Synthesis of Dendritic Polyglycerol Anions and Their Efficiency

ARTICLE pubs.acs.org/Biomac

Synthesis of Dendritic Polyglycerol Anions and Their Efficiency Toward L-Selectin Inhibition Marie Weinhart,§,† Dominic Gr€oger,§,† Sven Enders,‡ Jens Dernedde,‡ and Rainer Haag*,† † ‡

Institute of Chemistry and Biochemistry, Freie Universitaet Berlin, Takustr. 3, 14195 Berlin, Germany Zentralinstitut f€ur Laboratoriumsmedizin und Pathobiochemie, Charite-Universitaetsmedizin Berlin, CBF, Hindenburgdamm 30, 12203 Berlin, Germany

bS Supporting Information ABSTRACT: A versatile route for the synthesis of highly functionalized, polyanionic macromolecules based on dendritic polyglycerol was applied by means of the Huisgen
Sharpless
Meldal 1,3-dipolar cycloaddition (“click-reaction”) of polyglycerolazide precursors and alkyne-functionalized anions such as sulfonates, carboxylates, phosphonates, and bisphosphonates. In addition, the corresponding polyglycerol phosphate has been synthesized via direct hydroxyl interconversion of polyglycerol to the corresponding phosphate with a degree of functionalization >80% by analogy to the synthesis of previously reported polyglycerol sulfates (dPGS). On the basis of the finding that dPGS exhibits high affinity for L- and P-selectin, the potential of these novel polyanionic, multivalent macromolecules of varying anionic nature as L-selectin inhibitors has been evaluated in vitro by means of a competitive concentration dependent binding assay. Affinity of all polyanions toward L-selectin was demonstrated with distinct IC50 values ranging from the low nanomolar to the high micromolar range. The efficiency of L-selectin inhibition increases in the order carboxylate < phosphate < phosphonate ≈ sulfonate < bisphosphonate < sulfate. Additional DLS and ζ-potential measurements of these polyanions were performed to correlate their binding affinity toward L-selectin with their anionic nature. However, a direct correlation of effective charge and particle size with the determined IC50 values turned out to require further in-depth studies on the microstructure of the polyanions but clearly indicate an exceptional position of dPGS among the studied dendritic polyelectrolytes.

’ INTRODUCTION Acute inflammation is the common immune response of the body to potentially harmful stimuli such as microbial infection, toxins, injury, or foreign bodies to eliminate and inhibit the propagation of pathogens as well as to initiate the healing process of damaged tissue.1 An essential step within this immune response is the recruitment of leukocytes from the bloodstream to the vessel walls and further to the inflamed tissue. This recruitment proceeds in a cascade-like fashion including initial capture, decelerated rolling, and firm adhesion of leukocytes to the endothelium of the blood vessel, which finally results in transmigration of leukocytes through the endothelium into the tissue.2
4 Within this complex cascade, a distinct set of cell adhesion molecules (CAMs)5 expressed on both endothelial and leukocytic cells as well as soluble signaling factors like cytokines and in particular chemokines6 have been identified to operate well-orchestrated in promoting leukocyte extravasation. Among the group of CAMs are carbohydrate binding L-, P-, and E-selectin and their corresponding fucosylated and sialylated glycoprotein ligands.7 Nowadays, a variety of naturally occurring selectin specific ligands has been identified with P-selectin glycoprotein ligand 1 (PSGL-1) as the one that exhibits affinity to all three selectins.7 However, only a minority of the natural r 2011 American Chemical Society

ligands has been completely characterized on the molecular level, but it is generally suggested that they all bear the structural motif of the tetrasaccharide sialyl Lewis X (SiaLex).3 Additional sulfation of the carbohydrate moiety at distinct positions has been shown to enhance binding toward L- and P-selectins but not E-selectin.8 This is also true for the sulfation of tyrosine residues within the ligands' peptide sequence, as shown for PSGL-1.9 Chronic inflammation may result from the persistence of an antigen and leads to a dysfunction of the immune system. In the case of typical autoimmune diseases such as psoriasis, rheumatoid arthritis, or multiple sclerosis, there is a sustained proinflammatory reaction. In such settings, the ongoing extravasation of leukocytes causes tissue damage while resolution of inflammation is not achieved. Because of their crucial role in initiating the adhesion cascade, selectins have become a promising target for anti-inflammatory therapy.10
12 However, because selectins are essential for the innate immune response, a permanent suppression of the immune system by selectin inhibitors is of course not a proper therapy in the long term but definitively a Received: February 23, 2011 Revised: April 23, 2011 Published: May 20, 2011 2502

dx.doi.org/10.1021/bm200250f | Biomacromolecules 2011, 12, 2502–2511

Biomacromolecules valuable target for short-term applications, in particular, in ischemiareperfusion-type situations.13,14 In addition, selectin-targeted diagnostics are promising, especially because inflammation gains increased awareness as a concomitant event in many diseases such as cancer and selectin-mediated tumor metastasis.15,16 Dendritic polyglycerol sulfate (dPGS), a polyanionic polymer based on a polydisperse, dendritic polyglycerol scaffold with the outer hydroxyl groups converted to sulfate groups, was initially described as a heparin analog in 2004 with respect to its anticoagulant and anticomplementary activity in vitro.17 Further in vitro studies revealed this fully synthetic heparin mimetic to be a potent inhibitor of L- and P-selectin with IC50 values in the nanomolar range, which exerts its anti-inflammatory activity also in vivo, as demonstrated by reduced leukocyte extravasation in an acute dermatitis model.18 In contrast, synthetic carbohydrate (SiaLex)-based selectin inhibitors, which usually show IC50 values in the millimolar range, revealed low affinity and unfavorable pharmacokinetics once applied in vivo.7 This lack of efficacy in vivo further demonstrates the complexity and cooperativity of the leukocyte recruitment. Sulfation in various positions of carbohydrate-based inhibitors as well as the presence of other anionic groups such as carboxylates, phosphates, and phosphonates have often shown an enhanced binding to selectins in vitro, as indicated by lowered IC50 values compared with the nonionic analogues.3 In addition, it was demonstrated that the inhibitor efficiency increases upon a multiple presentation of the actual binding site of the ligand in a polymeric manner because the weak protein
carbohydrate interaction is not only multiplied but also increases exponentially because of the multivalency effect,19 unfortunately often at the cost of specificity.20
22 However, such SiaLex-based inhibitors or mimetics thereof are tedious to synthesize, whereas the natural anti-inflammatory compound heparin suffers from its animal origin and the related risk of disease transmission to humans as well as from its strong anticoagulant property. Therefore, lots of effort in the field of combinatorial chemistry has been made to identify new alternatives for carbohydrate based selectin inhibitors on one hand. As a result, for instance, efomycine, a noncarbohydrate-based low-molecular-weight SiaLex structure mimetic, has been identified.23 On the other hand, a more detailed knowledge of the actual binding mechanism of selectins and their respective ligands could help us to conclude rational design guidelines for new inhibitory lead structures. This includes the identification and isolation of specific biological ligands and more importantly the characterization of such ligands on the molecular level, which is challenging to both biologists and chemists. Nowadays, it is widely accepted that both L- and P-selectin exhibit an area of cationic surface potential24 in addition to the carbohydrate binding site, which can be independently addressed by anionic/acidic carbohydrate-based molecules such as sulfatide,25,26 dextran sulfate,27 heparin,28 fucoidin,13,14,27 or phosphomannon29 via ionic interaction. The cocrystal structure of P-selectin with its ligand, the PSGL-1 peptide, clearly shows that besides the SiaLex epitope, further interactions of sulfates derived from tyrosine residues of the peptide contribute to high affinity binding by contacting basic residues (His114 and Arg85) of P-selectin.30 Because it was further demonstrated by means of dPGS that carbohydrates are not necessarily required to gain an effective binding of synthetic ligands to selectins, as long as the assumed binding motif is presented multivalently, we herein want to focus on the nature of the anionic species of the ligand.18 Therefore, we chose a dendritic polyglycerol (dPG)31 scaffold

ARTICLE

because of its excellent biocompatibility, as demonstrated in vitro as well as in vivo.31 Its multiple hydroxyl groups on the surface of the polymer can be converted to the desired anionic groups to gain a multivalent binding site and well comparable structures. We herein present the synthesis and characterization of different polydisperse, dendritic polyglycerol anions that were obtained by a versatile click chemistry approach in the case of carboxylates (dPGC), sulfonates (dPGSn), phosphonates (dPGPn), and bisphosphonates (dPGBP). In addition, dendritic polyglycerol phosphate (dPGP) was synthesized by direct hydroxyl interconversion to the phosphate group by analogy to the synthesis of dPGS. These polyanionic multivalent structures were evaluated in vitro with respect to their efficiency as L-selectin inhibitors in a surface plasmon resonance (SPR)-based competitive binding assay. The determined IC50 values were compared with the avidity of previously reported dPGS.18 Dynamic light scattering (DLS) and ζ potential measurements of these polyanions were performed to correlate their size and overall effective charge with their binding efficacy.

’ MATERIALS AND METHODS Materials. All chemicals were reagent grade and used as received from Acros Organics (Geel, Belgium) or Sigma-Aldrich (Schnelldorf, Germany) unless stated otherwise. Milli-Q water was prepared using a Millipore water purification system with a minimum resistivity of 18.0 MΩcm. Dialysis was performed in benzoylated cellulose dialysis tubings (Sigma Aldrich, MWCO 1000 g mol
1) in a 1 L beaker, changing the solvent at least three times over a period of 48 h. Ultrafiltration was performed in solvent-resistant stirred cells from Millipore (Billerica, MA) with Ultracel regenerated cellulose membranes (MWCO 1000 g mol
1). 1 H NMR, 13C NMR, and 31P NMR spectra were recorded on a Jeol ECX 400 operating at 400, 101, and 162 MHz, respectively, or on a Bruker Biospin operating at 700 and 176 MHz, respectively, using the deuterated solvent peak as the internal standard. IR spectra were recorded with a Nicolet AVATAR 320 FT-IR 5 SXC equipped with a DTGS detector from 4000 to 650 cm
1 and evaluated with the software EZ OMNIC ESP. The intensities of signals are assigned as strong (s), medium (m), and weak (w). Synthesis. Dendritic polyglycerol (dPG) was synthesized according to literature with a molecular weight of Mn = 3000 and 6000 g 3 mol
1, respectively, a PDI < 1.8, and a degree of branching of ∼60% corresponding to 13.5 mmol OH groups per gram polyglycerol.32 Hence, dPG (3000 g 3 mol
1) bears ∼40 hydroxyl groups and dPG (6000 g 3 mol
1) 80 hydroxyl groups per molecule. Dendritic polyglycerol sulfate (P11, P12 dPGS) was prepared as published previously by sulfation of dPG with SO3 3 pyridine complex.17 Synthesis of ionic alkynes 1, 2, and 4 was performed as described in the Supporting Information, whereas 4-pentynoic acid (3) is commercially available. Azide functionalized dendritic polyglycerol (dPG) was prepared according to literature via mesylation and subsequent azidation of the hydroxyl groups of dPG.33 1H NMR spectroscopy revealed the conversion of the dPG hydroxyl groups to the corresponding mesylates to be 95% by integration of the mesyl protons versus the dPG backbone, whereas subsequent azidation proceeded with quantitative conversion of the mesylate groups. Azide-functionalized dPG with a degree of functionalization of 95% corresponds to ∼38 azide groups per dPG (3000 g 3 mol
1) and 77 azide groups per dPG (6000 g 3 mol
1), respectively. Molecular weights of all further functionalized dPG derivatives were calculated from the respective conversion, as determined by 1H NMR spectroscopy or from the sulfur content by combustional analysis in the case of dPGS. General Procedure for the Click-Coupling of Ionic Alkynes to Polyglycerol Azide. Dendritic polyglycerol azide (7900 g 3 mol
1, 2503

dx.doi.org/10.1021/bm200250f |Biomacromolecules 2011, 12, 2502–2511

Biomacromolecules 77 N3-groups per molecule or 3950 g 3 mol
1, 38 N3-groups per molecule) was dissolved in tetrahydrofuran (THF), whereas the ionic alkyne (1.1 eq/N3-group) was dissolved in an equal volume of water. Copper sulfate pentahydrate (CuSO4 3 5 H2O) (5 mol % of the alkyne) and sodium ascorbate (25 mol % of the alkyne) were added to the aqueous phase. The resulting yellow aqueous solution was mixed with the organic phase, and the pH of the resulting solution was adjusted to pH 8 with aqueous NaOH. The mixture was degassed by means of repeated freeze
pump
thaw cycles and was stirred afterward for 48 h at room temperature under exclusion of air. Subsequently, the solvent was removed and the crude product was subjected to ultrafiltration in water for 48 h to separate from low-molecular-weight compounds. For effective removal of residual copper ions, a saturated solution of EDTA in water was added during the first three runs of ultrafiltration. Freeze-drying yielded the respective polyelectrolytes in their sodium salt form as colorless salts. All click reactions were performed on a 0.1 to 1 g scale. Dendritic Polyglycerol Sulfonate Sodium Salt (P1, P2, dPGSn). Dendritic polyglycerol sulfonate sodium salt (P1, P2, dPGSn) was synthesized according to the general procedure applying 1 as the ionic alkyne. P1: Mn (dPG core) = 3000 g 3 mol
1, Mn (dPGSn) = 9150 g 3 mol
1, dF= 89%, yield: 91%. P2: Mn (dPG core) = 6000 g 3 mol
1, Mn (dPGSn) = 18 400 g 3 mol
1, dF = 91%, yield: 93%. 1H NMR (700 MHz, D2O, δ): 8.46
7.58 (m, 1 H, NCHdC); 5.50
4.28 (m, 3 H, CHtriazole, CH2-triazole); 4.27
4.11 (m, 2 H, CH2SO3); 4.10
3.01 (m, 5.5 H, PG-backbone). 13C NMR (176 MHz, D2O, δ): 139.4 (NCdC); 126.4, 125.4 (NCHdC); 78.1, 70.9, 69.5, 68.5 (PG-backbone); 61.2 (Csec.H-triazole); 50.9 (Cprim.H2-triazole); 47.4 (CH2S). IR (KBr) νmax 3465 (s), 3141 (m), 2994 (m), 2944 (m), 2881 (m), 2567 (w), 1652 (m), 1556 (w), 1463 (m), 1414 (m), 1357 (m), 1200 (s), 1128 (s), 1047 (s), 913 (w), 833 (w), 778 (m), 743 (m), 675 (m) cm
1. Dendritic Polyglycerol Carboxylate Sodium Salt (P3, P4, dPGC). Dendritic polyglycerol carboxylate sodium salt (P3, P4, dPGC) was prepared according to the general procedure using commercially available 4-pentynoic acid for click coupling to polyglycerol azide. P3: Mn (dPG core) = 3000 g 3 mol
1, Mn (dPGC) = 7800 g mol
1, dF = 81%, yield: 91%. P4: Mn (dPG core) = 6000 g 3 mol
1, Mn (dPGC) = 16 700 g mol
1, dF = 91%, yield: 96%. 1H NMR (700 MHz, D2O, δ): 8.00
7.05 (m, 1 H, NCHdC); 5.45
3.02 (m, 5.5 H, PG-backbone); 3.00
2.59 (m, 2 H, CH2CH2CdO); 2.59
2.15 (m, 2 H, CH2CH2CdO). 13C NMR (176 MHz, D2O, δ): 181.3 (CdO); 147.9 (NCdC); 123.4, 122.4 (NCHdC); 78.4; 71.2; 69.6; 68.6 (PG-backbone); 60.9 (Csec.Htriazole); 50.6 (Cprim.H2-triazole); 36.9 (CH2CH2CdO); 21.7 (CH2CH2CdO). IR (KBr) νmax 3420 (s), 3140 (m), 2929 (m), 1567 (s), 1411 (s), 1310 (w), 1223 (m), 1124 (m), 1058 (m), 936 (w), 814 (w), 69 (m) cm
1. Dendritic Polyglycerol Phosphonate Sodium Salt (P5, P6, dPGPn). Dendritic polyglycerol phosphonate sodium salt (P5, P6, dPGPn) was synthesized according to the general procedure applying 2 as the ionic alkyne. P5: Mn (dPG core) = 3000 g 3 mol
1, Mn (dPGPn) = 10 150 g 3 mol
1, dF = 81%, yield: 89%. P6: Mn (dPG core) = 6000 g 3 mol
1, Mn (dPGPn) = 26 000 g 3 mol
1, dF = 93%, yield: 94%. 1H NMR (700 MHz, D2O, δ): 8.35
7.55 (m, 1 H, NCHdC); 5.43
4.70 (m, 1 H, CHtriazole); 4.69
4.47 (m, 2 H, CH2OCH2P); 4.47
3.13 (m, 6.3 H, CH2triazole, PG-backbone, CH2OCH2P). 13C NMR (176 MHz, D2O, δ): 144.5 (NCd); 125.4, 124.4 (NCHd); 78.2, 71.5, 69.6, 68.7, 67.8, 64.5 (CH2OCH2P, PG-backbone); 61.0 (CH2OCH2P); 60.8 (Csec.Htriazole); 50.5 (Cprim.H2-triazole). 31P NMR (162 MHz, D2O, δ): 13.3. IR (KBr) νmax 3218 (m), 2989 (m), 2923 (m), 2889 (m), 2532 (w), 1654 (w), 1451 (m), 1350 (w), 1225 (m), 1068 (s), 973 (s), 758 (m) cm
1. Dendritic Polyglycerol Bisphosphonate Sodium Salt (P7, P8, dPGBP). Dendritic polyglycerol bisphosphonate sodium salt (P7, P8, dPGBP) was synthesized according to the general procedure applying 4 as the ionic alkyne. P7: Mn (dPG core) = 3000 g 3 mol
1, Mn (dPGBP) = 15 400 g mol
1, dF = 94%, yield: 88%. P8: Mn

ARTICLE

(dPG core) = 6000 g 3 mol
1, Mn (dPGBP) = 30 050 g 3 mol
1, dF = 95%, yield: 90%. 1H NMR (700 MHz, D2O, δ): 8.18
7.36 (m, 1 H, NCHdC); 5.39
5.10 (m, 1 H, CH-triazole); 4.57
3.29 (m, 4.2 H, PG-backbone); 3.28
2.87 (m, 2 H, triazole-CH2CHP2); 2.47
2.01 (m, 1 H, CH2CHP2). 13C NMR (176 MHz, D2O, δ): 148.0 (NCdCH); 124.3, 123.3 (NCHdC); 78.0, 71.2, 69.7, 68.8 (PG-backbone); 60.6 (Csec.H-triazole); 50.7 (Cprim.H2-triazole); 39.7 (t, 1JCP = 111.4 Hz, CP2); 22.0 (triazole-CH2CHP2). 31P NMR (162 MHz, D2O, δ): 18.9. IR (KBr) νmax 3414 (s), 2346 (m), 1654 (m), 1557 (w), 1460 (m), 1086 (s), 975 (m), 881 (m), 738 (w), 669 (w) cm
1. Dendritic Polyglycerol Phosphate Diethylester (on a 3000 and a 6000 g 3 mol
1 dPG Scaffold). Dendritic polyglycerol (32.4 mmol OHgroups, 2.40 g, 3000 or 6000 g 3 mol
1) was dissolved in dry DMF (50 mL), and N,N-diisopropylethyl amine (DIPEA) (64.8 mmol, 8.38 g, 11.0 mL, 2 eq/OH-group) was added. At
40 °C, diethyl chlorophosphite (64.8 mmol, 10.2 g, 9.3 mL, 2 eq/OH-group) was added dropwise to the solution over a period of 30 min, and the resulting suspension was allowed to warm to rt and was stirred for a further 24 h. Subsequently, the solution was cooled to 0 °C, and tert-butyl hydroperoxide (70% in H2O, 97.2 mmol, 8.76 g, 9.35 mL, 3 eq/OH-group) was added over a period of 30 min. The mixture was stirred at rt overnight, the solvent was evaporated, and the residue was dissolved in DCM (80 mL). The solution was extracted twice with water (2  100 mL), the phases were separated, and the combined organic phase was dried over Na2SO4. After filtration, the solvent was evaporated, and the residue was dialyzed in MeOH/CHCl3 1:1 for 24 h. The title compound was obtained as a yellowish oil in 73% (4.4 g) yield with a degree of functionalization of 82
83% of the hydroxyl groups, as determined by 1H NMR. 1H NMR (400 MHz, CDCl3, δ): 5.70
4.13 (m, 1 H, CHOP(O)(OEt)2); 4.12
3.97 (m, 4 H, POCH2CH3); 3.90
3.10 (m, 4.8 H, CH2OP(O)(OEt)2, PG-backbone); 1.26 (t, 6 H, 3JHH = 7.0 Hz, POCH2CH3). 13C NMR (101 MHz, CDCl3, δ): 78.6, 75.8, 74.8, 71.2, 69.8, 68.7, 66.2 (PGbackbone); 63.7 (OCH2CH3); 61.7 (CH2OH); 15.9 (OCH2CH3). 31P NMR (162 MHz, CDCl3, δ):
0.6,
1.2. IR (KBr) νmax 3454 (s), 2986 (s), 1643 (m), 1479 (m), 1446 (m), 1395 (m), 1370 (m), 1260 (s), 1030 (s), 860 (w), 820 (m), 746 (w) cm
1. Dendritic polyglycerol phosphate sodium salt (P9, P10, dPGP). Dendritic polyglycerol phosphate diethylester (0.20 mmol, 1.5 g, 7490 g 3 mol
1, 6.6 mmol phosphate groups or 99.3 μmol, 1.5 g, 15 100 g 3 mol
1, 6.6 mmol phosphate groups) was dissolved in anhydrous DCM (60 mL) and cooled to 0 °C. Trimethylsilyl bromide (TMSBr) (33.0 mmol, 5.05 g, 4.4 mL, 5 eq per phosphate group) was added dropwise over a period of 1 h. The reaction was allowed to warm to room temperature and was stirred for a further 12 h. Water (50 mL) was added, and the mixture was repeatedly extracted with DCM (2  70 mL). The pH of the separated aqueous phase was adjusted to pH 7 by the addition of aqueous NaOH. Further ultrafiltration in water for 3 days yielded the title compound (70%, 1.0 g) as a colorless solid. P9: Mn (dPG core) = 3000 g 3 mol
1, Mn (dPGP) = 7100 g 3 mol
1, dF = 82%, yield: 70%. P10: Mn (dPG core) = 6000 g 3 mol
1, Mn (dPGP) = 14 300 g 3 mol
1, dF = 83%, yield: 70%. 1H NMR (400 MHz, D2O, δ): 4.50
4.09 (m, 1 H, CHOP); 4.08
3.22 (m, 4.8 H, PG backbone, CH2OP). 13C NMR (101 MHz, D2O, δ): 78.0, 75.3, 72.3, 71.6, 71.2, 69.9, 69.3, 66.2, 65.3, 64.1, 62.8 (PG-backbone, COPO32-). 31P NMR (162 MHz, D2O, δ): 4.8, 4.2. IR (KBr) νmax 3425 (s), 2309 (w), 1654 (m), 1476 (w), 1358 (w), 1100 (s), 978 (m), 766 (w) cm
1. Dendritic Polyglycerol Sulfate (P11, P12, dPGS). A stirred solution of polyglycerol (1.0 g, 13.5 mmol OH groups, Mn = 3000 or 6000 g 3 mol
1) in dry DMF (10 mL) was heated to 60 °C, and a solution of SO3 pyridine complex (16.2 mmol, 2.6 g, 1.2 eq/OH-group) in dry DMF (5 mL) was added over a period of 4 h. The mixture was stirred for 24 h at 60 °C, then water (15 mL) was added, and the pH was adjusted to 7.5 by means of aq NaOH (1 M). Concentration in vacuo yielded the crude product, which was subjected to ultrafiltration in water. After 2504

dx.doi.org/10.1021/bm200250f |Biomacromolecules 2011, 12, 2502–2511

Biomacromolecules freeze-drying, the dendritic polyglycerol sulfate was obtained as colorless, amorphous solid. The degree of sulfation of the dPGS samples was determined via the sulfur content from combustional analysis. P11: Mn (dPG core) = 3000 g 3 mol
1, Mn (dPGS) = 6850 g 3 mol
1, dF = 92%, yield: 90%. P12: Mn (dPG core) = 6000 g 3 mol
1, Mn (dPGS) = 13000 g 3 mol
1, dF = 83%, yield: 80%. 1H NMR (400 MHz, D2O, δ): 4.85
4.63 (m, 1 H, Csec.H-OSO3Na), 4.47
4.15 (m, 2 H, Cprim.H2OSO3Na), 4.14
3.45 (m, 4 H, PG-backbone). 13C NMR (176 MHz, D2O, δ): 78.5, 77.5, 76.1, (Csec.H-OSO3Na), 71.2, 70.5, 69.7, 68.9, 68.5, 67.8, 67.2 (PG-backbone). IR (KBr) νmax: 3502 (m), 2919 (m), 2850 (m), 2359 (w), 2331 (w), 2095 (w), 1653 (m), 1472 (m), 1259 (s), 1075 (m), 1048 (m), 939 (m), 778 (m) cm
1. Sulfur content from combustional analysis: P11: 17.62% S, P12: 16.42% S. DLS and ζ Potential Measurements. DLS and ζ potential measurements were performed with a Zetasizer Nano ZS (Malvern Instruments, Malvern, U.K.) equipped with a 4 mW He
Ne laser (λ = 633 nm). Particle size was measured in UV-transparent disposable cuvettes (Sarstedt, N€umbrecht, Germany). The samples were dissolved in Dulbecco’s phosphate-buffered saline (DPBS, 1, wo Ca2þ and Mg2þ, pH 7.4) (PAA Laboratories GmbH, Pasching, Austria) and were filtered through a 0.2 μm cellulose acetate filter (Whatman Schleicher & Schuell, Dassel, Germany). Usually, samples were prepared at a concentration of 1 mg 3 mL
1, but in the case of small particle sizes (Mn = 3000 g 3 mol
1), the concentration for the measurement had to be increased to 3 mg 3 mL
1 to observe acceptable count rates. Prior to the measurement, the solutions were filtered again through a 0.2 μm Rotilabo PTFE syringe filter (Roth, Karlsruhe, Germany) and treated by ultrasonic for several minutes. The samples were equilibrated for 2 min at 25 °C; subsequently, the measurement was performed with 15 scans per sample. The stated values are the mean from at least five independent measurements. ζ potentials were determined in 10 mM phosphate buffer (0.411 g 3 L
1 Na2HPO4, 0.178 g 3 L
1 KH2PO4) at pH 7.4. The samples were prepared at concentrations of 1 mg 3 mL
1 with freshly prepared buffer. The solutions were subsequently filtered through 0.2 μm Rotilabo PTFE syringe filters (Roth), ultrasonicated for 2 min, and measured by applying an electric field across the polymer solutions using the technique of laser Doppler anemometry at 25 °C in folded DTS 1060 capillary cells (Malvern Instruments). The stated values are the mean from at least five independent measurements with 10 scans, respectively. Data evaluation was performed with Malvern Zetasizer Software 6.12.

Surface Plasmon Resonance (SPR) Measurements as In Vitro L-Selectin Binding Assay. Surface plasmon resonance experiments were carried out on a BIAcore X instrument (GE Healthcare) at 25 °C. The competitive selectin ligand binding assay has been previously described in detail.34 L-selectin-IgG chimera (R&D Systems GmbH, Wiesbaden-Nordenstadt, Germany) was immobilized on Protein-Acoated Au nanoparticles (Ø 15 nm, Biotrend Chemikalien GmbH, Cologne, Germany) to mimic leukocytes with their natural L-selectin ligands on the surface. As a model to the L-selectin ligand-presenting endothelium, a BIAcore sensor chip surface was coated with a standard ligand composed of a poly(N-(2-hydroxyethyl)acrylamide) (PAA) backbone conjugated with both binding motifs, the carbohydrate-based Sialyl-Lewisx (SiaLex, 20 mol %) and the ionic sulfated tyrosine (sTyr, 5 mol %) moiety. Therefore, biotinylated SiaLex-PAA-sTyr (20 mol %/5 mol %) (Lectinity Holdings, Moscow, Russia) was immobilized on a sensor chip SA (GE Healthcare Europe GmbH, Freiburg, Germany) with a surface comprising a carboxymethylated dextran matrix and covalently coupled streptavidin. For reference purposes, biotinylated N-acetyllactosamine-PAA (LacNAc-PAA) (Lectinity Holdings) was immobilized on a second lane of the same chip. L-Selectin coated Au nanoparticles dispersed in 20 mM HEPES buffer (pH 7.4, with 150 mM NaCl and 1 mM CaCl2) were then passed over the ligand presenting and the reference lane of the sensor chip within the BIAcore device at a constant flow rate of 20 μL 3 min
1. The resulting binding signal in

ARTICLE

response units after correction from the nonspecific interaction as simultaneously determined from the LacNAc-PAA coated reference lane was set to 100% binding of L-selectin to the ligand on the surface and served as a control. To evaluate L-selectin binding of potential dendritic polyanionic inhibitors, a defined preincubation step with the selectin nanoparticles was performed before its passage over the sensor chip. Reduction of the binding signal with respect to the inhibitor concentration was recorded and calculated as X% binding of the control. The inhibitor concentration that caused 50% reduction of binding was referred to as IC50 value. Each concentration was applied at least in duplicate.

’ RESULTS AND DISCUSSION Dendritic polyglycerol (dPG) (Scheme 1 A) is a globular polymer that can rapidly be synthesized by a one-step anionic multibranching ring-opening polymerization of glycidol on a polyol initiator such as pentaerythritol or 1,1,1-tris(hydroxymethyl) propane with low PDIs (typically PDI < 1.8).32 In addition, this polymer offers multiple hydroxyl groups on the surface of its flexible polyether backbone for further derivatization to present ligands multivalently. Such multivalent presentation of ligands on a synthetic polymer scaffold, to enhance binding affinity, is a common practice, as demonstrated by sulfated βlactose, which was conjugated to a dendrimer-like poly(ethylene glycol) (PEG) scaffold.35 Thereby the multivalently presented sulfated β-lactose ligand (12mer) revealed L-selectin specific binding affinity similar to heparin, but conjugates were inactive when the ligands were presented only as tri- or tetramers. In contrast with the multistep synthesis of a dendrimer-like PEG scaffold, dPG is easily accessible also on large scale and presents many more surface hydroxyl groups than a dendrimer-like PEG scaffold (second-generation dendrimer-like PEG: Mn ≈ 52 kDa, 12 surface hydroxyl groups; dPG: Mn ≈ 6 kDa, 80 surface hydroxyl groups). In addition, dPG exhibits similar or even better biocompatibility than PEG, which makes them superior scaffolds for studying multivalent effects with biological entities.36 However, a high degree of functionalization of the surface hydroxyl groups of dPG is needed to observe the highest binding affinity toward L-selectin, as previously demonstrated with dPGS of various degree of sulfation on the same dPG scaffold.18 In contrast with the synthesis of dPGS, which is straightforward by using SO3 3 pyridine complex as the sulfation reagent, it turned out that the incorporation of other anionic species with a high degree of functionalization is rather challenging because of solubility issues. dPG as a highly polar polymer is only soluble in polar protic or aprotic solvents such as water or dimethylformamide (DMF). If aprotic solvents are required for reactions under anhydrous conditions, then the anionic charged precursors, which are intended to be coupled, exhibit only low solubility and thus result in insufficient coupling efficiency. Gode et al. have reported on the synthesis of sulfonated dendritic poly(3-ethyl3-(hydroxymethyl) oxetane) by deprotonation of the polymers’ surface hydroxyl groups with sodium hydride and subsequent ring-opening reaction with 1,4-butane sultone.37 However, the highest degree of functionalization reported was ∼50%. In accordance with these results, we were never able to achieve high degrees of functionalization on a dPG scaffold when applying this synthetic route of introducing sulfonate groups (ESI, SI
P1 and SI
P2). This is somewhat expected because upon full deprotonation of the hydroxyl groups, the polymer precipitates from solution. Thereby no increase in solubility in aprotic solvents, which are required in such reactions, is achieved 2505

dx.doi.org/10.1021/bm200250f |Biomacromolecules 2011, 12, 2502–2511

Biomacromolecules

ARTICLE

Scheme 1. Synthesis and Structure of Highly Functionalized Dendritic Polyglycerol (dPG)-Based Polyanions with a Degree of Functionalization (dF) > 80%a

a

(A) Structure of dendritic polyglycerol and ionic alkynes 1, 2, 3, and 4. (B) Click coupling of dPG-Azide with ionic alkynes 1, 2, 3, and 4. (C) Synthesis of dPGP and dPGS via direct hydroxyl interconversion of dPG.

upon reaction with the sultone because another anionic charge is generated on the polymer. In contrast, when sulfating dPG with SO3 3 pyridine complex (Scheme 1 C), the generated organic pyridinium counterion keeps the anionic polymer in solution and therefore allows high degrees of sulfation. Consequently, Michael-type addition reactions of the hydroxyl groups of dPG with anionic vinyl compounds like vinyl sulfonic acid sodium salt

or vinyl carboxylic acid sodium salt always resulted in a low degree of functionalization (95%) of the dPG hydroxyl groups and yielded the corresponding carboxylate upon saponification of the nitrile groups.39 However, in our hands, a concomitant polymerization of acrylonitrile to a low extent was observed; therefore, the desired polyglycerol carboxylate (dPGC) contained inseparable traces of poly(acrylic acid) after quantitative saponification of the nitrile groups (data not shown). In addition, a strong yellow color of the usually colorless product was observed; therefore, this synthetic route did not qualify for our purposes. Alternatively, high conversions of the hydroxyl groups to carboxylates can be achieved upon reaction with succinic anhydride to yield dPGC but with a labile ester linker.40 A promising route to dendritic polyglycerol phosphonate (dPGPn) seemed to be a Mannich-type reaction of previously aminated polyglycerol with formaldehyde and phosphoric acid, which was analogously reported on dendritic poly (amido amine) tethered on silica.41 Unfortunately, in our hands, this reaction failed and yielded a multitude of phosphor species on the polymer, as indicated by the 31P NMR spectra after isolation (data not shown). In summary, these synthetic routes failed or resulted in an insufficient degree of functionalization. In 2008, a versatile click chemistry approach was reported by Sousa-Herves et al. on a dendritic PEG-based block copolymer with terminal azide groups and certain sulfated, sulfonated, and carboxylated alkynes, which yielded high degrees of functionalization.42 Applying this strategy to dendritic polyglycerol, we prepared azide-functionalized dPG and coupled anionic alkynes bearing carboxylate, sulfonate, phosphonate, and bisphosphonate groups, respectively, which yielded the anionic polymers with a high degree of functionalization (Scheme 1A,B). This can be attributed to the aqueous conditions under which 1,3-dipolar

cycloadditions can be performed, thereby overcoming all solubility issues. The hydroxyl groups of dPG were converted to azide groups with a degree of functionalization of 95% according to literature by mesylation and subsequent reaction with sodium azide.33 Anionic alkynes in their sodium salt state were coupled to the polyglycerol azide scaffold applying click chemistry. The alkyne carboxylate 3 is commercial, whereas the alkyne sulfonate 1 was synthesized according to literature (ESI).43 Alkyne phosphonate 2 and bisphosphonate 4 were synthesized from their respective literature-known ethyl ester precursors44
47 by ester hydrolysis applying TMSBr and ion exchange with sodium hydroxide (ESI). The click coupling was performed according to standard conditions in a water/tetrahydrofuran (1:1) mixture because azide-functionalized dPG is only soluble in organic solvents, whereas the alkyne and the catalyst system made up of 5 mol % copper(II) sulfate and 25 mol % sodium ascorbate as the reducing agent is soluble under aqueous conditions. After click coupling, which usually was performed for 48 h to ensure a high degree of functionalization, all polymers were purified by ultrafiltration. Dendritic polyglycerol phosphates, in contrast, were synthesized by a direct hydroxyl to phosphate group interconversion applying chloro diethylphosphite, in situ oxidation, and subsequent hydrolysis of the ester groups by TMSBr, followed by ion exchange (Scheme 1C).48
50 All polyanionic polymers were synthesized by applying dPG scaffolds of both 3000 and 6000 g 3 mol
1, which resulted in similar degrees of functionalization for both core sizes, as determined via NMR. Chemical specification of the different polyanions is summarized in Table 1. This set of polyanions is well-suited for further investigation of their inhibitory effect on L-selectin binding to a synthetic model ligand because they are all highly functionalized with similar degrees of functionalization, >80% of which corresponds to a number of functional, anionic 2507

dx.doi.org/10.1021/bm200250f |Biomacromolecules 2011, 12, 2502–2511

Biomacromolecules

Figure 1. (A) Schematic illustration of the SPR based L-selectin inhibition assay. (B,C) Plots of concentration dependent relative L-selectin binding of different polyanions based on a 3000 (B) and a 6000 g 3 mol
1 (C) dPG core.

groups ranging from 32 to 38 per dPG (3000 g 3 mol
1) and 67 to 77 per dPG (6000 g 3 mol
1) scaffold, respectively. For the evaluation of L-selectin binding, a well-established competitive SPR-based in vitro assay was applied.34 The experimental setup is schematically displayed in Figure 1A. L-selectincoated Au nanoparticles are passed over a SPR sensor chip surface that presents a synthetic standard ligand for L-selectin. The ligand comprises both ionic tyrosine-sulfate- and carbohydrate-based SiaLex binding moieties tethered on a polymeric backbone. Selectin-coated nanoparticles without preincubation of potential inhibitors were initially passed over the sensor chip surface. The resulting binding signal, which indicates nanoparticle adhesion, was set to 100% binding and serves as the control value. Subsequently, L-selectin-coated nanoparticles preincubated with the distinct dendritic polyanions at given concentrations were passed over the sensor chip surface. Thereby, a compound that blocks selectin
ligand interaction and reduces the binding is considered to be an inhibitor. The recorded binding signal was calculated as % binding of the control. This experimental in vitro setup mimics in vivo leukocyte binding to the endothelium in the presence of a synthetic inhibitor in good approximation. In addition, the assay is performed under

ARTICLE

constant shear stress in the microfluidic flow chamber of the BIAcore SPR device as it occurs in vivo and which has been shown to be a crucial prerequisite for leukocyte capture by endothelium.34,51 Survey plots of the inhibitor concentration versus the relative reduced binding of the control are given in Figure 1B for dendritic polyanions on a 3000 g 3 mol
1 dPG scaffold in comparison with a nonionic, bare dPG scaffold. Figure 1C analogously illustrates inhibition plots for dendritic polyanions on a 6000 g 3 mol
1 dPG scaffold. From Figure 1B,C, the respective inhibitor concentrations are accessible, at which 50% binding of the control is observed (IC50) (Table 1). Neutral dPG revealed no inhibition toward L-selectin binding within the applied concentration range, whereas all anionically modified polyglycerols on a 6000 g 3 mol
1 dPG scaffold showed a distinct, concentration-dependent inhibitory potential within the low nanomolar to high micromolar range. By comparing the IC50 values of the different polyanions on a 6000 g 3 mol
1 dPG scaffold, the inhibitory potential decreases in the order sulfate (P12) > bisphosphonate (P8) > sulfonate (P2) ≈ phosphonate (P6) > phosphate (P10) > carboxylate (P4). Basically the same order results when comparing the IC50 values of the different polyanions on a 3000 g 3 mol
1 dPG scaffold. Thereby the carboxylate inhibitory concentration, however, already exceeded the screened concentration range and the IC50 value is therefore assigned as not inhibitory (n.I.) for dPGC (P3). From the initiated decrease in binding observed for P3 in the high micromolar range (Figure 1B), the IC50 value is presumably located in the millimolar range. By comparing the same anionic species on the two different applied dPG core sizes, it becomes obvious that an increase in the core size has a beneficial effect on the L-selectin binding avidity of dPGS, dPGBP, dPGPn, dPGP, and dPGC. A similar observation was reported for sulfated amylose as the inhibitor in a competitive static binding assay that revealed enhanced L-selectin binding affinity with increasing molecular weight of the inhibitor.52 This was attributed to an increased clustering of anionic groups with rising molecular weight of the inhibitor, which in turn maximizes binding affinity. Such clustering of anionic groups occurs because of the flexible polymer scaffold that bears the anionic groups. This results in a high local charge density by a structural rearrangement of the inhibitor for optimized binding to the cationic surface domains of selectins. However, such a beneficial molecular weight effect of the dPG core was not observed for dPGSn (P5, P6) because their respective IC50 values are approximately the same. In general, the inhibitory effect of the sulfated dPG (P11, P12) with an IC50 value in the low nanomolar range is significantly higher (approximately an order of magnitude) than the ones of sulfonated (P1, P2), phosphonated (P5, P6), phosphorylated (P9, P10), and carboxylated (P3, P4) dPG. In between these two sets of dimensions operates dPGBP (P7, P8) with an IC50 value of 360 and 260 nM, respectively. The reproducibility and stability of the IC50 values determined by this assay are further supported by the IC50 values obtained from dPGSn (SI
P1, SI
P2) and dPGPn (SI
P3, SI
P4). The latter ones were synthesized via an alternative route, as described in the Synthesis section, without click chemistry yielding lower degrees of functionalization (ESI). The IC50 values of the phosphonates SI
P3 and SI
P4 are assumed to be located in the millimolar range. Because this exceeded the screened concentration range, they were assigned as not inhibitory. The 2508

dx.doi.org/10.1021/bm200250f |Biomacromolecules 2011, 12, 2502–2511

Biomacromolecules determined IC50 values of these polyanions (ESI, Figure S5 and S6 of the Supporting Information) were in the same order of magnitude as their respective analogs synthesized via click chemistry. However, these polyanions showed somewhat higher IC50 values probably because of the lower degree of functionalization. This is in accordance with the previous observation that L-selectin inhibition by dPGS is enhanced with an increasing degree of sulfation.18 In addition, the present data suggest that L-selectin binding of dPG-based anions is independent of the linking chemistry between dPG and the respective anionic group as long as the degree of functionalization is the same. L-selectin bears a cationic binding site apart from the carbohydrate binding site, and both are suggested to be addressed simultaneously in the case of the physiological L-selectin ligands, which results in enhanced binding strength. In the case of dendritic polyanions, it is assumed that solely the cationic binding site is addressed because of the lack of sugars moieties. Therefore, the hypothesis was that any kind of multivalent anionic polymer can electrostatically bind to L-selectin with a simultaneous sterical shielding of the carbohydrate recognition site that inhibits the physiological L-selectin-ligand binding. Our results clearly support this hypothesis but at the same time raise the question for the reason behind the observed differences in the inhibitory efficiency of the distinct polyanionic species. At first glance, the rank of anions with rising L-selectin binding affinity seems to increase with the acidity of the anionic group. However, this can merely account for the huge differences in IC50 values of dPGS and dPGSn. Therefore, DLS and ζ potential measurements were performed in buffered solution at pH 7.4, which resembles the conditions of the biological experiment. The obtained hydrodynamic sizes as well as the ζ potentials are listed in Table 1 and graphically plotted in Figure 2. In general, the hydrodynamic size of a polyelectrolyte depends on the nature of the ionic group, its surface charge density, the ionic strength as well as the pH of the aqueous solution. These factors together with the respective pKa value of the polyelectrolyte govern the degree of protonation of the polyanions, including conformational changes and partial aggregation due to intramolecular hydrogen bonding. All factors that influence the hydrodynamic size will also affect the ζ potential and thus the L-selectin binding efficiency of polyanions and remain hidden when looking only at the molecular weight.30 By comparing the distribution of size by volume, as obtained from DLS measurements of the different polymers (Figure 2A), it becomes obvious that functionalization of the respective dPG core with anionic groups expectedly leads to an increased hydrodynamic diameter. This can be attributed to electrosteric repulsion of the anionic groups on the surface of dPG, which leads to a more “stretched and swollen” configuration to gain maximal distance between the repulsive anionic surface charges. Moreover, the surface charges are stabilized by the generation of an ionic double layer of hydrated counterions, which screens the charges and contributes to the hydrodynamic size.53 However, a significant osmotic swelling of the polyanions due to confined counterions, as observed under salt-free conditions, should be negligible because DLS measurements were performed under relatively high ionic strength conditions.54 In contrast with sulfated, carboxylated, phosphonated, and phosphorylated polyglycerols for which the size tends to increase with the applied dPG core, sulfonated (P1, P2) and bisphosphonated (P7, P8) polyglycerols resulted in the same hydrodynamic size after functionalization of both dPG scaffolds, respectively (Figure 2A). This is difficult to explain and might be due to a conformational change of

ARTICLE

Figure 2. (A) Plot of the hydrodynamic size by volume of polyanions on a 3000 g 3 mol
1 and a 6000 g 3 mol
1 dPG scaffold, respectively, as determined via DLS in PBS at pH 7.4. (B) Plot of the ζ potential of polyanions on a 3000 and a 6000 g 3 mol
1 dPG scaffold, respectively, in 10 mM phosphate-buffered solution at pH 7.4. (The subscripts in the labeling of the x axes indicate the molecular weight of the dPG core in kDa and the number of the respective multivalently presented anionic group).

the respective higher molecular weight polyanion by which the charged groups partially fold back toward the interior polyether scaffold to minimize electrosteric repulsion. A similar observation was recently made in the case of very high molecular weight dPGS samples on a 250 and a 480 kDa dPG scaffold, which resulted in the same hydrodynamic size upon sulfation (manuscript in preparation). The observed constant hydrodynamic size in the case of dPGSn could be the reason for the similar IC50 values of P1 and P2, which in contrast with the other polyanions do not show a trend to enhanced L-selectin inhibition with increasing size of the respective dPG core (Table 1). However, dPGBP P8 showed enhanced L-selectin avidity compared with its lower molecular weight analog P7 despite having similar hydrodynamic size. Hence, particle size and surface charge are two complex interdigitating factors that both contribute to L-selectin avidity of polyanionic macromolecules. When comparing the ζ potential of the different polyanions (Figure 2B), no obvious correlation with their respective IC50 values can be stated. The effective charge of all dendritic polyanions except for dPGS remains the same within the range of error for the respective set of polyanions on the 3000 and 6000 2509

dx.doi.org/10.1021/bm200250f |Biomacromolecules 2011, 12, 2502–2511

Biomacromolecules g 3 mol
1 dPG scaffold, although the nominal amount of charges approximately doubles. In strong contrast is the ζ potential of dPGS that increases with the nominal amount of charges per molecule from approximately
5 mV for P11 to
14 mV for P12. Moreover, the measured higher effective charge of dPGC and dPGSn compared with the ζ potential of dPGS is rather unexpected because the ζ potential depends on the Nernst potential of the particle, which in turn is related to the surface charge density on a particle.53 Because dPGC (P3, P4), which is the weakest polyelectrolyte within this comparative study, is likely to be partially protonated under the applied conditions, the determined higher effective charge as compared with dPGS is surprising at first glance. Because the presence of ions, however, also strongly influences the ζ potential, the counterion binding strength of the various anionic groups might change the thickness of the ionic double layer and hence influence the measured ζ potential. Recently, a Hofmeister analog ranking of anionic surfactant head groups toward their capability of forming close ion pairs with various monovalent counterions was suggested.55 According to this series, dPGC should have the smallest ionic double layer because of much stronger (closer) sodium counterion binding as compared with dPGS. This in turn might be the reason for the determined and initially unexpected lower effective charge of dPGS as compared with dPGC and dPGSn. The sodium counterion is solely considered here because it is the one that is present in highest concentration under the conditions of the ζ potential measurements. In summary, dPGS seems to take an exceptional position among the studied sets of dendritic polyanions with similar degree of functionalization. The sulfated nanoparticles revealed the lowest effective charge and the smallest hydrodynamic size with a uniquely strong L-selectin inhibition property. However, further detailed studies on the microstructure including conformation and local pH at the surface of the polyelectrolytes are required to correlate generally the inhibitory potential of the polyanions with the nature of the anionic groups.

’ CONCLUSION Herein we report on the synthesis of highly functionalized polyanions based on a dendritic polyglycerol scaffold via a versatile click chemistry approach applying anionic sulfonate, carboxylate, phosphonate, and bisphosphonate alkynes. In addition, the synthesis of dendritic polyglycerol phosphate was accomplished via a direct hydroxyl group interconversion with chloro diethylphosphite, in situ oxidation, and subsequent ethyl ester hydrolysis. All polyanions were synthesized on a 3000 and a 6000 g 3 mol
1 dPG scaffold with a comparable, high degree of functionalization (>80%), respectively, and were subjected to a competitive L-selectin binding assay. To the best of our knowledge, no comparative study on the influence of the nature of the anionic group exists in literature. Thereby all dendritic anions on the 6000 g 3 mol
1 scaffold revealed a concentration-dependent inhibitory effect on the L-selectin binding to its ligand comprising the low nanomolar to high micromolar range. The inhibitor potential increased in the order carboxylate < phosphate < phosphonate ≈ sulfonate < bisphosphonate < sulfate and hence tends to increase with the acidity of the anionic group. This study further demonstrates that the inhibition of L-selectin-ligand binding tends to be more efficient for the respective anion on the 6000 g 3 mol
1 dPG scaffold compared with the 3000 g 3 mol
1 dPG core. Both hydrodynamic size and effective charge

ARTICLE

indicate an exceptional position of dPGS among the studied polyanions, which is also reflected by the remarkably low IC50 value from the L-selectin inhibition assay. However, further indepth investigations of the conformation, distinct protonation (pKa values), and local pH on the surface of the polyanions under the buffered conditions of the biological experiment are required to understand fully the observed differences in L-selectin inhibition. In addition, the strength of counterion binding of the respective anionic group according to the Hofmeister series might have a significant influence on selectin binding avidity of polyanions.

’ ASSOCIATED CONTENT

bS

Supporting Information. Synthesis and characterization of ionic alkynes and additional L-selectin binding data of alternatively synthesized polyanions. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Author Contributions §

These authors contributed equally.

’ ACKNOWLEDGMENT SFB765 of the Deutsche Forschungsgemeinschaft is kindly acknowledged for financial support. ’ REFERENCES (1) Inflammation: Basic Principles and Clinical Correlates; Gallin, J. I., Goldstein, I. M., Snyderman, R., Eds.; Raven Press: New York, 1988. (2) Butcher, E. C. Cell 1991, 67, 1033–1036. (3) Simanek, E. E.; McGarvey, G. J.; Jablonowski, J. A.; Wong, C.-H. Chem. Rev. 1998, 98, 833–862. (4) Springer, T. A. Annu. Rev. Physiol. 1995, 57, 827–872. (5) Springer, T. A. Nature 1990, 346, 425–434. (6) Epstein, F. H.; Luster, A. D. N. Engl. J. Med. 1998, 338, 436–445. (7) Ley, K. Trends Mol. Med. 2003, 9, 263–268. (8) Koenig, A.; Jain, R.; Vig, R.; Norgard-Sumnicht, K. E.; Matta, K. L.; Varki, A. Glycobiology 1997, 7, 79–93. (9) Sako, D.; Comess, K. M.; Barone, K. M.; Camphausen, R. T.; Cumming, D. A.; Shaw, G. D. Cell 1995, 83, 323–331. (10) Boehncke, W. H.; Sch€ on, M. P.; Giromolomi, G.; Bos, J. D.; Thestrup-Pedersen, K.; Cavani, A.; Nestle, F.; Bonish, B. K.; Campbell, J. J.; Nickoloff, B. Exp. Dermatol. 2005, 14, 70–80. (11) Bendas, G. Mini-Rev. Med. Chem. 2005, 5, 575–584. (12) Rychly, J.; Nebe, B. Curr. Pharm. Des. 2006, 12, 3799–3806. (13) Grace, P. A. Br. J. Surg. 1994, 81, 637–647. (14) Lefer, A. M.; Weyrich, A. S.; Buerke, M. Cardiovasc. Res. 1994, 28, 289–294. (15) Coussens, L. M.; Werb, Z. Nature 2002, 420, 860–867. (16) Brooks, S. A.; Lomax-Browne, H. J.; Carter, T. M.; Kinch, C. E.; Hall, D. M. S. Acta Histochem. 2010, 112, 3–25. (17) T€urk, H.; Haag, R.; Alban, S. Bioconjugate Chem. 2004, 15, 162– 167. (18) Dernedde, J.; Rausch, A.; Weinhart, M.; Enders, S.; Tauber, R.; Licha, K.; Schirner, M.; Zuegel, U.; von Bonin, A.; Haag, R. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 19679–19684. (19) Mammen, M.; Choi, S.-K.; Whitesides, G. M. Angew. Chem., Int. Ed. 1998, 37, 2754–2794. 2510

dx.doi.org/10.1021/bm200250f |Biomacromolecules 2011, 12, 2502–2511

Biomacromolecules (20) Toppila, S.; Lauronen, J.; Mattila, P.; Turunen, J. P.; Penttil€a, L.; Paavonen, T.; Renkonen, O.; Renkonen, R. Eur. J. Immunol. 1997, 27, 1360–1365. (21) Roy, R.; Park, W. K. C.; Srivastava, O. P.; Foxall, C. Bioorg. Med. Chem. Lett. 1996, 6, 1399–1402. (22) Thoma, G.; Duthaler, R. O.; Magnani, J. L.; Patton, J. T. J. Am. Chem. Soc. 2001, 123, 10113–10114. (23) Sch€on, M. P.; Krahn, T.; Sch€on, M.; Rodriguez, M.-L.; Antonicek, H.; Schultz, J. E.; Ludwig, R. J.; Zollner, T. M.; Bischoff, E.; Bremm, K.-D.; Schramm, M.; Henninger, K.; Kaufmann, R.; Gollnick, H. P. M.; Parker, C. M.; Boehncke, W. H. Nat. Med. 2002, 8, 366–372. (24) Malhotra, R.; Taylor, N. R.; Bird, M. I. Biochem. J. 1996, 314, 297–303. (25) Suzuki, Y.; Toda, Y.; Tamatani, T.; Watanabe, T.; Suzuki, T.; Nakao, T.; Murase, K.; Kiso, M.; Hasegawa, A.; Tadanoaritomi, K.; Ishizuka, I.; Miyasaka, M. Biochem. Biophys. Res. Commun. 1993, 190, 426–434. (26) Aruffo, A.; Kolanus, W.; Walz, G.; Fredman, P.; Seed, B. Cell 1991, 67, 35–44. (27) Skinner, M. P.; Lucas, C. M.; Burns, G. F.; Chesterman, C. N.; Berndt, M. C. J. Biol. Chem. 1991, 266, 5371–5374. (28) Simonis, D.; Christ, K.; Alban, S.; Bendas, G. Semin. Thromb. Hemostasis 2007, 33, 534–539. (29) Stoolman, L. M.; Tenforde, T. S.; Rosen, S. D. J. Cell Biol. 1984, 99, 1535–1540. (30) Somers, W. S.; Tang, J.; Shaw, G. D.; Camphausen, R. T. Cell 2000, 103, 467–479. (31) Calderon, M.; Quadir, M. A.; Sharma, S. K.; Haag, R. Adv. Mater. 2010, 22, 190–218. (32) Sunder, A.; Hanselmann, R.; Frey, H.; M€uhlhaupt, R. Macromolecules 1999, 32, 4240–4246. (33) Roller, S.; Zhou, H.; Haag, R. Mol. Diversity 2005, 9, 305–316. (34) Enders, S.; Bernhard, G.; Zakrzewicz, A.; Tauber, R. Biochim. Biophys. Acta, Gen. Subj. 2007, 1770, 1441–1449. (35) Rele, S. M.; Cui, W.; Wang, L.; Hou, S.; Barr-Zarse, G.; Tatton, D.; Gnanou, Y.; Esko, J. D.; Chaikof, E. L. J. Am. Chem. Soc. 2005, 127, 10132–10133. (36) Kainthan, R. K.; Janzen, J.; Levin, E.; Devine, D. V.; Brooks, D. E. Biomacromolecules 2006, 7, 703–709. (37) Gode, P.; Hult, A.; Jannasch, P.; Johansson, M.; Karlsson, L. E.; Lindbergh, G.; Malmstrom, E.; Sandquist, D. Solid State Ionics 2006, 177, 787–794. (38) Haag, R.; Dernedde, J.; Tauber, R.; Gesche, B.; Enders, S.; Weinhart, H. Dendritic Polyglycerol Sulfates and Sulfonates and Their Use for Inflammatory Diseases. PCT International Application WO 2008015015, 2008. (39) Barriau, E.; Frey, H.; Kiry, A.; Stamm, M.; Gr€ohn, F. Colloid Polym. Sci. 2006, 284, 1293–1301. (40) Tziveleka, L.-A.; Kontoyianni, C.; Sideratou, Z.; Tsiourvas, D.; Paleos, C. M. Macromol. Biosci. 2006, 6, 161–169. (41) Punyacharoennon, P.; Charuchinda, S.; Srikulkit, K. J. Appl. Polym. Sci. 2008, 110, 3336–3347. (42) Sousa-Herves, A.; Fernandez-Megia, E.; Riguera, R. Chem. Commun. 2008, 3136–3138. (43) Ryu, E.-H.; Zhao, Y. Org. Lett. 2005, 7, 1035–1037. (44) Bulman Page, P. C.; Moore, J. P. G.; Mansfield, I.; McKenzie, M. J.; Bowler, W. B.; Gallagher, J. A. Tetrahedron 2001, 57, 1837–1847. (45) Sawicki, M.; Siaugue, J.-M.; Jacopin, C.; Moulin, C.; Bailly, T.; Burgada, R.; Meunier, S.; Baret, P.; Pierre, J.-L.; Taran, F. Chem.—Eur. J. 2005, 11, 3689–3697. (46) Bansal, G.; Wright, J. E. I.; Kucharski, C.; Uludag, H. Angew. Chem., Int. Ed. 2005, 44, 3710–3714. (47) Delain-Bioton, L.; Villemin, D.; Jaffres, P.-A. Eur. J. Org. Chem. 2007, 1274–1286. (48) Mills, S. J.; Potter, B. V. L. Bioorg. Med. Mat. 2003, 11, 4245–4253. (49) Xu, Y.; Liu, X.-h.; Prestwich, G. D. Tetrahedron Lett. 2005, 46, 8311–8314. (50) Chung, S.-K.; Kwon, Y.-U.; Shin, J.-H.; Chang, Y.-T.; Lee, C.; Shin, B.-G.; Kim, K.-C.; Kim, M.-J. J. Org. Chem. 2002, 67, 5626–5637.

ARTICLE

(51) Finger, E. B.; Purl, K. D.; Alon, R.; Lawrence, M. B.; von Andrian, U. H.; Springer, T. A. Nature 1996, 379, 266–269. (52) Yoshida, T.; Fennie, C.; Lasky, L. A.; Lee, Y. C. Eur. J. Biochem. 1994, 222, 703–709. (53) Koetz, J.; Kosmella, S. Polyelectrolytes and Nanoparticles; Springer-Verlag: Berlin, 2007. (54) Ballauff, M. Prog. Polym. Sci. 2007, 32, 1135–1151. (55) Vlachy, N.; Jagoda-Cwiklik, B.; Vacha, R.; Touraud, D.; Jungwirth, P.; Kunz, W. Adv. Colloid Interface Sci. 2009, 146, 42–47.

2511

dx.doi.org/10.1021/bm200250f |Biomacromolecules 2011, 12, 2502–2511