Nanomedicines for DNA and interference RNA co-delivery: Combined gene therapy for Fabry disease

Abstract

Fabry disease (FD) is a multisystemic rare disorder caused by mutations in the GLA gene encoding α-Galactosidase A (α-Gal A) enzyme. The deficiency of this enzyme leads to progressive lysosomal accumulation of glycosphingolipids, especially globotriaosylceramide (Gb3) and lyso-Gb3. Current therapies remain limited in efficacy and do not fully prevent disease progression. In this work, we designed solid lipid nanoparticles incorporating 5 nm gold nanoparticles (“golden SLNs”) for the co-delivery of GLA plasmid DNA and siRNA against Gb3 synthase to simultaneously promote α-Gal A production and inhibit Gb3 synthesis, as a dual therapeutic strategy for FD. Nanomedicines were prepared by an optimized electrostatic assembly process, combining both the DNA and siRNA with golden SLNs, the peptide protamine and polysaccharides. They were characterized in terms of size, surface charge and nucleic-acid binding, and evaluated in an endothelial cell model of FD (IMFE-1 cells) to assess intracellular trafficking, α-Gal A expression, Gb3 synthase downregulation, Gb3 reduction, cell viability and hemocompatibility. The golden SLN-based nanomedicines, with a particle size from 200 to 300 nm and positive surface charge, increased α-Gal A activity and reduced Gb3 synthase expression and Gb3 levels for at least 15 days. Protamine and gold nanoparticles played key roles in nucleic-acid condensation and intracellular release, while lipid composition and polysaccharide coating determined efficacy and safety. This study presents the first co-delivery nanomedicine enabling simultaneous gene supplementation and genetic substrate reduction therapy, broadening the therapeutic scope for FD.

Graphical abstract

1. Introduction

Nucleic acid-based therapies offer different strategies with revolutionary potential for the treatment of pathological conditions that lack effective treatments. Their mechanism of action and high specificity make them a potential therapeutic option for viral infections, genetic disorders with unmet clinical needs, various cancers (Deverman et al., 2018) and rare diseases (Schambach et al., 2024); presenting promising advantages compared with conventional drugs (Dunbar et al., 2018). Nonetheless, nucleic acid-based therapies require the development of innovative delivery systems that preserve nucleic acids from degradation; afford specificity to target the desired cell and an appropriate intracellular distribution; preventing off-targets and the activation of the immune response. Considering the rise of gene therapy, there is a growing need for developing safe, efficient and specific gene therapy delivery systems. Progress in the science of materials and nanotechnology has resulted in a major boost to the development of non-viral vectors, with a better safety profile and advantages for industrial manufacturing; although they lack the efficacy of viral vectors (del Pozo-Rodríguez et al., 2020). In this sense, lipid-based systems are the most employed delivery systems in the market or in clinical research, followed by antibody-drug conjugates, polymer-drug/protein conjugates and inorganic nanoparticles (NP), including gold nanoparticles (GNs) (Shan et al., 2022). Recent milestones achieved with lipid nanoparticles (LNPs) such as the commercialization of COVID-19 vaccines and the first small interfering RNA (siRNA) marketed product ONPATTRO™, have accelerated the development of lipid-based nanodelivery systems for nucleic acid delivery and will foster the advance of these strategies towards more advanced stages of clinical development (Samaridou et al., 2020). Apart from LNPs, solid lipid nanoparticles (SLNs) have demonstrated their potential as nucleic acid delivery systems and are currently considered one of the most promising non-viral platforms. SLNs are spherical nanometric particles formed by a core composed of a solid lipid surrounded by a layer of surfactants. These versatile nanoplatforms allow the functionalization of their surface with ligands to target the loaded nucleic acid to specific cells; as well as to modulate the intracellular trafficking and the nucleic acid release, increasing their efficacy and safety (Gómez-Aguado et al., 2020; Mendes et al., 2022; Rodríguez-Castejón et al., 2022).

Fabry disease (FD) is a rare metabolic disease that, due to its condition of monogenic disease, can be considered a good candidate to be approached by nucleic acid based-therapies (Shaimardanova et al., 2023). FD is a lysosomal storage disorder (LSD) caused by mutations in the GLA gene, which is responsible for encoding the lysosomal enzyme α-Galactosidase A (α-Gal A). As a result, there is a progressive accumulation of the glycosphingolipid globotriaosylceramide (Gb3) and its deacylated derivate globotriaosylsphingosine (lyso-Gb3) in the lysosome of cells, especially in vascular endothelial and smooth muscle cells, leading to cardiac, renal and cerebrovascular manifestations (Tuttolomondo et al., 2021). Currently, two treatment strategies are available for the treatment of FD: intravenous (i.v.) enzyme replacement therapy (ERT) with two recombinant enzymes, agalsidase α (Replagal®) or agalsidase β (Fabrazyme®), and the oral chaperone Migalastat (Galafold®). In 2023, pegunigalsidase alfa (Elfabrio®) was approved by the EMA and the FDA as a new ERT alternative for FD (European Medicines Agency. Science Medicines Health, n.d.; Elfabrio, n.d.). Although existing therapies have demonstrated to improve the quality of life of patients, none have been able to completely revert clinical manifestations and there are still many clinical needs to be met (Felis et al., 2020). For instance, ERT shows limited efficacy due to low tissue penetration, infusion-associated reactions and not crossing the blood-brain barrier (Azevedo et al., 2021), and migalastat is only effective in amenable mutations (Müntze et al., 2023). Consequently, new strategies should be developed. In this sense, different options have been proposed as an alternative to those treatments, including gene supplementation therapy and substrate reduction therapy (SRT)(Kant and Atta, 2020).

At present, several clinical studies are evaluating the safety of gene therapy for FD, which differ in their therapeutic approaches (Lenders and Brand, 2021). Delivery of GLA sequence by plasmid DNA (pDNA) or messenger RNA (mRNA) to express α-Gal A in native cells has been proposed to overcome the limitations of current therapies (Domm et al., 2021). Endogenously produced α-Gal A by gene supplementation is advantageous because it involves natural translational and post-translational modifications; enhancing stability and reducing immunogenicity as compared to recombinant α-Gal A (Rodríguez-Castejón et al., 2022). In addition to gene supplementation approach, SRT has been proposed to reduce Gb3 accumulation in FD by inhibiting the Gb3 synthase (Gb3S) enzyme, thereby specifically suppressing Gb3 synthesis. Recently, genetic SRT (gSRT) mediated by siRNA has been explored in vitro as a FD-specific therapy (Beraza-Millor et al., 2024).

In the present work, we propose a combined nucleic acid therapy based on the co-delivery, within the same formulation, of a GLA pDNA to restore α-galactosidase A expression and an anti-Gb3S siRNA to reduce substrate synthesis, with the aim of achieving a dual therapeutic effect in the same target cell (Fig. 1). The co-delivery of pDNA and siRNA using a single non-viral vector remains particularly challenging, as it requires the coordinated control of cellular uptake, endosomal escape and intracellular trafficking, together with the condensation and bioavailability of two cargos with distinct physicochemical properties.

Fig. 1.

Schematic representation of combined gene therapy for FD. α-Gal A: α-Galactosidase A. pDNA: Plasmid DNA. Gb3S: Globotriaosylceramide synthase. siRNA: Small interference RNA. FD: Fabry disease. gSRT: Genetic substrate reduction therapy. Gb3: Globotriaosylceramide.

The platform designed in this work is based on SLNs, designed to promote efficient internalization and endosomal escape, combined with GN. The incorporation of GNs provides an additional adsorption interface for nucleic acids, through gold-nucleobase interactions, contributing to the stabilization of pDNA/siRNA co-loading and to a more favorable balance between condensation and intracellular release (Arral and Whitehead, 2026; Dutour and Bruylants, 2025) By enabling the delivery of GLA pDNA and Gb3S siRNA to the same target cell, the resulting vector is intended to promote a coordinated and sustained reduction of intracellular Gb3 levels, improving what can be achieved with single-nucleic acid approaches.

Therefore, this study aims to design and evaluate a hybrid nanomedicine based on SLNs and GNs capable of co-delivering GLA plasmid and siRNA against Gb3S. To this end, an hemocompatible golden SLN platform has been designed and characterized in terms of its physicochemical properties and nucleic acid co-loading, to further assess the intracellular delivery, functional transfection, silencing efficacy and Gb3 reduction durability in a FD endothelial model, IMFE-1 cells.

2. Materials and methods

2.1. Materials

For SLNs preparation, 1,2-dioleoyl-3-trimethylammonium-propane chloride salt (DOTAP) and 1,2-dioleoyl-3-dimethylammonium-propane (DODAP) were obtained from Avanti Polar-lipids, Inc. (Alabaster, AL, USA), while D-Lin-MC3-DMA (MC3) lipid was purchased in BroadPharm (San Diego, CA, USA). Precirol® ATO 5 (glyceryl palmitostearate) was kindly provided by Gattefossé (Madrid, Spain). Tween 80 and dichloromethane were obtained from Panreac (Madrid, Spain). Sigma-Aldrich (Madrid, Spain) provided protamine sulfate salt from salmon (Grade X) (P), dextran (Mn of 3260 Da) (Dx), D-Galacto-D-mannan from Ceratonia siliqua (Mr ∼ 200,000) (Gal), 5 nm and Nile Red. Lifecore Biomedical (Chaska, MN, USA) supplied hyaluronic acid (HA) (Mw of 100 KDa). GNs with a mean diameter of 5 nm were obtained from Sigma-Aldrich. GNs were supplied as a stabilized aqueous suspension with an optical density of 1 in citrate buffer, where citrate acts as the surface stabilizing agent. According to the manufacturer, the approximate particle concentration was ∼5.5 × 10¹³ particles/mL. GNs were stored at 2–8 °C and used as received, unless otherwise specified.

siRNAs against human Gb3S mRNA were obtained from Dharmacon™ (Cambridge, United Kingdom), as well as Dharma®FECT transfection reagent. The functional siRNA against human Gb3S mRNA with the following sequence (from 5′ to 3′) was employed: GCACACGGACUUCAUUGUU (J-016315-05, Mw 13,430.1 g/mol). A scrambled sequence (UGGUUUACAUGUCGACUAA) was used as negative control.

Plasmid pcDNA3-EGFP (6.1 kb) encoding the green fluorescent protein (GFP) was kindly provided by the laboratory of Professor B.H.F. Weber (University of Regensburg, Regensburg, Germany). Plasmid pCMV6-AC-αGLA (7.1 kb) encoding the α-Gal A was purchased from Origene (Rockville, MD, USA).

Materials used on agarose gel electrophoresis assay were provided from Bio-Rad (Madrid, Spain). Agarose, deoxyribonuclease I (DNAse I) and sodium dodecyl sulfate (SDS) were purchased from Sigma-Aldrich, GelRed™ from Biotium (Fremont, CA, USA) and GeneRuler Ultra Low Range DNA Ladder and 1 Kb pDNA ladder from Gibco (Thermo Fisher Scientific, MA, USA).

Immortalized Fabry Endothelial Cells line-1 (IMFE-1) were kindly provided by Dr. Kanenski (National Institute of Neurological Disorders and Stroke National Institutes of Health, Bethesda, MD, USA) and Dr. Shen (Institute of Metabolic Disease, Baylor Research Institute, Dallas, TX, USA). EGM®-2 medium supplemented with 2% fetal bovine serum (FBS), hydrocortisone, human epidermal growth factor (hEGF), vascular endothelial growth factor (VEGF), human fibroblastic growth factor (hFGF-B), insulin-like growth factor 1 (R2-IGF-1), gentamicin sulfate-amphotericin (GA-1000), ascorbic acid and heparin were obtained from Lonza (Basel, Switzerland). Cell culture reagents, including trypsin/EDTA and phosphate buffered saline (PBS) were obtained from Gibco (Thermo Fisher Scientific, Madrid, Spain), while glucose was obtained from Sigma-Aldrich.

To study the endocytosis activity in IMFE-1 cells, AlexaFluor488-Transferrin, AlexaFluor488-Cholera Toxin and Fluorescein-70,000 MV Dx were obtained from ThermoFisher Scientific (Madrid, Spain).

4-methylumbelliferyl-α-D-galactopyranoside (4-MU-α-Gal), N-acetyl-D-galactosamine and 4-methylumbelliferone (4-MU), employed for the quantification of α-Gal A activity, were obtained from Sigma-Aldrich. Micro BCA™ Protein Assay Kit was acquired from Thermo Fisher Scientific (Madrid, Spain).

To study the exposure of pDNA and siRNA on the surface of the vectors, the Quant-it™ Picogreen and Quant-it™ Ribogreen kits were obtained from ThermoFisher (Madrid, Spain).

A quantitative reverse transcription PCR (RT-qPCR) was performed to quantify mRNA expression. The materials employed for it, as well as the LightCycler® 2.0 System were provided by Roche Diagnostics (Mannheim, Germany): High Pure RNA Isolation kit, First Strand cDNA Synthesis Kit, LightCycler® FastStart DNA Master SYBR Green I, and the specific primers.

To detect the expression of Gb3S, anti-human Gb3S rabbit antibodies were provided by Sigma-Aldrich (Madrid, Spain), as well as goat serum and Triton™ X-100. Goat-anti-rabbit IgG cross-adsorbed secondary antibody (Alexa Fluor™ 488) was purchased from Gibco (Thermo Fisher Scientific, Madrid, Spain). 4′,6-diamidine-2′-phenylindole dihydrochloride (DAPI)-fluoromount-G® from Southern Biotech (Birmingham, AL, USA).

To determine the accumulation of Gb3, mouse anti-human Gb3 monoclonal antibody was purchased from BioLegend (San Diego, CA, USA) and pure Gb3 from Matreya (State College, PA, USA).

For viability assay, thiazolyl blue tetrazolium bromide (MTT) was obtained from Sigma-Aldrich (Madrid, Spain).

2.2. Preparation of SLNs and non-viral formulations

SLN_EE were elaborated as previously described (Gómez-Aguado et al., 2021). Precirol® ATO 5 was dissolved in dichloromethane (5% w/v) and mixed with an aqueous solution containing Tween 80 (0.1% w/v) and the cationic/ionizable lipid (0.4% w/v). The emulsion was obtained by sonication (Branson Sonifier 250, Danbury, Connecticut, USA) for 30 s at 50 W. The organic solvent was evaporated by stirring for 1 h and 45 min of vacuum. SLN_HM were prepared as previously described (Rodríguez-Castejón et al., 2022). Briefly, the oily phase (composed only by Precirol® ATO 5) and the aqueous phase were heated at 80 °C separately. Then, the aqueous phase was added to the oily phase and sonicated for 30 min.

siRNA vectors were prepared by the combination of the SLNs with different components: siRNA, GNs, P, and one polysaccharide, either Dx, Gal or HA, as previously described (Beraza-Millor et al., 2023). According to the optimization of the GNs amount in the previous study, a siRNA solution was mixed for 15 min with 5 nm GNs; two proportions of GNs were used, corresponding to 2.75 × 10¹¹ (GN5) and 1.38 × 10¹¹ (GN2.5) gold molecules, respectively. Then, P was added in a P:siRNA ratio of 2:1 (w:w) or 3:1 (w:w) and mixed for 5 min. Next, a solution of either Gal, Dx or HA was added and mixed for 15 min in a polysaccharide:siRNA ratio of 0.1:1 (w:w), 1:1 (w:w)or 0.5:1 (w:w), respectively. Finally, the SLNs were added to the complexes to get the final vector in a SLN:siRNA ratio of 5:1 (w:w). In the case of pDNA-vectors the same process was followed, except that siRNA was replaced by pcDNA3-EGFP or pCMV6-AC-α-GLA. For some studies, pcDNA3-EGFP-vectors were prepared without GN.

To prepare the co-delivery vectors, which contain pDNA, siRNA, GNs, P, and the polysaccharide, the order of addition of these components was optimized (Fig. S1), as well as the amount of GNs and P (Table S1). In all formulations, siRNA:pDNA:SLN ratio (w:w:w) was 0.05:1:5. The Dx:pDNA ratio was 1:1 (w:w) and the HA:pDNA ratio was 0.5:1 (w:w).

2.3. Characterization of SLNs and vectors: Size, polidispersity index and ζ-potential measurements

The size and polidispersity index were determined by Dynamic Light Scattering (DLS). ζ-potential was measured by Laser Doppler Velocimetry (LDV). Measurements were carried out in a ZetaSizer Nano ZS (Malvern Instruments, Worcestershire, UK) at 25 °C. DLS measurements were performed under highly diluted conditions in MilliQ™ water, to ensure single scattering and minimize interparticle interactions (Bhattacharjee, 2016; Stetefeld et al., 2016).

2.4. Stability study

Vectors containing SLN produced by the EE method (SLN_ET and SLN_EMC3) were stored as aqueous suspensions at 4 °C for 30 days, in polypropylene microcentrifuge tubes, sealed with Parafilm™ to minimize evaporative losses, and kept in a closed refrigerator, thereby reducing exposure to ambient light. Then, particle size, PDI and ζ-potential were measured.

2.5. Agarose gel electrophoresis assay

The capacity of the vectors to bind the pDNA and siRNA, to protect pDNA against DNAse I digestion, and to release pDNA and siRNA was studied by 0.7% (w/v) agarose gel electrophoresis labeled with GelRed™. The gel was run for 30 min at 120 V and analyzed with an Uvitec Uvidoc D-55-LCD.20 M Auto transilluminator (Cambridge, UK). The binding capacity of the systems was evaluated by adding the co-delivery vector diluted in Milli-Q™ directly in the gel in a final concentration of 0.03 μg pDNA/μL and 0.002 μg siRNA/μL. To study the protection against DNase I digestion, the vector at the same concentration was incubated for 30 min at 37 °C with 1.5 U DNase I/2.5 μg DNA. Then, a solution of SDS (4% w/v) was added to a final concentration of 1% (w/v) and incubated at room temperature. In the release study, the SDS solution was added to the vector to unbind the nucleic acids. As controls, 1 Kb pDNA ladder and untreated pR-M10-αGal A and RiboRuler Ultra low Range DNA Ladder were used.

2.6. pDNA and siRNA exposure on the nanoparticle surface

The exposure of pDNA and siRNA on the surface of co-delivery vectors was evaluated by using Quant-it™ PicoGreen and Quant-it™ RiboGreen kits, based on dyes that become intensely fluorescent when bound to pDNA and RNA, respectively. Briefly, a calibration curve was prepared for each nucleic acid as described in the respective user guides. Co-delivery vectors were diluted in 1 x TE buffer and incubated for 5 min with the corresponding working reagent before reading at λ = 480–530 nm using a Glomax®-Multi Detection System (Promega, Madison, WI, USA) microplate reader.

2.7. In vitro studies

2.7.1. Cell culture

For in vitro studies, a cellular model of FD (IMFE-1 cells) was employed (Shen et al., 2007). Cells were incubated at 37 °C, in a 5% CO₂ air atmosphere, changing the medium every 2 or 3 days and subculturing every 7 days.

2.7.2. Transfection efficacy

2.7.2.1. EGFP transfection

IMFE-1 cells were cultured in 24-well plates at a density of 30,000 cells per well 72 h before the addition of the vectors. Then, 0.5 mL of the medium was removed, leaving enough volume to cover the cells, and a total volume of 75 μL of each vector diluted in glucose (5% v/v) (equivalent to 2.5 μg of pDNA and 25 nM siRNA) was added to each well. After 4 h of incubation, the medium was refreshed with 2 mL of new complete medium. Cells were incubated for 48 h.

In order to know the efficacy of transfection in terms of percentage of cells transfected and production of the transgene protein, the vectors were prepared with the pcDNA3-EGFP plasmid. To analyze transfection efficacy, cells were washed twice in PBS and detached by incubation trypsin-EDTA. After centrifugation of cell suspensions at 1000 rpm for 5 min, pellets were resuspended in PBS and analyzed on a CytoFLEX flow cytometer (Beckman Coulter) at 525 nm (FITC channel). Transfection efficiency was calculated as the percentage of EGFP-positive cells among total cells, and EGFP expression was quantified by mean fluorescence intensity. Each sample included 10,000 recorded events.

2.7.2.2. α-Gal A transfection

Before transfection, cells were cultured in 6-well plates at a density of 200,000 cells per well. After 72 h of incubation, 1 mL of medium was removed and 75 μL of each vector diluted in glucose 5% (v/v) were added (equivalent to 2.5 μg of pDNA and 25 nM siRNA). Four hours after, the medium was refreshed and cells were incubated for additional 48 h. Then, cells were washed twice with PBS, scrapped and collected for 5 min of centrifugation at 5000 rpm. Then, the cell pellets were resuspended in 100 μL of Milli-Q™ water and sonicated.

A fluorometric assay of α-Gal A activity was performed to detect the conversion of 4-MU-α-Gal into the product 4-MU (Rodríguez-Castejón et al., 2022). An aliquot of each sample was incubated with 4-MU-α-Gal (5 mM) and 100 mM of the specific α-galactosidase B inhibitor α-N-acetylgalactosamine in 0.1 M sodium citrate buffer (pH = 4.4) under agitation at 37 °C. The reaction was stopped 30 min later with a 0.1 M glycine-NaOH buffer (pH = 10.4). The formation of the product 4-MU was determined with the fluorescence signal (λexcitation = 360 nm; λemission = 450 nm) on a Glomax®-Multi Detection System (Promega, Madison, WI, USA). Protein concentrations were determined by Micro BCA™ Protein Assay (Thermo Fisher Scientific, Rockford, IL, USA). One unit of α-Gal A activity is equivalent to the hydrolysis of 1 nmol of substrate in 1 h at 37 °C. α-Gal A activity was expressed as 4-MU nmol/h/mg total protein.

2.7.3. Endocytosis activity in IMFE-1 cells

To study the endocytosis activity of IMFE-1 cells, they were cultured in 24 well-plates at a density of 30,000 cells per well. After 72 h of incubation, 0.5 mL of medium was removed and different endocytosis markers were added to the cell: transferrin (75 μg), cholera toxin (5 μg) and 70,000 kDa Dx (500 μg). After 30 min or 2 h, cells were washed twice with PBS and detached with Trypsin-EDTA. After centrifugation of cell suspensions at 1000 rpm for 5 min, pellets were resuspended in PBS and analyzed on a CytoFLEX flow cytometer (Beckman Coulter) at 525 nm (FITC channel). For each sample, 10,000 events were collected.

2.7.4. Cellular association

For cellular association study, SLNs were labeled with the fluorescent dye Nile Red (λ = 590 nm).

IMFE-1 cells were seeded in 24-well plates at a density of 60,000 cells per well and incubated for 48 h. Before transfection, 0.5 mL of the medium was removed and a total volume of 75 μL of each vector diluted in glucose (5% v/v) (equivalent to 2.5 μg of pDNA and 25 nM siRNA) was added to each well. After 2 h of incubation, the culture medium was removed, cells were washed twice with PBS and detached from plates. After centrifugation of cell suspensions at 1000 rpm for 5 min, pellets were resuspended in PBS and cytometry analyses were carried out, using a CytoFLEX flow cytometer (Beckman Coulter) at 610 nm (ECD). For each sample, 10,000 events were collected.

Calibration curves of Nile red-labeled SLN_ET and SLN_EMC3 were prepared in PBS (acellular conditions), with SLNs amounts corresponding to 10, 20, 30, 40 and 50 μg of Precirol® ATO 5. The fluorescence signal (λexcitation = 520 nm; λemission = 580 nm) was measured using a GloMax®-Multi Detection System (Promega, Madison, WI, USA).

2.7.5. Intracellular disposition of pDNA

Cells were seeded in 4 well-Millicell EZ slides (Millipore) at a density of 35,000 cells in 1 mL of medium per well and they were incubated for 24 h. Co-delivery vectors were prepared with Label IT® Cy5-labeled pDNA. Cells were treated with 75 μL of each vector diluted in glucose (5% v/v) (equivalent to 2.5 μg of pDNA and 25 nM siRNA). Two hours after transfection, the slides were washed twice with PBS, fixed with paraformaldehyde (PFA) 4% and covered with the mounting fluid DAPI-fluoromount-G®. The slides were analyzed using a Leica DM IL LED Fluo inverted microscope.

2.7.6. Silencing efficacy

The ability of the siRNA vectors and the co-delivery vectors to silence Gb3S mRNA was evaluated in IMFE-1 cells. Cells were seeded at a density of 30,000 cells per well, in 12-well plates 72 h before transfection. One mL of medium was removed and 75 μL of each vector diluted in glucose (5% v/v) corresponding to a siRNA dose of 25 nM and 2.5 μg of pDNA was added to each well. Four hours later, the medium was refreshed with 2 mL of new complete medium and cells were incubated for additional 48 h.

Gb3S mRNA expression in IMFE-1 cells was measured by RT-qPCR as previously described (Zumbrun et al., 2010). Total RNA was extracted from cells using a High Pure RNA Isolation Kit with DNAse digestion and reverse-transcribed using the Transcriptor First Strand cDNA Synthesis Kit. The cDNA was amplified using the LightCycler® FastStart DNA Master SYBR Green I with specific primers sets to quantify the human Gb3S transgene and a primer set specific to the β-actin gene as the endogenous reference. The primer set specific for Gb3S gene were 5’-GGCAACATCTTCTTCCTGGAGACTTC-3′ (sense) and 5’-CGAACTTCCACATGAGTGCGATCC-3′ (antisense) and those corresponding to β-actin were 5’-CATTGTGATGGACTCCGGAGACGG-3′ (sense) and 5’-CATCTCCTGCTCGAAGTCTAGAGC-3′ (antisense). The thermal cycling conditions were 95 °C for 600 s, followed by 45 cycles of 95 °C for 5 s, 62 °C for 10 s and 72 °C for 20 s. Finally, melting curve was performed by a single cycle of 95 °C for 60 s, 70 °C for 60 s and 95 °C for 1 s. mRNA quantification, and thus the silencing inhibition potential, was calculated employing the ∆∆Ct method. As negative control of gene silencing, ON-TARGETplus™ non-targeting siRNA was employed (scramble) and ON-TARGETplus™ Cyclophilin B Control siRNA (Human) was used as positive control.

2.7.7. Gb3S enzyme expression

Gb3S enzyme expression was studied by immunocytochemistry in IMFE-1 cells. A density of 35,000 cells were cultured in 4-well Millicell EZ slides (Millipore, Burlington, Massachusetts, USA) and incubate for 24 h. Before transfection, 500 μL of medium were removed. Cells were transfected with 75 μL of each vector (corresponding to 2.5 μg of pDNA and/or 25 nM of siRNA). After 4 h of incubation, wells were refreshed with complete medium. After 7 or 15 days of incubation, cells were washed twice with PBS and fixed with PFA 4%. They were incubated with a primary rabbit antibody specific for Gb3S (Sigma-Aldrich, Madrid, Spain) 12 h at 4 °C. Then, goat-anti-rabbit IgG secondary antibody was added for 1 h. Finally, they were covered with the mounting fluid DAPI-fluoromount-G®. Slides were analyzed using a Leica DM IL LED Fluo inverted microscope.

2.7.8. Gb3 quantification

In order to quantify Gb3 in IMFE-1 cells, the method developed by Wee-Ren et al. (Ng and Narayanan, 2021) was followed. This method is based on the use of a specific FITC-conjugated mouse anti-human Gb3 monoclonal antibody (BioLegend, USA). This antibody interacts with all isoforms of Gb3 in a total lipid extract, and therefore, allows the quantification of the amount of Gb3 present in the sample, regardless of the varying isoforms or species (Krüger et al., 2010; Park et al., 2003).

Total lipid from IMFE-1 cells was extracted with ice-cold chloroform and treated with FITC-conjugated mouse anti-human Gb3 monoclonal antibody. Gb3 was quantified by measuring the fluorescence signal at excitation and emission wavelength of 480 nm and 520 nm, respectively in a GloMax®-Multi Detection System (Promega, Madison, WI, USA). Concentrations were calculated by extrapolating the fluorescence signal in a standard curve prepared with pure Gb3.

More detailed information about the method is available in Supplementary material – Section Validation of Gb3 quantification technique.

2.7.9. Cell viability

MTT analysis was performed for viability analysis of the IMFE-1 cells treated with vectors. Cells were cultured in 96-well plate, at a density of 1000 cells per well in a total volume of 100 μL of medium. Twenty-four hours later, the medium was removed and replaced with the vector suspension and incubated for 4 h. Twenty-four hours later, cells were treated with 50 μL of MTT reagent and, after 2 h of incubation, metabolic activity was measured using absorbance analysis by a GloMax®-Multi Detection System (Promega, Madison, WI, USA).

2.8. Interaction with erythrocytes: Hemolysis and hemagglutination

Hemolytic and hemagglutination effect of the siRNA vectors and co-delivery vectors were assessed following the protocol described by Kurosaki et al. (Kurosaki et al., 2010). Fresh blood was centrifuged at 4000 rpm during 5 min and the plasma and the buffy coat were discarded. Then, erythrocytes were washed three times with PBS by centrifugation and finally they were diluted in PBS to a final concentration of 5% (v/v) for the hemolysis study and 2% (v/v) for agglutination evaluation. For hemolysis assay, vectors were mixed with erythrocytes at a ratio 1:1 (v/v) and incubated for 60 min. After incubation, they were centrifuged at 4000 rpm for 5 min and hemoglobin release in the supernatant was measured at 545 nm using a GloMax®-Multi Detection System (Promega, Madison, WI, USA) microplate reader. Hemolysis values were expressed as percent of lysis buffer positive control (100%).

For agglutination study, vectors were incubated with erythrocytes at a ratio 1:1 (v/v) for 15 min. Then, 15 μL of the mixture were placed on a microscope slide and images were captured in a Leica DM IL LED Fluo inverted microscope. A vector prepared with Poly-l-Lysine was employed as positive control.

2.9. Statistical analysis

Results are reported as mean values ± standard deviation (SD). Statistical analysis was performed using IBM® SPSS® Statics 28 (IBM Corp, Armonk, NY, USA). The normal distribution of samples was assessed by Shapiro-Wilk test and homogeneity of variance by Levene test. The different formulations were compared with ANOVA and Student's t-test.

3. Results

3.1. Characterization of SLNs and vectors: Size, polydispersity index and ζ-potential

Table 1 shows the composition, particle size, polydispersity index (PDI) and ζ-potential of the SLNs. Two different techniques were employed to prepare SLN_S: solvent evaporation/emulsification (SLN_EE) and hot-melt emulsification (SLN_HM). The SLNs were composed of Precirol® ATO 5, Tween 80 and the cationic lipid DOTAP (SLN_ET and SLN_HT) or a mixture of DOTAP and either the ionizable lipid MC3 (SLN_EMC3 and SLN_HMC3) or DODAP (SLN_ED and SLN_HD). Particle size ranged from 90 to 230 nm and PDI values were lower than 0.3, indicating size homogeneity. The surface charge, expressed by ζ-potential, was always positive. SLN_HM presented a lower particle size and a higher ζ-potential than SLN_EE (p < 0.001). Considering the lipid composition, the inclusion of MC3 or DODAP resulted in higher particle size respect to those containing only DOTAP (p < 0.001).

Table 1.

Composition and physicochemical characteristics of the SLNs.

Type of SLN	Preparation protocol	Cationic/ionizable Lipid (%)			Tween 80 (%)	Size (d.nm)	PDI	Ζ-Potential (mV)
		DOTAP	MC3	DODAP
HT	HM	0.4			0.1	91.1 ± 1.8	0.27 ± 0.01	+66.2 ± 0.8
HMC3	HM	0.2	0.2		0.1	137.1 ± 0.8	0.27 ± 0.01	+65.5 ± 0.6
HD	HM	0.2		0.2	0.1	135.0 ± 2.2	0.28 ± 0.01	+62.0 ± 0.9
ET	EE	0.4			0.1	184.9 ± 1.3	0.24 ± 0.01	+56.6 ± 2.0
EMC3	EE	0.2	0.2		0.1	230.1 ± 1.4	0.28 ± 0.01	+58.7 ± 1.1
ED	EE	0.2		0.2	0.1	201.4 ± 3.5	0.28 ± 0.01	+56.9 ± 2.1

HM: Hot melt, EE: emulsification/evaporation, SLN: solid lipid nanoparticle, HT: SLN prepared by HM with DOTAP, HMC3: SLN prepared by HM with DOTAP/MC3, HD: SLN prepared by HM with DOTAP/DODAP, ET: SLN prepared by EE with DOTAP, EMC3: SLN prepared by EE with DOTAP/MC3, ED: SLN prepared by EE with DOTAP/DODAP. Data are expressed as mean ± standard deviation; n = 3. Significant differences (p < 0.001) in size and ζ-potential were observed among all formulations, except in size between HMC3 and HD.

Table S1 shows the size, PDI and ζ-Potential of siRNA vectors, pDNA vectors and co-delivery vectors. Figs. S2, S3 and S4 depict size and ζ-Potential to enable comparison across formulations.

The order of addition of the components to prepare the nanomedicines for co-delivery was found to be critical. Out of the four preparation protocols (A, B, C and D) (Fig. S1), only protocols A and B resulted in stable co-delivery vectors able to maintain homogeneous nanometric features for at least one week stored at 4 °C. The physicochemical features of the nanomedicines prepared by protocols A and B (Table S1) depended on the SLNs composition, with a particle size ranging from 108 nm to 300 nm, positive ζ-potential (+30 to +44 mV) and PDI lower than 0.4. All co-delivery vectors showed a lower ζ-potential (p < 0.001) compared to SLNs used for their preparation. Vectors formed by protocol A (Fig. S2) presented larger size than those formed by protocol B (Fig. 2). The highest difference was observed with DxP3GN5-ET (301 nm vs 244 nm) and HAP3GN5-HMC3 (335 nm vs 185 nm).

Fig. 2.

Size and ζ-potential of co-delivery vectors prepared by protocol B, containing pcDNA3-EGFP and siRNA. Data are expressed as mean ± standard deviation; n = 3. (A): SLN_DOTAP-based co-delivery vectors. (B): SLN_DOTAP/MC3-based co-delivery vectors. * p < 0.001 among formulations with the same composition but different polysaccharide.

The amount of P in the co-delivery vectors was optimized, and P2 (P:pDNA ratio of 2:1) and P3 (P:pDNA ratio of 3:1) were selected for the preparation of the final nanosystems. The type of polysaccharide affected the stability and the particle size and PDI of the vectors. The presence of HA generally resulted in larger vectors than those containing Dx, particularly in SLN_HT-based formulations. Vectors prepared with Gal lost their uniform nanoscale characteristics in less than a week.

3.2. Transfection efficacy of co-delivery vectors containing pcDNA3-EGFP plasmid

Transfection efficacy in IMFE-1 cells was analyzed by flow cytometry 2 days after the addition of the co-delivery vectors bearing the pcDNA3-EGFP and siRNA. The transfection efficacy of the nanomedicines prepared with SLN_EE depended on the composition of cationic and ionizable lipids, with transfections rates of around 20% for SLN_ET, 40% for SLN_EMC3 (Fig. 3) and 5% for SLN_ED (Fig. S5). The mean of fluorescence intensity, indicative of protein production, was higher for SLN_EMC3-based co-delivery vectors (Fig. 3). When vectors were formulated with SLN_HM, transfection was below 2% and the mean fluorescence intensity was nearly 0, regardless of the presence of the ionizable lipids. Therefore, they were discarded for further studies.

Fig. 3.

Transfection efficacy in IMFE-1 cells after transfection with co-delivery vectors. (A) Percentage of GFP-positive cells. (B) Mean fluorescence intensity. Data are expressed as mean ± standard deviation; n = 3. A.U.: arbitrary unit. GFP: green fluorescent protein. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Co-delivery vectors prepared by protocol B with SLN_EE resulted in higher transfection efficacy than pDNA vectors with the same composition but without siRNA (16% to 41%) vs (10% to 32%). In addition, the mean fluorescence signal was also higher with the co-delivery formulations (Fig. 3 vs Fig. S6).

The transfection efficacy of co-delivery vectors prepared with SLN_EE by protocol A (<20% transfected cells; Fig. S7), was lower than that obtained with protocol B. Considering these results, as well as the characterization in terms of size, ζ-potential and PDI, co-delivery vectors formed by protocol A were discarded for further studies.

After the selection of SLN_EE prepared by protocol B for the preparation of the co-delivery vectors, the influence of the GNs ratio on their physicochemical properties was evaluated. Two GNs proportions, GN5 and GN2.5, were used. As shown in Fig. 4 and Table S1, both proportions enabled the formation of vectors with suitable physicochemical characteristics. However, in some formulations, the lower GNs proportion resulted in smaller particles sizes (p < 0.001).

Fig. 4.

Influence of the GNs proportion (GN5 or GN2.5) in the physicochemical characteristics of co-delivery vectors. Data are expressed as mean ± standard deviation; n = 3. (A): ET-based co-delivery vectors. (B): EMC3-based co-delivery vectors. * p < 0.001.

3.3. Stability study

Fig. 5 features the physicochemical characteristics of co-delivery vectors prepared by Protocol B and containing SLN_EE, stored at 4 °C up to 30 days. In general, all vectors maintained good stability in terms of size, ζ-potential and PDI, except for HAP2GN2.5-ET and HAP3GN2.5-EMC3, which increased in size by 90 nm and 170 nm at day 30, respectively.

Fig. 5.

Stability of co-delivery vectors during one month. (ET) SLN_ET-vectors. (EMC3) SLN_EMC3-vectors. Data are expressed as mean ± standard deviation; n = 3.

3.4. Agarose gel electrophoresis assay

Fig. 6 illustrates the ability of the SLN_EE-based co-delivery vectors to bind, protect and release both siRNA and the pDNA, evaluated by agarose gel electrophoresis. Fig. 6A and C correspond to Dx formulations prepared with P3 and P2, respectively, while Fig. 6B and D correspond to HA formulations prepared with P3 and P2, respectively.

Fig. 6.

Binding, protection and release capacity of co-delivery vectors with SLNs containing DOTAP or DOTAP/MC3. In all gels, lane 1 corresponds to 1 Kb pDNA ladder. Lane 2 to naked pCMV6-AC-αGLA, lane 3 to naked pCMV6-AC-αGLA + SDS, lane 4 to naked siRNA, lane 5 to siRNA + SDS, lane 6 to siRNA-GN + SDS, lane 11 to naked pR-M10-αGal A and lane 12 to Naked siRNA + DNase + SDS. (A) DxP3GNSLNEE co-delivery vectors: (7) DxP3GN5-ET. (8) DxP3GN2.5-ET. (9) DxP3GN5-EMC3. (10) DxP3GN2.5-EMC3. (13) DxP3GN5-ET vector + DNase + SDS. (14) DxP3GN2.5-ET vector + DNase + SDS. (15) DxP3GN5-EMC3 + DNase + SDS. (16) DxP3GN2.5-EMC3 + DNase + SDS. (17) DxP3GN5-ET + SDS. (18) DxP3GN2.5-ET + SDS. (19) DxP3GN5-EMC3 + SDS. (20) DxP3GN2.5-EMC3 + SDS. (B) HAP3GNSLNEE co-delivery vectors: (7) HAP3GN5-ET. (8). HAP3GN2.5-ET. (9) HAP3GN5-EMC3. (10) HAP3GN2.5-EMC3. (13) HAP3GN5-ET + DNase + SDS. (14) HAP3GN2.5-ET + DNase + SDS. (15) HAP3GN5-EMC3 + DNase + SDS. (16) HAP3GN2.5-EMC3 + DNase + SDS. (17) HAP3GN5-ET + SDS. (18) HAP3GN2.5-ET + SDS. (19) HAP3GN5-EMC3 + SDS. (20) HAP3GN2.5-EMC3 + SDS. (C) DxP2GNSLNEE co-delivery vectors: (7) DxP2GN5-ET. (8). DxP2GN2.5-ET. (9) DxP2GN5-EMC3. (10) DxP2GN2.5-EMC3. (13) DxP3GN5-ET + DNase + SDS. (14) DxP3GN2.5-ET + DNase + SDS. (15) DxP2GN5-EMC3 + DNase + SDS. (16) DxP2GN2.5-EMC3 + DNase + SDS. (17) DxP2GN5-ET + SDS. (18) DxP2GN2.5-ET + SDS. (19) DxP2GN5-EMC3 + SDS. (20) DxP2GN2.5-EMC3 + SDS. (D) HAP2GNSLEE co-delivery vectors: (7) HAP2GN5-ET. (8). HAP2GN2.5-ET. (9) HAP2GN5-EMC3. (10) HAP2GN2.5-EMC3. (13) HAP2GN5-ET + DNase + SDS. (14) HAP2GN2.5-ET + DNase + SDS. (15) HAP2GN5-EMC3 + DNase + SDS. (16) HAP2GN2.5-EMC3 + DNase + SDS. (17) HAP2GN5-ET + SDS. (18) HAP2GN2.5-ET + SDS. (19) HAP2GN5-EMC3 + SDS. (20) HAP2GN2.5-EMC3 + SDS.

The vectors fully complexed both nucleic acids, as evidenced by the material retained in the sample-loading wells (lanes 7–10) and the absence of migrating bands, indicating that neither siRNA nor pDNA was able to diffuse through the gel. Upon SDS treatment (lanes 17–20), both nucleic acids were efficiently released. The siRNA bands (at the bottom of the gel) were more intense in formulations prepared with the lower polymer proportion (P2), suggesting a greater release capacity under these conditions.

The vectors also protected the pDNA from DNase degradation (lanes 13–16), whereas free pDNA was completely digested in the presence of DNase (lane 11). Nevertheless, differences in pDNA conformation after DNAse exposure were evident depending on the formulation: Dx-based vectors exhibited stronger protection against enzymatic degradation, while HA-based vectors showed a more intense open-circular pDNA band (upper bands in Fig. 6B and D, lanes 12–16).

3.5. pDNA and siRNA exposure on the nanoparticle surface

The use of the fluorescent dyes, PicoGreen™ and RiboGreen™, which selectively bind free DNA and RNA, respectively (Ban and Kim, 2024; Jones et al., 1998), confirmed that both pDNA and siRNA were strongly condensed within the vectors (Fig. S8). Although each dye is designed to preferentially recognize either pDNA or siRNA, the structural similarities between the two nucleic acids allow partial cross-binding. In the siRNA-Dx-P2GN5-EM3 control formulation (containing only siRNA), PicoGreen™ produced a very low fluorescence signal (<5%), whereas in the pDNA-Dx-P2-EMC3 control formulation (containing only pDNA), the fluorescence detected with RiboGreen™ reached approximately 20%.

3.6. Endocytosis activity

Fig. 7 features the results of the endocytosis activity assay performed in IMFE-1 cells. Exposure of the cells to transferrin, cholera toxin, and Dx 70,000 kDa—each a specific marker of clathrin-mediated endocytosis, lipid raft/caveolae-mediated endocytosis, and macropinocytosis, respectively—confirmed that all three pathways are active in this cell line. Thirty minutes after incubation with the markers, only about 10% of the cells showed a positive signal; however, after 2 h, this percentage increased markedly, exceeding 80%.

Fig. 7.

Endocytosis activity in IMFE-1 cells. Data are expressed as mean ± standard deviation; n = 3.

3.7. Cellular association

Cellular association was assessed using vectors labeled with Nile Red, a procedure that did not modify their particle size or zeta potential (p > 0.05) (Table S2). First, we confirmed that the fluorescence intensity of Nile Red-labeled vectors was comparable between SLN_ET and SLN_EMC3 (Fig. S9).

Fig. 8 shows the degree of cellular association 2 h after the transfection with the co-delivery vectors. All formulations achieved close to 100% positive cells, with the exception of HAP3GN5-ET, which exhibited an average of 80% (Fig. 8A). In terms of fluorescence intensity, nanovectors formulated with SLN_ET displayed lower fluorescence intensity levels compared with those prepared using SLN_EMC3 (Fig. 8B). Consistent with its lower cell association rate, HAP3GN5-ET also showed the weakest fluorescence signal among all evaluated formulations.

Fig. 8.

Cellular association of co-delivery vectors in IMFE-1 cells 2 h after transfection. Data are expressed as mean ± standard deviation; n = 3. (A) Percentage of positive cells after transfection with co-delivery vectors. (B) Mean fluorescence intensity after transfection with co-delivery vectors. Data are expressed as mean ± standard deviation; n = 3. A.U.: arbitrary unit. * p < 0.001.

3.8. Intracellular disposition of pDNA

Fig. 9 shows the intracellular disposition of pDNA labeled with Label IT® Cy5 in IMFE-1 cells after the treatment with selected co-delivery vectors. As expected, no intracellular fluorescence was detected when cells were exposed to naked pDNA or naked siRNA (data not shown). Vectors formulated with the lower amount of GNs (GN2.5) produced a stronger fluorescence signal than their counterparts containing a higher proportion of GNs (GN5), likely reflecting the greater DNA condensation capacity associated with higher GNs content. The fluorescence signal of Dx-based nanovectors was notably weaker than that of HA-containing vectors, consistent with the ability of HA to counteract the strong DNA condensation induced by P. The most pronounced intracellular fluorescence was observed in cells treated with HAP2GN2.5-ET, HAP2GN2.5-EMC3, HAP3GN2.5-ET and HAP3GN2.5-EMC3. Regarding SLN composition, vectors based on SLN_EMC3 induced higher fluorescence levels, which may be attributed to their lower surface charge which results in a reduced DNA condensation capacity.

Fig. 9.

Intracellular disposition of pDNA in IMFE-1 cells after transfection with different co-delivery vectors. Blue: nuclei labeled with DAPI. Red: Fluorescence signal of nucleic acid labeled with Label IT® Cy5. Scale bar = 15 μm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.9. α-Galactosidase A activity

To evaluate the capacity of the co-delivery vectors to induce α-Gal A synthesis in vitro, intracellular α-Gal A activity was quantified in IMFE-1 cells 48 h after transfection with vectors carrying pCMV6-AC-αGLA (Fig. 10). Basal α-Gal A activity in untreated cells was 13 nmol/h/mg protein, and results were benchmarked against pCMV6-AC-αGLA formulated with Lipofectamine® 2000 (positive control). All formulations except DxP3GN5-ET, HAP2GN5-ET and HAP3GN5-ET significantly increased intracellular α-Gal A activity. HAP2GN2.5-EMC3 achieved the highest activity (118 nmol/h/mg), followed by DxP2GN2.5-EMC3 and HAP3GN2.5-EMC3 (82 and 77 nmol/h/mg, respectively). Reducing the GNs content resulted in a significantly higher α-Gal A activity for the DxP3GN2.5-ET, HAP2GN2.5-ET, DxP2GN2.5-EMC3, and HAP2GN2.5-EMC3 formulations compared with their respective GN5 counterparts.

Fig. 10.

α-Gal A activity in IMFE-1 cells after their treatment with co-delivery vectors. Data are expressed as mean ± standard deviation; n = 3. * p < 0.001. # p < 0.001 with respect to the same vector prepared with the different SLN.

To determine whether the presence of siRNA in the co-delivery systems influences the transfection efficiency to induce α-Gal A expression, enzymatic activity was compared with that obtained after transfection with vectors prepared with pDNA alone. The α-Gal A activity achieved with the SLN_EMC3 vectors prepared without siRNA was comparable to that achieved with the corresponding co-delivery vector (Fig. 10 and Fig. S10). For GN5-based vectors, α-Gal A activity was higher when formulated with pDNA alone. In contrast, within the co-delivery vectors, α-Gal A activity was greater for GN2.5 formulations. This behavior was consistent regardless of the polysaccharide employed, Dx or HA.

3.10. Silencing efficacy

Fig. 11 shows the silencing efficacy of the SLN_EE-based co-delivery vectors in IMFE-1 cells, determined by quantifying Gb3S mRNA expression by RT-qPCR 48 h post-transfection. A scrambled siRNA sequence which did not induce gene silencing was formulated in SLN-based co-delivery vectors prepared with pEGFP, as a negative control; this did not result in any reduction in Gb3S levels, which confirms that the observed silencing efficacy stems from the therapeutic siRNA.

Fig. 11.

Gb3S gene silencing efficacy of the co-delivery vectors in IMFE-1. Data are expressed as mean ± standard deviation; n = 4.* p < 0.001 between the formulations. # p < 0.001 between the vectors prepared with either DOTAP or DOTAP/ MC3.

Reducing the GNs amount in the nanovectors enhanced the silencing effect in all SLN_ET and SLN_EMC3–based formulations prepared with P3. For Dx-vectors containing P3 and GN5 (DxP3GN5-ET and DxP3GN5-EMC3), the silencing efficiency was around 10%. Incorporation of SLN_EMC3 increased the silencing activity in DxP3GN2.5-EMC3 and HAP2GN5-EMC3 vectors compared with their SLN_ET-based counterparts. Overall, the formulations showing the highest silencing activity (≈70%) were: DxP2GN2.5-ET, HAP2GN2.5-ET, HAP3GN2.5-ET, DxP2GN2.5-EMC3, DxP3GN2.5-EMC3, HAP2GN5-EMC3, HAP3GN2.5-EMC3 and HAP3GN2.5-EMC3. The silencing efficacy of vectors formulated with siRNA alone (without pDNA) was also evaluated (Fig. S11). In this case, vectors containing SLN_ET achieved silencing levels of approximately 80–90%, exceeding those obtained of the corresponding SLN_ET co-delivery formulations. For SLN_EMC3 vectors prepared only with siRNA, the silencing capacity was around 70%.

3.11. Gb3S enzyme expression

The expression of Gb3S enzyme was studied by immunocytochemistry in IMFE-1 cells 7 and 15 days post-treatment. Fluorescence microscopy images of non-treated cells and cells treated with selected co-delivery vectors at both time points are shown in Fig. 12, while images for all tested formulations are provided in Fig. S12. In non-treated cells, green fluorescence corresponding to Gb3S was very abundant and it was detected in every cell. In contrast, cells treated with the co-delivery vectors displayed only minimal Gb3S fluorescence 7 days post-transfection. At day 15, Gb3S levels remained low across all conditions, although fluorescence intensity was higher than at day 7.

Fig. 12.

Immunocytochemistry assay of Gb3S in IMFE-1 cells transfected with co-delivery vectors. Blue: Nuclei labeled with DAPI. Green: Fluorescence signal of Gb3S. Magnification: 20 x. Scale bar: 100 μm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

To assess whether the presence of pDNA influenced the silencing efficacy, Gb3S expression was also evaluated following treatment with vectors containing siRNA alone. In these samples (Fig. S13), Gb3S fluorescence was barely detectable at day 7, and although the signal increased by day 15, its intensity remained clearly lower than in non-treated cells.

3.12. Gb3 quantification

The capacity of co-delivery vectors to reduce Gb3 synthesis was also quantified. Gb3 levels were quantified 7 and 15 days after transfection of IMFE-1. At both time points, all vectors significantly reduced Gb3 levels (Fig. 13A), and with some formulations, Gb3 levels at day 7 were below the lower limit of quantification. At day 15, the most effective vectors in reducing Gb3 were DxP3GN2.5EMC3 and HAP2GN2.5EMC3.

Fig. 13.

Gb3 quantification in IMFE-1 cells after the treatment with different vectors. (A) Gb3 quantification in IMFE-1 cells after the treatment with different co-delivery vectors. (B) Gb3 quantification in IMFE-1 cells 7 and 15 days after the treatment with different pDNA or siRNA vectors. NT: Non-treated cells. Data are expressed as mean ± standard deviation; n = 2. * p < 0.001 with respect to the cells treated with the formulations at 7 days. # p < 0.001 with respect to the cells treated with the formulations at 15 days. Dotted line: lower limit of quantification.

Fig. 13B shows Gb3 levels in IMFE-1 cells treated with vectors containing either pDNA or siRNA separately. Overall, pDNA-vectors produced a greater reduction in Gb3 than siRNA-vectors. In general, single–nucleic acid vectors (pDNA or siRNA) exhibited a lower capacity to reduce Gb3 levels compared with co-delivery vectors of the same composition (P, polysaccharide, and SLN).

To better track the evolution of Gb3 levels over time, the change from day 7 to day 15 is shown in Fig. 14. Although Gb3 levels increased with most from 7 to 15 days, the rate of increase was generally lower than in non-treated cells.

Fig. 14.

Change in Gb3 concentrations from 7 to 15 days after the treatment with different co-delivery vectors. (A) Gb3 concentration 7 and 15 days after the treatment with SLN_ET co-delivery vectors. (B) Gb3 concentration 7 and 15 days after the treatment with SLN_EMC3 co-delivery vectors. Gb3 levels in cells treated with pDNA and siRNA vectors are also included. Dotted line: lower limit of quantification.

3.13. Cell viability

Cell viability was measured with the MTT assay 24 h after the transfection of co-delivery vectors in IMFE-1 cells. In Fig. 15, it is observed that the co-delivery vectors resulted in a reduction of cell viability, although it was always greater or equal than 80%. Those vectors containing SLN_EMC3 provided lower cell viability than SLN_ET-vectors_, especially the formulations DxP3GN2.5-EMC3 and HAP3GN2.5-EMC3. In the case of siRNA-vectors, cell viability was around 100% (Fig. S14).

Fig. 15.

Viability in IMFE-1 cells 24 h after the addition of co-delivery vectors. NT: Non-treated cells. Data are expressed as mean ± standard deviation; n = 3. * p < 0.001 between the formulations.

3.14. Interaction with erythrocytes: Hemolysis and hemagglutination

The hemagglutination capacity of the co-delivery vectors after incubation with erythrocytes is shown in Fig. 16. Vectors containing Dx induced red blood cell agglutination, particularly those formulated with SLN_ET. For these vectors, increasing the P proportion (P3) reduced the agglutination effect. In contrast, no agglutination was observed with vectors containing HA.

Fig. 16.

Hemagglutination effect of co-delivery vectors. Positive control: Poly-l-Lysine. Scale bar: 50 μm.

As observed in Fig. 17, the hemolytic activity of the co-delivery vectors was lower than 15% with all formulations.

Fig. 17.

Hemolytic activity of co-delivery vectors. Lysis buffer represents 100% hemolysis. Results are shown as the mean ± standard deviation; n = 3.

4. Discussion

Nanotechnology possesses the opportunity to design therapeutic systems tailored to the nucleic acid, disease, and target cell, enabling more effective and personalized medicines (Beraza-Millor et al., 2024). In this context, vectors capable of co-delivering multiple nucleic acids within a single platform represent a promising strategy for complex disorders, such as LSDs, while also reduce manufacturing and regulatory burdens (Ball et al., 2018). Previous studies have explored the co-delivery of pDNA and siRNA across diverse therapeutic contexts using various nanomaterial platforms; including arginine-rich oligopeptide-grafted branched polyethylenimine (PEI) for breast cancer (Lu et al., 2015), poly-γ-glutamic acid into PEI complexes against hepatoma (Peng et al., 2017) or GN coated with degradable polymers for brain cancer treatment (Bishop et al., 2015).

In this work, co-delivery nanosystems based on SLNs and GNs (golden SLNs) were developed, establishing a foundation for an efficient combined gene therapy platform. Co-delivering pDNA and siRNA within a single nanocarrier requires the control of cellular uptake and endosomal escape, as well as the balanced condensation and intracellular release of two nucleic acids with markedly different sizes and binding thermodynamics (Scholz and Wagner, 2012; Zhu et al., 2022). Compared to other systems based on lipid nanoparticles or inorganic carriers (Li et al., 2022; Zhu et al., 2022), the combination of SLNs and GNs takes advantage of SLNs to promote the cellular uptake and endosomal escape, and the ability of GNs to provide a complementary condensation and stabilization of the nucleic acids.

These nanosystems were applied to FD, enabling the simultaneous delivery of α-Gal A-encoding pDNA and Gb3S-targeting siRNA for combined gene supplementation and gSRT-mediated suppression strategy. The functional relevance of the designed nanomedicines was evaluated in IMFE-1 cells. These cells accumulate Gb3 within lysosomes due to reduced α-Gal A activity, and provide a relevant platform for studying FD mechanisms and evaluating potential therapeutic strategies (Ruiz-de Garibay et al., 2015; Shen et al., 2007).

SLN-based vectors have been studied for different applications (del Pozo-Rodríguez et al., 2016; Gómez-Aguado et al., 2021), including gene supplementation for FD (Rodríguez-Castejón et al., 2025, Rodríguez-Castejón et al., 2022, Rodríguez-Castejón et al., 2021; Ruiz-de Garibay et al., 2015) and, recently, for gSRT (Beraza-Millor et al., 2023). The present study shows that the preparation sequence is critical to obtain stable SLN-based codelivery systems; the optimal protocol consisted of combining siRNA, GNs, and SLNs first, followed by the addition of pDNA complexed with P and polysaccharides. This procedure ensured efficient nucleic acid condensation, protection, and release, as demonstrated by agarose gel electrophoresis together with PicoGreen™ and RiboGreen™ assays.

SLNs were prepared by different methods, with HM producing smaller nanoparticles likely due to high-energy homogenization (Graván et al., 2023). The inclusion of ionizable lipids (MC3 or DODAP) increased particle size compared with formulations containing only DOTAP as the cationic lipid, while all SLNs maintained a positive ζ-potential, necessary to promote nucleic-acid adsorption.

The coating of SLNs with different ligands via electrostatic interactions is known to produce stable nanomedicines in which the nucleic acids remain adsorbed on the positively charged surface (Beraza-Millor et al., 2023; Delgado et al., 2011; Gómez-Aguado et al., 2021; Rodríguez-Castejón et al., 2025). The polycationic peptide P is useful because of its ability to enhance nuclear uptake of the genetic material (Delgado et al., 2012); in the co-delivery nanosytems P promoted pDNA condensation but limited siRNA release. Polysaccharides condition vector stability and provide additional benefits, such as stealth properties and modulation of cellular uptake mechanisms (Apaolaza et al., 2014; Delgado et al., 2012; Gómez-Aguado et al., 2020). Herein, HA improved intracellular pDNA availability in IMFE-1 cells compared with Dx (Fig. 9). The final structures allowed vectors to present a complete nucleic-acid loading, as evidenced by the absence of free cargo in agarose gels (Fig. 6) and Picogreen/Ribogreen assays (Fig. S8) together with the consistent biological responses observed in vitro. The formation of golden SLNs-based vectors bearing siRNA, as well as the final structure were characterized by TEM and cryo-TEM in a previous work, in which the presence and distribution of GNs within the lipid nanocarriers were clearly demonstrated (Beraza-Millor et al., 2023).

The reduction or no significant variation in particle size after vector assembly observed here has been also reported in previous studies evaluating SLNs as delivery systems for pDNA. This behavior has been attributed to the strong nucleic acid-condensing capacity of compact nanostructures (del Pozo-Rodríguez et al., 2009; Rodríguez-Castejón et al., 2025; Vicente-Pascual et al., 2020). In the present work, GNs additionally contribute to cargo compaction through their known interactions with nucleic acids. The combined action of P-mediated condensation and GNs assisted nucleic acid binding can therefore lead to tighter packing of the surface associated nucleic acids and partial collapse of surface exposed flexible domains resulting in slightly smaller hydrodynamic sizes despite additional components being incorporated (Graczyk et al., 2020).

GNs played a key role in co-formulating multiple nucleic acids within a single nanomedicine. Indeed, in a previous work GNs were necessary for stable assembly of SLN-based siRNA systems (Beraza-Millor et al., 2023). Formulation performance is affected by nucleic-acid size, dose, and nucleobase composition, since nucleotide adsorption onto GNs occurs through specific base-surface interactions, with strength order adenine > cytosine > guanine > thymine ≈ uracil (Dutour and Bruylants, 2025; Zhang et al., 2012). Fluorescence microscopy images (Fig. 9) shows that increasing GNs content in co-delivery vectors enhanced pDNA condensation, and regarding siRNA, strong condensation in golden SLNs by GNs, P and cationic lipids limits its intracellular detection, as previously reported (Beraza-Millor et al., 2023). The degree of condensation of the nucleic acids may justify the fact that the efficacy of the co-delivery vectors in IMFE-1 cells, considering α-Gal A activity, Gb3S downregulation, and intracellular Gb3 levels, was higher in formulations with lower P and GN content.

SLN preparation also strongly influenced transfection efficiency; SLN_HM nanomedicines, despite identical composition, exhibited very low transfection (<2%) compared with SLN_EE systems, highlighting the critical role of technological factors in determining SLN physicochemical properties and their impact on cellular behavior.

The cationic ζ-potential enhances electrostatic interactions with negatively charged membrane glycoproteins and proteoglycans (Fröhlich, 2012). In this sense, IMFE-1 cell-association rate of golden SLN-based nanomedicines was close to 100%, and a higher fluorescence intensity per cell was detected with SLN_EMC3 formulations, reflecting enhanced internalization efficiency (Fig. 8B). Internalization mechanism and uptake rates, influenced by vector physicochemical properties, directly affect transfection efficiency (Delgado et al., 2012). Lipid-based nanoparticles are mainly internalized via clathrin- and caveolae-mediated endocytosis, macropinocytosis and phagocytosis (Rennick et al., 2021; Schlich et al., 2021), although energy-independent mechanisms may also contribute, with pathway relevance depending on the cell type (Duncan and Richardson, 2012). In IMFE-1 cells, clathrin-, caveolae-mediated routes, and micropinocytosis may participate in co-delivery vectors uptake (Fig. 7).

After uptake, nanoparticles traffic to early endosomes, where acidification and enzymatic activity can degrade nucleic acids, making efficient endosomal escape essential for effective intracellular delivery and transfection (Schlich et al., 2021). Lipid composition further modulates intracellular delivery; ionizable lipids DODAP and MC3 remain neutral at physiological pH, acquiring positive charge in acidic endosomes which facilitates nucleic-acid release by membrane destabilization (Basha et al., 2011; Gilleron et al., 2013; Cheng and Lee, 2016; Sun and Lu, 2023). These mechanistic features align with our experimental observations, as labeled pDNA was more readily detected in the cytoplasm of cells treated with SLN_EMC3-co-delivery vectors, consistent with enhanced endosomal escape and higher transfection efficacy (Fig. 3 and Fig. 10), reinforcing the improved intracellular availability achieved with these formulations. However, DODAP-based vectors showed lower EGFP expression than MC3, which could be due to differences in the structure of the hydrophobic domain and in lipid pKa (6.44 for MC3 vs. 5.59 for DODAP) (Gómez-Aguado et al., 2022), that influence endosomal destabilization capacity (Basha et al., 2011; Gilleron et al., 2013; Jayaraman et al., 2012). Early endosomal escape is also necessary to achieve sufficient silencing activity (Patel et al., 2017), since the amount of siRNA reaching the cytosol directly determines therapeutic efficacy (Samaridou et al., 2020). In fact, Gb3S gene silencing, with effects lasting for at least 15 days, confirmed SLN_EMC3 vectors as the most effective.

These results demonstrate, for the first time, the efficacy of a nanomedicine combining gene supplementation and gSRT-mediated suppression for the treatment of FD. Vectors containing the αGLA-pDNA without siRNA reduced Gb3 more effectively than vectors containing only siRNA, yet co-delivery vectors achieved greater Gb3 reduction than either single-nucleic acid vector of identical composition, highlighting the additive potential of combining gene supplementation and gSRT. This strategy may be particularly relevant for heart and kidney, the most severely affected organs in FD (Branton et al., 2002; Felis et al., 2020). Indeed, one of the primary objectives of FD therapies is to reduce Gb3 deposits specifically in these organs (Kant and Atta, 2020; Miller et al., 2020; van der Veen et al., 2020). Delivering pDNA and siRNA from the same nanomedicine guarantees the effect of both nucleic acids in the same cell, with α-Gal A activity reducing existing Gb3 deposits and gSRT decreasing de novo Gb3 synthesis, while also enabling cross correction. Supplementation gene therapy allows cross correction because secreted α-Gal A can be taken up by neighboring or distant cells via mannose-6-phosphate receptors and trafficked to lysosomes, extending its therapeutic effect beyond the initially transfected cells (Domm et al., 2021; Pacienza et al., 2012).

In this context, although recent studies have reported intrinsic therapeutic effects of ultrasmall ligand-decorated GNs in kidney fibrosis (Chan et al., 2023); neurodegeneration (Lee et al., 2026); and lung inflammation (Liu et al., 2026), in the present FD platform GNs were incorporated at low loading within DOTAP/MC3-SLNs primarily as a technological component to stabilize the co-loading and release of pDNA and siRNA. Accordingly, the observed functional outcomes, namely increased α-gal A activity and Gb3S knockdown, are cargo-driven. Nevertheless, the possibility that GN-mediated organ-specific effects could further contribute in tissues affected in FD, such as kidney or the central nervous system, warrants future investigation.

In the transfection studies, the co-delivery vectors remained effective even in presence of serum, without compromising IMFE-1 cells viability. In biological environments, nanoparticles are rapidly coated by biomacromolecules, primarily proteins, forming a protein corona that influences toxicity, biodistribution, clearance, and immune response (Kobos and Shannahan, 2020; Kopac, 2021; Peigneux et al., 2020; Pinals et al., 2020; Westmeier et al., 2016). Transfection in serum-containing media better mimics in vivo conditions, highlighting the relevance of evaluating vector performance under physiologically realistic settings. Since the golden SLN-based co-delivery nanomedicine is intended to be administered intravenously, early evaluation of hematocompatibility is essential. Nanoparticles interact with blood and immune components, where thrombosis and immune activation are commonly reported side effects (Halamoda-Kenzaoui et al., 2019; Halamoda-Kenzaoui and Bremer-Hoffmann, 2018). Features such as particle size, surface charge, and surface ligands determine protein corona formation, which in turn influences hematological toxicity (Moraru et al., 2020; Scioli Montoto et al., 2020).

Co-delivery vectors showed no hemolytic activity, whereas the type of polysaccharide had an impact on erythrocyte agglutination. Only Dx-vectors exhibited some hemagglutination, likely reflecting the superior stealth properties of HA. In a previous work, TEM images of HA-SLN vectors revealed a smooth, well-formed spherical layer surrounding the particle surface, possibly due to HA segment loops (Apaolaza et al., 2014). Furthermore, Dx-based vectors containing the ionizable lipid MC3 displayed lower hemagglutination than DOTAP formulations, consistent with the reduced number of positive charges on the surface of MC3 vectors. Nevertheless, in previous studies with Dx-SLN vectors prepared with DOTAP bearing a single type of nucleic acid (pDNA or siRNA), no hemagglutination was detected (Beraza-Millor et al., 2023; Delgado et al., 2011). These observations indicate that the strategy of combining more than one type of nucleic acid in a vector affects not only therapeutic efficacy but also the safety profile, emphasizing the importance of rational design in co-delivery nanomedicines.

This study was conceived as a development and optimization stage for the dual co-delivery nanomedicine. Given the number of interdependent formulation variables (assembly sequence, lipid composition, polysaccharide coating, GNs loading and P ratio), a relevant in vitro model of FD (IMFE-1 cells) was prioritized to efficiently characterize and evaluate the designed therapeutic strategy. In vivo evaluation in FD mouse models, addressing biodistribution, kidney and heart pharmacodynamics, and safety-represents the immediate next step beyond the scope of the present work. Importantly, prior studies from the research group have already demonstrated the feasibility of SLN-based gene delivery in Fabry mice (Rodríguez-Castejón et al., 2025, Rodríguez-Castejón et al., 2022, Rodríguez-Castejón et al., 2021), supporting the planned in vivo transition. Future studies performed with this new platform will therefore focus on confirming organ-level Gb3 reduction and systemic safety in vivo.

5. Conclusions

We report a hybrid platform based on GNs and SLNs that co-delivers GLA pDNA and siRNA against Gb3S, enabling same-cell gene supplementation and substrate-reduction mechanisms in a FD setting. The process of production and the composition were critical. On the one hand, nanovectors obtained by emulsification/evaporation outperformed in efficacy those prepared by hot-melt emulsification; on the other hand, MC3-containing systems provided the best balance between uptake and intracellular availability. Protamine and GNs governed nucleic-acid condensation and release; notably, lower loadings of GNs improved functional readouts, indicating that fine-tuning condensation is essential for cytosolic delivery. In IMFE-1 cells, the optimized co-delivery vectors increased α-gal A activity, downregulated the synthase target and lowered intracellular Gb3 for at least15 days, with cell viability over 80% and low hemolysis; HA coatings further enhanced hemocompatibility. Collectively, these data validate a co-delivery platform that integrates complementary gene therapy approaches within a single nanomedicine and support its potential to expand therapeutic avenues for FD. The main limitation of this study is its in vitro scope; therefore, in vivo evaluation in Fabry mice constitutes the immediate next step to confirm organ level Gb3 reduction and safety.

CRediT authorship contribution statement

Marina Beraza-Millor: Writing – original draft, Methodology, Investigation, Formal analysis, Conceptualization. Paula Fernández-Muro: Writing – review & editing, Methodology, Investigation. Madalen Arribas-Galarreta: Writing – review & editing, Methodology. Ana del Pozo-Rodríguez: Writing – review & editing, Software, Resources, Methodology, Funding acquisition, Formal analysis. Alicia Rodríguez-Gascón: Writing – original draft, Validation, Supervision, Resources, Project administration, Funding acquisition, Formal analysis, Conceptualization. María Ángeles Solinís: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Funding

Beraza-Millor and Fernández-Muro were supported by grants of the University of the Basque Country UPV/EHU (PIF21/61 and PIF 23/45, respectively) and Arribas-Galarreta by the Basque Government (PR E_2022_1_00144). This work was funded by MCIU/AEI/FEDER, UE (RTI2018–098672-B-I00) and by the Basque Government (IT1587-22).

Declaration of competing interest

No conflicts of interest exist in the submission of this manuscript, and the manuscript has been approved by all authors for publication. This work described was original research that has not been published previously and is not under consideration for publication elsewhere, in whole or in part. All the authors listed have approved the manuscript for publication. If accepted, the article will not be published elsewhere in the same form, in English or in any other language, including electronically, without the written consent of the copyright-holder.

Data availability

Data will be made available on request.