Use of parallel-synthesis combinatorial libraries for rapid identification of potent FKBP12 inhibitors

Article abstract of DOI:10.1016/S0960-894X(02)00147-6

Using simple, inexpensive equipment, we have used solution-phase parallel synthesis to rapidly prepare hundreds of sulfonamide- and urea-containing FKBP inhibitors, resulting in rapid identification of extremely potent compounds in these series.

Full text of DOI:10.1016/S0960-894X(02)00147-6

Bioorganic & Medicinal Chemistry Letters 12 (2002) 1421–1428
Use of Parallel-Synthesis Combinatorial Libraries for Rapid
Identification of Potent FKBP12 Inhibitors
Chi Choi, Jia-He Li, Mark Vaal, Christine Thomas, David Limburg, Yong-Qian Wu,*
Yi Chen, Raj Soni, Chad Scott, Douglas T. Ross, Hong Guo, Pamela Howorth,
Heather Valentine, Shi Liang, Dawn Spicer, Mike Fuller, Joseph Steiner
and Gregory S. Hamilton
Guilford Pharmaceuticals Inc., Research Department, 6611 Tributary Street, Baltimore, MD 21224, USA
Received 2 November 2001; accepted 11 February 2002
Abstract—Using simple, inexpensive equipment, we have used solution-phase parallel synthesis to rapidly prepare hundreds of
sulfonamide- and urea-containing FKBP inhibitors, resulting in rapid identification of extremely potent compounds in these series.
# 2002 Elsevier Science Ltd. All rights reserved.
Immunophilins are ‘receptors’ for immunosupressant
drugs such as FK506 and cyclosporin A.1 Immunophi-
lins possess peptidyl-prolyl isomerase (rotamase, or
PPIase) activity. Recently it has been discovered that
the immunophilin FKBP12, which binds FK506, is
enriched 10- to 40-fold more in the brain than in the
immune tissues.2 Immunosuppressant drugs such as
FK506 have been shown to promote neuronal process
extension in vitro and regrowth of damaged peripheral
nerves in vivo.2,3 It has recently been demonstrated that
nonimmunosuppressive analogues of these drugs retain
these neurotrophic properties.4
isosteric mimetics (e.g., sulfonamides such as 1 and
ureas such as 2) have been reported in the literature,8,9
but no extensive development of the SARof these
compounds as FKBP12 inhibitors has appeared. Our
discovery that compounds such as GPI 1046 are a novel
class of potent neuroregenerative agents prompted us to
explore other structural classes of FKBP12 ligands.
We have recently described the neurotrophic actions of
simple N-glyoxyl prolyl and pipecolyl esters.5,6 These
small molecule FKBP12 ligands are devoid of immuno-
suppressive action but potently promote neurite out-
growth in neuronal cultures. Compounds such as GPI
1046 (Fig. 1) have been shown to possess neuroprotec-
tive and neuroregenerative properties in a variety of
animal models of neurodegeneration, thus suggesting
broad therapeutic potential for FKBP12 ligands in
nervous system disorders.6,7
FKBP12 ligands in which the a-ketoamide (N-glyoxyl)
moiety of FK506 and GPI 1046 has been replaced with
*Corresponding author. Tel.: +1-410-631-6823; fax: +1-410-631-
6848; e-mail: wuy@guilfordpharm.com
Figure 1. Ligands for FKBP12.
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00147-6

1422
C. Choi et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1421–1428
One of the approaches that we have deployed in
exploring other classes of FKBP12 ligands is solution-
phase, parallel synthesis. In an economical fashion, we
have quickly generated several series of compounds for
SARstudies and lead optimization using a Reacti-
Therm stirring-heating module, an expandable multi-
channel pipette, and a set up for solid-phase extrac-
tion.10 Reacti-Therm modules have been widely used by
pharmacokineticists and protein chemists. The module
provides heating and stirring functionalities. When
connected to a nitrogen outlet, the module can act as a
blow dryer. It can also provide an inert reaction atmo-
sphere when used in conjunction with proper septa.
Vials used with the module are readily amendable to the
solution transfer by the Matrix expandable multi-chan-
nel pipette.11 In addition to allowing quick reagent
addition, the multi-channel pipette facilitates product
transfer from vials to 96-well plates. The cost of the
module, 27 5-mL vials, caps, and septa is about $1500.
With the additional cost of the expandable multi-chan-
nel pipette, which is about $1000, the total cost of the
set up is no more than $3000. The low cost of the set up
makes it possible for chemists to carry out parallel
syntheses on the bench-top without having to invest
prematurely in expensive robotics.
isocyanates were scavenged by Tentagel-S-NH2 resin
after stirring for 16 h. The reactions were worked up by
passing the crude mixture through silica solid-phase
extraction columns. For optimization of the aryl alkyl
moiety, esters were added to each of the three vials in a
row, and sulfonyl chlorides, isocyanates, or thioisocy-
anates to each of the nine vials in a column. All pro-
ducts were checked for purity with HPLC. One-third of
the reaction mixtures were checked by FABMS for the
presence of products. The products (10–20 mg) were
collected with an average purity of 70%. An example of
a library is shown in Figure 2. Representative HPLC
traces are shown in Figure 3.
Compounds were tested for their ability to inhibit the
peptidyl-prolyl isomerase (PPIase, or rotamase) activity
of FKBP12 using the method of Kofron, as previously
described.12 Data are reported as apparent inhibition
constants (Ki’s). Although the products from the pre-
sently described parallel synthesis were used only in in
vitro studies, we have reported on the activity of these
compounds in in vivo models of neurodegeneration
elsewhere.13
We began by synthesizing libraries in which the sulfo-
namide, urea, and thiourea substituents were varied.
Adamantyl, cyclohexyl, and benzyl groups were found
to be particularly effective substituents in the urea,
thiourea, and sulfonamide series, respectively (19, 21,
and 48, Figs. 4 and 5). These moieties were held con-
stant in subsequent libraries while the ester side chains
were varied. Potent compounds in each series, such as
compounds 42, 43, and 56, were rapidly generated by
this method.
Compounds were synthesized as shown in Scheme 1.
N-(tert-Butyloxycarbonyl) proline or N-(tert-butyloxy-
carbonyl)pipecolic acid were esterified with various
alcohols in round bottom flasks using standard coupling
conditions to provide Boc-protected proline or pipecolic
acid ester. For optimization of the sulfonamide, urea, or
thiourea moieties, 2 mL of 1.48 mM esters solutions in
methylene chloride were pipetted to each of the nine
vials in a row. Deprotection of the Boc protecting group
with trifluoroacetic acid afforded proline or pipecolic
acid ester trifluoroacetic salts in 2 h. Solvent was then
removed by means of blow-drying. Solutions of sulfonyl
chlorides, isocyanates, or thioisocyanates in methylene
chloride (2 mL of 2.96 mM) were transferred to each of
the three vials in the column. Triethylamine was then
added. Excess sulfonyl chlorides, isocyanates, or thio-
Structure–activity trends were observed in both the sul-
fonamide and urea series. Examples of the trends
observed are shown in Tables 1 and 2. In general,
monosubstitution of the ring of the arylalkyl ester chain
with lipophilic substituents enhanced activity in the sul-
fonamides. Among the 4-halophenyl propyl side chains,
4-flourophenyl propyl was generally better than the
Scheme 1. Reagent and conditions: (a) R1OH, DCC, DMAP, CH2Cl2, rt; (b) trifluoroacetic acid; (c) 2 equiv R2SO2Cl or R2NCO or R2NCS, Et3N,
CH2Cl2.

C. Choi et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1421–1428
1423
Figure 2. An example of the library.
Figure 3. HPLC traces of two compounds from the parallel synthesis.

1424
C. Choi et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1421–1428
Figure 4. Procedure showing the lead optimization in the urea and thiourea series.

C. Choi et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1421–1428
1425
Figure 5. Procedure showing the lead optimization in the sulfonamide series.

1426
C. Choi et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1421–1428
4-chlorophenyl propyl, which was better than the 4-
bromophenyl propyl (Tables 1a and 2a). Best activities
were generally found with compounds with alkyl chains
length of 3–4 carbons (Table 1b). The presence of
electron donating groups such as methoxy was favor-
able for rotamase inhibition activity, and the 3,4- and
2,4-dimethoxy substitution patterns were particularly
active in the sulfonamide and urea series, respectively
Table 1. SARexamples in proline urea/thiourea series
(a)
58
1200 nM
59
2300 nM
60
3100 nM
61
4400 nM
62
6000 nM
63
>10,000 nM
64
850 nM
65
1500 nM
66
6300 nM
(b)
67
24,000 nM
68
1300 nM
69
820nM
70
2100 nM
71
1300 nM
72
1000 nM
73
470 nM
74
18,000 nM
75
47,000 nM
76
5200 nM
77
7200 nM

C. Choi et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1421–1428
1427
(e.g., 57 and 42). Benzylsulfonyl, adamantylcarbamoyl,
and cyclohexylthiocarbamoyl groups are found to be
the most efficacious N-substitutions in both proline and
pipecolic acid series. In the sulfonamide series, the order
of preference for sulfonamide substituent was con-
sistent, benzyl being greater than phenyl, and phenyl
greater than p-toluyl (Table 2b).
We have demonstrated the utility of a simple, inexpen-
sive parallel synthesis set up that has allowed us to
Table 2. SARexamples of pipecolic acid sulfonamide series
(a)
78
740 nM
79
3800 nM
80
8000 nM
81
1400 nM
82
4400 nM
83
320 nM
84
420 nM
85
2700 nM
(b)
86
260 nM
87
480 nM
88
180 nM
89
220 nM
90
400 nM
91
92 nM
92
120 nM
57
47 nM

1428
C. Choi et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1421–1428
rapidly generate and identify extremely potent com-
pounds from hundreds of sulfonamides and urea rota-
mase inhibitors. The set up has also enabled us to
rapidly develop structure–activity relationships in these
series. Compounds are generated in quantity and purity
that are suitable for biological evaluation. This rather
economical and simple set up makes it possible to
implement parallel synthesis combinatorial libraries as a
routine tool used by all medicinal chemists.
6. Hamilton, G. S.; Steiner, J. P. Curr. Pharm. Des. 1997, 3,
405.
7. Steiner, J. P.; Hamilton, G. S.; Ross, D. T.; Valentine,
H. L.; Guo, H.; Connolly, M. A.; Liang, S.; Ramsey, C.; Li,
J.-H.; Huang, W.; Howorth, P.; Soni, R.; Fuller, M.; Sauer,
H.; Nowotnick, A.; Suzdak, P. D. Proc. Natl. Acad. Sci.
U.S.A. 1997, 94, 2019.
8. Holt, D. A.; Konialian-Beck, A. L.; Oh, H.-J.; Yen, H.-K.;
Rozamus, L. W.; Krog, A. J.; Erhard, K. F.; Ortiz, E.; Levy,
M. A.; Brandt, M.; Bossard, M. J.; Luengo, J. I. Bioorg. Med.
Chem. Lett. 1994, 4, 315.
9. Dragovich, P. S.; Barker, J. E.; French, J.; Imbacuan, M.;
Kalish, V. J.; Kissinger, C. R.; Knighton, D. R.; Lewis, C. T.;
Moomaw, E. W.; Parge, H. E.; Pelletier, L. A. K.; Prins, T. J.;
Showalter, R. E.; Tatlock, J. H.; Tucker, K. D.; Villafranca,
J. E. J. Med. Chem. 1996, 39, 1872.
References and Notes
1. Schreiber, S. L. Science 1991, 251, 283.
2. Sabatini, D. M.; Snyder, S. H. Nat. Med. 1995, 1, 32.
3. Snyder, S. H.; Sabatini, D. M.; Lai, M. M.; Steiner, J. P.;
Hamilton, G. S.; Suzdak, P. D. Trends Pharm. Sci. 1998, 19, 21.
4. Steiner, J. P.; Connolly, M. A.; Valentine, H. L.; Hamilton,
G. S.; Dawson, T. M.; Hester, L.; Snyder, S. H. Nat. Med.
1997, 3, 421.
10. Available from Pierce Instruments: http://www.pierce
net.com.
11. Available from Matrix Technologies: http://www.
matrixtechcorp.com.
12. Kofron, J. L.; Kuzmic, P.; Kishore, V.; Colon-Bonilla, E.;
Rich, D. H. Biochemistry 1991, 30, 6127.
5. Hamilton, G. S.; Huang, W.; Connolly, M. A.; Ross, D. T.;
Guo, H.; Valentine, H. L.; Suzdak, P. D.; Steiner, J. P. Bioorg.
Med. Chem. Lett. 1997, 7, 1785.
13. Li, J.-H.; Hamilton, G. S.; Huang, W.; Connolly, M. A.;
Ross, D. T.; Guo, H.; Valentine, H. L.; Steiner, J. P. 214th
ACS National Meeting, Las Vegas, NV; 1997; MEDI 183.