3D-printed versatile biliary stents with nanoengineered surface for anti-hyperplasia and antibiofilm formation

Abstract

Biliary strictures are characterized by the narrowing of the bile duct lumen, usually caused by surgical biliary injury, cancer, inflammation, and scarring from gallstones. Endoscopic stent placement is a well-established method for the management of biliary strictures. However, maintaining optimal mechanical properties of stents and designing surfaces that can prevent stent-induced tissue hyperplasia and biofilm formation are challenges in the fabrication of biodegradable biliary stents (BBSs) for customized treatment. This study proposes a novel approach to fabricating functionalized polymer BBSs with nanoengineered surfaces using 3D printing. The 3D printed stents, fabricated from bioactive silica poly(ε-carprolactone) (PCL) via a sol–gel method, exhibited tunable mechanical properties suitable for supporting the bile duct while ensuring biocompatibility. Furthermore, a nanoengineered surface layer was successfully created on a sirolimus (SRL)-coated functionalized PCL (fPCL) stent using Zn ion sputtering-based plasma immersion ion implantation (S-PIII) treatment to enhance the performance of the stent. The nanoengineered surface of the SRL-coated fPCL stent effectively reduced bacterial responses and remarkably inhibited fibroblast proliferation and initial burst release of SRL in vitro systems. The physicochemical properties and biological behaviors, including in vitro biocompatibility and in vivo therapeutic efficacy in the rabbit bile duct, of the Zn-SRL@fPCL stent demonstrated its potential as a versatile platform for clinical applications in bile duct tissue engineering.

Graphical abstract

Highlights

• •Versatile biliary stents were successfully prepared via 3D printing and nanoengineering.
• •Functionalized PCL stent by silica exhibited controlled mechanical properties and biodegradability.
• •Nanoengineered surface reduced bacteria responses and remarkably inhibited fibroblast proliferation.
• •This strategy served to demonstrate the great potential for bile duct tissue engineering.

1. Introduction

The bile duct transports bile juice from the liver and gallbladder to the second portion of the duodenum, thereby playing a crucial role in the digestion process [1]. Benign and malignant biliary obstructions or strictures, which can lead to the narrowing of the bile duct, can result from various conditions, such as biliary injuries during surgery, cancer, inflammation, and scarring from cholelithiasis [2,3]. Benign or malignant biliary obstruction can significantly induce jaundice, liver failure, pruritus, cholangitis, and sepsis [4]. The diagnosis of malignant biliary obstruction with the initial goal of identifying candidates for curative surgery is one of the most important factors in treating biliary obstruction. The more than 70% of patients with malignant biliary obstruction are diagnosed at advanced-stage of the tumors [5]. Surgical resection is limited to fewer than 30% patients with malignant biliary obstruction, resulting in a high mortality rate [6,7]. Nonbiodegradable endobiliary plastic and self-expandable metallic stents are clinically used for biliary drainage in patients with malignant or benign obstruction to relieve their symptoms, such as jaundice, diarrhea, anorexia, and sleep disorders, and improve their quality of life [[8], [9], [10]]. Although stent placement is a safe and effective modality requiring a shorter hospital stay, reduced costs, and faster relief of symptoms compared with those with surgical approaches, stent-related re-obstruction frequently occurs because of biofilm formation, stent-induced granulation tissue formation, and growth of tumors within or around stent mesh [[9], [10], [11], [12]]. These complications require reinterventions such as stent removal, stent replacement, and placement of additional stent, exposing the patient to additional damage to the bile duct during repeated interventions [[13], [14], [15]]. Therefore, the ideal biliary stent should have improved biocompatibility, excellent expandability, and complete biodegradability after the stenosis is resolved [3,16] A potential solution to address this situation is the use of biodegradable polymeric stents that degrade naturally, thereby eliminating the need for further interventions and an external biliary drainage catheter for the treatment of biliary strictures [17]. Moreover, because the bile duct conditions vary based on the lesion, gender, age, and condition of the patient, customized biliary stents can treat biliary strictures more effectively. Recently, stents manufactured via extrusion-based three-dimensional (3D) printing with the introduction of an extruder and mandrel along the xƟz axis have been investigated to customize their biliary lesion and size [[17], [18], [19], [20]].

In addition to patient-specific customization, tuning the mechanical properties of 3D-printed polymeric stents also remains a challenge. Several Food and Drug Administration-approved biodegradable polymers, such as poly(ε-caprolactone) (PCL), poly(l-lactic acid) (PLLA), polydioxanone, and poly(lactic-co-glycolic acid), have been explored as polymer-based biodegradable stent materials or coating materials for drug eluting stents owing to their biodegradability, nontoxicity, and biocompatibility with various drugs [[21], [22], [23], [24], [25], [26]]. Ensuring the patency of bile ducts is critical for the effective management of biliary strictures, particularly after liver transplantation, because the majority of complications arise within the first three months. Therefore, the use of simple polymers alone may not be suitable for the management of biliary strictures [27]. Fabricating polymer-based and functionalized nanocomposites is one of the most facile and effective strategies for incorporating and combining the functions and properties of inorganic nanoparticles (NPs), otherwise not attainable by polymers alone [28]. Various inorganic NPs have been introduced to facilitate diverse functionalized polymer designs in which NP reinforcements are stabilized and immobilized via physical or chemical interactions within the polymer matrix [[29], [30], [31]]. Recently, silica (SiO₂) NPs, which are known as representative metal oxide-based bioactive nanotherapeutics, have been shown to exhibit drug loading ability, biocompatibility, and mechanical reinforcement effect [[32], [33], [34]]. Silica incorporated with polymers serves as a mechanical reinforcement for various polymer systems such as PCL, PLLA, and polyurethane, enabling enhanced mechanical performance suitable for clinical use [35,36].

Generally, coated antiproliferative or immunosuppressive therapeutics on the biliary stent locally treat the target biliary lesion, suppress the proliferation of fibroblasts, and inhibit neointima formation. Particularly, in-stent restenosis has been known to be substantially reduced with the use of drug eluting stents combined with the most widely used drugs, such as paclitaxel or sirolimus [[37], [38], [39]]. However, the initial burst of drugs is inevitable when using polymer-based drug delivery. For effective therapeutic delivery at the impaired bile duct, prolonged therapy is preferable to the excessive burst-type delivery of drugs and molecules to remain within the therapeutic dosage range throughout the bile duct treatment. Previously reported cases show that metal ion-embedded (sputtering-based plasma immersion ion implantation (S-PIII)) or -implanted surfaces (target-ion induced plasma sputtering) improved in vitro and in vivo cellular performances for biomedical fields including stent applications [[40], [41], [42], [43], [44], [45]]. Particularly, the S-PIII technique has been introduced as a facile yet effective polymeric surface modification strategy via massive metal ion, such as tantalum ion, implantation using accelerated metallic ions by high-density plasma [45]. S-PIII-treated layers exhibit excellent adhesion stability and biocompatibility and act as a physical barrier to suppress the initial burst of drugs. Among numerous candidates of metallic ions, Zn ions have recently been highlighted because of their distinctive antibacterial properties and biocompatibility [[46], [47], [48], [49]]. Thus, we hypothesize that 1) S-PIII-treated Zn ions on the 3D-printed biliary stent, known to have antibacterial properties, can reduce bacterial adhesion and biofilm formation on the surface of the biliary stent while maintaining its biocompatibility and 2) also suppress the initial antiproliferative drug burst, consequently prolong the bile duct patency without any adverse effect [[50], [51], [52], [53]].

To test these hypotheses, we designed 3D-printed polymer stents with silica coupled with implanted Zn ions for the management of biliary strictures. We used a 3D-printed biodegradable biliary stent (BBS) strategy based on sol–gel-derived bioactive silica-PCL nanocomposite and coated sirolimus embedded with Zn ions via the S-PIII treatment, as shown in Scheme 1. The mechanical performance and biocompatibility of 3D-printed versatile biliary stents are expected to be improved by combining the advantages of inorganic and organic biomaterials. Furthermore, a nanoengineered surface layer, coated with sirolimus/PLLA and embedded with the Zn ions, on the functionalized PCL stent can synergistically inhibit hyperplasia and biofilm formation, offering great potential for its use in a wide range of bile duct therapies. The physicochemical properties and biological behaviors, such as in vitro biocompatibility and in vivo bile duct therapy ability of the 3D-printed BBS, are estimated and discussed in detail.

Scheme 1.

(A) Schematic of the manufacturing process of customized biodegradable biliary stent (BBS) combining 3D printing and nanoengineering. (B) The elliptical strut facilitates the flow of bile to suppress the formation of precipitates and inhibits bacterial adhesion and biofilm formation by the hydrophilic surface and antibacterial Zn ions. The surface coated with sirolimus/PLLA embedded with the Zn ions controls drug release and effectively prevents fibroblast cells proliferation.

2. Materials and methods

2.1. Materials

The following materials were used in this study. Poly(ε-carprolactone) (PCL; Mn 45,000), dichloromethane (DCM), tetramethyl orthosilicate, Tween-20, antibiotic antimycotic solution (AA, 100×), glutaraldehyde, 1,1,1,3,3,3-hexamethyldisilazane, Triton X-100, and bovine serum albumin purchased from Sigma Aldrich, USA; fetal bovine serum (FBS) and 0.05% Trypsin-EDTA (1×) purchased from Gibco, USA; phosphate buffered saline (PBS) and alpha-minimum essential medium (α-MEM) purchased from Welgene, Korea; bladder epithelial cell basal medium and bladder epithelial growth kit were purchased from ATCC, USA; hydrochloric acid (HCl), 1,4-dioxane, and 4% paraformaldehyde purchased from Daejung, Korea; poly-l-lactic acid granules (PLLA, Mn = 100,000) purchased from Pureco, Korea; sirolimus purchased from LC Laboratory, USA; Zn target (purity 99.99%) purchased from Kojundo Chemical Lab, Japan; holey carbon grid (200 mesh Cu) purchased from Structure Probe Inc., USA; Luria–Bertani (LB) broth and LB agar purchased from BD DifcoTM, USA; Alexa Fluor™ 555 Phalloidin and 4′, 5-diamidino-2-phenylindole purchased from Invitrogen, Thermo Fisher Scientific, USA; Cell Counting Kit-8 (CCK-8) purchased from Dong-in LS, Korea; ox bile powder purchased from Nutricology, USA; potassium chloride purchased from JW Pharm, Korea; 18G angiocatheter with needle (BD Angiocath Plus) purchased from Becton Dickinson, USA; 0.014-inch micro-guide wire (Transend) purchased from Boston Scientific/Medi-Tech, USA; contrast medium (Telebrix Gastro) purchased from Guerbet, Villepinte, France; 3-mm-diameter and 10-mm-long balloon catheter purchased from Genoss, Korea; ketorolac trometamol and gentamicin purchased from Hana Pharm, Korea; terminal deoxynucleotidyl transferase mediated dUTP nick end labeling (TUNEL) purchased from Biogene, Germany; α-smooth muscle actin (α-SMA) antibody and Ki67 antibody. Other than these, we also used alanine aminotransferase (ALT), alkaline phosphatase (ALP), gamma (γ)-glutamyl transferase (GGT), and total bilirubin (TB)/hematoxylin and eosin (H&E) and Masson's trichrome (MT).

2.2. Methods

2.2.1. Synthesis of functionalized PCL with silica

To prepare a 10 w/v% polymer solution, PCL was added to DCM. Silica precursor was obtained by mixing tetramethyl orthosilicate, deionized water (DW), HCl, and DCM at a volume ratio of 2.5:0.5:0.01:3 [35]. After stirring for 30 min to achieve a single-phase solution, a silica solution at concentrations ranging from 0 to 30 wt% was added to the PCL solution. After drying the solution on a plate for 24 h, dried samples were chopped to fabricate chips for 3D printing.

2.2.2. Fabrication of 3D-printed functionalized PCL stents

The functionalized PCL (fPCL) stents were fabricated using a 3D stent printing system (CGBIO, Korea), as shown in Scheme 1. The unit cell of the stent strut was repeated seven times in the longitudinal direction and six to eight times in the horizontal direction. The unit cells were connected by three links. The rotation direction coordinates x, y, and z of the designed structure were transformed into g-code, which was then sent to the 3D printer using the instrument software. The printing system consisted of a translational stage with three axes (x-Ɵ-z), dispenser, nozzle, compression/heat controller, and software system. The fPCL chips were melted in a heating dispenser at 100 °C for 20 min, and the temperature of the nozzle heater was maintained at 110 °C. After the PCL was melted, a constant air pressure of 300–350 kPa was applied to the dispenser to extrude the material through a 200-μm nozzle. Eventually, stents with a diameter of 3 mm and a length of 9 mm were printed. Rectangular material specimens with dimensions of 10 × 10 × 2 mm³ were also printed via a 3D-bioprinter (4D6, Rokit, Korea) for physicochemical tests. The printing conditions were 110–130 °C with 5 mm/s of head speed.

2.2.3. Nanoengineering of the 3D-printed fPCL stents

To prepare the ultrasonic spraying solution, PLLA was heated in a furnace at 200 °C and quenched. The heat-treated PLLA was dissolved in 1,4-dioxane at a concentration of 1.5 w/v% and a sirolimus was supplemented to prepared solution with a concentration of 33 wt%. Immediately after printing the fPCL stents, sirolimus-loaded PLLA coating was applied to their surfaces using an ultrasonic spray nozzle fitted in the 3D stent printing system. The spray and rotation speed for spray coating were 100 μg/ml and 30 rpm, respectively. The coated stents were dried at 4 °C and removed from the Ɵ-axis mandrel of the 3D stent printer. Subsequently, the stents were cleaned with ethanol and treated with ultraviolet (UV) for 10 min, followed by their Zn S-PIII treatment.

The direct current magnetron sputter gun housing (Ultech Co., Ltd., Korea) included a Zn target with a diameter of 75 mm and a thickness of 5 mm. To set up the samples for the S-PIII treatment, the Zn target was aligned parallel to the long axis of the printed stents, which were fastened to a stainless-steel plate. Subsequently, a vacuum chamber containing specimens was pumped to a pressure of 5 × 10⁴ Pa using rotary and diffusion pumps. A negative voltage of 2000 V for 10 s under a 7 mTorr of Ar gas pressure was used. A target current of 50 mA was applied without any additional heat. The same procedure followed to fabricate the 3D-printed rectangular specimens to form nanoengineered surfaces.

2.2.4. Characterization of 3D-printed fPCL

Dynamic light scattering (DLS; Zetasizer Nano ZS, Malvern Instruments Ltd., Malvern, UK) analysis, with a 633-nm laser set at a scattering angle of 173° and performed at 25 °C was used to obtain the size distribution of silica in PCL. Five measurements were obtained and averaged to get the final value. The size and chemical composition of the silica in PCL were analyzed using transmission electron microscopy (TEM; JEM-2100F, JEOL, Japan) coupled with energy dispersive spectroscopy (EDS) at an acceleration voltage of 200 kV. TEM samples were prepared by homogenizing the hydrogels to granules and diluted with DW before loading them on a holey carbon grid. The TEM samples were dried in vacuum for 24 h before further analysis.

The chemical bonds in the synthesized fPCL were identified using Fourier-transform infrared (FT-IR) spectroscopy (Cary 630 FT-IR, Agilent, USA). Between 400 and 4000 cm⁻¹, FT-IR spectra of pristine PCL, sol–gel derived silica, and fPCL_Si20 were obtained. We used differential scanning calorimetry (DSC; PerkinElmer DSC 4000, Massachusetts, USA) and thermogravimetric analysis (TGA; TGA2, Mettler-Toledo, Germany) for the thermal analysis of the synthesized fPCL based on the silica contents. During DSC analysis, 2–3 mg of materials were put in an aluminum pan and heated from 30 to 100 °C at a rate of 10 °C/min in Ar flow. TGA was performed at a heating rate of 10 °C/min from 30 to 900 °C. A flow rate of 50 mL/min in an environment of N₂ was used to determine the thermal stability of 5 mg of the prepared sample.

Field emission scanning electron microscopy (FE-SEM; S-4800, HITACHI, Japan) was used to study the microstructure and cross section of the 3D-printed fPCL stents. Furthermore, EDS was utilized to study the elemental distribution. The surface and cross-sectional morphologies of the pristine PCL and fPCL stents were analyzed. The mechanical performances of the 3D-printed fPCLs with various silica contents were investigated via tension and compression tests. The mechanical properties of the stents, including tensile and compressive properties, were measured using a universal testing machine (MTDI INC., Korea) operated at a crosshead speed of 1 mm/min, in accordance with ASTM D695 and D638.

2.2.5. Characterization of 3D-printed stents with nanoengineered surface

The surface composition of specimens was evaluated by sectioning their cross section using a focused ion beam (FIB; Helios 650, FEI, USA) installed at National Center for Inter-university Research Facilities. Before milling a specimen, it was coated with Pt using the Pt coater (108 auto sputter, Crissington) at a current of 20 mA for 60 s. FIB milling was performed at an acceleration voltage of 30 kV for milling and at 2 kV with 39 pA for the final cleaning step. Cross-section samples were observed with a TEM coupled with EDS at an acceleration voltage of 200 kV. X-ray photoelectron spectrometry (XPS; Axis SupraTM, Kratos, England) was used to examine the chemical composition and presence of nanoengineered layers. The hydrophilicity of samples was assessed by measuring the contact angle formed on their surfaces using a drop of distilled water with the Phoenix 300 (Surface Electro Optics Co., Ltd., Korea) contact angle analyzer.

The mechanical tests were performed to measure the radial force and recovery rate of 3D-printed stents using a universal testing machine. The radial force was analyzed when the diameter was reduced to 50% at a speed of 10 mm/min, and the test was performed using a load–unload compressive test. In the recovery rate experiment, the diameters of stents were pressed to half of their initial values and the final diameters were checked after 6 h and the following equation:

$$\text{Recovery}\text{rate}(\%)=\frac{D_b}{D_a}\times 100$$

where D_a is the initial diameter of the specimen and D_b is its diameter 6 h after remaining in 50% compression.

2.2.6. In vitro degradation and molecules release analysis

In vitro degradation analysis was performed to investigate the change in weight, surface morphology, and mechanical characteristics of the 3D-printed stents. 3D-printed stents with a diameter of 3 mm and a length of 9 mm were immersed in PBS for 2, 4, 8, and 12 wk and then rinsed with DW. Next, the specimens were freeze dried for 24 h. The following equation was used to calculate the percentage of weight loss:

$$\text{Comparative}\text{weight}(\%)=\frac{W_t}{W_O}\times 100$$

where W_O is the initial weight of the stent, and W_t is the weight of the dried stent after time t. Thereafter, FE-SEM was used to analyze the surface morphologies of the dried specimens.

The 3D-printed stents were submerged in 3 ml of PBS including 0.05% of Tween-20 at 37 °C for 12 wk to analyze their drug release behavior. After being collected at regular intervals, the sirolimus released medium was filtered through 0.2 μm filters. After combining the PBS containing the filtered sirolimus with anhydrous grade methanol, the optical density of the filtered medium was determined using nanodrop spectrophotometry (Nanodrop, Thermo Fisher, USA) at a wavelength of 278 nm.

Initially, all the solutions were collected after 4 h and 12 h to confirm the initial release of Zn and Si ions. Subsequently, the interval was increased to 1, 3, 5 d and 1, 2, 3, 4, 6, 8, 10, and 12 wk to complete the collection using an inductively coupled plasma mass spectrometer (ICP-MS; ICAP Q, Thermo Fisher, USA). Thereafter, the solution was replaced with 3 mL of fresh PBS.

2.2.7. In vitro evaluation of anti-adhesion and bactericidal properties

Prior to culturing Escherichia coli (E.coli; ATCC 8739, Rockville, MD, USA) and gram-positive Staphylococcus aureus (S. aureus; ATCC 6538, Rockville, MD, USA) on the specimens, the bacterial solution was prepared by culturing for 18 h after inoculating 50 μl of the stock solution in 3 ml of the LB broth. The prepared E. coli and S. aureus solutions was diluted to 1 × 10⁸ CFU/ml and 1 × 10⁴ CFU/ml, respectively.

• 1)Anti-adhesion test: Sterilized specimens (10 × 10 × 2 mm³) were cultured in a diluted bacteria solution for 1 d at 200 rpm at 37 °C in a shaking incubator (SI-300R, Lab Companion, USA). For the in vitro bile flow model, 50 ml of the bile solution containing 10 wt% ox bile powder and 1 × 10⁵ CFU/ml of E. coli was prepared. The fabricated stents were placed in autoclaved tubes, which were connected between the circulation pump (SciQ 323, Watson Marlow, UK) and bile solution at 100 rpm stirring to prevent sedimentation. The pumping speed was set to 0.5 ml/min and the temperature was maintained at 37 °C. After 5 d, stents were removed from the tubes. To analyze the adhered bacteria and biofilm on the stent surface, the bacteria-adhered specimen was immersed in a 2.5% glutaraldehyde solution. Furthermore, the specimens were dehydrated using EtOH at various concentrations (50%, 60%, 70%, 80%, 90%, and 100%) with 5 min immersion for each condition and 1,1,1,3,3,3-hexamethyldisilazane.
• 2)Bactericidal test: Sterilized specimens (10 × 10 × 2 mm³) were fixed in culture plates. Subsequently, 60 μl each of the prepared E. coli and S. aureus solutions with the same concentration as in the anti-adhesion test was poured onto the specimens. After incubation in a shaking incubator at 200 rpm and 37 °C for 3 h, live bacteria were detached from the specimens in 3 ml of the LB broth by vortex for 1 min. Next, 100 μl of a 1/1000 diluted solution was spread on an agar plate (LB agar) and incubated in a humidified oven at 37 °C for 1 d. The average colony numbers of each bacteria were analyzed based on optical images.

2.2.8. In vitro evaluation of cell cytocompatibility

To evaluate cell attachment and viability, fibroblast cells, namely L929; CCL-1, Mus musculus, were cultivated in α-MEM, containing 10% fetal bovine serum and 1% AA and bladder epithelial cells were cultivated in bladder epithelial cell basal medium supplemented with bladder epithelial growth kit in an incubator with 5% CO₂ at 37 °C [54]. Fibroblasts and bladder epithelial cells with a density of 3 × 10⁴ cells/ml was used for cell attachment and 2 × 10⁴ cells/ml for cell viability testing on PCL, fPCL, SRL@fPCL, and Zn-SRL@fPCL specimens. Before fluorescence observation, the cultured cells on day 1 were fixed in 4% paraformaldehyde for 10 min and rinsed twice with PBS. The remaining cells were then permeabilized for 5 min with 0.1% Triton X-100 and 1% bovine serum albumin to block the nonspecific site before staining the F-actin and nuclei of the cells with Alexa Fluor™ 555 Phalloidin and 4′, 5-diamidino-2-phenylindole, respectively. Cell attachment was observed using a microscope (Nikon, Japan), and the cell viability was measured by CCK-8. After 3 and 5 d of incubation, the samples were collected, rinsed with PBS, and transferred to a new cell culture plate replenished with fresh media containing 10% of the CCK-8 reagent. After 2 h, the solutions were transferred to 96 well plates and absorbance was measured using a wavelength of 450 nm through a microplate reader (Model 550, Biorad, USA). The fibroblast cells and bladder epithelial cells cultured on PCL were used as a control, and the cell viability (%) was calculated using the following equation:

$$\text{Cell}\text{viability}(\%)=\frac{{OD}_{@490\text{nm}}\left(\text{Experimental}\text{group}\right)}{{OD}_{@490\text{nm}}\left(\text{Control}\right)}$$

2.2.9. Animal study design

The animal study was approved by the Institutional Animal Care and Use Committee of the institute and conformed to the US National Institutes of Health guidelines for the humane handling of laboratory animals (IACUC No. 2022-14-303). Twenty New Zealand white rabbits, weighing 3.2–4.1 kg (male; mean weight, 3.59 kg; JA BIO, Korea), were randomly divided into four groups of five each. The control and fPCL groups received pristine PCL and fPCL stents, respectively. The SRL@fPCL and Zn-SRL@fPCL groups received SRL-coated fPCL and Zn-SRL-loaded fPCL stents, respectively. All rabbits were housed at 24 ± 2 °C with a 12-h day/night cycle and free access to water and food. The body weight was monitored weekly until the sacrifice of the rabbits, and the percentage weight losses were evaluated using the below given formula. All rabbits were sacrificed 4 wk after stent placement, by an overdose of potassium chloride.

$$\text{Weight}\text{loss}(\%)=100\times \left(\frac{\text{weekly}\text{measured}\text{weight}-\text{baseline}\text{weight}}{\text{baseline}\text{weight}}\right)$$

2.2.10. Biliary stent placement

Under general anesthesia, laparotomy was performed to expose the bile duct, as previously described [55,56]. Subsequently, a 2-cm longitudinal epigastrium incision was made to identify the ampulla of the Vater of the duodenum. The duodenum was slightly punctured using an 18G angiocatheter (with needle). The needle was removed while the angiocatheter was held in place. Under fluoroscopic guidance, a 0.014-in microguide wire was negotiated into the common bile duct (CBD) through the angiocatheter. The angiocatheter was advanced over the guide wire across the ampulla of Vater into the CBD. Pre-procedural cholangiography was performed via the angiocatheter using a contrast medium to check the anatomical variables of the CBD. The diameter of the stent was selected considering the mean diameter (2.31 ± 0.51 mm) of the normal CBD in rabbits weighting 2.8–3.4 kg [55,56]. The 3-mm-diameter and 10-mm-long balloon catheter, which was a crimped PCL stent, was inserted over the guide wire into the CBD. The balloon catheter was fully inflated utilizing a contrast medium to 6 atm, as determined by a pressure-gauge monitor under continuous fluoroscopic guidance. The balloon catheter and guide wire were pulled out of the CBD after placing the stent. To confirm the stent location and identify any complications related to its placement, post-procedural cholangiography was performed immediately after the placement. Analgesics (ketorolac trometamol, 30 mg/1 mL) and antibiotics (gentamicin, 80 mg/2 mL) were routinely employed for 3 d to control postoperative pain.

2.2.11. Follow-up cholangiographic and hematological examinations

Before sacrificing the rabbits, follow-up cholangiography was performed at 4 wk after stent placement to check stent patency and position. Immediately before stent placement and sacrifice, a 3 mL intravenous blood sample was collected via the ear vein of the rabbits. The results of normal hematologic values were used as a reference in the analysis of the experimental group. The samples were separated into serum and stored in a deep freezer at −80 °C. Blood samples of all groups were analyzed for hepatobiliary functions, including ALT, ALP, GGT, and TB levels.

2.2.12. Gross and histological examinations

Surgical exploration of the duodenum, liver, and biliary duct was conducted. The degrees of granulation tissue formation and biliary sludge were evaluated by gross examination [55,56]. Tissue samples were fixed in 10% neutral buffered formalin for 48 h, and the middle portion of the stented CBD was transversely sectioned for histological analysis. The slides were stained with H&E and MT. Histological evaluation using H&E determined the degree of submucosal inflammatory cell infiltration, thickness of submucosal fibrosis by eight points around circumference average, and percentage of granulation tissue formation of ductal cross-sectional area stenosis was calculated as follows [[57], [58], [59]]:

$$Granulationtissueformation(\%)=100\times \left(\frac{stenoticareawithstent}{originalareawithstent}\right)$$

The percentage of the connective tissue area was calculated as

$$Connevtivetissuearea(\%)=100\times \left(\frac{\text{connective}\text{area}}{\text{original}\text{area}}\right)$$

and degree of collagen deposition was estimated using MT-stained slices [58]. The degrees of inflammatory cell infiltration and collagen deposition were subjectively determined using a 5-point scale: 1 indicates mild; 2, mild to moderate; 3, moderate; 4, moderate to severe; and 5, severe. All samples were analyzed using a digital slide scanner (Pannoramic 250 FLASH III, 3D HISTECH Ltd., Hungary), and measurements were taken with a digital microscope viewer (CaseViewer, 3D HISTECH Ltd., Hungary).

2.2.13. Immunohistochemistry

Immunohistochemistry analysis was performed on paraffin-embedded slices using TUNEL, α-SMA (1:200), and Ki67 (1:250) as primary antibodies. The extents of TUNEL, α-SMA, and Ki67-positive deposition were subjectively determined based on cell density and distribution: 1, mild; 2, mild to moderate; 3, moderate; 4, moderate to severe; and 5, severe.

2.2.14. Statistical analyses

Data were expressed as a mean ± standard deviation. Differences between the groups were analyzed using the one-way ANOVA or Kruskal–Wallis or Mann–Whitney U test, as appropriate. P-values <0.05 were considered statistically significant. For p < 0.05, a Bonferroni-corrected Mann–Whitney U test was used to identify the group causing the differences (p < 0.008 as statistically significant). SPSS software (ver 27.0, IBM, USA) was used to perform all statistical analyses.

3. Results and discussion

3.1. 3D printing of functionalized PCL stents

Although PCL, one of the candidate materials for BBSs that can function as a temporary scaffold during bile duct treatment, is biocompatible and biodegradable without any residual product, multitudinous materials, including organic and inorganic biomaterials, are combined with PCL to address its innate hydrophobicity and insufficient mechanical properties. To overcome these disadvantages, the application of metal oxides in the polymer matrix has been used in various biomedical fields. Metal oxides act as a drug delivery platform and reinforcement for enhancing physical and biological properties. Consequently, this approach has been used to fabricate the stent and coat stent for speedy formation of the endothelium, resulting in sufficient long-term stent outcomes. In our previous study, we explored how the electro-spun PCL and coated PCL can ameliorate the biological responses and mechanical properties through collaborative use of PCL with bioceramics [35]. Particularly, silica prepared via sol–gel process exhibited excellent wettability, bioactivity, and ease to embed it into the polymer-based materials without any segregation or phase separation. Consequently, a uniform nanostructured PCL/silica fiber or coating layer was attained with improved mechanical strength, elasticity, hydrophilicity, and cellular affinity. Based on these results, we designed and fabricated a novel 3D-printed BBS, containing a functionalized PCL by sol–gel derived silica with multifunction. The incorporated silica in PCL is expected to provide sufficient mechanical stability and controlled the degradation rate, whereas the exposed silica from PCL can provide biocompatibility.

To measure the nanoparticle size of silica in PCL solution, DLS and high resolution TEM were used. As shown in Fig. 1A, a hydrodynamic size distribution analyzed by DLS was found a peak in the 70–90 nm range, in silica in PCL solution, and this result agreed well with the TEM result which showed that the 60–90 nm range of spherical NPs were dispersed in the PCL matrix. A scanning transmission electron microscope (STEM) images and elemental mapping analyses using EDS proved that spherical NPs were silica contained Si and O. In addition to the Si and O signals from the silica, the C and O element signals in trace of typical polymer were detected and uniformly dispersed throughout the fPCLs, indicating that sol–gel derived silica NPs were well-dispersed in PCL matrix (Fig. 1B and Fig. S1).

Fig. 1.

Structural and chemical characterizations of 3D-printed functionalized PCL (fPCL) stents. (A) Size distribution of sol–gel derived silica NPs in PCL measured via DLS analysis and (B) TEM, STEM and elemental mapping images of sol–gel derived silica NPs in PCL. (C) FT-IR spectra of sol–gel derived silica, PCL, fPCL_Si20. (●: Si–O–Si stretching, ○: Si–OH stretching, □: O–H stretching, ■: C–O–C stretching, ▲: C–C and C–O stretching, ♦: C O stretching, ♣: CH₂ stretching). and (D) TGA and (E) DSC curves of fPCL with various silica contents of 0–30 wt%. (F) Images of 3D-printed fPCL_Si20 stents with various sizes and geometries, and (G) representative FE-SEM images of 3D-printed fPCL_Si20 stent with the entire surface, exterior surface and cross section of the stent strut. FE-SEM-EDS analyzed chemical compositions of 3D-printed pristine PCL and fPCL_Si20 stents.

FT-IR analysis was conducted to explore the functional bonding of fPCL. The FT-IR spectra of the sol–gel derived silica (Fig. 1C) exhibited characteristic peaks corresponding to the silicate network, including the asymmetric stretching vibration of Si–O–Si at 1064 cm⁻¹, the asymmetric bending and stretching vibration of Si–OH at 943 cm⁻¹, the symmetric stretching vibration of Si–O–Si at 833 cm⁻¹, and the stretching vibration of O–H at 3309 cm⁻¹ [60,61]. The pristine PCL spectrum revealed prominent peaks related to specific bonds, such as the C–O–C stretching vibration at 1167 cm⁻¹, asymmetric C–O–C stretching vibration at 1239 cm⁻¹, C–C and C–O stretching vibration at 1295 cm⁻¹, asymmetric C O stretching vibration at 1722 cm⁻¹, and asymmetric and symmetric CH₂ stretching vibrations at 2926 cm⁻¹ and 2855 cm⁻¹, respectively [35]. In the case of fPCL, we also observed the peaks of the combined silica and PCL, confirming the successful incorporation of sol–gel derived silica into PCL. However, compared with the FT-IR spectra of silica and fPCL, as shown inFig. S2A, the hydroxyl group (O–H bond) peak shifted from 3309 cm⁻¹ to 3294 cm⁻¹, whereas the carbonyl group (C O bond) peak in fPCL shifted from 1722 cm⁻¹ to 1720 cm⁻¹ compared to that of pristine PCL, manifesting that intermolecular hydrogen bonding was formed between the hydroxyl group of silica and carboxyl group of PCL. A similar phenomenon has been reported by other researchers who investigated the polymer-based hybrid materials and nanocomposites incorporated with silica NPs [[62], [63], [64]]. Owens et al. classified the inorganic–organic hybrid materials into Class I and Class II hybrids based on the chemical interaction between inorganic and organic materials. Class I hybrids generally possess only weak interactions, such as the hydrogen bond, van der Waals bond, π–π interaction, and electrostatic forces between each inorganic and organic phases. Class II hybrids contain strong bonds such as the covalent or ionic-covalent bond between each phase [65]. Class I hybrids, which are synthesized by hydrolysis–condensation of metal alkoxides in organic polymers, have been shown well-dispersed nanocomposites with hydrogen bonding between the carboxyl, carbonyl, and amide groups attached to the biopolymer and hydroxyl groups of sol–gel-derived metal oxides, consequently affecting the physical properties of materials. Based on our FT-IR results, we cautiously conclude that the synthesized fPCL might have hydrogen-bonded silica and PCL, which led to enhanced mechanical and physicochemical properties (Fig. S2B).

The thermal tolerance performance of fPCL with various contents of silica was examined using TGA and DSC. From the TGA graphs of the functionalized polymers, as shown in Fig. 1D, it can be verified that all of the specimens have gone through single-step degradation. However, pristine PCL did not show any remarkable weight change up to decomposition temperature (∼350 °C), whereas various fPCLs were not thermally stable with weight loss before degradation of PCL. This state was certainly observed as silica contents increased owing to the condensation of silanol groups in silica, resulting in the loss of moisture during heating [66]. Notably, as shown in Fig. S3, there was a remarkable increase in the decomposition temperature of fPCLs with an increase in the sol–gel derived silica content, demonstrating the PCL functionalized by silica possessed higher thermal stability compared with that of the pristine PCL. The physical or chemical reinforcements within a polymer matrix can affect the thermal properties such as the degradation temperature [[67], [68], [69]]. In this study, the thermal stability of fPCL increased with an increase in sol–gel derived silica loading. This enhanced thermal stability could provide a favorable environment, such as a wide 3D printing temperature window, for extrusion and 3D printing for commercial processing. Moreover, the decomposition process was accomplished above 500 °C, and the remaining fraction at 600 °C increased from 2.4 wt% for the pristine PCL to 28.9 wt% for the fPCL_Si30, suggesting that the silica contents were in line with the designed composition values. The melting temperature (T_m) analyzed using DSC was also slightly increased from 62.2 °C to 64.5 °C (i.e., 2.3 °C higher) for the pristine PCL and fPCL_Si30 (Fig. 1E), attributed to a marginal change in the crystallization kinetics of fPCL [70]. Generally, reinforcement in a polymer matrix can cause heterogeneous crystallization with incomplete crystals, consequently increasing T_m [71]. Thus, interactions between silica and PCL (i.e., hydrogen bonding between them) can lead to further deterioration in crystallinity, resulting in an increase in the T_m of fPCLs.

Hilar cholangiocarcinoma, which occurs at the point where the left and right intrahepatic bile ducts and the CBD meet, infiltrates in at least six complex patterns at that point [72]. Although various types and shapes of biliary stents for hilar cholangiocarcinoma have been developed and investigated, there is still insufficient consensus on optimal stent application [73]. Therefore, successful biliary stent placement in hilar cholangiocarcinoma of various patterns requires patient-customized and functionalized 3D printed stents according to the pattern of the lesion. The synthesized fPCL was used to fabricate BBSs with various geometries and sizes via the 3D printing process. Pristine PCL and fPCL with silica contents ranging from 10 to 30 wt% were heated and extruded from the 3D printer by adjusting the parameters as illustrated inTable S1. With the 0°-180°-0°reciprocating movement of the stent printer, the fPCL stent struts with various 3D trajectory designs could be extruded using the stainless-steel mandrel, as depicted in Fig. S4. The tubular architecture of the 3D-printed fPCL stents was maintained, and stents with various diameters, lengths, and designs could be fabricated as desired (Fig. 1F). As shown in Fig. 1G, the 3D-printed stents exhibited a typical stent structure with uniformly arranged bridges as designed using CAD. In relatively high magnification, a dense and smooth microstructure was observed in a 3D-printed fPCL stent. The 3D-printed stent strut had an elliptical shape with approximately 280 μm long axis diameter and approximately 200 μm short axis diameter. The EDS results also confirmed the presence of silica in fPCL; approximately 2.5 at% of silica was homogeneously distributed throughout the 3D-printed stent strut. This result is in good agreement with TEM-EDS results and suggests that the fPCL prepared through sol–gel process enabled the silica NPs to be well distributed within the PCL matrix.

3.2. Characterization of 3D-printed fPCL stents

Although the biodegradability and biocompatibility of the fPCLs are of primary consideration, appropriate mechanical properties are essential for the desired function and performance of the 3D-printed stents while expanding and supporting the bile ducts. Therefore, preliminary tension and compression evaluations were performed to quantitatively compare the 3D-printed fPCLs with various silica contents in terms of their resistances to extrinsic bile duct deformations. As depicted in Figs. S5A and S5B, the tensile properties of the fPCLs incorporated with silica were characterized to verify both the ability to withstand the architecture and the stress applied during the expansion of the ducts. All 3D-printed fPCLs exhibited a typical ductile semicrystalline polymer profile under tension: an initial linear elastic response deviating into a plastic flow behavior known as yielding; middle extended strain softening followed by a strain hardening region resulting from the strengthening of the fibrillar morphology axially oriented along the drawing force; and finally a sharp stress drop regime (Fig. S5A) [74,75]. The result of tensile tests demonstrated that the mechanical response of 3D-printed specimens shifted from flexible and weak to stiff and strong with an increase in sol–gel derived silica NPs. The tensile yield strength of all the 3D-printed fPCLs (ranging between 15 and 16 MPa) was 30% higher than that of the bare PCL (11.7 ± 0.2 MPa), demonstrating that the incorporation of well-dispersed, homogeneous, and fine silica NPs enhanced the yield strength (Fig. S5C). The effect of incorporated silica NPs in PCL was evident in tensile modulus values; as the silica contents increased from 0 to 30 wt%, the tensile modulus of the 3D-printed specimen increased from 272 ± 4 MPa to 399 ± 5 MPa, an approximately 1.5 times difference as illustrated in Fig. S5D. However, due to sufficient strain hardening effects, significant difference in ultimate tensile strength within the range of 16–18 MPa between all the 3D-printed specimens did not occur for 3D-printed fPCL specimens (Fig. S5E). For the 3D-printed PCL specimens, the steady plastic deformation occurring after the necking phase increased up to 220% elongation and then underwent rapid strain hardening (induced strain crystallization) until breakage at 500% elongation. However, both the end point of plastic deformation and the break point sharply decreased as the contents of silica increased (Fig. S5F). This result showed good agreement with our previous tensile results and confirmed that the ductile fracture behavior of the pristine PCL changed to relatively brittle fracture behavior with excessive amounts of silica NPs [28]. Although the introduction of sol–gel derived silica to PCL has been found to necessitate a tradeoff in the yield strength and stiffness besides the reduction in ductility, yield point (∼7%) and ultimate tensile strength values of each specimen did not noticeably change and the mechanical properties did not critically affect to the clinical use in terms of applied strain to the curved stent strut region during stent expansion (Fig. S5) [44,76]. Moreover, considering that the tensile strength, stiffness, and elongation of human blood vessel are 1.4–13.1 MPa, 4.5–130 MPa, and 11–242 %, respectively, the fPCL with enhanced tensile properties ensures flexibility in the stent design including the strut thickness reduction when compared with the bare PCL [77,78]. Consequently, fPCL can reduce the biliary side effects that could arise from the mechanical mismatch and be suited for the 3D-printed stent for use in a bile duct application.

To further investigate the mechanical properties of the 3D-printed fPCL specimens, compression tests were performed. The representative stress–strain behaviors of the 3D-printed fPCL are shown in Fig. S6A. All 3D-printed specimens exhibited a typical ductile thermoplastic behavior under compression, comprising three distinct regions: initial linear elastic response, roughly constant stress plateau region, and densification region. Notably, the compressive stress versus strain curves of 3D-printed fPCL specimens were significantly different from that of PCL, and all of these functionalized composites were stiffer and considerably stronger than the pristine PCL, as shown in Fig. S6A. The 3D-printed pristine PCL specimen had the lowest compressive yield strength among the four specimens (15.6 ± 3.0 MPa). As the silica content in PCL increased from 10 to 30 wt%, that of each fPCL specimen increased from 22.8 ± 1.1 to 35.6 ± 2.5 MPa. Thus, the 3D-printed fPCL with a silica content of 30 wt% (fPCL_Si30) exhibited compressive yield strengths exceeding double that of the bare PCL specimen (Fig. S6B). The silica-incorporated PCL also presented progressively stiffer property. As the silica content in PCL increased, the modulus value gradually increased. The fPCL_Si30 specimen exhibited a compressive modulus (322.2 ± 8.0 MPa) approximately double that of the pristine PCL (166.5 ± 6.9 MPa). In this study, the fPCL interaction via hydrogen bonding efficiently increased the interaction between sol–gel derived silica NPs and PCL without any noticeable agglomeration, increasing the compressive properties. Although the 3D-printed fPCL with high contents of silica remarkably enhanced most of the mechanical properties, the reduction in ductility of the polymer was inevitable even without aggregated silica NPs. This brittleness can cause the failure of the stent during the crimping and expansion processes even though its ductility is higher than that of the commercial PLLA, which is commonly used as a stent material. Therefore, fPCL_Si20 was selected as the base 3D printing material for further nanoengineered treatment and characterization.

3.3. Characterization of 3D-printed stent with nanoengineered surface

In this study, Zn ion implantation was applied onto the surface of the sirolimus-coated PCL stent to evade biofilm formation by antibacterial effects and suppress the initial burst of drug [79,80]. Despite their excellent antibacterial properties, the direct application of Zn as a coating material on the polymer surface is still limited because they have different structures and mechanical properties from those of polymer substrates. The proposed S-PIII process is an effective surface-modification technique; high-energy ions are physically embedded into the surface of the polymer substrate, resulting in their dispersion within a specific depth of the substrate without the need for an additional layer while addressing the limitations (e.g., cracking and detachment) of conventional coatings. With a negative bias (2000 V) on the substrate holder, Zn ions with high energy (10 keV) could be accelerated enough to be implanted on the near surface of the sirolimus-loaded PLLA on fPCL, leading to an implanted Zn layer on the inside surface.

To compare the Zn nanolayer before and after the Zn S-PIII treatment, surface and cross-sectional morphologies were observed using FE-SEM, FIB-TEM, and EDS, as shown in Figs. S7 and 2A. A microscopic view of each 3D-printed specimen showed slightly different surface microstructure. The microstructure of the PCL stent was much like the typical 3D-printed PCL strut with a rough surface because of micropores which is in agreement with the previously reported findings (Fig. S7) [81,82]. Furthermore, only carbon and oxygen were distributed over the whole surface of PCL. However, the 3D-printed fPCL had a relatively smooth surface with an additional Si element, which enabled the fPCL as a 3D printing material to fill pores during printing owing to its higher viscosity. The smooth surface was more clearly observed in SRL@fPCL by partially dissolving and coating it with a sirolimus-loaded PLLA/1,4-dioxane solution. The level of Si elemental distribution decreased due to the newly formed sirolimus layer. The Zn S-PIII treated 3D-printed SRL@fPCL stent was not significantly different from the SRL@fPCL specimen. EDS mapping data showed a small amount of Zn uniformly distributed with approximately 5 at% in the surface of the Zn-SRL@fPCL stent.

The cross-sectional morphology of the specimen was observed using FIB-TEM, EDS, and XPS to further evaluate the thickness and chemical composition of the Zn layer after the S-PIII treatment (Fig. 2A). Only two distinct regions composed of the unmodified fPCL substrate and platinum layer were confirmed by the TEM image and EDS line profile analyses. After the Zn S-PIII process, a relatively brighter contrast region with a thickness of 20 nm was observed at the topmost surface of fPCL, suggesting that Zn ions were implanted onto the fPCL surface within the depth of 20 nm without a distinct interface between the Zn layer and fPCL substrate. The parabolic line profile exhibited that the Zn elemental concentration increased up to approximately 2.7 at% before decreasing slowly with an increase in the depth from the top surface (Fig. S8). XPS was used for further analyses of the chemical compositions of 3D-printed fPCL surface before and after the Zn S-PIII treatment. As shown in Fig. 2A, the characteristic peaks for O 1s (532 eV) and C 1s (285 eV), which correspond to Si–OH, Si–O–Si, O C–O, and C–C, were present in the surface of all specimens regardless of the Zn S-PIII treatment [[83], [84], [85]]. Moreover, all specimens exhibited the peaks of the binding energies at 153 and 103 eV, which were assigned to Si 2s and Si 2p from silica in fPCL. In particular, the Si 2p peak is attributed to Si–OH (103.2 eV) and Si–O (102.4 eV) [84]. In contrast, Zn-related peaks including the Zn 2p peaks at 1045 eV and 1022 eV, Zn 3s peak at 140 eV, Zn 3p peak at 90 eV, and Zn 3d peak at 10 eV were noticeable in Zn-SRL@fPCL. In the high-resolution XPS results shown in Fig. S9, Zn2p_1/2 and Zn2p_3/2 peaks are also evident. These peaks correspond to the characteristic peaks of ZnO, which shows a successfully incorporated Zn layer on the near surface [[86], [87], [88]].

Fig. 2.

Chemical and physical characterization of 3D-printed fPCL stents with nanosurface. (A) Cross-sectional STEM images and STEM-EDS elemental profiles with C, O, Si, and Zn of 3D-printed fPCL stents before and after Zn S-PIII treatment. Red lines in STEM images indicate the region where the EDS analysis was performed. XPS full spectra of before and after Zn S-PIII treatment. (B) Wettability of various 3D-printed stents (**p < 0.01, ***p < 0.005, and ****p < 0.001). In vitro mechanical properties of (C) radial force curves for various 3D-printed stents and (D) their quantitative values including radial force, and recovery rate (***p < 0.005, ****p < 0.001).

Hydrophilicities of fabricated stents were evaluated by measuring the contact angle between water and their surface. As demonstrated in Fig. 2B, the contact angle decreased in fPCL to 50 ± 2° when compared with 77 ± 2° for PCL, which was attributed to hydrophilic silica NPs embedded in the PCL matrix [35,61,89]. However, the contact angle increased to 69 ± 2° after applying the SRL-loaded PLLA coating on fPCL, and this reduced hydrophilicity was caused by the coverage of PLLA on the top of the fPCL surface, similar to results reported previously [90,91]. A comparable contact angle was observed in the Zn-SRL@fPCL specimen (71 ± 1°) compared with that of the SRL@fPCL specimen.

The 3D-printed stent as a clinical biliary stent must have sufficient mechanical properties including radial force and recovery rate for its successful fixation and deployment. Therefore, to quantitatively compare the various 3D-printed stents in respect of their mechanical resistances, a compressive radial force analysis was performed, as shown in Fig. 2C. All 3D-printed stents exhibited elastic properties after 1.5 mm of compression–expansion cycles. Their load versus displacement graphs exhibited almost full recovery. After functionalizing PCL, the compressive force was 5.9 ± 0.2 N, and this value was approximately 50% higher than that of the 3D-printed PCL stent (3.3 ± 0.1 N, p < 0.001), suggesting that the radial force of the 3D-printed fPCL stent increased to a similar level compared with the compression test results (Fig. 2C and D). After coating with sirolimus and treating with Zn S-PIII, the values of radial force marginally increased to 5.8 ± 0.2 N and 5.9 ± 0.3 N, respectively. However, there were no significant differences in these values compared with those of the 3D-printed fPCL stent, as outlined in Table S2. Note that, as depicted in Fig. S7, ultra-spraying of the sirolimus-loaded PLLA solution acted as an organic solvent post-treatment, which remarkably reduced the stress concentration resulting from the surface defects or void, consequently slightly enhancing the mechanical properties [92,93]. The radial force of the 3D-printed Zn-SRL@fPCL stent was comparable to that of commercially used biliary self-expandable metallic stents [94]. All stents recovered to their initial strut without any noticeable sign of damage or permanent deformation, which is a key requirement of catheter-assisted stent implantation. Recovery rates after releasing the load from all the fPCL-based stents slightly decreased to below 90% compared with that of the PCL stent without statistical difference (p > 0.05), which could offer the necessary intra-biliary support for surgical implantation. Based on the tension and compression tests (Figs. S5 and S6), the improved stiffness due to the increase in silica content is expected to lead to a reduction in the recovery rate.

3.4. In vitro degradation of 3D-printed Zn-SRL@fPCL stents

For real-world applications, 3D-printed biodegradable stents must support and reopen the narrowed bile ducts and exhibit an appropriate degradation rate such that they degrade after the damaged bile duct tissues have recovered. Therefore, the degradation profiles the stent specimens were characterized by monitoring the changes in their surface morphologies, weights, and radial forces up to 12 wk, which is sufficient time to heal the impaired bile duct. As shown in Fig. 3A, the physical change was assessed by analyzing the stent surface using FE-SEM. For the 3D-printed PCL and fPCL, the surfaces were mainly degraded with initial rough surface by micropores. Particularly, the boundaries of fPCL were degraded considerably as the incubation time increased. A conspicuous intaglio surface was detected at week 12. The initially smooth surface of SRL@fPCL became randomly broken and cracked over time. In contrast, relatively fewer voids appeared on the surface of the Zn-SRL@fPCL stent. Note that the homogeneously distributed Zn element was detected over the entire surface of the stent even until 12 wk of the immersion test, suggesting that the Zn S-PIII treatment ensured excellent surface stability. In terms of weight loss after the degradation test (Fig. 3B), the PCL stent exhibited minimal weight loss (∼98.7%) after week 12, whereas 93.3 ± 3.5% of the initial weight remained for the fPCL stent after week 12. It is well known that sol–gel derived silica NPs degrade by themselves, resulting in a relatively faster decline in the weight loss of fPCL (Fig. 3B) during the long-term degradation test [35]. Incorporated sol-gel based silica nanoparticles containing polysiloxane network are well-known degradable materials in water. When sol-gel based silica nanoparticles meet water, following reaction occurs [95,96]:

$$\text{Si}{\mathrm{O}}_2\left(\mathrm{s}\right)+2{\mathrm{H}}_2\mathrm{O}\to \text{Si}{\left(\text{OH}\right)}_4\left(\text{aq}\right)$$

Fig. 3.

Degradation evaluation of various 3D-printed stents in vitro over 12 wk. (A) Representative surface morphologies as obtained using FE-SEM (Scale bar: 50 μm), and (B) weight changes in 3D-printed PCL, fPCL, SRL@fPCL, and Zn-SRL@fPCL stents at 0–12 wk. For Zn-SRL@fPCL stent, surface morphology and Zn elemental changes (insets) in the surface were observed. (C) Accumulated concentrations of released Si ions from surfaces of 3D-printed fPCL, SRL@fPCL, and Zn-SRL@fPCL stents. (D) Accumulated concentrations of released Zn ions from surfaces of 3D-printed Zn-SRL@fPCL stent over 7 d, and (E) sirolimus release behavior from 3D-printed SRL@fPCL and Zn-SRL@fPCL stents.

Thus, the rate of weight decrease in fPCL is markedly higher than that of pristine PCL.

On the contrary, the weights of the 3D-printed SRL@fPCL and Zn-SRL@fPCL stents remained nearly unchanged (98.5 ± 1.5% and 98.1 ± 2.3%), suggesting that the nanoengineered surface treatment suppressed the release of silica. Because the regeneration of bile ducts takes approximately 12 wk, the 3D-printed Zn-SRL@fPCL stents are capable of maintaining structural integrity and increase the degradation rate of PCL to more closely match the regeneration rate of impaired bile ducts.

3.5. In vitro metal ion and drug release in 3D-printed Zn-SRL@fPCL stents

For effective therapeutic delivery at impaired bile ducts, a prolonged delivery of the therapeutic drug/molecules is preferred over the burst-type delivery of drug/molecules to maintain an optimal dose during the biliary stricture treatment period. Along with the degradation properties of the 3D-printed Zn-SRL@fPCL stents, the in vitro release behaviors of both metal ions and sirolimus were monitored by immersion the stents in an artificial plasma solution for up to 12 wk.

The release of Si ions from the stents after the functionalization of PCL was monitored and compared (see Fig. 3C). The initial burst releases of Si ions were noted in fPCL and SRL@fPCL stents to be 28.4 ± 4.2 mg/L and 21.9 ± 5.0 mg/L after 12 h, respectively. Compared with fPCL and SRL@fPCL, Zn-SRL@fPCL exhibited a smaller initial burst release (4.9 ± 1.1 mg/L at 12 h). After 1 d, almost identical release profiles were noted in all experimental groups (Fig. S10A). Consequently, the order of accumulative Si ion release amount was fPCL > SRL@fPCL > Zn-SRL@fPCL. After 12 wk, the accumulated Si ion concentrations were 68.5 ± 10.8 mg/L (fPCL), 61.3 ± 10.4 mg/L (SRL@fPCL), and 40.6 ± 6.3 mg/L (Zn-SRL@fPCL). The slow release of Si ions from the SRL@fPCL and Zn-SRL@fPCL stents compared with that of the fPCL stents was mainly attributed to the additional PLLA coating on the fPCL stent and Zn ion implantation. The release path of Si ions should be longer because of the supplemented sirolimus containing PLLA coating, and this type of additional polymeric coating has been acknowledged to act as a barrier for the release of ions or drugs [39,97,98]. The Si ion release rate was further suppressed in the Zn-SRL@fPCL stents, which was attributed to the formation of a Zn ion-based nanolayer in the PLLA coating. The metal ion-based layer acts as a physical obstacle in releasing the drug molecules because of the decreased effective spacing between surface polymer chains [40].

The stability of the nanosurface was further evaluated by measuring the concentration of Zn ions released in the same condition as used for the Si ion release test, as shown in Fig. 3D. Compared with the amounts of the released Si ions, relatively tiny amounts of Zn ions were released from the 3D-printed Zn-SRL@fPCL stent; the implanted Zn ions were released over time with an initial burst phenomenon. After 1 wk, almost 95% of Zn release was noted in the release profile (Fig. S10B). The total amount of released Zn ions was only 18.1 ± 3.1 μg/L over 4 wk. Findings from FE-SEM and EDS analyses, as shown in Fig. 3A, indicate no observable disparities in surface morphologies between pre- and post-immersion states after 12 wk. Moreover, a considerable amount of Zn was still detected on the surface of the 3D-printed Zn-SRL@fPCL stent. The combination of its excellent chemical and mechanical stabilities makes it a promising for providing long-term protection against the occurrence of clinical complications in biliary stents.

Nanotechnology and drug agents with biliary stents have been actively investigated to overcome current biliary stent-related limitations in clinical applications [56,[99], [100], [101]]. Various functionalized biliary stents with nanotechnology and antiproliferative drugs demonstrate promising treatment methods with improved clinical outcomes; however, further investigations are warranted to confirm their efficacy and safety through preclinical studies [[101], [102], [103]]. Moreover, additional research on the development of novel multifunctionalized biliary stents is required to address existing clinical requirements [104]. Fig. 3E depicts the sirolimus release profile of the 3D-printed SRL@fPCL and Zn-SRL@fPCL stents immersed in an artificial plasma solution for 12 wk. The release of sirolimus was complete within approximately 12 for all the 3D-printed stents wk. The initial burst effect of the 3D-printed SRL@fPCL stent occurred on the first day of immersion, releasing a considerable amount of sirolimus (∼12%). After 1 wk, only 49.3% of residual sirolimus was remained. In contrast, sirolimus could be sustainably released from the 3D-printed Zn-SRL@fPCL stent owing to the relatively linear release behavior of this stent. At the initial stage (up to day 7), it exhibited a sustained sirolimus release rate compared with that of the 3D-printed SRL@fPCL stent. After immersion for 1 wk, the release rate of sirolimus apparently slowed down (Fig. S10C). After 4 wk, the loaded sirolimus was almost completely released from the 3D-printed SRL@fPCL stent, whereas approximately 33% of sirolimus still remained in the nanoengineered layer of the 3D-printed Zn-SRL@fPCL stent. This significantly extended period of relatively linear release of the drug from the nanosurface is mainly attributed to the Zn ion-implanted nanolayer on the surface of sirolimus and PLLA coating. Therefore, we expect that the Zn S-PIII treated nanosurface can synergistically suppress cell activity and biofilm formation by suppressing the excessive initial burst of sirolimus.

3.6. In vitro antibacterial properties of 3D-printed Zn-SRL@fPCL stent

One of the most commonly occurring postoperative problems attributed to the reduced patency of biliary stent implantation is sludge formation due to the biofilm formation on the surface of the stent [3]. Bacteria with a high multiplication rate adhering to the stent form concrete biofilms. Thus, recent studies on biliary stents have adopted antibacterial strategies to suppress the biofilm formation [3,55,[104], [105], [106], [107], [108], [109], [110]]. Among various candidates, Zn-ion-releasing materials have been widely explored as antibacterial agents owing to their nonspecific antimicrobial characteristics and long-lasting release of ions [46,[111], [112], [113]]. The antibacterial effectiveness of the developed Zn-SRL@fPCL stent was assessed in two types of experiments: 1) antiadhesive and 2) bactericidal property experiments, which represent bacteria adhesion and proliferation as essential steps for biofilm formation. First, the antiadhesive property of the specimens was estimated by culturing E. coli and S. aureus on the top of the experimental specimens for 1 d. Many widespread bacteria were observed in PCL, fPCL, and SRL@fPCL, whereas fewer bacteria were observed in Zn-SRL@fPCL, as demonstrated in Fig. 4A and B. The biofilm formation test was performed using an in vitro biliary mimetic system (Fig. S11). A formulated solution containing ox bile and E. coli was circulated through a tube containing the prepared stents, and the structure of the formed biofilm was observed (Fig. 4C). The biofilms formed by the agglomerated ox bile and E. coli considerabily covered the surface of PCL, fPCL, and SRL@fPCL. Although biofilms were observed on the Zn-SRL@fPCL stents, their densities were considerably lower than those of the other experimental groups. The variations in the biofilm coverages were larger in low magnification images (inset in Fig. 4C). The quantitatively calculated biofilm areas were 56 ± 6% for PCL, 52 ± 9% for fPCL, 55 ± 9% for SRL@fPCL, and 6 ± 3% for Zn-SRL@fPCL, indicating significantly inhibited biofilm formation in the Zn-SRL@fPCL group.

Fig. 4.

Antiadhesive properties against bacteria. Morphologies of attached (A) E. coli and (B) S. aureus. (C) SEM images of biofilms composed of E. Coli and ox bile inside the stent and biofilm area at 5 d of incubation (insets: low magnification images) (**p < 0.01).

Next, the bactericidal ability of the specimens was assessed through the colony formation experiment, as shown in Fig. 5. The bactericidal properties exhibited an identical trend as that of the antiadhesive properties, i.e., the number of colonies (less than 10% compared with those of the other groups) formed by the detached E. coli (Fig. 5A and B) and S. aureus (Fig. 5C and D) considerably decreased for Zn-SRL@fPCL. Even though there were additional components, including silica, PLLA, and sirolimus, in fPCL and SRL@fPCL, these components had minimal effects on antibacterial properties of the specimens [114,115]. The aforementioned antiadhesive and bactericidal effects of Zn-SRL@fPCL are attributed to the incorporated Zn layers. The release of the Zn ions from the Zn nanolayer is attributed to the reduction of bacterial adhesion and viability on the surface. It has been previously reported that Zn ions inhibit glycolytic enzymes by the oxidation of the thiol group [116]. In addition, electrostatic interactions between positively charged Zn ions and negatively charged bacterial surfaces composed of teichoic and lipoteichoic acids are known to disrupt the bacterial surface [117].

Fig. 5.

Assessment of bactericidal activity. (A) Optical images and (B) average colony numbers of E. coli after 1 d of culturing on agar plate using diluted E. coli solution (***p < 0.005). (C) Optical images and (D) average colony numbers of S. aureus after 1 d of culturing on agar plate using diluted S. aureus solution (**p < 0.01).

3.7. In vitro cellular responses of 3D-printed Zn-SRL@fPCL stent

Fibroblast cell line was used to compare the ability of the inhibiting hyperplasia, as portal fibroblasts in bile duct epithelia, which is associated with biliary fibrosis, are similar to other fibroblast [118,119]. The adhesion and viability of fibroblasts on the surfaces of 3D-printed stents were compared after culturing for 24 h and 3 and 5 d, respectively. First, the adhesion morphology of fibroblasts on the Zn-SRL@fPCL stent was investigated by CLSM observation and compared with that of PCL, fPCL, and SRL@fPCL surfaces, as shown in Fig. 6A. After 24 h of culturing, fibroblasts on the PCL surface adhered well and demonstrated spread-out morphologies with many filopodia. The fPCL sample had a significant number of and well-spread out fibroblasts adhered to its surface, suggesting that the released Si ions slightly promote cell adhesion owing to better hydrophilicity because of incorporating silica in PCL.

Fig. 6.

(A) CLSM micrographs of fibroblasts on 3D-printed PCL, SRL@fPCL, and Zn-SRL@fPCL specimens after 24 h of culturing and (B) proliferation of fibroblasts on 3D-printed PCL, fPCL, SRL@fPCL, and Zn-SRL@fPCL specimens after 3 and 5 d of culturing (**p < 0.01, ***p < 0.005).

However, the numbers of fibroblasts adhered to the sirolimus-loaded stent surfaces (SRL@fPCL and Zn-SRL@fPCL) were fewer when compared with that of PCL and fPCL stents. These results showed that the introduction of sirolimus as an antiproliferative drug effectively suppressed the proliferation of fibroblasts. The in vitro cell viability of sirolimus-loaded stent surfaces was considerably inhibited, as shown in Fig. 6B. The fPCL stent surface exhibited the highest viability of fibroblasts on days 3 and 5 of culturing (117.2 ± 16.5% and 186.1 ± 14.6%, respectively), demonstrating thrombosis. Both the SRL@fPCL and Zn-SRL@fPCL surfaces revealed substantially suppressed proliferation of fibroblasts when compared with that of the PCL surface on days 3 and 5. This finding agrees well with the results of the cell adhesion experiment. However, the cell viability score of Zn-SRL@fPCL was higher than that of SRL@fPCL because of the good cell compatibility of Zn; however, there was no significant difference [120]. This result agrees well with the sirolimus release test results (Fig. 3). The Zn nanolayer existed on the surface of the sirolimus coating controlled initial burst of drug, thus suppressing hyperplasia in the bile duct stricture. Although surface nanoengineering required only a few tens of seconds, it enabled massive Zn ion implantation onto the fPCL stent at markedly high doses; therefore, it is possible to modify the nanosurface properties of the 3D-printed Zn-SRL@fPCL stent. The mechanical forces of the placed stent induce constant stimulation on the bile duct wall, leading to early inflammatory reaction and late tissue hyperplasia [121,122]. The late proliferative phase is characterized by the proliferation of fibroblasts and myofibroblasts with a decline of inflammatory cells. Stent-induced tissue hyperplasia causing in-stent restenosis occurs as an excessive fibroblast cells response to mechanical injuries after stent placement [119]. In addition, in vitro tests using bladder epithelial cells (representative of epithelial cells) were also conducted to reveal effectiveness of silica NPs and Zn nanolayer. As shown in Fig. S12A, similar morphology of attached bladder epithelial cells was monitored. And any sign of cytotoxicity was not observed in all of the experimental groups (Fig. S12B). Thus, incorporated silica NPs and Zn nanolayer attributed to better cytocompatibility.

3.8. Procedural outcomes of stent placement in rabbit CBD

Before the use of the fabricated stents in the bile duct, stents were crimped on balloons for their insertion and inflation into the bile duct (Fig. S13A). Crimping of stents was performed using commercial balloons, as illustrated in Fig. S13B. After crimping, the stretch of the stents after inflating the balloon was tested to check its possibility of failure (Fig. S13C and Movie S1). Notably, failure was not observed during the crimping and inflation of the stents, which is ascribed to their excellent mechanical properties, as depicted in Fig. 2C and D. The placements of the balloon-expandable PCL stents in all the rabbits were successful without any procedural complications, such as ductal perforation (Figs. S13D–S13F). All the rabbits survived until the end of the study. However, jaundice was observed at 10–18 d (mean 14 d) in six rabbits (PCL: 3, fPCL: 2, and SRL@fPCL: 1) (Fig. 7A and B). Furthermore, rapid weight loss was observed in all the rabbits during the first week after the surgical procedure and biliary intervention. The mean percentages of the weight losses in the PCL (−14.23 ± 3.78%) and fPCL (−13.21 ± 8.32%) groups were significantly lower than those of the SRL@fPCL (−3.29 ± 6.82%) and Zn-SRL@fPCL (0.82 ± 3.04%) groups at four weeks (Fig. S14). Weight loss and jaundice in patients are common indicators of acute or chronic biliary obstruction in a clinical setting [123]. Although all the rabbits experienced rapid weight loss after the surgical procedure, only the Zn-SRL@fPCL group gradually recovered at the end of the study without evidence of jaundice in any of the rabbits. Our results indicate that the application of the SRL and Zn ions to biodegradable PCL stents fabricated using 3D printing and nanoengineering successfully prevented stent-induced tissue hyperplasia and sludge formation caused by bacterial reactions. Although additional studies are required to evaluate the long-term efficacy of the multifunctionalized PCL stent, the Zn-SRL@fPCL stents maintained their patency for 4 wk after placement without stent-related complications.

Fig. 7.

Procedural, cholangiographic, and hematological outcomes of stent placement in rabbit CBD. (A) The original colors of rabbits were maintained in the all Zn-SRL@fPCL stent groups, (B) whereas the pristine PCL group rabbits showed evidence of jaundice through yellowing eyes and ears. (C) Pre-cholangiography showing the entire common bile duct of rabbit. Follow-up cholangiographies showing relatively good patency of (D) the Zn-SRL@fPCL stent (arrowheads) and dilated intrahepatic bile duct caused by occluded and (E) pristine PCL stent (arrowheads) due to tissue hyperplasia and sludge formation. The hematology assay showing changes in (F) ALT, ALP, GGT, and (G) TB levels in all experimental groups (*p < 0.05, **p < 0.01, ***p < 0.005 and ****p < 0.001).

3.9. Cholangiographic and hematological findings

Representative cholangiography findings are shown in Fig. 7. Cholangiography revealed the dilated CBD and intrahepatic duct in the PCL and fPCL stent groups caused by the luminal narrowing of the placed stent due to the stent-induced granulation tissue and sludge formation. However, the stent patency of the SRL@fPCL and Zn- SRL@fPCL stent groups were relatively well preserved at 4 wk of the follow-up cholangiography. The degree of the dilated intrahepatic duct was substantially higher in the PCL and fPCL stent groups than that in the SRL@fPCL and Zn- SRL@fPCL stent groups. Hematology assays were performed using the collected blood serum to evaluate hepatobiliary functions and hematological findings, as shown in Fig. 7F and G. The mean ALT, ALP, GGT, and TB levels considerably varied for the different groups. Compared with the PCL and fPCL stent groups, the mean ALT, ALP, GGT, and TB levels were considerably lower in the SRL@PCL and Zn-SRL@PCL stent groups (Table S3). However, there were no significant differences between the PCL and fPCL stent groups and SRL@PCL and Zn-SRL@PCL stent groups (Table S3). Hematology assays of the ALT, ALP, GGT, and TB levels are commonly used as indicators of the chemistry of liver function and disorders, including bile duct obstruction, cholestasis, and jaundice [55,124]. ALT, which is mainly present in the cell cytoplasm, is highly sensitive to liver cell injury and is a reliable marker for detecting both acute and subacute hepatocellular injuries [125,126]. ALP and GGT enzymes associated with cholestasis and originating from the hepatobiliary system are primarily present in epithelial cells, including those of the bile duct [125]. Moreover, TB is eliminated from the hepatocytes through the canalicular membrane and forms a network of tubules between adjacent hepatocytes, representing the key location for the synthesis of bile [127,128]. In this study, the liver function deteriorated in the PCL and fPCL stent groups owing to the luminal narrowing in the stented CBD with bacterial colonization and cellular proliferation. Consistent with the cholangiography findings, the multifunctionalized PCL stent with antibacterial and antiproliferation properties exhibited a relatively good patency and stable liver functions compared with those the PCL and fPCL stent groups. The ALT, ALP, GGT, and TB levels of the Zn-SRL@fPCL stent group were the lowest among the study groups; however, there was no statistical difference when compared with the levels of the SRL@fPCL group. This was attributed to a relatively short follow-up period to produce any noticeable effect of Zn ions. Compared with the PCL and fPCL stent, the SRL@PCL and Zn-SRL@PCL stent groups exhibited consistently better hepatobiliary functions in this study. As an antiproliferative drug, SRL successfully prevented stent-induced tissue hyperplasia in this study. Moreover, the Zn ions prevented sludge formation, caused by its antibacterial activity, on the surface of the PCL stent.

3.10. Gross and histological findings

Representative gross findings of this study are shown in Fig. 8A. Compared with the Zn-SRL@fPCL stent group, the sludge formation due to bacterial reaction was more salient in the PCL, fPCL, and SRL@fPCL stent groups because of the prominent biofilm formation adjacent to the stents. The degree of sludge formation in the SRL@fPCL group was higher than that in the Zn-SRL@fPCL group. The SRL@fPCL can prevent stent-induced tissue hyperplasia, but it does not have an anti-bacterial effect. Although relatively high sludge around the stent were formed in the SRL@fPCL group, stent patency was well maintained until the end of the study. The histological findings are summarized in Table S4. The mean degrees of submucosal inflammatory cell infiltrations, thicknesses of submucosal fibrosis, percentages of granulation tissue formation areas, percentages of connective tissue areas, degrees of collagen depositions, degrees of α-SMA-positive depositions, and degrees of Ki67-positive depositions substantially differed between the groups (Fig. 8B and Table S2). However, the mean degrees of TUNEL-positive depositions were not remarkably different between the groups (p = 0.051). The mean degrees of submucosal inflammatory cell infiltrations, thicknesses of submucosal fibrosis, percentages of granulation tissue formation areas, percentages of connective tissue areas, degrees of collagen depositions, degrees of α-SMA-positive depositions, and degrees of Ki67-positive depositions were considerably lower in the SRL@fPCL and Zn-SRL@fPCL stent groups than those of the PCL and fPCL stent groups. However, there were no statistical differences between the PCL and fPCL stent groups and SRL@fPCL and Zn-SRL@fPCL stent groups (p > 0.05). These findings suggest that the SRL@fPCL and Zn-SRL@fPCL stent groups successfully suppressed bacterial and cellular proliferation responses in the stented rabbit CBD. Furthermore, from the macroscopic observations, the histological analyses of the stented CBD showed a notable reduction in tissue hyperplasia and bacterial activities in the SRL@fPCL and Zn-SRL@fPCL stent groups compared with those of the pristine PCL and fPCL stent groups. With antiproliferative properties, SRL is a successful, potent immunosuppressive agent and macrocyclic lactone that inhibits the restenosis cascade, such as smooth muscle cell migration, collagen synthesis, and myofibroblast growth, by disrupting the transition in the G1–S phase [[129], [130], [131]]. Zn ions, which are nanosized metal oxides, realized antibacterial activity by oxidizing cellular contents from the penetrated particles to the bacteria membrane [132,133]. Herein, stent-induced tissue hyperplasia-related variables considerably decreased in the SRL@fPCL and Zn-SRL@fPCL stent groups. Unexpanded connective tissue and reduced fibrosis were observed in the SRL@fPCL and Zn-SRL@fPCL stent groups. The TUNEL assay, which is the principal marker of the apoptotic/necrotic cells, showed no differences among the various groups. Furthermore, compared with the pristine PCL and fPCL stent groups, decreased myofibroblasts in the α-SMA-positive deposition and decreased mitotic cells in the Ki67-positive deposition were observed in the SRL@fPCL and Zn-SRL@fPCL stent groups. Overall, histological factors were better in Zn-SRL@fPCL; however, the follow-up at 4 wk is a relatively short time for monitoring actual in vivo response. Our findings demonstrated that stent-induced tissue hyperplasia and sludge formation adjacent to the PCL stent in rabbit CBD were effectively suppressed using the versatile BBS.

Fig. 8.

Gross and histological findings. (A) Representative macroscopic and microscopic images 4 wk after stent placement in rabbit CBD. (B) Serial histological results of 3D-printed PCL, fPCL, SRL@fPCL, and Zn-SRL@fPCL stent groups in the stented bile duct (*p < 0.05, **p < 0.01, ***p < 0.005 and ****p < 0.001).

4. Conclusion

This study presents a novel approach for fabricating BBSs to achieve more effective treatment of biliary strictures with reduced biofilm formation and hyperplasia. This strategy involves the fabrication of stent struts using fPCL with silica NPs along with the application of Zn S-PIII to the sirolimus-coated surface of the 3D-printed stents, resulting in versatile nanosurface properties. The 3D-printed stents with various geometries and sizes demonstrated excellent shape fidelity and printability with improved mechanical properties for expansion and support of the stenosed bile ducts as the silica NP content in the fPCL increased. The Zn nanolayers formed on the sirolimus/PLLA-coated surface of the 3D-printed fPCL stents through the Zn S-PIII treatment effectively controlled the degradation rate and facilitated the release of Si ions and sirolimus. The surface of the 3D-printed biliary stents exhibited enhanced antibacterial properties and inhibited fibroblast proliferation, thereby preventing rapid sirolimus release through the Zn nanolayer and minimizing hyperplasia. An in vivo study also confirmed that the Zn-SRL@fPCL stent considerably inhibited the granulation tissue and biofilm formation with superior stent patency without hepatobiliary dysfunctions in the rabbit CBD. Although long-term follow-up study is necessary to evaluate the functional and effective changes in the proposed biliary stent in vivo, this study demonstrates that these highly biocompatible biliary stents have therapeutic potential in preventing stent-related complications while eliminating the requirement for stent removal.

Approval for animal experiments

The use of animals in this study was approved by the Institutional Animal Care and Use Committee of the institute and conformed to the US National Institutes of Health guidelines for the humane handling of laboratory animals (IACUC No. 2022-14-303).

Data availability

The raw data required to reproduce these findings can be downloaded from [INSERT PERMANENT WEB LINK(s)]. The processed data required to reproduce these findings are available to download from [INSERT PERMANENT WEB LINK(s)].

Ethics approval and consent to participate

All in vivo procedures including animal selection, surgical protocol, management, and sacrificial procedures were approved by US National Institutes of Health guidelines for the humane handling of laboratory animals (IACUC No. 2022-14-303).

CRediT authorship contribution statement

Hyun Lee: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Conceptualization. Dong-Sung Won: Writing – review & editing, Writing – original draft, Visualization, Validation, Investigation, Formal analysis, Conceptualization. Sinwoo Park: Validation, Formal analysis, Data curation, Conceptualization. Yubeen Park: Validation, Investigation, Formal analysis, Data curation. Ji Won Kim: Validation, Investigation, Formal analysis. Ginam Han: Validation, Methodology, Investigation, Formal analysis. Yuhyun Na: Validation, Investigation, Formal analysis. Min-Ho Kang: Writing – original draft, Visualization, Validation, Formal analysis. Seok Beom Kim: Visualization, Supervision, Software, Investigation, Formal analysis. Heemin Kang: Writing – review & editing, Supervision, Investigation, Formal analysis, Data curation. Jun-Kyu Park: Visualization, Supervision, Investigation. Tae-Sik Jang: Visualization, Validation, Investigation, Data curation. Sang Jin Lee: Visualization, Supervision, Conceptualization. Su A. Park: Visualization, Supervision, Formal analysis. Sang Soo Lee: Visualization, Validation, Investigation, Formal analysis. Jung-Hoon Park: Writing – review & editing, Writing – original draft, Validation, Supervision, Formal analysis, Data curation, Conceptualization. Hyun-Do Jung: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Methodology, Investigation, Formal analysis, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article.