Sterilization Effects on Poly(glycerol dodecanedioate), a Biodegradable Shape Memory Elastomer for Biomedical Applications

Abstract

Biodegradable shape memory polymers provide unique regenerative medicine approaches in minimally invasive surgeries. Once heated, thermally responsive shape memory polymer devices can be compressed, programmed to fit within a small profile, delivered in the cold programmed state, and expanded when heated to body temperature. We have previously developed a biodegradable shape memory elastomer (SME), poly (glycerol dodecanedioate) (PGD), with transition temperatures near 37°C exhibiting nonlinear elastic properties like numerous soft tissues. Using SMEs in the clinic requires disinfection and sterilization methods that conserve physiochemical, thermomechanical, and shape recovery properties. We evaluated disinfection protocols using 70% ethanol and UV254nm for research applications and ethylene oxide (EtO) gas sterilization for clinical applications. Samples disinfected with ethanol for 0.5 and 1 minute showed no changes in physiochemical material properties, but after 15 minutes showed slower recovery rates than controls (p <0.05). EtO sterilization at 54.4°C decreased transition temperatures and shape recovery rate compared to EtO sterilization at 37.8°C (p < 0.01) and controls (p < 0.05). Aging samples for nine months in a vacuum desiccator significantly reduced shape recovery, and the recovery rate in EtO sterilized samples compared to controls (p <0.001). Cytotoxicity testing (ISO-10993.5C:2012) revealed media extractions from EtO sterilized samples, sterilized at 37.8°C, and high-density polyethylene (HDPE) negative control samples exhibit lower cytotoxicity (IC50) than Ethanol 1 min, UV 2hr, and EtO 54.4°C. Cell viability of NIH3T3 fibroblasts on sterilized surfaces was equivalent on EtO 37.7°C, EtO 54.4°C and Ethanol sterilized substrates. Finally, chromogenic bacterial endotoxin testing showed endotoxin levels were below the FDA prescribed levels for devices contacting blood and lymphatic tissues for ethanol 1min, UV 120min, EtO 37.7°C, EtO 54.4°C. These findings outline various disinfection and sterilization processes for research and pre-clinical application and provide a pathway for developing custom sterilization cycles for the translation of biomedical devices utilizing PGD shape memory polymers.

Introduction

Shape memory polymers(SMPs) are an emerging class of biomaterials with the potential to address numerous clinical challenges^(1)–(4). Biomedical SMPs often rely on a thermal transition, usually at or below 37°C, allowing recovery from a programmed delivery shape to an original permanent shape when deployed in vivo. These polymers are well suited for use in minimally invasive surgeries(MIS) requiring delivery of compressed geometries through catheter sheaths or trocars prior to full expansion in vivo^(5)–(8). Minimally invasive procedures are increasingly used across surgical disciplines driven by reduced patient trauma, reduced length of stay, adverse events, recovery time, and hospitalization costs^(9)–(12). Recently, percutaneously delivered polyurethane foams for aneurysm occlusion gained FDA approval as the first SMP device on the market and numerous permanent and biodegradable products in the pipeline^(5)–(8).

Biodegradable shape memory polymers (bSMP) are gaining prominence for their potential to address numerous tissue engineering and regenerative medicine challenges^(2),(4). While also suited for delivery in MIS, bSMP can provide structural support and tissue growth upon implantation, subsequently degrading into non-toxic byproducts and ultimately replaced by patient tissues. We developed a biodegradable shape memory polymer, poly(glycerol dodecanedioate) (PGD), and tailored the physiochemical properties to transition at or below 37°C⁽¹³⁾. This polymer exhibits nonlinear elastic properties similar to various soft tissues when placed in an environment above the transition temperature^(13),(14). Polymer mechanical properties are tunable to model surrounding soft tissues, and degradation rates are appropriate for numerous tissue repair applications^(14)–(16). Effective translation of this material for tissue engineering requires the development of suitable aseptic disinfection procedures and industrial-grade sterilization methods.

Early material research and product development require rapid aseptic processes that inactivate bacterial proteins and microbes bound to polymers. Aseptic procedures commonly employ immersion in 70% ethanol or UV irradiation at 254nm. Immersion in 70% ethanol effectively ruptures bacterial cell walls, denaturing proteins, and neutralizes fungi and viruses. Duration of exposure to ethanol impacts the morphology, molecular weight(MW), and thermal properties of various medical polyesters⁽¹⁷⁾. Duration of exposure to UV 254nm irradiation can also effectively disinfect the material while preserving the morphology. UV irradiation reduces MW, while ethanol immersion changes morphology and device geometry^(18),(19). Although both processes damage bacterial, microbial, and viral organisms, they are mainly ineffective against spores. Complete inactivation requires high temperatures and steam or gas sterilization processes.

The actuation temperatures of PGD present unique challenges for clinical-grade sterilization. Medical devices intended for surgical implantation require aseptic processing and sterilization prior to human use (FDA 56 Code of Federal Regulations - CFR 51354). Guidance on medical device manufacture and Pre-Market Approval (PMA) or Pre-Market Notification (510k) requires validated sterilization processes and associated documentation of established sterilization methods utilizing steam, irradiation, or chemical gas sterilization^(19),(20). Amongst these processes, ethylene oxide(EtO) gas sterilization and gamma irradiation makeup nearly 50% and 40% of sterilized devices, respectively⁽²¹⁾. Concerns using high temperature or radiation-based sterilization methods on polymers center on how sterilization processes alter polymer morphology, molecular architecture, and ultimately the device’s intended function. Many polymers used in biomedical applications have melting temperatures (T_m) between 60–120°C and require sterilization processes below these temperatures. Additionally, most polymers used in medical devices contain polyesters that undergo hydrolytic degradation when subjected to high heat and steam sterilization^{(17),(18),(21)–(25)}. Moreover, gamma irradiation of polymer causes polyester chain scission resulting in reduced MW. For example, gamma irradiation can reduce MWand tensile strength of polycaprolactone ⁽²²⁾.

Sterilization processes utilizing high heat and moisture also hydrolyze ester crosslinks in PGD responsible for actuation temperatures. Reduced PGD crosslinking shifts transition temperatures and increases crystallinity impacting the mechanical properties, shape recovery, and shape fixity^(15),(16). Traditional electron beam and gamma sterilization cleave polymer chains, impacting MW and architecture leading to changes in mechanical and shape recovery properties, including transition temperature (T_trans) and recovery rate (dRr/dt) ^{(21),(26)–(28)}. EtO at low temperatures and humidity poses a tenable approach for PGD sterilization, suggesting no impact on MW, crystallinity, or mechanical properties of polyesters^{(19),(21),(26),(29)–(31)}. Although EtO cycle refinement and process control for devices utilizing PGD will be specific to each application, identifying available research and commercial grade sterilization processes can streamline the overall material and device development process.

This study investigates the impact of the duration of exposure to 70% ethanol and UV 254nm on the physiochemical, thermomechanical, and biological properties of research-grade disinfected PGD (Figure 1). Furthermore, this study evaluates two EtO sterilization cycles and their impact on physiochemical, thermomechanical, and biological properties of PGD (Figure 1) along with shelf-life considerations germane to commercialization.

Figure 1.

Overview of experimental methods used to evaluate the composition, morphology, thermomechanics, and biocompatibility of research and commercial grade sterilization of PGD.

Disinfection and sterilization processes impacting physiochemical and thermomechanical properties such as roughness and stiffness can also affect the biological properties of the material. Additionally, residual sterilization agents entrapped within the polymer also impact cytotoxicity of the sterilized polymer. Addressing the toxicity of soluble products is a critical component of biocompatibility testing guidance (ISO10993:5C). In addition to verifying the toxicity of soluble products, it is also important to assess cell viability in direct contact with sterilized substrates. Finally, bacterial endotoxins within medical devices are also a concern since only high-temperature processes unsuitable for polymer sterilization can remove these contaminants. Manufacturing processes that limit initial contamination can prevent the incorporation of these endotoxins.

Finally, the practical challenge of long-term storage of bSMP devices was considered by naturally aging sterilized materials in vacuum desiccators for 4 and 9 months. The shelf-life of bMSP devices is a crucial factor impacting commercialization because it impacts both manufacturing and distribution. If sterilization impacts the physiochemical or thermomechanical properties of the polymer, sterilized polymers can consequently age differently based on the method of sterilization⁽³⁰⁾. Sterilized shape memory polyurethane foams increased glass transition temperature, reduced fixity, and recovery^(32),(33). As sterilized polyesters age, they undergo hydrolytic degradation from ambient humidity resulting in reduced MW, increased crystallinity, reduced ductility, and toughness⁽³⁰⁾. Consequently, this study also investigates the impact of aging sterilized material on the physiochemical, thermomechanical, and shape memory properties to identify the upper limit of shelf life for clinical use.

Materials and Methods

Poly (glycerol dodecanedioate) synthesis and preparation

Polyglycerol dodecanedioate was synthesized as previously described^(13),(14). 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. Pre-polymer was subsequently cooled and stored in a vacuum desiccator until further use. Sheets of 2 mm thick PGD were prepared by heating the pre-polymer to 90°C and pouring it into a rectangular silicone mold. All molds were subsequently cured for 72 h at 120°C under vacuum at 90 mTorr. Specimens were removed from the mold, cooled, dried, and stored in a vacuum desiccator until they were laser cut to 8 mm discs for rheometry, 6 mm discs for Fourier-Transform Infrared Spectroscopy(FTIR) and scanning electron microscopy(SEM), 4mm discs for Differential Scanning Calorimetry(DSC), 30mm × 5 mm rectangular bars for shape recovery, 25 mm × 5 mm rectangular bars for cytotoxicity testing.

Sterilization of Shape Memory Polymers

Samples were sterilized under UV254nm in a Class II A2 biosafety cabinet for 3 and 120 minutes as previously described⁽³⁴⁾. Samples were sterilized in 70% ethanol for 0.5, 1, and 15 minutes at 25°C and washed in deionized water for 1 min three times. Samples were shipped to Case Western Reserve University Hospitals to run two different EtO sterilization cycles. Samples were sterilized in 100% EtO gas at 37.8°C for 4 hours and outgassed for 48 hours or sterilized for 2 hours at 55.4°C and outgassed for 24 hours. Samples and their respective controls are provided in Table 1.

Table 1:

Experimental groups for the investigation of research and commercial grade sterilization of PGD

Research Grade Disinfection
Control	Untreated
UV254nm3min	UV irradiation at 254nm for 3 minutes
UV254nm120min	UV irradiation at 254nm for 120 minutes
Ethanol 0.5min	Immersion in 70% ethanol for 0.5 minutes
Ethanol 1min	Immersion in 70% ethanol for 1 minute
Ethanol 15min	Immersion in 70% ethanol for 15 minutes
Commercial Grade Sterilization
Transport Control	Not sterilized but transported and stored alongside treated samples
EtO 37.8	Sterilized using ethylene oxide at 37.8°C
EtO 54.4	Sterilized using ethylene oxide at 54.4°C
HDPE	High density polyethylene(HDPE) standard used in ISO10993:5C cytotoxicity assays

Surface characterization of PGD samples

Post-sterilization polymer samples were dried in vacuum desiccators overnight and mounted to SEM sample holders using adhesive tape. Ethanol, UV, and EtO sterilized samples were sputter-coated with a thin gold film for 60 s using a Quorum Q-150T ES (Quorum Technologies, East Sussex, England) and imaged using a LEO 1530 SEM (Carl Zeiss AG, Oberkochen, Germany‎) at 6 kV acceleration voltage. Four images were taken at two different magnifications of each sample. FTIR was performed on Ethanol, UV, and EtO sterilized sample using a Shimadzu Prestige 21 Infrared Spectrometer(Shimadzu Kyoto, Japan) with 16 accumulated scans per spectrum at 4 cm⁻¹. Spectra were collected from 3 regions on each sample (n=3). Representative spectra from each group are presented with peak analysis.

Thermomechanical Evaluation of PGD samples

According to the Mather classification of SMPs, a Class II chemically crosslinked elastomer exhibits transition temperatures corresponding with the melt transitions. Differential Scanning Calorimetry (DSC) was conducted using a Discovery Q250 with an RCS90 cooling system (TA instruments, New Castle DE) to quantify changes in thermal transitions within polymers subjected to different sterilization and disinfection processes. Samples were dried overnight in a vacuum desiccator, weighed (5–7 mg), and placed in a Tzero^® pan. Samples (n=3/group) were heated to 90°C from 25°C at a rate of 20°C/min to remove thermal history. After an isothermal hold for 3 min, samples were cooled at a rate of 10°C/min to −50°C, and after an isothermal hold for 3 min, heated up to 70°C at a rate of 5°C/min. The reported transition temperatures Tm and Tc and enthalpy ΔH_fusion were measured and calculated using the TRIOS software version 4.1 (TA Instruments). Rheometry was conducted using an MCR 302 rotational rheometer (Anton Paar GmbH, Graz, Austria) equipped with a PPTD-200 and an 8 mm parallel plate sandblasted measuring set. An oscillatory rheometer was used to measure changes in shear modulus as a function of temperature, indicating the onset and completion of the physical transition from plastic to an elastomer which results in a two-fold drop in shear modulus. A temperature sweep was run on PGD samples (8 mm diameter × 2 mm thickness, n=3) from −10°C to 80°C at a constant rate of 1°C/min with a constant angular frequency (ϖ = 10 rad/s) and a constant strain (5%). The storage modulus(G’), complex shear modulus G* and loss factor (tan δ) were evaluated using RheoCompass version 1.2 (Anton Paar) and the transition temperatures (T trans) were calculated at the inflection point (tan δ’). The storage modulus (G’) was used to infer relative changes in crosslink density across treatment groups^(35),(36).

Shape recovery assessment

Shape fixity and shape recovery were calculated using Equations (2) and (3) respectively, as previously described but adapted to this angular recovery test^(1),(15).

$$R_f=\frac{{\theta }_{\text{u}}\left(\text{N}\right)}{{\theta }_{\text{m}}}\times 100$$ (2)

$$R_r=\frac{{\theta }_m-{\theta }_p\left(N\right)}{{\theta }_m-{\theta }_p(N-1)}\times 100$$ (3)

Equation 2 was used to determine shape fixity and Equation 3 was used to determine shape recovery, where θ_m is the intended programming angle resulting from mechanical force application applied to the transitioned specimen, θ_u (N) is the actual programming angle of the temporary shape after the specimen has been cooled and the force has been removed, θ_m− θ_p (N) is the change in the angle during the course of recovery, and θ_m− θ_p (N−1) is the change in the angle during the course of programming. Recovery rate (degrees/min) was also calculated by the following equation and subsequently converted to radians/min.

$$\frac{{dR}_r}{dt}=\frac{{\theta }_m-{\theta }_p\left(N\right)}{\Delta t}$$ (4)

Cytotoxicity testing - ISO-10993.5C

The ISO10993 test methodology was used to evaluate cytotoxicity in the research lab setting. NIH3T3 fibroblasts (ATCC) were cultured to 80–90% confluency in complete media Dulbecco’s Modified Eagle Media containing 10% fetal bovine serum and 1% Penicillin/Streptomycin. The cytotoxicity of PGD was assessed according to the ISO10993.5C 2012 standard for cytotoxicity testing detailed in section 5 and Appendix C using an MTT assay. Cytotoxicity samples (25 × 5 × 2 mm, n=3/group were extracted for at 37°C for 48 h in Dulbecco’s Modified Eagle Medium (DMEM) with 10% Fetal Bovine Serum (FBS) and 1% penicillin-streptomycin at an extraction ratio of 3cm²/mL. Cells were plated at a density of 10,000 cells/well in Corning 96 well tissue culture treated microplates. Media containing material extractions was diluted (100% −0.1%) and each dilution was added in triplicate to cells in the 96well plate. NIH3T3 fibroblasts were incubated for 24 h at 37°C at 5% CO₂. Media was aspirated from wells, and MTT reagent with phosphate-buffered saline (PBS) and phenol red-free DMEM was added to wells according to the manufacturer’s instructions. Cells were incubated at 5% CO₂ for 2 hours. The solution was aspirated, and isopropanol with 0.04 M HCl was added to each well. Absorbances were obtained using a plate reader at 570 nm and 650 nm. Nonlinear curve fitting to determine the half-maximal cell growth inhibitory concertation (IC₅₀) using Graphpad Prism 9. Experiments were conducted in triplicate for statistical analysis of IC₅₀ values.

Direct cytotoxicity testing Live-dead assay

PGD samples (8mm diameter × 2mm height) treated with ethanol or sterilized using EtO were incubated in complete media overnight in non-treated tissue culture plates. NIH3T3 fibroblasts were seeded at a density of 10,000 cells/cm² onto PGD discs. A live-dead assay (Thermofisher Scientific) was conducted after 24 hours of cell adhesion to substrates following manufacturer protocols. A Zeiss 700 Laser Scanning Confocal microscope was used to image calcein, and ethidium bromide labeled cells on PGD substrates. A total of 4 fields of view were collected for each sample, and Image J particle analysis plugin was used to count live cells labeled with calcein and dead cells labeled with ethidium bromide.

Quantitative endotoxin testing

A chromogenic LAL endotoxin assay kit was used to quantify endotoxin units (GenScript Biotech Corp Piscataway, NJ). Samples (25 mm × 5 mm × 2mm) were placed in endotoxin-free glass vials and extracted in non-pyrogenic water overnight at 37°C on a shaker at 200RPM in accordance with manufacturer recommendation. Standard curve was constructed using serial dilutions of supplied endotoxin stock at (0.1, 0.05, 0.025, 0.0125 and 0 endotoxin units/mL). Extracted solutions and standards were combined with LAL reagent for 30 minutes, incubated in a chromogenic substrate for 6 minutes, and the reaction was subsequently stopped, colorimetric stabilizers were added, and absorbance was measured at 545nm.

Statistical Analysis

Statistical analyses were conducted using Prism 9 (GraphPad Software Inc., San Diego CA). JMP 13.1 (SAS Inc., Atlanta GA). One-way ANOVA using Bonferonni post-hoc tests for all pairwise comparisons was conducted to analyze transition temperatures, relative crystallinity, fixity, recovery, recovery rate. Two-way ANOVA using Tukey post-hoc tests for all comparisons was conducted to analyze the impact of aging and sterilization on the material’s thermomechanical and shape memory properties.

Results

Physiochemical and thermomechanical effects of UV sterilization

PGD samples sterilized with UV at 254 nm for 120 min exhibited visibly different surface morphologies than samples sterilized with UV treatment for 3 min. PGD surfaces appeared smoother due to UV sterilization compared to control surfaces (Figure 2A). The surface chemistry of the PGD polymer evaluated by FTIR was unchanged by the UV sterilization process (Figure 2B). All samples had the characteristic alkene stretch at 2918 cm⁻¹ and 2850 cm⁻¹, the carbonyl stretch at 1737 cm⁻¹, the alkoxy stretch at 1167 cm⁻¹, and the alcohol stretch at 3466 cm⁻¹. There were no additional peaks in the region of the alcohol stretch or the fingerprint region of the FTIR spectra (Figure 2B). DSC analysis revealed no significant differences between the peak Tm or Tc, and the representative curves did not demonstrate any drift between controls and UV-treated samples (Table 2 and Figure 2C). Oscillatory temperature sweeps using the rheometer revealed that transition temperatures did not vary across treatment groups. Complex shear modulus(G*) and the storage modulus (G’) above the transition temperature of UV treated samples were not different from controls (Table 2 and Figure 2D). Recovery rates of the UV sterilized polymer were also not significantly different from controls.

Figure 2.

Physiochemical effects of UV254nm treatment at 25oC on PGD. A) SEM images of UV sterilized samples. Blue bars 200μm, red bars 10μm B)FTIR spectra of UV sterilized samples C) DSC thermogram of the melt transition(Tm) D) Rheometry temperature sweep evaluating shear modulus.

Table 2.

Physiochemical evaluation of UV₂₅₄ and Ethanol_75% treated PGD samples at 25°C

	Control	UV254nm			Ethanol70%
		2 min	120 min	0.5 min	1 min	15min
DSC
Tc (°C)	17.4 ±1.7	16.2 ± 0.3	16.88 ± 1.6	19.1 ± 2.5	17.1 ± 1.5	19.2 ± 2.1
Tm - Ttrans (°C)	34.7 ±1.2	34.4 ± 0.6	34.4 ± 0.5	35.9 ± 1.3	34.4 ± 0.7	35.2 ± 0.6
ΔHfusion (J/g)	45.3 ± 3.5	41.3 ± 1.62	55.0 ± 15.1	42.6 ± 3.2	39.9 ± 2.9	40.3 ± 2.1
Rheometry
tan δ’ - Ttrans (°C)	36.2 ± 6.7	37.2 ± 1.6	37.9 ± 1.3	35.7 ± 2.1	37.9 ± 1.3	36.3 ± 0.5
G* (MPa)	0.5 ± 0.1	0.38 ± 0.2	0.32 ± 0.11	0.3 ± 0	0.5 ± 0.1	0.4 ± 0.1
G’ (MPa)	0.53 ± 0.07	0.44 ± 0.06	0.40 ± 0.08	0.41 ± 0.03	0.47 ± 0.09	0.28 ± 0.12*
Shape Recovery
$R_f$%	97.2 ± 1.6	98.8 ± 0.6	97.4 ± 1.4	95.1±3.2	96.5±1.7	97.6 ± 1.7
$R_r$ %	97.2 ± 1.8	100.0 ± 0	100.0 ± 0.0	97.2±1.8	99.1±1.6	87.0 ± 4.2^
$\frac{{dR}_r}{dt}$(rad/min)	4.6 ± 0.6	4.4 ± 1.5	3.8 ± 0.5	1.8±0.5	2.7±0.6	1.6 ± 0.8#

^*(p < 0.05) significant difference from Control,

^{^}(p < 0.05) significant difference from 1 min Ethanol_70%,

^#(p < 0.05) significant difference from Controls

Physiochemical and thermomechanical effects of ethanol sterilization

PGD immersed in 70% ethanol exhibited differences in surface characteristics compared to untreated controls at the 15 min time point (Figure 3A). Surface morphology became smoother, suggesting some erosion at the 15 min time point (Figure 3A). Evaluation of surface chemistry by FTIR (Figure 3B) indicated characteristic vinylic hydrogen stretch at 2918 cm⁻¹ and 2850 cm⁻¹, the carbonyl stretch at 1737 cm⁻¹, the alkoxy stretch at 1167 cm⁻¹, and the alcohol stretch at 3466 cm⁻¹. However, samples immersed in ethanol for 15 min exhibited some notable differences in the fingerprint region, indicated by a peak at 806 cm⁻¹, suggesting a difference in the alkene sp2 C-H bend. The alcohol stretch was largely unaffected by ethanol immersion, indicating overall preservation of surface hydroxyl groups regardless of the duration of sterilization. Thermal analysis using DSC showed T_m of PGD post-ethanol sterilization was like controls, indicating preservation of the intrinsic material properties arising from the conservation of molecular architecture (Figure 3C). There was no significant difference in transition temperatures or shear modulus across the different oscillatory temperature sweeps on the rheometer. Storage modulus revealed lower values indicative of lower density of crosslinks compared to control group (Table 2, p < 0.05). PGD samples immersed in ethanol for 15 min exhibit reduced recovery and recovery rate compared to controls (Table 2, p < 0.05).

Figure 3.

Physiochemical effects of ethanol treatment at 25oC on PGD A) SEM images of ethanol sterilized samples blue bars 200μm, red bars 10μm B) FTIR spectra of ethanol sterilized samples C) DSC thermogram of the melt transition(Tm) D) Rheometry temperature sweep evaluating shear modulus.

Physiochemical and thermomechanical effects of EtO

Like ethanol sterilization, EtO sterilization at high temperatures (Figure 4A) resulted in smoother surfaces than transport controls (Figure 4A). Evaluation of the surface chemistry by FTIR revealed no significant differences in eEtO sterilized samples compared to transport control (Figure 4B). There were slight shifts in the fingerprint region compared to controls between 950–700 cm⁻¹, indicating similar conformational differences to ethanol sterilization in the sp2 C-H bend (Figure 4B). Melt temperature (T_m) evaluated by DSC of EtO sterilized samples at 54.4°C revealed a downshift in transition temperatures (Figure 4C and Table 3). Thermomechanical transition temperatures evaluated by peak tan δ also revealed significant differences between control and EtO sterilized samples at higher temperatures (Figure 4D and Table 3). Shear modulus (G*) above the transition temperature did not reveal any significant differences across all samples (Table 3). Storage modulus in the rubber plateau above the transition was lower than controls but not statistically significant. Although fixity and overall recovery remained unaffected compared to controls, the shape recovery rate of EtO sterilized samples at 54.4°C was slower than samples sterilized at 37.8°C (Table 3, p < 0.05).

Figure 4.

Physiochemical effects of ethylene oxide sterilization on PGD A)SEM images of EtO sterilized samples blue bars 200μm, red bars 10μm B) FTIR spectra of untreated negative controls, transport controls that traveled with the EtO treated samples, and EtO treated samples sterilized at 37.8oC and 54.4oC C) DSC thermogram of the melt transition (Tm) D) Rheometry temperature sweep evaluating shear modulus.

Table 3.

Physiochemical evaluation of Ethylene oxide treated PGD samples

	Transport Control	Ethylene oxide
		37.8°C	54.4°C
DSC
Tc(°C)	19.3 ± 1.5	17.7 ± 1.4	18.3 ± 0.7
Tm = Ttrans (°C)	34.7 ± 1.2	34.2 ± 1.2	34.2 ± 2.4
ΔHfusion (J/g)	45.3 ± 3.5	47.5 ± 4.4	46.7 ± 5.7
Rheometry
tan δ’= Ttrans (°C)	37.6 ± 2.6	35.2 ± 2.8	35.8 ± 1.3
G* (MPa)	0.5 ± 0.1	0.4 ± 0.1	0.4 ± 0.1
G’ (MPa)	0.49 ± 0.12	0.43 ± 0.08	0.38 ± 0.02
Shape Recovery
$R_f$%	96.9±2.0	95.6±2.8	94.1±0.8
$R_r$%	97.4±1.2	100.0±0	94.4±5.6
$\frac{{dR}_r}{dt}$(rad/min)	4.8±1.2	5.0±0.1	2.7±0.5*

^*(p < 0.05) significant difference from negative controls and EtO 37.8 °C

Cytotoxicity, cell viability and endotoxin testing

A selected subset of sterilization conditions demonstrating limited impact on material properties, were selected for cell viability and cytotoxicity testing. Amongst these groups were Ethanol sterilization for 1 min, UV sterilization for 120 min, ethylene oxide sterilization at 37.7°C and 54.4°C. High-density polyethylene (Figure 5A), commonly used as a control for ISO-10993.5C:2012 cytotoxicity testing, did not exhibit any decrease in cell numbers at any extraction percentage. Similarly, samples sterilized by EtO at 37.8°C (Figure 5B) did not exhibit cytotoxicity at the highest extraction concentration. Samples sterilized by UV254nm for 120 min, samples sterilized by Ethanol immersion for 1 min, and samples sterilized by EtO for 54.4°C exhibited a 50% decrease in cell survival (IC50) when incubated with 24%, 64.1%, and 50.5% of the DMEM extract (Figure 5B). Samples sterilized using UV for 120 minutes exhibited significantly lower IC₅₀ values compared to ethanol and EtO, suggesting that components of soluble factors from the polymer as opposed to residuals may be impacting the cytotoxicity (Figure 5B, p < 0.01). FDA endotoxin limit for various medical devices is dependent on the intended device end-use and the contacting tissues and fluids⁽³⁷⁾. The FDA has adopted the USP Endotoxin Reference Standard to evaluate medical devices’ sterilization^(38),(39). The endotoxin limit is 0.5 endotoxin units/mL or 20 endotoxin units per device. Untreated controls, UV and ethanol-treated samples, and EtO sterilized samples were below the acceptable endotoxin limit (Figure 5C). There were no significant differences in endotoxin levels between EtO sterilization at 54.4°C or 37.8°C. Based on the endotoxin and cytotoxicity testing, ethanol immersion for 1 minute and both EtO sterilization groups were selected to test cell viability. Live/dead assay exhibited both live and dead cell populations on ethanol-treated and EtO sterilized samples (Figure 5D). There were no significant differences amongst live or dead cell fractions in ethanol-treated, and EtO sterilized samples (Figure 5E).

Figure 5.

Cytotoxicity, cell viability and endotoxin testing of sterilized and disinfected PGD samples. A) ISO109935C: Cytotoxicity test using an MTT assay to evaluate cell survival when exposed to media extracts from highdensity polyethylene(HDPE) negative control, PGD treated by UV254nm 120 min, 70% ethanol immersion 1 min, 37.8oC EtO cycle, and a 54.4oC EtO cycle B) ISO 10993.5C cytotoxicity testing IC50 values from 3 trials. C) Chromogenic LAL endotoxin quantitative assay and USP Bacterial endotoxin limit for medical devices contacting cardiovascular and lymphatic systems D) Live/dead assay of NIH3T3 cells on PGD substrates treated with Ethanol 70%, sterilized by ethylene oxide at 37.8oC, and sterilized by ethylene oxide at 54.4oC scale bar:250 μm E) Image J quantification of live/dead assay using particle analysis of calcein and ethidium bromide labeled cells

Aging effects on EtO sterilized poly (glycerol dodecanedioate)

Shelf-life of devices is a critical concern for commercialization. Consequently, we only investigated the impact of polymer aging on EtO sterilized materials. The characteristic symmetric and asymmetric CH stretch at 2918cm⁻¹ and 2850cm⁻¹ and the carbonyl stretch at 1737cm⁻¹ are evident in EtO 54.4 sterilized samples after nine months (Figure 6A) 1167cm⁻¹ C-C-O stretch of saturated esters. AT-FTIR spectra at nine months exhibited a greater carboxylic acid O-H hydrogen bond stretch (3466 cm⁻¹) in EtO sterilized samples compared to 0.5- and 4-month sterilized samples indicative of surface degradation of ester bonds. Melt temperatures (T_m) corresponding to the transition temperatures (T_trans) at nine months were higher than the melt transition at 0.5 months (Figure 6B, p < 0.01). EtO 37.8 °C at 9 months was greater than EtO 37.8°C at 0.5 months (p < 0.05). Enthalpy of fusion, proportional to % crystallinity, is higher at four months (Figure 6C, p < 0.05) and 9 (p <0.001) months in EtO 54.4 °C sterilized samples compared to controls. Complex shear modulus at 37°C was greater at 4 months compared to 0.5 months (p < 0.001). Specifically, both EtO sterilization cycles increased the shear modulus of the implants after four months (p < 0.01) and nine months (p < 0.05). There was an increase in fixity from 0.5 months compared to 9 months (p < 0.01) and from 4 months to 9 months for negative controls (p < 0.001). Specifically, there was an increase in fixity at nine months for EtO 54.4 °C samples compared to 0.5 months (p < 0.05). Shape recovery decreased in EtO 54.4 °C sterilized samples at 4 and 9 months compared to controls (Figure 6F, p < 0.001). EtO 54.4 °C sterilized controls exhibited lower Rr after 9 months compared to 0.5 months (p < 0.05). Recovery rate (Figure 6G) of all PGD samples was lower after 4 (p <0.001) months and 9 months (p < 0.001) compared to 0.5 months. Finally, the recovery rate of EtO 54.4 °C sterilized samples was decreased at four months (p < 0.001) and nine months (p < 0.01) compared to controls.

Figure 6.

Shelf-life characterization of ethylene oxide sterilized PGD. A) FTIR spectra of PGD controls at 0, 4, and 9 months. B) Peak melt transition temperature (Tm) C) Enthalpy of fusion D) Shear modulus evaluated at 37oC E) Shape fixity(%) F) Shape recovery (Rr %) G) Shape recovery rate (dRr/dt).

Discussion

Biodegradable shape memory polymers are emerging biomaterials with promising regenerative medicine applications addressing numerous clinical pathologies. Their competitive advantage over conventional polymers for minimally invasive procedures also presents various new clinical device development and delivery questions. As with many polymers used in surgical applications and implantable devices, these materials require validated sterilization processes approved by the FDA. High temperatures and moisture are used in conventional sterilization processes, resulting in surface modification, plasticization, bulk modification, or overall degradation. More specifically, for shape memory polymers, the critical molecular architecture, and crosslinks that allow fixity and recovery require sterilization temperatures at or below the transition onset to avoid altering this essential functionality. In addition to commercial use, these polymers need to be disinfected for various research applications for early product development ranging from in vitro cell culture to subcutaneous implantation in small animal models.

This study has successfully identified various sterilization and disinfection processes that preserve the overall molecular architectures of the polymer while providing necessary sterilization parameters. Specimens used for sterilization studies were 2 mm thick samples that represent the upper range of material geometries for minimally invasive delivery applications. Minimally invasive procedure instrumentation requires delivery through cathethers(2–5mm inner diameter) and trochars having a inner diameters of 5–10mm. When rolling or compressing structures for loading and delivery, even a1 mm thick expandable sheet can take on a cross section of 5–10mm depending on length. Compression and loading of expandable constructs for the devices requires thinner geometries..

UV sterilization of PGD samples resulted in smoother surfaces, indicating surface erosion or plasticization. Similarly, slight differences in surface properties of 50:50 poly-lactic-co-glycolic acid surfaces exposed to UV irradiation were previously described, indicating the surface altering properties of UV irradiation⁽³⁴⁾. After 1 hour of UV irradiation, PLGA 50:50 and PLGA 85:15 both exhibited changes in surface morphology⁽¹⁸⁾. These changes in surface properties should be considered when testing UV sterilized PGD constructs. The overall thermomechanical properties were unchanged, except for more prominent melt transition in the samples UV treated for 120 min. This more prominent peak in the representative thermograms of UV treated PGD indicates a higher enthalpy of fusion than controls. However, these differences were not statistically significant because of the high variance across these samples. The variance can also be attributed to limitations in controlling the uniformity of UV radiation dosage and how different regions of the polymer could be impacted differently due to this difficulty in controlling variance.

Interestingly, UV treating PGD for 120 min resulted in a slower shape recovery rate than controls and the 3 min sterilization process. These differences were also observed in increasing transition temperatures determined from rheometry temperature sweeps but were insignificant (Table 1). The transition temperature of PGD is inversely correlated with the total number of crosslinks, where PGD with more crosslinks will exhibit lower transition temperatures. The preservation of transition temperature measured by DSC and storage modulus(G’) measured by rheometry do not suggest differences in crosslink density.

However, the non-uniformity associated with UV irradiation may be driving the significant variance of measured thermal transitions and the consequent reduction in shape recovery rate. Spatial variability driven by non-uniform irradiation may confound these undiscernible thermomechanical changes from revealing the discernable impact on shape memory properties. The reduction in dRr/dt without any observed physiochemical effects suggests that surface and bulk properties of the polymer may be altered with longer exposure times. PLGA polymers similarly exhibit no discernable dimensional changes but indicate a reduction in the polymer MW⁽²⁵⁾. A thermoset polymer like PGD may similarly undergo changes in polyester crosslink density within the bulk, driving the decreased recovery rate.

Immersion of polymers in 70% ethanol is commonly used as a disinfection method for research-grade applications. After prolonged immersion in 70% ethanol, there were chemical differences and differences in surface morphology at the 15 min time point. Similarly, ethanol sterilized samples in PLGA and PCL exhibited changes in surface properties after a 30 min incubation period (18). Moreover, more prolonged incubation of thermoplastic polyesters in 70% ethanol impacts the morphology and dimensions caused by device degradation^{(17),(18),(25),(30)}. The polymer visibly swells during immersion for prolonged periods (>10 min). Thermosetting polymers like PGD are generally more resistant to changes propagated by swelling or solvent degradation than thermoplastic polymers due to the high degree of crosslinking. For instance, electrospun poly(glycerol sebacate)-poly(caprolactone) sterilized with ethanol overnight did not exhibit an altered overall structure, and the molecular and scaffold architecture was preserved⁽⁴⁰⁾. Rheometry, DSC, and FTIR suggest that the composition and molecular architecture of 70% ethanol sterilization for 0.5 and 1 minute is largely preserved. However, there are differences in the storage modulus(G’) and shape memory properties (R_f, R_r, and dR_r/dt) in PGD samples treated with ethanol for 15 minutes compared to controls. These differences suggest that chemical and thermomechanical properties may be impacted by molecular changes in the bulk more prominently than the surface. Although there are reductions in the shape fixity, recovery and recovery rate, the degree of change makes all ethanol treatment methods viable for various research-grade studies. For instance, increasing the recovery temperature or programming temperature of shape memory polymers could offset significant drops in recovery and recovery rate for research applications in vitro and in vivo⁽²⁾. The uniformity of ethanol immersion, the minimized impact on physiochemical and thermomechanical properties, and the general availability of ethanol reagents make this the preferred research grade disinfection method compared to UV.

EtO sterilized samples processed near the transition temperature (37.8°C) were sterilized and outgassed for a longer duration than samples processed above the transition temperature (54.4°C). Temperature plays a significant role in EtO sterilization. Both cycles were run at a relative humidity of 35%. Exposure to heat was significantly greater in the higher temperature cycle, but exposure to moisture was longer in the lower temperature cycle. Initially, EtO cycles did not impact the thermomechanical or physiochemical properties of the polymer. However, the recovery rate of samples sterilized at 54.4°C was slower than for samples sterilized at 37.8°C, likely driven by the differences in bulk crosslink density. These data suggest EtO sterilization at or below the transition temperature with longer outgassing periods provide ideal sterilization conditions that preserve the molecular architecture and shape recovery rate. However, these changes become more apparent as the polymer ages. EtO sterilization also impacted the degree of aging (Figure 6). The surface composition of PGD suggests more significant degradation at nine months (Figure 6A), while significant changes in shape memory properties occurred within the first four months of aging based on storage conditions at 25°C under vacuum. Ester crosslinks in the polymer that form net points are susceptible to degradation, causing an upward shift in the transition temperatures. Previously, hydrolytic and in vivo degradation of PGD increased Tm and enthalpy of fusion⁽¹⁶⁾. Decreased enthalpy of fusion coupled with increased transition temperatures may correspond with decreased crosslink density and a consequent increase in crystallinity. Aging PGD samples sterilized at EtO54.4 exhibited increased crystallinity compared to samples sterilized at EtO 37.8°C and controls by four months (Figure 6). This increased crystallinity suggests a decrease in PGD crosslinks arising from EtO sterilization.

Increased fixity from aging is also observed in various shape memory polymers. As the crosslinks degrade, fewer net points hold together the switching segments. Net points allow the polymer to return to its original permanent shape. When heated beyond T_trans, the switching segments increase in mobility. The switching segments are crystalline domains in semicrystalline thermoset SMPs like PGD undergo a melt transition. Physical energy imparted on the polymer to program a shape is dissipated within the switching segments when above the melt transition phase. This energy redistribution occurs more favorably with fewer net points. When cooled, the polymer is locked in a metastable energy state where it is kinetically unfavorable to revert to the original permanent ground state, and therefore remains in this transient programmed state. PGD with fewer crosslinks (net points) has more chain mobility to dissipate strain energy and fix the polymer shape. However, when the programmed polymer is heated past the transition temperature, the reduced net points rearrange the polymer in several thermally favorable ground states, thereby reducing overall recovery, driving the consequent drop in R_r and dR_r/dt (Figure 6F, G). Although a four-month storage threshold may seem acceptable from production to clinic, the end-use application, product distribution channels or product sales cycles can dramatically impact the ability to commercialize these materials having a short shelf-life. Taken together, the data reveal current storage limitations of PGD post-sterilization and provide further insight towards challenges that will need to be addressed to successfully commercialize biomedical devices comprised of PGD.

Finally, evaluating the biocompatibility of medical devices is critical for clearing regulatory approval. Implantable devices must undergo exhaustive ISO10993 biocompatibility testing to assess the safety of the device for biological applications. The use of PGD in a medical device, possibly in combination with other components, would require custom sterilization cycle development and appropriate biocompatibility testing pertinent to the clinical design requirement and design control intended for the end-use application. Identifying compatible gas sterilization processes and evaluating cytotoxicity using translatable methods can reduce costs, time, and risk associated with comprehensive biocompatibility testing of a new polymer.

Exposure of cells to media containing polymer byproducts is a standard test method used to evaluate cytotoxicity. The ISO10993–5C media extraction test evaluates toxicity of soluble products released from the device. Surface to volume ratio of specimen, extraction media, and incubation conditions (pH, temperature, flow) are components of this assay tailored to the intended end-use application. Disinfection and sterilization impact these media extracts in two ways: 1) residuals that may become trapped within the material and slowly released upon media extraction or 2) impacting material properties giving rise to different concentrations of soluble products. The presence of residuals is commonly described in the sterilization of various elastomers and shape memory polymers. For instance, poly(glycerol sebacate) elastomers and poly(ethylene glycol di-methacrylate) shape memory elastomers exhibited residuals but no reported impact on cytotoxicity^{(19),(28),(31)}.

The USO certified high-density polyethylene (HDPE) standard and EtO sterilized samples at 37°C resulted in no observable cytotoxicity even at the highest extraction percentages. Samples treated with UV, Ethanol, and EtO cycle at 54.4°C exhibited IC50 values of 24%, 64.1%, and 50.0%, respectively, indicating cell death. Since there were no discernable differences in surface composition indicative of residuals, the cause of the increased cytotoxicity may result from increased soluble byproducts. Therefore, compositional differences in the bulk may be causing the increased solubility of byproducts from UV 120 min, Ethanol 1min, and EtO 54.4°C treatments. Moreover, there was a significant increase in extract cytotoxicity from UV-treated samples after 120min compared to 70% ethanol immersion for 1 minute and EtO 54.4°C. Although PGD degradation byproducts are non-toxic, the solution concentration in this static environment can decrease the pH of the media, as with many polyester degradation products, and ultimately drive cytotoxicity. The live dead assay further supports NIT3T3 cell viability when cultured directly on treated PGD substrates. There was no significant difference in cell viability on Ethanol, EtO 37.8°C, and EtO 54.4°C samples, indicating that the soluble products extracted from the material are mainly driving cytotoxicity (Figure 5).

Medical device sterilization methods used in animal studies and human implantation require methods capable of inactivating bacteria, viruses, and fungi while maintaining acceptable bacterial endotoxin levels—control of material sourcing, manufacturing process, material handling, and packaging limits initial contamination. Reducing or eliminating contamination in these processes is essential to reducing endotoxins in medical devices. Consequently, there was no significant difference amongst sterilization groups. Endotoxins limits for neural and cerebrospinal fluid contacting devices(0.1EU/mL) will require further post-processing and clearance to reduce endotoxin levels.

Conclusion

Shape memory biodegradable polymers require early evaluation of compatible sterilization processes to streamline product development for FDA clearance and approval. Early assessment of the effects of sterilization on thermomechanical and biological properties of shape memory materials can provide insight into which sterilization methods can be used to maintain shape memory behavior and how long this behavior is retained post sterilization. The findings from this study support the use of ethanol disinfection for 1 min and EtO sterilization using the 37oC cycle for for research and commercial grade applications utilizing PGD. Using PGD in a medical device requires further studies conducted at good lab practice (GLP) compliant contract research organizations specific to the end-use application of the polymer containing device. Limited physiochemical impact on PGD combined with favorable cytotoxicity, cell viability, and endotoxin levels provide a pathway towards pursuing validated EtO sterilization cycles and comprehensive ISO10993 biocompatibility testing.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.