Synthesis of three series of derivatives of terfenadone, dehydroterfenadone and ebastine
The choice of terfenadone, 1
, and ebastine as starting points for the design of high-affinity inhibitors of CYP2J2 was based on: (i) the high regioselectivity of the CYP2J2-catalyzed hydroxylations of 1
and ebastine, in favor of the least reactive part of these substrates (), which implies their strict positioning in the CYP2J2 active site to keep their t-butyl group in close proximity of the heme iron for transfer of an oxygen atom from O2
, and (ii) the high affinity of 1
and ebastine for CYP2J2, as indicated by the IC50
value of 1
for CYP2J2 inhibition (0.7 μM) [28
] and the K
m of CYP2J2-catalyzed hydroxylation of ebastine (1 μM) [24
Sixteen derivatives of terfenadone were synthesized and compared to terfenadone and the drug terfenadine, as CYP2J2 inhibitors (). Most of them derive from terfenadone by replacement of its t-butyl group with various R
groups of different size and polarity. This includes R
groups bearing chemical functions well known to lead to suicide inactivation of cytochrome P450 after in situ
]. This is the case of the terminal double bond of compound 5
, of the CHF2
function of compound 12
, and of the benzo-1,3-dioxole function of compound 13
(). The structure of compounds 5
was chosen so that the CYP2J2-catalyzed hydroxylation occurs at the site leading to inactivating metabolites, assuming that hydroxylation of these compounds should occur on the homobenzylic position, as the hydroxylation of terfenadone [28
], terfenadine [27
] and ebastine [24
]. The general synthetic route used for the preparation of the terfenadone and ebastine derivatives has been described previously [28
]; it is recalled in . Compounds 16
, in which the keto group of 1
was replaced with a CH2
group, were also synthesized to evaluate the importance of this keto group in CYP2J2 inhibition. Compounds 5
were obtained from reaction of 11
with allytributyltin and imidazole, respectively. In the ebastine series, compounds 20
were prepared from reactions of the methanesulfonate of 19
with sodium ethanolate and sodium thiomethoxide, respectively (see Materials and Methods). Finally, compounds 22
, that are derived from dehydro-terfenadone, were obtained by treatment of compounds 1
and terfenadine by HCl in boiling water. All compounds were characterized by 1
H NMR spectroscopy and mass spectrometry. 1
H NMR spectroscopy in the presence of an internal standard showed that all these compounds were more than 95 % pure.
Comparison of the inhibitory effects of the terfenadone and ebastine derivatives towards CYP2J2
The fourth columns of , and compare the IC50 values found for the inhibition of ebastine hydroxylation catalyzed by recombinant CYP2J2 expressed in baculovirus-infected Sf9 insect cells. These IC50 values vary from 0.4 to 23 μM.
Comparison of the inhibitory effects of ebastine derivatives towards CYP2J2, 2B6, 2C8, 2C9 and 3A4.
Comparison of the inhibitory effects of dehydro-terfenadone derivatives and of compound 25 towards CYP2J2, 2B6, 2C8, 2C9 and 3A4.
The presence of a terminal hydrophobic group (Ph2CHO-, Ph2C(OH)- or Ph2C=) in ω position relative to the hydroxylation site is important for the affinity of the inhibitors. This is indicated by the IC50 value of 25, which is 19-fold higher than that of terfenadone 1 (13 and 0.7 μM, respectively). The hydrophobicity of this terminal group appears to be the most important factor for the affinity, as the removal of the tertiary alcohol function of 1 or of the ether function of ebastine resulting in 22 does not lead to any significant loss of affinity (IC50 of 0.7 and 1 μM for 1 and ebastine to be compared to an IC50 of 0.9 μM for 22).
The presence of the keto function para to the R group is also important for the affinity of the compounds. This is shown by the marked increase of the IC50 value observed after reduction of the CO function of 1 into CHOH or CH2 functions (0.7, 8 and 3.6 μM for 1, terfenadine and 16 respectively). Similar increases of IC50 by a factor of about 10 were also found when passing from 4 to 17 and from 22 to 24 ( and ).
The nature of the R
substituent has a great influence on the affinity of the inhibitors towards CYP2J2. The best results were obtained with hydrophobic C3
alkyl chains, as the lowest IC50
values were observed for compounds 4
(0.4 μM) for which R
is a propyl or an allyl group. Any increase or decrease of the chain length, as in 6
respectively, led to an increase of the IC50
value. Moreover, the introduction of polar groups into R
always led to a marked increase of the IC50
(compare for instance 7
, and 15
, and 10
). In that regard, the low IC50
value observed for compound 14
, for which R
is a quite polar imidazole group, is a particular case. Its good affinity for CYP2J2 is due to the presence of the imidazole heterocycle, which is a well-known ligand of P450 iron [52
] and should strongly bind to the iron of CYP2J2. Accordingly, addition of 14
to a suspension of microsomes from insect cells expressing CYP2J2 led, in difference visible spectroscopy, to a spectrum characterized by a peak at 432 nm and a trough at 415 nm (type II difference spectrum [57
]) (data not shown). These data confirmed that 14
binds to CYP2J2 iron through the accessible nitrogen atom of its imidazole moiety.
Selectivity of the inhibitors towards CYP2J2 by comparison with other vascular P450s
– also compare the inhibitory effects of the terfenadone and ebastine derivatives towards the other main CYPs that have been reported to be present in the blood vessels, namely CYP2B6, 2C8, 2C9 and 3A4 [58
]. The reference activities that were followed to measure these inhibitory effects were 7-benzyloxyresorufin O-debenzylase [42
], paclitaxel 6α-hydroxylase [43
], diclofenac 4′-hydroxylase [62
] and testosterone 6β-hydroxylase [47
] respectively. The substrate concentrations used in these experiments were equal to the K
m of the activity followed for each CYP (see Materials and Methods). Microsomes of the W(R)fur1 yeast strain expressing each CYP and overexpressing yeast cytochrome P450 reductase [36
] were used. We have confirmed that the IC50
values measured with these systems were very similar to those measured by using microsomes of insect cells expressing CYP 2B6, 2C8, 2C9, and 3A4 in the case of compounds 1
(less than 15 % variation of the IC50
Among all the studied compounds, the imidazole derivative 14 exhibited a particular behavior, as it acted as a good, but non selective inhibitor of all the studied CYPs, with IC50 values from 0.4 to 5.2 μM. However, it is noteworthy that it exhibited the best affinity for CYP2J2 (IC50 = 0.4 μM). Even though 14 is not a selective inhibitor for CYP2J2, it could be useful to inhibit all the vascular CYPs and, consequently, all the activities of arachidonic acid epoxidation at the vascular level.
As far as all the other compounds mentioned in – are concerned, the main conclusions drawn from those tables are the following ones:
- none of the studied compounds led to a significant inhibition of CYP2C8 (IC50 > 100 μM). In order to confirm these results, we have also studied the effects of the terfenadone and ebastine derivatives towards another usual CYP2C8 activity, the N-deethylation of amodiaquine . Both CYP2C8-catalyzed activities were not significantly inhibited by compounds 1–25, their IC50 values always being higher than 100 μM.
- CYP2C9 was inhibited with IC50 values between 10 and 69 μM, which were, in general, about 10 fold higher than those found for CYP2J2.
- the previous conclusion concerning CYP2C9 is globally valid for CYP2B6, with IC50 values varying from 7 to 90 μM, if one excludes compound 23 that appeared to be a very good inhibitor of CYP2B6. This inhibitory effect of 23 towards CYP2B6, and to a lesser extent towards CYP2J2, could be due to a strong binding of its terminal NH2 function to P450 iron.
- the studied compounds inhibited CYP3A4 with IC50 values from 0.9 to 68 μM, that were generally intermediate between those found for CYP2J2 and for CYP2C9 (or 2B6).
The best inhibitors of CYP2J2, 4 and 5, are reasonably selective towards this cytochrome as their IC50 values towards the other studied CYPs are at least 20-fold (for 4) and 14-fold (for 5) higher than those found for CYP2J2.
This preliminary study of the IC50 values of the terfenadone and ebastine derivatives allowed us to select the following compounds for further, more detailed studies. Compound 4 was chosen because of its IC50 value of 0.4 μM, which was the lowest one and of its good selectivity towards CYP2J2. The choice of compound 14 was made because of its low IC50 and because it could be considered as an inhibitor for all the main vascular CYPs. Finally, compounds 5 and 13 were selected because preliminary experiments showed us that their inhibitory effects increased as a function of the incubation time, suggesting that they could be mechanism-based inhibitors of CYP2J2.
Study of the mode of inhibition of CYP2J2 by 4
Kinetic studies of the inhibition of CYP2J2-catalyzed hydroxylation of ebastine by 4
were performed at various ebastine (0.2 to 5 μM) and 4
(0 to 2 μM) concentrations. The Lineweaver-Burk plots of the reciprocal of the reaction rate vs
the reciprocal of ebastine concentration, at different concentrations of 4
, indicated that 4
acts as a competitive inhibitor of CYP2J2 (intercept of the straight lines on the y axis [63
], data not shown). The Dixon plot of 1/v vs
the concentration of 4
led to a Ki
value for the inhibition of CYP2J2 by 4
of 0.16 ± 0.05 μM (). This value was in excellent agreement with the Ki
value (0.2 μM) that may be calculated from the IC50
value of , assuming that Ki
/2 for a competitive inhibitor [64
Dixon plots obtained from a kinetic study of CYP2J2-catalyzed hydroxylation of ebastine in the presence of different concentrations of compound 4
Actually, 4 is a substrate of CYP2J2; it is hydroxylated at the level of its propyl group, as expected from its great analogy with terfenadone (P. Lafite et al., publication in preparation). Thus, compound 4 is a competitive inhibitor and an alternative substrate of CYP2J2. We have confirmed that the concentration of 4 did not vary in a significant manner (less than 10 % consumption) under the conditions used in kinetic experiments of study of its inhibitory effects towards CYP2J2-catalyzed hydroxylation of ebastine (very low CYP2J2 concentration of 1 nM and short incubation times, 2–4 min).
Characterization of the inhibition of CYP2J2 by 14
In order to analyze the type of inhibition of CYP2J2 by 14
, the Lineweaver-Burk plots of the reciprocal of the reaction rate vs
the reciprocal of ebastine concentration were drawn at different 14
concentrations. shows that 14
acts as a mixed-type inhibitor of CYP2J2, with a competitive and a non-competitive component [65
]. The Dixon plot drawn for the inhibition of CYP2J2-catalyzed hydroxylation of ebastine by 14
led to an evaluation of the competitive and the non-competitive inhibition constants
, respectively (data not shown). The values deduced from the Dixon plot,
, suggest that 14
is a mixed-type inhibitor, with a strong competitive component, as
is 10-fold lower than
Lineweaver-Burk plots obtained from a kinetic study of CYP2J2-catalyzed hydroxylation of ebastine in the presence of different concentrations of compound 14
Mechanism of inhibition of CYP2J2 by compounds 5 and 13
To further analyze the time-dependent variation of the inhibitory effects of 5 and 13, insect microsomes containing recombinant CYP2J2 were preincubated with 5 or 13 in the presence or absence of NADPH (i.e. under catalysis or non catalysis conditions), and the remaining enzymatic activity was measured using ebastine as substrate, as a function of the preincubation time.
Incubation of microsomes with 5 in the presence of NADPH led to a progressive loss of CYP2J2 activity as a function of time (). With 20 μM 5, 50 % of the activity was lost after 7 min and only 20 % remained after 20 min. Loss of CYP2J2 activity was faster after incubation with identical concentrations of 13, as 50 % of the activity was lost after 1.5 min and only 20 % remained after 4 min in the presence of 20 μM 13 (). Incubations under identical conditions but in the absence of NADPH did not lead to any significant loss of CYP2J2 activity. In the absence of either 5 or 13, less than 10 % of the activity was lost after 10 min (data not shown). These results confirmed the existence of a catalysis-dependent inactivation of CYP2J2 upon oxidation of 5 and 13.
Kinetics of inactivation of CYP2J2 upon NADPH-dependent oxidation of compounds 5 (A) and 13 (B)
Kinetics of CYP2J2 inactivation by 5 and 13
shows that the loss of CYP2J2 activity as a function of time after incubation in the presence of NADPH and various 5
concentrations followed the classical kinetics previously described for other CYP suicide-substrates [54
]. The time required for half-maximal inactivation, t1/2
, and the apparent first-order constant, kinact
, were calculated from the logarithmic transformation of the remaining activity as a function of time, as depicted in . Plots of t1/2
versus reciprocal 5
concentration (, inset) led to the kinetic constants of the inactivation process (see Materials and Methods). From extrapolation to infinite 5
concentration, the time required to inactivate half of the enzyme, at the maximal rate, t1/2max
, and the maximal kinact
were 8.7 ± 2.2 min and 0.08 ± 0.02 min−1
, respectively (). The inhibition constant, KI
, was found to be 0.45 ± 0.05 μM, and the second-order rate constant kinact
, a proper index of the in vitro
effectiveness of a compound as inactivator [66
], was found to be 2960 ± 1000 L.mol−1
Comparison of the kinetic parameters calculated for inactivation of CYP2J2 by compounds 5 and 13 (this work) with those published for other suicide substrates of CYPs.
Identical experiments were performed in the case of compound 13
() and led to the kinetic constants reported in . The efficiencies of 5
as inactivators of CYP2J2 are similar, if one compares their kinact
values (2960 and 2700 L.mol−1
, respectively). In fact, the inactivation rate of CYP2J2 by 13
is 6-times higher than that found in the case of 5
(0.47 ± 0.05 instead of 0.08 ± 0.02 min−1
). At the opposite, the affinity of 5
for CYP2J2 seems to be higher than that of 13
, as suggested by the KI
(0.45 ± 0.05 and 2.9 ± 0.2 μM respectively, ) and IC50
values (0.4 ± 0.1 and 6.7 ± 2 μM, respectively, ) found for these compounds. also compares kinetic constants reported for other CYP mechanism-based inactivators [56
]. It indicates that 5
are reasonably efficient mechanism-based inactivators of CYP2J2, as judged from their kinact
Study of the molecular mechanism responsible for CYP2J2 inactivation by 13
has been designed to be a mechanism-based inhibitor of CYP2J2, as we expected that its hydroxylation by this enzyme would mainly occur at the level of its benzodioxole CH2
group. Such P450-dependent oxidations of benzo-1,3-dioxole derivatives are well known to lead to the formation of the corresponding catechol metabolites and to P450 inactivation due to the formation of very stable P450 iron-benzodioxole-derived carbene complexes [52
] (). The generally admitted mechanism of these reactions involves the free-radical abstraction of an hydrogen atom of the benzodioxole CH2
group by the high-valent P450 iron-oxo active species. The resulting radical may either undergo an oxidative transfer of the OH ligand of the P450 Fe(IV)-OH intermediate, with formation of an unstable orthoformiate that is eventually hydrolyzed to the corresponding catechol, or bind to P450 iron leading to a very stable P450 iron-carbene complex after elimination of H2
Possible mechanisms of the CYP2J2-catalyzed oxidation of 13 and of the inactivation of this cytochrome.
Addition of NADPH to microsomes of insect cells expressing CYP2J2 containing 100 μM 13
led to the progressive appearance of a difference visible spectrum characterized by two peaks at 428 and 457 nm (). This difference spectrum is characteristic of the formation of a P450 iron-benzodioxole-derived carbene complex [52
]. Under the used conditions (0.1 μM CYP2J2, 100 μM 13
), the difference spectrum reached its maximal intensity 10 min after the addition of 100μM NADPH. If one considers the ε455–490
nm values reported in the literature for the difference spectra of these P450 iron-carbene complexes, which vary from 50 000 to 75 000 M.cm−1
], one may estimate that 66 to 99 % of CYP2J2 is engaged in an iron-carbene complex derived from 13
Difference absorption spectra observed during CYP2J2-catalyzed oxidation of 13 in the presence of NADPH
Reaction mixtures from incubation of 13 with microsomes from insect cells expressing CYP2J2 in the presence of NADPH were studied by HPLC coupled to mass spectrometry. The major metabolite detected by this method exhibited a mass spectrum characterized by a molecular ion at m/z = 446 (M-12 if M is the molecular ion of 13). This ion well corresponded to the one expected for the catechol derived from 13. Moreover, the main fragments appearing in the mass spectrum of the metabolite were also in complete agreement with what could be expected for the catechol derived from 13. Thus, most fragments exhibited m/z values equal to those of 13 minus 12, except the fragments that do not contain the benzodioxole moiety which exhibited m/z values identical to chose of the corresponding fragments of 13 (). Finally, comparison of the UV spectra of 13 and of its main metabolite showed very similar characteristics with two peaks at 280 and 305–310 nm, and a small blueshift of the 305–310 nm peak for the metabolite of 13. A comparison of the UV spectra of analog compounds, 3′,4′-(methylenedioxy)-propiophenone and the corresponding catechol, led to very similar characteristics (data not shown).
Mass spectra of compound 13 and its metabolite formed after oxidation by CYP2J2
The aforementioned data confirmed that CYP2J2-catalyzed oxidation of 13 mainly occurred at the benzodioxole CH2 group with formation of the corresponding catechol metabolite and of a CYP2J2 iron-carbene complex that leads to the inactivation of this cytochrome.
Further studies of the characteristics of CYP2J2 inactivation by 13
shows that the loss of CYP2J2 activity upon incubation with 2 μM 13
for increasing times paralleled catechol metabolite formation, as expected for mechanism-based inhibition [48
]. illustrates the correlation between the CYP2J2 activity remaining after oxidation of 13
and the amount of catechol metabolite formed for various concentrations of 13
and incubation times. The linear relationship observed allowed us to estimate the partition ratio of the inactivation process, r
, which represents the number of productive turnovers (leading to the catechol metabolite) divided by the number of inactivating events [48
]. Actually, extrapolation to 0 % remaining activity in led to a r
value of 18 ± 3.
Relationship between CYP2J2 inactivation and efficient catalysis of 13 oxidation (catechol metabolite formation) (A) and determination of the partition ratio (B)
The presence of 4
, a good competitive inhibitor and alternative substrate of CYP2J2, in incubations of CYP2J2 with 13
and NADPH led to a clear decrease of the rate of inactivation of CYP2J2 (see figure in Supplementary Materials
An important property of efficient mechanism-based inhibitors is to generate reactive species that will rapidly react within the active site rather than diffuse out into solution. The presence of 5 mM reduced glutathione in incubations of CYP2J2 with 13
and NADPH did not have any significant effects on the rate of CYP2J2 inactivation (see figure in Supplementary Materials
). This result indicates that the reactive intermediate of 13
, which is responsible for CYP2J2 inactivation is not an electrophilic metabolite released in the medium. It is in agreement with the mechanism shown in , in which the free radical intermediate from 13
is rapidly trapped by CYP2J2 iron, with the eventual formation of the iron-carbene complex, and is not released in the medium.