizing” or “randomized” or “diversified” or “diversifying” as used herein refers to making at least one substitution, insertion or deletion in a polynucleotide or polypeptide sequence.
“Variant” as used herein refers to a polypeptide or a polynucleotide that differs from a reference polypeptide or a reference polynucleotide by one or more modifications for example, substitutions, insertions or deletions.
The term “specifically binds” or “specific binding” as used herein refers to the ability of the FN3 domain of the invention to bind to a predetermined antigen with a dissociation constant (KD) of 1×10−6 M or less, for example 1×10−7 M or less, 1×10−8 M or less, 1×10−9 M or less, 1×10−10 M or less, 1×10−11 M or less, 1×10−12 M or less, or 1×10−13 M or less. Typically the FN3 domain of the invention binds to a predetermined antigen (i.e. EGFR or c-Met) with a KD that is at least ten fold less than its KD for a nonspecific antigen (for example BSA or casein) as measured by surface plasmon resonance using for example a Proteon Instrument (BioRad). Thus, a bispecific EGFR/c-Met FN3 domain containing molecule of the invention specifically binds to each EGFR and c-Met with a binding affinity (KD) of at least 1×10−6 M or less for both EGFR and c-Met. The isolated FN3 domain of the invention that specifically binds to a predetermined antigen may, however, have cross-reactivity to other related antigens, for example to the same predetermined antigen from other species (homologs).
The term “library” refers to a collection of variants. The library may be composed of polypeptide or polynucleotide variants.
The term “stability” as used herein refers to the ability of a molecule to maintain a folded state under physiological conditions such that it retains at least one of its normal functional activities, for example, binding to a predetermined antigen such as EGFR or c-Met.
“Epidermal growth factor receptor” or “EGFR” as used here refers to the human EGFR (also known as HER-1 or Erb-B1 (Ullrich et al., Nature 309:418-425, 1984) having the sequence shown in SEQ ID NO: 73 and in GenBank accession number NP_005219, as well as naturally-occurring variants thereof. Such variants include the well known EGFRvIII and other alternatively spliced variants (e.g., as identified by SwissProt Accession numbers P00533-1, P00533-2, P00533-3, P00533-4), variants GLN-98, ARG-266, Lys-521, ILE-674, GLY-962, and PRO-988 (Livingston et al., NIEHS-SNPs, environmental genome project, NIEHS ES15478).
“EGFR ligand” as used herein encompasses all (e.g., physiological) ligands for EGFR, including EGF, TGF-α, heparin binding EGF (HB-EGF), amphiregulin (AR), and epiregulin (EPI).
“Epidermal growth factor” (EGF) as used herein refers to the well known 53 amino acid human EGF having an amino acid sequence shown in SEQ ID NO: 74.
“Hepatocyte growth factor receptor” or “c-Met” as used herein refers to the human c-Met having the amino acid sequence shown in SEQ ID NO: 101 or in GenBank Accession No: NP_001120972 and natural variants thereof.
“Hepatocyte growth factor” (HGF) as used herein refers to the well known human HGF having the amino acid sequence shown in SEQ ID NO: 102 which is cleaved to form a dimer of an alpha and beta chain linked by a disulfide bond.
“Blocks binding” or “inhibits binding”, as used herein interchangeably refers to the ability of the FN3 domains of the invention of the bispecific EGFR/c-Met FN3 domain containing molecule to block or inhibit binding of the EGFR ligand such as EGF to EGFR and/or HGF to c-Met, and encompass both partial and complete blocking/inhibition. The blocking/inhibition of EGFR ligand such as EGF to EGFR and/or HGF to c-Met by the FN3 domain or the bispecific EGFR/c-Met FN3 domain containing molecule of the invention reduces partially or completely the normal level of EGFR signaling and/or c-Met signaling when compared to the EGFR ligand binding to EGFR and/or HGF binding to c-Met without blocking or inhibition. The FN3 domain or the bispecific EGFR/c-Met FN3 domain containing molecule of the invention “blocks binding” of the EGFR ligand such as EGF to EGFR and/or HGF to c-Met when the inhibition is at least 30%, 35%, 40%/0, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%/0, 85%, 90%, 91%, 92%/0, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100%. Inhibition of binding can be measured using well known methods, for example by measuring inhibition of binding of biotinylated EGF on EGFR expressing A431 cells exposed to the FN3 domain or the bispecific EGFR/c-Met FN3 domain containing molecule of the invention using FACS, and using methods described herein, or measuring inhibition of binding of biotinylated HGF on c-Met extracellular domain using well known methods and methods described herein.
The term “EGFR signaling” refers to signal transduction induced by EGFR ligand binding to EGFR resulting in autophosphorylation of at least one tyrosine residue in the EGFR. An exemplary EGFR ligand is EGF.
“Neutralizes EGFR signaling” as used herein refers to the ability of the FN3 domain of the invention to inhibit EGFR signaling induced by EGFR ligand such as EGF by at least 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100%.
The term “c-Met signaling” refers to signal transduction induced by HGF binding to c-Met resulting in autophosphorylation of at least one tyrosine residue in the c-Met. Typically at least one tyrosine residue at positions 1230, 1234 or 1235 is autophosphorylated upon HGF binding.
“Neutralizes c-Met signaling” as used herein refers to the ability of the FN3 domain of the invention to inhibit c-Met signaling induced by HGF by at least 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100%.
“Overexpress, “overexpressed” and “overexpressing” as used herein interchangeably refer to a cancer or malignant cell that has measurably higher levels of EGFR and/or c-Met on the surface compared to a normal cell of the same tissue type. Such overexpression may be caused by gene amplification or by increased transcription or translation. EGFR and/or c-Met expression and overexpression can be measured using well know assays using for example ELISA, immunofluorescence, flow cytometry or radioimmunoassay on live or lysed cells. Alternatively, or additionally, levels of EGFR and/or c-Met-encoding nucleic acid molecules may be measured in the cell for example using fluorescent in situ hybridization, Southern blotting, or PCR techniques. EGFR and/or c-Met is overexpressed when the level of EGFR and/or c-Met on the surface of the cell is at least 1.5-fold higher when compared to the normal cell.
“Tencon” as used herein refers to the synthetic fibronectin type III (FN3) domain having the sequence shown in SEQ ID NO: 1 and described in U.S. Pat. Publ. No. US2010/0216708.
A “cancer cell” or a “tumor cell” as used herein refers to a cancerous, pre-cancerous or transformed cell, either in vivo, ex vivo, and in tissue culture, that has spontaneous or induced phenotypic changes that do not necessarily involve the uptake of new genetic material. Although transformation can arise from infection with a transforming virus and incorporation of new genomic nucleic acid, or uptake of exogenous nucleic acid, it can also arise spontaneously or following exposure to a carcinogen, thereby mutating an endogenous gene. Transformation/cancer is exemplified by, e.g., morphological changes, immortalization of cells, aberrant growth control, foci formation, proliferation, malignancy, tumor specific markers levels, invasiveness, tumor growth or suppression in suitable animal hosts such as nude mice, and the like, in vitro, in vivo, and ex vivo (Freshney, Culture of Animal Cells: A Manual of Basic Technique (3rd ed. 1994)).
The term “vector” means a polynucleotide capable of being duplicated within a biological system or that can be moved between such systems. Vector polynucleotides typically contain elements, such as origins of replication, polyadenylation signal or selection markers that function to facilitate the duplication or maintenance of these polynucleotides in a biological system. Examples of such biological systems may include a cell, virus, animal, plant, and reconstituted biological systems utilizing biological components capable of duplicating a vector. The polynucleotide comprising a vector may be DNA or RNA molecules or a hybrid of these.
The term “expression vector” means a vector that can be utilized in a biological system or in a reconstituted biological system to direct the translation of a polypeptide encoded by a polynucleotide sequence present in the expression vector.
The term “polynucleotide” means a molecule comprising a chain of nucleotides covalently linked by a sugar-phosphate backbone or other equivalent covalent chemistry. Double and single-stranded DNAs and RNAs are typical examples of polynucleotides.
The term “polypeptide” or “protein” means a molecule that comprises at least two amino acid residues linked by a peptide bond to form a polypeptide. Small polypeptides of less than about 50 amino acids may be referred to as “peptides”.
The term “bispecific EGFR/c-Met molecule” or “bispecific EGFR/c-Met FN3 domain containing molecule” as used herein refers to a molecule comprising an EGFR binding FN3 domain and a distinct c-Met binding FN3 domain that are covalently linked together either directly or via a linker. An exemplary bispecific EGFR/c-Met binding molecule comprises a first FN3 domain specifically binding EGFR and a second FN3 domain specifically binding c-Met.
“Valent” as used herein refers to the presence of a specified number of binding sites specific for an antigen in a molecule. As such, the terms “monovalent”, “bivalent”, “tetravalent”, and “hexavalent” refer to the presence of one, two, four and six binding sites, respectively, specific for an antigen in a molecule.
“Mixture” as used herein refers to a sample or preparation of two or more FN3 domains not covalently linked together. A mixture may consist of two or more identical FN3 domains or distinct FN3 domains.
Compositions of Matter
The present invention provides cysteine engineered monospecific and bispecific EGFR and/or c-Met binding FN3 domain containing molecules and methods of making and using them.
Monospecific EGFR Binding Molecules
The present invention provides fibronectin type III (FN3) domains that bind specifically to epidermal growth factor receptor (EGFR) and block binding of epidermal growth factor (EGF) to EGFR, and thus can be widely used in therapeutic and diagnostic applications. The present invention provides polynucleotides encoding the FN3 domains of the invention or complementary nucleic acids thereof, vectors, host cells, and methods of making and using them.
The FN3 domains of the invention bind EGFR with high affinity and inhibit EGFR signaling, and may provide a benefit in terms of specificity and reduced off-target toxicity when compared to small molecule EGFR inhibitors, and improved tissue penetration when compared to conventional antibody therapeutics.
One embodiment of the invention an isolated fibronectin type III (FN3) domain that specifically binds epidermal growth factor receptor (EGFR) and blocks binding of epidermal growth factor (EGF) to EGFR.
The FN3 domains of the invention may block EGF binding to the EGFR with an IC50 value of less than about 1×10−7 M, less than about 1×10−8 M, less than about 1×10−9 M, less than about 1×10−10 M, less than about 1×10−11 M, or less than about 1×10−12 M in a competition assay employing A431 cells and detecting amount of fluorescence from bound biotinylated EGF using streptavidin-phycoerythrin conjugate at 600 nM on A431 cells incubated with or without the FN3 domains of the invention. Exemplary FN3 domains may block EGF binding to the EGFR with an IC50 value between about 1×10−9 M to about 1×10−7 M, such as EGFR binding FN3 domains having the amino acid sequence of SEQ ID NOs: 18-29, 107-110, or 122-137. The FN3 domains of the invention may block EGF binding to the EGFR by at least 30%, 35%, 40%, 45%, 50%, 55%, 60%/0, 65%, 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%/0 or 100% when compared to binding of EGF to the EGFR in the absence of the FN3 domains of the invention using the same assay conditions.
The FN3 domain of the invention may inhibit EGFR signaling by at least 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% when compared to the level of signaling in the absence of FN3 domains of the invention using the same assay conditions.
Binding of a ligand such as EGF to EGFR stimulates receptor dimerization, autophosphorylation, activation of the receptor's internal, cytoplasmic tyrosine kinase domain, and initiation of multiple signal transduction and transactivation pathways involved in regulation of DNA synthesis (gene activation) and cell cycle progression or division. Inhibition of EGFR signaling may result in inhibition in one or more EGFR downstream signaling pathways and therefore neutralizing EGFR may have various effects, including inhibition of cell proliferation and differentiation, angiogenesis, cell motility and metastasis.
EGFR signaling may be measured using various well know methods, for example measuring the autophosphorylation of the receptor at any of the tyrosines Y1068, Y1148, and Y1173 (Downward et al., Nature 311:483-5, 1984) and/or phosphorylation of natural or synthetic substrates. Phosphorylation can be detected using well known methods such as an ELISA assay or a western blot using a phosphotyrosine specific antibody. Exemplary assays can be found in Panek et al., J Pharmacol Exp Thera 283:1433-44, 1997 and Batley et al., Life Sci 62:143-50, 1998.
In one embodiment, the FN3 domain of the invention inhibits EGF-induced EGFR phosphorylation at EGFR residue position Tyrosine 1173 with an IC50 value of less than about 2.5×10−6 M, for example less than about 1×10−6 M, less than about 1×10−7 M, less than about 1×10−8 M, less than about 1×10−9 M, less than about 1×10−10 M, less than about 1×10−11 M, or less than about 1×10−12 M when measured in A431 cells using 50 ng/mL human EGF.
In one embodiment, the FN3 domain of the invention inhibits EGF-induced EGFR phosphorylation at EGFR residue position Tyrosine 1173 with an IC50 value between about 1.8×10−8 M to about 2.5×10−6 M when measured in A431 cells using 50 ng/mL human EGF. Such exemplary FN3 domains are those having the amino acid sequence of SEQ ID NOs: 18-29, 107-110, or 122-137.
In one embodiment, the FN3 domain of the invention binds human EGFR with a dissociation constant (KD) of less than about 1×10−8 M, for example less than about 1×10−9 M, less than about 1×10−10 M, less than about 1×10−11 M, less than about 1×10−12 M, or less than about 1×10−13 M as determined by surface plasmon resonance or the Kinexa method, as practiced by those of skill in the art. In some embodiments, the FN3 domain of the invention binds human EGFR with a KD of between about 2×10−10 to about 1×10−8 M. The affinity of a FN3 domain for EGFR can be determined experimentally using any suitable method. (See, for example, Berzofsky, et al., “Antibody-Antigen Interactions,” In Fundamental Immunology, Paul, W. E., Ed., Raven Press: New York, N.Y. (1984); Kuby, Janis Immunology, W. H. Freeman and Company: New York, N.Y. (1992); and methods described herein). The measured affinity of a particular FN3 domain-antigen interaction can vary if measured under different conditions (e.g., osmolarity, pH). Thus, measurements of affinity and other antigen-binding parameters (e.g., KD, Kon, Koff) are preferably made with standardized solutions of protein scaffold and antigen, and a standardized buffer, such as the buffer described herein.
Exemplary FN3 domains of the invention that bind EGFR include FN3 domains of SEQ ID NOs: 18-29, 107-110, or 122-137.
In one embodiment, the FN3 domain that specifically binds EGFR comprises an amino acid sequence at least 87% identical to the amino acid sequence of SEQ ID NO: 27.
In one embodiment, the FN3 domain that specifically binds EGFR comprises
an FG loop comprising the sequence HNVYKDTNX9RGL (SEQ ID NO: 179) or the sequence LGSYVFEHDVML (SEQ ID NO: 180), wherein X9 is M or I; and
a BC loop comprising the sequence X1X2X3X4X5X6X7X8 (SEQ ID NO: 181).
wherein
The FN3 domains of the invention that specifically bind EGFR and inhibit autophosphorylation of EGFR may comprise as a structural feature an FG loop comprising the sequence HNVYKDTNX9RGL (SEQ ID NO: 179) or the sequence LGSYVFEHDVML (SEQ ID NO: 180), wherein X9 is M or I. Such FN3 domains may further comprise a BC loop of 8 or 9 amino acids in length and defined by the sequence X1X2X3X4X5X6X7X8 (SEQ ID NO: 181), and inhibit EGFR autophosphorylation with an IC50 value of less than about 2.5×10−6 M, and with an IC50 value of between about between about 1.8×10−8 M to about 2.5×10−6 M when measured in A431 cells using 50 ng/mL human EGF.
The FN3 domains of the invention that specifically bind EGFR and inhibit autophosphorylation of EGFR further comprise the sequence of
or the sequence
wherein
The EGFR binding FN3 domains can be generated and tested for their ability to inhibit EGFR autophosphorylation using well known methods and methods described herein.
Another embodiment of the invention is an isolated FN3 domain that specifically binds EGFR, wherein the FN3 domain comprises the sequence shown in SEQ ID NOs: 18-29, 107-110, or 122-137.
In some embodiments, the EGFR binding FN3 domains comprise an initiator methionine (Met) linked to the N-terminus or a cysteine (Cys) linked to a C-terminus of a particular FN3 domain, for example to facilitate expression and/or conjugation of half-life extending molecules.
Another embodiment of the invention is an isolated fibronectin type III (FN3) domain that specifically binds EGFR and blocks binding of EGF to the EGFR, wherein the FN3 domain is isolated from a library designed based on Tencon sequence of SEQ ID NO: 1.
Monospecific c-Met Binding Molecules
The present invention provides fibronectin type III (FN3) domains that bind specifically to hepatocyte growth factor receptor (c-Met) and block binding of hepatocyte growth factor (HGF) to c-Met, and thus can be widely used in therapeutic and diagnostic applications. The present invention provides polynucleotides encoding the FN3 domains of the invention or complementary nucleic acids thereof, vectors, host cells, and methods of making and using them.
The FN3 domains of the invention bind c-Met with high affinity and inhibit c-Met signaling, and may provide a benefit in terms of specificity and reduced off-target toxicity when compared to small molecule c-Met inhibitors, and improved tissue penetration when compared to conventional antibody therapeutics. The FN3 domains of the invention are monovalent, therefore preventing unwanted receptor clustering and activation that may occur with other bivalent molecules.
One embodiment of the invention an isolated fibronectin type III (FN3) domain that specifically binds hepatocyte growth factor receptor (c-Met) and blocks binding of hepatocyte growth factor (HGF) to c-Met.
The FN3 domains of the invention may block HGF binding to c-Met with an IC50 value of about less than about 1×10−7 M, less than about 1×10−8 M, less than about 1×10−9 M, less than about 1×10−10 M, less than about 1×10−11 M, or less than about 1×10−12 M in an assay detecting inhibition of binding of biotinylated HGF to c-Met-Fc fusion protein in the presence of the FN3 domains of the invention. Exemplary FN3 domains my block HGF binding to the c-Met with an IC50 value between about 2×10−10 M to about 6×10−8. The FN3 domains of the invention may block HGF binding to c-Met by at least 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 91%, 92%/0, 93%, 94%, 95%, 96%, 97%, 98%, 99%/0 or 100% when compared to binding of HGF to c-Met in the absence of the FN3 domains of the invention using the same assay conditions.
The FN3 domain of the invention may inhibit c-Met signaling by at least 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% when compared to the level of signaling in the absence of FN3 domains of the invention using the same assay conditions.
Binding of HGF to c-Met stimulates receptor dimerization, autophosphorylation, activation of the receptor's internal, cytoplasmic tyrosine kinase domain, and initiation of multiple signal transduction and transactivation pathways involved in regulation of DNA synthesis (gene activation) and cell cycle progression or division. Inhibition of c-Met signaling may result in inhibition in one or more c-Met downstream signaling pathways and therefore neutralizing c-Met may have various effects, including inhibition of cell proliferation and differentiation, angiogenesis, cell motility and metastasis.
c-Met signaling may be measured using various well know methods, for example measuring the autophosphorylation of the receptor on at least one tyrosine residues Y1230, Y1234 or 11235, and/or phosphorylation of natural or synthetic substrates.
Phosphorylation can be detected, for example, using an antibody specific for phosphotyrosine in an ELISA assay or on a western blot. Some assays for tyrosine kinase activity (Panek et al., J Pharmacol Exp Thera 283:1433-44, 1997; Batley et al., Life Sci 62:143-50, 1998).
In one embodiment, the FN3 domain of the invention inhibits HGF-induced c-Met phosphorylation at c-Met residue position 1349 with an IC50 value of less than about 1×10−6 M, less than about 1×10−7 M, less than about 1×10−8 M, less than about 1×10−9 M, less than about 1×10−10 M, less than about 1×10−11 M, or less than about 1×10−12 M when measured in NCI-H441 cells using 100 ng/mL recombinant human HGF.
In one embodiment, the FN3 domain of the invention inhibits HGF-induced c-Met phosphorylation at c-Met tyrosine Y1349 with an IC50 value between about 4×10−9 M to about 1×10−6 M when measured in NCI-H441 cells using 100 ng/mL recombinant human HGF.
In one embodiment, the FN3 domain of the invention binds human c-Met with an dissociation constant (KD) of equal to or less than about 1×10−7 M, 1×10−8 M, 1×10−9 M, 1×10−10 M, 1×10−11 M, 1×10−12 M, 1×10−13 M, 1×10−14 M, or 1×10−15 M as determined by surface plasmon resonance or the Kinexa method, as practiced by those of skill in the art. I some embodiments, the FN3 domain of the invention binds human c-Met with a KD of between about 3×10−10 to about 5×10−8 M. The affinity of a FN3 domain for c-Met can be determined experimentally using any suitable method. (See, for example, Berzofsky, et al., “Antibody-Antigen Interactions,” In Fundamental Immunology, Paul, W. E., Ed., Raven Press: New York, N.Y. (1984); Kuby, Janis Immunology, W. H. Freeman and Company: New York, N.Y. (1992); and methods described herein). The measured affinity of a particular FN3 domain-antigen interaction can vary if measured under different conditions (e.g., osmolarity, pH). Thus, measurements of affinity and other antigen-binding parameters (e.g., KD, Kon, Koff) are preferably made with standardized solutions of protein scaffold and antigen, and a standardized buffer, such as the buffer described herein.
Exemplary FN3 domains of the invention that bind c-Met include FN3 domains having the amino acid sequence of SEQ ID NOs: 32-49 or 111-114.
In one embodiment, the FN3 domain that specifically binds c-Met comprises an amino acid sequence at least 83% identical to the amino acid sequence of SEQ ID NO: 41.
In one embodiment, the FN3 domain that specifically binds c-Met comprises
The FN3 domains of the invention that specifically bind c-Met and inhibit autophosphorylation of c-Met further comprises the sequence:
wherein
Another embodiment of the invention is an isolated FN3 domain that specifically binds c-Met, wherein the FN3 domain comprises the sequence shown in SEQ ID NOs: 32-49 or 111-114.
Another embodiment of the invention is an isolated fibronectin type III (FN3) domain that specifically binds c-Met and blocks binding of HGF to the c-Met, wherein the FN3 domain is isolated from a library designed based on Tencon sequence of SEQ ID NO: 1.
Isolation of EGFR or c-Met FN3 Domains from a Library Based on Tencon Sequence
Tencon (SEQ ID NO: 1) is a non-naturally occurring fibronectin type III (FN3) domain designed from a consensus sequence of fifteen FN3 domains from human tenascin-C (Jacobs et al., Protein Engineering, Design, and Selection, 25:107-117, 2012; U.S. Pat. Publ. No. 2010/0216708). The crystal structure of Tencon shows six surface-exposed loops that connect seven beta-strands as is characteristic to the FN3 domains, the beta-strands referred to as A, B, C, D, E, F, and G, and the loops referred to as AB, BC, CD, DE, EF, and FG loops (Bork and Doolittle, Proc Natl Acad Sci USA 89:8990-8992, 1992; U.S. Pat. No. 6,673,901). These loops, or selected residues within each loop, can be randomized in order to construct libraries of fibronectin type III (FN3) domains that can be used to select novel molecules that bind EGFR. Table 1 shows positions and sequences of each loop and beta-strand in Tencon (SEQ ID NO: 1).
Library designed based on Tencon sequence may thus have randomized FG loop, or randomized BC and FG loops, such as libraries TCL1 or TCL2 as described below. The Tencon BC loop is 7 amino acids long, thus 1, 2, 3, 4, 5, 6 or 7 amino acids may be randomized in the library diversified at the BC loop and designed based on Tencon sequence. The Tencon FG loop is 7 amino acids long, thus 1, 2, 3, 4, 5, 6 or 7 amino acids may be randomized in the library diversified at the FG loop and designed based on Tencon sequence. Further diversity at loops in the Tencon libraries may be achieved by insertion and/or deletions of residues at loops. For example, the FG and/or BC loops may be extended by 1-22 amino acids, or decreased by 1-3 amino acids. The FG loop in Tencon is 7 amino acids long, whereas the corresponding loop in antibody heavy chains ranges from 4-28 residues. To provide maximum diversity, the FG loop may be diversified in sequence as well as in length to correspond to the antibody CDR3 length range of 4-28 residues. For example, the FG loop can further be diversified in length by extending the loop by additional 1, 2, 3, 4 or 5 amino acids.
Library designed based on Tencon sequence may also have randomized alternative surfaces that form on a side of the FN3 domain and comprise two or more beta strands, and at least one loop. One such alternative surface is formed by amino acids in the C and the F beta-strands and the CD and the FG loops (a C-CD-F-FG surface). A library design based on Tencon alternative C-CD-F-FG surface and is shown in
Library designed based on Tencon sequence also includes libraries designed based on Tencon variants, such as Tencon variants having substitutions at residues positions 11, 14, 17, 37, 46, 73, or 86 (residue numbering corresponding to SEQ ID NO: 1), and which variants display improve thermal stability. Exemplary Tencon variants are described in US Pat. Publ. No. 2011/0274623, and include Tencon27 (SEQ ID NO: 99) having substitutions E11R, L17A, N46V, E86I when compared to Tencon of SEQ ID NO: 1.
Tencon and other FN3 sequence based libraries can be randomized at chosen residue positions using a random or defined set of amino acids. For example, variants in the library having random substitutions can be generated using NNK codons, which encode all 20 naturally occurring amino acids. In other diversification schemes, DVK codons can be used to encode amino acids Ala, Trp, Tyr, Lys, Thr, Asn, Lys, Ser, Arg, Asp, Glu, Gly, and Cys. Alternatively, NNS codons can be used to give rise to all 20 amino acid residues and simultaneously reducing the frequency of stop codons. Libraries of FN3 domains with biased amino acid distribution at positions to be diversified can be synthesized for example using Slonomics® technology (http:_//www_sloning_com). This technology uses a library of pre-made double stranded triplets that act as universal building blocks sufficient for thousands of gene synthesis processes. The triplet library represents all possible sequence combinations necessary to build any desired DNA molecule. The codon designations are according to the well known IUB code.
The FN3 domains specifically binding EGFR or c-Met of the invention can be isolated by producing the FN3 library such as the Tencon library using cis display to ligate DNA fragments encoding the scaffold proteins to a DNA fragment encoding RepA to generate a pool of protein-DNA complexes formed after in vitro translation wherein each protein is stably associated with the DNA that encodes it (U.S. Pat. No. 7,842,476; Odegrip et al., Proc Natl Acad Sci USA 101, 2806-2810, 2004), and assaying the library for specific binding to EGFR and/or c-Met by any method known in the art and described in the Example. Exemplary well known methods which can be used are ELISA, sandwich immunoassays, and competitive and non-competitive assays (see, e.g., Ausubel et al., eds, 1994, Current Protocols in Molecular Biology, Vol. 1, John Wiley & Sons, Inc., New York). The identified FN3 domains specifically binding EGFR or c-Met are further characterized for their ability to block EGFR ligand such as EGF binding to EGFR, or HGF binding to c-Met, and for their ability to inhibit EGFR and/or c-Met signaling using methods described herein.
The FN3 domains specifically binding to EGFR or c-Met of the invention can be generated using any FN3 domain as a template to generate a library and screening the library for molecules specifically binding EGFR or c-Met using methods provided within. Exemplar FN3 domains that can be used are the 3rd FN3 domain of tenascin C (TN3) (SEQ ID NO: 75), Fibcon (SEQ ID NO: 76), and the 10th FN3 domain of fibronectin (FN10) (SEQ ID NO: 77). Standard cloning and expression techniques are used to clone the libraries into a vector or synthesize double stranded cDNA cassettes of the library, to express, or to translate the libraries in vitro. For example ribosome display (Hanes and Pluckthun, Proc Natl Acad Sci USA, 94, 4937-4942, 1997), mRNA display (Roberts and Szostak, Proc Natl Acad Sci USA, 94, 12297-12302, 1997), or other cell-free systems (U.S. Pat. No. 5,643,768) can be used. The libraries of the FN3 domain variants may be expressed as fusion proteins displayed on the surface for example of any suitable bacteriophage. Methods for displaying fusion polypeptides on the surface of a bacteriophage are well known (U.S. Pat. Publ. No. 2011/0118144; Int. Pat. Publ. No. WO2009/085462; U.S. Pat. Nos. 6,969,108; 6,172,197; 5,223,409; 6,582,915; 6,472,147).
The FN3 domains specifically binding EGFR or c-Met of the invention can be modified to improve their properties such as improve thermal stability and reversibility of thermal folding and unfolding. Several methods have been applied to increase the apparent thermal stability of proteins and enzymes, including rational design based on comparison to highly similar thermostable sequences, design of stabilizing disulfide bridges, mutations to increase alpha-helix propensity, engineering of salt bridges, alteration of the surface charge of the protein, directed evolution, and composition of consensus sequences (Lehmann and Wyss, Curr Opin Biotechnol, 12, 371-375, 2001). High thermal stability may increase the yield of the expressed protein, improve solubility or activity, decrease immunogenicity, and minimize the need of a cold chain in manufacturing. Residues that can be substituted to improve thermal stability of Tencon (SEQ ID NO: 1) are residue positions 11, 14, 17, 37, 46, 73, or 86, and are described in US Pat. Publ. No. 2011/0274623. Substitutions corresponding to these residues can be incorporated to the FN3 domains or the bispecific FN3 domain containing molecules of the invention.
Another embodiment of the invention is an isolated FN3 domain that specifically binds EGFR and blocks binding of EGF to EGFR, comprising the sequence shown in SEQ ID NOs: 18-29, 107-110, 122-137, further comprising substitutions at one or more residue positions corresponding to positions 11, 14, 17, 37, 46, 73, and 86 in Tencon (SEQ ID NO: 1).
Another embodiment of the invention is an isolated FN3 domain that specifically binds c-Met and blocks binding of HGF to c-Met, comprising the sequence shown in SEQ ID NOs: 32-49 or 111-114, further comprising substitutions at one or more residue positions corresponding to positions 11, 14, 17, 37, 46, 73, and 86 in Tencon (SEQ ID NO: 1).
Exemplary substitutions are substitutions E11N, E14P, L17A, E37P, N46V, G73Y and E86I (numbering according to SEQ ID NO: 1).
In some embodiments, the FN3 domains of the invention comprise substitutions corresponding to substitutions L17A, N46V, and E86I in Tencon (SEQ ID NO: 1).
The FN3 domains specifically binding EGFR (
Another embodiment of the invention is an isolated FN3 domain that specifically binds EGFR and blocks binding of EGF to EGFR comprising the amino acid sequence shown in SEQ ID NOs: 18-29, 107-110, or 122-137, optionally having substitutions corresponding to substitutions L17A, N46V, and E86I in Tencon (SEQ ID NO: 1).
Another embodiment of the invention is an isolated FN3 domain that specifically binds c-Met and blocks binding of HGF to c-Met comprising the amino acid sequence shown in SEQ ID NOs: 32-49 or 111-114, optionally having substitutions corresponding to substitutions L17A, N46V, and E86I in Tencon (SEQ ID NO: 1).
Measurement of protein stability and protein lability can be viewed as the same or different aspects of protein integrity. Proteins are sensitive or “labile” to denaturation caused by heat, by ultraviolet or ionizing radiation, changes in the ambient osmolarity and pH if in liquid solution, mechanical shear force imposed by small pore-size filtration, ultraviolet radiation, ionizing radiation, such as by gamma irradiation, chemical or heat dehydration, or any other action or force that may cause protein structure disruption. The stability of the molecule can be determined using standard methods. For example, the stability of a molecule can be determined by measuring the thermal melting (“TM”) temperature, the temperature in ° Celsius (° C.) at which half of the molecules become unfolded, using standard methods. Typically, the higher the TM, the more stable the molecule. In addition to heat, the chemical environment also changes the ability of the protein to maintain a particular three dimensional structure.
In one embodiment, the FN3 domains binding EGFR or c-Met of the invention exhibit increased stability by at least 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%/0, 85%, 90%, or 95% or more compared to the same domain prior to engineering measured by the increase in the TM.
Chemical denaturation can likewise be measured by a variety of methods. Chemical denaturants include guanidinium hydrochloride, guanidinium thiocyanate, urea, acetone, organic solvents (DMF, benzene, acetonitrile), salts (ammonium sulfate lithium bromide, lithium chloride, sodium bromide, calcium chloride, sodium chloride); reducing agents (e.g. dithiothreitol, beta-mercaptoethanol, dinitrothiobenzene, and hydrides, such as sodium borohydride), non-ionic and ionic detergents, acids (e.g. hydrochloric acid (HCl), acetic acid (CH3COOH), halogenated acetic acids), hydrophobic molecules (e.g. phosopholipids), and targeted denaturants. Quantitation of the extent of denaturation can rely on loss of a functional property, such as ability to bind a target molecule, or by physiochemical properties, such as tendency to aggregation, exposure of formerly solvent inaccessible residues, or disruption or formation of disulfide bonds.
In one embodiment, the FN3 domains of the invention binding EGFR or c-Met exhibit increased stability by at least 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%/0, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%/0, 85%, 90%, or 95% or more compared to the same scaffold prior to engineering measured by using guanidinium hydrochloride as a chemical denaturant. Increased stability can be measured as a function of decreased tryptophan fluorescence upon treatment with increasing concentrations of guanidine hydrochloride using well known methods.
The FN3 domains of the invention may be generated as monomers, dimers, or multimers, for example, as a means to increase the valency and thus the avidity of target molecule binding, or to generate bi- or multispecific scaffolds simultaneously binding two or more different target molecules. The dimers and multimers may be generated by linking monospecific, bi- or multispecific protein scaffolds, for example, by the inclusion of an amino acid linker, for example a linker containing poly-glycine, glycine and serine, or alanine and proline. Exemplary linker include (GS)2, (SEQ ID NO: 78), (GGGGS)5 (SEQ ID NO: 79), (AP)2 (SEQ ID NO: 80), (AP)5 (SEQ ID NO: 81), (AP)10 (SEQ ID NO: 82), (AP)20 (SEQ ID NO: 83), A(EAAAK)5AAA (SEQ ID NO: 84), linkers. The dimers and multimers may be linked to each other in a N- to C-direction. The use of naturally occurring as well as artificial peptide linkers to connect polypeptides into novel linked fusion polypeptides is well known in the literature (Hallewell et al., J Biol Chem 264, 5260-5268, 1989; Alfthan et al., Protein Eng. 8, 725-731, 1995; Robinson & Sauer, Biochemistry 35, 109-116, 1996; U.S. Pat. No. 5,856,456).
Bispecific EGFR/c/Met Binding Molecules
The bispecific EGFR/c-Met FN3 domain containing molecules of the invention may provide a benefit in terms of specificity and reduced off-target toxicity when compared to small molecule EGFR inhibitors, and improved tissue penetration when compared to conventional antibody therapeutics. The present invention is based at least in part on the surprising finding that the bispecific EGFR/c-Met FN3 domain containing molecules of the invention provide a significantly improved synergistic inhibitory effect when compared to a mixture of EGFR-binding and c-Met-binding FN3 domains. The molecules may be tailored to specific affinity towards both EGFR and c-Met to maximize tumor penetration and retention.
One embodiment of the invention is an isolated bispecific FN3 domain containing molecule comprising a first fibronectin type III (FN3) domain and a second FN3 domain, wherein the first FN3 domain specifically binds epidermal growth factor receptor (EGFR) and blocks binding of epidermal growth factor (EGF) to EGFR, and the second FN3 domain specifically binds hepatocyte growth factor receptor (c-Met), and blocks binding of hepatocyte growth factor (HGF) to c-Met.
The bispecific EGFR/c-Met FN3 domain containing molecules of the invention can be generated by covalently linking any EGFR-binding FN3 domain and any c-Met-binding FN3 domain of the invention directly or via a linker. Therefore, the first FN3 domain of the bispecific molecule may have characteristics as described above for the EGFR-binding FN3 domains, and the second FN3 domain of the bispecific molecule may have characteristics as described above for the c-Met-binding FN3 domains.
In one embodiment, the first FN3 domain of the bispecific EGFR/c-Met FN3 domain containing molecule inhibits EGF-induced EGFR phosphorylation at EGFR residue Tyrosine 1173 with an IC50 value of less than about 2.5×10−8 M when measured in A431 cells using 50 ng/mL human EGF, and the second FN3 domain of the bispecific EGFR/c-Met FN3 domain containing molecule inhibits HGF-induced c-Met phosphorylation at c-Met residue Tyrosine 1349 with an IC50 value of less than about 1.5×10−6 M when measured in NCI-H441 cells using 100 ng/mL human HGF.
In another embodiment, the first FN3 domain of the bispecific EGFR/c-Met FN3 domain containing molecule inhibits EGF-induced EGFR phosphorylation at EGFR residue Tyrosine 1173 with an IC50 value of between about 1.8×10−8 M to about 2.5×10−6 M when measured in A431 cells using 50 ng/mL human EGF, and the second FN3 domain of the bispecific EGFR/c-Met FN3 domain containing molecule inhibits HGF-induced c-Met phosphorylation at c-Met residue Tyrosine 1349 with an IC50 value between about 4×10−9 M to about 1.5×10−6 M when measured in NCI-H441 cells using 100 ng/mL human HGF.
In another embodiment, the first FN3 domain of the bispecific EGFR/c-Met FN3 domain containing molecule binds human EGFR with a dissociation constant (KD) of less than about 1×10−8 M, and the second FN3 domain of the bispecific EGFR/c-Met FN3 domain containing molecule binds human c-Met with a KD of less than about 5×10−8 M.
In the bispecific molecule binding both EGFR and c-Met, the first FN3 domain binds human EGFR with a KD of between about 2×10−10 to about 1×10−8 M, and the second FN3 domain binds human c-Met with a KD of between about 3×10−10 to about 5×10−8 M.
The affinity of the bispecific EGFR/c-Met molecule for EGFR and c-Met can be determined as described above for the monospecific molecules.
The first FN3 domain in the bispecific EGFR/c-Met molecule of the invention may block EGF binding to EGFR with an IC50 value of between about 1×10−9 M to about 1.5×10−7 M in an assay employing A431 cells and detecting amount of fluorescence from bound biotinylated EGF using streptavidin-phycoerythrin conjugate at 600 nM on A431 cells incubated with or without the first FN3 domain. The first FN3 domain in the bispecific EGFR/c-Met molecule of the invention may block EGF binding to the EGFR by at least 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% when compared to binding of EGF to EGFR in the absence of the first FN3 domains using the same assay conditions.
The second FN3 domain in the bispecific EGFR/c-Met molecule of the invention may block HGF binding to c-Met with an IC50 value of between about 2×10−10 M to about 6×10−8 M in an assay detecting inhibition of binding of biotinylated HGF to c-Met-Fc fusion protein in the presence of the second FN3 domain. The second FN3 domain in the bispecific EGFR/c-Met molecule may block HGF binding to c-Met by at least 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 91%, 920/0, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% when compared to binding of HGF to c-Met in the absence of the second FN3 domain using the same assay conditions.
The bispecific EGFR/c-Met molecule of the invention may inhibit EGFR and/or c-Met signaling by at least 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% when compared to the level of signaling in the absence of the bispecific EGFR/c-Met molecule of the invention using the same assay conditions.
EGFR and c-Met signaling may be measured using various well know methods as described above for the monospecific molecules.
The bispecific EGFR/c-Met molecules of the invention comprising the first FN3 domain specifically binding EGFR and the second FN3 domain specifically binding c-Met provide a significantly increased synergistic inhibition of EGFR and c/Met signaling and tumor cell proliferation when compared to the synergistic inhibition observed by a mixture of the first and the second FN3 domain. Synergistic inhibition can be assessed for example by measuring inhibition of ERK phosphorylation by the bispecific EGFR/c-Met FN3 domain containing molecules and by a mixture of two monospecific molecules, one binding EGFR and the other c-Met. The bispecific EGFR/c-Met molecules of the invention may inhibit ERK phosphorylation with an IC50 value at least about 100 fold smaller, for example at least 500, 1000, 5000 or 10,000 fold smaller when compared to the IC50 value for a mixture of two monospecific FN3 domains, indicating at least 100 fold increased potency for the bispecific EGFR/c-Met FN3 domain containing molecules when compared to the mixture of two monospecific FN3 domains. Exemplary bispecific EGFR-c-Met FN3 domain containing molecules may inhibit ERK phosphorylation with and IC50 value of about 5×10−9 M or less. ERK phosphorylation can be measured using standard methods and methods described herein.
The bispecific EGFR/c-Met FN3 domain containing molecule of the invention may inhibit H292 cell proliferation with an IC50 value that is at least 30-fold less when compared to the IC50 value of inhibition of H292 cell growth with a mixture of the first FN3 domain and the second FN3, wherein the cell proliferation is induced with medium containing 10% FBS supplemented with 7.5 ng/mL HGF. The bispecific molecule of the invention may inhibit tumor cell proliferation with an IC50 value that is about 30, 40, 50, 60, 70, 80, 90, 100, 150, 200, 300, 400, 500, 600, 700, 800, or about 1000 fold less when compared to the IC50 value of inhibition of tumor cell proliferation with a mixture of the first FN3 domain and the second FN3 domain. Inhibition of tumor cell proliferation can be measured using standard methods and methods described herein.
Another embodiment of the invention is a bispecific FN3 domain containing molecule comprising a first fibronectin type III (FN3) domain and a second FN3 domain, wherein the first FN3 domain specifically binds epidermal growth factor receptor (EGFR) and blocks binding of epidermal growth factor (EGF) to EGFR, and the second FN3 domain specifically binds hepatocyte growth factor receptor (c-Met), and blocks binding of hepatocyte growth factor (HGF) to c-Met, wherein
the first FN3 domain comprises
the second FN3 domain comprises
In another embodiment, the bispecific molecule comprises the first FN3 domain that binds EGFR comprising the sequence:
or the sequence
wherein in the SEQ ID NOs: X and X;
In another embodiment, the bispecific molecule comprises the second FN3 domain that binds c-Met comprising the sequence
wherein
Exemplary bispecific EGFR/c-Met FN3 domain containing molecules comprise the amino acid sequence shown in SEQ ID NOs: 50-72, 106, 118-121, or 138-165.
The bispecific EGFR/c-Met molecules of the invention comprise certain structural characteristics associated with their functional characteristics, such as inhibition of EGFR autophosphorylation, such as the FG loop of the first FN3 domain that binds EGFR comprising the sequence HNVYKDTNX9RGL (SEQ ID NO: 179) or the sequence LGSYVFEHDVML (SEQ ID NO: 180), wherein X9 is M or I.
In one embodiment, the bispecific EGFR/c-Met FN3 domain containing molecules of the invention
In another embodiment, the bispecific EGFR/c-Met FN3 domain containing molecules of the invention
In one embodiment, bispecific EGFR/c-Met molecules comprise the EGFR-binding FN3 domain comprising the sequence
wherein
the c-Met-binding FN3 domain comprising the sequence
wherein
Exemplary bispecific EGFR/c-Met molecules are those having the sequence shown in SEQ ID NOs: 57, 61, 62, 63, 64, 65, 66, 67 and 68.
The bispecific molecules of the invention may further comprise substitutions at one or more residue positions in the first FN3 domain and/or the second FN3 domain corresponding to positions 11, 14, 17, 37, 46, 73, and 86 in Tencon (SEQ ID NO: 1) as described above, and a substitution at position 29. Exemplary substitutions are substitutions E11N, E14P, L17A, E37P, N46V, G73Y, E86I and D29E (numbering according to SEQ ID NO: 1). Skilled in the art will appreciate that other amino acids can be used for substitutions, such as amino acids within a family of amino acids that are related in their side chains as described infra. The generated variants can be tested for their stability and binding to EGFR and/or c-Met using methods herein.
In one embodiment, the bispecific EGFR/c-Met FN3 domain containing molecule comprises the first FN3 domain that binds specifically EGFR and the second FN3 domain that binds specifically c-Met, wherein the first FN3 domain comprises the sequence:
wherein
X24 is E, N or R;
X25 is E or P;
X26 is L or A;
X27 is H or W;
X28 is E or D;
X29 is E or P;
X30 is N or V;
X31 is G or Y;
X32 is M or I; and
X33 is E or I;
and the second FN3 domain comprises the sequence:
wherein
X34 is E, N or R;
X35 is E or P;
X36 is L or A;
X37 is E or P;
X38 is V or L;
X39 is G or S;
X40 is S or K;
X41 is E or D;
X42 is N or V;
X43 is L or M;
X44 is G or S;
X45 is S or K; and
X46 is E or I.
In other embodiments, the bispecific EGFR/c-Met FN3 domain containing molecule comprises the first FN3 domain comprising an amino acid sequence at least 87% identical to the amino acid sequence of SEQ ID NO: 27, and the second FN3 domain comprising an amino acid sequence at least 83% identical to the amino acid sequence of SEQ ID NO: 41.
The bispecific EGFR/c-Met FN3 domain containing molecules of the invention may be tailored to a specific affinity towards EGFR and c-Met to maximize tumor accumulation.
Another embodiment of the invention is an isolated bispecific FN3 domain containing molecule comprising a first fibronectin type III (FN3) domain and a second FN3 domain, wherein the first FN3 domain specifically binds epidermal growth factor receptor (EGFR) and blocks binding of epidermal growth factor (EGF) to EGFR, and the second FN3 domain specifically binds hepatocyte growth factor receptor (c-Met), and blocks binding of hepatocyte growth factor (HGF) to c-Met, wherein the first FN3 domain and the second FN3 domain is isolated from a library designed based on Tencon sequence of SEQ ID NO: 1.
The bispecific EGFR/c-Met FN3 domain containing molecule of the invention can be generated by covalently coupling the EGFR-binding FN3 domain and the c-Met binding FN3 domain of the invention using well known methods. The FN3 domains may be linked via a linker, for example a linker containing poly-glycine, glycine and serine, or alanine and proline. Exemplary linker include (GS)2, (SEQ ID NO: 78), (GGGGS)5 (SEQ ID NO: 79), (AP)2 (SEQ ID NO: 80), (AP)5 (SEQ ID NO: 81), (AP)10 (SEQ ID NO: 82), (AP)20 (SEQ ID NO: 83), A(EAAAK)5AAA (SEQ ID NO: 84), linkers. The use of naturally occurring as well as artificial peptide linkers to connect polypeptides into novel linked fusion polypeptides is well known in the literature (Hallewell et al., J Biol Chem 264, 5260-5268, 1989; Alfthan et al., Protein Eng. 8, 725-731, 1995; Robinson & Sauer, Biochemistry 35, 109-116, 1996; U.S. Pat. No. 5,856,456). The bispecific EGFR/c-Met molecules of the invention may be linked together from a C-terminus of the first FN3 domain to the N-terminus of the second FN3 domain, or from the C-terminus of the second FN3 domain to the N-terminus of the first FN3 domain. Any EGFR-binding FN3 domain may be covalently linked to a c-Met-binding FN3 domain. Exemplary EGFR-binding FN3 domains are domains having the amino acid sequence shown in SEQ ID NOs: 18-29, 107-110, and 122-137, and exemplary c-Met binding FN3 domains are domains having the amino acid sequence shown in SEQ ID NOs: 32-49 and 111-114. The EGFR-binding FN3 domains to be coupled to a bispecific molecule may additionally comprise an initiator methionine (Met) at their N-terminus.
Variants of the bispecific EGFR/c-Met FN3 domain containing molecules are within the scope of the invention. For example, substitutions can be made in the bispecific EGFR/c-Met FN3 domain containing molecule as long as the resulting variant retains similar selectivity and potency towards EGFR and c-Met when compared to the parent molecule. Exemplary modifications are for example conservative substitutions that will result in variants with similar characteristics to those of the parent molecules. Conservative replacements are those that take place within a family of amino acids that are related in their side chains. Genetically encoded amino acids can be divided into four families: (1) acidic (aspartate, glutamate); (2) basic (lysine, arginine, histidine); (3) nonpolar (alanine, valine, leucine, isoleucine, proline, phenylalanine, methionine, tryptophan); and (4) uncharged polar (glycine, asparagine, glutamine, cysteine, serine, threonine, tyrosine). Phenylalanine, tryptophan, and tyrosine are sometimes classified jointly as aromatic amino acids. Alternatively, the amino acid repertoire can be grouped as (1) acidic (aspartate, glutamate); (2) basic (lysine, arginine histidine), (3) aliphatic (glycine, alanine, valine, leucine, isoleucine, serine, threonine), with serine and threonine optionally be grouped separately as aliphatic-hydroxyl; (4) aromatic (phenylalanine, tyrosine, tryptophan); (5) amide (asparagine, glutamine); and (6) sulfur-containing (cysteine and methionine) (Stryer (ed.), Biochemistry, 2nd ed, WH Freeman and Co., 1981). Non-conservative substitutions can be made to the bispecific EGFR/c-Met FN3 domain containing molecule that involves substitutions of amino acid residues between different classes of amino acids to improve properties of the bispecific molecules. Whether a change in the amino acid sequence of a polypeptide or fragment thereof results in a functional homolog can be readily determined by assessing the ability of the modified polypeptide or fragment to produce a response in a fashion similar to the unmodified polypeptide or fragment using the assays described herein. Peptides, polypeptides or proteins in which more than one replacement has taken place can readily be tested in the same manner.
The bispecific EGFR/c-Met FN3 domain containing molecules of the invention may be generated as dimers or multimers, for example, as a means to increase the valency and thus the avidity of target molecule binding. The multimers may be generated by linking one or more EGFR-binding FN3 domains and one or more c-Met-binding FN3 domain to form molecules comprising at least three individual FN3 domains that are at least bispecific for either EGFR or c-Met, for example by the inclusion of an amino acid linker using well known methods.
Another embodiment of the invention is a bispecific FN3 domain containing molecule comprising a first fibronectin type III (FN3) domain and a second FN3 domain, wherein the first FN3 domain specifically binds epidermal growth factor receptor (EGFR) and blocks binding of epidermal growth factor (EGF) to EGFR, and the second FN3 domain specifically binds hepatocyte growth factor receptor (c-Met), and blocks binding of hepatocyte growth factor (HGF) to c-Met comprising the amino acid sequence shown in SEQ ID NOs: 50-72 or 106.
Half-Life Extending Moieties
The bispecific EGFR/c-Met FN3 domain containing molecules or the monospecific EGFR or c-Met binding FN3 domains of the present invention may incorporate other subunits for example via covalent interaction. In one aspect of the invention, the bispecific EGFR/c-Met FN3 domain containing molecules of the invention further comprise a half-life extending moiety. Exemplary half-life extending moieties are albumin, albumin-binding proteins and/or domains, transferrin and fragments and analogues thereof, and Fc regions. An exemplary albumin-binding domain is shown in SEQ ID NO: 117.
All or a portion of an antibody constant region may be attached to the molecules of the invention to impart antibody-like properties, especially those properties associated with the Fc region, such as Fc effector functions such as C1q binding, complement dependent cytotoxicity (CDC), Fc receptor binding, antibody-dependent cell-mediated cytotoxicity (ADCC), phagocytosis, down regulation of cell surface receptors (e.g., B cell receptor, BCR), and can be further modified by modifying residues in the Fc responsible for these activities (for review; see Strohl, Curr Opin Biotechnol. 20, 685-691, 2009).
Additional moieties may be incorporated into the bispecific molecules of the invention such as polyethylene glycol (PEG) molecules, such as PEG5000 or PEG20,000, fatty acids and fatty acid esters of different chain lengths, for example laurate, myristate, stearate, arachidate, behenate, oleate, arachidonate, octanedioic acid, tetradecanedioic acid, octadecanedioic acid, docosanedioic acid, and the like, polylysine, octane, carbohydrates (dextran, cellulose, oligo- or polysaccharides) for desired properties. These moieties may be direct fusions with the protein scaffold coding sequences and may be generated by standard cloning and expression techniques. Alternatively, well known chemical coupling methods may be used to attach the moieties to recombinantly produced molecules of the invention.
A pegyl moiety may for example be added to the bispecific or monospecific molecules of the invention by incorporating a cysteine residue to the C-terminus of the molecule and attaching a pegyl group to the cysteine using well known methods. Exemplary bispecific molecules with the C-terminal cysteine are those having the amino acid sequence shown in SEQ IN NO: 170-178.
Monospecific and bispecific molecules of the invention incorporating additional moieties may be compared for functionality by several well known assays. For example, altered properties of monospecific and/or bispecific molecules due to incorporation of Fc domains and/or Fc domain variants may be assayed in Fc receptor binding assays using soluble forms of the receptors, such as the FcγRI, FcγRII, FcγRIII or FcRn receptors, or using well known cell-based assays measuring for example ADCC or CDC, or evaluating pharmacokinetic properties of the molecules of the invention in in vivo models.
Polynucleotides, Vectors, Host Cells
The invention provides for nucleic acids encoding the EGFR-binding or c-Met binding FN3 domains or the bispecific EGFR/c-Met FN3 domain containing molecules of the invention as isolated polynucleotides or as portions of expression vectors or as portions of linear DNA sequences, including linear DNA sequences used for in vitro transcription/translation, vectors compatible with prokaryotic, eukaryotic or filamentous phage expression, secretion and/or display of the compositions or directed mutagens thereof. Certain exemplary polynucleotides are disclosed herein, however, other polynucleotides which, given the degeneracy of the genetic code or codon preferences in a given expression system, encode the protein scaffolds and libraries of the protein scaffolds of the invention are also within the scope of the invention.
One embodiment of the invention is an isolated polynucleotide encoding the FN3 domain specifically binding EGFR having the amino acid sequence of SEQ ID NOs: 18-29, 107-110, or 122-137.
One embodiment of the invention is an isolated polynucleotide comprising the polynucleotide sequence of SEQ ID NOs: 97-98 or 168-169.
One embodiment of the invention is an isolated polynucleotide encoding the FN3 domain specifically binding c-Met having the amino acid sequence of the sequence shown in SEQ ID NOs: 32-49 or 111-114.
One embodiment of the invention is an isolated polynucleotide encoding the bispecific EGFR/-c-Met FN3 domain containing molecule having the amino acid sequence of SEQ ID NOs: 50-72, 106, 118-121 or 138-165.
One embodiment of the invention is an isolated polynucleotide comprising the polynucleotide sequence of SEQ ID NOs: 115-116 or 166-167.
The polynucleotides of the invention may be produced by chemical synthesis such as solid phase polynucleotide synthesis on an automated polynucleotide synthesizer and assembled into complete single or double stranded molecules. Alternatively, the polynucleotides of the invention may be produced by other techniques such a PCR followed by routine cloning. Techniques for producing or obtaining polynucleotides of a given known sequence are well known in the art.
The polynucleotides of the invention may comprise at least one non-coding sequence, such as a promoter or enhancer sequence, intron, polyadenylation signal, a cis sequence facilitating RepA binding, and the like. The polynucleotide sequences may also comprise additional sequences encoding additional amino acids that encode for example a marker or a tag sequence such as a histidine tag or an HA tag to facilitate purification or detection of the protein, a signal sequence, a fusion protein partner such as RepA, Fc or bacteriophage coat protein such as pIX or pIII.
Another embodiment of the invention is a vector comprising at least one polynucleotide of the invention. Such vectors may be plasmid vectors, viral vectors, vectors for baculovirus expression, transposon based vectors or any other vector suitable for introduction of the polynucleotides of the invention into a given organism or genetic background by any means. Such vectors may be expression vectors comprising nucleic acid sequence elements that can control, regulate, cause or permit expression of a polypeptide encoded by such a vector. Such elements may comprise transcriptional enhancer binding sites, RNA polymerase initiation sites, ribosome binding sites, and other sites that facilitate the expression of encoded polypeptides in a given expression system. Such expression systems may be cell-based, or cell-free systems well known in the art.
Another embodiment of the invention is a host cell comprising the vector of the invention. A monospecific EGFR-binding or c-Met binding FN3 domain or bispecific EGFR/c-Met FN3 domain containing molecule of the invention can be optionally produced by a cell line, a mixed cell line, an immortalized cell or clonal population of immortalized cells, as well known in the art. See, e.g., Ausubel, et al., ed., Current Protocols in Molecular Biology, John Wiley & Sons, Inc., NY, N.Y. (1987-2001); Sambrook, et al., Molecular Cloning: A Laboratory Manual, 2nd Edition, Cold Spring Harbor, N.Y. (1989); Harlow and Lane, Antibodies, a Laboratory Manual, Cold Spring Harbor, N.Y. (1989); Colligan, et al., eds., Current Protocols in Immunology, John Wiley & Sons, Inc., NY (1994-2001); Colligan et al., Current Protocols in Protein Science, John Wiley & Sons, NY, N.Y., (1997-2001).
The host cell chosen for expression may be of mammalian origin or may be selected from COS-1, COS-7, HEK293, BHK21, CHO, BSC-1, He G2, SP2/0, HeLa, myeloma, lymphoma, yeast, insect or plant cells, or any derivative, immortalized or transformed cell thereof. Alternatively, the host cell may be selected from a species or organism incapable of glycosylating polypeptides, e.g. a prokaryotic cell or organism, such as BL21, BL21(DE3), BL21-GOLD(DE3), XL1-Blue, JM109, HMS174, HMS174(DE3), and any of the natural or engineered E. coli spp, Klebsiella spp., or Pseudomonas spp strains.
Another embodiment of the invention is a method of producing the isolated FN3 domain specifically binding EGFR or c-Met of the invention or the isolated bispecific EGFR/c-Met FN3 domain containing molecule of the invention, comprising culturing the isolated host cell of the invention under conditions such that the isolated FN3 domain specifically binding EGFR or c-Met or the isolated bispecific EGFR-c-Met FN3 domain containing molecule is expressed, and purifying the domain or molecule.
The FN3 domain specifically binding EGFR or c-Met or the isolated bispecific EGFR/c-Met FN3 domain containing molecule of the invention can be purified from recombinant cell cultures by well-known methods, for example by protein A purification, ammonium sulfate or ethanol precipitation, acid extraction, anion or cation exchange chromatography, phosphocellulose chromatography, hydrophobic interaction chromatography, affinity chromatography, hydroxylapatite chromatography and lectin chromatography, or high performance liquid chromatography (HPLC).
Uses of Bispecific EGFR/c-Met FN3 Domain Containing Molecules and EGFR-Binding or c-Met Binding FN3 Domains of the Invention
The bispecific EGFR/c-Met FN3 domain containing molecules, the EGFR binding FN3 domains or the c-Met binding FN3 domains of the invention may be used to diagnose, monitor, modulate, treat, alleviate, help prevent the incidence of, or reduce the symptoms of human disease or specific pathologies in cells, tissues, organs, fluid, or, generally, a host. The methods of the invention may be used to treat an animal patient belonging to any classification. Examples of such animals include mammals such as humans, rodents, dogs, cats and farm animals.
One aspect of the invention is a method for inhibiting growth or proliferation of cells that express EGFR and/or c-Met, comprising contacting the cells with the isolated bispecific EGFR/c-Met FN3 domain containing molecule, the EGFR binding FN3 domain or the c-Met binding FN3 domain of the invention.
Another aspect of the invention is a method for inhibiting growth or metastasis of EGFR and/or c-Met-expressing tumor or cancer cells in a subject comprising administering to the subject an effective amount of the isolated bispecific EGFR/c-Met FN3 domain containing molecule, the EGFR binding FN3 domain or the c-Met binding FN3 domain of the invention so that the growth or metastasis of EGFR- and/or c-Met-expressing tumor or cancer cell is inhibited.
The bispecific EGFR/c-Met FN3 domain containing molecule, the EGFR binding FN3 domain or the c-Met binding FN3 domain of the invention may be used for treatment of any disease or disorder characterized by abnormal activation or production of EGFR, c-Met, EGF or other EGFR ligand or HGF, or disorder related to EGFR or c-Met expression, which may or may not involve malignancy or cancer, where abnormal activation and/or production of EGFR, c-Met, EGF or other EGFR ligand, or HGF is occurring in cells or tissues of a subject having, or predisposed to, the disease or disorder.
The bispecific EGFR/c-Met FN3 domain containing molecule of the invention may be used for treatment of tumors, including cancers and benign tumors. Cancers that are amenable to treatment by the bispecific molecules of the invention include those that overexpress EGFR and/or c-Met. Exemplary cancers that are amenable to treatment by the bispecific molecules of the invention include epithelial cell cancers, breast cancer, ovarian cancer, lung cancer, non-small cell lung cancer (NSCLC), lung adenocarcinoma, colorectal cancer, anal cancer, prostate cancer, kidney cancer, bladder cancer, head and neck cancer, ovarian cancer, pancreatic cancer, skin cancer, oral cancer, esophageal cancer, vaginal cancer, cervical cancer, cancer of the spleen, testicular cancer, gastric cancer, cancer of the thymus, colon cancer, thyroid cancer, liver cancer, or sporadic or hereditary papillary renal carcinoma (PRCC).
The FN3 domains that specifically bind c-Met and block binding of HGF to c-Met of the invention may be for treatment of tumors, including cancers and benign tumors. Cancers that are amenable to treatment by the c-Met binding FN3 domains of the invention include those that overexpress c-Met. Exemplary cancers that are amenable to treatment by the FN3 domains of the invention include epithelial cell cancers, breast cancer, ovarian cancer, lung cancer, colorectal cancer, anal cancer, prostate cancer, kidney cancer, bladder cancer, head and neck cancer, ovarian cancer, pancreatic cancer, skin cancer, oral cancer, esophageal cancer, vaginal cancer, cervical cancer, cancer of the spleen, testicular cancer, and cancer of the thymus.
The FN3 domains that specifically bind EGFR and blocks binding of EGF to the EGFR of the invention may be used for treatment of tumors, including cancers and benign tumors. Cancers that are amenable to treatment by the FN3 domains of the invention include those that overexpress EGFR or variants. Exemplary cancers that are amenable to treatment by the FN3 domains of the invention include epithelial cell cancers, breast cancer, ovarian cancer, lung cancer, colorectal cancer, anal cancer, prostate cancer, kidney cancer, bladder cancer, head and neck cancer, ovarian cancer, pancreatic cancer, skin cancer, oral cancer, esophageal cancer, vaginal cancer, cervical cancer, cancer of the spleen, testicular cancer, and cancer of the thymus.
Administration/Pharmaceutical Compositions
For therapeutic use, the bispecific EGFR/c-Met FN3 domain containing molecules, the EGFR-binding FN3 domains or the c-Met-binding FN3 domains of the invention may be prepared as pharmaceutical compositions containing an effective amount of the domain or molecule as an active ingredient in a pharmaceutically acceptable carrier. The term “carrier” refers to a diluent, adjuvant, excipient, or vehicle with which the active compound is administered. Such vehicles can be liquids, such as water and oils, including those of petroleum, animal, vegetable or synthetic origin, such as peanut oil, soybean oil, mineral oil, sesame oil and the like. For example, 0.4% saline and 0.3% glycine can be used. These solutions are sterile and generally free of particulate matter. They may be sterilized by conventional, well-known sterilization techniques (e.g., filtration). The compositions may contain pharmaceutically acceptable auxiliary substances as required to approximate physiological conditions such as pH adjusting and buffering agents, stabilizing, thickening, lubricating and coloring agents, etc. The concentration of the molecules of the invention in such pharmaceutical formulation can vary widely, i.e., from less than about 0.5%, usually at or at least about 1% to as much as 15 or 20% by weight and will be selected primarily based on required dose, fluid volumes, viscosities, etc., according to the particular mode of administration selected. Suitable vehicles and formulations, inclusive of other human proteins, e.g., human serum albumin, are described, for example, in e.g. Remington: The Science and Practice of Pharmacy, 21st Edition, Troy, D. B. ed., Lipincott Williams and Wilkins, Philadelphia, Pa. 2006, Part 5, Pharmaceutical Manufacturing pp 691-1092, See especially pp. 958-989.
The mode of administration for therapeutic use of the bispecific EGFR/c-Met FN3 domain containing molecules, the EGFR binding FN3 domains or the c-Met binding FN3 domains of the invention may be any suitable route that delivers the agent to the host, such as parenteral administration, e.g., intradermal, intramuscular, intraperitoneal, intravenous or subcutaneous, pulmonary; transmucosal (oral, intranasal, intravaginal, rectal); using a formulation in a tablet, capsule, solution, powder, gel, particle; and contained in a syringe, an implanted device, osmotic pump, cartridge, micropump; or other means appreciated by the skilled artisan, as well known in the art. Site specific administration may be achieved by for example intrarticular, intrabronchial, intraabdominal, intracapsular, intracartilaginous, intracavitary, intracelial, intracerebellar, intracerebroventricular, intracolic, intracervical, intragastric, intrahepatic, intracardial, intraosteal, intrapelvic, intrapericardiac, intraperitoneal, intrapleural, intraprostatic, intrapulmonary, intrarectal, intrarenal, intraretinal, intraspinal, intrasynovial, intrathoracic, intrauterine, intravascular, intravesical, intralesional, vaginal, rectal, buccal, sublingual, intranasal, or transdermal delivery.
Thus, a pharmaceutical composition of the invention for intramuscular injection could be prepared to contain 1 ml sterile buffered water, and between about 1 ng to about 100 mg, e.g. about 50 ng to about 30 mg or more preferably, about 5 mg to about 25 mg, of the FN3 domain of the invention. Similarly, a pharmaceutical composition of the invention for intravenous infusion could be made up to contain about 250 ml of sterile Ringer's solution, and about 1 mg to about 30 mg, e.g. about 5 mg to about 25 mg of the bispecific EGFR/c-Met FN3 domain containing molecule, the EGFR binding FN3 domain or the c-Met binding FN3 domain of the invention. Actual methods for preparing parenterally administrable compositions are well known and are described in more detail in, for example, “Remington's Pharmaceutical Science”, 15th ed., Mack Publishing Company, Easton, Pa.
The bispecific EGFR/c-Met FN3 domain containing molecules, the EGFR-binding FN3 domains or the c-Met-binding FN3 domains of the invention can be lyophilized for storage and reconstituted in a suitable carrier prior to use. This technique has been shown to be effective with conventional protein preparations and art-known lyophilization and reconstitution techniques can be employed.
The bispecific EGFR/c-Met FN3 domain containing molecules, the EGFR-binding FN3 domains or the c-Met-binding FN3 domains may be administered to a subject in a single dose or the administration may be repeated, e.g. after one day, two days, three days, five days, six days, one week, two weeks, three weeks, one month, five weeks, six weeks, seven weeks, two months or three months. The repeated administration can be at the same dose or at a different dose. The administration can be repeated once, twice, three times, four times, five times, six times, seven times, eight times, nine times, ten times, or more.
The bispecific EGFR/c-Met FN3 domain containing molecules, the EGFR-binding FN3 domains or the c-Met-binding FN3 domains may be administered in combination with a second therapeutic agent simultaneously, sequentially or separately. The second therapeutic agent may be a chemotherapeutic agent, an anti-angiogenic agent, or a cytotoxic drug. When used for treating cancer, the bispecific EGFR/c-Met FN3 domain containing molecules, the EGFR-binding FN3 domains or the c-Met-binding FN3 domains may be used in combination with conventional cancer therapies, such as surgery, radiotherapy, chemotherapy or combinations thereof. Exemplary agents that can be used in combination with the FN3 domains of the invention are antagonists of HER2, HER3, HER4, VEGF, and protein tyrosine kinase inhibitors such as Iressa® (gefitinib) and Tarceva (erlotinib).
While having described the invention in general terms, the embodiments of the invention will be further disclosed in the following examples that should not be construed as limiting the scope of the claims.
Tencon (SEQ ID NO: 1) is an immunoglobulin-like scaffold, fibronectin type III (FN3) domain, designed from a consensus sequence of fifteen FN3 domains from human tenascin-C (Jacobs et al., Protein Engineering, Design, and Selection, 25:107-117, 2012; U.S. Pat. Publ. No. 2010/0216708). The crystal structure of Tencon shows six surface-exposed loops that connect seven beta-strands. These loops, or selected residues within each loop, can be randomized in order to construct libraries of fibronectin type III (FN3) domains that can be used to select novel molecules that bind to specific targets.
Construction of TCL1 Library
A library designed to randomize only the FG loop of Tencon (SEQ ID NO: 1), TCL1, was constructed for use with the cis-display system (Jacobs et al., Protein Engineering, Design, and Selection, 25:107-117, 2012). In this system, a single-strand DNA incorporating sequences for a Tac promoter, Tencon library coding sequence, RepA coding sequence, cis-element, and ori element is produced. Upon expression in an in vitro transcription/translation system, a complex is produced of the Tencon-RepA fusion protein bound in cis to the DNA from which it is encoded. Complexes that bind to a target molecule are then isolated and amplified by polymerase chain reaction (PCR), as described below.
Construction of the TCL1 library for use with cis-display was achieved by successive rounds of PCR to produce the final linear, double-stranded DNA molecules in two halves; the 5′ fragment contains the promoter and Tencon sequences, while the 3′ fragment contains the repA gene and the cis- and ori elements. These two halves are combined by restriction digest in order to produce the entire construct. The TCL1 library was designed to incorporate random amino acids only in the FG loop of Tencon, KGGHRSN (SEQ ID NO: 86). NNS codons were used in the construction of this library, resulting in the possible incorporation of all 20 amino acids and one STOP codon into the FG loop. The TCL1 library contains six separate sub-libraries, each having a different randomized FG loop length, from 7 to 12 residues, in order to further increase diversity. Design of tencon-based libraries are shown in Table 2.
To construct the TCL1 library, successive rounds of PCR were performed to append the Tac promoter, build degeneracy into the F:G loop, and add necessary restriction sites for final assembly. First, a DNA sequence containing the promoter sequence and Tencon sequence 5′ of the FG loop was generated by PCR in two steps. DNA corresponding to the full Tencon gene sequence was used as a PCR template with primers POP2220 (SEQID NO: 2) and TC5′ to FG (SEQID NO: 3). The resulting PCR product from this reaction was used as a template for the next round of PCR amplification with primers 130mer (SEQID NO: 4) and Tc5′ to FG to complete the appending of the 5′ and promoter sequences to Tencon. Next, diversity was introduced into the F:G loop by amplifying the DNA product produced in the first step with forward primer POP2222 (SEQID NO: 5), and reverse primers TCF7 (SEQID NO: 6), TCF8 (SEQID NO: 7), TCF9 (SEQID NO: 8), TCF10 (SEQID NO: 9), TCF11 (SEQID N NO: 10), or TCF12 (SEQID NO: 11), which contain degenerate nucleotides. At least eight 100 μL PCR reactions were performed for each sub-library to minimize PCR cycles and maximize the diversity of the library. At least 5×g of this PCR product were gel-purified and used in a subsequent PCR step, with primers POP2222 (SEQ ID NO: 5) and POP2234 (SEQID NO: 12), resulting in the attachment of a 6×His tag and NotI restriction site to the 3′ end of the Tencon sequence. This PCR reaction was carried out using only fifteen PCR cycles and at least 500 ng of template DNA. The resulting PCR product was gel-purified, digested with NotI restriction enzyme, and purified by Qiagen column.
The 3′ fragment of the library is a constant DNA sequence containing elements for display, including a PspOMI restriction site, the coding region of the repA gene, and the cis- and ori elements. PCR reactions were performed using a plasmid (pCR4Blunt) (Invitrogen) containing this DNA fragment with M13 Forward and M13 Reverse primers. The resulting PCR products were digested by PspOMI overnight and gel-purified. To ligate the 5′ portion of library DNA to the 3′ DNA containing the repA gene, 2 pmol of 5′ DNA were ligated to an equal molar amount of 3′ repA DNA in the presence of NotI and PspOMI enzymes and T4 ligase. After overnight ligation at 37° C., a small portion of the ligated DNA was run on a gel to check ligation efficiency. The ligated library product was split into twelve PCR amplifications and a 12-cycle PCR reaction was run with primer pair POP2250 (SEQID NO: 13) and DidLigRev (SEQID NO: 14). The DNA yield for each sub-library of TCL1 library ranged from 32-34 μg.
To assess the quality of the library, a small portion of the working library was amplified with primers Tcon5new2 (SEQID NO: 15) and Tcon6 (SEQID NO: 16), and was cloned into a modified pET vector via ligase-independent cloning. The plasmid DNA was transformed into BL21-GOLD (DE3) competent cells (Stratagene) and 96 randomly picked colonies were sequenced using a T7 promoter primer. No duplicate sequences were found. Overall, approximately 70-85% of clones had a complete promoter and Tencon coding sequence without frame-shift mutation. The functional sequence rate, which excludes clones with STOP codons, was between 59% and 80%.
Construction of TCL2 Library
TCL2 library was constructed in which both the BC and FG loops of Tencon were randomized and the distribution of amino acids at each position was strictly controlled. Table 3 shows the amino acid distribution at desired loop positions in the TCL2 library. The designed amino acid distribution had two aims. First, the library was biased toward residues that were predicted to be structurally important for Tencon folding and stability based on analysis of the Tencon crystal structure and/or from homology modeling. For example, position 29 was fixed to be only a subset of hydrophobic amino acids, as this residue was buried in the hydrophobic core of the Tencon fold. A second layer of design included biasing the amino acid distribution toward that of residues preferentially found in the heavy chain HCDR3 of antibodies, to efficiently produce high-affinity binders (Birtalan et al., J Mol Biol 377:1518-28, 2008; Olson et al., Protein Sci 16:476-84, 2007). Towards this goal, the “designed distribution” of Table 3 refers to the distribution as follows: 6% alanine, 6% arginine, 3.9% asparagine, 7.5% aspartic acid, 2.5% glutamic acid, 1.5% glutamine, 15% glycine, 2.3% histidine, 2.5% isoleucine, 5% leucine, 1.5% lysine, 2.5% phenylalanine, 4% proline, 10% serine, 4.5% threonine, 4% tryptophan, 17.3% tyrosine, and 4% valine. This distribution is devoid of methionine, cysteine, and STOP codons.
The 5′ fragment of the TCL2 library contained the promoter and the coding region of Tencon (SEQ ID NO: 1), which was chemically synthesized as a library pool (Sloning Biotechnology). This pool of DNA contained at least 1×1011 different members. At the end of the fragment, a BsaI restriction site was included in the design for ligation to RepA.
The 3′ fragment of the library was a constant DNA sequence containing elements for display including a 6×His tag, the coding region of the repA gene, and the cis-element. The DNA was prepared by PCR reaction using an existing DNA template (above), and primers LS1008 (SEQID NO: 17) and DidLigRev (SEQID NO: 14). To assemble the complete TCL2 library, a total of 1 μg of BsaI-digested 5′ Tencon library DNA was ligated to 3.5 μg of the 3′ fragment that was prepared by restriction digestion with the same enzyme. After overnight ligation, the DNA was purified by Qiagen column and the DNA was quantified by measuring absorbance at 260 nm. The ligated library product was amplified by a 12-cycle PCR reaction with primer pair POP2250 (SEQID NO: 13) and DidLigRev (SEQID NO: 14). A total of 72 reactions were performed, each containing 50 ng of ligated DNA products as a template. The total yield of TCL2 working library DNA was about 100 μg. A small portion of the working library was sub-cloned and sequenced, as described above for library TCL1. No duplicate sequences were found. About 80% of the sequences contained complete promoter and Tencon coding sequences with no frame-shift mutations.
Construction of TCL14 Library
The top (BC, DE, and FG) and the bottom (AB, CD, and EF) loops, e.g., the reported binding surfaces in the FN3 domains are separated by the beta-strands that form the center of the FN3 structure. Alternative surfaces residing on the two “sides” of the FN3 domains having different shapes than the surfaces formed by loops only are formed at one side of the FN3 domain by two anti-parallel beta-strands, the C and the F beta-strands, and the CD and FG loops, and is herein called the C-CD-F-FG surface.
A library randomizing an alternative surface of Tencon was generated by randomizing select surface exposed residues of the C and F strands, as well as portions of the CD and FG loops as shown in
Library Screening
Cis-display was used to select EGFR binding domains from the TCL1 and TCL2 libraries. A recombinant human extracellular domain of EGFR fused to an IgG1 Fc (R&D Systems) was biotinylated using standard methods and used for panning (residues 25-645 of full length EGFR of SEQ ID NO: 73). For in vitro transcription and translation (ITT), 2-6 μg of library DNA were incubated with 0.1 mM complete amino acids, 1×S30 premix components, and 30 μL of S30 extract (Promega) in a total volume of 100 μL and incubated at 30° C. After 1 hour, 450 μL of blocking solution (PBS pH 7.4, supplemented with 2% bovine serum albumin, 100 μg/mL herring sperm DNA, and 1 mg/mL heparin) were added and the reaction was incubated on ice for 15 minutes. EGFR-Fc:EGF complexes were assembled at molar ratios of 1:1 and 10:1 EGFR to EGF by mixing recombinant human EGF (R&D Systems) with biotinylated recombinant EGFR-Fc in blocking solution for 1 hour at room temperature. For binding, 500 μL of blocked ITT reactions were mixed with 100 μL of EGFR-Fc:EGF complexes and incubated for 1 hour at room temperature, after which bound complexes were pulled down with magnetic neutravidin or streptavidin beads (Seradyne). Unbound library members were removed by successive washes with PBST and PBS. After washing, DNA was eluted from the bound complexes by heating to 65° C. for 10 minutes, amplified by PCR, and attached to a DNA fragment encoding RepA by restriction digestion and ligation for further rounds of panning. High affinity binders were isolated by successively lowering the concentration of target EGFR-Fc during each round from 200 nM to 50 nM and increasing the washing stringency. In rounds 4 and 5, unbound and weakly bound FN3 domains were removed by washing in the presence of a 10-fold molar excess of non-biotinylated EGFR-Fc overnight in PBS.
Following panning, selected FN3 domains were amplified by PCR using oligos Tcon5new2 (SEQID NO: 15) and Tcon6 (SEQID NO: 16), subcloned into a pET vector modified to include a ligase independent cloning site, and transformed into BL21-GOLD (DE3) (Stratagene) cells for soluble expression in E. coli using standard molecular biology techniques. A gene sequence encoding a C-terminal poly-histidine tag was added to each FN3 domain to enable purification and detection. Cultures were grown to an optical density of 0.6-0.8 in 2YT medium supplemented with 100 μg/mL carbenicillin in 1-mL 96-well blocks at 37° C. before the addition of IPTG to 1 mM, at which point the temperature was reduced to 30° C. Cells were harvested approximately 16 hours later by centrifugation and frozen at −20° C. Cell lysis was achieved by incubating each pellet in 0.6 mL of BugBuster® HT lysis buffer (Novagen EMD Biosciences) with shaking at room temperature for 45 minutes.
Selection of FN3 Domains that Bind EGFR on Cells
To assess the ability of different FN3 domains to bind EGFR in a more physiological context, their ability to bind A431 cells was measured. A431 cells (American Type Culture Collection, cat. # CRL-1555) over-express EGFR with ˜2×106 receptors per cell. Cells were plated at 5,000/well in opaque black 96-well plates and allowed to attach overnight at 37° C., in a humidified 5% CO2 atmosphere. FN3 domain-expressing bacterial lysates were diluted 1,000-fold into FACS stain buffer (Becton Dickinson) and incubated for 1 hour at room temperature in triplicate plates. Lysates were removed and cells were washed 3 times with 150 μL/well of FACS stain buffer. Cells were incubated with 50 μL/well of anti-penta His-Alexa488 antibody conjugate (Qiagen) diluted 1:100 in FACS stain buffer for 20 minutes at room temperature. Cells were washed 3 times with 150 μL/well of FACS stain buffer, after which wells were filled with 100 μL of FACS stain buffer and read for fluorescence at 488 nm using an Acumen eX3 reader. Bacterial lysates containing FN3 domains were screened for their ability to bind A431 cells (1320 crude bacterial lysates for TCL1 and TCL2 libraries) and 516 positive clones were identified, where binding was ≥10-fold over the background signal. 300 lysates from the TCL14 library were screened for binding, resulting in 58 unique FN3 domain sequences that bind to EGFR.
Selection of FN3 Domains that Inhibit EGF Binding to EGFR on Cells
To better characterize the mechanism of EGFR binding, the ability of various identified FN3 domain clones to bind EGFR in an EGF-competitive manner was measured using A431 cells and run in parallel with the A431 binding assay screen. A431 cells were plated at 5,000/well in opaque black 96-well plates and allowed to attach overnight at 37° C., in a humidified 5% CO2 atmosphere. Cells were incubated with 50 μL/well of 1:1,000 diluted bacterial lysate for 1 hour at room temperature in triplicate plates. Biotinylated EGF (Invitrogen, cat. # E-3477) was added to each well to give a final concentration of 30 ng/mL and incubated for 10 minutes at room temperature. Cells were washed 3 times with 150 μL/well of FACS stain buffer. Cells were incubated with 50 μL/well of streptavidin-phycoerythrin conjugate (Invitrogen) diluted 1:100 in FACS stain buffer for 20 minutes at room temperature. Cells were washed 3 times with 150 μL/well of FACS stain buffer, after which wells were filled with 100 μL of FACS stain buffer and read for fluorescence at 600 nm using an Acumen eX3 reader.
Bacterial lysates containing the FN3 domains were screened in the EGF competition assay described above. 1320 crude bacterial lysates from TCL1 and TCL2 libraries were screened resulting in 451 positive clones that inhibited EGF binding by >50%.
Expression and Purification of Identified FN3 Domains Binding EGFR
His-tagged FN3 domains were purified from clarified E. coli lysates with His MultiTrap™ HP plates (GE Healthcare) and eluted in buffer containing 20 mM sodium phosphate, 500 mM sodium chloride, and 250 mM imidazole at pH 7.4. Purified samples were exchanged into PBS pH 7.4 for analysis using PD MultiTrap™ G-25 plates (GE Healthcare).
Size Exclusion Chromatography Analysis
Size exclusion chromatography was used to determine the aggregation state of the FN3 domains binding EGFR. Aliquots (10 μL) of each purified FN3 domain were injected onto a Superdex 75 5/150 column (GE Healthcare) at a flow rate of 0.3 mL/min in a mobile phase of PBS pH 7.4. Elution from the column was monitored by absorbance at 280 nm. Centyrins that exhibited high levels of aggregation by SEC were excluded from further analysis.
Off-Rate of Selected EGFR-Binding FN3 Domains from EGFR-Fc
Select EGFR-binding FN3 domains were screened to identify those with slow off-rates (koff) in binding to EGFR-Fc on a ProteOn XPR-36 instrument (Bio-Rad) to facilitate selection of high affinity binders. Goat anti-human Fc IgG (R&D systems), at a concentration of 5 μg/mL, was directly immobilized via amine coupling (at pH 5.0) on all 6 ligand channels in horizontal orientation on the chip with a flow rate of 30 μL/min in PBS containing 0.005% Tween-20. The immobilization densities averaged about 1500 Response Units (RU) with less than 5% variation among different channels. EGFR-Fc was captured on the anti-human Fc IgG surface to a density around 600 RU in vertical ligand orientation. All tested FN3 domains were normalized to a concentration of 1 μM and tested for their binding in horizontal orientation. All 6 analyte channels were used for the FN3 domains to maximize screening throughput. The dissociation phase was monitored for 10 minutes at a flow rate of 100 μL/min. The inter-spot binding signals were used as references to monitor non-specific binding between analytes and the immobilized IgG surface, and were subtracted from all binding responses. The processed binding data were locally fit to a 1:1 simple Langmuir binding model to extract the koff for each FN3 domain binding to captured EGFR-Fc.
Inhibition of EGF-Stimulated EGFR Phosphorylation
Purified EGFR-binding FN3 domains were tested for their ability to inhibit EGF-stimulated phosphorylation of EGFR in A431 cells at a single concentration. EGFR phosphorylation was monitored using the EGFR phospho(Tyr1173) kit (Meso Scale Discovery). Cells were plated at 20,000/well in clear 96-well tissue culture-treated plates (Nunc) in 100 μL/well of RPMI medium (Gibco) containing GlutaMAX™ with 10% fetal bovine serum (FBS) (Gibco) and allowed to attach overnight at 37° C. in a humidified 5% CO2 atmosphere. Culture medium was removed completely and cells were starved overnight in 100 μL/well of medium containing no FBS at 37° C. in a humidified 5% CO2 atmosphere. Cells were then treated with 100 μL/well of pre-warmed (37° C.) starvation medium containing EGFR-binding FN3 domains at a concentration of 2 μM for 1 hour at 37° C. in a humidified 5% CO2 atmosphere. Controls were treated with starvation medium only. Cells were stimulated by the addition and gentle mixing of 100 μL/well of pre-warmed (37° C.) starvation medium containing 100 ng/mL recombinant human EGF (R&D Systems, cat. #236-EG), for final concentrations of 50 ng/mL EGF and 1 μM EGFR-binding FN3 domain, and incubation at 37° C., 5% CO2 for 15 minutes. One set of control wells was left un-stimulated as negative controls. Medium was completely removed and cells were lysed with 100 μL/well of Complete Lysis Buffer (Meso Scale Discovery) for 10 minutes at room temperature with shaking, as per the manufacturer's instructions. Assay plates configured for measuring EGFR phosphorylated on tyrosine 1173 (Meso Scale Discovery) were blocked with the provided blocking solution as per the manufacturer's instructions at room temperature for 1.5-2 hours. Plates were then washed 4 times with 200 μL/well of 1× Tris Wash Buffer (Meso Scale Discovery). Aliquots of cell lysate (30 μL/well) were transferred to assay plates, which were covered with plate sealing film (VWR) and incubated at room temperature with shaking for 1 hour. Assay plates were washed 4 times with 200 μL/well of Tris Wash Buffer, after which 25 μL of ice-cold Detection Antibody Solution (Meso Scale Discovery) were added to each well, being careful not to introduce bubbles. Plates were incubated at room temperature with shaking for 1 hour, followed by 4 washes with 200 μL/well of Tris Wash Buffer. Signals were detected by addition of 150 μL/well of Read Buffer (Meso Scale Discovery) and reading on a SECTOR® Imager 6000 instrument (Meso Scale Discovery) using manufacturer-installed assay-specific default settings. Percent inhibition of the EGF-stimulated positive control signal was calculated for each EGFR-binding FN3 domain.
Inhibition of EGF-stimulated EGFR phosphorylation was measured for 232 identified clones from the TCL1 and TCL2 libraries. 22 of these clones inhibited EGFR phosphorylation by ≥50% at 1 μM concentration. After removal of clones that either expressed poorly or were judged to be multimeric by size exclusion chromatography, nine clones were carried forward for further biological characterization. The BC and FG loop sequences of these clones are shown in Table 4. Eight of the nine selected clones had a common FG loop sequence (HNVYKDTNMRGL; SEQ ID NO: 95) and areas of significant similarity were seen between several clones in their BC loop sequences.
Large-scale Expression, Purification, and Endotoxin Removal
The 9 FN3 domains shown in Table 4 were scaled up to provide more material for detailed characterization. An overnight culture containing each EGFR-binding FN3 domain variant was used to inoculate 0.8 L of Terrific broth medium supplemented with 100 μg/mL ampicillin at a 1/80 dilution of overnight culture into fresh medium, and incubated with shaking at 37° C. The culture was induced when the optical density at 600 nm reached ˜1.2-1.5 by addition of IPTG to a final concentration of 1 mM and the temperature was reduced to 30° C. After 4 hours, cells were collected by centrifugation and the cell pellet was stored at −80° C. until needed.
For cell lysis, the thawed pellet was resuspended in 1× BugBuster® supplemented with 25 U/mL Benzonase® (Sigma-Aldrich) and 1 kU/mL rLysozyme™ (Novagen EMD Biosciences) at a ratio of 5 mL of BugBuster® per gram of pellet. Lysis proceeded for 1 hour at room temperature with gentle agitation, followed by centrifugation at 56,000×g for 50 minutes at 4° C. The supernatant was collected and filtered through a 0.2 un filter, then loaded on to a 5-mL HisTrap FF column pre-equilibrated with Buffer A (50 mM Tris-HCl pH 7.5, 500 mM NaCl, 10 mM imidazole) using a GE Healthcare ÄKTAexplorer 100s chromatography system. The column was washed with 20 column volumes of Buffer A and further washed with 16% Buffer B (50 mM Tris-HCl pH7.5, 500 mM NaCl, 250 mM imidazole) for 6 column volumes. The FN3 domains were eluted with 50% B for 10 column volumes, followed by a gradient from 50-100% B over 6 column volumes. Fractions containing the FN3 domain protein were pooled, concentrated using a Millipore 10K MWCO concentrator, and filtered before loading onto a HiLoad™ 16/60 Superdex™ 75 column (GE Healthcare) pre-equilibrated with PBS. The protein monomer peak eluting from the size exclusion column was retained.
Endotoxins were removed using a batch approach with ActiClean Etox resin (Sterogene Bioseparations). Prior to endotoxin removal, the resin was pre-treated with 1 N NaOH for 2 hours at 37° C. (or overnight at 4° C.) and washed extensively with PBS until the pH had stabilized to ˜7 as measured with pH indicator paper. The purified protein was filtered through a 0.2 μm filter before adding to 1 mL of Etox resin at a ratio of 10 mL of protein to 1 mL of resin. The binding of endotoxin to resin was allowed to proceed at room temperature for at least 2 hours with gentle rotation. The resin was removed by centrifugation at 500×g for 2 minutes and the protein supernatant was retained. Endotoxin levels were measured using EndoSafe®-PTS™ cartridges and analyzed on an EndoSafe®-MCS reader (Charles River). If endotoxin levels were above 5 EU/mg after the first Etox treatment, the above procedure was repeated until endotoxin levels were decreased to ≤5 EU/mg. In cases where the endotoxin level was above 5 EU/mg and stabilized after two consecutive treatments with Etox, anion exchange or hydrophobic interaction chromatography conditions were established for the protein to remove the remaining endotoxins.
Affinity Determination of Selected EGFR-Binding FN3 Domains to EGFR-Fc (EGFR-Fc Affinity)
Binding affinity of selected EGFR-binding FN3 domains to recombinant EGFR extracellular domain was further characterized by surface Plasmon resonance methods using a Proteon Instrument (BioRad). The assay set-up (chip preparation, EGFR-Fc capture) was similar to that described above for off-rate analysis. Selected EGFR binding FN3 domains were tested at 1 μM concentration in 3-fold dilution series in the horizontal orientation. A buffer sample was also injected to monitor the baseline stability. The dissociation phase for all concentrations of each EGFR-binding FN3 domain was monitored at a flow rate of 100 μL/min for 30 minutes (for those with koff˜10−2 s−1 from off-rate screening), or 1 hour (for those with koff˜10−3 s−1 or slower). Two sets of reference data were subtracted from the response data: 1) the inter-spot signals to correct for the non-specific interactions between the EGFR-binding FN3 domain and the immobilized IgG surface; 2) the buffer channel signals to correct for baseline drifting due to the dissociation of captured EGFR-Fc surface over time. The processed binding data at all concentrations for each FN3 domain were globally fit to a 1:1 simple Langmuir binding model to extract estimates of the kinetic (kon, koff) and affinity (KD) constants. Table 5 shows the kinetic constants for each of the constructs, with the affinity varying from 200 pM to 9.6 nM.
Binding of Selected EGFR-Binding FN3 Domains to EGFR on Cells (A431 Cell Binding Assay)
A431 cells were plated at 5,000/well in opaque black 96-well plates and allowed to attach overnight at 37° C., in a humidified 5% CO2 atmosphere. Purified EGFR-binding FN3 domains (1.5 nM to 30 μM) were added to the cells (in 50 uL) for 1 hour at room temperature in triplicate plates. Supernatant was removed and cells were washed 3 times with 150 μL/well of FACS stain buffer. Cells were incubated with 50 μL/well of anti-penta His-Alexa488 antibody conjugate (Qiagen) diluted 1:100 in FACS stain buffer for 20 minutes at room temperature. Cells were washed 3 times with 150 μL/well of FACS stain buffer, after which wells were filled with 100 μL of FACS stain buffer and read for fluorescence at 488 nm using an Acumen eX3 reader. Data were plotted as raw fluorescence signal against the logarithm of the FN3 domain molar concentration and fitted to a sigmoidal dose-response curve with variable slope using GraphPad Prism 4 (GraphPad Software) to calculate EC50 values. Table 5 reports the EC50 for each of the constructs ranging from 2.2 to >20 μM.
Inhibition of EGF Binding to EGFR on Cells Using Selected EGFR-Binding FN3 Domains (A431 Cell EGF Competition Assay)
A431 cells were plated at 5,000/well in opaque black 96-well plates and allowed to attach overnight at 37° C., in a humidified 5% CO2 atmosphere. Purified EGFR-binding FN3 domains (1.5 nM to 30 μM) were added to the cells (50 μL/well) for 1 hour at room temperature in triplicate plates. Biotinylated EGF (Invitrogen, Cat #: E-3477) was added to each well to give a final concentration of 30 ng/mL and incubated for 10 minutes at room temperature. Cells were washed 3 times with 150 μL/well of FACS stain buffer. Cells were incubated with 50 μL/well of streptavidin-phycoerythrin conjugate (Invitrogen) diluted 1:100 in FACS stain buffer for 20 minutes at room temperature. Cells were washed 3 times with 150 μL/well of FACS stain buffer, after which wells were filled with 100 μL of FACS stain buffer and read for fluorescence at 600 nm using an Acumen eX3 reader. Data were plotted as the raw fluorescence signal against the logarithm of FN3 domain molar concentration and fitted to a sigmoidal dose-response curve with variable slope using GraphPad Prism 4 (GraphPad Software) to calculate IC50 values. Table 5 reports the IC50 values ranging from 1.8 to 121 nM.
Inhibition of EGF-Stimulated EGFR Phosphorylation (Phoshpo-EGRF Assay)
Select FN3 domains that significantly inhibited EGF-stimulated EGFR phosphorylation were assessed more completely by measuring IC50 values for inhibition. Inhibition of EGF-stimulated EGFR phosphorylation was assessed at varying FN3 domain concentrations (0.5 nM to 10 μM) as described above in “inhibition of EGF stimulated EGFR phosphorylation”. Data were plotted as electrochemiluminescence signal against the logarithm of the FN3 domain molar concentration and IC50 values were determined by fitting data to a sigmoidal dose response with variable slope using GraphPad Prism 4 (GraphPad Software). Table 5 reports the IC50 values ranging from 18 nM to >2.5 μM.
Inhibition of Human Tumor Cell Growth (NCI-H292 Growth and NCI-H322 Growth Assay)
Inhibition of EGFR-dependent cell growth was assessed by measuring viability of the EGFR over-expressing human tumor cell lines, NCI-H292 and NCI-H322 (American Type Culture Collection, cat. # CRL-1848 & # CRL-5806, respectively), following exposure to EGFR-binding FN3 domains. Cells were plated at 500 cells/well (NCI-H292) or 1,000 cells/well (NCI-H322) in opaque white 96-well tissue culture-treated plates (Nunc) in 100 μL/well of RPMI medium (Gibco) containing GlutaMAX™ and 10 mM HEPES, supplemented with 10% heat inactivated fetal bovine serum (Gibco) and 1% penicillin/streptomycin (Gibco), and allowed to attach overnight at 37° C. in a humidified 5% CO2 atmosphere. Cells were treated by addition of 5 μL/well of phosphate-buffered saline (PBS) containing a concentration range of EGFR-binding FN3 domains. Controls were treated with 5 μL/well of PBS only or 25 mM ethylenediaminetetraacetic acid in PBS. Cells were incubated at 37° C., 5% CO2 for 120 hours. Viable cells were detected by addition of 75 μL/well of CellTiter-Glo® reagent (Promega), followed by mixing on a plate shaker for 2 minutes, and incubation in the dark at room temperature for a further 10 minutes. Plates were read on a SpectraMax M5 plate reader (Molecular Devices) set to luminescence mode, with a read time of 0.5 seconds/well against a blank of medium only. Data were plotted as a percentage of PBS-treated cell growth against the logarithm of FN3 domain molar concentration. IC50 values were determined by fitting data to the equation for a sigmoidal dose response with variable slope using GraphPad Prism 4 (GraphPad Software). Table 5 shows IC50 values ranging from 5.9 nM to 1.15 μM and 9.2 nM to >3.1 μM, using the NCI-H292 and NCI-H322 cells respectively.
A subset of the EGFR binding FN3 domains was engineered to increase the conformational stability of each molecule. The mutations L17A, N46V, and E86I (described in US Pat. Publ. No. 2011/0274623) were incorporated into clones P54AR4-83, P54CR4-31, and P54AR4-37 by DNA synthesis. The new mutants, P54AR4-83v2, P54CR431-v2, and P54AR4-37v2 were expressed and purified as described above. Differential scanning calorimetry in PBS was used to assess the stability of each mutant in order to compare it to that of the corresponding parent molecule. Table 6 shows that each clone was stabilized significantly, with an average increase in the Tm of 18.5° C.
Cysteine mutants of FN3 domains are made from the base Tencon molecule and variants thereof that do not have cysteine residues. These mutations may be made using standard molecular biology techniques known in the art to incorporate a unique cysteine residue into the base Tencon sequence (SEQ ID NO: 1) or other FN3 domains in order to serve as a site for chemical conjugation of small molecule drugs, fluorescent tags, polyethylene glycol, or any number of other chemical entities. The site of mutation to be selected should meet certain criteria. For example, the Tencon molecule mutated to contain the free cysteine should: (i) be highly expressed in E. coli, (ii) maintain a high level of solubility and thermal stability, and (iii) maintain binding to the target antigen upon conjugation. Since the Tencon scaffold is only ˜90-95 amino acids, single-cysteine variants can easily be constructed at every position of the scaffold to rigorously determine the ideal position(s) for chemical conjugation.
Each individual amino acid residue, from positions 1-95 (or 2-96 when the N-terminal methionine is present) of the P54AR4-83v2 mutant (SEQ ID NO: 27), which binds EGFR, was mutated to cysteine to assess the best chemical conjugation sites.
Construction, Expression and Purification
The amino acid sequence of each individual cysteine variant of P54AR4-83v2 was reverse translated into nucleic acid sequences encoding the proteins using preferred codons for E. coli expression and a synthetic gene was produced (DNA 2.0). These genes were cloned into a pJexpress401 vector (DNA 2.0) for expression driven by a T5 promoter sequence and transformed into E. coli strain BL21 (Agilent). The P54AR4-83v2 “cys scan” library was provided as glycerol stocks arrayed into a 96-well plate and the expression and purification of each followed the same procedure described in Example 2.
Chemical Conjugation
For the P54AR4-83v2 “cys scan” library, conjugation was integrated into the purification process. Cysteine variants in clarified lysate were bound to Ni-NTA resin in 96-well format using His-trap HP plates (catalog #28-4008-29, GE Healthcare) by adding lysate to the wells and centrifugation at 100×g for 5 min. The resin was washed 3 times with buffer A, and then N-ethyl maleimide (NEM) was added as 500 μL of a 1.5 mM solution. Following a one-hour room temperature incubation on a rotisserie shaker, excess NEM was removed by centrifugation and three washes with buffer A. Conjugated cysteine variants was eluted with 2×150 μL of buffer B and exchanged into PBS with MultiScreen Filter Plates with Ultracel-10 membrane (catalog # MAUF1010, Millipore) or with 96-well PD MultiTrap plates (catalog #28-9180-06, GE Healthcare). Conjugates were characterized by mass spectrometry (Table 7). Cysteine variants that expressed poorly (less than 0.1 mg of protein obtained from a 5 mL culture or no protein detected by mass spectrometry) or conjugated poorly to NEM (less than 80% conjugated, as determined by mass spectrometry) were excluded from further analysis. This eliminated L1C, W21C, Q36C, E37C, A44C, D57C, L61C, Y67C, and F92C due to poor expression and A17C, L19C, 133C, Y35C, Y56C, L58C, T65C, V69C, 171C, and T94C due to low conjugation efficiency.
Analytical Size-Exclusion Chromatography
Size exclusion chromatography for each NEM-conjugated cysteine variants of P54AR4-83v2 was performed as described in Example 2. Table 8 summarizes the results. The percent monomer for each protein was determined by integrating the Abs280 signal and comparing the peak in the monomer region (5.5-6 minutes) to the peaks in the oligomer region (4-5.3 minutes).
EGFR Binding Assay
Relative binding affinity of the NEM-conjugated cysteine variants of P54AR4-83v2 to EGFR was assessed as described in Example 2. Table 9 summarizes the data showing the ratios of each cysteine variant EGFR binding affinity relative to the P54AR4-83v2 parent protein. Cysteine conjugates that had reduced binding to EGFR (<65% of the signal observed with P54AR4-83v2 parent when treated with 10 nM protein) as determined by the ELISA assay were excluded from further analysis: P2C, A3C, P4C, K5C, L7C, D23C, W25C, F27C, Y28C, F31C, S55C, G73C, H75C, V77C, Y78C, T81C, N82C, M83C, and G85C.
Thermal Stability
The thermal stability of cysteine-NEM conjugates was assessed by differential scanning calorimetry (DSC). The only the conjugates tested were those determined to express at high levels, conjugate efficiently, and retain EGFR binding. Additionally, cysteine variants within the BC and FG loops were excluded. Stability data was generated by heating a 400 μL aliquot of the variant from 25° C. to 100° C. at a scan rate of 1° C. per minute in a VP-DSC instrument (MicroCal). A second identical scan was completed on the sample in order to assess the reversibility of thermal folding/unfolding. Data was fitted to a 2-state unfolding model in order to calculate the melting temperature (Table 10). Cys variants with reduced melting temperatures (≤63° C., or >8° C. below the P54AR4-83v2 parent) or that demonstrated irreversible unfolding were excluded from further analysis: V9C, V12C, T13C, R18C, E29C, E39C, G42C, V49C, P50C, G51C, P63C.
Cytotoxicity Assay
P54AR4-83v2 cysteine variants were conjugated to the cytotoxic tubulin inhibitor momomethyl auristatin F (MMAF) via an enzyme-cleavable Val-Cit linker or a non-cleavable PEG4 linker (VC-MMAF; see
Cell killing was assessed by measuring viability of the EGFR-overexpressing human tumor cell line H1573 following exposure to the cysteine variant-cytotoxin conjugates. Cells were plated in black-well, clear bottomed, tissue culture-treated plates (Falcon 353219) at 7000/well in 100 μL/well of phenol red free RPMI media (Gibco 11835-030) with 5% fetal bovine serum (Gibco). Cells were allowed to attach overnight at 37° C. in a humidified 5% CO2 atmosphere. Medium was aspirated from 96-well plate and cells were treated with 50 uL of fresh media and 50 uL of 2× inhibitor made up in fresh media. Cell viability was determined by an endpoint assay with Cell TiterGlo (Promega) at 70 hours. IC50 values were determined by fitting data to the equation for a sigmoidal dose response with variable slope using GraphPad Prism 5 (GraphPad Software). Table 11 reports IC50 values obtained from analysis of the CellTiter Glo data. The average IC50 of two replicates of the 83v2-cys/vcMMAF conjugate was 0.7 nM. Four of the 32 conjugates tested had IC50 values more than two times that of the parent (above 1.4 nM) and were discarded: L32C, T68C, Y72C, and V74C. Additionally, three conjugates gave IC50 values over two times more potent than the parent and may be especially suitable for formatting into drug conjugates: N6C, E53C, and T93C.
Final Cysteine Variants
Of the 96 positions tested, 28 of the cysteine variants were found to meet the criteria of retention of high expression level in E. coli, efficient conjugation via thiol-maleimide chemistry, retention of binding to target antigen EGFR, retention of thermostability and reversible unfolding properties, and retention of killing of cells with high EGFR expression when the cysteine variant is conjugated to a cytotoxic drug. These positions are: N6C (SEQ ID NOS: 210 and 248), V8C (SEQ ID NOS: 189 and 227), S10C (SEQ ID NOS: 190 and 228), E11C (SEQ ID NOS: 191 and 229), E14C (SEQ ID NOS: 192 and 230), D15C (SEQ ID NOS: 193 and 231), S16C (SEQ ID NOS: 194 and 232), S20C (SEQ ID NOS: 195 and 233), S30C (SEQ ID NOS: 196 and 234), Q34C (SEQ ID NOS: 197 and 235), S38C (SEQ ID NOS: 198 and 236), K40C (SEQ ID NOS: 199 and 237), V41C (SEQ ID NOS: 200 and 238), I45C (SEQ ID NOS: 201 and 239), L47C (SEQ ID NOS: 202 and 240), T48C (SEQ ID NOS: 203 and 241), E53C (SEQ ID NOS: 204 and 242), R54C (SEQ ID NOS: 205 and 243), T59C (SEQ ID NOS: 206 and 244), G60C (SEQ ID NOS: 207 and 245), K62C (SEQ ID S 208 and 246), G64C (SEQ ID NOS: 209 and 247), T68C (SEQ ID NOS: 210 and 248), S70C (SEQ ID NOS: 211 and 249), L88C (SEQ ID NOS: 212 and 250), S89C (SEQ ID NOS: 213 and 251), A90C (SEQ ID NOS: 214 and 252), 191C (SEQ ID NOS: 215 and 253), and T93C (SEQ ID NOS: 216 and 254). The locations of these 28 positions within the structure of the 83v2 protein are shown in
Panning on Human c-Met
The TCL14 library was screened against biotinylated-human c-Met extracellular domain (bt-c-Met) to identify FN3 domains capable of specifically binding c-Met. For selections, 3 μg of TCL14 library was in vitro transcribed and translated (IVTT) in E. Coli S30 Linear Extract (Promega, Madison, Wis.) and the expressed library blocked with Cis Block (2% BSA (Sigma-Aldrich, St. Louis, Mo.), 100 μg/ml Herring Sperm DNA (Promega), 1 mg/mL heparin (Sigma-Aldrich)). For selections, bt-c-Met was added at concentrations of 400 nM (Round 1), 200 nM (Rounds 2 and 3) and 100 nM (Rounds 4 and 5). Bound library members were recovered using neutravidin magnetic beads (Thermo Fisher, Rockford, Ill.) (Rounds 1, 3, and 5) or streptavidin magnetic beads (Promega) (Rounds 2 and 4) and unbound library members were removed by washing the beads 5-14 times with 500 uL PBS-T followed by 2 washes with 500 μL PBS.
Additional selection rounds were performed to identify FN3 domains molecules with improved affinities. Briefly, outputs from round 5 were prepared as described above and subjected to additional iterative rounds of selection with the following changes: incubation with bt-c-Met was decreased from 1 hour to 15 minutes and bead capture was decreased from 20 minutes to 15 minutes, bt-c-Met decreased to 25 nM (Rounds 6 and 7) or 2.5 nM (Rounds 8 and 9), and an additional 1 hour wash was performed in the presence of an excess of non-biotinylated c-Met. The goal of these changes was to simultaneously select for binders with a potentially faster on-rate and a slower off-rate yielding a substantially lower KD.
Rounds 5, 7 and 9 outputs were PCR cloned into a modified pET15 vector (EMD Biosciences, Gibbstown, N.J.) containing a ligase independent cloning site (pET15-LIC) using TCON6 (SEQID No. 30) and TCON5 E86I short (SEQID No. 31) primers, and the proteins were expressed as C-terminal His6-tagged proteins after transformations and IPTG induction (1 mM final, 30° C. for 16 hours) using standard protocols. The cells were harvested by centrifugation and subsequently lysed with Bugbuster HT (EMD Biosciences) supplemented with 0.2 mg/mL Chicken Egg White Lysozyme (Sigma-Aldrich). The bacterial lysates were clarified by centrifugation and the supernatants were transferred to new 96 deep-well plates.
Screening for FN3 Domains that Inhibit HGF Binding to c-Met
FN3 domains present in E. coli lysates were screened for their ability to inhibit HGF binding to purified c-Met extracellular domain in a biochemical format. Recombinant human c-Met Fc chimera (0.5 μg/mL in PBS, 100 μL/well) was coated on 96-well White Maxisorp Plates (Nunc) and incubated overnight at 4° C. The plates were washed two times with 300 μl/well of Tris-buffered saline with 0.05% Tween 20 (TBS-T, Sigma-Aldrich) on a Biotek plate washer. Assay plates were blocked with StartingBlock T20 (200 μL/well, Thermo Fisher Scientific, Rockland, Ill.) for 1 hour at room temperature (RT) with shaking and again washed twice with 300 μl of TBS-T. FN3 domain lysates were diluted in StartingBlock T20 (from 1:10 to 1:100,000) using the Hamilton STARplus robotics system. Lysates (50 μL/well) were incubated on assay plates for 1 hour at RT with shaking. Without washing the plates, bt-HGF (1 μg/mL in StartingBlock T20, 50 μL/well, biotinylated) was added to the plate for 30 min at RT while shaking. Control wells containing Tencon27 lysates received either Starting Block T20 or diluted bt-HGF. Plates were then washed four times with 300 μl/well of TBS-T and incubated with 100 μl/well of Streptavidin-HRP (1:2000 in TBS-T, Jackson Immunoresearch, West Grove, Pa.) for 30-40 minutes at RT with shaking. Again the plates were washed four times with TBS-T. To develop signal, POD Chemiluminescence Substrate (50 μL/well, Roche Diagnostics, Indianapolis, Ind.), prepared according to manufacturer's instructions, was added to the plate and within approximately 3 minutes luminescence was read on the Molecular Devices M5 using SoftMax Pro. Percent inhibition was determined using the following calculation: 100−((RLUsample−Mean RLUNo bt-HGF control)/(Mean RLUbt-HGF control−Mean RLUNo bt-HGF control)*100). Percent inhibition values of 50% or greater were considered hits.
High-Throughput Expression and Purification of FN3 Domains
His-tagged FN3 domains were purified from clarified E. coli lysates with His MultiTrap™ HP plates (GE Healthcare) and eluted in buffer containing 20 mM sodium phosphate, 500 mM sodium chloride, and 250 mM imidazole at pH 7.4. Purified samples were exchanged into PBS pH 7.4 for analysis using PD MultiTrap™ G-25 plates (GE Healthcare).
IC50 Determination of Inhibition of HGF Binding to c-Met
Select FN3 domains were further characterized in the HGF competition assay. Dose response curves for purified FN3 domains were generated utilizing the assay described above (starting concentrations of 5 μM). Percent inhibition values were calculated. The data were plotted as % inhibition against the logarithm of FN3 domain molar concentrations and IC50 values were determined by fitting data to a sigmoidal dose response with variable slope using GraphPad Prism 4.
35 unique sequences were identified from Round 5 to exhibit activity at dilutions of 1:10, with IC50 values ranging from 0.5 to 1500 nM. Round 7 yielded 39 unique sequences with activity at dilutions of 1:100 and IC50 values ranging from 0.16 to 2.9 nM. 66 unique sequences were identified from Round 9, where hits were defined as being active at dilutions of 1:1000. IC50 values as low as 0.2 nM were observed in Round 9 (Table 13).
FN3 domains were expressed and purified as described above in Example 2. Size exclusion chromatography and kinetic analysis was done as described above in Examples 1 and 2, respectively. Table 12 shows the sequences of the C-strand, CD loop, F-strand, and FG loop, and a SEQ ID NO: for the entire amino acid sequence for each domain.
C loop residues correspond to residues 28-37 of indicated SEQ ID NO:
CD strand residues correspond to residues 38-43 of indicated SEQ ID NO:
F loop residues correspond to residues 65-74 of indicated SEQ ID NO:
FG strand residues correspond to residues 75-81 of indicated SEQ ID NO:
Binding of Selected c-Met-Binding FN3 Domains to c-Met on Cells
NCI-H441 cells (Cat # HTB-174, American Type Culture Collection, Manassas, Va.) were plated at 20,000 cells per well in Poly-D-lysine coated black clear bottom 96-well plates (BD Biosciences, San Jose, Calif.) and allowed to attach overnight at 37° C., 5% CO2. Purified FN3 domains (50 μL/well; 0 to 1000 nM) were added to the cells for 1 hour at 4° C. in duplicate plates. Supernatant was removed and cells were washed three times with FACS stain buffer (150 μL/well, BD Biosciences, cat #554657). Cells were incubated with biotinylated-anti HIS antibody (diluted 1:160 in FACS stain buffer, 50 μL/well, R&D Systems, cat # BAM050) for 30 minutes at 4° C. Cells were washed three times with FACS stain buffer (150 μL/well), after which wells were incubated with anti mouse IgG1-Alexa 488 conjugated antibody (diluted 1:80 in FACS stain buffer, 50 μL/well, Life Technologies, cat # A21121) for 30 minutes at 4° C. Cells were washed three times with FACS stain buffer (150 μL/well) and left in FACS stain buffer (50 μL/well). Total fluorescence was determined using an Acumen eX3 reader. Data were plotted as raw fluorescence signal against the logarithm of the FN3 domain molar concentration and fitted to a sigmoidal dose-response curve with variable slope using GraphPad Prism 4 (GraphPad Software) to calculate EC50 values. FN3 domains were found to exhibit a range of binding activities, with EC50 values between 1.4 and 22.0, as shown in Table 13.
Inhibition of HGF-Stimulated c-Met Phosphorylation
Purified FN3 domains were tested for their ability to inhibit HGF-stimulated phosphorylation of c-Met in NCI-H441, using the c-Met phospho(Tyr1349) kit from Meso Scale Discovery (Gaithersburg, Md.). Cells were plated at 20,000/well in clear 96-well tissue culture-treated plates in 100 μL/well of RPMI medium (containing Glutamax and HEPES, Life Technologies) with 10% fetal bovine serum (FBS; Life Technologies) and allowed to attach overnight at 37° C., 5% CO2. Culture medium was removed completely and cells were starved overnight in serum-free RPMI medium (100 μL/well) at 37° C., 5% CO2. Cells were then replenished with fresh serum-free RPMI medium (100 μL/well) containing FN3 domains at a concentration of 20 μM and below for 1 hour at 37° C., 5% CO2. Controls were treated with medium only. Cells were stimulated with 100 ng/mL recombinant human HGF (100 μL/well, R&D Systems cat #294-HGN) and incubated at 37° C., 5% CO2 for 15 minutes. One set of control wells was left un-stimulated as negative controls. Medium was then completely removed and cells were lysed with Complete Lysis Buffer (50 μL/well, Meso Scale Discovery) for 10 minutes at RT with shaking, as per manufacturer's instructions. Assay plates configured for measuring phosphorylated c-Met were blocked with the provided blocking solution as per the manufacturer's instructions at room temperature for 1 hour. Plates were then washed three times with Tris Wash Buffer (200 μL/well, Meso Scale Discovery). Cell lysates (30 μL/well) were transferred to assay plates, and incubated at RT with shaking for 1 hour. Assay plates were then washed four times with Tris Wash Buffer, after which ice-cold Detection Antibody Solution (25 μL/well, Meso Scale Discovery) was added to each well for 1 hr at RT with shaking. Plates were again rinsed four times with Tris Wash Buffer. Signals were detected by addition of 150 Read Buffer (150 μL/well, Meso Scale Discovery) and reading on a SECTOR® Imager 6000 instrument (Meso Scale Discovery) using manufacturer-installed assay-specific default settings. Data were plotted as electrochemiluminescence signal against the logarithm of FN3 domain molar concentration and IC50 values were determined by fitting data to a sigmoidal dose response with variable slope using GraphPad Prism 4. FN3 domains were found to inhibit phosphorylated c-Met with IC50 values ranging from 4.6 to 1415 nM as shown in Table 13.
Inhibition of Human Tumor Cell Growth
Inhibition of c-Met-dependent cell growth was assessed by measuring viability of U87-MG cells (American Type Culture Collection, cat # HTB-14), following exposure to c-Met-binding FN3 domains. Cells were plated at 8000 cells per well in opaque white 96-well tissue culture-treated plates (Nunc) in 100 μL/well of RPMI medium, supplemented with 10% FBS and allowed to attach overnight at 37° C., 5% CO2. Twenty-four hours after plating, medium was aspirated and cells were replenished with serum-free RPMI medium.
Twenty-four hours after serum starvation, cells were treated by addition of serum-free medium containing c-Met-binding FN3 domains (30 μL/well). Cells were incubated at 37° C., 5% CO2 for 72 hours. Viable cells were detected by addition of 100 μL/well of CellTiter-Glo® reagent (Promega), followed by mixing on a plate shaker for 10 minutes. Plates were read on a SpectraMax M5 plate reader (Molecular Devices) set to luminescence mode, with a read time of 0.5 seconds/well. Data were plotted as raw luminescence units (RLU) against the logarithm of FN3 domain molar concentration. IC50 values were determined by fitting data to an equation for a sigmoidal dose response with variable slope using GraphPad Prism 4. Table 13 reports IC50 values ranging from 1 nM to >1000 nM.
Thermal Stability of c-Met-Binding FN3 Domains
Differential scanning calorimetry in PBS was used to assess the stability of each FN3 domain. Results of the experiment are shown in Table 14.
Generation of Bispecific EGFR/c-Met Molecules
Numerous combinations of the EGFR and c-Met-binding FN3 domains described in Examples 1-6 were joined into bispecific molecules capable of binding to both EGFR and c-Met. Additionally, EGFR-binding FN3 domains having amino acid sequences shown in SEQ ID NOs: 107-110 and c-Met binding FN3 domains having amino acid sequences shown in SEQ ID NOs: 111-114 were made and joined into bispecific molecules. Synthetic genes were created to encode for the amino acid sequences described in SEQID No. 50-72 and 106 (Table 15) such that the following format was maintained: EGFR-binding FN3 domain followed by a peptide linker followed by a c-Met-binding FN3 domain. A poly-histidine tag was incorporated at the C-terminus to aid purification. In addition to those molecules described in Table 15, the linker between the two FN3 domains was varied according to length, sequence composition and structure according to those listed in Table 16. It is envisioned that a number of other linkers could be used to link such FN3 domains Bispecific EGFR/c-Met molecules were expressed and purified from E. coli as described for monospecific EGFR or c-Met FN3 domains using IMAC and gel filtration chromatography steps.
Bispecific EGFR/c-Met Molecules Enhance Potency Compared to Monospecific Molecules Alone, Suggesting Avidity
NCI-H292 cells were plated in 96 well plates in RPMI medium containing 10% FBS. 24 hours later, medium was replaced with serum free RPMI. 24 hours after serum starvation, cells were treated with varying concentrations of FN3 domains: either a high affinity monospecific EGFR FN3 domain (P54AR4-83v2), a weak affinity monospecific c-Met FN3 domain (P114AR5P74-A5), the mixture of the two monospecific EGFR and c-Met FN3 domains, or a bispecific EGFR/c-Met molecules comprised of the low affinity c-Met FN3 domain linked to the high affinity EGFR FN3 domain (ECB1). Cells were treated for 1 h with the monospecific or bispecific molecules and then stimulated with EGF, HGF, or a combination of EGF and HGF for 15 minutes at 37° C., 5% CO2. Cells were lysed with MSD Lysis Buffer and cell signaling was assessed using appropriate MSD Assay plates, according to manufacturer's instructions, as described above.
The low affinity c-Met FN3 domain inhibited phosphorylation of c-Met with an IC50 of 610 nM (
The potential for the bispecific EGFR/c-Met molecule to enhance the inhibition of c-Met and/or EGFR phosphorylation through an avidity effect was evaluated in multiple cell types with variable c-Met and EGFR densities and ratios (
The potential for enhanced potency with a bi-specific EGFR/c-Met molecule was evaluated in a c-Met phosphorylation assay using a molecule with a high affinity to EGFR (0.26 nM) and medium affinity to c-Met (10.1 nM). In both NCI-H292 and NCI-H596 cells, the inhibition of phosphorylation of c-Met was enhanced with the bispecific molecule compared to the monospecific c-Met-binding FN3 domain, by 134 and 1012 fold, respectively (
It was verified that the enhanced potency for inhibition of EGFR and c-Met phosphorylation with the bispecific EGFR/c-Met molecules translated into an enhanced inhibition of signaling and proliferation. For these experiments, the mixture of FN3 EGFR-binding and c-Met-binding FN3 domains was compared to a bispecific EGFR/c-Met molecule. As described in Tables 17 and 18, the IC50 values for ERK phosphorylation (Table 17) and proliferation of H292 cells (Table 18) were decreased when cells were treated with the bispecific EGFR/c-Met molecule compared to the mixture of the monospecific binders. The IC50 for inhibition of ERK phosphorylation for the bi-specific EGFR/c-Met molecule was 143-fold lower relative to the mixture of the two monospecific EGFR and c-Met FN3 domains, showing the effect of avidity to the potency of the molecules in this assay. In Table 17, the monospecific EGFR- and c-Met binding FN3 domains do not fully inhibit activity and therefore the IC50 values shown should be considered lower limits. The proliferation assay was completed using different combinations EGFR and c-Met binding FN3 domains either as a mixture or linked in a bispecific format. The IC50 for inhibition of proliferation for the bispecific EGFR/c-Met molecule was 34-236-fold lower relative to the mixture of the monospecific parent EGFR or c-Met binding FN3 domains. This confirmed that the avidity effect observed at the level of the receptors (
In Vive Tumor Xenografts: PK/PD
In order to determine efficacy of the monospecific and bispecific FN3 domain molecules in vivo, tumor cells were engineered to secrete human HGF (murine HGF does not bind to human HGF). Human HGF was stably expressed in NCI-H292 cells using lentiviral infection (Lentiviral DNA vector expressing human HGF (Accession # X16322) and lentiviral packaging kit from Genecopoeia). After infection, HGF-expressing cells were selected with 4 μg/mL puromycin (Invitrogen). Human HGF protein was detected in the conditioned medium of pooled cells using assay plates from MesoScale Discovery.
SCID Beige mice were subcutaneously inoculated with NCI-H292 cells expressing human HGF (2.0×106 cells in Cultrex (Trevigen) in a volume of 200 μL) on the dorsal flank of each animal. Tumor measurements were taken twice weekly until tumor volumes ranged between 150-250 mm3. Mice were then given a single IP dose of bispecific EGFR/c-Met molecules (linked to an albumin binding domain to increase half-life) or PBS vehicle. At 6 h or 72 h after dosing, tumors were extracted and immediately frozen in liquid nitrogen. Blood samples were collected via cardiac puncture into 3.8% citrate containing protease inhibitors. Immediately after collection, the blood samples were centrifuged and the resulting plasma was transferred to sample tubes and stored at −80° C. Tumors were weighed, cut into small pieces, and lysed in Lysing Matrix A tubes (LMA) containing RIPA buffer with HALT protease/phosphatase inhibitors (Pierce), 50 mM sodium fluoride (Sigma), 2 mM activated sodium orthovanadate (Sigma), and 1 mM PMSF (MesoScale Discovery). Lysates were removed from LMA matrix and centrifuged to remove insoluble protein. The soluble tumor protein was quantified with a BCA protein assay and diluted to equivalent protein levels in tumor lysis buffer. Phosphorylated c-Met, EGFR and ERK were measured using assay plates from MesoScale Discovery (according to Manufacturer's protocol and as described above).
The concentration of each bispecific EGFR/c-Met molecule was measured at 6 and 72 hours after dosing in the blood and in the tumor (
Tumor Efficacy Studies with Bispecific EGFR/c-Met Molecules
SCID Beige mice were subcutaneously inoculated with NCI-H292 cells expressing human HGF (2.0×106 cells in Cultrex (Trevigen) in 200 μL) in the dorsal flank of each animal. One week after implantation, mice were stratified into groups with equivalent tumor volumes (mean tumor volume=77.9+/−1.7 mm3). Mice were dosed three times per week with the bispecific molecules and tumor volumes were recorded twice weekly. Tumor growth inhibition (TGI) was observed with four different bispecific molecules, with variable affinities for c-Met and EGFR.
Efficacy of Bispecfic Molecule and Other Inhibitors of EGFR and c-Met
The in vivo therapeutic efficacies of a bispecific EGFR/c-Met molecule (ECB38) and the small molecule inhibitors crizotinib (c-Met inhibitor) and erlotinib (EGFR inhibitor), cetuximab (anti-EGFR antibody), each as a single agent, and the combination of crizotinib and erlotinib, were evaluated in the treatment of subcutaneous H292-HGF human lung cancer xenograft model in SCID/Beige mice (
The H292-HGF cells were maintained in vitro in RPMI1640 medium supplemented with fetal bovine serum (10% v/v), and L-glutamine (2 mM) at 37° C. in an atmosphere of 5% CO2 in air. The cells were routinely subcultured twice weekly by trypsin-EDTA treatment. The cells growing in an exponential growth phase were harvested and counted for tumor inoculation.
Each mouse was inoculated subcutaneously at the right flank region with H292-HGF tumor cells (2×106) in 0.1 ml of PBS with cultrex (1:1) for tumor development. The treatments were started when the mean tumor size reached 139 mm3. The test article administration and the animal numbers in each study group were shown in the following experimental design table (Table 26). The date of tumor cell inoculation was denoted as day 0.
Before commencement of treatment, all animals were weighed and the tumor volumes were measured. Since the tumor volume can affect the effectiveness of any given treatment, mice were assigned into groups using randomized block design based upon their tumor volumes. This ensures that all the groups are comparable at the baseline. The randomized block design was used to assign experimental animals to groups. First, the experimental animals were divided into homogeneous blocks according to their initial tumor volume. Secondly, within each block, randomization of experimental animals to treatments was conducted. Using randomized block design to assign experimental animals ensured that each animal had the same probability of being assigned to a given treatment and therefore systematic error was reduced.
At the time of routine monitoring, the animals were checked for any effects of tumor growth and treatments on normal behavior, such as mobility, visual estimation of food and water consumption, body weight gain/loss (body weights were measured twice weekly), eye/hair matting and any other abnormal effect.
The major endpoint was whether tumor growth can be delayed or tumor bearing mice can be cured. Tumor size was measured twice weekly in two dimensions using a caliper, and the volume was expressed in mm3 using the formula: V=0.5a×b2 where a and b are the long and short diameters of the tumor, respectively. The tumor size was then used for calculations of both T-C and T/C values. T-C was calculated with T as the time (in days) required for the mean tumor size of the treatment group to reach 1000 mm3, and C was the time (in days) for the mean tumor size of the control group to reach the same size. The T/C value (in percent) was an indication of antitumor efficacy; T and C were the mean volume of the treated and control groups, respectively, on a given day. Complete tumor regression (CR) is defined as tumors that are reduced to below the limit of palpation (62.5 mm3). Partial tumor regression (PR) is defined as tumors that are reduced from initial tumor volume. A minimum duration of CR or PR in 3 or more successive tumor measurements is required for a CP or PR to be considered durable.
Animals for which the body weight loss exceeded 20%, or for which the mean tumor size of the group exceeds 2000 mm3 were euthanized. The study was terminated after two weeks of observation after the final dose.
Summary statistics, including mean and the standard error of the mean (SEM), are provided for the tumor volume of each group at each time point (shown in Table 19 below). Statistical analyses of difference in tumor volume among the groups were evaluated using a one-way ANOVA followed by individual comparisons using Games-Howell (equal variance not assumed). All data were analyzed using SPSS 18.0. p<0.05 was considered to be statistically significant.
The mean tumor size of the vehicle treated group (Group 1) reached 1,758 mm3 at day 23 after tumor inoculation. Treatment with the bispecific EGFR/c-Met molecule at 25 mg/kg dose level (Group 2) led to complete tumor regression (CR) in all mice which were durable in >3 successive tumor measurements (average TV=23 mm3, T/C value=1%, p=0.004 compared with the vehicle group at day 23).
Treatment with Crizotinib as a single agent at 50 mg/kg dose level (Group 3) showed no antitumor activity; the mean tumor size was 2,102 mm3 at day 23 (T/C value=120%, p=0.944 compared with the vehicle group).
Treatment with Erlotinib as a single agent at 50 mg/kg dosing level (Group 4) showed minor antitumor activity, but no significant difference was found compared with the vehicle group; the mean tumor size was 1,122 mm3 at day 23 (T/C value=64%, p=0.429 compared with the vehicle group), with 4 days of tumor growth delay at tumor size of 1,000 mm3 compared with the vehicle group.
The combination of Crizotinib (50 mg/kg, Group 5) and Erlotinib (50 mg/kg, Group 5) showed significant antitumor activity; the mean tumor size was 265 mm3 at day 23 (T/C=15%; p=0.008), with 17 days of tumor growth delay at tumor size of 1,000 mm3 compared with the vehicle group.
Cetuximab at 1 mg/mouse dosing level as a single agent (Group 6) showed significant antitumor activities; the mean tumor size was 485 mm3 at day 23 (T/C=28%; p=0.018), with 17 days of tumor growth delay at tumor size of 1,000 mm3 compared with the vehicle group.
Medium to severe body weight loss was observed in the vehicle group which might be due to the increasing tumor burden; 3 mice died and 1 mouse were euthanized when BWL>20% by day 23. Slight toxicity of the bispecific EGFR/c-Met molecule was observed in Group 2; 3 mice were euthanized when BWL>20% during the treatment period; the body weight was gradually recovered when the treatment was withdrawn during the 2 weeks of observation period. More severe body weight loss was observed in the Crizotinib or Erlotinib monotherapy group compared to the vehicle group, suggesting the treatment related toxicity. The combination of Crizotinib and Erlotinib was generally tolerated during the dosing phase, but severe body weight loss was observed at the end of the study, which might be due to the resumption of the fast tumor growth during the non-treatment period. The monotherapy of Cetuximab was well tolerated in the study; body weight loss was only observed at the end of the study due to the resume of the tumor growth.
In summary, the bispecific EGFR/c-Met molecule at 25 mg/kg (3 times/week×3 weeks) produced a complete response in H292-HGF human lung cancer xenograft model in SCID/Beige mice. The treatment was tolerated in 7 out of 10 mice, and resulted in severe body weight loss in 3 out of 10 mice.
Numerous methods have been described to reduce kidney filtration and thus extend the serum half-life of proteins including modification with polyethylene glycol (PEG) or other polymers, binding to albumin, fusion to protein domains which bind to albumin or other serum proteins, genetic fusion to albumin, fusion to IgG Fc domains, and fusion to long, unstructured amino acid sequences.
Bispecific EGFR/c-Met molecules were modified with PEG in order to increase the hydrodynamic radius by incorporating a free cysteine at the C-terminus of the molecule. Most commonly, the free thiol group of the cysteine residue is used to attach PEG molecules that are functionalized with maleimide or iodoacetemide groups using standard methods. Various forms of PEG can be used to modify the protein including linear PEG of 1000, 2000, 5000, 10,000, 20,000, or 40,000 kDa. Branched PEG molecules of these molecular weights can also be used for modification. PEG groups may also be attached through primary amines in the bispecific EGFR/c-Met molecules in some instances.
In addition to PEGylation, the half-life of bispecific EGFR/c-Met molecules was extended by producing these proteins as fusion molecules with a naturally occurring 3-helix bundle serum albumin binding domain (ABD) or a consensus albumin binding domain (ABDCon). These protein domains were linked to the C-terminus of the c-Met-binding FN3 domain via any of the linkers described in Table 16. The ABD or ABDCon domain may also be placed between the EGFR-binding FN3 domain and the c-Met binding FN3 domain in the primary sequence.
Select EGFR/c-Met molecules were characterized for their affinity to both EGFR and c-Met, their ability to inhibit EGFR and c-Met autophosphorylation, and their effect on proliferation of HGF cells. Binding affinity of the bispecific EGFR/c-Met molecules to recombinant EGFR and/or c-Met extracellular domain was further by surface Plasmon resonance methods using a Proteon Instrument (BioRad) according to protocol described in Example 3. Results of the characterization are shown in Table 21.
Generation of Bispecific EGFR/c-Met Molecules
Based on the data generated from the cysteine scanning of the P54AR4-83v2 mutant (Example 5), cysteine mutants were also designed in a bispecific anti-EGFR/c-Met molecule denoted ECB147 (SEQ ID NOS: 218 and 256), which consists of the P54AR4-83v2 (SEQ ID NO: 27), the cMet binder P114AR7P95-C5v2 (SEQ ID NO: 114), and an albumin binding domain for half-life extension. These three domains are connected by (Ala-Pro)5 linkers (SEQ ID NO: 81). Variants with one, two, or four cysteines were designed with substitutions at the C-terminus, in the linker regions, or at the Lys-62 position of one of the FN3 domains (SEQ ID NOS: 219-225 and 257-263). Another bispecific variant, ECB82cys (SEQ ID NOS: 226 and 264) consists of P54AR4-83v2 (SEQ ID NO: 27), P114AR7P94-A3v22 (SEQ ID NO: 111), and a variant of the albumin-binding domain, all three domains connected by AP linkers, and a single C-terminal cysteine. An additional cysteine variant of the non-targeted Tencon scaffold (SEQ ID NO: 265) was also used for the construction of control conjugates. All the variants were constructed, expressed, and purified as described in previous examples. Purity was assessed by SDS-PAGE analysis. Analytical size exclusion chromatography using a Superdex 75 5/150 column (GE Healthcare) shows that the FN3 domain preparations are free of aggregates and elute at a time consistent with a monomeric protein. Mass spectrometry determined the masses to be in agreement with the theoretical masses (Table 22).
Chemical Conjugation
To chemically conjugate the purified bispecific cysteine variants to maleimide-containing molecules, the proteins were first reduced with TCEP to generate free thiols. 1-2 mg of each bispecific cysteine variant was mixed with an excess of TCEP at neutral pH (Sigma catalog #646547) and incubated at RT for 30-60 minutes. TCEP was removed by adding 3 volumes of saturated ammonium sulfate solution (4.02 M) to precipitate the cysteine variants. After centrifugation at 16000-20000×g at 4° C. for 20 min and removal of the supernatant, the protein pellet was dissolved in PBS or sodium phosphate buffer and mixed immediately with a 5- to 10-fold excess of the maleimide-containing molecule. The reaction was incubated for 30-60 minutes at room temperature and then quenched with an excess of a free thiol, such as cysteine or β-mercaptoethanol, to scavenge excess maleimide. The unbound maleimide was removed with Zeba desalting columns (Thermo catalog #89890), by preparative SEC with a Tosoh G3000SW×1 column (# P4619-14N; 7.8 mm×30 cm; 5 Gun), or by binding the cysteine variant to Ni-NTA resin, washing, and eluting essentially as described above. Conjugates were characterized by SDS-PAGE and mass spectrometry. This general method was used to conjugate bispecific cysteine variants to fluorescein maleimide (Thermo catalog #62245), PEG24-maleimide (Quanta Biodesign catalog #10319), and maleimide-cytotoxin molecules with a variety of linkers (see structures in
Inhibition of EGF-Stimulated EGFR Phosphorylation
Purified bispecific PEG24-maleimide conjugates were tested for their ability to inhibit EGF-stimulated phosphorylation of EGFR in the human tumor cell line NCI-H292 (American Type Culture Collection, cat. # CRL-1848) using the EGFR phospho(Tyr1173) kit from Meso Scale Discovery (Gaithersburg, Md.) and as described in Example 3. The conjugates were compared to unmodified ECB38 (SEQID No. 109), which differs from ECB147 by two amino acids. The conjugates and ECB38 inhibited EGFR with similar IC50 values, as shown in Table 23, demonstrating that modification at the designed sites does not significantly affect target binding.
Inhibition of HGF-Stimulated c-Met Phosphorylation
Purified bispecific PEG24-maleimide conjugates were also tested for their ability to inhibit HGF-stimulated phosphorylation of c-Met in NCI-H292 cells, using the c-Met phosphor (Tyr1349) kit from Meso Scale Discovery (Gaithersburg, Md.), and as described in Example 7. The conjugates and ECB38 inhibited cMet with similar IC50 values as shown in Table 24, demonstrating that modification at these sites does not significantly alter target binding.
Cytotoxicity Assay
Conjugates consisting of ECB147 cysteine variants, 83v2-cys, or Tencon-cys linked to a cytotoxic tubulin inhibitor from the auristatin family (
Homo
sapiens
Homo sapiens
Homo
sapiens
Cysteine-Engineered EGFR/c-Met Bispecific Molecules (with Methionine Included at N-Terminus)
E. Coli
E. Coli
E. Coli
E. Coli
E. Coli
E. Coli
E. Coli
E. Coli
This application claims the benefit of U.S. Provisional Application Ser. No. 62/352,776, filed 21 Jun. 2016. The entire contents of the aforementioned application are incorporated herein by reference in their entirety.
| Number | Name | Date | Kind |
|---|---|---|---|
| 5223409 | Ladner et al. | Jun 1993 | A |
| 5643763 | Dunn et al. | Jul 1997 | A |
| 5643768 | Kawasaki | Jul 1997 | A |
| 5846456 | Liu | Dec 1998 | A |
| 5856456 | Whitlow et al. | Jan 1999 | A |
| 6172197 | McCafferty et al. | Jan 2001 | B1 |
| 6472147 | Janda et al. | Oct 2002 | B1 |
| 6582915 | Griffiths et al. | Jun 2003 | B1 |
| 6673901 | Koide | Jan 2004 | B2 |
| 6969108 | Fukumoto et al. | Nov 2005 | B2 |
| 7842476 | McGregor et al. | Nov 2010 | B2 |
| 10196446 | Goldberg | Feb 2019 | B2 |
| 20040197332 | Ullrich et al. | Oct 2004 | A1 |
| 20050004029 | Garcia | Jan 2005 | A1 |
| 20050272083 | Seshagiri | Dec 2005 | A1 |
| 20080241159 | Gerritsen et al. | Oct 2008 | A1 |
| 20090042906 | Huang et al. | Feb 2009 | A1 |
| 20090274693 | Gilmer et al. | Nov 2009 | A1 |
| 20090299040 | Camphausen | Dec 2009 | A1 |
| 20090311803 | Way et al. | Dec 2009 | A1 |
| 20100179094 | Emanuel et al. | Jul 2010 | A1 |
| 20100216708 | Jacobs et al. | Aug 2010 | A1 |
| 20100254989 | Bossenmaier et al. | Oct 2010 | A1 |
| 20110038866 | Hastewell et al. | Feb 2011 | A1 |
| 20110053842 | Camphausen et al. | Mar 2011 | A1 |
| 20110118144 | Hyun et al. | May 2011 | A1 |
| 20110274623 | Jacobs | Nov 2011 | A1 |
| 20110287009 | Scheer et al. | Nov 2011 | A1 |
| 20120225870 | Janne et al. | Sep 2012 | A1 |
| 20120244164 | Beste et al. | Sep 2012 | A1 |
| 20120263723 | Davies et al. | Oct 2012 | A1 |
| 20130012435 | Camphausen | Jan 2013 | A1 |
| 20130039927 | Dewhurst et al. | Feb 2013 | A1 |
| 20130184212 | Camphausen et al. | Jul 2013 | A1 |
| 20130226834 | Gannalo, II | Aug 2013 | A1 |
| 20140141000 | Chiu | May 2014 | A1 |
| 20140155325 | Mark | Jun 2014 | A1 |
| 20140155326 | Mark | Jun 2014 | A1 |
| 20140255408 | Chiu | Sep 2014 | A1 |
| 20140349929 | Camphausen | Nov 2014 | A1 |
| 20150104808 | Goldberg et al. | Apr 2015 | A1 |
| 20170348397 | Diem | Dec 2017 | A1 |
| Number | Date | Country |
|---|---|---|
| 102076713 | May 2011 | CN |
| 2011507543 | Mar 2011 | JP |
| 2011517314 | Jun 2011 | JP |
| 2011520961 | Jul 2011 | JP |
| 2011522517 | Aug 2011 | JP |
| 9638557 | Dec 1996 | WO |
| 2008127710 | Oct 2008 | WO |
| 2009083804 | Jul 2009 | WO |
| 2009085462 | Jul 2009 | WO |
| 2009102421 | Aug 2009 | WO |
| 2009111691 | Sep 2009 | WO |
| 2009126834 | Oct 2009 | WO |
| 2009133208 | Nov 2009 | WO |
| 2009142773 | Nov 2009 | WO |
| 2010039248 | Apr 2010 | WO |
| 2010115551 | Oct 2010 | WO |
| 2011110642 | Sep 2011 | WO |
| 2011130324 | Oct 2011 | WO |
| 2011131746 | Oct 2011 | WO |
| 2011151412 | Dec 2011 | WO |
| 2014081954 | May 2014 | WO |
| 2014165093 | Oct 2014 | WO |
| Entry |
|---|
| Diem et al. Selection of high-affinity Centyrin FN3 domains from a simple library diversified ata combination of strand and loop positions. Protein Engineer Design Selection 27(10): 419-429, 2014. |
| Goldberg et al. Engineering a targeted delivery platform using Centyrins. Protein Engineer Design Selet 29(12): 563-572, 2016. |
| Klein et al. Bispecific centyrin simultaneously targeting EGFR and c-Met demonstrates improved activity compared to the mixture of single agents. Cancer Res 73(8 Suppl): Abstract LB-312, Apr. 2013. |
| Mamluk et al. Anti-tumor effect of CT-322 as an adnectin inhibitor of vascular endothelial growth factor receptor-2. mAbs 2(2): 199-208, 2010. |
| Adjei et al., “Early Clinical Development of ARQ197, a Selective, Non-ADP-Competitive Inhibitor Targeting MET Tyrosine Kinase for the Treatment of Advanced Cancers,” The Oncologist, vol. 16, pp. 788-799 (2011). |
| Alfthan et al, “Properties of a single-chain antibody containing different linker peptides,” Protein Engineering, vol. 8, No. 7, pp. 725-731 (1995). |
| Baselga et al., “Critical Update and Emerging Trends in Epidermal Growth Factor Receptor Targeting in Cancer,” Journal of Clinical Oncology, vol. 23, No. 11, pp. 2445-2459 (2005). |
| Batley et al., “Inhibition of FGF-1 Receptor Tyrosine Kinase Activity by PD 161570, a New Protein-Tyrosine Kinase Inhibitor,” Life Sciences, vol. 62, No. 20, pp. 143-150 (1998). |
| Bean et al., “MET amplification occurs with or without T790M mutations in EGFR mutant lung tumors with acquired resistance to gefitinib or erlotinib,” Proceedings of the National Academy of Science, vol. 104, No. 52, pp. 20932-20937 (2007). |
| Berzofsky et al., “Antigen-antibody Interaction,” Chapter 23, Fundamental Immunology, edited by W.E. Paul, Raven Press, New York, 595-644 (1984). |
| Birtalan et al., “The Intrinsic Contributions of Tyrosine, Serine, Glycine and Arginine to the Affinity and Specificity of Antibodies,” Journal of Molecular Biology, vol. 377, pp. 1518-1528 (2008). |
| Bork et al., “Proposed acquisition of an animal protein domain by bacteria,” Proceedings of the National Academy of Science USA, vol. 89, pp. 8990-8994 (1992). |
| Cappuzzo et al., “Epidermal Growth Factor Receptor Gene and Protein and Gefitinib Sensitivity in Non-small-Cell Lung Cancer,” Journal of the National Cancer Institute, vol. 97, pp. 643-655 (2005). |
| Christensen et al., “c-Met as a target for human cancer and characterization of inhibitors for therapeutic intervention,” Cancer Letters, vol. 225, pp. 1-26 (2005). |
| Cooper et al., “Molecular cloning of a new transforming gene from a chemically transformed human cell line,” Nature, vol. 311, pp. 29-33 (1984). |
| DeRoock et al., “Effects of KRAS, BRAF, NRAS, and PIK3CA mutations on the efficacy of cetuximab plus chemotherapy in chemotherapy-refractory metastatic colorectal cancer: a retrospective consortium analysis,” Lancet Oncology, vol. 11, pp. 753-762 (2010). |
| Downward et al., “Autophosphorylation sites on the epidermal growth factor receptor,” Nature, vol. 311, pp. 483-485 (1984). |
| Engelman et al., “MET Amplification Leads to Gefitinib Resistance in Lung Cancer by Activating ERBB3 Signaling,” Science, vol. 316, pp. 1039-1043 (2007). |
| Ferguson, Kathryn M., “Structure-Based View of Epidermal Growth Factor Receptor Regulation,” Annual Review of Biophysics, vol. 37, pp. 535-373 (2008). |
| NCBI Reference Sequence NP_0052192, “Epidermal Growth Factor Receptor Isoform a Precursor [Homo sapiens],” pp. 1-14 (May 18, 2014). |
| Gill et al., “Monoclonal Anti-epidermal Growth Factor Receptor Antibodies Which Are Inhibitors of Epidermal Growth Factor Binding and Antagonists of Epidermal Growth Factor-stimulated tyrosine Protein Kinase Activity, The Journal of Biological Chemistry,” vol. 259, No. 12, pp. 7755-7760 (1984). |
| Goldstein et al., “Biological efficacy of a chimeric antibody to the epidermal growth factor receptor in a human tumor xenograft model,” Clinical Cancer Research, vol. 1, pp. 1311-1318 (1995). |
| Grünwald et al., “Developing Inhibitors of the Epidermal Growth Factor Receptor for Cancer Treatment,” Journal of the National Cancer Institute, vol. 95, No. 12, pp. 851-867 (2003). |
| Hallewell et al., “Genetically Engineered Polymers of Human CuZN Superoxide Dismutase,” The Journal of Biological Chemistry, vol. 264, No. 9, pp. 5260-5268 (1989). |
| Hanes et al., “In vitro selection and evolution of functional proteins by using ribosome display,” Proceedings of the National Academy of Sciences USA, vol. 94, pp. 4937-4942 (1997). |
| Hirsch et al, “Combination of EGFR gene copy number and protein expression predicts outcome for advanced non-small-cell lung cancer patients treated with gefitnib,” Annals of Oncology, vol. 18, pp. 752-760 (2007). |
| Ichimura et al., “Expression of c-met/HGF Receptor in Human Non-small Cell Lung Carcinomas in vitro and in vivo and Its Prognostic Significance,” Japan Journal of Cancer Research, vol. 87. pp. 1063-1069 (1996). |
| Jacobs et al. “Design of novel FN3 domains with high stability by a consensus sequence approach,” Protein _ Engineering, Design & Selection, vol. 25, No. 3, pp. 107-117, 2012. |
| Lehmann et al., “Engineering proteins for thermostability: the use of sequence alignments versus rational design and directed evolution,” Current Opinion in Biotechnology, vol. 12, pp. 371-375 (2001). |
| Linardou et al., “Somatic EGFR mutations and efficacy of tyrosine kinase inhibitors in NSCLC,” National Review of Clinical Oncology, vol. 6, pp. 352-366 (2009). |
| Li et al., “Skin toxicities associated with epidermal growth factor receptor inhibitors,” Target Oncology, vol. 4, pp. 107-119 (2009). |
| Ma et al., “c-Met: Structure, functions and potential for therapeutic inhibition,” Cancer and Metastasis Reviews, vol. 22, pp. 309-325 (2003). |
| Meinke et al., “Cellulose-Binding Polypeptides from Cellulomonas fimi: Endoglucanase D (CenD), a Family A b-1,4-Glucanase,” Journal of Bacteriology, vol. 175, No. 7, pp. 1910-1918 (1993). |
| Mendelsohn et al., “Epidermal Growth Factor Receptor Targeting in Cancer,” Seminars in Oncology, vol. 33, pp. 369-385 (2006). |
| Mendelsohn et al., “The EGF receptor family as targets for cancer therapy,” Oncogene, vol. 19, pp. 6550-6565 (2000). |
| Odegrip et al., “CIS display: In vitro selection of peptides from libraries of protein-DNA complexes,” Proceedings of the National Academy of Science USA, vol. 101, No. 9, pp. 2806-2810 (2004). |
| Olson et al., “Design, expression, and stability of a diverse protein library based on the human fibronectin type III domain,” Protein Science, vol. 16, pp. 476-484 (2007). |
| Panek et al.,“In Vitro Pharmacological Characterization of PD 166285, a New Nanomolar Potent and Broadly Active Protein Tyrosine Kinase Inhibitor,” The Journal of Pharmacology and Experimental Therapeutics, vol. 283, No. 3, pp. 1433-1444 (1997). |
| Peters et al., “MET: a promising anticancer therapeutic target,” Nature Reviews Clinical Oncology, vol. 9, pp. 314-326 (2012). |
| Prewett et al., “Mouse-Human chimeric Anti-Epidermal Growth Factor Receptor Antibody C225 Inhibits the Growth of Human Renal Cell Carcinoma Xenografts in Nude Mice,” Clinical Cancer Research, vol. 4, pp. 2957-2966 (1998). |
| Riely et al., “Clinical Course of Patients with Non-Small Cell Lung Cancer and Epidermal Growth Factor Receptor Exon 19 and Exon 21 Mutations Treated with Gefitinib or Erlotinib,” Clinical Cancer Research, vol. 12, No. 3, pp. 839-844 (2006). |
| Roberts et al., “RNA-peptide fusions for the in vitro selection of peptides and proteins,” Proceedings of the National Academy of Science USA, vol. 94, pp. 12297-12302 (1997). |
| Robinson et al., “Covalent Attachment of Arc Repressor Subunits by a Peptide Linker Enhances Affinity for Operator DNA,” Biochemistry, vol. 35, pp. 109-116 (1996). |
| Sakakura et al., “Gains, Losses, and Amplifications of Genomic Materials in Primary Gastric Cancers Analyzed by Comparative Genomic Hybridization,” Genes, Chromosomes & Cancer, vol. 24, pp. 299-305 (1999). |
| Schmidt et al., “Novel mutations of the MET proto-oncogene in papillary rental carcinomas,” Oncogene, vol. 18, pp. 2343-2350 (1999). |
| Siegfried et al., “The Clinical Significance of Hepatocyte Growth Factor for Non-Small Cell Lung Cancer,” Annals of Thoracic Surgery, vol. 66, pp. 1915-1918 (1998). |
| Sierra et al., “c-MET as a potential therapeutic target and biomarker in cancer,” Therapeutic Advances in Medical Oncology, vol. 3, No. 51, pp. 521-535 (2011). |
| Stamos et al., “Crystal structure of the HGF b-chain in complex with the Sema domain of the Met receptor,” The EMBO Journal, vol. 23, pp. 2325-2335 (2004). |
| Strohl, William R., “Optimization of Fc-mediated effector functions of monoclonal antibodies,” Current Opinion in Biotechnology, vol. 20, pp. 685-691 (2009). |
| Turke et al., “Preexistence and Clonal Selection of MET Amplification in EGFR Mutant NSCLC,” Cancer Cell, vol. 17, pp. 77-88 (2010). |
| Ullrich et al., “Human epidermal growth factor receptor cDNA sequence and aberrant expression of the amplified gene in A431 epidermoid carcinoma cells,” Nature, vol. 309, pp. 418-425 (1984). |
| Watanabe et al., “Gene Cloning of Chitinase A1 from Bacillus circulans WL-12 Revealed Its Evolutionary Relationship to Serratia Chitinase and to the Type III Homology Units of Fibronectin,” Journal of Biological Chemistry, vol. 265, pp. 15659-15665 (1990). |
| Zhang et al., “Complete disulfide bond assignment of a recombinant immunoglobulin G4 monoclonal antibody,” Analytical Biochemistry, vol. 311, pp. 1-9 (2002). |
| Määttä et al., “Proteolytic Cleavage and Phosphorylation of a Tumor-associated ErbB4 Isoform Promote Ligand-independent Survival and Cancer Cell Growth,” Molecular Biology, vol. 17, pp. 67-79 (2006). |
| Jänne et al., “Effect of Epidermal Growth Factor Receptor Tyrosine Kinase Domain Mutations on the Outcome of Patients with Non-small Cell Lung Cancer Treated with Epidermal Growth Factor Receptor Tyrosine Kinase Inhibitors,” Clinical Cancer Research, vol. 12, No. 14 Suppl, pp. 4416s-4420s (2006). |
| Hynes et al., “ERBB Receptors and Cancer: the Complexity of Targeted Inhibitors,” Nature Reviews, vol. 5, pp. 341-356 (2005). |
| SwissProt Accession No. P00533.2, “Epidermal Growth Factor Receptor,” pp. 1-49 (Jun. 11, 2014). |
| GenBank Accession No. NP 001120972, Jun. 30, 2018. |
| Spigel et al., “Final efficacy results from OAM4558g, a randomized phase II study evaluating MetMAb or placebo in combination with erlotinib in advanced NSCLC”, Journal of Clinical Oncology 29, No. 15, pp. 7505-7505, May 2011 (Abstract Only). |
| Jacobs et al., “FN3 Domain Engineering”, Protein Engineering, pp. 145-162, 2012. |
| Jacobs et al., “Fusion to a highly stable consensus albumin binding domain allows for tunable pharmacokinetics”, Protein Engineering, Design & Selection, vol. 28, No. 10, pp. 385-393, 2015. |
| Number | Date | Country | |
|---|---|---|---|
| 20170362301 A1 | Dec 2017 | US |
| Number | Date | Country | |
|---|---|---|---|
| 62352776 | Jun 2016 | US |
