Boosting ribosomal translation via ionizable lipid-hydrogel microplexes for localized mRNA therapy

Abstract

mRNA therapy holds immense promise for regenerative medicine; however, localized endoplasmic reticulum stress (ERS) in damaged tissues can impair the critical process of ribosomal translation. Here, we developed an in situ injectable lipid nanoparticle (LNP)/microsphere complex, also referred to as a lipid-hydrogel microplex (iLMP), with ERS-alleviating functionality to increase ribosomal translation. A vitamin E-derived ionizable lipid was synthesized to replace conventional ionizable lipids in LNPs, whereas porous hydrogel microspheres stabilized the LNPs via physical adsorption. In vitro studies revealed that the iLMPs codelivered vitamin E and mRNA, mitigating ERS and reducing eIF2α phosphorylation, a key translational barrier. Additionally, iLMPs injected in situ rapidly reconstructed the extracellular matrix, promoting tissue repair. In a bone defect animal model, iLMPs significantly enhanced BMP-2 mRNA translation, promoting osteogenesis. In summary, we present a novel in situ injectable mRNA delivery platform that enhances ribosomal translation, offering a promising strategy for tissue regeneration.

Graphical abstract

Currently, no studies have focused on how ERS-induced ribosomal translation limitations may affect the efficacy of mRNA therapeutics. In this work, we address this challenge by introducing a novel ERS-alleviating ionizable lipid and designing a specialized lipid-hydrogel microplex system that integrates LNPs with microspheres. This composite system is specifically engineered to overcome the limitations of mRNA transfection in inflammatory microenvironments.

Highlights

• •Innovative iVES Design: iVES serves as the core component of iLMP, not only alleviating ERS but also enabling mRNA condensation and endosomal escape.
• •iLMP Construction: iLMP contributes to the storage stability of mRNA therapeutics, further enhancing their efficacy.
• •Exploration of ERS on mRNA Therapeutics: This work is the first to focus on the impact of ERS during the in-situ application of mRNA.

1. Introduction

Currently, mRNA therapy is a disruptive technology in medical fields, including regenerative medicine. Once exogenous mRNA is delivered into cells, it can directly initiate protein translation in the cytoplasm without nuclear import or reverse transcription, thereby enabling the regulation of cell fate [1]. However, during inflammatory tissue repair delay, cells usually experience an increase in endoplasmic reticulum stress (ERS) and actively initiate the unfolded protein response (UPR), which can significantly reduce mRNA translation in ribosomes [2,3]. For example, in steroid-associated necrosis of the femoral head (SANFH), ischemia–hypoxia, oxidative and inflammatory stress, and lipid dysregulation converge to amplify ERS/UPR, suppressing translation and necessitating ERS alleviation to restore osteogenesis and mRNA therapy efficacy. Unfortunately, current mRNA therapies do not fully account for this limitation on ERS-related mRNA translation. Therefore, it is necessary to develop an ERS alleviation strategy to enhance mRNA therapy by accelerating ribosomal translation and global protein synthesis.

Normally, the UPR, including the downregulation of mRNA translation, is a protective mechanism by which cells respond to pathological increases in ERS [4,5]. During tissue regeneration, cells face various external stresses, such as inflammatory and hypoxic microenvironments, which can disrupt intracellular protein homeostasis. The accumulation of unfolded or misfolded proteins in the endoplasmic reticulum (ER), the primary site for protein folding and maturation, further exacerbates ERS [6,7]. Under ERS conditions, cells can initiate the UPR to reduce the ER load by many mechanisms. Among them, downregulating mRNA translation is an active strategy to decrease new protein synthesis and reduce the accumulation of unfolded or misfolded proteins. Additionally, by downregulating mRNA translation, cells can mitigate ERS-related damage and apoptosis risk, thereby increasing their survival rate under stressful conditions [8]. Although certain stress response proteins (such as ATF4 and XBP1) that help restore ER homeostasis are selectively upregulated, the global downregulation of mRNA translation in cells is indisputable. Multiple UPR-activated signaling pathways are involved in the process of mRNA translation downregulation, with the protein kinase R-like ER kinase (PERK) pathway playing the most critical role. Here, PERK could be considered a key sensor of the UPR. When ERS is activated, PERK undergoes autophosphorylation, activating its kinase activity. Activated PERK then phosphorylates the Ser51 site of eukaryotic translation initiation factor 2α (eIF2α). Finally, the phosphorylation of eIF2α directly inhibits whole-cell mRNA translation. Additionally, the inositol-requiring enzyme 1 (IRE1) and activating transcription factor 6 (ATF6) pathways also participate in related mechanisms, but they are not directly associated with mRNA translation levels [9]. Therefore, to increase the global mRNA translation level, alleviating ERS at the source and reducing the degree of cellular reliance on the protective UPR mechanism are essential.

Lipid nanoparticle (LNP)-based gene transfection tools have enabled the translation of meaningful mRNA therapies from the laboratory to clinical applications [10,11]. Currently, typical LNP formulations include ionizable lipids, PEGylated lipids, cholesterol, and helper lipids. Among them, ionizable lipids are the most critical component, enabling stable mRNA loading and efficient lysosomal escape through pH-dependent reversible protonation, which significantly affects the mRNA delivery efficiency of LNPs [12]. Currently, developing novel ionizable lipid-based LNP systems is a key approach to expanding mRNA therapies. For example, Yang et al. constructed ionizable lipids through lipid-modified guanidine-based cationic polydisulfides for LNP formulation to improve the mRNA transfection efficiency in dendritic cells by endowing them with an inflammation-related reactive oxygen species (ROS) scavenging ability [12]. Additionally, Zhang et al. designed a single-component ionizable cationic lipid enriched with secondary amines for selective mRNA delivery to the spleen and T cells, addressing the challenge of extrahepatic delivery faced by traditional LNPs [13]. Therefore, the design of ionizable lipids can further endow mRNA LNP systems with additional capacities.

In previous studies, we directly replaced traditional phospholipid and lipid alkyl chains with various lipophilic drug molecules to construct amphiphilic conjugates as formulation components, thereby creating functionalized liposomes and vesicles [[14], [15], [16], [17]]. For example, we reported a series of artemether-phospholipid and cationic lipid conjugates that maintained the drug activity of artemether while fully replacing traditional phospholipids in the construction of liposomal drug/gene delivery systems [18,19]. These data demonstrate that amphiphilic conjugates possess physicochemical properties similar to those of traditional phospholipids and lipids and can completely replace them in structural assembly. Vitamin E, a natural fat-soluble active molecule, has been shown to alleviate ERS and activate translation-related pathways across different cell types. Yao et al. reported that vitamin E alleviated ERS and promoted fatty acid synthesis in grass carp muscle cells [20]. Likewise, Shi et al. reported that vitamin E alleviated oxidative stress-induced hepatocellular injury by inhibiting ERS [21]. Although some mRNA-LNP vaccine studies reported the adjuvant activity of vitamin E derivatives, specifically, vitamin E does not directly activate the innate immune system but plays a crucial regulatory role in supporting and enhancing innate immunity through its antioxidant capabilitie [20,21]. Therefore, introducing vitamin E-based ionizable lipids into LNP formulations is a feasible approach to endow LNPs with an ERS-relieving ability to increase the level of translation for mRNA therapy.

Injectable hydrogel microspheres, renowned for their high biocompatibility and ability to remodel the microenvironment, have found widespread applications in tissue repair. In our previous studies, we explored the application of hydrogel microspheres in gene delivery [10,22,23]. In this work, we synthesized a novel vitamin E-derived ionizable lipid (iVES) based on the structure of classical Dlin-MC3-DMA, which was employed to construct lipid-hydrogel microcomplexes (iLMPs) with the ability of mRNA deliver and ERS alleviating. iLMPs could enhance ribosomal translation according to previous ERS mechanism studies, thereby improving the transfection of the mRNA cargos (Scheme 1) [24]. First, iVES was synthesized by an esterification reaction as a dimer with an ionizable tertiary amine structure, which has physicochemical properties similar to the classic Dlin-MC3-DMA ionizable lipids, including similar acid dissociation constant (pKa) values. Then, using traditional LNP and hydrogel microsphere microfluidic systems, iVES-based mRNA LNPs (iLNPs) and iLMPs were sequentially prepared. Physical adsorption enables the sustainable release of LNPs after in situ injection. In an in vitro ERS-positive cell model, iLMPs significantly reduced PERK pathway activation and decreased eIF2α phosphorylation (p-eIF2α). The improved ribosomal translation of ERS-positive cells was clearly observed via a O-Propargyl-Puromycin (OPP) click chemistry method. Finally, we applied a SANFH animal model, a condition characterized by ischemia–hypoxia, oxidative and inflammatory stress, and sustained ERS/UPR activation, to verify the efficacy of iLMPs by observing the significantly relieved SANFH symptoms, and the improved BMP-2 mRNA translation. In conclusion, injectable iLMPs could serve as an enhanced mRNA therapeutic platform for tissue regeneration by synergistically improving cellular mRNA translation.

Scheme 1.

Schematic illustration of how ERS-attenuating iLMPs enhance exogenous mRNA translation for in situ tissue regeneration. (A) Chemical structure and ERS alleviation efficacy of iVES. (B) General schematic illustration of iLNP fabrication through microfluidic mixing with water and the organic phase. (C, D) Microfluidic fabrication of in situ injectable iLMPs. (E) ERS attenuation and translation improvement mechanisms of BMP2@iLMPs.

2. Materials and methods

2.1. Preparation of iVES

Vitamin E succinate (2.65 g, 5.0 mmol) was dissolved in 30 mL of dichloromethane and activated by slowly adding DCC (1.13 g, 5.5 mmol). 3-Dimethylamino-1,2-propanediol (0.24 g, 2.0 mmol) and DMAP (0.25, 2.0 mmol) were subsequently added, and the mixture was stirred at room temperature for 24 h. Finally, the mixture was separated by column chromatography to obtain colorless, oily iVES. ¹H NMR (300 MHz, CDCl3) showed the following results: δ 5.30–5.20 (m, 1H); 4.47–4.12 (m, 2H); 2.91 (td, J = 6.6, 2.0 Hz, 4H); 2.84–2.73 (m, 4H); 2.59 (q, J = 8.1, 7.4 Hz, 4H); 2.54–2.38 (m, 2H); 2.26 (s, 6H); 2.10 (s, 6H); 2.03 (s, 6H); 1.99 (s, 6H); 1.79 (q, J = 7.1 Hz, 4H); 1.55 (td, J = 7.6, 6.5, 5.1 Hz, 6H); 1.47–1.33 (m, 10H); 1.30–1.04 (m, 32H); and 0.90–0.85 (m, 24H). ¹³C NMR (125 MHz, CDCl3) showed the following results: δ 171.97, 171.70, 170.85, 149.44, 140.49, 126.70, 124.95, 123.01, 117.37, 75.05, 69.91, 64.36, 59.20, 46.01, 39.39, 37.47, 37.44, 37.30, 32.80, 32.72, 31.11, 29.21, 28.95, 28.75, 27.98, 24.80, 24.45, 22.71, 22.62, 21.03, 20.59, 19.75, 19.65, 12.93, 12.09, 12.07, and 11.79.

2.2. Preparation of iLNPs

iLNPs were prepared with a microfluidic device. Fifty micrograms of BMP-2 mRNA was dissolved in 450 μL of citrate buffer (20 mM, pH = 4) to form an aqueous-phase solution. Then, iVES, DSPC, cholesterol, and DSPE-PEG-RGD were weighed according to a fixed molar ratio and dissolved in 150 μL of anhydrous ethanol to form an ethanol-phase solution. The ethanol and aqueous phases were mixed at a 1:3 vol ratio using a microfluidic apparatus (Suzhou Unigen Biotech Co., China). The resulting mixture was transferred to an ultracentrifuge tube containing 10 times the volume of PBS (pH = 7.4) and centrifuged to remove the ethanol phase, yielding prepared iLNPs.

2.3. Characterization of iLNPs

To observe the characteristics of the iLNPs, the particle size and zeta potential of the iLNPs were first examined by dynamic light scattering (DLS) (Zetasizer Nano S, UK). iLNPs were collected and diluted with PBS, and the diluted iLNP suspension was placed in a quartz cuvette to observe the particle size distribution of the iLNPs. The encapsulation rate of the iLNPs was subsequently measured with an RNA Quant quantification kit (Nanjing Vazyme Biotech Co., Ltd., China). The morphology of the iLNPs was observed by transmission electron microscopy (TEM) (Hitachi HT 7800, Japan). Finally, the stability of iLNPs in PBS at pH 7.4 and 4 °C environment was tested. The prepared iLNPs were placed on a shaker at 4 °C, and the size of the iLNPs was measured.

2.4. Preparation of iLMPs

iLMPs were prepared by HAMA-based microsphere and iLNPs. Firstly, HAMA was synthesized. Briefly, 5 g of HA (molecular weight ≈74 kDa, Bloomage Freda Biopharmaceutical Co., Ltd., China) was reacted with 10.4 g of methacrylic anhydride, dialyzed for 4 days, lyophilized and stored at −80 °C. Then, hydrogel microspheres were prepared by a microfluidic technique. The aqueous phase consisted of a hydrogel mixture (4 wt% HAMA and 0.4 wt% photoinitiator), and the oil phase consisted of paraffin oil mixed with 5 wt% Span 80. These phases were injected into the inlet of the microfluidic device from a syringe, and the flow rate ratio of the aqueous phase to the oil phase was adjusted. The generated droplets were frozen at −30 °C and crosslinked under UV light to obtain HAMA, after which the excess paraffin oil was removed and the product was lyophilized at −80 °C.

Then, iLNPs were diluted in deionized water to a final mRNA concentration of 2 μg/mL, after which lyophilized microspheres were dispersed in deionized water to a final concentration of 0.1 mg/mL. The suspension was gently rotated at 30 rpm and incubated at 4 °C for 15 min, allowing iLNPs to adsorb onto the surface and pores of the microspheres via physical interactions. Unloaded iLNPs were removed by centrifugation at 400g for 3 min, followed by a single gentle wash with deionized water. The resulting iLMPs were lyophilized and stored at −80 °C. Here, BMP2@iLNP loaded iLMP was named as B/iLMP.

2.5. Measurement of swelling ratio

The iLMPs samples were thoroughly dried, and their initial dry mass (W_d) was recorded. The dried microspheres were then immersed in PBS and placed in a thermostatic water bath. The samples were retrieved at predetermined time points (1, 2, and 3 h), gently blotted with filter paper to remove surface moisture, and promptly weighed to obtain the swollen mass (W_s). Accordingly, the swelling ratio (%) was calculated using the following formula:

$$\text{Swelling}\text{ratio}(\%)=\left({\mathrm{W}}_{\mathrm{s}}-{\mathrm{W}}_{\mathrm{d}}\right)/{\mathrm{W}}_{\mathrm{s}}\times 100\%$$

where W_d denotes the initial dry weight of the sample, and W_s represents the weight of the sample after swelling. All measurements were performed in triplicate (n = 3).

2.6. Transfection efficiency of iLMPs

We next aimed to clarify the transfection efficiency of iLMPs in osteoblast precursor (MC3T3-E1) cells. We chose mCherry mRNA as a model, prepared mRNA-iLMPs, coincubated them with MC3T3-E1 cells, and observed the transfection efficiency by fluorescence microscopy (Nikon ECLIPSE Ts2R, Japan). Briefly, 48 h prior to transfection, MC3T3-E1 cells were inoculated in 24-well plates at a density of approximately 1 × 10⁵ cells/well. After 24 h of culture to allow adhesion, the cells reached 80 % confluence and were transfected. One microgram of mCherry mRNA iLMPs was added to the corresponding wells, and MC3T3-E1-transfected cells were observed under a fluorescence microscope after 24, 48 and 72 h of incubation.

2.7. In vitro stability assessment of iLNPs and iLMPs

CLSM was used for in vitro stability assessment of iLNPs and iLMPs, where mCherry mRNA was selected model mRNA for detectability of LNP transfection under CLSM. Briefly, MC3T3-E1 cells were seeded into 24-well plates and cultured until they adhered and reached 80 % confluence. Subsequently, mCherry-loaded iLMPs lyophilized powder kept under −80 °C for 12 h and 1 week were co-incubated with MC3T3-E1 cells for 48 h with concentration of 10 μg/mL. Freshly prepared mCherry-loaded iLMPs were used as a control. The mCherry expression was detected under CLSM.

Additionally, mCherry-loaded iLNPs and iLMPs were suspended in PBS and stored at 4 °C for different period, then added to MC3T3-E1 cells for co-incubation with LNP equivalent concentration of 0.1 μg/mL. Intracellular fluorescence intensity was measured to assess the effect of microsphere encapsulation on the 4 °C storage stability of iLNPs.

2.8. Characterization of B/iLMPs

The particle size and degradation of B/iLMPs were measured by an inverted microscope (Nikon, Japan). A total of 10 mg of B/iLMPs was placed in PBS containing 0.2 U/mL collagenase at pH = 7.4 and then placed on a shaker at 37 °C; the solution was replaced with PBS solution containing fresh hyaluronidase every two days. Changes in the morphology of the hydrogel microspheres were observed, and the weight of the remaining microspheres was measured at specific time points and compared with the initial weight to calculate the percentage of microsphere degradation. Three milligrams of lyophilized B/iLMPs was combined with 1 mL of PBS, and the weight of the B/iLMPs and the Eppendorf tube were measured. Measurements were taken every 30 min for 3 h, and the samples were drained before each weighing. Finally, the morphology of the freeze-dried microspheres was observed by scanning electron microscopy (ZEISS, Germany).

2.9. Loading and release analysis of iLNPs from iLMPs

To clarify the loading and release of iLNPs from iLMPs, a confocal microscope and a multifunctional enzyme marker were used, respectively. To facilitate the observation of iLNPs, we introduced a trace amount of DiL dye into the iLNP preparation and observed their distribution by confocal microscopy. First, the iLNPs were stained with DiL (Beyotime Co., Ltd., China), and then the lyophilized HAMA microspheres were placed in a suspension of iLNPs, which adsorbed onto the lyophilized HAMA microspheres by physical electrostatic/hydrogen bonding. The absorbance values of free iLNPs in the supernatant were measured after microsphere precipitation, and the loading rate of iLNPs onto iLMPs was determined by standard curve calculation. The adsorption of iLNPs was observed by confocal microscopy. Next, iLMPs were placed in PBS at pH = 7.4 on a shaker at 37 °C. The absorbance values of iLNPs in the supernatant at different time points were measured using a multifunctional enzyme marker, and the release of the iLNPs was calculated using standard curves.

2.10. In vitro biocompatibility evaluation

To clarify the biosafety of the iLMPs, we used cell counting kit-8 (CCK-8) and calcein acetoxymethyl/propidium iodide (calcein AM/PI) cell viability/cytotoxicity assay kits. 5 groups were established: control group, MS group, LMP group, B/LMP group, and B/iLMP group. Specifically, the MS group consists of HAMA hydrogel microspheres alone; the LMP group consists of microspheres combined with unloaded LNPs; the B/LMP group consists of microspheres loaded with LNPs encapsulating BMP-2 mRNA; and the B/iLMP group consists of microspheres combined with iLNPs encapsulating BMP-2 mRNA. The LNPs were prepared based on Dlin-MC3-DMA, while the iLNPs were prepared based on iVES. MC3T3-E1 cells were inoculated in 24-well plates and cocultured with the corresponding extracts at 37 °C and 5 % CO₂. At 24 h and 72 h, the cells were incubated with calcein AM/PI buffer (Beyotime Co., Ltd., China) for 15 min, followed by fluorescence microscopy. MC3T3-E1 cells were inoculated in 96-well plates and cocultured with the corresponding extracts at 37 °C and 5 % CO₂. After 24 h and 72 h of incubation, MC3T3-E1 cells were incubated in medium supplemented with 10 % CCK-8 for 1 h, after which the absorbance at 450 nm was measured with an enzyme marker (Molecular Devices, Japan).

2.11. Hemolysis assay

The in vitro safety was further evaluated by a hemolysis assay. Briefly, whole blood was collected from SD rats into heparinized tubes and centrifuged at 1500 g for 10 min at 4 °C to remove plasma and the buffy coat, retaining the erythrocytes. The cells were washed three times with sterile PBS and resuspended to obtain a 2 % (v/v) erythrocyte suspension. Four experimental groups were established: PBS group (erythrocytes incubated with PBS, negative control), Triton group (erythrocytes incubated with 1 % Triton X-100 in PBS, positive control), B/LMP group (erythrocytes incubated with B/LMPs), and B/iLMP group (erythrocytes incubated with B/iLMPs). For each group, 0.5 mL of the erythrocyte suspension was mixed with 0.5 mL of the respective test or control solution and incubated at 37 °C for 1 h with gentle inversion every 10 min, followed by centrifugation at 1500 g for 10 min at 4 °C. The absorbance of the supernatant was measured at 541 nm to assess hemolysis.

2.12. In vivo assessment of systemic leakage of iLMPs

Under isoflurane anesthesia, 8-week-old SD rats were prepared by making an incision in the skin and performing blunt dissection of the surrounding muscles to expose the greater trochanter and hip joint. A drill was used to create an entry point at the greater trochanter, extending into the femoral neck to establish an intramedullary channel. Through this channel, 10 μL of Cy5.5-loaded iLMP solution was injected into the marrow cavity. Seventy-two hours after the surgery, rats were euthanized under isoflurane anesthesia, and the femur along with key organs were collected. In vivo bioluminescence imaging was performed to assess potential systemic leakage.

2.13. Wound-healing assay

MC3T3-E1 cells in the logarithmic growth phase were seeded into 6-well plates (2x10⁶ cells/well). After 70 % confluence was reached in each group, the control, MS, LMP, B/LMP and B/iLMP materials were added. After 20 h of treatment, a 200 μL pipette tip was used to make a scratch in the cell monolayer. The wells were washed three times with PBS to remove floating cells, and 2 mL of serum-free culture medium was added to each group. The plates were subsequently placed in an incubator at 37 °C with 5 % CO₂. After 24 h, cell migration was observed under an inverted light microscope, and the wells were photographed to measure the scratch width using ImageJ.

2.14. In vitro evaluation of the ability of B/iLMPs to scavenge ROS and relieve ERS

The effects of B/iLMPs on ERS in MC3T3-E1 cells were analyzed by fluorescence. MC3T3-E1 cells were inoculated in confocal Petri dishes. The control, model, MS, LMP, B/LMP and B/iLMP groups were used in this experiment. All groups except the control group were exposed to 400 μmol/L H₂O₂ for 24 h. After 24 h, the control and model groups were not treated, and the remaining groups were incubated with the corresponding extracts at 37 °C. The cells were washed with PBS three times, and DCFH-DA reagent and OPP reagent were added to detect the ROS level and ERS level in MC3T3-E1 cells, respectively.

2.15. Protein synthesis assay

O-propargyl-puromycin (OPP, from ClickChemistryTools Co.) was used to evaluate nascent protein synthesis. MC3T3-E1 cells were incubated with MS, LMPs, B/LMPs, or B/iLMPs for 24, 48, and 72 h. Following treatment, the cells were exposed to 20 μM OPP working solution for 30 min at 37 °C to allow incorporation into newly synthesized polypeptides. After incubation, the cells were fixed with 4 % paraformaldehyde and permeabilized with 0.1 % Triton X-100. The incorporated OPP was then fluorescently labeled using a Click-iT reaction cocktail according to the manufacturer's protocol. Nuclei were counterstained with DAPI. Fluorescence microscopy and flow cytometry were employed to visualize and quantify global protein synthesis levels.

2.16. Flow cytometry analysis of intracellular mCherry fluorescence

To evaluate the transfection efficiency of iLMPs in RAW264.7 and MC3T3-E1 cells, iLMPs loaded with mCherry mRNA were employed; cells were seeded into 24-well plates and cultured until they reached approximately 80 % confluence, at which point iLMPs were added and the cultures were further incubated for 48 h. Following incubation, the cells were harvested and transferred into flow cytometry tubes; intracellular mCherry fluorescence was subsequently detected using a flow cytometer with excitation at 561 nm, thereby allowing quantitative assessment of mCherry expression levels under these experimental conditions.

2.17. Western blot analysis of protein expression

The expression of Bip, p-PERK, p-eIF2α and BMP-2 in each group of MC3T3-E1 cells was detected by Western blot analysis. After scraping the treated MC3T3-E1 cells, the cells were lysed on ice for 0.5 h in RIPA lysis solution containing protease inhibitors (Servicebio Co., Ltd., China), and the supernatant was collected by centrifugation. Proteins in the supernatant were separated by SDS‒PAGE and transferred to a PVDF membrane (Millipore, Billerica, MA), which was then blocked with 5 % BSA (Servicebio Co., Ltd., China) at room temperature for 1 h. The membrane was incubated with the primary antibody at 4 °C overnight. After incubation with the corresponding horseradish peroxidase-labeled secondary antibodies (Protein Technology, Inc., USA) for 1 h, the blots were detected by a chemiluminescence system, and optical density values were measured using ImageJ software.

2.18. qPCR analysis of osteoblast RNA expression

MC3T3-E1 cells were cultured to 80 % confluence, after which the medium was replaced with osteogenic induction medium containing 300 μmol/L H₂O₂. Cells were treated with iLMPs, B/iLMPs, or an equal volume of PBS (control). The medium was refreshed every 3 days, and cells were harvested on day 7. Total RNA was extracted using TRIzol, and its concentration and purity were determined. One microgram of RNA was reverse-transcribed into cDNA. qPCR was performed in a 20 μL reaction mixture containing SYBR Green Master Mix, primers, and cDNA template, with cycling conditions of 95 °C for 3 min, followed by 40 cycles, and a melting curve analysis from 65 °C to 95 °C.

2.19. ALP and ARS staining

MC3T3-E1 cells were cultured until approximately 90 % confluence and then switched to osteogenic induction medium containing MS, LMPs, B/LMPs, iLMPs or B/iLMPs. For ALP staining, on day 7 of induction, the cells were rinsed with PBS and fixed with 4 % paraformaldehyde. After fixation, ALP staining solution was added, and the samples were incubated for 30 min. The formation of blue–purple precipitates was observed to evaluate early osteogenic activity. For ARS staining, on day 21, the cells were fixed with 4 % paraformaldehyde, washed with PBS, and incubated with ARS solution for 30 min. After staining, the cells were rinsed thoroughly with distilled water. Calcium deposition was evaluated microscopically.

2.20. In vivo modeling and in situ bone regeneration

All animal experiments were approved by the Animal Ethics Committee of Shanghai Rising Biotechnology Co., Ltd. (Approve No.: 2023-09-RJYY-CYY-076). In accordance with previously reported protocols for establishing an animal model, eight-week-old male Sprague‒Dawley rats were adaptively fed for one week. Every three days, they received an intragluteal injection of methylprednisolone sodium (MPA) succinate (ChongQing Huapontpharm. Co., Ltd., China) at a dose of 40 mg/(kg·d), for a total of six injections. The rats in the control group were injected with an equal volume of physiological saline. 1 week later, the rats were anesthetized with isoflurane, and a longitudinal skin incision was made to expose the greater trochanter and hip joint by blunt dissection of the surrounding muscles. A drill was used to create an entry point at the greater trochanter, extending into the femoral neck to establish an intramedullary canal. Through this canal, 20 μL of MS, LMPs, B/LMPs, or B/iLMPs solution (at a concentration of 0.5 mg/mL) was injected into the medullary cavity. The incision was then sutured, and the surgical site was disinfected with povidone-iodine. At the 8th week post-treatment, the rats were euthanized under isoflurane anesthesia, and the femurs were collected and fixed in 4 % paraformaldehyde solution. The experimental groups were as follows: the control group, model group, MS group, LMP group, B/LMP group, and B/iLMP group.

2.21. X-ray and Micro-CT evaluation

The bone tissue parameters of the rat femoral head were analyzed using X-ray and Micro-CT. Anteroposterior and lateral X-rays were obtained using a Faxitron X-ray system at 59 kV with an exposure time of 5 s. The femoral heads were scanned and analyzed using a Micro-CT system (SkyScan, Belgium) with a spatial resolution of 4 μm, a reconstructed pixel size of 0.6 μm, and a scan precision of 30 μm, under a voltage of 50 kV and a source current of 500 μA. Following micro-CT scanning, 3-dimensional (3D) reconstruction and quantitative analysis were performed using dedicated software. The region of interest (ROI) was defined within the trabecular compartment of the femoral head, specifically beneath the articular cartilage and above the growth plate. Within this ROI, the following morphometric parameters were quantified: (1) bone mineral density (BMD), calibrated with hydroxyapatite phantoms; (2) trabecular number (Tb. N), representing the number of trabeculae per unit length; (3) bone volume fraction (BV/TV), the ratio of bone volume to total tissue volume; (4) trabecular separation (Tb. Sp), the mean distance between adjacent trabeculae; and (5) trabecular thickness (Tb. Th), the average thickness of individual trabeculae.

2.22. H&E staining, Masson staining, and immunofluorescence analysis

The femoral head samples were fixed in 4 % paraformaldehyde solution for 48 h, followed by decalcification in 10 % ethylenediaminetetraacetic acid (EDTA) solution for 4 weeks. The samples were subsequently dehydrated, embedded in paraffin, sectioned, and subjected to H&E staining and Masson's trichrome staining. Prior to immunofluorescence staining, the bone sections were blocked with 5 % BSA at room temperature for 0.5 h. The tissue sections were then incubated overnight with primary antibodies against Bip, BMP-2, and RUNX2, followed by incubation for 1 h with secondary antibodies. Images were captured using a confocal microscope, and immunofluorescence was analyzed using ImageJ software.

2.23. FICAT stage system analysis

The FICAT staging system is utilized to assess the severity of femoral head osteonecrosis. It is primarily based on the evaluation of X-ray images, clinical manifestations, and lesion progression to categorize the stages of femoral head necrosis. The FICAT system consists of four stages: 1. Stage I: This is the early stage, where X-rays typically reveal no obvious changes. The lesion is primarily detected through advanced imaging modalities such as MRI or CT. 2. Stage II: At this stage, the femoral head begins to undergo morphological changes but has not yet collapsed completely. X-rays may reveal mild swelling or deformation of the femoral head. 3. Stage III: The femoral head exhibits noticeable collapse, with X-rays showing significant deformities and cartilage damage. 4. Stage IV: This is the late stage, where the femoral head experiences severe collapse. The joint surface may display significant bony changes, which could lead to a loss of function between the femoral head and the acetabulum.

2.24. Statistical analysis

Each group in the experiments comprised at least three replicates. Means with standard deviations (SDs) are presented for normally distributed data. Student's t-test and the Mann‒Whitney test were performed to compare different results between the two groups. One- or two-way analysis of variance (ANOVA) was used to compare the results between multiple groups. Differences were considered statistically significant when p < 0.05. All statistical analyses were performed using SPSS version 21.

3. Results and discussion

3.1. Synthesis and characterization of iVES

We first synthesized iVES as a substitute for traditional ionizable lipids. As shown in Fig. 1A, iVES was synthesized through an esterification reaction between vitamin E succinate and 3-dimethylamino-1,2-propanediol, which was catalyzed by dicyclohexylcarbodiimide (DCC) and dimethylaminopyridine (DMAP). The high-resolution mass spectrum (HRMS) data are shown in Fig. 1B. The molecular ion peaks were observed at m/z 1144.8739 and 1166.8558, corresponding to the calculated values (1144.87, [M+H]⁺; 1166.86, [M+Na]⁺). In addition, the ¹H and ¹³C NMR data are shown in Fig. 1C, Figure S1 and Figure S2. All the chemical shift peaks can be assigned to the chemical structure of iVES. In Fig. 1C, the peak corresponding to the methyl groups in the phospholipid head appear as a characteristic singlet at approximately 2.3 ppm. Moreover, the peaks corresponding to the methylene groups on vitamin E and the multiple methyl groups on the alkyl chains are clearly observed at approximately 1.3 ppm and 0.8 ppm, respectively. The fourier transform infrared (FT-IR) spectrum further confirmed the structure of iVES (Fig. S3).

Fig. 1.

Synthesis of iVES and preparation of iLNPs. (A) Synthesis route of iVES from VES. (B) HRMS of iVES. (C) ¹H NMR spectrum of iVES. (D) HPLC elution curves of iVES and VES, as well as iVES under acidic conditions (iVES+H⁺). (E) The release of VES from iVES. (F) The pKa value of iVES, defined as the pH at half-maximal fluorescence intensity. (G) Different formulations (A1-A4) of iLNPs. (H) Schematic illustration of BMP2@iLNP preparation using the five components. Zeta potential (I), size (J) and encapsulation efficiency (EE%) (K) of the optimized BMP2@iLNPs prepared through a microfluidic device at a high pump rate. (L) mRNA binding analysis of BMP2@iLNPs (A3) performed by agarose gel electrophoresis. (M) TEM image of BMP2@iLNPs (A3) stained with a uranyl acetate solution (scale bar, 200 nm). The experiments were carried out 3 times. The data are presented as the means ± SDs.

The release properties of VES from iVES in a mildly acidic medium were investigated using high-performance liquid chromatography (HPLC). Here, iVES functions not only as an ionizable lipid but also as a prodrug of VES. The cleavable ester bond endows it with the ability to activate drugs under mildly acidic conditions, which facilitates the lysosomal escape process. As shown in Fig. 1D, the retention times of iVES and VES in the HPLC elution curve were 3.7 and 9.0 min, respectively (peaks in blue and red curves). Under a mildly acidic condition with a pH value of 5.0, the VES peak clearly appeared in the elution curve of iVES (green curve). The emergence of a new VES peak indicated the cleavage of ester bonds to release the parent drug of the iVES prodrug. Further quantitative analysis was conducted to record this VES release behavior under acidic conditions. As shown in Fig. 1E, VES was released from iVES at an almost constant rate, reaching approximately 60 % after 30 min. Thereafter, the release rate decreased.

We hypothesize that these two ester bonds have different hydrolysis rates, which may be attributed to the combined effects of hydrophobic structure and steric hindrance. To further verify our hypothesis, we performed additional LC-MS and relevant quantitative analysis of the degradation products of iVES under pH 5.0 and esterase conditions. As shown in Fig. S4, three major peaks were identified in the LC elution curve. The peak at 3.45–4.06 min corresponded to a mass signal of 1144.87 (Fig. S5A), representing the [M+H]⁺ of intact iVES. The peak at 8.91–9.97 min showed a mass signal of 553.38 (Fig. S5B), which was assigned to the [M+Na]⁺ of VES. The peak at 10.58–11.63 min corresponded to a mass signal of 429.37 (Fig. S5C), corresponding to the [M−H]^- of vitamin E (tocopherol). Based on peak area comparison, the area of VES peak was higher than that of vitamin E. Therefore, our hypothesis was confirmed, and the VES structure was retained.

In addition, the pKa value of iVES, which is crucial for its ionizable properties, was determined using a fluorescence labeling method. The fluorescence curve is shown in Fig. 1F. The measured pKa value of iVES was approximately 6.36, which is very close to that of commercially available ionizable lipids, such as DLin-MC3-DMA. This provides strong evidence for iVES as a potential replacement for traditional ionizable lipids.

These data indicate that the majority of covalently conjugated VES can be degraded and released within a short period of time. Therefore, it can be inferred that in the lysosomal environment, iVES not only undergoes significant protonation but also experiences accelerated degradation. This not only facilitates the release of the VES drug but also promotes lysosomal escape of the entire LNP, enabling faster delivery of mRNA to the cytoplasm.

ERS inhibitors broadly include chemical chaperones (e.g., 4-phenylbutyrate), UPR pathway modulators (PERK/IRE1/ATF6; e.g., GSK2606414), calcium homeostasis regulators (e.g., dantrolene), antioxidants and mitochondria–ER axis modulators (e.g., VES), and regulators of protein degradation or autophagy (e.g., rapamycin). In principle, all can lower ERS, attenuate UPR, and thereby enhance mRNA translation. We selected VES for three reasons. First, VES is a derivative of an endogenous vitamin with extensive safety data. Second, as a broad-spectrum antioxidant, VES modulates ERS indirectly through multiple synergistic pathways, including autophagy, which improves the applicability of iLMPs. Third, the formulation goal of iVES is to replace traditional ionizable lipids, which requires suitable lipophilicity, chemical tractability, and steric characteristics; VES meets these criteria and has prior use in phospholipid-like modifications.

3.2. Preparation and characterization of iLNPs

Owing to the similar structure and pKa value of iVES compared with traditional ionizable lipids, we attempted to directly substitute it for traditional lipids in the preparation of ionizable lipid based LNPs (iLNPs). According to the classic 4-component formulation, we tested 4 different formulation systems based on iVES (formulations A1-A4, as shown in Fig. 1G). Since iVES is a novel ionizable lipid, we focused on adjusting the ratio of iVES to DSPC to achieve a balance between iLNP formation and mRNA encapsulation efficiency. Here, BMP2@iLNPs were prepared using a microfluidic device, as illustrated in the schematic diagram shown in Fig. 1H. Briefly, BMP-2 mRNA was dissolved in citrate buffer (pH = 4) to form an aqueous phase (water phase), and iVES, DSPC, cholesterol, and DSPE-PEG-RGD were dissolved in anhydrous ethanol to form an ethanol phase (organic phase). As an in situ delivery system, to increase the probability of iLNPs being recognized by integrin receptors and improve their ability to target specific cells, we further modified the traditional 1.5 % PEG component with RGD short-peptide (DSPE-PEG-RGD). This short-peptide-conjugated lipid simplified the formulation, and allowed for the direct mixing in ethanol, avoiding the risk of antibody denaturation in organic solvents. In systemic delivery, PEGylation is employed to achieve a balance between prolonged circulation and targeted delivery. For in situ injectable formulations, the RGD ligand targets potential effector cells such as bone marrow cells and pre-osteoblasts by binding to integrins. Additionally, as a short-peptide with a defined structure, the RGD ligand avoids the issue of unclear antibody conjugation sites. The data in Fig. S6 further demonstrated that mCherry exhibited stronger fluorescence intensity in MC3T3-E1 cells, suggesting that iLNPs functionalized with DSPE-PEG-RGD are more effective in delivering mCherry mRNA to osteoblasts (MC3T3-E1) with high integrin receptor expression.

The zeta potential and particle size of BMP2@iLNPs with different formulations (A1-A4) were measured, and the results are shown in Fig. 1I and J. Overall, there were no significant differences in zeta potential among the various formulations. The zeta potential was approximately between −1 and −2 mV, and the particle size was approximately 200 nm. Among the 4 iLNPs formulations, the A3 formulation exhibited the most favorable physicochemical properties, with a significantly smaller particle size compared to the other three groups and a higher drug encapsulation efficiency (Fig. 1K). Therefore, A3 formulation was used for the following analysis for iLNPs preparation.

The agarose gel electrophoresis analysis was used to detect the mRNA loading efficiency. As shown in Fig. 1L, no mRNA signals were observed in all BMP2@iLNPs with different mRNA concentrations, which indicated that the mRNA was almost completely encapsulated. Moreover, TEM was used to further observer the morphology of BMP2@iLNP. As shown in Fig. 1M, BMP2@iLNP exhibited a distinct spherical shape, and the particle size information was consistent with the results obtained from DLS.

In general, a higher absolute zeta potential improves colloidal stability. In this study, our LNPs are near-neutral for three reasons: (1) The LNPs are incorporated into microspheres rather than used alone; the matrix improves storage stability; (2) We use in situ injection to achieve rapid local transfection, avoiding prolonged blood circulation and thus lessening the need for high colloidal stability in blood; (3) Near-neutral charge is advantageous here: it reduces undesirable protein corona effects from strong negative charge, improves cellular uptake, and lowers the cytotoxicity associated with highly positive charge.

We further assessed thermal stability of iLNPs at 4, 25, and 37 °C by measuring particle size distribution and encapsulation efficiency. As shown in Fig. S7A, lyophilized iLNPs stored below 4 °C maintained a stable particle size after 8 weeks. However, the particle size of lyophilized iLNPs stored at 25 °C continuously increased after the 4th week. For those stored at 37 °C, particle size continuously increased after the 3rd week. As shown in Fig. S7B, using Ribogreen to measure the encapsulation efficiency and mRNA content of lyophilized iLNPs, it was found that, except for iLNPs stored at 37 °C, there was basically no significant change at different temperatures over time. After storing lyophilized iLNPs at 4 °C for 8 weeks, at 25 °C for 4 weeks, and at 37 °C for 3 weeks, we characterized the morphological structure of iLNPs using TEM. As shown in Figure Fig. S7C–E, iLNPs still maintained a spherical structure.

3.3. Preparation and characterization of iLMPs

Hydrogel microspheres are an ideal tool for loading LNPs, significantly enhancing their efficiency for in situ delivery. On the one hand, microspheres improve the stability of LNPs (both their storage stability and in vivo stability). On the other hand, the porous microsphere system can rapidly fill defect sites, efficiently remodel the physical environment of the extracellular matrix, and provide adhesion sites for cells.

Hyaluronic acid methacrylated (HAMA) is a classic raw material for microsphere preparation that was developed for drug and cell loading in our previous studies. The microspheres were first prepared according to previous methods, as shown in Fig. 2A [25,26]. HAMA was dissolved in the aqueous phase, with liquid paraffin as the oil phase. Using optical microscopy, uniformly dispersed microspheres with consistent particle sizes (150.2 ± 8.0 μm) were observed, and the “oil-in-water” method was applied to ensure the roundness of the microspheres (Fig. S8). For convenient long-term storage and enhanced adsorption capacity, the microspheres were freeze-dried. The freeze-dried microspheres were nearly spherical in shape and appeared as solid black, opaque particles under optical microscopy due to dehydration (Fig. 2C). However, the freeze-dried HAMA microspheres had a mesoporous network structure, allowing them to rapidly absorb liquid and swell (Fig. 2D).

Fig. 2.

Preparation and characterization of iLMPs. (A) Microfluidic fabrication of injectable HAMA microspheres. (B) Preparation of iLMPs from HAMA microspheres with iLNPs adsorption on the surface. (C) Optical microscopy images of the microspheres after freeze-drying and swelling. (D) The swelling ratio of iLMPs. (E, F) SEM images of iLMPs. (G) LSCM image of DiL-stained LNPs loaded with iLMPs and illustration of iLMPs. (H) Morphological changes in iLMPs observed during degradation tests. (I) Degradation profile of iLMPs. (J) Release curve of iLNPs from iLMPs. The experiments were carried out 3 times. The data are presented as the means ± SDs.

Lipid-hydrogel microplexes (LMPs) were further prepared based on these blank HAMA microspheres. The LMPs prepared using iLNPs were named iLMPs. The preparation of iLMPs was achieved directly through physical adsorption, as shown in Fig. 2B. First, HAMA microspheres were fully resuspended in deionized water. Then, BMP-2-loaded iLNPs were added. Owing to RGD modification, physical adsorption interactions such as hydrogen bonding and van der Waals forces occurred between the iLNPs and the HAMA microspheres, ensuring effective loading of LNPs and other nanoparticles. The reduced negative charge of the LNPs decreases electrostatic repulsion with the microspheres, thereby enhancing the stability of the complexes. In addition, the pH of the deionized water suspending the HAMA microspheres was approximately 7. Given that the pKa of iVES is 6.36, which is lower than 7, iVES would remain in its nonionic form, thereby preserving its stability. Besides, the preparation process was completed within a very short time, and rapid lyophilization was employed to avoid prolonged exposure to deionized water. The storage as lyophilized powder form could improve their stability as well.

In addition, to further address concerns regarding storage stability at −80 and 4 °C. We did a transfection efficiency assay of lyophilized iLMPs after 12 h and 1 week storage at −80 °C. mCherry mRNA was used for fluorescence microscopy observation after protein expression. As shown in Fig. S9, mCherry mRNA iLMPs remained comparable to that of freshly prepared formulations, with no significant decline observed. To further evaluate the effect of microspheres on the storage stability of iLNPs at 4 °C, we compared the transfection efficiency of iLNPs and iLMPs after storage at 4 °C for different durations. On days 7 and 10, iLMPs exhibited higher fluorescence expression levels. This suggests that the microspheres help reduce the hydrolytic and oxidative degradation of iLNPs, protecting them from enzymatic attack and thereby enhancing their stability (Fig. S10).

Furthermore, we observed the morphology of the microsphere surface through scanning electron microscopy (SEM). The microspheres appeared as regular, porous, nearly spherical microparticles, likely due to shrinkage during freeze-drying (Fig. 2E). Upon closer magnification, the mesoporous structure on the microsphere surface could be observed, facilitating iLNP adsorption and release (Fig. 2F). The nanoscale lipid nanoparticles uniformly adsorbed into the mesopores of the microspheres through noncovalent and physical interactions. Additionally, after the iLNPs were labeled with DiL dye, the iLMPs exhibited a uniform red coating under confocal microscopy, indicating effective adsorption of the iLNPs onto the microspheres (Fig. 2G). Finally, iLMPs were incubated at 37 °C in PBS containing hyaluronidase, which gradually degraded HAMA via enzymatic action. Inverted microscopy revealed that the degradation period for the HAMA hydrogel microspheres was 8 weeks, and the weight measurement results closely matched the degradation observed under inverted microscopy (Fig. 2H–J).

3.4. In vitro biocompatibility and cytotoxicity evaluation of iLMPs

Because iLMPs are designed for direct injection into the lesion site, the cytotoxicity of iLMPs is of critical importance. First, live/dead staining was applied to evaluate the cytotoxicity of iLMPs. This study focused on bone defects; therefore, the preosteoblastic MC3T3-E1 cell line was chosen as the model. Briefly, MC3T3-E1 cells were cocultured with extract solutions of microspheres (MSs), LNP microspheres (LMPs), BMP2@LMPs (B/LMPs), or BMP2@iLMPs (B/iLMPs) for 24 and 72 h. The live/dead staining results are shown in Fig. 3A. Significant cell proliferation was clearly observed in all the groups, and the cells exhibited a very high survival rate (Fig. S11).

Fig. 3.

In vitro biocompatibility evaluations of iLMPs. (A) Live/dead analysis of MC3T3-E1 cells after different treatments for 24 and 72 h via calcein-AM/PI staining. Scale bar, 50 μm. (B) Representative images of the different groups in the scratch experiment. Images of the migration of MC3T3-E1 cells after different treatments. Scale bar, 50 μm. (C) Semiquantitative analysis of the scratch test results. (D) CCK-8 assay to detect the cytotoxicity of B/iLMPs. *n.s.: not significant (p > 0.05).

Moreover, a scratch assay was performed to evaluate the ability of B/iLMPs to inhibit the migration of MC3T3-E1 cells. After MC3T3-E1 cells were cocultured with the microsphere solutions for 24, 48, and 72 h, no significant difference in the migration area was detected in the MC3T3-E1 cells (Fig. 3B and C). Then, a CCK-8 assay was used to further evaluate the proliferation of the cocultured MC3T3-E1 cells. As shown in Fig. 3D, after 3 days of coculture, the cells in each group exhibited nearly 3-fold greater proliferation than they did on day one, with no significant differences among the groups (p > 0.05). A hemolysis assay was conducted, and no significant increase in absorbance was observed in the B/LMP and B/iLMP groups compared to the control group (p > 0.05) (Fig. S12). Additionally, the diameter of HAMA microspheres ranges from 120 to 180 μm, which is much larger than the diameter of capillaries, preventing vascular dissemination. In vivo bioluminescence imaging revealed that iLMP localized predominantly to the femoral head and intertrochanteric region in rats, without signs of systemic dissemination (Fig. S13).

It is evident from the above results that neither the microspheres nor the complexes of microspheres and LNPs (including iLMPs and B/iLMPs) exhibited significant cytotoxicity or inhibited cell proliferation. This outcome is consistent with expectations, as most of the raw materials used for both the microspheres and LNPs have been extensively validated [27]. Among these materials, only iVES is a newly synthesized compound. Although positively charged molecules typically exhibit some degree of toxicity, iVES is biologically safe because its pKa is lower than normal physiological levels, which allows it to present a neutral charge under physiological conditions [28]. The safety of other related ionizable lipids has also been rigorously evaluated [29].

To evaluate the potential immunogenicity of iVES, we used an LPS-stimulated RAW 264.7 macrophage model and incubated cells with VES or iVES (5 μM, 24 h). qPCR analysis showed that both VES and iVES significantly reduced the LPS-induced mRNA expression of Mmp13, Il6, and Tnfα. Notably, Il6 and Tnfα levels did not differ significantly between VES and iVES, indicating comparable pharmacological activity and no evidence of heightened inflammatory transcriptional responses with iVES (Fig. S14). These data support a low immunogenicity profile for iVES at the cellular level.

In this study, we did not observe a proliferative effect of BMP-2 on MC3T3-E1 cells, which is also debated in the literature. Possible reasons include: (1) The inherently high proliferation rate of MC3T3-E1 cells may mask subtle effects of BMP-2 mRNA, especially in short-term cultures; (2) mRNA-based delivery differs from recombinant protein in kinetics and efficiency, often resulting in sub-threshold protein levels at early time points due to factors like mRNA modification, LNP release, and cellular uptake; (3) BMP-2 mRNA significantly upregulated osteogenic markers (ALP, Runx2), indicating active differentiation, during which proliferation typically decreases and may obscure early proliferative effects.

3.5. In vitro evaluation of iLMP-mediated ERS attenuation

Under certain pathological conditions, misfolded proteins within cells can induce ERS. In response to this phenomenon, cells spontaneously activate the unfolded protein response (UPR) to restore protein homeostasis. During the UPR, the global downregulation of ribosomal mRNA translation is inevitable [30,31]. In general, cells encounter various environmental stresses, such as inflammatory, ischemic, and hypoxic microenvironments, which can disrupt intracellular protein homeostasis and lead to ERS [6]. In SANFH, excessive ROS-induced oxidative stress is a key factor that triggers ERS in osteoblasts. Thus, we used H₂O₂ to establish an in vitro model with oxidative stress and inflammation. Upregulation of ERS markers (GRP78, CHOP) and increased phosphorylation of PERK and eIF2α by H₂O₂ treatment is representative, according to previous literatures [32]. Accordingly, MC3T3-E1 cells treated with 300 μmol/L H₂O₂ were selected as the in vitro model, providing a valid system for evaluating the efficacy of the iLMP formulation [33].

According to previous studies, activation of the Bip-PERK-eIF2α pathway in the UPR directly inhibits the translation of downstream proteins, as shown in Fig. 4A [34]. Clearly, this is not desirable for mRNA-based therapeutic studies, as such ribosomal translation suppression inevitably affects the efficacy of mRNA therapeutics.

Fig. 4.

In vitro evaluation of iLMP-mediated ERS attenuation. (A) Mechanism of reduced mRNA translation efficiency during ERS. (B) Western blot analysis of Bip, p-PERK and p-eIF2α protein expression in MC3T3-E1 cells after different treatments. (C–E) Semiquantitative analysis of Bip, p-PERK and p-eIF2α expression by Western blot analysis. (F) Schematic diagram of the OPP assay for detecting ribosomal translation levels. (G) Fluorescence images of OPP assay results in MC3T3-E1 cells and immunofluorescence staining of ROS production in MC3T3-E1 cells after transfection with iLMPs. Scale bar, 50 μm. (H, I) Fluorescence images and semiquantitative analysis of mCherry-positive MC3T3-E1 cells transfected with iLMPs obtained by flow cytometry. Scale bar, 50 μm **p < 0.01.

To clarify the attenuation of ERS and the underlying mechanism of B/iLMPs, we first examined the expression of key proteins in the Bip–PERK–eIF2α pathway using Western blot analysis [35]. The results showed that treatment with B/iLMPs reduced the expression of Bip, p-PERK, and p-eIF2α, indicating that B/iLMPs suppressed the excessive activation of the Bip–PERK–eIF2α pathway (p < 0.01) (Fig. 4B–E). Moreover, the ratios of p-PERK/PERK and p-eIF2α/eIF2α were significantly decreased, suggesting a reduced proportion of activated PERK and eIF2α (p < 0.01) (Fig. S15). These findings collectively demonstrate that the Bip–PERK–eIF2α pathway was attenuated or inhibited, leading to a reduction in misfolded proteins, blockade of UPR overactivation, alleviation of protein-folding stress within MC3T3-E1 cells, and ultimately, relief of ERS. Notably, the reduction in p-eIF2α directly suggests an increase in ribosomal translation, which is considered the most closely associated indicator [34].

The OPP assay is a novel method for detecting ribosomal translation levels (Fig. 4F). This method avoids the use of traditional radioactive labeling, making it very safe and efficient. OPP is a derivative of puromycin, which contains an alkyne group at its terminus that can enter the ribosomal receptor site and become incorporated into nascent polypeptide chains. After undergoing a copper-catalyzed azide‒alkyne cycloaddition reaction, it emits red fluorescence, allowing for the quantification of protein synthesis within the cells [36]. Numerous previous experiments have confirmed that the OPP assay can accurately measure translation levels [37]. Here, we employed an OPP assay to assess protein translation levels. As shown in Fig. 4G and S16, OPP-related red fluorescence could directly indicate protein translation levels. Among all the experimental groups, the B/iLMP group presented significantly enhanced red fluorescence. The red fluorescence within the cells was further quantified using flow cytometry, as shown in Fig. S17. Moreover, the ROS levels and OPP appeared to have a clear negative correlation, as shown in Fig. 4G. Generally, an increase in cellular translation leads to a simultaneous increase in the expression of exogenous mRNA. To verify this, we additionally introduced mCherry mRNA to detect the enhanced translation of such exogenous mRNA. Compared with the LMP group, the iLMP group clearly presented greater mCherry expression, as shown in Fig. 4H and I, and S18 This result corresponds to the previously observed increase in cellular translation levels.

These results indicate that B/iLMPs effectively alleviate cellular ERS and inhibit activation of the protein translation-related Bip-PERK-eIF2α pathway. The final concentration of iLNPs used in this study was 2 μg/mL, with an active vitamin E content of 0.94 μg/mL, which is within the normal therapeutic range [21]. Moreover, the significant increase in nascent protein expression detected by the OPP assay clearly indicates an increase in ribosomal translation. Finally, this increase in translation leads to a similar increase in the expression of exogenous mRNA. These findings provide the strongest experimental evidence that iLMPs enhance the therapeutic efficacy of mRNA-based drugs.

Moreover, we detected other canonical UPR downstream genes by qPCR. As shown in Fig. S19, in H₂O₂-treated MC3T3-E1 cells, iVES reduced the H₂O₂-induced upregulation of DNAJB9, GRP94, and PDIA4, suggesting attenuation of IRE1α/ATF6 signaling beyond the PERK branch. Multiple UPR suppressing factors appear to contribute to the action of iVES, and we infer that its primary efficacy arises from indirect modulation rather than direct clearance of ER stress.

Beyond direct antioxidant activity, VES can indirectly modulate ERS via effects on mitochondria, inflammation, and autophagy. VES preferentially accumulates in mitochondria, alleviates ROS and calcium dysregulation, and influences ER–mitochondria crosstalk to modulate UPR pathways. By reducing inflammation-driven redox imbalance and adjusting autophagic flux, α-TOS can further lower ER load. Overall, VES regulates ERS primarily through indirect mechanisms, and this modulatory effect is widely recognized; in oxidative stress models, VES is consistently associated with reduced ERS markers.

3.6. In vitro osteogenic evaluation of B/iLMPs

In cases of inflammatory bone defects and other tissue injuries, cells often exhibit increased intracellular ROS levels, ERS, and PERK pathway activation. More importantly, the phosphorylation level of eIF-2α is generally elevated, which inevitably leads to the suppression of ribosomal translation. Undoubtedly, the delivery of mRNA to related cells is affected. BMP-2 is the most common osteoinductive protein and has the widest application in bone repair. Here, we used iLMPs as a nucleic acid delivery tool loaded with BMP-2 mRNA and evaluated its effectiveness in bone repair. MC3T3-E1 cells were coincubated with drugs such as B/iLMPs. The transfection and translation efficiency of BMP-2 mRNA were used to collectively evaluate the osteogenic process [[38], [39], [40], [41], [42]]. Therefore, by detecting osteogenic markers, the significance of iLMPs in regulating ribosomal translation levels could be indirectly assessed.

First, the expression of two important osteogenic proteins, BMP-2 and RUNX2, in MC3T3-E1 cells after B/iLMP treatment was detected by Western blotting. As shown in Fig. 5A–C, the B/iLMP group presented significantly increased BMP-2 and RUNX2 expression compared with the other groups (p < 0.05). Here, the expression of BMP-2 is partly attributed to the transfection of exogenous mRNA. Following the induction of osteogenic differentiation in MC3T3-E1 cells, the expression of endogenous BMP-2 is correspondingly upregulated. Thus, the observed BMP-2 expression may represent a cumulative effect. RUNX2 expression, in turn, is a result of this combined osteogenic differentiation induction. Overall, the high expression of BMP-2 and RUNX2 directly reflects the molecular biological processes of cellular osteogenic differentiation.

Fig. 5.

In vitro evaluation of iLMP-mediated osteogenesis. (A) Western blot analysis of BMP-2 and RUNX2 protein expression in MC3T3-E1 cells. (B, C) Semiquantitative analysis of BMP-2 and RUNX2 expression by Western blot analysis. (D–F) Osteogenic differentiation assessed by ALP staining and semiquantitative analysis. Scale bar, 50 μm. (G–I) Osteogenic differentiation assessed by Alizarin red S staining and semiquantitative analysis. Scale bar, 50 μm **p < 0.01, n.s.: not significant (p > 0.05). ALP: alkaline phosphatase. ARS: Alizarin red s.

Moreover, the osteogenic differentiation level of MC3T3-E1 cells was further characterized by alkaline phosphatase (ALP) activity and Alizarin red S (ARS) staining. After 7 days of induction, ALP staining was significantly greater in the B/iLMP group than in the other groups, including the B/LMP group (Fig. 5D–F). In addition, ARS staining revealed similar results, indicating increased intracellular calcium salt deposition after B/iLMP treatment (Fig. 5G–I). Generally, ALP and ARS staining represent the early and late stages of osteogenic differentiation, respectively. Therefore, from the results mentioned above, it can be inferred that B/iLMPs rapidly affect MC3T3-E1 cells during the early stage of osteogenesis. Moreover, this differentiation induction persists, ultimately accelerating osteogenic mineralization.

To clarify the role of iLMPs, we conducted an additional in vitro experiment about the osteoinductive effect of iLMPs and B/iLMPs using qPCR and ALP staining assay. As shown in Fig. S20A and 20B, the mRNA expression of RUNX2 and ALP in B/iLMPs group is significantly higher than those of iLMPs group (p < 0.01). Meanwhile, iLMPs group also shows high expression of both genes (p < 0.05). Besides, ALP staining showed the similar results with qPCR analysis (Fig. S20C). From these data, the limited improvement of osteoinductive effect of iLMPs is confirmed.

These data suggest that B/iLMPs can accelerate osteogenic differentiation in vitro. Notably, compared with that induced by B/LMPs, this acceleration of osteogenic differentiation was significantly greater. This is evidently related to the unique bioactivity of iLMPs. The previously validated increase in ribosomal translation clearly accelerates the translation of cargo BMP-2 mRNA. Such regulation of translation levels is crucial for the efficacy of osteogenic mRNA therapeutics in treating bone defects.

3.7. Establishment of the animal model

Long-term use of glucocorticoids (GCs) is a primary cause of SANFH, as it leads to the generation of ROS and free radicals, which ultimately induce oxidative stress in bone metabolism-related cells, such as osteoblasts, mesenchymal stem cells, and osteocytes, thereby disrupting bone tissue homeostasis [43]. The use of animal models for in vivo evaluation of osteogenic functionality will more intuitively demonstrate the therapeutic value of the novel in situ injectable B/iLMPs. SANFH is a typical inflammatory disease characterized by high intracellular ERS. The long-term administration of GCs is a major etiological factor in the development of SANFH. GCs can exacerbate the protein folding burden, disrupt calcium homeostasis, and increase ROS levels, thereby activating three principal ERS signaling pathways, PERK, IRE1, and ATF6. Among these pathways, the PERK–eIF2α pathway plays a particularly critical role in mediating osteocyte apoptosis. Evidence indicates that targeted intervention in ERS signaling can significantly alleviate osteocytic injury and delay the progression of SANFH [44,45].

In addition, glucocorticoids induce oxidative stress in bone metabolism-related cells, such as osteoblasts, mesenchymal stem cells, and osteocytes. Oxidative stress is widely recognized as a key upstream mechanism in the initiation of ERS under various pathological conditions. It promotes the accumulation of misfolded proteins and triggers ERS by disrupting the protein-folding environment within the endoplasmic reticulum, disturbing calcium homeostasis, and directly activating ERS sensors such as PERK, IRE1, and ATF6 [6,46]. Currently, there is a lack of mRNA therapeutic strategies for SANFH, which may also be related to limitations in the level of cellular translation under ERS. Therefore, animal models of SANFH are very suitable for evaluating the therapeutic potential of B/iLMPs. All animal experiments were approved by the Animal Ethics Committee of Shanghai Rising Biotechnology Co. Ltd., with approval number 2023-09-RJYY-CYY-076.

Before B/iLMP treatment, we used MPA succinate to establish the SANFH model in rats (Fig. 6A) [47,48]. After 2 weeks of treatment, the SANFH model was established. We performed two in situ injections into the femoral head during the 3rd and 4th weeks, with each injection consisting of 20 μL (with iLMP concentration of 0.5 mg/mL). 8 weeks after MPA injection, all the animals were sacrificed, and the femoral heads were extracted for further analysis.

Fig. 6.

Establishment of the animal model and pathological analysis. (A) Schematic illustration of SANFH establishment by methylprednisolone sodium (MPA), administration modality, and therapeutic timeline. (B) Representative images of H&E-stained femoral head samples. Scale bars:1 mm (top), 100 μm (bottom). (C) Representative images of Masson's trichrome-stained samples from the different groups. Upper scale bar, 1 mm; lower scale bar, 100 μm. (D) Micro-CT images of the sagittal and transverse views of femoral head samples. (E) Ficat stages of femoral head necrosis after different treatments. (F–J) Quantitative micro-CT analysis of the bone mineralization density (BMD), trabecular number (Tb.N), bone volume fraction (BV/TV), trabecular separation (Tb.Sp), and trabecular thickness (Tb.Th) of femoral head samples. *p < 0.05, **p < 0.01, n.s.: not significant (p > 0.05).

Here, we administered two injections within the first two weeks after model establishment, based on a comprehensive consideration of the LNP release rate from iLMPs and the limited translation window of mRNA. Because protein expression from linear mRNA is short-lived, this two-dose regimen was designed to maximize mRNA availability during the early phase of bone repair and thereby initiate the regenerative cascade. Employing circRNA-based technologies could effectively prolong the duration of expression. In future studies, we will refine mRNA modification strategies or introduce circRNA approaches to evaluate whether a single injection can achieve comparable efficacy.

3.8. In vivo pathological analysis

First, we performed histopathological analysis of the femoral head tissue using sectioning and Micro-CT techniques. The femoral head samples were sectioned after fixation, decalcification, and embedding, followed by pathological analysis using H&E and Masson staining. H&E staining (Fig. 6B) revealed that, compared with the B/LMP group, the B/iLMP group presented a significant increase in new bone tissue and trabeculae in the femoral head, a reduction in fat cell infiltration, and a marked decrease in the degree of femoral head necrosis (p < 0.05). Furthermore, in the B/iLMP group, Masson staining (Fig. 6C) revealed a significant effect on bone tissue repair in the femoral head, with notable increases in the amount and area of the collagen fiber network, indicating better bone tissue repair outcomes. In addition, Micro-CT analysis was applied to further evaluate the pathological structures and bone tissue parameters of the femoral head samples. The 3D reconstructions are shown in Fig. 6D and S21. The results revealed that at 8 weeks post-modeling, the rats in the model group exhibited sparse, fractured, and missing trabeculae in the femoral head, with visibility of the bone marrow cavity. Compared with the model group, the B/iLMP group presented more new bone formation, an increased number of trabecular structures, and denser bone tissue.

According to the established scoring criteria, we analyzed the femoral head necrosis stages in each group [49]. As shown in Fig. 6E, the femoral head necrosis stage was significantly reduced in the B/iLMP-treated group. The results of the analysis of bone tissue parameters by Micro-CT are shown in Fig. 6F–J. The CT data showed that the bone mineral density (BMD), trabecular number (Tb.N), bone volume fraction (BV/TV), and trabecular separation (Tb.Sp) were significantly greater in the B/iLMP group than in the MS, LMP, and B/LMP groups, indicating marked improvement in bone destruction and osteoporosis (p < 0.05). All the pathological examination results demonstrated a significant improvement in the pathological effects on bone structure.

3.9. In vivo osteogenic molecular biology evaluation

Whether the improvement in the pathological state of bone tissue in SANFH could be attributed to enhanced ribosomal translation and the corresponding increase in BMP-2 mRNA transfection efficiency was further investigated. First, considering the importance of ERS, the UPR and the Bip-PERK-eIF2α pathway in mRNA translation, evaluating these processes could be used to directly determine the efficacy of exogenous BMP-2 mRNA at the lesion site (Fig. 7A). During these processes, high oxidative stress impairs endoplasmic reticulum function, leading to the accumulation of unfolded or misfolded proteins in the ER lumen as well as the dysregulation of calcium homeostasis, which ultimately leads to ERS and activation of the UPR. During ERS, PERK is kept inactive by binding to BiP on the ER luminal side. When misfolded proteins accumulate, BiP dissociates, allowing PERK to dimerize and autophosphorylate, activating its kinase domain. Activated PERK then phosphorylates eIF2α, leading to inhibition of translation initiation and suppression of global protein synthesis [35]. The in vitro data revealed that iLMPs containing the novel iVES demonstrated excellent ERS alleviation capability and enhanced mRNA translation. We hypothesized that the same effect occurs in vivo.

Fig. 7.

In vivo evaluation of B/iLMP-mediated ERS attenuation. (A) Mechanism of decreased mRNA translation efficiency during ERS. (B–D) Immunohistochemical staining for ERS-related proteins (Bip, p-PERK and p-eIF2α) in femoral head samples. Scale bars: 200 μm (top), 50 μm (bottom). (E–G) Semiquantitative analysis of Bip, p-PERK and p-eIF2α protein levels via immunohistochemical staining. *p < 0.05, **p < 0.01.

We further conducted immunofluorescence analysis of pathological sections of the femoral head. Staining for three key proteins (Bip, p-PERK and p-eIF2α) was performed separately. Among them, the phosphorylated proteins (p-PERK and p-eIF2α) are critical markers of the activity of this UPR-related pathway. As shown in Fig. 7B, the B/iLMP group presented extremely low Bip expression (p < 0.05) that was similar to that in the control group. Moreover, all the other groups, including the B/LMP group, presented significantly greater Bip expression levels, with no significant differences among them (p > 0.05). As shown in Fig. 7B and D, the results for p-PERK and p-eIF2α expression were similar. Generally, the B/iLMP group presented lower expression levels of the characteristic proteins (p < 0.05). The semiquantitative analysis results are shown in Fig. 7E–G. These data suggest that B/iLMP could significantly downregulate the activation of the Bip-PERK-eIF2α pathway, which this would clearly benefit the ERS-related UPR in femoral head lesions.

As ERS and the UPR at attenuated, ribosomal translation would recover accordingly, resulting in an increase in the expression of exogenous mRNA. Thus, B/iLMP treatment would effectively increase BMP2 expression levels in transfected cells. BMP-2 is essential for bone repair and regeneration by targeting mesenchymal stem cells and osteoblasts. It activates the Smad pathway, leading to phosphorylation of Smad1/5/8, which form a complex that enters the nucleus and promotes RUNX2 transcription. RUNX2 then drives bone matrix deposition and mineralization, facilitating bone formation and repair (Fig. 8A) [50].

Fig. 8.

In vivo evaluation of B/iLMP-mediated bone regeneration. (A) Schematic illustration of bone regeneration. (B, C) Immunohistochemical staining for BMP-2 and RUNX2 in femoral head samples. Scale bars: 40 μm (top) and 12 μm (bottom). (D, E) Semiquantitative analysis of BMP-2 and RUNX2 expression by immunohistochemical staining. *p < 0.05, **p < 0.01, n.s.: not significant (p > 0.05).

We further analyzed femoral head sections using immunofluorescence staining to detect BMP-2 and Runx2 expression. As shown in Fig. 8B, the protein expression of BMP-2 in the femoral head of the B/iLMP group was significantly higher than in the other experimental groups, including the B/LMP group. In the nucleus, we also observed a significant increase in the expression of Runx2 protein, as shown in Fig. 8C. The corresponding semiquantitative results are shown in Fig. 8D–E.

Owing to the significantly greater expression of BMP-2 in the B/iLMP group than in the B/LMP group, although BMP-2 can be effectively expressed without the involvement of iLMPs, its expression level remains limited. Only with iLMPs can the ribosomal translation inhibited by ERS be restored, thereby unleashing the true efficacy of exogenous mRNA. These data clearly demonstrate the critical role of iVES and iLMPs in mRNA expression under such pathological-like conditions. The expression of Runx2 is clearly positively correlated with that of BMP-2. The pathway through which BMP-2 influences Runx2 is relatively complex, and this influence partially overlaps with other osteogenic induction effects. Overall, iLMPs and iVES play crucial roles in promoting BMP-2 mRNA expression and osteogenic induction, especially in the complex pathological environment of SANFH.

Undoubtedly, the novel ERS-attenuating lipid-hydrogel microplexes will serve as a powerful tool for the in situ delivery of therapeutic mRNA. By significantly improving the efficacy of mRNA drugs in pathological states with disrupted protein homeostasis, the iLMP platform holds great potential to be extended for the treatment of a broader range of tissue defects. As a common feature of chronic diseases, ER stress is not only a key pathological mechanism but also affects mRNA translation, thereby influencing the development of mRNA therapeutics. In this work, our aim is not merely to present a new mRNA delivery strategy for bone defects; rather, we use bone defect as a model to explore a more general mRNA delivery platform for the treatment of chronic diseases on a wider scale.

4. Conclusion

In this work, we synthesized a novel iVES for the preparation of LNP-based iLMPs, with the aim of creating an injectable mRNA delivery system to alleviate ERS and ultimately enhance the ribosomal translation efficiency of exogenous mRNA. iVES possesses physicochemical properties similar to those of commercial ionizable lipids. As a substitute, the LNP system prepared with iVES still exhibited comparable mRNA encapsulation efficiency. The subsequently prepared iLMPs, as an in situ mRNA delivery tool, effectively addresses the ribosomal translation limitations caused by the UPR induced by ERS. Both in vitro and in vivo, ERS and ribosomal translation levels were significantly improved. Moreover, after loading BMP-2 mRNA, the osteogenic induction effect on cells under pathological conditions was also significantly enhanced. In summary, iLMPs, as an in situ delivery tool, has significant potential for the treatment of chronic diseases with mRNA.

CRediT authorship contribution statement

Yanyang Chen: Writing – original draft, Validation, Formal analysis, Data curation. Wei He: Methodology, Investigation, Data curation. Shifeng Ling: Validation, Formal analysis, Data curation. Yang Zhou: Software, Investigation, Formal analysis. Jie Chen: Software, Methodology. Yawei Du: Writing – review & editing, Project administration. Ran Mo: Writing – review & editing, Project administration, Conceptualization. Wenguo Cui: Writing – review & editing, Funding acquisition, Conceptualization.

Ethics approval and consent to participate

All the animal experiments in this study were approved by the Animal Ethics Committee of Shanghai Rising Bio-technology Co. Ltd. (Ethical Approval No. 2023-09-RJYY-CYY-076). All the authors were in compliance with all relevant ethical regulations.

Declaration of AI assistance

The author team independently developed this manuscript and employed ChatGPT exclusively for linguistic improvement after draft completion. Following this assistance, the authors reviewed and edited all content, accepting full responsibility for the final publication.

Declarations of competing interest

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

Appendix A. Supplementary data

The following is the Supplementary data to this article: