Fluorinated Steroids: A Survey of Recent Synthetic Methods for Direct Fluorination of Steroids

Selected examples of halogenated corticosteroids.

            Most of us are familiar with drug products such as Flonase®, Flovent and Advair, which are used for the management of asthma and chronic obstructive pulmonary disease (COPD). In 2013, Advair racked up over $4.8 billion dollars in sales. Flonase and Advair contain a 6,9-a-difluorinated corticosteroid derivative called fluticasone propionate (see scheme below, structure 8), which is responsible for inducing the potent anti-inflammatory effects of the drugs. The 9a-halogenated glucocorticosteroid series was first discovered in 1953 by Fried and Sabo (J. Am. Chem. Soc. 1953, 75, 2273), who noted that anti-inflammatory bioactivity was inversely correlated with the size of the halogen atom substituted at C9. The activity trend in the C9 series was: I < Br < Cl < F, with the 9a-fluoro derivatives (representative examples shown above) exhibiting a >10-fold increase in binding affinity to glucocorticoid receptors compared to the parent hormones. The 9a-fluoro substituent also impedes oxidation of the proximal 11-hydroxy group, providing increased duration of action. Incorporation of a 6a-fluoro prevents hydroxylation at this position. Interestingly, the rarely encountered fluorinated 20-thioester moiety of fluticasone propionate was incorporated to exploit a hepatic inactivation mechanism that affords its corresponding carboxylic acid. The lack of bioactivity associated with this carboxylic acid metabolite results in vastly reduced undesired systemic exposure of the parent drug.
            It is now well known that the fluorine atom’s electronegativity, size, omniphobicity/lipophilicity and electrostatic interactions can dramatically influence the binding affinity of a fluorinated ligand for its biological protein target (for an excellent medicinal chemistry resource, see this). Indeed, a single substitution of a fluorine atom in exchange for hydrogen can completely change the biological properties of a small molecule. In fact, the so-called ‘fluorine scan’ is currently a routine approach used in the development of a drug candidate.

            As described in a recent patent by Hovione Ltd., fluticasone propionate is manufactured by means of a rather lengthy semisynthetic derivatization process, wherein fluorine atoms are sequentially introduced by functional group interconversion reactions. For example, the 6a-fluoro substituent is installed by electrophilic fluorination of a dienol benzoate intermediate (2 à 3), itself derived from prednisone acetate. The embedded 9a-fluoro is then incorporated by subsequent ring-opening of a 9,11-oxirane ring with aqueous hydrofluoric acid. Finally, a 2014 Organic Process Research and Development report disclosed an effective method for introduction of the fluorinated 20-thioester moiety involving decarboxylative fluorination of a carboxylic acid intermediate (7 à 8) using silver nitrate and Selectfluor. The reaction to produce fluticasone propionate was demonstrated on 95-gram scale.

            The overall API manufacturing process for fluticasone propionate production is generally quite efficient, but somewhat linear in nature. A more concise synthetic entry into the halogenated corticosteroid structural framework is therefore highly sought after. Unfortunately, Nature does not produce fluorinated steroids and so enzymatic biotransformation, in this case, does not represent a viable solution. Synthetic methods for direct fluorination of the steroid nucleus, with ‘enzyme-like’ precision over regio- and stereoselectivity, would provide great utility for the efficient production of clinically relevant halogenated glucocorticoids such as fluticasone. A brief survey of recently reported fluorination protocols that do not require extensive pre-functionalization of steroidal substrates (i.e. ‘direct’ fluorination methods) is provided below.

            In 2012, the laboratory of John T. Groves at Princeton University reported a cytochrome P450-inspired method for oxidative C-H fluorination of cycloalkanes that is catalyzed by a unique manganese porphyrin complex, in combination with stoichiometric additives. The authors applied their technology to the steroid 5a-androstan-17-one, which contains no less than 28 unactivated sp³ C-H bonds. Presumably due to electronic deactivation of the D ring, in conjunction with steric hindrance of rings B and C, only the C2 and C3 positions on the A ring were fluorinated with good conversion and high stereoselectivity. The diastereoselectivity is likely the result of steric shielding of the b-face by the axial C19 methyl substituent. Mechanistic analysis of the transformation suggested that C-H bond cleavage of the cycloalkane is mediated by an oxomanganese(V) catalytic intermediate [oxo-Mn(V), structure shown above], produced by oxidation of the pre-catalyst, manganese(III)TMP-chloride [Mn(TMP)Cl]. The oxo-Mn(V) species abstracts a hydrogen atom from the substrate to produce a transient carbon-centered radical. The radical is then trapped by F delivery from an unusual trans-difluoromanganese(IV)TMP complex, culminating in hydrocarbon fluorination with extrusion of a manganese(III) intermediate back into the catalytic cycle. The putative fluorinating agent, Mn(IV)(TMP)F₂, was isolated and structurally characterized by X-ray crystallography. As described in a 2013 Nature Protocols manuscript, Groves’ C-H fluorination procedure requires only a typical laboratory fume hood and ordinary glassware.

            The direct regio- and stereocontrolled a-fluorination of allo-pregnanedione (shown above) was demonstrated in 2011 by MacMillan’s group. A primary amine-functionalized Cinchona alkaloid organocatalyst selectively activates the ketone carbonyl of the substrate’s A-ring, giving rise to a conformationally restricted enamine, which succumbs to electrophilic fluorination in a stereocontrolled fashion. Catalyst controlled regio- and stereoselectivity are particularly impressive in this instance given that allo-pregnanedione contains two ketones and three a-methylene sites. The same methodology was also applied to the steroid cholestanone and, again, fluorination was achieved with complete regiocontrol and outstanding efficiency.

            Ritter’s group at Harvard has developed an operationally simple ipso deoxyfluorination reaction that provides expedient access to 3-fluoro-estrone in a single operation, without the need for pre-functionalization. The reaction proceeds via the intermediacy of a 2-aryloxyimidazolium bifluoride salt (bracketed intermediate above), formed by condensation of the estrone A-ring phenol with Ritter’s stoichiometric fluorinating reagent. Proximity-induced nucleophilic fluorination by hydrogen bonded fluoride then produces the steroidal fluoro-arene product. Late-stage steroid fluorination techniques such as this one could potentially provide access to ¹⁸F-radiolabelled fluorosteroids which have shown promise as radiotracer imaging agents for cancer.

            While none of the protocols described above will immediately revolutionize the manner in which halogenated corticosteroids are manufactured, this general line of research, which has been initiated with some very exiting preliminary results, may one day lead to more step-economical syntheses of important APIs such as fluticasone propionate.