*Article Agaricus bisporus* **By-Products as a Source of Chitin-Glucan Complex Enriched Dietary Fibre with Potential Bioactivity**

#### **Sara M. Fraga and Fernando M. Nunes \***

CQ-VR, Chemistry Research Centre—Vila Real, Food and Wine Chemistry Lab, Chemistry Department, Universidade de Trás-os-Montes e Alto Douro, Apartado 1013, 5000-801 Vila Real, Portugal; sarafraga\_@hotmail.com

**\*** Correspondence: fnunes@utad.pt

Received: 24 February 2020; Accepted: 13 March 2020; Published: 25 March 2020

**Featured Application: A novel, simple and environmentally friendly method was developed and optimize to produce a dietary fiber ingredient from** *Agaricus bisporus* **by-products, with suitable characteristics for food application and potential biological activity, as a mean for upgrading mushroom industry wastes.**

**Abstract:** Mushroom production generates large amounts of by-products whose disposal creates environmental problems. The high abundance of biological active non-starch polysaccharides in mushroom cell walls makes these by-products attractive for dietary fiber-based ingredient (DFI) production. Traditional methods of dietary fiber preparation didn't allow to obtain a DFI with suitable chemical and functional properties. In this work a simple and environmentally friendly method was developed and optimized for DFI production using a central composite design with treatment time, hydrogen peroxide and sodium hydroxide concentration as factors and chemical composition, chromatic and functional properties as dependent variables. The chemical composition of the DFI was strongly influenced by the process parameters and its functional and color properties were dependent on its fiber and protein content, respectively. The method developed is simple, uses food grade and low-cost reagents and procedures yielding a DFI with white color, no odor and a high concentration of dietary fiber (>60%) with an identical sugar composition to the original mushroom fiber. Due to the high water and oil retention capacity, this DFI may be used not only for dietary fiber enrichment and reduction of the food energy value but also as a functional ingredient with potential bioactivity.

**Keywords:** dietary fiber; mushroom; *Agaricus bisporus*; dietary fiber ingredient; chemical composition; functional properties; optimization; central composite design

#### **1. Introduction**

Mushrooms consumption and production over the last decades has shown a phenomenal growth with several times increase in tonnage [1]. Some of the more common cultivated mushroom species are the button mushroom (*Agaricus bisporus*) which is widely cultivated in Europe and comprising about 32% of the world mushroom production, the Shiitake mushroom (*Lentinus edodes*) which is grown for centuries in China and other oriental countries and the oyster mushroom (*Pleurotus ostreatus*) cultivated in several countries around the world [2]. This increase is related to the increasing awareness of consumers for the importance of a healthy diet [3,4]. Indeed, a variety of substances present in mushrooms have been shown to present beneficial biological effects, from these, polysaccharides comprising the β-D-glucans and heteroglucans are the best known and most potent mushroom-derived substances with antitumor and immunomodulating properties [5–8]. The fungal cell walls are

composed by an alkali-insoluble structural skeleton mainly composed of β-(1→3)-glucan covalently linked to chitin [9,10], forming a chitin-glucan complex (CGC). Chitin and CGC have been shown to have very interesting biological activities [11–15]. In addition, several studies have shown that polysaccharides from a variety of mushrooms, including those of *A. bisporus*, have been successfully used as prebiotics [16–22].

Nevertheless, mushroom production generates a large amount of by-products including waste and off-grade mushrooms with no suitable commercial use that ranges between 5% and 20% of production volume [23]. Mushrooms and mushroom by-products, especially those from *A. bisporus*, are rapidly perishable products [24], and they suffer rapid and deleterious transformations resulting in darker products due to the tyrosinase activity and synthesis of melanins [25] and with an unpleasant smell, creating environmental problems for their disposal. The European regulation of waste management, Directive 2008/98/ EC ('Waste Framework Directive'), focused on the reduction of waste generation by 30% in 2025, requires that waste should be managed without endangering human health and harming the environment, in particular without risk to water, air, soil, plants or animals, and without causing a nuisance through noise or odors [26]. To reduce the environmental impact of the agro-food industries and the dependence on raw materials, the implementation of valorization procedures for these materials is stimulated. Some strategies have already been evaluated, for example, the use of *A. bisporus* by-products for non-animal chitin and chitosan production [27]. Nevertheless, most of the methods employed for the extraction of fungal cell wall polysaccharides are tedious and time-consuming involving the use of high temperatures and high concentrated alkaline and acid solutions or using specific enzymatic treatment in combination with synthetic detergents [28–31]. In addition, the use of concentrated reagents can deteriorate the native properties of the obtained products [32]. The production of dietary fiber ingredients (DFI) has been successfully used to upgrade agricultural products and by-products of cereals, fruits, and vegetables [33,34], therefore, the production of a high added value food ingredient based on mushroom dietary fiber from by-products and off-grade mushrooms can be an economical alternative to the simple waste disposal. Nevertheless for the food industry, beyond their nutritional and health benefits, dietary fibers also have technological properties that can be used in food formulations, resulting in texture modification and enhancement of food stability during production and storage, and more importantly, added fibers cannot alter the sensorial properties of foods where they are used. Altogether, these factors will determine their successful use in foods.

The purpose of this work was to develop and optimize a simple and green method for the production of mushroom CGC enriched dietary fiber from *A. bisporus* off-grade mushrooms and in situation of excessive mushroom production, using a response surface methodology based on a central composite design to evaluate the influence of the process variables in the chemical composition, nutritional and functional properties to evaluate its suitability for use as a DFI in food formulations.

#### **2. Materials and Methods**

#### *2.1. Materials*

As the method developed in this work is intended to be used for the valorization of off-grade mushrooms with no commercial value and the excesses of mushroom production, for reasons of simplicity and easier management, the *A. bisporus* mushrooms were bought on the local market for obtaining the DFI. Samples were stored in dark conditions at 4 ◦C until experiments began for a maximum of one day. All reagents used were of analytical grade. All reported values, unless otherwise stated, are expressed on a dry weight basis and are the average values of the analysis of at least two different replicates.

#### *2.2. Method for Preparation of Mushroom DFI*

Preliminary experiments in our lab allow concluding (detailed later) that to obtain a food ingredient based on dietary fiber from mushrooms with good technological properties, i.e., with a white color and no aroma, an alkaline oxidative treatment using sodium hydroxide and hydrogen peroxide was the most appropriate, and this method was further optimized concerning the treatment time, sodium hydroxide concentration and hydrogen peroxide concentration (Table 1). For obtaining mushroom DFI, mushrooms (300 g) were sliced and ground in a Waring blender (3 min) and transferred to the treatment solution (1 L) whose composition was varied according to Table 1. The material was stirred at 300 rpm at room temperature (20–22 ◦C) during the time for each specific treatment. After the treatment, the solution pH was neutralized with concentrated sulfuric acid to pH 6–7, and hydrogen peroxide was destroyed by adding NaHSO3. The material was filtrated, re-suspended in water and filtered again. The material was freeze-dried and the solid was ground and weighted. The yield, chemical and functional properties of the prepared mushroom DFI were determined to evaluate the impact of the studied process parameters.

#### *2.3. Dietary Fiber, Protein, Moisture, Lipids, Ash and Caloric Value Content of the DFI*

Due to the expected presence of chitin in the DFI, the content and sugar composition of dietary fiber of the mushroom DFI was determined as non-starch polysaccharides (NSP) by the method of Englyst et al. [35,36], being determined by the sum of sugars released after acid hydrolysis, and quantified by anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD) [37]. Protein content was determined by quantification of the nitrogen content of the mushroom DFI after correction for the nitrogen from glucosamine present and multiplication by 6.25. The total nitrogen content of the DFI was determined by the Dumas method. Moisture, lipids and ash content of the mushroom DFI were determined according to AOAC [38]. Total carbohydrates were calculated by difference. Caloric values, on a dry basis, were computed using the Attwater coefficients [39], corrected for 2% of ash.

#### *2.4. FTIR Analysis*

A Unicam Research Series FTIR spectrometer was used. The spectra were recorded in the range of wavenumbers 4000–450 cm−<sup>1</sup> with 50 scans being taken at 2 cm−<sup>1</sup> resolution. Pellets were prepared by thoroughly mixing DFI sample or reference polysaccharides (chitin, chitosan, and curdlan) and KBr at a 1:200 sample/KBr weight ratio in a small size agate mortar. The resulting mixture was placed in a Perkin–Elmer manual hydraulic press, and a force of 15 tons was applied for 10 min. The spectra obtained were background corrected and smoothed using the Savitzky–Golay algorithm using PeakFit v4 (AISN Software Inc., Oregon, United States, 1995). Chitin from crab shells (Sigma), Chitosan 66% deacetylation degree (Sigma) and curdlan (Sigma) were used as reference polysaccharides.

#### *2.5. DFI Chromatic Characteristics*

The chromatic characteristics of the DFI were evaluated with a Chroma Meter CR-300 Minolta (Osaka, Japan). CIE Lab coordinates were obtained using D65 illuminant a 10 observer as a reference system. L\*, a\*, and b\* parameters were calculated from the average of five color measurements. The equipment was calibrated with a white standard (L\* = 97.71; a\* = −0.59 and b\* = 2.31).

#### *2.6. Water Retention Capacity and Oil Retention Capacity*

Water retention capacity (WRC) of the mushroom DFI obtained under the different experimental conditions was determined under external centrifugal force using the method described by Robertson et al. [40]. Fiber (~1 g) was hydrated in water (30 mL) for 18 h, centrifuged (3000× *g*; 20 min), drained, and dried. WRC was calculated as the amount of water retained per g of dry fiber residue. For determination of the oil retention capacity (ORC) of the mushroom DFI, the same procedure described previously was used, but instead of water 30 mL of corn oil were used.

#### *2.7. Experimental Design*

A rotatable central composite design of the experiments was performed with k = 3—treatment time, hydrogen peroxide and sodium hydroxide concentrations as independent variables. Central composite experiments consisted of three sets of experimental points [41]: (1) a factorial design with 2k points, k is the number of xi variables (factors) with coded levels +1 and −1 for each; (2) a star for 2k points, coded as +α and −α on the axis of the system at a distance of 2k/4 from the origin, that accounts for non-linearity; (3) central points, which are replicated to provide an estimate of the lack of fit of the linear statistical model obtained as well as the pure error of the experiments (due to unreliability in the measurement of the dependent variable) [42]. The main advantage of this methodology is to decrease the number of experimental trials needed to evaluate multiple parameters and their interactions. The established ranges were: sodium hydroxide concentration (*CNaOH*) 0.1–0.5 M, hydrogen peroxide concentration (*CH*2*O*<sup>2</sup> )—1.5–4.5%, treatment time (*T*)—2–6 h. Table 1 shows the coded and uncoded experimental design.

#### *2.8. Statistical Analysis*

Data was fit to second-order polynomial Equation (1) for each dependent Y variable, through multiple regression analysis using Statistic® vs. 10 software.

$$\mathbf{Y} = \boldsymbol{\beta}\_0 + \boldsymbol{\beta}\_1 \mathbf{X}\_1 + \boldsymbol{\beta}\_2 \mathbf{X}\_2 + \boldsymbol{\beta}\_3 \mathbf{X}\_3 + \boldsymbol{\beta}\_4 \mathbf{X}\_1^2 + \boldsymbol{\beta}\_5 \mathbf{X}\_2^2 + \boldsymbol{\beta}\_6 \mathbf{X}\_3^2 + \boldsymbol{\beta}\_7 \mathbf{X}\_1 \mathbf{X}\_2 + \boldsymbol{\beta}\_8 \mathbf{X}\_1 \mathbf{X}\_3 + \boldsymbol{\beta}\_9 \mathbf{X}\_2 \mathbf{X}\_3,\tag{1}$$

β<sup>n</sup> are regression equation coefficients and Xn the independent variables. Based on the predicted model equations surface plots were generated. The analysis of the variance was performed to determine the lack of fit and the significance of the effects of each of the three independent factors, using the mean square pure error as the error term. This provides a more sensitive test of model fit because the effects of the additional higher-order terms are removed from the error.

#### *2.9. Principal Component Analysis (PCA)*

PCA is one of the most often used chemometric methods for data reduction and exploratory analysis of high-dimensional data sets [43]. PCA decomposes the original matrix into multiplication of loading (chemical composition, color, and functional properties) and score (dietary fiber samples) matrices. The principal components are linear combinations of the original variables. The principal components are uncorrelated and account for the total variance of the original variables. PCA is an unsupervised method of pattern recognition in the sense that no grouping of the data has to be known before the analysis. The new sub-space defined by the principal components leads to a model that is easier to interpret than the original data set. From these results, it should be possible to highlight several characteristics and correlate them to the chemical composition of the different DFI produced.

#### **3. Results**

#### *3.1. Development and Optimization of a Method for Obtaining Mushroom DFI*

Preliminary experiments using methods previously described for the production of DFIs from other sources, including the simple hot water extraction [44], ethanol extraction of the low molecular weight material [45] and enzymatic removal of protein [46], rendered a deep yellow-brown material with an unpleasant smell (results not shown). This prompts us to develop an efficient method for the production of mushroom DFI that could render a food ingredient with desirable characteristics for their application in foods. This could be accomplished by using a chemical treatment with an alkaline solution containing hydrogen peroxide at room temperature. Under these conditions, the product obtained presented a white color and with no perceived odor. The method developed for the production of mushroom dietary fiber was optimized considering three process variables: treatment time (*T*), hydrogen peroxide concentration (*CH*2*O*<sup>2</sup> ) and sodium hydroxide concentration (*CNaOH*). A rotatable central composite design with α = 1.68 was used for the optimization of the parameters for the production of mushroom DFI as well as optimization of the DFI physico-chemical, nutritional and functional properties. Table 1 shows the design matrix of factors for the rotatable central composite design. DFI yield, chemical and nutritional composition, color, and functional properties were measured as the output variables.

#### *3.2. Yield of the Mushroom DFI*

The yield of mushroom DFI obtained varied between a minimum of 15% to a maximum of 53% of the mushroom dry weight (0.77% to 2.8% of the fresh mushroom weight for a mushroom water content of 94.7%) (Table 1). Based on the central composite design, each of the three factors (*T*, *CH*2*O*<sup>2</sup> , and *CNaOH*) had a significant effect on the mushroom DFI yield (Table 2). There was also observed a significant interaction between *T* and *CNaOH*. Using multiple regression analysis, a second-order polynomial equation model based on codded levels was used to fit the experimental results. This model which consists of the factors found to be significant is shown in Table A1. The ANOVA results of the model obtained indicated an adequate performance with R<sup>2</sup> = 0.64, implying that 64% of the variations observed for the mushroom DFI yield are explained by the factors considered. It was therefore considered that the model provided a good description of the experimental data. Nevertheless, the F-test for the lack of fit was also significant (*p* < 0.05) therefore a more complicated model or additional factors (for example variability of the mushrooms used in each experiment) are required to a higher fit of the experimental data [41]. Figure 1 shows the three-dimension response surface curves of mushroom DFI yield for each pair of factors by keeping the third factor constant at the level where it presented the maximum value. The factor *T* presented the highest positive effect (Table 2, Figure 1a) on the mushroom DFI yield (maximum for the +1.682 level), followed by the factor *CH*2*O*<sup>2</sup> were it was observed a positive effect (Table 2; Figure 1b) for the linear effect and a negative effect for the quadratic effect (maximum value for the *CH*2*O*<sup>2</sup> > 0 level). The interaction between the factor *T* and *CNaOH* and the linear effect of the factor *CNaOH* both had a positive effect on the DFI yield (Table 2, Figure 1a).


*Appl. Sci.* **2020** , *10*, 2232



*Appl. Sci.* **2020** , *10*, 2232

**Figure 1.** Response surfaces for mushroom DFI yield as a function of: (**a**) time and hydrogen peroxide concentration (at sodium hydroxide concentration level of 1.682); (**b**) time and sodium hydroxide concentration (at hydrogen peroxide concentration level of 1); (**c**) hydrogen peroxide concentration and sodium hydroxide concentration (at time level of 1.682).

These results show that during the treatment of the mushroom material with the alkaline solution containing hydrogen peroxide there was observed an insolubilization of material with treatment time, this insolubilization being promoted by higher NaOH concentrations. According to the response surface analysis, the predicted maximum mushroom DFI yield in the range studied is 2.9% of the mushroom fresh weight (95% prediction interval: 2.1–3.6%) when the factor *T* is +1.682, the factor *CH*2*O*<sup>2</sup> is +0 and the factor *CNaOH* is +1.682, corresponding to 55% of the mushroom solids. To understand what is being insolubilized during the treatment of the mushroom material, and as the composition of DFIs are important from a legal perspective [47] but also for its nutritional and functional properties, the fiber, protein, lipid and ash contents of the DFI were determined.

#### *3.3. Fibre Content and Composition*

As can be observed in Table 1, fiber was the main component of the mushroom DFI for all the conditions employed and ranged from 45.9% to 70.4%. Only the factor *T* had a significant negative effect on the DFI fiber content (Table 2), therefore although higher treatment times resulted in higher DFI yields, there was also observed a decrease in the relative abundance of dietary fiber in the DFI. The previously described increase in the DFI yield with *T* is not related to an increase in fiber retention during the treatment as there was not observed a correlation between fiber yield and treatment time (R2

= 0.2092). For obtaining a mushroom DFI with higher fiber contents the lower processing times assure this value (Table 1). The yield of dietary fiber obtained ranged between 0.50% of the mushroom's fresh weight to 1.3%, corresponding to a yield of fiber between 31% and 81% of original mushroom fiber content (on average 54%). This yield is explained by the fact that some of the *A. bisporus* polysaccharides are water and alkali-soluble [48].

In Table 3 it is shown the sugar composition of the mushroom DFI obtained for the different experimental conditions tested. In addition, the sugar composition of the fiber of the freeze-dried mushrooms is shown. The method applied resulted in a relative increase of the fiber in the DFI from 1.9 to 2.9 times in comparison with the original mushroom, showing the efficiency of the method employed for enriching the DFI with the original mushroom fiber. As can be observed, the main sugars present in the mushroom DFI were glucose followed by glucosamine and xylose that together account for more than 91% of the total sugars present. Smaller amounts of mannose, glucuronic acid, fucose, and galacturonic acid were also present. This sugar composition is similar to that described in previous works that have shown that *A. bisporus* cell walls are composed predominantly by β(1→3)-linked glucan containing some β(1→6) linkages containing also chitin [27,48,49]. FTIR analysis of the DFI confirms the presence of chitin (Figure 2). The spectra had characteristic bands at 3400–3480 cm−<sup>1</sup> that responded to OH-3 and CH2OH-6 intra- and intermolecular hydrogen bonds, bands at 1650 cm−<sup>1</sup> for amide I, and 1557 cm−<sup>1</sup> for amide II vibrational mode. The chitin present in the mushroom DFI is in an antiparallel α-conformation [50] as there is observed a split of the amide I vibration band at 1655 cm−1, identical to the reference crustacean chitin (Figure 2). These results confirmed that the glycosaminoglycans of DFI from *A. bisporus* were in highly acetylated form. The ratio of intensities of the bands at ~1379 and ~2920 cm−<sup>1</sup> has been suggested as the crystallinity index for chitin and chitosan [51]. The crystallinity index of the DFI chitin was similar to that of the reference chitin (1.07 vs. 1.12, respectively) and higher than that observed for the reference chitosan (0.78). The crystallinity index obtained mushroom DFI was higher than that obtained by Wu et al. [27], this being probably due to the preparation method used by these authors (1M NaOH at 95 ◦C during 30 min followed by 2% acetic acid at 95 ◦C during 6 h). A clear spectrum of the glucan in DFI could not be observed, due to the overlapping of chitin bands and the lack of unique bands in β-glucans when compared to chitin (Figure 3). This composition of dietary fiber is similar to that previously described [52], although in our work the amount of glucosamine was lower, and this can be related with different times after harvesting of the mushrooms [53] or to different *A. bisporus* strains used [54], nevertheless the chitin content of the *A. bisporus* obtained in this work is in agreement with previous works [27,48,49,55]. Although the relative abundance of the different sugars in the mushroom DFI is similar to that observed for the original mushroom there was a decrease in the relative abundance of fucose (−60%), galactose (−70%), and glucuronic acid (−60%). The other sugars present increased their relative abundance by 10% for glucose, 70% for xylose and 30% for mannose. Galacturonic acid was not detected in the original mushrooms but was present in most of the DFI obtained, and the relative abundance of glucosamine was on average the same as that found in the original mushroom dietary fiber.


*Appl. Sci.* **2020**, *10*, 2232

**Table 3.** Sugar composition

 of starting material (g/100 g) and mushroom dietary fiber ingredient (g/100 g ashless basis).

ANOVA and Tukey post-hoc test).

**Figure 2.** FTIR spectra of the DFI obtained in treatment nº 4 and of the reference polysaccharides chitin, chitosan and curdlan.

**Figure 3.** Dietary fiber ingredient obtained for the treatment nº 4. *CNaOH* = 0.419 M; *CH*2*O*<sup>2</sup> = 3.9% and *T* = 2.81 h.

#### *3.4. Protein, Fat, Ash Content and Energy*

Protein was the second most abundant component of the mushroom DFI, ranging from 11 to 40% (Table 1). The yield of protein obtained varied between a minimum of 0.081% of the mushroom fresh weight to 0.88%, representing between 6.7% and 73% of the original mushroom protein content. Only *T* had a significant effect on the yield of protein from the mushroom. There was observed an increase in the amount of protein recovered in the DFI with increasing *T*, nevertheless, the model (Table 3) is

not enough to explain the variations observed in the amount of protein recovered in the DFI (R<sup>2</sup> = 0.33) during the chemical treatment of the mushroom material. Lower *T* assure lower amounts of protein in the DFI. This higher recovery of protein in the ingredient with increasing *T* was probably due to the denaturation of protein during the treatment in alkaline solution [56], to the crosslinking of protein due to the hydrogen peroxide treatment [57] or due to the alkaline conditions employed leading to the formation of lysinoalanine crosslinks [58].

The fat recovered in the mushroom DFI varied between a minimum of 0.016% of the mushroom fresh weight, to 0.057%, representing between 4.9% and 17.4% of the original mushroom fat content. There is observed an increase in the amount of fat recovered in the DFI with the increasing *T* and increasing *CNaOH* being observed a significant interaction between these two factors. This higher recovery of fat in the final product was probably due to the precipitation or adsorption of fatty acids released from triglycerides during the alkaline treatment, as also higher *CNaOH* and increasing *T* resulted in higher amounts of fat recovered.

The caloric values were calculated for each mushroom DFI produced under the different conditions employed and ranged from 261 to 316 kcal on a dry basis (Table 1). As can be observed, and as expected, the DFI with a higher amount of dietary fiber presented the lower caloric values.

#### *3.5. Colour of Mushroom DFI*

The mushroom DFI presented a white/yellowish color depending on the treatment conditions (Table 4).


**Table 4.** Yield, water and oil retention capacity and color parameters <sup>a</sup> obtained for the different runs used for the optimization process.

<sup>a</sup> L\* defines lightness, a\* denotes the red (+)/green (−) value and b\* the yellow (+)/blue (−) value.

In theory, the perfect colorless white (white point) has the values L\* = 100, a\* = 0, b\* = 0, therefore, processing conditions rendering DFI with values close to these theoretical values will allow to obtain DFI with a white color. Lightness (L\*) was significantly affected by the three factors, being also observed a significant interaction between *T* and *CH*2*O*<sup>2</sup> and between *CH*2*O*<sup>2</sup> and *CNaOH* (Table 2). Taking all these effects into account, for obtaining a DFI with a high L\* value, the optimum *CH*2*O*<sup>2</sup> . level is 0 and aiming having a DFI with higher fiber values (Section 3.2), the *CNaOH* level should be 1.682 and

the level for the treatment time factor should be −1 (predicted L\* = 103.79; 88.90 to 118.18). For the a\* value (red-green coordinate of the color space), all factors, either linear or quadratic effects were significant. In addition, there was observed a significant interaction between all factors (Table 2). There are several combinations of independent variables (*T*, *CH*2*O*<sup>2</sup> , *CNaOH*) that allow to obtain the desired value of a\* = 0 as there is observed a significant interaction between all factors (Table 2). Optimum values for a\* for factor *T* − 1, important for obtaining high fiber percentage in the DFI, can be obtained with the following combination: levels of *CH*2*O*<sup>2</sup> of 1 and of *CNaOH* of 1 (predicted a\* value of 0.785 and prediction interval of −5.760 to 7.330). For the b\* value (yellow-blue coordinate of the color space), the linear terms of *T* and *CNaOH* and all quadratic terms of the factors had a significant effect (Table 2). Additionally, there was observed a significant interaction between *T* and *CH*2*O*<sup>2</sup> and *CH*2*O*<sup>2</sup> and *CNaOH* (Table 2). For the desired value of b\*, all factors should remain at the lower level yielding a b\* value of 0.131 (95% prediction interval of −2.814 to 3.076). These results show that the method developed allowed us to obtain a DFI with desirable neutral color values for their application as a food ingredient (Figure 3).

#### *3.6. Mushroom DFI Water and Oil Retention Capacity*

The water retention capacity (WRC) of mushroom DFI is shown in Table 4. WRC ranged from a minimum of 4.2 g water/g of DFI to a maximum of 21.3 g water/g of DFI. The highest WRC values obtained for the mushroom DFI are higher than that obtained for dietary fiber concentrates obtained from orange, peach, artichoke and asparagus, mango peel, sugar beet (10–14 g/g), apple and pear (6–7 g/g) and much higher than that observed for wheat and oat bran, carrot, and pea dietary fiber concentrates (3–4 g/g) [59–61] and in the range of that observed for potato fiber (23–25 g/g) [62]. As can be observed in Figure 4a, the WRC of the mushroom DFI was strongly dependent on *T*, being observed a decrease in the WRC with increasing *T*, whatever the *CH*2*O*<sup>2</sup> .

For the *CH*2*O*<sup>2</sup> there was observed a minimum for the WRC at the central value of this factor, with lower and higher levels showing a positive effect in the WRC of the DFI. For the *CNaOH* there were observed to different situations (Figure 4b), for lower treatment times there was observed an increase in the WRC with increasing *CNaOH*, the reverse was true for higher treatment times. Keeping the level of factor *T* at the minimum, the effect of *CNaOH* on the WRC of DFI was much more important than the *CH*2*O*<sup>2</sup> (Figure 4c). Whatever the *CH*2*O*<sup>2</sup> the higher WRC was observed for the higher *CNaOH*. For a higher WRC of the DFI the optimum *CH*2*O*<sup>2</sup> level was 1.682, the *CNaOH* level should be 1.682 and the level for factor *T* should be −1.682 (predicted WRC of 36.0 g/g of DFI with a 95% prediction interval of 29.6 to 42.4 g/g).

The oil retention capacity (ORC) of mushroom DFI is shown in Table 4. ORC varied from a minimum of 6.0 g oil/g of DFI to a maximum of 18.9 g oil/g of DFI. The ORC values obtained for the mushroom DFI were higher than that obtained for dietary fiber concentrates obtained from apple, pea, wheat, carrot (1–2.3 g/g) and sugar beet (5 g/g) [59], apple pomace and citrus by-products (0.6–1.8 mL oil/g) [63], unripe banana flour (~2 mL oil/g) [64]; carrot pulp dried at 50 ◦C (~6 mL oil/g) [65] and asparagus by-products (5.5–8.5 mL oil/g) [66]. As can be observed in Figure 5a, the ORC of the mushroom DFI was strongly dependent on the *CH*2*O*<sup>2</sup> and *T*, being observed the higher value of ORC for lower *T* and higher *CH*2*O*<sup>2</sup> .

**Figure 4.** Response surfaces for mushroom DFI water retention capacity (WRC) as a function of: (**a**) time and hydrogen peroxide concentration (at sodium hydroxide concentration level of 1.682); (**b**) time and sodium hydroxide concentration (at hydrogen peroxide concentration level of 1.682); (**c**) hydrogen peroxide concentration and sodium hydroxide concentration (at time level of −1.682).

The same effect was observed for factor *CNaOH*, for lower *T* there was observed an increase in the ORC with increasing *CNaOH*, the reverse was true for high *T* (Figure 5b). When the effect of the factors *CNaOH* and *CH*2*O*<sup>2</sup> are represented, there was observed that the maximum ORC was obtained or for high *CNaOH* and *CH*2*O*<sup>2</sup> when the treatment time was at the lower level (Figure 5c) the reversed being observed for longer treatment times (result not showed). Using the factor levels where the maximum value is observed (*T* = −1.682; *CH*2*O*<sup>2</sup> and *CNaOH* = 1.682) the predicted value of ORC of the DFI is 35.8 g/g of DFI (95% prediction interval between 30.8 to 40.8 g/g).

**Figure 5.** Response surfaces for mushroom DFI oil retention capacity (ORC) as a function of: (**a**) time and hydrogen peroxide concentration (at sodium hydroxide concentration level of 1.682); (**b**) time and sodium hydroxide concentration (at hydrogen peroxide concentration level of 1.682); (**c**) hydrogen peroxide concentration and sodium hydroxide concentration (at time level of −1.682).

#### *3.7. E*ff*ect of Mushroom DFI Chemical Composition on the Colour and Functional Properties*

To understand the effect of the chemical composition of the mushroom DFI obtained under the different processing conditions on the color and functional properties of mushroom DFI, the data obtained was analyzed by principal component analysis (PCA). The first three principal components obtained explained >70% of the total variance in the original data set. The loadings express how well the new PCs correlate with the original variables (Figure 6a and Table 2). The first PC, which explains 44.4% of the total variance, correlates positively with total fiber, glucosamine, glucose, xylose, WRC and ORC and negatively with protein and the b\* value. The second PC, which explains 14.6% of the total variance, correlates positively with galacturonic acid content and negatively with glucuronic acid content, and PC3, which explains 13.5% of the total variance correlates positively with the a\* value and negatively with lightness (L\*). These results show that the WRC and ORC of the DFI are correlated with the total sugar content of the DFI (Figure 6a) and the b\* value, related to the yellowness of the DFI is correlated with its protein content.

**Figure 6.** (**a**) Factor loading plot based on correlations of the mushroom DFI chemical composition, color characteristics and functional properties projected on the space of PC1 vs. PC3; (**b**) scores plot of the mushroom DFI samples projected on the space of PC1 vs. PC3.

The scatter plot of the sample scores on the PC1 and PC3 scores (Figure 6b) shows the formation of two distinct clusters along the PC1. Samples with positive PC1 scores have a high relative content of total sugars and also a high WRC and low protein content and were less yellow, the opposite being true for the samples with negative PC1 scores. The WRC of dietary fibers is dependent on its structure [62,67] and chemical composition [68,69]. The water retained by dietary fiber material generally comprise three types of water, retained by three mechanisms: water bound by the hydrophilic polysaccharides of the fiber, dependent on the chemical composition of the fiber; water held by the fiber in the fiber matrix, mainly dependent on the pore size distribution of the fiber matrix; and water associated with fiber other than bound or matrix water, trapped within the cell wall lumen, dependent on the fiber source, method of preparation and method of measurement. Contrarily the ORC is in part related to its chemical composition but is more largely a function of the porosity of the fiber structure rather than the affinity of the fiber molecule for oil [70].

Samples with high scores on PC3 contain a relatively high a\* value and lower L\* value. As can be observed in Figure 6b, the group composed by the samples with a high PC1 score, corresponding to the DFI resulting from the 5 first treatments applied, are subdivided according to PC3 forming two groups, samples 3 and 5 contain a higher a\* and lower L\* values when compared to samples 1, 2 and 4.

#### **4. Discussion**

The production of a DFI from *A. bisporus* by-products, besides being a good strategy to reduce wastes generated in the mushroom agro-industry, can yield a DFI that besides having prebiotic activity [19,21,22] can also present a range of other very interesting biological activities. Several studies have shown that *A. bisporus* polysaccharides show immunostimulatory [71–74], antioxidant [74,75], antitumor [76–78], anti-inflammatory [7,79] and anti-sepsis activities [80], as well as antinociceptive inhibition [7,80]. The method developed allowed to concentrate the *A. bisporus* fiber polysaccharides up to 2.9 times, with a chemical composition similar to the initial cell wall, and so the biological activities of these polysaccharides are expected to be enhanced due to the concentration observed during the production process. In addition, it is expected that the original CGC present in *A. bisporus* cell walls is maintained [27,48,49]. CGC has been shown beneficial effects concerning the development of obesity and associated metabolic diabetes and hepatic steatosis, through a mechanism related to the restoration of the composition and/or the activity of gut bacteria, namely, bacteria from clostridial cluster XIVa [81]. Furthermore, CGC has potential beneficial effects concerning the development of atherosclerosis, mainly related to improving the antioxidant status [11,12]. On the other hand, chitin consumption can reduce triglyceride and cholesterol levels in liver and increase excretion of triglycerides in feces [13] and reduce cholesterol levels [14], and linear β-(1→3)-glucans have shown hypoglycemic activity accompanied by promotion of metabolism and inhibition of inflammation, through suppressing SGLT-1 expression and possibly associated with alteration of gut microbiota [15]. Nevertheless, to be acceptable, a DFI added to a food product must perform in a satisfactorily as a food ingredient [82], namely be bland in taste, color, and odor. Besides the well-established nutritional benefits of adding dietary fiber ingredients to food products, the use of DFI can also have important technological advantages [59]. Of the various technological benefits, the increases in the water retention capacity (WRC) and oil retention capacity (ORC) of foods are one of the main advantages of using DFIs [59,83]. Both WRC and ORC can increase the technological yield of food. WRC of DFIs can be advantageous in sauces and soups, but also for their textural properties, enhancing the flow properties and avoiding lump formation in powdered mixes (e.g., ready-to-eat sauces, mixes of spices, flavoring agents). ORC can be exploited in foods (cooked meat products) to enhance their retention of fat that is normally lost during cooking, being also beneficial for flavor retention [59]. The WRC can have a nutritional interest as well, as increase in water retention has been related to an increase in orocaecal transit time. The water-holding capacity of dietary fiber has been proposed to be valuable in the diet to alter stool bulking [84]. Increased stool weight can cause shorter gut transit times limiting the exposure of the gut to secondary bile acids and other toxins [85,86].

The use of a sodium hydroxide concentration of 0.419 mol/L and hydrogen peroxide of 3.9% during 2.81 h, at room temperature, allow to obtain the highest yield of DFI (1.22% in relation to the fresh weight) with one of the highest dietary fiber (63.7 g/100 g) and lower protein (12.0 g/100 g) contents and highest WRC (21.3 g/g of DFI) and ORC (18.9 g/f of DFI). In addition, the chromatic characteristics of the DFI obtained show that it has a good lightness (L\* = 90.6) and neutral (white) color (a\* = −0.470 and b\* = 13.71).

The by-products of *A. bisporus* production can be successfully used for the production of a DFI using a simple method at room temperature and using food-grade materials, being a good strategy to reduce wastes in the mushroom agro-industry. Due to the simplicity and the efficiency shown in the production of *A. bisporus* DFI, there are no anticipated problems in applying this technology to other abundant mushrooms wastes as those derived from the production of *Lentinus edodes* and the oyster mushroom (*Pleurotus ostreatus*).

**Author Contributions:** Conceptualization, F.M.N.; investigation, S.M.F.; resources, F.M.N.; writing—original draft preparation, F.M.N.; writing—review and editing, F.M.N. and S.M.F.; supervision, F.M.N.; project administration, F.M.N.; funding acquisition, F.M.N. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Centro de Química—Vila Real (CQ-VR), through the Fundação para a Ciência e Tecnologia (FCT)—Portugal, grant number UIDB/00616/2020 and the APC was funded by Fundação para a Ciência e Tecnologia (FCT)—Portugal, grant number UIDB/00616/2020.

**Acknowledgments:** The authors would like to acknowledge to João Coutinho for the nitrogen analysis.

**Conflicts of Interest:** The authors declare no conflicts of interest.



**Table A1.** Regression equation coefficients for response parameters of mushrooms DFI.

X1= Time; X2 = *CH*2*O*2 ; X3= *CNaOH*.


**Table 2.** Factor-variable correlations (factor loadings), based on correlations.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Processed Fruiting Bodies of** *Lentinus edodes* **as a Source of Biologically Active Polysaccharides**

**Marta Ziaja-Sołtys 1,\*, Wojciech Radzki 2, Jakub Nowak 3, Jolanta Topolska 1, Ewa Jabło ´nska-Ry´s 2, Aneta Sławi ´nska 2, Katarzyna Skrzypczak 2, Andrzej Kuczumow <sup>3</sup> and Anna Bogucka-Kocka <sup>1</sup>**


Received: 18 December 2019; Accepted: 6 January 2020; Published: 8 January 2020

**Abstract:** Water soluble polysaccharides (WSP) were isolated from *Lentinus edodes* fruiting bodies. The mushrooms were previously subjected to various processing techniques which included blanching, boiling, and fermenting with lactic acid bacteria. Therefore, the impact of processing on the content and biological activities of WSP was established. Non-processed fruiting bodies contained 10.70 ± 0.09 mg/g fw. Boiling caused ~12% decrease in the amount of WSP, while blanched and fermented mushrooms showed ~6% decline. Fourier transform infrared spectroscopy analysis (FTIR) confirmed the presence of β-glycosidic links, whereas due to size exclusion chromatography 216 kDa and 11 kDa molecules were detected. WSP exhibited antioxidant potential in FRAP (ferric ion reducing antioxidant power) and ABTS (2,2 -azino-bis(3-ethylbenzothiazoline-6-sulfonic acid)) assays. Cytotoxic properties were determined on MCF-7 and T47D human breast cell lines using MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) test. Both biological activities decreased as the result of boiling and fermenting.

**Keywords:** polysaccharides; *Lentinus edodes*; antioxidant; cytotoxicity; processing; mushrooms; LAB (lactic acid bacteria); fermenting

#### **1. Introduction**

Among thousands of mushrooms only about 20 species are cultivated commercially for culinary purposes. Japanese mushroom *Lentinus edodes*, commonly known as Shiitake, is cultivated for both its culinary and medicinal applications [1]. It has been reported that consumption of edible mushrooms provides a significant health improvement as they are low in calories, sodium, fat and cholesterol, but rich in proteins, carbohydrates, fibre, vitamins and minerals [2–4]. Commonly consumed mushroom species exposed to a source of ultraviolet (UV) radiation can generate nutritionally relevant amounts of vitamin D [5]. Providing considerable amount of iron and well absorbed proteins they are often called "the meat of forest" [2]. Mushrooms are rich in immunomodulating compounds which, unlike traditional chemical drugs, do not cause any harmful effect or allergic reactions and put no additional stress on the body [6,7].

Out of all mushroom-derived substances, polysaccharides are known to have the most potent antitumor, antioxidative and immunomodulating properties [6,8,9] but their biological activities differ greatly depending on the structural and physical features [10]. For example, both in vitro and in vivo antitumor activities of mushroom extracts arise from the presence of β-glucans, especially containing β-1,3 bounds in the polysaccharide chain [2,11]. Studies confirmed that lentinan, the high molecular weight polysaccharide β-1,3-d-glucan with β-1,6-glucopiranoside branches extracted from *Lentinus edodes* fruiting bodies not only has immunomodulating properties but also can suppress the growth of cancer cells and induce them to apoptosis [10,12,13]. However, the details of molecular mechanisms of these processes remain unclear [14,15]. Chen et al. [16] stated that among eight studied species of medicinal mushrooms, *Lentinus edodes* polysaccharides showed the strongest scavenging activity for hydroxyl radicals and were able to inhibit the proliferation of MCF-7 tumor cells. Antibacterial and antifungal properties of Shiitake extracts have also been reported [4,17].

As edible mushrooms are characterized by a short shelf life, 1–3 days at room temperature, they should be consumed directly after harvesting because their nutritional value is then the best then. Additionally, most of mushroom species need to be processed before the consumption. In the literature there are some information about the influence of processing like cooking, baking, drying or freezing on the contents of health promoting compounds [18,19] but little is known about the effect of these processes on mushroom-derived polysaccharides antioxidant and antiproliferative activities [20].

Contemporary consumers are more aware of what they eat and drink, they more often choose home made products or those without any artificial stabilizers. Noticing this tendency the aim of this article was to verify the impact of some processing methods on the content, chemical composition, antioxidant and antiproliferative activity of water soluble polysaccharides (WSP) obtained from *Lentinus edodes*. The processes that were chosen are easy to conduct and commonly used at home, including boiling, blanching and fermenting with lactic acid bacteria (*Lactobacillus plantarum*).

#### **2. Materials and Methods**

#### *2.1. Biological Material*

*Lentinus edodes* fruiting bodies were provided by a private producer (AgRoN, Z ˛abki, Poland). Harvested mushrooms were from the same crop and were transported in five kg plastic trays. Fresh fruiting bodies were stored at 5 ◦C up to five hours before the analysis.

The bacterial strain used in the experiment (*Lactobacillus plantarum* IBB76) was obtained from IBB Central Collection of Strains (Warsaw, Poland). This strain was used previously in fermentation of mushrooms [21–23].

#### *2.2. Processing of Mushrooms*

Fruiting bodies were divided into four portions. The first one was blanched for five minutes in citric acid solution (0.5% *w*/*v*) at 95 ◦C. The second portion was boiled in water for fifteen minutes at 100 ◦C. The third group was blanched as above and fermented with lactic acid bacteria strain, as in the previous study [23]. The last portion (control group) was not processed. All the four portions were lyophilized (Alpha 1–2 LD plus, Christ, Germany) prior the extraction of polysaccharides.

#### *2.3. Extraction of Water Soluble Polysaccharides (WSP)*

Lyophilized fruiting bodies were milled into fine powder, extracted with ethanol (80 ◦C, 1 h) and centrifuged. Insoluble part was rinsed with alcohol and subjected to water extraction (115 ◦C, 1 h). The obtained water extract was concentred and mixed with 2-propanol to precipitate polysaccharides. Precipitates were then washed three times with alcohol, lyophilized and weighed. The extraction was carried out in triplicate and extraction yields were calculated.

#### *2.4. Chemical Characteristics of Polysaccharides*

The extracted WSP were redissolved in water and subjected to three colorimetric assays. Total carbohydrate content was determined with the phenol–sulfuric acid method [24]. Protein content was determined with Bradford reagent [25] and total phenolics content was quantified with

Phenol-Cicalteau reagent [26]. Glucose, bovine albumin and gallic acid (respectively) were used to construct calibration curves.

Fourier transform infrared spectra of the samples were recorded on Nicolet NXR 9650 spectrometer (Thermo, Waltham, MA, USA) equipped with an ATR (attenuated total reflection) module. The samples were scanned within the range of 400–4000 cm<sup>−</sup>1.

Gel permeation chromatography (GPC) was employed to determine the molecular weight. The analysis was conducted according to the methodology described by Malinowska et al. [27]. Briefly, the samples were dissolved in NaN3 water solution and applied to the following TSK-GEL columns: G5000PWXL, G3000PWXL andG2500PWXL (Tosoh, Tokyo, Japan). Samples were detected with a Refracto Monitor IV refractive index detector (LDC Analytical, Riviera Beach, FL, USA) and compared with pullulans standards.

#### *2.5. Antioxidant Assays*

ABTS (2,2 -azino-bis(3-ethylbenzothiazoline-6-sulfonic acid)) radical scavenging activity assay was conducted according to Re et al. [28]. The samples (25 μL) were mixed with ABTS + solution (975 μL) and measured at 734 nm after 15 min. Ferric reducing antioxidant power (FRAP) analysis was done according to the methodology of Benzie and Strain [29]. FRAP reagent (1900 μL) was added to the samples (100 μL) and after 90 min of incubation the absorbance was measured at 593 nm. The results of both antioxidant assays were compared with calibration curves made with Trolox and expressed as micromoles of Trolox equivalent (TE) per 1 g of mushroom dry weight.

#### *2.6. Cytotoxic Properties*

The MCF-7 adenocarcinoma, cell line was obtained from the American Type Culture Collection (ATCC) cat. no. HTB-22. The T47D ductal, epithelial cancer cell line was obtained from the European Collection of Authenticated Cell Cultures (ECACC) no. 85102201. These two human breast cancer cell lines were grown at standard conditions (37 ◦C, 5% CO2, 95% humidity) in RPMI 1640 medium (Sigma) supplemented with 10% fetal bovine serum (FBS, Sigma), 100 U/mL penicillin (Polfa, Poland) and 100 μg/mL streptomycin (Polfa, Warsaw, Poland). For MCF-7 cells 50 μg/mL of bovine insulin (Sigma) was also added. Cytotoxicity effect of tested WSP was measured with a quantitative colorimetric toxicity MTT assay based on the transformation of yellow, soluble tetrazolium salts (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) to purple-blue insoluble formazan by mitochondrial succinate dehydrogenase of living, metabolically active cells [30,31].

T47D and MCF-7 cells (100 <sup>μ</sup>L/well) were seeded on 96-well plates at concentrations of 3 <sup>×</sup> 10<sup>5</sup> cells/mL and 2 <sup>×</sup> <sup>10</sup><sup>5</sup> cells/mL, respectively. After 24 h incubation at standard conditions, when cells in each well reached 75–80% confluence, the growth medium was replaced with WSP dilutions in RPMI 1640 with 2% FBS. The concentrations ranged from 25 μg/mL to 250 μg/mL. Cells were incubated with polysaccharides for 24 h, followed by addition of 25 μL of MTT (Sigma) solution (5 mg/mL in PBS) per well. After 3 h incubation at 37 ◦C formazan crystals were solubilized with 100 μL of lysis buffer (10% SDS in 0.01 M HCl) per well. Plates were incubated overnight at standard conditions. The absorbance was read at 540 nm with a plate reader (680 XR, Bio-Rad), and the mean value for each concentrations was calculated. The percentage of viable cells were calculated from the absorbance.

#### *2.7. Statistical Analysis*

The measurements were done in triplicate and results were expressed as the mean ± standard deviation. The collected data were evaluated using analysis of variance ANOVA with a level of significance set at *p* < 0.05. The LSD test was performed to assess statistically different results. Different letters on graphs and tables (a, b, c, etc.) show significant differences among the data according to LSD test (*p* < 0.05).

#### **3. Results and Discussion**

#### *3.1. The Content of Water Soluble Polysaccharides*

As shown in Table 1, mushrooms belonging to the control group contained 96.9 ± 0.8 mg/g dw of WSP. Blanching did not cause statistically relevant changes, whereas boiling led to the increase in the amount (112.0 ± 2.0 mg/g dw). The observed increase in the case of boiled fruiting bodies is with agreement with the previous study conducted on *Hypsizygus marmoreus* fungi [32] or button mushroom [22]. The reason for this could be the leaking of easy soluble substances into the brine and changing the ratio between high and low weight molecular weight compounds. Blanched and fermented mushrooms contained the lowest quantity of WSP (86.7 ± 3.1 mg/g dw).

**Table 1.** The content of the isolated water soluble polysaccharides (WSPs).


The results were also expressed in mg per g of fresh weight in order to consider the loss of the mass. As can be seen, the highest drop (~12%) was observed in the case of boiled fruiting bodies and slight decrease (~6%) in the case of blanched and fermented samples. Similar results were noticed. Boiling of mushrooms may lead to the changes in their mycelial structure and therefore facilitate the extraction of water soluble polysaccharides [32].

#### *3.2. Chemical Characteristics of Water Soluble Polysaccharides*

In the isolated WSPs the level of carbohydrate, protein and phenolics were determined (Table 2). Polysaccharides extracted from non-processed mushrooms contained 72.35 ± 3.77% of carbohydrate, 4.90 ± 0.46% of protein and 0.59 ± 0.02% of phenolic compounds. Previous research demonstrated a similar quantity of carbohydrate in *L. edodes* polysaccharides which ranged from 72% [33] to 78% [34]. Other authors who used different extraction method reported higher quantity of carbohydrate (87%), lower amount of protein (~3.5%) and no phenolic compounds [35]. In accordance with the presented results, previous studies have demonstrated similar content of protein but twice higher amount of phenolics [34]. Mushrooms are highly abundant in phenolics and these compounds have tendencies to bind to polysaccharides with hydrogen or even covalent bounds [36,37]. According to research phenolic antioxidants are released from polysaccharidic matrix by probiotic bacteria and then are absorbed into bloodstream [38].

**Table 2.** Chemical composition of the isolated WSPs.


The processing of mushrooms caused various changes in chemical composition. Blanching in citric acid solution resulted in higher carbohydrate content (79.67 ± 3.45%) along with lower protein (4.03 ± 0.28%) and phenolics content (0.49 ± 0.04%). Interestingly, boiling for 15 min affected only (negatively) protein quantity when compared with control samples. The most evident changes were observed in the case of blanching and fermenting, where significant decreases of phenolics and protein (0.25 ± 0.02% and 2.30 ± 0.35%, respectively) were noticed. In contrast, the level of

carbohydrate markedly rose reaching 93.00 ± 3.59%. These data is in accordance with previous experiment conducted on *Agaricus bisporus* [22]. The loss of protein and phenolics during lactic acid fermentation of mushrooms can be attributed to enzymes that are produced by bacteria: proteases degrading protein moieties [39] and esterases which release phenolic compounds [40]. Additionally, aqua-thermal processing may cause solubilization or even degradation of low molecular weight phenolics [41].

Fourier transform infrared spectroscopy enables us to detect functional groups which are present in a sample and is often used in the analysis of plant and fungi based polysaccharides [42–44]. FTIR coupled with a ATR accessory a is rapid and non-destructive tool which does not require sample preparation [45]. FTIR analysis may be useful in determination of polymers purity as well as for the detection of the type of glycosidic bond [46]. FTIR analysis of the isolated WSP is presented in Figure 1. The obtained spectra shows bands which are commonly present in polysaccharides mushroom origin. Broad bands at 3000–3500 cm−<sup>1</sup> and 2800–3000 cm−<sup>1</sup> are associated with stretching vibrations of O-H, N-H and C-H, respectively. Two signals at ~1645 cm−<sup>1</sup> (Amide I) and ~1535 cm−<sup>1</sup> (Amide II) indicate the presence of protein. They are result of C = O stretching vibrations (Amide I) and bending vibrations of N-H groups (Amide II) and were reported previously in *L. edodes* β-glucans [47]. Interestingly, Amide II band is barely visible in WSP isolated from blanched and fermented mushrooms. This is in accordance with the observed sharp drop in protein content shown due to Bradford assay. The signal at ~1520 cm−<sup>1</sup> can be attributed to C-C stretching of aromatic ring and suggests the presence of phenolic compounds [48]. This signal is the weakest in the case of WSP obtained from blanched and fermented mushrooms and may confirm the significant loss in phenolics. In the region of 1300–1450 cm−<sup>1</sup> there are signals which can be assigned to bending vibrations of O-H, C-O-H and CH2 [49], whereas intense peaks between 950–1190 cm−<sup>1</sup> are due to stretching vibrations of C-O-C, C-O-H, C-C [50]. FTIR analysis allowed to identify the type of glycosidic bonds dominating in the samples. Spectra showed absorbance at ~890 cm−<sup>1</sup> which is indicative of β-glucans [51,52].

**Figure 1.** FTIR spectra of WSP.

In order to establish molecular weight of the extracted polysaccharides, size exclusion chromatography was used (Figure 2). All the tested WSP gave intense signals at ~18.8 mL. These sharp peaks correspond to molecules of 216.3 kDa. This finding was also reported by Chen et al. [53] who isolated proteoglycan from *L. edodes*, having the molecular weight of 220 kDa and capable of scavenging free radicals. It is also consistent with the previous research in which 200 kDa polysaccharide (showing immunomodulating activities) was obtained from *L. edodes* water extract [54]. However, this study was unable to demonstrate the presence of larger molecules exceeding 600 kDa as was earlier reported by other authors [55,56]. Chromatogram of the control sample also contains the peak at 22.6 mL. It can be attributed to the smaller compound having the molecular weight of 11 kDa. This peak is removed due to the processing and a possible explanation for this might be that this compound was solubilized and extracted during blanching or boiling.

**Figure 2.** Size exclusion chromatography of WSP.

#### *3.3. Antioxidant Potential*

Mushrooms are rich in low molecular weight antioxidants, mainly phenolic acids. Such compounds are mainly present in free form and can be easily extracted with both water and organic solvents. However, some antioxidants bind to larger molecules and become less or non-extractable [57]. These compounds are often bound to mushroom derived polysaccharides and affect their antioxidant capacities. All the isolated WSP exhibited antioxidant activity in both assays (Figure 3). The highest antioxidant parameters were noticed in the case of the control samples (23.53 ± 0.30 μmoles of TE/g dw for ABTS and 12.29 ± 0.87 μmoles of TE/g dw for FRAP). Antioxidant activity of *L. edodes* polysaccharides was reported before and this study confirms earlier findings [16,35,58]. Blanching caused a significant drop in FRAP test (7.59 ± 0.75 μmoles of TE/g dw), while no relevant change in ABTS parameter was noticed. Boiling resulted in the decrease of ABTS activity (20.83 ± 1.24 μmoles of TE/g dw), whereas no further change in the case of FRAP test was observed. The most profound changes caused blanching and fermenting (7.78 ± 1.50 μmoles of TE/g dw for ABTS and 3.56 ± 0.33 μmoles of TE/g dw for FRAP). The results clearly showed that aqua-thermal processing may have negative impact on antioxidant parameters of polysaccharides. Previous findings showed that antioxidant activity of polysaccharides is related to total phenolics and protein content [59–61]. In this study, a high correlation was observed between ABTS values and total phenolics (R = 0.93) as well as protein content (R = 0.85). In the case of the FRAP parameter, the correlation coefficients were similar to ABTS (R = 0.96 and R = 0.82, respectively). Therefore, the reason for the drop in antioxidant capacity is very likely caused by the removal of phenolics and protein. Further studies should be carried out to isolate and identify antioxidant compounds which are bound to polysaccharidic matrix by covalent bonds [57].

#### *3.4. Cytotoxic Properties of WSP*

It has been reported that many mushroom polysaccharides show direct cytotoxicity to cancer cells [62]. To determine the response of cells to water soluble mushroom polysaccharides (WSP) we used MTT assay that measures the reduction in cell viability when exposed to cytotoxic substances. The control T47D and MCF-7 cells viability was assessed as 100%. Our results showed a dose dependent cytotoxic effect of crude polysaccharide extract from *Lentinus edodes* on MCF-7 cells

(Figure 4a). The control, a not processed WSP sample, caused a 73% and 41.7 ± 4.7% decrease of MCF-7 cell viability after treatment with 50 and 250 μg/mL concentration respectively. The processing of *Lentinus edodes* fruiting bodies particularly blanching and blanching with further fermentation had suppressed cytotoxic activity of applied WSP. After incubation with the highest polysaccharides concentration MCF-7 cells viability reached 74.0 ± 5.7% and 82.5 ± 3.5%, respectively. Boiling in water changed cytotoxic properties of polysaccharides in the least degree. MCF-7 cells viability after treating with (250 μg/mL) WSP from boiled mushrooms was 57.9 ± 5.8% [63]. Israilides et.al. achieved a 50% reduction in viability of MCF-7 cells, however the time of incubation with 73 ± 14 μg/mL of water extract from *Lentinus edodes* took 48 h. Moreover, it is known that whole mushroom extracts contain a lot of substances other than polysaccharides with very well documented direct cytotoxic activities achieved 50% reduction in viability of MCF-7 cells, however, the time of incubation with 73 ± 14 μg/mL of water extract from *Lentinus edodes* took 48 h. It is known that whole mushroom extracts contain a lot of substances other than polysaccharides with very well documented direct cytotoxic activities [64].

**Figure 4.** Cytotoxic effect of water soluble polysaccharides (WSP) from blanched, boiled, blanched and fermented *Lentinus edodes* on MCF-7 (**a**) and T47D (**b**) cell lines.

In the cell line T47D (Figure 4b), we have assessed no significant sensitivity to polysaccharides fractions from processed and control (*Lentinus edodes* fruiting bodies) mushrooms. Viability of T47D cells treated with maximal WSP concentration was consecutively 87.2 ± 4.5% for blanched, 84.9 ± 0.6% for blanched with further fermentation and 83.7 ± 1.8% for control sample. Interestingly boiling in water have led to complete loss of WSP antiproliferative activity (cells viability was 97.0 ± 0.5%).

These observations are consistent with the results of previous studies [23] carried out on *Pleurotus ostreatus* fungus where it was stated that hydro-thermal processing of mushrooms cause the loss of direct cytotoxic activity of mushroom derived polysaccharides. Similarly other authors reported that different ways of mushrooms conservation treatment and cooking affect the composition of nutritional values and their antioxidant properties [18]. To sum up, from both studied breast cancer cell lines MCF-7 and T47D, the first one seemed to be more sensitive to water soluble polysaccharides. The strongest antiproliferative activity to MCF-7 cells was stated for water soluble polysaccharides from non-processed *Lentinus edodes* fruiting bodies.

#### **4. Conclusions**

Mushrooms fruiting bodies are rarely eaten raw and certain processing needs to be applied prior the consumption. The main purpose of the present paper was to determine the effect of three processing methods on the content and biological activity of water soluble polysaccharides. The study has shown that boiling and fermenting with lactic acid bacteria may lead to small decrease in the amount of polysaccharides. Moreover, all the techniques, namely blanching, boiling and fermenting cause the decrease in both antioxidant and antiproliferative capacities. However, these processes do not diminish biological activity completely and polysaccharides still retain their properties to some extent. A natural progression of this work would be to verify the impact of different processing methods e.g., baking, frying or microwaves. In addition, further studies should be conducted to understand mechanisms beyond antiproliferative activities of polysaccharides and residues which are attached to them.

**Author Contributions:** Conceptualization, M.Z.-S., W.R. and A.K.; Investigation, M.Z.-S., W.R., J.N., J.T., E.J.-R., A.S. and K.S.; Methodology, M.Z.-S., W.R. and J.N.; Software, W.R.; Writing—original draft, M.Z.-S. and W.R.; Writing—review & editing, A.B.-K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflicts of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Review* **Formulation Strategies to Improve Oral Bioavailability of Ellagic Acid**

**Guendalina Zuccari \*, Sara Baldassari, Giorgia Ailuno, Federica Turrini, Silvana Alfei and Gabriele Caviglioli**

Department of Pharmacy, Università di Genova, 16147 Genova, Italy; baldassari@difar.unige.it (S.B.); ailuno@difar.unige.it (G.A.); turrini@difar.unige.it (F.T.); alfei@difar.unige.it (S.A.); caviglioli@difar.unige.it (G.C.)

**\*** Correspondence: zuccari@difar.unige.it; Tel.: +39-010-3352627

Received: 6 April 2020; Accepted: 8 May 2020; Published: 12 May 2020

#### **Featured Application: An updated description of pursued approaches for e**ffi**ciently resolving the low bioavailability issue of ellagic acid.**

**Abstract:** Ellagic acid, a polyphenolic compound present in fruit and berries, has recently been the object of extensive research for its antioxidant activity, which might be useful for the prevention and treatment of cancer, cardiovascular pathologies, and neurodegenerative disorders. Its protective role justifies numerous attempts to include it in functional food preparations and in dietary supplements, and not only to limit the unpleasant collateral effects of chemotherapy. However, ellagic acid use as a chemopreventive agent has been debated because of its poor bioavailability associated with low solubility, limited permeability, first pass effect, and interindividual variability in gut microbial transformations. To overcome these drawbacks, various strategies for oral administration including solid dispersions, micro and nanoparticles, inclusion complexes, self-emulsifying systems, and polymorphs were proposed. Here, we listed an updated description of pursued micro and nanotechnological approaches focusing on the fabrication processes and the features of the obtained products, as well as on the positive results yielded by in vitro and in vivo studies in comparison to the raw material. The micro and nanosized formulations here described might be exploited for pharmaceutical delivery of this active, as well as for the production of nutritional supplements or for the enrichment of novel foods.

**Keywords:** ellagic acid; oral administration; bioavailability; microformulations; nanoformulations; solubility enhancement

#### **1. Introduction**

Pomegranate, *Punica granatum*, is well-known as a traditional medicinal fruit mentioned in the Old Testament of the Bible, the Qur'an, the Jewish Torah, and the Babylonian Talmud as a sacred fruit harbinger of fertility, abundance, and luck. Historically and currently, pomegranate has been used for various purposes [1]. The bioactive compound mainly responsible for the health effects of pomegranate is ellagic acid (EA), as it represents one of the most potent dietary antioxidants.

EA is a polyphenol compound which derives from ellagitannins (ET), a family of molecules in which hexahydroxydyphenic acid residues are esterified with glucose or quinic acid. Following a hydrolytic process, the hexahydroxydyphenic acid group is released, dehydrates and spontaneously lactonizes, forming EA. Punicalagin isomers are the most representative of total tannins extracted from pomegranate (Figure 1).

**Figure 1.** Hydrolysis of punicalagin (a pomegranate ellagitannin) to produce ellagic acid.

Present also in a large variety of tropical fruit, nuts, berries (strawberries, raspberries, blackberries, cranberries, goji berries) and in a type of edible mushroom (Fistulina epatica), EA has recently attracted deep interest, since many research studies showed the role of oxidative stress in different pathologic conditions such as cancer, diabetes, cardiovascular and autoimmune diseases, obesity, and neurodegenerative disorders [2–7]. In this regard, EA may be a useful adjunct in Parkinson's disease treatment, as it proved to protect dopamine from oxidation and, by reducing inflammation and oxidative stress, to exert a neuroprotective effect [8]. Besides being able to reduce the pathological levels of NF-kB, EA may play a beneficial role in diseases associated with CNS inflammation, such as multiple sclerosis. EA has also been proposed as antidepressant and anxiolytic drug for its capability to modulate the monoaminergic system and to increase the endogenous levels of brain-derived neurotrophic factors [6]. In the last decade, several studies investigated EA effects on many types of cancer, showing its ability to arrest cell cycle progression, modulate pathways linked to cell viability, and inhibit angiogenesis via inactivating metalloproteinases [9–18]. For example, EA resulted in inducing apoptosis and modulating gene expression in different colon cancer cell lines at the concentrations achievable in the intestinal lumen from the diet, suggesting a potential role in cancer chemoprevention [19]. Moreover, EA may counterbalance the dangerous side effects of antitumoral drugs such as cisplatin, doxorubicin, and cyclosporine [20–22]. In addition, EA proved to possess antimicrobial properties inhibiting the growth of methicillin-resistant Staphylococcus aureus and Salmonella [23]. These properties were explained through the ability of EA of either coupling with proteins of the bacterial wall or inhibiting gyrase activity, which cleaves the DNA strand during bacterial replication. Furthermore, EA seemed to counteract the HIV-1 virus replication, by binding envelope proteins and inhibiting reverse transcriptase [24]. Pomegranate extract has been traditionally used for treating gastrointestinal bleeding and various types of wounds, due to EA's ability to promote blood coagulation by the activation of the Hageman factor (Factor XII) [25]. Moreover, preliminary results suggest potential anti-inflammatory properties [26,27] and an important role in prevention of cardiovascular diseases [28–30]. Finally, EA finds application also in cosmetic and nutraceutical industries, mainly for antiaging and prophylactic purposes respectively [31–33]. An overview of EA activities is shown in Figure 2.

However, though EA has the potential to become an efficacious agent in the prevention and therapy of several diseases, as supported by several papers present in the literature, the working efficacy of EA dietary and therapeutic formulations have been mainly examined at preclinical level. To date, few clinical studies have been performed to evaluate EA beneficial effects in humans, with often a limited number of patients, and most of them concern the administration of pomegranate juice or extract in cancer patient diet or deal with the evaluation of the enhancement of cognitive/functional recovery after stroke [34,35]. It was suggested that EA'S low water solubility and rapid metabolism might hinder the progress towards translational research. In our opinion, further studies focused on the application of micro and nanotechnology could provide more encouraging data, opening the path to new strategies able to overcome EA's poor biopharmaceutical characteristics. Hence, the goal of this review is to provide to scientists, embarking on the research involving EA, a scenario of all pursued formulation attempts. Therefore, starting from a description of EA bioavailability and solubility,

a summary of micro and nanosystem-based strategies suitable for improving orally administered EA efficacy is provided.

#### **2. EA Chemical Structure and Solubility**

EA is a chromene-dione derivative (2,3,7,8-tetrahydroxy-chromeno [5,4,3-cde]chromene-5,10-dione), encompassing both a hydrophilic moiety with 4 hydroxyl groups and 2 lactone groups, and a planar lipophilic moiety with 2 biphenyl rings. This particular structure has both hydrogen bonding acceptor (lactone) and donor (–OH) sites (Figure 1). Due to the weak acidic nature of its four phenolic groups (pKa1 = 5.6 at 37 ◦C), around neutral pH it is mainly deprotonated on positions 8 and 8 , while above pH 9.6 lactone rings open to give a carboxyl derivative [36]. EA low oral bioavailability is mostly due to its poor water solubility (9.7 μg/mL), which increases with pH, as well as the antioxidant action [37]. However, in basic solutions, phenolic compounds lack stability as these molecules, under ionic form, undergo extensive transformations or are converted to quinones as a result of oxidation. A stability study on pomegranate fruit peel extract demonstrated that EA content significantly decreases in few weeks regardless of the pH of the solution, due to the hydrolysis of the ester group with hexahydroxydiphenic acid formation, suggesting that EA should not be stored in aqueous medium [38]; this aspect reinforces the need to develop novel systems also for EA stabilization.

Prior to facing the route of the development of a new delivery system, the knowledge of solvents, co-solvents and substances in which EA could be consistently soluble and stable represents an indispensable step for designing a good formulation. In this regard, with the aim of providing a useful guidance for further research studies, Table 1 gathers published data concerning EA solubility. Concerning organic solvents, EA is slightly soluble in methanol, soluble in DMSO and shows maximum solubility in N-methyl-2-pyrrolidone (NMP), confirming the effect of basic pH on EA dissolution. An analogous trend is also observable in aqueous solutions, where results implied that the solubility of EA depends on the pH values of the media. While EA is almost insoluble in acidic media and distilled water, its water solubility is significantly improved by basic pH. As highlighted further ahead, one of the mostly exploited vehicles is polyethylene glycol (PEG) 400, as it is endowed with satisfactory biocompatibility and, at the same time, is miscible with both aqueous and organic solvents. EA solubility in oils and surfactants is also provided, helpful for developing emulsifying-based techniques.

#### **3. EA Dietary Assumption**

In foods and beverages, EA is present in several forms, such as unmodified, glycosylated and/or acetylated or inside the structure of hydrolysable ET polymers, usually esterified with glucose moieties. However, only a small fraction of free EA is absorbed in the stomach, since ET are resistant to acidic pH. ET hydrolysis occurs in the small intestine, leading to the release of EA, which can be absorbed mainly by passive diffusion, although an in vitro experiment on Caco-2 cells monolayer model suggested the involvement of a protein-mediated transport [39]. Once reached the systemic circulation, EA undergoes a massive first pass effect, being transformed in methyl esters, dimethyl esters or glucuronides, detectable in human plasma from 1 to 5 h after ET ingestion [40]. Meanwhile, unabsorbed ET and EA fractions are mostly converted to a family of metabolites called urolithins by colon microflora. Urolithins contain a common structure of dibenzopyranone and are more lipophilic, showing higher absorption rate across colon epithelium compared to EA, thus resulting 25–80 times more bioavailable [41]. Naturally, there is huge variability in microbial metabolism of EA among individuals depending on differences in gut microbiota composition. According to the metabotype (M-0, M-A and M-B), humans may produce no urolithins, highly active urolithins or less active urolithin, so that EA consumption may not exert equal health benefits in all subjects [42–44].

EA oral bioavailability was studied in human subjects after 180 mL of pomegranate juice consumption. In this study, the maximum EA blood concentration (Cmax) was 32 ng/mL which was rapidly metabolized in the next 4 h, confirming the extremely poor intestinal absorption [41]. Other pharmacokinetic studies supported the hypothesis that higher EA-to-ET ratio in the oral dose

corresponds to higher EA plasma concentrations, but, intriguingly, no enhancement in bioavailability was observed increasing the dose of free EA up to 524 mg, suggesting that a saturation condition in the absorption process was achieved [45]. In addition to these unfavorable conditions, it must be considered that polyphenols are not consumed in sufficient quantities among population in Western Europe and developing countries, because of inadequate fruit and vegetable intake, though an increased awareness of the benefits of a balanced diet is fostering the assumption of dietary supplements [46].

#### **4. Formulation Strategies for Improving EA Oral Bioavailability**

The Biopharmaceutics Classification System (BCS) is a scientific framework for classifying drug substances based on their aqueous solubility and intestinal permeability. It is a drug-development tool that divides drugs into high/low solubility and permeability classes. Based on its low aqueous solubility and low permeability (apparent permeability coefficient = 0.13 <sup>×</sup> 10−<sup>6</sup> cm/s [47]), EA is classified as class IV drug according to the BCS, i.e., endowed with low solubility and low permeability, features that limit a lot its clinical application. Micro and nanotechnology approaches widely demonstrated great potential in modifying pharmacokinetic, bioavailability and stability of several drugs, including phytochemicals used in cancer chemoprevention or in dietary supplements, contributing to preserve the properties of the active ingredients or to mask bad tastes and odors [48–50]. From a formulation point of view, a good EA dosage form would grant easier handling, transport and storage, increase its light resistance, hamper undesirable reactions such as oxidation and hydrolysis, and enhance solubility.

Focusing on preparative methods, features and biological results of the developed EA-loaded formulations, an exhaustive description of the employed strategies is thereafter presented. The various approaches were divided according to the kind of the dosage form and reported in Tables 2 and 3 depending on the micro or nanosize of the final product. Moreover, in the final paragraph, examples of fixed combinations containing EA were also described and discussed.

#### *4.1. Micronized EA (m-EA)*

According to Noyes-Whitney law, an effective way to increase dissolution rate and oral absorption of a substance is the size reduction of powder particles [51]. For this purpose, anti-solvent precipitation may represent a valid processing technique. In comparison to more popular micro and nanocarriers that will be discussed in the following sections, this approach presents several advantages, such as easy realization and rapidity. In addition, the process allows the reduction of organic solvents use and the one-step processing favors scalability. The anti-solvent precipitation of a poorly water-soluble substance is simply obtained by dissolving it in a solvent, followed by rapid mixing with a solvent-miscible anti-solvent. After mixing, the substance precipitates in micro-sized particles, since a supersaturated solution forms. The critical step of this method is to produce crystal nuclei rapidly, avoiding their excessive growth. For process optimization, precipitation time and temperature, addition speed of the solution to the anti-solvent, their volume ratio, drug concentration and stirring intensity are the keys parameters that have to be checked.

Li et al. set the best conditions to produce m-EA by anti-solvent precipitation with an accurate preformulative study [52]. An EA solution in NMP (30 mg/mL) was added to water, used as anti-solvent, at a rate of 30 mL/h with 2 min precipitation time, at 3 ◦C and 2500 rpm stirring. Then, the precipitate was dispersed in a maltodextrin aqueous solution to have a solid concentration of 1 mg/mL and was lyophilized. The authors verified that residual NMP (class 2 residual solvent) in the final product was in agreement with the established International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH) limit [53]. The m-EA freeze-dried powder, whose mean particle diameter was 428 nm, in a comparison to raw EA and to a physical mixture of EA in maltodextrin, showed the fastest dissolution rate. Surely, maltodextrin acted as hydrotrope, but the result could be attributed exclusively to the high specific surface of the lyophilized powder and to the low crystallinity of the incorporated EA. All in all EA water dissolution was improved up to 11.67 μg/mL at 37 ◦C. In addition, the 2, 2-diphenyl-1-1-pycrylhydrazyl (DPPH)-scavenging activity was higher than that of not-micronized EA and oral bioavailability in rats increased twice in comparison to raw EA.

In another study, Beshbishy et al. obtained m-EA using the anti-solvent precipitation method by injecting into deionized water a 10 mg/mL ethanolic solution of EA at a fixed flow rate [54]. m-EA particles, collected by filtration and vacuum dried, showed modified morphology with respect to the parent material and appeared as needle-shaped or rod-like structures. In vitro and in vivo experiments, performed by intraperitoneal injection in mice, showed a significantly improved m-EA activity against blood parasites as Babesia and Theileria. However, the authors concluded that EA activity requires further investigations to understand whether EA inhibits parasites grow or causes their death.

A similar approach was performed using supercritical CO2 as anti-solvent. In this procedure a solution of EA in NMP was sprayed into a vessel where a CO2 stream was continuously flowed under supercritical conditions. Solution atomization led to a fast mass transfer between supercritical CO2 and the solution itself, which caused supersaturation and subsequent rapid precipitation of EA in the form of micro-sized particles [55]. The study focused on the influence of process parameters and EA concentration on size, polydispersity index (PDI) and morphology. Moreover, a co-precipitation with Eudragit® L 100 was also carried out and, starting from EA:Eudragit® 1:1 weight ratio, a drug loading (DL) of 49% was achieved. Experimental results showed that EA concentration should be, if possible, close to saturation, thus allowing obtaining particles in the form of sticks and flowers and with a mean diameter of about 3 μm. Also, in this case residual NMP was far lower than the mandatory established limit. m-EA showed higher dissolution rate with respect to raw EA due to the small particle size, reaching higher concentrations even at pH 1.2. Finally, the coprecipitates with Eudragit® showed the fastest dissolution rates, probably because the dispersion of EA within the polymer led to smaller particles and, surprisingly, the presence of the polymer did not delay the pomegranate acid delivery, possibly since EA particles were obtained in crystalline form, confirming that the product was a solid dispersion in which the pomegranate acid particles were not encapsulated.

#### *4.2. EA in Spray Dried and Lyophilized Powders*

Solid dispersions are semi-crystalline or amorphous dispersions (ASDs) of a molecule in an inert matrix. ASDs represent a valid alternative for the development of microsystems entrapping poorly water-soluble molecules. ASDs can be prepared from a solution by spray drying, freeze drying, co-precipitation, rotary evaporation, film casting, or hot-melt extrusion. In all cases, a suitable polymeric excipient should have functional groups able to form H-bonds with EA, thus allowing the formation of metastable amorphous EA dispersions with adequate stability over time.

In this context, a first attempt was made with maltodextrins, a saccharide-based excipient extensively used in food industry. Using a freeze drying technique, maltodextrin DE5-8 or DE18.5 were applied to form a solid network around a cloudberry phenolic extract [56]. The amorphous matrices were prepared starting from a maltodextrin (9% *w*/*w*) and a cloudberry extract (1% *w*/*w*) aqueous solution, mixed under slight heating for about 0.5 h to avoid ET degradation. DE5-8 complexed phenolics in a more efficient way, rationally due to its higher molecular weight. The microencapsulation improved the extract storage stability to hydrolysis, and also in this case the improvement contribution of DE5-8 was better than that of DE18.5. Microencapsulation always resulted in a decrease of EA formation by ET hydrolysis during storage, but the stability was strongly dependent on relative vapor pressure (RVP). At 66% RVP and 25 ◦C maltodextrin DE5-8 protected phenolics up to 32 days, then the degradation reactions occurred in a major extent. On the contrary, the antioxidant properties did not alter during storage either in the extract alone or in the presence of DE5-8, probably because the loss of original phenolics was compensated by new formed molecules with equal antioxidant activity.

In another study, polymers as the hydrophilic hydroxypropylmethyl cellulose acetate succinate (HPMCAS), the rather hydrophobic carboxymethyl cellulose acetate butyrate (CMCAB), and the very hydrophobic cellulose acetate adipate propionate (CAAdP) were compared with polyvinylpyrrolidone (PVP), taken as reference excipient being widely used in ASD formulations [57]. Mixtures of EA and

PVP, CMCAB or HPMCAS at different weight ratios (1:9 and 1:3), dissolved in acetone:ethanol (1:4 *v*/*v*), were spray dried with yields of 50%–60%. ASD from CAAdP were instead prepared by co-precipitation method, due to the limited quantity of polymer available. EA was amorphous up to 25% *w*/*w* in solid dispersions with PVP and HPMCAS, and up to 10% with CMCAB and CAAdP. Consequently, EA-PVP and EA-HPMCAS reached the highest solution concentration of 1500 and 280 μg/mL, respectively. In a stability study performed in aqueous buffer (pH 6.8) for 24 h at room temperature (rt), pure EA showed a degradation degree by lactone ring opening of 20%, while lower values—16% for CMCAB, 14% for CAAdP, 6% for PVP and just 5% for HPMCAS—were found in the ASD samples. Dissolution rate at pH 6.8 was remarkably superior to that of pure EA, both from PVP and HPMCAS ASDs, while from CAAdP or CMCAB ASDs it was very slow and did not achieve adequate EA concentration. As expected, under acidic conditions EA release from PVP was quite fast, whereas from cellulose esters it was minimal, but this advantage was nullified by crystallization of a large amount of the EA released from PVP. Since the 1:9 ASDs with HPMCAS effectively stabilized EA against crystallization and degradation, the authors concluded that this could be considered the most promising formulation.

To prepare an ingredient for functional foods, a fine microdispersed powder containing EA was obtained by spray drying using low methoxyl pectin, characterized by high biocompatibility, good palatability, taste, and prebiotic properties [58]. Pectin represents a common natural food additive (E440) authorized in several food categories [59]. By varying EA:pectin weight ratio, different formulations were prepared. The best microdispersed powder, obtained with 1:4.5 EA:pectin ratio, showed a mean particle diameter of about 10 μm and a DL equal to 21%, allowing solubilizing 63 μg/mL of EA in water. Moreover, the drug loading of EA in the powder remained stable for one year of storage in a desiccator at 25 ◦C.

Recently, the search for promising systems suitable for dietary supplements enlarged the research field to other natural polymers, such as alginic acid (ALG) [60]. EA-ALG nanoparticles (NPs) were prepared by dissolving both EA and ALG at the concentration of 250 μg/mL in an aqueous medium at pH 8.5. The solution was spray dried and the obtained powder was added to a CaCl2 solution in order to induce reticulation by ionic gelation. The resulting NPs had a mean diameter of 670 nm with zeta potential around −27 mV. Thanks to the high degree of EA solubilization in the basic medium, the encapsulation efficiency (EE) was 50%, much higher than the EE of NPs prepared from a neutral medium. As rational remark to this procedure performed in basic environment, it may be highlighted that the possible degradation of EA in alkaline medium during preparation was not taken into consideration. This EA-ALG system showed a biphasic release profile with a fast phase during the initial 3 h, justified by EA adsorbed on NPs surface, followed by a sustained, complete release until 8 h. A mice model of epilepsy was used to assess in vivo EA-ALG NPs efficacy as anticonvulsant and neuroprotective agent by oral treatment every other day for 33 days. Results indicated that EA contained in this dosage form prevented seizures during the experimental period, and had a more pronounced effect, compared to free EA, by reducing oxidative stress damage, inhibiting apoptosis and down-regulating cytokine levels.

#### *4.3. Inclusion Complexes*

#### 4.3.1. EA Inclusion in Cyclodextrins (CDs)

CDs are able to form inclusion complexes with a variety of molecules by hosting a guest drug with suitable polarity and dimension inside their lipophilic cavity, thus modifying its physico-chemical behavior. Among CDs, hydroxypropyl-β-CD (HP-β-CD) is characterized by better inclusion ability and higher complex solubility. To investigate the feasibility and the stoichiometry of inclusion complex formation of EA in HP-β-CD, phase solubility studies were carried out in water at 30 ◦C for 72 h to reach the equilibrium [61]. The obtained solubility curve of EA showed a linear increase with HP-β-CD concentration up to 12 mM, followed by a negative deviation from linearity up to 24 mM. This trend belongs to AN type and was explained through the formation of a soluble 1:2

EA:HP-β-CD inclusion complex with stability constants K1:1 and K1:2 of 201.61 and 18.91 M-1, respectively. A 15 mM HP-β-CD solution allowed to solubilize 54.40 μg/mL of EA in water at 30 ◦C. Subsequently, the inclusion complex was prepared by freeze drying a solution of EA:HP-β-CD in 1:2 molar ratio upon its filtration. In a dissolution test performed at pH 6.8 and 37 ◦C, HP-β-CD released about 55% of the loaded drug in 15 min, more than 5-fold the amount dissolved from pure EA powder. Furthermore, the anti-inflammatory effect of complexed EA was tested in a rat model of inflammation. Oral treatment with 10 or 20 mg/kg EA-HP-β-CD resulted more efficacious than the administration of EA alone or of 10 mg/kg indomethacin as indicated by the measure of paw edema volume. The feasibility of using β-CDs as complexing agent for EA was also investigated with less satisfactory results [62]. The complex was less soluble, leading to a total EA concentration up to 39.14 μg/mL, lower than that obtained in the previous study with HP-β-CD, due to the lower β-CD solubility.

By following a different preparative procedure, β-CD was again chosen as EA complexing agent in another study [63]. A saturated aqueous β-CD solution was slowly mixed with 125 mg/mL of EA ethanolic solution and stirred for 2 h at rt, then the mixture was filtered and finally freeze-dried. As highlighted by SEM images, the obtained powder was different from those of the pure components in terms of morphology and dimensions; EE was 35% and in vitro dissolution studies (pH from 6.8 to 7.4) showed a multiphase release profile: a little amount of EA (3%) was released in a slow initial stage during the first 4–8 h, followed by a rapid phase between 8 and 24 h with 73% of drug released, and by a final slow phase. The effect of EA-β-CD complexes on proliferation of human liver carcinoma cells was evaluated by MTT assay and, though the lack of data relating to the treatment of cells with EA alone do not allow a comparison, a remarkable inhibition was observed.

Another study to increase EA solubility regarded EA complexation with β-CD nanosponges (NS) [64]. NS are hyper-cross-linked CDs obtainable either from a mixture of CDs, or CDs conjugated with relevant amounts of linear dextrin. The cross-linking leads to a cage-like structure, where CDs cavities and nanochannels are strictly connected to form a porous network. Consequently, NS are able to incorporate drugs both as inclusion and non-inclusion complexes, thus improving the overall solubilizing ability of the starting CD. In this work the β-CD NS were obtained using dimethyl carbonate as a cross-linker. By a solubility study the authors demonstrated that the β-CD NS solubilized EA in water up to 49.79 μg/mL. The mean particle size of EA-β-CD NS was about 423 nm, PDI was found of 0.409 and the zeta potential equal to −34 mV, thus sufficiently high to produce a stable suspension. In this regard, it has to be remembered that very low Z-potential values (ζ < 5 mV) usually translates in particles with high tendency to agglomerate that means physical instability in solution. On the contrary, increasing the charge associated with the particle i.e., ζ ± 30 mV or higher, usually prevents particles sticking after collision [65,66].

Moreover, EE was around 69% and, as highlighted by X-ray diffraction, in the polymeric structure EA was present either in the amorphous form or as a solid-state solution. The release profiles showed that only about 20% of loaded EA was released after 24 h, suggesting strong retention of EA in the NS network. A pharmacokinetic study in rabbits indicated that oral administration of EA-β-CD NS increased EA bioavailability by more than 2-fold if compared to the free EA suspension.

In a more recent work, a useful analysis of the preparation processes and experimental conditions for obtaining inclusion complexes was reported [67]. Three inclusion methods were considered, i.e., stirring-ultrasonic, ultrasonic, and grinding. In the first method, which led to the highest EE (84%), EA, dissolved in a small amount of ethanol, was added by slowly dripping into HP-β-CD aqueous solution under stirring, then the mixture was sonicated for 20 min, and finally kept under stirring at rt for further 18 h. The suspension was filtered, concentrated with rotary evaporator and vacuum dried to obtain the inclusion compound. Process conditions were optimized by an orthogonal test design, in which EA:HP-β-CD molar ratio, ethanol concentration, inclusion time and temperature were considered to be experimental parameters. The experimental design showed that EA:HP-β-CD

molar ratio of 1:2, ethanol concentration of 60%, inclusion time of 36 h, and inclusion temperature of 20 ◦C were the best conditions.

Recently, Gontijo et al. failed in realizing an effective complexation in HP-β-CD as the first study did, possibly because in this case the phase solubility studies were performed at rt instead of 30 ◦C [68]. In fact, a linear increase in EA solubilization was achieved only in alkaline solution, where raising HP-β-CD concentrations up to 300 mM led to a total EA concentration only 2.8-times higher than that without CD. However, the low stability constant of the complex (K1:1 = 4.5 M-1) resulted in a low complexation efficiency with an EA:CD stoichiometric ratio equal to 1:45. Concerning the procedure, the inclusion complex was prepared by spray drying a suspension of EA:CD 1:1 molar ratio in 0.1 M NaOH solution. Furthermore, also methyl-β-CD was used without success, since no complexing ability was found either in water or in basic solution. Finally, Caco-2 cells were employed in an in vitro model of invasive Candida albicans infection to test the bioactivity of EA complexes, but no statistically significant difference was observed in terms of antifungal activity between free or complexed EA, suggesting that complexation apparently did not improve antimycotic efficacy.

#### 4.3.2. EA Inclusion in Metalla-Cages

A recent strategy for allowing EA administration in a more bioavailable form consisted of a self-assembled metallo-supramolecular architecture able to host EA inside three-dimensional cages [69]. These vectors, engineered via coordination-driven self-assembly, present several advantages. In particular, the metal-guest interactions are very strong and highly directed, leading to the formation of a stable rigid complex. In this regard, EA was encapsulated in arene-ruthelium metalla-prism cages with a yield of about 90% and its anticancer activity was investigated. The cytotoxic effect against A549 human lung cancer cells was significant, overall considering that free EA lacked any activity. In an attempt to investigate if EA effect was macrophage-mediated, a study on RAW264.7 murine leukemia monocytes was performed, highlighting that EA encapsulated in metalla-cages exhibited an inhibitory effect for cancer cells via G-CSF induction and Rantes inhibition at both mRNA and protein levels.

#### *4.4. EA Encapsulated in Polymeric Carriers*

#### 4.4.1. Eudragit® Microspheres

EA was proposed as a prophylactic agent for the treatment of inflammatory bowel disease. Jeong et al. developed microspheres for EA site-specific delivery to the ileum using Eudragit® P-4135, a copolymer of methacrylic acid, methyl acrylate and methyl methacrylate able to dissolve at pH > 7.2, which corresponds to the pH of the distal part of the small intestine [70]. The adopted preparation method was a solvent evaporation process in oil phase using an acetone/light liquid paraffin system. Different polymer amounts, ranging from 0.5 to 1.5 g, were dissolved in acetone and mixed with 0.5 mg of EA, previously dispersed in the same solvent and sonicated. Then, the resultant solution was added to light liquid paraffin containing Span 80 (3% *v*/*v*), and the emulsion was stirred up to complete evaporation of acetone. The microspheres, harvested by filtration and washed with hexane to remove paraffin, yielded an EE ranging from 57 to 99% and a DL from 33 to 36%. Increasing polymer amounts resulted in higher EE and size increase from 98 to 168 μm. As expected, EA release was pH-dependent and at pH 6.8 an initial burst effect of about 25% was followed by a sustained release, with approx. 40% of loaded drug released after 6 h. In this case, the burst effect would not be critical, as it occurred at the target site, as demonstrated by the negligible EA in vitro release at more acidic pH. For in vivo studies fluorescein was added to the preparative mixture to accurately measure microsphere dissolution rate. The obtained microspheres were orally administered to rats and blood samples were collected every hour. The comparison between collected images of microspheres in small intestine of euthanized rats and fluorescein plasma concentration evidenced that the plasmatic peak occurred 3 h after administration, when microspheres reached the ileocaecal junction. Therefore, this Eudragit®-based system represents an effective vehicle to target EA release to the terminal ileum,

providing enhanced absorption in the region and thus preventing the variability connected to EA conversion into urolithins.

#### 4.4.2. Poly (Lactic-Co-Glycolic Acid) (PLGA) and Poly (ε-Caprolactone) (PCL) Nanospheres

Among the various encapsulating polymers, PLGA has been extensively used due to its biocompatibility, biodegradability, and ability to protect the drug from degradation and in obtaining sustained drug release [71]. To increase EA bioavailability, PLGA-based formulations suitable for oral administration were developed [72]. The particles were prepared by using the emulsion-diffusion-evaporation technique. PLGA was dissolved in ethyl acetate under stirring for 2 h, then EA, previously dissolved in PEG 400, was added, and a fine EA dispersion was obtained. The dispersion was emulsified in an aqueous phase containing a stabilizer (1% *w*/*v*), such as polyvinyl alcohol (PVA), PVA:chitosan (CS) in 80:20 ratio, or didodecyldimethylammonium bromide (DMAB). After with stirred the emulsion at 1000 rpm for 3 h, homogenization at 15,000 rpm for 5 min was performed to obtain a more homogeneous and smaller size distribution. Then, the addition of water caused the diffusion of the organic solvent into the aqueous phase, and the evaporation of the organic phase by heating led to EA-polymer precipitation. The use of DMAB yielded smaller particles, while the other two stabilizers led to bigger sizes and higher PDI, probably due to the increased viscosity and irregular entanglement associated with the long polymeric chains. The zeta potential was highly positive for NPs containing DMAB (about +75 mV), equal to +25 mV for the ones with PVA-CS, and slightly negative for those with PVA. The highest EE (55%) was obtained with PVA-CS, the lowest with DMAB (42%). Concerning EA release at pH 7.4, all products exhibited a rapid initial release followed by a slower sustained phase, with PVA NPs being the fastest, because PVA hydrophilic groups favor water penetration into the NPs compared to the more lipophilic DMAB. However, the release study was not performed simulating the pH conditions of the whole gastrointestinal tract. To study the permeability of NPs, an in situ intestinal study was performed on rats based on the close loop method, even though the disappearance of drug from the luminal side, used for determining permeation, not always correlates with the rate of absorption of the drug in the systemic circulation. EA-DMAB NPs showed the highest uptake (87%) in accordance with their smaller size and higher hydrophobicity. The antioxidant activity was assessed in cell free system and in yeast cell culture model, confirming the good scavenging activity of the encapsulated pomegranate acid. The same research group also investigated the feasibility of EA-loaded PCL NPs in comparison to PLGA NPs using the same emulsion-diffusion-evaporation method, but with an important modification: during the primary emulsion step EA underwent a rapid precipitation that caused its low entrapment, so the authors added the stabilizer to the polymer solution to increase EA solubility [22]. As expected, EE increased from 42% to 52% for DMAB-PLGA NPs and from 55% to 62% for PVA-PLGA NPs, not much different from EE of DMAB-PCL and PVA-PCL NPs (47% and 57% respectively). Since PCL is a slower degrading polymer compared to PLGA, its release profiles yielded slower release rates. The antioxidant potential of EA-DMAB NPs was assessed by evaluating their protective action against cyclosporine A-induced nephrotoxicity in rats: both PLGA- and PCL-based NPs were 3-fold more effective than free EA in kidney protection, thus indicating the improved bioavailability of the encapsulated EA.

Unfortunately, PLGA NPs major drawback is rapid opsonization, so that their systemic circulation time is too short. To address this issue, EA was encapsulated in PLGA NPs grafted with poly (ethylene glycol) (PEG) chains to escape to the reticuloendothelial system [73]. In a proof of principle study, NPs were prepared by the double emulsion-solvent evaporation method. Two dichloromethane stock solutions, containing PLGA-PEG and EA respectively, were mixed, added to 200 μL of phosphate-buffered saline solution (PBS) and sonicated. This primary *w*/*o* emulsion was emulsified in PVA 1% *w*/*v* aqueous solution by sonication. Subsequently, this *w*/*o*/*w* emulsion was first diluted with PVA 1%, then heated for dichloromethane evaporation and finally the obtained NP were dialyzed and lyophilized. The NP showed a diameter of 175 nm with PDI of 0.14. In vitro biological studies

on MCF-7 and Hs578T breast cancer cell lines proved a significant activity enhancement of EA in the encapsulated form, being its IC50 values more than 2-fold lower than those of the free form.

In a further attempt to curb the phagocytic activity, a combination of CS and PEG was considered to be surface coating material for EA-PLGA NPs [74]. The NPs were prepared by o/w single emulsion-solvent evaporation method as follows. EA and PLGA were separately dissolved in chloroform and emulsified with PVA 2% *w*/*v* aqueous solution containing 12% *w*/*w* CS and 5% *w*/*w* PEG 2 kDa through sonication. The emulsion was stirred overnight up to complete evaporation of the organic solvent. The coated CS-PEG-PLGA NPs were compared to PLGA NPs and PEG-PLGA NPs prepared in the same way. EE was always about 65%, while the mean hydrodynamic diameter was 220 nm for PLGA NPs and CS-PLGA NPs, and 255 nm for CS-PEG-PLGA NPs. As expected, positive charges were present on the surface of CS-PLGA NPs and CS-PEG-PLGA NPs (ζ = +27 and ζ = +38 mV, respectively); on the contrary, a negative zeta potential characterized PLGA NPs (ζ = −10 mV). Release studies, carried out at pH 7.4, showed an initial faster release of about 35% during the first 10 h, followed by a more controlled release phase, with further 55% of drug released in 24 h. In vitro anticancer activity against HepG2 hepatocellular carcinoma cells and HCT 116 colorectal cancer cells revealed a significant cytotoxic effect of EA loaded in CS-PEG-PLGA NP vs. free EA at every tested dosage (up to 100 μM), suggesting an efficient NPs uptake by the cells. Also Mady et al. used PCL for successful entrapment of EA through the emulsion-evaporation method [75]. In this case, EA, dissolved in PEG 200, was mixed with PCL, solubilized in ethyl acetate, and the mixture was emulsified by homogenization with an aqueous solution containing a stabilizer (PVA, DMAB or poloxamer 407 ranging from 0.1% to 1% *w*/*v*). The resulting *o*/*w* emulsion was sonicated for 1 min, then diluted in water and kept under stirring for 6 h up to complete evaporation of the organic solvent. The obtained NPs were centrifuged, washed with distilled water to remove unentrapped EA, and freeze-dried. As previously observed, 1% DMAB yielded the smallest mean NP size (193 nm). In fact, by creating a positive charge layer around the internal phase, DMAB provided electrostatic repulsion, while PVA and poloxamer, being uncharged, induced only a steric stabilizing effect. The type of stabilizer affected also EE and DL. DMAB gave the lowest EE (66.4%–73.5%) and DL (56.8%–63.9%), whereas PVA provided the highest EE (78.3%–90.3%) and DL (68.1%–78.6%). These differences in EA entrapment were probably linked to the unequal EA aqueous dissolution granted by the stabilizer. Release studies at pH 7.4 illustrated the typical two-step profile, with an initial fast release driven by diffusion for the first 12 h, and a second slow release due to the retarded diffusion of EA from the NPs inner layers. Finally, the release of EA from DMAB-PCL NPs reached up to 48% of the loaded amount in 8 days. The anticancer efficacy of EA-PCL NPs tested on HCT-116 colon adenocarcinoma cells was 6.9-fold enhanced in comparison to free EA, meaning that encapsulated EA was significantly superior in reducing HCT-116 cell survival. Moreover, a pharmacokinetic study by oral administration in rabbits showed that the nano-sized formulation improved the relative EA bioavailability by 3.6 times.

#### 4.4.3. Chitosan Micro/Nanospheres

CS is a polysaccharide obtained by partial deacetylation of chitin, endowed by important features from the biomedical point of view. It is mucoadhesive, non-toxic, biocompatible, biodegradable and exerts slight antimicrobial properties. CS is a weak base (pKa value ~6.5) positively charged at physiologic pH. In a study by Arulmozhi et al., EA-CS NP were prepared by ionic gelation method using sodium tripolyphosphate (TPP) as gelating agent, with CS to TPP ratio of 4:1. CS was dissolved in 1% *v*/*v* acetic acid under magnetic stirring overnight and the final pH was adjusted to 5.0 [76]. Different EA amounts (25, 50, 100 mg) dissolved in DMSO (0.5% *v*/*v*) were added to the CS solution, and finally an equal volume of 1 mg/mL TPP solution was added dropwise under mild stirring. The mixture was kept under stirring till NPs formation. The lowest EA amount added led to the highest EE and DL, of 94% and 33% respectively. Spherical NP were obtained with a mean diameter of 176 nm and a slightly positive zeta potential (ζ = +4 mV), which could be due to the adsorption of EA on the particle surface, thus masking CS positive groups. As usually, the EA release showed a two-step profile, where 55% of

loaded EA was rapidly released during the first 8 h and the remaining 20% was released more slowly until 48 h. Finally, EA-CS NPs were investigated as anticancer agents and proved higher cytotoxicity than free EA on KB human oral cancer cell line at almost all the tested doses. Successively, in another study by Gopalakrishnan et al., EA-CS NPs were been investigated as anti-hemorrhagic agents [77]. Also in this case, the NPs were prepared by ionic gelation procedure with 1:12 EA:CS ratio. They sized around 80 nm, being 50% their maximum EE. Release studies showed an EA fast dissolution (84% in 12 h), suitable for clotting purposes. In fact, EA-CS NPs yielded a significantly faster clotting time when compared to free EA, void CS NPs, and an EA:CS physical mixture.

A recent study by Ding et al. focused on the preparation of EA microspheres based on CS and sodium alginate as coating materials [78]. CS-alginate microparticles (MPs) were prepared by ionotropic gelation. The polycationic CS and Ca2<sup>+</sup> were added to the EA alginate solution, thus a polyelectrolyte complex coating formed; MPs were further stabilized with 2% glutaraldehyde solution. SEM images revealed a mean size of 4.36 μm, with some EA crystals on the surface, while EE was around 30%. Regarding dissolution profiles, they reflected the structure of the MPs, because EA was quickly released at 0–3 h and 6–24 h, while it was slowly released from 3 to 6 h because of the necessary deconstruction of the strong coating, suggesting the effective dispersion of EA inside the MPs. Therefore, this formulation seems to be more suitable if a sustained release is desired. Finally, in vitro experiments on 3T3-L1 adipocytes showed a consistent inhibition of adipogenic differentiation by entrapped EA vs. EA alone, suggesting its possible use in obesity treatment.

**Figure 2.** An overview of different ellagic acid's activities versus human diseases.

#### 4.4.4. Zein Nanocapsules

Among the various formulations that have been designed to ameliorate EA bioavailability, only one attempt concerns nanocapsules (NCs). NCs, because of their hollow structure, display large surface-to-volume ratio, low density, short solid-state drug diffusion length, and good surface permeability. On the other hand, NCs present some disadvantages, such as weak mechanical properties and complex production procedures. Recently, EA has been successfully encapsulated into NCs made of zein, a group of water insoluble proteins (prolamins) extracted from corn gluten meal [79]. Ruan et al. developed a system, characterized by EA and a sacrifice template to form the core and by an outer film of zein mixed with a plasticizer to increase film flexibility. NCs were prepared by antisolvent coprecipitation in water using Na2CO3 as core-forming template. Firstly, EA was dissolved in 0.1 M NaOH solution and mixed with Na2CO3 aqueous solution (1% *w*/*w*). EA-Na2CO3 cores were formed by coprecipitation in 70% ethanol, then an ethanolic solution of triethyl citrate and zein was added to

the core suspension, and finally the mixture was poured in distilled water to induce NCs precipitation and Na2CO3 diffusion outwards. Spherical NCs with an average diameter of 72 nm were obtained. EA content (326 mg/g) resulted much higher than that of NPs prepared without Na2CO3. Surprisingly, not remarkable differences were observed in the release behavior between solid and hollow NPs. Both formulations did not show a burst effect, achieving a well-controlled release. An in vitro permeability study on Caco-2 cells showed improved transporting ability of EA NCs compared to pure EA or solid zein NPs, because EA, either when free or present on NPs surface, tended to bind DNA and proteins, causing accumulation inside the cells. In addition, the superior anti-inflammatory activity of orally administered EA NCs was confirmed in a rat model and explained by pharmacokinetic experiments showing AUC 2.5 and 8.7 times higher than the ones of EA solid NPs and pure EA, respectively.

#### *4.5. Dendrimers*

Dendrimers are branched polymers characterized by a tree-like structure of globular shape of nanometric dimension endowed with inner cavities, able to host lipophilic drugs, and a peripheral shell rich in functionalized groups [80,81]. Among various typologies of dendrimers, the well-known polyamidoamine dendrimers (PAMAM) and branched polyethyleneimine polymers (b-PEI) are in vitro well-performant, but mainly because of their extremely cationic inner framework and poor biodegradability, they can exert hemolytic effect, which hindered their clinical translation. To reduce dendrimers cytotoxicity in the study by Alfei et al., PAMAM dendrimers were avoided and EA has been physically encapsulated into cationic dendrimers characterized by an uncharged polyester-based hydrolysable inner core peripherally esterified with a selection of four different amino acids [58]. Histidine, lysine, N-methyl-glycine and N, N-dimethylglycine were chosen to provide the carrier with primary, secondary, tertiary protonable amino groups and with the guanidine residue that is known to promote cellular up-take. Two dendrimers, one hydrophilic and one encompassing in the structure a C-18 chain that provides an amphiphilic character, were employed to encapsulate EA. Dendrimers were dissolved in MeOH and EA was added obtaining a conspicuous suspension maintained at acidic pH under magnetic stirring at room temperature for 48 h. The non-complexed EA was removed by filtration and the final EA dendrimer nanodispersions were brought to a constant weight under vacuum. The values of DL were of 46% (hydrophilic) and 53% (amphiphilic). NPs mean dimension was of 70 nm for both compounds while EA water solubility increased up to 1.7 mg/mL in the case of the amphiphilic dendrimer and up to 4.8 mg/mL in the case of the hydrophilic one.

#### *4.6. Peptide Microtubes*

In the last decade, there has been an increased focus on biomaterial fabrication, of which peptide micro and nanotubes provide a good example. These ordered structures origin from a dipeptide, usually L-diphenylalanine (FF), by self-assembling. Six FF units form cyclic hexamers, which can further be stacked to produce narrow channels with a diameter of approximately 10 Å. Various peptides can be used as building blocks, thereby these vehicles can be tailored according to the desired application [82]. In a study by Barnaby et al., EA was encapsulated by the self-assembling of the synthetic precursor bolaamphiphile bis (N-amido threonine)-1, 5-pentane dicarboxylate, with a pentyl chain between the head groups of the two aminoacids, according to the following procedure [83]. First, the bolaamphilile solution was allowed to assemble for 2–3 weeks with a higher yield at pH 6, then EA was added at a concentration of 0.06 mM, which maximizes EE, as at elevated concentrations EA forms discotic liquid crystal assemblies upon aggregation, while lower EA amounts efficiently insert into the nanotubes by capillarity. The self-assembled EA-loaded nanostructure had an average diameter ranging from 0.5 to 1 μm with an EE of about 80%. EA release was affected by pH, as expected, with about 50% of loaded acid released within 2 h at pH 6 and 7. EA antibacterial property was assessed in vitro against Gram-positive S. aureus and Gram-negative E. coli, and in both cases EA-loaded microtubes were significantly more efficacious than free EA. That was an unexpected result, since EA alone did not exert any action against S. aureus, but probably the explanation relies on the antibacterial activity of threonine itself.

#### *4.7. Functionalized Graphene Oxide (GO) Carriers*

GO is prepared by chemical oxidation of graphite and represent a promising pharmaceutical material due to its small size, biocompatibility, possible functionalization, and low cost. GO offers interesting opportunities for loading and delivering aromatic molecules such as flavonoids via simple physisorption [84]. To further improve GO aqueous solubility, it was covalently functionalized with hydrophilic excipients such as poloxamer F38, Tween 80 and maltodextrin through an esterification reaction. In a study by Kakran et al., EA loading was carried out by mixing 0.5 mg/mL functionalized GO aqueous solution with 50 mg/mL EA in 0.01 M NaOH solution [85]. After stirring overnight at rt, undissolved EA was removed by centrifugation, while the small amount of free dissolved EA was separated by dialysis. Therefore, EA was loaded by simple mixing, due to adsorption supported by hydrogen bonding, π-π stacking and hydrophobic interactions between the acid and the aromatic regions of graphene sheets [85]. Tween 80-GO exhibited the highest DL, with about 1.22 g of EA per g of excipient, as probably the poloxamer hindered GO surface with its higher molecular dimensions, differently Tween 80-GO exhibited a larger area available for the drug attachment. Release studies showed not differences among the three GO derivatives, probably due to the stronger hydrophobic interactions and hydrogen bonding between EA and the carriers, in any case an increasing release was observed raising pH values, which was attributed to the increased hydrophilicity and solubility of EA at higher pH. In vitro cytotoxic activity against MCF7 human breast cancer cells and HT29 human adenocarcinoma cells was found to be higher than the one of free EA, while the antioxidant property of EA-GO formulations was similar to that of EA alone, proving no interference of the carriers on EA scavenging activity as confirmed by DPPH assay.

#### *4.8. Lipid-Based Carriers*

#### 4.8.1. Solid Lipid Nanoparticles (SLNs)

SLNs are colloidal delivery systems based on a high melting point lipid core coated by surfactants. In recent years, SLNs have been considered to be promising delivery systems due to the ease of manufacturing processes, scale up capability, biocompatibility and also biodegradability of formulation constituents [86]. SLNs also have some disadvantages, such as low drug loading efficiency and the possibility of drug expulsion due to its crystallization during storage [87]. Besides the initial burst release of drug which usually occurs represent an additional concern with these formulations [88]. These systems were proposed as carriers for improving the anticancer activity of EA against prostate cancer by Hajipour et al. [89]. EA-loaded SLNs were prepared by the homogenization method as follows. EA was added after dissolution in DMSO to the melted lipid phase, consisting of Precirol® and Tween 80, then the aqueous phase with poloxamer 407 was added dropwise at the same temperature under high speed homogenization. Finally, the obtained nanoemulsion was cooled for NP solidification. Tween 80 seemed to increase EA encapsulation, while the poloxamer at higher concentrations yielded unsuitable sizes. The optimal formulation showed an average diameter around 100 nm with a narrow size distribution (PDI = 0.28) and a negative zeta potential (ζ = −20 mV). The EE was about 89% and DL 36%. Release studies performed at pH 7.4 demonstrated a noteworthy initial burst effect due to the drug present on or close to the surface, followed by a sustained release phase. Like polymer-based particulate systems, SLNs often lack uniformity of distribution of the drug in the nanostructure, which can explain the initial burst effect. In vitro biological studies revealed that EA loaded in SLNs was more efficacious than free EA on PC3 human prostate cancer cells, being the IC50 values of 61 and 82 μM, respectively.

#### 4.8.2. Liposomes (LPs)

With the aim of producing an effective EA dietary supplement, a formulation based on phospholipids was developed by Murugan et al. [90]. EA and hydrogenated soy phosphatidylcholine (HSPC) in 1 to 4 molar ratio were dissolved in dichloromethane and refluxed at a temperature not exceeding 60 ◦C for 2 h. The solution was concentrated and then n-hexane was added for inducing the precipitation of the EA-LPs by anti-solvent method with a yield of encapsulation of 89% *w*/*w*. The resulting EE was around 29% and SEM images showed that EA was intercalated in the spherical lipid layer and that EA-LPs had a mean diameter ranging from 1 to 3 μm. In an in vivo CCl4-induced hepatic damage model, rats orally treated with EA-LPs exhibited a remarkably reduced antioxidant liver damage from oxidative stress. The same effect was obtained only with a double amount of free EA. Moreover, a pharmacokinetic study on the rats confirmed the enhancement of bioavailability exerted by the complexation as the AUC value resulted increased by 2.8-fold if compared to free EA administration.

With the aim of developing a nutritional support during chemotherapy with cyclophosphamide, a recent study by Stojiljkovi´c et al. evaluated the protective effect of EA encapsulated into nanoliposomes, in a cyclophosphamide-induced rat liver-damage model [91]. EA was incubated overnight with a suspension of LPs in 1:10 ratio, obtaining an EE equal to 60%. Stability studies were performed at 37 ◦C at different pH (from 4.5 to 9.9) or in the presence of metal ions. During the 3-h study, EA in the encapsulated form always resulted less degraded, whereas free EA showed higher chelating ability, thus indicating the effectiveness of encapsulation. Subsequently, the loaded liposomes were essayed in a cyclophosphamide-induced rat liver-damage model. The animals were treated for 5 days either with free or encapsulated EA, and only on the third day cyclophosphamide administration occurred. The authors determined serum liver-damage parameters, oxidative tissue damage parameters, and morphological changes of liver cells through histopathological observations. The results indicated that EA nanoliposome formulation could represent an efficacious tool in adjuvant cancer chemotherapy.

#### 4.8.3. Self-Emulsifying Delivery Systems (SNEDDS)

Among the several approaches that were suggested to achieve better EA oral bioavailability, SNEDDS represent an attractive opportunity. SNEDDS are isotropic mixtures of active molecules, oils, surfactants, and coemulsifiers or solubilizers able to form an *o*/*w* nanoemulsion. Once in the gastrointestinal fluid, the fine droplet dispersion enhances water solubility of lipophilic molecules, thus providing more effective absorption. Though SNEDDS are not suitable for exerting controlled release, they demonstrated high potential in increasing oral bioavailability of several therapeutic agents, such as curcumin [92]. However, the poor palatability of the components and low compatibility with other excipients might be the major drawbacks in case of oral delivery. In this context, Avachat et al. developed a formulation containing EA starting from a mixture of EA and soy lecithin at 1:1 molar ratio [93]. This preliminary step was necessary to facilitate the incorporation of a larger amount of EA in the lipid phase and to increase its solubility. The EE was 95% and the water and n-octanol solubilities of complexed EA at 25 ◦C increased by about 3 and 6 times, respectively. Well-formulated SNEDDS should disperse in a few seconds under gentle stirring. Therefore, after a preliminary screening, the optimized composition was selected with the help of ternary phase diagrams. The final formulation included EA PL, glyceryl triacetate 40% *v*/*v* as oil, polyoxyl 40 hydrogenated castor oil 40% *v*/*v* as surfactant, PEG 400 20% *v*/*v* as co-surfactant. To prevent EA oxidation, tocopherol was added in the oil phase. This composition led to the smallest mean diameter (106 nm, PDI of 0.3388) and to a negative zeta potential (ζ = −13 mV). As expected, EA release from SNEDDS was very fast at pH 6.8 (nearly 95% after 1 h), and a bit slower at pH 1.2, but this largest amount of dissolved EA may strongly increase gastrointestinal absorption. To support the hypothesis of a potential enhancement of oral bioavailability, an ex vivo permeability study was also carried out, which confirmed the good permeation of the SNEDDS through rat stomach and intestine.

In a more recent work by Wang et al., EA was entrapped in SNEDDS made of food-grade components to prepare a highly biocompatible formulation suitable for nutraceuticals, performing the procedure previously described [94]. Firstly, a preliminary study investigated the solubility of EA in different oils, surfactants, and co-surfactants, then ternary diagrams were constructed. The best formulation was selected on the basis of the weight of water added dropwise to the mixed ingredients without causing turbidity, and the best formula corresponded to EA 2.5 mg/mL, PEG 400 45% *v*/*v* as co-surfactant, Tween 80 45% *v*/*v* as surfactant, and caprylic/capric triacylglycerols 10% *v*/*v* as oil. This mixture was able to load EA in a concentration 250 times higher than its aqueous solubility and to release it completely within 1 h at pH 6.8. In addition, the authors developed a pharmacokinetic study on rats to evaluate in vivo oral bioavailability of EA in SNEDDS and the AUC value of the emulsified EA was 6.6 times higher than the one of the EA suspension.

An issue associated with SNEDDS is the possibility for the dissolved molecule to undergo precipitation if administered in high dose [95]. Therefore, a slight but important modification consists of adding a precipitation inhibitor, such as PVP or HPMC, capable of increasing drug loading and stability over time. On this purpose a supersaturable self-nanoemulsifying drug delivery system (S-SNEDDS) containing EA was developed by Zheng et al. [96]. As usual, according to the results of solubility studies, those vehicles highly mixable with EA were chosen for formulation optimization through ternary phase diagrams. The final formulation consisted of Tween 80, PEG 400 and ethyl oleate at the ratio of 67.5/22.5/10% *w*/*w*, which can be fully diluted with water without phase separation. In addition, to avoid EA precipitation during storage, 0.5% PVP K30 was added. DL was equal to 4 mg/g, while EA S-SNEDS droplet mean diameter was 45 nm and the zeta potential was negative (ζ = −23 mV). The dissolution profiles were characterized by a slightly less rapid release, compared to EA-SNEDDS, since the complete delivery at pH 6.8 was reached after 4 h. Finally, in vitro and in vivo antioxidant activity was studied by DPPH assay or by detection of malondialdehyde, superoxide dismutase and glutathione in mice liver. The results indicated that the antioxidant ability of EA was noticeably improved by the S-SNEDDS vs. EA suspension, but it was of a lesser extent compared to vitamin C.

#### *4.9. EA Formulations in Fixed Combination with Other Bioactive Molecules*

With the aim to prevent or treat diseases associated with increased oxidative stress such as hyperglycemia, obesity, atherosclerosis, the industry of nutritional supplement has rapidly and greatly expanded over recent decades [97]. In a study by Ratnam et al., EA was combined with Coenzyme Q10 (CoQ10) in a PLGA nanoparticulate formulation to prevent hyperlipidemia [98]. The NP were prepared by emulsion-diffusion-evaporation method, dissolving PLGA and CoQ10 in ethyl acetate, and EA in PEG 400. The two solutions were mixed and emulsified in 1% PVA aqueous solution. The mean particle size obtained was around 259 nm with EE of 70% for EA and of 72% for CoQ10. In an in vivo experiment, rats were fed with a high fat diet and daily administered with oral suspension of free drugs or once in three days with EA-CoQ10 NP. The effects on reducing glucose and triglyceride levels were similar, but only EA-CoQ10 NP exerted a prolonged control on cholesterol levels for up to 2 weeks after treatment suspension.

With the aim at decreasing the risk of chemotherapy complications, further investigation concerning co-delivery systems involved EA and chemotherapy drugs, such as paclitaxel (PTX), for a combined anticancer therapy [99]. The idea of combining these actives was based on the evidence that PTX resistance relies on the NF-kB-dependent pathway, while EA hinders NF-kB. The two molecules were formulated with the temperature-sensitive amphiphilic copolymer poly (N-isopropylacrylamide-PEG acrylate), stable under physiological conditions, being its lower critical solution temperature higher than 37 ◦C. The NP were loaded directly by using the dialysis method at different temperatures. Several formulations were prepared, but the best one, in terms of EE and efficacy, contained the copolymer, PTX and EA in 20:1:1 weight ratio. The mixture was solubilized in DMF and dialyzed against deionized water, in which the copolymer self-assembled, developing core-shell nanoparticles. Since the drugs molecular weight is lower than the dialysis membrane cut-off

(12 kDa), a slight loss of EA and PTX was unavoidable. To limit this phenomenon and to achieve a high EE, the loading reaction was performed at 4 ◦C and an EE of 92% for PTX and of 98% for EA were achieved. After completion of the dialysis, the solution was filtered and lyophilized. NP size was below 200 nm and the release was well-controlled, as only the 8% of both drugs was released after 2 h, while the 61% of PTX and the 88% of EA were released after 48 h. In vitro cytotoxicity tests were performed against MCF-7 breast cell line. After 48 h PTX/EA-loaded NP showed slightly higher cytotoxicity compared to free PTX, deriving from an improved cellular uptake. However, further studies would be needed to assess an eventual synergism between the two loaded drugs.



<sup>1</sup> Caprylic/capric triacylglycerols <sup>2</sup> Polyoxyl 40 hydrogenated castor oil.

In a multiple drug co-delivery study by Fahmy et al., fluvastatin (FLV), alpha lipoic acid (ALA) and EA were chosen for their anticancer properties [100]. As dosage form, nanostructured lipid carriers (NLCs) were taken into consideration. NLCs are a relatively new generation delivery system, developed to overcome SLNs drawbacks such as low payload, crystallization of the lipid matrix, and drug expulsion during storage. In fact, NLCs, being composed of both solid and liquid lipids, do not undergo crystallization, allowing higher drug loading and consequent better bioavailability of low soluble drugs. Furthermore, NLCs are composed of highly biocompatible and biodegradable lipids, so they have a wider range of applications. As an example, NLCs are useful nutraceutical delivery systems with high drug loading, stability, capability in increasing bioavailability of bioactive compounds and consumer acceptability. In addition, they may provide controlled release of the encapsulated materials [101]. In this work, NLCs were prepared by hot emulsification–ultrasonication method starting from 1% almond oil, 3% glyceryl dibehenate, 0.5% L-α-phosphatidylcholine phospholipid, 0.25% FLV, 0.02% EA and 0.3% ALA. The mixture was melted heating up to 50 ◦C in chloroform:ethyl acetate 1:1. To this lipid phase 1% Gelucire® 44/14 aqueous solution was added and the resultant emulsion was sonicated, cooled down and collected by ultracentrifugation. The NPs had a mean particle

size of 85 nm (PDI = 0.58) and a negative zeta potential (ζ = −25 mV). The EE was 98% for FLV, 92% for ALA, and 96% for EA. The release rate was rather slow after an initial burst effect. The formulation was evaluated for its in vitro cytoxicity against PC3 prostate carcinoma cells. The co-delivery of FLV, ALA and EA from NLCs always provided better results in comparison with free drugs both when administered individually and mixed together, with significant differences in terms of cell survival, caspase-3 expression and incidence of apoptosis.

In another interesting study, Abd Elwakil et al. focused on the preparation of an inhalable spray dried powder for targeted co-delivery of EA and doxorubicin (DOX) to lung carcinoma [102]. This system consisted of nanocomposites comprised of drug-loaded NP and excipients like sugars just to reach the micro-range size necessary to ensure deposition in deep lung tissue. Once at the alveolar surface, the carbohydrate fraction rapidly dissolves, releasing the NP components apt to be internalized by the cancer cells. NP matrix was composed of lactoferrin (Lf), a cationic glycoprotein chosen for its ability to bind transferrin receptors overexpressed on cancer cells, and chondroitin sulfate (ChS), a polyanionic glycosaminoglycan able to bind hyaluronic acid CD44 receptors. The first step of this procedure was the nanocrystallization of raw EA via antisolvent precipitation in order to increase EA incorporation in the hydrophilic matrix. A methanolic EA solution was added in a 1:10 volume ratio to an aqueous phase containing 0.5% *w*/*v* poloxamer F188 as a stabilizer at 4 ◦C and the mixture was stirred for 15 min. The resulting EA nanocrystals had an average size of 148 nm (PDI of 0.185) and an aqueous solubility 33.3-fold higher than raw EA. The loaded Lf-ChS NPs exploited the polyelectrolyte electrostatic complexation. DOX dissolved in distilled water was dropped gradually into 2% *w*/*v* ChS solution containing lyophilized nano-sized EA; this dispersion was added by controlled dripping to 2% *w*/*v* Lf solution, at a suitable pH to allow formation and stabilization of NPs. The resulting product was harvested by centrifugation. The loaded NPs showed a size of around 192 nm and a negative zeta potential (ζ = −27 mV). The EE was 91% for DOX and 96% for EA. The in vitro anticancer efficacy was tested against A495 human lung cancer cell line. Co-encapsulated DOX and EA Lf–ChS NPs enhanced the potency of the drug co-administrated in mixture, as demonstrated by the significant reduction of IC50 value by 3.8-fold compared with combined free drug solution. This system was not expensive, scalable, and made of natural and highly biocompatible components, characteristics that make it potentially useful for preparation of EA-enriched foods.





*Appl. Sci.* **2020**, *10*, 3353


**Table3.***Cont*.

#### **5. Conclusions**

In recent decades, a growing interest in the administration of polyphenolic compounds for prevention of several diseases has arisen, since epidemiological studies have revealed a correlation between dietary habits and disease risks. In particular, EA exerts various health-promoting activities, suggesting that it may play an important role in dietary supplements. Furthermore, a few research studies proved that EA is endowed with a wide spectrum of therapeutic effects against oxidation-linked chronic illnesses such as, above all, diabetes, cancer, neurodegenerative disorders, and cardiovascular diseases. Unfortunately, EA presents unsuitable biopharmaceutical features, including poor bioavailability and interindividual variability that hamper its successful employment in prophylaxis and disease treatment. To date, on the market, EA is present in beverages, capsules, and tablets, which obviously do not overcome the problems associated with low oral EA intake; furthermore liquid formulation submits it to fast degradation. From this background, in order to address the many drawbacks associated with EA in vivo absorption, strategies mainly consisting of micro and nanotechnology approaches were designed and performed. The obtained EA formulations demonstrated modifying its release and improving its solubility, stability during storage, and bioavailability in animal models. However, from the evaluation of the various developed formulations, some considerations may be made. For example, approaches starting from basic EA solutions lead to higher EA content, but the preparation timing may represent a crucial factor as undesirable reactions such as oxidation and hydrolysis may occur. As a matter of fact, the choice of a proper vehicle for dissolving EA is challenging, because EA results rather insoluble in most common solvents. The attempt of reducing the size of EA powder through anti-solvent precipitation could be a recommended preliminary step as reported by different authors. Regarding micro and nanosystems preparative methods, encapsulation in biodegradable PLGA or PCL micro and nanospheres represents a valid route when effective EA protection, long circulation and controlled release are required, as, despite an initial burst effect, these systems provided a sustained release over one week. Thanks to its biocompatibility, chitosan was also extensively used, providing a rather fast EA release (50% within 8 h). Concerning lipid carriers, self-emulsifying systems are the most investigated ones, since they were usually endowed with high EA loading capacity and gastrointestinal release within 1 h. Although in the work of Wang et al. food-grade excipients were used, the effects of surfactants concentration on intestinal epithelial integrity has to be taken into consideration. In this regard, it is of great importance to develop formulations with residual organic solvents or surfactants below the recommended maximum levels indicated by regulatory agencies. The most recent, innovative, and highly biocompatible EA formulations consisted of pectin spray dried dispersion, cyclodextrin-based nanosponges, zein nanocapsules, chitosan/alginate microspheres, lactoferrin/chondroitin sulfate nanoparticles, and supersaturatable self-microemulsifying delivery systems. With these promising advances, novel and more effective strategies could be applied for allowing extensive investigations on EA in vivo beneficial effects. In this regard, in the last pharmacokinetic study in humans, it was reported that after pomegranate extracts consumption, the key factors hampering EA effectiveness are: its low solubility at the gastric pH, its bounding to intestinal epithelium, the saturable transcellular transport and the interindividual variability to produce urolithins. All these drawbacks may be overcome at least to a large extent by applying micro or nanocarrier-based approaches, suggesting that future pharmacokinetic studies will provide more encouraging results if they are performed by using an optimized pomegranate acid delivery system.

**Author Contributions:** Conceptualization, G.Z., S.B. and G.C.; resources, G.Z., F.T., S.A. and S.B.; writing—original draft preparation, G.Z.; writing—review and editing, G.Z., S.B., G.C., G.A., S.A. and F.T. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
