Discovery of 1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid diamides that increase CFTR mediated chloride transport

Article abstract of DOI:10.1016/j.bmcl.2005.02.041

A series of 1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid diamides that increase chloride transport in cells expressing mutant cystic fibrosis transmembrane conductance regulator (CFTR) protein has been identified from our compound library. Analoging efforts and the resulting structure-activity relationships uncovered are detailed. Compound potency was improved over 30-fold from the original lead, yielding several analogs with EC₅₀ values below 10 nM in our cellular chloride transport assay.

Full text of DOI:10.1016/j.bmcl.2005.02.041

Bioorganic & Medicinal Chemistry Letters 15 (2005) 2087–2091
Discovery of 1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid
diamides that increase CFTR mediated chloride transport
Bradford H. Hirth,* Shuang Qiao, Lisa M. Cuff, Brian M. Cochran, Marko J. Pregel,
Jill S. Gregory, Scott F. Sneddon and John L. Kane, Jr.
Genzyme Drug Discovery and Development, Genzyme Corp., One Kendall Sq., Cambridge, MA 02139, USA
Received 4 January 2005; revised 10 February 2005; accepted 14 February 2005
Available online 16 March 2005
Abstract—A series of 1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid diamides that increase chloride transport in cells expressing
mutant cystic fibrosis transmembrane conductance regulator (CFTR) protein has been identified from our compound library. Ana-
loging efforts and the resulting structure–activity relationships uncovered are detailed. Compound potency was improved over 30-
fold from the original lead, yielding several analogs with EC50 values below 10 nM in our cellular chloride transport assay.
Ó 2005 Elsevier Ltd. All rights reserved.
Cystic fibrosis (CF) is the most common lethal genetic
disease among the Caucasian population, affecting more
than 30,000 people in the US alone.1 The underlying
cause of the disease is a mutation in the gene encoding
the cystic fibrosis transmembrane conductance regulator
(CFTR) protein.2 This cAMP-gated chloride channel is
expressed in the apical membrane of epithelial cells in
the airways, pancreas, intestines, and testes.3 The most
common CFTR mutation in CF is the deletion of phen-
ylalanine 508 (DF508).4 DF508 mutant protein fails to
fold properly and is largely recognized and destroyed
by cellular quality control machinery. A small amount
of mutant CFTR reaches the apical membrane and re-
mains capable of ion transport, although the efficiency
of this process is reduced significantly.5 This ultimately
results in a variety of complications due to decreased
water transport, most notably, chronic infection and
inflammation of the airways leading to progressive dete-
rioration of lung function, the primary cause of mortal-
ity in CFpatients.
and traffick to the cell surface (e.g., via chemical chaper-
ones) or by increasing the activity of the channel once in
place at the membrane. The former class of bioactives
has been termed CFTR correctors and the latter, CFTR
activators or openers.6 Agents operating via either
mechanism can be detected using one of several cell-
based fluorescence assays measuring chloride efflux. To
date, several classes of small molecule compounds have
been identified as CFTR correctors or activators, includ-
ing flavonoids,7 xanthines,8 benzoquinoliziniums,9 benz-
imidazolones,10 and isoxazoles.11 This paper describes
our discovery and subsequent medicinal chemistry
investigation of a novel class of CFTR modulators,
1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid (Tic)
diamides. While the CFTR activity of these compounds
is apparent even at low concentrations (<10 nM), their
mechanism of action has yet to be elucidated.
Our CFdrug discovery program was powered by a high-
throughput screen, which utilized C127 murine mam-
mary epithelial cells stably transfected with recombinant
DF508 CFTR.12 Chloride efflux was measured by moni-
toring the fluorescence of a halide sensitive dye, MQAE
(N-(ethoxycarbonylmethyl)-6-methoxyquinolinium bro-
mide), and normalized by scaling between the relative
responses of untreated cells (0%) and cells maintained
at low temperature (28–30 °C; 100%), a condition which
is known to increase the level of functional DF508
CFTR.13 Analogs of our lead compounds were tested
at multiple concentrations to generate dose–response
One potential therapeutic approach to CFis to identify
agents which can enhance the limited chloride transport
capability of mutant CFTR. This may be achieved by
inducing more of the defective protein to fold properly
Keywords: CFTR; Cystic fibrosis.
*
Corresponding author. Fax: +1 617 252 7550; e-mail:
bradford.hirth@genzyme.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.02.041

2088
B. H. Hirth et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2087–2091
curves and associated EC50 values. The peak level of
activity from this curve is referred to here as a Ômaxi-
mum responseÕ percentage (MR).
Crude analog arrays were generated utilizing standard
EDC (1-[3-(dimethylamino)propyl]-3-ethylcarbodiimide)
amidation chemistry.14 Intermediate 2 was coupled
with a collection of 240 commercially available carbox-
ylic acids, including alkyl, aryl, and arylalkyl types
with substituents of varying size and polarity. Likewise,
intermediate 3 was coupled to a diverse set of 550
primary and secondary amines. After confirming that
neither starting material was active in the CFTR
assay, analog data was examined to reveal several strong
SAR trends. It was found that only a select group of
benzoic acids coupled with amine 2 to yield diamides
of significant activity, specifically those featuring moder-
ately polar to non-polar substituents in 3-position. The
most potent of these was the original acid component
of 1, 3-phenoxybenzoic acid. Lesser activity was ob-
served with a small number of 3,4-disubstituted benzoic
acids. Alkanoic acids, substituted or otherwise, did not
yield products of significant activity. In the case of the
analogs prepared from acid 3, activity spiked sharply
around a certain set of arylamine inputs. While alkyl
and secondary amines produced only a few weakly ac-
tive diamides, anilines bearing non-polar to moderately
polar substituents in the 4-position gave robust activity.
In particular, several 4-alkylanilines afforded potent
analogs, though activity seemed to be concentrated in
the straight chain variants. Anilines bearing other
modestly polar 4-substituents (halides, alkoxides, esters,
N-alkyl-N-acylamino, etc.) returned products with
varying levels of notable activity. This gross survey of
structural preferences seemed to indicate that SAR in
the Tic diamide series is fairly sharp with activity tightly
clustered around the structure of 1. In light of this find-
ing, subsequent analoging efforts focused on com-
pounds, which mirrored the positional and electronic
substitution patterns of this lead compound.
Screening of our in-house compound library, which was
prepared and tested as mixtures (generally 100 theoreti-
cal compounds per sample), yielded several hits. Individ-
ual active compounds were identified via a stepwise
deconvolution/retesting process and confirmed by deter-
mining EC50 values for the final purified and character-
ized compounds. This effort yielded several interesting
actives including the 1,2,3,4-tetrahydroisoquinoline-3-
carboxylic acid (Tic) diamide, 1.
Cl
HN
O
N
O
OPh
1
EC50: 294 nM
Max. Resp. 72%
Analog preparation around this initial lead was carried
out in two stages. First, a large array of single, unpuri-
fied Tic diamides was prepared in one step (amide cou-
pling) from one of two purified Tic precursors (Scheme
1). While the products were not subjected to purity anal-
ysis, pilot reactions for this simple conversion indicated
yields were generally robust. The objective of this exer-
cise was to quickly survey the structural space around
compound 1 and identify, if evident, general SAR
trends. This information was used to guide our second,
more deliberate, phase of analoging, where selected
target molecules were purified and characterized prior
to biological testing. The structures of the crude array
analogs are not detailed here, but discussed generally
with regards to SAR findings (vide infra). The structures
of the purified analogs are shown in the tables that
follow.
The purified Tic diamides prepared featured changes to
both sidechains of compound 1.15 Many of the requisite
amines and carboxylic acids were commercially avail-
able; the remainder were prepared by simple, known
methods (syntheses not shown). Initial analogs were pre-
pared as racemates, but, as efforts continued, single
enantiomers were examined. Activity was found to re-
side exclusively in the (S)-isomer and later analogs were
prepared only in this form.
Structures and data for aniline modified compounds are
shown in Table 1. Halogen substituents, as demonstrated
by compounds 4–7, are generally acceptable in the 4-po-
sition. The favorability of straight chain alkyl groups is
confirmed by analogs 8–12. Interestingly, there appears
to be a discernible preference in chain length, with activ-
ity peaking at five carbons and falling off in either direc-
tion. Branching, even four carbons out from the
aromatic ring, continues to have a detrimental effect on
activity as exemplified by compound 13. Alkoxy groups
are also acceptable in the 4-position (e.g., 14, 15),
although activity is reduced relative to alkyl variants. Po-
lar amino groups, as anticipated, are not tolerated (e.g.,
16). Although hinted at in our crude arrays, the activity
of N-alkyl-acetanilide compounds 17 and 18 and the pyr-
idine variants 19–21 is pleasantly surprising. In the end,
Cl
O
RCO2H, EDC 240 analogs
N
H
(crude)
CHCl3
NH
2
CO2H
O
RNH2 or RR'NH, EDC
CHCl3
N
550 analogs
(crude)
3
OPh
Scheme 1.

B. H. Hirth et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2087–2091
Table 2. Benzoic acid modified Tic diamides
2089
Table 1. Amine modified Tic diamides
1
1
X
R
R
HN
HN
*
N
*
N
O
O
O
O
2
R
OPh
Compd *R/S R1
R2
EC50 MRa
(nM) (%)
Compd *R/S R1
X
EC50 MRa
(nM) (%)
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
R/S
R/S
R/S
R/S
R/S
R/S
R/S
R/S
R/S
R/S
S
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
n-Pn
2-OPh
4-OPh
3-Bz
338
NA
336
291
394
167
282
590
NA
268
NA
NA
176
26
55
4
R/S
S
FCH 202
Cl
Br
92
NA
82
5
CH 119
CH 292
CH 314
CH 202
61
65
67
68
72
75
75
63
95
39
64
NA
66
71
41
66
53
6
R/S
R/S
R/S
R/S
S
3-Bn
3-SPh
67
55
7
I
8
n-Butyl
n-Pentyl
n-Pentyl
n-Hexyl
n-Heptyl
3-(4-(OMe)OPh)
3-(4-(CF3)OPh)
3-(4-(F)OPh)
3-(3,4-(Cl)2OPh)
3-O-(2-pyridyl)
H
92
69
9
CH
CH
61
23
10
11
12
13
14
15
16
17
18
19
20
21
70
R/S
R/S
R/S
R/S
R/S
R/S
R/S
R/S
R/S
R/S
S
CH 100
CH 230
CH 160
CH 420
CH 145
CH NA
CH 150
NA
74
4-Me-Pentyl
OEt
OBu
NA
NA
93
R/S
S
n-Pn 3-OH
n-Pn 3-OMe
n-Pn 3-OEt
OCH2CH2N(CH3)2
N-Acetyl-N-butyl-amino
N-Acetyl-N-pentyl-amino CH
S
73
S
n-Pn 3-O-i-Pr
n-Pn 3-O-t-Bu
n-Pn 3-O-Cyclopentyl
n-Pn 3-O-Cyclohexyl
6.3 69
58
197
119
62
S
10
15
20
78
74
71
NA
52
40
64
H
Cl
N
N
N
S
S
n-Pentyl
S
n-Pn 3-O-(N-Me-piperidin-4-yl) NA
a Maximum response relative to cold temperature control (100%).
S
n-Pn 3-Cl
n-Pn 3,4-Cl2
51
82
12
S
S
n-Pn 3-OPh-4-OMe
n-Pn 3-O-i-Pr-4-OMe
n-Pn 3-CH2 (N(Me)2)-4-OMe
S
3.5 72
S
NA
NA
NA
however, the optimal substituent for the aniline unit ap-
pears to be the 4-n-pentyl chain and all our later analogs
feature this group. This change results in a fivefold
improvement in potency versus the lead compound 1.
S
n-Pn 3-O-i-Pr-4-CH2 (N(Me)2) NA
a Maximum response relative to cold temperature control (100%).
Analogs featuring changes to the acid component are
shown in Tables 2 and 3. Although the 2-phenoxy-
benzoyl analog of 1 does exhibit reasonable activity,
compounds 22 and 23 confirm the preference for substi-
tution at the 3-position. There is considerable flexibility
in the form of the 3-substituent, however. A variety of
linker modified (e.g., 24–26) and phenoxy substituted
compounds (e.g., 27–29) return very respectable, and
sometimes improved, activity. At the same time, the di-
chloro analog 30 demonstrates that there are limits to
acceptable substitution. Cutting back the 3-substituent
to either hydrogen or hydroxyl is not allowed (e.g., 32,
33), although smaller ethers at this position do result
in good activity (i.e., the methoxy group of 34). From
this point the 3-substituent was expanded into several
larger alkoxy and cycloalkoxy variants (e.g., 35–39),
all of which exhibited impressive potency. The one
exception was compound 40, which contains an embed-
ded tertiary amine. Halogens, as exemplified by analogs
41 and 42, are also acceptable substituents, although
activity suffers somewhat relative to alkoxy analogs. An-
other key observation was that a small 4-substituent
(i.e., methoxy) can augment the potency of a 3-substitu-
ent. This finding led to our most potent analog,
compound 44, which boasts a near sevenfold increase
in activity versus the corresponding phenoxy derivative.
As a final area of investigation, we examined aza ana-
logs of selected compounds (Table 3). Compounds 47–
53 clearly demonstrate this is an acceptable
modification.
Finally, as an initial attempt to characterize the activity
of Tic diamides, a series of assay timing experiments was
carried out. It was determined that an exposure time on
cells of 17–24 h was required for maximal compound
activity. Half-maximal activity was observed after a 6–8 h
incubation period. Because agents directly targeting
membrane bound CFTR (activators) should, in general,
have a rapid onset of action, this observation seems to
be more consistent with a corrector mechanism, though
additional studies will be needed to confirm this hypoth-
esis. Furthermore, it appears that the activity of these
compounds extends beyond the DF508 form of CFTR.
In an assay analogous to our high-throughput screen,
compound 1 produced a robust enhancement of chloride
efflux in C127 cells stably transfected with wild-type pro-
tein. If Tic diamides do, indeed, act to promote DF508
CFTR trafficking, the same mechanism may apply to

2090
B. H. Hirth et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2087–2091
Table 3. Azine carboxylic acid modified Tic diamides
4. Kerem, B. S.; Rommens, J. M.; Buchanan, D.; Mark-
iewicz, T. K.; Cox, A.; Chakravarti, M.; Buchwald, M.;
Tsui, L. C. Science 1989, 245, 1073–1080.
5. Welsh, M. J.; Tsui, L.-C.; Boat, T. F.; Beaudet, A. L. In
The Metabolic and Molecular Basis of Inherited Disease;
Scriver, C. R., Beaudet, A. L., Sly, W. S., Valle, D., Eds.;
McGraw-Hill: New York, 1995; pp 3799–3876.
HN
*
N
O
O
6. Galietta, L. J. V.; Moran, O. Curr. Opin. Pharmacol. 2004,
4, 497–503.
X
Y
7. (a) Illek, B.; Fischer, H.; Santos, G. F.; Widdicombe, J.
H.; Machen, T. E.; Reenstra, W. W. Am. J. Physiol. 1995,
268, C886–C893; (b) Illek, B.; Lizarzaburu, M. E.; Lee, V.;
Nantz, M. H.; Kurth, M. J.; Fischer, H. Am. J. Physiol.
2000, 279, C1838–C1846.
2
Z
R
Compd
*R/S
R2
X/Y/Z
EC50
(nM)
MRa
(%)
47
48
49
50
51
52
53
S
S
S
S
S
S
S
OPh
O-i-Pr
N/CH/CH
N/CH/CH
N/CH/CH
N/CH/CH
CH/N/CH
CH/N/CH
CH/CH/N
20
6.0
9.4
79
35
16
11
85
74
83
55
71
55
51
8. (a) Haws, C. M.; Nepomuceno, I. B.; Krouse, M. E.;
Wakelee, H.; Law, T.; Xia, Y.; Nguyen, H.; Wine, J. J.
Am. J. Physiol. 1996, 270, C1544–C1555; (b) Chappe, V.;
Mettey, Y.; Vierfond, J. M.; Hanrahan, J. W.; Gola, M.;
Verrier, B.; Becq, F. Br. J. Pharmacol. 1998, 123, 683–
693.
9. Marivingt-Mounir, C.; Norez, C.; Derand, R.; Bulteau-
Pignoux, L.; Nguyen-Huy, D.; Viossat, B.; Morgant, G.;
Becq, F.; Vierfond, J.-M.; Mettey, Y. J. Med. Chem. 2004,
47, 962–972.
O-Cyclohexyl
Cl
O-Cyclohexyl
O-i-Pr
O-i-Pr
a Maximum response relative to cold temperature control (100%).
10. (a) Devor, D. C.; Bridges, R. J.; Pilewski, J. M. Am. J.
Physiol. 2000, 279, C461–C479; (b) Caci, E.; Folli, C.;
Zegarra-Moran, O.; Ma, T.; Springsteel, M. F.; Sammel-
son, R. E.; Nantz, M. H.; Kurth, M. J.; Verkman, A. S.;
Galietta, L. J. V. Am. J. Physiol. 2003, 285, L180–L188.
11. Sammelson, R. E.; Ma, T.; Galietta, L. J. V.; Verkman, A.
S.; Kurth, M. J. Bioorg. Med. Chem. Lett. 2003, 13, 2509–
2512.
the wild-type protein, which is also known to be fairly
inefficiently processed, with only an estimated 20–25%
of the immature protein reaching the cell surface.16
In summary, screening of our in-house compound li-
brary led to the discovery of a novel class of CFTR
modulators capable of stimulating chloride secretion,
Tic diamides. Subsequent analoging was carried out in
two stages: rapid array synthesis (unpurified analogs)
was used to quickly and broadly assess SAR around
the lead while a more focused, traditional medicinal
chemistry effort followed. By optimizing both the acid
and amine regions of the lead compound, we were able
to improve CFTR activity of the series more than 30-
fold.
12. Murine C127 cells expressing DF508 CFTR were grown in
microtiter plates under a 37 °C, 5% CO2 atmosphere
(DulbeccoÕs Modified Eagle Medium supplemented with
fetal bovine serum). After three days, the cultures were
switched to serum-free medium and treated overnight with
test compounds (final concentration, 12.5 lM) and
MQAE (10 mM). The cells were then washed with a
chloride-containing buffer to remove extracellular MQAE
and incubated to quench intracellular MQAE fluores-
cence. Chloride buffer was replaced with nitrate containing
buffer with added forskolin (10 lM) and isobutylmethyl-
xanthine (100 lM). The buffers were adjusted to pH 7.4
and contained (mM): HEPES (10), K2HPO4 (2.4),
KH2PO4 (0.6), CaSO4 (1.0), MgSO4 (1.0), NaCl (150),
or NaNO3 (150), glucose (25). CFTR chloride transport
was determined by measuring the rate of increase in
MQAE fluorescence over the first five minutes (excitation
360 nM, emission 460 nM). False positives (i.e., com-
pounds which induce non-specific chloride leakage) were
identified by measuring chloride flux in cells treated with
nitrate containing buffer without added forskolin and
isobutylmethylxanthine.
Acknowledgements
This work was supported by a grant from the Cystic
Fibrosis Foundation.
References and notes
13. Denning, G. M.; Anderson, M. P.; Amara, J. F.; Marshall,
J.; Smith, A. E.; Welsh, M. J. Nature London 1992, 358,
761–764.
1. Cystic Fibrosis Foundation.org. http://www.cff.org
(accessed December 20, 2004).
2. Pilewski, J. M.; Frizzell, R. A. Physiol. Rev. 1999, 79,
S215–S255.
14. The racemic intermediates 2 and 3 were prepared from
commercially available N-(tert-butoxycarbonyl)-Tic and
Tic methyl ester, respectively. EDC amidation of the
former starting material with 4-chloroaniline followed by
trifluoroacetic acid mediated amine deprotection afforded
2. Amidation of the latter starting material with 3-
phenoxybenzoic acid followed by hydrolysis of the ester
furnished 3. F or2: 1H NMR (CDCl3) d 9.42 (s, 1H), 7.65–
7.53 (m, 2H), 7.36–7.04 (m, 6H), 4.12–3.97 (m, 2H), 3.68
(dd, J = 10.4, 5.4 Hz, 1H), 3.35 (dd, J = 16.3, 5.4 Hz, 1H),
2.90 (dd, J = 16.3, 10.4 Hz, 1H) ppm. For 3 (6:4 mixture of
3. (a) Riordan, J. R.; Rommens, J. M.; Kerem, B.-S.; Alon,
N.; Rozmahel, R.; Grzelczak, Z.; Zielenski, J.; Lok, S.;
Plavsic, N.; Chou, J.-L.; Drumm, L.; Iannuzzi, M. C.;
Collins, F. S.; Tsui, L.-C. Science 1989, 245, 1066–1072;
(b) Crawford, I.; Maloney, P. C.; Zeitlin, P. L.; Guggino,
W. B.; Hyde, S. C.; Turley, H.; Gatter, K. C.; Harris, A.;
Higgins, C. F. Proc. Natl. Acad. Sci. U.S.A. 1991, 88,
9262–9266; (c) Denning, G. M.; Ostedgaard, L. S.; Cheng,
S. H.; Smith, A. E.; Welsh, M. J. J. Clin. Invest. 1992, 89,
339–349.
1
rotamers): H NMR (CDCl3) d 7.44–6.78 (m, 13H), 5.29

B. H. Hirth et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2087–2091
2091
(t, J = 5.7 Hz, 0.6H), 5.18 (d, J = 17.9 Hz, 0.4H), 4.81–
4.73 (m, 0.4H), 4.62–4.47 (m, 1.6H), 3.30–3.15 (m, 1.6H),
3.13–3.01 (m, 0.4H) ppm.
purity by 1H NMR. As example, compound 12 (8:2
mixture of rotamers): H NMR (CDCl3) d 8.81 (s, 0.8H),
7.54 (br s, 0.2H), 7.51–6.74 (m, 17H), 5.27 (t, J = 6.4 Hz,
0.8H), 5.16 (d, J = 16.5 Hz, 0.2H), 4.85–4.74 (m, 0.2H),
4.72–4.52 (m, 1H), 4.44–4.31 (m, 0.8H), 3.65–3.52 (m,
0.8H), 3.51–3.37 (m, 0.2H), 3.19–3.03 (m, 1H), 2.55 (t,
J = 7.6 Hz, 2H), 1.64–1.49 (m, 2H), 1.35–1.12 (m, 8H),
0.87 (t, J = 6.6 Hz, 3H) ppm.
1
15. As with intermediates 2 and 3, preparation of purified Tic
diamide analogs relied on standard aminoacid chemistry.
Thus, racemic or L-N-(tert-butoxycarbonyl)-Tic was trea-
ted with the amine component and EDC. The resulting
amide was N-deprotected with trifluoroacetic acid and
then amide coupled to the carboxylic acid component. The
final products were purified to homogeneity by column
chromatography and deemed to be of 95% or greater
16. Santis, G. Basic Molecular Genetics. In Cystic Fibrosis;
Hodson, M. E., Geddes, D. M., Eds.; Arnold: London,
2000; pp 27–47.