Modulating nonlinear elastic behavior of biodegradable shape memory elastomer and small intestinal submucosa(SIS) composites for soft tissue repair

Abstract

Structural repair of soft tissue for regenerative therapies can be advanced by developing bioCompatible and bioresorbable materials with mechanical properties similar to the tissue targeted for therapy. Developing new materials modeling soft tissue mechanics can mitigate many limitations of material based therapies, specifically concerning the mechanical stress and deformation the material imposes on surrounding tissue structures. However, many elastomeric materials used in soft tissue repair lack the ability to be delivered through minimally invasive surgical (MIS) or transcatheter routes and require open surgical approaches for placement and application. We have developed a bioCompatible and fully biodegradable shape memory elastomer, poly-(glycerol dodecanedioate) (PGD), which fulfills the requirements for hyperleasticity and exhibits shape memory behavior to serve as a novel substrate material for regenerative therapy in minimally invasive clinical procedures. Our previous work demonstrated control over the tangent modulus at 12.5% compressive strain between 1–3 MPa by increasing the crosslinking density in the polymer. In order to improve control over a broader range of mechanical properties, nonlinear behavior, and toughness, we 1) varied PGD physical crosslink density, 2) incorporated sheets of porcine small intestinal submucosa (SIS, Cook Biotech, Inc.) with varying thickness, and 3) mixed lyophilized SIS particulates into PGD at different weight percentages. Tensile testing (ASTM D412a) revealed PGD containing SIS sheets of were stiffer than controls (p < 0.01). Incorporating lyophilized SIS particulates into PGD increased the strain to failure (p < 0.001) compared to PGD controls. Test specimens with 1 ply sheets had greater tear strength (ASTM D624c) compared to PGD tear specimens prepared control specimens (p < 0.001). However, incorporating SIS particulates decreased tear strength of PGD-SIS 0.5wt% particulate composites (p < 0.01) compared to PGD controls. Incorporating 2 ply and 4 ply sheets and 0.5wt% particulates into PGD decreased the fixity and recovery of composite materials compared to controls (p < 0.01). Nonlinear modeling of stress strain curves under uniaxial tension demonstrated tunability of PGD-SIS composite materials to model various nonlinear soft tissues. These findings support the use of shape memory PGD-SIS composite materials towards the design of implantable devices for a variety of soft tissue regeneration applications by minimally invasive surgery.

Graphical Abstract

Introduction

Current clinical therapies for soft tissue repair utilizing biomaterials are limited in meeting mechanical requirements over the entire regenerative time scale while concurrently providing options for minimally invasive delivery methods(Chahal et al., 2012; Mann et al., 2016; Weinberger et al., 2017). Metallic transcather devices and injectable hydrogel therapies have gained prevalence in clinical practice, but challenges remain with respect to the suitability of their mechanical properties upon delivery. Metallic devices are often considerably stiffer than surrounding tissues and hydrogels exhibit nonlinear and viscoelastic properties similar to or often much less than soft tissues. Moreover, limited tissue ingrowth, and consequent drift or embolization leading to secondary surgical interventions are reported in various soft tissue regeneration applications (Akagi, 2014; Divekar et al., 2005; Riegler et al., 2015; Zhu et al., 2017, 2012). For example, cardiac device failure can often be attributed in part due to the mismatch between the device mechanical properties and the nonlinear mechanical properties of the surrounding heart tissue. The growing need to reduce allograft immunogenicity and autograft donor site morbidity during musculoskeletal soft tissue regeneration applications also emphasizes the urgency for novel material-based tissue engineering approaches to meet the mechanical requirements of surrounding soft tissues (Chahal et al., 2012; Walden et al., 2017; Yang et al., 2013).

Soft tissues exhibit nonlinear elastic or viscoelastic behavior arising from the heterogeneity in molecular composition and organization of extracellular matrix architectures. Stress strain curves from various uniaxial tensile testing results of the achilles and superspinatus tendons, meniscal attachments, and facial cartilage have been fit to various nonlinear models to develop tissue grafting substitutes (Abraham et al., 2011; Morales-Orcajo et al., 2017; Szczesny et al., 2012; Zopf et al., 2015). Ventricular free-wall myoCardium of guinea pigs has also been mechanically tested and fit to constitutive models in order to predict ventricular wall stress and compressibility during the cardiac cycle (Hassaballah et al., 2013; Hassan et al., 2012).The material parameters derived from these constitutive models can be used as biomaterial design targets for respective soft tissue applications.

Various families of biomaterials have been used for structural soft tissue repair ranging from thermoplastic polymers of α-hydroxy esters to natural and synthetic hydrogels (Atala et al., 2012; Freed et al., 1994; He et al., 2010; Lee et al., 2014; Seidi et al., 2011; Temenoff and Mikos, 2008). Intrinsic material properties in conjunction with various biofabrication methods have been used to tailor the mechanical properties of these materials to match the mechanical properties of native soft tissues. These materials are initially required to fill the defect site and attach to surrounding tissue margins. Current therapies addressing soft tissue repair are challenged by the initial functional and mechanical properties as well as the dynamically changing mechanical characteristics of the resorbable polymer. Synthetic polymers such as PCL, PLGA, and PLA exhibit linear elastic behavior with moduli in the hundreds of MPa to GPa range, resulting in a mismatch of mechanical properties with surrounding tissue leading to device failure due to tissue erosion and adverse tissue remodeling. Hydrogels exhibit nonlinear viscoelastic behavior similar to soft tissues, but are subject to migration or rapid resorption requiring multiple proCedural applications.

A new class of biodegradable and biocompatible elastomers exhibit bulk properties ranging between natural hydrogels and synthetic polymers, which gives rise to new design possibilities for various soft tissue repair applications. These elastomers include poly-(glycerol sebacate), poly-octanediol co-citrate, and alginate-based supramolecular ionic polyurethanes(Daemi et al., 2016; Kang et al., 2006; Wang et al., 2002). Among these materials, poly-(glycerol sebacate) and its modified variations with other polymer composites and surface modifications have shown great promise for a variety of soft tissue engineering applications (Nijst et al., 2007; Rai et al., 2013, 2012). Additionally, various manufacturing methods have been used to further tune the nonlinear elasticity of these polymers for structural soft tissue repair applications. Particulate leaching, eximer laser micropatterning, electrospinning, and 3D printing poly-(glycerol sebacate) are a few of the methods used to further control microstructure and mechanical behavior of scaffolds for soft tissue reconstruction (Engelmayr et al., 2008; Ifkovits et al., 2008; Jeffries et al., 2015; Mitsak et al., 2012). Although these approaches demonstrate the ability to provide appropriate mechanical properties, tissue ingrowth challenges remain for implantable device designs for minimally invasive surgical applications, which are increasingly becoming the standard of care for a variety of soft tissue regeneration applications. Fracture or failure of conventional polymers during delivery, compaction, and tearing of elastomers remain challenges for the device delivery for MIS repair of soft tissues.

To address the mechanical requirements for tissue repair and the clinical requirements for device delivery, we developed a shape memory biodegradable elastomer poly(glycerol dodecanedioate) (PGD), previously referred to as poly(glycerol dodecanoate) but corrected here. This novel polymer is comprised of the bioCompatible monomers glycerol and dodecanedioic acid, which is an elastic material at room temperature but exhibits nonlinear elastic behavior as an elastomer at body temperature(Migneco et al., 2009). Shape memory polymers are of increasing interest for biomedical applications because of their ability to maintain a rigid shape at room temperature that is appropriate for implantation through commonly used transcatheter delivery sheaths or trochars in MIS applications to the targeted anatomic site. The biodegradable device then undergoes a phase change, returning to its permanent pre-programmed shape within the tissue (Figure 1b,c) when heated beyond its transition temperature (T_trans). Previously, we described the shape memory behavior of this material, bioCompatibility and biodegradation, and the ability to tune the mechanical properties of the material under various thermal curing conditions (Ramaraju et al., 2020; Solorio et al., 2017). This allows us to control the mechanical properties of PGD by varying the crosslink density in order to match the mechanical properties of various soft tissues. However, the tunability of the materials using crosslink density still exhibited limited range of mechanical properties and constrained the range of transition temperatures (T_trans) required for the shape memory cycle (Figure 1b,c).

Figure 1.

Synthesis and shape recovery behavior of the shape memory polymer, poly-(glycerol dodecanedioate, PGD). A) Polycondensation reaction of poly-(glycerol dodecanedioate) pre-polymer. The pre-polymer is synthesized from glycerol and dodecanedioic acid in a 2-step process: 24 hours at 120°C under dry compressed N_2, followed by another 24 hours under a 30mTorr vacuum at 120°C. B) Schematic of a thermally activated shape recovery cycle where a cylinder is heated beyond the transition temperature to program a shape under applied force, cooled to store the programmed shape, and finally heated again beyond the transition temperature to once again resume its permanent shape. C) Example of a flat PGD sheet that was rolled up and cooled to store a programmed shape and heated at T_t = 37°C to resume its original flat shape.

To expand the range of controllable mechanical properties of PGD and model anisotropic elastic behavior of surrounding tissues, we incorporated commercially available porcine small intestinal submucosa (SIS, Cook Biotech, Inc.) into thermally cured variants of PGD. SIS is approved for a variety of soft tissue repair applications including cardiovascular, tendon, dura mater, and carotid artery(Feng et al., 2011; Gamradt, n.d.; Mosala Nezhad et al., 2016). Although SIS is indicated for use in a variety of clinical applications, challenges remain with mechanical performance in resorbable applications for soft tissue repair (Murphy et al., 2008; Witt et al., 2013). By incorporating SIS into PGD, we propose to increase tensile strength and tear resistance while tailoring composite material properties to mimic the nonlinear elastic behavior of various soft tissues, consequently reducing device failure and improving handling. To achieve these objectives, we incorporated SIS sheets of varying thicknesses (1 ply, 2 ply, 3 ply, and 4 ply) into PGD to improve tear resistance and tensile strength of the composite material, and to verify the modification of the nonlinear elastic properties to meet the requirements of various soft tissues. We also included SIS particulates to increase elasticity to meet the mechanical requirements of softer tissues.

2. Materials and Methods

2.1. PGD synthesis and PGD-SIS composite specimen fabrication

Poly-(glycerol dodecanedioate) was synthesized as previously described (Migneco et al., 2009; Solorio et al., 2017). Briefly, an equimolar amount of glycerol (MP Biomedical, LLC, Solon, OH) and dodecanedioic acid (Sigma-Aldrich, St. Louis, MO) were mixed at 120°C under nitrogen for 24 h. The reaction was then switched to vacuum at 30 mTorr at 120°C for an additional 24 h (Figure 1a). Pre-polymer was subsequently cooled and stored in a dessicator until further use. Tensile specimens were prepared by pre-heating the pre-polymer to 90°C until completely melted, and pouring into a pre-heated PTFE-mold (S1. a), shaped according to ASTM standard D412a. Tear specimens were prepared by pouring pre-polymer into a PTFE-mold shaped, according to ASTM standard D624c. Tensile and tear specimens with SIS sheets were prepared by cutting 1 ply, 2 ply, 3 ply, and 4 ply SIS sheets (Cook Biotech, West La Fayette, IN) to appropriate shapes, pre-heating in the tensile molds, and subsequently pouring in pre-heated PGD pre-polymer (S1. b–d). PGD-SIS particulate composite specimens were prepared by mixing 0.1wt% and 0.5wt% particulates into the preheated pre-polymer and pouring into a tensile or tear specimen mold. All molds were subsequently cured under vacuum at 90 mTorr for 72 h at 120°C for the low cure condition or 48 h at 130°C for the high cure condition. Specimens were removed from the mold, cooled, dried, and stored in a vacuum dessicator until they were tested.

2.2. Swelling test

Samples were laser cut to 6mm diameter and 1.5mm thickness. Samples were weighed to obtain the initial dry weight (m_i), and subsequently immersed in THF and incubated at 37°C for three days (n=6). Samples were weighed daily for any observable weight differences. After four days of incubation in THF, samples were removed and weighed for the equilibrium weight measurement (m_eq). After 2 days of drying in the fume hood, samples were weighed again for the dry mass measurement (m_d). The swelling ratio was calculated according to Equation (1) as previously described (Solorio et al., 2017)

$$Swellingratio=\frac{m_{eq}-m_d}{m_d}$$ (1)

2.3. Differential Scanning Calorimetry

Differential Scanning Calorimetry was conducted using a Discovery Q250 with an RCS90 cooling system (TA instruments, New Castle, DE). DSC samples (7g) from PGD and PGD-SIS particulates were trimmed from the cross section of an untested specimen. PGD-SIS sheet samples were trimmed from the full thickness cross section of an untested specimen. The samples (n=4/group) were dried overnight in a vacuum dessicator, weighed, and placed in a Tzero^® pan. Samples were heated to 90°C from 25°C at a rate of 10°C /min to remove thermal history. After an isothermal hold for 3 min, samples were cooled at a rate of 5°C /min to −50°C and heated back up to 70°C at a rate of 5°C /min. Melting transition (T_m) and enthalpy of fusion (ΔH_fus) were measured using the TA Instruments analysis software Trios Version 4.4.

2.4. Mechanical testing

Tensile specimens (S1. b,c; n=6/group) comprised of PGD, PGD-SIS sheets, and PGD-SIS particulates (n=6/group) were tested to failure using an MTS RT/Alliance mechanical test frame with a 500N load cell (MTS Systems Corp) and tangent modulus at 5% strain, failure strain, and nonlinear elastic properties were assessed using dogbone specimens at a loading rate of 50mm/min according to ASTM standard D412a. Tear testing specimens were conducted at a testing rate of 5mm/min according to ASTM standard D624c. Tensile and tear tests were conducted at 37°C in a custom temperature control chamber.

2.5. Scanning Electron Microscopy

Tensile and tear specimens were cut at the fracture interface, and mounted to SEM sample holders. Samples were sputter coated with a thin gold film for 60s using a Quorum Q-150T ES (Quorum Technologies, East Sussex, England) and imaged using a LEO 1530 SEM at 6kV. Representative images of high magnification (1000×) and low magnification (100×) for each group were chosen from a total of 4 samples per group, 4 field of views (FOV) per sample, and 2 different magnifications per FOV.

2.6. Shape Recovery analysis

Composites of PGD and SIS sheets and particulates were cured as previously described in 100mm × 900mm × 2mm rectangular molds. Low cure and high cure PGD rectangular coupons (n=10/group and programming angle) comprised of sheets (1 ply, 2 ply, 3 ply, and 4 ply) and particulates (0.1wt% and 0.5wt%) were laser cut to 30mm × 5mm × 2mm rectangular bars for shape recovery testing. Samples were heated to 60°C and bent to either 145°C or 160°C angles. Samples were allowed to cool to room temperature and the programmed angle was measured at a stress free condition. Samples were then held in a water bath at 37°C, and time to recover as well as final recovery angle was recorded (Figure 8a). Shape fixity and shape recovery were calculated using Equations (2) and (3) respectively, as previously described but adapted to this angular recovery test (Behl and Lendlein, 2007).

$$R_f=\frac{{\text{θ}}_{\text{u}}(\text{N})}{{\text{θ}}_{\text{m}}}$$ (2)

$$R_r=\frac{{\theta }_m-{\theta }_p(N)}{{\theta }_m-{\theta }_p(N-1)}$$ (3)

Equation 2 was used to determine shape fixity and Equation 3 was used to determine shape recovery, where θ_m is the mechanical deformation that causes the programming of the temporary shape ${\theta }_u(N),{\theta }_m-{\theta }_p(N)$ is the change in the angle during the course of recovery, and ${\theta }_m-{\theta }_p(N-1)$ is the change in the angle during the course of programming. Recovery rate (radians/min) was also calculated by the following equation.

Figure 8.

Shape recovery testing of PGD-SIS sheets and particulates (mean ± SD, n=10). A) Shape recovery of low cure PGD with 0.1wt% SIS particulates at 37°C in aqueous environment. B) Shape fixity of PGD-SIS sheets (^p<0.01 compared to high cure controls and #p<0.01 compared to 1 ply high cure PGD). C) Shape recovery of PGD-SIS sheets (#p<0.01 compared to low cure controls). D) Recovery rate of PGD-SIS sheets (#p<0.001 compared to controls). E) Shape fixity of PGD-SIS particulates (^p<0.001 compared to high cure controls and #p<0.05 compared to 1 ply high cure PGD). F) Shape recovery of PGD-SIS particulates. G) Recovery rate of PGD-SIS particulates (#p<0.001 compared to high cure controls, ^p<0.033 compared to 0.5wt% high cure PGD, and *p<0.01 compared to high cure controls). Error bars represent standard deviation.

$$\frac{d{R}_r}{dt}=\frac{{\theta }_m-{\theta }_p(N)}{\text{Δ}t}$$ (4)

2.7. Constitutive modeling of materials and tissues

Stress-strain measurements from uniaxial tensile tests were fit to 1-term and 4-term nonlinear Ogden strain energy functions (Ogden et al., 2004), shown in Equation (5). PGD, PGD-SIS 1 ply sheets, and PGD particulates were fit to a 1-term Ogden strain energy function, and PGD polymer with 2 ply, 3 ply, and 4 ply SIS sheets were fit to a 4-term Ogden strain energy function.

$$W\left({\lambda }_1,{\lambda }_2,{\lambda }_3\right)=\sum _{i=1}^N\frac{{\mu }_i}{{\alpha }_i}\left({\lambda }_1^{{\alpha }_i}+{\lambda }_2^{{\alpha }_i}+{\lambda }_3^{{\alpha }_i}-3\right)$$ (5)

This equation represents the strain energy function, where W is the strain energy function, μ_i is the shear material coefficient, α_i is a material coefficient, and λ_1, λ_2, λ₃ are the principal stretch ratios in the x, y, and z directions respectively. The 1^st Piola-Kirchoff (PK) stress tensor components in the principal directions for an incompressible material are then expressed as

$$T_i=-p\frac{1}{{\lambda }_i}+\frac{\partial W}{\partial {\lambda }_i}$$ (6)

where λ_i represents the stretch ratios, p is the hydrostatic pressure, T_i represents the principal 1^st PK stress components, and W is the strain energy function from Equation 5. After solving for the hydrostatic pressure p from a zero traction boundary condition, the T₃₃ 1^st PK stress component has the form

$$T_{33}={\sum }_{n=1}^N{\mu }_n\left({\lambda }^{{\alpha }_n-1}-{\lambda }^{-1-\frac{{\alpha }_n}{2}}\right)$$ (7)

where the 1-term model has 2 constants ${\mu }_1,{\alpha }_1$ and the 4-term model has 8 constants $\left({\mu }_1,{\mu }_2,{\mu }_3,{\mu }_4,{\alpha }_1,{\alpha }_2,{\alpha }_3,{\alpha }_4\right)$. The Baker-Ericksen inequality (Knowles and Sternberg, 1976) was also imposed on the materials as a stability requirement, which requires the highest stress in the direction of highest strain. This is mathematically represented as

$$\left(T_{33}-T_{22}\right)\left({\lambda }_3-{\lambda }_2\right)\ge 0$$ (8)

where under tension, the stretch ratio in the principal stretch direction λ₃, is greater than the other stretch ratio ${\lambda }_3$, and subsequently the 1^st PK stress component in the principal stretch direction $\left(T_{33}\right)$ is also greater than $\left(T_{22}\right)$. Models and all plots were generated in Matlab (Mathworks, Natick, MA). A least squares method was used to optimize parameters, utilizing the fmincon() function. Subsequently, R² values were calculated and plots of 1- or 4-term models against experimental data were produced in Matlab.

Models of tensile specimens were created in Solidworks 2016 (Dassault Systems, Velizy-Villacoublay, FR), meshed in 3-Matic 12.0 (Materialise NV, Leuven, Belgium) to contain a 10-node tetrahedral element mesh with 20k-100k elements, and imported into FEBio Preview 2.0. Tensile testing models were created for an unconstrained Ogden model (Figure 9a) and a tensile test was run with a sliding elastic contact, including tensile contact between the grips and specimen with a step size of 0.01 in FEBio 2.7.1 (Figure 9b) (Maas et al., 2012). The parameter optimization module was then used to import experimental data points for each test specimen, and a Levenberg-Marquardt damped least squares method was used to determine material constants based on goodness of fit to experimental data. Resulting force-displacement curves from optimized Ogden FEA models were compared to experimental data as well as 1-term and multiterm Ogden values for meniscal attachments(Abraham et al., 2011) and ventricular myocardium(Hassaballah et al., 2013), or tendon(Morales-Orcajo et al., 2017, 2016) and chordae tendinae(Zuo et al., 2016).

Figure 9.

Finite Element Analysis (FEA) of PGD-SIS and soft tissues fit to Ogden 1-term and multiterm models compared to experimental (Exp) values. A) Tensile model of 30k node low cure PGD with sliding-sliding elastic contacts (Preview 2.0.0, FEBio). B) Stress distribution in 30k node low cure PGD tensile model (Post-View 2.2.0, FEBio). C) High cure 1-term model of PGD and PGD with SIS particulates compared to myocardium and meniscal fiber attachments. D) Low cure 1-term model of PGD and PGD with SIS particulates compared to myocardium and meniscal attachments. E) High cure multiterm model of PGD and PGD with SIS sheets compared to chordae tendinae and tibialis tendon. F) Low cure multiterm model of PGD and PGD with SIS sheets compared to chordae tendinae and tibialis tendon.

2.8. Statistical and Numerical Methods

Means ± standard deviation are depicted in all bar graphs and tables. All statistical analysis were conducted using JMP 13.1 (SAS Inc., Atlanta, GA). Mean and standard deviations(SD) were provided for all graphs and tables. Two-way ANOVA on ranks using Tukey post-hoc tests for all pairwise comparisons were done in Graphpad Prism (Graphpad Software Inc., La Jolla, CA) to analyze swelling ratios, tensile modulus, strain at break, shape fixity, shape recovery, and recovery rate.

3. Results

Incorporation of sheets into tensile specimens did not reveal significant differences in material properties of PGD between regions proximal and distal to the incorporated SIS sheet (S2. a,b). Raman spectroscopy of the tensile specimen gauge cross section revealed no significant differences at 1064 cm⁻¹ (C-H), 1108 cm⁻¹ (C-O-C), 1296 cm⁻¹ (C-O,C-H), 1439 cm⁻¹ (O-H), and the characteristic antisymmetric carbonyl ester stretch at 1727 cm⁻¹ (C=O) in PGD proximal and distal to embedded SIS sheets (S2. c). SEM micrographs of SIS particulates exhibited non-uniform particulate size ranging from 5–150 microns with a variety of geometries (S3. a). SEM micrographs of PGD-SIS particulate tensile specimen cross sections exhibited a number of void spaces (S3. b), while the surface exhibited characteristic microroughness (S3. c). Incorporation of 0.1wt% PGD particulate cross sections (S3. d) and 0.5wt% PGD particulate cross sections (S3. e) and surfaces (S3. f) within PGD samples revealed uniform distribution of nitrogen (S3. g) throughout the polymer. The thermomechanical behavior of these sheet and composite material variants, coupled with the changing crosslinking density of the base polymer, gave rise to a broader range of mechanical properties and nonlinear mechanical behaviors.

3.1. Swelling

As previously described (Solorio et al., 2017), cure conditions affected the swelling ratio of PGD controls (Figure 2a, p<0.01). Samples containing SIS sheets exhibited differences between cure conditions (Figure 2a, p<0.01 for 1 ply and p<0.001 for 2 ply). Incorporation of SIS sheets within the PGD matrix increased the swelling of PGD (Figure 2a, p<0.001). Incorporation of 0.5wt% SIS particulates into the low cured PGD polymer increased the swelling ratio compared to controls (Figure 2b, p<0.001) and high cure condition (Figure 2b, p<0.001).

Figure. 2 -.

Swelling ratio of PGD with SIS sheets and PGD with SIS particulates in THF for 3 days at 37°C (mean±SD, n=4). A) Low cure and high cure PGD embedded with SIS sheets revealed significant differences in swelling ratios (*p<0.01 and **p<0.001 between cure conditions, and ^p<0.05 compared to controls). B) Particulate wt% and cure condition revealed significant differences in swelling ratios (*p<0.001 amongst cure conditions and ^p<0.001 across particulate incorporation).

3.2. Differential Scanning Calorimetry

Differential scanning calorimetery indicated melt transition coinciding with the shape transition temperate between 31°C and 37°C for all groups (Table 1; Figure 3). As expected, high cure PGD exhibits a lower peak T_m than low cure PGD (p <0.0001). Interestingly, incorporating SIS sheets into PGD decreased the T_m compared to controls in both high cure and low cure conditions (Table 1, p < 0.001). SIS particulate incorporation into PGD resulted in lower T_m compared to controls in the low cure condition (p < 0.01). As expected, ΔH_fus of low cure PGD was greater than high cure PGD in both PGD with SIS sheet and particulate incorporations (p < 0.001). The thickness of the SIS sheets incorporated into high cure and low cure PGD did exhibit significant changes in ΔH_fus. Incorporation of particulates into high cure PGD did not affect ΔH_fus in comparison to PGD with sheets. Particulate incorporation into low cure PGD increased ΔH_fus compared to controls (p<0.001).

Table 1.

DSC analysis of PGD with SIS sheets and particulates.

Ply	Low Cure		High Cure
	Tm (°C)	ΔHfus (J/g)	Tm (°C)	ΔHfus(J/g)
Control	37.9 ± 0.8	36.9 ± 0.8	35.1 ± 0.3	32.2 ± 1.9
1	35.9 ± 0.1	42.1 ± 2.4	31.3 ± 1.1	28.6 ± 2.8
2	35.7 ± 0.5	42.2 ± 0.5	31.9 ± 1.3	28.3 ± 3.3
3	36.9 ± 0.4	43.0 ± 1.8	33.1 ± 0.9	35.2 ± 2.3
4	36.2 ± 0.6	41.6 ± 1.8	34.5 ± 1.1	30.9 ± 2.0
Wt%	Tm (°C)	ΔHfus (J/g)	Tm (°C)	ΔHfus (J/g)
0.1	35.9 ± 0.4	44.8 ± 3.5	35.4 ± 2.5	33.1 ± 2.4
0.5	34.7 ± 0.4	41.6 ± 2.3	35.7 ± 0.4	33.5 ± 1.4

Figure 3.

Representative DSC thermograms of low cure and high cure PGD sheets (A,B) and particulates (C,D). The representative figures, chosen from 4 thermograms per group, depict similar peaks and melting transition temperatures with the incorporation of sheets and particulates, resulting in comparable enthalpies of fusion.

3.3. Mechanical testing

Incorporation of SIS sheets into PGD increased stiffness (tangent modulus calculated at 5% strain) of both high cure and low cure conditions compared to controls (Figure 4). There was an increase in modulus after incorporation of 1 ply SIS sheets compared to controls (Figure 4c, p<0.01) for high cure conditions. Similarly, PGD with 2 ply sheets exhibited greater modulus than controls (Figure 4c, p<0.001) and 1 ply sheets (Figure 4c, p<0.01) for low cure conditions. PGD with 2 ply sheets also exhibited increased stiffness compared to controls (Figure 4c, p<0.01) for high cure conditions. Incorporation of 3 ply and 4 ply sheets into PGD tensile specimens increased the modulus compared to controls (Figure 4c, p<0.0001) and 1 ply (Figure 4c, p<0.001) for both low and high cure conditions. Additionally, in high cure specimens, incorporation of 3 ply and 4 ply sheets into PGD tensile specimens increased the modulus compared to 2 ply sheets (Figure 4c, p<0.022). The maximum strain was lower for PGD with 1 ply, 2 ply, 3 ply, and 4 ply sheet thicknesses compared to controls (Figure 4d, p<0.001) across both cure conditions.

Figure 4.

Uniaxial tensile testing of low cure and high cure PGD with varying SIS sheet thicknesses at 37°C following ASTM standard D412a (mean ± SD, n=6). A) Representative stress-strain curves for low cure PGD with sheets and controls. B) Representative stress-strain curves for high cure PGD with sheets and controls. (C) Tangent modulus at 5% strain of PGD-SIS sheets and controls (#p<0.001 compared to low cure controls, ^p<0.01 compared to low cured 1 ply, †p<0.001 compared to high cure controls, and ‡p<0.01 compared to high cure 1 ply and 2 ply). (D) Strain at break of PGD-SIS sheets and controls; PGD with 1 ply, 2 ply, 3 ply, and 4 ply sheet thicknesses compared to controls across both cure conditions (*p<0.001 for low cure and #p<0.001 for high cure).

Incorporation of SIS particulates decreased the tensile modulus of high cure PGD (Figure 5). Representative images of low cure and high cure PGD curves show decreasing modulus and increasing strain at break (Figure 5a,b). Low cure PGD decreased in modulus with incorporation of 0.5wt% SIS particulates (Figure 5c, p<0.01). The modulus of high cure PGD was greater than the modulus of high cure PGD mixed with 0.1wt% (Figure 5c, p<0.01) or 0.5wt% particulates (Figure 5c, p<0.001). Similarly, there was an increase in strain at break with particulate incorporation. There was an increase in strain at break of low cure PGD with 0.1wt% (Figure 5d, p<0.01) and 0.5wt% (Figure 5d, p<0.001) compared to controls. There was also an increase in strain at break of high cure PGD with 0.5wt% particulates compared to controls and 0.1wt% particulates (Figure 5d, p<0.001). Finally, the medium cure PGD with 0.1wt% particulates exhibited greater strain at break than the high cure PGD at 0.1wt% particulates (p<0.01).

Figure 5.

Uniaxial tensile testing of low cure and high cure PGD and PGD-SIS particulate composites (mean ± SD, n=6) at 37°C in accordance with ASTM D412a. A) Representative stress-strain curve of low cure PGD with 0.1wt% and 0.5wt% SIS particulates incorporated in tensile dogbone specimens. B) Representative stress-strain curve of high cure PGD with 0.1wt% and 0.5wt% SIS particulates incorporated in tensile dogbone specimens. C) Tensile modulus of PGD with various particulate incorporation (#p<0.01 compared to controls for low cure, ^p<0.01 compared to controls for high cure, ^^p<0.001 compared to controls for high cure). D) Strain at break of PGD with various particulate incorporations (*p<0.01 compared to controls for low cure, **p<0.001 compared to controls for low cure, ‡p<0.001 compared to controls for high cure, and #p<0.01 for 0.1wt% low cure compared to high cure). Error bars represent standard deviation.

PGD with thinner SIS sheets (ie. fewer SIS plys) exhibited greater tear strength compared to PGD with higher sheet thickness (Figure 6a). Test specimens with 1 ply sheets had greater tear strength compared to PGD tear specimens prepared with 2 ply, 3 ply, and control specimens (Figure 6a, p<0.001). Tear strength of PGD with 2 ply sheets was greater than 3 ply (Figure 6a, p<0.01 for low cure and p<0.05 for high cure) and controls (Figure 6a, p<0.001). Incorporating 0.5wt% particulates into PGD low cure specimens reduced tear strength compared to 0.1wt% and controls (Figure 6b, p<0.001). Tear strength of low cure PGD with 0.5wt% SIS particulates was lower compared to high cure PGD (Figure 6b, p<0.05).

Figure 6.

Tear testing of PGD with SIS at 37°C following ASTM D624c (mean ± SD, n=6). A) Tear strength of PGD-SIS sheet tear specimens with 1 ply, 2 ply and 3 ply sheet thickness (*p<0.001 compared to control, ^p<0.001 compared to 1 ply, ##p<0.01 compared to 2 ply, and #p<0.05 compared to 2 ply). B) Tear strength of PGD-SIS particulate tear speciments with 0 wt%, 0.1wt% and 0.5wt% SIS particulates (#p<0.05, *p<0.001 compared to 0, and ^p<0.001 compared to 0.1wt%). Error bars represent standard deviation.

3.4. Scanning Electron Microscopy and Energy Dispersive X-ray

PGD tensile specimen fracture interface shows anisotropic crack propagation in response to uniaxial tensile loading (Figure 7a–d). PGD with embedded SIS sheets shows crack propagation orthogonal to embedded sheets (Figure 7e–h, indicated by white arrows). PGD with SIS particulates exhibit crack propagation along the polymer-particulate interface (Figure 7i–l, indicated by black arrows).

Figure 7.

Scanning electron microscopy (SEM) images of post-tensile testing fracture surfaces of high cure and low cure PGD with SIS sheets and particulates at 100× magnification (A-C, G-I) and at 1000x magnification (D-F, J-L). White scale bars are 100μm, and black scale bars are 10μm. White arrows indicate stress accumulation orthogonal to sheets, and black arrows indicate crack propagation at the interface of PGD and the incorporated particulates.

3.5. Shape Recovery Analysis

Incorporation of 2 ply and 4 ply sheets (Figure 8b) or 0.5wt% particulates (Figure 8e) into PGD decreased the shape fixity compared to high cure PGD alone (p<0.01). Recovery of PGD SIS sheets (Figure 8c) decreased with 2 ply and 4 ply incorporation across both high cure and low cure PGD compared to controls (p<0.001), whereas particulate incorporation (Figure 8f) has no significant effect on shape recovery. Incorporation of 2 ply SIS sheets decreased the recovery rate of low cure PGD compared to no SIS incorporation (Figure 8d, p<0.001) and 1 ply SIS (Figure 8d, p<0.05). Incorporation of particulates increased recovery rate of high cure PGD compared to no particulate incorporation and 0.5wt% particulates (Figure 8g, p<0.001). Incorporation of 0.1wt% particulates increased recovery rate of high cure PGD compared to low cure PGD. Multiple shape programming and recovery cycles did not affect the shape recovery rate (S4. a). A higher shape memory angle increased the shape recovery rate of high cure PGD (S4. b, p<0.05).

3.6. Nonlinear Model Fitting

Experimental tensile testing data was fit to Ogden models satisfying the Baker-Eriksen inequality (Equation 8). Representative images of high cure PGD with SIS particulates (S5. a) and low cure PGD with SIS particulates (S5. b) indicate the accuracy of model fitting. PGD and PGD with particulates were modeled accurately by the 1-term Ogden model (Table 2). Tensile testing data of PGD with 1 ply SIS sheets fit to a 1-term Ogden model exhibited higher μ₁ values compared to samples with no sheets and samples with 0.1wt% and 0.5wt% particulates (Table 2, p<0.001). Constitutive modeling of PGD with multiple SIS sheets required a 4-term Ogden model (Table 3). Resulting curves from a 4-term Ogden model fit to experimental tensile data of high cure PGD with SIS sheets (S5. c) and low cure PGD with SIS sheets (S5. d) indicate an accurate fit. The μ₁ values increased linearly from 1 ply to 4 ply across both low and high cure conditions. All PGD and PGD composite nonlinear elastic behavior was fit well with the Ogden model, with R² values ranging from 0.974 to 0.997 (S5; Table 2; Table 3). Parameter optimization of uniaxial tensile tests fit to FEBio finite element models of tensile specimens revealed similar mechanical properties compared to various soft tissues (Figure 9a,b). Incorporating particulates in high cure PGD can be used to tailor material properties to model Ogden 1-term nonlinear behavior of ventricular myocardium and meniscal attachments (Figure 9c). Incorporating SIS sheets in high cure and low cure PGD can be used to model mechanical properties of mitral valve chordae tendinea and the tibialis tendon (Figure 9e,f).

Table 2.

1-Term Ogden model of PGD incorporated with SIS sheets and particulates at different cure conditions.

High Cure
	μ1	α1	R2
0	4.43 ± 0.1	0.095 ± 0.01	0.997 ± 0.001
0.1 wt%	4.23 ± 0.4	0.1 ± 0.01	0.998 ± 0.043
0.5 wt%	3.63 ± 0.2	0.08 ± 0.01	0.997 ± 0.001
*1 ply	21.9 ± 9.7	0.158 ± 0.1	0.977 ± 0.003
Low Cure
	μ1	α1	R2
0	4.29 ± 0.16	0.111 ± 0.01	0.998 ± 0.001
0.1 wt%	3.72 ± 0.3	0.09 ± 0.02	0.998 ± 0.001
0.5 wt%	3.59 ± 0.1	0.07 ± 0.04	0.998 ± 0.001
*1 ply	20.1 ± 0.1	0.247 ± 0.13	0.974 ± 0.001

^*p < 0.001 compared to 0 wt%, 0.1 wt%, and 0.5 wt% particulate incorporation.

Table 3.

4-Term Ogden model of PGD incorporated with SIS sheets at different cure conditions.

High Cure
I	1 ply		2 ply		3 ply		4 ply
	μi	αi	μi	αi	μi	αi	μi	αi
1	2.78 ± 0.11	5.48 ± 0.02	5.93 ± 0.82	12.5 ± 1.89	6.74 ± 0.15	7.98 ± 4.32	13.3 ± 2.83	3.34 ± 10.28
2	−0.592 ± 0.06	1.94 ± 0.02	−4.75 ± 2.36	3.8 ± 1.00	−1.42 ± 4.51	2.81 ± 1.31	−2.01 ± 30.62	2.9 ± 0.77
3	−1.01 ± 0.05	7.82 ± 0.05	−1.56 ± 0.71	7.76 ± 0.6	0.07 ± 0.73	7.36 ± 2.49	−2.26 ± 8.15	−6.47 ± 16.68
4	−0.193 ± 0.08	9.5 ± 0.01	−2.18 ± 0.1	10.3 ± 6.19	−3 ± 0.84	11.6 ± 2.85	−6.06 ± 5.75	6.23 ± 18.65
R2	0.988 ± 0.006		0.996 ± 0.003		0.996 ± 0.005		0.997 ± 0.001
Low Cure
I	1 ply		2 ply		3 ply		4 ply
	μi	αi	μi	αi	μi	αi	μi	αi
1	2.18 ± 0.08	5.53 ± 0.1	5.38 ± 2.69	14.8 ± 13.47	17.7 ± 0.09	−0.922 ± 0.61	54.4 ± 0.45	2.33 ± 3.33
2	−0.96 ± 0.05	1.92 ± 0.17	−11 ± 15.36	1.85 ± 0	−41.9 ± 0.06	−2.56 ± 0.13	80.7 ± 0.1	−1.51 ± 2.11
3	−1.19 ± 0.19	7.54 ± 0.03	2.65 ± 4.78	10.83 ± 3.78	20.1 ± 0.01	−7.9 ± 0.02	98.5 ± 0.47	0.016 ± 0.04
4	0.12 ± 0.07	9.21 ± 0.33	4.94 ± 5.98	16.5 ± 9.93	−18.6 ±0.06	−3.6 ± 0.35	−0.16 ± 0.46	28.5 ± 0.2
R2	0.993 ± 0.004		0.997 ± 0.01		0.994 ± 0.005		0.994 ± 0.015

Discussion

Composites of PGD with SIS sheets and PGD with SIS particulates provide a unique platform to tune the bulk mechanical properties to match the nonlinear elastic properties of various soft tissues. While SIS sheets improve the overall stiffness and toughness of the material, SIS particulates increase elasticity. Raman spectra indicated uniform molecular structure in PGD with SIS sheets. The shift from 1640 to 1727 indicates esterification and decrease of free carboxylic acid groups which was present in both proximal and distal regions of sampled polymer(Grasselli et al., 1980; Maliger et al., 2013). Additionally, the remaining aliphatic, carboxyl, and hydroxyl stretches indicate no significant differences between proximal and distal sampling regions, reflecting uniform molecular structure through each specimen (S2. c). Amongst sheet and particulate groups, SIS inclusions concomitantly affect intrinsic material properties that impact shape recovery properties, and can in turn impact in vivo performance. The increased mechanical tunability in the context of material properties can provide boundary conditions for the design of implantable devices for soft tissue repair. Additionally, incorporation within PGD can improve surgical handling of various SIS mediated repair applications in MIS procedures.

4.1. Swelling ratio

When placed in organic solvents, the degree of swelling of thermoset elastomers such as PGD is inversely proportional to the crosslink density. As expected, there was greater swelling in low cure PGD compared to high cured PGD. Incorporation of sheets within the polymer matrix added to this difference in crosslink density between high cure PGD and low cure PGD. This could be a result of greater crosslink density in high cure PGD when SIS is incorporated into the polymer matrix. The differential influence of sheet thickness indicates thermal or physical interference with crosslinking mechanisms. The 2 ply sheet may be forcing the polymer matrix into a smaller volume space within the Teflon rectangular mold, thereby increasing the degree of crosslinks. The incorporation of particulates within the PGD polymer matrix decreased the number of crosslinks for low cure condition. Particulate incorporation of 0.5wt% into polymer matrix subsequently increased the swelling ratio. The particulates may also be interfering with crosslinking dynamics by physically separating the polymer chains and limiting formation of crosslinks.

4.2. Differential Scanning Calorimetry

The melting transition temperature (T_m) and enthalpy of fusion (ΔH_fus) indicate changes in polymer crosslinking density and crystallinity respectively. Both structural characteristics give rise to the overall mechanical properties of the polymer. As expected, we see a greater number of crosslinks and a lower peak T_m in high cure PGD controls compared to low cure PGD controls. Overall, we also see a lower peak T_m in high cure PGD with SIS sheets, suggestive of greater crosslinking than low cure PGD with sheets. These trends in crosslink density between high and low cure PGD were also reflected by the swelling assay as previously discussed (Figure 2a).

Interestingly, we also see lower transition temperatures and consequently higher number of crosslinks in high cure PGD with sheets compared to controls, which is also reflected by the swelling data. This may also be explained by the presence of a continuous sheet in the polymer matrix causing differences in thermal diffusivity in the polymer during the curing process. These changes in crosslink density within PGD-SIS sheet composite materials are suggestive of more favorable thermal conductivity within SIS compared to PGD alone. Since thermal conductivity of organic materials beyond 55°C increases nonlinearly at higher temperatures, SIS in the high cure condition may be far more conductive of heat than at the low cure conditions(Bhattacharya and Mahajan, 2003). Moreover, since the fraction of PGD within samples containing sheets is much lower, there may be better heat transfer in these samples through the SIS sheets overall, driving the increased crosslink density in high cure PGD sheets compared to controls and low cure PGD sheets. Additionally, we do not see these same improvements in thermal conductivity in low cure PGD because of the lower heating rate and longer cure time. Low cure PGD with SIS particulates suggests a high crosslink density compared to controls, which is not supported by the swelling data. However, the presence of voids throughout the cross section of the polymer (S3. b,e) explains the discrepancy between the measured crosslinking density and the swelling ratios.

The overall trends in relative crystallinity between low cure and high cure PGD were largely expected, with a few notable discoveries. As a polymer becomes more crosslinked, a drop in crystallinity is expected due to the reduction in chain mobility and entropy. This is reflected by the generally higher relative crystallinity of high cure PGD than low cure PGD across all variations. Although crosslink density of low cure PGD with SIS sheets remained relatively unchanged, the crystallinity of the composite materials was far greater than controls, suggesting more favorable organization of the polymer chains into lamellar structures in these composites. This may be due to the reduced volume of the polymers requiring more efficient packing or an SIS-mediated organization. The discrepancy between the higher crosslink density of low cure PGD polymer with SIS sheets and higher crystallinity of the polymer may be explained by the notion that the particulates perhaps facilitate the formation and aggregation of lamellar structures. These changes indicate a softer and tougher material for various tissue engineering applications.

4.3. Mechanical testing

SIS sheets increase overall stiffness and tear resistance across cure conditions compared to controls. The linear increase in stiffness indicates the greater contribution of SIS sheets to stiffness compared to PGD with increasing sheet thickness. However, increasing sheet thickness decreased tear resistance, likely driven by a greater number of voids between thicker sheets and the surrounding PGD material compared to thinner sheets. Particulate incorporation decreased stiffness of both cure conditions in large part due to void spaces evident in the SEM micrographs of PGD-SIS particulate tensile specimen cross sections (S3). These void spaces also contributed to the nonlinear elastic behavior of the tensile specimens while reducing the tear resistance by providing more crack initiation sites. The high cure condition abrogated the effect of particulate incorporation on tear resistance, but not overall tensile strength.

4.4. Scanning Electron Microscopy

Chain length, configuration, and architecture contribute to overall failure mechanics of all polymers. Additionally, when incorporating particulates and sheets, the expected differences in overall material architecture arise from differences at the filler polymer interface that result in different mechanical properties of the polymer. Taken together with the swelling data, incorporation of sheets and particulates affects the crosslink density and void spaces, consequently impacting overall architecture of the polymer matrix. The incorporation of SIS may modulate heat flow through the polymer, which is consequently affecting the crosslink density. However, the increases in tear strength and stiffness are driven by the incorporation of SIS sheets, where tearing of the sheet is observed at the failure interface, but lines orthogonal to the sheet indicate ductile failure of the polymer matrix. Similarly, the incorporation of particulates decreased crosslink density and overall polymer mechanical properties. The decreased mechanical properties could be attributed to increased distribution of void spaces throughout the polymer matrix at the particulate-polymer interface. The crack propagation consequently appears to be taking place along this interface.

4.5. Shape Recovery

Incorporating SIS sheets and particulates affected the shape fixity and shape recovery differently across cure conditions. Higher cured PGD with sheets has lower fixity and recovery rate compared to controls because of the effect of incorporating sheets with varying crosslink densities. Shape recovery rate is directly related to the crosslink density of the material. Trends in the shape recovery rate of PGD composite materials inversely correlate with the swelling ratio of composite materials. Increasing the sheet thickness increases the swelling ratio of PGD-SIS sheet samples, which relates to decreasing crosslink density. Consequently, we see a reduction in shape recovery rate of 2 ply, 3 ply, and 4 ply sheets across both cure conditions. Similarly, 0.5wt% particulate incorporation in low cure PGD, which had the highest swelling ratio exhibited the slowest recovery rate. There were also similar recovery rate differences between low and high cure conditions in the 0.1wt% particulate group, indicting particulate distribution in conjunction with thermal effects to modulate crosslink density of the material. The significant increase in crosslink density of the high cure 0.1wt% particulate materials compared to controls also indicate an improvement of crosslink density in line with particulate incorporation into PGD. Taken together with the mechanical testing data, particulate incorporation and thermal curing can independently control mechanical and shape recovery properties of the composite material, providing additional control parameters for the design of shape memory biomedical implantable devices comprised of PGD-SIS composite materials.

4.6. Nonlinear constitutive modeling of PGD-SIS composites

Soft biological tissues, are predominantly anisotropic and withstand non-uniform multi-axial extensional and compressive loads. Various nonlinear models have been used to describe mechanical properties of soft tissue deformation(Morales-Orcajo et al., 2017, 2016). Uniaxial tensile properties of PGD and PGD-SIS particulates demonstrated nonlinear behavior similar to meniscal attachments and ventricular myocardium (Figure 9c,d). Utilizing SIS particulates in the design of these materials could further tailor the mechanical properties of these composite materials to match intrinsic tissue material properties. Incorporating SIS sheets resulted in stiffer material that resembled tissue material properties of tendons and structural heart tissues (Figure 9e,f). Tendons, like many musculoskeletal tissues, are hierarchically organized bundles of fibers primarily comprised of fibrillar collagen. In the epitenon comprised of tendon fasciles, heterotrimeric collagen ColI organizes into fibrils and combines with tendon fibroblasts and surrounding proteoglycans. Modulation of directionally dependent crosslinking and concentration of proteoglycans by tendon fibroblasts gives rise to nonlinear anisotropic mechanical properties dependent on tissue loCation. Incorporation of sheets within PGD matrix combined with thermal crosslink variants could further mimic the mechanical contributions of the fibers and the crosslinking of the surrounding matrix. Similarly, myocardium primarily comprised of ColI fibers (about 80–90%) and ColIII fibers contains an entirely different architecture surrounding cardiomyocytes. Myocardium fiber organization also exhibits transmural anisotropy with longitudinal alignment in the endocardial surface rotating to a circumferential alignment at the epicardial surface. PGD-incorporated SIS particulates demonstrates nonlinear uniaxial tensile properties similar to myocardium (Figure 9c), and can be arranged along various axes to mimic aniosotropy of cardiac myofibers. Incorporating SIS sheets into PGD also improved tear resistance of the material. These findings suggest a potential role of PGD-SIS composite materials to meet the mechanical requirements across a range of soft tissue repair applications requiring minimally invasive repair using biodegradable implantable devices.

5. Conclusions

Poly-(glycerol dodecanedioate) can be cured under various conditions and combined with SIS sheets and particulates to create composite materials with a range of mechanical properties for many soft tissue reconstruction applications. In addition to matching tensile moduli of various soft tissues, the nonlinear elastic behavior can also closely mimic various extracellular matrix architectures of soft tissues, which suggests control of mechanical performance when used in implantable devices. In addition to tenability of mechanical properties, various formulations demonstrate preservation of the shape memory effect, which is critical for device delivery and surgical handling. Moreover, by incorporating SIS into the bioresorbable polymer, we can provide the structural and biological advantages of a biological extracellular matrix and impart finer control of resorption rates, which may mitigate reported chronic immune responses associated with SIS.

It is also important to evaluate how the degradation rate impacts polymer network structure, shape memory behavior, and mechanical performance in the context of physiologically releveant in vivo models. Our prior studies investigating the degradation behavior of PGD revealed in vitro hydrolytic degradation of PGD with an initial mass loss of 20% after 2 weeks in PBS, with another 20% after 8 weeks and an additional 30–50% degradation after 18 weeks(Ramaraju et al., 2020). A similar degradation rate is observed in vivo high cure PGD exhibits nearly 50% mass loss after 16 weeks in a mouse subcutaneous implant. We see faster degradation of amorphous regions in high cure PGD compared to low cure PGD in vivo over 1 month and faster degradation of crystalline regions in high cured PGD compared to low cured PGD after 4 months.

By incorporating SIS sheets and particulates within the polymer, we demonstrated changes in relative crystallinity and crosslink density of the overall polymer network(Figure 3, Table 1). Moreover, by incorporating particulates within the polymer, there is a visible increase in surface roughness (Figures 7F, S3C) absent in the control polymers or polymers with embedded sheets. We anticipate these differences in polymer network and surface morphology to differentially impact hydrolytic and enzymatic degradation of the polymer which will need to be methodically investigated for each intended clinical application using relevant in vitro and physiologically relevant in vivo models. Future studies will include biaxial evaluation of PGD-SIS and mimicking the inherent anisotropy in various biological soft tissues.