Phenolic compounds from Actinidia deliciosa leaves: Caco-2 permeability, enzyme inhibitory activity and cell protein profile studies

2.1 General experimental procedure

All chemicals were analytical grade. MTT (thiazolyl tetrazolium bromide), bovine serum albumin, rutin, acetylcholinesterase (AChE), acetylthiocholine (AChI) and 5,5-dithio-bis-(2-nitrobenzoic acid) (DTNB) were obtained from Sigma (Barcelona, Spain). Trifluoroacetic acid, quercitrin and methanol were obtained from Merck (VWR, International, Darmstadt, Germany). Ethanol was obtained from Carlo Erba (Peypin, France). Roswell Park Memorial Institute (RPMI) medium, Hank’s balanced salt solution (HBSS), glutamine, Pen-Strep (penicillin and streptomycin mixture), and fetal bovine serum (FBS) were bought from Lonza (Verviers, Belgium). For the one-dimensional electrophoresis gel GE Healthcare Amersham™ ECL™ was used in a 4–12% concentration. The marker protein sample buffer (5X SDS PAGE) and staining solution, BlueSafe, were purchased from NZYTech®.

2.2 Plant material

Fully expanded leaves from Actinidia deliciosa (A. Chev) C.F. Liang et A.R. Ferguson (Actinidiaceae) were collected in December 2016 from plants cultivated at the region of Beira Baixa, Portugal (lat. 39° 46′ 47,783″ N; long. 7° 48′ 20,171″ W, 244 m alt). Fresh leaves were washed and immediately used for extraction as decoction, 48 h after collection.

2.3 Preparation of plant extracts

Aqueous plant extracts were prepared as decoction by using 10 g of grounded leaves in 100 mL of distilled water, boiling for 10 min. The decoctions were filtered through number 1 Whatman grade paper and lyophilized. The yield of extraction was 114 mg of extract/g of plant. For the phenolic compounds isolation, the precipitation of fibers was developed as described in Ghanem et al. (2010), with small modifications (Henriques et al., 2017).

2.4 Reverse Phase High Performance Liquid Chromatography (RP-HPLC-DAD)

The chromatographic analysis was carried out using reverse phase liquid chromatography with a diode array detector (RP-HPLC-DAD) as described in Falé et al. (2013b). The quantification of rutin and quercitrin was carried out using the RP-HPLC-DAD by injecting standard solutions and preparing a calibration curve under the experimental conditions (Falé et al., 2013b). The results were analysed using Microsoft® Excel 2007.

2.5 Cytotoxicity studies in Caco-2 cells

This study was carried out as described in Falé et al. (2013a, 2013b) using the MTT test.

2.6 Permeation studies in Caco-2 cell line

Caco-2 cells (ATCC#HTB-37) a human colorectal adenocarcinoma epithelial cell line, were cultured as described in Falé et al. (2014). Non-purified extract and the fibres-free extract were introduced in the apical chamber and samples were withdrawn after 6 h from the apical and basolateral compartment. These samples were immediately analysed for the phenolic compounds by RP-HPLC-DAD as described in Falé et al. (2013b) and the cells were frozen at −80 °C for protein analysis. The permeation was measured as % of the amount introduced in the apical chamber and as apparent permeability coefficients (Papp). These were determined as indicated in Falé et al. (2014).

2.7 Polyacrylamide gel Electrophoresis (SDS–PAGE)

The cells (previously stored at −80 °C) that were under contact with the extracts and the control were disrupted by sonication during 10 min and then centrifuged at 4°C, 12,000g for 10 min. The supernatant was removed (150 µL) and proteins were quantified using the Bradford method (Bio-Rad®). A volume containing approximately 20 µg protein was used to obtain protein by precipitation with two volumes of acetone and left in the cold (4 °C) for 12 h. The samples were centrifuged at 4 °C, 12,000g for 10 min and then dried under nitrogen current. Electrophoresis was performed on a horizontal homogeneous precast gel 4–12% acrylamide (45 × 80 × 1.4 mm) using a GE Healthcare Life Sciences system and following the instructions provided. The precipitated protein from each sample, blank, cells under the influence of extract and mucilage-free extract, were dissolved with sample buffer and water in a ratio of 1/4. In each gel, well was added 20 µL of the sample (20 µg protein). Twenty microliters of molecular weight marker and sample buffer were also applied. Electrophoretic separation was carried out at 160 V for 80 min. The gel was stained with BlueSafe™ for one hour and photographed using ImageQuant LAS 500 system, GE Healthcare Life Sciences. The gel bands were treated and quantified using ImageJ software.

2.8 Protein identification by mass spectrometry

The protein identification was carried out as a service at Mass-Spectrometry Unit, ITQB-UNL, Oeiras, Portugal.

2.9 Acetylcholinesterase inhibition

Acetylcholinesterase inhibition was measured as described by Falé et al. (2013b) using the extract and the fibre-free extract.

2.10 Data analysis

The software used was Microsoft® Excel 2010, and the results were expressed as mean ± standard deviation. Additional analysis of variance (ANOVA) was performed with α = 0.05 using the software developed by Microsoft®.

3.1 Polyphenols from A.deliciosa decoction

Previous work on the extraction of phenolic compounds from A.deliciosa leaves (Henriques et al., 2017) indicated that these agricultural waste was a good promising source of bioactive compounds. Leaves from this plant collected at the end of December, the pruning period in Portugal, were used for the extraction of phenolic compounds with boiling water (decoction), followed by purification with ethanol precipitation, as described previously (Henriques et al., 2017). Both complete extract and fibers-free extract were analyzed by RP-HPLC-DAD. The chromatographic profile was similar for both extracts, as shown in Fig. 1. The major compounds, peaks 1 and 2 were identified by comparing their retention times and UV–Vis spectra rutin and quercitrin standard compounds. Both compounds are quercetin glycoside derivatives, Fig. 2. These flavonoids were quantified using the external standard method, Table 1. The amount of quercitrin in the extract is twice the value of rutin (significant differences at α = 0.05). Discarding fibers from the extract caused an increase in approximately 50% in the rutin and quercitrin content, keeping the proportionally between the two flavonoids, Table 1. Several phenolic compounds were also determined in the leaves of other plants (Alabri et al., 2014; Iqbal et al., 2015).

Close

Fig. 1

RP-HPLC-DAD of Actinidia deliciosa leaves’decoction. (……) extract; (–) fiber-free extract.

Export to PPT

Close

Fig. 2

Chemical structure of: (a): Rutin; (b): Quercitrin.

Export to PPT

3.2 Permeation of flavonoid derivatives through Caco-2 cells

To study the phenolic compound permeation the toxicity of the extract towards Caco-2 cells has to be evaluated. Neither the complete extract nor the fibers-free extract were toxic to these cells according to Okonogi et al. (2007) (Table 1). In order to evaluate the permeation of polyphenols through the intestinal barrier, and the effect of fibers on their permeation, Caco-2 cells were allowed to grow and differentiate as a monolayer on a Transwell™ system (Falé et al., 2013a, 2014). Extract and fiber-free extract were introduced in the apical chamber, with the same amount of polyphenols, incubated for six hours, after which samples were withdrawn from the apical and basolateral compartments. Only rutin and quercitrin were detected in the basolateral compartment by RP-HPLC-DAD, Fig. 3, and the quantification is shown in Table 2. Permeation of rutin and quercitrin through a monolayer of Caco-2 cells is similar for both flavonoid derivatives, around 7.6 and 8.1 × 10⁻⁶ cm/s, corresponding to approximately 10% of the initial amount detected in the basolateral chamber (not statistically different at α = 0.05). These values are in accordance with P_app values determined in other studies (D’Archivio et al., 2010). When using the fiber-free extract, the permeation values for both rutin and quercitrin, were also similar 3.6 and 3.4 × 10⁻⁶ cm/s, (α = 0.05), but the permeation without fiber represents a decrease of 41% and 56% comparatively to the complete extract (significantly different, α = 0.05). Our results showed the role of fibers on the permeation of flavonoids through the intestine, increasing the permeation of these compounds. Previous works have also shown that the bioavailability of phenolic compounds can be affected by the food matrix (D’Archivio et al., 2010; Pérez-Jiméne et al., 2009). Pectin, a glycoside-based polymer, also favored the permeation of phenolic compounds (Tamura et al., 2007). Permeation of rutin through Caco-2 cells monolayer can differ when the compound is within an extract containing several other phenolic compounds or when isolated, as a pure compound (Falé et al., 2013a, 2014).

Close

Fig. 3

RP-HPLC-DAD of the total extract and mucilage-free extract in basolateral side after 6 h of permeation through Caco-2 monolayers.

Export to PPT

Cells were lysed after the permeation experiments to analyze the intracellular content in flavonoid derivatives. Only 0.09% and 0.14% of the initial amount of rutin and quercitrin were detected inside Caco-2 cells using the complete extract, Table 2. Previous work also detected low quantities of rutin inside cells (Falé et al., 2013a). Within fiber-free extract, rutin and quercitrin showed an intake of 0.36 and 0.47%, respectively, representing an increase of 300% and 235% comparatively to the complete extract. Although our results indicated an increase in flavonoid permeation in the presence of fibers, the effect of these polymers on the permeation of phenolic compounds is still a matter of discussion and research (Bohn, 2014). The increase in permeation and the smaller intracellular amount observed with fibers (polymers having glycosylic units) is in agreement with other works showing that mucin, proteins containing glycosylic groups may enlarge tight junctions between the cells, facilitating the paracellular transport of compounds (Lis et al., 2013). Several studies point out the impact of mucus-secretory cells on the permeation of nutritional food compounds, increasing it (Pelaseyed et al., 2014). As a whole, the compounds could permeate the apical membrane of Caco-2 cells, corresponding to the luminal surface of the intestinal barrier.

3.3 Effect of phenolic compounds on protein-profile Caco-2 cells

As flavonoids were detected inside the cells, the effect on the cell proteins was also investigated. The cells that had been disrupted for the flavonoid analysis were also analyzed in relation to the protein profile. A strong effect on the protein content can be noticed by SDS-PAGE analysis, indicating that flavonoids can modify the protein profile, Fig. 4. Protein with molecular weight around 52 kDa showed an increase in 8% intensity under the effect of extract and the same protein increased 164% intensity relatively to the blank when the fiber-free extract was used. Similar results were obtained using aqueous extract of leaves from a medicinal plant (Falé et al., 2012) and it is known (Salucci et al., 2002) that flavonoids interfere with Caco-2 cell cycle which means that they cause modifications in protein expression inside the cell.

Close

Fig. 4

Gel obtained from SDS-PAGE of protein cells under the effect of: (a) control; (b) extract; (c) fiber-free extract; (d) proteins marker.

Export to PPT

Proteins present in higher amount in the treated cells were identified by mass spectrometry. The protein with molecular weight 52 kDa was identified as albumin, and the one with molecular weight of approximately 61 kDa was identified as transketolase (E.C. 2.2.1.1). Albumin is coming from the culture medium, as intestinal cells do not produce this protein and it is the main protein in the culture medium. It is relevant that the cells in contact with the polyphenols have a much higher amount of this protein, although all the cells had been washed before disruption and extraction of the proteins. This fact may be explained by the binding of polyphenols to proteins (Rawel et al., 2005; Falé et al., 2012). Polyphenols can permeate the cell membrane and be found inside the cells, as was shown in Table 2 due to several membrane transporter proteins (Gonzales et al., 2015). It is possible that medium proteins bound to polyphenols may enter the cells, explaining the higher amount of albumin in treated cells.

Transketolase (TKT) is involved in the phosphate pentose pathway (PPP), and recent research work has disclosed its relationship with cancer (Patra and Hay, 2014; De Preter et al., 2015; Xu et al., 2016; Kowalik et al., 2017). Cancer cells have an overproduction of this enzyme, and we observed that the presence of polyphenols increased its biosynthesis. This metabolic pathway is involved in energy production, with concomitant production of NADPH and reducing power (Xu et al., 2016). This molecule can participate in the decrease of oxidative stress found in cells. In cancer cells, it has been found that the knocking-out this enzyme avoids cells division, decreasing proliferation of tumour cells (De Preter et al., 2015). Decreasing oxidative stress in normal cells will help cells to proceed their normal biochemical activity, preventing diseases like cancer, diabetes and atherosclerosis (Reuter et al., 2010). The over production of TKT will mean increased synthesis of antioxidant power (NADPH), fighting oxidative stress. This situation is advantageous in normal cells but not for cancer cells.

These results also help to interpret cell toxicity. It is reported in this study that the leaves’ extract were not toxic to the cells, this means also that the compounds present cannot kill cancer cells. In fact, recent research has shown that knock-out of TKT will decrease cancer cells (Xu et al., 2016). As these polyphenols did not decrease TKT, but increased its production, they are not toxic to cancer cells neither to normal cells, but may be beneficial when a decrease in oxidative stress is needed. Therefore our research shows that polyphenols may favour replication of cancer cells, but may prevent oxidative stress-related diseases in normal cells.

3.4 Acetylcholinesterase inhibitory activity

In order to evaluate whether the compounds could inhibit some of the enzymes involved in the treatment of several important diseases, AChE was chosen as a model enzyme. IC₅₀ value for the extract and the fiber-free extract are indicated in Table 1. Mucilage-free extract originated a product with higher inhibitory activity. An increase in 30% of activity that can be attributed to a higher amount of rutin and quercitrin determined in this extract, Table 1. According to previous studies, plant extracts containing rutin and quercitrin also inhibit AChE activity (Hernandez et al., 2010; Neagu et al., 2016; Ressaissi et al., 2016), although with a lower inhibitory activity than the drug galantamine (Hernandez et al., 2010). A.delicosa leaves’ extract has AChE inhibitory capacity within the values found for plant extracts (Kamal et al., 2015). In the present work, these compounds justify approximately 25% of the inhibitory activity of the fiber-free extract, suggesting that the enzyme inhibitory activity is the added contribution of all the phenolic compounds present, several flavonoid derivatives and a triterpene acid-O-hexoside (Henriques et al., 2017).