Inhibition of exendin-4-induced steatosis by protein kinase Ain cultured HepG2 human hepatoma cells

Alice Y. Chen-Liaw1,2& Gabrielle Hammel1 & George Gomez1,3

Received: 27 April 2017 /Accepted: 20 June 2017 /Published online: 13 July 2017 / Editor: Tetsuji Okamoto# The Society for In Vitro Biology 2017

Abstract Nonalcoholic fatty liver is characterized by the ab-normal accumulation of triglycerides within hepatocytes,resulting in a steatotic liver. Glucagon-like peptide 1 and itsanalog exendin-4 can ameliorate certain aspects of this syn-drome by inducing weight loss and reducing hepatic triglyc-eride accumulation, but it is unclear whether these effectsresult from the effects of glucagon-like peptide 1 on the pan-creas, or from direct action on the liver. This study investigatedthe direct action and putative cellular mechanism of exendin-4on steatotic hepatocytes in culture. Steatosis was induced incultured HepG2 human hepatoma cells by incubation in me-dia supplemented with 2 mM each of linoleic acid and oleicacid. Steatotic hepatocytes were then pre-incubated in the pro-tein kinase A inhibitor H89 for 30 min, then treated withexendin-4 over a period of 24 h. Cell viability and triglyceridecontent were characterized by a TUNEL assay and AdipoRedstaining, respectively. Our results showed that steatotic cellsmaintained high levels of intracellular triglycerides (80%)compared to lean controls (25%). Exendin-4 treatment causeda significant reduction in intracellular triglyceride content after12 h that persisted through 24 h, while protein kinase A in-hibitors abolished the effects of exendin-4. The results dem-onstrate the exendin-4 induces a partial reduction in triglycer-ides in steatotic hepatocytes within 12 h via the GLP-1 recep-tor-mediated activation of protein kinase A. Thus, the

reduction in hepatocyte triglyceride accumulation is likelydriven primarily by downregulation of lipogenesis and upreg-ulation of β-oxidation of free fatty acids.

Keywords Hepatocyte . GLP-1 receptor . Intracellulartriglycerides . AdipoRed assay . Cell culture

Introduction

Nonalcoholic fatty liver disease (NAFLD) is a common andserious form of chronic liver disease that is strongly associatedwith obesity, insulin resistance, and type II diabetes mellitus(Angulo 2002). In diabetic conditions, insulin insensitivity inadipocytes may impair insulin-mediated suppression of lipol-ysis leading to an increase of circulating free fatty acids (FFA),which are subsequently metabolized by the liver (Jacome-Sosa and Parks 2014), which re-esterifies a portion of theFFAs as triglyceride (TG)-rich lipoproteins. Consequently, amain histological manifestation of NAFLD is the accumula-tion of TG within hepatocytes in the presence of excess TGdietary loads (Portillo-Sanchez and Cusi 2016).

Regulation of insulin-based TG metabolism in the liver ismediated by two mechanisms: (1) the phosphorylation of theforkhead box O1 (FoxO1) transcription factor, which preventsits entry into the nucleus and thus downregulates the genesrequired for gluconeogenesis, resulting in a decrease in glu-cose output from the liver (Matsumoto et al. 2007); and (2) theactivation of sterol receptor binding protein 1-c (Utzschneiderand Kahn 2006), which stimulates gene transcription leadingto TG synthesis and accumulation in hepatocytes (Cawanoand Cohen 2013, Ress and Kaser 2016). Thus, in type IIdiabetes, while dietary glucose increases insulin secretion,the downregulation of FoxO1 sensitivity leads to sustainedgluconeogenesis and increased glucose output by the liver

* George Gomezgeorge.gomez@scranton.edu

1 Biology Department, University of Scranton, 800 Linden Street,Scranton, PA 18510, USA

2 Present address: Mt. Sinai School of Medicine, New York, NY, USA3 Loyola Science Center 395, 204 Monroe Ave, Scranton, PA 18510,

USA

In Vitro Cell.Dev.Biol.—Animal (2017) 53:721–727DOI 10.1007/s11626-017-0181-y

(Khan et al. 2016). In contrast, insulin continues to activate thesterol receptor binding protein 1-c (Moon et al. 2014), leadingto accelerated FFA synthesis and TG accumulation in the liver.Collectively, these result in the classic type II diabetic triad ofsymptoms: hypertriglyceridemia, hyperinsulinemia, and hy-perglycemia (Brown and Goldstein 2008).

Pharmacological treatments have therefore targeted hepaticinsulin sensitivity. Targets such as the glucagon-like peptide 1(GLP-1) secreted from intestinal L cells have been shown toboth stimulate insulin secretion in the pancreas and increasehepatic insulin sensitivity (Taher et al. 2014). Since the pres-ence of GLP-1 receptors (GLP-1R) in hepatocytes has recent-ly been confirmed (Gupta et al. 2010), pharmacological agentstargeting these receptors present a potential therapeutic targetfor ameliorating NAFLD. One such compound is exendin-4(Göke et al. 1993), a GLP-1 mimetic that has a half-life ofapproximately 3–4 h, which makes it possible for clinicaltreatments that require the compound to persist in the systemfor extended periods of time. Other compounds such asliraglutide (a human GLP-1 analog) have a half-life of approx-imately 11–15 h (Armstrong et al. 2016) and show similarpharmacological effects to exendin-4. Obese mice treated withexendin-4 or other GLP-1R agonists show weight loss anddecreased hepatic TG accumulation (Ding et al. 2006).However, in these studies, it was not clear whether the weightloss was induced by improved NAFLD and liver function orvice versa. Therefore, our studies focused on the direct actionof exendin-4 on hepatocyte TG accumulation in steatotic he-patocytes in culture at the single-cell level in isolation from thebroader spectrum of systemic interactions in vivo.

We focused our studies on the action of exendin-4 at thesingle-cell level and tested whether exendin-4 acts directly onhepatocytes and attempted to identify the cellular pathway me-diating exendin-4-induced GLP-1 activity. We used HepG2 celllines, the most commonly used hepatocyte cell model (Guguen-Guillouzo and Guillouzo 2010), and measured intracellular TGaccumulation. GLP-1 has been shown to be expressed (Guptaet al. 2010) and glucagon is known tomodulate gene expression(Thonpho et al. 2010) and regulate hepatic lipid metabolism inthis cell line (Wang et al. 2016), making this an ideal modelsystem for study. Since the canonical pathwaymediating GLP-1activity involves adenylyl cyclase and cAMP (Baggio andDrucker 2007), we used protein kinase inhibitors to assess thepossible involvement of this cellular pathway in GLP-1 signal-ing. We hypothesized that exendin-4 would reduce TG contentin steatotic hepatocytes via the cyclic adenosinemonophosphate(cAMP) and protein kinase A (PKA) signaling pathway.

Methods and Materials

Materials Unless otherwise indicated, all materials were pur-chased from Sigma-Aldrich, St. Louis, MO. Cell culture

media consisted of Iscove’s Modified Dulbecco’s Mediumwith 10% fetal bovine serum and 1% penicillin-streptomycin.Stock solutions were prepared as follows: Exendin-4 (Abcam,Cambridge, MA) was mixed at 10 μM in cell culture media;H89 dihydrochloride (LC Laboratories, Woburn, MA) wasprepared as a 2 mM stock in DMSO; 8-Bromoadenosine3′:5′-cyclic monophosphate (Bromo-cAMP) was mixed as a100 μM stock in cell culture media. All stock solutions werestored in 20 μl aliquots, stored at −20°C, and thawed andmixed to the working concentrations immediately prior to use.

Cell cultures and validationHepG2/C3A (contact-inhibited)cells were purchased from American Type Culture Collection(Manassas, VA) and cultured at 37°C with 5% CO2. For eachrun of experiments, cells were thawed from frozen stocks,passaged twice, and used immediately. Possible mycoplasmacontamination was detected using Hoechst-33342 (Invitrogen,Carlsbad, CA) in conjunction with cell viability assays (seebelow) as described previously (Battaglia et al. 1994); cellsthat showed signs of possible mycoplasma contaminationwere discarded and not included in the study.

For each study, cells were passaged into six-well plates;after 1 d in vitro, culture media were completely replaced witheither fresh media (control) or culture medium supplementedwith 2 mM oleic acid, 2 mM linoleic acid, and 4% (W/V)bovine serum albumin (Bsteatotic^) to induce steatosis (Nativet al. 2014). Cells were cultured two more d. Cells werethen incubated in their respective media supplemented with10 nM exendin-4 in the presence or absence of 2 μMH89 foran additional 1, 3, 6, 12, or 24 h. For assays involving directactivation of PKA, steatotic cells were incubated with 100 nMBromo-cAMP at t = 0. At each time point, cells were fixedwith 4% paraformaldehyde for 10 min, then rinsed thrice withphosphate buffered saline (PBS) for 10 min each prior to eachassay.

Viability and apoptosis assays To measure cell viability, fol-lowing the treatments and prior to fixation, cells were washedwith PBS and dead cells were quantified following incubationfor 20 min at 37°C in 4.29 μg/ml ethidium homodimer-1(Ethd-1; Invitrogen) to stain the nuclei of dead cells, and1 μg/ml Hoechst-33342 (Invitrogen) to stain the nuclei of allcells (Nativ et al. 2014). Dead cells were visually identifiedwith Ethd-1-labeled (red) nuclei, while live cells were visuallyidentified with Hoechst-stained (blue) nuclei under fluores-cence illumination. In addition, the media were monitoredfor floating Ethd-1-labeled cells to identify possible dead cellsthat detached from the substrate. The viability was qualitative-ly compared for cells under various steatosis-reducingconditions.

Apoptosis was determined using the In Situ Cell DeathDetection Kit, Fluorescein (Roche Diagnostics, Indianapolis,IN) following the manufacturer’s instructions. Briefly, cells

722 CHEN-LIAW ETAL.

were permeabilized in PBS + Triton X100 (0.5%) for 15 min.Cells were then washed with PBS, incubated in the TUNELreaction mixture (TUNEL reaction enzyme TdT + TUNELlabeling solution fluorescein-dUTP, 1:10) for 1 h, washedthrice with PBS, and mounted on slides using Vectashield(Vector Labs, Burlingame, CA).

Fluorescent lipophilic AdipoRed™ staining and data anal-ysis Cells were stained with 1 μg/ml of 4′,6-diaminidino-2-phenylindole (DAPI) followed by incubation in AdipoRed™Assay Reagent (Lonza, Walkersville, MD) stain for 20 min.Cells were then washed three more times in PBS and mountedon slides using Vectashield.

AdipoRed-stained cells were viewed under oil immersion at×100 using a Nikon Eclipse E400 epifluorescence microscopeand photographed with a Hamamatsu Orca digital camera.Slides were viewed and photographed using the following fil-ters: B-2E/C (FITC, 465–495 nm excitation; 515–555 nm emis-sion) to visualize intracellular TG droplets as well as forTUNEL staining; G-2E/C filter (TRITC, 528–553 nm excita-tion, 600–660 nm emission) to visualize total cellular lipids(including intracellular lipid droplets, hydrophobic lipidmembranes of vesicles, and the plasma membrane; Greenspanet al. 1985). Cells were concurrently photographed under phasecontrast for visual reference. Exposure time (200 ms) and exci-tation light intensity were held constant for all conditions. Atotal of five fields of view per slide were taken.

Staining was analyzed using ImageJ (National Institutes ofHealth, Bethesda, MD) to quantify relative percentage of in-tracellular TG in response to each treatment. Imagesphotographed under the FITC (to quantify intracellular TG)and TRITC (to quantify total cellular lipids) filters wereanalyzed as follows: the total pixel count for all pixelsat intensities above a background (determined using thethreshold function in ImageJ) was measured for each FITCor TRITC image. The area covered by the lipids was deter-mined using the total number of suprathreshold pixels underFITC or TRITC illumination. Since our pilot studies showedthat overall lipid accumulation was proportional to the size ofthe lipid droplets within each hepatocyte, TG accumulationfor each cell was determined using the ratio of the overall areaTG accumulation (total number of pixels under FITC illumi-nation) to the area of total lipids within each cell (total numberof pixels under TRITC illumination). Values were averagedwithin each treatment, compared between individual treat-ments and time points using t tests (p < 0.05). Statistics werecalculated using GraphPad (LaJolla, CA).

Results

Cells were plated at 15% confluence and typically grew to65–70% confluence by the end of the trials (5 d). HepG2

cells attained their characteristic morphology and cell den-sity (Fig. 1A, C, E); under phase contrast illumination, TGdroplets were readily visible. Addition of lipids or treatmentwith pharmacological agents did not result in significantchanges in cell density (determined by cell counts); thus,all data analyses were conducted without adjustment forcell number. In conjunction with each treatment, lean andsteatotic hepatocytes were stained with Ethd-1 andHoechst-33342 to monitor viability. The lean hepatocytes(negative control) had a very small proportion of dead cells(<1%) while the steatotic hepatocytes (positive control)contained the highest percentage of dead cells (1%); however,these rates were not significantly different from each other.Visual examination showed no dead cells floating in the me-dia, suggesting that both the live and dead cells were sub-strate-adherent.

TUNEL staining for apoptotic nuclei (Fig. 1B, D, F)showed that untreated cells had a low rate of cell death(0.46%). Induction of steatosis resulted in a significant in-crease in cell death rates (2.56%, T8 = 5.003, p = .0007), andco-incubation of cells in steatotic media and exendin-4 re-duced the cell death rate to a value (1.03%; T10 = 1.4909,p = .1669) that was not significantly different from that ofthe control.

When stained with AdipoRed, TG accumulation in the he-patocytes was readily quantified (Fig. 2). The visually distinctTG granules fluoresced bright green and showed little stainingin lean cells (Fig. 2A) and extensive staining in steatotic cells(Fig. 2B). Under red fluorescence, the total lipid content (in-tracellular compartments and vesicles) of the cells was visible.In steatotic cells, the total lipid content increased compared tothat of the lean cells (Fig. 2B); to control for this overall in-crease, the proportion of TG (green fluorescence) to total lipidcontent (red fluorescence) was therefore used to determineintracellular TG.

Hepatocytes grown in lipid-enriched media showed in-creased intracellular TG accumulation (Fig. 3). Control(lean) and steatotic cells maintained intracellular TG contentsof 25 and 77%, respectively, of overall cellular lipids. We areshowing data from three different concentrations of exendin-4: 1, 10, and 100 nM. Steatotic cells treated with 1 nMexendin-4 did not show any changes in TG content, whilesteatotic cells treated with 10 or 100 nM exendin-4 showeda decrease in intracellular TG from 77 to 48% after 6 h; thesevalues did not change over time and were significantly differ-ent (T51 = 3.22, p = .0022) from those of the control, untreatedsteatotic cells, and steatotic cells treated with 1 nM exendin-4.Concentrations lower than 1 nM did not show any effect, andconcentrations higher than 100 nM did not show a more ex-acerbated effect (data not shown for clarity). There was nosignificant difference between 10 and 100 nM exendin-4 treat-ments. Thus, we used 10 nM exendin-4 for the subsequentstages of our study.

PKA AND STEATOSIS IN HEPG2 723

To investigate the signaling pathway through whichexendin-4 induces the reduction of intracellular triglycerides,we used Bromo-cAMP and H89 for protein kinase activationand inhibition, respectively. Steatotic hepatocytes were pre-incubated with the agents for 30 min before the addition ofexendin-4; TG accumulation was measured as described pre-viously. After 1–3 h, none of the treatments resulted in a sig-nificant change in intracellular TG content (Fig. 4), suggestingthat the treatments did not induce rapid changes in lipid me-tabolism. After 6 h onwards, exendin-4-treated steatotic hepa-tocytes that were pre-incubated with H89 were significantlydifferent from cells treated with exendin-4 only (T51 = 3.488;p = 0.001) and not significantly different from untreatedsteatotic cells (T51 = 1.526; p = 0.1332). Steatotic cells treatedwith Bromo-cAMP only did not show a significantly differentTG accumulation rate compared with untreated steatotic cells(T51 = 0.63; p = 0.53), suggesting the need for the activation ofPKA by the GLP-1 receptor; and thus, there are other elements

of the cAMP signal transduction cascade needed to induce thereduction in intracellular TG. Collectively, these results sug-gest that the effects exendin-4 are mediated primarily but notexclusively through a signaling cascade involving PKA.

Discussion

It has been proposed that one of the effects of exendin-4 isactivation of downstream signaling elements involved in theregulation of fatty acid synthesis (Gupta et al. 2010). Studieson lipogenesis regulation by GLP-1 report contradictory find-ings: GLP-1 has been shown to promote hepatic accumulationof glycogen and reduce cAMP (Lopez-Delgado et al. 1998).In contrast, the GLP-1R agonist exendin-4 was shown to sup-press hepatic lipogenesis in isolated hepatocytes via increasedcAMP levels (Ding et al. 2006). Our results suggest that thereduction of TG by exendin-4 was blocked by PKA inhibitors(Fig. 4), supporting the notion that exendin-4 acts through thecAMP/PKA pathway, the downstream targets of which in-clude enzymes for fatty acid synthesis, gluconeogenesis, andcholesterol synthesis (Horton and Goldstein 2002). Our stud-ies with the PKA activator Bromo-cAMP show that exclusiveactivation of PKA does not result in a reduction in TG.Together, these results suggest that while exendin-4 actsthrough a signaling cascade involving PKA, other elementsof the cAMP signaling transduction cascade are necessary toinduce a reduction in TG. What is likely driving the TG re-duction is exendin-4-induced downregulation of TG synthesisand upregulation of β-oxidation.

Since our study focused on hepatocytes isolated from therest of the somatic function, we could not account for anychanges in glucose metabolism, which could be a salient fac-tor in overall TG metabolism. In pancreatic islet β cells, it iswell-established that GLP-1 exerts incretin effects via thecAMP/AMPK pathway (Ben-Shlomo et al. 2011).pAMPKα is a potent suppressor of lipogenesis and is a down-stream target of PKA in pancreatic islet β cells and hepato-cytes (Ben-Shlomo et al. 2011). Exendin-4 has also beenshown to increase the amount of pAMPKα and GLP-1RmRNA in both mouth hepatocytes in vivo and HepG2in vitro in a dose-dependent manner, exerting its maximaleffect at 100 nM (Lee et al. 2012). Our study shows that lowdoses (10 to 100 nM) of exendin-4 can still exert a significanteffect on TG content (Fig. 3), perhaps reflecting the method-ological differences between in vitro and in vivo approaches.

We acknowledge that there are limitations to in vitromodels such as HepG2, and that our study is clearly limitedin scope to cellular and subcellular functions in the absence ofsystemic modulation. Lipid metabolism in vivo is a complexprocess involving several cell types from multiple tissuesources that involve complex hormone interactions; thus, theeffects we observed at the single-cell level may differ when

Figure 1. Sample phase-contrast photomicrographs of HepG2 cellsshow that visible TG accumulation changes with different cultureconditions. All images shown are after 24 h treatment. (A) Lean(untreated) hepatocytes show characteristic morphology with fewvisible phase-dark globules (black arrow) that denote intracellular TG.(B) TUNEL staining shows a low number of apoptotic cells; in our study,the average cell death was 0.46%. (C) Steatotic hepatocytes had a visuallydistinct increase in TG accumulation. (D) While steatotic treatment didnot result in an increase in cell numbers, there was a marked (though notstatistically significant) increase in apoptosis (2.56%). (E) Steatotic cellstreated with exendin-4 did not show a marked visual difference inintracellular from untreated cells. (F) Apoptosis resulting from steatotictreatments coupled with excendin-4 incubations was not significantlydifferent (1.03%) from controls. Scale bar = 50 μm.

724 CHEN-LIAW ETAL.

Figure 2. AdipoRed stainingallowed for visualization andquantification of intracellular TG.All images shown are 2 h post-treatment; leftmost panels showthe phase-contrast imagecorresponding to green(visualizing intracellular TG,middle panels) and red(visualizing all intracellular lipids,rightmost panels) fluorescenceillumination. TG accumulationwas measured by quantifying theratio of green staining to redstaining as described in themethods. (A) Lean cells showedlittle TG accumulation. (B)Steatotic cells had a higheramount of intracellular TG and anoverall visible (but notsignificant) increase in totalcellular lipids. (C) Steatotic cellstreated with exendin-4 showed adecrease in intracellular TG andan overall reduction in cellularlipids.

Figure 4. Activation of GLP-1wasmediated by protein kinase A. HepG2cells were cultured under different lipid conditions, stained, andquantified as in Fig. 3. The data for lean, untreated steatotic, and 10 nMexendin-4 are identical to those shown in Fig. 3 and are shown forcomparison. For clarity of display only, the time points on the X-axis ofBEx4 + H89^ and BBr cAMP^ are offset. Incubation with exendin-4 withthe simultaneous inhibition of protein kinase A with H89 (+Ex 4 andH89) resulted in the abolition of the effect of exendin-4 (+Ex 4),suggesting the involvement of protein kinase A in the GLP-1-mediatedsignaling cascade. However, the direct activation of protein kinase AwithBromo-cAMP (+Br cAMP) in the absence of exendin-4 did not result in asignificant difference from steatotic treatments, suggesting theinvolvement of other elements of the cAMP-medicated signaltransduction cascade in steatosis.

Figure 3. Exendin-4 at 10 nMwas effective at reducing intracellular TG.HepG2 cells were cultured under lean, steatotic, or steatotic supplementedwith different concentrations of exendin-4 as described in the methods. Atfixed time points (1, 3, 6, 12, and 24 h), a subset of cells was fixed andstained with AdipoRed to measure the proportion of intracellular TGcompared to total cell lipid as shown in Fig. 2. For clarity of displaypurposes only, the time points on the X-axis for BEx4 100 nM^ andBsteatotic^ are offset. Lean (untreated) cells maintained a low level ofintracellular TG of about 25–30% of total cell lipids, while steatoticcells maintained a high proportion of intracellular TG of about 75–80%over the 24-h trial. When treated with 10 or 100 nM exendin-4, thesteatotic cells showed a steady decline in intracellular TG that wassignificantly different (marked by asterisks) at 12 h onwards from bothlean and steatotic treatments. Incubation of steatotic cells with 1 nMexendin-4 did not show a significant decrease in TG accumulation,suggesting that this concentration of exendin-4 was ineffective. Thus,we selected 10 nM exendin-4 for the subsequent procedures.

PKA AND STEATOSIS IN HEPG2 725

applied to the whole organism. While this study focused onelucidating the effects of exendin-4 on HepG2 TG accumula-tion in vitro, there is ample evidence that these are broadlyapplicable at the whole organism level. Transgenic obese micetreated with exendin-4 for 60 d demonstrated a reduction innet weight gain, reduced serum glucose and hepatic steatosis,and improved insulin sensitivity (Ding et al. 2006). In high-fat(HF) diet-induced obese mice, treatment with exendin-4 re-sulted in decreased body weight, decreased serum FFA, andimproved HF-induced hepatic TG accumulation and inflam-mation (Lee et al. 2012). Histological staining showed thereduction in hepatic TG in obese mouse livers (Lee et al.2012). However, it is not clear from these studies whetherNAFLD improvement resulted fromweight loss or vice versa.Together, this study shows that exendin-4 exerts a significanteffect on hepatocyte TG accumulation (Figs. 3 and 4) evenwhile it might induce incretin effects in in vivo models (Ben-Shlomo et al. 2011).

GLP-1R agonists have been suggested as potential treat-ments for patients with type II diabetes mellitus and NAFLD,especially for their pleiotropic effects, which include incretineffects, changes in lipid metabolism, suppression of appetite,and improvement of insulin sensitivity in hepatocytes andadipocytes (Portillo-Sanchez and Cusi 2016). In a recent studythat examined liraglutide (a human GLP-1 receptor analogsimilar to exendin-4) effects and action in NAFLD, 39% ofnonalcoholic steatohepatitis patients treated for 48 weeks withliraglutide (1.8 mg/d) exhibited a reversal of nonalcoholicsteatohepatitis, which was defined in this study as the disap-pearance of hepatocyte ballooning (Armstrong et al. 2016).Patients also experienced a significant reduction in bodyweight and plasma glucose levels, reflecting studies in mousemodels. Hepatocytes were also shown to have the improvedability to handle FFA, de novo lipogenesis, lipid transport, andβ-oxidation, which may serve to slow or reverse the patho-genesis of NAFLD.

In addition to global metabolic shifts, insulin resistance inadipocytes and hepatocytes is another contributing factor toNAFLD. As a result of liraglutide treatment, insulin sensitiv-ity was found to increase in hepatocytes but not in adipocytes.However, these improvements were modest and confoundedwith weight loss (Armstrong et al. 2016). The experiments inthis present study provide additional evidence that exendin-4directly affects hepatic TG accumulation but do not examineinsulin resistance and whether exendin-4 had any influence onthat. Future research should focus on the relationship betweenhepatic lipid metabolism, weight loss, and insulin sensitivity.An approach that considers the cellular manifestations of typeII diabetes as well as NAFLD pathogenesis would contributegreatly to understanding mechanisms that underlie the actionof GLP-1R agonists, such as exendin-4 or liraglutide, and torealizing their therapeutic potential for treating NAFLD andcertain symptoms of type II diabetes.

Conclusion

The results demonstrate exendin-4 induces a partial reductionin triglycerides in steatotic hepatocytes within 12 h via theGLP-1 receptor-mediated activation of protein kinase A.This mechanism operating at the single-cell level may contrib-ute to the overall downregulation of lipogenesis and upregu-lation of β-oxidation of free fatty acids in vivo.

Acknowledgements This work was pioneered by AC as part of theCellular Bioengineering Summer Undergraduate Research Program atRutgers University (supported by the NSF EEC 1262924). This workwas supported by the Ellen Miller Casey Award to AC, Research as aHigh Impact Practice student funding program awarded to GH, and in-ternal funds from the University of Scranton awarded to GG.

References

Angulo P (2002) Nonalcoholic fatty liver disease. NE J Med 346:1221–1231

Armstrong MJ, Gaunt P, Aithal GP, Barton D, Hull D, Parker R,Hazlehurst JM, Guo K, Abouda G, Aldersley MA, Stocken D(2016) Liraglutide safety and efficacy in patients with non-alcoholic steatohepatitis (LEAN): a multicentre, double-blind,randomised, placebo-controlled phase 2 study. Lancet 387:679–690

Baggio LL, Drucker DJ (2007) Biology of incretins: GLP-1 and GIP.Gastroenterol 132:2131–2157

Battaglia M, Pozzi D, Crimaldi S, Parasassi T (1994) Hoechst 33258staining for detecting mycoplasma contamination in cell cultures: amethod for reducing fluorescence photobleaching. BiotechHistochem 69:152–156

Ben-Shlomo S, Zvibel I, Shnell M, Shlomai A, Chepurko E, Halpern Z,Barzilai N, Oren R, Fishman S (2011) Glucagon-like peptide-1 re-duces hepatic lipogenesis via activation of AMP-activated proteinkinase. J Hepatol 54:1214–1223

Brown MS, Goldstein JL (2008) Selective versus total insulin resistance:a pathogenic paradox. Cell Metab 7:95–96

Cawano Y, Cohen DE (2013) Mechanisms of hepatic triglyceride accu-mulation in non-alcoholic fatty liver disease. J Gastroenterol 48:434–441

Ding X, Saxena NK, Lin S, Gupta NA, Anania FA (2006) Exendin-4, aglucagon-like protein-1 (GLP-1) receptor agonist, reverses hepaticsteatosis in ob/ob mice. Hepatology 43:173–181

Göke R, Fehmann HC, Linn T, Schmidt H, Krause M, Eng J, Göke B(1993) Exendin-4 is a high potency agonist and truncated exendin-(9-39)-amide an antagonist at the glucagon-like peptide 1-(7-36)-amide receptor of insulin-secreting beta-cells. J Biol Chem 268:19650–19655

Greenspan P, Mayer EP, Fowler SD (1985) Nile red: a selective fluores-cent stain for intracellular lipid droplets. J Cell Biol 100:965–973

Guguen-Guillouzo C, Guillouzo A (2010) General review on in vitrohepatocyte models and their applications. In: Maurel P (ed)Hepatocytes: methods and protocols. Springer, New York, pp 1–40

Gupta NA, Mells J, Dunham RM, Grakoui A, Handy J, Saxena NK,Anania FA (2010) Glucagon-like peptide receptor is present on hu-man hepatocytes and has a direct role in decreasing hepatic steatosisin vitro by modulating element of the insulin signaling pathway.Hepatology 51:1584–1592

Horton JD, Goldstein JL (2002) SREBPs: activators of the completeprogram of cholesterol and fatty acid synthesis in the liver. J ClinInvest 109:1255–1131

726 CHEN-LIAW ETAL.

Jacome-Sosa M, Parks EJ (2014) Fatty acid sources and their fluxes asthey contribute to plasma triglyceride concentrations and fatty liverin humans. Curr Opin Lipidol 25:213–220

Khan S, Kumar S, Jena G (2016) Valproic acid reduces insulin-resistance,fat deposition and FOXO1-mediated gluconeogenesis in type-2 di-abetic rat. Biochimie 125:42–52

Lee J, Hong SW, Chae SW, KimDH, Choi JH, Bae JC, Park SE, Rhee EJ,Park CY, Oh KW, Park SW (2012) Exendin-4 improvessteatohepatitis by increasing Sirt1 expression in high-fat diet-in-duced obese C57BL/6J mice. PLoS One 7:e31394

López-Delgado MI, Morales M, Villanueva-Peñacarrillo ML, MalaisseWJ, Valverde I (1998) Effects of glucagon-like peptide 1 on thekinetics of glycogen synthase A in hepatocytes from normal anddiabetic rats. Endocrinol 139:2811–2817

MatsumotoM, Pocai A, Rossetti L, DePinho RA, Accili D (2007) Impairedregulation of hepatic glucose production in mice lacking the forkheadtranscription factor Foxo1 in liver. Cell Metabol 6:208–216

Moon YA, Ochoa CR, Mitsche MA, Hammer RE, Horton JD (2014)Deletion of ELOVL6 blocks the synthesis of oleic acid but doesnot prevent the development of fatty liver or insulin resistance. JLipid Res 55:2597–2605

Nativ NI, Yarmush G, So A, Barminko J, Maguire TJ, Schloss R,Berthiaume F, Yarmush ML (2014) Elevated sensitivity of

macrosteatotic hepatocytes to hypoxia/reoxygenation stress is re-versed by a novel defatting protocol. Liver Transpl 20:1000–1011

Portillo-Sanchez P, Cusi K (2016) Treatment of nonalcoholic fatty liverdisease (NAFLD) in patients with type 2 diabetes mellitus. ClinDiabet Endocrinol 2:9

Ress C, Kaser S (2016) Mechanisms of intrahepatic triglyceride accumu-lation. World J Gastroenterol 22:1664–1673

Taher J, Baker CL, Cuizon C, Masoudpour H, Zhang R, Farr S, NaplesM, Bourbon C, Pausova Z, Adeli K (2014) GLP-1 receptor agonismameliorates hepatic VLDL overproduction and de novo lipogenesisin insulin resistance. Molec Metabol 3:823–833

Thonpho A, Sereeruk C, Rojvirat P, Jitrapakdee S (2010) Identification ofthe cyclic AMP responsive element (CRE) that mediates transcrip-tional regulation of the pyruvate carboxylase gene in HepG2 cells.Biochem Biophys Res Comm 393:714–719

Utzschneider KR, Kahn SE (2006) The role of insulin resistance in non-alcoholic fatty liver disease. J Endocrin Metab 91:4753–4761

WangH, ZhaoM, SudN, Christian P, Shen J, SongY, Pashaj A, ZhangK,Carr T, SuQ (2016) Glucagon regulates hepatic lipid metabolism viacAMP and Insig-2 signaling: implication for the pathogenesis ofhypertriglyceridemia and hepatic steatosis. Sci Rep 6:32246

PKA AND STEATOSIS IN HEPG2 727