Amino Acid-Based Poly(ester urea) Biodegradable Membrane for Guided Bone Regeneration

Abstract

Barrier membranes (BM) for guided bone regeneration (GBR) aim to support the osteogenic healing process of a defined bony defect by excluding epithelial (gingival) ingrowth and enabling osteoprogenitor and stem cells to proliferate and differentiate into bone tissue. Currently, the most widely used membranes for these approaches are collagen-derived, and there is a discrepancy in defining the optimal collagen membrane in terms of biocompatibility, strength, and degradation rates. Motivated by these clinical observations, we designed a collagen-free membrane based on l-valine-co-l-phenylalanine-poly(ester urea) (PEU) copolymer via electrospinning. Degradation and mechanical properties of these membranes were performed on as-spun and water-aged samples. Alveolar-bone-derived stem cells (AvBMSCs) were seeded on the PEU BM to assess their cell compatibility and osteogenic characteristics, including cell viability, attachment/spreading, proliferation, and mineralized tissue-associated gene expression. In vivo, PEU BMs were subcutaneously implanted in rats to evaluate their potential to cause inflammatory responses and facilitate angiogenesis. Finally, critical-size calvarial defects and a periodontal model were used to assess the regenerative capacity of the electrospun PEU BM compared to clinically available Cytoflex synthetic membranes. PEU BM demonstrated equal biocompatibility to Cytoflex with superior mechanical performance in strength and elasticity. Additionally, after 14 days, PEU BM exhibited a higher expression of BGLAP/osteocalcin and superior in vivo performance–less inflammation and increased CD31 and VWF expression over time. When placed in critical-sized defects in the calvaria of rats, the PEU BM led to robust bone formation with high expression of osteogenesis and angiogenesis markers. Moreover, our membrane enhanced alveolar bone and cementum regeneration in an established periodontal model after 8 weeks. We demonstrate that the PEU BM exhibits favorable clinical properties, including mechanical stability, cytocompatibility, and facilitated bone formation in vitro and in vivo. This highlights its suitability for GBR in periodontal and craniofacial bone defects.

Graphical Abstract

1. INTRODUCTION

Reconstructing alveolar bone and craniofacial defects is a complex challenge for oral surgeons and periodontists,¹ leading to over 500,000 reconstructive surgeries annually in the US and costs exceeding $2.5 billion.² These defects, caused by infection, injuries, congenital malformations, and cancer, pose significant health, financial, and quality of life burdens. Guided Bone Regeneration (GBR), commonly used for clinical treatment, involves placing a barrier membrane (BM) in the defect area. This membrane separates soft tissues from bone defects, creating an environment ideal for regenerating bone and periodontal tissues.^3,4 This separation is crucial since different cell types migrate at varying rates during healing. The BM maintains space for slower-moving osteogenic cells to reach the defect site while it blocks faster-moving epithelial and connective tissue cells, ensuring targeted and effective regeneration.^5,6

Many BM are available on the market; however, most present challenges, as their clinical success largely depends on operator skills and technique.^2,7–10 These BM can be categorized into nonresorbable and resorbable varieties. Nonresorbable membranes, like those crafted from expanded polytetrafluoroethylene, claim bioinert properties that minimize the risk of adverse biological reactions and inflammation. However, their use necessitates a second surgery for removal, escalating patient discomfort, costs, and morbidity concerns.^10,11 Conversely, despite being widely used, resorbable membranes, such as those made from collagen or poly(lactic acid) (PLA), have associated drawbacks, including short resorption periods and the potential for inflammation due to acidic byproducts.^12,13 Collagen membranes are also subject to production variability due to their animal-derived composition.¹⁴

The limitations of current collagen-based and synthetic resorbable membranes (i.e., short degradation rates and adverse inflammatory responses) have driven the development of new resorbable BM to address these challenges, meeting a critical clinical need in craniomaxillofacial (CMF) bone regeneration. In recent years, significant advancements in their development have been achieved through electrospinning, a powerful and versatile technique for creating nanofibrous structures with diameters ranging from a few nanometers to several micrometers. These nanofibers, characterized by their high surface area-to-volume ratios, tunable porosity, and unique mechanical properties, hold great promise as BM for GBR.¹⁵ In addition to the manufacturing technique, l-valine-co-l-phenylalanine (PEU) copolymer is a promising biomaterial for GBR applications due to its tunable mechanical properties and controlled degradation rate, as demonstrated by previous research.^16–19 Using amino acid–based polyester derivatives, such as PEUs, is particularly advantageous because their degradation byproducts are essential amino acids, which do not result in acidic build-up. This reduces toxicity and prevents inflammation and bone resorption, ensuring a more stable and healthier environment for bone regeneration.

This study focused on a specific l-valine-co-l-phenylalanine PEU composition, selected based on a stable electrospinning jet, which ensures consistent production and optimal microstructural features vital for nutrient infiltration and osteogenesis. Herein, we report the fabrication of a novel PEU BM uniquely designed to prevent soft tissue invasion while facilitating nutrient infiltration, with critical design criteria: extended resorption periods, collagen-free, avoidance of acidic byproducts that cause tissue inflammation, ease of handling, and promotion of favorable bone regeneration outcomes in clinically relevant models. Since collagen-based membranes exhibit fundamentally better biological properties, a commercially available (Cytoflex) microporous polymeric membrane made of synthetic, bioresorbable polyglycolide/polylactide (PGA/PLA) copolymer was used as a clinical reference to offer a fair assessment in terms of mechanical competence and biological outcomes to our developed BM.

2. EXPERIMENTAL SECTION

2.1. Synthesis of PEU (Poly(1-Phe-6)₅₀-co-(1-Val-10)₅₀).

PEU (Poly(ester urea)) copolymers were synthesized using a copolymerization strategy by reacting stoichiometric mixtures of the di-ptoluenesulfonic acid salts of bis-L-amino acid-diesters together with triphosgene and sodium bicarbonate. PEU homopolymers are written as poly(1-A-D), where A stands for amino acid and D stands for diol chain length. A homopolymer synthesized using l-Phe with a diol chain length of six methylene units would be written as poly(1-Phe-6). For copolymers, they are abbreviated as poly[(1-A-D)_S-co-(1-A’-D’)_S’] where A stands for amino acid, D stands for diol chain length, and S represents the stoichiometric ratio of the monomers. A copolymer synthesized from a monomer salt of l-Phe with a diol chain length of six methylene units added at 30% mole fraction to a monomer salt of l-Val with a diol chain length of eight methylene units added at 70% mole fraction is abbreviated as poly[(1-Phe-6)₃₀-co-(1-Val-8)₇₀]. For this work, 50% poly(1-Phe-6)₅₀-co-(1-Val-10)₅₀ or 50% P6 V10 was synthesized using the following materials: Monomer (1-PHE-6) (50.00 g, 0.07 mol, 0.5 equiv), Monomer (1-VAL-10) (47.40 g, 0.07 mol, 0.5 equiv), sodium carbonate (43.40 g, 0.41 mol, 3.1 equiv), and triphosgene (15.70 g, 0.05 mol, 0.4 equiv) (Mn = 69.9 kDa, Mw = 134.4 kDa, ĐM = 1.9, Tg = 38.4 °C, Td = 235.2 °C) by an interfacial step growth polymerization process outlined in previous studies.^17,20 This particular formulation was chosen as the higher mass fraction of the phenylalanine subunit led to a glass transition temperature above physiological temperature (Supporting Information, Figure S1A) and prevented the collapse of the fiber mesh into a film in vivo. The process involved combining chloroform and water in a 2 L reactor equipped with an overhead mechanical stirrer. The reactor contained the specified molar equivalents of di-p-toluenesulfonic (1 equiv) acid monomer salts and sodium carbonate (3.15 equiv) to initiate the polymerization reaction. Approximately 500 mL of hot distilled water at 100 °C was poured into the flask and stirred at 200 rpm until complete dissolution. Afterward, the flask was cooled using an ice bath until it reached a temperature of 10 °C. Upon achieving this temperature, triphosgene (0.43 equiv) was dissolved in 500 mL of chloroform and introduced into the solution. Instantly, the solution turned white upon adding the organic solution, and the stirring continued for 48 h at 400 rpm. The resulting mixture was transferred into a separatory funnel and rinsed with 500 mL of deionized water. The chloroform layer was carefully added dropwise into 4 L of hot water. In instances where any remaining salt traces were detected in the H NMR analysis, the raw product was dissolved in a 2 L volume of acetone and subsequently precipitated into 4 L of hot water. This method ensured the thorough removal of any residual salts. The ¹H NMR spectrum (500 MHz, 303 K, DMSO-d₆) depicted distinct peaks from the valine and phenylalanine monomer components labeled X and Y, respectively (Supporting Information, Figure S1B). The recorded shifts are as follows: δ = 0.81–0.85 (m, 12H, –CH(CH₃)₂), 1.16–1.19 (b, 2YH, –COOCH₂CH₂(CH₂)YCH₂CH₂OOC–), 1.18–1.26 (br, 2XH, –COOCH₂CH₂(CH₂)XCH₂CH₂OOC–) 1.43–1.44 (b, 4H, –COOCH₂CH₂(CH₂)YCH₂CH₂OOC–), 1.53–1.54 (br, 4H, –COOCH₂CH₂(CH₂)XCH₂CH₂OOC–), 1.95–1.98 (m, 2H, –CH(CH₃)₂), 2.50 (s, DMSO), 2.90–2.92 (m, 2H, -NHCH(CH₂Ph)COO–), 3.29–3.33 (H₂O), 3.95 (t, 2H, –COOCH₂CH₂(CH₂)YCH₂CH₂OOC–), 4.03–4.06 (m, 6H, –CHCOOCH₂(CH₂)XCH₂OOCCH–) 4.36–4.39 (m, 2H, -NHCH(CH₂Ph)COO–), 6.36–6.38 (d, 2H, -NH–), 6.48–6.52 (d, 2H, -NH–), 7.13–7.25 (m, 10H, -C₆H₅) ppm.

2.2. Electrospinning of PEU BM.

All reagents and solvents were acquired from Sigma-Aldrich (St. Louis, MO) and used without further purification. PEU pellets were dissolved in hexafluoroisopropanol (HFIP) at a concentration of 12% (wt./v) and stirred continuously for 24 h. The resulting polymer solution was electrospun using Matregenix’s industrial-scale roll-to-roll electrospinning system (Model# ES800H4, Mission Viejo, CA). In this system, the polymer solution is loaded into a closed container and withdrawn using peristaltic pumps. The solution is then passed through branched silicon tubes and fed into four steel rods, each equipped with 40 convergent-divergent nozzles. Fibers are electrospun and collected onto a winding polyethylene terephthalate (PET) substrate in a bottom-up spinning configuration. The electrospinning process parameters were carefully optimized as follows: flow rate of 90 mL/h., screen distance of 107 mm, applied voltage of +65 kV on the solution and –10 kV on the collecting target, winding speed of 1.6 m/min, substrate horizontal movement of 80 mm, and movement speed of 30 mm/sec, ambient humidity maintained at 25%, and temperature maintained at 25 °C. These parameters were fine-tuned to enhance the stability of the electrospinning jet and facilitate consistent fiber formation.

2.3. Characterization of PEU (Poly Ester Urea).

For all analyses conducted hereafter, Cytoflex (Unicare Biomedical, Inc. Laguna Hills, CA, USA), a microporous membrane made of synthetic, bioresorbable polyglycolide (PGA), polylactide (PLA) copolymer, was used as the control. Mechanical properties (tensile strength, Young’s modulus, and elongation at break) of the BM were assessed using uniaxial tensile testing on an eXpert 5601 machine from ADMET (Norwood, MA, USA) with a 1 kN load cell from the same manufacturer (SM-250–961—250 lbf). Rectangular samples sized 25 × 3 mm, with a gauge length of 19 mm, were used for testing, with ten samples per group. Testing was performed under two conditions: dry, where samples were immediately tested upon receipt, and wet, where samples were immersed in a PBS solution for up to 60 days before testing at different intervals. Testing was conducted at a crosshead speed of 1 mm/min. The degradation of PEU BM and Cytoflex was examined by submerging samples measuring 10 × 10 mm (n = 6) into PBS and lipase (150 U/L) solutions at 37 °C. Their weight changes were continuously observed over 60 days.²¹ At specific intervals, the samples were removed from the solution, gently dried using low-lint wipes (Kimwipes, Irving, TX, USA), rinsed twice with deionized water, and vacuum-dried for 24 h before noting their weights. The degradation pattern was calculated using the following formula: Degradation ratio (%) = Wt/W₀ × 100.

2.4. Cell Culture and Viability Evaluation.

A primary culture of alveolar bone-derived mesenchymal stem cells (AvBMSCs) was obtained from human alveolar bone and characterized by flow cytometry to investigate the presence of CD90+, CD70+, and CD105+ markers as previously described.²² Cells were subcultured to passages 3–7 for the following experiments. Samples (10 × 10 mm²) were sterilized in UV light (1 h on each side), and placed in 24-well plates (Ultra-Low Attachment Microplates, Corning, ME, USA) under sterile stainless steel rings to standardize the seeding area (Supporting Information, Figure S2),²³ washed with 70% ethanol (10 min), rinsed twice with PBS, and soaked in minimum essential medium (α-MEM) containing 15% fetal bovine serum, 1% penicillin-streptomycin (all from Gibco). AvBMSCs were seeded on the membrane surface and cultured with or without differentiation medium (0.1 μM dexamethasone, 10 mM glycerophosphate, and 50 μg/mL ascorbic acid). Cells (3 × 10⁴) were seeded on the membrane surface for cell viability, proliferation, and morphology. For the differentiation assay, the seeding density was 1 × 10⁵ cells/sample, and the medium was replaced with osteogenic medium after 24 h. For all assays, the medium was refreshed every 2 days. Cells seeded on commercial barrier membrane Cytoflex were used as a control.

Barrier membranes (n = 4), seeded with AvBMSCs, underwent a 7-day incubation period. Cell viability was assessed employing a resazurin-based solution (alamarBlue, Invitrogen). The membrane/cell samples were collected and immersed in alamarBlue solution (1:10) for 3 h at each time point. Following the incubation, the supernatant was gathered, and fluorescence was measured at 560 nm excitation/590 nm emission using a Spectra Max Id3 spectrophotometer. The fluorescence readings were normalized against the negative control (blank). The percentage of cell viability was determined by comparing the values to the commercial barrier membrane Cytoflex at day one, which was considered 100%. For the qualitative assessment of cell morphology, AvBMSCs seeded on the membranes underwent examination using a scanning electron microscope (SEM) (MIRA3, TESCAN Brno, Kohoutovice, Czech Republic) with a sample size of n = 2. Following 24 and 72 h of culture, the cell-seeded membranes were fixed in formalin and subjected to dehydration through a graded ethanol series. Each sample was imaged at 2000× magnification, capturing four images from distinct areas.

2.5. In Vitro Osteogenesis Evaluation of PEU (Poly Ester Urea).

RNA extraction was performed to evaluate gene expression using the TRIzol method from AvBMSCs following osteogenic differentiation for 7 and 14 days (n = 6). The optical density of the samples, measured at 260 nm, determined RNA concentration. Subsequently, 600 nanograms of RNA underwent reverse transcription for first-strand cDNA synthesis using SuperScript (Thermo Fisher). Real-time PCR quantification of mRNA targeted osteoblastic markers, including ALPL (Hs01029144_m1), COL1A1 (Hs01076756_g1), RUNX2 (Hs01587814_g1), and BGLAP (Hs01047973_m1), was conducted using Taqman Gene Expression Assays (Applied Biosystems). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) served as the housekeeping gene. Samples were read in triplicate, and the results were expressed as fold change relative to the expression of AvBMSCs seeded on the commercial barrier membrane Cytoflex.

2.6. In Vivo Biocompatibility of PEU (Poly Ester Urea).

Twelve male Fischer 344 rats (Envigo RMS, Inc., Oxford, MI, USA) aged 10 weeks and weighing 250–290 g were used. They were divided into four groups corresponding to different time points: 1 week, 1 month, 2 months, and 3 months. All animal procedures adhered to the guidelines for reporting animal research and were approved by the local Institutional Animal Care and Use Committee (IACUC, PRO00010329). Surgical procedures were conducted under general anesthesia induced by administering 50 mg/kg of ketamine (Hospira, Inc., Lake Forest, IL, USA) and 5 mg/kg of xylazine (Akorn, Inc., Lake Forest, IL, USA) intraperitoneally. An incision of 2 cm was made in a head-to-tail orientation using a size 15 scalpel blade. Subsequently, four small subcutaneous pockets were created to accommodate the placement of two square-shaped samples (10 mm height ×10 mm width ×1 mm depth) of PEU BM and two Cytoflex in a randomized manner (six scaffolds per group at each time point). The animals were allowed to recover after wound closure with Coated Vicryl polyglactin 910 (Ethicon Endo-Surgery, Inc., Cincinnati, OH, USA). At specific time intervals, the animals were euthanized using CO₂ inhalation. The implanted BM and surrounding tissue were retrieved, fixed in 10% buffered formalin overnight, embedded in paraffin, and sectioned into 6 μm-thick slices. These slices were stained with hematoxylin and eosin (H&E) to assess the presence of inflammatory cells, neovascularization, and degradation.²⁴ The quantification of inflammatory cells was conducted utilizing the ImageJ software (U.S. National Institutes of Health, Bethesda, Maryland, USA) at 10× magnification. Succinctly, three separate images per sample were transposed from their original format to grayscale by adjusting their color channels to an 8-bit scale. Subsequently, the images were subjected to thresholding using an optimal threshold value to distinguish inflammatory cells from the background. Following the application of the threshold, the images underwent particle analysis, with specific criteria for size and circularity being established by modifying the relevant parameters in the Analyze Particles dialogue box.^25,26

Additionally, tissue slices underwent immunolabeling to evaluate angiogenesis using anti-CD31/PECAM1 (00055, Fortis Life Sciences, Waltham, Massachusetts, USA) and anti-Factor VIII Related Antigen/von Willebrand Factor (RB281A, Epredia, Portsmouth, New Hampshire, USA).²⁷ Moreover, the macrophage polarization profile was assessed through proinflammatory M1 macrophage staining (anti-iNOS, rabbit recombinant multiclonal, ab283655, Abcam, Cambridge, MA, USA) and anti-inflammatory M2 macrophage staining (anti-CD163, rabbit monoclonal, ab182422, Abcam).²⁸ The secondary antibodies, Alexa Fluor 568 (goat antirabbit IgG H&L, ab175471, Abcam) and Alexa Fluor 488 (goat antirabbit IgG H&L, ab150077, Abcam), were applied subsequently. The former was used for VWF/M2 macrophages, while the latter was applied for CD31/PECAM1/M1 macrophages. The cell nuclei underwent staining with DAPI using VECTASHIELD Antifade mounting media (Vector Laboratories, Newark, CA, USA). Subsequently, ImageJ software was utilized to measure the areas of positive immunofluorescence staining from six randomly chosen images per group.^27,29

2.7. In Vivo Osteogenesis of PEU (Poly Ester Urea) – Calvaria Defect.

Thirty male Fischer 344 rats aged 10 weeks, weighing 250–290 g, were employed following the approved IACUC protocol (PRO00010329). A single 6 mm-diameter calvarial defect was created using a trephine bur (REF 04948501, Ace Surgical Supply Co., Brockton, Massachusetts, USA) under general anesthesia induced with inhalation isoflurane (Piramal Critical Care Inc., Bethlehem, PA, USA) at 4–5% for induction and 1–3% for maintenance. The rats were divided into three groups (five per group at each time point): no treatment (SHAM), PEU BM, and Cytoflex. At four- and eight-week intervals postimplantation, the animals were euthanized via CO₂ inhalation. Samples were collected, fixed in 10% buffered formalin overnight, and subjected to μCT analysis (using a voxel size of 18 μm, 70 kV, 114 μA). The reconstructed 3D images were drawn to the original defect’s circumference, labeled the region of interest (ROI), and analyzed for bone volume fraction (BV/TV). Following μCT analysis, the samples underwent decalcification in 10% EDTA for 4 weeks, paraffin embedding, and sectioning into 5 μm-thick slices. These slices were stained with H&E and Masson’s Trichrome. Additionally, immunolabeling was performed to evaluate angiogenesis using anti-CD31/PECAM1 and anti-Factor VIII Related Antigen/von Willebrand Factor. Furthermore, osteogenic potential was assessed using antiosteopontin (Rabbit polyclonal, ab216402, Abcam, Cambridge, Massachusetts, USA) and anti-RUNX2 (Rabbit monoclonal, ab92336, Abcam). A universal secondary antibody, Alexa Fluor 488, was applied, and cell nuclei were stained with DAPI using VECTASHIELD Antifade mounting media. ImageJ software (National Institutes of Health) was used to quantify positive immunofluorescence staining areas from six randomly selected images per group.^27,29

2.8. In Vivo Osteogenesis of PEU (Poly Ester Urea) – Periodontal Defect.

The functionality of membranes in vivo was assessed using a rat periodontal defect model, following established procedures.^30,31 Twenty-four male Fischer 344 rats, aged 10 weeks and weighing 250–290 g, were employed following the approved IACUC animal protocol - PRO00010329. To simulate the periodontal defect, mucoperiosteal flaps were raised, exposing the lingual side of the right and left maxillary first molars’ alveolar bone. With continuous saline irrigation, a dental bur removed the alveolar bone covering the tooth roots on the lingual side. The resulting periodontal defect measured 1.5 mm in width, 3 mm in length, and 2 mm in depth, into which the BMs were implanted. The rats were grouped into three categories (eight defects per group at each time point): no treatment (SHAM), PEU BM, and Cytoflex. At four- and eight-weeks postimplantation, the animals were euthanized via CO₂ inhalation, and the samples were collected, fixed overnight in 10% buffered formalin, and subjected to μCT analysis (using a voxel size of 12 μm, 70 kV, 114 μA). Vertical bone loss at each defect site was determined by measuring the space between the alveolar bone crest and cemento-enamel junction (CEJ). Afterward, the collected samples were decalcified in 10% EDTA for 8 weeks, embedded in paraffin, sliced into 5 μm-thick sections, and stained with H&E. The stained sections were then examined under a light microscope at 4× and 10× magnifications to assess the presence of an inflammatory profile, regenerative performance, and membrane degradation at an orthotopic periodontal site.

2.9. Statistical Analysis.

Statistical analyses were conducted using GraphPad Prism 10. The data first underwent Shapiro–Wilk tests for normality and Levene tests for variance homogeneity. Subsequently, one-way ANOVA was applied, followed by Tukey’s or Games–Howell post hoc tests as appropriate. Two-way ANOVA was utilized to evaluate cell viability, the mechanical properties of wet membranes, and in vivo model results. All statistical conclusions were drawn with a significance threshold set at 5%.

3. RESULTS AND DISCUSSION

3.1. Characterization of PEU (Poly Ester Urea) Barrier Membrane.

Mechanical and degradation in vitro tests assessed the membrane strength and resorption profile over time. Collectively, the PEU BM showed stable in vitro degradation both in phosphate buffer saline (PBS) and lipase-enriched solution, an enzyme employed to replicate the natural degradation process within living organisms, facilitating the hydrolysis of ester bonds in polymeric scaffolds,^32,33 with ca. 95% of the original mass remaining after 30 days and ca. 90% of the original mass after 60 days (Figure 1A). The mechanical analysis indicated that the PEU BM could withstand a tensile strength of up to 6 MPa before breaking, compared to 4 MPa for Cytoflex (P = 0.0003). Furthermore, its elasticity was approximately 120%, significantly higher than Cytoflex’s elasticity of around 17% (P < 0.0001), as shown in Figure 1B. Lastly, mechanical analyses were done to assess the membrane strength over time to elucidate its overall mechanical competence and potential collapse into the bone defect. The tensile strength of PEU BM was statistically higher compared to those of Cytoflex for time points up to the thirtieth day, starting from 5.7 MPa on the first day (P < 0.0001), 4 MPa on the seventh day (P = 0.0003), and 3.4 MPa (P = 0.0102) on the 15th day, compared to the practically constant 2.5 MPa of Cytoflex for the same time points. On the thirtieth day, PEU BM showed 2.3 vs 1.5 MPa of Cytoflex (P = 0.0016). When comparing the Young’s modulus, PEU BM was superior to Cytoflex on the first day (50 MPa compared to 35 MPa, P = 0.0011) and at the 30th day (47 MPa compared to 35 MPa, P = 0.0169).

Figure 1.

Comparative Characterization of Cytoflex and PEU BM Membranes. (A) Degradation profiles of Cytoflex and PEU BM over 60 days on both PBS and lipase solutions, illustrating the percentage of remaining barrier membrane mass (%). Asterisk (*) = P < 0.0001 for materials compared within time point. (B) Mechanical properties of the as-received (dry) membranes, including tensile strength, Young’s modulus, and elongation at break, were analyzed using an unpaired t test to identify significant differences. (C) Tensile strength, Young’s modulus, and elongation at break after incubation in PBS at 37 °C for extended time points (up to 60 days), using two-way ANOVA with Sidak’s post hoc test for statistical analysis. P values highlight statistically significant differences, while ns. denotes not significant.

Regarding flexibility, PEU BM consistently outperformed Cytoflex across all the studied time points. The significant difference observed on the first day was particularly noteworthy, making it highly malleable and effectively conforming to bone defects. Overall, our PEU BM demonstrated significantly greater mechanical strength over the first 30 days of incubation (Figure 1C), which is critical to guarantee the initial phase of bone healing, mainly when applied to large defect sites such as those where GBR is employed. The membrane’s stable in vitro degradation profile also suggests that the membrane has a consistent resorption rate that can be predictable over time. Moreover, the mechanical strength over time shows that the membrane holds its mechanical strength better than Cytoflex during the initial phase of bone healing, meeting the requirements for extensive bone repairs while maintaining proper healing over time.¹⁷ In sum, the dry PEU BM outperforms the Cytoflex in terms of tensile strength and elasticity (elongation at break). When evaluated under wet conditions, the PEU BM surpasses the Cytoflex in elongation at break and tensile strength across various time points. Moreover, the Young’s Modulus of the PEU BM matches or exceeds that of the Cytoflex, further emphasizing its superior mechanical performance, which should translate into better regenerative outcomes in significant bone defects.

The observed decrease in mechanical properties over time in solution is mainly due to the surface erosion of the poly(ester urea) (PEU) membrane. Unlike bulk-degrading materials such as Cytoflex, PEU degrades through hydrolysis of its outer layers, gradually reducing its mechanical strength. The erosion mechanism weakens the material’s structural integrity without significant mass loss, as tensile strength data show. Despite retaining about 90% of its mass after 60 days, the PEU’s mechanical performance declines as surface erosion progresses, with minimal early stage degradation products and less impact on molecular weight. In GBR, the membrane’s durability and sustained barrier functionality are crucial, particularly within the critical time frame of 4–24 weeks.³⁴ This duration aligns with bone tissue regeneration, stressing the need for a durable BM that isolates the bone defect, hinders fast-growing cells, and fosters an environment for osteoblast activity. Apart from durability, the BM must allow easy coverage over bone defects and seamlessly conform to adjacent bone contours. Elasticity is crucial to preventing permanent deformation, preserving structural integrity, and ensuring proper functionality.³⁵ Our PEU BM meets these multifaceted criteria, promising advancement in pursuing successful and efficient GBR techniques.

3.2. In Vitro Viability and Osteogenic Differentiation Ability.

The electrospinning process produced a BM with an isotropic fibrous structure characterized by the absence of morphological defects or bead formation (Figure 2A). Our data demonstrate that our PEU BM provides a favorable substrate for cells without affecting their viability. Following AvBMSCs seeding, despite PEU BM initially exhibiting a nearly 30% decrease on day 1 compared to Cytoflex, the remaining viable cells surpassed 70%, indicating a cytocompatible behavior in line with the ISO 10993–5 standard. Furthermore, at later times, the observed PEU BM cell viability (Figure 2B) was on par with the commercial membrane Cytoflex, highlighting sustained and comparable support for cell viability and proliferation by PEU over time. The hydrophobic behavior of PEU membranes can decrease the initial cell adhesion and proliferation,³⁶ presenting lower cell viability when compared with Cytoflex (PLA/PGA), which presents a more hydrophilic behavior due to the presence of PGA compared to PEU.³⁷

Figure 2.

Cell-Membrane Interactions. (A) Scanning Electron Microscopy/SEM images depicting membrane morphology both in the absence and presence of AvBMSCs. (B) Cell viability of AvBMSCs cultured on the membranes over intervals of 1, 3, and 7 days. Note that the PEU BM morphology favored cell attachment and differentiation, with over 70% viable cell proliferation on the membrane, indicating cytocompatible behavior. A significant difference between the groups at each time point is indicated by an asterisk (*), analyzed through Two-way ANOVA with Sidak’s posthoc test. (C) RT-PCR for ALPL, COL1A1, and RUNX2 genes after 7 days of culture. (D) RT-PCR analysis to include ALPL, COL1A1, RUNX2, and BGLAP genes after 14 days. RT-PCR data are presented as mean fold change ± standard deviation and analyzed using an unpaired t test. P values highlight statistically significant differences, while ns. denotes not significant.

Regarding the osteogenesis potential, we evaluated the expression of ALP, COL1A1, and RUNX2 genes at 7 days, enriched by adding BGLAP at 14 days. At the initial time point, none of the genes investigated showed a statistically significant difference in expression levels relative to the control (Cytoflex) (Figure 2C). However, following 14 days of culture, there was a decrease in expression of the genes ALP (alkaline phosphatase) and COL1A1 (collagen type 1 alpha), both of which are early markers of osteogenesis. The late osteogenic marker BGLAP (osteocalcin) gene, representing mature osteogenic differentiation, showed a significant upregulation in the PEU BM group compared to the control (Figure 2D). The lack of disparity in RUNX2 expression suggests the potential involvement of alternative pathways or factors influencing osteocalcin production, such as OSX/SP7 and ATF4,^38,39 which were not assessed in the present investigation. While interpreting these findings, one should exercise caution due to the inherent limitations of studying these data within an in vitro microenvironment. Altogether, the PEU membrane offers adequate morphological features (i.e., three-dimensional fibrous morphology closely resembling that of the collagen fibers of the native extracellular matrix ECM of bone tissue) that support osteogenic differentiation of cultured AvBMSCs. These outcomes indicate that PEU BM could be a potential alternative GBR membrane due to similar gene expression and superior performance across multiple other material-based parameters compared to Cytoflex. This may lead to significant regenerative differences when applied to large bone defects.

3.3. In Vivo Biocompatibility.

For several reasons, implanting membranes into rat subcutaneous tissue is widely used as the initial biocompatibility assessment for new materials.^40–42 First, it enables the evaluation of biocompatibility, mimicking interactions similar to human tissue and predicting potential reactions or adverse effects. Additionally, this method facilitates the study of host responses, including inflammatory reactions, foreign body responses, tissue integration, and the organism’s tolerance to the new material. Moreover, it allows for monitoring the degradation process over time, observing breakdown and potential effects on surrounding tissues.

Examining the BM’s biodegradation, we found a notable resemblance between their in vitro degradation patterns and the observations from subcutaneous implantation in rats (Figure 3A). Particularly intriguing was the behavior of the PEU BM, which exhibited a distinctive separation among its fiber layers, suggesting an accelerated biodegradation process. Surprisingly, this separation did not facilitate cell infiltration within the membrane, presenting an interesting dual effect.

Figure 3.

In vivo biocompatibility and degradation postimplantation on rat subcutaneous tissue after 1-week, 1-, 2-, and 3-months, n = 6. (A) Representative Hematoxylin & Eosin slices for implanted membranes (m) at 4 ×, 10 ×, and 20× magnification (scale bar = 400, 200, and 100 μm, respectively). Black dashed squares highlight the areas selected for high magnification. At 7 days postimplantation, there is evidence of moderate inflammation, partly due to the surgical process involved in implanting the membranes. However, beginning from the first month, there is a significant decrease in the inflammatory process, which remains consistent until the third month. The expansion of the spaces occupied by the membrane can be observed over time, with signs of membrane resorption in the middle. (B) Bar graphs for inflammatory total cell count (mean ± SD), analyzed through Two-way ANOVA with Sidak’s posthoc test. P values highlight statistically significant differences.

Moreover, a comprehensive comparison of the inflammatory cell counts surrounding and infiltrating PEU and Cytoflex membranes reveals consistent and statistically significant differences favoring PEU BM across all time points. After 1 week, PEU-BM exhibited a 41% reduction in inflammatory cells compared to Cytoflex, with a mean of 725 ± 74 inflammatory cells versus Cytoflex’s significantly higher count of 1773 ± 92 cells. This reduction persisted at one month, showing a 34% decrease in PEU BM with 414 cells ±88 against Cytoflex’s higher count of 1220 cells ±44. The trend continued at two months, displaying a 44% reduction in PEU BM with 352 ± 57 cells compared to Cytoflex’s count of 799 ± 41. At the 3-month end point mark, PEU BM maintained a 40% reduction, registering 333 ± 57 inflammatory cells compared to Cytoflex’s count of 825 ± 71. These consistent reductions across time points indicate PEU’s notable superiority in eliciting significantly lower inflammatory responses than Cytoflex, demonstrating its potential as a more favorable biomaterial for GBR applications (Figure 3A–B).

Beyond inducing a controlled inflammatory response, the angiogenesis capacity is crucial for membranes in GBR. This process significantly contributes to establishing vascular networks, aiding tissue regeneration and overall success. Overall, PEU BM supported adequate angiogenesis after 1-, 2-, and 3-months postimplantation, as evidenced by augmented immunolabeling for CD31/PECAM1 after one month (p < 0.0001) and Von Willebrand factor after two months (p = 0.0374) (Figure 4A–B). CD31 and VWF are critical contributors to angiogenesis, although through distinct mechanisms. CD31, found in endothelial cells, plays a pivotal role in angiogenesis by facilitating endothelial cell–cell interactions, fostering the formation of stable blood vessels, and regulating endothelial cell migration during the angiogenic process. Moreover, CD31’s involvement in signaling pathways influences endothelial cell survival and proliferation, further supporting angiogenesis.⁴³ Conversely, besides its role in coagulation and platelet aggregation, VWF acts as a bridge between the extracellular matrix and endothelial cell surfaces, promoting endothelial cell adhesion and migration, crucial steps in developing new blood vessels.^44,45 CD31 and VWF play complementary roles in mediating endothelial cell interactions, adhesion, migration, and the establishment of stable vasculature, thus significantly contributing to the intricate process of vascular development and regeneration.

Figure 4.

In vivo immunolabeling for angiogenesis markers (A) CD31/PECAM1 with secondary antibody Alexa Fluor 488 (green) and (B) VWF with secondary antibody Alexa Fluor 568 (red) after 1-, 2-, and 3-months postimplantation (10× magnification). Note a significant increase in the angiogenesis markers CD31/PECAM1 after 1 month and VWF after 2 months for the PEU BM compared to Cytoflex. Bar graphs for immunolabeling expressions of CD31/PECAM1 and VWF (mean ± SD), analyzed through Two-way ANOVA with Sidak’s posthoc test. P values highlight statistically significant differences, while ns. denotes not significant.

The fundamental role of macrophages in assessing new membranes for GBR lies in their exhibition of distinct phenotypes, specifically the polarization into M1 and M2 types, which provides essential insights for biomaterial evaluation. M1 macrophages, pro-inflammatory in nature, initiate the tissue response by releasing cytokines like TNF-alpha, IL-6, and IL-1β. They are adept at phagocytosis and breaking down foreign materials.^46–48 Conversely, M2 macrophages, which are anti-inflammatory, are associated with tissue repair and regeneration through the production of growth factors such as TGF-β and express markers like CD206. These M2 macrophages play a role in resolving inflammation and supporting tissue healing.^49–51 When examining new membranes for GBR, understanding how these macrophage subsets interact with the material is vital.

In the present study, M1/M2 macrophage polarization around membranes revealed that while PEU BM and Cytoflex showed similar M2 expression (anti-inflammatory phenotype), PEU has significantly lower M1 (pro-inflammatory phenotype) levels after two months (P = 0.0003). This suggests that our PEU BM may have superior tissue repair/regenerative capabilities due to its reduced inflammatory response compared to Cytoflex (Figure 5A–B).

Figure 5.

In vivo immunolabeling for macrophage polarization markers (A) iNOS (M1) with secondary antibody Alexa Fluor 488 (green) and (B) CD163 (M2) with secondary antibody Alexa Fluor 568 (red) after 1-, 2-, and 3-months postimplantation (10× magnification). Note a significant decrease in the M1 macrophage marker for PEU BM after 2 months compared to Cytoflex. Bar graphs for immunolabeling expressions of iNOS and CD163 (mean ± SD), analyzed through Two-way ANOVA with Sidak’s posthoc test. P values highlight statistically significant differences.

3.4. In Vivo Osteogenesis of PEU (Poly Ester Urea) – Calvaria Defect.

The PEU BM’s ability to effectively work on a nonself-healing model mimics challenging clinical situations, allowing quantitative evaluation of regenerative potential, biocompatibility, and mechanistic insights (Figure 6A).

Figure 6.

Rat critical size calvarial defect model. (A) Photographs showing the section of the parietal bone with the trephine, the intact appearance of the brain, and the positioning of the membrane before wound closure. (B) Microcomputed tomography (μCT) of the bone defects with Cytoflex, PEU, or SHAM (empty) after 4- and 8-week postimplantation. White dashed circles highlight the areas assessed for bone quantification. Scale bar: 6 and 1 mm. (C) Ratio between bone volume and total volume defect analyzed through Two-way ANOVA with Tukey posthoc test. P values highlight statistically significant differences. Observe a statistically significant higher bone formation evoked by both membranes than the SHAM, without differences between them.

Following μCT analysis at 4 weeks, the SHAM group demonstrated a bone fill volume of 16.0 ± 3.2%. Meanwhile, the PEU BM exhibited a slightly higher BV/TV at 17.4 ± 4.3%, and the Cytoflex group showed a higher volume at 21.9 ± 3.7%. However, no statistical differences were observed between the PEU BM and Cytoflex groups at this time point. Transitioning to the 8-week evaluation, the SHAM group experienced a marginal increase in BV/TV to 17.5 ± 0.7% (Figures 6B–C). Notably, the PEU BM group showed a considerable improvement, reaching a bone fill volume of 30.6 ± 5.6%, while the Cytoflex group also demonstrated progress, achieving 33.4 ± 6.8%. Both applied BMs exhibited statistical superiority compared to the SHAM group at this stage. These results highlight the effectiveness of PEU BM and Cytoflex in treating critical bone defects, showing significantly better outcomes compared to the SHAM control group (Figure 6C). Histologically, the investigation revealed the beginning of bone formation within the defect area, mirroring the original marginal bone thickness when treated with BM’s (Figures 7A–B). In contrast, the SHAM group displayed newly formed bone with a thickness approximately half that of the original layer. Notably, at the 8-week mark, consistent bone thickness was observed across all groups. Nevertheless, PEU BM and Cytoflex demonstrated significantly more expansive bone formation than the empty defect, encompassing nearly the entire circumference of the circular defect (Figure 7).

Figure 7.

Hematoxylin/Eosin and Masson’s trichrome-stained slices of rat calvaria critical-size defects after (A) 4- and (B) 8-weeks postimplantation. New bone formation was observed in all groups, with a higher volume for Cytoflex and PEU. Black dashed squares highlight the areas selected for further magnification. New bone is indicated with (nb), original bone (ob), 4×, and 10× magnification, with 600 and 300 μm scale bars, respectively.

Regarding BM’s reabsorption, a consistent pattern in the subcutaneous model persisted. The PEU BM exhibited increased separation of fiber layers, concurrently impeding cellular infiltration between them. This feature holds significant importance as it emphasizes the objective of BMs to prevent the infiltration of specific cell types, particularly rapidly proliferating epithelial and connective tissue cells. By doing so, it fosters an environment conducive to the development of slower-growing cells that possess the capability to generate bone tissue.¹² When angiogenesis immunomarkers were assessed, a consistent pattern emerged for CD31 and VWF expressions. At the 4-week mark, comparable expression levels were observed across all experimental groups. However, by the 8-week point, a significant 2-fold increase in CD31 and VWF expression was noted in the PEU BM compared to both the SHAM (p = 0.0003/CD31 and p = 0.0004/VWF) and Cytoflex (p = 0.001/CD31 and p = 0.0015/VWF) groups. The time-dependent analysis indicated stable angiogenesis marker expression for the SHAM and Cytoflex groups. However, the PEU BM exhibited a marked increase in expression over the 4 to 8-week period (Figure 8).

Figure 8.

Immunohistochemistry of (A) CD31/PECAM1 and (B) VWF was conducted to evaluate the angiogenesis potential of membranes on rat critical-size calvarial defects. The secondary antibody Alexa Fluor 488 (green) was used for both primary antibodies. Observe the higher expression led by PEU BM for the two markers 8 weeks postimplantation, denoting an outstanding angiogenic behavior. Bar graphs for immunolabeling expressions of CD31/PECAM1 and VWF (mean ± SD), were analyzed using Two-way ANOVA with Sidak’s posthoc test, with P values indicating statistically significant differences.

Optimal angiogenesis properties elicited by the PEU BM play a crucial role in the success of the GBR technique since it actively supports and promotes the formation of new blood vessels in the treated area. This is essential for developing bone tissue with the necessary nutrients and oxygen, accelerating the regenerative process.⁵² The heightened expression in CD31 and VWF after 8 weeks observed in the PEU BM suggests enhanced vascularization, underscoring its clinical potential.

Furthermore, PEU BM demonstrated statistically higher levels of RUNX2 immunolabeling after 8 weeks compared to both SHAM and Cytoflex (p < 0.0001). Additionally, increased expression of OPN was noted at 4 weeks compared to SHAM (p = 0.0162) and Cytoflex (p = 0.0227) and at 8 weeks compared to SHAM (p = 0.0302) postimplantation (Figure 9). These findings underscore PEU’s potential for bone regeneration, suggesting its ability to stimulate osteogenic activity and contribute to successful bone tissue regeneration. Osteopontin, recognized as a multifunctional protein involved in bone mineralization and remodeling, and RUNX2, a crucial transcription factor in osteoblast differentiation, are pivotal markers of osteogenic commitment. The observed elevation in OPN expression implies enhanced bone matrix formation, while the increased expression of RUNX2 indicates osteoblast differentiation and maturation. Consequently, PEU BM in GBR holds promise due to its capacity to create a conducive microenvironment for effective osteogenesis and the regeneration of functional bone tissue.

Figure 9.

Immunohistochemistry of (A) OPN and (B) RUNX2 was conducted to evaluate the osteogenic potential of membranes on rat critical-size calvarial defects. The secondary antibody Alexa Fluor 488 (green) was used for both primary antibodies. Notably, PEU BM membranes showed significantly higher expression of both markers 8 weeks postimplantation, indicating superior osteogenic activity. Bar graphs representing the immunolabeling expressions of OPN and RUNX2 (mean ± SD) were analyzed using Two-way ANOVA with Sidak’s posthoc test, with P values indicating statistically significant differences.

3.5. In Vivo Osteogenesis of PEU (Poly Ester Urea) – Periodontal Defect.

Periodontal surgeries can use BM on GBR to prevent gingival epithelium and connective tissue ingrowth and promote regeneration and augmentation of bone tissue in areas where bone loss has occurred due to periodontal disease or other dental issues. This can be particularly useful in cases where there is insufficient bone volume to support dental implants or other restorative procedures.⁵³ By creating a controlled environment for bone regeneration, GBR enhances the body’s natural healing processes and facilitates the formation of new, functional bone in the periodontal defect, ultimately improving dental interventions’ overall success and longevity.⁵⁴ In our study, we conducted in vivo investigations to assess the performance of the developed PEU BM using a rodent periodontal defect model, allowing us to observe and analyze its behavior in a biologically relevant context (Figure 10A). In addition, valuable information about the handling properties, regenerative performance, and degradation profile of the membrane was acquired.

Figure 10.

In vivo evaluation of PEU BM in a rat periodontal defect model. (A) Photographs showing the surgical procedure depicting the defect creation and placement of the PEU BM before wound closure. (B) MicroCT evaluation of alveolar bone and cementum neoformation for control (no defect), SHAM (defect without treatment), and PEU BM. Note a diminished alveolar bone crest (ABC) to the cementoenamel junction (CEJ) distance in the PEU-treated group compared to the nontreated defect. (C) Bar graph showing the measured CEJ-ABC distance, presented as mean and standard deviation, analyzed through Two-way ANOVA with Tukey posthoc test. P values highlight statistically significant differences.

As noted by the surgical team, the BM demonstrated excellent handling properties, characterized by its ease of manipulation, adaptability to the defect site, and resistance to tearing during placement. Initially, after 4 weeks, no significant differences were found between groups. However, after 8 weeks, the data confirmed that both Cytoflex and PEU BM hindered soft tissue ingrowth into the defect while stimulating in vivo progenitor cells to regenerate new alveolar bone and cementum, with PEU BM recovering ca. 31% of the alveolar bone loss compared to sham animals, as evidenced by the measured CEJ-ABC distance, p = 0.0086 (Figure 10B–C). Histologic staining of the PEU BM, employing both H&E and Masson’s Trichrome, exhibited several promising outcomes. Notably, there was observable evidence indicating the preservation of the gingival sulcus. Moreover, there was clear evidence of new bone formation adjacent to the created defect, signifying a positive response to the PEU BM. Additionally, the presence of Sharpey’s fibers was noted, contributing to anchoring the newly formed alveolar bone to the cementum. The membranes exhibited consistent degradation throughout the study, corroborating our in vitro findings. After 8 weeks, the membranes effectively prevent the penetration of soft tissue cells, thereby supporting bone healing and regeneration (Figure 11).

Figure 11.

In vivo evaluation of PEU BM in a rat periodontal defect model at 8 weeks postimplantation, as demonstrated by Hematoxylin & Eosin histologic staining. The dashed yellow square indicates the area selected for higher magnification. Key features include membrane remnants (m), new bone formation (nb), original bone (ob), and Sharpey’s fibers (yellow arrow). Images are presented at 4× and 10× magnification, with scale bars of 300 and 200 μm, respectively. Note the reduced bone recovery in the SHAM group and the intensified inflammatory reaction associated with Cytoflex compared to PEU.

The observed mild bone regeneration can be attributed to two key factors. First, the typical time frame for complete bone augmentation in periodontal surgery ranges from 16 to 24 weeks, and the relatively short duration of our study may not have covered the full extent of regenerative potential.⁵⁴ Second, it is crucial to note that in clinical practice, BMs for GBR are often coupled with bone graft materials, a component not incorporated into our model. While acknowledging that our model might not precisely replicate optimal conditions regarding diminished size, the findings suggest a promising outlook for PEU BM success. Scaling up our present study into a larger animal model (e.g., canine or porcine) before thoroughly investigating its efficacy in a clinical trial in humans would better predict the safety and performance of our PEU BM in guided bone regeneration.

4. CONCLUSIONS

This work successfully developed a collagen-free membrane via electrospinning based on l-valine-co-l-phenylalanine-poly(ester urea) (PEU) copolymer. Compared to a clinically available synthetic polymeric membrane made of PGA/PLA copolymer (Cytoflex), our PEU-based BM demonstrated equal biocompatibility to Cytoflex with superior mechanical performance in strength and elasticity. Additionally, after 14 days, PEU BM exhibited a higher expression of BGLAP/osteocalcin and superior in vivo performance–less inflammation and increased angiogenic capacity. When placed in critical-sized defects in the calvaria of rats, the PEU BM led to robust bone formation with high expression of osteogenesis and angiogenesis markers. Moreover, the developed membrane enhanced alveolar bone and cementum regeneration in an established periodontal model. Our findings demonstrated that the PEU BM exhibits favorable clinical properties, including mechanical stability, cytocompatibility, and facilitated bone formation in vitro and in vivo. This highlights its suitability for GBR in periodontal and craniofacial bone defects.

Supplementary Material

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsami.4c09742.

Supporting Information includes Nuclear magnetic resonance (NMR) spectra for all the 50% PHE6(1-VAL-10) polymer compounds (S1) and additional details about the cell-based experimental setup (S2) (PDF)