3D-printed TCP-HA scaffolds delivering MicroRNA-302a-3p improve bone regeneration in a mouse calvarial model

The study addresses the challenge of regenerating critical-sized dentoalveolar bone defects that do not heal spontaneously, which is vital for restoring function and aesthetics prior to dental prostheses or implants. Conventional graft options (autografts, allografts, xenografts) have limitations in availability, cost, and immunogenic risks. Synthetic calcium phosphate-based substitutes, especially biphasic mixtures of tricalcium phosphate (TCP) and hydroxyapatite (HA), can be 3D-printed into personalized scaffolds and offer osteoconductivity and volume stability. Enhancing osteoinduction via bioactive molecules is a key goal; microRNA-302a-3p has been shown to promote osteoblast differentiation by repressing COUP-TFII, a repressor of RUNX2 and activator of RANKL. However, miRNAs are labile and require delivery systems. HA nanoparticles (HA-NPs), especially when surface-modified with APTES to increase positive charge, can condense and deliver miRNAs effectively and are biocompatible and osteoconductive. The hypothesis is that a 3D-printed TCP/HA scaffold functionalized with HA-NPs-APTES delivering miR-302a-3p will enhance bone regeneration in a critical-sized mouse calvarial defect.

The paper reviews limitations of traditional grafting materials and advantages of synthetic, 3D-printed biphasic calcium phosphate (TCP/HA) scaffolds for craniofacial bone regeneration, including better outcomes with block versus particulate forms and the ability to personalize scaffolds based on imaging. It summarizes the osteoconductive and mechanical roles of HA and the resorptive, osteogenic microenvironment provided by TCP, with optimal HA/TCP ratios facilitating angiogenesis and bone regeneration comparable to autografts. The authors discuss bioactive augmentation strategies using growth factors (BMP2, VEGF, FGF) and focus on miRNAs as post-transcriptional regulators. Prior work shows miR-302a-3p promotes osteogenesis by repressing COUP-TFII, which otherwise suppresses RUNX2 and increases RANKL. HA-NPs can bind nucleic acids via calcium-phosphate interactions; APTES functionalization increases surface charge, improving miRNA condensation and uptake. Previous in vitro studies demonstrated HA-NPs-APTES delivering miR-302a-3p enhanced osteogenic differentiation and mineralization, motivating in vivo evaluation.

In vitro: - HA-NPs-APTES synthesis and labeling: Hydroxyapatite nanoparticles were synthesized by mixing 0.25 M calcium nitrate and 0.15 M phosphate at pH 10, hydrothermally treating at 120 °C for 10 h, washing, and drying. Surface modification used 2.5% v/v APTES in anhydrous toluene (0.2 g HA-NPs in 20 mL) for 3 h, followed by washing and drying at 60 °C for 24 h. For imaging, HA-NPs-APTES were FITC-labeled by incubation in 0.2 mg/mL FITC in ethanol for 24 h, then washed and dried. - 3D-printed TCP/HA scaffolds: TCP/HA cement (Plotter-Paste-CPC) was printed (regenHU 3D Discovery) at 1 bar, room temperature, into cylinders consisting of 4 orthogonal layers of 250 μm rods with 200 μm inter-rod spacing. Scaffolds were hardened by immersion in sterile water for 48 h and air-dried. - Surface modification methods: • Method 1 (M1): Harden scaffolds directly in a suspension containing HA-NPs-APTES (50 μg/mL) and miR-302a-3p (5 nM) prepared in RNase-free/sterile water for 48 h. • Method 2 (M2): Harden scaffolds in sterile water for 48 h, remove water, then drop 10 μL of HA-NPs-APTES (50 μg/mL) and miR-302a-3p (5 nM) onto scaffold surfaces and air-dry 48 h. • Additional control: scaffolds directly conjugated with miR-302a-3p without nanoparticles (S-Mi) by hardening in 5 nM miR solution for 48 h. - Cell culture: Human osteosarcoma lines HOS and MG-63 and primary human mandibular-derived osteoblasts (HmOBs; passages 5–8) were cultured in DMEM + 10% FBS + antibiotics at 37 °C/5% CO2. For biocompatibility, cells (3 × 10^5) were seeded onto scaffolds in 24-well plates, allowed to adhere 24 h, then cultured up to 21 days with media changes every 2 days. - Assays: • Resazurin reduction assay on days 4, 7, and 11 to assess metabolic activity. • Fluorescence microscopy: FITC-tagged HA-NPs-APTES distribution on scaffold surfaces and cross-sections; FM4-64 membrane stain and fixation for cell uptake imaging. • qRT-PCR: RNA extracted on days 1 and 6; miScript II RT for cDNA; Quantitect SYBR for miR-302a-3p (normalized to RNU6-2); SensiFAST for mRNAs COUP-TFII, RUNX2, ALP, OCN, OSX (normalized to GAPDH). Relative quantification by ΔΔCt vs cells on culture plates. Primers as listed in Table 1. In vivo: - Ethics: ARRIVE 2.0 compliant; protocol approved (2173015). Power analysis targeted 0.8 power, alpha 0.05. - Animals: 48 male C57BL/6 mice, 8 weeks old; acclimated 2 weeks. - Groups (n=4 per group per time point): (1) no scaffold (Control); (2) bare scaffold; (3) scaffold+HA-NPs-APTES (M2) without miR; (4) scaffold+HA-NPs-APTES-miR-302a-3p (M2). - Surgery: Under anesthesia (tiletamine-zolazepam 20 mg/kg + xylazine 2 mg/kg, with ocular lubrication), local mepivacaine with epinephrine; 1 cm scalp incision; 4 mm circular defect in left calvarium with trephine; allocated scaffold placed without fixation; suturing for retention. Perioperative enrofloxacin 5 mg/kg and carprofen 5 mg/kg. Daily monitoring; exclusion if weight loss >15% in 7 days. - Time points: Euthanasia at 2, 4, and 6 weeks; samples for micro-CT and histology. - Micro-CT: Fixed in 4% PFA; scanned (Scanco μCT35) at 70 kV, 2375 ms, 16.259 μm voxel; blinded analysis. Outcomes: BV/TV within scaffold area; pore analysis: count whole pores (initially 74) and compute percentage unfilled by new bone. - Histology and histomorphometry: Decalcification, paraffin embedding, sagittal sections through center; H&E and Masson’s trichrome. Digital scanning (KEYENCE VHX-6000). Regions defined across 4 mm scaffold: border (0–1 mm and 3–4 mm from edge) and center (1–3 mm). Quantify new bone area fraction per region with blinded assessment. - Statistics: Shapiro–Wilk normality; variance homogeneity; one-way ANOVA with Tukey post hoc; significance at p<0.05. Data reported as mean ± SD/SEM.

In vitro: - HA-NPs-APTES were present on both superficial and deep scaffold layers for M1 and M2, and internalized by HOS cells. - Biocompatibility: Initial HOS proliferation on M1 was ~25% lower than control at early time points but recovered by day 21; overall M1 and M2 were biocompatible. - miRNA delivery: All modified scaffolds increased miR-302a-3p levels by day 6 in HOS, MG-63, and HmOBs; M1 and especially M2 achieved higher miR levels than scaffolds directly conjugated with miR (S-Mi). - Gene expression: miR-302a-3p delivery led to COUP-TFII downregulation and RUNX2 upregulation, most pronounced with M2. In MG-63 cells, COUP-TFII decreased ~5-fold with S-Mi, S-Mi-M1, S-Mi-M2; RUNX2 increased 6.34±0.99-fold (M1) and 7.38±0.68-fold (M2) vs control (both p<0.001). ALP, OCN, and OSX were upregulated notably with M2; S-Mi often did not differ from scaffold alone. In vivo (mouse calvarial 4 mm critical defects): - Micro-CT BV/TV at 2 weeks: bare scaffold 0.566±0.042; Scaffold+HA-NPs-APTES 0.062±0.049; Scaffold+HA-NPs-APTES-miR 0.653±0.034 (p<0.05 vs scaffold alone). With HA-NPs-APTES-miR, BV/TV increased to 0.656±0.042 at 4 weeks and 0.714±0.037 at 6 weeks, while control conditions showed little change. - Pore filling (% unfilled pores out of 74): At 2 weeks: Scaffold+HA-NPs-APTES-miR 45.5±4.1 vs scaffold 62.16±4.9 and Scaffold+HA-NPs-APTES 55.41±4.9. At 4 weeks: miR group 2.7±2.58 vs scaffold 49.55±16.01 and HA-NPs-APTES 24.77±13.26. At 6 weeks: miR group 0; HA-NPs-APTES 8.11±3.58; scaffold 21.62±3.6. - Histology/histomorphometry: Earlier and more extensive new bone and vascular ingrowth with HA-NPs-APTES-miR, including center-directed growth by 2 weeks. New bone area (%) within entire scaffold: at 2 weeks: scaffold 2.29±0.21; HA-NPs-APTES 3.7±1.7; HA-NPs-APTES-miR 5.35±0.15. At 4 weeks: 2.48±0.85; 3.26±0.7; 8.3±1.05 (miR significantly higher). At 6 weeks: 3.86±0.33; 5.08±0.46; 14.9±3.65 (miR significantly higher). - Center region new bone (% area): 2 weeks: miR 7.5±2.7 vs HA-NPs-APTES 1.41±1.2 and scaffold 0.38±0.3. 4 weeks: miR 9.25±2.67 vs HA-NPs-APTES 1±0.79 and scaffold 3.4±1.97. 6 weeks: miR 17±5.77 vs ~2% in both controls. The relative contribution of center region to total new bone in miR group rose from ~50% to ~60% between weeks 4 and 6, versus ~20% in HA-NPs-APTES without miR.

The findings support the hypothesis that functionalizing a 3D-printed osteoconductive TCP/HA scaffold with HA-NPs-APTES to deliver miR-302a-3p enhances osteoinduction and bone regeneration in critical-sized defects. In vitro, M2 surface adsorption yielded superior miRNA delivery and biological effects compared with M1 incorporation or direct miR conjugation, reflected by COUP-TFII repression and RUNX2 and other osteogenic markers upregulation. In vivo, miR-loaded scaffolds accelerated bone ingrowth, improved BV/TV, and promoted earlier and greater central pore filling, indicating enhanced osseoconduction and likely osteoinduction. The approach leverages the mechanical and architectural advantages of TCP/HA scaffolds while adding a gene-regulatory cue to drive osteoblast differentiation. These results are relevant to craniofacial bone regeneration where defect geometry can be matched by 3D printing and biological healing enhanced by miRNA delivery.

A 3D-printed TCP/HA scaffold functionalized with APTES-modified hydroxyapatite nanoparticles effectively delivered miR-302a-3p to bone cells, downregulated COUP-TFII, upregulated RUNX2 and other osteogenic genes, and significantly enhanced bone regeneration in a mouse critical-sized calvarial defect, particularly accelerating central defect filling. Method 2 (post-setting adsorption) outperformed Method 1 for miRNA delivery and regenerative outcomes. This strategy augments the inherent osteoconductivity of TCP/HA with osteoinductive signaling from miRNA cargo. Future work should optimize loading and release kinetics, compare additional miRNAs or combination therapies, evaluate long-term scaffold resorption and remodeling, assess dose-response and safety, and translate to larger animal models and patient-specific defect geometries.

- Small sample size per group (n=4) and use of a single mouse model limit generalizability. - Short follow-up (up to 6 weeks); no evidence of scaffold material resorption within this timeframe. - Potential inconsistency in BV/TV values among control groups at early time points; limited detail on release kinetics and in situ miRNA persistence. - Mechanistic differences between M1 and M2 loading approaches were not systematically investigated. - Only one miRNA dose (5 nM) and nanoparticle concentration (50 μg/mL) were tested; S-Mi (miR without NPs) was not included in vivo due to limited in vitro efficacy.