Naphthyridinone (NTD) integrase inhibitors 4. Investigating N1 acetamide substituent effects with C3 amide groups

Article abstract of DOI:10.1016/j.bmcl.2014.05.011

A series of N1 acetamide substituted naphthyridinone HIV-1 integrase inhibitors have been explored to understand structure-activity relationships (SAR) with various C3 amide groups. Investigations were evaluated using integrase enzyme inhibition, antiviral activity and protein binding effects to optimize the sub-structures. Lipophilicity was also incorporated to understand ligand lipophilic efficiency as a function of the structural modifications. Three representative analogs were further examined in a peripheral blood mononuclear cell (PBMC) antiviral assay as well as in vitro and in vivo drug metabolism and pharmacokinetic studies.

Full text of DOI:10.1016/j.bmcl.2014.05.011

Accepted Manuscript
Naphthyridinone (NTD) integrase inhibitors 4. Investigating N1 acetamide sub-
stituent effects with C3 amide groups
Brian A. Johns, Takashi Kawasuji, Jason G. Weatherhead, Eric E. Boros, James
B. Thompson, Cecilia S. Koble, Edward P. Garvey, Scott A. Foster, Jerry L.
Jeffrey, Tamio Fujiwara
PII:
S0960-894X(14)00498-3
DOI:
http://dx.doi.org/10.1016/j.bmcl.2014.05.011
BMCL 21624
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
21 February 2014
29 April 2014
5 May 2014
Please cite this article as: Johns, B.A., Kawasuji, T., Weatherhead, J.G., Boros, E.E., Thompson, J.B., Koble, C.S.,
Garvey, E.P., Foster, S.A., Jeffrey, J.L., Fujiwara, T., Naphthyridinone (NTD) integrase inhibitors 4. Investigating
N1 acetamide substituent effects with C3 amide groups, Bioorganic & Medicinal Chemistry Letters (2014), doi:
http://dx.doi.org/10.1016/j.bmcl.2014.05.011
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Naphthyridinone (NTD) integrase inhibitors 4. Investigating N1
acetamide substituent effects with C3 amide groups
Brian A. Johns,a* Takashi Kawasuji,b Jason G. Weatherhead,a Eric E. Boros,a James B.
Thompson,a Cecilia S. Koble,a Edward P. Garvey,a† Scott A. Foster,a† Jerry L. Jeffrey,a
Tamio Fujiwara b
GlaxoSmithKline Research & Development, Infectious Diseases Therapy Area Unit,a Five Moore Drive, Research Triangle
Park, NC 27709, USA, and Shionogi Pharmaceutical Research Center, Shionogi & Co., Ltd.,b 3-1-1, Futaba-cho,
Toyonaka-shi, Osaka 561-0825, Japan
This is where the receipt/accepted dates will go; Received Month XX, 2000; Accepted Month XX, 2000 [BMCL RECEIPT]
Abstract— A series of N1 acetamide substituted naphthyridinone HIV-1 integrase inhibitors have been explored to understand structure-
activity relationships (SAR) with various C3 amide groups. Investigations were evaluated using integrase enzyme inhibition, antiviral
activity and protein binding effects to optimize the sub-structures. Lipophilicity was also incorporated to understand ligand lipophilic
efficiency as a function of the structural modifications. Three representative analogs were further examined in a peripheral blood
mononuclear cell (PBMC) antiviral assay as well as in vitro and in vivo drug metabolism and pharmacokinetic studies.
We have previously reported on 7-benzyl substituted
naphthyridinone (NTD) HIV-1 integrase inhibitors.1,2,3
This hetereocyclic template contains the requisite two-
metal chelation elements necessary to bind to the essential
enzyme active site Mg2+ metals and has been shown to be
intrinsically potent as an integrase strand-transfer inhibitor.
This provided a strong baseline to justify further study,
however in order to be of use as a potential medicine,
numerous other considerations required careful
optimization. Our strategy was to explore the impact of
changes to the N1 and C3 positions of the NTD template
which were convenient chemical handles for tractable
modifications, while not altering the basic metal-binding
pharmacophore or 7-benzyl elements. A methodical
evaluation of N1 protio and methyl substituents1 resulted
in the discovery of GSK3647354,5 (1) which was
progressed into early stage clinical trials. During these
structure-activity relationship (SAR) studies, we became
interested in exploring additional functionality at the N1
and C3 positions; namely how these combinations would
impact both the potency of inhibiting the viral enzyme
target and also the drug properties of the molecules
including the pharmacokinetic profiles, as well as ligand
efficiency measurements.6 The focus of this report is N1
acetamide substitutions (e.g. 2) that were envisioned to
provide improved solubility and an overall reduction in the
hydrophobicity of the NTD drug molecules (Figure 1).
Figure
1 Naphthyridinone (NTD) HIV-1 integrase strand-transfer
inhibitors.
The synthetic methods used to prepare the title compounds
in Tables 1-3 began with the substituted 3-amino-2-
pyridinecarboxylate ester 32,7 (Scheme 1). Formation of
the Schiff base with glyoxal followed by a one pot
reduction using NaBH3CN provided the acetic acid adduct
4. Amide formation using HATU coupling conditions
smoothly provided amide 5 which was acylated using
chloro ethyl malonate under thermal conditions in the
absence of a base. The intermediate mixed malonate ester
amide was subjected to NaOEt in ethanol to facilitate a
Dieckmann cyclization to give the naphthyridinone ring
E-mail address: brian.a.johns@gsk.com
Tel: 919-483-6006
Fax: 919-483-6053

system 6. This was converted to the C3 amide through
direct transamination using a primary amine and either oil
bath or microwave heating. In some cases the amines were
used neat or ethanol was used as the solvent to afford the
final neutral free acid NTD target molecules 7-42.
substitutions (10-14, 16) provided similar antiviral
potencies and protein-adjusted activities. The lack of
improvement as a result of the added lipophilicity was
detrimental to the ligand efficiency estimates making these
analogs less attractive. The N-cyclopropyl analog 15
showed a lower protein shift and increased LLE, however
the trend did not continue with the homologated example
17, which had a 4-fold lower PAIC50. The electron
withdrawing effects of the trifluoroethyl analog 18 did not
impact either activity or lipophilicity substantially
compared to the ethyl derivative 10. Finally the addition
of neutral polar groups in example 19 and 21 resulted in
good activity while decreasing the clogP values and an
increase in the LLE assessment. However, the basic tail in
the dimethylaminoethyl analog 20 had a negative impact
on the enzymatic and antiviral potencies but did reduce the
protein shift to 2-fold.
Table 1. N1 acetamide/C3 methyl amide SAR.
Scheme 1. Reagents and conditions: (a) EtOH, Δ; NaBH3CN, r.t. (b)
HNRR’, HATU, Et3N, DMF (c) ClC(O)CH2CO2Et; (d) DCE, ; (d)
NaOEt, EtOH; (e) R3NH2, EtOH, orW.
Initially we performed a survey of N1 acetamide groups
while maintaining an N-methyl amide at the 3 position of
the NTD ring system (Table 1). We elected to maintain the
7-(4-fluoro)benzyl moiety consistently throughout the
study as this group was believed to be a suitably optimal
group from our previous naphthyridinone studies as well as
related systems8,9. This also minimized the number of
variables to contend with as we explored substituent
effects at the N1 and C3 positions. The assays used to
evaluate new compounds consisted of a biochemical
enzymatic integrase strand-transfer assay (STIC50)10 and a
nM)
pHIVIC50 pHIVPAIC50 LLE
Compd
STIC50
13
7
8.1
7.6
1.7
2.2
20
56
5.8
6.6
-
8
9
9.3
9.3
8.2
4.4
3.3
5.9
2.9
9.2
22
19
11
17
5.6
5.3
5.2
4.5
single-round pseudotyped antiviral assay utilizing
a
10
11
luciferase reporter (pHIVIC50).11 Additionally, the pHIV
assay provided a convenient means to estimate protein
binding effects (i.e. fold shift) on the antiviral activity
through the addition of purified human serum albumin
12
13
13
4.9
16
4.8
(HSA), giving rise to
(
a
protein-adjusted IC50
9.2
5.5
14
4.6
pHIVPAIC50).12 All compounds in the series had a
substantial therapeutic index when comparing antiviral
potency versus cellular toxicity (data not shown). We have
also used ligand lipophilicity efficiency (LLE = pIC50-
clogP)13,14 which is a key measure of true potency
improvement and reduced toxicity risk. In general, it is
desirable to have the LLE >5.
14
15
7.8
4.4
3.2
2.6
12
4.9
5.4
7.6
16
17
6.9
8.0
4.8
5.9
16
33
4.6
4.7
We began by systematically examining the simplest
primary, secondary and tertiary amides (7-9). All three
variations delivered potent enzyme activity. Likewise the
antiviral activities were within 2-fold amongst the three
amides, however the protein-shift values varied from 28-
fold for the primary amide to only 3-fold for the tertiary
amide. The less lipophilic primary amide also had the
highest LLE as a result of a decreased clogP value.
Additional simple alkyl secondary and tertiary amide
18
19
20
21
7.1
6.8
22
1.5
3.7
42
15
91
83
13
5.0
5.3
4.6
5.6
5.7
3.3

This initial assessment suggested it would be feasible to
replace the N1-Me of the original analog 1 with more polar
functionality and retain similar or improved potency.
These effects were further probed through modifications of
the C3 amide moiety. The dimethyl amide group in 9
appeared to have a good balance of potency and
lipophilicity as indicated in the LLE value and was chosen
as the basis for the amide survey.
basic amine 31 had a slightly lower antiviral activity, but
the lack of any shift in the presence of added serum
albumin was noteworthy.
Table 3. NTD acetamide/R3 methoxyethyl amide SAR .
Table 2. N,N-dimethylacetamide/C3 amide SAR.
nM)
pHIVIC50 pHIVPAIC50 LLE
Compd
STIC50
24
32
33
4.0
7.7
9.0
3.1
4.3
3.0
9.0
22
14
5.4
5.9
5.5
nM)
pHIVIC50 pHIVPAIC50 LLE
Compd
STIC50
9.3
9
5.9
9.3
19
5.3
5.1
34
35
7.3
8.5
3.0
3.3
18
39
5.0
4.5
22
3.0
193
23
24
5.7
4.0
17
35
6.0
5.4
36
37
6.2
11
3.2
2.4
57
20
4.8
4.6
3.1
9.0
25
26
27
8.6
7.1
6.5
31
17
40
24
82
5.6
5.6
5.2
38
39
40
41
42
45
42
27
3.8
3.6
3.5
5.3
124
11
4.1
4.3
4.4
5.0
4.0
3.6
4.7
7.6
28
12
6.7
13
28
23
7.2
24
3.9
29
30
9.2
11
70
5.5
The C3 2-methoxymethyl amide 24 was of particular
interest due to a low (3-fold) protein shift and a PAIC50
under 10 nM. We decided to re-examine the impact of the
acetamide substitution in the context of the 2-
methoxyethyl amide as shown in Table 3. The rank
ordering of potency was consistent with the N-methyl
amides in Table 1 wherein the primary amide (32) had the
least potent PAIC50 followed by the secondary amide 33
and then the tertiary amide 24 with the most potent
PAIC50. The SAR findings largely followed the results in
Table 1 wherein additional secondary acetamides did not
provide any substantial changes in protein-adjusted
potency while the basic amine group (e.g. 20 and 38) again
led to the largest loss in potency. A group of cyclic
tertiary amides were also studied resulting in similar
potency results compared to 24 but offering little
discernible advantage in terms of ligand lipophilicity
efficiency measures.
4.1
7.1
15
5.8
31
8.6
19
25
4.9
The N-cyclopropyl group again had an influence on the
protein-binding effects, although this time a near 20-fold
reduction in potency was observed for compound 22
(Table 2). The more polar 2-hydroxyethyl amide similar to
that of 1 was installed to give 23. The overall activity was
very similar with a protein-adjusted activity of 35 nM (vs.
20 nM for GSK735), however the lower clogP value for
the methyl to acetamide change resulted in an increase in
ligand lipophilicity efficiency. Capping the hydroxyl
group as the 2-methoxyethyl amide resulted in an
improvement in the protein-adjusted potency for 24. Other
hydroxyethyl derivatives (25-26, 28-30) showed lower
activity or did not improve over the protein-adjusted
activity of 24. The geminal-dimethyl analog 27 was
similarly potent to the methoxymethyl analog 24. The
Three analogs (9, 24, and 33) were selected for further
antiviral evaluation as well as in vitro and in vivo
pharmacokinetic studies. Table 4 shows the antiviral

activity in
a
7-day multi-round peripheral blood
33
33
33
rat
dog
cyno
0.5
13.8
2.0
3.7
1.2
2.3
51%
111%
19%
56
0.1
0.8
mononuclear cell (PBMC) assay. All three analogs had
slightly improved antiviral IC50 values in the PBL assay
when compared to the screening pHIV system. Applying
the fold-shift from the pHIV conditions provided a
predicted PBLPAIC50. Additionally, a 4-fold multiplier was
applied to empirically translate the IC50 to IC90 and thus
predict a clinical trough concentration target PAIC90.
ND = no drug detected at 24h
The results created further interest in exploring other
functionality at the N1 position. Further SAR studies to
optimize the potency, protein binding, ligand efficiencies,
and pharmacokinetics of this series and achieve a potential
for once-daily dosing will be the focus of subsequent
reports.
Table 4. Peripheral blood mononuclear cell (PBMC) antiviral activity
and determination of protein adjusted IC90 clinical trough concentration
targets.
PBLIC50
(nM)
Fold PBLPAIC50 PBLPAIC90
Shifta
3.3
(nM)b
14
(nM)c
55
Compd
References and Notes
9
4.2
4.3
2.3
24
33
2.9
13
50
† Current address for EPG is Viamet Pharmaceuticals, Inc.,
Durham, NC; and current address for SAF is Chimerix,
Inc., Durham, NC.
4.5
10
41
a Fold shift determined from pHIV assay.
b PBLPAIC50 PBLIC50 x fold shift
c PBLPAIC90 = 4 x PBLPAIC50
=
1
For the previous account in this series see: Johns, B. A.;
Kawasuji, T.; Weatherhead, J. G.; Boros, E. E.; Thompson,
J. B.; Koble, C. K.; Garvey, E. G.; Foster, S. A.; Jeffrey, J.
L.; Fujiwara, T. Bioorg. Med. Chem. Lett. 2013, 23, 422.
Compounds 9 and 24 showed no inhibition of CYP450’s
up to 33µM while 33 inhibited CYP1A2 at a level of
4.7µM. All three had good metabolic stability across rat,
dog, cyno and human S9 preparations. Each of the three
compounds were dosed in rats, beagle dogs and
cynomolgus monkeys via IV (1mg/kg) and oral (5 mg/kg)
routes of administration (Table 5). Oral dosing was
2
Boros, E. E.; Edwards, C. E.; Foster, S. A.; Fuji, M.;
Fujiwara, T.; Garvey, E. P.; Golden, P. L.; Hazen, R. J.;
Jeffrey, J. L.; Johns, B. A.; Kawasuji, T.; Kiyama, R.;
Koble, C. S.; Kurose, N.; Miller, W. H.; Mote, A. L.;
Murai, H.; Sato, A.; Thompson, J. B.; Woodward, M. C.;
Yoshinaga, T. J. Med. Chem. 2009, 52, 2754.
performed using
meglumine vehicle solution formulation.
a
10/10/80 DMSO/solutol/0.05M
All three
3 Johns, B. A.; Kawasuji, T.; Weatherhead, J. G.; Boros, E.
E.; Thompson, J. B.; Garvey, E. P.; Foster, S. A.; Jeffrey,
J. L.; Miller, W. H.; Kurose, N.; Matsumura, K.; Fujiwara,
T. Bioorg. Med. Chem. Lett. 2011, 21, 6461.
compounds had low clearance in rats while in dogs they
had slightly higher clearance ranging from 19% to 35% of
hepatic blood flow. Oral bioavailability values for 9 were
the most consistent across the three species suggesting no
permeability or solubility limitations in a therapeutic dose
range. However, the two methoxyethyl amide analogs 24
and 33 had a much larger range of the fraction absorbed
across the species with no clear pattern. We had a keen
interest in the concentration of drug at 24 hours post dose
(C24) relative to the PBLPAIC90. This serves as a crude
estimation of the potential inhibitory quotient (IQ)15 and
ultimately a property that would be expected to be a key
driver of efficacy. For the compounds studied, these
values were encouraging in rodent studies, especially in
the case for 33. However in higher species, the half-life of
each of these was simply not sufficient to support robust
coverage of the potency target suggesting these molecules
would likely require a twice-daily dose.
4 Garvey, E. P.; Johns, B. A.; Gartland, M. J.; Foster, S. A.;
Miller, W. H.; Ferris, R. G.; Hazen, R. J.; Underwood, M.
R.; Boros, E. E.; Thompson, J. B.; Weatherhead, J. G.;
Koble, C. S.; Allen, S. H.; Schaller, L. T.; Sherrill, R. G.;
Yoshinaga, T.; Kobayashi, M.; Wakasa-Morimoto, C.;
Miki, S.; Nakahara, K.; Noshi, T.; Sato, A.; Fujiwara, T.
Antimicrob. Agents Chemother. 2008, 52, 901.
5
Nakahara, K.; Wakasa-Morimoto, C.; Kobayashi, M.;
Miki, S.; Noshi, T.; Seki, T.; Kanamori-Koyama, M.;
Kawauchi, S.; Suyama, A.; Fujishita, T.; Yoshinaga, T.;
Garvey, E. P.; Johns, B. A.; Foster, S. A.; Underwood, M.
R.; Sato, A.; Fujiwara, T. Antiviral Res. 2009, 81, 141.
5
Kobayashi, M.; Nakahara, K.; Seki, T.; Miki, S.;
Kawauchi, S.; Suyama, A.; Wakasa-Morimoto, C.;
Kodama, M.; Endoh, T.; Oosugi, E.; Matsushita, Y.;
Murai, H.; Fujishita, T.; Yoshinaga, T.; Garvey, E.; Foster,
S.; Underwood, M.; Johns, B.; Sato, A.; Fujiwara, T.
Antiviral Res. 2008, 213.
Table 5. Rat in vivo pharmacokinetic screening results16
Clp
(mL/min/kg)
IV T1/2
(hr)
Compd Species
%F
C24/PBLPAIC90
6
Leeson, P. D.; Empfield, J. R. Ann. Rep. Med. Chem.
9
9
rat
dog
cyno
rat
0.7
10.9
6.3
13.8
1.3
69%
75%
60%
59%
7%
2.4
0.1
0.7
0.5
ND
0.6
2010, 45, 393.
7 Boros, E. E.; Burova, S. A.; Erickson, G. A.; Johns, B.
A.; Koble, C. S.; Kurose, N.; Sharp, M. J.; Tabet, E. A.;
Thompson, J. B.; Toczko, M. A. Org. Process Res.
Dev.2007, 11, 899.
9
5.0
24
24
24
3.6
12.5
1.8
dog
cyno
19.5
9.2
8
Johns, B. A.; Svolto, A. M. Expert Opin. Ther. Patents
5.4
31%
2008, 18, 1225

9
Pendri, A.; Meanwell, N. A.; Peese, K. M.; Walker, M.
A. Expert Opin. Ther. Patents 2011, 21, 1173.
10
Recombinant HIV-1 integrase strand transfer assay.
Boros, E. E.; Johns, B. A.; Garvey, E. P.; Koble, C. S.;
Miller, W. H. Bioorg. Med. Chem. Lett. 2006, 16, 5668.
11 Pseudo-type HIV assay (PHIV). See reference 4.
12
pHIVPAIC50 is determined using 40 mg/mL of purified
human serum albumin in the pHIV assay.
13
Leeson, P. D.; Springthorpe, B. Nat. Rev. Drug. Disc.
2007, 6, 881.
STPIC50 values were used for LLE
calculations.
14
LLE is also referred to as LipE, see Ryckmans, R.;
Edwards, M. P.; Horne, V. A>; Correia, A. M.; Owen, D.
R.; Thompson, L. R.; Tran, I.; Tutt, M. F.; Young, T.
Bioorg. Med. Chem. Lett. 2009, 19, 4406.
15 Inhibitory quotient (IQ) is defined as ratio of the drug
trough concentration in plasma and a measurement of viral
susceptibility. For the purposes of this study we have
chosen to define viral susceptibility as the PAIC90 (An IQ
of 1 would be a trough drug concentration equal to the
PAIC90).
16 Oral doses of 5 mg/kg were delivered as a
DMSO/solutol/water solution formulation. IV data was
from a 1 mg/kg bolus injection. Studies were performed in
fasted animals.

Naphthyridinone (NTD) integrase inhibitors 4. Investigating N1
acetamide substituent effects with C3 amide groups
Brian A. Johns,* Takashi Kawasuji, Jason G. Weatherhead, Eric E.
Boros, James B. Thompson, Cecilia S. Koble, Edward P. Garvey,
Scott A. Foster, Jerry L. Jeffrey, Tamio Fujiwara