Solid Lipid Nanoparticles Containing Morin: Preparation, Characterization, and Ex Vivo Permeation Studies

In recent years, bioactive compounds have been the focus of much interest in scientific research, due to their low toxicity and extraordinary properties. However, they possess poor solubility, low chemical stability, and unsustainable bioavailability. New drug delivery systems, and among them solid lipid nanoparticles (SLNs), could minimize these drawbacks. In this work, morin (MRN)-loaded SLNs (MRN-SLNs) were prepared using a solvent emulsification/diffusion method, using two different lipids, Compritol® 888 ATO (COM) or Phospholipon® 80H (PHO). SLNs were investigated for their physical–chemical, morphological, and technological (encapsulation parameters and in vitro release) properties. We obtained spherical and non-aggregated nanoparticles with hydrodynamic radii ranging from 60 to 70 nm and negative zeta potentials (about −30 mV and −22 mV for MRN-SLNs-COM and MRN-SLNs-PHO, respectively). The interaction of MRN with the lipids was demonstrated via μ-Raman spectroscopy, X-ray diffraction, and DSC analysis. High encapsulation efficiency was obtained for all formulations (about 99%, w/w), particularly for the SLNs prepared starting from a 10% (w/w) theoretical MRN amount. In vitro release studies showed that about 60% of MRN was released within 24 h and there was a subsequent sustained release within 10 days. Finally, ex vivo permeation studies with excised bovine nasal mucosa demonstrated the ability of SLNs to act as a penetration enhancer for MRN due to the intimate contact and interaction of the carrier with the mucosa.


Introduction
Neurodegenerative diseases such as Alzheimer's, Parkinson's, Huntington's, multiple sclerosis, and ischemic stroke represent a threat to world health, and the incidence of these pathologies dramatically increases with age [1,2]. These disorders commonly lead to the irreversible destruction of neuronal networks, resulting in permanent damage to the central drug, and above all to obtain a prolonged release of MRN in the brain. MRN-SLNs were obtained via a solvent emulsification/diffusion method using Compritol ® 888 ATO (COM) or Phospholipon ® 80H (PHO) as the lipid and Polysorbate 80 as the surfactant. A deep characterization study was carried out in order to achieve the best formulation in terms of sizes, electric surface properties, encapsulation efficiency, drug loading, and release rate. Scanning electron microscopy (SEM) was used for morphological characterization. A detailed physical-chemical characterization of the systems was conducted via µ-Raman spectroscopy, differential scanning calorimetry (DSC), and thermogravimetric (TG) analyses. Finally, ex vivo permeation studies of excised bovine nasal mucosa were carried out to evaluate the ability of the delivery systems to improve and prolong the permeation of MRN through the mucosa.

Preparation of Blank SLNs and MRN-SLNs
The preparation of the blank SLNs was performed according to the solvent emulsification/diffusion method reported by Butani et al. [48], with slight modification.
Briefly, two phases were prepared: (i) a lipid phase, formed of 50 mg of PHO or 50 mg of COM solubilized in 2.5 mL methanol, (ii) an aqueous phase, formed of 11.25 mL of water containing 170 µL of Tween ® 80 (0.5% w/v) [49]. Both solutions were heated in a water bath; in particular, the COM solution was heated at a temperature above 80 • C and the PHO solution was heated at a temperature above 60 • C. The lipid phase was added dropwise during the aqueous phase, during high-speed homogenization, using Ultra-Turrax T 25 (IKA-Werke, Staufen, Germany) at 11,000 rpm for 7 min. To reduce the sizes of the samples, the suspensions underwent ultrasonication (Bandelin Sonorex ® RK 514, amplitude delivered ranges between 3 and 5 µm, Berlin, Germany) for 5 min. After that, they were allowed to cool to room temperature under magnetic stirring (500 rpm) for 24 h to permit the evaporation of the organic solvent. Purification of the samples was carried out via centrifugation at 5000 rpm for 30 min (Heraeus Megafuge 16, ThermoFisher scientific, Waltham, MA, USA) to remove the lipids that did not interact. The obtained pellets were discarded, while the supernatant, containing the SLNs, was collected; 5% (w/v) trehalose as a cryoprotectant was added, and it was freeze-dried for 72 h (BenchTop K Series Freeze Dryers, VirTis, Gardiner, NY, USA).
MRN-SLNs were prepared according to the aforementioned procedure, with the addition of different concentrations of MRN to the lipid phase (2%, 5%, 8%, and 10%, w/w, based on lipid). The pellets obtained from the centrifugation at 5000 rpm for 30 min, and containing the lipids that did not interact and the drug that was not encapsulated, were collected and used for the indirect determination of MRN encapsulated in SLNs. Trehalose (5%, w/v) was added to the supernatant and the MRN-SLNs were lyophilized for 72 h.

Technological Characterization of MRN-SLNs
Freeze-dried blank SLNs and MRN-SLNs based on COM (MRN-SLNs-COM) or on PHO (MRN-SLNs-PHO) were weighed, and the yield percentage was determined following the formula in Equation (1): Yield (%) = (Effective yield/Theoretical yield) × 100 (1) The encapsulation efficiency (E.E.) and drug content (D.C.) percentage of the MRN-SLNs were indirectly determined via the quantification of the drug in the pellets; then, the following equations, Equations (2)  To determine the amount of free MRN, the pellets derived from the centrifugation at 5000 rpm were treated with cold methanol, since this organic solvent permits the solubilization of the drug, leaving the lipid insoluble. Hence, this suspension was filtered and analyzed via HPLC (Section 2.11).
The results are reported as the mean of three separate measurements on three different batches ± the standard deviation (S.D.).

Thermogravimetric Analysis
Thermogravimetric analysis (TGA) was performed using the TAQ500 instrument (TA Instruments, New Castle, DE, USA) under argon flow at a flow rate of 100 mL/min. All samples tested, after a vacuum drying procedure, were heated up to 700 • C with a heating rate of 10 • C/min.

Differential Scanning Calorimetry Analysis
Differential scanning calorimetry (DSC) was performed using the TAQ500 instrument (TA Instruments, New Castle, DE, USA) under nitrogen flow at a flow rate of 50 mL/min, from room temperature to 350 • C, with a heating rate of 5 • C/min.

X-ray Diffraction Analysis
All X-ray diffraction (XRD) experiments were performed at room temperature using a Bruker D8 Advance diffractometer (Bruker, Karlsruhe, Germany) and a Bragg-Brentano theta-2theta configuration, with Cu Ka radiation (40 V, 40 mA). XRD patterns were collected in the range of 5-80 • with a step of 0.2 • /s. Diffraction peaks were compared with those in the Joint Committee on Powdered Diffraction Standards (JCPDS) database.

Scanning Electron Microscopy (SEM) Analysis
The morphology of the samples was investigated using SEM analysis (Gemini field emission SEM, Carl Zeiss Supra 25). The acceleration voltage was varied in the range of 2-5 kV and the analyses were performed using Everhart-Thornley and in-lens detectors. The freeze-dried SLNs were dispersed in deionized water; then, a drop of the suspensions was cast onto a conductive substrate and dried in air. To limit the local charging induced by the electron beam, the samples were coated with a thin gold film (~10 nm) via sputtering deposition.

µ-Raman Spectroscopy Analysis
µ-Raman analysis was conducted on MRN, PHO, COM, MRN-SLNs-PHO, and MRN-SLNs-COM solid systems using a "BTR 111 Mini-RamTM" (B&W Tek, Inc., Newark, NJ, USA) portable spectrometer, with a diode laser source operating at 785 nm, maximum laser power at the excitation port of 280 mW, and a charge-coupled device (CCD) detector (thermoelectric cooled, TE). Spectra were recorded in the 500-1800 cm −1 range using a resolution of 10 cm −1 , and an acquisition time of 10 s × 32 scans. Calibration was performed by means of the peak of a silicon chip at 520.6 cm −1 , thus guaranteeing the best instrument performance. The system was equipped with a BAC151B Raman microscope. An 80× objective was used, with a working distance of 1.25 mm and laser beam spot size of 26 µm. The maximum power delivered to the samples was~120 mW.

In Vitro Release Studies
The release of MRN from SLNs was carried out using the most versatile and popular dialysis bag method [50][51][52], already described by our research group [44]. A weighted amount of freeze-dried MRN-SLNs-PHO and MRN-SLNs-COM formulations (both containing 5 mg of MRN) was suspended in 5 mL of phosphate-buffered solution (PBS, pH 7.4) and placed into the dialysis bag, previously wetted in PBS (pH 7.4) for 24 h. Hence, the bag was immersed in 150 mL of PBS (pH 7.4) containing 15% (v/v) ethanol as the dialysis medium and left under magnetic stirring (300 rpm) at 37.0 ± 0.5 • C (Telesystem 15.40 thermostatic bath with a Telemodul 40 • C control unit). At fixed time intervals (0.5, 1, 3, 5, 24, 48, 72, 120, 168, 240 h), the dialysis medium was removed and replenished with a fresh PBS/ethanol (15%, v/v) mixture, at the same temperature, to maintain the sink conditions. The dosage of the released MRN was performed via HPLC (Section 2.11).

Ex Vivo Permeation Studies in Excised Bovine Nasal Mucosa
Ex vivo permeation studies were carried out using Franz-type diffusion cells (LGA, Berkeley, CA, USA). The bovine nasal mucosa was excised from healthy bovines (15 ± 3 months of age) immediately after slaughtering and was maintained for the appropriate time in a PBS (pH 7.4) solution containing heparin. The integrity of the nasal mucosa was monitored by measuring transepithelial electrical resistance (TEER) before and after the in vitro permeation study using a voltmeter (Millicell ERS). Each mucosa section was mounted in a Franz diffusion cell with the external face side up between the donor and the receptor compartment. The latter was filled with a hydroalcoholic acceptor solution (PBS (pH 7.4):EtOH, 60:40, v/v) to enhance the affinity of MRN for the receptor phase, guaranteeing a pseudo-sink condition (MRN solubility in the medium, at 37 ± 0.5 • C, was 4.73 mg/mL). During the entire duration of the experiments, the receptor solution was constantly stirred by a small magnetic stirring bar and maintained at 37 ± 0.5 • C. MRN (13 mM), MRN-SLNs-COM, and MRN-SLNs-PHO both containing the same amount of MRN were suspended in PBS (pH 7.4); then, 200 µL [55] of each sample was poured in the donor compartment of Franz's cells. At prefixed times, 300 µL of the receiving solution was withdrawn and the sample volumes were replaced with the same amounts of the fresh receptor phase. All recovered aliquots were analyzed via HPLC to determine the concentration of the permeated MRN.
At the end of the permeation experiment, the amount of MRN accumulated in the mucosa was measured. The mucosa was washed with water to remove any remaining formulation, and subsequently comminuted using a surgical blade, resuspended in 5 mL of water, and homogenized at 11,000 rpm for 5 min using IKA Ultra-Turrax 1 (IKA1Werke GmbH & Co. KG, Staufen, Germany). After that, an additional minute of homogenization was performed, adding 5 mL of methanol. Each sample was centrifuged at 13,000 rpm for 30 min. The pellet was discarded, and the supernatant was collected. Then, the solvent was removed under reduced pressure, yielding residues which were added to 2 mL of methanol to solubilize the MRN. After filtration (0.22 µm Millipore filter), the solutions were injected into HPLC to determine the MRN [56,57]. The described procedure was also performed on non-treated mucosa and the final solution was injected into HPLC to determine the specificity of the method.

HPLC Analysis
The quantification of MRN was performed via reverse-phase HPLC (HPLC Prominence LC-20AB pump, Shimadzu North America, Columbia, MD, USA), following a method reported in the literature [58], with slight modification. Briefly, 20 µL of each sample was injected into a C-18 column (Discovery C18 column, 250 mm × 4.6 mm inner diameter (i.d.), 5 µm, Supelco) and the analysis was performed at a flow rate of 1 mL/min using a mobile phase consisting of 0.5% acetic acid in water/acetonitrile (80:20, v/v). MRN was detected at 355 nm using a UV-Vis detector (Shimadzu UV-Vis detector SPD-20A). A chromatogram of MRN is reported in the Supplementary Materials ( Figure S1), as an example.

Analytical Method Validation
The HPLC method was validated for linearity, specificity, sensitivity, and repeatability. A stock solution of MRN was prepared by dissolving 10 mg of the drug in 100 mL of methanol in a calibrated flask. Standard solutions were prepared at concentrations ranging from 2 µg/mL to 400 µg/mL, starting from stock solution. Six replicates of each concentration were prepared and analyzed in triplicate. The calibration curve was linear in the range of 10-400 µg/mL, obtaining an R 2 value of 0.998 ( Figure S2 in Supplementary Materials). The limit of detection (LOD) and limit of quantitation (LOQ) were determined via the signal-to-noise ratio [59], obtaining 3 µg/mL (corresponding to 3:1 signal-to-noise ratio) and 10 µg/mL (corresponding to 10:1 signal-to-noise ratio) for LOD and LOQ, respectively. The precision of the method was determined intraday and interday by injecting three of the calibration standards (10, 200, 400 µg/mL) six times during the same day and for the next five days. High intraday (RSD, 0.55%) and interday (RSE, 0.45%) precision was obtained. To determine the specificity of the method, a methanol solution derived from the extraction procedure on treated mucosa or non-treated mucosa (see Section 2.10) and pure solvent were analyzed in triplicate. No interference was observed with the peak of MRN, demonstrating that the method was specific for the drug. No degradation of MRN was observed during analysis.

Statistical Analysis
All values are expressed as mean ± standard deviation (SD) and each analysis was performed three times. The results were analyzed via one-way and two-way analysis of variance (ANOVA) followed by a Bonferroni post hoc test for multiple comparisons. A value of p < 0.05 was considered to be significant.
COM is a lipid consisting of mono-, di-, and triesters of behenic acid, with the diester fraction being predominant. Recently, it has received great interest as a lipid for SLN fabrication due to its low cytotoxicity, compared to other lipids, and ability to encapsulate a high amount of hydrophobic and hydrophilic molecules, producing their sustained release [60][61][62][63]. PHO is a natural emulsifier derived from soya oil. Despite this compound principally being used as a surfactant or as a phospholipid for liposome preparation [64], different papers reported PHO as being a co-lipid for SLN production [65,66]. Furthermore, in a previous study, we demonstrated the feasibility of PHO as a unique lipid constituent of SLNs, obtaining small particles able to produce the sustained release of the natural active rutin [44].
Before proceeding with the preparation of the systems, the stability of the drug in methanol (60 • C and 80 • C) and in a PBS (pH 7.4)/ethanol (15%, v/v) mixture was evaluated. In particular, the stability in methanol was investigated within 2 h, because it represents the time required to complete the preparation of the systems; instead, the stability of MRN in the release medium (PBS (pH 7.4)/ethanol (15%, v/v)) was carried out in 10 days at 37.0 ± 0.5 • C, leaving the drug in the dark and in the light. At fixed times, the solutions were analyzed via HPLC at 355 nm. As also reported by other authors, no significant degradation of MRN was detected for the experimental conditions, nor in the release medium [26].
In Table 1, the encapsulation parameters of the prepared SLNs are reported. Comparable yields, encapsulation efficiency, and drug content percentage were obtained for both of the systems with all of the MRN theoretical amounts tested. For all of the samples, good yields and optimum encapsulation efficiency were obtained. In particular, the best result was obtained using the highest MRN theoretical amount, with encapsulation efficiency (E.E. %) ≥ 99% (w/w) and drug content (D.C. %) of 4% (w/w), indicating the great affinity of the drug for both of the lipids used in this study. Other authors synthesized SLNs loaded with MRN using lipids other than COM and PHO, obtaining a lower E.E.% of MRN with respect to our systems (about 87% and 80%, w/w) [46,47].  Based on the good values obtained for the encapsulation parameters, to continue the successive studies, the samples with 10% theoretical amount of MRN were chosen (MRN-SLNs-COM 10 and MRN-SLNs-PHO 10 ). In Table 2, we report the hydrodynamic radius (R H ), polydispersity index (P.D.I.), and zeta potential (ζ) values of both of the formulations, comparative to blank SLNs-COM and SLNs-PHO. The blank SLNs showed small particle sizes and homogeneous distribution (see the low values of P.D.I. in Table 2), particularly for the SLNs prepared using PHO. A slight increase in sizes was observed for both of the formulations in the presence of the drug; however, this did not affect the suitability of the delivery systems for the proposed therapeutic objective. All of the formulations showed high negative ζ values, useful for guaranteeing the high physical stability of the systems. Table 2. Hydrodynamic radius, (R H ), polydispersity index (PDI), and zeta potential (ζ) values of blank SLNs, MRN-SLNs-COM 10 , and MRN-SLNs-PHO 10 samples. 10 76 10 72.95 ± 6.2 0.298 ± 0.05

MRN-SLNs-COM MRN-SLNs-PHO
The morphology of the SLNs was investigated via SEM. In Figure 1, we show the pictures obtained for re-dispersed MRN-SLNs-COM 10 powder. We observed spherical particles with a smooth surface and homogeneous sizes, with a very low grade of aggregation. Similar results were observed for MRN-SLNs-PHO 10 . formulations, comparative to blank SLNs-COM and SLNs-PHO. The blank SLNs showed small particle sizes and homogeneous distribution (see the low values of P.D.I. in Table 2), particularly for the SLNs prepared using PHO. A slight increase in sizes was observed for both of the formulations in the presence of the drug; however, this did not affect the suitability of the delivery systems for the proposed therapeutic objective. All of the formulations showed high negative ζ values, useful for guaranteeing the high physical stability of the systems. The morphology of the SLNs was investigated via SEM. In Figure 1, we show the pictures obtained for re-dispersed MRN-SLNs-COM10 powder. We observed spherical particles with a smooth surface and homogeneous sizes, with a very low grade of aggregation. Similar results were observed for MRN-SLNs-PHO10.

Physical-Chemical Characterization of the SLNs
The crystalline structure and polymorphic shapes of the lipids are decisive for the efficient incorporation of the drug into the SLNs and its subsequent release. These two parameters determine the effective applicability of the SLNs as drug delivery systems. Then, an exhaustive physical-chemical characterization of the systems is essential [44]. Differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), and X-ray scattering are widely used to investigate the status of the lipids into SLNs.

Physical-Chemical Characterization of the SLNs
The crystalline structure and polymorphic shapes of the lipids are decisive for the efficient incorporation of the drug into the SLNs and its subsequent release. These two parameters determine the effective applicability of the SLNs as drug delivery systems. Then, an exhaustive physical-chemical characterization of the systems is essential [44]. Differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), and X-ray scattering are widely used to investigate the status of the lipids into SLNs. the crystallization of the drug. The first endothermic peak refers to the loss of water from the sample; the second instead refers to the melting. As observed by other authors for SLNs with different compositions [46], MRN-SLNs-COM ( Figure 2a) and MRN-SLNs-PHO (Figure 2b) show the disappearance of the MRN melting peak (at about 290 • C), evidencing the interaction between the drug and the lipids, and the formation of hydrogen bonds or van der Waals interactions. the thermograms of COM and PHO are attributable to the melting of the different components. As a consequence of the SLNs' formation, a shift toward a higher temperature of these peaks was observed. The thermogram of MRN shows two endothermic peaks at about 170 °C and 290 °C and an exothermal peak at about 200 °C. The latter evidences the crystallization of the drug. The first endothermic peak refers to the loss of water from the sample; the second instead refers to the melting. As observed by other authors for SLNs with different compositions [46],

Thermogravimetric Analysis
In this study, TGA analyses were performed in an inert environment, at the heating rate of 10 °C/min. For each sample, TGA analysis was evaluated up to 700 °C in order to investigate the thermal stability and purity of the blank SLNs and MRN-loaded SLNs. In Figure 3, we report the TGA profiles of the COM-based systems ( Figure 3a) and PHObased systems ( Figure 3b).
As shown in Figure 3a, the COM sample is stable up to 350 °C; then, degradation in the range of 350-450 °C with a residual weight of 5% can be observed. The drug shows a first weight loss of about 4% in the 85-125 °C range, a second weight loss of 37% in the 250-350 °C range, and a final residual weight of 35% at the final temperature of 650 °C. The blank SNLs-COMs are stable up to 50 °C; in the 55-110 °C range, they show a first weight loss of 11%, almost certainly corresponding to the loss of absorbed water, and a second weight loss occurs in the 270-480 °C range of 56%, attributable to the SLNs-COM degradation, with a residual weight at the final temperature of 21%. The MRN-SLNs-COM10 sample shows an initial loss in weight in the 45-80 °C range of 10%, also attributable to absorbed water. After this first phase, it maintains a constant profile up to 200 °C; then, in the 205-550 °C range, it loses 97%, with a residual weight at the final temperature of 3%.
The graph in Figure 3b shows the weight loss profile of the MRN-SLNs-PHO10 systems. The PHO sample is stable up to 200 °C; after which, there is a massive weight loss in the 205-460 °C range of 90%, maintaining a residual weight at the final temperature of 10%. Three different stages of degradation are observed for the blank SLNs-PHO: the first is at 70-100 °C, attributable to water loss, the second is at 190-300 °C, with a maximum decomposition at 254 °C, and the third is at 300-400 °C, with a maximum decomposition temperature at 371 °C and a complete degradation of the sample with a total weight loss

Thermogravimetric Analysis
In this study, TGA analyses were performed in an inert environment, at the heating rate of 10 • C/min. For each sample, TGA analysis was evaluated up to 700 • C in order to investigate the thermal stability and purity of the blank SLNs and MRN-loaded SLNs. In Figure 3, we report the TGA profiles of the COM-based systems (Figure 3a) and PHO-based systems (Figure 3b).

X-ray Diffraction
Both of the investigated systems show similar diffractograms (Figure 4). We observed the presence of sharp peaks at 2θ = 10.0°, 15.0°, 25.0°, and 28.0° in the diffractogram of the free drug, evidencing its crystalline status. A single diffraction peak was observed at about 2θ = 21.0° for both COM (Figure 4a) and PHO (Figure 4b). Sharp peaks were observed in the diffractograms of blank SLNs, demonstrating a different crystalline structure of the lipids in the matrix with respect to bulk solid lipids. The disappearance of the principal peaks of MRN in the diffractograms of the drug-loaded SLNs could be due to the incor- As shown in Figure 3a, the COM sample is stable up to 350 • C; then, degradation in the range of 350-450 • C with a residual weight of 5% can be observed. The drug shows a first weight loss of about 4% in the 85-125 • C range, a second weight loss of 37% in the 250-350 • C range, and a final residual weight of 35% at the final temperature of 650 • C. The blank SNLs-COMs are stable up to 50 • C; in the 55-110 • C range, they show a first weight loss of 11%, almost certainly corresponding to the loss of absorbed water, and a second weight loss occurs in the 270-480 • C range of 56%, attributable to the SLNs-COM degradation, with a residual weight at the final temperature of 21%. The MRN-SLNs-COM 10 sample shows an initial loss in weight in the 45-80 • C range of 10%, also attributable to absorbed water. After this first phase, it maintains a constant profile up to 200 • C; then, in the 205-550 • C range, it loses 97%, with a residual weight at the final temperature of 3%.
The graph in Figure 3b shows the weight loss profile of the MRN-SLNs-PHO 10 systems. The PHO sample is stable up to 200 • C; after which, there is a massive weight loss in the 205-460 • C range of 90%, maintaining a residual weight at the final temperature of 10%. Three different stages of degradation are observed for the blank SLNs-PHO: the first is at 70-100 • C, attributable to water loss, the second is at 190-300 • C, with a maximum decomposition at 254 • C, and the third is at 300-400 • C, with a maximum decomposition temperature at 371 • C and a complete degradation of the sample with a total weight loss of about 90%. The MRN-SLNs-PHO 10 sample exhibits an initial weight loss in the 50-110 • C range of 10%, attributable to absorbed water; after this first phase, it maintains a constant profile up to 250 • C. Then, in the range between 255 • C and 550 • C, it loses 65%, with a residual weight at the final temperature of 33%. of about 90%. The MRN-SLNs-PHO10 sample exhibits an initial weight loss in the 50-110 °C range of 10%, attributable to absorbed water; after this first phase, it maintains a constant profile up to 250 °C. Then, in the range between 255 °C and 550 °C, it loses 65%, with a residual weight at the final temperature of 33%.

X-ray Diffraction
Both of the investigated systems show similar diffractograms (Figure 4). We observed the presence of sharp peaks at 2θ = 10.0°, 15.0°, 25.0°, and 28.0° in the diffractogram of the free drug, evidencing its crystalline status. A single diffraction peak was observed at about 2θ = 21.0° for both COM (Figure 4a) and PHO (Figure 4b). Sharp peaks were observed in the diffractograms of blank SLNs, demonstrating a different crystalline structure of the lipids in the matrix with respect to bulk solid lipids. The disappearance of the principal peaks of MRN in the diffractograms of the drug-loaded SLNs could be due to the incorporation of MRN into the crystal lattice of the lipids [67] and/or to the formation of noncovalent MRN-lipid interactions already observed via DSC analysis.

Raman Spectroscopy
µ-Raman spectroscopy allowed us to characterize, in a solid state, the molecular interactions driving the MRN-SLNs' formation. In particular, the used approach, already successfully applied for similar systems [68,69], consisted of monitoring the spectral changes that the vibrational bands characteristic of the free MRN undergo when the drug is loaded in the SLNs (both based on COM and PHO), as a consequence of the activation of interac-tions involving specific functional groups. Figure 5 shows the µ-Raman spectrum of MRN together with its molecular structure, in the range between 500 and 1800 cm −1 , where the majority of the MRN vibrational features fall.

Raman Spectroscopy
μ-Raman spectroscopy allowed us to characterize, in a solid state, the molecular interactions driving the MRN-SLNs' formation. In particular, the used approach, already successfully applied for similar systems [68,69], consisted of monitoring the spectral changes that the vibrational bands characteristic of the free MRN undergo when the drug is loaded in the SLNs (both based on COM and PHO), as a consequence of the activation of interactions involving specific functional groups. Figure 5 shows the μ-Raman spectrum of MRN together with its molecular structure, in the range between 500 and 1800 cm −1 , where the majority of the MRN vibrational features fall. Based on a comparison with the literature [70], the high-intensity peaks at ~582 cm −1 and ~637 cm −1 can be ascribed to the C-C-C and C-C-O bending modes of the γ-pirone/benzene R3/R2 and benzene R2 rings, respectively. Furthermore, the peak at ~748 cm −1 results from the C3*-C4*-C5* bending mode of R2, while the contribution at ~874 cm −1 is attributable to the H3-C3*-C2* bending vibration. In addition, the contribution falling at ~1208 cm −1 is due to different C-O stretching vibrations of the flavonoid rings, particularly the C2*-O (in the benzene ring R2), C3-O (in the γ-pirone ring R3), and C5-O (in the benzene ring R1) ones. The bending vibrations of the C4*-OH and C2*-OH groups of the benzene ring R2 give rise, respectively, to the contributions at ~1239 cm −1 and ~1312 cm −1 , while the triplet of prominent bands falling within the 1500-1750 cm −1 range can be assigned to the C=O (~1578 cm −1 ) stretching vibrations of the benzene ring R2, to the O-C (~1622 cm −1 ) stretching vibrations of the benzene and γ-pirone rings (R1 and R3), and to the C=C (~1665 cm −1 ) stretching vibrations.
To obtain insights into the interactions involved in the formation of the MRN-SLNs, the regions extending from 1500 cm −1 to 1750 cm −1 and from 560 cm −1 to 800 cm −1 were focused on in detail. The choice of these spectral ranges was mainly due to the fact that they are free of vibrational contributions coming from SLNs, and hence their inspection allows one to directly monitor the variations exhibited by the vibrational bands of MRN when loaded within the SLNs. Figure 6 shows the experimental μ-Raman spectra, in the 1500-1750 cm −1 (a) and 560-800 cm −1 (b) ranges, respectively, for MRN, MRN-SLNs-PHO10, and MRN-SLNs-COM10 systems. Based on a comparison with the literature [70], the high-intensity peaks at~582 cm −1 and 637 cm −1 can be ascribed to the C-C-C and C-C-O bending modes of the γ-pirone/benzene R 3 /R 2 and benzene R 2 rings, respectively. Furthermore, the peak at~748 cm −1 results from the C3*-C4*-C5* bending mode of R 2 , while the contribution at~874 cm −1 is attributable to the H3-C3*-C2* bending vibration. In addition, the contribution falling at~1208 cm −1 is due to different C-O stretching vibrations of the flavonoid rings, particularly the C2*-O (in the benzene ring R 2 ), C3-O (in the γ-pirone ring R 3 ), and C5-O (in the benzene ring R 1 ) ones. The bending vibrations of the C4*-OH and C2*-OH groups of the benzene ring R 2 give rise, respectively, to the contributions at~1239 cm −1 and~1312 cm −1 , while the triplet of prominent bands falling within the 1500-1750 cm −1 range can be assigned to the C=O (~1578 cm −1 ) stretching vibrations of the benzene ring R 2 , to the O-C (~1622 cm −1 ) stretching vibrations of the benzene and γ-pirone rings (R 1 and R 3 ), and to the C=C (~1665 cm −1 ) stretching vibrations.
To obtain insights into the interactions involved in the formation of the MRN-SLNs, the regions extending from 1500 cm −1 to 1750 cm −1 and from 560 cm −1 to 800 cm −1 were focused on in detail. The choice of these spectral ranges was mainly due to the fact that they are free of vibrational contributions coming from SLNs, and hence their inspection allows one to directly monitor the variations exhibited by the vibrational bands of MRN when loaded within the SLNs. Figure 6 shows the experimental µ-Raman spectra, in the 1500-1750 cm −1 (a) and 560-800 cm −1 (b) ranges, respectively, for MRN, MRN-SLNs-PHO 10  From the inspection of Figure 6a, a flattening of the contributions centered at ~1578 cm −1 (C=O stretching mode), ~1622 cm −1 (O-C stretching mode), and ~1665 cm −1 (C=C stretching mode) passing from the pristine MRN to both MRN-SLNs-PHO10 and MRN-SLNs-COM10 can be clearly observed. Such an occurrence testifies the involvement of the corresponding functional groups in the drug-nanoparticle encapsulation process, and consequently the establishment of new molecular interactions. In particular, the observed spectral modifications indicate a hindering of the aforementioned vibrational modes due the more constrained molecular environment, resulting from interactions activated during NP formation.
In addition, the analysis of the lower frequency region extending from 560 to 800 cm −1 (Figure 6b) also revealed a flattening of the MRN vibration bands at ~ 582 cm−1 (C-C-C bending mode), ~637 cm −1 (C-C-O bending mode), and 748 cm −1 (C3*-C4*-C5* bending mode) as a consequence of the drug encapsulation into SLNs. Such an occurrence testifies, in agreement with the analysis of the 1500-1750 cm −1 spectral region, a hindering of the aforementioned vibrational modes upon complexation, due to the establishment of new intermolecular linkages between the drug and SLNs, thus representing good evidence of complex formation.

In Vitro MRN Release from SLNs
The release of MRN from SLNs was evaluated via the dialysis method, using a mixture of PBS (pH 7.4) and ethanol (15% v/v) as the dialysis medium. Ethanol was added to ensure MRN solubilization (MRN solubility in the dialysis medium, 1.48 mg/mL). As shown in Figure 7, free MRN quantitatively crossed the synthetic membrane within 5 h, demonstrating the suitability of the chosen dialysis bag in terms of molecular weight cutoff. The fast passage of free MRN through the synthetic membrane shows that the release of the encapsulated MRN is SLN-controlled. SLNs were retained in the bag and the drugs released from the delivery systems immediately crossed the synthetic membrane.
In Figure 7, we report the release profiles of the MRN from the two realized formulations. The most amount of encapsulated MRN was released from the SLNs within the first 24 h from the beginning of the experiment (about 18% and about 60%, w/w, was released from both of the formulations within 5 h and 24 h, respectively); the remaining amount of the drug was quantitatively released from both of the systems within 10 days. This trend could be related to the internal structure of MRN-SLNs, which presents a nonhomogeneous matrix, and the shell enriched with the drug [71]. In this way, due to the From the inspection of Figure 6a, a flattening of the contributions centered at~1578 cm −1 (C=O stretching mode),~1622 cm −1 (O-C stretching mode), and~1665 cm −1 (C=C stretching mode) passing from the pristine MRN to both MRN-SLNs-PHO 10 and MRN-SLNs-COM 10 can be clearly observed. Such an occurrence testifies the involvement of the corresponding functional groups in the drug-nanoparticle encapsulation process, and consequently the establishment of new molecular interactions. In particular, the observed spectral modifications indicate a hindering of the aforementioned vibrational modes due the more constrained molecular environment, resulting from interactions activated during NP formation.
In addition, the analysis of the lower frequency region extending from 560 to 800 cm −1 (Figure 6b) also revealed a flattening of the MRN vibration bands at~582 cm−1 (C-C-C bending mode),~637 cm −1 (C-C-O bending mode), and 748 cm −1 (C3*-C4*-C5* bending mode) as a consequence of the drug encapsulation into SLNs. Such an occurrence testifies, in agreement with the analysis of the 1500-1750 cm −1 spectral region, a hindering of the aforementioned vibrational modes upon complexation, due to the establishment of new intermolecular linkages between the drug and SLNs, thus representing good evidence of complex formation.

In Vitro MRN Release from SLNs
The release of MRN from SLNs was evaluated via the dialysis method, using a mixture of PBS (pH 7.4) and ethanol (15% v/v) as the dialysis medium. Ethanol was added to ensure MRN solubilization (MRN solubility in the dialysis medium, 1.48 mg/mL). As shown in Figure 7, free MRN quantitatively crossed the synthetic membrane within 5 h, demonstrating the suitability of the chosen dialysis bag in terms of molecular weight cutoff. The fast passage of free MRN through the synthetic membrane shows that the release of the encapsulated MRN is SLN-controlled. SLNs were retained in the bag and the drugs released from the delivery systems immediately crossed the synthetic membrane.
In Figure 7, we report the release profiles of the MRN from the two realized formulations. The most amount of encapsulated MRN was released from the SLNs within the first 24 h from the beginning of the experiment (about 18% and about 60%, w/w, was released from both of the formulations within 5 h and 24 h, respectively); the remaining amount of the drug was quantitatively released from both of the systems within 10 days. This trend could be related to the internal structure of MRN-SLNs, which presents a non-homogeneous matrix, and the shell enriched with the drug [71]. In this way, due to the limited distance, the drug present in the shell was quickly released in the medium. The prolonged subsequent release of the remaining fraction of MRN (about 40%, w/w) could be due to a homogeneous distribution within the matrix of this fraction and to the increased distance which MRN must run across to go out of the systems. Furthermore, the high affinity of the drug for the used lipids and the consequential existence of a drug-lipid association must be considered. To migrate through the lipid matrix, MRN-lipid dissociation is needed, with a consequent retarding of the diffusion process [72].
limited distance, the drug present in the shell was quickly released in the medium. The prolonged subsequent release of the remaining fraction of MRN (about 40%, w/w) could be due to a homogeneous distribution within the matrix of this fraction and to the increased distance which MRN must run across to go out of the systems. Furthermore, the high affinity of the drug for the used lipids and the consequential existence of a druglipid association must be considered. To migrate through the lipid matrix, MRN-lipid dissociation is needed, with a consequent retarding of the diffusion process [72].
Due to the ability of SNLs to interact with the mucus and quickly be adsorbed by the nasal mucosa [73], SLNs can avoid mucociliary clearance and translate into the brain via different routes [30]. In this way, the amount of MRN released within the first few hours could assure the initial therapeutical effect of MRN when SLNs reach the brain. The subsequent sustained release could be useful to prolong the therapeutical activity of MRN and avoid frequent administrations, increasing the patient's compliance. To investigate the MRN release mechanism from the SLNs, the release data were treated according to different kinetic models [74][75][76]. The best-fit model was achieved by comparing the regression coefficient (R 2 ) value of all of the models; in general, the closer the R 2 value was to 1, the better fit or relationship between the two factors ( Table 3). The first-order kinetic model shows higher R 2 values than the zero-order kinetic model and Higuchi model for both of the formulations. These values are nearer to 1; so, this model was considered to be a good fit [77]. As reported in [78], based on the composition and morphology of nanoparticles, a different mechanism of drug release could be expected, that is, drug diffusion, matrix degradation, or swelling. Because the in vitro release studies of MRN from SLNs were performed in the absence of enzymes, the degradation of the lipidic matrix cannot be considered as a mechanism for MRN release nor swelling due to the lipophilic nature of the matrix. It is conceivable that drug diffusion is the predominant mechanism that occurs in vitro for MRN release from SLNs. On the other hand, the release profiles reported in Figure 7 demonstrate an SLN structure characterized by the presence of a shell enriched with MRN. Therefore, first-order diffusion was expected, depending on the residue drug within the system, the diffusion distance inside the nanosystems, and the reversible drug-lipid association [72]. Due to the ability of SNLs to interact with the mucus and quickly be adsorbed by the nasal mucosa [73], SLNs can avoid mucociliary clearance and translate into the brain via different routes [30]. In this way, the amount of MRN released within the first few hours could assure the initial therapeutical effect of MRN when SLNs reach the brain. The subsequent sustained release could be useful to prolong the therapeutical activity of MRN and avoid frequent administrations, increasing the patient's compliance.
To investigate the MRN release mechanism from the SLNs, the release data were treated according to different kinetic models [74][75][76]. The best-fit model was achieved by comparing the regression coefficient (R 2 ) value of all of the models; in general, the closer the R 2 value was to 1, the better fit or relationship between the two factors ( Table 3). The first-order kinetic model shows higher R 2 values than the zero-order kinetic model and Higuchi model for both of the formulations. These values are nearer to 1; so, this model was considered to be a good fit [77]. As reported in [78], based on the composition and morphology of nanoparticles, a different mechanism of drug release could be expected, that is, drug diffusion, matrix degradation, or swelling. Because the in vitro release studies of MRN from SLNs were performed in the absence of enzymes, the degradation of the lipidic matrix cannot be considered as a mechanism for MRN release nor swelling due to the lipophilic nature of the matrix. It is conceivable that drug diffusion is the predominant mechanism that occurs in vitro for MRN release from SLNs. On the other hand, the release profiles reported in Figure 7 demonstrate an SLN structure characterized by the presence of a shell enriched with MRN. Therefore, first-order diffusion was expected, depending on the residue drug within the system, the diffusion distance inside the nanosystems, and the reversible drug-lipid association [72]. Table 3. Regression coefficient (R 2 ), rate constant (K i , i = 0 for zero-order, 1 for first order, and H for Higuchi model, respectively), and transport exponent (n) of Korsmeyer-Peppas model, as obtained from release data of MRN from SLNs.

Zero-Order
First-Order Higuchi Korsmeyer-Peppas According to the Korsmeyer-Peppas model, data obtained from in vitro drug release studies were plotted as log cumulative percentages of MRN released versus log time. The value of transport exponent (n) was used to predict the drug release mechanism. We obtained n in the range of 0.45 < n < 0.89 for both of the systems, evidencing that non-Fickian diffusion occurs as the predominant release mechanism for MRN [76].

Ex Vivo Permeation Experiments
To evaluate the ability of SLNs to improve the permeation of MRN through the biomembranes, an ex vivo permeation study was performed on excised bovine nasal mucosa mounted on Franz's cells. Studies are present in the literature that demonstrate the potentiality of this method to explore the mechanistic aspects of nasal transport and the metabolism of drugs, representing a valid ex vivo model for drug screening before their application in in vivo studies [79][80][81].
In Figure 8, we showed the permeation profiles obtained for free MRN and MRN encapsulated into SLNs over the experimental time (48 h). All of the samples were prepared in PBS (pH 7.4); so, not only were the formulations in suspension, but the free MRN was too, due to its low solubility in this solvent (55 µg/mL in PBS pH 7.4 [26]). For this reason, the ability of free MRN to cross mucosa is related to its physical-chemical properties, i.e., its intrinsic solubility, dissolution rate, and lipophilicity [82]. The biopharmaceutical classification system in fact described the aqueous solubility and membrane permeability of a drug as the main factors that limit drug absorption via the viable membrane [83]. As demonstrated by Li et al. [84], at pH 7.04, MRN has a Log D lower than 1 and a percentage of unionized form of about 1.21%, suggesting that the poor solubility and low permeability of MRN are the limiting steps to its absorption through the biomembranes [84].
These observations permitted us to exploit the permeation profile obtained for the free drug in our experiment. The permeation observed for free MRN in the first hour of the experiment (about 15%, w/w) could be related to the low amount of MRN in the solution which, due to its lipophilicity, was available to cross the mucosa. After that, the permeation proceeded very slowly and could reflect the low rate of MRN dissolution. To confirm this hypothesis, the accumulated MRN in the treated mucosa was quantified at the end of the experiment. A very low value was obtained of about 10% (w/w) of the assayed dose (Figure 9), demonstrating the scarce availability of the free drug to cross the biomembranes.
Different results were obtained for both of the formulations. They showed a progressive increase in the permeated MRN, and within 48 h, about 40% (w/w) and 50% (w/w) of MRN was detected in the receptor compartment for MRN-SLNs-COM 10 and MRN-SLNs-PHO 10 , respectively. This result could be related to the amount of the free drug released by the formulations and immediately available for permeation. In fact, higher in vitro drug release for MRN-SLNs-PHO 10 with respect to MRN-SLNs-COM 10 corresponded to a higher percentage of permeated MRN. Due to the lipidic composition of the realized delivery systems, an intimate contact with the mucosa could not be excluded, also improved by the presence of surfactant on the particle surfaces [85], resulting in increased permeation of encapsulated MRN with respect to the free drug. Furthermore, based on the little sizes of our formulations (<200 nm), the internalization of the MRN-SLNs into the mucosa could not be excluded [86]. In this way, MRN-SLNs accumulated within the nasal mucosa could represent a deposit able to produce the prolonged release of the drug. This last hypothesis could be confirmed by the MRN accumulated in the mucosa after 48 h treatment with the two formulations; it was higher with respect to that observed for the free drug (about 20% and 35%, w/w for MRN-SLNs-COM 10 and MRN-SLNs-PHO 10 , respectively) ( Figure 9). On the other hand, various studies [85][86][87] have demonstrated SLNs as being an optimum carrier for increasing the biodistribution of drugs into the brain via nasal administration, due to their ability to protect drugs from chemical and enzymatic degradation [88], to improve permeation through nasal mucosa due to their lipids and surfactant composition, and to promote active or passive targeting due to their small sizes and superficial charge [86]. sizes of our formulations (<200 nm), the internalization of the MRN-SLNs into the mucosa could not be excluded [86]. In this way, MRN-SLNs accumulated within the nasal mucosa could represent a deposit able to produce the prolonged release of the drug. This last hypothesis could be confirmed by the MRN accumulated in the mucosa after 48 h treatment with the two formulations; it was higher with respect to that observed for the free drug (about 20% and 35%, w/w for MRN-SLNs-COM10 and MRN-SLNs-PHO10, respectively) ( Figure 9). On the other hand, various studies [85][86][87] have demonstrated SLNs as being an optimum carrier for increasing the biodistribution of drugs into the brain via nasal administration, due to their ability to protect drugs from chemical and enzymatic degradation [88], to improve permeation through nasal mucosa due to their lipids and surfactant composition, and to promote active or passive targeting due to their small sizes and superficial charge [86].    sizes of our formulations (<200 nm), the internalization of the MRN-SLNs into the mucosa could not be excluded [86]. In this way, MRN-SLNs accumulated within the nasal mucosa could represent a deposit able to produce the prolonged release of the drug. This last hypothesis could be confirmed by the MRN accumulated in the mucosa after 48 h treatment with the two formulations; it was higher with respect to that observed for the free drug (about 20% and 35%, w/w for MRN-SLNs-COM10 and MRN-SLNs-PHO10, respectively) ( Figure 9). On the other hand, various studies [85][86][87] have demonstrated SLNs as being an optimum carrier for increasing the biodistribution of drugs into the brain via nasal administration, due to their ability to protect drugs from chemical and enzymatic degradation [88], to improve permeation through nasal mucosa due to their lipids and surfactant composition, and to promote active or passive targeting due to their small sizes and superficial charge [86].

Conclusions
In this study, we performed a preformulation study to optimize SLN formulations intended for the nose-to-brain targeting of MRN. We used two different lipids, that is, PHO and COM, obtaining similar results concerning sizes, zeta potential, encapsulation efficiency, and drug loading. The small sizes and spherical morphology of the realized SLNs makes them suitable for nasal administration. Both of the formulations showed sustained drug release and prolonged permeation through the excised bovine nasal mucosa, compared to the pure drug. Furthermore, the MRN-SLNs allowed for a greater accumulation of the drug inside the mucosa with respect to the free MRN, probably due to the internalization of the SLNs into the mucosa.
Although biological in vitro and in vivo studies are needed, based on our data, we believe that MRN-SLNs are promising systems for the potential treatment of neurodegenerative diseases based on oxidative injury.