Cholesterol analogs modulate lipid nanoparticle performance for mRNA delivery after lyophilization and enable ocular disease therapy

Abstract

Lipid nanoparticles (LNPs) are essential nucleic acid delivery carriers, with cholesterol being crucial for their structural integrity and intracellular transport. Although C24-alkyl phytosterols (cholesterol analogs) promote gene transfection in HeLa cells, their broader impact on mRNA-LNP performance across organs and organ-derived cell lines remains unknown. We investigated the effects of C24-alkyl phytosterols on mRNA-LNP morphology, stability, transfection efficiency, and in vivo biodistribution before and after lyophilization. Freshly prepared β-sitosterol–containing mRNA-LNPs exhibited the highest transfection efficiency in most cell lines. Stigmasterol-containing mRNA-LNPs demonstrated superior performance after lyophilization. In vivo, compared with cholesterol-based controls, β-sitosterol–formulated ALC-0315-based mRNA-LNPs and fucosterol–formulated MC3-based LNPs yielded increased bioluminescence in target organs. β-sitosterol–mRNA-LNPs delivering (MERTK)mRNA robustly restored visual function in a retinal degeneration rat model. These findings highlight how subtle structural modifications of cholesterol analogs can influence mRNA-LNP therapeutic efficacy, providing ideas for the rational design of next-generation gene delivery systems. We demonstrated the therapeutic effects of mRNA-LNPs in an animal model of eye disease (RCS rats), paving the way for application in ocular gene therapy.

Graphical abstract

1. Introduction

Lipid nanoparticles (LNPs) encapsulating messenger RNA (mRNA), which can transfect genes into patients or host cells without integration, have emerged as revolutionary platforms for gene delivery [1], particularly in the fields of regenerative medicine and vaccine development [[2], [3], [4], [5]]. These nanoscale delivery vehicles enable the efficient transportation of fragile mRNA molecules into cells, protecting them from enzymatic degradation while ensuring their stability during storage and administration [4,6]. Their modular design allows the optimization of key structural components, such as ionizable lipids [7], cholesterol [6,8], helper lipids [9,10], and PEGylated lipids [11], which collectively determine their encapsulation efficiency, cellular uptake, and eventual therapeutic performance [12,13].

Cholesterol, a key component of mRNA-LNPs, plays an important role in maintaining the structural integrity and fluidity of lipid bilayers [8,12]. In addition to its stabilizing function, cholesterol has been shown to enhance the fusion of mRNA-LNPs with cellular membranes, thereby facilitating intracellular mRNA delivery [8,14]. Recent studies have highlighted the profound impact of cholesterol analog optimization on the delivery efficiency and stability of mRNA-LNP formulations in vitro [8,14].

C24-alkyl sterols—a structurally diverse group of naturally occurring phytosterols—are promising alternatives to conventional cholesterol in the design of mRNA-LNPs [14]. These sterols, including β-sitosterol, fucosterol, campesterol, and stigmastanol, differ from cholesterol in the length and saturation of their alkyl side chains. The ability of these compounds to induce crystalline defects in lipid membranes increases with side chain length, which enhances membrane fluidity and disrupts lipid packing [8,14]. This feature has been exploited to improve nanoparticle architecture and performance. Notably, even subtle modifications to the sterol structure—such as differences in alkyl side chain length or molecular rigidity—can significantly affect mRNA-LNP morphology, encapsulation efficiency, and cellular uptake (e.g., cholesterol < campesterol < β-sitosterol), further underscoring the importance of rational sterol selection in mRNA-LNP engineering [8,15]. However, previous researchers have investigated the transfection efficiency of mRNA-LNPs with cholesterol analogs only in specific cell lines [8,14], such as HeLa cells, fibroblasts, macrophages, dendritic cells, adenocarcinoma cells, epithelial cells and HEK293 cells. They did not evaluate the transfection efficiency of mRNA-LNPs with cholesterol analogs into a variety of organ cells, such as lung cells (A549 cells), kidney cells (HEK-293T cells), cervical cells (HeLa cells), colon cells (Colo205 cells), hepatocyte cells (HepG2 cells), and retinal epithelial cells (ARPE19 cells) (Fig. 1), as we did in this study. Furthermore, the reported studies [16,17] did not evaluate the biodistribution following systemic or subretinal administration of various cholesterol analogs containing mRNA-LNPs into C57BL/6J mice, and vision rescue efficacy in the animal disease model of Royal College of Surgeons (RCS) rats, whereas we performed these in vivo experiments in this study.

Fig. 1.

Schematic diagram illustrating the preparation of mRNA-LNPs, evaluation of their translatability, and in vivo validation of the therapeutic efficacy of the optimal mRNA-LNPs encoding the MERTK protein for the treatment of retinal degeneration.

Lyophilization (freeze-drying) has emerged as a pivotal technique for the stabilization and long-term storage of mRNA-LNP systems [15,[18], [19], [20]]. By removing the water content under controlled conditions, lyophilization minimizes hydrolytic degradation and preserves particle integrity [20]. Recent studies have reported excellent lyophilization of mRNA-LNPs [19,[21], [22], [23]], which can maintain systemic immunogenicity as well as particle stability after reconstitution.

Despite advances in the lyophilization of mRNA-LNP systems, significant knowledge gaps remain regarding the impact of cholesterol analogs on the lyophilization of mRNA-LNPs. The results of incorporating various C24-alkyl cholesterol analogs into mRNA-LNP formulations, such as their structural and functional effects during and after lyophilization, have yet to be fully elucidated. Moreover, the influence of these analogs on the transfection efficiency across diverse organ-derived cell lines remains poorly understood. Gaining insights into these aspects could facilitate the rational design of mRNA–LNP systems optimized for specific therapeutic applications. Additionally, in our database studies, no researcher has evaluated the therapeutic ability of these phytosterol-containing LNPs in ocular disease models [24]. Existing research has focused primarily on their ability to transfect ocular tissues following intravitreal or subretinal injection [[25], [26], [27], [28], [29]], without assessing functional or therapeutic outcomes.

To address these knowledge gaps, we investigated the structural and functional properties of mRNA-LNP formulations incorporating naturally occurring C24-alkyl phytosterols—including β-sitosterol (Sito), fucosterol (Fuco), campesterol (Camp), and stigmastanol (Stig)—both before and after lyophilization. We systematically evaluated the in vitro transfection efficiency of mRNA-LNPs composed of various cholesterol analogs across a panel of human organ-derived cell lines in vitro (Fig. 1), assessed their in vivo transfection performance in C57BL/6J mice, and examined their therapeutic efficacy in a rat model of retinal degenerative disease.

Our findings revealed that incorporating cholesterol analogs significantly altered mRNA-LNP morphology during lyophilization process, improved the transfection efficiency in specific cells in vitro or organs in vivo, and enhanced the visual function of a retinal degeneration animal model of RCS rats. This is the first study to confirm the vision rescue of Sito-containing mRNA-LNPs in the retinal degeneration animal model of RCS rats by nonintegrating gene therapy (mRNA-LNP treatment). These findings provide extensive insights into the structural modifications of LNPs that drive improved mRNA transfection efficiency, paving the way for the design of next-generation mRNA-LNP systems for ocular gene therapy and regenerative medicine.

2. Materials and methods

2.1. Materials and animal models

Cholesterol, 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC), and 1,2-dimyristoyl-rac-glycero-3-methoxypolyethylene glycol-2000 (DMG-PEG₂₀₀₀) were obtained from Sigma‒Aldrich Trading Co. Ltd. (St. Louis, MO, USA). mRNAs encoding enhanced green fluorescence protein ((eGFP), CAG eGFP mRNA (N′-Me-Pseudo UTP), DD4503-02) and firefly luciferase (FLuc mRNA (N1-Me-Pseudo UTP), DD4511-02) were obtained from Vazyme Biotech Co., Ltd. (Nanjing, China). The Quant-iT RiboGreen Assay kits (cat. #R11490) were purchased from Thermo Fisher Scientific (Waltham, MA, USA). Bright-Glo luciferase assay substrate was purchased from Promega (Madison, WI, USA). HEK-293T, HepG2, A549, PC12, and Y79 cells were purchased from Cellverse (iCell) Bioscience Technology Co., Ltd. (Shanghai, China), and each cell line was authenticated and tested for mycoplasma contamination. The ARPE-19 cell line was a gift from Prof. Quankui Lin at the Eye Hospital of Wenzhou Medical University. Phosphate-buffered saline (PBS), Dulbecco's modified Eagle's medium (DMEM), fetal bovine serum (FBS), and other biomolecules were purchased from Sigma‒Aldrich Trading Co. Ltd. (St. Louis, MO, USA).

RCS rats (rdy−/rdy−) were donated by Prof. Guoping Fan and obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). C57BL/6J mice were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). All animal procedures were approved by the Institutional Animal Care and Use Committee and conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.

2.2. Construction of the (MERTK)mRNA sequence

The coding DNA sequence (CDS) of MERTK, an open reading frame with both start and stop codons (3024 nucleotides containing the ATG start and TGA stop codons, encoding 994 amino acids of the MERTK protein and 8 amino acids of the FLAG protein), was chosen for mRNA synthesis. The sequence information was obtained from the NCBI GenBank database (accession number: AF208235.1), corresponding to the (MERTK)mRNA construct modified with the Cap 1 structure and N1-methylpseudouridine. The sense strand sequence of the (MERTK)mRNA utilized in this investigation is shown in Supplementary Fig. 1. The complete sequence data were sent to Nanjing GenScript Biotech Co., Ltd. (Nanjing, China) for the synthesis of in vitro transcribed (IVT) mRNAs capable of expressing the target MERTK protein. The synthesized (MERTK)mRNA was dissolved in pH 6.4 sodium citrate buffer at a final concentration of 1 mg/mL and stored at −80 °C. The certificate of analysis validating the quality and quantifying the synthesized (MERTK)mRNA is provided in Supplementary Fig. 2.

2.3. Preparation and assay of mRNA-LNPs

FLuc-, eGFP-, or MERTK-encoding mRNAs were entrapped in LNPs via a previously reported method [30,31]. Briefly, an acidic aqueous phase (citrate buffer, 10 mM) in which mRNA (27.4 μg/mL) was dissolved at pH 4.5 was generated, and the organic phase was generated by solubilizing ionizable lipids ((10Z, 13Z)-1-(9Z, 12Z)-9,12-octadecadien-1-yl-10,13-nonadecadien-1-yl ester (DLin-MC3-DMA, MC3) or ((4-hydroxybutyl)azanediyl)bis (hexane-6,1-diyl)bis (2-hexyldecanoate) (ALC-0315)), helper phospholipids (DSPCs), Chol or its analogs (Camp, Fuco, Sito, Stig), and lipid-anchored PEG (DMG-PEG₂₀₀₀) in ethanol at a molar ratio of 50:10:38.5:1.5. Then, the two phases were rapid mixed together at a volume ratio (aqueous solution:ethanol) of 3:1 by either mechanical pipetting mixing or a microfluidic device (Model No.: PG-SYN-FS, PreciGenome LLC, San Jose, CA, USA) to create mRNA-LNPs. Pipette mixing was primarily used for in vitro screening and subretinal injections, where only small volumes (∼1 μL per eye) and high mRNA concentrations (∼200 ng/μL) were required. In this method, lipid and mRNA solutions were rapidly combined in pre-chilled 1.5 mL tubes at room temperature and incubated for 30 min to allow particle self-assembly. Although microfluidic mixing can generate more uniform mRNA-LNPs and is advantageous for large-scale production—particularly for in vivo biodistribution studies after intravenous administration—pipette mixing is more practical for small-volume preparations. It minimizes material loss and yields mRNA-LNPs with comparable size, polydispersity index (PDI), and encapsulation efficiency. However, we note that scale-up using microfluidics may produce results that differ from those obtained with the pipette-mixing method. The mRNA and ionizable lipids were combined at an N:P ratio of 6:1. The generated mRNA-LNPs were then lyophilized via the process described in the next section or stored at 4 °C.

The particle size and PDI value of the mRNA-LNPs were investigated via a NanoZS Zetasizer (Malvern, Worcestershire, U.K.) with a scattering angle of 173° at a temperature of 25 °C. The freshly prepared samples were evaluated immediately after preparation. Lyophilized mRNA-LNPs were reconstituted via the rapid addition of 90 μL of pure water (nuclease-free) and gently mixed. The particle size evaluation was performed with a 10 s run duration, and the number of runs was manually determined. The efficiency of mRNA encapsulation was investigated via a Quant-iT RiboGreen assay kit (Thermo Fisher Scientific, Washington, DC, USA), as previously reported [22].

2.4. Lyophilization of the mRNA-LNPs

We first prepared PBS solutions containing 40% (w/v) cryoprotectant (sucrose). Freshly generated mRNA-LNP solutions were subsequently mixed with PBS containing 40% sucrose (1:1, V/V). The final sucrose concentration of the mRNA-LNP solutions was 20% (w/v). The mRNA-LNP solutions were then separated into two portions. One portion of each freshly generated sample was utilized for experiments immediately. The other portion was lyophilized in a glass container for a half day utilizing a lyophilizer (SCIENTZ-10/C, Ningbo Scientz Biotechnology, Ningbo, China) with a lyophilization process as reported in previous studies [18]. Briefly, the sample was frozen at −30 °C for 3 h, followed by a primary dry cycle at −25 °C in vacuo (5–10 Pa) for 17 h. During the second dry cycle, the sample was subsequently heated at 25 °C in vacuo (20 Pa) for 5 h. Then, lyophilized white samples of the mRNA-LNPs were prepared.

2.5. Fragment analysis by capillary gel electrophoresis

mRNA integrity was assessed using the Fragment Analyzer™ Automated Capillary Electrophoresis System (12-capillary, Agilent, CA, USA) following the manufacturer's protocol, together with a High-Sensitivity RNA Analysis Kit (Agilent Technologies, Santa Clara, CA, USA). Briefly, mRNA samples were diluted to 10 ng/μL and mixed with the provided diluent marker. Samples and RNA ladder were denatured at 70 °C for 2 min and immediately cooled on ice prior to loading. RNA identity was confirmed by the disappearance of the peak following RNase digestion. mRNA integrity was quantified using ProSize Data Analysis Software (Agilent, CA, USA). The area under the curve (AUC) corresponding to the full-length peak (defined as peak length ± 300 bp) was calculated and expressed as a percentage of the total mRNA AUC.

2.6. Cell cultivation

mRNA-LNPs were transfected in vitro into the HEK-293T, HeLa, HepG2, Colo205, ARPE-19, A549, and PC12 cell lines, which were obtained from iCell Bioscience, Inc. (Shanghai, China). Each cell line was cultivated in humidified 5% CO₂ at 37 °C during the investigation. Detailed information on the cell cultivation media, centrifugation conditions and enzyme digestion period for each cell line utilized in this research is summarized in Supplementary Table 1. Cells at passages 4–20 after purchase were utilized for the experiments.

2.7. Investigation of the cellular uptake and mRNA expression of the mRNA-LNPs in vitro

HEK-293T cells exhibiting an epithelial morphology were originally derived from human embryonic kidney tissue and were cultured in DMEM supplemented with 10% FBS at 37 °C in a 5% CO₂ environment via a standard procedure. The cells were detached from polystyrene dishes (TCP, Cat. No. 707003, Wuxi NEST Biotechnology Co., Jiangsu, China) utilizing 0.25% trypsin with EDTA and washed with DMEM by a standard procedure. Then, the HEK-293T cells were inoculated into 24-well TCP dishes (Catalog No. 702002, Wuxi NEST Biotechnology Co., Jiangsu, China) at a density of 2 × 10⁵ cells per well. The cells were cultivated overnight and then transfected with mRNA (eGFP)-LNPs with 800 ng of mRNA per well. eGFP expression and cellular uptake efficiency were evaluated via fluorescence microscopy (Zeiss Model Axio Observer A1; Carl Zeiss, Germany). The fluorescence intensity and transfection efficiency were measured via flow cytometry (BD Accuri™ C6 Plus, BD Biosciences, USA) after two days of transfection with each type of mRNA-LNP. HEK-293T cells treated with mRNA utilizing a commercial reagent (Lipofectamine 2000, Thermo Fisher Scientific, Waltham, MA, USA) were also utilized as a positive control. Following this method, we also evaluated the mRNA transfection efficiencies of the mRNA-LNPs in several other cell lines (Colo205, HeLa, PC12, Y79, ARPE-19 and A549 cells).

2.8. Determination of the in vivo distribution properties of the mRNA-LNPs

All procedures involving animals were reviewed and approved by the Ethics Committee of Eye Hospital, Wenzhou Medical University (Approval No: YSG24110705). The experiments were performed in strict adherence to institutional and national guidelines governing the ethical use of animals in research. C57BL/6J mice, sourced from Beijing Vital River Laboratory Animal Technology (Beijing, China), were used in the study. Lyophilized mRNA (FLuc)-LNPs, previously stored at 4 °C, were rehydrated with UltraPure™ DNase/RNase-Free Distilled Water (Thermo Fisher Scientific, 10,977,023; Waltham, MA, USA) to yield a final mRNA concentration of 0.5 mg/mL. A total volume of 50 μL was administered to each mouse via tail vein injection via an insulin syringe (Covidien). For comparison, an equivalent volume of freshly prepared mRNA-LNP solution (0.5 mg/mL) was administered to a separate group of mice (n = 3 per group).

Six hours following injection, the mice received an intraperitoneal dose of D-luciferin (30 mg/mL) at 150 mg/kg body weight. The animals were euthanized via a CO₂ chamber in combination with 3% isoflurane (Piramal Health Care Limited, Mumbai, India) as an anesthetic at 10–20 min after substrate administration. Major organs—including the heart, liver, spleen, lungs, kidneys, and eyes—were collected post-mortem. Bioluminescence signals from these tissues were immediately captured via the IVIS® in vivo imaging system (PerkinElmer, Waltham, MA, USA) and quantified with Living Image Software v4.7.4. Signal analysis was performed via Living Image v4.5.5.

2.9. Subretinal injection of (mCherry)mRNA-LNPs into C57BL/6J mice

Adult C57BL/6J mice (6–7 weeks old) were anesthetized with an intraperitoneal injection of ketamine (80 mg/kg) and xylazine (10 mg/kg). Pupils were dilated with 1% tropicamide, and subretinal injections were performed using a 33-gauge Hamilton syringe under a surgical microscope. Each eye received 1 μL of mRNA-LNP formulation containing 0.2 μg of reporter mRNA ((mCherry)mRNA). Eyes were treated with topical antibiotic ointment and monitored for retinal detachment recovery after injection. All injections were performed under a surgical microscope by trained personnel, and PBS-injected eyes and untreated eyes were used as controls to differentiate injection-related effects from mRNA-LNPs related responses. Ocular tissues were collected at 48 h post-injection for in vivo fluorescence imaging, frozen section preparation, and immunofluorescence analysis to evaluate the transfection pattern and intensity. The expression levels and distribution of the reporter were compared among those injected with mRNA-LNPs containing cholesterol and its structural analogs.

2.10. Subretinal injection of (MERTK)mRNA-LNPs into RCS rats

Three-week-old RCS rats, a well-established model for studying retinal degeneration [32,33], were used for this investigation. RCS rats were randomly assigned to treatment groups: untreated control (n = 3), (MERTK)mRNA-LNPs(Chol)–treated (n = 3), and (MERTK)mRNA-LNPs(Sito)–treated (n = 4). The animals were housed under a 12-h light–dark cycle with unrestricted access to food and water and were kept in transparent cages that allowed free movement and observation. To perform subretinal delivery of (MERTK)mRNA encapsulated in Sito-containing ALC-0315-based LNPs, the rats were first anesthetized via the intraperitoneal administration of 1% pentobarbital sodium at a dosage of 30 mg/kg. Pupil dilation was achieved via the use of compound tropicamide eye drops (Qiukang, Handan Kangye Pharmaceutical, Hebei, China), followed by topical application of proparacaine hydrochloride (Alcaine, Alcon, Geneva, Switzerland) for local anesthesia. To prevent ocular surface desiccation and potential infection, ofloxacin eye ointment (Dikeluo, Sinqi Pharmaceutical, China) was applied before the procedure.

Under a surgical microscope (M620F20, Leica Microsystems, Wetzlar, Germany), an initial scleral incision was made approximately 1–2 mm posterior to the limbus via a 29-gauge insulin needle. A Hamilton syringe fitted with a 33-gauge blunt needle was then inserted through the incision to deliver 2 μL of (MERTK)mRNA-LNP solution (200 ng/μL) into the subretinal space. To confirm successful subretinal delivery and retinal detachment, a 2% fluorescein solution was added to both the PBS and the mRNA-LNP formulations. Following the injection, the animals were maintained on a heated platform at 37 °C until full recovery.

2.11. Western blot analysis

For validation of MERTK expression in vitro, ARPE-19 and HEK293T cells were exposed to (MERTK)mRNA–LNPs (Sito), with untreated cells serving as negative controls. After 24 h, cells were harvested, and whole-cell protein extracts were prepared using a RIPA-based lysis buffer supplemented with protease inhibitors. Total protein concentration was measured using a Bicinchoninic Acid (BCA) protein quantification assay. Equivalent amounts (40 μg) of protein were subjected to SDS-PAGE analysis and transferred to membranes using standard blotting procedures. Membranes were blocked and incubated with primary antibodies against FLAG (Rabbit Anti-FLAG, ab1162) or MERTK (Rabbit Anti-MERTK, 27900-1-AP). Following incubation with an HRP-conjugated secondary antibody (Goat Anti-Rabbit IgG (H + L), 1,706,515, Bio-Rad), signals were detected using a digital chemiluminescence imaging system (iBright FL1000, Thermo Fisher Scientific).

2.12. Electroretinography evaluation of RCS rats

Three weeks after (MERTK)mRNA-LNP administration, retinal function in the treated RCS rats was evaluated via electroretinography (ERG) (RETI-Port21, Roland, Germany). To prepare for the test, the rats were dark-adapted overnight under light-controlled conditions. Prior to ERG measurements, the animals were anesthetized via an intraperitoneal injection of 1% pentobarbital sodium (30 mg/kg) and placed on a thermostatically controlled heating pad (37 °C) to maintain core temperature during recording. Pupil dilation was achieved by instilling compound tropicamide eye drops (NCD: H20044926, Handan Kangye Pharmaceutical Co., Ltd., Hebei, China) at 5-min intervals for three consecutive applications. Proparacaine hydrochloride eye drops (NCD: HJ20160133, s.a. Alcon-Couvreur n.v., Puurs Belgium) were used to anesthetize the ocular surface, and ofloxacin eye ointment (NCD: H10940177, SHENYANG XINGQI PHARMACEUTICAL CO., LTD., Liaoning, China) was applied to reduce the risk of infection and corneal dryness. Gold wire loop electrodes were positioned on the corneal surface of both eyes to capture the retinal response (active electrodes), whereas subdermal needle electrodes were placed on the cheeks as reference leads. A grounding electrode was inserted subcutaneously into the tail. Scotopic (dark-adapted) ERG recordings were obtained via a series of increasing flash intensities: 0.01, 3.0, and 10.0 cd/m². After data acquisition, the animals were allowed to recover on a 37 °C warming plate. All ERG recordings were conducted under dim red light to preserve dark adaptation.

2.13. Immunofluorescence of retinal cryosections

Eyes from C57BL/6J mice collected 48 h after transplantation and eyes from RCS rats collected 4 weeks after transplantation were fixed in 4% paraformaldehyde, cryoprotected in 10%, 20% and 30% sucrose solutions, embedded in Neg-50 frozen section medium (Epredia), and stored at −80 °C. Whole eyes were cryosectioned near the optic nerve head, and every 2–3 sections (12–14 μm) were collected for analysis. In regions distal to the optic nerve, every 5–6 sections were collected. Retinal cryosections were permeabilized with 0.3% Triton X-100 and blocked before immunostaining.

For MERTK detection, sections near the injection site were selected and incubated with MERTK Rabbit Antibody (Proteintech, 27900-1-AP) and RPE65 Mouse Antibody (Invitrogen, MA1-16578) for 1 h at 25 °C, followed by Alexa Fluor 594–conjugated Goat Anti-Mouse IgG and Alexa Fluor 488–conjugated Goat Anti-Rabbit IgG (1:500; 1 h, RT).

For FLAG detection, sections near the injection site were incubated with FLAG Rabbit Antibody (Abcam, ab1162) for 1 h at 25 °C, followed by Alexa Fluor 488–conjugated Goat Anti-Rabbit IgG (1:500; 1 h, RT).

For inflammatory marker staining, sections adjacent to the injection site—but confirmed under a fluorescence microscope to lack mCherry expression before DAPI counterstaining—were selected. These sections were incubated with GFAP Mouse Antibody (Abcam, ab279290) and Iba1 Rabbit Antibody (Abcam, ab178846) for 1 h at 25 °C, followed by Alexa Fluor 555–conjugated Goat Anti-Mouse IgG and Alexa Fluor 488–conjugated Goat Anti-Rabbit IgG (1:500; 1 h, RT).

All sections were counterstained with DAPI and mounted using antifade medium. Images were acquired with a Zeiss LSM 900 confocal microscope (20 × objectives; Z-stack mode) or detected by the SLIDEVIEW VS200 Research Slide Scanner (Olympus, Tokyo, Japan).

2.14. Hematoxylin and eosin (H&E) staining

The eye sections were first cleared in xylene and gradually brought to an aqueous state using decreasing concentrations of ethanol. Eye sections were then exposed to hematoxylin to label nuclei, followed by rinsing and color development. Afterward, tissues were briefly differentiated and blued before applying eosin to mark cytoplasmic and extracellular structures. Finally, the slides were passed through increasing ethanol concentrations for dehydration, cleared again in xylene, and coverslipped for microscopic examination.

2.15. Statistics and reproducibility

Data analysis was conducted via GraphPad Prism (v10.0.0; GraphPad Software, Boston, MA, USA www.graphpad.com). For comparisons involving multiple groups, ordinary two-way analysis of variance (ANOVA) was used. The results are expressed as the means ± standard deviations (SDs), with statistical significance defined as p < 0.05.

3. Results

3.1. Modification of naturally occurring C24-alkyl cholesterol analogs modulates the structure and morphology of mRNA-LNPs after and before lyophilization

Previous studies have revealed that the incorporation of C24 alkyl derivatives of cholesterol into LNPs can influence their physicochemical properties and significantly increase the efficiency of their in vitro transfection into fibroblasts, 293T cells, HeLa cells, and macrophages [8]. Notably, this enhancement appears to be independent of the cell type used, the specific ionizable lipid used, and the nucleic acid cargo. However, the effects of these C24 alkyl cholesterol derivatives (Sito, Fuco, Camp, and Stig) on LNP stability during lyophilization and their efficiency in applications in vivo remain unclear. To investigate this issue, we formulated mRNA-LNPs incorporating various C24 alkyl cholesterol derivatives, 4-(dimethylamino)-butanoic acid, MC3, DSPC, and DMG-PEG₂₀₀₀, at a molar ratio of 38.5:50:10:1.5 (Supplementary Fig. 3A) and encapsulated them with (eGFP)mRNA via a standard pipette mixing method or a microfluidic device (Model No.: PG-SYN-FS, PreciGenome LLC, San Jose, CA, USA), where standard pipette mixing method was used for preparation of small amount of mRNA-LNPs such as the first screening experiments. Physicochemical properties of the freshly generated and lyophilized mRNA-LNPs prepared with MC3 were investigated, and the results are shown in Supplementary Fig. 3.

We found that the inclusion of C24 alkyl groups in cholesterol analogs appears to greatly influence the stabilities of mRNA-LNP during lyophilization (Supplementary Fig. 3B–3G), and induce minor defects in mRNA-LNP bilayer organization (Supplementary Fig. 3, Supplementary Table 2). These defects may compromise the structural integrity of mRNA-LNPs during lyophilization.

We also prepared mRNA-LNPs using another different ionizable lipid, ALC-0315, which is designed as ALC-0315-based mRNA-LNPs to determine whether the effects of various C-24 alkyl cholesterol derivatives (Fuco, Sito, Camp, Stig, and Chol (control)) on the physicochemical characteristics and transfection efficiencies of mRNA-LNPs were dependent on the type of ionizable lipid both before and after lyophilization (Fig. 2A).

Fig. 2.

Lipids used to fabricate mRNA-LNPs and the basic physicochemical properties of LNP-encapsulated (eGFP)-mRNA with various naturally occurring C-24 alkyl cholesterol derivatives before and after lyophilization, in which ALC-0315 was used as an ionizable lipid. (A) Structural features of naturally occurring C-24 alkyl cholesterol derivatives. (B-F) The nanoparticle size (B), PDI values (C), zeta potential (D), morphology (E) and encapsulation efficiencies (F) of LNPs encapsulating (eGFP)-mRNA with ALC-0315 as the ionizable lipid and various naturally occurring C-24 alkyl cholesterol derivatives before and after lyophilization, as observed via DLS, TEM, and the Quant-iT RiboGreen Assay kit. Scale bar: 100 nm. All the data are shown as the means ± s.d.s (n ≥ 3). Statistical significance was evaluated using two-way ANOVA followed by Sidak’s multiple comparisons test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, ns indicates no significant difference. Three representative investigations are presented. (G) Fragment analysis of RNA identity and integrity for in vitro–transcribed (eGFP)-mRNA encapsulated in ALC-0315–based mRNA–LNPs formulated with Chol, Camp, Stig, Sito, or Fuco before and after lyophilization. All samples exhibited a dominant full-length mRNA peak (∼1000 nt, left peak) without smeared patterns indicative of degradation. A minor secondary peak at ∼1800–2000 nt (right peak) was observed in some formulations. LM, lower marker. In these mRNA-LNPs, the molar ratio of the four lipid components, ALC-0315:DMG-PEG₂₀₀₀:DSPC:cholesterol analogs, was 50:1.5:10:38.5.

The diameters of the freshly generated ALC-0315-based mRNA-LNPs formulated with Sito, Fuco, Camp, Stig, or Chol ranged from 60 nm to 133 nm (Fig. 2B), with average PDI values below 0.3 for all formulations except mRNA–LNPs(Fuco), which showed a slightly higher PDI of ∼0.4 (Fig. 2C). The diameters of all reconstituted ALC-0315-based mRNA-LNPs were markedly greater after lyophilization than those of freshly generated mRNA-LNPs (Fig. 2B). Despite this increase, formulations containing Camp, Stig, or Sito maintained average PDI values below 0.3, indicating relatively stable particle distributions after rehydration (Fig. 2C). In contrast, ALC-0315-based mRNA-LNPs(Fuco) exhibited the greatest change in size, increasing from 111 nm to 557 nm in average diameter after lyophilization, with sharp increases in PDI after reconstitution, rising from ∼0.2 when freshly prepared to as high as ∼0.94, suggesting significant particle destabilization upon reconstitution (Fig. 2C).

Moreover, the zeta potential of ALC-0315-based mRNA-LNPs(Fuco) shifted from −2 mV to 5 mV after lyophilization (Fig. 2D), accompanied by the most pronounced increase in particle size. These results indicate that ALC-0315-based mRNA-LNPs(Fuco) are unstable under lyophilization stress and likely undergo surface molecular rearrangement and particle aggregation. The diameters of lyophilized and reconstituted ALC-0315-based mRNA-LNPs formulated with Sito, Camp, Stig, and Chol (control) ranged between 148 nm and 266 nm. Among these mRNA-LNPs, ALC-0315-based mRNA-LNPs(Stig) showed minimal change in size after lyophilization, which indicated that the structures of ALC-0315-based mRNA-LNPs(Stig) are more stable than those of the other LNPs under lyophilization-induced stress.

TEM images revealed that freshly prepared ALC-0315-based mRNA-LNPs(Chol) (control) maintained a regular spherical morphology (Fig. 2E, Supplementary Fig. 4). However, the morphology of reconstituted ALC-0315-based mRNA-LNPs(Chol) was not observed via TEM after lyophilization, suggesting severe structural disruption. In contrast, ALC-0315-based mRNA-LNPs containing other C-24 alkyl cholesterol derivatives (mRNA-LNPs(Camp), mRNA-LNPs(Stig), mRNA-LNPs(Sito), and mRNA-LNPs(Fuco)) displayed faceted morphologies both before and after lyophilization, which is similar to the morphological properties of corresponding mRNA-LNPs prepared with MC3 as an ionizable lipid (Supplementary Fig. 3E).

Encapsulation efficiency analysis further supported lyophilization-induced structural alteration (Fig. 2F). Freshly prepared LNPs exhibited encapsulation efficiency values ranging from 40 to 80%, whereas reconstituted lyophilized LNPs showed dramatically reduced encapsulation efficiency, with values declining to below 20%.

To directly assess mRNA integrity, mRNA was extracted from LNP formulations after lyophilization and analyzed using capillary gel electrophoresis. All samples exhibited a dominant full-length mRNA peak (∼1000 nt) and lacked smeared signals typically associated with RNA degradation (Fig. 2G). A minor peak at ∼1800–2000 nt was detected in some formulations, consistent with dimeric or multimeric mRNA species rather than degradation products. The ratios of the dominant full-length mRNA peak (∼1000 nt) and the minor peak (∼1800–2000 nt) are summarized in Supplementary Table 3. Notably, the proportion of the minor peak (∼1800–2000 nt) increased substantially after reconstitution, suggesting a tendency toward aggregation. These results indicate that the reduced encapsulation efficiencies observed after lyophilization are primarily attributable to nanoparticle structural destabilization rather than mRNA degradation.

3.2. Influence of cholesterol analogs and ionizable lipids in mRNA-LNPs on the transfection efficiency of several cell lines derived from different organs

To evaluate how these physicochemical differences translate into biological performance, we next examined the transfection efficiency of the mRNA–LNP formulations in HEK-293T cell line. An obviously greater eGFP expression intensity was observed in the HEK-293T cells transfected with freshly prepared mRNA-LNPs(Camp), mRNA-LNPs(Stig), mRNA-LNPs(Sito), or mRNA-LNPs(Fuco), which all presented higher eGFP expression intensities compared to the corresponding mRNA-LNPs(Chol) (control) (Fig. 3A–D). The eGFP expression intensity in the HEK-293T cells treated with freshly prepared mRNA-LNPs(Fuco) was the highest among the cells treated with ALC-0315-based mRNA-LNP formulations (Fig. 3D).

Fig. 3.

The expression of eGFP (green) in HEK-293T cells transfected with designed ALC-0315 based freshly generated (fresh) or lyophilized (eGFP)mRNA-LNPs with various cholesterol analogs before and after lyophilization. (A) Fluorescence images of HEK-293T cells treated with each designed mRNA-LNP. The cell nuclei of HEK-293T cells were stained with DAPI (blue). The scale bar indicates 50 μm. (B) Representative flow cytometry analysis showing the transfection efficiency of each designed mRNA-LNP. (C & D) Transfection efficiency (C) and eGFP expression intensity (D) in HEK-293T cells treated with each designed mRNA-LNP, as evaluated by flow cytometry. All the data are presented as the mean ± s.d. (n = 3). Statistical significance was analyzed by one-way ANOVA. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗∗p < 0.0001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

The transfection efficiencies of all freshly prepared ALC-0315-based mRNA-LNPs (mRNA-LNPs(Chol) (control), mRNA-LNPs(Camp), mRNA-LNPs(Stig), mRNA-LNPs(Sito), and mRNA-LNPs(Fuco)) into HEK-293T cells were nearly 100% (Fig. 3C).

The eGFP expression intensities in the HEK-293T cells transfected with lyophilized and subsequently reconstituted forms of ALC-0315-based mRNA-LNPs(Fuco), mRNA-LNPs(Stig) and mRNA-LNPs(Camp) were greater than those in the cells treated with mRNA-LNPs(Chol) (control) (Fig. 3D). The transfection efficiencies of the ALC-0315-based mRNA-LNPs (Chol), mRNA-LNPs(Camp), mRNA-LNPs(Stig), mRNA-LNPs(Sito), and mRNA-LNPs(Fuco) into HEK-293T cells ranged from 80% to 92% following lyophilization and reconstitution (Fig. 3C). The HEK-293T cells transfected with lyophilized and subsequently reconstituted ALC-0315-based mRNA-LNPs(Fuco) also presented the highest transfection efficiency, as did the cells transfected with freshly prepared mRNA-LNPs(Fuco).

To further investigate the potential relationship between the lipid composition of mRNA-LNPs and their ability to be transfected selectively into organs or tissues, we evaluated the transfection efficiency of mRNA-LNPs formulated with various ionizable lipids (MC3 or ALC-0315) and cholesterol analogs (Chol, Camp, Stig, Sito, and Fuco) in a panel of cell lines derived from various organs (uterus, liver, colon, eye, lung, and neuron) (Fig. 4A). This approach was designed to identify lipid formulations that enable efficient transfection of specific organ-derived cells, thereby providing promising mRNA‒LNP combinations for future research on localized drug delivery therapies.

Fig. 4.

Comparison of the transfection efficiencies of mRNA-LNPs formulated with various ionizable lipids (MC3 or ALC-0315) and cholesterol analogs (Camp, Stig, Sito, Fuco, and Chol (control)) in various cell lines obtained from different organs (uterus, colon, liver, eye, lung, and nervous system). (A) Schematic diagram of selected cell lines obtained from different organs. (B–I) eGFP expression intensity (B-G) and transfection efficiency (H and I) in each cell type—human lung adenocarcinoma cells (A549), human hepatocellular carcinoma cells (HepG2), human colon cancer cells (Colo205), rat adrenal medulla-derived pheochromocytoma cells (PC12, neuronal), and human retinal pigment epithelium cells (ARPE-19)—transfected with freshly prepared (H) or lyophilized and reconstituted (I) MC3-based mRNA-LNPs. (J–Q) eGFP expression intensity (J-O) and transfection efficiency (P and Q) in each cell type (HeLa, HepG2, Colo205, ARPE-19, A549, and PC12) transfected with freshly prepared (P) or lyophilized and reconstituted (Q) ALC-0315-based mRNA-LNPs. All the data are shown as the means ± s.d.s (n = 3).

The following cell lines were selected to represent different organs (Fig. 4A): ARPE-19 (human retinal pigment epithelium), A549 (human lung adenocarcinoma), HepG2 (human hepatocellular carcinoma), Colo205 (human colon cancer), HeLa (human cervical cancer), and PC12 (rat adrenal medulla-derived pheochromocytoma), a cell line typically used as model nerve cells [34,35]. Transfection efficiencies and eGFP expression intensities were evaluated for all selected cell types. Fig. 4B–Q shows eGFP expression and transfection efficiency in each cell line transfected with mRNA-LNPs containing various combinations of ionizable lipids (MC3 or ALC-0315) and cholesterol analogs (Chol, Camp, Stig, Sito, and Fuco). Fig. 4H, I, P, and 4Q summarized the transfection efficiency of different cholesterol-analogue–containing mRNA-LNPs across multiple organ-derived cell lines before and after lyophilization, presented as radar plots.

For the HeLa and HepG2 cell lines (Fig. 4, Supplementary Fig. 5), the transfection efficiencies of all the tested mRNA-LNPs were nearly 100%, regardless of the ionizable lipid type (MC3 or ALC-0315) or cholesterol analog (Chol, Camp, Stig, Sito, or Fuco), both before and after lyophilization. The HeLa cells transfected with freshly prepared mRNA-LNPs(Sito) presented the highest eGFP intensity (Fig. 4B and J, Supplementary Fig. 5A), whereas the HeLa cells transfected with freshly prepared mRNA-LNPs(Chol) presented the lowest eGFP intensity (Fig. 4B and J, Supplementary Fig. 5A and 5B), regardless of the type of ionizable lipid (MC3 or ALC-0315). This trend in HeLa cells was consistent with that in HEK-293T cells treated with freshly generated mRNA-LNPs formulated with ionizable lipids (ALC-0315) and cholesterol analogs (Chol, Camp, Stig, Sito, and Fuco) (Fig. 3B).

The HeLa cells transfected with lyophilized and reconstituted mRNA-LNPs(Chol) yielded the highest eGFP intensity (Fig. 4B and J), and no other formulation surpassed this level when MC3 was utilized as an ionizable lipid (Supplementary Fig. 5A). When ALC-0315 was utilized as an ionizable lipid, HeLa cells treated with lyophilized and reconstituted mRNA-LNPs(Sito) yielded similar eGFP expression intensities to those obtained using lyophilized and reconstituted mRNA-LNPs(Chol) (control) (Fig. 4J, Supplementary Fig. 5B).

In HepG2 (liver cancer) cell lines (Fig. 4C and K), treatment with freshly prepared mRNA-LNPs(Sito) resulted in the highest eGFP intensity, which was similar to that of freshly prepared mRNA-LNPs(Chol) (Fig. 4C and K, Supplementary Fig. 5C and 5D), whereas cells treated with freshly prepared mRNA-LNPs(Stig) presented the lowest eGFP intensity, irrespective of the ionizable lipid (MC3 or ALC-0315) used.

For Colo205 (colon cancer) cells treated with freshly prepared MC3-based mRNA-LNPs (Fig. 4D, H and 4I), the transfection efficiencies of MC3-based mRNA-LNPs(Sito, Fuco, or Camp) were all nearly 100% (Fig. 4H), markedly greater than those of MC3-based mRNA-LNPs(Chol) (control) and MC3-based mRNA-LNPs(Stig), with transfection efficiencies of 0.5% and 39.1%, respectively (Fig. 4H). After lyophilization (Fig. 4I), reconstituted and lyophilized MC3-based mRNA-LNPs (Fuco, β-Sito, Stig, Camp, or Chol (control)) all presented significantly reduced transfection efficiencies (1–14.4%) into Colo205 cells, but MC3-based mRNA-LNPs(Stig) presented the highest transfection efficiency (14.4%), and MC3-based mRNA-LNPs(Chol) presented the lowest transfection efficiency (1%) into Colo205 cells.

Colo205 cells treated with freshly prepared ALC-0315-based mRNA-LNPs (Fuco, β-Sito, Stig, or Camp) all expressed greater eGFP intensity than those treated with mRNA-LNPs(Chol) (control) (Fig. 4L), and Colo205 cells treated with freshly prepared ALC-0315-based mRNA-LNPs(Sito) expressed the highest eGFP intensity (Fig. 4L). The transfection efficiencies of all freshly prepared ALC-0315-based mRNA-LNPs (Fuco, β-Sito, Stig, or Camp) in Colo205 cells were nearly 100% (Fig. 4P), which was obviously greater than that of the mRNA-LNPs(Chol) (control), which had a transfection efficiency of 67.9%. After lyophilization, the transfection efficiencies of all reconstituted and lyophilized ALC-0315-based mRNA-LNPs (Fuco, Sito, Stig, Camp, or Chol (control)) into Colo205 cells decreased, ranging from 5.4% to 56.3% (Fig. 4Q), and Colo205 cells transfected with lyophilized and reconstituted ALC-0315-based mRNA-LNPs(Stig) presented the highest eGFP expression among the cells treated with any other type of lyophilized and reconstituted ALC-0315-based mRNA-LNPs (Fuco, β-sito, Camp, or Chol) (Fig. 4L).

For ARPE-19 cells treated with MC3-based mRNA-LNPs (Fig. 4E, H and 4I, Supplementary Fig. 5E), the transfection efficiencies of freshly prepared MC3-based mRNA-LNPs(Fuco) and mRNA-LNPs(β-Sito) were both nearly 100% and greater than those of MC3-based mRNA-LNPs(Chol) (control), MC3-based mRNA-LNPs(Stig), and MC3-based mRNA-LNPs(Camp), which had transfection efficiencies of 44.1%, 38.3% and 29.0%, respectively (Fig. 4H). After lyophilization (Fig. 4I), the reconstituted lyophilized MC3-based mRNA-LNPs(Stig) presented the highest transfection efficiency into ARPE-19 cells (50.6%), whereas the transfection efficiencies of the reconstituted lyophilized mRNA-LNPs (Chol, Camp, Sito, and Fuco) significantly decreased after lyophilization, ranging from 2% to 8.6%.

Among ARPE-19 cells treated with ALC-0315-based mRNA-LNPs (Fig. 4M, P and 4Q, Supplementary Fig. 5F), ARPE-19 cells transfected with freshly prepared ALC-0315-based mRNA-LNPs (Fuco or Camp) showed eGFP intensities comparable to those of cells transfected with freshly prepared ALC-0315-based mRNA-LNPs(Chol) (control) (Fig. 4M). After lyophilization, ARPE-19 cells treated with reconstituted lyophilized ALC-0315-based mRNA-LNPs (β-Sito and Stig) expressed significantly higher eGFP intensities than those of cells treated with ALC-0315-based mRNA-LNPs(Chol) (Fig. 4M, Supplementary Fig. 5F). The transfection efficiency of all ALC-0315-based mRNA-LNPs (Chol, Camp, Stig, Sito, and Fuco) was nearly 100% in ARPE-19 cells both before and after lyophilization (Fig. 4P and Q).

In A549 cells treated with MC3-based mRNA-LNPs (Fig. 4F, H and 4I), freshly prepared MC3-based mRNA-LNPs (Fuco, β-Sito, Stig, or Camp) outperformed MC3-based mRNA-LNPs(Chol), which presented negligible transfection efficiency (3.6%) in A549 cells (Fig. 4F). Among the freshly prepared MC3-based mRNA-LNPs tested, only mRNA-LNPs(Sito) had nearly 100% transfection efficiency into A549 cells (Fig. 4H). After lyophilization, reconstituted MC3-based mRNA-LNPs(Camp) and mRNA-LNPs(Stig) presented relatively high transfection efficiency (19.1% and 9.6%, respectively) in A549 cells (Fig. 4I), whereas reconstituted MC3-based mRNA-LNPs (Fuco, β-Sito, Stig, or Chol) presented negligible transfection capability in A549 cells.

In A549 cells treated with ALC-0315-based mRNA-LNPs (Fig. 4N, P and 4Q), the transfection efficiencies of freshly prepared ALC-0315-based mRNA-LNPs(Stig), mRNA-LNPs(Sito), and mRNA-LNPs(Camp) were 84.8%, 94.4%, and 100%, respectively (Fig. 4P), which were obviously higher than those of ALC-0315-based mRNA-LNPs(Chol) (control), which were 31.1%. However, the transfection efficiency of reconstituted lyophilized ALC-0315-based mRNA-LNPs(Sito) into A549 cells dramatically decreased from 94.4% to 30.7% after lyophilization, whereas the transfection efficiency of reconstituted lyophilized ALC-0315-based mRNA-LNPs(Camp) was 100% even after lyophilization (Fig. 4Q).

In PC12 cells (Fig. 4G-I and 4O-Q), the transfection efficiencies of all freshly prepared or lyophilized and reconstituted mRNA-LNPs (Fuco, β-Sito, Stig, Camp, or Chol (control)) into PC12 cells were less than 68%, regardless of the ionizable lipid type (MC3 or ALC-0315). Overall, ALC-0315-based mRNA-LNPs (Fuco, β-Sito, Stig, Camp, or Chol (control)) generally presented higher transfection efficiencies in this neuronal cell line than MC3-based mRNA-LNPs did, both before and after lyophilization.

Overall, we found that ALC-0315-based mRNA-LNPs generally demonstrated superior transfection performance over MC3-based mRNA-LNPs, regardless of the type of cholesterol analog (Fuco, β-Sito, Stig, Camp, or Chol (control)) used, both before and after lyophilization. This conclusion was based on comparative analysis of the transfection efficiencies of multiple cell types (HeLa, HepG2, Colo205, ARPE-19, A549, and PC12) derived from various tissues (uterus, liver, colon, eye, lung, and neuro). Compared with the freshly prepared mRNA-LNPs formulated with other C-24 alkyl cholesterol derivatives (Fuco, Stig, Camp, or Chol (control)), regardless of the type of ionizable lipids (MC3 or ALC-0315) used, the cholesterol analogs Sito exhibited the highest transfection efficiency in most of the tested cell lines (HeLa, HepG2, Colo205, and A549). On the other hand, after lyophilization and reconstitution, MC3- or ALC-0315-based mRNA-LNPs(Stig) presented the highest transfection efficiencies in most of the tested cell lines (HepG2, Colo205, and ARPE-19) compared with those of the mRNA-LNPs formulated with other C-24 alkyl cholesterol derivatives, including cholesterol.

3.3. In vivo distribution of mRNA-LNPs formulated with various cholesterol analogs and ionizable lipids

We evaluated the in vivo distribution of firefly luciferase (FLuc)-encoded mRNA-LNPs formulated with various ionizable lipids (MC3 or ALC-0315) and cholesterol analogs (Chol, Camp, Stig, Sito, and Fuco) following intravenous (i.v.) administration in C57BL/6J mice (Fig. 5A). Bioluminescence signals in several main organs—including the spleen, liver, heart, lungs, eyes, and kidneys (Supplementary Figs. 6 and 7)—were assessed via an in vivo imaging system (IVIS) to evaluate the bioluminescent signals at 6 h post injection. All the mRNA-LNP formulations were predominantly transfected into the spleen and liver both before and after lyophilization, with no detectable signals observed in the eyes, kidneys, or heart (Fig. 5B and C). Notably, certain mRNA-LNP formulations were also transfected into the lungs, particularly lyophilized and reconstituted MC3-based mRNA-LNPs(Fuco).

Fig. 5.

In vivo biodistribution and bioluminescence quantification of mRNA(FLuc)-LNPs containing various cholesterol analogs and ionizable lipids. (A) Schematic overview. (B–C) Representative IVIS images of C57BL/6J mice 6 h after intravenous (i.v.) injection of freshly prepared (B) or lyophilized and reconstituted (C) mRNA-LNPs formulated with either MC3 or ALC-0315 and different cholesterol analogs. (D–G) Quantification of average radiance (photons/s/cm²/sr) in the spleen, liver, and lungs of mice treated with freshly prepared (D and F) or lyophilized and reconstituted (E and G) MC3-based mRNA-LNPs (D and E) and ALC-0315-based mRNA-LNPs (F and G). (H–K) Normalized average radiance values in each organ from mice receiving C24-alkyl-substituted cholesterol analog-containing MC3-based mRNA-LNPs (H and I) and ALC-0315-based mRNA-LNPs (J and K), which were freshly prepared (H and J) or lyophilized and reconstituted (I and K), expressed relative to cholesterol-based LNPs. The data are shown as the means ± s.d.s (n = 3 biologically independent mice per group). Statistical significance was evaluated via two-way ANOVA followed by Tukey's post hoc test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Quantitative analysis of the extracted organs was conducted to measure the average radiance (Avg Radiance [p/s/cm²/sr]) in the liver, spleen, and lungs (Fig. 5D–G, and Supplementary Fig. 8). In mice treated with freshly prepared MC3-based mRNA-LNPs (Fig. 5D), the average radiance ranged from 7.8 × 10⁵ to 7.0 × 10⁶ in the liver, from 1.6 × 10⁵ to 1.5 × 10⁶ in the spleen, and from 3.5 × 10⁴ to 1.1 × 10⁵ in the lungs. For freshly prepared ALC-0315-based mRNA-LNPs (Fig. 5F), the average radiance ranged from 6.6 × 10⁶ to 2.5 × 10⁷ in the liver, from 6.1 × 10⁵ to 3.2 × 10⁶ in the spleen, and from 2.9 × 10⁴ to 2.9 × 10⁵ in the lungs. The total radiance in the liver, spleen, and lungs indicated that ALC-0315-based mRNA-LNPs induced stronger bioluminescence signal expression than MC3-based mRNA-LNPs did (Supplementary Fig. 8) both before and after lyophilization.

To assess the role of C-24 alkyl cholesterol analogs (Fuco, β-Sito, Stig, Camp, or Chol (control)) in the in vivo translatable efficiencies of mRNA-LNPs, the Avg Radiance of the liver, spleen, and lungs transfected with mRNA-LNPs (Fuco, β-Sito, Stig, or Camp) was normalized to that of mRNA-LNPs(Chol (control)) (Fig. 5H–K). For MC3-based mRNA-LNPs, freshly prepared MC3-based mRNA-LNPs(Fuco) were used to transfect the liver, and the bioluminescence signal in the liver was significantly greater than that in the liver transfected with MC3-based mRNA-LNPs (β-Sito, Stig, Camp, or Chol (control)) (Fig. 5H). After lyophilization, lyophilized and reconstituted MC3-based mRNA-LNPs(Fuco) could also transfect lungs after i.v. injection (Fig. 5I). The normalized bioluminescence signal in lungs transfected with lyophilized and reconstituted MC3-based mRNA-LNPs(Fuco) was significantly greater than that in lungs transfected with other lyophilized and reconstituted MC3-based mRNA-LNPs (β-Sito, Stig, Camp, or Chol (control)) (Fig. 5I).

The total bioluminescence signal expression levels in the liver, spleen, and lungs (Supplementary Fig. 8) further confirmed the superior performance of freshly prepared MC3-based mRNA-LNPs(Fuco) over MC3-based mRNA-LNPs (β-Sito, Stig, Camp, or Chol (control)).

For ALC-0315-based mRNA-LNPs (Fig. 5J and K), the normalized bioluminescence signals (Fig. 5J) in lungs transfected with freshly prepared ALC-0315-based mRNA-LNPs(Sito) were markedly greater than those in lungs transfected with mRNA-LNPs(Chol) (control). In addition, the normalized bioluminescence signal level (Fig. 5K) in lungs transfected with lyophilized and reconstituted ALC-0315-based mRNA-LNPs(Sito) was obviously greater than that in lungs transfected with mRNA-LNPs(Chol) (control).

The total bioluminescence signals (Supplementary Fig. 8) in the liver, spleen and lungs further indicated that only transfection with freshly prepared ALC-0315-based mRNA-LNPs(Sito) had a comparable in vivo translatable ability to that of mRNA-LNPs(Chol) (control), and the total bioluminescence signals for other freshly prepared ALC-0315-based mRNA-LNPs (Stig, Camp, or Fuco) presented lower total bioluminescence signals in the organs than did those for mRNA-LNPs(Chol) (control).

In conclusion, the in vivo translatable ability of MC3-based mRNA-LNPs(Fuco) and ALC-0315-based mRNA-LNPs(Sito) was obviously superior to that of their respective mRNA-LNPs(Chol) controls, especially in terms of their translatability to the lungs. Compared with MC3-based mRNA-LNPs, ALC-0315-based mRNA-LNPs generally exhibited superior in vivo translatability.

3.4. Retinal distribution of designed (mCherry)mRNA–LNP formulations following subretinal injection

To determine whether the transfection performance of the designed mRNA–LNP formulations differs between systemic and local delivery routes, we compared representative mRNA–LNPs following subretinal injection in C57BL/6J mice. LNPs containing (mCherry)mRNA and incorporating distinct cholesterol analogs were administered at equivalent mRNA doses (200 ng) (Fig. 6A and Supplementary Fig. 9).

Fig. 6.

Retinal biodistribution and inflammatory response followed by subretinal delivery of ALC-0315 based mRNA(mCherry)-LNPs containing various cholesterol analogs. (A) Schematic illustration of subretinal delivery of (mCherry) mRNA–LNPs to the retina and the downstream experimental steps. (B) Representative confocal images ( × 20) of retinal cross-sections collected 48 h after subretinal injection of ALC-0315 based mRNA (mCherry)-LNPs containing various cholesterol analogs. mCherry fluorescence (red) and DAPI nuclear staining (blue) indicate the spatial distribution of reporter expression across retinal layers. The scale bar represents 20 μm. (C) Representative confocal images ( × 20) of retinal cross sections immunostained for Iba1 (microglia; green), GFAP (astrocytes; magenta), and DAPI (blue) to assess local inflammatory responses. Iba1 immunostaining shows microglial activation levels. GFAP immunostaining demonstrates Müller cell gliosis assessment. The scale bar represents 20 μm. Sample sizes: Untreated (n = 3); Chol and Sito (n = 3); Camp, Stig, and Fuco (n = 3). GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Confocal imaging of retinal cryosections (Fig. 6B) demonstrated that the ALC-0315–based Sito-containing mRNA–LNP, which was identified previously as a top-performing system under intravenous administration, also achieved strong mCherry expression in the outer nuclear layer (ONL) and retinal pigment epithelium (RPE) after subretinal delivery. The ALC-0315–based Fuco-containing mRNA–LNP mediated similarly prominent mCherry expression, largely localized to the layer of photoreceptor outer segments. In contrast, ALC-0315-based mRNA–LNPs incorporating Stig, Camp, or Chol produced only modest mCherry expression in the retina.

For MC3-based mRNA–LNP formulations incorporating the same cholesterol analogs (Camp, Stig, Sito, Fuco, or Chol), mCherry expression was overall consistently lower than that of their ALC-0315–based counterparts (Supplementary Fig. 10A and Supplementary Fig. 11). Among the MC3 formulations, Fuco-containing mRNA–LNPs yielded the strongest expression in the layer of photoreceptor outer segments and RPE layer. These findings suggest that the relative transfection preference of cholesterol analogs observed after systemic administration is largely preserved in the subretinal space, although the local ocular environment modulates overall expression magnitude.

To evaluate retinal immune responses, sections were stained for Iba1 (microglia) and GFAP (astrocytes) (Fig. 6C, Supplementary Fig. 10B). Minimal microglial or astrocytic activation was observed in eyes treated with ALC-0315–based Sito- or Chol-containing mRNA-LNPs, indicating favorable biocompatibility. In contrast, Camp- and Fuco-containing ALC-0315-based mRNA-LNPs induced pronounced microglial activation across the retina. Stig-containing LNPs triggered astrocyte activation in the ganglion cell layer (GCL) and RPE, whereas Fuco-containing mRNA-LNPs elicited GFAP upregulation in the GCL. Notably, MC3–based Fuco-containing mRNA-LNPs produced the strongest and most widespread immune activation (Supplementary Fig. 10B), which indicates a higher immunogenic potential.

H&E staining was performed to assess retinal cytoarchitecture and to evaluate potential tissue responses to subretinal delivery of the cholesterol-modified mRNA-LNPs (Supplementary Fig. 12). We can find that overall retinal lamination was preserved across all treatment groups. The GCL, INL, ONL, and photoreceptor outer segment remained clearly distinguishable. Small, vacuole-like spaces at the photoreceptor–RPE interface appeared in some sections; these are consistent with subretinal bleb formation during injection and are commonly observed technical artifacts rather than indicators of tissue damage. Importantly, these spaces are not accompanied by RPE disruption, inflammatory infiltrates, or structural degeneration, which indicates that subretinal administration of the cholesterol-modified mRNA–LNPs did not induce detectable retinal toxicity.

Overall, the ALC-0315–based Sito-containing mRNA–LNP emerged as the formulation that achieved robust retinal transfection by maintaining minimal immune activation with no detectable structural toxicity, based on integrated evaluations of mCherry expression, retinal inflammation, and histological outcomes. Consequently, this system was selected for subsequent therapeutic testing in vision-rescue studies.

3.5. In vivo validation of the therapeutic effect of optimal LNPs encapsulating mRNAs encoding the MERTK protein for the treatment of retinal degeneration

To deeply assess the therapeutic ability of the screened optimal ALC-0315-based mRNA-LNPs(Sito) for delivering mRNA in specific eye disease treatment, a classical retinal degeneration disease model, RCS rats carried a spontaneous recessive mutation in the MER proto-oncogene tyrosine kinase (MERTK) gene was selected [36]. Restoration of MERTK expression in RPE cells has demonstrated significant rescue of photoreceptor survival in preclinical models [37]. Therefore, MERTK was selected as the therapeutic target to evaluate the efficacy of mRNA-LNP–mediated gene restoration in MERTK-deficient retinal degeneration.

We formulated LNPs by encapsulating (MERTK)mRNA with ALC-0315, and β-Sito, DSPC, and DMG-PEG2K as the ionizable, structural, and PEG lipids, respectively. As a control, we prepared (MERTK)mRNA encapsulated by LNPs formulated with ALC-0315, Chol, DSPC, and DMG-PEG2K.

To enable specific detection of exogenous MERTK translated from the administered IVT (MERTK)mRNA, a FLAG tag was incorporated immediately downstream of the MERTK coding sequence. Western blotting was performed using ARPE-19 and HEK293T cells transfected with (MERTK)mRNA-LNPs (Sito) at first. Clear MERTK and FLAG protein bands were detected in both cell lines 24 h after transfection, whereas no FLAG signal appeared in untreated controls (Supplementary Fig. 13). These findings confirm that exogenous MERTK protein is expressed and detectable in vitro. Future studies will further explore the durability of expression across longer time frames. The biodistribution of mRNA-LNPs following i.v. administration primarily results in delivery to the liver, spleen, and lungs (Fig. 5), as indicated by the use of luciferase mRNA as a reporter. Therefore, i.v. injection is considered unsuitable for the delivery of (MERTK)mRNA to the targeted retina of the eye. Our studies demonstrated that subretinal administration of sito-containing (mCherry)mRNA-LNPs leads to an increase in the mCherry signal at 48 h post administration [38]. Thus, subretinal injection is considered a more direct and appropriate method for treating retinal diseases in the RCS animal model and was used in subsequent studies (Fig. 7A).

Fig. 7.

Subretinal delivery of (MERTK)mRNA-LNPs and visual functional rescue in a retinal degeneration disease model in RCS rats. (A) Schematic illustration of subretinal injection for targeted delivery of (MERTK)mRNA-LNPs to the retina. (B) Representative OCT images confirming successful subretinal delivery, indicated by the presence of a retinal bleb at the injection site. (C) Scotopic ERG responses measured at 3 weeks post injection showing greater b-wave amplitudes in RCS rats treated with (MERTK)mRNA-LNPs(Sito) than in those treated with (MERTK)mRNA-LNPs(Chol) and noninjected controls at stimulus intensities of 10.0, 3.0, and 0.01 cd s/m². (D) Quantitative analysis of b-wave amplitudes showing a significant improvement in visual function in the (MERTK)mRNA-LNP(Sito)-treated rats, particularly at 3.0 and 10.0 cd s/m² (∗p < 0.05). Statistical significance was evaluated via paired comparison followed by Holm-Bonferronl test. (E) Immunofluorescence staining of retinal cross-sections revealed robust MERTK protein expression in both the RPE and ONL layers in the (MERTK)mRNA-LNP(Sito)-treated group at 4 weeks post injection, whereas no MERTK expression was detected in noninjected controls or (MERTK)mRNA-LNPs(Chol)-treated group. The scale bar represents 20 μm in panels without individual scale annotations. Sample sizes: untreated group (n = 3); (MERTK)mRNA-LNPs(Chol)-treated group (n = 3); (MERTK)mRNA-LNP(Sito)-treated group (n = 4). GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer.

Optical coherence tomography (OCT) imaging was utilized after subretinal injection to confirm the successful localization of (MERTK)mRNA-LNPs in the subretinal space and to assess the retinal structure (Fig. 7B). The appearance of a bleb in the OCT images confirmed the successful subretinal delivery of (MERTK)mRNA-LNPs (Fig. 7B).

To evaluate visual function, we recorded scotopic electroretinography (ERG) waves from age-matched RCS rats at 3 weeks post injection. The animals received (MERTK)mRNA-LNPs formulated with either β-sitosterol or cholesterol or were not injected (control). The ERG b-wave amplitude was evaluated from the trough of the first negative wave to the peak of the first positive wave. Compared with noninjected control and (MERTK)mRNA-LNP(Chol)-treated rats, (MERTK)mRNA-LNP(Sito)-treated rats presented enhanced ERG responses at all tested stimulus intensities (10.0, 3.0, and 0.01 cd s/m²) (Fig. 7C). Notably, at the stimulus intensity of 3.0 cd s/m² and 10.0 cd s/m², the (MERTK)mRNA-LNPs(Sito) group exhibited significantly stronger b-wave responses than the no-injection group did (p < 0.05). In contrast, the (MERTK)mRNA-LNPs(Chol) group did not significantly differ from the noninjected group (p > 0.05), indicating that (MERTK)mRNA-LNPs(Sito) treatment has greater potential for preserving visual function than does (MERTK)mRNA-LNPs(Chol) treatment (Fig. 7C and D).

Immunofluorescence staining of retinal cross-sections further validated these findings (Fig. 7E). In the (MERTK)mRNA-LNPs(Sito)-treated RCS rat eyes, MERTK protein expression colocalized with RPE65, confirming successful expression of the MERTK protein in the RPE layer. Additionally, MERTK expression was detected in the outer nuclear layer (ONL) in the (MERTK)mRNA-LNPs(Sito) treated RCS rat eyes. In contrast, no MERTK expression was detected in the retinas of noninjected RCS rats nor (MERTK)mRNA-LNPs(Chol)-treated group. The stronger and more widespread MERTK expression was observed in RCS rat eyes treated with (MERTK)mRNA-LNPs(Sito) further demonstrates the superior delivery and therapeutic potential of Sito-formulated LNPs encapsulating (MERTK)mRNA drugs (MERTK)mRNA-LNPs(Sito).

Retinal cross sections collected 4 weeks after subretinal injection were subjected to immunofluorescence staining using an anti-FLAG antibody as well. Robust FLAG-positive signals were observed in the (MERTK)mRNA-LNPs(Sito) group, indicating the existence of the exogenous MERTK protein at the 4-week time point (Supplementary Fig. 14). In contrast, no FLAG signal was detected in the untreated controls or in eyes treated with (MERTK)mRNA-LNPs(Chol). These results demonstrate that Sito-containing mRNA-LNP formulations achieve markedly superior and more persistent delivery of IVT (MERTK)mRNA in vivo compared with the cholesterol-based formulation.

To assess potential retinal immune responses, retinal cross sections were also stained for Iba1 (microglia) and GFAP (astrocytes) (Fig. 8A). Minimal microglial or astrocytic activation was observed in untreated eyes and in eyes treated with (MERTK)mRNA-LNPs(Sito), indicating favorable biocompatibility. In contrast, the (MERTK)mRNA-LNPs(Chol) induced pronounced microglial activation within the layer of photoreceptor outer segments, suggesting a higher immunogenic potential.

Fig. 8.

Immunohistochemical and H&E staining analysis for ocular toxicity following subretinal injection of phytosterol-containing LNPs. (A) Representative confocal images ( × 20) of retinal cross sections immunostained for Iba1 (microglia; green), GFAP (astrocytes; magenta), and DAPI (blue) to assess local inflammatory responses. The scale bar represents 20 μm. Iba1 immunostaining showed microglial activation levels. GFAP immunostaining demonstrated Müller cell gliosis assessment. (B) Representative H&E-stained retinal cross sections from untreated, Chol-containing mRNA-LNP treated, or Sito-containing mRNA-LNP treated eyes. The scale bar represents 50 μm. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

H&E staining was performed to assess retinal cytoarchitecture and we evaluated potential tissue responses to subretinal delivery of (MERTK)mRNA-LNPs(Sito) (Fig. 8B). The outer nuclear layer (ONL) in eyes treated with (MERTK)mRNA-LNPs(Sito) or (MERTK)mRNA-LNPs(Chol) showed markedly increased thickness at 4 weeks compared with the untreated group. No RPE disruption, inflammatory infiltrates, or structural degeneration were observed. These findings indicate that subretinal administration of cholesterol-modified mRNA–LNPs did not produce detectable retinal toxicity.

4. Discussion and conclusion

LNPs have become the leading platform for delivering gene therapeutics, particularly mRNAs, owing to their efficiency and modularity [39]. Much of the current focus has focused on optimizing ionizable lipids, which are widely considered essential for forming lipid–mRNA complexes and facilitating intracellular delivery [[40], [41], [42], [43], [44], [45]]. However, the role of cholesterol and its structural analogs remains comparatively underexplored, especially in the context of lyophilization and tissue-specific delivery outcomes.

Previous work by Sahay et al. identified naturally occurring C24-alkyl-substituted cholesterol analogs that increased the cellular uptake and transfection of freshly prepared mRNA-LNPs into several cell types (HeLa, fibroblasts, macrophages, dendritic cells, adenocarcinoma cells, epithelial cells and HEK293 cells) [8,14]. However, their study did not evaluate lyophilized formulations or explore the cell type specificity that may emerge under more extensive conditions [8]. Furthermore, they did not evaluate the transfection efficiency of their mRNA-LNPs in each organ in vivo, and we proceeded with in vivo experiments in this investigation. In our study, we systematically studied the performance of C24-alkyl-substituted cholesterol analogs (Camp, Stig, Sito, and Fuco) in LNPs on multiple organ-derived cell lines (HeLa, HepG2, Colo205, ARPE-19, A549, and PC12) both after and before lyophilization.

Our investigations confirmed that (β-Sito)-formulated mRNA-LNPs achieve superior transfection in freshly prepared formulations compared with other (Camp, Stig, Fuco or Chol (control))-formulated mRNA-LNPs, which is consistent with prior reports by Sahay et al. [8]. However, following lyophilization, (Stig)-formulated mRNA-LNPs presented the highest transfection efficiency in several cell lines, including HepG2, Colo205, and ARPE-19, suggesting that the performance of mRNA-LNPs with cholesterol analogs is context dependent and can be substantially altered by lyophilization.

Interestingly, freshly prepared mRNA-LNPs composed of cholesterol analogs did not universally improve delivery into all the cell lines evaluated in this study. In particular, compared with the freshly prepared Chol-formulated mRNA-LNP formulation, the freshly prepared Stig-formulated mRNA-LNP formulation underperformed in the context of transfection into HepG2 and ARPE-19 cells, indicating that endosomal escape and cellular uptake efficiencies vary with both the lipid composition of the mRNA-LNPs and the cell type used for transfection. These findings underscore the need for careful matching of cholesterol analogs on mRNA-LNPs for transfection into specific tissues or therapeutic sites.

Importantly, we identified MC3-based Fuco-formulated mRNA-LNPs as a promising platform for lung-targeted delivery. After lyophilization and reconstitution, robust bioluminescence protein expression was observed in lungs transfected with Fuco-formulated MC3-based mRNA-LNPs following intravenous administration in vivo. These investigations indicate the superior translational potential of MC3-based Fuco-formulated mRNA-LNPs over β-Sito-formulated mRNA-LNPs, which were recently reported to induce mucosal and systemic immunity after pulmonary delivery [15]. Our findings indicate that Fuco-formulated mRNA-LNPs could be further developed to obtain lung-targeted gene therapies, including pulmonary vaccination and treatment of respiratory disorders.

We also explored the therapeutic potential of β-Sito-formulated ALC-0315-based LNPs in the context of inherited retinal degeneration in an animal model. Delivery of (MERTK)mRNA-LNPs formulated with β-Sito via subretinal injection led to notable recovery of visual function in RCS rats within three weeks. To our knowledge, this study provides the first demonstration of mRNA-LNP-based therapy for vision rescue in the retinal degeneration animal model of RCS rats and highlights a nonintegrating alternative to the reported AAV-mediated gene delivery method [46].

Despite these promising results, our study has several limitations. Therapeutic assessment was conducted over a relatively short timeframe, and intermediate expression kinetics as well as detailed pharmacodynamic characterization were not comprehensively evaluated. Longer-term and longitudinal studies will be necessary to define the full temporal expression profile and evaluate safety, and the potential for repeated dosing. Furthermore, mechanistic studies are warranted to delineate how cholesterol analogs influence mRNA-LNP structural stability, intracellular trafficking, and membrane interactions at the molecular level.

Overall, our findings reveal that cholesterol analogs in mRNA-LNPs can modulate the cell type-specific properties of mRNA-LNPs and underscore their underappreciated role in determining nanoparticle performance under physiologically and pharmaceutically relevant conditions. In particular, the identification of fucosterol (Fuco), a high-performing component of lyophilized mRNA-LNPs for lung-targeted delivery, opens new avenues for the development of next-generation mRNA therapeutics with improved stability and tissue specificity. Furthermore, this study is the first demonstration of the application of C24-alkyl phytosterol-formulated ALC-0315 LNPs for mRNA-mediated restoration of MERTK function in RCS rats.

CRediT authorship contribution statement

Ting Wang: Conceptualization, Formal analysis, Funding acquisition, Project administration, Supervision, Writing – original draft. Wanqi Li: Data curation, Methodology. Jianyang Chen: Formal analysis, Software. Sitian Cheng: Formal analysis, Validation. Min Gao: Visualization. Chengyu Jiang: Formal analysis. Chen Jiang: Formal analysis. Bingyi Hu: Formal analysis. Tzu-Cheng Sung: Formal analysis. Akon Higuchi: Funding acquisition, Project administration, Validation, Visualization, Writing – review & editing.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Ting Wang reports financial support was provided by Zhejiang Provincial Natural Science Foundation of China. Ting Wang reports financial support was provided by Wenzhou Municipal Science and Technology Bureau. Ting Wang reports financial support was provided by Start-up Foundation for Scientific Research, Eye Hospital, Wenzhou Medical University. Akon Higuchi reports financial support was provided by National Natural Science Foundation of China. Akon Higuchi reports financial support was provided by National Key Research and Development Program of China. If there are other authors, they 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 ASupplementary data

The following is the supplementary data to this article.

Data availability

Data will be made available on request.