**Evaluation of the Extraction Temperature Influence on Polyphenolic Profiles of Vine-Canes (***Vitis vinifera***) Subcritical Water Extracts**

**Olena Dorosh 1,**† **, Manuela M. Moreira 1,**† **, Diana Pinto <sup>1</sup> , Andreia F. Peixoto <sup>2</sup> , Cristina Freire <sup>2</sup> , Paulo Costa <sup>3</sup> , Francisca Rodrigues 1,\* and Cristina Delerue-Matos <sup>1</sup>**


Received: 28 May 2020; Accepted: 1 July 2020; Published: 3 July 2020

**Abstract:** This work focused on evaluating the possibility of using vineyard pruning wastes from two Portuguese *Vitis vinifera* varieties; Touriga Nacional (TN) and Tinta Roriz (TR), as new potential ingredients for the nutraceutical industry. An environmentally friendly extraction technique; namely subcritical-water extraction (SWE), was employed. The overall results indicate that phenolic acids were the major class of compounds quantified; being gallic acid the principal one. The highest value for total phenolic content (TPC) was obtained for the TR extract at 250 ◦C (181 ± 12 mg GAE/g dw). In terms of antioxidant activity; the DPPH values for the extracts obtained at 250 ◦C were approximately 4-fold higher than the ones obtained at 125 ◦C; with TR extract presenting the highest value (203 ± 22 mg TE/g dw). Thus, the TR extract obtained through SWE at 250 ◦C was selected to evaluate the scavenging activity and the in vitro effects on cells due to the best results achieved in the previous assays. This extract presented the ability to scavenge reactive oxygen species (O<sup>2</sup> •- , HOCl and ROO• ). No adverse effects were observed in HFF-1 viability after exposure to extract concentrations below 100 µg/mL. This work demonstrated that vine-canes extracts could be a potential ingredient to nutraceutical industry

**Keywords:** vine-canes; subcritical water extraction; scavenging capacity; in-vitro assays; byproducts

#### **1. Introduction**

Contemporary society is tightly bonded to over-consumerism being characterized by a mass-production of goods and consequently their over consumption. This linear economy depends on two basic assumptions: (i) there will always be resources that can be extracted and (ii) there will always be a place where it is possible to get rid of the materials that are not wanted anymore. Nevertheless, currently, due to the development of science and environmental awareness, there is a growing conscience that these two assumptions are not real and that the linear economy is not sustainable [1]. In fact, human population has grown exponentially in the last two centuries increasing the resources consumption [2]. The increased demand over natural resources has been negatively

affecting Earth's overshoot day. To reach a more sustainable world status and to preserve what is still left for future generations, it is imperative the transition from an economy based on fossil resources to a concept of circular economy. A growing number of companies, including the food industry, are working to overcome this challenge and transform their process of production in a more environmentally ethical practice. Food and beverage industries are the principal manufacturing sector in the European Union (EU) generating high amounts of byproducts for which profitable solutions need to be found [3,4]. Until now, the general application for the agro-food industry waste has been animal feed (that may not adjust to the nutritional requirements), combustion feedstock or fertilizers, causing major environmental issues [5,6]. Nevertheless, these byproducts can be used as renewable natural resources for many applications, such as low-cost adsorbents, nutraceuticals, supplement food products and ready meals, leading these industries to a concept more related to a circular economy [6,7].

Grapes (*Vitis vinifera*) are one of the principal fruits produced around the world [8]. According to OIV (International Organization of Vine and Wine), in the last years, Portugal is one of the principal world wine producers, representing this sector as a huge impact on the economy [9]. Consequently, large amounts of wine wastes, namely skins, seeds and stems are produced every year, representing approximately 20% of the total weight of processed grapes [10]. These undervalued byproducts are rich in bioactive compounds, particularly polyphenols [11–15], that could be used in several applications such as antioxidants in food, cosmetic or even pharmaceutical industries [16]. Depending on the vine varieties, around 1.75 tons of vine-cane wastes are produced for each hectare of vineyard. After harvesting season, many agricultural byproducts, including vine-canes, are usually incorporated in the soil, enhancing the soil health due to the degradation of organic matter and reducing the necessities of organic fertilizers and/or correctives [17]. Following the need to find a better end for vine-canes, different approaches were already explored, such as the biochar production, biofuels, pulp for paper sheets and particle board [18–20].

Vine-canes are constituted by two main fractions, holocellulose (cellulose and hemicellulose) and lignin, which correspond to approximately 68% and 20% of the total vine-canes weight, respectively [21]. The lignin degradation process releases phenolic compounds of low molecular weight, alcohols, aldehydes, ketones or acids [16]. By employing the appropriate extraction technique, considerable amounts of these valuable compounds can be recovered from the lignin fraction of vine-canes and afterwards used in added valued products. Subcritical water extraction (SWE) can be exploited as a sustainable and clean technique to achieve this goal. SWE is a pressurized liquid extraction technique that uses water at high temperatures (100–374 ◦C) and pressures (1–22.1 MPa), but below its critical point (374 ◦C and 22.1 MPa). The use of water as extracting solvent makes this technique safe, cost-effective and environmentally interesting, particularly for the extraction of phenolic compounds to be used in products for human consumption. For example, Gabaston et al. investigated the effect of different times (5, 15 and 30 min) and temperatures (100, 130, 160 and 190 ◦C) in SWE [22]. All extractions were conducted with 5 g of vine-canes powder in a 34 mL cartridge. According to the authors, the best results were obtained at 160 ◦C for 5 min (3.62 g of stilbenes/kg dry weight (dw)) [22]. Indeed, in our previous study, we compared three extraction techniques, namely microwave-assisted extraction (MAE), SWE and conventional extraction (CE), in what concerns to phenolic compounds extracted from vine-canes [12]. The obtained results revealed the advantages and potentialities of SWE, when compared to the other techniques. In this study, we focused on the extraction of vine-canes from two Portuguese varieties (Touriga Nacional, TN and Tinta Roriz, TR) from the Dão region using SWE performed at two different temperatures (125 and 250 ◦C) in order to maximize the bioactive compounds extraction. Opposite to our previous study, where we establish a specific temperature for the SWE (150 ◦C) in this work we aim to compare different extraction temperatures taking in consideration the results reported by other authors highlight temperature as the most influencing parameter on SWE [23–25]. Thus, based on the previous studies, we decided to explore the thermo-chemical conversion reactions effects on the recovery of antioxidant rich products using subcritical water conditions. Based on this, the antioxidant activities and the phenolic profile of the two

vine-canes varieties were investigated. The best extract was selected for further assays, being screened the scavenging activity against oxygen radical species as well as the cell viability effects. Overall, this work follows a sustainable approach for the valorization of vine-canes from the grape industry in order to promote their added value and circular economy.

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

### *2.1. Samples Collection and Preparation*

Vine-canes samples were kindly provided by Sogrape Vinhos, S.A. (Portugal). Both *V. vinifera* vine-canes varieties, namely TN and TR, were obtained in Quinta dos Carvalhais, located in Mangualde (North of Portugal), in November 2015 by randomized selection. Samples were oven-dried (Model no. 2000208, J.P. Selecta, Barcelona, Spain) at 50 ◦C for 24 h and milled (Retsch ZM200) to a particle size smaller than 1 mm. The fine particles obtained after milling were stored in sealed bags at room temperature until use.

#### *2.2. Subcritical Water Extraction*

SWE was conductedin a Parr Series 4560 Reactor connected to the Parr 4848 Reactor Controller (Figure 1).

**Figure 1.** Subcritical water extractor used.

The extractions were performed using 40 g of milled vine-canes and 400 mL of water. Two different extraction temperatures were tested, 125 and 250 ◦C, for 50 min after the sample reached the desired temperature at 250 rpm. After extraction, the system was cooled down and the extract was filtered and centrifuged at 15,763× *g* (Heraeus Megafuge 16 Centrifuge Series, Thermo Scientific, Waltham, MA, USA) for 15 min at 4 ◦C. Then, the extract was stored at −80 ◦C and lyophilized (Edwards lyophilizer) for 48 h, after being stored at 4 ◦C until further use.

−

–

#### *2.3. Determination of Total Phenolic and Flavonoid Contents*

The total phenolic content (TPC) was determined according to Singleton and Rossi [26], with minor modifications described by Paz et al. [27]. The reaction solution consisted of 25 µL of deionized water (blank), standard or sample, 75 µL of deionized water and 25 µL of diluted Folin–Ciocalteu reagent (1:1). After 6 minutes in the dark, 100 µL of a sodium carbonate solution (0.708 M) were added to each well. The microplate (BioTek Instruments, Inc., Winooski, VT, USA) was kept away from the light for 90 min and then the absorbance was measured at 760 nm. Calibration curves were done using gallic acid (GA) as standard. Results were expressed as milligrams of gallic acid equivalents (GAE) per gram of dw (mg GAE/g dw).

The total flavonoid content (TFC) was performed according to Paz et al. [27]. The procedure consisted in adding to each well 100 µL of deionized water, 10 µL of sodium nitrite solution (0.725 M) and 25 µL of deionized water (blank) or standard or sample. After 5 min in the dark, 15 µL of aluminum chloride (0.75 M) were added to each well, and after 1 minute of reaction, also in the dark, 50 µL of sodium hydroxide (1.0 M) were added. The absorbance was measured at 510 nm. Epicatechin was used as standard. Results were expressed as mg of epicatechin equivalents (EE) per gram of dw (mg EE/g dw).

#### *2.4. Determination of Antioxidant Activity and DPPH Free Radical Scavenging Assay*

The FRAP assay was based on the protocol reported by Benzie and Strain [28], with minor modifications described by Paz et al. [27]. FRAP reagent was prepared using a mixture of acetate buffer (pH 3.6; 0.3 M), 2,4,6-Tri(2-pyridyl)-s-triazine (TPTZ; 0.01 M) in HCl solution (0.04 M) and FeCl3.6H2O (0.27 M) in a 10:1:1 ratio. One hundred and eighty microliters of FRAP reagent were added to each well of the microplate along with 20 µL of deionized water (blank) or standard or sample. Absorbance was measured at 593 nm, after incubating in the dark at 37 ◦C for 10 min. A calibration curve was prepared with ascorbic acid (AA). Results were expressed as milligrams of AA equivalents (AAE) per gram of dw (mg AAE/g dw).

DPPH-RSA was performed following the protocol described by Paz et al. [27]. For that, 200 µL of an ethanolic solution of DPPH (0.1 M) were added to 25 µL of the standard or sample. The blank contained 225 µL of ethanol and the control 225 µL of the DPPH reagent. The reaction solution was incubated for 30 min in the dark. DPPH-RSA was determined spectrophotometrically at 517 nm. Calibration curve was made with Trolox. Results were expressed in milligrams of Trolox equivalents (TE) per gram of dw (mg TE/g dw).

#### *2.5. Qualitative and Quantitative Polyphenol Characterization*

The phenolic profile of subcritical water extracts was obtained by high performance liquid chromatography (HPLC) with photodiode array (PDA) detection employing the method described by Moreira et al. [12]. A Shimadzu HPLC system equipped with a Phenomenex Gemini C<sup>18</sup> column (250 mm × 4.6 mm, 5 µm) and a guard column with the same characteristics, that were kept at 25 ◦C, were used. Individual phenolic compounds were prepared in methanol and their mixtures for calibration curves construction were obtained by dilution of appropriate amounts of the stock solutions in methanol:water 50:50 (*v*/*v*). The mobile phase was composed by methanol (A) and water (B) both with 0.1% formic acid, which were previously filtered (0.20 µm nylon filter, Supelco, Bellefonte, PA, USA) and degassed for 15 min in an ultrasonic bath (Raypa® trade, Terrassa, Spain). A gradient program, at a flow rate of 1.0 mL/min and 20 µL of injection volume were used and the identification of detected peaks in subcritical water extracts was performed by comparing their retention time and UV-vis spectra with the ones of pure standards. GA, protocatechuic acid, (+)-catechin, 4-hydroxyphenilacetic acid, 4-hydroxybenzoic acid, 4-hydroxybenzaldehyde, vanillic acid, syringic acid, (-)-epicatechin, naringin, phloridzin, cinnamic acid, naringenin, phloretin and pinocembrin were quantified at 280 nm; chlorogenic acid, caffeic acid, *p*-coumaric acid, ferulic acid, sinapic acid, resveratrol and tiliroside at 320 nm and quercetin-3-*O*-glucopyranoside, rutin, ellagic acid, myricetin, kaempferol-3-*O*-glucoside, kaempferol-3-*O*-rutinoside, quercetin and kaempferol at 360 nm and their amount was expressed as mg/100 g dw.

#### *2.6. Reactive Oxygen Species Scavenging Capacity*

#### 2.6.1. Superoxide Radical Scavenging Assay

The superoxide radical (O<sup>2</sup> •- ) scavenging assay was performed according to Pistón et al. [29]. The reaction mixture was prepared by adding to each well the following reagents dissolved in phosphate buffer (19 × 10−<sup>3</sup> M, pH 7.4): 50 µL of NADH (166 × 10−<sup>6</sup> M); 150 µL of nitroblue tetrazolium (NBT; 43 × 10−<sup>6</sup> M); 50 µL of tested extract at different concentrations or phosphate buffer for the blank or positive controls and finally 50 µL of PMS (2.7 × 10−<sup>6</sup> M). Absorbance was read at 560 nm for 6 min at 37 ◦C in the microplate reader. GA and catechin were used as positive controls. The observed effects

were expressed as inhibition percentages of the NBT reduction to diformazan. Results were expressed as the necessary concentration of controls and subcritical water extract of TR variety obtained at 250 ◦C to inhibit 50%, IC50, of the NBT reduction to diformazan.

#### 2.6.2. Hypochlorous Acid Scavenging Activity

Hypochlorous acid (HOCl) scavenging activity was measured using a fluorescent methodology previously described by Pistón et al., based on the HOCl-induced oxidation of dihydrorhodamine (DHR) to rhodamine [29]. GA and catechin were used as positive controls. A 1% (*m*/*v*) NaOCl solution was used to prepare the HOCl solution by adjusting the pH to 6.2 with addition of H2SO4. The assay was directly performed in a 96-well microplate and the reagents previously dissolved in phosphate buffer (100 × 10−<sup>3</sup> M, pH 7.4). In each well, the reaction mixture consisted of: 150 µL of phosphate buffer (100 × 10−<sup>3</sup> M); 50 µL of tested extract at different concentrations or phosphate buffer for the blank or positive controls; 50 µL of DHR (5 × 10−<sup>6</sup> M) and finally 50 µL of HOCl (5 × 10−<sup>6</sup> M). Absorbance was read at 560 nm for 6 min at 37 ◦C in the microplate reader. Results were expressed as the inhibition, in IC50, of HOCl-induced oxidation of DHR.

#### 2.6.3. Peroxyl Radical Scavenging Activity

Peroxyl radical (ROO• ) was generated by thermo-decomposition of AAPH at 37 ◦C. The ROO• scavenging activity, also known as the oxygen radical absorbance capacity (ORAC) assay, was measured by the fluorescence decay of fluorescein as previously described by Ou et al. [30]. Reaction mixtures contained the following reagents dissolved in potassium phosphate buffer (75 × 10−<sup>3</sup> M, pH 7.4): 150 µL of fluorescein (61.2 × 10−<sup>9</sup> M); 25 µL of the tested extract at different concentrations or phosphate buffer for the blank or positive controls and 25 µL of 2,2'-Azobis(2-amidinopropane) dihydrochloride (AAPH; 19.1 × 10−<sup>3</sup> M). After preparing the reaction mixture, the 96-well plate was incubated in the microplate reader during 2 h at 37 ◦C, where the decrease in fluorescence was measured every minute. The excitation and emission wavelengths used were 528 ± 20 nm and 485 ± 20 nm, respectively. GA and catechin were used as positive controls, while Trolox was employed as standard. Results were expressed in ratio values: slope of the sample/slope obtained for Trolox.

#### *2.7. Cell Viability Assay*

The 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTT) assay was employed to access the cell viability after exposure to the extract. HFF-1 was purchased from ATCC (ATCC Number: SCRC-1041; ATCC, Manassas, VA, USA). Passages 24–26 were used. The assay was performed according to Rodrigues et al. [31]. HFF-1 was grown in DMEM medium, previously fortified with 10% of heat inactivated fetal calf serum, 1% of non-essential amino acids and 1% of antibiotic. Cells were maintained in an incubator (CO<sup>2</sup> Incubator MCO-18AC, Panasonic, Osaka, Japan) at 5% CO<sup>2</sup> and 37 ◦C and the culture medium was changed every 2 days until cells reached a good confluence. Afterwards, cells were cultured in 96-well microplates at a density of 2.5 × 10<sup>4</sup> cells/mL for 24 h at 37 ◦C with 5% of CO2, in order to provide conditions for an exponential cell growth. After the period of multiplication and adherence, the medium was removed, and cells were washed with phosphate-buffer saline (PBS) solution. Following, cells were incubated with different concentrations of extract (0.1–1000 µg/mL) dissolved in the DMEM medium for 24 h at 37 ◦C. After the incubation period, extracts were removed and cells were washed again with PBS and subsequently MTT was added to each well. The microplate was incubated for 4 h at 37 ◦C with 5% of CO<sup>2</sup> in the dark. After that period of incubation, DMSO was added and the microplate was put into agitation for 10 min to solubilize the MTT crystals. Positive control was made by incubating the cells only in culture medium, while negative control was made by incubating the cells only in Triton X-100. The absorbance was measured at 490 nm and at 630 nm to subtract the background. Results were expressed as percentages of cell viability.

#### *2.8. Statistical Analysis*

The statistical analysis was performed using IBM SPSS Statistics for Windows software (Version 24.0, IBM Corp., Armonk, NY, USA). Data was reported as mean ± standard deviation (SD) of three replications. The normal distribution and the homogeneity of variances were assessed by Shapiro–Wilk's and Levene's tests, respectively. For all assays, the data were normal and the homogeneity of variances confirmed. To evaluate the differences between samples, the one-way ANOVA was used. Tukey's HSD test was employed for the post hoc comparisons of the means, being *p* < 0.05 accepted as denoting significance. To compare the same sample at different temperatures, a *t*-test was employed, being *p* < 0.05 accepted as denoting significance. GraphPad Prism 5 software (GraphPad, La Jolla, CA, USA) was employed to calculate the IC<sup>50</sup> values of ROS scavenging activity.

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

#### *3.1. Total Phenolic and Flavonoid Contents, Antioxidant and Antiradical Activities*

In the present study, the influence of the extraction temperature parameter was evaluated, since previous studies have demonstrated that the temperature increase in SWE has more impact in the extraction efficiency than the solid–liquid ratio and the extraction time employed [23–25]. For example, Joki´c et al. reported that the best result for the TPC assay was obtained when the SWE temperature was 250 ◦C [23]. Taking into consideration the published studies, two extreme values of temperature were employed in the present work aiming to understand their influence in the phenolic profile characterization and antioxidant activity. Table 1 summarizes the values obtained for TPC, TFC, DPPH-RSA and FRAP assays for extracts obtained by SWE performed in TN and TR vine-canes varieties, with two different extraction temperatures (125 and 250 ◦C).

**Table 1.** Total phenolic content (TPC, results expressed in mg gallic acid equivalents (GAE)/g dw), total flavonoid content (TFC, results expressed in mg epicatechin equivalents (EE)/g dw), 2,2-diphenyl-1-picrylhydrazyl radical scavenging activity (DPPH-RSA, results expressed in milligram Trolox equivalents (TE)/g dw) and ferric reduction antioxidant power (FRAP, results expressed in milligram ascorbic acid equivalents (AAE)/g dw) of subcritical water extracts from Touriga Nacional (TN) and Tinta Roriz (TR) vine-canes varieties. Results were expressed as mean ± standard deviation.


For each assay (TPC, TFC, DPPH-RSA and FRAP), # for the same extraction temperature means that the vine-cane variety produces statistically significant differences (*p* < 0.05); \* for the same vine-cane variety means that different extraction temperatures produce statistically significant differences (*p* < 0.05).

According to Table 1, the increase of temperature from 125 to 250 ◦C resulted at least in 3.3-fold higher TPC values for TN and TR vine-canes varieties. Significant differences (*p* < 0.05) between the quantity of phenolic compounds extracted at 125 and 250 ◦C were observed. Considering the different vine-canes varieties studied, TR presented the highest TPC values (*p* < 0.05) at both tested temperatures. The highest value obtained was 181 ± 12 mg GAE/g dw for the TR extract obtained at 250 ◦C. These results are in line with Dorosh et al. as well as Moreira et al. that observed the same tendency for vine-canes extraction using SWE and other advanced extraction technologies: TR extracts present higher TPC values than the TN variety [12,32].

Considering the two variables studied, the values obtained for TFC were in accordance with the ones obtained for the TPC assay. As shown in Table 1, significant differences (*p* < 0.05) were observed between the TFC values obtained at 125 and 250 ◦C for the same vine-cane variety, as well as for the two different vine-canes varieties using the same extraction temperatures. The highest TFC was reported for TR extract obtained at 250 ◦C (51 ± 6 mg EE/g dw). However, the TFC values of the present study are not in agreement with the ones reported by Moreira et al. [12] since a higher TFC was reported by the authors for TN extracts obtained by SWE at 150 ◦C. This difference could be due to the environmental conditions observed in the harvest year, for example, as reported by other authors for other seasonal matrices also produced in the same region [33].

In what concerns to DPPH-radical scavenging assay, the capacity to reduce this radical was similar (*p* > 0.05) for both vine-canes extracts obtained at the highest temperature studied (250 ◦C). Nevertheless, at 125 ◦C, TN and TR extracts presented significant differences (*p* < 0.05). Regarding the comparison of different extraction temperatures for the same vine-cane variety, the DPPH values of the extracts obtained at 250 ◦C were approximately 4-fold higher than the ones obtained at 125 ◦C, with TR extract presenting the highest value (203 ± 22 mg TE/g dw). In fact, the increase of SWE temperature seems to be the major factor influencing the antioxidant capacity of the extracts, as previously reported by other authors [23–25]. The employment of high temperatures probably resulted in the formation of new compounds, which could justify the higher antioxidant activity observed [16,34]. Additionally, it was observed that the extracts with the highest TPC and TFC showed also higher DPPH• scavenging capacity. The increase in antioxidant capacity caused by a higher polyphenols content was also previously reported by several other authors [11,12,35].

According to the results presented in Table 1, the obtained values for FRAP assay were in accordance to the ones obtained for DPPH-RSA assay. For instance, considering the same vine-cane variety, the results obtained for the extracts at 250 ◦C were almost 4-fold higher than the ones obtained at 125 ◦C (*p* < 0.05). Additionally, no significant differences (*p* > 0.05) were observed between the values obtained for TN and TR for the highest extraction temperature tested (250 ◦C). Nevertheless, significant differences (*p* < 0.05) between the two varieties were reported for extracts performed at 125 ◦C. The highest FRAP value was obtained for TR extract at 250 ◦C (202 ± 14 mg AAE/g dw), which is the same subcritical water extract that presented the highest results for the three assays previously discussed. In fact, TR variety had higher antioxidant activity than TN, which was also previously reported by Moreira et al. [12].

Based on the data previously discussed, it is possible to conclude that higher extraction temperatures resulted in higher amounts of polyphenols and flavonoids as well as higher antioxidant properties. Further, it was also demonstrated that the vine-cane variety used exerts a significant influence in the obtained results, with TR variety being a better matrix to recover bioactive compounds.

#### *3.2. Identification of Phenolic Compounds by HPLC-PDA*

Due to the significant differences in the TPC, TFC, DPPH-RSA and FRAP results of the extracts obtained at 125 and 250 ◦C, HPLC-PDA analyses were performed to understand which phenolic compounds were the main contributors to the antioxidant properties. Figure 2 presents the HPLC chromatograms obtained for the polyphenol's standard mixture (Figure 2A) and for TR extract obtained at 250 ◦C (Figure 2B). Table 2 summarizes the obtained results.

According to Table 2, phenolic acids were the major class of compounds identified and quantified, corresponding to 43% and 78% and to 38% and 80% of the total amount of phenolic compounds for TN and TR extracts obtained at 125 and 250 ◦C, respectively. GA was the main phenolic acid quantified, especially in the extracts obtained at 250 ◦C (891 ± 45 and 1066 ± 53 mg of GA/100 g dw for TN and TR respectively). In fact, GA is a compound commonly found in wines, being responsible for its characteristic astringency [14,16]. In a recent study, we reported a higher amount of phenolic acids for TN subcritical water extract obtained at 150 ◦C (790 ± 40 mg of phenolic acids/100 g dw) than for the CE and MAE extracts (77.3 ± 3.9 and 265 ± 13 mg of phenolic acids/100 g dw, respectively) [12]. In the same study we also observed higher amounts of phenolic compounds for TN than for TR variety contrary to the results presented in Table 2 [12].

11) (−) **Figure 2.** HPLC chromatograms at 280 nm for (**A**) polyphenols standard mixture and (**B**) Tinta Roriz subcritical water extract at 250 ◦C; peak identification: (1) gallic acid, (2) protocatechuic acid, (3) (+)-catechin, (4) 4-hydroxyphenilacetic acid, (5) 4-hydroxybenzoic acid, (6) 4-hydroxybenzaldehyde, (7) chlorogenic acid, (8) vanillic acid, (9) caffeic acid, (10) syringic acid, (11) (−)-epicatechin, (12) *p*-coumaric acid, (13) ferulic acid, (14) sinapic acid, (15) naringin, (16) rutin, (17) resveratrol, (18) quercetin-3-*O*-glucopyranoside, (19) phloridzin, (20) cinnamic acid, (21) ellagic acid, (22) myricetin, (23) kaempferol-3-*O*-glucoside, (24) kaempferol-3-*O*-rutinoside, (25) naringenin, (26) quercetin, (27) phloretin, (28) tiliroside, (29) kaempferol and (30) pinocembrin.

The results discussed above suggest that the applied extraction technique significantly influences the amount of recovered bioactive compounds. Besides phenolic acids, flavanols and flavonols were the two main subclasses of flavonoid compounds present in higher amounts in the analyzed extracts. Moreover, comparing the amount of flavonols extracted by SWE at 125 ◦C in the present study with the amount reported in our last study at 150 ◦C, a higher content of flavonols was recovered when lower extraction temperature was employed, demonstrating that temperatures higher than 125 ◦C may cause degradation of these compounds [12]. Gabaston et al. focused on the extraction of stilbenes using SWE and obtained an amount of 362 mg of stilbenes/100 g dw, when performing the extraction at 160 ◦C for 5 min [22]. For the same conditions, the amount of resveratrol recovered was 130 mg/100 g dw, which is approximately 8 and 10 times more than the obtained for the extraction at 250 ◦C with TN and TR, respectively. Additionally, the authors pointed out that high temperatures and long periods of extraction may lead to the degradation of these compounds, which can justify the results obtained for TN and TR and presented in this paper.


**Table 2.** Content of the identified phenolic compounds in Touriga Nacional (TN) and Tinta Roriz (TR) extracts obtained through subcritical water extraction (SWE) at 125 and 250 ◦C. Results were expressed as mean ± standard deviations (milligrams of compound/100g dw, *n*=3).

<sup>a</sup> ND: not detected; <sup>b</sup> LOD: limit of detection.

Thus, the HPLC results showed that the amount of phenolic compounds quantified for the TR vine-cane variety was at least 1.31 times higher than the ones reported for TN. Regarding the two extraction temperatures tested (125 and 250 ◦C), the extracts obtained with the highest temperature presented higher amounts of phenolic compounds, which was already expected as the values obtained for the TPC, TFC, DPPH-RSA and FRAP assays for these extracts were also higher. In this way, the TR extract obtained at 250 ◦C was selected for the further assays.

#### *3.3. Capacity of Scavenging Reactive Oxygen Species*

Reactive oxygen species (ROS) are produced as side products of metabolic reactions. Nevertheless, in some cases these compounds are not naturally neutralized by cells, interacting with other molecules

and interfering in metabolic pathways, which results in oxidative damage to cellular biomolecules [36]. Therefore, it is critical to perform assays that evaluate the scavenging capacity of ROS by extracts. Table 3 summarizes the results obtained for ROS scavenging assays.


**Table 3.** Superoxide (O<sup>2</sup> •- ), hypochlorous acid (HOCl) and peroxyl radical (ROO• ) scavenging activities of Tinta Roriz (TR) subcritical water extract obtained at 250 ◦C.

IC<sup>50</sup> = in-vitro inhibitory concentration, expressed in µg/mL, required to scavenge 50% of the generated reactive oxygen species (mean ± SD, *n* = 3). <sup>a</sup> Results for ROO• scavenging activity are expressed as slop ratio between samples or positive controls and Trolox. Ssample = slope of extract/positive controls curves and STrolox = slope of Trolox curve.

#### 3.3.1. Superoxide Radical Scavenging Assay

As can be observed in Table 3, the TR extract presented a lower scavenging capacity of O<sup>2</sup> •- (IC<sup>50</sup> = 83.67 ± 5.84 µg/mL) than the positive controls employed (GA and catechin), which means that higher concentrations of extract would be needed to obtain the same results of the positive controls. Farhadi et al. assessed the percentage of O<sup>2</sup> • scavenging activity of skins, pulps, seeds, canes and leaves of five native grape cultivars in west Azerbaijan province (Iran) [37]. Even though the extracts obtained from grape skins exhibited the highest O<sup>2</sup> • scavenging activity (with values ranging from 86.15% to 89.92% of inhibition), vine-canes also proved to be a promising matrix (with inhibition percentages ranging from 81.46% to 86.34%). In another study performed by Barros et al., grape stem extracts were also proposed as good O<sup>2</sup> • scavengers [38]. The authors analyzed grape stems from seven Portuguese *V. vinifera* L. varieties, four red and three white. The IC<sup>50</sup> values ranged from 970 to 2010 µg/mL for the Rabigato and Tinta Barroca grape stems varieties, respectively, being 11-fold higher than the result obtained in the present work, which demonstrates that extracts obtained from vine-canes through SWE were more efficient in scavenging O<sup>2</sup> •- .

#### 3.3.2. Hypochlorous Acid Scavenging Assay

Regarding the HOCl scavenging assay, catechin was the best scavenger presenting an IC<sup>50</sup> value of 0.18 ± 0.01 µg/mL. Similarly to the O<sup>2</sup> •- assay, the TR extract showed a lower HOCl quenching power (IC<sup>50</sup> = 33.94 ± 2.95 µg/mL) than catechin and GA (IC<sup>50</sup> = 1.25 ± 0.05 µg/mL). Wada et al. evaluated the scavenging capacity of hypochlorite ion (ClO−) by grape seed extracts using a concentration range of 0.02–2 mg/mL [39]. Two commercial grape seed extracts purchased from Mitsubishi Chemical Corporation (Kanagawa) were used in this study: extract A (proanthocyanin 99%) and B (proanthocyanin >80%). The ClO<sup>−</sup> quenching capacity of grape seed extracts A and B at 1.0 mg/mL were 27.7% ± 4.2% and 22.0% ± 3.7%, respectively. The authors used five reference compounds to compare the results obtained for the extracts, namely trans-resveratrol, chalcone, cyanidin, delphinidin and pelargonidin. These grape seeds extracts exhibited lower ClO<sup>−</sup> quenching effects than trans-resveratrol, cyanidin and delphinidin, but higher scavenging efficiency compared to chalcone [39].

#### 3.3.3. Peroxyl Radical Scavenging Assay

According to the obtained results (Table 3), the TR extract showed a low quenching capacity against ROO• with Ssample/STrolox lower than 1 (Ssample/STrolox = 0.024 ± 0.001). Noteworthy, catechin presented the highest scavenging ability for ROO• (Ssample/STrolox = 7.592 ± 0.074), followed by GA (Ssample/STrolox = 1.119 ± 0.005). In fact, the obtained subcritical water extracts were mainly composed by gallic acid and catechin, which were the main contributors to the phenolic profile of TR. However, despite their higher amount in comparison to the other phenolic compounds identified and quantified in the extracts, the levels found and recovered were lower to produce the same effect as the positive standards alone. Additionally, a synergistic effect of all the compounds present in TR extract could also negatively affect the capacity to scavenge the peroxyl radical (ROO• ).

Although the TR extract exhibited a low capacity to quench ROO• , previous studies described no scavenging activity of different byproducts for this radical [40,41]. Tournour et al. also determined the ORAC values of extracts from grape pomace of four Douro's *V. vinifera* L. varieties, including TN and TR [42]. The obtained results corresponded to 2337 ± 368 and 1054 ± 199 µmol TE/g of dw, for the TN and TR varieties, respectively. According to the authors, TN presented the best ORAC values, being therefore also better than the ones obtained for TR. Nevertheless, the highest TPC and the best result for the DPPH-RSA assay, obtained in the study aforementioned, also corresponded to the TN variety, which suggest that the radical scavenging capacity is directly related to the results obtained by both assays. In the present study, the TR variety was the one that achieved the highest values for the TPC and DPPH assays, which supports the choice made to use this extract for further studies. Barros et al. also assessed the ORAC values of seven Portuguese grape stems varieties extracts, obtained through sonication baths [38]. The results obtained ranged from 40.26 to 150.79 mM Trolox/100 g of dried grape stems. A direct comparison with the values from the present work cannot be made, but the results reported by these authors supported the ones obtained in O<sup>2</sup> • scavenging assay, where higher values were reported for the red vine varieties.

Among the ROS studied, the highest scavenging efficiencies of the TR extract were observed for HOCl (IC<sup>50</sup> = 33.94 ± 2.95 µg/mL) and O<sup>2</sup> •- (IC<sup>50</sup> = 83.67 ± 5.84 µg/mL). According to the results observed for these radical scavenging assays, TR vine-canes extracts obtained by SWE at 250 ◦C showed promising results for applications in health-related products.

#### *3.4. Cell Viability Studies*

The cell viability results are represented in Figure 3 and were obtained after exposure of HFF-1 to different concentrations of TR extract (0.1; 1.0; 10; 100 and 1000 µg/mL).

As it is possible to observe, extract concentrations under 100 µg/mL did not result in a reduction of HFF-1 cellular viability. For these concentrations, the cell viability was higher than 100%. However, at a concentration of 1000 µg/mL there was a considerable reduction of cell viability to 52.15% (*p* < 0.05). Manca et al. incorporated grape seed and vine-canes extracts in vesicular systems designed for topical applications and analyzed their cell viability in HFF-1 [43]. In the MTT test, the authors incubated the cells solely with the extracts to serve as a comparison value and with the vesicles in four different concentrations, namely 0.2, 2, 20 and 40 µg/mL, for 48 h. After the incubation time, the best values were obtained for the cells incubated only with the grape seed and vine-canes extracts (>100%). These results are in accordance with the ones obtained for the TR extract, since until the concentration of 100 µg/mL the cell viability was also above 100%.

**Figure 3.** Effect of different concentrations of TR extract on the metabolic activity of HFF-1 cells, measured by the MTT assay. Values were expressed as mean ± SD (*n*=4). \* means significant differences (*p* < 0.05).

#### **4. Conclusions**

The present work proposes an alternative application for the reuse of vine-canes, an agro-industrial byproduct derived from grape production. Grapes are one of the major fruit crops produced throughout the world and after harvesting season the vines need to be pruned, which generates every year large amounts of wastes. The challenge is to find alternative solutions to the presently used applications (incorporation in the soil), considering not only economic profits but, mostly, environmental impacts. The results obtained proved that SWE is a suitable environmentally friendly technique for vine-canes extraction of bioactive compounds with antioxidant properties. The TR extract obtained the highest results with the highest temperature tested (250 ◦C), achieving a TPC of 181 ± 12 mg GAE/g dw and an antioxidant activity through the FRAP assay of 202 ± 14 mg AAE/g dw. In general, a tendency was observed: extracts with higher phenolic content also presented higher flavonoid content, higher DPPH-RSA and FRAP values. Indeed, extracts obtained at 250 ◦C presented higher results than the ones obtained at 125 ◦C.

Regarding the HPLC-PDA analysis, the TR variety presented the higher amount of individual phenolic compounds for the highest extraction temperature tested. The phenolic compound present in higher amounts was GA (1145 ± 57 and 1502 ± 75 mg of GA/100 g of dw for the TN and TR varieties extracts obtained at 250 ◦C, respectively), which seems to be the main compound contributing for the high antioxidant activity of these extracts. Taking into consideration the results obtained, TR extract prepared at 250 ◦C was selected for the radicals scavenging and cell assays, revealing a good scavenging capacity of oxygen species as well as no adverse effects on HFF-1 were detected until a concentration of 100 µg/mL.

Considering the portion that grape production occupies in cultivation of food crops worldwide, the use of vine-canes in added value products could have a significant impact in the sustainability of food and beverage industries. Indeed, final applications in functional foods or cosmetics would be of extreme interest. As future perspectives, it would be interesting to test a scale up for SWE, since at the industrial level it might be more profitable.

**Author Contributions:** Conceptualization, M.M.M., F.R. and C.D.-M.; Funding acquisition, A.F.P., C.F., F.R., P.C. and C.D.-M.; Investigation, O.D., D.P., M.M.M., F.R., P.C. and A.F.P.; Methodology, O.D. and D.P.; Project administration, A.F.P., C.F., F.R. and C.D.-M.; Resources, A.F.P., C.F., F.R. and C.D.-M.; Supervision, M.M.M., F.R. and C.D.-M.; Writing—original draft preparation, O.D., D.P., M.M.M., F.R.; Writing—review and editing, M.M.M, F.R., A.F.P., C.F., P.C. and C.D.-M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by FCT/MCTES through national funds (UIDB/50006/2020). This work was also financed by the FEDER Funds through the Operational Competitiveness Factors Program—COMPETE and by National Funds through FCT within the scope of the project "PTDC/BII-BIO/30884/2017—POCI-01-0145-FEDER-030884" and project "PTDC/ASP-AGR/29277/2017-POCI-01-0145-FEDER-029277".

**Acknowledgments:** O.D. is thankful for the research grant from project PTDC/BII-BIO/30884/2017—POCI-01-0145-FEDER-030884. D.P. is thankful for the PhD grant (SFRH/BD/144534/2019) financed by POPH-QREN and subsidized by the European Science Foundation and Ministério da Ciência, Tecnologia e Ensino Superior. M.M.M. (project CEECIND/02702/2017) and A.F.P. (DL57/2016–Norma transitória) are grateful for the financial support financed by national funds through FCT—Fundação para a Ciência e a Tecnologia, I.P. and to REQUIMTE/LAQV. The supply of the vineyard pruning is acknowledged to Sogrape, S.A. The authors are grateful to João Vasconcellos Porto for the provided identification of the studied plant material.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses or interpretation of data; in the writing of the manuscript or in the decision to publish the results.

### **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* **Antioxidant, Antidiabetic, and Antiobesity Properties, TC7-Cell Cytotoxicity and Uptake of** *Achyrocline satureioides* **(Marcela) Conventional and High Pressure-Assisted Extracts**

**Adriana Maite Fernández-Fernández <sup>1</sup> , Eliane Dumay <sup>2</sup> , Françoise Lazennec <sup>2</sup> , Ignacio Migues <sup>3</sup> , Horacio Heinzen <sup>3</sup> , Patricia Lema <sup>4</sup> , Tomás López-Pedemonte <sup>1</sup> and Alejandra Medrano-Fernandez 1,\***


**Abstract:** The growing incidence of non-communicable diseases makes the search for natural sources of bioactive compounds a priority for such disease prevention/control. *Achyrocline satureioides* ('marcela'), a plant rich in polyphenols and native to Brazil, Uruguay, Paraguay, and Argentina, could be used for this purpose. Data on its antidiabetic/antiobesity properties and cellular uptake of bioactive compounds are lacking. The potentiality of non-thermal technologies such as highhydrostatic pressure (HP) to enhance polyphenol extraction retains attention. Thus, in the present study aqueous and ethanolic marcela extracts with/without assisted-HP processing were chemically characterized and assessed for their in vitro antioxidant capacity, antidiabetic and antiobesity activities, as well as cellular cytotoxicity and uptake on intestinal cell monolayers (TC7-cells, a clone of Caco-2 cells). Aqueous and ethanolic conventional extracts presented different polyphenolic profiles characterized mainly by phenolic acids or flavonoids, respectively, as stated by reverse phase-high-performance liquid chromatography (RP-HPLC) analyses. In general, ethanolic extracts presented the strongest bioactive properties and HP had none or a negative effect on in vitro bioactivities comparing to conventional extracts. TC7-cell viability and cellular uptake demonstrated in conventional and HP-assisted extracts, highlighted the biological effects of marcela bioactive compounds on TC7-cell monolayers. TC7-cell studies showed no HP-induced cytotoxicity. In sum, marcela extracts have great potential as functional ingredients for the prevention/treatment of chronic diseases such as diabetes.

**Keywords:** bioactive compounds; cell metabolism; flavonoids; high-hydrostatic pressure; marcela; phenolic compounds; TC7-cellular uptake

#### **1. Introduction**

*Achyrocline satureioides* (known by the popular name of 'marcela') could be used for the prevention/treatment of non-communicable chronic diseases including cardiovascular diseases, cancers, respiratory diseases, and diabetes [1], which are the main cause of deaths in the current times. Thus, the search for antioxidant, antidiabetic, and antiobesity natural sources is of great importance for their prevention/treatment. Marcela has been studied for its antioxidant, cell cytoprotective effect against oxidants [2], anti-inflammatory,

**Citation:** Fernández-Fernández, A.M.; Dumay, E.; Lazennec, F.; Migues, I.; Heinzen, H.; Lema, P.; López-Pedemonte, T.; Medrano-Fernandez, A. Antioxidant, Antidiabetic, and Antiobesity Properties, TC7-Cell Cytotoxicity and Uptake of *Achyrocline satureioides* (Marcela) Conventional and High Pressure-Assisted Extracts. *Foods* **2021**, *10*, 893. https://doi.org/ 10.3390/foods10040893

Academic Editors: Francisca Rodrigues and Cristina Delerue-Matos

Received: 3 March 2021 Accepted: 12 April 2021 Published: 19 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 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 (https:// creativecommons.org/licenses/by/ 4.0/).

analgesic, antispasmodic, constipating, sedative, immunomodulatory, antiviral, antiherpetic, choleretic and hepatoprotective actions, whereas partial cytotoxicity in mice and rats has been found for aqueous and ethanolic extracts [3]. It is a plant native to Brazil, Uruguay, Paraguay, and Argentina, commonly used as herbal tea [2–4]. Recently, marcela proved to present anti-cancer activity against glioma cell lines (U87, U251 and C6) and to be less cytotoxic to brain cell than gliomas [5]. However, no scientific studies regarding its antidiabetic and antiobesity underlying mechanisms have been reported. Marcela extracts are composed of flavonoids such as quercetin, luteolin and 3-*O*-methylquercetin in their glycosylated and aglycone forms [2], found in both ethanolic [6] and aqueous [2] extracts. These compounds possess several bioactive properties such as cytoprotective activity against oxidant agents [2], but there are no reports on the bioavailability and/or absorption experiments, neither about its cytotoxicity on intestinal cells (as a means to elucidate the effect after their ingestion), which are necessary to assess the potential effectiveness of the marcela bioactive compounds. Once absorbed, these compounds may exert the above-mentioned bioactivities.

Aqueous and ethanolic extracts have shown different polyphenolic profiles as a consequence of different polyphenols solubility correspondent to solvent polarity, with subsequently different bioactive properties and/or biological effectiveness. High hydrostatic pressure (HP) technology proved to increase polyphenolic extraction yields [7] and plant cell membrane damage [8]. HP can also disrupt weak bonds such as hydrophobic bonds subsequently generating conformational changes as well as denaturating cell proteins, which could lead to enhance compounds accessibility during extraction [8]. HP technology could be a resourceful procedure for *Achyrocline satureioides* polyphenols extraction by the use of moderate or no heat treatment [7,9], being especially useful for thermolabile compounds extraction [10]. However, these compounds could suffer modifications during the process. Thus, studies regarding bioactivity, absorption and cytotoxicity are needed in order to state if this innovative technology presents advantages related to conventional extraction procedures as well as to ensure this novel extracts are safe for consumption.

The aim of the present work is to evaluate *Achyrocline satureioides* antioxidant, antidiabetic and antiobesity properties of aqueous and ethanolic extracts compared to HP-assisted extracts, along with the exposure to cultures of intestinal cells in order to elucidate the degree of cytotoxicity (as assessed on cellular metabolic activity and cell membrane integrity) and uptake/absorption of extracted bioactive compounds.

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

#### *2.1. Raw Material and Chemicals*

*Achyrocline satureioides* (marcela) commercial samples were purchased in a pharmacy store (La Botica del Señor, Montevideo, Uruguay), and milled with a domestic coffee mill. All the reagents used in physicochemical characterization analyses were of reagent grade. Phenolic acids (gallic, chlorogenic and caffeic acids) and quercetin standards were purchased from Sigma-Aldrich (St. Louis, MO, USA) and used for marcela extract composition by reverse phase-high-performance liquid chromatography (RP-HPLC) and reverse phase-ultra-high-performance liquid chromatography (RP-UHPLC) analyses. Antioxidant assays reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA): Folin reagent, 2,20-azinobis-(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS), 6-hydroxy-2,5,7,8-tetramethylch-roman-2-acid (Trolox), fluorescein (FL) disodium salt, 2,20-azobis (2-methylpropionamidine) dihydrochloride (AAPH). Antidiabetic assays reagents were also purchased from Sigma-Aldrich (St. Louis, MO, USA): bovine serum albumin (BSA), methylglyoxal (MGO), aminoguanidine (AG), α-glucosidase (rat intestine acetone powder), acarbose, 4-methylumbelliferyl-α-D-glucopyranoside, human saliva α-amylase (Type IX-A), starch, maltose standard, 3,5-dinitrosalicylic acid. Antiobesity assay reagents were the following: lipase from porcine pancreas, 4-methylumbelliferyl oleate (4-MUO), and dimethyl sulfoxide (DMSO), which were purchased from SigmaAldrich (St. Louis, MO, USA), and orlistat standard was purchased from Alfa Aesar (Haverhill, MA, USA).

#### *2.2. TC7-Cells and Reagents for Cell Culture*

TC7-cells (a clone of Caco-2 cells) was kindly provided by Dr. Rousset (Centre de Recherche des Cordeliers, UMR S872, Paris, France). For cell culture, the following reagents were used.

High-glucose Dulbecco's modified Eagle medium (DMEM) with L-glutamine and pyruvate (Phenol red-DMEM), high-glucose Dulbecco's modified Eagle medium without L-glutamine and pyruvate (Phenol red-free DMEM), Dulbecco's phosphate-buffered saline (DPBS) + Ca2+ and Mg2+, Hank's Balanced Salt Solution (HBSS), penicillinstreptomycin mixture, MEM non-essential amino acid and foetal bovine serum (FBS) from GibcoTM were purchased from Life Technologies (Villebon-sur-Yvette, France). For MTT-assay, 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl tetrazolium bromide (MTT) and dimethyl sulfoxide (DMSO) were purchased from Sigma-Aldrich (St-Quentin Fallavier, France). The β-nicotinamide adenine dinucleotide hydrate (NAD), Trizma® base (Tris), L-(+)-lactic acid needed for LDH-assay, and Quercetin (96% dry matter, 95% purity) came from Sigma-Aldrich (St-Quentin Fallavier, France). Triton® X-100 came from Merck (Darmstadt, Germany).

#### *2.3. Methods*

#### 2.3.1. Preparation of Marcela Extracts

Marcela aqueous extract (Mac) was obtained by adding 100 g of milled marcela powder to 1000 mL of distilled water. The mix was boiled for 1 h, filtered with paper (Whatman n◦1) and the liquid was freeze-dried until constant weight (96 h).

Marcela ethanolic extract (Me) was obtained by adding 100 g of milled marcela powder to 1000 mL of ethanol (95%) followed by maceration at 20 ◦C for 24 h. The mixture was filtered with paper (Whatman n◦1) then rotavaporated (60 ◦C, 120 rpm, under reduced pressure, approximately 13 kPa) to dryness and 30 mL of distilled water was added to recuperate polyphenols. Afterwards, the liquid was freeze-dried until constant weight (96 h).

High pressure (HP) extracts were prepared using a high-pressure unit Model S-IL-100-250-09W (HP Food Processor, Stansted Fluid Power, Ltd., Harlow, UK) located in Laboratorio Tecnológico del Uruguay pilot food plant (Montevideo, Uruguay). The pressure chamber (2 L volume, 100 mm bore internal diameter, 250 mm long) has inside the canister to hold samples. The vessel body and the pressure-transmitting fluid (water) were kept at treatment temperature (25 ◦C) by circulating water through an internal heat transfer jacket fitted to the outside of the high-pressure barrel assembly. The temperature of the pressure-transmitting fluid was monitored with a thermocouple positioned at the chamber bottom. Before treatment, samples were individually packed in Cryovac® pouches (Sealed Air®, Charlotte, North Carolina, USA) by adding 10 g of milled marcela powder to 100 mL of phosphate buffer pH 7.9 for aqueous HP extract, or to 100 mL of ethanol (95%) for ethanolic HP extract, then vacuum sealed. Samples were introduced in the pressurization chamber previously thermostated at 25 ◦C then submitted to 400 MPa and 25 ◦C for 1 min, in the case of ethanolic HP extract (Me HP), or 200 MPa and 25 ◦C, at pH 7.9 for 1 min for aqueous HP extract (Mac HP). These conditions were selected in previous trials studying optimum conditions for marcela antioxidant compounds extraction through HP procedure. Pressure was raised from 0.1 MPa at a rate of 100 MPa per 30 s, maintained at the desired pressure level for 1 min then reduced down to 0.1 MPa in less than 30 s. Sample blanks were also prepared in the same way at 0.1 MPa and 25 ◦C for 1 min but without the pressure-processing step, by adding 10 g of milled marcela powder to 100 mL of phosphate buffer pH 7.9 for aqueous HP blank extract (Mac HP BL), or to 100 mL of ethanol (95%) for ethanolic HP blank extract (Me HP BL). Afterwards, liquid samples were freeze-dried until constant weight (96 h). All the extracts were stored at −20 ◦C for further analyses.

#### 2.3.2. Proximate Analysis

To characterize the initial sample or raw material (marcela powder), different parameters were determined: fat, protein, ashes, dietary fiber, moisture, and total carbohydrates (by difference using protein, moisture, ashes, and fat content). All determinations were performed at least in triplicate as in Association of Official Analytical Chemists (AOAC) [11] methods. Briefly, protein content was determined by Kjeldhal method using the conversion factor 6.25, moisture was determined using a conventional oven at 105 ◦C till constant weight, ashes was determined by using a furnace at 525 ◦C for 8 h, and fat content was obtained by Soxhlet method for 6 h using petroleum ether.

#### 2.3.3. RP-HPLC and RP-UHPLC Analyses

To obtain the polyphenolic profile of extracts, each of the extract samples were eluted by RP-HPLC (Shimadzu, SPD-20A detector and LC-10AT pump) according to De Souza et al. [6] with detection at 370 nm in a Jupiter C18 reverse phase column and an isocratic flow program of 1 mL/min. The mobile phase was composed by methanol:0.16 M phosphoric acid, in a ratio of 53:47 (*v*/*v*). The injection volume was 20 µL.

Each of the samples were also eluted by RP-UHPLC according to Reza et al. [12] with some modifications. RP-UHPLC UltiMate 3000 (Thermo Fisher Scientific, Massachusetts, USA) was used with a diode array detection (DAD detector). The reverse phase column was a Thermo Scientific BDS Hypersil C18 (150 × 3 mm, 3 µm particle size) used at 1 mL/min flow rate. Mobile phase was composed by methanol, phosphoric acid (pH 2.81) and acetonitrile in gradient: time 0 min, 5% acetonitrile and 95% phosphoric acid (initial condition); time 10 min, 10% acetonitrile, 10% methanol and 80% phosphoric acid; time 20 min, 20% acetonitrile, 20% methanol and 60% phosphoric acid; time 40 min, 20% acetonitrile, 20% methanol and 60% phosphoric acid; time 45 min, 100% acetonitrile; time 50 min, 100% acetonitrile; time 55 min, 5% acetonitrile and 95% phosphoric acid (to return to the initial conditions). The duration of each run was 55 min. The injection volume was 20 µL. The software used was Dionex Chromeleon 7.1 SR2. Phenolic acids were quantified by detection at 290 nm and quercetin was quantified at 370 nm. Phenolic acids and quercetin were identified and quantified by the use of pure standards and the construction of calibration curves through the detection at 290 and 370 nm for phenolic acids and quercetin, respectively.

#### 2.3.4. Antioxidant Capacity

Total polyphenol content was performed by Folin–Ciocalteau method [13] as described by Fernández-Fernández, Iriondo-DeHond, Dellacassa, Medrano-Fernandez, and del Castillo [14], preparing sample solutions in distilled water and using a gallic acid standard curve (0.05–1.0 mg/mL). Results were expressed as mg GAE/g extract.

The ABTS method [15] was performed as described by Fernández-Fernández et al. [14], preparing samples in phosphate buffer (pH 7.4) and using a Trolox calibration curve (0.25–1.5 mM). Results were expressed as µmol TE/mg extract.

Oxygen radical antioxidant capacity-fluorescein (ORAC-FL) assay was performed by the method of Ou, Hampsch-Woodill, and Prior [16] modified by Dávalos, Bartolomé, and Gómez-Cordovés [17] as described in Fernández-Fernández et al. [14]. The area under the curve (AUC) of fluorescence vs. time were calculated and normalized to the AUC of the blank as follows: AUCantioxidant (trolox or sample)-AUCblank. Trolox calibration curve (AUC vs. [Trolox]) was constructed and results were expressed as µmol TE/mg extract.

All samples were prepared in triplicate and each one of the preparations was measured in triplicate.

#### 2.3.5. Antidiabetic and Antiobesity Activities

α-glucosidase inhibition capacity was evaluated as described by Fernández-Fernández et al. [14] as an antidiabetic strategy. Briefly, fluorescence measurements were displayed at 37 ◦C for 30 min (each minute) at 360 and 460 nm of excitation and emission wave lengths, respectively. Acarbose was used as reference with probed inhibition capacity.

α-amylase inhibition assay was evaluated as another antidiabetic strategy and performed as reported by Li, Yao, Du, Deng, and Li [18] with some modifications described by Fernández-Fernández et al. [19]. The inhibition capacity was calculated by taking positive control as 100% of enzyme activity.

Fluorescent advanced glycation end products (AGEs) formation was evaluated by determining BSA-MGO formation inhibition (antiglycant capacity) as another antidiabetic strategy, obtaining the IC<sup>50</sup> value [20]. Briefly, sample mixtures consisted of 500 µL BSA stock solution (2 mg/mL in PBS, 1 mg/mL final concentration), 25/50 µL 5/10 mM MGO stock solution (200 mM in PBS, 5 mM final concentration), different volumes of extracts from a sample stock solution of 50 mg/mL (concentrations 0.1-5 mg/mL) plus sufficient volume of PBS 10 mM pH 7.4 with 0.02% sodium azide to achieve 1 mL of the mixture final volume. Sample blanks consisted of the samples (different concentrations) with sufficient volume of PBS to achieve 1 mL of the mixture final volume (intrinsic fluorescence of the samples). Positive control was prepared by mixing 500 µL BSA, 25/50 µL 5/10mM MGO and 475/450 µL PBS, as previously explained. Aminoguanidine (AG) was used as reference (1, 4 and 8 mM final concentrations) mixed with BSA, MGO and PBS. All stock solutions were prepared in PBS 10 mM pH 7.4 with 0.02% sodium azide. Eppendorf tubes were incubated at 37 ◦C for 7 days. Fluorescence measurements were performed at 340 and 420 nm of excitation and emission wavelengths, respectively, and inhibition percentages were calculated by taking positive control as 100% of AGEs formation.

Pancreatic lipase inhibition capacity was determined as described in Fernández-Fernández et al. [14]. Measurements were determined after 30 min incubation at 25 ◦C by fluorescence measurements at 360 nm and 460 nm of excitation and emission wavelengths, respectively.

#### 2.3.6. TC7-Cell Culture and Sample Deposits

TC7-cells were routinely grown according to Benzaria et al. [21,22] with minor changes in 75 cm<sup>2</sup> sterile cell culture flasks in phenol red-DMEM culture medium. TC7-cells (passages 41-49) were seeded in sterile 12-well plates (3.5 cm2/well; Nunc, VWR, Fontenaysous-Bois, France) at a density of 2.5 × 10<sup>5</sup> cells/well (1 mL of cell suspension/well) then cultivated at 37 ◦C in controlled atmosphere (8% CO2, 92% air, 100% relative humidity, RH) (Thermo Scientific 8000 incubator, Thermo Electron, St-Herblain, France) for 19–20 days to reach cell-confluence and obtain differentiated cells, the culture medium (phenol red-DMEM supplemented with 20% *v/v* heat-inactivated FBS, 1% *v*/*v* penicillin-streptomycin and 1% *v/v* MEM non-essential amino acids), being changed every 2 days. Cell confluence was assessed by transepithelial electrical resistance (TEER; Millicell®-ERS volt-ohm meter, Millipore, St-Quentin-en-Yvelines, France) measurements before deposing extract samples onto the cells. For TEER measurements, cells were grown in sterile Transwell plates with ThinCert inserts (3 µm pore size; 1.13 cm2/well; Greiner Bio-one, VWR International, Fontenay-sous-Bois, France) at a density of 2.5 × 10<sup>5</sup> cells/well, obtaining TEER values of 750–800 Ω cm−<sup>2</sup> . Cell confluency of the cell cultures was also checked by inverse phase microscope examination.

After washing using Phenol red-free DMEM, differentiated TC7-cells were incubated for 3 h or 22 h at 37 ◦C in controlled atmosphere (8% CO2, 92% air, 100% HR) with 500 µL of extract mixture or control sample. Exposure times (3 h or 22 h) were chosen on the basis of previous experiments [22], and taking into account the open time necessary to prepare cell series. All cell seeding and sample deposit experiments were carried in sterile conditions under a laminar flow cabinet (PSM MSC Advantage, ThermoFisher Scientific, St-Herblain, France), using 0.2 µm filtrated media, sterile solutions and sterile plastic material (pipets, tips, flasks, plates, microplates, Eppendorf® and Falcon® tubes).

Ethanolic and aqueous extracts were assayed on TC7-cells for a range of lyophilized extract concentrations in the cell deposit medium. For this, a 56.6 mg of lyophilized

ethanolic extract (Me) was solubilized into 1 mL of 80% Ethanol (Ethanol/Water 80:20, *v*/*v*), and 28.3 mg of lyophilized aqueous extract (Mac) into 1 mL of sterile distilled water, due to its solubility. Various concentrations of lyophilized extracts were then prepared ranging from 0.88 to 56.6 mg/mL in 80% Ethanol for Me, and ranging from 3.54 to 28.3 mg/mL in sterile distilled water for Mac. In the case of HP extracts, samples were solubilized in the range 0.88–14.2 mg/mL in 80% Ethanol for Me HP, and in the range 3.54–28.3 mg/mL in sterile distilled water for Mac HP. Mixtures of 100 µL of the latter extract solutions and 1.9 mL of Phenol red-free DMEM were then prepared for cell deposit in 12-well plates. A 500 µL of mixture per well was deposed onto TC7-cells in triplicate. Control samples were also prepared using Phenol red-free DMEM alone, or 1.9 mL Phenol red-free DMEM plus 100 µL of one of the following solutions: distilled water, 80% Ethanol, 0.1% Triton X100, or purified quercetin at 0.156–0.625 mg/mL in 80% Ethanol.

#### 2.3.7. Determination of In Vitro TC7-Cell Membrane Integrity and Cell Metabolic Activity

After 3 h or 22 h of exposure time with Me or Mac extracts, or control samples, the apical TC7-culture supernatants were collected on ice to determine the lactate dehydrogenase (LDH) activity, i.e., LDH-release from cytosols in cellular apical media. TC7-cells were recovered for the MTT colorimetric determination (i.e., evaluation of cellular metabolic activity or cell viability), both as described by Benzaria et al. [22] with minor modifications. LDH-leakage outside TC7-cells was determined to evaluate cellular membrane damage after exposure to extracts, as an indicative of further cell death. Apical TC7 media were collected then four-fold diluted (1/4) in Phenol red-free DMEM. A 25 µL of the latter solutions were added to 96-well plates (8 replicates for each apical diluted medium). Then, 250 µL of pH 9.3 NAD reagent (1.65 mM NAD, 165 mM KCl, 54 mM L-lactic acid, 108 mM Tris, final) was added per well. LDH induced the lactate oxidation into pyruvate with the simultaneous reduction of NAD to NADH. NADH absorbance was therefore measured in plate wells at 340 nm and 37 ◦C over 10 min (Multiskan Spectrum microplate reader, Thermo Electron, Vintaa, Finland). LDH activity was expressed as the difference in absorbance values taken at 0 and 10 min. Results were the means of eight absorbance determinations for each apical cell medium. A positive control was included in the series, corresponding to high LDH release by exposure of TC7-cells to Triton® X-100 in Phenol red-free DMEM (0.005% final, *v*/*v*).

TC7-cells in plate wells were then recovered for MTT-assay to evaluate cell viability after 3 h or 22 h of exposure time to Me or Mac extracts at various concentrations, or to control samples. After washing with Phenol red-free DMEM, cells were incubated for 3 h with 500 µL/well of MTT reagent (0.15 mg/mL MTT in FBS-free Phenol red-free DMEM) at 37 ◦C. MTT is reduced into Formazan® by a succinate dehydrogenase in living cells. After removing MTT solution, Formazan® was recovered by cell-lysing for 30 min at 37 ◦C using 1000 µL DMSO per well. Amounts of 100 µL were then transferred into 96-well plates to measure Formazan® absorbance at 570 nm (Multiskan Spectrum microplate reader) (8 replicates for each apical cell lysate). The cell ability to reduce MTT provides an indication of mitochondrial integrity, and therefore of cell metabolic activity or cell viability. Results were expressed as the means of 8 absorbance determinations for each apical cell lysate sample.

For each series ("3 h or 22 h of exposure time"), data were pooled from 4 independent cell culture experiments involving different TC7-cell passages. For Me and Mac extracts, 3 to 4 independent cell culture experiments were carried out on different days, and 1 to 3 for Me HP and Mac HP, each with currently 2-3 apical cell media analyzed per studied extract concentration.

#### 2.3.8. Marcela Bioactive Compounds' Uptake

Cellular uptake of marcela compounds was determined after TC7-cell exposure to Me or Mac extracts for 3 h or 22 h, in triplicate, as described in Section 2.3.6. Me and Me HP samples were deposed at 0.088–0.354 mg/mL extract onto TC7-cells; Mac

and Mac HP samples, at 0.71–1.42 mg/mL extract. After cell incubation, the apical culture media were taken off and the plate-wells washed with DPBS+. TC7-cells were scratched with 500 µL of cold acidified methanol (methanol with 0.1% *v/v* acetic acid) and transferred into Eppendorf tubes for centrifugation at 10,000 rpm and 4 ◦C for 5 min. Methanolic supernatants of centrifugation were kept in brown vials at 4 ◦C until further analysis by RP-UHPLC as described in Section 2.3.3. Prior to sample injection (20 µL) in RP-UHPLC, supernatants were dried at 40 ◦C in a conventional stove and re-suspended in 250 µL of methanol.

#### 2.3.9. Statistical Analysis

All experiments were performed in triplicate and cell studies were performed at least in three different passages. The statistical analysis was established by analysis of variance (ANOVA). Results were expressed as means ± standard deviation (SD) (*n* = 3). Tukey test was applied to determine significant differences between values (*p* < 0.05) using Infostat v. 2015 program. Different letters state significant differences when *p* < 0.05.

In the case of cell studies, the statistical analysis was carried out using Sigma Plot vs. 11.0 program: the pooled data were analyzed by one-way analysis of variance (ANOVA) with all pairwise multiple comparison procedure (Holm–Sidak test) and an overall significance level of 0.05.

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

#### *3.1. Chemical Composition*

Aqueous and ethanolic extracts resulted in a yield of 6.9 and 3.1% *w*/*w*, respectively. The results of the proximate analysis showed in marcela powder (raw material) high fiber and carbohydrates contents. Marcela powder contained per 100 g: 4.77 ± 0.02 g proteins, 4.52 ± 0.18 g lipids, 21.01 ± 0.79 g carbohydrates (without fiber), 57.22 ± 0.73 g fiber, 4.94 ± 0.05 g ashes, and 7.54 ± 0.13 g moisture.

As to RP-HPLC (Figure 1A,B), and RP-UHPLC results (Figure 1C,D), the extracts presented a typical chromatogram as previously reported by De Souza et al. [6] in *Achyrocline satureioides* preparations. According to De Souza et al. [6] the three prominent peaks correspond to quercetin, luteolin and 3-*O*-methylquercetin, in order of appearance in the chromatogram (Figure 1B). In the present study, retention times (RT) were lower because of using a higher flow (1 mL/min) than 0.6 mL/min. Furthermore, the last prominent peak at RT 19 min in Me chromatogram (Figure 1B) could correspond to achyrobichalcone according to Zorzi et al. [23]. Polyphenol profile of marcela extracts in the current study are also in agreement with those reported by Martínez-Busi et al. [24].

Chromatograms at 370 nm showed that Me extract presents mainly flavonoids such as 3-*O*-methylquercetin (30% in the extract as stated in relative area, RA), being in a close proportion to quercetin (22% RA) (Figure 1B). In contrast, Mac chromatogram at 290 nm presented mainly phenolic acids (over 65% RA, Figure 1C) compared to flavonoids in which quercetin (5% RA) is in lower proportion than 3-*O*-methylquercetin (7–8% RA) as shown by Mac chromatogram at 370 nm (Figure 1A). These results agree with those reported by Polydoro et al. [25] where the aqueous extract of *Achyrocline satureioides* presented the lowest contents of quercetin, luteolin and 3-*O*-methylquercetin. Moreover, they found higher concentrations of the latter compounds in the extract with higher proportion of ethanol (80%) with similar ratio of quercetin to 3-*O*-methylquercetin. In the present study, quercetin was quantified in Mac and Me extracts by RP-HPLC obtaining 1.98 ± 0.13 and 88.9 ± 6.36 mg of quercetin/g extract, respectively. Calculating from quercetin calibration curve, 3-*O*-methylquercetin was estimated to 3.3 ± 0.3 mg and 127.1 ± 9.1 mg/g extract for Mac and Me, respectively. The Me extract in the present study showed greater quercetin content than the marcela aqueous extracts prepared by maceration and ultrasound-assisted extraction reported by Guss et al. [26].

**Figure 1.** RP-HPLC chromatogram at 370 nm of samples aqueous extract (Mac 10 mg/mL) (**A**) and ethanolic extract (Me 1 mg/mL) (**B**). RP-UHPLC chromatogram at 290 nm of aqueous extract (Mac 2 mg/mL) (**C**) and at 370 nm of ethanolic extract (Me 1 mg/mL) (**D**). In order of appearance according to the retention times: GA, gallic acid; Cl, chlorogenic acid; CA, caffeic acid; Q, quercetin; Lu, luteolin; 3-*O*-MQ, 3-*O*-methylquercetin. 24

In parallel, Me and Mac extracts were eluted by RP-UHPLC method (Figure 1C,D) in which, phenolic acids were identified and quantified. The composition in phenolic compounds is associated with the type of solvent used. Mac was characterized by the presence of gallic acid (RT 1.1 min), chlorogenic acid (RT 4.6 min) and caffeic acid (RT 5.4 min) with 11.9, 15.1 and 18.7% RA, respectively (Figure 1C) which corresponds to 3.76 ± 0.25 mg of gallic acid, 14.28 ± 0.01 mg of chlorogenic acid and 3.28 ± 0.62 mg of caffeic acid per g of Mac extract, in agreement with previous reports [27,28].

#### *3.2. Antioxidant Capacity*

As to total polyphenol content, aqueous extracts (Mac, Mac HP BL and Mac HP) presented lower content than ethanolic ones (Me, Me HP BL and Me HP) indicating ethanol favors polyphenols extraction (Table 1). For ABTS antioxidant capacity (Table 1), the tendency was different resulting Me as the best, followed by all the other extracts with no significant differences (*p* > 0.05). For ORAC-FL antioxidant capacity (Table 1), the highest antioxidant potential were Mac and Me, followed by ethanolic extracts Me HP BL and Me HP. When analyzing HP-assisted extracts, both aqueous and ethanolic HP extracts (Mac HP and Me HP) presented a lower polyphenol content by 3.2% and 7.6%, respectively (although non-significant, Table 1) when compared to their respective blanks (Mac HP BL and Me HP BL). Thus, high hydrostatic pressure did not significantly affect total polyphenol content nor antioxidant capacity when applied at the tested conditions to the crude extracts dispersed in phosphate buffer or 95% ethanol (Section 2.3.1).



Results are expressed as mean values ± SD (*n* = 3). ANOVA analysis was performed by column using Tukey test to state significant differences. Different letters indicate significant differences (*p* < 0.05) between values in the same column. Sample solutions were prepared in triplicate and assayed in triplicate. Marcela aqueous (Mac) and marcela ethanolic (Me) extracts. Marcela aqueous high pressure-assisted (Mac HP) and marcela ethanolic high pressure-assisted (Me HP) extracts. Blank of marcela aqueous HP (Mac HP BL) and marcela ethanolic HP (Me HP BL) extracts (Section 2.3.1).

> Compared to other studies, these extracts presented similar total polyphenol content to the marcela extracts reported by Ferraro et al. [29] (23.0-112.6 mg GAE/g) with the highest polyphenol content, observing a correlation with antioxidant capacity determined by DPPH. In contrast, Guss et al. [26] reported greater polyphenol content and ABTS antioxidant capacity (i.e., lower IC<sup>50</sup> value) of marcela maceration and ultrasound-assisted ethanolic extracts. In the current work, Me IC<sup>50</sup> value was of 354 ± 25 µg/mL compared to 21.8 ± 0.8 and 21.3 ± 0.4 µg/mL for marcela maceration and ultrasound-assisted ethanolic extracts [26]. Marcela ethanolic extract (Me) presented higher total polyphenol content when compared to other medicinal herbs such as *Mentha x piperita* L., *Peumus boldus* Mol. and *Baccharis trimera* Iless. aqueous and ethanolic extracts [30], as well as antioxidant capacity. In contrast with Irazusta et al. [30] results, marcela aqueous extract showed lower antioxidant capacity than ethanolic extract. Antioxidant capacity of crude methanolic extracts of native Australian mint and common spearmint showed 398.5 ± 19.3 and 403.5 ± 14.8 µmol TE/g extract for ABTS, and 1727.2 ± 183.5 and 1551.1 ± 137.4 µmol TE/g extract for ORAC-FL, respectively [31], presenting lower antioxidant capacity than Mac and Me extracts in the present study (Table 1).

#### *3.3. Antidiabetic Activities*

α-amylase and α-glucosidase inhibition capacities were assessed (Figure 2A,B) as a strategy for post-prandial plasma glucose level regulation through delay/inhibition of complex carbohydrates hydrolysis during digestion, such as starch, causing lower glucose absorption [14]. For α-amylase inhibition (Figure 2A), acarbose and quercetin presented the highest inhibitions with IC<sup>50</sup> values of 34.1 ± 0.8 and 2.4 ± 0.2 µg/mL, respectively. As to the extracts, aqueous extracts presented very low inhibition at the tested concentrations (up to 25 mg/mL, data not shown), in contrast with ethanolic extracts which showed IC<sup>50</sup> values of 515 ± 44 (Me), 2900 ± 51 (Me HP BL) and 7974 ± 422 µg/mL (Me HP), demonstrating HP negatively affects α-amylase inhibition capacity. Moreover, quercetin seems to be one of the responsible for ethanolic extracts inhibition capacity, because of being one of the main compounds present in the latter extracts. ‒

response curves of α ) and α used as α amylase and α **Figure 2.** (**A**,**B**) Dose-response curves of α-amylase (**A**) and α-glucosidase (**B**) inhibition capacities expressed as % inhibition vs. sample concentration (mg/mL). (**C**) Inhibition capacity (%) of fluorescent AGEs formation with methylglyoxal (MGO) at 5 or 10 mM by different sample concentrations (mg/mL). (**D**) Dose-response curves of pancreatic lipase inhibition capacity expressed as % inhibition vs. sample concentration (mg/mL). Samples are: marcela aqueous (Mac) and ethanolic (Me) extracts, marcela aqueous HP (Mac HP) and ethanolic HP (Me HP) extracts. Blank of marcela aqueous HP (Mac HP BL) and ethanolic HP (Me HP BL) extracts. Acarbose was used as α-amylase and α-glucosidase inhibitory agent (**A**,**B**). Orlistat was used as lipase inhibitory agent (**D**). Quercetin, caffeic acid, gallic acid and chlorogenic acid were used as standards (**A**,**B**,**D**).

quercetin (25 μg/mL) has For α-glucosidase inhibition (Figure 2B), acarbose (IC<sup>50</sup> = 4.0 ± 0.3 µg/mL) and chlorogenic acid (IC<sup>50</sup> = 69.1 ± 1.6 µg/mL) presented the highest inhibition capacity (lowest IC<sup>50</sup> value). Mac extract presented an IC<sup>50</sup> value of 150.8 ± 54.0 µg/mL, and 157.6 ± 23.3 µg/mL was found for Me extract with no significant differences (*p* > 0.05). For Mac HP and its blank (Mac HP BL), IC<sup>50</sup> values were 2973.1 ± 403.2 and 5392.0 ± 437.1 µg/mL respectively (significant difference for *p* < 0.05), stating bioactive compound extraction was favored by high hydrostatic pressure. In the case of Me HP and its blank (Me HP BL), IC<sup>50</sup> values were 2587.3 ± 214.5 and 2211.1 ± 196.0 µg/mL,

respectively, negatively affecting bioactivity by HP but with no significant differences (*p* > 0.05).

In accordance with the present work, quercetin has shown to possess more α-amylase inhibition capacity than acarbose [32]. Furthermore, other extracts from medicinal plants/herbs (Vietnamese and Amazonian plants, *Agrimonia asiatica*, species of *Myrcia* genus and *Euphorbia hirta* herbs) possessing phenolic acids (e.g., gallic acid) and flavonoids (quercetin and/or quercetin derivatives, and luteolin) like marcela extracts, have shown antidiabetic properties (α-amylase and α-glucosidase inhibition capacity) [33–36]. Particularly, *Euphorbia hirta* L. extract has shown to lower fast blood glucose level after 4 h and a significant reduction after 15 days treatment in streptozotocin-diabetic mice [36]. Guava (*Psidium guajava* L.) leaves possessing gallic, caffeic and chlorogenic acids, and quercetin, among others, have also shown antidiabetic properties [37]. The latter reports show the potential that marcela extracts could have as functional ingredients.

As another strategy for diabetes complications prevention/treatment, there is the inhibition of AGEs formation. Figure 2C showed maximum inhibition (close to 100%) of AGEs formation for Me extract at all the concentrations tested (0.1–5 mg/mL) in contrast with Mac extract that presented an increased inhibition trend with increasing concentration, although not significant (*p* > 0.05). This indicates that Me presents higher antiglycant capacity than Mac. Moreover, extracts did not present any differences when compared to methylglyoxal (MGO) at 5 and 10 mM. Inhibition capacity was not affected by MGO concentration at the tested concentrations (5 and 10 mM). Other medicinal herbs used as infusions, like marcela, have shown to inhibit AGEs formation such as *Mentha x piperita* L., *Peumus boldus Mol.* and *Baccharis trimera Iless.* [30] in a similar level as marcela extracts. Ethanolic extracts of ten common household condiments/herbs (*Allium sativum*, *Zingiber officinale*, *Thymus vulgaris*, *Petroselinum crispum*, *Murraya koenigii Spreng*, *Mentha piperita* L., *Curcuma longa* L., *Allium cepa* L., *Allium fistulosum* and *Coriandrum sativum* L.) showed correlation between total polyphenol content, antioxidant capacity and anti-glycant capacity [38], showing the same tendency when comparing marcela aqueous and ethanolic conventional extracts. Ethanolic extract (Me) showed higher total polyphenol content, antioxidant and anti-glycant capacity than aqueous extract (Mac).

#### *3.4. Antiobesity Activity*

Lipase inhibition capacity was determined (Figure 2D) as a strategy for post-prandial fat absorption control during digestion by delay/inhibition of triglycerides hydrolysis into free fatty acids, leading to lower fat absorption [14]. Mac extract presented an IC<sup>50</sup> value of 1.471 ± 0.103 mg/mL and ethanolic extract (Me) 0.219 ± 0.028 mg/mL, the latter having no significant difference (*p* > 0.05) with Orlistat IC<sup>50</sup> value (1.9 ± 0.2 µg/mL). Mac HP extract presented very little inhibition at the tested concentrations (0.1–10 mg/mL) and lower for its blank without pressure (Mac HP BL) (data not shown). Me HP extract presented an IC<sup>50</sup> value of 2.025 ± 0.053 mg/mL and its blank without HP (Me HP BL) of 1.634 ± 0.038 mg/mL, the latter having no significant differences with Mac (*p* > 0.05). Considering all extracts, ethanolic extracts presented the best inhibition capacity, although HP negatively affected the inhibition capacity when compared to the blank (increased IC<sup>50</sup> value of Me HP compared to Me HP BL) with significant differences (*p* < 0.05). Aqueous extracts seems to be more bioactive with applied high temperature (boiling extraction, Section 2.3.1) and for ethanolic extracts, it seems as if time was a key factor for bioactive compounds solvent extraction.

In parallel, polyphenol standards were tested finding IC<sup>50</sup> values of 4.566 ± 0.231, 0.332 ± 0.032 and 0.012 ± 0.001 mg/mL for caffeic, gallic and chlorogenic acids, respectively, stating gallic and chlorogenic acids as the main responsible for Mac antiobesity activity. Chlorogenic acid presented no significant differences with Orlistat (*p* > 0.05), followed by gallic acid, with no significant differences with Me (*p* > 0.05), and by quercetin with an IC<sup>50</sup> value of 1.105 ± 0.065 mg/mL (data not shown). In accordance with the present work, quercetin (25 µg/mL) has already been reported for inhibiting porcine pancreatic lipase by

a 27.4% [39]. Caffeic acid presented the lowest inhibition capacity of the samples tested in the current work (*p* < 0.05).

#### *3.5. Cell Studies*

Results of cell membrane integrity through LDH-assay after 3 h of exposure time to Me extract (Figure 3A) showed a significant increase in LDH activity up to a maximum being of Triton level (positive control for cell membrane disruption), for 0.35 and 0.71 mg/mL of final extract in cell deposit media, which corresponds to 0.104 and 0.208 µM/mL quercetin in Me, respectively. From 0.71 to 2.83 mg/mL of final Me extract in cell deposit media, LDH activity decreased down to DMEM value (negative control) which could be explained by a solubility loss of Me constituents (initially soluble in 80% Ethanol) at the highest concentrations in the cell deposit medium (i.e., in DMEM) during incubation. Indeed, it was checked by absorbance measurement at 370 nm of Me deposit mixtures, as carried out before and after centrifugation (1200 rpm for 4 min, 30 ◦C), a decrease in absorbance by 13.6%, 21.4% and 39.4%, for the 0.71, 1.42 and 2.83 mg/mL extract concentrations, respectively, due to some precipitate formation. Such precipitate could correspond to the most hydrophobic compounds and/or marcela fibers contained in Me extract (initially soluble in 80% ethanol, but no more in DMEM, i.e., an aqueous dispersion of amino-acids, vitamins, salts and glucose). Such precipitate on cell monolayers could limit the access of harmful compounds to the cell membrane, and therefore membrane damage. We have checked that the deposit of Me extract at the highest concentrations after centrifugation to exclude insoluble compounds displayed similar LDH activity values (data not shown) than that obtained without centrifugation (Figure 3). Such precipitate was not observed for Mac extract at the studied concentrations.

For Mac extract, values of LDH activity after 3 h incubation at the tested concentrations (0.177–1.42 mg/mL in cell deposit media) were maintained below 25% of Triton® value indicating no noticeable cell-membrane damage compared to control samples (DMEM ± water or ethanol) (Figure 3A).

For Me HP sample, LDH activity increased with the extract concentration as for the non-HP processed sample but significantly less steeply, presenting a maximum at 0.71 mg/mL, close to Triton level, as for the non-HP processed Me. In contrast, Mac HP sample maintained cell membrane integrity at all tested concentrations (0.177–1.42 mg/mL in cell deposit media) such as quercetin solutions (0.026–103 µM/mL), and close to DMEM level (negative control). A lower extraction of polyphenols during HP aqueous extraction might be the reason for the maintenance of cell membrane integrity, supported by total polyphenol content and bioactivity results shown above.

Generally speaking, LDH activity was higher after 22 h than 3 h of exposure time to all extract samples, and especially in the case of Me and Me HP samples (Figure 3A,B). Me induced a marked increase in LDH activity being of Triton level from lower concentrations (0.088–0.35 mg/mL in the cell deposit medium) than after 3 h incubation, followed by a significant decrease in LDH activity at ≥1.42 mg/mL extract. As previously explained for 3 h incubation, a decrease in Me constituent solubility at the highest studied concentrations (1.42–2.83 mg/mL) in cell deposit media probably explained the observed decrease in LDH activity. Such decrease in LDH activity was associated to a positive level of cell metabolic activity as evaluated by MTT-assay (Figure 3C,D).

For Mac extract, LDH activity after 22 h incubation was ≥ to that observed after 3 h, remaining ≤31% of Triton level and showing no noticeable or little cell membrane damage. For Me HP, the cell membrane integrity loss was significantly higher than after 3 h incubation and close to that observed for the non-HP processed sample. Mac HP and quercetin standard showed low LDH activity, ≤29% of Triton level indicating no or little cell membrane damage.

ining similar quercetin amounts (0.088‒0.354

mg/mL Me extract with 0.026‒0.104 μM/mL quercetin). Me and Me HP extracts displayed

Dulbecco's modified Eagle medium **Figure 3.** TC7-cell membrane integrity determined by LDH activity after 3 h (**A**) or 22 h (**B**) incubation, and TC7-cell viability assessed by MTT-assay after 3 h (**C**) or 22 h (**D**) incubation. Incubation of TC7-cells in the presence of marcela ethanolic extract (Me) or marcela aqueous extract (Mac) with or without HP-processing (HP), or purified quercetin (Q). Concentrations are expressed in mg/mL of marcela extracts or purified quercetin in the apical cell medium. Dulbecco's modified Eagle medium (DMEM) (±water or ethanol) was used as negative control and Triton as positive control. Bars and error bars represent the mean values and standard deviation, respectively. For each figure, the different letters on bars state significant differences for *p* < 0.05.

h of exposure to Me extract at its lowest studied concentrations (0.044‒0.088 mg/mL in cell (namely 0.177‒0.71 mg/mL) probably due As for cell metabolic activity, MTT-assay after 3 h of exposure time (Figure 3C) showed significant increases in cell viability for Me and Mac samples comparing with control samples (DMEM ± water or ethanol), indicating some benefit effect of both extracts on TC7-cells. For Me extract, a high metabolic activity was maintained with increasing concentration with a maximum at 0.177 mg/mL extract in the cell deposit medium, corresponding to 0.052 µM/mL quercetin. For Mac extract, cell viability progressively increased with the extract concentration in cell deposit media reaching a maximal value for a higher concentration (1.42 mg/mL) compared to Me (0.177 mg/mL), suggesting different metabolic mechanisms for both extracts due to their composition. In contrast, purified quercetin deposed at 0.026 to 0.103 µM/mL (i.e., 0.0078 to 0.031 mg/mL) in apical cell media did not induced some significant improvement in cell viability compared to control samples (DMEM ± water or ethanol), and remained well below Me extract deposits containing similar quercetin amounts (0.088–0.354 mg/mL Me extract with 0.026–0.104 µM/mL quercetin). Me and Me HP extracts displayed similar MTT-profiles as a function of extract concentration, indicating no particular benefit or detrimental effect from HP-process. In the opposite, Mac HP presented the same tendency as Mac but with lower values remaining at the quercetin or DMEM level, possibly due to lower total polyphenol content (Table 1).

Higher increases in cell metabolic activity was observed (Figure 3D) after 22 h than 3 h of exposure to Me extract at its lowest studied concentrations (0.044–0.088 mg/mL in cell deposit media), but not at higher concentrations (namely 0.177–0.71 mg/mL) probably due to an excessive cell membrane damage as assessed by cellular LDH-leakage (Figure 3B). Figure 3D shows similar viability profiles for Me and Me HP samples after 22 h than 3 h of exposure, suggesting no detrimental effect from HP-process on ethanolic extraction. Higher increase in cell metabolic activity was observed after 22 h than 3 h of incubation with Mac extract in the range 0.354–1.42 mg/mL in apical cell media, due to the longer exposure time without noticeable cell membrane damage. In the case of Mac HP, the longer incubation time (22 h) did not improve the observed cell metabolic activity comparing with DMEM controls, possibly due to low polyphenolic content shown in Table 1.

Taking into account the whole results, the ethanolic extract displayed higher effects on TC7-cells than the aqueous extract at similar tested concentrations, with a dose-time dependence coupling mechanisms both inside the cells after constituent uptake (i.e., cell viability) and at the cell membrane surface (i.e., membrane integrity). It would be interesting to identify both kinds of active constituents.

Me and Me HP appears beneficial to TC7-cell metabolic activity at the lowest tested concentrations (0.044–0.177 mg/mL extract) and exposure time (3 h). However, higher concentrations (0.35–0.71 mg/mL extract) and exposure time (22 h) induced dramatic LDHleakage. Indeed, a loss of cell-membrane integrity leads to further cell dysfunction then death. In contrast, Mac extract induced increased cell metabolic activity with increasing extract concentrations and exposure time, without noticeable loss of membrane integrity. However, the HP process led to a significant loss of its beneficial bioactive properties.

The fact that Me induced high levels of membrane damage and, simultaneously a high metabolic activity could result from a lag time between both mechanisms: membrane damage and decrease in mitochondrial activity; mitochondrial activity was still operating while the membrane started to be significantly damaged.

Purified quercetin deposed on TC7-cells at 0.026 to 0.103 µM/mL did not significantly increase cell metabolic activity as evaluated by MTT-assay, which does not highlight some prominent role of aglycone quercetin.

Previous reports of Polydoro et al. [25] showed ethanolic extracts (40 and 80% of ethanol) cytotoxicity assessed on Sertoli cells from Wistar rats at a concentration of 0.125 mg/mL with less than 80% of cell viability. Quercetin showed cytotoxicity [25] at high concentration (0.25 mg/mL or 0.827 µM/mL) which is higher than in the current study (0.026–0.103 µM/mL). Hence, the cytotoxicity observed by cell membrane disruption (LDH-assay) induced by ethanolic extracts in the present study might be displayed by other bioactive compounds than aglycone quercetin.

The RP-UHPLC analysis of TC7-cell methanolic extracts (Section 2.3.8) was carried out to detect the possible compound uptake by TC7-cells through their exposure to Me or Mac extracts compared to control DMEM. What was shown by RP-UHPLC chromatograms at 290 nm and/or 370 nm (Figure 4A–J) is outlined below.

A marked peak clearly appeared at 5.3 min on chromatograms at 290 nm, for most cellextract samples after 3 h or 22 h incubation, included DMEM control samples. The latter peak detected at 290 nm but not at 370 nm, and that absorb in the UV 200–295 nm range with a maximum at 285 nm (Figure 4E), could correspond to aromatic aminoacids/peptides and protein material present in the living cells. A set of 5 to 6 'intermediate peaks' that could be also interpreted as cellular metabolites were detected in the 11.6–17.4 min range at 290 nm but not at 370 nm (Figure 4E,F,I,J) for some cell-extract samples included the DMEM control sample after 22 h. Consequently, looking at the chromatograms at 370 nm that exclude the latter peaks highlights the possible presence of flavonoids that absorb at 370 nm.

' compounds – **Figure 4.** TC7-Cell uptake of marcela aqueous (Mac) and ethanolic (Me) extracts' compounds after 3 or 22 h of incubation. (**A**) TC7-cell methanolic extract (Section 2.3.8) for control DMEM. (**B**) Crude marcela aqueous extract (2 mg/mL) given for comparison. (**C**–**J**) TC7-cell methanolic extracts (Section 2.3.8) for Mac 1.42 mg/mL (**C**), Mac HP 1.42 mg/mL (**D**), Me 0.177 mg/mL and UV 200-295 nm spectrum for the 5.3 min peak allegedly amino-acids/protein material (**E**), Me 0.177 mg/mL (**F**), Me HP 0.177 mg/mL (**G**), Me 0.354 mg/mL (**H**), Me 0.088 mg/mL (**I**,**J**). Mac and Me concentrations in TC7-Cell apical media, incubation times, and elution wavelength are indicated. Figures can be amplified on the screen.

Chromatograms of Mac TC7-cell-methanolic extracts at 370 nm (Figure 4C,D) did not show the typical peaks corresponding to standards of gallic, chlorogenic and caffeic acids (visible at 290 nm, Figure 1C and 370 nm Figure 4B) which presented retention times of 1.1 min, 4.6 min, 5.4 min, respectively, nor the set of specific peaks visible in the crude aqueous extract in the 12-16.4 min RT range (Figure 4B).

In the opposite, the 3 peaks of Quercetin (18.0 min), Luteolin (18.3 min) and 3-*O*methylquercetin (19.4 min) detected in Me crude extract at 290 nm (not shown) and more strongly at 370 nm (Figure 1D), were revealed in Me and Me HP cell-methanolic extracts analyzed at 370 nm comparing with the flat DMEM chromatogram (Figure 4A): indeed, traces for quercetin and luteolin at 18.1–18.4 min, plus a clear emerging peak at 19.4 min attributed to 3-*O*-methylquercetin were visible on Figure 4F–H,J. The lower ratio of quercetin to 3-*O*-methylquercetin found in Me TC7-cell-methanolic extracts comparing with the crude Me (Figure 1D) suggested that quercetin was poorly absorbed into TC7-cells, or further degraded/excreted in cell supernatants. Such a result was in accordance with the absence of visible peak on chromatograms at 370 nm, in the case of cell exposure to purified aglycone quercetin for 3 h or 22 h (not shown). It has been demonstrated that methylated flavonoids are better absorbed into Caco-2 cells and present a higher resistance to microsomal oxidation than their corresponding non-methylated aglycone forms [40,41] which could simply explain the present results. As indicative of the compound uptake by TC7-cells shown in Figure 4, and as estimated on the basis of chromatogram peak areas, the retained quercetin represented 0.44 ± 0.08%, 0.24 ± 0.01% and 0.16 ± 0.03% of the quercetin present in Me deposits in the case of Figure 4F,H,J, respectively. Similarly estimated, the uptake of 3-*O*-methylquecetin represented 6.86 ± 0.72%, 1.91 ± 0.15% and 4.91 ± 0.43% of the 3-*O*-methylquecetin contained in Me deposits in the case of Figure 4F,H,J, respectively. Small peaks at 22.5 and 23.5 min were also noticed after 3 h incubation (Figure 4F,G,J) suggesting the presence of newly formed derivatives at higher RT values (Figure 1B,D). More experiments are needed to achieve a quantitative evaluation of cellular uptake and thorough identification of the retained molecules.

By comparison, Mac and Mac HP samples, although deposed onto TC7-cells at higher extract concentrations (0.71–1.42 mg/mL) than Me and Me HP samples (0.088–0.354 mg/mL), revealed no peak or non-quantifiable traces on chromatograms at 370 nm in the RT range characteristic of flavonols (Figure 4C,D). These findings were in accordance with UV-Vis spectra of marcela crude extracts and TC7-cell methanolic extracts (Figure 5A,B). Indeed, while a main band (band 1) characteristic of flavonols [6,24,42] was observed at 350–365 nm in both crude Me (Figure 5A) and Me TC7-cellmethanolic extracts (Figure 5B), such a band was not found in crude Mac and Mac TC7-cell-methanolic extracts. This agrees with the fact that Mac contains low amounts of quercetin (2 mg/g aqueous extract) compared to Me (89 mg/g ethanolic extract), as well as much lower amounts of 3-*O*-methylquercetin (Figure 1). Mac TC7-cellmethanolic extracts displayed high UV absorption at 220 nm plus a broad peak with a maximum at 260–263 nm (Figure 5B). Such UV-bands also present for control DMEM could correspond to cellular material (hydrophobic amino acids/nucleic acids) solubilized by methanol during the cell-extraction step. The higher UV-light absorption in the 220–280 nm range observed for Mac and Me TC7-cell-methanolic extracts compared to control DMEM (Figure 5B) may be an indicative of enhanced TC7-cell metabolic activity induced by marcela extracts as demonstrated by MTT-assay (Figure 3C).

**Figure 5.** (**A**) UV-Vis spectra of crude aqueous extract (Mac), crude ethanolic extract (Me) and purified aglycone quercetin from Sigma, diluted in methanol at the indicated concentrations for absorbance measurement. Absorption maxima characteristic of: (1) B-ring absorption (band I) of flavonols (glycosylated Q, 3-*O*-MQ) and flavones (Lu); (2) hydroxycinnamic acid shoulder, flavanones, all phenolic compounds; (3) shoulder for most flavonols and flavones; (4) A-ring absorption (Band II) of flavonols, flavones; (5) hydroxycinnamic acids; (6) hydroxybenzoic acids and flavanols. (**B**) UV-Vis spectra of TC7-cell methanolic extracts after 3 h incubation with control DMEM, Mac or Me mixtures deposed at the indicated concentrations in TC7-cell apical media.

It is known that quercetin has highly variable and poor bioavailability, still quercetin aglycone intestinal absorption in Caco-2 cells occurs by passive diffusion and organic anion transporting polypeptide. In contrast, glycosylated form of quercetin are deglycosylated at the small intestine prior to absorption followed by quercetin metabolization through Phase II conjugation at the small intestine involving methylation, glucuronidation, and sulfation. Moreover, quercetin glucoside has been found to possess greater bioavailability due to the presence of the glucoside moiety when compared to quercetin aglycone [43].

Small intestine cell permeability for quercetin and luteolin has already been reported [40], making marcela extracts, mainly ethanolic extracts, rich in potentially bioavailable compounds. In addition, methylated flavones that show improved transport through biological membranes and increased metabolic stability compared to unmethylated flavones could present a greater oral bioavailability [40,41].

Still, bioaccessibility studies should be assessed in order to determine the stability/bioactivity of marcela bioactive compounds after digestion conditions and whether quercetin is still present in the bioaccessible fraction to be absorbed in the small intestine. Should this be the case, it would be reasonable to encapsulate the bioactive compounds into delivery systems such as liposomes or food emulsions to protect them from the extreme conditions of the gastro-intestinal tract and assess both delivery efficiency and cytotoxicity.

The incorporation of herbal extracts into traditional foods such as yogurts, cookies and meat sausages has been previously studied [44–47]; evidence suggests that food matrix and processing conditions must also be taken into account as factors that may influence bioaccessibility of the polyphenol compounds [48].

#### **4. Conclusions**

*Achyrocline satureioides* aqueous and ethanolic extracts presented different polyphenolic composition being characterized by phenolic acids and flavonoids, respectively. The extracts presented high polyphenol content and great antioxidant capacity determined by ABTS and ORAC-FL when compared to other medicinal plants, as well as antidiabetic (α-amylase, α-glucosidase and AGEs formation inhibition capacity) and antiobesity (pancreatic lipase inhibition capacity) activities. High hydrostatic pressure applied in the experimented conditions of pressure, pressurization duration and temperature did not prove to enhance marcela bioactive compounds extraction. Moreover, high hydrostatic pressure resulted in negative effects on some marcela bioactive properties. TC7-cell studies showed different tendencies for aqueous and ethanolic extracts as determined by LDH and MTT-assays, finding no cytotoxicity for Mac extracts at the tested concentrations (0.177–1.42 mg/mL of extract in apical cell media) compared to conventional ethanolic extracts that presented increased cell membrane disruption with increasing extract concentration. However, the lowest tested Me concentrations (0.044–0.177 mg/mL of extract in apical cell media) allowed high TC7-cell metabolic activity with limited cellular membrane damage. Cellular uptake studies revealed the presence of mainly 3-*O*-methylquercetin in Me and Me HP TC7-cell-methanolic extracts analyzed by RP-UHPLC at 370 nm, demonstrating the uptake of marcela bioactive flavonoids (mainly flavonols) into intestinal cell monolayers, in the particular case of ethanolic extracts. This suggests that marcela extracts present great potential as functional food ingredients for the prevention and/or treatment of chronic diseases.

**Author Contributions:** Conceptualization, A.M.F.-F. and E.D.; methodology, A.M.F.-F., E.D., A.M.-F. and T.L.-P.; formal analysis, A.M.F.-F. (chemical and biological assays), F.L. (biological assays) and I.M. (RP-UHPLC analysis); investigation, A.M.F.-F. and E.D.; data curation, A.M.F.-F. and E.D.; writing original draft preparation, A.M.F.-F.; writing—review and editing, A.M.F.-F., E.D., H.H., P.L., T.L.-P. and A.M.-F.; supervision, E.D. and A.M.-F.; project administration, E.D., T.L.-P. and A.M.-F.; funding acquisition, E.D., T.L.-P. and A.M.-F. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Agency for Research and Innovation (Grant POS\_NAC\_2013\_1\_11655, ANII), Programa de Desarrollo de las Ciencias Básicas (PEDECIBA-

UDELAR), Comisión Sectorial de Investigación Científica (Project 2023—CSIC-i+d 2018- UDELAR), Campus France (Grant 185URYB150012), and by ECOS-Sud Committee funded project U08B01 (Procedimientos Innovadores y valorización de compuestos bioactivos destinados a la industria alimentaria, con particular atención en la industria láctea).

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

#### **References**


### *Article* **In Vitro Study of Two Edible Polygonoideae Plants: Phenolic Profile, Cytotoxicity, and Modulation of Keap1-Nrf2 Gene Expression**

**Marina Jovanovi´c 1,\* , Dina Tenji <sup>2</sup> , Biljana Nikoli´c <sup>3</sup> , Tatjana Srdi´c-Raji´c <sup>4</sup> , Emilija Svirˇcev <sup>2</sup> and Dragana Miti´c-Culafi´c ´ <sup>3</sup>**


**Abstract:** *Polygonum aviculare* and *Persicaria amphibia* (subfam. Polygonoideae) are used in traditional cuisines and folk medicine in various cultures. Previous studies indicated that phytochemicals obtained from Polygonoideae plants could sensitize chemoresistant cancer cells and enhance the efficacy of some cytostatics. Here, the cytotoxic properties of chemically characterized ethanol extracts obtained from *P. aviculare* and *P. amphibia*, individually and in combination with doxorubicin (D), were determined against hepatocarcinoma HepG2 cells. Phenolic composition, cell viability, cell cycle, apoptosis, and the expression of Keap1 and Nrf2 were examined by following methods: LC-MS/MS, LC-DAD-MS, MTT, flow cytometry, and qRT-PCR. Extracts were rich in dietary polyphenolics. Synergistic cytotoxicity was detected for extracts combined with D. The observed synergisms are linked to the interference with apoptosis, cell cycle, and expression of Keap1-Nrf2 genes involved in cytoprotection. The combined approach of extracts and D could emerge as a potential pathway of chemotherapy improvement.

**Keywords:** edible plants; Polygonoideae; phenolic profile; doxorubicin; apoptosis; cell cycle; Keap1- Nrf2 expression

#### **1. Introduction**

Widespread throughout Europe, Asia and the Americas, wild plants *Polygonum aviculare* and *Persicaria amphibia* (syn. *Polygonum amphibium*), subfamily Polygonoideae, are used in traditional cuisines and folk medicine in various cultures [1,2]. *P. aviculare*, known as the common knotweed, is edible and used as a Korean salad plant, an Australian honey plant, and a traditional Vietnam culinary herb [3–5]. In the USA, *P. amphibia,* popularly known as water smart weed, has been utilized in soft drink preparation [2]. Described as healing weeds, these plants are widely used as a home remedy to treat ailments such as stomach pains and diarrhea [1,2,6,7]. Concerning Serbia, *P. aviculare* is mainly used as an appetite stimulant [8]. Importantly, in the folk medicine of China and Austria *P. aviculare* and *P. amphibia* are employed to treat some types of cancer [9,10]. Chemical properties of these plants have been thoroughly investigated in recent decades and acquired data showed that their extracts are rich in flavonoids, sesquiterpenoids, and tannins, thus, justifying their ethnopharmacological use and contributing to their recognition in contemporary pharmacology [1,2,11,12]. Although a limited number of pharmacological studies regarding these herbs are available, some of them indicate that these plants and their active compounds could be used for the treatment of various diseases in clinical medicine, including diabetes

**Citation:** Jovanovi´c, M.; Tenji, D.; Nikoli´c, B.; Srdi´c-Raji´c, T.; Svirˇcev, E.; Miti´c-Culafi´c, D. In Vitro Study of ´ Two Edible Polygonoideae Plants: Phenolic Profile, Cytotoxicity, and Modulation of Keap1-Nrf2 Gene Expression. *Foods* **2021**, *10*, 811. https://doi.org/10.3390/foods10040811

Academic Editors: Francisca Rodrigues and Verica Dragovi´c-Uzelac

Received: 26 February 2021 Accepted: 7 April 2021 Published: 9 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 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 (https:// creativecommons.org/licenses/by/ 4.0/).

and some types of cancer [13–15]. Furthermore, epidemiological evidence has demonstrated that a diet rich in natural bioactive compounds could decrease the risk of cancer development and could be used in chemoprevention. The discovery of plant-derived drugs has emerged as a potential pathway in the search for chemotherapeutics owing to the accepted assumption that plant medicaments are safer than their synthetic counterparts. In addition, toxic and other unfavorable effects of synthetic anticancer drugs have been widely noted [2,16,17]. Chemotherapy treatment with anthracycline drugs, such as doxorubicin (D), results in high hepatotoxicity. Apart from the numerous side effects, the medical application of D is also limited due to the frequent development of resistance in tumor cells [18,19]. Considering that D could rely on an increase in the free radical production to exhibit its effect, the reduction of antioxidant defense could initially make the cancerous cells susceptible to chemotherapeutics [20,21]. Importantly, numerous cancerous cells possess increased endogenous antioxidant defense due to the constitutive overexpression of the nuclear factor erythroid 2-related factor 2 (Nrf2) related to the disruption of Kelchlike ECH-associated protein 1 (Keap1) [19,22]. Keap1 acts as a negative regulator of Nrf2, and hence, it may act as a tumor suppressor in cancer cells. Nrf2 is the redox-sensitive transcription activator that regulates the expression of a large number of cytoprotective enzymes [23]. Thereby, Nrf2 has been proposed as a novel therapeutic target to overcome chemoresistance in various types of cancer, including hepatocellular carcinoma (HCC) [22]. Moreover, it has been observed that some phytochemicals have the potential to sensitize chemoresistant HCC through the suppression of Nrf2 [22].

Therefore, in this work, the phenolic profiles and cytotoxic properties of ethanol extracts of aerial parts of *P. aviculare* (POA) and *P. amphibia* (PEA) were explored, as well as their potential to modulate the response of human hepatocellular carcinoma cells (HepG2) to D, the most widely used cytostatic in HCC treatment [16]. The cytotoxic properties of extracts, alone and combined with D, were estimated by MTT assay and flow cytometric analysis. The potential of co-treatments of extracts and D to influence Nrf2 and Keap1 expression was assessed by qRT-PCR. Thus, this study, in a comprehensive manner, investigated cytotoxic properties of POA and PEA. Using both plant extracts and cytostatic D, it aimed to present the benefits of the combined approach in order to make an initial step in chemotherapy improvement.

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

#### *2.1. Materials*

Reference standards of the secondary metabolites used in LC-MS/MS analysis were obtained from Sigma–Aldrich Chem (Steinheim, Germany): 4-hydroxy-benzoic acid, 2,5 dihydroxybenzoic acid, vanillic acid, gallic acid, cinnamic acid, caffeic acid, trans-ferulic acid, 3,4-dimethoxycinnamic acid, D-(−)-quinic acid, umbelliferon, matairesinol, secoisolariciresinol, chlorogenic acid, predominantly trans, quercetin dihydrat, (+)-catechin hydrate, baicalein, genistein, daidzein, baicalin, syringic acid, p-coumaric acid (predominantly trans isomer), 2-hydroxycinnamic acid (predominantly trans), sinapic acid (predominantly trans isomer), scopoletin, (−)-epicatechin, quercetin-3-*O*-beta-D-glucoside, quercitrin-hydrate, (−)-epigallocatechin gallate; Roth/Carl Roth GmbH/Rotichrom®: protocatechuic acid, esculetin, apigenin, apigenin-7-*O*-glucoside, hyperoside, chrysoeriol, amentoflavone trihydrate, apiin; Chromadex (Santa Ana, CA, USA): kaempherol, kaempferol 3-*O*-glucoside, naringenin, isorhamnetin; Extrasynthese Genay Cedex France: luteolin, luteolin-7-*O*glucoside, and from Fluka Chemie GmbH (Buchs, Switzerland): myricetin, vitexin, rutin, trihydrate. HPLC gradient-grade methanol was purchased from J. T. Baker (Deventer, The Netherlands) and p.a. formic acid from Merck (Darmstadt, Germany). Folin and Ciocalteu's Phenol Reagent (FC) was provided by Sigma–Aldrich, while sodium carbonate and aluminium(III) chloride were purchased from Centrohem (Stara Pazova, Serbia) and Kemika (Zagreb, Croatia), respectively. William's medium, fetal bovine serum (FBS), penicillin-streptomycin mixture, phosphate-buffered saline (PBS), trypsin from porcine pancreas, dimethyl sulfoxide (DMSO), protease inhibitor cocktails, Triton® X-100, and

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) were purchased from Sigma-Aldrich (Steinheim, Germany). Reagents for apoptosis and cell cycle assay were obtained from Invitrogen Life TechnologiesTM (FITC-AnexinV, Binding puffers 2x, Rnase A Pure LinkTM, Waltham, MA, USA) and 7- amino actinomycin D was provided from PharmingenTM (Franklin Lakes, NJ, USA). Trizol reagent, Power SYBR green PCR master mix and specific primers for qRT-PCR were obtained from Invitrogen Life TechnologiesTM (Carlsbad, CA, USA). Doxorubicin (D, Cas. No. 25316-40-9) was provided by Actavis, S.C. Sindan-Pharma S.R.L. (Buchures,ti, Romania). All the other chemicals and reagents were purchased from local companies and were of a molecular biology grade.

#### *2.2. Plant Material, Extracts Preparation, and Chemical Analysis*

Aerial parts of *P. aviculare* and *P. amphibia* were collected at Vlasina Lake (N42◦42′40.09" E22◦20′32.942") in Serbia. Plant materials were identified and the voucher specimens were deposited at the Herbarium of Department of Biology and Ecology, Faculty of Natural Sciences, University of Novi Sad, Serbia (BUNS Herbarium; voucher numbers for *P. aviculare* and *P. amphibia* are 2-1669 and 2-1691, respectively).

Extracts were prepared by the maceration of air dried and powdered aboveground plant material (10 g) with 80% ethanol (100 mL) for 72 h under constant stirring at room temperature. Extracts were removed from plant material by filtration and after vacuum drying, yield of dry raw extracts were: 1.31 g/13.1% (*P. aviculare*) and 1.32 g/13.2% (*P. amphibia*). Raw extracts were suspended in water and purified by liquid–liquid extraction with petroleum ether, to remove chlorophyll and other ballasts. Defatting of the extracts with petroleum ether can lead to losses of the compounds of interest [24,25], so the petroleum ether layer was washed with methanol and methanol fraction pooled with water layer. After purification, herb extracts yield decreased for 1.1% (1.20 g; *P*. *aviculare*) and for 0.8% (1.23 g; *P*. *amphibia*). Vacuum dried (<45 ◦C) purified extracts were dissolved in DMSO to a final concentration of 100 mg/mL.

The phytochemical profiles of *P. aviculare* and *P. amphibia* were evaluated by measuring the total phenolic content (by means of Folin–Ciocalteu reagent under alkaline conditions) and total flavonoids content (based on aluminum–flavonoids complex formation). Detailed procedure of these two spectrophotometric methods was previously published by Beara et al. [26]. Quantitative LC-MS/MS analysis of selected 45 secondary metabolites was carried out according to the previously reported method [27]. Standard mixture (containing 45 phenolics) was double diluted with mobile phase solvents: A (0.05% aqueous formic acid): B (methanol), in 1:1 ratio, to obtain fifteen working standards (from 25,000 ng/mL to 1.53 ng/mL). Extracts were diluted also with solvents A:B (1:1) to a final concentration of 2 mg/mL. Samples and standards were analyzed using Agilent Technologies 1200 Series high-performance liquid chromatograph coupled with Agilent Technologies 6410A Triple Quad tandem mass spectrometer with electrospray ion source (ESI), and controlled by Agilent Technologies MassHunter Workstation software—Data Acquisition (ver. B.03.01). The detailed procedure and method validation were published previously [27] (Supplementary material).

Qualitative LC-DAD-MS analysis of extracts was performed on Agilent Technologies 1200 Series HPLC with DAD, coupled with Agilent Technologies 6410A Triple Quad tandem mass spectrometer with electrospray ion source, and controlled by Agilent Technologies MassHunter Workstation software—Data Acquisition (ver. B.03.01). Working solution of standard mixture-45 (1.56 µg/mL) and 5 µL of extracts (20 mg/mL diluted with A:B (1:1)) were injected into the system, with Zorbax Eclipse XDB-C18 (50 mm, 4.6 mm, 1.8 µm) rapid resolution column held at 50 ◦C. Mobile phase A (0.05% aqueous formic acid) and B (methanol) was delivered at flow rate of 0.8 mL/min in a gradient mode (0 min 20% B, 6.67 min 60% B, 8.33 min 100% B, 12.5 min 100% B, re-equilibration time 4 min). Eluted components were firstly recorded on diode array detector (DAD), full spectra in 190–700 nm range, chromatograms were acquired at 254 nm, 340 nm and 430 nm; and secondly on triple quadrupole mass spectrometer, using MS2Scan run mode (both, positive

and negative ionization, m/z range of 120–1000 and fragmentor voltage of 80 V). ESI ion source parameters were as follows: nebulization gas (N2) pressure 40 psi, drying gas (N2) flow 9 L/min and temperature 350 ◦C, capillary voltage 4 kV.

#### *2.3. Human Cell Line*

The human cell line used in this study was hepatocellular carcinoma HepG2 (ATCC HB-8065, Manassas, VA, USA). HepG2 cells were grown in William's medium, with 15% fetal bovine serum, 1% penicillin/streptomycin, and 2 mM of L-glutamine. The cell line was maintained in an incubator at 37 ◦C with 5.0% CO<sup>2</sup> in a humidified atmosphere. The cells were sub-cultured at 90% confluence, twice a week, using 0.1% trypsin. Cell viability was determined by the trypan blue dye exclusion method. Cells in the logarithmic growth phase were used in all experiments.

#### *2.4. Cytotoxicity and Drug Synergism Analysis*

The cytotoxic effects of plant extracts and D, both as single compounds and in a mixture, were assessed by MTT assay, as described by Jovanovi´c et al. [28]. HepG2 cells were seeded into 96-well plates at a density 2 × 10<sup>4</sup> cells/well and incubated overnight with 5% CO<sup>2</sup> at 37 ◦C. Further on, the cells were exposed to a series of two-fold dilutions of extracts and D in the ranges 4000–125 µg/mL and 22.8–0.712 µg/mL, respectively. To prepare mixtures of extracts and D, the highest concentrations of each substance were combined and subsequently diluted two-fold. This process was repeated until reaching 125 µg/mL and 0.712 µg/mL of extracts and D, respectively. After the incubation for 24 h, the medium with test substances was replaced with MTT (final concentration 0.5 mg/mL) and incubated for additional 3 h. At the end of incubation with MTT, the medium was removed, and the formazan crystals were dissolved in DMSO. The optical density was measured at 570 nm, using a micro-plate reading spectrophotometer (Multiskan FC, Thermo Scientific, Shanghai, China). Three independent experiments were conducted.

To evaluate the nature of interaction between extracts and D, combination index (CI) analysis was used, providing quantitative definition for the additive effect (CI = 1), synergism (CI < 1), and antagonism (CI > 1) in drug combinations [29]. The CI was calculated for IC<sup>25</sup> and IC<sup>50</sup> values of the mixtures, using the formula: CI = CA/IC<sup>A</sup> + CB/ICB, where C<sup>A</sup> is the concentration of the first test substance in the binary mixture; IC<sup>A</sup> is the concentration of the first test substance alone; C<sup>B</sup> is the concentration of the second test substance in the binary mixture; and IC<sup>B</sup> is the concentration of the second test substance alone.

#### *2.5. Flow Cytometry Analysis of Apoptosis and Cell Cycle Phase Distribution*

Apoptotic cell death and analysis of the cell cycle phase distribution were analyzed using a fluorescence-activated cell sorting flow cytometer (FACS) (Calibur Becton Dickinson, Heidelberg, Germany) and Cell Quest computer software, according to manufacturer's protocol. HepG2 (1 × 10<sup>6</sup> cells/well) was cultured with plant extracts with and without D. Concentrations of tested substances were selected in accordance with the results of the MTT assay. IC<sup>50</sup> values of extracts and D, individually and combined, were tested. Apoptotic or necrotic cell death was assessed after the 24 h treatment. As described by Srdic-Rajic et al. [30], cells were harvested, washed with PBS, and stained with Annexin V FITC and 7- amino actinomycin D (7-AAD). In brief, Annexin V FITC binds to the exposed phosphatidylserine of the early apoptotic cells, whereas 7-AAD labels the late apoptotic/necrotic cells, containing damaged membrane. The numbers of viable (annexin V FITC <sup>−</sup> 7AAD−), early apoptotic (annexin V FITC<sup>+</sup> 7AAD−), and late apoptotic/necrotic (annexin V FITC<sup>+</sup> 7AAD<sup>+</sup> ) cells were determined.

The quantitative analysis of the proportion of cells in different cell cycle phases was performed after the treatment and incubation for 24 h. Cells were harvested and fixed with ice-cold 70% ethanol at −20 ◦C for 30 min. Subsequently, cells were resuspended in PBS

containing propidium iodide and RNase A and incubated for 30 min at room temperature. The distribution of the cells was measured by FACS analysis, as previously described.

#### *2.6. Real-Time Quantitative PCR (qRT-PCR) Analysis*

In order to detect the expression pattern of Keap1 and Nrf2 genes in HepG2 cells, qRT-PCR analysis was conducted as described in Kaisarevic et al. [31], with minor modifications. For the experiment, the cells were seeded into 12-well plate (10<sup>6</sup> cells/well) and, after 24 h, exposed in duplicates to the selected concentrations of combined extracts and D. The selected concentrations for this assay were the ones that induced 25% inhibitions of cell survival (IC25), considering that the test procedure requires high cell viability. After the 24 h treatment, the medium was removed, the cells were washed by PBS, and total RNA was extracted using trizol reagent according to supplier's instructions. The quality and quantity of RNA was determined spectrophotometrically by BioSpecnano (Schimadzu Corporation, Kyoto, Japan). Reverse transcription of each total RNA sample (2 µg) to cDNA was conducted using High-Capacity cDNA Reverse Transcription Kit with RNase inhibitor (Applied Biosystems). The reverse transcription reaction was conducted in the Veriti Thermal Cycler (Applied Biosystems), under the following incubation conditions: 10 min at 25 ◦C, 120 min at 37 ◦C, and 5 min at 85 ◦C. The expression level of Keap1 and Nrf2 were quantified by qPCR, which was conducted on Mastercycler® ep realplex (Eppendorf, Germany). Each PCR system contained cDNA (15 ng) and 500 nM of specific primers for the target mRNA, and the reaction was catalyzed by Power SYBR Green PCR Master Mix, according to the manufacturer's instruction. Cycling conditions were as follows: 50 ◦C for 2 min, 95 ◦C for 10 min, 40 cycles of 95 ◦C for 15 s, and 60 ◦C for 1 min. Keap1 and Nrf2 expression was detected with the amplification by 40 cycles. The following primers were used: 5′ -GACAGCCTCTGACAACACAAC-3 ′ (forward for Keap1), 5′ -GAAATCAAAGAACCTGTGGC-3′ (reverse for Keap1); 5′ - CCTCAACTATAGCGATGCTGAATCT-3′ (forward for Nrf2), 5′ -AGGAGTTGGGCATGAG TGAGTAG-3′ (reverse for Nrf2); 5′ -AGAGCTACGAGCTGCCTGAC-3′ (forward for βactin), 5′ -AGCACTGTGTTGGCGTACAG-3′ (reverse for β-actin). Data were analyzed by GraphPad Prism software with β-actin as a reference gene, and its expression was not altered by any of the treatments. The relative expression levels of each target were calculated based on the cycle threshold (Ct) method, as described by Voelker et al. [32].

#### *2.7. Statistical Analysis*

The values obtained from the following tests, MTT assay, apoptosis, cell cycle and qRT-PCR, were analyzed by analysis of variance (One-way ANOVA, Dunnett's multiple comparisons test) using GraphPad Prism 6.0 (GraphPad Software Inc. San Diego, CA, USA). The level of statistical significance was defined as *p* ≤ 0.05. To describe the type of pharmacokinetic interactions between extracts and D, the combination index (CI) was calculated, and data from MTT assay were employed. The values of CI being lower, equal, or higher than 1 (CI < 1, CI = 1, CI > 1) indicated the synergistic, additive, and antagonistic effect, respectively.

#### **3. Results**

#### *3.1. Identification of Compounds in the Extracts*

The results of spectrophotometric measurement of total phenolics and flavonoids content of the extracts were expressed as equivalents of gallic acid per g of dry extract (eq GA/g DE) and equivalents of quercetin per g of dry extract (eq Querc/g DE), respectively. They were determined to be 282.8 ± 73 mg eq GA/g DE and 306.9 ± 43 mg eq GA/1g DE, and 28.9 ± 0.5 mg eq Querc/1g DE and 38.5 ± 2.0 mg eq Querc/1g DE, for POA and PEA, respectively. The comparison of data concerning phenolics content of POA and PEA is presented in Table 1. The results of the LC-MS/MS analysis (Figure 1) showed that both extracts are rich in phenolic acids and flavonoids. POA is rich in quinic acid (8.72 mg/g DE), kaempherol-3-*O*-glucoside (1.33 mg/g DE), quercetin-3-*O*-glucoside (1.38 mg/g DE),

and quercetin-3-*O*-galactoside (3.02 mg/g DE). PEA is characterized by a high content of aglycone, such as quercetin (5.50 mg/g DE) and a high content of quercetin derivatives: quercetin-3-*O*-galactoside (11.90 mg/g DE), quercetin-3-*O*-L-rhamnoside (9.79 mg/g DE), and quercetin-3-*O*-glucoside (1.49 mg/g DE). PEA is also rich in free gallic acid (3.49 mg/g DE) and epigallocatechin gallate (1.28 mg/g DE).

**Table 1.** Concentrations of phenolics found in *Polygonum aviculare* (POA) and *Persicaria amphibia* (PEA) ethanol extracts (expressed as µg of phenolics per gram of dry extract).



**Table 1.** *Cont.*

<sup>a</sup> Numbers are used as labels on given chromatograms bellow. <sup>b</sup> From the method validation published in Orˇci´c et al. [27]. <sup>c</sup> Calculated from the instrument quantification limit (Orˇci´c et al. [27]) and sample dilution. <sup>d</sup> Results are given as the concentration (µg/g of dry extract) ± relative standard deviation of repeatability (as determined by method validation [27]). LoQ–limit of quantitation; the standard curves were provided in Supporting materials (Figure S1).

μ

č ć

č ć

**Figure 1.** MRM chromatograms of standard compounds (3.125 µg/mL each standard), ( μ **a**); of *P. aviculare* herb ethanol extract, POA, (**b**); and of *P. amphibia* herb ethanol extract, PEA, (**c**); 1: Quinic acid, 2: Gallic acid, 3: Protocatechuic acid, 4: Catechin, 5: 2,5-dihydroxybenzoic acid, 6: 5-*O*-caffeoylquinic acid, 7: Epigallocatechin gallate, 8: *p*-Hydroxybenzoic acid, 9: Esculetin, 10: Epicatechin, 11: Caffeic acid, 12: Vanillic acid, 13: Syringic acid, 14: *p*-Coumaric acid, 15: Umbelliferon, 16: Scopoletin, 17: Ferulic acid, 18: Sinapic acid, 19: Vitexin, 20: Luteolin-7-*O*-glucoside, 21: Quercetin-3-*O*-galactoside, 22: Rutin, 23: Quercetin-3-*O*-glucoside, 24: *o*-Coumaric acid, 25: Apiin, 26: Apigenin-7-*O*-glucoside, 27: Myricetin, 28: Quercetin-3-*O*-L-rhamnoside, 29: Secoisolariciresinol, 30: Kaempferol-3-*O*-glucoside, 31: 3,4-dimethoxycinnamic acid, 32: Baicalin, 33: Daidzein, 34: Matairesinol, 35: Quercetin, 36: Cinnamic acid, 37: Naringenin, 38: Luteolin, 39: Apigenin, 40: Kaempferol, 41: Baicalein, 42: Isorhamnetin, 43: Chrysoeriol, 44: Genistein, 45: Amentoflavone.

#### *3.2. Cytotoxicity and Drug Synergism Analysis of Herbal Extracts and Doxorubicin*

The evaluation of the cytotoxic effect of herbal extracts, alone and combined with D, was conducted on HepG2 cells. The IC<sup>25</sup> and IC<sup>50</sup> values of extracts and D, determined from the dose–response curves (Figure 2A–C), are presented in Table 2. Applied individually, PEA was more effective against HepG2 cells than POA. Applied in a mixture, in lower tested concentrations, POAD induced higher sensitivity of HepG2 cells than PEAD. To quantify the mode of interaction between tested substances, the combination index (CI) was calculated for IC<sup>25</sup> and IC<sup>50</sup> concentrations. Remarkable synergism for both mixtures, POAD (CI =0.62 and 0.13) and PEAD (CI = 0.89 and 0.39), was detected. Thus, the concentration required to inhibit cell viability for 25% and 50% for both agents in the mixtures has been remarkably reduced.

**Figure 2.** Inhibition rates of HepG2 cells treated with individual extracts (**A**); doxorubicin (**B**); and their combination (**C**) after 24 h.



\* The concentrations are expressed in µg/mL. *Polygonum aviculare* ethanol extract (POA); *Persicaria amphibia* ethanol extract (PEA); Doxorubicin (D); Co-treatment of POA and D (POAD); Co-treatment of PEA and D (PEAD); Combination index (CI). The concentrations in bold, individually and combined, were used in flow cytometry analysis.

#### *3.3. Effect of Herbal Extracts and Doxorubicin on Apoptosis and Cell Cycle*

To determine whether the cytotoxicity of individual agents and their combinations is related to apoptosis and mediated by cell cycle arrest, the flow cytometry was applied. A significant increase in early apoptosis was determined after treatment with POA (21%). (Figure 3A,B). In addition, both concentrations of D, used for the preparation of combinations with POA and PEA, also increased early apoptosis of cancer cells (36% and 48.65%). An increase in both early and late apoptosis was observed when the PEA (29% and 38%, respectively) and POAD (26% and 17% respectively) were applied. However, in the case of PEAD co-treatment, only an increase in late apoptosis was detected (46.51%).

The analysis of the cell cycle phase distribution of HepG2 treated cells showed a significant cell cycle arrest in G2/M phase when individual treatment, POA (33%), and both concentration of D (41% and 43.45%), as well as co-treatments POAD (46%) and PEAD (42.69%) were applied (Figure 3C,D). Additionally, a significant increase in HepG2 cells in S phase was observed after individual treatment with PEA (37.4%) and co-treatment with PEAD (31.84%).

#### *3.4. Effect of Herbal Extracts and Doxorubicin on Keap1 and Nrf2 Genes Expression*

In malignant cells, alterations of the expression of Keap1 and Nrf2 genes are not rare. Here, the expression of Nrf2 and Keap1 was examined in HepG2 cells. Both co-treatments significantly increased Keap1 and simultaneously decreased Nrf2 gene expression (Figure 4).

#### **4. Discussion**

≤ Polygonaceae species are rich sources of valuable secondary metabolites, mainly flavonoids. Data concerning chemical composition of *P. aviculare* is abundant, while *P. amphibia* has been less studied. Several studies gave important contributions in elucidating chemical composition of *P. aviculare* extracts [1,33–42], but Granica et al. [43–45], and Cai et al. [46] stands out. Granica et al. [43,44] focused on developing new standardization HPLC methods for *P. aviculare*. Taking into account the results of Granica [43], we have focused on a targeted search for given flavonol (myricetin (M), quercetin (Q), kaempferol (K), isorhamnetin (IR), kaempferide (KD)) glucuronides (U) and their acetylated derivatives (acU), as some of these tend to be the major compounds occurring in *P. aviculare* (Q-3-*O*glucuronide or kaempherol-3-*O*-glucuronide). For HPLC separation conditions used in our work there is a pattern of elucidation order, as follows: MU, QU, MacU, KU, IRU, QacU1, QacU2, KacU1, IRacU1, KacU2, IRacU2, KDU, KDacU (Figures 5 and 6) This way of analysis revealed significant differences between *P*. *aviculare* and *P*. *amphibia* plant extracts. Although both plant extracts contain myricetin-glucuronide and quercetin-glucuronide in significant amounts (not quantified), in PEA extracts the following compounds were not found: MacU, IRU, QacU1, QacU2, KacU1, IRacU1, KacU2, IRacU2, KDU, KDacU. Considering the scarce data on chemical composition of *P. amphibia* [34,35,42,47] this work, with quantitative and tentative qualitative HPLC analysis, creates a notable contribution.

≤

≤ **Figure 4.** The effect of extracts and doxorubicin (D) combined (IC25) on the expression of (**A**) Keap1 and (**B**) Nrf2 genes in HepG2 cells evaluated by the qRT-PCR, \* *p* ≤ 0.05.

μ **Figure 5.** ESI BPC chromatograms, negative ionization mode (MS2Scan) of standard mix-45 (1.56 µg/mL, each compound), (**a**); of *Poligonum aviculare* ethanol herb extract, POA, (**b**); and *Persicaria amphibia* ethanol herb extract, PEA, (**c**); with labeled phenolics that were confirmed by quantitative LC-MS/MS analysis, and tentatively determined -glucuronides (U), and acetylglucuronides (acU) derivatives of Myricetin (M), Quercetin (Q), Kaempferol (K), Isorhamenetin (IR) and Kaempferide (KD), e.g., MU-myricetin-glucuronide, MacU-myricetinacetylglucuronide (Tables S1 and S2).

**Figure 6.** HPLC-DAD chromatograms (255 nm) of *Poligonum aviculare* ethanol herb extract, POA, (**a**); and *Persicaria amphibia* ethanol herb extract, PEA, (**b**); with labeled phenolics that were confirmed by quantitative LC-MS-MS analysis. Tentatively determined -glucuronides (U), and acetylglucuronides (acU) derivatives of Myricetin (M), Quercetin (Q), Kaempferol (K), Isorhamenetin (IR) and Kaempferide (KD) are also labeled, e.g., MU-myricetin-glucuronide, MacUmyricetinacetylglucuronide (Figure S2).

Ć Ć Ć Ć Ć Ć Ć On the other hand, to overcome the problem of overall toxicity and resistance of cancer cells to chemotherapeutics, a combined approach that employs both commercial cytostatic and herbal extracts was subjected to the analysis. The benefit of this approach involves the induction of more diverse mechanisms of action that hinder the development of resistance and allow for the reduction of cytostatic doses. In this study, we examined in vitro cytotoxic properties of POA and PEA, alone and combined with commercial cytostatic D. Our investigation provided corroborative evidence that POA and PEA extract could potentiate D cytotoxicity in hepatocarcinoma (HepG2) cells. This result is in accordance with our previous findings [28], which demonstrated a synergistic interaction between *Polygonum maritimum* extract and D in HepG2 cells. Likewise, Ghazali et al. [48] demonstrated that herb extracts obtained from *Polygonum minus* have an antiproliferative effect on HepG2 cells. It has also been reported that the *Polygonum cuspidatum* extract has an antiproliferative effect on hepatocarcinoma cells Bel-7402 and Hepa 1–6 [49], whereas extracts obtained from *Polygonum glabrum* and *Polygonum orientale* exhibited the protective activity on normal hepatocytes in vivo [50,51].

č It is well known that inhibition of cancer cell proliferation may be a result of a proapoptotic effect and a cell cycle disruption. Consequently, the effect of extracts and their combinations with D on apoptosis and cell cycle arrest were monitored. Herein, it was confirmed that apoptosis induction plays an important role in D-induced cell death of HepG2 cells. Also, all treatments, but particularly with POA, applied alone and in a combination, were capable of inducing early or late apoptosis. Similar to this finding, Habibi et al. [7] have reported that methanol extract of *P. aviculare* induced apoptosis in breast cancer MCF-7 cells. As for the impact of *Polygonum* spp. extracts on the molecular mechanism of apoptosis, up-regulation of the apoptotic gene *p53* and down-regulation of the anti-apoptotic *Bcl-2* gene were demonstrated [7]. The pro-apoptotic effect controlled by p53 is accompanied by cell cycle arrest in G2/M phase [52]. Further on, the arrest of the cell cycle in G2/M phase is a well-established feature of D [53]. Importantly, the G2/M checkpoint serves to prevent damaged cells from entering mitosis and proliferate. Not only D, but also the extracts, particularly POA alone and combined with D, increased the number of HepG2 cells in the G2/M phase. Likewise, recent studies have shown that

various *Polygonum* spp. extracts and their active compounds induced the pro-apoptotic effect and arrested HepG2 cells in G2/M phase [54,55]. Moreover, this study demonstrated that PEA alone and combined with D induced cell cycle arrest in the S phase. The obtained results are in line with Ghazali et al. [48], implying that *Polygonum minus* extracts induced S phase cell cycle arrest in HepG2 cells. Therefore, the observed synergism between D and tested extracts in cancer cells might be attributed to the interference with pathways involved in the regulation of apoptosis and cell cycles.

Data reported in the literature indicates that pro-apoptotic effects and cell cycle arrest induced by plant extracts could be attributed to their chemical composition. Searching for possible active compounds among the main constituents of the tested extracts pointed to free gallic acid, as well as quercetin and its derivatives, since they have been well documented to possess cytotoxicity linked to pro-apoptotic effects and the ability to induce cell cycle arrest [56–58]. Thus, quercetin caused the cell cycle arrest in G2/M phase, which was followed by a decrease in cell numbers in the G0/G1 phase [58]. Furthermore, gallic acid induced cell cycle arrest in malignant cells, contributing to inhibition of cancerous cell proliferation [59]. Moreover, quercetin induced apoptosis in various cancer cells [60,61].

Beside the growth-inhibiting and apoptosis-inducing effects, phytochemicals are capable of modulating Nrf2 expression. Furthermore, due to the overexpression of Nrf2, malignant cells are frequently highly resistant to different chemotherapeutics [19,21]. Therefore, Nrf2 is an important pharmacological target of effective chemotherapy. This study demonstrated that both co-treatments decreased Nrf2 expression in HepG2 cells. As expected, this response was followed by increased Keap1 gene expression. Similar results were obtained for resveratrol, an active compound of *Polygonum cuspidatum*, which was shown to modulate Nrf2 expression in a concentration- and time-dependent way [62]. Comprehensively observed, Nrf2 has a functional link with numerous genes reported to play specific roles in the development of drug resistance. For instance, Nrf2 influences the regulation of phase II-detoxifying enzymes, antioxidant defense enzymes, and multidrug resistance-associated proteins 1-6 (MRP 1-6) [23,63]. Altogether, regulation of Nrf2 is responsible, at least partially, for chemotherapy resistance, indicating the importance of determining Nrf2 inhibitors, such as POAD and PEAD.

In conclusion, since synergistic cytotoxicity in hepatocarcinoma cells was observed, the combined approach that employs *Polygonum aviculare* and *Persicaria amphibia* ethanol extracts and cytostatic D, could serve as a good starting point in the search for hepatocarcinoma chemotherapy improvement.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/foods10040811/s1, Figure S1: Calibration curves of standards (34 out of 45) which presence was detected in analysed extracts, Figure S2: Extract Ion Chromatograms: -EIC: 493-,477-,461-,491-,535-, 519-,475-,503-,533-,517-,317-,301-,285-,315-,299- indicating the significant differences in flavonol-glucronides composition in *P. aviculare* and *P. amphibia* species, Table S1: Identification of the main compounds from *Polygonum aviculare* ethanol extracts (POA) by HPLC-DAD-MS, Table S2 Identification of the main compounds from *Persicaria amphibia* ethanol extracts (PEA) by HPLC-DAD-MS.

**Author Contributions:** Conceptualization, D.M.-C., B.N., and M.J.; methodology, T.S.-R., E.S., D.T., ´ and M.J.; validation, B.N.; formal analysis, T.S.-R., E.S., D.T., and M.J.; investigation, M.J., D.M.-C., ´ and B.N.; resources, D.M.-C.; data curation, E.S., T.S.-R., and D.T.; writing—original draft preparation, ´ M.J.; writing—review and editing, B.N., D.M.-C., and E.S.; visualization, M.J.; supervision, D.M.- ´ C., ´ B.N., and T.S.-R.; project administration, D.M.-C.; funding acquisition, D.M.- ´ C., and M.J.; All authors ´ have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by Ministry of Education, Science, and Technological Development of Republic of Serbia (451-03-68/2020-14/200178 and 200051).

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** The authors are grateful to Bojana Žegura, National Institute of Biology, Slovenia, for providing HepG2 cells for this work and to Goran Anaˇckov, University of Novi Sad—Faculty of Sciences, Serbia, for identifying plant material.

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

#### **References**


### *Article* **Quality and Antioxidant Properties of Cold-Pressed Oil from Blanched and Microwave-Pretreated Pomegranate Seed**

**Tafadzwa Kaseke 1,2, Umezuruike Linus Opara 2,\* and Olaniyi Amos Fawole 2,3,\***


**Abstract:** The present research studied the influence of blanching and microwave pretreatment of seeds on the quality of pomegranate seed oil (PSO) extracted by cold pressing. Pomegranate seeds (cv. Acco) were independently blanched (95 ± 2 ◦C/3 min) and microwave heated (261 W/102 s) before cold pressing. The quality of the extracted oil was evaluated with respect to oxidation indices, refractive index, yellowness index, total carotenoids content, total phenolic content, flavor compounds, fatty acid composition, and 2.2-diphenyl-1-picryl hydrazyl (DPPH) and 2.2-azino-bis (3 ethylbenzothiazoline-6-sulfonic acid) (ABTS) radical scavenging capacity. Blanching and microwave pretreatments of seeds before pressing enhanced oil yield, total phenolic content, flavor compounds, and DPPH and ABTS radical scavenging capacity. Although the levels of oxidation indices, including the peroxide value, free fatty acids, acid value, ρ-anisidine value, and total oxidation value, also increased, and the oil quality conformed to the requirements of the Codex Alimentarius Commission (CODEX STAN 19-1981) standard for cold-pressed vegetable oils. On the other hand, blanching and microwave heating of seeds decreased the pomegranate seed oil's yellowness index, whilst the refractive index was not significantly (*p* > 0.05) affected. Even though both blanching and microwave pretreatment of seeds added value to the cold-pressed PSO, the oil extracted from blanched seeds exhibited lower oxidation indices. Regarding fatty acids, microwave pretreatment of seeds before cold pressing significantly increased palmitic acid, oleic acid, and linoleic acid, whilst it decreased the level of punicic acid. On the contrary, blanching of seeds did not significantly affect the fatty acid composition of PSO, indicating that the nutritional quality of the oil was not significantly affected. Therefore, blanching of seeds is an appropriate and valuable step that could be incorporated into the mechanical processing of PSO.

**Keywords:** pomegranate seed; oil; pretreatment; cold pressing; total phenolic content; antiradical activity

#### **1. Introduction**

Pomegranates (*Punica granatum* L.) are consumed as fresh fruits and processed into products such as jam, juice, jelly, wine, and dried snacks [1]. In addition to increased production volumes, the inconvenience associated with fresh pomegranate consumption due to the fruit complexity has promoted the fruit's processing into these ready to eat and convenient products [2]. Moreover, the consumption of the fruit is related to its medicinal properties. The fruit's pharmacological value can be traced back to ancient times when the fruit was used as a traditional medicine to treat different ailments [3].

The literature has shown that every part of the fruit contains compounds with health benefits. The juice and peels contain punicalagins, hydrolyzable tannins, anthocyanins, and ellagic acid [4]. Pomegranate seeds, one of the waste products from the processing of

**Citation:** Kaseke, T.; Opara, U.L.; Fawole, O.A. Quality and Antioxidant Properties of Cold-Pressed Oil from Blanched and Microwave-Pretreated Pomegranate Seed. *Foods* **2021**, *10*, 712. https:// doi.org/10.3390/foods10040712

Academic Editors: Francisca Rodrigues and Cristina Delerue-Matos

Received: 27 February 2021 Accepted: 18 March 2021 Published: 26 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 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 (https:// creativecommons.org/licenses/by/ 4.0/).

the fruit, serve as a rich source of oil (12–20%) high in tocopherols, polyphenols, sterols, and punicic acid [5]. It has been demonstrated that these bioactive phytochemicals are implicated in pomegranate seed oil's chemopreventive activities such as anti-mutagenicity, antihypertension, antioxidative potential, and reduction in liver injury [6]. In line with this, pomegranate seed can be considered for value-added products. Further, processing the seeds into specialty oil is a profitable alternative to managing the postharvest waste from pomegranate fruit processing. Pomegranate seed oil can be extracted from the seeds using various techniques such as cold pressing and solvent, supercritical carbon dioxide, and ultrasound-assisted aqueous enzymatic extraction [7,8]. Prior research has indicated that the extraction technique is a major determinant of seed oil quality [9].

Seed oil extraction by cold pressing is the most preferred by processors and consumers because of the low production costs and high concentration of bioactive compounds such as essential fatty acids, tocopherols, phenols, carotenoids, and phytosterols in the oil [10]. The retention of antioxidant compounds may provide cold-pressed oils with acceptable oxidative stability and better health properties [11]. Cold-pressed oils are obtained mechanically using either a hydraulic or screw press without the application of heat, solvents, or chemical treatments, which makes the process environmentally friendly and the extracted oil safer for human consumption [12]. Therefore, there is a growing demand for cold-pressed oil, such as cold-pressed pomegranate seed oil. The maximum temperature of cold-pressed oil should not exceed 50 ◦C [13,14]. The cold-pressed oil may be physically purified through filtration, sedimentation, or centrifugation processes, which do not degrade the oil quality [10].

Despite the many advantages of cold pressing, the low-cost and sustainable oil extraction technique suffers from low oil yield due to a significant amount of oil that remains trapped in the pressed meal, which has hindered its development and commercial viability [15]. Nonetheless, this may be improved by blanching or microwave heating the oil-bearing seeds before pressing. According to Kaseke et al. [9], blanching seeds improved the pomegranate seed oil yield and bioactive compounds recovery with ethanol. Moreover, seed blanching is a novel technique that presents a sustainable strategy capable of improving seed oil quality whilst significantly reducing the oil extraction time and energy consumption during cold pressing [16]. Blanching significantly changes the seed matrix's structural integrity by disintegrating the cell walls and membranes, which may enhance the extractability of the intracellular material by cold pressing [17]. Nevertheless, microwave pretreatment is the commonly used technique to improve oil yield and bioactive compounds recovery in cold-pressed oils [18–21], due to its uniform energy delivery, high thermal conductivity to the interior of the material, energy saving, and precise process control [22]. Although the influence of seed pretreatment on the oil recovery efficiency of mechanical pressing has been studied, comparative studies on seed pretreatment techniques' potential to improve the quality of cold-pressed oil are limited.

In this regard, the current study aimed to investigate the effect of blanching and microwave heating pomegranate seeds on the quality and functional properties of oil extracted by cold pressing.

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

#### *2.1. Plant Material*

'Acco' pomegranates were harvested in April during the 2019 season from Blydeverwacht Farm (33◦48′0"S, 19◦53′0"E) in Western Cape Province, South Africa, at the commercial maturity stage (total soluble solids: 14.02–16.61◦Brix). The seeds were separated from the peels, membranes, and juice before they were thoroughly cleaned with tap water.

#### *2.2. Sample Preparation and Pretreatments*

#### 2.2.1. Blanching

Freshly extracted and clean pomegranate seeds (PS) were blanched in a water bath (Scientific, South Africa) at 95 ± 2 ◦C for 3 min [9]. After blanching, samples were cooled promptly in an ice water bath, drained off, and then oven dried at 55 ± 2 ◦C to 10% (*w/w*) moisture content. The thermogravimetric technique was applied to measure the moisture content using a moisture analyzer set at 100 ◦C (KERN, DBS60-3, Balingen, Germany).

#### 2.2.2. Microwave Pretreatment

Fifty grams of oven-dried PS were uniformly spread in a glass Petri dish (190 mm in diameter) inside a calibrated domestic microwave oven (Model: DMO 351, Defy Appliances, Cape Town, South Africa) with a nominal power of 900 W. The microwave oven was calibrated following the method described by Rekas et al. [23]. The samples were microwave heated at 2450 MHz and 261 W for 102 s [24]. The microwave-heated seeds were cooled at ambient temperature (25–27 ◦C) and thoroughly mixed to uniform samples. The moisture content of the seeds after microwave heating was adjusted to 10%.

#### *2.3. Cold Pressing*

PS (250 g) were pressed using a single-screw press (Farmet UNO, Ceska Skalice, Czech Republic) equipped with a 10 mm diameter die. The capacity of the expeller press is about 8– 12 kg seed/h. The press head was heated to 60 ± 5 ◦C before oil pressing using a removable heating element, and the temperature of the outflowing oil was 50 ± 5 ◦C. Temperature was measured using a type-K thermocouple connected to a digital temperature sensor (KIMO Instruments, Wilmington, NC, USA). The pressed oil was centrifuged at 4000 rpm for 15 min (Centrifuge 5810R, Eppendorf, Horsholm, Germany) to remove solid particles. Pomegranate seed oil (PSO) extraction yield was defined as gram per hundred gram of pomegranate seed (g/100 g seed). The oil samples were packed in brown bottles and stored at −20 ◦C before further analyses to minimize oxidation.

#### *2.4. Determination of Pomegranate Seed Oil Quality Indices*

#### 2.4.1. Yellowness and Refractive Index

The refractive index was evaluated at 25 ◦C using a calibrated Abbe 5 refractometer (Bellingham + Stanley, Kent, United Kingdom). The yellowness index (YI) was calculated from the PSO color properties, including lightness (L\*) and yellowness (b\*) values, which were measured using a calibrated Chroma meter CR-410 (Konica Minolta, INC, Tokyo, Japan).

$$\text{YI} = \frac{142.86 \text{b}^\*}{\text{L}^\*} \tag{1}$$

#### 2.4.2. Oxidation Indices

Free fatty acids (FFA) and acid value (AV) were measured following the AOCS standard [25]. The modified ferrous oxidation-xylenol orange (FOX) method was used to determine peroxide value (PV) [26]. ρ-Anisidine value (AnV) was determined according to [25]. Total oxidation (TOTOX) value was calculated from the PV and AnV using Equation (2).

$$\text{TOTOX} = 2PV + AnV \tag{2}$$

#### *2.5. Determination of Bioactive Compounds and Antiradical Activity*

2.5.1. Total Carotenoids Content and Total Phenolic Content

The method described by Ranjith et al. [27] was used to determine total carotenoids content (TCC). PSO (0.2 g) was dissolved in hexane (5 mL) and 0.5 mL sodium chloride (NaCl) (0.5%, *w/w*) and thoroughly vortexed before being centrifuged (Centrifuge 5810R, Eppendorf, Horsholm, Germany) at 4000 rpm for 5 min. The absorbance was measured at 460 nm using a UV spectrophotometer (Spectrum Instruments, United Scientific, Cape

Town, South Africa), and the results were expressed as mgβ-carotene/g of PSO. The Folin– Ciocalteu method was applied to evaluate the total phenolic content (TPC) [28]. Briefly, 200 µL of PSO methanol extracts, 250 µL of the Folin–Ciocalteau reagent, 750 µL of 2% (*w/v*) sodium carbonate, and 3 mL of distilled water were sequentially mixed, and the mixtures were vortexed and incubated in the dark for 40 min. The absorbances were measured at 760 nm using a UV spectrophotometer (Spectrum Instruments, United Scientific, Cape Town, South Africa), and the results were reported as milligram gallic acid equivalent per g PSO (mg GAE/g PSO).

#### 2.5.2. Antiradical Activity

The PSO antiradical activity was evaluated using 2.2-azino-bis (3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) and 2.2-diphenyl-1-picryl hydrazyl (DPPH) assays. Briefly, ABTS radical cation (ABTS<sup>+</sup> ) stock solution, prepared by mixing equal volumes of 2.2-azino-bis (3 ethylbenzothiazoline-6-sulfonic acid) (ABTS) solution (7.4 mM) and potassium persulfate solution (2.6 mM), was kept in the dark for 12–16 h. The absorbance was adjusted to 0.7 ± 0.02 at 750 nm after the incubation period, using 80% (*v/v*) methanol [29]. Three hundred microliters of methanol PSO extracts were mixed with 300 µL of the freshly prepared ABTS<sup>+</sup> solution and the samples were incubated for 10 min in the dark. The absorbances of the samples were measured at 750 nm using a microplate reader (Thermo Fisher Scientific, Shanghai, China). The results were reported as mmol Trolox/g of PSO.

The method described by Siano et al. [30] was used to determine the DPPH radical scavenging capacity of PSO. Aliquots of 200 µL of methanol PSO extracts were added to 2.5 mL of 0.04% (*w/v*) DPPH in 80% (*v/v*) methanol and vortexed before incubation in the dark for 60 min. The absorbance was measured using a UV spectrophotometer (Spectrum Instruments, United Scientific, Cape Town, South Africa) at 517 nm. Results were expressed as mmol Trolox/g of PSO.

#### *2.6. Fatty Acid Composition*

The gas chromatography-mass spectrometry (GC-MS) method was used to determine the PSO fatty acid composition [31]. PSO (0.1 g), 2.0 mL hexane, 50 µL heptadecanoic acid (1000 ppm, internal standard), and 1.0 mL of 20% (*v/v*) H2SO<sup>4</sup> in methanol were sequentially mixed, vortexed, and incubated at 80 ◦C for 1 h in an oven. To the cooled samples, 3 mL of saturated NaCl was added, and the mixture was further vortexed before centrifugation (Centrifuge 5810R, Eppendorf, Horsholm, Germany) at 4000 rpm for 3 min. The supernatant (hexane extract) was analyzed using a gas chromatograph (6890N, Agilent Technologies, Palo Alto, CA, USA) coupled to a flame ionization detector (FID). The fatty acid methyl esters were separated on a polar RT-2560 (100 m, 0.25 mm ID, 0.20 µm film thickness) (Restek, Bellefonte, PA, USA) capillary column and helium (1 mL/min) was used as the carrier gas. The sample (1µL) was injected in a 5:1 split ratio and at 240 ◦C. The oven temperature was programmed as 60 ◦C/min and increased to 120 ◦C at a rate of 8 ◦C/min, then to 245 ◦C at 1.5 ◦C/min, and finally to 250 ◦C at 20 ◦C/min for 2 min. Gas-chromatographic peaks of FAME were identified by comparison with a commercial mixture of standards, and the NIST library was used to identify the pomegranate seed oil fatty acids profiles. The relative content (%) of each fatty acid was calculated by dividing the peak area of each fatty acid by the total peak area of all the fatty acids identified.

#### *2.7. Determination of Volatile Compounds*

Volatile compounds were analyzed by HS-SPME-GC-MS [32]. One thousand microliters of oil samples was put in 20 mL SPME vials and 1 µL was injected into the SPME-GC-MS system. Separation was performed on a gas chromatograph (6890N, Agilent technologies network) coupled to an Agilent technologies inert XL EI/CI Mass Selective Detector (MSD) (5975B, Agilent Technologies Inc., Palo Alto, CA). The GC-MS system was coupled to a CTC Analytics PAL autosampler. Separation of the oil volatiles was performed on a ZBWaxPlus (30 m, 0.25 mm ID, 0.25 µm film thickness) capillary column. Helium

was used as the carrier gas at a flow rate of 1 mL/min. The injector temperature was maintained at 250 ◦C. Injection was performed in splitless mode. The oven temperature was programmed as follows: 35 ◦C for 5 min, followed by a ramping rate of 5 ◦C/min until 50 ◦C and held for 3 min, ramped again at a rate of 5 ◦C/min until 120 ◦C and held for 3 min, and finally ramped up to 240 ◦C at a rate of 10 ◦C/min for 3 min. The MSD was operated in a full scan mode, and the source and quad temperatures were maintained at 230 ◦C and 150 ◦C, respectively. The transfer line temperature was maintained at 250 ◦C. The mass spectrometer was operated under electron impact (EI) mode at an ionization energy of 70 eV, scanning from 25 to 650 m/z. Compound identification was based on mass spectral data of samples with the standard NIST and Wiley Library and with the comparison of retention indices. The relative content (%) of each volatile compound was calculated by dividing the peak area of each component by the total peak area of all the compounds identified.

#### *2.8. Statistical Analysis*

The results of all the studied variables are presented as mean ± SD (standard deviation). One-way analysis of variance (ANOVA) was carried out using Statistica software (Statistical v13, TIBC, Palo Alto, CA 94304, USA) and the mean values were separated according to Duncan's multiple range test. Graphical presentations were made using Microsoft Excel (Version: 16.0.13029.20344, Microsoft Cooperation, Washington, DC, USA).

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

#### *3.1. Oil Yield*

Oil yield is an essential factor in maximizing the gross income for seed oil processers. In this regard, pretreatment of the oleaginous material is crucial in promoting oil release from the seed matrix. The results in Figure 1 show that blanching and microwave heating of seeds enhanced the PSO yield (55 and 91%, respectively), with blanched seeds exhibiting a significantly higher oil yield (6.12%) than microwave-heated seeds (4.97%). The initial yield of PSO from unpretreated seeds was 3.20%. The finding that blanching and microwave heating of seeds significantly improved the PSO yield could be attributed to altering the pomegranate seed cellular structures, which increased the permeability of the cell walls and mass transfer of lipids during pressing [9,19]. Prior research has also reported significant enhancement of cold press oil extraction efficiency by thermal seed pretreatment [18,33,34]. For instance, seed microwave pretreatment doubled the yield of cold-pressed black cumin seed oil [35]. In a recent study, Lee et al. [36] observed a 2.3- to 2.4-fold increase in coldpressed perilla seed oil yield after steam pretreatment of the seeds. On the other hand, in the absence of seed pretreatment, which provides cell disintegration, the permeability of the pomegranate seed to oil could have been limited, hence the lower oil yield from unpretreated seeds (Figure 1) [37]. The oil yield in the current study is 1.1- to 2.2-fold lower than the one reported by Khoddami et al. [7], a fact that could be explained by differences in seed variety, moisture content, oil press equipment, fruit maturity, and geographical location, among other factors.

#### *3.2. Yellowness and Refractive Index*

The color and appearance of foods, including seed oil, constitute the first set of sensory attributes and therefore affect the consumer perception of quality. The color of food may be attributed to natural pigments or biochemical or chemical products developed during processing such as seed thermal pretreatment [38]. The effect of processing on food product color can be determined by measuring the YI. Table 1 depicts the changes in the PSO yellowness index due to seed blanching and microwave heating. Blanching and microwave heating of PS significantly decreased the oil YI by 7%. With respect to PSO from blanched seeds, the decrease in the YI could be ascribed to the reduction in total carotenoids content due to the conversion of trans-carotenoids, which are the usual configuration, to cisisomers, hence decreasing the oil yellowness (Figure 2a) [39]. According to Kha et al. [40], the extensive conjugated and trans-configured double bond system in carotenoids absorbs light in the visible region and provides foods such as seed oil with color. The pomegranate seed oil YI ranged from 37.30 to 40.21 and was lower than the one observed by Khoddami et al. [7] from cold-pressed oil (81.15–91.55) of different pomegranate cultivars.

**Figure 1.** Oil yield of cold-pressed pomegranate seed oil from unpretreated, blanched (95 ± 2 ◦C/3 min), and microwave-heated (261 W for 102 s) seeds. Columns followed by different letters are significantly different (*p* < 0.05) according to Duncan's multiple range test. Vertical bars indicate the standard deviation of the mean.

**Table 1.** Physicochemical properties of cold-pressed pomegranate seed oil (PSO) from unpretreated, blanched (95 ± 2 ◦C/3 min), and microwave-heated (261 W for 102 s) seeds.


Means ± standard deviation of analysis (*n* = 3). Different superscript letters in the same column indicate significant difference (*p* < 0.05) according to Duncan's multiple range test. RI = refractive index (25 ◦C), YI = yellowness index, FFA = free fatty acid as punicic acid (%), AV = acid value (mg KOH/g PSO), PV = peroxide value (meqO2/kg PSO, meqO2/kg = milli-equivalents of active oxygen per kg), AnV = ρ-anisidine value, TOTOX = total oxidation value.

> The refractive index is often applied to identify and characterize food materials, including seed oil. The relationships between RI and fatty acid chain length as well as the degree of unsaturation have been reported [41]. The refractive index of the oil did not significantly (*p* > 0.05) change after seed pretreatment, regardless of the significant change in fatty acid content after seed microwave pretreatment. In this sense, the interpretation of RI results in the present study should be made with caution. The RI narrowly ranged between 1.5194 and 1.5197, values typical of PSO and indicative of its high unsaturation (Table 1) [41]. These values agree with those reported by Costa et al. [42] (1.5091–1.5177) from cold-pressed PSO.

#### *3.3. Peroxide Value, Free Fatty Acids, Acid Value, ρ-Anisidine, and Total Oxidation Value*

The PV is used as the quantity measurement for peroxides, which are intermediate products of lipid oxidation. The PV test is a good way to determine the amount of primary oxidation products in freshly extracted seed oil. The PV of cold-pressed PSO from unpretreated seeds was relatively low (0.73 meqO2/kg PSO). After seed blanching and microwave heating, the PV significantly (*p* < 0.05) increased by 11 and 18%, although no significant differences were observed between the PV of oils extracted from microwave-

heated and blanched seeds (Table 1). Nevertheless, the PVs (0.73–0.86 meqO2/kg PSO) from all oil samples were far below the level (15 meqO2/kg oil) established by the World Health Organization (WHO) under the Codex Alimentarius Commission, indicating that the oils were of good quality and acceptable at the international market [43]. The lower values of peroxides in the oil samples may result from the lower extraction temperatures during oil pressing.

**Figure 2.** (**a**) Total carotenoids content (TCC) and (**b**) total phenolic content (TPC) of cold-pressed pomegranate seed oil (PSO) from unpretreated, blanched (95 ± 2 ◦C/3 min), and microwave-heated (261 W for 102 s) seeds. Columns followed by different letters are significantly different (*p* < 0.05) according to Duncan's multiple range test. Vertical bars indicate the standard deviation of the mean.

Free fatty acids and the acid value may be used to indicate lipase activity in the seed oil [7]. In this sense, higher FFA and AV in seed oil indicate a higher magnitude of hydrolytic deterioration and lower-quality oil product. As shown in Table 1, PSO from blanched and microwave-heated seeds had a relatively higher FFA and AV than PSO from unpretreated seeds. The increase in FFA and AV after seed blanching and microwave heating ranged between 7 and 54%. According to the quality requirements as recommended by the Codex Alimentarius Commission, cold-pressed oils should have a maximum of 4.0 mg KOH/g oil of AV [43]. Regardless of the increase after seed blanching and microwave heating, the AVs (1.19–1.83 mg KOH/g PSO) were within the standardized requirements (Table 1). The FFA (0.60–0.92%) in the present study were lower than those reported in previous studies. For instance, Khoddami et al. [7] reported FFA values of cold-pressed PSO ranging from 0.65 to 1.39%, which were 1.1- to 1.5-fold higher than our results.

The ρ-anisidine value measures the aldehyde and ketonic breakdown products of peroxides. These secondary products of oxidation are responsible for the development of rancidity in oils and fats. As shown in Table 1, the AnV of PSO from unpretreated, blanched, and microwave-heated seeds were 1.97, 2.71, and 3.02, respectively, which were 6

to 7 times lower than those reported by Costa et al. [42]. The result that microwave heating of seeds significantly increased the AV of PSO by 53% while blanching had an insignificant effect on AV indicates the difference in the pretreatment methods' mode of action. Despite the unavailability of an internationally recognized seed oil quality standard on AnV, there is a general agreement among researchers that for seed oil to be still acceptable, the AnV should be less than 10 [42,44].

The total oxidation value of PSO was determined using the PV and AnV values, representing the information for primary and secondary oxidation products. Therefore, the TOTOX value indicates both the oxidation history and further oxidation potential of the oils [45]. The changes in TOTOX values due to pomegranate seed pretreatment are shown in Table 1. The TOTOX value for PSO from unpretreated seeds was 3.42, which significantly increased to 4.33 and 4.74 after seed blanching and microwave heating, respectively. The results suggest that blanching and microwave heating of seeds could have promoted lipase enzyme activity and hydrolytic oxidation of the oil. The literature has reported increased activity of lipolytic enzymes on fat and oil in damaged cells [46].

#### *3.4. Total Carotenoids Content, Total Phenolic Content, and Antiradical Activity*

While the consumption of foods rich in carotenoids has been strongly linked to the reduction in incidences of diseases such as cancers, cardiovascular diseases, age-related macular degeneration, and cataracts, these thermolabile antioxidant compounds might be affected by processing [6]. According to Figure 2a, TCC significantly decreased (32%) after pomegranate seed blanching. Nevertheless, it was not significantly (*p* > 0.05) affected by seed microwave heating. The decrease in TCC after seed blanching could be explained by the breakdown of carotenoid molecules through isomerization and thermal oxidation [20]. These values were higher than TCC values reported in previous studies. For example, Costa et al. [42] reported TCC values ranging between 0.010 and 0.015 mg β-carotene/g PSO. Moreover, other studies failed to detect carotenoids in PSO [5]. The variation in the results could be due to dissimilarities in cultivars, fruit maturity, geographical location, and oil extraction process, among other factors [47]. It should also be highlighted that the absorbance in the spectrophotometric method might be increased by compounds other than carotenoids, which are active in the carotenoids' spectral range (400–500 nm) [48]. The TCC from other fruit seeds such as passion fruit and sour cherry ranged between 0.01 and 1.20 mg β-carotene/g oil [49,50]. The large disparity in the TCC of oil from different fruit seeds could reflect differences in the sensitivity of the methods of analysis, and it is, therefore, suggested that TCC calculated from the sum of individual carotenoids could be more reliable.

The total phenolic contents of PSO from unpretreated, blanched, and microwaveheated seeds are presented in Figure 2b. Blanching and microwave heating of pomegranate seeds significantly improved the TPC of cold-pressed oil by 21 and 37%, respectively. The findings suggest that blanching and microwaving of seeds facilitated the dissociation of glycosylated and esterified phenolic compounds, enhancing the amount of free phenolic compounds available for extraction [51]. The results coincide with Mazaheri et al. [20] and Lee et al. [36], who also reported improvement in TPC of cold-pressed black cumin and perilla seed oils after seed microwave and steam pretreatments, respectively. The levels of TPC from blanched and microwaved seeds did not significantly differ (*p* > 0.05). Given the potential bioactivity of phenolic compounds and the possible application of PSO as a functional food, enhancement of TPC after seed pretreatment was a desirable development. While the study of Zaouay et al. [52] reported TPC ranging from 0.03 to 0.07 mgGAE/g PSO, the TPC values in the current study varied from 1.33 to 1.83 mgGAE/g PSO (Figure 2b). Among other factors, the observed variation could be due to the selective nature of solvent extraction towards the phenolic compounds, hence the lower TPC values compared to the cold-pressed oil. On the contrary, Khoddami et al. [7] cold pressed oil from three different pomegranate cultivars and reported TPC values ranging from 8.52 to 10.44 mgGAE/g PSO that were 5.7 to 6.4 times higher than our results. These dissimilarities highlight

the importance of preharvest and processing factors consideration in PSO processing and quality.

The antiradical radical activity of PSO was determined using the DPPH and ABTS assays. The DPPH radical scavenging activity of the cold-pressed PSO from unpretreated, blanched, and microwaved seeds is given in Figure 3a. While blanching seeds significantly improved the DPPH radical scavenging activity of the oil by 37%, microwave heating did not significantly (*p* > 0.05) change the DPPH radical scavenging activity of the cold-pressed PSO. Despite the insignificant effect of seed microwave pretreatment on the DPPH radical scavenging activity of the oil, previous studies on purslane and rape seed have shown increased DPPH radical scavenging activity in cold-pressed oil after seed microwave heating [53,54]. It is nevertheless noteworthy that seed physical and cellular structures that vary among different types of seeds and cultivars play a vital role in the efficiency of seed pretreatment, cold pressing, and recovery of the antioxidant compounds. Considering that antioxidant properties of oil have a major effect on its oxidative stability behavior, the PSO from blanched seeds might exhibit better stability and improved shelf life. The ABTS radical scavenging activity of the oil samples ranged between 10.95 and 11.55 mmol Trolox/g PSO, with oil from microwaved seeds exhibiting significantly higher ABTS scavenging activity than oil from blanched and unpretreated seeds (3 and 5%, respectively). The variation in the oil samples' (microwaved and blanched seeds) DPPH and ABTS radicals scavenging suggests that the antioxidant compounds react differently with the different radicals, due to factors such as synergism [55]. The high ABTS scavenging activity (10.95–11.55 mmol Trolox/g PSO) in the present study could be attributed to the high levels of phenols in the cold-pressed oils and their synergistic effect with other antioxidant compounds such as tocopherols, which are also abundantly found in PSO [55,56].

#### *3.5. Fatty Acid Composition*

Fatty acid composition is one of the most critical quality characteristics of oilseeds, considering that the suitability of the oil for food, nutraceutical, or pharmaceutical applications may be governed by the type of fatty acids. Table 2 shows the fatty acid composition of cold-pressed PSO from unpretreated, blanched, and microwaved seeds. Chromatograms of FAMES for the treatments are presented in Supplementary Figure S1. Ten different types of fatty acids were identified in PSO, with palmitic acid, oleic acid, linoleic acid, and punicic acid being the primary fatty acids and representing 7.73–9.22%, 9.53–10.48%, 15.93– 17.11%, and 54.12–58.32% of the total composition, respectively. Other fatty acids identified but in minor quantities (0.06–4.32%) were arachidic acid, stearic acid, heneicosanoic acid, docosanoic acid, docosenoic acid, and linolenic acid. Generally, thermal pretreatment of oilseeds may alter the fatty acids composition due to the sensitivity of polyunsaturated fatty acids [33]. While microwave heating of seeds significantly decreased punicic acid by 7%, blanching did not significantly (*p* > 0.05) affect the fatty acid. In a similar study, Ozcan et al. [21] observed a 14 and 11% decrease in punicic acid after pomegranate seed roasting (150 ◦C for 10 min) and microwave heating (750 W for 7.5 min), respectively. Considering that punicic acid is implicated in most PSO biochemical properties, the decrease in punicic acid after microwave heating of the seeds in the current study was not desirable. Although punicic acid has been reported in other seeds such as bitter gourd [6], pomegranate seed remains the major source of this bioactive lipid. Compared to the literature, the levels of punicic acid (54.12–58.32%) in the current study are comparable to those reported by Costa et al. [42] (55.24–60.62%) from cold-pressed PSO. Nevertheless, some previous studies on cold-pressed PSO reported values that were higher (75.23–78.23%) than in the present study [7,57]. The dissimilarities in the punicic acid content could be ascribed to variation in processing techniques and pomegranate cultivars, among other factors. Linoleic acid and γ-linolenic acid, essential fatty acids, significantly increased by 1.1- and 3.4-fold after microwaving the pomegranate seed, whilst the blanching of the seeds did not significantly change the respective fatty acids. This indicates differences in microwaving and blanching modes of action and their impact on the fatty acids. Owing to the absence of appropriate

enzymes, the human body cannot synthesize these essential fatty acids, and therefore their maximum extraction from oilseeds is essential [58]. Oleic acid, the major monosaturated fatty acid in PSO, insignificantly varied from 9.68% to 9.53% and 10.48% after blanching and microwave heating the seeds, respectively (Table 2). Although the concentration of palmitic acid and arachidic acid, the main saturated fatty acids, increased between 1.4 and 19% and 42 and 43%, respectively, after seed blanching and microwave heating, the levels of stearic acid, heneicosanoic acid, and docosanoic acid were not significantly changed by seed pretreatment. The insignificant effect of seed thermal pretreatment on some fatty acids has also been reported in prior research [19].

**Figure 3.** (**a**) DPPH and (**b**) ABTS radical scavenging capacity of cold-pressed pomegranate seed oil (PSO) from unpretreated, blanched (95 ± 2 ◦C/3 min), and microwave-heated (261 W for 102 s) seeds. Columns followed by different letters are significantly different (*p* < 0.05) according to Duncan's multiple range test. Vertical bars indicate the standard deviation of the mean.

Regarding total saturated fatty acids (SFA), blanching and microwave heating of seeds significantly increased the SFA by 9 and 18%, respectively. A 9% decrease in total monosaturated fatty acids (MUFA) was observed in PSO from blanched seeds, and this could be due to the significant decrease (7-fold) in docosenoic acid. No significant (*p* > 0.05) variation in MUFA of PSO pressed from microwaved seeds was observed. The total polyunsaturated fatty acids (PUFA) of oil from unpretreated seeds were 74.32% (Table 2). After seed microwave pretreatment, the level significantly decreased by 6%, whilst it insignificantly decreased in PSO from blanched seeds, indicating increased heat penetration

and oxidation of polyunsaturated fatty acids during seed microwave heating (Table 1). The MUFA/PUFA index, which could be used as an indicator of the PSO stability to oxidation, among other factors [59], did not significantly vary after seed pretreatment. The finding implies that seed pretreatment did not affect the balance between the monosaturated and polyunsaturated fatty acids. However, the unsaturated fatty acids (UFA) to SFA index decreased after seed pretreatment, which could be explained by the significant increase in SFA after seed pretreatment.

**Table 2.** Fatty acid composition of cold-pressed pomegranate seed oil from unpretreated, blanched (95 ± 2 ◦C/3 min), and microwave-heated (261 W/102 s) seeds.


Values (mean ± SD, *n* = 3) in the same row and followed by different superscript letters are significantly different (*p* < 0.05) according to Duncan's multiple range test, SFA = saturated fatty acids, MUFA = monounsaturated fatty acids, PUFA = polyunsaturated fatty acids, UFA = unsaturated fatty acids, ∑ = sum of SFA, MUFA, or PUFA.

#### *3.6. Volatile Compounds*

The results of volatile compounds of cold-pressed PSO from unpretreated, blanched, and microwave-heated seeds are presented in Table 3. A typical chromatogram of volatiles from the investigated pomegranate seed oil is presented in Supplementary Figure S2. Volatile compounds that can be perceived by humans have a greater influence on PSO flavor. These are the primary volatile flavor substances and constitute the characteristic flavor of PSO. The PSO samples showed varied volatile compounds belonging to the following chemical classes: alcohols, aldehydes, ketones, esters, carboxylic acids, and hydrocarbons. The groups of volatile compounds were comparable to the findings of Costa et al. [42] and Dun et al. [60] from cold-pressed pomegranate seed and peanut oils, respectively.

Esters, which are derived from the esterification of free fatty acids and alcohols, occur naturally in many fruits and enhance their flavors. Pentyl pentanoate, the only ester observed in the oil samples, was significantly higher in oil from blanched and microwaved seeds (10- and 1.5-fold, respectively) than in unpretreated seeds, suggesting that blanching and microwaving the seeds can enhance the oil flavor [61]. Ren et al. [62] also reported the enhancement of ester compounds in rapeseed oil after microwave pretreatment of the seeds. In addition, blanching and microwave heating of seeds may induce heterocyclic compounds through the Maillard reaction, which enhances the positive flavors. Furan and its derivatives belong to heterocyclic compounds and correlate with the flavor of foods [62]. In the present study, 2.5-dimethyltetrahydrofuran was significantly higher (69%) in PSO from microwaved seeds when compared to unpretreated seeds.

**Table 3.** Volatile compounds of cold-pressed pomegranate seed oil from unpretreated, blanched (95 ± 2 ◦C/3 min), and microwave-heated (261 W/102 s) seeds.


Means ± standard deviation of analysis (*n* = 3). Different superscript letters in the same row indicate significant difference (*p* < 0.05) according to Duncan's multiple range test. ND = non-detected.

> Moreover, other furans including 2-pentylfuran and 2-butylfuran were only detected in oil from blanched and microwaved seeds. This phenomenon indicates that the flavor of PSO may be improved by blanching and microwave pretreatment of seeds. Pentanol, the primary alcohol observed, was 25 to 27% higher in PSO from blanched seeds than microwave-heated and unpretreated seeds (Table 3). Likewise, butanol and cycloheptanol manifested higher in PSO extracts from blanched than microwaved and unpretreated seeds. Other alcohol compounds observed in lower concentrations such as ethanol and octanol

were not significantly affected by seed blanching and microwave heating. Alcohols have also been reported in previous studies as important contributors to seed oil flavor [63].

Among the aldehydes, pentanal was the major compound observed in the cold-pressed PSO and was significantly higher in PSO extracts from blanched and microwave-heated seeds than unpretreated seeds. Pentanal is characterized by a nutty and fruity flavor and has been naturally found in other seed oils such as sesame, olive, and peanut [63]. Other compounds including hexanal, 3-methylbutanal, 2-heptenal, and nonanal were also significantly higher in oil extracts from blanched and microwaved pomegranate seeds (Table 2). Aldehydes in seed oil are primarily produced either through the lipoxygenase pathway during oilseed cell fragmentation or automatic oxidation of the oil during production [64]. Hexanal is a typical oxidation volatile and has been commonly used as a quality indicator for lipid oxidation in seed oils. It is characterized by green, oily, and fruity odors [60]. Its level has been positively correlated with rancid taste. As shown in Table 3, blanching and microwave pretreatment of seeds may promote the oxidative degradation of the oil. Our results, therefore, indicate higher oxidation liability of oil from pretreated pomegranate seeds compared with unpretreated seeds. The PV, AV, and AnV results found in this study support these findings (Table 1). While 2-propanone did not significantly differ in all oil samples, other ketones such as 5-butyltetrahydro-2-furanone and 5-butyl-5H-furan-2-one were only detected in oil from pretreated seeds. Saturated fatty acids including hexanoic acid, acetic acid, pentanoic acid, formic acid, and butanoic acid did not significantly (*p* > 0.05) vary among the PSO samples. These free fatty acids, which are linked to sour and pungent sensations synonymous with sensory defects, could have been produced from the oxidation of their respective aldehydes [65]. It can be stated that, although seed blanching and microwave heating may augment the positive flavor of cold-pressed PSO, they may also promote the development of undesirable flavors.

#### **4. Conclusions**

In the current study, the effect of blanching and microwave pretreatment of seeds on the quality of cold-pressed PSO was investigated. Blanching and microwave pretreatment of seeds prior to pressing improved oil yield, total phenolic content, flavor compounds, and DPPH and ABTS radical scavenging capacity. The findings are desirable to pomegranate seed oil processors and consumers along the value chain, given that cold pressing is also a greener and safer technology compared to the use of solvents such as hexane. The levels of oxidation indices including the peroxide value, free fatty acids, acid value, ρ-anisidine value, and total oxidation value also increased. Nevertheless, the oil quality conformed to the requirements of the Codex Alimentarius Commission standard (CODEX STAN 19-1981) on cold-pressed vegetable oils.

On the other hand, blanching and microwave heating of seeds decreased the pomegranate seed oil's yellowness index, whilst the refractive index was not significantly affected. Although both blanching and microwave pretreatment of seeds added value to the coldpressed PSO, the oil extracted from blanched seeds exhibited lower oxidation indices. The finding affirms that the processing technique is one of the important seed oil quality determinants. Microwave pretreatment of seeds before cold pressing significantly increased palmitic acid, oleic acid, and linoleic acid, whilst it decreased the level of punicic acid, highlighting increased heat penetration and oxidation of the conjugated fatty acid. On the contrary, blanching of seeds did not significantly affect the fatty acid composition of PSO, an indication that the nutritional quality of the oil was not significantly affected. In conclusion, blanching of seeds is a practical step that could be incorporated into mechanical production of PSO.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/foods10040712/s1, Figure S1: Chromatograms of FAMES of oil from unpretreated, blanched and microwave heated pomegranate seed, Figure S2: Typical chromatograms of volatiles from the pomegranate seed oil.

**Author Contributions:** Conceptualization, T.K., O.A.F., and U.L.O.; formal analysis, T.K.; funding acquisition, O.A.F. and U.L.O.; investigation, T.K.; methodology, T.K.; supervision, O.A.F. and U.L.O.; validation, O.A.F. and U.L.O.; visualization, T.K.; writing—original draft, T.K.; writing—review and editing, T.K., O.A.F., and U.L.O. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Research Foundation of South Africa grant number 64813 and the APC was partly funded by Stellenbosch University.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. Some of the data is contained within Supplementary Materials.

**Acknowledgments:** This work is based on the research supported wholly or in part by the National Research Foundation of South Africa (Grant Numbers: 64813). The opinions, findings, and conclusions or recommendations expressed are those of the author(s) alone, and the NRF accepts no liability whatsoever in this regard. The authors are grateful to the Cape Peninsula University of Technology, Agrifood Technology Station in South Africa for providing the oil press that was used to cold press the pomegranate seed oil.

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

#### **References**


### *Article* **Biological Effect of Different Spinach Extracts in Comparison with the Individual Components of the Phytocomplex**

**Laura Arru 1,2,\* ,† , Francesca Mussi 1,3,†, Luca Forti <sup>1</sup> and Annamaria Buschini 3,4,\***

	- 43124 Parma, Italy

**Abstract:** The Mediterranean-style diet is rich in fruit and vegetables and has a great impact on the prevention of major chronic diseases, such as cardiovascular diseases and cancer. In this work we investigated the ability of spinach extracts obtained by different extraction methods and of the single main components of the phytocomplex, alone or mixed, to modulate proliferation, antioxidant defense, and genotoxicity of HT29 human colorectal cells. Spinach extracts show dose-dependent activity, increasing the level of intracellular endogenous reactive oxygen species (ROS) when tested at higher doses. In the presence of oxidative stress, the activity is related to the oxidizing agent involved (H2O<sup>2</sup> or menadione) and by the extraction method. The single components of the phytocomplex, alone or mixed, do not alter the intracellular endogenous level of ROS but again, in the presence of an oxidative insult, the modulation of antioxidant defense depends on the oxidizing agent used. The application of the phytocomplex extracts seem to be more effective than the application of the single phytocomplex components.

**Keywords:** phytocomplex; spinach extracts; colon cancer cell line; phytochemicals; antioxidants

#### **1. Introduction**

The lifestyle of the most industrialized countries brings many benefits but can induce potential risks that can worsen the quality of life. Sedentary lifestyle, improper nutrition, unbalanced diet, chaotic pace of today's life, just to name a few, can have negative effects on human health; especially a fat-rich diet leads to oxidative stress which in turn can contribute to the onset of degenerative diseases [1].

There is increasing evidence of a close correlation between diet and risk of cancer, both in positive (prevention) and negative (development of the disease) sense [2]. The introduction of flavonoids, carotenoids, omega-3 fatty acids, vitamins, minerals, antioxidants through fruit and vegetables seems to have positive effects in reducing some types of cancer and chronic diseases, thanks to the ability of these molecules to reduce the damage caused by reactive oxygen species (ROS) [2–4].

Polyphenols, and antioxidants in general, act as free radical scavengers and metal chelators, helping the physiological cell response in counteracting the damage induced by ROS.

Intracellular ROS are normally generated during the cellular biochemical processes, and several cellular signaling pathways are regulated by ROS [5]. However, when their level happens to be increased by external agents (i.e., ionizing radiation, pollutants with chlorinated compounds or metal ions that may directly or indirectly generate ROS) they

**Citation:** Arru, L.; Mussi, F.; Forti, L.; Buschini, A. Biological Effect of Different Spinach Extracts in Comparison with the Individual Components of the Phytocomplex. *Foods* **2021**, *10*, 382. https://doi.org/ 10.3390/foods10020382

Academic Editor: Francisca Rodrigues

Received: 15 January 2021 Accepted: 4 February 2021 Published: 9 February 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 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 (https:// creativecommons.org/licenses/by/ 4.0/).

can damage proteins, lipids, and DNA, leading to impaired physiological functions with a decreased proliferative response, defective host defense and cell death [6].

Oxidative stress occurs when there is an imbalance between the intracellular levels of ROS and the cell defense systems; this can help the insurgence of diseases such as cardiovascular diseases, neurodegenerative diseases, and cancer [7]. Cells need to maintain the physiological homeostasis by balancing the ROS levels and the antioxidant defenses [6].

Not all antioxidant molecules have the same protective effect, but it is also known that the effect of a phytocomplex, a set of active ingredients contained in vegetable food, is often synergistic and greater than the effect that the single components can have. This can be in part explained because there is not a singular molecular target for a disease, but often disease is a result of a multi-factorial causality [8].

Furthermore, it should also be considered that the relative percentages of the constituents of the phytocomplex could play a decisive role in determining its effectiveness. Often, scientific attention focuses on a single active molecule, or on a few known constituents. However, the synergistic effect of the phytocomplex can be lost when testing single molecules or when, in the effort of extracting the phytocomplex, part of its minor components is lost [8–11].

Spinach (*Spinacia oleracea* L.) belongs to the family of Chenopodiaceae and it is a proven source of essential nutrients such as carotene (a precursor of vitamin A), ascorbic acid, and several types of minerals. According to the Agricultural Research Service of the U.S. Department of Agriculture, 100 g of fresh spinach provides at least 20% or more of the recommended dietary intake of β-carotene (provitamin A), lutein, folate (vitamin B9), α-tocopherol (vitamin E) and ascorbic acid (vitamin C). Moreover, spinach leaves contain flavonoids [12] and phenolic acids such as ferulic acid, ortho-coumaric and para-coumaric acids [13]. In 2009, Hait-Darshan and colleagues isolated from spinach leaves a mixture of antioxidants defined NAO (natural antioxidant) that contains aromatic polyphenols, including the phenolic acids and the derivatives of the glucuronic acid [14]. NAO can effectively counteract free radicals [15,16] resulting in an antiproliferative and anti-inflammatory potential, in vivo and in vitro [17].

In a previous study [18], we have already shown the ability of spinach leaf juice to inhibit the proliferation of the human HT29 colon cancer cell line in a time and dosedependent manner. The juice significantly also reduced the damage induced by a known oxidant agent up to 80%.

In this study, we evaluated the biological effects of different spinach extracts (hydrophilic, liquid nitrogen, and water extraction) and of some of the main components, alone or mixed together, on the human colorectal adenocarcinoma HT29, which is a cell line representative of the gastrointestinal tract and a recognized good model for the study of the correlation between diet and carcinogenesis [19]. We chose this approach as a step forward our previous research [18]: the spinach juice from fresh leaves proved to have both anti-proliferative and antioxidant effect. This time we aimed (1) to investigate if different extraction methods could positively influence this outcome; (2) to test the effect of the main components of the phytocomplex previously identified, when submitted alone or mixed in a sort of artificial simplified phytocomplex.

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

#### *2.1. Spinacia oleracea Extracts*

In this study, three extraction methods have been used to obtain a phytocomplex rich in different polyphenols and other relevant biological active molecules fractions. To obtain an extract rich in hydrophilic compounds (hydrophilic extract, HE), water was added to spinach leaves in a 1:1 (*w/v*) ratio [15]. The leaves were ground in a mortar, filtered with sterile gauze, the liquid transferred into a micro-tube, and centrifuged twice for 12 min at 15,000× *g*. The collected supernatant was concentrated (Speed Vacuum Concentrator, Eppendorf 5301, Eppendorf, AG, Hamburg, Germany) to 25%, then added a 9:1 volume of acetone in water, vortexed for 5 min and centrifuged for 12 min at 15,000× *g*. The

supernatant was carefully transferred to a fresh tube and completely dried. The dry extract was chilled on ice and stored at −20 ◦C until use. Water extract (WE) was obtained by simply adding sterile distilled water in a 1:1 ratio in the first step of the above-described procedure. Liquid nitrogen extract (NE) was obtained by grinding leaf samples in a mortar with liquid nitrogen, mashed material weighed, and transferred in a 10 mL syringe to be filtered with a 60 µm nylon filter. The extract was centrifuged for 10 min at 8000× *g* (4◦C) and the supernatant was filtered (0.2 µm filter).

#### *2.2. HT29 Cell Line*

Cells were thawed and grown in tissue culture flasks as a monolayer in DMEM (Dulbecco's Modified Eagle Medium), supplemented with 1% L-glutamine (2 mM), 1% penicillin (5000 U/mL)/streptomycin (5000 µg/mL) and 10% fetal bovine serum (FBS) at 37 ◦C in a humidified CO<sup>2</sup> (5%) incubator. The cultured cells were trypsinized with trypsin/EDTA for a maximum of 8 min and seeded with a subcultivation ratio of 1:3–1:8. Determination of cell numbers and viabilities was performed with the trypan blue exclusion test.

#### *2.3. Modulation of the Proliferation*

#### 2.3.1. MTS Assay

To determine cell viability, in the exponential phase of the growth cells were seeded at 5 × 104/mL in 96-well plates in medium supplemented with 1% glutamine, 0.5% penicillin/streptomycin, and 5% fetal bovine serum. After seeding (24 h), cells were treated, in quadruplicate, with increasing concentrations of phytochemicals (1–500 µM) and incubated for 24 and 48 h. Ascorbic acid, 20-hydroxyecdysone, ferulic acid, 2-hydroxycinnamic acid, *p*-coumaric acid, β-carotene, and lutein were from Sigma-Aldrich Company Ltd. (Milan, Italy) and resuspended in dimethylsulfoxide. The cytotoxicity assay was performed by adding a small amount of the CellTiter 96R AQueous One Solution Cell Proliferation Assay (Promega Corporation, Madison, WI, USA) directly to culture wells, incubating for 4 h and then recording the absorbance at 450 nm with a 96-well plate reader (MULTISKAN EX, Thermo Electron Corporation, Vantaa, Finland). The percentage of cell growth is calculated as:

$$\text{growth\%} = 100 - \left[1 - \text{(OD450 treated/OD450 untreated)}\right] \times 100\tag{1}$$

#### 2.3.2. Trypan Blue Exclusion Method

Different concentrations of spinach extracts (1%, 5%, 10%, 50%) were added to the cells medium. Cells were seeded in 6-well plates (2 mL/well) at the density of 2 × 10<sup>5</sup> cell/well. After 24 or 48 h of treatment, cells were trypsinized and resuspended in DMEM; a 1:1 dilution of the cell suspension was obtained using a 0.4% trypan blue solution (BioWhittake®, Lonza, Walkersville, MD, USA). The dilution was loaded on a counting chamber of a hemocytometer: since the dye freely passes only through the permeabilized membranes of dead cells, the percentage of viable cells can be evaluated. For each sample, 100 cells were scored.

#### 2.3.3. Comet Assay

Cells were seeded at 1 × 105/mL in 6-well plates in DMEM supplemented with 1% glutamine, 0.5% penicillin/streptomycin, and 10% fetal bovine serum. After seeding (24 h), HT29 cells were treated with single phytochemicals, a mixture of them at the lower concentration, and spinach extracts (1%, 5%, 10%). The phytochemical concentration was chosen according to the amount that can be found in 100 g of fresh spinach. A concentration tenfold higher was also tested to evidence possible activity variation directly related to the concentration (Table 1). To assess possible synergic or antagonist effects, the activity of a mixture of the phytochemicals at the lower dosage was also evaluated. After 24 h of incubation at 37 ◦C, the cells were trypsinized and resuspended in DMEM at a concentration of 5 × 10<sup>4</sup> cell/mL; centrifuged (1 min, 800× *g*) and the cell pellet resuspended in 90 µL Low Melting Agarose 0.7% (LMA), before being transferred onto degreased microscope

slides previously dipped in 1% normal melting agarose (NMA) for the first layer. The agarose was allowed to set for 15 min at 4 ◦C before the addition of a final layer of LMA. Cell lysis was carried out at 4 ◦C overnight in lysis buffer (2.5 M NaCl, 100 mM Na2EDTA, 8 mM Tris-HCl, 1% Triton X-100, and 10% DMSO, pH 10). The electrophoretic migration was performed in alkaline buffer (1 mM Na2EDTA, 300 mM NaOH, 0 ◦C) at pH > 13 (DNA unwinding: 20 min; electrophoresis: 20 min, 0.78 Vcm−<sup>1</sup> , 300 mA). DNA was stained with 75 µL ethidium bromide (10 µg/mL) before the examination at 400× *g* magnification under a Leica DMLS fluorescence microscope (excitation filter BP 515–560 nm, barrier filter LP 580 nm), using an automatic image analysis system (Comet Assay III Perceptive Instruments Ltd., Bury St Edmunds, UK). Total fluorescence % in tail (TI, tail intensity) provided representative data on genotoxic effects. For each sample, coded and evaluated blind, at least three independent experiments were performed, 100 cells were analyzed, and the median value of TI was calculated.


2-Hydroxycinnamic acid 2.8 2 µM 20 µM

**Table 1.** Concentration of pure phytochemicals tested in comet assay.

#### 2.3.4. Comet Assay—Antioxidant Activity

Cells were seeded and incubated with phytochemicals/extracts as described above. After incubation and trypsinization, cells were resuspended in DMEM (supplemented with 1% glutamine, 0.5% penicillin/streptomycin and 10% fetal bovine serum) at a concentration 1 × 10<sup>5</sup> cell/mL for further treatment in suspension before to perform the Comet assay, with H2O<sup>2</sup> (100 µM) on ice for 5 min. The suspensions were then centrifuged twice (1 min, 800× *g*) to wash and recover the cells. The slides were prepared and analyzed as reported above.

#### *2.4. Measurement of Reactive Oxygen Species (ROS) Production*

Cells were seeded at 1 × 105/mL in 24-well plates in DMEM supplemented as described above. After seeding (24 h), cells were treated for 24 h with ascorbic acid 3 and 30 µM, 20-hydroxyecdysone 1.5 and 15 µM, ferulic acid 1 and 10 µM, 2-hydroxycinnamic acid 2 and 20 µM, *p*-coumaric acid 0.1 and 1 µM, β-carotene 1.5 and 15 µM, lutein 2 and 20 µM, with the phytochemicals mixture (at their lowest concentration) and with spinach extracts at 1–5–10%. After 24 h of treatment, cells were washed with PBS and pre-incubated for 30 min (37 ◦C) in the dark with DCFH-DA 10 µM diluted in PBS (pH 7.4). Cells were washed with PBS to remove extracellular DCFH-DA, resuspended in DMEM, and treated 30 min at 37 ◦C with menadione 100 µM, a known oxidant agent [20]. The medium was removed and a lysis solution (Tris-HCL 50 mM, 0.5% TritonX pH 7.4; cell dissociation solution, Sigma Aldrich, St. Louis, MO, USA) was added for 10 min. Cell lysates were scraped from the dishes and the extracts were centrifuged. The supernatant was collected, and the fluorescence was read with a fluorescence spectrophotometer (Spectra Fluor Plus, Tecan Group Ltd., Männedorf, Switzerland) looking at the fluorescence peak between 510 and 550 nm. Each experiment was performed in triplicate.

#### *2.5. Statistical Analysis*

Data were analyzed by univariate analysis of variance (ANOVA) with the Bonferroni multiple comparison post-hoc test through the SPSS 18.0 software (SPSS Inc., Chicago, IL, USA). For each experiment, performed in triplicate, the significance was accepted for *p* < 0.05. multiple comparison post-hoc test through the SPSS 18.0 software (SPSS Inc., Chicago, IL, USA). For each experiment, performed in triplicate, the significance was accepted for *p* < 0.05.

Data were analyzed by univariate analysis of variance (ANOVA) with the Bonferroni

and the fluorescence was read with a fluorescence spectrophotometer (Spectra Fluor Plus, Tecan Group Ltd., Männedorf, Switzerland) looking at the fluorescence peak between 510

#### **3. Results 3. Results**

#### *3.1. Modulation of the Proliferation 3.1. Modulation of the Proliferation*

*2.5. Statistical Analysis* 

3.1.1. MTS Assay 3.1.1. MTS Assay

This colorimetric assay allows to evaluate if and to what extent the single phytochemicals tested can affect the proliferation of the HT29 cells, quantifying the number of cells in active proliferation. After 24 h and 48 h of treatment with increasing concentrations (1–500 µM) of ascorbic acid and hydroxycinnamic acids (ferulic-, *p*-coumaric- and 2-hydroxycinnamic acid) no variations have been found in the number of viable cells comparing treated and untreated samples (Figure 1). However, a significant decrease of cell proliferation has been recorded at the higher concentration tested (500 µM) after treatment with β-carotene, 20-hydroxyecdysone and lutein (Figure 1). This colorimetric assay allows to evaluate if and to what extent the single phytochemicals tested can affect the proliferation of the HT29 cells, quantifying the number of cells in active proliferation. After 24 h and 48 h of treatment with increasing concentrations (1– 500 µM) of ascorbic acid and hydroxycinnamic acids (ferulic-, *p*-coumaric- and 2-hydroxycinnamic acid) no variations have been found in the number of viable cells comparing treated and untreated samples (Figure 1). However, a significant decrease of cell proliferation has been recorded at the higher concentration tested (500 µM) after treatment with β-carotene, 20-hydroxyecdysone and lutein (Figure 1).

*Foods* **2021**, *10*, x FOR PEER REVIEW 5 of 16

and 550 nm. Each experiment was performed in triplicate.

**Figure 1.** Modulation of the proliferation of HT29 cell line: number of cells/well on concentration after 24 and 48 h of treatment with increasing phytochemical concentration (\* *p* < 0.05). **Figure 1.** Modulation of the proliferation of HT29 cell line: number of cells/well on concentration after 24 and 48 h of treatment with increasing phytochemical concentration (\* *p* < 0.05).

3.1.2. Trypan Blue Exclusion Method 3.1.2. Trypan Blue Exclusion Method

The MTS assay is not recommended for testing cell viability to spinach extract, since the presence of fibers and debris can interfere with the assay. In this case, the trypan blue exclusion method has been chosen to evaluate the modulation of cell proliferation. After 24 and 48 h of treatment with increasing concentration (1%, 5%, 10%, 50%) of The MTS assay is not recommended for testing cell viability to spinach extract, since the presence of fibers and debris can interfere with the assay. In this case, the trypan blue exclusion method has been chosen to evaluate the modulation of cell proliferation.

water (WE) and liquid nitrogen (NE) extracts, data indicate an antiproliferative activity related to highest concentrations. After 24 and 48 h of treatment with increasing concentration (1%, 5%, 10%, 50%) of water (WE) and liquid nitrogen (NE) extracts, data indicate an antiproliferative activity related to highest concentrations.

The hydrophilic extract (HE), at 48 h of treatment, shows a dose-dependent inhibition of proliferation and the highest concentration tested (50%) not only induces a reduction in cell division but also a strong cytotoxic effect. (Figure 2).

**Figure 2.** Modulation of the proliferation of HT29 cell line: number of cells/mL after 24 (dark grey) and 48 (light grey) hours of treatment with increasing concentrations of spinach extract. WE = water extract, NE = liquid nitrogen extract, HE = hydrophilic extract (\* *p* < 0.05 taking into account the growth at 0 concentration). Seeding is the number of cells at time 0.

#### *3.2. Genotoxic Activity*

A Comet assay has been carried out to evaluate if the single phytochemicals, their mixture, or the spinach extracts exert a genotoxic effect on the HT29 cell line. This assay considers the onset of possible DNA damage by evaluating the presence, after electrophoresis, of fragmented DNA outside the core of the cell nucleus. Each phytochemical was tested considering the quantity that can be found in 100 g of fresh spinach as approximate mean in standard growth conditions (considering that cultivar, production method, and growing season can all impact on the nutrient composition), and at a concentration tenfold higher (Table 1); the mixture was prepared considering the lower concentration; the extracts were tested in the concentrations of 1%, 5%, and 10%.

After 24 h of treatment with the single phytochemicals, no genotoxic effect was observed except for lutein 20 µM that showed a significant increase in tail intensity (TI, Table 2). This could explain the antiproliferative activity observed at this concentration with the MTS assay (Figure 1), related to DNA damage somehow induced by lutein. The treatment with the mixture of phytochemicals does not lead to any observed genotoxic effect as well (Table 2). Apart WE, the 24 h treatment with NE and HE led to genotoxic effect when tested at higher concentration (Table 2). This behavior could be partially responsible of the results reported in the Comet assay after oxidative injury (Figure 3).


**Table 2.** Genotoxic effects of the different phytochemicals and extracts.

<sup>1</sup> Total fluorescence % in tail. <sup>2</sup> Not treated.

(**a**)

**Figure 3.** *Cont*.

**Figure 3.** Percentage of reduction of DNA damage induced by H2O2 after a 24 h pre-treatment with different concentration (\* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001): (**a**) pure phytochemicals; (**b**) extracts. Among the antiproliferative phytochemicals, DNA damage reduction up to 40% was **Figure 3.** Percentage of reduction of DNA damage induced by H2O<sup>2</sup> after a 24 h pre-treatment with different concentration (\* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001): (**a**) pure phytochemicals; (**b**) extracts.

#### recorded with the highest dose of 20-hydroxyecdysone (1.5 and 15 µM) and, among the carotenoids, up to 50% with lutein 2 µM. β-carotene (1.5 and 15 µM) reduces DNA dam-*3.3. Antioxidant Activity*

#### age of about 15–20%. Despite its genotoxicity, also lutein 20 µM seems to induce a certain level of protection (~30%); the same level (~32%) observed after pre-treatment with the 3.3.1. Comet Assay

mixture of phytochemicals (Figure 3). Comet Assay with Spinach Extracts All the extracts have been tested at 1%, 5%, and 10%. The WE shows a significant dose-dependent reduction of DNA damage: the 5% concentration seems to be the most active, with a DNA damage reduction of about 78%. The NE significantly reduces DNA The antioxidant activity was evaluated as the ability of the samples to increase the endocellular defenses against an external oxidative stress. The Comet assay was carried out after 24 h of treatment to measure the protection against oxidative DNA damage induced by H2O<sup>2</sup> (100 µM).

damage up to 60%, without differences of activity among the concentrations tested. HE

#### does not show the ability to counteract the damage induced by hydrogen peroxide: on the contrary, at the highest dosage, it shows a pro-oxidant activity (Figure 3). Comet Assay with Phytochemicals

3.3.2. Measurement of Variation in Reactive Oxygen Species (ROS) Concentration The samples' ability to counteract the increase of ROS induced by an oxidizing agent has been investigated. For this purpose, menadione (vitamin K3) was used, a synthetic derivative of the natural vitamins K1 and K2 with a degree of toxicity against a wide variety of cancer cells. It can act directly, through the formation of reactive oxygen species, The chemicals that showed non-antiproliferative activity also exhibit a significant DNA damage reduction after H2O<sup>2</sup> oxidative stress: up to 60% for ascorbic acid (3 and 30 µM) and 75% for ferulic acid (1 and 10 µM) without variations between concentrations tested. 2-hydroxycinnamic acid gives a DNA damage reduction of about 25% (2 and 20 µM). Only *p*-coumaric acid does not show any antioxidant activity; this might be related to the very low concentrations tested (0.1 and 1 µM) or to the fact that the molecule, acting as a scavenger, may have another main target (i.e., superoxide anion) (Figure 3).

Among the antiproliferative phytochemicals, DNA damage reduction up to 40% was recorded with the highest dose of 20-hydroxyecdysone (1.5 and 15 µM) and, among the carotenoids, up to 50% with lutein 2 µM. β-carotene (1.5 and 15 µM) reduces DNA damage of about 15–20%. Despite its genotoxicity, also lutein 20 µM seems to induce a certain level of protection (~30%); the same level (~32%) observed after pre-treatment with the mixture of phytochemicals (Figure 3).

#### Comet Assay with Spinach Extracts

All the extracts have been tested at 1%, 5%, and 10%. The WE shows a significant dose-dependent reduction of DNA damage: the 5% concentration seems to be the most active, with a DNA damage reduction of about 78%. The NE significantly reduces DNA damage up to 60%, without differences of activity among the concentrations tested. HE does not show the ability to counteract the damage induced by hydrogen peroxide: on the contrary, at the highest dosage, it shows a pro-oxidant activity (Figure 3).

#### 3.3.2. Measurement of Variation in Reactive Oxygen Species (ROS) Concentration

The samples' ability to counteract the increase of ROS induced by an oxidizing agent has been investigated. For this purpose, menadione (vitamin K3) was used, a synthetic derivative of the natural vitamins K1 and K2 with a degree of toxicity against a wide variety of cancer cells. It can act directly, through the formation of reactive oxygen species, or indirectly through the depletion of the most important endogenous antioxidant, glutathione (GSH). The level of ROS was measured by a fluorescence assay with 2′ ,7′ -dichlorofluorescein-diacetate (DCFH-DA), a non-fluorescent compound that crosses cell membranes. Once in the cytoplasm, esterases remove the acetates to produce 2 ′ ,7′ -dichlorofluorescein (DCFH), which is not cell permeable anymore. DCFH is easily oxidized to 2′ ,7′ -dichlorofluorescein (DCF), a highly fluorescent compound. *Foods* **2021**, *10*, x FOR PEER REVIEW 10 of 16 or indirectly through the depletion of the most important endogenous antioxidant, glutathione (GSH). The level of ROS was measured by a fluorescence assay with 2′,7′-dichlorofluorescein-diacetate (DCFH-DA), a non-fluorescent compound that crosses cell membranes. Once in the cytoplasm, esterases remove the acetates to produce 2′,7′-dichlorofluorescein (DCFH), which is not cell permeable anymore. DCFH is easily oxidized to 2′,7′ dichlorofluorescein (DCF), a highly fluorescent compound.

> Only ascorbic acid, ferulic acid 10 µM, β-carotene at the highest dose and lutein can significantly inhibit the production of ROS, while 15 µM of 20-hydroxyecdysone weakly counteracts the increase of ROS levels induced by menadione (Figure 4). Interestingly, the synthetic mixture of the major components of the spinach complex presents no antioxidant activity, showing the presence of antagonistic effects (Figure 5). Among the extracts, only the lowest concentration of the liquid nitrogen extract (1%) and the highest concentration of the hydrophilic extract (10%) show the ability to significantly counteract the oxidative stress induced by this oxidizing agent (Figure 5). Only ascorbic acid, ferulic acid 10 µM, β-carotene at the highest dose and lutein can significantly inhibit the production of ROS, while 15 µM of 20-hydroxyecdysone weakly counteracts the increase of ROS levels induced by menadione (Figure 4). Interestingly, the synthetic mixture of the major components of the spinach complex presents no antioxidant activity, showing the presence of antagonistic effects (Figure 5). Among the extracts, only the lowest concentration of the liquid nitrogen extract (1%) and the highest concentration of the hydrophilic extract (10%) show the ability to significantly counteract the oxidative stress induced by this oxidizing agent (Figure 5).

**Figure 4.** Evaluation of antioxidant ability of the phytochemicals to counteract the intracellular ROS levels caused by 24 h treatment with menadione oxidative insult (100 µM) on HT29 cell line, expressed as ROS increment % (\* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001). **Figure 4.** Evaluation of antioxidant ability of the phytochemicals to counteract the intracellular ROS levels caused by 24 h treatment with menadione oxidative insult (100 µM) on HT29 cell line, expressed as ROS increment % (\* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001).

**Figure 5.** Evaluation of different spinach extract (left) and of single component mix (right) antioxidant ability to counteract menadione (100 µM) oxidative insult after 24 h of treatment expressed as ROS increment % (\*\*\* *p* < 0.001).

#### **4. Discussion**

*Spinacia oleracea* has a good antioxidant activity related to the presence of a pool of known phytochemicals such as ascorbic acid, carotenoids (β-carotene and lutein), hydroxycinnamic acids (ferulic acid, *p*-coumaric acid, 2-hydroxycinnamic acid), and 20 hydroxyecdysone [18]. The mechanism of action of these phytochemicals is not yet fully understood, even if it seems to be related both to their concentrations and the different types of cell lines involved. In this work, the ability has been evaluated of different extracts to modulate the proliferation of the human adenocarcinoma cell line (HT29). It has also tested the effect of the main phytochemicals belonging to the *Spinacia oleracea* phytocomplex as previously identified [18], submitted alone or merged together, in a molarity calculated as an approximate mean of concentration present in a standard condition of 100 g of fresh spinach (Table 1), considering that cultivar, production method, and growing season can all impact on nutrient composition. The activity of the phytochemicals seems to be related to the concentration applied, and on how the oxidative stress has been induced on the cells. In addition, in the case of the spinach extracts, the biological activity also reflects the method of extraction used, suggesting a greater level of complexity.

Considering the single antioxidants tested, hydroxycinnamic acids seem not to have antiproliferative activity on this cell line (Figure 1), independently from the concentration (1–500 µM) or the length time of the treatment (24 h or 48 h), in agreement also with Martini et al. [21]. The same considerations can be made for ascorbic acid, which does not induce any change in HT29 cells viability in the tested concentrations range (Figure 1), according to what observed by Fernandes et al. in 2017 on bone cancer cells cell line [22].

On the other hand, treating the cells with β-carotene (500 µM) leads to a severe inhibition of the proliferation (Figure 1). Similar results were obtained by Park et al. [23] on gastric cancer cells. Upadhyaya [24] reported a reduction of proliferation after 12 h of treatment with β-carotene starting from a concentration of 20 µM on the leukemic cell lines U937 and HL60. Compared to them, the HT29 cell line seems to be less sensitive to the effect of β-carotene, suggesting different effects of the same molecule on various cell lines. Different effects can also be found observing the response to lutein: according to the results presented in this paper, the HT29 cell line seems to be less sensitive (Figure 1) than the human breast cancer cell lines MCF7 and MDA-MB-157, as it undergoes a significant concentration-dependent reduction of viability after 24 h of treatment with lutein 5–120 µM [25].

On the other hand, all the spinach extracts demonstrate an antiproliferative activity on the HT29 cell line regardless of the type of extraction (Figure 2). While the HE shows dosedependent inhibition of the proliferation, WE and NE show both time and dose-dependent effect, with a cytotoxic activity only at the highest concentrations tested. These results

agree with what was observed by [26] on the colon cancer cell line Caco2 treated with *Amaranthus gangeticus* L. (red spinach) aqueous extracts and by Fornaciari et al. [18] on the HT29 cell line treated with *Spinacia oleracea* extracts.

Besides the antiproliferative activity, also the possible genotoxic effect of the single phytochemicals, of their mixture, and of the different spinach extracts have been evaluated by Comet assay, a useful approach to study the effect of nutrients and micronutrients. The data obtained with the hydroxycinnamic acids integrate what was already been reported by Ferguson et al. [27], namely the absence of genotoxicity at concentrations (0.5 and 1 mM) and time of treatment (7 days) higher than those assessed in the present article. Regarding the dose-dependent genotoxicity of lutein, similar behavior has been reported by Kalariya and colleagues [28]: after 9 h of treatment with lutein metabolites starting from the concentration 10 µM, a significant increment in tail intensity was observed on the retinal pigmented epithelium (ARPE-19). The presence of a genotoxic effect at 20 µM may suggest the involvement of a direct action of lutein on DNA, resulting in the reduction of the proliferation previously observed at concentrations higher than 10 µM.

The protective effect of foods rich in antioxidants carried out as decreased sensitivity to the damage induced by known oxidizing agents has been widely demonstrated [2,29]. In this work, we compared the antioxidant activity of extracts and chemicals using a modified protocol of the Comet assay that allows to verify the ability to reduce the extent of the DNA damage induced by hydrogen peroxide.

The phytochemicals mixture shows a significant DNA damage reduction, by 32% (Figure 3). However, it seems that there is not an additive antioxidant effect when the single phytochemicals are mixed, since the single molecules reached higher percentages of DNA damage reduction. Comparing the mixture to the natural phytocomplex, WE and NE show a significantly higher DNA damage reduction up to 75% (Figure 3). A similar result was observed by Ko and colleagues in 2014 on HepG2 cells and human leukocytes treated with *Spinacia oleracea* water extracts [1]. The activity of NE seems to be independent by its concentration, with a persistent reduction near 60%. WE shows a dose-dependent activity at the lowest doses with a DNA damage reduction up to 75%, while it loses effectiveness at the highest concentration. HE does not show activity at the lowest dose, while it proves a prooxidant effect at the highest ones. These observations suggest that the method of extraction strongly influences the molecular content of the phytocomplex and therefore the biological activity of the resulting extract. The WE activity recalls what has been observed for ascorbic acid (and other antioxidant molecules), high concentrations minimize the antioxidant effect or induce a pro-oxidant effect [30]. The antioxidant activity was further investigated by measuring the ability of the samples to modulate the HT29 physiological levels of ROS. Among the phytochemicals, only the highest concentration of ferulic acid, lutein, and 20-hydroxyecdysone significantly increases the intracellular ROS levels disturbing the cell oxidative balance (data not shown). A similar pro-oxidant effect was reported for ferulic acid in two cervical cancer cell lines (HeLa and ME-180) by Karthikeyan et al. in 2011 [31]. The induced imbalance does not seem to alter cell proliferation except for the treatment with lutein that, at the higher concentration, can increase the intracellular ROS levels, to inhibit cell proliferation and to induce significant DNA damage. The phytochemicals mixture does not induce ROS level variations, according to the behavior of the single molecules (Figure 5). The spinach extracts show an alteration of the intracellular ROS levels only at the higher tested concentration (10%), with a significant increase of ROS (Figure 5). Only NE shows a similar pro-oxidant effect already at a lower assayed concentration (5%).

Thereafter, also, the ability of the samples to counteract the increase of ROS induced by menadione has been evaluated. Among the single phytochemicals, β-carotene 15 µM, lutein 2 µM, and ascorbic acid 3 and 30 µM (which do not induce significant variations of intracellular ROS levels in physiological conditions, data not shown) significantly counteract the activity of menadione (Figure 4). The 10 µM ferulic acid and the 20 µM lutein induce an increase of ROS levels in the absence of stress and strongly reduce them in the presence of stress (Figure 4). The mixture of phytochemicals does not induce any variation

of ROS levels (Figure 5); it seems that the simultaneous presence of all the phytochemicals can interfere with their activity. The different extracts action, able to exhibit pro-oxidant effect both in the absence and in the presence of oxidative stress and to significantly counteract the increase due to menadione, points out again conflicting results depending on the method of detection and on the oxidant agent used.

The *p*-coumaric acid alone does not reduce the oxidative damage induced either by hydrogen peroxide or by menadione. The *p*-coumaric acid normally acts as a scavenger of superoxide anions; this could explain its inability to counteract hydrogen peroxide. Moreover, since the *p*-coumaric acid is unable to counteract the effect of menadione, which can act reducing the levels of glutathione, it can be supposed that the antioxidant activity of the *p*-coumaric acid is closely associated with that of glutathione.

The 2-hydroxycinnamic acid does not alter the intracellular levels of ROS in physiological conditions, while, in the case of oxidative stress, its activity seems to be related to the type of oxidant agent involved: a pro-oxidant effect was observed in presence of menadione, suggesting an activity comparable to that of the *p*-coumaric acid; an antioxidant effect was observed in the presence of hydrogen peroxide, suggesting a scavenger activity towards a wider range of ROS.

In physiological conditions, the ferulic acid shows a dose-dependent activity, with a pro-oxidant effect at the highest dose. In case of stress, the ferulic acid has proved a dose-dependent antioxidant activity towards menadione and a greater dose-independent one towards the hydrogen peroxide. These observations suggest that the ferulic acid could act only directly in the presence of menadione, since it may interfere with the antioxidant endogenous systems. Moreover, the increased antioxidant activity showed towards the hydrogen peroxide, suggests that the ferulic acid can counteract this oxidative damage through a combination of a direct and indirect antioxidant activity.

In the absence of oxidative stress, 20 HE shows a dose-dependent activity, with a pro-oxidant effect at the highest concentration. In presence of stress, it shows a good antioxidant activity towards hydrogen peroxide and a dose-dependent activity towards menadione, counteracting it only at the highest dose. The β-carotene does not alter the intracellular levels of ROS in physiological conditions, while in the presence of stress it has shown a dose-independent antioxidant activity towards hydrogen peroxide and a dose-dependent activity towards menadione. Lutein counteracts both oxidant agents, even though in absence of stress it showed a pro-oxidant effect. The ascorbic acid does not induce alterations in the intracellular ROS levels in physiological conditions and it has shown significant antioxidant activity with both hydrogen peroxide and menadione.

In physiological conditions, the spinach extracts show a dose-dependent activity, increasing the intracellular ROS levels only at the highest doses. In the presence of oxidative stress, their activity is strongly related to the oxidant agent involved and to the method of extraction used. It is interesting to compare the biological activity of the single phytochemicals at the concentration that we can find in 100 g of fresh spinach, the "artificial" mixture of them, and the different spinach extracts at the lowest dose (1%), which corresponds to the available concentration coming from 100 g of fresh spinach. In physiological conditions, neither the single phytochemicals, nor the mixture, nor the spinach extracts induce alterations of the oxidative balance. In the presence of oxidative stress, a different behavior has been observed that is related to the kind of oxidative agent used. Almost all the single phytochemicals can reduce the oxidative damage induced by hydrogen peroxide, but when they are mixed there is not an additive antioxidant effect, since the single molecules reached higher percentages of DNA damage reduction. Considering the activity of the spinach extracts, they showed different activity towards the hydrogen peroxide depending on the method of extraction used and, consequently, on the bioactive molecules extracted. While the WE provides the same DNA damage reduction of the "artificial" mixture of phytochemicals, the NE shows improved effectiveness suggesting a synergic effect of the mixture of bioactive constituents contained, as often reported in the literature [8,9,32].

**Phytochemical** 

*Foods* **2021**, *10*, x FOR PEER REVIEW 14 of 16

teract the ROS levels increase, but when mixed with the other molecules they lose their ability. The spinach extracts also in this case show a different behavior depending on the

Against the oxidative damage induced by menadione, there is a more variable response. Among the single phytochemicals, only lutein and ascorbic acid are able to counteract the ROS levels increase, but when mixed with the other molecules they lose their ability. The spinach extracts also in this case show a different behavior depending on the method of extraction used: the HE is unable to counteract menadione, the WE shows a pro-oxidant effect and only NE reduces the intracellular ROS levels. Once again, it can be assumed that the biological effects of a plant extract are strongly related to a synergic effect of the bioactive molecules contained. A schematic overview of the activities described above is summarized in the following tables (Tables 3 and 4): method of extraction used: the HE is unable to counteract menadione, the WE shows a pro-oxidant effect and only NE reduces the intracellular ROS levels. Once again, it can be assumed that the biological effects of a plant extract are strongly related to a synergic effect of the bioactive molecules contained. A schematic overview of the activities described above is summarized in the following tables (Table 3 and Table 4): **Table 3.** Schematic overview of the antioxidant activity of the single phytochemicals and of their mixture (⇿ = no activity; ↑= increase of oxidative stress; ↓= reduction of oxidative stress). **(Physiological Dose) Antioxidant Activity**  Against the oxidative damage induced by menadione, there is a more variable response. Among the single phytochemicals, only lutein and ascorbic acid are able to counteract the ROS levels increase, but when mixed with the other molecules they lose their ability. The spinach extracts also in this case show a different behavior depending on the method of extraction used: the HE is unable to counteract menadione, the WE shows a pro-oxidant effect and only NE reduces the intracellular ROS levels. Once again, it can be assumed that the biological effects of a plant extract are strongly related to a synergic effect of the bioactive molecules contained. A schematic overview of the activities described above is summarized in the following tables (Table 3 and Table 4): Against the oxidative damage induced by menadione, there is a more variable response. Among the single phytochemicals, only lutein and ascorbic acid are able to counteract the ROS levels increase, but when mixed with the other molecules they lose their ability. The spinach extracts also in this case show a different behavior depending on the method of extraction used: the HE is unable to counteract menadione, the WE shows a pro-oxidant effect and only NE reduces the intracellular ROS levels. Once again, it can be assumed that the biological effects of a plant extract are strongly related to a synergic effect of the bioactive molecules contained. A schematic overview of the activities described above is summarized in the following tables (Table 3 and Table 4): Against the oxidative damage induced by menadione, there is a more variable response. Among the single phytochemicals, only lutein and ascorbic acid are able to counteract the ROS levels increase, but when mixed with the other molecules they lose their ability. The spinach extracts also in this case show a different behavior depending on the method of extraction used: the HE is unable to counteract menadione, the WE shows a pro-oxidant effect and only NE reduces the intracellular ROS levels. Once again, it can be assumed that the biological effects of a plant extract are strongly related to a synergic effect of the bioactive molecules contained. A schematic overview of the activities described *Foods* **2021**, *10*, x FOR PEER REVIEW 14 of 16 Against the oxidative damage induced by menadione, there is a more variable response. Among the single phytochemicals, only lutein and ascorbic acid are able to counteract the ROS levels increase, but when mixed with the other molecules they lose their ability. The spinach extracts also in this case show a different behavior depending on the method of extraction used: the HE is unable to counteract menadione, the WE shows a pro-oxidant effect and only NE reduces the intracellular ROS levels. Once again, it can be assumed that the biological effects of a plant extract are strongly related to a synergic ef-*Foods* **2021**, *10*, x FOR PEER REVIEW 14 of 16 Against the oxidative damage induced by menadione, there is a more variable response. Among the single phytochemicals, only lutein and ascorbic acid are able to counteract the ROS levels increase, but when mixed with the other molecules they lose their ability. The spinach extracts also in this case show a different behavior depending on the method of extraction used: the HE is unable to counteract menadione, the WE shows a pro-oxidant effect and only NE reduces the intracellular ROS levels. Once again, it can be *Foods* **2021**, *10*, x FOR PEER REVIEW 14 of 16 Against the oxidative damage induced by menadione, there is a more variable response. Among the single phytochemicals, only lutein and ascorbic acid are able to counteract the ROS levels increase, but when mixed with the other molecules they lose their ability. The spinach extracts also in this case show a different behavior depending on the method of extraction used: the HE is unable to counteract menadione, the WE shows a *Foods* **2021**, *10*, x FOR PEER REVIEW 14 of 16 Against the oxidative damage induced by menadione, there is a more variable response. Among the single phytochemicals, only lutein and ascorbic acid are able to counteract the ROS levels increase, but when mixed with the other molecules they lose their ability. The spinach extracts also in this case show a different behavior depending on the *Foods* **2021**, *10*, x FOR PEER REVIEW 14 of 16 Against the oxidative damage induced by menadione, there is a more variable response. Among the single phytochemicals, only lutein and ascorbic acid are able to counteract the ROS levels increase, but when mixed with the other molecules they lose their *Foods* **2021**, *10*, x FOR PEER REVIEW 14 of 16 Against the oxidative damage induced by menadione, there is a more variable response. Among the single phytochemicals, only lutein and ascorbic acid are able to counteract the ROS levels increase, but when mixed with the other molecules they lose their *Foods* **2021**, *10*, x FOR PEER REVIEW 14 of 16 Against the oxidative damage induced by menadione, there is a more variable response. Among the single phytochemicals, only lutein and ascorbic acid are able to coun-*Foods* **2021**, *10*, x FOR PEER REVIEW 14 of 16 Against the oxidative damage induced by menadione, there is a more variable response. Among the single phytochemicals, only lutein and ascorbic acid are able to coun-*Foods* **2021**, *10*, x FOR PEER REVIEW 14 of 16

*Foods* **2021**, *10*, x FOR PEER REVIEW 14 of 16

ability. The spinach extracts also in this case show a different behavior depending on the

teract the ROS levels increase, but when mixed with the other molecules they lose their

*Foods* **2021**, *10*, x FOR PEER REVIEW 14 of 16

*Foods* **2021**, *10*, x FOR PEER REVIEW 14 of 16

**Table 3.** Schematic overview of the antioxidant activity of the single phytochemicals and of their mixture (  **No Stress Menadione H2O<sup>2</sup>**  *p*-Coumaric acid ⇿ ↑ ⇿ = no activity; ↑ = increase of oxidative stress; ↓ = reduction of oxidative stress). **Table 3.** Schematic overview of the antioxidant activity of the single phytochemicals and of their **Table 3.** Schematic overview of the antioxidant activity of the single phytochemicals and of their above is summarized in the following tables (Table 3 and Table 4): fect of the bioactive molecules contained. A schematic overview of the activities described above is summarized in the following tables (Table 3 and Table 4): assumed that the biological effects of a plant extract are strongly related to a synergic effect of the bioactive molecules contained. A schematic overview of the activities described pro-oxidant effect and only NE reduces the intracellular ROS levels. Once again, it can be assumed that the biological effects of a plant extract are strongly related to a synergic efmethod of extraction used: the HE is unable to counteract menadione, the WE shows a pro-oxidant effect and only NE reduces the intracellular ROS levels. Once again, it can be assumed that the biological effects of a plant extract are strongly related to a synergic efmethod of extraction used: the HE is unable to counteract menadione, the WE shows a pro-oxidant effect and only NE reduces the intracellular ROS levels. Once again, it can be method of extraction used: the HE is unable to counteract menadione, the WE shows a pro-oxidant effect and only NE reduces the intracellular ROS levels. Once again, it can be ability. The spinach extracts also in this case show a different behavior depending on the method of extraction used: the HE is unable to counteract menadione, the WE shows a ability. The spinach extracts also in this case show a different behavior depending on the method of extraction used: the HE is unable to counteract menadione, the WE shows a Against the oxidative damage induced by menadione, there is a more variable re-

ability. The spinach extracts also in this case show a different behavior depending on the

teract the ROS levels increase, but when mixed with the other molecules they lose their


other compounds.

other compounds.

possible anticancer effects.

possible anticancer effects.

**5. Conclusions** 

**5. Conclusions** 

In this work, the biological activity of several phytochemicals applied to a human cell line at different concentrations was evaluated. The results lead to a confirmation of their beneficial properties, acting on ROS and lowering the oxidative damage. This highlights, among other possibilities, also that of considering the development of supplements including compounds derived from spinach extract, in order to defend health and to trigger

possible anticancer effects.

tiproliferative concentration, in this case, is lower than the biologically active one of the

proliferation. This latter activity is observed just for 20-hydroxyecdysone and β-carotene but at high concentrations (100 µM). Similar behavior is observed with lutein but the antiproliferative concentration, in this case, is lower than the biologically active one of the

pounds at low concentration an antioxidant activity can be noted but no inhibition of the proliferation. This latter activity is observed just for 20-hydroxyecdysone and β-carotene but at high concentrations (100 µM). Similar behavior is observed with lutein but the antiproliferative concentration, in this case, is lower than the biologically active one of the

but at high concentrations (100 µM). Similar behavior is observed with lutein but the antiproliferative concentration, in this case, is lower than the biologically active one of the

In this work, the biological activity of several phytochemicals applied to a human cell line at different concentrations was evaluated. The results lead to a confirmation of their beneficial properties, acting on ROS and lowering the oxidative damage. This highlights, among other possibilities, also that of considering the development of supplements including compounds derived from spinach extract, in order to defend health and to trigger

In this work, the biological activity of several phytochemicals applied to a human cell line at different concentrations was evaluated. The results lead to a confirmation of their beneficial properties, acting on ROS and lowering the oxidative damage. This highlights, among other possibilities, also that of considering the development of supplements including compounds derived from spinach extract, in order to defend health and to trigger

In this work, the biological activity of several phytochemicals applied to a human cell line at different concentrations was evaluated. The results lead to a confirmation of their beneficial properties, acting on ROS and lowering the oxidative damage. This highlights, among other possibilities, also that of considering the development of supplements including compounds derived from spinach extract, in order to defend health and to trigger

In this work, the biological activity of several phytochemicals applied to a human cell line at different concentrations was evaluated. The results lead to a confirmation of their beneficial properties, acting on ROS and lowering the oxidative damage. This highlights, among other possibilities, also that of considering the development of supplements including compounds derived from spinach extract, in order to defend health and to trigger

decreased proliferative response, but the relation is not so direct. With almost all the compounds at low concentration an antioxidant activity can be noted but no inhibition of the

decreased proliferative response, but the relation is not so direct. With almost all the com-

induce impaired physiological functions and a ROS reduction seems to be related to a

dative balance, any variation in the amount of intracellular reactive oxygen species can

possible anticancer effects.

cluding compounds derived from spinach extract, in order to defend health and to trigger

line at different concentrations was evaluated. The results lead to a confirmation of their

cluding compounds derived from spinach extract, in order to defend health and to trigger

other compounds.

possible anticancer effects.

**5. Conclusions** 

beneficial properties, acting on ROS and lowering the oxidative damage. This highlights, among other possibilities, also that of considering the development of supplements in-

line at different concentrations was evaluated. The results lead to a confirmation of their beneficial properties, acting on ROS and lowering the oxidative damage. This highlights,

line at different concentrations was evaluated. The results lead to a confirmation of their beneficial properties, acting on ROS and lowering the oxidative damage. This highlights,

decreased proliferative response, but the relation is not so direct. With almost all the compounds at low concentration an antioxidant activity can be noted but no inhibition of the

decreased proliferative response, but the relation is not so direct. With almost all the com-

tiproliferative concentration, in this case, is lower than the biologically active one of the

but at high concentrations (100 µM). Similar behavior is observed with lutein but the antiproliferative concentration, in this case, is lower than the biologically active one of the

In this work, the biological activity of several phytochemicals applied to a human cell line at different concentrations was evaluated. The results lead to a confirmation of their beneficial properties, acting on ROS and lowering the oxidative damage. This highlights, among other possibilities, also that of considering the development of supplements including compounds derived from spinach extract, in order to defend health and to trigger

In this work, the biological activity of several phytochemicals applied to a human cell line at different concentrations was evaluated. The results lead to a confirmation of their beneficial properties, acting on ROS and lowering the oxidative damage. This highlights, among other possibilities, also that of considering the development of supplements including compounds derived from spinach extract, in order to defend health and to trigger

line at different concentrations was evaluated. The results lead to a confirmation of their

decreased proliferative response, but the relation is not so direct. With almost all the compounds at low concentration an antioxidant activity can be noted but no inhibition of the proliferation. This latter activity is observed just for 20-hydroxyecdysone and β-carotene but at high concentrations (100 µM). Similar behavior is observed with lutein but the antiproliferative concentration, in this case, is lower than the biologically active one of the

In this work, the biological activity of several phytochemicals applied to a human cell line at different concentrations was evaluated. The results lead to a confirmation of their beneficial properties, acting on ROS and lowering the oxidative damage. This highlights, among other possibilities, also that of considering the development of supplements including compounds derived from spinach extract, in order to defend health and to trigger

cluding compounds derived from spinach extract, in order to defend health and to trigger

among other possibilities, also that of considering the development of supplements including compounds derived from spinach extract, in order to defend health and to trigger

possible anticancer effects.

other compounds.

possible anticancer effects.

**5. Conclusions** 

possible anticancer effects.

possible anticancer effects.

possible anticancer effects.

possible anticancer effects.

other compounds.

other compounds.

other compounds.

other compounds.

possible anticancer effects.

possible anticancer effects.

possible anticancer effects.

possible anticancer effects.

**5. Conclusions** 

**5. Conclusions** 

**5. Conclusions** 

**5. Conclusions** 

**Author Contributions:** Conceptualization, A.B. and L.A.; methodology, F.M.; validation, A.B., and L.F.; formal analysis, A.B.; investigation, F.M.; writing—original draft preparation, F.M.; writing—review and editing, L.A. and L.F.; supervision, A.B. All authors have read and agreed to the published version of the manuscript.

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

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The remaining data are available on request from the corresponding author.

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

#### **References**


### *Article* **Mayten Tree Seed Oil: Nutritional Value Evaluation According to Antioxidant Capacity and Bioactive Properties**

**Rosanna Ginocchio 1,2, Eduardo Muñoz-Carvajal <sup>3</sup> , Patricia Velásquez 3,4 , Ady Giordano 3,\* , Gloria Montenegro <sup>4</sup> , Germán Colque-Perez <sup>5</sup> and César Sáez-Navarrete 5,6**


**Abstract:** The Mayten tree (*Maytenus boaria* Mol.), a native plant of Chile that grows under environmentally limiting conditions, was historically harvested to extract an edible oil, and may represent an opportunity to expand current vegetable oil production. Seeds were collected from Mayten trees in north-central Chile, and seed oil was extracted by solvent extraction. The seed oil showed a reddish coloration, with quality parameters similar to those of other vegetable oils. The fatty acid composition revealed high levels of monounsaturated and polyunsaturated fatty acids. Oleic and linoleic acids, which are relevant to the human diet, were well represented in the extracted Mayten tree seed oil. The oil displayed an antioxidant capacity due to the high contents of antioxidant compounds (polyphenols and carotenoids) and may have potential health benefits for diseases associated with oxidative stress.

**Keywords:** antioxidant capacity; DPPH; solvent extraction; nutrition; *Maytenus boaria*; ABTS

### **1. Introduction**

In recent decades, the oil crop sector has been one of the most dynamic agricultural segments worldwide, with a 4.3% per annum (p.a.) growth rate compared with an average of 2.1% p.a. for all agriculture [1]. Worldwide production, consumption, and trade in this sector have increasingly become dominated by a small number of crops [2,3], as palm, soybean, and rapeseed oils have represented 82% of the total global oil crop production since 1990 (according to oil equivalent measurements). Secondary oil crops represent major elements of the food supply and food security in several countries, including sesame oil (e.g., in Sudan, Uganda, Ethiopia, and Myanmar), groundnut oil (e.g., in Sudan, Ghana, Myanmar, Vietnam, Senegal, the United Republic of Tanzania, and Benin), coconut oil (e.g., in the Philippines, Sri Lanka, Vietnam, and Mexico), olive oil (Mediterranean countries), and cottonseed oil (e.g., in Central Asia, the Sahel, Pakistan, Turkey, and the Syrian Arabic Republic) [1].

Vegetable oils are derived from the seeds and fruits of plants. Among the vegetable oils that are derived from seeds (seed oils), most are currently obtained from only a few commercially significant species (i.e., soybeans, sunflowers, rapeseed, flax, oil palm nuts,

**Citation:** Ginocchio, R.; Muñoz-Carvajal, E.; Velásquez, P.; Giordano, A.; Montenegro, G.; Colque-Perez, G.; Sáez-Navarrete, C. Mayten Tree Seed Oil: Nutritional Value Evaluation According to Antioxidant Capacity and Bioactive Properties. *Foods* **2021**, *10*, 729. https://doi.org/10.3390/foods10040729

Academic Editors: Francisca Rodrigues and Cristina Delerue-Matos

Received: 26 February 2021 Accepted: 26 March 2021 Published: 30 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 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 (https:// creativecommons.org/licenses/by/ 4.0/).

castor beans, groundnuts, cottonseed, and shea nuts) [4]. The most common energy-rich compounds contained in the endosperm or cotyledons of seeds are carbohydrates (starches) and fatty acids (oils) [4,5]. However, several plant species that grow under environmentally limiting conditions (i.e., arid and semiarid climates or nutrient-poor soils) worldwide that are not currently used as oil crops are known to feature oil-bearing seeds [6]. These crops may constitute an opportunity for expanding vegetable oil production to regions where crop production is not feasible, either currently or in the future, due to global climate change scenarios. Therefore, many indigenous tree species, which may be more resistant than current agricultural crops to limiting environmental factors (i.e., heat, water stress, salinity, frosts, and pests) but are not yet grown commercially, are becoming increasingly recognized as potentially valuable sources of vegetable oils [7], such as the Mayten tree (*Maytenus boaria* Mol.).

The Mayten tree is a medium-sized evergreen tree (up to 15 m in height) that is native to Chile, Argentina, Peru, and Brazil [8]. In Chile, the species has a wide geographic distribution (28◦ to 45◦ Southern latitude and 15 to 1800 msl) [9,10], presenting great adaptability to different site conditions, such as precipitation levels (mean annual values from 355 to 2351 mm), soil pH (neutral to acidic), and soil water availability (dry to saturated) [8]. These trees have been described as having high seed oil contents. Scattered historical information has suggested that this seed oil was extracted from seeds collected from wild trees and used for human consumption during Colonial times (from 1600 to 1810) in central Chile [11]. Large-scale seed oil extraction and exportation to France has been described during the 19th century [12], likely for lamp oil use [13]. However, scarce information is available regarding the amounts produced, the extraction methods used, and the specific physicochemical characteristics of this seed oil. According to the work of Vicente Bustillos, from 1846 [13], Mayten tree seeds are easy to grind and press (cold or heat press), and large amounts of oil can be obtained (approx. 25% of total weight); the oil is transparent, reddish-yellowish in color, featuring a bitter taste and a density similar to that of olive oil (specific weight of 0.92), and it begins to freeze at 4–5 ◦C. Recent literature has indicated high oil contents in the seeds of the Mayten tree (40%) [14], which may be used for cooking [15] and industrial purposes (i.e., as a substitute for linseed oil) [16]. However, none of these applications are currently in use.

The Mayten tree belongs to the *Celastraceae* family, a plant group known to be rich in secondary metabolites (i.e., sesquiterpenes and agafurans), some of which have been reported to demonstrate interesting pharmacological behaviors [17], antiseptic properties [18] or pesticidal activities [19]. Other chemical compounds (i.e., agafurans) that have been isolated from the Mayten leaves feature biopesticidal or natural insecticidal properties [20]. Most of the studies that have been performed to explore secondary metabolites and bioactive properties have examined vegetative aerial tissues (leaves and stems); however, no studies have examined seeds or seed oil [8,14,21,22]. Therefore, the primary objective of the present study was to perform a physicochemical characterization of Mayten tree seed oil, including the evaluation of antioxidant capacity and bioactive properties to determine the nutritional value.

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

#### *2.1. Seed Materials and Seed Oil Extraction*

Seeds from Mayten trees (Figure S1) were collected in north-central Chile (Elqui Valley, Santiago Metropolitan Park, and Quirihue) from February 2016 to April 2016. The seeds were hand-cleaned, allowed to air dry, and stored at 5 ◦C. Seeds were ground using an electric coffee grinder, and the oil was extracted with solvents (2:4:2 *w*:*v*:*v* ratio of seeds:methanol:chloroform), according to the method described by Bligh and Dyer [23]. The mixture of ground seeds and solvents was agitated at 200 rpm for 3 h. The method described by Bligh and Dyer was modified by adding water only in the second stage of the extraction. Distilled water and chloroform were then added (1.8:2.0:2.0, *v*:*v*:*v*, mixture:water:chloroform), by forming a ternary system, and the mixture was vacuum filtered

through a 7 µm pore filter. Two phases were collected, the phase of methanol–water composition (top), and the chloroform–seed oil (bottom). The chloroform was evaporated using a rotary evaporator (40 ◦C for 30 min), and the seed oil was stored at 4 ◦C for 30 min in the dark (Figure S1).

#### *2.2. Physicochemical Analysis*

Determination of the seed oil color was performed according to the method described by Popa and Doran [24] using a colorimeter (WR10, DANSTLEE). Each extract was placed in a quartz cuvette for measurement. Coordinate values were obtained, where L represents lightness and varies between 0 (black) and 100 (white), a expresses red (+) or green (−), and b indicates yellow (+) or blue (−). Chroma (C◦ ) and hue (H◦ ) values were obtained from the following Equations:

$$\mathbf{C}^{\diamond} = \sqrt{\mathbf{a}^2 + \mathbf{b}^2} \tag{1}$$

$$\mathbf{H}^{\diamond} = \arctan{\frac{\mathbf{a}}{\mathbf{b}}}\tag{2}$$

The seed oil density was determined using pre-calibrated volumetric flasks [25]. The peroxide index, which measures the initial oxidation state, is expressed in milliequivalent of oxygen per kilogram of oil. The iodine value indicates the degree of unsaturation, expressed in gram of iodine absorbed by 100 g of oil. Free acidity, is the percentage of free acid present in the oil, expressed in oleic acid percentage. Rancimet 743 measures the stability oxidation of the oil, corresponds to the induction period, expressed in hours, and thiobarbituric acid reactive substances (TBARS) measure malondialdehyde (MDA) present in the sample. All analyses were determined according to the American Oil Chemists' Society (AOCS) standard method [26], as described by Petropoulus et al. (2020) [27].

#### *2.3. Fatty Acid Profile*

Gas chromatography (GC) coupled with a flame ionization detector (FID) was used according to the AOCS standard method [26]. The following standard measurements were used: butanoic acid (C4:0); caproic acid (C6:0), caprylic acid (C8:0); capric acid (C10:0); undecylic acid (C11:0); lauric acid (C12:0); tridecylic acid (C13:0); myristic acid (C14:0); pentadecylic acid (C15:0); palmitic acid (C16:0); margaric acid (C17:0); stearic acid (C18:0); arachidic acid (C20:0); japonic acid (C21:0); behenic acid (C22:0); tetrasanoic acid (C24:0); tetradecene acid (C14:1); pentadecylic acid (C15:1); palmitoleic acid (C16:1); heptadecene acid (C17:1); oleic acid (C18:1); timnodonic acid (C20:1n9); erucic acid (C22:1n9); tetrasaenoic acid (C24:1); linoleic acid (C18:2n6); γ-linoleic acid (C18:3n6); α-linoleic acid (C18:3n3); gondoic acid (C20:2n6); di-homo-γ–linoleic acid (C20:3n6); dihomolinolenic acid (C20:3n3); eicosatetranoic acid (C20:4n6); eicosapentaenoic acid (EPA) (C20:5n3); docosadienenoic acid (C22:2); and docosahexaenoic acid (DHA) (C22:6n3).

#### *2.4. Lipidic Indices*

The qualities of the lipids in the seed oil were determined by evaluating the ratio of polyunsaturated fatty acids (PUFAs) to saturated fatty acids (SFAs) [28]. Two indices of coronary heart disease risk—the atherogenic index (AI), used to obtain the standard cardiac risk estimation; and the thrombogenic index (TI), which indicates the tendency to form clots in the blood vessels—were calculated as functions of the composition of monounsaturated fatty acids (MUFAs), PUFAs, SFAs, and specific fatty acids [29], as follows:

$$\text{AI} = \frac{(\text{C12} : 0 + 4 \times \text{C14} : 0 + \text{C16} : 0)}{(\text{MUFA} + \text{PUFA})} \tag{3}$$

$$\text{TI} = \frac{(\text{C14} : 0 + \text{C16} : 0 + \text{C18} : 0)}{\left(0.5 \times \text{MUFA} + 0.5 \times \text{n} - 6 \text{PUFA} + 3 \times \text{n} - 3 \text{PUFA} \times \left(\frac{\text{n} - 3 \text{PUFA}}{\text{n} - 6 \text{PUFA}}\right)\right)} \tag{4}$$

#### *2.5. Extraction of Antioxidant Compounds*

Seed oil was mixed with methanol and hexane (1:5:1, *v*:*v*:*v*, oil:methanol:hexane) and maintained in an ultrasonic bath at 20 Hz for 20 min at a room temperature bath [30]. The solution was then centrifuged at 4000 rpm for 20 min at room temperature. The mixture was stored at 4 ◦C for 1 h, and the supernatant was then filtered through a 0.22 µm membrane with a syringe filter for the antioxidant extraction.

#### *2.6. Quantification of Phenols and Flavonoids, and Determination of Antioxidant Capacity*

Total polyphenolic contents (TPC) were estimated using the Folin–Ciocalteu (FC) method based on the antioxidant extract from oil, as described by Velásquez et al. (2019) [31]. Gallic acid (GA) was used as the standard; therefore, the resulting values are expressed in mg GAE (100 g)−<sup>1</sup> of seed oil.

Total flavonoid contents (TFC) were estimated using the aluminum chloride method based on the antioxidant extract from oil, as described by Velásquez et al. (2019) [31]. Quercetin (Q) was used as the standard, and results are expressed in mg QE (100 g)−<sup>1</sup> of seed oil. Some specific phenolic acids and flavonoids, such as 4-hydroxy benzoic acid, apigenin, caffeic acid, catechin, chlorogenic acid, cinnamic acid, chrysin, coumaric acid, epigenin, ferulic acid, gallic acid, kaempferol, luteolin, pinocembrin, quercetin, resveratrol, rutin, sinapic acid, syringic acid, vanillic acid, and phytohormone abscisic acid were quantified through using ultra-performance liquid chromatography (UPLC)-tandem mass spectrometry (MS/MS), according to the method described by Giordano et al. (2019) [32], by interpolation from standard calibration curves.

Antioxidant capacity was determined using the ferric reducing antioxidant power (FRAP) and 2,2′ -azino-bis-(3-ethylbenzothiazoline-6-sulfonic acid (ABTS) radical-stabilization methods, according to the methods described by Diniyah et al. (2020) [33]. The antioxidant extract was measured using a FeSO<sup>4</sup> standard and is expressed in mg FeSO4/100 g of seed oil, whereas Trolox (T) was used as the standard for radical stabilization, which is expressed in T equivalents (TE 100 g)−<sup>1</sup> of seed oil. The 2,2-diphenyl-1-picrylhydrazyl (DPPH) radical was also assessed as described by Diniyah et al. (2020) [33], with some modifications, using several antioxidant extract dilutions. A curve was generated comparing the inhibition percentage against the tested dilutions (logarithmic relationship), interpolating the half-maximal inhibitory concentration (IC50), which is expressed in g of seed oil per mL−<sup>1</sup> of methanolic extract. The Agilent 8453 UV-visible spectrophotometer (Palo Alto, CA, USA) was used for these analyses.

#### *2.7. Extraction and Quantification of Carotenoids and β-Carotene*

A carotene extract was obtained through seed oil saponification, as described by Varzakas and Kiokias (2016) [34]. The quantification of total carotenes was performed according to Bihler et al. (2010) [35], using β-carotene as the standard and expressed as mg β-carotene/100 g of seed oil. Absorbance was measured using the Agilent 8453 UV-visible spectrophotometer (Palo Alto, CA, USA).

The carotene profile and β-carotene levels were obtained using a high-performance liquid chromatography (HPLC)-diode-array detector (DAD). A reverse phase C18 column was used (150 mm × 4 mm × 5 µm), with a mobile phase flow of 1.5 mL min−<sup>1</sup> . Methanol (A), water (B), and n-butanol (C) were used as solvents in the following concentration gradient: 0 min 3% A, 92% B, and 5% C; 4 min 0% A, 92% B, and 8% C; 18.1 min 3% A, 92% B, and 5% C, until 23 min. The carotene profile was measured at 440 nm. The β-carotene level was interpolated from the calibration curve of a 10–100 mg/L β-carotene standard.

#### *2.8. Statistical Analysis*

All analyses were realized in triplicate (*n* = 3), and the data were expressed as the mean ± standard deviation using the statistical software Minitab 19.

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

#### *3.1. Seed Oil Extraction*

A yield of 61.77 ± 8.24% *w*/*w* seed oil was extracted from Mayten tree seeds, which is a yield larger than those for seed oils extracted from sunflower (51.00% *w*/*w*), sesame (48,00% *w*/*w*), and pumpkin seeds (46.00% *w*/*w*) [36].

#### *3.2. Physicochemical Characteristics*

The seed oil extracted from Mayten tree seeds had a reddish coloration, according to the tone angle (H: 59.40◦ ), which is categorized among the purple–red colors [37]. The oil had a lower yellow and a higher red color composition (Table 1) compared with palm oil (b\* = 56.87–74.50; a\* = 2.21–13.0) [37]. According to the clarity analysis, Mayten tree seed oil is closer to white, based on a 1 to 100 scale, and is slightly darker than palm oil [37].

**Table 1.** Physicochemical characterization of Mayten tree seed oil.


\* adimentional value. All values reported in this table correspond to mean ± standard deviation.

The Blight and Dyer solvent extraction method allows for the extraction of high molecular weight lipophilic compounds that exist in a solid state at 25 ◦C. This method resulted in oil dispersion [37,38], with a density greater than 1 (1.06 ± 0.07 g mL−<sup>1</sup> ); as reported in Table 1, this is a value similar to the density of pine oil (1.042–1.224 g mL−<sup>1</sup> ) extracted using organic solvents [25]. The seed oil density value was higher than that reported for palm oil (1.06 g mL−<sup>1</sup> versus 0.892–0.899 g mL−<sup>1</sup> ); such a value could indicate the presence of high molecular weight compounds, such as carotenoids and fatty acid [38].

The peroxide value identified for the Mayten tree seed oil was 5.10 ± 0.18 meq O2/kg oil (Table 1), a value similar to that reported for olive oil [23], and was within the range reported for palm oil (1.0–10.0 mg eq O<sup>2</sup> g <sup>−</sup><sup>1</sup> oil) [38]. In addition, this peroxide value is below that allowed by the Chilean Sanitary Regulations of Food for the year 2015 (10.0 mg eq O<sup>2</sup> g <sup>−</sup><sup>1</sup> oil) and also lower than that obtained from the seeds of the maqui berry [23].

The iodine value of 57.63 identified for Mayten tree seed oil was similar to the value for palm oil (46.0–56.0), but higher than that for coconut oil (6.0–11.0) [38,39], showing a similar degree of unsaturation with these oils. The free acidity of Mayten tree seed oil was within the range identified for palm oil (3.7–5%, Table 1) and within the range allowed by the Chilean Sanitary Regulations of Food for the year 2015 [23].

The TBARS index of the Mayten tree seed oil had a lower value than those for avocado and olive seed oils [40]. When examining the oxidative stability under acceleration conditions (Rancimet analysis), the Mayten tree seed oil demonstrated a longer induction period (52.15 ± 2.15 h) than soybean oil (11.2 h) or extra virgin olive oil (24–26 h) [41]. According to the obtained physicochemical properties, Mayten tree see oil has values similar to other commercial oils indicating that it could be used for human consumption.

#### *3.3. Fatty Acid Profile and Nutritional Quality*

The fatty acid profile and nutritional qualities of Mayten tree seed oil were as shown in Figure 1 and Table 2, respectively. Mayten tree seed oil had a high unsaturated fatty acid content (Figure 1). The MUFA concentration of this seed oil (25.70 g/100 g oil) (Table 2) was lower than that for avocado seed oil (49–75.96%) [40,42,43], palm oil (37.1–39.2%) [38], olive oil (80.53%) [40], and extra virgin olive oil (71.32–79.62%) [41]. However, the PUFA composition of 20.98 g/100 g oil (20.98%) (Table 2) was greater than that for palm oil (8.1–10.5%) [38] and olive oil (5.43%) [40] and was within the range of that for avocado seed oil (11.75–37.13%) [40,42,43]. Based on these results, Mayten tree seed oils would likely provide health benefits, as unsaturated fatty acids favor specific enzymatic reactions such as cyclooxygenases, lipoxygenases and cytochrome P450 enzymes that resolve different processes of inflammation and protect against brain or renal dysfunctions [43,44], and PUFAs can provide the essential fatty acids that must be obtained through the diet.

**Figure 1.** Profile of fatty acids of Mayten tree seed oil.


**Table 2.** Composition of the relevant groups of fatty acids and the nutritional qualities of Mayten tree seed oil.

The ingestion of foods with PUFA/SFA ratios in the range of 1.25 to 2.4 has been described to confer beneficial effects for the prevention of cardiovascular diseases (CVDs) [45]. As Mayten tree seed oil was found to have a PUFA/SFA value of 1.97 (Table 2), this oil would provide a comparative advantage over other commonly consumed seed oils, such as olive oil and menhaden oil [28].

The oleic acid (C18:1), linoleic acid (C18:2n6), and palmitic acid (C16:0) contents were high (in decreasing order) in Mayten tree seed oil (Figure 1). The oleic acid contents of Mayten tree seed oil were higher than in sunflower oil, similar to contents in soybean oil [46,47], and lower than in palm oil [38]. Oleic acid (C18:1) is an omega-9 MUFA with anti-thrombosis and other bioactive properties, which has been used in cosmetic and pharmacological applications [48].

The linoleic acid (C18:2n6) content in Mayten tree seed oil was higher than that in olive oil and palm oil [38]. Linoleic acid is often described as a precursor for other lipid mediators with anti-inflammatory properties [43]. Linoleic acid (C18:2n6) is an essential omega-6 fatty acid that has been shown to be involved in the maintenance of the skin's permeable barrier against water, and the topical application of linoleic acid (C18:2n6) has been shown to improve dermatitis. Furthermore, linoleic acid is metabolized to arachidonic acid, which is a precursor of the eicosanoid compounds that regulate a range of physiological processes [49].

Palmitoleic acid (C16:1) and α-linoleic acid (C18:3n3) were found among the unsaturated fatty acid composition of Mayten tree seed oil (Figure 1), at higher levels than are found in sesame, sunflower, and rice brain [50]. Canola seed oil contains 3.4 g of palmitic acid per 100 g of oil [51], whereas the palmitic acid contents of Mayten tree seed oil were two-fold higher that in canola seed oil. α

Two indices of coronary heart disease risk, AI and TI [29], were calculated for Mayten tree seed oil and other commonly consumed seed oils, as shown in Figure 2. The Mayten tree seed oil's AI value was lower than for rice bran oil (Figure 2) and menhaden oil [28] but higher than for sunflower oil, canola oil, and olive oil (Figure 2). By contrast, the TI value for Mayten tree seed oil was lower than values found in other commonly used seed oils, such as olive, sunflower [28], rice bran, and canola oils [50] (Figure 2), likely due to the Mayten tree seed oil's higher contents of oleic acid, a compound that reduces the risks of thrombosis [52]. These results indicate that Mayten tree seed oils could potentially reduce the risks of generating thrombi in blood vessels compared with other commonly consumed seed oils.

**Figure 2.** Atherogenic index (AI) and thrombogenic index (TI) values among Mayten tree seed oil and other commonly consumed seed oils (adapted from ref. [50]).

#### *3.4. Polyphenol and Flavonoid Compounds*

Polyphenol compounds (Table 3) were co-extracted with fatty acids during the seed oil extraction process from Mayten tree seeds, similar to other seed oils [23]. The TPC value (16.1 mg GAE/100 g oil) in Mayten tree seed oil represents 0.016% of the total oil contents, which is a larger proportion than is found in canola (11.26 mg GAE/100 g oil) [53] and olive

oils (2–13 mg GAE/100 g oil) [54]. Flavonoids are a type of polyphenol component that dissolves in oil, and Mayten tree seed oil was found to contain a TFC of 15.5 mg EQ/100 g seed oil, which is similar to that reported for grape, rice bran, and chia seed oils but higher than those reported for sunflower, canola, soybean, and cottonseed oils [55].


**Table 3.** Phenolic and carotenoid compound contents in Mayten tree seed oil.

All values reported in this table correspond to mean ± standard deviation.

Flavonoids, such as quercetin and myricetin; phenolic acids, such as ferulic, coumaric, and syringic acids; and the sesquiterpene abscisic acid (Table 3) were found in Mayten tree seed oil. Ferulic acid, the most abundant compound in mayten tree seed oil, has been demonstrated to display various pharmacological properties, including anti-inflammatory activities [56,57]. Ferulic acid has been shown to reduce cholesterol synthesis in the liver, followed by an increase in sterol acid secretion, and has been shown to act as a chemoprotective agent against coronary heart disease, preventing thrombi and sclerosis production [57,58]. Ferulic acid also inhibits the population growth of influenza, syncytial, and human immunodeficiency viruses [56] and has been shown to exert anticancer activity against colon and rectal cancer [57]. Similar to ferulic acid, the other polyphenolic compounds identified in this seed oil are known to provide health benefits, such as antioxidant properties against free radical formation and the prevention of diseases and infections due to antimicrobial activities [59].

#### *3.5. Total Carotenoids and β-Carotene*

The reddish coloration observed for Mayten tree seed oil was likely derived from its high carotene content (Table 3), which is present in the seed aril (Figure S1) [42,43]. This relationship between seed oil color and presence of carotenoids has also been described for palm and cucumber seed oils [60,61]. Mayten tree seed oil is darker in color because it has a total carotene content (TCC) that is three times higher than that of palm seed oil [38].

Approximately 70% of the total carotene content in Mayten tree seed oil is represented by β-carotene (Table 3). As a precursor of vitamin A, β-carotene plays an important role in the prevention of cataracts and macular hatching, in addition to improving night blindness and dry eyes [62]. The difference between the TCC and β-carotene contents in Mayten tree seed oil is expected to indicate the presence of other carotenoid compounds that were not specifically identified in the present study. The high carotene contents, together with the presence of polyphenols, may give this seed oil resistance to fatty acid peroxidation, which would prevent rancidity [63].

#### *3.6. Antioxidant Capacity*

The resistance against factors that cause rancidity in seeds oils is a function of the antioxidant capacity of seed oils, which can provide nutritional value without requiring the incorporation of synthetic antioxidants, such as those used by the food industry [64]. The antioxidant capacity of Mayten tree seed oil against the ABTS radical appeared to be lower than the capacities found in other seed oils (Table 4). However, this characteristic is typically determined based on the presence of polyphenolic compounds (as in the present study) and carotenoids because these molecules also react to the ABTS radical [65].


**Table 4.** Antioxidants capacity in Mayten tree seed oil.

All values reported in this table correspond to mean ± standard deviation.

The FRAP value is a measure of the antioxidant capability based on the evaluation of electron donation that occurs due to the activity of antioxidant compounds (Table 4) [66]. For Mayten tree seed oil, the FRAP value was three-fold higher that of butylhydroxytoluene (BHT), an artificial antioxidant commonly used by the food industry [64], indicating that Mayten tree seed oil displays a high antioxidant capacity. According to the IC<sup>50</sup> value, a low amount of Mayten tree seed oil dissolved in methanol was necessary to inhibit 50% of the DPPH radical [64], indicating a better response against the proton donation mechanism of oxidated compounds. Based on the FRAP results, the antioxidant capacity of Mayten tree seed oil was found to be three-fold that of rapeseed oil and 15-fold those for sunflower, rice brain, and olive oils [50,67]. The percentage of DPPH radical inhibition of a decreasing methanol dilution range of Mayten tree seed oil is shown in Figure 3b. Initially, the percentage of inhibition increased as the contents of the seed oil increases, followed by a stabilization when 60–70% inhibition of the DPPH radical was achieved. Therefore, the examined Mayten tree seed oil appears to have a higher antioxidant capacity than the BHT supplement commonly used by the food industry (Figure 3).

**Figure 3.** (**a**) Ferric reducing antioxidant power (FRAP) values for Mayten tree seed oil, sunflower oil, brain rice oil, canola oil, and olive oil. (Adapted with permission from Szydłowska-Czerniak, A. et al. [67], Copyright 2008 Elsevier); (**b**) Inhibition of the 2,2-diphenyl-1-picrylhydrazyl (DPPH) radical by Mayten tree seed oil. Values are represent as mean ± standard deviation.

#### **4. Conclusions**

Mayten tree seed oil was obtained through the application of the Blight and Dyer solvent (methanol–chloroform–water) extraction method, resulting in a 61.77% yield, which may be further improved using other extraction procedures. This seed oil was found to have high omega-6 and omega-9 fatty acid contents, with a higher PUFA content than that found in most major commercial vegetable oils.

The analyzed quality parameters indicated that Mayten tree seed oil has a high resistance against rancidity, with values similar to those of other commercial seed oils and within the requirements of Chilean Health Regulations. The high TPC and TFC values for this seed oil provide protection against lipid peroxidation and result in high ABTS, DPPH, and FRAP antioxidant capacities, which may be beneficial for human health and pharmacological uses.

The reddish coloration of Mayten tree seed oil is associated with a high content of carotenoids, which may constitute a comparative advantage relative to other vegetable oils currently used for human consumption. β-carotene is the dominant carotenoid in this seed oil, potentially providing elements that can contribute to the functional properties of this oil.

The results of the present study indicate that Mayten tree seed oil has nutritional value, based on the antioxidant capacity and bioactive properties identified for this oil. This seed oil would be an interesting alternative to other vegetable oils intended for human consumption, particularly since it could be produced in areas affected by global climate change with higher yields than other traditional oils. It could be considered as a functional food, a carotenoids supplement or an antioxidants additive ingredient for the food industry. However, studies exploring the domestication of this tree species are necessary.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/foods10040729/s1. Figure S1: (a) Fruits and seeds (orange-reddish color) of Mayten tree, (b) seed oil of Mayten tree extracted with solvents (methanol and chloroform).

**Author Contributions:** Conceptualization, R.G. and A.G.; methodology, E.M.-C., G.C.-P., A.G. and G.M.; validation, C.S.-N.; formal analysis, E.M.-C., P.V. and A.G.; investigation, E.M.-C.; writing original draft preparation, E.M.-C., P.V. and A.G.; writing—review and editing, R.G. and C.S.-N.; supervision, R.G., A.G., G.M. and C.S.-N.; funding acquisition, R.G. and C.S.-N. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Proyecto Interdisciplina VRI-UC N◦6/2015 and FONDEQUIP under Grants EQM 130032 and EQM160042.

**Acknowledgments:** The authors acknowledgment the analytical support of the Instituto de Nutrición y Tecnología de Alimentos (INTA) for determination of fatty acids and quality analysis. Authors thanks to Víctor Ahumada for his contribution to the antimicrobial capacity analysis.

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

#### **References**


### *Article* **Bioactive Compounds in Wild Nettle (***Urtica dioica* **L.) Leaves and Stalks: Polyphenols and Pigments upon Seasonal and Habitat Variations**

**Maja Repaji´c <sup>1</sup> , Ena Cegledi <sup>1</sup> , Zoran Zori´c <sup>1</sup> , Sandra Pedisi´c <sup>1</sup> , Ivona Elez Garofuli´c 1,\*, Sanja Radman <sup>2</sup> , Igor Palˇci´c <sup>3</sup> and Verica Dragovi´c-Uzelac <sup>1</sup>**


**Abstract:** This study evaluated the presence of bioactives in wild nettle leaves and stalks during the phenological stage and in the context of natural habitat diversity. Thus, wild nettle samples collected before flowering, during flowering and after flowering from 14 habitats situated in three different regions (continental, mountain and seaside) were analyzed for low molecular weight polyphenols, carotenoids and chlorophylls using UPLC-MS/MS and HPLC analysis, while the ORAC method was performed for the antioxidant capacity measurement. Statistical analysis showed that, when compared to the stalks, nettle leaves contained significantly higher amounts of analyzed compounds which accumulated in the highest yields before flowering (polyphenols) and at the flowering stage (pigments). Moreover, nettle habitat variations greatly influenced the amounts of analyzed bioactives, where samples from the continental area contained higher levels of polyphenols, while seaside region samples were more abundant with pigments. The levels of ORAC followed the same pattern, being higher in leaves samples collected before and during flowering from the continental habitats. Hence, in order to provide the product's maximum value for consumers' benefit, a multidisciplinary approach is important for the selection of a plant part as well as its phenological stage with the highest accumulation of bioactive compounds.

**Keywords:** nettle leaves and stalks; phenological stage; location; accelerated solvent extraction; UPLC-MS/MS; polyphenols; chlorophylls; carotenoids; antioxidant capacity; ORAC

#### **1. Introduction**

Nettle (*Urtica dioica* L.) is a perennial wild plant of the Urticaceae family, genus Urtica, which is widespread in Europe, Asia, America and part of Africa, and has been adapted to different climatic conditions [1,2]. Nettle has long been used in the food, cosmetic and pharmaceutical industries due to its nutritional and health potential, as all parts of nettle (leaves, stalks and roots) show a rich composition of bioactive compounds with high antioxidant capacity [2,3] Previous studies have shown that nettle leaves and stalks are a rich source of vitamins A, B and C, minerals (iron, potassium, calcium, magnesium), polyphenols such as phenolic acids and flavonoids as well as pigments, especially chlorophyll and carotenoids [4–11]. In accordance with the above, aerial parts of nettle have anti-inflammatory and therapeutic effects; these nettle parts are used in the treatment of arthritis, anemia, allergies, joint pain and urinary tract infections, have a diuretic effect and are used to strengthen hair [3,12]. Besides aerial parts, nettle root also presents a rich source of various compounds such as protein lectin, sterols, polysaccharides, lignans and phenols [5,7,13,14] and is mostly used in the treatment of benign prostatic hyperplasia [15]. Apart from medicinal use, other applications of nettle include food preparation, where it

**Citation:** Repaji´c, M.; Cegledi, E.; Zori´c, Z.; Pedisi´c, S.; Elez Garofuli´c, I.; Radman, S.; Palˇci´c, I.; Dragovi´c-Uzelac, V. Bioactive Compounds in Wild Nettle (*Urtica dioica* L.) Leaves and Stalks: Polyphenols and Pigments upon Seasonal and Habitat Variations. *Foods* **2021**, *10*, 190. https://doi.org/ 10.3390/foods10010190

Received: 23 November 2020 Accepted: 14 January 2021 Published: 18 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 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 (https:// creativecommons.org/licenses/by/ 4.0/).

is consumed in the form of tea, soup, stew or salad [3], or for commercial extraction of chlorophyll, which is used as a green coloring agent (E140) [16].

For medicinal purpose and medicinal preparations, nettle is mostly often used in the form of liquid or dry extract; thus, it is important to apply extraction method that will give a highly stabile extract with the greatest possible content of bioactive ingredients. Therefore, new extraction methods are increasingly being used and one of them is accelerated solvent extraction (ASE). In addition to being an efficient method, it uses less solvent, shortens the extraction time and more effectively isolates the target components [17].

Aside from extraction method, extract quality and richness in bioactives also depends on used plant material, either wild or cultivated, where its chemical composition and consequently antioxidant capacity are influenced by environmental, genotypic and phenotypic factors.

Different parts of plant may contain different amounts of particular compounds, e.g., nettle leaves accumulate higher amounts of polyphenols and chlorophylls in comparison with stalks [6,7,18]. In general, leaves are the richest part of a nettle in bioactive compounds, therefore they are mostly used in processing. However, changes in chemical composition and compounds' distribution occur with plant's maturity, where bioactive compounds are present in different proportions during different phenological stages. For example, the content of polyphenols decreases with growth and maturity of the plant [19]. Bioactive compounds are produced in response to different forms of (a)biotic stresses, as well as to fulfil important physiological tasks (attracting pollinators, establishing symbiosis, providing structural components to lignified cell walls of vascular tissues, etc.) [20]. These processes are often connected to specific phenological stages. Hence, harvest time depends on the type of final product. Although opinion on nettle optimal harvest time differs among various authors [3], Moore (1993) [21] stated that for juices and other fresh preparations, nettle leaves are best picked in spring or early summer (before flowering), and according to Upton (2013) [3] for dried preparations, it is best to harvest from mid-spring to late summer. If nettle is used for food purposes, the recommended harvest should be at the pre-flowering and flowering phases, certainly before the appearance of the seeds when it contains the least bioactive ingredients [3].

Nevertheless, nettle herb is mostly wild-harvested [3]. Concerning the natural habitat and climate, nettle is a quite adaptable plant. It grows in areas characterized by mild to temperate climates and prefers open or partly shady habitats with plenty of moisture such as forests, by rivers or streams and on roadsides [2]. Still, accumulation of polyphenols and pigments varies upon climate and habitat diversity. Plants grown in cold climates often show greater antioxidant properties, as a result of oxidative stress defense [22], while pigments synthesis is enhanced due to exposure to higher temperatures and more sunlight [23,24].

Although mentioned scientific literature provides data regarding nettle chemical composition, to our best knowledge there are no comprehensive studies on polyphenols and pigments constituents and their accumulation in wild nettle leaves and stalks during different vegetation periods of growing across diverse regions. These cognitions could be beneficial input data in a production of liquid and dry extracts. Therefore, the current study aimed to examine the presence and profile of low molecular weight polyphenols, carotenoids and chlorophylls as well as to determine antioxidant capacity in wild nettle leaves and stalks collected during three phenological stages (before flowering, during flowering and after flowering) from 14 different natural habitats situated in three regions in Croatia.

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

#### *2.1. Chemicals*

HPLC grade acetonitrile was procured from J.T. Baker Chemicals (Deventer, Netherlands). Purified water was obtained by a Milli-Q water purification system (Millipore, Bedford, MA, USA). Ethanol (96%) was purchased from Gram–mol d.o.o. (Zagreb, Croa-

tia) and formic acid (98–100%) from T.T.T. d.o.o. (Sveta Nedjelja, Croatia). Commercial standards of quercetin-3-glucoside, kaempferol-3-rutinoside, myricetin, caffeic acid, gallic acid, ferulic acid, sinapic acid, quinic acid, chlorogenic acid, *p*-coumaric acid, esculetin, scopoletin, α-carotene, β-carotene, chlorophyll *a* and chlorophyll *b* were purchased from Sigma–Aldrich (St. Louis, MO, USA). Epicatechin, catechin, epigallocatechin gallate, epicatechin gallate, apigenin, luteolin and naringenin were obtained from Extrasynthese (Genay, France), while quercetin-3-rutinoside was procured from Acros Organics (Thermo Fisher Scientific, Geel, Belgium). Apigenin was dissolved in ethanol with 0.5% (*v/v*) dimethyl sulfoxide, standards of carotenoids and chlorophylls in *n*-hexane. Other standards were prepared as a stock solution in methanol, and working standard solutions were prepared by diluting the stock solutions to yield five concentrations.

#### *2.2. Plant Material*

Samples of wild nettle (*Urtica dioica* L.) were collected at three phenological stages [(I) before flowering, (II) during flowering and (III) after flowering] during 2019 from different habitats in Croatia belonging to three regions (continental, mountain and seaside) (Table 1). Plant material was identified by using usual keys and iconographies with support of Department of Vegetable Crops, Faculty of Agriculture, University of Zagreb (Croatia). Immediately after harvesting, leaves were separated from stalks and samples were stored at −18 ◦C, freeze-dried (Alpha 1-4 LSCPlus, Martin Christ Gefriertrocknungsanlagen GmbH, Osterode am Harz, Germany) and afterwards grinded into fine powder using a commercial grinder (GT11, Tefal, Rumilly, France). Obtained powders were immediately analyzed for total solids by drying at 103 ± 2 ◦C to constant mass [25] and further used for the extraction. Content of dry matter in samples was >95%.




**Table 1.** *Cont.*

a.d. T = average day temperature, T min = minimal day temperature, T max = maximal day temperature, a.p. = accumulated precipitation.

#### *2.3. Extraction Conditions*

Extraction of polyphenols and pigments from dry nettle leaves and stalks was carried out by ASE. Extraction conditions and procedure were adopted from the study of Repaji´c et al. (2020) [11]: extraction was performed in Dionex™ ASE™ 350 Accelerated Solvent Extractor (Thermo Fisher Scientific Inc., Sunnyvale, CA, USA) using ethanol (96%) as the extraction solvent. Extraction was accomplished in 34 mL stainless steel cells fitted with 2 cellulose filters (Dionex™ 350/150 Extraction Cell Filters, Thermo Fisher Scientific Inc., Sunnyvale, CA, USA), within which 1 g of sample was mixed with 2 g of diatomaceous earth, placed in cell and filled up with diatomaceous earth to the full cell volume. Extraction parameters differed for leaves and stalks: leaves were extracted under 110 ◦C with 10 min of static extraction time and 4 cycles, while stalk extracts were obtained at 80 ◦C, 5 min of static extraction time and 4 cycles (parameters previously optimized). Other extraction parameters remained fixed for the extraction of both plant parts, namely pressure 10.34 MPa, 30 s of purge with nitrogen and 50% of flushing. Obtained extracts were collected in 250 mL glass vessel with Teflon septa, transferred into 50 mL volume flask and made up to volume with the extraction solvent. All extracts were filtered through a 0.45 µm membrane filter (Macherey-Nagel GmbH, Düren, Germany) prior to further analysis. All extracts have been prepared in a duplicate (n = 2).

#### *2.4. UPLC-MS/MS Conditions*

Identification and quantification of phenolics were performed on UPLC–MS/MS in both ionization modes on a 6430 QQQ mass spectrometer Agilent Technologies (Agilent, Santa Clara, CA, USA). Analytes were ionized using ESI ion source with nitrogen as desolvation and collision gas (temperature 300 ◦C, flow 11 L h−<sup>1</sup> ), capillary voltage, +4 −3.5 kV−<sup>1</sup> and the pressure of nebulizer was set at 40 psi. The mass spectrometer was linked to UPLC system (Agilent series 1290 RRLC instrument) consisted of binary pump, autosampler and a column compartment thermostat. Reversed phase separation was performed on a Zorbax Eclipse Plus C18 column 100 × 2.1 mm with 1.8 µm particle size (Agilent). Column temperature was set at 35 ◦C and the injection volume was 2.5 µL. The solvent compositions and the gradient conditions used were as described previously by Elez Garofuli´c et al. (2018) [26]. For instrument control and data processing, Agilent MassHunter Workstation Software (ver. B.04.01) was used. Quantitative determination was carried out using the calibration curves of the standards, where protocatechuic acid, gentisic acid, syringic acid and *p*-hydroxybenzoic acid were calculated as gallic acid equivalents and cinnamic acid according to *p*-coumaric acid. Isorhamnetin rutinoside, quercetin rhamnoside, quercetin, isorhamnetin, quercetin pentoside, quercetin acetylhexoside, quercetin acetylrutinoside and quercetin pentosylhexoside were calculated according to quercetin-3-glucoside, kaempferol hexoside, kaempferol pentoside, kaempferol rhamnoside, kaempferol pentosylhexoside and kaempferol according to kaempferol-3-rutinoside, apigenin hexoside and genistein according to apigenin, while umbelliferone was expressed as scopoletin equivalents. All analyses have been performed in a duplicate and concentrations of analyzed compounds are expressed as mg 100 g−<sup>1</sup> of dry matter (dm) (N = 4).

#### *2.5. HPLC-UV-VIS/PDA Conditions*

The carotenoids and chlorophylls identification and quantification were performed using Agilent Infinity 1260 system equipped with Agilent 1260 photodiode array detector (PDA; Agilent, Santa Clara, CA, USA) with an automatic injector and Chemstation software (ver. C.01.03).

The separation of carotenoids and chlorophylls was performed using Develosil RP-Aqueus C 30 column (250 × 4.6 mm i.d. 3 µm, Phenomenex, Torrance, CA, USA). The solvent composition and the used gradient conditions were described previously by Castro– Puyana et al. (2017) [27]. The mixture of MeOH:MTBE:water (90:7:3, *v/v/v*) (A) and MeOH:MTBE (10:90, *v/v*) (B) formed the mobile phase. The injection volume was 10 µL and the flow rate was kept at 0.8 mL min−<sup>1</sup> . The chromatogram was monitored by scanning from 240 to 770 nm and the signal intensities detected at 450 nm and 660 nm were used for carotenoid and chlorophyll quantitation. Identification was carried out by comparing retention times and spectral data with those of the authentic standards (α- and β-carotene, chlorophyll *a* and *b*) or in case of unavailability of standards by comparing the absorption spectra reported in the literature [28,29]. Quantifications were made by the external standard calculation, using calibration curves of the standards α-carotene, β-carotene, chlorophyll *a* and chlorophyll *b*. The quantification of individual carotenoid compounds (neoxantine, violaxantine, lutein and its derivatives, derivative of zeaxantine and lycopene) was calculated as β-carotene equivalents and derivatives of chlorophylls as chlorophyll *a* and *b* equivalents using the equation based on the calibration curves, respectively. All determinations have been performed in a duplicate and results are expressed as mg 100 g−<sup>1</sup> dm (N = 4).

#### *2.6. ORAC Determination*

The procedure was based on a previously reported method [30,31] with slight modifications. Briefly, a 96 wells black microplate was prepared containing 150 µL of fluorescein solution (70.30 nM) and 25 µL of blank (75 µM phosphate buffer, pH 7.4), Trolox standard (3.24–130.88 µM) or sample (appropriate diluted) were added. The plate was incubated for 30 min at 37 ◦C. After the first three cycles (representing the baseline signal), AAPH

(240 mM) was injected into each well to initiate the peroxyl radical generation. Fluorescence intensity (excitation at 485 nm and emission at 528 nm) was monitored every 90 sec over a total measurement period of 120 min using an automated plate reader (BMG LABTECH, Offenburg, Germany) and data were analyzed by MARS 2.0 software. The results were expressed as mmol Trolox equivalent (TE) 100 g−<sup>1</sup> of dm. Determinations were carried out in duplicate (N = 4).

#### *2.7. Statistical Analysis*

Statistica ver. 10.0 software (Statsoft Inc., Tulsa, OK, USA) was applied for the statistical analysis. Full factorial randomized design was designated for the experimental part and descriptive statistic was used for the basic data evaluation. Continuous variables (polyphenols, pigments and antioxidant capacity) were analyzed by multifactorial analysis of variance (MANOVA) and marginal mean values were compared with Tukey's HSD test. Relationships between determined compounds and antioxidant capacity were examined by calculated Pearson's correlation coefficients, while possible grouping of the samples according to the examined sources of variations was tested using Principal Component Analysis (PCA). Significance level *p* ≤ 0.05 was assigned for all tests.

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

This study examined the influence of plant part (leaves and stalks), phenological stage (before flowering, during flowering and after flowering) and habitat (Table 1) on the concentrations of polyphenols and pigments in wild nettle grown in Croatia. A total of 84 nettle samples were analyzed, where target compounds (polyphenols and pigments) were extracted using ASE and their identification/quantification was assessed by UPLC-MS/MS (polyphenols) and HPLC-UV-VIS/PDA (pigments). Moreover, obtained extracts were characterized for their antioxidant capacity by the ORAC method.

#### *3.1. Influence of Phenological Stage and Habitat on Polyphenols in Nettle Leaves and Stalks*

Table 2 shows detailed polyphenolic profile and mass spectrometric data obtained by UPLC-MS/MS analysis of nettle leaves and stalks. A total of 41 polyphenolic compounds were identified, belonging to the classes of benzoic, cinnamic and other phenolic acids, flavonols, flavan-3-ols, flavones, isoflavones, flavanones and coumarins (Supplementary files 1 & 2). Among the benzoic acids, compound **35** was identified as gallic acid by comparison of its retention time and mass spectra data with those of an authentic standard. Other benzoic acids were tentatively identified according to their mass fragmentation patterns. Compounds **2** and **14** showed same fragmentation pattern with molecular ion at *m/z* 153 and fragment ion at *m/z* 109, corresponding to the loss of carbon dioxide moiety and implicating the structure of dihydroxybenzoic acids and were therefore according to their polarity tentatively identified as protocatehuic (3,4-dihydroxybenzoic acid) and gentisic acid (1,3-dihydroxybenzoic acid), respectively [32]. Compound **31** showed precursor ion at *m/z* 197 and fragmentation loss of −15 amu corresponding to the loss of methyl radical characteristic for methoxylated phenolic acids and was tentatively identified as syringic acid. Compound **34** showed precursor ion at *m/z* 137 and characteristic fragmentation pattern for deprotonated phenolic acid with loos of −44 amu due to decarboxylation [33] and was assigned as *p*-hydroxybenzoic acid. The composition of benzoic acids in nettle leaves and stalks is in accordance with previous reports [9,14]. Among the cinnamic acids, compounds **12**, **15**, **19**, **25** and **32** were identified using authentic standards as caffeic, chlorogenic, *p*-coumaric, ferulic and sinapic acid, respectively. Compound **21** was presented with precursor ion at *m/z* 147, and fragment ion at *m/z* 103 as a result of decarboxylation and was due to its mass spectra data assigned as cinnamic acid [32]. Compound **16** was identified as quinic acid comparing its spectral data and retention time with those of an authentic standard. The composition of cinnamic acids is in accordance with previous reports by Orˇci´c et al. (2014) [14] and Franciškovi´c et al. (2017) [34] with the exception of cinnamic acid which was not detected in their research, but was reported previously

in composition of nettle leaves by Zekovi´c et al. (2017) [35]. The most numerous class of flavonoid polyphenols identified in nettle were flavonols and their glycosides. Compounds **4**, **8**, **17** and **18** were identified by the authentic standard comparison as kaempferol-3 rutinoside, myricetin, quercetin-3-glucoside and quercetin-3-rutinoside, respectively. Other compounds were tentatively identified according to their mass spectra and characteristic fragmentation patterns reported previously. Among the aglycones, compounds **10**, **24** and **41** were assigned as quercetin, isorhamnetin and kaempferol due to characteristic molecular ion at *m/z* 301, *m/z* 315 and *m/z* 285 [36]. The presence of this aglycones in nettle aerial parts was confirmed previously by Bucar et al. (2006) [37]. Flavonol glycosides lacking authentic standards were tentatively identified according to the characteristic loss of sugar moiety and formation of aglycon fragment ion. Therefore, because of fragment ion at *m/z* 317, compound **3** was assigned as isorhamnetin glycoside. Precursor ion at *m/z* 625 implicated glycosylation with rhamnose (+146 amu) and glucose (+162 amu), so it was assigned as isorhametin rutinoside.

**Table 2.** Mass spectrometric data and identification of polyphenols.



**Table 2.** *Cont.*

\* Identification confirmed using authentic standards.

Its presence in nettle leaves and stalks was reported previously by Pinelli et al. (2008) [6]. Compounds **5**, **28**, **30**, **33** and **36** were identified as quercetin glycosides due to MS/MS ion at *m/z* 303 and were assigned as quercetin rhamnoside, quercetin pentoside, quercetin acetylhexoside, quercetin acetylrutinoside and quercetin pentosylhexoside due to fragmentation losses corresponding to rhamnose (−146 amu), pentose (−132 amu), hexose with acetyl residue (−162 and −42 amu), rutinose with acetyl residue (−308 and −42 amu) and pentose with hexose moiety (−132 and −162 amu) [38]. Previous reports on quercetin glycosides composition in nettle mostly included quercetin glucoside [6,14,34] and quercetin rutinoside [8,14,34], while not reporting the presence of acylated glycosides and diglycosides identified in this study. The latter provides the valuable contribution to detailed insight into nettle polyphenolic profile. Because of the characteristic fragment ion at *m/z* 287 corresponding to the kaempferol aglycon, compounds **6**, **27**, **29** and **39** were assigned as kaempferol hexoside, pentoside, rhamnoside and pentosylhexoside, respectively, due to fragment losses of corresponding sugar moieties. Similar to the previous literature reports on quercetin glycosides, the ones on kaempferol glycosides mostly only include kaempferol rutinoside [6,8] or glucoside [14,34,39], while not reporting the presence of kaempferol pentoside, rhamnoside and pentosylhexoside which are therefore being confirmed here for the first time. All compounds belonging to the class of flavan-3-ols (**23**, **37**, **38** and **40**), namely epigallocatechingallate, epicatechin, catechin and epicatechingallate were identified and confirmed according to the authentic standard. Orˇci´c et al. (2014) [14] identified catechin in nettle stalks, epicatechin was reported by Proestos et al. (2006) [40] in leaves, while there are no available reports on previous identification of epicatechingallate and epigallocatechingallate. Compounds **9** and **22** were assigned as luteolin and apigenin due to molecular ions at *m/z* 287 and *m/z* 271 and confirmed by comparison with standards, while compound **7** was tentatively identified as apigenin hexoside based on fragment ion at *m/z* 271 and fragmentation loss of -162 amu specific for hexose residue. Nencu et al. (2012) [41] reported the polyphenolic composition of nettle leaves including aglycones luteolin and apigenin, which is in accordance with our findings, while literature reports on flavone aglycones are scarce. Compound **20** showed precursor ion at *m/z* 269 and fragment ion at *m/z* 133, corresponding to the previously reported fragmentation mechanism of genistein anion [42], confirmed in nettle leaves extract by Zekovi´c et al. (2017) [35]. Compounds **11**, **13** and **26** were identified by its corresponding authentic standards as naringenin, esculetin and scopoletin, while compound **1** was tentatively assigned as umbelliferone due to molecular ion at *m/z* 161 and fragment ion at *m/z* 133 formed after the loss of one carbon

monoxide molecule [43]. The composition of flavanones and coumarines reported in our study is in accordance with previous literature data [14,34,35].

To examine the influence of phenological stage and habitat on the content of polyphenols in nettle leaves and stalks, identified polyphenols were arranged in corresponding classes, following which their individual concentrations accordingly summarized and subjected to statistical analysis, as shown in Table 3. Total polyphenols grand mean (GM) was 380.90 mg 100 g−<sup>1</sup> dm, among which cinnamic acids were the most abundant group (GM 179.22 mg 100 g−<sup>1</sup> dm), followed by flavonols (GM 134.60 mg 100 g−<sup>1</sup> dm), flavones (GM 24.56 mg 100 g−<sup>1</sup> dm), flavan-3-ols (GM 20.70 mg 100 g−<sup>1</sup> dm) and benzoic acids (GM 10.20 mg 100 g−<sup>1</sup> dm). Coumarins, isoflavones and other acids were present in lower concentrations: GM values 5.31, 3.09 and 2.88 mg 100 g−<sup>1</sup> dm, respectively, while the least represented group of polyphenols were flavanones (GM 0.34 mg 100 g−<sup>1</sup> dm). Moreover, obtained results are in accordance with the results of other authors [6,8,11,14], who reported quite similar phenolic profile in nettle extracts where cinnamic acids accounted for the most of presented total polyphenols.

As can be observed, the plant part, phenological stage and habitat had a significant influence (*p* < 0.01) on amounts of all polyphenols' groups. When comparing amounts of polyphenols between nettle leaves and stalks, it can be seen that leaves accumulated significantly higher concentrations of all polyphenols' groups (Table 3). Otles and Yalcin (2012) [7] also documented higher polyphenols content in wild nettle leaves extracts when compared to stalks extracts, as well as Pinelli et al. (2008) [6] who studied the content of polyphenols in cultivated and wild nettle and reported higher total polyphenols in leaves of both types of nettle (cultivated 7.364 mg g−<sup>1</sup> fw, wild 2.58 mg g−<sup>1</sup> fw) as opposed to nettle stalks (cultivated 3.670 mg g−<sup>1</sup> fw, wild 0.750 mg g−<sup>1</sup> fw).

Same authors documented the abundance of nettle stalks with fibers, consisting of several components of the lignin. However, in study of Orˇci´c et al. (2014) [14], who examined nettle samples picked at three different locations, several identified polyphenols were recorded in higher levels in stalks, but the cinnamic acids presented in their study with chlorogenic acid were also more abundant in leaves.

Considering the phenological stage, it can be noticed that the 1st phenological stage (before flowering) resulted with higher concentrations of all polyphenols, except flavan-3 ols which were significantly higher during the 2nd phenological stage (flowering) (Table 3). Overall, total polyphenols decreased for almost 50% by the 3rd phenological stage. Similar to our results, in two studies of Nencu et al. (2012, 2013) [41,44], it was concluded that the optimal time for nettle leaves harvest was March, since the polyphenols content greatly decreased (over 80%) by June and September, respectively. Authors reported that the total polyphenols decrease is due to the decrease of non-tannin phenols (phenolcarboxylic acids and flavonoids), which are the most important compounds from nettle leaves. This was also confirmed by Roslon et al. (2003) [45] who reported a sudden drop of phenolcarboxylic acids in leaves harvested at the plant flowering stage. Furthermore, the results of Biesiada et al. (2009, 2010) [46,47] and K˝oszegi et al. (2020) [19] also indicated that the beginning of the nettle vegetation period was optimal for harvesting, giving the highest yield of polyphenols, which then decreased by autumn for over 50%. Therefore, in order to obtain extracts with the highest polyphenols content, the optimal time to harvest the aerial parts of the nettle is spring (before the flowering of the plant). It can be assumed that the total polyphenols decrease starting at the flowering stage is a result of the physiological switch from the vegetative to the generative phase and the formation of flowers [48].


**Table 3.** The differences in polyphenols content (mg 100 g<sup>−</sup><sup>1</sup> dm) in wild nettle (*Urtica dioica*L.) due to the plant part, phenological stage and habitat.

C = continental, M = mountain, S = seaside. \* Statistically significant variable at *p*≤0.05. Results are expressed as mean±SE (N = 4). Values with different letters within column are statistically different at *p*≤0.05.

Habitats of wild nettle samples differed according to the climate conditions and could be grouped into three different regions: continental, mountain and seaside (Table 1). As presented in Table 3, habitats significantly (*p* < 0.01) differed regarding polyphenols content, with no uniform pattern regarding individual polyphenolic groups. Thus, Žakanje, belonging to the continental region, was characterized with the highest concentrations of total polyphenols (513.12 ± 1.03 mg 100 g−<sup>1</sup> dm), benzoic (19.39 ± 0.10 mg 100 g−<sup>1</sup> dm) and cinnamic acids (227.10 ± 0.70 mg 100 g−<sup>1</sup> dm), flavan-3-ols (32.13 ± 0.11 mg 100 g−<sup>1</sup> dm), flavones (42.44 ± 0.12 mg 100 g−<sup>1</sup> dm) and isoflavones (5.29 ± 0.05 mg 100 g−<sup>1</sup> dm). Contrarily, Ogulin, situated in mountain areas, was characterized with the highest amounts of other acids (4.62 ± 0.07 mg 100 g−<sup>1</sup> dm), flavonols (182.65 ± 0.38 mg 100 g−<sup>1</sup> dm), flavanones (0.56 ± 0.02 mg 100 g−<sup>1</sup> dm) and coumarins (6.77 ± 0.04 mg 100 g−<sup>1</sup> dm). Moreover, seaside habitats generally showed the lowest presence of all polyphenols. Still, based on total polyphenols content, a difference between seaside samples and ones from other two regions can be observed, where continental and mountain samples showed significantly higher levels of total polyphenols when compared to the samples from seaside zone. This could be explained as a plant's self-defense against oxidative stress caused by lower temperatures. According to Di Virgillo et al. (2015) [1] habitat greatly affects the accumulation of polyphenolic compounds in nettle. Just as in the current study, other authors also confirmed a diversity in nettle polyphenols content in growing areas [7,14].

#### *3.2. Influence of Phenological Stage and Habitat on Pigments in Nettle Leaves and Stalks*

The presence of nettle natural color carriers, carotenoids and chlorophylls was monitored by HPLC analysis, which has detected a total of 13 carotenoids and 9 chlorophylls in wild nettle leaves and stalks, namely neoxanthin and its two derivatives, violaxanthin and its two derivatives, 13′ -*cis*-lutein, lutein 5,6-epoxide, lutein, zeaxanthin, 9′ -*cis*-lutein, α-carotene, β-carotene, chlorophyll *a* and its six derivatives and chlorophyll *b* and its derivative (Figure 1, Supplementary file 1). A similar chlorophylls and carotenoids composition was previously reported [4,11]. For statistical purposes, identified pigments were grouped and analyzed as total carotenoids and total chlorophylls, as well as their sum (total pigments) (Table 4). Total pigments GM was 644.22 mg 100 g−<sup>1</sup> dm, most of which were chlorophylls (GM 611.19 mg 100 g−<sup>1</sup> dm), while carotenoids were less present (GM 33.03 mg 100 g−<sup>1</sup> dm). Other authors also reported higher chlorophylls content in nettle leaves extracts in comparison with the content of carotenoids [9,11,47,49].

As presented in Table 4, all sources of variation significantly (*p* < 0.01) affected both groups of pigments as well as their sum. When comparing the pigments distribution in examined plant parts, abundance in pigments was expectedly higher in leaves since they are major photosynthesis organs [50]. Accordingly, Hojnik et al. (2007) [18] also reported a much higher concentration of chlorophylls in nettle leaves in comparison with stalks (147.1 vs. 16 mg g−<sup>1</sup> extract). Furthermore, determined values for total chlorophylls in leaves were similar to previously reported results by Biesiada et al. (2010) [46], Zeipin, a et al. (2014) [49] and Repaji´c et al. (2020) [11], but were higher than in Ðurovi´c et al.'s (2017) [9] study. Also, the obtained total carotenoids content was in accordance with the values documented in Repaji´c et al.'s (2020) [11] study, but it showed dissimilarity in comparison with the data of other authors [4,9,46,49], probably due to environmental differences.

č ′ ′ α β **Figure 1.** HPLC UV-VIS/PDA detection of pigments in wild nettle leaves (*Urtica dioica* L.) collected from Poreˇc before flowering: (**a**) at 450 nm (1 = violaxanthin derivative 1, 2 = neoxanthin derivative 1, 3 = neoxanthin, 4 = violaxanthin, 5 = violaxanthin derivative 2, 6 = 13′ -*cis*-lutein, 7 = neoxanthin derivative 2, 8 = lutein 5,6-epoxide, 9 = lutein, 10 = zeaxanthin, 11 = 9′ -*cis*-lutein, 12 = α-carotene, 13 = β-carotene); (**b**) at 660 nm (1 = chlorophyll *a* derivative 1, 2 = chlorophyll *a* derivative 2, 3 = chlorophyll *b*, 4 = chlorophyll *b* derivative 1, 5 = chlorophyll *a* derivative 3, 6 = chlorophyll *a* derivative 4, 7 = chlorophyll *a*, 8 = chlorophyll *a* derivative 5, 9 = chlorophyll *a* derivative 6).

− − β μ <sup>−</sup> ć Regarding the phenological stage, the highest amounts of all analyzed pigments were observed during the 2nd stage (flowering), where chlorophylls were the dominant pigments present in almost a 19-fold higher concentration (691.46 mg 100 g−<sup>1</sup> dm) when compared to the amount of carotenoids (36.97 mg 100 g−<sup>1</sup> dm). Similarly, Biesiada et al. (2009) [47] reported increased content of chlorophylls and carotenoids in nettle leaves when harvested in July in comparison with the harvest in May. Additionally, Marchetti et al. (2018) [10] observed that the highest lutein and β-carotene concentrations in nettle leaves occurred during the flowering stage (184 and 6.7 µg g−<sup>1</sup> dm, respectively). Pajevi´c et al. (1999) [51] also determined the maximum levels of chlorophylls and carotenoids in leaves of five alfalfa (*Medicago sativa* L.) genotypes just before and during the flowering stage. These similar patterns can be explained by enhanced production of secondary metabolites, such as plant pigments, during the flowering stage as a plant mechanism for fulfilling important physiological tasks like attracting pollinators [20].

When observing the differences in nettle pigments among the examined habitats, generally samples grown in seaside regions (particularly in the Limski zaljev and Bale habitats) had the highest pigments content. As this area was generally characterized by higher temperatures and lower accumulated precipitation (Table 1), these results are expected since the level of pigments in nettle is influenced by environmental factors, primarily the climate and growing location, where exposure to higher temperatures and more solar energy will result in a higher pigments content [24]. The results of Candido et al.'s (2015) [52] study, in which they examined carotenoid content in buriti palms pulp grown in two different regions (Amazon and Cerrado, Brazil), supported the aforementioned

results. They concluded that a higher content of carotenoids was measured in samples from the Amazon area, characterized by higher temperatures and humidity which prevent photodegradation of fruit pigments.

**Table 4.** The differences in pigments content (mg 100 g−<sup>1</sup> dm) and ORAC values (mmol TE 100 g−<sup>1</sup> dm) in wild nettle (*Urtica dioica* L.) upon plant part, phenological stage and habitat.


C = continental, M = mountain, S = seaside. \* Statistically significant variable at *p* ≤ 0.05. Results are expressed as mean ± SE (N = 4). Values with different letters within column are statistically different at *p* ≤ 0.05.

#### *3.3. Influence of Phenological Stage and Habitat on Antioxidant Capacity in Nettle Leaves and Stalks*

The results of nettle antioxidant capacity measured by the ORAC method are given in Table 4 and Supplementary file 1. ORAC GM was 9.67 mmol TE 100 g−<sup>1</sup> dm. Moreover, the nettle antioxidant capacity was significantly influenced (*p* < 0.01) by all examined sources of variation. Nettle leaves showed higher antioxidant capacity in comparison with stalks (11.96 mmol TE 100 g−<sup>1</sup> dm vs. 7.37 mmol TE 100 g−<sup>1</sup> dm). Similar ORAC values in nettle leaves were recorded in study of Repaji´c et al. (2020) [11], while Ceslova et al. ˇ (2016) [53] obtained the same results by measuring the antioxidant capacity of different nettle parts infusions, where nettle leaves gained higher DPPH levels when compared to stalks. In support, Kırca and Arslan (2008) [54] concluded that leaves and flowers of different examined plants had a higher antioxidant capacity when compared to stalks and seeds.

When observing the influence of phenological stage, the highest ORAC value was observed during flowering, after which it significantly decreased and was the lowest after flowering. Similar to the results of the current study, other authors [19,46] documented that the antioxidant capacity of nettle leaves was higher in the earliest periods (April/May and June/July), after which it decreased (September/October).

Nettle samples showed diversity in antioxidant capacity upon habitat variations. As can be observed, samples from the continental and mountain part were described with the highest ORAC levels as opposed to nettles grown in seaside areas, which were

characterized with the lowest antioxidant capacity levels. These results are in accordance with previously discussed contents of polyphenols and pigments, where a certain grouping of the samples according to the presence of polyphenols and pigments by the growing area is evident. Moreover, calculated correlation coefficients supported this observation, since they showed a strong correlation between ORAC values and cinnamic acids, flavonols and total phenols (Table 5).

**Table 5.** Pearson's correlations between analyzed compounds (mg 100 g−<sup>1</sup> dm) and ORAC values (mmol TE 100 g−<sup>1</sup> dm).


\* *p* ≤ 0.05.

Obtained results clearly demonstrated the importance of the appropriate plant part selection as well as its phenological stage with the presence of the highest bioactive compounds accumulation in order to obtain the maximally enriched product, which will be beneficial for consumers.

#### *3.4. PCA Analysis*

Additionally, in order to examine a possible grouping of the nettle samples according to the applied sources of variations, PCA was carried out and obtained results are presented in Figure 2.

According to the preliminary PCA, a communality value of ≥0.5 described all 14 variables, thus they were all included in the test. The first two components (PC1 and PC2) explained 71.31% of total variance, where PC1 accounted for 53.47% of total variance, while PC2 attributed to 17.84% of total variance. Since PC1 strongly/very strongly negatively correlated (−0.77 ≤ r ≤ −0.96) with benzoic and cinnamic acids, flavonols, flavan-3-ols, flavones, ORAC values and total polyphenols, while PC2 had a strong/very strong correlation with carotenoids, chlorophylls and total pigments (−0.79 ≤ r ≤ −0.81), these variables could be considered as the most discriminating variables.

As can be seen in Figure 2a, separation of the samples clearly occurs based on the plant part. Most of the leaf samples were distributed at negative PC2 values, while all samples of stalks were situated at positive PC2 values. Regarding the phenological stage, a certain grouping appeared between samples from the 1st and 3rd phenological stage, where samples collected before flowering were mainly situated at negative PC1 values and almost all of the post-flowering samples were located at the positive PC1 values (Figure 2b). a partial grouping of nettle samples is visible in Figure 2c based on the growing region, where the most of separation can be seen to be present between continental and seaside samples, although this did not completely occur.

**Figure 2.** Distribution of wild nettle samples in two-dimensional coordinate system defined by the first two principal components (PC1 and PC2) according to the (**a**) plant part; (**b**) phenological stage; (**c**) growing region (1 = isoflavones, 2 = flavanones, 3 = flavones, 4 = benzoic acids, 5 = cinnamic acids, 6 = total polyphenols, 7 = flavonols, 8 = flavan-3-ols, 9 = other acids, 10 = coumarins, 11 = ORAC, 12 = carotenoids, 13 = chlorophylls, 14 = total pigments).

#### **4. Conclusions**

The current study confirmed the abundance of wild nettle with diverse bioactive molecules such as low molecular weight polyphenols and pigments, where 41 phenolic compounds, 13 carotenoids and 9 chlorophylls were documented. By using applied extraction conditions, cinnamic acids and flavonols were found to be the dominant classes of identified polyphenols (33.10–519.81 mg 100 g−<sup>1</sup> dm and 57.44–383.25 mg 100 g−<sup>1</sup> dm, respectively), while chlorophylls were the most abundant natural pigments (4.26–1934.38 mg 100 g−<sup>1</sup> dm). Moreover, the ORAC values of obtained nettle extracts ranged from 3.05 to 19.83 mmol TE 100 g−<sup>1</sup> dm. However, in order to obtain high valuable wild nettle extracts that are abundant in natural antioxidants, it is of the utmost importance to select appropriate plant parts as well as an appropriate harvest time. Obtained results evidenced that the highest levels of nettle bioactives accompanied by high antioxidant capacity were present in leaves, which should be collected during the early phenological period (before and at the flowering stage). Moreover, the amounts of wild nettle polyphenols and pigments greatly differed based on the natural habitat, as samples from the seaside region were characterized with elevated accumulation of pigments, while higher polyphenols amounts were present in habitats located in continental and mountain areas. This research will surely contribute to the selection of plant part and phenological stage for nettle optimal harvest, as well as to designate nettle natural habitats that have been shown to be a source of valuable plant material. These findings present the basis for the production of nettle seedlings with high bioactives content, which could further be used in the production of liquid and dry extracts. Furthermore, they showed the importance of a multidisciplinary approach for the selection of a plant part as well as its phenological stage in order to provide highly enriched products intended for the benefit of consumers.

In addition, besides low molecular weight polyphenols and pigments covered by this research, future studies could also include other beneficial compounds present in nettle such as oligomers and polymers as well as sterols, to provide a full insight into the nettle's bioactive potential.

**Supplementary Materials:** The following figures and tables are available online at https://www. mdpi.com/2304-8158/10/1/190/s1, (file 1) Figure S1: LC-MS/MS chromatogram in dMRM acquisition from the extract of wild nettle leaves (*Urtica dioica* L.) collected from Poreˇc before flowering, (file 2) Tables S1–S3: Concentrations of individual compounds and ORAC values in nettle (*Urtica dioica* L.) samples.

**Author Contributions:** Conceptualization, M.R.; Data curation, M.R., Z.Z., S.P. and I.E.G.; Formal analysis, M.R., E.C., Z.Z. and S.P.; Methodology, M.R. and I.E.G.; Project administration, V.D.-U.; Resources, S.R. and I.P.; Supervision, M.R. and V.D.-U.; Visualization, M.R.; Writing—original draft, M.R., E.C. and I.E.G.; Writing—review & editing, M.R., S.P., I.E.G., S.R., I.P. and V.D.-U. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Croatian Science Foundation project, grant number IP-01-2018-4924.

**Acknowledgments:** Authors wish to thank Valentina Kruk from mag.ing.techn.aliment for her assistance during experimental work and the Meteorological and Hydrological Institute of Croatia for providing the meteorological data.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


### *Article* **Investigation of Phenolic Composition and Anticancer Properties of Ethanolic Extracts of Japanese Quince Leaves**

**Vaidotas Zvikas <sup>1</sup> , Ieva Urbanaviciute <sup>2</sup> , Rasa Bernotiene <sup>3</sup> , Deimante Kulakauskiene <sup>1</sup> , Urte Morkunaite <sup>1</sup> , Zbigniev Balion 1,3, Daiva Majiene <sup>3</sup> , Mindaugas Liaudanskas 1,4, Pranas Viskelis 1,2 , Aiste Jekabsone 1,5 and Valdas Jakstas 1,4,\***


**Abstract:** Glioblastoma multiforme is an aggressive and invasive disease with no efficient therapy available, and there is a great need for finding alternative treatment strategies. This study aimed to investigate anticancer activity of the extracts of the Japanese quince (JQ) cultivars 'Darius', 'Rondo', and 'Rasa' leaf extracts on glioblastoma C6 and HROG36 cells. As identified by ultra high performance liquid chromatography electrospray ionization tandem mass spectrometry, the extracts contained three prevailing groups of phenols: hydroxycinnamic acid derivatives; flavan-3-ols; and flavonols. Sixteen phenols were detected; the predominant compound was chlorogenic acid. The sum of detected phenols varied significantly between the cultivars ranging from 9322 µg/g ('Rondo') to 17,048 µg/g DW ('Darius'). Incubation with the extracts decreased the viability of glioblastoma HROG36 cells with an efficiency similar to temozolomide, a drug used for glioblastoma treatment. In the case of C6 glioblastoma cells, the extracts were even more efficient than temozolomide. Interestingly, primary cerebellar neuronal-glial cells were significantly less sensitive to the extracts compared to the cancer cell lines. The results showed that JQ leaf ethanol extracts are rich in phenolic compounds, can efficiently reduce glioblastoma cell viability while preserving non-cancerous cells, and are worth further investigations as potential anticancer drugs.

**Keywords:** *Chaenomeles japonica* leaves; phenolic compounds; glioblastoma; anticancer activity

#### **1. Introduction**

Japanese quince (*Chaenomeles japonica* (Thunb.) Lindl. ex Spach), a representative of the *Rosaceae* Juss. family, has already been known in oriental folk medicine for about 3000 years [1]. This plant is a great source of secondary metabolites possessing various biological effects including anticancer activity [2,3]. For example, quince extracts have high amounts of triterpenes (such as ursolic and oleanolic acids) that are reported to decrease the viability of colon, breast, melanoma, lung, hepatic carcinoma, and other cancer cell types [4–7]. Furthermore, Japanese quince fruit extracts are rich in phenolic compounds, mostly flavonoids [8], that are also known for preventive and therapeutic anticancer potential [9]. Procyanidins and flavanols from Japanese quince fruits induce apoptosis and suppress invasiveness in human colon, prostate, and breast cancer cell cultures [10–12]. A recent study revealed that the extract of Japanese quince leaves reduces viability of

**Citation:** Zvikas, V.; Urbanaviciute, I.; Bernotiene, R.; Kulakauskiene, D.; Morkunaite, U.; Balion, Z.; Majiene, D.; Liaudanskas, M.; Viskelis, P.; Jekabsone, A.; et al. Investigation of Phenolic Composition and Anticancer Properties of Ethanolic Extracts of Japanese Quince Leaves. *Foods* **2021**, *10*, 18. https://dx.doi.org/10.3390/foods 10010018

Received: 30 November 2020 Accepted: 20 December 2020 Published: 23 December 2020

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 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 (https://creativecommons.org/ licenses/by/4.0/).

colon cancer cells SW-480 and HT-29 to a greater extent compared to normal intestinal cells CCD-18 Co and CCD 841 CoN [13]. Such results encouraged us to investigate quince leaf extract efficiency on other cancer cell types.

Glioblastoma multiforme is one of the most aggressive and invasive cancerous diseases, and there are no efficient treatment options available [14]. The most common therapy is temozolomide, however the treatment is accompanied by severe side effects and the efficiency is poor [14]. Some promising results are achieved by applying plant-derived anticancer substances [15]. Therefore, it is important to continue investigating the new plant sources in order to find more efficient treatment or therapy complement for glioblastoma. Our previous research revealed that the leaves of three Japanese quince cultivars 'Darius', 'Rondo', and 'Rasa' are rich in phenols and triterpenes suggesting that the extracts might possess anticancer activity [16]. The current study aimed to perform a broader phenol analysis of the leaves of the same cultivars, and to investigate the effect of the extracts on the viability of glioblastoma HROG36 and C6 cells. In addition, to predict the potential level of cytotoxicity on healthy brain tissue, the effect of the extracts on viability of primary non-cancerous cultured brain cells was evaluated.

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

#### *2.1. Chemicals*

All the solvents, reagents, and standards used were of analytical grade. The following substances were used in the study: Ethanol 96% (*v*/*v*) (AB Strumbras, Kaunas, Lithuania), procyanidin C1, procyanidin B2, quercetin, hyperoside, avicularin, quercitrin, kaempherol 3-O-glucoside, luteolin 7-O-glucoside, phloridzin, formic acid, acetonitrile, (+)-catechin, (-)-epicatechin, rutin, isoquercitrin, chlorogenic acid, p-coumaric acid, caffeic acid, hydrochloric acid, Hoechst33342, propidium iodide, glucose, temozolomide, DMSO and KCl (Sigma-Aldrich, Steinheim, Germany), Dulbecco's Modified Eagle Medium (DMEM) with Glutamax, foetal bovine serum, penicillin-streptomycin, Versene solution, antibioticantimycotic solution (Anti-Anti) were of Gibco brand and purchased from Thermo Fisher Scientific, Waltham, MA, USA. During the study, we used purified de-ionized water prepared with the Milli–Q® (Millipore, Bedford, MA, USA) water purification system.

#### *2.2. Plant Material and Extract Preparation*

Japanese quince (*C. japonica*) leaves were collected in September 2018, after ripe fruits were harvested, from the garden of the Institute of Horticulture, Lithuanian Research Centre for Agriculture and Forestry, Babtai (55◦60′ N, 23◦48′ E). The leaves of each cultivar were collected from five shrubs, and frozen (at −40 ◦C) in a freezer with air circulation, and then lyophilized with a sublimator Zirbus 3 × 4 × 5 (ZIRBUS technology GmbH, Bad Grund, Germany), at a pressure of 0.01 mbar (temperature of condenser −85 ◦C) for 24 h. The lyophilized leaves were grounded to a fine powder with a knife mill GM (Retsch GmbH, Haan, Germany). Powdered leaf sample of each cultivar (2.5 g) was mixed with 50 mL of 40% ethanol, and extracted with ultrasonic bath Sonorex Digital 10 P (Bandelin Electronic GmbH & Co. KG, Berlin, Germany) for 40 min, at 60 ◦C, using 480 W ultrasonic power. The extracted samples were centrifuged and then filtered through filter paper (Watman no. 1). The Japanese quince ethanolic extracts (5 g/100 mL) were kept in a freezer (at −40 ◦C) in hermetically sealed containers for one week until further tests.

#### *2.3. Evaluation of Phenolic Compound Composition (UPLC-ESI-MS/MS Conditions)*

The variability in the qualitative and quantitative content of phenolic compounds in Japanese quince leaf samples was evaluated by applying validated UPLC-ESI-MS/MS method [17]. Samples were analyzed with Acquity H-class UPLC system (Waters Corporation, Milford, MA, USA) coupled with triple quadrupole tandem mass spectrometer (Xevo, Waters Corporation, Milford, MA, USA). To obtain MS/MS data an electrospray ionization source (ESI) was used. Compounds of interest were separated with YMC Triart C18 (100 × 2.0 mm; 1.9 µm) column (YMC Europe Gmbh, Dislanken, Germany). Constant

temperature of 40 ◦C and flow rate of 0.5 mL·min−<sup>1</sup> were maintained during analysis. Mobile phase consisted of 0.1% formic acid solution in water (solvent A) and acetonitrile (solvent B). Gradient profile was applied with following proportions of solvent A: Initially 95% for 1 min followed by linear increase to 70% over 4 min; 50% over next 3 min and to 95% over last 2 min. Analysis was performed in negative electrospray ionization mode. Capillary voltage was set to negative 2 kV. Temperature in ion source was maintained at 150 ◦C. Nitrogen gas temperature was set to 400 ◦C and flow rate to 700 L·h −1 . Cone gas flow rate was set to 20 L·h −1 . Each compound of interest had a specific collision energy and cone voltage selected. The selected mass spectrometry parameters for this method are presented in Table 1. The validation characteristics of the developed method are presented as supplementary Table S1.


**Table 1.** Mass spectrometry parameters for the analysis of phenolic compounds.

#### *2.4. C6 and HROG36 Cell Culture*

The C6 and HROG36 cell lines were purchased from the Cell Lines Service GmbH (Germany). The cells suspended in DMEM with 10% of foetal bovine serum, 100 U/mL penicillin and streptomycin, seeded in 75 cm<sup>2</sup> flasks, and incubated at 37 ◦C, with 5% CO<sup>2</sup> and saturated humidity. The cells were reseeded to new flasks every 3 days. Twenty-four hours prior to the treatment with quince leaf extracts, the cells were transferred to 96 well plates (VWR) at density of 0.2 × 10<sup>6</sup> cells/cm<sup>2</sup> .

#### *2.5. Primary Neuronal-Glial Cell Culture*

Primary rat cerebellar-glial cell culture was prepared as described previously [18]. Briefly, the cerebella were isolated, minced, and triturated in Versene solution (1:5000) to a single-cell suspension. The suspension was centrifuged at 270× *g* for 5 min and resuspended in DMEM with Glutamax supplemented with 5% horse serum, 5% foetal calf serum, 38 mM glucose, 25 mM KCl, and antibiotic-antimycotic solution. The cells were plated at a density of 0.25 × 10<sup>6</sup> cells/cm<sup>2</sup> in 96-well plates (VWR) coated with 0.0001% poly-L-lysine and kept in a humidified incubator containing 5% CO<sup>2</sup> at 37 ◦C. The cultures were subjected to treatment after 7 days in vitro.

#### *2.6. Treatments of the Cells with Quince Leaf Extracts*

HROG36, C6, and primary cerebellar neuronal glial cells were treated with 0.88, 1.25, 1.63, 2.00, 2.38, 2.75, 3.13 and 3.75 mg/mL ethanolic extracts made from leaves of Japanese quince cultivars 'Rasa', 'Darius' or 'Rondo' for 24 h. The controls with the same volume of the solvent (ethanol) were made in parallel. In addition, the C6 and HROG36 cells were treated with temozolomide concentration range from 0.02 to 4.85 mg/mL, chlorogenic acid (range 5–500 g/mL), epicatechin (5–300 g/mL), hyperoside (5–333 g/mL), and quercitrin (2–220 g/mL). After treatment, the cells were subjected to viability evaluation.

#### *2.7. Evaluation of Cellular Viability*

Viability of C6, HROG36, and primary cerebellar neuronal glial cells after treatments was evaluated according to metabolic activity by means of PrestoBlue™ Cell Viability Reagent (Thermo Fisher Scientific). The fluorescence of resorufin produced after PrestoBlue reagent cleavage was measured in a plate reader Infinite M Plex (Tecan Austria, Salzburg, Austria) at excitation and emission wavelengths of 560 and 590 nm, respectively. The results were expressed as percentage of the untreated control fluorescence level.

In addition, C6 and primary cerebellar neuronal-glial cells were evaluated for necrosis by double-staining with Hoechst 33,342 (15 µg/mL, Merck) and propidium iodide (PI; 5 µg/mL, Merck). After 15 min incubation with the dyes in dark at room temperature, the nuclear fluorescence was assessed under fluorescent microscope OLYMPUS IX71SIF-3 (Olympus Corporation, Tokyo, Japan). Hoechst33342-only-positive nuclei exhibiting blue fluorescence were considered viable, and Hoechst3334-plus-PI-positive nuclei stained magenta (because of blue and red signal overlay) were identified as necrotic.

#### *2.8. Statistical Analysis*

The phenolic compound content of each cultivar was expressed as means ± SD (standard deviation) of three replicates. The significant differences (*p* ≤ 0.05) between means were evaluated using Tukey's HSD (Honest Significant Difference test). Cellular viability and metabolic activity results are presented as means ± standard deviation of 5 experimental repeats, each of 3 technical repeats. The data are expressed as percentage of the untreated control. Statistical analysis was performed by one-way analysis of variance (ANOVA) with the Dunnett's post-test by SigmaPlot 13.0 software (Systat Software Inc., Surrey, UK). A value of *p* < 0.05 was taken as the level of significance. EC<sup>50</sup> was calculated by SigmaPlot 13.0 (Systat Software Inc., Surrey, UK) software by means of four-parameter logistic function. Correlations were analyzed by Microsoft Office Excel 2010 (Microsoft, Redmond, WA, USA) software Correlation function.

#### **3. Results**

#### *3.1. Phenolic Compound Composition of Japanese Quince Leaves*

The sum of detected phenols varied significantly between cultivars, the highest amount was found in 'Darius', and the lowest in 'Rondo' leaves (Table 2). Sixteen phenolic compounds were identified in the leaves of 'Rondo', while 15 in 'Darius', and 14 in 'Rasa'. The majority of the identified phenols belong to three groups: Hydroxycinnamic acid derivatives, flavonols, and flavan-3-ols. There were also others, such as flavone luteolin 7-O-glucoside, and dihydrochalcone phloridzin. Total amount of hydroxycinnamic acid derivatives ranged from 5533 ('Darius') to 5839 ('Rasa') µg/g, and consisted of chlorogenic acid, p-coumaric acid, and caffeic acid. The flavan-3-ol group members were (-)-epicatechin, procyanidin B2, procyanidin C1, and (+)-catechin, and their total amount ranged from 700.4 ('Rondo') to 6426 ('Darius') µg/g. The flavonols found in the extracts were rutin, isoquercitrin, avicularin, kaempferol 3-O-glucoside, hyperoside, and quercetin. The total amount of the flavonols ranged from 2506 ('Rondo') to 4872 ('Darius') µg/g. The predominant group of phenols in 'Rondo' and 'Rasa' was the hydroxycinnamic acids, but in 'Darius' was flavan-3-ols. However, the total amount of hydroxycinnamic acids did not differ between the cultivars.


**Table 2.** Quantitative composition of phenols in Japanese quince leaves, µg/g DW.

Value is average ± SD (*n* = 3); Different letters in the same line indicate a statistically significant difference (*p* ≤ 0.05); DW: dry weight; ND: not detected.

> The total flavan-3-ol levels varied significantly between the cultivars. The highest amount was detected in 'Darius' leaves (6426 ± 145 µg/g), while 'Rasa' and 'Rondo' had around two and nine-fold less of the flavan-3-ols (3652 ± 73.6 µg/g and 700.4 ± 10.7 µg/g, respectively).

> The total amount of flavonols was also significantly different between the cultivars. In addition, the distribution of individual compounds of this phenol group was different, too. The leaves of 'Darius' had significantly more isoquercitrin and quercitrin, 'Rondo' had more kaempferol 3-O-glucoside and luteolin 7-O-glucoside, and 'Rasa' had higher rutin amount. The main phenolic glycosides found in the extracts were isoquercitrin, hyperoside, and quercitrin.

#### *3.2. The Effect of Quince Leaf Extracts on Viability of C6 and HROG36 Glioblastoma Cells*

Next in the study, we evaluated the effect of the extracts from leaves of Japanese quince cultivars 'Darius', 'Rasa', and 'Rondo' on metabolic activity of rat C6 and human HROG36 glioblastoma cells (Figure 1).

**Figure 1.** The effect of the extracts from leaves of different Japanese quince cultivars (**a**,**b**) and some phenolic compounds found in the extracts (**c**,**d**) on viability of glioblastoma C6 (**a**,**c**) and HROG36 (**b**,**d**) cells evaluated by metabolic activity assay with PrestoBlue reagent. For Ethanol, the concentrations were the same as in other samples at the indicated concentration point, i.e., 35, 50, 65, 80, 95, 110, 125, and 150 L/mL. In both (**a**,**c**), \* indicates significant difference compared with ethanol-only treated samples; in (**a**), ˆ significant difference compared with temozolomide; in (**c**), ˆ significant difference compared with Epicatechin, + with Chlorogenic acid, # with Hyperoside. In (**b**), all extract-treated samples were statistically significantly different from Ethanol starting from concentration 0.25 mg/mL, and Temozolomide–starting from 0.5 mg/mL. in (**d**), all the samples were statistically significantly different from Ethanol starting from the concentration of 7 g/mL, and there was statistically significant difference between Epicatechin and other samples at 7 and 9 g/mL. The level of significance *p* < 0.05.

After 24-h application of 1.25 mg/mL quince leaf extracts, the viability of C6 cells (assessed as cellular metabolic activity by PrestoBlue assay) was significantly decreased compared to ethanol control. The viability was by 14%, 11%, and 13% lower after treatment with 'Rondo', 'Rasa', and 'Darius', respectively (Figure 1a). The metabolic activity of the cells continued to decrease with increase in the concentration of the extracts and reached 8% of control in 3.125 mg/mL 'Rondo'- and 'Darius'-treated samples, and 10% in 3.125 mg/mL 'Rasa"-treated samples. The difference between ethanol control and the 3.125 mg/mL extract-treated samples were 43%, 41%, and 43% for 'Rondo', 'Rasa', and 'Darius', respectively. The metabolic activity of C6 cells was more sensitive to the extracts compared to the effect of temozolomide, the drug used for glioblastoma treatment. The metabolic activity of C6 cells treated with extracts at 1.25 mg/mL and higher concentration was significantly lower than in the C6 samples treated by the same concentrations of temozolomide. This was also reflected in calculated EC50; the values of 'Rondo', 'Rasa', and 'Darius' were 71%, 67%, and 74% lower compared to the EC<sup>50</sup> value of temozolomide for C6 cells (Table 3).


**Table 3.** Calculated EC<sup>50</sup> of extracts from leaves of quince cultivars 'Rondo', 'Rasa', and 'Darius' and of some phenolic compounds found in the extracts for C6 and HROG36 cells.

> Human glioblastoma cells HROG36 were more sensitive to quince leaf extract treatment compared to C6 cells (Figure 1b). 'Rondo', 'Rasa', and 'Darius' applied at 0.25 mg/mL significantly decreased metabolic activity of HROG36 cells compared to ethanol control by 37%, 24%, and 34%, respectively. After treatment with 0.75 mg/mL extracts, metabolic activity of HROG36 cells was less than 5% and about 91% lower compared to the ethanol control. The effect of temozolomide on HROG36 cell metabolic activity was similar to that of the extracts, and there were no statistically significant differences detected. However, the EC<sup>50</sup> value of temozolomide calculated from the average titration data was by 157.3–162.0 µg/mL higher compared to the EC<sup>50</sup> of the extracts (Table 3).

> For the next step in the study, we have investigated the toxicity of the phenolic compounds identified in the quince leaf extracts on C6 and HROG36 cells. Chlorogenic acid was selected as a representative of hydroxycinnamic acids, hyperoside and quercitrin from flavonols, and epicatechin from flavan-3-ols. C6 cells were most sensitive to quercitrin and chlorogenic acid, as presented in the titration curves in Figure 1c and EC<sup>50</sup> values in Table 3. The next least toxic compound was hyperoside, and the least toxic was epicatechin. Similarly to the case of extract treatment, HROG36 cells were more sensitive to the phenolic compounds compared to the C6 cells (Figure 1c,d). After treatment with 23 µg/mL of each compound, the metabolic activity of HROG36 cells was 8% (for quercitrin)–26% (for epicatechin) of the untreated control. For comparison, after similar 30 µg/mL treatment in C6 cell samples, the metabolic activity was either unchanged (chlorogenic acid, epicatechin), or decreased only to 90% (hyperoside) and 75% (quercitrin) of untreated control. The most toxic for HROG36 cells were quercitrin and hyperoside, although chlorogenic acid was also very close to that level. A slightly lower toxicity was caused by epicatechin; there was statistically significant difference between epicatechin and other investigated phenolic compounds at 7 and 9 g/mL.

> The potential input of each group of phenolic compounds and some individual phenols in the extracts was estimated by correlation analysis (Table 4).


**Table 4.** The values of coefficient for correlation between viability (metabolic activity) of C6 or HROG36 cells and amount of phenolic compounds in the extracts from leaves of quince cultivars 'Rondo', 'Rasa', and 'Darius'.

> Strong negative correlation with r value close to −1 was found between viability level of both glioblastoma cell types and amounts of hydroxycinnamic acids and chlorogenic acid. In addition, strong negative correlation was between metabolic activity of C6 and hyperoside, and between metabolic activity of HROG36 and quercitrin. Moderate negative correlation was between C6 viability and the amount of quercitrin, and low negative between HROG36 viability and total contents of flavonols and phenols. The analysis allows to suggest that the toxicity of the extracts was most likely mediated by chlorogenic acid.

> Metabolic activity assays such as PrestoBlue reflects not only changes in cell viability, but also differences in proliferation rate and metabolic disturbances. Therefore, we have

additionally investigated viability of C6 cells after quince leaf extract treatments by double nuclear fluorescence staining that allows to detect necrotic cells with lost membrane integrity (Figure 2).

**Figure 2.** The effect of the extracts from leaves of different Japanese quince cultivars on viability of glioblastoma C6 cells. (**a**–**g**) Characteristic images of Hoechst/propidium iodide staining of the C6 cell nuclei after extract treatments; (**a**) samples treated with 125 µL/mL of the solvent ethanol (amount corresponds to 1.88 mg/mL extract treatment); (**b**–**d**) treated with 1.19 mg/mL extracts from 'Rondo', 'Rasa', and 'Darius' cultivars, respectively; (**e**–**g**) samples, after treatment with 1.56 mg/mL extracts from 'Rondo', 'Rasa', and 'Darius' cultivars, respectively; (**h**) quantitative summary of viability data presented as averages with standard deviation, \* indicates significant difference compared with ethanol-only treated samples, *p* < 0.05.

The extracts did not significantly affect C6 cell viability up to the concentration of 1.19 mg/mL (Figure 2b–d,h). Treatment with 1.19 mg/mL extract from leaves of 'Darius' induced small yet significant decrease in percentage of viable C6 cells (Figure 2d,h). The average level of viable cells in 'Darius' extract-treated samples was by 14% lower compared with samples treated with the same amount of ethanol. Further increase in extract concentration to 1.38 mg/mL caused a remarkable drop in C6 viability in all three cultivar groups (Figure 2h). The average level of viable cells decreased by 93%, 78%, and 88% after treatment with extracts from 'Rondo', 'Rasa', and 'Darius', respectively. After treatment with 1.88 mg/mL extracts, the number of viable cells in C6 samples was close to 'zero' in all three cultivar groups. There was no significant decrease in C6 cell viability, observed after treatment with ethanol up to 1.88 mg/mL. Calculated levels of EC<sup>50</sup> from double

nuclear staining experiments were 1.26 mg/mL, 13.0 mg/mL, and 1.26 mg/mL for 'Rondo', 'Rasa', and 'Darius', respectively; the values slightly lower yet similar to those revealed by metabolic activity assay.

#### *3.3. The Effect of Quince Leaf Extracts on Viability of Primary Non-Cancerous Brain Cells*

All the Japanese quince leaf extracts investigated in the study were toxic to human rat glioblastoma C6 cells at concentrations equal to or higher than 1.38 mg/mL. Human glioblastoma HROG36 cells were even more sensitive to the treatments. The next step in this study was to investigate whether primary non-cancerous brain cells have similar susceptibility to the same extract treatment (Figure 3).

**Figure 3.** The effect of the extracts from leaves of different Japanese quince cultivars on viability of cultivated primary rat cerebellar neuronal-glial cells. (**a**–**e**) Representative images of nuclei stained with Hoechst33342 and propidium iodide after extract treatments: (**a**) Untreated control, (**b**) sample treated with 125 µL/mL of the solvent ethanol; (**c**–**e**) samples treated with 1.19 mg/mL extracts from 'Rondo', 'Rasa', and 'Darius' cultivars, respectively; (**f**) quantitative summary of viability data presented as averages with standard deviation, \* indicates significant difference compared with the samples treated with the same amount of ethanol.

For these experiments, rat cerebellar neuronal-glial cell cultures consisting of approximately of 81 ± 4% granule neurons, 14 ± 3% astrocytes, and 6 ± 2% microglial cells [18]

were used. The extracts up to the concentration of 1.0 mg/mL did not cause significant decrease in viability of primary cerebellar cells compared to the respective ethanol control (Figure 3f). However, increase in concentration up to 1.19 mg/mL caused viability drop by 13% ('Rondo'), 9% ('Rasa'), and 10% ('Darius') compared with the respective ethanol treatment. Further increase in extract concentration continued to lower the number of viable cells, and after treatment with 1.88 mg/mL extracts, the average numbers such cells in primary neuronal-glial cultures were 38%, 47%, and 43% for 'Rondo', 'Rasa', and 'Darius', respectively. In the case of C6 glioblastoma cells, the extracts applied at concentration 1.56 mg/mL induced 80% and higher loss of viable cells, and there were almost no viable cells left after treatment with 1.88 mg/mL extracts (Figure 2). In primary brain cell cultures, the level of viable cells remained higher than 50% after treatment with 1.56 mg/mL extracts and close to 50% after treatment with 1.88 mg/mL extracts. Thus, primary neuronal-glial cells from rat cerebella were less sensitive to the toxic effect of the 1.56–1.88 mg/mL extracts from quince leaves compared to the glioblastoma C6 cells. This was also confirmed by the calculated EC<sup>50</sup> of the extracts for primary cerebellar cells; the values were 1.58 mg/mL, 1.72 mg/mL, and 1.64 mg/mL for 'Rondo', 'Rasa', and 'Darius', respectively, and they were higher than the values for C6 cells calculated from the double nuclear staining assay data. However, primary neuronal-glial cells were more sensitive to the solvent ethanol compared to C6 cell line. There was significant decrease in primary cerebellar cell viability observed after treatment with 80 µL/mL ethanol; the average level of viable cells after this treatment was by 25% lower compared to the untreated control. After treatment with 150 µL/mL ethanol, the percentage of viable cells were by 30% lower than in control samples.

#### **4. Discussion**

Japanese quince leaf extract consisted of three major phenol groups: Hydroxycinnamic acids, flavan-3-ols, and flavonols. However, distribution of these groups between cultivars was significantly different. For example, in 'Rondo', more than half of total phenols consisted of hydroxycinnamic acid derivatives (62.6%), while 'Darius' and 'Rasa' contained only 32.5%, and 43% of these compounds, respectively. Similar amounts (from 42.90% to 50.90%) of hydroxycinnamic acids in quince leaves were found in a recent study of Chojnacka and co-authors [13]. The predominant compound of this group in all cultivars was chlorogenic acid, and this is in agreement with other studies [13,19]. The amount of flavonols in all cultivars was around 30% of total phenols. The majority of the flavonols consisted of isoquercitrin, hyperoside, and quercitrin, all the three are known for anti-cancer properties [20–22]. Saccharide moiety of the compounds mediates the toxicity interacting with membranes of cancer cells and promoting active uptake of the compounds [23,24]. The distribution of flavan-3-ols between cultivars was most diverse and varied from 7.5% ('Rondo') to 37.7% ('Darius').

In the present study, we have found that ethanolic quince leave extracts were cytotoxic similarly as (in the case of human glioblastoma HROG36 cells) or even more than (in rat C6 cell case) temozolomide, the drug used for glioblastoma treatment in the clinical practice. The viability of both investigated cell types had strong negative correlation with the amount of hydroxycinnamic acid derivatives and chlorogenic acid, and the cytotoxicity of the chlorogenic acid was also demonstrated in both C6 and HROG36 cell cultures. Although other investigated compounds were also toxic to the cells, especially flavonols quercitrin and hyperoside, the amounts of the compounds present in the extracts were far too small to mediate the toxicity for C6 cells. However, for HROG36 cells, the flavonols could contribute to the toxicity of chlorogenic acid in the extracts because the EC<sup>50</sup> values of the compounds for the cells were close to the levels found in the extracts. The anticancer activity of hydroxycinnamic acids is also reported by others. These compounds promote apoptosis, arrest cell cycle, and prevent metastasis of different types of breast, lung, colon, gastric, liver, pancreatic, and prostate cancer cells [25–27]. Ekbatan and colleagues have shown that chlorogenic acid and its metabolites caffeic acid, 3-phenyl propionic acid, and benzoic acid cause cell cycle arrest and apoptosis of colon cancer cells Caco 2 [28]. Another group of scientists has demonstrated that chlorogenic acid disrupts cytoskeleton organization and mTORC2 signalling of both adenocarcinomic human alveolar basal epithelial cells A549 and human hepatocyte carcinoma HepG2 [29]. The above-mentioned findings are in line with another extensive study performed by Huang and co-workers, who investigated anticancer activity of chlorogenic acid on human cancerous lung, liver, kidney, colon, and brain cells including human glioblastoma lines U87MG and M059J, rat C6, and mouse G422 [30]. The study revealed that chlorogenic acid promotes all cancer cell (including glioblastoma) cycle arrest and differentiation to maturation phenotype via miR-17 family downregulation, p21 upregulation and mitochondrial suppression. The efficiency of chlorogenic acid was comparable to that of temozolomide.

Comparison of viability of primary rat non-cancerous brain cells and rat glioblastoma C6 cells evaluated by double nuclear staining revealed the significantly higher sensitivity of glioblastoma for the quince leave extracts. Although the EC<sup>50</sup> values of the extracts for the non-cancerous cells were similar to those calculated for C6 cells from the metabolic PrestoBlue evaluation data, to our opinion, it would not be relevant to compare the data obtained from different viability assessment assays. PrestoBlue assay monitors the rate of cellular metabolic activity, which is proportional to the number of viable cells. However, the metabolic activity might be directly influenced by the investigated compound without causing cell death, e.g., chlorogenic acid, the main component of the quince leaf extract, is reported to decrease mitochondrial activity of glioblastoma [30]. The number of viable cells might be lower not, or not only, due to the increase in cell death, but also due to the suppression of proliferation, because hydroxycinnamic acids present in the extracts might induce cell cycle arrest [26]. Thus, the metabolic assay is different from the evaluation of viability by double nuclear staining which gives information about percentage of necrotic and viable cells without sensing the metabolic activity or proliferation rate of them.

Similar finding about lower susceptibility to quince leaf extract of non-cancerous cells compared to cancer cells was recently described by Chojnacka and co-authors [13]. Such higher sensitivity of cancer cells to bioactive compounds of quince leaf extracts could be related to specific biology of these cells. Usually, cancer cells have higher proliferation rate, migration capacity and altered energy metabolism compared to the surrounding noncancerous cells [31,32]. Such a difference opens a niche for selective targeting of cancer cells with lower risk to destroy healthy non-cancerous cells. Ethanolic extracts from Japanese quince leaves have several compounds that interfere with cancer cell-specific pathways related to proliferation, migration and energy metabolism [4,29,30]. This might at least partially explain cancer cell-selective cytotoxicity of quince leaf extracts.

Thinking about quince leaf extracts as complementary therapy for glioblastoma, it is important to evaluate the ability of the extract compounds to cross the blood-brain barrier (BBB). One of the best studied pathways for plant phenolic metabolites to cross BBB is passive permeation, however active transport might be also possible. Lee and co-authors have found that intraperitoneally administered chlorogenic acid ameliorates brain damage and oedema after cerebral ischaemia in rats [33]. In another study, chlorogenic acid was found both in the blood and brain after intraperitoneal administration in mice, and was safe even at very high doses (1000 mg/kg) [30]. A pharmacokinetics and brain penetration study shows that chlorogenic acid is rapidly absorbed in plasma after both intranasal and intravenous administration of 10 mg/kg and reaches brain and cerebrospinal fluid [34]. The concentration of chlorogenic acid in the brain after 30 min of intranasal administration was about 250 g/mL and remained about 25 g/mL after 6 h. Such data allow to suggest that chlorogenic acid and other compounds from quince leave extracts might be applied as complementary therapy for glioblastoma. However, future research should focus on an effective and safe dose, biologically active compound absorption, distribution, and excretion.

#### **5. Conclusions**

The main compound in ethanolic extracts from Japanese quince cultivars 'Rondo', 'Rasa', and 'Darius' is chlorogenic acid, and the amount of this compound is similar in all three cultivars. The amount of other phenolic compounds is more variable between the cultivars. The extracts of the leaves of all three cultivars significantly decrease viability of C6 and HROG36 glioblastoma cells; in the case of the HROG36 cells, the extracts are equally toxic, but in the C6 cells, the extracts are more toxic than glioblastoma drug temozolomide. The effect on viability has strong correlation with the level of chlorogenic acid for both cell types. In addition, quince leaf extracts exert significantly higher toxicity on rat C6 glioblastoma cells compared to primary rat neuronal-glial cerebellar cells. This finding suggests that Japanese quince leaves could be further investigated as anticancer drugs for glioblastoma treatment.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/2304-815 8/10/1/18/s1, Table S1: Validation characteristics of developed UPLC-ESI-MS/MS method.

**Author Contributions:** Conceptualization, P.V., V.J., and A.J.; methodology, V.Z., I.U., R.B., D.M., and Z.B.; formal analysis, I.U., R.B., D.K., U.M., D.M., and M.L.; investigation, V.Z., I.U., R.B., U.M., and Z.B.; resources, P.V., A.J., and V.J.; writing—original draft preparation, I.U. and A.J.; writing—review and editing, I.U., M.L., P.V., A.J., and V.J.; visualization, I.U., R.B., M.L., and A.J.; supervision, P.V., A.J., and V.J.; All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was financed by the Institute of Horticulture, Lithuanian Research Centre for Agriculture and Forestry and the EUREKA Network Project E! 13496 "OHMDRINKS" (No. 01.2.2- MITA-K-702-08-003).

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. The data are not publicly available due to the institutional data policy.

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

#### **References**


*Article*
