Derivatives of R-aminocarnitine without ammonium moiety as liver carnitine palmitoyltransferase I (L-CPTI) Inhibitors

Full text of DOI:10.1002/cmdc.201100289

DOI: 10.1002/cmdc.201100289
Derivatives of R-Aminocarnitine without Ammonium Moiety as Liver
Carnitine Palmitoyltransferase I (L-CPT I) Inhibitors
Emanuela Tassoni,*[a] Roberto Conti,[b] Grazia Gallo,[a] Silvia Vincenti,[a] Lucilla Mastrofrancesco,[a]
Tiziana Brunetti,[a] Walter Cabri,[a] and Fabio Giannessi*[b]
Type II diabetes is a chronic and progressive metabolic disor-
der with potentially life-threatening consequences, usually aris-
ing due to resistance to insulin action in the setting of an inad-
equate compensatory insulin secretory response.[1] In type II
diabetes, increased hepatic fatty acid oxidation by formation
of high levels of acetyl-coenzyme A (CoA), ATP and NADH,
over-regulates gluconeogenesis and thus hepatic glucose pro-
duction.[2]
hyperketotic/antihyperglycemic activity. It was also reported
that inhibition of hypothalamic lipid oxidation via intracerebro-
ventricular infusion of ST1326 results in cellular accumulation
of long-chain fatty acyl-CoA within the arcuate nucleus of the
hypothalamus, leading to marked inhibition of food intake and
glucose production,[10] with a significant anorectic response
mirrored by an early occurrence of satiety onset.[11] Unfortu-
nately, the physicochemical properties of ST1326 and the other
previously synthesized betaine compounds were not very ap-
propriate for crossing the blood–brain barrier (BBB).
The transport of fatty acids into the mitochondria is regulat-
ed by two membrane-bound, carnitine-dependent, long-chain
acyltranferases, known as carnitine palmitoyltransferases (CPT I
and CPT II).[3] CPT I, the outer mitochondrial membrane
enzyme—present in two isoforms: liver (L-CPT I) and muscle
(M-CPT I)—catalyzes the formation of long-chain acylcarnitines.
CPT II, the inner mitochondrial membrane enzyme present in a
single isoform, reconverts long-chain acylcarnitines into long-
chain acyl CoA thioesters.
Here, we report the synthesis and biological characterization
of new derivatives of R-aminocarnitine, where the quaternary
ammonium moiety of the aminocarnitine backbone of the lead
compounds ST1326 and ST2425 is replaced by groups without
permanent cationic charges.[12] Such compounds, with tertiary
amino or alkyl groups, were designed to increase the ability of
the agent to cross the BBB with the intention of achieving CPT
inhibition in the central nervous system (CNS), with the aim of
overcoming the limitations explained above. Owing to the lack
of the quaternary ammonium cation, the new compounds
were also expected to be less tensioactive; however, their bio-
logical activity, liver/heart selectivity and CPT II inhibition re-
mained to be evaluated.
Oxirane carboxylates, such as etomoxir and methyl 2-tetra-
decylglycidate,[4] were reported to be irreversible inhibitors of
CPT systems, exhibiting good activity as hypoglycemic agents,
but at the time they were discontinued probably owing to car-
diac hypertrophy due to a lack of selectivity toward the L-CPT I
isoform.[5] More recently, this assumption has been questioned
since inhibition of the ubiquitously present CPT II isoform has
been put forward as a therapeutic target for the treatment of
diabetes and associated pathologies.[6]
Amino analogues 8–11 were prepared from starting material
1 according to known procedures (Scheme 1).[13] When the ap-
propriate isocyanate used in step d was not commercially
available, it was prepared from the corresponding carboxylic
acid via a Curtius reaction; 4-[(3-hexyloxy)phenoxy]butylisocya-
nate is not commercially available and so was prepared as re-
ported in the patent literature.[9b] Using the same starting ma-
terial (1), tert-butyl analogues were also synthesized following
the procedure described in Reference [10], where nucleophilic
substitution of 1 was carried out by adding lithium di-tert-
butyl cuprate (Scheme 2). Compounds 15 and 16 were then
obtained in about 70% yield from ester 14 after reaction with
the appropriate isocyanate and subsequent acid hydrolysis.
Compounds 17 and 18, where the ammonium group of the
aminocarnitine backbone is replaced by an isopropyl group,
were synthesized from the commercially available l-b-homo-
leucine hydrochloride and the appropriate isocyanate in 80%
and 28% yield, respectively (Scheme 3).
We previously published the identification of selective L-
CPT I inhibitor ST1326 (teglicar)[7,8] and the backup analogue
ST2425,[9] both with carnitine-related backbones, and their anti-
[a] Dr. E. Tassoni, Dr. G. Gallo, Dr. S. Vincenti, Dr. L. Mastrofrancesco,
Dr. T. Brunetti, Dr. W. Cabri
For in vitro efficacy and selectivity evaluation, IC50 values
against L-CPT I were determined in rat liver mitochondria for
all new compounds. The most active derivatives were further
investigated for their inhibition of M-CPT I using heart mito-
chondria. Furthermore, these same derivatives, i.e. those that
exhibited selectivity for the liver isoform, were assayed for
their ED50 values for ketogenesis inhibition in hepatocytes;
since b-hydroxybutyrate (b-HBA) is the main product of liver
Chemistry & Analytical Department, Sigma-Tau S.p.A.
Via Pontina Km 30.400, 00040 Pomezia (Italy)
E-mail: emanuela.tassoni@sigma-tau.it
[b] Dr. R. Conti, Dr. F. Giannessi
Endocrinology & Metabolism Department, Sigma-Tau S.p.A.
Via Pontina Km 30.400, 00040 Pomezia (Italy)
E-mail: fabio.giannessi@sigma-tau.it
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201100289.
ChemMedChem 2011, 6, 1977 – 1980
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1977

MED
fatty acid oxidation during fast-
ing, a CPT I inhibitor is expected
to significantly reduce its level in
vitro and in vivo. Finally, CPT II
inhibition in rat liver mitochon-
dria was also evaluated. The re-
sults of the biological evalua-
tions are given in Table 1.
The experimental data show
that the highest inhibitory activi-
ty in liver mitochondria was ex-
hibited by compound 8 (IC50 =
0.63 mm), which has a “nor-ami-
nocarnitine” structure with the
resorcinol in the side chain,
while the straight-chain ana-
logue 10, a nor derivative of te-
glicar, exhibited a more modest
activity (IC50 =5.7 mm). Similarly,
the resorcinol-substituted benzyl
derivative 9 exhibited an IC50
value of 2.1 mm. Complete loss
or a significant reduction in in-
Scheme 1. Synthesis of 8–11. Reagents and conditions: a) NHRMe, MeOH, 18h, RT, 2: 72%, 3: 79%; b) PhOH, aq
HBr, 18h, 1308C, 4: 95%, 5: 90%; c) SOCl2, anhyd MeOH, 18 h, RT, 6: 71%, 7: 96%; d) R1NCO, Et3N, MeOH, RT, 24–
48 h, yields for methyl ester of: 8: 88%, 9: 66%, 10: 77%, 11: 91%; e) aq HCl/THF (1:3), 48 h, RT, 8: 36%, 9: 66%,
10: 38%, 11: 21%.
hibitory activity resulted when the ammonium group of the
carnitine backbone was replaced with a tert-butyl moiety, as in
the straight-chain or resorcinol-substituted derivatives 15 and
16, respectively, or with the corresponding isopropyl group as
in analogues 17 and 18. It is worth noting that, in the pres-
ence of the evidently inadequate alkyl moieties, the resorcinol-
substituted chain is able to preserve some inhibitory effect,
further confirming the trend observed when the same structur-
al modification was introduced to obtain ST2425 from
ST1326.[9]
For the most active compounds 8, 9 and 10, inhibition of M-
CPT I evaluated in hearth mitochondria evidenced their very
good selectivity for the liver isoform, especially for derivatives
8 and 9 (Table 1). No or weak (ꢀ20% at 20 mm) CPT II inhibition
was evidenced for 8 and 9, respectively, by measuring the
oxygen consumption of rat liver mitochondria in the presence
of l-palmitoylcarnitine as described in the Supporting Informa-
tion. In ketogenesis and glucose production in hepatocytes,
the determined IC50 values confirmed good efficacy, without
modification of hepatocytes ATP content, with compound 8
exhibiting the most interesting profile.
Scheme 2. Synthesis of 15 and 16. Reagents and conditions: a) (tBu)2CuLi,
THF, ꢀ408C, 6 h, 12: 52%; b) PhOH, aq HBr, 18 h, 1308C, 13: 73%; c) SOCl2,
anhyd MeOH, 18 h, 408C, 14: 93%; d) R1NCO, Et3N, MeOH, RT, 24–48 h, yield
for methyl ester of: 15: 81%, 16: 73%; e) aq HCl/THF (1:3), 48 h, RT, 15:
72%, 16: 68%.
Furthermore, as tensioactivity is an important parameter and
an element of potential concern for compounds conceived for
oral administration, the critical micellar concentration (CMC)
was measured for compounds 8 and 9 and compared to that
of the parent ammonium analogue ST2425 (in the case of
ST2425, the CMC was determined in water, while for 8 and 9 it
was measured in 0.01n hydrochloric acid owing to lack of sol-
ubility; the lower the CMC, the higher the tensioactivity). For
compound 8, a CMC value greater than 2ꢁ10ꢀ3 m was found,
which is higher than that of ST2425 (7.96ꢁ10ꢀ4 m). For methyl-
benzylamino derivative 9, a more similar value of 5.4ꢁ10ꢀ4
m
Scheme 3. Synthesis of 17 and 18. Reagents and conditions: a) R1NCO, aq
NaOH, 08C!RT, 4 h, 17: 80%, 18: 28%.
was recorded.
1978
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 1977 – 1980

Table 1. Docking scores, theoretical LogBB values and preliminary biological data against L-CPT I for all new compounds, ST1326 and ST2425. Selected de-
rivatives were further evaluated against M-CPT I and in assays to determine b-hydroxybutyrate production and gluconeogenesis in hepatocytes. For com-
pound 8, ST1326 and ST2425, ED50 values against b-hydroxybutyrate were determined in CD1 mice.
Compd
Docking score
LogBB
IC50[a] [mm]
b-hydroxybutyrate
ED50 [mgkgꢀ1
]
b-hydroxybutyrate
in CD1 (gavage)[c]
L-CPT I
M-CPT I
Gluconeogenesis
in hepatocytes[b]
in hepatocytes
ST1326 (teglicar)
ST2425
8
9
10
11
15
16
17
18
ꢀ10.2
ꢀ11.6
ꢀ11.1
ꢀ13.1
ꢀ9.0
ꢀ9.2
ꢀ7.8
ꢀ9.6
ꢀ7.8
ꢀ9.0
ꢀ1.01
ꢀ1.50
ꢀ0.60
ꢀ0.39
ꢀ0.19
0.03
0.16
ꢀ0.33
0.09
0.68ꢁ0.13
0.13ꢁ0.03
0.63ꢁ0.05
2.10ꢁ0.98
5.70ꢁ0.81
9.70ꢁ1.22
50.1ꢁ5.9
0.77ꢁ0.2
3.19ꢁ1.0
12.5
4.7
11.1
–
–
–
–
–
–
–
5.44ꢁ0.3
0.044ꢁ0.001
0.29ꢁ0.025
60ꢁ1.4
0.40ꢁ0.1
0.83ꢁ0.1
118ꢁ2.3
1.0ꢁ0.16
3.42ꢁ0.6
70ꢁ1.6
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
>200
70.06ꢁ2.30
>200
66.60ꢁ1.72
ꢀ0.41
[a] IC50 values represent the mean ꢁSD of three determinations performed in duplicate; – not determined; [b] Gluconeogenesis is defined and calculated
as the b-oxidation-dependent fraction; [c] n=4 animals per dose.
To confirm inhibition of liver CPT I in vivo, the most interest-
ing derivative (8) was orally administered to 24 h fasted CD1
mice. Compound 8 showed good activity in this animal model;
the ED50 value calculated from the results was approximately
11 mg kgꢀ1, comparable to that of ST1326 (12.5 mgkgꢀ1) and
only slightly higher than ST2425 (4.7 mgkgꢀ1).
A previously proposed three-dimensional structural model
for L-CPT I[9a] was used for docking studies of the novel deriva-
tives (Figure 1; see Supporting Information for further discus-
sion and details). The results (GScore), reported in Table 1,
allow the classification of the ligands in agreement with their
activity data, with the most active compounds giving GScore
values less than ꢀ10. Regarding the lower GScore value calcu-
lated for compound 9 compared with the more active com-
pound 8, it should be considered that 9 has a predicted pKa
value of 8.20, while compound 8 is predicted to have a pKa
value of 9.16. This property, indicating that the amino group of
8 is ionized in a greater proportion at physiological pH com-
pared with compound 9 (and therefore, is more ammonium-
like), should give compound 8 an advantage in terms of enzy-
matic interaction, resulting in higher potency in vitro and in
vivo.
The potential ability of the new structures to cross the BBB
was evaluated by theoretical calculations of the brain/blood
concentration ratio (LogBB; Table 1). From these calculations,
all compounds are expected to exhibit slightly higher penetra-
tion of the CNS compared with the lead compounds, with de-
rivative 8 predicted to show approximately a 2.5-fold increase
in brain/blood concentration ratio with respect to ST1326, and
eightfold with respect to the more structurally related ST2425.
The greatest increase (~15-fold over ST1326) was predicted for
inactive compound 15, while for derivative 9, approximately a
fourfold increase was calculated versus ST1326, and a 13-fold
increase versus ST2425. The structural characteristics of com-
pound 8 could, therefore, give rise to an improved balance be-
tween L-CPT I inhibitory efficacy and the brain/blood concen-
tration ratio.
In conclusion, we have reported the identification of new
CPT I inhibitors with the ammonium group of aminocarnitine
replaced by amino or alkyl moieties, in order to avoid the pre-
viously present permanent cationic charge and improve CNS
penetration without loss of activity and selectivity. The nor de-
Figure 1. The hydrophobic (orange)/hydrophilic (turquoise) profile of complexes a) L-CPT I–8 and b) L-CPT I–ST1326, and c) protein L-CPT I alone. The protein
is shown as purple ribbon; protein residues are shown as sticks (grey); the inhibitor is shown as ball and stick (green); heteroatoms are indicated: oxygen
(red), nitrogen (blue).
ChemMedChem 2011, 6, 1977 – 1980
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1979

MED
rivative 8 was selected for further evaluation because of its effi-
cacy on L-CPT I; using molecular modeling, the dimethylamino
group was predicted to maintain adequate interactions in the
catalytic site of the CPT I enzyme, and the resorcinol group in
the side chain assisted its activity as already observed in the
second generation compound ST2425 with respect to
ST1326.[9] Compound 8 also exhibited good selectivity for L-
CPT I and good efficacy both in vitro and in vivo in terms of
antiketogenic activity, as well as of glucose production in hep-
atocytes. A favorable CMC value was also determined with re-
spect to ST2425. The ability of compound 8 to cross the BBB
was predicted to be approximately 2.5-fold better than that of
ST1326, and about eightfold better than that of the more
structurally similar ST2425. Further investigation in mice
models of type II diabetes, together with determination of the
effects on food intake and experimental BBB penetration, are
needed to better assess the profile of this newly identified L-
CPT I inhibitor.
D’Amore and Dr. Marco Quaglia for analytical support, and Ales-
sandro Peschechera, Dr. Rita Ricciolini and Dr. Pierino Chiodi for
support with data elaboration and statistical evaluation.
Keywords: aminocarnitine analogues · antidiabetic agents ·
carnitine palmitoyltransferase I (CPT I) · inhibitors · teglicar
[1] a) J. E. Gerich, Mayo Clin. Proc. 2003, 78, 447–456; b) S. E. Kahn, Diabeto-
logia 2003, 46, 3–19.
[2] R. A. DeFronzo, R. C. Bonadonna, E. Ferrannini, Diabetes Care 1992, 15,
318–368.
[3] J. D. McGarry, N. F. Brown, Eur. J. Biochem. 1997, 244, 1–14.
[4] J. E. Foley, Diabetes Care 1992, 15, 773–784.
[5] R. C. Anderson, Curr. Pharm. Des. 1998, 4, 1–16.
[6] M. Andjelkovic, S. M. Ceccarelli, O. Chomienne, G. Lewis Kaplan, P.
Mattei, J. Wright Tilley, (Hoffmann–La Roche, Nutley, USA), US2009/
0270505 A1, 2009; [CAN 151:508869].
[7] F. Giannessi, P. Pessotto, E. Tassoni, P. Chiodi, R. Conti, F. De Angelis, N.
Dell’Uomo, R. Catini, R. Deias, M. O. Tinti, P. Carminati, A. Arduini, J. Med.
Chem. 2003, 46, 303–309.
[8] R. Conti, E. Mannucci, P. Pessotto, E. Tassoni, P. Carminati, F. Giannessi,
A. Arduini, Diabetes 2011, 60, 644–651.
Experimental Section
[9] a) E. Tassoni, R. Conti, G. Gallo, S. Vincenti, N. Dell’Uomo, L. Mastrofran-
cesco, R. Ricciolini, W. Cabri, P. Carminati, F. Giannessi, ChemMedChem
2010, 5, 666–669; b) F. Giannessi, E. Tassoni, N. Dell’Uomo, R. Conti,
M. O. Tinti, (Sigma-Tau S.p.A., Rome, Italy), WO2008/015081 A1, 2008;
[CAN 148:238879].
Authorization for the use of animals at Sigma-Tau laboratories was
obtained from the Italian Health Authority. The care and husbandry
of animals are in accordance with European Directives 86/609 and
with the Italian Regulatory system (D.L. no. 116, Art. 6, January 27,
1992). Experimental details and analytical data for the synthesis of
intermediates and final products are reported in the Supporting In-
formation along with experimental protocols for molecular model-
ing, critical micellar concentration determination, calculation of
theoretical physicochemical properties, and in vitro and in vivo
biological assays.
[10] S. Obici, Z. Feng, A. Arduini, R. Conti, L. Rossetti, Nat. Med. 2003, 9,
756–761.
[11] R. Coccurello, A. Caprioli, S. Bellantuono, F. R. D’Amato, R. Conti, F. Gian-
nessi, F. Borsini, A. Moles, Psychopharmacology 2010, 210, 85–95.
[12] F. Giannessi, E. Tassoni, M. O. Tinti, R. Conti, N. Dell’Uomo, T. Brunetti,
(Sigma-Tau S.p.A., Rome, Italy), WO/2006/092204 A1, 2006; [CAN
145:293349].
[13] a) C. W. Jefford, J. McNulty, Helv. Chim. Acta 1994, 77, 2142–2146;
b) C. W. Jefford, J. McNulty, L. Zhi-Hui, J. B. Wang, Helv. Chim. Acta 1996,
79, 1203–1216.
Acknowledgements
The authors wish to thank Emanuele Pani, Giuseppina Calvisi,
Roberto Catini, Berardino Ciani, Rita Liberati, Loredana Magnani-
mi and Samanta Prati for technical assistance, Gianfranco
Received: June 6, 2011
Revised: July 22, 2011
Published online on August 10, 2011
1980
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 1977 – 1980