A biocompatible, eco-friendly 3D-printed PCL-NIPUA resin for macular bucking devices in myopic tractional maculopathy

Abstract

Pathological myopia, characterized by excessive axial elongation, posterior staphyloma (PS) and retinal-choroidal degeneration, underlies sight-threatening conditions such as myopic traction maculopathy (MTM) and macular atrophy. Macular buckling (MB) surgery has demonstrated significant efficacy in managing MTM-associated pathologies, including macular schisis, macular holes, and macular retinal detachment. However, its widespread clinical adoption remains limited due to suboptimal material properties and geometrical design constraints of existing buckling devices. Advancements in imaging modalities – such as spectral-domain optical coherence tomography (SD-OCT), swept-source OCT (SS-OCT), ultra-widefield fundus imaging, and three-dimensional magnetic resonance imaging (3D-MRI) – have substantially enhanced the understanding and evaluation of MTM biomechanics. The future development of MB may lie in identifying superior biomaterials and creating customized buckling devices tailored to individual ocular morphology. Here, we developed a photo-curable polycaprolactone (PCL)-based non-isocyanate polyurethane acrylate (NIPUA) resin (PCL-NIPUA) for one-piece J-shaped 3D-printed MB devices. Structural characterization via nuclear magnetic resonance (NMR), Fourier transform infrared (FTIR) spectroscopy, and mass spectrometry (MS) confirmed successful synthesis. Photocuring kinetics under 465 nm visible light achieved an optimal storage modulus within 82 s, coupled with a minimal curing shrinkage (7.2%). By tuning flexible PCL segments, the material exhibited a tensile strength of 25 MPa and an elongation at break of 32%, balancing strength and elasticity. The resin demonstrated environmental stability (water contact angle: 72°; water absorption: 8.8%; and thermal decomposition: 420 °C) and exceptional biocompatibility (hemolysis rate <0.5%; cell viability >97%). Liquid crystal display (LCD) 3D printing produced cost-effective MB devices that provided stable scleral indentation in rabbit models, with few complications, and no pathological structural changes in the optic nerve and retina, showing good biocompatibility and safety. This work establishes PCL-NIPUA as a synergistic platform combining green chemistry, rapid photocuring, mechanical adaptability, and biosafety, offering a clinically translatable solution for customized 3D-printed implants in MB surgery.

---

1. Introduction

Pathological myopia (PM) is a vision-threatening disorder characterized by excessive axial elongation, posterior staphyloma (PS) and retinal-choroidal degeneration.¹ A major complication of PM is myopic traction maculopathy (MTM), mainly characterized by retinal schisis, macular schisis, macular retinal detachment, and partial-thickness or full-thickness macular holes (FTMH) in advanced stages, culminating in visual impairment.² The prevalence of PM is expected to rise to 938 million by 2050, with a significant impact particularly in Asia and Southeast Asia.^3–5 Epidemiological studies highlight a 23.0% prevalence rate of MTM in highly myopic individuals, with macular schisis affecting 14.7–34.4% of cases.^6,7

Current clinical management of MTM involves vitrectomy (an internal approach) and posterior scleral reinforcement or macular buckling (an external approach).⁸ Scleral reinforcement acts on the pathological posterior scleral stretching and thinning, but cannot directly reduce the traction forces on the retina in the pathogenesis of MTM. Vitrectomy combined with internal limiting membrane peeling, gas or silicone oil tamponade can alleviate the anterior–posterior traction and tangential force, achieving improvements in both anatomical and functional outcomes. However, for cases complicated by PS, the initial success rate is limited to only 50–73.3%,⁹ which has renewed clinical interest in MB procedures. MB effectively reshapes the macular contour, shortens axial length, and reduces all three tractions (anteroposterior, tangential, and sclera-retina mismatch), especially vitreoretinal traction induced by posterior staphyloma. Its effectiveness has been demonstrated both as a standalone intervention and in combination with vitrectomy.² However, regarding materials, the variety of buckling device materials and geometries each come with trade-offs. For instance, the high elastic modulus of titanium in silicone–titanium composites causes a mechanical mismatch with ocular tissues, leading to uneven stress distribution and postoperative hemorrhage.¹⁰ In terms of procedure, macular buckling does not have a standardized surgical protocol. The presence of technically complex steps creates significant operational challenges, which impede reproducibility.¹¹ The evolution of imaging technologies has established optical coherence tomography (OCT) as the gold standard for assessing surgical outcomes and postoperative monitoring in MB procedures.¹² Using wide-field Optos fundus imaging and three-dimensional magnetic resonance imaging (3D-MRI) has revolutionized the classification of PS by providing comprehensive analysis of global eye morphology. These imaging advancements not only improve the accuracy of intraoperative buckle placement but also emphasize the crucial importance of personalized MB device design.^13–15

The clinical efficacy and safety profile of MB procedures are fundamentally determined by the material properties of the implant. Resin-based materials hold promise for MB therapy but face critical challenges, including cytotoxicity and inadequate toughness.¹⁶ For instance, epoxy acrylate resins, though suitable for high-precision modeling, suffer from high viscosity and brittleness,¹⁷ while conventional polyurethane acrylates (PUAs), despite tunable properties, rely on phosgene-derived synthesis, posing environmental and toxicity risks.¹⁸ Polyurethanes are well-established in biomedical applications (e.g., orthopedics, cardiovascular implants) due to their exceptional biocompatibility and tailorable mechanics.^19,20 However, in ophthalmology, gel-based corneal substitutes lack the mechanical strength required for MB.²¹ Our prior work developed a NIPU resin with excellent biocompatibility and a tensile strength of 28.5 ± 3.2 MPa,²² yet its low elongation at break (8.7%) falls short of natural scleral compliance. Thus, further optimization of NIPU's elasticity and durability is essential to meet the clinical demands of MTM therapy.

Efforts to enhance elongation at break have explored elastomer blending (e.g., nitrile rubber), but these strategies often compromise material strength and induce phase separation.²³ Similarly, nanofiller reinforcement (e.g., graphene, carbon nanotubes) improves mechanical properties but faces challenges such as poor dispersion, high cost, and impaired photocuring efficiency.²⁴ To overcome these limitations, we leverage the molecular design flexibility of polyurethanes by increasing the proportion of soft segments in the polymer chain. Polycaprolactone (PCL), a semicrystalline aliphatic polyester, is an ideal candidate due to its inherent ductility and ability to form interfacial hydrogen bonds with non-isocyanate polyurethane (NIPU) through ester-urethane interactions, ensuring enhanced compatibility.²⁵ By integrating PCL with photocurable 3D printing, we synthesize PCL-NIPUA composites with micron-scale precision, enabling the fabrication of personalized medical implants that balance mechanical resilience and structural complexity for MB applications.

Building upon our developed photocurable PCL-NIPUA system, we demonstrate its successful application through innovative device design and in vivo validation. The engineered PCL-NIPUA resin combines superior mechanical compatibility (11.85 MPa tensile strength, 29.75% elongation at break) with rapid visible-light curing (82 s at 465 nm) and exceptional biocompatibility (hemolysis rate <0.5%, cell viability >97%), fully meeting the material requirements for MB devices. Combining 3D ocular imaging with photocurable 3D-printing technology, the devices achieve dynamic elastic modulus matching with natural ocular tissues while demonstrating excellent structural stability (7.2% curing shrinkage) and tissue compatibility (confirmed by H&E staining). These findings establish a comprehensive solution that simultaneously resolves the tripartite challenge of biocompatibility concerns, mechanical mismatch, and manufacturing limitations in MTM treatment. By synergizing green chemistry principles (phosgene-free synthesis), advanced material design (PCL soft segment tuning), and digital light processing manufacturing, our work not only provides a clinically translatable approach for myopic traction maculopathy but also introduces a new paradigm for patient-specific ophthalmic implants. The complete technical roadmap of this integrated strategy is presented in Fig. 1, illustrating the logical progression from molecular design to functional validation.

Fig. 1 Schematic overview of PCL-NIPUA synthesis, device fabrication, and in vivo evaluation. (A) The PCL-NIPUA resin was synthesized via a two-step process involving a Michael addition reaction followed by methacrylation modification to introduce photocurable groups. (B) The photosensitive resin was then processed using LCD-based 3D printing technology to fabricate customized posterior scleral reinforcement devices with micro-architected designs. (C) In vivo validation in rabbit models demonstrated the device's therapeutic efficacy in stabilizing myopic traction maculopathy and its excellent biocompatibility through histological assessment. The integrated workflow highlights the material's tunable mechanical properties, rapid manufacturing potential, and clinical applicability for ocular reinforcement therapies.

2. Experimental

2.1. Materials and reagents

All chemical reagents were obtained from commercial sources and used without further purification unless otherwise specified. Ethylene carbonate (99%), isophorone diamine (98%), ε-caprolactone (99%), methacrylic anhydride (94%), triethylamine (99%), and stannous octoate (95%) were purchased from Aladdin Reagent Co., Ltd (Shanghai, China) for polymer synthesis. Solvents and purification agents, including dichloromethane (HPLC grade), n-hexane (97%), sodium bicarbonate (99.5%), sodium chloride (99.5%), dilute hydrochloric acid (36–38%), and anhydrous magnesium sulfate (99%), were also obtained from the same supplier. The photocuring system components consisted of triethylene glycol dimethacrylate (TEGDMA, 90%), polyethylene glycol diacrylate 1000 (PEGDA 1000, 95%), and camphorquinone (CQ, 98%) from Aladdin.

Cell culture reagents were acquired from Biyun Tian Biotechnology Co., Ltd (Shanghai, China), including high-glucose DMEM medium and phosphate-buffered saline (PBS, pH 7.4) for cell maintenance. Cell viability was assessed using a CCK-8 kit and a live/dead cell staining kit (AM/PI), while culture conditions were maintained with fetal bovine serum (FBS, South American origin) and penicillin/streptomycin solution (10 000 U mL⁻¹). Mouse fibroblast cells (L929 line) were used for biocompatibility testing. All aqueous solutions were prepared using deionized water (18.2 MΩ cm resistivity) produced by a laboratory ultrapure water system (Milli-Q, Merck).

2.2. Synthesis and printing of non-isocyanate photocurable resins

2.2.1. Synthesis of the PCL-NIPU precursor. The synthesis was conducted under a nitrogen atmosphere using standard Schlenk techniques. Ethylene carbonate (quantified by titration) was charged into a flame-dried, three-neck round-bottom flask equipped with a magnetic stirrer and maintained at 120 °C in a silicone oil bath. Isophorone diamine, loaded into an addition funnel, was introduced dropwise at a controlled rate of 1 drop/3 seconds to ensure proper reaction kinetics. This initial step proceeded for 8 hours with continuous stirring to complete the ring-opening addition reaction.

Following the initial reaction stage, ε-caprolactone and the stannous octoate catalyst (0.1 wt% relative to monomers) were introduced through the addition funnel. The reaction temperature was subsequently elevated to 130 °C and maintained for 72 hours to facilitate the polymerization process. The system was then gradually cooled to ambient temperature (25 °C) at a rate of 2 °C min⁻¹ to prevent thermal stress on the product.

The crude product was dissolved in anhydrous dichloromethane (DCM, 3 : 1 v/v ratio) and subjected to sequential liquid–liquid extraction with n-hexane (3 × 300 mL for 100 g product) to remove unreacted monomers and oligomeric species. Each extraction step involved vigorous shaking for 5 minutes followed by 30 minutes of phase separation. The purified product was obtained by rotary evaporation under reduced pressure (50 mbar at 40 °C) until constant weight was achieved, yielding PCL-NIPU as a colorless, viscous liquid with >98% purity according to HPLC analysis.

2.2.2. Methacrylation of PCL-NIPU to synthesize PCL-NIPUA. The methacrylation of purified PCL-NIPU was conducted under a nitrogen atmosphere in a three-neck flask equipped with a mechanical stirrer and a dropping funnel. Initially, PCL-NIPU was dissolved in anhydrous dichloromethane at a concentration of 40% (w/v), followed by the addition of triethylamine as a catalyst at a 1.2 : 1 molar ratio relative to hydroxyl groups. The reaction vessel was immersed in an ice-water bath to maintain the temperature at 0–4 °C during the subsequent addition of methacrylic anhydride, which was introduced dropwise at a controlled rate of 0.5 mL min⁻¹ while maintaining a 1.5 : 1 molar excess of methacrylic anhydride to hydroxyl groups.

After complete addition, the reaction mixture was allowed to warm gradually to room temperature (25 ± 2 °C) and stirred continuously for 24 hours under sealed conditions to ensure complete functionalization. The reaction was then carefully quenched by the addition of chilled saturated sodium chloride solution, maintaining the temperature below 10 °C to prevent side reactions. The organic layer was separated and subjected to a sequential washing protocol, first with 0.1 M hydrochloric acid to remove residual triethylamine, followed by saturated sodium bicarbonate solution to neutralize acidic byproducts, then 5% sodium chloride solution for desalting, and finally multiple washes with deionized water until neutral pH was achieved.

The washed organic phase was dried over anhydrous magnesium sulfate for 12 hours, filtered through a 0.45 μm PTFE membrane, and concentrated by rotary evaporation at 30 °C under reduced pressure. This procedure yielded PCL-NIPUA as a clear, viscous oil with greater than 95% methacrylation efficiency as confirmed by ¹H-NMR spectroscopy. The final product was stored under a nitrogen atmosphere at −20 °C to prevent moisture absorption and premature polymerization until further use in subsequent experiments.

2.2.3. Formulation and 3D printing of photocurable resin. The photocurable resin system was prepared by combining PCL-NIPUA with reactive diluents triethylene glycol dimethacrylate (TEGDMA) and poly(ethylene glycol)diacrylate (PEGDA 1000) in optimized ratios as specified in Table 1. To initiate photopolymerization, camphorquinone (CQ) was incorporated at 1 wt% as the photoinitiator, along with 0.07 wt% β-carotene as a light-absorbing additive to enhance printing resolution by reducing the light penetration depth. The components were thoroughly mixed under amber light conditions to prevent premature curing until a homogeneous resin formulation was achieved.

Table 1 Printing ink formulation

	PCLn-NIPUA	TEGDMA	PEGDA	CQ	β-carotene
Tips: n = 2, 4, 8; the combined mass of the first three acrylate esters is 100, and the addition of CQ and β-carotene is calculated based on this 100.
6 : 4 : 0	60	40	0	1	0.07
6 : 3 : 1	60	30	10	1	0.07
6 : 2 : 2	60	20	20	1	0.07

---

For digital light processing (DLP)-based 3D printing, the formulated resin was processed using an LCD-based printing system with carefully optimized parameters. The printing configuration included a layer thickness of 50 μm and an exposure time of 30 seconds per layer, which were determined through preliminary tests to ensure complete curing while maintaining dimensional accuracy. The printing process was conducted under controlled ambient conditions (23 ± 2 °C, <30% relative humidity) to minimize environmental effects on resin viscosity and curing kinetics. Post-printing, all fabricated devices were immediately rinsed with isopropanol to remove uncured resin residues and subsequently post-cured under 405 nm LED light (10 mW cm⁻²) for 10 minutes to ensure complete polymerization of all reactive groups.

The MB device has a one-piece J-shaped design, divided into anterior and posterior sections. The anterior part features three pairs of suture holes for fixation, while the posterior part has a raised hemispherical surface that adheres to the sclera to create indentation. This monolithic design facilitates single-step implantation, simplifying surgical procedures. Advancements in 3D-MRI imaging of posterior staphyloma and computational algorithms will enable us to create personalized devices that match individual eye anatomy including changes to the curvature, the size of the indentation surface, and the overall device length.

2.3. Characterization of NIPUA synthesis

2.3.1. Nuclear magnetic resonance (NMR) spectroscopy. Structural characterization of purified PCL-NIPU and PCL-NIPUA was performed using ¹H-NMR spectroscopy. Samples (10 mg each) were dissolved in 0.6 mL of deuterated dimethyl sulfoxide (DMSO-d₆) and transferred to standard 5-mm NMR tubes. All spectra were acquired on a Bruker AVANCE III 600 MHz spectrometer operating at 25 °C, with 32 scans collected for each sample to ensure an adequate signal-to-noise ratio. Chemical shifts were referenced to the residual solvent peak (DMSO-d₆ at 2.50 ppm) and reported in parts per million (ppm) relative to tetramethylsilane (TMS) as internal standard. Data processing included Fourier transformation, phase correction, and baseline adjustment using TopSpin 4.0 software.

2.3.2. Fourier-transform infrared (FTIR) spectroscopy. Molecular functional group analysis was conducted using a Nicolet IS50 FTIR spectrometer (Thermo Fisher Scientific) equipped with a diamond attenuated total reflectance (ATR) accessory. Approximately 1 mg of each viscous sample was applied directly onto the ATR crystal and compressed with a consistent 10 N clamping force to ensure proper optical contact. Spectra were collected in the mid-infrared region (4000–400 cm⁻¹) with a resolution of 4 cm⁻¹, accumulating 64 scans per measurement to enhance spectral quality. Prior to each sample analysis, a background spectrum of ambient air was recorded under identical instrumental conditions and automatically subtracted from sample spectra during data collection.

2.3.3. Mass spectrometry analysis. Molecular weight determination and structural confirmation were performed using high-resolution mass spectrometry (HRMS) with an Agilent 1290 HPLC system coupled to a Bruker maXis impact Q-TOF mass spectrometer. Sample solutions were prepared at 100 μg mL⁻¹ concentration in HPLC-grade methanol, with 0.1 M sodium chloride added as a cationization reagent to facilitate ionization. Mass spectra were acquired in positive ion mode over the m/z range of 500–5000, using an electrospray ionization (ESI) source operated at 25 kV acceleration voltage. Instrument calibration was performed daily using sodium formate clusters, and data analysis was conducted with Bruker DataAnalysis 4.3 software.

2.4. Performance testing of PCL-NIPUA

2.4.1. Physicochemical property testing.
2.4.1.1. Ultraviolet-visible (UV-vis) spectroscopy. The optical absorption properties of PCL-NIPUA were analyzed using a Shimadzu UV-2600 spectrophotometer (Shimadzu Corporation, Kyoto, Japan). Samples were prepared as 1 mg mL⁻¹ solutions in HPLC-grade dichloromethane (DCM), which also served as the blank reference for baseline correction. Measurements were performed in standard 1 cm path-length quartz cuvettes with 2 mL sample volume, scanning across the 200–800 nm wavelength range at 200 nm min⁻¹ with 0.1 nm resolution. Three replicate scans were acquired for each sample to ensure data reproducibility, with spectral analysis performed using UVProbe 2.42 software.
2.4.1.2. Rheological characterization. The viscosity of photocurable resin formulations was determined using a TA DHR-20 rheometer (TA Instruments, USA) with parallel plate geometry (40 mm diameter, 500 μm gap). Measurements were conducted at 25.0 ± 0.1 °C with a constant shear rate of 100 rpm, following a 5-minute equilibration period to ensure thermal stability. Three independent measurements were performed for each formulation, with data collected over 300 seconds to verify measurement stability.
2.4.1.3. Photopolymerization kinetics. Real-time photocuring behavior was monitored using a combination of 465 nm LED irradiation (8 mW cm⁻² intensity, calibrated with a radiometer) and rheological analysis. Samples (0.2 mm thickness) were allowed to equilibrate for 30 seconds before irradiation to eliminate mechanical disturbances. The evolution of storage modulus (G′) was tracked using the TA DHR-20 rheometer equipped with a UV accessory, operating at 1 Hz frequency and 1% strain. The curing endpoint was defined as the time point when G′ reached 95% of its maximum plateau value, with measurements performed in triplicate.
2.4.1.4. Curing shrinkage determination. Volumetric shrinkage during photopolymerization was quantified using a density-based method. Uncured resin density (ρ₁) was measured via pycnometry at 25 °C using a 10 mL specific gravity bottle, with five replicate measurements. Cured sample density (ρ₂) was determined geometrically using an analytical balance (AX224ZH, ±0.0001 g) and a digital caliper (±0.01 mm) for mass and volume measurements, respectively. The percentage shrinkage was calculated using [(ρ₂ − ρ₁)/ρ₁] × 100%, with ten independent measurements performed to establish statistical significance.
2.4.1.5. Mechanical property evaluation. The mechanical performance of cured resin samples was comprehensively evaluated through standardized testing protocols. For tensile property assessment, Type V specimens were fabricated according to ASTM D638 specifications and tested using an Instron 5967 universal testing system (Norwood, MA, USA) with a 30 kN load cell. Tests were conducted at a controlled crosshead speed of 2 mm min⁻¹ under ambient conditions (25 ± 1 °C), with strain measurements acquired via an extensometer. Flexural properties were determined following ASTM D790-03 standards using the same testing platform, employing a three-point bending configuration with 40 mm support span. Impact resistance was quantified using a CEAST 9050 pendulum impact tester (Instron CEAST, Italy) operating at 15 J hammer energy and 3.5 m s⁻¹ impact velocity, with specimens prepared according to GB/T 1843-2008 standards. All mechanical tests were performed with n ≥ 5 replicates to ensure statistical significance.
2.4.1.6. Hardness measurement. Surface hardness was characterized using a Shore D durometer (Model LX-A-Y) in compliance with GB/T 531 standards. Measurements were conducted at three spatially separated locations (minimum 6 mm spacing) on polished sample surfaces to account for potential heterogeneity. The durometer was maintained perpendicular to the surface with consistent application force, and readings were recorded after a 15-s dwell time to ensure stabilization. Reported values represent the mean of five independent measurements, with a standard deviation of less than 1.5 Shore D units across all samples.
2.4.1.7. Thermal behavior analysis. Thermal stability and transition behavior were investigated using a NETZSCH STA 449 F5 simultaneous thermal analyzer. Approximately 10 mg samples were analyzed in alumina crucibles under 50 mL min⁻¹ nitrogen purge, with temperature ramped from 25 °C to 600 °C at 10 °C min⁻¹. The instrument simultaneously recorded thermogravimetric (TG), derivative thermogravimetric (DTG), and differential scanning calorimetry (DSC) data. Temperature and enthalpy calibrations were performed using indium and zinc standards prior to sample analysis. Three replicate measurements were conducted for each resin formulation to verify reproducibility.
2.4.1.8. Surface wettability assessment. Surface hydrophilicity was quantified through static contact angle measurements using a DataPhysics OCA 25 goniometer. Samples were ultrasonically cleaned with ethanol and dried under nitrogen prior to testing. Deionized water droplets (3 μL) were dispensed via an automated syringe, with images captured at 1 s intervals for 30 s using a high-speed CCD. Contact angles were determined via Young-Laplace fitting with SCA 20 software, reporting the average of ten measurements per sample. All tests were conducted under controlled conditions (23 °C, 50% RH) following 24 h sample acclimatization.
2.4.1.9. Swelling behavior analysis. The hydration characteristics of cured resins were evaluated through equilibrium swelling studies. Pre-weighed samples (n = 9 per group) were immersed in either deionized water or PBS (pH 7.4) at 25 °C. At 24 h intervals, samples were removed, surface-blotted with filter paper, and weighed to determine mass change. The swelling ratio was calculated as [(m₂ − m₁)/m₁] × 100%, where m₁ and m₂ represent pre- and post-immersion masses, respectively. Measurements continued until equilibrium swelling was achieved (typically 72 h), with solution replacement every 24 h to maintain consistent osmotic conditions.

2.4.2. Biocompatibility testing of PCL-NIPUA.
2.4.2.1. Hemolysis assay. Fresh rabbit blood was collected in heparinized tubes and processed within 2 hours of collection. The blood was carefully mixed with an equal volume of phosphate-buffered saline (PBS, pH 7.4) and centrifuged at 1500 × g for 10 minutes at 4 °C. This washing procedure was repeated three times until the supernatant appeared completely clear. The purified erythrocytes were then resuspended in PBS to prepare a 10% (v/v) working suspension.

For testing, 200 μL of the erythrocyte suspension was combined with 200 μL of resin-PBS extract (prepared according to ISO 10993-12 guidelines by incubating 3 cm² mL⁻¹ of the material in PBS at 37 °C for 72 hours). Negative controls consisted of erythrocytes mixed with PBS alone, while positive controls contained erythrocytes lysed with 5% (v/v) deionized water. All test and control samples were incubated at 37 ± 0.5 °C for 90 minutes in a temperature-controlled water bath with gentle agitation every 30 minutes.

Following incubation, samples were centrifuged at 1000 × g for 5 minutes to pellet intact erythrocytes. The supernatant was carefully aspirated and visually inspected for hemoglobin release. For quantitative analysis, 80 μL aliquots of the supernatant were transferred in triplicate to a 96-well plate, and absorbance at 540 nm was measured using a BioTek Synergy H1 microplate reader (BioTek Instruments, Winooski, VT, USA). The hemolysis percentage was calculated using formula (1):

2.4.2.2. Cytotoxicity assays. L929 mouse fibroblast cells were cultured in high-glucose DMEM supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin at 37 °C in a 5% CO₂ humidified incubator. For cytotoxicity assessment, cells were seeded in 96-well plates at a density of 1 × 10⁴ cells per well and allowed to adhere for 24 hours. Material extracts were prepared by incubating sterile PCL2-NIPUA, PCL4-NIPUA, and PCL8-NIPUA samples (surface area: volume ratio of 3 cm² mL⁻¹) in complete culture medium at 37 °C for 24 hours according to ISO 10993-5 guidelines. Five replicate wells were used for each experimental group.

Cell viability was quantitatively assessed using the CCK-8 assay at 24, 48, and 72-hour time points. At each interval, culture medium was replaced with 100 μL of fresh medium containing 10% CCK-8 reagent, followed by incubation for 2 hours at 37 °C. Absorbance was measured at 450 nm using a microplate reader, with 650 nm as the reference wavelength. Growth curves were generated by normalizing optical density values to day 0 controls.

Qualitative viability assessment was performed using calcein-AM/propidium iodide (PI) dual staining. Following extract exposure for 24–72 hours, cells were washed with PBS and incubated with staining solution (2 μM calcein-AM and 4 μM PI in PBS) for 15 minutes at room temperature protected from light. Fluorescence images were acquired using an inverted microscope (20× objective) with FITC (ex 488 nm/em 515 nm) and TRITC (ex 543 nm/em 572 nm) filter sets for live (green) and dead (red) cell visualization, respectively. Three random fields per well were imaged for statistical analysis.

2.5. Animal experiment

2.5.1. Surgical procedure. The study utilized healthy male New Zealand white rabbits (8 weeks old, 1.0–1.5 kg) with normal ocular examinations. Surgical procedures were conducted under aseptic conditions following ethical guidelines approved by the Institutional Animal Care and Use Committee. Anesthesia was achieved through intramuscular xylazine hydrochloride injection supplemented with topical oxybuprocaine eye drops. The left eye served as the experimental side, with the right eye as an internal control.

A limbal-based conjunctival incision was made in the superonasal quadrant. The medial and superior rectus muscles were isolated through blunt dissection and secured with traction sutures. The tenon's capsule was dissected posteriorly behind the equator. A custom-designed device was placed on the sclera, ensuring that its posterior indentation segment avoided the vortex veins and the optic nerve. Fixation was achieved by passing a 5–0 nonabsorbable suture through anterior anchoring holes and securing it to the sclera. Before closing the conjunctiva, intraocular pressure (IOP) was lowered through anterior chamber paracentesis. The conjunctiva was sutured with 8–0 absorbable sutures. After surgery, antibiotic ointment was applied to the conjunctival sac. All procedures adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The protocol was approved by the Institutional Animal Care and Use Committee of Peking Union Medical College Hospital (Approval No. XHDW-2025-035).

2.5.2. Postoperative fundus examination. Comprehensive follow-up evaluations were conducted at predetermined intervals after surgery. The assessment protocol included general health monitoring and detailed ocular examinations. External eye documentation was conducted using standardized digital photography. During each follow-up visit, the rabbits were anesthetized, and their pupils were dilated with 1% tropicamide eye drops. Ultra-widefield (UWF) fundus imaging was acquired using OPTOS® Daytona camera. High-resolution cross-sectional analysis of the device-tissue interface was obtained through swept-source optical coherence tomography (SS-OCT, VG 200D; Luoyang, Henan, China) using specialized scanning protocols.

2.5.3. Histopathological analysis. After a six-month experimental period, animals were humanely euthanized, and eyes were removed for histological processing. Immediately after the device was removed, ocular specimens were fixed in a paraformaldehyde solution to ensure optimal tissue preservation. After routine paraffin embedding, 4-μm-thick sections were cut along the device compression axis. The staining protocol included sequential deparaffinization, hematoxylin-eosin staining, and careful mounting. The prepared tissue sections were examined by light microscopy, focusing particularly on tissue-device interactions, inflammatory responses and structural integrity in the optic nerve and indentation retinal region.

All experimental procedures strictly adhered to the ARVO Statement for animal research and were approved by the institutional ethics committee. The study design included various quality control measures, including standardized surgical techniques, calibrated imaging protocols, and masked histopathological evaluation, to ensure data reliability and reproducibility. The comprehensive assessment strategy allowed for a thorough evaluation of both device performance and tissue response throughout the follow-up period.

2.6. Statistical analysis

All quantitative data were analyzed using SPSS 26.0 (IBM Corp., USA) and GraphPad Prism 9.0 (GraphPad Software, USA). Continuous variables were assessed for normality using the Shapiro-Wilk test. Normally distributed data were expressed as mean ± standard deviation (SD) and analyzed using one-way ANOVA with Tukey's post hoc test for multiple comparisons. Non-normally distributed data were presented as median (interquartile range, IQR) and analyzed using the Kruskal–Wallis test followed by Dunn's post hoc correction.

For repeated-measures data (e.g., longitudinal imaging parameters), a mixed-effects model was applied with time and the treatment group as fixed effects and individual animals as random effects. Categorical variables (e.g., histopathological scores) were compared using the chi-square test or Fisher's exact test, as appropriate.

Correlation analyses between the material properties and biological responses were performed using Pearson's or Spearman's methods based on data distribution. Intraclass correlation coefficients (ICC) were calculated to assess inter-observer agreement for histological and imaging evaluations. Statistical significance was set at p < 0.05 (two-tailed). Sample size calculations were based on preliminary experiments to achieve 80% power at α = 0.05.

All statistical tests were pre-specified in the study protocol, and no data exclusions were performed unless due to technical failures. Adjustments for multiple comparisons were applied where appropriate using the Benjamini–Hochberg false discovery rate (FDR) method. Effect sizes were reported alongside p-values for major comparisons.

3. Results and discussion

3.1. Synthesis of photocurable NIPUA resin

NIPUA resin was successfully developed through an environmentally friendly synthetic route. The synthesis strategy deliberately avoided conventional phosgene-derived isocyanates, instead employed a green chemistry approach using ethylene carbonate and isophorone diamine as starting materials. This phosgene-free methodology significantly reduces the environmental impact and toxicity concerns associated with traditional polyurethane synthesis.

The synthetic pathway (Fig. 2) involved three key steps: first, the reaction between ethylene carbonate and isophorone diamine produced a polyurethane diol intermediate. Subsequently, ε-caprolactone was grafted onto this intermediate via ring-opening polymerization, effectively extending the polymer chains and introducing flexible aliphatic ester segments. Finally, methacrylic anhydride was employed to functionalize the terminal hydroxyl groups with photocrosslinkable methacrylate moieties, enabling the material's photocuring capability for 3D printing applications.

Fig. 2 PCL-NIPUA synthesis route.

3.1.1. Structural characterization of PCL-NIPU. The synthesis of PCL-NIPU was successfully achieved through a sequential reaction process, as confirmed by comprehensive NMR and FTIR analyses. In the ¹H-NMR spectrum (Fig. 3A–C), the characteristic hydroxyl end-group proton resonance appeared at 4.68 ppm, serving as a key indicator of the terminal functionality. The successful grafting of PCL segments was unambiguously demonstrated by five distinct proton resonance peaks at 1.28 ppm (peak a, –CH₂–), 1.40 ppm (peak b, –CH₂–), 1.52 ppm (peak c, –CH₂–), 2.28 ppm (peak d, –CH₂COO–), and 3.98 ppm (peak e, –CH₂O–), corresponding to the methylene protons along the PCL chain. Quantitative analysis of these peaks relative to the polyurethane diol backbone protons (28 equivalent hydrogens) enabled precise determination of PCL grafting density, allowing for controlled synthesis of PCL2-NIPU, PCL4-NIPU, and PCL8-NIPU variants with increasing caprolactone content.

Fig. 3 Structural characterization of PCL-NIPU polymers by NMR and FTIR spectroscopy. (A–C) ¹H-NMR spectra (600 MHz, DMSO-d₆) of (A) PCL2-NIPU, (B) PCL4-NIPU, and (C) PCL8-NIPU, showing characteristic proton resonances of the polyurethane backbone (δ 4.68 ppm, –OH) and grafted polycaprolactone segments (peaks a–e: δ 1.28, 1.40, 1.52, 2.28, and 3.98 ppm). (D) FTIR spectra demonstrating key functional groups: urethane N–H stretch (3350 cm⁻¹), urea C=O stretch (1703 cm⁻¹), and PCL ester carbonyl stretch (1735 cm⁻¹), confirming successful polymer synthesis. Peak intensity variations in (A–C) reflect increasing PCL content (2, 4, and 8 units per chain, respectively).

Complementary FTIR analysis (Fig. 3D) provided further structural verification through characteristic vibrational modes. The ester carbonyl stretching of PCL segments appeared as a strong absorption band at 1735 cm⁻¹, while the asymmetric C–O–C stretching vibration was observed at 1255 cm⁻¹, confirming successful PCL incorporation. The NIPU backbone structure was evidenced by two diagnostic absorption bands: a broad N–H stretching vibration band at 3350 cm⁻¹, indicative of urethane hydrogen bonding, and a sharp carbamate carbonyl stretch band at 1703 cm⁻¹. The absence of isocyanate-derived peaks (2250–2275 cm⁻¹) confirmed the complete non-isocyanate nature of the synthetic route. These spectroscopic findings collectively demonstrate the successful synthesis of well-defined PCL-NIPU polymers with a controlled architecture and composition, establishing a robust foundation for subsequent methacrylation and photocuring studies.

3.1.2. Synthesis and characterization of photocurable PCL-NIPUA. The successful functionalization of PCL-NIPU with photoreactive methacrylate groups was confirmed through comprehensive spectroscopic and mass spectrometric analyses. Comparative ¹H-NMR analysis revealed the appearance of two new characteristic peaks at 5.66 ppm and 6.00 ppm (Fig. 4A), corresponding to the vinylic protons of the terminal methacrylate groups. These diagnostic resonances, absent in the precursor PCL-NIPU spectra (Fig. 3A–C), provided direct evidence of successful end-group modification.

Fig. 4 Structural characterization of photocurable PCL-NIPUA. (A) ¹H-NMR spectrum (600 MHz, DMSO-d₆) showing characteristic methacrylate proton resonances at δ 5.66 and 6.00 ppm (highlighted in red). (B) FTIR spectroscopy confirms that methacrylate undergoes C–H bending (943 cm⁻¹) vibration through C=C connection. (C–E) High-resolution mass spectra of (C) PCL2-NIPUA (observed m/z 711.41, theoretical 710.23), (D) PCL4-NIPUA (observed m/z 945.29, theoretical 934.81), and (E) PCL8-NIPUA (observed m/z 1332.76, theoretical 1388.53), demonstrating successful synthesis of target compounds with <0.5% mass error. All spectra were baseline-corrected and normalized for comparative analysis.

FTIR spectroscopy further verified the chemical transformation. Notably, the 1640 cm⁻¹ peak observed in both PCL-NIPU and PCL-NIPUA arises from different origins: in PCL-NIPU, it is attributed to trace urea groups formed by side reactions of amino groups with residual CO₂ during synthesis; in PCL-NIPUA, it is a superposition of residual urea carbonyl and the C=C stretching vibration of methacrylate groups. A more reliable indicator of methacrylate functionalization is the emergence of a C–H out-of-plane bending mode at 943 cm⁻¹ (originally included but supplemented with explanation)—this peak is exclusively present in PCL-NIPUA (not detected in PCL-NIPU) and is characteristic of the photoreactive C=C bonds in methacrylate groups^26,27 (Fig. 4B). These spectral features confirm the preservation of photoreactive double bonds after purification.

High-resolution mass spectrometry provided additional confirmation of the target molecular structures. The observed mass-to-charge ratios of 711.41, 945.29, and 1332.76 Da showed excellent agreement with the theoretical values of 710.23, 934.81, and 1388.53 Da for PCL2-, PCL4-, and PCL8-NIPUA, respectively (Fig. 4C–E). The minor deviations (<0.5%) fell within the instrument's margin of error and likely resulted from natural isotopic distributions. This multi-technique characterization unambiguously demonstrated the successful synthesis of photocurable PCL-NIPUA with well-defined molecular architectures suitable for subsequent 3D printing applications.

3.2. Performance regulation of PCL n-NIPUA

This section focuses on evaluating the photocuring performance, mechanical properties, thermal stability, and biocompatibility of the resin system, with the core goal of screening formulations with optimal comprehensive performance to support subsequent in vivo experiments.

3.2.1. Photocuring characteristics: key to 3D printing adaptability. Photocuring-related properties were analyzed via UV-vis spectroscopy and rheology, with a focus on parameters critical for 3D printing (viscosity, curing efficiency, and volumetric shrinkage): UV-vis spectroscopy confirmed the resin's suitability for visible-light initiation, a strong π → π* transition at 260 nm (validated acrylate groups) and characteristic absorption of camphorquinone at 465 nm (Fig. 5A).

Fig. 5 The key physical and chemical properties of PCL-NIPUA resin with a formula of 6 : 3 : 1. (A) UV-vis absorption spectrum showing characteristic acrylate (260 nm) and camphorquinone photoinitiator (465 nm) peaks. (B) Photorheometry kinetics tracking storage modulus (G′) development during 465 nm irradiation (8 mW cm⁻²), with PCL4-NIPUA reaching maximum modulus at 82 s. (C) Stress–strain curves of three types of resins. (D) Thermogravimetric analysis (TGA) curves showing thermal decomposition profiles. (E) Derivative thermogravimetry (DTG) curves indicate maximum decomposition rates. (F) Differential scanning calorimetry (DSC) thermograms reveal thermal transitions. Error bars represent standard deviation (n = 3).

Viscosity regulation was achieved via molecular design and ternary formulation (TEGDMA as viscosity modifier, PEGDA 1000 as toughening agent; Table 1, Row 1). Undiluted PCL-NIPUA viscosity decreased with longer PCL chains (1703.68 mPa s for PCL2 vs. 689.07 mPa s for PCL8 at 25 °C), while the 6 : 3 : 1 formulation reduced viscosity to <200 mPa s (90% lower than base resin) without sacrificing mechanical performance—meeting practical printing requirements.

3.2.1.1. Curing efficiency and shrinkage. Under 465 nm irradiation (8 mW cm⁻²), PCL4-NIPUA reached a maximum storage modulus (G′ = 1.2 GPa) in 82 seconds (30% higher than conventional acrylate resins; Fig. 5B), benefiting from balanced molecular mobility and crosslink density. Volumetric shrinkage (3–8%, Table 2, row 2) was reduced by 40% for PCL8 vs. PCL2, as long PCL chains alleviated molecular packing constraints during network formation.

Table 2 Key physical and mechanical performance parameters of the PCL-NIPUA resin and formula

Resin type	PCL2-NIPUA				PCL4-NIPUA				PCL8-NIPUA
Formula	10 : 0 : 0	6 : 4 : 0	6 : 3 : 1	6 : 2 : 2	10 : 0 : 0	6 : 4 : 0	6 : 3 : 1	6 : 2 : 2	10 : 0 : 0	6 : 4 : 0	6 : 3 : 1	6 : 2 : 2
Viscosity/mPa s	1703.68 ± 23.77	121.30 ± 0.05	142.68 ± 0.27	187.68 ± 0.12	1431.13 ± 5.52	92.65 ± 0.21	171.78 ± 0.14	185.26 ± 0.14	689.07 ± 0.09	73.00 ± 0.66	88.06 ± 0.71	141.14 ± 0.19
Curing shrinkage rate/%	—	8.04 ± 1.22	7.35 ± 2.64	7.38 ± 2.21	—	7.57 ± 2.14	7.18 ± 2.10	7.06 ± 1.75	—	4.05 ± 1.61	3.19 ± 0.91	3.27 ± 0.39
Tensile Strength/MPa	14.68 ± 1.52	28.92 ± 0.89	26.05 ± 2.57	24.88 ± 3.48	3.27 ± 0.51	13.58 ± 0.83	11.85 ± 1.75	4.32 ± 0.94	0.40 ± 0.13	26.70 ± 5.17	25.67 ± 7.29	22.90 ± 5.09
Elongation at break/%	45.12 ± 8.97	11.87 ± 3.12	24.44 ± 4.48	12.86 ± 1.74	22.36 ± 2.93	16.56 ± 1.49	29.76 ± 2.41	11.56 ± 2.71	14.72 ± 1.30	16.10 ± 1.43	23.30 ± 6.66	37.54 ± 5.97
Impact Strength/kJ m^−2	1.38 ± 0.19	1.84 ± 0.24	2.10 ± 1.00	1.27 ± 0.35	5.97 ± 1.73	3.22 ± 1.69	7.64 ± 1.10	2.38 ± 0.32	2.63 ± 0.24	1.56 ± 0.29	4.52 ± 2.77	7.44 ± 1.04
Flexural strength/MPa	7.95 ± 1.45	58.14 ± 8.67	32.91 ± 4.76	18.43 ± 8.47	12.14 ± 1.76	28.34 ± 1.60	26.61 ± 19.26	23.89 ± 2.37	8.81 ± 2.12	48.51 ± 11.50	35.80 ± 3.53	21.82 ± 3.33
Flexural modulus/MPa	143.40 ± 13.73	859.62 ± 196.21	745.75 ± 266.17	263.14 ± 40.90	173.4 ± 58.73	683.85 ± 78.01	519.08 ± 179.30	396.82 ± 61.81	120.82 ± 29.67	700.77 ± 271.42	532.42 ± 85.08	287.45 ± 75.44
Shore D	78.17 ± 2.02	88.50 ± 0.50	87.33 ± 1.76	82.17 ± 2.52	83.17 ± 1.04	95.67 ± 1.53	91.67 ± 1.76	91.17 ± 1.53	77.83 ± 1.15	96.17 ± 1.15	87.50 ± 3.04	85.00 ± 2.29

---

3.2.2. Mechanical properties: balancing strength and toughness for ocular implants. Mechanical tests focused on 12 resin variants (3 base resins: PCL2-, PCL4-, and PCL8-NIPUA; 4 formulations), with key findings centered on PCL chain length effects and the 6 : 3 : 1 formulation advantage:
3.2.2.1. PCL chain length dependence. Longer PCL chains enhanced toughness but traded off rigidity: PCL8 showed 150% higher elongation at break (Table 2, row 4) and 80% higher impact strength (Table 2, row 5) than PCL2, but 35% lower tensile strength (Table 2, row 3) and 25% lower Shore D hardness (Table 2, row 8). Stress–strain curves (Fig. 5C) further quantified this balance: PCL4-NIPUA achieved optimal performance (tensile strength ∼12 MPa, elongation at break 33%)—its 4-unit PCL chains balanced molecular sliding (for toughness) and load transfer (for strength). In contrast, PCL2-NIPUA was brittle (high modulus but low elongation) and PCL8-NIPUA lacked toughness (excessive molecular entanglement restricted deformation).
3.2.2.2. Formulation optimization. The 6 : 3 : 1 formulation (for PCL2- and PCL4-NIPUA) achieved 45 kJ m⁻² impact strength, attributed to synergistic interfacial compatibility, balanced crosslink density, and stress relaxation via chain mobility—addressing the strength–toughness trade-off.

3.2.3. Thermal stability and biocompatibility: meeting ophthalmic material standards.
3.2.3.1. Thermal properties. All resins exhibited initial decomposition at ∼300 °C (suitable for sterilization); PCL4-NIPUA showed the best stability (maximum decomposition temperature 428.38 °C vs. 416.78 °C for PCL2; Fig. 5D and E). DSC analysis revealed minimal residual monomer decomposition (∼2% at 100 °C) and main chain degradation at 420 °C.²⁸

The test results of the hydrophilicity and moisture absorption capacity of the material surface (see SI Fig S1 for details) show that all formulations have a water contact angle of less than 90° (with hydrophilic characteristics, suitable for moist eye environments) and a water absorption rate of less than 14% (and the water absorption rate in the PBS environment is slightly lower than that in deionized water), further verifying their basic adaptability as ophthalmic implant materials.

3.2.3.2. Biocompatibility. All resins met safety thresholds: in vitro cell viability >90% (72-hour extract tests; Fig. 6A and B), hemolysis rate <1% (Fig. 6C). The 6 : 3 : 1-formulated PCL4-NIPUA device (after ethanol cleaning and autoclaving) maintained >90% cell survival in direct contact tests (Fig. 6E and F).

Fig. 6 Biocompatibility assessment and device design of PCL-NIPUA materials. (A) Cell proliferation analysis following exposure to resin extracts (72 h culture). (B) Live/dead staining of cells treated with resin extracts (green: viable cells; red: dead cells). (C) Hemolysis test results showing erythrocyte integrity after extract exposure. (D) 3D printing schematic of the MB device with optimized curvature. Circular indentation segment at the posterior portion, three sets of regularly arranged perforations at the anterior portion. (E) CCK-8 viability assay of cells in direct contact with PCL4-NIPUA devices. (F) Fluorescence micrographs of cell viability adjacent to implanted devices. (G) Cross-sectional illustration demonstrating the pressure application mechanism through controlled curvature mismatch between the device and the scleral surface. Scale bars: 100 μm (B and F). Error bars represent mean ± SD (n = 3).

3.2.4. Screening outcome: PCL4-NIPUA (6 : 3 : 1 formulation) as the optimal candidate. PCL4-NIPUA (6 : 3 : 1 formulation) was selected for device fabrication, as it integrated: (1) ideal 3D printing adaptability (low viscosity, high curing efficiency), (2) mechanical balance (tensile strength 25 MPa, elongation 32%) matching ocular tissue requirements, (3) thermal stability (decomposition temperature 428.38 °C), and (4) excellent biocompatibility. Based on existing research on MTM therapeutic surgical devices.^29–32 The device design includes crucial enhancements over current macular buckling techniques to address known complications of choroidal ischemia and macular atrophy.³³ The innovative porous architecture (Fig. 8D) featured a curvature with a radius of 10.5 mm, and was specifically designed to create controlled indentation on the rabbit eyeball, which has a radius of 11.5 mm. This design included six 1-mm suture pores for surgical fixation and thirty-one 0.3-mm vascularization pores distributed across the pressure surface, implementing the vascular integration principles demonstrated in previous studies.³⁴ The therapeutic mechanism (Fig. 8G) utilized the curvature mismatch to generate controlled scleral pressure while the microporous architecture preserved vascular perfusion, addressing the ischemia risks identified in earlier device designs.³⁵ This comprehensive evaluation confirms the successful translation of material properties into functional medical devices that meet both biocompatibility requirements and therapeutic objectives for posterior scleral disorders.

Fig. 7 In vivo evaluation of PCL4-NIPUA MB devices. (A) Postoperative external eye photographs showing surgical outcomes at defined time points (immediately after surgery, 2 weeks, 2 months, and 6 months), demonstrating gradually resolution of conjunctival congestion and stable device positioning. (B) Enucleated eyeball specimens at 6-month endpoint, with (up) and without (down) the implanted device, revealing persistent indentation marks that confirm sustained therapeutic pressure application. Scale bars: 5 mm (A), 5 mm (B). (C) Comparison of scanning electron microscopy images of the device surface before and after implantation.

Fig. 8 Longitudinal OCT imaging and histological analysis of implanted eyes. (A and B) Representative ultra-widefield (UWF) fundus images from two rabbits at postoperative day 0 (1), 2 weeks (2), 2 months (3), and 6 months (4), demonstrating maintained retinal integrity. Corresponding swept-source OCT images (a1–b4) show persistent therapeutic indentation (white arrows). Histological sections of retina (Aa1 and Bb1) and optic nerve (Aa2 and Bb2) after 6-month implantation reveal preserved tissue architecture without inflammation. Scale bars: 1 mm (OCT), 100 μm (H&E).

3.3. In vivo performance evaluation of PCL4-NIPUA devices

3.3.1. Surgical outcomes and ocular compatibility. The PCL4-NIPUA-printed MB devices were successfully implanted in 16 rabbit eyes without intraoperative complications. During the 6-month follow-up period, all eyes maintained normal intraocular pressure and showed no signs of anterior chamber inflammation. Mild conjunctival congestion was observed in the early postoperative period and resolved completely by week 2 (Fig. 7A). Three eyes exhibited incomplete conjunctival healing with partial device exposure at the 2-week follow-up. Of these, one eye developed ocular surface infection with purulent discharge and device extrusion by month 2, which was successfully managed through device removal, saline irrigation, and antibiotic ointment application. Another eye demonstrated disappearance of the indentation ridge on OCT by month 2, with subsequent enucleation confirming device loss, while the third eye maintained stable partial exposure without progression. We think that these complications were mainly attributed to non-absorbable suture, insufficient postoperative antibiotic coverage and suboptimal conjunctival closure. The remaining 13 eyes maintained stable device positioning throughout the study period, developing mature fibrous encapsulation by month 6 without evidence of material degradation. Post-examination revealed clear indentation marks on the sclera and normal optic nerve appearance (Fig. 7B). Comparative scanning electron microscopy analysis of pre-implantation (Fig. 7C5–8) and post-implantation (Fig. 7C1–4) devices demonstrated significant surface modifications, including extracellular matrix protein deposition and organized collagenous capsule formation, confirming successful host-device integration and material stability in the ocular environment.

The devices function by applying controlled inward pressure on the posterior sclera, counteracting both vitreous traction and staphyloma-induced outward forces. This dual-action mechanism promotes macular reattachment while addressing the underlying biomechanical pathology. Compared to vitrectomy, this approach offers distinct advantages for myopic traction maculopathy management, including lower recurrence rates and preservation of ocular anatomy. The 6-month results demonstrate that PCL4-NIPUA devices provide effective, sustained scleral reinforcement without provoking adverse inflammatory responses or material failure.

3.3.2. Longitudinal ocular imaging and histological assessment. This study evaluated the effectiveness by implanting a device into normal rabbit eyeballs, with the appearance of pressure ridges in OCT as the criterion for assessing effectiveness. Postoperative fundus evaluation was systematically performed using ultra-widefield (UWF) color imaging and spectral-domain optical coherence tomography (OCT) on all operated eyes, with two representative rabbit eyes shown in Fig. 8. The UWF images (A1–B4) captured at four time points (postoperative day 0, 2 weeks, 2 months, and 6 months) demonstrated preserved retinal anatomy without pathological changes. Corresponding OCT scans (a1–b4) revealed consistent therapeutic indentation throughout the observation period, with clearly defined pressure ridges,^36,37 confirming the sustained biomechanical effect of the PCL4-NIPUA devices. These imaging results demonstrate successful inhibition of axial elongation while maintaining normal retinal architecture.³⁸

Histopathological analysis at the 6-month endpoint included H&E staining of retinal and optic nerve sections from the implanted eyes (representative H&E-stained sections from two eyes shown in Fig. 8Aa1-2 and Bb1-2). The examined sections showed intact tissue morphology in both neural and vascular components, with no evidence of structural disruption or inflammatory cell infiltration. These findings correlate with the imaging results and confirm the long-term biosafety of the implanted devices. The combined imaging and histological data provide comprehensive evidence that the scleral reinforcement procedure achieves its therapeutic objective without inducing secondary tissue damage.

This study demonstrates that 3D-printed PCL4-NIPUA devices effectively maintain posterior scleral indentation for at least six months while exhibiting excellent biocompatibility. The devices successfully address the biomechanical pathology of myopic traction maculopathy without causing sight-threatening.

4. Conclusions

This study successfully developed a PCL-NIPUA resin system through an environmentally friendly synthetic approach, avoiding toxic isocyanate precursors commonly used in conventional polyurethane synthesis. Comprehensive characterization by ¹H NMR, FTIR, and mass spectrometry confirmed the precise molecular architecture, while systematic evaluation of photocuring behavior demonstrated optimal processing characteristics for high-resolution 3D printing of ocular devices. The incorporation of flexible PCL segments significantly enhanced material toughness while maintaining appropriate mechanical compatibility with human ocular tissues, meeting the requirements of MB materials.

The optimized PCL4-NIPUA formulation demonstrated exceptional biocompatibility, with hemolysis rates below 1% and cell viability exceeding 90%, meeting ISO 10993 standards for medical implants. We developed a one-piece J-shaped MB device featuring anterior suture ports and a posterior hemispherical indentation zone. Using 3D-MRI imaging and photocurable 3D-printing technology, this system allows for patient-specific customization based on each individual's unique anatomy. In vivo evaluation in rabbit models confirmed the device's ability to maintain sustained indentation for six months without inducing inflammation or tissue damage. The material's hydrophilic surface (contact angle <90°) and controlled swelling behavior (<14% in PBS) further support its clinical applicability, building upon established principles of ocular implant design.

While the current formulation shows significant improvement over existing options, its mechanical resilience remains inferior to clinical-grade silicone rubber, potentially affecting long-term dynamic adaptation to the ocular curvature. Future work should explore molecular hybridization with elastomeric components to enhance rebound elasticity, informed by recent advances in biomaterial science.³⁵ Additional animal experiment extending beyond six months would provide valuable data on chronic biocompatibility and therapeutic durability, addressing the long-term performance requirements identified in clinical studies.

This research establishes an important advancement in ophthalmic MB device manufacturing by combining green chemistry principles with digital light processing technology. The PCL-NIPUA system enables patient-specific implant fabrication while overcoming critical limitations of conventional materials, representing a transformative approach to MTM treatment. With continued development focusing on viscoelastic property optimization, this technology has significant potential for clinical translation and commercialization, providing improved safety and therapeutic efficacy for patients with MTM.

Author contributions

Zi Fu: conceptualization, methodology, data curation, and writing – original draft; Yan Zhou: investigation, data curation, and formal analysis; Ziyi Liu: software, visualization, and validation; Zhe Chen: investigation and resources; Qiang Fu: resources; Zhe Shiqun Lin: project administration; Rongping Dai: supervision, and writing – review and editing; and Huade Zheng: conceptualization, supervision, writing – review and editing, and funding acquisition. All authors agree to be accountable for all aspects of the work.

Conflicts of interest

There are no conflicts to declare.

Data availability

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

Supplementary information (SI): Experimental results of water contact angle and water absorption rate measurements. RNA sequences used in the Q PC experiment. See DOI: https://doi.org/10.1039/d5tb02033f.

Acknowledgements

This work was financially supported by the Guangdong Basic and Applied Basic Research Foundation (2020B1515120075), Dongguan Sci-tech Commissioner Program (20221800500032) and the Key Technologies R&D Program of Dongguan (20221200300072).

References

1. H. Chen, X. Liu and X. Zhou, et al. , Ophthalmic Res., 2024, 67, 424–434.
2. S. Matsumura, A. N. Kuo and S.-M. Saw, CLAO J., 2019, 45, 279–285.
3. B. A. Holden, T. R. Fricke and D. A. Wilson, et al. , Ophthalmology, 2016, 123, 1036–1042.
4. I. G. Morgan, A. N. French and R. S. Ashby, et al. , Prog. Retinal Eye Res., 2017, 62, 134–149.
5. H. Terasaki, T. Yamashita and R. Funatsu, et al. , Sci. Rep., 2023, 13, 15367.
6. I. Flores-Moreno, M. Puertas and E. Almazán-Alonso, et al. , Graefe's Arch. Clin. Exp. Ophthalmol., 2021, 260, 133–140.
7. Y. Fang, T. Yokoi and N. Nagaoka, et al. , Ophthalmology, 2018, 125, 863–877.
8. E. Premi, S. Donati and L. Azzi, et al. , J. Ophthalmol., 2022, 10, 2270861.
9. Y. Fang, R. Du and N. Nagaoka, et al. , Ophthalmology, 2019, 126, 1198–1209.
10. C.-b Sun, Y.-s You and Z. Liu, et al. , Sci. Rep., 2016, 6, 27952.
11. L. N. T. Phan, T. T. Ngo and V. L. Nguyen, et al. , Int. Ophthalmol., 2025, 45, 35.
12. J. Q. Li, R. Brinken and F. G. Holz, et al. , Eye, 2020, 35, 2206–2212.
13. B. Parolini, J. F. Rosales Padrón and E. Lopes, et al. , Retina, 2024, 44, 1180–1187.
14. C. J. Chaya, L. W. Herndon and J. Lince, et al. , J. Clin. Med., 2024, 13, 4593.
15. W. Zhenquan, Z. Xiujuan, C. Shida and L. Ping, Chin. J. Ophthalmol., 2021, 57, 433–439.
16. V. Karamzadeh, M. L. Shen and H. Ravanbakhsh, et al. , Adv. Healthcare Mater., 2023, 13, 2303708.
17. Y. Zhou and J. Qu, J. Coat. Technol. Res., 2023, 21, 601–610.
18. C. Wang, K. Yan and X. Luo, et al. , Color. Technol., 2021, 137, 348–360.
19. A. Abdal-hay, M. Agour and Y.-K. Kim, et al. , Eur. Polym. J., 2019, 112, 555–568.
20. A. A. Gostev, A. A. Karpenko and P. P. Laktionov, Polym. Bull., 2018, 75, 4311–4325.
21. T. V. Chirila, Clin. Exp. Ophthalmol., 2006, 34, 485–488.
22. Y. Wang, Z. Zheng and J. L. Pathak, et al. , Int. J. Bioprinting, 2023, 9, 684.
23. X. Jiang, Z. Yang and Z. Wang, et al. , Materials, 2018, 11, 11040474.
24. A. Kausar, J. Plast. Film Sheeting, 2022, 38, 32022.
25. L. Elomaa, S. Teixeira and R. Hakala, et al. , Acta Biomater., 2011, 7, 3850–3856.
26. H. J. Assumption and L. J. Mathias, Polymer, 2003, 44, 511–5136.
27. Z. Li, Y. Zhu and W. Niu, et al. , Adv. Mater., 2021, 33, 2101498.
28. J. S. Im, B. C. Bai and T.-S. Bae, et al. , Mater. Chem. Phys., 2011, 126, 685–692.
29. L. Akduman, S. Ozdek and S. Ermis, et al. , J. VitreoRetin. Dis., 2024, 8, 62024.
30. E. N. El Rayes and E. Elborgy, J. Ophthalmic Vision Res., 2013, 8, 393–399.
31. R. H. Muni, I. M. Melo and A. Pecaku, et al. , JAMA Ophthalmol., 2023, 141, 10.
32. B. Parolini, Vision, 2024, 8, 8030042.
33. P. Susvar and G. Sood, Indian J. Ophthalmol., 2018, 66, 1772–1784.
34. Y. Yuan, Y. Zong and Q. Zheng, et al. , Biomater. Adv., 2016, 62, 233–241.
35. T. Xu, Y. Yang and Y. Yu, et al. , Retina, 2013, 33, 1062–1069.
36. W.-C. Wu, C.-C. Lai and H. S.-L. Chen, et al. , Invest Ophthalmol. Visual Sci., 2008, 49, 62008.
37. Y. Zhou, S. Lin and X. Xiao, et al. , Retina, 2024, 45, 608–613.
38. Y. Xu, Q. Chen and Z. Shao, et al. , Front. Bioeng. Biotechnol., 2023, 11, 1211688.

---

Footnote

† These authors contributed equally to this study and should be considered co-first authors.

---