Hydroxylation of the drug ebastine by recombinant CYP2J2⁷ expressed in baculovirus-infected Sf9 insect cells³ was used as an assay to find CYP2J2 inhibitors. During a first screening for such inhibitors, compound 1, derived from the drug terfenadine by oxidation of its benzylic alcohol function, was found to inhibit CYP2J2¹¹ with an IC₅₀ value of 0.7 ± 0.1 μM. This value was much lower than the IC₅₀ found for terfenadine itself (8.1 ± 0.4 μM) (Table 1), that was previously described as an inhibitor of CYP2J2.¹² Then, compound 1, called terfenadone in the following, was used as a starting point for the design of high-affinity inhibitors of CYP2J2. Ebastine,⁷ terfenadine,¹³ and compound 1 are all hydroxylated by CYP2J2 at a site that is weakly reactive from a chemical standpoint, a CH₃ of the t-butyl group (Scheme 1). This regioselectivity in favor of the least reactive part of these substrates implies their strict positioning in the CYP2J2 active site in order to maintain the t-butyl group in close proximity of the heme iron for transfer of an oxygen atom from O₂. Therefore, a series of compounds derived from 1 by replacement of its t-butyl group with various R groups of different size and polarity was synthesized (Scheme 2). This included compounds bearing functions previously known to lead to suicide inactivation of cytochromes P450 after in situ oxidation.¹⁴ This is the case for the terminal double bond of compound 5, since terminal alkenes act as mechanism-based inhibitors of cytochromes P450 after N-alkylation of the heme by an intermediate derived from P450-catalyzed oxidation of the substrate double bond.¹⁴ The choice of the CHF₂ and benzodioxole functions of compounds 12 and 13 was also made on the basis of literature data on suicide substrates of cytochrome P450.¹⁴ In situ hydroxylation of the C–H bond of CHF₂ groups leads to an electrophilic intermediate able to acylate the P450 protein, whereas inactivation of cytochrome P450 by benzodioxole derivatives is due to the formation of an iron–carbene bond after oxidation of the benzodioxole CH₂ group.¹⁴ The structure of compounds 5, 12, and 13 has been chosen in order that CYP2J2-catalyzed oxidations occur at the site leading to inactivating metabolites, assuming that hydroxylation of 5, 12, and 13 should occur on the homobenzylic position as the hydroxylation of compound 1 and terfenadine.

The general synthetic route used for the preparation of terfenadone derivatives (Scheme 2) with R = (CH₂)_nCH₃ (with n = 0–3), (CH₂)_nOH and (CH₂)_nOCOCH₃ (with n = 2 or 3), Br, CH₂CHF₂, OCH₂O–, and OCH₃ involved an acylation of the benzenic starting compound with 4-chlorobutanoic acid chloride in the presence of FeCl₃, AlCl₃ or SnCl₄ and reaction of the corresponding product with α,α-diphenyl-4-piperidinomethanol. In the case of 7 and 8, the starting compounds were the acetates of 2-phenylethanol and 3-phenylpropanol, respectively. Deprotection of the alcohol function was done as the last step of the synthesis. Compound 5 was obtained from reaction of 11 with allyltributyltin, in the presence of tetrakis(triphenylphosphine)palladium(0).

The structures of all the terfenadone derivatives listed in Table 1 were completely established from their ¹H NMR and mass spectra; ¹H NMR spectroscopy analysis in the presence of an internal standard showed that all these compounds were more than 95% pure.

Table 1 compares the IC₅₀ values measured for the inhibition of ebastine hydroxylation catalyzed by recombinant CYP2J2. It shows that most of the synthesized terfenadone derivatives are good CYP2J2 inhibitors with IC₅₀ values at the low μM range. Compounds 4 and 5 had the highest affinity with an IC₅₀ value of 0.4 μM. In fact, increasing the chain length from R = methyl to R = propyl results in a gradual decrease of the IC₅₀ value, whereas a further increase of the chain length (R = butyl) leads to a loss of affinity. Introduction of a polar function in the R substituent generally leads to a decrease in the affinity of the inhibitors. Compounds such as 10 and 13 in which oxygen atoms have been introduced at benzylic positions exhibit IC₅₀ values one order of magnitude greater than those observed for compounds bearing an alkyl chain (R = Et or Pr, 3 or 4 for instance). Compounds such as 7, 8, 9, and 12 in which an OH, OAc or F substituent have been introduced in the R-chain farther from the phenyl ring exhibit intermediate IC₅₀ values, around 2 μM. Thus, the best inhibitors (in terms of IC₅₀ value) were compounds 4 and 5. Preliminary experiments showed that compound 4 is a competitive inhibitor of CYP2J2-catalyzed hydroxylation of ebastine with a K_i of 160 ± 30 nM and also a competitive substrate of CYP2J2. Compound 5 seems to be a time-dependent inhibitor, as expected for a compound bearing a terminal double bond.¹⁴ Interestingly, compounds 12 and 13 involving a CHF₂ and benzodioxole function, respectively, also led to time-dependent inhibitory effects that suggest a mechanism-based type of inhibition.

Table 2 compares the inhibitory effects of the best inhibitors found for CYP2J2, compounds 4 and 5, toward the other main human cytochromes P450 that are present in the cardiovascular system, CYP2C8, CYP2C9, CYP2B6, and CYP3A4.²¹ The data clearly show that compounds 4 and 5 are selective inhibitors of CYP2J2, as they are nearly inactive toward CYP2C8 and their IC₅₀ values for CYP2C9, CYP2B6, and CYP3A4 are 1–3 orders of magnitude higher than those observed for CYP2J2.