2,6,9-Trisubstituted purine derivatives as protein A mimetics for the treatment of autoimmune diseases

Article abstract of DOI:10.1016/j.bmcl.2008.10.116

A series of 9-substituted and 2,9-disubstituted 6-(3-aminophenylamino) purines were synthesized and evaluated for their ability to mimic protein A binding to human IgG antibody. The structure-activity relationship (SAR) demonstrates that the 6-(3-aminoani

Full text of DOI:10.1016/j.bmcl.2008.10.116

Bioorganic & Medicinal Chemistry Letters 19 (2009) 242–246
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
2,6,9-Trisubstituted purine derivatives as protein A mimetics
for the treatment of autoimmune diseases
*
Boulos Zacharie , Daniel Fortin, Nicole Wilb, Jean-François Bienvenu,
Michel Asselin, Brigitte Grouix, Christopher Penney
ProMetic BioSciences Inc., 500 Boulevard Cartier Ouest, Bureau 150, Laval, Que., Canada H7V 5B7
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of 9-substituted and 2,9-disubstituted 6-(3-aminophenylamino) purines were synthesized and
evaluated for their ability to mimic protein A binding to human IgG antibody. The structure–activity rela-
tionship (SAR) demonstrates that the 6-(3-aminoanilinyl) purine component was essential for activity.
Purine 14 demonstrated significant activity, compared to protein A. These compounds may prove useful
for the treatment of autoimmune diseases.
Received 8 September 2008
Revised 17 October 2008
Accepted 20 October 2008
Available online 31 October 2008
Ó 2008 Elsevier Ltd. All rights reserved.
Keywords:
Purines
Protein A mimetics
IgG
Autoimmune diseases
Staphylococcal protein A (SPA) is a cell-surface component of
the bacterial pathogen Staphylococcus aureus. This protein, with a
molecular weight of 42,000, contains five domains which selec-
tively bind the tail (Fc) portion of human and mouse antibodies.1
The interaction between SPA and immunoglobulin G (IgG) is one
of the most studied protein–protein interactions.2 The binding
properties of protein A makes it commercially important for the
purification of monoclonal antibodies. In general, antibodies are
purified by classical (e.g., ion-exchange) column or affinity chro-
matography with protein A attached to a solid-phase support.3
However, purification with SPA affinity absorbents is costly and
these adsorbents have been shown to leak protein A and contam-
inate purified immunoglobulin products.4 As a result, there is sig-
nificant interest in developing a synthetic mimetic of SPA, or a
protein A mimetic, which is more economical and stable. However,
few molecules have been reported in the literature which mimic
bacterial protein A and can be used to bind IgG. They are small syn-
thetic compounds covalently attached to a solid-phase support3a,5
or larger peptides6 or polypeptides.7 However, this approach is
limited to molecules attached to solid-phase matrices.
requires the use of protein A covalently linked to an inert silica
matrix whereby the patient’s blood is passed through this column
in a manner similar to kidney dialysis (apheresis). Unfortunately,
this treatment requires multiple visits to the doctor’s office and
the outcome after treatment is often unpredictable.9 Subsequently,
there is a need for a novel, safe, non toxic small molecule mimetic of
protein A which can be administered as a drug.
The purine scaffold has been exploited for designing biologi-
cally relevant molecules with broad biomedical value as thera-
peutics.10 These compounds have attracted interest as molecular
probes or lead molecules for drug development including inhibi-
tors of kinase signaling.10e As part of our effort toward the discov-
ery of therapeutics for the treatment of inflammatory conditions,
we examined small molecules that are mimetics of protein A. In
this letter, the synthesis and anti-inflammatory activities of di-
and trisubstituted purines are reported as exemplified by the
general structure 1.
H2N
NH
N
Protein A also has therapeutic utility but its toxicity and cost
N
R2
N
limit its therapeutic use. Nonetheless,
a protein A column
(ProsorbaÒ) from Cypress was approved by the US FDA for the treat-
ment of autoimmune diseases; immune thrombocytopenic purpura
in 1987 and severe rheumatoid arthritis in 1999.8 This treatment
N
n
R
1
The general synthetic procedure for purine compounds is out-
lined in Schemes 1 and 2. Scheme 1 illustrates the two synthetic
* Corresponding author. Tel.: +1 450 781 1394x2251; fax: +1 450 781 1403.
E-mail address: b.zacharie@prometic.com (B. Zacharie).
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.10.116

B. Zacharie et al. / Bioorg. Med. Chem. Lett. 19 (2009) 242–246
243
49) display significant activity comparable to protein A (Tables
1–3). As shown in Table 1, analogues of disubstituted purine 4
were synthesized in which different aliphatic (10–13) or aro-
matic (14–16) amines were varied at position 6 while maintain-
ing the 9-(4-aminophenethylamino)purine component. Only 3-
aminoaniline 14 showed activity. Clearly, the 3-amino function
of the aniline appears to be essential and activity declines upon
substitution with other groups (3-OMe 15 or 3-fluoro 16). On
the basis of this result, 14 was selected as a core structure upon
which to define a structure–activity relationship for a trisubsti-
tuted purine scaffold. Table 2 illustrates the variation at the 9
position (R1) on the purine ring. Prior work15 indicated the
importance of 4-aminophenethyl amine at this position. This
fragment enhances good in vitro antibody binding activity and
compound 14 supports this finding. Our results showed that
4-carboxymethylbenzyl (23) or 4-aminobenzyl (17) substituents
display the same in vitro activity as protein A. Also, the corre-
sponding analogues with halogens demonstrate reasonable
activity. For example, the 3-chloro (28) and the 4-chloro (27)
benzyl analogues display as good activity as 23 and 17 while
activity is reduced with 4-fluoro (20) and 3-fluoro (21) benzyl
derivatives. Homologation of the chain with a one-carbon spacer
between the 9-substituted purine ring and the phenyl (29) led
to a drop in activity. Substitution of the phenyl ring of the 9-
phenethyl group (R1) with polar functions gave different activi-
ties. For example, 4-amino (14), 3-fluoro (25) and 4-fluorophen-
ethyl (24) analogues demonstrate good activity. However,
replacement of the 4-fluoro with 4-chloro (26) or 3-amino
groups (19) results in a loss of activity. The corresponding ana-
logue with a 4-methyl substituent (30) retains the activity of
compound 14. Interestingly, methyl substitution of the two-
carbon spacer of the 9-phenethyl group (22) showed a signifi-
cant increase in activity; equipotent with SPA. Also, analogues
were synthesized in which a heteroatom was added to the
two-carbon spacer of 29. As a result, the oxygen derivatives
(32–34) demonstrated good activity. Introduction of rigidity,
for example, 31, diminishes the activity relative to the
9-phenethyl derivative (29). Surprisingly, 9-substituted alkyl
(36) possesses comparable activity with SPA.
Cl
N
N
N
N
N
NR2
R1
c
Cl
N
a
b
2
N
N
N
N
NR2
N
N
R1
N
H
d
N
4
N
N
H
3
Scheme 1. Reagents and conditions: (a) aralkyl alcohol, DIAD, Ph3P, THF or alkyl
bromide, K2CO3, DMF, 25 °C, overnight; (b) arylamine or alkylamine, ethanol, 50 °C,
21 h; (c) arylamine, DIEA, n-butanol, 90 °C, overnight; (d) alkyl bromide, K2CO3,
DMF, 100 °C, 20 min, microwave.
Cl
N
Cl
Cl
N
N
N
N
N
N
N
N
d
b
a
or c
H2N
N
H
F
N
N
H
F
N
N
R1
7
6
NR2
NR2
N
N
R3
e
F
N
N
N
N
R1
R1
8
9
Scheme 2. Reagents and conditions: (a) NaNO2, HBF4; (b) aralkyl alcohol, DIAD,
Ph3P, THF, 25 °C, overnight; (c) alkyl iodide, Cs2CO3, dioxane, 25 °C, overnight; (d)
arylamine, DIEA, n-butanol, 65 °C, overnight; (e) ammonia or alkylamine, DIEA, n-
butanol, 110 °C, overnight.
routes employed for the disubstituted purine derivative 4. The
first approach was the Mitsunobu reaction11 of 6-chloropurine
with the aralkyl alcohol to give 9-substituted aralkyl derivative
2. This reaction proceeded in high yield (>80%) and is general
for different aralkyl alcohols. In addition, the desired N-9 adduct
was the major product with only traces of N-7 adduct. Com-
pound 2 was also prepared by reacting 6-chloropurine with alkyl
bromide in the presence of potassium carbonate.12 The chloro
intermediate 2 was then treated with different alkyl or aryl-
amines to give the final product 4. Other routes were also exam-
ined for the preparation of disubstituted purine 4. For example,
6-chloropurine was treated with arylamine in ethanol at 50 °C
to give 6-substituted purine 3 in respectable yield. The latter
was treated with alkyl bromide in presence of a base at 100 °C
for 20 min by microwave to afford compound 4. Both procedures
were practical and used to generate 6,9-disubstituted analogues
of 4. Scheme 2 demonstrates the synthetic method for the tri-
substituted purine derivative 9 where the key intermediates is
the 2-fluoro derivative 7. This compound was prepared by con-
verting 2-amino-6-chloropurine to the corresponding fluoro
derivative 613 followed by a Mitsunobu reaction with the aralkyl
alcohol to give compound 7 in high yield (>85%). Again, the N-9
adduct was the major product. The chloro derivative 7 was then
reacted with aryl amines in hot butanol in the presence of base
to give the expected fluoro derivative 8 in good yield. Displace-
ment of fluorine with different alkyl or aralkyl amines gave the
trisubstituted product 9. The reaction is practical and amenable
to scale-up starting from commercially available 2-amino-6-chlo-
ropurine. Also, it offers the advantage of a general procedure for
the preparation of a variety of analogues of 2,6,9-trisubstituted
purines.
It was next decided to synthesize trisubstituted purines.
Therefore, analogues of the lead compound (14) were made in
which different groups were introduced at position 2 of the pur-
ine ring. Table 3 summarizes the results. In general, it can be
seen that a rigid amine at R2 is preferred since activity is
dependent upon the distance and relative degrees of freedom
Table 1
IC50 (lM) of 6-substituted-9-(4-aminophenylethyl)-purine relative to protein A as
ascertained by a protein A human IgG competitive ELISA
R
N
N
NH2
N
N
Compound
R
IC50
(lM) Compound
R
IC50
(lM)
in PBS
in PBS
N
Protein A
0.2
13
Inactive
1.0
NH
HN
NH2
OMe
F
HN
10
Inactive
Inactive
Inactive
14
15
16
NH
NH
HN
HN
HN
HN
11
12
Inactive
Inactive
A number of purine derivatives of 4 and 9 were synthe-
sized14 and evaluated as SPA mimetics in a competitive protein
A binding ELISA assay. Six compounds (17, 22, 23, 32, 34 and
NH2

244
B. Zacharie et al. / Bioorg. Med. Chem. Lett. 19 (2009) 242–246
Table 2
IC50 (lM) of 9-substituted-6-(3-aminophenylamino)-purine relative to protein A as ascertained by a protein A human IgG competitive ELISA
H2N
NH
N
N
N
N
R1
Compound
Protein A
R1
IC50
0.2
(lM) in PBS
Compound
R1
IC50
3.6
(lM) in PBS
H2C
26
Cl
H2C
H2C
H2C
H2C
14
17
18
19
20
21
22
23
24
25
1
27
28
29
30
31
32
33
34
35
36
0.5
0.4
2.3
1.5
12.5
0.3
0.5
0.2
0.6
0.4
NH2
Cl
Cl
0.2
NH2
H2C
H2C
H2C
Inactive
5.1
OH
NH2
H2C
H2C
H2C
H2C
1.7
F
F
O
O
H2C
1.8
F
CH3
H2C
0.2
C
F
H
Cl
H2C
H2C
H2C
H2C
O
0.3
COOCH3
O
0.6
F
F
0.5
H2C
H2C
at this position. Cyclopentylamine (49) is the most effective sub-
stituent among rigid amines evaluated such as isopropyl (45)
and cyclobutyl (48). Activity was not improved, relative to 14,
in the case of benzyl amine derivatives (51) and (52). While
hydrogen substitution, compound 14, results in a reasonable
activity, the 2-amino (37) and the 2-fluoro (38) analogues are
less active.
On the basis of the above results, it was of interest to determine
if this activity would translate into an in vivo effect. Therefore, the
lead compound 14 was tested in two inflammatory disease mod-
els. Table 4 shows the effect of intravenous administration of 14
on oxazolone-induced delayed-type hypersensitivity (DTH) in
mice.16 This compound induces a significant reduction of inflam-
mation as seen by reduced ear thickness. It displays similar po-
tency to methotrexate and reduces redness, crust formation and
ear swelling. Also, the effect of oral administration of 14 was stud-
ied on DTH. Figure 1 represents the activity of this compound in
comparison with hydrocortisone as the positive control.17 Again,
14 induces a significant reduction in inflammation (p = 0.002,
day 14) as seen by reduced ear thickness in both challenges 1
and 2. Encouraged by these results, it was of interest to determine
the effect of 14 on Freund’s adjuvant-induced arthritis.18 As illus-
trated in Figure 2 (100% of the animals rapidly developed a syno-
vitis. A significant reduction (20%, day 21) in the severity of
arthritis (inflammatory index) was observed by intravenous injec-
tion of methotrexate (positive control) by day 13 and over. A sig-
nificant reduction (up to 25%) of the inflammatory index was also
observed with intravenous injection of 14 from day 1 to day 21 at
a dose of 25 mg/kg. Compound 14 also displayed reduction (up to
30%) in inflammation when the experiment was repeated with
compound administered orally at a dose of 50 mg/kg (data not
shown). Animals treated with 14 showed less inflammation than
the control group and the activity of this compound is similar to
methotrexate.
A first-in-class series of low molecular weight synthetic mole-
cules is described that mimic the ability of protein A to bind to hu-
man IgG antibody. The SAR studies demonstrate the importance of
the presence of a 1,3-phenylenediamine substituent. The hydro-
phobicity of these purines is important for binding to the IgG tail
portion. These compounds, represented by lead compound 14,
show good in vivo activity in standard models of inflammation
when administered by oral and intravenous routes. They offer a
unique approach for the treatment of autoimmune diseases by vir-
tue of their novel biochemical target.

B. Zacharie et al. / Bioorg. Med. Chem. Lett. 19 (2009) 242–246
245
Table 3
IC50 M) of 2-substituted-6-(3-aminophenylamino)-9-(4-aminophenylethyl)-purine relative to protein A as ascertained by a protein A human IgG competitive ELISA
(l
H2N
NH
N
N
N
R2
NH2
N
Compound
Protein A
R2
IC50
0.2
(lM) in PBS
Compound
R2
IC50
6
(lM) in PBS
HN
44
14
37
38
39
40
41
42
43
H
1
45
46
47
48
49
50
51
52
1.1
13
HN
HN
NH2
F
13
Inactive
Inactive
64
0.6
2.3
0.2
9
HN
NH2
OH
HN
HN
HN
HN
HN
OH
Inactive
4.8
HN
HN
OH
1.8
14
HN
HN
NH2
22
HN
490
440
390
340
290
240
Table 4
Effect of intravenous administration of compound 14 on DTH in mice
Inflammation/ear thickness (mm)
Challenge 1
Challenge 2
Control
Methotrexate
Compound 14
0.29 0.10
0.17 0.04
0.19 0.06
0.83 0.46
0.46 0.17
0.56 0.12
p = 0.001
p = 0.008
p = 0.02
p = 0.05
1
2
3
4 5 6 7 8 11 12 13 14 15 16 17 18 19 20 21
Day
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
Compound 14
Control
MTX
Figure 2. Effect of intravenous administration of compound 14 on adjuvant-
induced arthritis in rats.
Acknowledgments
We thank Dr. Jean Barabé for revising the manuscript. We also
acknowledge Marc-Alexandre Duceppe, Marie-Ève Fafard, Dan-
nyck Gaudreau, Joël Lechasseur, Angela Moutzoukis, Natalie St-
Amant and François Sarra-Bournet for their technical assistance.
We thank Lyne Marcil for the skillful preparation of this
manuscript.
5
6
7
12
4
14
3
11
13
10
Day
Control
Compound 14 (150)
Hydrocortisone
Compound 14 (50)
p < 0.05
Figure 1. Effect of oral administration of compound 14 on DTH in mice.
References and notes
1. (a) Advances in Immunology: Langone, J.J., Dixon, F.J., Kuhnel, H.G. Eds. New
York, 1982, p. 157. Vol. 32.; (b) Nilsson, B.; Abrahmsén, L. Methods Enzymol.
1990, 185, 144; (c) Moks, T.; Abrahmsén, L.; Nilsson, B.; Hellman, U.; Sjoquist,
J.; Uhlén, M. Eur. J. Biochem. 1986, 156, 637.
2. Delano, W.; Ultsch, M.; de Vos, A. M.; Wells, J. Science 2000, 287, 1279. and
references cited therein.
Registry numbers: 14, 917460-18-5; 17, 917381-70-5; 19,
917381-75-0; 24, 917381-77-2; 37, 917381-65-8; 40, 917381-
71-6; 42, 917381-67-0; 43, 917381-74-9; 44, 917381-73-8; 46,
917381-68-1; 47, 917381-66-9; 49, 917381-76-1.

246
B. Zacharie et al. / Bioorg. Med. Chem. Lett. 19 (2009) 242–246
3. (a) Li, R.; Dowd, V.; Stewart, D. J.; Burton, J.; Lowe, C. R. Nat. Biotechnol. 1998, 16,
(s, 1H), 7.80–7.20 (m, 7H), 7.51 (t, J = 8.02 Hz, 1H), 5.66 (s, 2H); LRMS (ESI): m/z
332 (MH+), 354 (MNa+). Data for 22 brown solid 1H NMR (400 MHz, D2O): d
8.26 (s, 1H), 8.09 (s, 1H), 7.49–7.48 (m, 1H), 7.40–7.36 (m, 2H), 7.23–7.18 (m,
1H), 7.08–7.06 (m, 1H), 6.95–6.86 (m, 3H), 5.73 (dd, J = 7.04 and 14.28 Hz, 1H),
1.81 (d, J = 7.04 Hz, 3H); LRMS (ESI): m/z 349 (MH+). Data for 49 brown solid;
mp: 250 °C (decomposition); 1H NMR (400 MHz, CD3OD): d 8.45 (s, 1H), 8.13 (s,
1H), 7.90 (m, 1H), 7.57 (t, J = 8.22 Hz, 1H), 7.40 (m, 4H), 7.20 (m, 1H), 4.57 (t,
J = 7.04 Hz, 2H), 4.35 (q, J = 6.65 Hz, 1H), 3.30 (m, 2H), 2.10 (m, 2H), 1.75 (m,
6H); LRMS (ESI): m/z 429 (MH+), 451 (MNa+).
190; (b) Fassina, G.; Ruvo, M.; Palombo, G.; Verdoliva, A.; Marino, M. J. Biochem.
Biophys. Methods 2001, 49, 481.
4. Godfrey, M. A.; Kwasowski, P.; Clift, R.; Marks, V. J. Immunol. Methods 1993, 160,
97.
5. Fern Teng, S.; Sproule, K.; Hussain, A.; Lowe, C. R. J. Mol. Recognit. 1999, 12, 67.
6. (a) Delano, W. L.; Ultsch, M. H.; de Vos, A. M.; Wells, J. A. Science 2000, 287,
1279; (b) Ehrlich, G. K.; Bailon, P. J. Biochem. Biophys. Methods 2001, 49, 443; (c)
Fassina, G.; Verdoliva, A.; Oderna, M. R.; Ruvo, M.; Cassini, G. J. Mol. Recognit.
1996, 9, 564.
15. Penney, C. et al. WO 2006/024175.
7. Johnson, C. P.; Jensen, I. E.; Prakasam, A.; Vijayendran, R.; Leckband, D.
Bioconjugate Chem. 2003, 14, 974.
8. (a) Sasso, E. H.; Merrill, C.; Furst, D. E. Ther. Apher. 2001, 5, 84; (b) Brunner, J.;
Kern, P. M.; Gaipl, U. S.; Voll, R. E.; Kalden, J. R.; Wiesenhutter, C. W.; Hermann,
M. Mod. Rheumatol. 2005, 15, 9.
9. Silverman, G.; Goodyear, C. S.; Siegel, D. L. Transfusion 2005, 45, 274.
10. (a) Wright, G. E. Acta Biochim. Pol. 1996, 43, 115; (b) Lou, B. Drug Discov. Today
2001, 24, 1288; (c) Davis, J. T. Angew. Chem., Int. Ed. 2004, 43, 668; (d)
16. On day 0, mice were sensitized with 100 ll of oxazolone in 5% acetone. On day
0, 1 and 2, mice were treated by intravenous administration of the vehicle
(control) or methotrexate (MTX; positive control) or the compound at 50 mg/
kg. Mice were challenged with an application of 50 ll of oxazolone on the
surface of the right ear (first challenge, day 3; second challenge, day 10). Ear
thickness was measured on day 4 to day 7, and on day 11 to 14. Redness and
crust formation was also observed. Mice were sacrificed on day 14. TDTH (CD4)
cells play an important role in regulating the intensity of the DTH response.
17. Oral administration of compound 14 was undertaken following the same
protocol as Ref. 15 with the exception that compound 14 was orally
administered at 50 or 150 mg/kg from day 0 to day 13.
18. Adjuvant-induced arthritis was induced in female Lewis rats by the injection of
lyophilized Mycobacterium butyricum suspended in mineral oil into the
footpad. The development of arthritis was monitored over a 3-week period
post-adjuvant injection. Inflammation peaks at day 3 following the adjuvant
administration. Immune activation appears around day 14. Compounds were
injected iv day À3, À2 and À1 pre-adjuvant injection and at day 10, 11 and 12
post-adjuvant injection. Body weight was recorded. The arthritis index, which
Br
ꢀndvang, M.; Gundersen, L.-L. Bioorg. Med. Chem. 2007, 15, 7144; (e) Sabat,
M.; VanRens, J. C.; Clark, M. P.; Brugel, T. A.; Maier, J.; Bookland, R. G.;
Laufersweiler, M. J.; Laughlin, S. K.; Golebiowski, A.; De, B.; Hsieh, L. C.; Walter,
R. L.; Mekel, M. J.; Janusz, M. J. Bioorg. Med. Chem. Lett. 2006, 16, 4360.
11. Chang, Y.-T.; Gray, N. S.; Rosania, G. R.; Sutherlin, D. P.; Kwon, S.; Norman, T. C.;
Sarohia, R.; Leost, M.; Meijer, L.; Schultz, P. G. Chem. Biol. 1999, 6, 361.
12. Langli, G.; Gundersen, L.-L.; Rise, F. Tetrahedron 1996, 52, 5625.
13. Gray, N. S.; Kwon, S.; Schultz, G. P. Tetrahedron Lett. 1997, 38, 1161.
14. Zacharie, B. et al. WO 2006/136005. The purity and identity of all the products
were monitored by LC/MS at 210 or 250 nm (Agilent 1100) using an analytical
C18 column (75 Â 4.6 mm,
5
l
m) with different gradients of CH3CN–H2O
is a measure of inflammation (edema), redness, and stiffness of the
containing 0.01% TFA as the eluant. All compounds demonstrated >98% purity.
1H NMR and ESI-MS characterization data were consistent with the expected
structure. For example: 14 brown solid; mp: 167 °C; 1H NMR (400 MHz,
CD3OD): d 8.50 (s, 1H), 8.30 (s, 1H), 8.20–8.15 (m, 1H), 7.80–7.70 (m, 1H), 7.58
(t, J = 8.24 Hz, 1H), 7.40–7.20 (m, 5H), 4.63 (t, J = 7.04 Hz, 2H), 3.31 (m, 2H);
LRMS (ESI): m/z 346 (MH+), 368 (MNa+). Data for 17 brown solid; mp: 225 °C
(decomposition); 1H NMR (400 MHz, CD3OD): d 8.75 (s, 1H), 8.53 (s, 1H), 7.98
articulations, was used to monitor the development of the disease. The
degree of arthritis was determined by measuring two perpendicular diameters
of the ankles in the mediolateral and dorsoventral planes using a caliper. Joint
circumference in millimeters is then calculated using a geometric formula.
Both the incidence and severity of the arthritis was evaluated. Incidence is
defined as the number of rats with clinical evidence of joint inflammation
during the study period.