Lipidomic dissection of nonalcoholic steatohepatitis: Moving beyond foie gras to fat traffic


  • Nathan M. Bass

    Corresponding author
    1. Division of Gastroenterology, University of California, San Francisco, San Francisco, CA
    • Box 0538, Division of Gastroenterology, UCSF, 513 Parnassus Avenue, Room 357S, San Francisco, CA 94143-0538
    Search for more papers by this author
    • fax: 415-476-0659

  • See Article in Volume 50, Issue 6, Page 1827

  • Potential conflict of interest: Nothing to report.

In a rather remarkable letter to HEPATOLOGY almost 20 years ago, a proposed role of hepatocyte swelling from extensive triglyceride accumulation as a cause of elevated portal pressure was challenged.1 To make their point, the authors studied portal hemodynamics in ducks used to produce the delicacy of foie gras de canard. With their massive, iatrogenic burden of hepatocellular triglyceride, the livers were dramatically steatotic and enlarged, but the measured sinusoidal pressures were normal. In addition, the state of feeding-induced fat accumulation did not result in chronic necroinflammation or fibrosis that might have produced a chronic abnormality in structure and hemodynamic integrity. The moral of this story is that appearances can be misleading; the most morphologically dramatic histological feature of a disease is not necessarily the proximate cause of the problem in hand. So, it appears, is the case in general regarding triglyceride, the most pathologically visible form of hepatic fat accumulation in nonalcoholic fatty liver disease (NAFLD) and the development of nonalcoholic steatohepatitis (NASH).


16:0, palmitic acid; 16:1 n7, palmitoleic acid; 18:1 n7, vaccenic acid; 18:0, stearic acid; 18:1 Δ9, oleic acid; COX, cyclooxygenase; DGAT2, acyl-coenzyme A:diacylglycerol acyltransferase 2; DHA, docosahexaenoic acid; DPA, docosapentaenoic acid; ER, endoplasmic reticulum; FFA, free fatty acids; HETEs, hydroxyeicosatetraenoic acids; LOX, lipoxygenase; MCD, methionine-choline-deficient; MUFA, monounsaturated fatty acids; NAFL, nonalcoholic fatty liver; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; PPARα, peroxisome proliferator activated receptor alpha; PUFA, polyunsaturated fatty acids; Δ9 SCD, Δ 9 steroyl-CoA desaturase; SCD1, steroyl-CoA desaturase-1; SFA, saturated fatty acids; VLDL, very-low density lipoprotein.

Hepatocellular steatosis has defined the histology of NASH and has not only provided a name for this disease, but has also directed thinking in terms of its pathogenesis (see reviews by Carter-Kent et al.,2 Schreuder et al.,3 Parekh et al.,4 and Neuschwander-Tetri5). The quest for understanding the mechanisms whereby the accumulation of fat in hepatocytes progresses to steatohepatitis and ultimately, in some individuals, to advanced fibrosis, has identified insulin resistance and resultant hyperinsulinemia—usually associated with obesity—as responsible for the promotion of increased traffic of fatty acids into the hepatocyte. This traffic arises from three sources: excessive postprandial lipolysis within an enlarged adipocyte mass (the major contributor), de novo lipogenesis in the hepatocyte, and dietary fat.6, 7 The response of the liver to this situation is to increase triacylglycerol ester (triglyceride) synthesis from fatty acids and glycerol. In the absence of very low density lipoprotein secretion that can match this flux by exporting triglyceride to the circulation, readily visible droplets of triglyceride accumulate in hepatocytes. In earlier reflections, the idea that the triglyceride droplets provide a substrate for lipid peroxidation and oxidative stress leading to hepatocyte injury, inflammation, cell death, and fibrogenesis was the basis for the acceptance of the conceptually sequential “two hit” hypothesis into the theoretical canon of the pathogenesis of NASH.8 New evidence has, not surprisingly, made it clear that there is more going on than meets the eye, and that triglyceride may be less the villain of the piece than a marker of the presence of more nefarious molecular actors. Indeed, the main role of hepatic triglyceride in fatty liver diseases may be an adaptive and protective one in the setting of a more dynamic but histologically invisible traffic of toxic fatty acids and their derivatives (Fig. 1). In favor of this argument, it is amply recognized that histological triglyceride may fade from the liver during the progression of NASH to cirrhosis,9 whereas elegant experimental studies have shown that pathways that direct fatty acids to triglyceride may protect against liver injury. Conversely, manipulations that attenuate triglyceride biosynthesis in the face of increased fatty acid flux can lead to decidedly worse damage. The expression of steroyl-coenzyme A (CoA) desaturase-1 (SCD1, the murine homolog of human Δ 9 steroyl-CoA desaturase), the enzyme responsible for the conversion of saturated fatty acids (SFA), e.g., palmitic acid and stearic acid, to their monounsaturated fatty acid (MUFA) derivatives, palmitoleic acid and oleic acid, respectively, which are then preferentially incorporated into triglycerides, is increased in experimental steatosis resulting from feeding a high fat diet.10 Steatosis is also produced in mice after feeding a methionine-choline–deficient (MCD) diet, but in mice lacking SCD1, this diet results in less steatosis but markedly increased hepatocellular apoptosis, liver injury, and fibrosis.10 The greater cytotoxic impact of SFA compared to MUFA in inducing hepatic toxicity and apoptosis is well-recognized.11, 12 These findings thus point to an important role for SCD1 in partitioning the fatty acids in the liver between more easily esterified MUFA and more toxic—and less readily disposed—SFA.10 Further support for the role of fatty acid traffic directed away from triglyceride synthesis in producing liver damage in NASH comes from the db/db leptin-resistant mouse model. Feeding an MCD diet to these obese, diabetic animals induces NASH and liver fibrosis.13 Inhibition of acyl-coA:diacylglycerol acyltransferase 2 (DGAT2), the enzyme that catalyzes the final and only commited step in triglyceride synthesis, while feeding these mice an MCD diet results in improvement in hepatic steatosis, but worsening in oxidative stress, liver injury, and fibrosis.13

Figure 1.

Hepatic pathways of fatty acid metabolism that contribute to the plasma lipidome. The diagram illustrates selected major pathways of fatty acid flux and metabolism in the liver and the key hepatic lipid precursors and products that constitute the plasma (circulating) lipidome. Abbreviations: 16:0, palmitic acid; 16:1 n7, palmitoleic acid; 18:1 n7, vaccenic acid; 18:0, stearic acid; 18:1 Δ9, oleic acid; COX, cyclooxygensase; DGAT2, acyl-coA:diacylglycerol acyltransferase 2; ER, endoplasmic reticulum; FFA, free fatty acids; HETEs, hydroxyeicosatetraenoic acids; LOX, lipoxygenase; MUFA, monounsaturated fatty acids; Δ9 SCD, Δ 9 steroyl-CoA desaturase (SCD); SFA, saturated fatty acids; VLDL, very-low density lipoprotein.

In considering the role of fatty acids as fundamental mediators of lipotoxic liver injury, there are more details to consider with respect to which among the diverse family of fatty acid acyl and ester derivatives might be most at fault. Although substantial evidence, including that noted above, point to SFA playing a role,10, 11 it is important to consider that fatty acids give rise to plethora of biologically active metabolites that includes a range of desaturated products, elongation products, diacylglycerides, phospholipids, ceramides, sphingolipids, and eicosanoids, all of which may impact in a deleterious—and, in some cases, protective—fashion on cell membrane integrity, pathways of cell signaling and viability, and the generation of stress and reactive oxygen species within the endoplasmic reticulum and mitochondria.11, 12, 14

It is in this complex context of the role of fatty acids and their derivatives and the derangements that may occur in their hepatic formation and disposal in NASH that we must view the study from Puri et al.15 in this issue of HEPATOLOGY. The approach used in this study was to quantitate the absolute and relative amounts of different lipid classes in the plasma, including free fatty acids, triglycerides, phospholipids, and eicosanoids, and compare the distribution of fatty acids within each of these lipid classes in normals, lean controls, patients with histological simple hepatic steatosis (NAFL) and patients with NASH. The authors' intention was, in this manner, to characterize a comprehensive, disease stage–specific plasma lipid signature or “lipidome”, that might yield clues to the derangement of lipid traffic and metabolism and the role of specific metabolites and pathways in the pathogenesis of NASH.

In an earlier work, the authors characterized the lipidomic profile in liver biopsy tissue in normal controls and in patients with NAFL and NASH.16 In that study, there were important and even surprising observations. Total free fatty acids were no different in concentration among the three groups, whereas free cholesterol, another potential hepatic lipotoxin, showed a stepwise increase from normal through NAFL to NASH livers. Numerous other NAFL/NASH-related changes were observed in absolute and relative tissue levels of phospholipids and polyunsaturated fatty acids (PUFA), the meaning of which still await explanation.

In the present study, the authors have applied the same sophisticated technology to the quantitation and characterization of the plasma lipidome across the spectrum of lean normal controls and patients with NAFL and NASH. The yield of data from this exercise, given the sheer number of measurable species and metabolites, is, as expected, prodigious, and summarizing these in a succinct manner presents a robust challenge. In the first instance, the authors show that differences do indeed exist along the spectrum of normal through NASH in serum lipidomic changes, in some regards even more pronounced than seen in the earlier liver tissue measurements. Overall, the findings support the presence of increased lipogenesis, Δ 9 steroyl-CoA desaturase (Δ9 SCD) activity (with an increased MUFA:SFA ratio) and triglyceride synthesis, and raise the question, given the higher degree of toxicity attributed to SFA compared to MUFA,11, 12 whether these changes constitute evidence of an adaptive response to increased fatty acid traffic in NAFLD. Based on a stepwise increase in relative abundance of hydroxyeicosatetraenoic acids (HETEs) from normal through NASH with little change in the levels of cyclooxygenase-generated prostaglandins, the data also suggest a relative activation of the lipoxygenase (LOX) pathway, and also favor overproduction of the proinflammatory (e.g., 15-HETE) rather than anti-inflammatory (e.g., lipoxin) products of LOX. This observation lends support to the existence of a state of increased proinflammatory lipid production in NASH. Of note, the data also provide new evidence for a global impairment of peroxisomal function maximally expressed in subjects with NASH that includes an increase in plasma docosapentaenoic acid (DPA) with a decrease in docosahexaenoic acid (DHA, which is synthesized in peroxisomes from DPA) and lowered levels of plasmalogen lipids, that are also synthesized in peroxisomes. Indeed, an alternative explanation for the increases measured in HETE levels may be impaired catabolism of these fatty acids via the peroxisomal pathway as well.17 Peroxisomal β-oxidation of fatty acids has been viewed as a peroxisome proliferator activated receptor α–activated protective response to fatty acid overload in the liver.14, 18, 19 Thus, these findings suggest the possibility of a stunted protective response in NASH to extant lipotoxic conditions.

As an exploratory undertaking alone, the work is impressive, but the authors also go beyond the complex phenomenology of their findings and devote considerable thought toward their interpretation, using expository illustrations to summarize the shifts in pathways implied by the differences in circulating lipid levels. They conclude their article with a novel summary diagram deriving a model of the “circulating lipidome”, essentially a proposal for the fundamental changes in lipid metabolism that may occur as the liver transitions from normalcy, through NAFL, to NASH. The model incorporates interesting and useful concepts and ultimately a complex hypothesis that can now be tested on many levels, using both human and animal experimental approaches. The main caveats in setting up such a complex edifice based on the results of their study is that, at this point, it is unclear how much of the lipidomic changes that occur during the evolution of NAFLD are important in promoting the progression of disease as opposed to secondary phenomena or even epiphenomena of little consequence. However, several of the findings as well as existing knowledge suggest that the altered lipid profiles may indeed be important in pathogenesis, such as the increase in proinflammatory eicosanoids with disease progression, whereas others point to adaptive responses that may provide some protection against lipotoxicity. Still others point to potential protective mechanisms that may be failing to respond adequately to this threat. What is certain is that we now have a plethora of new data to fuel productive hypotheses for the ongoing investigation of a complex and fascinating disease.

It is noteworthy that the value of lipidomic studies in the study of NAFLD lies not only in the potential of this approach for providing a better understanding of the pathogenesis of this disease. The lipidomic signature of NASH also requires further evaluation as a biomarker for the noninvasive identification of patients who either have, or are at risk for developing, progressive liver injury and fibrosis. Finally, this powerful technology bears promise for illuminating molecular targets and pathways that may offer opportunities for therapeutic intervention aimed at alleviating the damage resulting from lipotoxic traffic in NAFLD.