Cytotoxicity of EPL, PPC and PI in HepG2 and HepaRG cell lines

Addition of EPL to HepG2 cell line cultures at concentrations ≥ 0.375 mg/ml resulted in a marked and concentration-dependent reduction in cell viability (Table 1). Cell viability in HepG2 cells ranged, on average, from 93.0%–102.6% (as a percentage of untreated controls) with PPC at concentrations ≤ 0.9 mg/ml; at higher concentrations there was a modest reduction in cell viability. For PI addition to HepG2 cell line cultures, cell viability ranged, on average, from 92.3%–93.3% (as a percentage of untreated controls) with concentrations ≤ 0.073 mg/ml; PI concentrations 0.36–7.3 mg/ml resulted in a modest decrease in cell viability, whereas PI at 36.3 mg/ml had a severe impact on cell viability (Table 1).

Table 1 Cytotoxicity of EPL, PPC and PI in the HepG2 cell line

Mean fluorescence of resorufin in the culture medium ranged from 101.0%–123.0% (EPL), 95.9%–129.1% (PPC), and 98.8%–123.1% of control values (untreated HepaRG cells) following incubation for 48 h with EPL, PPC, or PI, respectively. Thus, none of the evaluated concentrations (0.01–20 mg/ml) of EPL, PPC and PI showed cytotoxicity in the HepaRG cell line in culture (Supplementary Table S1).

Based on these data, the concentrations of PLs selected for further evaluations in hepatocyte cell lines were 0.1 and 0.25 mg/ml EPL, 0.1 and 1 mg/ml PPC, and 0.1 and 1 mg/ml PI, as these concentrations were considered to be the highest concentrations without major cytotoxicity.

Effect of EPL, PPC and PI on membrane fluidity in HepG2, HepaRG and steatotic HepaRG cell lines

Figure 1 and Supplementary Table S2 depict anisotropy measurements in HepG2 cells (decreased anisotropy values signify increased membrane fluidity). EPL addition resulted in a concentration-dependent and significant decrease in anisotropy measurements versus untreated cells; least-square (LS) mean differences (95% confidence intervals [CI]) versus untreated cells were –0.038 (–0.048 to –0.027; P < 0.001) and –0.058 (–0.041 to –0.047; P < 0.001) with 0.1 and 0.25 mg/ml EPL, respectively. PPC also significantly reduced anisotropy measurements, at both concentrations, in HepG2 cells; LS mean differences (95% CI) versus untreated cells were –0.030 (–0.041 to –0.019; P < 0.001, 0.1 mg/ml PPC) and –0.048 (–0.059 to –0.038; P < 0.001, 1 mg/ml PPC). Membrane fluidity (decreased anisotropy) was also significantly increased in HepG2 cells by incubation with PI; LS mean differences (95% CI) versus untreated cells in anisotropy values were –0.041 (–0.052 to –0.030; P < 0.001, 0.1 mg/ml PI) and –0.068 (–0.078 to –0.057; P < 0.001, 1 mg/ml PI). At 0.1 mg/ml, the largest decrease in anisotropy from untreated cells was seen with PI (68% decrease) and the lowest decrease was with PPC (50% decrease), and EPL resulted in a 63% decrease.

Fig. 1

Effect of EPL, PPC and PI on anisotropy in the HepG2 cell line. Values shown are mean ± SE for 4 separate experiments; n = 1 well for each concentration of each compound per experiment. ***P < 0.001 versus untreated cells. ANI, anisotropy; EPL, essential phospholipids; PI, phosphatidylinositol; PPC, polyenylphosphatidylcholine; SE, standard error. Supplementary Table S2 shows the statistical analyses

In HepaRG cells, the only significant effect on anisotropy was a decrease with 1 mg/ml PI versus untreated controls (LS mean difference [95% CI] –0.032 [–0.056 to –0.008], P < 0.01) (Supplementary Fig. S2A, Supplementary Table S2). EPL and PPC had no significant impact on membrane fluidity in HepaRG cells at the concentrations evaluated. In steatotic HepaRG cells, EPL and PPC did not affect membrane fluidity, whereas PI addition did significantly decrease anisotropy (LS mean difference [95% CI] –0.009 [–0.018 to –0.0004], P < 0.05) (Supplementary Fig. S2B, Supplementary Table S2).

Analyses of cell apoptosisHepG2 cells

Figure 2A and Supplementary Table S3 show the effects of tamoxifen (positive control for apoptosis), EPL, PPC, and PI, and the combination of tamoxifen and each PL on apoptosis induction in HepG2 cells, as assessed by caspase-3/-7 fluorescence intensity of Sytox-negative cells. Tamoxifen resulted in a slight increase in caspase-3/-7 staining at the highest concentration (55 µM) only compared with untreated cells. The concentrations of EPL, PPC and PI evaluated did not induce apoptosis in HepG2 cells (i.e., in the absence of tamoxifen), as there were no significant differences in fluorescence intensity compared with untreated cells. EPL, PPC, or PI addition to HepG2 cells with 42 µM tamoxifen also had no impact on fluorescence staining compared with non-PL exposed cells exposed to 42 µM tamoxifen. EPL significantly reduced 55 µM tamoxifen-induced apoptosis at both concentrations evaluated, i.e., LS mean difference (95% CI) –38.8 (–63.4 to –14.1), P < 0.001 with 0.1 mg/ml EPL, and –43.2 (–67.8 to –18.51), P < 0.001 with 0.25 mg/ml EPL. PPC significantly reduced tamoxifen-induced apoptosis (55 µM) only at the highest concentration evaluated i.e., 1 mg/ml: LS mean difference (95% CI) –39.3 (–64.0 to –14.7), P < 0.001. PI addition to HepG2 cells significantly reduced tamoxifen-induced apoptosis (55 µM) at both concentrations evaluated; 0.1 mg/ml LS mean difference (95% CI) –26.5 (–51.1 to –1.8), P < 0.05; 1 mg/ml: –49.1 (–73.8 to –24.5), P < 0.001.

Fig. 2

Effect of EPL, PPC and PI on apoptosis in the HepG2 cell line. Values shown are mean ± SE (as % of untreated cells) for 2 separate experiments; n = 2 wells for each concentration of each compound per experiment. ns: not significant, *P < 0.05, **P < 0.01, ***P < 0.001 versus untreated cells. Note: for untreated HepG2 cells, 1.43% of cells were found to be Sytox positive (dead cells). Here, these values are presented as percentages, as the results are normalized to untreated cells. AU, arbitrary units; EPL, essential phospholipids; ns, not significant; MFI, median fluorescence intensity; PI, phosphatidylinositol; PPC, polyenylphosphatidylcholine; SE, standard error. Supplementary Table S3 shows the statistical analyses

To confirm the effects of the PLs on tamoxifen-induced apoptosis, an evaluation was also conducted of Sytox-positive cells (Fig. 2B, Supplementary Table S3), as these cells are dead and should only increase in number as a result of apoptosis induction. The number of dead cells was normalized against Sytox-positive untreated cells (without PLs or tamoxifen) and the normalized percentage of dead Sytox-positive cells was compared to that in Sytox-negative cells. Tamoxifen produced a marked concentration-dependent increase in the percentage of dead cells in untreated HepG2 cells. In the absence of tamoxifen, EPL, PPC, and PI had no significant effect on Sytox-positive cells in HepG2 cells. Tamoxifen-induced Sytox-positive cells (42 µM) were not affected by either 0.1 or 0.25 mg/ml EPL, or by either 0.1 or 1 mg/ml PPC. In contrast, the percentage of tamoxifen-induced (42 µM) dead HepG2 cells was significantly decreased in the presence of 1 mg/ml PI (LS mean difference [95% CI] –364.0 [–663.0 to –64.9], P < 0.01), but not by the lower PI concentration of 0.1 mg/ml. In the presence of tamoxifen at 55 µM, EPL reduced the percentage of dead cells at both concentrations evaluated, which was significant for 0.1 mg/ml EPL (LS mean difference [95% CI] –452.8 [–751.8 to –153.8], P < 0.001) but not for 0.25 mg/ml EPL (LS mean difference [95% CI] –293.7 [–592.7 to –5.3], P = 0.0599). PPC significantly reduced tamoxifen-induced Sytox-positive cells (55 µM) at both PPC concentrations, i.e., 0.1 mg/ml (LS mean difference [95% CI] –372.4 [–671.4 to –73.4], P < 0.01), and 1 mg/ml (LS mean difference [95% CI] –416.1 [–715.1 to –117.1], P < 0.001). For tamoxifen at 55 µM, PI significantly reduced apoptosis at both 0.1 mg/ml (LS mean difference [95% CI] –388.1 [–687.2 to –89.1], P < 0.01), and 1 mg/ml (LS mean difference [95% CI] –1045.3 [–1334.3 to –746.2], P < 0.001).

HepaRG cells

Supplementary Fig. S3A and Supplementary Table S4 show the effects of tamoxifen and each PL on apoptosis induction in HepaRG cells (caspase-3/-7 fluorescence intensity of Sytox-negative cells). Tamoxifen addition to HepaRG cells increased caspase-3/-7 staining at both concentrations evaluated compared with untreated cells. In the absence of tamoxifen, none of the PLs induced apoptosis in HepaRG cells. EPL, PPC, and PI at the concentrations tested had no significant effect on caspase-3/-7 staining in presence of either tamoxifen concentration.

The effects of tamoxifen, EPL, PPC, and PI on the percentage of dead HepaRG cells are shown in Supplementary Fig. S3B and Supplementary Table S4. Tamoxifen induced a concentration-dependent increase in the percentage of dead cells. In the absence of tamoxifen, EPL and PPC had no significant effect on the percentage of dead HepaRG cells at the concentrations evaluated. Addition of PI to untreated HepaRG cells had no significant effect at 0.1 mg/ml and slightly increased the percentage of dead cells at 1 mg/ml (LS mean difference [95% CI] 85.7 [2.6–168.7], P < 0.05). EPL addition to HepaRG cells had no significant effects on tamoxifen-induced (45 µM) cell death. PPC addition to HepaRG cells at 0.1 mg/ml significantly increased the percentage of dead cells in the presence of 45 µM tamoxifen (LS mean difference [95% CI] 116.9 [33.8–199.9], P < 0.001); although PPC at 1 mg/ml had no effect. With PI, the percentage of dead cells in the presence of 45 µM tamoxifen (LS mean difference [95% CI] 90.0 [7.0–173.1], P < 0.05), whereas 1 mg/ml PI had no effect on this endpoint. With tamoxifen at 60 µM in HepaRG cells, neither EPL nor PI at the concentrations evaluated had any effect on the percentage of dead cells. PPC addition to HepaRG cells at 0.1 mg/ml significantly decreased the percentage of dead cells induced by 60 µM tamoxifen (LS mean difference [95% CI] 100.1 [33.8–199.9], P < 0.001); whereas 1 mg/ml PPC had no effect on this endpoint.

Steatotic HepaRG cells

Tamoxifen addition to steatotic HepaRG cells slightly increased apoptosis at both concentrations evaluated (Supplementary Fig. S4A, Supplementary Table S5). In this cell line, EPL, PPC, and PI at the concentrations tested had no significant effects on caspase-3/-7 staining in the absence or presence of tamoxifen (Supplementary Fig. S4A, Supplementary Table S5).

Tamoxifen slightly increased the percentage of dead HepaRG cells in untreated cells. This increase was concentration-dependent; overall tamoxifen levels effect P < 0.0001 (Supplementary Fig. 4B, Supplementary Table S5). In the absence of tamoxifen, EPL, PPC, and PI had no significant effect on the percentage of dead steatotic HepaRG cells. For both concentrations of tamoxifen used (45 and 60 µM), EPL, PPC, and PI had no significant effect on the percentage of dead cells at any concentration evaluated (Supplementary Fig. S4B, Supplementary Table S5).

Analyses of hepatocyte transport functionBCRP

Intracellular accumulation of the model substrate, mitoxantrone, was used to assess the activity of BCRP in cell culture, i.e., decreased intracellular concentrations of mitoxantrone indicate increased activity of BCRP in the extracellular transport of this substrate. As a positive control, the BCRP inhibitor KO143 was added to each cell line in the absence of PL.

As a percentage of untreated HepG2 cells, the mean (SD) intracellular concentration of mitoxantrone was 123.4 (7.33; n = 5 replicate wells in 1 experiment) in the presence of KO143, indicating BCRP inhibition. In HepG2 cell cultures, compared with untreated controls, EPL statistically significantly decreased mitoxantrone accumulation at both 0.1 mg/ml (LS mean difference [95% CI] –16.2 [–26.0 to –6.4], P < 0.001) and 0.25 mg/ml (LS mean difference [95% CI] –31.8 [–41.6 to –22.0], P < 0.001) (Fig. 3A, Supplementary Table S6). PI also statistically significantly decreased mitoxantrone accumulation at both 0.1 mg/ml (LS mean difference [95% CI] –22.2 [–32.0 to –12.4], P < 0.001) and 1 mg/ml (LS mean difference [95% CI] –36.0 [–45.8 to –26.2], P < 0.001) versus untreated controls, whereas PPC addition to HepG2 cells had no significant effect on the BCRP activity at either concentration tested (Fig. 3A, Supplementary Table S6).

Fig. 3

Effect of EPL, PPC and PI on hepatocellular transport protein activity in HepG2 cell line. Values shown are mean ± SE (cellular substrate accumulation as percentage of untreated cells) for 4 (MRP-2, BSEP, P-GP) or 5 (BCRP) experiments; n = 1 (BCRP), 4 (MRP-2, BSEP) or 6 (P-GP) wells for each concentration of each compound/experiment. ns: not significant; ***P < 0.001 versus untreated cells. BCRP, breast cancer resistance protein; BSEP, bile salt export protein; EPL, essential phospholipids; MRP-2, multidrug resistance-associated protein 2; P-GP, P-glycoprotein; PI, phosphatidylinositol; PPC, polyenylphosphatidylcholine; SE, standard error. Supplementary Table S6 shows the statistical analyses

As a percentage of untreated cells, the mean (SD) intracellular concentrations of mitoxantrone were 111.4 (7.05; n = 4 experiments) and 117.4 (3.44; n = 4 experiments) in HepaRG and steatotic HepaRG cells, respectively, in the presence of KO143, indicating BCRP inhibition. Addition of EPL, PPC, or PI had no significant effects on BCRP activity in vitro in either HepaRG cells (Supplementary Fig. S5A, Supplementary Table S7) or steatotic HepaRG cells (Supplementary Fig. S6A, Supplementary Table S8).

MRP-2

For assessing MRP-2 activity in vitro, the model substrate used was sulforhodamine 101. Decreased intracellular concentrations of this substrate indicate increased activity of MRP-2 in extracellular transport. As a positive control, the MRP-2 inhibitor MK571 was added to each cell line in the absence of PL.

As a percentage of untreated HepG2 cells, the mean (SD) intracellular concentration of sulforhodamine 101 was 168.9 (28.54; n = 4 experiments) in the presence of MK571, indicating MRP-2 inhibition. There was no significant effect of EPL, PPC, or PI addition to HepG2 cell cultures on sulforhodamine 101 accumulation at the concentrations evaluated versus untreated controls (Fig. 3B, Supplementary Table S6).

The mean (SD) intracellular concentration of sulforhodamine 101 was 141.0% (5.5; n = 4 experiments) of untreated HepaRG controls, in the presence of MK571, indicating MRP-2 inhibition. In HepaRG cells, intracellular accumulation of sulforhodamine 101 was significantly reduced by EPL addition at both 0.1 mg/ml (LS mean difference [95% CI] –10.4 [–19.1 to –1.7], P < 0.05) and 0.25 mg/ml (LS mean difference [95% CI] –12.9 [–21.6 to –4.2], P < 0.01) versus untreated cells (Supplementary Fig. S5B, Supplementary Table S7). MRP-2 activity in HepaRG cells was significantly increased by the addition of 1 mg/ml PPC (LS mean difference [95% CI] –14.9 [–23.6 to –6.2], P < 0.001) versus untreated cells, whereas the 0.1 mg/ml concentration had no significant effect. PI addition to HepaRG cells also significantly increased MRP-2 activity in vitro at 0.1 mg/ml (LS mean difference [95% CI] –9.6 [–18.3 to –0.9], P < 0.05) and 1 mg/ml (LS mean difference [95% CI] –22.3 [–31.0 to –13.5], P < 0.001) versus untreated cells (Supplementary Fig. S5B, Supplementary Table S7).

MRP-2 inhibition by MK571 was achieved in steatotic HepaRG cells, as the mean (SD) intracellular concentration of sulforhodamine 101 was 152.9% (21.07; n = 4 experiments) of untreated steatotic HepaRG controls. In steatotic HepaRG cells, EPL or PPC addition did not significantly impact MRP-2 activity compared with untreated controls (Supplementary Fig. S6B, Supplementary Table S8). Only the highest concentration of PI (1 mg/ml) evaluated in steatotic HepaRG cells significantly increased MRP-2 activity versus untreated controls (LS mean difference [95% CI] –12.4 [–20.9 to –3.8], P < 0.01) (Supplementary Fig. S6B, Supplementary Table S8).

BSEP

Accumulation of 7-β-(4-nitrobenzo-2-oxa-1,3-diazol [NBD])-taurocholate in the canaliculi of cultured cells was used to assess the impact of PL addition on the bile salt export; thus, increased concentrations of this substrate in the canaliculi demonstrate increased BSEP activity.

In HepG2 cells, EPL at 0.25 mg/ml significantly increased 7-β-NBD-taurocholate in the canaliculi compared with untreated controls (LS mean difference [95% CI] 96.8 [42.3–151.3], P < 0.001) (Fig. 3C, Supplementary Table S6). The addition of PPC to HepG2 cells resulted in a marked increase in bile salt export versus controls at both concentrations evaluated (LS mean difference [95% CI]: 0.1 mg/ml 107.7 [53.3–162.2], P < 0.001; 1 mg/ml 168.4 [114.0–222.9], P < 0.001) (Fig. 3C, Supplementary Table S6). Figure 4 depicts the accumulation of BSEP in the canaliculi of HepG2 cells exposed to 0.25 mg/ml EPL versus controls. PI addition to HepG2 cells had no significant effect on bile salt export (Fig. 3C, Supplementary Table S6).

Fig. 4

Visualization of BSEP in 0.25 mg/ml EPL treated HepG2 cells. Representative HepG2 cell culture treated with 0.25 mg/ml EPL and visualization of BSEP as accumulation of the substrate 7-beta-NBD-taurocholate in canaliculi. Yellow lines show the individual cells. Red arrows highlight some example spots of canaliculi structures with accumulated fluorescent substrate. BSEP, bile salt export protein; EPL, essential phospholipids

There was no significant effect of EPL, PPC, or PI at the concentrations evaluated in HepaRG cell cultures on BSEP activity versus untreated controls (Supplementary Fig. S5C, Supplementary Table S7). In steatotic HepaRG cells, EPL at 0.25 mg/ml significantly increased 7-beta-NBD-taurocholate in the canaliculi versus controls (LS mean difference [95% CI] 116.5 [21.0–212.0], P < 0.05) (Supplementary Fig. S6C, Supplementary Table S8).

The addition of PPC to steatotic HepaRG cells increased bile salt export versus controls at both concentrations evaluated (LS mean difference [95% CI]: 0.1 mg/ml 120.3 [24.8–215.8], P < 0.05; 1 mg/ml 183.5 [87.9–279.0], P < 0.001; Supplementary Fig. S6C, Supplementary Table S8). Figure 5 depicts accumulation of BSEP in the canaliculi of steatotic HepaRG cells exposed to 0.25 mg/ml EPL. PI addition to steatotic HepaRG cells had no significant effect on bile salt export (Supplementary Fig. S6C, Supplementary Table S8).

Fig. 5

Visualization of BSEP in 0.25 mg/ml EPL treated steatotic HepaRG cells. Representative steatotic HepaRG cell cultures treated with 0.25 mg/ml EPL and visualization of BSEP as accumulation of the substrate 7-beta-NBD-taurocholate in canaliculi. Yellow lines show the individual cells. Red arrows highlight some example spots of canaliculi structures with accumulated fluorescent substrate. The blue circled area with high fluorescence intensity is a dead cell where the fluorescence substrate accumulates. BSEP, bile salt export protein; EPL, essential phospholipids

P-GP

The model substrate used for evaluating P-GP transport activity was calcein acetoxymethyl (calcein-am). Increased P-GP activity is demonstrated by decreased intracellular concentrations of this substrate. As a positive control, the P-GP inhibitor PSC833 was added to each cell line in the absence of PL.

As a percentage of untreated HepG2 cells, the mean (SD) intracellular concentration of calcein-am was 140.17 (46.05; n = 4 experiments) in the presence of PSC833, indicating P-GP inhibition. All 3 PL preparations added to HepG2 cells in vitro significantly increased P-GP-mediated transport at all concentrations evaluated compared with control (Fig. 3D, Supplementary Table S6). With EPL addition, LS mean differences (95% CI) were: 0.1 mg/ml –32.3 (–42.4 to –22.2), P < 0.001; 0.25 mg/ml –41.7 (–51.8 to –31.6). Following PPC addition, LS mean differences (95% CI) were: 0.1 mg/ml –20.1 (–30.2 to –10.0), P < 0.001; 1.0 mg/ml –47.2 (–57.3 to –37.1). For PI addition, LS mean differences (95% CI) were: 0.1 mg/ml –32.4 (–42.5 to –22.3), P < 0.001; 1.0 mg/ml –43.3 (–53.4 to –33.2).

The mean (SD) intracellular concentration of calcein-am was 207.92% (81.45; n = 4 experiments) of untreated HepaRG controls, in the presence of PSC833, indicating P-GP inhibition. Addition of either EPL or PPC at the concentrations evaluated had no significant effects on P-GP transport activity in HepaRG cells versus controls (Supplementary Fig. S5D, Supplementary Table S7). PI addition to HepaRG cells significantly increased P-GP activity only at the 1 mg/ml concentration (LS mean difference [95% CI] –23.2 [–36.1 to –10.4], P < 0.001) compared with controls (Supplementary Fig. S5D, Supplementary Table S7).

P-GP inhibition by PSC833 was achieved in steatotic HepaRG cells, as the mean (SD) intracellular concentration of calcein-am was 262.1% (9.86; n = 4 experiments) of untreated steatotic HepaRG controls. In steatotic HepaRG cells, EPL and PPC at the concentrations evaluated had no significant effects on P-GP transport activity versus controls (Supplementary Fig. S6D, Supplementary Table S8). PI addition to steatotic HepaRG cells significantly increased P-GP activity only at the 1 mg/ml concentration (LS mean difference [95% CI] –32.6 [–57.0 to –8.1], P < 0.01) compared with controls (Supplementary Fig. S6D, Supplementary Table S8).