Synthesis of Avatrombopag

Avatrombopag is a non-chiral compound that contains 6 rings in the molecule, 5 of which are heterocycles. The 4 six-membered rings are cyclohexane, piperazine, pyridine, and piperidine; the 2 five-membered rings consist of thiophene and thiazole. The rings are directly connected to one another or through an amide bond. The key steps for the synthesis of Avatrombopag involve a selective nucleophilic aromatic substitution reaction and an amide bond formation reaction.

The synthesis begins with an α-keto halogenation of the thiophene to give the corresponding bromide.

The α-keto halogenation reaction usually takes place under acidic or basic conditions in an aqueous medium with the corresponding elemental halogen (chloride, bromide, or iodide). The reason why the alpha position to the carbonyl group in a ketone is easily halogenated is the keto-enol tautomerism. The enol or the enolate form of the ketone attacks the halogen, and the newly introduced halogen decreases the basicity of the carbonyl oxygen while making the remaining hydrogens more acidic. Therefore, in an acidic solution, only one alpha hydrogen is replaced by a halogen because each successive halogenation is slower than the previous. However, in a basic solution successive halogenations are more likely due to the inductive electron withdrawal by the halogen.

Condensation with thiourea produces the thiazolamine ring of the molecule in 46% yield over two steps. This is the Hantzsch thiazole synthesis.

When a thioamide or urea reacts with an α-haloketone, both the nitrogen and the sulfur act as nucleophiles, so which one attacks the oxygen and which one attacks the halogen? Carbonyl groups are ‘hard’ electrophiles and react well with basic ‘hard’ nucleophiles. Alkyl halides are ‘soft’ electrophiles and react best with large, ‘soft’, uncharged nucleophiles from lower down in the periodic table. So the ketone reacts with the nitrogen, and the alkyl halide reacts with the sulfur. A possible mechanism starts with nucleophilic substitution of the alpha-haloketone; the sulfur attacks the alpha-carbon of the ketone displacing the halide. The resulting isothiourea attacks the carbonyl carbon, giving a five-membered ring. Finally, a molecule of water is eliminated and the thiazole is formed.

The thiazolamine is then brominated with N-bromosuccinimide in DMF. While bromothiazoles can be produced via direct bromination, thiazole is both less aromatic and considerably less electron rich than thiophene. As a consequence, electrophilic aromatic substitution of thiazole is much less facile. However, the rate of electrophilic aromatic substitution reactions is greatly affected by the groups attached to the ring. Activating groups that donate electron density to the ring make the aromatic ring more electron-rich and the reaction faster. Thus, activation of the thiazole ring by the presence of an amino group at C-2 and deactivation of the thiophene by the presence of the electron-withdrawing chlorine atom may justify the selectivity observed in this bromination step.

Subsequent nucleophilic aromatic substitution with cyclohexylpiperazine provides the desired amine in 34% overall yield.

Amide bond formation with dichloronicotinic acid is accomplished by activation with phosphorus oxychloride to give the corresponding nicotinamide in 83% yield.

A second nucleophilic aromatic substitution with the piperidine, followed by hydrolysis, and salt formation using maleic acid gives avatrombopag maleate in 85% yield.

Due to its electron-poor nature, nucleophilic aromatic substitution (SNAr) on pyridine is relatively easy (faster than benzene but slower than pyrimidine). The addition is facilitated by the electron-deficiency at the alpha- and gamma-carbons, which is further increased by the inductive withdrawal of electrons by the halogen substituent. The ability of the heteroatom to accommodate negative charge in the intermediates produced (Meisenheimer complexes) by inductive and mesomeric effects favor the SNAr at the 2,4, and 6 positions over C-3 and C-5.

Reference work:

J. Med. Chem. 2020, 63, 10652.