Heparan Sulfate Mimicking Glycopolymer Prevents Pancreatic β Cell Destruction and Suppresses Inflammatory Cytokine Expression in Islets under the Challenge of Upregulated Heparanase

Abstract

Diabetes is a chronic disease in which the levels of blood glucose are too high because the body does not effectively produce insulin to meet its needs or is resistant to insulin. β Cells in human pancreatic islets produce insulin, which signals glucogen production by the liver and causes muscles and fat to uptake glucose. Progressive loss of insulin-producing β cells is the main cause of both type 1 and type 2 diabetes. Heparan sulfate (HS) is a ubiquitous polysaccharide found at the cell surface and in the extracellular matrix (ECM) of a variety of tissues. HS binds to and assembles proteins in ECM, thus playing important roles in the integrity of ECM (particularly basement membrane), barrier function, and ECM–cell interactions. Islet HS is highly expressed by the pancreatic β cells and critical for the survival of β cells. Heparanase is an endoglycosidase and cleaves islet HS in the pancreas, resulting in β-cell death and oxidative stress. Heparanase could also accelerate β-cell death by promoting cytokine release from ECM and secretion by activated inflammatory and endothelial cells. We demonstrate that HS-mimicking glycopolymer, a potent heparanase inhibitor, improves the survival of cultured mouse pancreatic β cells and protects HS contents under the challenge of heparanase in human pancreatic islets. Moreover, this HS-mimicking glycopolymer reduces the expression levels of cytokines (IL8, IL1β, and TNFα) and the gene encoding Toll-like Receptor 2 (TLR2) in human pancreatic islets.

Graphical Abstract

INTRODUCTION

Diabetes is a disorder that impairs the body’s ability to produce enough insulin for processing blood glucose or to effectively use the insulin it makes.¹ β Cells in pancreatic islets produce insulin, which, in turn, signals glycogen production by the liver and causes muscles and fat to uptake glucose. Individuals with type I diabetes (T1D) cannot process glucose needed for energy supply, and glucose accumulates in the blood because the insulin-producing pancreatic β cells have been destroyed. As such, T1D patients require daily insulin injections to lower their blood glucose levels.² People with type 2 diabetes (T2D) cannot make enough insulin to convert glucose into energy or are resistant to the insulin they do produce.² Therefore, T2D patients eventually need insulin shots to allow their body to process glucose and prevent complications.

Heparan sulfate proteoglycans (HSPGs) are found at the cell surface and in the extracellular matrix (ECM) of a variety of tissues.³ HSPGs contain several heparan sulfate (HS) chains that are attached to the core proteins. Heparan sulfate (1) is a negatively charged linear polysaccharide composed of repeating glucosamine (GlcN) and uronic acid [glucuronic acid (GlcA) or iduronic acid (IdoA)] disaccharide moieties (Figure 1). HS is modified by various degrees of N-sulfation/acetylation on glucosamine residues and O-sulfation at various sites (Figure 1).³ HS interacts with ECM proteins, therefore involving in ECM integrity, barrier function, and ECM–cell communications. Further, HS regulates the activity of bioactive molecules (growth factors, cytokines, chemokines, and blood coagulation factors) at the cell surface and in the ECM, thereby regulating diverse normal and pathological processes.^3,4

Figure 1.

Structure of heparan sulfate chains.

Intra-islet HS is highly expressed within the basement membrane of pancreatic islets. Islet HS protects peri-islet from attack by destructive insulitis mononuclear cells.⁵ Therefore, the structural integrity of intra-islet HS is critical for the survival of islet β cells in the pancreas.⁶ Flow cytometry studies confirmed that a decrease in intra-islet HS correlates with a significant amount of islet β-cell death.⁷ HS also acts as an antioxidant to protect the β cells from free-radical-mediated damage.⁶

Heparanase, a mammalian enzyme capable of cleaving HS chains,⁸ has been implicated in diabetes and its complications.⁹ High levels of heparanase have been detected in the plasma and urine of diabetic individuals.^10,11 Under hyperglycemic conditions, a significant increase in heparanase levels in endothelial cells has also been detected in individuals with diabetes.^11,12 Heparanase involvement in intra-islet HS cleavage to promote β-cell death has been evidenced by showing that autoimmune destruction of pancreatic islets was associated with increased levels of heparanase expression by infiltrating mononuclear cells.⁶ Another possible mechanism for β-cell destruction involves an increased release of cytokines, such as TNFα, IFNγ, and interleukins, from activated inflammatory and endothelial cells.¹³ Although these cytokines are produced by inflammatory and endothelial cells, intra-islet HS could serve as a supplementary source for these cytokines.¹³ Therefore, HS degradation by heparanase could lead to β-cell destruction by creating a burst of cytokine release.

Heparanase cleaves the internal glycosidic bond that connects the C1-carbon of glucuronic acid (GlcA) to the C4-hydroxyl of sulfated glucosamine (GlcN) of HS (Figure 2).¹⁴ Degradation of HS by heparanase leads to disassembly of the ECM and BM and is therefore involved in fundamental biological phenomena associated with tissue remodeling and cell migration.^15-17 The heparanase mRNA encodes a 61.2 kDa protein of 543 amino acids. This pro-enzyme is post-translationally cleaved by cathepsin L, into 8 and 50 kDa subunits that noncovalently associate to form active heparanase.¹⁸ The binding site of heparanase is flanked by two heparin-binding domains (HBD1 and HBD2).¹⁹ Amino acid residues, Glu225 and Glu343, located in the active site of heparanase are responsible for the hydrolysis of β-(1,4)-linkage between GlcA at −1 subsite and GlcN at +1 subsite (Figure 2).¹⁹ The co-crystallized structures of heparanase when bound to HS ligands show that a trisaccharide, GlcN-GlcA-GlcN, spanning – 2, −1, and +1 subsites of heparanase is a minimal recognition sequence.¹⁹ Extracellular heparanase enters pancreatic β cells likely through interaction with HS on the cell surface followed by endocytosis.²⁰

Figure 2.

Structure basis of mammalian heparanase and HS substrate interactions.

Small molecules have been developed as potent inhibitors of heparanase.²¹ However, only sulfated carbohydrate molecules have reached clinical trials.²² Except for PG545 (a highly sulfated tetrasaccharide bound to a lipophilic cholestanol),²³ these sugar compounds have varying sizes and sulfation patterns leading to cross-bioactivity with many heparin-binding proteins and unforeseen adverse effects. As a result, most of these carbohydrate molecules were either terminated or halted after reaching clinical trials.²² Among these heparanase inhibitors, PI-88, which is a mixture of sulfated, mono-phosphorylated oligosaccharides ranging in size from two to six mannose units,²⁴ reduced T1D incidences in nonobese diabetic mice.⁶ However, the anticoagulant activity of PI-88 can result in excessive bleeding when used in patients with diabetes.²⁵ PI-88 also binds to platelet factor 4 protein, resulting in the generation of antibody that causes heparin-induced thrombocytopenia associated with life-threatening thrombosis.²⁶ Due to these adverse effects, there was no further follow-up of PI-88 for the treatment of individuals with diabetes.

We recently reported that HS-mimicking glycopolymer 2 (commonly known as GPM2) with 12 repeating disaccharide units (n = 12) was not only the potent heparanase inhibitor with minimal cross-bioactivity, but was also not hydrolyzed by heparanase (Figure 3).²⁷ Herein, we systematically studied the degree of polymerization (n) and the varying sulfation patterns on synthetic glycopolymers^28-32 to modulate heparanase activity. We identified glycopolymer 2 as not only the most potent inhibitor of heparanase but also improved the survival of cultured mouse pancreatic β cells under the challenge of high-level heparanase. Glycopolymer 2 also reduces expression levels of cytokines (IL8, IL1β, and TNFα) and a major macrophage inflammatory receptor (TLR2) in human pancreatic islets.

Figure 3.

HS-mimicking glycopolymer 2 (GPM2) as a potent inhibitor of heparanase.

RESULTS AND DISCUSSION

Synthesis of Glycopolymers with Varying Lengths and Well-Defined Sulfation Domains.

In the synthesis of glycopolymer 2 and its derivatives with varying degrees of polymerization and sulfation pattern, the key challenge is the stereoselective construction of the α-1,2-cis-2-amino linkage-forming disaccharide core 5 (Scheme 1) in large quantities with high levels of selectivity. Previously, the coupling of C4-hydroxyl acceptor 4 with N-phenyltrifluoroacetimidate donor 3 to form 5 (79%, α only) was achieved using 15 mol % of nickel triflate, Ni(OTf)₂, at 35 °C for 12 h.^27,33,34 We recently discovered that triflic acid released from Ni(OTf)₂ is the active catalyst to stereoselectively promote the formation of α-1,2-cis-2-amino linkages.^35-37 The triflic acid-catalyzed coupling reaction offers several advantages, including operationally simple and mild conditions, short reaction time, synthetically useful yields, and excellent diastereoselectivity. As such, we investigated the coupling of 4 with 3 catalyzed by 5 mol % triflic acid. The reaction was completed within 3 h at 25 °C providing disaccharide 5 in good yield (76%) as a single α-isomer, compatible with the previous Ni(OTf)₂-catalyzed conditions.^27,33,34

Scheme 1. Synthesis of Monomer via Nickel Triflate-Catalyzed Selective Glycosylation^a.

^aConditions and reagents: (a) 5 mol % TfOH, CH₂Cl₂, 25 °C, 3 h, 76%, α only; (b) NaOMe, MeOH; (c) HCl, acetone; (d) SO₃·Me₃N, Et₃N, DMF, 55 °C, 66%, 3 steps; (e) DDQ, PBS/CH₂Cl₂, 52%; (f) CuI, DBU, DMF, 50 °C, 64%; (g) LiOH, THF, H₂O, 94%.

With the ability to facile access to disaccharide 5 in large quantities and stereocontrol, we next employed the previously established conditions for the synthesis of monomer 8 (Scheme 1).^27,33,34 Several key observations were obtained from this synthesis: (1) the C6-acetyl group in 5 could be chemoselectively removed in the presence of the C3-acetyl group if the C2-benzylidene functionality remains intact. It is likely because the acetyl group at C3 is sterically blocked by both the hindered 2-naphthylmethyl (NAP) group at C4 and the benzylidene group at C2. This chemoselective hydrolysis also allows access to other glycopolymers with varying sulfation patterns (Figure 4, vide infra); (2) removal of the NAP group prior to the CuAAC “click” with polymerizable oxonorbornene scaffold 7 provided triazole 8 in good yield; (3) removal of the NAP group provided the desired product in good yield if the reaction was conducted using pH 7 PBS in combination with CH₂Cl₂ as a solvent. Finally, treatment of compound 8 with LiOH hydrolyzed both the carboxylate methyl ester and the acetyl group at C3 to afford monomer 9 (Scheme 1).

Figure 4.

Glycopolymers bearing varying sulfation patterns at the C6 and C3-positions of glucosamine.

Next, sulfated monomer 8 was subjected to the ring-opening metathesis polymerization (ROMP) using the third-generation of Grubbs catalyst 10 (Table 1).³⁸ By varying the amount of catalyst (9, 11.5, and 20 mol %, Table 1), the length of glycopolymers could be tailored to the desired degree of polymerization (DP) with n = 5, 9, and 12, respectively. As illustrated in Table 1, increasing the catalyst loading decreased the degree of polymerization. For the ROMP reaction to proceed smoothly with catalyst 10, the counterion for the N-sulfate group in the monomer 8 should be sodium cation (Na⁺). Further, the ratio of the solvent system, dichloroethane (DCE) and trifluoroethanol (TFE), was also screened for solubility of both monomer 8 and catalyst 10 as well as to prevent premature precipitation during the polymerization. It was determined that a 2.5:1 ratio of the DCE/TFE system was optimal for full conversion of monomer 8 to the corresponding polymers 11, 12, and 13 (Table 1). Due to its amphiphilic nature, glycopolymers 11, 12, and 13 aggregates to form micelles after polymerization. As a result, the degree of polymerization (DP) and molecular weight (M_n) of glycopolymers were determined by ¹H NMR end group analysis (Table 1) as opposed to the standard gel permeation chromatography (GPC). Glycopolymers of the same scaffold showed no variation in ¹H NMR signals, only varying in the ratio of the GlcN C1-anomeric hydrogen peak (~5.5 ppm) and the phenyl end group hydrogens (~7.4 ppm), which were used to determine the DP (see the Supporting Information for detailed information how DP of glycopolymers were calculated based on the ¹H NMR integrations). Hydrolysis of 11, 12, and 13 followed by dialysis afforded glycopolymers 2, 14, and 15, respectively (Table 1). The sulfate content of glycopolymers was determined using the turbidimetric assay (see Figure S1).³⁹

Table 1.

Polymerization of Sulfated Monomer 8^a

entry	Grubbs III 10 (mol %)	DCE/TFE	DPnb (n)	conv. (%)	yield (%)c	polymer	Mn (theoretical)	Mn (NMR)	hydrolysis yield (%)c
3	9	2.5:1	12	100	89	11	12,075	13,030	2: 82
2	11.5	2.5:1	9	100	97	12	9200	9765	14: 83
1	20	2.5:1	5	100	81	13	5425	5425	15: 80

^aThe polymerization reaction was conducted with 0.0166 mmol of glycomonomer 8 in a degassed mixture of 2.5:1 DCE/TFE with final concentration = 0.025 M.

^bDP (n) was determined by ¹H NMR end group analysis.

^cIsolated yield.

We also studied glycopolymers 16–18 with different sulfation patterns on the oxygen of C6 and C3 positions (Figure 4) to determine the effect of sulfation pattern on heparanase activity. Based on recent findings,^19,40 we hypothesize that C6-O and C3-O-SO₃⁻ groups on the GlcN unit play a critical role in HS-heparanase interaction. While glycopolymer 16 will determine whether the C6-O-SO₃⁻ group is critical for heparanase recognition, glycopolymer 17 will examine whether C6- or C3-O-SO₃⁻ group is more important. On the other hand, highly sulfated glycopolymer 18 will give a clear indication of whether additional sulfate groups could increase or decrease inhibitory activity against heparanase. Glycopolymer 19 will determine whether N-SO₃⁻ group on the glucosamine could be essential for interaction with heparanase.

In the syntheses of the disaccharide units of glycopolymers 16–19 (Scheme 2), disaccharide 5 was the core starting material. Removal of the N-benzylidene group in 5 afforded ammonium salt 20. Sequential N-sulfation, NAP group deprotection, and O-deacetylation afford N-sulfated disaccharide 21. We observed that the deacetylation step would proceed smoothly only if the counterion of the N-sulfate group was Na⁺ cation rather than Et₃NH⁺ cation. In addition, the use of Na⁺ cation reduced the elimination product that resulted from the abstraction of the acidic C5-hydrogen of the glucuronic acid (GlcA). On the other hand, disaccharide 5 could be functionalized by chemoselective 6-O-deacetylation followed by protection of the resulting C6-hydroxyl group to afford NAP ether 22. Sequential N-benzylidene removal followed by C3-O-deacetylation, O- and N-sulfation, and removal of the NAP ether groups produced disaccharide 23. For the synthesis of sulfated 25, disaccharide 5 was first modified by chemoselective C6-O-deacetylation. The N-benzylidene and C3-O-acetyl group of the intermediate 24 was sequentially deprotected. Simultaneous O- and N-sulfation followed by global removal of the NAP groups ultimately produced disaccharide 25. For the synthesis of 26, N-trifluoroacetylation of ammonium salt 20 followed by sequential O-deacetylation, O-sulfation, and NAP group deprotection led to the formation of disaccharide 26.

Scheme 2. Synthesis of Differentially Sulfated Glycopolymers^a.

^aConditions and reagents: (a) 12N HCl, acetone; (b) NaOMe, MeOH; (c) SO₃·Me₃N, Et₃N, DMF; (d) DDQ, PBS/CH₂Cl₂; (e) 7, CuI, DBU, DMF, 50 °C; (f) 9 mol % Grubbs III catalyst 10, TFE/DCE, 55 °C, <1 h; (g) LiOH, THF/H₂O; (h) NAP-Br, TBAI, Ag₂O, CH₂Cl₂; (i) (CF₃-Ac)₂O, Et₃N, DMAP, CH₂Cl₂.

With the four differentially sulfated disaccharides 21, 23, 25, and 26 available, their azide group individually underwent click reaction with the terminal alkyne of the polymerizable scaffold 7 (Scheme 2). The triazole-containing intermediates were then subjected to the polymerizable catalyst 10 in a mixture of 1,2-dichloroethane/trifluoroethanol. This solvent system increases the solubility of the triazole monomers and prevents catalyst deactivation. The solvent ratio was accordingly modified depending on the number of sulfate and hydroxyl groups present on the monomers. We chose to keep a similar degree of polymerization for all glycopolymers so that we can compare their inhibitory activity against heparanase. Accordingly, each differentially sulfated monomer individually underwent polymerization in the presence of 9 mol % catalyst 10 so that the desired glycopolymers could be generated with similar DP_n. Glycopolymers were produced within 1 h with n = 10–12 (Scheme 2). For the synthesis of polymer 19, the labile trifluoroacetyl (TFA) group was hydrolyzed after polymerization. Like previous glycopolymers, ¹H NMR end group analysis was used to determine the DP and molecular weight (M_n) of the glycopolymers. Following polymerization, the resulting glycopolymers were subjected to hydrolysis and then dialysis to ultimately lead to the formation of glycopolymers 16–19 (Scheme 2).

Evaluation of HS-like Glycopolymers as Potential Inhibitors of Heparanase.

With all glycopolymers in hand, we first examined the binding specificity of glycopolymers 2, 14, and 15 for heparanase to determine whether heparanase binds glycopolymers in a length-dependent manner. Although a fondaparinux (Arixtra) inhibition assay could be used to evaluate the capacity of glycopolymers as potent inhibitors of heparanase (Figure S2),^41,42 fondaparinux is a homogeneous heparin pentasaccharide and represents only a portion of heparin and even less of heparan sulfate. As a result, we determined the heparanase-inhibiting activity of glycopolymers utilizing a TR-FRET assay which is the combination of the time-resolved (TR) measurement of fluorescence and the technology of fluorescence resonance energy transfer (FRET).⁴³ We first evaluated the inhibitory activity of glycopolymer 2 (GPM2) with this TR-FRET assay. When the heparanase inhibition exerted by GPM2 was fitted on the saturation curve, we observed a high degree of variation. With this standard hyperbolic fit, the IC₅₀ for glycopolymer 2 is comparable to the concentration of heparanase.⁴⁴ As a result, the inhibition for GPM2 was fitted to a tight-binding equation (see the Supporting Information), which resulted in little variation from the data; an IC₅₀ value was extrapolated to be 0.10 ± 0.036 nM for GPM2 (Figures 5 and S3). The ability of heparin to inhibit heparanase activity was also measured, and its IC₅₀ value was determined to be 0.54 ± 0.028 nM. The degree of polymerization (DP_n) has a major impact on the inhibitory activity of glycopolymers against heparanase, which is consistent with previously reported glycopolymer studies.^45,46 For example, decreasing the DP_n of the glycopolymer to 9 repeating disaccharides (n = 9) decreased the binding affinity (14: IC₅₀ = 4.43 ± 0.13 nM, Figures 5 and S3). Further decrease of the DP_n resulted in lower binding affinity. For instance, glycopolymer 15 with five repeating disaccharides (n = 5) was determined to have an IC₅₀ = 7.49 ± 0.48 nM (Figures 4 and S3). To illustrate the multivalent structure of glycopolymers amplified the affinity to heparanase, the heparanase inhibition activity of monomer 9 was also assessed. Monomer 9 exhibited 1000-fold less potency (IC₅₀ = 11.59 ± 0.95 μM, Figure S3) than its corresponding glycopolymers 2, 14, and 15. Furthermore, the data illustrate that glycopolymer 2 with 12 repeating disaccharide cores has the highest inhibitory activity against heparanase. Since GPM2 inhibits heparanase in the picomolar range, it is unlikely that increasing DP_n will provide improvements. Additionally, we observed that increasing the DP to n = 18 significantly decreased the yield of the resulting glycopolymer to 40%; as such, we did not pursue larger glycopolymers with n > 12.

Figure 5.

Inhibition of heparanase by glycopolymers 2 (GPM2), 14, and 15 using TR-FRET assay. Data are shown as mean ± SEM (n = 3). ***p < 0.0005 and ****p < 0.0001.

Next, we examined the effect of glycopolymer’s sulfation domains on modulating heparanase activity. Accordingly, we evaluated the inhibitory activity of synthetic glycopolymers 16–19 using a TR-FRET assay (Figures 6 and S4). The data obtained from the TR-FRET assay indicate that the sulfation pattern of the glucosamine (GlcN) plays a critical role in modulating heparanase activity. The C6-O and C2-N positions of the GlcN must be sulfated to achieve the highest potency. For instance, removal of the C6-O-SO₃⁻ significantly decreased the potency of the glycopolymer against heparanase (16: IC₅₀ = 17.89 ± 0.954 nM). While recent study reports that heparanase can recognize the GlcN bearing either C6- or C3-O-SO₃⁻,⁴⁰ we found that glycopolymer 17 (IC₅₀ = 4.041 ± 0.156 nM) bearing C3-O-SO₃⁻ is less potent (IC₅₀ = 4.041 ± 0.156 nM) than glycopolymer 2 (IC₅₀ = 0.10 ± 0.036 nM) having C6-O-SO₃⁻. Interestingly, the incorporation of an additional sulfate to the C3-O position of the GlcN, generating glycopolymer 18 (IC₅₀ = 5.48 ± 0.31 nM), only resulted in decreasing the inhibitory activity against heparanase. This result indicates that the sulfation pattern on the GlcN unit of glycopolymers is very critical for interactions with heparanase. Finally, exchanging the N-sulfate for ammonium (19: IC₅₀ = 8.83 ± 0.52 nM) decreased the binding affinity (Figure 6). Overall, the data from this TR-FRET assay show that while C2-N-sulfate is important for the binding, it is not as important as C6-O-sulfate for heparinase recognition. Collectively, the results obtained in Figures 5 and 6 show that glycopolymer 2 (GPM2) is the most potent inhibitor of heparanase among the synthetic glycopolymers evaluated.

Figure 6.

Inhibition of heparanase by glycopolymers 16–19 as well as glycopolymer 2 (GPM2) using a time-resolved fluorescence resonance energy transfer (TR-FRET) assay. Data are shown as mean ± SEM (n = 3). *p < 0.05, ***p < 0.0005, and ****p < 0.0001.

Evaluation of Cross-Bioactivity of Glycopolymer 2.

Since glycopolymer 2 (GPM2) is the most potent heparanase inhibitor, we next examined its cross-bioactivity with particular heparin-binding proteins ATIII and PF4. Although heparin is a potent inhibitor of heparanase, its anticoagulant activity increases the risk of bleeding when it is used for the treatment of cancer and diabetes.⁴⁷ Heparin interacts with antithrombin III (ATIII) to induce a conformational change of ATIII, resulting in potentiation of the anticoagulant effect of ATIII protein.⁴⁸ The activated ATIII then inactivates serine proteases, factor Xa (FXa), and FIIa. As such, we evaluated the anticoagulant activity of glycopolymer 2 (GPM2) compared to low-molecular-weight heparin (LWMH, MW = 3 kDa) and heparin (18 kDa) using chromogenic substrate assays.⁴⁸ GPM2 exhibited no anticoagulant activity against proteases (FXa: IC₅₀ > 4500 nM and FIIa: IC₅₀ > 4500 nM).²⁷ In contrast, heparin exhibited high levels of anticoagulant activity against FIIa (IC₅₀ = 4.63 ± 0.22 nM) and LMWH for FXa (IC₅₀ = 266 ± 11.5 nM).

Next, a solution-based biolayer interferometry (BLI) competition assay was employed to assess the ability of glycopolymer 2 to complete with heparin attached to the BLI biosensor in binding to platelet factor 4 (PF4) (Figure 7).⁴⁹ It has been reported that about 3% of patients who were treated with heparin experienced heparin-induced thrombocytopenia (HIT) after five or more days of treatment.⁴⁷ This side effect is difficult to eliminate because the methods used to obtain heparin are not as refined as they are for many other drugs. Interaction of heparin with PF4 triggers an autoimmune response to promote HIT.⁵⁰ Thrombocytopenia is a risk factor of thrombotic complications,⁵¹ the primary reason why heparin-based heparanase inhibitors were suspended at clinical trials.⁵² For instance, in a Phase I clinical study, three out of 42 patients developed HIT due to the formation of a PF4-PI-88 complex.²⁶ As such, we evaluated the ability of heparin and GPM2 to form a complex with PF4 using the BLI competition assay (Figure 7). In this assay, we measured the response for the leftover PF4 interacting with immobilized heparin. As such, we measured the % activity of the leftover PF4 and used Graphpad Prism to calculate IC₅₀ values. Heparin exhibited a markedly low IC₅₀ value in nanomolar concentration (1.46 ± 0.09 nM), suggesting excellent binding strength to PF4. This result is consistent with the observation that a significant number of patients experience HIT when treated with heparin.⁵⁰ In comparison, GPM2 binding to PF4 showed nearly 1000-fold lower affinity, IC₅₀ = 1290 ± 0.44 nM. GPM2 possesses a significantly lower binding affinity to PF4 than that of heparin (Figure 7), partly due to heparin having multiple iduronic acid (IdoA) residues while glycopolymer 2 does not contain any IdoA units. Studies show that the highly sulfated IdoA-GlcN region of heparin is responsible for binding to PF4.⁵³

Figure 7.

Binding affinity interactions of PF4 with heparin and glycopolymer 2 (GPM2) using solution-based biolayer interferometry (BLI) competition assay.

Glycopolymer 2 Protects Pancreatic β Cells and Suppresses Islet Inflammation.

β Cells in the pancreas contain high levels of intra-islet HS that is cleaved by heparanase in T1D patients.⁵⁴ In T1D, increased levels of heparanase represent a major challenge to β-cell function and survival through degradation of HS contents.⁶ The β cells rely on extracellular and intracellular HS for survival, with HS protecting them from migrating immune cell attack and/or reactive oxygen species (ROS)-caused damage. As such, we examined the effect of glycopolymer 2 (GPM2) to protect β cells in the pancreas under the challenge of heparanase. Accordingly, mouse pancreatic Min-6 β-cell line was treated with vehicle as a control, heparanase (5 ng/mL) alone, heparanase (5 ng/mL) plus GPM2 (0.3 μM), or GPM2 alone (0.3 μM) for 24 h (Figure 8). The morphology of the cells after the treatments was then recorded by B/W phase microscopy, and the survived viable cells were counted. The results demonstrate that treatment with heparanase markedly decreased the survival of cultured mouse pancreatic β cells, as indicated by the cell number and morphology following treatment of the Min-6 β cells with heparanase (Hpse) vs vehicle control (PBS) (Figure 8). In contrast, β cells treated with heparanase plus GPM2 revealed survival rates comparable to β cells treated with vehicle alone. Further, the heparanase-treated β cells grew to a lower density and exhibited irregular shape compared to the higher-density and islet-like colonies formed by Min-6 β cells treated with the vehicle control or with heparanase plus GPM2. Moreover, treatment with GPM2 alone had no toxic effect on mouse β-cell growth or colony formation.

Figure 8.

Mouse pancreatic Min-6 β cells were treated with vehicle control (PBS), heparanase only (5 ng/mL), heparanase plus GPM2 (0.3 μM), or GPM2 (0.3 μM) alone for 24 h. Morphology and number of cells that survived after the treatments were determined. The experiment was repeated three times, and the representative images are shown. Survival of cells (cells per mL) was determined by the Trypan Blue exclusion test of cell viability. Data are shown as mean ± SEM (n = 5). * The p value is <0.05 for Hpse vs veh.

Next, the MTT cell proliferation assay was conducted to quantitatively determine cell viability in the presence of the glycopolymer 2 (GPM2). While heparanase treatment significantly reduced the Min-6 cell proliferation/viability compared to vehicle control, treatment of GPM2 protected the Min-6 β cells from heparanase-caused reduction in cell viability as illustrated in Figure 9. Furthermore, treatment of Min-6 cells with GPM2 alone exhibited no toxic effect on Min-6 cell proliferation or viability. Together, these results confirmed the protective effect of GPM2 on mouse pancreatic β cell survival in the presence of the heparanase challenge.

Figure 9.

MTT cell proliferation viability. Min-6 cells were treated with vehicle (PBS), heparanase (5 ng/mL), heparanase (5 ng/mL) plus GPM2 (0.3 μM), or GPM2 (0.3 μM) alone for 24 h, followed by MTT assay to determine cell proliferation viability. The viability/proliferation activity of Min-6 cells treated with the vehicle was used as the control to determine the degree of cell viability of all of the groups, shown by % of cell viability. The average of all of the controls was set as 100%, which was used to calculate % of cell viability for all of the other treatment groups. Data are shown as mean ± SEM (n = 6). ** The p value is <0.01 for Hpse vs veh. Heparanase = hpse and vehicle = veh.

We also examined the expression of genes encoding major regulators in mitochondrial stress response, including m-AAA protease 1 (OMA1) and dynamin-related protein 1 (DRP1),⁵⁵ in the pancreatic Min-6 β cells treated with heparanase or with heparanase plus glycopolymer 2 (Figure 10). Although GPM2 did not significantly reduce the expression of genes encoding OMA1 (Figure 10A) or DRP1 (Figure 10B) when incubated with heparanase (Hpse), incubation of Min-6 β cells with glycopolymer 2 (GPM2) alone significantly suppressed expression of both OMA1 and DRP1 (Figure 10). As for the expression profiles of mitochondrial stress genes, OMA1 and Drp1, we performed statistical significance for the effects of GPM2 in the absence or presence of Hpse. However, p values vary from significance (*p < 0.05) to nonsignificance (ns). It is possible that the Hpse treatment for 24 h did not significantly affect intracellular stress response, although it did change the extracellular HS contents in the cellular matrix. Nevertheless, the data indicate a potential role of glycopolymer 2 (GPM2) to protect Min-6 β cells from the damages caused by the mitochondrial stress response.

Figure 10.

Mouse pancreatic Min-6 cells were treated with vehicle control (PBS), heparanase (5 ng/mL), heparanase plus GPM2 (0.3 μM), or GPM2 (0.3 μM) alone for 24 h. Levels of the transcripts encoding the mitochondrial stress response mediators OMA1 (A) and Drp1 (B) in pancreatic Min-6 β cells treated with vehicle, heparanase, and/or GPM2 were analyzed by quantitative real-time PCR (qPCR). Mean ± SEM (n = 3 biological replicates), * The p value is <0.05 for GPM2 vs veh and GPM2 vs Hpse.

To further evaluate the protective effect of glycopolymer 2 (GPM2) against heparanase-caused damage, we treated ex vivo cultured, insulin-producing human pancreatic islets with vehicle or heparanase (10 ng/mL), heparanase (10 ng/mL) plus GPM2 (0.3 μM), or GPM2 (0.3 μM) alone (Figure 11). Human pancreatic islets have high heparan sulfate contents. As a result, the control was stained because Alcian blue detected the negatively charge HS polysaccharide. Administration of heparanase to human pancreatic islets resulted in no Alcian blue staining because heparanase degrades HS into small oligosaccharides which cannot be detected by Alican stain. Next, human pancreatic islets were treated with GPM2 and heparanase. Prior to staining with Alcian blue, pancreatic islets were washed thoroughly with NaCl to remove excess GPM2. The result showed that administration of GPM2 and heparanase to pancreatic islets restored Alcian blue staining. Since GPM2 is a negatively charged glycopolymer, it is likely to be stained by Alcian blue. Therefore, human pancreatic islets were treated with GPM2 alone. The result showed that GPM2 was more stained than HS as islets incubated with GPM2 alone exhibited high levels of Alcian blue stain compared to islets incubated with the PBS vehicle. Collectively, the result, when islets incubated with heparanase and GPM2 showed more Alcian blue stain than the control, indicates that glycopolymer 2 likely interacts with pancreatic islets and protects them from degradation and destruction by heparanase.

Figure 11.

Human pancreatic islets were treated with vehicle (PBS), heparanase (10 ng/mL), heparanase (Hpse) plus GPM2 (0.3 μM), or GPM2 (0.3 μM) alone for 24 h. Heparan sulfate in human pancreatic islets was stained by Alcian blue histochemistry.

To further validate the effect of glycopolymer 2 (GPM2) on protecting human pancreatic islets from degradation and destruction by heparanase, human pancreatic islets were treated with PBS, heparanase (Hpse), heparanase plus GPM2, or GPM2 alone, and then incubated with an FITC-conjugated anti-heparan sulfate antibody (10E4) to label intact heparan sulfate contents remained in human islets with fluorescence (Figure 12A). The 10E4 antibody is commonly used to detect heparan sulfate as it recognizes a common glucosamine epitope on heparan sulfate.^56,57 As shown by heparan sulfate-specific fluorescent imaging, the pancreatic islets treated with heparanase lost heparan sulfate contents on the surface and exhibited smaller sizes and irregular shapes, compared to the islets treated with vehicle, heparanase plus GPM2, or GPM2 alone (Figure 12A). Quantitative analysis of the fluorescent signals indicated that, while the heparanase treatment eliminated heparan sulfate contents on the surface of human pancreatic islets, the treatment of glycopolymer 2 (GPM2) effectively preserved heparan sulfate contents of the islets under the heparanase challenge (Figure 12B).

Figure 12.

Human pancreatic islets were treated with vehicle, heparanase (10 ng/mL), heparanase (Hpse) plus GPM2 (0.3 μM), or GPM2 (0.3 μM) alone for 24 h. (A) The islets were incubated with an FITC-conjugated monoclonal anti-heparan sulfate antibody to label heparan sulfate contents with fluorescence. Images and fluorescent signals of the stained islets were obtained using a confocal fluorescent microscopy reader. (B) Heparan sulfate fluorescent intensities were quantified with ImageJ software. The average fluorescent intensity of all of the vehicle controls was set as 100%, which was used to calculate % of heparan sulfate contents for all of the treatment groups. Data are shown as mean ± SEM (n = 5 or 6). ** The p value is <0.01 for Hpse vs veh or Hpse+GPM2 vs Hpse.

Next, we evaluated the effect of glycopolymer 2 (GPM2) on islet inflammatory response associated with the challenge of heparanase. Accordingly, expression levels of the genes encoding major pro-inflammatory molecules were examined (Figure 13). Quantitative real-time PCR (qPCR) analysis shows that expression levels of TNFα, IL1β, and IL8 in human islets treated with GPM2 were modestly reduced compared to islets incubated with the vehicle, in the presence or absence of the heparanase challenge (Figure 13). The expression of Toll-like Receptor 2 (TLR2), a major macrophage pro-inflammatory receptor, was not significantly decreased in response to GPM2 (Figure 13). The effects of GPM2 on suppressing the expression of IL8, IL-1β, TNFα, and TLR2 in human islets change from statistical significance (*p < 0.05) to nonsignificance (ns) regardless of whether heparanase is present or absent. As GPM2 protects HS contents in the islet extracellular matrix, it is possible that intracellular inflammatory pathways, mediated by individual inflammatory molecules, may differentially respond to the heparanase challenge or GPM2 treatment. For the future study, we will perform high-throughput approaches, such as RNA-sequencing and proteomics analyses, to profile the specific stress response or inflammatory pathways modulated by GPM2 in pancreas islets in the presence or absence of heparanase challenge. Nevertheless, the data suggest that GPM2 may play a role in suppressing islet inflammatory response under the challenge of high-level heparanase.

Figure 13.

Human pancreas islets were treated with vehicle (PBS), heparanase (10 ng/mL), heparanase plus GPM2 (0.3 μM), or GPM2 alone for 24 h. Levels of the transcripts encoding the inflammatory molecules IL-8, IL-1β, TNFα, and TLR2 in human islets treated with vehicle, heparanase, and/or GPM2 were determined by quantitative real-time PCR (qPCR). Mean ± SEM (n = 3 biological replicates). * The p value is <0.05.

CONCLUSIONS

We have shown that the HS-mimicking glycopolymer of 12 repeating sulfated disaccharide units is a potent heparanase inhibitor. This glycopolymer 2 (GPM2) also has low affinity to serine proteases in the coagulation cascade and platelet factor 4 compared to heparin. Further, this glycopolymer improved the survival of cultured mouse pancreatic β cells under the heparanase challenge. Our data illustrate that treatment with the glycopolymer alone did not show any toxic effect on mouse β-cell growth and colony formation. The HS-like glycopolymer 2 also protects ex vivo cultured, insulin-producing human islets from damage by heparanase. Expression levels of inflammatory molecules, including IL8, IL-1β, TNFα, and TLR2, in human pancreatic islets treated with GPM2, were reduced. Based on our initial investigation of the effects of HS-mimicking glycopolymer 2, we envision that further tailoring its structures could lead to the discovery of new heparanase inhibitors that could be highly effective at protecting pancreatic β cell destruction and suppressing inflammatory cytokine expression in human islets under the challenge of heparanase.

MATERIALS AND METHODS

Glycopolymer Formation.^27,33

To a 10 mL Schlenk flask were added synthetic disaccharide-containing oxonorbornene scaffold and a mixture of 2,2,2-trifluoroethanol and 1,2-dichloroethane solvent under an inert atmosphere. The resulting solution was stirred at room temperature for 5 min, and a solution of the third-generation Grubbs catalyst 10 was then added. The resulting mixture was stirred at 55 °C for 1 h. Once ¹H NMR indicated that the polymerization was completed, the reaction mixture was diluted with methanol and then triturated by diethyl ether to give the desired glycopolymer. The protected glycopolymer was treated with LiOH in a mixture of water and THF for 24 h. The crude glycopolymer was dialyzed (3.5K MWCO) against 0.9% NaCl solution (3 buffer changes) and deionized water (3 buffer changes).

Biolayer Interferometry Assay.

This competition assay was performed on an Octet Red Instrument at 25 °C. Immobilization and binding analysis were conducted at 1000 rpm using HBS-EP buffer, which consists of 10 mM HEPES, pH 7.4, 150 mM NaCl, 3.0 mM EDTA, and 0.005% (v/v) surfactant Tween 20. Affinities of ligands in this BLI assay were determined based on SPR analysis.⁴⁹ In this BLI assay, platelet factor 4 (35 nM) is mixed with various concentrations of 18 kDa heparin and glycopolymer 2 (GPM2). Free platelet factor 4 protein in this mixture is evaluated for binding to immobilized heparin (platelet factor 4 protein was purchased from R&D Systems). Biotinylated heparin (18 kDa, 5 μg/mL) was attached to streptavidin biosensors for 5 min. Binding experiments were performed under mass transport conditions. The binding response of the mixtures to immobilized heparin was measured by Octet Red Instrument. The binding response decreases as glycopolymer 2 (GPM2) or heparin compete with immobilized heparin to bind PF4, resulting in signal reduction in a concentration-dependent manner. The assay was repeated twice for each ligand.

Fondaparinux (Arixtra) Heparanase Inhibition Assay.⁴¹

This colorimetric assay was initially carried out to evaluate the inhibitory effect of glycopolymer 2 (GPM2) against heparanase. This assay is based on the degradation of pentasaccharide fondaparinux by heparanase. This assay measures the appearance of the disaccharide hemiacetal product, which is produced from the cleavage of fondaparinux by heparanase and colorimetrically reacts with the tetrazolium salt WST-1. UV absorbance of the disaccharide-WST-1 complex is measured at 584 nm. The assay solution (100 μL) consists of pH 5 sodium acetate buffer (40 mM), fondaparinux (100 mM), and with or without increasing concentrations of glycopolymer 2 (GPM2). Recombinant heparanase with a final concentration of 140 pM is then added to the assay solution. The plates are incubated at 37 °C for 18 h and the reaction is stopped by the addition of 100 μL of a solution containing 1.69 mM WST-1 in 0.1 M NaOH. The plates are developed at 60 °C for 1 h, and the absorbance (OD) is measured at 584 nm. This fondaparinux experiment was repeated twice.

TR-FRET Heparanase Inhibition Assay.

This assay is based on energy transfer between a donor [Eu(K)] and an acceptor (Streptavidin-XL665). Degradation of biotin-HS-Eu(K) by heparanase enzyme prevents energy transfer from a donor fluorophore to an acceptor fluorophore, thus resulting in no FRET signal. On the other hand, if glycopolymers present as heparanase inhibitors, biotin-HS-Eu(K) remains intact resulting in high FRET signal. In this assay, microtubes were charged with a solution of glycopolymer (42 μL) in water (0.00016–4000 μM) or water only as a control water and a solution of heparanase (42 μL, 5.3 nM) in pH 7.5 Tris buffer (consisting of 20 mM Tris HCl, 0.15 M NaCl, and 0.1% CHAPS) or pH 7.5 Tris buffer as blank. The total concentration of heparanase was 0.5 nM. The resulting mixture was preincubated at 37 °C for 10 min. A solution of 84 μL of biotin-HS-europium cryptate (58.6 ng in pH 5.5 0.2 M NaCH₃CO₂ buffer) was then added to the mixture. The resulting mixture was incubated for 1 h at 37 °C. A solution of 168 μL of Streptavidin-XLent! (1.0 μg/mL) in pH 7.5 buffer (made of 0.1 M NaPO₄, 0.8 M KF, 0.1% BSA) was added to the reaction mixture. The mixture was stirred at room temperature for 15 min. A 100 μL (per well) solution of the reaction mixture was transferred to a 96-well microplate in triplicate. Homogeneous time-resolved fluorescence emissions at 616 and 665 nm were measured at 340 nm using SpectraMax i3x Microplate Reader. This TR-FRET assay was repeated three times. In the TR-FRET assay, suramin was used as a positive control. The IC₅₀ value of suramin was determined to be 5–10 μM against heparanase.

Two-Stage Chromogenic Anticoagulation Assay to Evaluate FXa and FIIa Activity.

In this assay, antithrombin inhibition was measured by monitoring the cleavage of 4-nitroaniline at its absorbance at 405 nm from specific peptide substrates.⁴⁸

Factor Xa Activity.

Low-molecular-weight heparin (LMWH) or GPM2 with different concentrations (0.0002–166.7 μM; 40 μL) and antithrombin III (0.04 IU; 40 μL) were added to a deep-well block (Nunc 96 DeepWell 1.0 mL/well, clear). The mixture was mixed and incubated at 37 °C for 2 min. Protease FXa (0.32 μg; 40 μL) was then added to the mixture. The resulting mixture was incubated at 37 °C for another 2 min (Stage 1). FXa-specific chromogenic substrate (0.048 mmol; 40 μL) was added to the mixture. After the reaction mixture had been stirred at room temperature for 2 min, citric acid (240 μL; 20 g/L) was then added. A 100 μL solution of the reaction mixture was then transferred to a 96-well microplate in triplicate, and absorbance at 405 nm was measured with a SpectraMax i3x Microplate Reader.

Factor IIa Activity.

Heparin (0.0023–2.4 μM; 40 μL, 18 kDa) or GPM2 with different concentrations (90–1.4 μM; 40 μL) and antithrombin III (0.01 IU; 40 μL) were added to a deep-well block (Nunc 96 DeepWell 1.0 mL/well, clear). The mixture was mixed and incubated at 37 °C for 2 min. Protease FIIa (1.2 nkat; 40 μL) was then added to the mixture. The resulting mixture was incubated at 37 °C for 2 min (Stage 1). FIIa-specific chromogenic substrate (0.048 mmol; 40 μL) was then added to the mixture. After the reaction mixture had been stirred at room temperature for 2 min, citric acid (240 μL; 20 g/L) was then added. A 100 μL solution of the reaction mixture was then transferred to a 96-well microplate in triplicate, and absorbance at 405 nm was measured with SpectraMax i3x Microplate Reader.

Mouse Pancreatic β Cell Line Min-6 Culture and Treatment.

Mouse pancreatic β cell line (insulinoma) Min-6 was cultured as we previously described.⁵⁸ Specifically, Min-6 cells were cultured with high-glucose DMEM supplemented with 10% fetal bovine serum (FBS), 50 μM β-mercaptoethanol, and antibiotics. Min-6 cells were cultured in six-well plates or seeded in four-well slide chambers for gene expression and immunofluorescence analyses, respectively. When the cells reached approximately 50% confluency, they were treated with vehicle (PBS), heparanase (5 ng/mL), heparanase plus GPM2 (0.3 μM), or GPM2 (0.3 μM) alone for 24 h. Morphology and the number of cells that survived after the treatments were determined. The experiment was repeated three times. Cell survival rates were evaluated by counting living and death cells distinguished by Trypan Blue staining. For each treatment, living cells in three to five random areas were counted, and mean values were calculated as the survival cell counts.

MTT Cell Proliferation Viability Assay.

Mouse pancreatic β cell line Min-6 were cultured in 96-well plates and then treated with vehicle (PBS), heparanase (5 ng/mL), heparanase (5 ng/mL) plus GPM2 (0.3 μM), or GPM2 (0.3 μM) for 24 h. After the treatments, 10 μL of MTT labeling reagent (0.5 mg/mL, Cayman Chemical, MI) was added into the wells, and the cells were incubated for 4 h at 37 °C in a CO₂ incubator, followed by the addition of 100 μL of solubilization solution and incubation for 18 h. The absorbance of each sample at 570 nm, which represents cell viable/proliferation activity, was measured using a microplate reader. The viability proliferation activity of Min-6 cells treated with the vehicle was used as the control to determine the degree of cell viability in response to heparanase treatment in the presence or absence of GPM2, shown by % of cell viability. The average of all of the controls was set as 100%, which was used to calculate % of cell viability for all of the other treatments including the vehicle controls.

Quantitative Real-Time PCR (qPCR).

Total RNAs from human pancreatic islets were extracted utilizing Trizol according to the manufacturer’s instructions. Single-stranded cDNA was prepared from total RNA (50 ng) using High Capacity cDNA Reverse Transcription Kit. The total mRNA content was measured by quantitative real-time PCR using the SYBR Green PCR Master mix. The sequences of the primers for human transcripts encoding: IL-8: 5′ - T C T G C A G C T C T G T G T G A A G G - 3′ and 5 ′ -TGGGGTGGAAAGGTTTGGAG-3′; IL-1β: 5′-GCTCGCCAGT-GAAATGATGG-3′ and 5′-GTCCTGGAAGGAGCACTTCAT-3′; TNFα: 5′-CCCATCTATCTGGGAGGGGT-3′ and 5′-ATCCC-A A A G T A G A C C T G C C C - 3 ′; TLR2: 5′-C T G T G C T C T G T T C C T G C T G A - 3 ′ and 5 ′- GAGCTTTCCTGGGCTTCCTT-3′. The sequences of the primers for mouse transcripts encoding: OMA1: 5 ′-TGTACTGTGCGGGGTGATTC-3′; and 5′-GGCCTCTAT-T A C G G C A G GA C - 3 ′; and Drp1 : 5 ′ -C C A G A G - GAACTGGTGTGGTC-3′ and 5′-ACTCTCCAGCTCTGAAATTTACCA-3′.

Ex Vivo Cultured, Insulin-Producing Human Pancreatic Islets.

Healthy human islets were isolated using the Prodo Labs protocol (https://prodolabs.com/). The insulin-secreting human islets were cultured in PIM medium supplemented with 5% human AB serum and 1% glutamine/glutathione as previously described.⁵⁹

Treatment of Human Islet with Heparanase and/or Glycopolymer.

Human pancreatic islets were treated with vehicle (PBS), recombinant human active heparanase (10 μg/mL), heparanase plus GPM2 (300 nM), or GPM2 alone for 24 h. Human pancreatic islets were then subjected to biochemical, gene expression, and histological analyses.

Alcian Blue and Immunofluorescent Staining of Heparan Sulfate Contents in Human Islets.

Isolated human pancreatic islets were stained with 1% Alcian blue in HCl for 10 s. Counterstaining was performed for 1 min with nuclear-fast red. To quantitatively analyze intact heparan sulfate contents in human islets via fluorescent staining, isolated human islets were blocked with 5% normal donkey serum in TBST buffer containing 0.1% Tween 20, 2.4 g/L Tris base, and 8 g/L NaCl, followed by incubation with an FITC-conjugated monoclonal anti-heparan sulfate antibody (10E4 epitope, US Biological, Inc.) in 1:50 dilution overnight at 4 °C. The islets were washed three times with TBST wash buffer containing Tween 20 (0.1%) and NaCl (8 g/L). Images and fluorescent signals of the stained islets were obtained using an Agilent BioTek confocal imaging reader (Agilent Technologies, Santa Clara, CA). ImageJ software was utilized to quantitatively analyze heparan sulfate fluorescent intensity. The fluorescent intensity of vehicle PBS-treated islets was used as the control to determine the percentage of heparan sulfate content that remained in the islets after the treatments. The average of all of the controls was set as 100%, which was used to calculate % of heparan sulfate contents for all of the samples including the controls.

Statistics.

Experimental data are illustrated as mean ± SEM. All in vitro experiments were repeated at least three times. Mean values for biochemical data from the experimental groups were compared by unpaired two-tailed Student’s t tests, unless otherwise indicated. A value of p < 0.05 was considered statistically significant.