Thyroid receptor ligands. Part 8: Thyromimetics derived from N-acylated-α-amino acid derivatives displaying modulated pharmacological selectivity compared with KB-141

Article abstract of DOI:10.1016/j.bmcl.2007.05.049

Based on the scaffold of the pharmacologically selective thyromimetic 2b, structurally a close analog to KB-141 (2a), a number of novel N-acylated-α-amino acid derivatives were synthesized and tested in a TR radioligand binding assay as well as in a repor

Full text of DOI:10.1016/j.bmcl.2007.05.049

Bioorganic & Medicinal Chemistry Letters 17 (2007) 4131–4134
Thyroid receptor ligands. Part 8: Thyromimetics derived
from N-acylated-a-amino acid derivatives displaying
modulated pharmacological selectivity compared with KB-141
Neeraj Garg,a Yi-Lin Li,a Ana Maria Garcia Collazo,a Chris Litten,a Denis E. Ryono,b
Minsheng Zhang,d Yolanda Caringal,b Robert P. Brigance,b Wei Meng,b
William N. Washburn,b Peter Agback,a Karin Mellstro¨m,a Stefan Rehnmark,a
Mahmoud Rahimi-Ghadim,a Thomas Norin,a Marlena Grynfarb,a Johnny Sandberg,a
Gary Groverc and Johan Malma,*
aKaro Bio AB, Novum, Huddinge S-141 57, Sweden
bPharmaceutical Research Institute, Bristol-Myers Squibb, Princeton, NJ 08543-5400, USA
cDepartment of Pharmacology, Eurofins Scientific-PSL, 2394 Rt. 130, Dayton, NJ 08810, USA
dChemical and Screening Sciences, Wyeth Research, Princeton, NJ 08543-8000, USA
Received 9 April 2007; revised 15 May 2007; accepted 17 May 2007
Available online 21 May 2007
Abstract—Based on the scaffold of the pharmacologically selective thyromimetic 2b, structurally a close analog to KB-141 (2a), a
number of novel N-acylated-a-amino acid derivatives were synthesized and tested in a TR radioligand binding assay as well as in a
reporter cell assay. On the basis of TRb1-isoform selectivity and affinity, as well as affinity to the reporter cell assay, 3d was selected
for further studies in the cholesterol-fed rat model. In this model 3d revealed an improved therapeutic window between cholesterol
and TSH lowering but decreased margins versus tachycardia compared with 2a.
ꢀ 2007 Elsevier Ltd. All rights reserved.
R3
Hyperlipidemia is a well-established risk factor for ath-
erogenesis and cardiovascular disease, and thus one of
the leading causes for morbidity and mortality in the
developed world.1 The active endogenous thyroid hor-
mone (L-T3, 1) (Fig. 1) is a potent lipid lowering agent
but cannot be used therapeutically in patients with
hyperlipidemia, mainly due to the side-effect of tachy-
cardia. There exist two subtypes of thyroid hormone
receptors (TRs), a and b, unequally distributed in the
body. TRa is most abundant in the heart. Most effects
of 1 on the heart are mediated through TRa, while the
majority of actions in the liver are mediated through
TRb. The prospects for the treatment of metabolic dis-
eases such as dyslipidemia with TRb-selective ligands
would be enhanced by avoidance of cardiovascular
acceleration mediated through TRa.2 Moreover, as evi-
I
2'
5'
3
4
I
1' O
3'
O
2
1
NH2
CO2H
4'
5
CO2H
6'
HO
I
HO
R5
6
2a: R3,R5= Cl
1
2b: R3,R5= Br
Figure 1. Chemical structure of L-T3 (1) including ring-numbering,
and KB-141 (2a) and 2b.
dent from the older literature, tissue selectivity can also
contribute significantly to pharmacological selectivity.3
We have previously shown that TRb1-selectivity can be
achieved both in vitro and in vivo when the R1-amino
acid side-chain of 1 is truncated to acetic acid and the
iodine atoms at the R3 and R5-positions are substituted
with chlorine or bromine to generate KB-141 (2a) and
2b (Fig. 1). These compounds were 14 and 10 times
selective for TRb over TRa, respectively, ratios that
were further propagated in vivo.4
Keywords: KB-141; Thyromimetic; a-Amino acid; Binding assay; TRb-
selective; Cholesterol-fed rat model; Cholesterol; Thyroid stimulating
hormone; Tachycardia.
*
Corresponding author. E-mail: johanm@excite.com
0960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.05.049

4132
N. Garg et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4131–4134
Me
Br
3d, R= (S)-i-Pr
3a, R=H
O
3e, R= (R)-i-Pr
(a), (b)
(c), (d)
Me
Me
O
O
R
R
3b, R= (S)-Me
3c, R= (R)-Me
2b
3f, R=(S)-Phenyl
3g, R=(R)-Phenyl
HO
Br
3a-g
N
H
COOH
Me
Br
O
4a, R=(S)-CH2COOH
4b, R=(S)-CH2CH2OH
2b
HO
Br
4a-b
N
H
COOH
Me
Br
Me
Br
Me
Br
O
O
O
Me
HO
O
(h), (c), (i)
Me
(e), (f), (g)
Me
COOMe
COOMe
Br
N
COOH
MeO
Br
MeO
Br
H
F
F
5
7
6
Me
Br
Me
Br
O
O
(c), (h)
Me
HO
O
Me
HO
(j)
2b
COOH
Br
N
H
COOH
Br
Cl
Cl
9
8
Me
Br
Me
Br
O
O
(l), (c), (h)
(k)
(h)
Me
Me
2b
COOMe
COOMe
HO
Br
CHO
11
HO
Br
10
Me
Br
O
Me
O
HO
Br
N
COOH
H
Me
12
Scheme 1. Reagents and conditions: (a) EDCI, HBT, CH2Cl2 + carboxylic acid ester of amino acid, TEA, 40 ꢁC; (b) NaOH, EtOH, 40 ꢁC; (c) EDCI,
HBT, DMF + carboxylic acid ester of amino acid, DMF, rt; (d) NaOH, EtOH, rt; (e) HNO3, C6H6, rt; (f) sodiumhydrosulfite, EtOH, D; (g) NOBF4,
CH2Cl2, o-xylen, D; (h) LiOH, THF, rt; (i) BF3ÆSMe2, CH2Cl2, rt; (j) benzyltrimethylammonium tetrachloroiodate acetic acid, rt; (k) SnCl4, trioctyl
amine, toluene, rt + paraformaldehyde; (l) hydrazine hydrate, EtOH, rt + KOH, ethylene glycol, 150 ꢁC.
As 2a had a similar tissue-distribution as 1, its selectivity
for cholesterol lowering versus tachycardia could be
explained by virtue of its TRb1-selectivity.5 On the other
hand, TRb1 activation can be expected to reduce thyroid
stimulating hormone (TSH) levels and, therefore, reduce
the endogenous production of 1. This may be an issue in
tissues not accessible to TRb1-selective agonists, leading
to a paradoxical local tissue hypothyroidism.6 Indeed,
TSH was lowered in vivo with 2a, at doses even lower
than for cholesterol lowering.5 Consequently, in order
to avoid TSH-lowering induced by both 1 and 2a, a
selective agent may require some degree of liver selectiv-
ity (or lack of systemic effects) or, alternatively, selective
receptor modulation.
ics was prepared as outlined in Scheme 1 where various
a-amino acids were N-acylated with the R1-acetic acid
group of 2b by expedient coupling with protected amino
acids to generate the end products 3a–g and 4a and 4b i0n
low to moderate yields (18–52%). Introduction of R5 -
substituents was achieved by more elaborate routes:
R5 -F ð7Þ from intermediate 60 via a nitration–reduc-
0
tion–diazotization sequence; R5 -Cl ð9Þ from intermedi-
0
ate 8 via regioselective chlorination; and R5 -Me ð12Þ
from intermediate 11 via regioselective formylation fol-
lowed by reduction. Total yields for the preparation of
7, 9, and 12 were 8%, 18%, and 18%, respectively.
The results of a radioligand binding assay for the human
TRa1 and TRb1, as well as a reporter cell assay employ-
ing CHOK1-cells (Chinese hamster ovary cells) stably
transfected with hTRa1 or hTRb1 and an alkaline phos-
phatase reporter gene downstream thyroid response ele-
ment (TRAFa1 and TRAFb1),7 are summarized in
Table 1. All new ligands (3a–g, 4a, 4b, 7, 9, and 12)
exhibited weaker TRb1-binding (IC50s) as well as
TRAFb1 affinities (EC50s) compared with 2b. However,
in general TRb1-binding selectivity was conserved due to
a concomitant reduction in affinity for TRa1. The natu-
rally occurring amino acid stereochemistry (S-form) was
moderately favored in isomeric pairs 3b and 3c and 3d
In the present study, we wanted to investigate both the
in vitro and in vivo consequences when the R1-acetic
acid group was condensed with a-amino acids. As is evi-
dent from the literature, since the TR receptor is highly
accommodative of a variety of R1-side-chains, this flexi-
bility can be expected to encompass larger R1-chains
than acetic acid. Since the therapeutic ratio of 2b was
almost within the same range as 2a, we also decided to
maintain R3 and R5-position as bromines particularly
as the starting material could be prepared with facility
and in larger yields.4 Consequently, a set of thyromimet-

N. Garg et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4131–4134
4133
Table 1. Thyroid hormone receptor binding affinities (IC50), TRAF affinities (EC50), and efficacy of 1, and synthetic thyromimetics 2a, 2b, 4a–i, and
a
9a–c
b
b
c
d
d
Compound
hTRa1 IC50
hTRb1 IC50
a1/b1
0.54
14
TRAFa EC50
TRAFb EC50
1
0.24
0.26
1.1 0.029
—
1.7
—
0.78
2a
2b
3a
3b
3c
3d
3e
3f
3g
4a
4b
7
25 1.7
1.4 0.44
26 1.2
10
0.095 0.23
2.7 0.41
1.3
8.7
5.9
4.5
8.3
9.4
8.2
5.7
6.0
0.9
3.8
4.3
0.29
0.34
0.38
0.20
103 (75)
7.2 (81)
15 (83)
1.4 (78)
2.3 (88)
Nt
42 (85)
2.6 (92)
7.6 (100)
0.77 (92)
1.2 (100)
Nt
83
5.9
4.3 0.22
9.8 1.0
7.5
0.27 0.0018
0.71 0.067
0.78
5.5 0.29
460 38
168
0.60 0.070
290 23
26
1.5 (65)
1500 (72)
66 (74)
17 (78)
65 (21)
170 (13)
0.73 (85)
1300 (84)
38 (88)
11 (88)
210 (31)
1100 (27)
88 12
670 50
1500 73
12 0.50
1400 16
2600 240
9
12
a All IC50 and EC50 values are expressed as nM and the data for 1, 2a, and 2b are taken from Ref. 4.
b The IC50 values for 2a, 2b, 3a, 3d, 3e, 3g, 4a, 7, 9, and 12 are expressed as mean IC50
of two runs with variability of the measurements with an average of 25%.
s
SE, while IC50s for the other compounds are average means
c Normalized selectivity: (IC50 hTRa1)/(IC50 hTRb1 · 1.7). For an explanation, see Ref. 4.
d EC50 values are expressed as nM and are all calculated means of duplicate runs with an average variability of the measurements of approximately
25%. Values within parentheses denote % agonism.
and 3e; but no preference was observed for 3f and 3g.
Anyway, only the (S)-enantiomers of analogs 4a, 4b,
7, 9, and 12 were investigated. The introduction of addi-
tional polar groups in the amino acid chain, either H-
bond acceptors (4a) or H-bond donators (4b), resulted
in significant loss of affinity for both TR-binding and
TRAF-binding, but unexpectedly retained agonist effi-
cacy in TRAF.
(dose causing 50% suppression from vehicle), and TSH
suppression (dose causing 30% suppression from vehi-
cle). Compared with 2a, potency for cholesterol lower-
ing was retained with 3d, slightly increased for
tachycardia, and decreased for TSH-lowering. The
potency ratios (= ‘therapeutic windows’) clearly show
the selectivity for these compounds by dividing the
ED15 for heart rate/ED50 cholesterol and ED30 for
TSH/ED50 cholesterol. When normalized in this
manner, it can be seen that 2a has a larger margin for
cholesterol lowering versus tachycardia, while 3d has a
larger margin for cholesterol versus TSH-lowering. In
contrast, 1 exhibits absolutely no separation between
cholesterol and TSH-lowering versus tachycardia in this
animal model.5
Affinity and agonist efficacy progressively diminished as
0
the size of the R5 -substituent increased suggesting that
there exists a maximal size of a ligand that can be accom-
modated by the receptor with retained TR agonism (F
(7) < Cl (9) < Me (12)).8 In keeping with this hypothesis,
evaluation of 12 for antagonism in TRAF revealed that
1100 nM concentration of 12 produced 50% inhibition
of L-T3 and 70% efficacy as an antagonist.9
Consideration of the potency ratios above suggests that
3d might be a selective thyroid hormone receptor mod-
ulator (STRM) in relation to 2a, since increased liver-
selectivity for 3d alone would not be sufficient to explain
the increased separation between TSH and cholesterol-
lowering as the separation versus tachycardia actually
decreased.
Overall, based on optimal affinity for TR-binding,
TRAF-affinity as well as TRb1-selectivity, 3d was
selected for in vivo studies in the cholesterol-fed rat
model.5
Dose–response data for 1, 2a,5 and 3d are shown in
Table 2. The data are shown as ED15 for heart rate (dose
causing 15% increase from vehicle), ED50 for cholesterol
In conclusion, based on the scaffold of 2b, a number
of a-amino acid derivatives were prepared. Among
these compounds 3d was found to exhibit optimal
affinity, TRb1-selectivity, and agonist activity. In cho-
lesterol-fed rats, 3d displayed potential STRM behav-
ior in relation to 2a, which warrants further
examination of this compound in relevant in vitro
and in vivo models.
Table 2. ED-values and potency ratios for 1, 2a, and 3d in the
cholesterol-fed rat modela
Compound ED15 HR/ED50 Chol
ED30 TSH/ED50 Chol
6.7/21 = 0.3
1
31/21 = 1.5
2a
3d
2900/72 = 40 (27-fold vs 1) 32/72 = 0.4 (1.3-fold vs 1)
1580/82 = 19 (13-fold vs 1) 242/82 = 3.0 (10-fold vs 1)
a ED-values are expressed as nmol/kg/day. Administration was per oral
and the experimental procedures in Ref. 5 were followed. Full dose–
response data and procedure for calculation of ED-values can be
found in ‘Supplementary data’.
Supplementary data
Supplementary data, including experimental procedures,
dose–response data in rats, and analytical data,

4134
N. Garg et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4131–4134
Egan, D.; Baxter, J. D. Proc. Natl. Acad. Sci. U.S.A. 2003,
100, 10067.
6. Grover, G. J.; Malm, J. Drug Discovery Today, Ther. Strat.
2005, 2, 137.
associated with this article can be found, in the online
version, at doi:10.1016/j.bmcl.2007.05.049.
References and notes
7. TR-binding assay, vector constructs, generation of reporter
cell lines (TRAF), and assay procedures have been
described in detail in Ref. 4 and 0references therein.
1. (a) Talbert, R. L. Am. J. Health. Syst. Pharm. 2003, 60, S3,
quiz S25; (b) Friday, K. E. Exp. Biol. Med. (Maywood).
2003, 228, 769; (c) Grundy, S. M.; Cleeman, J. I.; Merz, C.
N., ; Brewer, H. B., Jr.; Clark, L. T.; Hunninghake, D. B.;
Pasternak, R. C.; Smith, S. C., Jr.; Stone, N. J. J. Am. Coll.
Cardiol. 2004, 44, 720.
8. The inductive effect of the R5 -position might also be
0
considered. When the R5 -position is 0 substituted with a
fluorine atom, then the pKa of the R4 -hydroxyl group is
lowered by approximately 1.5 log units (HL/L+H = 8.99,
ACD/pKa DB: 4.04, http://www.acdlabs.com). Considering
0
2. (a) Malm, J. Curr. Pharm. Des. 2004, 10, 3522; (b) Webb, P.
Exp. Opin. Invest. Drugs 2004, 13, 489; (c) Malm, J.; Grover,
the importance of the R4 -hydroxyl group for functional
activity (agonism/antagonism), a change of its pKa may
have consequences. The cell-assay (TRAF) experiment is,
˚
G. J.; Fa¨rnegardh, M. Mini-Rev. Med. Chem. 2007, 7, 79.
3. (a) Yokoyama, N.; Walker, G. N.; Main, A. J.; Stanton, J.
L.; Morrissey, M. M.; Boehm, C.; Engle, A.; Neubert, A.
D.; Wasvary, J. M.; Stephan, Z. F.; Steele, R. E. J. Med.
Chem. 1995, 38, 695; (b) Barlow, J. W.; Raggatt, L. E.; Lim,
C.; Kolliniatis, E.; Topliss, D. J.; Stockigt, J. R. Acta
Endocrinol. 1991, 124, 37.
however, conducted in a cell culture media at pH 7.4.
Consequently, the larger fraction of the R4 -hydroxyl group
is still unionized.
0
9. For the use of the TRAF-assay in ‘antagonist-mode’, see:
(a) Carlsson, B.; Singh, B. N.; Temciuc, M.; Nilsson, S.; Li,
Y.-L.; Mellin, C.; Malm, J. J. Med. Chem 2002, 45, 623; (b)
4. Ye, L.; Li, Y.-L.; Mellstro¨m, K.; Bladh, L.-G.; Koehler, K.;
Garg, N.; Garcia Collazo, A. M.; Litten, C.; Husman, B.;
Persson, K.; Ljunggren, J.; Grover, G.; Sleph, P. G.;
George, R.; Malm, J. J. Med. Chem. 2003, 46, 1580.
5. Grover, G. J.; Mellstro¨m, K.; Ye, L.; Malm, J.; Li, Y.-L.;
Bladh, L.-G.; Sleph, P. G.; Smith, M. A.; George, R.;
Vennstro¨m, B.; Mookhtiar, K.; Horvath, R.; Speelman, J.;
Koehler, K.; Gordon, S.; Brandt, P.; Carlsson, B.; Backs-
¨
bro-Saeidi, A.; Apelqvist, T.; Agback, P.; Grover, G. J.;
˚
Nelson, W.; Grynfarb, M.; Farnegardh, M.; Rehnmark, S.;
Malm, J. J. Med. Chem. 2006, 49, 6635; (c) Hedfors, A.;
¨
˚
Appelqvist, T.; Carlsson, B.; Bladh, L.-B.; Litten, C.;
Agback, P.; Grynfarb, M.; Koehler, K. F.; Malm, J.
J. Med. Chem. 2005, 48, 3114.