HK1208471B - Improved antagonist antibodies against gdf-8 and uses therefor - Google Patents

Description

Improved anti-GDF-8 antagonist antibodies and uses thereof

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims the benefit of U.S. Provisional Application No. 61/660,232, filed June 15, 2012, the contents of which are incorporated herein by reference in their entirety.

Sequence Listing Reference

The Sequence Listing, filed concurrently in computer readable form (CRF) under 37 CFR §1.821 via EFS-Web with the file name PC071914_SEQLIST_ST25.txt, is incorporated herein by reference. The electronic copy of the Sequence Listing was created on May 14, 2013 and has a file size of 71 KB.

Biological Deposits

Representative materials of the present invention were deposited with the American Type Culture Collection ("ATCC"), 10801 University Boulevard, Manassas, VA 20110-2209, USA, on June 14, 2012. Vector OGD1.0.0-HC, having ATCC accession number PTA-12980, is a polynucleotide encoding the heavy chain variable region of OGD1.0.0, and vector OGD1.0.0-LC, having ATCC accession number PTA-12981, is a polynucleotide encoding the light chain variable region of OGD1.0.0.

The deposit was made under the provisions of the Budapest Treaty on the International Recognition of the Deposit of Microorganisms for Purposes of Patent Procedure and the regulations thereunder (the "Budapest Treaty"). This guarantees that living cultures of the deposit can be maintained for 30 years from the date of deposit. The deposit will be made available to ATCC under the terms of the Budapest Treaty and subject to an agreement between Pfizer and ATCC, which guarantees permanent and unrestricted public access to progeny of the deposited cultures upon the issuance of the relevant U.S. patent or upon disclosure to the public of any U.S. or foreign patent application, whichever comes first, and guarantees access to said progeny to persons determined by the U.S. Patent and Trademark Office under 35 U.S.C. § 122 and the Rules of the Commission thereunder (including 37 CFR § 1.14, particularly with reference to 886 OG 638).

The patentees of this application have agreed that if a culture of the deposited material dies or is lost or destroyed when grown under suitable conditions, another culture of the material will be promptly replaced and notice will be given. The availability of the deposited material should not be construed as permission to practice the invention in violation of the rights granted by any governmental agency under its patent laws.

Background Art

Growth differentiation factor-8 (GDF-8), also known as myostatin, is a member of the transforming growth factor-β (TGF-β) superfamily of secreted proteins and structurally related growth factors. Members of this superfamily have growth regulatory and morphogenetic properties (Kingsley et al. (1994) Genes Dev. 8:133-46; Hoodless et al. (1998) Curr. Topics Microbiol. Immunol. 228:235-72). Human GDF-8 is synthesized as a 375-amino acid precursor protein that forms a homodimeric complex. During treatment, an amino-terminal propeptide called "negative associated peptide" (LAP) is cleaved and can remain non-covalently associated with the homodimer, forming an inactive complex called the "small latent complex" (Miyazono et al. (1988) J. Biol. Chem. 263:6407-15; Wakefield et al. (1988) J. Biol. Chem. 263:7646-54; Brown et al. (1999) Growth Factors 3:35-43; Thies et al. (2001) Growth Factors 18:251-59; Gentry et al. (1990) Biochemistry 29:6851-57; Derynck et al. (1995) Nature 316:701-05; Massague (1990) Ann. Rev. Cell Biol. 12:597-641). Proteins such as follistatin and its relatives also bind to mature GDF-8 homodimers and inhibit GDF-8 biological activity (Gamer et al. (1999) Dev. Biol. 208:222-32).

Comparison of deduced GDF-8 amino acid sequences from different species reveals that GDF-8 is highly conserved (McPherron et al. (1997) Proc. Natl. Acad. Sci. U.S.A. 94:12457-61). The sequences of human, mouse, rat, pig, and chicken GDF-8 are 100% identical in the C-terminal region, while baboon, cattle, and sheep GDF-8 differ only by three amino acids at the C-terminus. This high degree of conservation across species suggests that GDF-8 has essential physiological functions.

GDF-8 has been shown to play an important role in regulating muscle development and homeostasis by inhibiting the proliferation and differentiation of myoblasts and satellite cells (Lee and McPherron (1999) Curr. Opin. Genet. Dev. 9:604-07; McCroskery et al. (2003) J. Cell. Biol. 162:1135-47). It is expressed early in developing skeletal muscle and continues to be expressed in adult skeletal muscle, preferentially in fast-twitch muscle types. In addition, overexpression of GDF-8 in adult mice causes significant muscle loss (Zimmers et al. (2002) Science 296:1486-88). Similarly, natural mutations that inactivate the GDF-8 gene have been shown to cause hypertrophy and hyperplasia in animals and humans (Lee and McPherron (1997), supra). For example, GDF-8 knockout transgenic mice are characterized by significant hypertrophy and proliferation of skeletal muscle and altered cortical bone structure (McPherron et al. (1997) Nature 387:83-90; Hamrick et al. (2000) Bone 27:343-49). Similar increases in skeletal muscle mass are seen in cattle with natural GDF-8 mutations (Ashmore et al. (1974) Growth 38:501-07; Swatland et al. (1994) J. Anim. Sci. 38:752-57; McPherron et al., supra; Kambadur et al. (1997) Genome Res. 7:910-15). In addition, various studies have shown that increased GDF-8 expression is associated with HIV-induced muscle loss (Gonzalez-Cadavid et al. (1998) Proc. Natl. Acad Sci. U.S.A. 95:14938-43). GDF-8 has also been shown to be involved in the production of muscle-specific enzymes (e.g., creatine kinase) and myoblast proliferation (WO 00/43781).

In addition to its growth regulatory and morphogenic properties, GDF-8 is believed to be involved in a variety of other physiological processes, including glucose homeostasis during the development of type 2 diabetes, impaired glucose tolerance, metabolic syndrome (i.e., a syndrome such as Syndrome X, which involves the simultaneous occurrence of a group of health conditions that may include insulin resistance, abdominal obesity, dyslipidemia, hypertension, chronic inflammation, a thrombotic obstructive state, etc., which puts an individual at high risk for type 2 diabetes and/or heart disease), insulin resistance (e.g., resistance induced by trauma such as trauma or nitrogen dysregulation), and adipose tissue disorders (e.g., obesity, dyslipidemia, non-alcoholic fatty liver disease, etc.) (Kim et al. (2000) Biochem. Biophys. Res. Comm. 281:902-06).

A variety of human and animal disorders are associated with impaired muscle tissue function, including amyotrophic lateral sclerosis (ALS), muscular dystrophy (MD), including Duchenne muscular dystrophy, muscle atrophy, organ wasting, frailty, congestive obstructive pulmonary disease (COPD), sarcopenia, cachexia, and muscle wasting syndromes caused by other diseases and conditions. Currently, there are few reliable or effective therapies to treat these disorders. The pathology of these diseases suggests a potential role for GDF-8 signaling as a target for treatment of these diseases.

ALS is a late-onset, fatal neurodegenerative disease characterized by central nervous system degeneration and muscle atrophy. ALS typically begins with gait abnormalities and loss of flexibility, then progresses to limb and diaphragm paralysis. Although most ALS cases are sporadic and the cause is unknown, 5-10% of cases have been shown to be inherited from a dominant familial (FALS) pattern. Approximately 10-20% of FALS cases are attributed to mutations in the Cu/Zn superoxide dismutase (SOD1) gene (reviewed in Bruijn et al. (2004) Ann. Rev. Neurosci. 27:723-49). SOD1 is a heterodimeric metalloprotein that catalyzes the reaction of superoxide to hydrogen peroxide and diatomic oxygen, and since loss of SOD1 does not cause motor neuron disease (Reaume et al. (1996) Nat. Genet. 13:43-47), it is believed that toxic gain of function can induce the disease (reviewed in Bruijn et al., supra). The specific mechanism of SOD1-induced neuronal cell death is not clear and may involve alterations in axonal transport, cellular responses to misfolded proteins, mitochondrial dysfunction, and excitotoxicity (Bruijn et al., supra).

The motor neuron degeneration observed in ALS can occur through a variety of mechanisms, including impaired uptake or transport of trophic factors by motor neurons (reviewed in Holzbaur (2004) Trends Cell Biol. 14: 233-40). Therefore, ALS can be treated with therapies that rejuvenate degenerating neurons by providing an optimal survival environment. The neural environment includes non-neuronal cells such as glial and muscle cells that are innervated by motor neurons. This environment provides nutrients and growth factors that are endocytosed by neurons and transported to the cell body via retrograde axonal transport (Chao (2003) Neuron 39: 1-2; Holzbaur, supra).

FALS has been modeled in both mice and rats by overexpression of mutant SOD1 (Howland et al. (2002) Proc. Natl. Acad. Sci. U.S.A. 99: 1604-09). Transgenic mice overexpressing the G93A form of mutant SOD1 show muscle weakness and atrophy by 90 to 100 days of age and typically die around 130 days of age (Gurney et al. (1994) Science 264: 1772-75). However, potential SODG93A-induced pathology, including decreased grip strength and loss of neuromuscular junctions, is evident as early as 50 days of age (Frey et al. (2000) J. Neurosci. 20:2534-42; Fisher et al. (2004) Exp. Neuro. 185:232-40; Ligon et al. (2005) NeuroReport 16:533-36; Wooley et al. (2005) Muscle Nerve 32:43-50). Transgenic rats expressing the SODG93A mutation follow a similar degeneration time course (Howland et al., supra). Recent studies have suggested that the development of pathology is not cell autonomous, consistent with the hypothesis that the motor neuron degeneration observed in ALS occurs through multiple mechanisms, including disruption of trophic factor uptake and transport by motor neurons (see above). Clement and collaborators used chimeric mice to show that wild-type non-neuronal cells can prolong the survival of motor neurons expressing mutant SOD1 (Clement et al. (2003) Science 302: 113-17). These observations have prompted research into therapies that might slow neuronal degeneration by providing an optimal survival microenvironment. For example, treatment of SODG93A mice with direct intramuscular injection of virally expressed growth factors (including IGF-1, GDNF, and VEGF) prolonged animal survival (Kaspar et al. (2003) Science 301: 839-42; Azzouz et al. (2004) Nature 429: 413-17; Wang et al. (2002) J. Neurosci. 22: 6920-28). In addition, in the SODG93A transgenic mouse model, muscle-specific expression of a local IGF-1 specific isoform (mIGF-1) stabilizes neuromuscular junctions, promotes motor neuron survival, and delays the onset and progression of the disease, suggesting that direct effects on muscle can affect disease onset and progression in transgenic SOD1 animals (Dobrowolny et al. (2005) J. Cell Biol. 168: 193-99). An association between muscle hypermetabolism and motor neuron vulnerability has also been reported in ALS mice, supporting the hypothesis that muscle defects can contribute to disease etiology (Dupois et al. (2004) Proc. Natl. Acad Sci. U.S.A. 101: 11159-64). Therefore, promoting muscle growth should improve local support for motor neurons and thus produce therapeutic benefits.

Inhibition of myostatin expression leads to muscle hypertrophy and hyperplasia (Lee and McPherron, supra; McPherron et al., supra). Myostatin negatively regulates muscle regeneration after injury, and in GDF-8 null mice, the absence of myostatin leads to accelerated muscle regeneration (McCroskery et al., (2005) J. Cell. Sci. 118:3531-41). Myostatin neutralizing antibodies increase body weight, skeletal muscle mass, and skeletal muscle size and strength in wild-type mice (Whittemore et al. (2003) Biochem. Biophys. Res. Commun. 300:965-71) and mdx mice (a model of muscular dystrophy, Bogdanovich et al. (2002) Nature 420:418-21; Wagner et al. (2002) Ann. Neurol. 52:832-36). Furthermore, in these mice, myostatin antibodies reduced damage to the diaphragm (a muscle also targeted during ALS pathogenesis). It has been hypothesized that the effects of growth factors such as HGF on muscle may be due to inhibition of myostatin expression (McCroskery et al. (2005), supra), thereby helping to shift the "push and pull" or balance between regeneration and degeneration. Therefore, GDF-8 inhibition presents a potential pharmacological target for treating ALS, muscular dystrophy (MD), and other GDF-8-related disorders (e.g., neuromuscular disorders) where it is desirable to increase muscle mass, strength, size, etc. The availability of multiple animal models for ALS (mouse and rat) allows for testing therapeutic agents in two different species, thereby increasing confidence in therapeutic efficacy in humans.

In addition to human neuromuscular disorders, there are growth factor-related conditions associated with bone loss, such as osteoporosis and osteoarthritis, which primarily affect the elderly and/or postmenopausal women. In addition, metabolic bone diseases and conditions include low bone mass due to long-term glucocorticoid therapy, premature gonadal failure, androgen suppression, vitamin D deficiency, secondary hyperparathyroidism, nutritional deficiencies, and anorexia nervosa. Although many current therapies for these conditions work by inhibiting bone resorption, therapies that promote bone formation may be a useful alternative treatment. Because GDF-8 plays a role in bone development as well as muscle development, regulation of GDF-8 is also an excellent pharmacological target for treating bone degeneration disorders.

Among other biological effects, murine monoclonal antibodies that specifically antagonize GDF-8 have previously been described as increasing muscle mass and strength in rodent models of ALS (Holzbaur, EL et al., Myostatin inhibition slows muscle atrophy in rodent models of amyotrophic lateral sclerosis, Neurobiology of Disease (2006) 23(3): 697-707). Therefore, it is expected that mouse antibodies and their humanized counterparts will be effective in increasing muscle mass and strength in patients with ALS, as well as in patients affected by other diseases and conditions characterized by or mediated by excess GDF-8, such as those described above.

The humanized versions of the mouse anti-GDF-8 antibodies mentioned above (such as various monoclonal antibodies and other protein-based therapeutics) are difficult and expensive to manufacture because their production typically requires production in living mammalian cells. Therefore, increasing the yield of this antibody, or other antibodies with similar specificity, would allow for the production of equivalent amounts of active drug with less investment. This would have the dual benefit of reducing manufacturing costs while freeing up limited manufacturing facilities for the production of other biopharmaceuticals. Both benefits would further the goal of increasing the availability of therapeutic anti-GDF-8 antibodies and other biologics to patients. Therefore, there is a need in the art for improved forms of anti-GDF-8 antibodies that have higher yields in mammalian cells.

Summary of the Invention

The present invention provides humanized anti-GDF-8 antibodies or antigen-binding fragments thereof that can be expressed in host cells at higher levels than previous forms of these antibodies that share the same complementarity determining regions (CDRs). Compositions comprising these antibodies for use in the methods of the invention are also provided.

In certain embodiments, these antibodies have a heavy chain variable (VH) region in which CDR1 is defined by the amino acid sequence of SEQ ID NO: 10 or SEQ ID NO: 20, CDR2 is defined by the amino acid sequence of SEQ ID NO: 11 or SEQ ID NO: 21, and CDR3 is defined by the amino acid sequence of SEQ ID NO: 12, and wherein the VH region is modified such that the amino acid at Kabat position 108 is leucine rather than methionine. In other embodiments, these antibodies have a VH region comprising the same CDRs, and wherein the fourth framework region of the VH region comprises amino acids 106-116 of SEQ ID NO: 44. In other embodiments of these antibodies, the CDRs are grafted onto the human germline VH gene segment DP-47 and then linked to the JH4 human heavy chain J segment gene. In other embodiments, the VH region of these antibodies comprises the amino acid sequence of SEQ ID NO: 44.

Antibodies of the invention having a leucine at Kabat position 108 (e.g., those exemplified above) are characterized by increased expression levels compared to forms in which a methionine is present at the same position. In certain embodiments, the former form is expressed at levels greater than about 50%, 100%, 150%, 200%, 250%, 300%, 400%, 500%, 600%, 700%, 800%, 900%, 1000%, 1200%, 1400%, 1600%, 1800%, or even 2000% greater than the latter form under similar conditions.

According to other embodiments of the antibodies of the present invention, the VH region described in the previous paragraph can be paired with a light chain variable (VL) region in which CDR1 is defined by the amino acid sequence of SEQ ID NO: 13, CDR2 is defined by the amino acid sequence of SEQ ID NO: 14, and CDR3 is defined by the amino acid sequence of SEQ ID NO: 15, and wherein the amino acid at Kabat position 100 in the VL region is glycine or glutamine. In other embodiments, the VH region is paired with a VL region in which the light chain CDRs are grafted to the human germ cell VL gene segment DPK-9 and then connected to the JK4 human light chain J segment gene. According to some other embodiments of these antibodies, the previously described VH region is paired with a VL region having the previously described VL region CDRs and wherein the fourth framework region of the VL region comprises amino acids 98-107 of SEQ ID NO: 9 or SEQ ID NO: 46. In other embodiments, the previously described VH domains are paired with VL domains comprising the amino acid sequence of SEQ ID NO:9 or SEQ ID NO:46.

In some other embodiments, the VH region described above is connected to a heavy chain constant region from a human antibody subtype (including IgA, IgG, IgD, IgE or IgM). If the heavy chain constant region is from IgG, then in other embodiments, the antibody heavy chain constant region is from an IgG subtype of IgG1, IgG2, IgG3 or IgG4. The heavy chain constant region can be modified, for example, to eliminate one or more Fc domain effector functions, as exemplified in SEQ ID NO: 57. In other embodiments, the VL region described above is connected to a light chain constant region that can have a λ or κ subtype.

According to other embodiments, the VH region of SEQ ID NO:44 is linked to a modified heavy chain region of amino acid sequence SEQ ID NO:57 to produce a full-length antibody heavy chain of amino acid sequence SEQ ID NO:58. Conversely, in other embodiments, the VL region of SEQ ID NO:46 can be linked to a kappa constant light chain of amino acid sequence SEQ ID NO:17 to produce a full-length antibody light chain of amino acid sequence SEQ ID NO:59. In other embodiments, two of each full-length antibody heavy and light chain can be combined to produce an anti-GDF-8 antibody with two antigen-binding sites. In other embodiments, the antibodies can comprise fragments or derivatives of these full-length intact antibodies, including, for example, Fab, F(ab') ₂ , scFv, scFv-Fc, scFv-CH, scFab, scFv-zipper, diabodies, triabodies, tetrabodies, minibodies, Fv, and bispecific antibodies.

Antibodies of the invention can have a range of binding affinities for GDF-8, e.g., about 5000 nM or even higher, e.g., at least about 4000 nM, 3000 nM, 2000 nM, 1000 nM, 900 nM, 800 nM, 700 nM, 600 nM, 500 nM, 400 nM, 300 nM, 200 nM, 100 nM, 90 nM, 80 nM, 70 nM, 60 nM, 50 nM, 40 nM, 30 nM, 25 nM, 20 nM, 15 nM, 10 nM, 7 nM, 6 nM, 5 nM, 4 nM, 3 nM, 2 nM, 1 nM, 0.1 nM, 0.01 nM, or 0.001 nM.

The present invention also provides nucleic acids encoding anti-GDF-8 antibodies, expression vectors comprising these nucleic acid sequences, and host cells for expressing these antibodies.

The present invention also provides methods for treating or preventing muscle disorders characterized by decreased muscle mass or strength in a patient. These methods are performed by administering to a patient in need of treatment or prevention of such a disorder a therapeutically or prophylactically effective amount of a composition comprising an antibody or antigen-binding fragment thereof that specifically binds GDF-8. Antibodies useful in these methods include those described above and herein.

According to certain embodiments, the antibody compositions of the present invention can be administered in a therapeutically or prophylactically effective amount to a patient in need of treatment or prevention of a muscle disorder, including a disorder selected from the group consisting of muscular dystrophy, amyotrophy, sarcopenia, cachexia, muscle wasting syndrome, age-related loss of muscle mass or strength, and frailty. In other embodiments, the muscle disorder is cachexia caused by cancer. In other embodiments, the muscle disorder is Duchenne muscular dystrophy. In certain embodiments of the latter, administration of an anti-GDF-8 antibody can effectively improve a patient's performance on a 6-minute walk test.

In some embodiments, the compositions comprising the antibodies of the invention can also be used to treat patients with muscular dystrophy (e.g., Duchenne muscular dystrophy) by administering before, simultaneously with, or after another agent effective to treat muscular dystrophy. In certain embodiments, the agent is a glucocorticoid, such as prednisone.

Also provided are methods of increasing muscle mass or strength in a mammal by administering to the mammal a composition comprising an anti-GDF-8 antibody of the invention in an amount effective to increase muscle mass or strength in the mammal. In various embodiments, the muscle is skeletal muscle (including one or more skeletal muscles that play a role in respiration) or cardiac muscle.

Also provided are methods of treating or preventing neuromuscular disorders by administering to a subject in need thereof a therapeutically or prophylactically effective amount of an anti-GDF-8 antibody of the invention. In certain embodiments, the neuromuscular disorder to be treated or prevented is ALS.

Also provided are methods for treating or preventing metabolic disorders by administering a therapeutically or prophylactically effective amount of an anti-GDF-8 antibody of the invention to a subject in need of treatment or prevention of a metabolic disorder. In various embodiments, the metabolic disorder to be treated or prevented includes type 2 diabetes, metabolic syndrome, syndrome X, insulin resistance, and impaired glucose tolerance.

Also provided are methods for treating or preventing adipose tissue disorders by administering a therapeutically or prophylactically effective amount of an anti-GDF-8 antibody of the invention to a subject in need thereof. In various embodiments, the adipose tissue disorders to be treated or prevented include obesity and abnormally high body mass index.

Also provided are methods for treating or preventing bone loss disorders by administering a therapeutically or prophylactically effective amount of an anti-GDF-8 antibody of the invention to a subject in need of such treatment or prevention. In various embodiments, the bone loss disorders to be treated or prevented include osteoporosis, osteopenia, osteoarthritis, and osteoporosis-related fractures.

In some other embodiments, the anti-GDF-8 antibodies useful in the methods of the invention include antibodies conjugated to moieties that usefully alter their function or characteristics (e.g., but not limited to, increased serum half-life). In other embodiments, amino acid changes can be achieved for similar or other purposes.

The antibody compositions used in the methods of the invention can be prepared in various dosage forms, including but not limited to aqueous suspensions, for administration by a variety of routes, including but not limited to subcutaneous, intravenous, intramuscular, intraperitoneal, infusion, or bolus injection.

In some embodiments, an effective dosage of an anti-GDF-8 antibody of the invention ranges from 0.001 mg/kg to about 250 mg/kg, which can be given in a single administration or in multiple, spaced-apart administrations.

The present invention also provides pharmaceutical kits for use by clinicians and other personnel to assist in administering anti-GDF-8 antibody compositions to patients. In some embodiments, the kits include an anti-GDF-8 antibody of the invention in lyophilized form or in aqueous solution, a diluent (e.g., pharmaceutical-grade water or buffer), and a device for administering the anti-progastrin antibody (e.g., a syringe and needle). In other embodiments, the kits may further include a second therapeutic agent.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1A provides an alignment of the amino acid sequences of the VH regions of certain anti-GDF-8 antibodies of the present invention, including murine antibody VH regions and two humanized antibody VH regions (i.e., VH0 and VH1) generated by grafting murine heavy chain CDRs onto the human germline VH region DP-47. Also provided is an alignment of the amino acid sequences of further humanized VH regions VH2-VH5. The amino acid sequences of the heavy chain CDRs are highlighted in bold and underlined font. The amino acid at Kabat position 108 is highlighted in bold font below an asterisk.

FIG1B provides an alignment of the amino acid sequences of the VL regions of certain anti-GDF-8 antibodies of the present invention, including murine antibody VL regions and two humanized antibody VL regions (i.e., VL0 and VL1) generated by grafting murine light chain CDRs onto the human germline VL region DPK-9. Also provided is an alignment of the amino acid sequences of further humanized VL regions VL2-VL5. The amino acid sequences of the light chain CDRs are highlighted in bold and underlined font. The amino acid at Kabat position 100 is highlighted in bold font below an asterisk.

Figure 2A provides a graph showing the increase in gastrocnemius (GAS) and quadriceps (QUAD) mass from C57B1/6 mice treated with 10 mg/kg OGD1.0.0 antibody for two weeks compared to vehicle controls. Figure 2B reports the same data in Figure 2A as the percent increase in muscle mass in mice treated with OGD1.0.0 antibody relative to controls. Figure 2B also shows the percent increase in muscle mass in the extensor digitorum longus (EDL) from the same group of antibody-treated and control mice.

Figure 3 provides a graph showing the increase in total tetanic force generated by EDL muscles from C57B1/6 mice treated with 10 mg/kg OGD1.0.0 antibody for two weeks compared to vehicle controls.

Figure 4A provides a graph showing a dose-responsive increase in gastrocnemius and quadriceps muscle mass from C57B1/6 mice treated weekly with PBS vehicle and 0.3, 1, 3, 10, and 30 mg/kg OGD1.0.0 antibody for 4 weeks. The data represent muscle mass measured at the end of the 4 weeks. Figure 4B reports the same data in Figure 4A as the percent increase in gastrocnemius and quadriceps muscle mass of mice treated with OGD1.0.0 antibody relative to controls.

Figure 5A provides a graph showing a dose-responsive increase in whole body lean mass in C57B1/6 mice treated weekly with PBS vehicle and 0.3, 1, 3, 10, and 30 mg/kg OGD1.0.0 antibody for 4 weeks. The data represent the lean mass measured at the end of each week for 4 weeks. Figure 5B provides a graph showing an increase in whole body lean mass in C57B1/6 mice treated weekly with PBS vehicle and 0.3, 1, 3, 10, and 30 mg/kg OGD1.0.0 antibody for 4 weeks at the end of the 4-week study.

Figure 6A provides a graph showing the increase in whole body lean mass in mdx mice treated with 10 mg/kg OGD1.0.0 antibody once weekly for 8 weeks relative to control mdx mice administered PBS vehicle. Figure 6B provides a graph showing the increase in maximum peak grip strength in mdx mice treated with 10 mg/kg OGD1.0.0 antibody once weekly for 8 weeks relative to control mdx mice administered PBS vehicle.

Figure 7A provides a graph showing the increase in gastrocnemius and quadriceps muscle mass from mdx mice and C57B1/6 mice treated with 10 mg/kg OGD1.0.0 antibody once weekly for 8 weeks relative to control mice administered PBS vehicle. The data represent muscle mass measured at the end of the 8 weeks. Figure 7B reports the same data in Figure 7A as the percent increase in gastrocnemius and quadriceps muscle mass from mdx mice treated with OGD1.0.0 antibody relative to controls.

Figure 8 provides graphs showing dose-responsive increases in whole body and leg lean mass in cynomolgus monkeys treated with 0, 3, 10, and 30 mg/kg OGD1.0.0 antibody.

Figure 9 provides a graph showing that increases in whole body lean mass in cynomolgus monkeys treated with 10 mg/kg and 30 mg/kg OGD1.0.0 antibody persist for several weeks after discontinuation of antibody treatment.

Figure 10A provides a graph showing that the volume of supraaxial muscle covering the L3-L5 vertebrae is increased in cynomolgus monkeys treated with 10 mg/kg and 30 mg/kg OGD1.0.0 antibody for 8 weeks relative to control animals administered PBS vehicle.

Figure 11 provides 3D images of epiaxial muscle from an exemplary animal subject before and after 4 weeks of OGD1.0.0 antibody treatment.

FIG12A provides the amino acid sequence of an exemplary anti-GDF-8 antibody heavy chain containing a heavy chain variable region designated herein as VHO.

FIG12B provides the amino acid sequence of an exemplary anti-GDF-8 antibody light chain containing a light chain variable region designated herein as VLO.

Detailed description

The present invention provides improved forms of humanized anti-GDF-8 antibodies that are capable of being expressed in cells at significantly higher levels compared to previous forms of the antibodies. Thus, it is expected that the improved forms of the anti-GDF-8 antibodies described herein can be produced in larger quantities and at a lower cost of goods with the same investment compared to previous forms. The present invention also describes various methods of using the improved antibodies for treatment or prevention. Thus, in certain exemplary, non-limiting embodiments, the improved anti-GDF-8 antibodies can be used to treat muscular dystrophy, cachexia, and other conditions where increasing muscle mass or strength in a subject is expected to yield therapeutic benefit.

Antibody structure and diversity

As used herein, the term antibody refers to an intact immunoglobulin (Ig) or any antigen-binding fragment, part or portion thereof, and particularly encompasses any polypeptide comprising a complete or partial antigen-binding site and retaining at least some antigen-binding specificity. Antibodies may also refer to a combination of an antigen-binding fragment, part or portion derived from an intact antibody and another molecule (including a different antibody, a protein from the Ig superfamily, or a protein or other type of molecule not derived from the immune system). These antibody derivatives may include non-proteinaceous portions or moieties.

An antigen is a substance, protein, or other substance that can be specifically bound by an antibody. An antigen may have more than one antigenic determinant or epitope, which is the part of the antigen that is bound by an antibody.

Immunoglobulins (Ig) are heterotetrameric proteins comprising two heavy chains, each approximately 50 kDa, and two light chains, each approximately 25 kDa. Each chain comprises multiple Ig domains. Starting from the amino terminus, the heavy chain contains a single variable region (VH), followed by three or four constant regions, designated CH1, CH2, CH3, and (when present) CH4, depending on the Ig subtype. Similarly, in the light chain, a single variable region (VL) is located at the amino terminus of the polypeptide, followed by a single constant region (CL). Between the CH1 and CH2 regions is a hinge region of variable length, which, depending on the isotype, imparts flexibility to the molecule. The heavy chain carboxyl terminus of CH1 (including the hinge), CH2, CH3, and (when present) CH4 constitute the Fc region. Each variable or constant region comprises a single Ig domain.

Ig light chains are bound to Ig heavy chains via disulfide bonds, and pairs of Ig heavy chains are bound to each other via disulfide bonds. Non-covalent interactions can also help stabilize the interchain quaternary structure between heavy and light chains and between paired heavy chains. In a complete Ig molecule, the VH and VL regions of the paired heavy and light chains are positioned adjacent to each other and interact and cooperate to form an antigen binding site. Because a complete Ig molecule contains two pairs of paired heavy and light chains, i.e., two heavy chains and two light chains in total, the Ig molecule contains two antigen binding sites. The presence of the hinge region provides flexibility between the antigen binding site and the rest of the molecule.

The heavy and light chain constant regions do not directly participate in antigen recognition. However, the heavy chain, especially the Fc region, contains sequences that can interact with effector molecules and cells of the immune system, allowing the heavy chain constant region to mediate important biological functions of Ig molecules.

The variable regions of the heavy and light chains contain three spaced regions of increased amino acid variability (called hypervariable regions or complementarity determining regions (CDRs)), which vary widely between different Ig molecules compared to the more conserved framework regions (FRs) surrounding the CDRs. Starting from the amino terminus of the variable region, the sequential order and numbering of the FRs and CDRs are FR1, CDR1, FR2, CDR2, FR3, CDR3, and FR4. The framework regions are primarily responsible for determining the tertiary structure of the variable region Ig domain. In contrast, the CDRs form loops extending outward from each variable region. The CDRs of adjacent VH and VL regions cooperate to form the antigen binding surface, which is primarily responsible for defining the antigen binding specificity of a particular Ig molecule.

Researchers studying antibody structure and function have developed different schemes to identify the heavy and light chain CDRs within the amino acid sequence of any specific VH or VL region. Many of these schemes identify CDRs based on an invariant or nearly invariant pattern associated with the surrounding framework of the heavy and light chain variable regions. CDRs are then defined using numerical ranges corresponding to the positions of their constituent residues within the VH and VL regions. Because CDRs, especially the third CDR, can vary in length, schemes sometimes also use letters to define the constituent residues. One of the first of these schemes is called the Kabat numbering system, which is based on aligning subsequently known VH and VL sequences to determine the position of variable CDR subsequences within the more conserved framework regions. Other schemes for defining CDRs include the AbM numbering system and the Chothia numbering system. Other schemes are also possible. For example, CDRs can be defined as those variable region residues that contact the antigen, even if these residues do not fully conform to the more formal definitions of CDRs (e.g., the Kabat or Chothia numbering schemes). For example, see Y. Ofran et al., Automated identification of complementarity determining regions (CDRs) reveals peculiar characteristics of CDRs and B cell epitopes, J Immunol. 2008 Nov 1; 181(9): 6230-5, which is incorporated by reference. The Kabat numbering scheme and certain other antibody numbering schemes are described in more detail in, for example, Handbook of Therapeutic Antibodies (2007), Stefan Dubel, ed., Wiley-VCH Verlag GmbH & Co. KgaA, Weinheim, which is incorporated by reference.

The amino acids within the CDRs of the heavy and light chain variable regions contact residues in the antigen and are primarily responsible for defining the antibody's binding specificity for the antigen. Depending on the antibody-antigen pair being studied, all or less than all of the CDR residues may directly contact the antigen. Furthermore, certain contacts may contribute more to defining specificity and/or affinity than other contacts.

The identity of the contact residues in the antibody and antigen can be determined using x-ray crystallography or other methods known to those skilled in the art. Mutations in these contact residues will typically (but not always) negatively affect antigen binding specificity and/or affinity. In contrast, non-contact CDR residues and FR residues may be mutated without significantly affecting antigen binding specificity or affinity. Although conservative amino acid changes are expected to be more likely to preserve antigen binding specificity and affinity, the actual effect of any particular CDR or framework mutation can be determined empirically using techniques familiar to those skilled in the art.

Although the CDRs in the VH and VL regions, supported by their respective framework regions, are generally responsible for establishing antigen binding specificity and affinity, there may be exceptions. For example, in some Ig molecules, FR residues may also contribute to antigen binding, while in some other Ig molecules, one or more CDRs may not directly contact the antigen. Furthermore, in other Ig molecules, the CDRs in the isolated VH and VL regions may have significant antigen binding specificity even in the absence of the corresponding variable regions that would normally be paired with them. The ability of some isolated Ig molecule variable regions to specifically bind antigen is similar to the antigen binding specificity of shark or camelid antibodies, which contain paired heavy chains but no light chains.

Ig molecules of certain species can be classified according to different isotypes. For example, in humans, Ig isotypes include IgA, IgG, IgD, IgE, and IgM. In addition, IgA and IgG isotypes can be classified into multiple subtypes, which are respectively referred to as IgA1 and IgA2, and IgG1, IgG2, IgG3, and IgG4. Isotypes and subtypes are defined according to the differences in the amino acid sequences of the heavy chain constant regions. Therefore, different isotypes and subtypes can interact with different effector molecules on different immune cells, thereby producing different effector functions. For example, IgA molecules contribute to mucosal immunity, while IgE molecules contribute to immunity against certain parasites. The heavy chains of IgM and IgE contain four CH Ig domains arranged in series, which are numbered as CH1, CH2, CH3, and CH4 starting from the amino-terminal CH region. However, IgA, IgD, and IgG only contain three CH regions arranged in series. The light chain constant region comprises two isotypes (called kappa and lambda), which have no known biological effector functions. A native Ig molecule will have only a single light chain constant region isotype.

The genes that express the heavy and light chains of antibodies are constructed in vivo through multiple gene rearrangements called V(D)J recombination. This process is responsible for generating a large library of antigen-binding proteins from a relatively limited pool of gene sequences in the genome. More information about this process is described in Abbas, A.K., Lichtman, A.H., and Pillai, S., 2010, Cellular and Molecular Immunology, 6th edition, Chapter 8, Saunders, Philadelphia, PA, which is incorporated by reference in its entirety.

In human germline DNA, three separate loci encode the exons required to construct immunoglobulin heavy chains, kappa light chains, and lambda light chains. The heavy chain locus is located on chromosome 14, the kappa chain locus is located on chromosome 2, and the lambda chain locus is located on chromosome 22. At the 5' end of each locus are multiple variable (V) gene segments, each approximately 300 base pairs long, which encode most of the amino acids that make up the variable regions of the antibody heavy and light chains, including the first and second complementarity determining regions (CDR1 and CDR2). In humans, there are approximately 100 V genes in the heavy chain locus, approximately 35 V genes in the kappa chain locus, and approximately 30 V genes in the lambda chain locus. The V gene segments are separated from each other by introns.

Located downstream of the V segments and upstream of the constant (C) gene segments in the human heavy chain locus and the kappa light chain locus is a cluster of joining (J) segments, which are typically about 30-50 base pairs long and separated from each other and from adjacent V and C genes by non-coding sequences. The heavy chain locus contains a cluster of 6 functional J segments upstream of 9 functional C genes associated with different Ig isotypes, and the kappa light chain locus contains a cluster of 5 J segments upstream of a single C _kappa gene. The human lambda light chain locus also contains 4 functional J segments, but each is located 5' of one of the 4 corresponding functional C _lambda genes. The human heavy chain locus also contains a cluster of more than 20 diversity (D) gene segments located downstream of the V gene and upstream of the J segment cluster. Both light chain loci do not contain D gene segments.

In mature Ig light chain genes, the V region is encoded by the V and J gene segments, while in Ig heavy chains, the V region is encoded by the V, J, and D gene segments. The CDR1 and CDR2 in the heavy and light chains are encoded by the V gene segments. However, constructing CDR3 is more complicated. For the heavy chain, CDR3 is encoded by the VDJ junction comprising the D and J segments and connecting residues. Similarly, the CDR3 of the light chain is encoded by the VJ junction comprising the J segment and connecting residues.

In immature B cells, all V, D and J gene segments are located alone in the germ cell line and cannot be used to express functional Ig proteins. Instead, as B cells mature, these gene segments undergo a complex DNA rearrangement process called V (D) J recombination, which brings randomly selected heavy chain V, D and J gene segments or light chain V and J gene segments into close proximity. During the connection of the V, D and J gene segments, the molecules responsible for implementing V (D) J recombination randomly add or remove nucleotides between the segments. In this way, complete variable region exons are generated in the genome of mature B cells, which are then combined with other exons (including those encoding C regions) in the mRNA encoding functional Ig heavy and light chain proteins.

The random combination of different V, D, and J gene segments to construct V region exons, and the random addition or removal of nucleotides between connected gene segments, are two important mechanisms by which the immune system generates highly diverse antigen-binding sites. These phenomena are called combinatorial diversity and junctional diversity, respectively. Because CDR3 is formed from sequences contributed by V, D, and J segments (for heavy chains) or V and J segments (for light chains), junctional diversity explains why CDR3 is the most variable of the three CDRs and generally forms the most extensive contacts with antigen.

Because the structure of Ig molecules is essentially modular, and different regions perform different functions, it is possible to prepare fragments or derivatives of anti-GDF-8 antibodies that retain GDF-8 binding ability. These fragments or derivatives are encompassed by the term antibody as used herein. Non-limiting examples of antigen-binding fragments or derivatives prepared from Ig molecules include Fab fragments, which are monovalent fragments comprising VH, CH1, VL, and CL regions; F(ab')2, which are bivalent fragments comprising two Fabs linked to each other via a hinge region; Fd fragments, which comprise VH and CH1 regions; Fv fragments, which comprise VL and VH regions; and dAb fragments, which comprise either VH or VL regions. Another example is a single-chain Fv region (scFv), which comprises VH and VL regions arranged in tandem in a single polypeptide chain and separated by a polypeptide linker that allows the variable regions to bind and form a monovalent antigen-binding site. The single-chain Fv region can be designed such that the VH region precedes the VL region, or alternatively, such that the VL region precedes the VH region. A non-limiting example of a linker is a 15-residue (Gly ₄ Ser) ₃ peptide (SEQ ID NO: 34). Other linkers are also possible. Other fragments or derivatives include Fab', surrobodies, disulfide-stabilized Fv antibodies (dsFv), diabodies, triabodies, and single-domain antibodies (e.g., shark antibodies, camelidized antibodies, or nanobodies). Other fragments or derivatives are also possible. Antigen-binding fragments, sites, or portions (such as those described herein) can be produced recombinantly or by enzymatic or chemical cleavage of intact antibodies.

Exemplary Anti-GDF-8 Antibodies

GDF-8 refers to growth differentiation factor-8, which is a member of the TGF-β superfamily. The amino acid sequence of mature human GDF-8 is shown in SEQ ID NO: 1.

Previous studies have identified a murine monoclonal antibody that specifically binds to GDF-8 and neutralizes its biological activity. This antibody has been shown to increase muscle mass and strength in mice, including in a mouse model of amyotrophic lateral sclerosis (ALS). See WO 2007/024535, which is incorporated by reference in its entirety. The VH region of the mouse antibody has the amino acid sequence of SEQ ID NO: 3, and its VL region has the amino acid sequence of SEQ ID NO: 4. These VH and VL regions are shown in Figures 1A and 1B, respectively, with the amino acid sequences of the respective VH and VL region CDRs depicted in bold. The SEQ ID NOs associated with each VH and VL CDR according to the Kabat and AbM numbering systems are listed in Table 1 below. CDRs H1, H2, and H3, defined according to the Kabat numbering system, are designated as SEQ ID NOs: 10, 11, and 12, respectively, while CDRs L1, L2, and L3 are designated as SEQ ID NOs: 13, 14, and 15, respectively. According to the AbM numbering system, CDRs H1, H2, and H3 are designated as SEQ ID NOs: 20, 21, and 22, respectively, and CDRs L1, L2, and L3 are designated as SEQ ID NOs: 23, 24, and 25, respectively.

Table 1: Nucleic acid and amino acid sequences of the present invention identified by sequence identification number

As further explained in WO 2007/024535, murine antibodies are humanized by CDR grafting. Specifically, the murine VH region is humanized by using the human germline heavy chain variable (VH) gene DP47 (VH3-23; GenBank Accession No. AB019439) as a human acceptor framework onto which the murine VH CDRs are grafted. The amino acid sequence of DP47 (SEQ ID NO: 33) is shown in Figure 1A. The murine VL region is humanized using the human germline kappa light chain variable (VL) gene DPK9 (O12m Vk1; GenBank Accession No. X59315) as a human acceptor framework onto which the murine VL CDRs are grafted. The amino acid sequence of DPK9 (SEQ ID NO: 32) is shown in Figure 1B.

Because the DP47 and DPK9 V region sequences are derived from germline cells rather than recombinant V region genes, the humanization process also required the selection of human J gene segments to encode the amino acid sequences of each portion of the humanized VH and VL regions at the carboxyl terminus of the CDR3. As described in WO 2007/024535, humanization was accomplished using the JH3 heavy chain J segment (SEQ ID NO: 35, i.e., DP47/JH3) for the VH region and the JK1 light chain J segment (SEQ ID NO: 39, i.e., DPK9/JK1) for the VL region. In the sequence alignments shown in Figures 1A and 1B, the amino acid sequences encoded by these J segment genes appear directly after the VH CDR3 and VL CDR3, respectively.

As used herein, the VH region of the humanized anti-GDF-8 antibody constructed using DP47 and JH3 described in WO 2007/024535 is referred to as VH1 (SEQ ID NO: 7), while the VL region of the humanized anti-GDF-8 antibody constructed using DPK9 and JK1 is referred to as VL1 (SEQ ID NO: 9). Also as used herein, the humanized anti-GDF-8 antibody comprising VH1 and VL1 is referred to as OGD1.1.1. In this nomenclature, the VH region number immediately follows the antibody name "OGD1," and the VL region number immediately follows the VH region number. Thus, for example, the antibody name OGD1.0.1 would refer to an antibody having a VH0 region and a VL1 region, while the antibody name OGD1.1.0 would refer to an antibody having a VH1 region and a VL0 region. The alignment between the mouse VH and VL regions, the humanized VH1 and VL1 regions, and the amino acids encoded by the DPK9 and DP47 gene sequences is illustrated in Figures 1A and 1B.

Novel forms of humanized anti-GDF-8 antibodies are described herein that possess the surprising property of being expressed by cells at levels significantly higher than OGD1.1.1 while retaining the ability to specifically bind GDF-8 with high affinity and neutralize GDF-8 activity.

In certain embodiments of these novel antibodies, a different heavy chain J segment, JH4 (GenBank Accession No. J00256, SEQ ID NO: 37), is used after CDR3 in the VH region. Due to this change, Leu (L) replaces Met (M) at position 108 (using the Kabat numbering scheme) in the VH region compared to VH1. As used herein, this novel humanized VH region in which L appears at Kabat position 108 is referred to as VH0. The amino acid sequence of VH0 (SEQ ID NO: 44) is illustrated in the sequence alignment of Figure 1A.

Because the Kabat numbering scheme uses letters appended to certain identical numbers to indicate CDRs of varying lengths, there is not necessarily a one-to-one correspondence between the Kabat numbers of residues and their physical positions in the residue sequence of a polypeptide. Thus, Kabat position 108 of a VH region corresponds to amino acid number 111 in the amino acid sequences of SEQ ID NO: 44 (i.e., VH0) and other humanized VH regions disclosed herein.

In other embodiments of the antibodies of the invention, a different light chain J segment, JK4 (GenBank Accession No. J00242, SEQ ID NO: 41), is used after CDR3 in the VL region. Due to this change, Gly (G) replaces Gln (Q) at position 100 (using the Kabat numbering scheme) in the VL region compared to VL1. As used herein, this novel humanized VL region with G appearing at Kabat position 100 is referred to as VL0. The amino acid sequence of VL0 (SEQ ID NO: 46) is illustrated in the sequence alignment of Figure 1B.

As used herein, a humanized anti-GDF-8 antibody comprising a VH0-containing heavy chain and a VL0-containing light chain is referred to as OGD1.0.0.

The gene segments, sequences and nomenclature associated with the above-mentioned VH and VL embodiments are summarized in Table 2 below.

Table 2: Overview of humanized VH and VL regions

Humanized variable region Human receptor framework Human J-segment VH1 (Seq ID No: 7) DP47 (Seq ID No: 33) JH3 (Seq ID No: 36) VH0 (Seq ID No: 44) DP47 (Seq ID No: 33) JH4 (Seq ID No: 38) VL1 (Seq ID No: 9) DPK9 (Seq ID No: 32) JK1 (Seq ID No: 40) VL0 (Seq ID No: 46) DPK9 (Seq ID No: 32) JK4 (Seq ID No: 42)

As further described in the Examples, it has been surprisingly discovered that anti-GDF-8 antibodies comprising VH0 are expressed by cells at significantly higher levels than anti-GDF-8 antibodies comprising VH1. For example, in one non-limiting embodiment described in Example 1, it was surprisingly shown that an intact immunoglobulin comprising VH0 and VL0 (i.e., OGD1.0.0) was transiently expressed at 12-fold higher levels than a similar antibody comprising VH1 and VL1 (i.e., OGD1.1.1). As discussed in Example 2, stable expression levels were also significantly higher. Interestingly, as explored in Example 3, it was found that the enhanced expression was attributable to the presence of VH0, as enhanced expression occurred regardless of whether VH0 was paired with VL0 or VL1, but only when VL0 was paired with VH0, not VH1.

In certain embodiments, the antibodies of the invention are complete heterotetrameric Ig molecules comprising full-length heavy and light chains, wherein the heavy chain variable region is VH0 and the light chain variable region is VL0 (OGD1.0.0) or VL1 (OGD1.0.1), while in other embodiments, the antibodies are GDF-8-specific binding fragments or derivatives of these full-length antibodies.

According to some embodiments, the VH region of an antibody of the invention comprises three heavy chain CDRs (i.e., CDRH1, CDRH2, and CDRH3) set forth in the amino acid sequence of SEQ ID NO:44 or its mouse counterpart SEQ ID NO:3, wherein the amino acid at Kabat position 108 is leucine. In other embodiments, the VH region comprises the amino acid sequence of SEQ ID NO:44 (i.e., VH0). In other embodiments, the VL region of an antibody of the invention comprises three light chain CDRs (i.e., CDRL1, CDRL2, and CDRL3) set forth in the amino acid sequence of SEQ ID NO:46 or its mouse counterpart SEQ ID NO:5, wherein the amino acid at Kabat position 100 is glycine or glutamine. In other embodiments, the VL region comprises the amino acid sequence of SEQ ID NO:46 (i.e., VL0) or SEQ ID NO:48 (i.e., VL1).

In the antibodies of the present invention, the antibody heavy chain isotype can be any human Ig isotype or subtype, i.e., IgA1, IgA2, IgD, IgE, IgG1, IgG2, IgG3, IgG4, or IgM. The antibody light chain isotype can be κ or λ. In specific non-limiting embodiments, the antibody heavy chain constant region is the amino acid sequence SEQ ID NO: 19 or SEQ ID NO: 57, both of which are IgG1 subtypes. SEQ ID NO: 19 contains two substitution mutations in the hinge region that prevent binding to Fc receptors on immune cells, while SEQ ID NO: 57 contains another hinge region mutation with a similar phenotype (three mutations in total). In another specific non-limiting embodiment, the light chain CH region is the amino acid sequence SEQ ID NO: 17, which is a κ isotype.

In a specific, non-limiting embodiment of the invention, an anti-GDF-8 antibody comprises a full-length antibody heavy chain according to the amino acid sequence of SEQ ID NO: 58 and a full-length antibody light chain according to the amino acid sequence of SEQ ID NO: 59. The former sequence comprises VH0 and a heavy chain constant region of the amino acid sequence of SEQ ID NO: 57, while the latter sequence comprises VL0 and a light chain kappa constant region of the amino acid sequence of SEQ ID NO: 17. According to another exemplary, non-limiting embodiment, an anti-GDF-8 antibody of the invention comprises a complete heterotetrameric antibody (i.e., OGD1.0.0) consisting of two heavy chains and two light chains according to the amino acid sequences of SEQ ID NO: 58 and SEQ ID NO: 59.

As described above, in other embodiments, the antibodies of the present invention include antigen-binding fragments or derivatives of anti-GDF-8 immunoglobulins comprising VH0. In certain embodiments of these fragments or derivatives, VH0 can be paired with VL0 or VL1. Non-limiting examples of fragments or derivatives of the present invention include Fab', F(ab') _{2, Fab, Fv, scFv, dsFv, diabodies, triabodies, and single-domain antibodies (e.g., shark antibodies, camelized antibodies, or nanobodies) comprising VH0} . Other fragments or derivatives are also possible. Specific non-limiting examples of Ig derivatives of the present invention include SEQ ID NO:63, which is an scFv in which VL0 is arranged in tandem at the amino terminus of VH0. Another non-limiting example is SEQ ID NO:65, which is an scFv in which the V regions are inverted and VH0 is arranged in tandem at the amino terminus of VL0.

Although the antibodies of the present invention are exemplified as immunoglobulins in which the heavy chain CDRs of a murine anti-GDF-8 antibody are grafted onto the human germline VH region DP47, the humanized anti-GDF-8 antibodies of the present invention are not limited to use of only such variable regions. Thus, for example, the antibodies also include intact immunoglobulins and fragments or derivatives thereof in which the murine heavy chain CDRs (i.e., SEQ ID NOs: 10-12 or 20-22) are grafted onto a human VH region other than DP47 and further modified such that the resulting VH region polypeptide includes Leu (L) at Kabat position 108. The sequences of other human germline VH regions can be found by searching GenBank or various publicly accessible internet databases, including VBASE (http://vbase.mrc-cpe.cam.ac.uk/) or VBASE2 (http://www.vbase2.org/).

As further described in Example 10, co-crystal structures of OGD1.0.0 bound to GDF-8 and a chimeric anti-GDF-8 antibody comprising murine VH and VL regions of SEQ ID NO: 3 and SEQ ID NO: 5, respectively, were solved and used to identify contact residues in the antibody responsible for antigen binding. Using this information and as further explained in Example 11, the VH and VL regions were further humanized by mutating non-contact residues in the CDRs to match residues at the same positions in human germline variable sequences. As shown in FIG1A , the further humanized heavy chain variable regions are designated VH2, VH3, VH4, and VH5. And as shown in FIG1B , the further humanized light chain variable regions are designated VL2, VL3, VL4, and VL5.

In certain embodiments of the antibodies of the present invention, any humanized VH domain can be paired with any humanized VL domain to generate a complete anti-GDF-8 antibody, or antigen-binding fragment or derivative thereof. For example, in certain embodiments, VH0 can be paired with any of the VL domains VL0, VL1, VL2, VL3, VL4, or VL5. In other embodiments, VH1 can be paired with any of the VL domains VL0, VL1, VL2, VL3, VL4, or VL5. In other embodiments, VH2 can be paired with any of the VL domains VL0, VL1, VL2, VL3, VL4, or VL5. In other embodiments, VH3 can be paired with any of the VL domains VL0, VL1, VL2, VL3, VL4, or VL5. In other embodiments, VH4 can be paired with any of the VL domains VL0, VL1, VL2, VL3, VL4, or VL5. And in certain other embodiments, VH5 can be paired with any of the VL regions VLO, VLl, VL2, VL3, VL4, or VL5.

As explained above, mutations in non-contact residues within the CDRs and framework regions are expected to have minimal effects on GDF-8 binding specificity and/or affinity, while mutations in contact residues are expected to have greater effects. Although mutations, particularly in contact residues, can reduce binding specificity and/or affinity, in some cases, mutations may be observed to increase specificity and/or affinity for GDF-8. The actual effect of any particular mutation on specificity or affinity can be determined using techniques familiar to those skilled in the art (e.g., surface plasmon resonance or other techniques).

Based on the above principles, in certain embodiments, one, two, three, or more non-contact residues within one or more VH and/or VL CDRs or framework regions of the antibodies of the present invention may be conservatively or non-conservatively substituted with different amino acid residues while retaining significant specificity and binding affinity for GDF-8. In other embodiments, one, two, three, or more contact residues within one or more VH and/or VL CDRs may be conservatively substituted while retaining significant specificity and binding affinity for GDF-8. In other embodiments, mutation of non-contact residues or contact residues results in improved specificity and/or affinity for GDF-8.

In other embodiments of the antibodies of the invention, the amino acid sequences of the VH and/or VL regions may differ by varying percentages from the sequences explicitly recited herein and retain significant or even improved specificity and/or affinity for GDF-8. Thus, in certain embodiments, the VH region of an anti-GDF-8 antibody of the invention may differ by 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% from the amino acid sequence of VH0, VH1, VH2, VH3, VH4, or VH5. In other embodiments, the VL region of an anti-GDF-8 antibody of the invention may differ by 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% from the amino acid sequence of VL0, VH1, VH2, VH3, VH4, or VH5. And in other embodiments, the VH and VL regions of the anti-GDF-8 antibodies may differ by similar percentages to those explicitly recited herein while retaining significant or even improved GDF-8 specificity and/or binding affinity.

The antibodies of the present invention may also be derivatized, covalently modified, or conjugated to other molecules to alter their properties or improve their function. For example, but not limitation, derivatized antibodies include antibodies that have been modified by, for example, glycosylation, fucosylation, acetylation, pegylation, phosphorylation, amidation, formylation, derivatization with known protecting/blocking groups, attachment to cellular ligands or other proteins, etc.

In some embodiments, the C-terminal lysine of the heavy chain of an anti-GDF-8 antibody of the invention can be cleaved and removed. Thus, for example, in certain embodiments of the invention, an anti-GDF-8 antibody comprises a heavy chain constant region of SEQ ID NO: 19 or SEQ ID NO: 57 lacking the C-terminal lysine, or may comprise an antibody heavy chain of SEQ ID NO: 58 lacking the C-terminal lysine.

Certain modifications to the structure of anti-GDF-8 antibodies may occur naturally due to the type of cells in which they are produced. In a non-limiting example, antibody synthesis in mammalian cells (e.g., CHO cells) may result in glycosylation at one or more amino acids in the antibody chain. In an exemplary, non-limiting embodiment of an anti-GDF-8 antibody, amino acid N296 in the heavy chain is glycosylated. Other sites may also be glycosylated. It will be appreciated by those skilled in the art that antibody production in some other cell types (e.g., bacterial cells) may result in non-glycosylated antibody chains. Other types of antibody modifications may occur naturally or non-naturally via chemical or enzymatic modifications performed during or after antibody purification.

Alternatively, specific amino acids in the variable or constant regions can be altered to alter or improve function. In one non-limiting example, amino acid residues in the Fc region of an antibody can be altered to increase the serum half-life of the antibody by enhancing its binding to FcRn. See, for example, WO 2000/009560, which is incorporated herein by reference. In other non-limiting examples, antibody amino acids can be altered to reduce binding to one or more Fc receptors, complement, or other immune receptors that mediate Ig bioeffector functions. In another non-limiting example, amino acids in the CDRs, framework regions, or constant regions can be altered to increase GDF-8 binding affinity or reduce immunogenicity. In a specific non-limiting example, certain human framework residues in the VH or VL regions of an antibody of the invention can be altered back to their murine counterparts as shown in the amino acid sequences of SEQ ID NO:26 and SEQ ID NO:27, respectively.

In other embodiments, antibodies can be labeled with a detectable portion according to methods familiar to those skilled in the art and can be detected according to methods familiar to those skilled in the art. These labels can be directly or indirectly conjugated to antibodies of the present invention. Labels can be directly detectable (e.g., radionuclides or fluorescent labels) or can be indirectly detected (e.g., enzyme labels that catalyze substrates to produce directly detectable products) by their ability to generate detectable molecules. Examples of detectable labels include enzymes (e.g., horseradish peroxidase, beta-galactosidase, luciferase, alkaline phosphatase, etc.), prosthetic groups (e.g., biotin, etc.), fluorescent dyes or moieties (e.g., FITC, rhodamine, lanthanide phosphors), luminescent moieties, bioluminescent moieties, radionuclides (e.g., ³ H, ¹⁴ C, ¹⁵ N, ³⁵ S, ⁹⁰ Y, ⁹⁹ Tc, ¹¹¹ In, ¹²⁵ I, ¹³¹ I, etc.), positron emitting atoms or ions, magnetic atoms or ions, paramagnetic metal atoms or ions, or peptide epitopes that can be specifically bound by other antibodies. In some embodiments, labels can be linked using spacers of varying lengths to reduce or prevent potential steric hindrance to the antigen binding site.

The antibodies of the present invention can be expressed in culture or animals by any cell type capable of sustained expression of mammalian proteins. Non-limiting examples include human cells, mouse, rat or other rodent cells, other mammalian cells, CHO cells, yeast cells or other fungal cells, plant cells or bacterial cells. Techniques for cloning DNA encoding Ig molecules or fragments or derivatives thereof into expression vectors and subsequently transiently or stably transfecting cells with these vectors are well known in the art. Culture conditions can be varied to maximize antibody expression levels. Antibodies can also be expressed in animals using techniques familiar to those skilled in the art and then purified from milk or other body fluids. Antibodies can also be synthesized in whole or in part.

The antibodies of the present invention bind to GDF-8 with high affinity, for example, with an equilibrium dissociation constant ( _KD ) of at least about 1× ^10-6 M, 1× ^10-7 M, 1× ^10-8 , 1× ^10-9 , 1× ^10-10 , 1× ^10-11 M, or higher. The _KD of an anti-GDF-8 antibody for GDF-8 can be determined by various methods familiar to those skilled in the art. Non-limiting examples of such techniques include surface plasmon resonance (SPR) and ELISA. It is well known to those skilled in the art that, due to avidity effects, anti-GDF-8 antibodies having two or more antigen-binding sites can have an apparent binding affinity greater than that of antibody fragments having a single antigen-binding site.

Although the antibodies of the present invention are specific for GDF-8, these antibodies may also bind with high affinity to the closely related growth differentiation factor called GDF-11, depending on the epitope they recognize. Therefore, antibodies specific for GDF-8 do not necessarily exclude antibodies that can bind to the GDF-11 molecule.

As used herein, a neutralizing anti-GDF-8 antibody is an antibody that reduces the biological activity of GDF-8 compared to a nonspecific control antibody or other appropriate control. Without wishing to be bound by any particular theory of operation, at least one way in which an anti-GDF-8 antibody can neutralize a biological function mediated by GDF-8 is by preventing mature GDF-8 from binding to its high-affinity receptor (e.g., ActRIIB) or one or more of its low-affinity receptors. However, other mechanisms by which an anti-GDF-8 neutralizing antibody can interfere with the biological activity of GDF-8 are possible.

Various biological activities mediated by GDF-8 that can be reduced by the neutralizing antibodies of the present invention are known in the art. Non-limiting examples include the binding of GDF-8 to ActRIIB, which can be measured, for example, using an ELISA-based assay. Another example includes activating its cell signaling pathway by GDF-8, which can be detected, for example, using a transfected reporter gene including a so-called CAGA element. For example, see Lee et al., Regulation of muscle growth by multiple ligands signaling through activin type II receptors, PNAS (2005) 102: 18117-18122; and Thies et al., GDF-8 Propeptide Binds to GDF-8 and Antagonizes Biological Activity by Inhibiting GDF-8 Receptor Binding, Growth Factors (2001) 18: 251-59, which are incorporated by reference. Another example includes phosphorylation of SMAD proteins responsible for transferring GDF-8-mediated signals from the GDF-8 receptor on the cell surface to the nucleus. For example, see Philip et al., Regulation of GDF-8 signaling by the p38 MAPK, Cellular Signalling (2005) 17:365–375, which is incorporated by reference. Phosphorylation of SMAD proteins can be detected, for example, by quantitative Western blotting using anti-phospho-SMAD antibodies. Regulation of gene expression downstream of genes normally activated or inhibited by GDF-8 can also be detected. Another example of a GDF-8-mediated activity that can be reduced using the neutralizing antibodies of the present invention is negative regulation of muscle mass or strength. Other activities are also possible.

Neutralizing antibodies of the invention can reduce GDF-8-mediated biological activities to varying degrees, depending on variables familiar to those skilled in the art (e.g., antibody and antigen concentrations and binding affinity, among other variables). Exemplary, non-limiting percentage reductions in GDF-8-mediated biological activities caused by antibodies that bind to GDF-8 include reductions of at least about 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, 99% or more compared to a suitable control.

Inhibition of GDF-8-mediated biological activity by anti-GDF-8 antibodies can be conveniently expressed as the concentration of these antibodies that inhibits 50% of the biological activity under any of the assay conditions chosen. This concentration is also referred to as the _IC50 . In certain embodiments, the _IC50 value of the anti-GDF-8 antibodies of the invention is equal to or less than about 500 nM, 250 nM, 100 nM, 75 nM, 50 nM, 40 nM, 30 nM, 20 nM, 10 nM, 5 nM, 1 nM, 0.5 nM, 0.1 nM, or less.

secretory leader sequence

According to certain embodiments, the genes encoding the heavy and light chains of the antibodies may be provided with sequences encoding amino-terminal secretory leader peptides that direct the newly synthesized protein to the secretory compartment. Post-translational processing then removes the leader peptide, after which the mature antibody is secreted from the cell. In a specific non-limiting embodiment of the antibodies of the present invention, the VH0, VH1, VL0, and VL1 regions are provided with secretory leader peptides of 19 amino acids in length. These V regions including the leader sequences are designated by the following sequence identifier numbers: VH0 (SEQ ID NO: 50); VL0 (SEQ ID NO: 52); VH1 (SEQ ID NO: 54); and VL1 (SEQ ID NO: 56). Other secretory leader sequences may also be used. Non-limiting examples include the first 19 amino acids of the murine VH region and its leader sequence (SEQ ID NO: 29) and the first 20 amino acids of the murine VL region and its leader sequence (SEQ ID NO: 31).

Anti-GDF-8 antibody expression

As explained in more detail in the Examples, OGD1.0.0 is expressed in mammalian cells at significantly higher levels than OGD1.1.1. For example, when OGD1.0.0 and OGD1.1.1 were expressed in transiently transfected COS cells, OGD1.0.0 was expressed at approximately 12-fold the level of OGD1.1.1. Similarly, when these antibodies were expressed in stably transfected CHO cells, OGD1.0.0 was expressed at approximately 6-fold the level of OGD1.1.1. As also explained in the Examples, this difference in expression levels appears to be primarily due to the presence of VH0 rather than VH1 in the antibodies.

Thus, when expressed under similar conditions, antibodies of the present invention comprising VH0 exhibit expression higher than similar antibodies comprising VH1. For example, in certain embodiments, the expression level of an antibody comprising VH0 is at least about 1.5 times, 2 times, 3 times, 4 times, 5 times, 6 times, 7 times, 8 times, 9 times, 10 times, 11 times, 12 times, 13 times, 14 times, 15 times, 20 times, 30 times or more higher than a similar antibody comprising VH1 expressed under similar conditions. The degree of difference between the antibody expression levels can depend, for example, on the type of host cell used to express these antibodies (e.g., COS cells or CHO cells), or whether the host cell is transiently transfected or stably transfected. The comparative expression level of antibodies containing VH0 or VH1 heavy chain variable regions can vary with other growth conditions and can be determined using methods familiar to those skilled in the art.

In other embodiments, the expression level of the VH0 antibody is at least about 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 75%, 100%, 150%, 200%, 250%, 300%, 400%, 500%, 600%, 700%, 800%, 900%, 1000%, 1100%, 1200%, 1300%, 1400%, 1500%, 1600%, 1700%, 1800%, 1900%, 2000%, 3000%, 4000%, 5000% or more than a similar antibody containing VH1 expressed under similar conditions. Other differences in expression levels (e.g., differences between the percentages recited herein) are also possible.

Antibody expression levels can be measured using techniques familiar to those skilled in the art. In a non-limiting example, antibody expression levels can be measured using quantitative ELISA assays. Other quantitative assays can also be used according to the knowledge of those skilled in the art.

Nucleic acid molecules encoding anti-GDF-8 antibodies

The present invention also provides nucleic acid molecules or polynucleotides encoding anti-GDF-8 antibodies. The nucleic acid may comprise DNA or RNA in which U replaces T in the DNA nucleobase sequence. The nucleic acid may also contain modifications, such as non-standard nucleobases (e.g., 5-methylcytosine) or modified backbones (e.g., phosphorothioates). Other modifications are possible. The nucleic acid may be single-stranded or double-stranded. The nucleic acid may be obtained from a natural source, such as a cell or a whole organism. Non-limiting examples of nucleic acids of natural origin include genomic DNA, amplified plasmid DNA, or mRNA. Alternatively, the nucleic acid may be a synthetic nucleic acid. Non-limiting examples of synthetic nucleic acids include cDNA, PCR products, or nucleic acids synthesized on a nucleic acid synthesis machine.

In certain embodiments, the nucleic acids of the present invention encode an amino acid sequence comprising a VH0 antibody heavy chain or a fragment or derivative thereof. In other embodiments, the nucleic acids encode an amino acid sequence comprising a VL0 antibody light chain or a derivative or fragment thereof. In other embodiments, the nucleic acid sequences encoding the VH0 and VL regions (e.g., VL0 or VL1) are present in different or identical isolated polynucleotides.

While the present invention provides specific nucleic acid sequences encoding anti-GDF-8 antibodies, or fragments or derivatives thereof, those skilled in the art will appreciate that, due to the degeneracy of the genetic code, these sequences are exemplary only and should not be considered limiting. Thus, exemplary nucleic acid sequences encoding the VH and VL regions of a murine anti-GDF-8 antibody are SEQ ID NO:2 and SEQ ID NO:4, respectively. Exemplary nucleic acid sequences encoding VH1 are SEQ ID NO:6 and SEQ ID NO:47. Exemplary nucleic acid sequences encoding VL1 are SEQ ID NO:8 and SEQ ID NO:48. An exemplary nucleic acid sequence encoding VH0 is SEQ ID NO:43. An exemplary nucleic acid sequence encoding VL0 is SEQ ID NO:45. An exemplary nucleic acid sequence encoding the amino acid sequence of SEQ ID NO:19, comprising the CH region of an IgG1 containing two hinge region mutations, is SEQ ID NO:18. An exemplary nucleic acid sequence encoding the amino acid sequence of SEQ ID NO:17, comprising the human κCL region, is SEQ ID NO:16. Exemplary nucleic acid sequences encoding the VH and VL regions of a murine anti-GDF-8 antibody following the leader sequence are SEQ ID NO: 28 and SEQ ID NO: 30, respectively. Exemplary nucleic acid sequences encoding the amino acid sequences of the VH0, VL0, VH1, and VL1 regions following the leader sequence are SEQ ID NO: 49, SEQ ID NO: 51, SEQ ID NO: 53, and SEQ ID NO: 55, respectively.

According to certain embodiments, the nucleic acid molecule comprises a nucleic acid sequence encoding any of the following amino acids: 7, 9, 10, 11, 12, 13, 14, 15, 17, 19, 20, 21, 22, 23, 24, 25, 26, 27, 36, 38, 40, 42, 44, 46, 50, 52, 54, 56, 57, 58, 59, 63, or 65. In other embodiments, the nucleic acid molecule comprises a nucleic acid sequence that encodes an amino acid sequence that is at least 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identical to SEQ ID NO: 7, 9, 10, 11, 12, 13, 14, 15, 17, 19, 20, 21, 22, 23, 24, 25, 26, 27, 36, 38, 40, 42, 44, 46, 50, 52, 54, 56, 57, 58, 59, 63, or 65. In other embodiments, the nucleic acid molecule comprises a nucleic acid sequence that encodes an amino acid sequence that is at least 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identical to VH0 (SEQ ID NO:44) and wherein Kabat position 108 is leucine.

According to certain embodiments, the nucleic acid molecule comprises a nucleic acid sequence of any of the following SEQ ID NOs: 6, 8, 16, 18, 35, 37, 39, 41, 43, 45, 47, 48, 49, 51, 53, 55, 62, or 64. In other embodiments, the nucleic acid molecule comprises a nucleic acid sequence that is at least 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identical to the following nucleic acid sequence SEQ ID NOs: 6, 8, 16, 18, 35, 37, 39, 41, 43, 45, 47, 48, 49, 51, 53, 55, 62, or 64. In other embodiments, the nucleic acid molecule comprises a nucleic acid sequence that hybridizes under high stringency conditions to the following nucleic acid sequence SEQ ID NO: 6, 8, 16, 18, 35, 37, 39, 41, 43, 45, 47, 48, 49, 51, 53, 55, 62, or 64.

Non-limiting examples of high stringency hybridization conditions are incubation of the hybridizing nucleic acids in 1×SSC at 65° C. or in 1×SSC and 50% formamide at 42° C., followed by a wash in 0.3×SSC at 65° C. Additional examples of stringent conditions are provided in Sambrook et al., Molecular Cloning: A Laboratory Manual, Chapters 9 and 11, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY (1989), which is incorporated herein by reference.

It will be appreciated by those skilled in the art that certain nucleic acids of the present invention can be linked together in frame to create a composite nucleic acid sequence. For example, a nucleic acid encoding a VH0 region can be linked in frame with a nucleic acid encoding a CH region to create a composite nucleic acid encoding a complete heavy chain. In a non-limiting example, the nucleic acid sequence SEQ ID NO:43 encoding a VH0 region can be linked in frame with a nucleic acid sequence encoding the amino acid sequence SEQ ID NO:57 (the constant portion of a human IgG1 heavy chain containing three mutations that affect effector function). Similar linkages to create light chains are possible, as are other linkages to create other complexes of the nucleic acids described herein.

carrier

The nucleic acid of the present invention can be incorporated into a vector using techniques well known to those skilled in the art. In certain embodiments, the vector includes a plasmid, typically a bacterial plasmid, a eukaryotic episome, a yeast artificial chromosome, and a viral genome. Exemplary non-limiting viruses include retroviruses, adenoviruses, adeno-associated viruses (AAV), and plant viruses (e.g., cauliflower mosaic virus and tobacco mosaic virus). Other types of vectors are possible. In some embodiments, the vector can replicate autonomously in a suitable host. In other embodiments, the vector is maintained in the host in an extrachromosomal form, or it can be integrated into the host genome so that the vector can replicate together with the host genome. A vector comprising a gene and a control sequence sufficient to maintain the transcription and translation of the gene is referred to as an expression vector. The vector of the present invention can be selected or designed according to the knowledge of those skilled in the art to function in any cell type that can support Ig gene expression, including bacterial cells, other prokaryotic cells, yeast cells, other fungal cells, plant cells, animal cells, insect cells, mammalian cells, Chinese hamster ovary celIs (CHO) cells, and human cells or other cells.

The vector may optionally contain one or more control sequences. Certain control sequences allow replication, such as origins of replication. Other control sequences control or regulate transcription, such as promoters, enhancers, and transcription termination sites. Non-limiting examples of promoters or enhancers are those derived from retroviral LTRs, cytomegalovirus (CMV), simian virus 40 (SV40), adenovirus (e.g., adenovirus major late promoter (AdMLP)), or polyoma virus. Other examples include tissue-specific promoters and enhancers, constitutively active promoters and enhancers, inducible promoters and enhancers, Ig gene promoters and enhancers, and actin promoters and enhancers. Other promoters and enhancers are also possible.

Certain control sequences control or regulate post-transcriptional RNA processing, such as splicing and polyadenylation signals or signals that increase or decrease mRNA stability. Other control sequences control or regulate protein translation (e.g., translation initiation sequences (e.g., Kozak consensus sequences)), post-translational processing (e.g., signal peptide sequences that direct secretion of the gene product out of the host cell), or protein stability. The signal peptide sequence can be derived from an immunoglobulin or a secreted protein that is not part of the Ig superfamily. Other control sequences are also possible.

Vectors may also include selectable marker genes that allow selection of host cells that have taken up these vectors. Non-limiting examples include selectable marker genes that confer a drug resistance phenotype, such as the dihydrofolate reductase gene (DHFR) (for dhfr ^- host cells, allowing selection with methotrexate), the neo gene (allowing selection with G418 or similar drugs), the hph gene (allowing selection with hygromycin B), and the glutamate synthetase gene (allowing selection with methionine sulfonimide).

In some embodiments, the vector may contain nucleic acid sequences encoding a single Ig heavy or light chain, or antigen-binding fragment thereof, but not both chains in the same vector. Typically, expression of whole antibodies from these vectors involves introducing separate vectors containing the heavy and light chains into the same cell. In other embodiments, the vector may contain nucleic acid sequences encoding both heavy and light Ig chains, or antigen-binding fragments thereof, in the same vector.

Nucleic acid molecules of the present invention or vectors comprising these nucleic acids can be introduced into one or more types of host cells capable of supporting antibody expression. Methods for introducing nucleic acids or vectors into suitable host cells are well known to those skilled in the art. Non-limiting examples include transient and stable transfection, transformation, transduction, and viral infection of target host cells. Other examples include dextran-mediated transfection, calcium phosphate precipitation, polybrene-mediated transfection, protoplast fusion, electroporation, encapsulation of polynucleotides in liposomes, and direct microinjection of DNA into the nucleus. Exemplary non-limiting methods are discussed in, for example, U.S. Patents Nos. 4,399,216, 4,912,040, 4,740,461, and 4,959,455, which are incorporated by reference. Methods for transforming plant cells are also well known in the art, including, for example, Agrobacterium-mediated transformation, biolistic transformation, direct injection, electroporation, and viral transformation. Methods for transforming bacterial and yeast cells are also well known in the art.

host cells

In certain embodiments, nucleic acids encoding antibodies of the present invention, or fragments or derivatives thereof, are introduced into suitable host cells for expression. Cells capable of expressing antibodies include bacterial, fungal, plant, animal, and mammalian cells. Other cell types may also be used, as known to those skilled in the art.

Mammalian cells suitable as antibody expression hosts are known in the art. Exemplary non-limiting examples include some immortalized cell lines available from American Type Culture Collection (ATCC) or other sources, including Chinese hamster ovary (CHO) cells, NSO cells, SP2 cells, HEK-293T cells, NIH-3T3 cells, HeLa cells, baby hamster kidney (BHK) cells, African green monkey kidney cells (e.g., COS, CV-1 or Vero cells), human hepatocellular carcinoma cells (e.g., HepG2), A549 cells, A431 cells, HeLa cells, L cells, BHK21 cells, HL-60 cells, U937 cells, HaK cells, Jurkat cells and other cells. Other animals, insects or mammalian cells suitable as antibody expression hosts are possible.

In other embodiments, cell lines from insects, plants, bacteria or fungi can be used. Exemplary non-restrictive insect cells include Sf9 or Sf21 cells, which are typically used in conjunction with baculovirus vector expression systems. Exemplary non-restrictive plant cells include those from tobacco, Arabidopsis thaliana, duckweed, corn, wheat and potato species. Exemplary non-restrictive bacteria include Escherichia coli, Bacillus subtilis, Salmonella typhimurium and Streptomyces strains. Exemplary non-restrictive fungi include Schizosaccharomyces pombe, Saccharomyces cerevisiae, Pichia pastoris, Kluyveromyces yeast strains and Candida yeast strains. Other insects, plants, bacteria and fungal cells are possible.

Methods for growing and maintaining different types of host cells under conditions conducive to antibody expression are well known in the art. After antibody expression occurs, the antibodies thus expressed can then be purified from the host cells according to the knowledge of those skilled in the art. For example, secreted antibodies can be purified from the culture medium in which the host cells are grown. Alternatively, in some embodiments, particularly when the host cells have cell walls, the host cells can be mechanically, chemically, or enzymatically broken open to release the expressed antibodies sequestered within the cells. Exemplary, non-limiting methods for antibody purification include ion exchange chromatography, salt precipitation, and gel filtration. In other embodiments, affinity chromatography can be used. For example, a mouse antibody that recognizes a human constant region sequence can be immobilized on a purification column. Alternatively, the antibody can be expressed fused to an epitope tag or a larger affinity tag (e.g., maltose binding protein, glutathione S-transferase, and thioredoxin) to allow purification using specific antibodies or other molecules that bind tightly to the affinity tag. Thereafter, the epitope tag or affinity tag can be cleaved using techniques familiar to those skilled in the art and the antibody can be purified using other techniques (e.g., those disclosed herein). Other techniques for purifying antibodies from culture medium and host cells are also possible. According to good manufacturing practices or other regulatory requirements, the antibodies may be subjected to other processing steps to further purify the antibodies. Suitable purification methods and their implementation procedures are known to those skilled in the art.

genetically modified animals and plants

The antibodies of the present invention can also be produced in genetically modified non-human animals or plants. Expression of antibodies in these organisms can be constitutive or inducible. The antibodies expressed in these organisms can then be isolated using techniques known to those skilled in the art. Methods for expressing antibodies and antigen-binding fragments thereof in transgenic non-human organisms are well known in the art. In a non-limiting example, the antibodies of the present invention can be produced in and recovered from the milk of goats, cows, or other non-human mammals. For example, see U.S. Patents Nos. 5,827,690, 5,756,687, 5,750,172, and 5,741,957, which are incorporated herein by reference. Other non-limiting examples of transgenic mammals in which antibodies can be expressed are mice, rats, sheep, pigs, or horses. Another non-limiting example of a body fluid from which antibodies can be isolated is blood. Other body fluids are also possible. The antibodies of the present invention can also be produced in and recovered from plants. See, for example, US Patent Nos. 6,417,429, 6,046,037, and 5,959,177, which are incorporated herein by reference.

Pharmaceutical composition

For use in the therapeutic and preventive methods of the present invention, the antibodies disclosed herein can be formulated as compositions. Alternatively, the compositions can include one or more other agents therapeutically or prophylactically effective against GDF-8-mediated conditions, including antibodies that bind to different GDF-8 epitopes than those disclosed herein. Compositions will typically be supplied as part of a sterile pharmaceutical composition, which will typically include a pharmaceutically acceptable carrier. The compositions can be in any suitable form depending on the desired method of administration to the patient.

The antibodies of the present invention can be administered to a subject by a variety of routes, typically parenterally, for example, by subcutaneous, intravenous, intraperitoneal or intramuscular injection. Administration can be achieved by one or more bolus injections or by one or more infusions. Other routes of administration are also possible according to the knowledge of those skilled in the art. The most suitable route of administration in any given situation may depend on the specific composition to be administered and the characteristics of the subject, such as the condition to be treated, age or sex.

The pharmaceutical composition can be conveniently presented in a unit dosage form containing a predetermined amount of antibody per dose. The unit may contain, for example, but not limited to, 5 mg to 5 g, 10 mg to 1 g, or 20 to 50 mg. The pharmaceutically acceptable carrier for use in the present invention can be in a variety of forms depending, for example, on the route of administration.

The pharmaceutical compositions of the present invention can be prepared as lyophilized preparations or aqueous solutions for storage by mixing an antibody having the desired purity with optional pharmaceutically acceptable carriers, excipients, or stabilizers commonly used in the art (each of which is referred to herein as a "carrier," i.e., a buffer, stabilizer, preservative, isotonicity agent, nonionic detergent, antioxidant, and other various additives). See Remington's Pharmaceutical Sciences, 16th edition (Osol ed. 1980). These additives must be non-toxic to the recipient at the dosages and concentrations employed.

Buffering agents help maintain the pH within a range close to physiological conditions. They can be present at a concentration ranging from about 2 mM to about 50 mM. Buffers suitable for use in the present invention include organic and inorganic acids and salts thereof, such as citrate buffers (e.g., monosodium citrate-disodium citrate mixtures, citric acid-trisodium citrate mixtures, citric acid-monosodium citrate mixtures, etc.), succinate buffers (e.g., succinic acid-monosodium succinate mixtures, succinic acid-sodium hydroxide mixtures, succinic acid-disodium succinate mixtures, etc.), tartrate buffers (e.g., tartaric acid-sodium tartrate mixtures, tartaric acid-potassium tartrate mixtures, tartaric acid-sodium hydroxide mixtures, etc.), fumarate buffers (e.g., fumaric acid-monosodium fumarate mixtures, In some embodiments, the present invention provides a buffer solution containing at least one of the following components (e.g., a mixture of fumaric acid and disodium fumarate, a monosodium fumarate and disodium fumarate mixture, etc.), a gluconate buffer (e.g., a gluconic acid and sodium gluconate mixture, a gluconic acid and sodium hydroxide mixture, a gluconic acid and potassium gluconate mixture, etc.), an oxalate buffer (e.g., an oxalic acid and sodium oxalate mixture, an oxalic acid and sodium hydroxide mixture, an oxalic acid and potassium oxalate mixture, etc.), a lactate buffer (e.g., a lactic acid and sodium lactate mixture, a lactic acid and sodium hydroxide mixture, a lactic acid and potassium lactate mixture, etc.), and an acetate buffer (e.g., an acetic acid and sodium acetate mixture, an acetic acid and sodium hydroxide mixture, etc.). In addition, a phosphate buffer, a histidine buffer, and a trimethylamine salt (e.g., Tris) can be used.

Preservatives may be added to slow the growth of microorganisms and may be added in amounts ranging from 0.2% to 4% (w/v). Suitable preservatives for use in the present invention include phenol, benzyl alcohol, m-cresol, methylparaben, propylparaben, octadecyldimethylbenzyl ammonium chloride, benzalconium halides (e.g., benzalkonium chloride, benzalkonium bromide, and benzalkonium iodide), hexamethonium chloride, and alkylparabens (e.g., methylparaben or propylparaben), catechol, resorcinol, cyclohexanol, and 3-pentanol. Isotonicity agents, sometimes referred to as "stabilizers," may be added to ensure the isotonicity of the liquid compositions of the present invention and include polyols, such as triols, or higher sugar alcohols, such as glycerol, erythritol, arabitol, xylitol, sorbitol, and mannitol. Stabilizers refer to a broad class of excipients whose functions can range from self-swelling agents to additives that solubilize the therapeutic agent or help prevent denaturation or adhesion to the container wall. Typical stabilizers can be polyols (listed above); amino acids such as arginine, lysine, glycine, glutamine, asparagine, histidine, alanine, ornithine, L-leucine, 2-phenylalanine, glutamic acid, threonine, etc.; organic sugars or sugar alcohols such as lactose, trehalose, stachyose, mannitol, sorbitol, xylitol, ribitol, myoinisitol, galactitol, glycerol, and the like, including cyclitols such as inositol; polyethylene glycol; amino acid polymers The stabilizer may be present in an amount ranging from 0.1 parts by weight per part active protein to 10,000 parts by weight per part active protein.

A nonionic surfactant or detergent (also referred to as a "wetting agent") can be added to help dissolve the antibody and any other therapeutic agents that may be included to counteract agitation-induced aggregation, and it also allows the composition to be exposed to shear surface stress without causing protein denaturation. Suitable nonionic surfactants include polysorbates (20, 80, etc.), polyoxamers (184, 188, etc.), polyols (Pluronic polyol), polyoxyethylene sorbitan monoether (etc.). The nonionic surfactant can be present in a range of about 0.05 mg/ml to about 1.0 mg/ml, for example, about 0.07 mg/ml to about 0.2 mg/ml.

Other miscellaneous excipients may include chelating agents (eg, EDTA), antioxidants (eg, ascorbic acid, methionine, vitamin E), and co-solvents.

In an exemplary, non-limiting embodiment, an antibody of the invention is formulated in a solution comprising 20 mM L-histidine, 85 mg/ml sucrose, 0.2 mg/ml PS-80, and 0.05 mg/ml EDTA (pH 5.8). In other embodiments, the concentration of the antibody in the formulation is 100 mg/ml. In other embodiments, the antibody in the formulation is lyophilized.

Drug kit

In certain embodiments, the present invention provides pharmaceutical kits for use by clinicians or other personnel. A pharmaceutical kit is a package comprising an anti-GDF-8 antibody of the invention (e.g., in lyophilized form or in aqueous solution) and one or more of the following: at least one second therapeutic agent as described elsewhere herein; a device for administering the antibody, such as a needle and/or syringe; and, if the antibody is in lyophilized or concentrated form, pharmaceutical-grade water or buffer for resuspending or diluting the antibody. The kit may also include instructions for preparing the antibody composition and/or administering the composition to a patient.

Each unit dose of an anti-GDF-8 antibody composition can be packaged individually, and the kit can contain one or more unit doses (e.g., 2 unit doses, 3 unit doses, 4 unit doses, 5 unit doses, 7 unit doses, 8 unit doses, 10 unit doses, or more). In one embodiment, the one or more unit doses are each contained in a syringe, and in another embodiment, the one or more unit doses are each contained in a bag or similar container suitable for connection to an intravenous line.

Treatment and prevention methods

The present invention provides methods for treating and preventing conditions and disorders in which reducing GDF-8 activity directly or indirectly results in a therapeutic benefit. These methods comprise administering to a subject an effective amount of a composition comprising an anti-GDF-8 antibody. In certain of these embodiments, the antibody is OGD1.0.0 or a GDF-8-binding fragment, portion, part, or derivative thereof.

The subject to whom the anti-GDF-8 antibody composition can be administered can be a mammal, such as a non-primate (e.g., cow, pig, horse, cat, dog, rat, etc.) or a primate (e.g., monkey, chimpanzee, ape, or human). The subject can be a human, such as an adult patient or a pediatric patient.

Conditions and disorders treatable using the antibody compositions of the invention are those that are mediated, at least in part, by GDF-8 or for which there is scientific rationale suggesting that decreasing GDF-8 activity in a subject would result in therapeutic benefit.

While the therapeutic benefit will depend in part on the specific condition or disorder, a therapeutic benefit exists when reducing GDF-8 activity in a subject results in any improvement in the symptoms, signs, or severity of the condition or disorder, or halts or slows the gradual worsening of such symptoms, signs, or severity. A therapeutic benefit further exists when reducing GDF-8 activity improves the life expectancy, comfort, or quality of life of the subject. A therapeutic benefit also exists when reducing GDF-8 activity improves or halts or slows the deterioration of one or more physical or physiological functions of the subject, or the subject's performance on tests reflecting such functions.

The therapeutic benefit can be inferred by observing the subject perform certain tasks, asking the subject questions about their feelings, or performing one or more tests on the subject at the bedside, or performing one or more tests on a sample obtained from the subject in a laboratory. The therapeutic benefit can also be demonstrated by markers of GDF-8 inhibition. For example, but not limited to, a biopsy can be performed on the muscle of the treated subject and tested for the presence or absence of markers associated with the downregulation of the signaling pathway stimulated by GDF-8 (e.g., a decrease in phosphorylated SMAD2 or SMAD3 protein levels). Other tests suitable for detecting the therapeutic benefit of a subject treated with a composition comprising an antibody of the present invention are known to those skilled in the art. Although it is desired that the condition or disorder being treated or prevented is completely cured or reversed, it is not necessary for the therapeutic benefit to exist.

In certain embodiments, compositions comprising antibodies of the invention can be used to treat or prevent conditions or disorders characterized by loss of skeletal muscle mass and/or strength, or in which increasing such muscle mass and/or strength yields therapeutic benefit. In certain of these embodiments, the antibody is OGD1.0.0 or a GDF-8 binding fragment or derivative thereof.

Conditions or disorders associated with loss of muscle mass and/or strength that can be treated or prevented by administering the antibodies disclosed herein include, but are not limited to, age-related loss of muscle mass or strength, frailty, sarcopenia, and loss of muscle mass or strength caused by muscle atrophy, immobilization, or disuse (e.g., after injury, denervation, or sustained exposure to zero gravity). In other embodiments, conditions or disorders that can be treated or prevented include bone fractures, particularly in the elderly or others who are prone to bone fractures (e.g., hip fractures or other bone fractures) or are susceptible to stable joint replacement surgery. In some other embodiments, conditions or disorders that can be treated or prevented are muscle wasting syndromes, including those attributed to primary disease processes. Non-limiting examples of muscle wasting syndromes include cachexia, such as caused by cancer, anorexia, or other types of malnutrition; and muscle wasting caused by AIDS, sepsis, burns, chronic renal failure, congestive heart failure (CHF), and chronic obstructive pulmonary disease (COPD).

A therapeutic benefit to a subject administered a composition comprising an antibody of the invention can be demonstrated, for example, but not limited to, through an increase in muscle mass or strength in general or specific muscles. Non-limiting examples of muscles whose mass and/or strength can be increased by treatment with an anti-GDF-8 antibody include skeletal and cardiac muscles. Other examples include muscles that control breathing, including the diaphragm and intercostal muscles; and accessory muscles of inspiration, including the sternocleidomastoid muscles, scalenes, and others. Other examples of skeletal muscles include the gastrocnemius, tibialis posterior, soleus, tibialis anterior, longus, brevis, gluteus maximus, biceps femoris, semitendinosus, semimembranosus, iliopsoas, quadriceps femoris, hip adductors, levator scapulae, trapezius, rectus abdominis, transverse abdominis, external obliques, internal obliques, erector spinae, pectoralis major, biceps brachii, triceps brachii, brachialis, pronator teres, brachioradialis, rhomboids, deltoids, and latissimus dorsi. Other skeletal muscles are also possible whose mass and/or strength may be increased by treatment with an anti-GDF-8 antibody of the invention.

Increases in muscle mass or strength can be assessed directly, for example, by observing a subject's ability to resist or lift a weight, or indirectly, for example, by scanning the subject's body using MRI, CT, or dual-energy X-ray absorptiometry (DEXA). Other techniques are also possible.

Alternatively, a therapeutic benefit may be inferred from a reduction in the severity of symptoms that would otherwise gradually worsen. Benefit may also be demonstrated using physiological tests of muscle function (e.g., electromyography), histopathological testing of biopsied muscle structure, and biochemical tests (e.g., the presence of serum creatine kinase, an enzyme released by damaged muscle). Other tests of muscle structure and function that can be used to detect therapeutic benefit are also possible.

In other embodiments related to muscle mass and/or strength, the present invention provides methods for treating and preventing muscular dystrophy ("MD") by administering a composition comprising an anti-GDF-8 antibody to a patient in need of treatment or prevention of these diseases. In some embodiments of these methods, the antibody is OGD1.0.0 or an anti-GDF-8 binding fragment or derivative thereof. According to certain embodiments, the subject is a human pediatric patient with muscular dystrophy, and in other embodiments, the subject is a human adult patient with muscular dystrophy.

As is known in the art, there are different types of muscular dystrophy, which differ in the nature of the genetic damage that causes the disease and the phenotype resulting from the underlying genetic defect. Non-limiting examples of the types of muscular dystrophy that can be treated or prevented by administering a composition comprising an antibody of the invention include Duchenne muscular dystrophy (DMD) (also known as Duchenne MD), Becker muscular dystrophy (BMD), Emery-Dreifuss muscular dystrophy (EDMD), limb-girdle muscular dystrophy (LGMD), facioscapulohumeral muscular dystrophy (FSH or FSHD) (also known as Landouzy-Dejerine MD), myotonic dystrophy (MMD) (also known as DM or Steinert Disease), oculopharyngeal muscular dystrophy (OPMD), distal muscular dystrophy (DD) (also known as Miyoshi MD), and congenital muscular dystrophy (CMD).

In addition to the improved techniques for assessing muscle mass and/or strength described above, the therapeutic benefit of administering a composition comprising an antibody of the invention to a subject with MD can be quantified using the 6-minute walk test ("6MWT"). See, for example, McDonald et al., The 6-minute walk test as a new outcome measure in Duchenne muscular dystrophy, Muscle Nerve (2010) 41:500-510, which is incorporated herein by reference.

In the 6MWT, subjects are tested to determine how far they can walk along a preset route in 6 minutes. Subjects are typically tested before starting treatment to establish a baseline and then tested at intervals as treatment progresses. A therapeutic benefit is observed when the MD subject's performance on the 6MWT remains constant or actually improves with treatment, or when the subject's performance declines more slowly than the average untreated subject. Exemplary, non-limiting improvements in performance on the 6MWT include approximately 4%, 5%, 6%, 7%, 8%, 9%, 10%, 11%, 12%, 13%, 14%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 60%, 70%, 80%, 90%, 100% or greater percentage improvements compared to an untreated or placebo-treated control. The 6MWT can also be used to detect the therapeutic benefit of subjects being treated for conditions or disorders other than MD that affect ambulation.

In other embodiments, the present invention provides methods for treating and preventing motor neuron diseases by administering a composition comprising an anti-GDF-8 antibody to a patient in need of treatment or prevention of such diseases. In some embodiments of these methods, the antibody is OGD1.0.0 or an anti-GDF-8 binding fragment or derivative thereof. Non-limiting examples of motor neuron diseases that can be treated or prevented by administering a composition comprising an antibody of the invention include amyotrophic lateral sclerosis (ALS) (also known as Lou Gehrig's disease), spinal muscular atrophy type 1 (SMA1) (also known as Wernig-Hoffmann disease), spinal muscular atrophy type 2 (SMA2), spinal muscular atrophy type 3 (SMA3) (also known as Kugelberg-Welander disease), and spinal bulbar muscular atrophy (SBMA) (also known as Kennedy disease).

Previous work has demonstrated that murine anti-GDF-8 antibodies are effective in increasing muscle mass and strength in SOD1 mice and rats, a small animal model of human ALS. See WO2007/024535 and Holzbauer et al., Myostatininhibition slows muscle atrophy in rodent models of amyotrophic lateral sclerosis, Neurobiology of Disease (2006) 23:697-707, which are incorporated herein by reference. As reported, treatment with murine antibodies increased muscle mass in the diaphragm and skeletal muscle of SOD1 mice and rats compared to PBS-treated controls. Similarly, antibody treatment reduced muscle atrophy in the gastrocnemius and diaphragm of SOD1 mice compared to controls receiving PBS. The nutritional effects of antibody treatment were primarily evident during the early stages of the disease process, rather than the final stages, but inhibition of diaphragm atrophy was evident in both phases. Compared to vehicle-only treated control animals, antibody treatment was also observed to increase limb muscle strength and total body weight in SOD1 mice and rats, but the antibody did not prolong survival. Because the humanized anti-GDF-8 antibodies of the present invention (e.g., OGD1.0.0) share the same antigen-binding determinants as the murine antibodies discussed above, they are also expected to be effective in treating or preventing ALS in humans.

Other congenital or acquired diseases and disorders of the muscle, central nervous system, and peripheral nervous system that affect muscle mass, function, and/or strength can also be treated or prevented by administering to a subject in need thereof a composition comprising an antibody of the invention.

The present invention also provides methods of treating and preventing metabolic disorders by administering a composition comprising an anti-GDF-8 antibody to a patient in need of treatment for a metabolic disorder. In certain of these embodiments, the antibody is OGD 1.0.0 or a GDF-8 binding fragment or derivative thereof.

Non-limiting examples of metabolic disorders that can be treated or prevented by administering an antibody of the invention include type 2 diabetes, metabolic syndrome (eg, syndrome X), insulin resistance, and impaired glucose tolerance.

In other embodiments, the present invention provides methods for treating and preventing adipose tissue disorders by administering a composition comprising an anti-GDF-8 antibody to a patient in need of treatment for these disorders. In certain of these embodiments, the antibody is OGD1.0.0 or a GDF-8 binding fragment or derivative thereof.

Non-limiting examples of adipose tissue disorders that can be treated or prevented by administering the antibodies of the invention include obesity and a higher than normal body mass index (BMI) for the particular subject's sex, age, and height.

In other embodiments, the present invention provides methods for treating and preventing bone loss disorders by administering a composition comprising an anti-GDF-8 antibody to a patient in need of treatment for such disorders. In certain of these embodiments, the antibody is OGD1.0.0 or a GDF-8 binding fragment or derivative thereof.

Non-limiting examples of bone loss disorders that can be treated or prevented by administering the antibodies of the invention include osteoporosis, hormone-related osteoporosis, osteopenia, osteoarthritis, and osteoporosis-related fractures.

Combination therapy

According to certain embodiments of the methods of the present invention, the anti-GDF-8 antibody can be administered in a composition as a monotherapy or as a combination therapy with at least one second therapeutic agent. Typically, but not in all cases, the second therapeutic agent is selected to treat or prevent the same condition or disorder targeted by the anti-GDF-8 antibody. However, in other embodiments, the second agent can be selected to treat or prevent a different condition or disorder. The dosages of the antibody and second therapeutic agent used in the combination therapy are selected according to the knowledge of those skilled in the art to maximize efficacy and minimize side effects.

The anti-GDF-8 antibody compositions of the present invention can be administered to a subject using the same or different modes of administration as the second therapeutic agent. Depending on their chemical and physical characteristics, the anti-GDF-8 antibody and the second therapeutic agent can be combined in the same composition. In alternative embodiments, they are administered as separate compositions. The combination of the antibody and the second therapeutic agent can be conveniently included in a kit of the present invention.

If administered as a combination therapy, the antibody and the second therapeutic agent may be administered simultaneously, sequentially or separately.

When two or more agents are administered at the same time, even if individual administrations overlap, but start or end at different times, simultaneous administration occurs. When two or more agents are administered to a subject on the same day (e.g., during the same clinical visit) but not at the same time, sequential administration occurs.

Sequential administrations can be separated by 1, 2, 3, 4, 5, 6, 7, 8 hours or more. The anti-GDF-8 antibody composition can be administered first, followed by the second agent, or vice versa.

Separate administration occurs when the agents are administered to a subject on different days. Exemplary intervals between separate administrations of the agents can be 1 day, 2 days, 3 days, 4 days, 5 days, 6 days, 1 week, 2 weeks, 3 weeks, or 1 month or longer. As with sequential administration, the anti-GDF-8 antibody composition can be administered before or after the separate administration of the second agent.

In certain other embodiments of the invention, whether administered sequentially or separately, the anti-GDF-8 antibody composition and the second therapeutic agent may be administered repeatedly in an alternating pattern.

In methods for treating or preventing metabolic disorders, the anti-GDF-8 antibodies of the present invention can be combined with a second agent that is effective in treating or preventing these disorders. Non-limiting examples of second agents effective for this purpose include metformin, sulfonylureas, insulin, pramlintide, thiazolidinediones (e.g., rosiglitazone and pioglitazone), GLP-1 analogs (e.g., exenatide), and DPP-IV inhibitors (e.g., vildagliptin).

In methods for treating or preventing bone loss disorders, the anti-GDF-8 antibodies of the present invention can be combined with a second agent that is effective in treating or preventing these disorders. Non-limiting examples of second agents effective for this purpose include bisphosphonates (e.g., alendronate and risedronate), calcitonin, raloxifene, and hormonal agents (e.g., estrogen or parathyroid hormone (PTH)).

In methods for treating or preventing muscular dystrophy, the anti-GDF-8 antibodies of the present invention can be combined with a second agent (e.g., a corticosteroid) that is effective in treating or preventing muscular dystrophy. Other agents effective in treating or preventing muscular dystrophy are known in the art. Non-limiting examples of corticosteroids effective in treating muscular dystrophy include methylprednisolone, deflazacort, betamethasone, prednisolone, hydrocortisone, cortisone, beclomethasone, budesonide, cortisol, dexamethasone, fluticasone, prednisone, mometasone, triamcinolone, and their derivatives. In other embodiments, the anti-GDF-8 antibodies can be administered with an agent used to treat cardiomyopathy, which often occurs in DMD patients, particularly elderly DMD patients. These agents include, but are not necessarily limited to, beta-adrenergic blockers and angiotensin-converting enzyme inhibitors.

In methods for treating or preventing ALS, the anti-GDF-8 antibodies of the invention can be combined with a second agent effective for treating or preventing ALS, including, but not necessarily limited to, riluzole, talampanel, glycopyrrolate, benztropine, scopolamine, atropine, trihexyphenidyl hydrochloride, amitriptyline, fluvoxamine, baclofen, tizanidine, dantrolene, diazepam, quinine, phenytoin, benzodiazepine, gabapentin, antispasmodics, antidepressants, and morphine or other analgesics.

According to other embodiments, compositions comprising an anti-GDF-8 antibody of the invention can be administered in conjunction with non-drug based therapies, including, for example, but not limited to, exercise, physical therapy, respiratory therapy, ventilatory support, cardiac therapy, and nutritional supplementation.

Effective dose

As described above, compositions comprising the anti-GDF-8 antibodies of the invention can be administered to a subject in need of treatment or prevention of a condition or disorder at a dosage effective to at least partially achieve the desired therapeutic benefit.

It is not necessary to bind all GDF-8 to achieve therapeutic efficacy. Rather, reducing the concentration of mature active GDF-8 in body fluids (e.g., blood or serum) or in body tissues (e.g., muscle or other body tissues or organs) can also be effective.

According to the knowledge of one skilled in the art, the dosage of the anti-GDF-8 antibody composition can be gradually increased in a patient to reduce the concentration of active GDF-8 in a tissue or body fluid of interest by at least about 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100%, or by about 5% to 10%, about 10% to 45%, about 15% to 50%, or about 10% to 45%. 20%, about 20%-25%, about 25%-30%, about 30%-35%, about 35%-40%, about 40%-45%, about 45%-50%, about 50%-55%, about 55%-60%, about 60%-65%, about 65%-70%, about 70%-75%, about 75%-80%, about 80%-85%, about 85%-90%, about 90%-95%, about 95%-99%, or a percentage decrease in active GDF-8 concentration between any of the above values.

The amount of anti-GDF-8 antibody administered to a subject will depend on a variety of factors, including the condition or disorder to be treated or prevented, the size and weight of the subject, the form, route, and site of administration, the treatment regimen (e.g., whether a second therapeutic agent is used), the age and condition of the particular subject, the amount of active GDF-8 detected in the subject's tissue or fluid of interest prior to the start of treatment, and the subject's responsiveness or sensitivity to the effects of the antibody composition. An appropriate dosage can be readily determined by one skilled in the art. Ultimately, a clinician or similar care provider will determine the appropriate dosage to use. This dosage can be repeated at an appropriate frequency. If side effects occur, the amount and/or frequency of the dosage can be altered or reduced according to normal clinical practice. Appropriate dosage and treatment regimen can be established by monitoring the progress of treatment using methods known to those skilled in the art.

An effective dose can be initially estimated based on in vitro assays. For example, an initial dose for use in an animal can be formulated to achieve circulating blood or serum concentrations of an anti-GDF-8 antibody that are equal to or greater than the binding affinity of the antibody for GDF-8 as measured in vitro. One skilled in the art is well-versed in taking into account the bioavailability of a particular antibody when calculating doses to achieve these circulating blood or serum concentrations. As a guide, the reader is referred to Part 1: General Principles in "Goodman and Gilman's The Pharmacological Basis of Therapeutics", 11th edition, Hardman, J.G. et al., eds., McGraw-Hill Professional, and references cited therein. Initial doses can also be estimated based on in vivo data (e.g., animal models). One skilled in the art can routinely adapt this information to determine a dose suitable for human administration.

In certain embodiments, the dosage for an individual subject can be determined by measuring the concentration of active GDF-8 in serum, muscle, or other body fluid or tissue of interest multiple times over several days to weeks prior to administration of the antibody composition to calculate a saturable amount of anti-GDF-8 antibody (i.e., an amount sufficient to bind substantially all active GDF-8). Those skilled in the art will appreciate that the amount of any specific antibody required to achieve saturation for a given amount of GDF-8 in serum, muscle, or other location will depend in part on the affinity of the specific antibody for GDF-8. Methods for calculating the saturating amount of a specific anti-GDF-8 antibody (taking into account the pharmacokinetic properties and bioavailability of the specific antibody, as needed) are well known in the art. To ensure saturation, an amount greater than the calculated saturating amount may be administered, for example, an amount at least 2, 3, 4, 5, 6, 7, 8, 9, or even 10 times the calculated saturating amount may be administered.

Depending on the factors discussed above (e.g., the condition or disorder to be treated or prevented, etc.), an effective dosage of an anti-GDF-8 antibody composition can range from about 0.01 mg/kg to about 250 mg/kg per single (e.g., bolus) administration, multiple administrations, or continuous (e.g., infusion) administration, or any effective range or value therein.

In certain embodiments, each dose may be within the following ranges: about 0.1 mg/kg to about 0.5 mg/kg; about 0.25 mg/kg to about 0.75 mg/kg; about 0.5 mg/kg to about 1 mg/kg; about 2 mg/kg; about 1.5 mg/kg to about 2.5 mg/kg; about 2 mg/kg to about 3 mg/kg; about 2.5 mg/kg to about 3.5 mg/kg; about 3 mg/kg to about 4 mg/kg; about 3.5 mg/kg to about 4.5 mg/kg; about 4 mg/kg g to about 5 mg/kg; about 5 mg/kg to about 7 mg/kg; about 6 mg/kg to about 8 mg/kg; about 7 mg/kg to about 9 mg/kg; about 8 mg/kg to about 10 mg/kg; about 10 mg/kg to about 15 mg/kg; about 12.5 mg/kg to about 17.5 mg/kg; about 15 mg/kg to about 20 mg/kg; about 17.5 mg/kg to about 22.5 mg/kg; about 20 mg/kg to about 25 mg/kg; about 22.5 mg/kg to about 27.5 mg/kg; about 25 mg/kg to about 30 mg/kg; about 30 mg/kg to about 40 mg/kg; about 35 mg/kg to about 45 mg/kg; about 40 mg/kg to about 50 mg/kg; about 45 mg/kg to about 55 mg/kg; about 50 mg/kg to about 60 mg/kg; about 55 mg/kg to about 65 mg/kg; about 60 mg/kg to about 70 mg/kg; about 65 mg/kg to about 75 mg/kg; about 70 mg/kg to about 80 mg/kg; about 75 [0014] The present invention provides an exemplary embodiment of the present invention wherein the dosage range is from about 1 mg/kg to about 85 mg/kg; about 80 mg/kg to about 90 mg/kg; about 85 mg/kg to about 95 mg/kg; about 90 mg/kg to about 100 mg/kg; about 95 mg/kg to about 105 mg/kg; about 100 mg/kg to about 150 mg/kg; about 125 mg/kg to about 175 mg/kg; about 150 mg/kg to about 200 mg/kg; about 175 mg/kg to about 225 mg/kg; about 200 mg/kg to about 250 mg/kg. Other dosage ranges are also possible.

In some embodiments, the amount of the treatment regimen used, the frequency and the duration of the treatment regimen used will depend on multiple factors, such as the age, body weight and the disease state of the object. Therefore, in non-limiting examples, the treatment regimen used can be continuous 1 day or longer, 2 days or longer, 3 days or longer, 4 days or longer, 5 days or longer, 6 days or longer, 1 week or longer, 2 weeks to an uncertain time, continue 2 weeks to 6 months, 3 months to 5 years, 6 months to 1 or 2 years, 8 months to 18 months or the like. Alternatively, the treatment regimen provides repeated administration, such as twice a day, once a day, every 2 days, 3 days, 4 days, 5 days, 6 days once, once a week, every two weeks or once a month. Can be repeated administration with the same dose or with different doses. Use can be repeated 1 time, 2 times, 3 times, 4 times, 5 times, 6 times, 7 times, 8 times, 9 times, 10 times or more times. A therapeutically effective amount of an anti-GDF-8 antibody composition can be administered as a single dose or over the course of a treatment regimen (e.g., over the course of 1 week, 2 weeks, 3 weeks, 1 month, 3 months, 6 months, 1 year, or longer). What constitutes an effective dose of an anti-GDF-8 antibody composition in a particular subject may vary over time, as the subject's condition changes, or as other health issues arise.

Example

Example 1

Transient expression analysis of OGD1.0.0 and OGD1.1.1

Transient expression of intact heterotetrameric OGD1.1.1 and OGD1.0.0 was tested in COS-1M6 cells and showed that OGD1.0.0 was expressed at significantly higher levels.

Briefly, DNA encoding VH0 and VH1 (SEQ ID NOs: 49 and 53, respectively) was cloned into a mammalian IgG expression vector such that each VH region was linked in frame to a nucleic acid sequence encoding a heavy chain constant region of human IgG1 comprising three mutations that abolished effector function (SEQ ID NO: 57) to express a full-length antibody heavy chain comprising VH0 or VH1. Similarly, DNA encoding VL0 and VL1 (SEQ ID NOs: 51 and 55, respectively) was cloned into a mammalian IgG expression vector such that each VL region was linked in frame to a nucleic acid sequence encoding a human kappa light chain constant region of SEQ ID NO: 17 to express a full-length antibody light chain comprising VL0 or VL1.

After the expression vector is generated, DNA is prepared in large quantities using standard techniques. The cells are plated onto a 100 mm tissue culture dish and subsequently transiently co-transfected with heavy and light chain expression vectors (i.e., combining VH0 and VL0 in one plate and combining VH1 and VL1 in another plate). TransIT (Mirus MIR2306) transfection reagent (40 μl) is added to 2 ml OptiMEM growth medium with glutamine (2 mM final concentration) at room temperature, mixed by vortexing, and then incubated at room temperature for 15 minutes. The prepared DNA (8 μg each of heavy and light chain DNA) is added to the mixture and incubated at room temperature for 15 minutes. The transfection solution is then added to a tissue culture dish containing 8 ml of growth medium (DMEM, HIFBS, penicillin, streptomycin, glutamine). After incubation for 24 hours at 37°C, 10% CO ₂ , the cells were washed with R1CD1 serum-free growth medium and then grown for 48 hours at 37°C, 10% CO ₂ in 10 ml of R1CD1 (supplemented with penicillin, streptomycin, glutamine). The conditioned medium was removed from the cells, centrifuged to pellet any debris, and the supernatant was transferred to a fresh tube.

The concentration of the antibody produced by transiently transfected COS-1 cells is quantified using total human IgG-Fc specific ELISA. In brief, by adding 100 μl of the antibody (1 μg/ml) in PBS to each well and incubating overnight at room temperature, flat-bottomed ELISA plates are coated with goat anti-human IgG (Pierce 31125). Plates are blocked for 3 to 24 hours and then washed at room temperature with 0.02% casein PBS solution in 100 μl/well. Standards and samples are serially diluted in assay buffer (0.5% BSA, 0.02% Tween-20 in PBS), assigned to ELISA plates (100 μl/well) and incubated for 3 to 24 hours at room temperature. After washing, (100 μl/well) are assigned to goat anti-human IgG (Pierce 31413) diluted in assay buffer at 1:5000, and then the plate is incubated at room temperature for 15 minutes. After washing, the plates were developed by adding BioFX TMB (TMBW-0100-01) (100 μl/well). After stopping the reaction with 0.18 NH ₂ SO ₄ (100 μl/well), the plates were read at 450 nm using a Molecular Devices vMax plate reader. The sample concentrations were calculated using the linear range of the curve determined by the dilution series of the standards.

The results of transient transfection experiments are shown in the table below, where POI represents the peak of interest after purification by size exclusion chromatography (SEC) of Protein A. POI represents the proportion of intact, full-length antibody expressed by the cells, as opposed to high molecular weight aggregates or degradation products.

Surprisingly, under the same conditions after transient transfection, the OGD1.0.0 antibody was expressed at a significantly higher level (i.e., more than 10-fold) than the OGD1.1.1 antibody. Importantly, as indicated by the POI value, the significantly increased expression levels observed were almost entirely associated with the intact full-length antibody rather than high molecular weight complexes or degradation products. This expression difference is even more striking in view of the fact that between OGD1.0.0 and OGD1.1.1, there is only one amino acid difference at Kabat position 108 in the VH region (i.e., residue numbering 111 of SEQ ID NOs: 44 and 7) and one amino acid difference at Kabat position 100 in the VL region (i.e., residue numbering 100 of SEQ ID NOs: 46 and 9). See Figures 1A and 1B.

As explained above, these structural and functional differences are attributed to the use of different J segments in VH0 and VL0 compared to VH1 and VL1, respectively. As explained below, the most important differences appear to be changes in the VH region. Notably, this is believed to be the first demonstration that the choice of J segment used to construct a humanized antibody can comprehensively impact antibody expression levels, not to mention the significant improvement observed here. This discovery is particularly important because it is expected to significantly reduce the cost of goods required to produce OGD1.0.0. Without this discovery, this antibody would not be economically produced in the quantities required to market it, which would be detrimental to the patient population that could benefit from treatment with this antibody.

Table 3: Comparison of OGD1.0.0 and OGD1.1.1 transiently expressed in COS cells.

Antibody Transient expression in COS-1 cells POI OGD1.0.0 28.45 μg/ml >99% OGD1.1.1 2.35 μg/ml >99%

Example 2

Stable expression analysis of OGD1.0.0 and OGD1.1.1

Stable expression of OGD1.0.0 and OGD1.1.1 was tested in CHO-DUKX cells. Briefly, cells were grown to 80% confluence and then co-transfected with 25 μg each of the heavy and light chain expression vectors described in the previous example (50 μg total) using lipofectamine transfection reagent (i.e., VH0 and VL0 for one group of cells, and VH1 and VL1 for another group of cells). After transfection, spent medium was replaced with fresh R1CD1 medium supplemented with 10% FBS every 3 to 4 days while establishing a stable pool.

After establishing stable transfectants, the cells were tested for their ability to express anti-GDF-8 antibodies when grown as adherent cells in serum-free R5CD1 medium. Under these conditions, cells expressing OGD1.0.0 expressed 47.3 mg/L of antibody after 96 hours of growth, while cells expressing OGD1.1.1 expressed 41 mg/L of antibody after 72 hours of growth. The antibodies were purified using a 1 mL protein A column, and the concentration was quantified using HPLC.

After the attached cells adapted to suspension growth in serum-free medium, expression of the OGD1.0.0 and OGD1.1.1 antibodies was re-assayed. 3.0×10 ⁵ viable cells/mL were seeded in AS1 serum-free medium and incubated at 37°C. On day 4, the pH was adjusted to 7.3, a concentrated feed was added, and the incubation temperature was lowered to 31°C for an additional 3 days of growth. OGD1.0.0-expressing cells were grown in a 100-L culture volume, while OGD1.1.1-expressing cells were grown in a 50-mL culture volume. All other growth conditions were identical between the cells. On day 7, as determined by protein A purification and HPLC quantification, cells expressing OGD1.0.0 expressed 66.12 mg/L of antibody, while cells expressing OGD1.1.1 expressed 10.6 mg/L of antibody. When the experiment using OGD1.0.0 cells was repeated by growing the cells in 100 L culture for 9 days (including 5 days at 31° C.), the antibody concentration increased to 207.2 mg/L. In another experiment in which OGD1.0.0-expressing cells were grown in 25 L culture for 11 days (including 7 days at 31° C.), the antibody concentration was 145 mg/L. In a separate experiment in which OGD1.1.1-expressing cells were grown in 50 mL culture in serum-free R5CD1 medium for 7 days (including 3 days at 31° C.), the antibody concentration was 39.3 mg/L.

The results of the stable transfection experiments are shown in the table below, where POI represents the peak of interest after protein A purification by size exclusion chromatography (SEC). Consistent with the results obtained when expressing OGD1.0.0 and OGD1.1.1 in transiently transfected COS-1 cells, under similar conditions, the expression level of the OGD1.0.0 antibody in stably transfected CHO cells was significantly higher than that of OGD1.1.1. This surprising result is consistent with the increased expression levels observed in the transient transfection experiments above. The results also indicate that neither the mode of culturing the expressing cells (attached or suspended) nor the cell type significantly affects the increased expression of the OGD1.0.0 antibody relative to OGD1.1.1.

Table 4: Comparison of expression levels of OGD1.0.0 and OGD1.1.1 stably expressed in CHO cells.

Antibody CHO stabilization pool OGD1.0.0 66.1mg/L OGD1.1.1 10.6mg/L

Example 3

Transient expression analysis of OGD1.0.1 and OGD1.1.0

Because changes were made to the J segments of each of the heavy and light chain variable regions, it is unclear whether a single change alone is sufficient to cause the significantly higher expression levels observed for OGD1.0.0, or whether both changes may contribute to increased antibody expression.

To investigate this issue, the applicant repeated the above transient transfection experiments, but combined VH0 and VL1 constructs in one plate and VH1 and VL0 constructs in another plate, and then quantified antibody expression levels using ELISA. The results of the experiment are shown in the table below, which shows that replacing the JH3J segment with JH4 in the VH region is sufficient to produce the huge increase in antibody expression observed by the applicant. In contrast, changing the kappa J segment (i.e., replacing JK1 with JK4) does not seem to significantly affect expression levels.

Table 5: Effects of combining VH1 with VL0 and VH0 with VL1 on antibody expression

Antibody Express OGD1.0.0 28.5 μg/ml OGD1.0.1 27.6 μg/ml OGD1.1.0 1.9 μg/ml

OGD1.1.1 2.4 μg/ml

Example 4

GDF-8 binding of anti-GDF-8 antibodies

GDF-8 binding was analyzed using quantitative ELISA and surface plasmon resonance (SPR) for the parental murine antibody, a chimeric mouse-human antibody (murine variable domains and human constant domains), and the humanized antibodies OGD1.0.0 and OGD1.1.1. In the ELISA assay, the ability of the antibodies to inhibit GDF-8 binding to its cognate high-affinity receptor, ActRIIB, was determined by calculating _IC50 values. Apparent _KD values were calculated using SPR analysis. The results are shown in Table 6.

For ELISA, ActRIIB-Fc fusion protein (1 μg/ml in 0.2 M sodium carbonate buffer) was coated onto 96-well flat-bottom assay plates at 4°C overnight. The coated plates were then blocked with 1 mg/ml BSA (200 μl/well) in PBS 0.1% Tween for 1 hour at room temperature or overnight at 4°C, followed by washing. Different concentrations of antibodies were combined with 10 ng/ml of GDF-8 conjugated to biotin and incubated at room temperature for 45 minutes. After incubation, the test solution was added to the blocked ELISA plate (100 μl/well) and incubated at room temperature for another hour. After washing each well, the amount of GDF-8 bound to the immobilized ActRIIB-Fc was detected relative to the control using streptavidin-horseradish peroxidase (incubation for 30 minutes) and TMB. Colorimetric measurements at 450 nm were recorded in a microplate reader. Experiments using murine and chimeric antibodies were repeated four times and the average value was taken.

SPR was performed using a BIACORE 3000 (GE Healthcare) machine at 25°C. Anti-mouse IgG antibodies were used to capture murine antibodies, while protein A was used to capture humanized antibodies. Protein A was immobilized on all four flow cells of a CM5 sensor chip using amino coupling chemistry. The surface was activated by injecting a solution of 0.2M N-ethyl-N-dimethyl-amino-propyl-carbodiimide (EDC) and 50mM N-hydroxysuccinimide (NHS) for 7 minutes. Protein A was diluted to 50 μg/ml in 10mM sodium acetate buffer (pH 5.0) and injected at a flow rate of 10 μl/min for 3 minutes. The surface was then blocked with 1M ethanolamine (ETH) for 7 minutes. The final immobilization level of protein A was between 1000-1200 response units (RU). After the immobilization procedure, the surface was washed several times with running buffer (0.01M HEPES pH 7.4, 0.15M NaCl, 3mM EDTA, 0.005% P20) to equilibrate the surface. The antibodies were diluted to 0.25 μg/ml in HBS-EP buffer. A solution (5 μl) of each antibody was injected through flow cells 2, 3, or 4 coated with Protein A at a rate of 10 μl/min, generating approximately 200 RU of captured antibody. A GDF-8 titration series (2-fold dilutions from 4.0 nM to 0.125 nM) was prepared in 0.01M sodium acetate (pH 5.0, 0.15M NaCl, 3mM EDTA, 0.005% P20). The sodium acetate solution also served as the running buffer. The GDF-8 solution was injected over the captured antibody at a flow rate of 50 μl/min for 2 minutes and allowed to dissociate for 30 minutes. After each injection and capture cycle, the sensor chip surface was regenerated with 30 μl of 10 mM NaPO ₄ , 0.5 M NaCl (pH 2.5) at a flow rate of 50 μl/min. Data were analyzed using BIA Evaluation Software (version 4.1.1, GE Healthcare). Data were double-referenced by subtracting the signal contributed by the buffer and reference surface. Sensorgram data were globally fitted using a Langmuir 1:1 model and _K values were calculated. Experiments using OGD 1.0.0 were repeated three times and averaged.

_IC50 values determined using ELISA were comparable between the antibody formats tested.In the more precise SPR assay, OGD1.0.0 had a significantly greater binding affinity for GDF-8 than OGD1.1.1.

Table 6: GDF-8 Binding of Anti-GDF-8 Antibodies

Antibody Mouse antibodies 0.165nM 21.83pM chimeric antibodies 0.165nM 2.99pM OGD1.0.0 0.140nM 2.59pM OGD1.1.1 0.140nM 7.25pM

Example 5

GDF-8 neutralization ability of OGD1.0.0 antibody

A reporter gene assay was used to confirm the ability of anti-GDF-8 antibodies to neutralize GDF-8-mediated signaling. A reporter gene construct, designated pGL3(CAGA) ₁₂ , was constructed by placing 12 CAGA boxes upstream of the TATA box and transcription start site from the adenovirus major late promoter in the luciferase reporter gene vector pGL3 (Promega). The CAGA box found in the PAI-1 gene promoter is a TGFβ response element that is also responsive to GDF-8. The human rhabdomyosarcoma cell line A204 (ATCC HTB-82) was transiently transfected with pGL3(CAGA) ₁₂ and cultured for 16 hours in 96-well plates in McCoy's 5A medium supplemented with 2 mM glutamine, 100 U/ml streptomycin, 100 μg/ml penicillin, and 10% fetal bovine serum. The antibody was preincubated with GDF-8 (10 ng/ml) in medium supplemented with 1 mg/ml BSA for 1 hour at room temperature. Cells were then treated with the test samples and controls excluding GDF-8 and containing GDF-8 (10 ng/ml) but without antibody for 6 hours at 37°C. Luciferase activity was measured using a luciferase assay system (Promega). Experiments using murine and chimeric antibodies were each repeated twice and averaged, while experiments using OGD1.0.0 were repeated three times and averaged. _EC50 values determined using a reporter gene assay were comparable among the antibody formats tested.

Table 7: GDF-8 neutralizing activity of anti-GDF-8 antibodies

Antibody CAGA EC50nM Mouse antibodies 33.50nM chimeric antibodies 24.25nM OGD1.0.0 27.30nM OGD1.1.1 26.00nM

Example 6

OGD1.0.0 antibody increases muscle mass, strength and lean mass in mice

Eight-week-old male C57B1/6 mice were administered OGD1.0.0 (10 mg/kg) or vehicle control (PBS) intraperitoneally (IP) once weekly for two weeks. A total of eight mice were used per group. On day 14, whole-body lean mass was determined by small animal NMR imaging. After determining lean mass, the animals were euthanized, and the gastrocnemius, quadriceps, and extensor digitorum longus (EDL) muscles were dissected and weighed. The EDL muscle was also tested for its ability to generate force in vitro.

After two weeks of treatment, control animals increased lean mass by 1.66 ± 0.56 g, while OGD 1.0.0-treated animals increased lean mass by 3.36 ± 0.62 g, representing a 102% increase relative to controls.

As shown in Figures 2A and 2B, in animals treated with the OGD1.0.0 antibody, quadriceps muscle mass increased by 14.8%, gastrocnemius muscle mass increased by 10.3%, and EDL muscle mass increased by 10.8% relative to controls. As shown in Figure 3, the total tetanic force exerted by the EDL muscle in animals treated with the OGD1.0.0 antibody increased by 14.8% relative to the force generated by the EDL muscle in vehicle-treated mice. Data are presented as mean ± SEM.

The dose responsiveness of whole body lean mass and quadriceps and gastrocnemius muscle mass to OGD1.0.0 treatment was also determined. In these experiments, 12-week-old female C57Bl/6 mice were divided into groups (n=6) and treated with 0.3, 1, 3, 10 or 30 mg/kg vehicle or OGD1.0.0 weekly for 4 weeks. At 7, 14, 21 and 28 days of the treatment phase, lean mass was determined using NMR imaging. At the end of the study phase, after the test animals were euthanized, the quadriceps and gastrocnemius muscles were dissected and weighed. As shown in Figures 4A and 4B, the muscle mass of the quadriceps and gastrocnemius increased with increasing antibody doses until the antibody dose was about 10 mg/kg. Similarly, as shown in Figures 5A and 5B, whole body lean mass increased with increasing antibody doses until the antibody dose was about 10 mg/kg. Data are shown as mean ± SEM.

Example 7

OGD1.0.0 antibody increases muscle mass and lean mass in mdx mice

The mdx mutation of the X-linked dystrophin gene (Dmd) spontaneously arises in C57BL/10ScSn mice and results in a point mutation within exon 3185 of the gene, converting a glutamine codon to a stop codon and leading to premature termination of dystrophin production. Consequently, mdx mice lack functional dystrophin and are used as a small animal model for human Duchenne muscular dystrophy. Muscle necrosis begins at approximately three weeks of age, accompanied by some pronounced muscle weakness. Although limb skeletal muscle is characterized by sustained and progressive degeneration and necrosis, this is compensated by a regenerative response driven by satellite cell activation and muscle hypertrophy. Muscle elasticity is generally reduced in mdx mutants, making them more susceptible to injury from prolonged activation. Leg muscles in mutant mice initially develop normally, but differentiation of regenerating myotubes into fast and slow fiber types is significantly inhibited. The relatively mild phenotype of mdx mice can be attributed in part to the compensatory function of the dystrophin-related protein utrophin, which is significantly upregulated in regenerating muscle fibers of adult mdx mutants. In contrast to limb muscles, diaphragm muscles of mdx mice do not undergo a significant regeneration period, and thus, continued malnutrition weakens these muscles with age. Specific twitch tension, specific tetanic tension, and maximal power are all reduced in mdx mutant diaphragms.

8-week-old male mdx and control C57Bl/6 mice were given OGD1.0.0 (10 mg/kg) or vehicle control (PBS) intraperitoneally (IP) once a week for 8 weeks. In these experiments, 10 mdx mice were treated with the antibody, 8 mice were given vehicle control, and 6 C57Bl/6 mice were treated with the antibody or PBS. At the end of the treatment period, whole body lean mass, grip strength, and muscle mass were measured. Whole body lean mass was determined by small animal NMR imaging. Grip strength was tested by placing the test animal on a wire grid so that it gripped the mesh with all limbs, then pulling the tail and measuring the maximum peak force when the animal released its grip. The data for each animal were averaged from 3-5 trials. After measuring body lean mass and grip strength, the mice were euthanized, and the quadriceps and gastrocnemius muscles were dissected and weighed.

As shown in Figure 6A, OGD1.0.0 antibody treatment increased lean mass in mdx mice by an average of 7.28 ± 0.4 g, compared to an average of 4.83 ± 0.4 g in PBS-treated mdx mice. The difference was statistically significant (p < 0.05). Thus, over the 8-week study, lean mass in mdx mice increased by 50% ± 8.2% relative to vehicle-treated controls. As shown in Figure 6B, antibody treatment also increased grip strength in mdx mice compared to vehicle-treated controls. The difference was statistically significant (p < 0.05).

As shown in Figure 7A, antibody treatment increased the mass of the gastrocnemius and quadriceps muscles of mdx mice and C57Bl/6 mice compared to the same type of mice treated with PBS. These increases were statistically significant (p = 0.005 and p = 0.002 for the mdx quadriceps and gastrocnemius, respectively; p = 0.001 and p = 0.003 for the C57Bl/6 quadriceps and gastrocnemius, respectively). As shown in Figure 7B, the mass of the gastrocnemius and quadriceps muscles from antibody-treated mdx mice increased by 12.2% and 12.1%, respectively, compared to vehicle-treated control mice. After antibody treatment, the mass of the same type of muscle from C57Bl/6 mice also increased by 15.2% and 12.8%, compared to vehicle-treated controls (not shown).

Example 8

OGD1.0.0 antibody increases muscle volume and lean mass in non-human primates

Two studies were designed and conducted to investigate the effects of administration of the OGD1.0.0 antibody on lean body mass and muscle volume in cynomolgus monkeys.

Each study lasts for 8 weeks, wherein the animals are given antibodies weekly by intravenous administration and excess food is provided to ensure positive nitrogen balance. In the first study, PBS vehicle or OGD1.0.0 antibody was administered to each of three male and three female subjects at a dosage of 3.0 mg/kg, 10 mg/kg, and 30 mg/kg. Before the first treatment and then at the 4th and 8th weeks, the animals were anesthetized and subsequently imaged using dual-energy X-ray absorptiometry (DEXA), computerized x-ray tomography (CT), and magnetic resonance imaging (MRI) to detect and measure body composition, including lean mass and fat content. The subject animals of Study 1 were then euthanized and autopsied. In the second study, only male subjects were used and they only received vehicle (n=5) or OGD1.0.0 antibody at a dosage of 10 mg/kg and 30 mg/kg (respectively, n=5 subjects and n=3 subjects). At 8 weeks, the subject animals were imaged as in the first study. Thereafter, animals were maintained on the supplemented diet and imaged again at 12, 17, and 26 weeks.

8 weeks of data on body lean mass measured by DEXA from both studies were combined and analyzed. The results are shown in Figure 8, which show that after 8 weeks of treatment with OGD1.0.0 antibody, there was a dose-responsive increase in body lean mass and leg lean mass. The data are expressed as mean ± SEM. The number of subjects included in the study was n = 11 (for vehicle only), n = 6 (for 3 mg / kg antibody), n = 10 (for 10 mg / kg), and n = 8 (for 30 mg / kg). At all antibody doses tested, the increase in body lean mass and leg lean mass relative to vehicle-treated controls was statistically significant (p < 0.05). In addition, the increase in leg lean mass of subjects treated with 30 mg / kg relative to 10 mg / kg was also found to be statistically significant (p < 0.05).

Interestingly, as shown in FIG9 , in the second study, increases in lean body mass persisted for several weeks after the last antibody dose at week 7 between subjects treated with 10 mg/kg and 30 mg/kg OGD1.0.0 antibody. The data are shown as the difference compared to the PBS vehicle at each time point. The increases relative to the control at the higher dose were statistically significant (p<0.05) at all weeks shown. The increases at the lower dose were statistically significant at weeks 4 and 8 (p<0.09).

The effect of OGD1.0.0 antibody treatment on the dorsal spinal muscle volume at lumbar vertebrae L3-L5 was measured by CT scan. As shown in Figures 10A and 10B, compared to the PBS control, the 10 mg/kg (n=5) and 30 mg/kg (n=3) OGD1.0.0 antibody treatments for 8 weeks significantly increased the supraaxial muscle volume. The increase in muscle volume was statistically significant (p<0.05).

Figure 11 is a 3D representation of epiaxial muscle from an exemplary test subject after 4 weeks of treatment with 30 mg/kg OGD1.0.0 antibody. The resulting muscle volume increased by 22% compared to baseline (left panel), which was visually evident (right panel).

Example 9

OGD1.0.0 antibody lacks Fc domain effector function

Surface plasmon resonance was used to test the binding of the OGD1.0.0 antibody, which includes three mutations in the Fc region known to abrogate Fcγ receptor (FcγR) binding to a panel of Fcγ receptors. All experiments were performed using a Biacore T200 instrument (GE Healthcare). Briefly, 100 RU of GDF-8 was captured on a Sensor Chip-SA via the biotin tag to which approximately 100 RU of the OGD1.0.0 antibody was captured. FcγRs were then passed over concentrations ranging from 0 to 21 μM (for CD32a-131H, CD16a-158V, and CD32b) and 0 to 270 nM (for CD64). For each FcγR binding experiment, injections were performed continuously using a single-cycle kinetic mode. The association and dissociation phases each lasted 120 s. At the end of the dissociation phase, following the last injection, the GDF-8-containing surface was regenerated using a 20 s pulse of 0.1% TFA solution.

No binding was observed to CD16, CD32a, and CD64. In contrast, binding to CD32b was observed, but only at the highest concentration tested (21 μM). Due to the lack of data points above 21 μM, the exact Kd was not determined, but it can be assumed to be above 21 μM, which is considered extremely weak compared to the Kd of wild-type IgG1 molecules (i.e., 2-4 μM). These results suggest that the OGD1.0.0 antibody will have no or significantly reduced ability to induce effector function.

Example 10

Crystal structure of anti-GDF-8 antibody bound to GDF-8

As explained in this Example, the crystal structures of chimeric mouse and humanized anti-GDF-8 antibodies bound to human GDF-8 were solved and used to determine the identities of the amino acids within the antibody and GDF-8 that contact each other.

Fab fragments were prepared from a chimeric anti-GDF-8 antibody containing murine VH and VL regions linked to human IgG1 constant regions (SEQ ID NO: 3 and SEQ ID NO: 5, respectively). The Fab fragments were then mixed with human GDF-8 protein to form a binding complex. The protein complex was concentrated to 10.75 mg/mL in 50 mM tris hydrochloride (pH 7.5) and 100 mM sodium chloride. Crystals were formed using the hanging drop method equilibrated at 18°C in a solution containing 20% PEG MME 5000 and 100 mM bis-tris (pH 6.5). Crystals containing Fab prepared from humanized anti-GDF-8 antibody OGD1.0.0 and GDF-8 were prepared in a similar manner, except that the protein solution was equilibrated in an unbuffered solution containing 20% PEG 3350 and 200 mM sodium chloride.

Single-wavelength data for each crystal were collected online at the ID beamline of the SER-CAT Advanced Photon Source, Argonne National Laboratory. Each data set used a single crystal cooled to -180°C. Data were processed using DENZO and Scalepack (Z. Otwinowski and W. Minor, "Processing of X-ray Diffraction Data Collected in Oscillation Mode," Methods in Enzymology, Vol. 276: Macromolecular Crystallography, Part A, pp. 307-326, 1997, eds. C. W. Carter, Jr. and R. M. Sweet, Academic Press (New York)) (incorporated by reference). The structure of the chimeric antibody in complex with GDF-8 was solved by molecular replacement using the program AMORE (Navaza, J. (2001). Implementation of molecular replacement in A Mo Re. Acta Crystallogr., Sect. D: Biol. Crystallogr. 57, 1367-1372) (incorporated by reference). The probe used in the molecular replacement search was PDB entry 1HZH. Prior to refinement, 5% of the data was randomly selected and designated the R- _free test set to monitor refinement progress. The structure of each complex was then reconstructed in Coot using a series of approximations (Emsley, P. and Cowtan, K. (2004) Coot: model-building tools for molecular graphics. Acta Crystallogr., Sect. D: Biol. Crystallogr. 60, 2126–2132) (incorporated by reference). Refinement statistics are listed in Table 8. The structure of humanized OGD1.0.0 in complex with GDF-8 was solved in a similar manner, except that the probe used was the structure of the chimeric antibody.

Table 8: Refinement statistics of antibody:GDF-8 co-crystal structure

The residues in the antibody and GDF-8 that contact each other, as inferred based on the co-crystal structure, are listed in Table 9, with antibody residues from the VH chain preceded by "H" and numbered with reference to SEQ ID NO: 3. Residues from the VL chain are numbered after "L" and with reference to SEQ ID NO: 5. Residues from mature human GDF-8 are numbered with reference to SEQ ID NO: 1. Residues are defined as contacting if they contain at least one pair of contacting atoms. Atoms are defined as contacting if their contact ratio, C, is <1.3, where C = _D12 ÷ ( _R1 + _R2 ), where _D12 is the distance between atoms, _R1 is the vdW radius of atom 1, and _R2 is the vdW radius of atom 2. In practice, the average distance between contacting atoms is approximately , but in any particular case, the actual distance varies depending on the type of atom in question.

Table 9: Residue contacts between antibodies and GDF-8 observed in the co-crystal structure

Example 11

Further humanization of antibody VH and VL regions

Based on sequence analysis and structure of the anti-GDF-8 antibody co-crystallized with GDF-8, the antibody VH and VL regions were modified in an attempt to further humanize their sequences. A sequence alignment of the further humanized VH regions is shown in Figure 1A. A sequence alignment of the further humanized VL regions is shown in Figure 1B. After generating expression constructs containing the novel VH and VL regions, the antibodies were produced in transiently transfected COS-1 cells and purified using standard techniques. The antibodies were then tested for binding affinity and neutralizing activity against GDF-8 as described herein. The results are reported in Table 10.

The CDR2 amino acid sequences from the humanized VH0 and VL0 domains (derived from murine antibodies) were compared to the CDR2 sequences of the human germline VH domain DP-47 and VL domain DPK-9, respectively. All residues in the VH0 and VL0 CDR2 sequences that differed from the human sequences were converted to human residues. The new VH and VL domains were designated VH2 (SEQ ID NO: 66) and VL2 (SEQ ID NO: 67), respectively. Intact antibodies generated using VH2 with VL0 domains and VH0 with VL2 domains were tested for binding to GDF-8 in competitive ELISA assays. The results showed that the fully humanized VH CDR2 significantly reduced GDF-8 binding, while the fully humanized VL CDR2 did not reduce antigen binding.

Further humanization of the VH and VL regions was based on the co-crystal structure. Here, mouse-derived residues in the CDRs of the VH and VL regions were retained only if they were observed to contact GDF-8 residues in the co-crystal structure. Otherwise, all VH and VL CDR residues were changed to the corresponding human residues in DP-47 and DPK9, respectively, to generate VH3 (SEQ ID NO: 68) and VL3 (SEQ ID NO: 69). In competitive ELISA experiments, the antibody containing VH3 and VL0 (i.e., OGD1.3.0) showed a significant loss of activity, while the antibody containing VH0 and VL3 (i.e., OGD1.0.3) appeared to retain full activity.

Based on the sequence of CDR2 in VL2, position 50 of VL3 was substituted with alanine (i.e., S50A) to generate VL4 (SEQ ID NO: 71). An antibody comprising VH0 and VL4 prepared using this VL region (i.e., OGD1.0.4) retained substantial activity. A different mutation, W96L, was introduced into CDR3 of VL3 to generate VL5 (SEQ ID NO: 73). However, an antibody comprising VH0 and VL5 prepared using this VL region (i.e., OGD1.0.5) showed reduced activity compared to OGD1.0.0.

New mutations were also introduced into the heavy chain variable region. Two substitutions (i.e., M99F and N101D) were introduced into CDR3 to form VH4 (SEQ ID NO: 70). Antibodies comprising VH4 and VLO (i.e., OGD1.4.0) prepared using this VH region had significantly reduced activity, which correlated with reduced binding affinity for GDF-8. In CDR2, a G53S substitution was introduced to generate VH5 (SEQ ID NO: 72). Antibodies comprising VH5 and VLO (i.e., OGD1.5.0) prepared using this VH region retained substantial activity.

Table 10: Expression and activity of further humanized antibody VH and VL regions

All publications, patents, patent applications, and other documents cited in this application are hereby incorporated by reference in their entirety for all purposes to the same extent as if each individual publication, patent, patent application, or other document were individually indicated to be incorporated by reference for all purposes.

While particular embodiments have been illustrated and described, it will be appreciated that various changes can be made without departing from the spirit and scope of the invention.

Claims (59)

1. An antibody that specifically binds to GDF-8, said antibody comprising: The antibody heavy chain variable (VH) region, wherein VH CDR1 is composed of SEQ ID NO:10 or SEQ ID NO:20, VH CDR2 is composed of SEQ ID NO:11 or SEQ ID NO:21, and VH CDR3 is composed of SEQ ID NO:12; and The antibody light chain variable (VL) region, wherein VL CDR1 is composed of SEQ ID NO:13, VL CDR2 is composed of SEQ ID NO:14, and VL CDR3 is composed of SEQ ID NO:15; In the VH region, the amino acid position corresponding to residue number 111 (Kabat position 108) of SEQ ID NO:44 is leucine. 2. The antibody of claim 1, wherein the fourth frame region (FR4) of the VH region comprises amino acids 106 to 116 of SEQ ID NO:44. 3. The antibody of claim 1, wherein in the VL region, the amino acid position corresponding to residue number 100 (Kabat position 100) of SEQ ID NO:46 is glycine. 4. The antibody of claim 1, wherein the amino acid position corresponding to residue number 100 (Kabat position 100) of SEQ ID NO:46 in the VL region is glutamine. 5. The antibody of claim 3, wherein the fourth frame region of the VL region comprises amino acids 98 to 107 of SEQ ID NO:46. 6. The antibody of claim 4, wherein the fourth frame region of the VL region comprises amino acids 98 to 107 of SEQ ID NO:9. 7. The antibody of claim 1, wherein the VH region comprises the amino acid sequence SEQ ID NO:44. 8. The antibody of claim 7, wherein the VL region is composed of the amino acid sequence SEQ ID NO:46. 9. The antibody of claim 7, wherein the VL region is composed of the amino acid sequence SEQ ID NO:9. 10. The antibody of claim 1, further comprising an antibody heavy chain constant (CH) region derived from an antibody subtype selected from IgA, IgG, IgD, IgE or IgM. 11. The antibody of claim 10, wherein the IgG subtype is further selected from IgG1, IgG2, IgG3 and IgG4. 12. The antibody of claim 11, wherein the IgG subtype is IgG1. 13. The antibody of claim 10, wherein the heavy chain of the antibody comprises the amino acid sequence SEQ ID NO:57. 14. The antibody of claim 1, further comprising an antibody light chain constant (CL) region. 15. The antibody of claim 14, wherein the CL region is a κ or λCL region. 16. The antibody of claim 14, wherein the light chain of the antibody comprises the amino acid sequence SEQ ID NO:17. 17. The antibody of claim 7, wherein the VH region is encoded by the nucleic acid sequence SEQ ID NO:43. 18. The antibody of claim 1, wherein the antibody binds to GDF-8 with a Kd of at least ^10⁻⁶ M, ^10⁻⁷ M, ^10⁻⁸ M, ^10⁻⁹ M, ^10⁻¹⁰ M or ^10⁻¹¹ M. 19. The antibody of claim 1, wherein the antibody is selected from Fab, F(ab') ₂ , scFv, scFv-Fc, scFv-CH, scFab, scFv-zipper, biantibody, triple-chain antibody, quadruple-chain antibody, microantibody, Fv and bispecific antibody. 20. The antibody of claim 1, wherein the amount of said antibody produced by transfected cells is greater than the amount of antibody produced under similar conditions wherein the amino acid at position 108 of the Kabat region in the VH region is methionine, and all other aspects are identical. 21. The antibody of claim 20, wherein the amount of said antibody produced by transfected cells exceeds the amount of said other identical antibodies produced under similar conditions by at least the following amounts: 1.5 times, 2 times, 3 times, 4 times, 5 times, 6 times, 7 times, 8 times, 9 times, 10 times, 11 times, 12 times, 13 times, 14 times, 15 times, 20 times, and 30 times. 22. An antibody that specifically binds to GDF-8, said antibody comprising: a VH region consisting of the amino acid sequence SEQ ID NO:44, and a VL region consisting of the amino acid sequence SEQ ID NO:46, wherein the amino acid at position 108 of Kabat is leucine. 23. An antibody that specifically binds to GDF-8, comprising an antibody heavy chain defined by the amino acid sequence SEQ ID NO:58 and an antibody light chain defined by the amino acid sequence SEQ ID NO:59. 24. The antibody of claim 23, wherein the antibody comprises two antibody heavy chains, each defined by a free amino acid sequence SEQ ID NO:58, and two antibody light chains, each defined by a free amino acid sequence SEQ ID NO:59. 25. A pharmaceutical composition comprising the antibody of claim 1 and a pharmaceutically acceptable carrier. 26. An isolated polynucleotide comprising both a heavy chain and a light chain nucleic acid sequence encoding an antibody that specifically binds to GDF-8, said antibody comprising: The antibody heavy chain variable (VH) region, wherein VH CDR1 is composed of SEQ ID NO:10 or SEQ ID NO:20, VH CDR2 is composed of SEQ ID NO:11 or SEQ ID NO:21, and VH CDR3 is composed of SEQ ID NO:12. In the VH region, the amino acid position corresponding to residue number 111 (Kabat position 108) of SEQ ID NO:44 is leucine; and The antibody light chain variable (VL) region, wherein VL CDR1 is composed of SEQ ID NO:13, VL CDR2 is composed of SEQ ID NO:14, and VL CDR3 is composed of SEQ ID NO:15. In the VL region, the amino acid position corresponding to residue number 100 (Kabat position 100) of SEQ ID NO:46 is glycine. 27. The isolated polynucleotide of claim 26, wherein the nucleic acid sequence encoding the VH region is nucleic acid sequence SEQ ID NO:43 or SEQ ID NO:49, and wherein the nucleic acid sequence encoding the VL region is nucleic acid sequence SEQ ID NO:45 or SEQ ID NO:51. 28. The isolated polynucleotide of claim 26, further comprising a nucleic acid sequence encoding the CH region of a human antibody. 29. The isolated polynucleotide of claim 26, further comprising a nucleic acid sequence encoding the CL region of a human antibody. 30. The isolated polynucleotide of claim 29, wherein the CL region is encoded by the nucleic acid sequence SEQ ID NO:16. 31. An expression vector comprising the polynucleotide of claim 26, wherein the expression vector expresses both the heavy chain and the light chain. 32. A host cell comprising the polynucleotide of claim 26 operablely linked to a regulatory sequence. 33. A method for generating an antibody that specifically binds to GDF-8, comprising the steps of culturing the host cell of claim 32 and recovering the antibody thereby generated. 34. Use of the antibody of claim 1 in the preparation of a pharmaceutical composition that increases muscle mass or strength in mammals. 35. The use of claim 34, wherein the muscle is skeletal muscle or cardiac muscle. 36. The use of claim 35, wherein the skeletal muscle plays a role in respiration. 37. Use of the antibody of claim 1 in the preparation of a pharmaceutical composition for treating muscle diseases in objects requiring treatment of muscle diseases. 38. The use of claim 37, wherein the muscle condition is selected from muscular dystrophy, muscle atrophy, sarcopenia, cachexia, muscle wasting syndrome, age-related loss of muscle mass or strength, and asthenia. 39. The use of claim 38, wherein the muscle condition is muscular dystrophy. 40. The use of claim 39, wherein the muscular dystrophy is Duchenne muscular dystrophy. 41. The use of claim 40, wherein the treatment is to effectively improve the performance of said subject in a 6-minute walk test. 42. The use of claim 40, wherein the antibody is administered to a subject who is still being treated with glucocorticoids. 43. Use of the antibody of claim 1 in the preparation of a pharmaceutical composition for the prevention of muscle diseases in objects requiring prevention of muscle diseases. 44. The use of claim 43, wherein the muscle condition is selected from muscular dystrophy, muscle atrophy, sarcopenia, cachexia, muscle wasting syndrome, age-related loss of muscle mass or strength, and asthenia. 45. The use of claim 44, wherein the muscle condition is muscular dystrophy. 46. The use of claim 45, wherein the muscular dystrophy is Duchenne muscular dystrophy. 47. Use of the antibody of claim 1 in the preparation of a pharmaceutical composition for treating or preventing neuromuscular diseases in subjects requiring treatment. 48. The use of claim 47, wherein the neuromuscular disorder is ALS. 49. Use of the antibody of claim 1 in the preparation of a pharmaceutical composition for treating or preventing metabolic disorders in subjects requiring treatment of metabolic disorders. 50. The use of claim 49, wherein the metabolic disorder is selected from type 2 diabetes, metabolic syndrome, syndrome X, insulin resistance, and impaired glucose tolerance. 51. Use of the antibody of claim 1 in the preparation of a pharmaceutical composition for treating or preventing adipose tissue disorders in subjects requiring treatment. 52. The use of claim 51, wherein the adipose tissue condition is obesity. 53. Use of the antibody of claim 1 in the preparation of a pharmaceutical composition for treating or preventing bone loss syndrome in subjects requiring treatment of bone loss syndrome. 54. The use of claim 53, wherein the bone loss condition is selected from osteoporosis, osteopenia, osteoarthritis, and osteoporosis-related fractures. 55. An antibody that specifically binds to GDF-8, comprising: The antibody VH region composed of SEQ ID NO:44 and the antibody VL region composed of SEQ ID NO:46 The antibody expressed at a higher level under similar conditions than the second antibody, which is otherwise identical to the first antibody, which contains the VH region of SEQ ID NO:7 and the VL region of SEQ ID NO:9. 56. The antibody of claim 55, wherein the expression level of the antibody is at least 1.5 times, 2 times, 3 times, 4 times, 5 times, 6 times, 7 times, 8 times, 9 times, 10 times, 11 times, 12 times, 13 times, 14 times, 15 times, 20 times or 30 times higher than that of the second antibody. 57. The antibody of claim 56, wherein when expressed in COS cells, the antibody is expressed at least 12 times the amount of the second antibody. 58. The antibody of claim 56, wherein when expressed in CHO cells, the antibody is expressed at least 6 times the amount of the second antibody. 59. A host cell comprising a nucleic acid sequence encoding a heavy chain polypeptide and a light chain polypeptide of an antibody that specifically binds to GDF-8, said antibody comprising: The antibody heavy chain variable (VH) region, wherein VH CDR1 is composed of SEQ ID NO: 10 or SEQ ID NO: 20, VH CDR2 is composed of SEQ ID NO: 11 or SEQ ID NO: 21, and VH CDR3 is composed of SEQ ID NO: 12. Specifically, in the VH region, the amino acid position corresponding to residue number 111 (Kabat position 108) of SEQ ID NO: 44 is leucine; and The antibody light chain variable (VL) region, wherein VL CDR1 is composed of SEQ ID NO: 13, VL CDR2 is composed of SEQ ID NO: 14, and VL CDR3 is composed of SEQ ID NO: 15. Specifically, in the VL region, the amino acid position corresponding to residue number 100 (Kabat position 100) of SEQ ID NO: 46 is glycine. The nucleic acid sequence is operatively linked to one or more regulatory sequences.