Synthesis and biological evaluation of lisofylline (LSF) analogs as a potential treatment for Type 1 diabetes

Article abstract of DOI:10.1016/j.bmcl.2006.04.036

Lisofylline (LSF, 1-(5-R-hydroxyhexyl)-3,7-dimethylxanthine) is an anti-inflammatory agent that protects β-cells from Th1 cytokine-induced dysfunction and reduces the onset of Type 1 diabetes in non-obese diabetic (NOD) mice. Due to its low potency, poor oral bioavailability, and short half-life, the widespread clinical utility of LSF may be limited. Our goal has been to develop new agents based on the LSF structural motif that resolve the potency and pharmacokinetic liabilities of LSF. In this study, we have generated a focused library of LSF analogs that maintain the side chain (5-R-hydroxyhexyl) constant, while substituting a variety of nitrogen-containing heterocyclic substructures for the xanthine moiety of LSF. This library includes the xanthine-like (5-aza-7-deazaxanthine), as well as non-xanthine-like skeletons. The LSF analogs were evaluated in a pancreatic β-cell line for the effects on apoptosis protection and insulin release. The metabolic stability of selected compounds was also tested.

Full text of DOI:10.1016/j.bmcl.2006.04.036

Bioorganic & Medicinal Chemistry Letters 16 (2006) 3401–3405
Synthesis and biological evaluation of lisofylline (LSF) analogs
as a potential treatment for Type 1 diabetes
Peng Cui,a,* Timothy L. Macdonald,a Meng Chenb and Jerry L. Nadlerb
aDepartment of Chemistry, University of Virginia, Charlottesville, VA 22904, USA
bDivision of Endocrinology and Metabolism, Department of Medicine, University of Virginia, Charlottesville, VA 22908, USA
Received 21 February 2006; revised 5 April 2006; accepted 5 April 2006
Available online 2 May 2006
Abstract—Lisofylline (LSF, 1-(5-R-hydroxyhexyl)-3,7-dimethylxanthine) is an anti-inflammatory agent that protects b-cells from
Th1 cytokine-induced dysfunction and reduces the onset of Type 1 diabetes in non-obese diabetic (NOD) mice. Due to its low poten-
cy, poor oral bioavailability, and short half-life, the widespread clinical utility of LSF may be limited. Our goal has been to develop
new agents based on the LSF structural motif that resolve the potency and pharmacokinetic liabilities of LSF. In this study, we have
generated a focused library of LSF analogs that maintain the side chain (5-R-hydroxyhexyl) constant, while substituting a variety of
nitrogen-containing heterocyclic substructures for the xanthine moiety of LSF. This library includes the xanthine-like (5-aza-7-
deazaxanthine), as well as non-xanthine-like skeletons. The LSF analogs were evaluated in a pancreatic b-cell line for the effects
on apoptosis protection and insulin release. The metabolic stability of selected compounds was also tested.
Ó 2006 Elsevier Ltd. All rights reserved.
OH
O
Type 1 diabetes, a common and widespread disease
occurring in every part of the world, is an autoimmune
disorder that results from the immune-mediated inflam-
matory destruction of insulin-producing b-cells in pan-
creatic islets. There are an estimated 500,000 to 1
million people with Type 1 diabetes in the US today.
Although the specific pathogenic mechanisms in Type
1 diabetes are not defined, it is clear that activated T
cells and macrophages are required for the initiation.
Once activated, macrophages secrete several inflamma-
tory cytokines, such as interleukin 1b (IL-1b), interleu-
kin 12 (IL-12), and tumor necrosis factor a (TNF-a),
and trigger interferon-c (IFN-c) production from acti-
vated T cells.1 These cytokines are cytotoxic to b-cells
by inducing the formation of oxygen free radicals, nitric
oxide, and lipid peroxides within b-cells and enhance
Th1-cell-mediated inflammatory responses, which are
CH3
N
H3C
N
N
O
N
CH3
Figure 1. Structure of Lisofylline (LSF, (1-(5-R-hydroxyhexyl)-3,7-
dimethylxanthine)).
of graft vs. host disease and to prevent the onset of exper-
imental autoimmune encephalomyelitis by blocking IL-
12-induced T helper 1 differentiation.3,4 It abrogated re-
lease of inflammatory cytokines during oxidative lung
injury and reduced inflammatory cytokine release in re-
sponse to cytotoxic cancer chemotherapy.5 LSF also
showed beneficial effects in several inflammatory disor-
ders such as sepsis, hypoxia, and hemorrhagic organ inju-
ry.6,7 LSF has also been shown to protect b-cells from
multiple inflammatory cytokine-mediated injuries by its
ability to maintain insulin secretory capability and cell
viability.2 Although not established, the mechanism of
LSF-induced protection may be due to promotion of b-
cell mitochondrial metabolism, normalizing mitochon-
drial membrane potential and stimulating energy produc-
tion. This unique spectrum of activity suggests that
agents such as LSF could have clinical utility in prevent-
ing b-cell damage during the development of Type 1 dia-
betes or after islet cell transplantation. This hypothesis
was supported by studies that showed LSF could signifi-
2
responsible for b-cell destruction.
Lisofylline (LSF) (Fig. 1), 1-(5-R-hydroxyhexyl)-3,7-
dimethylxanthine, is a novel anti-inflammatory com-
pound that was originally used to reduce the incidence
Keywords: Type 1 diabetes; Lisofylline; LSF; LSF analogs; 5-Aza-7-
deazaxanthine.
*
Corresponding author. Tel.: +1 434 924 0595; fax: +1 434 982
2302; e-mail: pc3n@virginia.edu
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.04.036

3402
P. Cui et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3401–3405
cantly reduce spontaneous Type 1 diabetes development
in the non-obese diabetic (NOD) mouse.1
containing heterocyclic substructures for the xanthine
moiety of LSF.
LSF has disadvantages that may limit its clinical devel-
opment: LSF exhibits low (to non-existent) oral bio-
availability, has an extremely short half-life, requiring
constant intravenous infusion in humans, and has rela-
tively low potency for direct islet protective effects,
requiring concentrations of 20 lM in vitro and at least
25 mg/kg body weight in rodents to see beneficial effects.
In efforts directed at enhancing the potency of LSF,
Klein et al.8 prepared a library of LSF analogs based
on the xanthine framework and studied the SAR for
suppression of Th1 differentiation through IL-12 block-
ade. Despite considerable structural variation, only rela-
tively small differences in activity were observed among
the analogs studied. Importantly, it was found that the
enantiomeric S-LSF (which possesses the S-alcohol con-
figuration) was inactive, implicating a highly specific
interaction of the R isomer with a binding point on
the target site.
The synthetic route to 5-aza-7-deazaxanthine-LSF 1
(Scheme 1) commenced with coupling of glycine 2
and acetic anhydride 3 in presence of pyridine afford-
ing acetamidoacetone 4,10 which was hydrolyzed in
HCl to form aminoacetone hydrochloride 5. The ke-
tone was then transformed to its acetal by treatment
with a mixture of trimethyl orthoformate and metha-
nol in the presence of p-toluenesulfonic acid. The free
amine 6 was obtained by treatment with sodium
hydroxide.11 Compound 6 was ready for later reac-
tions with a protected carbonyl group and a free
Table 1. The selected heterocyclic compounds being investigated
O
H
N
O
H
N
Br
O
NH
NH
O
NH
NH
N
N
N
H
O
N
O
O
13
14
15
16
17
The aim of our investigation was to synthesize a focused
library of small molecules based on the LSF structure to
elucidate the SAR for potency and pharmacokinetic
parameters. In one approach, we synthesized analogs
of the 5-aza-7-deazaxanthine family (Fig. 2). The 5-
aza-7-deazaxanthine structure is isosteric with xanthine,
but has a distinct electronic character as evidenced by its
lower basicity and its photophysical properties.9 In a
second approach, we substituted a series of nitrogen-
O
O
H
N
H
N
O
H3C
O
NH
NH
NH
N
S
N
H
O
O
N
H
CH3
O
O
18
19
20
21
22
H
N
H
N
H
N
N
H
O
S
O
OH
O
23
24
25
26
CH3
H3C
N
N
N
O
N
N
H
N
H
CH3
27
28
Figure 2. 5-Aza-7-deazaxanthine LSF (5,7-isoLSF) 1.
CH3 CH3
O
OH
iii
O
O
O
H
N
i
H2N
ii
O
O
CH3
NH2.HCl
O
+
NH2
H3C
H3C
O
CH3
H3C
H3C
O
2
3
5
6
4
O
O
O
Cl
CH3
HN
N
HN
N
iv
v
HN
N
vi
N
N
CH3
O
N
H
Cl
O
N
H
N
H
N
O
N
H
Cl
N
Cl
O
7
8
9
10
O
CH3
OAc
O
N
OH
O
CH3
CH3
vii
HN
N
ix
viii
H3C
N
N
H3C
N
N
N
O
N
N
N
O
O
N
CH3
CH3
CH3
11
1
12
Scheme 1. Reagents and conditions: (i) pyridine, reflux, 6 h, 78%; (ii) HCl, H2O, reflux, 6 h, 60%; (iii) trimethylorthoformate, MeOH, p-
toluenesulfonic acid monohydrate, reflux 24 h, then 3 N NaOH, 49%; (iv) NaOH, H2O; (v) 6, H2O, reflux, 3 h, 17%; (vi) H2SO4, 95 °C, 1.5 h, 35%;
(vii) NaOH, CH3I, H2O, acetone, rt, 24 h, 34%; (viii) 5-(R)-acetoxy-1-chlorohexane, NaH, DMSO, 70–80 °C, overnight; (ix) 1 M HCl in ether,
MeOH, rt, 12 h, 60% two steps.

P. Cui et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3401–3405
3403
amine. Compound 8 was prepared by controlled
hydrolysis of cyanuric chloride 7 with sodium hydrox-
ide in aqueous solution.12 Condensation of compound
8 with compound 6 in aqueous media at reflux tem-
perature furnished compound 9 and the hydrolysis
of the protecting acetal groups was achieved in situ.
The ring annulation was accomplished by heating
with concentrated sulfuric acid.13 Compound 10 was
methylated with methyl iodide in basic solution to
give 11, which was then coupled with 5-(R)-acetoxy-
1-chlorohexane14 followed by removing the acetate
group to afford compound 1.
cyclic compounds with 5-(R)-acetoxy-1-chlorohexane,
followed by hydrolysis of the resulting acetates. For
the synthesis of LSF analogs based on phthalhydraz-
ide 13 (Scheme 2), the byproduct 32 was also isolated.
This compound proved valuable in elucidating the
structure of 31 by demonstrating that alkylation oc-
curred on the oxygen.
We investigated the effects of LSF analogs on b-cells in
the mouse insulin-secreting INS-1 cell line.15 Evidence
showed that LSF and its analogs could protect b-cells
from cytokine-induced cell death. INS-1 cells were treat-
ed with a combination of recombinant IL-1b, INFc, and
TNFa with 10À5–103 lM LSF and its analogs. At a
concentration of 10À2 lM, LSF showed a maximum
protective effect (Fig. 3a). There was no further increase
in b-cell protection when the LSF concentration was
increased from 10À2 to 102 lM. Most of the LSF
In our second study, the side-chain moiety (5-R-
hydroxyhexyl) was held constant, while a series of
nitrogen-containing heterocyclic scaffolds were substi-
tuted for the xanthine substructure (Table 1). The
LSF analogs were synthesized by coupling the hetero-
O
O
O
OAc
O
NH
N
i
N
N
NH
NH
+
O
O
OAc
2
30
OAc
29
ii
ii
O
O
O
O
OH
NH
N
N
N
+
OH
OH
31
32
Scheme 2. Reagents and conditions: (i) NaH, DMF, then 5-(R)-acetoxy-1-chlorohexane, tetrabutylammonium iodide, 70–80 °C, overnight; (ii) HCl
(1 M in ether), MeOH, rt, overnight, 31, 22%, two steps.
0.06
0.05
0.04
0.03
0.02
0.06
0.05
0.04
0.03
0.02
b
a
10-5 10-4 10-3 10-2 10-1 100 101 102 103
10-5 10-4 10-3 10-2 10-1 100 101 102 103
(μM)
(μM)
0.06
c
0.05
0.04
0.03
0.02
10-5 10-4 10-3 10-2 10-1 100 101 102 103
(μM)
Figure 3. b-cell protection in INS-1 cells: (a) LSF, (b) compound 1, and (c) compound 31.

3404
P. Cui et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3401–3405
4
3
2
1
3
a
b
2
1
0
0
4
3
2
1
c
0
Figure 4. Insulin secretion in INS-1 cells: (a) LSF, (b) compound 1, and (c) compound 31.(+CÀG3 = w/compound, w/o cytokines, 3 mM glucose;
ÀCÀG3 = w/o compound, w/o cytokines, 3 mM glucose; +CÀG28 = w/compound, w/o cytokine, 28 mM glucose; ÀCÀG28 = w/o compound, w/o
cytokine, 28 mM glucose; +C+G3 = w/ compound, w/cytokines, 3 mM glucose; ÀC+G3 = w/o compound, w/ cytokines, 3 mM glucose;
+C+G28 = w/compound, w/cytokines, 28 mM glucose; ÀC+G28 = w/o compound, w/cytokines, 28 mM glucose).
analogs showed some protective effects on b-cells (data
not shown), especially compounds 1 (Fig. 3b) and 31
(Fig. 3c), which proved effective at low concentrations.
Both of these two compounds showed comparative
activity with LSF at concentration of 10À2 lM. No fur-
ther increases in b-cell protection were observed when
analog concentrations were increased.16
lived LSF analogs, since LSF is completely metabolized
by the liver with half-life of less than 3 min in vivo.
In summary, we have synthesized a series of LSF ana-
logs that maintain the 5(R)-hydroxyhexyl side chain,
while varying the xanthine core substructure. Two lead
compounds 1 and 31 have been identified that protect
b-cells from cytokine-induced injury and maintain insu-
lin secretory capability. In addition, these lead com-
pounds showed metabolic stability in vitro. Ongoing
studies of the pharmacokinetic profiles of these mole-
cules will determine if the goal of generating an LSF
analog with improved potency, metabolic stability, and
oral bioavailability has been achieved.
We also examined the effects of LSF and selected ana-
logs on insulin secretion in INS-1 cells with and without
inflammatory cytokines.17 A maintenance in basal
(3 mM) and glucose-stimulated (28 mM) insulin release
was observed by 20 lM of LSF and its analogs (Fig. 4).
Table 2 shows results of metabolic stability experiments
for compounds 1 and 31. This assay evaluated the stabil-
ity of 1 M concentrations of 1 and 31 in human liver
microsomes at pH 7.4. Compound levels were measured
by LC/MS. Three reference drugs were used: warfarin
(high stability), propranolol (lower stability), and testos-
terone (poor stability). Both of the compounds showed
acceptable stability values under these in vitro condi-
tions. These results are relevant to our search for longer
Acknowledgment
This work was supported by grants from the NIH
(R21DK 063521).
References and notes
1. Yang, Z. D.; Chen, M.; Wu, R.; McDuffie, M.; Nadler, J.
L. Diabetologia 2002, 45, 1307.
2. Chen, M.; Yang, Z. D.; Wu, R.; Nadler, J. L. Endocri-
nology 2002, 143, 2341.
3. Coon, M. E.; Diegel, M.; Leshinsky, N.; Klaus, S. J.
J. Immunol. 1999, 163, 6567.
4. Bright, J. J.; Du, C.; Coon, M.; Sriram, S.; Klaus, S. J.
J. Immunol. 1998, 161, 7015.
Table 2. Metabolic stability of LSF analogs
Warfarin Propranolol Testosterone
105 34
1
31
Metabolic
stability
% remaining
9
82 75

P. Cui et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3401–3405
3405
5. De Vries, P.; Singer, J. W. Exp. Hematol. 2000, 28, 916.
6. Rice, G. C.; Brown, P. C.; Nelson, R. J.; Bianco, J. A.;
Singer, J. W.; Bursten, S. Proc. Natl. Acad. Sci. U.S.A.
1994, 91, 3857.
7. Wattanasirichaigoon, S.; Menconi, M. J.; Fink, M. P.
Crit. Care Med. 2000, 28, 1540.
8. Klein, J. P.; Klaus, S. J.; Kumar, A. M.; Gong, B. Int
Patent under PCT, WO 00/61583, 2000; Chem. Abstr.
2000, 742096.
9. Rao, P.; Benner, S. A. J. Org. Chem. 2001, 66, 5012.
10. Hepworth, J. D. Org. Synth. Coll. 1973, 5, 27.
11. Calabretta, R.; Giordano, C.; Gallina, C.; Morea, V.;
Consalvi, V.; Scandurra, R. Eur. J. Med. Chem. 1995, 30,
931.
days. The cells were plated at a density of 105/cm2.
Culture vessels were coated with poly-D-lysine and
gelatin (Sigma, St. Louis, MO) to retain detached and
dead cells so that seeding cell numbers reflect the actual
cell numbers after all treatment conditions. INS-1 cells
were treated with the combination of recombinant
mouse IL-1b (5 ng/ml), IFNc (100 ng/ml), and TNFa
(10 ng/ml; R&D Systems, Inc., Minneapolis, MN) sus-
pended in complete RPMI medium. LSF (provided by
Cell Therapeutics, Inc., Seattle, WA) and analogs were
added simultaneously with the cytokines in complete
RPMI medium. All treatments were performed for 18 h.
16. Beta cells were treated with apoptosis detecting dye for 2–
3 h at room temperature. Apoptotic cells were recognized
with purple-red color under a microscope. After washing
to eliminate free dye and adding dye release reagent, color
density was quantified by reading at OD 450 nM.
17. At the end of treatment, cells were washed with Krebs–
Ringer-bicarbonate–HEPES buffer (KRB) containing
134 mm NaCl, 4.7 mm KCl, 1.2 mm KH2PO4, 1.2 mm
MgSO4, 1.0 mm CaCl2, 10 mm HEPES, and 0.1% BSA at
37 °C, pH 7.4. They were preincubated in the same buffer
for 30 min, followed by 60-min incubation in KRB
supplemented with 15 mM D-glucose (J. T. Baker, Phil-
lipsburg, NJ). The supernatant was harvested and sub-
jected to centrifugation to eliminate residue cells. Insulin
secreted into the supernatant was measured by RIA with
mouse insulin as a standard.
12. Horrobin, S. J. Chem. Soc. 1963, 4130.
13. Kim, S. H.; Bartholomew, D. G.; Allen, L. B.; Robins, R.
K.; Revankar, G. R.; Dea, P. J. Med. Chem. 1978, 21, 883.
14. Klein, J. P.; Leigh, A. J.; Michnick, J.; Kumar, A. M.;
Underiner, G. E. US Patent, US005629423A, 1997; Chem.
Abstr. 1997, 342002.
15. Cells from passages 10–30 were maintained in RPMI
1640 medium(LifeTechnologies, Inc., Gaithersburg, MD)
supplemented with 10% heat-inactivated FBS, 10 mm
HEPES, 200 lm L-glutamine, 1 mm sodium pyruvate,
5 nm 2-mercaptoethanol, 50 U/ml penicillin, and 50 lg/
ml streptomycin at pH 7.4. The cells were cultured at
37 °C in a humidified incubator supplied with 5%
carbon dioxide. Fresh medium was replaced every 2