Synthesis of new potent and selective aromatase inhibitors based on long-chained diarylalkylimidazole and diarylalkyltriazole molecule skeletons

Article abstract of DOI:10.1016/S0928-0987(00)00074-9

A series of long-chained diarylalkylimidazoles and diarylalkyltriazoles were synthesized and evaluated for the inhibitory potency for aromatase (estrogen synthetase) activity in human placental microsomes. The relative specificity of inhibition was evaluated by measuring the inhibition of cholesterol side-chain cleavage enzyme (desmolase) in human placental mitochondria and the inhibition of 7-ethoxycoumarin O-deethylase (a typical drug-metabolizing enzyme activity) in rat liver microsomes. The structural requirements including substituent effects for the strongest potency and for the highest specificity were delineated. α,ω-Diarylalkyltriazoles and imidazoles were the most interesting molecules, in which the geometric and optical isomerism displayed remarkable selectivity for aromatase inhibition. Copyright (C) 2000 Elsevier Science B.V.

Full text of DOI:10.1016/S0928-0987(00)00074-9

European Journal of Pharmaceutical Sciences 11 (2000) 109–131
www.elsevier.nl/locate/ejps
Synthesis of new potent and selective aromatase inhibitors based on long-
chained diarylalkylimidazole and diarylalkyltriazole molecule skeletons
a,1
a,1
b
a,2
Arto Karjalainena , , Arja Kalapudas , Marja Sodervall , Olavi Pelkonen , Risto Lammintausta
*
¨
aDepartment of Synthetic Chemistry, Orion Corporation, Orion Pharma, R&D, P.O. Box 65, Espoo, Finland
bDepartment of Pharmacology and Toxicology, University of Oulu, PL 5000, FIN-90401 Oulu, Finland
Received 1 July 1999; received in revised form 15 December 1999; accepted 18 January 2000
Abstract
A series of long-chained diarylalkylimidazoles and diarylalkyltriazoles were synthesized and evaluated for the inhibitory potency for
aromatase (estrogen synthetase) activity in human placental microsomes. The relative specificity of inhibition was evaluated by measuring
the inhibition of cholesterol side-chain cleavage enzyme (desmolase) in human placental mitochondria and the inhibition of 7-
ethoxycoumarin O-deethylase (a typical drug-metabolizing enzyme activity) in rat liver microsomes. The structural requirements
including substituent effects for the strongest potency and for the highest specificity were delineated. a,v-Diarylalkyltriazoles and
imidazoles were the most interesting molecules, in which the geometric and optical isomerism displayed remarkable selectivity for
aromatase inhibition.
2000 Elsevier Science B.V. All rights reserved.
Keywords: Aromatase (inhibitors); Breast cancer; Desmolase; Cholesterol side chain cleavage
1. Introduction
steroid hormone-synthesizing pathways. For example,
glucocorticoids and mineralocorticoids in the adrenal
glands are synthesized through pathways in which par-
ticular P450 enzymes participate as rate-limiting steps. Of
particular importance in all steroidogenic tissues is choles-
terol side-chain cleavage step catalyzed by P450 SCC or
desmolase (CYP11A1), because cholesterol is a precursor
for all steroid hormones (Miller, 1988). If desmolase is
inhibited, it leads to a decrease of production of all steroid
hormones, including gluco- and mineralocorticoids, with
all consequent side-effects and problems. The probability
of inhibition of different P450 enzymes by any particular
inhibitor is rather high, because different P450 enzymes
and especially their active sites show considerable homol-
ogy; their structures resemble each other to a considerable
extent. This relative similarity of all P450 enzymes is
consequently the principal source of potential problems
with aromatase inhibitors. A molecular model of aromatase
based on structures of bacterial P450s has been published
(Graham-Lorence et al., 1995), but its stability to predict
affinities and turnover numbers of potential substrates and
inhibitors has not been surveyed.
Aromatase inhibitors (Bhatnagar et al., 1990a; Vanden
Bossche et al., 1994) are potentially useful in treating
some estrogen-dependent cancers, such as breast cancer,
because they inhibit the enzyme, aromatase, that catalyses
the last step in the pathway to the formation of hormonally
active estrogens (estradiol and others) from androgenic
precursors (such as androstenedione or testosterone).
Aromatase inhibition is the prerequisite for the therapeutic
effect and according to the current opinion estrogen
synthesis must be inhibited in all tissues capable of the
synthesis (Goss and Gwyn, 1994).
Aromatase (CYP19) belongs to a group of cytochrome
P450 (CYP) enzymes, named so because the terminal
oxygen activator in the enzyme complex is the heme
protein, cytochrome P450 (Simpson et al., 1994). The
P450 enzymes are important catalysts in many other
*Corresponding author. Tel.: 1358-9-4292-963; fax: 1358-9-4293-
600.
E-mail address: arto.karjalainen@orion.fi (A. Karjalainen)
1Present address: Hormos Medical Ltd., University of Oulu, Department
of Chemistry, PL 3000, FIN-90401 Oulu, Finland.
2Present address: Hormos Medical Ltd., BioCity, Turku, FIN-20520
Turku, Finland.
Perhaps the most important step in which aromatase
inhibitors may cause inadvertent, not-sought-for inhibition,
is cholesterol side-chain cleavage. Naturally, the goal is to
synthesize inhibitors that display as much specificity as
0928-0987/00/$ – see front matter
PII: S0928-0987(00)00074-9
2000 Elsevier Science B.V. All rights reserved.

110
A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
The synthetic program was devised to be based on the
structure–activity relationships, which encompassed the
screening of our compounds by both aromatase and
cholesterol side-chain cleavage assays and in the selected
cases by 7-ethoxycoumarin O-deethylase assay represent-
ing a xenobiotic-metabolizing P450 enzyme.
The expectation that the long-chained 4(5)-
arylalkylimidazoles might be aromatase inhibitors, proved
correct. Compound (1) was a moderate inhibitor.
A new interesting lead was then provided by hybridizing
the structural features of Eli Lilly’s compounds and of
compound (1). The diarylmethylene moiety connected into
imidazole by a lengthened carbon chain led to rather potent
and selective compounds (Karjalainen et al., 1992). For
example, 4(5)-(4,4-diphenylbutyl)-1H-imidazole (2) (see
Fig. 1. Eli Lilly’s substances displaying aromatase inhibition.
possible in terms of different P450 enzymes; i.e., selectivi-
ty is one of the most important goals in attempts to create
new aromatase inhibitors. As an example, amino-
glutethimide (Fig. 3) (Jones et al., 1990), the only widely
used aromatase inhibitor thus far, displays equal inhibitory
potencies towards aromatase and desmolase and actually
produces harmful consequences based on the inhibition of
all other steroid-synthesizing pathways (Santen et al.,
1990). Our project for the development of potent and
selective aromatase inhibitors started in 1986 after the
observation that Eli Lilly (Hirsch and Taylor, 1984) had
filed a patent application including 4,49-difluoro- and 4,49-
dichloro-substituted 4(5)-(a,a-diphenylmethyl)-imidazoles
covered by our patent (Karjalainen and Kurkela, 1981)
(Fig. 1). Eli Lilly’s research team had observed that these
structures had yet another interesting pharmacological
effect, in addition to the a-receptor effects we had
observed with the derivatives: they were aromatase in-
hibitors. The intended indication for the use of these
compounds was the treatment of breast cancer.
Fig. 2) was a relatively potent aromatase inhibitor (IC50
2
mM), but less potent in inhibiting desmolase (IC50 between
66 and 320 mM). We prepared skeletal modifications, in
which the other aryl group was at various positions along
the hydrocarbon bridge (Karjalainen et al., 1995a,b, 1996).
In this study we found that in general these kind of
structures had activity on aromatase inhibition and the
substituents in the phenyl rings had influence on potency
and selectivity. Also, the double bond in the carbon bridge
of the a,v-diarylalkyl derivatives brought additional bene-
ficial effects especially in terms of selectivity.
An extensive screening program over these long-chained
diarylalkyl substances was carried out later on with 2- and
1-substituted imidazoles and then logically with triazoles
(Karjalainen et al., 1997), because the attachment of a
short diarylmethylene group in the 1-position in imidazole
and triazole had earlier been successful in producing potent
and selective aromatase inhibitors as well. Examples of
these are Ciba Geigy’s CGS-18320B and Letrozole (CGS
20267) (Fig. 3) (Bhatnagar et al., 1990b).
Another interesting and stimulating point was that
previously we had also synthesized long-chained 4(5)-
arylalkylimidazoles,
for
example
4(5)-[4-(2,6-di-
methylphenyl)butyl]-1H-imidazole (1) (see Fig. 2) (Kar-
jalainen and Kurkela, 1986), which we had found to have
antimycotic and antibacterial properties. Because these
substances and those of Eli Lilly (Fig. l) resembled
structurally ordinary antimycotic imidazole derivatives,
which inhibit P450 enzyme participating in ergosterol
synthesis in yeast, we hypothesized that these long-chained
compounds might inhibit aromatase as well.
Other known imidazole or triazole aromatase inhibitors
are Ciba Geigy’s Fadrozole (CGS 16949A) (Browne et al.,
1991), Jansen’s Vorozole (Vanden Bossche et al., 1994)
and Zeneca’s Anastrozole (Dukes et al., 1996) (Fig. 3).
It was found earlier, for example by Ciba-Geigy and by
Zeneca, that the substituents such as F or CN were
favourable as for the activity. Substitution of our long-
chained diarylalkylimidazoles and diarylalkyltriazoles with
the fluoro and cyano substituents gave similar results. One
of the most interesting members of these series was the
diastereomeric triazole 4-[3-(4-fluorophenyl)-2-hydroxy-1-
[1,2,4]triazol-1-ylpropyl]benzonitrile, finrozole (MPV-2213
ad) (3) (see Fig. 4), in which the optical isomerism
achieved by introducing a hydroxyl group into the alkyl
chain brought conspicuous additional effect on selectivity.
In this report we give a summary of the syntheses and
the results of the structure–activity screening of these
long-chained 4(5)-diarylalkyl imidazoles (I and II in Fig.
4) and 1-diarylalkylimidazoles and triazoles (III in Fig. 4)
leading to new groups of selective aromatase inhibitors, of
which the most interesting representatives are the com-
Fig. 2. Chemical structures of two early lead compounds.

A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
111
Fig. 3. Chemical structures of aminoglutethimide and known imidazole or triazole derivatives of aromatase inhibitors.
Fig. 4. Chemical structures of three main series of synthesised imidazole and triazole compounds and those of MPV-2213 ad and MPV-1837 AV b, which
demonstrate a high degree of potency and selectivity with respect to aromatase inhibition.

112
A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
pounds MPV-2213 ad (3) and 4-[4-(4-fluorophenyl)-1-
(1H-imidazol-4-yl)-but-1-enyl]benzonitrile (MPV-1837 A
V b) (4) (Fig. 4).
synthesis of the starting materials, used especially for the
preparation of 4(5)-(v,v-diarylalkyl)-1H-imidazoles with a
four-carbon chain bridge, was carried out as depicted in
Scheme B.
In this method a benzophenone, by applying the Re-
formatsky reaction, was allowed to react with ethyl
bromoacetate to yield the hydroxyl ester B I. The ethyl
ester B II could be achieved from this by dehydration and
hydrogenation. The reduction of the ester group to alcohol
and the replacement of hydroxyl with bromine produced
3,3-diarylpropylbromide B III.
2. Chemistry
2.1. Syntheses
2.1.1. Synthesis of 4(5)-substituted imidazoles
In the course of our earlier projects, in which we
developed new selective a2-agonists (e.g., detomidine,
medetomidine and dexmedetomidine) (Karjalainen, 1981;
Karjalainen et al., 1982, 1989; Karjalainen and Kurkela,
1985) we synthesized a lot of variously substituted 4(5)-
arylalkyl-imidazoles. In that connection we developed a
new useful method to produce the 4(5)-substituted aryl
alkyl imidazole skeleton starting from 4(5)-imidazole
carbaldehyde by allowing it to react in the Grignard
reaction with an arylmagnesiumbromide or with an
arylalkylmagnesiumbromide in tetrahydrofuran at elevated
temperature (Karjalainen, 1981). Because of the difficulties
encountered in producing 4(5)-imidazole carbaldehyde in
large quantities it was found more favourable to use the
protected imidazole, 1-benzyl-5-imidazole carbaldehyde as
the starting reagent (Parhi, 1985).
Also in this study 1-benzyl-5-imidazole carbaldehyde
was a useful starting material in the synthesis of 4(5)-
substituted diarylalkylimidazoles by various methods (Kar-
jalainen et al., 1992, 1995a,b, 1996, 1997). In the follow-
ing chapters we present the main lines of the used
methodology. Our referred patents give detailed descrip-
tion of the syntheses with the analytical data and in the
experimental part of this report we have included only
some typical examples from these.
The 4(5)-(v,v-diarylalkyl)-1H-imidazoles, in which the
substituents in the aromatic rings are the same, were
synthesized easily according to the Scheme C. 1-Benzyl-5-
imidazole carbaldehyde was allowed firstly to react with
malonic acid in pyridine (Doebner modification of the
Knoevenagel reaction) to 5-(1-benzylimidazole)-acrylic
acid C I. The N-benzyl group of imidazole was removed
and the double bond in the alkyl chain was reduced
simultaneously by catalytic hydrogenation in acidic con-
ditions and the propionic acid intermediate was esterified
to achieve 4(5)-imidazole propionic acid ethyl ester C II.
Alternatively the hydrogenation of C I was carried out in
neutral conditions to preserve the protection and thus to
produce 1-benzyl-5-imidazole propionic acid ethyl ester C
V. The compounds C II or C V reacted further with aryl
magnesium bromide to afford (3,3-diaryl-3-hydroxy-
propyl)-imidazoles C III or C VI, respectively. The end
products C VIII could be achieved by a sequence of
reactions involving dehydration, debenzylation by hydro-
genation over palladium as well as reduction of the double
bond as described previously.
To produce symmetrically substituted diaryl derivatives
with a 4-alkyl carbon chain bridge, the side chain exten-
sion with an appropriate carboxylic functionality in the end
of the chain was needed. This was achieved through a
series of conventional reactions as outlined in Scheme D.
These reactions comprised the reduction of the ester
group of the intermediate C V to the alcohol with lithium
aluminum hydride followed by halogenation. The substitu-
tion of the halogen was then accomplished through the
reaction with sodium cyanide and the carboxylic ester
functionality was introduced through the hydrolysis fol-
lowed by esterification.
2.1.1.1. Synthesis of 4(5)-(v,v-diarylalkyl and -alkenyl)-
1H-imidazoles
Scheme A presents the basic method, which we have
used to produce of the 4(5)-(v,v-diarylalkyl)-1H-imida-
zole skeleton.
Here 1-benzyl-5-imidazole carbaldehyde reacted with
diarylalkylmagnesium halides to afford the structures A I.
The end products A III and AV were prepared from A I by
a successive sequence of reactions comprising dehydration
and catalytic hydrogenation. When synthesizing the unsatu-
rated compounds A III, the N-benzyl group was eliminated
by catalytic hydrogenation or by hydrogen transfer reaction
in ethanol before the dehydration to give the intermediates
A II. Alternatively the synthesis of A V directly from A I
could be achieved by hydrogen transfer reaction with
ammonium formate and Pd/C by heating in acetic acid
(Ram and Spicer, 1987).
A special method used to prepare a group of differential-
ly substituted 4(5)-3,3-diarylpropyl imidazoles is outlined
in Scheme E.
According to this procedure the aldol condensation of
1-benzyl-5-imidazole carbaldehyde with substituted
acetophenone gave the unsaturated ketone E I. The double
bond was hydrogenated selectively under neutral con-
ditions at room temperature to produce the ketone E II.
The diarylalkylimidazole E III was obtained by the re-
action of the ketone E II with an aryl magnesium bromide.
4(5)-3,3-Diarylpropylimidazole E IV could be prepared as
described in the latter part of Scheme C.
Diarylalkylhalides were either commercially available or
they were prepared according to the general methods. The

A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
113
Scheme A.
2.1.1.2. Preparation of 4(5)-(a,v-diarylalkyl- and
alkenyl)-1H-imidazoles
oxidation with manganese dioxide. The resulting ketone F
II was converted to the a,v-diarylalkylimidazole F III in
the Grignard reaction followed by deprotection, dehydra-
tion and hydrogenation affording the end product F V. The
hydrogenation of the alcohol F III with ammonium formate
and palladium on carbon as catalyst gave a convenient
straight entry to the end product F V. The geometric
The 4(5)-(a,v-diarylalkyl- and -alkenyl)-1H-imidazoles
were synthesized according to the general routes F–G.
According to route F the reaction of 1-benzyl-5-imidazole
carbaldehyde with appropriate arylalkylmagnesium bro-
mide gave the alcohol F I. This was followed by the

114
A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
Scheme B.
isomers of F IV were separated and the assignment of Z
and E configurations were made by means of 1D NOE
difference and NOESY NMR spectroscopic techniques
(Kalapudas et al., 1993) (Scheme F).
hydrogenation was used as such a precursor. The t-butyl
amino carbonyl moiety could be converted to the cyano
group by treatment with thionyl chloride, phosphorous
trichloride or phosphorous pentachloride.
As an alternative approach for the preparation of the
intermediate F III (Scheme G), imidazolyl aryl ketone was
allowed to react with arylalkylmagnesium bromide.
Yet another method for the preparation of 4(5)-a,v-
diarylalkenylimidazoles and a,v-diarylalkylimidazoles uti-
lized the McMurry reaction (McMurry, 1989), in which
the reductive carbonyl coupling of the imidazolyl aryl
ketone with an appropriate aldehyde in the presence of
titanium(0) resulted directly in the formation of 4(5)-a,v-
diarylalkenylimidazole as outlined in Scheme H. 4(5)-
(b,v-Diarylalkenyl- and -alkyl)-1H-imidazoles could be
prepared by the same method starting from 1-benzyl-5-
imidazole carbaldehyde and an appropriate diaryl ketone.
2.1.2. Preparation of N-substituted (a,v-diarylalkyl-
and -alkenyl)imidazoles and triazoles
Saturated N-substituted a,v-diarylalkylimidazoles and
triazoles J II were prepared according to Scheme J
(Karjalainen et al., 1997). In this procedure imidazole or
triazole underwent a direct N-alkylation with an appro-
priately substituted benzyl halide to give J I. Treatment of
compound J I with n-butyl lithium followed by alkylation
with an appropriate aryl alkyl halide gave the end product
J II.
An alternative method used to synthesize the compound
J II was the direct N-alkylation of imidazole or triazole
with diarylalkyl halide. N-Substituted v,v-diarylalkyl
imidazoles and triazoles could also be prepared with this
method starting from appropriate diarylalkyl halides.
1-Substituted a,v-diaryl-b-hydroxyalkyltriazoles and
imidazoles K II and unsaturated N-a,v-diaryltriazoles and
imidazoles K III were synthesized starting from N-benzyl
triazoles and imidazoles, which first were treated with
n-butyl lithium to produce the carbanion of the benzylic
carbon. Then the reaction with appropriately substituted
arylalkyl aldehydes afforded KII and the following elimi-
2.1.1.3. Preparation of para-cyanoderivatives
The methods used in the synthesis of 4(5)-(diarylalkyl-
and -alkenyl)imidazoles include reaction conditions in
which the cyano substituent is unstable. That is why the
strategy for the preparation of the cyano-substituted end
products was based on the methodology to produce the
cyano group from an appropriate precursor group in a late
stage of the reaction pathway. t-Butyl amino carbonyl,
which is rather stable towards organometal reagents and

A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
115
Scheme C.

116
A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
Scheme D.
nation of water by the treatment with acid yielded K III.
Alternatively the nucleophilic addition of the carbanion to
an appropriately substituted arylalkyl esters gave K I,
which was then reduced with sodiumborohydride to K II
(Scheme K).
also by the treatment of a ketone L I with thionyl
bisimidazole (Scheme L) (Ogata et al., 1987).
1-(1,3-Diaryl-3-hydroxypropyl)triazoles M III were pre-
pared by heating 1,3-diarylpropenones M I and triazole in
the presence of Triton B followed by reduction with
sodiumborohydride (Scheme M).
Unsaturated imidazole derivatives L II were prepared
Scheme E.

A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
117
Scheme F.
3. Experimental section
reference in the indicated solvent at an ambient tempera-
ture. Chemical shifts were measured downfield from TMS.
MS spectral data were obtained on a Kratos MS 80 RF
Autoconsole instrument. The spectra were considered
consistent with the assigned structures. The estimation of
the isomeric purity was based on the H NMR spectra. The
separation of enantiomers and measurement of optical
purity was done by chiral HPLC using Hewlett-Packard
3.1. Chemistry
Melting points were determined using a Buchi 510 glass
capillary melting point apparatus. Melting and boiling
points are uncorrected. NMR spectra were recorded with a
Bruker AC-P300 spectrometer using TMS as the internal
1
Scheme G.

118
A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
Scheme H.
HP 1090 apparatus. The silic acid used in the flash
chromatography was silica gel 60, 230–400 mesh.
3.2. Scheme A
3.2.1. 4-(4,4-Diphenylbutyl)-1H-imidazole (2)
3.1.1. General methods for the preparation of
diarylalkylimidazole and diarylalkyltriazole molecular
skeleton
3.2.1.1. 1-Benzyl-5-(1-hydroxy-4,4-diphenylbutyl)-1H-
imidazole
The detailed syntheses and the analytical parameters of
the compounds of this study are presented in our patent
publications, which are referred in the text above and in
the following just a few selected examples of these are
given to illustrate the principal methods used to prepare
these structures as outlined in Schemes A–M.
Two grams of magnesium turnings are covered with 60
ml of dry tetrahydrofuran. To the mixture is then added
dropwise a solution of 1-bromo-3,3-diphenylpropane (22.9
g) in 20 ml of dry tetrahydrofuran at such a rate that a
smooth reaction is maintained. After the addition is
complete, the reaction mixture is refluxed for one addition-
Scheme J.

A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
119
Scheme K.
al hour and cooled to room temperature. The reaction
3.2.1.2. 1-Benzyl-5-(4,4-diphenyl-1-butenyl)-1H-
imidazole
mixture is then added dropwise to a solution of 1-benzyl-5-
imidazolecarbaldehyde (7.35 g) in 80 ml of tetrahydro-
furan at 608C. After the addition is complete, the reaction
mixture is refluxed for 2 h, cooled and poured into cold
water. Tetrahydrofuran is evaporated and to the solution is
added conc. hydrochloric acid. The solution is cooled and
the precipitate which contains the product as hydrochloride
salt is removed by filtration, washed with water and dried.
Yield 14.1 g. M.p. 160–1688C: 1H NMR (as HCl-salt,
MeOH-d4) d 1.50–2.30 (m, 4H), 3.82 (t, 1H), 4.65 (t,
1H), 5.43 (s, 2H), 7 05–7.50 (m, 16H), 8.62 (d, 1H).
1 - Benzyl - 5 - (1 - hydroxy - 4,4 - diphenylbutyl) - 1H - im-
idazole hydrochloride (5.0 g) and 30.0 g of anhydrous
potassium hydrogen sulphate are heated at 1508C for 4 h.
The mixture is cooled, 90 ml of ethanol is added to
dissolve the product. The mixture is then filtered and the
filtrate is evaporated to minor volume. Water is added and
the mixture is made alkaline with sodium hydroxide. The
product is extracted in methylene chloride, washed with
water and evaporated to dryness. The product is then made
to hydrochloride salt with dry hydrochloric acid in dry
Scheme L.

120
A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
Scheme M.
1
ethylacetate. Yield is 2.9 g. M.p. 204–2068C: H NMR (as
HCl-salt, MeOH-d4) d 2.88–3.05 (m, 2H), 4.08 (t, 1H),
5.29 (s, 2H), 6.22–6.32 (m, 2H), 7.00–7.50 (m, 16H), 8.87
(d, 1H).
ethyl acetate using dry hydrogen chloride. Yield 0.2 g.
M.p. 204–2068C: 1H NMR (as HCl-salt, MeOH-d4) d
1.40–1.90 (m, 2H), 1.90–2.30 (m, 2H), 2.75 (t, 2H), 3.95
(t, 1H), 7.00–7.40 (m, 11H), 8.72 (d, 1H); HRMS Calcd
for C19H20N2: (M1) 276.1622. Found: (M1) 276.1626.
3.2.1.3. 1-Benzyl-5-(4,4-diphenylbutyl)-1H-imidazole
1-Benzyl-5-(4,4-diphenyl-1-butenyl)-1H-imidazole hy-
drochloride (2.0 g) is dissolved in ethanol and a catalytic
amount of Pd/C (10%) is added. The reaction mixture is
agitated vigorously at room temperature in a hydrogen
atmosphere until the uptake of hydrogen ceases. The
mixture is filtered and the filtrate is evaporated to dryness.
The residue which is the product is purified by flash
chromatography eluting with the mixture of methylene
chloride–methanol. Yield 1.3 g. M.p. of the hydrochloride
salt is 200–2028C: 1H NMR (as HCl-salt, MeOH-d4) d
1.30–1.70 (m, 2H), 1.85–2.20 (m, 2H), 2.61 (t, 2H), 3.83
(t, 1H), 5.35 (s, 2H), 7.05–7.50 (m, 16H), 8.89 (d, 1H).
3.3. Scheme F
3.3.1. 4-[1-(4-Cyanophenyl)-4-(4-fluorophenyl)-1-
butenyl]-1H-imidazole, isomers a (77) and b (4)
3.3.1.1. 1-Benzyl-5-[4-(4-fluorophenyl)-1-hydroxybutyl]-
1H-imidazole
Magnesium turnings (5.4 g) are covered with 50 ml of
dry tetrahydrofuran. A solution of 1-bromo-3-(4-fluoro-
phenyl)propane (48.7 g) in 200 ml of dry tetrahydrofuran
is then added dropwise to the mixture at such a rate that a
smooth reaction is maintained. After the addition is
complete, the reaction mixture is refluxed for one addition-
al hour and cooled to room temperature. To the reaction
mixture is then added dropwise the solution of 1-benzyl-5-
imidazolecarbaldehyde (20.9 g) in 200 ml of tetrahydro-
furan. After the addition is complete, the reaction mixture
is refluxed for 2 h, cooled and poured into cold saturated
ammonium chloride solution. Tetrahydrofuran phase is
removed and the water phase is extracted three times with
ethyl acetate. Organic phases are combined, dried and
evaporated to dryness. The residue is suspended with
3.2.1.4. 4-(4,4-Diphenylbutyl)-1H-imidazole
1-Benzyl-5-(4,4-diphenylbutyl)-1H-imidazole
hydro-
chloride (0.6 g) is hydrogenated in the mixture of 20 ml of
2 N hydrochloric acid and 10 ml of ethanol at 808C Pd/C
(10%) as catalyst. When the uptake of the hydrogen ceases,
the reaction mixture is cooled, filtered and evaporated to
dryness. Water is added and the mixture is made alkaline
with sodium hydroxide. The product is then extracted to
methylene chloride which is washed with water, dried with
sodium sulphate and evaporated to dryness. The residue is
the product as base and it is made to its hydrochloride in
1
diethyl ether and filtered. Yield 86%; H NMR (as base,
CDCl3) d 1.5–1.75 (m, 2H), 1.76–1.84 (m, 2H), 2.53 (t,

A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
121
2H), 4.49 (t, 1H), 5.20 and 5.26 (AB q, 2H), 6.91–7.1 (m,
7H), 7.30–7.35 (m, 3H), 7.48 (s, 1H).
Thionyl chloride is evaporated and the residue is dissolved
to ethyl acetate, washed with 2 M sodium hydroxide
solution and water, dried and evaporated to dryness. The
residue is crystallized from ethyl acetate as hydrogen
chloride salt (isomer a). The mother liquid is washed with
2 M sodium hydroxide solution and water, dried and
concentrated. Diethyl ether is added to the solution and the
precipitated product is filtered (isomer b). Yields 53%
isomer a and 17% isomer b; 1H NMR (as HCl-salt,
MeOH-d4) isomer a d 2.4 (q, 2H), 2.77 (t, 2H), 6.47 (t,
1H), 6.92–7.00 (m, 2H), 7.03 (s, 1H), 7.08–7.13 (m, 3H),
7.21–7.24 (m, 2H), 7.76–7.78 (m, 2H), 8.87 (d, 1H);
isomer b d 2.61 (q, 2H), 2.85 (t, 2H), 6.61 (t, 1H),
6.96–7.01 (m, 2H), 7.16–7.21 (m, 3H), 7.34 (d, 1H), 7.4
(d, 2H), 7.71 (d, 2H), 8.93 (d, 1H); HRMS Calcd for
C20H16FN3: (M1) 317.1325. Found: (M1) 317.1328.
3.3.1.2. 1-Benzyl-5-[4-(4-fluorophenyl)-1-oxobutyl]-1H-
imidazole
The mixture of 36.1 g of 1-benzyl-5-(4-(4-fluoro-
phenyl)-1-hydroxybutyl]-1H-imidazole and 53 g manga-
nese dioxide in 550 ml of tetrachloroethylene is refluxed
stirring for 4 h. The reaction mixture is filtered through
siliceous earth and the filtrate is evaporated to dryness. The
product is crystallized from acetone as hydrochloride salt:
1H NMR (as base, CDCl3) d 1.9–2.0 (m, 2H), 2.58 (t,
2H), 2.76 (t, 2H), 5.52 (s, 2H), 6.9–7.32 (m, 9H), 7.63 (s,
1H), 7.75 (s, 1H).
3.3.1.3. 1-Benzyl-5-[4-(4-fluorophenyl)-1-hydroxy-1-(4-
tert-butylaminocarbonylphenyl)butyl]-1H-imidazole
p-tert-Butylaminocarbonylphenyl bromide (9.9 g) is
dissolved in 200 ml of dry tetrahydrofuran and cooled to
2708C. To the solution is added dropwise n-butyl lithium
(5.3 g) in hexane and the mixture is stirred for 1 h.
1-Benzyl-5-[4-(4-fluorophenyl)-1-oxobutyl]-1H-imidazole
(10.4 g) in 200 ml of tetrahydrofuran is added to the
mixture at 708C and the mixture is allowed to warm to
room temperature and stirring is continued overnight.
Saturated ammonium chloride is added to the mixture and
the solution is extracted with ethyl acetate. Ethyl acetate
fractions are combined and evaporated to dryness: 1H
NMR (as base, CDCl31MeOH-d4) d 1.15–1.35 (m, 1H),
1.47 (s, 9H), 1.6–1.8 (m, 1H), 2.15–2.25 (m, 2H), 2.52 (t,
2H), 4.74 and 4.80 (AB q, 2H), 6.80–7.31 (m, 13H), 7.57
(d, 2H).
3.4. Scheme H
3.4.1. 4-(2,3-Diphenylpropyl)-1H-imidazole (16)
3.4.1.1. 1-Benzyl-5-(2,3-diphenyl-1-propenyl)-1H-
imidazole
A Flask is charged with Zn (39.8 g, 0.162 mol) and 200
ml of tetrahydrofuran. TiCl4 (57.3 g, 0.306 mol) is added
dropwise to the mixture at 0–108C. Then the mixture is
refluxed for 1 h. Ten grams (0.051 mol) desoxybenzoin
and 11.39 g (0.061 mol) 1-benzyl-5-imidazolyl aldehyde in
100 ml of THF is added to the mixture at room tempera-
ture. The mixture is heated to boiling and the refluxing is
continued for 3 h. After cooling to room temperature the
mixture is poured to 10% K2CO3 solution. Toluene is
added and the mixture is filtered through siliceous earth.
Toluene phase is separated and the water layer is extracted
again with toluene. Toluene extracts are combined, washed
with water, dried with MgSO4 and evaporated to dryness.
The crude product is purified by flash chromatography
with methylene chloride and methanol (9.75:0.25) as
eluent: MS m/z (%) 350 (M1) (8), 259 (4), 197 (57), 105
(100), 92 (6), 77 (58).
3.3.1.4. 4-[4-(4-Fluorophenyl)-1-hydroxy-1-(4-tert-butyl
aminocarbonylphenyl)butyl]-1H-imidazole
1-Benzyl-5-[4-(4-fluorophenyl)-1-hydroxy-1-(4-tert-
butylaminocarbonylphenyl)butyl]-1H-imidazole
hydro-
chloride (12.95 g) is dissolved into aqueous ethanol (400
ml) and 1.3 g of 10% Pd/C is added. Ammonium formate
(8.17 g) is added to the boiling solution in small portions.
The mixture is refluxed for 3 h. Then the reaction mixture
is filtered through siliceous earth and the filtrate is evapo-
rated to dryness. The residue is dissolved to methylene
chloride, washed with 2 M sodium hydroxide and water,
3.4.1.2. 4-(2,3-Diphenylpropyl)-1H-imidazole
1-Benzyl-5-(2,3-diphenyl-1-propenyl)-1H-imidazole
(0.67 g, 1.9 mmol) is dissolved to ethanol–water solution
(25:15). Pd/C (10%, 0.07 g) and 0.6 g (9.5 mmol) of
ammonium formate in 15 ml of water is added to the
mixture. After refluxing for 2 h the mixture is filtered and
the filtrate is evaporated to dryness. The residue is
dissolved into methylene chloride and washed many times
with water. Methylene chloride is evaporated and the
residue is dissolved into 2 M hydrogen chloride solution.
The solution is extracted twice with diethyl ether. The
water layer is made alkaline and extracted with methylene
chloride. The methylene chloride phase is dried and
evaporated to dryness. The product is then made to
1
dried and evaporated to dryness. Yield 79%; H NMR (as
base, CDCl3) d 1.3-1.5 (m, 1H), 1.47 (s, 9H), 1.6–1.8 (m,
1H), 2.0–2.3 (m, 2H), 2.45–2.6 (m, 2H), 6.78 (s, 1H), 6.91
(t, 2H), 6.94–7.07 (m, 2H), 7.43 (d, 2H), 7.50 (s, 1H), 7.61
(d, 2H).
3.3.1.5. 4-[1-(4-Cyanophenyl)-4-(4-fluorophenyl)-1-
butenyl]-1H-imidazole, isomers a and b
4 - [4 - (4 - Fluorophenyl) - 1 - hydroxy - 1 - (4 - tert - butyl-
aminocarbonylphenyl)butyl]-1H-imidazole (7.6 g) is dis-
solved into thionyl chloride (75 ml) and refluxed for 2 h.

122
A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
hydrochloride salt with dry hydrogen chloride gas in
diethyl ether. Yield 16%; MS m/z (%) 262 (M1.) (10),
197 (3), 181 (22), 171 (18), 82 (100); HRMS Calcd for
C18H18N2: (M1) 262.1466. Found: (M1) 262.1470.
(s, 1H); HRMS Calcd for C19H16FN3: (M1) 305.1325.
Found: (M1) 305.1352.
3.6. Scheme K
3.5. Scheme J
3.6.1. 1-[1-(4-Cyanophenyl)-4-(4-fluorophenyl)-1-
butenyl]-1,2,4-triazole, isomers a (82) and b (83)
3.5.1. 1-[1-(4-Cyanophenyl)-4-(4-fluorophenyl)butyl]-
1H-imidazole (73)
3.6.1.1. 1-[1-(4-Cyanophenyl)-4-(4-fluorophenyl)-2-
hydroxybutyl]-1,2,4-triazole, diastereomers a1d and b1c
1-[1-(4-Cyanophenyl)-4-(4-fluorophenyl)-2-hydroxy-
butyl]-1,2,4-triazole is prepared from 1-(4-cyanobenzyl)-
1,2,4-triazole (10.8 mmol), n-BuLi (10.8 mmol) and 3-(4-
fluorophenyl)propionaldehyde (13 mmol) according to
Scheme J. The product is purified by flash chromatog-
1-(4-Cyanobenzyl)-imidazole (1 g, 5.4 mmol) is dis-
solved into dry tetrahydrofuran (30 ml) and cooled to
2708C. n-BuLi in hexane (5.4 mmol) is added dropwise
into the reaction mixture. After stirring for additional 30
min at 2708C, 3-(4-fluorophenyl)-propyl bromide (1.5 g,
6.9 mmol) in THF (10 ml) is added to the mixture and
stirring is continued for 2 h. Then the mixture is allowed to
warm to room temperature. Saturated aqueous ammonium
chloride solution is added to the mixture, shaken and then
the layers are separated. THF-phase is dried and evapo-
rated to dryness. The residue is crystallized from iso-
propanol as hydrogen chloride salt. The filtrate is purified
1
raphy: H NMR (base, CDCl3): diastereomer a1d d 0.5–
1.7 (m, 2H), 2.6–2.73 (m, 1H), 2.8–2.9 (m, 1H), 4.4–4.5
(m, 1H), 5.23 (d, 1H), 6.96 (t, 2H), 7.11 (dd, 2H), 7.48 (d,
2H), 7.66 (d, 2H), 8.05 (s, 1H), 8.08 (s, 1H); diastereomer
b1c d 1.5–1.7 (m, 2H), 2.63–2.73 (m, 1H), 2.8–2.9 (m,
1H), 4.3–4.4 (m, 1H), 5.26 (d, 1H), 6.95 (t, 2H), 7.05 (dd,
2H), 7.38 (d, 2H), 7.65 (d, 2H), 8.07 (s, 1H), 8.12 (s, 1H).
1
by flash chromatography: H NMR (HCl-salt, MeOH-d4) d
1.5–1.63 (quintet, 2H), 2.3–2.5 (m, 2H), 2.7 (t, 2H), 5.75
(t, 1H), 6.94–7.00 (m, 2H), 7.14–7.19 (m, 2H), 7.62 (d,
2H), 7.63 (s, 1H), 7.78 (s, 1H), 7.8 (d, 2H), 9.6 (s, 1H);
HRMS Calcd for C20H18FN3: (M1) 319.1481. Found:
(M1) 319.1485.
3.6.1.2. 1-[1-(4-Cyanophenyl)-4-(4-fluorophenyl)-1-
butenyl]-1,2,4-triazole, isomers a and b
1-[1-(4-Cyano-phenyl)-4-(4-fluorophenyl)-2-hydroxy-
butyl]-1,2,4-triazole (0.42 g, 1.32 mmol) is dissolved into
acetonitrile. Phosphorous pentachloride (0.27 g, 1.3 mmol)
is added into the solution and the mixture is refluxed for 2
h. Acetonitrile is evaporated, the residue is dissolved with
2 M aqueous sodium hydroxide and extracted with methyl-
ene chloride. Methylene chloride is dried and the product
is crystallized from ethyl acetate as hydrogen chloride salt:
1H NMR (HCl-salt, MeOH-d4) isomer a d 2.42 (q, 2H),
2.85 (t, 2H), 6.85 (t, 1H), 7.0 (t, 2H), 7.16–7.21 (m, 2H),
7.37 (d, 2H), 7.74 (d, 2H), 8.82 (s, 1H), 9.38 (s, 1H);
HRMS Calcd for C19H15FN4: (M1) 318.1278. Found:
(M1) 318.1281; isomer b d 2.53 (q, 2H), 2.83 (t, 2H),
6.63 (t, 1H), 6.98 (t, 2H), 7.12–7.17 (m, 2H), 7.33 (d, 2H),
7.80 (d, 2H), 8.62 (s, 1H), 9.33 (s, 1H).
3.5.2. 1-[1-(4-Cyanophenyl)-3-(4-fluorophenyl)propyl]-
1,2,4-triazole (72)
The mixture of 1-chloro-1-(4-cyano-phenyl)-3-(4-fluoro-
phenyl)propane (4.18 g, 15 mmol) and 1,2,4-triazole
sodium derivative (1.37 g, 15 mmol) in DMF (30 ml) is
heated mildly for 4 h. DMF is evaporated. The residue is
dissolved in ethyl acetate and washed with water. The
organic layer is dried and the solvent is evaporated. The
product is purified with flash chromatography (Silica gel
60 mesh 230–400, eluent: methylene chloride–methanol
99:1); 1H NMR (HCl-salt, MeOH-d4) d 2.55–2.65 (m,
3H), 2.78–2.84 (m, 1H), 5.83 (dd, 1H), 7.00 (t, 2H), 1.17
(dd, 2H), 7.68 (d, 2H), 6.78 (d, 2H), 8.75 (s, 1H), 9.69 (s,
1H); HRMS Calcd for C18H15FN4: (M1) 306.1278.
Found: (M11) 307.1359.
3.7. Scheme M
3.7.1. 1-[1-(4-Cyanophenyl)-3-(4-fluorophenyl)-3-
hydroxypropyl]-1,2,4-triazole (93) and (94)
3.5.3. 1-[1-(4-Cyanophenyl)-3-(4-fluorophenyl)propyl]-
1H-imidazole (71)
1-[1-(4-Cyanophenyl)-3-(4-fluorophenyl)propyl]-1H-
imidazole (71) is prepared according to the same pro-
cedure using 1-chloro-1-(4-cyanophenyl)-3-(4-fluoro-
phenyl)propane and 1H-imidazole sodium derivative as
starting materials: H NMR (HCl-salt, MeOH-d4) d 2.59–
2.80 (m, 4H), 5.72 (m, 1H), 7.00 (t, 2H), 7.19 (dd, 2H),
7.63 (d, 2H), 7.67 (s, 1H), 7.81 (d, 2H), 7.84 (s, 1H), 9.23
3.7.1.1. 3-(4-Cyanophenyl)-1-(4-fluorophenyl)prop-2-
en-1-one
4-Cyanobenzaldehyde (0.1 mol) and 4-fluoro-
acetophenone (0.1 mol) are dissolved in methanol (150
ml) and solid sodium hydroxide is added to make the
solution alkaline. The mixture is stirred in room tempera-
ture for 6 h. The product is filtered off and washed with
1

A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
123
1
methanol: H NMR (CDCl3) d 7.21 (t, 2H), 7.59 (d, 1H),
3.8. Biochemical assays
3.8.1. Chemical sources
7.73 (s, 4H), 7.79 (d, 1H), 8.08 (dd, 2H).
All chemicals were from commercial sources and were
of the purest grade available.
3.7.1.2. 3-(4-Cyanophenyl)-1-(4-fluorophenyl)-3-(1-
triazolyl)propanone
3-(4-Cyanophenyl)-1-(4-fluorophenyl)-prop-2-en-1-
one (2.5 g, 10 mmol), 1,2,4-triazole (0.7 g, 10 mmol) and
one drop of Triton B are heated to solution. The cooled
mixture is diluted with ether and the product is filtered off.
The product is used for the next step without further
purification: H NMR (CDCl3) d 3.61 (dd, 1H), 4.37 (dd,
1H), 6.25 (dd, 1H), 7.15 (t, 2H), 7.55 (d, 2H), 7.68 (d,
2H), 7.95 (s, 1H), 7.99 (m, 2H), 8.23 (s, 1H).
3.8.2. Enzyme sources
Human placentas were obtained after normal delivery at
term. Tissue was rinsed with ice-cold buffer, coagulated
blood and connective tissue were excised from trophoblas-
tic tissue and the selected area was first cut into small
pieces with scissors and then homogenized in a Waring
Blendor for 20 s. The resulting homogenate was further
homogenized with a Potter-Elvehjem homogenizer with
four strokes each lasting about 10 s. Mitochondrial and
microsomal pellets (10 0003g and 100 0003g pellets,
respectively, were isolated by differential centrifugation
(Pelkonen and Pasanen, 1981). Pellets were resuspended in
phosphate buffer, pH 7.4, into the final protein concen-
tration of about 20 mg/ml.
1
3.7.1.3. 1-[1-(4-Cyanophenyl)-3-(4-fluorophenyl)-3-
hydroxypropyl]-1,2,4-triazole
3-(4-Cyanophenyl)-1-(4-fluorophenyl)-3-(1-triazolyl)-
propanone (25 mmol) is dissolved in methanol. Sodium
borohydride (12.6) is added and the mixture is stirred in
308C for 1 h. The mixture is rendered with 2 M hydrochlo-
ric acid and the solvent is evaporated. The residue is
dissolved into ethyl acetate, washed with dilute sodium
hydroxide and water, dried and the solvent is evaporated.
The mixture is rendered with 2 M hydrochloric acid and
the solvent is evaporated. The residue is dissolved into
ethyl acetate, washed with dilute sodium hydroxide and
water, dried and the solvent is evaporated. The product is
purified by flash chromatography as a mixture of dia-
stereomers (a1d:b1c, 2:1): 1H NMR (base, CDCl3) d
2.27–2.37 and 2.54–2.63 (2m, together 1H), 2.76–2.88
(m, 1H), 4.26 and 4.41 (2dd, together 1H), 5.62 and 5.91
(2dd, together 1H), 7.03 and 7.04 (2t, together 2H), 7.22–
7.31 (m, 2H), 7.50 and 7.55 (2d, together 2H), 7.65 and
7.69 (2d, together 2H), 7.94 and 8.04 (2s, together 1H),
8.05 and 8.22 (2s, together 1H).
For the ECOD assay, liver microsomes from untreated
Sprague–Dawley rats were prepared by the standard
ultracentrifugation method.
3.8.3. Aromatase assay
Aromatase activity was measured by the method of
Pasanen (1985). It involved evaporation of an acetone
solution containing the substrate (10 mM androstenedione,
containing 100 000 dpm of [3H]androstenedione) and 0.5
mg Tween 80 to dryness under a nitrogen stream. The
residue was then dissolved in 1 ml of an incubation
mixture containing 10 mM MgCl2, 2 mM NADPH, 25
mM phosphate buffer, pH 7.4, and 1 mg microsomal
protein containing aromatase enzyme. After incubation for
40 min at 1378C the enzyme reaction was stopped by
adding 100 ml ice-cold 33.3% TCA. After centrifugation
for 2 min the supernatant (1 ml) was taken into a 1-ml
plastic injection syringe and forced through a water-equili-
brated Sep-Pak into a counting vial. Ten ml of scintilla-
tion liquid were added and samples were counted in an
LKB Rack Betascintillation counter. The tritiated water
formed during the aromatase reaction was quantitated by
subtracting the control value from the sample value.
The mixture of diastereomers is triturated with diethyl-
ether and filtered. The diastereomer a1d (93) is enriched
in the insoluble material (.90%) and the diastereomer
b1c (94) in the filtrate (.80%). Both diastereomers are
1
further purified by recrystallization from toluene: H NMR
(HCl-salt, MeOH-d4) diastereomer a1d d 2.67 (ddd, 1H),
2.84 (dd, 1H), 4.54 (dd, 1H), 6.13 (dd, 1 H), 7.04 (t, 2H),
7.33 (dd, 2H), 7.77 (d, 2H), 7.81 (d, 2H), 8.79 (s, 1H),
9.86 (s, 1H); diastereomer b1c d 2.43 (ddd, 1H), 2.94
(ddd, 1H), 4.33 (dd, 1H), 6.14 (dd, 1H), 7.05 (t, 2H), 7.34
(dd, 2H), 7.66 (d, 2H), 7.75 (d, 2H), 8.69 (s, 1H), 9.62 (s,
1H); HRMS Calcd for C18H15FN4O: (M1) 322.1227.
Found: (M11) 323.1292.
3.8.4. Assay for cholesterol side-chain cleavage enzyme
(desmolase)
Cholesterol side-chain cleavage activity was measured
by the method of Pasanen and Pelkonen (1984). The
substrate was prepared according to Hanukoglu and Jef-
coate (1980). The mixture containing the following sub-
stances, 5 mM cholesterol (including 100 000 dpm
[3H]cholesterol), 5 mM cyanoketone, 10 mM MgCl2, 2
mM NADPH, 25 mM phosphate buffer, pH 7.4, and 1 mg
mitochondrial protein, was incubated for 30 min at 1378C
and the reaction was stopped by adding 900 ml methanol.
14C-Labelled pregnenolone was added to each incubation
3.7.2. Reference compounds
Fadrozole (Browne, 1985; Browne et al., 1991), let-
rozole (Bowman et al., 1990), CGS 18320B (Bowman et
al., 1990) and the compound (23) (Hirsch and Taylor,
1984) were prepared according to the literature.

124
A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
vial as an internal standard and the methanol-precipitated
protein was centrifuged. The clear supernatant was loaded
onto a pre-equilibrated Sep-Pak column (75% methanol)
and the pregnenolone formed in the reaction was eluted
into the counting vials with 80% methanol. Ten ml of
Scintillation liquid were added and samples were counted
in an LKB Rack Beta scintillation counter with a double
label program. The pregnenolone formed during the incu-
bation was quantitated by subtracting the control value
from the sample value.
monoarylalkyl imidazoles, in which the number of the
bridge carbons varied from one to five (compounds (5–9)
in Table 1). It was observed that the optimal length of the
carbon bridge, optimizing the effect on the aromatase
inhibition, is three or four carbons (see compounds (7–8)
for aromatase: IC50 35 and 12 mM, respectively).
While screening diarylalkyl imidazoles, an important
finding was made: an additional aryl group increased
aromatase activity considerably for example in compounds
(9) and (5) as compared to corresponding mono aryl
structures (13) and (10), respectively. The lengthening of
the bridge carbon chain to comprise three or four carbons
will produce the most potent aromatase inhibitors (compare
compounds (2, 10–14)) in this series too. It was of some
interest that 4(5)-(diphenylmethyl)imidazole (10) (Table
1), the basic skeleton of one of the reference compounds,
4(5)-[bis(4-fluorophenyl)methyl]-1H-imidazoles (23) (in
Table 7) was about 10-fold less potent than the diarylalkyl
imidazole (2) containing four bridge carbons.
Interestingly, the long-chained 4(5)-diarylalkyl imida-
zoles (e.g. compound (2)) showed good selectivity for
aromatase over desmolase as well, which was the crucial
matter as to the continuation of the synthetic program.
Diarylalkyl derivatives, in which the position of one of
the aryl rings was varied in the bridge, gave also a good
effect on aromatase inhibition. For example, 4(5)-(1,3-
diphenylpropyl)-1H-imidazole (15), had a IC50 value of
4.5 mM for aromatase.
3.8.5. 7-Ethoxycoumarin O-deethylase assay (ECOD)
The ability of the compounds to inhibit the enzyme
ECOD was investigated by the in vitro assay method
according to Aitio (1978). The incubation mixture includ-
ing cofactors (200 mM KCl, 10 mM MgCl2, 2.5 mM
NADP, 6 mM glucose-6-phosphate and 4 U glucose-6-
phosphate dehydrogenase) and 0.5 mg microsomal protein
from rat liver in 0.1 M sodium–potassium phosphate
buffer, pH 7.4 (total volume 0.49 ml), was preincubated at
1378C for 2 min and the reaction was started by adding 10
ml of 5 mM 7-ethoxycoumarin (final concentration 0.1
mM). After 10-min incubation, the reaction was stopped
by adding 0.5 ml 6% trichloroacetate and the mixture was
spun at 3000 rpm for 10 min. From the clear supernatant,
0.5 ml was taken to a fluorometer cuvette containing 2 ml
1.6 M glycine–NaOH buffer, pH 10.4, just before measur-
ing the fluorescence with the Hitachi spectrophotofl-
uorometer with settings excitation 390 nm and emission
440 nm. Authentic 7-hydroxycoumarin was used for the
calibration and for the construction of the standard curve.
The substitution of the 1-position of imidazole with the
diarylalkyl moiety preserved some activity (see compound
(22); IC50 for aromatase 16 mM), but the substitution of
the 2-position of imidazole reduced activity radically (for
example 2-(4,4-diphenylbutyl)-2H-imidazole: IC50 over
1000 mM).
3.8.6. Inhibition experiments
In inhibition experiments, the studied substance in a
final concentration from 0.1 mM to 1 mM was added into
incubation mixture in a volume of 5–10 ml, usually as an
ethanol or DMSO solution. The same volume of the solute
was added into control incubation tubes (especially DMSO
is slightly inhibitory itself). The IC50 values (concentration
causing a 50% inhibition) were determined graphically on
a semilogarithmic paper. Enzymes from different indi-
vidual placentas gave approximately similar IC50 values
(data not illustrated).
4.2. Screening the effect of substituents
With respect to substituent effects, especially two series
of compounds were screened: v,v-diarylalkyl 4(5)-imida-
zoles with three or four carbons in the bridge (compounds
(24–52) Table 2).
As a rule in these skeletons the substituents had surpris-
ingly little effect with respect to the aromatase inhibition
and selectivity. There are, however, exceptions. In the
three-carbon bridge series the para-cyano or the para-
cyano combined with the para-fluoro substitution in-
creased the potency of the skeleton considerably, as
compared the compound (12) (IC5015 mM) (Table 1) with
the compounds (26) (IC50 0.5 mM) and (38) (IC50 0.5
mM) (Table 2). However, also in this case, substituents
had no significant effect on the selectivity.
The effect of the fluoro and cyano substituents in the
a,v- and b,v-diarylalkyl framework could be observed to
be much more favourable than in the v,v-framework. With
the cyano group at the para-position of the a-phenyl and
hydrogen or fluorine at the para-position of the v-phenyl,
4. Results and discussion
4.1. Screening of basic skeletons
The observation that MPV-400 A III (1) (Fig. 1) was a
relatively potent aromatase inhibitor (IC5015 mM) started
our rational modification of the long-chained 4(5)-arylalkyl
imidazole and 4(5)-diarylalkyl imidazole skeletons (2) and
(5–22) (Table 1).
The effect of the length of the carbon chain on the
aromatase inhibition was tested first with a series of

A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
125
Table 1
Effects on aromatase (A) and desmolase (D) inhibition for a series of the basic carbon skeleton modifications in the mono- and diarylalkylimidazoles and
triazoles
Compound
R
m
n
X
Y
A: IC50
D: IC50
D/A
Method of
synthesis
References
(mM)a
(mM)a
5
6
7
8
H
H
H
H
0
1
2
3
4
0
1
2
3
4
5
0
1
1
2
0
3
2
3
0
0
0
0
0
0
0
0
0
0
0
2
1
2
1
4
1
0
0
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
C
C
400.0
300.0
35.0
12.0
50.0
25.0
49.0
15.0
2.0
1.7
14.5
4.5
45.0
5.5
7.5
ND
ND
ND
105
ND
ND
9.0
ND
320.0
68.0
61.0
7.5
18.0
12.0
19.0
ND
Karjalainen (1981)
Karjalainen (1981)
A
A
A
9
9
H
10
11
12
2
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Karjalainen (1981)
,1
H
C
A
A
A
F
H
H
E
A
B
J
Karjalainen et al. (1992)
Karjalainen et al. (1995b)
Karjalainen et al. (1995b)
Karjalainen et al. (1995b)
Karjalainen et al. (1995b)
160
40
4
2
,1
2
13
14
15
16
17
18
19
20
21
22
Karjalainen et al. (1995a)
Karjalainen et al. (1992)
Karjalainen et al. (1995b)
Karjalainen et al. (1995b)
3
25.0
4.7
20.0
16.0
27.0
28.0
52.5
6
1
3
J
a IC50 is the concentration of inhibitor required to give 50% inhibition.
both the potency and the selectivity of these structures
were increased as seen in the connection of the compound
(60) (Table 3). Also on the b,v-diaryl skeleton, the tested
fluorine and cyano were beneficial (compare compound
(17) in Table 1 with compound (66) in Table 3). Surpris-
ingly, however, the selectivity was not so impressive with
this framework.
We connected a,v-diaryl framework with substituents
also to the 1-position of imidazole and triazole, and both
the potency and the selectivity increased now considerably
(see compounds (71–74) in Table 4).
Also with 1-triazoles and with 1-imidazoles this effect is
rather conspicuous as compared the isomers of 4-[4-(4-
fluorophenyl)-1-[1,2,4]triazol-1-ylbut-1-enyl]benzonitrile
((82) and (83)) and 4-[4-(4-fluorophenyl)-1-imidazol-1-
ylbut-1-enyl]benzonitrile ((84) and (85)) (Table 5).
The effect of relative positions of fluoro and cyano
groups was also studied and observed to be important as
for the activity and selectivity. It was somewhat surprising
to find that dicyano substitution did not give this selectivity
difference (see compounds (78) and (79) in Table 5).
With 1-substituted imidazoles the geometric isomerism
gave similar selectivity difference between the geometric
isomers (compare compounds (84) and (85) in Table 5).
Interestingly, however, the introduction of a double bond
in this case did not seem to bring any improvement in the
selectivity compared to the selectivity of the corresponding
saturated structure 4-[4-(4-fluorophenyl)-1-imidazol-1-
ylbutyl]benzonitrile (73) (Table 4).
4.3. Geometric isomerism
Introduction of a double bond into the appropriately
substituted a,v-diarylalkyl framework linked to imidazole
or triazole and separation of E- and Z-isomers offered an
interesting possibility to affect selectivity. The isomer
possessing favourable spatial configuration proved to have
many-fold increased selectivity when compared to the
other isomer. Especially this increase in selectivity was
very pronounced with the four-carbon bridged a,v-
diarylalkene framework, connected to 4(5)-imidazole moi-
ety, as seen by comparing the effects of the geometric
isomers of the compound 4-[4-(4-fluorophenyl)-1-(1H-im-
idazol-4-yl)but-1-enyl]benzonitrile, (4) and (77) (Table 5).
4.4. Triazoles and optical isomerism
When our best skeletal framework was attached to
1-position of triazole ring (compounds (69), (70), (72),
(74), (80–83) in Tables 4 and 5), it was found that all the
important structure–activity characteristics for activity and

126
A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
Table 2
Effects on aromatase (A) and desmolase (D) inhibition for various substituent modifications in the basic v,v-diarylalkylimidazoles
Compound
R1
R2
n
A: IC50
D: IC50
D/A
Method of
synthesis
References
(mM)a
(mM)a
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
H
H
H
p-Me
p-F
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
7.1
10.1
0.48
33.0
20.0
14.0
8.6
3.2
8.6
19.0
5.0
5.2
32.0
2.4
0.5
0.85
0.86
3.4
2.5
16.0
2.8
8.5
28.0
3.5
50.0
440.0
14.0
8.2
29.0
48.0
32.0
36.0
26.0
20.5
30.0
110.0
ND
61.0
21.0
33.0
36.0
29.0
26.0
38.0
80.0
65.0
48.0
110.0
340.0
210.0
37.0
175.0
47.0
7
44
29
,1
14
3
4
11
3
E
E
E
C
C
C
C
C
C
C
E
E
C
E
E
E
C
A
A
A
A
A
A
A
A
C
C
A
B
Karjalainen (1992)
Karjalainen (1992)
Karjalainen et al. (1996)
Karjalainen (1992)
Karjalainen (1992)
Karjalainen (1992)
Karjalainen (1992)
Karjalainen (1992)
Karjalainen (1992)
Karjalainen (1992)
Karjalainen (1992)
p-CN
o-Me
m-Me
p-Me
o-Me
o-Me
p-OMe
m-F
o-Me
m-Me
p-Me
o-Me
o-Me
p-OMe
m-F
1
6
21
p-F
p-F
p-F
p-CF3
p-CF3
p-CF3
p-CN
p-CN
p-CN
H
H
H
H
H
m-Me
p-Me
p-CF3
p-NH2
p-NO2
p-F
p-CF3
p-CN
p-F
p-OMe
p-CN
o-Me
m-Me
o-F
25
42
39
19
8
10
2
29
8
2
31
3
8
2
Karjalainen et al. (1996)
Karjalainen et al. (1996)
Karjalainen et al. (1996)
Karjalainen et al. (1996)
Karjalainen et al. (1995b)
Karjalainen et al. (1995b)
Karjalainen et al. (1995b)
Karjalainen et al. (1995b)
Karjalainen et al. (1995b)
Karjalainen et al. (1995b)
Karjalainen et al. (1995b)
p-F
p-Et
m-Me
p-Me
p-CF3
p-NH2
p-NO2
p-F
120.0
26.0
20.0
3.3
Karjalainen et al. (1995b)
Karjalainen et al. (1995b)
Karjalainen et al. (1995b)
Karjalainen et al. (1996)
53
29
p-CN
p-CN
1.6
a IC50 is the concentration of inhibitor required to give 50% inhibition.
selectivity found by screening 4(5)-imidazoles, were valid
with 1-triazoles as well.
modifications the crucial matters with respect to the
favourable effects.
In the course of this research we observed that in vivo
activity was better with triazoles than with imidazoles and
so our focus in the structure–activity screening, when
making further modifications in the skeleton, was shifted
to 1-triazoles.
We synthesized 1-substituted a,v-diarylalkyl triazoles
with a hydroxyl group in the b-carbon of the bridge,
creating thus an additional chiral centre giving a new
aspect as to achieving selectivity (see compounds (3) and
(86–94) in Table 6). The separated diastereomers dis-
played substantial additional selectivity differences. The
length of the bridge carbon chain containing three carbons
and appropriate substitution pattern were also with these
4.5. Comparison with reference compounds
The comparison of the potency and selectivity of our
best structures, e.g., compounds (3), (4) and (89), with the
selected known inhibitors was made and is presented in
Table 7. As reference compounds were used letrozole,
fadrozole, CGS 18320B and 4-[bis(4-fluorophenyl)-
methyl]-1H-imidazole (23), which was included in Eli
Lilly’s patent application. The reference compounds were
prepared according to the literature and studied in conjunc-
tion in our own compounds.
The potency of compounds (3), (4) and (89) to inhibit

A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
127
Table 3
Effects on aromatase (A) and desmolase (D) inhibition for various substituent modifications in the basic diarylalkylimidazoles
Compound
R1
R2
m
n
A: IC50
D: IC50
D/A
Method of
synthesis
References
(mM)a
(mM)a
53
54
55
56
57
58
59
60
61
62
63
64
65
66
F
F
F
F
F
CN
F
CN
F
F
H
0
0
0
0
0
0
0
0
0
0
1
1
1
1
4
4
4
4
3
3
3
3
2
2
1
2
2
2
210.0
3.5
3.8
8.5
0.8
0.21
0.63
0.23
0.7
2.2
4.4
19.0
16.0
57.0
170.0
31.0
12.0
29.0
20.0
22.0
22.0
3.4
9
5
F
F
F
F
F,G
F
H,G
F
H,G
G
H
Karjalainen et al. (1995b)
Karjalainen et al. (1995b)
Karjalainen et al. (1995b)
Karjalainen et al. (1995b)
Karjalainen et al. (1995b)
Karjalainen et al. (1996)
o-Me
m-Me
p-Me
H
15
20
41
57
46
87
31
10
1
H
p-F
p-F
H
p-F
H
H
p-F
p-F
Karjalainen et al. (1996)
Karjalainen et al. (1995b)
F
F
F
CN
2.2
1.3
0.5
28.0
20.0
6.6
13
15
13
H
H
H
Karjalainen et al. (1995a)
Karjalainen et al. (1995a)
a IC50 is the concentration of inhibitor required to give 50% inhibition.
aromatase proved to be superior to aminoglutethimide and
they were at least equally potent as the other reference
compounds. The best representatives of our structures gave
a very high selectivity ratio towards desmolase, with
compounds (3) and (4) selectivity ratio being over 3850
and 5268, respectively.
The compound, 4-[(4-cyanophenyl)(1H-imidazol-4-
yl)methyl]benzonitrile (95), a structural relative of let-
rozole, was synthesized (according to the method C) for
comparison of selectivity and activity between 1-substi-
tuted and 4(5)-substituted diarylalkyl imidazoles. Rather
low aromatase inhibition potency and poor selectivity of
Table 4
Effects on aromatase (A) and desmolase (D) inhibition for substituent modifications in some a,v- and v,v-diarylalkylimidazoles and triazoles
Compound
R1
m
n
X
A: IC50
D: IC50
D/A
Method of
synthesis
References
(mM)a
(mM)a
67
68
69
70
71
72
73
74
H
F
F
F
CN
CN
CN
CN
2
2
2
0
0
0
0
0
0
0
0
2
2
2
3
3
C
C
N
N
C
N
C
N
5.0
2.2
36.0
4.0
0.05
0.19
0.04
0.12
21.0
33.0
36.0
11.0
11.8
20.0
17.0
27.0
4
15
1
J
J
J
J
J
J
J
J
3
246
105
236
225
Karjalainen et al. (1997)
a IC50 is the concentration of inhibitor required to give 50% inhibition.

128
A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
Table 5
Effects on aromatase (A) and desmolase (D) inhibition for Z,E-isomerism in some a,v-diarylalkylimidazoles and triazoles
Compound
Isomer
R1
R2
n
X
Y
A: IC50
D: IC50
D/A
Method of
synthesis
References
(mM)a
(mM)a
75
76
77
4
78
79
80
81
82
83
84
85
a
b
a
b
a
b
a
b
a
b
a
b
CN
CN
CN
CN
CN
CN
CN
CN
CN
CN
CN
CN
F
F
F
F
CN
CN
F
F
F
1
1
2
2
2
2
1
1
2
2
2
2
C
C
C
C
C
C
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
C
C
0.19
0.33
0.21
0.19
1.1
36.0
3.4
33.0
189
10
157
5268
6
11
48
3300
25
2140
18
F
F
F
F
F
F
K
K
K
K
K
K
Karjalainen et al. (1996)
Karjalainen et al. (1996)
Karjalainen et al. (1996)
Karjalainen et al. (1996)
Karjalainen et al. (1996)
Karjalainen et al. (1996)
Karjalainen et al. (1997)
Karjalainen et al. (1997)
Karjalainen et al. (1997)
Karjalainen et al. (1997)
Karjalainen et al. (1997)
Karjalainen et al. (1997)
.1000
6.3
5.4
5.7
165.0
16.0
300.0
4.0
0.5
0.12
0.05
0.65
0.14
0.34
0.06
F
F
F
13.5
225
a IC50 is the concentration of inhibitor required to give 50% inhibition.
(95) pointed out that with 4(5)-substituted imidazoles the
lengthened carbon backbone is the essential structural
feature for the high potency and the great selectivity.
4.6. Effects on drug-metabolizing cytochrome P450
In the light of potential drug interactions, it is important
Table 6
Effects on aromatase (A) and desmolase (D) inhibition for optical isomerism in a,v-diarylalkylimidazoles and triazoles
Compound
Isomer
m
n
X
A: IC50
D: IC50
D/A
Method of
synthesis
References
(mM)a
(mM)a
86
87
3
88
89
90
91
92
93
94
a1d
b1c
a1d
b1c
a1d
b1c
a1d
b1c
a1d
b1c
0
0
0
0
0
0
0
0
1
1
0
0
1
1
2
2
2
2
0
0
N
N
N
N
N
N
C
C
N
N
42.0
17.0
0.22
0.5
0.95
3.9
0.18
0.5
0.37
0.7
.1000
215.0
.1000
305
.1000
350
24
13
3850
610
1054
90
272
130
851
600
K
K
K
K
K
K
K
K
M
M
Karjalainen et al. (1997)
Karjalainen et al. (1997)
Karjalainen et al. (1997)
Karjalainen et al. (1997)
Karjalainen et al. (1997)
Karjalainen et al. (1997)
Karjalainen et al. (1997)
Karjalainen et al. (1997)
49.0
65
315
420
a IC50 is the concentration of inhibitor required to give 50% inhibition.

A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
129
Table 7
Comparative effects on aromatase (A), desmolase (D) and ECOD inhibition for some known and some of our novel a,v-diarylalkylimidazoles and triazoles
Compound Structure
A: IC50 (mM)a D: IC50 (mM)a ECOD: IC50 (mM)a
a IC50 is the concentration of inhibitor required to give 50% inhibition.

130
A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
to screen affinities of chemicals to drug-metabolizing
enzymes, especially to cytochrome P450 enzymes. In this
study, rat liver microsomal 7-ethoxycoumarin O-deeth-
ylase (ECOD) activity, representative for many different
P450 enzymes, was employed. The general conclusion is
that most imidazole derivatives are rather potent inhibitors
of ECOD, whereas triazoles exhibit much less inhibition.
However, taking into consideration that many inhibitors
display rather restricted CYP selectivity (see Pelkonen et
al., 1998), this conclusion has to be regarded as very
preliminary.
work, Pirjo Korpi and Riitta Heikkinen for their assistance
in the biochemical testing. Research program of Olavi
Pelkonen was supported also by The Academy of Finland
Medical Research Council.
References
Ahokoski, O., Irjala, K., Huupponen, R., Halonen, K., Salminen, E.,
Scheinin, H., 1998. Hormonal effects of MPV-2213ad, a new selective
aromatase inhibitor, in healthy male subjects. Br. J. Clin. Pharmacol.
45, 141–146.
Especially the most potent and selective compounds (3)
and (4) were virtually devoid of any affinity to rat liver
P450 enzymes (see Table 7), and later similar findings
were observed with human liver microsomes (unpublished
results). Interestingly also compound (4), an imidazole
derivative, was very selective with respect to ECOD. The
geometric isomerism plays an important role in this
selectivity too: the Z-isomer (4) is about nine times more
selective than the E-isomer (77) (Table 5). The same kind
of comparison between the diastereomers (3) and (88)
(Table 6) reveals that the optical isomerism has a strong
effect on selectivity in the case of ECOD as well, although
both diastereomers are very selective.
Aitio, A., 1978. A simple and sensitive assay of 7-ethoxycoumarin
deethylation. Anal. Biochem. 85, 488–491.
Bhatnagar, A.S., Haeusler, A., Schieweck, K., 1990a. Inhibition of
aromatase in vitro and in vivo by aromatase inhibitors. J. Enzyme
Inhib. 4 (2), 179–186.
Bhatnagar, A.S., Haeusler, A., Schieweck, K., Lang, M., Bowman, R.,
1990b. Highly selective inhibition of estrogen biosynthesis by CGS
20267, a new non-steroidal aromatase inhibitor. J. Steroid Biochem.
Mol. Biol. 37 (6), 1021–1027.
Bowman, R.M., Steele, R.E., Browne, L.J., 1990. Alpha-heterocycle
substituted tolunitriles. US 4,937,250. US: Ciba-Geigy Corp., USA.
Browne, L.J., 1985. Substituirte bicyclische Verbindungen. EP 0 165 904
A2.
Browne, L.J., Gude, C., Rodriguez, H., Steele, R.E., Bhatnager, A., 1991.
Fadrozole hydrochloride: a potent, selective, nonsteroidal inhibitor of
aromatase for the treatment of estrogen-dependent disease. J. Med.
Chem. 34 (2), 725–736.
Dukes, M., Edwards, P.N., Large, M., Smith, I.K., Boyle, T., 1996. The
preclinical pharmacology of ‘Arimedex’ (anastrozole; ZD1033) — a
potent, selective aromatase inhibitor. J. Steroid Biochem. Mol. Biol.
58 (4), 439–445.
Graham-Lorence, S., Amarneh, B., White, R.E., Peterson, J.A., Simpson,
E.R., 1995. A three-dimensional model of aromatase cytochrome
P450. Protein Science 4, 1065–1080.
Goss, P.E., Gwyn, K.M.E.H., 1994. Current perspectives on aromatase
inhibitors in breast cancer. J. Clin. Oncol. 12, 2460–2470.
Hirsch, K.S., Taylor, H.M., 1984. 4(5)-Substituted imidazoles. EP 0168
965 A1.
Jones, C.D., Winter, M.A., Hirsch, K.S. et al., 1990. Estrogen synthetase
inhibitors. 2. Comparison of the in vitro heterocycles substituted with
diarylmethane or diarylmethanol groups. J. Med. Chem. 33 (1), 416–
429.
5. Conclusions
Some interim conclusions could be drawn on the basis
of results described above: a good aromatase inhibition is
achieved with the structures including a framework of
diarylalkyl moiety comprising a lengthened carbon chain
linked to the 1- or 4(5)-positions of imidazole or 1-position
of triazole. An especially favourable framework seems to
be the a,v-diarylalkyl with three or four carbon atoms in
the bridge, which connects the framework to the heterocy-
cle. The best substituents, as for the aromatase activity, are
the cyano group in the a-phenyl ring and fluorine atom in
the v-phenyl ring. Although the substituents bring some
selectivity, particularly geometric and optical isomerism
give rise to the high selectivity.
The outcome of this synthetic program was a large
group of potent and selective aromatase inhibitors, of
which the most selective representatives are the structures
MPV-2213 C II a1d (3) and MPV-1837 A V b (4). The
MPV-2213 C II a1d (3) also demonstrated the desired
effect, i.e., the inhibition in estrogen synthesis in ex-
perimental animals (unpublished results) and in human
volunteers in vivo (Ahokoski et al., 1998).
¨
Kalapudas, A., Sodervall, M., Rahkamaa, E., 1993. In: Poster in XVI
National NMR Symposium, Turku, Finland.
Karjalainen, A.J., 1981. On the Design and Synthesis of Antihypertensive
4(5)-Substituted Imidazole Derivatives. Acta Universitatis Ouluensis;
Series A (no. 125).
Karjalainen, A.J., Kurkela, K.O.A., 1981. 4-Benzyl- and 4-ben-
zoylimidazole derivatives. EP 24829.
Karjalainen, A., Kurkela, K., 1985. Antihypertensive substituted imida-
zole derivatives. US 4544664.
Karjalainen, A., Kurkela, K., 1986. Substituted imidazoles and their use.
US 4568686.
Karjalainen, A.J., Kurkela, K.O.A., Parhi, S.S.L., 1982. Substituted
imidazole derivatives and their use. EP 58047 A1
Karjalainen, A.J., Virtanen, R.E., Savolainen E.J., 1989. Preparation and
testing of medetomidine enantiomers as CNS agents. GB 2206880 A1.
Karjalainen, A., Kangas, L., Kurkela, K., 1992. Aromatase inhibiting
4(5)-imidazoles. EP 0311447.
Acknowledgements
¨
Karjalainen, A.J., Kalapudas, A.M., Pelkonen, R.O., Sodervall, M.L.,
Lahde, M.A., Lammintausta, R.A.S., 1995a. Aromatase-inhibiting
4(5)-imidazoles. EP 0 476 944 B 1.
The authors are grateful to Mirja-Liisa Vesa, Marja
Fisher, Anita Puputti, Riitta Ristikari and Anja Kok-
koniemi for their technical assistance in the synthetic
¨
Karjalainen, A., Pelkonen, O., Sodervall, M.-L. et al., 1995b. Aromatase
inhibiting 4(5)-imidazoles. US Pat. 5,439,928.

A. Karjalainen et al. / European Journal of Pharmaceutical Sciences 11 (2000) 109–131
131
¨
¨
¨¨
Karjalainen, A., Sodervall, M.-L., Kalapudas, A. et al., 1996. Novel
Pelkonen, O., Maenpaa, J., Taavitsainen, P., Rautio, A., Raunio, H., 1998.
Inhibition and induction of human cytochrome P450 (CYP) enzymes.
Xenobiotica 28, 1203–1253.
selective aromatase inhibiting compounds. US Pat. 5,559,141.
¨
Karjalainen, A., Sodervall, M.-L., Kalapudas, A. et al., 1997. Novel
selective aromatase inhibiting compounds. GB 2,273, 704.
McMurry, J.E., 1989. Carbonyl-coupling reactions using low-valent
titanium. Chem. Rev. 89, 1513–1524.
Pelkonen, O., Pasanen, M., 1981. Effect of heparin on the subcellular
distribution of human placental 7-ethoxycoumarin O-deethylase activi-
ty. Biochemistry 30, 3254–3256.
Miller, W.L., 1988. Molecular biology of steroid hormone synthesis.
Endocr. Rev. 9, 295–318.
Ogata, M., Matsumoto, H., Shimizu, S., Kida, S., Shiro, M., Tawara, K.,
1987. Synthesis and antifungal activity of new 1-vinylimidazoles. J.
Med. Chem. 30, 1348–1354.
Parhi, S.S.L., 1985. Substituted 2-mercaptoimidazoles and their use. EP
146228 A1.
Pasanen, M., 1985. Human placental aromatase activity: use of a C18
reversed-phase cartridge for separation of tritiated water or steroid
metabolites in placentas from both smoking and non-smoking mothers
in vitro. Biol. Res. Pregnancy 6, 94–99.
Ram, S., Spicer, L.D., 1987. Rapid debenzylation of N-benzylamino
derivatives to amino-derivatives using ammoniumformate as catalytic
hydrogen transfer agent. Tetrahedron Lett. 28 (5), 515–516.
Santen, R.J., Manni, A., Harvey, H., Redmond, C., 1990. Endocrine
treatment of breast cancer in women. Endocr. Rev. 11, 221–265.
Simpson, E.R., Mahendroo, M.S., Means, G.D., Kilgore, M.W., Hinshel-
wood, M.M., Graham-Lorence, S., Amarneh, B., Ito, Y., Fisher, C.R.,
Michael, M.D., Mendelson, C.R., Bulun, S.E., 1994. Aromatase
cytochrome P450, the enzyme responsible for estrogen biosynthesis.
Endocr. Rev. 15, 342–355.
Vanden Bossche, H., Moereels, H., Koymans, L.M.H., 1994. Aromatase
inhibitors — mechanisms for non-steroidal inhibitors. Breast Cancer
Res. Treat. 30 (1), 43–55.
Pasanen, M., Pelkonen, O., 1984. Cholesterol side-chain cleavage activity
in human placenta and bovine adrenals: An one step method for
separation of pregnenolone formed in vitro. Steroids 43, 517–527.