*3.1. Screening of Experimental Variables on Nanoparticle Properties*

The pharmaceutical interest of cellulose derivatives has increased in recent years, due to economic factors (usually due to the low-cost of the excipients, compared with synthetic compounds) and as they originate from renewable resources [24]. EC was selected as the LPNC polymer, as it has been approved by the Food and Drug Administration (FDA) for medical and food applications (i.e., vitamin and mineral tablets), and it has shown good biocompatibility in ocular drug delivery systems [25–27]. Furthermore, EC is cheaper than other commonly used polymers, such as PLGA or polycaprolactone. EC is commonly used in the pharmaceutical industry as a coating material for different drug delivery applications, due to its controlled release properties caused by the porous structure, where the drugs can be trapped and diffuse [24].

The choice of organic phase is a parameter to consider, in terms of toxicity. Chloroform and dichloromethane are organic solvents that are commonly used in the Emulsion Solvent Evaporation method; however, due to their toxicity, it is advisable to avoid them when possible. In this case, ethyl acetate and ethanol were selected, as they have been classified as class III (or low toxic potential) solvents by the European Medicines Agency (EMA) [28]. In the Emulsion Solvent Evaporation method, the viscosity of the organic phase influences the diffusion to the aqueous phase. An organic phase that is too viscous will not diffuse adequately, forming large aggregates that cannot be stabilized in the final formulation. This may be attributed to the increase in viscosity of the dispersed phase, resulting in a reduction of the net shear stress and prompting bigger nanodroplets. The ratio between the continuous and dispersed phases is also an important factor: In the solvent evaporation emulsion method, a high proportion of dispersed phase causes an increase in the size of the nanoparticles, due to the increase in droplet size. Furthermore, due to the hydrophobic nature of DEX, the burst effect would increase, and the encapsulation efficiency would decrease. To avoid this problem, dilutions of EC were made in different binary mixtures of ethanol and ethyl acetate and the viscosities of the different organic phases were characterized. The LPNCs were designed with MCT as the oil core, to increase the %EE and reduce the DEX burst release effect. Benzalkonium chloride was incorporated as a preservative.

Table 3 shows the viscosity results for the different binary mixtures of ethyl acetate (EA) and ethanol (ET) with EC at 2% *w*/*w*. The 1:1 and 5:1 ratios were selected as optimal, as they showed the lowest viscosity values. This means a decrease in the viscous forces resisting droplet breakdown and, thus, smaller oil droplets were formed, resulting in a decreased particle size. When LPNCs were produced at the 1:1 ratio, aggregates were observed; so, finally the 5:1 EA:ET ratio was selected as definitive, as it did not show aggregates and obtained low viscosity. To increase the loading capacity of the LPNCs, the EC concentration was increased to the maximum concentration possible to obtain LPNCs with adequate properties, in term of size and PDI [29]. The final EC concentration was fixed at 2.33% *w*/*w* in the organic phase (0.35% *w*/*w* in the final formulation) and the MCT concentration was set at 0.2% *w*/*w*, since at higher concentrations aggregates began to appear.

The stability of LPNCs is directly related to their size, PDI, and Z-potential. Surfactants have an important role in these characteristics and, so, the effect of surfactant concentration on the physicochemical properties of LPNCs was studied. As in any colloidal suspension, lipomers can undergo destabilization processes, such as Ostwald ripening. This process can be avoided by steric stabilization using Tween 80 [30]. Due to its HLB (14.9), Tween 80 acts

as a stabilizer of the aqueous phase, coating the surfaces of the LPNCs and preventing their aggregation. Due to the more lipophilic nature of Span 60 (HLB 4.7), it becomes dispersed in the LPNC core when MCT is used [31]. In this case, the percentages of Tween 80 and Span 60 were modified between 1.5% and 2.5% *w*/*w* and 0.16% and 0.32% *w*/*w*, respectively. As it can be seen from Table 4, four batches were produced and the Z-average, PDI, Z-potential, and *%EE* were characterized for each of them.


**Table 3.** Organic phase ethyl cellulose (EC) mixture viscosities.

**Table 4.** Influence of Span 60 and Tween 80 on the physicochemical parameters.


A general linear model regression using the Minitab software was performed (with a significance level of α = 0.05) to evaluate the effect of the tested experimental variables. The levels of Tween 80 and Span 60 were not significant for the hydrodynamic diameter, the PDI, or for the *%EE*, but they were significant for the Z-potential (*p* = 0.006 and *p* = 0.017, respectively). In Figure 1, the effects of these surfactants on the surface charge of the LPNCs can be observed. By increasing the amount of surfactant, the positive charge was reduced, as the hydroxyl groups of Tween 80 and Span 60 shielded the positive charge of the nanoparticles.

Thus, formulation LP03 was selected as the final formulation, with lower levels of both surfactants (1.5% *w*/*w* of Tween 80 and 0.16% *w*/*w* of Span 60). The %LC of batch LP03 was calculated following Equation (2), which resulted in 30.22%, thus confirming the high loading capacity of this type of nanoencapsulation system [5] compared to other drug delivery systems like liposomes, in which Amin et al. reached an LC of 12.6% [16].

Table 5 shows the residual EA and ET found after the evaporation process. EA has a vapor pressure at 25 ◦C of 1.23 atm, which is higher than the vapor pressure of ethanol at the same temperature (0.08 atm). Due to the higher volatility of EA than ET, after 5 min of rotary evaporation at 40 ◦C, the residual quantity was already below the limit of quantification (LOQ). Regarding ethanol, with the mildest evaporation conditions, more residues were found, but they were well below the limit established for residual solvents of class 3 (5000 ppm) in the guideline for residual solvents (ICH Q3C) of the European Medicines Agency. Thus, the evaporation conditions were fixed at 5 min and 40 ◦C.

**Table 5.** Residual solvents after evaporation at 40 ◦C.


**Figure 1.** Plots of the main effects of Tween 80 and Span 60 on the Z-potential of the lipid core polymeric nanocapsules (LPNCs). **Figure 1.** Plots of the main effects of Tween 80 and Span 60 on the Z-potential of the lipid core polymeric nanocapsules (LPNCs).

#### Thus, formulation LP03 was selected as the final formulation, with lower levels of *3.2. Physicochemical Characterization of the LPNCs*

both surfactants (1.5% *w/w* of Tween 80 and 0.16% *w/w* of Span 60). The %LC of batch LP03 was calculated following Equation (2), which resulted in 30.22%, thus confirming the high loading capacity of this type of nanoencapsulation system [5] compared to other drug delivery systems like liposomes, in which Amin et al. reached an LC of 12.6% [16]. Table 5 shows the residual EA and ET found after the evaporation process. EA has a Thanks to the high resolution of the TEM images (Figure 2), it is possible to appreciate the lipid nucleus (brighter area) and distinguish it from the polymeric framework (darker area), as can be observed in Figure 2C. Furthermore, the spherical morphology was confirmed. A total of 180 particles were analyzed and a mean size of 126.40 ± 34.93 nm was obtained. *Pharmaceutics* **2021**, *13*, x 10 of 21

vapor pressure at 25 °C of 1.23 atm, which is higher than the vapor pressure of ethanol at the

**Figure 2.** Transmission Electron Microscopy (TEM) images of negative-stained lipid–polymer hybrid LPNCs with 2% uranyl acetate at (**a**) ×250,000 magnification; and (**b**) ×300,000 magnification. (**c**) Individual nanoparticle representation zoom. **Figure 2.** Transmission Electron Microscopy (TEM) images of negative-stained lipid–polymer hybrid LPNCs with 2% uranyl acetate at (**a**) 250,000× magnification; and (**b**) 300,000× magnification. (**c**) Individual nanoparticle representation zoom.

**Figure 3.** Histogram of lipomers from the measurement of the diameter of the particles (*n* = 180)

310

340

370

The release data of the free DEX and LPNCs can be seen in Figure 4. After 24 h, the DEX formulated in an ethanolic solution (FREE-DEX) was released at 100%, while the release of DEX loaded in LPNCs (LPNCs-DEX) was more sustained, due to diffusion through the polymeric framework and the affinity of the active for the lipid cores of the

The diameter distribution of the particles, as measured by TEM, is shown in Figure 3

*3.3. In Vitro Release Test of DEX-Loaded LPNCs and Free DEX* 

using ImageJ software.

10

40

70

100

130

160

190

**Diameter (nm)**

220

250

280

**Frequency**

particles.

The diameter distribution of the particles, as measured by TEM, is shown in Figure 3 as a histogram. The average obtained was similar using both TEM and DLS techniques and the %BIAS error for the diameter was 10.78%, which is low. The diameter distribution of the particles, as measured by TEM, is shown in Figure 3 as a histogram. The average obtained was similar using both TEM and DLS techniques and the %BIAS error for the diameter was 10.78%, which is low.

**Figure 2.** Transmission Electron Microscopy (TEM) images of negative-stained lipid–polymer hybrid LPNCs with 2% uranyl acetate at (**a**) ×250,000 magnification; and (**b**) ×300,000 magnification.

using ImageJ software.

(**c**) Individual nanoparticle representation zoom.

*Pharmaceutics* **2021**, *13*, x 10 of 21

**Figure 3.** Histogram of lipomers from the measurement of the diameter of the particles (*n* = 180) **Figure 3.** Histogram of lipomers from the measurement of the diameter of the particles (*n* = 180) using ImageJ software.
