Horseradish Peroxidase-Catalyzed Crosslinking of Fibrin Microthread Scaffolds

Abstract

Horseradish peroxidase (HRP) has been investigated as a catalyst to crosslink tissue-engineered hydrogels because of its mild reaction conditions and ability to modulate the mechanical properties of the matrix. Here, we report the results of the first study investigating the use of HRP to crosslink fibrin scaffolds. We examined the effect of varying HRP and hydrogen peroxide (H₂O₂) incorporation strategies on the resulting crosslink density and structural properties of fibrin in a microthread scaffold format. Primary (1°) and secondary (2°) scaffold modification techniques were evaluated to crosslink fibrin microthread scaffolds. A primary scaffold modification technique was defined as incorporating crosslinking agents into the microthread precursor solutions during extrusion. A secondary scaffold modification technique was defined as incubating the microthreads in a postprocessing crosslinker bath. Fibrin microthreads were enzymatically crosslinked through primary, secondary, or a combination of both approaches. All fibrin microthread scaffolds crosslinked with HRP and H₂O₂ via primary and/or secondary methods exhibited an increase in dityrosine crosslink density compared with uncrosslinked control microthreads, demonstrated by scaffold fluorescence. Fourier transform infrared spectroscopy indicated the formation of isodityrosine bonds in 1° HRP crosslinked microthreads. Characterization of tensile mechanical properties revealed that all HRP crosslinked microthreads were significantly stronger than control microthreads. Primary (1°) HRP crosslinked microthreads also demonstrated significantly slower degradation than control microthreads, suggesting that incorporating HRP and H₂O₂ during extrusion yields scaffolds with increased resistance to proteolytic degradation. Finally, cells seeded on HRP crosslinked microthreads retained a high degree of viability, demonstrating that HRP crosslinking yields biocompatible scaffolds that are suitable for tissue engineering. The goal of this work was to facilitate the logical design of enzymatically crosslinked fibrin microthreads with tunable structural properties, enabling their application for engineered tissue constructs with varied mechanical and structural properties.

Impact statement

This study is the first to report the use of horseradish peroxidase to dityrosine crosslink a fibrin scaffold. We demonstrate the strategic engineering of fibrin microthread scaffolds with tunable biophysical properties by a facile method of varying crosslinker incorporation. The incorporation of crosslinking agents into precursor solutions during microthread extrusion was considered a primary method, whereas soaking microthreads in a postprocessing crosslinker bath was considered a secondary method. The ability to generate tunable scaffold mechanics and degradation rates will enable the application of fibrin microthreads toward the design of engineered tissues with varying architectures, mechanical properties, and functional requirements.

Introduction

Native tissues possess extracellular matrix (ECM) architectural cues that are highly complex and dictate tissue-specific functions. The primary structural components of ECM are fibrillar elements, arranged into hierarchically ordered structures from the nano- to macro-scale. The organization of these fibers varies from tissue to tissue to optimize the mechanical properties, contact guidance cues, and function of specific tissues. In tendon and ligament, type I, II, and V collagen fibrils are organized into dense and highly aligned cable-like bundles to facilitate uniaxial force transduction.¹ In skin, type I and III collagen fibers form sheet-like arrays in various orientations that are capable of resisting tension in multiple directions.² This structure–function relationship has motivated the rational design and development of microfiber scaffolds that mimic this complex, fibrous tissue architecture.

Microthreads are discrete, fibrous scaffolds comprising naturally derived biopolymers, including fibrin,^3–7 collagen,^8–11 silk,^12,13 and chitosan.^14,15 Their structure is analogous to native ECM fibrils, and they can be similarly arranged to create hierarchically complex scaffolds to recapitulate a variety of tissue architectures.¹⁶ Microfiber scaffolds also provide provisional structural and mechanical support, topographic cues, and biochemical signaling cues to direct cell-mediated tissue regeneration.¹⁶

Fibrin microthreads are a microfiber scaffold whose cylindrical form mimics native ECM fibrils, promoting cellular alignment and guided tissue regeneration.^{3,4,6,7,17–20} Fibrin, the bioactive provisional matrix protein responsible for blood clotting, recruits cells that direct wound healing and functional tissue regeneration, making it a beneficial scaffold material.²¹ Fibrin microthreads are fabricated by extruding solutions of fibrinogen and thrombin, which triggers the polymerization of a fibrin matrix.³ These scaffolds can be modified to deliver therapeutic proteins such as fibroblast growth factor 2²² and hepatocyte growth factor^17,23 to enhance cellular proliferation and migration. Discrete fibrin microthreads hierarchically organized into bundles or aligned composite patches mimic the organization and mechanical properties of muscle and cardiac tissue, respectively.^4,7,17 Fibrin microthreads have robust mechanical properties compared with fibrin hydrogel-based systems, whereas uncrosslinked fibrin scaffolds remain susceptible to rapid degradation by fibrinolytic proteases when implanted in vivo, limiting the regenerative capacity of these matrices.^3,5,6 This motivated the investigation of production and postprocessing modifications to increase the structural stability and persistence of these scaffolds.

Toward this, the mechanical properties of fibrin microthread scaffolds have been modified through both production⁶ and postprocessing crosslinking techniques.^3,5 Increasing the degree to which fibrin microthreads are stretched during production yields scaffolds with significantly greater tensile strength and stiffness, but it does not decrease the scaffold degradation rate.⁶ Fibrin microthreads have also been ultraviolet (UV)³ and carbodiimide crosslinked⁵ via postprocessing methods. The facile method of UV crosslinking produced microthreads with a two-fold greater ultimate tensile strength and moduli, but attenuated fibroblast proliferation.³ This significantly limits the utility of UV crosslinking for implantable tissue engineering scaffolds. Carbodiimide crosslinking of fibrin microthreads also resulted in tunable scaffold mechanics.⁵ However, when carbodiimide crosslinked microthreads were implanted into a murine muscle injury, scaffolds demonstrated minimal signs of degradation, even 60 days after implantation.¹⁷ The results of this study suggest that the high degree of crosslinking limited scaffold degradation and cell mediated scaffold remodeling. Thus, there is a significant unmet need for a crosslinking strategy that tunes fibrin microthread mechanical properties and degradation while maintaining cell viability.

One alternative method of crosslinking biomaterial scaffolds is through dityrosine crosslinking.^24,25 Dityrosine crosslinks are responsible for stabilizing biopolymers in vivo^26,27 by exploiting naturally occurring tyrosine residues within biomaterial scaffolds.²⁴ Fibrin has the requisite phenol groups, containing ∼3.3 mol% tyrosine, to enable utilization of this crosslinking method.^28,29 Formation of dityrosine bonds can be facilitated by fenton-like, photo-initiated, and enzymatic reactions.²⁴ Researchers investigated the use of both fenton and photo-initiated reactions to dityrosine crosslink fibrin.^29–32 Fibrinogen was crosslinked through a fenton reaction by exposure to hematin and nonthermal plasma treatment, which generates hydrogen peroxide (H₂O₂).³² H₂O₂ and the ferrous ion in hematin are requisites for fenton-like dityrosine crosslinking, and they resulted in the formation of dityrosine bonds in fibrinogen.³² Photo-initiated reactions to dityrosine crosslink fibrin with ruthenium sodium persulfate have also been investigated as a means of enhancing structural stability and limiting cellular remodeling and degradation.^29–31 Finally, enzymatic reactions to dityrosine crosslink scaffolds take place in aqueous solutions and at neutral pH, making this method amenable to biopolymer scaffold modifications. In addition, enzymes have high specificity, limiting deleterious side-chain reactions that are common with other dityrosine crosslinking methods such as photo-initiated reactions.

Horseradish peroxidase (HRP) is one of the most commonly used enzymes, with H₂O₂ frequently used as the substrate. H₂O₂ interacts with the Fe(III) core of HRP, generating a high oxidation state intermediate compound. This intermediary consists of an Fe(IV) oxyferryl center and porphyrin-based cation radial.³³ In the presence of a phenol that acts as a reducing substrate, the intermediate compound undergoes two consecutive reduction steps to return HRP to its resting state. Resulting phenol radicals create a covalent dityrosine bond between polymer-phenol conjugates. Because the catalytic cycle of HRP consumes one H₂O₂ molecule and generates two phenolic radicals (and thus one crosslink), varying H₂O₂ concentration has been utilized to easily tune polymer scaffold mechanical properties, swelling ratio, gelation time, and degradation rate in a variety of phenol-containing biomaterial scaffolds.^34–45 In addition, biomaterials crosslinked with HRP are biocompatible, and they have demonstrated desired cellular responses in applications such as cartilage regeneration, wound healing, and stem cell differentiation and delivery.^{38,41,43,46–52} Although enzymatic dityrosine crosslinking has been used for many biomaterials, to our knowledge it has yet to be used to crosslink fibrin.

Here, we report the results of the first study to investigate the HRP-catalyzed dityrosine crosslinking of a fibrin scaffold. We examined the effect of varying HRP and H₂O₂ incorporation approaches on the resulting crosslink density and structural properties of fibrin microthreads. We evaluated both primary and secondary scaffold modification techniques to crosslink fibrin microthreads. The incorporation of crosslinking agents into the precursor solutions during microthread extrusion was considered a primary method, whereas soaking microthreads in a postprocessing crosslinker bath was considered a secondary method of scaffold modification. Fibrin microthreads were enzymatically crosslinked through primary, secondary, or both approaches. Resulting microthreads were evaluated for crosslink density, mechanical properties, resistance to proteolytic degradation, and their ability to support cell viability. We hypothesized that strategically varying the incorporation strategy of HRP and H₂O₂ through extrusion or postprocessing modifications would yield scaffolds with tunable crosslink densities, tensile mechanical properties, and degradation rates. The purpose of this work was to enable the rational design of enzymatically crosslinked fibrin microthreads with tunable structural properties, facilitating their use for tissue engineering applications that target a broad range of tissues with varying mechanical and structural properties.

Materials and Methods

Fibrin microthread extrusion

Fibrin microthreads were generated by co-extruding fibrinogen and thrombin as previously described.³ Briefly, fibrinogen isolated from bovine plasma (MP Biomedicals, Irvine, CA) was dissolved in HEPES (N-[2-Hydroxyethyl]piperazine-N’-[2-ethanesulfonic acid]) buffered saline (HBS, 20 mM HEPES, 0.9% NaCl, pH 7.4) to a concentration of 70 mg/mL and stored at −20°C until use. Thrombin isolated from bovine plasma (Sigma Aldrich, St. Louis, MO) was dissolved in HBS at 40 U/mL and stored at −20°C until use. To produce uncrosslinked (UNX) control microthreads, fibrinogen and thrombin solutions were thawed to room temperature, and thrombin was mixed with a 40 mM CaCl₂ solution to form a working solution of 6 U/mL thrombin and 34 mM CaCl₂. Equal volumes of fibrinogen and thrombin solutions were aspirated into separate 1 mL syringes and inserted into the end of a blending applicator (Micromedics, Inc., St. Paul, MN; SA-3670). The solutions were mixed in the blending applicator and extruded through 0.86 mm inner diameter polyethylene tubing (Becton Dickinson, Sparks, MD) into a 10 mM HEPES buffer bath (pH 7.4) in a Teflon coated pan. Threads were extruded at 0.225 mL/m by using a dual syringe pump. After 15 m incubation to allow for polymerization, the microthread scaffolds were removed, stretched to 300% of their initial drawn length, and hung overnight to dry under the tension of their own weight. Previous work demonstrated that stretching microthreads enhances mechanical properties compared with unstretched microthreads.⁶ The UNX control microthreads are defined as microthreads that were not subject to additional processing or crosslinking beyond thrombin-mediated polymerization.

Preparation of horseradish peroxidase crosslinked microthreads

Three different crosslinking methods were developed to evaluate the effect of scaffold crosslinking techniques on fibrin microthread crosslink densities, mechanical properties, and degradation rates. Primary extrusion and secondary postprocessing scaffold modification techniques were utilized to crosslink fibrin microthreads. HRP and H₂O₂ were either included in precursor solutions before microthread extrusion, referred to as a primary (1°) incorporation method, included in a postprocessing bath, referred to as a secondary (2°) method, or a combination of both approaches (1°/2°) (Fig. 1B). A 1000 U/mL HRP stock solution was prepared by reconstituting type VI lyophilized HRP powder (Sigma Aldrich) in deionized (DI) water and stored at −20°C until use. A stock solution of 0.165 M H₂O₂ was made by diluting H₂O₂ (Sigma Aldrich) in DI water and stored at 4°C until use.

FIG. 1.

Method of HRP-catalyzed dityrosine crosslinking. (A) Schematic of HRP-catalyzed oxidation reaction of tyrosine residues on fibrin, yielding the formation of dityrosine and isodityrosine bonds. (B) Incorporation strategies utilized to crosslink fibrin microthreads. The HRP and H₂O₂ were either included in precursor solutions before microthread extrusion, referred to as a primary (1°) incorporation method, or included in a postprocessing bath, referred to as a secondary (2°) method. HRP, horseradish peroxidase. Color images are available online.

For the first crosslinking method, 2° HRP crosslinking, fibrin microthreads were extruded and then crosslinked through a postprocessing bath. Fibrin microthreads were extruded into a bath of HBS, stretched, and dried overnight as described earlier to generate UNX microthreads. These microthreads were secured in one-well plates (Nunc; Thermo Fisher Scientific) by suspending them taught between two slatted polydimethylsiloxane (PDMS) columns placed on the edges of the one-well plates, as previously described.⁵³ Approximately thirteen 5-cm-long microthreads were suspended in each one-well plate. Microthreads then underwent postprocessing by first hydrating with 30 mL of DI water for 30 m, then replacing DI water with 30 mL of HRP (0.02 U/mL) and H₂O₂ (4.83 μM) solution, and incubating at 37°C on a shaker plate for 4 h. After 4 h, threads were rinsed three times with DI water and dried under tension. An additional control was included, UNX 2° H₂O₂, where microthreads were incubated in a bath containing only H₂O₂ (4.83 μM) solution with no HRP. Without the addition of HRP to catalyze the reaction, these UNX 2° H₂O₂ microthreads are considered uncrosslinked.

For the second crosslinking method, 1°/2° HRP crosslinking, HRP was added to the fibrinogen solution before extrusion (1°), and H₂O₂ was introduced via a postprocessing bath (2°). We chose to incorporate HRP into the fibrinogen precursor solution based on previous work crosslinking fibrinogen with hematin, a molecule of similar size and structure to HRP, which showed that fibrinogen maintained its activity with this molecule incorporated.³² A solution of fibrinogen (70 mg/mL) and HRP (0.22 U/mL) was coextruded with thrombin (6 U/mL) into HBS. Microthreads were stretched, dried, and affixed within a one-well plate suspended between PDMS columns. Microthreads were then postprocessed by incubation in an H₂O₂ (4.83 μM) bath at 37°C on a shaker plate for 4 h. After 4 h, threads were rinsed three times with DI water and dried in tension. The intermediary product of fibrin microthreads extruded with HRP (UNX 1° HRP) served as an additional uncrosslinked control for this study.

For the third crosslinking method, 1° HRP crosslinking, both HRP and H₂O₂ were mixed into precursor fibrinogen and thrombin solutions, respectively, immediately before microthread extrusion. A fibrinogen (70 mg/mL) and HRP (0.22 U/mL) solution was coextruded with a thrombin (6 U/mL) and H₂O₂ (52.5 μM) solution into HBS. Before stretching, threads were incubated at 37°C in HBS for 4 h, then stretched, and finally dried.

For all methods, the concentration of H₂O₂ in the preparations was calculated to be 10x greater than the theoretical amount required to react with all the tyrosine groups in the fibrin microthreads. The HRP concentration was determined by holding the ratio of H₂O₂: HRP constant, 240 μM H₂O₂: 1 U/mL HRP, based on previously published literature that sought to optimize HRP crosslinking parameters.³⁵

Experiment

Experimental design

Fluorescence spectroscopy

UV fluorescence microscopy using a Zeiss Axiovert 200 M microscope (Carl Zeiss, Germany) was performed with uniform exposure time to assess the fluorescence of dityrosine crosslinked scaffolds, where excitation of a dityrosine crosslinked material at 315 nm results in an emission centered at 400 nm.⁵⁴ To quantify changes in fluorescence between microthread conditions, pixel intensities of images from randomly selected regions were quantified using ImageJ by converting the blue channel to an 8-bit gray scale and determining the mean gray value from three regions on the microthread. Mean gray values of the image background were subtracted to normalize data (n ≥ 9).

Fourier transform infrared spectroscopy

To further characterize HRP-mediated isodityrosine bonds in fibrin microthreads, analyses of samples were conducted with Fourier transform infrared spectroscopy (FTIR). Twelve −2 cm-long microthread samples were positioned onto the attenuated total reflectance crystal of a Bruker Vertex 70 instrument (Bruker, Billerica, MA) with a Golden Gate ATC accessory (Specac, Swedesboro, NJ). The FTIR absorbance spectra data were collected in the mid-IR range, 4500–800 cm⁻¹, and obtained by averaging 1024 scans. Backgrounds were subtracted from each spectrum. Three batches of microthreads for each condition were analyzed and averaged to obtain a representative spectrum for comparison to other conditions. Baseline correction of absorbance was performed by normalizing data at 4200 cm⁻¹, a region of the spectra with no characteristic peaks. For quantitative comparison of isodityrosine peaks, ratios of the 1040 cm⁻¹ characteristic peak to C-H stretch (2955 cm⁻¹) for each sample were compared and statistically evaluated (n ≥ 3).

Mechanical characterization

Mechanical analysis of HRP crosslinked fibrin microthreads was performed using methods previously described by our laboratory.^5,55 Briefly, individual microthreads were attached with medical-grade silicone adhesive (Factor II, Lakeside, AZ) to vellum paper frames with precut windows of 2.0 cm, which defined the region of loading. The 2.0 cm distance was the space between the adhesive spots on the two edges of the vellum frame window and was used as the initial gauge length for testing. The microthreads that adhered to vellum frames were hydrated in phosphate-buffered saline (PBS) (pH 7.4, room temperature) for 1 h before testing. The hydrated microthread diameters were measured by using a calibrated reticule with a 10 × objective on a Nikon Eclipse E600 upright microscope (Nikon, Melville, NY). To define the cross-sectional area, the microthreads were assumed to be cylindrical. Three measurements were taken along the length of each microthread and averaged. The hydrated microthreads within the vellum frame were mounted in the grip of an ElectroPuls E1000 uniaxial testing machine (Instron, Norwood, MA) with a 1 N load cell. Once secured, the two edges of each vellum frame were cut, and the microthreads were uniaxially loaded until failure at a strain rate of 10 mm/min. Force and displacement data were recorded every 0.1 s through each test.

The engineering stress was calculated as the recorded force divided by the initial cross-sectional area of the microthread. The strain was calculated as the increased extension from the initial 2.0 cm gauge length. A MATLAB (MathWorks, Natick, MA) script was written to analyze the ultimate tensile strength, maximum tangent modulus (MTM), strain at failure, and load at failure for each sample. The MTM was defined as the highest linear region of the stress–strain curve for each sample that corresponds to at least 20% of the total length of each test. Outliers were identified as datum that fell outside the mean ± two standard deviations within each data set and were subsequently removed from further analysis (n ≥ 23).

Degradation assay

To assess proteolytic degradation of HRP crosslinked microthreads, an in vitro plasmin degradation assay was performed, as previously described.⁵ Single crosslinked microthreads were cut to 1 cm in length and secured to the bottom of 48-well plates with medical-grade silicone adhesive. Each experimental condition was run in triplicate. Brightfield images of microthreads were taken on a Nikon Eclipse TS100 inverted microscope both before and after hydration (d₀) in 0.5 mL tris buffered saline (TBS, 25 mM Tris-HCl, 0.9% NaCl, pH 7.5). After 1 h hydration, TBS was replaced with 0.5 mL of 0.5 U/mL human plasmin (EMD Biosciences, San Diego, CA) in TBS and incubated at 37°C. The microthreads were imaged at 0, 1, 3, 6, 9 h, 1, 2, 3, and 4 days after addition of plasmin. Each image was analyzed with ImageJ to measure microthread diameter at three different positions along the length of the microthread. Data were plotted as a ratio to the initial diameter (d/d₀) (n ≥ 9).

Swelling ratio

Fibrin microthread swelling ratios were determined by measuring microthread diameter measurements both before and after hydration, using a well-characterized tool for evaluating microthread swelling ratio.^3,5,6,9,56 Microthread diameter measurements made both before and after 1 h hydration in TBS (before addition of plasmin) in the degradation assay were used to calculate swelling ratio, based on previous work.⁵ In a preliminary experiment, we observed no appreciable changes in diameter beyond 1 h, and therefore, this was deemed adequate to fully hydrate the microthreads. The three measured diameters for each microthread were averaged. Swelling ratio was defined as the ratio of the wet diameter of a microthreads to its dry diameter (wet/dry diameter) (n ≥ 9).

Cell culture

C2C12 immortalized mouse myoblasts (ATCC CRL-1772, Manassas, VA) were cultured in complete medium, consisting of a 1:1 (v/v) ratio of high-glucose Dulbecco's modified Eagle's medium (DMEM; Gibco, Thermo Fisher Scientific) and Ham's F12 (Gibco), supplemented with 4 mM L-glutamine, 10% fetal bovine serum (FBS, HyClone, Logan, UT), 1% penicillin-streptomycin (Thermo Fisher Scientific, Waltham, MA), and 1% amphotericin-B (Thermo Fisher Scientific), as previously described.⁵ Cells were incubated at 37°C with 5% CO₂ and maintained at a density below 70% confluence by using standard cell culture techniques. Cell passage was carried out by using 0.25% trypsin-EDTA (Corning, Corning, NY).

Myoblast viability

To facilitate cell attachment, five microthreads were bundled together by hydrating scaffolds dropwise with PBS until microthreads adhered together, as previously described.⁵⁷ The microthread bundles were adhered with medical-grade silicone adhesive to Thermanox™ coverslips (Thermo Fisher Scientific) cut to 2.5 cm by 2.5 cm. Individual coverslips with adhered microthread bundles were placed in six-well plates, sterilized with ethylene oxide (EtO), and allowed to aerate for at least 24 h to remove residual EtO. Immediately before seeding, the microthread bundles were rehydrated with 1 mL of PBS for 1 h. After 1 h, PBS was aspirated and 500 μL of a 250,000 cells/mL cell suspension in complete medium supplemented with 50 μg/mL aprotinin was added to the top of the coverslip. After 4 h of incubation, medium and unattached cells were aspirated from each well immediately followed by replacing with 2 mL of fresh medium to completely submerge the microthread bundles. Seeded microthreads were cultured for 3 d. Cell viability was qualitatively assessed with a LIVE/DEAD staining kit (Molecular Probes, Eugene, OR) according to the manufacturer's protocol. After staining, microthreads bundles were removed from the culture coverslips and placed on glass slides for fluorescence imaging, to ensure that only cells that are directly attached to the microthread bundles are imaged. The microthreads were imaged by using a fluorescence microscope (Zeiss Axiovert 200M microscope, Carl Zeiss, Germany) (n ≥ 6).

Statistical analyses

Statistical analyses were performed by using Graphpad Prism 7 software (Graphpad, Software, La Jolla, CA). Statistical differences between means of the fluorescence pixel intensity and normalized FTIR peak intensity were determined by one-way ANOVA (p < 0.05) with Tukey's multiple-comparisons post hoc analysis. When ANOVA assumptions were not met, a nonparametric equivalent Kruskal Wallis test was run with Tukey's multiple-comparison post hoc analysis; this was performed to analyze statistical differences between the medians for swelling ratio and mechanical data. Two-way ANOVA (p < 0.05) with Tukey's multiple-comparisons post hoc analysis was used to evaluate degradation. Values reported are means ± standard error of the mean (SEM) unless otherwise stated. Significance is indicated as *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, and ****p ≤ 0.0001. In addition, γ indicates significance (p ≤ 0.05) from all other groups.

Experimental Results

Altering horseradish peroxidase incorporation strategy yields microthreads with varying crosslink density

Three methods of crosslinking fibrin microthreads were investigated by varying the incorporation strategy of crosslinking reagents HRP and H₂O₂. Crosslinking reagents were either (1) incorporated with fibrinogen and thrombin precursor solutions before microthread extrusion (1° HRP crosslinking), (2) mixed into a postprocessing bath (2° HRP crosslinking), or (3) combined by using both extrusion and postprocessing strategies (1°/2° HRP crosslinking) (Fig. 1B). The addition of HRP and H₂O₂ to fibrin microthread extrusion and/or postprocessing resulted in the formation of covalent bonds between tyrosine residues of fibrin (Fig. 1A), which was confirmed by quantitative fluorescence microscopy and FTIR (Fig. 2). The UV fluorescence microscopy and subsequent pixel intensity quantification was performed to assess the fluorescence properties of dityrosine crosslinked scaffolds. All fibrin microthread scaffolds crosslinked with HRP and H₂O₂ via primary and/or secondary methods exhibited increased fluorescence compared with UNX controls, indicating an increase in dityrosine crosslink density (Fig. 2A). Pixel intensity analyses of fluorescence images were performed to further quantify this finding (Fig. 2B). Pixel intensity quantification suggests increasing dityrosine bond formation in all HRP crosslinked fibrin microthreads compared with UNX control microthreads.

FIG. 2.

Analyses of crosslink density for HRP crosslinked fibrin microthreads. (A) Representative images of microthread fluorescence indicate dityrosine bond formation. (B) Pixel intensity quantification of microthread fluorescence suggests increasing dityrosine bond formation in HRP crosslinked fibrin microthreads. (C) Average normalized FTIR spectrum of control and HRP crosslinked microthreads. The peak at 1040 cm⁻¹, indicative of an ether bond, indicates the addition of isodityrosine bonds in 1° HRP crosslinked microthreads. (D) Peak normalization indicates a significant increase in isodityrosine bonds in 1° HRP crosslinked microthreads. Scale bars are 50 μm. FTIR, fourier transform infrared spectroscopy. Color images are available online.

To further characterize the effect of HRP-catalyzed crosslinking on fibrin microthread secondary structure, FTIR was performed (Fig. 2C). In addition to dityrosine bonds, HRP-catalyzed crosslinking can yield the formation of isodityrosine, a diphenyl ether bond.⁵⁸ Isodityrosine bonds were characterized by the formation of peaks in FTIR spectra centered at 1040 cm⁻¹, indicative of ether bonds. Only 1° HRP crosslinked microthreads exhibited a peak in the FTIR spectra at 1040 cm⁻¹ (Fig. 2C). A significant increase in the relative amount of this bond compared with the C-H stretch (2955 cm⁻¹) was observed in 1° HRP crosslinked microthreads compared with UNX controls, suggesting that only this method of crosslinking microthreads yielded the formation of isodityrosine bonds (Fig. 2D).

HRP-mediated crosslinking modulates the tensile mechanical properties of fibrin microthreads

Uniaxial tensile testing was performed on HRP crosslinked microthreads to characterize the effect of crosslink density on scaffold mechanical properties. After crosslinking, fibrin microthreads were mounted onto vellum frames, hydrated, and uniaxially loaded until failure to determine their tensile mechanical properties. Characteristic stress–strain curves of UNX microthreads revealed initial toe regions where there was a small increase in stress with increasing elongation, similar to previous work evaluating fibrin microthread tensile properties (Fig. 3A).⁶ The HRP crosslinked microthreads displayed shorter toe regions with a more rapid increase in stress until failure than UNX microthreads (Fig. 3A). Regardless of whether HRP and H₂O₂ were incorporated during extrusion (1°), postprocessing (2°), or a combination of both methods (1°/2°), all crosslinked microthreads exhibited significantly greater ultimate tensile strengths compared with UNX controls (p < 0.01) (Fig. 3B). Ultimate tensile strengths of HRP crosslinked microthreads ranged between 6 and 9.5 MPa. The HRP crosslinking yielded varying MTM values depending on HRP and H₂O₂ incorporation strategies (Fig. 3C). The MTM of 1°/2° HRP crosslinked microthreads was ∼4.5-fold greater than UNX control threads (p < 0.0001) (Fig. 3C). However, no significant increase in MTM was observed for both 1° and 2° HRP crosslinked microthreads compared with UNX controls. The significant differences between tangent moduli for each HRP crosslinking method suggests the dependence of scaffold mechanical properties on crosslinker incorporation strategy. Although all HRP crosslinked microthreads displayed trends in lower strain at failure, 1°/2° HRP crosslinked microthreads were the only crosslinked condition with significantly reduced strain at failure compared with UNX controls (p < 0.001) (Fig. 3D).

FIG. 3.

Tensile mechanical properties of HRP crosslinked microthreads. (A) Representative stress–strain curves from uniaxial tensile testing of fibrin microthreads to failure show that crosslinked microthreads exhibit shorter toe regions with a more rapid increase in stress until failure than UNX control microthreads. (B) Ultimate tensile strength of all HRP crosslinked scaffolds regardless of incorporation strategy was increased compared with control UNX microthreads. (C) Strain at failure is significantly decreased in 1°/2° HRP crosslinked microthreads compared with control microthreads. (D) MTM is significantly higher in 1°/2° and 2° HRP crosslinked microthreads compared with control. γ indicates significance (p ≤ 0.05) from all other groups. MTM, maximum tangent modulus. Color images are available online.

Moreover, the addition of either HRP or H₂O₂ alone to fibrin microthreads affected scaffold mechanical properties. Both UNX 1° HRP and UNX 2° H₂O₂ microthreads exhibited significantly higher ultimate tensile strengths and MTM compared with UNX microthreads (Fig. 3B, C). In addition, UNX 1° HRP microthreads possessed significantly reduced strains at failure compared with UNX control microthreads (Fig. 3D). All mechanical data are summarized in Table 1.

Table 1.

Mechanical and Structural Properties of Fibrin Microthreads Crosslinked with Horseradish Peroxidase

	Diameter (μm)		Swelling ratio	UTS (MPa)	MTM (MPa)	SAF (mm/mm)	Load (mN)
	Dry	Hydrated
UNX	56.5 ± 9.1 (9)	117.1 ± 18.2 (9)	2.08 ± 0.12 (18)	3.1 ± 1.7 (40)	20.3 ± 28.3 (41)	0.50 ± 0.22 (40)	15.0 ± 3.4 (41)
UNX HRP inc	50.0 ± 10.1 (9)	96.6 ± 23.7 (9)	1.92 ± 0.18 (9)	7.8 ± 4.5 (30)	62.3 ± 46.9 (29)	0.28 ± 0.18 (29)	12.9 ± 6.7 (29)
UNX H2O2 bath	48.1 ± 15.8 (9)	86.6 ± 33.5 (9)	1.78 ± 0.13 (9)	5.6 ± 2.7 (34)	34.1 ± 22.3 (33)	0.37 ± 0.21 (36)	16.9 ± 5.9 (34)
1° HRP x-linked	57.5 ± 13.0 (9)	115.7 ± 32.9 (9)	1.99 ± 0.18 (9)	7.2 ± 5.4 (30)	29.9 ± 22.3 (29)	0.41 ± 0.18 (32)	25.8 ± 8.5 (30)
1°/2° HRP x-linked	49.9 ± 10.5 (9)	94.1 ± 22.5 (9)	1.88 ± 0.19 (9)	9.5 ± 4.0 (24)	88.2 ± 45.6 (23)	0.27 ± 0.15 (25)	13.1 ± 4.9 (24)
2° HRP x-linked	44.5 ± 9.7 (9)	82.1 ± 21.9 (9)	1.83 ± 0.14 (9)	6.0 ± 2.7 (30)	42.4 ± 41.6 (30)	0.44 ± 0.24 (32)	19.8 ± 7.0 (31)

Data presented as mean ± SEM (n #).

HRP, horseradish peroxidase; MTM, maximum tangent modulus.

Microthread swelling ratio is influenced by HRP incorporation strategy

To further elucidate the effect of HRP-catalyzed crosslinking on microthread crosslink density, the swelling ratios of the scaffolds were analyzed. Microthreads were hydrated in TBS for 1 h and imaged on an inverted microscope both before and after hydration to measure microthread diameters. These measurements were used to calculate swelling ratios.^5,9,56 A trend of decreased swelling ratios relative to UNX microthreads was observed for all microthread scaffolds processed with HRP and/or H₂O₂, although not all swelling ratios showed significant reductions (Fig. 4). Scaffolds crosslinked by means of postprocessing (2° HRP crosslinked microthreads), as well as UNX 2° H₂O₂ microthreads, swelled significantly less than UNX microthreads (Fig. 4). All microthread diameters and swelling ratios are summarized in Table 1.

FIG. 4.

Swelling ratios of fibrin microthreads as a function of HRP and H₂O₂ incorporation strategy. Both the UNX 2° H₂O₂ and 2° HRP crosslinked microthreads had significantly lower swelling ratios compared with UNX microthreads, suggesting an increase in crosslink density. Color images are available online.

Prolonged degradation of 1° HRP crosslinked fibrin microthreads

The structural integrity of HRP crosslinked microthreads was evaluated by performing an in vitro plasmin degradation assay. Plasmin is the primary enzyme responsible for fibrinolysis in vivo.⁵⁹ Hydrated fibrin microthread diameters were measured both before (d₀) and after the addition of plasmin (d). Diameter measurements of degrading microthreads were made up to 4 days after the addition of plasmin and used to calculate the ratio d/d₀. Degradation profiles of all plasmin-treated microthreads over time exhibited an initial rapid decrease in diameter within the first 24 h, followed by slower degradation through day 4 (Fig. 5A). In contrast, UNX microthreads not exposed to plasmin (negative control) did not decrease in diameter over time, validating that observed degradation is solely plasmin mediated (Fig. 5A).

FIG. 5.

In vitro plasmin degradation assay. The structural integrity of HRP crosslinked microthreads was evaluated as a function of time by measuring the diameters of degrading microthreads (d) relative to their initial hydrated diameters before the addition of plasmin (d0) by calculating the ratio d/d0. (A) Degradation profiles of all plasmin-treated microthreads indicate an initial rapid decrease in diameter. (B) The same data was plotted as a column graph to compare differences between conditions at discrete time points. At all time points after 1 h, UNX microthreads with no plasmin exposure had significantly less degradation than all plasmin-treated conditions. In addition, 1° HRP crosslinked microthreads had significantly lower degradation than UNX microthreads at early time points before 24 h, suggesting that incorporating crosslinkers during extrusion yields scaffolds with increased resistance to proteolytic plasmin degradation. Color images are available online.

To further elucidate differences in degradation between control and crosslinking conditions at discrete time points, the data were also plotted as a column graph (Fig. 5B). At all time points beyond 1 h, UNX microthreads with no plasmin exposure had significantly less degradation than all other plasmin-treated conditions at that time point (Fig. 5B). Notably, at 3, 6, and 9 h after the addition of plasmin, 1° HRP crosslinked microthreads had significantly less degradation than UNX microthreads. This suggests that incorporating crosslinkers during microthread extrusion yields scaffolds with increased resistance to proteolytic degradation (Fig. 5B).

Cellular viability is retained on crosslinked scaffolds

To verify that HRP crosslinked fibrin microthreads were suitable to support extended cell viability, C2C12 murine myoblasts were seeded on the surface of crosslinked microthread bundles. Seeded microthread bundles were cultured for 3 days before live/dead staining, and fluorescence imaging was performed to qualitatively evaluate cell viability. Cells grown on tissue culture plastic (TCP) served as a positive control, and cell growth on TCP and exposed to 70% ethanol for 1 h immediately before staining served as a negative control. One hundred percent viability was observed for the live cell TCP positive controls, and no living cells were observed for the negative controls (Fig. 6). To account for fibrin microthread autofluorescence, an unseeded thread bundle also served as a negative control to denote the diffuse background fluorescence observed in several seeded conditions, such as the 1° HRP crosslinked condition (Fig. 6). Although it is possible that this background fluorescence may attenuate the ability to visually quantify dead cells, a high degree of cell viability was observed for all control and HRP crosslinked microthread bundles (Fig. 6).

FIG. 6.

Cellular viability on HRP crosslinked fibrin microthreads. C2C12 murine myoblasts demonstrate high viability after 3 days on control and HRP crosslinked fibrin microthread bundles, as demonstrated by fluorescence imaging of live/dead staining. Only a few dead cells could be identified on the surface of microthreads bundles (white arrows). Scale bars are 100 μm. Color images are available online.

Discussion

The goal of this study was to dityrosine crosslink fibrin microthreads to facilitate tunable scaffold crosslink density, mechanical properties, and degradation rates. Dityrosine crosslinking of fibrin has been achieved by photochemical^29–31 and fenton-like reactions,³² whereas the use of HRP-catalyzed dityrosine crosslinking had never been investigated for crosslinking fibrin scaffolds. The HRP and H₂O₂-mediated crosslinking strategies were varied through primary and/or secondary scaffold modification techniques by mixing reagents with precursor solutions before extrusion (1°), by adding reagents into a postprocessing bath (2°), or through a combination of both methods (1°/2°). Utilization of primary and/or secondary incorporation strategies yielded microthreads with a range of crosslink densities, evaluated by FTIR and dityrosine fluorescence analyses. Characterization of tensile mechanical properties revealed that all HRP crosslinked microthreads were significantly stronger than UNX control microthreads and exhibited a range of MTM and strain at failure values. In addition, 1° HRP crosslinked microthreads demonstrated significantly slower plasmin degradation rates than UNX microthreads, suggesting that incorporating HRP and H₂O₂ during extrusion yields scaffolds with increased resistance to proteolytic degradation. Finally, cells seeded on all crosslinked microthreads exhibited a high degree of viability, demonstrating that HRP crosslinking yields biocompatible scaffolds that are suitable for tissue engineering applications. Taken together, these findings support the strategic design of engineered fibrin microthreads with a range of biophysical properties, using a facile method of varying crosslinker incorporation strategy. Tunable scaffold mechanics and degradation rates enable the application of fibrin microthreads to a diverse array of target tissues with varying structure, mechanical properties, and functions, such as skeletal or cardiac muscle, tendon, ligament, or skin.

Most notably, altering HRP incorporation strategy resulted in fibrin microthreads with varying degrees of crosslink density. In both primary and secondary crosslinking methods, important parameters that govern this reaction, including temperature, pH, protein concentration, reaction time, and HRP:H₂O₂ ratio, were held constant.^35,37 To account for differences between the crosslinking methods, concentrations of H₂O₂ were normalized for each method by calculating the concentration needed to react 10 times the theoretical amount of tyrosine in the fibrin microthreads. Quantification of dityrosine fluorescence within fibrin microthreads indicated that 1° HRP crosslinked microthreads had the highest degree of dityrosine crosslinking. We believe that the fluorescence gradient observed in 1° HRP crosslinked can be attributed to fibrin autofluorescence and imaging inhomogeneities related to the cylindrical geometries of the scaffolds, rather than an uneven distribution of crosslinking throughout the microthread.

The initiation of crosslinking was further substantiated by FTIR analyses, which confirmed the formation of isodityrosine chemical bonds on 1° HRP crosslinked microthreads, which occur between phenol groups of fibrin.^58,60,61 Interestingly, 2° and 1°/2° HRP crosslinked microthreads did not appear to contain detectable levels of isodityrosine bonds. Isodityrosine bonds are a less common reaction product of this enzymatic reaction, and they may be dependent on factors such as reaction pH, steric hinderance of tyrosine residues within fibrin, and the degree of surface exposure of tyrosines.⁶² Taken together, these results confirm that HRP-catalyzed crosslinking of fibrin occurs as hypothesized, and that the crosslinking is dependent on whether HRP and H₂O₂ are incorporated during extrusion or postprocessing steps. Future analyses of crosslink density with liquid chromatography and/or mass spectroscopy may provide additional information about this reaction, including the percentage of tyrosines participating in the crosslinking reaction.^24,63 These data would enable further tuning of this system to optimize crosslink density toward the precise engineering of an instructive biomaterial scaffold.

Differences in the extent of crosslinking observed between HRP incorporation strategies may likely be explained by the distribution of HRP and H₂O₂ within the microthreads. When crosslinked through primary means, HRP and H₂O₂ were mixed with fibrinogen and thrombin precursor solutions, respectively, immediately before extrusion. This ensured a uniform distribution of enzyme and substrate within the fibrin during polymerization, which may facilitate the generation of more phenol radicals. In addition, this crosslinking method took place during polymerization when we anticipate a higher degree of polymer chain mobility, which is important for permitting radical collisions that result in covalent crosslinks.³⁵ In contrast, the secondary crosslinking strategy of soaking microthreads in a postprocessing bath relied on diffusion of HRP and H₂O₂ from the surface of the material into the microthreads. Fibrin microthreads are dense structures made with markedly higher concentrations of fibrinogen than fibrin hydrogels, which may limit the diffusion of HRP into the microthreads. We hypothesize that this limited diffusion may result in crosslinking occurring predominantly on the surface of the microthreads where it comes in direct contact with the crosslinking bath, significantly reducing the generation of phenolic radicals and thus crosslinks. In addition, this postprocessing technique took place after microthreads underwent polymerization and likely exhibit lower polymer chain mobility, which may also limit radical collisions. This hypothesis has previously been proposed by researchers who crosslinked electrospun gelatin and collagen fibers with glutaraldehyde.^64,65 They speculated that a lack of glutaraldehyde diffusion into the collagen and gelatin fibers resulted in uneven crosslinking and the generation of a highly crosslinked outer shell on the fiber scaffolds.^64,65 This hypothesis may explain the discrepancies observed between crosslink density and the swelling ratio of microthreads. Secondary (2°) HRP crosslinked microthreads were the only crosslinked scaffold that had significantly lower swelling ratio than UNX microthreads. This may be due to a high degree of crosslinking along the periphery of the microthread, preventing diffusion of buffer into the scaffold and limiting swelling.^64,65 Although 1° HRP crosslinking generated the most dityrosine and isodityrosine crosslinks, we did not observe a significant decrease in the swelling ratio of 1° HRP crosslinked microthreads compared with UNX controls. This may be a result of the uniform distribution of crosslinks throughout the microthread, allowing buffer to diffuse into the microthread.

The HRP crosslinking of fibrin microthreads yielded enhanced biophysical properties, including increased tensile strength and stiffness. The HRP crosslinking produced ultimate tensile strengths approximately two to three times greater than UNX microthreads, depending on crosslinking strategy. These results are consistent with previous literature performing photochemical dityrosine crosslinking on fibrin hydrogel scaffolds. Work by Elvin and colleagues irradiated fibrinogen in the presence of ruthenium trisbipyridyl chloride and sodium persulfate to generate covalent dityrosine crosslinks.²⁹ They found that crosslinked fibrinogen reached an ultimate tensile strength of 45 kPa, representing a 60% increase compared with their previously reported fibrin hydrogel crosslinked with thrombin and factor XIII.^29,66 In addition, we found that the MTM values of fibrin microthreads increased by ∼1.5–3.5 times by utilizing varying HRP crosslinking strategies. Other researchers reported a two-fold increase in fibrin hydrogel compressive stiffness after photochemical crosslinking to initiate dityrosine bond formation.³¹ Finally, the conservation of the 1° HRP crosslinked and UNX 1° HRP microthreads structural and mechanical properties suggests that the addition of H₂O₂ to thrombin before extrusion did not inhibit the enzyme's functional activity.

Interestingly, the addition of HRP or H₂O₂ alone to fibrin microthreads enhanced the mechanical properties of the scaffolds. Both the UNX 1° HRP microthreads and UNX 2° H₂O₂ microthreads had significantly increased ultimate tensile strength and MTM values, relative to control UNX microthreads. We hypothesize that trace quantities of native reactive oxygen species from serum are among the nonspecific molecules within the fibrinogen preparation, and they are activated by the HRP in UNX 1° HRP microthreads to initiate crosslinking.⁶⁷ The enhanced mechanical properties of UNX 2° H₂O₂ microthreads may be explained by the peroxide-generating free radical side groups in the fibrin, which can trigger nonspecific crosslinking events.⁶⁸ Alternatively, trace quantities of native serum heme peroxidases such as myeloperoxidase may also be present in the fibrinogen preparation.⁶⁹ Myeloperoxidase is present in serum and has been shown to crosslink tyrosine residues in the presence of H₂O₂.^70,71 Additional native serum peroxidases may activate H₂O₂ in the UNX 2° H₂O₂ microthreads and initiate dityrosine crosslinking. These hypotheses of crosslinking within UNX 2° H₂O₂ microthreads are further strengthened by observations of a significant increase in dityrosine autofluorescence, as well as a significant decrease in swelling ratio compared with control UNX microthreads. Future work will seek to elucidate the mechanisms by which these HRP and H₂O₂ control microthreads are able to seemingly undergo crosslinking and yield enhanced tensile mechanical properties.

The range of microthread ultimate tensile strengths and moduli permitted by utilizing primary and/or secondary HRP crosslinking strategies enables the application of these scaffolds to an array of target tissues that possess varying structures, mechanical demands, and functions. Tunable mechanical stiffnesses of biomaterial scaffolds can regulate an array of cellular processes, including adhesion, proliferation, migration, and differentiation.⁷² To further exploit this, we plan on conducting future studies that systematically vary fibrinogen, HRP, and H₂O₂ concentrations; it is well recognized that modifying these concentrations yields scaffolds with a range of mechanical properties, including modulus and strength.^34–40 However, an excessive increase of H₂O₂ concentration inhibits the enzymatic activity of HRP, so these modifications must be done with discretion.^35,37 In addition, future work to hierarchically assemble these microthreads into complex 3D scaffolds such as composites, braids, and bundles will allow us to further optimize bulk scaffold mechanics and mimic tissue-specific ECM architecture such as skeletal muscle, ventricular myocardium, tendon, ligament, or skin.¹⁶

HRP-catalyzed crosslinking further mediated the resistance of fibrin microthreads to proteolytic degradation. Although fibrin microthreads have robust mechanical properties compared with fibrin hydrogels, fibrin remains susceptible to cell-mediated compaction and rapid degradation by proteinases when implanted in vivo.^3,5,6,31 Despite significant increases in crosslink density and tensile mechanical properties, both 1°/2° and 2° HRP crosslinked microthreads did not exhibit increased resistance to proteolytic degradation compared with UNX control microthreads. In contrast, 1° HRP crosslinked microthreads, which exhibited the highest degree of crosslink density, demonstrated prolonged degradation at early time points compared with control microthreads. One possible explanation may be that the increased crosslink density inhibits plasmin from accessing its cleavage sites, which was previously hypothesized for carbodiimide crosslinked microthreads.⁵ This may be a result of the location of dityrosine crosslinks between and within monomer units. Tyrosine residues are most abundant on fibrinogen's β-chain (4.9%) and γ-chain (5.6%), both of which are cleaved by plasmin to form fibrin degradation products.²⁸ The HRP-mediated crosslinking appears to have a less pronounced effect on fibrin microthread degradation than carbodiimide crosslinking.⁵ This result is desired, as carbodiimide crosslinked microthreads implanted into a murine muscle defects showed minimal signs of degradation at the 60 day terminal time point, suggesting that the high crosslink density limited cell-mediated tissue remodeling.¹⁷ For applications in tissue regeneration, scaffold degradation should match the rate of new tissue ingrowth and ECM deposition.⁷³ HRP-catalyzed crosslinking has been previously shown to prolong the enzyme-mediated degradation of hyaluronic acid (HA)-tyramine and silk-HA composite hydrogels.^36,40,41 Tunable scaffold degradation was achieved by altering H₂O₂ concentration,³⁶ percent concentration of composite components,⁴⁰ and delivery method.⁴¹ Future work varying fibrinogen, HRP, and H₂O₂ concentrations within our crosslinked microthreads will allow us to develop more robust changes in scaffold degradation kinetics.

In addition to controlled degradation, tissue-engineered scaffolds must support cellular viability and not elicit an immunological response. We qualitatively observed robust myobalst viability after 3 days of culture on all HRP crosslinked fibrin microthreads. The C2C12 myoblasts were chosen as an immortalized musculoskeletal cell line to perform initial cellular viability assessments. These results are consistent with previous literature supporting extended viability and proliferation of cells seeded on or encapsulated within HRP crosslinked scaffolds fabricated from a large range of biomaterials, including silk,^34,40 HA,^49,52 silk-HA composites,⁴⁰ alginate,⁴³ chitosan,⁴⁷ gelatin,^48,50,51,74 and dextran.⁴⁶

Despite this, concerns remain regarding the in vivo immunological response to plant-derived HRP, which could potentially hinder its efficacy as a crosslinker for implantable biomaterial scaffolds.^37,75,76 In two in vivo rodent models, HRP immunization with subsequent boosters containing Freund's adjuvant elicited the production of antibodies against HRP.^75,76 Although immunization with Freund's adjuvants likely heighted this immune response, the efficacy of HRP-based treatments still must be rigorously evaluated before its clinical application can be realized. Some groups have addressed this limitation by generating enzyme-free scaffolds whereby HRP is immobilized to magnetic or silica beads.^77,78 By passing the hydrogel solutions through a syringe filled with beads, immobilized HRP is able to catalyze the crosslinking reaction and, subsequently be contained within the syringe, eliminating it from the final crosslinked hydrogel.^77,78 Alternatively, the use of human peroxidases or oxidoreductases that are capable of catalyzing phenolic coupling could alleviate the aforementioned concerns with immunogenicity and clinical translation.^37,79–81 One example is myeloperoxidase, which has been shown to crosslink tyrosine residues in the presence of H₂O₂.⁷⁰ Alternative human-derived catalysts such as hematin, obtained through decomposition of hemoglobin, have been evaluated for their ability to catalyze dityrosine crosslinking.^32,82,83

In addition to HRP's immunogenicity, additional concerns regarding cytocompatibility of H₂O₂ have also been raised.^37,39,84,85 Both Gardner et al. and Gulden et al. determined that increasing concentration of and/or exposure time to H₂O₂ in vitro resulted in increasing cytotoxicity in murine fibroblasts and C6 glioma cells, respectively.⁸⁵ Others demonstrated senescence-like growth arrest on exposure of human diploid fibroblasts to low concentrations (200 μM) of H₂O_2.⁸⁶ We anticipate cytotoxicity resulting from H₂O₂ to be minimal in our HRP crosslinked fibrin microthreads, because these scaffolds are crosslinked and, subsequently rinsed thoroughly before being seeded with cells, removing residual H₂O_2. This is unlike many HRP crosslinked hydrogel scaffolds that encapsulate cells during the crosslinking process, where cells are directly exposed to H₂O₂. Nonetheless, the immunological and cytotoxicity concerns of HRP and H₂O₂ emphasize the importance of thoughtful scaffold design to minimize deleterious effects from crosslinking and warrant further investigation. Future scaffold modifications could include careful consideration of HRP and H₂O₂ concentrations and exposure times, the development of enzyme-free systems, and use of human peroxidases or HRP-mimetics.

Conclusions

In this study, we demonstrated that fibrin microthread scaffolds can be dityrosine crosslinked by using HRP as a catalyst. To our knowledge, this is the first investigation to apply HRP-catalyzed dityrosine crosslinking to fibrin-based scaffolds. We investigated the effect of varying HRP and H₂O₂ incorporation strategies through primary extrusion and/or secondary postprocessing scaffold modification techniques to facilitate the generation of tunable scaffold crosslink density, mechanical properties, and degradation rate. The HRP crosslinking of fibrin microthreads resulted in the generation of dityrosine bonds and increased crosslink density. Characterization of tensile mechanical properties revealed that all HRP crosslinked microthreads were significantly stronger than control microthreads and demonstrated a range of moduli and strain at failure values. Scaffolds crosslinked by primary processing demonstrated significantly slower degradation rates than control microthreads, suggesting that incorporating HRP and H₂O₂ during microthread extrusion yields scaffolds with increased resistance to proteolytic degradation. Lastly, cells seeded on HRP crosslinked microthreads retained a high degree of viability. Together, these findings demonstrate that HRP crosslinking yields biocompatible scaffolds that are suitable for tissue engineering applications. Further, these results demonstrate the strategic engineering of fibrin microthread scaffolds with a tunable range of microthread biophysical properties by a facile method of varying crosslinker incorporation strategy. The ability to generate tunable scaffold mechanics and degradation rates will enable the application of fibrin microthreads for the biomimetic design of engineered tissues with varying tissue architectures, mechanical properties, and functional requirements.

Disclosure Statement

The authors declare the following competing financial interest(s): G.D.P. discloses that he is a co-founder and has equity interest in Vitathreads L.L.C., a company that aims at commercializing scaffolds composed of fibrin microthreads.

Funding Information

This work was funded in part by NIH R15 HL137145, NSF DGE IGERT 1144804, and NSF EEC REU 1559819.