Fetal obesity syndrome: Maternal nutrition as a cause of nonalcoholic steatohepatitis


  • Michael Charlton M.D., F.R.C.P.

    Corresponding author
    1. Division of Gastroenterology and Hepatology, Mayo Clinic, Rochester, MN
    • Department of Gastroenterology and Hepatology, Mayo Clinic, 200 First Street S.W., Rochester, MN 55905
    Search for more papers by this author
    • fax: 507-266-1856.

  • See Article on Page 1796

  • Potential conflict of interest: Nothing to report.

Having children has its blessings and its downsides. The blessings are usually obvious and typically greatly outweigh the downsides. The downsides, however, are not trivial, and they are sometimes unexpected. My son was recently suspended from his soccer team for a verbal exchange with an opposing coach, which revolved principally around the alacrity of what the coach was saying to the referee (we have a school system where it can be heroic to question Darwin, but the opinion of a soccer coach is strictly off limits). My son should have known better, but rather than apologize or express remorse, as he came off the field he blamed his altercation and suspension on me for passing on the genes of my famously mercurial father. It was tempting to point out that if he had the time to bemoan the parts of his genetic code that had made it impossible for him to suffer the petty indignities of school sports in silence, he should also consider being more grateful for those genes that were responsible for him having a total of 20 digits and two functioning eyes. As I struggled to suppress a reflexive apoplexy it occurred to me that, at some level, he had been right. Just as children are a mixed blessing, so are parents. If anything, my son had understated the biological impact of parents on their children.


BMI, body mass index; ETC, electron transport chain; FFA, free fatty acid; HF, high fat; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; NHANES, National Health and Nutrition Education Survey; TGF-β1, transforming growth factor-beta 1; TNF-α tumor necrosis factor alpha; TRAIL, tumor necrosis factor–related apoptosis-inducing ligand.

The rising global prevalence of obesity has been well documented. The World Health Organization estimates that, worldwide, 1.6 billion adults are overweight (defined as body mass index [BMI] >25 kg/m2) and 400 million are obese (BMI >30 kg/m2). According to the most recent National Health and Nutrition Education Survey (NHANES), two-thirds of adults in the United States are overweight or obese.1 There are more than 1 million Americans with a BMI > 60 kg/m2 (NHANES III data estimates).1 The average adult gained ≈10 pounds in the 1990s. Although the overall prevalence of obesity has recently begun to plateau, the prevalences of the most severe classes of obesity are rising sharply, with a dramatic shift to the right in BMI distribution. Class III obesity (BMI ≥40 kg/m2) quadrupled and BMI ≥50 kg/m2 quintupled2 between 1986 and 2000. Mercifully, although the prevalence and severity of obesity are high, only a small proportion of obese individuals develop nonalcoholic steatohepatitis (NASH) with progressive fibrosis. A conservative estimate of the frequency of nonalcoholic fatty liver disease (NAFLD) as the underlying cause of liver disease among liver transplant recipients is 5%-10% (equivalent to 350-700 total patients, http://www.ustransplant.org/)—a tiny fraction of patients with NAFLD. The basis for disparate histological findings seen in the livers of patients with obesity is one of the great scientific questions facing academic hepatologists.

The physiology of steatosis is no mystery. Simple overnutrition can produce a net accumulation of hepatic fat (just order foie gras). Similarly, oxidative stress is an inevitable byproduct of metabolizing carbohydrates and lipids.3 The concept and some of the mechanisms of lipoapoptosis, a consequence of lipid-induced oxidative injury, are increasingly well established. Intrahepatic free fatty acids (FFAs) are potent inducers of the apoptosis receptors Fas and tumor necrosis factor–related apoptosis-inducing ligand (TRAIL).4 Overnutrition, even with simple carbohydrate excess, sensitizes hepatocytes to apoptosis through increased expression of Fas.5 Further contributing to oxidative stress and injury, intrahepatic FFA excess results in mitochondrial dysfunction associated with release of cathepsin B from lysozomes.6 Cathepsin B knockout mice are protected against overnutrition-associated oxidative hepatocellular injury, suggesting a possible mechanistic role for cathepsin B in steatohepatitis. In addition, reactive oxygen species mediate release of tumor necrosis factor alpha (TNF-α) by Kupffer cells and hepatocytes.7 TNF-α increases mitochondrial permeability, impairs mitochondrial respiration, and causes depletion of mitochondrial cytochrome c and further release of cathepsin B. Both TNF-α–induced caspase activation and hepatocyte death (apoptosis) are increased in NAFLD.8 Finally, lipid peroxidation results in the production of malondialdehyde and 4-hydroxynonenal, which serve as chemoattractants for neutrophils (necroinflammation), stimulate hepatic stellate cells (fibrosis), and up-regulate transforming growth factor beta 1 (TGF-β1) expression in macrophages (fibrosis).9 And yet, for all of the mechanistic insights into the physiologic basis of steatohepatitis, most patients with decades of medically complicated obesity, with or without insulin resistance, do not have clinically meaningful amounts of hepatic fibrosis. We have barely a clue about the basis of steatofibrosis.

Although the physiological basis of susceptibility to lipotoxicity is not known, there is emerging evidence of heritable/genetic susceptibility to steatosis associated with elevated aminotransferases. A genomewide association survey in the ethnically diverse individuals participating in the Dallas Heart Study found that hepatic fat content is more than two-fold higher in PNPLA3 (adiponutrin) rs738409[G] homozygotes than in noncarriers,10 suggesting an ancestry-related basis for interindividual differences in hepatic fat content and susceptibility to NAFLD. PNPLA3 has transacetylase activity, which catalyzes triglyceride synthesis in adipocytes,11 and is up-regulated by insulin12 and carbohydrate refeeding,13 suggesting a lipogenic effect in the liver. In a recent analysis, the rs738409[G] allele in PNPLA3 was associated with increased quantitative measures of liver fat content and serum aspartate aminotransferase concentrations independently of age, sex, and BMI.14 Although the association of rs738409[G] allele in PNPLA3 with steatosis and elevated aminotransferases is very exciting, about half of individuals with the rs738409[G] allele have entirely normal liver fat content, highlighting the polygenic nature of NAFLD phenotypes. In addition, there are no reports of histological associations of the rs738409[G] allele, e.g., with hepatic inflammation or fibrosis. Our genotype is thus highly unlikely to explain the basis of steatohepatitis and subsequent fibrosis. Our parents, however, are far from off the hook.

In this issue of HEPATOLOGY, Bruce et al.15 present intriguing new evidence that we may be able to blame our parents. Bruce et al. fed female mice either a high fat (HF) or control chow (C) diet prior to and during gestation and lactation. The resulting offspring were either fed a C or HF diet postweaning to generate four offspring groups: HF/HF, HF/C, C/HF, C/C. At 15 weeks of age, liver histology revealed that whereas the C/HF offspring developed simple steatosis, the HF/HF offspring developed NASH. At 30 weeks, histological analysis and Kleiner scoring revealed that both the HF/C and C/HF groups had NAFL, whereas the HF/HF group had a more severe form of NASH. The authors reasonably surmise that exposure to an HF diet in utero and during lactation contributes to NAFLD progression to NASH. Some mechanistic insights were also produced. At 15 weeks of age, hepatic mitochondrial electron transport chain (ETC) enzyme complex activity (I, II/III, and IV) was reduced in both groups of offspring from HF-fed mothers (HF/C and HF/HF). In addition, measurement of hepatic gene expression indicated that lipogenesis, oxidative stress, and inflammatory pathways were up-regulated in the 15-week-old HF/C and HF/HF offspring. The authors conclude that maternal fat intake contributes to the development of NASH in adult offspring, an effect that is mediated through impaired hepatic mitochondrial metabolism and up-regulated hepatic lipogenesis.

These are potentially highly significant findings. To many, the concept of maternal behavior affecting offspring in adulthood may seem unlikely. It shouldn't. Metabolic programming in utero is well established for other complications of the metabolic syndrome in humans, such as coronary artery disease.16–21 Forsen et al.16 demonstrated that, in multivariate analysis, maternal obesity in women who are short of stature is strongly associated with coronary heart disease in adult offspring. The effect is most pronounced in the offspring of women who have a history of chronic malnutrition who subsequently become obese before pregnancy, which may explain the disproportionately high prevalence of metabolic syndrome among obese individuals in emerging economies, such as India.22, 23 A physiological basis for this observation was made by Mitrani et al.,24 who observed that metabolic programming of hyperinsulinemia throughout adulthood of offspring of obese rats produces changes in autonomic activity, including increased parasympathetic and decreased sympathetic activity, in the postnatal maintenance of hyperinsulinemia in response to the high caloric diet in utero. The article by Bruce et al. in this issue of HEPATOLOGY yields several new insights that suggest that NAFLD may not be just part of what I will call fetal obesity syndrome, but NAFLD may be part of the cause of the metabolic consequences of maternal obesity. The first mechanistic insight is that pathways of oxidative stress and inflammation, with associated mitochondrial dysfunction (potentially a cause and effect of oxidative stress), persist into murine adulthood, suggesting an additional mechanism. Bruce et al. observed that although mitochondrial numbers were similar between study groups, mitochondrial function, as measured by ETC activity, was impaired in the offspring of HF/HF dams. Impaired mitochondrial ETC function will have made the liver unable to increase beta oxidation of excess intracellular acyl-coenzyme A, leading inevitably to increased triglyceride synthesis. Our ability to correct mitochondrial protein and mitochondrial DNA defects is very small. Damage done in utero may thus persist for a lifetime. These data suggest that not only is it possible to have your mother's ankles, you may well also have her liver.

Studies that include genomics usually provoke a tedious list of caveats. I am not going to bother. The study by Bruce et al., although preliminary by virtue of the results being produced, by necessity, in a murine model rather than in humans, should make us begin to rethink our approach to NAFLD. These results and an increasing number of others in humans and animals suggest that risk for NAFLD and the metabolic syndrome seem likely to begin in utero and to be present at birth. Genes and behavior (and probably environmental obesogens25) are undoubtedly important too. I am not advocating requiring a pregnancy test before getting served at McDonalds, although I hope and expect the cheap calorie industrial complex to join cigarette manufacturers on the list of modern human scourges. The study by Bruce et al., however, adds to the body of evidence that, in total, is enough for us to educate women of the possible consequences of overnutrition and obesity during pregnancy. In addition to the needed mechanistic and prevention studies, we should, in my opinion, begin to consider earlier identification and intervention of individuals at risk of NAFLD. Earlier, as in the maternity ward.