Whey Proteins as a Potential Co-Surfactant with Aesculus hippocastanum L. as a Stabilizer in Nanoemulsions Derived from Hempseed Oil

2.2. Droplet Size Studies

Droplet size and PDI (polydisperisty index) are key factors affecting emulsion stability. Based on the experimental design used for preliminary studies (Table 3), different concentrations of HSO, AHE and whey protein were used to find the optimal droplet size distribution of nanoemulsion formulations. According to our previous studies [9,32], we decided to eliminate runs S06, S12 and S15 from the experimental design model (Table 1) for the evaluation of the droplet size. Analysis of the results presented in Table 3 for the center points of the Box–Behnken design were performed on samples S01, S05, S06, S12, S15, and similar results for Z-ave, PDI, average peak maximum by intensity were recorded for the S01 and S05 samples (see the standard deviation range of three replications).

In contrast to previous studies [33,34], we determined the droplet size of the emulsion systems as delivered after preparing the emulsion without additional dilution since dilution may affect the emulsion stability. Furthermore, based on the principles of the dynamic light scattering (DLS) method used herein and apparatus limitations, we chose to round droplet size numbers up to a full number. Moreover, based on previous experience from studies with different high-PDI systems [35], we considered droplet/particle size by intensity and number. Detailed studies of droplet size distributions clearly showed that even a small fraction of larger particles/droplets or accidental contaminations impacted the Z-ave and size distribution. Results presented in Table 2 exhibit that the major percentage of droplets were much smaller than average size distribution by intensity peak maximum. For example, the average maximum of the peak by number for sample S08 was 23 nm; however, its light scattering intensity was 318 nm.

According to the regression analysis and the ANOVA, we were able to predict Z-Ave for samples using a quadratic model (p < 0.05). The relative significance in order of individual factors according to the greatest effect was determined to be as follows: HSO (%) > whey (%) > AHE. The interaction of HSO and whey was a significant factor affecting Z-ave of samples (p < 0.05). The significant interaction effect of HSO and a poloxamer used as a surfactant to produce an HSO-based nanoemulsion was previously reported [36]. The ANOVA and the regression analysis also showed that the PDI of nanoemulsion samples could be fit to a linear model with HSO (%) as the only significant factor (p < 0.05). To find an optimal and stable formulation and further possible applications of the emulsion system, other characterizations such as droplet size distribution, visual, centrifugation and microscopic properties of the emulsion need to be considered in addition to Z-ave and the PDI. Additional stability studies, such as identification of destabilization mechanisms including creaming, sedimentation, flocculation, coalescence can also confirm the stability potential of an emulsion system. Based on the preliminary stability tests as well as droplet size distribution, we selected four samples for further analysis (Figure 1B).

Figure 2 presents detailed analysis of droplet size distribution from four selected samples. All the selected samples contained a whey concentration of 4, except for S14, which had a whey concentration of 2.5%. According to the DLS results of S14 (Figure 2D), the scattered light intensity and droplet size by number were not different. This result was confirmed by a low PDI value (0.190). Microscopic observation of sample S14 (Figure 2D) also confirmed homogeneity of the samples. In addition, in comparison with other samples, single droplets were easy to distinguish at low magnification (e.g., ×40).

Based on our previous observations for the HSO emulsion system stabilized by the AHE [12], we decided to evaluate emulsion droplet distribution and sample homogeneity using light microscopy with an inverted microscope. To prevent fast evaporation of samples, the emulsions were inserted in a corvette equipped with slide channels. The images presented in Figure 3 were taken a few minutes after the emulsion injection. This method prevents the flow of injected liquid artefacts. For unstable systems, a short break in time enabled the observed coalescence of the droplets. With lower magnification, high homogeneity of samples S02, S08 and S13 was observed (Figure 3A–C). Only a small fraction of larger droplets could be distinguished. This observation corresponded with the DLS results in case of droplet size distribution by intensity, where a more intensive signal was detected for the samples with larger hydrodynamic diameters (Figure 2A–C). The presence of the smaller droplets fraction determined by DLS was probably caused by two-step processing using ultrasound treatment for preparing this emulsion. Some of the studies suggested that ultrasound treatment as well as ultrahigh-pressure homogenization caused an increase in the α-helix structure content of whey protein and a decrease in the β-sheet component of whey protein [37,38].

Microscopic imaging for nanoemulsion systems with proteins presented by Zhu et al. [39] corresponded to our findings. In their studies, optical microscopic imaging was compared with confocal laser scanning microscopy (CLSM). It was suggested that using CLSM allowed for the determination of a core–shell structure in the emulsion systems where proteins were located at the surface of the emulsion droplets. More detailed studies are needed for our nanoemulsion systems. First of all, signals from all the fluorescent structures should be identified and separated (i.e., from HSO, AHE) [12,31]. Then, additional studies on the impact of whey protein on the fluorescent behavior of the components should be investigated. Ren and Giusti [40] showed that anthocyanin-rich extracts decreased the fluorescence intensity of whey protein while increasing λ_max. The study concluded that thermally induced whey protein was effective in protecting anthocyanin from color degradation. Using an optical microscope, we focused only on the verification of the homogeneity of the samples as well as possible coalescence (which was not observed in the optimal nanoemulsions). Nevertheless, we strongly recommend using more than one technique for the analysis of the emulsion systems droplets size.

2.3. Rheological Tests

Figure 4a shows the flow curves of the various samples, which illustrate the samples’ rheological properties. The flow curves showed a non-Newtonian behavior, with a decreasing slope (viscosity) up to a certain cut-off, suggesting typical pseudoplastic behavior. However, the viscosity increased after a certain shear rate, which is a typical property of a dilatant fluid. The changes in the dynamic viscosity of the samples with increasing shear rate are shown in Figure 4b. It was apparent that the S08 sample had the highest viscosity, followed by S13 and S02, with S14 being least viscous at various shear rates. All the samples exhibited two types of behavior: first, shear thinning, and then shear thickening, after a particular shear rate cut-off. This became more apparent in Figure 4c when the power law was used to model these curves to evaluate the coefficient that depicted clearly the two zones, with a cut-off close to 130 Hz. As a result, a broken power law model (Equation (1)) was used to describe the flow behavior of the samples.

where η is the dynamic viscosity, K₁ and K₂ are the consistency coefficients,

is the shear rate and n₁ and n₂ are the flow behavior indices.

Table 4 shows the broken power law model parameters. The consistency coefficient K₁ at shear rates less than 130 Hz was found to be highest for S08, while those of other samples were not significantly different. This measure is an indicator of the initial system viscosity [41], suggesting that sample S08 had higher viscosity to begin with, which was sustained even at changing shear rates.

The higher viscosity of S08 as compared to the other samples could be attributed to the highest HSO (5%) content amongst the four tested samples. Furthermore, S14 was found to possess the lowest viscosity on account of having the lowest HSO content (1%). This showed that a higher oil loading was associated with higher viscosity, which is consistent with observations of other researchers [42,43,44]. Rha [45] and Jarzebski et al. [12] attributed this phenomenon of increasing viscosity with increasing oil loading to the greater formation of interphase layers, creating a larger barrier between the emulsion components.

At a shear rate of 130 Hz, all the samples exhibited a transition from shear thinning behavior to shear thickening behavior. Again, samples S08, having the highest oil loading, and sample S14, with the lowest oil and whey loading, demonstrated the highest and the lowest viscosity, respectively. This transition from shear thinning to the shear thickening behavior could be attributed to the phenomenon that at very high shear rates, tremendous turbulence occurs. When such turbulence occurs after a certain shear rate cut-off (which was 130 Hz in our case), any increase in the shear rate will result in increased turbulence, resulting in increased viscous dissipation and higher resistance to flow, which in turn makes the flow appear as shear-thickening. This effect is a typical example of the Taylor vortex flow in shear-thinning fluids [46,47]. The interpretation of Chhabra and Richardson [48] could be used to explain such transitionary behavior of our shear-thinning samples. At rest, the emulsion has sufficient interfacial tension to be stable. At low shear rates, lubrication for particle motion of the continuous phase between the plates is provided by the dispersed oil phase resulting in decreased stress with increasing shear (shear-thinning behavior). However, at high shear rates, the emulsion breaks, and the dispersed phase is completely separated from the continuous phase due to centrifugal forces. Furthermore, continuous and dispersed phases expand or dilate slightly under increasing shear strain, resulting in increased friction and shear stress, causing the dynamic viscosity to increase rapidly with shear rate.

2.4. FTIR

The FTIR spectra were recorded with air or water as the background. Both series of results are presented in Figure 5.

The HSO spectrum shows clear signals originating from the C–H bonds (around 2950 cm⁻¹) and C = O (at 1700 cm⁻¹), respectively. The absence of signals that are characteristic of hydroxyl group bonds proves that there were no significant amounts of free fatty acids and hydroxylated acids in the oil. A relatively weak but distinctive signal slightly above 3000 cm⁻¹ confirms the existence of double bonds between carbon atoms, which most likely originated from unsaturated fatty acid residues in HSO, as confirmed by others [49].

Among the signals common to all emulsion samples, those arising from the O–H bonds (at about 3250 cm⁻¹), aliphatic C–H (at 2800–2950 cm⁻¹) and C = O (at 1700 cm⁻¹) should be distinguished. Although the differences in the intensities of the individual bands can be explained by the different composition of the emulsions, it is important to note that these differences were not only in intensity, but also in the position of the signals in the 1050–1150 cm⁻¹ range. They could originate from vibrations of the C–O bonds (to a lesser extent, of C–N, which occurs in the protein structure). A similar shift, visible as a change in shape, is present for vibrations from the O–H group. This shift may indicate a variation in the strength and configuration of the hydrogen bonds between polar components of the tested emulsions, suggesting the complex interactions between the emulsion components, which may then have impact on their physicochemical properties [50,51] and plausible biological activity of the samples [52,53].

2.5. Color and Refractive Index Analysis

The results of the L*a*b* color analysis of samples in Figure 6 showed that the S02 sample represented more lightness, followed by S13, S08 and S14. Considering the ratio of ingredients in samples S02, S13 and S08, it was determined that the L* value decreased with higher concentrations of HSO or the AHE, respectively. However, sample S14 having a low content of HSO (1%) and the AHE (1%) featured a low L* value. We hypothesize that other factors might affect the lightness of samples in addition to the concentration of ingredients. Based on droplet size distribution studies, the PDI and Z-ave of S14 were lower compared to other samples (Table 1). In addition, according to McClements and Demetriades (1998), interactions between the ingredients within the emulsion and some other factors such as transmission, reflection, scattering and absorption can also affect the color and appearance of the emulsion [54].

It was also found that the difference between the a* values was negligible. In this analysis, sample S14 showed lower b* values. This observation may also relate to the higher ratio of whey to HSO in this formulation, which caused a shift in favor of a yellow shade. To monitor the color of the emulsions with a unique index, the WI values of the samples were calculated based on the L*, a* and b* values. The WI values were found to be 42.68, 35.86, 40.00 and 29.87 for S02, S08, S13 and S14, respectively, indicating marked difference between sample appearance.

The refractive index recorded for the dispersed phase represented the average refractive indices of individual droplets (Table 5). The theory that the thickness of the phase interface is small enough comparative to the wavelength of the related light would indicate that the interface of droplets plus the oil phase act as an individual dispersed phase [55].

The total appearance of an emulsion can be determined by both light scattering and absorption. Scattering is mainly accountable by recording the turbidity or lightness of an emulsion, while absorption determines the chromatic properties (redness, blueness, greenness, etc.) [56]. The relationship between the color of an emulsion and the refractive index ratio has an important role in foods containing high proportions of the aqueous phase. The refractive index of the samples (Table 4) was similar to the RI of the dispersed phase (water). This property allowed us to use an optical transmission method to analyze emulsions. The level of lipid oxidation in an emulsion, for example, can be determined by adding some chemicals (such as glycerol) to adjust the RI to 1.0, then measuring the spectra of the absorption [57].

2.6. Biological Activity

To assess the microbiological safety of optimized emulsions, we determined the effect of emulsions on the Lactobacillus sp. 2675 strain, a common probiotic (Figure 7). As a reference sample, cultures without an emulsion were considered (with 0% changes in cell metabolic activity). The highest increase in the bacterial cell metabolic activity (84%) was noticed for sample S08. The observed increase for sample S13 was relatively lower (66%). The increase in metabolic activities were lowest for samples S14 and S02 (37% and 31%, respectively). It can be assumed that higher concentrations of the AHE promoted higher probiotic bacteria growth; however, the small number of experiments needs further confirmation of results.

Nevertheless, the critical observations were that all samples showed a positive effect on the metabolic activity of Lactobaccilus sp. 2675, indication non-toxic properties of the emulsions. Our results correspond directly with results presented by Gharehcheshmeh et al. [58], who reported no effect of a sweet almond and sesame oil nanoemulsions on the growth of L. delbrueckii subsp. bulgaricus. There are however, few studies showing an effect of AHE on probiotic bacteria. However, former studies showing an impact on probiotic bacteria of polyphenol-rich extracts from blueberry [57,58] or apples [59] agree with our results.

Nevertheless, the crucial observations are that all the samples exhibited a positive effect on the metabolic activity, which proves the nontoxic properties of the emulsions. Our results correspond directly with the results presented by Gharehcheshmeh et al. who studied the impact of sweet almond and sesame oil nanoemulsions on the growth of L. delbrueckii subsp. bulgaricus. They did not observe any inhibitory effect of nanoemulsions.

It is worth adding that the blueberry (Vaccinium corymbosum) extract [59,60] also showed beneficial effects on probiotic microorganisms due to the presence of polyphenols, which is also a distinguishing feature of the AHE we used [31]. It should also be noted that plant-derived substances can become an additional nutrient for probiotic microorganisms while providing a protective substance against the adverse effects of the external environment. Ahmad et al. [61] obtained promising results using polyphenols extracted from apples to protect the Bifidobacterium lactis bacteria during freezing. In the light of these studies, we can see the great potential of our obtained emulsions to stimulate growth and protect probiotic bacteria.

3.2. Emulsion Preparation

Emulsion samples with a volume of 20 mL were prepared in sterile 50 mL plastic laboratory tubes. The two-step process used [9] had small modifications, such as in the first step where the components were homogenized using a hand homogenizer CAT X120 equipped with a T10 shaft. The samples were mixed at 10,000 rpm for 600 s. In the second step, the samples were homogenized (sonicator Sonoplus, Bandelin, Berlin, Germany) in the following conditions: 10 min, in 10 s/10 s action/break cycles, amplitude of 16%, and cooled with tap water.

3.3. Experimental Design

In this study, a three level Box–Behnken experiment design with response surface methodology (RSM) (Design Expert software version 13.0) was used for determining the nanoemulsion criteria. The variables included the AHE, WPI and HSO concentrations (% w/v) at three coded levels, i.e., −1, 0, +1. The specific ranges of the variables were selected according to prior knowledge and preliminary studies performed on the formulation of the hempseed oil nanoemulsion. The Box-Behnken design provided 17 experimental trials with five replications for the center point. The optimal formulations for further stability tests and visual observations were chosen based on the droplet size distribution results of the experimental trails.