Flavonoid and Non-Flavonoid Compounds of Autumn Royal and Egnatia Grape Skin Extracts Affect Membrane PUFA’s Profile and Cell Morphology in Human Colon Cancer Cell Lines

Grapes contain many flavonoid and non-flavonoid compounds with anticancer effects. In this work we fully characterized the polyphenolic profile of two grape skin extracts (GSEs), Autumn Royal and Egnatia, and assessed their effects on Polyunsaturated Fatty Acid (PUFA) membrane levels of Caco2 and SW480 human colon cancer cell lines. Gene expression of 15-lipoxygenase-1 (15-LOX-1), and peroxisome proliferator-activated receptor gamma (PPAR-γ), as well as cell morphology, were evaluated. The polyphenolic composition was analyzed by Ultra-High-Performance Liquid Chromatography/Quadrupole-Time of Flight mass spectrometry (UHPLC/QTOF) analysis. PUFA levels were evaluated by gas chromatography, and gene expression levels of 15-LOX-1 and PPAR-γ were analyzed by real-time Polymerase Chain Reaction (PCR). Morphological cell changes caused by GSEs were identified by field emission scanning electron microscope (FE-SEM) and photomicrograph examination. We detected a different profile of flavonoid and non-flavonoid compounds in Autumn Royal and Egnatia GSEs. Cultured cells showed an increase of total PUFA levels mainly after treatment with Autumn Royal grape, and were richer in flavonoids when compared with the Egnatia variety. Both GSEs were able to affect 15-LOX-1 and PPAR-γ gene expression and cell morphology. Our results highlighted a new antitumor mechanism of GSEs that involves membrane PUFAs and their downstream pathways.

The anthocyanin profiles of Autumn Royal and Egnatia GSEs were very similar and characterized by a higher percentage of anthocyanin-3-O-monoglucoside compounds (43% in both samples) (Figure 1). These molecules have different antitumor effects depending on the B-ring substituents [29]. Anthocyanins, whose structure contains an o-dihydroxy (catechol) B-ring, such as cyanidin-3-O-monoglucoside and delphinidin-3-O-monoglucoside, are characterized by tumorigenesis inhibition activity [30].  Within each class of flavonoids, a structural variation exists in their basic 15-carbon skeleton, leading to different physicochemical properties. An improvement in the antitumor biological activities of flavonoids is due to the 3′-and 4′-hydroxyl groups (o-diphenol groups), known as Signals of phenolic acids were relevant in both varieties (22% and 35% for Autumn Royal and Egnatia, respectively), while total stilbene signals intensity was 2% in Autumn Royal and 5% in Egnatia extracts ( Figure 2). Both samples showed flavonols and flavanonols as main signals, with total signal intensity 53% and 54% in Autumn Royal and Egnatia, respectively. In particular, high intensity of the B-ring trisubstituted flavonols, such as myricetin and syringetin glycosides, was found. Instead, signals of flavan-3-ols and proanthocyanidins were higher in Autumn Royal (23%) compared to Egnatia extract (6%). Several studies showed that flavan-3-ol oligomers (proanthocyanidins) are potent antioxidants and free radical scavengers also characterized by anticancer properties [31,32]. In our previous study, we found a higher antioxidant activity in Autumn Royal compared to Egnatia, and this result could also be due to the greater content of flavan-3-ols and proanthocyanidins found in Autumn Royal GSEs [28].
Within each class of flavonoids, a structural variation exists in their basic 15-carbon skeleton, leading to different physicochemical properties. An improvement in the antitumor biological activities of flavonoids is due to the 3 -and 4 -hydroxyl groups (o-diphenol groups), known as catechol groups, present on the B-ring [33]. Compared to the Egnatia variety, Autumn Royal presented a higher catechol percentage, calculated by summing the intensities of the signals of those metabolites whose structure includes one or more o-diphenol groups (100% versus 57.6%, respectively). Table 1 shows the o-diphenol compounds identified in both GSEs.
The number and position of hydroxyl groups also influence the interactions of flavonoids with the cell membrane lipid bilayer. The hydrophilic flavonoids, which contain more hydroxyl groups, interact with the polar head groups by hydrogen bonds, inducing membrane rigidification. On the contrary, the hydrophobic flavonoids pass through cell membranes, causing a modification of their permeability and fluidity [6,34,35]. Therefore, the contribution of each phenolic compound can be different, and their synergistic/antagonist interactions might influence their biological effects. To investigate the effects of Autumn Royal and Egnatia GSEs on the n-3 and n-6 PUFAs membrane composition, Caco2 and SW480 human colon cancer cell lines were treated with increasing concentrations of GSEs (20, 50 and 80 µg/mL), and the lipidomic profile was analyzed after 48 h of treatment (Table 2a,b). The choice to use these GSEs concentrations and this experimental time (48 h), was dictated by our previous study, in which we demonstrated that the greatest antiproliferative effect of GSEs in Caco2 and SW480 cell lines was observed in these experimental conditions [16].
Compared to the untreated control group (CTR), in Caco2 cells, the treatment with Autumn Royal and Egnatia GSEs caused an increase of both essential fatty acids (EFAs), linoleic acid (LA) and α-linolenic acid (ALA), already starting from the concentration of 20 µg/mL (Table 2a), and these increases did not correspond to a modification in the n-3 and n-6 fatty acids downstream pathways (Table 2a). Moreover, the increase in total PUFA levels, already at the lowest concentration (20 µg/mL) of both GSEs, was essentially due to the contribution of LA and ALA (Table 2a).
In the SW480 cell line, the treatment with Autumn Royal GSE increased LA levels at 20 µg/mL of concentration, whereas for Egnatia GSE, a higher concentration (50 µg/mL) was necessary (Table 2b). Compared to CTR, only a decrease in arachidonic acid (AA) levels was observed after exposure of both GSEs at 50 µg/mL (Table 2b). As regards the n-3 PUFAs pathway, ALA levels were induced after exposure to the two grape polyphenols, in particular after treatment with 80 µg/mL of Autumn Royal GSE and 20 µg/mL of Egnatia GSE (Table 2b). The increase in total PUFAs obtained after Autumn Royal treatment, at the concentration of 20 µg/mL, was due exclusively to the contribution of LA (Table 2b). The reduction of AA levels found at the concentration of 50 µg/mL of Autumn Royal was more pronounced than the increase of LA obtained at the same concentration of extract, thus balancing the levels of total PUFAs at 50 and 80 µg/mL (Table 2b).
Moreover, no alteration in n-6/n-3 ratio levels was found in all two cell lines and with both treatments (data not shown).
We used two human colon adenocarcinoma cell lines with different degrees of differentiation, Caco2 and SW480, in order to investigate the variations of the membrane PUFA levels induced by the quality of polyphenols contained in the two table grape varieties used. Cancer cells are in continuous proliferation and need large quantities of fatty acids and phospholipids to generate new cellular membranes [19,36]. Previously, we demonstrated the ability of Autumn Royal and Egnatia GSEs to influence membrane fluidity in Caco2 and SW480, through the inhibition of the enzyme SCD1 [16]. This enzyme, that converts saturated fatty acids (SFAs) into monounsaturated fatty acids (MUFAs), was reduced mainly in the Caco2 cell line after GSEs treatment. In this study, we demonstrated that Autumn Royal and Egnatia GSEs were also able to influence the membrane levels of total PUFAs. Different basal levels of PUFAs were found in untreated Caco2 and SW80 cells, probably due to cell type, developmental and growth stage of cells [37]. In SW480 cells with a lower degree of differentiation, the high AA levels found contribute to increasing the total PUFAs levels present in these cells, preparing the cellular pathways towards more inflammatory outcomes.  Both GSEs induced the levels of EFAs, LA and ALA, in all two cell lines studied. EFAs are important structural components of cell membranes that the animal cells must exclusively obtain from their environment [37]. Therefore, the increase of LA and ALA detected in treated Caco2 and SW80 cells was certainly attributable to the fatty acids contained in the grape skins.
Several studies have shown that LA, γ-linolenic acid (GLA), and dihomo-γ-linoleic acid (DGLA) have anticancer effects, unlike AA, which has been associated with the inflammation and with onset of the tumor [38][39][40][41]. Treatment with increasing concentrations of both extracts increased LA levels in SW480 cells. In addition, in these cell lines, both extracts led to a reduction in AA compared to untreated cells, exerting an anti-inflammatory effect inside cells. As regards the n-3 fatty acids pathway, no variation was observed. Hanikoglu A. et al. found differences in the reorganization of fatty acids in cell membranes of two different breast cancer cell lines, MCF-7 and MDA-MB231, after treatment with somatostatin, curcumin, and quercetin, alone or in combination [14]. Cancer cells, according to the degree of differentiation, behave differently to treatment with drugs and/or natural compounds, and this feature could explain the different response of SW480 to treatment with GSEs, with respect to Caco2 cells.

Effects of GSE Treatments on the Gene Expression of 15-LOX-1 and PPAR-γ in Human Colon Cancer Cell Lines
To better investigate the effects of Autumn Royal and Egnatia GSEs on membrane PUFA levels and their antitumoral and antimetastatic action, the gene expression of 15-LOX-1 and PPAR-γ, markers involved in the onset of colorectal cancer (CRC), was studied. Figure (Figure 3a). 15-LOX-1, through its product 13-S-HODE, activates PPAR-γ by inhibiting colorectal tumorigenesis. Therefore, possible changes in PPAR-γ gene expression after exposure to increasing concentrations of GSEs were investigated. Compared to CTR, Autumn Royal GSE treatment exerted an up-regulation of PPAR-γ mRNA levels, starting from 20 µg/mL in both Caco2 and SW480 cells (Figure 3b), whereas a higher concentration of Egnatia (80 µg/mL) was need to obtain the same significant increase in PPAR-γ gene expression (Figure 3b).
Carcinogenesis is known to be also caused by changes in PUFA levels of the cell membrane, including colon cancer formation [19]. 15-LOX-1 is able to oxygenate both n-3 and n-6 PUFAs. The main substrate of 15-LOX-1 is represented by LA, leading to the formation of 13(S)-HODE that activates PPAR-γ, an antimetastatic and anti-inflammatory factor in CRC [26,27].
The increase in expression of 15-LOX-1 found in the cell lines studied after GSE treatment confirms the antitumor effect exerted by these extracts, and the data obtained highlight a new mechanism of action through which GSEs inhibit colon tumorigenesis. Previously, we demonstrated the ability of GSEs to block cell migration and motility by inhibiting SCD1 and some components of the cytoskeleton [16]. The data obtained in this work show that the antiproliferative effect of GSEs also occurs through the induction of the expression of 15-LOX-1 that can be used for therapeutic purposes in CRC.
Cimen I. et al. have shown that 15-LOX-1 indirectly inhibits NF-kB through 13(S)-HODE-mediated PPAR-γ activation in HCT-116 and HT29 CRC cell lines, thus blocking cell proliferation [26]. Moreover, again in HCT-116 and HT29 cell lines, the expression of 15-LOX-1 reduced the ability of cells to adhere to fibronectin, thus inhibiting cell motility [42]. Molecules 2020, 25, x 8 of 18 Carcinogenesis is known to be also caused by changes in PUFA levels of the cell membrane, including colon cancer formation [19]. 15-LOX-1 is able to oxygenate both n-3 and n-6 PUFAs. The main substrate of 15-LOX-1 is represented by LA, leading to the formation of 13(S)-HODE that activates PPAR-γ, an antimetastatic and anti-inflammatory factor in CRC [26,27]. Previous in vitro studies have shown that grape extracts can act differently on proliferation and apoptotic pathways [43,44]. These different biological effects of GSEs could depend both on the type of cancer cell and on the different polyphenolic content of grape extracts [43,45,46]. In fact, it is known that there are cell lines more sensitive to treatment with polyphenols than others in relation to cellular differentiation degree. Moreover, certain flavonoids and non-flavonoids contained in grape extracts can act together synergistically to give particular antiproliferative effects on cancer cells [46,47]. These considerations suggest that the quality of the polyphenolic content in a grape cultivar is an important factor that must be considered. Figures 4 and 5 show the cell morphology of Caco2 and SW480 cell lines, respectively, treated with increasing concentrations of Autumn Royal and Egnatia GSEs (10, 20, 50, and 80 µg/mL) for 24 (T1) and 48 h (T2), analyzed by field emission scanning electron microscope (FE-SEM). To highlight the cellular morphological differences before and after each time GSEs exposure, untreated cells were used as control (CTR) at T0, T1, and T2 (Figures 4 and 5).

Autumn Royal and Egnatia GSEs Induce Cell Morphological Changes
Caco2 CTR cells appeared firmly adherent and covered with abundant microvilli, with a visible cytoplasm and at the center a notable nucleus region without shrinkage (Figure 4). Both at T1 and T2, treatment with Autumn Royal GSE induced visible characteristic morphological changes, such as shrinkage of membrane cells, starting from low concentrations of extract (10 µg/mL), showing a typical state of cell suffering (Figure 4a). In comparison to CTR, the increase of the Autumn Royal concentration induced evident and characteristic changes in the cells, as the cytoplasmic contraction and cell membrane collapse. Moreover, at the highest concentrations of Autumn Royal GSE, it was no longer possible to distinguish cell structures ( Figure 4a). As regards the treatment with Egnatia GSE, at T1, the maximum concentration (80 µg/mL) was necessary to observe clear signs of apoptosis, while at T2, already at 50 µg/mL, membrane blebbing and cell shrinkage were noted (Figure 4b).
The FE-SEM micrographs reported in Figure 5 show the untreated controls (CTR) of the SW480 cell line that appeared flat and adherent to the substrate with an evident central nucleus. The morphological changes induced by Autumn Royal GSE on SW480 cells were visible at T1 starting from the highest treatment concentrations (50 and 80 µg/mL), given that at the lowest concentrations the cells were morphologically similar to the untreated cells (Figure 5a). At T2, the proapoptotic effect of the polyphenols contained in Autumn Royal GSE was appreciated, starting from the lowest concentrations (10 µg/mL) (Figure 5a). At the concentration of 80 µg/mL of GSE, the cellular structures did not appear very detailed and appreciable, when compared to the CTR; in addition, the apparent break of the surface of the cell membrane caused cell death (Figure 5a). Both at T1 and T2, SW480 cells treated with low concentrations of Egnatia GSE were visibly adherent with a round and abundant cytoplasm (Figure 5b). Furthermore, these cells were connected with neighboring cells and extended in all directions. The higher concentrations of GSE (50 and 80 µg/mL) induced characteristic changes, as well as the reduction of cell cytoplasm and a decrease of surface microvilli, which led to cellular apoptosis (Figure 5b).
To our knowledge, this is the first study describing cell morphological changes induced by polyphenols using FE-SEM micrographs. A previous study of Wang S. et al. demonstrated the ability of Trollius chinensis flavonoids to induce apoptosis in human breast cancer MCF-7 cells using SEM analysis [48]. In these cells treated with high concentrations of flavonoids, the microvilli on the cellular surface completely disappeared and cell membranes collapsed. Other morphological changes, in particular cell shrinkage and membrane blebbing, have been found in HCT-15 colon cancer cells after treatment with diet-derived gallic acid [49]. However, our study confirms previous data obtained about antiproliferative and proapoptotic effects in human colon cancer cell lines treated with both GSEs [28] and, for the first time, demonstrates that the beneficial effects of GSE polyphenols are also due to their ability to induce morphological changes in cancer cells, preventing their growth and proliferation.     Table grapes cultivar Autumn Royal, a seedless black grape variety, and Egnatia, a new red seedless genotype, were planted and grown in Apulia region at the Research Center for Viticulture and Enology of the Council for Agricultural Research and Economy (CREA-VE, Turi, BA, Italy). Grape samples were harvested at maturity in summer 2019 and berries were randomly collected and frozen at −20 • C. Approximately 100 frozen berries were manually peeled. To prepare the extracts, 250 mg of dry skin powder were mixed with 5 mL extraction solution of ethanol:water:hydrogen chloride 37% (70:30:1 v/v/v). After 24 h of complete darkness, the mixture was centrifuged, and the supernatant recovered, concentrated in a SpeedVac concentrator (Savant®SPD131DDA, Thermo Fisher Scientific, Waltham, MA, USA) for 90 min at 25 • C and 1.5 atmospheres of pressure and analyzed.

Total Polyphenolic Content
Total phenolic content was determined by Folin-Ciocalteu micro scale protocol with slight modification, as previously described [50]. Briefly, 1 mL of water, 0.02 mL of extract sample, 0.2 mL of the Folin reagent, and 0.8 mL of 10% sodium carbonate solution were mixed and brought to 3 mL. The absorbance was measured at 765 nm after 90 min. Results were expressed as mg GAE/g dw using calibration curves with standard gallic acid.

UHPLC/QTOF Mass Spectrometry
The extracts were three-fold diluted with H 2 O/CH 3

Cell Culture and Treatment
Human colon adenocarcinoma derived Caco2 cell line (ATCC: HTB 37) (well-differentiated) (G1-2) (from adenocarcinoma) and SW480 cell line (ATCC: CCL 228) (poorly-differentiated) (G3-4) (from adenocarcinoma grades III-IV) were purchased from the American Type Culture Collection (ATCC) Cell Bank (Manassas, VA, USA). Cells were cultured in Roswell Park Memorial Institute (RPMI) 1640 medium, for Caco2 cells, and Dulbecco's Modified Eagle Medium (DMEM) for SW480 cells. All cell culture medium and reagents were purchased from Gibco, Life Technologies Limited, Paisley, UK. Culture medium was supplemented with 10% fetal bovine serum (FBS), 2 mM glutamine, 100 U/mL penicillin, and 100 µg/mL streptomycin and incubated at 37 • C in a humidified atmosphere containing 5% CO 2 in air. Autumn Royal and Egnatia GSEs were added to the medium at increasing concentrations dissolved in 100 µL of solvent, composed by ethanol:water:hydrogen chloride 37% (70:30:1 v/v/v) (Sigma Aldrich, Milan, Italy), whereas the cells of control group received the same amount of the solvent. The cells were then incubated at 37 • C in a humidified 5% CO 2 incubator for 24 and 48 h.

Lipids Extraction and PUFAs Analysis
Cell membrane fatty acids were extracted after 48 h of Autumn Royal or Egnatia GSEs treatment at 20, 50, and 80 µg/mL of concentrations. Untreated cells were used as control. Lipids from cell lysate were extracted using the Folch extraction method with some modifications [52]. PUFA analysis was assessed by a gas chromatograph (ThermoFisher Scientific, Focus GC, Milan, Italy) using ChromQuest 4.1 software (Thermo Fisher Scientific, Focus GC, Milan, Italy), as previously described [16].

RNA Extraction and Quantitative Real-Time PCR
After 48 h of treatment with Autumn Royal or Egnatia GSEs (20, 50, and 80 µg/mL), total RNA was extracted from Caco2 and SW480 cells using the Qiagen RNeasy Mini Kit (Qiagen, Hilden, Germany), according to the manufacturer's instructions. The control sample was represented by untreated cells. Samples were retro-transcribed and analyzed using real-time Polymerase Chain Reaction (PCR) for the evaluation of 15-LOX-1 and PPAR-γ expressions on a CFX96 Touch Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA) according to the manufacturer's instructions. Table 3 shows gene-specific primer sets used (Bio-Rad Laboratories); β-actin gene was chosen as the reference gene. The ∆∆Ct method was used for relative quantification by CFX Manager software 2.1 (Bio-Rad Laboratories).

FE-SEM Investigation
Caco2 and SW480 cells were seeded on 1 cm 2 silicon chips (Ted Pella Inc., Redding, CA, USA) at a density of 10 4 cells per well in eight-well chamber slides and, at a sub-confluent density, untreated control cells (CTR) were fixed at T0. To observe cell morphological changes, cell lines were exposed to increasing concentrations of Autumn Royal and Egnatia GSEs (10, 20, 50, and 80 µg/mL) for 24 (T1) and 48 h (T2). Cell lines were treated as previously described [53]. Briefly, each experimental time had a CTR group to highlight the cellular morphological differences before and after each time GSE exposure. As regards fixation procedures, the cells were treated with 3% glutaraldehyde in PBS for 1 h at 4 • C and then incubated for 1 h at room temperature in 1% osmium tetroxide (OsO 4 ). Samples were washed several times with 0.005 M of sodium cacodylate pH 7.2, and the dehydration was completed by passing the samples in increasing concentrations of acetone, from 20% to 100%, for 10 min each step. After completing the drying, samples were coated with a thin Au film using a sputter coater (208HR High Resolution Sputter Coater, Ted Pella Inc.). The fixed cells were observed by using a Zeiss Sigma FE-SEM (Carl Zeiss, Oberkochen, Germany), equipped with an in-lens secondary electron detector. The samples deposited on silicon chips were fixed to stainless-steel sample holders by using double-sided carbon tape. A uniform Au metal coating of few nanometers was deposited on the samples placed on silicon chips by using a turbomolecular pumped SC7620 Mini Sputter/Glow Discharge System of Quorum Technologies (Quorum Technologies Ltd, Lewes, UK). The overall FE-SEM measurements on the samples were acquired at constant Extra-High Tension (EHT) value of 3 kV and at working distance (WD) ranging from 1.8 to 3 mm. The FE-SEM micrographs presented for each experiment were selected as representative of a series of images collected on each sample. Each experiment was performed in triplicate.

Statistical Analysis
Data on total polyphenolic content of GSEs were analyzed by paired Student t-test. The significance of the differences between the control and each treated experimental group was evaluated with one-way analysis of variance (ANOVA) and Dunnett's posttest. Differences were considered as statistically significant with a p-value < 0.05. Data, expressed as mean ± Standard Deviation (SD), were analyzed using STATA statistical software, version 15.1 (StataCorp, 4905 Lakeway Drive, College Station, TX 77845, USA).

Conclusions
Our in vitro data wanted to emphasize the effects of table grape polyphenolic compounds on some molecular mechanisms involved in tumorigenesis. Moreover, the lipidomic approach followed in this study provided valuable information for understanding the protective effect of two GSEs studied on human cell metabolism. Given their ability to influence cell morphology, the flavonoid and non-flavonoid compounds present in table grapes could become a promising dietary source for cancer prevention and treatment.