Reversible carnitine palmitoyltransferase inhibitors with broad chemical diversity as potential antidiabetic agents

Article abstract of DOI:10.1021/jm010889+

A series of carnitine related compounds of general formula XCH₂CHZRCH₂Y were evaluated as CPT I inhibitors in intact rat liver (L-CPT I) and heart mitochondria (M-CPT I). Derivative 27 (ZR = -HNSO₂R, R = C₁₂, X

Full text of DOI:10.1021/jm010889+

© Copyright 2001 by the American Chemical Society
Volume 44, Number 15
J uly 19, 2001
Letters
hyperglycemia and hyperinsulinemia on hepatic glucose
production. It is also thought that the high level of
plasma free fatty acids (FFA) leads to an increase of
liver mitochondrial â-oxidation, which is crucial to drive
gluconeogenesis at higher rates.2 The mitochondrial
oxidation of long-chain FFA requires the intervention
of two membrane-bound carnitine-dependent long-chain
acyltransferases, also known as carnitine palmitoyl-
transferases (CPT).3 CPT I, the outer mitochondrial
membrane enzyme, catalyzes the formation of long-
chain acylcarnitines. Two genes encode for what are
known as liver (L-CPT I) and muscle (M-CPT I) isoforms
of CPT I. CPT II, the inner mitochondrial membrane
enzyme, reconverts long-chain acylcarnitines into long-
chain acyl coenzyme A esters. At variance of CPT I, CPT
II is the product of a single gene. Long-chain acyl-CoAs
are then â-oxidized to acetyl-coenzyme A, which acti-
vates a key gluconeogenetic enzyme, pyruvate carboxy-
lase.4 CPT inhibitors, by lowering the level of acetyl-
coenzyme A, indirectly reduce liver gluconeogenesis and
are hence helpful in the treatment of type II diabetes
as hypoglycemic agents.5,6 Methyl 2-tetradecylglycidate
is the first of a class of mitochondrial irreversible CPT
I inhibitors containing the oxirane carboxylate moiety.5
The clinical development of these compounds as hy-
poglycemic agents was discontinued, possibly because
they are equally active in inhibiting both L- and M-CPT
I isoforms, and side effects have been observed such as
cardiac hypertrophy.7
Rever sible Ca r n itin e
P a lm itoyltr a n sfer a se In h ibitor s w ith
Br oa d Ch em ica l Diver sity a s P oten tia l
An tid ia betic Agen ts
Fabio Giannessi,*,† Piero Chiodi,‡ Mauro Marzi,†
Patrizia Minetti,† Pompeo Pessotto,‡
Francesco De Angelis,§ Emanuela Tassoni,†
Roberto Conti,‡ Fabrizio Giorgi,† Massimo Mabilia,|
Natalina Dell’Uomo,† Sandra Muck, Maria O. Tinti,†
Paolo Carminati,‡ and Arduino Arduini*,‡
Departments of Chemical Research, Endocrinology &
Metabolism, and Analytical Research, Sigma Tau
Pharmaceutical Industries S.p.A., Via Pontina Km 30.400,
00040 Pomezia, Italy, Department of Chemistry, Chemical
Engineering and Materials, University of L’Aquila,
L’Aquila, Italy, and S.IN-Soluzioni Informatiche S.a.s.,
Vicenza, Italy
Received April 2, 2001
Abstr a ct: A series of carnitine related compounds of general
formula XCH2CHZRCH2Y were evaluated as CPT I inhibitors
in intact rat liver (L-CPT I) and heart mitochondria (M-CPT
I). Derivative 27 (ZR ) -HNSO2R, R ) C12, X ) trimethyl-
ammonium, Y ) carboxylate, (R) form) showed the highest
activity (IC50 ) 0.7 µM) along with a good selectivity (M-CPT
I/L-CPTI IC50 ratio ) 4.86). Diabetic db/db mice treated orally
with 27 showed a significant reduction of serum glucose levels.
In tr od u ction . The contribution of hepatic gluconeo-
genesis to the fasting and post-absorptive hyperglycemia
of type II diabetes is well established.1 This appears to
be related to the lack of suppressive effects of both
Recent studies have reported on new carnitine de-
rivatives as reversible CPT I inhibitors and good hy-
poglycemic agents.8,9 These compounds were designed
as transition state mimics, assuming a mechanism
similar to that reported for the acyl transfer with the
highly homologous enzyme carnitine acetyltransferase.
Reversible CPT I inhibition, selective for L-CPT I versus
M-CPT I isoform, resulted in an improved approach for
pharmacological intervention on fatty acid oxidation,
with indirect glycemic control without induction of
myocardial hypertrophy.8
* To whom correspondence should be addressed. F. Giannessi:
phone, 39-06-91393612; fax, 39-06-91393638; e-mail, fabio.giannessi@
sigma-tau.it. A. Arduini: phone, 39-06-91393843; fax, 39-06-91393988;
e-mail, arduino.arduini@sigma.tau.it.
† Department of Chemical Research, Sigma Tau Pharmaceutical
Industries S.p.A.
‡ Department of Endocrinology & Metabolism, Sigma Tau Pharma-
ceutical Industries S.p.A.
§ University of L’Aquila.
| S.IN-Soluzioni Informatiche S.a.s.
Following this general strategy, and considering that
no crystallographic data are available for CPT I, we
Department of Analytical Research, Sigma Tau Pharmaceutical
Industries S.p.A.
10.1021/jm010889+ CCC: $20.00 © 2001 American Chemical Society
Published on Web 06/19/2001

2384 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 15
Letters
Sch em e 1a
F igu r e 1. General formula of racemic compounds; X, Y, and
Z groups were chosen on the basis of their chemical diversity
and synthetic accessibility. X: trimethylammonium, trimeth-
ylphosphonium, pyridinium, quinuclidinium, guanidinium.
Y: carboxylate, phosphonate, tetrazolate. Z: oxygen, amino,
ureido, thioureido, carbamate, carbonate, sulfonamide, sulfa-
mide. R: C8-C10 and C14 linear alkyl.
F igu r e 2. Structures of compounds 17-20.14
reasoned that carnitine analogues, with medium to long
alkyl chains, could occupy the catalytic site of the
enzyme, thus slowing down the rate of conversion of
long-chain acyl-CoA into long-chain acylcarnitines.10,11
Thus, we have designed, synthesized, and evaluated the
pharmacological properties of a series of racemic car-
nitine related compounds as potential reversible CPT I
inhibitors.
Resu lts a n d Discu ssion . Figure 1 shows the general
formula of carnitine related compounds, where X, Y, and
Z groups were chosen on the basis of their chemical
diversity and synthetic accessibility. Only 18 compounds
were selected of the possible 270 combinations, also
employing the experimental D-optimal design ap-
proach,12 and finally synthesized.13 The effects of chiral-
ity were also investigated for a few selected compounds.
The structures of the compounds as well as the synthetic
procedures are summarized in Figure 2 and in Schemes
1-4.
As reported in Scheme 1, aminocarnitine derivatives
have been synthesized, in the case of compounds 2 and
3, from racemic aminocarnitine 1,14 or from its isobutyl
ester 4 for compounds 5-9. The synthesis of the carni-
tine analogues 11 and 13, obtained from (R,S)-4-iodo-
3-hydroxybutyrate esters 1015 and 12,16 and of com-
pounds 15 and 16 starting from racemic carnitine per-
chlorate benzyl ester 14,17 is reported in Scheme 2. The
structures of compounds 17-20 are reported in Fig-
ure 2.18 Starting from epibromhydrine 21, derivatives
23-25 have been obtained as illustrated in Scheme 3.
IC50 (µM) values for L-CPT I and M-CPT I are
reported in Table 1. M-CPT I inhibition was determined
only for compounds characterized by IC50 < 25 µM in
rat liver mitochondria. It should be noted that the
M-CPT I isoform is expressed in both cardiac and
skeletal muscle. Aminocarnitine derivatives (X ) trim-
ethylammonium, Y ) carboxylate) with Z ) carbamate
(6), ureido (2), sulfonamide (9), and sulfamide (5) were
shown to be the most active compounds. The IC50 values
for the liver isoform were better than, or comparable
to, that of the reference compound SDZ-CPI 975,9 with
a
Reagents, conditions, and yields: (a) 2 CH3(CH2)8NCO, DMSO,
40 °C, 60 h, 68%; (b) 2 CH3(CH2)8NCS, DMSO, 40 °C, 60 h, 53%;
(c) HCl(g) in isobutyl alcohol, 130 °C, 18 h, 95%; (d) 3 SO2Cl2, 4
NEt3, CH2Cl2, 3 h, then 2 NEt3, 2 CH3(CH2)8NH2, 18 h; (e) 1 N
NaOH, 18 h, 50% (from step d); (f) 2.5 CH3(CH2)9SO2Cl, 5.5
NEt3,CH2Cl2, 72 h; (g) 1 N NaOH, 18 h, 44% (from step f); (h) 1.5
CH3(CH2)7OCOCl, 9 NEt3, CH2Cl2, 4.5 h, 50%; (i) IRA 402 resin
(OH-), 94%; (j) 1.1 R1CHO, MeOH, CH3COOH, Pd/C H2 (30 psi),
18 h, 47% (R1 ) CH3(CH2)6 for 7, and R1 ) CH3(CH2)12 for 8); (k)
IRA 402 resin (OH-), 70% (R ) CH3(CH2)7 for 7 and R )
CH3(CH2)13 for 8).
selectivity for heart versus liver isoform (M-CPT I/L-
CPT I IC50 ratio) ranging from 1.4 (compound 9) to 9.5
(compound 6). The ureido derivative 2 showed a 4.2
selectivity, while for SDZ-CPI 975 a 3.6 value was found.
When X ) trimethylammonium and Y ) carboxylate
(aminocarnitine and carnitine derivatives), the great
influence of the Z group became clear by comparing the
inhibitory activity of the two carbamate moieties NH-
COO (6, IC50 ) 22,3 µM) and OCONH (16, IC50 > 300
µM), of the two ureido (Z ) NHCONH, 2, IC50 ) 19.5
µM) and thioureido moieties (Z ) NHCSNH, 3, IC50
122 µM), and of carbonate too (15, IC50 > 300 µM).
)
On the grounds of these encouraging results, we
decided to investigate the effects of chirality and alkyl
chain length on the most active L-CPT I inhibitors,
namely sulfonamide 9 and sulfamide 5. Thus, starting
from aminocarnitine of the corresponding configura-
tion14 (Scheme 4), we have synthesized derivatives 26
and 27, the S and R sulfonamide compounds (R ) C12),
respectively, and 28 and 29, the S and R sulfamide
compounds, respectively (R ) C11). Table 2 shows that
the R derivatives 27 and 29 are more effective L-CPT I
inhibitors than the S ones (compounds 26 and 28), even
by comparison with their original racemic derivatives.
In addition, an improvement of selectivity of R forms
toward the L-CPT I isoform was also observed. The
M-CPT I/L-CPT I IC50 ratios were 4.8 (27) and 7.2 (29).
Even if these results were not totally unexpected, they
confirm that R derivatives are more active than S ones
with all the examined functional groups, and that the

Letters
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 15 2385
Sch em e 2a
Ta ble 1. Effect of CPT I Inhibitors on Mitochondrial CPT I
Activitya
b
b
chain length
IC50
IC50
compd
(carbon atoms)
liver (µM)
heart (µM)
2
3
9
9
19.5
122
83.2
5
9
2.3
3.8
6
7
8
9
11
13
15
16
17
18
19
20
23
24
25
8
14
8
10
9
14
9
9
8
14
14
14
9
22.3
230
211
>300
1.6
2.3
>300
>300
>300
>300
>300
47
>300
not determined (insoluble)
>300
>300
>300
9
14
14
SDZ-CPI9757,8
17.4
62
a
The activity of CPT 1 was measured radiochemically using
14C]palmitoylCoA and carnitine itself by evaluating the incorpora-
[
tion of [14C]palmitoyl residue on carnitine in intact fresh liver and
b
heart mitochondria from Sprague-Dawley male rats.21,22 The
a
IC50 values (µM) were calculated from the inhibition curve. Data
are means of two determinations in triplicate. The variation
between experiments was less than 8%.
Reagents, conditions, and yields: (a) 1.2 (Me)3P, THF, 120 h,
71%; (b) 2 CH3(CH2)8NCO, DMF, 110 °C, 7 d, 34%; (c) 1 N HCl,
70 °C, 3 h; (d) Amberlyst A-21, 49% (from step c); (e) 1 quinuclidine,
CH3CN, 60 °C, 20 h, 50%; (f) 1.5 CH3(CH2)13OCOCl, 1.5 DMAP,
CH2Cl2, 20 h; (g) Amberlyst A-21 (HCl form), 58% (from step f);
(h) trifluoroacetic acid, 1 h; (i) IRA 402 (Cl-), SiO2 gel flash
chromatography, 55% (from step h); (j) 1 DMAP, 1 CH3(CH2)8OCO-
Cl, DMF, 72 h; (k) A-21 resin (HCl form), 32% (from step j); (l)
Pd/C 10%, H2 (47 psi), MeOH, 2 h, 68%; (m) 2 C9H19NCO, toluene,
reflux, 120 h, then 0.5 C9H19NCO, toluene, reflux, 120 h, 38%; (n)
Pd/C 10%, H2 (47 psi), 4 h, MeOH; (o) Amberlyst A-21 resin, 88%
(from step n).
Sch em e 4a
Sch em e 3a
a
Reagents, conditions, and yields: (a) 2.5 CH3(CH2)11SO2Cl, 5.5
NEt3, CH2Cl2, 72 h; (b) 1 N NaOH, 18 h, 40% (from step a); (c) 3
SO2Cl2, 4 NEt3, CH2Cl2, 3 h, then 2 NEt3, 2 CH3(CH2)10NH2,18 h;
(d) 1 N NaOH, 18 h, 35% (from step c).
Ta ble 2. Effect of Sulfonamidic and Sulfamidic Aminocarnitine
Derivatives as Inhibitors of Mitochondrial CPT I Activity
b
b
chain length
IC50
IC50
compd
chirality
(carbon atoms) liver (µM) heart (µM)
26
27
28
29
S
R
S
R
12
12
11
11
6
0.7
6
>10
3.4
>10
5.8
0.8
a
The IC50 values (µM) were calculated from the inhibition curve.
Data are means of two determinations in triplicate. The variation
between experiments was less than 8%.
27 is a reversible, mixed inhibitor of L-CPTI with
respect to palmitoyl-CoA (apparent Ki ) 0.25 ( 0.01
µM).
a
Reagents, conditions, and yields: (a) 1 HPO(OBn)2, 1 BuLi, 1
BF3Et2O, THF, -70 °C, 3 h, 60%; (b) 2 CH3(CH2)8NCO, 1.25
BF3Et2O, CH2Cl2, 30′, 85%; (c) trimethylamine (gas), TBAI cat.
DMF, 50 °C, 24 h, 25%; (d) Pd/C 10%, H2 (60 psi), MeOH, 18 h,
99%; (e) 2 CH3(CH2)8NCO, 1.25 BF3Et2O, CH2Cl2, 30′, 85%; (f)
TBAI cat., pyridine, 50 °C, 24 h, 20%; (g) 2 (CH3)3SiI, CH2Cl2, 30′,
then H2O, IRA 402 (Cl-), 80%; (h) 1 CH3(CH2)13OCOCl, 1 DMPA,
2 TEA, CHCl3, 24 h, 75%; (i) 2 quinuclidine, TBAI cat., DMF, 50
°C, 24 h, 15%; (j) Pd/C 10%, H2 (60 psi), MeOH, 18 h, 99%.
An initial in vivo test designed to evaluate the
efficiency of a CPT I inhibitor can be conveniently
carried out by determining plasma â-hydroxybutyrate
(BHB) levels in fasted animals. Since BHB is a major
product of liver fatty acid oxidation during fasting,19
a
liver specific CPT I inhibitor would significantly reduce
circulating BHB levels. According to the in vitro findings
on L-CPT I inhibition, when compound 27 was orally
longer alkyl chains seem to induce stronger inhibitory
capability. Furthermore, a kinetic study revealed that

2386 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 15
Letters
Villhauer, E. B. Antidiabetic Agents: A New Class of Reversible
Carnitine Palmitoyltranferase I Inhibitors. J . Med. Chem. 1995,
38, 3448-3450.
administered to 24 h fasted normal rats, a reduction of
plasma BHB at 6 h postdose with an ED50 of 20 mg/kg
occurred. This finding prompted us to test the antidia-
betic activity of 27 on db/db mice, a model of type II
diabetes.20 The oral administration of 27 for 30 days at
a dose of 100 mg/kg daily caused a significant reduction
of serum glucose levels with respect to untreated db/db
mice (528 ( 45.8 vs 734 ( 47.6 mg/dL; mean ( SE; p <
0.01). In addition, such a treatment did not affect either
body weight or circulating concentrations of insulin,
triglyceride, urea, alanine aminotransferase, and cho-
lesterol compared to controls. Moreover, cardiac and
liver sizes (heart and liver/body weight ratios) were
similar in both experimental groups.
Further investigation will be devoted to explore the
effects of chain length and chirality, in terms of activity
and selectivity, in the case of different functional groups
(for example, ureidic and carbamic ones) and the in vivo
effects of selected compounds as antichetotic and po-
tential antidiabetic agents in several animal models.
(10) J enkins, D. L.; Griffith, O. W. Antiketogenic and Hypoglycemic
Effects of Aminocarnitine and Acylaminocarnitines. Proc. Natl.
Acad. Sci. U.S.A. 1986, 83, 290-294.
(11) Savle, P. S.; Pande, S. V.; Lee, T. S.; Gandour, R. D. Stereoiso-
meric Acylamidomorpholinium Carnitine Analogues: Selective
Inhibitors of Carnitine Palmitoyltransferase I and II. Bioorg.
Med. Chem. Lett. 1999, 3099-3102.
(12) We calculated the following descriptors: 3D autocorrelograms
(Bin r ) ∑(xi)(xj)), where x are properties of atoms i and j spaced
by between r - 1 and r Å; the properties are vertex order
(topological index), positive and negative charge for X and Y
groups and vertex order, atomic volume, HB donor and acceptor
indexes, and numbers of N and O atoms for Z group. The Bin r
values, truncated to the minimum common distance, were
submitted to a principal component analysis (PCA), and two PC
for X and Y descriptions and three PC for Z description were
extracted (which globally explain 87.7, 95.8, and 86.2%, respec-
tively, of the total variance) (software TSAR, OMG). The
resulting PC’s descriptors for the three groups were processed
with a D-optimal experimental design to allow the selection of
the molecules (software DESDOP, MIA).
(13) Giannessi, F.; Marzi, M.; Minetti, P.; Tinti, M. O.; De Angelis,
F.; Chiodi, P.; Arduini, A. Compounds Having Reversible
Inhibiting Activity of Carnitine Palmitoyl-Transferase. WO99/
59957, 1999; Chem. Abstr. 1999, 131, 351673.
(14) Castagnani, R.; De Angelis, F.; De Fusco, E.; Giannessi, F.;
Misiti, D.; Meloni, D.; Tinti, M. O. Stereospecific Synthesis of
(R)-Aminocarnitine (Emeriamine) Starting from (R)-Carnitine
via Double Inversion of Configuration. J . Org. Chem, 1995, 60,
8318-8319. Racemic aminocarnitine was prepared starting from
racemic carnitine.
Ack n ow led gm en t. The authors thank Luigi Al-
monte, Giuseppina Calvisi, Roberto Catini, Enrico De
Fusco, Isabella Lustrati, and Ornella Petrella for tech-
nical assistance.
(15) Larcheveque, M.; Henrot, S. Enantiomerically Pure â, γ-Ep-
oxyesters from â-Hydroxylactones: Synthesis of â-Hydroxyesters
and (-)-GABOB. Tetrahedron, 1990, 46, 4277-4282. Racemic
â-hydroxy-γ-butyrolactone was prepared from R, S-carnitine:
Calvisi, G.; Catini, R.; Chiariotti, W.; Giannessi, F.; Muck, S.;
Tinti, M. O.; De Angelis, F. Single Step Conversion of Chiral
Carnitine and Derivatives into (S)- and (R)-â-Substituted-γ-
Butyrolactones. Synlett 1997, 1, 71-74.
Su p p or tin g In for m a tion Ava ila ble: Analytical data are
available free of charge via the Internet at http://pubs.acs.org.
Refer en ces
(1) DeFronzo, R. A.; Bonadonna, R. C.; Ferranini, E. Pathogenesis
of NIDDM: a balanced overview. Diabetes Care 1992, 15, 318-
368.
(16) Boots, S. B.; Boots, M. R. New Synthesis of (R, S)-Carnitine
Chloride. J . Pharm. Sci. 1975, 64, 1262-1264.
(2) Golay, A.; Swislocki, A. L. M.; Chen, Y. D.; Reaven, G. M.
Relationships between plasma-free fatty acid concentration,
endogenous glucose production, and fasting hyperglycemia in
normal and noninsulin-dependent diabetic individuals. Metabo-
lism 1987, 36, 692-696.
(17) Brendler, K.; Bressler, R. The Resolution of (()-Carnitine and
the synthesis of Acylcarnitine. Biochim. Biophys. Acta 1967, 137,
98-106. The transformation into the soluble perchlorate form
was realized in quantitative yield by means of IRA-402 (ClO4-).
(18) The preparation of compounds 17-20 is part of a more compre-
hensive study on carnitine ether analogues, which will be
reported in a future publication.
(3) McGarry, J . D.; Brown, N. F. The mitochondrial carnitine
palmitoyltransferase system. From concept to molecular analy-
sis. Eur. J . Biochem. 1997, 244, 1-14.
(4) Attwood, P. A.; Keech, D. B. Pyruvate carboxylase. Curr. Top.
Cell Regul. 1984, 23, 1-55.
(19) Zammit, V. Regulation of ketone body metabolism. Diabetes Rev.
1994, 2, 132-154.
(5) Foley, J . E. Rationale and Application of Fatty Acid Oxidation
Inhibitors in Treatment of Diabetes Mellitus. Diabetes Care
1992, 15, 773-784.
(20) Kodama, H.; Fujita, M.; Yamaguchi, I. Development of hyper-
glycaemia and insulin resistance in conscious genetically diabetic
(C57BL/KsJ db/db) mice. Diabetologia 1994, 37, 739-744.
(21) Kerner, J .; Bieber, L. Isolation of a malonyl-CoA sensitive CPT/
beta-oxidation enzyme complex from heart mitochondria. Bio-
chemistry 1990, 29, 4326-4334.
(6) Anderson, R. Carnitine Palmitoyltransferase: A Viable Target
for the Treatment of NIDDM? Curr. Pharm. Des. 1998, 4, 1-15.
(7) Bressler, R.; Gay, R.; Copeland, J . G.; Bahl, J . J .; Bedotto, J .;
Goldman, S. Chronic inhibition of fatty acid oxidation: New
model of diastolic dysfunction. Life Sci. 1989, 44, 1897-1906.
(8) Deems, R. O.; Anderson, R. C.; Foley, J . E. Hypoglycemic effects
of a novel fatty acid oxidation inhibitor in rats and monkeys.
Am. J . Physiol. 1998, 274, R524-R528.
(22) Hovius, R.; Lambrescts, H.; Nicolay, K.; De Kruijff, B. Improved
methods to isolate and subfractionate rat liver mitochondria.
Lipid composition of the inner and outer membrane. Biochim.
Biophys. Acta 1990, 1021, 217-226.
(9) Anderson, R. C.; Balestra, M.; Bell, P. A.; Deems, R. O.; Fillers,
W. S.; Foley, J . E.; Fraser, J . D.; Mann, W. R.; Rudin, M.;
J M010889+