1. Introduction
Multiple sclerosis (MS) is a chronic, immune-mediated disease of the central nervous system (CNS), characterized by inflammatory demyelination and progressive neuroaxonal degeneration, ultimately leading to cognitive, sensory, and motor impairment [ 1 , 2 , 3 ]. Although its exact etiology remains incompletely understood, neuroinflammation and oxidative stress are recognized as major contributors to disease progression [ 4 , 5 ]. Accordingly, several therapeutic strategies currently employed in MS aim to modulate peripheral immune responses while limiting CNS damage and slowing disease progression [ 6 ].
Dimethyl fumarate (DMF), a fumaric acid ester approved by the EMA (Tecfidera TM ) for the treatment of relapsing-remitting MS, is among the most widely used disease-modifying therapies for this purpose. Besides its established clinical efficacy, DMF exhibits potent anti-inflammatory, antioxidant, and neuroprotective properties mainly associated with activation of the nuclear factor erythroid 2-related factor 2 (Nrf2) pathway and the subsequent induction of cytoprotective antioxidant responses [ 7 , 8 ]. These properties have attracted considerable interest in the development of novel strategies aimed at maximizing its therapeutic potential within the CNS.
Following oral administration, however, DMF undergoes extensive and rapid hydrolysis to monomethyl fumarate (MMF), which becomes the predominant species in both humans and animals. Consequently, no measurable plasma concentrations of the parent DMF can generally be detected after oral administration [ 9 , 10 , 11 ]. MMF is widely regarded as the main active metabolite responsible for the therapeutic effects of DMF [ 10 , 12 ]. Recent studies have further demonstrated that DMF is rapidly converted to MMF even following intravenous administration in rats [ 13 ]. Moreover, it is known that only ~5% of administered DMF can be recovered in the bloodstream as MMF after both intravenous and oral administrations [ 13 , 14 ].
MMF, which has itself been approved by the FDA (Bafiertam TM ) for relapsing MS, is known to cross the blood–brain barrier (BBB), although CNS concentrations remain substantially lower than those measured in the systemic circulation [ 13 , 15 ]. Nevertheless, increasing evidence suggests that the parent drug may exert pharmacological activities distinct from those of its metabolite. In particular, DMF displays greater membrane permeability, broader intracellular reactivity, and stronger neuroprotective and immunomodulatory effects than MMF in several in vitro models, including a greater ability to suppress the production of pro-inflammatory cytokines [ 16 ]. Furthermore, unlike MMF, whose intracellular activity largely depends on receptor-mediated mechanisms, DMF readily penetrates cellular membranes and interacts with multiple intracellular targets [ 12 , 16 , 17 , 18 , 19 ]. These observations suggest that increasing direct CNS exposure to DMF may provide additional therapeutic benefits beyond those achievable through systemic exposure to MMF alone. Another important limitation of DMF therapy is the occurrence of gastrointestinal adverse effects, which frequently reduce patient compliance and may compromise long-term treatment efficacy [ 20 ]. Alternative delivery approaches capable of enhancing CNS exposure while minimizing systemic and gastrointestinal side effects are therefore highly desirable. Among these approaches, intranasal administration has emerged as a promising non-invasive strategy for CNS drug delivery. By exploiting the olfactory and trigeminal pathways, drugs may bypass the BBB and reach the CNS through direct nose-to-brain transport mechanisms, including delivery to the cerebrospinal fluid (CSF) via permeation across the olfactory mucosa and neuronal transport through intracellular and extracellular routes [ 21 , 22 , 23 , 24 ]. In addition to its potential for CNS targeting, the nasal route offers several practical advantages, including rapid absorption, ease of administration, and avoidance of first-pass metabolism [ 25 ]. These characteristics make intranasal delivery particularly attractive for DMF, whose therapeutic potential may be limited by rapid systemic biotransformation. However, successful nose-to-brain delivery requires formulations capable of enhancing drug solubility and stability, promoting efficient permeation across the nasal mucosa, and ensuring appropriate release profiles to maximize therapeutic efficacy [ 26 ]. Several formulation approaches have recently been investigated to overcome the poor aqueous solubility and limited stability of DMF, including lipid-based carriers and nanotechnological delivery systems [ 13 , 27 , 28 , 29 , 30 ].
Among these, nanoemulsions (NEs) have attracted considerable interest because of their ability to solubilize lipophilic compounds, improve formulation stability, and enhance drug interaction with biological membranes. NEs are colloidal systems composed of oil and aqueous phases stabilized by surfactants, typically exhibiting droplet sizes in the range of 20–500 nm [ 31 , 32 ]. Due to their small droplet size, NEs provide a large interfacial area that can facilitate mucosal interaction, drug absorption, and rapid release, making them particularly suitable for intranasal drug delivery applications [ 33 , 34 , 35 , 36 , 37 , 38 ].
Geraniol (GER), a naturally occurring monoterpene alcohol widely present in essential oils, has recently attracted attention for its antioxidant, anti-inflammatory, and neuroprotective properties [ 39 , 40 , 41 , 42 , 43 ]. Previous studies demonstrated that incorporating GER into colloidal formulations enhances its permeation across biological membranes [ 39 , 44 ], thereby potentially improving its pharmacological efficacy. Moreover, nasal formulations based on cyclodextrin complexes and nanoparticulate systems have already been proposed to improve its brain delivery [ 42 , 43 ]. Beyond its pharmacological activity, GER represents a suitable lipophilic phase for the incorporation of DMF, potentially contributing both to drug solubilization and to the therapeutic profile of the final formulation. The combination of DMF and GER may therefore represent an innovative strategy to simultaneously enhance drug delivery and provide complementary neuroprotective effects.
Chitosan, a biocompatible and biodegradable polysaccharide, is widely used in drug delivery due to its mucoadhesive, antimicrobial, and permeability-enhancing properties. Its cationic nature promotes interaction with negatively charged biological membranes and transient opening of tight junctions, thereby enhancing paracellular transport and mucosal retention, improving paracellular permeability and drug retention at mucosal sites [ 45 , 46 ]. Among its derivatives, chitosan oleate (CSOA), an amphiphilic ionic complex of chitosan and oleic acid, has previously demonstrated suitability as both a self-assembling carrier to obtain polymeric micelles [ 47 ] and a stabilizer for NE systems [ 33 ]. In the first case, CSOA is exploited for its auto-assembling behavior, enhancing hydrophobic drug loading in core hydrophobic domains. In NEs, CSOA contributes to stabilization, enhancing the physical stability of the dispersed phase, while preserving the mucoadhesive and penetration-enhancing properties of chitosan, potentially improving nasal absorption through the combined effects of both polysaccharide and oleic acid [ 44 ]. Based on these considerations, the aim of the present study was to develop and characterize a GER-based NE containing DMF for intranasal administration and potential nose-to-brain delivery. The formulation was stabilized using CSOA and characterized in terms of physicochemical properties, drug loading, stability, release profile, antioxidant activity, and permeation across RPMI 2650 nasal epithelial cells. Finally, based on the results of the first phase of pharmaceutical development, a NE formulation was prepared for in vivo administration, and an in vivo proof of concept study in rats was performed to evaluate the ability of the formulation to promote the delivery of DMF and GER to the CNS following intranasal administration.
2. Materials and Methods
2.1. Materials
Dimethyl fumarate (DMF, 99%) was purchased from AK Scientific Inc. (Union City, CA, USA). Monomethyl fumarate (MMF, 97%), geraniol (GER, 98%), ferulic acid (Fer), carbazole, carbamazepin (CBZ), cellulose membrane (cut-off 14 kDa, 1.96 mL/cm), ascorbic acid (AA, 2,2-diphenyl-1-picrylhydrazyl (DPPH), 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) (MTT) and bovine serum albumin (BSA) were purchased from Sigma Aldrich, Co. (St. Louis, MO, USA). Chitosan base (CS, batch number: STBF8219V, deacetylation degree: 75–85%, molecular weight about 187 kDa, viscosity: 20–300 mPas) was acquired from Sigma-Aldrich (Milan, Italy), while low-molecular-weight chitosan-HCl (CS-HCl, batch number: 312-090216-01, molecular weight: about 165 kDa, viscosity < 20 mPas) was acquired from Heppe Medical Chitosan GmbH (Halle, Germany). Both chitosans are obtained by the deacetylation of chitin. The chitin for HMC’s chitosan comes mainly from crustacea ( Chionoecetes opilio ).
Oleic acid was purchased from Fluka Analytical (Neu-Ulm, Germany). Acetic acid glacial, formic acid (99%), ethanol (96°), methanol (99%) and acetonitrile (99%) were acquired from Carlo Erba Reagents (Val-de-Reuil, France).
Ultrapure water was obtained from a MilliQ ® apparatus (Millipore, Burlington, MA, USA).
2.2. Physicochemical Characterization
2.2.1. Solubility Studies of Dimethyl Fumarate in Geraniol
Solubility studies of DMF in GER were carried out by preparing a supersaturated solution at 25 ± 1 °C. DMF was gradually and incrementally added to 1 mL of GER under continuous magnetic stirring, with a total of approximately 50 mg of DMF introduced. The solution was then stirred overnight at 300 rpm, until equilibrium was achieved. The resulting solution was then centrifuged at 10,000 rpm for 3 min, and the supernatant was collected, filtered and analyzed for DMF content. The solubility results were reported as mean ± standard deviation (SD).
2.2.2. Fourier Transform Infrared Spectroscopy
Fourier Transform Infrared (FT-IR) spectra were recorded using a Perkin Elmer Spectrum One spectrophotometer (Perkin Elmer, Monza, Italy) equipped with an ATR accessory (MIRacle™ device, Pike Technologies, Madison, WI, USA) fitted with a zinc selenide (ZnSe) crystal. The spectra were acquired in transmittance mode, within the mid-infrared (Mid-IR) range of 650 to 4000 cm −1 , using 64 scans at a resolution of 4 cm −1 at ambient conditions. Each measurement was collected at least in triplicate.
2.2.3. Thermal Analysis
Differential scanning calorimetry (DSC) analyses were carried out on DMF, GER, and on their physical mixtures at different weight ratios, using a Mettler Star e System (Mettler Toledo, Milan, Italy) coupled with a DSC 821 e Module and an Intracooler device for sub-ambient temperature analysis (Julabo FT 900, Seelbach, Germany). The instrument was calibrated with Indium as a reference standard. Samples (2–12 mg), accurately weighed with a Mettler M3 Microbalance, were hermetically sealed in aluminum crucibles (40 μL capacity) with lids. Measurements were performed under a nitrogen atmosphere (50 mL/min) at a heating rate of 10 K/min, in a temperature range of 30–350 °C.
Thermogravimetric analysis (TGA) was performed on the same samples using a Mettler Star e System TGA/DSC1 (Mettler Toledo, Milan, Italy), also calibrated with Indium as a reference standard. Samples (3–20 mg) were placed in alumina crucibles (70 μL capacity) with enclosed lids. All measurements were performed at least in triplicate.
2.2.4. Preparation of Chitosan Oleate
Chitosan oleate (CSOA) was prepared in 1:0.5 and 1:1 (mol/mol) ratios, following a modified protocol from previous studies [ 44 ]. Briefly, 1.5 g of chitosan base (CS) was accurately weighed using an analytical balance (ADJ 100-4, max 120 g, d = 0.1 mg, Kern & Sohn, Balingen, Germany) and added to a reaction flask with 7.5 mL of ethanol and 0.3 g of acetic acid. The mixture was stirred continuously using a magnetic stirrer (ARE Heating Magnetic Stirrer, Velp Scientifica, Usmate Velate, Italy). In a second flask, either 1.8 g or 0.9 g of oleic acid (corresponding to 1:0.5 and 1:1 chitosan-to-oleic acid ratios, respectively) were dissolved in 7.5 mL of ethanol under magnetic stirring until complete solubilization. The final solution was gradually added to the first reaction mixture and then stirred overnight. Ethanol was removed the following day using a rotary evaporator (Laborota 4000, Heidolph, Schwabach, Germany) at 45 °C and 150 rpm for approximately 20 min. To remove excess oleic acid, the resulting solid was divided into two 14 mL conical tubes, each containing 6 mL of ethanol; the tubes were centrifuged (SL8, Thermo Scientific, Waltham, MA, USA) at 3000 rpm for 15 min, and the supernatant was discarded. This washing step was repeated once more. The remaining pellet was dried in a crystallizer for 24 h, yielding the CSOA powder.
2.2.5. Preparation of Blank and Dimethyl Fumarate-Loaded Nanoemulsions
In this study, both blank and DMF-loaded nanoemulsions (NEs) were prepared using two different preparative methods: phase inversion emulsification and spontaneous emulsification. All NEs were prepared with the same main components: GER as the oil phase, ultrapure water as the aqueous phase, and CSOA as the amphiphilic surfactant.
2.2.6. Phase Inversion Emulsification Method
The first method used for NE preparation involves a combination of phase inversion emulsification (PI) with a high-energy method using an Ultra-Turrax T18 basic homogenizer (Janke & Kunkel, IKA Labortechnik, Staufen, Germany). A water-in-oil (W/O) emulsion was initially formed by adding CSOA at chitosan-to-oleic acid molar ratios of 1:0.5 and 1:1, at final concentrations of 2% and 3%, along with acetic acid (for chitosan solubilization), into the oil phase. The minimum volume of MilliQ ® ultrapure water needed to achieve a W/O dispersion was then gradually added. Emulsification was performed using the Ultra-Turrax, starting at 3600 rpm for the first 3 min while the remaining MilliQ ® ultrapure water was added dropwise, inducing phase inversion. The resulting emulsion was centrifuged at 7000 rpm for 3 min (Mini centrifuge LLG-uniCFUGE5, LLG Labware, Meckenheim, Germany), and subsequently sonicated (Sonic Dismembrator, Fisher Scientific, Waltham, MA, USA, 3 mm probe) for 18 min at 20% amplitude, with a pulsed cycle of 10 s on and 10 s off. To ensure a stable temperature, an ice bath was used to dissipate any heat generated during this process.
DMF-loaded NEs were prepared following the same protocol, with 5 mg of DMF added to 250 mg of the oil phase at the start of the process.
2.2.7. Spontaneous Emulsification Method
In this method, 6 mg of chitosan-HCl (CS-HCl) was dissolved in 6 mL of MilliQ ® ultrapure water. The lipophilic phase was prepared by dissolving 8.4 mg of oleic acid (maintaining a 1.4 mg:1 mg ratio with CS-HCl) and 14.4 mg of GER (to achieve a CSOA:GER, ratio of 1:1) in the minimum ethanol volume required for complete solubilization. The organic phase was then added dropwise to the aqueous phase under continuous magnetic stirring at 500 rpm. Diffusion of the organic phase into the external aqueous phase induced spontaneous formation of an oil-in-water (O/W) NE. The emulsion was stirred overnight to allow partial ethanol evaporation, with any residual solvent removed under a nitrogen stream. The final sample underwent ultrasonication (Sonic Dismembrator, Fisher Scientific, 3 mm probe) for 15 min at 20% amplitude, using a pulsed cycle of 10 s on and 10 s off.
The DMF-loaded NE was prepared following the same procedure by adding different amounts of DMF, as indicated later in the results section, which were incorporated into the organic phase before emulsification.
2.2.8. Determination of Size and Zeta Potential
The droplet size of the NEs was determined by photon correlation spectroscopy (PCS) with an N5 submicron particle analyzer (Beckman Coulter, Milan, Italy) and by Dynamic Light Scattering (DLS) with a Zetasizer Nano ZS90 (Malvern Instruments, Worcestershire, UK). PCS measurements were carried out at 25 °C with a fixed scattering angle of 90°, following appropriate sample dilution in MilliQ ® ultrapure water (1:10 to 1:100, v / v ). The zeta potential of the NEs was measured by DLS using an Omega cuvette (DTS1070, Malvern Instruments, Worcestershire, UK) under the same temperature and detection angle conditions. All measurements were performed in triplicate, and results are reported as mean ± SD.
2.2.9. Transmission Electron Microscopy
The ultrastructural characterization of unloaded and loaded NEs was carried out using transmission electron microscopy (TEM, Jeol-Jem-1200EXII, Tokyo, Japan). For sample preparation, a drop of DMF-loaded NE or its unloaded counterpart was placed on a copper grid and air-dried at 25 °C for 2 min. The grid was then stained with uranyl acetate prior to imaging.
2.2.10. Stability Studies
The stability of the different NEs was evaluated by monitoring droplet size over time. Particle size was measured weekly by PCS for up to six weeks in formulations prepared by the phase inversion and spontaneous emulsification methods. NEs were aliquoted and stored under two different conditions: refrigerated storage (4 ± 1 °C) and room temperature (25 ± 1 °C).
At each time point, measurements were performed in triplicate, and results were expressed as mean ± SD.
2.2.11. Quantification of Dimethyl Fumarate and Geraniol in Nanoemulsions
The samples of NEs (100 μL) were first diluted with 45 μL of acetate buffer 100 mM (pH 4) and with 855 μL of acetonitrile, the latter being added after vortexing. These samples were further diluted 1:10 using a mixture of acetonitrile:buffer acetate 95:5 ( v / v ) and analyzed by the high-performance liquid chromatography (HPLC) system below described, after filtration through regenerated cellulose filters (0.45 μm), for the detection and quantification of DMF and GER. For the preliminary characterization of the NEs, a specific HPLC method was adapted for the detection of both compounds in the same chromatogram [ 48 , 49 ].
In particular, the analyses were performed using an HPLC system (1100 Series, Agilent, Santa Clara, CA, USA) equipped with an endcapped C18 column (150 mm × 4.6 mm, 3 μm). The mobile phase consisted of H 2 O:formic acid (pH 3.0 ± 0.1; DL40GP, MemoTitrator, Mettler, Milan, Italy) and acetonitrile (60:40 v / v ) with a flow rate of 0.8 mL/min. Chromatographic separations were carried out at room temperature (25 ± 1 °C) under a pressure of 112 bar. Under these conditions, the retention times were 4.8 min for DMF and 13.18 min for GER. The injection volume was 20 μL. To avoid cross-contamination, the injection needle was rinsed with acetonitrile after each sample run. Detection was performed at a wavelength of 210 nm [ 49 , 50 ].
Calibration curves were constructed for both analytes. For DMF, a six-point calibration curve was prepared from a 1 mg/mL stock solution in methanol diluted in the mixture of acetonitrile:buffer acetate 95:5 ( v / v ) to cover the concentration range of 0.2–1.8 μg/mL, with carbamazepine (CBZ) (final concentration 8 μg/mL) as an internal standard. The assay showed good linearity within this range (y = 0.855x + 0.001; R 2 = 0.9977).
Similarly, for GER, a six-point calibration curve was generated from a 1 mg/mL stock solution in methanol diluted in the mixture of acetonitrile:buffer acetate 95:5 ( v / v ) to cover a range of concentrations of 2–50 μg/mL. Each sample also contained CBZ (8 μg/mL) as an internal standard. Also, for GER, linearity was confirmed within the selected range (y = 0.0304x − 0.0255; R 2 = 0.9971).
2.2.12. Recovery and Loading Capacity
Recovery% ( Rec %) and drug loading ( DL %) were determined on the loaded formulation by HPLC analysis using the conditions described above, using the following Equations (1) and (2) [ 51 ]: (1) R e c % = L D M F T D M F × 100 (2) D L % = L D M F T f o r m u l a t i o n × 100 where L DMF is the amount of DMF loaded in the NE, T DMF is the total amount of DMF added to the formulation, and T formulation is the total weight of the formulation (CSOA, DMF and GER).
All determinations were performed in triplicate in independent experiments, and results are expressed as mean ± SD.
2.2.13. In Vitro Release Studies
The in vitro release profile of the formulation containing DMF in GER was evaluated using a dialysis cellulose membrane with a molecular weight cut-off of 14 kDa. The membrane was filled with the sample and immersed in 140 mL of phosphate-buffered saline (PBS, pH 4) prepared according to the European Pharmacopoeia (Ph. Eur. 10.5 Ed.). The receiving medium was maintained at 34 ± 1 °C under constant magnetic stirring at 500 rpm. At predetermined time intervals (ranging from 5 to 240 min), 1 mL aliquots were withdrawn from the receiving medium and immediately replaced with an equal volume of fresh PBS to maintain sink conditions for both DMF and GER. Their concentrations in the release medium were determined by HPLC, as previously described. All experiments were conducted in triplicate, with results expressed as mean ± SD.
2.2.14. ROS-Scavenging Activity
The reactive oxygen species (ROS)-scavenging activity of NEs was evaluated by the 2,2-diphenyl-1-picrylhydrazyl (DPPH) colorimetric assay (Sigma Aldrich, Milan, Italy). Each formulation was tested after dilution at 1:3, 1:5 and 1:10 in MilliQ ® ultrapure water. For each measurement, a 100 μL aliquot of the sample was mixed with 900 μL of a DPPH solution (0.0112% in 70% v / v methanol) and incubated in the dark for 1 h at room temperature (25 ± 1 °C). A negative control, consisting of a methanol–DPPH mixture without NEs, was prepared for comparison. Following incubation, the samples were centrifuged at 15,000 rpm for 10 min, and the absorbance of the supernatant was recorded at 517 nm using a microplate reader (Synergy HT, BioTek, Swindon, UK). The ROS-scavenging activity percentage was calculated using the following Equation (3): (3) A b s o r b a n c e % = A c t r l − A s a m p A c t r l × 100 where A ctrl is the absorbance of the negative control, and A samp is the absorbance of the tested sample, corrected for the corresponding blank. Ascorbic acid (3.125–100 μg/mL) was used as a positive control. All experiments were carried out at least in triplicate and the results reported as mean ± SD.
2.2.15. Cell Culture
Immortalized human nasal epithelial cell line RPMI 2650, derived from an anaplastic squamous cell carcinoma of the human nasal septum, was obtained from the American Type Culture Collection (ATCC). Cells were used between passages P3 and P5 for in vitro cytotoxicity and permeability assays. The cells were maintained in T75 polystyrene culture flasks (SPL Life Sciences, Pocheon-si, South Korea) at a seeding density of 5 × 10 5 cells/cm 2 [ 52 ] and incubated at 37 °C with 5% CO 2 . The culture medium was refreshed every 2–3 days. Cell counts were determined using a Dual Fluorescence Cell Counter (LUNA-FL™, Logos Biosystems, Anyang-si, South Korea). Cell maintenance was performed in αMEM medium (PAN Biotech, Aidenbach, Germany), supplemented with 10% fetal bovine serum (FBS), 1% sodium pyruvate, 1% glutamine, and 1% penicillin-streptomycin-amphotericin [ 53 ].
Upon reaching 80% confluency, cells were detached with 1X Trypsin with calcium, magnesium, and phenol red (Biosigma, Cona, Italy) at 37 °C. For the cytotoxicity assay, suspended cells were then seeded at a density of 1.5 × 10 5 cells/cm 2 in 96-well plates, while for permeability experiments, cells were seeded at 5 × 10 5 cells/cm 2 in 24-well Transwell plates.
2.2.16. Cytotoxicity Studies
The in vitro cytotoxicity of DMF, DMF-loaded and unloaded NEs was assessed on RPMI 2650 cells to determine their effects and the highest non-cytotoxic concentration for subsequent permeability studies. Cytotoxicity evaluation was performed using the MTT assay (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide). Cells were seeded in 96-well plates at a density of 1.5 × 10 5 cells/cm 2 and cultured for approximately 6 days [ 54 ]. After incubation, the culture medium was replaced with 100 μL of the tested sample. Cells were exposed to the samples for 2, 6, 24, and 48 h. Following incubation, the samples were discarded, and 100 μL of MTT solution (0.5 mg/mL) was added to each well and incubated at 37 °C with 5% CO 2 for 3 h. After incubation, the MTT solution was removed, and 100 μL of dimethyl sulfoxide (DMSO) was added to dissolve the formazan crystals formed. Cytotoxicity was evaluated across six increasing concentrations of DMF (5, 10, 20, 30, 50, and 100 μM), both as a free compound (stock solution prepared at a concentration of 1670 µM in methanol:culture medium 30:70 ( v / v )) and formulated in NEs, all diluted in culture medium. Blank NEs were subjected to the same dilution series as DMF-loaded NEs. Untreated cells served as controls (100% metabolic activity). The optical density (OD) of the solutions was measured at 570 nm, with a reference wavelength of 670 nm, using a microplate reader (Synergy HT, Agilent Technologies, BioTek, Winooski, VT, USA) [ 55 ].
Cell viability (%) was calculated according to the following Equation (4): (4) C e l l v i a b i l i t y % = O D s a m p l e O D c o n t r o l × 100
All experiments were performed in triplicate and the results expressed as mean ± SD.
2.2.17. Permeability Studies
Permeability assessments were performed on both DMF alone and DMF-loaded NEs prepared by the spontaneous emulsification method (containing 10% DMF in the oil phase) at the concentration selected by the MMT assay (50 μM). DMF alone was dissolved in a mixture of methanol:culture medium (volume ratio 30:70) at a concentration of 1670 µM; the stock solution was then diluted in culture medium (DMF final concentration: 30 µM, methanol final concentration: 0.5%). To assess the relevance of NE particle size, permeability was evaluated under the same conditions also on the NE prepared by the high-energy method. RPMI 2650 cells were seeded at a density of 5 × 10 5 cells/cm 2 [ 52 ] in 24-well Corning Costar Transwell plates (Sigma Aldrich, Milan, Italy, Lot#1701) with polycarbonate membranes of 0.4 μm pore size. The cells were cultured at 37 °C with 5% CO 2 for 3 weeks, with the medium replaced every 2–3 days. During the first 8 days, cells were cultured under liquid–liquid conditions, with medium present in both the apical (donor) and basal (acceptor) compartments of the Transwells. Thereafter, the apical medium was removed, and cells were maintained at the air–liquid interface (ALI), as previously reported by Kreft et al. [ 53 ].
The transepithelial electrical resistance (TEER) was monitored weekly using a Millicell ERS-2 voltmeter (Millicell, Merck Millipore, Darmstadt, Germany). Permeability testing was conducted 21 days post-seeding, when the TEER value reached 57 ± 2 Ω·cm 2 [ 54 ], consistent with the expected range for human nasal mucosal cells (40–120 Ω·cm 2 ) [ 56 , 57 ].
For permeability evaluation, the culture medium from both compartments was removed, and wells were washed twice with PBS. Subsequently, 600 μL of phenol red-free, FBS-free αMEM medium was added to the basal compartment, while 200 μL of each test sample was introduced into the apical compartment. At designated time points (30, 60, 120, and 180 min), 300 μL of medium was withdrawn from the basal compartment for DMF quantification and immediately replaced with fresh medium.
The apparent permeability coefficient, Papp (cm s −1 ), measures drug permeability across the cell monolayer, and is calculated according to the following Equation (5): (5) P a p p = d Q d t A · C 0 where dQ (μg) is the amount of drug permeated from the apical to the basolateral side during the incubation time t (s), A is the surface area of the insert (cm 2 ), and C 0 is the initial concentration of DMF on the apical side (μg/mL) [ 54 , 58 , 59 ].
All collected samples were analyzed by HPLC, according to the method previously described, and the results were expressed as the mean percentage of permeated DMF ± SD for each experimental condition considered.
All cell culture procedures were performed under aseptic conditions using sterile culture materials, media, and reagents. The nanoemulsion formulations were not subjected to terminal sterilization before addition to the cultures.
2.2.18. Quantification of Monomethyl Fumarate, Dimethyl Fumarate and Geraniol in the Nanoemulsions Formulated for In Vivo Studies
Samples (100 μL) of the NEs formulated for in vivo studies with a DMF theoretical concentration of 2.5 mg/mL (NE2.5) were first diluted with 45 μL of acetate buffer 100 mM (pH 4) and with 855 μL of acetonitrile, the latter being added after vortexing. These samples were further diluted 1:10 using a mixture of acetonitrile:buffer acetate 95:5 ( v / v ), filtered through regenerated cellulose filters (0.45 μm) and injected (10 μL) into the same HPLC systems used for the quantification of monomethyl fumarate (MMF), DMF and GER, described in the section below.
2.2.19. Nasal Administration of Nanoemulsions to Rats
Adult male Sprague–Dawley rats (200–250 g; Charles River Laboratories, Calco, Italy) were used for the in vivo experiments. Animals were housed in groups of five per cage under controlled temperature and relative humidity conditions and a regular 12 h light/dark cycle, with food and water available ad libitum and environmental enrichment. Following arrival at the animal facility, animals were allowed to acclimatize for at least one week before the experimental procedures. Animals were monitored daily throughout the study, and a body condition score was recorded to assess their general health status. Humane endpoints included a body weight loss greater than 20% and signs of marked distress, including persistent isolation in a corner of the cage or an abnormal flattened posture. Only clinically healthy animals were included in the study.
A total of eight animals were included and allocated to two independent analytical cohorts ( n = 4 each): one for the pharmacokinetic analysis of DMF/MMF and the other for the pharmacokinetic analysis of GER. The individual animal represented the experimental unit. Given the exploratory proof-of-concept nature of the in vivo pharmacokinetic study, no formal a priori sample-size calculation was performed. The sample size was selected based on previous pharmacokinetic studies performed by our group using comparable experimental approaches, while minimizing animal use in accordance with the principle of reduction. No additional predefined exclusion criteria were applied, and no animals or experimental data were excluded from the analyses. No formal randomization procedure was used for allocation to the two analytical cohorts, since both cohorts received the same intranasal treatment and differed only in the analytes subsequently quantified. The investigator performing the HPLC analyses was blinded to the experimental allocation of the samples.
Rats were anesthetized with isoflurane (2–5%) and placed in the supine position. Each animal received 50 μL of NE2.5 into each nostril (total administration volume, 100 μL), corresponding to approximately 0.5 mg/kg DMF and 0.14 mg/kg GER. Intranasal administration was performed using a semiautomatic pipette connected to a short polyethylene tube inserted approximately 0.6 cm into the nostril. NE2.5 was freshly prepared and handled under controlled laboratory conditions prior to administration. The formulation was not subjected to terminal sterilization. Given the single-dose design and the short pharmacokinetic observation period (150 min), no terminal sterilization step was included in the experimental protocol. During the experimental procedures, animals were continuously observed for changes in respiratory pattern as indicators of adverse reactions. No adverse events related to intranasal administration or sampling procedures were observed, and none of the animals reached the predefined humane endpoints.
At predetermined time points, CSF (30 μL) and blood (100 μL) samples were collected. CSF collection was performed according to the cisternal puncture method described by van den Berg et al. [ 60 ], requiring a single needle insertion and allowing the collection of virtually blood-free serial CSF samples (30–40 μL) [ 61 ]. No more than 120 μL of CSF per rat (i.e., four 30 μL samples per rat) was collected within 150 min. Sampling time points ( n = 3–4 per time point) were selected taking into account the maximum of four CSF collections per animal, thereby allowing physiological restoration of CSF volume between sampling procedures.
Blood samples were serially collected from the tail vein using a butterfly needle to minimize animal stress and facilitate repeated sampling. CSF samples (10 μL) were immediately injected into the appropriate HPLC system (see below) for compound quantification.
For DMF and MMF quantification, blood samples from the DMF/MMF cohort ( n = 4) were immediately mixed with 250 μL of ice-cold acetonitrile, followed by the addition of 50 μL of 100 μM Fer (dissolved in water/acetonitrile, 25:75, v / v ) as the internal standard. Samples were centrifuged twice at 16,000× g for 5 min at room temperature. DMF and MMF were quantified by injecting 10 μL of the resulting supernatant into the appropriate HPLC system (see below).
For GER quantification, blood samples from the GER cohort ( n = 4) were immediately mixed with 250 μL of ice-cold acetonitrile, followed by the addition of 50 μL of 100 μM carbazole (dissolved in water/acetonitrile, 25:75, v / v ) as the internal standard. Samples were centrifuged twice at 16,000× g for 5 min at room temperature. GER was quantified by injecting 10 μL of the resulting supernatant into the appropriate HPLC system (see below).
The areas under the concentration–time curves (AUCs; μg·min·mL −1 ) for DMF, MMF, and GER in blood and CSF following intranasal administration of NE2.5 were calculated using the trapezoidal method.
Monomethyl Fumarate and Dimethyl Fumarate Quantification by HPLC Analysis
The HPLC system used for the MMF and DMF quantification in rat blood and CSF samples was a modular chromatographic system, consisting of a pump and a diode-array detector (DAD), models LC-40D and SPD-M40 (Shimadzu, Kyoto, Japan), respectively. This system was paired with an injection valve containing a 20 μL sample loop (model 7725; Rheodyne, IDEX, Torrance, CA, USA). Data acquisition and processing were performed with LabSolutions Software, version 5.110 (Shimadzu). A 5 μm Hypersil BDS C18 column (150 mm × 4.6 mm i.d.; ThermoFisher Scientific Srl, Milan, Italy), protected by a guard column filled with the same stationary phase, was used for separations that were performed at room temperature. A water:acetonitrile 80:20 ( v / v ) mixture acidified with 0.1% ( v / v ) TFA constituted the mobile phase, which was eluted under isocratic conditions at a flow rate of 0.8 mL/min. Chromatograms were recorded at 216 nm for the quantification of MMF and DMF, and at 320 nm for Fer, used as the internal standard in the blood extraction procedures. The retention times for MMF, Fer, and DMF were 3.8, 6.4, and 8.2 min, respectively.
The chromatographic precision for each compound was obtained by repeated analysis ( n = 6) of their 50 μM solutions (6.66 μg/mL MMF; 7.21 μg/mL DMF) in a mixture of water:acetonitrile 25:75 ( v / v ) and it was represented by relative standard deviation (RSD) values ranging from 0.89 to 0.96 for the tested compounds.
The calibration of DMF and MMF in CSF was performed by using a simulation fluid, constituted by a balanced Dulbecco’s phosphate-buffered saline (DPBS) solution depleted of calcium and magnesium and supplemented with 0.45 mg/mL BSA [ 62 , 63 ]. The concentration range of calibration in simulated CSF was 0.3–100 μM for both MMF and DMF (0.04–13.13 μg/mL for MMF and 0.04–14.41 μg/mL for DMF) and was found to be linear ( n = 6, r = 0.998, p < 0.0001). Preliminary injections of blank rat CSF and whole blood samples showed no detectable interfering peaks at the retention times of MMF, DMF and the internal standard. The analytical procedure was developed and assessed for its suitability for the present exploratory preclinical study rather than subjected to a complete regulatory bioanalytical validation. Importantly, CSF samples were analyzed immediately after collection without freezing or prolonged storage, thereby minimizing possible degradation of DMF before chromatographic analysis. To further assess the identity and chromatographic specificity of the DMF peak detected in CSF, the original HPLC-DAD data were retrospectively examined using LabSolutions software. The UV spectral profile of the chromatographic peak at the DMF retention time (8.2 min) in a representative CSF sample collected 30 min after NE2.5 administration was compared with that obtained from an authentic 100 μM DMF standard prepared in simulated CSF and analyzed under the same chromatographic conditions. In addition, spectral homogeneity across the CSF peak was evaluated by DAD peak-purity analysis using the peak-purity function of the acquisition software. The corresponding chromatograms, UV spectra, and peak-purity analysis are reported in the Supplementary Materials (Figures S1–S5) .
MMF recovery from whole blood samples was approximately 84%. This value was obtained by comparing the peak areas derived from 50 μM (6.66 μg/mL) MMF blood test samples ( n = 4) to those obtained by injecting an equivalent concentration of MMF dissolved in a water:acetonitrile 25:75 ( v / v ) mixture. MMF blood concentration was therefore calculated using the peak area ratio with respect to the internal standard (Fer). The MMF calibration in rat whole blood (4 °C) was obtained at six different concentrations ranging from 0.5 to 30 μM (0.07–3.99 μg/mL), producing a linear relationship ( n = 6, r = 0.995, p < 0.0001).
Geraniol Quantification by HPLC Analysis
The HPLC system used for the GER quantification in rat blood and CSF samples was a modular chromatographic system, consisting of a pump and a variable wavelength UV-Vis detector, models LC-10 AD VD and SPD-10A VP (Shimadzu, Kyoto, Japan), respectively. Also, this system was paired with an injection valve containing a 20 μL sample loop (model 7725; Rheodyne, IDEX, Torrance, CA, USA). Data acquisition and processing were performed with CLASS-VP Software, version 7.2.1 (Shimadzu). A 5 μm Hypersil BDS C18 column (150 mm × 4.6 mm i.d.; ThermoFisher Scientific Srl, Milan, Italy), protected by a guard column filled with the same stationary phase, was used for separations that were performed at room temperature. A water:acetonitrile 53:47 ( v / v ) mixture constituted the mobile phase, which was eluted under isocratic conditions at a flow rate of 1.0 mL/min. Chromatograms were recorded at 210 nm for the quantification of GER and carbazole, the latter being used as the internal standard in the blood extraction procedures. The retention times for GER and carbazole were 6.7 and 9.0 min, respectively.
The chromatographic precision for GER was obtained by repeated analysis ( n = 6) of its 20 μM solution (3.08 μg/mL) in a mixture of water:acetonitrile 25:75 ( v / v ) and it was represented by a relative standard deviation (RSD) value of 0.91.
The calibration of GER in CSF was performed by using the same simulation fluid described above for the calibration of DMF and MMF. The concentration range of GER calibration in simulated CSF was 0.5–20 μM (0.077–3.086 μg/mL) and was found to be linear ( n = 6, r = 0.999, p < 0.0001). Preliminary injections of blank rat CSF and whole blood samples showed no detectable interfering peaks at the retention times of GER and the internal standard. The analytical procedure was developed and assessed for its suitability for the present exploratory preclinical study rather than subjected to a complete regulatory bioanalytical validation.
GER recovery from whole blood samples was approximately 70%. This value was obtained by comparing the peak areas derived from 10 μM (1.543 μg/mL) GER blood test samples (n = 4) to those obtained by injecting an equivalent concentration of GER dissolved in a water:acetonitrile 25:75 ( v / v ) mixture. GER blood concentration was therefore calculated using the peak area ratio with respect to the internal standard (carbazole). The GER calibration in rat whole blood (4 °C) was obtained at six different concentrations ranging from 0.5 to 30 μM (0.077–4.629 μg/mL), producing a linear relationship ( n = 6, r = 0.997, p < 0.0001).
Selectivity of the analytical method was additionally supported by the markedly different chromatographic behavior expected for the low-molecular-weight analytes and the highly lipophilic [ 64 , 65 ] or polymeric components [ 66 ] of the nanoemulsion.
2.2.20. Statistical Analysis
Statistical analysis was performed using Stat Graphics Centurion 18.0 (Statgraphics Technologies, Inc., Rockville, MD, USA).
Results are expressed as mean ± SD. Comparisons between two groups were performed using Student’s t -test, whereas comparisons among more than two groups were carried out by one-way analysis of variance (ANOVA) followed by Fisher’s least significant difference (LSD) post hoc test. One-way ANOVA was applied to evaluate the effects of formulation variables on NEs size, polydispersity index, recovery, drug loading, release behavior, antioxidant activity, permeability, and apparent permeability coefficients (Papp). Student’s t -test was used for direct comparisons between two formulations when appropriate.
For the in vivo pharmacokinetic studies, concentrations of DMF, MMF and GER in blood and CSF are reported as mean ± SD. Areas under the concentration–time curves (AUC) were calculated using the trapezoidal rule with GraphPad Prism version 7.0 (GraphPad Software, San Diego, CA, USA).
Differences were considered statistically significant at p < 0.05.
3. Results and Discussion
3.1. Solubility of Dimethyl Fumarate in Geraniol
HPLC-based solubility studies demonstrated that DMF exhibits a higher solubility in GER than in water at room temperature (25 ± 1 °C), with 27.5 ± 0.97 mg of DMF dissolving in 1 mL of the oil phase. This value is approximately twenty times greater than its previously reported aqueous solubility under identical experimental conditions (about 1.4 mg/mL) [ 28 ], indicating an affinity of DMF for the lipophilic environment provided by GER. These results support the use of GER as a suitable solubilizing component for the development of DMF-loaded NEs.
3.2. Thermal Analysis
Thermogravimetric analysis (TGA) was performed to study the influence of the oil phase on DMF thermal stability.
In Figure 1 , the TGA curves and the respective first derivatives (DTG) of DMF, GER, and their physical mixtures (PMs) at different DMF:GER weight ratios (2:1, 1:1, and 1:2, respectively) were compared.
The TGA curve recorded on pure DMF showed a single mass loss at about 125 °C, associated with the sublimation of the anhydrous sample. The PM with a 2:1 (w/w) DMF:GER ratio exhibited two different mass losses, the first of which was 25% and occurred between 57 and 143 °C, and could be attributed to the sublimation of excess DMF that was not completely stabilized by the oil phase, followed by a second mass loss of about 70%, in the range between 143 and 312 °C, corresponding to the decomposition of the remaining PM.
In contrast, the PM at the ratio 1:1 (w/w) exhibited only a mass loss of 90% starting at around 158 °C, and therefore higher than the sublimation temperature of pure DMF. This shift is consistent with a stabilizing effect exerted by the oil phase on DMF.
Also, the TGA of the PM at the ratio 1:2 (w/w) showed a shift in the DMF decomposition onset temperature to a higher value (T onset = 200 °C), supporting the hypothesis that the oil phase exerts a stabilizing effect, directly proportional to its concentration in the formulation.
As already highlighted in our previous study on NEs with DMF and carvacrol [ 38 ], in all three samples the amount of DMF exceeds the solubility determined at 25 °C, although an increase in solubility may occur with the temperature increase during the experiment. The results therefore support the conclusion that DMF exhibits enhanced thermal stability at elevated temperatures when solubilized in the oil phase as compared to the pure drug.
DSC analysis was performed on DMF and on the same PMs. DMF showed a typical thermal profile of an anhydrous crystalline compound, with a sharp endothermic peak at 103.2 ± 0.7 °C, corresponding to its melting point (T onset = 100.5 ± 0.5 °C; ∆H fus = 208 ± 2 J/g). All DSC curves of the PMs displayed similar thermal profiles with a broadening and a shift to lower temperatures of the DMF melting peak as the amount of GER increased, consistent with its behavior as an ‘impurity’ within the system. Nevertheless, the normalized enthalpy of DMF remained nearly constant across the mixtures (∆H fus = 200 ± 4 J/g) and comparable to that of the pure DMF, confirming the homogeneity of the mixtures ( Figure 2 ).
3.3. Preparation of the Nanoemulsions
NEs can be prepared using either high-energy or low-energy emulsification methods, each offering distinct advantages [ 67 ]. High-energy methods, including ultrasonication and homogenization, require mechanical forces to break down larger droplets and achieve smaller particle sizes [ 33 ]. Ultrasonication uses high-frequency sound waves (≥20 kHz) to create cavitation, forming and collapsing microdroplets that generate nano-sized emulsions [ 68 ]. This method effectively reduces droplet size while preserving NE integrity [ 69 ]. Similarly, high-pressure homogenization applies intense mechanical forces to break down droplets, aiding in homogenization, emulsification, and suspension production [ 70 ]. While effective, homogenization faces challenges in automation and scalability. By optimizing these methods, stable and efficient NEs can be developed for drug delivery applications, including the intranasal administration of therapeutics targeting the CNS.
In contrast, low-energy methods, such as phase inversion and spontaneous emulsification, leverage the physicochemical properties of components to form emulsions without external energy input [ 71 ]. Considering the very poor stability of DMF [ 13 , 27 ], we preferred these methods over the high-energy ones. Phase inversion emulsification can be induced by either composition changes (PIC) or temperature shifts (PIT), leading to applications in the pharmaceutical, cosmetic, and food industries [ 72 , 73 ]. Spontaneous emulsification relies on surfactant diffusion between phases and has been used to develop delivery systems for bioactive compounds, achieving particle sizes below 200 nm [ 74 , 75 ]. While these NEs can be stable at room temperature, co-surfactants may enhance their thermal stability [ 76 ]. Given its simplicity and cost-effectiveness, this method holds promise for pharmaceutical and biomedical applications, including nasal drug delivery [ 77 ].
The NEs were prepared using CSOA as a surfactant, taking into account that chitosan-coated formulations further enhance drug stability, controlled release, and bioavailability across various routes. It is well known that chitosan-coated nanoparticles exhibit high encapsulation efficiency and positive zeta potential, which allow them to improve their interaction with biological membranes [ 78 ], making them promising for drug delivery applications [ 79 ].
3.3.1. Phase Inversion Emulsification Method
The first method used for the preparation of NEs was the phase inversion method combined with an Ultra-Turrax homogenizer (PI). The droplet size of the internal phase was evaluated through a unimodal analysis combined with a numerical size distribution, both obtained from PCS.
In Figure 3 , the results are given as mean droplet size and PDI. The effect of two different chitosan (CS) and oleic acid (OA) molar ratios (1:0.5 and 1:1) on particle size was first assessed by PCS. As illustrated in Figure 3 A, in the case of size distribution in intensity, the mean diameter of the internal phase droplets of the NE prepared with CSOA in a 1:1 ratio was slightly smaller (736 ± 15 nm) compared to that of the formulation with a CSOA 1:0.5 ratio (805 ± 7 nm). On the other hand, the PDI values were significantly lower in the formulation with a CSOA 1:1 ratio, indicating a narrower size distribution and improved homogeneity (Student’s t -test, p < 0.05).
The discrepancy observed between the mean diameter values obtained by intensity and by number is therefore probably attributable to the higher polydispersity of the sample with CSOA 1:0.5 compared to the 1:1 formulation. Based on these results, the CSOA molar ratio of 1:1 was selected for further study, as it resulted in a more uniform and homogeneous NE, in terms of droplet size distribution.
Subsequent studies instead focused on comparing NE formulations with two different concentrations of 1:1 CSOA, namely 2% and 3% (w/v) ( Figure 3 B). Increasing the CSOA concentration to 3% led to a significant increase (nearly twofold) in the mean diameter of the internal phase droplets, with a particle size of 1111 ± 14 nm compared to 764 ± 15 nm for the 2% CSOA formulation, in terms of intensity (unimodal analysis). Similarly, the size distribution by number revealed larger droplets for the 3% CSOA formulation (403 ± 261 nm) than those in the 2% formulation (311 ± 27 nm), although the high variability observed in the 3% samples made this difference statistically insignificant. This behavior may be attributed to the aggregation of the excess surfactant molecules in the continuous phase, which reduces the effective amount of surfactant available to stabilize the oil–water interface, leading to droplet coalescence and size increase [ 80 ]. In addition, the NE prepared with the 2% CSOA displayed the lowest PDI, thereby justifying the selection of this concentration for further formulation.
Figure 3 C, by contrast, showed the effect of reducing the oil phase concentration (GER) from 10% (500 mg) to 5% (250 mg). A significant decrease in the NE mean droplet size was observed at the lower concentration of oil, as evidenced by both unimodal and numerical size analyses (Student’s t -test, p < 0.05). Although the PDI of the 5% GER formulation was slightly higher, it remained within the acceptable range, with an average value of 0.22. Based on these results, the 5% GER concentration was selected for further studies. These results agree with the literature [ 81 , 82 ] indicating that an increase in oil content in NEs increases the viscosity of the oil phase, thereby reducing its ability to disperse effectively within the aqueous phase.
The NE, based on a CSOA 1:1 ratio with CSOA at 2% and GER at 5%, was therefore loaded with DMF at 2% relative to the oil phase. The influence of this loading on the droplet size of the dispersed phase was then evaluated. Accordingly, dimensional analyses were performed on both blank and 2% DMF-loaded NEs ( Figure 3 D). The results showed a slight increase in mean droplet size when the oil phase of the formulation was loaded with DMF. The PDI values remained comparable between the two formulations, indicating preserved droplet uniformity. Nonetheless, the differences in particle size and PDI were statistically significant (ANOVA, Fisher’s post hoc test, p < 0.05), indicating that DMF loading influences the dimensional characteristics of the NE system.
The formulation identified as optimal using the PI method was subsequently subjected to a stability study to evaluate its physicochemical robustness over time.
Dimensional Stability Studies
Dimensional stability analyses were performed on both loaded and unloaded NEs, stored for six weeks at 4 ± 1 °C. As shown in Figure 4 , the DMF-free NEs (i.e., blank NE) exhibited an increase in droplet size during the first week of storage (<200 nm), after which the mean diameter remained substantially unchanged throughout the remainder of the study. In contrast, the DMF-loaded NEs showed a marked increase in droplet size, from approximately 450 nm at preparation to 943 nm after four weeks of storage.
3.3.2. Spontaneous Emulsification Method
To improve the stability of the formulation, a second preparation method was investigated, specifically spontaneous emulsification (SE). The NEs obtained with this method had a translucent appearance and a bluish color, according to the Tyndall effect characteristic of nano-dispersed systems. As shown in Figure 5 , the droplet size obtained for the blank NE prepared by SE was smaller than that obtained using the PI method; while the PDI remained around 0.3, indicating good size homogeneity.
A statistically significant difference was observed in the intensity and PDI values between NEs prepared by the two methods (Student’s t -test, p < 0.05). Based on these results, SE-prepared formulations were selected for further investigations.
Considering the very poor stability of DMF [ 13 , 27 , 28 ], some advantages can be envisaged for this method over the high-energy one.
Figure 6 shows a comparison between unloaded and DMF-loaded NEs prepared via SE, at 2% (v/v) of DMF in the oil phase and at a higher concentration of up to 10% (v/v). Both loaded NEs exhibited a slight increase in droplet size (observed in both unimodal and numerical analyses) compared to the corresponding blank NEs. The lower PDI of the DMF-loaded NEs indicated a more monodisperse and homogeneous droplet population. Statistical analysis evidenced that the differences in unimodal analysis between unloaded and loaded formulations, although quite small, were statistically significant, while no statistical differences were found in the case of number analysis (one-way ANOVA followed by Fisher’s LSD post hoc test, p < 0.05).
Considering the good dimensional results, the subsequent characterization was performed on the NE loaded with 10% DMF. This formulation corresponded to a theoretical DMF concentration of 0.24 mg/mL and a GER concentration of 2.4 mg/mL. The quantification of DMF and GER in this formulation after preparation was verified by HPLC and resulted in a recovery of about 74 ± 8% and 82 ± 3% for DMF and GER, respectively.
3.4. Fourier Transform Infrared Analysis
The interactions occurring between the substances within the NEs were further characterized by FT-IR. First, to identify possible interactions between the two compounds, the FT-IR spectra of pure GER, pure DMF, and their PMs containing 10% (w/w) DMF in the oil phase were compared ( Figure 7 ).
According to the literature, the FT-IR spectrum of GER shows a broad band at 3323 cm −1 due to the stretching vibration of the terminal -OH group. Additional bands at 2913 and 1376 cm −1 are attributed to -CH 2 stretching vibrations, while the band at 1668 cm −1 corresponds to the conjugated double bond C=C. The band at 996 cm −1 is associated with the out-of-plane bending vibration of the -CH group [ 83 , 84 ].
In the DMF spectrum, characteristic bands are observed for the carbonyl C=O stretch at 1714 cm −1 , the C=C stretch at 1634 cm −1 , and C-O ester stretching vibrations at 1154 and 1303 cm −1 . Additional bands at 773 and 671 cm −1 correspond to -CH bending vibrations.
Some of these bands shift to higher wavenumbers in the spectrum of the PMs, suggesting the occurrence of molecular interactions between the two components. Specifically, the carbonyl stretching band of DMF at 1714 cm −1 shifts to 1731 cm −1 in formulations containing 10% DMF. Similarly, the bands at 1303 and 1154 cm −1 shift to 1305 and 1157 cm −1 , respectively. These spectral changes indicate alterations in the chemical environment of the two components, consistent with interactions between DMF and GER.
FT-IR spectra of the blank NEs and NEs containing 10% DMF, are presented in Figure 8 . In both spectra, the predominant absorption region appeared between 3600 and 3200 cm −1 , with a peak at around 3330 cm −1 , corresponding to the stretching vibration of the water O-H groups and reflecting the high hydration level of the O/W NE. This was followed by a band at 1636 cm −1 attributable to the bending vibration of O-H groups, indicating inter- and intramolecular hydrogen bonds occurring during NE formation.
In the DMF-loaded NE spectrum, an additional band at 1737 cm −1 , corresponding to the characteristic carbonyl C=O stretching vibration, was observed, confirming the incorporation of the active ingredient into the formulation.
3.5. Dimensional Stability Studies
The dimensional stability of the DMF-loaded NE was evaluated over three months of storage under refrigerated conditions (4 ± 1 °C). During this period, the formulations were monitored for changes in droplet size distribution. The particle size differences among NEs (blank vs. DMF-loaded) prepared by the SE method were statistically significant ( p < 0.05, one-way ANOVA).
In general, formulations characterized by smaller droplet sizes exhibited greater stability during storage, which is in agreement with previous studies [ 81 ]. Moreover, all formulations remained stable during the first four weeks, and droplet sizes did not exceed 400 nm.
Among the three formulations, the NE containing 0.24 mg/mL DMF (corresponding to 10% DMF in the oil phase) remained the most stable during the first four weeks, exhibiting a smaller droplet size (<350 nm) compared to the other formulations as reported in Figure 9 .
TEM analysis was performed to characterize the unloaded and loaded NEs in terms of size and their morphology. The NE images obtained confirmed the droplet size (~170 nm) and the spherical shape of the internal phase droplets for both blank and loaded NEs ( Figure 10 ).
Specifically, the blank NEs appeared as spherical globules with definite dark boundaries and a lighter inner region (A), whereas the DMF-loaded NEs exhibited spherical droplets with a dark inner core. Overall, the droplet size observed by TEM for both types of NEs was found to be comparable to that obtained by PCS analysis, confirming the validity of the size distribution data.
3.6. Zeta Potential and pH of Nanoemulsion
Zeta potential measurements were carried out to assess the long-term stability of NE formulations and to understand the state of the surface morphology. The blank NE exhibited a zeta potential of 20 ± 8 mV, while the DMF-loaded NEs exhibited a comparable value of 22 ± 10 mV. The positive zeta potential is consistent with the presence of CSOA on the particle surface. This is indicative of a good physical stability of the dispersed system by preventing droplet aggregation through electrostatic repulsion. Moreover, the positive surface charge is relevant for interaction with mucin and cell membranes, potentially enhancing mucoadhesion and improving bioavailability, as indicated by the literature [ 85 , 86 ].
Finally, the pH of this NE formulation was assessed, showing a value of 5.5 ± 0.5, which is compatible with the physiological pH range of the nasal mucosa (4.5–6.5) [ 87 ].
3.7. In Vitro Release Studies
The in vitro dialyzing membrane model was used to evaluate the release kinetics of DMF and GER from the NEs. Analysis of the release profile obtained revealed a rapid initial release phase within the first 120 min, during which 85 ± 4% of the total encapsulated DMF was released in good accordance with the previous results obtained with a NE formulation combining DMF and carvacrol [ 88 ]. The cumulative release reached 87 ± 4% at 210 min. When compared with free DMF, which under the same experimental conditions showed a cumulative release of 85 ± 2% after 60 min in a previous study, the NE formulation exhibited a slightly slower release profile [ 88 ]. However, the release remains compatible with the nasal application, which requires relatively fast release to optimize absorption during the short contact time of the formulation on the nasal mucosa.
Analogous considerations can be made for GER, which showed a maximum cumulative release of 84 ± 5% after 210 min.
Overall, these results confirm that the release behavior is suitable for nasal administration, considering that the nasal mucociliary clearance occurs in about 10–20 min, but this time can be prolonged due to the chitosan mucoadhesion effect. Nose-to-brain absorption is based to a relevant extent on the possible extraneuronal pathway to the CNS, by which the drugs can be transported along olfactory neurons in a short time (<30 min). This pathway should be active also for the cargo released in the nasal olfactory epithelium without the nanocarrier endocytosis. However, in the case of endocytosis, the transport of the carrier would be much slower, taking several hours or days [ 89 , 90 ]. In the present case, considering that the release is completed in a few hours, it is likely that the passage to the CNS involves the NE cargoes, after their release.
3.8. ROS-Scavenging Activity
The results of the antioxidant activity of DMF-loaded NE (NE DMF), blank NE (NE GER) and CSOA are shown in Figure 11 . In general, the samples showed a statistically significant dose-dependent ROS-scavenging activity (one-way ANOVA, Fisher’s least significant difference (LSD) procedure, p < 0.05). For all the samples, statistically significant increase was found between the 1:10 and 1:5 dilutions ( p < 0.05). Notably, the CSOA and DMF-loaded NE reached their maximum activity already at a 1:5 dilution, as the difference between the 1:5 and 1:3 dilutions is not statistically significant for these formulations. For the blank NE GER, a comparable effect with NE DMF was found only with the highest concentration (1:3 dilution). This difference is probably due to the combined presence of DMF and GER in the loaded NEs. The essential oil is an acyclic isoprenoid monoterpene, which, as demonstrated in the literature, already exhibits intrinsic antioxidant properties, possibly due to its hydroxyl group responsible for ROS deactivation [ 91 , 92 ]. Furthermore, DMF itself is known to exert a Nrf2-mediated antioxidant effect, especially in neuroinflammation [ 93 ].
3.9. Cytotoxicity Studies
Cellular metabolic activity was evaluated using the MTT assay on RPMI 2650 cell line treated with DMF-loaded NE (obtained with the SE method), blank NE, and free drug at different incubation times (2, 6, 24, and 48 h) and DMF concentrations (5, 10, 20, 30, 50, 100 μM). The results are shown in Figure 12 . Overall, the data demonstrated that the cytotoxicity profile was concentration- and time-dependent. Good short-term cytocompatibility was observed at concentrations up to 50 μM, whereas a marked loss of cell viability occurred at the highest concentration tested (100 μM), particularly after 24 h of exposure.
Although a general dose-dependent decrease in cellular metabolic activity was observed, RPMI 2650 cells exhibited higher metabolic activity after 48 h of incubation with both blank and DMF-loaded NEs compared with the corresponding values measured after 24 h. This observation suggests that prolonged exposure to the formulations did not impair cellular metabolic activity and may even promote metabolic activity in RPMI 2650 cells under the tested conditions.
The lowest viability value (10 ± 3%) was observed after 24 h for the formulation with the highest DMF concentration (100 μM).
The results demonstrated high cytocompatibility of the loaded NE, with cell viability values above 87 ± 3% up to a 50 μM after 6 h of incubation. Based on these results, the samples containing 50 μM DMF were selected for the following in vitro permeability assay.
3.10. Permeability Studies
To evaluate the ability of NEs prepared with the two different methods to permeate the nasal epithelium, permeability studies were conducted using RPMI 2650 cells cultured at the air–liquid interface, a model that is, according to the literature, widely recognized as an in vitro analog of the human nasal epithelial barrier [ 56 ].
The DMF solubilized in a methanol:culture medium mixture (30:70 v / v ) and then diluted in culture medium to obtain a 50 μM concentration was tested as a control. The permeability data are shown in Figure 13 A. After 60 min, free DMF exhibited the highest permeation, reaching nearly 70%, while the NE produced by the SE method (NE SE) and that produced by PI (NE PI) reached 59% and 35%, respectively. This trend was maintained up to 120 and 180 min, with free DMF permeability around 70%, NE SE remaining stable at 59%, and NE PI progressively increasing to 45%.
Free DMF exhibited the greatest variability (higher values of SD), particularly at 60 and 120 min, probably due to its low intrinsic stability, tendency to sublime, and poor solubility.
Free DMF showed the highest apparent permeability under the experimental conditions, indicating that incorporation into the NE did not enhance DMF permeation relative to the unformulated drug. However, among the two NE formulations, NE-SE showed consistently greater permeability than NE-PI. Since the two formulations had the same qualitative composition but differed in their preparation method and resulting droplet characteristics, the higher permeability of NE-SE may be related, at least in part, to its smaller droplet size and narrower size distribution.
The calculated Papp values quantified these results and showed that free DMF solubilized in methanol demonstrated the highest permeability compared to the other two NE formulations ( Figure 13 B). Generally, a compound solubilized in an organic solvent, like methanol in this case, displays higher apparent permeability, due to the solvent’s ability to enhance solubility and decrease mass transfer resistance across the cell monolayer [ 94 ].
However, the difference between the percentage permeated for free DMF solubilized in methanol and DMF delivered by NE SE at 30 and 60 min was not statistically significant. It is important to underline that DMF’s poor solubility limits its bioavailability if not properly formulated.
Moreover, the results clearly highlighted that the droplet size of the dispersed phase plays a critical role in the delivery of the drug encapsulated in the GER phase. The smaller NE SE dimensions are a factor recognized as relevant for the improvement of drug absorption by the nose-to-brain route [ 86 ]. It can be envisaged that smaller dimensions correspond to a higher specific surface area of the droplets, larger contact area of the NE with the epithelium, and easier partition from the NE to the epithelium of both the oil and the drug loaded in the NE, resulting in faster absorption. After 3 h, however, the absorption percentage for the two NE samples became comparable.
A similar positive effect of smaller oil droplet dimensions was observed for GER absorption, as shown in Figure 14 A. After one hour of incubation, GER from NE SE reached a maximum permeability of 65%, increasing to 79% and 87% at 120 and 180 min, respectively. In contrast, GER from NE PI reached only 34%, as a consequence of the larger droplet size of this formulation. Papp values calculated for GER at 60 min, and reported in Figure 14 B, confirmed again the more marked difference between samples with different droplet sizes.
3.11. Formulation for In Vivo Study
As a proof of concept, a preliminary in vivo evaluation was performed in a rat model to evaluate the impact of the delivery system on DMF bioavailability in the CNS when administered by the nasal route together with GER.
To enable the in vivo evaluation, additional formulation studies were performed to increase the final DMF concentration in the NE based on the results obtained in the first part of the work. The target DMF concentrations were selected taking into account previous pharmacokinetic studies on intranasal DMF administration in rats, in which a dose of 1 mg/kg was employed [ 27 ]. The theoretical GER concentration was maintained constant at 5% ( v / v ), whereas the target DMF concentration was progressively increased from the formulation previously investigated (0.24 mg/mL; NE0.24) to 1.0 mg/mL (NE1) and 2.5 mg/mL (NE2.5).
Following preparation, DMF and GER recovery (Rec%), drug loading (DL%), and final concentrations were determined by HPLC. The results are reported in Figure 15 . Data obtained for the formulation containing the lowest DMF concentration (NE0.24) are also included for comparison.
The results indicate that increasing the target DMF concentration in the formulation led to higher DMF recovery and drug loading, while a progressive reduction in GER recovery was observed. Since the CSOA concentration was kept constant in all formulations, these findings suggest that increasing DMF incorporation may affect the retention of GER within the NE structure. This behavior could be related to a limited capacity of the colloidal system to simultaneously accommodate increasing amounts of both compounds. Further studies investigating higher CSOA concentrations may help clarify whether the loading capacity of the system can be further improved. Despite the reduction in GER content, this compound is known to exhibit efficient transport following intranasal administration when formulated in CSOA-stabilized emulsions, likely due to the combined stabilizing and permeation-enhancing properties of CSOA [ 44 ].
Neuroprotective effects of GER have been reported in mouse models following oral administration of relatively high doses (50–200 mg/kg) [ 41 , 95 ]. However, pharmacokinetic studies performed in mice after oral administration of 200 mg/kg GER showed low systemic exposure, with a blood C max of approximately 0.05 µg/mL and only trace amounts detected in brain tissue [ 96 ]. In contrast, oral administration of 50 mg/kg GER to rats, formulated as a coarse glycerol emulsion, resulted in markedly higher bioavailability (92%), yielding peak concentrations of approximately 280 µg/mL in blood and 2.5 µg/mL in CSF [ 49 ]. These findings highlight the strong influence of species and formulation on GER pharmacokinetics. Considering the potential advantages of the intranasal route for CNS delivery, it was considered worthwhile to investigate whether lower amounts of GER, such as those contained in the NE2.5 formulation, could still provide measurable CSF exposure following nasal administration. Based on these considerations, NE2.5 was selected for the proof-of-concept in vivo study because it provided the highest experimentally achieved DMF content while retaining GER and maintaining a nanoscale droplet size and low polydispersity. A complete biopharmaceutical and safety characterization comparable to that performed for the lower-loaded formulation was not carried out for NE2.5 and should therefore be considered a limitation of the present study, although the concentration of the stabilizing amphiphilic chitosan was the same as in the previously tested formulations, and the dimensions of the NE were comparable with those previously obtained with the selected preparation method.
Because the permeation test highlighted the importance of the NE dimensions for this aspect, before administration the NE2.5 formulation was characterized in terms of particle size (182 ± 2 nm, PDI 0.021 ± 0.027) and active contents: 1.08 ± 0.02 mg/mL DMF, 0.03 ± 0.01 mg/mL MMF and 0.28 ± 0.01 mg/mL GER. MMF appears to be derived from the degradation of DMF with an amount corresponding to ~3% of the DMF detected in the NE2.5 formulation. A similar degradation amount was recently evidenced in DMF-loaded nanoformulations designed for nasal administration [ 13 ].
Nasal Administration of the Formulation NE2.5
GER and DMF were administered nasally to rats with doses of 0.14 mg/kg (~0.035 mg/rat) and 0.5 mg/kg (~0.13 mg/rat), respectively, by using the NE2.5 formulation (50 μL inserted in each nostril of the rats). It is well known that GER can induce dermal and respiratory side effects, including damage to the nasal mucosa, but these effects do not occur if the amount of GER administered into the nose of rats does not exceed 20 μL [ 39 , 42 ]. The GER amount contained in the NE2.5 formulation administered to rats is ~0.035 mg, corresponding to ~0.04 μL, an amount ~500 times lower than the toxicity threshold in the airways of rats.
Following intranasal administration of NE2.5, GER was detected in both the bloodstream and CSF, whereas DMF was detected in CSF but not in blood, where MMF was instead measurable ( Figure 16 ). In our previous study, intranasal administration of free DMF at 1 mg/kg did not result in detectable DMF or MMF in CSF, whereas nanoparticle-associated DMF reached the CSF [ 13 ]. Moreover, intravenous administration of free DMF at 2 mg/kg did not result in detectable DMF in CSF [ 13 ], supporting the view that the CSF exposure observed after NE2.5 administration is unlikely to be explained simply by systemic redistribution of free DMF. Although cross-study comparisons should be interpreted cautiously, these findings support a contribution of the nanostructured formulation to DMF CSF exposure. The present study, however, does not allow the individual contributions of GER, CSOA, and the nanoemulsion architecture to be distinguished.
Figure 16 A reports, in particular, the GER and MMF profiles detected in the bloodstream, showing for GER a C max value of 1.69 ± 0.13 μg/mL at 60 min (T max ) which decreased to zero within 150 min after the nasal administration; the AUC value of this profile was 126.5 ± 8.5 μg min⋅mL −1 . The C max value of MMF, detected at the same T max of GER (60 min) was 0.99 ± 0.05 μg/mL and decreased to zero within 150 min; the AUC value of this profile was 62.6. ± 5.3 μg⋅min⋅mL −1 .
A comparison between the GER profiles in the bloodstream obtained after the nasal administration of NE2.5 and an oral administration previously performed in rats with a dose of 12.5 mg (50 mg/kg) showed that this compound after oral administration was completely eliminated within 1 h [ 49 ], whereas a more prolonged GER exposure profile was observed after intranasal NE2.5 administration. This behavior seems in good agreement with the in vitro release results described above (see Section 3.7 ), which evidenced a maximum cumulative release of GER from NE within a few hours.
The absence of DMF in the bloodstream of rats after nasal administration of NE2.5, is not surprising considering that we have recently demonstrated that DMF is instantly hydrolyzed to MMF in the blood of rats and that, after intravenous administration of DMF to rats, only MMF can be detected in the bloodstream of rats at amounts corresponding to ~5% of the total drug administered. In particular, after the intravenous administration of 2.0 mg/kg DMF to rats, the MMF profile in the bloodstream was characterized by an AUC value of 74.7 ± 3.4 μg⋅min⋅mL −1 [ 13 ]. Interestingly, nasal administration of NE2.5 (0.5 mg/kg DMF) to rats was associated with an AUC value of MMF in the bloodstream of 62.6 ± 5.3 μg⋅min⋅mL −1 ( Figure 16 A). For contextual comparison, in our previous study intravenous administration of DMF (2 mg/kg) resulted in an MMF blood AUC of 74.7 ± 3.4 μg·min·mL −1 [ 13 ], whereas in the present study intranasal administration of NE2.5 containing 0.5 mg/kg DMF resulted in an MMF AUC of 62.6 ± 5.3 μg·min·mL −1 . These values should not be interpreted as a formal bioavailability comparison, since MMF is a metabolite of DMF and its systemic exposure depends on route- and input-dependent absorption, hydrolysis, formation, and disposition processes. The different MMF exposure profiles may partly reflect the different input kinetics of DMF following intranasal administration of the NE and intravenous infusion, together with the rapid hydrolysis of DMF in blood.
It is important to remark that after the oral administration of DMF at therapeutic doses to humans (240 mg), this drug was absent in the bloodstream, unlike its metabolite MMF, whose C max appeared at near 1 μg/mL [ 10 ], a value very similar to that obtained in the bloodstream of rats after the nasal administration of the NE2.5 formulation ( Figure 16 A).
Following nasal administration of NE2.5, DMF, MMF, and GER were detected in rat CSF, as reported in Figure 16 B. In particular, the DMF concentration at 20 min was 5.3 ± 0.4 μg/mL and increased at 30 min to values higher than 7 μg /mL, which were maintained within 90 min, and then the DMF concentration decreased to zero within 150 min. Moreover, relatively small amounts of MMF were also detected in the CSF of rats, with concentration values ranging from ~0.2 to ~0.3 μg/mL between 20 and 120 min after the nasal administration and then decreasing to zero within 150 min ( Figure 16 B). The AUC values obtained from the profiles of these compounds were 823 ± 28 μg·min·mL −1 and 30 ± 2 μg·min·mL −1 for DMF and MMF, respectively, suggesting that the small amounts of MMF detected in the CSF of rats (~3% of DMF) can be attributed to the hydrolysis of DMF in the NE2.5 formulation, in accordance with a similar behavior previously evidenced for other DMF nasal formulations [ 13 ].
A potential analytical limitation associated with DMF is its susceptibility to hydrolysis to MMF in biological media. In the present study, however, freshly collected CSF samples were immediately injected into the HPLC system and were not subjected to storage or freeze–thaw cycles. Therefore, degradation during prolonged sample handling was minimized. Moreover, any hydrolysis occurring between CSF collection and chromatographic injection would be expected to decrease the measured DMF concentration while increasing MMF, rather than generating an artifactual DMF signal. This interpretation is further supported by the chromatographic separation of MMF and DMF, which eluted at 3.8 and 8.2 min, respectively.
In addition, retrospective analysis of the original HPLC-DAD data was performed to further assess the identity and chromatographic specificity of the DMF peak detected in CSF. The UV spectral profile of the peak eluting at the DMF retention time (8.2 min) in a representative CSF sample collected 30 min after NE2.5 administration closely matched that obtained from an authentic 100 μM DMF standard analyzed under the same chromatographic conditions. DAD peak-purity analysis further showed spectral homogeneity across the CSF peak, with a peak-purity index of 1.000000, exceeding the software-defined threshold of 0.999923, and no evidence of detectable co-eluting impurities under the applied chromatographic conditions ( Figures S1–S5 ). These findings provide additional support for the assignment of the CSF peak to DMF.
Notably, DMF was readily detected in the CSF despite being undetectable in the bloodstream, where only its hydrolysis product MMF was observed. This marked dissociation between systemic and CSF profiles is consistent with the possibility that at least part of the administered DMF reaches the CSF before undergoing the rapid hydrolysis occurring in blood. This interpretation is further supported by our previous findings showing that intranasal administration of free DMF at 1 mg/kg did not result in detectable DMF or MMF in CSF, whereas DMF administered in a nanostructured formulation reached the CSF; moreover, intravenously administered free DMF was not detectable in CSF [ 13 ]. Nevertheless, CSF measurements alone cannot distinguish unequivocally between direct nose-to-brain transport and indirect routes involving systemic absorption and subsequent CNS distribution. Therefore, the present data should be interpreted as evidence of substantial and sustained DMF exposure in CSF following intranasal NE2.5 administration, with a pharmacokinetic pattern consistent with a contribution of nose-to-brain transport, rather than as a definitive demonstration of direct nose-to-brain delivery.
Figure 16 C reports a comparison between two DMF profiles obtained in the CSF of rats after the nasal administration of two different formulations: the green profile is related to the NE2.5 formulation, whereas the blue profile was previously obtained by us using a nasal formulation based on lipid–polymer hybrid nanoparticles (LPNs). The DMF dose in NE2.5 was 0.5 mg/kg and produced an AUC value of 823 ± 28 μg·min·mL −1 , very similar to the AUC value obtained by LPNs (844 ± 34 μg·min·mL −1 ), whose DMF dose was 1 mg/kg [ 13 ]. Interestingly, despite the lower DMF dose administered with NE2.5 (0.5 mg/kg) compared with the previously investigated LPN formulation (1 mg/kg), substantial and prolonged DMF concentrations were detected in CSF. In our previous study, intranasal administration of free DMF at 1 mg/kg did not result in detectable DMF or MMF in CSF, whereas DMF administered in the previously investigated LPN formulation reached the CSF. Although comparisons between separate studies should be interpreted cautiously, these previous findings support the hypothesis that formulation of DMF in nanostructured nasal delivery systems contributes to its CSF exposure.
A DMF-tocopherol acetate nanolipidic carrier has been previously shown by other authors to improve the bioavailability of MMF in both the bloodstream and brain after oral administration to rats, in comparison to the free DMF [ 97 ]. NE2.5 differs from this type of carrier, being associated with a prolonged exposure of DMF itself in the CSF, together with systemic MMF concentrations in the range reported as therapeutic in humans [ 10 ]. Considering that DMF possesses stronger immunomodulatory properties and higher membrane permeability than MMF [ 16 , 18 ], sustained DMF exposure in the CSF may warrant further investigation in the context of MS treatment.”
In addition to DMF and MMF, the nasal administration of the NE2.5 formulation also allowed the quantification of GER in the CSF of rats, with concentrations ranging from ~1.0 to ~1.6 μg/mL between 20 and 120 min, as reported in Figure 16 B. Interestingly, these values have the same order of magnitude as GER concentrations detected in the CSF of rats after receiving an oral dose of 50 mg/kg [ 49 ], which is normally used to study the in vivo neuroprotective effects of this compound [ 41 , 96 ]. In particular, the C max value of GER in the CSF of rats, related to this oral dose, was ~2.5 μg/mL, but its concentration rapidly decreased to zero within 1 h, whereas GER remained detectable in CSF for a longer period following intranasal NE2.5 administration.
3.12. Limitations of the Study
Several limitations of the present proof-of-concept study should be acknowledged. First, the absence of concurrent free-DMF, DMF-GER, blank-NE, and systemic-route comparator groups prevents attribution of the observed CSF exposure specifically to the NE and does not allow formal quantification of the relative contribution of direct nose-to-brain transport. Relevant intravenous and intranasal free-DMF data obtained under comparable experimental conditions were available from our previous study and provide useful contextual support, but they cannot replace concurrent randomized comparator groups. Second, drug concentrations were determined in CSF but not in individual brain regions; therefore, CSF exposure should not be considered equivalent to direct evidence of drug distribution within brain parenchyma. Third, although NE2.5 was characterized immediately before administration with respect to droplet size, polydispersity, and actual DMF, MMF, and GER content, it was not subjected to the complete set of in vitro biopharmaceutical and local safety studies performed during formulation optimization. In particular, direct mucin-binding, mucosal-retention, ciliotoxicity, and nasal-histopathology assessments were not performed. Fourth, the DPPH assay provides information on chemical radical-scavenging activity but does not demonstrate pharmacological synergy between DMF and GER. In addition, the cytotoxicity data showed concentration- and time-dependent effects at the higher DMF concentrations and therefore do not support a generalized definition of the formulation as non-cytotoxic. Regarding the in vitro permeability study, post-exposure TEER measurements, paracellular permeability markers, donor/receiver mass-balance analysis, and a dedicated 0.5% methanol vehicle control were not included. Therefore, mechanistic interpretation of the observed permeability differences is limited, and a possible contribution of the solvent to the permeability observed with free DMF cannot be excluded. Finally, the present in vivo study was pharmacokinetic in nature and did not evaluate therapeutic efficacy or safety in an experimental model of multiple sclerosis. These issues require dedicated comparative pharmacokinetic, pharmacodynamic, and local safety studies.
4. Conclusions
In the present study, an oil-in-water NE formulation was successfully developed and characterized as a promising delivery system for the administration of DMF and GER via the nose-to-brain route, aimed at the treatment of MS.
The formulation took advantage of the use of CSOA as an amphiphilic mucoadhesive stabilizer and absorption enhancer, which efficiently stabilized the lipophilic phase. Among the preparative methods evaluated, the spontaneous emulsification (SE) technique demonstrated superiority to the phase inversion (PI) one, yielding NEs with optimal physicochemical properties, including a small mean droplet size (~170 nm), a low polydispersity index (PDI < 0.3), and a positive zeta potential (>20 mV) that ensures long-term physical stability.
Thermal analyses supported by FT-IR confirmed the chemical compatibility between DMF and GER, highlighting a mutual stabilization effect.
In vitro characterization demonstrated a highly favorable biopharmaceutical profile: the NE provided a rapid drug release (exceeding 80% of the drug within 2 h) and exerted a distinct, dose-dependent ROS-scavenging activity due to the antioxidant properties of both DMF and GER. Furthermore, in vitro cell culture studies on human nasal epithelial RPMI 2650 cells demonstrated the high biocompatibility and non-cytotoxic nature of the formulation, together with good mucosal permeability, which depended strictly on the small size of the droplets.
The proof-of-concept in vivo study showed that intranasal administration of NE2.5 resulted in measurable and relatively sustained exposure of both DMF and GER in rat CSF. Notably, DMF reached concentrations of approximately 8 μg/mL in the CSF while remaining undetectable in the bloodstream, where its hydrolysis product MMF was instead observed. This distinct CSF/systemic profile, together with our previously published findings showing that free DMF administered intravenously did not reach detectable levels in CSF and that intranasal free DMF did not produce detectable DMF or MMF in CSF, supports the possibility that intranasal formulation contributes to DMF access to the CSF before extensive systemic hydrolysis. However, because the present study did not include concurrent free-drug, blank-formulation, or systemic-route comparator groups, and drug concentrations were not determined in brain tissue, the relative contribution of direct nose-to-brain transport cannot be established from the present data. Accordingly, these findings should be considered pharmacokinetic proof-of-concept evidence supporting further investigation of NE2.5 rather than a definitive demonstration of BBB bypass, direct nose-to-brain transport, or therapeutic efficacy.
Acknowledgments
Authors thank Francesco Saverio Robustelli della Cuna, Environmental Research Center, Istituti Clinici Scientifici Maugeri IRCCS, 27100 Pavia, Italy, for his support in the development of the HPLC method used during the formulation development. The authors thank Massimo Boiocchi of Centro Grandi Strumenti and Gloria Milanesi of the Biology and Biotechnology Department (University of Pavia) for their support and assistance in TEM analyses.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Supplementary Materials
The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/pharmaceutics18091168/s1 , Figure S1: Representative chromatogram recorded at 216 nm of a 100 µM (14.41 µg/mL) DMF standard solution in CSF simulation fluid constituted by a balanced Dulbecco’s Phosphate-Buffered Saline (DPBS) solution depleted of calcium and magnesium and supplemented with 0.45 mg/mL BSA; Figure S2: Representative UV spectrum of the DMF peak (R t 8.2 min) following the injection in the HPLC-DAD system of a 100 µM (14.41 µg/mL) DMF standard solution in CSF simulation fluid constituted by a balanced Dulbecco’s Phosphate-Buffered Saline (DPBS) solution depleted of calcium and magnesium and supplemented with 0.45 mg/mL BSA; Figure S3: Representative chromatogram recorded at 216 nm of a CSF sample collected 30 min after nasal administration of the NE2.5 formulation; Figure S4: Representative UV spectrum of the DMF peak (R t 8.2 min) following the injection in the HPLC-DAD system of a CSF sample collected 30 min after nasal administration of the NE2.5 formulation; Figure S5: Peak purity of the DMF peak (R t 8.2 min) following the injection in the HPLC-DAD system of a CSF sample collected 30 min after nasal administration of the NE2.5 formulation. The purity curve (black) remains aligned with the zero line (red) across the entire peak width. No impurities were detected, confirmed by a peak impurity index of 1.000000 (single-point threshold: 0.999923) and a minimum peak purity index of 76, indicating spectral homogeneity of the chromatographic peak and no detectable evidence of co-eluting impurities under the applied HPLC-DAD conditions.
Author Contributions
Conceptualization: M.C.B., M.S., A.D., L.F., L.C. and S.P.; Visualization: E.S.C., A.D., M.S., M.C.B. and L.C.; Material preparation, data collection, and formal analysis: E.S.C., G.B. and F.K.T.; Supervision: M.C.B., M.S., L.C., S.P., A.D., L.F. and S.B.; Data curation: L.C., M.S., M.C.B., S.P., G.B., S.B. and L.F.; Writing—original draft preparation: E.S.C., L.C., M.C.B. and M.S.; Writing—review and editing: L.C., M.C.B., S.P., M.S., A.D., S.B. and L.F.; Funding acquisition: A.D. All authors have read and agreed to the published version of the manuscript.
Institutional Review Board Statement
The in vivo experiments were performed under current Italian legislation (D.L. 26/2014) that allows experimentation on laboratory animals only after approval by the Ministry of Health (Rome, Italy; authorisation n° 512/2024-PR, approved date 13 October 2024), and in strict accordance with the European council directives on animal use in research (n. 2010/63/EU). Every effort was made to reduce the number of animals and their suffering. The experimental protocol was established before the study but was not preregistered.
Informed Consent Statement
Not applicable.
Data Availability Statement
The original contributions presented in this study are included in the article. Further inquiries can be directed to the corresponding author.
Conflicts of Interest
The authors declare no conflicts of interest.
References
1. Montalban X. Gold R. Thompson A.J. Otero-Romero S. Amato M.P. Chandraratna D. Clanet M. Comi G. Derfuss T. Fazekas F. ECTRIMS/EAN Guideline on the pharmacological treatment of people with multiple sclerosis Mult. Scler. J. 2018 24 96 120 10.1177/1352458517751049 29353550
2. Thompson A. Kermode A. Wicks D. Macmanus D. Kendall B. Kingsley D. Mcdonald W. Major differences in the dynamics of primary and secondary progressive multiple-sclerosis Ann. Neurol. 1991 29 53 62 10.1002/ana.410290111 1996879
3. Lassmann H. The pathology of multiple sclerosis and its evolution Philos. Trans. R. Soc. Lond. Ser. B-Biol. Sci. 1999 354 1635 1640 10.1098/rstb.1999.0508 PMC1692680 10603616
4. Comi G. Bar-Or A. Lassmann H. Uccelli A. Hartung H. Montalban X. Sorensen P. Hohlfeld R. Hauser S. Fdn E.C. Role of B Cells in Multiple Sclerosis and Related Disorders Ann. Neurol. 2021 89 13 23 10.1002/ana.25927 PMC8007167 33091175
5. Correale J. Farez M. The role of astrocytes in multiple sclerosis progression Front. Neurol. 2015 6 180 10.3389/fneur.2015.00180 PMC4539519 26347709
6. Wei W. Ma D. Li L. Zhang L. Progress in the Application of Drugs for the Treatment of Multiple Sclerosis Front. Pharmacol. 2021 12 724718 10.3389/fphar.2021.724718 PMC8313804 34326775
7. Burness C. Deeks E. Dimethyl Fumarate: A Review of Its Use in Patients with Relapsing-Remitting Multiple Sclerosis CNS Drugs 2014 28 373 387 10.1007/s40263-014-0155-5 24623127
8. Havrdova E. Hutchinson M. Kurukulasuriya N. Raghupathi K. Sweetser M. Dawson K. Gold R. Oral BG-12 (dimethyl fumarate) for relapsing-remitting multiple sclerosis: A review of DEFINE and CONFIRM Expert Opin. Pharmacother. 2013 14 2145 2156 10.1517/14656566.2013.826190 23971970
9. Litjens N. Burggraaf J. van Strijen E. van Gulpen C. Mattie H. Schoemaker R. van Dissel J. Thio H. Nibbering P. Pharmacokinetics of oral fumarates in healthy subjects Br. J. Clin. Pharmacol. 2004 58 429 432 10.1111/j.1365-2125.2004.02145.x PMC1884599 15373936
10. Lategan T. Wang L. Sprague T. Rousseau F. Pharmacokinetics and Bioavailability of Monomethyl Fumarate Following a Single Oral Dose of Bafiertam™ (Monomethyl Fumarate) or Tecfidera(R) (Dimethyl Fumarate) CNS Drugs 2021 35 567 574 10.1007/s40263-021-00799-9 PMC8144082 33797063
11. Dibbert S. Clement B. Skak-Nielsen T. Mrowietz U. Rostami-Yazdi M. Detection of fumarate-glutathione adducts in the portal vein blood of rats: Evidence for rapid dimethylfumarate metabolism Arch. Dermatol. Res. 2013 305 447 451 10.1007/s00403-013-1332-y 23525570
12. Mrowietz U. Morrison P. Suhrkamp I. Kumanova M. Clement B. The Pharmacokinetics of Fumaric Acid Esters Reveal Their In Vivo Effects Trends Pharmacol. Sci. 2018 39 1 12 10.1016/j.tips.2017.11.002 29183628
13. Serri C. Cossu M. Guarino V. Cruz-Maya I. Botti G. Ferraro L. Giunchedi P. Rassu G. Gavini E. Dalpiaz A. Intranasal delivery of dimethyl fumarate and monomethyl fumarate using hyaluronic acid-based hybrid nanoparticles to enhance CNS bioavailability J. Drug Deliv. Sci. Technol. 2025 114 107617 10.1016/j.jddst.2025.107617
14. Xu L. Peng C. Dawson K. Stecher S. Woodworth J. Prakash C. Metabolism, pharmacokinetics and excretion of [14C]dimethyl fumarate in healthy volunteers: An example of xenobiotic biotransformation following endogenous metabolic pathways Xenobiotica 2023 53 163 172 10.1080/00498254.2023.2217506 37216617
15. Wipke B. Hoepner R. Strassburger-Krogias K. Thomas A. Gianni D. Szak S. Brennan M. Pistor M. Gold R. Chan A. Different Fumaric Acid Esters Elicit Distinct Pharmacologic Responses Neurol.-Neuroimmunol. Neuroinflamm. 2021 8 e950 10.1212/NXI.0000000000000950 PMC7862084 33468560
16. Dello Russo C. Scott K. Pirmohamed M. Dimethyl fumarate induced lymphopenia in multiple sclerosis: A review of the literature Pharmacol. Ther. 2021 219 107710 10.1016/j.pharmthera.2020.107710 33091427
17. Blair H. Dimethyl Fumarate: A Review in Relapsing-Remitting MS Drugs 2019 79 1965 1976 10.1007/s40265-019-01229-3 31784875
18. Yadav S. Soin D. Ito K. Dhib-Jalbut S. Insight into the mechanism of action of dimethyl fumarate in multiple sclerosis J. Mol. Med. 2019 97 463 472 10.1007/s00109-019-01761-5 30820593
19. Kopincova J. Bernatova I. Dimethyl Fumarate vs. Monomethyl Fumarate: Unresolved Pharmacologic Issues Pharmaceutics 2025 17 1506 10.3390/pharmaceutics17121506 PMC12736242 41471022
20. Narapureddy B. Dubey D. Clinical evaluation of dimethyl fumarate for the treatment of relapsing-remitting multiple sclerosis: Efficacy, safety, patient experience and adherence Patient Prefer. Adherence 2019 13 1655 1666 10.2147/PPA.S187529 PMC6778444 31631980
21. Crowe T. Greenlee M. Kanthasamy A. Hsu W. Mechanism of intranasal drug delivery directly to the brain Life Sci. 2018 195 44 52 10.1016/j.lfs.2017.12.025 29277310
22. Frey W. Liu J. Chen X. Thorne R. Fawcett J. Ala T. Rahman Y. Delivery of I-125-NGF to the brain via the olfactory route Drug Deliv. 1997 4 87 92 10.3109/10717549709051878
23. Illum L. Transport of drugs from the nasal cavity to the central nervous system Eur. J. Pharm. Sci. 2000 11 1 18 10.1016/S0928-0987(00)00087-7 10913748
24. Selvaraj K. Gowthamarajan K. Karri V. Nose to brain transport pathways an overview: Potential of nanostructured lipid carriers in nose to brain targeting Artif. Cells Nanomed. Biotechnol. 2018 46 2088 2095 10.1080/21691401.2017.1420073 29282995
25. Xu K. Duan S. Wang W. Ouyang Q. Qin F. Guo P. Hou J. He Z. Wei W. Qin M. Nose-to-brain delivery of nanotherapeutics: Transport mechanisms and applications Wiley Interdiscip. Rev.-Nanomed. Nanobiotechnol. 2024 16 e1956 10.1002/wnan.1956 38558503
26. Gänger S. Schindowski K. Tailoring Formulations for Intranasal Nose-to-Brain Delivery: A Review on Architecture, Physico-Chemical Characteristics and Mucociliary Clearance of the Nasal Olfactory Mucosa Pharmaceutics 2018 10 116 10.3390/pharmaceutics10030116 PMC6161189 30081536
27. Serri C. Piccioni M. Guarino V. Santonicola P. Cruz-Maya I. Crispi S. Di Cagno M. Ferraro L. Dalpiaz A. Botti G. Hyaluronic Acid-Based Hybrid Nanoparticles as Promising Carriers for the Intranasal Administration of Dimethyl Fumarate Int. J. Nanomed. 2025 20 71 89 10.2147/IJN.S481917 PMC11717652 39802381
28. Cama E. Catenacci L. Perteghella S. Sorrenti M. Caira M. Rassu G. Gavini E. Giunchedi P. Bonferoni M. Design and development of a chitosan-based nasal powder of dimethyl fumarate-cyclodextrin binary systems aimed at nose-to-brain administration. A stability study Int. J. Pharm. 2024 659 124216 10.1016/j.ijpharm.2024.124216 38734272
29. González N. Rassu G. Cossu M. Catenacci L. Sorrenti M. Cama E. Serri C. Giunchedi P. Gavini E. A thermosensitive chitosan hydrogel: An attempt for the nasal delivery of dimethyl fumarate Int. J. Biol. Macromol. 2024 278 134908 10.1016/j.ijbiomac.2024.134908 39181356
30. Pinto B. Ribeiro L. da Silva G. Freitas C. Kraemer L. Oliveira F. Clímaco M. Mourao F. Dos Santos G. Béla S. Inhalation of dimethyl fumarate-encapsulated sold lipid nanoparticles attenuate clinical signs of experimental autoimmune encephalomyelitis and pulmonary inflammatory dysfunction in mice Clin. Sci. 2022 136 81 101 10.1042/CS20210792 34904644
31. Gupta A. Eral H. Hatton T. Doyle P. Nanoemulsions: Formation, properties and applications Soft Matter 2016 12 2826 2841 10.1039/c5sm02958a 26924445
32. Gupta A. Nanoemulsion Eun Ji Chung L.L.a.C.R. Elsevier Amsterdam, The Netherlands 2020 371 384
33. Bonferoni M. Rossi S. Sandri G. Ferrari F. Gavini E. Rassu G. Giunchedi P. Nanoemulsions for “Nose-to-Brain” Drug Delivery Pharmaceutics 2019 11 84 10.3390/pharmaceutics11020084 PMC6409749 30781585
34. Rinaldi F. Oliva A. Sabatino M. Imbriano A. Hanieh P. Garzoli S. Mastroianni C. De Angelis M. Miele M. Arnaut M. Antimicrobial Essential Oil Formulation: Chitosan Coated Nanoemulsions for Nose to Brain Delivery Pharmaceutics 2020 12 678 10.3390/pharmaceutics12070678 PMC7407154 32709076
35. Otaegui L. Urgin T. Zaiter T. Zussy C. Vitalis M. Pellequer Y. Acar N. Vigor C. Galano J. Durand T. Nose-to-brain delivery of DHA-loaded nanoemulsions: A promising approach against Alzheimer’s disease Int. J. Pharm. 2025 670 125125 10.1016/j.ijpharm.2024.125125 39788398
36. Ahmed B. Rizwanullah M. Mir S. Akhtar M. Amin S. Development of cannabidiol nanoemulsion for direct nose to brain delivery: Statistical optimization, in vitro and in vivo evaluation Biomed. Mater. 2022 17 065009 10.1088/1748-605X/ac9267 36108625
37. Mahajan H. Patil N. Nanoemulsion containing a synergistic combination of curcumin and quercetin for nose-to-brain delivery: In vitro and in vivo studies Asian Pac. J. Trop. Biomed. 2021 11 510 518 10.4103/2221-1691.328058
38. Tareen F. Catenacci L. Perteghella S. Sorrenti M. Bonferoni M. Carvacrol Essential Oil as a Neuroprotective Agent: A Review of the Study Designs and Recent Advances Molecules 2025 30 104 10.3390/molecules30010104 PMC11721141 39795159
39. Soliman S. Sheta N. Ibrahim B. El-Shawwa M. Abd El-Halim S. Novel Intranasal Drug Delivery: Geraniol Charged Polymeric Mixed Micelles for Targeting Cerebral Insult as a Result of Ischaemia/Reperfusion Pharmaceutics 2020 12 76 10.3390/pharmaceutics12010076 PMC7022886 31963479
40. Rekha K. Selvakumar G. Sethupathy S. Santha K. Sivakamasundari R. Geraniol Ameliorates the Motor Behavior and Neurotrophic Factors Inadequacy in MPTP-Induced Mice Model of Parkinson’s Disease J. Mol. Neurosci. 2013 51 851 862 10.1007/s12031-013-0074-9 PMC3824202 23943375
41. Bagheri S. Salehi I. Ramezani-Aliakbari F. Kourosh-Arami M. Komaki A. Neuroprotective effect of geraniol on neurological disorders: A review article Mol. Biol. Rep. 2022 49 10865 10874 10.1007/s11033-022-07755-w 35900613
42. de Oliveira E.J. Truzzi E. Ferraro L. Fogagnolo M. Pavan B. Beggiato S. Rustichelli C. Maretti E. Lima E. Leo E. Nasal administration of nanoencapsulated geraniol/ursodeoxycholic acid conjugate: Towards a new approach for the management of Parkinson’s disease J. Control. Release 2020 321 540 552 10.1016/j.jconrel.2020.02.033 32092370
43. Truzzi E. Rustichelli C. de Oliveira E.R. Ferraro L. Maretti E. Graziani D. Botti G. Beggiato S. Iannuccelli V. Lima E.M. Nasal biocompatible powder of Geraniol oil complexed with cyclodextrins for neurodegenerative diseases: Physicochemical characterization and in vivo evidences of nose to brain delivery J. Control. Release 2021 335 191 202 10.1016/j.jconrel.2021.05.020 34019946
44. Bonferoni M. Ferraro L. Pavan B. Beggiato S. Cavalieri E. Giunchedi P. Dalpiaz A. Uptake in the Central Nervous System of Geraniol Oil Encapsulated in Chitosan Oleate Following Nasal and Oral Administration Pharmaceutics 2019 11 106 10.3390/pharmaceutics11030106 PMC6471858 30832389
45. Mohammed M. Mansell H. Shoker A. Wasan K. Wasan E. Development and in vitro characterization of chitosan-coated polymeric nanoparticles for oral delivery and sustained release of the immunosuppressant drug mycophenolate mofetil Drug Dev. Ind. Pharm. 2019 45 76 87 10.1080/03639045.2018.1518455 30169982
46. Vieira A. Chaves L. Pinheiro S. Pinto S. Pinheiro M. Lima S. Ferreira D. Sarmento B. Reis S. Mucoadhesive chitosan-coated solid lipid nanoparticles for better management of tuberculosis Int. J. Pharm. 2018 536 478 485 10.1016/j.ijpharm.2017.11.071 29203137
47. Bonferoni M. Sandri G. Dellera E. Rossi S. Ferrari F. Mori M. Caramella C. Ionic polymeric micelles based on chitosan and fatty acids and intended for wound healing. Comparison of linoleic and oleic acid Eur. J. Pharm. Biopharm. 2014 87 101 106 10.1016/j.ejpb.2013.12.018 24384070
48. Gennari O. Montesano D. Salzano A. Albrizio S. Grumetto L. Determination of dimethyl fumarate in desiccant and antimould sachets by reversed-phase liquid chromatography Biomed. Chromatogr. 2011 25 1315 1318 10.1002/bmc.1602 21308706
49. Pavan B. Dalpiaz A. Marani L. Beggiato S. Ferraro L. Canistro D. Paolini M. Vivarelli F. Valerii M. Comparone A. Geraniol Pharmacokinetics, Bioavailability and Its Multiple Effects on the Liver Antioxidant and Xenobiotic-Metabolizing Enzymes Front. Pharmacol. 2018 9 18 10.3389/fphar.2018.00018 PMC5788896 29422862
50. Kommareddy S. Nigam R. Kumar S. Development and validation of a stability-indicating RP-HPLC method for the identification, determination of related substances and assay of dimethyl fumarate in drug substance Ann. Pharm. Fr. 2021 79 179 193 10.1016/j.pharma.2020.10.002 33091397
51. Chou T. Nugroho D. Chang J. Cheng Y. Liang C. Deng M. Encapsulation and Characterization of Nanoemulsions Based on an Anti-oxidative Polymeric Amphiphile for Topical Apigenin Delivery Polymers 2021 13 1016 10.3390/polym13071016 PMC8037426 33806031
52. Kürti L. Veszelka S. Bocsik A. Ozsvári B. Puskás L. Kittel A. Szabó-Révész P. Deli M. Retinoic acid and hydrocortisone strengthen the barrier function of human RPMI 2650 cells, a model for nasal epithelial permeability Cytotechnology 2013 65 395 406 10.1007/s10616-012-9493-7 PMC3597180 22940916
53. Kreft M. Jerman U. Lasic E. Rizner T. Hevir-Kene N. Peternel L. Kristan K. The Characterization of the Human Nasal Epithelial Cell Line RPMI 2650 Under Different Culture Conditions and Their Optimization for an Appropriate in vitro Nasal Model Pharm. Res. 2015 32 665 679 10.1007/s11095-014-1494-0 25145337
54. Gonçalves V. Matias A. Poejo J. Serra A. Duarte C. Application of RPMI 2650 as a cell model to evaluate solid formulations for intranasal delivery of drugs Int. J. Pharm. 2016 515 1 10 10.1016/j.ijpharm.2016.09.086 27702697
55. Gambichler T. Brüggestrat L. Skrygan M. Scheel C. Susok L. Becker J. The Antineoplastic Effect of Dimethyl Fumarate on Virus-Negative Merkel Cell Carcinoma Cell Lines: Preliminary Results Cancers 2023 15 547 10.3390/cancers15020547 PMC9857057 36672496
56. Sibinovska N. Zakelj S. Kristan K. Suitability of RPMI 2650 cell models for nasal drug permeability prediction Eur. J. Pharm. Biopharm. 2019 145 85 95 10.1016/j.ejpb.2019.10.008 31639418
57. Martin J. Rittersberger R. Treitler S. Kopp P. Ibraimi A. Koslowski G. Sickinger M. Dabbars A. Schindowski K. Characterization of a primary cellular airway model for inhalative drug delivery in comparison with the established permanent cell lines CaLu3 and RPMI 2650 In Vitr. Model. 2024 3 183 203 10.1007/s44164-024-00079-y PMC11756470 39872698
58. Hubatsch I. Ragnarsson E. Artursson P. Determination of drug permeability and prediction of drug absorption in Caco-2 monolayers Nat. Protoc. 2007 2 2111 2119 10.1038/nprot.2007.303 17853866
59. Lee J. Zgair A. Taha D. Zang X. Kagan L. Kim T. Kim M. Yun H. Fischer P. Gershkovich P. Quantitative analysis of lab-to-lab variability in Caco-2 permeability assays Eur. J. Pharm. Biopharm. 2017 114 38 42 10.1016/j.ejpb.2016.12.027 28088633
60. van den Berg M. Romeijn S. Verhoef J. Merkus F. Serial cerebrospinal fluid sampling in a rat model to study drug uptake from the nasal cavity J. Neurosci. Methods 2002 116 99 107 10.1016/S0165-0270(02)00033-X 12007987
61. Dalpiaz A. Ferraro L. Perrone D. Leo E. Iannuccelli V. Pavan B. Paganetto G. Beggiato S. Scalia S. Brain Uptake of a Zidovudine Prodrug after Nasal Administration of Solid Lipid Microparticles Mol. Pharm. 2014 11 1550 1561 10.1021/mp400735c 24717116
62. Madu A. Cioffe C. Mian U. Burroughs M. Tuomanen E. Mayers M. Schwartz E. Miller M. Pharmacokinetics of fluconazole in cerebrospinal-fluid and serum of rabbits—Validation of an animal-model used to measure drug concentrations in cerebrospinal-fluid Antimicrob. Agents Chemother. 1994 38 2111 2115 10.1128/AAC.38.9.2111 PMC284693 7811028
63. Felgenhauer K. Protein Size and Cerebrospinal Fluid Composition J. Mol. Med. 1974 52 1158 1164 10.1007/bf01466734 4456012
64. Al-Rimawi F. Odeh I. Development and Validation of an HPLC-UV Method for Determination of Eight Phenolic Compounds in Date Palms J. Aoac Int. 2015 98 1335 1339 10.5740/jaoacint.15-010 26525252
65. Wang Y. Pan Z. Qin W. Jiao F. A simple and rapid HPLC method for the quantitative determination of sodium oleate in flotation study Miner. Eng. 2019 141 105842 10.1016/j.mineng.2019.105842
66. Yan X. Evenocheck H. Chitosan analysis using acid hydrolysis and HPLC/UV Carbohydr. Polym. 2012 87 1774 1778 10.1016/j.carbpol.2011.09.091
67. Jaiswal M. Dudhe R. Sharma P. Nanoemulsion: An advanced mode of drug delivery system 3 Biotech 2015 5 123 127 10.1007/s13205-014-0214-0 PMC4362737 28324579
68. Ahmad N. Alam M. Ahmad F. Sarafroz M. Ansari K. Sharma S. Amir M. Ultrasonication techniques used for the preparation of novel Eugenol-Nanoemulsion in the treatment of wounds healings and anti-inflammatory J. Drug Deliv. Sci. Technol. 2018 46 461 473 10.1016/j.jddst.2018.06.003
69. Singh Y. Meher J. Raval K. Khan F. Chaurasia M. Jain N. Chourasia M. Nanoemulsion: Concepts, development and applications in drug delivery J. Control. Release 2017 252 28 49 10.1016/j.jconrel.2017.03.008 28279798
70. Ding W. Shah N. Effect of Homogenization Techniques on Reducing the Size of Microcapsules and the Survival of Probiotic Bacteria Therein J. Food Sci. 2009 74 M231 M236 10.1111/j.1750-3841.2009.01195.x 19723206
71. Komaiko J. McClements D. Formation of Food-Grade Nanoemulsions Using Low-Energy Preparation Methods: A Review of Available Methods Compr. Rev. Food Sci. Food Saf. 2016 15 331 352 10.1111/1541-4337.12189 33371595
72. Kumar A. Li S. Cheng C. Lee D. Recent Developments in Phase Inversion Emulsification Ind. Eng. Chem. Res. 2015 54 8375 8396 10.1021/acs.iecr.5b01122
73. Perazzo A. Preziosi V. Guido S. Phase inversion emulsification: Current understanding and applications Adv. Colloid Interface Sci. 2015 222 581 599 10.1016/j.cis.2015.01.001 25632889
74. Feng J. Rodríguez-Abreu C. Esquena J. Solans C. A Concise Review on Nano-emulsion Formation by the Phase Inversion Composition (PIC) Method J. Surfactants Deterg. 2020 23 677 685 10.1002/jsde.12414
75. Gulotta A. Saberi A. Nicoli M. McClements D. Nanoemulsion-Based Delivery Systems for Polyunsaturated (ω-3) Oils: Formation Using a Spontaneous Emulsification Method J. Agric. Food Chem. 2014 62 1720 1725 10.1021/jf4054808 24475908
76. Guttoff M. Saberi A. McClements D. Formation of vitamin D nanoemulsion-based delivery systems by spontaneous emulsification: Factors affecting particle size and stability Food Chem. 2015 171 117 122 10.1016/j.foodchem.2014.08.087 25308650
77. Li Z. Xu D. Yuan Y. Wu H. Hou J. Kang W. Bai B. Advances of spontaneous emulsification and its important applications in enhanced oil recovery process Adv. Colloid Interface Sci. 2020 277 102119 10.1016/j.cis.2020.102119 32045722
78. Luo Y. Teng Z. Li Y. Wang Q. Solid lipid nanoparticles for oral drug delivery: Chitosan coating improves stability, controlled delivery, mucoadhesion and cellular uptake Carbohydr. Polym. 2015 122 221 229 10.1016/j.carbpol.2014.12.084 25817662
79. Paul P. Sengupta S. Mukherjee B. Shaw T. Gaonkar R. Debnath M. Chitosan-coated nanoparticles enhanced lung pharmacokinetic profile of voriconazole upon pulmonary delivery in mice Nanomedicine 2018 13 501 520 10.2217/nnm-2017-0291 29383985
80. Sarheed O. Dibi M. Ramesh K. Studies on the Effect of Oil and Surfactant on the Formation of Alginate-Based O/W Lidocaine Nanocarriers Using Nanoemulsion Template Pharmaceutics 2020 12 1223 10.3390/pharmaceutics12121223 PMC7766092 33348692
81. Tan S. Masoumi H. Karjiban R. Stanslas J. Kirby B. Basri M. Bin Basri H. Ultrasonic emulsification of parenteral valproic acid-loaded nanoemulsion with response surface methodology and evaluation of its stability Ultrason. Sonochemistry 2016 29 299 308 10.1016/j.ultsonch.2015.09.015 26585010
82. An Y. Yan X. Li B. Li Y. Microencapsulation of capsanthin by self-emulsifying nanoemulsions and stability evaluation Eur. Food Res. Technol. 2014 239 1077 1085 10.1007/s00217-014-2328-3
83. Wany A. Kumar A. Nallapeta S. Jha S. Nigam V. Pandey D. Extraction and characterization of essential oil components based on geraniol and citronellol from Java citronella (Cymbopogon winterianus Jowitt) Plant Growth Regul. 2014 73 133 145 10.1007/s10725-013-9875-7
84. de Andrade L. Guilger-Casagrande M. Germano-Costa T. de Lima R. Polymeric Nanorepellent Systems Containing Geraniol and Icaridin Aimed at Repelling Aedes aegypti Int. J. Mol. Sci. 2022 23 8317 10.3390/ijms23158317 PMC9368950 35955452
85. Hussain A. Altamimi M. Ramzan M. Mirza M. Khuroo T. GastroPlus- and HSPiP-Oriented Predictive Parameters as the Basis of Valproic Acid-Loaded Mucoadhesive Cationic Nanoemulsion Gel for Improved Nose-to-Brain Delivery to Control Convulsion in Humans Gels 2023 9 603 10.3390/gels9080603 PMC10453491 37623058
86. Comfort C. Garrastazu G. Pozzoli M. Sonvico F. Opportunities and Challenges for the Nasal Administration of Nanoemulsions Curr. Top. Med. Chem. 2015 15 356 368 10.2174/1568026615666150108144655 25579345
87. Jagdale S. Shewale N. Kuchekar B. Optimization of Thermoreversible In Situ Nasal Gel of Timolol Maleate Scientifica 2016 2016 6401267 10.1155/2016/6401267 PMC4884840 27293975
88. Tareen F. Perteghella S. Catenacci L. Ghiozzi G. Cama E. della Cuna F. Sorrenti M. Bonferoni M. Carvacrol-based nanoemulsions loaded with dimethyl fumarate intended for nose to brain delivery for treatment of multiple sclerosis Drug Deliv. Transl. Res. 2026 16 579 597 10.1007/s13346-025-01912-x 40617997
89. Nguyen T. Duong V. Advancements in Nanocarrier Systems for Nose-to-Brain Drug Delivery Pharmaceuticals 2025 18 615 10.3390/ph18050615 PMC12115159 40430435
90. Koo J. Lim C. Oh K. Recent Advances in Intranasal Administration for Brain-Targeting Delivery: A Comprehensive Review of Lipid-Based Nanoparticles and Stimuli-Responsive Gel Formulations Int. J. Nanomed. 2024 19 1767 1807 10.2147/IJN.S439181 PMC10898487 38414526
91. Khan A. Khan R. Qamar W. Lateef A. Rehman M. Tahir M. Ali F. Hamiza O. Hasan S. Sultana S. Geraniol attenuates 12-O-tetradecanoylphorbol-13-acetate (TPA)-induced oxidative stress and inflammation in mouse skin: Possible role of p38 MAP Kinase and NF-κB Exp. Mol. Pathol. 2013 94 419 429 10.1016/j.yexmp.2013.01.006 23399806
92. Peñuelas J. Munné-Bosch S. Isoprenoids: An evolutionary pool for photoprotection Trends Plant Sci. 2005 10 166 169 10.1016/j.tplants.2005.02.005 15817417
93. Scuderi S. Ardizzone A. Paterniti I. Esposito E. Campolo M. Antioxidant and Anti-inflammatory Effect of Nrf2 Inducer Dimethyl Fumarate in Neurodegenerative Diseases Antioxidants 2020 9 630 10.3390/antiox9070630 PMC7402174 32708926
94. Sun F. Chou Y. Huang Q. Hu J. Ting Y. Making Concentrated Pterostilbene Highly Bioavailable in Pressure Processed Phospholipid Nanoemulsion Processes 2021 9 294 10.3390/pr9020294
95. Rekha K. Selvakumar G. Santha K. Sivakamasundari R. Geraniol attenuates α-synuclein expression and neuromuscular impairment through increase dopamine content in MPTP intoxicated mice by dose dependent manner Biochem. Biophys. Res. Commun. 2013 440 664 670 10.1016/j.bbrc.2013.09.122 24103762
96. Moriki Y. Mitsugi R. Kayou T. Horikoshi J. Yamaguchi Y. Shibuya S. Shimizu T. Pharmacokinetics of Geraniol and Its Metabolites in Mice After Oral Administration Food Sci. Nutr. 2025 13 e4653 10.1002/fsn3.4653 PMC11716996 39803254
97. Kumar P. Sharma G. Kumar R. Malik R. Singh B. Katare O. Raza K. Enhanced Brain Delivery of Dimethyl Fumarate Employing Tocopherol-Acetate-Based Nanolipidic Carriers: Evidence from Pharmacokinetic, Biodistribution, and Cellular Uptake Studies ACS Chem. Neurosci. 2017 8 860 865 10.1021/acschemneuro.6b00428 27983793