Potent blockers of the monocarboxylate transporter MCT1: Novel immunomodulatory compounds

Article abstract of DOI:10.1016/j.bmcl.2006.01.024

A novel series of potent blockers of the monocarboxylate transporter, MCT1, is disclosed. From very potent but lipophilic lead compounds, systematic changes to all parts of the molecule, targeting reduction in log D, afforded compounds with significantly

Full text of DOI:10.1016/j.bmcl.2006.01.024

Bioorganic & Medicinal Chemistry Letters 16 (2006) 2260–2265
Potent blockers of the monocarboxylate transporter MCT1:
Novel immunomodulatory compounds
S. D. Guile,a,* J. R. Bantick,a D. R. Cheshire,a M. E. Cooper,a,ꢀ A. M. Davis,b
D. K. Donald,a R. Evans,a C. Eyssade,a D. D. Ferguson,b S. Hill,a R. Hutchinson,c
A. H. Ingall,a L. P. Kingston,a I. Martin,b,ꢁ B. P. Martin,a R. T. Mohammed,a C. Murray,c
M. W. D. Perry,a R. H. Reynolds,a P. V. Thorne,a D. J. Wilkinsona and J. Withnalla
aDepartment of Medicinal Chemistry, AstraZeneca R&D Charnwood, Bakewell Road, Loughborough, LE11 5RH, UK
bDepartment of Physical & Metabolic Sciences, AstraZeneca R&D Charnwood, Bakewell Road, Loughborough, LE11 5RH, UK
cDepartment of Discovery BioScience, AstraZeneca R&D Charnwood, Bakewell Road, Loughborough, LE11 5RH, UK
Received 9 December 2005; revised 5 January 2006; accepted 6 January 2006
Available online 7 February 2006
Abstract—A novel series of potent blockers of the monocarboxylate transporter, MCT1, is disclosed. From very potent but
lipophilic lead compounds, systematic changes to all parts of the molecule, targeting reduction in logD, afforded compounds
with significantly improved overall properties. These compounds show potent immunomodulatory activity.
Ó 2006 Elsevier Ltd. All rights reserved.
An important component of the immune response is
activation of T-cells following antigen challenge. How-
ever, undesirable activation can lead to graft rejection
following transplantation and to autoimmune diseases
such as rheumatoid arthritis.
Intrigued by these results, we chose to investigate this
area further. The synthesis and properties of a novel
series of compounds derived from 1 are presented below.
Importantly, we demonstrated that in primary T-lym-
phocytes these compounds do not exert their activity
through effects on IL-2 regulation. Moreover, through
compound-led target identification4 we were able to
show, by photoaffinity labelling and proteomic charac-
terisation, that the actual molecular target of these com-
pounds is the monocarboxylate transporter, MCT1.
This was supported by a strong correlation between
binding at MCT1 and in vitro immunomodulatory
activity in an assay measuring inhibition of PMA/
ionomycin-induced human PBMC proliferation.5
Upon activation, T-cells produce the autocrine growth
factor, IL-2, whose regulation relies on the nuclear
factor of activated T-cells (NFAT-1).1 TacrolimusTM
(FK-506) and cyclosporin, when bound to their immu-
nophilin (FKBP and cyclophilin, respectively), reduce
IL-2 production by inhibiting calcineurin, which in turn
reduces NFAT-1 translocation to the nucleus.2 While
both compounds are effective clinical agents, renal and
other toxicities limit their widespread use. Researchers
at Sterling Winthrop Pharmaceuticals pursued an
alternative, calcineurin independent, approach to IL-2
inhibition via NFAT-1 and identified a number of inter-
esting compounds including the pyrrolopyrimidinedione
1.3
The monocarboxylate transporters are a family of
proteins which transport lactate and other small mono-
carboxylates. We have shown that MCT1 expression is
rapidly upregulated upon T-lymphocyte activation in
order to meet the demand for lactate efflux resulting
from an increased glycolytic rate. Inhibition of lactate
efflux by potent blockade of lactate transport results in
the accumulation of lactate within the cell and feedback
inhibition of glycolysis. This suppression of cellular
metabolism results in the inability of T-lymphocytes to
sustain the rapid rate of cell division occurring during
the early immune response to antigen recognition,
without being cytotoxic. Blockade of MCT1 is thus a
Keywords: Monocarboxylate; Transporter; Immunomodulatory.
*
Corresponding author. Tel.: +44 1509 645308; fax: +44 1509
645512; e-mail: simon.guile@astrazeneca.com
Present address: 7TM Pharma A/S, Fremtidsvej 3, 2970 Horsholm,
Denmark.
ꢀ
ꢁ
Present address: Organon Laboratories Ltd, Newhouse, Lanark-
shire, ML1 5SH, UK.
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.01.024

S.D. Guile et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2260–2265
2261
novel mechanism of immunosuppression distinct from
current therapies.
which, following halogen–metal exchange, could be
reacted with a range of electrophiles. Reaction with sul-
fur provided a route to functionalised thioethers
(Scheme 2) while quenching with sulfur dioxide or car-
bon dioxide allowed the synthesis of sulfonamides and
amides (Scheme 3).
Initial chemical effort focussed on synthesising com-
pounds with changes to the heterocyclic scaffold.
Amongst numerous new cores investigated, those con-
tained in compounds 2 and 3, prepared as previously de-
scribed, were most interesting (Fig. 1).5 The new
compounds displayed excellent affinity for MCT1. How-
ever, these compounds were too lipophilic and, as a con-
sequence, suffered from poor aqueous solubility and
high in vitro metabolic clearance in human hepatocytes
as well as potent CYP2C9 inhibition (Table 1).
The homologous thioethers 10–14 showed significantly
improved in vitro metabolic stability over 4 and had
comparable affinities for MCT1, albeit less than that
of the starting compound. Reintroducing hydroxyl
groups on the cyclopentyl thioether gave a range of
compounds 15–18 which generally displayed good affin-
ity but also showed high intrinsic clearance in human
hepatocytes. The dihydroxypropyl derivative 19 was sig-
nificantly less active than the monohydroxypropyl lead 4
and showed no improvement in hepatic clearance
(Table 2).
In modern drug discovery programmes, it is well
recognised that compound optimisation often results
in an increase in both molecular complexity and lipo-
philicity, driven by the need to increase potency,
thereby defining the required lead properties and
resulting in the concept of ‘leadlikeness’.6,7 Given the
high starting potency in this case, our strategy, con-
versely, was to substantially reduce lipophilicity while
accepting some reduction in activity in order to im-
prove overall properties.
These results indicated that, although the hydroxy
group interacted favourably with the transporter giving
high activity, it was detrimental to metabolic stability. In
agreement with this, examination of the metabolite pro-
file following incubation of compound 3 in human hepa-
tocytes indicated that the primary route of metabolism
was glucuronidation on the hydroxyl group.
To explore 5- and 6-substituent changes (see Fig. 1 for
numbering), the pyrimidinothiophene ring system was
used as this facilitated introduction of diversity at both
positions. Initially 2-trifluoromethylphenylmethyl was
chosen as the 6-position (e.g., 4), as this gave similar
properties to naphthylmethyl (e.g., 3). The 5,6-unsubsti-
tuted pyrimidinothiophene 55 was selectively brominat-
ed at the 6-position. Following deprotonation with
LDA at the 5-position, a bromide migration occurred.8
The resulting anion 6 was reacted with 2-trifluorometh-
ylbenzaldehyde to provide the 5-bromo intermediate 7.
Oxidation of this compound gave the keto compound
8 which was now activated towards nucleophilic attack
at the 5-position. Following addition of the desired thiol
the keto group could be removed via a standard two-
step process (Scheme 1). Alternatively, deoxygenation
of bromothiophene 7 gave the key intermediate 9 having
the desired 6-substituent and a functionalised 5-position
Although reducing lipophilicity was an ultimate goal of
this work, in this initial phase it can be seen that
measured logD values remained high for all compounds,
within a range much narrower than calculated, and
displaying variations which did not follow the expected
trend (see measured vs calculated values, Table 2). For
example, addition of a second hydroxyl group in going
from 4 to 19 led to an increase in logD where a reduc-
tion would be expected. Similarly, in the non-hydroxyl-
ated compounds, sequentially increasing the number of
carbons from ethyl 10 to cyclohexyl 14 reduced logD
where a significant increase would be anticipated. It is
likely that this inconsistency is due to ‘hydrophobic
collapse,’ that is, lipophilic 5- and 6-substituents interact
intramolecularly leading to less than expected bulk
lipophilicity.9
O
O
O
S
OH
S
OH
S
OH
5
N
N
N
N
3
N
N
6
S
O
N
O
N
1
iBu
iBu
iBu
1
2
3
Figure 1. Structures of compounds 1–3.
Table 1. Comparison of heterocyclic scaffolds
Compound
MCT1 binding Kia
(nM)
Human Heps Clint
(lL/min/106cells)
CYP2C9 IC50
(lM)
Sol (mg/mL)
logD
1
2
3
0.33
0.10
0.28
nd
36
58
nd
0.08
<0.001
<0.001
4.4
4.8
4.8
2.60
0.50
nd, not determined.
a FB binding (Ref. 5).

2262
S.D. Guile et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2260–2265
O
O
O
Br
-
N
N
N
a,b
Br
Br
-
S
S
O
N
O
N
S
O
N
iBu
iBu
5
6
iBu
O
O
Br
OH
O
N
N
c
d
CF3
CF3
S
S
O
N
O
N
8
iBu
7
iBu
O
R
N
e,f,g
CF3
S
O
N
iBu
S
OH
S
*
S
S
S
*
*
*
*
4
10
11
13
14
Scheme 1. Reagents and condition: (a) Br2, DCM, 98%; (b) LDA, THF, ꢀ78 °C; (c) 2-CF3C6H4CHO, 71% (two steps); (d) tetrapropylammonium
˚
perruthenate, NMO, 4 A molecular sieves, DCM, 97%; (e) RSH, K2CO3, DMF, 65–97%; (f) NaBH4, MeOH, 77–90%; (g) TFA, Et3SiH, 48–93%.
O
O
Br
R
N
N
a
b,c,d
CF3
CF3
7
S
S
O
N
O
N
iBu
iBu
9
HO
OH
OH
S
S
S
S
*
*
*
*
12
(rac)-15
(rac)-16
(rac)-17
OH
OH
S
OH
S
*
SO2
OH
*
OH
*
18
(rac)-19
(rac)-20
Scheme 2. Reagents and condition: (a) TFA, Et3SiH, 83%; (b) i-PrMgCl, THF, 0 °C; (c) S8; (d) for 12: bromocyclobutane, NaBH4, NaHCO3, H2O,
THF, 36% (three steps); for (rac)-15: cyclopentane epoxide, NaBH4, NaHCO3, H2O, THF, 36% (three steps); for (rac)-16 and (rac)-17: i—2-
cyclopentenone, dil HCl(aq), THF; ii—NaBH4, MeOH, 4% and 16% (four steps); for 18: i—3-tosyloxycyclopentene, NaHCO3, H2O, THF; ii—OsO4,
NMO, THF, H2O, 7% (four steps); for (rac)-19: glycidol, NaBH4, NaHCO3, H2O, THF, 23% (three steps); for (rac)-20: mCPBA, DCM (on (rac)-17),
10% (four steps).
In an effort to significantly reduce measured lipophilici-
ty, changes to the thioether link were made, exemplified
here for the hydroxycyclopentyl series which ultimately
provided one of the most favoured 5-substituents.
Replacing thioether 17 with sulfone 20 gave a drop in
activity and very high CYP2C9 inhibition. Replacing
the cyclopentane with the equivalent pyrrolidine to give
the sulfonamide 21 improved activity, however,
CYP2C9 inhibition was again high. The analogous
amide 22 had the lowest logD in this set, displayed good
affinity and also showed improved human hepatocyte
stability and lower CYP2C9 inhibition relative to the
sulfone and sulfonamide analogues (Table 3).
More amide analogues were also evaluated (Table 4). In
this series, the non-hydroxylated 4, 5 and 6-membered

S.D. Guile et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2260–2265
OH
2263
OH
O
O
N
R
N
SO2
O
N
N
*
a,b
*
CF3
(rac)-21
22
23
*
8
S
OH
O
N
iBu
O
N
O
N
O
N
O
N
*
*
*
*
24
25
26
27
Scheme 3. Reagents and conditions: (a) i-PrMgCl, THF, 0 °C; (b) for 21: i—SO2; ii—NCS, HCl(aq), DCM; iii—(R)-3-hydroxypyrrolidine,
triethylamine, DCM, 70% (four steps); for 22–27: i—CO2, 0 °C, 75% (two steps); ii—oxalyl chloride, cat. DMF, DCM; iii—RR0NHÆHCl,
triethylamine, DCM, 22–67% (two steps).
Table 2. Comparison of 5-substituents
similar to that of the non-hydroxy derivative 23 but
was more metabolically stable, perhaps due to the alco-
hol being tertiary.
Compound MCT1
binding
Human
Heps Clint
logD Calculated
logDb
Kia (nM) (lL/min/106 cells)
Alternative groups at both the 1- and 3-positions were
also examined but no substituents significantly better
than the original methyl and isobutyl were found (data
not shown).
4
10
11
12
13
14
15
16
17
18
19
0.35
6.0
43
nd
2
4.1
4.5
4.4
5.7
6.6
6.9
2.2
13
3
>3.5 7.0
3.2
6.0
4
4.2
4.1
4.2
nd
7.5
8.1
5.8
5.7
5.7
4.7
4
Attention was then turned to the 6-substituent. In order
to prepare a range of 6-heterocyclic 5-amides, predicted
to have lower logD, the synthesis was modified to allow
incorporation of the 6-substituent at a later stage via key
intermediates 28 or 29 (Schemes 4 and 5).5
2.7
11
26
31
82
59
0.29
0.42
0.68
6.5
4.4
4.4
>4.5 4.7
nd, not determined.
a FB binding (Ref 5).
b Ref. 10.
While a range of 6-heteroaryl compounds were prepared
with various different 5-amides, results are again
presented here for the (R)-5-hydroxypyrrolidine amide
series derived from 22 as this was found to be amongst
the best 5-amides and the trends described are represen-
tative of those found with other amides (Table 5). The
quinolines 30 and 31 gave good potency and stability
along with an appreciably lower logD relative to the tri-
fluoromethylphenyl derivative 22, 30 being the better
isomer with respect to CYP2C9 inhibition. 4-Isoquino-
line 32 showed very potent CYP3A4 inhibition. The
indole 33 had properties similar to those of 22 but suffered
from high inhibition of both CYP2C9 and CYP3A4. The
azaindole 34 was more stable, had much reduced CYP
inhibition and had a lower logD than the indole 33 but
was also less active. In the imidazole series, the 2-methyla-
mino benzimidazole 35 gave the best balance of proper-
ties with good potency, very good stability, low CYP
inhibition and low logD. The methyl benzimidazole 36
and methyl imidazole 37 were both less active. It is note-
worthy that the combination of an amide at the 5-posi-
tion and a heteroaryl group at the 6-position was
required in order to reduce the logD to the desired range.
Table 3. Comparison of 5-linking group
Compound MCT1
binding
Human
Heps Clint
CYP2C9
IC50 (lM)
logD
Kia (nM) (lL/min/106 cells)
17
20
21
22
0.42
12
31
31
82
17
0.2
0.2
0.2
1.7
4.4
3.2
3.3
2.7
1.1
4.9
a FB binding (Ref. 5).
Table 4. Comparison of 5-amides
Compound
MCT1
binding
Kia(nM)
Human
Heps Clint
(lL/min/106 cells)
logD
22
23
24
25
26
27
4.9
9.2
5.5
8.7
37
17
18
27
nd
11
7
1.7
3.5
3.8
4.1
2.9
3.2
3.5
In summary, novel MCT1 blockers have been discov-
ered which show potent inhibition of PMA/ionomycin-
induced human PBMC proliferation. By reducing logD,
the drug-like properties of the series were significantly
improved with respect to solubility, plasma protein
binding (ppb), human metabolic stability and CYP2C9
inhibition (cf. 3 and 30, Table 6). Good potency was
nd, not determined.
a FB binding (Ref. 5).
rings (23, 24 and 25) showed similar properties, while
the smaller dimethyl amide 26 was less active. In this
case, the logD trend was in the expected direction.
The hydroxy-substituted azetidine 27 showed potency

2264
S.D. Guile et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2260–2265
OH
O
CO2Et
O
O
CO2Et
O
N
N
a,b,c
d,e
N
N
S
Ar
O
N
S
O
N
S
Ar
O
N
iBu
28
iBu
iBu
*
*
*
N
N
N
30
31
32
Scheme 4. Reagents and condition: (a) ArCHO, LDA, THF, ꢀ78 °C, 56–84%; (b) Ac2O, TEA, DMAP, DCM or MeSO2Cl, DMAP, TEA, THF; (c)
H2, Pd/C, NaHCO3, H2O, MeOH, 28–78% (two steps); (d) NaOH, H2O, THF, MeOH, 79–88%; (e) RR0NHÆHCl, EDCI, HOBt, DCM or PyBrOP,
DIPEA, THF, 62–97%.
OH
O
O
CO2Me
O
CO2Me
O
N
N
a
b,c
N
N
S
Ar
S
O
N
Br
O
N
S
Ar
O
N
iBu
iBu
29
iBu
*
*
*
*
*
N
N
N
Cl
MeNH
N
N
N
N
N
H
H
N
Cl
33
34
35
36
37
Scheme 5. Reagents: (a) indole, NaHCO3, H2O, CHCl3, 63% or azaindole, NaH, ZnCl2, toluene, 55% or ‘imidazole,’ NaH, DMF, 11–82%; (b)
NaOH, H2O, THF, MeOH; (c) RR0NHÆHCl, EDCI, HOBt, DCM or PyBrOP, DIPEA, THF, 7–58%.
Table 5. 6-Substituent changes
Compound
MCT1 binding Kia (nM)
Human Heps Clint
(lL/min/106 cells)
CYP2C9 IC50
(lM)
CYP3A4 IC50 (lM)
logD
22
30
31
32
33
34
35
36
37
1.9
4.8
5.3
3.9
0.79
4.7
6.6
16
17
6
1.7
16
nd
16
2.7
1.7
2.0
1.7
2.4
1.7
1.6
1.3
1.6
<3
nd
6
6.0
13
20
<0.07
1.0
6.7
1.4
35
<3
<3
<3
<3
11
>50
>50
>50
2.5
8.7
26
nd, not determined.
a SPA binding (Ref. 11).
Table 6. Summary of progress
Compound
MCT1 binding Kia
(nM)
Human Heps Clint
(lL/min/106 cells)
CYP2C9 IC50
(lM)
Sol (mg/mL)
logD
ppb (% free)
3
30
0.08
4.8
58
6
0.50
16
<0.001
0.52
4.8
1.7
0.1
36
a SPA binding (Ref. 11).

S.D. Guile et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2260–2265
2265
9. Rich, D. H.; Homes, O. W. In Practice of Medicinal
Chemistry; Wermuth, C. G., Ed.; Elsevier: London, 2003;
pp 373–386, and references therein.
10. ACDlogD is available from ACD Labs, Toronto, Canada,
<http://www.acdlabs.com>.
achieved due to the very high potency of the lead com-
pound. Unfortunately, many of the amides described
here, including 30, exist as a mixture of conformational
isomers resulting from restricted rotation about both the
amide carbon–nitrogen and the carbonyl–aryl bonds.
Further studies with these compounds, including atrop-
isomeric properties and implications, will be reported in
due course.
11. A radioligand-binding assay was developed using scintil-
lation proximity assay (SPA) technology (Bosworth, N.;
Towers, P. Nature 1989 341, 167) Wheatgerm agglutinin
SPA beads (Amersham) (0.2 mg/mL) coated with Jurkat
cell membranes were incubated with [125I]-38 (0.1 nM) in
the absence or presence of increasing concentrations of
test compound. Unlabelled 3 (1 lM) was used to define
non-specific binding. The test compounds were diluted in
DMSO and then assay buffer (50 mM HEPES, 0.1 mM
EDTA and 150 mM NaCl, pH 7.4) containing 0.5%
bovine serum albumin (BSA) such that the final concen-
tration of DMSO in the assay was 0.5%. The final
concentration of BSA in the total assay volume of 200 lL
was 0.05%. The assay reagents were incubated for 18 h at
room temperature and radioactivity bound to the mem-
brane-coated beads was determined using a Topcount
scintillation counter (Perkin Elmer).
References and notes
1. Shaw, J. P.; Utz, P. J.; Durand, D. B.; Toole, J. J.; Emmel,
E. A.; Crabtree, G. R. Science 1988, 241, 202.
2. Liu, J.; Farmer, J. D. J.; Lane, W. S.; Friedman, J.;
Weissman, I.; Schreiber, S. L. Cell 1991, 66, 807.
3. Michne, W. F.; Schroeder, J. D.; Guiles, J. W.; Trea-
surywala, A. M.; Weigelt, C. A.; Stansberry, M. F.;
McAvoy, E.; Shah, C. R.; Baine, Y.; Sawutz, D. G.;
Miller, P. B.; Stankunas, B. M.; Reid, J.; Bump, E.;
Schlegel, D. J. Med. Chem. 1995, 38, 2557.
4. Schreiber, S. L. Science 2000, 287, 1964.
The radioligand 125I-38 was prepared from the heterocycle
39 (Bantick, J.; Cooper, M.; Perry, M.; Thorne, P.
WO9929695) as shown below.
5. Murray, C. M.; Raymond Hutchinson, R.; Bantick, J.
R.; Belfield, G. P.; Benjamin, A. D.; Brazma, D.;
Bundick, R. V.; Cook, D. I.; Craggs, R. I.; Edwards, S.;
Evans, L. R.; Harrison, R.; Holness, E.; Jackson, A. P.;
Jackson, C. G.; Kingston, L. P.; Perry, M. W. D.; Ross,
A. R. J.; Rugman, P. A.; Sidhu, S. S.; Sullivan, M.;
Taylor-Fishwick, D. A.; Walker, P. C.; Whitehead, Y.
M.; Wilkinson, D. J.; Wright, A.; Donald, D. K. Nat.
Chem. Biol. 2005, 1, 371.
O
O
S
OH
125I
N
N
N
N
a,b,c,d
N
NH
6. Teague, S. J.; Davis, A. M.; Leeson, P. D.; Oprea, T.
Angew. Chem., Int. Ed. 1999, 38, 3743.
iBu
125I-38
iBu
39
7. Oprea, T. I.; Davis, A. M.; Teague, S. J.; Leeson, P. D.
J. Chem. Inf. Comput. Sci. 2001, 41, 1308.
8. Kano, S.; Yuasa, Y.; Yokomatsu, T.; Shibuya, S. Hetero-
cycles 1983, 20, 2035.
(a) i—2-chloromethyliodobenzene, Cs2CO3, DMF; ii—
LDA, TsS(CH2)3OTBDMS, THF; (b) (CH3)6Sn2,
Pd(PPh3)4, toluene; (c) chloramine-T, Na125I, methanol;
(d) TBAF, THF.