Synthesis and SAR of novel isoquinoline CXCR4 antagonists with potent anti-HIV activity

Article abstract of DOI:10.1016/j.bmcl.2010.03.118

Using AMD070 as a starting point for structural modification, a novel series of isoquinoline CXCR4 antagonists was developed. A structure-activity scan of alternate lower heterocycles led to the 3-isoquinolinyl moiety as an attractive replacement for benzimidazole. Side chain optimization in the isoquinoline series led to a number of compounds with low nanomolar anti-HIV activities and promising rat PK properties.

Full text of DOI:10.1016/j.bmcl.2010.03.118

Bioorganic & Medicinal Chemistry Letters 20 (2010) 3026–3030
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and SAR of novel isoquinoline CXCR4 antagonists with potent
anti-HIV activity
John F. Miller a, , Kristjan S. Gudmundsson , Leah D’Aurora Richardson , Stephen Jenkinson ,
a
a
d
*
a
b
c
Andrew Spaltenstein , Michael Thomson , Pat Wheelan
Department of Medicinal Chemistry, Infectious Diseases Center for Excellence in Drug Discovery, GlaxoSmithKline Research and Development, Five Moore Drive,
Research Triangle Park, NC 27709-3398, USA
a
b
Department of Virology, Infectious Diseases Center for Excellence in Drug Discovery, GlaxoSmithKline Research and Development, Five Moore Drive, Research Triangle Park,
NC 27709-3398, USA
c
Department of DMPK, Infectious Diseases Center for Excellence in Drug Discovery, GlaxoSmithKline Research and Development, Five Moore Drive, Research Triangle Park,
NC 27709-3398, USA
d
Department of Biochemical and Analytical Pharmacology, Infectious Diseases Center for Excellence in Drug Discovery, GlaxoSmithKline Research and Development,
Five Moore Drive, Research Triangle Park, NC 27709-3398, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Using AMD070 as a starting point for structural modification, a novel series of isoquinoline CXCR4 antag-
onists was developed. A structure–activity scan of alternate lower heterocycles led to the 3-isoquinolinyl
moiety as an attractive replacement for benzimidazole. Side chain optimization in the isoquinoline series
led to a number of compounds with low nanomolar anti-HIV activities and promising rat PK properties.
Ó 2010 Elsevier Ltd. All rights reserved.
Received 10 March 2010
Revised 30 March 2010
Accepted 31 March 2010
Available online 3 April 2010
Keywords:
CXCR4
Isoquinoline
Tetrahydroquinoline
AMD070
HIV
Antiviral
In recent years the medicinal chemist’s focus in the battle
against HIV/AIDS has broadened to include molecular targets be-
yond the HIV reverse transcriptase and protease enzymes, which
acid residues, points to a receptor–ligand binding model in which
charge-charge interactions play a prominent role. Not surpris-
4
ingly, the currently known small molecule CXCR4 antagonists are
also highly basic in nature. A particularly notable example is the
tetrahydroquinoline derivative AMD070 (Fig. 1) which was re-
cently shown to possess significant anti-HIV efficacy in human
are the cornerstones of the highly active antiretroviral therapy
1
(
HAART). One area that has received a great deal of attention is
the concept of blocking viral entry by targeting the chemokine
receptors CCR5 and CXCR4, which function as co-receptors, along
with CD4, to facilitate fusion of the viral membrane with the host
5
clinical studies. Key features of the AMD070 pharmacophore in-
clude a triad of basic nitrogen atoms and a distal amino group at-
tached to the central nitrogen by a 4-carbon tether.
2
cell. CXCR4 is a G-protein coupled 7-transmembrane receptor uti-
lized by T-tropic HIV strains to gain entry into T-cells. The appear-
ance of CXCR4 utilizing strains of HIV is associated with a decrease
in the number of T-cells and accelerated disease progression. In
We recently reported a structure–activity study toward improv-
ing the antiviral potency and/or ADME properties of AMD070
through iterative structural modifications. Specifically, we showed
that it was possible to shift the distal amine side chain from the
central nitrogen to either ring of the benzimidazole with retention
3
vitro studies have shown that addition of the natural CXCR4 ligand,
2
e
SDF-1, or small molecule antagonists, can block HIV infection.
SDF-1 is a highly basic protein with about 20% of its 68 amino acids
for SDF-1 ) being arginine, lysine or histidine. This observation,
6
of potent antiviral activity (Fig. 1, compounds 1 and 2). In a sepa-
(
a
rate report we described related efforts in which we replaced the
benzimidazole with an imidazopyridine in combination with the
along with the fact that the extracellular binding regions of the
CXCR4 receptor are particularly rich in aspartic acid and glutamic
7
side chain shift. The results of these studies prompted us to ex-
plore the possibility of replacing the benzimidazole with other het-
eroaromatic ring systems capable of maintaining the required
basic nitrogen triad.
*
Corresponding author. Tel.: +1 919 483 2750; fax: +1 919 315 6923.
E-mail address: john.6.miller@gsk.com (J.F. Miller).
0
960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.03.118

J. F. Miller et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3026–3030
3027
Table 1
distal
amino
group
Anti-HIV IC50s ± standard deviation (n) and CC50s for alternate lower heterocycle
derivatives
N
N
basic
nitrogen
triad
N
NH2
N
NH
AMD070
0 nM IC50
N
NH2
1
R
a
b
Compound
R
IC50 (nM)
CC50 (nM)
H
N
5c
23 ± 2 (4)
13,000
1
2
N
N
N
N
N
N
1
7 nM IC50
2 nM IC50
N
N
N
6d
7
70 ± 8 (4)
11,000
N
NH
N
>20,000 (1)
>20,000
N
N
8d
34 ± 7 (2)
12 ± 1 (2)
12,000
13,000
N
N
Figure 1. Tetrahydroquinolines with potent anti-HIV activity.
9
In order to identify viable alternative ring systems, we con-
ducted a scan of various heteroaryl fragments with the distal amine
side chain attached to the central nitrogen analogous to AMD070,
with the exception that this study was done in the racemic series.
Our synthetic approach is illustrated in Scheme 1. The key second-
ary amine intermediate 4 was synthesized by reductive amination
of ketone 3 with N-Boc-1,4-diaminobutane. The benzimidazole
derivative 5 was prepared via alkylation with N-Boc-2-(chloro-
methyl)benzimidazole followed by TFA deprotection. Compounds
10
11 ± 1 (2)
8200
8
N
Me
N
N
1e
1
120 ± 3 (2)
>20000 (2)
>20,000
>20,000
6–14 were prepared by reductive alkylations with the appropriate
heteroaryl aldehydes followed by HCl mediated cleavage of the
Boc protecting group. The benzothiazole analog 15 was synthesized
by alkylation with 2-(bromomethyl)-1,3-benzothiazole followed by
acidic deprotection.
Antiviral and cytotoxicity data for the alternate lower heterocy-
cle derivatives is shown in Table 1. Most of the ring systems stud-
ied retain appreciable antiviral activity with several in the same
potency range as the benzimidazole derivative 5 (racemic version
of AMD070). Exceptions include the 3-pyridyl and imidazole deriv-
atives 7 and 12, which due to the position of the basic ring nitro-
gens, are incapable of maintaining the required nitrogen triad
Me
N
12
N
N
13
42 ± 8 (2)
>20,000
9700
N
S
1
4d
340 ± 29 (2)
200 ± 12 (2)
N
S
1
5d
13,000
N
a
HOS cells expressing hCXCR4/hCCR5/hCD4/pHIV-LTR-luciferase, HIV-1, CXCR4
a
strain (IIIB). Compounds were tested for their ability to block infection of the HOS
cell line. IC50 is the concentration at which 50% efficacy in the antiviral assay was
N
N
9
observed.
b
O
N
CC50 is the concentration at which 50% cytotoxicity is observed in the HOS cell
3
4
H
NHBoc
line.
c
Compound previously reported in Ref. 10.
Compound previously reported in Ref. 11.
Compound previously reported in Ref. 12.
b,c for 5
d,e for 6-14
f,e for 15
d
e
N
N
geometry. The isoquinoline derivatives 8–10 were particularly
interesting with antiviral activities in the 10–30 nM range. A com-
parison of these compounds with the pyridyl derivative 6, shows
that benzo ring fusion confers a consistent improvement in activ-
ity. Compounds 14 and 15 show reduced activities relative to the
corresponding nitrogen analogs (i.e., 15 compared to 5), perhaps
due to the reduced basicity of the thiazole and benzothiazole ring
nitrogens.
NH2
R
5
-15
Scheme 1. Reagents and conditions: (a) N-Boc-1,4-diaminobutane, NaBH(OAc)
AcOH, 1,2-dichloroethane (73%); (b) N-Boc-2-(chloromethyl)benzimidazole, KI,
EtN, MeCN, rt (83%); (c) TFA, CH
,2-dichloroethane (55% for 6); (e) 4 N HCl/dioxane, MeOH (62% for 6); (f) 2-
3
,
(
1
(
iPr)
2
2
Cl
2
3
(71%); (d) RCHO, NaBH(OAc) , AcOH,
bromomethyl)-1,3-benzothiazole, KI, (iPr)
2
EtN, MeCN, 80 °C (59%).

3
028
J. F. Miller et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3026–3030
Having identified several suitable replacements for the benz-
imidazole ring system, we turned our attention to shifting the dis-
tal amine side chain from the central nitrogen to the lower
heterocycle. Based on a combination of attractive potency and syn-
thetic tractability, we chose the isoquinoline system 9 for further
study. We explored two different modes of attaching the amino
group to the 4-position of the isoquinoline ring system: 1- and 3-
carbon tethers (Fig. 2, general structure A).
N
Br
Me
N
N
CH3
N
N
a
N
18
H
CH3
N
b
Scheme 2 illustrates the synthesis of the 1-carbon primary
amine 23. In order to assess the pharmacodynamic effect of the
distal amine side chains, we also prepared the unsubstituted par-
ent compound 19 for comparison. The bromomethyl intermediate
16
17
19
c
1
7 was prepared by NBS bromination of commercially available
3
-methylisoquinoline 16. Reaction of 17 with the enantiopure tet-
Br
6
Me
N
N
rahydroquinoline derivative 18 gave 19 in high yield. The S abso-
lute configuration of 18 was chosen based on the clear activity
I
CN
N
preference observed in our previous studies.6 Substitution at
the 4-position of the isoquinoline was achieved by NIS iodination
to give intermediate 20, which was then subjected to Pd catalyzed
cyanation followed by NBS bromination. Benzylic bromide 21 was
reacted with 18 to afford nitrile 22 which was subjected to Raney
nickel reduction to give primary amine 23. Scheme 3 details the
subsequent side chain elaboration of 23 to give target compounds
,7
N
d,e
N
f
g
CH3
2
2
NH2
2
0
21
23
Scheme 2. Reagents and conditions: (a) NBS, AIBN, CCl
compound 18, (iPr) EtN, DMF, rt (98%); (c) NIS, glacial AcOH, 80 °C (56%); (d)
Zn(CN) Pd(Ph P) DMF, 120 °C (96%); (e) NBS, AIBN, CCl reflux (62%); (f)
compound 18, (iPr)
53%).
4
, reflux (66%); (b)
2
4–29.
Our synthetic route to the 3-carbon series is shown in
2
2
,
3
4
,
4
,
Scheme 4. Heck reaction of iodide 20 with acrylonitrile followed
by NBS bromination gave 30 as an E/Z isomer mixture. Base pro-
2
EtN, DMF, rt (92%); (g) H
2
3
(50 psi), Raney Ni, 2 M NH /MeOH
(
N
moted S 2 substitution of bromide 30 with 18 gave 31 in high
yield. Sodium borohydride reduction of 31 afforded nitrile 32,
which upon Raney nickel reduction gave 33. Primary amine 33
was then further elaborated as described in Scheme 4 to afford
compounds 34–37.
Table 2 shows SAR data for the 4-substituted isoquinoline
series. Interestingly, the unsubstituted analog 19 shows signifi-
cant activity even in the absence of a basic, distal amine side
chain. Attaching the 1-carbon primary amine moiety (compound
a
b,c
d
e,f,g
N
N
2
2
2
4
5
6
27
N
h,f,g
CH3
28
29
i,f,g
NH2
2
3) affords a sixfold increase in activity relative to 19. Dimethy-
lation of the amino group leads to a twofold loss in activity. A
comparison of 23 and 24 with the longer chain amino deriva-
tives 33 and 34, shows that the 3-carbon chain is favored, a re-
sult consistent with our previously reported chain length SAR in
the benzimidazole series.6 The equipotent guanidine derivatives
2
3
Scheme 3. Reagents and conditions: (a) 37% aqueous formaldehyde, NaBH(OAc)
AcOH, 1,2-dichloroethane (77%); (b) N,N -di-Boc-1H-pyrazole-1-carboxamidine, 1:1
THF/CH Cl2 (68%); (c) TFA, CH Cl2 (87%); (d) MeSO Cl, (iPr) EtN, CH Cl (70%); (e)
3
,
0
2
5 and 35 showed impressive activities and the best cytotoxicity
windows in the entire series. The equivalent activities of these
- and 3-carbon analogs contrasts with the modest but consis-
2
2
2
2
2
2
Boc-glycine, HATU, (iPr)
2
EtN, MeCN (72%); (f) TFA, CH
, AcOH, 1,2-dichloroethane (84% for 30); (h)
EtN, MeCN (63%); (i) Boc- -proline, HATU, (iPr) EtN,
2 2
Cl (89% for 30); (g) 37%
aqueous formaldehyde, NaBH(OAc)
Boc- -proline, HATU, (iPr)
MeCN (71%).
3
1
L
2
D
2
tent potency preference for the 3-carbon chain length that is
generally observed. Perhaps this is due to the ability of the gua-
nidine moiety to present two basic nitrogens in a 1,3-relation-
ship thereby allowing the proximal nitrogen of 35 to mimic
the distal nitrogen of 25. A surprising aspect of the SAR in this
series is the unexpectedly potent activity of the non-basic,
capped amine derivatives 26, 36 and 37. In fact, the 3-carbon
sulfonamide 37 is nearly as active as the corresponding primary
amine 33. Similarly interesting is the activity of nitrile deriva-
tives 22 and 32. This contrasts sharply with our previous obser-
vations in the benzimidazole series, where non-basic, acylated
amine side chains showed substantial reductions in activity.
Within the set of lower heterocycles that we have explored,
there is clearly something unique to the isoquinoline system
that apparently facilitates side chain-receptor interactions
through non-charged forces (i.e., hydrogen bond and/or dipole-
dipole interactions).
The amide derivatives 27–29 allowed us to examine 4-atom
tethers between the basic nitrogen and the ring system, albeit
in a more conformationally rigid manner. The dimethylaminogly-
cine analog 27 was quite potent with a 12 nM IC50. Proline
derivatives 28 and 29 show that conformationally constraining
the basic amine into a five-membered ring leads to fourfold
N
N
N
N
n=1,3
R1
N
N
NH2
(
CH2)n
N
R2
9
A
Figure 2.

J. F. Miller et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3026–3030
3029
Table 2
Anti-HIV IC50s ± standard deviation (n) and CC50
derivatives
s
for 4-substituted isoquinoline
Br
Me
N
N
I
CN
N
N
N
a,b
CH3
c
N
N
CN
N
R
2
0
30
3
1
a
b
f
Compound
19
R
IC50 (nM)
CC50 (nM)
3
4
N
N
H
CN
160 ± 19 (2)
72 ± 9 (2)
650
1300
g,h
22
N
35
d
e
CH3
3
1
32
i
j
23
NH2
N
27 ± 11 (2)
3100
3
3
6
7
NH2
24
48 ± 7 (5)
5900
NH
NH2
2
5
20 ± 5 (4)
41 ± 9 (4)
12 ± 4 (5)
9000
2000
480
3
3
N
H
Scheme 4. Reagents and conditions: (a) acrylonitrile, Pd(OAc)
00 °C (77%); (b) NBS, AIBN, CCl , reflux (57%); (c) compound 18, (iPr)
98%); (d) NaBH , iPrOH, rt, 5 days (71%); (e) H (1 atm), Raney Ni, 2 M NH
, 1,2-dichloroethane (81%); (g)
Cl (71%); (h) TFA, CH Cl
EtN, CH Cl (75%); (j) MeSO Cl,
2
, Ph
3
P, Et
EtN, DMF, rt
/MeOH
3
N, DMF,
O
S
1
(
(
4
2
26
27
N
H
CH3
4
2
3
70%); (f) 37% aqueous formaldehyde, NaBH(OAc)
3
O
O
0
N,N -di-Boc-1H-pyrazole-1-carboxamidine, 1:1 THF/CH
2
2
2
2
(
(
91%); (i) N,N-dimethylcarbamyl chloride, (iPr)
iPr) EtN, CH Cl (83%).
2
2
2
2
2
2
2
N
N
H
O
CH3
reduction in activity for the S-enantiomer 28 with the R-enantio-
mer 29 remaining unchanged relative to the open chain analog.
The 3-carbon amine derivative 34 is the most active analog in
the entire series with a 5 nM IC50 and a 130-fold cytotoxicity
window. In comparing the overall SAR of the isoquinoline series
2
2
8
9
N
44 ± 9 (2)
12 ± 5 (2)
540
690
N
H
O
CH3
N
6
N
H
with our previously reported benzimidazole and imidazopyri-
7
dine series, we were able to achieve similar antiviral activities,
CN
however, the isoquinolines show somewhat diminished cytotox-
icity windows.
32
21 ± 2 (2)
12 ± 2 (2)
950
3
3
3
4
NH2
1000
To further explore the therapeutic potential of the isoquinoline
series, a representative set of compounds was chosen for pharma-
cokinetic analysis in rats. Table 3 illustrates the PK parameters
associated with five selected analogs. All of the compounds
showed acceptable half life and clearance values and with the
exception of the guanidine derivative 25, they also showed at least
some level of systemic exposure. The dimethyl urea derivative 36
showed the best exposure with a 20% oral bioavailability, thus
demonstrating the potential of this structural class in regards to
lower species PK.
Using AMD070 as a starting point for structural modification,
we identified several viable heterocyclic replacements for the
benzimidazole ring system. The 3-isoquinolinyl moiety was chosen
for further structural manipulation, specifically by using our previ-
ously successful approach of shifting the distal amine side chain to
the lower heterocyclic ring system. This led to a number of com-
pounds with low nanomolar antiviral activities, and to the surpris-
ing observation of potent activity with non-basic side chain
derivatives. In addition, several analogs were shown to have
acceptable rat PK properties. While we were able to achieve po-
tency and pharmacokinetic profiles similar to our benzimidazole
and imidazopyridine series, the isoquinolines suffered from gener-
ally less favorable cytotoxicity windows. However, this exercise
effectively validates the concept of combining heterocyclic
replacements with side chain shifts to generate a pharmacologi-
cally viable and structurally novel chemotype in the tetrahydro-
quinoline series.
N
5.0 ± 2.0 (5)
17 ± 3 (2)
650
NH
O
35
6500
N
H
NH
2
36
38 ± 12 (4)
16 ± 2 (4)
750
690
N
H
N
O
S
3
7
N
H
CH3
O
a,b
See footnotes a and b from Table 1.
Table 3
Pharmacokinetic parameters of representative isoquinoline CXCR4 antagonists in rats
Compound
t
1/2 (h)
Cl (mL/min/kg)
V
d
(L/kg)
%F
2
2
34
36
5
7
12
16
11
6.1
16
9.5
9.0
1.5
2.9
1.7
1.4
0
10
6
20
4
1.5
2.9
3.1
5.8
3
7
d
Half life (t1/2), clearance (Cl) and volume of distribution (V ) calculated following
1 mg/kg IV doses. Percent oral bioavailability (%F) calculated following solution
doses of 3 mg/kg.

3
030
J. F. Miller et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3026–3030
5
.
.
Moyle, G.; DeJesus, E.; Boffito, M.; Wong, R. S.; Gibney, C.; Badel, K.; MacFarland,
R.; Calandra, G.; Bridger, G.; Becker, S. Clin. Infect. Dis. 2009, 48, 798.
Gudmundsson, K. S.; Sebahar, P. R.; Richardson, L. D.; Miller, J. F.; Turner, E. M.;
Catalano, J. G.; Spaltenstein, A.; Lawrence, W.; Thomson, M.; Jenkinson, S.
Bioorg. Med. Chem. Lett. 2009, 19, 5048.
Gudmundsson, K. S.; Boggs, S. D.; Catalano, J. G.; Svolto, A.; Spaltenstein, A.;
Thomson, M.; Wheelan, P.; Jenkinson, S. Bioorg. Med. Chem. Lett. 2009, 19, 6399.
Kelly, T. R.; Lebedev, R. L. J. Org. Chem. 2002, 67, 2197.
All compounds were also characterized in a cell fusion CXCR4 receptor binding
assay. The resulting SAR tracked closely with the HOS antiviral data. For
example, compounds 5, 27 and 34 showed IC50s in the receptor binding assay
of 36, 16, and 3.4 nM, respectively. In addition, representative compounds were
characterized in a cell based functional assay and were shown to be non-
competitive antagonists of the CXCR4 receptor. For assay protocols see:
Gudmundsson, K.; Miller, J. F.; Turner, E. M. International Patent WO 2006/
020415, 2006.
Acknowledgments
6
We thank Richard Hazen, Wendell Lawrence, David McCoy for
providing HIV-1 antiviral data.
7
.
8
9
.
.
References and notes
1
2
.
.
Agrawal, L.; Lu, X.; Jin, Q.; Alkhatib, G. Curr. Pharm. Des. 2006, 12, 2031.
(a) Qian, K.; Morris-Natschke, S. L.; Lee, K. Med. Res. Rev. 2009, 29, 369; (b)
Kuritzkes, D. R. Curr. Opin. HIV AIDS 2009, 4, 82; (c) Kazmierski, W. M.;
Gudmundsson, K. S.; Piscitelli, S. C. Ann. Rep. Med. Chem. 2007, 42, 301; (d)
Mosley, C. A.; Wilson, L. J.; Wiseman, J. M.; Skudlarek, J. W.; Liotta, D. C. Expert
Opin. Ther. Patents 2009, 19, 23; (e) Grande, F.; Garofalo, A.; Neamati, N. Curr.
Pharm. Des. 2008, 14, 385.
10. Bridger, G. J.; Skerlj, R. T.; Kaller, A.; Harwig, C.; Bogucki, D.; Wilson, T.;
Crawford, J.; McEachern, E. J.; Atsma, B.; Nan, S.; Zhou, Y.; Schols, D.; Smith, C.
D.; Di Fluri, R. M. International Patent WO 2003/055876, 2003.
11. Bridger, G.; McEachern, E. J.; Skerlj, R.; Schols, D. U.S. Patent US 2004/020992,
2004.
12. Bridger, G.; Kaller, A.; Harwig, C.; Skerlj, R.; Bogucki, D.; Wilson, T. R.; Crawford,
J.; McEachern, E. J.; Atsma, B.; Nan, S.; Zhou, Y.; Schols, D.; Smith, C. D.; Di Fluri,
M. R. U.S. Patent US 2004/0019058, 2004.
3
.
.
Schuitemaker, H.; Koot, M.; Kootstra, N. A.; Dercksen, M. W.; de Goede, R. E.;
van Steenwijk, R. P.; Lange, J. M.; Schattenkerk, J. K.; Miedema, F.; Tersmette, M.
J. Virol. 1992, 66, 1354.
(a) Dealwis, C.; Fernandez, E. J.; Thompson, D. A.; Simon, R. J.; Siani, M. A.; Lolis,
E. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 6941; (b) Hatse, S.; Princen, K.; Gerlach,
L.; Bridger, G.; Henson, G.; De Clercq, E.; Schwartz, T. W.; Schols, D. Mol. Pharm.
4
2
001, 60, 164.