Differential effect of oleic and palmitic acid on lipid accumulation and apoptosis in cultured hepatocytes


  • Partly supported by grants from MIUR Ministero Istruzione Università e Ricerca Scientifica -PRIN 2004061213_001.

Nicola Carulli and Paola Loria, Università degli Studi di Modena, Dipartimento di Medicina Interna, Endocrinologia, Metabolismo e Geriatria, Nuovo Ospedale Civile Sant’Agostino Estense di Baggiovara, Via Giardini Baggiovara, Modena 41100, Italy. Email: paola.loria@unimore.it


Background and Aim:  Studies have shown monounsaturated oleic acid to be less toxic than palmitic acid and to prevent/attenuate palmitic acid hepatocites toxicity in steatosis models in vitro. However, to what degree these effects are mediated by steatosis extent is unknown.

Methods:  We evaluated whether steatosis per se is associated with hepatocytes apoptosis and determined the role of oleic and palmitic acid, the most abundant fatty acids in western diets, on triglyceride accumulation and apoptosis in an in vitro model of steatosis induced in three hepatocytic cell lines (HepG2, HuH7, WRL68). The impact of incubation for 24 h with oleic (0.66 and 1.32 mM) and palmitic acid (0.33 and 0.66 mM), alone or combined (molar ratio 2 : 1) on steatosis, apoptosis, and insulin signalling, was evaluated.

Results:  Concurrent with PPARγ and SREBP-1 gene activation, steatosis extent was larger when cells were treated with oleic than with palmitic acid; the latter fatty acid was associated with increased PPARα expression. Cell apoptosis was inversely proportional to steatosis deposition. Moreover, palmitic, but not oleic acid, impaired insulin signalling. Despite the higher amount of fat resulting from incubation of the two fatty acids combined, the apoptosis rate and impaired insulin signalling were lower than in cells treated with palmitic acid alone, indicating a protective effect of oleic acid.

Conclusions:  Oleic acid is more steatogenic but less apoptotic than palmitic acid in hepatocityc cell cultures. These data may provide a biological basis for clinical findings on dietary patterns and pathogenetic models of nonalcoholic fatty liver disease.


4′,6-diamidino-2-phenylindole dihydrochloride


Fatty Acid


Free Fatty Acids


Insulin Resistance


Non esterified fatty acids


Oleic Acid


Palmitic Acid


Type 2 Diabetes




Very low density lipoproteins


Fatty acids (FAs) are major components of biological cell membranes that play important roles in intracellular signalling and as precursors for ligands that bind to nuclear receptors.1,2 FAs represent vital energy stores, but high-fat diets are associated with the development of obesity and type 2 diabetes (T2D).3,4 Several lines of evidence indicate the importance of both quantitative and qualitative (e.g. saturated vs unsaturated) changes in dietary FAs as relevant mechanisms for the development of nonalcoholic fatty liver disease (NAFLD) both in rodent models and in humans.5–9 Finally, in obese NAFLD patients, the decreased unsaturated/ saturated fatty acid ratio in serum, fat and liver tissue might have a pathogenetic role in the disease.10,11 Increased free fatty acids (FFA) levels are linked with the pathogenesis of insulin resistance (IR), which is considered a major determinant in the pathogenesis of NAFLD. Peripheral IR results in increased concentrations of circulating FFAs due to unopposed antilipolytic insulin action. Additionally, accumulation of FFA results in an impaired post-receptor insulin signalling12 contributing to IR and causing a further increase in FFA.

The hepatocytes are not a physiological site of lipid storage and development of steatosis is associated with cellular dysfunction and apoptosis.13,14 This phenomenon, which also occurs in the kidney, pancreas and heart, is referred to as lipotoxicity and is believed to play a significant role in the pathogenesis of tissue damage.15 In fact, the extent of apoptosis correlates with the severity of steatohepatitis and the stage of fibrosis in human NAFLD.16 Lipotoxicity also contributes to decreased insulin sensitivity, perhaps by promoting the accumulation of fat-derived metabolites that inhibit insulin signalling and action.3,17

Fatty acids are chemically classified as saturated and unsaturated (monounsaturated and polyunsaturated) and their structure affects their biological effects. This is of particular relevance for the most abundant FAs present in the diet and in serum, namely palmitic acid (PA), a saturated FA, and oleic acid (OA), a monounsaturated FA.18 In spite of the amount of information available for the pathogenesis of triglyceride accumulation, little is known on the metabolic effects of PA and OA as determinants of the pathophysiology of NAFLD.

Studies on the effect of FAs-induced steatosis on cellular apoptosis have demonstrated that PA and OA mixtures-induced steatosis is associated with apoptosis in hepatocyte cell cultures.19 Moreover, similar to other cell lines, monounsaturated FAs were less toxic also in hepatocytes20 and were able to prevent/attenuate PA toxicity.21 However, it remains unknown to what extent these effects were mediated by degree/type of lipid accumulated in the hepatocytes.

Aim of the present study was to evaluate the role of OA and PA, the major FA present in western diets, on steatogenesis, cell survival, FAs composition, gene expression and insulin signalling in an in vitro model of steatosis. Given that a priori it could not be ruled out that specific hepatocytic cell lines might differ in their susceptibility to steatosis/apoptosis, we carried out experiments in three different cell lines.


Hepatocyte cell cultures

Three cell lines with different characteristics were used: (i) HepG2 cells are derived from a well differentiated human hepatoblastoma cell line that retain many characteristics of normal differentiated quiescent hepatocytes, and are p53 wild-type;22,23 (ii) WRL-68 cells, a fetal liver cell line;24 and (iii) p53-mutated HuH7 cells, derived from a differentiated hepatocellular carcinoma.25 HepG2 and WRL68 cells were purchased from Istituto Zooprofilattico Sperimentale (Brescia, Italy) and HuH7 cells from Japanese Cancer Research Resources Bank (Osaka, Japan).

Long-chain FAs, palmitic (16:0) and oleic (18:1) were provided as sodium salts (Sigma-Aldrich, Milan, Italy). Palmitic acid and OA were dissolved in MetOH 99% (stock solution 100 mM). Stock solutions were kept at −20°C before the experiments. Solutions and reagents used for cell cultures were from GIBCO Life Technologies Ltd (Grand Island, NY, USA).

Protocol of the study—induction and evaluation of steatosis

Steatosis was induced by a slight modification of previously described methods.26,27 HepG2, WRL-68 and HuH-7 cell cultures were incubated with phenol red-free medium containing 10% of charcoal stripped fetal bovine serum (FBS; Cambrex, East Rutherford, NJ, USA), 1% bovine serum albumin (BSA), supplemented with FFA (oleic and palmitic acid alone or in association) at the following final concentrations: a) PA: 0.33 mM and 0.66 mM; b) OA: 0.66 mM and 1.32 mM and c) mixtures of OA and PA (ratio 2:1) at two different final concentrations: PA 0.33 mM + OA 0.66 mM (final fatty acids concentration 1 mM) and PA 0.66 mM + OA 1.32 mM (final fatty acids concentration 2 mM). Control cell cultures were incubated with plain medium or with medium added with the vehicle in which fatty acids were dissolved. After 24 h of incubation with PA and OA alone or in association, the extent of steatosis, apoptosis and gene expression were evaluated as detailed above. In agreement with previous studies,19,27 we found 24 h to be optimal incubation time. At 12 h or less the TG accumulation was low and no detectable variations in gene expression could be shown (data not shown). A more prolonged incubation time (e.g. 48 h and 72 h) while not providing a significant advantage in terms of intracellular TG accumulation was associated with a significant decrease in cell viability at higher FFA concentrations in our and others’ experience (data not shown,28).

After fixation with formaldehyde, neutral lipids were stained using 0.5% Oil-Red-O (Sigma-Aldrich) in isopropanol for 30 min and nuclei were stained with hematoxylin.

For the electron microscopy pictures cells were detached from the substrate with trypsin, fixed in 2.5% glutaraldeyde for 1 h and than post-fixed in tyrode 1% OsO4 (osmium tetraoxyde) for 1 h. Cells were then dehydrated in progressive concentrations of ethanol, dried in propylene oxide and embedded in Durcupan's resin. Sections 50 nm were cut on an Ultratatome (Reichert-Jung, Wetzlar, Germany), placed on mesh copper grids (S162, ASSING, Rome, Italy) and stained with 7% uranyl acetate and 2.66% lead citrate. Micro-photographs were taken by an electron microscope (JEOL, model 2011, Peabody, MA, USA) at 200 kV.

Intracellular triglyceride content was evaluated after lysis of the cells with NaOH 0.3 N. Triglyceride concentration (mg/dL) was determined by standard technique with an automatic analyzer (Roche, Milan, Italy) and normalised by protein content (mg/mL).

Total intracellular lipid content was evaluated by Nile Red staining (Adipored, Cambrex); briefly, cells were grown in 96 black-plates and treated with FAs. At the end of incubation cells were washed twice with phosphate buffered saline (PBS) and incubated with Adipored for 10 min. Fluorescence was evaluated as previously described.29

Evaluation of apoptosis

Apoptosis, assessed with DAPI (4',6-diamidino-2-phenylindole dihydrochloride, Sigma-Aldrich) staining and caspases 3/7 activity (Promega, Milan, Italy) were evaluated as previously described.29

Measurement of the fatty acid composition

Lipids were extracted from cells using the method of Folch et al.30 Aliquots were mildly saponified as previously described31 in order to obtain free fatty acids for high pressure liquid chromatography (HPLC) analysis. Separation of FAs was carried out with a Agilent 1100 HPLC system (Agilent, Palo Alto, CA, USA) equipped with a diode array detector. A C-18 Inertsil 5 ODS-2 Chrompack column (Chrompack International BV, Middleburg, The Netherlands), 5 µm particle size, 150 × 4.6 mm, was used with a mobile phase of CH 3 CN/H2O/CH 3 COOH (70/30/0.12, v/v/v) at a flow rate of 1.5 mL/min. Unsaturated fatty acids were detected at 200 nm. Spectra (195–315 nm) of the eluate were obtained every 1.28 s and were electronically stored.32 These spectra were taken to confirm the identification of the HPLC peaks.

Analysis of saturated FAs and further confirmation of unsaturated fatty acids was carried out by GC assay of FA methyl esters.

Free fatty acids obtained as described above were methylated by the addition of 14% BF3/CH3OH at room temperature and immediately extracted into a solvent consisting of n-hexane and water (4:3 ratio). After centrifugation to separate the two phases, the hexane phase was saved and the aqueous phase was further extracted by another round of hexane. The two hexane collections were combined, dried, and redissolved in 500 µL of n-hexane.33

The gas chromatograph (Model 6890, Agilent) was equipped with split ratio of 20:1 injection port, a flame ionization detector (FID), an autosampler (Model 7673, Agilent), a 100 m HP-88 fused capillary column (Agilent), and an Agilent ChemStation software system. The injector and detector temperatures were set at 250°C and 280°C respectively. Hydrogen served as carrier gas (1 mL/min), and the FID gases were H2 (30 mL/min), N2 (30 mL/min), and purified air (300 mL/min). The temperature program was as follows: initial temperature was 120°C, programmed at 10°C/min to 210°C and 5°C/min to 230°C, then programmed at 25°C/min to 250°C and held for 2 min.

Determination of cellular mRNA level

In HepG2 cell line total RNA was isolated using the RNeasy lipid tissue kit (Qiagen, Milan, Italy) and quantified/checked with RNA Nano LabChip (Agilent, Milan, Italy). About 1 µg of total RNA was reverse-transcribed with the high capacity cDNA Archive Kit (Applied Biosystems, Monza Italy). TaqMan polymerase chain reactions (PCR) were performed on cDNA samples using the TaqMan Universal PCR Master Mix (Applied Biosystems) according to PRISM 7900 HT Sequence Detection Systems.

The TaqMan strategies for each gene have been developed as Assay-on-Demand by Applied Biosystems. Gene expression profiling was achieved using the comparative cycle threshold (CT) method of relative quantification (the calibrator samples were non-treated cells, with 18S RNA used as endogenous control). Data are expressed as log2 of the relative quantity (RQ) defined also as ‘fold induction versus the controls’.

Western blot analysis of cellular proteins

Confluent, serum-starved (12 h) HepG2 were treated with fatty acid, after stimulation with 100 nM of insulin for 15′, the cells were quickly placed on ice, and washed with ice-cold PBS. The mono-layer was lysed in RIPA buffer (20 mM Tris·HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, 1% Nonidet P-40, 1 mM Na3VO4, 1 mM phenyl methyl sulfonyl fluoride, 0.01% protease inhibitor cocktail [Protease Inhibitor Cocktail, Sigma-Aldrich]) and transferred to micro-centrifuge tubes. Insoluble proteins were discarded by centrifugation at 12 000 rpm at 4°C. Protein concentration in the supernatant was measured in duplicate using a commercially available assay (Pierce, Rockford, IL, USA). Equal amounts of total cellular proteins were separated by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and analyzed by Western blot using primary antibodies as indicated. Detection was performed using a chemiluminescent substrate (ECL, Amersham, IL, USA). Primary antibodies were rabbit Phospho-Akt (Ser-473) Antibody (Cell Signalling Technology, Danvers, MA, USA) and mouse Anti-β-Actin antibody (Sigma-Aldrich).

Statistical analysis

Results were expressed as mean ± standard error (SE). All data represent a minimum of three experiments conducted in triplicate, unless otherwise specified. The significance of differences was assessed by Student's t-test for independent data. Linear regression analysis was performed by the least square method. The differences between slope values were evaluated by the analysis of variance. Significance was accepted at the P < 0.05 level. Statistical analysis was performed with the aid of SPSS statistical software (version 14.0 for Windows, SPSS Inc., Chicago, IL, USA).


OA and PA differentially affect triglyceride accumulation in hepatocytic cell lines

We first analyzed the development of lipid accumulation in an in vitro model of hepatic steatosis. HepG2 were exposed to increasing concentrations of OA or PA, or to a combination of the two fatty acids at a 2:1 ratio in favour of OA. After 24 h, lipid accumulation was evident in all cells exposed to FA, as indicated by staining with Nile Red staining (Fig. 1). The degree of fat accumulation was roughly proportional to the concentration of FA to which cells were exposed. Similar findings were obtained for WRL68 and HuH7 cell lines (data not shown). Finally, electron microscopy pictures displayed a massive increase in the number and size of lipid droplets into the cytosol of HepG2 cells incubated with FAs as compared to controls (Fig. 2).

Figure 1.

Effect of different fatty acids on lipid accumulation in HepG2 cells. After 24 h of incubation (for details see materials and methods) intracellular lipid droplets were measured by Nile Red. Data are expressed as means ± standard error of Relative Fluorescence Units (RFLU) per mg of protein. Three experiments conducted in duplicate. *P < 0.01 vs control; **P < 0.01 vs palmitic acid (PA) 0.33 mM; ***P < 0.01 vs PA 0.66 mM; ****P < 0.01 vs oleic acid (OA) 0.66 mM; *****P < 0.01 vs PO1 mM.

Figure 2.

Representative electron microscopy photograph. HepG2 cells in the absence (Panel a) or presence (Panel b) of fatty acids 1 mM (palmitic acid [PA] 0.33 mM + oleic acid [OA] 0.66 mM) for 24 h.

To evaluate quantitatively the relation between the type of FA and the degree of triglyceride accumulation, we measured triglyceride (TG) concentration in cells lysates (Fig. 3). All treatments with FA caused a significant increase in triglyceride content of all three cell lines tested in comparison to control cells. In addition, there was a clear dose-dependency for all conditions tested. However, OA induced a significantly higher triglyceride accumulation than PA when used at equimolar concentrations (Fig. 3). This effect was evident in all three lines, and was statistically significant for HepG2 and WRL-68. Incubation of all cell cultures with mixtures of both FAs showed a dose dependent increase in the amount of triglycerides which was almost equivalent to the sum of triglyceride levels induced by separate incubation with the two different FA. Finally, the concentration of triglyceride in HepG2 cells was always greater than that found in WRL-68 cells but lower than in HuH-7 treated with the same experimental schedule. To sum up, the three different cell lines display a more pronounced steatosis response in the order PA > PA + OA > OA; for each FA a dose-response relationship was observed. However, the absolute ability to accumulate triglycerides was decreasing in the order HuH-7 > HepG2 > WRL-68.

Figure 3.

Effect of different fatty acids on triglycerides accumulation. HepG2 (Panel a), WRL-68 (Panel b) and HuH-7 (Panel c) cell cultures were incubated with palmitic acid (PA) (0.33 and 0.66 mM), oleic acid (OA) (0.66 and 1.32 mM) or mixtures of the two fatty acids (1 mM or 2 mM) for 24 h. Triglyceride (TG) accumulation was evaluated as the concentration of TGs in cell lysates after NaOH lysis. Columns represent mean values ± standard error of three different experiments conducted in triplicate. *P < 0.05 vs control; **P < 0.05 vs PA 0.66 mM.

PA is a stronger apoptotic stimulus than OA

We next analyzed the effects of the different FAs treatments on programmed cell death in the three different lines of hepatocytic cells, using the percentage of cells showing nuclear condensation when stained with DAPI29 (Fig. 4). Palmitic acid at a concentration of 0.33 mM increased significantly (P < 0.05) the percentage of apoptotic cells only in HuH7 cell cultures while at 0.66 mM apoptosis was significantly increased in all cell lines tested. In contrast, OA, used at concentrations as high as 1.32 mM did not induce any increase in the number of apoptotic cells as compared to untreated controls. Remarkably, comparison of the effects of the two FA used at equimolar concentrations showed a significantly higher ability of PA to induce apoptosis in all cell lines. These data indicate that PA, but not OA induces apoptosis in cultured hepatocytes.

Figure 4.

Effect of fatty acids on apoptosis rate. In HepG2 (Panel a), WRL-68 (Panel b) and HuH-7 (Panel c) apoptosis was quantified by assessing the characteristic nuclear changes of apoptosis using the nuclear binding dye 4',6-diamidino-2-phenylindole dihydrochloride (DAPI) and counting cell numbers under a fluorescent microscopy. Columns represent mean values ± standard error of three different experiments conducted in triplicate. *P < 0.05 vs control; **P < 0.05 vs palmitic acid 0.66 mM.

Co-incubation with the two FAs at the lower concentrations (1 mM) did not modify the rate of apoptosis. At the higher concentrations (0.66 mM PA together with 1.32 mM OA) an increase in the percentage of apoptotic cells was observed in all cell lines. However, in two out of three lines, the increase in apoptosis was significantly lower in the presence of OA than when PA at a concentration of 0.66 mM was used alone. These data are strongly suggestive of a protective effect of OA on apoptosis induced by PA.

To add further evidence to the differential effects of the two fatty acids on apoptosis, we measured the activity of caspases 3/7, that belong to the group of the execution caspases, activated once the cell death program is fully operative (Fig. 5). In cells exposed to PA, the activity of caspases 3/7 increased in a dose-dependent fashion. In contrast, OA did not have any effects on caspase activity except in HepG2, where a significant increase was observed. However, it should be stressed that the increase in response to OA was only 1–2 fold as compared to control, while PA induced an 8–16 fold increase. As a result, caspases 3/7 activity in HepG2 exposed to OA was always significantly lower than in cells treated with PA.

Figure 5.

Effect of fatty acids on caspases 3/7 activity. In HepG2 (Panel a), WRL-68 (Panel b) and HuH-7 (Panel c) caspases 3/7 activity was quantified measuring the fluorescence from Z-DEVD-R110 cleavage and normalised by protein content in parallel plates. Columns represent mean values ± standard error of three different experiments conducted in triplicate. *P < 0.05 vs control; **P < 0.05 vs palmitic acid (PA) 0.66 mM; ***P < 0.05 vs PA 0.33 mM.

In co-incubation experiments, the addition of OA to PA decreased caspases activity in comparison to cells treated with PA alone in HepG2 and in WRL68 but not in HuH7 cell line. Thus, data on caspases activity are closely similar to those obtained with DAPI staining.

PA and OA are associated with different FA profiles in HepG2 cells

We hypothesized that the mechanisms of cell toxicity due to palmitic acid could be ascribed to different fatty acid profiles and to the esterified/unesterified ratio. To test this hypothesis, we compared individual and total fatty acid composition obtained by HPLC with the results of TG content assessed enzymatically.

The absolute FAs composition in steatotic HepG2 cultures after exposure to different treatments is shown in Table 1. All treatments induced significant changes in the concentration of various FAs. PA and OA concentration increased after incubation with the respective FA. In addition, while PA incubation was associated with an increase 16:1, 18:1c11 and 14:0, the incubation with OA induced an increase of 18:0, 16:0 and 14:0. As expected, the incubation with the two FAs induced a combined pattern and was also associated with a reduction in the concentration of 16:1 as compared to PA alone. The putative metabolic fate of FAs added to the medium based on results shown in Table 1 is depicted in Figure 6.

Table 1.  Fatty acid profile of HepG2 cells incubated with palmitic acid (PA), oleic acid (OA) or PA and OA combined
 20:522:616:120:418:222:418:118:1 c1112:014:015:016:018:020:022:0Total FAs
  • *

    P < 0.05 vs control.

  • Data are expressed as nmoles/mg of proteins.

P 0,31.1764.65347.159*8.18017.1436.69685.25274.4851.58412.0532.791150.288*25.6941.7711.646440.570
P 0,61.3275.33668.936*10.13821.9248.03799.72797.634*1.85915.323*3.405223.582*28.5331.8141.786584.360*
O 0,61.3625.21830.65110.06118.4848.441291.330*66.2241.31412.0833.404104.33522.1651.6161.772578.460*
O 1,31.4156.24035.75611.74520.58811.683594.991*92.4851.54217.634*3.447152.017*37.363*2.1202.303991.330*
PO 11.1675.06136.0389.46619.6368.433293.478*70.8541.62513.5053.084169.168*27.5141.6761.806662.510*
PO 21.4365.44549.472*10.30123.1229.594532.038*86.541*2.659*16.687*3.601280.682*33.767*1.4802.1541058.980*
Figure 6.

Putative metabolic pathway compatible with changes in fatty acid profile observed after adding palmitic acid (PA) and oleic acid (OA) to cell cultures base on findings shown in Table 1. PA (16:0) increases the concentration of 16:1, 18:1c11 presumably by stimulating delta 9 desaturation of PA yielding 16:1, further elongation (18:1c11). The increase of 14:0 may be ascribed to peroxisomal beta oxidation of PA. OA (18:1c9) increases 18:0, 16:0 and 14:0. 18:0 increase might be due to a partial inhibition of SCD1 with a consequent increase of 16:0 and 14:0 as products of peroxisomal beta oxidation of 18:0.

Total FA content was proportional to the moles of fatty acids added to the medium (data not shown). Contrasting with data on TG content, we found that OA and PA at equimolar doses induced a comparable accumulation of total fatty acids. This strongly suggests that after incubation with PA the concentration of TG is lower (Fig. 1) than total FA concentration due to the presence of a consistent FFA not being detected in the total TG assay. It could be indirectly estimated that the percentage of FFAs following incubation with PA is in the order of magnitude of 15% (vs 0–1% following incubation with OA).

FAs structure and not TG content is the key determinant of apoptosis

To better define the relationship between the degree of steatosis and induction of apoptosis, we plotted the intracellular TG concentration with the activity of caspases 3/7 (Fig. 7). While these parameters were significantly correlated for both FA, the slope of the curve obtained with PA was significantly steeper than the one observed for OA. It is of relevance that the regression for combined treatments fall in between PA and OA, again suggesting a protective effect of OA on PA induced cytotoxicity. Similar relationships were observed also plotting TG versus percentage of apoptosis evaluated by DAPI staining (data not shown). These data confirm that for a given amount of intracellular triglycerides, the degree of apoptosis depends on the FA used for the induction of steatosis being far higher for PA, followed by their association and by OA.

Figure 7.

Relationship between extent of triglyceride (TG) accumulation and caspases 3/7 activity. Regression lines represent the relationship between the two parameters after incubation of HepG2 cells respectively with palmitic acid (PA; ▾), PA + oleic acid (OA; ●) and OA (▪). Single points are the mean of three experiments that have in parallel evaluated TGs content and caspases activity. The slope values of PA vs PA + OA (t = 6.805; P < 0.0001)and OA vs PA + OA (t = −5.353; P < 0.0001) regression lines are significantly different.

Gene expression

The expression of PPARalpha, gamma and SREBP-1 constantly increased following exposure to FAs. However, as shown in Figure 8, PA but not OA induced a significant increase in PPARα expression vs controls (P < 0.05). In contrast, incubation of HepG2 with OA but not with PA was associated, at both concentrations used, with an increase in the expression of PPARγvs controls (P < 0.05). In addition OA at the highest concentration increased the expression of SREBP-1 vs controls (P < 0.05). When the effect of two equimolar doses of PA and OA were compared, the difference did not reach statistical significance.

Figure 8.

Effect of palmitic acid (PA) and oleic acid (OA) on PPARγ, SREBP-1 and PPARα gene expression. Expression was evaluated by real time-polymerase chain reaction (PCR; for details see materials and methods). Data are reported as mean ± standard error of log2(RQ) normalized to the controls. *P < 0.05 vs control.

PA inhibits signaling downstream of the insulin receptor

Peripheral IR, is caused, at least in part, by blunted transmission of the signal along the enzymatic cascade downstream of the insulin receptor. A critical step in this pathway is represented by phosphorylation and activation of the serine/threonine kinase Akt, a central regulator of glucose uptake, anabolic metabolism and antiapoptotic signals.34 We compared the effects of steatosis induced by the different FAs, alone or in association, on insulin-induced phosphorylation of Akt (Fig. 9). Insulin caused a rapid and robust phosphorylation of Akt on Ser473, an activation-specific residue, in HepG2 cells. Steatosis induced by PA was associated with a marked decrease of insulin-induced Akt phosphorylation, while in cells treated with OA Akt phosphorylation was maintained. Interestingly, in cells exposed to both FA, an intermediate level of phosphorylation was detected. No differences were found among different fatty acid treatments in the activity of ERK1-2 (data not shown). Taken together, these data demonstrate that PA and OA differentially affect the early signaling pathways downstream of the insulin receptor.

Figure 9.

Effect of fatty acids (FAs) on Akt phosphorylation. Akt phosphorylation was evaluated in confluent, serum-starved (12 h) HepG2 were treated with FAs, after stimulation with 100 nM of insulin for 15′ and analyzed by Western blot using Phospho-Akt (Ser-473) Antibody.


In this in vitro model of steatogenesis, three different lines of hepatocytes were exposed to increasing doses of FAs for comparison purposes. The effects of administering OA, PA or their combination on TG accumulation, cell survival, FAs composition, gene expression and insulin signalling were explored. Data showed that while the extent of steatosis was more severe in cells treated with OA than in those exposed to PA, opposite effects of the two FAs were found on cell apoptosis. The greater steatosis extent observed following exposure to OA was associated with prominent PPARγ and SREBP-1 activation. In contrast, PA supplementation was associated with PPARα activation. Moreover, PA, but not OA, impaired insulin signalling. Despite the higher steatosis extent resulting from incubation with the two FAs combined, apoptosis rate and insulin signalling impairment were lower than in cells treated with PA alone, indicating a protective effect of OA.

As expected, in our study the accumulation of TGs was proportional to the concentration of FAs in the culture medium. However, OA was a more powerful trigger for TG accumulation than PA irrespective of the cell line we used. This is a novel finding for hepatocytes and may represent a general property of OA, which has been previously reported to induce a more marked fat accumulation also in other cell types, such as islet β cells.35 It is tempting to associate the more effective steatogenic property of OA with the specific pattern of gene expression (PPAR-gamma and SREBP-1) shown in the present study. In contrast, the lesser steatosis extent and increased apoptosis are due to PPAR-alpha activation which results in enhanced β-oxidation and thus oxidative stress.36

The most intriguing finding of the present study is the dissociation between the effects on steatosis and apoptosis of the different FAs. Steatosis per se induces cell apoptosis.14,15 Other studies have indicated a greater pro-apoptotic effect of PA as compared to OA, but few studies have examined the relationship between steatosis extent and apoptosis in hepatic cell cultures.20,28 The finding that OA is less efficient in the induction of apoptosis is in agreement with the recent work by Mahli et al. who, however, at variance with our study, found no differences in the steatogenic effect of OA and PA.20 This discrepancy might probably be accounted for by differences in the experimental conditions notably including the lower (0.2 mM in Mahli's vs 0.66 mM in our study) FA concentrations used.19 In this study, we have shown a correlation between TG accumulation and apoptosis with both OA and PA. However, the slope of this function was markedly different comparing the two FAs indicating that, for a given amount of TG accumulation, the effects on apoptosis were dramatically enhanced when cells were treated with PA (Fig. 6).

A large number of molecular mechanisms have been implicated in PA-associated apoptosis: ceramide production, NO synthesis, suppression of antiapoptotic factors such as Bcl-2,15 reactive oxygen species generation,37 endoplasmic reticulum stress;37,38 nuclear factor-kB activation39 and decreased synthesis of cardiolipin.40 In hepatocytes, steatosis may enhance the expression of inflammatory cytokines, such as TNF-α, and increase apoptosis via Fas receptor activation.16,19,27 A JNK stimulated mitochondrial20 and a reticulum stress-mediated9,21 apoptotic pathways have been also reported. Death receptors are cell surface proteins, which can trigger apoptosis when bound by their ligands. Ribeiro & Cortez-Pinto41 observed enhanced expression of the death receptor Fas in alcoholic hepatitis and tumor necrosis factor (TNF) receptor-1 in NASH. Others have identified enhanced Fas expression in alcoholic and nonalcoholic steatohepatitis16,42 Thus steatohepatitis may sensitize hepatocytes to extracellular death ligands (e.g. Fas ligand, tumor necrosis factor-alpha) promoting an extrinsic pathway to apoptosis.19,27 NF-κB is a transcription factor that can upregulate both death receptors and ligands.43,44 NF-κB activation in Kupffer cells or infiltrating monocytes is proinflammatory and induces the expression of death ligands such as TNFα.

FFAs can directly activate the IKK-κ/ NF-κB pathway in hepatocytes via a lysosomal, cathepsin B– dependent mechanism.27 This pathway involves the translocation of Bax to lysosomes with subsequent lysosomal destabilization and release of the cysteine protease, cathepsin B, into the cytosol. This subsequently leads to the activation of NF-κB, via IKK-κ, and a subsequent increase in the expression of TNFα apoptosis has increasingly been linked to inflammation and hepatic fibrogenesis.45–47

More recently, Gores' group48 reported OA to sensitise hepatocytes in vitro to to the death ligand TRAIL. Oleic acid led to upregulation of the cognate TRAIL receptor death receptor 5 whose expression was enhanced in steatotic human liver samples. DR5 was responsible for FFA sensitisation to TRAIL killing mediated by a JNK dependent mechanism. Moreover, as previously reported, if Kupffer cells are activated they produce death ligands, including Fas ligand, and TNFα which, in turn, induce apoptosis in hepatocytes expressing death receptors,49 this might be also a condition in NASH patients. Given that inhibiting apoptosis may well prevent the feared consequences of chronic liver disease, cirrhosis, portal hypertension, and liver failure thus drugs are being developed to block apoptosis.50

The cells' ability to readily incorporate FAs into cytoplasmic TGs might serve as a protection against their pro-apoptotic effects.51 The finding that OA resulted in a greater steatosis extent implies that this FA is more readily incorporated into TG and therefore is associated with less apoptosis than PA which was less steatogenic. Along these lines, the cytoprotective effect of OA in vitro could thus be explained by this FA's ability to promote channelling of PA into TG synthesis, as demonstrated in cultured fibroblasts.51 The toxic effect of PA on hepatocytic cell cultures could be, at least, in part related to the amount of FFA associated with incubation with this FA. However, given that this explanation is based on in vitro indirect data, our hypothesis needs to be verified by further studies given that a recent report disproves the paradigm that FFAs are increased in human NAFLD.52 In addition to PA channelling into TG synthesis, changes in the composition of the intracellular FAs pool may represent an alternative strategy to protect cells from PA-dependent apoptosis. The addition of PA to the medium resulted in an increased proportion of the less damaging palmitoleic acid indicating the attempt to ‘detoxify’ PA. This detoxification might be achieved through increased steroyl-CoA desaturase 1 (Δ9) activity (Fig. 8), which protects from PA-induced apoptosis.53

Our data provide in vitro evidence for the principle that TG accumulation might be a defence mechanism against the toxicity of excess FFAs. Yamaguchi et al. recently reached the same conclusion in the obese mouse model, wherein inhibited triglyceride synthesis was associated with improved hepatic steatosis though to the expenses of worsened hepatic inflammatory and fibrotic changes.54 Observations that simple steatosis is rarely progressive in humans55–57 further confirm that FAs stored as inert triglycerides are harmless unless associated to another injuring factor, such as IR, or oxidative stress, that may switch on the inflammatory cascade. Familial heterozygous hypobetalipoproteinemia represents a naturally occurring model of human simple steatosis due to genetically impaired capacity to export VLDL from the hepatocytes into the bloodstream. Of interest, in the absence of IR, steatosis in these individuals, although massive, does not appear to be progressive.58 Studies have shown that PA impairs insulin signalling via increased JNK activity leading, in turn, to phosphorylation of insulin receptor substrate-1 and 2 at the inhibitory sites eventually leading to IR.59 Increased circulating FFAs concentrations represent an early correlate of peripheral IR, which may be demonstrated to occur before the development of obesity in experimental condition.12 In this connection, our data have shown that PA administration, in itself, impaired insulin signalling while OA partially restored it. This finding implies that the steatosis extent is not the only, nor perhaps the major, determinant of impaired insulin action. Indeed, rat models7,9 and human studies5,6,8 consistently support the concentrations and the chemical structure of FAs to be fundamental players in the development and progression of NAFLD. In human NAFLD, FAs that are incorporated in hepatic TG are mostly derived from the circulating NEFA pool and thus reflect the adipose tissue FAs composition.60 Individuals consuming western-type high-fat diets show a serum/tissue predominance of PA with a relative reduction in OA.6,10,11 In principle, characterization of FAs profile in the serum/liver may assist in predicting those individuals with progressive/severe NAFLD.20 Our data also highlight that variations in the desaturase activity might have a role in the type of FAs stored in hepatic TG and thus in the pathogenesis of human NAFLD. Conversely, PA deprivation and OA enrichment in diet might be a goal to prevent/treat the disease.

In conclusion, while a limitation of the current study is the use of three hepatocyte cultures of different origin but no data on primary hepatocyte culture, our study showing that OA is more steatogenic but less damaging than PA in hepatocyte cell cultures may provide a biological clue useful to a better understanding of animal models of NAFLD. As far as human NAFLD is concerned, our data might suggest two major clinical implications/research hypothesis. First, for the diagnosis of NASH, steatosis extent evaluated histologically might not be as relevant as its chemical composition. Second, in agreement with recent studies envisaging for saturated FAs a role of ‘second hit’,61 the chemical composition of steatosis might help differentiate those non-progressive (‘inert’) forms of NAFLD from those that, due to their enrichment in saturated FAs, are at a substantial risk of progression. These hypotheses are worthy being tested in specific studies.


We thank Dorval Ganazzi for the assistance in statistical analysis. We are also grateful to Dr. Cristiana Bertolani and Nadia Navari, coworkers of Prof. Marra, for their contribution.

Part of these data have been presented in abstract form at the European Association for the Study of the Liver, Wien, April 2006.