Introduction

The food industry is considered among the most important ones for the economy within the European Union, for its relevant impact in the collecting, processing, and manufacturing of the agro food area products. Plants are a relevant source of antioxidant substances [1]. Their beneficial effects may be taken from the use of their extracts, which can be used as a supplement in daily product foodstuff with the aim of obtaining several health beneficial effects [2, 3].

During the last years, the interest of consumers for “all natural”, considered more healthy foods in the collective imagination, raised exponentially. At the same time, it is also growing the interest with reference to the nutrients effects in preventing/delaying the onset of life style related diseases and improving physical and mental well-being. In this regard, the impact of supplements and functional foods on the market has been highly satisfactory in terms of producers, consumer’s interest and consumption.

Among functional foods, dairy-based functional foods account for nearly 43% of the market. It has been estimated that the functional foods represent about the 43% of the market of bioactive compounds added to food matrices to increase their intrinsic nutritional value [4]. One example is the use of dairy beverages that can be enriched with bioactives like omega-3 fatty acids, phytosterols, vitamins, minerals, fibers, phenolic compounds, with recognized beneficial effects to human health [5]. It is also to be noted that the addition of extracts obtained from plants like Panax quinquefolius L. (American ginseng), green tea, black tea and soybean into milk beverages has been reported [6]. The growing interest in herbal added dairy foods is due to their potential antioxidant effect in preventing and as aid of the conventional pharmacological therapy approach in health conditions like diabetes, obesity, cardiovascular diseases, inflammatory diseases [7,8,9].

Hibiscus sabdariffa (H. sabdariffa) plant is characterized by a high health beneficial compounds content (mainly phenolic acids and flavonoids like quercetin and anthocyanins). This peculiarity makes this Africa originating plant an interesting source of substances with assessed bioactivity [10], indicating also the plant extracts as a possible useful and suitable tool for the incorporation in dairy beverages to realize functional beverages. Phenolic compounds are, in general, poorly stable in presence of oxygen, light, high temperature, and this represent a limiting factor for their possible incorporation in food and especially in beverages. Moreover, they may undergo chemical modifications and/or suffer degradation in the gastrointestinal tract when taken orally. Moreover, when food are added with bioactive compounds, some technological problems can also occur because of the potential chemical reactions which may happen between bioactives and other food components during processing and storage. It is, therefore, of utmost importance to ensure the stability of these functional ingredients into food matrixes. Nanoencapsulation involves the incorporation of food ingredients into nanosized carriers, with the purpose of protecting them from gastrointestinal environment and from interaction with other components of the formulation. The ultimate aim is to increase the bioavailability of food ingredients by their intact delivery to their site of absorption/action. Among the different types of nanoparticles that have been exploited for oral administration of bioactive compounds, nanostructured lipid carriers (NLC) appear to combine several advantages, e.g., high loading capacity for poorly soluble and lipophilic compounds, modified-release profile, absence of cyto/genotoxicity and high in vitro/in vivo biocompatibility [11,12,13,14,15]. Their lipid composition acts as oral absorption enhancer thereby increasing the oral bioavailability of loaded compounds [16,17,18,19]. We anticipate that phenolic compounds extracted from H. sabdariffa can be loaded into NLC to improve their physicochemical stability for the further incorporation in a dairy beverage. The aim of this work has been the encapsulation of polyphenol-enriched extracts from H. sabdariffa into NLC to be further processed to obtain a dairy beverage (e.g. for the production of fortified milk) (Fig. 1). The two different approaches for the extracts obtainment have been: (i) microwave-assisted extraction (MAE); (ii) a pressurized liquid extraction (PLE). The long-term stability and texture properties of the obtained milk have been characterized, as well as the in vitro absorption of quercetin and anthocyanins as referenced phenolic compounds from H. sabdariffa.

Fig. 1
figure 1

Schematic representation of the preparation of H. sabdariffa samples by homogenization in a ultra-centrifuge mill, followed by the production of polyphenol-enriched extracts by two methods, i.e., microwave assisted extraction (MAE) or by pressurized liquid extraction (PLE). The extracts were loaded into NLC followed by the incorporation of the nanoparticles in skimmed milk

Full size image

Materials and methods

Materials

For the extraction, calyces of H. sabdariffa collected from Monteloeder (Elche, Spain) have been used as starting material. They have been then homogenized using a ZM200 (Retsch GmbH, Haan, Germany) ultra-centrifuge mill. Absolute ethanol was purchased from VWR international S.A.S. (Fontenay-sous-Bois, France); ultrapure water was purchased from Millipore (Millipore, Bedford, MA, USA). The sea sand was obtained from Panreac (Barcelona, Spain). In the nanostructured lipid carriers (NLC), the lipid phase was a mixture of solid Biogras™ Vegetal BM 297 ATO, provided by Gattefossé (Saint-Priest, France) and liquid soybean oil purchased from Sigma-Aldrich (Sintra, Portugal). The surfactants used were Tween 80 and Span 80, purchased from Merck (Hohenbrunn, Germany). The obtained nanoparticles were incorporate into skimmed milk (Mimosa, Porto, Portugal). For the in vitro release study, phosphate buffered saline (PBS) at pH 7.4 was prepared using disodium hydrogen phosphate and sodium chloride (Sigma, Darmstadt, Germany).

Polyphenol-enriched extracts of H. sabdariffa obtainment

Microwave assisted extraction (MAE) and pressurized liquid extraction (PLE) were performed to obtain two polyphenol-enriched extracts, as described by Pimentel-Moral et al. [20]. Briefly, a 60% ethanol (v/v) solution for the MAE extraction was used. The microwave apparatus was a microwave extraction reactor (Anton Paar GmbH, Graz, Austria); T was set at 164 ºC and the total extraction time was of 22 min.

For the PLE extraction, a Dionex ASE 350 extractor (Dionex Corp., Sunnyvale, CA, USA) was used. Temperature was set at 200 °C and a 100% (v/v) ethanol solution was used. Both extracts were then dried, stored at − 20 °C, and protected from light until further use.

Preparation of H. sabdariffa-loaded nanostructured lipid carriers (HS-NLC)

MAE and PLE extracts were loaded into the NLC using the method described by Pimentel-Moral et al. [21], and the two obtained preparations were named HS-MAE-NLC and HS-PLE-NLC, respectively.

The optimum percentages of lipid phase (2.21% (w/w)) and surfactant (1.93% (w/w)) were used. Briefly, 10 mg of H. sabdariffa extract from MAE or PLE were dissolved in 1.25 mL of ultrapure water. The aqueous phase was added to a melted lipid phase (T = 60 °C) composed of a liquid lipid (30% (v/v)) and a solid lipid (70% (w/w)) to produce a hot w/o emulsion by mechanical stirring at 11000 rpm for 90 s. The obtained w/o emulsion was then added to an external aqueous phase composed of a water-surfactants mixture of Tween 80 and Span 80 (90:10) to produce a w/o/w emulsion. The hot w/o/w emulsion was kept under mechanical stirring at 11000 rpm for 3 min. An IKA Ultra-Turrax T-25 Digital Homogenizer (Staufen, Germany) was used in both cases. Sonication at T = 90% was then applied using a sonication apparatus (Ultrasonic processor VCX500, Sonics, Switzerland), and HS-NLC were obtained under cooling of the hot w/o/w emulsion in ice bath.

Dairy-based beverages formulation

The dairy beverages were produced combining the developed HS-MAE-NLC and HS-PLE-NLC with the skimmed milk in a 70:20:10 ratio, i.e., each HS extract-loaded NLC beverage was prepared by combining 10 mL of HS extract-loaded NLC, 20 mL of water, and 70 mL of skimmed milk. These formulations were stirred for 10 min. To compare the in vitro delivery of H. sabdariffa from NLC against the free extract (NLC-free beverage), H. sabdariffa extract was incorporated in the milk in a free-form. To that end, 10 mg of powdered extract was firstly dispersed in 25 mL of water and then the NLC-free beverage was produced using the 70:20:10 ratio, as described previously.

Long-term stability of HS-NLC and HS-NLC dairy beverages

The long-term stability of blank NLC (without HS extracts), of HS-MAE-NLC and of HS-PLE-NLC was simulated ad carried out using the LUMiSizer (LUM GmbH, Dias de Sousa, Portugal). This analytic centrifuge accelerates the destabilization by sedimentation or creaming phenomena using the Space-and Time-resolved Extinction Profiles (STEP™−Technology). The samples were stored at 4 °C, 22 °C and 40 °C for 48 h, then placed in rectangular test-tubes (optical path of 2 mm), and centrifuged at 4000 rpm (T = 25 °C) and the variation of the extinction was measured at a wavelength of 865 nm. The samples were analyzed every 10 s allowing to obtain 1000 different profiles per sample. This approach allowed the evaluation of the temperature effect on the formulation stability in a accelerate time.

Texture analysis of new functional beverages

The texture of the dairy beverages has been determined using a texture analyzer equipped with a 7 mm blade (TA, XTplus; Stable Micro System, UK). The parameters of the texture analyzer were as follows: pretest and post-test speed, 2 mm/s; blade movement speed, 0.2 mm/s; moving distance, 5 mm, T = 10 °C. All measurements were determined in triplicate.

In vitro release study

In vitro release study was done using Franz glass cells were used. This method consists of donor and receptor chambers separated by cellulose membranes with an average pore size 0.6 µm [22, 23]. As receptor medium 5 mL of phosphate-buffered saline (PBS) at pH 7.4 was used; this procedure is commonly used as biological environment to assess the behavior of oral formulations. To simulate the surface intestinal temperature the receptor chamber was kept at 37 °C. Each cell was kept under magnetic stirring during the experiment. One mL of each sample (HS-MAE-NLC in milk, HS-PLE-NLC in milk, and free HS extract in milk) were placed in the donor chamber and a volume of 200 µL was sampled after 15, 30 min, 1, 2, 4, 6, 8 and 24 h. After each sample selection, the Franz cells were filled up with receptor medium. A Synergy Mx Monochromator-Based Multi-Mode Microplate reader by Bio-Tek (Winooski, VT, USA) was used to measure the absorbance of the collected samples at a wavelength of 300 nm and 520 nm for the determination of quercetin and anthocyanins, respectively. Four kinetic models (zero order, and first order kinetics, Higuchi and Korsmeyer-Peppas), have been tested in the evaluation of the release mechanism of quercetin and anthocyanin from NLC [24]. The selection of the most appropriate model was based on the obtained R2 values [25, 26].

Statistical analysis

Statistically significant differences between mean values were determined by ANOVA with Dunnet post-test. Results were considered significantly different if p < 0.05. The results are presented as the mean ± standard deviation (SD) of n = 3. Statistical analysis was performed using GraphPad Prism (version 8.0) (GraphPad Software, San Diego, CA, USA) [27].

Results and discussion

The long-term stability of nanoparticles in suspension analysis using LUMiSizer to measure the intensity of the transmitted light during centrifugation as a function of time and position is increasing in its use [28, 29]. The instability phenomena which are related to the migration of particles (sedimentation or creaming) when centrifugal acceleration is applied are related to the progression of the transmission profiles which provides information about the kinetics of the nanoparticle behavior and also makes it possible the predict the shelf-life. The transmission profiles of samples of blank NLC (without the addition of HS extract), HS-MAE-NLC and HS-PLE-NLC as well as of the dairy beverages containing these nanoparticles incorporated and stored for 48 h at different temperatures, are shown in Fig. 2. In comparison to HS-MAE-NLCs and HS-PLE-NLCs profiles, blank NLC without extract showed a multimodal particle size distribution; this recorded broad range of particle sizes indicated in comparison with the others an unstable formulation. On the opposite, when loaded with H. sabdariffa extracts, NLC exhibited longer-term stability. This result can be attributed to the interactions between the polyphenols and the NLC lipid part, since polyphenols may distribute at the oil/tensioactive interface providing this way a protective antioxidant effect in aqueous media and on lipid phase [20]. Similar findings were previously reported by Kumazawa [23] when studying a lipid model of epigallocatechin gallate (EGCG) interaction with dimyristoyl-phosphatidylcholine phosphate (DMPC) using 31P NMR [30]. The spacing between the profiles in NLC (Fig. 2a) decreased over time and the particles tend to migrate more slowly, since there is the possibility of particle aggregation. At 4 °C, creaming was the most prominent mechanism of destabilization, although at higher temperatures, samples showed a strong trend to form sediments. For HS-MAE-NLC and HS-PLE-NLC, a homogeneous particle size distribution was observed as indicated by a symmetrical spacing in the profiles. High stability was obtained in a faster time. Moreover, for HS-MAE-NLC similar profiles were observed at different values of the temperature, although there was a slightly higher tendency for creaming at 4 °C, which gradually diminished. The sedimentation increased at higher temperatures. On the opposite, as shown in Fig. 2c, this phenomenon was more pronounced for HS-PLE-NLC. With respect to functional dairy-based beverages, HS-MAE-NLC and HS-PLE-NLC milks showed similar behaviors. At 4 °C and 22 °C a slight creaming process was observed in both beverages. When T reached a value of 40 °C, the sedimentation mechanism increased. These instability mechanisms were less pronounced when nanoparticles were incorporated into a dairy beverage providing longer-term stability. Indeed, protein from milk matrix can stabilize the incorporated lipid nanoparticles. It has been demonstrated that many proteins are surface-active molecules which can be used as emulsifiers and improve the stability of emulsions [31]. Milk proteins can provide adsorbed layer onto the oil droplets, which create a physical barrier against coalescence. The results obtained measuring the instability index of these functional beverages indicated that higher temperatures slightly influenced the HS-MAE-NLC and HS-PLE-NLC milk. The HS-MAE-NLC dairy beverage demonstrated to be more stable than the HS-PLE-NLC as shown in Table 1.

Fig. 2
figure 2

Long-term stability of empty NLC (a), HS-MAE-NLC (b), HS-PLE-NLC (c), milk with HS-MAE-NLC (d) and milk with HS-PLE-NLC (e), at different storage temperatures, evaluating the transmission signal (%, YY-axis) over the position (in mm, XX axis)

Full size image
Table 1 Instability index of fortified milks as a function of temperature
Full size table

Among the key quality attributes evaluated in the food industry to assess a product acceptability, texture is one of the most relevant. For processed foods, texture properties are instrumental for the control of processing steps, e.g., heating, frying and drying to obtain the desired properties of the final product. Therefore, food formulation may be related to desirable or undesirable changes in texture. The texture can be evaluated using sensory and instrumental assays.

Nonetheless, sensory methods are subjective measured, which may imply disadvantages depending on the level of training received by the panel tests participants, on the time necessary and the high cost. These factors have limited their use compared to more appealing and lower cost instrumental approaches [32].

In this work, firmness, consistency, cohesiveness and index of viscosity of skimmed milk and skimmed milk with HS-MAE-NLC and HS-PLE-NLC have been analysed. As showed in Table 2, all recorded parameters increased when nanoparticles were incorporated into dairy beverage. Significant differences (p < 0.05) were observed between skimmed milk before and after the incorporation of nanoparticles. Higher values were found for HS-MAE-NLC milk compared to HS-PLE-NLC milk. Between both, significant differences were found for firmness, consistency and cohesiveness, but not for the index of viscosity.

Table 2 Textural properties of fortified milk.
Full size table

Supplementation and interaction of NLC with the protein matrix of the milk may strengthen the texture parameters [33]. These results are according to previous researches. In a previous work, a texture analysis of fortified yogurt with nano and micro sized calcium, iron and zinc was measured and the results showed that the yogurts fortified with calcium and zinc nanoparticles increased their consistency and firmness [34]. An important step for product development is the in vitro release testing, which is instrumental to obtain information on drug release mechanism and kinetics, to establish in vivo/in vitro correlations. In the present study, quercetin and anthocyanins release was compared before (Fig. 3) and after (Figs. 4 and 5) the loading of the extracts into NLC. When analyzing the non-loaded extract (Fig. 3), the release of anthocyanin was more delayed than the release of quercetin. Within 4 h ca. 60% of quercetin has been released against ca 25% of anthocyanin. The loading of extracts into NLC clearly reduced the cumulative percentage of release over the course of the 24 h for both polyphenols (Figs. 4 and 5). However, interesting was to realize that the matrix of NLC retained much longer quercetin (Fig. 4) than anthocyanins (Fig. 5).

Fig. 3
figure 3

In vitro release of quercetin versus anthocyanins from free extract

Full size image
Fig. 4
figure 4

In vitro release of quercetin from HS-MAE-NLC versus HS-PLE-NLC

Full size image
Fig. 5
figure 5

In vitro release of anthocyanins from HS-MAE-NLC versus HS-PLE-NLC

Full size image

Anthocyanins are water-soluble flavonoids and are less likely to be retained within lipophilic matrices dispersed in aqueous medium; within 5 h ca 100% of anthocyanins were quantified in the release medium. Quercetin, on the other hand, is practically insoluble in water, taking about 9 h to release ca. 49% from the PLE extract versus 45 ~ % from MAE extract.

The in vitro release of quercetin from NLC-containing milks was slower than when the milk contained the free form of the HF extract. At a pH value of 7.4 (intestine conditions), the quercetin release from beverage containing free HF extract experienced a relevant increase, namely from 5.2% to 53.1% in 4 h. Quercetin release from NLC-containing milks did not reach 40% within the same time-period, i.e., from 6.3% and 5.9% to 36.0% and 22.7%, respectively, for HS-PLE-NLC and HS-MAE-NLC milks. The slowest quercetin release was found for HS-MAE-NLC. Similar trends were observed for the anthocyanins. For this group of compounds, the release was slower when the HF extract was encapsulated. A 100% release was found after 4 h in HF non-encapsulated milk, while for HS-PLE-NLC and HS-MAE-NLC beverages, was released after 4 h, 92.5% and 31.1%, respectively. For anthocyanins, HS-MAE-NLC fortified milk demonstrated to have the slowest release as shown in Fig. 4. From literature data, it has been observed that quercetin has been loaded in cationic nanostructured lipid carrier (QR-CNLC) and an in vitro release assay was carried out. The release of free quercetin was much faster than QR-CNLC and only 50% of quercetin was released from QR-CNLC after 24 h. Similar results were found in our proposed experimental work: only 43% and 49% of quercetin was released from HS-MAE-NLC and HS-PLE-NLC dairy beverages after 24 h. Moreover, quercetin has been incorporated in SLN (QT-SLN) to enhance gastrointestinal absorption [35]. However, the in vitro release studies showed that 50% of quercetin from QT-SLN was released within 6 h, indicating a faster delivery compared to the results reported in the present study. This may be attributed to the higher loading capacity of NLC compared to SLN [17, 21, 36, 37].

Thus, NLC unlike SLN, contain lipid droplets that are partially crystallized and have a less-ordered crystalline structure or an amorphous solid structure preventing the risk of expulsion of drug entrapped. In a study by Kumari [31] has been described the loading of quercetin in poly-D,L-lactide (PLA) nanoparticles. The release profile indicated a fast burst release of quercetin (40–45%) within 0–0.5 h although the complete release was recorded at 72 h [38]. For the anthocyanins, the release behavior was shown to be dependent on the pH and temperature, as these compounds are easily degradable at high pH and high temperature values [39]. At 37 °C and pH 7.4 (intestinal conditions), anthocyanins are quickly released showing a complete delivery within 4 h for beverages not containing encapsulated HF. Nevertheless, for HS-MAE-NLC and HS-PLE-NLC dairy drinks, a complete release was observed within 6 h, with a slower release profile. Loypimai et al. [32] showed that the percent of release of anthocyanins from SLN was higher at pH = 7.4 than at pH = 3. In addition, the short-term stability was investigated in PBS (pH = 7.4) at 25, 40 and 50 °C showing that the most appropriate condition for anthocyanin storage is at lower temperatures (T < 25 °C). Anthocyanin SLN increased the apparent short-term accelerated stability of anthocyanins against relatively high pH and temperature. In our study, NLC have demonstrated to be a potential technological approach for the loading of flavonoids, such as quercetin and anthocyanins in a dairy product beverage, capable to exhibit a modified release profile, and which may be successfully administered in a liquid form.

To describe the release mechanism followed by the flavonoids from the lipid matrices formulated in milks, four mathematical fitting models have been used i.e., zero model, first model, Higuchi and Korsmeyers–Peppas model (Power Law), commonly used to describe the drug release from nanoparticles. Both free flavonoids followed an unconventional kinetics as linearity depicted high noise in all four models. With respect to the R2 values, the best fitting model was shown to be the Korsmeyers–Peppas model with a R2 of 0.8596 for quercetin (Fig. 6) and 0.7431 for anthocyanins (Fig. 7).

Fig. 6
figure 6

Mathematical fitting models of the release profile of quercetin from free (non-loaded) extract

Full size image
Fig. 7
figure 7

Mathematical fitting models of the release profile of anthocyanin from free (non-loaded) extract

Full size image

The loading of the extracts into NLC improved the regression coefficient. Except for the release of quercetin from HS-MEA-NLC which was closer to Higuchi model (Fig. 8), in general release of flavonoids from NLC clearly fitted to the powder law (\({M}_{t}/{M}_{\infty } = {k}^{\mathrm{^{\prime}}}{t}^{n}\)) of Korsmeyers–Peppas model (Figs. 9, 10 and 11).

Fig. 8
figure 8

Mathematical fitting models of the release profile of quercetin from HS-MAE-NLC

Full size image
Fig. 9
figure 9

Mathematical fitting models of the release profile of quercetin from HS-PLE-NLC

Full size image
Fig. 10
figure 10

Mathematical fitting models of the release profile of anthocyanin from HS-MAE-NLC

Full size image
Fig. 11
figure 11

Mathematical fitting models of the release profile of anthocyanin from HS-PAE-NLC

Full size image

Korsmeyers–Peppas model was developed to describe, specifically, the drug release from a matricial core, where\(M_{t}\) is the cumulative amount of drug released at the time t, \({M}_{\infty }\) is the cumulative amount of drug released at infinite time, \({k}^{\mathrm{^{\prime}}}\) is a constant governed by the physicochemical properties of the matrix, and n is the diffusional release exponent. It is called the powder law since it is the n value that describes the release mechanism of the drug. When n = 0.5, the release is governed by the Fickian diffusion, when 0.5 < n < 1.0 a non-Fickian diffusion is observed. For spherical particles, it has been described that drug release becomes independent of time and reaches zero-order release known as Case II transport, achieved as n approaches 1.0, typical of a non-Fickian transport.

Conclusions

In the present study, a new formulation based on enriched milk with H. sabdariffa-loaded nanostructured lipid carriers of food grade has been proposed for the oral delivery of antioxidant polyphenols as a dairy beverage. It is anticipated that the protein matrix from milk contributes for the stabilization of lipid nanoparticles, which showed to modify the release profile of quercetin and anthocyanins. The method to obtain the polyphenols-enriched extracts, namely, microwave-assisted extraction (MAE) and pressurized liquid extraction (PLE), influenced the in vitro release profile of quercetin and anthocyanins. Except for the release of quercetin from HS-MEA-NLC which was closer to Higuchi model, in general release of flavonoids from NLC fitted to the Korsmeyers–Peppas model. In general, the loading of the extracts into NLC improved the regression coefficient. Our results highlight the clear need for further research towards the development of functionalized foodstuff, to have a better understanding of the development of functional dairy beverages containing polyphenols with improved bioactivities.