*3.1. Characterization of Lipid Vesicles*

During the lipid vesicle design stage, different factors were considered that could potentially influence their properties, based on previous knowledge about this type of nanocarriers. In this sense, two essential factors are the composition and proportion of components, since they can affect important properties, such as the size, stability, or releasing ability of the drug. It is well known that by increasing the proportion of cholesterol located in the lipid bilayers, the particle size [68,69] and the vesicle rigidity also increase [69,70]. Considering that one of the main reasons why lipid vesicles improve drug absorption through the skin is their ability to deform and penetrate between the cells of the stratum corneum, we sought to produce flexible vesicles of the smallest possible size [71,72]. For this, a low ratio of cholesterol to phospholipid (molar ratio 1:17) was chosen for conventional liposomes, since it provides small and ultraflexible vesicles with enough stability. Besides, 85:15% *w/w* lipid:surfactant ratio was used since Ahad et al. demonstrated that it is the most suitable for transdermal delivery of eprosartan mesylate in transferosomes based on P90G:Tween 80 [43]. Additionally, different proportions of ethanol have been studied for preparing ethosomes (20–50% *w/w*). In this case, intermediate concentration (30% *w/w*) was used, considering that it is the most promising ratio for transdermal absorption purposes, according to the characterization results [73].

The initial results (day 0) of size, PDI, zeta potential, drug loading, and phospholipid content are reported in Table 2.



In accordance with other works, size measurements were in the expected range for these types of lipid vesicles. Wu et al. obtained conventional liposomes in a range of 236–374 nm, depending on the percentage of cholesterol included in the formulation [74]. The transferosomal and ethosomal particle dimensions obtained were smaller than those of conventional liposomes, because the presence of the edge-activating substances in the lipid bilayer of the vesicles causes a reduction in the surface tension [75]. Transferosome sizes obtained were similar to the values previously reported by Carreras et al. [75], and ethosomal batches reproduced at exactly the size described by Touitou et al., the original research for this vesicle type (30% *w/w* ethanol content) [44].

PDI values as a measure of the variability in size of the particle populations were always below 0.3, which is the limit value to consider homogeneous populations for lipid-based carriers in drug delivery applications [76,77]. bilization of the vesicles with such strategy. T1d and T2d presented an entrapment value within the expected range. At this point, the differences between both transferosomes and L2 could be attributed either to the differences in size, which allowed the incorporation of

we explored the option of formulating B12 in the organic phase to check any possible difference. In our case, the L2 formulation showed the maximum efficiency in the entrapment process compared to our other prototypes, being significantly different from the other formulations. The comparison between L2 and L1 analyzes the effect of including the B12 in every phase, and the result was higher when the addition was in the lipid one. We also tested the possible influence of differences in the initial dose by preparing L3 formulation. The poor entrapment efficiency in L3 formulation seems to indicate a desta-

*Pharmaceutics* **2021**, *13*, x FOR PEER REVIEW 9 of 23

Zeta-potential values of all formulations were negative, as expected, due to the negative charge of the phospholipids. Liposomes showed the strongest charge, probably due to the bigger size and higher phospholipid amount. It has also been reported that the surface charge decreases when increasing the level of cholesterol in a phospholipid membrane [78]. Consequently, our prototypes presented lower values in comparison with other reports where lipid vesicles contained cholesterol in higher proportion [75]. Regarding the transferosomes and ethosomes, they presented a size reduction (probably because of an edge-activator), which implies a lower concentration in phospholipids, thus reducing the negative charge. These results are consistent with the ranges described by Ahad et al. and Touitou et al., who reported values from −5.91 to −14.0 and 4.6 to −4.3 mV for the same type of transferosomes and ethosomes, respectively [43,44]. a higher amount of B12 in the vesicle core, or to the purification method used to remove the non-entrapped drug. While liposomal formulations were centrifuged, T1d and T2d were dialyzed over 24 h [47]. This latter reason is also checked through the EE% and PC% results of T1c and T2c. Both values were lower than expected, probably because of the unsuitability of the centrifugation method for transferosomes. In fact, the supernatant obtained after centrifugation was not completely clear, as it was in the conventional liposomes, and contained still an important amount of liposomes, as demonstrated by a poor phospholipid recovery. The specific encapsulation efficiency parameter was calculated by dividing EE% by PC% [82] (Figure 2b). We could assume that the centrifugation method

Following the work of Jain et al., we used the vesicle size reduction rate and volume loss after cold extrusion as an indirect measurement of the vesicles deformability capacity [54]. Results are presented in Figure 1a,b. Significant differences were obtained between liposomes and ultraflexible vesicles, as expected [79,80]. Liposomes were retained in the 100 nm filters and forced to split into smaller particles to pass the pores. On the contrary, transferosomes and ethosomes—whose initial size was higher than 100 nm—maintained their size during the passage thanks to their flexibility. Likewise, the final volume collected after cold extrusion was inversely related to the vesicle flexibility. As such, no differences were observed in transferosomal or ethosomal batches, while all liposomal prototypes presented a significant reduction compared to their initial volumes. Thus, both techniques confirm the flexibility of transferosomes and ethosomes. is adequate for liposomes, as this method is usually successfully used in our laboratories, and the expected amount of phospholipid was recovered after the process. Based on our results, we could conclude that centrifugation was only suitable for liposomal formulations. Ethosomes were not dialyzed because the dialysis process could extract the ethanol from the vesicles [75]. The recovered PC% in ethosomes was around 60% of the initial amount, and the EE% was reasonable for a hydrophilic molecule. In the other samples, phosphatidylcholine was not incorporated in a complete manner, presumably due to material loss during the manufacturing and extrusion processes (Table 2).

**Figure 1.** *Cont.*

**Figure 1.** (**a**) Size change after extrusion through 100 nm membrane; (**b**) Volume loss after extrusion through 100 nm membrane (9 passages). All results are expressed as mean ± SD (*n* = 3). \* means statistically significant differences (*p* < 0.05), n.s. means no statistically significant differences (*p* > **Figure 1.** (**a**) Size change after extrusion through 100 nm membrane; (**b**) Volume loss after extrusion through 100 nm membrane (9 passages). All results are expressed as mean ± SD (*n* = 3). \* means statistically significant differences (*p* < 0.05), n.s. means no statistically significant differences (*p* > 0.05) using one-way ANOVA followed by Tukey's multiple comparison test.

0.05) using one-way ANOVA followed by Tukey's multiple comparison test. However, it should be considered that the objective of this project was to obtain flexible lipid vesicles with the highest possible B12 dose, as saturated systems provide highest skin permeability rates [83]. Figure 2c represents the final amount of B12 per 10 mL of liposomal suspension. Here, we observed that, regardless the encapsulation efficiency values obtained, the highest B12 load was obtained when B12 was introduced in the aqueous phase in a saturated PBS solution (L1 and T1d) (Figure 2c). The higher water volume entrapped by the liposome core contains higher total amounts of B12 compared to the incorporation of the drug in the thin film layer. The entrapment efficiency percentage and total amount of encapsulated B12 are also graphically presented in Figure 2a,c. Hydrophilic compounds often show lower entrapment rates than lipophilic ones. This happens because the entrapment is more efficient if the compound is retained in the phospholipid bilayers, which depends on its primary affinity [48,81]. Arsalan et al. reported a maximum entrapment efficacy of 40% in B12 liposomes when maximizing the amount of lipid included during the vesicle formulation [40]. This result is clearly superior to our L1 data (Figure 2a). To overcome this issue and considering the ability of B12 to solve in aqueous and organic solvents such as methanol, we explored the option of formulating B12 in the organic phase to check any possible difference. In our case, the L2 formulation showed the maximum efficiency in the entrapment process compared to our other prototypes, being significantly different from the other formulations. The comparison between L2 and L1 analyzes the effect of including the B12 in every phase, and the result was higher when the addition was in the lipid one. We also tested the possible influence of differences in the initial dose by preparing L3 formulation. The poor entrapment efficiency in L3 formulation seems to indicate a destabilization of the vesicles with such strategy. T1d and T2d presented an entrapment value within the expected range. At this point, the differences between both transferosomes and L2 could be attributed either to the differences in size, which allowed the incorporation of a higher amount of B12 in the vesicle core, or to the purification method used to remove the non-entrapped drug. While liposomal formulations were centrifuged, T1d and T2d were dialyzed over 24 h [47]. This latter reason is also checked through the EE% and PC% results of T1c and T2c. Both values were lower than expected, probably because of the unsuitability of the centrifugation method for transferosomes. In fact, the supernatant obtained after centrifugation was not completely clear, as it was in the conventional liposomes, and contained still an important amount of liposomes, as demonstrated by a poor phospholipid recovery. The specific encapsulation efficiency parameter was calculated by dividing EE% by PC% [82] (Figure 2b). We could assume that the centrifugation method is adequate for liposomes, as this method is usually successfully used in our laboratories, and the expected amount of phospholipid was recovered after the process. Based on our results, we could conclude that centrifugation was only suitable for liposomal formulations. Ethosomes were not dialyzed because the dialysis process could extract the ethanol from the vesicles [75]. The recovered PC% in ethosomes was around 60% of the initial amount, and the EE% was reasonable for a hydrophilic molecule. In the other

87

samples, phosphatidylcholine was not incorporated in a complete manner, presumably due to material loss during the manufacturing and extrusion processes (Table 2). phase in a saturated PBS solution (L1 and T1d) (Figure 2c). The higher water volume entrapped by the liposome core contains higher total amounts of B12 compared to the incorporation of the drug in the thin film layer.

However, it should be considered that the objective of this project was to obtain flexible lipid vesicles with the highest possible B12 dose, as saturated systems provide highest skin permeability rates [83]. Figure 2c represents the final amount of B12 per 10 mL of liposomal suspension. Here, we observed that, regardless the encapsulation efficiency values obtained, the highest B12 load was obtained when B12 was introduced in the aqueous

**Figure 1.** (**a**) Size change after extrusion through 100 nm membrane; (**b**) Volume loss after extrusion through 100 nm membrane (9 passages). All results are expressed as mean ± SD (*n* = 3). \* means statistically significant differences (*p* < 0.05), n.s. means no statistically significant differences (*p* >

0.05) using one-way ANOVA followed by Tukey's multiple comparison test.

*Pharmaceutics* **2021**, *13*, x FOR PEER REVIEW 10 of 23

**Figure 2.** (**a**) Entrapment efficiency (EE) (expressed as a percentage) of the nine prototypes of liposomes, transferosomes, and ethosomes; (**b**) Specific entrapment efficiency rate of the nine prototypes of liposomes, transferosomes, and ethosomes; (**c**) Total amount of B12 encapsulated in 10 mL of liposomal suspension (μg). All results are expressed as mean ± SD (*n* = 3). \* means statistically sig-**Figure 2.** (**a**) Entrapment efficiency (EE) (expressed as a percentage) of the nine prototypes of liposomes, transferosomes, and ethosomes; (**b**) Specific entrapment efficiency rate of the nine prototypes of liposomes, transferosomes, and ethosomes; (**c**) Total amount of B12 encapsulated in 10 mL of liposomal suspension (µg). All results are expressed as mean ± SD (*n* = 3). \* means statistically significant differences (*p* < 0.05), n.s. means no statistically significant differences (*p* > 0.05) using one-way ANOVA followed by Tukey's multiple comparison test.

way ANOVA followed by Tukey's multiple comparison test. The short-term stability behavior of the formulations was studied by the Turbiscan Stability Index However, it should be considered that the objective of this project was to obtain flexible lipid vesicles with the highest possible B12 dose, as saturated systems provide

nificant differences (*p* < 0.05), n.s. means no statistically significant differences (*p* > 0.05) using one-

most stable formulation since no changes in light transmission and backscattering lines were reported over 24 h (Figure 3c,d). A flocculation phenomenon was observed for the ethosomal samples through the backscattering graph (Figure 3f). Its evolution over the whole height of the sample proves a global increase of the particles size (Figure 3e,f). Flocculates could be easily redispersed by gentle shaking. Other aggregative processes, such as coalescence, were discarded, given that no significant long-term size changes were observed after periodically redispersing the samples. On the contrary, liposomal vesicles experienced first a sedimentation process, as the backscattering increases at the bottom of the sample, due to an increase of the concentration in the dispersed phase (sediment) and a decrease at the top of the sample, due to a reduction of the concentration (clarified layer) (Figure 3a,b). Additionally, the smallest particles that remained at the clarified phase showed a tiny flocculation at the end of the investigated times (yellow-red frame). Transmission and backscattering profiles in the work of Cristiano et al. were similar and showed comparable phenomena of flocculation for ethosomes and stability for transferosomes [85]. Therefore, the global destabilization kinetics (Figure 4) confirmed the stability indexes about B12 vesicles (liposome < ethosome

< transferosome).

highest skin permeability rates [83]. Figure 2c represents the final amount of B12 per 10 mL of liposomal suspension. Here, we observed that, regardless the encapsulation efficiency values obtained, the highest B12 load was obtained when B12 was introduced in the aqueous phase in a saturated PBS solution (L1 and T1d) (Figure 2c). The higher water volume entrapped by the liposome core contains higher total amounts of B12 compared to the incorporation of the drug in the thin film layer.

The short-term stability behavior of the formulations was studied by the Turbiscan Stability Index (TSI), a parameter offered by TurbiscanTM LAB Stability Analyzer device. TSI allows the comparison of samples that present different progression phenomena [84]. Transferosomes were found to be the most stable formulation since no changes in light transmission and backscattering lines were reported over 24 h (Figure 3c,d). A flocculation phenomenon was observed for the ethosomal samples through the backscattering graph (Figure 3f). Its evolution over the whole height of the sample proves a global increase of the particles size (Figure 3e,f). Flocculates could be easily redispersed by gentle shaking. Other aggregative processes, such as coalescence, were discarded, given that no significant long-term size changes were observed after periodically redispersing the samples. On the contrary, liposomal vesicles experienced first a sedimentation process, as the backscattering increases at the bottom of the sample, due to an increase of the concentration in the dispersed phase (sediment) and a decrease at the top of the sample, due to a reduction of the concentration (clarified layer) (Figure 3a,b). Additionally, the smallest particles that remained at the clarified phase showed a tiny flocculation at the end of the investigated times (yellow-red frame). Transmission and backscattering profiles in the work of Cristiano et al. were similar and showed comparable phenomena of flocculation for ethosomes and stability for transferosomes [85]. Therefore, the global destabilization kinetics (Figure 4) confirmed the stability indexes about B12 vesicles (liposome < ethosome < transferosome).
