Sulfated Hyaluronan Alters the Interaction Profile of TIMP-3 with the

Sep 9, 2016 - Sulfated Hyaluronan Alters the Interaction Profile of TIMP-3 with the Endocytic Receptor LRP-1 Clusters II and IV and Increases the ...
1 downloads 0 Views 7MB Size
Article pubs.acs.org/Biomac

Sulfated Hyaluronan Alters the Interaction Profile of TIMP‑3 with the Endocytic Receptor LRP‑1 Clusters II and IV and Increases the Extracellular TIMP‑3 Level of Human Bone Marrow Stromal Cells Sandra Rother,† Sergey A. Samsonov,‡ Ute Hempel,§ Sarah Vogel,§ Stephanie Moeller,∥ Joanna Blaszkiewicz,⊥,# Sebastian Köhling,⊥,# Matthias Schnabelrauch,∥ Jörg Rademann,⊥,# M. Teresa Pisabarro,‡ Vera Hintze,† and Dieter Scharnweber*,† †

Institute of Materials Science, Max Bergmann Center of Biomaterials, TU Dresden, Budapester Str. 27, 01069 Dresden, Germany Structural Bioinformatics, BIOTEC TU Dresden, Tatzberg 47-51, 01307 Dresden, Germany § Institute of Physiological Chemistry, Carl Gustav Carus Faculty of Medicine, TU Dresden, Fiedlerstraße 42, 01307 Dresden, Germany ∥ Biomaterials Department, INNOVENT e.V., Prüssingstraße 27 B, 07745 Jena, Germany ⊥ Institute of Pharmacy & Institute of Chemistry and Biochemistry, Freie Universität Berlin, Königin-Luise-Str. 2, 14195 Berlin, Germany # Institute of Medical Physics and Biophysics, Universität Leipzig, Härtelstr. 16/18, 04107 Leipzig, Germany ‡

ABSTRACT: Sulfated glycosaminoglycans (sGAGs) modulate cellular processes via their interaction with extracellular matrix (ECM) proteins. We revealed a direct binding of tissue inhibitor of metalloproteinase-3 (TIMP-3) to the endocytic receptor low-density lipoprotein receptor-related protein (LRP-1) clusters II and IV using surface plasmon resonance. Sulfated hyaluronan (sHA) and chondroitin sulfate (sCS) derivatives interfered with TIMP-3/LRP-1 complex formation in a sulfation-dependent manner stronger than heparin. Electrostatic potential calculations suggested a competition between negatively charged GAGs and highly negatively charged complement-like domains of LRP-1 for the binding to a positively charged area of TIMP-3 as an underlying mechanism. In vitro studies revealed increased amounts of pericellular TIMP-3 in the presence of sHA as a consequence of the blocked protein uptake. GAG derivatives as part of biomaterials might post-translationally modulate TIMP-3 levels stronger than native GAGs, thus exhibiting catabolic effects on the ECM, which could prevent extensive pathological matrix degradation and promote wound healing.



INTRODUCTION To foster the development of new therapies toward patientspecific needs, a detailed understanding of the complex mechanisms regulating tissue homeostasis is required. The extracellular matrix (ECM), a complex meshwork surrounding cells, comprises several major components: structural proteins, glycoproteins, and proteoglycans with glycosaminoglycans (GAGs) as their functional parts.1 GAGs have various ECMrelated functions including the sequestering of growth factors and chemokines and the interaction with other matrix components and thereby play an important role in numerous signaling pathways.2 The ECM composition is highly dynamic because it underlies a continuous remodeling. Proteolytic enzymes such as matrix metalloproteinases (MMPs) are mainly responsible for the degradation of a broad range of ECM components.3 In particular, during wound healing, the activities of MMPs are crucial because they control processes, for example, cell migration. However, the activity of proteases requires strict regulatory mechanisms. Tissue inhibitors of © XXXX American Chemical Society

metalloproteinases (TIMPs) act as their native counterbalance, preventing an exceeded matrix-turnover.4 TIMP-3 is the only member of the TIMP family that is sequestered to the ECM by binding to sulfated GAGs, indicating its role as a major regulator of tissue homeostasis.5 An imbalance between MMPs and TIMPs is often associated with impaired wound healing such as chronic foot ulcers.6 Depending on their carbohydrate backbone, glycosidic linkage, and degree and pattern of sulfate residues, GAGs are likely to be involved in the alteration of mediator protein activities (e.g., cytokines, growth factors, and enzymes).7 In particular, sulfated hyaluronan derivatives (sHAs) are interesting model molecules to study the structure−function relationship of GAGs in terms of their interaction with mediator proteins and their corresponding receptors because they, in contrast with native GAGs, display Received: July 1, 2016 Revised: September 7, 2016

A

DOI: 10.1021/acs.biomac.6b00980 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules

of functional biomaterials to inhibit extensive tissue degradation.

defined sulfation degrees, patterns, and chain lengths with minimal batch-to-batch variabilities.8 Thus they are more appropriate for future clinical applications. Recent studies revealed a binding of TIMP-3 to native sulfated GAGs (sGAGs) like chondroitin sulfate E (CS E), heparin (HEP), and heparan sulfate (HS).9 Considering the broad spectrum of TIMP-3 functions in the ECM, the regulation of extracellular TIMP-3 levels is of great physiological relevance. TIMP-3 was recently reported to be a ligand of the endocytic receptor low-density lipoprotein receptor-related protein-1 (LRP-1),10 while the exact binding site of TIMP-3 to LRP-1 is unknown. The LRP-1 is a multifunctional cell-surface receptor widely expressed on several cell types like fibroblasts, keratinocytes, and macrophages.11 It is continuously involved in repeated cycles of ligand uptake with recycling times of ∼12 min.12 LRP-1 consists of a 515 kDa alpha-chain containing the ligand binding domains and an 85 kDa transmembrane beta-chain, which are noncovalently associated. The alpha-chain of this receptor binds a high number of structurally unrelated ligands such as proteinases, proteinase-inhibitor complexes, ECM proteins, and growth factors, thereby influencing biological processes like cell growth and migration.11 Ligand binding occurs at discrete regions of LRP-1 consisting of clusters of small (∼40 residues) Ca2+-binding domains named complement-like (CR) domains.13 The alpha-chain of LRP-1 is composed of four such cysteine-rich clusters termed I, II, III, and IV with 2, 8, 10, and 11 CR domains.14,15 The intracellular chaperone receptorassociated protein (RAP) competes with all known LRP-1 ligands, thereby preventing premature ligand binding. For RAP, a binding to complement-like repeats 7 (CR7) and 8 (CR8) of LRP-1 ligand binding cluster II was revealed.16,17 The subdomains CR3-CR7 of cluster II are reported to be essential for ligand binding in general.18 Previous findings on the influence of sHA on osteoblastderived matrix vesicles revealed a down-regulation of the LRP-1 protein amount compared with CS-treated matrix vesicles, suggesting a potential impact of sHA derivatives on TIMP-3 levels and LRP-1 function.19 This is in accordance with Troeberg et al., reporting an inhibition of TIMP-3 binding to the LRP-1 ectodomain in the presence of HS and CS E.9 However, little is known about the molecular details and the structural requirements of GAGs for establishing these interactions. Extensive ECM degradation and thus impaired tissue functions often occur during pathological conditions such as osteoarthritis, atherosclerosis, and cancer or in chronic wounds.9,20,21 Therefore, detailed analyses regarding the regulatory systems controlling the MMP/TIMP balance and thereby ECM stability and integrity are mandatory for the development of suitable strategies for the treatment of these conditions. The goal of this work was to elucidate the impact of sGAG derivatives in comparison with native GAGs on TIMP-3 homeostasis. Surface plasmon resonance (SPR) studies and molecular modeling techniques were used to determine whether TIMP-3 is able to directly bind to distinct LRP-1 clusters and to investigate how sGAGs could interfere with this complex formation. To reveal the molecular mechanism in detail, HA tetrasaccharides were examined. Mesenchymal stromal cells (hMSCs) were used to clarify if these findings result in modulated TIMP-3 levels in vitro, which could be favorable for further biomedical applications of sGAGs as part



MATERIALS AND METHODS

Materials. Native HA (from Streptococcus, Mw = 1100 kDa) was obtained from Aqua Biochem (Dessau, Germany), HEP (from porcine mucosa, Mw = 18 kDa), and biochemical reagents were from SigmaAldrich (Schnelldorf, Germany). CS (from porcine trachea, Mw = 20 kDa), 30% chondroitin-6-sulfate, 70% chondroitin-4-sulfate) was from Kraeber (Ellerbek, Germany). Recombinant human TIMP-3 (973TM-010), recombinant human LRP-1 cluster II/Fc chimera (2368-L2050; a disulfide linked homodimer; 68.1 kDa), as well as recombinant human LRP-1 cluster IV/Fc chimera (5395-L4-050; a disulfide linked homodimer; 76.7 kDa) were obtained from R&D Systems (Wiesbaden-Nordenstadt, Germany). Series S Sensor Chips CM5 were from GE Healthcare Europe (Freiburg, Germany). Preparation and Characterization of Polymeric and Oligomeric GAG Derivatives. Polymeric and oligomeric HA and CS derivatives were synthesized, purified, and characterized as previously described.22−24 The respective degree of sulfation (D.S.) determined by elemental analysis, the weight-average (Mw) molecular weight analyzed by gel permeation chromatography with laser light scattering detection, the degree of polymerization (dp), and the sulfate group distribution detected by nuclear magnetic resonance of the used GAG derivatives are summarized in Table 1. Figure 1 displays the repeating units of the utilized polymeric HA derivatives.

Table 1. Analytical Data of GAGs sample HA-48 kDa HA-88 kDa sHA1-31 kDa sHA1-42 kDa sHA3 CS sCS3 HEP HA (dp 4) sHA1 (dp 4) psHA (dp 4)

D.S.

1.0 1.5 2.9 0.8 3.1 2.2 1.0 4.0

Mw (Da)

dp

48 255 88 449 31 056 42 430 20 950 19 763 19 915 18 000 777 1005 1763

240 441 123 153 60 82 56 58 4 4 4

sulfate group distribution

6 62′3′ 462′, 463′, 62′3′ 4, 6 462′, 463′, 62′3′ 2′6, 2N, 6 6 462′3′

SPR Analysis of TIMP-3/GAG Complex Binding to LRP-1 Clusters II and IV. Binding analyses were performed using a Biacore T100 instrument (GE Healthcare, Freiburg, Germany) with TIMP buffer (50 mM Tris, 10 mM CaCl2, 150 mM NaCl, 0.05% Brij-L23, pH 7.5) according to ref 9 at 37 °C. LRP-1 clusters II and IV were immobilized onto Series S Sensor Chips CM5 via amine coupling at 25 °C as described by GE Healthcare, resulting in immobilization values of 500 RU or 380 RU, respectively. An activated, and afterward deactivated, flow cell without immobilized protein served as reference. For binding studies, samples (100 nM TIMP-3, 100 μM related to the molecular weight of disaccharide units (D.U.) GAGs or TIMP-3/GAG complexes after 60 min of preincubation) were injected for 120 s at 30 μL/min, followed by a 10 min dissociation phase in running buffer. The same molar concentrations of D.U. were chosen for GAG comparison because they represent the possible binding sites of interactions. HA-48 kDa and sHA1-31 kDa were used as polymeric HA derivatives during SPR measurements. Binding levels were recorded 10 s before the end of injection. Five M NaCl was used for regeneration. The baseline was allowed to stabilize for 1000 s with running buffer prior to next sample injection. The Biacore T100 evaluation software 2.03 was used to evaluate the binding and kinetic parameters. Specific Biacore sensorgrams were obtained by double referencing.25 Kinetic Analysis of TIMP-3/LRP-1 Clusters II and IV Interactions via SPR. Single-cycle kinetics were performed with immobilized LRP-1 cluster II or IV by sequentially injecting five B

DOI: 10.1021/acs.biomac.6b00980 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules

Figure 1. Structural characteristic of used native and chemically sulfated GAGs. increasing TIMP-3 concentrations (12.5−200 nM) for 220 s without regeneration after the first four TIMP-3 injections at a flow rate of 30 μL/min. After the end of sample injections, TIMP-3 was allowed to dissociate from the surface for 1000 s to determine the TIMP-3/LRP-1 cluster complex stability by fitting the dissociation phases of the sensorgram curves to the heterogeneous ligand model. Modeling of Complement-like Domains 7 and 8 of LRP-1 Cluster II. Structures of CR7 (Uniprot 1011-1054) and CR8 (Uniprot 1059-1100) were obtained from the Protein Data Bank (PDB): PDB IDs 1J8E (X-ray, 1.85 Å) and 1CR8 (NMR, model 1), respectively. Electrostatic potential isosurfaces were calculated using AMBER1426 and PBSA program from AmberTools with a grid spacing of 1 Å and visualized in VMD.27 Isolation and Cultivation of human Bone Marrow Stromal Cells (hBMSCs). hBMSCs were isolated from bone marrow aspirates according to ref 28 and cultured in basic medium (Dulbecco’s minimal essential medium (DMEM; Biochrom, Berlin, Germany) with 10% heat-inactivated fetal calf serum (FCS; BioWest via Th.Geyer, Hamburg, Germany) and antibiotics (20 U penicillin/20 μg streptomycin/ml; Biochrom)). The donors fulfilling the standards for bone marrow donation (e.g., free of HIV, HBV, and serious illness) were informed and gave their approval (votes of the local ethics commission EK263122004, EK114042009). The origin cell populations were all positive for CD73, CD90, and CD105 and negative for CD14, CD43, CD45, and CD133.28 For experiments, cells were seeded onto Petri dishes (ø 3 cm; 5.555 hBMSC/cm2). 2 h after seeding, the treatment with CS, HA-88 kDa, and sHA1-42 kDa (200 μg/mL; Table 1) was started. From day 4 after seeding the medium was additionally supplemented with 10 mM β-glycerophosphate (Sigma, Taufkirchen, Germany), 300 μM ascorbate (Sigma), and 10 nM dexamethasone (Sigma) to induce osteogenic differentiation.29 Western Blot Analysis. After 22 days cell culture, with medium changes twice per week and GAG derivatives added with each medium change, medium was aspirated and the cell layers were washed with PBS (Biochrom) and incubated in lysis buffer (20 mM Tris-HCl, pH 8.8, 2% Na-deoxycholate, 2 mM PMSF (phenylmethylsulfonylfluoride), 2 mM EDTA (ethylendiamine tetraacetate), 2 mM iodineacetic acid, 2 mM N-ethylmaleimide, 1% aprotinin and 0.1 mM Na3VO4) (all chemicals from Sigma, Taufkirchen, Germany) at 4 °C for 10 min. Lysates were centrifuged at 14 000g at 4 °C for 30 min. Ten μg of total protein was subjected to 5−15% SDS-polyacrylamide gel and

transferred afterward to nitrocellulose membranes (GE Healthcare, Freiburg Germany). The membranes were incubated at 4 °C with 5% dry milk (Roth, Karlsruhe, Germany) in 25 mM Tris-buffered saline, pH 8/0.5% Tween-20 (Sigma) for 2 h and mouse antihuman TIMP-3 IgG1 (MAB973 clone 183551, R&D BioTechne, Wiesbaden, Germany) in 25 mM Tris-buffered saline, pH 8/0.5% Tween-20/5% bovine serum albumin. Horseradish peroxidase-conjugated antimouse IgG (CST, Leiden, Netherlands) in 25 mM Tris-buffered saline, pH 8/ 0.5% Tween-20/5% dry milk was used as secondary antibody. Immune complexes were visualized by chemiluminescence detection with MFChemiBIS1.6 (Biostep, Jahnsdorf, Germany). Then, the membrane was reblotted with mouse antihuman glycerinaldhehyde-3-phosphat dehydrogenase antibody (CB1001 clone 6C5, Calbiochem via Merck Millipore, Darmstadt, Germany) Densitometric evaluation was performed with ImageQuant 5.1 software (GE Healthcare). Immunofluorescence Staining. 22 days after seeding, the hBMSCs were fixed in 4% paraformaldehyde (w/v) (Sigma) in PBS for 10 min, permeabilized with 0.1% Triton X-100 (Sigma) in PBS, and incubated with 1% bovine serum albumin (BSA) (w/v)/0.05% Tween-20 (Sigma) in PBS to block unspecific binding sites. For visualization of LRP-1, sheep antihuman LRP-1 cluster III-IgG (AF4824, R&D BioTechne) in PBS/1% BSA/0.05% Tween-20 was used as primary antibody and AlexaFluor-488 donkey antisheep-IgG (A11015, Invitrogen via ThermoFisher) in PBS/1% BSA/0.05% Tween-20 was used as a secondary antibody. For staining of TIMP-3, mouse antihuman TIMP-3-IgG1 (MAB973 clone 183551, R&D BioTechne) in PBS/1% BSA/0.05% Tween-20 was the primary and AlexaFluor-647 goat antimouse IgG (A-21235, Invitrogen via ThermoFisher, Darmstadt, Germany) in PBS/1% BSA/0.05% Tween-20 was the secondary antibody. Samples that were stained for TIMP-3 were costained for fibronectin (FN) with rabbit antihuman FN-IgG (F3648, Sigma) and AlexaFluor-488 donkey antirabbit IgG (A21206, Invitrogen). Mouse anti-TIMP-3 IgG1 (MAB973 clone 183551, R&D BioTechne) in PBS/1% BSA/0.05% Tween-20 was used as primary antibody and AlexaFluor-647 goat antimouse IgG (A-21235, Invitrogen via ThermoFisher, Darmstadt, Germany) in PBS/1% BSA/0.05% Tween-20 was used as a secondary antibody. Nuclei were stained with 0.2 μg DAPI (4′,6-diamidin-2phenylindole, Sigma)/mL in PBS. After embedding in Mowiol 4-88 (Sigma) cells were visualized using an AxioPhot fluorescence microscope (Carl Zeiss, Oberkochen, Germany) with the following C

DOI: 10.1021/acs.biomac.6b00980 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules Table 2. Primer Pairs Used for Real-Time PCR genea

accession no.

forward primer

reverse primer

binding position

product length (bp)b

actin gapdh rps26 lrp-1

NM_001101 NM_002046 NM_001029 NM_002332

aatgtggccgaggactttgattgc tgttcgtcatgggtgtgaacca caatggtcgtgccaaaaag ggtccttcggaacagcac

ttaggatggcaagggacttcctgt tgatggcatggactgtggtcat ttcacatacagcttgggaagc tggatgctctcgtcatagacc

1414−1508 563−718 286−474 5929−5988

95 156 189 60

Genes: actin, β-actin; gapdh, glycerinaldehyde-3-phosphate dehydrogenase; rps26, 40S ribosomal protein S26; lrp-1, low-density lipoprotein receptor-related protein. bConfirmed by sequencing of amplified DNA products with >97% identity.

a

filters: excitation 450−490 nm and emission 520 nm for AlexaFluor488; excitation 608−648 nm and emission 672−712 nm for AlexaFluor-647; and excitation 365 nm and emission 420 nm for DAPI. Digital images were obtained with an AxioCam MRm camera (Carl Zeiss; AxioVision 4.6). Analysis of Gene Expression. For the analysis of gene expression, RNA was prepared from cell lysates at day 22 using RNeasy Mini Kit (Qiagen, Hilden, Germany). cDNA was synthesized using the QuantiTect Reverse Transcription Kit (Qiagen) including a DNA digestion step. Real-time PCR reactions were performed using RotorGeneQ PCR machine (Qiagen) with the RotorGene SYBR Green PCR Kit (Qiagen). After the initial activation step at 95 °C for 5 min, 50 PCR cycles were carried out (denaturation at 95 °C for 5 s; annealing and synthesis at 60 °C for 10 s). Primers were constructed by use of Universal Probe Library (Roche, Mannheim, Germany) (see Table 2 for detailed information) and synthesized by Eurofines MWG Operon (Ebersberg, Germany). The same single-stranded cDNA was used to analyze the expression of the gene of interest and the housekeeping genes glycerinaldehyde-3-phosphate dehydrogenase (GAPDH), β-actin, and 40S ribosomal protein S26 (rps26). The relative expression values were calculated with the three house-keeping genes using the comparative quantification method of the RotorGene software. Statistical Analysis. For analysis of statistical significance, experiments were repeated three times. Cell culture data are derived from three independent experiments using three different hBMSC preparations (cells of the individual donors were not pooled and analyzed independently). Results are presented as mean ± standard deviation. A one-way ANOVA (analysis of variance) with Bonferroni’s post-test was applied. P values