Cellular Interactions with Hydrogel Microfibers Synthesized via Interfacial Tetrazine Ligation

Abstract

Fibrous proteins found in the natural extracellular matrix (ECM) function as host substrates for migration and growth of endogenous cells during wound healing and tissue repair processes. Although various fibrous scaffolds have been developed to recapitulate the microstructures of the native ECM, facile synthesis of hydrogel microfibers that are mechanically robust and biologically active have been elusive. Described herein is the use of interfacial bioorthogonal polymerization to create hydrogel-based microfibrous scaffolds via tetrazine ligation. Combination of a trifunctional strained trans-cyclooctene monomer and a difunctional s-tetrazine monomer at the oil-water interface led to the formation of microfibers that were stable under cell culture conditions. The bioorthogonal nature of the synthesis allows for direct incorporation of tetrazine-conjugated peptides or proteins with site-selectively, genetically encoded tetrazines. The microfibers provide physical guidance and biochemical signals to promote the attachment, division and migration of fibroblasts. Mechanistic investigations revealed that fiber-guided cell migration was both F-actin and microtubule-dependent, confirming contact guidance by the microfibers. Prolonged culture of fibroblasts in the presence of an isolated microfiber resulted in the formation of a multilayered cell sheet wrapping around the fiber core. A fibrous mesh provided a 3D template to promote cell infiltration and tissue-like growth. Overall, the bioorthogonal approach led to the straightforward synthesis of crosslinked hydrogel microfibers that can potentially be used as instructive materials for tissue repair and regeneration.

1. Introduction

For use as tissue engineering scaffolds, synthetic biomaterials must recapitulate the characteristics of the natural extracellular matrix (ECM). In addition to cytocompatibility and bioactivity, it has become increasingly evident that the morphological and mechanical properties of the scaffolds play a major role in directing cellular functions, including cell migration and cell division.^[1-3] Contact guidance is a phenomenon where cell polarity, morphology and migration are influenced by topographical features of the substrates.^[4-5] It is well known that cells of both epithelial and mesenchymal origins undergo directional migration, rather than random movement, on fibrous materials with oriented fibrillar topography.^[3] Directional migration driven by fibrillar cues is widely found in wound healing,^[6] therefore, scaffolding materials should incorporate ECM-mimicking fibrillar features to promote contact guidance and to facilitate the repair of damaged tissues.

Various methods, including wet spinning,^[7-8] extrusion,^[9] phase separation,^[10] electrospinning,^[11] self-assembly^[12] and polyelectrolyte complexation^[13-14] have been developed for the production of fibrous materials. Traditional polymer processing methods, such as wet spinning and melt extrusion, are incompatible with biomacromolecules and generally produce large diameter (micron to millimeter) fibers. Although phase separation is a straightforward approach to producing porous or fibrous structures, this strategy can only be applied to limited types of polymers. Owing to its simplicity and scalability, electrospinning has been the most widely explored approach to generate fibers with diameters ranging from less than 100 nanometers to several micrometers^[11,15] Semi-crystalline polymers, such as poly(ε-caprolactone) (PCL), poly(lactic acid) (PLA) and poly(lactic-co-glycolic acid) (PLGA), are popular candidates for electrospinning. Although hydrolytically degradable, these materials are chemically and biologically inert, thus unable to initiate intimate engagement with cells. Because these fibers have limited capacity to absorb water, further chemical modifications are required to improve their hydrophilicity and to introduce bioactive motifs.^[11,16-17] While hydrophilic natural polymers, such as hyaluronic acid and dextran, have been electrospun into fibers, additional chemical crosslinking is necessary to prevent their dissolution during cell culture studies.^[18-19]

Alternatively, polyelectrolyte complexation has been exploited for the production of cellladen hydrogel microfibers.^[14] In this method, individual fibers were drawn in parallel from polyionic droplets arranged side-by-side on a template; formation of higher structures is possible through continuous drawing and fiber fusion. These physically crosslinked fibers may not be structurally sound; tuning of the fiber properties, stable incorporation of biological motifs and fabrication of scaffolds with complex shape and geometry are not straightforward. Nanofibrous materials have been generated by self-assembly through concerted noncovalent interactions between preprogrammed individual atomic or molecular constituents.^[20] ECM-mimetic hydrogels containing entangled nanofilaments have been synthesized via the self-assembly of short peptides, such as peptide amphiphiles or β-hairpins.^[21-22] The weak mechanical strength of the resultant matrices and the limited scalability, however, prohibit the clinical translation of these materials.

Synthetic polymers, in the form of ropes, rather than microfibers, have been produced via interfacial polymerization.^[23-24] Tuning of molecular diffusivity and the rate at which the polymeric product is removed from the interface may lead to the production of fibers. However, classical interfacial polymerization reactions display poor functional group tolerance, making it difficult to incorporate biological functional groups during synthesis. In recent years, biomaterials research has been greatly enhanced through bioorthogonal chemistry—unnatural reactions that are chemoselective in the presence of biological functionality.^[25-32] Tetrazine ligation is an inverse-electron-demand Diels-Alder cycloaddition between s-tetrazines (Tz) and strained alkenes. With conformationally strained trans-cycloctene (sTCO) as the dienophile, this bioorthogonal reaction features fast kinetics, high selectivity at low concentration and compatibility with biological systems. When applied to materials synthesis via diffusion-controlled strategies, tetrazine ligation enabled the synthesis of hydrogel spheres with 3D spatial patterns^[31] and multiblock copolymers as meter-long polymer microfibers with an exceptionally high molecular weight.^[33] A limitation of our previously reported microfibers is limited stability in aqueous environment; fiber fragmentation was observed after 5 days of incubation under aqueous conditions.

Described herein is the use of interfacial bioorthogonal polymerization to engineer biomimetic hydrogel microfibers that can potentially be used to promote wound healing and tissue repair. Stably crosslinked hydrogel microfibers were obtained through a simple one-step synthesis using a hydrophobic, trifunctional sTCO monomer (tris-sTCO) and a poly(ethylene glycol) (PEG)-derived bifunctional tetrazine monomer (PEG-bis-Tz). Using Tz monomer with a dangling RGD peptide [PEG(RGD)-bis-Tz], hydrogel microfibers or fibrous scaffold displaying integrin binding motifs were prepared. A model protein genetically encoded with an unnatural tetrazine amino acid, when included in the aqueous phase during polymerization, was covalently immobilized in the fibers without compromising structural integrity. These hydrogel-like polymer fibers, with diameters of ~10 μm, are cytocompatible, biologically active and mechanically robust, effectively promoting the attachment, spreading and migration of fibroblasts. Actin and tubulin polymerization inhibition studies were carried out to understand the mechanism of fiber-guided cell migration. We further illustrated how the microfiber scaffold served as an instructive template for tissue growth.

2. Experimental Section

2.1. Fiber synthesis:

To synthesize stably crosslinked microfibers (Figure 1A), tris-sTCO was first dissolved in ethyl acetate at a concentration of 0.8 mg/mL (1.2 mM). Separately, PEG-based bis-Tz monomer, with or without the dangling RGD (Figure 1A), was dissolved in DI water at a concentration of 1 mg/mL (0.125 mM). For protein incorporation, a stock solution of tetrazine-conjugated superfolder green fluorescent protein (sfGFP-Tz) was added to the bis-Tz solution at a final concentration of 1 μM. To a 60-mm diameter petri dish was added 3 mL of the aqueous solution of the appropriate tetrazine molecules. The solution of tris-sTCO (2 mL) in ethyl acetate was carefully added over the aqueous phase without disturbing the interface. Upon contact, a polymer thin film formed at the interface and was grasped gently using sharp tweezers connected to a collecting frame that was driven by a motor at a speed of 10 RPM for continuous fiber collection (Figure 1B-D). Control fibers without any crosslinking were synthesized using PEG-bis-Tz and bis-sTCO.^[33] Detailed synthesis of Tz and TCO species can be found in the Supporting Information. Video S1 captured the fiber collection process. Crosshatched meshes were generated by changing the axis of rotation of the collecting frame.^[33] Fibers were imaged using a Nikon optical microscope (Nikon, Melville, NY) and a Zeiss LSM 880 confocal microscope (Carl Zeiss, Thornwood, NY).

Figure 1.

Synthesis and characterization of crosslinked hydrogel microfibers. (A). Structures of tris-sTCO, PEG-bis-Tz, PEG(RGD)-bis-Tz and sfGFP-Tz. (B). Microfibers were produced via interfacial tetrazine ligation using tris-sTCO and PEG-bis-Tz or PEG(RGD)-bis-Tz, dissolved in ethyl acetate (colorless, 1.2 mM) and water (pink, 0.125 mM), respectively. The hydrogel microfiber (blue arrow) was pulled out of the immiscible interface. (C, D). Photographs of the multi-roller wrapping machine (C) for microfiber collection and microfibers collected on a roller (D). (E). Confocal image of a sfGFP-containing fibrous mesh synthesized using tris-sTCO, PEG-bis-Tz and sfGFP-Tz. Images shown are maximum intensity projection of z-stacks using a 20× objective. Scale Bar: 50 μm. (F-G). Optical micrographs of microfibers immediately after exposure to water (F) and after incubation in water at 37 °C for 21 days. Scale Bar: 50 μm.

2.2. Mechanical properties:

Single fiber tensile tests were conducted via a micro-force tester, reported previously,^[34] modified for microfiber tensile testing. The device applied the same principles used in standard uniaxial tensile testing; however, the use of an in-line strain-based load cell was replaced by a soft-calibrated cantilevered beam (spring constant of 28.4 ± 0.6 μN/μm). One end of the microfibers was fixed to the cantilevered beam while the other end was taped to an X-Y manual positioning stage. Microfibers were aligned to the stretch axis (Z) and preloaded with a 150 μN force to ensure full fiber engagement; this configuration established the reference length of the microfibers 4.3 ± 0.9 mm. During testing, microfibers were stretched by driving the cantilevered beam via a piezoelectric stage (0-800 ± 0.0018 μm) along the stretch axis (Z). The displacement of microfibers was calculated by taking the difference of the stage travel and cantilevered beam deflection, measured with a capacitance sensor (0-150 ± 0.014 μm). The force was determined by multiplying the beam deflection by the beam constant. The microfibers were pre-conditioned by a single ramp from 0 to 3.5% strain and then underwent three consecutive ramps from 0.75 to 3% strain, with a crosshead speed of 350 μm/s or a nominal strain rate of ~8.2 %/s for both stretching and retraction. Following dry fiber testing, hydration was performed by total immersion of fibers in DI water on stage. The hydrated microfibers underwent three consecutive ramps from 0.75 to 12% strain using the same crosshead speed of 350 μm/s. Optical images of dry microfibers were taken with a Nikon MM-400/s microscope (with a Nikon Digital Sight DS-FI1 camera) with a 20 × objective prior to tensile testing. The diameter of each fiber was quantified digitally using ImageJ. The diameters of hydrated microfibers were calculated by multiplying the dry fiber diameters by the average diametric swelling ratios, determined by measuring optical images of dry and hydrated fibers using ImageJ. A total of 29 individual fibers, crosslinked or uncrosslinked, were characterized in this study and T-tests were conducted in Prism to compare the moduli for microfiber types and hydration conditions.

2.3. Cell maintenance:

Approximately 1×10⁶ GFP-labeled NIH3T3 fibroblasts (Cell Biolabs, San Diego, CA) were seeded in 75 cm² cell culture flasks (Celltreat, Shirley, MA), and maintained at 37 °C with 5% CO₂ in high glucose DMEM media with 10% fetal bovine serum (FBS), 0.1 mM nonessential amino acids and 1% penicillin-streptomycin, with media changes every three days. After reaching 95 % confluence, cells were washed with PBS before being lifted with 3 mL of 0.25 % trypsin for subsequent cell culture studies.

2.4. Cell culture with individual fibers.

2.4.1. Cell migration:

To prepare substrates for cell migration study, a small drop of Sylgard^® 184 poly(dimethyl siloxane) (PDMS) mixture was added to the center of the glass slide of a 35-mm glass bottom petri dish (MatTek, Ashland, MA). The mixture was cured at room temperature for 24 h to create a non-adhesive area in the center. A 3-mL cell suspension was added to the petri dish and cells were allowed to attach overnight. A single hydrogel microfiber collected on a stainless steel metal frame (diameter: 10 mm, thickness: 1 mm) was washed with PBS and cell culture media three times each, followed by exposure to germicidal UV light for 15 min. Subsequently, the fiber fixed by the frame was laid on top of cells across the PDMS center. Cells were incubated overnight in the presence of the immobilized microfiber and live cell imaging was performed using a Zeiss LSM 880. Z-stack and tile images were collected every 15 min over a 12-h period. Images were processed using Zen software (Carl Zeiss, Thornwood, NY) and videos were created with ImageJ. Latrunculin B (Lan B) and nocodazole were added in situ to a final concentration of 10 μM for F-actin and microtubule inhibition, respectively. After the inhibitor solutions were added, live cell confocal imaging was carried out for 3 h.

2.4.2. Immunostaining:

For F-actin and ki67 staining, cells were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100 for 5 min, and blocked with 1% bovine serum albumin (BSA) for 1 h. Cells were incubated with an anti-ki67 antibody (1:500 dilution, Abcam, Cambridge, MA) at room temperature (R.T.) overnight. Subsequent incubation with Alexa 647-conjugated secondary antibody (1:400 dilution, Life Technologies, Carlsbad, CA) at room temperature for 2 h completed the ki67 staining. Separately, filamentous F-actin was stained by phalloidin-Alexa 568 (1:400 dilution, 1 h, R.T.). Cell nuclei were counter stained with DAPI (1:1000 dilution, 5 min, R.T.). For inhibition studies, cells were rinsed with PBS and fixed with cold methanol at −20 °C for 20 min after live cell imaging. Cells were then fixed and permeabilized as described above. For microtubule staining, fixed and permeabilized cells were incubated with anti-β tubulin antibody (1:100 dilution, Invitrogen, Carlsbad, CA) overnight at 4 °C, followed by incubation with Alexa 488-conjugated secondary antibody (1:250 dilution, Life Technologies) for 2 h at room temperature. Following staining, samples were placed in PBS and imaged by a Zeiss LSM 880 confocal microscope. Z-stack images were taken using a 10 × and 40 × objective and processed using Huygens (Scientific Volume Imaging, the Netherlands) and Amira (FEI, Hillsboro, OR) for 3D deconvolution and reconstruction.

2.4.3. Quantification of cell migration:

Time-lapse video files were imported individually to the ImageJ image processing software then compiled into a single image stack. When necessary, a macro was used to stabilize the image by tracking x-y translations then adjusting each frame to correct for fiber drift. Using the Manual Tracking plugin, the paths of 10 cells were tracked by selecting the center of the cell while advancing the video frame by frame. The position of these points was recorded and from these tracks the distance traveled and velocity were calculated for each cell. Cell position at each time point relative to the initial position (origin) was plotted and the point-to-point velocities were averaged to obtain a mean velocity for each condition. Student’s t-tests were used for statistical analysis of velocities under normal condition versus under treatment with Lan B and nocodazole.

2.5. Cell culture with fibrous meshes:

Similar to the single fiber experiments, the surface of a glass-bottomed petri dish was coated with Sylgard^® 184 PDMS. A fibrous mesh fixed on the stainless steel frame was laid on top of the PDMS surface, washed with PBS and cell culture media and sterilized with germicidal UV light for 15 min before cell seeding. A cell suspension containing 0.4×10⁶ GFP-labeled NIH3T3 fibroblasts was added to the petri dish to allow for attachment and spreading. Two, four and seven days post seeding, cultures were imaged using a Zeiss LSM 880 confocal microscope. Z-stack images were taken and processed using Zen software.

3. Results and Discussion

3.1. Fiber synthesis.

Stably crosslinked hydrogel microfibers were synthesized via step growth polymerization employing a diffusion-controlled process. The chemical structures of tris-sTCO and the bis-Tz monomers, including PEG-bis-Tz and PEG(RGD)-bis-Tz are shown in Figure 1A. A fibronectin-mimicking peptide with a basic sequence of GRGDSP was strategically incorporated into the PEG-based monomer through the lysine residues to produce the bis-Tz monomer with dangling RGD for better binding with cells. The centrally positioned peptide was flanked with 3.5-kDa PEG so that PEG-bis-Tz and PEG(RGD)-bis-Tz had a comparable molecular weight. These building blocks were synthesized in gram scale with high purity. A water-soluble diphenyl-s-tetrazine was used owing to its aqueous stability and high reactivity towards sTCO (k₂: 284,000 M⁻¹s⁻¹ at 25 °C in water^[31]). For interfacial polymerization, an organic solution of sTCO species was overlaid on top of an aqueous solution of Tz monomers (Figure 1B). The exceptionally rapid kinetics led to the instantaneous formation of a crosslinked polymer thin film at the oil/water interface. The solvent-swollen film was pulled from the interface and collected as microfibers on a rotating mandrel driven by a stepper motor (see Video S1). During fiber collection, the reaction continued at the liquid interface, thereby producing a continuous fiber with excellent mechanical strength. The collector rotation speed was optimized to 10 RPM as faster rotation (> 20 RPM) resulted in fiber breakage. Our custom-made fiber-pulling device (Figure 1C-D) can simultaneously collect fibers from three separate reservoirs. In our experiment, the sTCO monomer in feed was in excess relative to the Tz monomer because some sTCO compound was ‘wicked away’ from the top oil phase as the fiber was drawn.

The bioorthogonal strategy not only permits straightforward synthesis of peptide-decorated microfibers, but also enables simultaneous incorporation of intact proteins during fiber pulling. As a proof-of-concept, superfolder green fluorescent protein (sfGFP) was used for in situ protein conjugation purposes. Following a reported procedure^[35], tetrazine was site-specifically encoded into sfGFP using a synthetic non-natural amino acid carrying methyl phenyl tetrazine (see methods in Supporting Information). Purified sfGFP-Tz (Figure S2B) was then added to the aqueous PEG-bis-Tz solution (at a final concentration of 1 μM), to which an organic layer containing tris-sTCO was introduced. Microfibers were collected similarly as described above. sfGFP-Tz combines with sTCO at a rate (k₂ of 87,000 M⁻¹s⁻¹ in PBS at pH 7, 21 °C ^[35]) that is competitive with the diphenyl counterpart, thereby enabling efficient protein conjugation to the polymer during fiber pulling. As shown in Figure 1E, the resultant microfibers displayed the characteristic GFP fluorescence homogeneously throughout the entire mesh, indicating successful protein incorporation in the hydrogel matrix without significant unfolding/denaturation. Physical adsorption of sfGFP to the microfiber is possible as sfGFP at the oil-water interface could non-covalently associate with the polymer film, thereby being removed from the interface during fiber pulling. To account for physical adsorption, interfacial polymerization was carried out using tetrazine-free sfGFP (Figure S2A). Analysis of the protein concentration (see methods in Supporting Information) in the aqueous reservoir before and after fiber pulling revealed that ~6% of the protein was consumed per mg of microfiber collected when Tz-free sfGFP was used. On the other hand, polymerization in the presence of sfGFP-Tz led to ~30% sfGFP consumption from the protein-containing aqueous reservoir. Analysis of protein release after 24-h incubation in a high salt (1 M NaCl), acidic buffer (pH 5.5) showed that fibers prepared using sfGFP and sfGFP-Tz contained a similar amount of physically incorporated protein (Figure S3). Thus, approximately 80% protein incorporated in fibers prepared in the presence of sfGFP-Tz was covalently conjugated. The bioorthogonal strategy can be readily applied to the immobilization of biologically active proteins, either individually or collectively, to recapitulate key elements of cellular microenvironment.

3.2. Fiber characterization.

3.2.1. Swelling, stability and crystallinity.

The stably crosslinked microfibers swelled spontaneously in water due to the presence of PEG segments. As assessed by microscopy (Figure 1F-G), the average diameter of dry fibers was 7.7 ± 0.5 μm, and upon swelling in water, the average diameter increased by 2.15 ± 0.07 fold. Control fibers synthesized using bis-sTCO had a larger diameter (10.5 ± 0.4 μm) and swelled more (2.54 ± 0.10 fold). Our observations are in line with the expectation of lower chain mobility in crosslinked networks. Crosslinked fibers maintained the structural integrity during long-term water exposure (Figure 1F). The fiber surface remained smooth and defect-free. The fiber diameter did not increase significantly over time, nor did the fibers degrade, making them suitable for long-term cell culture applications. Fibers derived from tetrazine monomers with or without the dangling RGD peptide exhibited a comparable stability. By contrast, microfibers derived from bis-sTCO do not exhibit desired aqueous stability. Immediately upon exposure to water, microfibers showed the presence of small voids/defects (arrows, Figure S4A). By day 5, crazes in the form of voids and fibrillary bridges perpendicular to the fiber length and spanning the entire width of the fibers were abundant (Figure S4B). As water diffuses through the fibers, a sharp boundary is created between the amorphous regions that are highly swollen and the crystalline regions that are not substantially swollen. The stresses at the swelling boundary cause the uncrosslinked microfiber to craze and fracture.^[36] Prolonged incubation of control fibers in water beyond 5 days led to fiber fragmentation, as a result of the untangling and disassociation of the PEG chains in the crystalline regions. Collectively, covalent crosslinking is necessary to stabilize the polymer chains within individual fibers.

Characterization by differential scanning calorimetry (DSC) (Figure S5A) confirmed the semicrystalline nature of the crosslinked fibers, as evidenced by the broad melting transitions (T_m) centered around 48 °C, with a melting enthalpy of 114 J/g. Results from Wide Angle X-Ray Diffraction (WXRD) (Figure S5C) showed sharp and intense peaks at 2Θ values of 19° and 23°, consistent with the WXRD pattern of pure PEG.^[37] Control bis-sTCO fibers exhibited a higher T_m (53 °C, Figure S5B), a lower melting enthalpy (96 J/g), and a similar WXRD pattern (Figure S5D). Thus, in both types of fibers, the crystalline domains originated from the PEG chains, whereas the hydrophobic sTCO building blocks resided in the amorphous regions. Crosslinking gave rise to smaller crystalline domains (thus lower T_m) by restricting the ordered packing of polymer chains.^[38] Covalent crosslinking, however, did not compromise the overall crystallinity.

3.2.2. Mechanical properties.

The mechanical properties of the microfibers were characterized using a micro-force tester that can detect forces with a 4 μN resolution.^[34] A single microfiber was carefully affixed to the cantilevered beam and the other end was taped to the X-Y stage, as shown in Figure 2A. After being aligned to the stretch axis (Z) and preloaded with a 150 μN force to ensure full fiber engagement, the microfibers were stretched by moving the cantilevered beam via a piezoelectric stage (0-800 μm) along the stretch axis (Z). The fiber displacement was calculated by taking the difference of the stage travel and cantilevered beam deflection, measured with a capacitance sensor. The force was determined by multiplying the beam deflection by the beam’s spring constant of 28.4 μN/μm. The initial length of the microfibers was measured following preloading and was found to be highly consistent across all tests, 4.3 ± 0.9 mm. The microfibers were then pre-conditioned by a single ramp from 0 to 3.5% strain and, finally, stretched in three consecutive ramps from 0.75 to 3% strain. After being tested under dry conditions, the same microfibers were fully hydrated by total immersion in a drop of water loaded on the stage as shown in Figure 2A. Similarly, the hydrated microfibers underwent three consecutive ramps from 0.75 to 12% strain. A total of 29 fibers of each type from two different batches were used for mechanical testing.

Figure 2.

Mechanical characterization of crosslinked microfibers. Control microfibers synthesized using bis-sTCO were included for comparison purposes. (A). Schematic of the micro-force tester used for single fiber tensile testing. Representative force-displacement curves for dry (B) and hydrated (C) fibers. Young’s moduli were determined by calculating the slope of the stress-strain curves on a fiber-by-fiber basis. Young’s moduli for individual dry (D) and hydrated (E) fibers were shown with error bars representing the uncertainty in the analysis, as described in the Supporting Information. The fiber populations mean ± 95% confidence interval are indicated. Statistical analysis of Young’s moduli between crosslinked and control fibers for dry and hydrated conditions were performed using student’s T-tests (#: p<0.01).

Representative force-displacement curves are displayed in Figure 2B and 2C for a dry and hydrated fiber, respectively. Hydration led to a 47 and 29-fold decrease in stiffness for fibers produced using tris- and bis-sTCO, respectively. The repeatability of the force-displacement curves for representative microfibers is shown in Figure S6. Our results indicate that microfibers can be stretched repeatedly without permanent deformation under these conditions. Linear fits to repeat force-displacement curves demonstrated that fiber stiffness varied by less than 5 % of the mean value. To make straightforward comparisons between fibers (i.e. adjust for differences in fiber diameter and initial length), stiffness values were converted into Young’s modulus (Δσ · Δε⁻¹). To calculate the change in fiber stress, the change in applied load was divided by the cross-sectional area of each fiber, where prior to tensile testing the diameter of each fiber was quantified using optical microscopy. The change in fiber strain was determined using the ratio of fiber stretch to its original measured length. Bis-sTCO-derived fibers had an average modulus (mean ± 95% confidence interval) of 229 ± 66 and 1.11 ± 0.40 MPa under dry and hydrated conditions, respectively. Tris-sTCO-derived fibers were significantly stiffer, both under dry (525 ± 161 MPa) and hydrated (3.93 ± 1.70 MPa) conditions.

Hydration dramatically reduced the stiffness for both types of microfibers (Figure S7); this is intuitive since the added water is not expected to contribute to mechanical properties. The marked softening effect of polymer hydration is thought to be due to the partial dissolution of the PEG crystallites during contact with water. ^[39] In addition, the hydrated polymer chains are inherently more slippery than dry polymer chains.^[40] Compared to the control microfibers, crosslinked fibers exhibited significantly higher Young’s moduli under both dry and hydrated conditions (Figure 2D-E). Thus, covalent crosslinking significantly increased fiber stiffness. Compared to bulk PEG gels synthesized using 40 wt% of PEG diacrylate with a molecular weight 10 kDa,^[41] the microfibers exhibit a higher Young’s modulus, possibly due to a higher crosslinking density, presence of crystalline domains and/or substantial chain alignment during fiber pulling.^[33]

3.3. Fiber-guided cell migration, proliferation and neotissue formation.

The crosslinked hydrogel microfibers are structurally sound and mechanically robust, thus can be easily manipulated as single fibers or fabricated into fibrous meshes. When decorated with appropriate biological signaling motifs, these fibers can be used to guide cell migration and to promote wound healing. Wound healing includes three distinct stages: inflammation, new tissue formation and tissue remodeling.^[42] During the initial stage, the damaged tissues are removed through inflammatory responses. Next, new tissues start to form through proliferation and migration of different cell types into the damaged region. Finally, new ECM proteins are laid down at the wound bed and the tissue is remodeled by the resident cells. Here, we demonstrate that the synthetic fibers not only facilitate cell attachment and proliferation but also guide cell migration and promote neotissue formation.

3.3.1. Cell attachment, proliferation and migration.

Using RGD-containing microfibers, we devised a single fiber migration experiment to test whether these fibers guided cell migration and proliferation. As shown in Figure 3A, NIH3T3 fibroblasts expressing GFP were seeded on a petri dish that contained a central PDMS-coated, non-adhesive area. After overnight culture, confocal microscopy revealed that a monolayer of attached cells had formed around the central non-adhesive area (Figure 3B). Then a single hydrogel microfiber fixed on a metal frame was laid on top of the cell layer carefully to connect the cell-rich regions through the central cell free area. For the RGD-free fiber, cells did not migrate onto the fiber above the non-adhesive area after 24 h cell culture (Figure 3C). By contrast, cell binding was observed on the RGD-containing microfiber spanning across the PDMS region. Within 24-72 h, cells had completely traversed the non-adhesive area (Figure 3D). The presence of cells on the microfiber above the cell-free zone of the culture surface suggests that cells from both ends have migrated and grown on the microfiber. Time-lapse confocal microscopy revealed that cells were motile on the microfiber surface (Video S3 and S5). These observations indicate that RGD-containing microfibers present the appropriate biophysical and biochemical cues for cell attachment and migration.

Figure 3.

Fiber-guided cell migration across a non-adhesive region. (A). Schematic of contact guidance provided by the RGD-containing microfiber. NIH3T3 fibroblasts were allowed to grow around a central non-adhesive region. Cells attached to and migrated along the microfiber suspended across the non-adhesive center. (B). Confocal image of NIH3T3/GFP fibroblasts attached around the PDMS-coated, non-adhesive circle. (C, D). Confocal images of NIH3T3/GFP cells cultured with an RGD-free (C) and RGD-containing (D) microfiber for 24 h. Images shown in C and D were made by stitching together maximum intensity projection of z-stacks from multiple tile scans using a 10× objective. DIC channel was overlapped with GFP channel for visualization of the RGD-free fiber in C. Scale bar: 200 μm.

Immunofluorescent staining for a cell proliferation marker ki67 showed positive staining on cells grown on 2D surface (arrows, Figure 4A) and on the microfiber (arrows, Figure 4B). The presence of a cell in anaphase indicates that cells are undergoing mitosis on the microfiber (arrowhead, Figure 4B). A mitotic cell in metaphase attached to the microfiber by a transient filopodia was distinctly seen rounded up, and then underwent cell division in the vicinity of the microfiber (Figure 4C). Following cell division, the two daughter cells migrated back on to the fiber. Filopodia are highly organized tightly cross-linked long bundles of unidirectional and parallel actin filaments.^[43] It is known that transient filopodia facilitate attachment of mitotic cells to their substrates,^[44] recognize topographical features and direct cell orientation and migration.^[45] It is interesting to note that following cell division, the filopodial structures reorganized into lamellipodia, which are thin sheet-like branched network of actin filaments involved in cell migration. Such transitions have been reported to control contact guidance of cells on nanofibrillar environments.^[45] In addition, time-lapse live cell recording over a 12-h period clearly showed cell mitosis at multiple locations along the microfiber (Video S2). Cells at different phases during cell cycle (such as metaphase and telophase) were also captured in a microtubule-stained confocal image (Figure S8).

Figure 4.

Fiber-supported cell proliferation. (A, B) Confocal images of fibroblasts attached to the surface outside PDMS-coated region (A) and on the microfiber (B). Arrows point to actively dividing cells stained red by ki67. Arrowhead denotes a cell in anaphase. Cell nuclei were counter stained blue by DAPI. (C). Images from a confocal microscopy time-lapse movie (Video S2) of GFP-expressing fibroblasts attached to the microfiber. The higher magnification images captured a cell detaching from the fiber, undergoing mitosis and returning to the fiber. Scale bar: 50 μm.

Inhibition experiments were conducted to gain mechanistic understanding of fiber-guided cell migration. Studies have shown that cell migration involves the development of leading and trailing edges. At the front, intense actin polymerization generates sufficient forces to promote cell protrusion, adhesion, contraction, and retraction.^[46-47] At the trailing edge, a highly dynamic microtubular network opposes the contractile forces to facilitate the retraction of the rear.^[48] To investigate the roles of actin and tubulin in cell migration, cells were allowed to migrate onto the fibers and Lan B^[49-50] or nocodazole ^[51] were added to the media to inhibit the development of stress fibers and microtubules, respectively. Video microscopy was utilized to monitor cell movement along the microfiber.

Addition of Lan B completely depolymerized actin filaments as revealed by fluorescent phalloidin staining. Compared to the control cells (Figure 5A), Lan B treatment effectively disrupted the long F-actin filament and replaced them with large aggregates of actin monomers (Figure 5B). Actin depolymerization resulted in the loss of cell morphology (Figure 5C-D). Cell morphology changed immediately upon the addition of Lan B and cells became completely round after 3 h of incubation (Video S4). Prior to Lan B treatment, bidirectional cell movement along the microfiber with extended cell morphology was obvious (Video S3). To quantify cell migration, representative cells were tracked and the distance each cell traveled relative to the origin in each frame was calculated (Figure 5E). Our analyses revealed an average migration velocity of 29.7 ± 2.4 and 12.2 ± 1.9 μm/h for tracked cells before and after Lan B treatment, respectively. These results supported a role for actin in fibroblast migration on the microfibers.

Figure 5.

Effects of Lan B treatment on cell migration along the microfiber. (A, B). Confocal images of fibroblasts cultured with an RGD-containing microfiber before (A) and after (B) Lan B treatment at 10 μM for 3 h. F-actin and nuclei were stained red and blue by phalloidin and DAPI, respectively. (C, D). Confocal live cell images of GFP-expressing fibroblasts cultured on an RGD-containing microfiber before (C) and after (D) Lan B treatment (10 μM, 3 h). Scale bar: 20 μm. (E). Quantification of cell migration before and after Lan B treatment from tracking of representative trajectories of 10 cells. The movement of individual cells was tracked every 15 min for 3 h (Video S3 and S4).

Without nocodazole treatment, tubulin assembled into filamentous intracellular structures that were parallel to the synthetic extracellular microfibers (Figure 6A). After cells were treated with 10 μM nocodazole for 3 h, tubulin disassembled and localized throughout cells homogeneously instead of existing as microtubule fibers (Figure 6B). The overall cell morphology didn’t change significantly upon nocodazole treatment (Figure 6C-D), in agreement with previous reports.^[52] Prior to nocodazole treatment, distinct bidirectional cell movement along the microfiber was observed (Video S5). Following nocodazole treatment, migration velocity and directionality were reduced (Video S6). Nocodazole treatment led to a reduction of velocity from 29.7 ± 2.4 to 15.4 ± 1.4 μm/h (Figure S9). Binding of nocodazole to tubulin affects the dynamic instability of microtubules, thereby compromising microtubule turnover and negatively impacting cell migration.^[53] Consistent with previous studies,^[54-55] the significantly reduced cell migration in nocodazole treated cells confirmed the involvement of microtubules in microfiber-guided cell migration.

Figure 6.

Effects of nocodazole treatment on cell migration along the microfiber. (A, B). Confocal images of fibroblasts cultured with an RGD-containing microfiber before (A) and after (B) nocodazole treatment at 10 μM for 3 h. F-actin and nuclei were stained red and blue by phalloidin and DAPI, respectively. (C, D). Confocal live cell images of GFP-expressing fibroblasts cultured on an RGD-containing microfiber before (C) and after (D) nocodazole treatment (10 μM, 3 h). Scale bar: 20 μm. (E). Quantification of cell migration before and after nocodazole treatment from tracking of representative trajectories of 10 cells. The movement of individual cells was tracked every 15 min for 3 h (Video S5 and S6).

Our results collectively show that the synthetic microfibers provide contact guidance to cells, eliciting multiple cellular processes simultaneously on the fiber, including cell proliferation, cell migration, turnover of substrate adhesion and reorganization of actin filopodia to lamellipodia. The ability of the mitotic cell to maintain contact with the microfiber through a filopodia is crucial for the maintenance of tissue integrity. These observations have not been reported previously in any synthetic fibrous materials and cannot be reproduced using natural ECM proteins owing to the challenges in purifying and manipulating these proteins. Inhibition of actin and tubulin polymerization with drugs again validate that movement, and proliferation on the fiber is actin and microtubule dependent. Thus, the RGD-containing microfibers provide topographical features similar to those present in the native tissue environment to promote cell migration, a critical process involved in wound healing.

3.3.2. Neotissue formation.

The crosslinked microfibers were designed for long-term cell culture purposes. To this end, cells were incubated with the microfiber for 7 days, stained for F-actin and nuclei and visualized by confocal microscopy (Figure 7A). Complete fiber coverage by fibroblasts was obvious. A higher magnification Z-stack image (Figure 7B) showed cells developed extensive stress fibers and stacked on top of each other. Detailed and pronounced cell structures were revealed by the reconstituted image as a heatmap (Figure 7C). Cells on the outer surface displayed cellular extensions beyond the confines of the fiber. A cross sectional view of the reconstructed image clearly showed that cells attached around the microfiber, not just on the top edge (Figure 7D). Cells formed a tissue-like, multicellular and multilayered structure with an estimated thickness of 40 μm encasing the 40-μm thick fiber core. Collectively, the RGD-containing microfibers provides an instructive track that not only promotes cell attachment and division but also facilitates the establishment of organized cellular structures.

Figure 7.

Fiber-supported formation of multilayered neotissue. (A). Confocal image of fibroblasts after a 7-day culture with an RGD-containing microfiber. F-actin and nuclei were stained red and blue with phalloidin and DAPI, respectively. The image was stitched from the maximum intensity projection of eight consecutive tiles obtained using a 10× objective. Scale bar: 200 μm. (B). Maximum intensity projection of a Z-stack confocal image showing the multilayered, multicellular structure. Scale bar: 20 μm. (C). 3D visualization of the cellular construct after deconvolution and reconstruction of Z-stack confocal images. (D). Cross sectional view of the cellular construct showing multiple layers of cells encasing a fiber core (appeared black in the image).

Microfibrous scaffolds consisting of multiple layers of orthogonally aligned fibers were prepared by alternating the axis of the collecting frame during the fiber pulling process. The microfibrous scaffold fixed on a metal frame were then laid on top of a PDMS-coated cell culture dish. In the presence of RGD-lacking fibers, the majority of cells remained suspended in media and were removed during washing or sample transfer. Although some cells were able to anchor to the edges of individual fibers, they could not populate the entire scaffold even after 7 days of culture (Figure 8A-C). Random cell aggregates with a diameter of 60-200 μm were seen on the culture dish, confirming the non-adhesive nature of the fibrous scaffold. By contrast, RGD-containing meshes served as a provisional matrix to promote cell attachment, proliferation and neotissue formation. Two days post seeding, a large number of cells were anchored on individual fibers, with individual cell bodies aligned along the fiber direction (Figure 8D, G). A base cell monolayer also formed on the underlying PDMS surface. In agreement with our previous observations, the cell adhesive microfibers served as contact points to bridge cells across the underlying cell repellent surface. By day 7, the entire scaffold became densely populated by cells to establish a 3D interconnected cellular network (Figure 8F, I). The 3D structures formed based on microfiber templates can be observed more clearly by Z-stack imaging of cells above the bottom cell monolayer (Figure 8 G-I). Maximum intensity projection of the z-stack confocal image over entire scaffolds (Figure S10) showed that a three-dimensional neotissue of 400-μm thick formed within the scaffolding material on day 7. We are not limited to such a tissue thickness as thicker scaffolds can be produced readily by extending the duration for fiber collection. Under our experimental conditions, micrfibers can be continuously collected until the monomers in the reservoirs were exhausted.

Figure 8.

Neotissue formation on microfiber scaffolds. Confocal live cell images of GFP-expressing fibroblasts cultured on RGD-free (A-C) and RGD-containing (D-I) scaffolds on day 2, 4 and 7. Scale bar: 200 μm. Images are maximum intensity projection of Z-stacks. Images in D-F show the entire cellular construct consisting of cells growing on individual fibers, as well as the cell monolayer underneath the scaffold. Images in G-I only shows cells growing on the scaffold.

In summary, interfacial polymerization/crosslinking enabled facile synthesis of robust hydrogel microfibers in one-step from monomeric building blocks without the need for postpolymerization processing steps. Unlike amorphous bulk hydrogel materials, the hydrogel microfibers are anisotropic, thus promoting the directional migration of cells through contact guidance. The fibrous scaffold provides an instructive framework that not only promotes cell attachment and division but also facilitates the establishment of organized cellular structures. The materials choice for the current fiber system is guided by known knowledge of PEG, a biocompatible “blank slate” widely used in various biomedical applications.^[56] The hydrophobic tris-sTCO monomer was used to introduce stable crosslinks and to alter the physical properties of the polymer fibers. The synthetic strategy overcomes limitations associated with animal-derived collagen scaffolds, ^[57-58] with added advantages of tunability in fiber composition. Both peptide fragments and intact proteins can be immobilized in the fiber during fiber pulling.

Under the current laboratory conditions, fiber production is low throughput, fiber diameter cannot be easily manipulated and scaffolds produced is pseudo-3D. The polymerization process can be scaled up through process optimization since interfacial polymerization is widely used for the synthesis of polyamide, polyurethane, polyester and polycarbonate, both at the laboratory and industrial scales.^[59-60] Production of submicron-sized fibers is possible via the combination of interfacial polymerization electrospinning or through the usage of microfluidic devices. Composite 3D constructs can be generated by adapting microscale 3D textile weaving techniques, and infiltrating the woven scaffold with a cell/gel mixture.^[61] With further development, we project utility of this type of materials in tissue repair and wound healing applications.

4. Conclusions

We have developed a novel strategy to produce crosslinked hydrogel microfibers via a diffusion-controlled, interfacial crosslinking mechanism. The enabling chemistry is the rapid tetrazine-TCO reaction. Fibers pulled out of the oil-water interface exhibit robust mechanical properties and excellent aqueous stability. This strategy allows straightforward incorporation of biological motifs using tetrazine-conjugated peptide or protein monomers without relying on chemical modification post fiber production. The RGD-containing microfiber not only supports cell attachment and facilitates cell proliferation, but also promotes directional migration of fibroblasts through a non-adhesive region. With excellent stability, due to the crosslinked microfiber structure, fibroblasts can build their own structures using microfibers as a 3D template over long-term culture. The bioorthogonally produced hydrogel microfiber scaffolds have the potential to serve as instructive matrix to facilitate tissue repair and regeneration.