Pistacia lentiscus extract enhances mammary epithelial cells’ productivity by modulating their oxidative status

We assessed the potential of phenolic compounds from Pistacia lentiscus (lentisk) to enhance production of milk constituents in bovine mammary epithelial cells (MEC). MEC were exposed to 0 (control), 1 or 10 ppm of polyphenols from lentisk ethanolic extract (PLEE) for 24 h. PLEE were absorbed by the MEC plasma membrane, but also penetrated the cell to accumulate in and around the nucleus. PLEE increased triglyceride content in the cell and its secretion to the medium, and significantly increased intracellular lipid droplet diameter. Compared to control, PLEE increased dose-dependently the lactose synthesis, secretion of whey proteins, and contents of casein. To evaluate mitochondrial activity under pro-oxidant load, MEC were preincubated with PLEE and exposed for 2 h to H2O2. Exposure to H2O2 increased the proportion of cells with impaired mitochondrial membrane potential twofold in controls, but not in PLEE-pre-treated cells. Accordingly, proton leakage was markedly decreased by PLEE, and coupling efficiency between the respiratory chain and ATP production was significantly enhanced. Thus, lentisk polyphenols divert energy to production of milk fat, protein and lactose, with less energy directed to cellular damage control; alternatively, PLEE enables MEC to maintain energy and oxidative status under extreme metabolic rate required for milk production and secretion, and reduces the limitation on energy required to support production.


Results
Doses of PLEE used in the present study were according to preliminary dose response experiment with 1, 10 and 100 ppm of PLEE, which resulted in 99.01% ± 12.8 and 100.52% ± 29.4 and 40.9% ± 33.3 live cells, respectively, compared to 100% living cells in control group. Doses were chosen according to their toxic effect on live cell percentage after 24 h of treatment and according to solubility in the medium.

P. lentiscus polyphenols penetrate MEC in culture and increase antioxidant capacity. To inves-
tigate the association between polyphenols and MEC, the cells were exposed to 1 and 10 ppm PLEE. After exposure, cells were washed and visualized by confocal microscopy, taking advantage of the auto-fluorescence of polyphenols. The polyphenols penetrated the MEC cytoplasm and nucleus and gathered around the nuclei (Fig. 1). The autofluorescence intensity was higher for the 10 ppm vs. 1 ppm PLEE treatment, whereas no autofluorescence was observed in control cells, which were not exposed to the plant extract.
After exposure to the different PLEE concentrations, MEC were washed to remove any unbound extract, and the cells' antioxidant capacity was determined by luminol-dependent chemiluminescence (LDCL) assay (Fig. 2). A ninefold increase in antioxidant capacity was found in treated cells (P = 0.046) compared to the untreated controls. PLEE treatment alters lipid production and secretion. The effect of PLEE on fat production and secretion into the medium was determined. Cells were harvested and medium was collected after 24 h of treatment with 1 or 10 ppm PLEE. Triglyceride (TG) content and phospholipid composition were determined by HPLC. PLEE treatment significantly increased the TG content in the cell, by 18-22% ( Fig. 4a; P = 0.044), and in the medium for 1 ppm by 35% ( Fig. 4b; P = 0.012) compared to controls. In addition, membrane phospholipid composition was determined, to assess membrane stability and the probability of lipid droplet size modifications due to enhancement of fusion events between intracellular lipid droplets. Phosphatidylcholine and sphingomyelin were the major phospholipids in the MEC membranes, accounting for approximately 40 and 30%, respectively, of all membrane phospholipids. The treatment did not affect the composition of the membrane phospholipids (Fig. 4c).

PLEE treatment alters protein secretion but not intracellular content by MEC.
We investigated the effect of PLEE on MEC protein synthesis and secretion in cell culture. After incubation without (control) or with 1 or 10 ppm PLEE, cells were harvested and medium was collected to determine the content and secretion of whey protein and caseins by HPLC. PLEE increased whey protein content in the medium in a dose-responsive manner, by 50 and 77% for 1 ppm and 10 ppm, respectively ( Fig. 5a; P = 0.009), and casein protein content by 82 and 100%, respectively ( Fig. 5b; P = 0.01), compared to controls. However, whey protein ( Fig. 5c; P = 0.82) and casein ( Fig. 5d; P = 0.24) contents in the cells did not differ between treatments.
PLEE treatment increases lactose production and secretion by MEC. PLEE treatment significantly increased lactose content in the cells by 22 and 27% for 1 and 10 ppm PLEE, respectively ( Fig. 6a; P = 0.02), and in the medium by 53 and 59%, respectively ( Fig. 6b; P = 0.001) compared to controls.

Effect of PLEE on gene expression in MEC.
The effect of PLEE on the expression of genes encoding enzymes in the production chain of lactose, fat and protein, and a marker for mitochondrial activity-NADH:ubiquinone oxidoreductase complex assembly factor 3 (NDUFAF3), a key enzyme in the oxidative phosphorylation chain-was determined. Treatment with 10 ppm PLEE increased β-casein gene expression by 1.75-fold compared to controls ( Fig. 7; P < 0.05). The expression of NDUFAF3 was marginally elevated by the 1 ppm treatment, and reduced by the 10 ppm treatment, resulting in a significant difference only between the 1 and 10 ppm treatments. The gene-expression levels of α-lactalbumin, a key enzyme in lactose biosynthesis, and of fatty acid-binding protein (FABP) were not modified by the PLEE treatment.

PLEE increases mitochondrial count and activity in MEC.
Cellular metabolic status can be determined by mitochondrial number, independent of the cell cycle. Therefore, MitoTracker red stain was used to evaluate the number of mitochondria in MEC incubated for 24 h without (control) or with 1 or 10 ppm PLEE. PLEE treatments increased mitochondrial amounts by 80 and 51%, respectively, compared to controls ( Fig. 8; To assess the role of polyphenols in protection against pro-oxidant challenge, MEC were treated with 0 (control), 1 or 10 ppm PLEE and then exposed to hydrogen peroxide (H 2 O 2 ) for 2 h. Cells were stained with 5,5′,6,6′-tetra-chloro-1,1′,3,3′-tetraethylbenzimidazolyl carbocyanine iodide fluorescent probe (JC-1) before the exposure (Fig. 9d), and after 1 h (Fig. 9e) and 2 h (Fig. 9f) Fig. 9b) membrane potential. Images were taken by fluorescence microscopy. At time zero (Fig. 9d), PLEE treatment did not affect the distribution of high (P = 0.69), impaired (P = 0.48), or intermediate (P = 0.1) membrane potential. Control MEC exposed to 2 h of H 2 O 2 without preincubation with PLEE showed an increasing proportion of cells with impaired mitochondrial membrane potential (from 11.8 to 25.6%; Fig. 9f). Preincubation with 1 ppm PLEE did not change the percentage of cells with damaged mitochondria compared to time 0 (P = 0.224). In MEC treated with 10 ppm PLEE, the proportion of cells with impaired membrane potential increased from 18% at time 0 to 37.5% 2 h after exposure to H 2 O 2 . Moreover, while exposure to H 2 O 2 for 2 h adversely affected control cells with high membrane potential, from 57 to 17% (P < 0.001), no significant changes were observed after 1 h of exposure (Fig. 9e).

PLEE treatment alters oxygen-consumption rates in MEC.
MEC incubated without (control) or with 1 or 10 ppm PLEE were analysed for oxygen-consumption rate. Coupling efficiency, non-mitochondrial oxygen consumption and ATP production were calculated by Wave software (2.6.1.53 version), according to the manufacturer's guidelines. Treatment with 1 ppm PLEE reduced oxygen-consumption rate compared to control and to the 10 ppm treatment ( Fig. 10a; P < 0.01). Accordingly, mitochondrial ATP production was higher in cells treated with 10 vs. 1 ppm PLEE ( Fig. 10b; P < 0.05). Spare respiratory capacity (Fig. 10c) and the maximal possible oxygen-consumption rate were measured after addition of carbonyl cyanide-4-(trifluoromethoxy) phenylhydrazone (FCCP), which artificially imposes supreme ATP demand but disperses the proton gradient and thus blocks conversion of ADP to ATP. While no differences were observed between PLEE treatments (P = 0.12) in maximal oxygen-consumption rate, the spare respiratory capacity in the 1 ppm treatment was 29.6 and 43.2% higher than in the 10 ppm treatment and controls, respectively ( Fig. 10c; P < 0.01). At the end of the experiment, rotenone and antimycin A were added as inhibitors of respiratory complexes I and III, respectively, to determine  (d) Figure 5. Effect of PLEE on intracellular content and secretion of whey protein and caseins. MEC were incubated without (control, white) or with 1 (grey) or 10 (black) ppm PLEE for 24 h. Whey protein and casein contents were determined in the medium to assess secretion ((a,b), respectively) and in the cell lysate, normalized to 1 million cells ((c,d), respectively); n = 4 for each replicate in each treatment. Different letters indicate significant difference at P ≤ 0.05. www.nature.com/scientificreports/ the non-mitochondrial oxygen consumption. MEC treated with 1 ppm PLEE had significantly enhanced nonmitochondrial oxygen consumption compared to control MEC ( Fig. 10d; P = 0.01). Despite the presence of oligomycin in the system, some mitochondrial respiration persisted due to the "uncoupling", which was expressed as proton leakage and coupling efficiency. Proton leakage was markedly decreased by PLEE, most notably by the 1 ppm treatment ( Fig. 10e; P < 0.001). Consequently, the coupling efficiency in the respiratory chain and ATP synthase were significantly more advanced in PLEE-treated vs. control MEC ( Fig. 10f; P < 0.001).

Discussion
We investigated the direct effect of lentisk extract on MEC with a focus on production traits. PLEE positively affected production and metabolism of primary bovine MEC. In particular, enhanced lipid production was concomitant with increased diameter of intracellular lipid droplets (the milk fat globule precursors); casein secretion was elevated; lactose synthesis and secretion was enhanced, and mitochondrial activity was more resilient to pro-oxidant load, as expressed by antioxidant activity, coupling efficiency and spare respiration capacity. In farm animals, the utilization of exogenous antioxidants was studied in several aspects: improvement of milk components and quality for human health 29,30 , vascular homeostasis during prooxidant challenge 31 , incidence of mastitis 32 semen quality 33 , plasma redox status 34 and weight gain of bull calves 35 . However, no cellular mechanism was investigated for these production traits till now.
We first wanted to explore how the phenolic compounds in PLEE interact with MEC in culture; several previous lines of evidence have suggested that the effect of phenolic compounds on cells can be mediated by their coating of, and incorporation into the plasma membrane, as found in erythrocytes 36 , endothelial cells 37 , rat cerebral membranes 38 and semen cells 33 . In the present study, indeed, using confocal fluorescence microscopy, we clearly demonstrated the presence of PLEE polyphenols near the plasma membrane of MEC, in the cytoplasm, Various antioxidants have been shown to have cytoprotective effects against oxidative stress. However, most of these studies were conducted on cell lines of mammary epithelium with limited capacity to synthesize and secrete milk constituents. For instance, exposure of a MAC-T line to resveratrol enhanced the expression of multiple antioxidant genes under normal and oxidative conditions via activation of Akt and ERK signalling pathways 16 . Moreover, the damage to MEC by activated neutrophils was totally inhibited by exposure of MAC-T cells to 50 µM catechin 39 . In primary culture of bovine MEC, treatment with tea polyphenols protected bovine MEC against oxidative stress by scavenging ROS and upregulating the expression of antioxidants and detoxifying enzymes 17 . In agreement with those results, we found a ninefold increase in the antioxidant capacity of MEC treated with PLEE (Fig. 2). Nevertheless, the mechanisms underlying the regulation of phenolic compounds and MEC production are not well understood.
In the present study, we used lipid droplet size as a proxy for fat production by MEC. PLEE treatments (both 1 and 10 ppm) increased lipid droplet size by 20% compared to controls. In general, lipid droplet size is positively correlated with cellular triacylglycerol (TG) content 40,41 , reduced lipolysis, and elevated capacity for lipogenesis 42 . Another mechanism which can contribute to lipid droplet size is fusion, which is tightly associated with cellular phospholipid composition 42 . To determine the dominant mechanism in the increase in lipid droplet size, we quantified the weight, content and composition of phospholipids, as the main membrane component altering membrane stability 43 and potentially modulating lipid droplet fusion 44 . We found no significant change in MEC phospholipid composition among treatments, suggesting that the increase in lipid droplet size under the PLEE treatment is not driven by membrane instability and fusion but, most likely, by greater accumulation of TG in the cell. We extended our test by measuring the lipid-secretion capacity of the MEC and found an average enhancement of 35% in cells treated with 1 ppm PLEE compared to controls (Fig. 4).
While cellular whey protein and casein contents did not differ with PLEE treatment, amounts of secreted whey protein and casein were approximately twofold greater in the 10 ppm PLEE treatment compared to the control (Fig. 5). Although the mechanism is not clear, the increase in casein content was in accordance with an increase in β-casein gene expression. β-Casein is one of the major compounds secreted by MEC, accounting for 24-28 g protein per L of milk 45 , and found to be associated with milk and protein yields in Holstein cows 46 . The fact that greater protein secretion occurred under the PLEE treatment may be attributed to greater non-mitochondrial oxygen consumption recorded in the treated MEC as the protein-folding process prior to secretion occurs in the endoplasmic reticulum requires oxygen 47 . We assume that PLEE increased the cell's oxygen consumption, which in turn increased endoplasmic reticulum activity, manifested by greater casein secretion.
PLEE increased cellular lactose content and elevated its secretion (Fig. 6), but it was not associated with any changes in α-lactalbumin gene expression. Lactose synthesis in MEC is a complex process, starting in the www.nature.com/scientificreports/ endoplasmic reticulum and continuing in the Golgi apparatus. The enzymes participating in this process are not always regulated in a coordinated manner. For instance, in bovine MEC cultured with 5-10 mmol/L glucose, the mRNA expression of B4GALT, the enzyme that transfers galactose to form lactose, and of lactose synthase was elevated, but that of α-lactalbumin was unchanged 48 . On the other hand, lactose synthesis can be allosterically modulated by the availability of its precursors. Two moles of glucose are required for every mole of lactose. The glucose available for this process is tightly associated with the redox status of the cell, because it can be used to generate NADPH in the pentose phosphate pathway. NADPH is a reducing agent, required for the reduction of oxidized glutathione to reduced glutathione. Therefore, increased cellular ROS production might enhance the utilization of glucose for NADPH production as a substrate for ROS quenching with oxidized glutathione by glutathione reductase 49 . In light of this, we suggest that PLEE has a glucose-sparing effect, hence releasing more glucose for lactose synthesis. Taken together, we report the combined effects of PLEE on lipid, lactose, whey protein and casein production, indicating a general effect on productivity, even though we are aware that the improved production traits were not always in agreement with gene-expression levels, suggesting a general bioenergetic effect, presumably through mitochondrial quantity and functionality. Accordingly, exposing MEC to 10 ppm PLEE increased basal respiration, which was expressed in elevated ATP production (Fig. 10). On the other hand, treatment with 1 ppm PLEE did not change basal respiration. However, both treatments reduced proton leakage and increased coupling efficiency. These results imply that the treated cells utilize oxygen more efficiently, and synthesize more ATP without changing their oxygen-consumption rate. Consequently, it is suggested that the greater production of www.nature.com/scientificreports/ milk constituents by the PLEE-treated cells is due to a greater ATP production capacity and most probably, less inhibition by ROS in the metabolic and production pathways. The main ROS producers in the cell are the mitochondria, with enhanced ATP production being coupled with higher ROS production 50 , and the endoplasmic reticulum during protein folding 41 . To further investigate the mitochondrial resistance to pro-oxidant load, we exposed MEC to H 2 O 2 as in Flora et al. 51 . We found that the mitochondria of the PLEE-treated cells could maintain high membrane potential, even under this exogenous oxidative challenge. The inner membrane of mitochondria in the PLEE-treated cells maintained higher potential and durability compared to controls, suggesting that lentisk polyphenols directly protect the energy-production mechanism of MEC from pro-oxidant load. These findings are in line with the previously reported antioxidant capacity of various polyphenols 52 . In particular, gallic acid, a main component in the lentisk extract 53 , is a strong electron donor to ROS 54 . For instance, in mouse, gallic acid provided indirect protection by elevating the hepatic levels of glutathione peroxidase and catalase, thereby potentially providing direct and indirect protection against the pernicious effects of H 2 O 2 55 . Changes in the expression of genes encoding antioxidant enzymes such as HO-1, Xct, Txnrd1, and NQO-1 through Nrf2 occur in a timeframe of 4-25 h 16 , and this was therefore not assessed in the present study, where exposure to oxidative challenge only lasted 2 h.
It should be noted that the PLEE polyphenols entered the MEC in a dose-dependent manner, as autofluorescence was more intense in the 10 ppm compared to 1 ppm treatment, however this was not necessarily correlated with cell function. Moreover, 2 h after exposure to H 2 O 2 , the proportion of cells with impaired membrane potential increased in MEC treated with 10 ppm PLEE, but not with 1 ppm-treated cells. Furthermore, 10 ppmtreated cells had higher ATP production and less spare respiratory capacity than those treated with 1 ppm. Taken together, these results imply that 10 ppm PLEE can be somewhat toxic to the cells for specific energetic processes, while remaining beneficial to other cellular processes.
The effect of exogenous phenolic compounds on milk production has mainly been studied in vivo, with its polyphenols administered as part of the diet. In vivo, dairy goats browsing on lentisk produced milk with more fat and protein and lower urea content compared to their counterparts fed a hay-based diet with undetectable tannin content 26 . Similar results were obtained when Lotus corniculatus was used as a phenolic compound source to feed dairy ewes 56 and cows 57 ; L. corniculatus elevated milk protein content and reduced milk urea content 28,58 . These findings are explained by the protein-binding activity of tannins, protecting the protein from bacterial degradation and hence increasing protein flow in the duodenum, and its availability for the synthesis of milk proteins 52 . However, low contents of dietary phenolic compounds such as quebracho tannins 59 and Acacia mearnsii tannin extract 60 supplemented to dairy cows affected milk fatty acid composition, which could probably be attributed to a systemic effect of the phenolic compounds after their absorption into the bloodstream, and not to a local effect in the gastrointestinal tract.
In summary, the assumption that dietary phenolic compounds exert a systemic effect post-absorption is supported by the finding that dietary phenolic compounds are absorbed into the bloodstream and transported into milk 61 , indicating that the mammary cells are directly exposed to these chemical compounds. The results of the current study shed light on the direct effect of phenolic compounds on MEC, and how exposure to these compounds modulates energy status of the cells, and probably utilization of reducing agents and nutrients (such as NADPH and glucose) for production instead of for damage control. Supplementation of lentisk polyphenols to the culture medium enhanced energy production, which was directed to lipid, protein and lactose production instead of oxidative-damage regulation. Such knowledge could provide the foundation for developing new profitable and sustainable nutritional strategies to enhance mammary gland productivity while coping with stress.
Experimental design. Primary MEC were plated at 50,000 cells per well in 6-well plates on glass cover slips for MitoTracker red, Nile red or JC-1 staining and MEC autofluorescence determination; at 150,000 cells per 60-mm plastic dish for cellular lipid, protein and lactose extraction, RNA extraction, and antioxidant capacity determination; and at 28,000 cells in XF24 cell-culture microplates (Seahorse Bioscience, North Billerica, MA, USA) for cellular metabolic flux analyses. Cells were grown in plastic culture dishes with DMEM/F12 supplemented with 10% FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, 0.25 μg/mL amphotericin B, 1 μg/mL insulin and 0.5 μg/mL hydrocortisone. When plates reached to 80% confluence, which was approximately 3 days post-plating, medium was replaced with DMEM/F12 without serum, containing 0.15% (w/v) free fatty acid-free BSA and insulin (1 μg/mL), hydrocortisone (0.5 μg/mL) and prolactin (1 μg/mL) for 48 h to induce milk lipid,

Scientific Reports
| (2020) 10:20985 | https://doi.org/10.1038/s41598-020-78065-z www.nature.com/scientificreports/ protein and lactose synthesis. Then cells were treated for 24 h with treatment medium that contained 1 ppm or 10 ppm of a 70% ethanolic extract of P. lentiscus foliage, all dissolved in DMEM/F12 supplemented with insulin (1 μg/mL), hydrocortisone (0.5 μg/mL), prolactin (1 μg/mL) and 0.1 M oleic acid (C18:1). Doses of PLEE used in the present study were selected according to the results of preliminary dose response experiment with 0, 1, 10 and 100 ppm of PLEE. Doses were chosen according to their toxic effect on live cell percentage after 24 h of treatment and according to solubility in the medium. Compared with control (0 ppm) no change in cells live cells number was visualized under 1 and 10 ppm. After 24 h, cells were harvested for cell counting, lipid extraction, protein and lactose quantification and antioxidant capacity determination, or for staining of intracellular lipid droplets or mitochondria, or JC-1 staining. The cells were not harvested for counting for real-time PCR; cells were harvested at the end of the oxygenconsumption assay for cell counting and normalization.
Primary culture. Primary culture of MEC was isolated from mammary biopsies according to a protocol established in our laboratory 44,62 with slight changes. Briefly, udder tissue was collected from three lactating cows at Rahat commercial slaughterhouse. Tissues were collected after cows were commercially slaughtered by a certified slaughterhouse worker, and after veterinarian inspection, and transferred to the laboratory. Only udders with normal structure from lactating cows were used. Tissues were immediately submerged in DMEM/F12 medium supplemented with 10% (w/v) FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, 0.25 μg/mL amphotericin B, 1 μg/mL insulin and 0.5 μg/mL hydrocortisone (growth medium). Tissue (10 g) was digested by shaking in 100 mL of growth medium supplemented with collagenase (1 mg/mL), hyaluronidase (1 mg/mL) and heparin (0.02 mg/mL) at 100 rpm for 3 h combined with two controlled breaks to enrich the suspended fraction, at 37 °C. After incubation, the suspension was filtered through a metal mesh (250 μm), and the filtrate was centrifuged at 350×g for 5 min. The sediment was treated with trypsin-EDTA and 0.04% (w/v) DNase. Cells were washed with growth medium supplemented with heparin and treated with DNase, filtered using a 100-μm cell strainer (BD Falcon, Bedford, MA, USA) and then washed with the growth medium. Cells were grown in plastic culture dishes with DMEM/F12 supplemented with 10% FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, 0.25 μg/mL amphotericin B, 1 μg/mL insulin and 0.5 μg/mL hydrocortisone.
Plant material collection, extraction and fractionation. Leaves from P. lentiscus were randomly collected from Ramat Hanadiv Nature Park (south of Carmel Heights, 32° 33′ N, 34° 56′ E) during the summer and spring. The lentisk leaf extract was prepared according to Azaizeh et al. 53 . Briefly, the leaves were dried at 50 °C for 24 h, ground and stored for further analysis at room temperature. Powdered leaf tissue samples (10 g each) were incubated with 100 mL of 70% ethanol at 35 °C for 24 h. The crude ethanol extract was filtered and evaporated under vacuum (Rotorvapor Hie-VAP; Hiedolph, Germany) at 45 °C to remove the ethanol and water. The extraction yield was calculated as gram extract per gram dry matter (plant leaves), and the dried extracts were stored at − 20 °C. On treatment day, dried extract was diluted with double-distilled water to 1 and 10 ppm and named PLEE (polyphenols from lentisk ethanolic extract). Dried plant extract of lentisk was analysed for phenolic compounds using HPLC as previously described 53 . Briefly, for hydrolysis, 250 mg of dried extract was dissolved in 25 mL of 1% HCl, and incubated at 70 °C for 8 h with shaking at 100 rpm. The presence of polyphenols and flavonoids in the hydrolysed and non-hydrolysed solutions was determined using reversed-phase HPLC analysis with binary gradient elution on a Thermo Scientific Finnigan Surveyor system equipped with a PDA Plus Detector (220-450 nm) (Thermo Fisher Scientific, Waltham, MA, USA). As recently published by Hadaya et al. 29 , the chromatographic separation was performed on a Gemini 5µ C6-Phenyl 110 Å column (250 × 4.60 mm) (Phenomenex, Torrance, CA, USA) at 30 °C. The mobile phase consisted of 0.1% acetic acid in water (A) and 0.1% acetic acid in methanol (B), and the run was programmed for 35 min. The elution conditions were: 0 to 5 min, 25% B; 5 to 15 min, 25 to 50% B; 15 to 20 min, 50% B; 20 to 29 min, 50 to 25% B; 29 to 35 min, 25% B. The flow rate was 1.0 mL/min and the injection volume was 15 µL. Standards for tyrosol, luteolin-7-O-glycoside, apigenin-7-O-glucoside, quercetin-7-O-rutinoside, luteolin, quercetin, gallic acid and rutin were from Sigma Aldrich Israel; and for hydroxytyrosol, oleuropein, chlorogenic acid, catechin and myricetin, from Cayman Chemicals (Ann Arbor, MI, USA). The polyphenols of the ethanolic phospholipid extract consisted mainly of galloyl derivatives (63.6%), flavonol glucosides (28.6%) and catechin (7.8%).
Lipid extraction and analysis. After the 24-h treatment with PLEE, total lipids were extracted from harvested cells with trypsin (0.05%), washed with phosphate buffered saline (PBS) and stored at − 20 °C until lipid extraction as previously described 62 . Briefly, a 5 mL of chloroform:methanol solution (2:1, v/v) was added to each sample. After incubation at room temperature, 1 mL of double-distilled H 2 O was added. After overnight incubation with cold extraction at 4 °C, the upper phase was removed, and the lower phase was filtered through glass wool. Samples were then evaporated under a nitrogen stream at 65 °C, diluted in 100 μL chloroform:methanol (97:3, v/v) and stored at − 20 °C until injection for HPLC analysis. Separation of polar and neutral lipids was performed on a silica column (Zorbax RX-SIL, 4.6 × 250 mm, Agilent Technologies, Santa Clara, CA, USA) by HPLC (HP 1200, Agilent Technologies) with an evaporative light-scattering detector (1200 series ELSD, Agilent Technologies). The column was heated to 40 °C, flow was 1 mL/min, and injection volume was 20 μL. The ELSD was heated to 65 °C, nitrogen pressure was 3.8 bars, the filter was set to 5, and gain (sensitivity) was set to 4 for the first 14 min, then changed to 12 until 21 min, and then to 7 until the end of the run to enable detection of differently abundant lipid components. The separation protocol consisted of a gradient of dichloromethane, methanol:ammonium mix (99:1, v/v) and double-distilled water. The separation process was managed by ChemStation software (Agilent Technologies), which permitted acquisition of data from the ELSD detector. The separated lipids were identified using external standards (Sigma Aldrich, Israel). Quantification was performed Scientific Reports | (2020) 10:20985 | https://doi.org/10.1038/s41598-020-78065-z www.nature.com/scientificreports/ against external standard curves and expressed as µg/per 10 6 live cells or as weight % out of the sum of phospholipids (µg) in the sample. Live cell number was determined with a haemocytometer after 5 min of Trypan blue staining.
Protein extraction and analysis. After the 24-h treatment with PLEE, cells were harvested with trypsin (0.05%), washed with phosphate buffered saline (PBS) and stored at − 20 °C until protein extraction, and 0.5 mL of medium were collected for further analysis. Cells and medium were lysed by 4 rotations of freezer and thaw cycles followed by 30 min of sonication. Then, cells were centrifuged in 350 g for 5 min and 0.   Real-time PCR analysis. Self-designed primers were produced using Primer-BLAST software (NCBI, http:// www.ncbi.nlm.nih.gov/tools /prime r-blast /index ), based on cDNA sequences published in the NCBI database and validated by PCR-product sequencing. Primers were synthesized by Sigma Aldrich Israel according to the sequences, as indicated (Table 1). cDNA was mixed with the primers and platinum SYBR Green qPCR Supermix-UDG without ROX (Invitrogen Corporation, Carlsbad, CA, USA). A Mx3000P Real-Time PCR System (Stratagene, La Jolla, CA, USA) was used. Analysis was performed by MxPro software version 4.10 (Stratagene). Dissociation-curve analysis was performed after each real-time experiment to confirm the presence of only one product. The efficiency of the reaction and the initial mRNA quantity in the sample were determined using LightCycler 96 software version 1.1.0.1320 (Roche, Basel, Switzerland), and the ΔΔCt method was used to calculate the relative expression of each gene. The efficiency of the reaction and the initial mRNA quantity in the sample were determined using DART-PCR software version 1.0. Data were normalized to the geometrical mean of two housekeeping genes: 18S and β2-microglobulin, and presented as fold change relative to the control treatment.

Lactose extraction and analysis.
Oxygen-consumption rate. Oxygen-consumption rate of live cells in a 24-well plate was measured in real time using a Seahorse Bioscience XF24 extracellular flux analyser. Cell number optimization was determined on 28,000 cells/mL which were seeded and grown for 24 h to 70-100% confluence before the metabolic flux analysis. Then, cell growth medium was replaced with XF assay medium (pH 7.4, Seahorse Biosciences) supplemented with 2 mM glutamine and 1 mM sodium pyruvate. Prior to the cell-respiration measurement, cells were incubated for 1 h at 37 °C without CO 2 . Basal oxygen consumption, maximal respiratory capacity and non-mitochondrial oxygen consumption were determined using the XF Cell Mito Stress Test Kit (Agilent Technologies, UK). The inhibitors of mitochondrial respiration, including oligomycin, carbonylcyanide www.nature.com/scientificreports/ p-trifluoromethoxyphenylhydrazone (FCCP) and rotenone/antimycin were auto-injected into the experimental wells after basal measurements. Oligomycin used to inhibit ATP synthase, FCCP used as a protonophore, Rotenone and antimycin blocked mitochondrial respiration of electron transport chain. Cell number determined by haemocytometer was used to normalize the oxygen-consumption rate values between wells (n = 4 for each treatment) and treatments.
Scanning electron microscopy (SEM). Primary MEC were fixed using the methanol method described by Talbot and White 66 . The sample was then dried in a critical point dryer (CPD-030, Bal-Tec/Leica, Wetzlar, Germany) and gold-coated in a sputter-coating unit (Quorum Technologies/Polaron, Laughton, UK). The sample was observed by low-vacuum SEM (JSM 5410 LV, Jeol, Tokyo, Japan). The SEM images were used to measure secretion area in the cell membrane.
Statistical analysis. All statistical procedures were performed using JMP software version 12.0.1 (SAS Institute, Cary, NC, USA). Reported data are means ± SE. Dependent variables were checked for homogeneous variance by unequal variances in JMP software, and if the variance was not homogeneous, a Welch-ANOVA test was performed. Effects and comparisons between treatments were tested by ANOVA followed by LS Mean Tukey-Kramer HSD multiple-comparison test. The distribution of cell phenotypes based on mitochondrial potential membrane categories was compared by Chi square test. Significance probe was set to 0.05.