Rational design of poly(peptide-ester) block copolymers for enzyme-specific surface resorption

Abstract

Tissue resorption and remodeling are pivotal steps in successful healing and regeneration, and it is important to design biomaterials that are responsive to regenerative processes in native tissue. The cell types responsible for remodeling, such as macrophages in the soft tissue wound environment and osteoclasts in the bone environment, utilize a class of enzymes called proteases to degrade the organic matrix. Many hydrophobic thermoplastics used in tissue regeneration are designed to degrade and resorb passively through hydrolytic mechanisms, leaving the potential of proteolytic-guided degradation underutilized. Here, we report the design and synthesis of a tyrosol-derived peptide-polyester block copolymer where protease-mediated resorption is tuned through changing the chemistry of the base polymer backbone and protease specificity is imparted through incorporation of specific peptide sequences. Quartz crystal microbalance was used to quantify polymer surface resorption upon exposure to various enzymes. Aqueous solubility of the diacids and the thermal properties of the resulting polymer had a significant effect on enzyme-mediated polymer resorption. While peptide incorporation at 2 mol% had little effect on the final thermal and physical properties of the block copolymers, its incorporation improved polymer resorption significantly in a peptide sequence- and protease-specific manner. To our knowledge, this is the first example of a peptide-incorporated linear thermoplastic with protease-specific sensitivity reported in the literature. The product is a modular system for engineering specificity in how polyesters can resorb under physiological conditions, thus providing a potential framework for improving vascularization and integration of biomaterials used in tissue engineering.

Introduction

Surface erosion of thermoplastic biodegradable polymers remains a highly desirable phenomenon in biomaterials design.^1–3 The success of biodegradable scaffolds in regenerative medicine requires precise control over their degradation and remodeling kinetics.⁴ In tissue engineering applications, surface erosion preserves the mechanical properties of the polymeric device for extended periods of time during tissue regeneration, which is especially critical in load-bearing areas of the body. Of the polymers that show sensitivity to water-mediated chain scission, only those with high susceptibility to hydrolytic degradation, such as polyanhydrides^5–7 and polyorthoesters,⁸ have demonstrated surface eroding tendencies. Many biodegradable polymers used in regenerative medicine, such as polyesters, undergo bulk erosion,^9–11 which leads to a loss of mechanical strength and a burst release of degradation products at the end of the device’s lifetime.

Tissue regeneration requires the coordination of many cell types, and a material whose properties can respond to cell-specific remodeling and recapitulate this complex environment would improve integration and promote healing in the native tissue.¹² In the body, the catalytic activity of enzymes contributes to tissue remodeling, and therefore, enzymes can also contribute to polymer degradation. Leveraging enzyme activity can provide precise control over how a biomaterial is remodeled by taking advantage of (1) the differential expression of enzymes by different cell types and (2) enzyme substrate specificity. Some enzymes, such as collagenases, degrade amide bonds of specific proteins (collagen). Others degrade specific bond types – lipases, for example, degrade ester bonds. In the bone environment, osteoclasts secrete cathepsin K to degrade type I collagen, providing a cell-specific, enzyme-driven pathway for resorption. Recent efforts have focused on matrix metalloproteinases (MMPs), which are enzymes that degrade proteins of the extracellular matrix with distinct peptide specificity.¹³ The use of MMP-specific peptide sequences to control spatial and temporal degradation^14–17 is most commonly used in hydrogels due to the water-swollen network increasing accessibility of the peptide sequence to water-soluble enzymes.^18–21 Recent studies have demonstrated incorporation of bioactive peptide into nylons,²² however, the use of enzyme-sensitive peptides to control resorption of high molecular weight thermoplastics remains unreported. Zorn and colleagues polymerized oligopeptide segments into polyurethanes²³ for the purpose of improving polymer biodegradation. Their work fully characterized the surface of these peptide-incorporated block copolymers, demonstrating that the surface is enriched with the most hydrophobic of the monomer units. Our goal was to expand upon this concept and utilize peptide sequence to give even greater control over biodegradation and resorption of polyesters.

Advantages of linear thermoplastic polymers over their crosslinked counterparts include the ability to process the material into devices through solvent-based and thermal mechanisms without additional post-processing steps. The mechanical strength of high molecular weight thermoplastics is another advantage over other biomaterials, such as hydrogels or naturally derived matrices, in orthopedic applications. Furthermore, their hydrophobicity reduces the rate of hydrolytic degradation to more closely match the timeline of bone healing.^24,25 It has been previously reported that the glass transition temperature of a polymer is a strong indicator of its ability to undergo surface erosion – those that are amorphous under physiological conditions have the ability to surface erode.²⁶ Surface erosion under physiological conditions has also been achieved through very precise control over chemical structure and crystallinity.^26–29 Thus, the chemical, thermal, and physical properties of the polymer must be taken into consideration when designing surface eroding polymers.^29,30

We have previously described a library of tyrosol-based poly(ester-arylates), of which the thermal properties, mechanical properties, and degradation kinetics can be tuned based on the chemical structure of the polymer.³¹ These polymers are composed of diphenol monomers derived from tyrosol, a natural antioxidant found in olive oil.³² They contain fewer acidic degradation products than other commonly used polyesters such as poly(lactic acid) and poly(glycolic acid), thus potentially reducing the likelihood of cytotoxic tissue acidosis during late stages of the device’s lifetime.³³ These tyrosol-based polymers have been used for 3D-printing scaffolds for bone regeneration³⁴ and as a vehicle for delivery of hydrophobic drugs.³⁵ To expand upon the utility of these polymers in bone healing applications, the bioactivity of a subset of the tyrosol-based polymer library was enhanced to control enzymatic resorption. In this study, we describe the design of a linear thermoplastic polymer that incorporates an enzymatically sensitive peptide sequence to achieve a mode of enzyme-specific material erosion (Fig. 1). To the best of our knowledge, no other study has linearly incorporated protease-sensitive peptides into high molecular weight biodegradable thermoplastics for the purpose of controlling their resorption. Incorporation of peptide sequences derived from a fragment of type I collagen was accomplished through a pre-polymer block synthesis followed by chain elongation with the peptide via Michael addition. Quartz crystal microbalance with dissipation (QCM-D) was used to evaluate the effect of polymer properties and peptide sequence on polymer resorption by proteinase K, a model enzyme whose cleavage preferences are similar to that of cathepsin K.^36,37 The polymer films were also challenged by other physiological enzymes to validate enzyme-specificity. RAW264.7 macrophages and RAW264.7-derived osteoclasts were cultured on the peptide–polymers surface, thus demonstrating its ability to support attachment and proliferation of cells involved in matrix remodeling. Together, these results demonstrate the utility of peptide sequences to impart an enzyme-specific mode of resorption to thermoplastics, which is a key step towards using the differential expression of enzymes by different cell types to achieve cell-specific biomaterial resorption.

Fig. 1.

Schematic diagram of study design of poly(peptide-ester) block copolymers for enzyme-specific resorption. A thiol-capped protease-sensitive peptide was linearly incorporated into a maleimide-capped pre-polymer block via Michael addition (left). Polymers were evaluated for protease-mediated resorption using quartz crystal microbalance with dissipation (right).

Experimental

Materials

Amino acids and O-(1H-6-chlorobenzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HCTU) were purchased from Aapptec. Proteinase K, chymotrypsin, lipase, collagenase type I, and bovine serum albumin (BSA) were purchased from Sigma Aldrich. All other chemicals used for synthesis were purchased from TCI America or Sigma Aldrich unless otherwise noted. All solvents used were reagent grade.

Cell culture media (basal media, Dulbecco’s phosphate buffered saline, trypsin-EDTA, water) were purchased from Gibco/Life Technologies. Fetal bovine serum (FBS) was purchased from Gibco/Life Technologies, and the same lot was used for all experiments. Gentamicin and dimethylsulfoxide (DMSO, cell culture grade) were purchased from Sigma Aldrich.

Tyrosol-derived monomers were synthesized via Fisher esterification as previously described.^31,34 Briefly, tyrosol (1.01 eq., 0.33 M final concentration) was added to 2-(4-hydroxy-phenyl)acetic acid (1.0 eq.) in toluene with 85 wt% phosphoric acid in water (0.10 eq.). The product formed is denoted as HTy. All diacids were purchased from Sigma Aldrich or TCI America.

Trityl-protected mercaptoproprionic acid was synthesized by adding a 1.1-fold excess of trityl chloride dissolved in dichloro-methane (DCM, 6% w/v) (Sigma Aldrich) to 1 equivalent of 3-mercaptoproprionic acid dissolved in DCM (12% w/v) dropwise over 1 hour. The reaction was stirred for another 12 hours. The white solid was filtered and washed with diethyl ether (DEE) twice, then dried under vacuum overnight at ambient temperature.

Polymer characterization

Chemical composition was determined by proton nuclear magnetic resonance (¹H NMR, Varian 500 MHz) using deuterated chloroform (CDCl₃) or deuterated dimethylsulfoxide (DMSO-d₆) as a solvent. Absolute molecular weight of polymer blocks was quantified using end group analysis by ¹NMR (CDCl₃).

Number average molecular weight (M_n) and weight average molecular weight (M_w) were measured using gel permeation chromatography (GPC). A Waters 717 system equipped with two Agilent PLGel 5 mm columns with pore sizes of 105 Å and 103 Å was used with a refractive index detector (Waters 410). Samples were run using either chloroform or dimethylformamide (DMF) (+0.1% trifluoroacetic acid, TFA) as an eluent. All measurements were calibrated to a polystyrene standard curve of molecular weights ranging from 580–920 000 Da. Analytical high performance liquid chromatography (HPLC) was performed on an Agilent HPLC equipped with a dual UV absorbance detector at 220 nm and 280 nm wavelengths and a C18 column (4.6 × 250 mm, 5 μm pore size). A 15 minute linear gradient method from 5%/95% acetonitrile (+0.1% TFA)/water (+0.1% TFA) to 95%/5% acetonitrile (+0.1% TFA)/water (+0.1% TFA) was used to achieve separation.

FTIR spectra were collected on a ThermoFisher Scientific Nicolet iS10 spectrometer equipped with a monolithic diamond attenuated total reflectance (ATR) crystal between 500 and 4000 cm⁻¹ at 4 cm⁻¹ resolution. A sample of dried polymer was placed to cover the aperture, and 32 scans were averaged across the spectral range to improve the signal-to-noise ratio.

Differential scanning calorimetry (DSC) was used to measure glass transition temperature and melting temperature (T_m) using a 5-step method. Briefly, samples were (1) heated from −50 to 250 °C at a ramp rate of 10 °C min⁻¹, (2) held at 250 °C for two minutes, (3) cooled to −50 °C at a ramp rate of 10 °C min⁻¹, (4) held at −50 °C for two minutes, then (5) reheated to 250 °C at a ramp rate of 10 °C min ⁻¹. T_g and T_m were recorded from the reheat step.

Thermogravimetric analysis (TGA) was used to measure volatiles and the decomposition temperature (T_d) of the polymers. Volatiles were measured by the step transition from 50 °C to 150 °C, and the T_d was calculated at the plateau region before and after mass loss.

Polymer crystallinity and crystallite size were calculated from XRD data, which was collected on a Philips Xpert diffractometer in the parafocus geometry with Cu Kα radiation. Polymer films were prepared by solvent casting. Solutions of polymer were prepared at a concentration of 40 mg mL⁻¹ in DCM and were equilibrated under orbital shaking for 16 hours. Solutions were cast into Teflon dishes and allowed to dry under a saturated DCM environment for 8 hours before being placed in a vacuum oven overnight to remove residual solvent. Films were fitted onto zero-background holders (single crystal of silicon; (100) plane cut 9 degrees towards (001)). The following parameters were used for each scan: 10–40° 2θ range, 8.0 s per 0.05° step size, no-spin. Pearson VII functions (exponential factor = 5, Lorentzian = 0, and skew = 0) were used to profile fit each spectra using JADE software (Materials Data Inc., CA). Calculations for determining polymer percent crystallinity and crystallite size are described in detail in the ESI.†

Poly(ester-arylate) synthesis

Polymerization of tyrosol-derived diols with diacid was performed using carbodiimide chemistry (Scheme 1). For each polymerization, the desired molecular weight was achieved by optimizing the molar ratio of diol to diacid according to the Carother’s equation. Polymerizations were performed in DCM. Diisopropylcarbodiimide (DIC) and dimethylaminopyridinium p-toluenesulfonate (DPTs) were used at a 2.7 and 0.33 eq. to the diacid, respectively. Polymers were precipitated in a 10-fold excess of isopropyl alcohol (IPA) three times to remove excess monomer and catalyst. Solid was washed briefly in deionized (DI) water, frozen, and lyophilized.

Scheme 1.

A three-step synthesis involving (1) the polymerization of a tyrosol-derived diphenol with a diacid using carbodiimide chemistry, (2) capping the diol-capped pre-polymer with maleimide groups using carbodiimide chemistry, and (3) chain extension of bismaleimide prepolymer with a dithiol peptide via Michael addition. Y is defined for each polymer in Table 1.

Polymer blocks were capped with 6-maleimidohexanoic acid (6-MHA) using the same carbodiimide chemistry used for polymerization (Scheme 1). Polymer blocks, DPTs, and 6-MHA were dissolved in anhydrous DCM (10% w/v). DIC (2.7 eq.) was added dropwise, and the reaction was allowed to proceed for 3 hours. Reaction was performed under argon and protected from light. Reaction was precipitated into 10-fold excess of IPA three times to remove excess catalyst and 6-MHA. Solid was washed with DI water, filtered, frozen, and lyophilized.

All peptide-free poly(ester-arylates) were synthesized using a diol : diacid ratio of 1 : 0.98–0.99 with a final molecular weight of 150–200 kDa as previously published.³¹ For peptide-containing polymers, pre-polymer blocks were synthesized as described below prior to further chain elongation with a bifunctional peptide.

Bifunctional peptide synthesis

A degradable peptide was designed with two cleavage sites for proteinase K and cathepsin K. With regards to general specificity, it has been shown that cathepsin K, like other cysteine proteases, has high preference for a hydrophobic residue in the P2 position, and less specificity in other positions. The cathepsin K recognition site was identified as G-P-X-G, where X represents serine, methionine, arginine, glutamine, or tryptophan.^20,37 It has also been demonstrated that proteinase K cleaves preferentially on the C-terminal side of aromatic and aliphatic amino acids.³⁶ A degradable peptide with the sequence (mercaptoproprionate-GGPMGPWGGC) was designed with two cleavage sites for both cathepsin K and proteinase K (after the methionine and tryptophan) with the addition of glycine spacers to improve flexibility of the peptide monomer and accessibility of the cleavage sites. A non-degradable control peptide (Pep(Ctrl)) with the sequence (mercaptoproprionate-GGPGGPGGGC) was designed to remove the enzyme recognition sites, thus rendering it insusceptible to degradation. The thiol groups on either termini were used for polymer chain elongation with the bismaleimide pre-polymer. Peptide was synthesized using solid phase peptide synthesis and purified using flash chromatography (Full methods in ESI†). Mass was confirmed using a Finnegan LCQ Duo mass spectrometer. All peptides were stored under argon at −20 °C until use.

Chain extension of bifunctional poly(ester-arylate) with bifunctional peptide

The final polymerization was performed under anhydrous conditions and protected from light. For further polymerization with peptide, the bismaleimide pre-polymer blocks were dissolved in anhydrous DCM at a concentration of 20% w/v. One molar equivalent of peptide was dissolved in an equivalent volume of anhydrous DMF containing one equivalent of triethylamine (TEA) and 0.1 eq. of lithium bromide. The peptide solution was added to the pre-polymer solution dropwise (Scheme 1). Molecular weight growth was monitored by GPC with chloroform (+0.1% TFA) as an eluent. All polymers had weight average molecular weights greater than 100 kDa. The reaction was precipitated twice in a 10-fold excess of isopropyl alcohol (IPA), washed once with DI water, then frozen and lyophilized. Peptide-incorporated polymers synthesized with the pre-polymer diol : diacid ratio of 1 : 0.95 and 1 : 0.8 are denoted HTy(diacid)2%Pep and HTy(diacid)8%Pep, respectively, to denote the mole percentage of peptide relative to the moles of monomer (HTy).

Peptide degradation kinetics

Degradation of CSTK-degradable peptide was confirmed by liquid chromatography-mass spectrometry (LCMS) on a Thermo Fisher Vanquish UPLC and LTQ-XL mass spectrometer. A Pron-toSIL C18 AQ, 120 Å, 3 μm, 2.0 × 50 mm HPLC Column was used for purification, running a gradient between 95% water and 5% acetonitrile to 100% acetonitrile, with all solvents containing 1% acetic acid. The peptide was dissolved in Tris buffered saline (TBS) (10 mM Tris, 5 mM CaCl₂, pH 7.5) at a concentration of 1 mg Ml⁻¹. Tris(2-carboxyethyl)phosphine (TCEP) was added at a final concentration of 1 mmol to prevent disulfide bond formation. To initiate degradation, proteinase K as added at a final concentration of 1 μg mL⁻¹. Degradation proceeded at room temperature and at designated time points, 10 μL of the reaction mixture was injected for analysis by LCMS for relative quantification of degradation products.

Quartz crystal microbalance with dissipation (QCM-D)

A 2% w/v polymer solution was prepared in DCM (base polymers) or N-methyl-2-pyrrolidone (peptide-containing polymers). Spin coating was performed under dry conditions (8% RH) using a Headway Research Inc. (Galway, TX) spin coater. 70 μL of polymer solution was added to the surface of a gold QSensor (Nanoscience Instruments) and spin coated at 3000 rpm for 30 seconds. Coated sensors were dried under vacuum for 16 hours prior to analysis.

Measurements were performed on a QSense Pro quartz crystal microbalance with dissipation. All solutions were sonicated and equilibrated at 37 °C before use. Sensors were primed with Tris buffered saline with 5 mM CaCl₂ (TBS) prior to ensure a stable baseline. Measurement runs consisted of the following steps: (1) TBS (10 min, 20 μL min⁻¹), (2) 10% FBS in TBS (30 min, 20 μL min⁻¹), (3) TBS (30 min, 20 μL min⁻¹), (4) BSA (0.1 mg mL⁻¹), proteinase K (0.1 mg mL⁻¹), trypsin (1 mg mL⁻¹), chymotrypsin (0.1 μg mL⁻¹) or lipase from porcine pancreas (0.1 mg mL⁻¹) (10 hours, 10 μL min⁻¹), (5) TBS (1 hour, 20 μL min⁻¹). At least three sensors were analyzed for each polymer and degradation condition. The Sauerbrey equation was used to calculate mass and thickness under the assumption and observation of negligible changes in dissipation.

Scanning electron microscopy (SEM)

Solutions of HTyDD2%Pep and HTyDD2%Pep(Ctrl) were prepared at a concentration of 40 mg mL⁻¹ in DCM and were equilibrated under orbital shaking for 16 hours. Solutions were cast into Teflon dishes and allowed to dry under a saturated DCM environment for 8 hours before being placed in a vacuum oven overnight to remove residual solvent. A 3 mm biopsy punch was used to punch uniform circular films. Films were incubated in a solution of proteinase K (1 mg mL⁻¹) at 37 °C, shaking, for three weeks. Enzyme solution was replaced every 2–3 days. After three weeks, films were rinsed thrice with DI water, and then briefly with 70% ethanol. Films were dried under vacuum for at least 24 hours. Films were sputter coated with gold/palladium (5 nm) and imaged (FEI Quanta 600 ESEM, voltage = 5 kV).

Cell compatibility assessment

RAW 264.7 macrophages (ATCC) were maintained in complete Dulbecco’s Modified Eagle Medium (DMEM) (Life Technologies) (DMEM supplemented with 10% FBS (Gibco) and 35 μg mL⁻¹ gentamicin). Cells were lifted by gentle scraping and passaged upon reaching 80% confluence. Cells between passage 6–12 were used for all experiments.

Spin coated polymer films were prepared with 2% w/v solutions of polymer. 50 μL of polymer solution was deposited on 10 mm glass coverslips and spin coated at 3000 RPM for 30 seconds. Coverslips were dried under vacuum for 16 hours. Films were sterilized under UV for 30 minutes on each side. Films were rinsed with sterile Dulbecco’s phosphate buffered saline (DPBS) (Life Technologies).

For evaluation of RAW264.7 cell proliferation on the polymer films, RAW264.7 cells were seeded at a density of 1.25 × 10⁴ cells per cm² onto 10 mm glass coverslips spin-coated with HTyDD or HTyDD2%Pep. Cells were seeded at the same density into a well of a tissue culture treated 48-well plate as a control. Three coverslips or wells were seeded in parallel. alamarBlue (Thermo Fisher Scientific) was used according to manufacturer’s instructions at Days 1, 4, and 7 to quantify changes in metabolic activity over time. Increases in metabolic activity over time were associated with increases in cell number.

Osteoclasts were generated from RAW264.7 macrophages as previously described.³⁸ RAW264.7 cells were seeded at a density of 1.25 × 10⁴ cells per cm². Media was supplemented with 20 ng mL⁻¹ mouse recombinant RANKL (R&D Systems). Media was changed every 48 hours. Osteoclasts were visible between 72–96 hours after initial seeding.

RAW264.7 macrophages or RAW264.7-derived osteoclasts were gently lifted using a cell scraper. RAW264.7 macrophages were seeded onto polymer-coated coverslips at a density of 1.25 × 10⁴ cells per cm². RAW264.7-derived osteoclasts were transferred to polymer-coated coverslips without dilution. Cells were cultured for an additional 48 hours on the polymer films in complete DMEM (RAW264.7 macrophages) or complete DMEM containing 20 ng mL⁻¹ RANKL (RAW264.7-derived osteoclasts).

Samples were fixed in 4% paraformaldehyde (Affymetrix) in PBS for 15 minutes to prepare them for imaging. Coverslips were washed in 0.05% Tween-20 in PBS (PBT) for five minutes three times, then incubated in a blocking solution (PBT + 1% BSA + 10% normal goat serum (Life Technologies)) for one hour. A staining solution of phalloidin-488 (Invitrogen, 1 : 50 dilution in blocking solution) was used to stain samples for actin expression. Samples were washed three times with blocking solution prior to incubation with Hoechst (AnaSpec) (1 : 10 000 in blocking solution) for five minutes. Samples were washed with DPBS three times prior to imaging. Images were acquired on a Zeiss fluorescent microscope and analyzed in ImageJ.

Results & discussion

Poly(ester-arylate) selection

Polymers were chosen from a library of tyrosol-derived poly(ester-arylates) previously designed in the laboratory. The HTy-derived poly(ester-arylates) have been used for applications in 3D-printing and bone regeneration,³⁴ and thus, a subset of the HTy-based polymers were chosen. The glass transition temperature (T_g) of a polymer can affect its sensitivity to enzyme-mediated hydrolysis in vivo. The propensity for polymer crystallization will increase if the polymer (1) has symmetrical or linear repeated segments and (2) is brought above its T_g during processing or after implantation. The crystallinity of the polymer in vivo can determine the accessibility of the labile bonds to the enzymes. For PLA, it has been documented that proteinase K has the ability to degrade amorphous, but not crystalline, regions of a polymeric device.³⁰ It was hypothesized that a similar effect would be observed with the tyrosol-based poly(ester-arylates), and therefore, different diacids were chosen to modulate these thermal properties and propensity for crystallization.

It has been shown that modulating the diacid length and symmetry can be used to control final polymer properties.³¹ For initial polymer screenings, the following diacids were chosen: glutaric acid (Glu), azelaic acid (Az), dodecanedioic acid (DD), and 1,4-phenylenediacetic acid (PDA). Polymer nomenclature is defined in Table 1. The solubility of the monomer units dictates how quickly the polymer will resorb after being degraded. For example, polymers containing glutaric acid, with a solubility of ~1600 mg mL⁻¹, are expected to resorb faster than those containing dodecanedioic acid, which has a solubility of 0.04 mg mL⁻¹ (Table 2). In general, increasing diacid length resulted in decreased glass transition temperature.³¹ The HTyGlu polymer has a significantly higher glass transition temperature (33 °C) than the HTyAz and HTyDD polymers (~4 °C). Polymers with an even number of carbons in the diacid, such as HTyDD, are more symmetric than those composed of diacids with an odd number of carbons in the diacid, such as HTyGlu and HTyAz. A higher degree of symmetry facilitates polymer chain stacking through aromatic π–π interactions, leading to crystallization. The additional bulky and rigid aromatic group in PDA also resulted in a higher glass transition temperature (51 °C) when copolymerized with HTy (HTyPDA). While the latter polymer has multiple sites for stacking, the asymmetry of HTy inhibits its ability to crystallize. Based on this data, it was predicted that HTyGlu and HTyPDA would exhibit the greatest sensitivity to enzymatic degradation by proteinase K, followed by HTyAz, then HTyDD. While the mechanical properties of the polymer are not expected to affect surface resorption in this study, they are important considerations when evaluating a polymer’s utility in tissue engineering. The tensile modulus of these polymers ranges from 280 MPa (HTyDD) to 1.12 GPa (HTyGlu), and we have previously shown that post-processing steps can be utilized to further increase polymer stiffness ~2–3-fold.³¹

Table 1.

Polymer nomenclature and molecular weights of pre-polymer blocks and final polymers based on Scheme 1

Polymer	Y	Polymer	Mwa (kDa)	Mna (kDa)
Poly(HTy glutarate)	–(CH2)3–	HTyGlu	157.6	90.9
Poly(HTy azelate)	–(ch2)7–	HTyAz (r = 0.8)	5.9	9.0
		HTyAz (r = 0.8) bismaleimide	11.0	15.0
		HTyAz8%Pep	270.1	142.1
		HTyAz	172.6	83.6
Poly(HTy phenylenediacetate)	–CH2–(Ar)–CH2–	HTyPDA (r = 0.8)	3.3	5.9
		HTyPDA (r = 0.8) bismaleimide	9.2	12.8
		HTyPDA8%Pepb	96.7	50.6
		HTyPDA	156.0	82.4
Poly(HTy dodecanedioate)	−(CH2)10−	HTyDD(r = 0.8)	10.2	7.0
		HTyDD (r = 0.8) bismaleimide	14.0	8.8
		HTyDD8%Pep	140.4	63.0
		HTyDD (r = 0.95)	45.6	26.4
		HTyDD (r = 0.95) bismaleimide	53.3	30.3
		HTyDD2%Pep(Ctrl)	140.1	57.2
		HTyDD2%Pep	112.8	76.7
		HTyDD	157.9	86.9

^aAll values were measured using polystyrene standards with chloroform + 0.1% TFA as an eluent unless otherwise noted.

^bMeasured using polystyrene standards with DMF + 0.1% TFA as an eluent.

Table 2.

Physical and thermal properties of base polymers

Polymer	Diacid solubility (mg mL−1)	Tg (°C)	Tm (°C)	% Crystallinity	Crystallite size (Å)
HTyGlu	1600b	3331	13831	32	122
HTyAz	2.41	431	7231	60	113
HTyPDA	0.16	5131	13331	Amorphous	N/A
HTyDD	0.041	631	8931	36	112
HTyDD2%Pep(Ctl)	—	7	88	40	97
HTyDD2%Pep	—	7	90	28	136
HTyDD8%Pep	—	12	79	a	a

^aNot quantified due to low solubility of polymer in DCM.

^bThe human metabolome database.

Synthesis & chemical structure

Synthesis of pre-polymer blocks with a diol:diacid ratio of 0.8 and 0.95 resulted in number average molecular weights of 8–10 kDa and 25–30 kDa, respectively (Table 1). Capping the pre-polymer blocks with 6-MHA resulted in a slight increase in molecular weight. Incorporation of the peptide into the polymer backbone required a co-solvent system of DCM and DMF due to differing solubilities between the base polymer and the peptide. DMF has also been shown to increase the rate of thiol–ene reactions.³⁹ Final molecular weights were 60–90 kDa (M_n) and 110–160 kDa (M_w) (Table 1). Successful incorporation of the degradable peptide into the polymer backbone was confirmed by NMR by ensuring a 1 : 1 ratio of methionine to pre-polymer block (Fig. 2 and Schemes S1–S9, Fig. S1–S4, ESI†). Peptide bond presence in the polymer was also confirmed through the presence of characteristic infrared spectra peaks of Amide A (3300–3500 cm⁻¹) and Amide I (1600–1800 cm⁻¹) in the degradable peptide and peptide-incorporated polymers (Fig. 2).⁴⁰ Amide peak magnitudes were greatest in the peptide, with decreasing peak amplitude with decreasing peptide incorporation into the polymer. The FTIR spectra HTyDD base polymer did not show any peaks in these regions.

Fig. 2.

(top) Spectral confirmation of peptide incorporation using FTIR. FTIR spectra of the degradable peptide, HTyDD, and HTyDD with 2% and 8% peptide incorporation are shown. Characteristic amide peaks are visible in the peptide and peptide-incorporated polymers between 3300–3500 cm⁻¹ (Amide A) and between 1600–1800 cm⁻¹ (Amide I). (bottom) Spectral confirmation of HTyDD (r = 0.95), HTyDD (r = 0.95) bismaleimide, and HTyDD2%Pep by ¹H NMR in CDCl₃. Peak assignments can be found in Schemes S1–S9 (ESI†). Maleimide peaks designated at 6.68 ppm (left) and methionine (peptide) peak designated at 2.09 ppm (right).

Thermal & physical properties

The measured diacid solubilities, T_g, T_m, percent crystallinity, and crystallite size are shown in Table 2. Polymers composed of diacids with a range of solubilities were chosen. The calculated solubility of 1,4-phenylenediacetic acid in PBS is 0.16 mg mL⁻¹. Low (2%) peptide incorporation into HTyDD did not change thermal properties significantly (Fig. S7, ESI†). Higher peptide incorporation (8%) increased the glass transition temperature from 6–7 °C to 12 °C, and it also reduced the melting temperature from 89–90 °C to 79 °C (Table 2). Increased peptide incorporation increases entropy in the system, making it more difficult for crystalline regions to form. Though 2% peptide incorporation decreased the decomposition temperature (T_d) slightly (Fig. S8, ESI†), the T_d remained >300 °C, retaining its ability to be thermally processed. The higher peptide incorporation resulted in decreased solubility of the final polymer in common solvents such as DCM and DMF, likely due to increased hydrogen bonding between and within the polymer chains. The propensity for crystallization of the base polymers was evaluated by measuring percent crystallinity and crystallite size in a solvent-casted polymer film (Fig. S9, ESI†). The film was cast at room temperature, below the T_g of HTyGlu and HTyPDA and above the T_g of HTyAz, HTyDD, and all peptide-incorporated HTyDD polymers. The films were allowed to dry slowly under a solvent-saturated environment, thus allowing time for crystals to form.

HTyPDA was the only polymer in the study to remain completely amorphous, because the temperature was not elevated above its glass transition temperature during handling, processing, or characterization. While it was expected that HTyDD would be most crystalline due to the long diacid chain providing the greatest polymer chain flexibility, HTyAz was the most crystalline of all the polymers. This is due to the symmetry of the HTyAz diol/diacid pair that enables greater chain packing than the asymmetric HTyDD diol/diacid pair. Incorporation of the degradable peptide in HTyDD decreased percent crystallinity but increased crystallite size. The side chains of the amino acids tryptophan and methionine likely interfere with chain stacking, whereas their replacement in the control peptide with glycine residues, whose side chains are a single hydrogen, resulted in increased hydrogen bonding and chain alignment, and thus, a slight increase in the crystallinity.

Peptide validation

Peptide degradation by proteinase K was confirmed using LCMS (Fig. 3). It was hypothesized that proteinase K would first cleave after the tryptophan and then after the methionine, because proteinase K cleaves after the carboxyl group of aliphatic and aromatic amino acids. The peptide was cleaved rapidly after the tryptophan, followed by a slower cleavage after the methionine. One degradation product was not detectable by HPLC, likely due to its high hydrophilicity minimizing its interaction with the C18 column, causing it to elute in the void volume. The control peptide did not degrade within the timeframe tested.

Fig. 3.

Proteinase K degrades degradable peptide (A), but not the control (nondegradable) peptide (E), demonstrating enzyme-specificity through peptide sequence modulation. Relative concentrations of degradation products monitored as a function of time by LCMS (left). Peptide degradation pathway and intermediate products (B, C and D) (right).

Polymer resorption evaluation

QCM-D is a tool that can quantify nanogram-level changes in mass per unit area by measuring changes in vibrational frequencies of an oscillating quartz sensor. When combined with a robotic autosampler, changes in polymer film mass through deposition of proteins or loss of mass from resorption can be quantified in a high throughput manner. It has been previously used to quantify enzymatic resorption of PLA.⁴¹ This technique overcomes the burdensome process of assessing polymer degradation and resorption through traditional long term in vitro degradation studies, which require meticulous and copious weighing of samples and frequent media changes. While mass loss can be recorded during bulk material degradation studies, surface changes are overshadowed by larger changes in the bulk, and most laboratory balances are not sensitive enough to detect mass loss due to the small degrees of surface erosion. QCM-D also allows for high throughput screening of resorption potential of polymer films. Thin film coatings use less material than traditional degradation studies, which allows for conservation of material during initial screening. It is worth noting that it is difficult to use this technique for assessing polymer degradation, as the amount of material deposited on the surface is not typically enough for downstream analysis by GPC for molecular weight loss or NMR for changes in chemical structure.

Polymer films on gold-coated quartz crystal sensors were first equilibrated with a buffered saline solution to hydrate the surface and establish a stable baseline. A solution of 10% fetal bovine serum was flowed over the crystals for 30 minutes to deposit a stable layer of serum proteins. While this adds complexity to the system, it better mimics the environment in vivo, where the enzyme of interest must compete with many other proteins to bind to its target. For experimental conditions, BSA (negative control) or enzyme was flowed over the crystals for ten hours. Frequency changes were monitored to reflect changes in mass, and dissipation changes were monitored to reflect changes in stiffness of the film. The deposition of serum proteins resulted in a decrease in frequency and increase in dissipation, which corresponds to an increase in mass deposition and decrease in stiffness, as the proteins are more flexible than the rigid polymer film. For most conditions, changes in dissipation were small, indicating a negligible change in the overall film thickness after the initial protein deposition. Incubation of the polymer film in BSA as the control for hydrolytic degradation did not result in a decrease in film thickness for any of the polymers tested (Fig. 4A). There was no significant difference in the final thickness of proteins on the HTyGlu, HTyAz, and HTyDD polymers; however, there was a significantly greater layer of adsorbed protein on the HTyPDA polymer.

Fig. 4.

BSA adsorption (A) and proteinase K-mediated changes (B) in base polymer film thickness over time. Thickness is normalized to maximum thickness after end of incubation with 10% FBS. Final thickness of polymer films after either BSA or proteinase K incubation for 10 hours (A and B right). Incorporation of proteinase K-sensitive peptide into HTyDD polyester resulted in increased resorption in the presence of proteinase K. Final thicknesses of polymer and polymer8%Pep films after proteinase K incubation for 10 hours (C). Final thickness of HTyDD2%Pep(Ctrl), HTyDD2%Pep, and HTyDD8%Pep after proteinase K incubation for 10 hours (D). Peptide-incorporated polymer is resorbed selectively in the presence of proteinase K. Calculated thickness of HTyDD (E) and HTyDD2%Pep (F) film in the presence of BSA or enzyme over time. Thickness is normalized to maximum thickness after end of incubation with 10% FBS. Final thicknesses of polymer films after incubation with different enzymes for 10 hours (G). Final thicknesses of HTyDD2%Pep films after incubation with different enzymes for 10 hours (H). Data presented as mean ± standard deviation (n = 3 sensors). One-way ANOVA with a post-hoc Tukey test was performed on relative final thicknesses (A–D and H). Two-way ANOVA with a post-hoc Tukey test was performed on relative final thicknesses (G). * denotes significance at p < 0.5. ** denotes significance at p < 0.01. *** denotes significance at p < 0.002. **** denotes significance at p < 0.001.

It was hypothesized that polymer sensitivity to proteinase K degradation would decrease with decreasing glass transition temperature and increasing propensity for crystallization because the polymer chains become less accessible in their crystalline form. It was also hypothesized that the rate of resorption would correlate directly with the solubility of the diacid monomer. Here, we show that controlling the thermal properties and solubility parameters of the final polymer allow for control over surface resorption of the polymer. As predicted, HTyGlu showed the greatest sensitivity to proteinase K degradation, with the greatest reduction in thickness (Fig. 4B). HTyPDA and HTyAz showed similar reductions in thickness in the presence of proteinase K, though significantly less than that of HTyGlu. Unlike HTyAz, the T_g of HTyPDA is above physiological temperature, which means the polymer is in its amorphous form during characterization, and theoretically it is more sensitive to protease degradation. However, 1,4-phenylenediacetic acid has a much lower solubility than azelaic acid, thus reducing its resorption rate. HTyDD showed an increase in mass and dissipation as measured by QCM-D, which translates to an overall increase in film thickness. This is likely due to water uptake in the polymer matrix. Dodecanedioic acid is minimally soluble in water, and thus oligomers of HTyDD must be fully degraded before mass loss will be observed. Because of the extent to which HTyGlu resorbed in the presence of proteinase K, this polymer was not included in the evaluation of the effect of peptide incorporation on selective enzymatic degradation.

A proteinase K-sensitive peptide was incorporated at 8 mol% into the HTyAz, HTyPDA, and HTyDD polymers. Incorporation of peptide at 8% into HTyAz and HTyPDA did not significantly reduce the final polymer thickness, indicating that it had little effect on polymer resorption (Fig. 4C). This could be due to the moderate sensitivity of the base polymer to resorption in the presence of proteinase K. Incorporation of the degradable peptide into HTyDD at 8 mol% increased polymer resorption significantly in the presence of proteinase K. When comparing HTyDD2%Pep(Ctrl) to HTyDD2%Pep and HTyDD8%Pep, there was no significant difference in resorption between either of the degradable peptide-incorporated polymers in the presence of proteinase K; however, both experimental peptide-incorporated polymers resorbed significantly more than the control peptide-incorporated polymer, indicating enzyme sensitivity can be controlled by tuning the peptide sequence (Fig. 4D). The HTyDD2%Pep polymer was used to evaluate the specificity of enzymatic resorption compared to the base polymer due to its improved solubility in common solvents.

BSA was used as a model protein to evaluate nonspecific protein adsorption to the polymer surface. BSA adsorption was similar on both HTyDD and HTyDD2%Pep, and a broader range of proteins would need to be evaluated to fully elucidate the effects of peptide incorporation on protein adsorption in the body. Peptide incorporation decreased the polymer’s sensitivity to lipase- and chymotrypsin-mediated resorption and increased its sensitivity to trypsin-mediated resorption (Fig. 4E–G). Lipases are a subclass of esterases, and therefore may be able to cleave the esters and arylates present in the backbone of the poly(ester-arylates) used in this work. Chymotrypsin cleaves on the carboxyl side of hydrophobic amino acids, and the large percentage of tyrosine-based monomers in the polymer renders the polymer sensitive to this enzyme. Peptide incorporation, even at 2%, reduces the efficacy of both degradation pathways. Peptide incorporation had no effect on collagenase type I-mediated resorption. When comparing all enzymes tested, proteinase K was the only enzyme capable of reducing the polymer film thickness to less than its starting thickness, thus validating enzyme-specific polymer resorption (Fig. 4H).

Changes in surface morphology upon incubation with proteinase K was visualized using SEM. After three weeks of incubation with the enzyme, an increase in surface roughness and pitting was observed in the HTyDD2%Pep polymer compared to the HTyDD2%Pep(Ctrl) polymer, indicating enzyme-driven surface erosion in a peptide-specific manner (Fig. 5).

Fig. 5.

SEM images of polymer film surfaces prior (top) and after (bottom) proteinase K incubation for three weeks. Scale bar = 100 μm (full size) and 30 μm (inset).

Biocompatibility of peptide–polymers

The HTyDD and HTyDD2%Pep polymers were chosen for cell compatibility evaluation due to their differences in resorption potential. RAW264.7 macrophages were cultured on the polymer films and their changes in metabolic activity were quantified by alamarBlue. Both polymers supported cell proliferation, as indicated by the increase in metabolic activity over time (Fig. 6A). RAW264.7 macrophages and RAW264.7-derived osteoclasts were cultured on polymer films of HTyDD and HTyDD2%Pep and stained to visualize actin and nuclei. Representative images are shown in Fig. 6B. RAW264.7-derived osteoclasts exhibit characteristic morphological features of osteoclasts, such as the presence of multiple nuclei and expression of actin circumferentially around the giant cell body.⁴² This demonstrates the ability of these polymers to support the attachment and survival of cell types responsible for resorption in vivo.

Fig. 6.

(A) Metabolic activity of RAW264.7 macrophages on HTyDD, HTyDD2%Pep, and tissue culture polystyrene (TCPS) over 7 days. Values at Day 4 and 7 are normalized to Day 1. Data are represented as n = 3 ± SD. (B) Fluorescence imaging of actin (green) and nuclei (blue) of RAW264.7 macrophages (top) and RAW264.7-derived osteoclasts (bottom) on HTyDD (left) and HTyDD2%Pep (right) polymers after 48 hours of culture. Scale bar = 100 microns.

Conclusions

This work was inspired by the body’s natural control over matrix protein remodeling through differential enzyme expression by cells. Here, we have demonstrated the ability to engineer specificity in the enzymatic resorption of a hydrophobic thermoplastic by incorporating a peptide sequence derived from type I collagen, and proteinase K was used as the model enzyme for this study. Block copolymers of maleimide-capped tyrosol-derived poly(ester arylates) and a thiol-capped enzyme-sensitive peptide were synthesized via Michael addition. It was confirmed that 2% peptide addition into HTyDD did not affect thermal properties or crystallinity significantly, but it was enough to render the polymer resorbable in the presence of proteinase K. QCM-D was used as a technique to measure mass loss of a thin film during enzyme incubation, and it proved to be much more sensitive at quantifying surface changes than traditional methods of evaluating in vitro polymer degradation and resorption. These peptide-incorporated polymers also support the attachment and survival of macrophages and osteoclasts – cell types responsible for resorption in vivo. Future work will confirm the cell-mediated resorption profiles of these polymers. From the preliminary work shown here, the base polymer and peptide sequence can be modified to generate libraries of polymers that are tailored to resorb in different biological environments, and using QCM-D, the efficiency of resorption can be evaluated in a high throughput manner. Together, this work motivates the need to let biology guide polymer design. It also highlights the potential for cell-specific remodeling of thermoplastics, which has implications in greatly improving the integration, vascularization, and overall success of these polymers used in regenerative applications.