In the early stages of nonalcoholic fatty liver disease (NAFLD), triglycerides accumulate in hepatocytes. Diacylglycerol acyltransferase 2 (DGAT2) catalyzes the final step in hepatocyte triglyceride biosynthesis. DGAT2 antisense oligonucleotide (ASO) treatment improved hepatic steatosis dramatically in a previous study of obese mice. According to the 2-hit hypothesis for progression of NAFLD, hepatic steatosis is a risk factor for nonalcoholic steatohepatitis (NASH) and fibrosis. To evaluate this hypothesis, we inhibited DGAT2 in a mouse model of NASH induced by a diet deficient in methionine and choline (MCD). Six-week-old genetically obese and diabetic male db/db mice were fed either the control or the MCD diet for 4 or 8 weeks. The MCD diet group was treated with either 25 mg/kg DGAT2 ASO or saline intraperitoneally twice weekly. Hepatic steatosis, injury, fibrosis, markers of lipid peroxidation/oxidant stress, and systemic insulin sensitivity were evaluated. Hepatic steatosis, necroinflammation, and fibrosis were increased in saline-treated MCD diet–fed mice compared to controls. Treating MCD diet–fed mice with DGAT2 ASO for 4 and 8 weeks decreased hepatic steatosis, but increased hepatic free fatty acids, cytochrome P4502E1, markers of lipid peroxidation/oxidant stress, lobular necroinflammation, and fibrosis. Progression of liver damage occurred despite reduced hepatic expression of tumor necrosis factor alpha, increased serum adiponectin, and striking improvement in systemic insulin sensitivity. Conclusion: Results from this mouse model would suggest accumulation of triglycerides may be a protective mechanism to prevent progressive liver damage in NAFLD. (HEPATOLOGY 2007.)
Nonalcoholic fatty liver disease (NAFLD) is one of the most common liver diseases in the world.1–3 Clinically, the most common disorder associated with NAFLD is insulin resistance.4 Accumulation of triglycerides in hepatocytes is the hallmark of NAFLD. Recent studies have demonstrated that acyl-coenzyme A:diacylglycerol acyltransferase 2 (DGAT2) plays an important role in hepatocyte triglyceride synthesis and hepatic steatosis.5, 6 Triglyceride synthesis is increased in the fatty livers that accompany obesity and type 2 diabetes in humans and mice.7 According to the 2-hit hypothesis for NAFLD progression, hepatic steatosis is a risk factor for nonalcoholic steatohepatitis (NASH) and fibrosis.8, 9 In a previous study, knocking down DGAT2 in the livers of mice with diet-induced obesity (DIO) and diabetes successfully prevented hepatic steatosis.10 Because only mild NASH and little fibrosis develop in mice with DIO, however, that model is not helpful for determining if inhibiting steatosis prevents progression of NAFLD to more advanced stages of liver damage (i.e., NASH and liver fibrosis).
To address this question, we studied a recently described model of progressive obesity-related NASH in db/db mice.11, 12 Db/db mice spontaneously develop obesity, severe type 2 diabetes, and fatty livers as a result of a functional defect in the long form of the leptin receptor, which plays a key role in the regulation of food intake, insulin sensitivity, and control of body weight. Feeding db/db mice MCD diets induces NASH and liver fibrosis within 4-8 weeks, providing a useful small-animal model for progressive obesity-related NAFLD. In this study, we tested if inhibiting hepatic DGAT2 expression by treatment with DGAT2 antisense oligonucleotide (ASO) would reduce hepatic steatosis and prevent development of NASH and liver fibrosis in db/db mice fed MCD diets.
We found that treatment with DGAT2 ASO successfully prevented hepatic steatosis but worsened liver injury and fibrosis. Interestingly, progressive liver damage occurred despite striking improvement in several parameters of systemic insulin resistance, but correlated with increased hepatic free fatty acid content, cytochrome P450 2E1 induction, and evidence of oxidative stress and lipid peroxidation. These findings suggest that in this model, triglyceride synthesis per se is not harmful to hepatocytes. Rather, it provides a useful mechanism for protecting the liver from lipotoxicity.
A mouse model of MCD diet–induced nonalcoholic steatohepatitis (NASH) was studied. Six-week-old male db/db mice (BKS.Cg-m+/+Leprdb/J) were purchased from Jackson Laboratories (Bar Harbor, ME), maintained in a temperature- and light-controlled facility, and permitted ad libitum consumption of water. Thirty-six mice were fed either the control diet (n = 6; cat no. 960441; ICN, Aurora, OH) or the MCD diet (n = 30; cat no. 960439; ICN, Aurora, OH) for 4 or 8 weeks. Half the MCD diet–fed mice were treated with 25 mg DGAT2 ASO/kg mouse body weight (Isis Phamaceuticals, Inc., Carlsbad, CA) intraperitoneally twice weekly; the remainder were injected with saline.10 All animal experiments fulfilled NIH and Duke University requirements for humane animal care.
Immunoblot was performed as described previously.13 Protein isolated from whole livers was loaded onto a 10% SDS-PAGE gel and transferred to PVDF membranes (Millipore, Billerica, MA). Membranes were probed with anti-Cyp2E1 (Stressgen, Victoria, Canada), anti-catalase (Abcam, Cambridge, MA) or β-actin (Sigma-Aldrich, St. Louis, MO) antibody followed by horseradish peroxidase (HRP)–conjugated antimouse or rabbit IgG (Amersham, UK). Antigens were demonstrated by ECL (Amersham, UK) and quantitated by AlphaEaseFC software (Alpha Innotech, San Leandro, CA).
Two-Step Real-Time RT-PCR.
Total RNA was extracted from whole livers with RNeasy kits (Qiagen, Valencia, CA), reverse-transcribed using random primer and Superscript RNase H-reverse transcriptase (Invitrogen, Carlsbad, CA), and analyzed by real-time PCR as previously described.14 Primers were designed by Primer Express Software (PE Applied Biosciences; Table 1). For all primer pairs, specificity was confirmed by sequencing PCR products. Target gene level is presented as a ratio of levels in the treated group versus the corresponding control group, according to the ΔΔCt method as previously reported.15 The magnitude of changes was determined using point and interval estimates.
Table 1. RT-PCR Primers for Analysis
Genbank accession number
Amplicon size (bp)
Immunohistochemistry and Analysis of Liver Architecture.
Serial sections were stained with H&E or Oil Red O using standard techniques. After deparaffinization, microwave antigen retrieval, and blocking endogenous peroxidase activity, other sections were incubated with a TdT-mediated dUTP-digoxigenin nick end labeling (TUNEL) reaction mixture containing terminal deoxyribonucleotidyl transferase (TdT) and fluorescein-dUTP (Roche Diagnostic, Indianapolis, IN), anti-α-smooth muscle actin (α-SMA; DakoCytomation, Carpinteria, CA), or anti-HNE-11s (Alpha Diagnostic International Inc., San Antonio TX) antibody. Antigens were demonstrated using a secondary antifluorescein antibody POD conjugate (Roche, Indianapolis, IN), antimouse or rabbit polymer HRP (Dako, Carpinteria, CA), and DAB chromagen (Dako, Carpinteria, CA), and a counterstaining with Gill's hematoxylin (Vector Laboratories). TUNEL-positive hepatocytes were quantitated in 3 randomly selected fields/section (20× magnification). The proportion of 4-hydroxy-2-nonenal (4-HNE)–positive hepatocytes was quantitated in 5 randomly selected fields/section (20× magnification) with Meta Morph software (Molecular Devices Corporation, Downington, PA) as previously described.16, 17
Quantification of Hepatic Collagen Content.
Liver sections were stained with picrosirius red (Sigma-Aldrich, St. Louis, MO) and counterstained with fast green (Sigma-Aldrich, St. Louis, MO). Sirius red staining was quantitated by Meta Morph software in 3 randomly selected fields/section (20× magnification). Hydroxyproline content was quantified colorimetrically. Freeze-dried liver specimens (20 mg) were hydrolyzed in 6N HCl at 110°C for 16 hours. After evaporation under vacuum, sediment was dissolved in distilled water, filtered, and incubated with chloramine-T (Sigma-Aldrich, St. Louis, MO) in acetate-citrate buffer and 50% isopropanol. To this was added Ehrlich's solution (4-dimethylaminobenzaldehyde in 60% perchloric acid with isopropanol, all from Sigma-Aldrich, St. Louis, MO). After incubating at 65°C for 15 minutes, absorbance was read at 561 nm. Hydroxyproline concentration was calculated from a standard curve prepared with high-purity hydroxyproline (Sigma-Aldrich, St. Louis, MO).
Tissue and Serum Biochemical Measurements.
Serum aspartate aminotransferase (AST), alanine aminotransferase (ALT), free fatty acid (FFA), glucose, insulin, and adiponectin levels were measured as previously described.10 Tissue FFA, triglyceride, and thiobarbituric acid-reactive substances (TBARS) were measured using the free fatty acid Half-Micro test (Roche, Indianapolis, IN), Triglyceride Detection Kit (Sigma-Aldrich, St. Louis, MO), and TBARS Assay kit (ZeptoMetrix Co., Buffalo, NY) according to the manufacturers' instructions.
The results are expressed as means ± SEMs. Significance was established using the Student t test and analysis of variance when appropriate. Differences were considered significant at P < 0.05.
An earlier study demonstrated that twice-weekly intraperitoneal injections of DGAT2 ASO effectively suppressed hepatic triglyceride accumulation in mice with genetic or diet-induced obesity, whereas treatment with DGAT1 ASO had no effect.10 In separate studies, we initially evaluated the effect of DGAT1 ASO in MCD diet–fed db/db mice and verified that DGAT1 inhibition also failed to improve hepatic triglycerides or reduce hepatic necroinflammation in our model (data not shown). Therefore, in the present study, we compared hepatic expression of DGAT2 in db/db mice that were fed MCD diets for 4 or 8 weeks while being treated with either saline or DGAT2 ASO. Results of both MCD-fed groups were compared to those in untreated db/db mice that were fed comparable methionine-sufficient diets (control). The MCD diet per se had no effect on hepatic expression of DGAT2 mRNA. As expected, treatment with DGAT2 ASO reduced DGAT2 mRNA expression by more than 90% at both time points (**P < 0.01; Fig. 1). As in the previous study,10 DGAT2 ASO had no effect on hepatic expression of DGAT1 (data not shown).
Inhibiting DGAT2 Reduced Hepatic Steatosis.
Because DGAT2 catalyzes the final step in triglyceride synthesis, inhibiting DGAT2 expression is expected to reduce hepatic steatosis. This was evaluated by H&E and Oil Red O staining. Steatosis was scored on H&E-stained sections according to Brunt's criteria,18 and photomicrographs from representative mice are shown (Fig. 2A). Control db/db mice had mild diffuse steatosis. MCD diets exacerbated this. Treatment with DGAT2 ASO attenuated the increases in hepatic steatosis that were caused by MCD diets, lowering the steatosis score from 4 to 1 at 4 weeks. These findings were confirmed by biochemical analysis of hepatic triglyceride content (Fig. 2B). Consumption of MCD diets more than doubled hepatic triglycerides after 4 weeks (**P < 0.01). Treatment with DGAT2 ASO significantly inhibited the early, MCD diet–induced accumulation of hepatic triglycerides (**P < 0.01 for MCD-DGAT2 ASO vs. MCD-saline groups). Interestingly, hepatic triglyceride content declined as the duration of MCD diet consumption was extended from 4 to 8 weeks. At the later time, liver triglyceride concentrations were similar in both groups of MCD-fed mice, suggesting that factors other than triglyceride synthesis determine hepatic triglyceride storage during prolonged methionine-choline deficiency.
Despite Decreased Steatosis and Lower TNFα Than MCD Controls, Mice Treated with DGAT2 ASO Had Worse Liver Injury.
Steatosis is thought to be a prerequisite for NASH and has also been identified as an independent risk factor for liver fibrosis. Thus, reducing hepatic steatosis was expected to protect db/db mice from MCD diet–induced hepatic inflammation and fibrosis. To address this issue, we compared various injury-related parameters among the 3 groups of mice. In many murine models of obesity and insulin resistance, as well as in humans with these conditions, tumor necrosis factor alpha (TNFα) has been implicated in the pathogenesis of NASH. MCD diets increased expression of TNFα mRNA in the livers of db/db mice (Fig. 3A). Treatment with DGAT2 ASO blocked this MCD diet–related induction of TNFα expression. As noted for hepatic steatosis, greater inhibition occurred after 4 weeks than after 8 weeks. Consistent with evidence that MCD diets increased hepatic steatosis and TNFα expression, saline-treated db/db mice that were fed MCD diets exhibited 6-fold higher serum ALT values than did db/db controls (Fig. 3B). Surprisingly, however, serum ALT values were even higher in MCD diet–fed mice treated with DGAT2 ASO. To determine if liver injury was, in fact, worsened by DGAT2 ASO treatment, liver sections were evaluated for lobular inflammation and hepatocyte death. Both the lobular inflammatory grade18 and the percentage of TUNEL-positive hepatocytes were higher in MCD diet–fed db/db mice than in db/db controls. The highest levels of lobular inflammation and hepatocyte death occurred in the MCD diet–fed group that received DGAT2 ASO (Fig. 3C–D). These results demonstrate that treatment with DGAT2 ASO enhanced liver injury despite reducing hepatic steatosis and TNFα expression.
Blocking Triglyceride Synthesis with DGAT2 ASO Increased Hepatic FFA Content, Induced Cyp2E1 Expression, and Increased Markers of Lipid Peroxidation/Oxidative Stress.
To determine why blocking triglyceride synthesis exacerbated liver damage, we compared liver FFA content, expression of nonmitochondrial fatty acid–oxidizing enzymes, and markers of lipotoxicity among the 3 groups of mice at the end of 4 and 8 weeks of treatment. Hepatic FFA content was similar in both groups of MCD diet–fed mice after 4 weeks, but it remained stable from 4 to 8 weeks in the MCD diet + saline group while doubling in the MCD diet–fed group treated with DGAT2 ASO. Thus, after 8 weeks, MCD diet–fed mice treated with DGAT2 ASO had significantly higher liver FFA content than did mice in the other groups (*P < 0.05; Fig. 4A). The higher hepatic FFA content in DGAT2 ASO–treated mice occurred despite these mice having the lowest serum FFA levels (*P < 0.05; Fig. 6B). Interfering with normal mechanisms for FA detoxification might induce other hepatic enzyme systems that oxidize FFA. Because MCD diets are known to inhibit mitochondrial oxidation of FFA,19 we evaluated expression of microsomal and peroxisomal FFA-oxidizing enzymes. Although catalase levels did not differ appreciably among the various groups, expression of Cyp2E1 was increased in both MCD diet–fed groups. Cyp2E1 levels were consistently greater in MCD diet–fed mice treated with DGAT2 ASO than in those that received saline (Fig. 4B). The db/db mice treated with DGAT2 ASO also demonstrated earlier increases in TBAR levels and had significantly more hepatic accumulation of 4-HNE at 8 weeks (Fig. 4C–D). Thus, inhibiting triglyceride synthesis with DGAT2 ASO increased hepatic FFA content, induced the microsomal FFA-oxidizing enzyme Cyp2E1, and exacerbated oxidative damage in MCD diet–fed db/db mice. Together, these findings suggest that triglyceride synthesis helps to protect the liver from lipotoxicity by buffering FFA accumulation.
Blocking Triglyceride Synthesis with DGAT2 ASO Did Not Prevent MCD Diet–Induced Fibrosis.
Hepatic steatosis is an independent risk factor for fibrosis in humans with chronic hepatitis C. Thus, although DGAT2 ASO–treated mice exhibited hepatic necroinflammation, it was conceivable that the DGAT2 ASO–mediated reduction in hepatic triglyceride content might have protected them from liver fibrosis. To address this issue, mRNA levels of various fibrosis markers were compared in the 3 groups of mice. Feeding db/db mice MCD diets significantly increased hepatic expression of transforming growth factor beta-1 (TGFβ-1), α-SMA, collagen, and tissue inhibitor of metalloproteinase 1 (TIMP-1). Treatment with DGAT2 ASO blunted the MCD diet–related induction of TGFβ-1, but failed to inhibit diet-related increases in collagen or TIMP-1 expression, and actually exacerbated the induction of α-SMA (Fig. 5A). To further assess the effects of MCD diets and DGAT2 ASO on liver fibrosis, hepatic hydroxyproline content was determined. As predicted from the fibrosis markers, MCD diets significantly increased hepatic hydroxyproline content in db/db mice. Interestingly, the highest liver hydroxyproline content occurred in the subgroup of MCD diet–fed mice that received DGAT2 ASO (Fig. 5B). The latter group also had the greatest area of Sirius red-stained fibrils demonstrated by liver morphometry (Fig. 5C). Thus, inhibiting hepatic triglyceride accumulation not only exacerbated liver necroinflammation and oxidative damage in MCD diet–fed db/db mice but also exacerbated liver fibrosis.
Liver Injury and Fibrosis Were Increased by MCD Diets Despite Reduced Obesity, Serum FFA, Glucose, and Insulin and Increased Adiponectin.
Previous studies have reported that steatohepatitis is associated with insulin resistance.4 Therefore, we investigated the impact of MCD diets and treatment with DGAT2 ASO on markers of systemic insulin sensitivity. db/db mice are genetically obese and severely insulin resistant.12 Feeding db/db mice MCD diets for 8 weeks significantly attenuated aging-related gains in body weight, but these mice remained obese. Adding DGAT2 ASO treatment to MCD diets further decreased body weight (Fig. 6A). Thus, both MCD diet–fed groups were less obese than the db/db controls. They also exhibited significant improvements in serum FFA, glucose, and insulin levels (Fig. 6B). Interestingly, serum adiponectin levels increased significantly in DGAT2 ASO–treated mice (Fig. 6B). Hence, MCD diets lessened obesity and systemic insulin resistance, despite exacerbating obesity-related hepatic necroinflammation and fibrosis. Moreover, although adiponectin is known to inhibit activation of hepatic stellate cells,20 DGAT2 ASO–related increases in serum adiponectin did not protect MCD db/db mice from activation of stellate cells or development of liver fibrosis.
Histology and various imaging techniques used clinically to diagnose and grade hepatic steatosis quantify hepatic triglyceride accumulation.18, 21 In humans, the severity (grade) of hepatic steatosis appears to correlate with stage of liver fibrosis.22 However, it remains unclear if (and how) hepatocyte triglyceride accumulation per se might incite necroinflammatory and fibrotic responses in the liver. Our results show that triglycerides themselves are not hepatotoxic, at least in mice with MCD diet–induced steatohepatitis. Rather, we demonstrate that in such mice, triglyceride synthesis actually helps to protect hepatocytes from lipotoxicity by buffering the accumulation of FFA.
FFA might evoke hepatocyte damage by several mechanisms. They are substrate for ROS-generating microsomal enzymes, such as Cyp2E1, and thus microsomal FA metabolism increases hepatocyte ROS production.23, 24 Oxidized FA, themselves, can also catalyze lipid peroxidation reactions that are directly cytotoxic.25, 26 Finally, because certain FA function as endogenous ligands for nuclear hormone receptors, such as the peroxisomal proliferator–activated receptors and hepatocyte nuclear factor 4-alpha, their accumulation might affect global changes in liver gene expression.27 We found that inhibiting triglyceride synthesis with DGAT2 ASO increased hepatic FFA content, up-regulated Cyp2E1, exacerbated hepatic oxidative damage, and increased hepatocyte death, liver inflammation and fibrosis in MCD diet–fed mice. Hence, liver damage accrued from the “invisible” lipids rather than from the “visible” fat (i.e., triglycerides) that was the basis for a diagnosis of NAFLD.
At first glance, our results differ from those of Yu et al., who reported that inhibiting triglyceride synthesis with DGAT2 ASO improved fatty liver disease in ob/ob mice and mice with high-fat diet-induced obesity.10 As do many obese, type 2 humans with diabetes, ob/ob mice, db/db mice, and mice with diet-induced obesity readily develop hepatic steatosis, but seldom experience progressive liver damage or fibrosis. The latter lesions are thought to require the imposition of secondary “hits” that induce sufficient oxidative and/or endoplasmic reticulum stress to incite hepatic inflammatory cell infiltration and stellate cell activation.8, 28 In obesity and insulin resistance, the major mechanisms driving hepatic triglyceride accumulation are increased delivery of FFA from peripheral adipose depots to liver and enhanced de novo synthesis of FA and triglyceride in the liver itself. Hepatic lipid disposal via mitochondrial beta oxidation and lipoprotein export is relatively intact.29, 30 Thus, the liver retains mechanisms to rid itself of potentially toxic FFA, even when triglyceride biosynthesis is inhibited by blocking DGAT2. Hence, reducing hepatic steatosis by inhibiting hepatic triglyceride synthesis is tolerated without appreciable acute hepatotoxicity.
MCD diets induce hepatic triglyceride accumulation by inhibiting mitochondrial beta-oxidation of fatty acids and blocking hepatic export of very-low-density lipoprotein.19 These inhibitory effects on lipid disposal are sufficient to cause hepatic triglyceride accumulation in lean mice. Interestingly, unlike high-fat diets, which produce relatively little hepatic necroinflammation or fibrosis, MCD diets are widely used to induce NASH with fibrosis. Although MCD diets inhibit mitochondrial oxidation of FA, thereby limiting that process as a source of ROS, these diets induce microsomal FA oxidizing enzymes, such as Cyp2E1. The latter may produce sufficient ROS to trigger the secondary “hits” that are required for progression from NAFLD to more advanced stages of fatty liver disease.28 Thus, MCD diets promote NASH with fibrosis, even in normal mice.
In db/db mice, in which genetic insulin resistance and obesity drive increased delivery of FFA to liver, as well as de novo hepatic FA synthesis,31 a study demonstrated that MCD diets dramatically exacerbate steatosis.11 In that study, MCD diets also promoted rapid progression from steatosis to NASH and liver fibrosis. Moreover, liver damage related to methionine-choline deficiency was worse in db/db mice than in wild-type controls. Thus, damage that resulted from inhibiting mitochondrial oxidation of FA and reducing hepatic lipoprotein release was worse in the db/db livers coping with increased FFA exposure. When we blocked the terminal enzyme in triglyceride synthesis by treatment with DGAT2 ASO, another major mechanism for FA disposal (i.e., triglyceride synthesis) was removed, and this exacerbated MCD diet–induced liver damage, despite reducing hepatic triglyceride content.
These results suggest that confusion about the natural history of fatty liver results, at least in part from the current diagnostic criteria for hepatic steatosis. The latter rely on static quantification of hepatic triglyceride content by histology or liver imaging. However, this approach ignores the importance of FA flux through various hepatocellular compartments, some of which are harmless (e.g., incorporation into trigylcerides), whereas others are potentially toxic (e.g., microsomal oxidation). Hence, the implications of hepatic steatosis (i.e., triglyceride accumulation) vary widely depending on the metabolic context in which this occurs. A low triglyceride content might reflect low hepatic FFA exposure and thus a low demand for triglyceride synthesis (good) or an inability to up-regulate triglyceride synthesis sufficiently to detoxify accumulating FFA (bad). Extension of this logic also helps to explain why peripheral insulin resistance has been linked to progressive fatty liver disease. Unlike db/db mice, some insulin-resistant individuals are apparently unable to compensate for impaired hepatic triglyceride synthesis by enhancing insulin signaling in peripheral adipose depots so that fatty acids are retained in fat. Thus, heir livers continue to be flooded with fat-derived FFAs, further stressing already-compromised hepatic mechanisms for FFA disposal and thereby exacerbating lipotoxicity.
Indeed, our findings in db/db mice suggest that blocking hepatic FA detoxification provides a strong stimulus for enhancing insulin sensitivity in peripheral tissues. The db/db mice are genetically resistant to insulin because they inherit leptin receptor mutations that cause resistance to leptin, a potent insulin-sensitizing hormone.32 Nevertheless, such mice become extremely sensitive to insulin when acquired insults (e.g., methionine-choline deficiency) limit efficient hepatic disposal of FA. Knocking down one of their remaining mechanisms for hepatic FA detoxification by blocking hepatocyte triglyceride synthesis with DGAT2 ASO intensifies the stimulus for peripheral insulin sensitivity. Thus, db/db mice treated with DGAT2 ASO while being fed MCD diets exhibited the least insulin resistance in our study. Further research is needed to identify the signals that mediate peripheral tissue sensitization to insulin in this setting. Adiponectin might be involved in this process because we found increased serum adiponectin levels in the group of db/db mice that received an MCD diet + DGAT2 ASO treatment. On the other hand, leptin appears to be relatively unimportant, because profound improvements in systemic insulin resistance were observed in the latter group despite genetic leptin resistance. Whether leptin plays a more significant role in humans who develop acquired hyperleptinemia because of obesity requires further study.
In summary, by progressively inhibiting different FA disposal mechanisms in a homogenous group of inbred obese mice with type 2 diabetes, we quickly and reproducibly converted simple steatosis to NASH and NASH to NASH with fibrosis. Progression of liver damage occurred despite striking improvements in systemic insulin resistance, demonstrating that hepatic lipotoxicity was the predominant factor driving liver damage in this model. Moreover, as the severity of hepatic lipotoxicity increased, systemic insulin resistance actually improved, suggesting that failure of hepatic FA detoxification stimulated compensatory sensitization to insulin in peripheral tissues. These findings suggest that variations in the hepatic capacity to detoxify FA, coupled with differences in the ability to respond to hepatic FFA retention by increasing insulin sensitivity in peripheral tissues, significantly influence the progression of liver damage in obesity-related NAFLD.