Mineralocorticoid receptor antagonists: Identification of heterocyclic amide replacements in the oxazolidinedione series

Article abstract of DOI:10.1016/j.bmcl.2014.02.057

Novel potent and selective mineralocorticoid receptor antagonists were identified, utilizing heterocyclic amide replacements in the oxazolidinedione series. Structure-activity relationship (SAR) efforts focused on improving lipophilic ligand efficiency (LLE) while maintaining nuclear hormone receptor selectivity and reasonable pharmacokinetic profiles.

Full text of DOI:10.1016/j.bmcl.2014.02.057

Bioorganic & Medicinal Chemistry Letters 24 (2014) 1681–1684
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Mineralocorticoid receptor antagonists: Identification of heterocyclic
amide replacements in the oxazolidinedione series
a
a
a
a
b
Jason M. Cox a, , Hong D. Chu , Christine Yang , Hong C. Shen , Zhicai Wu , Jaume Balsells ,
Alejandro Crespo c, Patricia Brown e, Beata Zamlynny e, Judyann Wiltsie d, Joseph Clemas d, Jack Gibson d,
Lisa Contino d, JeanMarie Lisnock d, Gaochao Zhou d, Margarita Garcia-Calvo d, Thomas Bateman f,
Ling Xu f, Xinchun Tong f, Martin Crook e, Peter Sinclair a
⇑
a Department of Discovery Chemistry, Merck Research Laboratories, 2000 Galloping Hill Rd, Kenilworth, NJ 07033, USA
b Department of Process Chemistry, Merck Research Laboratories, 2000 Galloping Hill Rd, Kenilworth, NJ 07033, USA
c Department of Chemistry Modeling & Informatics, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
d Department of In Vitro Pharmacology, Merck Research Laboratories, 2000 Galloping Hill Rd, Kenilworth, NJ 07033, USA
e Department of Cardiovascular Diseases, Merck Research Laboratories, 2000 Galloping Hill Rd, Kenilworth, NJ 07033, USA
f Department of Drug Metabolism, Merck Research Laboratories, 2000 Galloping Hill Rd, Kenilworth, NJ 07033, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Novel potent and selective mineralocorticoid receptor antagonists were identified, utilizing heterocyclic
amide replacements in the oxazolidinedione series. Structure–activity relationship (SAR) efforts focused
on improving lipophilic ligand efficiency (LLE) while maintaining nuclear hormone receptor selectivity
and reasonable pharmacokinetic profiles.
Received 7 January 2014
Revised 17 February 2014
Accepted 20 February 2014
Available online 1 March 2014
Ó 2014 Elsevier Ltd. All rights reserved.
Keywords:
Mineralocorticoid receptor antagonist
Hypertension
Oxazolidinedione
Lipophilic ligand efficiency
The mineralocorticoid receptor (MR) is a member of the nuclear
hormone receptor (NHR) super-family and regulates gene expres-
sion involved in cardiovascular disease.1 The natural endogenous
hormone aldosterone activates MR leading to chronic kidney
disease, hypertension and congestive heart failure via electrolyte
imbalance.2 The only two marketed MR antagonists are the
steroids spironolactone and eplerenone, which have demonstrated
positive effects in the treatment of patients with the aforemen-
tioned conditions (Fig. 1).3 However, selectivity for MR versus
other members in the NHR super-family (members include:
androgen receptor (AR), estrogen receptor (ER), glucocorticoid
receptor (GR), and progesterone receptor (PR)) is essential to avoid
undesired side effects. Spironolactone, for example, has sex
hormone related side effects (gynecomastia and menstrual irregu-
larities) due to AR and PR activity.4 In contrast, eplerenone is very
selective for MR versus the other NHR members, but at the price of
reduced MR potency and b.i.d dosing.5
O
O
O
O
H
O
H
H
SAc
H
H
O
O
CO2Me
Spironolactone
Eplerenone
Figure 1. Marketed steroidal MR antagonists.
Due to the limitations of the steroid MR antagonists, we6 and
others7 have initiated discovery programs targeting non-steroidal
MR antagonists. Herein we report our continuing investigation of
a novel oxazolidinedione (ODO) class of potent MR antagonists,
focusing on efforts toward the discovery of heterocyclic replace-
ments for the amide moiety in compound 1 (Fig. 2). Compound 1
was identified through an in-house high-throughput screen utiliz-
ing a commercial PathHunter™ assay8 and demonstrated modest
MR potency (IC50 = 6 l
M).9
As previously disclosed,10 initial efforts focused on amide sub-
stitution and identification of N-3 substitution, with appropriate
⇑
Corresponding author. Tel.: +1 908 740 0425.
E-mail address: jason_cox@merck.com (J.M. Cox).
http://dx.doi.org/10.1016/j.bmcl.2014.02.057
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.

1682
J. M. Cox et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1681–1684
R1
Y
Cyclization
O
O
X
R2
O
O
O
O
O
O
N
N
H
N
R
H
3
N
N
N
O
O
O
1
2
3
Figure 2. Evolution of ODO class MR antagonists.
O
O
O
a
b, c
O
O
O
OH
OEt
O
O
EtO
EtO
EtO
NCO
N
OEt
O
4
5
7
6
R1
N
R1
R2
O
X
O
R2
O
O
f
d, e
R1
O
N
H
N
N
O
O
NH2
HO
R2
9
10
8
X = O, N
Scheme 1. Synthesis of oxazole and imidazole ODO MR antagonists. Reagents and conditions: (a) CsF, DMF, Air, 40 °C, 3 days, 71%; (b) NaOH, 4 Å ms, THF, 0 °C, 5 min. then 6,
0 °C to rt, 4 h, 46%; (c) Chiral SFC (AD column); (d) 8, EtOH, 70 °C, 2 h; (e) DMP, DCM, rt, 30 min. 40–82% over two steps; (f) oxazole: I2, PPh3, TEA, 5 min, then 9, DCM, 1 h or
imidazole: NH4OAc, AcOH, 100 °C, 40 h, 25–90% over two steps.
R stereochemistry, to identify MR antagonists 2. We envisioned
cyclization of the amide moiety would afford novel antagonists
(3), with improved physicochemical, ADME, and safety profiles
by attempting to improve the lipophilic ligand efficiency (LLE).11
The synthesis of oxazole and imidazole MR antagonists is de-
picted in Scheme 1.12 Diethyl malonate derivative 4 was oxidized
by treatment with cesium fluoride in the presence of air to afford
racemic hydroxymalonate 5. The hydroxymalonate was converted
to enantiopure oxazolidinedione 7 by first condensation with chi-
ral isocyanate 6 in the presence of sodium hydroxide followed by
purification via chiral supercritical fluid chromatography (SFC).
Treatment of key intermediate 7 with substituted hydroxylethyl
amines 8, provided cyclization precursors 9 after oxidation of the
alcohol with Dess–Martin periodinane. The final oxazole targets
10 were prepared by adding intermediates 9 to a freshly prepared
solution of iodine, triphenylphosphine and triethylamine. Alterna-
tively, treatment of intermediates 9 with ammonium acetate in hot
acetic acid afforded imidazoles 10.
Highlighted in Scheme 2 is the synthesis of triazole and oxadi-
azole MR antagonists. The 1,3,4-triazole derivatives 12 were syn-
thesized in a single step from key intermediate 7 by heating with
amino-hydrazides in ethanol. 1,3,4-Oxadiazoles 13 were prepared
in three steps from 7 by first acyl-hydrazide formation using anhy-
drous hydrazine, followed by amide bond generation using a vari-
ety of acids. Finally, cyclization utilizing the procedure for
synthesizing the oxazoles (addition of the product from the previ-
ous step to a freshly prepared solution of iodine, triphenylphos-
phine and triethylamine) provided 1,3,4-oxadiazoles 13.
Our initial focus was to identify heterocyclic amide replace-
ments (Table 1) to simultaneously improve not only liver micros-
me (LM) stability but also LLE with respect to the previously
disclosed 3,5-dimethoxy benzyl amide analog (IC50 = 54 nM;
R
NH
N
a
O
O
N
7
LLE = 2.54; clogP = 4.73; H, R LM% at 1 lM incubation after
NH
N
0.5 h = 0, 0).6d While oxazole region-isomers 10a and 10b had a
significant loss in potency and LLE, gratifyingly, each compound
demonstrated an increase in microsomal stability when compared
to the amide analog. We could improve the intrinsic potency and
LLE by making the benzyl substitution (10c); however, not surpris-
ingly this came at the expense of the microsomal stability. Switch-
ing to phenyl imidazole 10d improved the intrinsic potency and
LLE, while showing similar LM stability compared to the oxazoles;
however, benzyl substituted imidazole 10e had a similar profile to
oxazole 10c. Addition of the 3,5-dimethoxy phenyl substitution
(10f) significantly improved the intrinsic potency and LLE with re-
spect to the other cyclized analogs, while maintaining some of the
microsomal stability improvements gained from cyclization.
Unfortunately, the oxazoles and imidazoles did not achieve our
goal of improved LLE compared to their respective amide analogs.
NH2
O
R
N
H
12
11
R
N
O
O
b, c, d
O
N
7
N
O
13
Scheme 2. Synthesis of triazole and oxadiazole ODO MR antagonists. Reagents and
conditions: (a) 11, DIEA, EtOH, 70 °C, 1 h, 30–75%; (b) anhydrous hydrazine, MeOH,
sonication, 1 h; (c) RCOOH, DIEA, HATU, DCM, rt, 1.5 h; (d) I2, PPh3, TEA, 5 min, then
product from step c, DCM, 1 h, 40–90% over three steps.

J. M. Cox et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1681–1684
1683
Table 1
While we observed improved microsomal stability with the
oxazoles, imidazoles, and 1,3,4-triazoles compared to the amide
analogs, we were still left with the challenge of improving the
LLE and reducing the clogP in order to be in a more desirable
drug-like space. By calculating clogP prior to synthesis, we knew
that the 1,3,4-oxadiazole amide replacement would, in general,
have a 0.5–1.5 clogP reduction when compared to the previous
heterocyclic and amide analogs. With this in mind, we synthesized
1,3,4-oxadiazoles as depicted in Scheme 2, with representative
examples highlighted in Table 3. In general, the 1,3,4-oxadiazoles
showed similar potency to the imidazoles; both had improved po-
tency relative to the 1,3,4-triazoles and oxazoles, with improved
LLE due to improved clogP (e.g. 13a and 13b vs 10a–e and
12a,b) as predicted. Unfortunately, these improvements in LLE
came at the price of generally reduced rat microsomal stability
compared to the 1,3,4-triazoles, however, the human microsomal
stability was maintained. Interestingly, while we saw similar
potency with the 3,4-dichlorophenyl 12c compared to the
unsubstituted phenyl 12a in the 1,3,4-triazoles, the 3,4-dichloro-
SAR of amide replacement: oxazole and imidazole
R1
X
R2
O
O
N
N
O
a
Compd
R1
R2
X
IC50
(
l
M)
LLEb
clogP
LMc
H, R
10a
10b
10c
10d
10e
H
Ph
H
H
H
H
O
O
O
NH
NH
1.70
1.11
0.24
0.24
0.27
1.21
1.4
2.32
1.91
2.12
4.55
4.55
4.29
4.70
4.44
73, 15
73, 29
0, 0
62, 4
0, 0
Ph
Bn
Ph
Bn
MeO
10f
H
NH
0.068
2.38
4.79
43, 4
MeO
phenyl-1,3,4-oxadiazole 13c had
compared to 13a.
a 20-fold loss in potency
a
b
c
Values are the average of two experiments, each in 10-point titrations.
LLE = pIC50—clogP.
% Remaining at 0.5 h and 1 lM incubation in human (H) and rat (R) liver
As was previously noted, addition of 3,5-methoxyphenyl substi-
tution provided a boost in potency, which also held true for 1,3,4-
oxadiazole 13d, leading to an improved LLE. Compound 13d was
microsomes (LM).
Table 2
Table 3
SAR of amide replacement: 1,3,4-triazole
SAR of amide replacement: 1,3,4-oxadiazole
R
R
O
NH
N
O
N
O
O
N
O
N
N
N
O
O
LLEb
clogP
LMc
H, R
a
Compd
R
IC50
(l
M)
Compd
R
IC50
(
lM)
LLEb
clogP
LMc
H, R
a
12a
12b
Ph
Bn
1.44
0.66
1.39
1.99
4.45
4.19
78, 29
63, 7
13a
13b
Ph
Bn
0.38
0.16
3.09
3.74
3.33
3.07
82, 0
7, 0
Cl
Cl
12c
1.32
0.58
5.30
88, 41
13c
6.74
0.5
4.67
—
Cl
Cl
MeO
MeO
MeO
12d
0.67
2.06
4.11
74, 37
13d
0.039
3.95
3.46
64, 28
MeO
F
a
b
c
Values are the average of two experiments, each in 10-point titrations.
LLE = pIC50—clogP.
% Remaining at 0.5 h and 1 lM incubation in human (H) and rat (R) liver
13e
13f
13g
0.27
1.53
0.13
2.92
3.79
4.12
3.65
2.03
2.76
74, 1
—
microsomes (LM).
F
N
With the increased LM stability imparted by cyclization (as
exemplified by analogs in Table 1), we next turned our attention
to improving the LLE profile, by attempting to maintain potency
while reducing clogP via introduction of a third heteroatom in
our cyclized ring. To this end, we first investigated the addition
of a nitrogen atom to generate 1,3,4-triazoles as depicted in Table 2.
In general, the analogs in Table 2 had lower clogP and improved
microsomal stability, but these attributes came at the cost of re-
duced intrinsic potency and reduced LLE. For example, 1,3,4-tria-
zole 12d when compared to imidazole 10f showed a 10-fold
reduction in intrinsic potency with a 0.68 log unit improvement
in clogP, leading to a 0.32 reduction in LLE. However, 12d demon-
strated an improved microsomal stability (H, R = 74, 37) versus 10f
(H, R = 43, 4) after 0.5 h microsomal incubation in both human and
rat microsomes.
MeO
N
N
31, 0
MeO
13h
13i
0.47
3.85
3.58
2.47
3.50
—
0.082
57, 2
a
b
c
Values are the average of two experiments, each in 10-point titrations.
LLE = pIC50—clogP.
% Remaining at 0.5 h and 1 lM incubation in human (H) and rat (R) liver
microsomes (LM).

1684
J. M. Cox et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1681–1684
Table 4
References and notes
Selectivity and PKa profile of select compounds 13d and 13i
1. (a) Fardella, C.; Miller, W. Annu. Rev. Nutr. 1996, 16, 443; (b) Moore, J. T.; Collins,
J. L.; Pearce, K. H. ChemMedChem 2006, 1, 504; (c) Gronemeyer, H.; Gustafsson,
J. A.; Laudet, V. Nat. Rev. Drug Disc. 2004, 3, 950.
Property
13d
13i
b
NHR selectivity IC50
(lM): AR, ER, GR, PR
>5
>5
2. (a) McKelvie, R. S.; Yusuf, S.; Pericak, D.; Avezum, A.; Burns, R. J.; Probstfield, J.;
Tsuyuki, R. T.; White, M.; Rouleau, J.; Latini, R.; Maggioni, A.; Young, J.; Pogue, J.
Circulation 1999, 100, 1056; (b) Catena, C.; Colussi, G.; Marzano, L.; Sechi, L. A.
Horm. Metab. Res. 2012, 44, 181; (c) Fourkiotis, V. G.; Hanslik, G.; Hanusch, F.;
Lepenies, J.; Quinkler, M. Horm. Metab. Res. 2012, 44, 194.
Ion channel IC50
(
l
M): Ca, IKr, Na
>13
>50
3.7
7.7
1.2
2.4
44
>6
CYP IC50
(l
M): 3A4, 2D6, 2C9
>19
4.4
7.9
1.0
2.8
89
AUCN iv (
l
M h kg/mg)
Cl (mL/min/kg)
Vd (L/kg)
3. (a) The, R. A. L. E. S. N. Eng. J. Med. 1999, 341, 709; (b) Zannad, F.; McMurray, J. J.
V.; Krum, H.; van Veldhuisen, D. J.; Swedberg, K.; Shi, H.; Vincent, J.; Pocock, S.
J.; Pitt, B. N. Eng. J. Med. 2003, 348, 1309; (c) Funder, J. W. Hypertens. Res. 2010,
33, 539; (d) Calhoun, D. A.; White, W. B. J. Am. Soc. Hypertens. 2008, 2, 462; (e)
The RALES Investigators Am. J. Cardiol. 1996, 78, 902; (f) Bomback, A. S.;
Muskala, P.; Bald, E.; Chwatko, G.; Nowicki, M. Clin. Nephrol. 2009, 72, 449; (g)
Williams, J. S. Nat. Rev. Endocrinol. 2010, 6, 248; (h) Nishizaka, M. K.; Calhoun,
D. A. Curr. Hypertens. Rep. 2005, 7, 343; (f) Gaddam, K.; Corros, C.; Pimenta, E.;
Ahmed, M.; Denney, T.; Aban, I.; Inusah, S.; Gupta, H.; Lloyd, S. G.; Oparil, S.;
Husain, A.; Dell’Italia, L. J.; Calhoun, D. A. Hypertension 2010, 55, 1137; (g)
Zannad, F.; McMuray, J. J.; Drexler, H.; Drum, H.; van Veldhuisen, D. J.;
Swedberg, K.; Shi, H.; Vincent, J.; Pitt, B. Eur. J. Heart Fail. 2010, 12, 617.
4. De Gasparo, M.; Whitebread, S. E.; Preiswerk, G.; Jeunemaitre, X.; Corvol, P.;
Menard, J. J. Steroid Biochem. 1989, 32, 223.
5. De Gasparo, M.; Joss, U.; Ramjoe, H.; Whitebread, S.; Haenni, H.; Schenkel, L.;
Kraehenbuehl, C.; Biollas, M.; Grob, J.; Schmidlin, J.; Wieland, P.; Wehrli, H. U. J.
Pharmacol. Exp. Ther. 1987, 240, 650.
6. (a) Shen, H.; Cox, J. M.; Yang, C.; Wu, Z. PCT Intl. Appl. 2013, WO2013055606.;
(b) Shen, H.; Yang, C.; Cox, J. M.; Liu, K. PCT Intl. Appl. 2013, WO2013055607.;
(c) Shen, H.; Cox, J. M.; Chu, H. D. PCT Intl. Appl. 2013, WO2013055608.; (d)
Yang, C.; Shen, H. C.; Wu, Z.; Chu, H. D.; Cox, J. M.; Balsells, J.; Crespo, A.; Brown,
P.; Zamlynny, B.; Wiltsie, J.; Clemas, J.; Gibson, J.; Contino, L.; Linsnock, J.-M.;
Zhou, G.; Garcia-Calvo, M.; Bateman, T.; Xu, L.; Tong, X.; Crook, M.; Sinclair, P.
Bioorg. Med. Chem. Lett. 2013, 23, 4388.
7. (a) Futatsugi, K.; Piotrowski, D. W.; Casimiro-Garcia, A.; Robinson, S.;
Sammons, M.; Loria, P. M.; Banker, M. E.; Petersen, D. N.; Schmidt, N. J.
Bioorg. Med. Chem. Lett. 2013, 23, 6239; (b) Piotrowski, D. W. J. Med. Chem. 2012,
55, 7957; (c) Leftheris, K.; Zheng, Y.; Lala, D. S. Annu. Rep. Med. Chem. 2011, 46,
89; (d) Barfacker, L.; Kuhl, A.; Hillisch, A.; Grosser, R.; Figueroa-Perez, S.;
Heckroth, H.; Nitsche, A.; Erguden, J. K.; Gielen-Haertwig, H.; Schlemmer, K. H.;
Mittendorf, J.; Paulsen, H.; Platzek, J.; Kolkhof, P. ChemMedChem 2012, 7, 1385;
(e) Hasui, T.; Ohra, T.; Ohyabu, N.; Asano, K.; Matsui, H.; Mizukami, A.; Habuka,
N.; Sogabe, S.; Endo, S.; Siedem, C. S.; Tang, T. P.; Gauthier, C.; De Meese, L. A.;
Boyd, S. A.; Fukumoto, S. Bioorg. Med. Chem. 2013, 21, 5983; (f) Nariai, T.; Fujita,
K.; Mori, M.; Katayama, S.; Hori, S.; Matsui, K. J. Cardiovasc. Pharmacol. 2012, 59,
458; (g) Coates, D. A.; Gavardinas, K.; Jadhav, P. K. PCT Int. Appl. 2010,
WO2010104721.; (h) Casimiro-Garcia, A.; Futatsugi, K.; Piotrowski, D. W. PCT
Int. Appl. 2011, WO2011141848.; (i) Meyers, M. J.; Arhancet, G. B.; Hockerman,
S. L.; Chen, X.; Long, S. A.; Mahoney, M. W.; Rico, J. R.; Garland, D. J.; Blinn, J. R.;
Collins, J. T.; Yang, S.; Huang, H. C.; McGee, K. F.; Wendling, J. M.; Dietz, J. D.;
Payne, M. A.; Homer, B. L.; Heron, M. I.; Reitz, D. B.; Hu, X. J. Med. Chem. 2010,
53, 5979.
T
(h)
1/2
F (%)
Rat plasma protein binding (% bound)
0.2
0.5
a
Formulation: 1 mg/mL PEG 200: water (70:30). iv dose: 1 mg/kg (n = 2). po
dose: 2 mg/kg (n = 3). Blood concentrations were determined by LC/MS/MS fol-
lowing protein precipitation with acetonitrile.
b
Values are the average of two experiments, each in 10-point titrations.
also the only compound that maintained similar human and rat
microsomal stability in this subclass, when compared to the
1,3,4-triazoles. With these positive attributes, we continued our
SAR evaluation by investigating alternate 3,5-phenyl substitution,
introduction of nitrogen, and a combination thereof as exemplified
by 13e, 13f, and 13g, respectively.13 Unfortunately, these analogs
had reduced intrinsic potency leading to reduced LLE (13e and
13f) and/or decreased rat microsomal stability (13e and 13g). We
investigated non-aromatic substitution, but in all cases this led
to significant loss in intrinsic potency with 13h being the most po-
tent and highest LLE non-aromatic substitution compound synthe-
sized. All other non-aromatic analogs demonstrated significant
erosion of LLE. In an effort to block the presumed benzylic metab-
olism of 13b, we introduced cyclopropyl substitution (13i), which
improved human metabolic stability and modestly increased po-
tency but at the expense of increased clogP.
Based on improved in vitro data for 13d in the phenyl 1,3,4-oxa-
diazole series and 13i in the benzyl 1,3,4-oxadiazole series, we
chose to profile both molecules in greater detail (Table 4). Both
13d and 13i had similar nuclear hormone receptor selectivity
(IC50 >5
lM), ion channel activity (IC50 >6
lM), and cytochrome
M). The rat pharmacokinetic (PK)
P450 (CYP) activity (IC50 >19
l
profile for both analogs were also very similar; each compound
had low clearance, moderate half-life, and reasonable oral
bioavailability.
8. http://www.discoverx.com/nhrs/prod-nhrs.php.
9. Average potency in PathHunter™ assay: spironolactone 11 nM and eplerenone
244 nM.
In conclusion, improved analog profiles, with respect to the
amides, was achieved by cyclization of the amide to form heterocy-
cles in the oxazolidinedione series. The heterocycles maintained
intrinsic MR potency while showing improved LLE, and increased
metabolic stability. The 1,3,4-oxadiazole amide replacement,
exemplified by both 13d and 13i, also demonstrated significant
selectivity versus other nuclear hormone receptors, ion channels,
and CYP’s while having reasonable rat PK profiles. Further study
of the ODO series will be reported in due course.
10. For complete SAR discussion see Ref. 6d.
11. (a) Hopkins, A. L.; Groom, C. R.; Alex, A. Drug Discovery Today 2004, 9, 430; (b)
Edwards, M. P.; Price, D. P. Annu. Rep. Med. Chem. 2010, 45, 381; (c) Leeson, P.
D.; Empfield, J. R. Annu. Rep. Med. Chem. 2010, 45, 393; (d) Leeson, P. D.;
Springthorpe, B. Nat. Rev. Drug. Disc. 2007, 6, 881.
12. For detailed synthetic procedures see Ref. 6c.
13. Other phenyl substitution, pyridine, pyridazine, pyrimidine, and pyrazine
substitutions were not well tolerated.