C-4 modified sialosides enhance binding to Siglec-2 (CD22): Towards potent Siglec inhibitors for immunoglycotherapy

Article abstract of DOI:10.1002/anie.201207267

Two changes for the better: A novel class of sialic acid derivatives is prepared by modifying both the C-4 and C-9 positions of Neu5Aα2Me (see structure). This approach gives a lead compound that has sub-micromolar affinity for Siglec-2 and may provide a pathway for immunoglycotherapy strategies for autoimmune diseases and B cell derived non-Hodgkin's lymphoma. Copyright

Full text of DOI:10.1002/anie.201207267

.
Angewandte
Communications
DOI: 10.1002/anie.201207267
Immunoglycotherapy
C-4 Modified Sialosides Enhance Binding to Siglec-2 (CD22): Towards
Potent Siglec Inhibitors for Immunoglycotherapy**
Sørge Kelm,* Paul Madge, Tasneem Islam, Ryan Bennett, Hendrik Koliwer-Brandl,
Mario Waespy, Mark von Itzstein, and Thomas Haselhorst*
The regulatory functions of Siglecs (sialic acid binding
immunoglobulin-like lectins) in the immune system provide
opportunities for innovative therapeutic strategies for a wide
range of immunological disorders or cancer (immunoglyco-
therapy).[1] Siglec-2 (CD22), as a consequence of its pivotal
role in B cell activation, has become an attractive target for
therapies of autoimmune diseases and B cell-derived non-
Hodgkinꢀs lymphoma (NHL). NHL is among the ten most
common cancers with over 20000 deaths in 2010 for the US
alone.[2] Siglec-2 binds with high preference to a(2,6)-linked
sialic acids (Sia),[3] such as Neu5Aca(2,6)lactosamine
(Scheme 1). Neu5Aca2Me (1) interacts with Siglec-2 mainly
through 1) the negative charge on its carboxylate group,
2) the C-5 N-acetamido substituent, and 3) the glycerol side
chain. Furthermore, replacement of the C-9 hydroxy group by
an amino group did not interfere with binding to Siglecs.[3]
Crystallographic studies on Siglec-1 (sialoadhesin, Sn)[4]
Scheme 1. N-acetylneuraminic acid derivatives (R=NHAc).
demonstrated that acylation of this amino group enhances
the overall affinity of the ligand for Siglecs by two to three
orders of magnitude.[5] The first breakthrough in the develop-
ment of potent Siglec-2 inhibitors was the design of 9-
biphenylcarboxamido Neu5Aca2Me (9-BPC-Neu5Aca2Me,
2) which has a more than two orders of magnitude higher
affinity to Siglec-2 than 1,[5d] and 2 has demonstrated potential
to modulate signal transduction in B cells. Furthermore,
based on 2, compounds were developed, which kill B cell
lymphoma cells.[6] Structural studies[4a,b] and modifications of
the C-5 N-acyl substituent and the C-2 aglycon moiety of N-
acetylneuraminic acid (Neu5Ac) have led to further improve-
ment in affinity.[5a–c,e,f]
Herein we report, for the first time the design, synthesis,
and evaluation of a novel class of disubstituted Neu5Ac
derivatives that is modified at the C-4 and C-9 positions of 1.
Our structure-based design approach resulted in a promising
novel lead compound 9-biphenylcarboxamido-4-m-nitrophe-
nylcarboxamido-4,9-dideoxy
Neu5Aca2Me
(9-BPC-4-
mNPC-Neu5Aca2Me, 6b) that has sub-micromolar affinity
for Siglec-2 and may provide a pathway for immunoglyco-
therapy strategies.
An evaluation of our homology model (see Supporting
Information) for Siglec-2 and other Siglecs led us to
hypothesize that substituents at C-4 may provide additional
interactions. To address this hypothesis we posed the follow-
ing questions:
[*] Prof. S. Kelm, Dr. H. Koliwer-Brandl, Dipl.-Chem. M. Waespy
Centre for Biomolecular Interactions Bremen, Department of
Biology and Chemistry, University of Bremen
28334 Bremen (Germany)
E-mail: skelm@uni-bremen.de
1) Can C-4 substituents enhance the interaction with Siglecs?
2) Do they interact specifically with the protein?
3) Do C-4 and C-9 modifications act synergistically?
4) Do the C-4 modified Neu5Ac derivatives bind to the same
binding site as other Sia, such as 1?
Homepage: http://www.cbib.uni-bremen.de
Prof. S. Kelm, P. Madge, Dr. T. Islam, R. Bennett,
Prof. M. von Itzstein, Dr. T. Haselhorst
Institute for Glycomics, Gold Coast Campus, Griffith University
Queensland, 4222 (Australia)
E-mail: t.haselhorst@griffith.edu.au
Homepage: http://www.griffith.edu.au/glycomics
To see if C-4-modified derivatives of Neu5Aca2Me (1)
enhance the interaction with Siglec-2 and other siglecs, novel
C-4 functionalized compounds of 1 (4b–4g) were prepared.
Thus, per-O-acetylated 4-amino-4-deoxy-Neu5Aca2Me (3)
was readily synthesized using a literature method,[7] and
subsequent treatment of 3 with acyl chlorides or sulfonyl
chloride followed by deprotection provided 4b–4g (Table 1,
Supporting Information).
[**] We gratefully acknowledge the financial support of the Volkswa-
genstiftung (S.K.), Deutsche Forschungsgemeinschaft (S.K.),
Tçnjes-Vagt-Stiftung (S.K.), and the Australian Research Council
(T.H., M.v.I.). S.K. is thankful for the Sir Allan Sewell Fellowship
awarded by Griffith University. We thank Mary Murphy (Reichert Inc,
Depew, NY (USA)) for the SPR analysis.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201207267.
3616
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 3616 –3620

Angewandte
Chemie
binding to Siglec-2, suggesting that the C-4 substituents in 4d
and 4e interact specifically with amino acids positioned
adjacent to the Sia binding site. The striking differences in
inhibition potencies prompted us to evaluate the interaction
of 4d with Siglec-2, Siglec-4, and Siglec-9 by saturation
transfer difference (STD) NMR spectroscopy (Figure 1).[8]
STD NMR signals of 4d were normalized to the methyl
protons of the N-acetamido group. The relative signal
intensity for Ho proton of the mNPC moiety is almost
double in Siglec-9:4d compared to that in Siglec-2:4d and
three times that in the Siglec-4:4d complex. These results are
in full agreement with the hypothesis that the C-4 mNPC
functionality in 4d significantly contributes to the com-
poundꢀs inhibitory potential observed in the inhibition assays.
The protons of the O-methyl group of 4d in Siglec-2:4d
receives 80% saturation, whereas the same protons receive
only 36% and 16% in Siglec-4:4d and Siglec-9:4d, respec-
tively.
Table 1: Inhibition of Siglec-4 (MAG), Siglec-2 (CD22), and Siglec-9 by C-
4-modified Sia.
Siglec-2
(CD22)
Siglec-4
(MAG)
Siglec-9
[a]
R’
rIC50
IC50
[mm][b]
1
OH
1.0
0.6
1.0
1.7
n.i.[c]
n.i.[c]
4a NH2
4b
2.2
0.2
n.i.[c]
A number of bifunctionalized Neu5Aca2Me (1) deriva-
tives with modifications at C-4 and C-9 were synthesized to
investigate whether such modifications act synergistically.
Initial protection of the amino group on C-4 of 3 with
butoxycarbonyl (Boc) was undertaken to minimize the
number of reaction steps. Thus, introduction of a C-9 azido
group provided 5,[9] which, after reduction to the C-9 amino
4c
1.7
0.6
n.i.[c]
4d
4e
15
10
0.4
0.5
6.5
n.i.[d]
4 f
0.5
0.7
n.i.[d]
n.i.[d]
n.i.[d]
n.i.[d]
precursor, enabled easy installation of a BPC group. Removal
of the C-4 Boc protecting group enabled coupling of
substituents to the free amine, providing the C-4/C-9 modified
compounds (6a–d; Supporting Information).
4g
[a] rIC50 values were calculated with 1 (IC50 2.0 mm) as reference
compound from IC50 values determined in at least three titrations,
standard deviations were within 15%. [b] rIC50 values could not be
calculated, since 1 was not inhibitory at up to 20 mm. [c] Not inhibitory at
20 mm. [d] Not inhibitory at 2 mm.
Table 2 outlines the inhibition potencies of compounds
6a–d in hapten inhibition assays showing that modifications at
C-4 significantly alter binding to Siglec-2. A naphthylamido
substituent (6a) enhances the binding to Siglec-2 approxi-
mately fourfold compared to 2 but the dibenzylamino
derivative (6d) decreases the binding affinity to Siglec-2 by
a factor of 13. The most significant increase in binding affinity
was observed for 4-mNPC (6b; rIC50 = 14) or 4-toluenesul-
phonylamido (6c; rIC50 = 20) derivatives. This is in excellent
agreement with the rIC50 values determined for the Sia
derivatives without the C-9 BPC moiety (Table 1) that show
the mNPC moiety at C-4 (4d) increases the binding affinity by
a factor of 15. Most striking is the synergistic effect of the C-4
and C-9 modifications resulting in an inhibition 9100 times
stronger for 6b compared to 1.
Absolute binding affinities were also determined using
surface plasmon resonance (SPR) measurements. Dissocia-
tion constants (KD) were obtained for the key compounds
1 (KD = 31 mm), 2(KD=7 mm), and 6b (KD = 660 nm; see
Supporting Information). STD NMR competition experi-
ments were undertaken to explore whether C-4 and/or C-9 Sia
derivatives bind to the same Siglec-2 Sia binding site as 1 and
2. Compound 4d was added incrementally to a mixture of
Siglec-2 (CD22), Siglec-4 (myelin-associated glycopro-
tein, MAG), and Siglec-9, were chosen to evaluate these
compounds as potential inhibitors in hapten inhibition assays
using immobilized fetuin as the target glycoprotein (Table 1).
Compounds 1 and 4a[7c] that contain a hydroxy or an amino
group, respectively, at the C-4 position bind only weakly to all
three Siglecs (IC50 ꢀ 20 mm or more).
An N-acetamido moiety at C-4 (compound 4b) led to an
increase in affinity for Siglec-2 but did not improve the
affinity for Siglec-9 and even decreases binding to Siglec-4.
The largest increase in binding affinity for Siglec-2 was
observed when
a meta-nitrophenylcarboxamido moiety
(mNPC) was installed to give 4d (IC50 = 0.13 mm; rIC50 =
15). This result is in contrast to the moderate affinity for
Siglec-9 (IC50 = 6.5 mm) or Siglec-4 (IC50 = 6.3 mm; rIC50 =
0.4). In addition the results show that compounds 4d and 4e
are specific inhibitors of Siglec-2. The presence of an aromatic
residue (4c, 4 f, or 4g) was not sufficient to enhance the
Angew. Chem. Int. Ed. 2013, 52, 3616 –3620
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3617

.
Angewandte
Communications
Table 2: rIC50 values of C-4/C-9-modified Sia for Siglec-2 (CD22).
[a]
Compound
R’
rIC50
2
1
2
OH
1.0
667
6a
3.8
2500
6b
6c
14
20
9100
13000
6d
0.1
50
[a] rIC50 values were calculated with 2 (IC50 =3.0 mm) or 1 (IC50 =2.0 mm)
as the reference compound from IC50 values determined in at least three
titrations, standard deviations were within 15%.
Figure 1. a) 1H NMR spectrum of 4d and STD NMR spectra of
500 mm 4d in complex with b) 5 mm Siglec-2, c) Siglec-4, and d) Siglec-
9.
5 mm Siglec-2 and 500 mm 1 (Figure 2a). STD NMR signal
intensities for the methyl protons of the N-acetamido group of
1 reveal that a concentration of 200 mm of 4d was sufficient to
reduce the STD NMR signals of 1 by about 50%. At an
equimolar ligand concentration, the STD NMR effects of
1 almost disappeared (Figure 2a). This result is in good
agreement with a rIC50 value for 4d of 15 times that of
1 (Table 1).
Figure 2. Competition STD NMR experiments of 5 mm Siglec-2 in
complex with a) 500 mm 1 and an increasing concentration of 4d,
b) 500 mm 2 and an increasing concentration of 6b. The H NMR
1
spectra of 1 and 2 are shown at the bottom of the respective panels.
A second competition experiment was carried out using
5 mm of Siglec-2 and 500 mm of 2, with a gradual increase in the
concentration of 6b. A very low 6b concentration (5 mm) was
sufficient to reduce the STD NMR signal of 2 by over 50%
(Figure 2b). At a concentration of 10 mm of 6b the STD NMR
effects of 2 were decreased that they could hardly be detected,
clearly demonstrating the very high affinity of 6b for Siglec-2.
This affinity of 6b for Siglec-2 became even more clear in
a competition experiment with 1 (data not shown). No
binding of 1 (500 mm) to Siglec-2 was detected at a very low
concentration (5 mm) of 6b, suggesting that at a concentration
equimolar to the protein, 6b occupies essentially all the
available Siglec-2 Sia binding sites.
The binding epitope of 6b when in complex with Siglec-2
was determined by STD NMR spectroscopy (Figure 3). The
STD NMR spectrum showed very strong STD NMR signals
for the mNPC substituent at C-4. The HoA and HmA protons
revealed a STD NMR effect of 181% and 192%, respectively,
suggesting a close contact to the protein. Interestingly, these
STD effects are stronger than those determined for 4d
3618
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 3616 –3620

Angewandte
Chemie
tained, the BPC substituent is located in a pocket. As targeted
in our compound design, the mNPC substituent is in close
contact with the loop between the F- and G-strands.
In summary, we have successfully developed novel lead
structures with sub-micromlolar affinities against Siglec-2.
These structures have been functionalized at the C-4 of
1 leading to enhanced interaction with Siglec-2 by at least one
order of magnitude as in compound 4d. Furthermore, these
substituents act synergistically with those at C-9 as in
compound 6b. We further conclude that 4d is likely to bind
to the same Sia binding site as 2 and 1, strongly supporting our
initial hypothesis that modifications at C-4 and C-9 act
synergistically and that the addition of suitable C-4 substitu-
ents significantly increases the binding to Siglec-2. Based on
published results[5,9] for Siglec-4 and Siglec-2, it can be
expected that the addition of suitable moieties at C-2 will
lead to a further increase in affinity as observed for 2. Similar
to antibodies, suitable polymers carrying high-affinity Siglec-2
inhibitors, such as compound 6b, can induce similar effects[5-
b,e,f] and will provide even more biologically active compounds
than those described for 2.[6b,10] Thus, our approach provides
new lead structures for the design of next generation Siglec
inhibitors as potential drugs for immunoglycotherapy.
Received: September 8, 2012
Revised: November 26, 2012
Published online: February 25, 2013
Keywords: antitumor agents · immunoglycotherapy ·
.
NMR spectroscopy · sialic acid · siglecs
[1] C. Jandus, H. U. Simon, S. von Gunten, Biochem. Pharmacol.
2011, 82, 323.
[2] A. Jemal, R. Siegel, J. Xu, E. Ward, Ca-Cancer J. Clin. 2010, 60,
277.
[3] S. Kelm, R. Brossmer, R. Isecke, H. J. Gross, K. Strenge, R.
Schauer, Eur. J. Biochem. 1998, 255, 663.
Figure 3. a) 1H NMR and b) STD NMR spectra of 500 mm 6b in
complex with 5 mm Siglec-2; ax=axial, eq=equatorial, Tris=2-amino-
2-hydroxymethylpropane-1,3-diol.
[4] a) N. R. Zaccai, A. P. May, R. C. Robinson, L. D. Burtnick, P. R.
Crocker, R. Brossmer, S. Kelm, E. Y. Jones, J. Mol. Biol. 2007,
365, 1469; b) N. R. Zaccai, K. Maenaka, T. Maenaka, P. R.
Crocker, R. Brossmer, S. Kelm, E. Y. Jones, Structure 2003, 11,
557; c) A. P. May, R. C. Robinson, M. Vinson, P. R. Crocker,
E. Y. Jones, Mol. Cell 1998, 1, 719.
(Figure 1). Similarly, the aromatic rings B and C of the BPC
moiety at C-9 receive a significantly higher degree of
saturation in 6b than in compound 2. This result suggests
that both the 4-mNPC and the 9-BPC moieties have intimate
contact with the protein surface and significantly contribute
to the ligands binding affinity. Overall STD NMR effects for
the Sia ring and glycerol side-chain protons of 6b bound to
Siglec-2 are clearly higher than those for the Sia ring protons
in 1, 2, and 4d bound to Siglec-2. This increase in saturation
can be either a consequence of a tighter contact of the Sia
scaffold of 6b with the protein or a result of the intra-
molecular transfer of saturation received from the aromatic
substituents at C-4 and C-9. Overall the epitope maps support
the hypothesis that the binding modes of our new bifunctional
Sia derivatives are similar, but not necessarily identical, to
that of 1 and 2. This observation is in very good agreement
with the result of a computer-based docking experiment
(Supporting Information). While all critical interactions
between the protein and Sia, through the Sia carboxylate,
its N-acetamido moiety, and glycerol side chain, are main-
[5] a) S. V. Shelke, B. Cutting, X. Jiang, H. Koliwer-Brandl, D. S.
Strasser, O. Schwardt, S. Kelm, B. Ernst, Angew. Chem. 2010,
122, 5857; Angew. Chem. Int. Ed. 2010, 49, 5721; b) S. Mesch, K.
Lemme, M. Wittwer, H. Koliwer-Brandl, O. Schwardt, S. Kelm,
B. Ernst, ChemMedChem 2012, 7, 134; c) S. Mesch, K. Lemme,
H. Koliwer-Brandl, D. S. Strasser, O. Schwardt, S. Kelm, B.
Ernst, Carbohydr. Res. 2010, 345, 1348; d) S. Kelm, J. Gerlach, R.
Brossmer, C. P. Danzer, L. Nitschke, J. Exp. Med. 2002, 195,
1207; e) H. H. Abdu-Allah, T. Tamanaka, J. Yu, L. Zhuoyuan,
M. Sadagopan, T. Adachi, T. Tsubata, S. Kelm, H. Ishida, M.
Kiso, J. Med. Chem. 2008, 51, 6665; f) H. H. Abdu-Allah, K.
Watanabe, G. C. Completo, M. Sadagopan, K. Hayashizaki, C.
Takaku, T. Tamanaka, H. Takematsu, Y. Kozutsumi, J. C.
Paulson, T. Tsubata, H. Ando, H. Ishida, M. Kiso, Bioorg.
Med. Chem. 2011, 19, 1966.
[6] a) W. C. Chen, D. S. Sigal, A. Saven, J. C. Paulson, Leuk.
Lymphoma 2012, 53, 208; b) W. C. Chen, G. C. Completo, D. S.
Sigal, P. R. Crocker, A. Saven, J. C. Paulson, Blood 2010, 115,
4778.
Angew. Chem. Int. Ed. 2013, 52, 3616 –3620
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3619

.
Angewandte
Communications
[7] a) E. Schreiner, E. Zbiral, R. G. Kleineindam, R. Schauer,
Liebigs Ann. Chem. 1991, 129, 0; b) M. von Itzstein, B. Jin, W. Y.
Wu, M. Chandler, Carbohydr. Res. 1993, 244, 181; c) S. Cicco-
toso, M. von Itzstein, Tetrahedron Lett. 1995, 36, 5405.
[8] a) B. Meyer, T. Peters, Angew. Chem. 2003, 115, 890; Angew.
Chem. Int. Ed. 2003, 42, 864; b) M. Mayer, B. Meyer, Angew.
Chem. 1999, 111, 1902; Angew. Chem. Int. Ed. 1999, 38, 1784;
c) T. Haselhorst, F. E. Fleming, J. C. Dyason, R. D. Hartnell, X.
Yu, G. Holloway, K. Santegoets, M. J. Kiefel, H. Blanchard, B. S.
Coulson, M. von Itzstein, Nat. Chem. Biol. 2009, 5, 91.
[9] S. V. Shelke, G. P. Gao, S. Mesch, H. Gꢁthje, S. Kelm, O.
Schwardt, B. Ernst, Bioorg. Med. Chem. 2007, 15, 4951.
[10] L. Cui, P. I. Kitov, G. C. Completo, J. C. Paulson, D. R. Bundle,
Bioconjugate Chem. 2011, 22, 546.
3620 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 3616 –3620