1562
A. S. Kalgutkar et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1559–1563
156 (+2H+)
NH
revealed ꢀ5-fold lower conversion of 5 to 6 when compared with
109
F
156.0
100
50
methoxylamine conjugate formation with 1.
Next, the possibility that lack of mutagenicity of 5 can occur via
substrate inhibition of the cytochrome P450’s responsible for reac-
tive metabolite formation was examined. In human liver micro-
O
F
N
N
170 (+2H+)
N
170.0 213.1
213 (+2H+)
somes (HLM), the half-lives of 1 (T1/2 = 8 min) and 5 (T1/2
=
5 (MH+ = 349)
20 min) were only increased when incubations were conducted
in the presence of the P4503A4 and P4501A2-selective inhibitors,
ketoconazole and furafylline, respectively, (T1/2 for 1 and 5 were
>90 min and 40–45 min with ketoconazole and furafylline, respec-
tively). This observation suggests that P4503A4 and P4501A2 are
principally involved in the metabolism of 1 and 5.19 Based on this
information, P4501A2 and 3A4 inhibition by 1 and 5 was evaluated
in HLM.20 Compound 5 inhibited P4501A2 and P4503A4 with IC50
109.0
139.1
349.2
0
100 140 180 220 260 300 340 380
m/z (amu)
values of 0.4
weak inhibitor of P4503A4 (IC50 of 14
no inhibitory effect on P4501A2 activity (IC50 > 50
l
M and 4.9
l
M, respectively. In contrast, 1 was a
M) and had virtually had
M). Consider-
l
l
Figure 3. Product ion spectrum obtained by the CID of the MH+ ion (m/z 349) of 5.
ing that ariclor 1254 induction in the rat results in marked in-
creases in activities of P450 isozymes (especially those
responsible for metabolism/bioactivation of 1 and 5: >35-fold
induction of P4501A and >2-fold induction of P4503A)17, it is quite
possible that at the high substrate concentrations used for the
Ames test 5 inhibits its own metabolism in the S9 segment result-
ing in reduction of reactive intermediates and consequently lack of
mutagenicity.
In summary, insights into the mechanisms of reactive metabo-
lite formation with the 5-HT2C receptor agonist 1 assisted us in the
design of follow-on compounds to eliminate genotoxic response.
Diazinylpiperazine 5 retained the attractive primary pharmacology
and pharmacokinetic properties of 1 and was devoid of in vitro
mutagenicity in the Salmonella assay. Despite the negative result
in the genotoxicity assay, the compound retained the ability of 1
in forming DNA-reactive metabolites derived from piperazine ring
bioactivation. Lack of mutagenicity with 5 despite forming reactive
metabolites may be rationalized on the basis of occurrence of sev-
eral events in tandem such as: (a) absence of quinone methide for-
mation (the formation of this metabolite may be essential along
with formation of the aldehyde and nitrone intermediates for
mutagenicity to occur), (b) slower rates of metabolism of 5 (when
compared with 1) and (c) self-catalyzed inhibition of its own
metabolism. The latter two phenomenon could potentially reduce
the ‘‘critical mass” of reactive metabolites required for DNA adduc-
tion leading to mutagenicity.
cause the fumarate and HCl salt forms of 1 and 5, respectively, used
in genotoxicity testing, possessed comparable solubility.15
There are several other biochemical hypotheses which can indi-
vidually or collectively account for the disconnect in mutagenicity
between the two compounds. For instance, it is possible that statis-
tically significant increases in reverse mutations via DNA adduc-
tion to electrophilic intermediates is dependent upon the
formation of aldehyde, nitrone and quinone-methide reactive
metabolites; a scenario only likely with 1. While S9/NADPH-
dependent covalent modification of DNA by [14C-1] (radiolabeled
on the pyrazine carbons) and the decrease in mutagenicity in the
Salmonella assay upon co-incubation with methoxylamine
strengthens the role of the aldehyde and nitrone as a DNA-reactive
intermediates3, the contribution of the liberated quinone-methide
derivative in the mutagenic response of 1 cannot be ruled out,
especially since addition of GSH in the Salmonella assay did result
in attenuation of the mutagenic response of 1.3
Apart from changes in bioactivation profile, other suitable
explanations for mutagenicity differences of 1 and 5 include: (a)
differences in rates of substrate consumption in S9 and/or (b) sub-
strate inhibition of P450 enzymes responsible for reactive metabo-
lite formation.16,17 The likelihood that greater metabolic resistance
of 5 (compared with 1) in S9 significantly decreases reactive alde-
hyde and nitrone formation and therefore abrogates mutagenicity
was explored by comparing half-lives of 1 and 5 in the metabolic
activation system.18 As shown in Figure 4, 5 was ꢀ3-fold more
resistant towards metabolism in NADPH-supplemented S9 when
compared with 1 (T1/2 for 5 ꢀ13–19 min; T1/2 for 1 ꢀ4–7 min). Fur-
thermore, LC–MS/MS analysis of the peak height ratios of meth-
oxylamine conjugate / parent compound with both 1 and 5
References and notes
1. Kim, B. S.; Margolin, B. H. Environ. Mol. Mutagen. 1999, 34, 297.
2. Zeiger, E. Regul. Toxicol. Pharmacol. 1998, 28, 85.
3. Kalgutkar, A. S.; Dalvie, D. K.; Aubrecht, J.; Smith, E. B.; Coffing, S. L.; Cheung, J.
R.; Vage, C.; Lame, M. E.; Chiang, P.; McClure, K. F.; Maurer, T. S.; Coelho, R. V.,
Jr.; Soliman, V. F.; Schildknegt, K. Drug Metab. Dispos. 2007, 35, 848.
4. Kalgutkar, A. S.; Fate, G.; Didiuk, M. T.; Bauman, J. Expert Rev. Clin. Pharmacol.
2008, 1, 515.
5. (a) Siuciak, J. A.; Chapin, D. S.; McCarthy, S. A.; Guanowsky, V.; Brown, J.;
Chiang, P.; Marala, R.; Patterson, T.; Seymour, P. A.; Swick, A.; Iredale, P. A.
Neuropharmacology 2007, 52, 276; b In primary in vitro screens for the 5-HT2
receptor subtypes, both 1 and 5 are potent, functional agonists at the 5-HT2C
receptor with EC50 values of 0.6 nM (96% maximal 5-HT activation) and 1.7 nM
(88% maximal 5-HT activation), respectively. The EC50 for 5-HT2A agonism by 1
and 5 was 381 nM (67% activation) and 4441 nM (62% activation), respectively.
While 1 also displayed weak agonism of the 5-HT2B receptor (EC50 = 201 nM,
26% activation), 5 was an antagonist of this receptor. In acute in vivo studies,
oral administration of 1 and/or 5 to Wistar rats resulted in the dose-dependent
inhibition of both spontaneous, nocturnal food intake and fasting-induced re-
feeding. Furthermore, oral administration of 1 and 5 at 30 mg/kg daily to
Wistar rats for four days resulted in ꢀ8–10% inhibition of cumulative food
intake and ꢀ25% reduction in body weight over the four-day period.
6. Predicted human blood clearance from liver microsomes for both 1 and 5 is
ꢀ4.0 mL/min/kg, respectively. Protocols for measuring half-lives in human liver
microsomes and subsequent scaling to blood clearance have been published.
See Obach, R. S. Drug Metab. Dispos. 1999, 27, 1350–1359.
25
20
15
10
5
0
1
5
Aroclor Rat S9 (mg/mL)
Figure 4. Differences in metabolic half-lives of 1 (open bars) and 5 (filled bars) in
NADPH-supplemented ariclor 1254-induced rat S9.