CRISABOROLE COCRYSTAL DERIVATIVES

Field of the Invention

The present invention pertains to the field of pharmaceutical solid forms of Crisaborole, particularly methods of making such solid forms with varying dissolution properties and uses thereof for various medical treatments. More particularly the pharmaceutical forms are Crisaborole cocrystals obtained with caffeine, picolinic acid, quinaldic acid and pyridoxine as coformers.

Background of the Invention

The present invention refers to derivatives of 4-((1 -Hydroxy-1 ,3- dihydrobenzo[c][1 ,2]oxaborol-5-yl)oxy)benzonitrile (Crisaborole). Crisaborole is a nonsteroidal topical medication used for the treatment of mild-to-moderate atopic dermatitis in adults and children. It is a phosphodiesterase 4 (PDE-4) inhibitor. PDE-4 inhibition results in increasing levels of intracellular cyclic adenosine monophosphate.

Table 1. General information of the pure drug.

Three crisaborole anhydrous forms were described, named Form I, Form

II and Form III. In spite of the reported data, the nature of the forms is still not clearly disclosed, and the same form is reported to be an anhydrous species (Cryst. Growth Des., 2018, 18, 4416) and a hydrate derivative (US2021002307A1 ).

Moving further, several solvates were described, named Form D, Form E and Form F; finally, additional species, such as Form III (labeled as Form V) and Form IV, the nature of which was not completely disclosed, were reported in some patents.

With regard to Crisaborole cocrystals, only one example was reported in literature, with the species arising from the reaction of Crisaborole with 4,4'- Bipyridine (Cryst. Growth Des. 2018, 18, 4416). No other derivatives were found in literature.

Cocrystals are crystalline molecular complexes of two or more nonvolatile compounds bound together in a crystal lattice by non-ionic interactions. Pharmaceutical cocrystals are cocrystals of a therapeutic compound, e.g., an active pharmaceutical ingredient (API), and one or more non-volatile compound(s) (referred to herein as coformer). A coformer in a pharmaceutical cocrystal is typically a non-toxic pharmaceutically acceptable molecule, such as, for example, food additives, preservatives, pharmaceutical excipients, or other APIs. A cocrystal of an API is a distinct chemical composition of the API and coformer(s) and generally possesses distinct crystallographic and spectroscopic properties when compared to those of the API and coformer(s) individually. Crystallographic and spectroscopic properties of crystalline forms are typically measured by X-ray powder diffraction (XRPD) and single crystal X-ray crystallography, among other techniques. Cocrystals often also exhibit distinct thermal behavior. Thermal behavior is measured in the laboratory by such techniques as capillary melting point, thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC).

As crystalline forms, cocrystals may possess more favorable solid state, physical, chemical, pharmaceutical and/or pharmacological properties or may be easier to process than known forms or formulations of the API. For example, a cocrystal may have different dissolution and/or solubility properties than the API, and can, therefore, be more effective in therapeutic delivery. In the case of Crisaborole this may include more rapid skin permeation due to enhanced aqueous dissolution to favor formulation or alternatively a controlled slower release formulation due to reduced aqueous dissolution. The skin is a multicomplex membrane and changes from an avascular and lipophilic structure (stratum corneum) to a more aqueous structure (the viable epidermis and dermis). Uncomplicated penetration of a substance requires both solubility in the lipophilic environment and the more aqueous environment (Xenobiotica, 1987, 17, (3), 325). A cocrystal may also affect other pharmaceutical parameters such as storage stability, compressibility and density (useful in formulation and product manufacturing), permeability, and hydrophilic or lipophilic character. New pharmaceutical compositions comprising a cocrystal of a given API, therefore, may have attractive or superior properties as compared to its natural state or existing drug formulations.

For the sake of clarity and comprehension of the description, the following Table comprises a Glossary of terms used along the same.

Table 2. Glossary: Abbreviation table

Brief description of the drawings

The features and advantages of certain embodiments will be more readily appreciated when considered in conjunction with the accompanying figures. The figures are not to be construed as limiting any of the preferred embodiments.

Figure 1 is the XRPD pattern of solid TRQ03-CAF NP01 isolated from SL experiment using CAF.

Figure 2 is the ¹ HNMR spectrum of TRQ03-CAF NP01 .

Figure 3 is the FT-IR spectrum of TRQ03-CAF NP01 .

Figure 4 shows the DSC profile of TRQ03-CAF NP01 .

Figure 5 is the TGA profile of TRQ03-CAF NP01 . Figure 6 is the XRPD pattern of solid TRQ03-PIC NP01 isolated from SL experiment using PIC.

Figure 7 is the ¹ HNMR spectrum of TRQ03-PIC NP01. The structures of Crisaborole and picolinic acid are included to specify some of the signals.

Figure 8 is the FT-IR spectrum of TRQ03-PIC NP01 .

Figure 9 is the DSC profile of TRQ03-PIC NP01 .

Figure 10 is the TGA profile of TRQ03-PIC NP01 .

Figure 11 is the XRPD pattern of solid TRQ03-QNL NP01 isolated from SL experiment using QNL.

Figure 12 is the ¹ HNMR spectrum of TRQ03-QNL NP01. The structures of Crisaborole and quinaldic acid are included to specify some of the signals.

Figure 13 is the FT-IR spectrum of TRQ03-QNL NP01 .

Figure 14 is the DSC profile of TRQ03-QNL NP01 .

Figure 15 is the TGA profile of TRQ03-QNL NP01 .

Figure 16 is the XRPD pattern of solid TRQ03-PYX NP01 isolated from SL experiment using PYX.

Figure 17 is the ¹ HNMR spectrum of TRQ03-PYX NP01. The structures of Crisaborole and pyridoxine are included to specify some of the signals.

Figure 18 is the FT-IR spectrum of TRQ03-PYX NP01 .

Figure 19 is the TGA profile of TRQ03-PYX NP01 .

Figure 20 is the DSC profile of TRQ03-PYX NP01 .

Figure 21 is a comparative dissolution profile of TRQ03-CAF NP01 , TRQ03-PIC NP01 , TRQ03-QNL NP01 , TRQ03-PYX NP01 , and Crisaborole as- is.

Detailed description of the invention Since no standardized methods and/or definite rules have been established in the prior art for assessment of physico-chemical properties of coformers necessary for co-crystal formation (Arun Kumar et al., A Review about Regulatory Status and Recent Patents of Pharmaceutical Co-Crystals, Adv. Pharm. Bull., 2018, 8(3), 355-363, doi:10.15171/apb.2018.042, http://apb.tbzmed.ac.ir, see paragraph entitled “Non-obviousness” on page 357), 30 different coformers were screened and satisfactory results were obtained with caffeine, picolinic acid, pyridoxine and quinaldic acid.

The object of the present invention is to provide new cocrystals of Crisaborole having varied dissolution properties over the pure drug, methods of making such solid forms and uses thereof for various medical treatments. In some embodiments the disease or condition is selected from: psoriasis and allergic dermatitis which are non-infectious inflammatory diseases with chronic and recurrent disease. Currently, although some treatments can be used to control these diseases, other treatments are still under investigation. Appropriate treatments can help relieve symptoms and prolong the interval. Cribboron (also known as Crisaborole, AN-2728) is a topical boron-containing anti-inflammatory compound developed by Anacor Pharmaceuticals Inc. that inhibits PDE4 activity, thereby inhibiting TNFalpha, IL-12, IL-23 and release of other cytokines. Crisaborole has a good therapeutic effect on skin diseases such as psoriasis and allergic dermatitis. It was approved by the US FDA on December 14, 2016.

EXPERIMENTAL

Preferred embodiments of the invention In a preferred embodiment of the invention, a Crisaborole cocrystal was prepared named TRQ03-CAF NP01 using caffeine as the coformer.

In a further preferred embodiment of the invention, a Crisaborole cocrystal was prepared named TRQ03-PIC NP01 using picolinic acid as the coformer.

In another preferred embodiment of the invention, a Crisaborole cocrystal was prepared named TRQ03-QNL NP01 using quinaldic acid as the coformer.

In another preferred embodiment of the invention, a Crisaborole cocrystal was prepared named TRQ03-PYX NP01 using pyridoxine as the coformer. Preparation and evaluation of the preferred cocrystals of the invention

1) TRQ03-CAF NP01

Crisaborole SM Form III Caffeine

Synthesis

Crisaborole (4 g) and 1 .2 equivalents of Caffeine (3.713 g) were weighed in a 100 mL reactor vessel equipped with an anchor stirrer. Ethanol/Water 1 :1 v/v mixture (80 mL) was added, and the reaction mixture was stirred at 300 rpm overnight. The almost complete dissolution of the suspended powders occurred after the solvent addition, followed by the immediate precipitation of a solid.

After approx. 16 hours, the solid was isolated by vacuum filtration using a buchner filter equipped with a 55 mm por. 42 paper filter, washed with 8 mL of solvent mixture, and dried on the filter for approx. 30 minutes. The recovery of the desired derivative was confirmed by XRPD analysis. The isolated solid was then dried at 50°C for approx. 16 hours, and the recovery of the desired derivative was again confirmed by XRPD analysis (Figure 1 ). 6.714 g of TRQ03-CAF NP01 were isolated (yield = 91.0%).

The ¹ HNMR spectrum (Figure 2) revealed the presence of Crisaborole and Caffeine in the stoichiometric ratio TRQ03 : CAF 1 :1 .

1 H-NMR (400 MHz, DMSO-d6) 5: 9.22 (s, 1 H), 8.00 (d, J1 = 0.6 Hz, 1 H), 7.88-7.83 (m, 2H), 7.80 (d, J1 = 8.0 Hz, 1 H), 7.18-7.13 (m, 3H), 7.10 (dd, J1 = 2.0 Hz, J2 = 8.0 Hz, 1 H), 4.97 (s, 2H), 3.87 (d, J1 = 0.6 Hz, 3H), 3.40 (s, 3H), 3.21 (s, 3H).

The structural integrity of the molecule was confirmed.

FT-IR

Figure 3 shows the FT-IR spectrum of TRQ03-CAF NP01 and the peaks list can be found in the table below.

Table 3. FT-IR peak list of TRQ03-CAF NP01

Thermal analysis (DSC and TGA)

The DSC profile (Figure 4) showed a broad endothermic event at 86.5°C (Onset 71 ,6°C).

From TG-EG analysis (Figure 5) a weight loss of 4.3% w/w was observed in correspondence to the event recorded by DSC. The nature of the gas evolved during the TG analysis was disclosed by means of combined EG analysis. No evolution of organic solvent was detected in correspondence to the observed weight loss, so the presence of water could be assumed. The observed weight loss was compatible with the recovery of a monohydrate derivative. Based on this observation, the observed DSC event could be ascribed to simple dehydration and amorphization, since no other events were highlighted before sample decomposition, which occurred above approx. 240°C.

2) TRQ03-PIC NP01

Crisaborole SM Form III Picolinic acid

Synthesis

Crisaborole (4 g) and 1.2 equivalents of picolinic acid (2.354 g) were weighed in a 100 mL reactor vessel equipped with an anchor stirrer. Methanol (80 mL) was added, and the reaction mixture was stirred at 300 rpm overnight. The almost complete dissolution of the suspended powders occurred after the solvent addition, followed by the immediate precipitation of a solid.

After approx. 16 hours, the solid was isolated by vacuum filtration using a buchner filter equipped with a 55 mm por. 42 paper filter, washed with 8 mL of solvent, and dried on the filter for approx. 15 minutes. The recovery of the desired derivative was confirmed by XRPD analysis.

The isolated solid was then dried at 50°C for approx. 16 hours, and the recovery of the desired derivative was again confirmed by XRPD analysis (Figure 6). 4.946 g of TRQ03-PIC NP01 were isolated (yield = 83.0%).

The ¹ HNMR spectrum (Figure 7) revealed the presence of two species in solution, probably in equilibrium between each other.

This evidence is corroborated by the split of the signals at approx. 5 ppm, ascribable to the (Ph)-CH2-O(B) group. As can be seen, in the described region of the spectrum, both a singlet and a double doublet were observed: the singlet is related to the Crisaborole molecule “as-is”, while the “dd” signal could be tentatively ascribed to a species in which the B atom of Crisaborole is somehow coordinated with the N atom of the pyridine ring of Picolinic Acid, thus leading to the formation of a structure in which the H atoms of the CH2 group are diastereotopic.

Because of this evidence, an unambiguous assignment of the signals, especially in the aromatic region, could not be done, and the stoichiometric ratio between the components of the species was calculated considering the integral of all the signals. For this reason, a value of 2 was assigned to the integral in the region 5.2-4.9 ppm, and the aromatic protons were observed to be 11 .

Based on this data, the stoichiometric ratio TRQ03 : PIC was calculated as 1 :1 .

FT-IR

Figure 8 shows the FT-IR spectrum of TRQ03-PIC NP01 and the peaks list can be found in the table below.

Table 4. FT-IR peak list of TRQ03-PIC NP01

Thermal analysis (DSC and TGA) The DSC profile (Figure 9) showed a sharp endothermic event at 244.5°C (Onset 242.5°C), ascribable to sample melting and decomposition.

TG-EG analysis (Figure 10) confirmed the recovery of an anhydrous compound, since no weight loss was observed before simple degradation, which occurred at approx. 260°C.

Moreover, combined TG-EG analysis confirmed the presence of picolinic acid in the isolated derivative, as massive evolution of carbon dioxide was recorded during sample decomposition along with the evolution of o- anisylcyanide, ascribable to decomposition of Crisaborole.

3) TRQ03-QNL NP01

Crisaborole SM Form III Quinaldic acid

Synthesis

Crisaborole (2.5 g) and 1 .2 equivalents of quinaldic acid (2.069 g) were weighed in a 100 mL reactor vessel equipped with an anchor stirrer. Methanol (50 mL) was added, and the reaction mixture was stirred at 300 rpm overnight. The almost complete dissolution of the suspended powders occurred after the solvent addition, followed by the immediate precipitation of a solid.

After approx. 16 hours, the solid was isolated by vacuum filtration using a buchner filter equipped with a 55 mm por. 42 paper filter, washed with 8 mL of solvent, and dried on the filter for approx. 15 minutes.

The isolated solid was then dried at 50 °C for approx. 16 hours, and the recovery of the desired derivative was confirmed by XRPD analysis (Figure 11 ). 3.564 g of TRQ03-QNL NP01 were isolated (yield = 84.4%). From the ¹ HNMR spectrum (Figure 12), it is possible to observe the presence of two species in solution, probably in equilibrium between each other, as already described for the TRQ03-PIC NP01 derivative.

In fact, also in this case, the split of the signals at approx. 5 ppm, ascribable to the (Ph)-CH2-O(B) group, was observed. As can be seen, in the described region of the spectrum, both a singlet and a double doublet were found: the singlet is related to the Crisaborole molecule “as-is”, while the dd signal could be tentatively ascribed to a species in which the B atom of Crisaborole is somehow coordinated with the N atom of the pyridine ring of quinaldic acid, thus leading to the formation of a structure in which the H atoms of the CH2 group are diastereotopic.

Because of this evidence, an unambiguous assignment of the signals, especially in the aromatic region, could not be done, and the stoichiometric ratio between the components of the species was calculated considering the integral of all the signals. For this reason, a value of 2 was assigned to the integral in the region 5.2-4.9 ppm, and the aromatic protons were observed to be 13.

Based on this data, the stoichiometric ratio TRQ03 : QNL was calculated as 1 :1 .

FT-IR

Figure 13 shows the FT-IR spectrum of TRQ03-QNL NP01 and the peaks list can be found in the table below.

Table 5. FT-IR peak list of TRQ03-QNL NP01

Thermal analysis (DSC and TGA) The DSC profile (Figure 14) showed a sharp endothermic event at 267.4°C (Onset 263.8°C), ascribable to sample melting and decomposition.

TG-EG analysis (Figure 15) confirmed the recovery of an anhydrous compound, since no weight loss was observed before sample degradation, which occurred at approx. 260°C.

Moreover, combined TG-EG analysis confirmed the presence of quinaldic acid in the isolated derivative as evolution of carbon dioxide was recorded during sample decomposition.

4) TRQ03-PYX NP01

Crisaborole SM Form III Pyridoxine

Synthesis

Crisaborole (4 g) and 1.2 equivalents of pyridoxine (3.235 g) were weighed in a 100 mL reactor vessel equipped with an anchor stirrer. Methanol (80 mL) was added, and the reaction mixture was stirred at 300 rpm overnight. The almost complete dissolution of the suspended powders occurred after the solvent addition, followed by the immediate precipitation of a solid.

After approx. 16 hours, the solid was isolated by vacuum filtration using a buchner filter equipped with a 55 mm por. 42 paper filter, washed with 8 mL of solvent, and dried on the filter for approx. 30 minutes. The recovery of TRQ03- PYX NP02 (some signals shift with respect to NP01 , dry sample) derivative was confirmed by XRPD analysis. The isolated solid was then dried at 50 °C for approx. 48 hours, and the recovery of the desired TRQ03-PYX NP01 derivative was confirmed by XRPD analysis (Figure 16). 6.602 g of TRQ03-PYX NP01 were isolated (yield 94.6%).

From the ¹ HNMR spectrum (Figure 17), it is possible to observe the presence of more species in solution in equilibrium between one another.

However, unlike what was observed for PIC and QNL derivatives, in this case the split of the signals at approx. 5 ppm could be ascribed to the pyridoxine moiety and not to the Crisaborole unit.

An equilibrium between two tautomeric forms of pyridoxine has been described in solution (J. Org. Chem. 2000, 65, 2716). As a result, both signals of the (Pyr)-CH2-O(H) groups of pyridoxine were split into couples, thus leading to the presence of 4 signals, along with the signal of the Crisaborole (Ph)-CH2- O(B) group.

As already described for other derivatives, because of this evidence, an unambiguous assignment of the signals, especially in the aromatic region, could not be done, and the stoichiometric ratio between the components of the species was calculated considering the integral of all the signals. For this reason, a value of 6 was assigned to the integral in the region 5.0-4.4 ppm, and the aromatic protons were observed to be 8.

Based on this data, the stoichiometric ratio TRQ03 : PYX was calculated as 1 :1 .

FT-IR

Figure 18 shows the FT-IR spectrum of TRQ03-PYX NP01 and the peaks list can be found in the table below.

Table 6. FT-IR peak list of TRQ03-PYX NP01

Thermal analysis (TGA and DSC)

TG analysis (Figure 19) recorded two consecutive weight losses partially overlapped of approx. 1.5% w/w in the region 25-100°C, and of approx. 4.2% w/w in the region 100-180°C. Nevertheless, no evolution of solvent was detected by EG analysis, suggesting the presence of water in the analyzed sample. The first weight loss could be so ascribed to the presence of imbibition water, while the second weight loss is compatible with the recovery of a monohydrated derivative.

The DSC profile (Figure 20) showed a broad endothermic event at 112.7°C (Onset 82.1 °C), followed by a second small endothermic event at 169.6°C (Onset 165.0°C) and an exothermic event at 176.0°C (Onset 171.9°C).

Calorimetric events above 160°C are not well explained by comparison with TGA/EGA as a crystallization event (exo at 172°C) should not be excluded, but melting points of single components occur near that temperature region. The dehydration process was found to be a potentially critical point for stability of the derivative.

Powder dissolution results

Powder dissolution experiments were carried out in potassium phosphate buffer - pH 6.8.

The solubility of the prepared Crisaborole derivatives was evaluated through powder dissolution experiments. In a typical powder dissolution experiment, approx. 225 mg of sample were added to 600 mL of dissolution medium (pH 6.8 potassium phosphate buffer) and the resulting mixture was stirred at 37 °C and 100 rpm. Aliquots were withdrawn from the flask at regular intervals and then injected in HPLC.

Powder dissolution experiments were carried out on Crisaborole as well as on all the scale-up derivatives, such as TRQ03-CAF NP01, TRQ03-PIC NP01, TRQ03-PYX NP01 and TRQ03-QNL NP01.

Table 7 summarizes the results of powder dissolution tests conducted with TRQ03-SM and scale-up cocrystals prepared with CAF, PIC, PYX and QNL.

Table 7

With reference to Figure 21 , the following observations arise:

Crisaborole: the pure drug showed a dissolution maximum under the investigated conditions after approx. 2 hours of analysis (14.1% of powder dissolved). After this time, a decrease of the concentration of the API was observed (6 hours, 10.5% of powder dissolved), then followed by a slow increase until the final value of 14.6% of powder dissolved.

TRQ03CAF NP01 : a dissolution maximum was observed after approx. 2 hours of analysis, and a value of 19.0% of powder dissolved. Moreover, after this time, a decrease of the concentration of Crisaborole was observed, with the minimum reached after approx. 6 hours (14.3% of powder dissolved) readily followed by an apparently slight increase of the dissolved powder concentration, until a final value of 19.6% of powder dissolved.

TRQ03PIC NP01 : the maximum dissolution was reached after approx. 4 hours of analysis (16.1% of powder dissolved), followed by a decrease (6 hours, 12.2% of powder dissolved) and a further increase of the amount of dissolved Crisaborole, until a final value of 17.3% of powder dissolved after 24 hours of analysis.

TRQ03PYX NP01 : the dissolution profile showed 15.2% of powder dissolved after approx. 2 hours of analysis, then the concentration of Crisaborole dissolved in the medium remained almost stable at 15.0% of powder dissolved until the end of the analysis.

TRQ03-QNL NP01 : the dissolution profile showed 6.2% of powder dissolved after approx. 2 hours of analysis, 15.3% after 6 hours, and the maximum value was observed at the end of the test (24.9% of powder dissolved). It is worth mentioning that this is the highest value observed within the performed experiments, suggesting that TRQ03-QNL NP01 is the most soluble species in the used medium after 24 hours.

General comments

- After 2 hours of dissolution, the caffeine derivative showed a faster dissolution rate with respect to Crisaborole as-is, even at the same order of concentration (only approx. 1.3-fold higher).

- The picolinic acid derivative showed a slower dissolution rate with respect to Crisaborole as-is (within 2 hours of analysis). - The pyridoxine derivative showed the same dissolution kinetics as compared to Crisaborole as-is, but showed a constant profile after 2 hours of analysis.

- The quinaldic acid derivative showed the lowest dissolution kinetics with reference to Crisaborole as-is, but an apparently constant increase in time.