Biomaterials doped with Bone Morphogenetic Protein 2 (BMP2) have been used as a new therapeutic strategy to heal non-union bone fractures. To overcome side effects resulting from an uncontrollable release of the factor, we propose a new strategy to site-directly immobilize the factor, thus creating materials with improved osteogenic capabilities.
Different therapeutic strategies for the treatment of non-healing long bone defects have been intensively investigated. Currently used treatments present several limitations that have led to the use of biomaterials in combination with osteogenic growth factors, such as bone morphogenetic proteins (BMPs). Commonly used absorption or encapsulation methods require supra-physiological amounts of BMP2, typically resulting in a so-called initial burst release effect that provokes several severe adverse side effects. A possible strategy to overcome these problems would be to covalently couple the protein to the scaffold. Moreover, coupling should be performed in a site-specific manner in order to guarantee a reproducible product outcome. Therefore, we created a BMP2 variant, in which an artificial amino acid (propargyl-L-lysine) was introduced into the mature part of the BMP2 protein by codon usage expansion (BMP2-K3Plk). BMP2-K3Plk was coupled to functionalized beads through copper catalyzed azide-alkyne cycloaddition (CuAAC). The biological activity of the coupled BMP2-K3Plk was proven in vitro and the osteogenic activity of the BMP2-K3Plk-functionalized beads was proven in cell based assays. The functionalized beads in contact with C2C12 cells were able to induce alkaline phosphatase (ALP) expression in locally restricted proximity of the bead. Thus, by this technique, functionalized scaffolds can be produced that can trigger cell differentiation towards an osteogenic lineage. Additionally, lower BMP2 doses are sufficient due to the controlled orientation of site-directed coupled BMP2. With this method, BMPs are always exposed to their receptors on the cell surface in the appropriate orientation, which is not the case if the factors are coupled via non-site-directed coupling techniques. The product outcome is highly controllable and, thus, results in materials with homogeneous properties, improving their applicability for the repair of critical size bone defects.
The ultimate goal of bone tissue engineering and bone regeneration is to overcome the disadvantages and limitations occurring during common treatments of non-union fractures. Auto- or allo-transplantations are predominantly used as current therapy strategies, even though they both have several drawbacks. The ideal bone graft should induce osteogenesis by osteoinduction as well as osteoconduction, leading to the osteointegration of the graft into the bone. Nowadays, only auto-transplantation is considered as the "gold standard" since it provides all characteristics of an ideal bone graft. Unfortunately, it also presents important negative aspects, such as long surgery times, and a second trauma site that usually entails more complications (e.g., chronic pain, hematoma formations, infections, cosmetic defects, etc.). Allogenic grafts, on the other hand have suboptimal characteristics for all general aspects1. Alternative bone graft technologies have been improved in the last few years, with the aim to produce scaffolds that are osteoinductive, osteoconductive, biocompatible, and bioresorbable. Since many biomaterials do not show all of these osteogenic characteristics, different growth factors, mainly BMP2 and BMP7, have been incorporated in order to improve the osteogenic potential of the particular scaffold2.
As an essential criterion, such growth factor delivery systems should provide a controlled dose release over time in order to facilitate the essential events like cell recruitment and attachment, cell ingrowth, and angiogenesis. However, BMPs as well as other osteogenic growth factors have been commonly immobilized non-covalently3. Entrapment and adsorption techniques require the use of supra-physiological amounts of protein due to an initial burst release, which leads to severe disadvantages in vivo, typically affecting the surrounding tissues by inducing bone overgrowth, osteolysis, swelling, and inflammation4. Thus, the retention of growth factors at the delivery site for longer periods of time can be achieved by covalent immobilization methods. Chemically modified BMP2 (succinylated5, acetylated6 or biotinylated7), engineered heterodimers8, or BMP2 derived oligopeptides9 have been designed and used to overcome the limitations related to absorption. However, the bio-activity of these constructs is not predictable since the arrangement potentially inhibits the binding of the immobilized ligand to the cellular receptors. As previously shown, it is essential that all four receptor chains involved in the formation of activated ligand-receptor complexes interact with the immobilized BMP2 in order to fully activate all downstream signaling cascades10.
To overcome the problems of an inhomogeneous product outcome with limitations in terms of bioactivity, stability, and bioavailability of the immobilized factor, we designed a BMP variant capable of covalently binding scaffolds in a site-directed manner. This variant, termed BMP2-K3Plk, comprises an artificial amino acid which was introduced by genetic codon expansion11. This variant has been successfully linked to scaffolds using a covalent coupling strategy while maintaining its biological activity.
1. Production of the BMP2 Variant BMP2-K3Plk
2. Optimization of Copper (I)-Catalyzed Alkyne-Azide Cycloaddition (CuAAC) Conditions
3. Covalent Coupling Technique of BMP2-K3Plk to Azide Functionalized Agarose Beads
4. Validating the Presence and Biological Activity of Immobilized BMP2-K3Plk Using Texas Red Labeled BMP Receptor I A Ectodomain (BMPR-IA EC )
5. Measuring Alkaline phosphatase (ALP) Expression to Prove the In Vitro Bioactivity of the Produced BMP2-K3Plk Before and After the Coupling Reaction.
In this article, we describe a method to covalently couple a new BMP2 variant, BMP2-K3Plk, to commercially available azide functionalized agarose beads (Figure 1). The bioactivity of the produced BMP2-K3Plk variant was validated by the induction of alkaline phosphatase (ALP) gene expression in C2C12 cells. The in vitro test shows similar ALP expression levels induced by wild type BMP2 (BMP2-WT) and BMP2-K3Plk (Figure 2).
The redox reactions between copper (II) sulfate (CuSO4) and sodium ascorbate (NaAsc) generate reactive oxygen species that might affect the structural integrity and thus affect the bioactivity of BMP2-K3Plk. To verify the structural integrity of the protein, we performed an overnight incubation with CuSO4 and NaAsc, showing BMP2-WT degradation in a concentration-dependent manner (visualized by SDS-PAGE under reduced conditions (Figure 3A1)) and the formation of multimers or aggregates visible at approximately 40 kDa (under non-reduced conditions) (Figure 3A2). To avoid protein degradation, Tris(3-hydroxypropyltriazolyl-methyl)amine (THPTA) was used as a protective agent (Figure 3B). The use of THPTA prevents BMP2 fragmentation, while the formation of higher molecular weight structures could not be prevented by the addition of THPTA. Further improvements of the reaction addressed the composition of the reaction buffer, the reaction temperature, and reaction time (data not shown). The protocol described here details the final achievements regarding reaction buffer composition, temperature, and reaction time.
To confirm that BMP2-K3Plk retains bioactivity after coupling, we used fluorescently labeled receptor ectodomain protein of the type I receptor (BMPR-IAEC) to demonstrate that the immobilized protein is still able to bind to this receptor. Beads coated with BMP2-K3Plk via CuAAC chemistry yielded fluorescence when incubated with the dye-labeled BMPR-IAEC (Figure 4). In contrast, non-coated beads or beads that were incubated with BMP2-WT showed no fluorescence signal above background levels (data not shown)14.
After confirming receptor binding capabilities of the immobilized ligand in vitro, we also tested whether biological responses can be triggered by the functionalized beads in a cell-based assay. ALP mediated staining occurred only in those cells which were in direct contact with the BMP2-K3Plk-functionalized beads (Figure 5). This confirms that the protein is indeed covalently linked to the beads and not just absorbed, since a more spread-out staining at larger distances to the beads would otherwise have been observed.
Figure 1: Depiction of BMP2-K3Plk. (A) Depiction of the BMP2-K3Plk variant with the localization of the introduced amino acid substitution. (B) Coupling scheme: BMP2-K3Plk CuAAC reaction with azide functionalized beads. Reprinted (adapted) with permission from Tabisz et al. (2017): Site-Directed Immobilization of BMP-2: Two Approaches for the Production of Innovative Osteoinductive Scaffolds. Biomacromolecules. 18 (3), 695-708. (Copyright 2017 American Chemical Society) Please click here to view a larger version of this figure.
Figure 2: Bioactivity of BMP2-variants. Comparison of the bioactivities (ALP assay) of BMP2-K3Plk and wildtype BMP2 (BMP2-WT). The x-axis represents the concentration of BMP2-WT or BMP2-K3Plk used in the assay. The EC50 data represent mean values and standard deviations of 4 individual experiments (N=4). Reprinted (adapted) with permission from Tabisz et al. (2017): Site-Directed Immobilization of BMP-2: Two Approaches for the Production of Innovative Osteoinductive Scaffolds. Biomacromolecules. 18 (3), 695-708. (Copyright 2017 American Chemical Society) Please click here to view a larger version of this figure.
Figure 3: Effect of the reducing agent on the integrity of the BMP2. (A) BMP2-WT was exposed to different molar ratios of sodium ascorbate (NaAsc) to copper (II) sulfate (CuSO4). The samples were analyzed by SDS-PAGE and Coomassie Brilliant Blue staining under reduced (Figure A1) and non-reduced conditions (Figure A2). In the reduced conditions (A1), the red arrows represent cleaved BMP2-WT. In the non-reduced conditions (A2), the red arrows indicate multimeric BMP2-WT. The bands representing cleaved BMP2 species are indicated by a blue arrow. Legend: NCR – reduced BMP2-WT untreated; NC – BMP2-WT untreated; 1:1 to 20:1 – increasing molar ratios of sodium ascorbate to CuSO4. (B) BMP2-K3Plk was coupled to 3-Azido-7-hydroxycoumarin using CuAAC reaction conditions supplemented with THPTA. Upon coupling 3-Azido-7-hydroxycoumarin becomes a fluorescent dye. Legend: NC – BMP2-WT reacted with 3-Azido-7-hydroxycoumarin; 7:1 to 20:1 increasing molar ratios of THPTA to CuSO4. Please click here to view a larger version of this figure.
Figure 4: Bioactivity of immobilized BMP2-K3Plk in vitro. Representative picture of the interaction of immobilized BMP2-K3Plk with Texas-Red-labelled BMPR-IAEC. Reprinted (adapted) with permission from Tabisz et al. (2017): Site-Directed Immobilization of BMP-2: Two Approaches for the Production of Innovative Osteoinductive Scaffolds. Biomacromolecules. 18 (3), 695-708. (Copyright 2017 American Chemical Society) Please click here to view a larger version of this figure.
Figure 5: Bioactivity of immobilized BMP2-K3Plk in a cell-based assay. (A) Representative picture of alkaline phosphatase (ALP) staining upon treatment with beads coupled to BMP2-K3Plk functionalized beads. (B) Alkaline phosphatase (ALP) staining upon treatment with soluble 25 nM BMP2-K3Plk. Reprinted (adapted) with permission from Tabisz et al. (2017): Site-Directed Immobilization of BMP-2: Two Approaches for the Production of Innovative Osteoinductive Scaffolds. Biomacromolecules. 18 (3), 695-708. (Copyright 2017 American Chemical Society) Please click here to view a larger version of this figure.
Generating tagged protein variants by genetic codon expansion allows the introduction of various non-natural amino acid analogs principally at any position of the primary protein sequence. In case of BMPs like BMP2, common tags such as a 6-Histidine (His) tag can only be introduced N-terminally, since the protein´s C-terminal end is buried within the tertiary protein structure, and is thus not accessible from the outside. At other positions, the size of the introduced tag may very likely cause structural alterations that consequently obliterate the BMP´s bioactivity. Further, introducing a mutation into the BMP2 sequence can affect the refolding efficacy, and can also change other protein parameters such as the isoelectric point of the modified protein. Thus, every single step in an established protein production protocol might need to be adapted according to the altered protein variant´s characteristics. The production of the BMP2-K3Plk indeed required a modification of the established method for the expression of wild type BMP2. Different culture times or propargyl-L-lysine (Plk) concentrations were tested (data not shown) in order to reach the highest expression yield. Refolding, separation and purification steps fortunately did not require further adaptions. We have already reported that in case of other BMP2 variants produced in our lab, some protein characteristics were altered significantly, which required the modification of several steps of the BMP production method18. The produced BMP2-K3Plk showed biological activity comparable to that of the wild type protein. A difference occurs at high BMP2 concentrations, probably as a consequence of a lower solubility of the BMP2-K3Plk variant in medium compared to that of wild type BMP2.
Despite the known advantages of copper-catalyzed azide-alkyne cycloaddition (CuAAC)19, CuAAC reactions should be carefully adapted to particular applications. We showed how BMP2 can be fragmented or create aggregates or multimers due to the strong reducing reaction conditions. Thus, all reaction parameters had to be analyzed in detail in order to find conditions that leave the protein unaffected.
Our approach to covalently couple the BMP2 variant to azide-functionalized agarose beads by click chemistry could be realized with high efficiency resulting in functionalized beads triggering osteogenic differentiation in vitro. When wild type BMP2 protein was used in the reaction with the beads instead, we could observe only weak staining of ALP expression, indicating that coupling indeed occurred highly specifically via the propargyl-L-lysin residue of BMP2-K3Plk.
This positional coupling specificity reflects a great improvement compared to the immobilization procedures involving classical NHS/EDC chemistry. In such cases, random coupling occurs, mainly involving the primary amine groups present in lysine residues. The coupled protein has a variable osteogenic activity, due to the multiplicity of possible connections to the scaffold, leading to different orientations of the protein towards cell receptors.
The use of site-direct immobilized BMP2 might overcome the mentioned drawbacks related to the high doses of soluble BMP2, which are needed to induce bone formation. In addition, the results clearly show that certain cellular responses, such as the osteogenic differentiation of C2C12 cells, are entirely initiated by immobilized BMP2, despite recent studies claiming that signal transduction requires endocytosis of the ligand/ligand-receptor complex20,21. Our findings are in agreement with other studies demonstrating that BMP2, which is coupled covalently (but not site-directed) to non-endocytosable surfaces, is still able to induce osteoblast differentiation22.
Considering that we have used supra-physiological concentrations of BMP2 for the coupling reaction, the amount of the immobilized BMP2 might be further reduced while still conserving the osteogenic properties of the beads. Prior to such optimization steps, however, the BMP2-K3Plk-functionalized scaffold needs to be tested in animal experiments to prove its osteogenic potential in vivo.
The authors have nothing to disclose.
The authors thank Dr. M. Rubini (Konstanz, Germany) for providing the plasmid encoding pyrrolysyl-tRNA and for providing pRSFduet-pyrtRNAsynth encoding the corresponding aminoacyl-tRNA synthetase.
Material | |||
1-Step NBT/BCIP | Thermo Fisher | 34042 | Add solution to cells |
3-Azido-7-hydroxycoumarin | BaseClick | BCFA-047-1 | Chemical used for click reaction |
Agarose low melting point | Biozym | 840101 | Agarose for ALP assay |
Azide agarose beads | Jena Bioscience | CLK-1038-2 | Beads used for reaction |
BamHI (Fast Digest enzyme) | Thermo Fisher Scientific | FD0054 | Restriction enzyme |
BMP receptor IA (BMPR-IAEC) | — | — | Produced in our lab |
Coomassie Brilliant Blue G-250 Dye | Thermo Fisher Scientific | 20279 | Chemical used for Coomassie Brilliant blue staining of SDS PAGE |
Copper (II) sulfate anhydrous (CuSO4) | Alfa Aesar | A13986 | Chemical used for click reaction |
DNA Polymerase and reaction buffer | Kapabiosystems | KK2102 | KAPA HiFi PCR Kit |
Dulbecco’s modified Eagle’s medium (DMEM) GlutaMAX | Gibco | 61965-026 | Cell culture media |
ethylenediaminetetraacetic acid (EDTA) | Sigma Aldrich GmbH | E5134-1kg | Chemical used to stop click reaction |
Isopropyl ß-D-1-thiogalactopyranoside (IPTG) | Carl Roth GmbH | 2316.5 | Bacteria induction (1mM final concentration) |
NdeI (Fast Digest enzyme) | Thermo Fisher Scientific | ER0581 | Restriction enzyme |
NHS-activated Texas Red | Life technologies | T6134 | Coupled to receptor |
P- Nitrophenyl Phosphate | Sigma Aldrich GmbH | N4645-1G | Alkaline Phosphatase |
p25N-hmBMP2 | — | — | Plasmid kindly provided from Walter Sebald to J. Nickel |
pET11a-pyrtRNA | — | — | Provided by the Chair for Pharmaceutics and Biopharmacy, University Wuerzburg |
propargyl-L-lysine (Plk) | — | — | Provided by the Chair for Pharmaceutics and Biopharmacy, University Wuerzburg |
pSRFduet-pyrtRNAsynth | — | — | Provided by the Chair for Pharmaceutics and Biopharmacy, University Wuerzburg |
Qiagen Gel Extraction Kit | Qiagen | 28704 | Gel Purification |
Qiagen PCR purification Kit | Qiagen | 28104 | PCR Purification |
Sodium L-ascorbate | Sigma Aldrich GmbH | A7631-100G | Chemical used for click reaction |
T4 DNA Ligase | ThermoScientific | EL0011 | Ligation |
tris(3-hydroxypropyltriazolylmethyl)amine (THPTA) | BaseClick | BCMI-006-100 | Chemical used for click reaction |
4-(1,1,3,3-Tetramethylbutyl)phenyl-polyethylene glycol | Sigma Aldrich GmbH | X100-1L | Triton X 100 |
Name | Company | Catalog Number | Comments |
Equipment | |||
Amicon concentrating cell 400 ml | Merck KGaA | UFSC40001 | Concentrating unit |
Amicon Ultra-15 Centrifugal Filter Units | Merck KGaA | UFC901024 | Concentrating centrifugal unit |
ÄKTA avant FPLC | ÄKTA | — | FPLC machine |
Avanti J-26XP | Beckman Coulter | 393124 | Centrifuge for bacterial culture |
Bacterial Shaking Incubator | Infors HT | Shaking incubator for bacterial culture | |
FluorChem Q system | proteinsimple | — | Imaging and analysis system for SDS-PAGE |
Fluorescent miscroscope | Keyence | BZ-9000 (BIOREVO) | |
Fractogel® EMD SO3– (M) | Merck KGaA | 116882 | Ion Exchange Chromatography column material |
Greiner CELLSTAR® 96 well plates | Sigma | M5811-40EA | 96 well plates for cell culture (ALP Assay) |
Heraeus Multifuge X1R | ThermoScientific | — | Centrifuge |
M-20 Microplate Swinging Bucket Rotor | ThermoScientific | 75003624 | Rotor for Microcentrifuge for plate during ALP staining |
Microcentrifuge – 5417R | Eppendorf | — | Centrifuge |
OriginPro 9.1 G | OriginLab | — | software for stastic analysis of ALP assay data |
Polysine Slides | ThermoScientific | 10143265 | microscope slides |
Rotor JA-10 | Beckman Coulter | — | rotor for Avanti J-26XP centrifuge |
Rotor JLA 8.1 | Beckman Coulter | — | rotor for Avanti J-26XP centrifuge |
Rotor JA 25.50 | Beckman Coulter | — | rotor for Avanti J-26XP centrifuge |
Tecan infinite M200 multiplate reader | Tecan Deutschland GmbH | — | Multiplate reader for ALP assay |
Thermocycler – Labcycler Gradient | SensoQuest GmbH | — | PCR |
TxRed – microscope filter | Keyence | Filter for fluorescent microscope | |
Ultrafiltration regenerated cellulose discs 3 kDa | Merck KGaA | PLBC04310 | used with amicon concentrating cell 400ml |
Ultrafiltration regenerated cellulose discs 10 kDa | Merck KGaA | PLGC04310 | used with amicon concentrating cell 400ml |