EA039118B1 - Method for treating fibrodysplasia ossificans progressiva - Google Patents

Abstract

A method for treating Fibrodysplasia Ossificans Progressiva (FOP) is provided, comprising administering to a subject having FOP an effective regime of an antibody against Activin A, wherein the antibody comprises the antibody heavy and light chain variable regions SEQ ID NOs: 1 and 5, SEQ ID NOs: 9 and 13, or SEQ ID NOs: 18 and 17, respectively. The antibody is chimeric, veneered, humanized or human antibody.

Description

Prerequisites for the creation of the invention

Fibrodysplasia ossificans progressive (FOP) is an autosomal dominant disorder characterized by early onset, episodic and progressive ossification of skeletal muscle and associated connective tissue. FOP is caused by mutations in the intracellular domain of ACVR1 (ALK2), where in the vast majority of cases, arginine at position 206 is replaced by histidine (R206H) (Pignolo R.J. et al., 2011, Orphanet J. Rare Dis.6:80). ACVR1 is a type 1 receptor for bone morphogenetic proteins (BMPs). It is believed that the R206H mutation, among other things, increases the sensitivity of this receptor to activation and its resistance to silencing. There is no effective medical treatment for FOP.

Statement of the essence of the claimed invention

The present invention provides methods of treating fibrodysplasia ossificans progressive (FOP) comprising administering an effective anti-Activin A antibody regimen to a subject with FOP. SEQ ID NOS: 9 and 13, or SEQ ID NOS: 18 and 17, respectively. In another embodiment, the antibody is a chimeric, venated, humanized, or human antibody. In yet another embodiment of the invention, the antibody is an intact antibody. According to one embodiment of the invention, the antibody is a human IgG1 kappa antibody. Also according to one embodiment of the invention, when the antibody contains the variable regions of the heavy and light chains of the antibody SEQ ID NOS: 1 and 5, SEQ ID NOS: 9 and 13, or SEQ ID NOS : 18 and 17 respectively, the antibody is a human IgG1 kappa antibody.

Brief description of the figures

In FIG. Figure 1 shows micro-CT data of mice that developed ectopic osteogenesis 6 weeks after the start of tamoxifen administration, treated and untreated with ACVR2A-Fc/ACVR2B-Fc. Nine of ten control mice treated with mFc (numbered 1-9) developed ectopic osteogenesis 6 weeks after tamoxifen administration. The most common localization of osteogenesis was the hind limb, neck area and sternum. Two of eleven mice treated with ACVR2A-Fc/ACVR2B-Fc (numbered 12 and 14) developed ectopic osteogenesis 6 weeks after tamoxifen administration. The ectopic bone lesions in these two mice were small compared to those seen in the mFc treatment group and both were located in the sternum.

In FIG. Figure 2 shows micro-CT data of mice that developed ectopic osteogenesis 4 weeks after the start of tamoxifen administration, treated or not treated with LDN212854. Numbers 1-8 indicate mice treated with the tamoxifen + vehicle combination. The formation of large ectopic foci of ossification was recorded in mice numbered 1, 2, 3, 4, 5 and 7, and the formation of small ectopic foci of ossification was recorded in mice numbered 6 and 8. Numbers from 9 to 16 indicate mice treated with the combination of tamoxifen + LDN212854 . The formation of small ectopic foci of ossification was recorded in mice numbered 9, 12 and 13. Ectopic foci of ossification in mice numbered 10, 11, 14, 15 or 16 were absent.

In FIG. 3 shows micro-CT data from mice predisposed to ectopic osteogenesis treated with an anti-activin A antibody, an irrelevant control antibody of the appropriate isotype, and ACVR2A-Fc. The anti-activin A antibody most effectively inhibited the formation of ectopic foci of ossification.

In FIG. 4 shows micro-CT data from mice predisposed to ectopic osteogenesis treated with an anti-activin A antibody, an irrelevant isotype control antibody, and an anti-Acvr2a/Acvr2b antibody. Antibodies to activin A and Acvr2a/Acvr2b inhibited the formation of ectopic foci of ossification.

In FIG. 5 shows micro-CT data of mice prone to ectopic osteogenesis treated with various doses of anti-activin A antibody compared to an irrelevant control antibody of the corresponding isotype. Doses ranging from 1-25 mg/kg have been shown to be effective, with the most effective dose tested being 10 mg/kg.

Definitions

Antagonists are usually supplied in isolated form. This means that the purity of the antagonist preparation is usually at least 50 wt.% in relation to contaminating proteins and other contaminants formed during production or purification, while it is possible to combine the antagonist with an excess of pharmaceutically acceptable carrier(s) or other carrier media, designed to facilitate the use of the drug. Occasionally, antagonist preparations are at least 60%, 70%, 80%, 90%, 95% or 99% pure by weight with respect to protein contaminants and other contaminants generated during manufacture or purification.

To classify amino acid substitutions into conservative or non-conservative amino acids grouped as follows: group I (hydrophobic side chains): met, ala, val, leu, ile; Group

- 1 039118

II (neutral hydrophilic side chains): cys, ser, thr; group III (acid side chains): asp, glu; group IV (main side chains): asn, gln, his, lys, arg; group V (residues affecting chain orientation): gly, pro; and group VI (aromatic side chains): trp, tyr, phe. Conservative substitutions include amino acid substitutions within the same class. Non-conservative substitutions involve replacing a representative of one of these classes with a representative of another class.

The percent sequence identity is determined by antibody sequences that maximally align according to Kabat numbering for the variable region or EC numbering for the constant region. For other proteins, sequence identity can be determined by aligning sequences using algorithms such as BESTFIT, FASTA, and TFASTA in Wisconsin Genetics Software Package Release 7.0 (Genetics Computer Group, 575 Science Dr., Madison, WI), with parameters skips by default or after learning and best alignment. After alignment, if the antibody region of interest (e.g., the entire mature heavy or light chain variable region) matches the same region of the control antibody, the percent sequence identity between the regions of interest and control antibody is the number of positions occupied by the same amino acids in the regions of interest and control antibody. divided by the total number of aligned positions of the two parcels, excluding gaps, multiplied by 100 to convert to a percentage.

Compositions or methods containing one or more of these elements may include other elements not specifically mentioned. For example, a composition that contains an antibody may contain the antibody alone or in combination with other ingredients.

A humanized antibody is a genetically engineered antibody in which CDRs from a non-human donor antibody are inserted into human acceptor antibody sequences (See, for example, Queen, US Pat. Nos. 5,530,101 and 5,585,089; Winter, US Pat. U.S. Patent No. 6407213; Adair U.S. Patent Nos. 5859205 and 6881557; Foote U.S. Patent No. 6881557). Antibody acceptor sequences can be, for example, mature human antibody sequences, a composition of such sequences, a human antibody consensus sequence, or a germline region sequence. Thus, a humanized antibody is an antibody with some or all of the CDRs that are wholly or substantially derived from the donor antibody, and the variable region framework sequences and constant region sequences, if present, are wholly or substantially represented by human antibody sequences. Similarly, a humanized heavy chain includes at least one, two, and typically three CDRs that are wholly or substantially derived from the heavy chain of the donor antibody, and a heavy chain variable region framework sequence and a heavy chain constant region, if present, essentially represented by human heavy chain variable region framework sequences and constant region sequences. Similarly, a humanized light chain includes at least one, two, and typically three CDRs that are wholly or substantially derived from the light chain of a donor antibody, and a light chain variable region framework sequence and a light chain constant region, if present, essentially represented by the frame sequences of the human variable region of the light chain and the sequences of the constant region. In addition to nanobodies and dAbs, the humanized antibody contains a humanized heavy chain and a humanized light chain. The CDR in a humanized antibody is taken substantially from the corresponding CDR in the non-human antibody, where at least 85%, 90%, 95%, or 100% of the corresponding residues (as determined by Kabat) are identical to the corresponding CDRs. The antibody chain variable region or antibody chain constant region framework sequences are substantially taken from the human variable region or human constant region framework sequence, respectively, where at least 85%, 90%, 95%, or 100% of the respective Kabat residues are identical.

Although humanized antibodies often include all six CDRs (preferably defined by Kabat) from a mouse antibody, they can also be made with less than all CDRs (e.g., at least 3, 4, or 5 CDRs from a mouse antibody) (e.g., Pascalis et al., J. Immunol. 169: 3076, 2002; Vajdos et al., Journal of Molecular Biology, 320: 415-428, 2002; Iwahashi et al., Mol. Immunol. 36: 1079-1091, 1999; Tamura et al., Journal of Immunology, 164: 1432-1441, 2000).

A chimeric antibody is an antibody in which the mature variable regions of the light and heavy chains of a non-human (eg, mouse) antibody are combined with the constant regions of the human light and heavy chains. Such antibodies essentially or completely retain the binding specificity of the mouse antibody and contain about two-thirds of the human sequence.

A veneered antibody is a type of humanized antibody that retains some, and usually all, of the CDRs and some of the non-human variable region framework residues of the non-human antibody, but with the replacement of other variable region framework residues that may be included in the B- or T cells, such as open residues (Pad

- 2 039118 lan, Mol. Immunol. 28: 489, 1991), with residues at the appropriate positions in the human antibody sequence. The result is an antibody in which the CDRs are wholly or substantially derived from a non-human antibody, and the non-human antibody variable region frameworks are more humanized by substitutions.

The human antibody may be derived from a human or otherwise obtained by expression of human immunoglobulin genes (eg, in a transgenic mouse, in vitro, or by phage display). Methods for making human antibodies include the triome technique of Oestberg et al., Cys muoma 2: 361-367 (1983); Oestberg, US patent No. 4634664; and Engleman et al., US Pat. No. 4,634,666. Monoclonal antibodies can also be produced by transgenic mice carrying genes for the human immune system, such as the Veloclmmune® mouse available from Regeneron Pharmaceuticals, Inc. (Murphy, PNAS 111 no. 14, 5153-5158 (2014), Xenomouse, Jakobovits, Nature Biotechnology 25, 1134-1143 (2007), or the HuMAb mouse available from Medarex, Inc. (Lonberg, Handbook Exp. Pharmacol. 181, 69-97 (2008); Lonberg et al., WO 93/12227 (1993); патенты США №№ 5877397, 5874299, 5814318, 5789650, 5770429, 5661016, 5633425, 5625126, 5569825, 5545806, Nature 148, 1547-1553 (1994), Nature Biotechnology 14, 826 (1996), Kucherlapati, WO 91/10741 (1991) Human antibodies can also be generated using phage display techniques (see, for example, Dower et al., WO 91/17271 and McCafferty et al., WO 92/01047, US Pat. Nos. 5,877,218; 5,871,907; 5,858,657; 5,837,242;

If an antagonist is indicated to retain the properties of the parent antibody from which it was derived, such retention of properties may be total or partial. Complete retention of activity means that the activity of the antagonist is identical, within experimental error, or greater than that of the molecule from which it was derived. Partial retention of activity means that the activity is well above the background level of the negative control (ie above the experimental error) and preferably is at least 50% of the corresponding activity of the molecule from which it was derived.

Two antibodies share the same epitope if all amino acid mutations in the antigen that reduce or block the binding of one antibody also reduce or block the binding of the other. Two antibodies have overlapping epitopes if certain amino acid mutations that reduce or block the binding of one antibody also reduce or block the binding of the other.

Antibody competition is determined by an assay in which the antibody under test inhibits the specific binding of a reference antibody to a common antigen (see, for example, Junghans et al., Cancer Res. 50:1495, 1990). The test antibody competes with the control antibody if an excess of the test antibody (e.g., at least 2x, 5x, 10x, 20x, or 100x) inhibits control antibody binding by at least 50%, but preferably 75, 90, or 99% as measured by the assay competitive binding. Antibodies identified by competitive analysis (competitive antibodies) include antibodies that bind to the same epitope as the control antibody and antibodies that bind to an adjacent epitope close enough to the epitope bound to the control antibody to create steric hindrance.

Detailed description

I. General description

Methods for the treatment of progressive fibrodysplasia ossificans (FOP) are provided. Such methods include administering to a subject suffering from FOP an effective regimen for administering an activin type 2A receptor antagonist (ACVR2A) and/or an activin type 2B receptor (ACVR2B) and/or an activin type 1 receptor antagonist (ACVR1) and/or an antagonist activin A. Such antagonists include fusion proteins containing one or more extracellular domains (ECDs) of ACVR2A, ACVR2B and/or ACVR1 and an immunoglobulin heavy chain Fc domain. Antagonist antibodies for ACVR2A, ACVR2B, ACVR1, or Activin A are also provided.

II. ACVR1, ACVR2A, ACVR2B and Activin A

The transforming growth factor β (TGFP) superfamily of ligands includes, for example, bone morphogenetic proteins (BMPs) and growth and differentiation factors (GDFs). The receptors for these ligands are heteromeric receptor complexes formed from transmembrane type I and type II serine/threonine kinase receptors. Examples of type I receptors include type IA activin receptors (ACTRIA, ACVR1 or ALK2), type IA BMP receptor, and type III BMP receptor. Examples of type II receptors include the type IIA and IIB activin receptors (ACTRIIA, or ACVR2A and ACTRIIB, or ACVR2B) and the type II BMP receptor. The ligands of the TGFP superfamily differ in each case with different affinities for different type I and type II receptors.

In both cases, type I and type II receptors contain an extracellular ligand binding domain (ECD) and an intracellular serine/threonine kinase domain. In addition, type I receptors contain a glycine/serine rich region (GS box) that precedes the kinase domain and an L45 loop before

- 3 039118 lax kinase domain. Both receptors act in concert for ligands that activate downstream signaling pathways, eg Smad signaling pathways and non-Smad signaling pathways. Activation includes ligand binding, ligand-receptor oligomerization, and transphosphorylation of the type I receptor GS block by a type II receptor kinase. The type II receptor kinase is constitutively active and is involved in ligand binding and activation of the type I receptor.

ACVR1, also known as activin type I receptor, ACVR1A, ACVRLK2, or ALK2, is the type I receptor for the TGFp ligand superfamily. ACVR1 has serine/threonine kinase activity and phosphorylates Smad proteins and activates downstream signaling pathways. ACVR1 is found in many body tissues, including skeletal muscle and cartilage, and helps control the growth and development of bones and muscles. As described elsewhere in this document, certain mutations in the ACVR1 gene cause FOP. Examples of ACVR1 activity include the ability to bind to ligands, the ability to form complexes with the type II receptor, or the ability to activate downstream signaling pathways, such as the Smad pathway.

ACVR2 is also known as the type II activin receptor for the TGFp ligand superfamily. There are at least two ACVR2 receptors, for example the type IIA activin receptor (ACVR2A or ACTRIIA) and the type IIB activin receptor (ACVR2B or ACTRIIB). Reference to ACVR2 includes either or both of ACVR2A and ACVR2B. Expression of ACVR2A and ACVR2B can occur in a variety of tissues, including skeletal muscle, stomach, heart, endometrium, testes, prostate, ovaries, and neural tissues.

Upon ligand binding, the ACVR2 receptor forms a complex with a type I receptor, such as ACVR1, and phosphorylates the GS block of the type I receptor, thereby enhancing the kinase activity of the type I receptor. I or the ability to phosphorylate the type I receptor.

An example form of human ACVR2A has accession number P27037 in Swiss Prot. Residues 1-19 are the signal peptide, residues 20-135 are the extracellular domain, residues 59116 are the domain of activin type I and II receptors, residues 136-161 are the transmembrane domain, and residues 162-513 are the cytoplasmic domain. An example form of the human ACVR2B is numbered Q13705 in Swiss Prot. Residues 1-18 are the signal sequence, residues 19-137 are the extracellular domain, residues 27-117 are the domain of activin type I and II receptors, residues 138-158 are the transmembrane domain, and residues 159-512 are the cytoplasmic domain. An example form of human ACVR1 has accession number Q04771 in Swiss Prot. Residues 1-20 are the signal sequence, residues 21-123 are the extracellular domain, residues 33-104 are the domain of activin type I and II receptors, residues 124-146 are the transmembrane domain, and residues 147-509 are the cytoplasmic domain. Reference to any of ACVR1, ACVR2A, and ACVR2B includes such exemplary forms, their known isoforms and polymorphisms, such as those listed in the Swiss Prot database, related forms from other species, and other variants with at least 90, 95, 96, 97, 98 or 99% sequence identity with example shape.

Residues of forms ACVR2A, ACVR2B, and ACVR1, other than the example sequences above, are numbered with maximum alignment with the corresponding example sequences, so that the aligned residues are given the same number. Substitutions from exemplary sequences may be conservative or non-conservative. Reference to ACVR1, ACVR2A, or ACVR2B also includes intact extracellular domains (e.g., residues 20-135, 19-137, or 21-123 of ACVR2A, ACVR2B, and ACVR1, respectively) or a portion thereof that contains no or substantially no transmembrane and cytoplasmic portion. Portions of the extracellular domain retain enough intact extracellular domain residues to bind at least one ligand or counter-receptor that binds the intact extracellular domain and thereby act as an antagonist of the respective receptor (e.g., residues 59-116, 27-117, or 33-104 ACVR2A, ACVR2B and ACVR1 respectively).

Activin A in the human body can exist in the form of a homo- or heterodimeric protein. A homodimeric protein contains a homodimeric pair of a beta A subunit. A heterodimeric protein contains a beta subunit and a beta B, beta C or beta E subunit (ie beta A beta B, beta A beta C or beta A beta E). Each of the subunits is expressed as precursor polypeptides, including a signal peptide, a propeptide, and a mature polypeptide. An example of a precursor form of the human beta A subunit is the 426 amino acid polypeptide number P08476 in Swiss Prot, in which residues 1-20 are the signal peptide, residues 21-310 are the propeptide, and residues 311-426 are the mature polypeptide. An example of the precursor form of the human beta B subunit is number P09529 in Swiss Prot, in which residues 1-28 are the signal peptide, residues 29-292 are the propeptide, and residues 293-407 are the mature polypeptide. An example of the precursor form of the human beta C subunit is number P55103 in Swiss Prot, in which residues 1-18 are the signal peptide, residues 19-236 are the propeptide, and residues

- 4 039118

237-352 are the mature polypeptide. An example form of the precursor of the human beta E subunit is number P58166 in Swiss Prot, in which residues 1-19 are the signal peptide, residues 20-236 are the propeptide, and residues 237-350 are the mature polypeptide. Several variants of such sequences are known and are described in the Swiss Prot database. Reference to activin A includes any of the beta A homodimer, beta A beta B, beta A beta C and beta A beta E heterodimeric forms, as well as their subunits, as well as their precursors, in which the subunits are associated with a propeptide and/or a signal peptide, which are defined by the example sequences given in Swiss Prot or by other naturally occurring human forms of these sequences. Activin A signals by binding to ACVR2A or ACVR2B, but there is no evidence that it is a ligand for ACVR1.

III. Antagonists of ACVR1, ACVR2A, ACVR2B

Antagonists of ACVR1 type I receptor proteins and ACVR2 type II receptor proteins (eg, ACVR2A and/or ACVR2B) are suggested for the treatment of FOP. Such antagonists may act on receptors, among other mechanisms, directly by binding to the receptor (as in the case of an antibody to ACVR1, ACVR2A or ACVR2B) or indirectly by binding to a ligand or counterreceptor and inhibiting the binding of the ligand or counterreceptor to ACVR1, ACVR2A or ACVR2B (as in fusion protein ACVR1, ACVR2A or ACVR2B). ACVR2A and ACVR2B antagonists can also bind to Activin A.

An ACVR1, ACVR2A or ACVR2B antagonist provided herein can inhibit or reduce the activity of ACVR1, ACVR2A and/or ACVR2B by at least 1, 5, 10, 20, 30, 40, 50, 60, 70, 80, 90, 95.99% or more relative to the control cell or animal model that did not receive the antagonist.

Any ACVR1, ACVR2A or ACVR2B antagonist can be used in FOP treatments. The antagonist may be, for example, an ACVR1, ACVR2A or ACVR2B polypeptide, such as an extracellular domain, an antagonist antibody, or a small molecule inhibitor.

A. Extracellular domains of ACVR1, ACVR2A and ACVR2B polypeptides

Antagonists include the ACVR1, ACVR2A, and ACVR2B proteins and fragments thereof effective to inhibit at least one form of ACVR1, ACVR2A, and ACVR2B activity, respectively. Such antagonists typically contain the extracellular domain of ACVR1, ACVR2A or ACVR2B or part of it. Preferably, such extracellular domains are wholly or substantially free of transmembrane and cytoplasmic regions (ie, all remaining residues from these regions do not significantly affect the functions of the extracellular domain). In other words, the ACVR2A, ACVR2B, or ACVR1 components of such antagonists comprise or substantially contain all or part of the extracellular domain of ACVR2A, ACVR2B, or ACVR1, as described above. Such antagonists optionally include other component(s) other than ACVR2A, ACVR2B or ACVR1, as described in more detail below. Such extracellular domains, which do not contain or essentially do not contain transmembrane and cytoplasmic domains, are soluble. Such extracellular domains can function as antagonists by binding to a soluble ligand or counterreceptor, effectively competing with the binding of the ligand or counterreceptor to the ACVR1, ACVR2A, or ACVR2B cell surface receptor, thereby modulating (reducing) the availability of the ligand or counterreceptor in vivo.

Soluble extracellular domains may initially be expressed with a signal sequence that is cleaved off during expression. The signal sequence may be a native ACVR1, ACVR2A or ACVR2B signal sequence, such as those described in US Pat. plasminogen activator (TRA). In an alternative embodiment, soluble extracellular ACVR1, ACVR2A or ACVR2B polypeptides can be synthesized or expressed without a signal sequence.

The ECD or ligand-binding domains of ACVR1, ACVR2A and ACVR2B are highly conserved across many species, including mice and humans. ECDs contain a cysteine-rich region and a C-terminal region. The ACVR1, ACVR2A, and ACVR2B ECDs bind to a diverse group of TGFP family ligands, including, for example, activin A, myostatin (GDF-8), GDF-11, and BMP. See, for example, Souza et al. (2008) Molecular Endocrinology 22(12): 2689-2702.

Examples of ACVR2A and ACVR2B polypeptides and soluble ACVR2A and ACVR2B polypeptides include those disclosed in US Pat. No. 7,842,633, US Pat. No. 7,960,343, and US Pat. No. 7,709,605, each of which is incorporated herein by reference in its entirety.

The ECD of ACVR1, ACVR2A, or ACVR2B polypeptides can be mutated such that the polypeptide variant has altered ligand binding properties (eg, binding specificity or affinity). Some ACVR1, ACVR2A, or ACVR2B polypeptide variants are characterized by an altered binding affinity (eg, increased or decreased) for a specific ligand.

- 5 039118

Variants have at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% sequence identity with naturally occurring ACVR1, ACVR2A, or ACVR2B sequences and retain biological activity, and thus have ACVR1 activity. , ACVR2A and ACVR2B as described elsewhere in this document. Active variants and fragments of ACVR2A and ACVR2B are described, for example, in US patent No. 7842633, US patent No. 7960343 and US patent No. 7709605, each of which is fully incorporated herein by reference.

Assay procedures for assessing ACVR1, ACVR2A, or ACVR2B activity are disclosed in, for example, US Pat. No. 7,842,633, US Pat. No. 7,960,343, and US Pat. No. 7,709,605. the ability to interfere with the binding of the ligand to the receptor protein ACVR1, ACVR2A or ACVR2B.

B. Fusion proteins

The ACVR1, ACVR2A and ACVR2B polypeptides described above may be expressed as fusion proteins comprising at least a portion of an ACVR1, ACVR2A and/or ACVR2B polypeptide and one or more fusion domains.

Fusion domains include an immunoglobulin heavy chain constant region (Fc), human serum albumin (HSA), glutathione S-transferase (GST), protein A, protein G, or any fusion domain that can be used to stabilize, solubilize, isolate, or multimerize a fusion squirrel.

The immunoglobulin heavy chain Fc domain is the preferred domain for fusion proteins. Fusions with the Fc portion of an immunoglobulin confer desirable pharmacokinetic properties on a wide range of proteins (eg, increase protein stability and/or serum half-life). Thus, the invention provides fusion proteins comprising at least one ACVR1, ACVR2A and/or ACVR2B ECD fused to an immunoglobulin Fc domain.

The Fc domain for use in the methods of the present invention may be from any immunoglobulin. Any of the classes of immunoglobulin may be used, including IgG, IgA, IgM, IgD and IgE. There are various subclasses or isotypes within the IgG class, including, for example, IgG1, IgG ₂ , IgG ₃ and IgG4. In one embodiment, the Fc fusion protein contains the Fc domain of an IgG molecule. In a further embodiment, the Fc domain is from an IgG1 molecule. The immunoglobulin molecule can be obtained from any animal, including, for example, a mammal, rodent, human, mouse, rat, hamster, or rabbit. In one embodiment, the Fc domain of the immunoglobulin is derived from a mammal. In another embodiment, the Fc domain is from a human. In yet another embodiment, the Fc domain is from a rodent, such as a mouse or rat. In a specific embodiment, the Fc domain of the fusion protein is from human IgG1.

The Fc fusion proteins provided herein may be prepared by any method known to those skilled in the art. Fc fusion proteins may contain at least the CH2 and CH3 regions, and typically at least part of the hinge region. Despite the possible presence of the CH1 region, it is usually omitted in fusion proteins.

The fusion may be at any site within the Fc portion of the immunoglobulin constant domain. Fusions can be made to the C-terminus of the Fc portion of the constant domain, or directly by linking the N-terminus to the CH1 region of the heavy chain. Specific sites may be selected to optimize the biological activity, secretion, or binding parameters of the Fc fusion protein.

In some instances, a nucleic acid encoding an ACVR1, ACVR2A and/or ACVR2B ECD is C-terminally fused to a nucleic acid encoding the N-terminus of an immunoglobulin constant domain sequence. In other cases, N-terminal fusions are also possible. There is also the possibility of fusion of the ACVR1, ACVR2A and/or ACVR2B ECDs to the N-terminus and C-terminus of an immunoglobulin constant domain sequence.

For immunoglobulin fusions, see also US Patent No. 5,428,130, US Patent No. 5,843,725, US Patent No. 6,018,026, and WO 2005/070966, each of which is incorporated herein by reference.

A fusion protein can be obtained, for example, by recombinant expression of a nucleic acid encoding a fusion protein. For example, a fusion protein can be prepared by fusing a nucleic acid encoding an ACVR1, ACVR2A and/or ACVR2B ECD with a nucleic acid encoding an Fc domain. The ACVR1, ACVR2A and/or ACVR2B ECD nucleic acid may be fused to the N-terminus of the nucleic acid encoding the F domain, or may be fused to the C-terminus of the gene encoding the Fc domain. In an alternative embodiment, the ECD may be fused to any position in the Fc domain.

ECD fusion proteins may also include a linker. In the case of Fc fusion proteins, the linker may be located between the ACVR1, ACVR2A or ACVR2B ECD and the Fc domain, optionally replacing part or all of the hinge region. The linker may be 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 30, 50 or more amino acids, which generally do not have a secondary structure. The linker may be saturated with glycine and proline residues and may, for example, contain repeating after

- 6 039118 threonine/serine and glycine ratios (eg TG4 or SG ₄ repeats).

Two or more ECD-Fc fusion proteins can be joined together by a linker. In such cases, a linker may be positioned between the ECDs, or a linker may be positioned between the Fc domains to bring the fusion proteins together. For example, 1, 2, 3, 4 or more Fc and ACVR1, ACVR2A and/or ACVR2B fusion proteins can be linked together.

Examples of ACVR2A and/or ACVR2B ECD fusion proteins have been previously described, such as those disclosed in US Pat. No. 7,842,633, US Pat. No. 7,960,343 and US Pat.

One example of an ACVR2A antagonist is known as sotatercept (also referred to as ACE-011). The sotatercept contains an ACVR2A ECD fused to the Fc domain of human IgG1 and is described in detail in Carrancio et al. (2014) British J Haematology. 165(6): 870-872, incorporated herein by reference in its entirety.

One example of an ACVR2B antagonist is known as ACE-031. ACE-031 contains an ACVR2B ECD fused to the Fc domain of human IgG1 and is described in detail in Sako et al., (2010) J. Biol. Chem. 285(27): 21037-21048, incorporated herein by reference in its entirety.

Examples of fusion proteins with ACVR1 ECD are known, such as those disclosed in Berasi, et al. (2011) Growth Factors, 29(4): 128-139; incorporated herein by reference in its entirety.

C. ECD Fusion Proteins

Fusion protein antagonists with fusion or multispecific ECD are also provided. Fusion ECD proteins may comprise a combination of two or more ACVR1, ACVR2A and/or ACVR2B ECDs. For example, fusion proteins may contain 1, 2, 3, 4 or more ACVR1, ACVR2A and/or ACVR2B ECD molecules. In one embodiment, the antagonist comprises an ACVR2A ECD coupled to an ACVR2B ECD. In a further embodiment, the antagonist further comprises an Fc domain.

In one embodiment, the fusion protein may comprise one or more ACVR2A ECDs and one or more ACVR2B ECDs. In another embodiment, the fusion protein may comprise one or more ACVR1 ECDs and one or more ACVR2A ECDs. In another embodiment, the fusion protein may comprise one or more ACVR1 ECDs and one or more ACVR2B ECDs.

In one embodiment, the fusion protein comprises one or more ACVR2A and Fc ECD fusion proteins and one or more ACVR2B and Fc ECD fusion proteins that are complexed together. In another embodiment, the fusion protein comprises one or more ACVR1 and Fc ECD fusion proteins and one or more ACVR2A and Fc ECD fusion proteins that are linked together in a complex. In another embodiment, the fusion protein comprises one or more ACVR1 and Fc ECD fusion proteins and one or more ACVR2B and Fc ECD fusion proteins that are linked together in a complex. In such cases, the fusion proteins can be linked together via their Fc domains, for example, via at least one disulfide bond or linker sequence. In an alternative embodiment, the ECD portions of the fusion proteins may be joined together by a linker sequence.

In one embodiment, the antagonist comprises an ACVR2A ECD fused to a first Fc domain and an ACVR2B ECD fused to a second Fc domain. In such cases, the Fc domains may be complexed with each other. In other embodiments, the antagonist comprises a linker between an ACVR2A ECD and an ACVR2B ECD, each of which is fused to an Fc domain.

Fusion proteins can be designed in such a way as to form ACVR1, ACVR2A and/or ACVR2B antagonists in tandem format. In one embodiment, the fusion protein contains two or more ACVR1, ACVR2A and/or ACVR2B ECDs in tandem followed by an Fc domain. In some cases, the ECDs organized in tandem are separated by a linker sequence. Such a tandem fusion protein may contain 1, 2, 3, 4 or more ACVR1, ACVR2A and/or ACVR2B ECDs.

D. Antagonist antibodies

Antagonists of ACVR1, ACVR2A or ACVR2B include antibodies to (in other words specifically binding to) any of these receptors, preferably antibodies containing an epitope within the extracellular domain. Specific binding of an antibody or fusion protein to its target antigen means an affinity of at least 106, 10 ⁷ , 108, 109 or 10 ¹⁰ M ^-1 . Specific binding is detectably higher in magnitude and distinct from non-specific binding to at least one foreign target. Methods for producing antibodies are known to those skilled in the art. See, for example, Kohler & Milstein (1975) Nature 256: 495-497 and Harlow & Lane (1988) Antibodies: a Laboratory Manual, Cold Spring Harbor Lab., Cold Spring Harbor, NY.

Any antibody that inhibits or reduces the activity of ACVR1, ACVR2A and/or ACVR2B can be used (eg, an antagonist antibody). Such antibodies for ACVR2A and ACVR2B include, for example, those disclosed in US Pat. No. 8,486,403; US Pat. No. 8,128,933; (2014) Mol. Cell biol. 34(4): 606-618, each of these documents is incorporated herein by reference in its entirety. Includes humanized, chimeric and ve

- 7 039118 nated forms of any of the listed antibodies, as well as antibodies competing for binding to them.

In one embodiment, the antibody is an anti-ACVR2A antibody. In another embodiment, the antibody is an anti-ACVR2B antibody. In other embodiments, the antibody may be a bispecific antibody to ACVR2A and ACVR2B. In another embodiment, the antibody is an anti-ACVR1 antibody. In other embodiments, the antibody may be a bispecific antibody to ACVR1 and ACVR2A, or to ACVR1 and ACVR2B.

The term antibody includes intact antibodies with two pairs of heavy and light chains, fragments of an antibody that can bind an antigen (for example, Fab, F(ab') ₂ , Fv, single chain antibodies, diabodies, antibody chimeras, hybrid antibodies, bispecific antibodies, humanized antibodies etc.), as well as recombinant peptides containing the above.

The term antibody fragments refers to a portion of an intact antibody, preferably the antigen-binding or variable region of an intact antibody. Examples of antibody fragments include Fab, F(ab') ₂ and Fv fragments; diabody; linear antibodies (Zapata et al. (1995) Protein Eng. 10: 10571062); single chain antibody molecules; and multispecific antibodies formed from antibody fragments.

The antibody may be monoclonal or polyclonal. The term monoclonal antibody refers to an antibody derived from a population of substantially homogeneous antibodies, i.e., the individual antibodies that make up the population are identical, apart from possible natural mutations that may be present in small amounts. Monoclonal antibodies are often highly specific, directed against a single antigenic site. In addition, unlike conventional (polyclonal) antibody preparations, which typically contain different antibodies directed against different determinants (epitopes), each monoclonal antibody is generally directed against a single determinant in the antigen. The definition of monoclonal refers to the characterization of an antibody that has been derived from a substantially homogeneous population of antibodies, such as those produced by a clonal population of B cells, and does not require the production of the antibody by any particular method.

Monoclonal antibodies to be used in accordance with the methods described herein can be obtained using the hybridoma method first described by Kohler et al. (1975) Nature 256:495, or modifications thereof. Typically, an animal, such as a mouse, is immunized with a solution containing an antigen (eg, an ACVR1, ACVR2A and/or ACVR2B polypeptide or Activin A, in particular the extracellular domain (in receptors) or part thereof).

Immunization can be performed by mixing or emulsifying the antigen-containing solution in saline, preferably Freund's complete adjuvant, and then parenteral injection of the mixture or emulsion. After the animal is immunized, its spleen (and additionally several large lymph nodes) are removed and separated into individual cells. Spleen cell screening can be performed by applying the cell suspension to a plate or to wells coated with an antigen of interest. B cells expressing membrane-bound immunoglobulin specific for the antigen bind to the plate and are not washed away. The resulting B cells or all dissociated spleen cells are then induced to fuse with myeloma cells to form hybridomas and cultured in a selective medium. The resulting cells are plated in serial dilutions and analyzed for the production of antibodies that specifically bind to the antigen of interest (and that do not bind to foreign antigens). Selected monoclonal antibody (mAb) secreting hybridomas are then cultured in vitro (eg, in tissue culture flasks or hollow fiber fermenters) or in vivo (as in mouse ascites).

In an alternative embodiment, monoclonal antibodies can be obtained using recombinant DNA methods (see, for example, US patent No. 4816567). Monoclonal antibodies can also be isolated from antibody phage libraries using the techniques described, for example, in Clackson et al. (1991) Nature 352:624-628; Marks et al. (1991) J. Mol. Biol. 222:581-597 and US Pat. No. 5,514,548.

The term antibodies includes chimeric, venated, humanized and human monoclonal antibodies to any of ACVR1, ACVR2A, ACVR2B and Activin A as defined above.

Depending on the amino acid sequence of the constant domain of their heavy chains, immunoglobulins can be divided into different classes. There are five main classes of immunoglobulins: IgA, IgD, IgE, IgG and IgM, and some of them can be further divided into subclasses (isotypes), such as IgG1, IgG ₂ , IgG ₃ , IgG ₄ , IgA1 and IgA ₂ . The heavy chain constant domains that correspond to the various classes of immunoglobulins are referred to as alpha, delta, epsilon, gamma, and mu, respectively. The subunit structures and three-dimensional configurations of the various classes of immunoglobulins are well known.

- 8 039118

As used herein, the monoclonal antibodies or Fc fusion proteins can be from any of a variety of classes of antibodies. In one embodiment, the monoclonal antibody is an IgG class antibody. In other embodiments, the monoclonal antibody may be of the IgM, IgE, IgD, or IgA classes. In specific embodiments, the antibody is an IgG isotype, eg IgG1, IgG2, IgG ₃ or IgG ₄ , in particular human IgG1, IgG2, IgG ₃ or IgG ₄ .

One or more amino acids at the amino or carboxyl terminus of the light and/or heavy chain, such as lysine at the C-terminus of the heavy chain, may be absent or present in a substituted form in some or all molecules. Substitutions can be made at constant sites to reduce or enhance effector function, such as complement-mediated cytotoxicity or ADCC (see, for example, Winter et al., US Pat. No. 5,624,821; Tso et al., US Pat. Proc. Natl. Acad. Sci. USA 103: 4005, 2006), or to increase half-life in humans (see, for example, Hinton et al., J. Biol. Chem. 279:6213, 2004). Examples of substitutions include Gln at position 250 and/or Leu at position 428 (EC numbering) to increase the half-life of the antibody. A substitution at any of positions 234, 235, 236 and/or 237 reduces the affinity for Fcy receptors, in particular the FcyRI receptor (see, for example, US Pat. No. 6,624,821). Possible substitutions at positions 234, 236 and/or 237 in human IgG2 with alanine and at position 235 with glutamine. (See, for example, US 5,624,821). Effector functions can also be weakened by replacing EFLG at positions 232-236 with PVA (see WO 14/121087). It is possible to change the S at position 428 to P, in particular in human IgG ₄ , to reduce the exchange between endogenous and exogenous immunoglobulins. Other options may include the addition or removal of post-translational modification sites, such as N-linked glycosylation at NXS/T motifs. Variants may also include the introduction of overhangs (i.e., replacing one or more amino acids with larger amino acids) or pits (i.e., replacing one or more amino acids with smaller amino acids) to promote the formation of heterodimers between different heavy chains to generate bispecific antibodies. Examples of ridge and trough pair substitutions include T336Y and Y407T, respectively (Ridgeway et al., Protein Engineering vol. 9 no. 7 pp. 617-621, 1996). Variants may also include mutations that impair interaction with the A protein (eg H435R and Y436F) in the EC numbering. Bispecific antibodies in which one heavy chain contains these variants and the other does not can be separated from their parent antibodies by protein A affinity chromatography.

Antibodies may also include antibodies that specifically bind to activin A. Such antibodies can specifically bind to any or all of the beta A beta A, beta A beta B, beta A beta C, and beta A beta E forms of activin A. Some antibodies specifically bind to only one of the listed forms (i.e. beta A beta A, beta A beta B, beta A beta C or beta A beta E). Specificity for beta A beta A, beta A beta B, beta A beta C, and beta A beta E may be determined by an epitope in the beta B, beta C, or beta E subunit, respectively, or by an epitope that involves both components of the heterodimer. Beta epitope specificity A beta can be determined by the epitope that both molecules participate in in the homodimer (eg, at a subunit interface). Some antibodies specifically bind to all such forms of activin A, in which case the epitope is typically located on the beta A subunit. Antibodies typically contain epitopes in the mature polypeptide component of precursor proteins. Some antibodies specifically bind to any or all forms of activin A without binding to human inhibin, which exists as alpha (Swiss Prot P05111) beta A or alpha beta B heterodimers. Some antibodies specifically bind to any or all of the forms of activin A and bind to either or both forms of human inhibin. While such antibodies are generally believed to inhibit activin A signaling through one or more of its counter-receptors, ACVR2A and/or ACVR2B and/or BMPR2, a precise understanding of the mechanism is not necessary for the use of such antibodies in methods of treating FOP.

Сообщалось о большом числе антител к активину А. Например, в US 2015/00373339 раскрываются человеческие антитела, обозначенные как Н4Н10423Р, Н4Н10424Р, Н4Н10426Р, Н4Н10429Р, Н4Н10430Р, Н4Н10432Р2, Н4Н10433Р2, Н4Н10436Р2, Н4Н10437Р2, Н4Н10438Р2, Н4Н10440Р2, Н4Н10442Р2, Н4Н10445Р2, Н4Н10446Р2, H4H10447R2, H4H10447R2, H4H10448R2, H4H10452R2. US Pat. No. 8,309,082 discloses human antibodies A1-A14. Mouse antibodies to Activin A are available from several commercial suppliers, such as MAB3381 from R&D Systems or 9H16 from Novus Biologicals or MM0074-7L18 (ab89307) AbCam.

Preferred antibodies have an affinity for activin A (measured at 25°C, as in example 3 of US2015/00373339) of at least 108 M ^-1 , 109 M ^-1 , 10 ¹⁰ M ^-1 , 10 ¹¹ M ^-1 , 10 ¹² M ^-1 or 10 ¹³ M ^-1 . Some antibodies have an affinity in the range of 109-10 ¹² M ^-1 . Preferred antibodies inhibit Activin A signaling with an IC ₅₀ of less than 4 nM and preferably less than 400 or 40 pM. Some antibodies inhibit signal transduction with an IC ₅₀ ranging from 4 nM to 10 pM or 3.5 nM to 35 pM.

- 9 039118

Signal transmission inhibition can be measured in the same manner as in Example 6 of US20150037339, which is summarized as follows. The A204 human rhabdomyosarcoma cell line was transfected with the Smad 2/3-luciferase reporter plasmid to obtain the A204/CAGAx12-Luc cell line. A204/CAGAx12-Luc cells were maintained in McCoy's 5A medium supplemented with 10% fetal bovine serum, penicillin/streptomycin/glutamine, and 250 μg/ml G418. For bioassay, A204/CAGAx12-Luc cells were seeded in a 96-well assay plate at 10,000 cells per well in low serum medium, 0.5% FBS and OPTIMEM, and incubated at 37°C and 5% CO ₂ for night. Activin A was serially diluted 1:3 from 100 to 0.002 nM and added to the cells, starting with a control containing no activin. Antibodies were serially diluted 1:3 ranging from 100 to 0.002 nM, 1000 to 0.02 nM, or 300 to 0.005 nM, including controls either containing the appropriate isotype control antibody or without antibody, and added to cells with constant concentration of 100 pM activin A.

Some antibodies inhibited activin A binding to ACVR2A and/or ACVR2B and/or BMPR2 by at least 1, 5, 1o, 20, 30, 40, 50, 60, 70, 80, 90, 95, 99%, which was evaluated by the results of receptor expression by the cell or by fusion of the extracellular domain with the Fc domain as a fusion protein and immobilization of the fusion protein on a substrate (eg, Biacore sensor chip). In such measurements, the antibody and activin A should be present in equimolar amounts, and the receptor or extracellular domain in excess.

Some antibodies bind to an epitope within full length Activin A residues 321-343 or 391-421, corresponding to C11-S33 and C81-E111 of the mature protein.

An example of an antibody that is used in the examples presented is the antibody designated H4H10446P in US2015037339. Its heavy chain variable region and heavy chain CDR1, CDR2 and CDR3 have the amino acid sequences of SEQ ID NOS: 162, 164, 166 and 168, respectively, in US2015/00373339 (SEQ ID NOS: 1-4, respectively, herein). Its light chain variable region and light chain CDRL1, CDRL2 and CDRL3 have the amino acid sequences of SEQ ID NOS: 146, 148, 150 and 152, respectively, in US2015/0037339 (SEQ ID NOS: 5-8, respectively, herein). H4H10446P inhibits activin A-mediated signaling through ACVR2A and/or ACVR2B, but does not markedly or at all inhibit activin A binding to ACRIIA or ACVR2B. Also included are other antibodies that compete with H4H10446P for binding to human activin A or binding to the same epitope on human activin A as H4H10446P, and include its involvement in signal inhibition.

Another example of an antibody that is used in the present methods is the antibody designated H4H10430P in US 2015037339. Its heavy chain variable region and heavy chain CDRH1, CDRH2 and CDRH3 have the amino acid sequences of SEQ ID NOS: 66, 68, 70 and 72, respectively, in US 2015/00373339 (SEQ ID NO: 9-12, respectively, herein). Its light chain variable region and light chain CDRL1, CDRL2 and CDRL3 have the amino acid sequences of SEQ ID NOS: 74, 76, 78 and 80, respectively, in US 2015/0037339 (SEQ ID NOS: 13-16, respectively, herein). Such an antibody inhibits Activin A binding to ACVR2A and/or ACVR2B and inhibits signal transduction through one or both of these receptors. Also included are other antibodies that compete with H4H10430P for binding to human activin A or binding to the same epitope on human activin A as H4H10430P, and also include its property to inhibit activin A binding to and signaling through ACVR2A and ACVR2B.

Another example of an antibody for use in the methods of the present invention are the antibodies designated A1 in US Pat. 18 respectively in this document). Its light chain CDRs CDRL1, CDRL2, and CDRL3 have the sequences of SEQ ID NOS: 11, 12, and 13, respectively, in US Pat. have the sequences of SEQ ID NOs: 62, 63 and 64, respectively, in US Pat. No. 8,309,082 (SEQ ID NOs: 22-24, respectively, herein). This also includes other antibodies that compete with H4H10430P for binding to human activin A or binding to the same epitope on human activin A as H4H10430P, and also includes its ability to inhibit activin A binding to and/or signaling through them. .

Other antibodies can be obtained by mutagenesis of cDNA encoding the heavy and light chains of any of the antibodies mentioned above. Also included within the scope of the present invention are monoclonal antibodies that are at least 90%, 95%, or 99% identical to any of the antibodies mentioned above in terms of the amino acid sequence of the mature heavy and/or light chain variable regions and retain its functional properties and/or that differ from corresponding antibody by a small number of functionally inconsistent amino acid substitutions (eg conservative substitutions), deletions or insertions. Also included are monoclonal antibodies in which at least 1, 2, 3, 4, 5, and preferably all six CDRs are 90%, 95%, 99%, or 100% identical to

- 10 039118 corresponding CDR of any of the antibodies given as an example. CDRs preferably correspond to the Kabat definition, but can be determined using any alternative definition, such as Chothia, composite Kabat-Chothia, contact definition or AbM definition (see web link bioinf.org.uk/abs).

E. Small molecule antagonists

Small molecule antagonists can also act as antagonists of ACVR1, ACVR2A and ACVR2B. Such small molecule antagonists can inhibit the activity of ACVR1, ACVR2A, ACVR2B, or Activin A. Small molecule ACVR1 antagonists include, for example, LDN-212854 described by Mohedas et al. (2013) ACS Chem. Biol. 8:1291-1302, incorporated herein by reference in its entirety.

IV. Screening studies

Screening for the activity of various ACVR1, ACVR2A and/or ACVR2B antagonists and variants or fragments thereof provided herein can be performed using a variety of analytical approaches. For example, ACVR1, ACVR2A and/or ACVR2B antagonists and variants thereof may be screened for their ability to bind to ligands or bind to ACVR1, ACVR2A or ACVR2B receptors, their ability to inhibit ligand binding to an ACVR1 and/or ACVR2 polypeptide and/or their ability to inhibit the activity of ACVR1 or ACVR2 receptors.

The activity of ACVR1 or ACVR2 antagonists, or variants or fragments thereof, can be tested in vitro or by cell-based assays. In vitro binding assays and assays for assessing inhibition of receptor activity are well known. Various methods for measuring the activity of ACVR1, ACVR2A or ACVR2B antagonists are described in detail in, for example, US Pat. No. 7,842,663, which is incorporated herein by reference in its entirety.

The ability of an antagonist to modulate the formation of a complex between an ACVR1 or ACVR2 polypeptide and its binding protein can be monitored using a variety of techniques. For example, modulation of complex formation can be quantified using, for example, detectable label proteins, such as radiolabeled (eg, ³² P, ³⁵ S, ¹⁴ C, or ³ H), fluorescently labeled (eg, FITC), or enzymatically labeled ACVR1 polypeptides, or ACVR2 or their binding protein, by immunoassay or chromatographic detection.

The ability of ACVR1 or ACVR2 antagonists to inhibit signal transduction mediated by the ACVR1 or ACVR2 receptor can be monitored. For example, the effects of subsequent signaling, such as Smad activation, can be monitored using a reporter gene responsive to Smad.

Screening for the activity of ACVR1 and/or ACVR2 antagonists and variants or fragments thereof can be performed using an in vivo assay. For example, ACVR1 or ACVR2 antagonists or variants thereof can be screened for their ability to treat FOP in a mouse model of FOP (eg, ability to reduce ectopic osteogenesis). Transgenic knockin mice have been bred with a conditional allele encoding Acvr1[R206H]. Such Acvr1 ^{[R206H]COIN/+} mice are described in US 14/207,320 and PCT/US2014/026582, which are incorporated herein by reference in their entirety. Such an allele expresses the R206H variant only after activation by the Cre recombinase. This provides Cre-dependent activation of Acvr1[R206H] expression in specific tissues and at a certain point in time due to the use of different directions of Cre activation. Due to this, the resulting mice are also not susceptible to perinatal lethality, which was noted for the unregulated knockin allele Acvr1[R206H]. Activation of Acvr1[R206H] expression in young or adult mice leads to ectopic osteogenesis. For example, mice have Acvr1 ^{[R206H]COIN/+} . Gt(ROSA26)Sor ^CreERt2/+ (wherein CreERt2 is a tamoxifen-regulated recombinase (see Feil et al. (1997) Biochem Biophys Res Commun. 237(3):7527), which is introduced into the Gt(ROSA26)Sor locus and is therefore constitutively and globally expressed) FOP develops after exposure to tamoxifen. Briefly, in the absence of tamoxifen, CreERt2 is inactivated. Tamoxifen upregulates Cre expression, which then acts on Acvr1 ^{[R206H]COIN/+} to convert it to Acvr1 ^[R206H]/+ , thereby transforming the mouse genotype to mirror that of FOP patients who have ACVR1[ R206H]. The Acvr1 ^[R206H] allele expresses Acvr1[R206H], and this is sufficient to cause the development of FOP in Acvr1 ^[R206H]/+ mice; Gt(ROSA26)Sor ^CreERt2/+ . Due to this, it is possible to avoid the lethality of embryos observed in common nokin mice Acvr1 ^[R206H] , Acvr1 ^tm1Emsh (http://www.informatics.jax.org/allele/key/828153). After administration of tamoxifen, ACVR1, ACVR2A and/or ACVR2B antagonists or controls can be administered to Acvr1 ^{[R206H]COIN/+} mice; Gt(ROSA26)Sor ^CreERt2/+ and monitor animals for ectopic osteogenesis. See Chakkalakal SA, et al. (20120 An Acvr1 R206H knock-in mouse has fibrodysplasia ossificans progressiva. J Bone Miner Res. 27(8): 1746-56. Such analysis is detailed in the examples below.

V. Progressive ossifying fibrodysplasia (FOP)

FOP is a rare hereditary disease in which heterotopic ossification (HO) results in the formation of histologically and biomechanically normal bones outside

- 11 039118 skeletal areas, for example in connective tissue. Although this impairment is episodic, it is cumulative and leads to permanent disability with increasing severity.

The worldwide rate for FOP is about 1/2,000,000. There is no ethnic, racial, gender, or geographic predisposition to FOP. It is not only a highly disabling disease, but also a condition with a greatly reduced life expectancy.

Manifestations of FOP include, for example, congenital deformities of the big toe, inflammation characterized by painful swelling of the soft tissues on the head, neck and/or back with inflammation and progressive formation of heterotopic bones due to intracartilaginous ossification.

FOP may be suspected clinically based on the presence of deformities in the big toe. Diagnostic tests, such as x-rays or bone scans, can detect abnormalities in the big toe and confirm the presence of heterotopic ossification. The diagnosis of FOP can also be confirmed by the results of genetic testing, for example, by detecting the 617 G-on-A (R206H) mutation in the ACVR1 gene.

FOP is often misdiagnosed as a number of other disorders, including other conditions of heterotopic ossification. FOP should be distinguished on the basis of a differential diagnosis from disorders such as, for example, isolated congenital malformations, lymphoedema, soft tissue sarcoma, desmoid tumors, aggressive juvenile fibromatosis, juvenile deformities, isolated brachydactyly, progressive bone heteroplasia, and heterotopic ossification. The presence of congenital deformities of the big toe and painful inflammation of the soft tissues can be used to differentiate FOP from other disorders.

FOP patients have congenital deformities of the big toe but otherwise appear healthy at birth. Inflammations associated with FOP begin to appear within the first decade of life. Inflammations can be caused, for example, by soft tissue trauma, falls, fatigue, viral infections, or intramuscular injections. Inflammation results in the transformation of soft tissues, such as ligaments, skeletal muscles, or tendons, into heterotopic bone.

Therapeutic treatments for FOP have not been available in the past. Preventive measures have been used to contain the development of FOP, such as improving safety and minimizing injury strategies, avoiding intramuscular injections, and focusing on dental procedures. Taking high doses of corticosteroids during the first 24 hours after inflammation may help reduce inflammation and swelling associated with inflammation. Surgical strategies to remove heterotopic bone are not recommended because they are counterproductive and cause new trauma-related heterotopic ossification.

FOP is caused by mutations in ACVR1 (also known as ALK2) that appear to destabilize the interaction of the GS domain with the FKBP12 inhibitor molecule (Groppe, J., et al., 2011, Cells Tissues Organs, 194: 291-295). FKBP12 is a negative modulator of ACVR1 and results in stabilization of the receptor in an inactivated conformation (Huse, M., et al., 1999, Cell, 96:425-436). See, for example, Kaplan, F.S., et al., 2012, Disease Models & Mechanisms, 5:756-762).

An example of a mutation in ACVR1 that is associated with FOP is the arginine 206 to histidine (R206H) mutation in the intracellular domain.

Subjects at risk of developing FOP include any subject with an ACVR1 R206H mutation or other FOP-associated mutation, subjects born with deformities of the big toe, or subjects with a family history of FOP who have not yet shown symptoms of FOP sufficient to diagnose FOP according to criteria recognized by experts in the field.

VI. Methods of treatment

Provided herein are methods of treating FOP comprising administering to a subject with FOP an effective regimen for administering an ACVR1, ACVR2A and/or ACVR2B antagonist. In one embodiment, an effective regimen for administering an ACVR2A antagonist and an ACVR2B antagonist is used. In a further embodiment, the ACVR2A antagonist is an Fc fusion protein and the ACVR2B antagonist is an Fc fusion protein. In another embodiment of the invention, FOP treatment is carried out as part of an effective regimen for administering an anti-Activin A antibody.

By treating a subject with FOP, it is meant to administer an effective treatment regimen administering an ACVR1, ACVR2A and/or ACVR2B antagonist or an anti-Activin A antibody to a subject with FOP to treat, cure, alleviate, alleviate, reverse, eliminate, improve, normalize, or affect the condition of one or more FOP symptoms.

By subject is meant any animal (i.e., mammals), such as humans, primates, rodents, such as mice and rats, agricultural and domesticated animals, such as dogs, cats, cattle, horses, pigs, sheep, etc., in which there is need for FOP treatment. In any of the methods presented, the subject may be a mammal, and preferably a human.

- 12 039118

An effective treatment regimen with an activin A, ACVR1, ACVR2A and/or ACVR2B antagonist or anti-Activin A antibody means using a combination of dose, frequency, and route of administration of the antagonist that provides a positive response to at least one sign or symptom of FOP. A positive response may include reducing, eliminating, alleviating, stopping the worsening, or delaying the worsening of at least one sign or symptom of FOP. Signs or symptoms of FOP that may respond positively include, for example, ectopic or heterotopic osteogenesis, inflammation of the FOP, or pain and swelling associated with inflammation. The scheme can be evaluated on the example of a single patient when comparing signs and symptoms before and after treatment. The regimen is considered effective if at least one sign or symptom shows a positive response after treatment. Alternatively or additionally, the regimen can be assessed by comparing the signs and symptoms of a population of subjects treated with one or another antagonist or antagonists of the present invention with a control population of subjects not receiving treatment. Subjects for such comparison may be an animal model or human subjects in a clinical trial (eg, phase I, phase II, IIa, IIb, or III). A regimen is considered effective if there is a statistically significant positive response between populations for at least one sign or symptom.

In some treatments for FOP, the subject does not develop other conditions that can be treated with ACVR1, ACVR2A and/or ACVR2B antagonists or anti-Activin A antibodies and is not at risk of developing them. For example, a subject may be free of all or any of the following diseases: type 2 diabetes, muscular dystrophy, amyotrophic lateral sclerosis (ALS), and osteoporosis.

A. Methods of administration

ACVR1, ACVR2A and/or ACVR2B antagonists or an anti-Activin A antibody are usually administered directly in the form of proteins or small molecules, but in the case of proteins, may also be administered in the form of a nucleic acid encoding such proteins. Such antagonists can be administered in a variety of ways, for example, by cell transfection, gene therapy, direct administration with a delivery vehicle or pharmaceutically acceptable carrier, indirect delivery using recombinant cells containing a nucleic acid encoding an ACVR1, ACVR2A and/or ACVR2B antagonist, or an anti-Activin A antibody. given in this document.

Various delivery systems can be used to administer the ACVR1, ACVR2A and/or ACVR2B antagonists or anti-Activin A antibodies provided herein, e.g. , Wu and Wu, 1987, J Biol Chem 262: 4429-4432), a nucleic acid construct as part of a retroviral or other vector, etc.

Routes of administration include enteral or parenteral and include intradermal, intramuscular, intraperitoneal, intravenous, subcutaneous, pulmonary, intranasal, intraocular, epidural and oral routes of administration. The compounds can be administered by any suitable route, for example by infusion or bolus injection, by absorption through epithelial or mucocutaneous membranes (eg, oral mucosa, rectal and intestinal mucosa, etc.) and can be used in conjunction with other biologically active agents. The introduction can be systemic or local. In addition, it may be desirable to administer the pharmaceutical compositions of the present invention to the central nervous system by any convenient means, including intraventricular injection and injection into the spinal canal region; wherein the intraventricular injection may be facilitated by the use of an intraventricular catheter, for example connected to a reservoir, such as an Omcana reservoir. Pulmonary delivery may also be used, for example using an inhaler or aerosol device, as well as a formulation containing an aerosolizing agent.

Pharmaceutical compositions of the present invention may be administered locally to the area in need of treatment; this can be achieved by, for example, local infusion during surgery, topical application, for example, by injection, by means of a catheter, or by using an implant, said implant being made of a porous, non-porous or gelatinous material, including membranes, such as silastic membranes, fibers or commercial leather substitutes.

In another embodiment, the active substance can be delivered in a vesicle, in particular in a liposome (see Langer (1990) Science 249: 1527-1533). In another embodiment, the active agent may be delivered in a controlled release system. In one embodiment, a pump may be used (see Langer (1990) above). In another embodiment, polymeric materials can be used (see Howard et al. (1989) J. Neurosurg. 71:105). In another embodiment, where the active substance of the present invention is a nucleic acid encoding a protein, such nucleic acid may be administered in vivo to promote the expression of the protein it encodes by constructing it as part of an appropriate nucleic acid expression vector and administering it so that it became intracellular, for example, when

- 13 039118 using a retroviral vector (see, for example, US patent No. 4980286), or by direct injection, or using microparticle bombardment, or lipid coating, or cell surface receptors, or transfection agents, or introducing it in conjunction with a homeobox-like a peptide known to be incorporated into the nucleus (see, for example, Joliot et al., 1991, Proc. Natl. Acad. Sci. USA 88:1864-1868), etc. In an alternative embodiment, the nucleic acid can be administered intracellularly and include in the DNA of the host cell for expression through homologous recombination.

B. Combination therapies

The ACVR1, ACVR2A and ACVR2B antagonists or anti-Activin A antibody provided herein may be administered in combination with each other or with other drugs. In one embodiment, a method for treating FOP comprises the co-administration of an ACVR2A antagonist and an ACVR2B antagonist. In another embodiment, a method for treating FOP comprises the co-administration of ACVR1, ACVR2A, and ACVR2B antagonists. In other embodiments, an ACVR1 antagonist may be administered together with an ACVR2A antagonist and/or an ACVR2B antagonist. Antagonists of ACVR1, ACVR2A and ACVR2B may be administered in the form of a separate pharmaceutical composition or may be administered in the form of a general pharmaceutical composition containing a combination of these substances. ACVR1, ACVR2A and/or ACVR2B antagonists or an anti-Activin A antibody, either alone or in combination, may be administered in combination with one or more drugs. Combination therapies may include simultaneous or alternate administration. In addition, combination therapies may include single or continuous administration.

C. Pharmaceutical compositions

The present invention also provides pharmaceutical compositions comprising an activin A, ACVR1, ACVR2A and/or ACVR2B antagonist or anti-Activin A antibody as provided herein and a pharmaceutically acceptable carrier. The term "pharmaceutically acceptable" means approved by a regulatory agency of the federal or state government or listed in the United States Pharmacopeia or other generally accepted pharmacopoeias for use in animals, and more particularly in humans. The term carrier refers to the diluents, adjuvants, excipients, or carrier vehicle with which the therapeutic agent is administered. Pharmaceutical carriers may be sterile liquids such as water or oils, including oils derived from petroleum, vegetable, animal, or synthetic oils, such as peanut, soybean, mineral, sesame, and the like. Suitable pharmaceutical excipients include starch, glucose, lactose, sucrose, gelatin, malt, rice, flour, chalk, silica gel, sodium stearate, glycerol monostearate, talc, sodium chloride, skimmed milk powder, glycerin, propylene glycol, water, ethanol and the like. . If desired, the composition may also contain small amounts of wetting agents or emulsifiers or buffering agents to maintain the pH.

Such preparations may take the form of solutions, suspensions, emulsions, tablets, pills, capsules, powders, sustained release dosage forms, and the like. The composition may take the form of a suppository with standard binders and carriers, such as triglycerides. The oral formulation may include standard carriers such as pharmaceutical grade mannitol, lactose, starch, magnesium stearate, sodium saccharin, cellulose, magnesium carbonate, and the like. Examples of suitable pharmaceutical carriers are described in Remington's Pharmaceutical Sciences, E.W. Martin.

In one exemplary implementation, the composition of the composition is determined in accordance with standard procedures for pharmaceutical compositions prepared for intravenous administration to the human body. If necessary, the composition may also include a solubilizing agent and a local anesthetic, such as lidocaine, to reduce pain at the injection site. If the formulation is to be administered by infusion, it may be supplied with an infusion bottle containing sterile pharmaceutical grade water or saline. If the composition is to be administered by injection, an ampoule of sterile water for injection or saline may be supplied so that the ingredients can be mixed prior to administration.

The Activin A, ACVR1, ACVR2A and/or ACVR2B antagonists or anti-Activin A antibody provided herein may be formulated in neutral or salt form. Pharmaceutically acceptable salts include salts formed with free amino groups, for example obtained from hydrochloric, phosphoric, acetic, oxalic, tartaric acids, etc., and formed with free carboxyl groups, for example obtained in the presence of sodium, potassium, ammonium, calcium, iron hydroxides, isopropylamine, triethylamine, 2-ethylaminoethanol, histidine, procaine, etc.

The amount and frequency of use of an Activin A antagonist, ACVR1, ACVR2A and/or ACVR2B antagonist, or anti-Activin A antibody administered in a particular manner effective for the treatment of FOP (e.g., effective regimen) can be determined by standard clinical procedures based on the description provided. In addition, in vitro assays can be used to determine the optimal dose range. The exact dose to be used in the composition also depends on the route of administration and the severity of the condition and should be determined based on judgment.

- 14 039118 the attending physician and the condition of each subject. However, suitable dose ranges for parenteral administration, preferably intravenous or subcutaneous, are usually about 20-50,000 micrograms of active substance per kilogram of body weight. For anti-Activin A antibodies, suitable dose ranges include 1-25 mg/kg, 2-20 mg/kg, 5-15 mg/kg, 8-12 mg/kg and 10 mg/kg.

Suitable dose ranges for intranasal administration typically range from about 0.01 pg/kg body weight to 1 mg/kg body weight. Effective doses can be extrapolated from a dose-response curve obtained in in vitro animal model test systems.

The frequency of administration also varies, among other factors, depending on the severity of the condition and the half-life of the substance, but usually varies from daily to quarterly, including, for example, twice a week, weekly, biweekly, monthly, once at two months. Substances may also be administered at irregular intervals, among other factors, taking into account the condition of the patient or the decrease in the serum level of the substance below the cut-off level.

All patent applications, websites, other publications, account numbers, and the like, mentioned above or below, are hereby incorporated by reference in their entirety for any purpose equally, as if a particular individual position were indicated separately and individually so incorporated by reference. If different versions of a particular sequence are associated with an account number at different times, the version associated with the account number on the effective filing date of this application is assumed. The effective filing date means the earlier of the actual filing date or the filing date of the priority application, where reference is made to the reference number (if applicable). Similarly, if different editions of a publication, website, etc. are published at different points in time, then, unless otherwise indicated, the most recent published version on the effective date of filing is meant. Unless specifically stated otherwise, any feature, step, element, embodiment, or aspect of the present invention may be used in combination with any other of the foregoing.

Examples

Example 1 Use of ACVR2A-Fc/ACVR2B-Fc to Suppress Ectopic Osteogenesis in the FOP Mouse Model.

Mice Acvr1 ^{[R206H]COIN/+} ; Gt(ROSA26)Sor ^CreERt2/+ were protected from ectopic osteogenesis after administration of tamoxifen due to the administration of ACVR2A-Fc/ACVR2B-Fc.

Mice in a FOP model called Acvr1 ^{[R206H]COIN/+} ; Gt(ROSA26)Sor ^CreERt2/+ , tamoxifen was administered at a dose of 1 mg/mouse ip for eight days. Eleven mice received 10 mg/kg ACVR2A-Fc and 10 mg/kg ACVR2B-Fc twice a week and ten mice received 10 mg/kg control mFc twice a week for 6 weeks. Mice were examined by micro-CT in vivo at baseline and 2, 4 and 6 weeks after the start of tamoxifen administration. After 6 weeks, 9 out of 10 mice in the mFc group had ectopic osteogenesis in at least one site; in contrast, only 2 out of 11 mice in the ACVR2A-Fc and ACVR2B-Fc groups developed ectopic bone, and this bone was characterized by small size. These results are shown in Fig. one.

Example 2 Use of Small Molecule ACVR1 Kinase Inhibitors to Suppress Ectopic Osteogenesis in the FOP Mouse Model.

Mice Acvr1 ^{[R206H]COIN/+} ; Gt(ROSA26)Sor ^CreERt2/+ were protected from ectopic osteogenesis after administration of tamoxifen by administration of the ACVRl kinase inhibitor LDN-212854.

Acvr1 ^{[R206H]COIN/+} mice. Gt(ROSA26)Sor ^CreERt2/+ was administered tamoxifen at a dose of 1 mg/mouse ip for eight days. Eight mice received 3 mg/kg of the ACVR1 kinase inhibitor LDN-212854 (Mohedas et al. (2013) ACS Chem. Biol. 8:1291-1302) twice daily for 4 weeks. Eight mice received vehicle control twice daily for 4 weeks. Mice were examined by in vivo microCT at baseline and 2 and 4 weeks after the start of tamoxifen administration. After 4 weeks, 8 out of 8 mice in the vehicle control group showed ectopic osteogenesis, 6 of these mice had large ectopic bone lesions. In contrast, in the LDN-212854 treated group, 3 out of 8 mice had ectopic osteogenesis, with 3 of these mice having smaller ectopic bone lesions compared to the vehicle treated control group. These results are shown in Fig. 2.

Example 3 Use of Anti-Activin A Antibody to Suppress Ectopic Osteogenesis in a FOP Mouse Model.

mice Acvr1 ^{[R206H]COIN/+} ; Gt(ROSA26)Sor ^CreERt2/+ was administered tamoxifen at a dose of 1 mg/mouse ip for eight days. Seven mice received 25 mg/kg isotype control antibody twice a week, eight mice received 25 mg/kg anti-Activin A (H4H10446P) antibody twice a week, and eight mice received 10 mg/kg ACVR2a-Fc twice a week for 3 weeks. Treatment with these substances began simultaneously with the introduction of tamoxifen. Mice were examined using computed microtomography (micro-CT) in vivo at baseline and 2 and 3 weeks after the start of tamoxifen administration. In FIG. 3 shows that after 3 weeks, all mice in the isotype control antibody group developed ectopic bone at at least one site; on the contrary, none of the mice from the anti-activin A group at that time showed the formation of ecto

- 15 039118 peak bone tissue. Two mice in the ACVR2a-Fc group developed ectopic bone after 3 weeks.

Example 4

Mice Acvr1 ^{[R206H]COIN/+} ; Gt(ROSA26)Sor ^CreERt2/+ were protected from ectopic osteogenesis after administration of tamoxifen by administration of a blocking antibody to Activin A and Acvr2a and b.

mice Acvr1 ^{[R206H]COIN/+} ; Gt(ROSA26)Sor ^CreERt2/+ was administered tamoxifen at a dose of 40 mg/kg ip for eight days. Eight mice received 10 mg/kg isotype control antibody (REGN1945), nine mice received 10 mg/kg anti-Activin A (H4H10446P) antibody (REGN2477), and nine mice received 10 mg/kg anti-Acvr2a/Acvr2b antibody twice weekly for 6 weeks. Mice were examined by micro-CT in vivo at baseline and 2, 3 and 4 weeks after the start of tamoxifen administration. In FIG. 4 shows that after 4 weeks, 7 of 8 mice in the isotype control antibody group developed ectopic bone in at least one site; in contrast, only one mouse in the anti-Activin A antibody group and three mice in the anti-Acvr2a/Acvr2b antibody group developed ectopic bone after 4 weeks. The size of ectopic bone tissue that formed in the antibody group was smaller than in the isotype control group.

Example 5

Mice Acvr1 ^{[R206H]COIN/+} ; Gt(ROSA26)Sor ^CreERt2/+ were protected from ectopic osteogenesis after administration of tamoxifen due to the administration of an Activin A blocking antibody.

mice Acvr1 ^{[R206H]COIN/+} ; Gt(ROSA26)Sor ^CreERt2/+ was administered tamoxifen at a dose of 40 mg/kg ip for eight days. Eight mice received 25 mg/kg isotype control antibody (REGN1945), nine mice received 25 mg/kg anti-activin A antibody (H4H10446P) (REGN2477), nine mice received 10 mg/kg anti-activin A antibody (REGN2477), and nine mice received 1 mg/kg anti-Activin A antibody (REGN2477) weekly for 6 weeks. Mice were examined by micro-CT in vivo at baseline and 2, 3, 4 and 6.5 weeks after the start of tamoxifen administration. The volume of ectopic bone in each mouse was calculated from micro-CT images. In FIG. 5 shows that after 4 weeks, all mice in the isotype control antibody group developed ectopic bone in at least one site, while only 2 mice from each of the anti-Activin A antibody groups. After 6.5 weeks, the mean total ectopic bone volume in the isotype group was 65.4 mm ³ compared with 1.87 mm ³ in the 25 mg/kg group, 0.3 mm ³ in the 10 mg/kg group and 7.3 mm ³ in the group treated with 1 mg/kg anti-Activin antibody.

Claims (8)

1. A method of treating fibrodysplasia ossificans progressive (FOP), comprising administering an effective anti-Activin A antibody regimen to a subject with FOP. 2. The method of claim 1, wherein the antibody comprises the antibody heavy and light chain variable regions of SEQ ID NOS: 1 and 5, SEQ ID NOS: 9 and 13, or SEQ ID NOS: 18 and 17, respectively. 3. The method of claim 1, wherein the antibody is a chimeric, veneered, humanized, or human antibody. 4. The method of claim 1, wherein the antibody is an intact antibody. 5. The method of claim 1 wherein the antibody is a human IgG1 kappa antibody. 6. The method of claim 2, wherein the antibody is a human IgG1 kappa antibody. 7. The method of claim 2, wherein the antibody comprises heavy and light chain variable regions of the antibody SEQ ID NOS: 1 and 5. 8. The method of claim 7 wherein the antibody is a human IgG1 kappa antibody.