GROWTH DIFFERENTIATION FACTOR 15 AGONIST COMPOUNDS AND METHODS OF USING THE SAME

Abstract

Compounds are provided herein for inducing weight loss and for treating diabetes, dyslipidemia, NASH and/or obesity. Also provided are pharmaceutical compositions containing such compounds and therapeutic uses of such compounds and compositions, where such compounds act as GDF15 agonists with extended time of action and other advantageous properties.

Claims

1. A compound comprising SEQ ID NO:2.

2. A compound consisting of SEQ ID NO:2.

3. A homodimer comprising two compounds of claim 1, wherein the compounds are linked via at least one interchain disulfide bond between a cysteine residue of the first compound and a cysteine residue of the second compound.

4. A pharmaceutical composition comprising the compound of claim 1 and one or more pharmaceutically acceptable excipients.

5. A method of inducing weight loss in an individual, the method comprising the step of: administering to the individual an effective amount of the compound of claim 1 or the composition of claim 4.

6. A method of treating type 2 diabetes in an individual, the method comprising the step of: administering to the individual an effective amount of the compound of claim 1 or the composition of claim 4.

7. A method of treating obesity in an individual, the method comprising the step of: administering to the individual an effective amount of the compound of claim 1 or the composition of claim 4.

8. A method of treating nonalcoholic steatohepatitis (NASH) in an individual, the methods comprising the step of: administering to the individual an effective amount of the compound of claim 1 or the composition of claim 4.

9. A method of treating dyslipidemia in an individual, the method comprising the step of: administering to the individual an effective amount of the compound of claim 1 or the composition of claim 4.

10. A compound produced by cultivating a mammalian cell comprising a cDNA molecule encoding a polypeptide comprising an amino acid sequence of SEQ ID NO:2 under such conditions that the polypeptide is expressed, and recovering the compound.

Description

EXAMPLES

[0034] The following non-limiting examples are offered for purposes of illustration, not limitation.

Example 1: In Vitro Receptor Affinity

[0035] In vitro binding parameters of the human GDF15 homodimer and exemplary Compound 2 to a human GFRAL receptor are determined by Surface Plasmon Resonance (SPR). The ECD of GFRAL includes three Cys-rich domains (D1, D2 and D3). The affinity of the homodimer of human GDF15 (des-ARN SEQ ID NO:1; that is, GDF15 lacking Ala, Arg and Asp from the N-terminal end) and Compound 2, a homodimer of the Fc-GDF15 Compound 1 (SEQ ID NO:2) to HIS-tagged human GFRAL ECD (SEQ ID NO:4) may be measured by SPR. The kinetics are summarized in Table 1.

[0036] To measure human GDF15 (des-ARN SEQ ID NO:1) or Compound 2 binding to human GFRAL ECD, a Biacore T200 Instrument is used to measure the binding kinetics of human GDF15 homodimer (des-ARN SEQ ID NO:1) and Compound 2 to the GFRAL ECD. Measurements are performed at 25 C. Samples are dissolved in HBS-EP+, pH 7.4 (GE Healthcare; 10 mM HEPES pH 7.4+150 mM NaCl+3 mM EDTA+0.05% surfactant P20). Goat anti-rabbit GDF15 antibody (R&D) is immobilized on the flow cells of a CM5 sensor chip using amine coupling chemistry. GDF15 samples are prepared at 25 g/ml by dilution into running buffer.

[0037] Human GFRAL ECD samples are prepared at final concentrations of 1000, 500, 250, 125, 62.5, 31.3, 15.6, 7.8, 3.9, 2.0 and 0 (blank) nM by dilution into running buffer. The SPR experiment includes the following steps: (1) capturing GDF15 samples on separate flow cells (Fc2, Fc3 and Fc4), (2) injecting 200 l of human GFRAL ECD over all flow cells at 100 l/min, (3) dissociating for 600 sec, (4) regenerating chip surfaces with a 10 l (30 sec) injection of glycine, pH 1.5, at 20 l/min, and (5) equilibrating chip surfaces with a 10 l (30 sec) injection of HBS-EP+ at 20 l/min. Each concentration of human GFRAL ECD is injected in duplicate.

[0038] Data are processed using standard double-referencing and fit to a 1:1 binding model using Biacore T200 Evaluation Software to determine the association rate (k.sub.on, M.sup.1s.sup.1), dissociation rate (k.sub.off, s.sup.1) and maximum response (Rmax; response units (RU)). The equilibrium dissociation constant (K.sub.D) is calculated from the relationship K.sub.D=k.sub.off/k.sub.on (M).

TABLE-US-00001 TABLE 1 Binding Kinetics of Human GDF15 Homodimer and Compound 2 Homodimer to Human GFRAL ECD (D1-D2-D3-ECD) Receptor at 25 C. Compound k.sub.on (1/Ms) k.sub.off (1/s) K.sub.D (nM) Human GDF15 2.92E+05 7.21E03 7.2 (homodimer of two compounds including des-ARN SEQ ID NO: 1) Compound 2 1.05E+06 8.71E03 8.3 (homodimer of two compounds including SEQ ID NO: 2)

[0039] The results show that the human GDF15 homodimer (two compounds comprising des-ARN SEQ ID NO:1) had affinity of 7.2 nM for human GFRAL ECD, and the Compound 2 homodimer (two compounds comprising SEQ ID NO:2) had affinity of 8.3 nM for human GFRAL ECD.

Example 2: In Vitro Functional Activity

[0040] GFRAL, the receptor for GDF15, is a member of the GDNF receptor family and contains a single transmembrane domain and no intracellular signaling domain. Members of the GDNF receptor family including GFRAL utilize the co-receptor RET to transduce a signal. Stimulation of the GDNF receptor/RET complex is known to activate signal transduction networks involving ERK, AKT and STAT phosphorylation.

[0041] To assess the activity and relative potency of mature human GDF15 variants, HEK293 cells (ATCC) are transfected with CMV-driven mammalian expression vectors containing human GFRAL (NCBI Ref. Sequence No. NP_997293.2) and human RET51 (NCBI Ref. Sequence No. NP_066124.1) and grown for three weeks in the presence of geneticin and puromycin (Life Technologies) to create a pool of cells stably expressing both proteins. Measurement of ERK1/2 phosphorylation is chosen as an endpoint to assess RET activation utilizing an AlphaLISA Surefire Ultra Phospho-ERK1/2 Assay Kit (Perkin Elmer), according to the manufacturer's instructions, and an EnVision Multilabel Microplate Reader Model 2102 (Perkin Elmer).

[0042] Briefly, HEK293-hGFRAL/hRET cells, between Passage 10-20, are seeded at 20,000 cells per well on poly-D-lysine-coated 96-well plates and grown for three days at 37 C. in a 5% CO.sub.2/95% O.sub.2 atmosphere. Cells are incubated in serum-free media for 4 hrs and then treated, as indicated, for 10 min, lysed and frozen until assayed. Cell lysates are thawed by incubating for 1 hr at 4 C., mixed on an orbital plate shaker, and transferred to a white 384-well plate. The AlphaLISA 2-plates/i-incubation method is utilized by adding Acceptor and Donor Mixes, supplied with the assay kit, to the lysates and incubating at room temperature in the dark for 8 hrs prior to measurement of AlphaLISA signal on the EnVision Plate Reader.

[0043] Each plate of cells used for an assay includes native human mature GDF15 as a reference standard for assay performance. ERK1/2 phosphorylation (pERK1/2) data from multiple plates are collected in a single graph by converting raw AlphaLISA signal data on the Y-axis to % Maxima relative to the human GDF15 reference standard utilizing the following formula:

% Maxima=100*(xmin)/(maxmin),

where min=0 and represents no stimulation, max=100 and represents the stimulation observed at a saturating dose of compound, and x=any given treatment. EC.sub.50 for each compound is determined using a variable slope sigmoidal dose-response curve with four parameter logistic regression in GraphPad Prism version 7.00. The EC.sub.50 listed for each compound in Table 2 is determined by generating the geometric mean of the EC.sub.50 for 3 independent assays of each compound using the following formula:

Geomean=X1*X2*X3

[0044] HEK293 cells expressing both human GFRAL and human RET51 are used to assess the potency of Compound 2. Native human mature GDF15 (SEQ ID NO:1), as a homodimer, is used as a reference standard on each plate of cells. Test Articles are run in quadruplicate in three independent experiments.

TABLE-US-00002 Compound 2 Potency Comparison to Human GDF15. Test Article EC.sub.50 (pM) SEM N Mature human GDF15 27.9 2.9 3 (homodimer of two compounds including SEQ ID NO: 1) Compound 2 86.2 5.8 3 (homodimer of two compounds including SEQ ID NO: 2) Data from a representative experiment are shown as mean SE N, number of independent experiments

Example 3: In Vivo Efficacy in Normal Rat Models

[0045] Obesity is a common co-morbidity associated with diabetes and insulin resistance. GDF15 has been reported to suppress food intake, induce weight loss and improve glucose homeostasis. Therefore, a GDF15 compound may be effective in therapy for inducing weight loss and for conditions such as T2D, dyslipidemia, NASH and/or obesity. Compound 2 is administered in male Sprague Dawley rats for a period of 15 days with daily body weight and food intake measurements.

[0046] Normal male Sprague-Dawley rats (Envigo; 14 weeks old) are maintained on a regular chow diet upon arrival at the facility (TD2014; Teklad, Madison, Wis.) and are used in the following studies. Animals are individually housed in a temperature-controlled (24 C.) facility with 12-hour light/dark cycle (lights on 0600) and free access to food (TD2014) and water.

[0047] After a minimum of one week of acclimation to the facility, the rats are randomized according to their body weight, so that each experimental group of animals would have similar body weight. Body weights ranged from 332 to 369 grams.

[0048] All groups contain five rats. Vehicle or Compound 2 (dose range 0.01 to 0.3 mg/kg) dissolved in vehicle (PBS) are administered by subcutaneous (SC) injection (1 mL/kg) to ad libitum fed normal rats every 3 days for 15 days.

[0049] Subcutaneous injections are made on Day 1, 4, 7, 10 and 13. Daily body weight and food intake measurements are made throughout the study period. Percent body weight is calculated daily from each animal's initial body weight as follows:

% Body Weight=(Daily Body Weight/Initial Body Weight)100.

[0050] All data are presented as the meanSEM of five animals per group. The effective dose concentrations for percent weight loss and cumulative food intake are determined in GraphPad Prism using the non-linear fit tool. Statistical analysis is performed using repeated measures ANOVA, followed by Dunnett's method for multiple comparisons. Significant differences are identified at *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

TABLE-US-00003 TABLE 3 Effect of Compound 2 Treatment on Percent Body Weight and Cumulative Food Intake in Normal SD Rats. Treatment Body Weight (%) Cumulative Food Intake (g) Vehicle (1 ml/kg, SC) 106.69 0.54 264.48 7.28 Compound 2 (0.01 mg/kg) 93.34 2.13* 193.04 8.54* Compound 2 (0.03 mg/kg) 85.18 2.60* 155.42 11.13* Compound 2 (0.06 mg/kg) 81.80 1.66* 122.54 14.05* Compound 2 (0.1 mg/kg) 78.30 3.71* 115.74 15.75* Compound 2 (0.3 mg/kg) 73.57 1.56* 91.22 8.57* All data are presented as the mean SEM of 5 animals per group from Day 14. Statistical analysis was performed using repeated measures ANOVA, followed by Dunnett's method for multiple comparisons. Significant differences were identified at *p < 0.05.

[0051] Chronic treatment with Compound 2 in normal SD male rats resulted in a statistically significant reduction in food intake and a statistically significant reduction in percent body weight at study completion compared to vehicle treated animals in a dose-dependent manner. The cumulative food intake is reduced from 264.48 to 91.22 at the highest dose tested and the body weight reduction ranges from a 7% to a 26% reduction. The data appear to indicate that Compound 2 has a statistically significant effect in reducing body weight and food intake in normal SD rats in a dose dependent manner.

Example 4: In Vivo Efficacy in Diet-Induced Obese (DIO) Mice

[0052] To investigate Compound 2's effect on weight loss, metabolism, body composition and hepatic steatosis, Compound 2 is chronically dosed to C57/B16 DIO mice.

[0053] Male C57/B16 DIO mice (Taconic) 24 to 25 weeks old, are maintained on a calorie rich diet upon arrival at the facility (TD95217; Teklad, Madison, Wis.) and are used in the following studies. Animals are individually housed in a temperature-controlled (24 C.) facility with 12-hour light/dark cycle (lights on 2200) and free access to food (TD95217) and water. After a minimum of 2 weeks acclimation to the facility, the mice are randomized according to their body weight, so that each experimental group of animals would have similar body weight. The body weights range from 40 to 50 grams.

[0054] All groups contain five mice. Vehicle or Compound 2 (0.002, 0.006, 0.02, 0.06, and 0.2 mg/kg) are dissolved in vehicle (PBS) and are administered by subcutaneous (SC) injection (10 mL/kg) to adlibitum-fed DIO mice 30 to 90 min prior to the onset of the dark cycle every 3 days for 21 days. SC injections are made on Day 1, 4, 7, 10, 13, 16, 19 and 22. Body weight and food intake are measured daily throughout the study.

[0055] Absolute changes in body weight are calculated by subtracting the body weight of the same animal prior to the first injection of molecule. On Day 17, animals are fasted at 0600 and at 1300, blood is collected to measure fasting blood glucose and insulin. Blood glucose is measured by AccuChek Glucometers (Roche Diabetes Care, Inc.; Indianapolis, Ind.). Insulin is measured by ELISA (MSD, Rockville, Md.). Insulin resistance index is calculated by the formula:

HOMA IR=fasting plasma insulin(mIUl/L)*(fasting blood glucose(mmol/L)/22.5.

[0056] On Day 24, animals are sacrificed prior to the dark photoperiod, and the livers are removed and frozen. Liver triglycerides, determined from homogenates of livers collected at sacrifice, and plasma analytes are measured on the Hitachi Modular P Clinical Analyzer.

TABLE-US-00004 TABLE 4 Effect of Compound 2 Treatment on Body Weight and Cumulative Food Intake in DIO Mice on Day 16. Body Weight Change From Cumulative Treatment Initial Measurement (g) Food Intake (g) Vehicle (10 ml/kg, SC) 0.68 0.14 41.30 0.94 Compound 2 (0.002 mg/kg) 4.58 0.71* 33.22 0.99* Compound 2 (0.006 mg/kg) 8.64 0.41* 29.24 1.41* Compound 2 (0.02 mg/kg) 11.82 0.57* 23.00 1.86* Compound 2 (0.06 mg/kg) 9.02 0.57* 23.86 0.96* Compound 2 (0.2 mg/kg) 9.08 0.59* 27.52 1.30* All data are presented as mean SEM of 5 animals per group. Values are presented as mean SEM with n = 5. *p < 0.05 compared to vehicle group; one-way ANOVA, followed by Dunnett's method for multiple comparisons.

TABLE-US-00005 TABLE 5 Effect of Compound 2 Treatment on Serum Glucose, Cholesterol, Alanine Aminotransferase (ALT) and Liver Triglycerides in DIO Mice on Day 24. Liver Glucose Cholesterol ALT Triglycerides Treatment (mg/dL) (mg/dL) (IU/L) (mg/g tissue) Vehicle 339.18 30.58 271.80 12.02 310.80 34.92 210.10 11.82 (10 ml/kg, SC) Compound 2 317.52 17.95 220.80 4.89* 231.60 25.39 69.88 18.88* (0.002 mg/kg) Compound 2 261.00 20.43* 177.00 7.82* 159.60 73.32* 46.56 13.85* (0.006 mg/kg) Compound 2 223.38 16.02* 172.80 18.51* 87.00 13.04* 24.82 8.80* (0.02 mg/kg) Compound 2 245.94 10.10* 199.20 15.34* 114.60 22.72* 58.90 14.45* (0.06 mg/kg) Compound 2 241.02 8.88* 167.40 8.72* 112.80 19.15* 34.62 10.88* (0.2 mg/kg) Values are presented as mean SEM with n = 5. *p < 0.05 compared to vehicle group; one-way ANOVA, followed by Dunnett's.

TABLE-US-00006 TABLE 6 Effect of Compound 2 Treatment on Fasting Blood Glucose, Insulin and HOMA-IR in DIO Mice on Day 17. Fasting Glucose Fasting Insulin Treatment (mmol/L) (mIU/L/mL) HOMA-IR Vehicle 6.80 0.28 85.64 13.68 25.78 4.13 (10 ml/kg, SC) Compound 2 6.64 0.28 45.94 4.66* 13.51 1.32* (0.002 mg/kg) Compound 2 5.16 0.53* 26.44 2.95* 6.22 1.15* (0.06 mg/kg) All data were presented as mean SEM of 5 animals per group from Day 17. Statistical analysis was performed using repeated measures ANOVA, followed by Dunnett's method for multiple comparisons. Significant differences were identified at p < 0.05.

[0057] Compound 2 dose-dependently reduces body weight and food intake in male DIO mice, which indicates it could be used in therapy for inducing weight loss and for obesity (Table 4). Compound 2 also reduces serum glucose, which indicates it could be used in therapy for diabetes (Table 5). Compound 2 further reduces cholesterol, which indicates it could be used in therapy for dyslipidemia (Table 5). Compound 2 improves liver health (demonstrated by decrease of serum ALT) and reduces liver triglycerides, which indicates it could be used in therapy for NASH (Table 5). Fasting blood glucose and insulin testing done on Day 17 demonstrates increased insulin sensitivity and a reduced insulin resistance index, which further support the possibility of using Compound 2 in therapy for diabetes (Table 6).

Example 5: Stability of Different Fc-GDF15 Constructs

[0058] Protein aggregation in a pharmaceutical product is not desired because it lowers overall production yield, and it could trigger an immune response when administered to individuals. Thus, maintaining the compound in the monomeric state, or in this case a covalently bound homodimer, is preferred.

[0059] To test the propensity of different GDF15 fusion proteins to aggregate and/or degrade, various Fc-GDF15 fusion proteins may be expressed in mammalian CHO cells. Fc-GDF15 fusion proteins may be captured onto Mab Select Protein A (GE; Piscataway, N.J.) with PBS (pH 7.4) running buffer; briefly may be washed with running buffer to remove non-specifically bound material; and may be eluted with 20 mM acetic acid and 5 mM citric acid, pH 3. Fractions containing Fc-GDF15 proteins are pooled, and pH is neutralized by adding 1/10 volume of 1 M Tris pH 8.0.

[0060] Protein A capture pool is then analyzed by size exclusion chromatography (SEC) to determine the % High Molecular Weight aggregate (HMW), indicating how much aggregation had occurred. The SEC separation method is performed on a Tosoh Bioscience 3000SWXL, 5 m Column with dimensions 30 cm0.78 cm. Mobile phase is PBS1, 350 mM NaCl (pH 7.4) at a flow rate of 0.5 mL/min. Initial low concentration samples are applied as 10 mcL injections and monitored at an absorbance wavelength of 214 nm. The % LMW species indicates chemical stability, or small molecular weight degradation products of a compound.

TABLE-US-00007 TABLE 7 Comparison of Physical and Chemical Stability of Different Fc-GDF15 Fusion Proteins. Compound % HMW % LMW Compound 2 4.1 2.3 (including monomeric Fc) Compound 3 60.8 3.6 (including tandem Fc) Compound 4 2.0 49.1 (including heterodimeric Fc)

[0061] A tandem Fc-GDF15 fusion protein construct (two similar Fc portions of antibody is attached to one another end-to-end, containing a hinge region, and then is attached to a GDF15 molecule where the GDF15 portions then self-associate to create a dimer), Compound 3 (homodimer of two molecules, each comprising SEQ ID NO:5) has a 60.8% HMW compared to the 4.1% HMW of Compound 2. This suggests that Compound 2 has better physical stability, with lower aggregation and less risk of immunogenic aggregates than a tandem Fc-GDF15 fusion protein construct.

[0062] A heterodimeric Fc-GDF15 (two different Fc portions of an antibody that associate with one another side-by-side and where one Fc portion (Chain B, SEQ ID NO:17) is associated with another Fc portion which is attached to a GDF15 molecule (Chain A, SEQ ID NO:6) and where the GDF15 portions then self-associate to create a homodimer), Compound 4 (a homodimer of two molecules, each molecule comprising a heterodimer of SEQ ID NO:6 & SEQ ID NO: 17) has a high 49.1% LMW species compared to Compound 2, which has only 2.3% LMW, suggesting that Compound 2 has better chemical stability and decreased propensity to chemical or proteolytic cleavages. Chemical stability and avoiding degradation helps to retain the potency of a compound over long periods of time. Also, less degradation minimizes the risk that degradation byproducts will cause further potency problems by interfering with the active ingredient in a formulation or that degradation byproducts will aggregate and cause immunogenicity issues.

Example 6: Stability of Double Versus Single Fc Mutations

[0063] The unique structure of the Fc portion of the compounds herein, with two specific mutations, unexpectedly exhibits better stability and higher solubility at high concentrations than an Fc construct with either one of the mutations alone. Two Fcs with a single mutation are generated to disrupt the CH3/CH3 association in the IgG4-Fc and to create a monomeric Fc construct. The first was Compound 7, with a glutamine substituted for a phenylalanine at position 405 in a full length human IgG4 heavy chain, corresponding to position 173 in SEQ ID NO:2 (position 175 in SEQ ID NO:9). The second single mutant generated was Compound 8, an Fc with a glutamic acid substituted for a tyrosine at position 407 in a full length human IgG4 heavy chain, corresponding to position 175 in SEQ ID NO:2 (position 177 in SEQ ID NO:10). A dual Fc mutant, Compound 5 (SEQ ID NO:7), also is generated, and it contains both the Fc mutations from Compounds 7 and 8. Compound 5 is the same sequence as Compound 2 except an additional two amino acids, AP, are present at the N-terminus of the Fc region in Compound 5. Compounds 2, 5, 7 and 8 are lacking the hinge region of ESKYGPPCPPCP (SEQ ID NO:8) to ensure formation of the monomeric IgG4 Fc.

[0064] The compounds in this example are expressed in mammalian CHO cells, and the harvested media is captured onto Mab Select Protein A (GE) with PBS (pH 7.4) running buffer; briefly washed with running buffer to remove non-specifically bound material; and eluted with 20 mM acetic acid and 5 mM citric acid, pH 3.0. Fractions containing GDF15 proteins are pooled and pH is neutralized by adding 1/10 volume of 1 M Tris pH 8.0.

[0065] Neutralized protein A capture pools are then buffer exchanged in 10 mM citrate pH 5.0 in Amicon Ultra 0.5 ml Centrifugal Concentrator with 10 MWCO. The single and double Fc-GDF15 mutants are next attempted to concentrate up to or above 10 mg/ml by using Amicon Ultra 0.5 ml Centrifugal Concentrator with 10 MWCO.

TABLE-US-00008 TABLE 8 Maximum Solubility of GDF15 Fusions at pH 5.0. Max solubility at pH Compound 5.0 (mg/ml) Visual observations Homodimer of 1.7 Unable to achieve 10 mg/ml; Compound 7 observed precipitation during (SEQ ID NO: 9) concentration Homodimer of 6.7 Unable to achieve 10 mg/ml, Compound 8 observed precipitation (SEQ ID NO: 10) during concentration Homodimer of 9.8 Clear solution Compound 5 (SEQ ID NO: 7) Monomeric IgG4 12.0 Clear solution Fc-only control

[0066] The results show that single Fc mutants, the homodimers of Compound 7 and Compound 8, had reduced solubility when compared to the double Fc mutant, the homodimer of Compound 5. The homodimer of Compound 7 could not achieve 10 mg/ml due to precipitation and material loss. The homodimer of Compound 8 showed improved solubility when compared to the homodimer of Compound 7; however, it was not able to achieve 10 mg/ml due to material loss during the concentration process. The homodimer of Compound 5, with the novel double mutation in the Fc, retained solubility at 9.8 mg/ml. A monomeric IgG4 Fc only control was soluble up to 12 mg/ml.

Example 7: Comparison of Chemical Stability for Flexible or Rigid Linkers

[0067] To test the chemical stability of GDF15 compounds herein, such compounds may be dialyzed overnight at 4 C. into the following buffers: 10 mM citrate, pH6 (C6), 10 mM citrate, pH7 (C7), 10 mM Tris, pH 7.0 (T7), 10 mM Tris, pH 8.0 (T8) or 1PBS, pH 7.4 (PBS). To assess long-term stability of Fc-GDF15 fusions 1.0 mg/mL protein solutions in respective buffers may be stored at 4 C., 25 C. and 40 C. for 4 weeks, and the % LMW is determined by analytical SEC.

[0068] The SEC separation method may be performed on a Tosoh Bioscience 3000SWXL, 5 m Column with dimensions 30 cm0.78 cm. Mobile phase may be PBS, 350 mM NaCl, pH 7.4 at a flow rate of 0.5 mL/min. Initial low concentration samples may be applied as 10 mcL injections and monitored at an absorbance wavelength of 214 nm.

[0069] Changes in chemical stability, measured by % LMW, are evaluated relative to the time 0 samples. As determined by studies conducted essentially as described above, the % LMW growth for GDF15 compounds are summarized in Table 9.

[0070] The compounds tested are Compound 2, a homodimer of Compound 1 (SEQ ID NO:2), with a rigid G4-AP10-G4 linker, a homodimer of Compound 5 (SEQ ID NO:7) with a flexible L5Q linker (GGGGQGGGGQGGGGQGGGGQGGGGQ (SEQ ID NO: 12)), and a homodimer of Compound 6 (SEQ ID NO: 11), with a rigid G4S-AP10-G4S linker.

TABLE-US-00009 TABLE 9 Growth in % LMW for GDF15 Compounds at 1 mg/ml as Determined by Analytical SEC. % LMW Growth at 1 mg/ml after 4 weeks of storage at 4 C., 25 C. and 40 C. Citrate pH 7.0 or Tris pH 8.0 or PBS Citrate pH 6.0 Tris pH 7.0 pH 7.4 4 C. 25 C. 40 C. 4 C. 25 C. 40 C. 4 C. 25 C. 40 C. Compound 2 0 0 0.75 0 0.36 3.45 0.15 0.99 7.09 Linker: GGGG- AP10-GGGG Homodimer of 0.37 0.81 3.99 0.56 1.8 6.63 0.71 2.8 10.5 Compound 5 Linker: L5Q Homodimer of 1.65 0.76 2.7 1.52 1.2 5.76 1.83 1.52 5.65 Compound 6 Linker: GGGGS- AP10-GGGGS

[0071] The L5Q linker has a higher % LMW after storage at 25 C. and 40 C. than either of the AP10 linkers, suggesting a higher propensity to chemical degradation for GDF15 fusion proteins comprising an L5Q flexible linker than a protein comprising an AP10 (APAPAPAPAPAPAPAPAPAP; SEQ ID NO:13) rigid linker. Also, the homodimer of Compound 6, with a linker comprising G4S (GGGGS; SEQ ID NO: 14), has a higher % LMW than Compound 2, with a linker comprising G4 (GGGG; SEQ ID NO:15), in most conditions, suggesting greater chemical stability of Compound 2 with its G4-AP10-G4 linker.

[0072] The chemical and physical stability of the GDF15 compounds with rigid (Compound 2 and Compound 6) and flexible linkers (Compound 5) may be assessed at high concentration of 10 mg/ml. Compounds may be first dialyzed O/N at 4 C. into following buffers: (a) 10 mM citrate pH 6.5; (b) 10 mM citrate pH 6.5, 150 mM NaCl; and (c) 10 mM citrate pH 6.5, 0.02% Polysorbate 80. Excipients such as 150 mM NaCl and 0.02% Polysorbate are added to the formulation to test their effect on chemical and physical stability of the GDF15 fusions. Next, compounds may be concentrated to 10 mg/ml using 10,000 MWCO, 15 ml Millipore Spin Concentrators. After concentration, the % HMW and % LMW may be determined by SEC using the concentrated protein (t=0).

[0073] The SEC separation method may be performed on a Tosoh Bioscience 3000SWXL, 5 m Column with dimensions 30 cm0.78 cm. Mobile phase was PBS, 350 mM NaCl, pH 7.4 at a flow rate of 0.5 mL/min. 10 mg/ml samples are diluted to 1 mg/ml in their respective buffers and then are applied as 10 mcL injections and monitored at an absorbance wavelength of 214 nm.

[0074] The 10 mg/ml formulations may be incubated for 4 weeks at 4 C., 25 C. and 40 C. to assess longer-term stability under stress conditions. The % LMW and % HMW may be determined again at 4 weeks time (t=4w) by SEC. Changes in chemical stability (% LMW) may be evaluated relative to the time 0 samples.

TABLE-US-00010 TABLE 10 Chemical Stability Represented as Growth in % LMW for GDF15 Compounds at 10 mg/ml. % LMW Growth at 10 mg/ml after 4 weeks of storage at 4 C., 25 C. and 40 C. Citrate pH 6.5 + Citrate pH 6.5 + Citrate pH 6.5 150 mM NaCl 0.02% Polysorbate 80 4 C. 25 C. 40 C. 4 C. 25 C. 40 C. 4 C. 25 C. 40 C. Homodimer of 0.52 1.03 6.39 0.25 0.86 5.11 0.27 1.04 6.2 Compound 5 Linker: L5Q Homodimer of 0 0.2 2.73 0 0.3 2.52 0 0.31 2.8 Compound 6 Linker: GGGGS- AP10-GGGGS

TABLE-US-00011 TABLE 11 Physical Stability Represented as Growth in % HMW for GDF15 Compounds at 10 mg/ml. % HMW Growth at 10 mg/ml after 4 weeks of storage at 4, 25 and 40 C. Citrate pH 6.5 + Citrate pH 6.5 + Citrate pH 6.5 150 mM NaCl 0.02% Polysorbate 80 4 C. 25 C. 40 C. 4 C. 25 C. 40 C. 4 C. 25 C. 40 C. Homodimer of 0.06 0.23 3.04 0.05 0.19 7.3 0.1 0.24 3.13 Compound 5 Linker: L5Q Homodimer of 0 0.1 2.98 0 0.33 7.18 0 0 2.85 Compound 6 Linker: GGGGS- AP10-GGGGS

[0075] As shown in Table 10, the homodimer of Compound 5 at 10 mg/ml, containing a flexible linker, showed a higher level of % LMW than the homodimer of Compound 6, with a rigid linker, at 25 C. and 40 C., which indicates lower chemical stability.

[0076] As shown in Table 11, the homodimer of Compound 5 at 10 mg/ml, containing a flexible linker, showed a higher level of % HMW than the homodimer of Compound 6, with a rigid linker, which indicates lower physical stability. Low chemical and/or physical stability can cause precipitation and possible immunogenicity issues.

Example 8: Pharmacokinetic Behavior of GDF15 Compounds

[0077] The pharmacokinetic behavior of Compound 2 is tested in monkeys. Plasma concentrations of GDF15 compounds are determined by an ELISA method at Eli Lilly and Company (Indianapolis, Ind.). The sandwich ELISA uses a polyclonal goat anti-human GDF15 (R&D Systems/AF957) or monoclonal mouse anti-human/primate GDF15 antibody (R&D Systems/MAB957) coated at 1 or 3 g/mL as the capture reagent. The GDF15 compound then is detected with either a mouse anti-human IgG4 pFc-HRP antibody (Southern Biotech/9190-05) or a mouse anti-human IgG Fc-HRP antibody (Southern Biotech/9200-05). The plates are developed using the tetramethylbenzidine (TMB) substrate system (SeraCare/5120), and the enzymatic reaction is quenched with TMB Stop Solution (KPL/50-85-04).

[0078] Male Cynomolgus monkeys are administered a single subcutaneous dose (0.3 mg/kg) of Compound 2 in PBS (pH 7.4) at a volume of 0.5 mL/kg. Blood is collected from each animal at pre-dose and at 6, 24, 48, 72, 96, 168, 240, 336, 408 and 504 hrs post-dose for pharmacokinetic characterization.

TABLE-US-00012 TABLE 12 Individual and Mean Pharmacokinetic Parameters Following a Single 0.3 mg/kg Subcutaneous Dose to Male Cynomolgus Monkeys. Compound Ani- t.sub.1/2 T.sub.max C.sub.max AUC.sub.0-inf CL/F (Dose) mal (hr) (hr) (g/mL) (hr*g/mL) (mL/hr/kg) Compound 2 1 118 48 1.96 206 1.45 (0.3 mg/kg) 2 116 48 2.44 184 1.63 Mean 117 48 2.20 195 1.54 Abbreviations: AUC.sub.0-inf = area under the curve from time 0 hours to infinity, CL/F = clearance/bioavailability, T.sub.max = time to maximal concentration, C.sub.max = maximum observed plasma concentration, t.sub.1/2 = half-life.

[0079] The data in Table 12 show that the t.sub.1/2 in monkeys is measured at 117 hrs. This is an unexpectedly long t.sub.1/2 for a monomeric Fc molecule. Also, the pharmacokinetic profile in monkeys is stable for a prolonged window of time, which suggests that the pharmacokinetic profile of Compound 2 in humans also could be demonstrably stable for a prolonged window of time in the therapeutic range to enable a steady level of drug in the bloodstream with once weekly dosing.

SEQUENCES

[0080] The following nucleic and/or amino acid sequences are referred to in this disclosure and are provided below for reference.

TABLE-US-00013 SEQIDNO:1 MaturehumanGDF15 ARNGDHCPLGPGRCCRLHTVRASLEDLGWADWVLSPREVQVTMCIGACPS QFRAANMHAQIKTSLHRLKPDTVPAPCCVPASYNPMVLIQKTDTGVSLQT YDDLLAKDCHCI SEQIDNO:2 Fusionprotein EAAGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSQEDPEVQFNWYVDGV EVHNAKTKPREEQFNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKGLPSSI EKTISKAKGQPREPQVYTLPPSQEEMTKNQVSLTCLVKGFYPSDIAVEWE SNGQPENNYKTTPPVLDSDGSFQLESRLTVDKSRWQEGNVFSCSVMHEAL HNHYTQKSLSLSLGGGGGAPAPAPAPAPAPAPAPAPAPGGGGGDHCPLGP GRCCRLHTVRASLEDLGWADWVLSPREVQVTMCIGACPSQFRAANMHAQI KTSLHRLKPDTVPAPCCVPASYNPMVLIQKTDTGVSLQTYDDLLAKDCHC I SEQIDNO:3 Leadersequence METDTLLLWVLLLWVPGSTG SEQIDNO:4 Artificialsequence QTNNCTYLREQCLRDANGCKHAWRVMEDACNDSDPGDPCKMRNSSYCNLS IQYLVESNFQFKECLCTDDFYCTVNKLLGKKCINKSDNVKEDKFKWNLTT RSHEGFKGMWSCLEVAEACVGDVVCNAQLASYLKACSANGNPCDLKQCQA AIRFFYQNIPFNIAQMLAFCDCAQSDIPCQQSKEALHSKTCAVNMVPPPT CLSVIRSCQNDELCRRHYRTFQSKCWQRVTRKCHEDENCISTLSKQDLTC SGSDDCKAAYIDILGTVLQVQCTCRTITQSEESLCKIFQHMLHRKSCFNY PTLSNVKGMALYTRKHANKHHHHHH SEQIDNO:5 Artificialsequence ERKSSVECPPCPAPPVAGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHE DPEVQFNWYVDGVEVHNAKTKPREEQFNSTERVVSVLTVVHQDWLNGKEY KCKVSNKGLPAPIEKTISKTKGQPREPQVYTLPPSREEMTKNQVSLTCLV KGFYPSDIAVEWESNGQPENNYKTTPPMLDSDGSFFLYSKLTVDKSRWQQ GNVFSCSVMHEALHNHYTQKSLSLSPGGGGGSGGGGSGGGGSGGGGSGGG GSGGGGSGGGGSGGGGSERKSSVECPPCPAPPVAGPSVFLFPPKPKDTLM ISRTPEVTCVVVDVSHEDPEVQFNWYVDGVEVHNAKTKPREEQFNSTFRV VSVLTVVHQDWLNGKEYKCKVSNKGLPAPIEKTISKTKGQPREPQVYTLP PSREEMTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPMLDSDG SFFLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGGGGSGG GGASARNGDHCPLGPGRCCRLHTVRASLEDLGWADWVLSPREVQVTMCIG ACPSQFRAANMHAQIKTSLHRLKPDTVPAPCCVPASYNPMVLIQKTDTGV SLQTYDDLLAKDCHCI SEQIDNO:6 Artificialsequence DKTHTCPPCPAPALGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDP EVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKC KVSNKALPAPIEKTISKAKGQPREPQVYTLPPSREEMTKNQVSLWCLVKG FYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSKLTVDKSRWQQGN VFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSGGGGSGGGG SNGTHCPLGPGRCCRLHTVRASLEDLGWADWVLSPREVQVTMCIGACPSQ FRAANMHAQIKTSLHRLKPDTVPAPCCVPASYNPMVLIQKTDTGVSLQTY DDLLAKDCHCI SEQIDNO:7 Artificialsequence APEAAGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSQEDPEVQFNWYVD GVEVHNAKTKPREEQFNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKGLPS SIEKTISKAKGQPREPQVYTLPPSQEEMTKNQVSLTCLVKGFYPSDIAVE WESNGQPENNYKTTPPVLDSDGSFQLESRLTVDKSRWQEGNVFSCSVMHE ALHNHYTQKSLSLSLGGGGGQGGGGQGGGGQGGGGQGGGGQGGGGDHCPL GPGRCCRLHTVRASLEDLGWADWVLSPREVQVTMCIGACPSQFRAANMHA QIKTSLHRLKPDTVPAPCCVPASYNPMVLIQKTDTGVSLQTYDDLLAKDC HCI SEQIDNO:8 Fcfragment ESKYGPPCPPCP SEQIDNO:9 Artificialsequence APEAAGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSQEDPEVQFNWYVD GVEVHNAKTKPREEQFNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKGLPS SIEKTISKAKGQPREPQVYTLPPSQEEMTKNQVSLTCLVKGFYPSDIAVE WESNGQPENNYKTTPPVLDSDGSFQLYSRLTVDKSRWQEGNVESCSVMHE ALHNHYTQKSLSLSLGGGGGQGGGGQGGGGQGGGGQGGGGQGGGGDHCPL GPGRCCRLHTVRASLEDLGWADWVLSPREVQVTMCIGACPSQFRAANMHA QIKTSLHRLKPDTVPAPCCVPASYNPMVLIQKTDTGVSLQTYDDLLAKDC HCI SEQIDNO:10 Artificialsequence APEAAGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSQEDPEVQFNWYVD GVEVHNAKTKPREEQFNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKGLPS SIEKTISKAKGQPREPQVYTLPPSQEEMTKNQVSLTCLVKGFYPSDIAVE WESNGQPENNYKTTPPVLDSDGSFFLESRLTVDKSRWQEGNVFSCSVMHE ALHNHYTQKSLSLSLGGGGGQGGGGQGGGGQGGGGQGGGGQGGGGDHCPL GPGRCCRLHTVRASLEDLGWADWVLSPREVQVTMCIGACPSQFRAANMHA QIKTSLHRLKPDTVPAPCCVPASYNPMVLIQKTDTGVSLQTYDDLLAKDC HCI SEQIDNO:11 Artificialsequence APEAAGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSQEDPEVQFNWYVD GVEVHNAKTKPREEQFNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKGLPS SIEKTISKAKGQPREPQVYTLPPSQEEMTKNQVSLTCLVKGFYPSDIAVE WESNGQPENNYKTTPPVLDSDGSFQLESRLTVDKSRWQEGNVFSCSVMHE ALHNHYTQKSLSLSLGGGGGSAPAPAPAPAPAPAPAPAPAPGGGGSGDHC PLGPGRCCRLHTVRASLEDLGWADWVLSPREVQVTMCIGACPSQFRAANM HAQIKTSLHRLKPDTVPAPCCVPASYNPMVLIQKTDTGVSLQTYDDLLAK DCHCI SEQIDNO:12 Artificialsequence GGGGQGGGGQGGGGQGGGGQGGGGQ SEQIDNO:13 Artificialsequence APAPAPAPAPAPAPAPAPAP SEQIDNO:14 Artificialsequence GGGGS SEQIDNO:15 Artificialsequence GGGG SEQIDNO:16 Artificialsequence gaggccgccgggggaccatcagtcttcctgttccccccaaaacccaagga cactctcatgatctcccggacccctgaggtcacgtgcgtggtggtggacg tgagccaggaagaccccgaggtccagttcaactggtacgtggatggcgtg gaggtgcataatgccaagacaaagccgcgggaggagcagttcaacagcac gtaccgtgtggtcagcgtcctcaccgtcctgcaccaggactggctgaacg gcaaggagtacaagtgcaaggtctccaacaaaggcctcccgtcctccatc gagaaaaccatctccaaagccaaagggcagccccgagagccacaggtgta caccctgcccccatcccaggaggagatgaccaagaaccaggtcagcctga cctgcctggtcaaaggcttctaccccagcgacatcgccgtggagtgggaa agcaatgggcagccggagaacaactacaagaccacgcctcccgtgctgga ctccgacggctccttccagctcgagagcaggctaaccgtggacaagagca ggtggcaggaggggaatgtcttctcatgctccgtgatgcatgaggctctg cacaaccactacacacagaagagcctctccctgtctctgggtggcggtgg tggagctccagccccagctcctgctcctgcaccagcacctgcccccgctc cagcacccgcacctggaggagggggcggtgaccactgtcccctggggcct ggccggtgttgcagactgcatacagtccgggcctccctggaagacctggg ttgggcagattgggtgctctcacctcgcgaggttcaggtgaccatgtgca ttggggcttgtccctcacaatttcgcgcagctaacatgcacgcccaaatc aaaacctccctgcaccggcttaaaccggatactgttcctgctccctgctg cgtgccagcatcctacaaccccatggtgctgatccagaagacggatacag gggtgtcactgcagacctatgatgacctcctggccaaagattgccactgt atc SEQIDNO:17 Artificialsequence DKTHTCPPCPAPALAGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHED PEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYK CKVSNKALPAPIEKTISKAKGQPREPQVYTLPPSREEMTKNQVSLSCAVK GFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLVSKLTVDKSRWQQG NVFSCSVMHEALHNHYTQKSLSLSPGK