Co-crystal forms of novobiocin analogues and proline

Abstract

To provide a compound for treating neurodegenerative disorders such as diabetic peripheral neuropathy.SOLUTION: Characterized by an X-ray powder diffractogram comprising the following peaks: 14.76, 16.86, 19.00 and 21.05 ° 2 θ ± 0.20 ° 2 θ, as determined on a diffractometer with Cu-K α radiation at a wavelength of 1.54178 Å. A cocrystal of N - (2 - (5 - (((2R, 3R, 4S, 5R) -3, 4-dihydroxy-5-methoxy-6, 6-dimethyltetra-hydro -2H - pyran-2-yl) oxy) -3' - fiuoro - [l, l' - biphenyl] -2-yl) ethyl) - acetamide and L-proline (1:2).SELECTED DRAWING: None

Description

[Technical Field]

[0001] Priority claims This patent application claims the benefit of priority to U.S. Patent Application No. 62 / 627,570, filed February 7, 2018, the entire contents of which are incorporated herein by reference. [Background technology]

[0002] background Approximately 26 million Americans suffer from either type 1 or type 2 diabetes. Despite the use of insulin and oral antidiabetic drugs to help maintain normal blood glucose levels, approximately 60-70% of these individuals develop diabetic peripheral neuropathy (DPN). See Veves, A.; Backonja, M.; Malik, RA, Pain Med. 9 (2008) 660-674 (Non-Patent Document 1). Numerous small molecules based on the novobiocin skeleton have been reported to inhibit heat shock protein 90 (Hsp90) and have been shown to have significant neuroprotective properties in animal models and to be useful in reversing DPN symptoms. See BR Kusuma et al., J. Med. Chem. 55 (2012) 5797-5812 (Non-Patent Document 2); U.S. Patent No. 9,422,320 (Patent Document 1).

[0003] One such novobiocin analog ("novologue") is N-(2-(5-(((3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide (4), which has been reported to exhibit high neuroprotective activity. Kusuma (2012) further reported that novologue 4 has effects that depend on the presence of another heat shock protein, Hsp70, while other effects are independent of Hsp70. The precise role of Hsp70 in the mechanism of action of novologue 4 and related compounds has not been fully characterized. J. Ma et al., ACS Chem. Neurosci. 6(9) (2015) 1637-1648 (Non-Patent Document 3).

[0004] The synthesis of Novolog 4 has been reported following a procedure that leads to an amorphous solid, the physicochemical characterization of which precludes a definitive assignment of the stereochemistry at the 2-position (Kusuma (2012); U.S. Pat. No. 9,422,320), thereby allowing in principle for the existence of two possible anomers, 4a and 4b, as shown below. TIFF2026012851000001.tif34134

[0005] The published synthesis of 4 (also known as KU-596) shows 95.6% HPLC purity but does not demonstrate anomeric purity of the amorphous solid, as evidenced by the fact that only the 2-position of noviose lacks a definitive stereochemical assignment (Kusuma (2012)). [Prior art documents] [Patent documents]

[0006] [Patent Document 1] U.S. Patent No. 9,422,320 [Non-patent literature]

[0007] [Non-Patent Document 1] Veves, A.; Backonja, M.; Malik, RA, Pain Med. 9 (2008) 660-674 [Non-patent document 2] BR Kusuma et al., J. Med. Chem. 55 (2012) 5797-5812 [Non-patent document 3] J. Ma et al., ACS Chem. Neurosci. 6(9) (2015) 1637-1648 Summary of the Invention

[0008] overview This disclosure is premised on the surprising discovery that co-crystalline forms of Novolog 4a with L-proline or D-proline can be achieved in high yield, purity, and anomeric purity. The present forms further demonstrate significant improvements in bioavailability over the known amorphous form 4.

[0009] Thus, one embodiment of the present disclosure is a co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide and L-proline (1:2). The co-crystal is characterized by a powder X-ray diffractogram containing the following peaks: 14.76, 16.86, 19.00, and 21.05 °2θ ± 0.20 °2θ as determined on a diffractometer using Cu-Kα radiation at a wavelength of 1.54178 Å. This co-crystal is referred to herein as "Form B."

[0010] Another embodiment is a co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide and L-proline (1:2). The co-crystal is characterized by a powder X-ray diffractogram containing the following peaks: 9.20, 16.19, 18.45, and 24.51 °2θ ± 0.2 °2θ as determined on a diffractometer using Cu-Kα radiation at a wavelength of 1.54178 Å. This co-crystal is referred to herein as "Form D."

[0011] Additionally, an embodiment is a co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide and L-proline (1:1:1) existing as an acetone solvate. The co-crystal is characterized by a powder X-ray diffractogram containing the following peaks: 14.64, 17.53, 18.91, and 21.33 °2θ ± 0.20 °2θ as determined with a diffractometer using Cu-Kα radiation at a wavelength of 1.54178 Å. This co-crystal is referred to herein as "Form C."

[0012] In a further aspect, the disclosure relates to a co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide and L-proline in a molar ratio of about 1:1.2:0.6:0.1, respectively, as a solvate of methyl ethyl ketone and pyrazine. The co-crystal is characterized by a powder X-ray diffractogram containing the following peaks: 10.42, 14.62, 19.28, and 21.14 °2θ ± 0.20 °2θ, as determined with a diffractometer using Cu-Kα radiation at a wavelength of 1.54178 Å. This co-crystal is referred to herein as "Form G."

[0013] This disclosure describes Cu-K at a wavelength of 1.5405929 Å. α1 Also provided is a co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide and D-proline (1:1), characterized by a powder X-ray diffractogram containing the following peaks as determined by a diffractometer using radiation: 11.77, 14.52, 19.54, and 21.23 °2θ ± 0.20 °2θ.

[0014] The present disclosure further provides Cu-K at a wavelength of 1.5405929 Å. α1 Provided is a co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide and L-proline (1:1), characterized by a powder X-ray diffractogram containing the following peaks as determined by a radiation diffractometer: 8.52, 16.33, 19.50, and 21.22°2θ±0.20°2θ.

[0015] In another aspect, the present disclosure relates to a pharmaceutical composition comprising any one of the co-crystal forms described herein. The composition further comprises a pharmaceutically acceptable solid carrier. In some embodiments, the composition further comprises one or more additional co-crystals.

[0016] Another aspect of the present disclosure is a method for inhibiting heat shock protein 90 (Hsp90) in a subject, the method comprising administering to the subject a therapeutically effective amount of a cocrystal described herein.

[0017] The present disclosure also embodies a method for treating or preventing a neurodegenerative disorder in a subject suffering from the neurodegenerative disorder. The method comprises administering to the subject a therapeutically effective amount of a cocrystal described herein. In some embodiments, the neurodegenerative disorder is diabetic peripheral neuropathy (DPN).

[0018] Alternatively, according to another aspect, the present disclosure provides a method for preventing or reducing the likelihood of developing diabetic peripheral neuropathy in a subject suffering from type 1 or type 2 diabetes, the method comprising administering to the subject a therapeutically effective amount of a cocrystal described herein.

[0019] According to another aspect, the present disclosure provides a method for making a co-crystal of Form B. The method comprises reacting a combination of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide (4a) and L-proline in a molar ratio of about 1:1 to about 1:2 with C 1-6 - heating in an alkyl alcohol to a first temperature to provide a solution, then cooling the solution to a second temperature of about 30°C or less to provide a slurry of the co-crystal, and then stirring the slurry at the second temperature for about 72 hours or less.

[0020] Another embodiment is a method of making Form D of the co-crystal. The method includes heating a combination of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3′-fluoro-[1,1′-biphenyl]-2-yl)ethyl)-acetamide (4a) and L-proline in about a 1:1 molar ratio in EtOH or acetonitrile to a first temperature, and then cooling the solution to a second temperature of about 30° C. or less, thereby providing a suspension of the co-crystal. The suspension is then stirred at the second temperature for about 72 hours or less.

[0021] The present disclosure is embodied in yet another method directed to making a co-crystal of Form C. The method optionally includes step (a) refluxing equimolar amounts of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3′-fluoro-[1,1′-biphenyl]-2-yl)ethyl)-acetamide (4a) and L-proline in EtOH to provide a solution, and cooling the solution to a temperature of about 30° C. or less, thereby providing a solid product. The product of step (a), or alternatively a combination of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3′-fluoro-[1,1′-biphenyl]-2-yl)ethyl)-acetamide (4a) and L-proline in an approximately 1:1 molar ratio, is then stirred in acetone at a temperature of about 30° C. or less for about 72 hours or less, thereby providing a co-crystal.

[0022] According to another aspect, the present disclosure provides a method of making a co-crystal of Form G. The method includes combining N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3′-fluoro-[1,1′-biphenyl]-2-yl)ethyl)-acetamide (4a), L-proline, and pyrazine in a molar ratio of about 1:1:20, respectively, in a mixed solvent of mesyl ethyl ketone (MEK) and MeOH to provide a solution, and then stirring the solution to thereby provide the co-crystal.

[0023] Additionally, the present disclosure provides a method for preparing a hydroxybenzoate comprising the steps of: (a) reacting a hydroxybenzoate with N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3′-fluoro-[1,1′-biphenyl]-2-yl)ethyl)-acetamide (4a) in a molar ratio of about 1:1 with D-proline; 1-6and cooling the solution to a second temperature of about 30° C. or less, thereby providing a suspension of the co-crystal.

[0024] In an additional embodiment, the present disclosure provides a method for increasing the concentration of N-(2-(5-(((2S,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3′-fluoro-[1,1′-biphenyl]-2-yl)ethyl)-acetamide (4a) relative to N-(2-(5-(((2S,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3′-fluoro-[1,1′-biphenyl]-2-yl)ethyl)-acetamide (4b) in a composition comprising 4a and 4b. The method includes contacting the composition with proline in a solvent and subjecting the composition, proline, and solvent to crystallization conditions under which a co-crystal of 4a and proline is produced. The bulk co-crystal exhibits a higher concentration of 4a than in a composition containing 4a and 4b. [The present invention 1001] Co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide and L-proline (1:2), characterized by a powder X-ray diffractogram containing the following peaks: 14.76, 16.86, 19.00, and 21.05°2θ ± 0.20°2θ as determined with a diffractometer using Cu-Kα radiation at a wavelength of 1.54178 Å. [The present invention 1002] A cocrystal of the present invention 1001, wherein the powder X-ray diffractogram further comprises peaks at 12.14, 17.51, 18.89, and 19.41°2θ±0.20°2θ. [The present invention 1003] A cocrystal of 1001 or 1002 of the present invention, having an X-ray powder diffractogram substantially as shown in Figure 1. [The present invention 1004] Any of the cocrystals of 1001 to 1003 of the present invention, characterized by a differential scanning calorimetry (DSC) thermogram including an exotherm at about 211°C. [The present invention 1005] 1004. A co-crystal of the present invention having a DSC thermogram substantially as shown in FIG. [The present invention 1006] Co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide and L-proline (1:2), characterized by a powder X-ray diffractogram containing the following peaks: 9.20, 16.19, 18.45, and 24.51°2θ ± 0.20°2θ as determined with a diffractometer using Cu-Kα radiation at a wavelength of 1.54178 Å. [The present invention 1007] 1006. A cocrystal of the present invention, wherein the powder X-ray diffractogram further comprises peaks at 11.83, 17.16, 20.15, and 25.34°2θ±0.2°2θ. [The present invention 1008] A cocrystal of 1006 or 1007 of the present invention, having an X-ray powder diffractogram substantially as shown in Figure 14. [The present invention 1009] Any of the cocrystals of 1006 to 1008 of the present invention, characterized by a differential scanning calorimetry (DSC) thermogram containing an endotherm with an onset temperature of about 211.2°C. [The present invention 1010] A co-crystal of 1009 of the present invention, having a DSC thermogram substantially as shown in Figure 15. [The present invention 1011] Acetone solvate (1:1:1) of a co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide and L-proline, characterized by a powder X-ray diffractogram containing the following peaks: 14.64, 17.53, 18.91, and 21.33°2θ ​​± 0.20°2θ as determined on a diffractometer using Cu-Kα radiation at a wavelength of 1.54178 Å. [The present invention 1012] A cocrystal of the present invention 1011, whose powder X-ray diffractogram further comprises peaks at 12.10, 15.14, 18.26 and 19.56°2θ±0.2°2θ. [The present invention 1013] A cocrystal of 1011 or 1012 of the present invention, having a powder X-ray diffractogram substantially as shown in Figure 10. [The present invention 1014] The cocrystal of any of 1011 to 1013 of the present invention, characterized by a thermogravimetric analysis (TGA) thermogram including weight loss steps ending at about 150°C and about 220°C. [The present invention 1015] A co-crystal of the present invention 1014, having a TGA thermogram substantially as shown in Figure 11. [The present invention 1016] A co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide with L-proline, methyl ethyl ketone, and pyrazine in a molar ratio of approximately 1:1.2:0.6:0.1, characterized by a powder X-ray diffractogram containing the following peaks: 10.42, 14.62, 19.28, and 21.14°2θ ± 0.20°2θ as determined by a diffractometer using Cu-Kα radiation at a wavelength of 1.54178 Å. [The present invention 1017] A cocrystal of the present invention 1016, wherein the powder X-ray diffractogram further comprises peaks at 11.85, 14.93, 17.40, and 19.28°2θ±0.2°2θ. [The present invention 1018] A co-crystal of 1016 or 1017 of the present invention, having an X-ray powder diffractogram substantially as shown in Figure 17. [The present invention 1019] Any of cocrystals 1016 to 1018 of the present invention characterized by unit cell dimensions as follows: a=10.975 Å, b=10.310 Å, c=15.704 Å, α=90°, β=108.56°, and γ=90°. [The present invention 1020] Cu-K at a wavelength of 1.5405929 Å α1 Co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide and D-proline (1:1), characterized by a powder X-ray diffractogram containing the following peaks as determined by a radiation diffractometer: 11.77, 14.52, 19.54, and 21.23°2θ ± 0.20°2θ. [The present invention 1021] A cocrystal of 1020 of the present invention, wherein the powder X-ray diffractogram further comprises peaks at 8.45, 13.18, 16.95, and 19.12°2θ±0.2°2θ. [The present invention 1022] A co-crystal of 1020 or 1021 of the present invention, having an X-ray powder diffractogram substantially as shown in Figure 18. [The present invention 1023] Any of the cocrystals 1020 to 1022 of the present invention, characterized by a differential scanning calorimetry (DSC) thermogram including an endotherm at about 130°C. [The present invention 1024] A co-crystal of 1023 of the present invention, having a DSC thermogram substantially as shown in Figure 19. [The present invention 1025] Cu-K at a wavelength of 1.5405929 Åα1 A co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide and L-proline (1:1), characterized by a powder X-ray diffractogram containing the following peaks as determined by a radiation diffractometer: 8.52, 16.33, 19.50, and 21.22°2θ±0.20°2θ. [The present invention 1026] A cocrystal of 1025 of the present invention, wherein the powder X-ray diffractogram further comprises peaks at 9.19, 13.22, 14.75, and 17.57°2θ±0.2°2θ. [The present invention 1027] A co-crystal of 1025 or 1026 of the present invention, having a powder X-ray diffractogram substantially as shown in Figure 23. [The present invention 1028] Any of the cocrystals of inventions 1025 to 1027, characterized by the following unit cell dimensions: a = 10.126 Å, b = 11.021 Å, c = 30.259 Å, α = 90°, β = 90°, and γ = 90°. [The present invention 1029] Any of the cocrystals 1025 to 1027 of the present invention, characterized by a differential scanning calorimetry (DSC) thermogram including an endotherm at about 145°C. [The present invention 1030] A co-crystal of 1029 of the present invention, having a DSC thermogram substantially as shown in Figure 24. [The present invention 1031] A pharmaceutical composition comprising any one of the cocrystals of the present inventions 1001 to 1030 and a pharmaceutically acceptable solid carrier. [The present invention 1032] A method for inhibiting heat shock protein 90 (Hsp90) in a subject, comprising administering to the subject a therapeutically effective amount of a cocrystal of any of the cocrystals of inventions 1001 to 1030. [The present invention 1033] A method for treating or preventing a neurodegenerative disorder in a subject suffering from the disorder, comprising administering to the subject a therapeutically effective amount of a cocrystal of any one of 1001 to 1030 of the present invention. [The present invention 1034] The method of claim 1033, wherein the neurodegenerative disorder is diabetic peripheral neuropathy. [This invention 1035] A method for preventing or reducing the likelihood of developing diabetic peripheral neuropathy in a subject, the method comprising administering to the subject a therapeutically effective amount of a cocrystal of any of 1001 to 1030 of the present invention, wherein the subject is suffering from type 1 or type 2 diabetes. [The present invention 1036] 1. A method for increasing the concentration of N-(2-(5-(((2S,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3′-fluoro-[1,1′-biphenyl]-2-yl)ethyl)-acetamide (4a) relative to N-(2-(5-(((2S,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3′-fluoro-[1,1′-biphenyl]-2-yl)ethyl)-acetamide (4b) in a starting composition comprising 4a and 4b, the method comprising the steps of contacting the starting composition with proline in a solvent and subjecting the starting composition, proline, and solvent to crystallization conditions under which a co-crystal of 4a and proline is produced, co-crystal) exhibits a higher concentration of 4a than in the starting composition comprising 4a and 4b. [This invention 1037] 1036. The method of claim 1036, wherein the proline is L-proline. [The present invention 1038] 1036. The method of claim 1036, wherein the proline is D-proline. [This invention 1039] The method of any one of claims 1036 to 1038, further comprising heating the starting composition, proline, and solvent. [The present invention 1040] 1039. The method of any one of claims 1036 to 1039, wherein the concentration of 4a is determined by HPLC. [The present invention 1041] The method of any of claims 1036 to 1040, wherein the concentration of 4a in the bulk co-crystal of 4a and proline is about 3 to about 20% (w / w) higher than in the starting composition. [The present invention 1042] The method of any of claims 1036 to 1041, wherein the concentration of 4a in the bulk co-crystal of 4a and proline is about 5 to about 15% (w / w) higher than in the starting composition. [This invention 1043] The method of any of claims 1036 to 1042, wherein the concentration of 4a in the bulk co-crystal of 4a with proline is increased by about 5%, about 10%, or about 15% (w / w). [This invention 1044] A cocrystal of any of 1001 to 1030 for treating or preventing a neurodegenerative disorder in a subject suffering from the neurodegenerative disorder. [This invention 1045] A cocrystal of any of 1001 to 1030 for preventing or reducing the likelihood of developing diabetic peripheral neuropathy in a subject suffering from type 1 or type 2 diabetes. [The present invention 1046] Use of any of the cocrystals of inventions 1001 to 1030 in the manufacture of a medicament for treating or preventing a neurodegenerative disorder in a subject suffering from the neurodegenerative disorder. [This invention 1047] Use of any of the cocrystals of inventions 1001 to 1030 in the manufacture of a medicament for preventing or reducing the likelihood of developing diabetic peripheral neuropathy in a subject suffering from type 1 or type 2 diabetes. [Brief explanation of the drawings]

[0025] [Figure 1] FIG. 1 presents the X-ray powder diffraction (XRPD) pattern of Form B. [Figure 2] FIG. 2 is a differential scanning calorimetry (DSC) curve of Form B. [Figure 3]FIG. 3 is a thermogravimetric analysis (TGA) curve of Form B. [Figure 4] FIG. 4 is the dynamic water vapor sorption (DVS) curve for Form B. [Figure 5] FIG. 5 is an infrared (IR) spectrum of Form B. [Figure 6] FIG. 6 is a Raman spectrum of Form B. [Figure 7] FIG. 7 is an atomic displacement ellipsoid diagram of Form B as determined by single crystal X-ray structure analysis. [Figure 8] FIG. 8 is a calculated XRPD pattern of Form B based on a single crystal structure determination. [Figure 9] FIG. 9 shows a comparison of the calculated XRPD pattern of Form B (lower trace) with the experimental XRPD pattern of Form B (upper trace). [Figure 10] FIG. 10 presents the X-ray powder diffraction (XRPD) pattern of Form C. [Figure 11] FIG. 11 is a thermogravimetric analysis (TGA) curve of Form C. [Figure 12] FIG. 12 is an atomic displacement ellipsoid diagram of Form C as determined by single crystal X-ray structure analysis. [Figure 13] FIG. 13 is a calculated XRPD pattern of Form C based on a single crystal structure determination. [Figure 14] FIG. 14 presents the X-ray powder diffraction (XRPD) pattern of Form D. [Figure 15] FIG. 15 shows the DSC curve (lower trace) and TGA curve (upper trace) of Form D. [Figure 16] FIG. 16 is the dynamic water vapor sorption (DVS) curve for Form D. [Figure 17] FIG. 17 presents the X-ray powder diffraction (XRPD) pattern of Form G. [Figure 18] FIG. 18 presents the X-ray powder diffraction (XRPD) pattern of the 4a / D-proline cocrystal. [Figure 19] FIG. 19 shows the DSC curve (lower trace) and TGA curve (upper trace) of the 4a / D-proline cocrystal. [Figure 20] FIG. 20 is a dynamic vapor sorption (DVS) curve for 4a / D-proline cocrystal. [Figure 21] FIG. 21 shows the mean plasma concentrations of 4a in mice after a single oral dose of substance A (●) and amorphous 4a (◯). [Figure 22] Figure 22 shows the mean plasma concentrations of 4a in monkeys after single oral doses of substance A (●) and amorphous 4a (◯) by oral gavage, and of loosely encapsulated substance A (▼) and amorphous 4a (△). [Figure 23] FIG. 23 presents the powder X-ray diffraction (XRPD) pattern of 4a / L-proline cocrystal substance A. [Figure 24] FIG. 24 shows the DSC curve (lower trace) and TGA curve (upper trace) of 4a / L-proline cocrystal substance A. [Figure 25] FIG. 25 is a dynamic vapor sorption (DVS) curve for 4a / L-proline cocrystal substance A. DETAILED DESCRIPTION OF THE INVENTION

[0026] Detailed Description definition Abbreviations, acronyms, and terms used throughout this disclosure have the following meanings: TIFF2026012851000002.tif187128

[0027] preface As summarized above, studies of Novolog 4 highlighted the compound's superior potency in inhibiting Hsp90 independently of Hsp70 (Kusuma (2012) and Ma (2015)). The studies revealed potential drawbacks to the compound's synthesis, including a tendency to produce a mixture of α-anomer 4a and β-anomer 4b and low overall yields. Additionally, while the reported column chromatographic purification method for 4 is suitable for small-scale studies, even when the compound was approximately 95% pure (HPLC), the method is not practical for producing large-scale, pharmaceutically pure quantities of α-anomer 4a for drug development.

[0028] Therefore, we undertook various crystallization strategies to isolate 4a. However, we did not find conditions under which 4a could be separated from 4b by crystallization. We subjected amorphous 4a to a co-crystallization screen consisting of 28 co-formers and surprisingly found that L-proline and D-proline selectively co-crystallized with the α-anomer 4a. Furthermore, we found that L-proline and D-proline were the only co-formers tested that yielded any crystalline material suitable for definitive characterization (see Example 3).

[0029] cocrystal form Contacting compound 4 with the coformers L-proline or D-proline surprisingly results in the selective cocrystallization of 4a with either coformer (see Examples 2 and 10). Thus, cocrystallization achieves an amount of 4 that is highly enriched in 4a relative to 4b, as determined, for example, by HPLC. Thus, in certain embodiments, selective cocrystallization of 4a with L-proline reduces the concentration of the β-anomer and facilitates the removal of trace impurities. Consequently, formation of the 4a / L-proline cocrystal improves the purity of 4a (HPLC) from about 90% to at least 95%, 96%, 97%, or 98%. Subsequent recrystallization of the 4a / L-proline cocrystal further improves the purity of 4a to at least 97%, 98%, or 99%.

[0030] Similarly, in other embodiments, cocrystallization of a starting composition of 4a and 4b with D-proline, e.g., cocrystallization of 4a / 4b and D-proline in approximately equimolar amounts, results in a 4a / D-proline cocrystal in which the purity (i.e., concentration) of 4a is improved by at least 15%, 10%, 5%, or 3% compared to the concentration of 4a in the starting composition, as determined by HPLC. Thus, for example, the starting composition of 4a / 4b contains 4a at a concentration of about 93%, and the cocrystal obtained after cocrystallization with D-proline contains 4a at a concentration of about 98%. In some embodiments, the 4a / D-proline cocrystal contains 4a with a final purity of at least 85%, 90%, 95%, 97%, 98%, or 99%.

[0031] In other embodiments, a quantity of α-anomer 4a is produced and purified as a 4a / D-proline cocrystal by contacting it with D-proline, e.g., in equimolar amounts. The resulting concentration of 4a in the bulk cocrystal is, e.g., at least 1%, 2%, 3%, 4%, or 5% (HPLC) higher than the concentration of 4a in the starting amount of 4a. Each of these embodiments contemplates one or more optional recrystallization steps to further improve the purity of 4a in a given cocrystal.

[0032] As summarized herein above, co-crystallization also produces various co-crystalline forms, which are identified and distinguished from one another by one or more analytical techniques, including X-ray powder diffraction (XRPD), differential scanning calorimetry (DSC), and thermogravimetric analysis (TGA).

[0033] Form B Thus, one embodiment, designated Form B, is a co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide (4a) and L-proline in a 1:2 molar ratio, respectively. The powder X-ray diffractogram contains characteristic peaks at 14.76, 16.86, 19.00, and 21.05 °2θ ± 0.2 °2θ as determined on a diffractometer using Cu-Kα radiation at a wavelength of 1.54178 Å. In certain embodiments, the powder X-ray diffractogram further contains peaks at 12.14, 17.51, 18.89, and 19.41 °2θ ± 0.2 °2θ. In yet another embodiment, Form B is further characterized by substantially its entire powder X-ray diffractogram (see Figure 1).

[0034] The DSC curve for Form B is characteristic of this co-crystal, exhibiting an exotherm at about 211° C. According to one embodiment, Form B is characterized by a full DSC thermogram substantially as shown in FIG.

[0035] The present disclosure is further embodied in a method for making Form B (see Example 4). The method comprises reacting a combination of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide (4a) and L-proline in a molar ratio of about 1:1 to about 1:2 with C 1-6 -alkyl alcohol to provide a solution. In some embodiments, 4a is present as pure 4a, while in other embodiments, 4a is present in combination with the β-anomer 4b, such as the combination provided in the published synthesis of 4 (Kusuma 2012, supra). For example, 4a is present at 95%, 96%, 97%, 98%, or 99% (w / w). The combination is heated to a first temperature ranging from about 50°C to about 80°C. Exemplary C 1-6Examples of alkyl alcohols include methanol, ethanol, and n- and i-propanol. In one embodiment, the alcohol is ethanol. In one embodiment, the suitable first temperature is the boiling point of the alcohol under standard pressure. Thus, for example, if ethanol is the alcohol, the first temperature is the boiling point, i.e., about 78°C.

[0036] The method further comprises cooling the solution of 4a and L-proline to a second temperature of about 30° C. or less, thereby providing a slurry of the co-crystal. The slurry is stirred at the second temperature for about 72 hours or less. In some embodiments, the slurry is filtered to isolate Form B.

[0037] Form D The present disclosure is further embodied in a co-crystal of 4a and L-proline present in a 1:2 molar ratio, respectively, which is designated as Form D. Form D is characterized by a powder X-ray diffractogram containing the following peaks: 9.20, 16.19, 18.45, and 24.51 °2θ ± 0.2 °2θ, as determined by a diffractometer using Cu-Kα radiation at a wavelength of 1.54178 Å. A further embodiment relates to additional characteristic peaks occurring at 11.83, 17.16, 20.15, and 25.34 °2θ ± 0.2 °2θ. Form D is further characterized by its powder X-ray diffractogram substantially as shown in Figure 14.

[0038] This co-crystal is also characterized in that the DSC curve for Form D exhibits an endotherm at about 212.2° C. with an onset temperature of about 211.2° C. According to certain embodiments, Form D is characterized by a full DSC thermogram substantially as shown in FIG.

[0039] Certain embodiments of the present disclosure also relate to methods of making Form D. The method comprises heating a combination of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide (4a) and L-proline in an approximately 1:1 molar ratio in EtOH or acetonitrile to a first temperature. In some embodiments, 4a is present as pure 4a, while in other embodiments, 4a is present in combination with the β-anomer 4b, such as that provided in the published synthesis of 4 (Kusuma 2012, supra). For example, 4a is present at 95%, 96%, 97%, 98%, or 99% (w / w). The first temperature is selected from the range of about 70°C to about 85°C. A suitable temperature is achieved, for example, by refluxing the combination, ie, at the boiling point of acetonitrile, about 82°C.

[0040] The method further includes cooling the solution to a second temperature of about 30° C. or less, thereby providing a suspension of co-crystals, and then stirring the suspension at the second temperature for about 72 hours or less. According to some embodiments, the slurry is filtered to isolate, for example, Form D.

[0041] Form C The present disclosure is further embodied in an acetone solvate of a co-crystal of 4a and L-proline, present in a 1:1:1 molar ratio, respectively, which is designated as Form C (see Example 6). The co-crystal is characterized by a powder X-ray diffractogram containing the following peaks: 14.64, 17.53, 18.91, and 21.33 °2θ ± 0.2 °2θ, as determined by a diffractometer using Cu-Kα radiation at a wavelength of 1.54178 Å. More specifically, according to another embodiment, the powder X-ray diffractogram contains additional peaks at 12.10, 15.14, 18.26, and 19.56 °2θ ± 0.2 °2θ. These and additional peaks characteristic of Form C are shown in its powder X-ray diffractogram, substantially as shown in FIG. 10.

[0042] Form C is further characterized by reference to its TGA thermogram, which includes weight loss steps ending at about 150° C. and about 220° C. One embodiment relates to a TGA thermogram of Form C substantially as shown in FIG.

[0043] Form C is made by a method according to various embodiments of the present disclosure. Thus, in one embodiment, the method comprises refluxing equimolar amounts of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide (4a) and L-proline in EtOH to provide a solution, and cooling the solution to a temperature of about 30°C or less, thereby providing a solid product. The solid product is then stirred in acetone at a temperature of about 30°C or less for about 72 hours or less, thereby providing Form C.

[0044] Alternatively, a combination of 4a and L-proline in about a 1:1 molar ratio is stirred in acetone at a temperature of about 30° C. or less for about 72 hours or less, thereby providing Form C. In either of these embodiments, 4a is present either as pure 4a or in combination with the β-anomer 4b, such as that produced by the published synthesis of 4. In further embodiments, Form C is isolated, such as by filtration.

[0045] Form G The present disclosure further relates to a cocrystal of 4a with L-proline, which exists as a solvate of methyl ethyl ketone and pyrazine, designated Form G (see Example 9). As described in the Examples, XRPD indexing of Form G is consistent with a 1:1 molar ratio of 4a:L-proline; however, the indexing does not distinguish between similarly sized MEK and pyrazine molecules, making it difficult to determine a definitive amount of solvent using this analytical technique. Proton NMR analysis of Form G, however, determined the molar ratios of 4a to L-proline to MEK to pyrazine to be approximately 1:1.2:0.6:0.1, respectively. Form G is thus characterized by its XRPD diffractogram, as determined with a diffractometer using Cu-Kα radiation at a wavelength of 1.54178 Å, with characteristic peaks at 10.42, 14.62, 19.28, and 21.14 °2θ ± 0.2 °2θ. Further embodiments provide additional peaks at 11.85, 14.93, 17.40, and 19.28 °2θ±0.2 °2θ. Form G can also be characterized by its complete XRPD diffractogram, substantially as shown in FIG.

[0046] According to another embodiment, XRPD analysis of Form G further determined the unit cell parameters characterizing the co-crystal: a=10.975 Å, b=10.310 Å, c=15.704 Å, α=90°, β=108.56°, and γ=90°.

[0047] The present disclosure further relates to a method for making Form G. The method comprises combining N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3′-fluoro-[1,1′-biphenyl]-2-yl)ethyl)-acetamide (4a), L-proline, and pyrazine in a molar ratio of about 1:1:20, respectively, in a mixed solvent of mesyl ethyl ketone (MEK) and MeOH to provide a solution. In some embodiments, 4a is present as pure 4a, while in other embodiments, 4a is present in combination with the β-anomer 4b, such as the combination provided in the published synthesis of 4 (Kusuma 2012, supra). For example, 4a is present at 95%, 96%, 97%, 98%, or 99% (w / w). Typically, the mixed solvent is embodied with an excess of MEK relative to MeOH. That is, an exemplary ratio of MEK to MeOH is about 9:1 (v / v). The solution is then stirred, thereby providing Form G.

[0048] Substance A Another aspect of the present disclosure is a co-crystal of 4a and L-proline present in a 1:1 molar ratio, respectively, which is designated as substance A. Substance A exhibits a Cu-K α1

[0013] Material A is characterized by an X-ray powder diffractogram containing the following peaks, as determined by a diffractometer using radioactive rays: 8.52, 16.33, 19.50, and 21.22°2θ±0.20°2θ. A further embodiment relates to additional characteristic peaks occurring at 9.19, 13.22, 14.75, and 17.57°2θ±0.2°2θ. Material A is further characterized by its X-ray powder diffractogram substantially as shown in Figure 23.

[0049] According to another embodiment, XRPD analysis of substance A further determined the unit cell parameters characterizing the co-crystal: a=10.126 Å, b=11.021 Å, c=30.259 Å, α=90°, β=90°, and γ=90°.

[0050] The DSC curve of Substance A is also characteristic of this co-crystal, exhibiting an endotherm at about 145° C. According to one embodiment, Substance A is characterized by an overall DSC thermogram substantially as shown in FIG.

[0051] Material A is further characterized by reference to its TGA thermogram, which includes weight loss steps ending at about 160° C. and about 230° C. One embodiment relates to a TGA thermogram of Material A substantially as shown in FIG.

[0052] 4a / D-proline cocrystal In another embodiment, the present disclosure also provides a co-crystal of 4a and D-proline present in a 1:1 molar ratio (see Example 11). The co-crystal exhibits Cu-K at a wavelength of 1.5405929 Å. α1 It is characterized by its XRPD diffractogram having characteristic peaks at 11.77, 14.52, 19.54, and 21.23°2θ±0.20°2θ, as determined by a diffractometer using radioactive rays. Additional characteristic peaks occur at 8.45, 13.18, 16.95, and 19.12°2θ±0.2°2θ. These and even additional peaks characteristic of the cocrystal are shown in its powder X-ray diffractogram, substantially as shown in FIG. 18.

[0053] The DSC curve of crystalline 4a / D-proline also exhibits an endotherm at about 130° C., which is characteristic of this cocrystal. According to one embodiment, the cocrystal is characterized by a full DSC thermogram substantially as shown in FIG.

[0054] The co-crystal is further characterized by reference to its TGA thermogram, which includes two weight loss steps ending at about 150-160° C. and about 230° C., respectively. One embodiment relates to a TGA thermogram of the co-crystal substantially as shown in FIG.

[0055] The present disclosure further relates to a method for producing 4a / D-proline crystals. The method comprises reacting a combination of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3′-fluoro-[1,1′-biphenyl]-2-yl)ethyl)-acetamide (4a) and D-proline in a molar ratio of about 1:1 with C 1-6 and (b) cooling the solution to a second temperature of about 30° C. or less, thereby providing a suspension of the co-crystal. In some embodiments, 4a is present as pure 4a, while in other embodiments, 4a is present in combination with the β-anomer 4b, such as the combination provided in the published synthesis of 4 (Kusuma 2012, supra). For example, 4a is present at 95%, 96%, 97%, 98%, or 99% (w / w).

[0056] Purification method The surprising discovery that 4a selectively co-crystallizes with L-proline and D-proline yields, in certain embodiments, a method for purifying 4a from bulk quantities of 4. That is, co-crystallization of 4a with proline enriches the concentration of 4a relative to 4b in the resulting bulk sample of 4a / proline cocrystal. The method for increasing the concentration of 4a first contemplates a starting composition of 4a and 4b. The starting composition can be a bulk solid resulting from a published synthesis of 4 or from one of many alternative synthetic routes to 4 known or reasonably contemplated by those skilled in the art of organic synthesis. In addition, the starting composition can be a bulk solid consisting predominantly of 4a resulting from other purification means, such as column chromatography. The inventors surprisingly discovered that 4a resisted all crystallization attempts. Indeed, there were no conditions under which 4a was observed to exist in crystalline form. In any of these examples, the starting composition contains at least some 4b, such as 0.5 to about 10% (w / w).

[0057] A molar excess of proline, such as 1 to about 2 equivalents, is mixed with the starting composition in a solvent. In some embodiments, the proline is L-proline, and in other embodiments, the proline is D-proline. A mixture of L-proline and D-proline can be used. Any solvent that can substantially dissolve proline and the starting composition is suitable for this purpose. Exemplary solvents, such as any of the solvents described herein, include C 2 O 4 , C 3 O 4 , C 4 O 4 , C 5 O 4 , C 6 O 4 , C 7 O 4 , C 8 O 4 , C 9 O 4 , C 10 O 4 , C 12 O 4 , C 14 O 4 , C 16 O 4 , C 18 ... 1-6 In some embodiments of the method, it is beneficial to heat the mixture of the starting composition, proline, and solvent to facilitate dissolution. A suitable temperature for this purpose is the reflux temperature of the solvent.

[0058] The combination of the starting composition, proline, and solvent is then subjected to crystallization conditions to achieve co-crystallization of 4a and proline. Various crystallization techniques, such as any of those described herein, are useful in this context. In an exemplary embodiment, the warmed solution of the starting composition and proline is allowed to cool to room temperature. External cooling is performed to cool the solution below room temperature, promoting co-crystallization. Alternatively, or in combination, the solvent may be slowly evaporated. Either of these measures, alone or in combination with each other, can shift the dissolution equilibrium toward crystallization.

[0059] The resulting bulk cocrystal of 4a and proline is thereby enriched in 4a relative to the concentration of 4a in the starting composition. The corresponding concentration of 4b is reduced. Additionally, the method purifies 4a from other impurities. While any analytical technique capable of resolving and quantifying the components present in a mixture, including gas chromatography (GC) performed on a chiral stationary phase, is suitable for this purpose, a suitable method for quantifying the concentration of 4a is by HPLC. That is, for example, the concentration of 4a in the bulk cocrystal is about 3 to about 20%, or about 5 to about 15% (w / w) higher than in the starting composition. Alternatively, the enrichment of 4a in the bulk cocrystal is at least about 5%, about 10%, or about 15% (w / w) relative to the concentration of 4a in the bulk starting composition. That is, for example, the starting composition of 4 contains about 93% 4a and about 6% 4b as determined by HPLC (see Example 10(A)). As directed by the methods of the present invention, after co-crystallization with L-proline, the amount of 4b in the resulting bulk co-crystal is reduced to about 2.5%. In any of these embodiments, subsequent recrystallization of the 4a / proline co-crystal can further reduce the amount of 4b in the bulk material.

[0060] This disclosure refers to patterns, such as XRPD patterns, in terms of their characteristic peaks. The collection of such peaks is unique to a given co-crystal form within the uncertainty due to particular instruments and experimental conditions. Thus, for example, each XRPD peak is disclosed in terms of angles 2θ with an allowable uncertainty of ±0.2° 2θ, with the understanding that variations in characteristic peaks within this uncertainty in no way detract from the identity of the co-crystal form with its corresponding collection of characteristic peaks.

[0061] Pharmaceutical Compositions As another aspect, the present disclosure also contemplates pharmaceutical compositions containing the cocrystals described herein. As illustrated in the Examples, the cocrystals of the present invention surprisingly exhibit much greater bioavailability than 4a alone, i.e., as an amorphous solid. Therefore, pharmaceutical compositions can be formulated to contain lower concentrations of the cocrystals to achieve the same therapeutic effect compared to formulations containing amorphous 4a. Due to this benefit of the cocrystals of the present invention, the therapeutically effective amount of the cocrystals in the pharmaceutical compositions provides a dose of about 0.1 mg to about 1000 mg, adjusted as needed depending on the subject's body weight. Typical dosages can vary from about 0.01 mg / kg to about 100 mg / kg per day.

[0062] In that the aggregates are components of a pharmaceutically acceptable carrier, the pharmaceutical composition will further comprise one or more pharmaceutically acceptable excipients, diluents, adjuvants, stabilizers, emulsifiers, preservatives, colorants, buffers, or flavoring agents, in accordance with pharmaceutical compounding practices. Generally, pharmaceutical compositions are prepared using conventional materials and techniques, such as mixing and blending. In principle, the pharmaceutically acceptable carrier can be liquid, so long as the cocrystal maintains structural and structural stability, e.g., by not dissolving in the carrier. Generally, however, the pharmaceutically acceptable carrier, and hence the composition as a whole, will be solid.

[0063] According to some embodiments, the pharmaceutical composition further comprises one or more additional co-crystal forms as disclosed herein. For example, the composition comprises two forms, three forms, or four forms. An exemplary composition comprises Form B and Form D. Binary compositions, i.e., compositions containing only two forms, provide the forms in various weight ratios ranging from about 0.05:1 to about 1:0.05. Intermediate ratios and ranges are also contemplated, such as, for example, 0.2:1 to about 1:0.2, and 0.5:1 to about 1:0.5.

[0064] For tablet compositions, the cocrystals of the present invention are mixed with non-toxic pharmaceutically acceptable excipients for tablet manufacture. Examples of such excipients include, but are not limited to, inert diluents such as calcium carbonate, sodium carbonate, lactose, calcium phosphate, or sodium phosphate; granulating and disintegrating agents such as corn starch or alginic acid; binders such as starch, gelatin, or acacia; and lubricants such as magnesium stearate, stearic acid, or talc. The tablets may be uncoated or may be coated by known coating techniques to delay disintegration and absorption in the gastrointestinal tract and thereby provide a sustained therapeutic effect over a desired period of time. For example, a time-delay material such as glyceryl monostearate or glyceryl distearate may be used.

[0065] Formulations for oral use may also be presented as hard gelatin capsules in which the active ingredient is mixed with an inert solid diluent, such as calcium carbonate, calcium phosphate, or kaolin, or as soft gelatin capsules in which the active ingredient is mixed with water or an oil medium, such as peanut oil, liquid paraffin, or olive oil. These formulations, and all other liquid formulations described herein, are subject to the limitations detailed above in order to preserve the compositional and structural integrity of the solid cocrystal.

[0066] According to embodiments described below, the pharmaceutical composition is presented as a suspension, which is referred to as a "stable suspension," meaning that a given co-crystal or combination of co-crystals maintains its properties, such as XRPD peaks, even while in contact with the other components of the suspension, i.e., by not dissolving in the liquid excipients of the suspension, by not converting to another co-crystal or amorphous form, or both.

[0067] For aqueous suspensions, the co-crystals of the present invention are mixed with suitable excipients to maintain a stable suspension, including, but not limited to, sodium carboxymethylcellulose, methylcellulose, hydropropylmethylcellulose, sodium alginate, polyvinylpyrrolidone, gum tragacanth, and gum acacia.

[0068] Oral suspensions can also contain dispersants or wetting agents such as naturally occurring phospholipids, such as lecithin; or condensation products of alkylene oxides and fatty acids, such as polyoxyethylene stearate; or condensation products of ethylene oxide and long-chain aliphatic alcohols, such as heptadecaethyleneoxycetanol; or condensation products of ethylene oxide and partial esters derived from fatty acids and hexitols, such as polyoxyethylene sorbitol monooleate; or condensation products of ethylene oxide and partial esters derived from fatty acids and hexitol anhydrides, such as polyoxyethylene sorbitan monooleate. Aqueous suspensions can also contain one or more preservatives, such as ethyl p-hydroxybenzoate or n-propyl p-hydroxybenzoate, one or more colorants, one or more flavorings, and one or more sweeteners, such as sucrose or saccharin.

[0069] Oily suspensions can be formulated by suspending the active ingredient in a vegetable oil, for example arachis oil, olive oil, sesame oil or coconut oil, or in a mineral oil such as liquid paraffin. The oily suspensions may contain a thickening agent, for example beeswax, hard paraffin or cetyl alcohol.

[0070] Sweetening agents such as those set forth above, and flavoring agents may be added to provide a palatable oral preparation.These compositions may be preserved by the addition of an antioxidant such as ascorbic acid.

[0071] Dispersible powders and dispersible granules suitable for preparing an aqueous suspension by adding water can provide the active ingredient in admixture with a dispersing or wetting agent, a suspending agent, and one or more preservatives.Suitable dispersing or wetting agents and suspending agents are exemplified by those already mentioned above.Additional excipients, such as sweeteners, flavorings, and coloring agents, may also be present.

[0072] The pharmaceutical composition of the present invention can also be in the form of oil-in-water emulsion.Oil phase can be vegetable oil, such as olive oil or peanut oil, or mineral oil, such as liquid paraffin, or their mixture.Suitable emulsifiers can be naturally occurring gum, such as gum acacia or gum tragacanth; naturally occurring phospholipid, such as soybean, lecithin; and the ester or partial ester derived from fatty acid and hexitol anhydride, such as sorbitan monooleate; and the condensation product of this partial ester and ethylene oxide, such as polyoxyethylenesorbitan monooleate.Emulsion can also contain sweeteners and flavoring agents.

[0073] Syrups and elixirs can be formulated with sweetening agents, such as glycerol, propylene glycol, sorbitol, or sucrose. Such formulations may also contain demulcents, preservatives, flavorings, and coloring agents. The pharmaceutical compositions may be in the form of a sterile injectable, aqueous, or oleaginous suspension. These suspensions may be formulated according to known techniques using suitable dispersing or wetting agents, as well as suspending agents, as described above. Sterile injectable preparations may also be sterile injectable solutions or suspensions in non-toxic, parenterally acceptable diluents or solvents, such as solutions in 1,3-butanediol. Acceptable vehicles and solvents that can be used include water, Ringer's solution, and physiological saline solution. Additionally, sterile, fixed oils are conventionally used as solvents or suspending media. For this purpose, any bland, fixed oil may be used, including synthetic mono- or diglycerides. Additionally, fatty acids, such as oleic acid, are used in the preparation of injectable solutions.

[0074] The cocrystals of the present invention can also be administered in the form of suppositories for rectal administration of the cocrystals. These compositions can be prepared by mixing the cocrystals with a suitable non-irritating excipient that is solid at ambient temperature but liquid at rectal temperature, and thus will melt in the rectum and release the cocrystals. Examples of such materials are cocoa butter and polyethylene glycol.

[0075] Compositions for parenteral administration are administered in a sterile medium. Depending on the vehicle and concentration of cocrystals used in the formulation, the parenteral formulation may be a suspension of the cocrystals, provided that the particle size distribution of the cocrystals is suitable for this mode of administration. Adjuvants such as local anesthetics, preservatives, and buffers may also be added to the parenteral composition.

[0076] How to use One surprising advantage offered by the cocrystals of the present invention is the ability to produce large quantities of 4a with very high diastereomeric and chemical purity, as evidenced by the accompanying examples. This is particularly important for developing 4a in compliance with, for example, Good Manufacturing Practice (GMP) regulations promulgated by the U.S. Food and Drug Administration. In contrast, the synthesis of amorphous 4a, even when isolated with high chemical and diastereomeric purity for GMP purposes, remains inefficient—requiring subsequent laborious separation techniques such as chromatography, and crystallization attempts have all failed, as noted above. For these reasons, the cocrystals of the present invention and the methods for making them provide large quantities of 4a that are useful for clinical trials and commercialization efforts.

[0077] Another advantage of the cocrystals of the present invention is the unexpectedly high bioavailability of cocrystal-derived 4a compared to amorphous 4a. More specifically, in vivo administration of the cocrystal increased the bioavailability of 4a by approximately 1.5-2 times relative to the same dose of amorphous 4a (see Examples 12 and 13). This feature of the cocrystals of the present invention is all the more surprising in light of the general consensus that amorphous forms of pharmaceuticals are, in fact, significantly more soluble than their crystalline counterparts and therefore more bioavailable. See BC Hancock et al., Pharm. Res. 17(4) (2000) 397-404; BC Hancock et al., J. Pharm. Sci. 86(1) (1997) 1-12.

[0078] In light of these advantages, the present disclosure further relates to the use of any of the cocrystal forms, including pharmaceutical compositions thereof, for treating or preventing a neurodegenerative disorder in a subject suffering from the disorder. Non-limiting examples of such neurodegenerative disorders include Alzheimer's disease, Parkinson's disease, amyotrophic lateral sclerosis (ALS), Huntington's disease, spinal muscular atrophy, spinocerebellar ataxia and forms of ataxia, and demyelinating neuropathies, including motor neuron diseases. In addition, the cocrystal forms and pharmaceutical compositions thereof are useful for treating diabetic neuropathy (including both painful and numb forms thereof) and other forms of neuropathy, including non-diabetic neuropathic pain.

[0079] Cocrystal forms and pharmaceutical compositions thereof may also be useful for treating neurological disorders involving mitochondrial dysfunction, oxidative stress, or inflammation, given the reported use of compound 4a to improve impaired mitochondrial function in neurons and reduce the expression of inflammatory markers in diabetic neurons (Ma (2015)). Because diabetic tissues experience significant oxidative stress, these results indicate that the cocrystal forms of the present invention and pharmaceutical compositions thereof may also be useful for treating other neurological disorders involving oxidative stress and chronic inflammation, such as epilepsy, multiple sclerosis, spinal cord injury, and psychiatric disorders, including schizophrenia, depression, bipolar disorder, autism and related disorders, and post-traumatic stress disorder. Compositions in cocrystal form may be used in combination with other therapies, particularly those that reduce oxidative stress, inflammation, and mitochondrial dysfunction through other mechanisms.

[0080] As used herein, the term "neurodegenerative disorder" refers to a disorder in which progressive loss of neurons occurs in the peripheral nervous system or in the central nervous system. Thus, in one aspect, the present disclosure provides a method for inhibiting Hsp90 in a subject, for example, during or following the treatment of a neurodegenerative disorder, by inhibiting the progressive deterioration of neurons, which leads to cell death.

[0081] The methods described herein include administering to a subject a therapeutically effective amount of a cocrystal of the present invention. Within the dosing guidelines set forth above, a "therapeutically effective amount" is an amount of cocrystal that wholly or partially inhibits the progression of a disorder or at least partially alleviates one or more symptoms of the disorder. A therapeutically effective amount can also be a prophylactically effective amount. The amount that is therapeutically effective will depend on the size and sex of the patient, the disorder to be treated, the severity of the disorder, and the result desired. For a given patient and disorder, the therapeutically effective amount can be determined by methods known to those of skill in the art.

[0082] In various embodiments, the method includes preventing a neurological disorder. As used herein, the terms "preventing" or "prevention" mean that the cocrystals of the invention are useful when administered to a subject who has not been diagnosed as possibly having the disorder at the time of administration, but who is generally expected to develop the disorder or is considered to be at increased risk for the disorder. The cocrystals of the invention delay the onset of disorder symptoms, delay the onset of the disorder, or prevent the subject from developing the disorder altogether. Prevention also contemplates administration of the cocrystals to a subject who is considered to be predisposed to the disorder due to age, family history, genetic or chromosomal abnormalities, and / or the presence of one or more biological markers for the disorder.

[0083] In other embodiments, the methods of the invention include "treatment" or "treating," meaning that the cocrystals are used on a subject who has received at least a provisional diagnosis of a disorder. Thus, the cocrystals of the invention slow or delay the progression of the disorder. In addition, the term "treatment" encompasses at least the amelioration of symptoms associated with the disorder, where amelioration is used broadly to refer to at least a reduction in the severity of a parameter associated with the condition being treated, e.g., a symptom. Thus, "treatment" also encompasses situations in which a disorder or at least a symptom associated therewith is completely suppressed, e.g., prevented from occurring, or arrested, e.g., terminated, such that the subject no longer suffers from the disorder or at least a symptom characterizing the disorder.

[0084] In one embodiment, neurodegenerative disorder is the glucose toxicity of sensory neurons, resulting from the hyperglycemia associated with diabetic condition, for example.For example, the subject suffers from type 1 or type 2 diabetes.More specifically, according to an embodiment, neurodegenerative disorder is diabetic peripheral neuropathy.Therefore, in an embodiment, the method of the present invention comprises preventing or reducing the possibility of developing diabetic peripheral neuropathy in the subject suffering from type 1 or type 2 diabetes.

[0085] In the context of the methods and uses of the present invention, a "subject" to be treated with a cocrystal of the present invention is an animal, preferably a mammal, such as a dog, cat, mouse, monkey, rat, rabbit, horse, cow, guinea pig, or sheep. In some embodiments, the subject is a human. [Example]

[0086] The following non-limiting examples are provided to illustrate additional aspects of the present disclosure.

[0087] For example, following published procedures (Kusuma (2012) and US Pat. No. 9,422,320), amorphous 4a is obtained with approximately 95% purity (HPLC).

[0088] I. General Crystallization Experimental Methods Crash Precipitation (CP): Solutions of 4a were prepared by stirring in various solvents using various coformers at given molar ratios. Aliquots of various antisolvents were added with stirring until precipitation occurred. The mixtures were stirred for specified periods. Where specified, additional crystallization techniques were used.

[0089] Fast Evaporation (FE): Solutions of 4a were prepared by stirring various coformers in the given molar ratios in various solvents or solvent systems. Unless otherwise specified, each solution was evaporated from an open vial at ambient conditions. If a solid was isolated by the specified method or additional crystallization techniques were used as specified, the solution was evaporated to dryness unless specified as a partial evaporation (solids present with a small amount of solvent remaining).

[0090] Manual grinding: Weighed amounts of 4a and various coformers were transferred to an agate mortar. A small amount of a given solvent was added to the solids, and the mixture was manually ground with an agate pestle for a given time.

[0091] Milling:A weighed amount of 4a and the given coformer was transferred to an agate grinding vessel. A small amount of the given solvent and agate mill balls were added to the vessel, which was then attached to a Retsch mill. The mixture was milled for a specified duration of 30 seconds. -1 Between cycles, solids were scraped off the walls of the jar.

[0092] Reaction Crystallization (RC): Mixtures of 4a and various coformers were prepared in a given solvent by adding solids of one component to a solution of the second component. When enough solids were added so that the solution contained different concentrations of each component (typically a 10- to 20-fold molar difference between one component and the other), the solution was stirred for an extended period of time. If no precipitate formed, additional solids of the higher concentration component were added, and the mixture was again stirred for an extended period of time. Any precipitated solids were isolated and analyzed.

[0093] Slow cooling (SC): Highly concentrated solutions of 4a were prepared using various coformers at given molar ratios in various solvent systems at elevated temperatures with stirring. Each vial was capped and placed on a hot plate. The hot plate was then turned off and the samples were allowed to cool slowly to ambient temperature. If no solids were present after cooling to ambient temperature, the samples were placed in a refrigerator (approximately 2-8 °C) and / or freezer (approximately -10 to -25 °C) for further cooling. If no solids were present, additional crystallization techniques were used as specified.

[0094] Slow Evaporation (SE): Solutions of 4a were prepared in various solvent systems using various coformers at given molar ratios. Each solution was evaporated at ambient conditions in a pinhole-punctured aluminum foil-covered vial. Solutions were evaporated to dryness unless specified as partial slow evaporation, which evaporates a portion of the solvent. The resulting solid was isolated by the specified technique, or, where specified, additional crystallization techniques were used.

[0095] Slurry Experiments:Suspensions of 4a and various coformers in specified molar ratios were prepared by adding enough solid to a given solvent system at ambient conditions so that undissolved solid was present. The mixture was then agitated (typically by stirring) in a sealed vial at specified conditions for an extended period of time. The solid was either collected by a specified technique, or, if specified, additional crystallization techniques were used.

[0096] Vapor Diffusion (VD): Highly concentrated solutions of a given starting material (either a given form of 4a / L-proline or a defined stoichiometric mixture of 4a and L-proline) were prepared in various solvents. In some cases, the solutions were filtered through a 0.2 μm nylon filter. Each solution was dispensed into a small vial, which was then placed inside a larger vial containing the given antisolvent. Where specified, seed crystals of the given form of 4a / L-proline were added to the solution. The small vial was left uncapped, while the larger vial was capped to allow vapor diffusion. Where specified, additional crystallization techniques were attempted.

[0097] Vacuum filtration: The solids were collected on filter paper or nylon filters by vacuum filtration and allowed to air dry on the filters under reduced pressure for a short period before being transferred to vials.

[0098] Interconversion slurry: A solution of a given starting material (either a given form of 4a / L-proline or a specified stoichiometric mixture of 4a and L-proline) was prepared by adding the solid to a given solvent system at a specified temperature. When a saturated solution was specified, the suspension was stirred at ambient temperature for an extended period of time to ensure saturation of the liquid phase. Seed crystals of the desired given form of 4a / L-proline were added to the prepared solution (or to the liquid phase filtered from the saturated solution) so that undissolved solids were present. Each mixture was then agitated (typically by stirring) in a sealed vial for a specified time at a specified temperature. The solid was isolated by vacuum filtration and analyzed.

[0099] II. X-ray powder diffraction (XRPD) peak identification Throughout this disclosure, X-ray diffraction patterns and tables accompany peak listings. Peaks in the range up to about 30°2θ were selected. A rounding algorithm was used to round each peak to the nearest 0.01°2θ. Peak positions along the x-axis (°2θ) in both figures and tables were determined using proprietary software (TRIADS™ v2.0) and rounded to two significant decimal places. Peak position variation is based on the variability in powder X-ray diffraction (United States Pharmacopoeia, USP 38-NF 33, including S2). <941> The d-spacing is given within ±0.2° 2θ based on recommendations outlined in the USP discussion on 12 / 1 / 2015. The accuracy and precision associated with any particular measurement method disclosed herein has not been determined. Furthermore, third-party measurements on independently prepared samples on different instruments may result in variations greater than ±0.2° 2θ. The wavelength used to calculate the d-spacing is Cu-K at 1.5405929 Å. α1 (Holzer, G.; Fritsch, M.; Deutsch, M.; Hartwig, J.; Forster, E. Phys. Rev. 1997, A56 (6), 4554-4568). The variability associated with the d-spacing estimates was calculated from the USP recommendations for each d-spacing and is provided in the respective data tables.

[0100] In accordance with USP guidelines, variable hydrates and solvates may exhibit peak dispersions greater than 0.2° 2θ, and therefore a peak dispersion of 0.2° 2θ is not applicable to these substances.

[0101] For samples with only one XRPD pattern and no other, and no other means of assessing whether the sample provides a good approximation of the powder average, the peak table contains data identified only as "prominent peaks." These peaks are a subset of the entire observed peak list. Prominent peaks are selected from the observed peaks by identifying low-angle peaks with strong intensity, preferably non-overlapping.

[0102] When multiple diffraction patterns are available, particle statistics (PS) and / or preferred orientation (PO) can be determined. Reproducibility between XRPD patterns from multiple samples analyzed on a single diffractometer indicates adequate particle statistics. Consistency in relative intensities between XRPD patterns from multiple diffractometers indicates good orientation statistics. Alternatively, when available, observed XRPD patterns can be compared to calculated XRPD patterns based on a single crystal structure. Two-dimensional scattering patterns using an area detector can also be used to assess PS / PO. If it is determined that the contributions of both PS and PO are negligible, the XRPD pattern is representative of the powder average intensity of the sample, and prominent peaks can be identified as "representative peaks." Generally, the more data collected to determine representative peaks, the more confident the classification of those peaks can be.

[0103] "Diagram peaks," to the extent they exist, are a subset of the characteristic peaks used to distinguish one crystalline polymorph from another (polymorphs, which are crystalline forms with the same chemical composition). Characteristic peaks are determined by assessing which characteristic peaks, if any, are present in one crystalline polymorph of a compound within ±0.2° 2θ relative to all other known crystalline polymorphs of that compound. Not all crystalline polymorphs of a compound necessarily have at least one characteristic peak.

[0104] equipment operation Differential scanning calorimetry (DSC): DSC was performed using a TA Instruments Q2000 Differential Scanning Calorimeter. Temperature calibration was performed using NIST-traceable indium metal. The sample was placed in an aluminum Tzero crimped DSC pan, covered with a lid, and the weight was accurately recorded. A weighed aluminum pan, designated as the sample pan, was placed on the reference side of the cell.

[0105] Dynamic Vapor Sorption (DVS):Dynamic water vapor sorption data were collected on a VTI SGA-100 water vapor sorption analyzer. NaCl and polyvinylpyrrolidone (PVP) were used as calibration standards. Samples were not dried prior to analysis. Sorption and desorption data were collected over the range of 5% to 95% RH in 10% RH increments under a nitrogen purge. The equilibrium criteria used for analysis were less than 0.0100% weight change in 5 minutes, with a maximum equilibration time of 3 hours. Data were not corrected for the initial moisture content of the samples.

[0106] EasyMax™ Reactor: Crystallization experiments were carried out using a Mettler Toledo EasyMax™ 102 with a Julabo F26 cooling / circulating device. Crystallizations were carried out in 20 mL glass tubes (capped) with magnetic stirring. The temperature was regulated using the jacket temperature (Tj) setting.

[0107] Elemental analysis: Elemental analyses were performed by Galbraith Laboratories, Knoxville, Tennessee.

[0108] Infrared spectroscopy: IR spectra were acquired on a Nicolet 6700 Fourier transform infrared (FT-IR) spectrophotometer (Thermo Nicolet) equipped with an Ever-Glo mid / long wavelength IR source, a potassium bromide (KBr) beam splitter, and a deuterated triglycine sulfate (DTGS) detector. Wavelength calibration was performed using NIST SRM 1921b (polystyrene). An attenuated total reflectance (ATR) accessory (Thunderdome™, Thermo Spectra-Tech) with a germanium (Ge) crystal was used for data collection. Each spectrum was captured at a 4 cm -1 The graph represents 256 co-added scans collected at a spectral resolution of 100 kHz. A background data set was acquired using a pure Ge crystal. The Log 1 / R (R = reflectance) spectrum was obtained by taking the ratio of these two data sets to each other.

[0109] Raman spectroscopy: Raman spectra were acquired with an FT-Raman module connected to a Nexus 670 FT-IR spectrophotometer (Thermo Nicolet) equipped with an indium gallium arsenide (InGaAs) detector. Wavelength calibration was performed using sulfur and cyclohexane. Each sample was prepared for analysis by placing it in a pellet holder. Approximately 0.514 W of Nd:YVO4 laser power (excitation wavelength of 1064 nm) was used to irradiate the sample. Each spectrum was acquired using a 4 cm -1 represents 256 co-weighted scans collected at a spectral resolution of .

[0110] Single Crystal X-ray Diffraction (SCXRD): The single crystal structures of 4a / L-proline forms B and C were determined at the Crystallography Laboratory at Purdue University.

[0111] Thermogravimetric analysis (TGA): TG analyses were performed using a TA Instruments 2050 Thermogravimetric Analyzer or a Discovery Thermogravimetric Analyzer. Temperature calibration was performed using nickel and Alumel™. Each sample was placed in an aluminum or platinum pan and inserted into the TG furnace. The furnace was heated from ambient temperature to 350°C at a heating rate of 10°C / min under a nitrogen purge.

[0112] Optical microscopy: Samples were observed under a Wolfe optical microscope with crossed polarizers and a 2x or 4x objective, or under a Leica stereomicroscope equipped with a first order red compensator and crossed polarizers and a 0.8x to 10x objective.

[0113] solution 1 H NMR spectroscopy: solution 1 H NMR spectra were obtained from Varian UNITYAcquisition was performed by Spectral Data Services, Champaign, IL, at 25°C using an INOVA-400 spectrometer. Samples were dissolved in DMSO-d6. The residual peak from incompletely deuterated DMSO is at approximately 2.5 ppm, and the relatively broad peak at approximately 3.3 ppm is due to water.

[0114] X-ray powder diffraction (XRPD) PANalytical X'PERT Pro MPD Diffractometer - Transmission Geometry (most samples): XRPD patterns were collected on a PANalytical X'Pert PRO MPD diffractometer using an incident beam of Cu radiation generated using an Optix long, fine-focus source. An elliptically graded multilayer mirror was used to focus the Cu Kα X-rays through the specimen onto the detector. Prior to analysis, a silicon specimen (NIST SRM 640e) was analyzed to verify that the observed Si 111 peak position matched the NIST-certified position. Sample specimens were sandwiched between 3 μm-thick membranes and analyzed in transmission geometry. A beam stop, a short antiscatter extension, and an antiscatter knife edge were used to minimize atmospheric background. Soller slits for the incident and diffracted beams were used to minimize divergence from the axial divergence. Diffraction patterns were collected using a scanning position-sensitive detector (X'Celerator) positioned 240 mm from the specimen and Data Collector software version 2.2b.

[0115] PANalytical X'PERT Pro MPD Diffractometer - Reflection Geometry (Limited Sample Volume):XRPD patterns were collected on a PANalytical X'Pert PRO MPD diffractometer using an incident beam of Cu Kα radiation generated using a long fine-focus source and a nickel filter. The diffractometer was configured using a symmetric Bragg-Brentano geometry. Prior to analysis, a silicon specimen (NIST SRM 640e) was analyzed to verify that the observed Si 111 peak position matched the NIST-certified position. Sample preparations were prepared as thin circular layers centered on silicon zero-background substrates. Antiscatter slits (SS) were used to minimize background caused by the atmosphere. Soller slits for the incident and diffracted beams were used to minimize spread from axial divergence. Diffraction patterns were collected using a scanning position-sensitive detector (X'Celerator) positioned 240 mm from the sample and Data Collector software version 2.2b.

[0116] HPLC method The parameters and conditions for the HPLC measurements described herein are presented in the table below. The reported HPLC purities of 4a and 4b do not take into account the peak for proline. TIFF2026012851000003.tif127142

[0117] Example 1: Synthesis of 4a / L-proline cocrystal (substance A) TIFF2026012851000004.tif35159

[0118] Compound 4a was obtained by chromatographic separation (HPLC: 4a 96.4% and 4b 1.2%, 500 mg) and mixed with L-proline (128 mg, 1 eq.) in EtOH (4 mL). The mixture was heated to reflux for 15 min. The hot solution was filtered through a cotton plug. The resulting clear filtrate was slowly cooled and kept at room temperature for 16 h. The precipitated solid was collected by filtration and air-dried at room temperature to give 4a / L-proline cocrystal (456 mg, 73% yield), designated as substance A, as a white solid. MP 203-205 °C. 1H NMR showed the ratio of 4a to L-proline as 1:1.1.

[0119] Substance A is a 1:1 4a / L-proline cocrystal and is likely an isostructural solvate. Substance A contains a minor L-proline component that was successfully indexed based on the XRPD pattern (Figure 23) (Table A1). XRPD indexing is typically successful for samples consisting primarily or exclusively of a single crystalline phase. However, an indexing solution was obtained for this mixture with the understanding that minor peaks / shoulders present in the XRPD pattern at 8.7°, 15.0°, and 18.0° 2θ were not consistent with the indexing solution and were likely due to L-proline. Selected unit cell parameters obtained from the indexing solution are presented in Table A2.

[0120] (Table A1) Peaks observed for 4a / L-proline substance A TIFF2026012851000005.tif49128TIFF2026012851000006.tif203101

[0121] Table A2: Unit cell parameters for 4a / L-proline substance A TIFF2026012851000007.tif56128

[0122] The unit cell volume is large enough to accommodate a solvated 1:1 4a / L-proline cocrystal. The free volume (or the unit cell volume remaining after the cocrystal is occupied) could conceivably be water and / or any of the solvents from which Substance A is generated, including EtOH, IPA, and THF.

[0123] Material A described above was further characterized by DSC, TGA, and DVS. An overlay of the DSC and TGA thermograms for this material is shown in Figure 24. The TGA thermogram exhibits two distinct weight loss steps: a first step (7 wt%) occurring between approximately 100 and 160 °C and a second step (21 wt%) occurring between 160 and 230 °C. A broad endotherm was observed by DSC with a peak maximum at 145 °C, which corresponds to the first TGA weight loss step. The relatively high temperature at which these events occur likely indicates loss of bound solvent / water as well as the gradual nature of the weight loss. The overlapping endothermic events between approximately 170 and approximately 240 °C by DSC correspond to the second weight loss step in the TGA thermogram, which likely corresponds to the simultaneous melting of the cocrystal and volatilization of the L-proline component. The sharp drop in the TGA thermogram above approximately 250 °C likely corresponds to decomposition.

[0124] The DVS isotherm for material A described above is presented in Figure 25. Because the material is characterized as a mixture with unreacted L-proline, it is unclear what effect, if any, excess L-proline had on the water vapor sorption behavior. The material exhibited significant hygroscopicity at or above 85% RH, accounting for approximately 6 wt% water vapor between 85% and 95% RH. The kinetic equilibrium for water vapor sorption timed out between 85% and 95% RH, indicating that the cocrystal could have taken up more water than measured if a longer equilibration time had been allowed. A relatively constant weight loss was observed upon desorption between 95% and 5% RH. The weight lost during desorption (approximately 8 wt%) was significantly greater than the weight gained during sorption, indicating that the material was likely solvated / hydrated at the start of the analysis. Analysis of the material after DVS by XRPD showed a decrease in crystallinity, although the solid form remained intact. The presence or absence of excess L-proline in the samples after DVS could not be confirmed due to disturbances in the XRPD patterns.

[0125] Example 2: Purification of 4a by co-crystallization with L-proline 500 mg of a mixture of 4 consisting of compounds 4a and 4b (HPLC: 92.0% 4a and 7.1% 4b) and L-proline (128 mg, 1 eq) were refluxed in ethanol (4 mL) for 15 minutes. Seed crystals of the co-crystal obtained in Example 1 were added to the mixture, and the mixture was allowed to cool and then maintained at room temperature for 18 hours. The white solid was filtered off while the mother liquor was used to transfer the remaining white solid from the reaction flask. The collected amount of 4a / L-proline co-crystal ( 1 The product (ratio 1:1 as determined by H-NMR, 497 mg, 79% yield) consisted of 98.0% 4a and 1.49% 4b (analyzed by HPLC).

[0126] Example 3: Co-crystallization screen Amorphous 4a was utilized in approximately 50 cocrystal screen experiments with 26 coformers other than l-proline and d-proline, as summarized in Table 1 below. A variety of crystallization techniques favoring cocrystal formation were used, including solvent-assisted grinding and manual milling, cooling, evaporation, slurrying, rapid precipitation, and reaction crystallization, in which a solution containing a high molar excess of one component is combined with another to promote a reaction equilibrium favoring cocrystal formation. A variety of coformers with functional groups capable of forming hydrogen bonds were utilized, including carboxylic acids, amino acids, sugars, amides, amines, and polyfunctional aromatic compounds. Although coformers were explored in this screen under a variety of conditions, 4a did not form any confirmed cocrystals with these common coformers.

[0127] (Table 1) TIFF2026012851000008.tif214170TIFF2026012851000009.tif202170TIFF2026012851000010.tif227170TIFF2026012851 000011.tif210170TIFF2026012851000012.tif226170TIFF2026012851000013.tif211170TIFF2026012851000014.tif94170 a,bVarious batches of seeds of an uncharacterized crystalline material containing 4a and pyroglutamic acid

[0128] Example 4: Preparation and characterization of 4a / L-proline form B Equimolar amounts of amorphous 4a and L-proline (1:1) were mixed in methanol and heated to approximately 68° C. The resulting solution was allowed to slowly cool to room temperature, at which point a white suspension formed. The suspension was then stirred at room temperature for 3 days, after which Form B was collected by filtration as a white solid and dried.

[0129] Alternatively, amorphous 4a and L-proline were combined in a 1:2 molar ratio in ethanol and heated to approximately 82°C to give a white suspension. The suspension was held at 82°C for approximately 5 minutes, slowly cooled to room temperature, and then stirred at room temperature for 3 days. Form B was collected by filtration as a white solid and dried.

[0130] Form B is an anhydrous 1:2 4a / L-proline cocrystal. Form B was characterized by XRPD (with indexing), DSC, TGA, DVS, Raman spectroscopy, IR spectroscopy, proton NMR, HPLC, and elemental analysis.

[0131] The XRPD pattern for Form B was successfully indexed (Table 2), which indicated that Form B consists primarily or exclusively of a single crystalline phase (Figure 1). The unit cell volume obtained from the indexing solution is consistent with an anhydrous 1:2 4a / L-proline cocrystal. The unit cell parameters are presented in Table 3 below.

[0132] Table 2: Peaks observed for form B of 4a / L-proline TIFF2026012851000015.tif224101

[0133] Table 3. Unit cell parameters for form B of 4a / L-proline TIFF2026012851000016.tif56128

[0134] A sample of Form B isolated from the MeOH slurry was characterized by proton NMR and HPLC. The proton NMR data indicates a 1:2 stoichiometry of 4a / L-proline with no residual solvent detected. The purity of 4a in the sample as determined by HPLC was 99.7%.

[0135] DSC and TGA thermograms for Form B are presented in Figures 2 and 3, respectively. Because different samples were analyzed for each technique, the data were plotted separately. The sample analyzed by TGA was isolated from the MeOH slurry, while the DSC sample was derived from a cocrystal formation experiment in MeOH. Virtually no weight loss was observed by TGA between ambient temperature and 190 °C, consistent with an anhydrous / unsolvated material. Similarly, DSC is consistent with this, showing no significant thermal events until the onset of an endothermic event at 208 °C, accompanied by a strong overlapping exothermic event. Of note, the sample was observed to be removed from the pan after this analysis, likely contributing to the magnitude of the exotherm. These events likely correspond to melting / dissociation of the cocrystal. Similarly, the sharp decrease in the TGA thermogram above 190 °C is likely due to the volatilization of some of the L-proline component of the cocrystal, which likely subsequently decomposes.

[0136] To further confirm the co-crystal stoichiometry, Form B was analyzed by elemental analysis for C, H, N, F, and O (Table 4). Comparison of the experimental and theoretical percent compositions for the 1:1 and 1:2 co-crystals indicates that the sample more closely matches the 1:2 co-crystal. This result is consistent with other characterization data.

[0137] Table 4: Elemental analysis of 4a / L-proline form B TIFF2026012851000017.tif48141

[0138] The dynamic water vapor sorption (DVS) isotherm for Form B is shown in Figure 4. A weight gain of 2.3 wt% was observed between 5% and 95% RH, with the majority of sorption occurring above 50% RH. All of this weight was lost during desorption, which exhibited little hysteresis. The kinetic equilibrium for water vapor sorption timed out during the sorption phase between 85% and 95% RH, indicating that the cocrystal could have taken up more water than measured if a longer equilibration time had been allowed. Analysis of the material after DVS by XRPD showed no observable change in morphology.

[0139] IR and Raman spectra were obtained for Form B and are presented in Figures 5 and 6, respectively.

[0140] Single-crystal X-ray structure determination of form B of active 5:4a / L-proline Data collection Form B of 4α / L-proline (C 34 H 48 FN3O 10 [C 24 H 30 A colorless platelet of FNO6, 2(C5H9NO2)] was mounted on a nylon loop in a random orientation. Preliminary testing and data collection were performed using Cu Kα radiation (λ = 1.54178 Å) on a Rigaku Rapid II diffractometer equipped with confocal optics. Refinement was performed using SHELX2013. (Sheldrick, GM Acta Cryst., 2008, A64, 112)

[0141] The cell constants and orientation matrix for data collection were obtained from least-squares refinement using set angles for 21,646 reflections in the range 4° < θ < 68°. The refined mosaicity from DENZO / SCALEPACK was 0.44°, indicating good crystal quality (Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307). The space group was determined using the program XPREP (Bruker, XPREP in SHELXTL v. 6.12., Bruker AXS Inc., Madison, WI, USA, 2002). From the systematic occurrence of the following conditions: 0k0 k = 2n, and subsequent least-squares refinement, the space group was determined to be P21 (no. 4).

[0142] Data were collected at room temperature up to a maximum diffraction angle (2θ) of 135.73°.

[0143] Data organization Frames were integrated with HKL3000 (Otwinowski (1997)). A total of 21,646 reflections were collected, of which 5,936 were unique. Lorentzian and polarization corrections were applied to the data. The linear absorption coefficient was 0.833 mm for Cu Kα radiation. -1 An empirical absorption correction using SCALEPACK (Otwinowski (1997)) was applied. Transmission coefficients ranged from 0.128 to 0.779. A secondary extinction correction was applied (Sheldrick (2008)). The final coefficient, refined by least squares, was 0.0157 (11) (in absolute units). The intensities of equivalent reflections were averaged. The agreement factor for averaging was 4.1% based on intensity.

[0144] Structure solution and refinement The structure was solved by direct methods using SHELXT (Sheldrick (2008)). The remaining atoms were located by subsequent difference Fourier synthesis. Hydrogen atoms were included in the refinement but were restricted depending on the atoms to which they were bonded. Structure is a function: TIFF2026012851000018.tif7128 was refined using full matrix least squares by minimizing the weight w. 2 (F o 2 )+(0.0640P) 2 +(0.5095P)](P=(F o 2 +2F c 2 ) / 3).

[0145] The scattering factors were taken from "International Tables for Crystallography" (International Tables for Crystallography, Vol. C, Kluwer Academic Publishers: Dordrecht, The Netherlands, 1992, Tables 4.2.6.8 and 6.1.1.4). Of the 5936 reflections used in the refinement, F o 2 >2σ(F o 2 ) reflections were used in the calculation of the fit residual R. A total of 5601 reflections were used in the calculation. The final cycle of refinement included 490 variable parameters, and The unweighted and weighted agreement factors converged to TIFF2026012851000019.tif20128 (maximum parameter shift was <0.01 times its estimated standard deviation).

[0146] The standard deviation of the unit weight observations (goodness of fit) was 1.062. The highest peak in the final difference Fourier transform was 0.187 e / Å. 3 The smallest negative peak had a height of -0.193 e / Å. 3 It had a height of

[0147] Calculated X-ray powder diffraction (XRPD) patterns Calculated XRPD patterns were generated for Cu radiation using Mercury (Macrae, C. F. Edgington, P. R. McCabe, P. Pidcock, E. Shields, G. P. Taylor, R. Towler M. and van de Streek, J.; J. Appl. Cryst., 2006, 39, 453-457) and atomic coordinates, space group, and unit cell from the single crystal structure.

[0148] Atomic displacement ellipsoid diagrams and filling diagrams Atomic displacement ellipsoid diagrams were generated using Mercury (Macrae (2006)). Atoms are represented by 50% probability anisotropic thermal ellipsoids. Packing diagrams and additional figures were generated using Mercury (Macrae (2006)). Hydrogen bonds are represented by dashed lines. Chiral center identification was performed using PLATON (Spek, AL PLATON. Molecular Graphics Program. Utrecht University, Utrecht, The Netherlands, 2008. Spek, AL, J. Appl. Cryst. 2003, 36, 7). The absolute configuration is assessed using the specifications of the rules of molecular chirality (Cahn, RS; Ingold, C; Prelog, V. Angew. Chem. Intern. Ed. Eng., 1966, 5, 385 and Prelog, V., Helmchen, G. Angew. Chem. Intern. Ed. Eng., 1982, 21, 567).

[0149] result The monoclinic lattice parameters and calculated volume are: a = 11.1270(4) Å, b = 10.1566(4) Å, c = 16.0790(6) Å, β = 109.309(2)° (α = γ = 90°), V = 1714.91(11) Å 3 The formula weight of the asymmetric unit in the crystal structure of Form B is 677.75 g mol with Z=2.-1 and the calculated density is 1.313 g cm -3 The space group was determined to be P21 (no. 4). A summary of the crystallographic data and crystallographic data collection parameters is provided in Table 5.

[0150] Table 5. Crystallographic data and data collection parameters for Form B of 4a / L-proline TIFF2026012851000020.tif213119 a Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307. b Flack, HD Acta Cryst., 1983 A39, 876. c Hooft, RWW, Straver, LH, and Spek, AL J Appl Cryst, 2008, 41, 96-103.

[0151] The space group and unit cell parameters are consistent with those obtained from the XRPD analysis of Form B (see Table 3 above).

[0152] The quality of the resulting structure is high, as indicated by the fit residual R of 0.0422 (4.22%). R values ​​in the range of 0.02 to 0.06 are quoted for the most reliably determined structures. (Glusker, Jenny Pickworth; Trueblood, Kenneth N. Crystal Structure Analysis: A Primer, 2 nd ed.; Oxford University press: New York, 1985; p.87)

[0153] An atomic displacement ellipsoid diagram for Form B is shown in Figure 7. The molecules observed in the asymmetric unit of the single crystal structure are consistent with the proposed molecular structure of 4a. The asymmetric unit shown in Figure 7 contains one 4a molecule and two L-proline molecules, which is consistent with a 1:2 4a:L-proline stoichiometry. Two protons representing the zwitterion were positioned separately on both proline nitrogen atoms and refined. Both L-proline molecules are disordered over two positions, which refine to 82 / 18% and 71 / 29% occupancy.

[0154] The absolute configuration can be determined by analyzing anomalous X-ray scattering from the crystal. The refinement parameter x, known as the Flack parameter (Flack, H.D.; Bernardinelli, G., Acta Cryst. 1999, A55, 908; Flack, H.D.; Bernardinelli, G., J. Appl. Cryst. 2000, 33, 1143; Flack, H.D. Acta Cryst. 1983, A39, 876; Parsons, S., Flack, H.D., Wagner, T., Acta Cryst. 2013, B69, 249-259), encodes the relative abundance of the two components in the reverse twin. The structure contains the fraction 1-x of the refined model and its reciprocal x. When low standard uncertainty is obtained, the Flack parameter should be close to 0 if the solved structure is correct and close to 1 if the reverse model is correct. The measured Flack parameter for the Form B structure shown in Figure 7 is -0.02, with a standard uncertainty of 0.11. This indicates a weak inversion-distinguishing power, and therefore the Flack parameter may not be interpreted correctly. The error in the standard uncertainty prevents an assignment based on the Flack factor alone.

[0155] Refinement of the Flack parameter (x) does not provide a quantitative statement about the assignment of absolute structure. However, a Bayesian statistical approach to the Bijvoet difference can provide a range of probabilities for different hypotheses about the absolute structure (Hooft, RWW, Straver, LH, and Spek, ALJ Appl. Cryst., 2008, 41, 96-103; Bijvoet, JM; Peerdeman, AF; van Bommel, AJ Nature 1951, 168, 271). This analysis provides the Flack equivalent (Hooft) parameter, as well as the probability that the absolute structure is correct, incorrect, or a racemic twin. For the current data set, the Flack equivalent (Hooft) parameter was determined to be -0.02(5). The probability that the structure is correct is 1.000, and the probability that the structure is incorrect is 0.9 × 10. -91 and the probability that the material is a racemic twin is 0.2 × 10 -24 Therefore, the absolute configuration of the model in Figure 7 is correct. This structure contains four chiral centers on 4a located at C7, C9, C11, and C12 (Figure 7), which are bonded in the R, R, S, and R configurations, respectively, and one chiral center on each proline molecule at C26 and C31, both of which are bonded in the S configuration.

[0156] Figure 8 shows the calculated XRPD pattern of Form B generated from the single crystal structure. The experimental XRPD pattern of Form B, previously indexed (Example 4), is shown above and matches the calculated XRPD pattern (Figure 9).

[0157] Example 6: Preparation and characterization of Form C of 4a / L-proline Equimolar amounts of amorphous 4a and L-proline (1:1) were slurried and then stirred in acetone at room temperature for 3 days. The slurry was filtered to collect Form C as a white solid.

[0158] Form C was also prepared by dissolving equimolar amounts of amorphous 4a and L-proline in EtOH, followed by precipitation of Form C by vapor diffusion (VD) with acetone.

[0159] Although acetone is not involved in hydrogen bonding, the data indicate that Form C consists of a 1:1 co-crystal with one mole of acetone present in the crystal lattice. The single crystal data provides confirmation of the chemical and solid state composition.

[0160] The XRPD pattern for Form C (Figure 10) was successfully indexed. The observed peaks are shown in Table 6.

[0161] Table 6: Peaks observed for Form C of 4a / L-proline TIFF2026012851000021.tif19598

[0162] The unit cell parameters from XRPD indexing are presented in Table 7 below.

[0163] Table 7. Unit cell parameters for form C of 4a / L-proline TIFF2026012851000022.tif56128

[0164] The TGA thermogram for Form C exhibits a gradual weight loss, consistent with the finding that the material consists of an acetone solvate (Figure 11). The acetone appears to volatilize in two separate steps. In the first step, a weight loss of 2.6% is observed between approximately 60 and 150 °C. Assuming the volatile is acetone, 2.6 wt% corresponds to 0.26 moles (or approximately 1 / 4 of the total acetone per single crystal structure). If acetone is assumed to be the only volatile, the second weight loss step between 150 and 220 °C corresponds to a weight loss of 6.7%, or 0.70 moles.

[0165] Single-crystal X-ray structure determination of form C of active 7:4a / L-proline C with dimensions of approximately 0.19 x 0.18 x 0.10 mm32 H 45 FN2O9[C 24 H 30 Colorless platelets of [FNO6, C5H9NO2, C3H6O] were mounted on the fiber in a random orientation. Preliminary testing and data collection were performed on a Rigaku Rapid II diffractometer equipped with confocal optics using Cu Kα radiation (λ = 1.54178 Å). Refinement was performed using SHELX2013 (Sheldrick (2008)).

[0166] The cell constants and orientation matrix for data collection were obtained from least-squares refinement using set angles for 12,615 reflections in the range 4° < θ < 59°. The refined mosaicity from DENZO / SCALEPACK was 0.25°, indicating good crystal quality (Otwinowski (1997)). The space group was determined by the program XPREP (Bruker (2002)). From the systematic occurrence of the following conditions: 0k0 k = 2n, and from subsequent least-squares refinement, the space group was determined to be P21 (no. 4).

[0167] Data were collected at room temperature up to a maximum diffraction angle (2θ) of 117.84°.

[0168] Frames were integrated with an HKL3000 (Bruker (2002)). A total of 12615 reflections were collected, of which 4368 were unique. Lorentzian and polarization corrections were applied to the data. The linear absorption coefficient was 0.788 mm for Cu Kα radiation. -1 An empirical absorption correction was applied using SCALEPACK (Bruker (2002)). Transmission coefficients ranged from 0.060 to 0.924. A quadratic extinction correction was applied (Sheldrick (2008)). The final coefficient, refined by least squares, was 0.0049 (7) (in absolute units). The intensities of equivalent reflections were averaged. The agreement factor for averaging was 4.8% based on the intensity.

[0169] Structure solution and refinement were performed in a manner similar to that described in Example 5 above. Of the 4368 reflections used in the refinement, Fo 2 >2σ(F o 2 ) reflections were used in the calculation of the fit residual R. A total of 3518 reflections were used in the calculation. The final cycle of refinement included 417 variable parameters, and The unweighted and weighted agreement factors converged to TIFF2026012851000023.tif20128 (maximum parameter shift was <0.01 times its estimated standard deviation).

[0170] The standard deviation of the unit weight observations (goodness of fit) was 1.078. The highest peak in the final difference Fourier transform was 0.271 e / Å. 3 The smallest negative peak had a height of -0.167 e / Å. 3 It had a height of

[0171] Calculated XRPD patterns and atomic displacement ellipsoid diagrams were generated according to the procedures in Example 5.

[0172] The monoclinic lattice parameters and calculated volume are: a = 10.9962(6) Å, b = 10.2721(6) Å, c = 15.3197(9) Å, β = 107.937(4)° (α = γ = 90°), V = 1646.32(17) Å 3 The formula weight of the asymmetric unit in the crystal structure of Form C is 620.70 g mol with Z=2. -1 and the calculated density is 1.252 g cm -3 The space group was determined to be P21 (no. 4). A summary of the crystallographic data and crystallographic data collection parameters is provided in Table 8. The space group and unit cell parameters are consistent with those previously obtained by XRPD indexing (Example 6).

[0173] Table 8. Crystallographic data and data collection parameters for Form C of 4a / L-proline TIFF2026012851000024.tif212121 a Otwinowski (1997). b Flack (1983). c Hooft (2008).

[0174] An atomic displacement ellipsoid diagram for Form C is shown in Figure 12. The molecules observed in the asymmetric unit of the single crystal structure are consistent with the proposed molecular structure of 4a / L-proline. The asymmetric unit shown in Figure 12 contains one 4a molecule, one L-proline molecule, and one acetone molecule. The two protons representing the zwitterion were refined by separately positioning them on the proline nitrogen atom.

[0175] Figure 13 shows the calculated XRPD pattern of Form C generated from a single crystal structure. As described above, the experimental XRPD pattern of the bulk material from which the crystals were obtained is shown in Figure 10. All peaks in the experimental pattern are present in the calculated XRPD pattern, indicating that the bulk material is likely to be single phase. Differences in intensity between the calculated and experimental powder diffraction patterns are often due to preferred orientation. Preferred orientation is when the crystals This preferred orientation of a sample can have a large effect on peak intensities, but not peak positions, in experimental powder diffraction patterns.

[0176] For the current data set, the Flack equivalent (Hooft) parameter was determined as described in Example 5 and was determined to be 0.09(9). The probability that the structure is correct is 1.000, and the probability that the structure is incorrect is 0.4×10 -21 and the probability that the material is a racemic twin is 0.4 × 10 -4 is.

[0177] Therefore, the absolute configuration of the model in Figure 12 is likely correct. This structure contains four chiral centers on 4a located at C7, C9, C11, and C12 (Figure 12) that are bonded in the R,R,S,R configuration, respectively, and one chiral center located on the proline at C26 that is bonded in the S configuration.

[0178] Example 8: Preparation and characterization of 4a / L-proline form D Equimolar amounts of amorphous 4a and L-proline were combined in acetonitrile to give a thin suspension, which was then heated to about 85° C. The suspension was cooled to about 71° C., seeded with a small amount of the material prepared in Example 1, and held at about 71° C. for about 15 minutes. The suspension was then allowed to cool slowly to room temperature and then stirred for 3 days. The resulting white suspension was filtered to give Form D, which was then dried.

[0179] Form D was also prepared by combining equimolar amounts of amorphous 4a and L-proline in EtOH, heating to about 82°C and holding at that temperature for about 15 minutes, cooling to about 76°C, seeding with a small amount of the material prepared in Example 1, allowing to cool slowly to temperature, and stirring for 3 days. The resulting off-white slurry was filtered, and the collected amount of Form D was dried under vacuum at about 48°C.

[0180] Form D consists of an anhydrous / unsolvated 1:2 4a / L-proline cocrystal. The XRPD pattern for Form D was successfully indexed, which indicated that the material consists primarily or exclusively of a single crystalline phase (Figure 14, Table 9). The unit cell volume obtained from the indexing solution (Table 10) is consistent with an anhydrous 1:2 4a / L-proline cocrystal.

[0181] Table 9: Peaks observed for form D of 4a / L-proline TIFF2026012851000025.tif227103

[0182] Table 10: Unit cell parameters for form D of 4a / L-proline TIFF2026012851000026.tif56128

[0183] Form D was characterized by thermal techniques, proton NMR, HPLC, and DVS. An overlay of DSC and TGA thermograms for dried Form D is shown in Figure 15. The slight weight loss by TGA up to 185°C and the broad desolvation endotherm by DSC are consistent with an anhydrous / unsolvated material. A small broad endotherm was observed at 132°C (peak maximum). The sharp endotherm occurring at 211°C corresponds to a sudden, gradual weight loss of 26 wt% in the TGA thermogram, likely corresponding to the simultaneous melting of the cocrystal and volatilization of L-proline.

[0184] Proton NMR data indicated a 1:2 stoichiometry of 4a / L-proline with 0.2 wt% residual EtOH detected. The purity of 4a in the sample by HPLC was 99.6%.

[0185] The DVS isotherm for Form D is shown in Figure 16. The material exhibited significant hygroscopicity, with a weight gain of 4.7 wt% observed between 5% and 95% RH, particularly above 55% RH. All of this weight was lost during desorption, which exhibited little hysteresis. Notably, the kinetic equilibrium for water vapor sorption timed out during the sorption stage between 85% and 95% RH, indicating that the cocrystal could have taken up more water than was measured if a longer equilibration time had been allowed.

[0186] Example 9: Preparation and characterization of form G of 4a / L-proline Amorphous 4a, pyrazine, and L-proline were combined in a molar ratio of 1:20:1 by first dissolving pyrazine in methyl ethyl ketone and MeOH (90:10, v / v), respectively. The pyrazine solution was added to the mixture of 4a and L-proline and stirred at room temperature for 2 days to give an opaque white suspension. Form G was isolated by vacuum filtration of the suspension.

[0187] Form G consists of a 4a / L-proline co-crystal in MEK solvate. Form G exhibited a distinctive crystalline pattern by XRPD (Figure 17) that was indexed (Table 11).

[0188] Table 11. Peaks observed for form G of 4a / L-proline TIFF2026012851000027.tif204103

[0189] The unit cell volumes obtained from indexing the XRPD pattern (Table 12) are consistent with a 1:1 4a / L-proline cocrystal with up to 1 mole of MEK or pyrazine present (MEK and pyrazine molecules are comparable in volume and cannot be distinguished by XRPD indexing). The unit cell parameters (Table 12) also indicate that Form G is isostructural with Forms B and C. Form G was confirmed by proton NMR to contain 4a, L-proline, MEK, and traces of residual pyrazine in a molar ratio of approximately 1:1.2:0.6:0.1.

[0190] Table 12: Unit cell parameters for form G of 4a / L-proline TIFF2026012851000028.tif55128

[0191] Single crystal analysis of Forms B and C showed that these forms are isostructural, with 4a and L-proline forming channels that accommodate an additional L-proline for Form B and acetone for Form C. Both the L-proline and acetone molecules in each channel do not form hydrogen bonds with the molecules that comprise the channel. As noted above, the space group and other unit cell parameters obtained for Form G indicated that it is isostructural with Forms B and C. While other explanations are possible, given what is known about the molecular packing for these forms and the nonstoichiometric equivalents of L-proline and MEK measured by proton NMR for Form G, it is highly likely that the channels in Form G can accommodate both L-proline and MEK in nonstoichiometric (and possibly variable) ratios due to the ease of exchange afforded by the lack of hydrogen bonding within the channels.

[0192] Example 10: Purification of 4a by co-crystal formation A. Purification using L-proline Cocrystallization of 4 according to the above procedure yielded Form B and Form D, thereby very efficiently reducing the amount of β-anomer 4b. A typical batch of amorphous 4 consisted of 93.2% 4a / 6.3% 4b as determined by HPLC. After the cocrystal formation step in a typical experiment, the level of 4b decreased from 6.3% to 2.4% (HPLC). The resulting 4a / L-proline cocrystal was recrystallized by mixing with MeOH (2 volumes), and the mixture was heated to reflux for 3 hours. The mixture was cooled to 0 ± 3 °C over 2.5 hours and then stirred overnight. The resulting solid was collected by filtration. After recrystallization, the amount of 4b was further reduced to 1.2%, and the purity of 4a improved to 98.7% (HPLC).

[0193] B. Purification using D-proline A separate amount of amorphous compound 4 contained 89.3% 4a and 10.1% 4b as determined by HPLC. Compound 4 (300 mg, 0.670 mmol) and D-proline (77.3 mg, 0.671 mmol) in EtOH (2.4 mL) were heated in a 90 °C oil bath. After refluxing for 15 min, the resulting solution was cooled to room temperature in a vial, and the vial was uncapped and kept for 24 h to allow slow evaporation of the EtOH at room temperature. The precipitated solid was collected by filtration and air-dried to give the 4a / D-proline cocrystal. The cocrystal contained 97.6% 4a and 2.1% 4b as determined by HPLC (150 mg as a white solid, 40% yield). 1 H NMR analysis showed a 1:1 molar ratio of 4a and D-proline.

[0194] Example 11: Preparation and characterization of 4a / D-proline cocrystals A mixture of compound 4a (100 mg, 0.223 mmol) and D-proline (25.8 mg, 0.224 mmol) in EtOH (0.8 mL) was heated in an oil bath at 90 °C. After refluxing for 15 min, the solution was cooled to room temperature and kept at room temperature in a capped vial for 24 h. The vial was then uncapped and the EtOH was allowed to slowly evaporate at room temperature. After 24 h, the precipitated solid was collected by filtration and then air-dried to give 4a / D-proline cocrystal as a white solid with 99+% purity (HPLC) (74 mg, 59% yield). 1 H NMR analysis showed that the co-crystal contained a 1 / 1 ratio of compound 4a and D-proline.

[0195] The 4a / D-proline cocrystal was characterized by XRPD, DSC, TGA, and DVS. The XRPD pattern for the 4a / D-proline cocrystal was successfully indexed, indicating that the material consisted primarily or exclusively of a single crystalline phase (Figure 18, Table 13).

[0196] Table 13: Peaks observed for 4a / D-proline TIFF2026012851000029.tif52128TIFF2026012851000030.tif221103TIFF2026012851000031.tif55128

[0197] A DSC / TGA overlay for the material is shown in Figure 19. The TGA thermogram for 4a / D-proline cocrystal exhibited two distinct weight loss steps: the first step (7.0% weight loss) occurring between approximately 100°C and 150-160°C, and the second step (20.0% weight loss) occurring between 150-230°C. A broad endotherm was observed by DSC with a peak maximum at 130°C, consistent with the first TGA weight loss step, likely representing loss of bound solvent / water. An overlapping endotherm was observed above approximately 170°C, likely due to the simultaneous melting / volatilization of the D-proline component of the cocrystal. The sharp drop in the TGA thermogram above approximately 250°C likely corresponds to decomposition.

[0198] The DVS isotherm for the d-proline cocrystal is shown in Figure 20. Upon sorption, the cocrystal grew 26 wt% between 5% and 95% RH, with the majority of the weight gain occurring between 85% and 95% RH. Kinetic equilibrium timed out during this stage, indicating that the cocrystal could have taken up more water than measured if a longer equilibration time had been allowed. Upon desorption, the cocrystal exhibited a relatively constant weight loss between 95% and 5% RH, losing more weight (29 wt%) than it gained during sorption, indicating that the material likely contained solvent / water at the start of the analysis. Of note, it was observed that the post-DVS sample of the d-proline cocrystal stuck to the pan and could not be recovered, indicating partial deliquescence during the experiment.

[0199] Example 12: Comparative administration of 4a / L-proline Substance A and amorphous 4a to mice In this example, the systemic exposure to 4a was evaluated in male C57BL / 6 mice after oral administration of a suspension formulation of substance A prepared in Example 1 compared to a suspension formulation of amorphous 4a.

[0200] Amorphous 4a or substance A was formulated in 0.5% CMC with 5% DMSO (to account for the presence of proline by weight) and administered by oral gavage to 35 male C57BL / 6 mice at 1000 mg / kg using a dose volume of 15 mL / kg. Blood samples for plasma isolation were collected 0.25, 0.5, 1, 2, 4, 8, and 24 hours after administration using separate groups of mice (n=5 / group / time point) for each time point, as summarized in Table 14 below. Plasma samples were analyzed for 4a concentration by LC / MS / MS techniques, as described below.

[0201] Table 14. Summary of mouse studies TIFF2026012851000032.tif63156

[0202] Reagents and supplies:All reagents and supplies were of high quality, and where appropriate, LC / MS grade, and were obtained from standard commercial suppliers.

[0203] Plasma sample preparation: Protein precipitation was used to extract 4a from K3EDTA-fortified plasma samples. 20 μL of a 2.5-ng / mL D3-4a (internal standard, IS) solution prepared in acetonitrile / HO (1:1) was added to wells of a 96-well polypropylene microplate, followed by 20 μL of plasma sample. The plate was sealed with sealing tape (Phenomenex, AH0-7362) and gently mixed on a vortex mixer for 1 minute. The solution was pipetted into wells of a 96-well polypropylene plate (Phenomenex, AH0-7194) containing 500 μL of methanol. The plate was sealed and vortexed for 5 minutes, then centrifuged at 3000 × g for 3 minutes at room temperature. A 300 μL aliquot of the supernatant was transferred to a new well containing 300 μL of deionized water. After gentle mixing, the plate was sealed and placed in an LC autosampler maintained at 12°C, and a 10 μL aliquot was injected into the LC / MS / MS system for quantitative analysis of 4a.

[0204] Chromatography and mass spectrometry conditions: Liquid chromatographic separation of 4a was achieved on a reversed-phase analytical column using a mobile phase solution containing HO, acetonitrile, and fumaric acid. Chromatographic analytes were detected on a Waters Xevo TQ-S triple quadrupole mass spectrometer operated in multiple reaction monitoring (MRM) mode. Chromatographic peak areas for quality control samples, calibration standards, and research samples were integrated using MassLynx software V4.1 (Waters Corp.).

[0205] Pharmacokinetic analysis:Pharmacokinetic parameter estimates were obtained from non-compartmental analysis of the mean 4a plasma concentration-time data for each dose group using WinNonlin™ software version 6.3 (Pharsight Corp., Cary, NC). The area under the plasma concentration-time curve (AUC ) from time zero to the time (t) at which the last measurable 4a concentration was measured was calculated. (0-t) ) was determined using the linear log trapezoidal rule. The time of the last measurable concentration was defined as the time after which the concentration of 4a was below the lower limit of quantitation (BLQ) in the majority of animals in each dose group.

[0206] result: The mean (±SD) 4a plasma concentration-time data following a single oral gavage dose of Substance A and amorphous 4a are presented in Table 15 below and displayed in Figure 21. Pharmacokinetic parameter estimates for 4a are presented in Table 16.

[0207] Table 15. Mean (±SD) plasma concentrations of 4a in mice after single oral administration of Substance A and amorphous 4a. TIFF2026012851000033.tif43130 a All samples in the group were BLQ.

[0208] Table 16. Plasma pharmacokinetic parameter estimates after a single oral dose in mice TIFF2026012851000034.tif20136

[0209] Example 13: Comparative administration of 4a / L-proline Substance A and amorphous 4a to monkeys In this example, the systemic exposure of the formulation of Substance A prepared in Example 1 and a formulation of amorphous 4a, respectively, was evaluated in male cynomolgus monkeys after a single oral gavage dose of 30 mg / kg using 0.5% carboxymethylcellulose (CMC) in sterile water as a vehicle (5 mL / kg) or 50 mg loosely packed in capsules (irrespective of body weight), as summarized in Table 17 below.

[0210] Table 17. Summary of monkey studies TIFF2026012851000035.tif72163 a Doses administered regardless of weight b NA, not applicable

[0211] Administration of Substance A took into account the presence of proline on a weight basis. Blood samples for plasma isolation were collected from each monkey at 0.25, 0.5, 1, 2, 4, 8, 12, and 24 hours after administration. Plasma samples were analyzed for 4a concentration by LC / MS / MS techniques, and pharmacokinetic analysis was performed as described in Example 12 above.

[0212] Results: Mean (±SD) plasma concentrations after a single oral gavage dose (30 mg / kg, 5 mL / kg) or capsule administration (50 mg) are presented in Table 18 below and displayed in Figure 22. Pharmacokinetic parameter estimates are presented in Table 19.

[0213] Table 18. Mean (±SD) 4a plasma concentrations in monkeys after oral administration of Substance A and amorphous 4a as suspension or capsules. TIFF2026012851000036.tif53135 a Two of the three values ​​for the intermediate time points (i.e., between the two time points with quantifiable values) were below the limit of quantitation (BLQ) and were included in the average as half the lower limit of quantitation (LLOQ) (i.e., 0.200 ng / mL). b All values ​​for intermediate time points (ie, between two time points with quantifiable values) were the BLQ and were reported as half the LLOQ (ie, 0.200 ng / mL). c One of the three values ​​was the BLQ, which was assigned a value half the LLOQ (ie, 0.200 ng / mL) to calculate the mean and standard deviation. d Since two or three samples were BLQ, the average was reported as BLQ. e All values ​​were BLQ.

[0214] Table 19. Mean (±SD) plasma concentrations in monkeys after administration of 4a by oral gavage of suspension or capsules. TIFF2026012851000037.tif41142 a Data are presented as median values.

[0215] Both capsule formulations (but not the oral suspension formulation) produced plasma concentrations of 4a that were below or near the lower limit of quantitation (i.e., 0.400 ng / mL) at 4 and 8 hours after administration, but showed a secondary peak in exposure at 24 hours with mean (±SD) 4a concentrations of 12.5±11.0 and 6.11±9.23 ng / mL for the Substance A formulation and amorphous 4a formulation, respectively.

Claims

1. Co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide and L-proline (1:2), characterized by a powder X-ray diffractogram containing the following peaks: 14.76, 16.86, 19.00, and 21.05°2θ ± 0.20°2θ as determined with a diffractometer using Cu-Kα radiation at a wavelength of 1.54178 Å.

2. 2. The co-crystal of claim 1, wherein the powder X-ray diffractogram further comprises peaks at 12.14, 17.51, 18.89, and 19.41 degrees 2θ ± 0.20 degrees 2θ.

3. 3. The co-crystal of claim 1 or 2, having an X-ray powder diffractogram substantially as shown in Figure 1.

4. 4. The co-crystal of any one of claims 1 to 3, characterized by a differential scanning calorimetry (DSC) thermogram comprising an exotherm at about 211°C.

5. 5. The co-crystal of claim 4, having a DSC thermogram substantially as shown in Figure 2.

6. Co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide and L-proline (1:2), characterized by a powder X-ray diffractogram containing the following peaks: 9.20, 16.19, 18.45, and 24.51°2θ ± 0.20°2θ as determined with a diffractometer using Cu-Kα radiation at a wavelength of 1.54178 Å.

7. 7. The co-crystal of claim 6, wherein the powder X-ray diffractogram further comprises peaks at 11.83, 17.16, 20.15, and 25.34 degrees 2θ ± 0.2 degrees 2θ.

8. 8. The co-crystal of claim 6 or 7, having an X-ray powder diffractogram substantially as shown in Figure 14.

9. 9. The co-crystal of any one of claims 6 to 8, characterized by a differential scanning calorimetry (DSC) thermogram comprising an endotherm with an onset temperature of about 211.2°C.

10. 10. The co-crystal of claim 9, having a DSC thermogram substantially as shown in Figure 15.

11. Acetone solvate of a co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide and L-proline (1:1:1), characterized by a powder X-ray diffractogram containing the following peaks as determined on a diffractometer using Cu-Kα radiation at a wavelength of 1.54178 Å: 14.64, 17.53, 18.91, and 21.33°2θ ​​± 0.20°2θ.

12. 12. The co-crystal of claim 11, wherein the powder X-ray diffractogram further comprises peaks at 12.10, 15.14, 18.26, and 19.56 degrees 2θ ± 0.2 degrees 2θ.

13. 13. The co-crystal of claim 11 or 12, having an X-ray powder diffractogram substantially as shown in Figure 10.

14. 14. The co-crystal of any one of claims 11 to 13, characterized by a thermogravimetric analysis (TGA) thermogram comprising weight loss steps ending at about 150°C and about 220°C.

15. 15. The co-crystal of claim 14, having a TGA thermogram substantially as shown in Figure 11.

16. A co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide with L-proline, methyl ethyl ketone, and pyrazine in a molar ratio of approximately 1:1.2:0.6:0.1, characterized by a powder X-ray diffractogram containing the following peaks: 10.42, 14.62, 19.28, and 21.14°2θ ± 0.20°2θ as determined by a diffractometer using Cu-Kα radiation at a wavelength of 1.54178 Å.

17. 17. The co-crystal of claim 16, wherein the powder X-ray diffractogram further comprises peaks at 11.85, 14.93, 17.40, and 19.28 degrees 2θ ± 0.2 degrees 2θ.

18. 18. The co-crystal of claim 16 or 17, having an X-ray powder diffractogram substantially as shown in Figure 17.

19. 19. The co-crystal of any one of claims 16-18, characterized by unit cell dimensions as follows: a=10.975 Å, b=10.310 Å, c=15.704 Å, α=90°, β=108.56°, and γ=90°.

20. Cu-K at a wavelength of 1.5405929 Å α1 Co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide and D-proline (1:1), characterized by a powder X-ray diffractogram containing the following peaks as determined by a radiation diffractometer: 11.77, 14.52, 19.54, and 21.23°2θ ± 0.20°2θ.

21. 21. The co-crystal of claim 20, wherein the powder X-ray diffractogram further comprises peaks at 8.45, 13.18, 16.95, and 19.12 degrees 2θ ± 0.2 degrees 2θ.

22. 22. The co-crystal of claim 20 or 21, having an X-ray powder diffractogram substantially as shown in Figure 18.

23. 23. The co-crystal of any one of claims 20 to 22, characterized by a differential scanning calorimetry (DSC) thermogram comprising an endotherm at about 130°C.

24. 24. The co-crystal of claim 23, having a DSC thermogram substantially as shown in Figure 19.

25. Cu-K at a wavelength of 1.5405929 Å α1 A co-crystal of N-(2-(5-(((2R,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3'-fluoro-[1,1'-biphenyl]-2-yl)ethyl)-acetamide and L-proline (1:1), characterized by a powder X-ray diffractogram containing the following peaks as determined by a radiation diffractometer: 8.52, 16.33, 19.50, and 21.22°2θ±0.20°2θ.

26. 26. The co-crystal of claim 25, wherein the powder X-ray diffractogram further comprises peaks at 9.19, 13.22, 14.75, and 17.57 degrees 2θ ± 0.2 degrees 2θ.

27. 27. The co-crystal of claim 25 or 26, having an X-ray powder diffractogram substantially as shown in Figure 23.

28. 28. The co-crystal of any one of claims 25-27, characterized by unit cell dimensions as follows: a=10.126 Å, b=11.021 Å, c=30.259 Å, α=90°, β=90°, and γ=90°.

29. 28. The co-crystal of any one of claims 25 to 27, characterized by a differential scanning calorimetry (DSC) thermogram comprising an endotherm at about 145°C.

30. 30. The co-crystal of claim 29, having a DSC thermogram substantially as shown in Figure 24.

31. 31. A pharmaceutical composition comprising the cocrystal of any one of claims 1 to 30 and a pharmaceutically acceptable solid carrier.

32. 31. A method for inhibiting heat shock protein 90 (Hsp90) in a subject, comprising administering to the subject a therapeutically effective amount of the cocrystal of any one of claims 1 to 30.

33. 31. A method for treating or preventing a neurodegenerative disorder in a subject suffering from the disorder, comprising administering to the subject a therapeutically effective amount of the cocrystal of any one of claims 1-30.

34. 34. The method of claim 33, wherein the neurodegenerative disorder is diabetic peripheral neuropathy.

35. 31. A method for preventing or reducing the likelihood of developing diabetic peripheral neuropathy in a subject, comprising administering to the subject a therapeutically effective amount of the cocrystal of any one of claims 1-30, wherein the subject is afflicted with type 1 or type 2 diabetes.

36. 1. A method for increasing the concentration of N-(2-(5-(((2S,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3′-fluoro-[1,1′-biphenyl]-2-yl)ethyl)-acetamide (4a) relative to N-(2-(5-(((2S,3R,4S,5R)-3,4-dihydroxy-5-methoxy-6,6-dimethyltetra-hydro-2H-pyran-2-yl)oxy)-3′-fluoro-[1,1′-biphenyl]-2-yl)ethyl)-acetamide (4b) in a starting composition comprising 4a and 4b, the method comprising the steps of contacting the starting composition with proline in a solvent and subjecting the starting composition, proline, and solvent to crystallization conditions under which a co-crystal of 4a and proline is produced, co-crystal) exhibits a higher concentration of 4a than in the starting composition comprising 4a and 4b.

37. 37. The method of claim 36, wherein the proline is L-proline.

38. 37. The method of claim 36, wherein the proline is D-proline.

39. 39. The method of any one of claims 36 to 38, further comprising heating the starting composition, proline and solvent.

40. 40. The method of any one of claims 36 to 39, wherein the concentration of 4a is determined by HPLC.

41. 41. The method of any one of claims 36-40, wherein the concentration of 4a in the bulk co-crystal of 4a and proline is about 3 to about 20% (w / w) higher than in the starting composition.

42. 42. The method of any one of claims 36-41, wherein the concentration of 4a in the bulk co-crystal of 4a and proline is about 5 to about 15% (w / w) higher than in the starting composition.

43. 43. The method of any one of claims 36-42, wherein the concentration of 4a in the bulk co-crystal of 4a and proline is increased by about 5%, about 10%, or about 15% (w / w).

44. 31. The cocrystal of any one of claims 1 to 30 for treating or preventing a neurodegenerative disorder in a subject suffering from the neurodegenerative disorder.

45. 31. The cocrystal of any one of claims 1 to 30 for preventing or reducing the likelihood of developing diabetic peripheral neuropathy in a subject suffering from type 1 or type 2 diabetes.

46. 31. Use of a cocrystal according to any one of claims 1 to 30 in the manufacture of a medicament for treating or preventing a neurodegenerative disorder in a subject suffering from the disorder.

47. 31. Use of a cocrystal according to any one of claims 1 to 30 in the manufacture of a medicament for preventing or reducing the likelihood of developing diabetic peripheral neuropathy in a subject suffering from type 1 or type 2 diabetes.

Images