Injectable pH-responsive adhesive hydrogels for bone tissue engineering inspired by the underwater attachment strategy of marine mussels

Abstract

A major challenge in tissue engineering is the development of alternatives to traditional bone autografts and allografts that can regenerate critical-sized bone defects. Here we present the design of injectable pH-responsive double-crosslinked adhesive hydrogels inspired by the molecular mechanism and environmental post-processing of marine mussel adhesive. Nine adhesive hydrogel formulations were developed through the conjugation of crosslinkable catechol functional groups (DOPA) and the synthetic oligomer oligo[poly(ethylene glycol) fumarate] (OPF), varying the DOPA content (w/w%) and molecular weight (MW) of the OPF backbone to produce formulations with a range of swelling ratios, porosities, and crosslink densities. DOPA incorporation altered the surface chemistry, mechanical properties, and surface topography of hydrogels, resulting in an increase in material stiffness, slower degradation, and enhanced pre-osteoblast cell attachment and proliferation. When injected within simulated bone defects, DOPA-mediated interfacial adhesive interactions also prevented the displacement of scaffolds, an effect that was maintained even after swelling within physiological conditions. Taken together, OPF-DOPA hydrogels represent a promising new material to enhanced tissue integration and the prevention of the post-implantation migration of scaffolds that can occur due to biomechanical loading in vivo.

1. Introduction

The need for transplants to replace bone lost to osteoporosis, trauma, and surgical removal continues to be a major problem, with over one million bone fractures imposing an estimated $3 billion USD economic burden on the United States each year [1,2]. Particularly difficult cases are criticalsized defects, where the amount of bone lost cannot be replaced by the natural healing process [3] without being filled by a material that provides mechanical support [4]. Bone autografts remain the standard of care in this case, necessitating that a patient’s own tissue be transplanted into the defect. However, harvesting this tissue requires an additional surgical procedure that cannot always supply the quantity of material needed and increases the chances of post-operative pain and complications [5]. As the rate of transplants increase to meet the demands of aging populations globally, cost-effective scaffolding materials that support bone regeneration are needed.

A variety of alternatives to traditional bone grafts have been proposed within the field of tissue engineering, including implants comprised of metals [6,7], ceramic/bioactive glass [8], synthetic polymers [9,10], natural materials [11], and synthetic/natural composites [12–14]. Successful scaffolding candidates must first and foremost support cell proliferation and differentiation [15], while also providing adequate space for vascularization and eventual bone regeneration and tissue integration [16–18]. Additionally, the biocompatibility and degradation kinetics of scaffolds must facilitate, rather than interfere with, the bone remodeling process [6,12–14,19], while also preventing bacterial infection [20,21] and tumor growth [22]. Similarly, bone tissue remodeling is particularly dependent on the precise timing and concentration of multiple growth factors, necessitating that scaffolds support their delivery during degradation [23]. Considering these challenges, promising candidates for the next generation of bone scaffolds are biocompatible and injectable synthetic polymers [24]. The surgical implantation of preformed scaffolds remains an invasive procedure that increases the likelihood of infection and postoperative movement of the implant [16,25]. Addressing the inflexibility of rigid implants, injectable polymers can crosslink in situ and form cellular scaffolds that closely match the shape and size of a defect via minimally invasive surgery.

One biocompatible, biodegradable, and injectable synthetic polymer that shows promise for bone tissue engineering applications is oligo [poly(ethylene glycol) fumarate] (OPF) [26]. Synthesized through the combination of fumaryl chloride and poly(ethylene glycol) (PEG), OPF is cost-effective, easy to manufacture, and can display a wide variety of chemical and physical properties through copolymerization [27–31]. Utilizing the unsaturated double bonds present in fumarate, OPF hydrogels can be formed after injection in situ using either a chemical, thermal, or UV crosslinking method, while the repeating ester bonds throughout the oligomer facilitate rapid biodegradation through hydrolysis under physiological conditions [32]. Our lab has produced several OPF-based bone scaffolds that support bone regeneration through the incorporation of bioactive molecules such as growth factors that facilitate osteoblast proliferation, differentiation, and implant mineralization [33–35]. However, despite these successes, one drawback to these injectable formulations is that they do not adhere to surfaces, particularly under wet-conditions. Instead, OPF hydrogels typically absorb water and swell under physiological conditions, increasing the likelihood that implants move away from the bone-implant interface after injection, especially when subjected to biomechanical loading [36–39].

In this study, we propose a solution to the post-implantation displacement of bone scaffolds through the development and evaluation of an injectable adhesive OPF hydrogel modeled after the underwater attachment strategy of marine mussels (Fig. 1A) [40]. Mussels adhere to the seafloor to resist dislodgement by ocean waves by secreting fibers (byssal threads) from their foot, which they tether to substrates using a protein-based adhesive (adhesive plaque, Fig. 1B). Molecular characterization of the proteins that make up the adhesive plaque suggest that the prevalence of l-3,4-dihydroxyphenylalanine (DOPA) residues (Fig. 1C) perform interfacial reactions (hydrogen bonding, coordination, etc.) with surfaces when reduced [41,42], and covalent crosslinks with each other when oxidized [43,44]. To form an effective adhesive, mussels control the sequencing of these two behaviors by tightly regulating the chemical environment surrounding the plaque [45] (Fig. 1B), maintaining an acidic environment (pH 2–3) with low ionic strength (0.15 M) during secretion [46–48], and then exposing the material to basic (pH 8.1), ionically saturated (0.7 M) seawater [49]. This transition acts as a molecular trigger during adhesive curing, increasing the adhesion strength of the plaque by increasing the cohesive strength of the protein network after the glue has made contact with a surface [50,51].

Fig. 1.

Conceptual diagrams of adhesive hydrogel design. (A) Marine mussels attach to the seafloor using byssal threads tipped with an adhesive plaque. (B) The adhesive plaque is made up of proteins that are deposited from secretory glands in the mollusk’s foot. Diagram of mussel foot protein (Mfp-5) adorned with adhesive l-DOPA residues that form interfacial bonds with surfaces and dopaquinone crosslinks (C). Mussel-inspired hydrogel comprised of synthetic polymers (D) that forms primary (1°) crosslinks upon injection and secondary (2°) crosslinks after pH adjustment and oxidation within the human body (E).

Mussel adhesive has inspired the development of an array of surface coatings and medical adhesives that display impressive adhesion to both wet and dry surfaces [52–57], antimicrobial activity [20,21,58,59], and self-healing properties [60]. Here we present an improvement upon previous methods by adapting the ‘molecular trigger’ model used by mussels during adhesive protein secretion to develop an injectable tissue implant system. By controlling the pH and redox conditions prior to injection, we describe the design and performance of injectable DOPA functionalized OPF hydrogels that are capable of rapidly crosslinking in situ. Post-injection, implants were able to successfully adhere and remain attached to simulated bone defects, undergoing secondary crosslinking and mechanical stiffening after exposure to physiological conditions. Using material characterization, mechanical testing, and in vitro cell compatibility assays we then show how the incorporation of DOPA alters the physical characteristics of hydrogels, preventing the displacement of scaffolds from bone defects, while also promoting the attachment and proliferation of pre-osteoblast cells. DOPA incorporation into OPF based polymers is a novel direction that, when used in combination with injectable formulations under physiological conditions, represent an exciting method for tissue implant generation with the potential to side step some of the commonly encountered pitfalls inherent in other systems.

2. Materials and methods

2.1. Polymer synthesis and characterization

Oligo[poly(ethylene glycol) fumarate] (OPF) was synthesized from poly (ethylene glycol) (PEG) and fumaryl chloride in methylene chloride using previously described methods [26,61,62]. Three OPF oligomers were synthesized by varying the number average molecular weight (M_n; hereafter referred to as the PEG MW) of the PEG starting material (1000; 3000; or 10,000). Dopamine methacrylate (DMA) was synthesized through the combination of dopamine hydrochloride and methacryloyl chloride in methanol following a modified version of the protocol outlined by Xu et al. (2012) [63]. Poly(ethylene glycol) diacrylate (PEG-DA) with a M_n of 575 was purchased from Sigma-Aldrich (Milwaukee, WI). The structure and molecular weight of synthesized polymers were confirmed with ¹H NMR (Figs. S1–4) and gel permeation chromatography (GPC; Table S1). Additional information regarding polymer synthesis and characterization is provided in Sections 1.1 and 1.2 of the Supplementary material.

2.2. Injectable OPF-DOPA hydrogel formulations

Nine OPF-DOPA hydrogel formulations were developed by varying the PEG MW used to synthesize OPF (x = 1 k, 3 k, or 10 k) and the w/w% of DMA to OPF (y = 0%, 10%, or 20%), hereafter referred to using the nomenclature OPF_x-DOPA_y (Table 1). Hydrogels were formed by dissolving OPF, DMA, and 10% w/w PEG-DA to OPF in acidified (pH 3) distilled and deionized water (ddH₂O) at a total w/v ratio of 1:2. Polymers were added to solvents under flowing nitrogen to ensure limited oxidation, and stirred with a stir bar until completely dissolved. Crosslinking was induced through the addition of 10 mM of both the crosslinking initiator ammonium persulfate (APS) and the crosslinking accelerator l-ascorbic acid (AA). In combination, AA/APS produce both ionic and organic free-radicals that initiate the polymerization process. The use of an AA/APS-Redox system has been found to reliably generate hydrogels from OPF and PEG-DA starting materials by our laboratory and others [64–66], and when used at concentrations of 10 mM have minimal impact on solution pH and the survival of marrow stromal cells and neonatal human dermal fibroblasts [66,67]. Immediately following the addition of redox reagents, hydrogel formulations were quickly (<10 s) injected onto a substrate or into molds to form hydrogels (Fig. 2A–B).

Table 1.

Injectable formulations of OPF-DOPA hydrogels. PEG nominal MW refers to the number average molecular weight of the PEG used to synthesize oligo(poly(ethylene glycol) fumarate) (OPF). Dopamine methacylate (DMA) is reported as the w/w%ofDMA to OPF. The mean ± SE is reported for each metric. Formulations that share letters were not significantly different as determined by a post-hoc Tukey HSD test (alpha = 0.5).

Hydrogel formulation	PEG nominal MW	DMA	Sol Fraction	Mass Swelling Ratio	Gel Point (s)	Crosslinked (min)
OPF1k	1000	–	17.1 ± 3.3	4.7 ± 0.6abc	24 ± 3a	5.9 ± 1.2a
OPF1k-DOPA10		10%	16.9 ± 1.1	4.6 ± 0.2ab	35 ± 5a	6.2 ± 0.7a
OPF1k-DOPA20		20%	17.1 ± 5.4	2.8 ± 0.4d	42 ± 5a	10.7 ± 2.8b
OPF3k	3000	–	16.9 ± 2.1	8.9 ± 0.3a	21 ± 6a	8.1 ± 0.7b
OPF3k-DOPA10		10%	25.1 ± 4.7	4.5 ± 0.3abc	24 ± 2a	8.6 ± 0.7b
OPF3k-DOPA20		20%	23.7 ± 4.0	3.6 ± 0.2d	55 ± 3b	11.4 ± 0.5b
OPF10k	10,000	–	15.2 ± 3.0	26.0 ± 4.5e	23 ± 4a	9.0 ± 0.3b
OPF10k-DOPA10		10%	17.0 ± 2.4	4.4 ± 0.1abc	46 ± 3b	8.0 ± 0.9b
OPF10k-DOPA20		20%	10.4 ± 2.8	4.2 ± 0.1bcd	76 ± 6c	14.1 ± 0.7c

Fig. 2.

(A) Injectable OPF hydrogel formulations with PEG-DA, DMA, ascorbic acid, and ammonium persulfate dissolved in acidified, nitrogen flushed H₂O. After injection, hydrogels with DMA adhered to surfaces within minutes (B) as primary (1°) chemical crosslinks were formed between the OPF backbone and PEG-DA. (C) ATR-FTIR spectrum showing DMA incorporation into the OPF_1k oligomer. (D) Rheological test measuring the storage modulus (G’, Pa), loss modulus (G”, Pa), and viscosity (tan δ) of hydrogel formulations as a function of time after injection (without environmental modification). Storage modulus (E), loss modulus (F), and viscosity (G) of hydrogel formulations made from PEG with different molecular weights (1k, 3k, or 10k) and crosslinked with either 0, 10, or 20 w/w% DMA. Asterisks mark formulations that are statistically different than the control.

2.3. Primary (1°) chemical crosslinking

The rheological properties of OPF-DOPA formulations were determined using a torsional dynamic mechanical analyzer (Discovery Hybrid Rheometer HR-1; TA Instruments, New Castle, DE), tracking the 1° crosslinking reaction after the addition of 10 mM AA/APS. Formulations were injected between two 10 mm plates and compressed with a constant force of 1.5 ± 0.1 N, measuring the storage (G’ and loss (G”) moduli as crosslinking took place. The viscosity (tan δ) of each formulation was defined as the ratio of G”/G’. The timepoint where formulations underwent gelation (gel point) was defined as the intersection of the storage and loss modulus curves, as marked by a steep reduction in viscosity (see Fig. 2C). Crosslinking time was measured as the timepoint where the derivate of the storage and loss moduli crossed zero. Results are presented as the average of 3–4 samples for each formulation.

2.4. Secondary (2°) environmental crosslinking

The 2° environmental crosslinking dynamics of OPF-DOPA hydrogels were approximated using OPF-DOPA copolymer solutions produced through thermal polymerization in dimethylformamide (DMF) with azobisisobutyronitrile (AIBN; Sigma-Aldrich). Additional information regarding copolymer synthesis and characterization is provided in Sections 1.1 and 1.2 of the Supplementary material. The environmental conditions sufficient to produce dopaquinone formation and DOPA crosslinking (phenol coupling) were investigated by monitoring changes in the absorbance spectrum of OPF-co-DMA (0.6 M) in phosphate buffered saline (PBS) over time following the protocol outlined in Xu et al. (2012) [63]. The pH (3.0, 5.0, 7.0, or 9.0) or redox conditions (NaIO₄:DOPA molar ratio = 0:1, 0.5:1, 1:1, 2:1, 4:1; pH 7.4) of copolymer solutions were adjusted through the addition of concentrated HCL or the oxidizing agent sodium periodate (NaIO₄), measuring the absorbance spectrum between 200 and 600 nm, 5 min and/or 24 h after adjustment using a SpectraMax Plus 384 UV/VIS spectrophotometer (Molecular Devices, San Jose, CA).

2.5. Mechanical and physical properties

The mechanical properties of hydrogels were determined using dynamic mechanical analysis (DMA). Hydrogels were produced through the addition of 10 mM AA/APS to OPF-DOPA formulations and subsequent injection between two polypropylene plates fitted with a 1 mm rubber spacer. Plates were bound with binder clips and incubated for 1 h at 37 °C. For mechanical analysis, resulting gels were then swollen in PBS_pH3 for 4 h, cut into 8 mm discs, and either returned to PBS_pH3 or soaked in PBS_7.4 or PBS_7.4 with 10 μM NaIO₄ for 24 h at 37 °C. The compressive moduli of swollen hydrogels were determined using a dynamic mechanical analyzer (DMA-2980, TA Instruments, New Caste, DE) fitted with an 8 mm plate that applied a load at a rate of 5 mm min⁻¹. A stress-strain curve was generated for each sample, measuring the compression modulus as the slope of the linear region of the curve. Results are presented as the average of 3–4 samples for each formulation.

The physical properties of OPF-DOPA hydrogels were investigated using a series of swelling experiments wherein the weight of 8 mm discs was determined before and after swelling in ddH₂O and hexane [65]. Using a modification of the Peppas-Merrill equation that assumes a Gaussian distribution of chain lengths between crosslinks [68,69], these results were then used to calculate the volumetric swelling ratio (Q), mesh size (ξ), molecular weight between crosslinks (Mc), and crosslink density (ρx) of each hydrogel formulation using methods previously validated for fumarate-based gel networks [70,71]. As mesh size is a measure of the linear distance between two adjacent crosslinks and is functionally linked to the rate of solute diffusion [72], and it was treated as a reliable measure of average pore size and an indicator of porosity more broadly [73]. Additional information about the procedure used to calculate the physical parameters of hydrogels is available in Section 1.3 of the Supplementary material.

2.6. Degradation kinetics

The degradation kinetics of OPF-DOPA hydrogels were investigated following the protocols outlined in Shin et al. [32], measuring the mass loss (%) and mass swelling ratio of formulations over time. Briefly, 8 mm hydrogel discs were produced as previously described, dried, and weighed before being placed in microfuge tubes containing 1.3 ml of PBS_pH7.4. Hydrogels were then incubated at 37 °C for up to 4 weeks, sampling at either 1–7, 14, 21, or 28 d (n = 3 per formulation, per timepoint). The PBS solution in each tube was changed daily during the first week, and then weekly thereafter, measuring the pH of the solution upon collection. The pH variation of the solution was defined as the difference between the measured value and a pH of 7.4. Additional information regarding the experimental design of degradation experiments can be found in Section 1.4 of the Supplementary material.

2.7. Adhesive testing

The ability of OPF-DOPA hydrogels to adhere to dry and wet surfaces was assessed using lap-shear testing following the ASTM standard protocol F2255-05. Efforts were taken to approximate the conditions present during surgical implantation, wherein an adhesive implant is applied under semi-dry conditions and is exposed to body fluids after wound closure. Hydrogel formulations mixed with the crosslinkers AA/APS were injected between two glass microscope slides (placed in a plastic holder to prevent movement) with a 1 cm overlap and a width of 2.5 cm. A 5 N load was then applied by gripping the two slides with a modified clothes pin. Adhesive hydrogels cured at 23 °C for 1 h before being separated into one of two groups. The first group was tested immediately (dry adhesion), while the second was submerged in PBS_pH7.4, incubated at 37 °C for 24 h, and then tested (wet adhesion). Testing was performed by loading the adhered glass slides into the grips of a tensiometer fitted with a 25 lb. load cell, removing the clothes pin, and pulling the slides apart at a rate of 5 mm min⁻¹. The maximum of the force reported was divided by the overlap area to calculate the maximum shear stress, reported as the average of 3–5 samples.

The ability of OPF-DOPA hydrogels to remain within bone defects after implantation was investigated with push-out testing following the protocol outlined in Spicer et al. [74]. Bovine femurs were obtained from local commercial sources with the bone marrow removed. The femoral shaft was then milled into 30 mm square pieces using a rotary saw, shaving pieces to a 2 mm thickness to mimic the size and shape of calvarial critical size defects [74]. One side of each fragment was closed with parafilm tape and defects were filled with hydrogel formulations, injected within the center of defects and allowing the material to flow towards edges (0.3 ml). Once injected, hydrogels were allowed to dry-cure for 1 h before being incubated in PBS_pH7.4 for 24 h at 37 °C. After removing the parafilm tape, compression testing was performed with a dynamic mechanical analyzer (DMA-2980, TA Instruments) fitted with an 8 mm plate that applied a load at a rate of 0.5 mm min⁻¹. The peak stress required to dislodge the material was recorded for each sample, with an average of 3–5 samples reported for each hydrogel formulation.

2.8. Surface characterization

The surface morphology of OPF-DOPA hydrogel formulations was compared using scanning electron microscopy (SEM). OPF-DOPA hydrogel discs were manufactured as previously described, swollen in PBS_pH3 for 24 h, and then lyophilized for 48 h. Dried samples were mounted to microscope stubs with carbon tape and sputter coated with Au/Pd before being imaged at a magnification of 25×, 200×, or 1000× at 5 kV with a Hitachi S-4700 Scanning Electron Microscope (Hitachi High Technologies, Tokyo, Japan).

Due to the challenges of accurately measuring the surface characteristics of hydrated materials with techniques like atomic force microscopy [75], the surface roughness of OPF-DOPA hydrogels was approximated using a protein adsorption assay [76]. Hydrogel discs were incubated in 10% FBS for 4 h (n = 5 per group), measuring the total adsorbed protein with the MicroBCA protein assay (Pierce, Rockforld, IL). To isolate the effect of DOPA mediated protein adsorption from the effect of DMA mediated alterations in hydrogel surface roughness, adsorption assays were conducted with discs swollen in PBS_pH3 (uncrosslinked) and PBS_pH7.4 + 10 μM NaIO₄ (crosslinked). Additional information regarding protein adsorption experiments can be found in Section 1.5 of the Supplementary Material.

2.9. Cytocompatibility and cytotoxicity

The cytocompatibility of OPF-DOPA hydrogel formulations was assessed by culturing cells on hydrogel disc swollen in PBS_pH3 (uncrosslinked) and PBS_pH7.4 + 10 μM NaIO4 (crosslinked) using the MC3T3-E1 pre-osteoblast cell line [77], measuring cell proliferation after 1, 3, or 7 d with the QuantiT PicoGreen dsDNA assay kit (Invitrogen). Cell coverage on the surface of hydrogels was also investigated using the LIVE/DEAD viability/cytotoxicity kit (Molecular Probes, Eugene, OR) at each timepoint, followed by imaging with an LSM 780 confocal microscope (Zeiss, Oberkochen, Germany); live cells were stained green while dead cells appeared red. Additional information regarding the measurement of cell proliferation can be found in Section 1.6 of the Supplementary Material.

The cytotoxicity of OPF-DOPA hydrogel formulations was investigated by exposing MC3T3-E1 cells grown on 12-well tissue culture polystyrene (TCP) plates to the leaching solution of hydrogel discs. Transwells chambers (mesh size, 3 μm) containing hydrogel discs (8 mm) were fitted in 12-well TCP plates seeded with 30,000 cells, measuring cell proliferation after 4 d with PicoGreen. The cytotoxicity of the oxidizer sodium periodate (NaIO₄) was also assessed by culturing cells on 12-well TCP plates in media containing either 0, 10, 25, 50, or 100 μM NaIO₄. Cell viability (%) was calculated as the ratio of a given treatment to the positive control and reported as the average of 3–4 replicates.

2.10. Statistical analysis

Statistical analyses were performed using R (Version 3.4.1; http://www.r-project.org/) with the RStudio IDE (Version 1.0.153; http://www.rstudio.com/). Analysis of variance (ANOVA) was used to investigate differences in response variables across polymer formulations. Models were constructed using a combination of PEG MW (3 levels: 1k, 3k, or 10k), w/w% DMA to OPF (3 levels: 0%, 10%, or 20%), time (3 levels: 1, 3, or 7 d), or oxidation treatment (2 levels: untreated or treated with NaIO₄) as factors. Before analysis, the assumptions of normality and homoscedasticity were confirmed using the Shapiro test and a visual assessment of Q-Q and residual-fitted plots. When normality was violated, as was often the case with results expressed as a percentage, variables were transformed using the Johnson transformation package [78]. For significant effects (α = 0.05), the agricolae package was used to perform pairwise comparisons of groups using the Tukey HSD post hoc test [79].

3. Results and discussion

Oligo(poly(ethylene glycol) fumarate) (OPF) has received interest for tissue engineering applications as a synthetic polymer because it is biocompatible, biodegradable, and has a low production cost [27,28,33–35,80,81]. However, without modification, OPF-based materials are smooth and neutrally charged, resulting in poor cell attachment, limited integration with surrounding tissues, and the migration of implants under biomechanical loading [82]. Here we outline the design of an injectable OPF-based bone scaffold that addresses these issues by introducing adhesive catechol motifs through the conjugation of OPF and 3,4-dihydroxyphenyl-l-alanine (DOPA) (Fig. 1). In recent years, a number of DOPA functionalized materials inspired by mussel adhesive proteins (Fig. 1C) have been developed that display impressive adhesive properties [54,57,83–86]. We propose an improvement upon these methods by mimicking the sequenced application undertaken by the animal in nature (Fig. 1B) [46,87], resulting in an injectable, double-crosslinked adhesive hydrogel (Fig. 1E) that stiffens under physiological conditions and can support osteoblast attachment and proliferation for tissue regeneration.

3.1. Preparation of OPF-DOPA hydrogels

Inspired by the molecular trigger model used by marine mussels, we designed OPF-DOPA hydrogels that can be injected into bone defects and stiffen in reaction to the pH and oxidative conditions present within the body (Fig. 1E). To do this, OPF was synthesized from fumaryl chloride and PEG with three different number average molecular weights (1k, 3k, and 10k) through a one-pot reaction in methylene chloride with K₂CO₃ added as a proton scavenger [26,61,62]. These reactions generated three OPF oligomers with diverse molecular weights (GPC, Table S1) but similar structures – arrays of linked-PEG chains with unsaturated double bonds incorporated at different intervals (¹H NMR, Figs. S1–S3). A crosslinkable form of DOPA, dopamine methacrylate (DMA), was then synthesized from dopamine hydrochloride and methacryloyl chloride in methanol with triethylamine, taking care to control the pH and redox conditions of the reaction to prevent inactivation of the hydroxyl groups adorning the phenyl ring (Fig. 1; GPC, Table S1; ¹H NMR, Fig. S4). The crosslinkability of OPF and DMA was subsequently confirmed through copolymerization with the thermal initiator AIBN and comparison of ATR-FTIR (Fig. 2c) and ¹H NMR spectra (Fig. S5).

From these constituents, nine OPF-DOPA hydrogel formulations were developed by varying the w/w% of DMA to OPF (0%, 10%, or 20%) and the PEG MW used to synthesize OPF (see Table 1). This factorial design produced formulations with different spacing of DOPA functional groups, a factor that impacts the adhesive properties and crosslinking behavior of the material [88]. Hydrogels were prepared by dissolving OPF and DMA in the correct proportion with 10% w/w poly(ethylene glycol) diacrylate (PEG-DA) in N₂-flushed, acidified ddH₂0 (1:2 w/v) mimicking the redox and pH conditions under which mussel adhesive is secreted onto surfaces [46,89]. After homogenization, formulations were placed in a syringe, mixed with 10 mM ascorbic acid (AA) and ammonium persulfate (APS), and injected onto surfaces, or into molds, to form discs (Fig. 2A).

The addition of chemical crosslinkers performed two functions. First, as free radicals are generated, acrylate groups present in PEG-DA interact with the unsaturated double bonds in the OPF backbone, forming primary (1°) crosslinks between oligomer chains and inducing gelation within seconds (Table 1, Fig. 2A–B). Monitored with a rheometer over time (Fig. 2D), crosslinking of the hydrogel network was achieved in 6–15 min depending on the PEG MW and DOPA content (Table 1). Second, the methacrylate group of DMA crosslinks with PEG-DA and OPF, leading to the formation of a DOPA adorned network (Fig. 1E). Due to the heavily protonated environment, the reaction of AA/APS at this concentration did not significantly impact the oxidation state of DOPA functional groups, as confirmed by UV/VIS spectra of an OPF-DOPA copolymer before and after addition (Fig. S6).

PEG MW and DOPA content influenced the reaction kinetics and physical properties of the gel network during the 1° crosslinking reaction. As PEG MW used to synthesize OPF increased, the storage (Fig. 2E; p < 0.001) and loss moduli (Fig. 2F; p < 0.001) of hydrogels decreased, while the gelation time (Table 1; p < 0.001), crosslinking time (Table 1; p = 0.031), and viscosity increased (Fig. 2G; p < 0.001). DMA content had a synonymous effect on all these parameters, with a notable decrease in the storage and loss moduli when DOPA was incorporated into the 1k and 10k formulations (compared to the control). These decreases may be due to competition for crosslinking sites between DMA and PEG-DA, but were minimal when compared with the effect of PEG MW. A complete list of the results of statistical analyses comparing these groups can be found in Table S2 in the Supplementary material.

3.2. Secondary (2°) crosslinking under physiological conditions

In the mussel adhesive system, oxidative crosslinking (phenol coupling) of DOPA functional groups is facilitated by an increase in environmental pH after plaque deposition [46,89]. For OPF-DOPA hydrogels, an analogous process occurs after wound closure when implants absorb fluid from surrounding tissues and experience an increase in environmental pH, oxygen, and temperature. To investigate the effect of this process, we monitored pH-induced changes in the UV/VIS spectra of an OPF-DOPA copolymer solution (Fig. 3), as well as the swelling dynamics (Table 1) and mechanical properties of OPF-DOPA hydrogel discs (Fig. 4). As the pH of the copolymer solution was step adjusted from 3 to 9, a whole absorption range increase from 250 to 350 nm was immediately evident (Fig. 3C), with the spectra of the pH 7 and 9 treatments clustering away from the acidic treatments after being held at 37 °C for 24 h (Fig. 3C). The presence of a widened absorbance peak at 268–274 nm was also evident in the pH 7 and 9 treatments, indicating the formation of multiple dicatechol variants via oxidation and phenol coupling over time [90].

Fig. 3.

Environmentally induced secondary (2°) crosslinking of OPF-DOPA hydrogels. (A) Diagram of phenol coupling between Dopa residues on chains within the OPF hydrogel network after pH adjustment and oxidation. (B) Conversion of DOPA to Dopaquinone under environmental conditions. (C) UV/VIS spectra of OPF-DOPA in pH adjusted PBS after 5 min and 24 h at 37 °C. (D) OPF-co-DMA dissolved in PBS_7.4 with increasing molar ratios of the oxidizer NaIO₄ with accompanying (E) UV/VIS spectra at 5 min.

Fig. 4.

(A) OPF-DOPA hydrogel discs made from PEG with different number average molecular weights (1k, 3k, or 10k) and either 0, 10, or 20 w/w% DOPA. Discs were soaked in PBS_pH3, PBS_pH7.4, or PBS_pH7.4 with 15 mM NaIO₄ for 24 h before testing. (C) Compressive modulus of OPF-DOPA hydrogel discs. Asterisks mark formulations that are statistically different than the control. Scale bar represents 1 cm.

The transition of OPF-DOPA copolymer solutions from a light yellow to dark tan color over 24 h is indicative of a process known as quinone tanning, wherein DOPA is oxidatively converted to DOPA quinone, the precursor to phenol coupling [91]. A demonstration of instantaneous DOPA quinone formation can be seen in Fig. 3D through the addition of the oxidizer sodium periodate (NaIO₄), resulting in a distinct peak at 425 nm after 5 min (Fig. 3E). It is worth noting that a distinct peak at 425 nm was not achieved through pH adjustment alone (data not shown). These results suggest that DOPA quinone formation occurs more slowly under environmental conditions, continuing over a 24 h period as was observed in the spectra of OFP-DOPA solutions adjusted to a pH 7 and 9 (Fig. 3C) and the yellow-orange tint of OPF-DOPA hydrogels soaked overnight in PBS_pH7.4 (Fig. 4A). However, the oxidation of DOPA functional groups within the OPF hydrogel network is undoubtedly incomplete over this timeframe, as can be seen by comparing the color of OPF-DOPA hydrogel discs swollen in PBS_pH7.4 overnight with those treated with 10 μM NaIO₄ overnight (Fig. 4A).

One advantage of this system is that the speed of 2° crosslink formation is tunable. Faster coordination between DOPA side chains is possible through the addition of metallic ions such as ferric iron (Fe³⁺). In mussel adhesive, the transition from an acidic to basic pH facilitates the formation of increasingly stable catechol-Fe coordination reactions within the distal region of mussel byssal threads [92]. This principle has inspired the design of numerous double crosslinked bioadhesives for wound closure [21,93], with the added benefit of providing the material with self-healing properties after load shearing [60].

3.3. Mechanical and physical properties of hydrogels

Despite being enzyme mediated [94], the formation of a fully crosslinked protein network in mussel adhesive can take days [50] due to the presence of antioxidants within the protein matrix [47] and an outer coating [95]. This same process of environmentally-induced stiffening was observed in the physical characteristics of OPF-DOPA hydrogels following crosslinking when swollen under physiological conditions. Compression testing of 8 mm OPF-DOPA hydrogel discs confirmed significant increases in the compressive modulus of the OPF_1k-DOPA₁₀, OPF_1k-DOPA₂₀, and OPF_3k-DOPA₁₀ formulations after 24 h in PBS_pH7.4 at 37 °C when compared with those held in PBS_pH3 (Table S3; Fig. 4B). A similar effect was also observed in samples oxidized with NaIO₄, indicating that phenol coupling likely results in hydrogel network stiffening. However, this effect was not observed in either of the OPF_10k formulations that incorporated DOPA (Fig. 4B). Instead, OPF_10k formulations had 33% fewer crosslinks per unit volume and more molecular weight between crosslinks (Table S4).

While DOPA functionalization impacted the mechanical properties of hydrogels, the main factor that determined hydrogel network structure was the molecular weight of the PEG starting material (PEG MW) used to synthesize OPF (summarized in Table S4, Fig. S7). The volumetric swelling ratio (Q, p = 0.018), mesh size (ξ, p = 0.003), and molecular weight between crosslinks (M_c, p = 0.002) all increased with PEG MW, while crosslink density (ρ_x, p = 0.026) decreased (Table S5). Across all formulations, mesh size (an indicator for porosity) was positively correlated with the molecular weight between crosslinks, while the volumetric swelling ratio was negatively correlated with crosslink density (Fig. S7). DOPA content did not significantly impact any of these parameters (Table S5), although a negative trend of reduced volumetric swelling and mesh size was observed in the OPF_1k and OPF_3k formulations that contained DOPA. Similarly, the crosslinking density within the OPF_1k formulations that contained DOPA were higher than the control, indicating that network stiffening after incubation in PBS_pH7.4 is likely due to DOPA crosslinking.

3.4. Degradation kinetics

The degradation of hydrogels under physiological conditions (PBS_7.4, T = 37 °C) was significantly impacted by both PEG MW and DOPA content (Table S6, Fig. 5). DOPA functionalization was negatively correlated with mass swelling ratio (p < 0.001), an effect that was consistent across formulations but was most pronounced within the OPF_10K formulations (summarized in Table 1; Fig. 5). A testament to the covalent nature of the phenol coupling reaction, this reduction was maintained even after incubation under physiological condition for up to 4 weeks, resulting in significantly slower degradation (reduction in % mass loss, p = 0.005) of DOPA containing hydrogels (Fig. 5).

Fig. 5.

Degradation of OPF-DOPA hydrogels made from PEG with different molecular weights (1k, 3k, or 10k) and either 0 (Ctrl), 10, or 20 w/w% DOPA. Hydrogels were placed in PBS (pH 7.4) for 4 weeks at 37 °C with daily solution exchange, over which the (A) mass swelling ratio, (B) mass loss (%), and (C) pH variation of the surrounding PBS solution (from pH 7.4) were measured over time.

DOPA also significantly impacted the pH of the surrounding solution (p < 0.001), resulting in a ~ 0.5 pH reduction regardless of DOPA concentration (Fig. S7). This result is likely the result of protons from catechol groups leaching out of hydrogels into the surrounding tissue during biodegradation. While acidic microenvironments are known to contribute to the development of tumors [96], it is worth noting that pH variability of the magnitude observed here did not impact cell viability (Fig. 9B), a result that is in agreement with other published work that demonstrates that polydopamine materials can support robust cell growth [97].

Fig. 9.

(A) Cell proliferation on OPF-DOPA hydrogels after 1, 4, and 7 d post-seeding. (B) Viability of cells exposed to the leaching solution of OPF-DOPA hydrogels. (C) Live/Dead-stained confocal images of MC3T3-E1 cells 4 d post-seeding on the surface of OPF-DOPA hydrogels made from PEG with different number average molecular weights (1k, 3k, or 10k), containing either 0 (ctrl), 10, or 20 w/w% DOPA. Asterisks mark timepoints that were statistically different (p < 0.05) than the control.

3.5. Adhesive properties of injectable OPF-DOPA hydrogels

While the spacing between DOPA residues in mussel adhesive is unknown, DOPA spacing likely modulates the adhesive strength of mussel adhesive by controlling post-curing crosslink density [98]. Taking this into account, we developed nine OPF-DOPA formulations with a range of mesh sizes (57–163 Å) and crosslink densities (0.177–0.711 × 10³ mol cm⁻³) in an effort to vary the number and spacing of DOPA sidechains available for adhesion (summarized in Table S4). The adhesive performance of each formulation was then investigated using two tests: lap-shear strength adhesion testing (Fig. 6) and push-out testing (Fig. 7). Lap-shear testing approximated the conditions present during bone fracture fusion [99], while push-out testing approximated the force required to dislodge implants from calvarial critical size defects [74]. For Lap-shear testing, hydrogels were first injected between two glass plates and dry-cured at room temperature for 1 h. Plates were then either pulled to failure immediately or incubated within physiological conditions (PBS_pH7.4) for 24 h and then tested. When cured under dry conditions all formulations performing equally well, with no observed effect of DOPA content nor PEG MW on adhesion strength (Table S7, Fig. 6C). However, after incubation in physiological conditions, DOPA modified hydrogels outperformed the controls for each PEG MW tested, irrespective of the amount w/w% DOPA added Fig. 6D).

Fig. 6.

Lap-shear adhesion testing of OPF-DOPA hydrogel formulations. (A) Diagram of experimental workflow including a representative stress-strain curve from mechanical tests in which hydrogels were (1) injected between two glass plates, either (2) allowed to air dry for 1 h and tested or (3) soaked in PBS_7.4 for 24 h and tested. (B) The tensile testing machine used to perform lap-shear adhesion testing. Maximum adhesion of OPF-DOPA hydrogel formulations after being cured for 1 h dry at RT (C) or 24 h in PBS_pH7.4 at 37 °C (D). Crosslinked OPF scaffold without DOPA modification was used as a control. Asterisks mark formulations that are statistically different (p < 0.05) than the control; Hashmarks indicate formulations that are statistically different (p < 0.05) at 1 h and 24 h.

Fig. 7.

Push-out testing of OPF-DOPA hydrogel formulations in bone defects. (A) Diagram of experimental workflow including a representative stress-strain curve. (B) The compression testing machine used to perform push-out tests. (C) Maximum push-out strength of hydrogel formulations after being cured for 24 h in PBS_pH7.4 at 37 °C. Crosslinked OPF scaffold without DOPA modification was used as a control. Asterisks mark formulations that are statistically different (p < 0.05) than the control.

Further analysis of OPF-DOPA hydrogel adhesion was performed using push-out testing. Critical-sized bone defects were simulated by drilling 10 mm holes in excised bovine femur cross-sections, which were subsequently filled with hydrogel formulations, dry-cured for 1 h, incubated in PBS_pH7.4 for 24 h at 37 °C, and then tested. For this test, hydrogels containing DOPA outperformed the control in all cases, with the OPF_1k-DOPA₁₀, OPF_3k-DOPA₁₀ and OPF_3k-DOPA₂₀ formulations requiring the most force to dislodge implants (Table S7, Fig. 7C). As was seen with wet lap-shear testing, the adhesion strength of the OPF control was greatly diminished in almost every case, rendering the unmodified OPF_10k treatment untestable due to hydrogels not remaining in the defect after incubation in PBS (Fig. 7C).

Taken together, the results of adhesion testing highlight two ways DOPA facilitates the adhesion of OPF implants to surfaces. First, while the initial interfacial reactions between DOPA and a surface are important, under dry conditions the 1° crosslinking reaction that generates the structure of the hydrogel is sufficient to adhere two glass plates, most likely due to hydrogen bonding between the surface and the polymers present in the hydrogel network. In this case, the advantage of the bidentate coordination between DOPA and silica was only evident after the hydrogel was incubated in PBS for 24 h and hydrogen bonding in the control treatments was interrupted by water absorption. Second, as was observed previously in degradation assays, the phenol coupling of DOPA under physiological redox and pH conditions greatly diminishes both the mass and volumetric swelling of hydrogels (Tables 1, S4, Fig. 5A), serving to prevent the migration of implants away from the substrate surface. This restricted movement can result from bidentate coordination and hydrogen bonding, as seen in lap-shear testing with glass, in addition to coordination reactions with metallic ions within composite materials [100,101]. In either case, the robust adhesion strength of OPF-DOPA hydrogels before and after exposure to physiological conditions suggests that DOPA functionalization is a promising material modification that can improve implant localization by maintaining the integrity of the implant-tissue interface after swelling. Similarly, synonymous adhesion strength using either 10 or 20 w/w% DOPA in most formulations, allowed adequate design flexibility so as to tune the stiffness of hydrogels in accordance with desired degradation kinetics – a significant advantage when designing bone scaffolds with different release profiles for targeted drug delivery [23,102].

3.6. Osteoblast attachment and proliferation

One major drawback of bone scaffolds comprised of synthetic polymers such as OPF is that they provide a smooth, neutrally charged surface that fails to support robust cell attachment or proliferation [82]. In contrast, due to the highly reactive nature of the hydroxy groups adorning the phenol ring of DOPA, polydopamine is a promising addition to bio-engineered surface coatings that aim to enhance cell proliferation or provide attachment sites for bioactive molecules [54,103,104]. For bone tissue engineering specifically, the addition of DOPA functional groups has also been shown to enhance hydroxyapatite crystal formation when exposed to simulated body fluid, a key factor thought to promote osteoblast attachment and proliferation [105]. In this study, DOPA functionalization significantly altered the surface topography of the OPF hydrogels when observed through scanning electron microscopy (SEM) (Fig. 8), introducing distinct ridges to the surface of the 10% w/w DOPA treatments (Fig. 8B). In the 20% w/w DOPA treatments, these ridges were deeper and more profound, forming dense ribbons with a spacing that was dependent on OPF MW (Fig. 8C).

Fig. 8.

Scanning electron micrograph (SEM) images (at two magnifications) of the surface of OPF-DOPA hydrogels made from PEG with different molecular weights (1k, 3k, or 10k) and crosslinked with either 0 (A), 10 (B), or 20 (C) w/w% DOPA.

To investigate the impact of DOPA-mediated adhesion and changes to the surface topography of hydrogels, the attachment and proliferation of MC3T3-E1 pre-osteoblast cells were monitored on 8 mm discs over 7 d. DOPA functionalization supported enhanced protein adsorption (p < 0.001, Fig. S10A), cell attachment (p < 0.001, Fig. S10B), and cell proliferation (p < 0.001, Fig. 9A), with a significant interaction between DOPA content and PEG MW (p < 0.001, summarized in Table S8). All OPF-DOPA formulations reporting higher total dsDNA than the unmodified OPF controls after 7 d, with the greatest proliferation reported on OPF_1kand OPF_3k discs (Fig. 9A). These results were in agreement with LIVE/DEAD staining assays and confocal imaging, wherein dense cultures of cells were seen growing along the surface ribbons of each formulation after 4 d in culture (Fig. 9C). OPF_10k formulations supported the least cell proliferation, a result that is consistent with other studies that employ OPF-based materials [33,106] and a likely consequence of the preference of osteoblasts for stiffer surfaces during settlement [107].

While striking, the magnitude of the effect that DOPA has on the surface architecture of OPF hydrogels provides a unique challenge when isolating the contribution of adhesive interactions to cell proliferation. For example, the surface roughness of titanium implants alone has been shown to affect the attachment, spreading, and proliferation of pre-osteoblasts, irrespective of surface chemistry [108,109]. To address this possibility, cell proliferation assays were repeated after the addition sodium periodate (NaIO₄), an oxidizer that serves to nullify the impact of DOPA-mediated adhesion by accelerating the crosslinking of DOPA functional groups. It should be noted that neither DOPA functionalization (Fig. 9B) nor NaIO₄ treatment at a concentration of 10 μM (Table S8, Fig. S8) alone were found to negatively impact cell proliferation. Cell proliferation on OPF-DOPA hydrogels was not significantly different when treated with NaIO₄ (p = 0.83, Table S9, Fig. S9). This result emphasizes the role of surface complexity over adhesive interactions as the motivating factor behind the enhanced cytocompatibility observed on OFP-DOPA hydrogels.

4. Conclusion

Here we reported the design and evaluation of an injectable bone scaffold that adheres to wet surfaces and undergoes material stiffening when exposed to the pH and redox conditions present within the body. Inspired by the environmentally triggered adhesive system of marine mussels, adhesive hydrogels were developed through the conjugation of crosslinkable catechol functional groups (DOPA) and the synthetic oligomer OPF and PEG-DA, varying the DOPA content (w/w%) and molecular weight (MW) of the OPF backbone to produce formulations with different swelling ratios, porosities, and crosslink densities. A double-crosslinked system, 1° crosslinking between OPF and PEG-DA was induced through the addition of the crosslinking initiator ammonium persulfate (APS) and the crosslinking accelerator l-ascorbic acid (AA). Stiffening of the network was subsequently achieved through the gradual oxidation and phenol coupling of DOPA quinone under physiological conditions, a process that reduced the swelling of implants and slowed their degradation. DOPA functionalization also enabled hydrogels to adhere to glass plates as assessed through lap-shear testing, as well increased the force required to dislodge implants from critical-size bone defects after 24 h under physiological conditions. Moreover, the inclusion of DOPA improved the cytocompatibility of OPF hydrogels, enhancing the attachment and proliferation of osteoblasts in vitro. In conclusion, OPF-DOPA hydrogels, when processed sequentially in a manner that mimics the natural adhesive system, are a promising injectable bone scaffolding material due to their tunable material, mechanical, and degradation properties, enhanced biocompatibility, and robust wet-adhesion capable of limiting post-implantation migration.