Multifunctional ions/drugs co-delivering nanocomposite hydrogel orchestrates neuro-osteogenic microenvironment for boosting osteoporotic osseointegration

Abstract

With population aging, the incidence of osteoporosis continuously elevates worldwide, resulting in increased fracture risks and clinical demand for orthopedic fixation. However, under osteoporotic conditions, the stability and longevity of implants are severely compromised by the pathological microenvironment, thus developing effective therapeutic interventions to achieve successful osteoporotic osseointegration remains a critical challenge in the regenerative medicine field. Herein, the parathyroid hormone (PTH) is encapsulated in Sr²⁺/Zn²⁺ co-doped mesoporous silica nanoparticles (SrZn-MSNs) to synthesize P@SrZn-MSNs, which are then incorporated into gelatin methacrylate (GelMA) containing calcitonin gene related peptide (CGRP) for realizing bioactive ions/drugs co-delivery. The obtained Gel/P@SrZn-MSNs/G nanocomposite hydrogels are characterized by excellent injectability and photosensitivity, 3D interconnected porous architecture, durable degradation and concurrent releasing of Si/Sr/Zn ions, CGRP and PTH controlled within proper doses. These favorable properties endow Gel/P@SrZn-MSNs/G with superior biological multifunctionality that not only promotes functionality of DRG sensory neurons and BMSCs directly, but also manipulates mutually beneficial interactions of DRG/BMSCs to further stimulate neurogenesis/osteogenesis. Specifically, Gel/P@SrZn-MSNs/G mainly increases NGF and BDNF expressions in BMSCs to stimulate axonal outgrowth and activation of DRG neurons, and in turn, the up-regulation of Sema3A in DRG neurons could be responsible for the neuro-inductive osteogenesis. Consistently, the in vivo outcomes further validate the potent and timely peri-implant innervation and bone formation induced by Gel/P@SrZn-MSNs/G when applied to femoral condyle defects of OVX rats. These results highlight the efficient ions/drugs co-delivery system renders Gel/P@SrZn-MSNs/G exerting synergy bioeffects to orchestrate desired neuro-osteogenic niches for boosting osteoporotic osseointegration, thus may inspiring new therapeutic strategies for osteoporotic bone healing.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04398-y.

Introduction

With constant aging of the population, osteoporosis has emerged as a debilitating degenerative bone disease that affects hundred millions worldwide, resulting in heavy medical burden and public health concerns [1]. As a systemic skeletal disorder characterized by deteriorated trabecula microarchitecture and decreased bone mass, osteoporosis greatly increases the fracture risks and thus boosts the demand for orthopedic fixation [2]. However, the stability and longevity of bone implants are severely compromised by the abnormal bone metabolism circumstance of osteoporosis, due to its hypoactive bone formation coupled with overactive bone resorption ability [3, 4]. Under osteoporotic conditions, bone implants such as titanium used in fracture fixations or joint replacements are frequently subjected to loosening, leading to unexpected implant failure, revision surgeries or delayed healing [5, 6]. Therefore, developing effective therapeutic interventions to achieve successful osteoporotic osseointegration remains a critical challenge in the regenerative medicine field [7, 8].

The key prerequisite to address osteoporotic osseointegration is a profound understanding of the microenvironmental hallmarks of osteoporosis [9]. It is acknowledged that bone microenvironment is synchronously regulated by an interactive network generated by nervous, vascular, immune and osteoprogenitor cell lineages [10]. In addition to imbalanced osteogenesis/osteoclastogenesis, pathological osteoporotic niches are yet tightly associated with impaired innervation especially for sensory nerves, which widely distribute from periosteum to trabeculae and exert indispensable effects to modulate bone homeostasis [11, 12]. Recently, emerging evidence highlights the pivotal roles of skeletal interoception on determining pathological bone diseases, wherein denervation or dysfunctional sensory nerves accompanied by diminished CGRP secretion immediately results in low bone mass and osteoporosis [13, 14]. On this basis, innovative strategies targeting neuro-inductive osteogenesis have been intensively applied by activating sensory nerves or direct supplement of CGRP for promoting bone healing, providing a promising therapeutic target towards ameliorative osteoporotic osseointegration [15, 16].

On the other hand, one major issue for inducing osteoporotic osseointegration is fabrication of advanced biomaterial systems that can deliver bioactive agents to bone-implants interface safely and effectively. A well established injectable photo-crosslinking hydrogel paradigm is highly appreciated since hydrogels (e.g. GelMA, hyaluronic acid) can not only fill the gap around implants to support osseointegration but also serve as biocompatible vehicles to release bioactive peptides (e.g. CGRP) for inducing innervated bone regeneration [5, 17, 18]. Moreover, the incorporation of nanocarries such as mesoporous silica nanoparticles (MSNs) can further endow hydrogels with biological multifunctionality via spatiotemporal releasing of diverse therapeutics [19]. As one typical silica-based nanomaterial, MSNs, have gained great interests and become a compelling candidate in bone tissue engineering mainly for satisfactory biocompatibility, specific mesostructure, tunable size/composition, excellent drug delivery and controlled release ability [20, 21]. These advantageous properties further contribute to the developments of ions/drugs co-delivery systems based on ions doped MSNs and drug encapsulation, which has shown great promise to improve bone repair outcomes by exerting synergy effects [22]. Thus, scholars have developed multifunctional MSNs by co-delivering bioactive ions (e.g. Ag, Sr) and biomolecules (e.g. BMP-2, vancomycin) for enhancing bone regeneration [23, 24]. Our previous research yet successfully synthesized novel Sr/Zn co-doped MSNs (SrZn-MSNs) that revealed superior biological multifunctionality to significantly promote osteoporotic osseointegration when combined with GelMA hydrogels, mainly owing to synergy effects of bioactive Sr²⁺/Zn²⁺ [25].

In this view, present study developed novel Gel/P@SrZn-MSNs/G nanocomposite hydrogels by simultaneously incorporating PTH loaded SrZn-MSNs (P@SrZn-MSNs) and CGRP (G) into GelMA (Gel), followed by photo-crosslinking with UV irradiation. PTH, an anabolic peptide approved by FDA for treating osteoporosis, was encapsulated in SrZn-MSNs to realize sustained and controlled release in local sites, which may actively induce osteogenesis and osseointegration under osteoporotic conditions [26, 27]. Meanwhile, the unencapsulated CGRP released relatively rapid from composite hydrogels for stimulating sensory innervation and bone formation [16]. According to the aforementioned cues, the designed Gel/P@SrZn-MSNs/G composite hydrogels may possess desired multifunctionality to orchestrate neuro-osteogenic niches for dictating functional osteoporotic osseointegration. For experimental investigations, the physico-chemical properties of materials were comprehensively characterized and the synergy actions of bioactive SrZn-MSNs and PTH/CGRP released from composite hydrogels were examined in vitro and in vivo.

Materials and methods

Synthesis of SrZn-MSNs and drug loading

The synthesis of SrZn-MSNs was conducted based on our prior research [25]. Briefly, 5 mL ammonium hydroxide (Sigma-Aldrich, USA) was mixed with 30 mL deionized water (Millipore, Germany) and 215 mL absolute ethanol (Sigma-Aldrich, USA) under room temperature. Then, 5 mL TEOS (Alfa Aesar, USA) was slowly added to the mixture under constant stirring, followed by supplement of Sr(NO₃)₂.4H₂O (Alfa Aesar, USA) and ZnCl₂ (Alfa Aesar, USA) with a designated molar ratio of Sr/Zn/Si to be approximately 1/2.08/70.56. After thorough reactions for 90 min, the resulting solid silica nanoparticles (sSiO₂) were purified by 75% ethanol and dried for further use. To develop mesoporous silica (mSiO₂) shell onto sSiO₂, 100 mg CTAB (Sigma-Aldrich, USA) was dissolved in a mixture of 20 mL absolute ethanol and 20 mL ultrapure water (Millipore, Germany), followed by supplement with 500 µL ammonium hydroxide (Sigma-Aldrich, USA) under constant stirring for 20 min. Thereafter, the synthesized sSiO2 (5 mg/mL) and 2 mL TEOS were added to the mixture under constant stirring, followed by supplement of Sr(NO₃)₂.4H₂O (Alfa Aesar, USA) and ZnCl₂ (Alfa Aesar, USA) with a designated Sr/Zn/Si molar ratio mentioned prior. Following overnight reactions, the solution underwent centrifugation to collect core-shell structural sSiO2@mSiO2 nanoparticles, which were then etched in 0.4 M Na₂CO₃ solution (Sigma-Aldrich, USA) with constant stirring under 50 °C. After 10 h etching, the products were purified by a mixture of 3 mL HCl/50 mL methanol and then centrifuged to yield hollow mesoporous structural SrZn-MSNs.

For the drug encapsulation into SrZn-MSNs, PTH peptide (P, Medchem Express, China) was dissolved in deionized water (Beyotime, China) at a concentration of 100 µg/mL and then mixed with 0.1 g SrZn-MSNs under continuous stirring for 24 h. The products were centrifuged and washed with PBS (Beyotime, China) twice to obtain PTH-delivered SrZn-MSNs (P@SrZn-MSNs). The residual PTH in supernatants was measured using a PTH ELISA kit (Phoenix Pharmaceuticals, USA) to calculate the drug loading amounts and encapsulation efficiency.

Synthesis of GelMA and composite hydrogels

The synthesis of GelMA and its mixture with SrZn-MSNs was performed based on our prior research [25]. In brief, gelatin (Sigma-Aldrich, USA) was dispersed in 10% PBS with constant stirring until complete dissolution, followed by dropwise addition of methacrylic anhydride (Sigma-Aldrich, USA) to react for 6 h. Thereafter, the reaction liquid was diluted by PBS and transferred into a dialysis bag (7000 kDa) for dialyzing against ultrapure water (Millipore, Germany) under 40 °C until 1 week, wherein ultrapure water was exchanged twice per day. The obtained solution was subjected to filtration and freeze-drying to yield pristine GelMA (Gel) for use.

To develop composite hydrogels, 3% (w/w) P@SrZn-MSNs were dissolved in 8% (w/v) GelMA solution containing 1 µM CGRP (G, Medchem Express, China) and 0.5% photoinitiator LAP (EFL-Tech, China), followed by sonication to form homogeneously dispersed hydrogel precursor. The precursor was crosslinked using UV irradiation (365 nm) to obtain pristine Gel and composite hydrogels (Gel/SrZn-MSNs, Gel/SrZn-MSNs/G, Gel/P@SrZn-MSNs/G). For better understanding the superiority of encapsulated PTH delivery, Gel/P@SrZn-MSNs were developed by incorporating P@SrZn-MSNs into GelMA without CGRP. Thereafter, based on the calculated encapsulation efficiency of P@SrZn-MSNs, an equal dose of PTH was directly incorporated into GelMA and then combined with SrZn-MSNs to develop Gel/SrZn-MSNs/P composite hydrogels.

Materials characterization

The morphology of synthesized nanoparticles and hydrogels was examined by scanning electron microscopy equipped with energy dispersive spectroscopy (SEM-EDS, Quanta 250, USA) at a voltage of 5 kV. The mesoporous architecture of synthesized nanoparticles was detected by transmission electron microscopy (TEM, HT7700, Japan) at a voltage of 100 kV. The specific pore size and surface area of synthesized nanoparticles were analyzed by N₂ adsorption-desorption isotherms generated by the gas adsorption analyzer (AS1-MP, USA). The particle size distribution was determined by a granularity analyzer (Zetasizer Nano ZS90, UK) under 532 nm laser. Compared to gelatin, the methylation degree of synthesized GelMA was examined by a ¹H NMR spectrometer (Mercury plus 400, USA) under 400 MHz. The components of hydrogels were analyzed by FTIR spectroscopy (Nicolet 6700, USA) within 400–4000 cm^− 1 wavelength and 4 cm^− 1 resolution.

An acknowledged liquid displacement method was applied to determine the porosity of hydrogels by immersing lyophilized samples in anhydrous ethanol until complete penetration. To investigate degradation behavior of hydrogels, samples were soaked in PBS solution (10 mL, pH 7.4) with 120 rpm stirring under 37 ℃. At designated intervals (1/3/7/14/21 days), residual samples were lyophilized and weighed to measure the weight loss ratios (n = 4). For drug release assays, Gel/SrZn-MSNs/G and Gel/P@SrZn-MSNs/G hydrogels were respectively incubated in PBS (5 mL, pH 7.4) with 120 rpm stirring under 37 ℃. At designated intervals (1/3/5/7/9/11/14/17/21/25/30 days), the solution was collected and replaced with fresh PBS. The concentrations of CGRP and PTH in the extracted solution (n = 4) were determined by a CGRP ELISA kit (Phoenix Pharmaceuticals, USA) and PTH ELISA kit (Phoenix Pharmaceuticals, USA) respectively. Likewise, the release kinetics of PTH from Gel/SrZn-MSNs/P and Gel/P@SrZn-MSNs hydrogels were examined for the initial one week (1/3/5/7 days).

To investigate ions releasing patterns, composite hydrogels were incubated in DMEM medium with 120 rpm stirring under 37 ℃ for 30 days. The supernatants were collected at designated intervals (1/3/5/7/14/21/30 days) to measure the released ions (Si/Sr/Zn) concentrations (n = 4) by ICP-AES (Varian 720, USA). Additionally, 1 mL of crosslinked hydrogels were incubated in 10 mL of culture medium with 120 rpm stirring under 37 ℃ for 24 h to collect extracted medium, which was then subjected to centrifugation and filtration using a filter membrane (0.22 μm, Millipore, Germany) to produce materials extracts used for biological experiments. The concentrations of ions and drugs in materials extracts (n = 4) were determined by ICP-AES and ELISA kits mentioned above.

Cell harvest from OVX rats and culture on coated Ti disks

Thirty SD rats (female, 180 ~ 200 g) purchased from SPF (Suzhou) Biotechnology Co., Ltd. received bilateral ovariectomy and were kept 2 months to create ovariectomized (OVX) rats. All animal experiments were approved by the Institutional Animal Care and Use Committee of Suzhou Jingzida Testing Service Co., Ltd. (IACUC-2023-0006-DI). As compared to sham rats (n = 3), femurs of OVX rats were analyzed by micro-CT (Skyscan 1275, USA) and CTAn software for assessing the trabecular number (Tb.N) and bone volume percentage (BV/TV). Thereafter, specimens (n = 3) were decalcified and subjected to immunohistofluorescence staining of CGRP (Abcam, UK) for evaluating the intrabony sensory innervation. The sections were examined using a fluorescence microscope (Olympus CKX53, Japan) and analyzed by ImageJ to quantify the fluorescence intensity.

Dorsal root ganglion (DRG) tissues were collected from paravertebral region in OVX rats and immediately incubated in DMEM (Gibco, USA) supplemented with 10% FBS (Gibco, USA). The tissues were subsequently minced into fine fragments and subjected to enzymatic digestion using trypsin (Beyotime, China) and collagenase A (Sigma-Aldrich, USA). Thereafter, isolated DRG sensory neurons were subjected to centrifugation and resuspended in Neurobasal A medium (NA, Gibco, USA) containing 1% L-glutamine (LG, Gibco, USA), 2% B-27 (Invitrogen, USA) and 1% penicillin/streptomycin (PS, Beyotime, China).

Prior to cell culture, Ti disks (Ti-6Al-4V, Φ 8 mm) purchased from Biovet Yunao Co., Ltd. were coated with various hydrogels and then crosslinked by UV irradiation. The DRG suspensions were seeded at a density of 3 × 10⁴ cells/mL onto coated Ti disks and maintained in NA/LG/B-27/PS medium supplemented with 20 ng/mL NGF (Sigma-Aldrich, USA). On the other hand, BMSCs were isolated from femoral marrow cavities of OVX rats and cultured in α-MEM (Gibco, USA) supplemented with 10% FBS (Gibco, USA). Following three passages, BMSCs were seeded at a density of 2 × 10⁴ cells/mL onto coated Ti disks for biological investigations. To perform osteogenic differentiation experiments, BMSCs were cultured in a specialized osteogenic induction medium (Cyagen, China).

DRG/BMSCs co-culture in microfluidic devices

As shown in Fig. 4A, B, microfluidics assays were performed using one commercial microfluidic device (Xona Microfluidics, USA), in which central and outer compartments were arrayed and connected by microgroove barriers (500 μm). By supplying materials extracts to either DRG or BMSCs, this microfluidic device enabled DRG/BMSCs interacted with each other by their conditioned medium. Therefore, 20 µL DRG suspensions were seeded at a density of 3 × 10⁶ cells/mL into central compartments, and 20 µL BMSCs suspensions were seeded at a density of 2 × 10⁶ cells/mL into outer compartments. DRG and BMSCs were allowed to attach for 5 min and then supplemented with 150 µL medium respectively. Following 24 h culture,

Fig. 4.

Interactive neurogenesis/osteogenesis of DRG/BMSCs detected by microfluidics assays. (A) The microfluidic device used for co-culture of DRG/BMSCs. (B) Cellular interaction of DRG/BMSCs linked by microchannels. (C) The axonal outgrowth (first row) and intracellular Ca²⁺ activity (second row) of DRG stimulated by conditioned medium of co-cultured BMSCs. (D) Quantitation of DRG axonal length. (E) Quantification of fluorescent Ca²⁺ intensity. (F) Osterix immunofluorescence of BMSCs stimulated by conditioned medium of co-cultured DRG (DRG-CM). (G) Quantitation of fluorescent osterix. (H) ALP staining of BMSCs stimulated by DRG-CM. (I) Quantitative semaphorin family gene expressions of DRG cultured in materials extracts for 7 days. (J) ALP staining of BMSCs stimulated by DRG-CM of Gel/P@SrZn-MSNs/G with/without neutralizing antibodies. Quantitative (K) ALP and (L) Runx2 gene expressions of BMSCs stimulated by DRG-CM of Gel/P@SrZn-MSNs/G with/without neutralizing antibodies. Data are presented for mean ± SD. n = 4. ^**P < 0.01, ^*P < 0.05

the mixture of materials extracts and culture medium (1:3 ratio) was supplied to DRG for investigating the neuro-inductive osteogenesis effects on BMSCs. On the contrary, BMSCs were cultured in materials extracts mixed with culture medium (1:3 ratio) to stimulate the neurogenesis of co-cultured DRG.

Cell adhesion and viability

The adhesion of DRG/BMSCs was detected via actin cytoskeleton and nuclear staining following 12 h culture on coated Ti disks. Firstly, DRG/BMSCs were rinsed twice with PBS and fixed with 4% paraformaldehyde (Beyotime, China), followed by double permeabilization via 0.1% Triton X-100 (Beyotime, China). Thereafter, DRG/BMSCs were stained with a phalloidin kit (Beyotime, China) and observed under a confocal microscope (Olympus FV1000, Japan). Finally, quantitative analysis of cell spreading areas (n = 4) was performed using ImageJ.

The viability of DRG cultured on coated Ti disks was observed by live/dead staining. Following 1 and 3 days culture, DRG were rinsed thrice with PBS and then stained by Calcein AM/PI (Beyotime, China) at a dose of 0.002 mmol/L for 15 min. Then, the live (green) and dead (red) cells were observed using a fluorescence microscope (Olympus CKX53, Japan). Additionally, DRG (n = 4) was stained by MTT (Beyotime, China) at a dose of 5 mg/mL after 1 and 3 days culture, and the optical densities (OD) at 490 nm were measured with a microplate reader (Thermo MK3, USA) to assess cell viability. Likewise, the viability of BMSCs was determined by live/dead staining and MTT assays for 1, 3 and 7 days culture on coated Ti disks.

Immunofluorescence staining

DRG cultured on coated Ti disks for 3 days was rinsed and fixed with 4% paraformaldehyde (Beyotime, China), followed by permeabilization via 0.5% Triton X-100 (Beyotime, China) and blocking with normal goat serum (Beyotime, China). DRG were then incubated with primary NF200 antibody (Proteintech, China) under 4 °C for 2 h and Alexa Fluor 488-conjugated secondary antibody (Proteintech, China) for 1 h. Thereafter, DRG were rinsed, incubated with primary Sema3A antibody (Santa Cruz, USA) under 4 °C for 2 h and Alexa Fluor 594-conjugated secondary antibody (Proteintech, China) for 1 h. Ultimately, DRG were stained with DAPI (Beyotime, China) and detected under a fluorescence microscope (Olympus CKX53, Japan). For characterization of neuron-specific markers, isolated DRG neurons were co-stained by TUBB3 (Proteintech, China) and NeuN (Proteintech, China) antibodies following 12 h culture in 24-well plates. Likewise, BMSCs were subjected to Osterix/OCN/DAPI immunofluorescence using Osterix (Abcam, UK), OCN (Santa Cruz, USA) and DAPI (Beyotime, China) antibodies following 3 days culture on coated Ti disks. Additionally, BMSCs were stained with Osterix (Abcam, UK) and DAPI (Beyotime, China) antibodies after 3 days co-culture in the microfluidic device. For quantitation of immunofluorescence (n = 4), the fluorescence intensity was analyzed using ImageJ.

Fluorescent Ca²⁺ probe and axonal outgrowth

After cultured on coated Ti disks or microfluidic device for 3 days, DRG was rinsed thrice with PBS and then stained with DAPI (Beyotime, China). Subsequently, DRG were incubated in Hanks’ balanced solution (Beyotime, China) supplemented with 4 µM Fluo-4 AM (Beyotime, China) and 1 µM Pluronic F-127 (Beyotime, China). DRG were re-rinsed with PBS and incubated again with Hanks’ balanced solution (Beyotime, China) for 20 min under 37 °C. The fluorescent Ca²⁺ signal in DRG (n = 4) was observed using a fluorescence microscope (Olympus CKX53, Japan) and quantified with ImageJ. For evaluating axonal outgrowth (n = 4), DRG were incubated with TUBB3 (Abcam, UK) and DAPI (Beyotime, China) antibodies after 3 days co-culture in the microfluidic device, and the neurites growing into microchannels were examined by a fluorescence microscope (Olympus CKX53, Japan).

Intracellular cAMP measurements

Isolated DRG was rinsed thrice with HEPES buffer (Beyotime, China) and then cultured on coated Ti disks for 30 min. Then, the medium was removed and DRG was.

incubated in Hanks’ balanced solution (Beyotime, China) containing 0.1 M HCl and 0.5 mM IBMX (Beyotime, China) on ice for 15 min. The lysate was collected and centrifuged for 10 min under 4 °C to obtain supernatants, which were then examined by a cAMP ELISA kit (Cayman Chemical, USA). The concentrations of cAMP (n = 4) were normalized to total protein contents per well by the Bradford assay.

Flow cytometry

To identify the isolated BMSCs, a specific CD45/CD29/CD90 triple marked flow cytometry was performed. Briefly, isolated BMSCs cultured in medium were centrifuged and resuspended at a concentration of 1 × 10⁶ cells/mL, followed by incubation with PE anti-rat CD45 antibody (1 µL, Biolegend, USA), FITC anti-rat CD29 antibody (2 µL, Biolegend, USA) and APC anti-rat CD90 antibody (0.3 µL, Biolegend, USA) under 4 ℃ for 30 min. Thereafter, cells were washed by PBS twice and then analyzed by a flow cytometer (Agilent, USA) to determine the expression of CD45/CD29/CD90 surface markers.

ALP staining

BMSCs cultured on coated Ti disks for 7 days were rinsed thrice and stained with a commercial ALP staining kit (Beyotime, China) to mark the production of ALP. For microfluidics assays, ALP staining of BMSCs was conducted after 7 days co-culture with DRG that was cultured in materials extracts. To investigate the key molecules responsible for neuro-inductive osteogenesis of Gel/P@SrZn-MSNs/G, DRG was cultured in materials extracts of Gel/P@SrZn-MSNs/G with/without neutralizing antibodies of Sema3A (Santa Cruz, USA), Sema4D (Santa Cruz, USA) or Sema7A (Santa Cruz, USA) for 7 days, then co-cultured BMSCs were stained with a commercial ALP staining kit (Beyotime, China).

RT-qPCR analysis

Total RNA of BMSCs cultured on coated Ti disks for 7 days was extracted and reverse transcribed to cDNA by trizol (Beyotime, China) and First Strand cDNA Synthesis Kit (Beyotime, China) respectively. Then, one Bio-Rad RT-qPCR system was utilized to perform RT-qPCR analysis and the expression of osteogenic genes (ALP, Col I, OPN, Runx2) was calculated using 2^−∆∆Ct methods (n = 4). For microfluidics assays, neurotrophic genes (NGF, BDNF, neurotrophin-3, bFGF) expressions of BMSCs cultured in materials extracts for 3 days and semaphorin family genes (Sema3A, Sema3E, Sema4D, Sema7A) expressions of DRG cultured in materials extracts for 7 days were determined by RT-qPCR (n = 4). To investigate the key signaling molecules responsible for neuro-inductive osteogenesis of Gel/P@SrZn-MSNs/G, DRG was cultured in materials extracts of Gel/P@SrZn-MSNs/G with/without neutralizing antibodies of Sema3A (Santa Cruz, USA), Sema4D (Santa Cruz, USA) or Sema7A (Santa Cruz, USA) for 7 days, then osteogenic genes (ALP, Runx2) expressions of co-cultured BMSCs were determined by RT-qPCR (n = 4). The sequences of primers used are provided in Table S1 (Supplementary Information).

Western blot assays

After stimulated by materials extracts for 3 days, co-cultured BMSCs were washed thrice using TBST (Beyotime, China), lysed using RIPA Lysis Buffer (Beyotime, China) on ice and then centrifuged to extract proteins. The collected proteins were delivered to PVDF films (Millipore, Germany), blocked using nonfat milk (5%, Servicebio, China) and incubated with primary antibodies against NGF (Affinity, USA), BDNF (Affinity, USA), neurotrophin-3 (Affinity, USA), bFGF (Affinity, USA) and GAPDH (Abcam, UK) overnight at 4 ℃, followed by 1 h incubation using anti-rabbit secondary antibody (Affinity, USA). Likewise, semaphorin secretions of DRG co-cultured in materials extracts for 7 days were examined by using antibodies against Sema3A (Santa Cruz, USA), Sema3E (Thermo, USA), Sema4D (Santa Cruz, USA), Sema7A (Santa Cruz, USA) and GAPDH (Abcam, UK). To investigate the key signaling molecules responsible for neuro-inductive osteogenesis of Gel/P@SrZn-MSNs/G, the medium of co-cultured BMSCs were supplemented with/without Sema3A inhibitor SM345431 (2 µM, MedChemExpress, USA) or Wnt/β-catenin inhibitor XAV939 (10 µM, Selleckchem, USA) for 7 days, then western blot assays were performed by using antibodies against Nrp1 (Affinity, USA), p-Rac1 (Affinity, USA), Rac1 (Affinity, USA), Wnt6 (Abcam, UK), β-catenin (Abcam, UK), ALP (Abcam, UK), Runx2 (Abcam, UK) and GAPDH (Abcam, UK). The expressed protein bands were detected by a chemiluminescence device (Beyotime, China) and then analyzed for gray values (n = 4) by using AlphaEaseFC software.

Materials implantation in OVX rats

An acknowledged model of femoral condyle defects in OVX rats was conducted. All animal experiments were approved by the Institutional Animal Care and Use Committee of Suzhou Jingzida Testing Service Co., Ltd. (IACUC-2023-0006-DI). Twenty-four healthy OVX rats were randomly allocated into four groups: (1) Gel, (2) Gel/SrZn-MSNs, (3) Gel/SrZn-MSNs/G, and (4) Gel/P@SrZn-MSNs/G. The implantation procedure was shown in Fig. 6A. Briefly, rats were fully anesthetized and subjected to a longitudinal incision approximately 15 mm at knee joints for exposing bilateral femoral condyle. A drill was used to develop an implantation channel in the femoral myelocavity, followed by injection of hydrogels per group and crosslinking with UV irradiation (365 nm). Thereafter, Ti screws (Ti-6Al-4 V, 6 mm length) purchased from Biovet Yunao Co., Ltd. were sterilized and properly implanted into defects. The subcutaneous tissues, muscles and skins were adequately sutured. At 6 weeks post-implantation, rats were euthanized to harvest the implants.

Fig. 6.

Osteoporotic osseointegration evaluation of implants at 6 weeks post-surgery. (A) Surgical procedures of materials implantation in OVX rats. (B) 3D reconstruction of Ti implants and the representative sagittal, median transverse and coronal scopes viewed by micro-CT. (C) Undecalcified H&E sections of implants with surrounding bone. (D) Quantification of bone volume/total volume (BV/TV). (E) Quantification of bone mineral density (BMD). (F) Quantification of trabecular thickness (Tb.Th). (G) Quantification of bone-implant contact (BIC) ratio. Red arrows indicate osseointegration around implants. Data are presented for mean ± SD. n = 4. ^**P < 0.01, ^*P < 0.05

Osteoporotic osseointegration evaluation

Implanted specimens were scanned and 3D reconstructed using micro-CT (Skyscan 1275, USA) and Dataviewer software. Quantitative analysis of bone parameters (n = 4) including bone volume percentage (BV/TV), trabecular thickness (Tb.Th) and bone mineral density (BMD) were conducted with CTAn software. For evaluating peri-implant osseointegration, specimens (n = 4) were subjected to undecalcified H&E (Beyotime, China) staining and the bone-implant contact (BIC) was quantified with ImageJ. Additionally, decalcified specimens (n = 4) were sectioned and stained by H&E (Beyotime, China) and immunohistofluorescence of CGRP (Abcam, UK), CD31 (Abcam, UK) and Col Ⅰ/Sema3A (Santa Cruz, USA) at center sites of implants. The sections were examined using a fluorescence microscope (Olympus CKX53, Japan) and analyzed by ImageJ to quantify the fluorescence intensity. For mechanical assessment, specimens (n = 4) were subjected to pull-out tests via the testing device (Shanghai Hengyi Precision Instrument Co., Ltd.) as reported [5].

Statistical analysis

All data were indicated as mean ± standard deviation (SD). Statistical analysis was performed by one-way variance analysis (ANOVA) and unpaired t-test with Graphpad Prism 8. The p-value < 0.05 was statistically significant (^**P < 0.01, ^*P < 0.05).

Results and discussion

Characterization of materials

As detected by SEM-EDS and TEM, synthesized SrZn-MSNs revealed regular spherical shape and core/shell mesoporous architecture, accompanied by the uniform distribution of Si/Sr/Zn elements in the MSN framework (Fig. 1A). The results indicated the successful Sr/Zn co-doping and hollow mesoporous structures of SrZn-MSNs, and the encapsulation of PTH (P@SrZn-MSNs) well retained these desired features (Fig. 1A). Moreover, both SrZn-MSNs and P@SrZn-MSNs exhibited typical type IV curves with hysteresis loops in the N₂ adsorption-desorption isotherms (Fig. 1B), which further confirmed their mesoporous structures. For specific surface area and mesopore distribution (inset) analysis (Fig. 1B), the BET surface area and average pore size of SrZn-MSNs and P@SrZn-MSNs were measured at 224.57 m²/g, 334.79 m²/g and 4.92 nm, 5.22 nm, respectively. Based on this hollow mesoporous architecture within suitable pore sizes, the encapsulation efficiency and loading amounts of PTH were measured as 46.35% and 6.45 µg/mg of P@SrZn-MSNs respectively, indicating the desired drug loading ability of synthesized SrZn-MSNs. The particle size distribution of SrZn-MSNs was calculated as 223.8 ± 27.9 nm, and encapsulation of PTH slightly increased the particle size to 242.2 ± 33.9 nm for P@SrZn-MSNs (Fig. 1C).

Fig. 1.

Characterization of synthesized nanoparticles and hydrogels. (A) Morphology and mesoporous structure of nanoparticles detected by SEM-EDS and TEM. (B) N₂ adsorption-desorption isotherms and mesopore distribution (inset) of nanoparticles. (C) Analysis of particle size distribution. (D) Samples of crosslinked hydrogels and topography of lyophilized scaffolds detected by SEM. (E) Extrusion of injectable hydrogels. (F) ¹H NMR analysis of Gelatin and GelMA. (G) FTIR spectra of hydrogels. (H) Porosity of hydrogels. (I) Degradation of hydrogels. (J) Cumulative release of CGRP from Gel/SrZn-MSNs/G hydrogels. (K) Cumulative release of CGRP and PTH from Gel/P@SrZn-MSNs/G hydrogels. Cumulative release of Si/Sr/Zn ions from (L) Gel/SrZn-MSNs, (M) Gel/SrZn-MSNs/G, and (N) Gel/P@SrZn-MSNs/G. Data are presented for mean ± SD. n = 4. ^**P < 0.01, ^*P < 0.05. ns represents no significant differences

As revealed in Fig. 1D and E, the incorporation of SrZn-MSNs and peptides maintained the fundamental injectability and photosensitivity of pristine GelMA (Gel) for realizing the extrusion and UV-crosslinking of composite hydrogels. Following freeze-drying, each group of lyophilized scaffolds kept a stable and porous 3D architecture, and the SEM further showed the interconnected honeycomb-like porous structure of hydrogels (Fig. 1D). The ¹H NMR spectra verified the efficient binding between methyl acrylamide with gelatin molecules for synthesis of GelMA, in which the elevated methyl group signal (1.8 ppm) and diminished arginine signal (2.9 ppm) were observed with respect to pristine gelatin mainly due to the reaction between arginine residues and methacrylic anhydride (Fig. 1F). Moreover, the peaks (5.6 and 5.3 ppm) representing double carbon-carbon atoms bonded in GelMA further indicated the accomplished methacrylation of gelatin (Fig. 1F). As examined by FTIR, the sharp peaks at 1610 cm^− 1 and 1360 cm⁻¹ represented the stretching vibration and symmetrical vibration of amide-I and CH₃ groups respectively, which originated from the GelMA components (Fig. 1G). On the other hand, the characteristic peaks at 539 cm⁻¹, 775 cm⁻¹ and 1080 cm⁻¹ corresponded to the vibrations of Si-O-Si bonds for SrZn-MSNs framework (Fig. 1G), suggesting the effective combination of GelMA and SrZn-MSNs in composite hydrogels.

As shown in Fig. 1H, the average porosity of hydrogels revealed no significant differences between groups, wherein Gel was calculated to be 70.06% and the porosity of composite hydrogels decreased to 61.57% (Gel/SrZn-MSNs), 63.48% (Gel/SrZn-MSNs/G) and 65.06% (Gel/P@SrZn-MSNs/G). Each group of hydrogels endured a gradual degradation period up to 21 days, and the cumulative weight loss of Gel, Gel/SrZn-MSNs, Gel/SrZn-MSNs/G and Gel/P@SrZn-MSNs/G reached 72.2%, 61.8%, 64.9% and 67.9% respectively, revealing durable stability of composite hydrogels (Fig. 1I). For drug release assays, the releasing of CGRP revealed quite rapid in Gel/SrZn-MSNs/G (Fig. 1J) and Gel/P@SrZn-MSNs/G (Fig. 1K), wherein CGRP released beyond 80% from Gel/SrZn-MSNs/G (81.8%) and Gel/P@SrZn-MSNs/G (80.2%) with 9 days. In comparison, a sustained and controlled release of PTH that cumulatively released about 65.4% (726.7 ± 45.6 ng/mL) up to 30 days was detected in Gel/P@SrZn-MSNs/G (Fig. 1K and Figure S1, Supplementary Information), mainly attributed to the encapsulation and mesoporous properties of SrZn-MSNs. The ions releasing behaviors of composite hydrogels revealed comparable to the drug release kinetics, wherein Si, Sr and Zn ions continuously released from Gel/SrZn-MSNs (Fig. 1L), Gel/SrZn-MSNs/G (Fig. 1M) and Gel/P@SrZn-MSNs/G (Fig. 1N) until 30 days, which further validated the durable degradation and stability of composite hydrogels to ensure sustained ions/drugs releasing for prolonged bone healing. In particular, the concentrations of ions/drugs were determined in materials extracts to be applied for in vitro assays (Table S2, Supplementary Information).

In perspective of various nanocarries, there remains certain limitations hindering their biological performances including potential cytotoxicity (e.g. iron oxide nanoparticles), insufficient delivery ability (e.g. gold nanoparticles), mechanical fragility (e.g. hydroxyapatite) and poor osteoconductivity (e.g. PLGA) [28]. In comparison, MSNs are highly appreciated for their superior properties and MSNs-based nanotherapeutics have been extensively and effectively applied in biomedical fields especially for bone repair [20, 29]. Moreover, advanced ions/drugs co-delivery strategies were adopted in present study for improving the bioactivity of MSNs despite its self-releasing of Si ions could benefit bone healing [22, 30]. Based on our recent research [25], co-doping of Sr/Zn ions into MSNs was successfully performed to synthesize SrZn-MSNs, followed by encapsulation of PTH peptides for constructing bioactive P@SrZn-MSNs. These modifications well retained the inherent MSNs framework and hollow mesoporous architecture within suitable pore sizes, as determined by SEM-EDS, TEM and BET analysis. The encapsulation of PTH slightly increased the particle size of P@SrZn-MSNs compared to SrZn-MSNs probably by altering dispersion characteristics of particles as reported [31], which may potentially increase the specific surface area to facilitate serum protein adsorption and cellular interactions [22, 32].

To further address the pathological osteoporotic osseointegration, synthesized nanoparticles and CGRP were incorporated into GelMA for developing multifunctional composite hydrogels. GelMA was utilized as a vehicle for its wide applications in regenerative medicine, owing to the excellent injectability, photosensitivity and biocompatibility [17, 25]. Our results confirmed the efficient integration of nanoparticles and GelMA in composite hydrogels characterized by 3D interconnected porous architecture and representative phase peaks. The incorporation of nanoparticles and CGRP revealed minimal impacts on the native properties of GelMA, thus guaranteeing the feasibility to extrude composite hydrogels before UV-crosslinking. The porosity of composite hydrogels decreased compared to pristine Gel, since the interspaces of hydrogel networks were occupied by nanoparticles to form denser structures [33]. Hence, the increased hydrogel density could further contribute to reinforced 3D structure for resisting matrix breakdown, which resulted in prolonged degradation of composite hydrogels and thus may better meet the long-period osteoporotic osseointegration [25, 34].

In present study, the composite hydrogels were developed by a sophisticated incorporation of 3% (w/w) P@SrZn-MSNs and 1 µM CGRP into 8% (w/v) GelMA solution. Our recent research has demonstrated this formulation allowed SrZn-MSNs components to fully exert therapeutic efficacy by proper releasing of Si/Sr/Zn ions [25]. Herein, the characterization of ions releasing profile along with materials extracts also indicated Si, Sr and Zn ions released continuously from composite hydrogels within appropriate concentrations conducive to bone healing [22, 25, 35]. On this basis, SrZn-MSNs were utilized to encapsulate PTH for its inherent mesoporous architecture, resulting in a sustained and controlled release of PTH reported earlier [36]. As examined by ELISA kits, PTH encapsulated in Gel/P@SrZn-MSNs/G initially released to materials extracts (Table S2, Supplementary Information) within moderate concentrations (149 ± 37 ng/mL) and cumulatively released to 726.7 ± 45.6 ng/mL until 30 days (Figure S1, Supplementary Information). As previously reported, PTH-related peptides were able to exert anabolic effects on osteoblasts at a concentration range of 50 ~ 400 ng/mL and revealed no significant cytotoxicity within 800 ng/mL concentrations during 5 days culture period [26]. In present study, DRG or BMSCs were maintained for 3 or 7 days culture on coated Ti samples for biological investigations, the concentrations of PTH released from Gel/P@SrZn-MSNs/G were considered within proper ranges and non-cytotoxic due to the controlled release kinetics rather than burst release. Moreover, the exposure doses of PTH to co-cultured DRG/BMSCs were further diluted by mixing materials extracts with culture medium (1:3 ratio), thus may resulting in a quite safe and effective PTH exposure in vitro. On the other hand, CGRP were directly incorporated into composite hydrogels to realize a rapid release, which could probably synergize with P@SrZn-MSNs to induce timely innervation and bone regeneration [16]. In this view, present results highlight the successful synthesis of composite hydrogels characterized by desired physico-chemical properties for biological applications.

Activation of sensory neurons

Isolated DRG neurons grew well and formed abundant neurofilament networks stained by specific neuronal markers TUBB3 (β-III Tubulin) and NeuN (Figure S2 and Figure S3, Supplementary Information), suggesting a high survival and purity of DRG neurons harvested from OVX rats. To evaluate the bioeffects of hydrogels on sensory neurons, DRG neurons were harvested and directly cultured on coated Ti samples (Fig. 2A). Consequently, DRG spontaneously adhered and exhibited specific neuronal dendrite and axone cytoskeleton for 12 h culture on coated Ti samples, and the quantitative cell spreading area increased in Gel/SrZn-MSNs/G and Gel/P@SrZn-MSNs/G, suggesting beneficial effects of CGRP on adhesion of DRG neurons (Fig. 2B, C). The live/dead staining and fluorescent Ca²⁺ probe results further indicated the desired biocompatibility and bioactivity of samples to support survival and activation of DRG neurons, especially for Gel/SrZn-MSNs/G and Gel/P@SrZn-MSNs/G (Fig. 2B). Consistently, the MTT assays verified the significantly enhanced viability of DRG for 3 days culture on Gel/SrZn-MSNs/G and Gel/P@SrZn-MSNs/G than Gel (Fig. 2D).

Fig. 2.

Activation of sensory neurons cultured on Ti samples coated by hydrogels. (A) Harvest and culture procedure of DRG neurons for biological investigation. (B) The adhesion (first row), viability (second and third row) and intracellular Ca²⁺ activity (fourth row) of DRG. (C) Quantitation of cell spreading. (D) Quantitation of cell viability by MTT assays. (E) Quantification of fluorescent Ca²⁺ intensity. (F) NF200/Sema3A immunofluorescence of DRG cultured for 3 days. (G) Quantitation of neurite length. (H) Quantification of fluorescent Sema3A. Yellow arrows indicate elongated neurite. Data are presented for mean ± SD. n = 4. ^**P < 0.01, ^*P < 0.05. ns represents no significant differences

The DRG intracellular Ca²⁺ activity quantified by fluorescent Ca²⁺ intensity revealed notably stronger in composite groups than pristine Gel, while Gel/SrZn-MSNs/G and Gel/P@SrZn-MSNs/G remarkably boosted the intracellular Ca²⁺ activity as compared to Gel/SrZn-MSNs (Fig. 2E). The intracellular cAMP levels of DRG significantly elevated in Gel/P@SrZn-MSNs/G than pristine Gel and Gel/SrZn-MSNs, while Gel/SrZn-MSNs/G revealed much stronger cAMP levels than other groups (Figure S4, Supplementary Information). The DRG neurite stained by NF200 immunofluorescence was detected much longer in composite groups than pristine Gel, and Gel/SrZn-MSNs/G further promoted the neurite elongation with respect to Gel/SrZn-MSNs (Fig. 2F, G). Additionally, the fluorescent Sema3A expressions significantly elevated in composite groups than pristine Gel (Fig. 2F, H).

It is widely known that sensory nerves are predominantly distributed in metabolically active areas to positively regulate osteogenesis, and the impaired or dysfunctional sensory nerves exactly occur in the initial stage of pathological bone diseases such as osteoporosis [12, 13, 37]. Thus, the bioeffects of implants on modulating sensory neuronal activities is a critical issue required to be investigated. To this end, DRG was isolated from OVX rats to collect sensory neurons, which were then directly seeded on coated Ti samples for mimicking in vivo cell-to-implant contact. Our results indicated each group of hydrogels well supported the cytoskeleton spreading and viability of DRG neurons probably due to the matrix properties of GelMA, which possessed proper elastic stiffness and also chemical signals such as RGD ligands to generate microenvironment beneficial for neurons developments [38]. On this basis, the incorporation of SrZn-MSNs into GelMA further promoted the survival and activation of DRG neurons, since MSNs has been applied as a delivery vehicle to effectively stimulate neuronal viability and neurite outgrowth [39]. On the other hand, divalent metal cations (e.g. Zn²⁺, Mg²⁺, Cu²⁺) have been reported able to induce bone formation by activating sensory nerves [40], thus the stimulatory effects of SrZn-MSNs on DRG neurons may be attributed to the bioactive Sr²⁺/Zn²⁺ as well.

In particular, present results highlight the principal roles of CGRP in composite hydrogels for activating sensory neurons, as indicated by the remarkable neurogenesis effects of Gel/SrZn-MSNs/G rather than Gel/P@SrZn-MSNs/G. It is acknowledged that sensory nerves positively regulate bone metabolism by secreting CGRP, which in turn can act a a neurotrophic signaling to support sensory functionality by autocrine and paracrine manners [11, 41]. Thus, increased CGRP expressions constantly keep company with sensory innervation and our result indicated CGRP rapidly releasing from composite hydrogels exerted potent bioeffects to benefit the spreading, viability and bioactivity of DRG neurons. In comparison, the controlled releasing of PTH from Gel/P@SrZn-MSNs/G appeared to somewhat compromise the activation of DRG neurons during 3 days culture, potentially by inhibiting the intracellular cAMP signal transduction (Figure S4, Supplementary Information). Actually, the exposure of PTH has been reported to suppress the intracellular cAMP signaling of DRG neurons by activating specific G-protein coupled receptors [42]. Regardless, the intracellular cAMP levels of Gel/P@SrZn-MSNs/G remained much higher than that of pristine Gel and Gel/SrZn-MSNs, suggesting CGRP rapidly released from Gel/P@SrZn-MSNs/G played predominate roles to potently activate DRG neurons. Therefore, Gel/SrZn-MSNs/G and Gel/P@SrZn-MSNs/G notably enhanced the bioactivity of DRG neurons as compared to pristine Gel and Gel/SrZn-MSNs, suggesting the efficient ions/drugs co-delivery in composite hydrogels for addressing sensory neurons activation.

Osteogenesis evaluation of BMSCs

As examined by a triple marked flow cytometry, the isolated BMSCs were characterized by highly specific CD29⁺/CD90⁺/CD45⁻ surface markers and thus could be identified as mesenchymal stem cells (Figure S5, Supplementary Information). As detected by phalloidin immunofluorescence, BMSCs fully spread and exhibited comparable spindle shaped cytoskeleton for 12 h culture on each group of coated Ti samples (Fig. 3A, B). For live/dead staining examinations, BMSCs vigorously grew with few dead cells observed after 1, 3 and 7 days culture on each sample (Fig. 3A), indicating the excellent biocompatibility of synthesized hydrogels. The quantitative MTT assays verified significantly enhanced viability of BMSCs cultured on Gel/SrZn-MSNs than Gel for 3 and 7 days, while Gel/SrZn-MSNs/G and Gel/P@SrZn-MSNs/G notably promoted the viability of BMSCs as compared to Gel and Gel/SrZn-MSNs on day 3 and 7 (Fig. 3C). The fluorescent osterix expressions were detected significantly higher in composite groups than pristine Gel, and Gel/P@SrZn-MSNs/G further facilitated osterix expressions to a greater level than Gel/SrZn-MSNs and Gel/SrZn-MSNs/G for 3 days culture (Fig. 3D, E). Meanwhile, the fluorescent OCN expressions distinctly elevated in Gel/SrZn-MSNs/G than Gel and remarkably increased in Gel/P@SrZn-MSNs/G than Gel and Gel/SrZn-MSNs for 3 days culture (Fig. 3D, F). Compared to pristine Gel, the ALP staining of BMSCs cultured for 7 days was observed obviously stronger in composite groups especially for Gel/P@SrZn-MSNs/G (Fig. 3G). Likewise, the RT-qPCR indicated Col I, ALP and Runx2 gene expressions were significantly up-regulated in composite groups than pristine Gel, and Gel/P@SrZn-MSNs/G promoted higher ALP and Runx2 gene expressions than Gel/SrZn-MSNs for 7 days culture (Fig. 3H, I). Additionally, OPN gene expressions were distinctly up-regulated in Gel/SrZn-MSNs/G and Gel/P@SrZn-MSNs/G than Gel, while Runx2 gene expressions also significantly increased in Gel/SrZn-MSNs/G than Gel/SrZn-MSNs and remarkably elevated in Gel/P@SrZn-MSNs/G than Gel/SrZn-MSNs/G (Fig. 3I).

Fig. 3.

Osteogenesis evaluation of BMSCs cultured on Ti samples coated by hydrogels. (A) The adhesion (first row) and viability (second to fourth row) of BMSCs. (B) Quantitation of cell spreading. (C) Quantitation of cell viability by MTT assays. (D) Osterix/OCN immunofluorescence of BMSCs cultured for 3 days. Quantitation of fluorescent (E) Osterix and (F) OCN. (G) ALP staining of BMSCs cultured for 7 days. Quantification of (H) Col Ⅰ, ALP and (I) OPN, Runx2 gene expressions by RT-qPCR for 7 days culture. Data are presented for mean ± SD. n = 4. ^**P < 0.01, ^*P < 0.05. ns represents no significant differences

By comparing Gel/SrZn-MSNs/P with Gel/P@SrZn-MSNs, an initial burst release was detected for unencapsulated PTH in Gel/SrZn-MSNs/P, while encapsulated PTH released form Gel/P@SrZn-MSNs in a controlled and sustained manner (Figure S6A, Supplementary Information). The ALP staining of BMSCs cultured for 7 days revealed obviously stronger in Gel/P@SrZn-MSNs than Gel/SrZn-MSNs/P (Figure S6B, Supplementary Information). Consistently, RT-qPCR results indicated Gel/P@SrZn-MSNs significantly up-regulated ALP gene expressions as compared to Gel/SrZn-MSNs/P (Figure S6C, Supplementary Information). Moreover, the fluorescent osterix (Figure S6D, E, Supplementary Information) and OCN (Figure S6F, G, Supplementary Information) expressions notably increased in Gel/P@SrZn-MSNs than Gel/SrZn-MSNs/P.

In accordance with culture of DRG neurons, BMSCs were harvested from femoral marrow cavities of OVX rats and then directly seeded on coated Ti samples for mimicking in vivo circumstances. The conducive adhesion and viability of BMSCs cultured on coated Ti samples further confirmed the excellent cellular affinity and biocompatibility of hydrogels to favor cellular interactions with Ti implants. Of note, the incorporation of SrZn-MSNs significantly enhanced the osteogenic ability of Gel/SrZn-MSNs than pristine Gel, while encapsulated PTH and CGRP in composite hydrogels further boosted the osteogenesis of BMSCs for prolonged culture. Considering each group of hydrogels were characterized by comparable topographic structures, the distinct outcomes observed in various groups were basically mediated by delivered components. Moreover, based on the durable degradation and stability of composite hydrogels, incorporated components exerted bioeffects in vitro probably by ions/drugs releasing. As previously reported, the combination of Si/Sr/Zn ions endowed bioactive glass or bioceramics with multifunctionality by proper releasing and thus strongly facilitated cellular developments and osteogenic differentiation of BMSCs for bone regeneration [35, 43]. Herein, the sophisticated incorporation of SrZn-MSNs into GelMA also realized a concurrent release of Si/Sr/Zn ions controlled within appropriate concentrations, thereby contributing to improved osteogenesis in composite hydrogels than pristine Gel. This observation largely coincides with our prior research [25] and hence further validates the desired bioactivity of SrZn-MSNs for promoting osteogenesis.

In addition to its own bioactivity, SrZn-MSNs are highly appreciated for its reliable drug delivery ability that achieved a sustained and controlled release of PTH due to the unique hollow mesoporous structure. As known, PTH is a bioactive peptide which positively participates in bone remodeling by exerting anabolic or catabolic effects, relying on the administration mode and dosage [44]. There is no doubt that intermittent administration of PTH is a quite effective therapy approved by FDA for treating osteoporosis, mainly by accelerating bone turnover to increase bone mass [45]. However, the inconvenient intermittent injection, high costs and potential complication risks (e.g. osteosarcoma, hypercalcemia) yet hinder the application of PTH especially for addressing osteoporotic osseointegration or bone defects repair [27, 46]. To realize broader application of PTH, scholars have developed specific delivery vehicles to conduct a sustained and controlled release of PTH for maintaining proper therapeutic concentrations at local sites, which efficiently avoids superphysiological doses caused by burst release and induces vigorous bone regeneration without systemic PTH exposure [47–49]. Likewise, in present study, the release kinetics of PTH encapsulated in Gel/P@SrZn-MSNs/G was properly controlled within proper concentrations rather than burst release, thus generating exciting anabolic effects to potently promote osteogenesis. For better understanding the superiority of encapsulated PTH delivery, PTH was directly incorporated into GelMA and then combined with SrZn-MSNs to develop Gel/SrZn-MSNs/P composite hydrogels. As compared to Gel/P@SrZn-MSNs, unencapsulated PTH directly released from Gel/SrZn-MSNs/P in an initial burst way, which probably resulted in relatively high-dose PTH exposure to BMSCs and thus greatly attenuated the anabolic effects of released PTH (Figure S6, Supplementary Information). These results indicate local delivery of PTH exerted pleiotropic effects on osteogenesis by a dose-dependent manner, which coincides with previous researches [47–49] and hence highlights the effective encapsulation of PTH in SrZn-MSNs for dictating osteogenesis.

On the other hand, the incorporation of CGRP into GelMA also notably enhanced the viability and osteogenic differentiation of BMSCs, which was consistent with previous studies that the direct and rapid release of CGRP from composite hydrogels was highly beneficial to support osteogenesis of BMSCs [16, 50]. Indeed, CGRP can bind to the transient receptor potential vanilloid 1 (TRPV1) receptor of BMSCs and then triggers the activation of Erk 1/2 pathway for mobilizing osteogenesis [51]. Overall, present results indicate the proper integration of SrZn-MSNs, CGRP and encapsulated PTH in composite hydrogels generated synergy bioeffects for inducing osteogenesis, thereby accounting for the superior bioactivity of Gel/P@SrZn-MSNs/G based on ions/drugs co-delivery.

Interactive neurogenesis/osteogenesis of DRG/BMSCs

For cellular interactions evaluation, the osteogenesis of BMSCs was affected by conditioned medium of DRG (DRG-CM) cultured in materials extracts via the interconnected microchannels in the microfluidic device, and vice versa for neurogenesis of DRG (Fig. 4A, B). After 3 days co-culture, the axons of DRG could be observed growing into microchannels as stimulated by conditioned medium of BMSCs (BMSCs-CM) cultured in materials extracts for each group, particularly increased DRG axonal outgrowth was stimulated by BMSCs-CM of Gel/SrZn-MSNs/G and Gel/P@SrZn-MSNs/G (Fig. 4C). Consequently, the axonal length of co-cultured DRG was significantly greater in Gel/SrZn-MSNs than Gel and remarkably larger in Gel/SrZn-MSNs/G and Gel/P@SrZn-MSNs/G than Gel and Gel/SrZn-MSNs (Fig. 4D). Likewise, the intracellular Ca²⁺ activity of co-cultured DRG revealed notably stronger in Gel/SrZn-MSNs/G and Gel/P@SrZn-MSNs/G than Gel and Gel/SrZn-MSNs (Fig. 4C, E). To investigate the neurotrophic factors secreted by BMSCs cultured for 3 days, the RT-qPCR analysis indicated Gel/P@SrZn-MSNs/G notably up-regulated the neurotrophic gene expressions of NGF and BDNF rather than neurotrophin-3 (NT3) and bFGF, as compared to pristine Gel (Figure S7, Supplementary Information).

For neuro-inductive osteogenesis evaluation, fluorescent osterix expressions of co-cultured BMSCs were significantly stronger in Gel/SrZn-MSNs/G and Gel/P@SrZn-MSNs/G than Gel and Gel/SrZn-MSNs, and Gel/P@SrZn-MSNs/G further promoted osterix expressions to a higher level than Gel/SrZn-MSNs/G for 3 days co-culture (Fig. 4F, G). As compared to pristine Gel, ALP staining of BMSCs co-cultured for 7 days was detected visibly stronger in composite groups especially for Gel/P@SrZn-MSNs/G (Fig. 4H). To investigate the key signaling molecule responsible for superior neuro-inductive osteogenesis of Gel/P@SrZn-MSNs/G, the related semaphorin family genes expressions of DRG cultured in materials extracts for 7 days were comprehensively analyzed by RT-qPCR. Consequently, Gel/P@SrZn-MSNs/G notably up-regulated the expressions of Sema3A, Sema4D and Sema7A genes compared to pristine Gel, while there was no significant difference for Sema3E gene expressions (Fig. 4I). Of note, ALP staining of BMSCs stimulated by DRG-CM of Gel/P@SrZn-MSNs/G for 7 days was obviously attenuated by the supplement of specific neutralizing antibodies especially for Sema3A (Fig. 4J). Furthermore, the ALP and Runx2 gene expressions of BMSCs stimulated by DRG-CM of Gel/P@SrZn-MSNs/G for 7 days were greatly blocked by the supplement of neutralizing antibodies against Sema3A rather than Sema4D or Sema7A (Fig. 4K, L).

As compared to pristine Gel, the western results further confirmed Gel/P@SrZn-MSNs/G significantly up-regulated neurotrophic factors (NGF/BDNF) levels in BMSCs (Fig. 5A, C) and remarkably up-regulated Sema3A protein expressions in DRG (Fig. 5B, D), which coincided with the RT-qPCR results shown in Figure S7 (Supplementary Information) and Fig. 4I respectively. For better understanding the potential signaling mechanism responsible for neuro-inductive osteogenesis of Gel/P@SrZn-MSNs/G, the SM345431 (Sema3A inhibitor) or XAV939 (Wnt/β-catenin inhibitor) was supplied to co-cultured BMSCs. The neuro-inductive signaling molecules expressions of BMSCs indicated Gel/P@SrZn-MSNs/G notably promoted the activation of Nrp1 (Sema3A receptor), p-Rac1 (phosphorylated Rac1) and osteogenesis-related proteins (Wnt6, β-catenin, ALP, Runx2) levels than pristine Gel for 7 days co-culture (Fig. 5E-H). In contrast, such superior neuro-inductive effects of Gel/P@SrZn-MSNs/G on co-cultured BMSCs were greatly attenuated by the supplement of either SM345431 or XAV939, resulting in dramatic reduction of Nrp1, p-Rac1 and osteogenesis-related proteins (Wnt6, β-catenin, ALP, Runx2) expressions (Fig. 5E-H).

Fig. 5.

Western blot analysis of interactive neurogenesis/osteogenesis orchestrated by Gel/P@SrZn-MSNs/G. (A) Neurotrophic factors expressions of BMSCs cultured in materials extracts for 3 days. (B) Semaphorin family proteins expressions of DRG cultured in materials extracts for 7 days. (C) Gray values quantification of neurotrophic factors secreted by BMSCs. (D) Gray values quantification of semaphorin family proteins secreted by DRG. (E) Neuro-inductive signaling molecules expressions of BMSCs supplemented with/without SM345431 or XAV939 for 7 days co-culture. Gray values quantification of (F) Nrp1, p-Rac1, (G) Wnt6, β-catenin, (H) ALP, Runx2 expressions of co-cultured BMSCs supplemented with/without SM345431 or XAV939. Data are presented for mean ± SD. n = 4. ^**P < 0.01, ^*P < 0.05

It is well-known that sensory nerves distribute throughout bony tissues and play essential roles for skeletal interoception and bone homeostasis, and in turn, skeleton provides ensemble of mechanical, chemical and electrical cues crucial for physiological function of sensory nerves [52]. Thereinto, DRG neurons can secrete neuropeptides and neurotransmitters to target bone lineage cells for regulating bone metabolism, and meanwhile bone lineage cells release bioactive signaling factors towards nerve fibers to dictate sensory innervation [11]. Of note, as an undifferentiated progenitor cell within bone, BMSCs exert pivotal effects for maintaining the bilateral nerve-bone homeostasis since they can not only differentiate into osteogenic-related cells but also potentially serve as important cell source for intrabony neurogenesis especially when nerve injury occurs [11, 53]. Hence, a normal crosstalk between DRG and BMSCs is critical to orchestrate a complex neuro-osteogenic microenvironment beneficial for mutual neurogenesis/osteogenesis, while an abnormal cellular crosstalk inevitably results in pathological bone disorders such as osteoporosis [51, 54]. However, this crosstalk is few investigated in bone biomaterials applied for osteoporotic osseointegration.

Herein, the nerve-bone crosstalk mediated by implants was probed by co-culture of DRG/BMSCs in a preferable microfluidic device. Given DRG and BMSCs were co-cultured without direct contact, the bilateral communication between DRG and BMSCs was basically executed by the soluble cytokines released into medium. As previously reported, BMSCs secreted abundant neurotrophic factors (e.g. NGF, BDNF, CNTF and bFGF) to trigger the endogenous MAPK and PI3K-Akt signaling activation in DRG neurons, thus greatly promoting the neuronal survival and neurite outgrowth in vitro [55]. Likewise, our results also revealed the indirect neurogenesis effects mediated by BMSCs cultured in materials extracts especially for Gel/P@SrZn-MSNs/G, which remarkably up-regulated the expression of NGF and BDNF in BMSCs to stimulate the axonal outgrowth and activation of DRG neurons. On the other hand, the osteogenic differentiation of BMSCs induced by co-cultured DRG revealed much stronger in Gel/P@SrZn-MSNs/G than other groups, highlighting the superior neuro-inductive osteogenesis effects orchestrated by ions/drugs co-delivery. These results indicate Gel/P@SrZn-MSNs/G not only promoted functionality of DRG sensory neurons and BMSCs directly, but also manipulated mutually beneficial interactions of DRG/BMSCs to further stimulate neurogenesis/osteogenesis. In particular, the superior neuro-inductive osteogenesis effects of Gel/P@SrZn-MSNs/G could be selectively blocked by supplying specific neutralizing antibodies against Sema3A. As shown, the up-regulation/withdrawl of Sema3A secreted from DRG cultured in Gel/P@SrZn-MSNs/G could be responsible for the enhanced/suppressed osteogenesis of BMSCs, thus highlighting the key roles of Sema3A to positively regulate the neuro-inductive osteogenesis.

Indeed, the semaphorin family molecules (e.g. Sema3A, Sema3E, Sema4D, Sema7A) are acting as vital neurotransmitters to guide intrabony innervation and also control bone homeostasis, wherein DRG-derived Sema3A can directly bind to the Neuropilin 1 (Nrp1) receptor and then specifically triggers the activation of Rac1, leading to the up-regulation of Wnt/β-catenin pathway and downstream osteogenic differentiation of BMSCs [51, 56]. Moreover, the activated Wnt/β-catenin pathway can generate a positive feedback loop to further promote the Nrp1 receptor expression, which is critical for exogenous Sema3A to target BMSCs [57]. Therefore, in addition to SM345431 (Sema3A inhibitor), the supplement of XAV939 (Wnt/β-catenin inhibitor) also significantly inhibited the signal transduction of Sema3A by down-regulating the Nrp1 protein expressions of BMSCs, resulting in decreased activation of Rac1 and osteogenic proteins expressions. These results suggest both DRG-derived Sema3A and intracellular Wnt/β-catenin signaling played key roles for osteogenic differentiation of co-cultured BMSCs. In this view, the Sema3A-dependent Wnt/β-catenin signaling axis could account for the superior neuro-inductive osteogenesis effects of Gel/P@SrZn-MSNs/G. Hence, present results further support the considerable effects of Gel/P@SrZn-MSNs/G to mediate close crosstalk between DRG and BMSCs, which may orchestrate desired neuro-osteogenic microenvironment for ameliorating osteoporotic osseointegration.

Osteoporotic osseointegration evaluation of implants

Compared to Sham rats, the BV/TV and Tb.N dramatically declined in OVX rats subjected to ovariectomy, accompanied by significantly reduced CGRP positively stained sensory nerves (Figure S8, Supplementary Information). Following 6 weeks post-implantation, the specimens were harvested without obvious fracture, infection or cellulitis. As evaluated by micro-CT, specimens were 3D reconstructed to detect Ti implants (green) and peri-implant bone formation (red) from sagittal, median transverse and coronal scopes (Fig. 6B). The quantitative BV/TV was significantly higher in composite groups than pristine Gel, and Gel/P@SrZn-MSNs/G further increased BV/TV to a greater level than Gel/SrZn-MSNs and Gel/SrZn-MSNs/G (Fig. 6D). Moreover, quantitative BMD revealed notably higher in Gel/P@SrZn-MSNs/G than the other groups (Fig. 6E), and quantitative Tb.Th also significantly elevated in Gel/P@SrZn-MSNs/G than Gel and Gel/SrZn-MSNs (Fig. 6F).

Consistently, the implants osseointegration (red arrows) exhibited in undecalcified H&E sections were visibly stronger in Gel/P@SrZn-MSNs/G than the other groups (Fig. 6C). For quantitative analysis, the bone-implant contact (BIC) ratios significantly increased in composite groups than pristine Gel, and Gel/P@SrZn-MSNs/G possessed higher BIC ratios than Gel/SrZn-MSNs and Gel/SrZn-MSNs/G (Fig. 6G). In addition, the pull-out force of Ti implants was significantly larger in Gel/SrZn-MSNs and Gel/SrZn-MSNs/G than pristine Gel, and Gel/P@SrZn-MSNs/G further enhanced the pull-out force of Ti implants to a higher level than the other groups (Figure S9, Supplementary Information).

The median transverse section of implants labeled in Fig. 7A was decalcified and subjected to H&E and immunofluorescence staining. As revealed in Fig. 7B, peri-implant bone formation (red arrows) was observed in each group especially for Gel/P@SrZn-MSNs/G, accompanied by ingrowth of CGRP⁺ sensory nerves (yellow arrows), CD31⁺ vascularization (yellow arrows) and Col I⁺/Sema3A⁺ cytokines expressions (yellow arrows). The fluorescent CGRP expressions significantly increased in composite groups than pristine Gel, while Gel/SrZn-MSNs/G and Gel/P@SrZn-MSNs/G notably promoted CGRP expressions than Gel/SrZn-MSNs (Fig. 7C). Likewise, the fluorescent CD31 expressions significantly elevated in Gel/SrZn-MSNs/G than Gel and remarkably increased in Gel/P@SrZn-MSNs/G than Gel and Gel/SrZn-MSNs (Fig. 7D). Besides, the fluorescent Col I/Sema3A expressions significantly increased in Gel/P@SrZn-MSNs/G than Gel (Fig. 7E, F).

Fig. 7.

Histological analysis of implants osseointegration at 6 weeks post-surgery. (A) The median transverse section of implants labeled by micro-CT. (B) Median transverse sections of implants stained by H&E (first row), CGRP (second and third row), CD31 (fourth and fifth row) and Col Ⅰ/Sema3A (sixth to eighth row) immunofluorescence. Quantitation of fluorescent (C) CGRP, (D) CD31, (E) Col Ⅰ, (F) Sema3A. Dashed lines denote bone-implant interface. Red arrows indicate osseointegration around implants. Yellow arrows indicate positive immunofluorescence staining. Data are presented for mean ± SD. n = 4. ^**P < 0.01, ^*P < 0.05

To determine the therapeutic potentials of hydrogels for osteoporotic osseointegration, femoral condyle defects were developed in OVX rats and then implanted by Ti screws surrounded by injectable hydrogels. Ti screws were preferably applied as fixation implants in present study for their frequent usage in orthopedic surgery, mainly depending on the excellent biomechanical properties, chemical stability and biocompatibility [58]. Nonetheless, the insufficient surface bioactivity and biological inertness of Ti implants remain a big issue to restrict osseointegration especially under osteoporotic conditions. To address this issue, the nanocomposite hydrogels were injected and crosslinked in situ based on the superior photosensitivity, aiming to fill the gap between implanted Ti and host bone. Thereafter, the durable degradation of composite hydrogels rendered bioactive ions/drugs continuously releasing to exert cumulative bioeffects around Ti implants. Consequently, the application of composite hydrogels induced stronger bonds of Ti screws than pristine Gel, probably by restoring the neuro-osteogenic niches at bone-implant interfaces. As reported, the pathological microenvironment of osteoprosis is tightly associated with impaired sensory innervation in addition to hypoactive osteogenic activities [11, 12]. Our results yet revealed the deteriorated bone microarchitecture accompanied by obviously reduced CGRP⁺ sensory nerves in OVX rats.

Under this circumstance, Gel/P@SrZn-MSNs/G efficiently induced functional osteoporotic osseointegration probably by orchestrating desired neuro-osteogenic niches, as indicated by the simultaneously reinforced peri-implant innervation, vascularization and bone formation. The in vivo outcomes exactly coincide with the in vitro observations, which further validates the effective ions/drugs co-delivery in Gel/P@SrZn-MSNs/G for inducing potent and timely innervation and bone healing. Actually, previous research has reported silicified collagen scaffolds released bioactive Si to actively stimulate the Sema3A expression in sensory nerves for improving bone regeneration [59]. Furthermore, MSNs-based co-delivery of therapeutic biomolecules/drugs in injectable hydrogels targets the Nemo-like kinase signaling in BMSCs to induce innervated bone regeneration [60]. Herein, based on the sophisticated integration of SrZn-MSNs, CGRP and encapsulated PTH in composite hydrogels, Gel/P@SrZn-MSNs/G exerted synergy bioeffects by concurrent releasing of bioactive ions/drugs controlled within proper doses. Overall, the efficient ions/drugs co-delivery system endowed Gel/P@SrZn-MSNs/G nanocomposite hydrogels with superior biological multifunctionality to orchestrate desired neuro-osteogenic niches for boosting osteoporotic osseointegration, thus may inspiring new therapeutic strategies for osteoporotic bone healing.

However, there remains one major issue to investigate the in-depth molecular mechanism responsible for the neurogenesis/osteogenesis effects of composite hydrogels. In addition, the antibacterial ability of composite hydrogels based on Zn ions releasing is worth exploring especially for infected bone repair [61].

Conclusions

In present study, novel Gel/P@SrZn-MSNs/G nanocomposite hydrogels were fabricated by simultaneously incorporating PTH loaded SrZn-MSNs (P@SrZn-MSNs) and CGRP into GelMA, followed by photo-crosslinking with UV irradiation. Based on the proper integration of each component, Gel/P@SrZn-MSNs/G possessed desired physico-chemical properties including excellent injectability and photosensitivity, 3D interconnected porous architecture, durable degradation and concurrent releasing of Si/Sr/Zn ions, CGRP and PTH controlled within proper doses. These favorable properties generated synergy bioeffects to greatly promote the neurogenesis/osteogenesis of DRG/BMSCs directly cultured on Gel/P@SrZn-MSNs/G. The microfluidics assays further exhibited the superior bioactivity of Gel/P@SrZn-MSNs/G to mediate close crosstalk between DRG and BMSCs, wherein increased NGF and BDNF expressions in BMSCs stimulated sensory innervation and the up-regulation of Sema3A in DRG neurons could be responsible for the neuro-inductive osteogenesis. Following implantation into femoral condyle defects of OVX rats, Gel/P@SrZn-MSNs/G efficiently induced functional osteoporotic osseointegration by stimulating potent and timely sensory innervation and bone formation around Ti screws. In this view, the efficient ions/drugs co-delivery system endowed Gel/P@SrZn-MSNs/G nanocomposite hydrogels with superior biological multifunctionality to orchestrate desired neuro-osteogenic niches for boosting osteoporotic osseointegration, thus may inspiring new therapeutic strategies for osteoporotic bone healing.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

JSY: Writing-original draft, Conceptualization, Data curation, Methodology, Funding acquisition. UI: Methodology, Conceptualization, Investigation. CZW: Writing-original draft, Data curation. GYW: Methodology, Conceptualization, Investigation. FT: Data curation, Visualization. YY: Methodology, Visualization. YQ: Investigation, Data curation. WJN: Data curation, Investigation. SDW: Conceptualization, Methodology. DZC: Methodology, Visualization. HZF: Funding acquisition, Project administration. CW: Funding acquisition, Project administration. SH: Conceptualization, Project administration. NJJ: Funding acquisition, Writing-review & editing, Project administration. YL: Project administration, Conceptualization, Writing-review & editing.

Funding

This study was supported by the National Natural Science Foundation of China (82302680, 82472426, 82202697), the Technology Innovation on Medicine and Health of Suzhou Science and Technology Bureau (SKY2023147 ), the Suzhou Basic Research Pilot Program (SSD2024048) and Entrepreneurship and Innovation Doctoral Program of Jiangsu Province (JSSCBS20230503).

Data availability

All data used or analyzed in current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

All animal experiments were approved by the Institutional Animal Care and Use Committee of Suzhou Jingzida Testing Service Co., Ltd. (IACUC-2023-0006-DI).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.