Injectable Nanocomposite Biomaterial for 3D Printing of Personalized Matrices and Their Use in Bioreactors for Bioengineering Advanced Cell Culture Models

For the bHA synthesis, a neutralization reaction was exploited. Three solutions were prepared: an acid solution containing 13.74 g of H₃PO₄ (Sigma-Aldrich, 85% pure) in 0.1 L of bidistilled water; a basic suspension containing 15.49 g of Ca­(OH)₂ (Sigma-Aldrich, 95% pure) and 2.02 g of MgCl₂·6H₂O (Sigma-Aldrich) in 0.16 L of bidistilled water; and a polymeric solution containing 5 g of gelatin (Gel) powder (Italgelatin, Italy) in 0.1 L of bidistilled water prepared under stirring conditions at 45 °C. The H₃PO₄ solution was mixed with the gelatin solution until perfect blending and then slowly added to the basic suspension at 25 °C under constant and vigorous stirring. The mixture was stirred for 1 h, followed by a 2 h settling period at 25 °C. After centrifugation at 6000 rpm for 10 min at 25 °C, the pellet was collected, washed three times with bidistilled water by centrifugation, freeze-dried, sieved at 150 μm, and micronized at 3 μm. The morphology of bHA was evaluated with electron scanning microscopy (SEM, Carl Zeiss Sigma NTS Gmbh Öberkochen, Germany) after Pt coating (QT150T, Quorum Technologies Ltd., UK). Inductively coupled plasma–optical emission spectrometry (ICP–OES 5100, vertical dual view apparatus, Agilent Technologies, Santa Clara, CA, USA) and attenuated total reflection–Fourier transform infrared spectroscopy (ATR–FTIR, Thermo Fisher Scientific Inc., Waltham, USA) were used to determine the chemical characteristics and the stoichiometric composition of bHA. In detail, ICP allows for the quantification of the Mg²⁺, Ca²⁺, and PO₄^3– ions, which constitute the inorganic mineral component; 10 mg of hybrid GelMgHA particles or 20 mg of scaffold was dissolved in 50 mL of a 2 wt % % HNO₃ solution prior to the analysis. The crystallographic identity and crystallinity degree of bHA were evaluated with X-ray diffraction (XRD, Panalytical X'Pert PRO, Bruker, Germany). The analysis employed Cu Kα radiation (λ = 1.54178 Å) at 40 kV and 40 mA. Spectra were recorded in the 2θ range from 20° to 80 °C, with a step size (2θ) of 0.02° and a counting time of 0.5 s. Thermogravimetric analysis (TGA, STA 449/C Jupiter, Netzsch, Germany) was used to evaluate bHA weight losses under thermal treatment. The analysis was conducted in alumina crucibles, from room temperature to 1100 °C, at a heating rate of 10 °C/min under a nitrogen flow. The sample weighed approximately 10 mg.

bHAGel injectable formulations were prepared by varying bHA loading content, achieving 1:1, 0.5:1, and 0:1 w/w ratios of bHA with respect to Gel and maintaining constant a 5% w/V of Gel concentration (Table). Specifically, bHAGel_0 refers to a formulation without the mineral phase. Instead, to prepare 10 mL of composite hydrogel, 0.5 or 0.25 g of bHA was tip-sonicated with Na₃C₆H₅O₇ (Sigma-Aldrich) in 10 mL of bidistilled water for 10 min in an ice bath at 20% amplitude. Then, 0.5 g of Gel powder was added, and the mixture was kept at 45 °C under magnetic stirring for 1 h. In summary, bHAGel_0.5 has a bHA:Gel ratio of 0.5:1 and bHAGel_1 has a bHA:Gel ratio of 1:1.

The obtained hydrogel was loaded in 3 mL printer cartridges (CellInk, Göteborg, Sweden), and it was kept overnight at 4 °C and then thermally conditioned at 27 °C, the printing temperature, for 1 h, before printing.

The rheological properties were investigated by comparing the behavior of the three bHAGel formulations with different bHA loadings, measuring their viscosity against a shear rate ramp in controlled stress mode and against temperature, as well as their printability.

All of the measurements were performed with a rotational rheometer (C-VOR 120, Bohlin Instruments, UK). The shear ramp test was performed using a plate/plate PP20 (Ø = 20 mm) geometry, increasing the shear stress from 1 Pa up to 10,000 Pa with a sweep time of 720 s. A solvent trap was used to prevent water evaporation during the test. A temperature ramp test was performed using a plate/plate PP20 (Ø = 20 mm) geometry, increasing the temperature from 20 up to 35 °C, with a rate of 2 °C/min. Before the test was started, the samples were left to rest for 15 min to reduce any possible influence on the measurement because of the solution handling.

The printability of the biomaterial inks was assessed through three main tests: filament drop, filament spreading, and buildability (E (BIO X, CellInk, Göteborg, Sweden). The printability tests were executed with a Ø_in of 0.25 metal needle. For the filament drop test, after thermal conditioning, the filament extrusion profile of each biomaterial ink, extruded at the same pressure, was qualitatively evaluated. For the filament spreading test, 2 cm × 2 cm and 2 layers in height scaffolds were printed. High-resolution pictures of the printed structures were taken by digital optical microscopy (Hirox RH-2000, 3D Digital Microscope, Hirox Europe) and their model fidelity was determined using ImageJ software. Buildability of each ink was assessed by printing two different 3D structures: a 10 mm × 20 mm hollow cylinder and a 10 mm × 20 mm hollow cone.

The 3D scaffolds were printed using a bioplotter (BIO X, CellInk, Göteborg, Sweden), with design and slicing software. The design consisted of a 2 cm × 2 cm square base and a 90° layer shift, resulting in a perpendicular mesh with 20% infill and a total height of 12 layers. Printing was performed with a Ø_in = 0.25 mm conical nozzle at an average pressure of 40 kPa and 5 mm/s speed. A thermocontrolled printhead (27 °C) and a cooled printbed (4 °C) were used. After printing, the obtained structures were freeze-dried (Freeze-drier, 5 Pascal Lio 3000 PLT, Rozzano, Milano, Italy) with a controlled freezing ramp of −50 °C/h until −40 °C, followed by a controlled heating of 5 °C/h from 40 °C to −5 °C and 3 °C/h until 20 °C, lasting approximately 3 days under a vacuum, P = 0.086 mbar. Finally, they were cross-linked via dehydrothermal treatment (DHT) at 160 °C for 72 h under vacuum 0.01 mbar (Vacuum Oven Heated Shelf 50 Lt, 5 Pascal Srl). Cylindrical scaffolds (8 mm in diameter) were punched out to fit the bioreactor chamber. Finally, the scaffolds were sterilized by 25 kGy of γ-ray irradiation.

The scaffold morphology was evaluated by environmental scanning electron microscopy (SEM TM Quanta 200, FEI, Thermo Fisher Scientific Inc.), set in High-Vacuum (P <10^–4 Torr) mode. The samples were fixed on aluminum stubs using carbon tape, and they were coated with Au using a Polaron Sputter Coater E5100 (Polaron Equipment, Watford, Hertfordshire, United Kingdom). The macropore and micropore dimensions of the printed scaffolds were quantitatively evaluated from SEM micrographs using ImageJ (NIH, USA). To characterize the inclusion of bHA particles within the polymeric matrix and its interaction with it, XRD, ATR–FTIR, TGA, and ICP analyses were performed. The swelling behavior was assessed by soaking samples in 1× PBS solution with 0.1% w/v of NaN₃ at 37 °C under shaking. Samples were weighed at different time points (0 h; 0.5 h; 3 h; 6 h; 24 h; 48 h) after 5 s rest on a nonabsorbent surface. The swelling ratio (S_r) was calculated asSr=Ws−WWwhere W_s is the swollen sample weight and W the dry sample weight before soaking.

Degradation analysis was performed in the incubator exposing the samples to a 1× PBS solution with 0.1% w/v NaN₃ at 37 °C at different time points (3, 7, 14, 21, 28 days), washing them three times with bidistilled water before freeze-drying and weighing them.

The degradation ratio (D_%) was calculated asD%=W−WdWd×100where W is the dried sample initial weight and W_d is the degraded sample weight at a specific time point.

Dynamic mechanical analysis (DMA) was performed with wet samples submersed in PBS and at 37 °C to better replicate physiological conditions. Compression test was performed at 37 °C by using a Q800 DMA (TA Instruments, USA). The samples, with dimensions of 8 × 3 mm (Ø × h), were incubated in PBS overnight prior to testing and were preloaded to 0.01 N to ensure full contact between the scaffold surfaces and the compression plates. Then, compression testing was performed compressed in force control using a force ramp rate of 0.5 N/min to the upper force limit of 5 N. The compressive moduli were calculated as the slope of the initial linear part of the stress–strain curve up to 15% strain.

Bone marrow (BM) aspirates (Table) were obtained during routine orthopedic surgical procedures involving exposure of the iliac crest, after informed consent from the patient and following protocol approval by the local ethical committee (Ethical Komission Beider Basel #78/07). BM mesenchymal stromal cells (BM-MSCs), selected based on adhesion and proliferation on the plastic substrate as previously described, were used after two passages of expansion.

Freshly isolated nucleated cells were plated at a density of 1*10⁵ cells/cm². Cell expansion was carried out in complete medium (CM), which consisted of α-modified Eagle's medium (α-MEM), 10% fetal bovine serum, 1× GlutMAX, 100 mM HEPES buffer solution, 1 mM sodium pyruvate, 100 U/ml penicillin, 100 mg/mL streptomycin (Thermofisher Gibco), and supplemented with 5 ng/mL fibroblast growth factor-2 (FGF-2; R&D Systems). The medium was changed twice a week. At confluence, human bone marrow mesenchymal cells (hBM-MSCs) were replated for expansion (seeding density of 3–5*10³ cells/cm²). Upon confluence, cells at passage 2 or 3 were enzymatically retrieved and counted for use in the 3D culture experiments.

Dynamic 3D cultures were established using the perfusion-based U-CUP bioreactor system (CELLEC Biotek AG), previously validated for uniform cell seeding and long-term culture within 3D scaffolds. A schematic overview of the experimental design and workflow, including scaffold installation, cell seeding, perfusion culture phases, and end point analyses, is provided in Figure.

In the first set of experiments, 1 × 10⁶ human BM-MSCs were suspended in 8 mL of CM supplemented with 5 ng/mL FGF2, 10 nM dexamethasone, and 0.1 mM l-ascorbic acid-2-phosphate (DAF medium), a preconditioning for attachment and proliferation.− The suspension was perfused through the bHAGel scaffold (Ø 8 mm × 4 mm) at a superficial velocity of 1 mm/s (corresponding to 3.00 mL/min). Following 5 days of dynamic culture, the medium was replaced with osteoinductive medium (OM), consisting of CM supplemented with 10 nM dexamethasone, 0.1 mM l-ascorbic acid-2-phosphate, and 10 mM β-glycerophosphate, and the culture was maintained for an additional 3 additional weeks.

To investigate osteoclast differentiation, 2 × 10⁶ freshly isolated CD14⁺ monocytes from buffy coats obtained from healthy donor's peripheral blood were seeded on osteo-induced MSC scaffolds and cocultured for 10 days under perfusion in CM supplemented with 10 nM vitamin D3 and 1 μM Prostaglandin E2 (PGE2). The medium was renewed every 2–3 days.

In parallel, static cocultures were performed to compare scaffold performance and cellular interactions. Briefly, 60 μL of a suspension containing 1 × 10⁵ MSCs was pipetted directly onto the bHAGel scaffolds and incubated for 40 min at 37 °C in low-attachment 12-well plates (Sarstedt, #4021721) to allow cell attachment. Scaffolds were then submerged in DAF medium and maintained for 5 days. Subsequently, the medium was switched to OM and the culture continued for 3 weeks. CD14⁺ cells (1 × 10⁶) were then seeded onto the constructs and allowed to attach for 40 min at 37 °C. Cocultures were maintained for 10 additional days in CM supplemented with 25 ng/mL M-CSF and either 10 nM vitamin D3 (osteotropic condition) or 50 ng/mL RANKL (osteoclastogenic condition). Experiments were conducted with BM-MSCs from three donors (BM187, BM231, BM272). Dynamic osteogenic cultures included n = 2 scaffolds per donor at day 0 and n = 3 at day 21; acellular scaffolds (n = 3) served as material controls. Each scaffold was cultured in an independent bioreactor unit. Osteoblast–osteoclast cocultures (donor BM272) were performed with n = 2–3 independent bioreactor replicates.

Engineered tissue constructs were fixed overnight at 4 °C in 0.1 M sodium cacodylate buffer containing 2% glutaraldehyde and then dehydrated through a graded ethanol series (30% to 100%, 15 min each). Samples were dried using critical point drying with liquid CO₂ (Autosamdri-815, Tousimis), mounted on aluminum stubs, and sputter-coated with 20 nm gold (LEICA EM ACE600). SEM imaging was performed on a ZEISS Gemini 2 microscope at 5 kV and 200 pA, with magnifications from 500 to 5000×. Surface microstructure was qualitatively evaluated on longitudinal and transverse sections.

Paraffin-embedded scaffold samples were sectioned at 7 μm thickness using a microtome (Microm HM 355S, Thermo Scientific). Sections were mounted on Superfrost Plus glass slides (Thermo Scientific), deparaffinized with Ultraclear (Bio-Optica), and rehydrated through a graded ethanol series (100%, 96%, 70%, and 50%; two washes per step, 10 min each) into distilled water.

Hematoxylin and eosin (H&E) staining was performed by immersing sections in Harris hematoxylin for 7 min, rinsing under running tap water, and counterstaining with eosin for 2 min. Slides were subsequently dehydrated, cleared, and mounted with coverslips.

TRAP staining was performed on separate section sets to detect osteoclast activity. Sections were incubated in 0.2 M acetate buffer containing TRAP5b staining solution (1 mg/mL Naphthol AS-MX Phosphate Disodium Salt and 1 mg/mL Fast Red Violet LB Salt; Sigma-Aldrich) at 37 °C for 1–4 h. After washing in distilled water, sections were counterstained with Mayer's hematoxylin for 1 min, rinsed in running water, air-dried, mounted with Pertex (Histolab), and imaged using a fluorescence microscope (Nikon TI2).

Immunohistochemistry (IHC) was conducted using a Ventana Discovery Ultra platform (Roche Diagnostics) with primary antibodies against bone sialoprotein (BSP; Abcam ab52128, 1:100) and Osteopontin (OPN; Proteintech 22952-1, 1:100). Antigen retrieval and detection followed the manufacturer's automated protocol. Counterstaining was performed with hematoxylin.

Paraffin-embedded scaffold sections were deparaffinized in xylene and rehydrated through a graded ethanol series (100%, 96%, 70%, and 50%; two washes per step, 10 min each). Antigen retrieval was performed by incubating slides at 95 °C for 30 min in citrate buffer (pH 6.00, ProTaqs, no. 400300692), followed by cooling at RT for 20 min. To permeabilize the tissue, sections were washed twice for 10 min with 1% goat serum (GS) in PBS containing 0.4% Triton X-100 (PBS-T). Nonspecific antibody binding was blocked with 5% GS in PBS-T (blocking solution, BS) for 30 min at RT.

Sections were incubated overnight at 4 °C with a rabbit antihuman PTX3 polyclonal antibody (in-house purified) diluted 1:200 in PBS-T containing 1% GS. The next day, sections were washed twice in the same buffer (10 min each), then incubated for 1 h at RT with an Alexa Flour 647-conjugated goat antirabbit IgG (H + L) polyclonal secondary antibody (Invitrogen, #A21245) diluted in PBS-T containing 1% GS. After two additional 10 min washes, nuclear counterstaining was performed using DAPI (1:1000 in PBS; BD Biosciences, #564907) for 5 min. Slides were rinsed in PBS, mounted, and imaged using a fluorescence microscope (Nikon TI2).

For the whole-mount immunostaining, selected 3D scaffold constructs were harvested at the end of the coculture period, cut in two halves, and fixed overnight in 4% paraformaldehyde at 4 °C. After washing in PBS, scaffolds were permeabilized with PBS-T for 10 min and blocked for 1 h at 4 °C in BS. Primary antibodies, rabbit anti-osteocalcin polyclonal (Abcam, #ab93876) or rabbit recombinant anti-TRAP monoclonal (Abcam, #ab240970), were diluted 1:200 in BS and incubated overnight at 4 °C. Samples were washed twice with PBS, then incubated for 1–4 h at RT with an Alexa Flour 647-conjugated goat antirabbit polyclonal secondary antibody (Abcam, #a21247) diluted 1:1000 in BS. After two PBS washes, nuclei were counterstained with DAPI (1:1000 dilution in PBS, 10 min, RT). Scaffolds were loaded onto multiwell imaging slides (ibidi, #80821) and imaged using a Nikon AXR confocal microscope.

Quantification of PTX3 and OPG in OB/OC coculture supernatants was performed using commercial ELISA kits, following the manufacturers' protocols (PTX3: Hycult Biotech; OPG: R&D Systems). Absorbance was measured using a VERSAmax Tunable Microplate Reader, and data were analyzed with SoftMax Pro 5.3 software (Molecular Devices, LLC.).

The concentrations of human MMP9 and BMP9 in the OB/OC coculture were assessed using the ELLA Automated Immunoassay System (Bio-Techne), in accordance with the manufacturer's instructions. The reported limits of detection (LODs) were 0.156 ng/mL for MMP9 and 0.086 pg/mL for BMP9.

Total RNA was extracted using the Trizol reagent (Invitrogen, Thermo Fisher) following mechanical homogenization with stainless steel beads in a TissueLyser II system (Qiagen). RNA concentration and purity were assessed via NanoDrop One spectrophotometry (Thermo Scientific). One microgram of RNA was reverse-transcribed using M-MLV reverse transcriptase (Invitrogen) according to the manufacturer's protocol. Quantitative real-time PCR (qRT-PCR) was performed using TaqMan Gene Expression Assays on a ViiA 7 Real-Time PCR System (Applied Biosystems). Target genes included Runx2, Osteocalcin (Bglap/Ocn), Bone Sialoprotein (Ibsp), Collagen type I (Col1a1), Bone Morphogenic Protein 2 (Bmp2), Tartrate Resistant Acid Phosphatase (Trap), Cathepsin K (Ctsk), Osteopontin (OPN), and Pentraxin-3 (Ptx3). GAPDH was used as the endogenous control, and the relative gene expression was calculated using the 2^–ΔCt method.

The naturally occurring biomineralization process was translated into an in-lab process to create a bioinspired and highly biomimetic hydroxyapatite (bHA). In detail, Mg²⁺-doped HA was nucleated and grown directly onto the Gel macromolecules. The in-lab biomineralization process was achieved by performing a neutralization reaction in an intentionally uncontrolled environment (air) at a low temperature (Troom), physiological pH conditions, in the presence of Mg²⁺ as doping ion and Gel as the organic template.

An open-air reaction environment favored spontaneous doping with CO₃^2– ions derived from CO₂, thus reproducing a physiological aspect of natural bone deposition. Mg²⁺ doping, typical of newly formed bone, promoted acceleration of HA nucleation kinetics, consequently reducing its crystallinity. The acidified gel provided nucleation centers for biomineralization, working in conjunction with the low synthesis temperature and doping ions to hinder crystallization, thereby enhancing the final biomimicry of the obtained bHA.

Conceptually inspired by the in-lab collagen biomineralization process reported by Tampieri et al. in 2008, this method instead utilizes Gel as a template, enabling the formation of biomineralized microparticles rather than a three-dimensional structure. These hybrid microparticles are well suited as mineral additives in ink formulations for 3D printing and bioprinting. The presence of Gel combined with low temperature allowed for the growth of nanoparticles of hydroxyapatite with reduced crystallinity. The hybrid microparticles appear as microsized flakes, and it is possible to observe the presence of Gel inside the particle (Figure S1↗), whereas the nanostructured surface featured needle-like hydroxyapatite nanoparticles (FigureA).

Crystallographic analysis (FigureB) revealed the presence of characteristic signals of hydroxyapatite (according to PDF card #09-0432), which appeared rather broad. This indicates the formation of poorly crystalline mineral phases, a consequence of the biomineralization process carried out at low temperature, that determines the crystal growth in interaction with gelatin molecules. This interaction hinders the formation of a perfectly ordered crystal structure, promoting the formation of small crystallites. Through ATR–FTIR analysis (FigureC), both inorganic and organic components were detected. The organic component, attributable to gelatin (Gel), was identified by the presence of characteristic amide I and II bands at around 1650 cm^–1. However, due to the low organic/inorganic ratio (20/80) of bHA, the Gel-related peaks appeared poorly defined, though still clearly detectable. On the other hand, the prevailing presence of the inorganic component, made of hydroxyapatite, was recognizable by the well-defined phosphate signals within the range of 450–1200 cm^–1. The thermogravimetric analysis confirmed the inorganic/organic ratio of 20/80 (FigureD).

The rheological assessment performed on the bHA ink indicated that this has a shear thinning behavior with a viscosity of 33.9 kPa at 27 °C, the printing temperature, making it suitable for an effective printing process.

The rheological analysis performed at 27 °C (FigureA) confirmed the shear-thinning behavior of all tested formulations, a critical property for extrusion-based printing. Increasing the content of biomimetic hydroxyapatite (bHA) did not alter the overall shear-thinning profile, indicating that all formulations maintained excellent flow characteristics under applied stress. This behavior ensures consistent extrusion performance and reduces the risk of nozzle clogginga key requirement for precision printing.

Moreover, bHA particles have previously been shown to enhance the rheology of gelatin-based systems, acting as multifunctional additives that not only provide osteogenic cues but also improve structural fidelity during printing. The viscosity–temperature profile (FigureB) further highlights the thermoresponsive nature of the Gel matrix, with a general decrease in viscosity observed as temperature increases. Interestingly, a modest initial increase in viscosity was detected at low-to-moderate temperatures before the expected thermal thinning. This behavior can be attributed to temperature-activated associative interactions between polymer chains and bHA nanoparticles, which transiently strengthen the network structure before the onset of thermal disruption at higher temperatures. This trend reflects the increased molecular mobility of Gel chains at elevated temperatures. Notably, the bHAGel_1 formulation exhibited a more stable viscosity profile across the temperature range, due to the higher mineral phase content. The abundant bHA appears to hinder Gel chain mobility, thereby moderating the viscosity drop and contributing to better temperature stability, a crucial feature for maintaining consistent printability across varying conditions.

The printability of the developed biomaterial inks was evaluated in terms of extrudability, buildability, and model fidelity, as summarized in FigureC–F. The filament drop test demonstrated a consistent and continuous flow across all formulations, with filament diameters matching the nozzle size. Notably, the bHAGel_1 sample showed minor swelling and curling, suggesting that the inclusion of 50% bHA provides adequate pregelation and an optimal printing temperature for stable extrusion. This test, which reflects the initial gelation state of the ink, is critical for achieving well-defined 3D structures with controlled porosity. Indeed, the cylindrical and conical constructs printed (FigureD) revealed that increasing the mineral content improved the buildability of the structures, preventing collapse or deformation. This suggests that the addition of bHA nanoparticles enhances the ink's flow behavior by shifting its viscoelastic response from elastic to plastic, without compromising the rapid temperature-triggered gelation typical of thermosensitive hydrogels. Filament spreading analysis (FigureE) was used to assess structural fidelity semiquantitatively. The Pr factor (pore shape factor) revealed that all tested inks generated pores with shapes close to the ideal square geometry (Pr ≈ 1), falling within the defined printability range of 0.9 ≤ Pr ≤ 1.1. This indicates that, despite differences in buildability, all inks are printable and capable of forming regular porous patterns. Interestingly, the slightly higher Pr values (>1) observed in some formulations suggest a more gelled state at the time of extrusion, whichwhile potentially reducing flowenhanced shape retention and structural accuracy postdeposition. As emphasized in previous studies, printability is inherently multifactorial, and no single parameter can fully capture the performance of a biomaterial ink. Considering these favorable rheological characteristics, the reliable printability, and the compositional and structural resemblance to the native bone tissue, the bHAGel_1 formulation was selected as the bioink of interest for this study. For simplicity, it will hereafter be referred to as bHAGel.

Gel was used both as a biomineralization template for bHA and an organic matrix to develop composite biomaterial ink. ⁴⁴

To this end, sodium citrate was employed to improve bHA colloidal stability by promoting its electrostatic interaction with the Gel matrix, preventing sedimentation and a consequent uneven distribution of the biomineralized phase within the layers during printing.This specifically considered the high loading of bHA (1:1 with respect to the organic matrix). , ^{45 46}

The development of dual-porosity scaffolds, achieved by combining 3D printing with freeze-drying, represents a crucial strategy for engineering constructs intended for bone regeneration in perfusion bioreactors. While 3D printing enables the fabrication of well-defined macroporous architectures with controlled filament orientation and interconnectivity, it alone does not generate the microporous texture necessary to fully mimic the native extracellular matrix of bone. To investigate the influence of drying methods on porosity, we compared freshly printed filaments, air-dried filaments, and freeze-dried filaments (Figure). In both freshly printed and air-dried samples, the observed porosity was limited to the macropores defined by the 3D printing design. In contrast, freeze-dried samples exhibited a secondary level of porosity, characterized by an interconnected microporous network within the filaments themselves, resulting from ice sublimation during the lyophilization process.

Quantitative image analysis (using ImageJ thresholding and contour detection calibrated on the scale bars) revealed that the macropores between printed filaments exhibited an average diameter of approximately 800–900 μm, while the internal micropores within each filament ranged between 80 and 100 μm. Such hierarchical organization closely matches the dual-scale porosity typically considered optimal for bone tissue engineering, where macropores (>300 μm) support vascularization and tissue ingrowth, and micropores (10–150 μm) enhance cell attachment, nutrient diffusion, and matrix deposition.This microporosity enhances the internal surface area, facilitates the diffusion of nutrients and oxygen, and improves cell infiltration and matrix deposition under dynamic culture conditions. Thus, the synergy between structural macroporosity from 3D printing and textural microporosity from freeze-drying is crucial for designing biofunctional scaffolds that are compatible with perfusion bioreactors and can support effective bone tissue regeneration. , ^{47 48 49}

A physicochemical characterization of the bHAGel composite was performed to confirm that the properties of bHA were preserved during the 3D printing process. XRD analysis (FigureA) demonstrated that the bHA particles retained their low crystallinity; the diffractogram is broader for the significant presence of gelatin. Thermogravimetric analysis (FigureB) revealed a distinct profile due to the addition of the Gel phase, confirming a bHA loading of 50% w/w. This composition resulted in an overall organic-to-inorganic ratio of 58:42 in the final injectable formulation. Interestingly, the ATR–FTIR profile showed, compared to that of bHA, increased peaks corresponding to the primary and secondary amides around 1650 cm^–1, typical of Gel, and consistent with the increased amount of Gel in the formulated composite biomaterial ink (FigureC).

Since the last aim of this construct is to be fitted in a bioreactor, testing its mechanical properties in bioreactor-like conditions becomes crucial to better understand and predict its performance in long-term cultures. bHAGel was tested overnight preconditioned in PBS at 37 °C and then tested in uniaxial submersion-compression mode at the same conditions (FigureD). The Young Modulus (E), determined as the angular coefficient of the stress–strain response from 5 to 15% strain, was found to be 21.1 kPa. This value is within the typical range reported for soft hydrogels used in tissue engineering applications and is particularly well-suited to mimicking the mechanical microenvironment of early bone marrow or trabecular bone niches. While not as stiff as the native cortical bone, which is in the GPa range, this level of stiffness is adequate for perfusion bioreactor culture and supports cell proliferation, migration, and extracellular matrix deposition under dynamic conditions.

Since the dried scaffold is intended for use under wet conditions, its swelling behavior was evaluated after incubation in medium at 37 °C for different time points, as well as its degradation profile under static conditions up to 28 days (FigureE,F). bHAGel exhibited a high swelling capacity reaching approximately 700% within the first hour. This confirms the successful generation of bimodal porosity, which facilitates efficient cell infiltration and nutrient diffusion. The pronounced swelling can be ascribed to the hydrophilic nature of the polymeric matrix and the presence of interconnected pores that allow rapid water uptake through capillary forces.− Importantly, the high swelling did not compromise scaffold stability, as demonstrated by a limited degradation of about 20% after 28 days under cell culture-like conditions. The relatively slow degradation rate is likely due to the stable cross-linked network that resists hydrolytic cleavage, ensuring the maintenance of structural integrity over prolonged incubation.− Overall, these features support the suitability of bHAGel for long-term in vitro applications, where a balance between water absorption and mechanical stability is crucial.