Investigation of Various Impurities in Febuxostat

Febuxostat is a novel, non-purine, selective inhibitor of xanthine oxidase for hyperuricemia in patients with gout. It is the first promising substitute for allopurinol in 40 years. Various synthetic routes to febuxostat, as well as polymorphic forms and impurities of the drug, are reported in the literature. The authors have also identified several impurities that result from the synthesis of febuxostat. This article describes the identification and control of all isomeric, carryover, and byproduct impurities of febuxostat and its intermediates.

Febuxostat is a novel, non-purine, selective inhibitor of xanthine oxidase for hyperuricemia in patients with gout (1). Febuxostat was discovered by Teijin and approved by FDA in February 2009 (2, 3). The drug reduces uric acid production by inhibiting the activity of xanthine oxidase, an enzyme that, in the last step of purine metabolism, converts xanthine to uric acid (4). Febuxostat has emerged as the foremost treatment alternative for gout and is considered the first promising substitute to allopurinol in more than 40 years. Research has shown febuxostat to be well tolerated in long-term treatment in patients with hyperuricemia, including those experiencing intolerance to allopurinol (5, 6).

Febuxostat is a 2-arylthiazole derivative with a methyl carboxyl group (-CH₂COOH). More than 50 polymorphic forms of febuxostat have been reported, including Crystal A and several others disclosed by Teijin (7). While various febuxostat synthesis routes starting from 4-hydroxybenzonitrile have been reported, far less information on isomeric, carryover, and byproduct impurities is available (8-11). The impurity profile of a drug substance is of increasing importance for ensuring the quality of drug products (11, 12). However, it is extremely challenging for an organic chemist to identify impurities that form in small quantities and particularly burdensome if the product is non-pharmacopoeial (13). This article describes the identification and synthesis of various impurities that form during the production of febuxostat and its intermediates as well as strategies for minimizing the formation of all isomeric, carryover, and by-product impurities of febuxostat and its intermediates.

Materials and methods
All chemicals and solvents were purchased from Avra Synthesis (Hyderabad), Neogen Chemicals (New Bombay), and Hangzhou Dayangchem (China). Hydrogen-1 nuclear magnetic resonance (¹H-NMR) was performed on a 300-MHz Fourier transform (FT)-NMR (Brucker) using either deuterated chloroform (CDCl₃) or deuterated dimethyl sulfoxide DMSO-d₆ or both as solvent, and tetramethylsilane (TMS) as the internal standard. Mass spectrometry (MS) was performed on a Quattro micro API mass spectrometer 0-800 Da in auto specifications. Infrared (IR) spectroscopy was carried out using a PerkinElmer 100 FT-IR. High-performance liquid chromatography (HPLC) was performed on a Shimadzu LC system with Inertsil C18 columns (150 & #215; 4.6 mm, 3.0 & #956;m); acetonitrile mobile phase; 80:20 buffer solution (1.36 g KH₂PO₄ in 1 L water, pH adjusted to 2.0 & #177; 0.05 with diluted H₃PO₄); and a flow rate of 1.0 mL/min. For Impurity XIX, HPLC was performed on a Waters LC system with Chiralpak IC columns (250 & #215; 4.6 mm, 5.0 & #956;m); n-hexane mobile phase; ethanol:trifluoroacetic acid (EtOH:TFA) buffer solution (95:5:0.1); and a flow rate of 1.5 mL/min.

2-Hydroxybenzenecarbothioa mide (Impurity VIII). Magnesium Chloride Hexahydrate (MgCl₂.6H₂O, 34.1 g, 0.167 mol) was added to a stirred solution of 2-cyanophenol (10.0 g, 0.084 mol) in dimethylformamide (DMF, 100 mL) at 25°C. To this solution, 30% sodium hydrosulfide (NaHS, 46.9 mL, 0.252 mol) was added at 25°C. The reaction mixture was heated to 45-50°C for 15-18 h. Reaction progress was monitored by thin layer chromatography (TLC). After completion of the reaction, the solution was cooled to 25°C, and 100 mL of water was added. The solution was adjusted to pH 1-2 using 5N hydrochloric acid (HCl). The product was extracted using ethyl acetate (3 & #215; 100 mL), with the combined ethyl acetate layer washed with water (2 & #215; 50 mL), and finally with brine (50 mL). The organic layer was evaporated to dryness on a rotavapor. The crude compound was dried in an air oven at 60°C for 12 h to obtain Impurity VIII (10.92 g, 85%).

Ethyl 2-(2-hydroxyphenyl)-4-methyl-1,3-thiazole-5-carboxylate (Impurity IX). Impurity VIII (5 g, 0.033 mol) in isopropyl alcohol (20 mL) was heated at 65°C, and ethyl-2-chloroacetoacetate (5 mL, 0.036 mol) was added dropwise for 10 min. The reaction mixture was refluxed for 1 h and cooled to 0-5°C for 1 h. The isolated solid was filtered and washed with cyclohexane (5 mL). The yellow solid was dried in an air oven to obtain Impurity IX (8.1 g, 95%).

Ethyl 2-(3,5-diformyl-4-hydroxyphenyl)-4-methyl-1,3-thiazole-5-carboxylate (Impurity X). Hexamine (26.6 g, 0.190 mol) was added to the stirred solution of Compound III
(10.0 g, 0.038 mol) in TFA (50 mL) at 25°C. The reaction mixture was heated at 100°C for 12 h, and cooled to 25°C. Following that, 250 mL of water was added and stirred for 1 h and the yellow solid was filtered. The crude compound was loaded on a silica gel column and eluted with ethyl acetate: hexane (15:85) to obtain Impurity X (3.81 g, 31%).

Ethyl 4-methyl-2-[4-(2-methylpropoxy)phenyl]-1,3-thiazole-5-carboxylate (Impurity XI). Potassium carbonate (21.0 g, 0.152 mol), potassium iodide (0.315 g, 0.002 mol), and isobutyl bromide (10.37 mL, 0.095 mol) were added to the solution of Compound III (10.0 g, 0.038 mol) in 50 mL DMF. The heterogeneous mixture was heated at 70-75°C for 10-12 h. When the reaction was completed, the reaction mass was cooled to 25°C, and water (150 mL) was added. The isolated solid was filtered, washed with water (50 mL), and dried in an air oven to get Impurity XI (8.1 g, 67%).

Ethyl 2-[3,5-diformyl-4-(2-methylpropoxy)phenyl]-4-methyl-1,3-thiazole-5-carboxylate (Impurity XII). The experimental procedure is similar to Impurity XI but with Impurity X used as the starting material.

Ethyl 2-(4-butoxy-3-formylphenyl)-4-methyl-1,3-thiazole-5-carboxylate (Impurity XIII). The experimental procedure is similar to Impurity XI but with Compound IV and n-butyl bromide used as the starting materials.

Ethyl 2-[3-formyl-4-(1-methylpropoxy)phenyl]-4-methyl-1,3-thiazole-5-carboxylate (Impurity XIV). The experimental procedure is similar to Impurity XI but the starting material was Compound IV and 2-methyl propyl bromide.

Ethyl 2-(4-butoxy-3-cyanophenyl)-4-methyl-1,3-thiazole-5-carboxylate (Impurity XV). Hydroxylamine hydrochloride (2.4 g, 0.035 mol) and sodium formate (3.13 mol, 0.046 mol) were added to a stirred solution of Impurity XIII (10.0 g, 0.029 mol) in formic acid (50 mL) and refluxed for 3-4 h (TLC). The reaction mass was cooled to 25°C and water (200 mL) was added. The solid was filtered, washed with water (100 mL), and dried in an air oven to obtain Impurity XV (8.4 g, 85%).

Ethyl 2-[3-cyano-4-(1-methylpropoxy)phenyl]-4-methyl-1,3-thiazole-5-carboxylate (Impurity XVI). The experimental procedure is similar to Impurity XV but with Impurity XIV used as the starting material.

Methyl 2-[3-cyano-4-(2-methylpropoxy)phenyl]-4-methyl-1,3-thiazole-5-carboxylate (Impurity XVII). Compound I (5.0 g, 0.016 mol) was suspended in methanol (25 mL). Thionyl chloride (3.5 mL, 0.047 mol) was slowly added. The reaction mixture was heated to reflux for 12 h. It was cooled to 25°C and water (250 mL) was added. The solid was filtered, washed with water (50 mL), and dried in an air oven to obtain Impurity XVII (5.0 g, 96%).

4-Methyl-2-[4-(2-methylpropoxy)phenyl]-1,3-thiazole-5-carboxylic acid (Impurity XVIII). MeOH:THF (1:1) (50 mL) was added to Impurity XI (5.0 g, 0.016 mol). NaOH (0.80 g, 0.020 mol) in water (25 mL) was added to this suspension at 25°C. The reaction mixture was heated at 45-50°C for 1-2 h and monitored by TLC. The reaction mixture was cooled to 25°C and water (25 mL) was added. The reaction mixture was adjusted to pH 1-2 by using 5N HCl (5-10 mL). The fall out solid was filtered, washed with water (10 mL), and dried in an air oven to obtain Impurity XVIII (3.5 g, 76%).

2-(4-Butoxy-3-cyanophenyl)-4-methyl-1,3-thiazole-5-carboxylic acid (Impurity XIX). The experimental procedure is similar to Impurity XVIII but with Impurity XV used as the starting material.

2-[3-Cyano-4-(1-methylpropoxy)phenyl]-4-methyl-1,3-thiazole-5-carboxylic acid (Impurity XX). The experimental procedure is similar to Impurity XVIII but Impurity XVI was used as the starting material.

2-[3-Carbamoyl-4-(2-methylpropoxy)phenyl]-4-methyl-1,3-thiazole-5-carboxylic acid (Impurity XXI). The experimental procedure is similar to Impurity XVIII but the starting material was Compound VII, NaOH (0.08 mol), and recrystallization in MeOH.

2-[3-Carboxy-4-(2-methylpropoxy)phenyl]-4-methyl-1,3-thiazole-5-carboxylic acid (Impurity XXII). The experimental procedure is similar to Impurity XVIII but the starting material was Compound VII in NaOH (0.08 mol), with recrystallization in MeOH.

Results and discussionFigure 1 describes synthesis of febuxostat (I) from 4-hydroxybenzonitrile (II) in six stages. The synthesis shown is a short, concise route and does not require use of poisonous reagents such as KCN (14). Compound II was converted to 4-hydroxybenzothioamide (III) with 85% yield using NaHS in the presence of hydrated magnesium chloride as Lewis acid. Intermediate III, on cyclization with ethyl-2-chloroacetoacetate, gave thiazole ester (IV) with quantitative yield. In these two stages, the source of potential impurities was identified as an ortho isomer (i.e., 2-hydroxybenzonitrile), which can lead to Impurity VIII and subsequently to Impurity IX (see Figure 2). Impurities VIII and IX can be controlled in starting material II with appropriate specification.

The ortho formylation of hydroxyl compound IV by using Duff condition (hexamine/TFA) gave aldehyde V (15). The major impurity identified in this reaction was dialdehyde X. Although we have used only 1.0 equivalence of hexamine with respect to Compound IV, the dialdehyde X impurity was formed to a 5-10% ratio in only 2.5 h. It is, therefore, impossible to get rid of this impurity during the reaction, and only effective recrystallization will eliminate it. Impurity X was minimized (≤ 2%) by recrystallization using IPA/H₂O (3:5) to get aldehyde V with 50% yield and & #8805; 97% HPLC purity.

Aldehyde V, on alkylation with isobutyl bromide in the presence of potassium carbonate base, gave compound VI with 90% yield. In this stage, Impurities XI and XII were alkylations of carryover Compound IV and dialdehyde, respectively. Two more isomeric impurities n-butyl-aldehyde XIII and 1-methyl propyl-aldehyde XIV were also identified in this stage. Both isomeric impurities can be controlled with appropriate specification for isobutyl bromide. The reaction of Compound VI with hydroxylamine hydrochloride and sodium formate in formic acid at reflux temperature gave Compound VII with 85% yield. Impurities XIII and XIV will also carry forward to impurities n-butyl-nitrile XV and 1-methyl propyl-nitrile XVI, respectively.

In the final step, Compound VII was hydrolyzed using sodium hydroxide in a MeOH:THF:H2O (1:1:1) solvent combination to yield febuxostat (85%). During saponification, methyl ester Impurity XVII was identified via trans-esterification. Its hydrolysis was comparatively slower than its ethyl isomer VII. One way to avoid Impurity XVII is to replace methanol with ethanol. Carryover impurities XI, XV, and XVI were also hydrolyzed to their respective acid derivatives impurities XVIII, XIX, and XX. However, the acid derivatives of impurities X and XII were unexpectedly absent as impurities. It is believed that, because they were present in low concentrations during workup, they were eliminated in the mother liquor. Two additional impurities, amide XXI and diacid XXII, formed by the side reaction of the febuxostat nitrile group with sodium hydroxide, were identified during saponification. The amide XXI and diacid XXII impurities can be controlled by using appropriate equivalence of sodium hydroxide and controlled reaction time. Febuxostat, on acetone recrystallization and seed Crystal A at 45°C, gave pure febuxostat with 75% yield.

S.No.

Impurity

1H-NMR (δ) ppm

IR (υ) cm-1

MS (M++ H) m/z

1

VIII

6.75-6.80 (t, 1H), 6.88-6.91 (d, 1H, J= 8.1 Hz), 7.17-7.23 (m, 1H), 8.04-8.07 (d, 1H), 10.10 (s, 1H)

3364.1, 3191.1, 1621.3. 1604.1

153.98

2

IX

1.17-1.22 (t, 3H), 2.67 (s, 3H), 4.14-4.21 (dd, 2H, J= 7.2 Hz each), 6.75-6.80 (t, 1H), 6.88-6.91 (d, 1H, J= 8.1 Hz), 7.17-7.23 (m, 1H), 8.04-8.07 (d, 1H)

3053.6, 2994.3, 2530.3, 1703.4, 1605.7, 1298.7

264.01

3

X

1.38-1.43 (m, 3H), 2.80 (s, 1H), 4.34-4.41 (dd, 2H, J= 7.2 Hz each), 8.58 (s, 2H), 10.32 (s, 2H), 11.89 (s, 1H)

2988.3, 1705.0, 1665.2, 1652.4, 1268.3

320.10

4

XI

1.04-1.06 (d, 6H), 1.39-1.42 (d, 3H, J= 7.2 Hz), 2.07-2.16 (m, 1H), 2.78 (s, 3H), 3.77-3.79 (d, 1H, J= 6.6 Hz), 4.32-4.39 (dd, 2H, J= 7.2 Hz each), 6.93-6.96 (d, 2H), 7.91-7.94 (d, 2H)

2970.0, 1711.1, 1606.7, 1261.5

320.24

5

XII

1.08-1.14 (m, 6H), 1.22-1.28 (m, 3H), 2.20-2.34 (m, 1H), 2.89 (s, 1H), 3.95-3.97 (d, 2H, J=6.3 Hz), 4.33-4.40 (dd, 2H, J=7.2 Hz each), 8.66 (s,  2H), 10.44 (s, 2H)

2958.1, 1710.1, 1684.0, 1607.4, 1258.3

376.16

6

XIII

0.99-1.04 (t, 3H), 1.37-1.42 (m, 3H), 1.51-1.62 (m, 2H), 1.83-1.92 (m, 2H), 2.77 (s, 1H), 4.15-4.19 (t, 2H, J= 7.2 Hz each), 4.32-4.39 (dd, 2H, J= 6.9 Hz each ), 7.05-7.08 (d,  1H,  J= 8.7 Hz), 8.19-8.23 (dd, 1H, J= 2.4 Hz each), 8.35-8.36 (d, 1H, J= 2.4 Hz), 10.52 (s, 1H)

2931.2, 1709.4, 1608.2, 1287.4

348.36

7

XIV

1.00-1.05 (t, 3H), 1.37-1.42 (m, 6H), 1.70-1.88 (m, 2H), 2.79 (s, 1H), 4.31-4.39 (dd, 2H, J= 7.2 Hz each ), 4.53-4.59 (dd, 1H, J= 6 Hz each ), 7.05-7.08 (d,  1H,  J= 8.7 Hz), 8.17-8.21 (dd, 1H, J= 2.4 Hz each), 8.35-8.36 (d, 1H, J= 2.4 Hz), 10.51 (s, 1H)

2872.2, 1711.9, 1608.2, 1605.4, 1102.3

348.34

8

XV

0.99-1.04 (t, 3H), 1.37-1.42 (m, 3H), 1.53-1.63 (m, 2H), 1.83-1.92 (m, 2H), 2.77 (s, 3H), 4.13-4.18 (t, 2H, J= 7.2 Hz each),  4.33-4.40 (dd, 1H, J= 6 Hz each), 7.02-7.04 (d, 1H, J= 8.7 Hz), 8.08-8.12 (dd, 1H, J= 2.4 Hz each), 8.18-8.19 (d, 1H, J= 2.4 Hz)

2971.2, 2226.7, 1711.0, 1262.2

345.18

9

XVI

1.00-1.06 (t, 3H), 1.37-1.42 (m, 6H), 1.69-1.91 (m, 2H), 2.77 (s, 3H), 4.33-4.40 (dd, 2H, J= 7.2 Hz each),  4.48-4.54 (dd, 1H, J= 6 Hz each), 7.00-7.03 (d,  1H, J= 8.7 Hz), 8.06-8.10 (dd, 1H, J= 2.4 Hz each), 8.17-8.18 (d, 1H, J= 2.1 Hz)

2970.0, 2226.8, 1711.1, 1261.5

345.16

10

XVII

1.08-1.11 (d, 6H), 2.17-2.26 (m, 1H), 2.78-2.80 (d, 3H), 3.90-3.92 (m, 5H), 3.80-3.93 (m, 2H), 7.00-7.05 (dd, 1H, J= 3.3 Hz each), 8.08-8.12 (dd, 1H, J= 2.4 Hz each), 8.18-8.22 (dd, 1H, J= 2.1 Hz each)

2874.8, 2226.7, 1710.5, 1299.6

331.08

11

XVIII

1.04-1.07 (d, 6H), 2.06-2.19 (m, 1H), 2.80 (s, 3H), 3.78-3.80 (d, 1H, J= 6.3 Hz), 6.95-6.98 (d, 2H, J= 8.7 Hz), 7.91-7.94 (d, 2H, J= 8.7 Hz)

2965.7, 2530.1, 1680.7, 1604.4

292.14

12

XIX

0.86-0.91 (d, 3H), 1.38-1.50 (m, 2H), 1.70-1.80 (m, 2H), 2.64 (s, 3H), 4.02-4.06 (t, 2H), 6.92-6.95 (d, 1H), 7.95-7.99 (dd, 1H), 8.05-8.06 (d, 1H)

2959.2, 2226.4, 1731.7

317.09

13

XX

0.92-0.97 (m, 3H), 1.25-1.31 (m, 3H), 1.63-1.74 (m, 2H), 2.65 (s, 3H), 4.64-4.74 (m, 1H), 7.37-7.40 (d, 1H, J= 9 Hz), 8.17-8.20 (dd, 1H, J= 2.4 Hz each), 8.25-8.26 (d, 1H, J= 2.4 Hz), 13.42 (bs, 1H)

2973.2, 2229.3, 1676.7

317.10

14

XXI

0.99-1.00 (d, 6H), 2.08-2.16 (m, 1H), 2.66 (s, 3H), 3.96-3.98 (d, 2H), 7.24-7.27 (d, 1H, J= 9 Hz), 7.59 and 7.77 (2 bs, 2H), 8.01-8.05 (dd, 1H, J= 2.4 Hz each), 8.33-8.34 (d, 1H, J= 2.4 Hz), 13.5 (bs, 1H)

3850.2, 2874.3, 1928.4, 1644.9

335.21

15

XXII

0.89-1.00 (2d, 6H, J= 6.3 Hz each), 1.99-2.08 (m, 1H), 2.65 (s, 3H), 3.87-3.89 (d, 2H, J= 6.3 Hz), 7.20-7.23 (d, 1H, J= 8.7 Hz), 8.03-8.06 (t, 1H), 8.21-8.22 (d, 1H, J= 2.1 Hz), 13.08 (bs, 1H)

3850.2, 2957.3, 1914.9, 1695.1

336.11

A total of 15 impurities of febuxostat and its intermediates were synthesized and characterized by 1H-NMR, MS, and FT-IR (see Table I). Figure 3 shows a HPLC chromatogram for various impurities (VII, XVII, XVIII, XX, XXI, and XXII) of febuxostat. Impurity XIX could not be separated with the same HPLC protocol, so a separate method was developed to detect its presence. Figure 4 shows a HPLC chromatogram for impurity XIX. The single known and single unknown impurity in pure febuxostat specification is & #8804; 0.10%, and total impurities should be & #8804; 0.50%.

Conclusion
In summary, 15 impurities of febuxostat and its intermediates were identified, synthesised, and characterised. All impurities in the final stage (>95% HPLC purity) exhibited a well-defined separation from the parent febuxostat in the HPLC chromatogram. Only a single impurity XIX needed a separate HPLC method. A method to eliminate or minimise these impurities at various stages during the synthesis of febuxostat was also demonstrated. This work will help in determining more stringent specifications for the non-pharmacopeial drug substance febuxostat and enable long-term surveillance on its safety.

References
1. M. Hu and B. Tomlinson, Ther. Clin. Risk Manag. 4 (6) 1209–1220 (2008).
2. S. Kondo, H. Fukushima, M. Hasegawa, M. Tsuchimoto, I. Nagata, Y. Osada, K. Komoriya, and H. Yamaguchi, & #8220;2-Arylthiazole derivatives and pharmaceutical composition thereof,” US Patent 5614520, March 1997.
3. C.A. Woods and O. Hilas, Drug Forecast 35 (2) 82-85 (2010).
4. K.K.C. Liu et al., Bioorg. Med. Chem. 19 1136-1154 (2011).
5. L.A. Sorbera et al., Drugs Future 26 32-98 (2001).
6. P.I. Hair, P.L. McCormack, and G.M. Keating, Drugs 68 1865-1874 (2008).
7. M. Iwai, K. Nakamura, M. Dohi, H. Mochizuki, and S. Mochizuki, & #8220;Solid preparation containing single crystal form,” US Patent 7361676, April 2008.
8. M. Hasegawa, Heterocyles 47 (2) 857-864 (1998).
9. J. Canivet et al., Org. Lett. 11 (8) 1733-1736 (2009).
10. T. Yamamoto et al., Chem. Eur. J. 17 10113-10122 (2011).
11. M.H. Kadivara et al., J. Pharm. Biomed. Analysis 56 749-757 (2011).
12. K.R. Wadekar et al., Pharm. Tech. 36 (2) 46-51 (2012).
13. G.D. Patil et al., Org. process res. Dev. 16 (8) 1422–1429 (2012).
14. K. Watanabe, T. Yarino, and T. Hiramatsu, & #8220;Production of aldehyde derivative,” Japan Patent 3836177, August 2006.
15. J.C. Duff and E.J. Bill, J. Chem. Soc. 1987-1988 (1932). PTE

About the Authors
Anand M. Lahoti, PhD, Research scientist-1, anandlahoti@neulandlabs.com, Neuland Laboratories Ltd, Research & Development Center, Hyderabad, AP, India
Ponnaiah Ravi, PhD, President–technical, Neuland Laboratories Ltd, Research & Development Center, Hyderabad, AP, India
Neela Praveen Kumar, PhD, General Manager, Neuland Laboratories Ltd, Research & Development Center, Hyderabad, AP, India
V. Innareddy is a Research Associate, Neuland Laboratories Ltd, Research & Development Center, Hyderabad, AP, India
P. S. Deepthi, Senior Research Associate, Neuland Laboratories Ltd, Research & Development Center, Hyderabad, AP, India
V. Shanmugam, Senior Research Associate, Neuland Laboratories Ltd, Research & Development Center, Hyderabad, AP, India
M. Sudhakar Rao, Research Scientist, Neuland Laboratories Ltd, Research & Development Center, Hyderabad, AP, India
Vivekananda Reddy, PhD, Research Scientist, Neuland Laboratories Ltd, Research & Development Center, Hyderabad, AP, India

*To whom all correspondence should be addressed. Submitted: 12 February 2013. Accepted: 23 Apr 2013