Salicylamide inhibitors of influenza virus fusion

Article abstract of DOI:10.1016/S0960-894X(00)00335-8

Structural variation of the quinolizidine heterocycle of the influenza fusion inhibitor BMY-27709 was examined by several topological dissections in order to illuminate the critical features of the ring system. This exercise resulted in the identification of a series of synthetically more accessible decahydroquinolines that retained the structural elements of BMY-27709 important for antiviral activity. The 2-methyl-cis-decahydroquinoline 6f was the most potent influenza inhibitor identified that demonstrated an EC₅₀ of 90 ng/mL in a plaque reduction assay. (C) 2000 Elsevier Science Ltd. All rights reserved.

Full text of DOI:10.1016/S0960-894X(00)00335-8

Bioorganic & Medicinal Chemistry Letters 10 (2000) 1649±1652
Salicylamide Inhibitors of In¯uenza Virus Fusion
Keith D. Combrink,a,* H. Belgin Gulgeze,a Kuo-Long Yu,a Bradley C. Pearce,a
Ashok K. Trehan,a Jianmei Wei,a Milind Deshpande,a Mark Krystal,b
Albert Torri,b Guangxiang Luo,b Christopher Cianci,b Stephanie Danetz,b
Laurence Tileyb,y and Nicholas A. Meanwella
aDepartment of Chemistry, The Bristol-Myers Squibb Pharmaceutical Research Institute,
5 Research Parkway, Wallingford, CT 05492, USA
bDepartment of Virology, The Bristol-Myers Squibb Pharmaceutical Research Institute,
5 Research Parkway, Wallingford, CT 05492, USA
Received 9 May 2000; accepted 17 May 2000
AbstractÐStructural variation of the quinolizidine heterocycle of the in¯uenza fusion inhibitor BMY-27709 was examined by several
topological dissections in order to illuminate the critical features of the ring system. This exercise resulted in the identi®cation of a
series of synthetically more accessible decahydroquinolines that retained the structural elements of BMY-27709 important for
antiviral activity. The 2-methyl-cis-decahydroquinoline 6f was the most potent in¯uenza inhibitor identi®ed that demonstrated an
EC50 of 90 ng/mL in a plaque reduction assay. # 2000 Elsevier Science Ltd. All rights reserved.
In¯uenza is responsible for over 20,000 deaths annually
in the US despite the annual seasonal vaccination cam-
paigns that have become familiar. Vaccination provides
limited protection based on the ability to predict the
exact strains of in¯uenza that will predominate a year in
advance of the in¯uenza season. The antiviral agents
amantadine and rimantadine have been available for some
time but do not protect against in¯uenza B infections and
resistance to these agents has been reported to develop
quickly.1,2 More recently, a new class of drugs that
inhibit in¯uenza neuraminidase has been licensed for
use in the US and European countries.3,4
identi®cation of the 5-methyl and 5-chloro-salicylamide
derivatives, (1b±c) as optimal.8 In this article, we
describe complementary studies that provide insight into
the SARs associated with the quinolizidine heterocycle
of BMY-27709 (1) as part of a broader e€ort to identify
structurally simpler elements that would facilitate optimi-
zation of potency and spectrum of activity. Conceptually,
the quinolizidine ring was systematically simpli®ed using
the bond disconnections and ring addition depicted in
Figure 1. Breaking bond c identi®ed a series of 4-amino-
piperidines 2 which could be elaborated into the tropanes
3a±b, examined in an e€ort to provide some insight into
the active conformation of the heterocyclic element. The
double disconnection of a and c produced the dialkya-
minoethyl derivatives 4 as targets while breaking only
the carbon±carbon bond de®ned by a led to a series of
2,6-dimethylpiperidines 5. Since the presence and orien-
tation of the methyl group of BMY-27709 (1) was known
to be of importance,8 a series of decahydroquinolines 6
and 7, in which this alkyl group is projected from a
heterocyclic ring with less conformational ¯exibility,
was identi®ed based on combining bond disconnection a
with the ring addition de®ned by b.9
We have recently described a new class of in¯uenza inhi-
bitor that interferes with the membrane fusion activity of
the virus surface protein hemagglutinin (HA)5 7 and have
disclosed fundamental aspects of the structure±activity
relationships associated with the salicylic acid moiety of
this new chemotype.8 Mechanistic studies with the proto-
type, BMY-27709 (1), indicate that the compound pre-
vents the low pH-induced conformational rearrangement
of HA that occurs in the endosomal compartment after
endocytosis of the virion. Variation of the substitution
pattern of the aromatic ring of BMY-27709 (1) led to the
Chemistry
*Corresponding author. Tel.: +1-203-677-6975; fax: +1-203-677-
7702; e-mail: keith.combrink@bms.com
yCurrent address: Centre for Veterinary Science, University of Cam-
bridge, Madingley Road, Cambridge, UK, CB3 0ES.
1,1-N,N-Diisopropylpropylenediamine was prepared by
alkylation of diisopropylamine with 3-bromopropionitrile
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00335-8

1650
K. D. Combrink et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1649±1652
Figure 1. Carbon±carbon bond disconnections and connections used to design compounds.
followed by reduction with LiAlH4. The N-ethyl- and
N-propylamino-2,6-dimethyl piperidine derivatives 5a±d
were synthesized by literature methods.10 N-sec-Butyl-4-
aminopiperidine was prepared by reductive amination
of the secondary amine with methyl ethyl ketone and
NaBH3CN.11 The decahydroquinoline derivatives were
obtained by a two step reduction of the corresponding
quinoline.12,13 The 3 carbon linker was introduced by
Michael addition of the amine to acrylonitrile followed
by reduction of the nitrile with LiAlH4. The 2 carbon
linker element was installed by acylation of the amine
with chloroacetyl chloride followed by displacement of
the chloride by azide and concomitant reduction of the
azide and amide using LiAlH4 (Scheme 1).
regenerated either by mild hydrolysis of the acetate or
nucleophilic demethylation of the methyl ether using
sodium methanethiolate (Scheme 2, method A and B).
However, the most convenient method for amide bond
formation was to protect the phenol and acid elements
simultaneously as an acetonide and heat this with an
excess of amine in toluene,14,15 as depicted in Scheme 2,
method C.
The compounds prepared as part of this study are
compiled in Table 1. All compounds were evaluated in a
cell protection assay using MDCK cells infected with
the A/WSN/33 strain of in¯uenza virus (H1N1 subtype)
and the concentration of drug a€ording 50% protection
(EC50) determined.5
Results and Discussion
The amino piperidine derivative 2 was inactive and activ-
ity was not restored by introducing the conformational
constraint a€orded by the tropanes 3a±b, examined as the
individual axial and equatorial isomers (only four repre-
sentatives of several examined are depicted in Table 1).16
The simplest structural representation of the quinolizidine
element, dialkyaminoalkyl derivatives 4 were also inactive
with the exception of the diisopropyl analogue 4, which
possesses antiviral activity comparable to that of the pro-
totype 1. Further elaboration of this chemotype into the
cis-dimethyl piperidines 5a±d provided active compounds
in which there is a preference for a 3 carbon linker (5a
and 5b) compared to the ethyl linkage found in 5c and 5d.
This trend is consistent with the topological disconnection
a depicted in Figure 1 and, taken together with the SAR
for compounds 2±4, suggests that non-collapsible struc-
tural elements distal to the amide moiety are required for
e€ective expression of antiviral activity. This theme was
Direct coupling of commercially available salicyclic acid
derivatives via standard methods sometimes gave low
yields of the desired product along with a complex
mixture of side products. In these cases, the phenol was
masked as an acetate or methyl ether, the acid chloride
generated and coupled with the amine. This procedure
was particularly useful for hindered amines such as the
axial amino tropanes 3a. The phenol was subsequently
Scheme 1. (a) n=1; (i) ClCH2COCl/Et3N; (ii) NaN3/THF/H2; (iii)
LiAlH4; ether; (b) n=2; (i) acrylonitrile; (ii) LiAlH4, ether.
Scheme 2. Method A: R=Me or Ac; (1) SOCl2; (2) amine, Et3N, CH2Cl2; (3) NaSEt or Na2CO3, respectively; Method B: R=H, EDC, HOBT,
DMF, amine; Method C: amine, toluene, Á.

K. D. Combrink et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1649±1652
1651
Table 1. Cell protection assay of cells infected with H1N1 subtype in¯uenza for compounds 1±7
Compound no.
n
X
R
Synthetic methoda
EC50 (mg/mL)b
CC50 (mg/mL)c
1c
2
3a±b
4
Cl
Me
Me or Cl
Me
Cl
0.808
nad
na
0.90
>200
100
100
>100
100
>100
30
>100
30
>100
43
90
35
40
>100
80
9.5
100
60
B
A
B
Me, iPr, s-Bu or Bn
5a
5b
5c
5d
6a
6b
6c
6d
6e
6f
3
3
2
2
3
2
2
3
2
2
2
2
2
2
2
2
2
A
A
A
A
A
A
A
A
C
B
A
A
A
A
C
C
C
1
Me
Cl
Me
Me
Me
Cl
Me
Me
Cl
Me
Me
Me
Me
Me
Cl
0.70 (Æ0.1)
4.5 (Æ1.5)
1.8 (Æ0.2)
na
0.65 (Æ0.5)
1.4 (Æ0.6)
na
0.19 (Æ0.05)
0.09 (Æ0.02)
6.0 (Æ0.5)
0.25
H
H
H
a-2-Me
a-2-Me
a-2-Me
b-2-Me
a-2-iPr
a-3-Me
a/b-4-Me
H
6g
6h
6i
na
na
6j
7a
7b
7c
0.35 (Æ0.15)
1.4 (Æ0.6)
1.5 (Æ0.5)
H
a-Me
40±45
20±30
Me
aSee Scheme 2 for the synthetic method.
b50% Inhibition of A/WSN/33 in¯uenza strain-induced toxicity in a cell protection assay in MDBK cells, values are mean of duplicate experiments;
standard deviation in parentheses.
cCytotoxicity of the compound in MDBK cell.
dna=not active at 50 mg/mL.
further explored in the context of the decahydroquino-
lines 6 and 7, several of which demonstrate potent
in¯uenza inhibitory activity in cell culture. Interestingly,
in this series a 2 carbon linker appears to be optimal
since 6a and 6d were found to be inactive. The two
unsubstituted cis-decahydroquinolines 6b and 6c are
active antiviral agents that can be improved by the
introduction of a methyl group at C-2, compounds 6e
and 6f. Activity is sensitive to the stereochemistry at C-
2, since the b-methyl isomer 6g is markedly less active,
and to the relative topological location, since the 3 and 4-
substituted methyl isomers 6i and 6j, respectively, are
devoid of antiviral properties. The single alkyl homologue
examined, 6h, retains good inhibitory activity. In the
trans-decahydroquinoline series, the activity associated
with the parent structures 7a and 7b are comparable to
that of their cis-con®gured counterparts 6b and 6c while
the single substituted analogue 7c is approximately 10-
fold weaker than 6e.
shortcoming inherent to 1 and structurally related series
of in¯uenza fusion inhibitors.8,17,18 Based on this ®nding
and the discovery of more potent in¯uenza inhibitors
within a series of cyclohexanol derivatives,14 further
examination of this series was not pursued.
References and Notes
1. Bean, B. Clin. Microbiol. Rev. 1992, 5, 146.
2. Meanwell, N. A.; Krystal, M. Drug Disc. Today 1996, 1,
316; ibid 1996, 1, 388.
3. Bethell, R. C.; Smith, P. W. Drugs Future 1998, 23, 1099.
4. Kim, C. U.; Chen, X.; Mendel, D. B. Antiviral Chem. Che-
mother. 1999, 10, 141.
5. Luo, G.-X.; Colonno, R.; Krystal, M. Virology 1996, 226,
66.
6. Luo, G.-X.; Torri, A.; Harte, W. E.; Danetz, S.; Cianci, C.;
Tiley, L.; Day, S.; Mullaney, D.; Yu, K.-L.; Quellet, C.; Dex-
traze, P.; Meanwell, N. A.; Colonno, R.; Krystal, M. J. Virol.
1997, 71, 4062.
7. Cianci, C.; Yu, K.-L.; Dischino, D. D.; Harte, W.; Desh-
pande, M.; Luo, G.-X.; Colonno, R. J.; Meanwell, N. A.;
Krystal, M. J. Virol. 1999, 73, 1785.
8. Yu, K. L.; Ruediger, E.; Luo, G.; Cianci, C.; Danetz, S.;
Tiley, L.; Trehan, A. K.; Monkovic, I.; Pearce, B.; Martel, A.;
Krystal, M.; Meanwell, N. A. Bioorg. Med. Chem. Lett. 1999,
9, 2177.
Whilst this study has identi®ed potent and e€ective inhibi-
tors of in¯uenza virus that are synthetically more accessible
than the quinolizidine heterocycle found in the prototype 1,
the antiviral activity resides predominantly in compounds
with only limited conformational constraint. As a con-
sequence, conclusions about the active conformation of 1
cannot be logically reached from this study.
9. Ring addition b could also be drawn to include the 5-
methyl group of the quinolizidine with disconnection a which
gives the same 2-methyldecahydroquinoline.
None of the compounds presented in Table 1 demon-
strated signi®cant inhibition of an in¯uenza H3 strain, a

1652
K. D. Combrink et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1649±1652
10. Roufos, I. H.; Dooley, S. J.; Schwarz, R. D.; Campbell, G.
W.; Probert, A. W. J. Med. Chem. 1994, 37, 268.
11. Borch, R. F.; Bernstein, M. D.; Durst, H. D. J. Am.
Chem. Soc. 1971, 93, 2897.
12. Meyers, A. I.; Milot, G. J. Am. Chem. Soc. 1993, 115, 6652.
13. Booth, H.; Griths, D. V.; Jozefowicz, M. L. J. Chem.
Soc. Perkin Trans 2 1976, 751.
14. Deshpande, M.; Wei, J.; Luo, G.-X.; Cianci, C.; Danetz,
S.; Tiley, L.; Pearce, B. C.; Krystal, M.; Yu, K.-L.; Meanwell,
N. A. Bioorg. Med. Chem. Lett., manuscript submitted.
15. Nishimura, M.; Hoshi, M.; Igawa, H. Japanese Patent
21608, 1997; Chem. Abstr. 1997, 127, 205354.
16. For an example where the quinolizidine was successfully
substituted with a tropane ring see: King, F. D.; Hadley, M.
S.; Joiner, K. T.; Martin, R. T.; Sanger, G. J.; Smith, D. M.;
Smith, G. E.; Smith, P.; Turner, D. H.; Watts, E. A. J. Med.
Chem. 1993, 36, 683.
17. Staschke, K. A.; Hatch, S. D.; Tang, J. C.; Hornback, W.
J.; Munroe, J. E.; Colacino, J. M.; Muesing, M. A. Virology
1998, 248, 264.
18. Plotch, S. J.; O'Hara, B.; Morin, J.; Palant, O.;
LaRocque, J.; Bloom, J. D.; Lang, S. A. Jr.; DiGrandi,
M. J.; Bradley, M.; Nilakantan, R.; Gluzman, Y. J.
Virol. 1999, 73, 140.