An in situ oxidation strategy towards overcoming hERG affinity

Article abstract of DOI:10.1016/j.bmcl.2010.09.089

In an effort to overcome hERG affinity with a lead compound, several S-oxide and N-oxide analogues were synthesised with a much improved hERG profile but low in vivo absorption. This led to the implementation of an in situ oxidation strategy wherein a sulfide was dosed orally and systemic levels of the corresponding sulfoxide and sulfone were monitored. SAR and pharmacokinetic data to support this as a possible strategy are presented, although ultimately the approach was shown not to be suitable due to very low levels of active circulating metabolites.

Full text of DOI:10.1016/j.bmcl.2010.09.089

Bioorganic & Medicinal Chemistry Letters 20 (2010) 6400–6404
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
An in situ oxidation strategy towards overcoming hERG affinity
David C. Pryde a, , Rhys Jones b, Donald S. Middleton a, Ben J. Laverty b, David R. Fenwick a, Helen J. Mason a,
⇑
Martin Corless a, Nick N. Smith a
a Worldwide Medicinal Chemistry, Pfizer Global Research and Development, Ramsgate Road, Sandwich, Kent CT13 9NJ, United Kingdom
b Pharmacokinetics, Dynamics and Metabolism, Pfizer Global Research and Development, Ramsgate Road, Sandwich, Kent CT13 9NJ, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
In an effort to overcome hERG affinity with a lead compound, several S-oxide and N-oxide analogues
were synthesised with a much improved hERG profile but low in vivo absorption. This led to the imple-
mentation of an in situ oxidation strategy wherein a sulfide was dosed orally and systemic levels of the
corresponding sulfoxide and sulfone were monitored. SAR and pharmacokinetic data to support this as a
possible strategy are presented, although ultimately the approach was shown not to be suitable due to
very low levels of active circulating metabolites.
Received 9 August 2010
Revised 14 September 2010
Accepted 15 September 2010
Available online 21 September 2010
Keywords:
CCR5
Ó 2010 Elsevier Ltd. All rights reserved.
hERG
In situ oxidation
Prodrug
HIV
The advent of Highly Active Anti-RetroViral Therapy (HAART)
has significantly changed the face of managing HIV infection,1
and its success in suppressing viral replication and reducing HIV
associated deaths is due to the combined use of agents which tar-
get up to three different viral targets.2 Nonetheless, the emergence
of resistance to some of the components of HAART, and the toxicity
of some of the currently available drugs has necessitated the con-
tinuance of the search for new effective antiviral agents, and in
particular of new mechanistic classes directed to both virus and
host targets with superior efficacy and/or safety profiles.3
One such target is CCR5, a host chemokine receptor expressed
on white blood cells, which functions as a co-receptor for the
HIV-1 virus.4 Homozygotes which lack the CCR5 receptor have
demonstrated very high resistance to infection, and blocking of this
receptor with antagonists has been shown to produce a potent
antiviral response in patients.5 Previous publications from Pfizer
have detailed the discovery programme in which the high through-
put screening lead 1 was optimised through to maraviroc 2 (see
Fig. 1), which was launched several years ago as the first-in-class
CCR5 antagonist for the treatment of HIV.6
Whilst the main direction of the programme took us from 1 to
2,6 we wish to describe in this Letter a parallel medicinal chemistry
effort in which attempts to improve the pharmacokinetic and car-
diac safety properties of 3 were made by specifically targeting po-
lar, particularly oxidised, analogues of the tetrahydropyran ring.
This effort was initiated by the observation that charged analogues
F
F
N
O
N
N
HN
N
N
N
1
N
2
O
During this programme, an extensive medicinal chemistry
effort initially identified the tetrahydropyran derivative 3 as a
promising lead compound on the basis of high antiviral activity,
but suffered from high metabolic instability and off-target affinity
for the hERG channel.7
O
HN
N
N
N
3
⇑
Corresponding author.
E-mail address: david.pryde@pfizer.com (D.C. Pryde).
Figure 1. A high throughput screening hit (1) for the Pfizer medicinal chemistry
programme which led to maraviroc (2) and an intermediate lead compound (3).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.09.089

D. C. Pryde et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6400–6404
6401
Table 1
O
S+
Oxygenated amide substituent SAR based on the lead 4
gp160 IC50 0.2 nM
RRCK 0.8 x 10-6 cm/s
log D 2.3
O
HLM Clint 38 µL/min/mg
hERG IC50 >10 µM
HN
O
N
HN
Rat PK (1mg/kg IV,
2mg/kg PO)
N
N
N
Cl 104 mL/min/kg
F
N
N
Vd 23 L/kg
T1/2 2.6 h
F <1 %
F
4
Compound Structure
c log P gp160 fusion inhibition (IC50,
Figure 2. Profile of a lead sulfoxide analogue 4. Abbreviations: RRCK, A sub clone of
MDCK cells known to express low levels of p-glycoprotein used to assess membrane
permeability; HLM, human liver microsomes; hERG, human ether-a-go-go related
gene; Cl, clearance; F, oral bioavailability; Vd, volume of distribution; Caco-2,
human epithelial cell line derived from a colon carcinoma; HPLC, high pressure
liquid chromatography. Measured partition coefficients were generated at pH7.4 in
mixtures of 1-octanol and 0.1 M sodium phosphate buffer using the shake-flask
method.
nM)
O
S+
4
1.8
1.7
0.2
0.2
O
S
O
5
6
O
O
of 3 such as the tetrahydrothiopyran S-oxide 4 displayed excellent
cell-based activity8 and selectivity over the hERG channel (Fig. 2).
Compound 4 was taken into an in vivo rat pharmacokinetic
study, but suffered from high clearance and very low oral bioavail-
ability. Several analogues of this sulfoxide group, for example, the
related sulfone and the analogous pyridine N-oxides were then
synthesised with the purpose of identifying an analogue with im-
proved in vivo pharmacokinetics, while retaining the favourable
potency and hERG window of 4.
The chemistry to prepare the analogues described in this paper
is entirely analogous to that described previously for the maraviroc
programme. S- and N-oxidations were preferably carried out at a
late stage on the corresponding sulfide or pyridine derivatives. In
the case of the sulfoxides, diastereoisomers were separated by re-
verse phase HPLC. Starting materials for the oxidations were con-
structed from the appropriate aldehyde A and amine B which
were prepared according to the published methods.9,10 An example
of a complete synthetic procedure is detailed in Scheme 1 below.
Simple oxidised analogues were targeted first and initially as-
sessed for potency in a cell–cell fusion assay8 as shown in Table
1. The sulfoxide 4 and sulfone 5 derivatives showed equivalent
activity. The alternative aromatic sulfone 6 was some 200-fold less
S
4.2
40
S+
7
8
3.2
3.1
>100
4
O
O
O
N+
N+
9
3.1
3.1
5
O
N+
10
>100
active, and the aliphatic sulfoxide 7 was more than 500 times less
potent than 4. The isomeric pyridine N-oxides 8–10 were all signif-
icantly less active than the starting point 4, with the o-isomer 10
significantly weaker than the m and p isomers 8 and 9.
S
BocNH
CHO
HN
O
i, ii, iii
N
N
HN
+
N
A
N
N
B
F
31
F
v
O
iv
O
O
S+
S
O
O
HN
HN
N
N
N
N
N
N
F
5
F
4
Scheme 1. Reagents and conditions: (i) NaBH(OAc)3, AcOH, DCM (80%); (ii) 4 M HCl in dioxane (95%); (iii) EDCI, HOBT, tetrahydro-2H-thiopyran-4-carboxylic acid, DCM
(65%); (iv) NaIO4, 4:1 MeOH/water (40%); (v) H2O2 (30 wt % in water) (77%).

6402
D. C. Pryde et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6400–6404
Table 2
whereas insertion of a fluorine atom in the same position in 13
Heterocyclic tropane SAR based on the lead 4
gave a 20-fold increase in potency. Interestingly the same function-
ality of compounds 12 and 13 applied to the 5-position (viz. 11 and
14) gave subtly different results, while the ethyl-substituted benz-
imidazole 15 gave improved potency, albeit in a more lipophilic
structure. Overall, all benzimidazole substituents in Table 2 re-
tained good levels of potency, and at this stage did not rule out
any analogues, but rather prompted us to investigate combinations
of the functionality described within Tables 1 and 2 to identify
optimal compounds. In this phase of the investigation, we also pro-
filed each compound for its affinity for the hERG channel as shown
in Table 3. In general, the 6-fluoro group exemplified in 13, 17, 22
and 29 supported higher fusion potency than the other tropane
substituents. Also in general, the phenyl sulfone, phenyl sulfoxide
and pyridine N-oxide groupings in 6, 8, 9 and 20–30 all carried
varying degrees of hERG affinity, with the p-pyridine N-oxide
group of compounds 8, 26, 27 and 28 showing the greatest affinity
amongst this set of compounds. These data focussed attention on
both the tetrahydrothiopyran S-oxide and S,S-dioxide analogues
of compounds 4, 5 and 12–19. These compounds were all 1 nM
or better inhibitors of cell-cell fusion with no measurable inhibi-
O
S+
O
HN
N
F
Compound
Structure
c log P
gp160 fusion inhibition (IC50, nM)
N
N
4
1.8
0.2
tion of the hERG channel at 10 lM concentration. To assess if
N
N
11
12
0.7
0.7
4
1
any of these compounds offered a more appealing profile than 4,
in vitro microsomal stability and permeability was evaluated for
this subset of compounds, as shown in Table 4.
N
From Tables 3 and 4, compound 17 had excellent potency and
hERG selectivity and subsequently found to possess reasonable
HLM stability (Fig. 3). While the Caco-2 and RRCK measured per-
meability was not ideal and arguably not significantly different
to that of the starting compound 4, it was decided to take 17 into
an in vivo study to obtain in vivo confirmation of its absorption
profile. The rat pharmacokinetics of 17 was assessed and 17 was
found to have similarly high Cl and low oral bioavailability to 4.
A portal vein/systemic crossover study in rat confirmed that the
absorption of 17 was very low, approximately 3%.
N
N
N
N
N
13
2.0
0.01
The in vitro data for both 4 and 17 was therefore, while encour-
aging, not borne out by sufficient in vivo absorption in rat. At this
point, we returned to the starting point of this programme of work,
namely compound 3, and considered alternative ways in which we
could access the favourable permeability properties of a compound
like 3 (Caco-2 AB/BA 23/42, RRCK 9) with the potency and hERG
selectivity profile of 4. This prompted us to look at the lower oxi-
dation analogue of 4, namely the sulfide 31.
Our thinking was that the lipophilic, membrane-permeable sul-
fide 31, albeit hERG-liable, could be dosed and following rapid and
complete absorption, would then undergo rapid first-pass hepatic
oxidation to a much more selective, metabolically robust deriva-
tive which would retain antiviral activity. This strategy is summa-
rised in Figure 4.
F
N
N
14
15
2.0
2.4
0.4
F
N
N
0.05
Compound 31 was indeed confirmed as possessing high in vitro
metabolic instability and high permeability, and as suspected, 31
also had significant hERG affinity. The in vitro microsomal oxida-
tion of 31 was also confirmed as resulting in the generation of both
4 and 5 as primary metabolites.
This initial tranche of compounds showed that good levels of
potency were attainable with a variety of amide substituents,
and excluded the aliphatic sulfoxide group in 7 and the o-pyridine
N-oxide in 10 from further consideration. Also apparent was that
the tetrahydrothiopyran template for either a sulfoxide or sulfone
was especially favoured for potency. Retaining the sulfoxide group
in this template, our attention turned to the benzimidazole group,
and a range of benzimidazole derivatives were synthesised and
investigated as detailed in Table 2.
It therefore appeared that the in situ oxidation strategy could be
well-tested using 31 as a prodrug starting point.
Compound 31 was dosed to the rat, and plasma samples ana-
lysed for the presence of parent compound 31, and for active circu-
lating metabolites 4 and 5. Data is shown in Table 5 and clearly
shows that 31 is rapidly metabolised in vivo, with minimal plasma
exposure of the parent compound detected.
These data showed that relatively small variations in the nature
of the substituent heterocycle could lead to significant differences
in cell-based potency. For example, inserting a nitrogen atom into
the 6-position of the benzimidazole ring in 4 gave the imidazopyri-
dine 12, some fivefold less potent than the starting compound,
Following oral administration of 31 to the rat, and taking into
account plasma protein binding, the sulfoxides 4 exhibited a free
AUC approximately 15-fold higher than parent in favour of diaste-
reoisomer 4A. However although the sulfoxides were observed at
concentrations above parent based on a metabolite clearance of

D. C. Pryde et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6400–6404
6403
Table 3
Composite SAR around lead compound 4 through combination of amide and heterocyclic substituents
O
HN
N
F
N
N
N
N
N
N
N
N
N
N
Compound gp160 (IC50, nM) [3H]-dof binding (IC50, nM)
N
F
F
O
S+
4
0.2
12
1
13
0.01
14
0.4
15
0.05
>10,000
>10,000
>10,000
>10,000
>10,000
O
5
0.2
16
0.2
17
0.01
18
0.03
19
0.3
S
O
>10,000
>10,000
>10,000
>10,000
>10,000
O
S+
20
40
NT
21
2
>10,000
22
1
8000
23
36
NT
NS
NS
O
O
S
6
40
855
24
1
4330
25
8
1180
NS
NS
O
O
N+
8
4
261
26
4
1480
27
0.1
327
28
24
663
9
5
29
0.5
30
1
N+
NS
NS
5920
>10,000
5240
NS = not synthesised.
approximately liver blood flow in the rat (70 mL/min/kg) oral
exposure indicated a bioavailability of <10%.
Table 4
Microsomal stability, permeability and selected physicochemical data for compounds
4, 5 and 12-19
It was unclear why was the exposure of the oxidised metabo-
lites was so low, given that the parent was highly permeable,
and highly susceptible to oxidation to give these metabolites. The
reasons for this observation are not fully understood but potential
explanations include gut wall metabolism of 31 to the sulfoxide(s)
Compound c log P
(log D)
HLM Clintd
min/mg)
(l
L/ Caco-2e AB/BA
RRCKf
(Â10À6 cm/
s)
(Â10À6 cm/s)
4Aa
4Bb
5
1.8 (2.3) 36
1.8 (2.3) 111
2/20
ND
<1/20
2/2
2/18
ND
ND
1
1
4
ND
ND
2
2
1
1.7 (2.8) 23
0.7 (1.7) 50
2.0 (ND) ND
2.0 (2.7) 93
2.4 (2.8) 89
0.6 (1.9) <8
1.9 (2.9) 10
1.9 (3.0) 14
2.2 (3.7) 25
12c
13c
14c
15c
16
O
O
S
17
gp160 IC50 0.01 nM
log D 2.9
ND
Caco-2 AB/BA 5/31 x10-6 cm/s
O
17
18
19
5/31
3/21
ND
2
2
3
HLM 10 µL/min/mg
RRCK 2 x10-6 cm/s
hERG IC50 >10 µM
HN
N
N
N
a
b
c
Rat PK (1mg/kg IV, 2mg/kg PO)
Estimated Cl ~85 mL/min/kg
Rat absorption
Diasteroisomer 1, unassigned absolute configuration.
Diasteroisomer 2, unassigned absolute configuration.
Mixture of diasteroisomers.
~3 %
F
d
Human liver microsomal stability (0.8 mg/mL protein,
1 lM substrate
concentration).
F
e
Flux (5
lM, pH7.4).
f
Permeability as measured in a Pfizer specific cell line (2
lM, pH 7.4).
Figure 3. Sulfone analogue 17 of the lead compound 4.

6404
D. C. Pryde et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6400–6404
O
S+
O
S
O
S+
O
S
O
HN
O
HN
[O]
O
[O]
O
HN
HN
N
N
N
N
N
N
N
N
N
F
4A + 4B
F
31
F
5
gp160 IC50 0.2 nM
log D 3.7
hERG IC50 740 nM
RRCK 29 x10-6 cm/s
RLM >510 µL/min/mg
HLM >320 µL/min/mg
Figure 4. Oxidation strategy starting with the sulfide 31.
metabolites, and prompted the project to investigate alternative
options.11–13
Table 5
Rat plasma pharmacokinetics of 31
Compound
31
References and notes
Dose (route)
1 mg/kg (iv)
3 mg/kg (po)
31
4A
4B
5
1. http://www.unaids.org/en/KnowledgeCentre/HIVData/GlobalReport/2008/.
2. See for example: Broder, S. Antiviral Res. 2010, 85, 1.
3. Carr, A. Nat. Rev. Drug Disc. 2003, 2, 624.
Clearance (mL/min/kg)
Clu (mL/min/kg)
Vdss (L/kg)
60
—
—
—
—
—
—
—
11
1.7
—
—
—
—
—
—
—
2
—
—
—
—
—
—
—
<1
<0.1
2431
2.0
81
0.8
—
—
270
6.7
—
—
—
—
1%
9%
10
0.25
4. (a) Huang, Y.; Paxton, W. A.; Wolinsky, S. M.; Neumann, A. U.; Zhang, L.; He, T.;
Kang, S.; Ceradini, D.; Jin, Z.; Yazdanbakhsh, K.; Kunstman, K.; Erickson, D.;
Dragon, E.; Landau, N. R.; Phair, J.; Ho, D. D.; Koup, R. A. Nat. Med. 1996, 2, 1240;
(b) Samson, M.; Libert, F.; Doranz, B. J.; Rucker, J.; Liesnard, C.; Farber, C. M.;
Saragosti, S.; Lapoumeroulie, C.; Cognaux, J.; Forceille, C.; Muyldermans, G.;
Verhofstede, C.; Burtonboy, G.; Georges, M.; Imai, T.; Rana, S.; Yi, Y.; Smyth, R.
J.; Collman, R. G.; Doms, R. W.; Vassart, G.; Parmentier, M. Nature 1996, 382,
722.
Vdu (L/kg)
Terminal T1/2 (h)
Bioavailability
Absorption
AUC (ng h/mL)
Free AUC (ng h/mL)
0.3
5. Olsen, W. C. J. Virol. 1999, 73, 4145.
6. Wood, A.; Armour, D. Prog. Med. Chem. 2005, 43, 239.
7. (a) Dorr, P.; Westby, M.; Dobbs, S.; Griffin, P.; Irvine, B.; Macartney, M.; Mori, J.;
Rickett, G.; Smith-Burchnell, C.; Napier, C.; Webster, R.; Armour, D.; Price, D.;
Stammen, B.; Wood, A.; Perros, M. Antimicrob. Agents Chemother. 2005, 49,
4721; (b) Armour, D.; de Groot, M. J.; Edwards, M.; Perros, M.; Price, D. A.;
Stammen, B. L.; Wood, A. ChemMedChem 2006, 1, 706; (c) Price, D. A.; Armour,
D.; de Groot, M.; Leishman, D.; Napier, C.; Perros, M.; Stammen, B. L.; Wood, A.
Bioorg. Med. Chem. Lett. 2006, 16, 4633.
8. Mansfield, R.; Able, S.; Griffin, P.; Irvine, B.; James, I.; Macartney, M.; Miller, K.;
Mills, J.; Napier, C.; Navratilova, I.; Perros, M.; Rickett, G.; Root, H.; van der Ryst,
E.; Westby, M.; Dorr, P. Methods Enzymol. 2009, 460, 17.
9. (a) Armour, D. R.; Price, D. A.; Stammen, B. L. C.; Wood, A.; Perros, M.; Edwards,
M. P. Eur. Pat. Appl., EP1013276, 2000.; (b) Huang, X.; O’Brien, E.; Thai, F.;
Cooper, G. Org. Process Res. Dev. 2010, 14, 592.
10. Armour, D. R.; Price, D. A.; Stammen, B. L. C.; Wood, A.; Perros, M.; Edwards, M.
P., PCT Int. Appl., WO2000038680, 2000.
11. Barber, C. G.; Blakemore, D. C.; Chiva, J.-Y.; Eastwood, R. L.; Middleton, D. S.;
Paradowski, K. A. Bioorg. Med. Chem. Lett. 2009, 19, 1075.
12. Barber, C. G.; Blakemore, D. C.; Chiva, J.-Y.; Eastwood, R. L.; Middleton, D. S.;
Paradowski, K. A. Bioorg. Med. Chem. Lett. 2009, 19, 1499.
13. Pryde, D. C.; Corless, M.; Fenwick, D. R.; Mason, H. J.; Stammen, B. L. C.;
Stephenson, P. T.; Ellis, D.; Bachelor, D.; Gordon, D.; Barber, C. G.; Wood, A.;
Middleton, D. S.; Blakemore, D. C.; Parsons, G. C.; Eastwood, R.; Platts, M. Y.;
Statham, K.; Paradowski, K. A.; Burt, C.; Klute, W. Bioorg. Med. Chem. Lett. 2009,
19, 1084.
followed by subsequent metabolism and/or uptake by the liver for
excretion. The precise fate of 31 was not explored further. Overall
the data shown above was sufficient to indicate the use of 31 as a
potential oxidisable precursor to be non-viable due to the limited
formation and systemic exposure of the active metabolite.
Through the course of this work, we were intrigued by the pos-
sibility that S- and N-oxide derivatives could offer an excellent
compromise between on the one hand high cell-based activity
and a wide hERG selectivity window and acceptable oral absorp-
tion on the other. We were able to identify several compounds
which were likely to come close to being optimal within the series,
but confirmed that oral bioavailability in the rat for two com-
pounds was low, as a result of very low absorption. We then inves-
tigated a prodrug strategy in which we sought to dose a highly
permeable, metabolically labile sulfide and monitor for active oxi-
dised metabolites in the systemic circulation. This approach was
also compromised by low exposure of both parent and active