Innovative 3D-printed porous piezoelectric poly(vinylidene fluoride) cages with accelerated spinal fusion

Abstract

Spinal fusion is an efficient treatment for degenerative disc disease, however, poor osseointegration capacity of spinal cages often cause clinical implant failure. Herein, we proposed an innovative piezoelectric interbody fusion cage with porous structure, load-carrying capacity and enhanced osseointegration effect via 3D printing technology. Firstly, poly(vinylidene fluoride) (PVDF), poly(lactic acid) PLA and ceramic barium titanate (BTO) were compounded to prepare the composite filament with improved 3D printability and good piezoelectric effect. It is found that both BTO and PLA are favored for β-phase transformation and PLA effectively solves the warping issue during the 3D-printing. Then, cages with triangular lattice with varying infill densities are designed and fabricated. It is found that triangular PVDF/PLA/BTO structure with lower infill density exhibits stronger piezoelectric effect, and all the samples exhibited compressive properties analogous to trabecular bone. Furthermore, the in vitro cellular assays and in vivo large animal model sheep experiments validated an obvious piezoelectric effect on osseointegration for PVDF/PLA/BTO cages. After three-month implantation, it presented promoted new bone formation with a bone volume of 76.8 % in contrast with that of 49.3 % in PEEK counterpart. Therefore, the current study highlights a promising material for intervertebral fusion cages with good biomimetic mechanical performance and spinal osseointegration effect.

Graphical abstract

1. Introduction

Spine vertebrae and intervertebral disc defects, which result from accident trauma or spinal diseases by tumor, spondylosis and degeneration, often lead to sustainable pain and even permanent disability [1,2]. Interbody fusion is considered as a critical surgical procedure to promote bone fusion between two or more vertebrae and releasing the compression of nerve roots [3]. The success of spine fusion is substantially determined by the properties of interbody fusion cage, therefore, the material and structure have sparked significant research enthusiasm over the past two decades [[4], [5], [6], [7]].

Currently, Titanium (Ti) and polyetherether ketone (PEEK) are the predominant materials [[8], [9], [10]]. It has been demonstrated that Ti-based cage possessed good bioactivity, corrosion resistance and high strength [8], whereas PEEK possess comparable elastic modulus with cancellous bone [9]. Despite these benefits, there are still drawbacks in clinical applications. The elastic modulus of Ti-based cages exceeded those of vertebral bone, causing high risk of stress shielding and subsequently prosthesis loosing and sinking [11]. On the other hand, the intrinsic biological inertness of PEEK as biomedical implants brings serous issue for osseointegration and thereby result in incomplete fusion and prolonged recovery periods [12,13]. Accordingly, conventional fusion cages face great challenges in insufficient osseointegration and satisfied mechanical properties [14].

To address these issues, innovative fusion cage strategies have been proposed. Firstly, three-dimensional (3D) interconnected porous structures have been utilized with a favorable impact on vascularization through nutrients, bone ingrowth and reduce overhigh elastic modulus [15,16], which are fabricated by additive manufacturing techniques. Fused deposition modeling (FDM) 3D printing has widely used due to the great advantages of low-cost, feasible printability, high precision and polymer designability, which support the rapid development for bone implants. Jia proposed a biomimetic porous 3D-printed PEEK cage in order to enhance the mechanical properties [17]. Animi also evaluated and compared the clinical performance of 3D-printed Ti-based versus PEEK cage [18]. These 3D-printed cage models also demonstrated the great application feasibility for cage manufacturing. Besides, bioactive coatings [18] or biological hydrogels [19] were incorporated on the bone implants to enhance the bioactivity and osseointegration capacity. These developments hold promise for facilitating tissue-tear healing [20], cartilage repair [21] and early intervertebral fusion.

Recently, the latest research proposed that bone itself is a natural piezoelectric material system, it could generate polarization and surface changes when suffered forces [19,20]. Triggered by biomimetic idea, piezoelectric materials with bioelectric stimulation effect offer a solution by mimicking bone native electromechanical properties. This electric stimulus is enable to regulate the osteoblast differentiation, enhance the vascular permeability and increase local blood flow, which exhibit a remarkable osseointegration capacity [21].

Poly(vinylidene fluoride) (PVDF), a thermoplastic polymer with excellent piezoelectric effect, has shown great potentials in energy harvesters [22], electronic skin [23], tendons, bone fixation [24] and scaffolds [25]. Especially, for biomedical applications, polarized PVDF-based bone implants could produce piezoelectric signals at the bone/implant interface, which showed effective osteogenic capacity and good biocompatibility [25]. As reported, the piezoelectric property of PVDF is greatly determined by the crystal polymorphs mainly including α (fluorine atoms alternatively arranged at both carbon backbone), β (fluorine and hydrogen atoms arranged parallelly on each side) and γ phases [26,27]. The conformations of the molecular chain were shown in Fig. 1. Of all the conformation forms, β-PVDF with this molecular structure possesses the highest dipole moment, resulting in a better piezoelectric effect, chemical stability and corrosion resistance [28]. In addition, PVDF-based composites also demonstrated the 3D-printability for custom-design and complicated parts [29,30]. Therefore, the combination of piezoelectric materials with porous structures could well meet the requirements of spinal implants, which provide an innovative strategy for fabricating 3D-printed piezoelectric interbody cages.

Fig. 1.

The piezoelectric effect of PVDF and the potential application as porous interbody fusion cage.

In the current work, a 3D-printed porous interfusion cage with good piezoelectric effect was designed and developed (shown in Fig. 1). Firstly, PVDF, barium titanate (BTO) and poly(lacitc acid) (PLA) were compounded and extruded into 3D-printing feedstock. The 3D printability, mechanical property, thermal behavior and piezoelectric characteristics of PVDF/PLA/BTO composites were studied. Subsequently, cages with internal porous structure were proposed, and the influence of infill density on the mechanical and piezoelectric effect were investigated. Finally, the 3D-printed piezoelectric cage with triangular lattice structure was implanted into sheep animal model to assess the osseointegration. The study highlights a favorable cage strategy that would facilitate better bone ingrowth and integration for spinal fusion.

2. Materials and methods

2.1. Materials

PVDF pellets (trade name FR906) with a weight-average molecular weight of approximately 5.3 × 10⁵ g/mol was purchased from Shanghai 3F New Material Co., Ltd, China. PLA (trade name 4032D) with a weight-average molecular weight of approximately 1.5 × 10⁵ g/mol was purchased from Natureworks, Inc. in the USA. Commercial barium titanate (BaTiO₃, 99.0 %, an average of 100 nm) was purchased from Shanghai Aladdin Biochemical Technology Co. Ltd. The low-glucose DMEM medium, RPMI-1640 medium, fetal bovine serum, penicillin, streptomycin, and 0.25 % EDTA were purchased from Gibco, Grand Island, New York, USA.

2.2. Fabrication of PVDF/PLA/BTO filaments

PVDF and PLA pellets were pre-dried in an oven at 60 °C for 8 h. PVDF, PLA and BTO were melt-compounded in the following mass ratios: (1) 97 % PVDF and 3 % BTO (2) 85 % PVDF and 15 % PLA; (3) 82 % PVDF, 12 % PLA and 3 % BTO. These materials were pre-mixed and sequentially fed into a co-rotating twin-screw extruder (Nanjing Giant Machinery Co., Ltd., SHJ-20, China). The temperatures from the fed zone to the fifth die zone were 170, 210, 215, 220 and 230 °C, respectively, and the screw speed was 300 rpm. The extrudates were cooled in a water bath, cut into pellets, collected, and dried. Afterwards, a single-screw extruder (Wellzoom line 2, Shenzhen Misda Technology Co., Ltd., China) was used to fabricate the 3D-printed filaments with a diameter of 1.75 mm ± 0.08 mm.

2.3. Printing of PVDF/PLA/BTO specimens, lattice structures and interbody fusion cages

The obtained composite filaments were fabricated by using a FDM 3D printer. The 3D printing processing parameters were adopted as follows: the printing temperature at 215 °C, bed temperature at 65 °C, printing speed of 15 mm/s, and the layer height of 0.2 mm.

The 3D-printed specimen for piezoelectric test was in a dimension of 20 mm × 20 mm × 4 mm. The porous sample with triangular infill pattern was constructed and sliced by CURA software. Thirdly, the interbody fusion cage was designed according to the geometry in Fig. 1.

2.4. Characterization

Scanning electron microscopy (SEM). The distribution of BTO, the surface and the cross-sectional microstructure of 3D-printed composites were observed by SEM (S-4800, Hitachi, Tokyo, Japan). The operating voltage was 3 kV, and the sample was sputtered with a thin gold layer under vacuum before observation. Energy dispersive spectrometer (EDS) equipped on SEM was used to determine the element distribution on the fabricated materials.

Differential scanning calorimetry (DSC). The melting point temperature (T_m), the endothermic melting enthalpy (${\Delta H}_m$) and the degree of crystallinity (%χ_c) of the samples were examined by DSC (TA Instrument, Q25, USA). The testing sample weighed about 10 mg was firstly heated from 0 to 200 °C to remove the thermal history, cooled to room temperature and re-heated to 200 °C for the second heating. Both the heating and cooling rate was 10 °C/min and conducted under a nitrogen atmosphere.

The specific formula for the crystallinity χ_c% was as follows [31]:

$${{\chi }_c\%}_c=\frac{100\%\times {\Delta H}_m}{{\omega \Delta H}_m^{°}}$$ (1)

where ${\Delta H}_m$ is the fusion/melting enthalpy (J/g), ω is the actual portion of polymer in the composites, and ${{\Delta H}_m}^0$ = 104.7 J is the melting enthalpy for a 100 % crystalline PVDF.

Rheological behavior. The rheological properties of the fabricated materials were carried out using a rotational rheometer (TA, AR2000ex, USA) with a parallel plate geometric diameter of 25 mm. The tested temperature was 215 °C with a frequency scan ranging from 0.1 to 100 rad/s and the selected strain value was 1 %.

Fourier transform infrared spectroscopy (FTIR). The presence of the electroactive β phase in PVDF was analyzed by FTIR (Nicolet IS50, Thermo Fisher Scientific, USA) in attenuated total reflectance (ATR) mode. The scanning range was from 1600 to 400 cm⁻¹ with a resolution of 4 cm⁻¹. The fraction of electroactive phase $\mathrm{F}\left(\mathrm{\beta }\right)$ was calculated from peaks located close to 840 cm⁻¹ (β and γ) and 763 cm⁻¹ (α) [32]:

$$\mathrm{F}\left(\mathrm{\beta }\right)=\frac{100\%\times A_{\beta }}{\left(\frac{K_{\beta }}{K_{\alpha }}\right)A_{\alpha }+A_{\beta }}$$ (2)

where $A_{\alpha }$ and $A_{\beta }$ represent the absorbance at 766 and 840 cm⁻¹respectively. $K_{\beta }$ and $K_{\alpha }$ as absorption coefficients are 6.1 × 10⁴ and 7.7 × 10⁴ cm²/mol, respectively the absorption coefficients at the.

Thermomechanical analysis (TMA). The coefficient of thermal expansion (CTE) of pure PVDF and PVDF-based materials were obtained by TMA (TA Instruments, Q400, USA). The testing temperatures was set from 25 to 125 °C with a heating rate of 2 °C/min.

Melt Flow Index (MFI). The MFI of pure PVDF and PVDF-based composites were tested using a melting finger apparatus (Chengde Jinjian Testing Instrument Co., Ltd., MFI-1221, China). The testing temperature was 230 °C with 2.16 kg.

Mechanical tests. The tensile property with dumbbell-shaped specimens were 3D-printed according to ISO 527standard. The tensile tests were tested using a universal testing machine (Wance TSE 105D, China) at a test speed of 1 mm/min. For flexural test, the specimens were prepared according to ISO 178, and the test speed was carried out at a speed of 5 mm/min. All mechanical measurements were repeated five times. The strength and modulus were calculated of the 3D-printed samples were calculated by the obtained stress-strain curve.

2.5. Piezoelectric signals

To test the piezoelectric properties, copper tape was adhered on both sides of the surface of 3D-printed PVDF-based specimens, including the rectangular splines and cubic porous structure. A cyclic loading static force of 100 N was applied by a compression machine to simulate the interbody fusion forces between spinal vertebrae. The resultant micro-deformation was captured by an electrical signal using an electrostatic motor (Keithley 6514), and the signal data were acquired via an NI PCI-6221 data card using LabVIEW software.

2.6. In vitro cell culture studies

2.6.1. Cell culture

Cell culture was performed using the mouse osteoblast MC3T3-E1(Shanghai Jinyuan Biotechnology Co., Ltd). The MC3T3-E1 was cultured using osteoblast medium and trypsin-EDTA solution (0.25 %) in a culture environment of 5 % CO₂ and 95 % humidity at 37 °C.

2.6.2. In vitro cell proliferation and live-dead staining

Four kinds of PVDF-based scaffolds (PVDF, PVDF/BTO, PVDF/PLA, PVDF/PLA/BTO) were 3D-printed into rectangular specimen with a thickness of 0.5 mm. Prior to cell seeding, the specimens were rinsed and sterilized by 70 % ethanol for 30 min and then washed repeatably with deionized water. After that, the specimens were dried under UV irradiation in 30 min for each side.

Subsequently, MC3T3-E1 murine pre-osteoblasts (4 × 10³ cells/mL) were initially pre-seeded onto the sterilized specimen in 96-well plates for 4 h to facilitate cell attachment in standard culture conditions. The in vitro cell viability was assessed for 24 h and 72 h respectively. Briefly, the culture medium was discarded and then washed for three times by PBS. Subsequently, 100 μL Cell Counting Kit 8 (CCK-8) reagent (10 % v/v in basal medium) was added to each well and incubated for additional 3 h. The optical density (OD) values at 450 nm were obtained using a microplate reader (BioTek Synergy H1, USA). The Calcein-AM/PI dual-staining solution for test was prepared according to the manufacture instructions. The cell-seeded specimens were stained in a dark and incubated for 30 min at 37 °C, and then the staining solution was discarded from the plates. Then, the cell proliferation in scaffolds were observed by inverted fluorescence microscope (Olympus IX83, Japan).

2.7. Animal experiments in vivo

2.7.1. Preparation of intervertebral disc defect model

For this study, our proposed piezoelectric fusion cage (PVDF/PLA/BTO) was designed and 3D-printed, and the 3D-printed porous PEEK cage with same infill structure was set as control. All prepared cage devices were sterilized with ethylene oxide prior to implantation.

The in vivo sheep experiment was performed under the authorization of the Institutional Animal Ethics Committee (IAEC) at Central Hospital of Dalian University of Technology, following the international standards on animal welfare (Approval YN 2023-042-03, date of approval: 2023, 22nd May). Four experimental sheep (60 ± 5.0 kg) were selected from Beijing Fulong Tengfei Experimental Animal Research Institute Co., Ltd to establish intervertebral disc defect model. Anesthesia was induced via intravenous bolus injection of propofol (5 mg/kg) administered through a scalp needle inserted into the jugular vein. Under X-ray guidance, the C2-C4 vertebral segments were localized. Followed routine disinfection and draping, a layered incision was made at the C2-C5 intervertebral spaces. Disc tissues were excised, removed and the 3D-printed fusion cages were implanted into the C2/C3 and C3/C4 intervertebral spaces, respectively. After filling the disc defects, the internal fixation was achieved using an anterior cervical 4-hole titanium plate and screw system of appropriate length to make the cages effectively fixed. The sheep was sacrificed at twelve weeks after the operation, and the specimens were fixed in a 4 % paraformaldehyde solution for subsequent experiments.

2.7.2. Clinical computed tomography (CT)

CT scanner was performed before the model animal operation to select proper cages using X-ray machine (SIEMENS SOMATOM Drive; SIEMENS Healthineers, Germany). After the intervertebral disc fusion operation, the defected and implanted area of animal model were re-scanned by CT.

2.7.3. Micro-CT

The internal structure of implanted cage and the bone ingrowth into the cage were detected by a Zeiss Xradia 610 Versa micro-CT scanner with a scanning voltage of 60 kV. Two-dimension (2D) slice images were captured and then constructed to 3D models using Dragonfly software.

2.7.4. Histological analysis

The new-bone regeneration into the cages were analyzed using hematoxylin-eosin (H&E) staining. After 12-week operation, the experimental sheep was euthanatized, and the implanted cages with surrounding tissues were fixed in a 4 % paraformaldehyde, embedded in paraffin, and made into hard tissue sections with a thickness of 30 μm without decalcification. A digital microscope was used to observe the bone regeneration, and the area of the new bone was calculated using Image J software for semi-quantitative analysis.

2.8. Statistical analysis

All data in this experiment were expressed as mean ± standard deviation for n ≥ 3. Statistical data analysis was carried out using a one-way analysis of variance with Turkey's test. p < 0.05 was considered statistically significant. Statistical differences were expressed as ∗ (p < 0.05), ∗∗(p < 0.01), and ∗∗∗(p < 0.001). All data were analyzed using SPSS analysis software.

2.9. Finite element analysis

The stress distribution of the spinal fusion device under compressive load was characterized using finite element analysis (FEA). The device model was constructed and imported into Abaqus for analysis. Rigid plates were positioned at the top and bottom of the model, and the structure was meshed with C3D8R eight-node hexahedral elements and analyzed using the explicit solver. The bottom rigid plate was fixed, while a prescribed displacement was applied to the top plate along the Z-axis. Contact between the spinal fusion device and the rigid plates was modeled using a penalty contact method with 0.2 friction coefficient and hard contact.

3. Results and discussion

3.1. Characterization of 3D printability of PVDF-based composite filaments

In this section, four composites including PVDF, PVDF/BTO, PVDF/PLA and PVDF/PLA/BTO were prepared into feedstocks for 3D printing. The microstructural observation, rheological properties, printability and mechanical performance were evaluated and compared. As shown in Fig. 2, the SEM images showed the BTO nanoparticles were well-dispersed in PVDF and PVDF/PLA matrix. Meanwhile, the element distribution including carbon (C), oxygen (O), fluorine (F), titanium (Ti) and barium (Ba) were examined. These elements distribution maps confirmed an evenly distribution for each component, which was beneficial for a stable and balanced performance [33].

Fig. 2.

The characterization of PVDF, PVDF/BTO, PVDF/PLA and PVDF/PLA/BTO samples, (a) the microstructural observation of SEM images, (b) the element distribution of C, O, F, Ti, Ba and element merge diagrams, DSC curves including (c1) the cooling and (c2) the second heating curves, and frequency dependence of (d1) dynamic modulus and (d2) complex viscosity.

Then, the non-isothermal crystalline properties of the PVDF-based composites were evaluated via DSC analysis. From Fig. 2b, the addition of PLA and BTO did not greatly change the crystalline and thermal behavior. The melting temperature T_m of PVDF, PVDF/BTO, PVDF/PLA, and PVDF/PLA/BTO were 168.9 °C, 171.0 °C, 169.3 °C and 169.1 °C, whereas the crystalline temperature (T_c) in the cooling curves displayed 136.9 °C, 137.2 °C, 136.9 °C and 136.9 °C respectively. Both T_m and T_c were assigned to PVDF phase and not greatly changed. Besides, T_c of PLA at 112.8 °C and 114.0 °C were detected in PVDF/PLA and PVDF/PLA/BTO respectively. The degree of crystallinity (χ_c%) was further calculated from the DSC curves. A slightly higher χ_c% in PVDF/BTO with 47.2 %, PVDF/PLA with 45.4 % and PVDF/PLA/BTO with 47.6 % were achieved in contrast with pure PVDF with 44.9 %. The above results indicated that the incorporated BTO dispersed in polymeric matrix had a nucleation effect on PLA and PVDF crystallization, whereas low content of PLA phase showed little effect on crystallization.

The rheology property of PVDF-based composites played a vital role in the flowability during printing extrusion. Fig. 2d1 and d2 showed the dynamic rheology results including the frequency dependency of dynamic modulus and complex viscosity. It could be seen that the incorporation of PLA exhibits a higher initial storage modulus and complex viscosity in low frequency area. This phenomenon could be explained by an enhanced chain entanglements between PVDF and PLA molecular chains, suggesting an enhancement in melt strength. Then, as the shear frequency increased, the difference became smaller, which was ascribed to the molecular chain disentanglement.

The 3D printing quality of the prepared composites played a key role in the comprehensive performance of fabricated components. As shown in Fig. 3, it could be shown that all kinds of PVDF feedstocks possessed homogeneous diameter of 1.75 ± 0.08 mm. However, there were differences in the printing quality among the four filaments. For pure PVDF, the warping was obvious and the printing stopped due to the detachment of printed resin. After adding BTO, the warping issue was slightly improved but the printing surface was still rough. When further incorporated with a content of 15 wt% PLA, the warping phenomenon of both PVDF/PLA and PVDF/PLA/BTO disappeared and the printing surface as well as the dimension stability was satisfied.

Fig. 3.

(a) The fabricated filaments, (b) 3D-printed sheets, (c) SEM images of cross-sectional samples, (d) the MFIs and (e) thermal expansion coefficients of PVDF, PVDF/BTO, PVDF/PLA and PVDF/PLA/BTO composites.

Fig. 3c showed the cross-sectional microstructures of four kinds printed samples by using SEM micrographs. PVDF exhibited a layered structure with some defects due to the layer-by-layer deposition mode and the shrinkage of PVDF crystallization. When extruded from the nozzle, the environmental temperature was declined, and then the crystallization took place during the cooling process. This was even more serious in the presence of BTO nanoparticles, which acted as a heterogeneous agent. The delamination was largely alleviated after PLA incorporating. The interfacial boundary became blurred and interfacial adhesion also became compacted. This was mainly attributed to the good printability of PLA with proper flowability and relatively low crystallization rate [34].

The reason for warping was mainly because the relaxation of the internal stress accumulated during the nozzle extrusion and the rest parts cooling, which caused anisotropic shrinkage and detached from the printing platform. So, to explain the printability between the PVDF-based composite filaments, the thermal expansion coefficient and melt flow index were investigated (Fig. 3d). The MFIs indicated that the addition of BTO might restricted the melting flowability whereas PLA increased the fluidity (Fig. 4b). The MFI of PVDF/BaTiO₃ was 23.70g/10 min, lower than that of PVDF at 28.92 g/10 min. The decreased MFI adversely affected the printing quality, which brought a rough printing surface to the 3D-printed parts. This was probably due to the existence of BTO inhibited the chain entanglement between PVDF and PLA. After adding, PVDF/PLA achieved MFI of 34.08 g/10 min and PVDF/PLA/BaTiO₃ reached the highest value of 39.24 g/10 min. Although the molecular chain entanglement might exist between PLA and PVDF, the good flowability of PLA under testing temperature contributed to an increased MFI. Accordingly, PLA well improved the processability and printability to PVDF matrix.

Fig. 4.

Tensile stress-strain diagrams of 3D-printed PVDF, PVDF/BTO, PVDF/PLA and PVDF/PLA/BTO tensile specimens.

On the other hand, the thermal expansion coefficient would influence the dimensional stability of the 3D-printed parts, particularly during the cooling process. Since the differences in thermal expansion property between polymers and ceramics, the different thermal expansion behavior might bring additional stress and deformation [35]. The variation curves in coefficient as a function of temperature were achieved (Fig. 3e). It could be seen that PVDF/BTO exhibited higher thermal expansion coefficient and followed by pure PVDF and then PVDF/PLA/BTO, which was more serious in high-temperature region. This might explain the reason for the warping and shrinkage for PVDF/BTO and PVDF. In contrast, PVDF/PLA blends showed the lowest expansion coefficients, which contributed to the dimensional stability during the 3D-printing process.

The mechanical performance of 3D-printed PVDF-based tensile specimens was shown in Fig. 4. It was shown that the tensile strength and Young's modulus (the slope of stress-strain curve) were enhanced by the addition of PLA due to the high stiffness and modulus of PLA. For pure PVDF sample, the tensile strength was 43.34 MPa. Among the four kinds of materials, the highest tensile strength was as high as 51.46 ± 2.4 MPa in PVDF/PLA/BTO samples, suggesting a reinforcing effect of PLA. But the presence of PLA decreased the elongation at break of the PVDF-based composites, which decreased from initial 14 % to about 7 %. As presented in Fig. 3b, the printing quality for PVDF and PVDF/PLA were not satisfied due to the warping issue, which deteriorated the mechanical performance. On the other hand, PVDF/PLA and PVDF/PLA/BTO presented good molding quality, and thereby exhibited improved tensile property. The small content addition of BTO did not contribute to an enhancement in modulus and strength. The superior mechanical property of PVDF/PLA/BTO could meet the biomechanic property of interbody fusion devices.

3.2. Piezoelectric effect of 3D-printed PVDF-based composites

In this section, the piezoelectric effect of PVDF-based composites was evaluated by FTIR and piezoelectric test. The piezoelectric properties of PVDF are primarily determined by the percentage of the β-phase, which could be obtained from FTIR curves and calculate according to Eq. (2). As marked in Fig. 5, the band located at 764 cm⁻¹ was corresponded to the formation of the α phase, whereas the presence of β phase can be identified by the band located at 842 cm⁻¹and 1279 cm⁻¹. For pure PVDF, the characteristic peaks at 764 cm⁻¹ were obvious, suggesting the melt blending mainly induced the formation of α-phase. After the addition of BTO and PLA separately, the peak intensity at 764 cm⁻¹ assigned to α-phase became weaken whereas the peak intensity at 842 cm⁻¹ attributed to β-phase became intensified [[36], [37]]. In Fig. 5b, compared to 50.5 % β-phase in pure PVDF, the content of the β-phase in PVDF/PLA/BTO was as high as 68.9 %, which would contribute to a piezoelectric performance. Moreover, the enhancement effect of PLA was superior to BTO nanoparticles. This might be due to the interaction between the -CF₂- groups of PVDF and the -OH or -CH₃ groups of PLA [34]. These interactions help to convert the conformation of α-phase PVDF to conformation at the interface, thereby enhancing the β-phase structure in PVDF-PLA binary polymer blends.

Fig. 5.

(a) FTIR spectra, (b) the calculated β-phase of PVDF, (c) the open-circuit voltage waveform diagrams and (d) piezoelectric signals bar charts for 3D-printed PVDF, PVDF/BTO, PVDF/PLA and PVDF/PLA/BTO.

The open-circuit voltage of a piezoelectric device was closely related to the dielectric constant of the piezoelectric material. The sample with 20 mm × 20 mm × 4 mm was fabricated via 3D printing. With regards to the working condition of intervertebral disc, it subjected to axial quasi-static pressure. Therefore, the piezoelectric test condition was set in quasi-static axial compression stress with a 100 N force. The piezoelectric voltages of PVDF, PVDF/BTO, PVDF/PLA, and PVDF/PLA/BTO were 25 mV, 31 mV, 42 mV, and 63 mV, respectively (Fig. 5d). Considering the negative impact of printing quality on piezoelectric test, the piezoelectric signal gap was widened. Consequently, the piezoelectric effect of PVDF/PLA/BTO was superior to pure PVDF, achieving significant enhancement in piezoelectric performance.

3.3. Compression and piezoelectric effect of 3D printed PVDF porous parts

In this section, triangular porous architecture was selected for infill pattern of the cage, and the effect of infill densities (40 %, 50 %, 60 %, and 70 %) on the piezoelectric property and mechanical behavior were investigated. The designed models and the corresponded 3D-printed samples were illustrated in Fig. 6. For the piezoelectric signals, standard samples achieved 465 mV, 421 mV, 397 mV and 248 mV for infill density of 40 %, 50 %, 60 % and 70 % respectively. These open-circuit voltage could stimulate calcification and accelerated the bone defect repair [38]. In contrast with solid cubic samples in Fig. 5c, it was found that the piezoelectric signals were greatly improved, and the signal intensity was negatively correlated with the infill density. In detail, the piezoelectric signal with 40 % infill density was 1.08 times higher than that with 70 % counterpart. This phenomenon was primarily determined by the sample deformation during the compression process. Samples with lower infill density corresponding to large porosity suffered larger deformation, which produced higher piezoelectric signal. Therefore, it is of great advantage for fabricating 3D-printed interbody cage with excellent piezoelectric effect and high porosity.

Fig. 6.

a) Schematic diagram of different fillings, b) piezoelectric voltage signals, and c) statistical bar charts for different fillings.

The compressive properties of 3D-printed porous PVDF/PLA/BTO samples with triangular infill pattern were further evaluated. As shown in Fig. 7, the standard specimens with a dimension of 20 mm × 20 mm × 20 mm were subjected to a compression force and the deformation process were recorded. It could be observed that the all the samples with different infill density were successfully designed, sliced and 3D-printed. When suffered compression, all samples kept intact shape within compression strain below 0.1. However, no structure disintegrated was observed for all samples, suggesting the implantation safety as biomedical implants. With the increase of compressive strain, the samples became squeezed and deformed during the compression. The maximum compressive strengths were 19.24 ± 0.15 MPa, 27.63 ± 1.85 MPa, 34.55 ± 3.14 MPa, 51.86 ± 2.17 MPa for samples with 40 %, 50 %, 60 % and 70 % infill density respectively. In previous publish work, Jia proposed 3D-printed porous cage in similar geometry using PEEK/tantalum (Ta) composites, and achieved compressive strength and modulus with ∼50 MPa and ∼100 MPa respectively [13]. The compressive performance of PVDF/PLA/BTO cage developed in our work was close to those of PEEK counterparts, suggesting the feasibility of PVDF/PLA/BTO as interbody cage. For another commonly-used cage material, the maximum compressive strength exceeded 40 MPa, and the average modulus ranged from 1 to 3.5 GPa in Ti6Al4V interbody fusion cage [39]. In comparison with PEEK and PVDF/PLA/BTO (this work), the compressive strength and modulus of porous Ti6Al4V were higher, which might indicate a potential stress-shielding issue. Although there are curtained difference in compressive parameters among three different materials, these data were in the range of compressive modulus (5–170 MPa) and E-modulus (0.5–20 GPa) of cancellous and cortical bone.

Fig. 7.

PVDF, PVDF/BTO, PVDF/PLA, PVDF/PLA/BTO porous samples (a) infill pattern and 3D-printed samples, (b) compressive process under different strains, (c) compressive stress-strain curves, and (d) compressive strength and modulus values.

Consequently, the porous sample with 60 % infill density achieve balanced performance of piezoelectric effect and sufficient load-bearing capacity. However, the porosity (which was opposite with infill density) has not been clearly defined, it is still a challenge topic for the cage porosity. Based on the above experimental results, an infill density of 60 % was selected for the fabrication of interbody fusion cage for the subsequent animal experiments.

3.4. Biological performance of 3D printed PVDF-based composites

It is crucial to verify the in vitro biocompatibility and in vivo osseointegration of piezoelectric PVDF/PLA/BTO as interbody cage. In this section, MC3T3 E1 was selected to evaluate the cell cytotoxicity of PVDF-based slices with a thickness of 0.5 mm. Fig. 8 showed the cytoskeletal morphology of cells on the cages. After culturing 24 h and 72 h, no significant cell viability was detected among the four kinds of PVDF-based materials. It was found that all samples achieved over 80 % cell viability, suggesting that the incorporation of BTO and PLA have no negative impact on the biocompatibility. Furthermore, fluorescence staining was used to evaluate the cell proliferation. As shown in Fig. 8b, cells were robustly proliferated after culturing 24 h and 72 h under the fluorescent observation. Therefore, the 3D-printed PVDF-based samples exhibited good biocompatibility, ensuring sustained cell health and proliferation.

Fig. 8.

PVDF, PVDF/BTO, PVDF/PLA, PVDF/PLA/BTO in MC3T3-E1 (a) CCK-8 and (b) live death staining assay results for 24 h and 72 h.

Furthermore, the interbody cage was designed, and the piezoelectric PVDF/PLA/BTO cage was then 3D-printed as illustrated in Fig. 9. Firstly, the geometric dimension of PVDF/PLA/BTO cage with a 60 % infill density was presented in Fig. 9 a. Then, the porous cage was successfully 3D-printed and the implanted cages was further sterilized by propylene oxide, which prevented potential implantation infections (Fig. 9b). The surface and internal microstructure was detected by SEM and micro-CT (Fig. 9c and d). The continuity and deposition of printing layers were clearly shown, and no obvious defects were detected, which suggested a high printing quality and feasibility of porous cages. To confirm the rationality of structural design, a finite element analysis (FEA) was utilized to simulate the stress concentration. As presented in Fig. 9e, the outer thick shell subjected to the main compression load, while the central filling part also carried some external force. No serious concentration situation was detected for the current proposed infill pattern. Consequently, internal microstructure is beneficial for load-bearing, providing an auxiliary support effect as interbody cages.

Fig. 9.

(a) The design geometry of the cage, (b) the 3D-printed porous cage sample and the sterilized implanted samples, (c) the micro-CT image of internal structure, (d) the surface morphology and (e) the FEA simulation of the stress state of proposed cage structure.

Afterwards, the porous PVDF/PLA/BTO cage was implanted in sheep animal model, and PEEK cage with the same structure was 3D-printed and set as control group. In Fig. 10, it was cleared that both PEEK and PVDF/PLA/BTO cages were well-fixed between two adjacent vertebral body. In the sliced CT scanning images, it was clearly observed the PVDF/PLA/BTO porous cage was perfected inserted between the vertebrae spines, which was expected to be well-carried the external compression force. By CT scanning, due to the radiopacity of BTO, it could be easily observed the PVDF/PLA/BTO porous cage in the transaxial and sagiital view whereas the PEEK counterpart was not visible.

Fig. 10.

(a) The real-time CT imaging of the C2-C5 intervertebral implanted images in transaxial and sagittal views, and (b) the intervertebral disc defects created and the implanted cages.

After three-month and five-month implantation, the sheep models were CT-scanned for location and then sacrificed, and the part of the cervical vertebrae containing the implanted PVDF/PLA/BTO and PEEK cages was cut and staining for further evaluation. As shown in Fig. 11, tissue sections revealed that more newly-formed bone was detected in the PVDF/PLA/BTO cage (Fig. 11b) than that in PEEK counterpart (Fig. 11d) after three-month implantation, exhibiting a positive effect by piezoelectric materials. Moreover, a closer contact with surrounding newly-formed tissues was observed, which suggested bone integration around the cage [40] (Fig. 11b and d). After five-month implantation, it could be seen that more fibrous tissue was grown into the cages. From the side section, the fibrous tissues have almost filled in the PVDF/PLA/BTO cages, and the direction of cell proliferation along the cage pores was also obvious, suggesting the piezoelectric signals had evident osseointegration effect (Fig. 11c and e). Meanwhile, micro-bubble pores were seen in the PVDF/PLA/BTO cages after five months, which was probably due to the biodegradation of PLA phase. In contrast, PEEK cages did not show a close-contact affinity to bone tissues (Fig. 11e). From the side-view tissue section, the differences in bone ingrowth into the two kinds of cages were distinguished. Although the newly-formed bone in PEEK cages after 5 months implantation was more than that in 3 months, more newly-formed bone and better spinal fusion were found into PVDF/PLA/BTO cages. The neostatic bone tissue surface area in PVDF/PLA/BTO and PEEK cages were conducted in semi-quantitative analysis (Fig. 11f). After 3-moth implantation, the bone surface area in PVDF/PLA/BTO covered 73 % whereas that of PEEK cage only reached 42 %. After 5 months, the neostatic bone tissue surface area was above 90 % in PVDF/PLA/BTO cages, demonstrating excellent bone integration and healing rate compared to PEEK control. Also, it was expectable that the osseointegration and new ingrowth bone would carry more external loading and act as natural piezoelectric system instead of the implanted cage.

Fig. 11.

(a) Cervical spine of sheep with cage embedded, the hard tissue HE staining of (b) PVDF/PLA/BTO cage after (b) three months and (c) five months, the contrast group: implanted PEEK cage after (d) three months and (e) five months. (B: new bone tissue, C: cage, and F: fibrous tissue), and (f) the quantitative comparison analysis between cages with three-month and five-month implantation period.

Furthermore, the embedded cage in the bone tissues were detected and reconstructed to 3D models by using micro-CT scanning and software processing. The cages and ingrowth bone were separated and assigned to yellow and dark blue color respectively. As shown in Fig. 12, it was clearly showed that the amount of new-generated bone in PVDF/PLA/BTO cages was obviously more than that in PEEK cage. In detail, the proportion of bone volume achieved 49.3 % in porous PEEK at targeted height, while the bone volume reached weigh higher of 76.8 % in piezoelectric cages. This result confirmed the superior osteoconductive properties of PVDF/PLA/BTO cages in hard-tissue HE staining results. Since PVDF materials was unable to directly promote bone healing, the osseointegration effect with enhanced bone ingrowth capacity was attributed to the electric charge on the contact surface from the cyclic loading at the intervertebral disc position. Then, the electrical stimulation activated the voltage-gated calcium channels, increased intracellular calcium levels, and thus trigger the signal pathways for osteogenic differentiation and mineralization [41]. Reported in the literature used as bone implanted materials, it could also stimulate the secretion of the growth factors like BMPs and TGF-β, which was crucial for the differentiating osteoblast cells. This promotion effect has been demonstrated by previous PVDF-based bone tissue scaffolds, which shed a light on the fabrication of piezoelectric porous cages [42,43]. In addition, PVDF/BaTiO₃ were demonstrated with long-term mechanical stability and beneficial biomineralization effect, which also show great advantages as biomedical implants [44,45]. Therefore, the current work proposes a new insight to fabricate biomimetic cage with porous structure and piezoelectric stimulation for bone integration.

Fig. 12.

The 3D micro-CT reconstruction images of printed PVDF/PLA/BTO and PEEK cages.

4. Conclusion

In the study, a novel piezoelectric porous cage with excellent osteointegration and mechanical performance were proposed and 3D-printed by composite PVDF/PLA/BTO filaments. By incorporating BTO and PLA, the printability was improved and the transformation of β-phase was stimulated. The enhanced piezoelectric effect and mechanical properties showed PVDF/PLA/BTO great potentials as interbody cage materials. With a triangular infill pattern, the relationship between the infill density with piezoelectric effect as well as compression performance were investigated. The results showed that 60 % infill density achieved a piezoelectric output voltage of 465 mV with a compressive stress of 34.5 MPa. The suitable electric signals and load-bearing capacity would be effective triggered osseointegration and a good supporting effect. Moreover, PVDF/PLA/BTO cage with triangular infill pattern exhibited good in vitro cell biocompatibility and accelerated inner osseointegration in sheep cervical fusion model in contrast with PEEK control. In brief, the proposed 3D-printed porous PVDF/PLA/BTO with piezoelectric effect has great potential as interbody fusion cages. The 3D-printing technique facilitates custom-design for geometry and infill pattern, which realize a bioinspired structure and mechanical modulating. The utilization of piezoelectric PVDF-based materials endows an enhanced osseointegration in spine fusion. In addition, this study presents a practical application for the design and manufacturing of piezoelectric PVDF-based composites through 3D printed biomimetic structures.

CRediT authorship contribution statement

Hao Zhang: Supervision, Resources, Project administration, Funding acquisition, Conceptualization. Guoqing Gao: Methodology, Investigation, Formal analysis, Data curation. Lihan Wang: Methodology, Data curation. Chunlu Li: Methodology, Investigation. Chang Xu: Investigation. Xufeng Dong: Supervision. Lin Sang: Writing – original draft, Validation, Supervision, Data curation.

Funding

This work was financially supported by Dalian Science and Technology Talent Innovation Fund (No. 2023RY022).

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.

Data availability

Data will be made available on request.