Bio-Inspired Copper Coordination Bonds Enable Tunable, Viscoelastic, and Cytocompatible Polyhydroxyalkanoates

Abstract

Biomaterials for repair and regeneration of soft tissues are ideally viscoelastic and processable. One way to obtain such materials is through dynamic cross-linking. Inspired by the high abundance of histidine residues in copper metalloproteins, we synthesized PHA copolymers containing imidazole side chains. The addition of a small amount of copper­(II) ions results in the formation of dynamically cross-linked materials whose properties are controlled by varying the copolymer composition and counterion identity. These materials are cytocompatible with human umbilical vein endothelial cells and exhibit promising degradation timelines. This work will inform the future use of dynamically cross-linked polyhydroxyalkanoates in soft tissue applications and provide a platform for accessing the wide range of material properties necessary for many organs in the human body.

Introduction

Biomaterials are essential for a variety of biomedical applications, such as joint replacements and skin grafts. − To fit a wide range of uses, the ability to tune mechanical properties, such as tensile strength and Young’s modulus, is critical to meeting individual requirements. Focusing on applications in soft tissue healing and regeneration, it is advantageous to be able to tailor the moduli and toughness of the polymer for the intended area of the body. Furthermore, soft and viscoelastic materials that are processable under mild conditions are highly desirable, enabling scaffold fabrication for soft tissue engineering.

One way to achieve tunable material properties is by incorporating dynamic bonding motifs, such as hydrogen bonds, − metal–ligand bonds, − or ionic bonds, to enable reversible cross-linking and self-healing properties. Dynamic linkages in other polymer systems have demonstrated better cell-growth conditions and mechanical properties mimicking native tissues. ,, Coordination bonds are of especially high interest, as their bonding strength can be modulated through metal lability, ligand type, and counterion coordination capability. ,, Dynamic networks have been explored for olefinic polymers and polyethers, but are underexplored in biodegradable systems like polyesters.

Polyhydroxyalkanoates (PHAs), a nontoxic and bioderived class of polyesters, are promising candidates for biomedical applications. PHAs can be synthesized biologically from carbon feedstocks in engineered E. coli, but the resultant polymers are too crystalline, making them unsuitable for soft tissue applications. These obstacles have been partially overcome by blending with other polyesters, such as poly­(lactic acid) and poly­(caprolactone), to break up crystallinity, enhance processability, and tune the rate of degradation. − Unfortunately, blending can result in phase separation and negative impacts for in vivo degradation, making copolymerization a more attractive strategy. Additionally, we sought to synthesize atactic PHAs due to their degradation timelines. Atactic PHAs degrade more quickly via chemical and enzymatic hydrolysis than the isotactic or syndiotactic forms. ,

The chemical synthesis of PHAs has been well established as an alternative method to produce this class of polyesters. PHAs can be synthesized in a chain growth fashion via the ring-opening polymerization (ROP) of 4-membered and 8-membered ring lactones. − We have focused on the synthesis of PHAs through the ROP of β-lactones due to their facile and scalable synthesis through the catalytic carbonylation of epoxides. − By employing a number of commercially available epoxides, PHAs with a wide range of functionalities have been accessed. However, the use of dynamic cross-linking to access PHA-based biomaterials has been underexplored in the field. The only example of a dynamically cross-linked PHA network was reported by Chen, Beckham, and co-workers, who used the biological synthesis of PHAs to access boronic ester-based vitrimers. These materials exhibit Young’s moduli of 1.7–2.5 MPa and elongation at break values up to 300%, but their processability is time-intensive due to long crystallization kinetics that may be detrimental to soft tissue engineering applications.

Moreover, we have reported a Schiff base containing polyester cross-linked with multiple metals, including Fe³⁺ and Cu²⁺. This material with a salicylaldimine ligand demonstrated elastomer-like behavior with limited hysteresis during cyclic testing to 20% strain and varied mechanical properties when changing the metal:ligand ratio. However, the material is difficult to tune due to kinetics of step-growth polymerization. For this reason, we targeted a tunable, dynamically cross-linked PHA material that is also melt-processable, soft, and viscoelastic for soft tissue applications.

Inspired by copper-containing metalloenzymes such as copper–zinc superoxide dismutase − and previous work by Harrington and Holten-Anderson, we chose to use a histamine-derived system, where metal–imidazole interactions are stronger than the previous salicylaldimine system. ,,, We hypothesized that the mechanical properties of the PHA dynamic network created by the imidazole–metal bond can be tuned via the metal ion and counterion. , Herein, we report a novel synthetic strategy to access dynamic metal-cross-linked PHAs with tunable properties via the ring-opening copolymerization of β-lactones and subsequent postpolymerization modifications (Figure ).

1.

Synthesis of dynamic metal-cross-linked PHA copolymers via ring-opening copolymerization, histamine postpolymerization modification via thiol–ene, and copper cross-linking to access new cytocompatible biomaterials.

Experimental Section

Materials

CaH₂ was purchased from Acros Organics (10–100 mm pieces) and ground into a fine powder using a mortar and pestle before use. Carbon monoxide was purchased from Matheson (research quality, >99.99% min purity) and used as received. Allyl glycidyl ether was purchased from Millipore Sigma and was dried over CaH₂ for 3 days under a nitrogen atmosphere. It was vacuum transferred and degassed via 3 freeze–pump–thaw cycles and stored in the glovebox freezer over 3 Å sieves. Diethyl zinc, copper­(II) nitrate trihydrate, and copper­(II) chloride dihydrate were purchased from Millipore Sigma and used as received. Copper­(II) triflate was purchased from Ambeed and used as received. Copper­(II) acetate monohydrate was purchased from Fisher Scientific and used as received. Copper­(II) tetrafluoroborate tetrahydrate was purchased from Oakwood Chemicals and used as received. Acetylhistamine was purchased from Millipore Sigma and used as received. [(TPP)­Al­(THF)₂]­[Co­(CO)₄] (where TPP = tetraphenylporphyrin) and N-(2-(1H-imidazol-4-yl)­ethyl)-4-mercaptobutanamide (hist-SH) were synthesized as previously reported. ,

β-Butyrolactone (β-BL) was purchased from TCI and purified according to a modified procedure of Kurcok and co-workers. β-BL was predried by stirring over CaH₂ under a nitrogen atmosphere for 24 h. This was fractionally distilled (active vacuum, 50 °C, −78 °C cooling on the receiving flask), and the first fraction (5% of total volume) was collected and removed. The second fraction was then oxidized over KMnO₄ (ca. 1 mg/mL) until a color change from purple to brown occurred (ca. 10–15 min). This was then distilled under gentle heat with care taken not to evaporate to dryness, leaving about 5% volume in the sending flask (active vacuum, 50 °C, −196 °C cooling on the receiving flask). Finally, β-BL was dried by stirring over CaH₂ under a nitrogen atmosphere for 3 days before being vacuum transferred and degassed by three freeze–pump–thaw cycles. This was brought into the glovebox, placed over 3 Å sieves for 24 h before use, and stored in the freezer (−25 °C).

Endothelial cell medium supplemented with 5% fetal bovine serum, growth factors, ascorbic acid, and hydrocortisone was purchased from Promocell and stored at 4 °C. Phosphate-buffered saline was purchased from ThermoFisher and used as received.

Methods

Nuclear Magnetic Resonance (NMR) Spectroscopy

All NMR spectra were acquired at 25 °C. ¹H and ¹³C­{¹H} spectra were recorded on a Bruker Avance III HD spectrometer with a broadband Prodigy cryoprobe at 25 °C (operating at 500 MHz for ¹H and 126 MHz for ¹³C) or a Bruker AVIII HD spectrometer with a BBFO probe (operating at 400 MHz for ¹H). For ¹H and ¹³C spectra, chemical shift data are reported in units of δ (ppm) relative to that of tetramethylsilane (TMS). ¹H NMR spectra were referenced to residual solvent (7.26 ppm for CDCl₃, 7.16 ppm for C₆D₆, and 3.31 ppm for CD₃OD), and ¹³C NMR spectra were referenced to the deuterated solvent itself (77.16 ppm for CDCl₃). NMR spectra were processed in MestReNova version 14.3.2. NMR spectroscopic data are reported as follows: chemical shift, multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, hept = heptet, m = multiplet, b = broad), coupling constants (Hz), and integration.

Size-Exclusion Chromatography (SEC)

Size-exclusion chromatography (SEC) analyses were carried out using an Agilent 1260 Infinity SEC system equipped with an Agilent 1260 Infinity autosampler and a refractive index detector. The Agilent SEC system was equipped with two Agilent PolyPore columns (5-μm, 300 mm × 4.6 mm) and a guard column (50 mm × 4.6 mm), which were eluted with THF at 30 °C at 0.3 mL/min and calibrated using monodisperse polystyrene standards. The samples were filtered through a 0.45 μm PTFE syringe filter.

General Procedure for Allyl Glycidyl Ether Carbonylation

In a nitrogen glovebox, a 100 mL Parr high-pressure reactor was placed in the freezer (−20 °C). [(TPP)­Al­(THF)₂]­[Co­(CO)₄] was added to a 100 mL glass insert (0.390 g, 0.408 mmol) and dissolved in 22.0 mL of THF. This glass insert was placed in the bottom of the reactor in the freezer for 30 min to prevent rearrangement to the methyl ketone. After 30 min, allyl glycidyl ether (stored in the freezer) (12.0 g, 102 mmol) was added. The reactor was removed from the glovebox and charged with 800 psi CO, followed by stirring and heating to 50 °C. After 18 h, the reactor was placed in an ice bath for 30 min and slowly vented in a well-ventilated fume hood. THF was removed in vacuo, and the lactone was subsequently diluted in CH₂Cl₂ (50 mL) and passed through a silica plug to remove the residual catalyst. CH₂Cl₂ was removed in vacuo, and the lactone was dried over CaH₂ and activated carbon under a nitrogen atmosphere for 3 days. It was vacuum transferred (active vacuum, 80 °C, −196 °C cooling for the receiving flask) and degassed by three freeze–pump–thaw cycles. Activated carbon is used to react with residual Co₂(CO)₈ to prevent any transfer to the receiving flask. 4-Allyloxymethyl-β-propiolactone was stored at −25 °C in the glovebox freezer over 3 Å sieves for 24 h before use. After purification, the product was isolated in 98% yield as a slightly viscous oil. ¹H NMR (500 MHz, CDCl₃): δ 3.43 (m, 2H, CH₂), 3.73 (m, 2H, CH₂), 4.05 (m, 2H, CH₂), 4.64 (m, 1H, CH), 5.24 (m, 2H, allyl), 5.87 (m, 1H, allyl). ¹³C­{¹H} NMR (126 MHz, CDCl₃): δ 40.0, 70.0, 73.0, 118.1, 134.3, 168.0. NMR data is in agreement with previous reports.

General Procedure for Copolymerization of β-BL and ABL

In a nitrogen glovebox, [EtZnO ⁱ Pr]₄ (5.50 mg, 0.036 mmol) was added to a 20 mL septum-capped vial equipped with a stir bar. Racemic β-BL and ABL were added sequentially (β-BL and ABL = 18 mmol). Finally, toluene was added to the vial to dilute the solution to a final concentration of 6.00 M in lactone. The vial was capped and heated to 60 °C in an aluminum heating block inside a glovebox for 18 h. The polymerization was then removed from the heater block, brought out of the glovebox, and quenched with CH₂Cl₂. An aliquot was removed for ¹H NMR spectroscopy and SEC analysis. The quenched reaction mixture was precipitated into 40 mL of diethyl ether at −20 °C. The polymer was dissolved in CH₂Cl₂ and reprecipitated from cold Et₂O to ensure removal of the catalyst and then dried in vacuo at 40 °C.

General Procedure for Thiol–Ene Functionalization of ^enePHA Copolymers

In a nitrogen glovebox, solid copolymer (1.49 g, 0.460 mmol, 1.00 equiv alkene) was dissolved in 50 mL of CH₂Cl₂. Hist-SH (0.392 g, 1.84 mmol, 4.00 equiv) was dissolved in 50 mL of MeOH. 2,2-Dimethoxy-2-phenylacetophenone (DMPA) (58.9 mg, 0.230 mmol, 0.500 equiv) was dissolved in 2.00 mL of 50/50 v/v% MeOH/CH₂Cl₂. The polymer and thiol solutions were added to a Schlenk flask equipped with a stir bar, followed by the DMPA solution. The Schlenk flask was transferred to a Schlenk line and irradiated with 365 nm light for 3 h under a positive pressure of nitrogen. Upon completion, the crude reaction mixture was added to dialysis tubing (Spectra/Por 3 Dialysis Tubing 3.5 kDa MWCO). The tubing was placed in a 1 L beaker containing 40% MeOH, 40% CH₂Cl₂, and 20% ethyl acetate for 24 h. After 24 h, the solution was changed to 100% MeOH for an additional 48 h. The tubing contents were then reduced in vacuo, dissolved in a minimal amount of CH₂Cl₂, and precipitated into cold Et₂O.

General Procedure for Copper Cross-Linking of ImPHA

In a 100 mL bottle, dried polymer (1.1 g, 0.54 mmol imidazole, 1.0 equiv) was dissolved in 56 mL of 50/50 v/v% MeOH/CH₂Cl₂ (2 wt % polymer solution). Cu­(NO₃)₂·3H₂O (32 mg, 0.13 mmol, 0.25 equiv) was dissolved in 2.0 mL of MeOH and CH₂Cl₂. The Cu²⁺ solution was added dropwise to the vigorously stirring polymer solution, resulting in the immediate appearance of a soft, blue, insoluble material. The solution was slightly shaken and then poured into a Teflon dish. Evaporation was allowed at room temperature in a fume hood overnight to reveal a film.

Preparation of Films for Tensile and In Vitro Testing

The ImPHA-Cu²⁺ film (vide supra) was dried for 8 h in vacuo to remove residual solvent and then folded and pressed between two Teflon sheets on a Carver press (45 °C, 1 metric ton load) to form a film of uniform thickness. Due to the dynamic nature of the cross-links, folded or broken pieces were reprocessed to generate uniform films. The films were then cut into dog bone samples for tensile testing or squares for the appropriate in vitro experiments. For cytocompatibility experiments, the Teflon casting dishes and Teflon sheets for pressing were autoclaved prior to use to add an additional sterilization step in the preparation.

Tensile Testing

Tensile tests were performed on a Shimadzu AGS-X series instrument with pneumatic grips and a 500 N load cell. Tensile specimens were prepared following the ASTM 37-4 standard. Tensile elongation measurements were performed under ambient conditions with a crosshead velocity of 10 mm/min until break. Cyclic tests were performed under ambient conditions with a crosshead velocity of 30 mm/min up to 20% strain and 120 s hold at 20% strain.

Preparation of Material Extracts for Cell Viability

^2.5ImPHA-Cu²⁺, ^5.0ImPHA-Cu²⁺, and polycaprolactone films were prepared via solvent casting into autoclaved Teflon casting dishes. Films were dried for 12 h in vacuo before being pressed into uniform films using previously described conditions on autoclaved Teflon sheets. Films were cut into 6 mm² squares, soaked in ethanol for 45 min, and subsequently rinsed with sterile water. Films were dried at 0.1–0.3 mbar for 24 h.

Cell Viability Assay Using Material Extracts

Material extracts were prepared for cytocompatibility testing in accordance with ISO:10993-12. Endothelial cell medium supplemented with 5% fetal bovine serum, growth factors, ascorbic acid, and hydrocortisone (Promocell C-22121) was incubated at 37 °C for 24 h on its own or with ^2.5ImPHA-Cu²⁺ films (8 mm²/mL), ^5.0ImPHA-Cu²⁺ films (8 mm²/mL), polycaprolactone (8 mm²/mL), or latex (5 mm²/mL). With the prepared extracts, cell viability experiments were carried out in accordance with ISO:10993-5. Human umbilical vein endothelial cells (HUVECs) (passage 6) were seeded and cultured in 96-well TCPS plates. Upon reaching ∼80% confluency, the media was removed and replaced with incubated media or material extracts. After 1 day, four samples from each group were subjected to the CellTiter-Glo 2.0 Assay (Promega) in accordance with manufacturer instructions. This was repeated after 2 days and 3 days. Absorbance measurements were recorded on a microplate reader.

Simulated Degradation of Thin Film Samples

Thin film samples of ^2.5ImPHA-Cu²⁺ and ^5.0ImPHA-Cu²⁺ were prepared via a previously described method for cytocompatibility. Films (4 mm² and 0.3 mm thick) were added to 10 mL of 1 M NaOH in PBS (phosphate-buffered saline) and incubated at 37 °C with agitation. After 1, 2, 3, and 6 days, the samples were removed in triplicate, rinsed with distilled water, and dried in vacuo before recording the final mass.

Results and Discussion

To design a tunable, dynamically cross-linked PHA material that is also melt-processable, soft, and viscoelastic for soft tissue applications, we first needed to create a PHA with a handle for postpolymerization modification (PPM). We chose to copolymerize 4-allyloxymethyl-β-propiolactone (ABL) with commercially available β-BL to afford a polymer containing terminal alkenes suitable for PPM. An optimized synthesis of ABL was achieved by catalytic carbonylation of allyl glycidyl ether (AGE) using [(TPP)­Al­(THF)₂]­[Co­(CO)₄] (TPP = tetraphenylporphyrin) as the catalyst due to its high activity and ease of preparation. ,,, The reaction of AGE with CO in the presence of 0.2 mol % of [(TPP)­Al­(THF)₂]­[Co­(CO)₄] resulted in the formation of ABL on a 12 g scale in 98% yield (see SI Section 2.1 for further experimental details).

With both monomers in hand, we sought to identify a catalyst for the copolymerization of β-BL and ABL. While a variety of β-lactone polymerization catalysts have been developed, many of these catalysts require multistep syntheses or are impractical for biomaterial applications. ,− For this reason, we targeted catalysts that are inexpensive, readily prepared, biocompatible, nonstereoselective, and have similar activity toward β-BL and ABL. We selected ethyl zinc isopropoxide [EtZnO ⁱ Pr]₄, a catalyst for the controlled ROP of β-BL to form atactic PHB. Additionally, due to its straightforward synthesis from isopropanol and diethyl zinc, the absence of an ancillary ligand facilitates easy removal of catalyst residues from the resulting polymer, limiting cytotoxicity for final applications.

The copolymerization of β-BL and ABL by [EtZnO ⁱ Pr]₄ was investigated. The copolymerization is well-controlled, allowing us to tune the molecular weight and amount of alkene incorporation. Prior reports demonstrated moderate molecular weight control for the polymerization of β-BL by [EtZnO ⁱ Pr]₄, and we confirmed this at a fixed ratio in the copolymerization (Table S2). To investigate the resulting copolymer incorporation of β-BL and ABL, we fixed the overall monomer-to-catalyst ratio at 500:1 and varied the ABL feed ratio from 2.5 to 30 mol %. As the feed ratio of ABL increased, we observed a proportional increase in the degree of alkene incorporation in the resulting polymer (^2.5–30enePHA), while maintaining excellent molecular weight control (Table , entries 1–5). Overall, this catalyst allowed us to obtain random, atactic copolymers with a wide range of allyl group incorporation. To study the copolymerization, we performed Fineman–Ross copolymerization parameter studies. Through the analysis of copolymers of varied monomer feeds, we determined r ₁ for β-BL to be 0.85 and r ₂ for ABL to be 0.87. At lower compositions of β-BL, there is some preference for β-BL, but the copolymer library has β-BL compositions of at least 70%, and therefore, we synthesize random copolymers in all cases (see SI Section 3.3 for further details on the Fineman–Ross analysis).

1. Copolymerization of β-BL and ABL Using [EtZnO ⁱ Pr]₄ .

entry	polymer	β-BL:ABL	ABL incorp (%)	M n, theo (kDa)	M n, SEC (kDa)	Đ
1	2.5enePHA	97.5:2.5	2	43.7	53.7	1.2
2	5.0enePHA	95:5	5	44.4	43.5	1.2
3	10enePHA	90:10	10	45.8	37.1	1.3
4	15enePHA	85:15	17	47.2	43.1	1.3
5	30enePHA	70:30	30	51.4	53.8	1.3

^aPolymerizations were run for 18 h at 60 °C at 6 M monomer concentration in toluene. Both monomers achieve >99% conversion determined via ¹H NMR spectroscopy.

^bIncorporation of β-BL and ABL was determined via ¹H NMR spectroscopy of a precipitated polymer sample.

^cDetermined by SEC in THF at 30 °C, calibrated with monodisperse polystyrene standards.

With our PHA library in hand, we developed a route to append ligands suitable for reversible binding. We selected a histamine functionality due to its prevalence in metalloprotein systems such as copper zinc superoxide dismutase, laccase, and lytic polysaccharide monooxygenase. ,− Some of these active sites exhibit catalytic activity with reactive oxygen species (ROS) to convert them to water and oxygen. This activity would be attractive in biomedical implants to control ROS concentrations in and around the implantation site. Harrington and Holten-Anderson have investigated metal–imidazole bonds in a PEG-based system and their variable tunability. ,, We hypothesized that the mechanical properties of a PHA dynamic network created via imidazole–metal bonds could be tuned via counterion identity and cross-linking density.

By subjecting ^2.5–30enePHA to the light-initiated thiol–ene click reaction with a histamine-derived thiol (hist-SH), imidazole-functionalized PHAs (^2.5–30ImPHA) were synthesized (Figure ). ¹H NMR spectroscopic analysis confirmed quantitative conversion of the pendant alkenes to imidazole functionalities (Figures S30–S34). Due to the polarity of the histamine groups, ^2.5–30ImPHA were insoluble in THF, precluding the use of SEC to determine any effects of molecular weight during PPM. The resulting ^2.5–30ImPHA were soluble in a 1:1 methanol and dichloromethane mixture, indicating no cross-linking occurred during PPM. We performed a control experiment with benzyl mercaptan to ensure no side reactions were occurring during the PPM. Full conversion to the thioether is observed with the SEC trace remaining narrow (Table S4 and Figure S29).

2.

Synthetic scheme of thiol–ene postpolymerization modification with hist-SH and subsequent copper cross-linking to access ImPHA-Cu²⁺ library. Ratio of 4 imidazole ligands to 1 Cu center is the optimized geometry and drawn here, but due to the diffusion-limited polymer system, it may vary.

The coordination chemistry of Cu­(II) histamine complexes was investigated via model studies using acetylhistamine as a small-molecule model. The reaction between acetylhistamine (AH) and CuX₂ yielded crystalline complexes. X-ray structural determination of the product of CuCl₂ and AH revealed that [Cu­(AH)₄]­Cl₂ had formed (Figure ). The complex has a distorted square planar geometry with 2 outer sphere Cl anions. This structure indicated the optimal stoichiometry for cross-linking the polymeric systems. It must, however, be noted that in the diffusion-limited polymer system, it is unlikely that four imidazole ligands are on a single copper center, and there are most likely many structures possible. The structure was additionally compared using UV–vis spectroscopy, where the λ_max of Cu­(AH)₄(NO₃)₂ matched that of a thin film of ¹⁰ImPHA-Cu²⁺ (Figure S55).

3.

X-ray structure determination of model compound. [Cu­(acetylhistamine)₄]­Cl₂ X-ray crystallographic structure. Hydrogen atoms, except those bound to nitrogen, and chloride anions are omitted for clarity.

The ImPHA library was cross-linked through dropwise addition of the Cu­(II) salt solution (0.25 equiv relative to mol % imidazole) to a stirred solution of ^2.5–30ImPHA. Successful cross-linking was confirmed by the formation of a blue, insoluble gel composed of Cu-cross-linked ImPHA (ImPHA-Cu²⁺). To further verify cross-linking, the gel fraction was determined to be 99% for all materials (see page S32 for additional experimental details). The reversibility of the Cu–imidazole bond was verified by adding a ligand that binds more strongly to Cu­(II) than imidazole. Acetylacetone (2 equiv per mole of Cu in film) was added to a cross-linked film in chloroform. After 1 h, the film was dissolved, confirming the lability of the Cu–imidazole bond (Figure S56). This result demonstrates the suitability of these materials for low-temperature melt pressing.

The mechanical properties of the library of ImPHA-Cu²⁺ polymers were then studied by using uniaxial and cyclic tensile testing of specimens made from solvent casting in Teflon dishes, followed by melt pressing. We first investigated how the coordinating strength of the Cu­(II) counterion affects the relaxation dynamics, modulus, and stretchability of the material. To evaluate different counterions, ¹⁵ImPHA-Cu²⁺ was used as the model polymer to maximize deviations in Young’s modulus (E). The ions selected exhibited the following trend in coordination affinity to Cu²⁺: acetylacetone > acetate > nitrate ≈ chloride > triflate > tetrafluoroborate. This trend can be rationalized as follows: during stretching, the imidazole(s) dissociate from Cu²⁺. A more coordinating counterion, such as nitrate, will lead to slower ligand exchange, leading to a higher modulus and less stretchability. Generally, decreased coordination affinity led to materials with a smaller E and increased elongation at break (ε_B) (Figure a). The nitrate and chloride counterions have similar coordination affinities, and the resulting networks exhibit similar Young’s moduli (ca. 0.034 MPa) and ε_B values (ca. 500%) (Figure a). With triflate, a further drop in E is observed to 0.011 MPa and an ε_B of 900%. Furthermore, for the least coordinating tetrafluoroborate counterion, a further decrease in E to 0.010 MPa is observed, and the elongation at break increases to 2000%. The acetate counterion, a more coordinating counterion of bidentate or monodentate behavior, results in a lower E of 0.014 MPa and a higher ε_B of 1500 ± 600%. Acetate ions more readily enable diffusion through the material, leading to a lower E and higher ductility than those of nitrate or chloride counterions with similar coordination affinity. Its competitive nature with the imidazole ligands leads to rapid association and dissociation, and thus the observed mechanical properties. Cu­(acac)₂ further demonstrates this effect. The bidentate anion, acetylacetone, is a stronger ligand than imidazole, resulting in a low cross-linking density. This is consistent with our observation of an extremely soft, yet stretchable material with an E of 0.0023 MPa and an elongation of 4000% without breaking.

4.

Mechanical property characterization data for copper-cross-linked PHA materials. (a) Representative E and ε_B results of varying Cu²⁺ counterions with ¹⁵ImPHA. (b) Representative tensile test curves of varied ^xImPHA-Cu²⁺. (c) Cyclic testing of ^2.5ImPHA-Cu ²⁺ and ^5.0ImPHA-Cu²⁺.

We next investigated the impact of the imidazole loading on mechanical properties. We selected the nitrate counterion due to the Young’s modulus closely resembling that of soft tissue (Figure a) and solubility in cross-linking conditions. ^2.5ImPHA-Cu²⁺ is highly stretchable and soft with an ε_B of 2000 ± 250% and a Young’s modulus (E) of 0.012 ± 0.008 MPa (Figure b). Similarly, ^5.0ImPHA-Cu²⁺ is a soft material with a slightly lower ε_B of 1000 ± 150% and a Young’s modulus (E) of 0.013 ± 0.001 MPa. These E values closely resemble those of human skin and other soft tissues, which possess Young’s moduli varying from 0.01 to 0.5 MPa. Increasing the imidazole content to 10 and 15% results in slightly larger E values of 0.038 and 0.034 MPa, respectively, while decreasing the ε_B values to 550% and 500%. Finally, ³⁰ImPHA-Cu²⁺ is a stiffer and more brittle material with an E of 0.159 MPa and an ε_B of 200% (Figure b). The uniaxial tensile testing results demonstrate that increasing imidazole leads to higher E and reduced ε_B (Figure S48 provides a magnified view up to 30% strain). As the imidazole ligand content increases, the materials become less stretchable due to the copper ions easily reforming bonds with new ligands. At lower imidazole loadings, there is a reduced frequency of Cu–imidazole (Cu–Im) bond breakage and reformation, leading to high ε_B. Overall, these data demonstrate that this platform enables the synthesis of biocompatible materials with a wide range of mechanical properties.

Cyclic testing was performed on this library of materials to evaluate the elasticity and hysteresis (Figure c). The ^2.5ImPHA-Cu²⁺ and ^5.0ImPHA-Cu²⁺ materials exhibit the least hysteresis after 10 cycles due to faster relaxation and rebinding of the Cu–Im bonds (Figure c). As imidazole content increased, the residual strain during cyclic testing was greater due to slower relaxation dynamics (see data in SI Section 5.2). These cyclic testing results demonstrate that increased imidazole content increases the viscous phase and attenuates the elastic phase of the material.

The mechanical testing of ^2.5ImPHA-Cu²⁺ and ^5.0ImPHA-Cu²⁺ materials demonstrated that they possess properties necessary for soft tissue applications. To evaluate the simulated in vitro degradation of these polymer materials, films were subjected to 1 M NaOH in phosphate-buffered saline (PBS) at 37 °C, and the mass% loss was quantified. After 1 day, both materials degraded between 3 and 8 mass%. After 2 days, a more significant mass change was observed with ^2.5ImPHA-Cu²⁺ decreasing by ∼20 mass% and ^5.0ImPHA-Cu²⁺ decreasing by ∼17 mass%. A more substantial reduction in mass occurred after 3 days for both samples of approximately 30 mass% (Figure a). Ultimately, after 6 days, ^2.5ImPHA-Cu²⁺ films lost ∼75% of their original mass, and ^5.0ImPHA-Cu²⁺ films lost 90% of their original mass. Analysis of the films over this chemical hydrolysis period indicated bulk degradation through the uniform shrinking of the films and a loss of sharp edges. Due to the atactic nature of the prepolymers, fast degradation was observed under these conditions. , This basic condition is used to simulate the long-term hydrolysis in a shorter time period. , NaOH-induced degradation allows us to investigate bulk degradation of the cross-linked materials, but it does have some limitations, as we cannot model cell-mediated or enzyme-mediated degradation pathways. The degradation timeline of these materials will be further investigated in vivo using a subcutaneous model.

5.

In vitro degradation and cytocompatibility. (a) Simulated degradation in vitro with incubation in 1 M NaOH in PBS at 37 °C. Mass loss at 1, 2, 3, and 6 days. (b) Cytocompatibility of ^2.5–5ImPHAs-Cu²⁺ with HUVECs. Statistical analysis was performed using ANOVA against the media control, where * is P ≤ 0.05, ** is P ≤ 0.01, *** is P ≤ 0.001, and **** is P ≤ 0.0001.

To evaluate the cytotoxicity of these materials, media extract experiments were performed with human umbilical vein endothelial cells (HUVECs). Films of ^2.5ImPHA-Cu²⁺ and ^5.0ImPHA-Cu²⁺ were placed in a HUVEC medium at 37 °C to prepare the extracts. HUVECs (Passage 6) were plated in 96-well plates, and at ∼80% confluency, the media were replaced with the sample extracts. To measure the ATP cellular activity, the CellTiter-Glo 2.0 assay was performed after 1, 2, and 3 days (Figure b). As compared to a media control, ^5.0ImPHA-Cu²⁺ exhibits ∼80% viability after 3 days, and ^2.5ImPHA-Cu²⁺ demonstrates slightly less viability of ∼70%. These materials perform similarly to polycaprolactone films prepared in the same manner and do not exhibit cytotoxicity like latex. We hypothesize that a small concentration of Cu²⁺ is leaching into the media extracts due to the lability of the coordination bonds, leading to a small decrease in cytocompatibility.

Conclusion

We have developed a new strategy for dynamically cross-linked polyhydroxyalkanoate polymers using imidazole ligands and copper­(II) ions. Using the industrially relevant and accessible catalyst, [EtZnO ⁱ Pr]₄, we have synthesized copolymers of β-butyrolactone and 4-allyloxymethyl-β-propiolactone with random monomer incorporation and controlled molecular weights. The resulting pendant-alkene-functionalized copolymers were subject to thiol–ene click chemistry with a histamine-derived thiol. The addition of copper ions cross-links the imidazole-containing polymer, producing a library of soft and stretchable ^xImPHA-Cu²⁺ polymers. By tuning the imidazole content and counterions of the Cu²⁺, a range of Young’s moduli from 0.012 to 0.160 MPa and strain from 200 to 2000% have been achieved. Finally, ^2.5ImPHA-Cu²⁺ and ^5.0ImPHA-Cu²⁺, with soft-tissue-like mechanical properties, exhibited cytocompatibility toward HUVECs and degradation via hydrolysis. This platform enables tailoring of mechanical properties for soft tissue repair and can be extended to other metal ions to further modulate Young’s moduli and strain. Future work will include in vivo studies on implant performance, soft tissue repair, and regeneration.

Glossary

Abbreviations

PHA

polyhydroxyalkanoate

ROP

ring-opening polymerization

PPM

postpolymerization modification

[EtZnO ⁱ Pr]₄

ethyl zinc isopropoxide

AGE

allyl glycidyl ether

AH

acetylhistamine

β-BL

β-butyrolactone

ABL

4-allyloxymethyl-β-propiolactone

SEC

size-exclusion chromatography

HUVEC

human umbilical vein endothelial cell

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acspolymersau.6c00022.

• Experimental procedures, SEC, DSC, tensile testing results, and other supporting data (PDF)

The manuscript was written through the contributions of all authors. All authors have given approval to the final version of the manuscript.

The authors declare no competing financial interest.

Published as part of ACS Polymers Au special issue, “POLY 75th Anniversary”.