2,4,6-Trisubstituted triazines as protein A mimetics for the treatment of autoimmune diseases

Article abstract of DOI:10.1021/jm901403r

A first-in-class series of low molecular weight trisubstituted triazines were synthesized and evaluated for their ability tomimic proteinAbinding to human IgGantibody. The structure-activity relationship (SAR) demonstrates that the 1,3-phenylenediamine component was essential for robust activity. Twenty-two compounds, represented by lead molecule 34, displayed significant activity compared to protein A. These compounds may prove useful for the treatment of autoimmune disease.

Full text of DOI:10.1021/jm901403r

1138 J. Med. Chem. 2010, 53, 1138–1145
DOI: 10.1021/jm901403r
2,4,6-Trisubstituted Triazines as Protein A Mimetics for the Treatment of Autoimmune Diseases
ꢀ
Boulos Zacharie,* Shaun D. Abbott, Jean-Franc-ois Bienvenu, Alan D. Cameron, Josee Cloutier, Jean-Simon Duceppe,
Abdallah Ezzitouni, Daniel Fortin, Karine Houde, Caroline Lauzon, Nancie Moreau, Valerie Perron, Nicole Wilb,
ꢀ
ꢁ
ꢀ
Michel Asselin, Andre Doucet, Marie-Eve Fafard, Dannyck Gaudreau, Brigitte Grouix, Franc-ois Sarra-Bournet,
Natalie St-Amant, Lyne Gagnon, and Christopher L. Penney
ProMetic BioSciences Inc., 500 Boulevard Cartier Ouest, Bureau 150, Laval, Queꢀbec H7V 5B7, Canada
Received September 21, 2009
A first-in-class series of low molecular weight trisubstituted triazines were synthesized and evaluated for
their ability to mimic protein A binding to human IgG antibody. The structure-activityrelationship(SAR)
demonstrates that the 1,3-phenylenediamine component was essential for robust activity. Twenty-two
compounds, represented by lead molecule 34, displayed significant activity compared to protein A. These
compounds may prove useful for the treatment of autoimmune disease.
Introduction
requires the use of protein A covalently linked to a silica
matrix whereby the patient’s blood is passed through an
external device in a manner similar to kidney dialysis
(apheresis). This treatment requires multiple visits to the
doctor’s office, and the outcome after treatment is often
unpredictable.10 Indeed, the manufacture of this product
was discontinued in 2008. Subsequently, there is need for a
novel, safe, nontoxic small molecule mimetic of protein A11
that can be administered as a drug.
1,3,5-Triazine compounds have been studied extensively and
are the subject of many reviews.12 Thetriazinescaffoldprovides
the basis for the design of biologically relevant molecules with
widespread application as therapeutics. For example, these
compounds possess potent antiprotozoal,13 anticancer,14 anti-
malarial,15 antiviral,16 and antimicrobial activity.17 As part of
our autoimmune disease program, small molecules were ex-
amined as potential mimetics of protein A. It was reported3a
that the aromatic side chains of the hydrophobic core dipeptide
Phe132-Tyr133 of protein A are able to bind the Fc portion of
IgG. This structural feature was therefore taken into considera-
tion when designing a protein A mimetic. Aniline and tyramine
substituents may mimic the side chain of Phe132-Tyr133, while a
triazine ring was used as a core structure to maintain optimal
orientation of the two mimetic groups. In this paper, the
synthesis and activities of trisubstituted triazines are reported
as exemplified by the general structure 1, where NR1, NR2,
NR3, and NR4 represent primary or secondary amines.
Staphylococcal protein A (MW=42 000) is displayed on
the cell membrane of the Gram-positive bacterial pathogen
S. aureus. This protein contains five domains that selectively
bind with high affinity the tail (Fca) portion of human and
mouse antibodies.1 The interaction between protein A and
immunoglobulin G (IgG) is one of the most studied protein-
protein interactions.2 The binding properties of protein A
make it the standard tool for the purification of monoclonal
antibodies. In general, antibodies are purified by ion-ex-
change or affinity column chromatography with protein A
attached to a solid-phase support.3 However, purification
with protein A affinity adsorbents is costly, as well as inade-
quate regarding purity and yield.4 Furthermore, these adsor-
bents have been shown to leak protein A and contaminate
purified immunoglobulin products.5 Consequently, there is
significant interest in developing a synthetic mimetic of pro-
tein A that is more economical and stable. However, few
molecules have been reported in the literature that mimic
bacterial protein A and can be used to bind IgG. These are
small synthetic compounds,3a larger peptides,6 or polypep-
tides7 covalently attached to a solid-phase support.3a,8
Protein A also has therapeutic utility, but its toxicity
and cost limit its use. Nonetheless, a protein A column from
Cypress was approved by the U.S. FDA for the treatment of
autoimmune diseases: immune thrombocytopenic purpura in
1987 and severe rheumatoid arthritis in 1999.9 This treatment
*To whom correspondence should be addressed. Phone: (450) 781-
1394, extension 2251. Fax: (450) 781-1403. E-mail: b.zacharie@
prometic.com.
a Abbreviations: dsDNA, double-standard deoxyribunocleic acid;
DTH, delayed-type hypersensitivity; ELISA, enzyme-linked immuno-
sorbent assay; F1, first generation; Fc, fragment crystallizable; FcγR,
fragment crystallizable type γ receptor; Fab, fragment antigen-binding;
FDA, Food and Drug Administration; HAG, heat aggregated human
immunoglobulin G; IgG, immunoglobulin G; MRL/lpr, a substrain of
Murphy Roths Large mice that is genetically predisposed to the develop-
ment of systemic lupus erythematosus-like syndrome; NZB/NZW,
cross-breed of New Zealand black and New Zealand white strains;
PBS, phosphate buffered saline; SLE, systemic lupus erythematosus.
Chemistry
The starting material for these compounds is cyanuric
chloride. This is an inexpensive commercially available re-
agent which makes its use attractive. The ease of displacement
r
pubs.acs.org/jmc
Published on Web 01/04/2010
2010 American Chemical Society

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1139
Scheme 1a
a For intermediates 4a-8a, NR2 = 3-(Boc-amino)aniline, NR3 = NR4 = 2-aminoethanol, and H2N-linker-NH2 = 4-aminophenethylamine.
Reagents and conditions: (a) cyanuric chloride, acetone/water, 0 °C; (b) hydroxyalkyl- or aminoarylamines, NaHCO3, THF/acetone/water; (c)
aminoalkyl- or aminoarylamines, THF, 60 °C; (d) NaHCO3, THF/acetone/water; (e) CH2Cl2, 4 M HCl/dioxane.
of chlorine atoms in cyanuric chloride by various nucleophiles
enhances the utility of this reagent for the preparation of
mono-, di-, and trisubstituted 1,3,5-triazines under controlled
temperature.12 The general synthetic procedure for triazine
compounds is outlined in Scheme 1. It illustrates the route
employedfor thetrisubstitutedtriazine compound8 where the
key intermediate is amino derivative 6. This compound was
prepared by two different approaches. The first method was
the treatment of m-substituted aniline2 with cyanuricchloride
at 0 °C to give dichlorotriazine derivative 3. This reaction
proceeded in excellent yield (>99%) and is general for
different aniline derivatives. The next step was the reaction
of compound 3 with different hydroxyalkylamines or amino-
arylamines in the presence of base to afford triazine 4 in high
yield (>90%). This compound was then reacted with differ-
ent aminoalkyl- or aminoarylamines to give the intermediate
6. The latter was also prepared by converting dichlorotriazine
3 tothe corresponding aminederivative5 followed by reaction
with hydroxyalkyl- or aminoarylamines. However, com-
pound 5 was only isolated in low yield (40%) because of the
formation of byproduct. Best results were obtained using the
initial approach that gave high yield and purity of amino
derivative 6. The next step was linkage of the two triazine
rings. This was achieved by slow addition of 4,6-dichlorotri-
azine bearing an aliphatic or aromatic amine on position 2 to
amino derivative 6 in the presence of sodium bicarbonate to
give the expected chloro derivative 7 in high yield (>80%).
This reaction is general for different aminoalkyl-, aminoaryl-,
or aminoaralkylamines and can be used to generate linked bis-
triazine analogues 7 with selected linkers. The final step was
displacement of the last chlorine atom by various alkyl-, aryl-,
or aralkylamines. Thus, treatment of compound 7 with
different amines at 60 °C gave the final intermediate 8 in high
yield (>80%). Subsequent acid-induced deprotection gave
the final compound 1 in quantitative yield. The synthesis of a
representative example of compound 1 (compound 10) is
described in detail. This synthetic route is practical, amenable
to scale-up, and offers a general procedure for the preparation
of a variety of analogues of triazine compound 1.
Results and Discussion
A number of triazine derivatives of structural type 1 were
synthesized and evaluated as protein A mimetics by compe-
titive ELISA. Twenty-two compounds (10-12, 14, 19-23, 26,
29, 31-36, 40-44) display significant activity comparable to
protein A (Tables 1-3). As shown in Table 1 (series I
compounds), analogues of 1 were synthesized in which NR1
and NR2 were different aliphatic or aromatic amines while
maintaining the linker as 4-aminophenethylamine and NR3
and NR4 as ethanolamine. The most active triazine com-
pounds are 10 and 15. Clearly, the 3-amino function of the
3-aminoaniline group (NR1) appears to be essential, and
activity declines upon replacement of 3-aminoaniline with
other groups such as 2-fluoro- (9) or 3-methylsulfonamido-
aniline (18) or aminoethylamidine (17). Analogues were also
synthesized in which 3-aminoaniline (NR2) was replaced by
othersubstituted anilines. Inthiscase, 3-fluoro- (11), 4-fluoro-
(12), 4-chloro- (13), and 3-methylsulfonylaniline (14) display
less activity then compounds 10 and 15. Replacement of
aminoaniline (10) or ethanolamine (15) at NR2 with ethyl-
enediamine (16) results in a loss of activity. On the basis of this
result, compound 10 was selected as a core structure upon
which to define a structure-activity relationship for triazine
scaffold 1.
Table 2 (series II compounds) illustrates the effect of the
linker between the two triazine rings in compound 1, where
NR1 and NR2 are 3-aminoaniline and NR3 and NR4 are
ethanolamine. Aliphatic linkers with a two- or three-carbon
spacer (19, 20, 22, or 23) demonstrate good activity compared
to protein A. However, an increase in chain length of one

1140 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3
Zacharie et al.
Table 1. IC50 (nM) of Series I Compounds Relative to Protein A As Ascertained by a Protein A Human IgG Competitive ELISA
a Compounds were assayed in both PBS and 20% DMSO/PBS.
carbon (21) results in reduction of activity. In contrast,
addition of an oxygen atom to the four-carbon spacer (24)
led to reasonable activity. Selection of the linker is critical,
since activity is dependent upon the distance between the
two triazine rings. Replacement of the alkyl spacer by
aromatic linkers gave different activities. For example, 4-
aminoaniline (26) demonstrates better activity than 3-ami-
noaniline (25). The best linkers were 4-aminophenethyl-
amine (10) or its cyclic structure (29). These analogues
display the same activity as protein A. Prior work18 indi-
cated the importance of 4-aminophenethylamine as a linker
at this position. This moiety enhances good in vitro anti-
body binding activity, and compound 10 supports this
finding. Interestingly, linkers such as 3-aminophenethyl-
amine (28) or 4-amino-R-methylbenzylamine (isomeric
attachment of the aminoethyl group) (27) displayed a
decrease in activity. Again, the distance and relative degrees
of freedom are important at this position.
It was next decided to synthesize triazine 1 in which NR1
and NR2 were 3-aminoaniline, NR3 was ethanolamine, and
NR4 was varied with different aromatic or aliphatic amines
(Table 3, series III compounds). The phenethylamine group
(30) results in a loss of activity. However, aminobenzylamine
(32) or 2-, 3-, or 4- aminophenethylamine (31, 33, or 34)
substituents display the same in vitro activity as protein A.
Also, hydroxy substitution of the two-carbon spacer of
3-aminophenethylamine (35) retains activity comparable with
that of compound 33. Surprisingly, 3-hydroxyphenethyl-
amine (36) possesses comparable activity with protein A while
4-hydroxy- (37), 4-fluoro- (38), and 4-dimethylaminophene-
thylamines (39) lose activity. Good activity is also obser-
ved with an aminoanilino substituent. For example, the

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1141
Table 2. IC50 (nM) of Series II Compounds Relative to Protein A As Ascertained by a Protein A Human IgG Competitive ELISA
a Compounds were assayed in both PBS and 20% DMSO/PBS. Results for the PBS assay are reported. Results in 20% DMSO/PBS were consistent
with results in PBS with the exception of compound 24 (5192 nM for 20% DMSO in PBS), probably because of solubility issues.
3-aminoaniline substituent (40) is more potent than protein A.
When NR4 is an aliphatic amine, the activities were variable
and a rigid amine is preferred at this position. Piperazine (44) is
the most effective substituent among these latter amines, e.g.,
N,N-dimethylaminoethylamine (42) and dihydropyrrole (43).
On the basis of the above results, triazine compound 34 was
selected as the lead structure. It was of interest to demonstrate
that this compound binds directly to IgG as in the case of
protein A. Indeed, compound 34 linked to a solid-phase
support was able to bind and purify IgG.19 Figure 1 shows a
fitting20a of compound34intothe binding siteofthe B domain
of protein A. The binding site is largely hydrophobic (green)
and contains only a few sites for polar contacts (purple).20b It
is possible to explain the significant activity of compound 34
compared to the inactive compound 37, perhaps because of a
process of protonation or deprotonation implicated in the
interaction of this compound and the binding site of protein
A. Furthermore, 34 was able to inhibit the binding of heat
aggregated human IgG to the FcγR-I (CD64, high affinity
receptor for monomeric and aggregated IgG), FcγR-IIa
(CD32a, lowaffinityreceptorforaggregatedIgGandimmune
complex), and FcγR-IIIb (CD16b, low affinity receptor for
immune complex) receptors, as illustrated in Table 4. In the
competitive binding ELISA, 34 demonstrated equipotent
affinity for IgG relative to protein A. However, a difference
in affinity for FcγR was observed relative to protein A. Both
protein A and 34 displayed almost equipotent affinity for
FcγR-I. A different affinity profile for FcγR-IIa and FcγR-
IIIb was observed forproteinA and34. This difference may be
explained by inhibition of binding by a multimeric protein
versus a low molecular weight compound.
relevant rodent model for autoimmune disease. For this
purpose, use was made of two strains of mice that replicate
the symptoms of systemic lupus erythematosus (SLE).21
Figure 2 illustrates the effect of weekly intravenous admini-
stration of 34 (50 mg/kg) on SLE in NZB/NZW-F1 mice.
Treatment with 34 induces up to a 2-fold reduction of anti-
dsDNA antibody without significant depletion of total IgG
compared to vehicle.19 Most probably, compound 34 increases
or facilitates the elimination of antibody immune complexes.
Another SLE experiment was performed using MRL/IPR
mice. Mice were treated by intravenous injection of 34 (50 mg/
kg, once a week) or vehicle once a week for 5 consecutive weeks
(between weeks 18 and 22).22 Mice were sacrificed, and
immune complex depositions were recorded in kidney tissue
by electron microscopy. Micrographs A and B (Figure 3)
represent kidney tissue from the control group (vehicle). In
Figure 3A (4000ꢀ), many small and large immune complex
depositions are observed in the epithelial (E) and mesangial
(M) area. In Figure 3B (7000ꢀ), abundant and very large
immune complex depositions are mostly observed in the
mesangial area. Micrographs C and D (Figure 3) represent
kidney tissue from the group treated with 34. In Figure 3C
(3500ꢀ), only a few immune complex depositions are observed
in the mesangial area. In Figure 3D (7000ꢀ), only one smaller
immune complex deposition is observed. However, mesangial
cells appear to be very active, as demonstrated by the increase
in lysosomes observed in an earlier stage of the pathology.
Conclusion
A first-in-class series of low molecular weight triazine
dimers is described that mimic the ability of protein A to bind
to murine and human IgG antibody. The SAR studies
It was next desired to demonstrate that the binding activity
of 34 would translate into a significant in vivo effect in a

1142 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3
Zacharie et al.
Table 3. IC50 (nM) of Series III Compounds Relative to Protein A As Ascertained by a Protein A Human IgG Competitive ELISA
a Compounds were assayed in both PBS and 20% DMSO/PBS. Results for the PBS assay are reported. Results in 20% DMSO/PBS were consistent
with results in PBS.
Table 4. IC50 (M) of Compound 34 Relative to Protein A As Ascer-
tained by a Heat Aggregated Human IgG (HAG) Binding Assay
demonstrate the importance of the presence of a 1,3-phenyl-
enediamine substituent. Lead structure 34 exhibits significant
in vivo activity in a standard rodent model for autoimmune
binding assay, IC50 (M)
disease where antibody plays an important role in disease
pathology. Additionally, 34 displays significant in vivo
activity in other rodent models for immunological disease
where antibody does not play a key role in disease pathology
(e.g., DTH and adjuvant arthritis; results to be published
elsewhere). In particular, compound 34 offers a novel ap-
proach to the treatment of autoimmune disease by virtue of a
novel biochemical target coupled with significant in vivo
activity.
protein A
competitive
FcγR-I-
HAG
FcγR-IIa-
HAG
FcγR-IIIb-
HAG
compd
protein A
34
1.6 ꢀ 10-7
4.5 ꢀ 10-7
1 ꢀ 10-10
>1 ꢀ 10-6
8.3 ꢀ 10-8
2.9 ꢀ 10-6
1.2 ꢀ 10-5
5.7 ꢀ 10-5
(75 mm ꢀ 4.6 mm, 5 μm), with gradients of acetonitrile/water
containing 0.01% TFA and a flow rate of 2 mL/min (30 °C).
Gradients used were as follows: gradient A, 1-40%/6 min;
gradient B, 15-99%/6 min; gradient C, 15-99%/10 min;
gradient D, 10-40%/6 min; gradient E, 1-20%/6 min. The
purity was g97% for all compounds and was determined by
LC-MS.
Experimental Section
Melting points were obtained on an Electrothermal MEL-
TEMP melting point apparatus. NMR spectra were recorded at
400 MHz for 1H NMR, 377MHz for 19FNMR, and100 MHz for
13C NMR. All HPLC and mass spectra were recorded on a HP
1100 LC-MS Agilent instrument using a diode array detector,
monitoring at 210 and 254 nm, and an analytical C18 column
2-([3-tert-Butoxycarbonylaminophenyl]amino)-4,6-dichloro-
[1,3,5]triazine (3). A solution of cyanuric chloride (2.37 g, 12.9
mmol) in acetone (16 mL) was cooled to 0 °C, and ice-water (6.6
mL) was added. The suspension was treated dropwise with a
solution of 2 (2.65 g, 12.7 mmol) in acetone (7.7 mL). The pH of
the resultant clear solution was adjusted to 7-8 by addition of

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1143
Figure 1. Postulated binding site of compound 34 (orange) onto
the Fc portion of IgG.
Figure 3. Electron micrographs of kidney tissue from SLE experi-
ment in MRL/LPR mice. Micrographs A and B are from control
group, and micrographs C and D are from the 34-treated group.
3.67 (t, J=5.6 Hz, 1H), 3.61 (broad, 1H), 3.46-3.52 (m, 1H),
1.51 (s, 9H). 13C NMR (CD3OD, 100 MHz) δ: 169.37, 168.56,
166.36, 166.19, 164.12, 163.93, 154.06, 153.98, 139.85, 139.76,
139.62, 139.24, 139.15, 128.92, 128.68, 128.62, 115.23, 114.76,
113.88, 111.36, 111.03, 79.77, 79.64, 65.75, 60.51, 60.38, 43.28,
43.04, 27.79, 27.60 (mixture of rotamers). LRMS (ESI): m/z
381.2 (M(35Cl)Hþ). HPLC: tR = 3.4 min (gradient B, purity
100% at 254 nm).
2-[4-[2-[4-Aminophenyl]ethylamino]-6-([3-tert-butoxycarbonyl-
aminophenyl]amino)[1,3,5]triazine-2-ylamino]ethanol (6a). A solu-
tion of 4a (214 mg, 0.56 mmol) in anhydrous tetrahydrofuran
(6 mL) was added to 4-aminophenethylamine (230 mg, 1.69
mmol), and the mixture was stirred at 60 °C overnight. The
suspension was dissolved by addition of methanol and was
evaporated onto silica gel. Purification on silica gel (elution with
10-75% ethyl acetate in hexanes) gave 5a asa whitesolid(220 mg,
81%). Mp=154-155 °C. 1H NMR (CD3OD, 400 MHz) δ: 7.26
(broad, 2H), 7.13 (t, J=7.8 Hz, 2H), 6.98 (d, J=8.2 Hz, 4H), 6.68
(d, J=2.5 Hz, 2H), 3.68 (broad, 2 H), 3.52 (broad, 4H), 2.72 (t, J=
7.4 Hz, 2H), 1.49 (s, 9H). LRMS (ESI): m/z 481.4 (MHþ). HPLC:
tR=1.7 min (gradient B, purity 100% at 254 nm).
2-[4-([3-tert-Butoxycarbonylaminophenyl]amino)-6-[2-[4-[4-
([3-tert-butoxycarbonylaminophenyl]amino)-6-chloro[1,3,5]tri-
azine-2-ylamino]phenyl]ethylamino][1,3,5]triazine-2-ylamino]-
ethanol (7a). A solution of 6a (220 mg, 0.46 mmol) in tetra-
hydrofuran/acetone/water (4:1:1, 6 mL) was treated with 3 (179
mg, 0.50 mmol) and 5% w/v aqueous sodium bicarbonate (0.4
mL), and the mixture was stirred at room temperature over-
night. The mixture was treated with water (5 mL) and extra-
cted with ethyl acetate (2 ꢀ 5 mL). Combined extracts were
washed with saturated sodium chloride (10 mL), dried over
sodium sulfate, and evaporated in vacuo to give the crude
product. Purification on silica gel (elution with 10-100% ethyl
acetate/hexanes) gave 7a as an off-white solid (303 mg, 83%).
Figure 2. Inhibition of serum dsDNA autoantibodies by com-
pound 34 in NZB/NZW mice.
5% w/v aqueous sodium bicarbonate (27 mL), and the mixture
was stirred at 0 °C. The resultant precipitate was collected by
filtration, washed with water (50 mL), and dried in vacuo to give
3 as a pale-brown solid (4.56 g, quantitative). Mp=172-173 °C.
1H NMR (CD3OD, 400 MHz) δ: 7.69-7.73 (m, 1H), 7.14-7.30
(m, 3H), 1.52 (s, 1H). 13C NMR (CD3OD, 100 MHz) δ: 170.32,
169.42, 164.52, 153.39, 140.72, 137.71, 129.53, 116.47, 115.78,
112.55, 79.78, 28.78. LRMS (ESI): m/z 356.0 (M(35Cl2)Hþ).
HPLC: tR=4.4 min (gradient B, purity 100% at 254 nm).
2-[4-([3-tert-Butoxycarbonylaminophenyl]amino)-6-chloro[1,3,5]-
triazine-2-ylamino]ethanol (4a). A solution of 3 (210 mg, 0.59
mmol) in tetrahydrofuran/acetone/H2O (4:1:1, 9 mL) was trea-
ted with ethanolamine (39 μL, 0.65 mmol) and 5% w/v aqueous
sodium bicarbonate (1.3 mL), and the mixture was stirred at
room temperature overnight. The mixture was treated with
water (5 mL) and extracted with ethyl acetate (2 ꢀ 5 mL).
Combined extracts were washed with saturated sodium chloride
(10 mL), dried over sodium sulfate, and evaporated in vacuo to
give the crude product. Purification on silica gel (elution
with 5-50% ethyl acetate/hexanes) gave 4a as an off-white solid
1
(214 mg, 95%). Mp = 157-159 °C. H NMR (CD3OD, 400
MHz) δ: 8.20 (broad, 1H), 7.76 (broad, 1H), 7.24-7.30 (m, 1H),
7.08-7.22 (m, 3H), 6.96 (broad, 1H), 3.72 (t, J=5.6 Hz, 1H),

1144 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3
Zacharie et al.
Mp=146-149 °C. 1H NMR (CD3OD, 400 MHz) δ: 7.75-7.72
(m, 3H), 6.88-7.32 (m, 9H), 3.68 (broad, 2H), 3.60 (broad, 2H),
3.54 (broad, 2H), 2.84 (t, J=7.2 Hz, 2H), 1.49 (s, 18H). 13C
NMR (CD3OD, 100 MHz) δ: 171.83, 165.94, 164.52, 154.02,
140.50, 139.55, 136.57, 135.22, 128.96, 128.71, 128.50, 120.94,
114.54, 112.85, 110.90, 79.85, 79.57, 61.35, 60.38, 42.85, 42.11,
35.35, 27.59, 27.58, 19.71, 13.30. LRMS (ESI): m/z 800.4
(M(35Cl)Hþ). HPLC: tR = 3.2 min (gradient B, purity 100%
at 254 nm).
2-(4-([3-tert-Butoxycarbonylaminophenyl]amino)-6-(4-(2-(4-
([3-tert-butoxycarbonylaminophenyl]amino)-6-(2-hydroxyethyl-
amino)[1,3,5]triazin-2-ylamino)ethyl)phenylamino)[1,3,5]triazin-2-
ylamino)ethanol (8a). A solution of 7a (46 mg, 0.057 mmol) in
anhydrous tetrahydrofuran (1 mL) was treated with ethanol-
amine (10 μL, 0.17 mmol), and the mixture was stirred at 60 °C
overnight. The suspension was dissolved by addition of metha-
nol and evaporated onto silica gel. Purification on silica gel
(elution with 10-100% ethyl acetate/hexanes) gave 8a as a
white solid (39 mg, 82%). LRMS (ESI): m/z 825.4 (MHþ),
847.4 (MNaþ). HPLC: tR=2.5 min (gradient B, purity 100%
at 254 nm).
extract (NE) antibodies, as well as SLE-related clinical mani-
festations including leukopenia, thrombocytopenia, protei-
nuria, and glomerulonephritis. These mice develop anti-DNA
antibodies after the age of 3 months, with a peak of anti-DNA
antibody response occurring at 7 months. Mice were treated by
intravenous injection of compound 34 or vehicle once a week
from week 12. Serum dsDNA autoantibodies were measured as
described by Marino et al.24
(2) MRL/lpr mice also spontaneously develop SLE-like syn-
drome. MRL/lpr mice have a homozygous fas mutation, which
accelerates autoimmunity. Mice were treated by intravenous
injection of compound 34 or vehicle once a week for 5 con-
secutive weeks. Mice were sacrificed, and immune com-
plex deposition was recorded in kidney tissue by electron
microscopy.
Acknowledgment. The authors thank Mona Saleh for data
management and administrative support, Marc-Alexandre
€
ꢀ
Duceppe, Joel Lechasseur, and Valerie Leduc for valuable
technical support, and Lyne Marcil for the preparation of this
manuscript.
2-(4-(3-Aminophenylamino)-6-(4-(2-(4-(3-aminophenylamino-6-
(2-hydroxyethylamino)[1,3,5]triazin-2-ylamino)ethyl)phenylamino)-
[1,3,5]triazin-2-ylamino)ethanol (10). A solution of 8a (39 mg,
0.047 mmol) in anhydrous dichloromethane (0.79 mL) was
cooled to 0 °C and treated with a 4 M solution of hydrochloric
acid in 1,4-dioxane (2.36 mL, 9.44 mmol). The solution was
stirred for 5.5 h, with slow warming to ambient temperature.
1,2-Dichloroethane (2 mL) was then added, and solvents were
evaporated in vacuo. The residue was triturated with diethyl
ether/hexanes (3:1, 4 mL) with sonication. This process was
repeated two more times, and the solid was dried in vacuo to give
10 as a white solid (41 mg, 99.9%). Mp=199-203 °C. 1H NMR
(CD3OD, 400 MHz) δ: 7.87-8.01 (m, 1H), 7.62-7.77 (m, 2H),
7.48-7.60 (m, 5H), 7.26-7.40 (m, 2H), 7.15-7.20 (m, 2H),
3.67-3.78 (m, 6H), 3.58-3.66 (m, 4H), 2.95-3.03 (m, 2H).
LRMS (ESI): m/z 625.4 (MHþ). HPLC: tR=3.3 min (gradient
A, purity 100% at 254 nm).
Protein A Human IgG Competitive ELISA. This assay evalu-
ates the ability of compounds to mimic protein A by inhibition
of binding to IgG. The competitive protein A binding assay was
performed on a 96-well plate Maxisorp surface to enhance the
binding of protein A to the bottom of the plate. The wells were
coated with 100 μL of protein A (0.8 μg) and incubated over-
night at 4 °C. After incubation, unbound protein A was removed
by three washes with phosphate buffered saline (PBS). The plate
was then incubated with 100 μL/well of a 2% solution of bovine
serum albumin (BSA) for 1 h at 37 °C to block nonspecific
protein binding. After incubation, the plate was washed three
times with PBS. An amount of 50 μL of test compounds or
protein A, diluted in PBS or PBS-20% DMSO at appropriate
concentration, was added to the wells followed by addition of
50 μL of peroxidase-conjugated human IgG (HRP-IgG). After
1 h of incubation at 37 °C, the plate was washed three times with
PBS to remove unbound HRP-IgG. Bound HRP-IgG was
detected by incubation with 100 μL of 2,20-azino-di[3-ethyl-
benzthiazoline sulfonate]diammonium salt (ABTS) solution for
20 min in the dark at room temperature. The plate was then read
at 405 nm on an EL 800 universal Microplate reader (Bio-Tek).
Data were analyzed using Microsoft Excel, and the concentra-
tion of compound that inhibits 50% binding of protein A (IC50)
was calculated using the Prism software.
Supporting Information Available: Characterization data for
compounds 9 and 11-44. This material is available free of
charge via the Internet at http://pubs.acs.org.
References
(1) (a) Langone, J. J. Protein A of Staphylococcus aureus and related
immunoglobulin receptors produced by streptococci and pneumo-
ꢀ
cocci. Adv. Immunol. 1982, 32, 157–252. (b) Nilsson, B.; Abrahmsen,
L. Fusions to staphylococcal protein A. Methods Enzymol. 1990, 185,
ꢀ
144–161. (c) Moks, T.; Abrahmsen, L.; Nilsson, B.; Hellman, U.;
ꢀ
Sjoquist, J.; Uhlen, M. Staphylococcal protein A consists of five IgG-
binding domains. Eur. J. Biochem. 1986, 156, 637–643.
(2) Delano, W.; Ultsch, M.; de Vos, A. M.; Wells, J. A. Convergent
solutions to binding at a protein-protein interface. Science 2000,
287, 1279–1283 and references cited therein.
(3) (a) Li, R.; Dowd, V.; Stewart, D. J.; Burton, J.; Lowe, C. R. Design,
Synthesis, and application of a protein A mimetic. Nat. Biotechnol.
1998, 16, 190–195. (b) Fassina, G.; Ruvo, M.; Palombo, G.; Verdoliva,
A.; Marino, M. Novel ligands for the affinity-chromatographic purifi-
cation of antibodies. J. Biochem. Biophys. Methods 2001, 49, 481–
490.
(4) McCormick, D. Bioseparations look ahead to the past. Pharm.
Technol. 2005, 29, 36–44.
(5) Godfrey, M. A.; Kwasowski, P.; Clift, R.; Marks, V. Assessment of
the suitability of commercially available SpA affinity solid phases
for the purification of murine monoclonal antibodies at process
scale. J. Immunol. Methods 1993, 160, 97–105.
(6) Teng, S.; Sproule, K.; Hussain, A.; Lowe, C. R. A strategy for the
generation of biomimetic ligands for affinity chromatography.
Combinatorial synthesis and biological evaluation of an IgG
binding ligand. J. Mol. Recognit. 1999, 12, 67–75.
(7) (a) Delano, W. L.; Ultsch, M. H.; de Vos, A. M.; Wells, J. A.
Convergent solutions to binding at a protein-protein interface.
Science 2000, 287, 1279–1283. (b) Ehrlich, G. K.; Bailon, P. Identifica-
tion of model peptides as affinity ligands for the purification
of humanized monoclonal antibodies by means of phage display.
J. Biochem. Biophys. Methods 2001, 49, 443–454. (c) Fassina, G.;
Verdoliva, A.; Oderna, M. R.; Ruvo, M.; Cassini, G. Protein A mimetic
peptide ligand for affinity purification of antibodies. J. Mol. Recognit.
1996, 9, 564–569.
(8) Johnson, C. P.; Jensen, I. E.; Prakasam, A.; Vijayendran, R.;
Leckband, D. Engineered Protein A for the orientational control
of immobilized proteins. Bioconjugate Chem. 2003, 14, 974–978.
(9) (a) Sasso, E. H.; Merrill, C.; Furst, D. E. Immunoglobulin binding
properties of the prosorba immunadsorption column in treatment
of rheumatoid arthritis. Ther. Apheresis 2001, 5, 84–91. (b) Brunner,
J.; Kern, P. M.; Gaipl, U. S.; Voll, R. E.; Kalden, J. R.; Wiesenhutter,
C. W.; Hermann, M. The low-throughput protein A adsorber: an
immune modulatory device. Hypothesis for the mechanism of ac-
tion in the treatment of rheumatoid arthritis. Mod. Rheumatol. 2005,
15, 9–18.
Heat Aggregated Human IgG Binding Assay. In vitro binding
assay on human Fcγ receptors was conducted on the basis of an
assay described in the literature.23
SLE Models. (1) New Zealand mice of the F1 hybrid cross
NZB/NZW develop most of the autoimmune abnormalities
seen in human SLE and die from SLE-like immune complex
(IC)-mediated glomerulonephritis. The mice develop high titers
of anti-DNA (double-strand and single-strand) and nuclear
(10) Silverman, G.; Goodyear, C. S.; Siegel, D. L. On the mechanism of
staphylococcal protein A immunomodulation. Transfusion 2005,
45, 274–280.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1145
(11) Zacharie, B.; Fortin, F.; Wilb, N.; Bienvenu, J.-F.; Asselin, M.;
Grouix, B.; Penney, C. 2,6,9-Trisubstituted purine derivatives as
protein A mimetics for the treatment of autoimmune diseases.
Bioorg. Med. Chem. 2009, 19, 242–246.
(12) Blotny, G. Recent applications of 2,4,6-trichloro-1,3,5-triazine and
its derivatives in organic synthesis. Tetrahedron 2006, 62, 9507–
9522 and references cited therein.
(13) Baliani, A.; Bueno, G. J.; Stewart, M. L.; Yardlev, V.; Brun, R.;
Barrett, M. P.; Gilbert, I. H. Design and synthesis of a series of
melamine-based nitroheterocycles with activity against Trypano-
somatid parasites. J. Med. Chem. 2005, 48, 5570–5579.
Bhanuprakash, K.; Harakishore, K.; Murthy, U. S. N.; Rao, V. J.
Synthesis and antibacterial activity of various substituted triazines.
Eur. J. Med. Chem. 2006, 41, 1240–1246.
(18) Penney, C. Zacharie, B.; Abbott, S. D.; Bienvenu, J.-F.; Cameron,
A. D.; Duceppe, J.-S.; Ezzitouni, A.; Fortin, D.; Houde, K.;
Moreau, N.; Wilb, N.; Grouix, B.; Gagnon, L. Triazines Dimmers
for the Treatment of Autoimmune Diseases. WO 2005/049607,
2005.
(19) Unpublished results.
(20) (a) The active site was defined by the collection residues of IgG Fc
˚
within 5 A of bound protein A (PDB code 1FC2). MOE was used to
(14) (a) Menicagli, R.; Samaritani, S.; Signore, G.; Vaglini, F.; Via,
L. D. In vitro cytotoxic activities of 2-alkyl-4,6-diheteroalkyl-1,3,5-
triazines: new molecules in anticancer research. J. Med. Chem.
2004, 47, 4649–4652. (b) Kawashima, S.; Matsuno, T.; Yaguchi, S.;
Sasahara, H.; Watanabe, T. Heterocyclic Compound and Antitumor
Agent Containing the Same as Active Ingredient U.S. Patent 7,071,189,
2006. (c) Matsuno, T.; Kato, M.; Sasahara, H.; Watanabe, T.; Inaba, M.;
Takahashi, M.; Yaguchi, S.-I.; Yoshioka, K.; Sakato, M.; Kawashima, S.
Synthesis and antitumor activity of benzimidazolyl-1,3,5-triazine and
benzimidazolylpyrimidine derivatives. Chem. Pharm. Bull. 2000, 48,
1778–1781. (d) Moon, H.-S.; Jacobson, E. M.; Khersonsky, S. M.;
Luzung, M. R.; Walsh, D. P.; Xiong, W.; Lee, J. W.; Parikh, P. B.; Lam,
J. C.; Kang, T.-W.; Rosania, G. R.; Schier, A. F.; Chang, Y.-T. A novel
microtubule destabilizing entity from orthogonal synthesis of triazine
library and zebrafish embryo screening. J. Am. Chem. Soc. 2002, 124,
11608–11609. (e) Arya, K.; Dandia, A. Synthesis and cytotoxic activity
of trisubstituted-1,3,5-triazines. Bioorg. Med. Chem. Lett. 2007, 17,
3298–3304.
(15) Melato, S.; Prosperi, D.; Coghi, P.; Basilico, B.; Monti, D. A
combinatorial approach to 2,4,6-trisubstituted triazines with po-
tent antimalarial activity: combining conventional synthesis and
microwave-assistance. ChemMedChem 2008, 3, 873–876 and refer-
ences cited therein.
(16) Xiong, Y.-Z.; Chen, F.-E.; Balzarini, J.; De clercq, E.; Pannecouque,
C. Non-nucleoside HIV-1 reverse transcriptase inhibitors. Part 11:
structural modulations of diaryltriazines with potent anti-HIV
activity. Eur. J. Med. Chem. 2008, 43, 1230–1236 and references cited
therein.
generate several potential docking modes, and a low energy one
was chosen. Molecular dynamics (500 K; heat, 100 ps; run, 200 ps;
cool, 100 ps) was used to explore the region, and the residues
defining the active site were also allowed to move. The structures
above show the final bound complex. It is likely that there are
multiple binding modes. (b) The binding site was specified to be the
same as for protein A: Deisenhofer, J. Crystallographic refinement
and atomic models of a human Fc fragment and its complex with
˚
fragment-b of protein A from Staphylococcus-aureus at 2.9 A and
˚
2.8 A resolution. Biochemistry 1981, 20 (9), 2361–2370.
(21) Werwitzke, S.; Trick, D.; Sondermann, P.; Kamino, K.;
Schlegelberger, B.; Kniesch, K.; Tiede, A.; Jacob, U.; Schmidt,
R. E.; Witte, T. Treatment of lupus-probe NZB/NZW F1 mice with
recombinant soluble Fcγ receptor II (CD32). Ann. Rheum. Dis.
2008, 67, 154–161.
(22) Passwell, J.; Schreiner, G. F.; Nonaka, M.; Beuscher, H. U.;
Colten, H. R. Local extrahepatic expression of complement genes
C3, factor B, C2, and C4 is increased in murine lupus nephritis.
J. Clin. Invest. 1988, 82, 1676–1684.
(23) (a) Shields, R. L.; Namenuk, A. K.; Hong, K.; Meng, Y. G.; Rae, J.;
Briggs, J.; Xie, D.; Lai, J.; Stadlen, A.; Li, B.; Fox, J. A.; Presta,
L. G. High resolution mapping of the binding site on human IgG1
for FcγRI, FcγRII, FcγRIII, and FcRn and design of IgG1
variants with improved binding to the FcγR. J. Biol. Chem. 2001,
276, 6591–6604. (b) Wines, B. D.; Gavin, A.; Powell, M. S.; Steinitz,
M.; Buchanan, R. R. C.; Hogarth, P. M. Soluble FcγRIIa inhibits
rheumatoid factor binding to immune complexes. Immunology 2003,
109, 246–254.
(17) (a) Zhou, C.; Min, J.; Liu, Z.; Young, Z.; Deshazer, H.; Gao, T.;
Chang, Y.-T.; Kallenbach, R. Synthesis and biological evaluation
of novel 1,3,5-triazine derivatives as antimicrobial agents. Bioorg.
Med. Chem. Lett. 2008, 18, 1308–1311. (b) Srinivas, K.; Srinivas, U.;
(24) Marino, M.; Ruvo, M.; De Falco, S.; Fassina, G. Prevention of
systemic lupus erythematosus in MRL/lpr mice by administration
of an immunoglobulin-binding peptide. Nat. Biotechnol. 2000, 18,
735–739.